Transduction Channels in Sensory Cells

Transduction Channels in Sensory Cells Edited by S. Frings and J. Bradley Transduction Channels in Sensory Cells. Edit...

Author: Stephan Frings | Jonathan Bradley

163 downloads 1147 Views 6MB Size Report

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

DOWNLOAD PDF

Transduction Channels in Sensory Cells Edited by S. Frings and J. Bradley

Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

Further Titles of Interest M. Futai, Y. Wada, J. Kaplan (eds.)

Handbook of ATPases – Biochemistry, Cell Biology, Pathophysiology 2004

ISBN 3-527-30689-7

C. M. Niemeyer, C. A. Mirkin (eds.)

Nanobiotechnology – Concepts, Applications and Perspectives 2004

ISBN 3-527-30658-7

Y. Yawata

Cell Membrane – The Red Blood Cell as a Model 2003

ISBN 3-537-30463-0

G. Krauss

Biochemistry of Signal Transduction and Regulation, 3rd edition 2003

ISBN 3-527-30591-2

O. von Bohlen und Halbach, R. Dermietzel (eds.)

Neurotransmitters and Neuromodulators 2002

ISBN 3-527-30318-9

Transduction Channels in Sensory Cells Edited by Stephan Frings and Jonathan Bradley

Prof. Dr. Stephan Frings University of Heidelberg Department of Molecular Physiology Im Neuenheimer Feld 230 69120 Heidelberg Germany [email protected]

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Dr. Jonathan Bradley Johns Hopkins University School of Medicine Department of Neuroscience 725 N. Wolfe St. Baltimore, MD 21205 USA [email protected]

Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data: A catalogue record for this book is available from the British Library. Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at http://dnb.ddb.de ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim All rights reserved (including those of translation in other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by lax. Cover Illustration Hirmer Verlag, Mu¨nchen Composition Mitterweger & Partner GmbH, Plankstadt Printing betz-druck GmbH, Darmstadt Bookbinding Großbuchbinderei J. Scha¨ffer GmbH & Co. KG, Gru¨nstadt Printed in the Federal Republic of Germany. Printed on acid-free paper ISBN 3-527-30836-9

V

Table of Contents Preface XIII List of Contributers 1

1.1 1.2 1.3 1.4 1.4.1 1.4.2 1.4.2.1 1.4.2.2 1.4.3 1.4.3.1 1.4.3.2 1.4.4 1.4.4.1 1.4.4.2 1.4.4.3 1.4.5 1.4.6 1.4.6.1 1.4.6.2 1.4.6.3

XV

The Molecular Basis of Touch Sensation as Modeled in Caenorhabditis elegans 1 Laura Bianchi and Monica Driscoll

Abstract 1 Introduction 2 Features of the C. elegans Model System 3 Mechanosensation Is a Major Mechanism by Which C. elegans Senses Its Environment 4 Gentle Body Touch 5 The Touch Receptor Neurons 5 Ultrastructural Features of the Touch Receptor Neurons 5 Touch Cell-specific Microtubules 5 The Extracellular Mantle 6 Genetic and Molecular Analysis of Body Touch 7 mec-4 and mec-10 Ion Channel Subunits Form Na+ Channels 7 MEC-4 at the Molecular Level 7 The Candidate Mechanotransducing Channel is a Heteromultimeric Complex 9 MEC-4 and MEC-10 Form a Functional Ion Channel 10 MEC-2 Is a Stomatin-like Protein That May Help Tether the MEC-4/ MEC-10 Channel to the Membrane Bilayer and/or the Cytoskeleton 10 MEC-6 Is a Transmembrane Paraoxonase-like Protein That Controls MEC Channel Activity 11 Intracellular Proteins Needed for Touch Transduction 13 Extracellular Proteins Needed for Touch Transduction 14 MEC-1 14 MEC-5 15 MEC-9 15

VI

Table of Contents

1.4.7 1.4.7.1 1.4.7.2 1.4.8 1.4.8.1 1.4.8.2 1.5 1.5.1 1.5.2 1.5.2.1 1.5.2.2 1.5.2.3 1.6

The MEC Channel Functions Specifically in Neuronal Responses to Gentle Touch 16 Test of a Key Hypothesis 16 Additional Insights Revealed by Imaging In Vivo Ca2+ Changes in Responding Touch Neurons 18 Summary: A Molecular Model for Gentle-touch Sensation 19 How Touch Is Sensed to Elicit a Specific Behavioral Response 19 Notes on the Working Model 19 The C. elegans Degenerin Family: A Global Role of Degenerin Channels in Mechanotransduction? 20 unc-105 20 unc-8 and del-1 21 A Stomatin Partner for the UNC-8 Channel Suggests a Common Composition of Degenerin Channels 21 Trp Channels May Also Contribute to Mechanosensory Functions in C. elegans 23 Fly and Mouse Neuronal DEG/ENaCs Influence Mechanotransduction, Supporting Conserved Roles for This Family of Proteins 24 Concluding Remarks 25 Acknowledgments 26 References 26

2

Transduction Channels in Hair Cells Robert Fettiplace

2.1 2.2 2.2.1 2.2.2 2.2.3 2.3 2.3.1 2.4 2.4.1 2.4.2 2.5 2.5.1 2.6 2.6.1 2.7

Introduction 31 Gating Mechanism: Channel Kinetics 32 Tip Links and Gating Springs 34 Gating Compliance 36 Three-state Channel Schemes 36 Ionic Selectivity 38 Blocking Compounds 40 MET Channel Adaptation 41 Ca2+ Regulation of Adaptation 42 The Function of Adaptation 44 Single-channel Conductance 45 Number of MET Channels Per Stereocilium 47 The MET Channel as a Member of the TRP Family Properties of TRPV Channels 49 Conclusions 49 Acknowledgments 51 References 51

31

48

Table of Contents

3

Acid-sensing Ion Channels 57 Kenneth A. Cushman and Edwin W. McCleskey

3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.7.1 3.7.2 3.7.3 3.8 3.9 3.10

Introduction 57 ASICs and the DEG/ENaC Superfamily Amino Acid Structure 61 Assembly Into Channels 61 Pharmacology 62 Gating 63 Proposed Sensory Functions 65 Pain/Nociception 65 Mechanosensation 66 Taste 67 CNS ASICs 67 Stroke 68 Other pH-activated Channels 68 References 69

4

Chemosensory Transduction in Caenorhabditis elegans 73 Noelle LEtoile

4.1 4.1.1 4.1.2 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.3 4.3.1 4.3.2 4.3.3 4.3.4 4.4 4.4.1 4.4.2 4.4.3 4.4.3.1 4.4.4 4.4.5 4.5 4.6

Introduction 73 The organism C. elegans 73 Introduction to the Channels 75 The Chemosensory Organs 75 The Amphid Organ 75 Phasmid Organ 78 Inner Labial 79 The Sensory Signaling Circuit 79 Behavioral Assays 79 Chemotaxis 80 Repulsion 81 Thermotaxis 82 Social Feeding or Bordering 82 How Is The Response to Each Stimulus Generated? 83 The Chemotaxis Olfactory Response 83 Chemotaxis to Water-soluble Compounds 85 Repellents 85 The ASH Polymodal Sensory Neuron 85 Thermotaxis 86 Feeding Behavior 87 Structure of the TAX, Cyclic Nucleotide-gated Channels of the Worm Channel Regulation 93 References 94

58

88

VII

VIII

Table of Contents

5

5.1 5.1.1 5.1.2 5.1.3 5.1.4 5.2 5.2.1 5.3 5.4 5.5 5.5.1 5.5.2 5.6 5.7 5.8 5.9 5.10 5.11 5.12 5.13 5.14 5.15

Vertebrate Olfactory Signal Transduction and the Interplay of Excitatory Anionic and Cationic Currents 99 Johannes Reisert and Jonathan Bradley

Abstract 99 Introduction 99 Tissue 99 Olfactory Receptor Neurons 100 Sustentacular Cells 101 Basal Cells 102 Recording Odor-induced Electrical Activity 102 The Electroolfactogram 102 Odorant Responses of Single Isolated Olfactory Receptor Neurons 103 Components of the Transduction Pathway 106 Cloning of G Proteins Expressed in the OE 108 Gaolf 108 Adenylyl Cyclase 108 Odorant Receptors 109 Cyclic Nucleotide-gated Channel in OE 110 Cloning of a CNG Channel Expressed in the OE 114 Negative Feedback by Ca2+ on the CNG Channel 115 The Olfactory Ca2+-activated Cl– Channel 118 Activation of the Cl– Conductance 119 Single Channel Properties and Channel Densities 121 Regulation of Cl– Channel Activity 122 Amplification of the CNG Current and Generation of the Cl– Current 123 Open Questions 126 References 127

6

Transduction Channels in the Vomeronasal Organ Emily R. Liman and Frank Zufall

6.1 6.2 6.3

Introduction 135 Anatomy of the Vomeronasal System 136 Sensory Responses Involve Generation of Action Potentials and Ca2+ Entry 137 Two Families of G-protein-coupled Receptors Mediate VNO Transduction 139 Signaling Downstream of G Proteins May Involve a PLC 140 Second Messengers for VNO Transduction: Functional Studies 140 Identification of the TRPC2 Ion Channel as a Candidate Transduction Channel for VNO Sensory Signaling 141 TRPC2 Is Essential for Pheromone Transduction 144 Mechanism of TRPC2 Activation 144 TRPC2 Knockout Mice: Behavioral Defects 146 Loss of VNO Signaling Components in Human Evolution 147 Summary: Is TRPC2 the VNO Transduction Channel? 149 Acknowledgements 150

6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12

135

Table of Contents

7

Transduction Mechanisms in Taste Cells Kathryn Medler and Sue C. Kinnamon

7.1 7.2 7.2.1 7.2.1.1 7.2.1.2 7.2.2 7.2.2.1 7.2.2.2 7.2.2.3 7.3 7.3.1 7.3.2 7.3.3 7.3.4 7.3.5 7.3.5.1 7.3.5.2 7.4

Introduction 153 Ionic Stimuli 155 Salt 155 Epithelial Sodium Channel 155 Amiloride-insensitive Pathway 158 Sour 159 Proton-permeable Channels 161 Proton-gated Channels 161 Proton-blocked Channels 162 Complex Stimuli 163 GPCR Signaling in Taste Cells 163 Store-operated Channels and TRPM5 Cyclic Nucleotide-regulated Channels Ligand-gated Channels 171 Miscellaneous Channels 173 Fat-modulated Channels 173 Water-activated channels 173 Conclusions 174

8

Invertebrate Phototransduction: Multimolecular Signaling Complexes and the Role of TRP and TRPL Channels 179 Armin Huber

8.1 8.2 8.3 8.4 8.5 8.5.1 8.5.2 8.5.3 8.5.4 8.6 8.7

153

164 169

Abstract 179 Introduction 180 Structure of the Drosophila Compound Eye and Its Visual Pigments 181 The Drosophila Phototransduction Cascade Is a Prototypical G-proteincoupled Signaling Pathway 184 Essential Components of the Transduction Pathway Are Organized into a Multimolecular Signaling Complex 186 TRP and TRPL, the Transduction Channels of Drosophila Photoreceptors 189 Identification and Characterization of TRP and TRPL 189 Possible Gating Mechanism 191 Transduction Channels in the Visual Systems of Other Invertebrates 193 Drosophila TRP Is the Founding Member of the TRP Family of Ion Channels 194 Light-dependent Relocation of TRPL Alters the Properties of the Photoreceptive Membrane 196 Concluding Remarks and Outlook 198 Acknowledgments 199

IX

X

Table of Contents

9

The Transduction Channels of Rod and Cone Photoreceptors U.B. Kaupp and D. Tra¨nkner

9.1 9.2 9.2.1 9.2.2 9.2.3 9.3 9.3.1 9.3.2 9.3.3 9.3.4 9.4 9.5 9.6 9.6.1 9.6.2 9.7 9.7.1 9.8 9.8.1 9.9 9.9.1 9.9.2

Introduction 207 Brief Overview 207 Ligand Sensitivity 207 Ion selectivity 208 Modulation 208 Function of CNG Channels in Phototransduction and Adaptation 209 Rod and Cone Photoreceptors 209 CNG Channels in Pinealocyte Photoreceptors 212 CNG Channels in Parietal Eye Photoreceptors 213 CNG Channels in Hyperpolarizing Photoreceptors of Invertebrates 214 Structure of Subunits 215 Transmembrane Topology and Subunit Stoichiometry 215 Interaction of CNG Channels With Other Proteins 219 The Glutamic Acid-rich Part (GARP) of B1 219 Interaction with the Na+/Ca2+-K+ Exchanger 220 Modulation 221 Modulation by Ca2+ 221 Phosphorylation 223 Retinal 223 Visual Dysfunction Caused by Mutant CNG Channel Genes 224 Mutations Associated with Retinitis Pigmentosa 225 Mutations Associated with Achromatopsia or Cone Dystrophy 227 Appendix 229

10

Ion Channels and Thermotransduction Michael J. Caterina

10.1 10.2

Introduction 235 Physiological Studies Provide Evidence for the Existence of Thermally Gated Ion Channels 236 Molecular Characterization of a Heat-gated Ion Channel, TRPV1 239 TRPV2 Is an Ion Channel Activated by Extremely Hot Temperatures 241 TRPV3 and TRPV4 Are Warmth-activated Channels 242 TRPM8 and ANKTM1 Are Activated by Cool and Cold Temperatures, Respectively 242 Non-TRP Channels Implicated in Mammalian Temperature Sensation 243 Temperature-sensing Proteins in Non-mammalian Species 244 Mechanisms of Temperature Transduction 245 Conclusions 246

10.3 10.4 10.5 10.6 10.7 10.8 10.9 10.10

207

235

Table of Contents

11

Pain Transduction: Gating and Modulation of Ion Channels 251 Peter A. McNaughton

11.1 11.2 11.2.1 11.2.2 11.2.3 11.2.4 11.2.5 11.2.6 11.2.7 11.3 11.3.1 11.3.1.1 11.3.1.2 11.3.2 11.3.3 11.3.4 11.3.4.1 11.3.4.2 11.3.4.3 11.3.4.4 11.3.4.5 11.3.5 11.3.6 11.3.7 11.4

Introduction 251 Ion Channels Gated by Noxious Stimuli 253 Ion Channels Gated by Noxious Heat 253 Ion Channels Gated by Noxious Cold 254 Ion Channels Gated by Acid 254 ATP-gated Ion Channels 255 Ion Channels Gated by Mechanical Stimuli 256 Initiation of Action Potentials by Noxious Stimuli 256 Summary Diagram of a Nociceptive Terminal 257 Sensitization by Inflammatory Mediators 257 Short-term Sensitization: Mediators and Pathways 258 Bradykinin and the PKC Pathway 258 Prostaglandins and the PKA Pathway 260 Nerve Growth Factor 262 Direct Modulation of TRPV1 by Protons 263 Other Modulators of Nociceptor Sensitivity 264 ATP 264 Proteases 264 Bv8/Prokineticin 264 Glutamate 265 Norepinephrine 265 Long-term Sensitization 265 Gene Expression Regulated by NGF 266 Gene Expression Regulated by GDNF 266 Conclusions 267

12

Transduction and Transmission in Electroreceptor Organs Robert C. Peters and Jean-Pierre Denizot

12.1 12.1.1 12.1.2 12.2 12.2.1 12.2.2 12.2.3 12.2.4 12.2.5 12.3 12.4 12.4.1

271

Abstract 271 Introduction 272 Transduction at Electroreceptor Cells 272 Favorite Species 273 Types of Electroreceptor Organs 274 General 274 The Sensory Mucous Glands in Monotremes 274 The Microampullary Organs 275 The Tuberous Organs 275 The Ampullae of Lorenzini 275 How is Transduction at Electroreceptor Cells Studied? 276 Current Views on Transduction and Transmission in Electroreceptor Organs 276 Spontaneous Activity and Modulation of Afferent Activity 276

XI

XII

Table of Contents

12.4.2 12.4.3 12.4.3.1 12.4.3.2 12.4.3.3 12.4.3.4 12.4.3.5 12.4.4 12.4.5 12.4.5.1 12.4.5.2 12.4.5.3 12.4.5.4 12.5 12.6

Monotreme Mucous Gland Electroreceptor Organs 277 Microampullary Organs in Freshwater Organisms 278 General 278 Patch-clamp Experiments 280 Indirect Pharmacological Evidence 281 The Synaptic Paradox 282 The Transduction Model Revisited 283 Tuberous Organs in Freshwater Fishes 285 Ampullae of Lorenzini in Marine Fish 286 General 286 Ampullae in Plotosus 287 Ampullae in Elasmobranchs 287 The Synaptic Paradox 287 Mucus and Transduction 290 Conclusions and Open Ends 291 Acknowledgments 293

Preface

She loves him, observes the tourist upon beholding the image of the Pharaoh and his wife in the Egyptian Museum in Cairo. And indeed, the intimate scene depicting Tutankhamun and his queen Ankhesenamun (shown on the cover of our book) confers that impression even more than 3000 years after its creation. For the sensory physiologist who recognizes Ankhesenamun’s gesture as a mechanosensory gentle touch, the sensation of a hand touching a shoulder is in molecular terms no simple process. In fact, more than 20 years of hard experimental work was necessary to shed some light on the molecular steps that convert, or transduce, physical contact into an electrical signal interpretable by the nervous system. As Laura Bianchi and Monica Driscoll outline in the first chapter of this book, the path toward understanding touch was paved by a creature much less noble than Tutankhamun: the soil nematode Caenorhabditis elegans. Painstaking genetic analysis of the worm’s response to being experimentally touched and probed with an eyelash led to identification of the transduction channels in mechanosensory neurons, known collectively now as the degenerin family of ion channels. It is these proteins that translate mechanical stimuli into electrical signals that can be processed by the sensory neurons and eventually are interpreted by the organism as a sensory experience. In all sensory cells transduction channels show fascinating adaptations to their task of reporting sensory stimuli. Imagine this: if Ankhesenamun speaks to her husband, or when the Pharaoh listens to his musicians playing cymbals and harp, tiny protein filaments tug at the transduction channels in his inner ear to excite mechanosensory hair cells and to produce a neuronal auditory signal. Robert Fettiplace describes in his chapter the biophysical examination of these exquisitely sensitive transduction channels.

XIV

Preface

In the world of chemoreception, transduction channels appear to be as numerous as the qualities of chemical stimuli. Acid-sensing ion channels respond to the simplest of all chemicals. They are opened by protons and probably serve multiple functions in the body, including the generation of heartache when ischemia turns things sour within the myocardium. Other chemoreception modalities are more conducive to Pharaoh’s bliss. In particular, the metabotropic transduction cascades of taste and smell form the molecular basis of sensory pleasures, which were so highly cherished by the ancient Egyptians that the hieroglyphic determinative for happiness was a nose. The chemoreception chapters in our book describe the state of knowledge about transduction channels in chemosensory cells. Here we meet an entire zoo of different transduction channels, including cyclic nucleotide-gated cation channels, calcium-activated chloride channels, and a channel family that plays an increasingly prominent role in sensory physiology: the transient receptor potential channels. Transient receptor potential channels mediate sensory transduction in systems as diverse as mouse pheromone receptors, insect ommatidia, and human thermoreceptors, apparently acting as one of nature’s multiple-purpose transduction components. Looking at his wife is probably what makes the Pharaoh really happy. And, indeed, the beautiful daughter of Nefertiti must have been an exceptional visual experience. Just look at how the rays of the sun seem to caress her and her husband with tiny hands of light! The old Egyptians surely had a way of representing sensory perception in art. In modern days, we have learned to understand how photoelectrical transduction works in the light-sensitive cells of the retina. Dimitri Tra¨nkner and Benjamin Kaupp describe the pivotal role that transduction channels play in such different visual tasks as looking at stars at night or beholding bright and colorful images such as the one on the book cover. And Armin Huber explains the ingenious method that flies use to achieve high temporal resolution in vision: the formation of multimolecular signaling complexes to rapidly drive transduction channels. If you have ever wanted to know why you can rarely catch the fly that annoys you, read this chapter. It won’t help you in catching the insect, but you will understand why the bug is so fast. In the concluding chapter, Robert C. Peters and Jean-Pierre Denizot discuss a sensory modality that Tutankhamun and Ankhesenamun did not use to perceive the world: electroreception. If not the Pharaoh, another denizen of Egypt is a master of electroreception. A small fish with a long nose, the elephant nose (Gnathonemus petersii), finds his way through murky waters by means of emitting and perceiving electrical signals. The elephant nose is one of the best-studied weakly electric fish, and it has slowly revealed how it does it. It is fascinating to read about the sensory equipment that electric fish employ to feel their way in the dark! Thus, the authors of this book cover many sensory modalities and explain the generation of receptor currents in a wide range of sensory cells. They address their chapters to students of biology, physiology, and medicine, as well as to scientists interested in signal transduction, sensory physiology, and perception. And who knows – even some aficionados of Egyptian archaeology may wish to know more about the Pharaoh’s senses. Baltimore and Heidelberg March 2004

Stephan Frings Jonathan Bradley

XV

List of Contributors

Laura Bianchi Department of Molecular Biology and Biochemistry Rutgers University New Brunswick, NJ 08855 USA Jonathan Bradley Department of Neuroscience Johns Hopkins University School of Medicine 725 N. Wolfe St. Baltimore, MD 21205 USA [email protected] Michael J. Caterina Department of Biological Chemistry and Neuroscience Johns Hopkins University School of Medicine 725 N Wolfe St. Baltimore, MD 21205 USA [email protected] Kenneth A. Cushman Vollum Institute Oregon Health & Science University Portland, OR 97239 USA

Jean-Pierre Denizot UNIC – Unit de Neurosciences Intgratives et Computationnelles CNRS UPR 2191 1 Avenue de la Terrasse 91198 Gif-sur-Yvette France Monica Driscoll Department of Molecular Biology and Biochemistry Rutgers University Nelson Biological Laboratories 604 Allison Road Piscataway, NJ 08855 USA [email protected] Robert Fettiplace Department of Physiology University of Wisconsin Medical School 1300 University Avenue Madison, WI 53706 USA [email protected] Stephan Frings Department of Molecular Physiology University of Heidelberg Im Neuenheimer Feld 230 69120 Heidelberg Germany [email protected]

XVI


Armin Huber Department of Cell and Neurobiology Institute of Zoology University of Karlsruhe Haid-und-Neu-Str. 9 76131 Karlsruhe Germany [email protected] U. Benjamin Kaupp Institut fu¨r Biologische Informationsverarbeitung Forschungszentrum Ju¨lich Leo-Brand-Straße 52425 Ju¨lich Germany [email protected] Sue C. Kinnamon Department of Biomedical Sciences Colorado State University Ft. Collins, CO 80523 USA [email protected] Noelle D. L’Etoile Department of Psychiatry University of California Center for Neuroscience 1544 Newton Ct. Davis, CA 95616 USA [email protected] Emily R. Liman Department of Biological Science University of Southern California 3614 Watt Way Los Angeles, CA 90089 USA [email protected]

Edwin W. McCleskey Vollum Institute Oregon Health & Science University Portland, OR 97239 USA [email protected] Peter McNaughton Department of Pharmacology University of Cambridge Tennis Court Road Cambridge CB2 1PD UK [email protected] Kathryn Medler Department of Biomedical Sciences Colorado State University Ft. Collins, CO 80523 USA Robert C. Peters Functional Neurobiology Utrecht University Padualaan 8 3584 CH Utrecht The Netherlands [email protected] Johannes Reisert Department of Neuroscience Johns Hopkins University School of Medicine 725 N. Wolfe St. Baltimore, MD 21205 USA [email protected]


Dimitri Tra¨nckner Institut fu¨r Biologische Informationsverarbeitung Forschungszentrum Ju¨lich Leo-Brand-Straße 52425 Ju¨lich Germany present address: University of California, HHMI 9500 Gilman Drive La Jolla, CA 92093 USA

Frank Zufall Department of Anatomy and Neurobiology University of Maryland School of Medicine Baltimore, MD 21201 USA

XVII

1

1

The Molecular Basis of Touch Sensation as Modeled in Caenorhabditis elegans Laura Bianchi and Monica Driscoll

Abstract

One of the looming mysteries in signal transduction today is the question of how mechanical signals, such as pressure or stretch, are sensed. Elegant electrophysiological studies in organisms ranging from bacteria to mammals support that mechanotransduction can be mediated by ion channels that gate in response to mechanical stimuli. Despite the importance of the molecular identification of these ion channels for elaborating mechanisms of mechanotransduction, genes encoding mechanosensitive ion channels eluded cloning efforts for a long time. Breakthroughs in the understanding of mechanosensitive channels have come from genetic analyses of touch sensation in Caenorhabditis elegans and Drosophila. In C. elegans, screens for touch-insensitive mutants identified two genes, mec-4 and mec-10, that encode channel subunits implicated in touch sensation and are postulated to be the core of a mechanotransducing ion channel complex. mec-4 and mec-10 encode proteins with similarity to subunits of the mammalian amiloride-sensitive epithelial Na+ channel (ENaC) that mediates sodium reabsorption in the kidney and lung. mec-4 is expressed exclusively in six neurons that laser ablation studies have identified as gentle-touch receptors, and mec-10 is expressed in these six neurons plus two pairs of touch receptors that are thought to sense harsher touch. The same genetic screens that identified mec-4 and mec-10 identified other genes required for normal touch sensation in the nematode. MEC-5, a novel collagen, and MEC-9, a protein that includes multiple Kunitz-type protease inhibitor repeats and EGF repeats, are extracellular matrix proteins that may interact with MEC-4/MEC-10 channel subunits on the extracellular side of the neuron to help exert gating tension on the channel. Inside the touch receptor, a specialized cytoskeleton is assembled that features 15-protofilament microtubules composed of MEC-12 a-tubulin and MEC-7 b-tubulin subunits. This cytoskeleton may be linked to tether MEC-4/MEC-10 on the intracellular side. When a mutant hyperactivated MEC-4(d) subunit is heterologously expressed in Xenopus oocytes, voltage-independent Na+ currents are produced that can be modulated in both amplitude and properties by two other proteins also identified by genetic screens as required for touch transduction: MEC-2, a stomatin-like protein, and Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

2

1 The Molecular Basis of Touch Sensation as Modeled in Caenorhabditis elegans

MEC-6, a protein that shares similarity with mammalian paraoxonases. The C. elegans genome encodes 28 members of the MEC-4 and MEC-10 channel family, called the degenerin family. We discuss here the global role of degenerins in mechanosensation, reporting findings on the function of three other degenerins (UNC-8, DEL-1, and UNC-105) in mechanosensitive and stretch-sensitive behaviors in the nematode, and we review studies addressing the role of mammalian homologues in touch sensation.

1.1

Introduction

The sense of touch is so profoundly important to our daily life that – when you actually think of it – the degree to which we take this sense for granted is unthinkable. We fully depend on our sense of touch to make and drink our morning coffee, to flip through the newspaper, to dress, and to move to the places where we type, phone, compute, pass paper, fold, sell, and manufacture things. Virtually no activities required for daily life (feeding, drinking, moving, protecting, communicating) can transpire without touch or mechanical sensation. Moreover, without touch sensation we would be unable to ensure the viability of our young. In addition to the obvious reasons for this, it is becoming increasingly clear that touch plays a critical role in both physical and emotional development. For example, hospitalized preterm infants show accelerated weight gain, enhanced activity, and faster development if they are gently stroked daily for 15 minutes – resulting in faster hospital discharge [5]. Despite widespread and fundamental importance, touch is the least understood of the senses, at both the cellular and molecular levels. The sense of touch is initiated by the perception of a mechanical stimulus such as pressure and the conversion of this signal into electrical signaling. Groundbreaking electrophysiological studies characterized ion channels that could be gated in response to pressure or stretch rather than voltage changes or ligand binding [39, 41, 58]. Such channels could be identified in specialized mechanoreceptors [24, 48], yet the genes encoding mechanically gated ion channels that mediate the senses of touch and hearing eluded cloning efforts for years (some genes, such as those encoding the hearing channel, remain unidentified even to this day; see Chapter 2). Technically, this might have been predicted, as there are no known reagents that specifically associate with mechanosensitive channel subunits at high affinity that could facilitate protein isolation and there is a remarkable paucity of mechanically gated channels even in specialized mechanotransducing structures such as the vertebrate cochlea. Moreover, given that these channels are likely to be tethered to accessory proteins that exert gating tension, reconstitution in heterologous systems is extremely difficult. Although the cloning of mechanically gated MscL and MscS channels from bacteria constituted major breakthroughs in the field of mechanical signaling [39], the MscL and MscS channel classes have no clear eukaryotic homologues, and thus their identification did not facilitate an immediate revolution in our understanding of mammalian mechanotransduction.

1.2 Features of the C. elegans Model System

Exciting advances in our understanding of the sense of touch have instead emerged from invertebrate genetics. Both nematode and fly mutants defective in touch sensation have facilitated the cloning of ion channels thought to act directly as mechanotransducing channels. More specifically, the DEG/ENaC Na+ channel subunits (named for the C. elegans degenerins and the related mammalian epithelial amiloride-sensitive Na+ channel) have been directly implicated in touch sensation in both invertebrates and vertebrates. Likewise, members of the transient receptor potential (TRP) channel family are mechanotransducing channels implicated in touch [84] and possibly hearing [50, 69] (see Chapter 2). Here we focus on reviewing the genetic, molecular, electrophysiological, and calcium-imaging studies conducted using the simple nematode C. elegans that have greatly advanced our understanding of touch sensation through the identification and the characterization of mechanically gated DEG/ENaC ion channels and accessory proteins. We discuss how these and TRP channels may work together to contribute to touch sensation and note how data from invertebrates has stimulated a successful search for analogous processes in higher organisms.

1.2

Features of the C. elegans Model System

The 1-mm long simple soil worm Caenorhabditis elegans is a facile system for experimental manipulation that features many developmental and behavioral pathways strikingly conserved between nematodes and mammals. C. elegans can be easily reared on an E. coli diet in the laboratory. This animal completes a reproductive life cycle in just 2.5 days at 25 8C, during which it progresses through embryonic development and four larval stages (L1-L4) before reaching sexual maturity. The most common sexual form is the hermaphrodite (XX), although males (X0), which can arise spontaneously by nondisjunction, can be easily propagated in the laboratory for use in genetic studies. The C. elegans body and eggshell are transparent so that each cell can be visualized by Normarski microscopy. In fact, the entire map of all cell divisions during the development of the animal has been constructed [44, 75]. The nervous system includes only 302 neurons, for which the pattern of synaptic connections, including circuits for specific mechanosensory behaviors, has been deduced using serial section electron microscopy [85]. Laser ablation experiments have helped define the importance of specific identified neurons in mechanosensory behaviors [14, 47]. The major advantage of using C. elegans as a model system for studying biological processes is that it is a powerful genetic system [8]. Mutations that affect development and behavior, including those affecting touch sensation, have been generated and mapped to specific genes. Sequence analysis of the C. elegans genome is complete [20], and powerful methods for generation of transgenic animals [26] and dsRNAimediated transcript disruption (RNAi) [27, 77] are routine. Despite the considerable advantages that C. elegans offers for studying gene function in vivo, this model has had certain limitations for electrophysiological analysis of chan-

3

4


nel function, especially for channels expressed in neurons. The tiny neurons (1–2 lm diameter) are embedded in poorly accessible tissues confined in a pressurized cuticle. However, recent technical improvements established in the field have led to the development of electrophysiological methods for characterizing channel function in C. elegans [35]. In addition, a recently developed method for culturing C. elegans cells now allows routine electrophysiological recordings from neurons, muscles, and other cell types [19, 76]. Finally, sophisticated methods for monitoring intracellular calcium concentration changes during channel activity in living and behaving nematodes have been developed and have led to important findings [49, 76]. Taken together, the identification of specialized mechanosensory neurons, the cloning of genes required for mechanosensitive responses, and the study of their function both in vivo and in vitro have led to significant insight into the molecular mechanisms of mechanotransduction in C. elegans.

1.3

Mechanosensation Is a Major Mechanism by Which C. elegans Senses Its Environment

C. elegans does not have a sense of sight and must evaluate its environment primarily by chemosensation and mechanosensation (see Chapter 4). C. elegans can respond to a range of mechanical stimuli encountered virtually anywhere on its body. The bestcharacterized mechanosensitive behavior is the movement away from a gentle brush of an eyelash hair delivered to the body, generally referred to as gentle-touch sensation [13]. Other mechanosensitive behaviors include response to head-on collision with an object (the nose touch response), response to light touch to the side of the nose (head withdrawal response), response to harsh touch delivered by a metal wire, and response to tapping on the plate on which the worms are reared. The process by which males mate most likely involves touch-mediated recognition of the hermaphrodite vulva. Mechanical stimuli also impact on locomotory behaviors, foraging, feeding, egg laying, and defecation circuits. Because the avoidance of gentle touch is the behavior most intensively investigated, here we will first focus on summarizing how the study of gentle touch has produced a detailed molecular model for a mechanically gated ion channel. Later, we will review what is known about the identities of genes that influence other mechanosensory behaviors, and we will consider emerging molecular themes in touch sensation.

1.4 Gentle Body Touch

1.4

Gentle Body Touch 1.4.1

The Touch Receptor Neurons

In the laboratory, C. elegans moves across an agar plate on its side with a readily observed sinusoidal motion. When stroked with an eyelash hair on the anterior body, the animal will reverse its direction and move backwards; if touched on the posterior body, it will move forward [13]. The neurons required for the sensation of the gentle-touch stimuli have been identified by laser ablation studies and genetic disruption. These six touch receptor neurons were initially called microtubule cells because their processes are filled with distinctive 15-protofilament microtubules. Their processes are embedded in the hypodermis adjacent to the cuticle (the worm “skin”) and run longitudinally along the body wall, a distribution that enables them to more or less “cover” the touch sensory field of most of the body. Two embryonically generated PLM neurons (posterior lateral microtubule cells) are situated in the posterior body, on the right and left sides; two embryonically generated ALM neurons (anterior lateral microtubule cell) are situated in the anterior, on the right and left sides. In the first larval stage, AVM (anterior ventral microtubule cell) and PVM (posterior ventral microtubule cell) are added to the body plan. Laser ablation of individual ALMs, PLMs, and AVM established roles for these neurons in gentle touch [14]. Although PVM looks identical to the other touch neurons, it does not initiate a behavioral response to gentle touch on its own, and thus it has been postulated to modulate other behavioral circuits that can be influenced by touch [14] (Fig. 1.1A).

1.4.2

Ultrastructural Features of the Touch Receptor Neurons Touch Cell-specific Microtubules Touch receptor processes are filled with bundles of wide-diameter (15-protofilament, pf) microtubules that are uniquely assembled in this group of six neurons [15, 16]. Most other nematode cells include 11-pf microtubules (Fig. 1.1C). The 15-pf microtubules are required for touch receptor function: if microtubules are disrupted by the microtubule assembly inhibitor colchicine or by genetic mutations, touch sensitivity is completely lost [13, 15]. Individual microtubules are not long enough to extend from end to end of the touch neuron. Rather, single microtubules (10-20 lm long) overlap with each other to span the full length of the touch cell processes (about 400-500 lm). Interestingly, the distal microtubule end is diffusely stained and is always situated outside of the microtubule bundle, often positioned adjacent to the plasma membrane. This ultrastructural feature suggests that the oriented microtubule network might associate with plasma membrane proteins, such as the mechanosensitive ion channels that sense touch, a hypothesis that remains to be tested [16] (see discussion of mechanotransduction model below). 1.4.2.1

5

6


A

C cuticle touch receptor process mantle

hypodermis 15-pf microtubules

B

D cuticle

mantle

touch receptor process

15-pf microtu

hypodermis

Fig. 1.1 C. elegans neurons that sense gentle body touch. (A) Diagram showing the position of the six neurons that in C. elegans sense the gentle stroke of an eyelash hair on the body; anterior body is to the left. There are two fields of touch sensitivity defined by the position of the touch neurons processes along the body axis. The ALMs and AVM sense touch to the anterior field, whereas PLMs sense touch to the posterior field. (B) Touch neurons are here visualized in a living nematode, by expression of the Green Fluorescent Protein under the control of the mec-4 promoter, which is active exclusively in these neurons. Arrows point to touch receptor cell bodies. (C) Electron micrograph of a cross-section of a touch receptor neuron process. The touch cell process, which is surrounded by the mantle and

embedded in the hypodermis, is filled with 15-pf microtubules and is in very close proximity to the cuticle. This anatomical arrangement is thought to ensure the transmission of the mechanical forces applied on the cuticle down to the touch neuron process. (D) Schematic representation of a touch receptor neuron EM cross-section, depicting its most important components. The darkly stained region, depicted here as a bar-shaded rectangle connecting the mantle and the cuticle, is the fibrous organelle (not visible in the electron micrograph). Such specializations occur periodically along the length of the touch receptor process and may serve to attach the process to the cuticle. Adapted from [78]

The Extracellular Mantle Touch receptor processes are surrounded by a specialized extracellular matrix, called the mantle, which appears to help maintain the touch receptor process in close association with the cuticle [13]. Cuticular structures resembling muscle attachment sites are positioned periodically along the length of the touch receptor process in close contact with the mantle and may be sites at which the touch receptor process is fixed to the cuticle (Fig. 1.1D). Although genetic mutations support that the integrity of the mantle is critical for touch receptor function, mutations in him-4 cause touch neurons to stray away from the cuticle, yet the mutants still sense touch [82]. Since detachment is variable in the him-4 background, it is possible that adequate contact is maintained 1.4.2.2


for some touch sensation; alternatively, any deflection of even a “loose” mechanoreceptor neuron might be sufficient to activate the behavioral avoidance response.

1.4.3

Genetic and Molecular Analysis of Body Touch

In pioneering studies on the genetics of touch sensation, Martin Chalfie and colleagues mutagenized animals and screened their progeny for the failure to respond to the gentle brush of an eyelash hair [11, 13]. The mutants selected exhibited grossly normal locomotion and were still able to respond to the prod of a metal wire, so that defects appeared to specifically alter gentle-touch sensitivity. Hundreds of touch-insensitive mutants, many of them designed as mec (mechanosensory abnormal), defined several genes that contribute specifically to touch cell development and function. It should be emphasized that since the criteria for mutant isolation demanded that other aspects of nematode locomotion and harsh-touch sensation be unaffected by the mutations, genes that encode proteins used for gentle-touch sensation but also used in other locomotory activities would not have been identified in this screen. Likewise, genes that encode functionally redundant proteins would be missed. Nonetheless, the genes identified in this screen provided a major breakthrough in our understanding of the molecules needed for touch sensation. mec-4 and mec-10 Ion Channel Subunits Form Na+ Channels mec-4 and mec-10 loss-of-function mutants are touch-insensitive, yet their touch receptor neurons appear to develop normally and share all apparent ultrastructural features of wild-type (WT) touch receptor neurons [13]. Cloning revealed that mec-4 and mec-10 encode homologous proteins related to subunits of the amiloride-sensitive, voltageindependent Na+ channel, which mediates Na+ reabsorption in vertebrate kidney, intestine, and lung epithelia (the ENaC channel [9, 10, 12, 22, 45, 52]). The mec-4 channel subunit is expressed only in the six touch receptor neurons (Fig. 1.1B [55]), and the mec-10 channel subunit is expressed in the six touch receptor neurons as well as in two other neuron pairs that may mediate stretch-sensitive or harsh-touch responses (FLPL/R and PVDL/R [45]). Because the MEC-4 and MEC-10 subunits are expressed exclusively in touch neurons and are clearly needed for the function of these neurons, and because no other channel genes were identified among touch-insensitive mutants, it was proposed that the MEC-4 and MEC-10 subunits assemble in vivo to create a mechanically gated channel that responds directly to touch. Progress toward addressing this hypothesis is outlined in more detail below. 1.4.3.1

MEC-4 at the Molecular Level There are many more mec-4 mutations than there are mec-10 mutations (perhaps suggesting that mec-4 plays a more central role in gentle touch), and thus MEC-4 structure/function is better understood. MEC-4 is a 768-amino-acid membrane protein that includes two membrane-spanning domains (MSDI, MSDII; see Fig. 1.2B). The chan1.4.3.2

7

8


Fig. 1.2 Degenerin MEC-4 structure/function. (A) Dendrogram of the 28 degenerins encoded by the C. elegans genome. The 28 genes encoding postulated degenerin subunits were identified by searching the C. elegans database, compiled by the C. elegans Genome Sequencing Consortium, for predicted proteins sharing homology with known degenerins. Black background indicates the most characterized degenerins, including MEC-4. (B) Structural features of a single MEC-4 subunit (likely four subunits form a channel). The MEC-4 polypeptide spans the membrane twice, leaving the Nand C-termini in the cytosol. The second membrane-spanning domain, which is longer than required for a single transmembrane pass, may loop back in the membrane to participate in the formation of the pore. Ala713, which when replaced by a bulkier amino acid results in necrotic cell death, is

indicated by the skull and crossbones icon. MEC-4 protein also features three cysteine-rich domains (CRDI, II, III) that are thought to be involved in protein-protein interactions, perhaps anchoring MEC-4 to extracellular matrix proteins. Other important domains include the putative extracellular regulatory domain, the neurotoxin-related domain, and the intracellular regulatory domain. (C) Model for mec-4(d)-induced toxicity. WT MEC-4 channels are able to open and close, but MEC-4(d) channels, which encode substitutions for a conserved alanine adjacent to MSDII, are thought to be “locked” in an open conformation due to steric hindrance. This is thought to result in excessive Na+ influx that triggers necrotic-like cell death, which manifests itself in the early stages as cell swelling (lower right panel)


nel subunit is positioned in the membrane such that relatively short N- and C-terminal domains project into the cytosol and a single large central loop extends extracellularly [52] (this is typical of all DEG/ENaC family members). The MEC-4 extracellular domain also includes three cysteine-rich domains (CRDI, CRDII, and CRDIII) and one region similar to venom neurotoxins (NTD) [79]. Understanding of structure/function relations in MEC-4 is still at an early stage, but studies on this and other members of the DEG/ENaC superfamily have highlighted three conserved regions important for function: (1) MSDII contributes to the channel pore [42]; (2) a short but highly conserved intracellular stretch adjacent to MSDI influences ion permeation and selectivity [36, 43]; and (3) the Cys-rich extracellular loop domains are important for function in some way, possibly mediating protein-protein interactions that may help tether the MEC-4 channel to the specialized extracellular matrix of the touch neuron. An unusual type of mec-4 mutation acts dominantly to induce swelling and neurodegeneration of the touch neurons. Substitution of large side-chain amino acids for a highly conserved Ala residue situated adjacent to channel pore MSDII (AA713; see Fig. 1.2B [22, 52]) generates MEC-4 mutant subunits (named MEC-4(d)) that induce necrotic-like death of the touch receptor neurons (Fig. 1.2C [11, 13]). The channel pore must be intact for neurodegeneration to occur [42], suggesting that ion influx is critical in the toxicity mechanism. Since large side-chain amino acids at the conserved “d” position are toxic but small ones are not, it was originally proposed that these large substitutions favor the channel-open conformation and hyperactivate ion influx [22, 42]. Indeed, the A713V substitution markedly enhances whole-cell currents when MEC-4(d) channel activity is measured in the Xenopus oocyte expression system [34]. Since other C. elegans family members (e.g., deg-1 and mec-10) can be altered by analogous amino acid substitutions to induce neurodegeneration [17, 45], the C. elegans branch of the gene family has been named the “degenerin” family (Fig. 1.2A). A mutant variant of neuronally expressed mammalian DEG/ENaC member MDEG (ASIC2), engineered to encode a large side-chain amino acid at the corresponding position, induces swelling and death when introduced in to Xenopus oocytes and hamster embryonic kidney cells [83]. The small amino acid normally situated at the “DEG” site can be modified with chemical reagents only when the channel is activated, supporting that conformational changes associated with an open channel involve this residue [1].

1.4.4

The Candidate Mechanotransducing Channel is a Heteromultimeric Complex

The subunit compositions and stoichiometry of DEG/ENaC channels remain somewhat uncertain. Electrophysiological assays of the rat ENaC channel reconstituted in Xenopus oocytes determined that at least three homologous subunits (a-, b-, and crENaC) must be co-expressed to form a channel with pharmacological properties similar to the in vivo channel [10]. Stoichiometries of four to nine subunits per ENaC channel have been supported [3, 6, 7, 21, 28, 51, 70]. Genetic interactions suggest that

9

10


MEC-4 and MEC-10, which cannot functionally complement one another and thus appear to perform distinct functions in vivo, form a heteromeric channel in touch neurons: engineered mec-10(d) subunit (harboring the substitution analogous to channel-activating, death-inducing MEC-4 substitution A713V) requires functional mec-4 to be toxic [37, 45]. MEC-4 and MEC-10 Form a Functional Ion Channel MEC-4 and MEC-10 co-assemble in Xenopus oocytes to form a Na+-selective channel sensitive to the ENaC-blocking agent amiloride. This channel exhibits a high permeability to lithium, as do other members of the DEG/ENaC superfamily. Interestingly, while MEC-4 can form channels of low conductance on its own, MEC-10 is not functional when expressed alone. However, the co-introduction of the MEC-10 subunit to the MEC-4(d) channel in oocytes affects the Kd for amiloride, consistent with MEC-10 being included in the same channel as MEC-4 [34]. Still, the introduction of MEC-10 in the oocyte system does little to change most properties of the MEC-4 channel, and the “MEC-10(d)” mutant subunit cannot conduct current on its own. Thus, the MEC-4 subunit appears most critical for channel properties. Importantly, in Xenopus oocytes the MEC-4/MEC-10 channel has not been demonstrated to be gated by mechanical forces (membrane stretch induced by hypotonic solutions), probably due to the lack of intracellular and extracellular proteins, normally present in vivo, that are essential for channel gating (see below). The MEC-4(d) subunit conducts much more current than the MEC-4(+) subunit (at least 10 times larger currents), consistent with the idea that the “d” substitutions next to the channel pore hyperactivate the channel. Most electrophysiological studies have therefore concentrated on the activated MEC-4(d) channel, which conducts markedly more robust current than the MEC-4(+) subunit. 1.4.4.1

1.4.4.2 MEC-2 Is a Stomatin-like Protein That May Help Tether the MEC-4/MEC-10 Channel to the Membrane Bilayer and/or the Cytoskeleton

MEC-2 May Participate in Several Protein Interactions in the Touch Channel Complex Mechanosensitive ion channels are thought to be gated by forces exerted upon the channels via associated protein attachments. MEC-2 is a candidate protein that might help exert gating tension on the MEC-4/MEC-10 channel from the membrane and/or from the intracellular side. mec-2 encodes a 481-amino-acid protein expressed in the touch receptor neurons and in a few additional neurons in the head [46]. There are three candidate protein interaction domains in MEC-2: (1) a cytoplasmically situated N-terminal domain (positioned between aa 42 and aa 118) needed for the localization of MEC-2::lacZ enzymatic activity to the touch receptor process; (2) a central domain that exhibits 65 % identity to the human red blood cell protein stomatin, an integral membrane protein that associates with the RBC cytoskeleton and affects ion balance via an unknown mechanism [74] (note that the stomatin-related domain includes a hydrophobic stretch that is membrane-associated, but most of this domain is thought to project into the cytoplasm); and (3) a C-terminal intracellular proline-rich region that


is similar to SH3-binding domains that mediate protein interactions. That MEC-2 might serve as part of a link between the channel and the specialized microtubule cytoskeleton is suggested by the facts that specific MEC-12 mutant forms (although not the MEC-12 null mutant) prevent MEC-2::lacZ activity from being localized out to the touch neuron processes (the normal MEC-2::lacZ distribution for the wild-type background [46]) and specific combinations of mutant MEC-2 stomatins and hyperactivated channel subunit MEC-10(d) alter the severity of MEC-10(d)-induced neurodegeneration [37]. MEC-2/stomatin can physically associate with MEC-4 and MEC-10 in heterologous assay systems ([34] and our unpublished results), but associations with MEC-12 a-tubulin have not been documented. MEC-2 also appears to homo-oligomerize in vivo because specific combinations of two distinct MEC-2 mutant proteins in heteroallelic strains can influence touch responses in different ways [23]; other stomatin family members have been shown to oligomerize [71]. In sum, MEC-2 might set up a membrane microdomain raft environment critical to the function of the touch channel complex. As noted above, whether MEC-2 actually participates in a link to the cytoskeleton remains to be determined. MEC-2 Enhances MEC-4/MEC-10 Channel Activity Based on the sequence similarity MEC-2 shares with stomatin, (known to influence ion balance in RBCs [72, 73]), it was postulated that MEC-2 may affect the function of the MEC-4/MEC-10 channel. Assays of the MEC-4(d) channel expressed in Xenopus oocytes confirm this hypothesis [18, 34] (Fig. 1.3). Co-expression of MEC-2 with MEC4 and MEC-10 (both WT and (d) variants) increases current amplitude up to 40-fold over channels lacking MEC-2, yet the presence of MEC-2 does not change the total number of channels present at the oocyte surface. While the MEC-2 stomatin-like domain alone is sufficient to boost channel activity, the intact protein is required for the full effect. Interestingly, human stomatin is also able to modulate the MEC-4/MEC-10 channel complex, suggesting that stomatins regulate DEG/ENaC channels by a conserved mechanism. MEC-2 seems to participate in, or influence the formation of, the channel pore, since co-expression of MEC-2 alters lithium permeability of the channel complex (channels become less permeable to Li+ but remain permeable to Na+). The mechanism by which MEC-2 regulates MEC-4/MEC-10 channel activity remains unclear and may involve effects on single-channel conductance and/or open probability.

MEC-6 Is a Transmembrane Paraoxonase-like Protein That Controls MEC Channel Activity Recessive mec-6 alleles can confer insensitivity to gentle touch and also can completely block mec-4(d)-induced death, suggesting that mec-6 is needed in some way for in vivo MEC channel activity [13, 18, 40]. mec-6 encodes a 377-amino-acid transmembrane protein with a small intracellular N-terminus and a large extracellular C-terminus that has the potential for glycosylation. Over a stretch of 250 amino acids at its C-terminus, MEC-6 exhibits 25 % identity and 45 % similarity to vertebrate paraoxonase/ arylesterases. Vertebrate serum paraoxonase (PON1), and perhaps other mammalian 1.4.4.3

11

12


Fig. 1.3 Summary of currents generated by expression of combinations of MEC proteins in Xenopus oocytes. Current amplitude at –85 mV is shown here for the subunit compositions indicated by the grid. Complementary RNA corresponding to each subunit was synthesized in vitro and injected into Xenopus oocytes (10 ng per oocyte for mec-4(d), mec-10(d) and mec-2 and 1 ng per oocyte for and mec-6). Currents were recorded in voltage-clamp 4–10 days after injection and data are expressed as mean SE. Adapted from [18]

paraoxonases, are esterases and are thought to protect against cellular damage from toxic agents such as oxidized lipids in the plasma like low-density lipoproteins (LDL). As such, mammalian paraoxonases have been implicated in the prevention of atherosclerosis and coronary heart disease. Nematode MEC-6 may modify some component of the MEC channel to influence its activity, although no evidence for such a role has yet been identified. MEC-6 is expressed broadly in neurons, muscles, and the canal cell (the kidney of the worm) in an expression pattern that overlaps with multiple degenerin family members, and its activity appears to be needed for toxicity of mec-4(d) expressed in many different cells [40]. Thus, mec-6 is likely required for function of many degenerin ion channels encoded by the C. elegans genome. Whether mammalian paraoxonases are needed for DEG/ENaC channel activity has yet to be determined.


MEC-6 Potentiates Channel Activity Co-expression of MEC-6 and MEC-4(d) in Xenopus oocytes increases current amplitude up to 24-fold greater than when MEC-4(d) is expressed alone [18] (Fig. 1.3). By contrast, MEC-6 does nothing to alter the inability of MEC-10(d) to form functional channels on its own, suggesting that the MEC-4 subunit is always required for the formation of conducting channels. Co-immunoprecipitation of heterologously expressed subunits in CHO cells suggests that MEC-6 can associate with MEC-4, MEC-10, and MEC-2, as would be expected if all subunits act together in a single functional channel complex. MEC-4, MEC-10, and MEC-2 protein levels at the plasma membrane are not affected by the addition of MEC-6 to an oocyte expression system. Thus, MEC-6 is likely to increase current amplitude acting at the single channel level, perhaps increasing conductance and/or open probability, but this still remains to be established. MEC-6 can affect channel selectivity lowering lithium permeability, indicating that it may directly participate in the channel pore or that it may affect pore structure through interaction with other sites. Interestingly, effects of MEC-2 and MEC-6 on current amplitude are synergistic: when these subunits are co-expressed with MEC-4(d), they increase current amplitude up to 400-fold (Fig. 1.3). MEC-10(d) does not have a major impact on current magnitude or properties when MEC-2, MEC-4(d), and MEC-6 are co-expressed (except for the Kd of amiloride). The synergy afforded by MEC-2 and MEC-6 suggests that these two subunits might increase current amplitude through different mechanisms, and the transmembrane topologies suggest that MEC-6 has potential for more interaction with extracellular MEC-4/MEC-10 domains, whereas MEC-2 would have to interact with intracellular channel domains. Since they both include membrane-associated regions, both MEC-2 and MEC-6 could influence the transmembrane pore via the membrane as well. Taken together, the model of the touch channel that emerges is that core-conducting subunit MEC-4 associates with degenerin MEC-10, stomatin-related MEC-2, and paraoxonase-related MEC-6. These findings reveal mutual interactions between channel subunits and accessory proteins that form a heteromultimeric complex whose regulation and mechanism of function is just starting to be deciphered.

1.4.5

Intracellular Proteins Needed for Touch Transduction

The touch receptor processes are filled with bundled 15-protofilament microtubules. Mutations in two genes, mec-7 and mec-12, disrupt the production of these microtubules, often eliminating them entirely [11, 13]. Still, even in the absence of the 15-pf microtubules, the touch receptor processes grow out normally and become filled with 11-pf microtubules [15]. Such touch receptors do not function, suggesting that the extensively cross-linked 15-pf microtubules contribute a specific role in touch transduction. mec-7 encodes a b-tubulin [64] expressed at high levels in the touch receptor neurons [38, 65]. Like other tubulins, MEC-7 is highly conserved: it differs from other b-tubu-

13

14


lins by only seven amino acids in the variable C-terminal domain. mec-7 mutations isolated in the screen for touch-insensitive mutants range in severity from recessive to strongly dominant, and many of the amino acid changes that disrupt MEC-7 function have been defined [64, 65]. Mutations that disrupt touch sensitivity affect the domain for GTP binding and hydrolysis, sites for heterodimerization with a-tubulin, and sites for higher-order microtubule assembly. mec-12 encodes an a-tubulin expressed in the touch receptor neurons but also expressed in several other neurons that do not assemble 15-pf microtubules [38]. MEC12 is acetylated (Lys40) in the touch neurons but not in other neurons, suggesting that the modification may be important for 15-pf microtubule assembly [30]. Acetylation of MEC-12 also occurs in cultured touch neurons (Bianchi and Driscoll, unpublished observations), suggesting that correct processing of this protein and assembly of 15-pf microtubules do not require the presence of the extracellular mantle proteins normally produced by the hypodermis, such as collagen MEC-5 [23]. As is the case for mec-7, many mec-12 mutations are semi-dominant or dominant and are likely to disrupt subunit interactions or protofilament assembly. Work on mec-7 and mec-12 strongly supports that unique a and b tubulins assemble to form the 15-pf microtubules required for touch receptor function. Whether these specialized microtubules play a direct role in the function of the mechanotransducing complex, perhaps by participating in a direct linkage to the channel, remains to be determined.

1.4.6

Extracellular Proteins Needed for Touch Transduction MEC-1 mec-1 is needed for proper formation of the mantle, the touch neuron’s specialized extracellular matrix. In mec-1 mutants [13] the mantle is almost completely absent and touch neuron processes fail to attach to the cuticle. The failure to attach the touch receptor to the cuticle might be the reason touch cannot be transduced, although the observation that touch cell processes detach sporadically in a him-4 mutant background without affecting touch sensitivity [82] does somewhat challenge this working hypothesis. mec-1 encodes a 1999-amino-acid polypeptide with an N-terminal signal sequence followed by a Kunitz-type domain, two EGF domains, 14 additional Kunitz-type domains, and a C-terminus of 160 amino acids. The Kunitz and EGF domains are likely to be protein-protein interaction domains. Mutations in the N-terminal region through the sixth Kunitz domain affect wrapping of the touch neuron by the surrounding hypodermis and disrupt touch sensitivity, whereas mutations in the C-terminal region affect only touch transduction. This indicates that the C-terminus of MEC-1 is essential for mechanosensory function but not for attachment and engulfment of the touch cell processes. 1.4.6.1


MEC-5 mec-5 encodes a novel collagen that is secreted by hypodermal cells and is a component of the touch neuron mantle [23]. mec-5 mutations affect the mantle in a subtle manner. While normally the mantle can be stained with peanut lectin, in mec-5 mutants this staining fails ([13]; E. Hedgecock and M. Chalfie, unpublished). The central portion of the mec-5 protein is made up of Pro-rich Gly-X-Y repeats. mec-5 mutations (many of which are temperature-sensitive, ts) cluster toward the carboxy terminus of the protein and affect these repeats. No touch-insensitive mutations map to the amino- and carboxy-termini; therefore, the role of these unique sequences in MEC-5 function is not known. Genetic interactions indicate that mec-5 influences MEC-4/MEC-10 channel function (for example, mec-4 and mec-10 mutations can enhance the mec-5(ts) mutant phenotype [37]). Thus, MEC-5 could physically interact with the touch receptor channel, perhaps acting to provide gating tension via an extracellular link. The probable importance of collagen/degenerin interactions in channel function is underscored by studies of unc-105, another degenerin family member that is expressed in muscle [53]. Semi-dominant gain-of-function mutations in unc-105 cause severe muscle hypercontraction [57]. Specific mutations in let-2/sup-20, which encodes a type IV basement membrane collagen, suppress the unc-105(sd) phenotype [53, 57]. Taken together, these observations suggest that degenerin/collagen interactions, which still await experimental verification, may emerge as a common theme in the function of this channel class. 1.4.6.2

MEC-9 mec-9 encodes a protein that is secreted by the touch receptor neurons [23] and it is likely to be part of the mantle. Despite this, mec-9 mutations do not alter mantle ultrastructure in a detectable manner [13]. The mec-9 gene encodes two transcripts, the larger of which encodes an 834-amino-acid protein (MEC-9L) that is expressed exclusively by the touch neurons. The predicted MEC-9L protein appears specialized for protein interactions and contains several domains related to the Kunitz-type serine protease inhibitor domain, the Ca2+-binding EGF repeat, the non-Ca2+-binding EGF repeat, and a glutamic acid-rich domain. (Agrin, a protein that localizes acetylcholine receptors, has a similarly specialized domain structure, including multiple EGF and Kazal-type serine protease inhibitor repeats [63].) Mutations that disrupt MEC-9 function affect the two Ca2+-binding EGF repeats, the sixth EGF repeat, and the third Kunitz-type domain, thus highlighting these regions as important in touch transduction. mec-9 mutations are dominant enhancers of a mec-5(ts) allele, indicating that MEC-5 and MEC-9 proteins may interact in the mantle [23, 37]. 1.4.6.3

15

16


1.4.7

The MEC Channel Functions Specifically in Neuronal Responses to Gentle Touch Test of a Key Hypothesis Although genetic studies provided clear breakthroughs in the identification of molecules required for gentle body touch sensation, the technical challenges that precluded direct electrophysiological recording from tiny nematode touch receptor neurons stalled testing of the critical hypothesis that the MEC-4 channel functioned specifically in the process of touch transduction. This next step in analysis of touch mechanisms was critical because although the genetic data indicate that each MEC protein is needed for a behavioral response to gentle touch, they cannot distinguish whether a given protein might be a true component of the mechanotransducing complex rather than being needed generally for the integrity of neuronal function or for neuronal signaling downstream of the initial perception event. This concern was particularly strong for the channel subunits, which could function in setting the resting membrane potential or propagating electrical signals, and for MEC-2 stomatin, which might be associated with a general leakiness of membranes (analogous to what occurs in hereditary stomatocytosis in which RBCs lack stomatin [29]). Proteins required for general maintenance of touch neurons would play a secondary role rather than being true touch-transducing molecules. Elegant answers to the question of specificity of function have been recently obtained using a gene-based, fluorescent calcium-binding reporter named “cameleon” [76]. In cameleon, the fluorophores cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) are linked by calmodulin and the calmodulin-binding peptide M13 [25, 56]. In low calcium, the two fluorophores are positioned far from each other, resulting in low fluorescence resonance energy transfer (FRET). In high calcium, calcium-bound calmodulin associates with M13 and the two fluorophores are brought close together, enabling FRET. Ratiometric CFP/YFP signals over time thus give a report of in vivo calcium fluxes. The power of the cameleon reagent is that it is genetically encoded and thus signals from live transgenic animals expressing the cameleon in touch neurons can be imaged to ask whether touch neurons mount calcium transients in response to touch, providing for the first time a physiological readout of touch receptor function. Experiments based on cameleon-reported Ca2+ transients revealed three key findings: (1) C. elegans touch receptor neurons respond to gentle body touch with transient calcium influx. (Although the MEC-4 channel appears to be a Na+ channel [34], the Ltype voltage-gated Ca2+ channel EGL-19 and the regulatory subunit UNC-36 are needed for normal calcium transients. An EGL-19-containing channel is thus likely gated by membrane depolarization associated with Na+ channel activation.) (2) mec-4, mec-2, and mec-6 null mutants lack any cameleon-detected response to gentle-touch stimuli (Fig. 1.4A). Thus, the MEC-4/MEC-2/MEC-6 channel is needed for calcium transients provoked by gentle-touch stimuli. (3) The touch neurons mount a distinct cameleondetected response to a harsh-touch stimulus, and mec-4 and mec-2 null mutants are normal for this response (mec-6 not tested) (Fig. 1.4A). Furthermore, electrophysiological studies on cultured neurons revealed that mec-4 and mec-2 mutant neurons 1.4.7.1


Body Touch Sensation Gentle Touch 20

20

5 10

20

30

40

10

fast

slow

5 0

0

10

20

30

40

Ratio change (%)

10 fast slow

00

15

15 Ratio change (%)

Ratio change( %)

A

20

buzz

buzz

15

0

0

mec-4(null)

WT

10

20

30

40

Time (sec)

Time (sec)

Time(sec)

buzz

10 fast slow 5

mec-2(null)

16

40

14 12

35

20 15

10

5 00

10

-5

20

30

Time(sec)

WT

40

30

8 6 4 2 00

10

20

30

40

Ratio change (%)

10

Ratio change (%)

Ratio change (%)

Harsh Touch 25

25 20 15 10 5 00

10

20

30

40

Time (sec)

Time (sec)

mec-4(null)

mec-2(null)

B MEC-5

MEC-10

MEC-6

EGL-36 MEC-9 MEC-4

EGL-19

- - + + +

MEC-2

MEC-12, MEC-7 Fig. 1.4 Molecular model for body touch sensation. (A) Studies using the genetically encoded calcium sensor cameleon revealed that touch neurons undergo transient changes in intracellular calcium concentration during touch stimulation [76]. As shown here in the upper left panel, large transients are generated upon stimulation by a constantly moving probe (buzz) and small ones are stimulated by slow or fast pokes. Middle and right upper panels show lack of calcium transients in touch neurons from mec-4 and mec-2 null backgrounds. In the lower panels calcium transients were apparent in all three strains when worms were stimulated by harsher touch, which was delivered at the time indicated by the arrow. (B) Cartoon of a touch-transducing complex in C. elegans touch-

receptor neurons (see text for discussion of the properties of the MEC proteins depicted) and the possible mechanism of calcium transient activation. In the absence of mechanical stimulation, the channel is closed and therefore the sensory neuron is at rest. Application of a mechanical force to the body of the worm results in distortion of a network of interacting molecules that opens the channel allowing Na+ influx. This in turn causes depolarization of the neuron, which activates the voltagegated L-type calcium channel EGL-19 (with likely accessory subunit UNC-36) that produces rapid and transient changes in intracellular calcium concentration, initiating the perceptory integration of the stimulus

17

18


have resting currents indistinguishable from wild-type neurons and also depolarize in response to high K+ like wild-type neurons. These last points clearly establish that touch neurons maintain normal physiological functions in the absence of mec-4 or mec-2 – in other words mec-4 and mec-2 are needed specifically for gentle-touch responses and are not generally needed for other touch neuron activities. Although cameleon data cannot establish that MEC-4 and MEC-2 are the actual touch transducers, these experiments position them very close to the actual transduction event and thus validate the attention given to them as components of a premier model of a touchtransducing channel. Additional Insights Revealed by Imaging In Vivo Ca2+ Changes in Responding Touch Neurons Use of the cameleon Ca2+ reporter also revealed previously unknown aspects of touch receptor biology. First, touch neurons appear to be rapidly adapting receptors that sense motion better than continuous pressure. A “press” stimulus in which a probe is moved a unit distance into the worm cuticle, held steady, and then pulled away elicits Ca2+ transients only when the stimulus is moved in or moved out—the neurons do not appear sensitive to continuous pressure. When the probe is moved rapidly back and forth against the body (a “buzz” stimulus), a large Ca2+ influx is observed (Fig. 1.4A). Thus, it appears that the touch neuron is tuned to mechanically sense motion, such as the brush of an eyelash hair, the original gentle-touch stimulus applied to the study of touch sensitive behavior. A second observation is that touch neurons respond when probed in their sensory fields and do so in a roughly “all or none” fashion. Gentle touch near a sensory process virtually always elicits a calcium transient, and touch far away from the process virtually never elicits a response. Interestingly, neurons occasionally respond when the animal is touched slightly posterior to their process, but when they do respond, the magnitude of the response is roughly the same as when the touch is delivered directly over the neuron process, suggesting that neurons do not respond in a markedly graded fashion according to the position of the stimulus. A third important point is that the six “gentle” body touch neurons can sense and respond to more than gentle touch. Harsh stimuli (such as a higher velocity jab with a rigid probe) elicit neuronal Ca2+ influxes that are distinct from those induced by gentletouch stimuli, and these responses to harsh touch do not depend upon the function of MEC-4 or MEC-2 (Fig. 1.4A). This strongly suggests that touch neurons can respond to distinct mechanical stimuli using distinct molecular machinery. What might the harsh-touch sensory channel be? Current data suggest that of the 20 or so degenerin channel subunits encoded by the C. elegans genome, only degenerins mec-4 and mec-10 are expressed in the touch neurons. It is possible that a monomeric MEC-10 channel responds to harsh touch, a hypothesis that can easily be tested when mec-10 null alleles are generated. Alternatively, a distinct channel type might be used. 1.4.7.2


1.4.8

Summary: A Molecular Model for Gentle-touch Sensation How Touch Is Sensed to Elicit a Specific Behavioral Response Genetic, molecular, and functional analysis of cloned touch cell structural genes suggests a working model of the touch receptor mechanotransducing complex (Fig. 1.4B; see [23, 37, 46, 52] for discussion). The central component of this model is the candidate mechanosensitive ion channel that includes multiple MEC-4 and MEC-10 subunits. These subunits assemble to form a channel pore that is lined by hydrophilic residues in MSDII of MEC-4. Accessory subunits MEC-2 and MEC-6 associate with the core channel and may influence channel pore properties, as they do in the oocyte expression system. The MEC-4 and MEC-10 channel subunits have a transmembrane topology in which the Cys-rich domains extend into the specialized extracellular matrix surrounding the touch cell (the mantle) and the amino- and carboxy- channel termini project into the cytoplasm. MEC-5 collagen and MEC-1 and MEC-9 interaction domain-rich proteins are components of the extracellular mantle that might serve to exert gating tension on the extracellular domains of the channel. Inside the touch neuron, a unique microtubule network may associate (likely via protein linkages) with the channel to help tether the intracellular channel domains. MEC-2 may also play a role in channel tethering or localization to a microdomain environment required for proper gating. Regulated gating is expected to depend on mechanical forces exerted on the channel by some or all these proteins. A touch stimulus could deform the microtubule network, or could perturb the mantle connections (or both), to deliver the gating stimulus. Na+ influx would activate the touch receptor to signal the appropriate locomotory response via a characterized neuronal circuit. Cameleon-based reporting of physiological responses to touch stimuli have revealed that one of the signals that is triggered by activation of the putative mechanosensory channel is elevation of intracellular calcium via a specific L-type voltage-gated calcium channel, EGL-19 (Fig. 1.4B). Upon conversion of a mechanical stimulus into an electrochemical response, the touch receptor neurons activate a simple reflex circuit [14]. The touch cells activate the interneurons (AVD for backward and PVC for forward locomotion) that in turn activate motor neurons. While the touch cells form gap junctions with agonist interneurons, they form apparent chemical synapses with the antagonist interneurons. This reciprocal pattern of connectivities (for example, AVD activated and PVC inhibited for backward movement) enables locomotion in the appropriate direction to be stimulated at the same time that locomotion in the inappropriate direction is inhibited. 1.4.8.1

Notes on the Working Model The working model for touch transduction accommodates all cloned MEC proteins to postulate a complex that may gate like the mechanosensory channels that respond to sound in the hair cells of the vertebrate inner ear (reviewed in [33]). Aspects of the model that remain to be tested include: 1.4.8.2

19

20

1 The Molecular Basis of Touch Sensation as Modeled in Caenorhabditis elegans *

*

*

Whether the channel is itself directly mechanically gated. Such demonstration will require direct assay of the channel in an in vivo context (cell culture or direct recording from neurons in the nematode). The channel cannot be mechanically gated as reconstituted in Xenopus oocytes (this was to be expected, as putative accessory proteins were not co-expressed and would be unlikely assemble properly even if they were [34]). Whether specific MEC proteins associate with the channel. Once interactions are identified, definition of residues involved in interactions will fill in molecular details of the model. Whether all MEC proteins are directly involved in the channel complex rather than carrying out other functions required for touch receptor function or signal propagation. In addition, deciphering single channel properties and determining precisely how subunits influence channel gating will extend our understanding of molecular mechanisms of touch.

1.5

The C. elegans Degenerin Family: A Global Role of Degenerin Channels in Mechanotransduction?

The C. elegans genome encodes a total of 28 degenerins. Our preliminary surveys of expression patterns suggest that degenerins are expressed in specific but overlapping cell groups. Initial work suggests that at least some of these ion channel subunits (unc105, unc-8, and del-1) may function in mechanical signaling, suggesting a potential common functional theme for members of this channel class in C. elegans.

1.5.1

unc-105

Semi-dominant unc-105 alleles induce hypercontraction of body wall muscles [57]. The unc-105(sd) mutations encode amino acid substitutions in the extracellular domain [54] that render the mutant UNC-105 subunits hyperactive [31], suggesting that hypercontraction is the result of excess ion influx. Interestingly, specific alleles of sup-20/let2, an essential type IV basement membrane collagen [68], suppress unc-105(sd) hypercontraction [53, 57], indicating that sup-20/let-2 collagen may function in muscle cells similarly to mec-5 collagen in touch neurons. The working model is that unc-105 may function as a stretch-responsive channel in body wall muscle that is gated via attachment to collagen in the extracellular matrix [53]. Interestingly, UNC-105 null mutations have no apparent phenotype [57], suggesting that another degenerin coexpressed with unc-105 could redundantly supply the same function in muscle or that the unc-105 null phenotype may be a subtle behavioral defect not readily detectable.

1.5 The C. elegans Degenerin Family: A Global Role of Degenerin Channels in Mechanotransduction?

1.5.2

unc-8 and del-1

Unusual semi-dominant gain-of-function unc-8 alleles induce transient neuronal swelling [67] and severe uncoordination [8, 57], with a particularly pronounced defect in backward locomotion. unc-8 encodes a degenerin expressed in several motor neuron classes, in some interneurons, and in ASH polymodal sensory neurons [80]. Another degenerin family member, del-1(for degenerin-like), is co-expressed with unc-8 in a subset of neurons (the VA and VB motor neurons and the FLP harsh touch neurons) and is likely to assemble into a channel complex with UNC-8 in these cells [80]. What is the role of the UNC-8 degenerin channel in locomotion? Important clues came from the unc-8 null mutant phenotype and from the neuronal anatomy of motor neurons expressing unc-8. unc-8 null mutants have locomotory defects: although they move in a sinusoidal trajectory (the normal locomotion mode for C. elegans), the tracks they carve in the bacterial lawn on which they live and travel through are markedly reduced in amplitude. Thus, unc-8 null mutants do not bend as deeply as wild-type worms (Fig. 1.5A). Interestingly, some motor neurons that co-express unc-8 and del-1 (namely, the VA and VB motor neurons) have processes that have been hypothesized to be stretch-sensitive because they lack synapses to muscle or other nerves over most of their processes (originally by R. L. Russell and L. Byerly and discussed in [85] and Fig. 1.5B). Given the homology of UNC-8 and DEL-1 to candidate mechanically gated channels, we have suggested that these subunits co-assemble into a stretch-sensitive channel that might be localized to the sensory regions of the motor neuron process. In this model UNC-8/DEL-1 channels respond to displacement of the bending neuronal process during locomotion, and when they do so, they enhance signaling at the distant neuromuscular junction. The increased strength and duration of the muscle contraction they contribute is needed for a full-sized body bend. In the absence of the stretch activation signal delivered by the UNC-8/DEL-1 channel, the body bend still occurs, but with reduced amplitude (Fig. 1.5B). Although this proprioception model is based on neuroanatomy, behavior of mutant UNC-8 subunits, and proposed functional homology with MEC-4/MEC-10 channel complexes, details of the working hypothesis still await experimental verification.

1.5.2.1 A Stomatin Partner for the UNC-8 Channel Suggests a Common Composition of Degenerin Channels Genetic screens for mutations that affect anesthetic sensitivity in C. elegans identified several genes that alter sensitivity to specific anesthetics. Among these was unc-1, a gene that encodes a stomatin-like protein [61, 62]. unc-1 mutants share some uncoordinated traits with unc-8(sd) mutants, which prompted testing of unc-8 mutants for anesthetic responses and for genetic interactions with unc-1. Importantly, allele-specific interactions between unc-1/stomatin and unc-8/degenerin were identified in this survey, strongly suggesting that the channel and the stomatin, which are co-expressed in many cells, physically interact [62]. Biochemical evidence for physical association has not yet been published.

21

22


1.5 The C. elegans Degenerin Family: A Global Role of Degenerin Channels in Mechanotransduction?

3 Fig. 1.5 Modulation of locomotion by the stretch-responsive channel UNC-8 in motor neurons (proprioception model). (A) Examples of tracks left on an E. coli lawn by WT and unc-8(lf ) nematodes. While WT animals inscribe wide sinusoidal marks, unc-8(lf ) animals produce much narrower tracks due to defects in the VB motor neurons’ function caused by the absence of UNC-8 channels. (B) Features and function of VB motor neurons in body stretch sensation. Two VB motor neurons in the ventral nerve cord are depicted here. A typical VB motor neuron makes synapses to muscle near to the cell body and possesses a long nerve terminal that is synapse-free and proposed to have stretch-sensory functions. The stretch-sensitive channels are postulated to be situated in these “undifferentiated” processes. The anterior signaling by VB is proposed to be potentiated by opening of the UNC-8 channels that monitor body stretch due to local muscle contraction. Following potentiation, this motor neuron will signal to the anterior muscles to become fully contracted. At the same time another motor neuron in the middle of the body does not experience stretch and therefore remains idle. Sequential activation of motor neurons distributed along the ventral nerve cord may amplify and propagate the sinusoidal body wave. (C) Cartoon of a stretch-sensitive channel complex in C. elegans VB motor neurons (see text for hypothesized channel subunits’ function). The channel complex, by analogy with the touch-sensitive MEC-4 channel expressed in touch neurons, is hypothesized to sense stretching forces and includes the degenerin subunits UNC-8 and DEL-1 and stomatin UNC-1. Stomatin UNC-24 is needed for UNC-1 to be properly localized and may be part of the channel complex. In the absence of stretch stimulation, the channel is closed and therefore the motor neuron is at rest. When the worm bends its body, the motor neuron experiences stretch forces along its undifferentiated nerve terminal where UNC-8 channels are located. Stretch forces may distort the network of interacting molecules that opens the UNC-8 channel allowing Na+ influx. Adapted from [80]

UNC-1 protein expression and distribution is dependent upon a second stomatin domain protein, UNC-24 [66]. The unc-24 gene encodes a protein that consists of a stomatin-like domain and a lipid-transfer domain [4]. Lipid-transfer proteins are mostly soluble proteins that transfer a wide variety of phospholipids and sterols from donor membranes, mainly vesicles, to acceptor membranes. Because UNC24 is membrane-anchored at the stomatin domain, its lipid-transfer domain can act only locally, possibly regulating the composition of small areas of the lipid bilayer. UNC-1 protein is normally distributed in a punctate pattern in neuronal processes. In unc-24 null mutants, UNC-1-related puncta are greatly diminished in number in unc-24 null animals. Interestingly, in almost all neurons there are only two puncta located on each side of the nucleus. These data indicate a role for UNC-24 in UNC-1 localization to the process and suggest that UNC-24 may heteromultimerize with UNC-1 and possibly other proteins in the UNC-8 channel complex (Fig. 1.5C). It is noteworthy that UNC-24 is expressed in touch neurons and in FLPs and PVDs that mediate harsh-touch responses, such as the ones delivered by a wire prod. Thus, UNC24 is likely to regulate multiple degenerin channels and contribute to multiple mechanosensory functions.

Trp Channels May Also Contribute to Mechanosensory Functions in C. elegans In the laboratory, C. elegans moves through a bacterial lawn on a petri dish with a readily observable sinusoidal motion. When the worm collides head-on with an object such as an eyelash hair, it initiates backward motion known as nose touch avoidance. Three classes of mechanosensory neurons act in parallel to mediate the avoidance 1.5.2.2

23

24


response, including ASH, FLP, and OLQ [47]. ASH neurons mediate two other wellcharacterized behaviors: osmotic avoidance (nematodes respond to high osmotic strength solutions by avoiding them) and adverse response to noxious stimuli. TRP channels (see Tab. 8.3) have been implicated in all ASH-mediated behavioral responses. Based on sequence conservation, the TRP superfamily can be divided into three subfamilies named TRPC, TRPV, and TRPM. The C. elegans TRPV subfamily includes OSM-9 and OCR-1, -2, -3, and -4. osm-9 and ocr-2 mutants are defective in all ASH-mediated behaviors. In addition, osm-9 also acts in other neuronal types, where its functions may be different. For example, in AWA neurons OSM-9 is proposed to form a G protein-regulated transduction channel required for odorant sensitivity, and in AWC it has a role in olfactory adaptation. Thus, studies on TRPV channels implicate this channel family in mechanisms of mechanotransduction in C. elegans but also document their involvement in other sensory behaviors, suggesting that they may be gated by stimuli besides mechanical forces [81]. One possibility is that TRPV channels may set the physiological stage so that the other channels (for example, DEG/ ENaCs) can function. Studies in Drosophila [50, 84] and zebrafish [69] have strongly implicated TRP channels in mechanosensory transduction. In a Drosophila screen for touch-insensitive mutants, mutations affecting a gene named nompC (no mechanoreceptor potential) were isolated. nompC encodes a TRP channel expressed in Drosophila mechanosensory organs, including bristles, and mutants lack or have abnormal mechanosensory currents in bristle neurons during bristle deflection. Zebrafish larvae treated with morpholino antisense oligonucleotides for the homologue of nompC are deaf, and sensory hair cells no longer respond to sound stimulation [69] (see Chapter 2).

1.5.2.3 Fly and Mouse Neuronal DEG/ENaCs Influence Mechanotransduction, Supporting Conserved Roles for This Family of Proteins C. elegans DEG/ENaCs are clearly directly implicated in mechanotransduction. A key question that arises, then, is whether this role is conserved across species such that mammalian DEG/ENaCs may contribute to mechanoperception. Recent studies in Drosophila and mouse support such a conserved role.

Fly The Drosophila DEG/ENaC Pickpocket is expressed early in dendritic varicosities of the da/md abdominal peripheral neurons that function in the adult as mechanoreceptors [1]. The PPK1 channel plays an essential role in controlling rhythmic locomotion, suggesting a potential role in mechanical signaling [2]. Mouse Based on sequence similarity, mammalian DEG/ENaCs can be grouped into two subfamilies: the ENaC subfamily that includes a, b, c, and d subunits involved in Na+ reabsorption in specialized kidney and lung epithelia [9] and the ASIC subfamily. The mammalian acid-gated ASIC subfamily includes five members, with two known

1.6 Concluding Remarks

splice variants for ASIC2, that are expressed in both the central and peripheral nervous system (see Chapter 3). ASIC2a, one of the ASIC2 splice variants, is expressed in medium and large diameter mechanosensory neurons of the dorsal root ganglia (DRG) and is localized to nerve termini that are known to function as cutaneous mechanosensors [32, 59]. Functional analysis of these mechanoreceptors from an ASIC2 knockout mouse revealed that two types of low-threshold (i.e., they sense light touch) mechanosensitive fibers (the rapidly adapting [RA] and, to a lesser extent, the slowly adapting [SA]) are not able to respond to stronger displacement force stimuli by increasing the frequency of action potentials, as fibers from WT mice do. This finding has biological relevance, since the dynamic sensitivity of RA and SA mechanosensors is thought to play an important role in how we perceive gentle touch. ASIC3 has also been implicated in mechanosensation. ASIC3 is also localized to several types of mechanosensory nerve terminals, and its expression pattern partly overlaps with ASIC2a [60]. In particular, both channels are expressed in RA mechanoreceptors. Studies of nerve-skin preparations from the ASIC3 knockout mouse revealed that both the threshold sensitivity and the response frequency of AM fiber mechano-nociceptors (which respond to high-threshold stimuli-like pinching) are reduced. Interestingly, RA mechanoreceptors from ASIC3 knockout mouse respond to stronger displacement forces, but they do so with doubled firing frequency as compared to WT. To conclude, ASIC3 is involved in responding to both gentle-touch and painful stimuli and it seems to do so in distinct ways. These findings support a conserved role for DEG/ENaC family members in touch sensation, but also highlight the higher complexity of mammalian touch perception mechanisms. In fact, ASIC2a and ASIC3 knockout effects are subtle. In both knockout mice no mechanoreceptor responses are completely eliminated, and for some mechanoreceptors that normally express the ASIC2 channel, responses are unaffected. Explanations for such differences may include redundancy (other members of the family may be expressed in the same mechanoreceptors and may functionally overlap with ASIC2a and ASIC3), modulatory function of ASIC2a and ASIC3 in a heteromultimeric channel complex (similar to MEC-10), and a more crucial role of other channel families (TRP for example) in touch sensation in mammals.

1.6

Concluding Remarks

DEG/ENaC and TRP ion channels have both been implicated in mechanosensory transduction mechanisms in invertebrates and vertebrates and both have been suggested to be the primary transducers of the mechanical stimuli. Looming questions in the field of molecular mechanotransduction include the following. Are TRPs and DEG/ENaCs functionally redundant, or do they “sense” distinct stimuli? Does one channel type merely set the physiological stage so that the other can function? Do the DEG/ENaCs and TRPs physically interact or cooperate to sense force? Addressing these questions will be clearly feasible once DEG/ENaC and TRP channel function, regulation, and gating are studied in detail in both invertebrates and vertebrates.

25

26


Studying the members of the two channel families that are co-expressed in the same mechanosensory cells should clarify relative roles in mechanotransduction.

Acknowledgments

We thank our colleagues cited herein, especially D. Hall for providing touch receptor electron micrographs. Some of the work reviewed here was supported by the National Institutes of Health NINDS (1R01-NS37955).

References 1

2

3

4

5 6

7

8

Adams, C. M., Anderson, M. G., Motto, D. G., Price, M. P., Johnson, W. A., and Welsh, M. J. Ripped pocket and pickpocket, novel Drosophila DEG/ENaC subunits expressed in early development and in mechanosensory neurons. J Cell Biol 1998, 140, 143–152. Ainsley, J. A., J. M. Pettus, D. Bosenko, C. E. Gerstein, N. Zinkevich, M. G. Anderson, C. M. Adams, M. J. Welsh, and W. A. Johnson, Enhanced locomotion caused by loss of the drosophila DEG/ENaC protein Pickpocket1. Curr Biol, 2003, 13(17), 1557–1563. Ausiello, D. A., Stow, J. L., Cantiello, H. F., de Almeida, J. B., and Benos, D. J. Purified epithelial Na+ channel complex contains the pertussis toxin-sensitive Gai-e protein. J Biol Chem 1992, 267, 4759–9472. Barnes, T. M., Jin, Y., Horvitz, H. R., Ruvkun, G., and Hekimi, S. 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 1996, 67, 46–57. Beachy, J. M. Premature infant massage in the NICU. Neonatal Netw 2003, 22, 39–45. Benos, D. J., Saccomani, G., and SaribanSohraby, S. The epithelial sodium channel: subunit number and location of the amiloride binding site. J Biol Chem 1987, 262, 10613–10618. Berdiev, B. K., K. H. Karlson, B. Jovov, P. J. Ripoll, R. Morris, D. Loffing-Cueni, P. Halpin, B. A. Stanton, T. R. Kleyman, and Ismailov, I. I. Subunit stoichiometry of a core conduction element in a cloned epithelial amiloride-sensitive Na+ channel. Biophys J, 1998, 75(5), 2292–2301. Brenner, S. The genetics of Caenorhabditis elegans. Genetics 1974, 77, 71–94.

9

10

11

12 13

14

15

16

17

Canessa, C. M., Horsiberger, J. D., and Rossier, B. C. Functional cloning of the epithelial sodium channel relation with genes involved in neurodegeneration. Nature 1993, 349, 588–593. Canessa, C. M., Schild, L., Buell, G., Thorens, B., Gautschi, I., Horisberger, J. D., and Rossier, B. C. Amiloride-sensitive epithelial Na+ channel is made of three homologous subunits. Nature 1994, 367, 412–413. Chalfie, M., and Au, M. Genetic control of differentiation of the Caenorhabditis elegans touch receptor neurons. Science 1989, 243, 1027–1033. Chalfie, M., Driscoll, M., and Huang, M. Degenerin similarities. Nature 1993, 361, 504. Chalfie, M., and Sulston, J. Developmental genetics of the mechanosensory neurons of Caenorhabditis elegans. Dev. Biol. 1981, 82, 358–370. Chalfie, M., Sulston, J. E., White, J. G., Southgate, E., Thomson, J. N., and Brenner, S. The neural circuit for touch sensitivity in Caenorhabditis elegans. J Neurosci 1985, 5, 956–964. Chalfie, M., and Thomson, J. N. Structural and functional diversity in the neuronal microtubules of Caenorhabditis elegans. J Cell Biol 1982, 93, 15–23. Chalfie, M., and Thomson, N. J. Organization of neuronal microtubules in the nematode Caenorhabditis elegans. J Cell Biol 1979, 82, 278–289. Chalfie, M., and Wolinsky, E. The identification and suppression of inherited neurodegeneration in Caenorhabditis elegans. Nature 1990, 345, 410–416.

1.6 Concluding Remarks 18

19

20

21

22

23

24

25

26

27

28

29

Chelur, D. S., G. G. Ernstrom, M.B. Goodman, C. A. Yao, L. Chen, R. O’Hagan and M. Chalfie, The mechanosensory protein MEC-6 is a subunit of the C. elegans touch-cell degenerin channel. Nature, 2002 420(6916), 669–673. Christensen, M., Estevez, A., Yin, X., Fox, R., Morrison, R., McDonnell, M., Gleason, C., Miller, D. M., and Strange, K. A primary culture system for functional analysis of C. elegans neurons and muscle cells. Neuron 2002, 33, 503–514. The C. elegens sequence consortium, Genome sequence of the nematode C. elegans: a platform for investigating biology. The C. elegans Sequencing Consortium. Science, 1998, 282(5396), 2012–2018. Dijkink, L., A. Hartog, C. H. van Os, and R. J. Bindels, The epithelial sodium channel (ENaC) is intracellularly located as a tetramer. Pflugers Arch, 2002, 444(4), 549–555. Driscoll, M., and Chalfie, M. The mec-4 gene is a member of a family of Caenorhabditis elegans genes that can mutate to induce neuronal degeneration. Nature 1991, 349, 588–593. Du, H., Gu, G., Williams, C., and Chalfie, M. Extracellular proteins needed for C. elegans mechanosensation. Neuron 1996, 16, 183–194. Dumont, R. A., and Gillespie, P. G. Ion channels: hearing aid. Nature 2003, 424, 28–29. Fan, G. Y., H. Fujisaki, A. Miyawaki, R. K. Tsay, R. Y. Tsien, and M. H. Ellisman, Videorate scanning two-photon excitation fluorescence microscopy and ratio imaging with cameleons. Biophys J, 1999, 76(5), 2412–2420 Fire, A. and Mello, C. C. DNA Transformation. Methods in Cell Biology. Caenorhabditis elegans: Modern Biological Analysis of an Organism. H.F. Epstein and D.C. Shakes (eds), Academic Press, Inc., San Diego. 1995, 48, 451–482. Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E., and Mello, C. C. Potent and specific genetic interference by doublestranded RNA in Caenorhabditis elegans. Nature 1998, 391, 806–811. Firsov, D., I. Gautschi, A. M. Merillat, B. C. Rossier, and L. Schild, The heterotetrameric architecture of the epithelial sodium channel (ENaC). Embo J, 1998, 17(2), 344–352. Fricke, B., Lints, R., Stewart, G., Drummond, H., Dodt, G., Driscoll, M., and von During, M. Epithelial Na+ channels and stomatin are expressed in rat trigeminal mechanosensory neurons. Cell Tissue Res 2000, 299, 327–334.

30

31

32

33

34

35

36

37

38

39

40

41

Fukushige, T., Siddiqui, Z. K., Chou, M., Culotti, J. G., Gogonea, C. B., Siddiqui, S. S., and Hamelin, M. MEC-12, an a-tubulin required for touch sensitivity in C. elegans. J Cell Sci 1999, 112, 395–403. Garcia-Anoveros, J., Garcia, J. A., Liu, J. D., and Corey, D. P. The nematode degenerin UNC-105 forms ion channels that are activated by degeneration- or hypercontractioncausing mutations. Neuron 1998, 20, 1231–1241. Garcia-Anoveros, J., Samad, T. A., ZuvelaJelaska, L., Woolf, C. J., and Corey, D. P. Transport and localization of the DEG/ENaC ion channel BNaC1alpha to peripheral mechanosensory terminals of dorsal root ganglia neurons. J Neurosci 2001, 21, 2678–2686. Gillespie, P. G., and Walker, R. G. Molecular basis of mechanosensory transduction. Nature 2001, 413, 194–202. Goodman, M. B., Ernstrom, G. G., Chelur, D. S., O’Hagan, R., Yao, C. A., and Chalfie, M. MEC-2 regulates C. elegans DEG/ENaC channels needed for mechanosensation. Nature 2002, 415, 1039–1042. Goodman, M. B., Hall, D. H., Avery, L., and Lockery, S. R. Active currents regulate sensitivity and dynamic range in C. elegans neurons. Neuron 1998, 20, 763–772. Grunder, S., Jaeger, N. F., Gautschi, I., Schild, L., and Rossier, B. C. Identification of a highly conserved sequence at the N-terminus of the epithelial Na+ channel alpha subunit involved in gating. Pflugers Arch 1999, 438, 709–715. Gu, G., Caldwell, G. A., and Chalfie, M. Genetic interactions affecting touch sensitivity in Caenorhabditis elegans. Proc Natl Acad Sci USA 1996, 93, 6577–6582. Hamelin, M., Scott, I. M., Way, J. C., and Culotti, J. G. The mec-7 beta-tubulin gene of Caenorhabditis elegans is expressed primarily in the touch receptor neurons. Embo J 1992, 11, 2885–2893. Hamill, O. P., and Martinac, B. Molecular basis of mechanotransduction in living cells. Physiol Rev 2001, 81, 685–740. Harbinder, S., Tavernarakis, N., Herndon, L., Kinnel, M., Xu, S., Fire, A., and Driscoll, M. Genetically targeted cell disruption in Caenorhabditis elegans. Proc Natl Acad Sci USA 1997, 94, 13128–13133. Harteneck, C., Plant, T. D., and Schultz, G. From worm to man: three subfamilies of TRP channels. Trends Neurosci 2000, 23, 159–166.

27

28

1 The Molecular Basis of Touch Sensation as Modeled in Caenorhabditis elegans 42

43

44

45

46

47

48

49

50

51

52

53

Hong, K., and Driscoll, M. A transmembrane domain of the putative channel subunit MEC4 influences mechanotransduction and neurodegeneration in C. elegans. Nature 1994, 367, 470–473. Hong, K., Mano, I., and Driscoll, M. In vivo structure-function analyses of Caenorhabditis elegans MEC-4, a candidate mechanosensory ion channel subunit. J Neurosci 2000, 20, 2575–2588. Horvitz, H. R., and Sultson, J. E. Isolation and genetic characterization of cell-lineage mutants of the nematode Chaenorhabditi elegans. Genetics 1980, 96, 435–454. Huang, M., and Chalfie, M. Gene interactions affecting mechanosensory transduction in Caenorhabditis elegans. Nature 1994, 367, 467–470. Huang, M., Gu, G., Ferguson, E. L., and Chalfie, M. A stomatin-like protein necessary for mechanosensation in C. elegans. Nature 1995, 378, 292–295. Kaplan, J. M. and Horvitz, H. R. A dual mechanosensory and chemosensory neuron in Caenorhabditis elegans. Proc Natl Acad Sci USA 1993, 90, 2227–2231. Kennedy, H. J., Evans, M. G., Crawford, A. C., and Fettiplace, R. Fast adaptation of mechanoelectrical transducer channels in mammalian cochlear hair cells. Nature Neurosci 2003, 6, 832–836. Kerr, R., Lev-Ram, V., Baird, G., Vincent, P., Tsien, R. Y., and Schafer, W. R. Optical imaging of calcium transients in neurons and pharyngeal muscle of C. elegans. Neuron 2000, 26, 583–594. 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. A TRPV family ion channel required for hearing in Drosophila. Nature 2003, 424, 81–84. Kosari, F., S. Sheng, J. Li, D. O. Mak, J. K. Foskett, and T. R. Kleyman, Subunit stoichiometry of the epithelial sodium channel. J Biol Chem, 1998. 273(22), 13469–13474. Lai, C. C., Hong, K., Kinnell, M., Chalfie, M., and Driscoll, M. Sequence and transmembrane topology of MEC-4, an ion channel subunit required for mechanotransduction in Caenorhabditis elegans. J Cell Biol 1996, 133, 1071–1081. Liu, J., Schrank, B., and Waterston, R. Interaction between a putative mechanosensory membrane channel and a collagen. Science 1996, 273, 361–364.

54

55

56

57

58

59

60

61

62

63

64

Liu, K. S., and Sternberg, P. W. Sensory regulation of male mating behavior in Caenorhabditis elegans. Neuron 1995, 14, 79–89. Mitani, S., Du, H., Hall, D. H., Driscoll, M., and Chalfie, M. Combinatorial control of touch receptor neuron expression in Caenorhabditis elegans. Development 1993, 119, 773–783. Miyawaki, A., O. Griesbeck, R. Heim, and R. Y. Tsien, Dynamic and quantitative Ca2+ measurements using improved cameleons. Proc Natl Acad Sci USA, 1999, 96(5), 2135–2140. Park, E., and Horvitz, H. R. Mutations with Dominant Effects on the Behavior and Morphology of the Nematode C. elegans. Genetics 1986, 113, 821–852. Patel, A. J., Lazdunski, M., and Honore, E. Lipid and mechano-gated 2P domain K(+) channels. Curr Opin Cell Biol 2001, 13, 422–428. 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., Tarr, D. E., Hrstka, R. F., Yang, B., Williamson, R. A., and Welsh, M. J. The mammalian sodium channel BNC1 is required for normal touch sensation. Nature 2000, 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. The DRASIC cation channel contributes to the detection of cutaneous touch and acid stimuli in mice. Neuron 2001, 32, 1071–1083. Rajaram, S., Sedensky, M. M., and Morgan, P. G. Unc-1: a stomatin homologue controls sensitivity to volatile anesthetics in Caenorhabditis elegans. Proc Natl Acad Sci U S A 1998, 95, 8761–8766. Rajaram, S., Spangler, T. L., Sedensky, M. M., and Morgan, P. G. A stomatin and a degenerin interact to control anesthetic sensitivity in Caenorhabditis elegans. Genetics 1999, 153, 1673–1682. Rupp, F., D. G. Payan, C. Magill-Solc, D. M. Cowan, and R. H. Scheller, Structure and expression of a rat agrin. Neuron, 1991, 6(5), 811–823. Savage, C., Hamelin, M., Culloti, J. G., Coulson, A., Albertson, D., and Chalfie, M. mec-7 is a beta tubulin gene required for the production of 15 protofilament microtubules in Caenorhabditis elegans. Genes Dev 1989, 3, 870–881.

1.6 Concluding Remarks 65

66

67

68

69

70

71

72 73

74

75

Schild, L., Canessa, C. M., Shimkets, R. A., Gautschi, I., Lifton, R. P., and Rossier, B. C. A mutation in the epithelial sodium channel causing Liddle disease increases channel activity in the Xenopus laevis oocyte expression system. Proc Natl Acad Sci USA 1995, 92, 5699–5703. Sedensky, M. M., Siefker, J. M., and Morgan, P. G. Model organisms: new insights into ion channel and transporter function. Stomatin homologues interact in Caenorhabditis elegans. Am J Physiol Cell Physiol 2001, 280, C1340–C1348. Shreffler, W., Magardino, T., Shekdar, K., and Wolinsky, E. The unc-8 and sup-40 genes regulate ion channel function in Caenorhabditis elegans motorneurons. Genetics 1995, 139, 1261–1272. Sibley, M.H., J.J. Johnson, C.C. Mello, and Kramer, J. M. Genetic identification, sequence, and alternative splicing of the Caenorhabditis elegans alpha 2(IV) collagen gene. J Cell Biol, 1993, 123(1), 255–264. Sidi, S., Friedrich, R. W., and Nicolson, T. NompC TRP channel required for vertebrate sensory hair cell mechanotransduction. Science 2003, 301, 96–99. Snyder, P. M., Cheng, C., Prince, L. S., Rogers, J. C., and Welsh, M. J. Electrophysiological and biochemical evidence that DEG/ ENaC cation channels are composed of nine subunits. J Biol Chem 1998, 273, 681–684. Snyers, L., Umlauf, E., and Prohaska, R. Oligomeric nature of the integral membrane protein stomatin. J Biol Chem 1998, 273, 17221–17226. Stewart, G. W. Stomatin. Int J Biochem Cell Biol 1997, 29, 271–274. Stewart, G. W., Argent, A. C., and Dash, B. C. Stomatin: a putative cation transport regulator in the red cell membrane. Biochim Biophys Acta 1993, 1225, 15–25. Stewart, G. W., and Turner, E. J. The hereditary stomatocytoses and allied disorders: congenital disorders of erythrocyte membrane permeability to Na and K. Baillieres Best Pract Res Clin Haematol 1999, 12, 707–727. Sulston, J. E., Schierenberg, E., White, J. G., and Thomson, J. N. The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev Biol 1983, 100, 64–119.

76

77

78

79

80

81

82

83

84

85

Suzuki, H., Kerr, R., Bianchi, L., FrokjarJensen, C., Slone, D., Xue, J., Gerstbrein, B., Driscoll, M., and Schafer, W. R. In vivo imaging of C. elegans mechanosensory neurons demonstrates a specific role for the mec4 channel in the process of gentle touch sensation. Neuron 2003, 39, 1005–1017. Tabara, H., Grishok, A., and Mello, C. C. RNAi in C. elegans: soaking in the genome sequence. Science 1998, 282, 430–431. Tavernarakis, N., and Driscoll, M. Molecular modeling of mechanotransduction in the nematode Caenorhabditis elegans. Annu Rev Physiol 1997, 59, 659–689. Tavernarakis, N., and Driscoll, M. Caenorhabditis elegans degenerins and vertebrate ENaC ion channels contain an extracellular domain related to venom neurotoxins. J Neurogenet 2000, 13, 257–264. Tavernarakis, N., Shreffler, W., Wang, S., and Driscoll, M. unc-8, a DEG/ENaC family member, encodes a subunit of a candidate mechanically gated channel that modulates C. elegans locomotion. Neuron 1997, 18, 107–119. Tobin, D. M., Madsen, D. M., Kahn-Kirby, A., Peckol, E. L., Moulder, G., Barstead, R., and Bargmann, C. I. Combinatorial expression of TRPV channel proteins defines their sensory functions and subcellular localization in C. elegans. Neuron 2002, 35, 307–318. Vogel, B. E., and Hedgecock, E. M. Hemicentin, a conserved extracellular member of the immunoglobulin superfamily, organizes epithelial and other cell attachments into oriented line-shaped junctions. Development 2001, 128, 883–894. Waldmann, R., Champigny, G., Voilley, N., Lauritzen, I., and Lazdunski, M. The mammalian degenerin MDEG, an amiloride-sensitive cation channel activated by mutations causing neurodegeneration in Caenorhabditis elegans. J. Biol. Chem. 1996, 271, 10433–10436. Walker, R. G., Willingham, A. T., and Zuker, C. S. A Drosophila mechanosensory transduction channel. Science 2000, 287, 2229–2234. White, J. G., Southgate, E., Thomson, J. N., and Brenner, S. The structure of the nervous system of Caenorhabditis elegans. Philos Trans R Soc Lond 1986, 314, 1–340.

29

31

2

Transduction Channels in Hair Cells Robert Fettiplace

2.1

Introduction

Hair cells are the sensory neurons of the vertebrate inner ear, comprising the organs of hearing (the cochlea) and balance (three semicircular canals, a saccule, and the utricle). They also serve as receptors for the lateral line neuromasts on the skin of fish and primitive amphibia. A hair cell detects mechanical stimuli by deflection of its hair bundle, an array of between 20 and 300 elongated actin-filled stereocilia projecting from the top of the cell [1]. In the cochlea, hair bundles are mechanically vibrated by motion of an overlying gelatinous flap known as the tectorial membrane to which they are attached. The stereocilia are arranged in several ranks of increasing height and are interconnected by extracellular filaments [2, 3] to ensure that all move in synchrony during bundle deflection. One class of filament, the tip link, extending from the apex of each stereocilium to the side of its taller neighbor [4, 5], is essential for transduction [6]. Thus, deflection of the bundle towards its taller edge, rotating the rigid stereocilia about their basal insertion into the cell apex [7], is proposed to exert tension on the tip links [4, 8]. This transmits force directly or indirectly to mechanoelectrical transduction (MET) channels located near the tips of the stereocilia (Fig. 2.1 [9]). When opened by hair bundle displacement, the transduction channels allow influx of cations to generate a depolarizing receptor potential at a cell resting potential of about –50 to –70 mV. Mainly because of the paucity of hair cells in the inner ear epithelia, the MET channel has been difficult to clone and a substantive clue to its molecular identity has only recently appeared [10]. It has been provisionally classified as a member of a new branch of the TRP (transient receptor potential) family, TRPN1. However it has not yet been expressed in a heterologous system and, unlike the mechanosensitive bacterial MscL channel [11], there is little information about its behavior under controlled conditions. As a consequence, the biophysical properties of the channel have been wholly derived from measuring transduction currents in intact hair cells during Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

32

2 Transduction Channels in Hair Cells

Fig. 2.1 Tip links and hair cell transduction. (A) Transmission electron micrograph showing the tip link from the apex of the shorter stereocilium of a guinea pig outer hair cell bundle to the electron-dense plaque on the side of the neighboring taller stereocilium. Note the “tenting” of the membrane, suggesting that the tip link (arrowed) tugs on the stereociliary membrane, and also the electron-dense contact region between the two stereocilia, which has been suggested as an alternative location for the MET channel (courtesy of Y. Katori, D.N. Furness, and C.M. Hackney). (B) Schematic of top of the hair cell, illustrating how deflection of the hair bundle causes the stereocilia to bend at their rootlets where a subset of actin filaments insert into the cuticular plate. Stereociliary displacement exerts force via the tip link to open the MET channel and, in the intact hair cell epithelium, allows influx of K+ and Ca2+

manipulation of the hair bundle. The interpretation of such responses (for example, in terms of gating kinetics or actions of Ca2+) may be complicated by the way in which the mechanical stimulus is coupled to the channel. The behavior of the macroscopic current may therefore reflect intrinsic channel properties as well as the micromechanics of channel connections. This review will consider both the performance and molecular identity of the MET channel. Topics to be covered include how it is gated and whether its sensitivity and kinetics are sufficient for the needs of transduction, especially in the cochlea. What are its intrinsic properties, ionic selectivity, pharmacological profile, and single-channel conductance, and do these identify or exclude candidates for channel structure? Finally, what is the role of Ca2+ in channel operation? Ca2+ permeates readily and also modulates channel gating at both external and internal faces, but the mechanism and significance of its action are still incompletely understood. 2.2

Gating Mechanism: Channel Kinetics

The fine mechanical sensitivity and fast kinetics that are hallmarks of hair cell transduction have been amply documented in both non-mammalian vertebrates [12, 13] and mammals [14–16]. The probability of channel opening, assayed by recording MET currents in individual hair cells, can be fully modulated by between 100 and 500 nm displacements of the tip of the hair bundle, depending on the preparation. A maximum displacement of 100 nm, although smaller than the diameter of a single

2.1 Introduction

Fig. 2.2 Activation of transduction currents in auditory hair cells. Rapid deflection (Dx) of the hair bundle with a piezoelectric stimulator connected to a rigid glass fiber evokes fast onset METcurrents. In the turtle (A), the activation time constant decreases from 400 ls to 50 ls with increasing stimulus amplitude. In the rat (B) the time course of current activation is faster than in the turtle and is

limited by the stimulus onset, time constant approximately 50 ls. The mammalian responses also show adaptation at a faster rate than in the turtle, dotted line denoting fit with a time constant of 300 ls. In both experiments the fluid bathing the hair bundle contained 0.05 mM Ca2+ and the hair cell had a holding potential of –80 mV at room temperature (22 8C)

stereocilium, would be achieved in the cochlea only at the very highest sound levels [17]. Step deflections of the bundle with abrupt onset evoke MET currents that develop with a sub-millisecond time course (Fig. 2.2). The insignificant delay (< 20 ls) between the mechanical stimulus and channel opening argues that external force is fed directly to the MET channels without any of the biochemical amplification that occurs in other sensory receptors [12]. As might be expected if the stimulus directly affects the channel’s opening rate, the MET current activates with a principal time constant that depends upon stimulus amplitude. The activation time constant decreases with the size of bundle displacement from 400 ls for small stimuli to less than 50 ls for saturating ones (Fig. 2.2 [12, 13]). The channel’s activation kinetics are rapid even though measured in non-mammalian vertebrates, i.e., frog or turtle. However, when corrected for differences in body temperature, they may be too slow to explain the speed of transduction in auditory hair cells of mammals with high-frequency hearing. An indirect measure of transduction in the mammalian cochlea can be obtained from the cochlear microphonic, the summed extracellular voltage that is dominated by outer hair cell transduction currents [18, 19]. Cochlear microphonics can be observed for tonal stimuli with frequencies of at least 50 kHz [20], implying that channel gating in the mammalian cochlea can occur on a cycle-by-cycle basis up to this frequency. Such a frequency response would require the channel to activate with an unprecedented time constant of (2p. 50)–1 ms or 3 ls. It is possible that there are kinetic variations attributable to differences in channel subunit composition among species and that the MET channel in mammalian cochlear hair cells is optimized for the rapid gating needed for high-

33

34


frequency hearing. Because of the bandwidth of the patch-clamp recording system and the limited speed with which a piezoelectric stimulator can deflect the bundles, the kinetic performance of MET currents in mammals is not known for certain. However, recent measurements suggest that they activate at room temperature in less than 50 ls even for small displacements, over five- to tenfold faster than in frogs or turtles under the same conditions (Fig. 2.2; [15]). The fast kinetics and high sensitivity of the hair cell MET channel distinguish it from those mechanically sensitive channels (for example, MscL) gated by changes in tension of the lipid bilayer [21]. These properties probably require the hair cell channel protein to be tethered to both the intracellular cytoskeleton and the extracellular tip link so that force can be applied directly across the channel without delays imposed by viscous damping.

2.2.1

Tip Links and Gating Springs

The tip links, stretching obliquely between the vertex of each stereocilium and the lateral membrane of its taller neighbor (Fig. 2.1; [4, 5]) are central to hair cell transduction. They explain theoretically the polarization of transduction, whereby only those movements of the hair bundle along its axis of symmetry are detected [22]. Deflections towards the taller edge of the bundle that tug on the tip links open the MET channels, whereas deflections of opposite polarity relax the links and close the channels. The importance of the tip links is supported by two lines of experimental evidence. Firstly, exposing hair cells to sub-micromolar Ca2+, buffered with BAPTA or EGTA, destroys the links and simultaneously abolishes transduction [6, 23, 24]. Secondly, experiments to localize the MET channels show that they are confined to the tips of the stereocilia in close proximity to tip-link insertion. These experiments include defining the location of maximum susceptibility to iontophoresis of the channel blocker dihydrostreptomycin [25] or visualizing the Ca2+ influx through open channels using intracellular calcium dyes and confocal microscopy [26, 27]. An alternative location for the channels is in the contact region just below the stereociliary tips where short lateral connections can be seen between the membranes of adjacent stereocilia (Fig. 2.1; [28]). Placement at this site is supported by immunogold labeling with an antibody raised against the binding site for amiloride, which blocks the transduction channels. A key premise of models for transduction is that the MET channels are opened by force applied by elastic elements, the “gating springs” that are stretched when the hair bundle is deflected (Fig. 2.3 [8]). One end of the spring is driven by bundle motion, and the other end is attached to the channel’s hypothetical gate. The gating springs have previously been identified with the tip links. The structure of the tip links is not known [29]; however, their coiled double-helical structure suggests that they are inextensible [30] and are more likely rigid connections for force transmission. Alternative sites for the series compliance of the gating springs lie in the intracellular cytoskeletal attachments of the channel or in the channel itself.

2.1 Introduction

Fig. 2.3 Cartoon of MET channel activation and adaptation. Hair bundle displacement tensions the tip link (T), which acts via gating springs that deliver force to the channel to increase its probability of being open. The gating springs are depicted as connecting to the channel both externally (Gs) and internally (Gs’). Ca2+ entering the stereocilium through the open channel may promote channel closure and evoke adaptation by binding at the inner face of the channel and/or by altering the stiffness of the internal gating spring Gs’. Gs’ may represent a cytoskeletal attachment

In the model’s simplest form [8], the channel occupies two states, closed and open (C$O), with the transition rates modulated by stimulus energy. The open probability of the channel (pO) is then a sigmoidal function of bundle displacement (X): pO(X) = [1 + exp (–z(X – XO)/kBT)]–1

(1)

where kB is the Boltzmann constant, T is absolute temperature, XO is the displacement for half-activation, and z is the single-channel gating force applied at the top of the bundle. The single-channel gating force is a measure of the channel’s mechanical sensitivity and has a value in frog or turtle of 0.3 pN [31, 32]. It can be used to specify the range of displacements over which the channel activates. Thus, the difference in bundle position from where the channel starts to activate (pO = 0.12) to where it is almost saturated (pO = 0.88) is equal to 4 kBT/z [31]. The operating range defined in this manner varies from 50 to 250 nm for different hair cells of the turtle cochlea [13, 32], frog saccule [8, 33], and mouse cochlea [34]. In the theoretical analysis, z is equal to c KGSd where KGS is the stiffness of the gating spring, d is a constant identified with the distance by which the gate moves during channel opening (5 nm), and c is a dimensionless factor (0.1) that depends on bundle geometry [35]. c symbolizes a transformer ratio, the reduction in displacement at the tip link compared to that at the top of the bundle. Variation in c in hair cells from different end organs and species can in theory generate a range of values for the single-channel gating force (0.1–0.8 pN [31]). Since c is inversely proportional to bundle height, a change in bundle dimensions provides a means of altering the sensitivity of transduction, matching it to the needs of the hair cell in a particular end-organ. The hair bundle may therefore be regarded as an accessory structure

35

36


for converting forces and displacements occurring during inner ear stimulation to the molecular scale appropriate for deforming the MET channel protein.

2.2.2

Gating Compliance

An important and surprising feature of the gating-spring model for channel activation is that it can be used to derive the effective stiffness, Kch, of the gating spring-channel complex: Kch = KGS – z2pO (1 – pO) / kBT

(2)

The second (negative) term represents a reduction in stiffness referred to as “gating compliance” [8] that depends on the channel opening. Thus, the change in conformation of the channel between closed and open states introduces a phenomenological compliance in series with the gating spring. This signifies reversibility in transduction: a mechanical stimulus opens the channel, but, conversely, opening the channel can itself generate force. Such reversibility contrasts with the rectification in those sensory receptors (for example, photoreceptors) where a signal is amplified in an enzymatic cascade prior to activation of the transduction channel. An implication of Eq. (2) is that the force-displacement relationship of the hair bundle is nonlinear over the range of bundle positions where the channel is activated: it predicts that the bundle will not behave like a simple Hookian spring. The existence of such nonlinearity has been verified in measurements of bundle stiffness (Fig. 2.4; [8]). These are performed by applying force stimuli, usually with a flexible glass fiber more compliant than the bundle, and monitoring the resultant motion by photodiode imaging of the bundle or the attached fiber [8, 32, 34, 36, 37]. Experimental observation of the appropriate nonlinearity supports the model for gating and furthermore allows independent estimates of several of the theoretical parameters (z, KGS, d). Thus, the gating-spring model implemented in terms of a two-state (C$O) channel is a powerful tool for extrapolating from experimental measurements of macroscopic transduction current and bundle stiffness to some fundamental properties of the MET channel.

2.2.3

Three-state Channel Schemes

Several aspects of MET channel gating are not fully described by the simple two-state channel scheme but are better fit with a three-state channel, with two closed states and one open state (C1$C2$O). The most significant discrepancy is that the activation curve (pO–X relationship) for the MET channel deduced from recording MET currents in response to bundle displacement often does not show the symmetrical shape expected of a single Boltzmann equation (Eq. (1)). Instead, it has an asymmetric form in which the current saturates more gradually for positive displacements than for nega-

2.1 Introduction

Fig. 2.4 Nonlinear stiffness of the hair bundle. (A) The hair bundle was displaced with a flexible glass fiber, delivering a family of force steps (top) that generated MET currents (middle) and associated bundle displacements (bottom). The largest force step was 175 pN. Note the adaptation in the MET current for small stimuli. (B) The force-displacement relationship (top) and the current-displacement relationship (bottom, filled circles) were de-

rived from the records in (A); both MET currents and bundle movements were measured 1 ms after the onset of the force step. Hair bundle stiffness was calculated by differentiating the force-displacement results and is plotted against displacement (bottom, open circles). Note the reduction in stiffness in the region over which the MET channel activates. Taken from [32] with permission

tive ones, where a sharp corner is seen (Fig. 2.4). This type of asymmetry can be better described with a three-state channel [12–14]. In one version of the three-state model [31], the gating spring was postulated to be less compliant when the channel is in the first closed state (C1) than in either of the other two states (C2 or O). This was realized by assuming that a portion of the spring is immobilized or “latched” when the channel is in C1. A second refinement, a necessary correlate of the model, is that the gating springs eventually slacken for large negative displacements and therefore the energy of the channel becomes independent of displacement. The model reproduced the observed asymmetry of the channel’s activation curve. It also predicted a difference in bundle stiffness between positive and negative displacements that is sometimes found experimentally. However, there may be other explanations for this result (for example, an effect of intracellular Ca2+ on the stiffness of the gating spring). An alternative thermodynamic treatment of the three-state scheme [38] does not require the gating-spring stiffness to differ between states but assumes that the three

37

38


states are engaged at different bundle positions, rather than being modulated over the entire stimulus range. Differential engagement was incorporated to account for the finite minimum open probability at bundle positions more negative than –50 nm. This treatment too successfully described the asymmetry in the pO-X relationship and predicted nonlinearity in the bundle mechanics. Differences in energies between the closed and open states in this model were found to be 5 kBT, somewhat smaller than the >10 kBT estimated from the two-state model [39]. An extension of the threestate model where intracellular Ca2+ affects the energy gaps between the different states predicts Ca2+-dependent adaptation in the channel opening [38]. The role of Ca2+ in channel adaptation will be discussed in Section 2.4.1.

2.3

Ionic Selectivity

The MET channel discriminates little between monovalent cations but is highly permeable to divalent cations, especially Ca2+ [40, 41]. The permeability ratios for monovalent cations determined from measuring reversal potentials for the transduction current are Li+, 1.14; Na+, 1.0; K+, 0.96; Rb+, 0.92; and Cs+, 0.79 [23, 41]. The equivalent permeabilities for divalent cations relative to Na+ are Ca2+, 3.8; Sr2+, 2.3; Ba2+, 2.2; Mg2+, 2.0; and Na+, 1.0, and they indicate only a modest preference for Ca2+ over monovalent cations [41]. However, they too were based on interpreting reversal potentials using the constant field equation, which assumes no interaction between different ionic species in the pore. This is unlikely to be the case, given the blocking action of external calcium on the flux of monovalent ions [23, 42] and the anomalous mole fraction effects seen with ion mixtures [43]. A more pertinent measure of the channel’s selectivity is obtained from the fraction of current carried by the different ions. This indicates an effective permeability ratio for Ca2+ over monovalent ions of more than 100:1 [42, 44]. The proportion of transduction current contributed by Ca2+ was inferred in the turtle [42] by loading hair cells with 1 mM of a calcium-sensitive dye and using the change in dye fluorescence to assay the Ca2+ influx. The calcium fluorescence was calibrated under conditions where Ca2+ was the sole charge carrier. This technique had previously been applied to the cyclic nucleotide-gated channel [45] (see Chapter 5). In saline with 2.8 mM Ca2+ and 130 mM Na+, over half the current was found to be carried by Ca2+ [42], and even in an endolymph-like solution containing only 50 lM Ca2+, Ca2+ carried at least one-tenth of the current (Fig. 2.5). Hair cells in vivo form a tight epithelium, separating two fluids of different ionic composition. The basolateral synaptic pole of the hair cells is exposed to Na+-rich perilymph similar to normal extracellular fluid, whereas the apical transducing pole is bathed in endolymph with high K+ and low Ca2+. The Ca2+ concentration in endolymph, in which the hair bundles are immersed, has been estimated in different preparations as between 20 and 60 lM [23, 46, 47]. The high flux of Ca2+ through the channel when exposed to low-Ca2+ endolymph is important physiologically because of the ion’s role in regulating adaptation of the MET channels. Although the channel is

2.3 Ionic Selectivity

Fig. 2.5 Fraction of MET current carried by Ca2+ in different external Ca2+ concentrations. (A) Fractional Ca2+ current determined by two methods: from measurements of Ca2+ influx using a cytoplasmic Ca2+-fluorescent indicator with Na+ (filled circles) or K+ (filled triangles) as the external monovalent ion; from the ratios of the MET current in Tris+ to the current in Na+ (open circles) assuming that Tris+ does not permeate the channel.

Note that in 0.06 mM Ca2+ (a concentration similar to that in endolymph) the divalent ion still carries 10–20 % of the total current. (B) Mean MET current carried by Na+ (filled circles) and Ca2+ (open circles). The dependence of Na+ current on extracellular Ca2+ was fitted with a single inhibitory-site model with inhibition constant KI = 1 mM. Modified from [42]

equally permeable to both Na+ and K+, substitution of K+ for Na+ as the chief monovalent ion increased the fractional current carried by Ca2+ and augmented its intracellular effect on adaptation (Fig. 2.5; [42]). In assessing the contributions of Ca2+ in vivo, there are conflicting requirements because the ion has a dual action on the channel: Ca2+ entry regulates its gating, but extracellular Ca2+ also blocks the channel and hence diminishes the total transduction current. Thus, the composition of endolymph, rich in K+ and unique among body fluids with a low level of Ca2+, may be specialized to compromise between the largest monovalent current and the sufficient Ca2+ influx to the control adaptation. A variety of larger organic cations can traverse the channel with permeability ratios relative to Na+ of 0.27 for choline, 0.16 for TMA, and 0.14 for TEA [41]. A recent finding is that the cationic styryl dye FM1-43 (molecular weight = 451) can also pass through the channel, its fluorescence on binding to intracellular membranes providing a sensitive indication of rapid influx [48, 49]. Channel permeability to such a large molecule has been explained by its elongated and linear structure with a cross-section not much greater than that of TEA [48]. This discovery is technically significant because intracellular accumulation of fluorescent FM1-43 can be used as a monitor of normal MET channel function without the need for invasive electrical recording. This method has been exploited to screen rapidly for genetic mutants lacking transduction [10].

39

40


2.3.1

Blocking Compounds

Both Ca2+ and Mg2+ block the MET channel at millimolar extracellular concentrations (Fig. 2.5; [23, 42]), as do other inorganic cations such as Gd3+ [50] and La3+ [41]. The channel is also blocked by a broad spectrum of organic compounds (Tab. 2.1), including the calcium channel antagonists D600 [51, 52], cis-diltiazem [51], and FM1-43 [48]. Familiar antagonists of hair cell transduction, such as aminoglycoside antibiotics like dihydrostreptomycin (DHS), are also polycations. These antagonists share a number of properties that indicate that they may compete with Ca2+ for a binding site in the external mouth of the pore. They are effective only from the extracellular surface, and their block is voltage-dependent and diminished by depolarization that opposes the entry of the cation into the pore [41, 53]. Furthermore, the block is subject to competition from extracellular Ca2+: raising the Ca2+ concentration alleviates channel block with FM1-43 [48] and increases the KI for DHS (Tab. 2.1). Despite the blocking action of extracellular Ca2+, the current-voltage (I-V) relationship for the MET channel does not show pronounced rectification in control salines containing 1–4 mM Ca2+ [13, 14, 41, 55] and is approximately ohmic for membrane potentials between –50 and +50 mV. I-V relations with a modest outward curvature at voltage extremes of 100 mV have been fit [14, 55] to a single energy barrier model: I = k(exp((1-d)(V–Vrev)/Vs) – exp(-d(V–Vrev)/Vs))

(3)

where Vrev is the reversal potential (0 mV), Vs is a measure of the steepness of rectification (50 mV), d is the fractional distance of the energy barrier from the outside boundary of the membrane’s electric field (0.5), and k is a proportionality constant. Tab. 2.1

Blocking agents of the hair cell transduction channel

Blocker

KI (lM)

nH

Reference

Ca2+ Gd3+ Tetracaine cis-Diltiazem D600 Dihydrostreptomycin Dihydrostreptomycin Dihydrostreptomycin Dihydrostreptomycin

1000 10 608 236 111 8 (0.25 mM Ca) 23 (2.5 mM Ca) 44 (5 mM Ca) 0.8 (low CF*, 0.25 mM Ca) 6 (high CF, 0.25 mM Ca) 50 53 24 6 2.3 2.4

1 1.1

42 50 51 51 51, 52 53 54 53 56

Amiloride Amiloride Amiloride Benzamil Curare FM1-43

1 0.9 1 1 1.7 2 1.6 1.0 1.2

57 55 56 55 58 48

KI: half-blocking concentration; nH: Hill coefficient; CF: characteristic frequency

2.4 MET Channel Adaptation

However, the I-V relationship does acquire significant outward rectification in the presence of external blockers such as DHS and amiloride, with greater attenuation for inward current as the cationic blocker is swept into the channel at negative membrane potentials [41, 53, 55]. In its ionic selectivity and pharmacological signature, the MET channel resembles the cyclic nucleotide-gated (CNG) channel. Both are non-selective cation channels with similar permeability sequence for alkali cations and high selectivity for Ca2+. Neither shows the same degree of discrimination for Ca2+ over monovalent cations found with voltage-dependent Ca2+ channels, but, nevertheless, their high permeability to Ca2+ has important physiological consequences. In each case Ca2+ influx through channels activated by the sensory stimulus modulates transduction and promotes adaptation (see Chapter 5). Both types of channels are also subject to block by 1 mM Ca2+ at the external face and are affected by some of the same antagonists, such as amiloride, diltiazem, D600, and tetracaine (for CNG channels: [59, 60]). This might imply a similar pore structure, but there are significant dissimilarities. The MET channel is blocked by aminoglycoside antibiotics and has a five- to tenfold larger unitary conductance. Most conspicuously, it is significantly permeable to organic cations such as choline, TMA, and TEA, none of which pass through the CNG channel (PCat/PNa < 0.019; [61]). Indeed the MET channel is anomalous compared to other channels in displaying a high selectivity for Ca2+ while still being substantially permeable to organic ions. For example, its choline permeability is twice that for the nicotinic acetylcholine receptor [62]. This implies a wide pore, a property that may be linked to the unusually large single-channel conductance (>100 pS).

2.4

MET Channel Adaptation

The high Ca2+ permeability of the MET channel implies that its opening by a mechanical stimulus promotes influx and intracellular accumulation of Ca2+. The local rise in Ca2+ acts as a feedback signal to regulate channel opening just as it does in other sensory receptors; control may be exerted either through resetting the mechanical input or by directly interacting with the channel to affect its gating. Such regulation is referred to as adaptation [63, 64] and is manifested as a decline in the transduction current during a sustained deflection of the bundle. This reflects a parallel translation of the pO-X relation along the displacement axis in the direction of the stimulus to keep the channels near their maximal sensitivity. Two main mechanisms can be distinguished based on their speed and range: slow adaptation, with a time constant (sA) of 10–100 ms and a working range of 1 lm [65, 66], and fast adaptation, with a sA of 0.3-5.0 ms [13, 67] and a working range of 0.1 lm. Slow adaptation has been proposed to operate through the action of an unconventional myosin that alters the mechanical input, possibly by adjusting the tension in the tip links [68]. In one scenario, the upper attachment point of the tip link is ferried up and down the stereocilium by attachment through the plasma membrane to an array of myosins that track along sub-membranous actin filaments [69]. In mouse vestibular

41

42


hair cells, myosin-1c has been implicated as the isoform driving slow adaptation [70]. Modification of the ATP-binding site on myosin-1c renders it susceptible to block by ADP analogues [71], and introducing such analogues through the recording pipette abolishes slow adaptation. In mouse cochlear hair cells, a distinct isoform, myosin7a, may perform a supportive role in slow adaptation [72]. In contrast, fast adaptation is unaffected by agents that interfere with the myosin ATPase [70, 73,], and furthermore it has kinetics too fast to be compatible with the slow cycle time of an ATPase. However, despite the differences in underlying mechanism, fast and slow adaptation mechanisms are both controlled by Ca2+ entering the stereocilia through the MET channels [13, 33, 42]. Furthermore, Ca2+ action in one or both components of adaptation may be implemented through binding to calmodulin [74], which has been shown immunohistochemically to be concentrated at the tips of the stereocilia [75, 76]. Besides these two types of adaptation, there is evidence for at least one other slower process affecting the operating point of the MET channels. Extracellular application of a membrane-permeant form of cyclic AMP causes a positive shift in the pO-X relation along the displacement axis [67]. This shift can be as large as 1 lm with no effect on fast adaptation or on the slope of the pO-X relation. The pathway may involve stimulation by cAMP (produced by action of a Ca2+-dependent adenyl cyclase) of protein kinase A, which in turn phosphorylates the channel or the myosin motor [77]. 2.4.1

Ca2+ Regulation of Adaptation

The Ca2+ dependence of fast adaptation has been extensively documented in turtle auditory hair cells. Experiments manipulating either extracellular or intracellular Ca2+ have shown it to have a dual effect on the adaptation time constant (sA) and the open probability (pO) of the MET channels at rest. Lowering extracellular Ca2+ slows sA and shifts the pO–X relation in the negative direction, thus increasing the resting pO [13, 78]. Similar effects are produced by loading hair cells with high concentrations of calcium buffer (1–5 mM BAPTA), which argues for an intracellular action of Ca2+. Because relatively large concentrations of BAPTA, which has a fast Ca2+-binding rate, were needed to affect adaptation, the site of calcium’s action is probably not far from the channel. By comparing the extent of the shift in the pO-X relation with different BAPTA concentrations, the distance Ca2+ diffuses to its target was assessed as no more than 15–35 nm from the mouth of the channel [78]. This distance was obtained by computing Ca2+ gradients along the stereocilia from the tip where the channel was assumed to be located. The Ca2+ gradient was steeper with higher buffer concentrations. Finally, the rate of adaptation (1/sA) is proportional to the amount of Ca2+ entering the stereocilia, assayed by Ca2+ imaging bundles filled with the fluorescent dye Ca-Green-1 [42]. The proportionality extended up to rates of 1.4 ms–1, which implies that the subsequent molecular events underlying adaptation must occur in less than 1 ms. The simplest hypothesis based on this collection of results is that fast adaptation requires a direct interaction of Ca2+ with the MET channels (possibly via a regulatory subunit) to modulate their probability of opening (Fig. 2.3; [13, 73, 78]. As extra sup-

2.4 MET Channel Adaptation

Fig. 2.6 Model of activation of MET channel. (A) Kinetic scheme in which the channel is converted from closed (C) to open (O) states by bundle displacement (x) modulating rate constants K1, K2, and K3 and can also bind two Ca2+ ions. Ki are of form exp[ai(X – Xa1)], where X is bundle displacement and ai and Xa1 are constants. (B) Open probability of the channel (Popen) for hair bundle displacements of different amplitude calculated using the kinetic scheme in (A). Ca2+ ions bind to the closed state with sequential affinities of 20 lM

and 10 lM and to the open state with affinities of 2 lM and 5 lM. Ca2+ is cleared from the internal mouth of the channel with a time constant of 0.1 ms. Note that channel activation accelerates for stimuli of increasing amplitudes and that lowering the channel current carried by Ca2+ (ICa) from 4 pA to 1 pA slows both activation and adaptation rates. The bundle’s resting position is indicated by the stimulus level following the step. Note that lowering Ca2+ increases Popen at this resting position

port for this notion, intracellular Ca2+ has been found to alter the time constant of channel activation as well as adaptation [79], indicating that its action is closely linked with channel gating. If Ca2+ acts by binding to calmodulin, then kinetics demands that calmodulin be constitutively bound to the channel. The interaction of Ca2+ with the MET channel can be expressed in terms of a kinetic scheme [Fig. 2.6; 80]) with the binding of multiple (two or more) Ca2+ ions being required to achieve the requisite sensitivity. This scheme can reproduce the effects of Ca2+ on the activation and adaptation kinetics of the channel solely by altering the Ca2+ influx (Fig. 2.6). An alternative site for Ca2+ to act is on the gating springs to reduce their stiffness (KGS) (Fig. 2.3; [81]). This latter mechanism acting alone would reduce the gating force z, and also the slope of the pO-X relationship, which would extend the channel’s operating range.

43

44


2.4.2

The Function of Adaptation

The accepted role of adaptation in sensory receptors is to preserve transducer sensitivity for small changes in stimulus about a larger background level. The limited operating range of the MET channels requires that transduction be endowed with one or more adaptation mechanisms to maintain the channels near their maximal sensitivity. In the inner ear, adaptive processes exist peripherally to the hair cells to prevent large static displacement being imposed on the cells, shielding them from over-stimulation and damage. These include the helicotrema, the shunt between the two perilymphatic compartments of the cochlea, which acts like a high-pass filter for sound frequencies below about 100 Hz [82, 83]. More precise control is exerted at the hair cell level by the slow myosin-based adaptation that can adjust the mechanical input to the MET channels. However, fast adaptation with its small dynamic range and rapid kinetics may have a more subtle function. A possible clue to the role of fast adaptation comes from the turtle auditory papilla, where the adaptation time constant (sA) varies inversely with hair cell characteristic frequency (CF) [67, 78]. (The CF is the sound frequency to which a given hair cell is most sensitive, and in all vertebrate cochleae, it changes systematically with location to generate a tonotopic map.) How the variation in sA originates will be considered in delineation of the single channel properties, but it is partly attributable to a change in stereociliary influx of Ca2+. The corner frequency (1/2psA) of the high-pass filter contributed by fast adaptation is approximately two-thirds of the CF. This suggests that the adaptation may in some way contribute to hair cell frequency selectivity, a cell’s ability to discriminate different frequency components in a sound stimulus. This idea is reinforced by the observation that in saline containing physiological (50 lM) Ca2+ concentrations, fast adaptation may display under-damped resonance at frequencies in the turtle’s auditory range [78]. In the mammalian cochlea, where the CFs are much higher than those in the turtle, sA is correspondingly smaller, with a value of 100 ls or less [15]. In the turtle cochlea the channel’s activation time constant also increases with CF, [79] and the activation and fast adaptation time constants therefore confer on transduction a variable band-pass filter matched to the CF. This filter is unlikely to be the major source of hair cell frequency selectivity in the turtle, which instead stems from a sharply-tuned electrical resonance [84] produced by interplay of a voltage-dependent Ca2+ current and a Ca2+-activated K+ current [85]. The hair cell’s Ca2+-activated K+ channels vary in number and kinetics with location along the cochlea in order to generate a range of CFs [86, 87]. In contrast, the transduction filter may provide a mechanism for actively restricting the bandwidth to improve the signal-to-noise ratio of transduction within the frequency range encoded by the hair cell [88]. This must be done on a cycleby-cycle basis and to be optimal should therefore vary with CF. (An intriguing developmental question is how the center frequency of the filter supplied by the MET channels is matched to the frequency of electrical resonance endowed by the Ca2+-activated K+ channels.) It has been previously argued [89, 90] that using an active filter to narrow the stimulus bandwidth is a way of extending the physical detection limits of sensory

2.5 Single-channel Conductance

transduction in the face of intrinsic thermal noise. Another way of viewing this active filter is in terms of the mechanical stimulus. As a result of the gating compliance, closing the MET channels through adaptation elicits a mechanical-output, active motion of the hair bundle [8, 36, 91, 92]. Thus, a positive bundle displacement opens the MET channels, leading to a rise in intracellular Ca2+, which recloses the channels producing negative recoil. Such active motion may effectively amplify the cell’s response for external stimuli near threshold [93, 94]. The presence of fast adaptation in mammalian outer hair cells [15] implies that the mechanism is retained as part of the amplification in the mammalian cochlea [94, 95].

2.5

Single-channel Conductance

A distinguishing feature of the MET channel in assigning it to a particular protein family is single-channel conductance, but this property has been difficult to define. Conventional methods of recording single channels in membrane patches, cell-attached or detached, are unavailable probably because of the sub-micron diameter of the stereocilia and the need to preserve extracellular mechanical links to observe gating with physiological stimuli. Approaches employing analysis of current fluctuations [96] or inference from measurement of the macroscopic current and the number of active stereocilia [26] are at best indirect. The only systematic method devised has been to destroy or inactivate the majority of channels in a bundle and to monitor the one or two channels remaining in whole-cell recording mode [23, 97]. The number of channels was reduced by brief exposure to sub-micromolar Ca2+ concentrations, which most likely works by severing the majority of the tip links [6, 24]. The technique carries the assumption that the remaining channels were not significantly modified by the isolation procedure. Some justification for this assumption is provided by the finding that the residual channel behaved similarly to the macroscopic transduction current: it was gated by small displacements of the hair bundle, displayed fast activation and adaptation, and was blocked by suitable doses of dihydrostreptomycin [97]. Conspicuously, changing external Ca2+ had a dual action on channel properties resembling the effects on the macroscopic current (Fig. 2.7). Thus, reducing Ca2+ increased the channel amplitude by removal of block and also slowed the time course of channel activation and adaptation. The latter outcome was manifested in both the slowing of the ensemble average current and the increase in the channel’s mean open time. Two main results emerged from the single-channel experiments. Firstly, the unitary conductance was found to be surprisingly large, between 80 and 160 pS in 2.8 mM extracellular Ca2+; the range of values nearly doubled to 150–300 pS on removing the blocking action of Ca2+ by lowering its concentration to 50 lM (Fig. 2.7 [97]). Since the KI for channel block by Ca2+ is 1 mM (Fig. 2.5; [42]), measurements in 50 lM Ca2+ should reflect the maximum conductance of the channel in its unblocked state. Secondly, variation in channel conductance, in either high or low Ca2+, was correlated with the CF of the hair cell: cells with higher CFs possessed channels with bigger conduc-

45

46


Fig. 2.7 Single MET channels and their modulation by Ca2+. (A) Four single-channel responses recorded in a turtle hair cell for hair bundle deflections (Dx) of 150 nm in 2.8 mM extracellular Ca2+. Middle trace is the ensemble average of 140 channel responses. At the bottom is the open-time histogram of channel events in the absence of bundle stimulation. C and O denote the closed and open levels of the channel. (B) Two single-channel responses in the same cell in 0.05 mM extracellular Ca2+. Middle trace is the ensemble average of 250

channel responses, and at the bottom is the histogram of open-times without stimulation. Reducing extracellular Ca2+ had two effects: it increased the mean single-channel amplitude from 7 pA to 15 pA and it slowed channel kinetics. The latter effect was indicated by the slower time course of activation and adaptation of the ensemble average current and the increase in mean open time (sO) (holding potential, –80 mV). Unitary conductance in the Ca2+-unblocked state is 190 pS

tance. MET channels of 100 pS conductance have also been observed in hair cells of chicks [41] and mammals [34]. It will be important to determine whether this conductance varies with hair cell CF, especially in the mammalian cochlea. The change in channel conductance with CF in the turtle cochlea is unlikely to result from differences in Ca2+ homeostasis within the stereocilia because virtually identical conductance ranges were found in high and low extracellular Ca2+ [97]. Changes in MET channel conductance are most simply explained if hair cells in different regions of the cochlea contain channels with unique structure or sub-unit composition. Further support for this notion comes from the finding that MET channels in hair cells with higher CFs also differ functionally in having faster activation kinetics [56, 79]. The tonotopic distribution of MET channel properties, conductance and ki-

2.5 Single-channel Conductance

netics, is unexpected, but it parallels gradients in the properties of the Ca2+-activated K+ channel that underlie variation of the electrical resonant frequency in the turtle cochlea [87, 98]. The latter variation may stem from differential expression of alternatively spliced isoforms of the Ca2+-activated K+ channel a-subunit combined with a cochlear gradient in an accessory b-subunit [86, 99, 100].

2.5.1

Number of MET Channels Per Stereocilium

The single-channel conductance can be used to infer the number of channels per stereocilium if the maximum size of the macroscopic transduction current and the number of stereocilia in the hair bundle are known. If the MET channels are uniformly distributed across the bundle, each stereocilium possesses between one and two channels irrespective of hair cell CF [96]. Increases in the stereociliary complement [101] and single-channel conductance with CF together produce a severalfold change in the peak amplitude of the macroscopic MET current between turtle hair cells tuned to low and high frequency [67]. Inferring the number of channels per stereocilium from the total current assumes that all available channels are active during macroscopic recordings. The channels or their mechanical connections (for example, the tip links) may be damaged when the cochlea is isolated, which would lead to an underestimate in the total MET current and hence in the number of channels per stereocilium. Denk et al. [26] used two-photon imaging of stereociliary Ca2+ transients to assess how many of the stereocilia were actively contributing to a macroscopic transduction current. From those experiments they too concluded that each stereocilium possesses two channels. Based on the observation that bundle stimulation generated Ca2+ signals in both the shortest and tallest stereociliary rows, it was argued [26] that channels occur at both ends of the tip link. Such an arrangement implies a negative cooperativity between pairs of channels at the two ends of the tip link where opening of a channel at one end of the link relieves the force on the channel at the opposite end of the link. However, there is no experimental support for an interaction of this kind. Furthermore, tip-link destruction in sub-micromolar Ca2+ during the single-channel experiments should have left two channels for each intact tip link, but often one channel remained after exposure to low Ca2+ [97]. These results suggest that the MET channels occur at only one end of the tip link, probably at its insertion into the top of the stereocilium. The tonotopic gradient in the time constant of fast adaptation may also be explicable in terms of the single-channel properties. Adaptation was evident in the ensemble averages of single-channel activity (Fig. 2.7) from which a time constant could be extracted. Systematic variation in this time constant was seen in the ensemble average currents: cells with higher CFs adapting more rapidly [97], just as with the macroscopic current. Since the rate of fast adaptation is directly proportional to Ca2+ influx [42], the increase in adaptation rate with hair cell CF may be partly explained by an increase in channel size: doubling the channel conductance doubles the amount of Ca2+ entering and thus halves the adaptation time constant. Changes in other channel properties

47

48


may also contribute in varying the adaptation time constant. There is no evidence that the increase in channel conductance is accompanied by a concomitant augmentation of Ca2+ permeability, which would theoretically speed up adaptation [56]. However, spectral analysis of transducer current noise has suggested a difference in channel kinetics between high- and low-frequency hair cells [56]. This difference is endorsed by measurements of activation kinetics obtained by fitting the onset of the MET current in response to rapid bundle deflections [79]. Speeding up the activation and deactivation kinetics of the channel might produce additional acceleration of adaptation rate over that realizable by changes in single-channel conductance.

2.6

The MET Channel as a Member of the TRP Family

There are several distinctive features of the MET channel gleaned from measurements on intact hair cells that may aid in its molecular classification. These include a high selectivity for Ca2+ over other cations, a broad spectrum of blocking agents, large unitary conductance, and regulation by intracellular Ca2+. These attributes eliminate several channel contenders [102]. One is the epithelial Na+ channel (ENaC), which has subunits orthologous to the MEC-4 and MEC-10 proteins that form a mechanoreceptor channel in touch neurons of the nematode worm Caenorhabditis elegans [103, 104] (see Chapter 1). Like the hair cell MET channel, ENaC is blocked by amiloride. However, known forms of ENaC are Na+-selective channels with low Ca2+ permeability and small unitary conductance (13–40 pS [105]. Moreover, the properties of their block by amiloride are substantially different from the MET channel, with a 100-fold greater affinity for the drug (KI = 0.5 lM) and a Hill coefficient of 1.0 [106]. Nevertheless, a member of the ENac family may be a constituent of the transduction channel in vertebrate cutaneous mechanoreceptors [107]. The most likely channel candidate on the basis of present evidence belongs to the TRP (transient receptor potential) superfamily, some members of which possess channel properties resembling the hair cell MET channel [108]. The MEC channel subunits were obtained from genetic screens of mutants lacking a touch response in C. elegans. Such screens also generated a separate channel, CeOSM-9, from animals with defects in osmotic avoidance and nose touch [109]. CeOSM-9 was the first TRP channel implicated in mechanosensitivity, but two relatives were subsequently cloned from Drosophila melanogaster (see Chapter 8). One is DmNOMPC, the mutation of which largely abolishes receptor potentials in the touch-sensitive bristle organs [110] but only mildly affects hearing in Drosophila [111]. The other is DmNanchung (DmNAN), which is localized to the ciliary neurons in Johnston’s organ and whose mutation deafens the insects [112]. Both DmNOMPC and DmNAN have structures consistent with TRP channels with six transmembrane domains, a pore region between S5 and S6, intracellular N- and C-termini, and multiple ankyrin repeats at the N-terminus (29 in DmNOMPC and 5 in DmNAN). The two proteins have been assigned to different subclasses of the TRP superfamily: DmNAN to TRPV, similar to CeOSM-9, and DmNOMPC to a new sub-class, TRPN [113]. A vertebrate NOMPC with 62 % simi-

2.7 Conclusions

larity at the amino acid level to the Drosophila version has recently been identified in hair cells of the zebra fish, Danio rerio [10]. Removal of NOMPC gene function by injection of morpholino antisense oligonucleotides into larvae caused initial deafness and imbalance that later reversed as the morpholino was diluted by the endogenous transcript. There is as yet no evidence for the occurrence of NOMPC in hair cells of other vertebrates, including mammals.

2.6.1

Properties of TRPV Channels

Although no channel information exists for NOMPC, DmNAN, when expressed in cultured Chinese hamster ovary (CHO) cells, forms Ca2+-permeable channels that confer responsiveness to osmotic stress, consistent with a role in mechanotransduction [112]. More detailed information on channel attributes is available for other TRP members [108], which behave as nonspecific cation channels highly permeable to Ca2+ with sizable single-channel conductance. The precise properties depend on channel type but may be illustrated for the TRPV subfamily [114–123] to which DmNAN belongs. This family can be subdivided in terms of Ca2+ selectivity: for TRPV1, TRPV3, and TRPV4 PCa/PNa is 10:1, whereas for TRPV5 and TRPV6 it is 100:1. The less Ca2+-selective TRPV1, TRPV3, and TRPV4 channels exhibit larger single-channel conductance (110, 137, 90, or 300 pS) compared to TRPV5 and TRPV6 (78, 50 pS). The functional category to which DmNAN belongs is not yet known, but clearly there are TRPV members with permeation properties encompassing those of the hair cell MET channel. Although there have been no systematic studies of TRP channel permeability to organic cations, TRPV1, like the hair cell MET channel, can pass the large FM1-43 dye [49]. Within the TRPV family only TRPV4 and DmNAN have been recognized as being mechanosensitive. Nevertheless, other channels in the TRP superfamily are capable of mechanotransduction, including the yeast vacuolar channel, Yvc1p [124], the polycystic kidney disease gene product, PKD-2 [125], and a renal stretchinhibitable channel [126]. These too are Ca2+ permeable and have large (>100 pS) unit conductance. As a final mark of similarity, TRP channel gating can be modulated by divalent ions. For example, TRPV5 is blocked in a voltage-dependent manner by extracellular Ca2+ [123], Mg2+, and low concentrations of La3+ [122]. It is also desensitized by elevation of intracellular Ca2+ [127]. TRPV4 is modulated by extracellular and intracellular Ca2+ [128] and is blocked by Gd3+ [119].

2.7

Conclusions

Measurements in intact hair cells have provided considerable insight into the properties of the transduction channels, including their fast sub-millisecond gating and exquisite nanometer mechanical responsiveness, their cationic permeability with preference for Ca2+, and their large single-channel conductance. The permeability properties

49

50


are consistent with the MET channel being a member of the TRP superfamily, a hypothesis reinforced by the identification of TRP channels underlying mechanosensitivity in the hearing systems of Drosophila and zebra fish. Whether or not a TRP channel eventually proves to be the hair cell transduction in all vertebrate inner ears, it is likely that multiple isoforms or accessory proteins will be needed to explain the distribution of MET channel properties encountered in the turtle. The availability of single-channel measurements in intact hair cells provides a norm for comparison with future cloned channels to define subunit composition, as was the case for olfactory cyclic nucleotide-gated channels [129]. From recordings in intact hair cells there is also evidence that Ca2+ regulates the channel and its mechanical input to mediate multiple components of adaptation. Neither the mechanism nor the role of adaptation is entirely clear. Fast adaptation may directly alter MET channel gating, whereas slow adaptation may reset the mechanical stimulus via the action of one or more unconventional myosins. For auditory hair cells, fast adaptation varies with hair cell CF, which may be important for maximizing the signal-to-noise ratio of transduction in a frequency band around CF. By generating active hair bundle motion, it may also amplify the extrinsic mechanical stimulus [94]. Cloning the MET channel may lead to an understanding of how the channel interacts with Ca2+ and with other subcellular components, including myosins. It may in addition reveal the existence of multiple channel isoforms with unitary conductance or kinetics specialized for operation in different frequency ranges. If the MET channel is localized to the osmiophilic cap seen in electron micrographs of the tips of the stereocilia (Fig. 2.1), it could well be part of a larger protein complex analogous to the transduciosome proposed in Drosophila photoreceptors [130, 131]. Other proteins would include those anchoring the channel to the internal cytoskeleton and to the extracellular links, one or more of the different myosin isoforms (1C, 3, 6, 7A and 15 [132–134] localized to hair cells), and a plasma membrane CaATPase [135], which is crucial for Ca2+ extrusion from the bundle. All may be cemented together by the PDZ -domain protein INAD, which localizes TRP channels in Drosophila photoreceptors [131, 136, 137] (see Chapter 8). INAD was recently identified in mammalian hair cells [134]. There is a significant lack of knowledge about how the mechanical input (for example, increased tension in the tip links) is coupled to channel gating, which may involve one or more accessory proteins. The molecular identities of the proteins that comprise the “gating spring” are also currently unknown. However, the existence of mechanical linker proteins has been established for the mechanotransducer complex in C. elegans. Touch mutations have shown that proper mechanical gating of the DEG/ENac channel requires multiple intracellular and extracellular accessory proteins [138] (see Chapter 1). Application of molecular genetics to mammalian hearing may similarly identify not only the hair cell MET channel but also the set of accessory proteins essential for the correct targeting and operation of the channel [139, 140].

2.7 Conclusions

Acknowledgments

This work was supported by the National Institutes on Deafness and other Communicative Disorders Grant RO1 DC 01362. I would like to thank Carole Hackney for commenting on the manuscript; Tony Ricci, Helen Kennedy, and Andrew Crawford, who took part in the experiments in Figures 2.2 and 2.7; and Gaurang Patel for help with modeling in Fig. 2.6.

References 1 2

3

4

5

6

7

8

9

10

11

Tilney LG, Tilney MS Functional organization of the cytoskeleton. Hear Res 1986, 22, 55–77. Bagger-Sjoback D, Wersall J The sensory hairs and tectorial membrane of the basilar papilla in the lizard Calotes versicolor. J Neurocytol 1973, 2, 329–350. Little KF, Neugebauer D-C Interconnections between the stereovilli of the fish inner ear. II Systematic investigation of saccular hair bundles of Rutilus rutilus (Teleostei). Cell Tissue Res 1985, 284, 473–479. Pickles JO, Comis SD, Osborne MP Crosslinks between stereocilia in the guinea pig organ of Corti, and their possible relation to sensory transduction. Hear Res 1984, 15, 103–112. Furness DN, Hackney CM Cross-links between stereocilia in the guinea pig cochlea. Hear Res 1985, 18, 177–188. Assad JA, Shepherd GM, Corey DP Tip-link integrity and mechanical transduction in vertebrate hair cells. Neuron 1991, 7, 985–994. Flock A, Flock B, Murray E. Studies on the sensory hairs of receptor cells in the inner ear. Acta Otolaryngol 1977, 83, 85–91. Howard J, Hudspeth AJ Compliance of the hair bundle associated with gating of mechanoelectrical transduction channels in the bullfrog’s saccular hair cell. Neuron 1988 1, 189–199. Hudspeth AJ Extracellular current flow and the site of transduction by vertebrate hair cells. J Neurosci 1982, 2, 1–10. Sidi, S., Friedrich, R.W. and Nicolson, T. NompC TRP channel required for vertebrate sensory hair cell mechanotransduction. Science 2003, 301, 96–99. Sukharev SI, Blount P, Martinac B, Kung C. Mechanosensitive channels of Escherichia coli: the MscL gene, protein, and activities. Annu Rev Physiol. 1997, 59, 633–657.

12

13

14

15

16

17

18

19

20

21

22

Corey DP, Hudspeth AJ Kinetics of the receptor current in bullfrog saccular hair cells. J Neurosci 1983, 3, 962–976. Crawford AC, Evans MG, Fettiplace R Activation and adaptation of transducer currents in turtle hair cells. J Physiol 1989, 419, 405–434. Kros CJ, R€ usch A, Richardson GP Mechanoelectrical transducer current in hair cells of the cultured neonatal mouse cochlea. Proc R Soc Lond B 1992, 249, 185– 193. Kennedy HJ, Evans MG, Crawford AC, Fettiplace R Fast adaptation of mechanoelectrical transducer channels in mammalian cochlear hair cells. Nat Neurosci 2003, 6, 832–836. Vollrath MA, Eatock RA Time course and extent of mechanotransducer adaptation in mouse utricular hair cells: comparison with frog saccular hair cells. J Neurophysiol 2003, 90, 2676–2689. Robles L, Ruggero MA Mechanics of the mammalian cochlea. Physiol Rev 2001, 81, 1305–1352. Dallos P, Cheatham MA Production of cochlear potentials by inner and outer hair cells. J Acoust Soc Am 1976, 60, 510–512. Patuzzi RB, Yates GK, Johnstone BM. The origin of the low-frequency microphonic in the first cochlear turn of guinea-pig. Hear Res 1989, 39, 177–188. Pollak G, Henson OW, Novick A Cochlear microphonic audiograms in the pure tone bat, Chilonycteris parnelli parnelli. Science 1972, 176, 66–68. Hamill OP, Martinac B. Molecular basis of mechanotransduction in living cells. Physiol Rev 2001, 81, 685–740. Shotwell SL, Jacobs R, Hudspeth AJ. Directional sensitivity of individual vertebrate hair cells to controlled deflection of their hair bundles. Ann N Y Acad Sci 1981, 374, 1–10.

51

52

2 Transduction Channels in Hair Cells 23

24

25

26

27

28

29

30

31

32

33

34

Crawford AC, Evans MG, Fettiplace R The actions of calcium on the mechano-electrical transducer current of turtle hair cells. J Physiol 1991, 434, 369–398. Hackney, C.M & Furness, D.N. Hair cell ultrastructure and mechanotransduction: morphological effects of low extracellular calcium on stereociliary bundles in the turtle cochlea. In Active Hearing, Flock A, Ottoson D, Ulfendahl M (eds.), Pergamon, London, pp 103–111. 1995. Jaramillo F, Hudspeth AJ Localization of the hair cell’s transduction channels at the hair bundle’s top by iontophoretic application of a channel blocker. Neuron 1991, 7, 409–420. Denk W, Holt JR, Shepherd GM, Corey DP Calcium imaging of single stereocilia in hair cells: localization of transduction channels at both ends of tip links. Neuron 1995, 15, 1311–1321. Lumpkin EA, Hudspeth AJ Detection of Ca2+ entry through mechanosensitive channels localizes the site of mechanoelectrical transduction in hair cells. Proc Natl Acad Sci USA 1995, 2, 10297–10301. Hackney CM, Furness DN, Benos DJ, Woodley JF, Barratt J Putative immunolocalization of the mechanoelectrical transduction channels in mammalian cochlear hair cells. Proc R Soc Lond B 1992, 248, 215–221. Goodyear R, Richardson G A novel antigen sensitive to calcium chelation that is associated with the tip links and kinocilial links of sensory hair bundles. J Neurosci 2003, 23, 4878–4887. Kachar B, Parakkal M, Kurc M, Zhao Y, Gillespie PG High-resolution structure of haircell tip links. Proc Natl Acad Sci USA 2000, 97, 13336–13341. Markin VS, Hudspeth AJ Gating-spring models of mechanoelectrical transduction by hair cells of the internal ear. Annu Rev Biophys Biomol Struct 1995, 24, 59–83. Ricci AJ, Crawford AC, Fettiplace R Mechanisms of active hair bundle motion in auditory hair cells. J Neurosci 2002, 22, 44–52. Assad JA, Hacohen N, Corey DP Voltage dependence of adaptation and active bundle movement in bullfrog saccular hair cells. Proc Natl Acad Sci USA 1989, 86, 2918–2922. Ge´le´oc GS, Lennan GW, Richardson GP, Kros CJ A quantitative comparison of mechanoelectrical transduction in vestibular and auditory hair cells of neonatal mice. Proc R Soc Lond B 1997, 264, 611–621.

35

36

37

38

39

40

41

42

43

44

45

46

47

48

Howard J, Roberts WM, Hudspeth AJ Mechanoelectrical transduction by hair cells. Annu Rev Biophys Biophys Chem 1988, 17, 99–124. Crawford AC, Fettiplace R The mechanical properties of ciliary bundles of turtle cochlear hair cells. J Physiol 1985, 364, 359–379. Russell IJ, Kossl M, Richardson GP Nonlinear mechanical responses of mouse cochlear hair bundles. Proc R Soc Lond B 1992, 250, 217–27. van Netten SM, Kros CJ Gating energies and forces of the mammalian hair cell transducer channel and related hair bundle mechanics. Proc R Soc Lond B 2000, 267, 1915–1923. Hudspeth AJ Hair-bundle mechanics and a model for mechanoelectrical transduction by hair cells. In Sensory transduction, DP Corey, SD Roper (eds.), Rockefeller University Press, New York, pp 357–370. 1992 Corey DP, Hudspeth AJ Ionic basis of the receptor potential in a vertebrate hair cell. Nature 1979, 281, 675–677. Ohmori H Mechano-electrical transduction currents in isolated vestibular hair cells of the chick. J Physiol 1985, 359, 189–217. Ricci AJ, Fettiplace R Calcium permeation of the turtle hair cell mechanotransducer channel and its relation to the composition of endolymph. J Physiol 1998, 506, 159–173. Lumpkin EA, Marquis RE, Hudspeth AJ The selectivity of the hair cell’s mechanoelectricaltransduction channel promotes Ca2+ flux at low Ca2+ concentrations. Proc Natl Acad Sci USA 1997, 94, 10997–11002. Jorgensen F, Kroese AB. Ca selectivity of the transduction channels in the hair cells of the frog sacculus. Acta Physiol Scand 1995, 155, 363–376. Frings S, Seifert R, Godde M, Kaupp UB. Profoundly different calcium permeation and blockage determine the specific function of distinct cyclic nucleotide-gated channels. Neuron 1995, 15, 169–179. Bosher SK, Warren RL Very low calcium content of cochlear endolymph, an extracellular fluid. Nature 1978, 273, 377–378. Salt AN, Inamura N, Thalmann R, Vora A Calcium gradients in inner ear endolymph. Am J Otolaryngol 1989, 10, 371–375. Gale JE, Marcotti W, Kennedy HJ, Kros CJ, Richardson GP FM1-43 dye behaves as a permeant blocker of the hair-cell mechanotransducer channel. J Neurosci 2001, 21, 7013–7025.

2.7 Conclusions 49

50

51

52

53

54

55

56

57

58

59

60

61

Meyers JR, MacDonald RB, Duggan A, Lenzi D, Standaert DG, Corwin JT, Corey DP. Lighting up the senses: FM1-43 loading of sensory cells through nonselective ion channels. J Neurosci 2003, 23, 4054–4065. Kimitsuki T, Nakagawa T, Hisashi K, Komune S, Komiyama S Gadolinium blocks mechanoelectric transducer current in chick cochlear hair cells. Hear Res 1996, 101, 75–80. Ricci, A.J. Pharmacological clues to the nature of the mechanoelectric transducer channel. Assoc Res Otolaryngol Abstr 2003, 26, 841. Baumann M, Roth A The Ca++ permeability of the apical membrane in neuromast hair cells. J Comp Physiol A 1986, 158, 681–688. Kroese AB, Das A, Hudspeth AJ Blockage of the transduction channels of hair cells in the bullfrog’s sacculus by aminoglycoside antibiotics. Hear Res 1989, 37, 203–217. Kimitsuki T, Ohmori H Dihydrostreptomycin modifies adaptation and blocks the mechanoelectric transducer in chick cochlear hair cells. Brain Res 1993, 624, 143–150. R€ usch A, Kros CJ, Richardson GP Block by amiloride and its derivatives of mechanoelectrical transduction in outer hair cells of mouse cochlear cultures. J Physiol 1994, 474, 75–86. Ricci A Differences in mechanotransducer channel kinetics underlie tonotopic distribution of fast adaptation in auditory hair cells. J Neurophysiol 2002, 87, 1738–1748. Jorgensen F, Ohmori H. Amiloride blocks the mechano-electrical transduction channel of hair cells of the chick. J Physiol 1988, 403, 577–588. Glowatzki E, Ruppersberg JP, Zenner HP, R€ usch A Mechanically and ATP-induced currents of mouse outer hair cells are independent and differentially blocked by d-tubocurarine. Neuropharmacol 1997, 36, 1269–1275. Frings S, Lynch JW, Lindemann B Properties of cyclic nucleotide-gated channels mediating olfactory transduction. Activation, selectivity and blockage. J Gen Physiol 1992, 100, 45–67. Fodor AA, Gordon SE, Zagotta WN Mechanism of tetracaine block of cyclic nucleotidegated channels. J Gen Physiol 1997, 109, 3–14. Picco C, Menini A The permeability of the cGMP-activated channel to organic cations in retinal rods of the tiger salamander. J Physiol 1993, 460, 741–758.

62

63 64

65

66

67

68

69 70

71

72

73

74

75

Dwyer TM, Adams DJ, Hille B The permeability of the endplate channel to organic cations in frog muscle. J Gen Physiol 1980, 75, 469–492. Eatock RA Adaptation in hair cells. Annu Rev Neurosci 2000, 23, 285–314. Fettiplace R, Ricci AJ Adaptation in auditory hair cells. Curr Opin Neurobiol 2003, 13, 446–451. Assad JA, Corey DP An active motor model for adaptation by vertebrate hair cells. J Neurosci 1992, 12, 3291–3309. Shepherd GM, Corey DP. The extent of adaptation in bullfrog saccular hair cells. J Neurosci 1994, 14, 6217–6229. Ricci AJ, Fettiplace R The effects of calcium buffering and cyclic AMP on mechano-electrical transduction in turtle auditory hair cells. J Physiol 1997, 501, 111–124. Howard J, Hudspeth AJ Mechanical relaxation of the hair bundle mediates adaptation in mechanoelectrical transduction by the bullfrog’s saccular hair cell. Proc Natl Acad Sci USA 1987, 84, 3064–3068. Gillespie PG, Corey DP Myosin and adaptation by hair cells. Neuron 1997, 19, 955–958. Holt JR, Gillespie SK, Provance DW, Shah K, Shokat KM, Corey DP, Mercer JA, Gillespie PG. A chemical-genetic strategy implicates myosin-1c in adaptation by hair cells. Cell 2002, 108, 371–381. Gillespie PG, Gillespie SK, Mercer JA, Shah K, Shokat KM Engineering of the myosinibeta nucleotide-binding pocket to create selective sensitivity to N(6)-modified ADP analogs. J Biol Chem 1999, 274, 31373–31381. Kros CJ, Marcotti W, van Netten SM, Self TJ, Libby RT, Brown DM, Richardson GP, Steel KP Reduced climbing and increased slipping adaptation in cochlear hair cells of mice with Myo7a mutations. Nature Neurosci 2002, 5, 41–47. Wu YC, Ricci AJ, Fettiplace R Two components of transducer adaptation in auditory hair cells. J Neurophysiol 1999, 82, 2171–2181. Walker RG, Hudspeth AJ Calmodulin controls adaptation of mechanoelectrical transduction by hair cells of the bullfrog’s sacculus. Proc Natl Acad Sci USA 1996, 93, 2203–2207. Furness DN, Karkanevatos A, West B, Hackney CM An immunogold investigation of the distribution of calmodulin in the apex of the cochlear hair cells. Hear Res 2002, 173, 10–20.

53

54

2 Transduction Channels in Hair Cells 76

77

78

79

80

81

82

83

84

85

86 87

88

89

Cyr JL, Dumont RA, Gillespie PG: Myosin 1-c interacts with hair-cell receptors through its calmodulin- binding IQ domains. J Neurosci 2002, 22, 2487–2495. Ge´le´oc G, Corey DP Modulation of mechanoelectrical transduction by protein kinase A in utricular hair cells of neonatal mice. Assoc Res Otolaryngol Abstr 2001, 24, 242. Ricci AJ, Wu YC, Fettiplace R The endogenous calcium buffer and the time course of transducer adaptation in auditory hair cells. J Neurosci 1998, 18, 8261–8277. Fettiplace, R., Crawford, A.C. and Ricci, A.J. The effects of calcium on mechanotransducer channel kinetics in auditory hair cells. In Biophysics of the Cochlea: from molecule to model, AW Gummer (ed.), World Scientific, Singapore, pp. 65–72. 2003. Choe Y, Magnasco MO, Hudspeth AJ A model for amplification of hair-bundle motion by cyclical binding of Ca2+ to mechanoelectrical-transduction channels. Proc Natl Acad Sci USA 1998, 95, 15321–15326. Martin P, Bozovic D, Choe Y, Hudspeth AJ. Spontaneous oscillation by hair bundles of the bullfrog’s sacculus. J Neurosci 2003, 23, 4533–48. Franke R, Dancer A. Cochlear mechanisms at low frequencies in the guinea pig. Arch Otorhinolaryngol 1982, 234, 213–8. Cheatham MA, Dallos P. Inner hair cell response patterns: implications for low-frequency hearing. J Acoust Soc Am 2001, 110, 2034–44. Crawford AC, Fettiplace R An electrical tuning mechanism in turtle cochlear hair cells. J Physiol 1981, 312, 377–412. Art JJ, Fettiplace R Variation of membrane properties in hair cells isolated from the turtle cochlea. J Physiol 1987, 385, 207–42. Fettiplace R, Fuchs PA Mechanisms of hair cell tuning. Ann Rev Physiol 1999, 61, 809–34. Wu YC, Art JJ, Goodman MB, Fettiplace R. A kinetic description of the calcium-activated potassium channel and its application to electrical tuning of hair cells. Prog Biophys Mol Biol 1995, 63, 131–58. Dinklo T., van Netten, M., Marcotti, W. & Kros, C.J. Signal processing by transducer channels in mammalian outer hair cells. In Biophysics of the Cochlea: from molecule to model, AW Gummer (ed.), World Scientific, Singapore, pp. 73–79. 2003. Bialek W Physical limits to sensation and perception. Ann Rev Biophys Biophys Chem 1987, 16, 455–478.

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

Block S.M. Biophysical principles of sensory transduction. In Sensory transduction, DP Corey, SD Roper (eds.) Rockefeller University Press, New York., pp 1–17, 1992. Benser ME, Marquis RE, Hudspeth AJ Rapid, active hair bundle movements in hair cells from the bullfrog’s sacculus. J Neurosci 1996, 16, 5629–5643. Ricci AJ, Crawford AC, Fettiplace R Active hair bundle motion linked to fast transducer adaptation in auditory hair cells. J Neurosci 2000, 20, 7131–7142. Martin P, Hudspeth AJ Active hair-bundle movements can amplify a hair cell’s response to oscillatory mechanical stimuli. Proc Natl Acad Sci USA 1999, 96, 14306–14311. Fettiplace R, Ricci AJ, Hackney CM Clues to the cochlear amplifier from the turtle ear. Trends Neurosci 2001, 24, 169–175. Hudspeth A Mechanical amplification of stimuli by hair cells. Curr Opin Neurobiol 1997, 7, 480–486. Holton T, Hudspeth AJ The transduction channel of hair cells from the bull-frog characterized by noise analysis. J Physiol 1986, 375, 195–227. Ricci, A.J., Crawford, A.C. & R. Fettiplace, R. Tonotopic variation in the conductance of the hair cell mechanotransducer channel. Neuron 2003, (in press). Art JJ, Wu YC, Fettiplace R. The calcium-activated potassium channels of turtle hair cells. J Gen Physiol 1995, 105, 49–72. Jones EM, Gray-Keller M, Fettiplace R. The role of Ca2+-activated K+ channel spliced variants in the tonotopic organization of the turtle cochlea. J Physiol 1999, 518, 653–65. Ramanathan K, Michael TH, Jiang GJ, Hiel H, Fuchs PA. A molecular mechanism for electrical tuning of cochlear hair cells. Science 1999, 283, 215–7. Hackney CM, Fettiplace R, Furness DN. The functional morphology of stereociliary bundles on turtle cochlear hair cells. Hear Res 1993, 69, 163–75. Strassmaier M, Gillespie PG The hair cell’s transduction channel. Curr Opin Neurobiol 2002, 12, 380–386. HuangM,ChalfieMGeneinteractionsaffecting mechanosensory transduction in Caenorhabditis elegans. Nature 1994, 367, 467–470. Goodman MB, Ernstrom GG, Chelur DS, O’Hagan R, Yao CA, Chalfie M MEC-2 regulates C. elegans DEG/ENaC channels needed for mechanosensation. Nature 2002, 415, 1039–1042.

2.7 Conclusions 105 Ismailov II, Awayda MS, Berdiev BK, Bubien

106

107

108

109

110

111

112

113 114

115

116

117

JK, Lucas JE, Fuller CM, Benos DJ Triplebarrel organization of ENaC, a cloned epithelial Na+ channel. J Biol Chem 1996, 271, 807–16. Sariban-Sohraby S, Benos DJ The amiloridesensitive sodium channel. Am J Physiol 1986, 250, C175–190. Price MP, McIlwrath SL, Xie J, Cheng C, Qiao J, Tarr DE, Sluka KA, Brennan TJ, Lewin GR, Welsh MJ The DRASIC cation channel contributes to the detection of cutaneous touch and acid stimuli in mice. Neuron 2001, 32, 1071–83. Minke B, Cook B TRP channel proteins and signal transduction. Physiol Rev 2002, 82, 429–72. Colbert HA, Smith TL, Bargmann CI OSM-9, a novel protein with structural similarity to channels, is required for olfaction, mechanosensation, and olfactory adaptation in Caenorhabditis elegans. J Neurosci 1997, 17, 8259–69. Walker RG, Willingham AT, Zuker, CS A Drosophila mechanosensory transduction channel. Science 2000, 87, 2229–2234. Eberl DF, Hardy RW, Kernan MJ. Genetically similar transduction mechanisms for touch and hearing in Drosophila. J Neurosci 2000, 20, 5981–8. Kim J, Chung YD, Park D-Y, Choi S, Shin, DW, Soh H, Lee HW., Son W, Yim J, Park CS, Kernan MJ, Kim C A TRPV family ion channel required for hearing in Drosophila. Nature 2003, 424, 81–84. Corey DP New TRP channels in hearing and mechanosensation. Neuron 2003, 39, 585–8. Caterina MJ, Schumacher MA, Tominaga M, Rosen TA, Levine JD, Julius D The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature 1997, 389, 816–24. Premkumar LS, Ahern GP. Induction of vanilloid receptor channel activity by protein kinase C. Nature 2000, 408, 985–90. Mohapatra DP, Wang SY, Wang GK, Nau C. A tyrosine residue in TM6 of the vanilloid receptor TRPV1 involved in desensitization and calcium permeability of capsaicin-activated currents. Mol Cell Neurosci 2003, 23, 314–324. Xu H, Ramsey IS, Kotecha SA, Moran MM, Chong JA, Lawson D, Ge P, Lilly J, SilosSantiago I, Xie Y, DiStefano PS, Curtis R, Clapham DE. TRPV3 is a calcium-permeable temperature-sensitive cation channel. Nature 2002, 418, 181–186.

118 Strotmann R, Harteneck C, Nunnenmacher

119

120

121

122

123

124

125

126

127

128

K, Schultz G, Plant TD OTRPC4, a nonselective cation channel that confers sensitivity to extracellular osmolarity. Nature Cell Biol 2000, 2, 695–702. Liedtke W, Choe Y, Marti-Renom MA, Bell AM, Denis CS, Sali A, Hudspeth AJ, Friedman JM, Heller S Vanilloid receptor-related osmotically activated channel (VR-OAC), a candidate vertebrate osmoreceptor. Cell 2000, 103, 525–535. Voets T, Prenen J, Vriens J, Watanabe H, Janssens A, Wissenbach U, Bodding M, Droogmans G, Nilius B. Molecular determinants of permeation through the cation channel TRPV4. J Biol Chem 2002, 277, 33704–33710. Vennekens R, Hoenderop JG, Prenen J, Stuiver M, Willems PH, Droogmans G, Nilius B, Bindels RJ. Permeation and gating properties of the novel epithelial Ca2+ channel. J Biol Chem 2000, 275, 3963–3969. Nilius B, Vennekens R, Prenen J, Hoenderop JG, Bindels RJ, Droogmans G Whole-cell and single channel monovalent cation currents through the novel rabbit epithelial Ca2+ channel ECaC. J Physiol 2000, 527, 239–248. Yue L, Peng JB, Hediger MA, Clapham DE CaT1 manifests the pore properties of the calcium-release-activated calcium channel. Nature 2001, 410, 705–709. Zhou XL, Batiza AF, Loukin SH, Palmer CP, Kung C, Saimi Y The transient receptor potential channel on the yeast vacuole is mechanosensitive. Proc Natl Acad Sci USA 2003, 100, 7105–7110. Nauli SM, Alenghat FJ, Luo Y, Williams E, Vassilev P, Li X, Elia AE, Lu W, Brown EM, Quinn SJ, Ingber DE, Zhou J Polycystins 1 and 2 mediate mechanosensation in the primary cilium of kidney cells. Nat Genet 2003, 33, 129–137. Suzuki M, Sato J, Kutsuwada K, Ooki G, Imai M Cloning of a stretch-inhibitable nonselective cation channel. J Biol Chem 1999, 274, 6330–6335. den Dekker E, Hoenderop JG, Nilius B, Bindels RJ The epithelial calcium channels, TRPV5 & TRPV6: from identification towards regulation. Cell Calcium 2003, 33, 497–507. Watanabe H, Vriens J, Janssens A, Wondergem R, Droogmans G, Nilius B Modulation of TRPV4 gating by intra- and extracellular Ca2+. Cell Calcium 2003, 33, 489–495.

55

56

2 Transduction Channels in Hair Cells 129 B€ onigk W, Bradley J, M€ uller F, Sesti F,

130

131

132

133

134

Boekhoff I, Ronnett V, Kaupp UB and Frings S The native rat olfactory nucleotide-gated channel is composed of three distinct subunits. J Neurosci 1999, 19, 5332–5347. Scott K, Zuker CS Assembly of the Drosophila phototransduction cascade into a signalling complex shapes elementary responses. Nature 1998, 395, 805–8. Huber A Scaffolding proteins organize multimolecular protein complexes for sensory signal transduction. Eur J Neurosci 2001, 14, 769–776. Hasson T, Gillespie PG, Garcia JA, MacDonald RB, Zhao Y, Yee AG, Mooseker MS, Corey DP: Unconventional myosins in innerear sensory epithelia. J Cell Biol 1997, 137, 1287–1307. Belyantseva IA, Azevedo RB, Fridell RA, Friedman TB, Kachar B Assoc Res Otoloaryngol Abstr 2002, 25, 157. Walsh T, Walsh V, Vreugde S, Hertzano R, Shahin H, Hsika S, Lee MK, Kanaan M, King M-C, Avraham K From flies’ eyes to our ear: mutations in a human class III myosin causes nonsyndromic hearing loss DFNB30. Proc Natl Acad Sci USA 2002, 99, 7518–7523.

135 Dumont RA, Lins U, Filoteo AG, Penniston

136

137

138

139

140

JT, Kachar B, Gillespie PG Plasma membrane Ca2+-ATPase isoform 2a is the PMCA of hair bundles. J Neurosci 2001, 21, 5066–5078. Shieh BH, Zhu MY. Regulation of the TRP Ca2+ channel by INAD in Drosophila photoreceptors. Neuron 1996 16, 991–8. Chevesich J, Kreuz AJ, Montell C Requirement for the PDZ domain protein, INAD, for localization of the TRP store-operated channel to a signaling complex. Neuron 1997 18, 95–105. Tavernarakis N, Driscoll M Molecular modeling of mechanotransduction in the nematode Caenorhabditis elegans. Ann Rev Physiol 1997, 59, 659–89. Steel KP, Kros CJ A genetic approach to understanding auditory function. Nat Genet 2001, 27, 143–9. Boda B, El-Amraoui A, Bahloul A, Goodyear R, Daviet L, Blanchard S, Perfettini I, Fath KR, Shorte S, Reiners J, et al: Myosin VIIa, harmonin and cadherin 23, three Usher I gene products that cooperate to shape the sensory hair bundle. EMBO J 2002, 21, 6689–6699.

57

3

Acid-sensing Ion Channels Kenneth A. Cushman and Edwin W. McCleskey

3.1

Introduction

Krishtal and Pidoplichko discovered acid-sensing ion channels in 1980 by using voltage clamp recordings from rat sensory neurons [1]. This and a series of subsequent papers [2–4] from the same group established the essential features of these molecules: (1) the channels open when pH drops below 7.0, (2) they pass Na+ over K+ about as well as voltage-gated Na+ channels do, (3) amiloride blocks them at rather high concentrations (tens of micromolars), and (4) distinct subtypes of the channel evident

Fig. 3.1 Different desensitization kinetics of acid-evoked currents observed in trigeminal ganglia cells. Currents recorded in response to pH stimulus of 6.2. From [2] Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

58

3 Acid-sensing Ion Channels

in different sensory neurons differ in their pH sensitivity and rates of activation and desensitization (Fig. 3.1). Krishtal argued that these acid-evoked currents were caused by unique ion channels, whereas another group thought they might be acid-modified Ca++ channels [5]. The debate was loud but the audience was small, and the subject fell from view until 1997. That was when Lazdunski and colleagues cloned an acid-sensing, amiloride-sensitive, Na+-selective channel and coined the acronym ASIC (acid-sensing ion channel) [6]. The work showed that ASICs were a subfamily of mammalian epithelial Na+ channels (ENaCs) and their relatives in C. elegans, degenerins (see Chapters 1 and 4), channels that are all Na+ selective and amiloride sensitive but that differ in their gating mechanisms. From the first, ASICs were proposed to be transducers for acid-evoked pain [3]. They also have been implicated in mechanosensation, taste transduction, and learning and memory. A variety of reviews discuss their various properties and proposed functions [7–9].

3.2

ASICs and the DEG/ENaC Superfamily

ASICs belong to the superfamily of epithelial sodium channels (ENaCs) and degenerins (DEGs) (Fig. 3.2). All channels in the DEG/ENaC family are sodium selective and voltage insensitive and are blocked by amiloride (albeit at very different concentrations). Some are constitutively open, protons gate some, a peptide (FMRFamide) gates one [10], and some may be mechanosensitive. The ASICs are defined by homology to each other rather than by their acid sensitivity. ASIC1 and ASIC3 are readily opened by small pH changes, but ASIC2 needs extreme pH (5.0) to open and ASIC4 may not be acid sensitive at all. The variability in acid sensitivity suggests that ASICs may prove to have functions that are unrelated to their name. ENaCs are among the most critical ion channels to mammalian biology. Expressed in epithelial cells, they control sodium reabsorption for fluid homeostasis. For example, by controlling sodium flux across kidney epithelia, they control the amount of blood in the body [11]. Amiloride can block ENaCs at tens of nanomolars, making it clinically useful for control of blood pressure. Concentrations at least 1000-fold greater are needed to block ASICs. The dominant “degenerin” mutations cause the neurons that express this protein in C. elegans to swell and lyse [12–14], due to a constitutive activity of the ion channel [15]. Most of the DEGs are expressed in mechanosensory cells, leading to “mechanosensory abnormal” (MEC) or uncoordinated (UNC) phenotypes in the mutants and raising the possibility that these channels are mechanosensors (see Chapter 1). The first ASIC cloned was ASIC2a, originally called MDEG (mammalian degenerin) and BNC1 (Brain Na+ Channel 1) [16, 17]. It has homology to DEGs (20–29 % identity) and is constitutively active with a degenerin mutation. Neither group that cloned it connected it with Krishtal’s acid-gated sodium channels; indeed, ASIC2a requires quite extreme acid (pH 5) to open.

3.2 ASICs and the DEG/ENaC Superfamily

Fig. 3.2 Phylogenetic tree of DEG/ENaC family members. Sequences aligned using ClustalW

ASIC1a was cloned by two groups: Corey’s, which named it BNaC2 (Brain Na+ Channel 2) [18], and Lazdunski’s, which realized its acid sensitivity and named it ASIC [6]. ASIC1a in rats is a 526-amino-acid protein that is expressed throughout the peripheral and central nervous system. An extracellular pH drop below 6.9 activates a transient amiloride-sensitive sodium current, with a half maximal activating pH (pH0.5) near 6.5 (Tab. 3.1) [6, 19, 20]. ASIC1a has a significant Ca++ permeability and is blocked by higher levels of Ca++. This channel has kinetics, expression pattern, pharmacology, and selectivity similar to some of the currents described by Krishtal and Pidoplichko. The Lazdunski group then cloned and characterized rat ASIC3 [21]. It was originally called DRASIC (dorsal root ASIC) because its mRNA was found only in dorsal root ganglion sensory neurons in rats. ASIC3 is more sensitive to acid than is ASIC1a, with a pH0.5 of 6.7 and a steeper activation curve [22]. ASIC3 desensitizes much faster than ASIC1a. Calcium ions carry current through ASIC3 in the absence of Na+ [22], but there is no reported Ca++ permeability in the presence of Na+. Lazdunski’s group then showed that ASIC2a is also a proton-gated ion channel [23]. ASIC2a is much less sensitive to protons than are either ASIC3 or ASIC1a, with a pH0.5 of 4.35 [24]. They also cloned a splice variant of ASIC2 with a different amino terminus, ASIC2b (MDEG2). ASIC2b is inactive as a homomer but modulates the kinetics, selectivity, and sensitivity of both ASIC2a and ASIC3 when co-expressed

59

60

3 Acid-sensing Ion Channels Tab. 3.1.

Summary of properties of ASIC homomers and heteromers

Channel

Other Names

PNa/PK

PNa/PCa

pH0.5

Sustained Location Current?

Modulating Peptides

ASIC1a

ASIC, ASICa, BNaC2 ASICb BNC1, BNaC1, MDEG1 MDEG2 DRASIC

7.8-13

2.5-18.5

5.3-6.6

N

C/S

FMRFRFRP1/2

14 10

high 20

5.8-5.9 4.35

N

S C/S

Inactive 13,5

– high

– 6.5-6.7

– Y

C/S S

– 4.8

–

C

ASIC2a/2b ASIC2a/3

Inactive – 7.2 36 Doesn’t form Doesn’t form * >1 for bothhigh

3.9 4.3-6.5

Y Y

ASIC3/2b

*

6.5-6.7

Y

ASIC1b ASIC2a ASIC2b ASIC3

ASIC4 SPASIC ASIC1a/2a ASIC1a/2b ASIC1a/3

NPFFFMRFNPSFRFRP1/2

NPFFFMRF

* = Sodium-selective initial current, nonselective sustained current S = Sensory neurons C = Central nervous system

[23]. ASIC2a and ASIC2b are both expressed throughout the central and peripheral nervous systems [20, 23]. Other labs have since cloned further genes and splice variants in this family. ASIC1b (ASIC-b) and ASIC-b2 are splice variants of ASIC1 [25–27]. ASIC1b lacks the Ca++ permeability of ASIC1a, has a decreased sensitivity to protons (pH0.5=5.84) [19], and is restricted to sensory neurons. ASIC-b2 does not form functional homomers but does associate with ASIC1b to decrease its affinity for protons. ASIC4 was the last member of the ASIC family to be cloned [28, 29]. It is expressed in the CNS and has not been shown to form a functional ion channel or to modulate the acid sensitivity of other ASICs. The human ASICs are different from the rat ASICs described above. They show different expression patterns, isoforms, and pharmacology. Human ASIC3, for example, has three isoforms and is in the CNS, PNS, and internal organs such as lung and kidney [30].

3.4 Assembly Into Channels

3.3

Amino Acid Structure

ASICs have two predicted transmembrane domains in their structure, with most of the protein being in the extracellular loop. There are two conserved cysteine-rich domains (CRDs) in the extracellular loop that are common to all DEG/ENaC family members. These CRDs in ENaCs play a role in proper expression and activation of the channel [31]. The two transmembrane domains are presumed to be analogous to the pore-forming alpha helices of other ion channels [32, 33]. The second (C-terminal) transmembrane domain and a short region (the “P-loop”) just upstream from it are considered the primary structures contributing to pores. This hypothesis has not been investigated in ASICs, but systematic mutation of residues in this domain of ENaCs alters Na+ selectivity [34, 35], in agreement with expectation. However, mutation of the residues just upstream of transmembrane domain 1 of ASIC2 also clearly modifies Na+ selectivity, arguing that this region also contributes to the pore [36].

3.4

Assembly Into Channels

There is no doubt that multiple ASIC proteins must assemble in order to form a functional channel, but the stoichiometry is debated. One group provides data arguing for a nonameric structure, whereas another argues for a tetrameric structure [37, 38]. A similar conflict exists in the ENaC field [39, 40]. It seems likely that ASICs assemble with the same stoichiometry as ENaCs given their conserved topology. Unlike ENaCs, whose functional channels must contain different channel subtypes [41], ASICs can form homomeric ion channels. ASIC1a, -1b, -2a, and -3 can form functional homomeric channels. ASIC2b can form heteromers with other ASICs to modify their activity but is inactive by itself. ASIC4 has been shown neither to form a functional ion channel nor to modulate other subunits. ASIC1a and ASIC2a share common patterns of expression in the central nervous system, suggesting they may colocalize to form heteromers. It has been shown in oocytes that the two subunits can form a heteromeric channel [42]. The heteromer is less selective than either homomer and has a decreased permeability to Ca++ and an intermediate proton sensitivity. This heteromer has also been observed in cultured hippocampal neurons [43]. ASIC2b is electrically silent as a homomer but can change the properties of ASIC2a and ASIC3 when co-expressed. ASIC2a loses sensitivity to protons and creates a sustained nonselective current when ASIC2b is present [23, 44]. Co-expression of ASIC2b with ASIC3 does not change the pH sensitivity of ASIC3; however, the normal sodium-selective sustained current seen with ASIC3 at low pH becomes nonselective [23]. ASIC2b has no significant effect when co-expressed with ASIC1a. ASIC2a and ASIC3 are co-expressed in many dorsal root ganglion (DRG) neurons. When expressed together in oocytes, they form a functional heteromeric channel with

61

62


novel properties [45]. There is a large increase in sustained current that is sodium selective, like the sustained current seen in ASIC3 homomers. The ASIC currents seen in mouse DRG are different from those evoked from cloned channels, suggesting the presence of heteromers in vivo. Benson reproduced the currents evoked from medium to large DRG cells by expressing three subunits together: ASIC1a, -2a, and -3 [46]. It is not clear whether the three subunits all combine to form one channel or whether native currents are due to a mix of heteromers composed of two subunits. In individual knockouts of ASIC1, ASIC2a, or ASIC3, DRG ASIC currents matched heteromeric currents of the remaining two subunits. ASIC3 seems necessary for the fast desensitization kinetics seen in DRG, since the ASIC3 knockouts had slow kinetics. ASIC1 in mouse seems significantly different from the rat clone. Rat ASIC1 is substantially less sensitive to pH than rat ASIC3, whereas the two mouse clones both activate at and just below pH 7.

3.5

Pharmacology

Like all members of the DEG/ENaC family, ASICs are blocked by amiloride. The amiloride block is voltage dependent [47], with an IC50 ranging from 10–100 lM depending on the subunit composition of the channel. ENaCs have a much higher affinity for amiloride, typically on the order of 100 nM [48]. At high enough concentrations, amiloride can block L- and T-type Ca++ channels [49, 50], Na+/Ca++ exchangers, Na+/H+ exchangers, Na+ pumps, Ca++ pumps [51], and other channels. The low affinity of amiloride for ASICs, coupled with its promiscuity as a blocker, makes it a weak pharmacological tool. The mechanosensing ion-channel blocker Gd3+ also blocks ASIC2a/3 heteromers [45], but once again, this is not a very selective blocker. Residues contributing to amiloride binding are well established in ENaCs; amiloride is thought to bind to two sites, one in the extracellular domain and one near the pore [48, 52, 53]. Insight into the reason that ASICs are so much less sensitive to amiloride than are ENaCs arises from a study that mutated the residues in ASICs that are analogous to the ENaC amiloride-binding site near the pore. The mutation eliminated block and unmasked a second action of amiloride: enhancement of the ASIC current [47]. The interpretation is that ASICs also have two amiloride-binding sites, one that activates and one, like in ENaCs, that blocks. The only known ASIC-specific blocker is Psalmotoxin 1 (PcTx1), which comes from the venom of a South American tarantula [54, 55]. PcTx1 is specific for ASIC1a homomers and blocks with an IC50 of 0.9 nM. This potent toxin has proven useful to distinguish ASIC1a homomers from heteromers in the CNS [43]. The divalent ion Zn++ potentiates current in ASIC2a-containing channels [56]. Zn++ does not affect ASIC3 or ASIC1a homomers, but it does potentiate heteromers formed with ASIC2a. The potentiation is greater at higher pH values, with a greater than sevenfold enhancement at pH 6.0 for ASIC2a and with an EC50 of 120 lM. The presence of Zn++ (300 lM) caused a leftward shift of the activation curve of ASIC1a/2a

3.6 Gating

heteromers (pH0.5 from 5.5 to 6.0). Two histidine residues are essential for this potentiation; when either one is mutated to alanine, the effect disappears. ASIC currents are very sensitive to Ca++ [2]. Ca++ decreases both single-channel conductance and open probability [57, 58]. Dropping extracellular Ca++ increases the current through ASICs and shifts its activation curve to more basic pHs [59]. The steadystate inactivation curves of ASICs are also shifted by Ca++, with higher levels of Ca++ preventing inactivation [19]. Ca++ has a central role in gating the channel, as will be discussed further. FMRFamide (Phe-Met-Arg-Phe-amide) gates one DEG/ENaC channel called FaNaCh, and ASICs have strong homology to it. This led to the question of whether FMRFamide or other peptides can modulate ASICs. Mammals do not have FMRFamide, but they have related peptides such as neuropeptide FF (NPFF). NPFF is involved in pain modulation via both opioid-dependent and -independent pathways [60]. Although the peptides do not gate ASICs, they do slow the rate of desensitization and introduce a sustained current [61]. Interestingly, the peptides need to be applied prior to proton activation to have an effect. NPFF heavily modulates ASIC3 and ASIC2a/3 heteromers in oocytes but has no effect on ASIC1a or -2a (Tab. 3.1). FMRFamide modulates all the above channels except homomeric ASIC2a. Interestingly, neither NPFF nor FMRFamide modulates human ASIC3 [62]. NPSF, another mammalian RFamide peptide, also potentiates ASIC3 and DRG acid-evoked currents [63]. Experiments with other RFamide-related peptides show slowed desensitization and increased peak amplitude in DRG and heterologously expressed ASIC1 and ASIC3 [64]. The ability of RFamide peptides to induce sustained currents in some ASICs raises the idea that they can increase the sensitivity of ASICs in sustained pain states.

3.6

Gating

Our lab has recently proposed an idea for ASIC gating (Fig. 3.3) that, at this early stage, should be considered controversial. ASIC gating is highly sensitive to calcium concentration in the extracellular medium [2, 19, 57, 58, 65]. This raises the possibility that the proton-binding site is a titratable calcium-binding site, more or less like calcium chelators such as EGTA. We tested this hypothesis on ASIC3 and, consistent with it, found (1) that Ca++ and H+ compete at the gating site and (2) that eliminating extracellular Ca++ opens the channel without any change in extracellular pH [65]. Like a Ca++ chelator, the affinity of the channel for Ca++ diminishes as pH drops. It has an apparent affinity of around 150 nM at pH 9, of 10 lM at pH 7.4, and of 100 lM at pH 7.0. Thus, protons essentially catalyze release of Ca++ from this binding site, and we suggest that this release is the event that opens the channel. At physiological Ca++ (mM) concentrations, negligibly few channels will be free of Ca++ at pH 7.4, whereas nearly 10 % will be free at pH 7.0, more or less the fraction of channels open at this pH. Multiple protons must bind for effective Ca++ release, conveniently explaining the very steep proton activation curve of ASIC3.

63

64

3 Acid-sensing Ion Channels Fig. 3.3 Schematic showing proposed gating mechanism for ASIC3. At rest, ASIC3 has a high affinity for Ca++, which blocks the channel. When H+ binds, it decreases the affinity for Ca++. When Ca++ is released, the channel is relieved from block and opens. From [65]

A mathematical model that fit all our gating data also quantitatively fit Ca++ block data on single channels. This led us to suggest that the Ca++ ion involved in gating simply blocks the pore – in other words, that the channel does not open due to a proton-induced conformation change but, rather, due to a relief of Ca++ block. In contrast, desensitization clearly appears to be a slow, H+-driven conformation change. The model seems appealing and explains a variety of data on ASIC3, but it needs to be critically examined on other ASICs. At present, one study in the literature seems supportive. Askwith et al. found that decreasing temperature slows ASIC desensitization with little effect on activation rate [66]. Because conformational changes are affected by temperature far more than channel block is, this supports the idea that channel activation is not a conformation change. Other studies have investigated gating mechanisms of ASICs using mutations or chemical reagents to reveal important residues. A study on ASIC2a found that when Gly430 is mutated to a bulkier amino acid, the proton sensitivity can increase more than two orders of magnitude [24]. Besides an increased proton sensitivity, the mutant channel also fails to desensitize. Gly430 is the same residue that causes the degenerin phenotype when mutated [16]. This residue in ASIC2a is also implicated in gating due to its MTS reactivity in open but not closed states [67]. It is accessible only when the channel is open, suggesting a conformational change exposing the residue. Clearly, this residue is very important for channel function, but its role in gating the channel is uncertain. ENaCs are not gated by ligands or voltage, but are constitutively active. They are regulated either through controlling surface expression or by modifying their open probability. ENaCs are also blocked by Ca++ [68], which can be removed with mechanical forces to open the channel [69].

3.7 Proposed Sensory Functions

3.7

Proposed Sensory Functions

The ability to sense pH makes ASICs candidates for transducing acid-evoked pain and taste sensations. Their relation to degenerins makes them appealing as possible mechanosensing ion channels. Expression patterns and channel properties may be consistent with these ideas, but definitive tests in whole animals are lacking. 3.7.1

Pain/Nociception

Krishtal and Pidoplichko found that 74 % of DRG neurons expressing an acid-sensitive current were smaller than 26 lm [3], demonstrating expression in neurons that make Ad and C axons that carry the bulk of nociceptive information [70]. Immunocytochemistry shows that ASIC1 co-expresses with vanilloid receptors, Substance P, and CGRP, all markers for nociceptors [71]. The distribution is consistent with ASICs playing a role in pain but does not mean this need be their only function. Experiments on humans have shown that amiloride diminishes the pain associated with injection of a pH 6.0 solution [72]. The VR1 antagonist capsazepine has no effect on the intensity of pain perceived at pH 6.0 but does reduce it at pH 5.0, although to a lesser extent than amiloride. This provides evidence for the role of ASICs in transducing pain from acidic stimuli. It suggests that VR1 plays little or no part in sensing acid within the physiological range. Inflammation causes transcript levels of ASICs to increase [73] through transcriptional control that involves serotoninergic and nerve growth factor pathways [74]. Nonsteroid anti-inflammatory drugs (NSAIDs) such as aspirin and ibuprofen suppress this increase. In addition to suppressing the increase in transcript levels, NSAIDs also directly inhibit the currents produced by ASICs in both DRG and ASIC3 or ASIC1a expressing COS cells. The sensitivity of ASIC expression to persistent pain conditions is further circumstantial evidence for a role in pain. Benson et al. fluorescently tagged sensory neurons that innervate heart muscle and found that they all expressed ASIC3-like current at exceedingly high levels [22, 75]. As a sensory system, the heart is unusual in that the only conscious sensation that arises from it is pain (angina) and the only trigger for this is ischemia, when the heart receives insufficient oxygen for its metabolic demand. Thus, the expression is consistent with ASIC3 being a sensor for ischemic pain by detecting lactic acid. Moreover, ASIC3 is clearly sensitive enough to detect the fairly small (1/2 pH unit) changes in extracellular acidity that occurs in ischemic muscle. ASIC3 also responds better to lactic acid than to other forms of acid [59], making it better for detecting the lactic acidosis that accompanies ischemia than the carbonic acidosis typical of most metabolic acidoses. Together these observations make a circumstantial case that ASIC3 is a sensor for angina and related ischemic pain. The idea is difficult to test in transgenic mice because there is no rodent model of ischemic pain. The clearest demonstration for a role of an ASIC in pain has come from Sluka, who created a persistent pain paradigm triggered by muscle acidity in mice. Two muscle

65

66


injections of acid some time apart create a long-lived hypersensitivity to touch of nearby skin [76]. This very reproducible syndrome simply does not occur in mice that lack ASIC3 [77]. The condition may be analogous to some forms of temporomandibular joint syndrome, in which muscle or joint problems cause chronic pain throughout the face. More traditional assays provide mixed messages. Two labs have made ASIC3 knockout mice that differ in their behavior. With i.p. injections of 0.6 % acetic acid, the Zimmer lab knockout shows an increase (not decrease) in writhing over wild type [78]. The Welsh lab knockout exhibits no change in paw licking after acid injections and a decrease in mechanical hyperalgesia after intramuscular acid injection [79]. With regard to noxious heat, the Zimmer knockout is more sensitive to temperatures above 50 8C in a hot-plate test [78]; the Welsh knockout shows a decreased response of heat sensitive C-fibers to noxious heat and normal paw withdrawal to radiant heat [79]. The Zimmer lab knockout differs from the wild type only in tests at high intensity, suggesting a role for ASIC3 in high-intensity pain in different modalities. The Welsh lab knockout demonstrates a role for ASIC3 in acid and carrageenan-evoked hyperalgesia. In both cases, neither knockout shows a severe phenotype, indicating that there are other players in all of these types of pain. It seems fair to summarize the data on knockout mice to indicate that they give subtle effects in pain assays and that different labs report different results. A crucial concern is that these experiments cannot address the hypothesis that ASICs are sensors for ischemic pain such as angina because there is no mouse model for such pain.

3.7.2

Mechanosensation

ASICs are considered as possible mechanosensing channels because of their homology to degenerins, which are expressed on mechanosensing C. elegans neurons (see Chapter 1). ASIC2a and ASIC3 are expressed on mechanosensory terminals in rat skin [79, 80]. ASIC2a knockouts show decreased sensitivity of rapidly adapting mechanoreceptors [81]. ASIC3 knockouts have altered mechanosensation; however, in this case the rapidly adapting mechanoreceptors become more sensitive [79]. Neither knockout has dramatically altered mechanosensation like the degenerin mutants. However, the degenerin mutation causes loss of the entire mechanosensing cell, not just a single molecule. How might mechanosensitivity arise? Is acid sensitivity relevant? Most models of mechanosensation surmise that the detecting channel is tethered to other proteins in the extracellular and intracellular media. The DEG proteins are thought to form ion channels in a multi-subunit complex necessary for transducing mechanical forces. Many proteins associate with DEGs to form this complex, but very few interacting proteins have been found that associate with ASICs. One protein that associates with ASIC3 is CIPP (channel-interacting PDZ domain protein). CIPP interacts with the carboxy terminus of ASIC3 using one of its four PDZ domains. When expressed with ASIC3 in COS cells, CIPP increases the current fivefold, and shifts

3.8 CNS ASICs

the activation curve towards more basic values [82]. CIPP may act as a scaffold protein to link ASIC3 to other intracellular proteins, since it has four PDZ domains.

3.7.3

Taste

Foods are perceived as sour due to an acid receptor in taste cells (see Chapter 7). Currents generated by acid in taste cells generally have a low amiloride and proton sensitivity. Recently it has been shown that both ASIC2a and -2b are present in a population of rat taste cells and that the heteromer is insensitive to amiloride at low pH [44, 83, 84]. ASIC2a/2b heteromers have a proton sensitivity similar to that seen for sour taste. Two other proton-activated channels have also been implicated in sour taste transduction: VR1 and HCN channels. They can both be activated by acid and are expressed in taste cells. It is unclear to what extent these channels participate in sour taste transduction or whether all are involved.

3.8

CNS ASICs

Immunolocalization studies show that ASIC1a is expressed in areas of high synaptic concentration such as cortex, olfactory bulb, hippocampus, amygdala, and cerebellum [85, 86]. This bolsters the hypothesis that ASICs have a synaptic function. ASICs may be activated in the CNS due to synaptic release. Vesicles become acidified when they are loaded with their neurotransmitter. When vesicles empty their transmitter, they also unload protons, thereby rapidly acidifying the synaptic cleft [87]. Alternatively, there may be some as yet undiscovered ligand of ASICs. ASICs in the CNS may enhance learning and memory formation. The Welsh group knocked out ASIC1 in mice and found deficits in LTP and learning in the knockouts [85, 88]. The mutant mice have decreased facilitation of EPSPs during high-frequency stimulation, deficits in spatial memory, impaired LTP, and impaired eye blink and fear conditioning. Even with all these defects, the mice appear normal and are capable of learning. Interestingly, all proton-gated currents are abolished in the CNS of knockout animals, suggesting that ASIC1a is necessary for proper expression of ASIC2a, the other common CNS ASIC. No role for ASICs in normal synaptic transmission has been demonstrated. Blockade of ASICs with amiloride or desensitization with low pH causes no evident effect on synaptic transmission [86]. Furthermore, the study did not find enrichment of ASIC1 in postsynaptic density (PSD) fractions of cell homogenates [86], unlike other synaptic proteins (e.g., NMDA receptors). However, ASIC2a is enriched in synaptic fractions from cerebellum [89].

67

68


3.9

Stroke

Since ASICs are proton-gated, they may have deleterious side effects in the CNS during pathological conditions such as stroke and seizure. During the ischemic conditions of a stroke or seizure, the pH can drop as low as 6.2 due to lactic acidosis [90], increasing extracellular lactate and proton concentrations. In addition to the pH change, the extracellular Ca++ concentration can drop from 1.2 mM to 0.1 mM [91], due to the translocation of Ca++ into cells. Synaptic release occurring during these pathologies can also release large amounts of Zn++ into the extracellular medium. All of these ionic conditions – increased H+, Zn++, and lactate and decreased Ca++ – are known potentiators of ASIC currents. If ASICs are activated strongly in the CNS, they could conceivably lead to neuronal death, much as constitutively active degenerins kill cells expressing them [12]. Hyperactive ASICs are capable of loading a cell with sodium, creating an osmotic stress and thereby killing the cell. ASIC1a has a significant Ca++ permeability, which could let excess Ca++ into the cell, thereby killing it via Ca++-activated apoptotic pathways. For either of these to occur, substantial activation would be necessary. The transient nature of ASIC currents may protect against such damage, but small sustained ASIC currents occur at many pH levels. Allen and Atwell subjected cerebellar purkinje cells to many different conditions occurring during CNS ischemia [92]. The ASICs in the cerebellar slices are activated when pH falls below 6.8 with a pH0.5 of 6.4. Many ischemic factors enhanced both the maximal and sustained ASIC current in these cells. Arachidonic acid, cell swelling, and lactate all potentiate the acid-gated current, often inducing a sustained component. In a rat model of global ischemia, Johnson et al. found that ASIC2a is upregulated in surviving cells [93]. Since the ASIC1a/2a heteromers are less sensitive to acid than ASIC1a, the increase in ASIC2a may act to suppress activity of ASIC1a homomers by increasing the proportion of ASIC1a in heteromers. Using a pilocarpine model of epilepsy, both ASIC1a and 2b are downregulated [94]. This may be a neuroprotective mechanism to prevent neuronal death.

3.10

Other pH-activated Channels

ASICs may be the most sensitive, but they are not the only channels activated by protons. The vanilloid receptor (VR1 or TRPV1), a member of the transient receptor potential (TRP) family of ion channels, is opened by many types of stimuli, including heat, protons, and capsaicin [95]. Protons activate VR1 receptors by lowering their temperature threshold. The pH needs to drop below 5.9 at room temperature to activate the channel. The current through VR1 receptors is easily distinguished from ASICs since it does not desensitize and is a nonselective cation channel. VR1 is thought to transduce noxious heat and capsaicin sensitivity in vivo [96] (see Chapters 10 and 11). Hyperpolarization-activated and cyclic nucleotide-gated channels (HCNs)

3.10 Other pH-activated Channels

are also activated by protons. These cation channels are normally opened by hyperpolarization; however, both cyclic nucleotides and extracellular protons shift the activation curve to more positive voltages [97, 98]. At very low pH values, this shifts to above the resting potential of the cell, thereby activating the channel. As sensory transduction channels, HCN1 and HCN4 are thought to participate in sour taste transduction. There is also a family of potassium channels that shows proton gating. The TASKs (TWIK-related acid-sensitive K+ channels) are members of the tandem pore potassium channel family. Instead of being activated by low pH, TASK1 and TASK3 are inactivated by acid. The IC50 for protons of TASK1 and TASK3 are pH 7.3 and 6.3, respectively, putting them within the physiological range of pH changes, especially TASK1 [99]. Low pH excites cells expressing these channels by eliminating a resting potassium conductance.

References 1

2

3

4

5

6

7

8

9

10

Krishtal, O.A. and V.I. Pidoplichko, A receptor for protons in the nerve cell membrane. Neuroscience, 1980. 5(12): p. 2325–7. Krishtal, O.A. and V.I. Pidoplichko, Receptor for protons in the membrane of sensory neurons. Brain Research, 1981. 214: p. 150–154. Krishtal, O.A. and V.I. Pidoplichko, A “receptor” for protons in small neurons of trigeminal ganglia: possible role in nociception. Neuroscience Letters, 1981. 24: p. 243–246. Korkushko, A.O. and O.A. Krishtal, Blocking of proton-activated sodium permeability of the membranes of trigeminal ganglion neurons in the rat by organic cations. Neirofiziologiia, 1984. 16(4): p. 557–561. Morad, M., Proton-induced transformation in gating and selectivity of the calcium channel in neurons. Proton passage across cell membranes, 1988: p. 187–200. Waldmann, R., et al., A proton-gated cation channel involved in acid-sensing. Nature, 1997. 386(6621): p. 173–7. Krishtal, O., The ASICs: Signaling molecules? Modulators? Trends in Neurosciences, 2003. 26(9): p. 477–483. Kellenberger, S. and L. Schild, Epithelial sodium channel/degenerin family of ion channels: a variety of functions for a shared structure. Physiological Reviews, 2002. 82: p. 735–767. de la Rosa, D.A., et al., Structure and Regulation of Amiloride-Sensitive Sodium Channels. Annual Review of Physiology, 2000. 62(1): p. 573–594. Lingueglia, E., et al., Cloning of the amiloridesensitive FMRFamide peptide-gated sodium channel. Nature, 1995. 378: p. 730–733.

11

12

13

14

15

16

17

18

Benos, D.J. and B.A. Stanton, Functional domains within the degenerin/epithelial sodium channel (Deg/ENaC) superfamily of ion channels. J Physiol (Lond), 1999. 520(3): p. 631–644. Chalfie, M. and E. Wolinsky, The identification and suppression of inherited neurodegeneration in Caenorhabditis elegans. Nature, 1990. 345(6274): p. 410–6. Hall, D.H., et al., Neuropathology of degenerative cell death in Caenorhabditis elegans. J Neurosci, 1997. 17(3): p. 1033–45. Garcia-Anoveros, J., et al., The nematode degenerin UNC-105 forms ion channels that are activated by degeneration- or hypercontractioncausing mutations. Neuron, 1998. 20(6): p. 1231–41. Hong, K. and M. Driscoll, A transmembrane domain of the putative channel subunit MEC-4 influences mechanotransduction and neurodegeneration in C. elegans. Nature, 1994. 367(6462): p. 470–473. Waldmann, R., et al., The mammalian degenerin MDEG, an amiloride-sensitive cation channel activated by mutations causing neurodegeneration in Caenorhabditis elegans. J Biol Chem, 1996. 271(18): p. 10433–6. Price, M.P., P.M. Snyder, and M.J. Welsh, Cloning and expression of a novel human brain Na+ channel. J Biol Chem, 1996. 271(14): p. 7879–82. Garcia-Anoveros, J., et al., BNaC1 and BNaC2 constitute a new family of human neuronal sodium channels related to degenerins and epithelial sodium channels. Proc Natl Acad Sci U S A, 1997. 94(4): p. 1459–64.

69

70

3 Acid-sensing Ion Channels 19

20

21

22

23

24

25

26

27

28

29

30

31

Babini, E., et al., Alternative splicing and interaction with di- and polyvalent cations control the dynamic range of acid-sensing ion channel (ASIC) 1. J. Biol. Chem., 2002: p. M205877200. Alvarez de la Rosa, D., et al., Functional implications of the localization and activity of acidsensitive channels in rat peripheral nervous system. Proc Natl Acad Sci U S A, 2002. 99(4): p. 2326–31. Waldmann, R., et al., Molecular cloning of a non-inactivating proton-gated Na+ channel specific for sensory neurons. J Biol Chem, 1997. 272(34): p. 20975–8. Sutherland, S.P., et al., Acid-sensing ion channel 3 matches the acid-gated current in cardiac ischemia-sensing neurons. Proc Natl Acad Sci U S A, 2001. 98(2): p. 711–6. Lingueglia, E., et al., A modulatory subunit of acid sensing ion channels in brain and dorsal root ganglion cells. J Biol Chem, 1997. 272(47): p. 29778–83. Champigny, G., et al., Mutations Causing Neurodegeneration in Caenorhabditis elegans Drastically Alter the pH Sensitivity and Inactivation of the Mammalian H+-gated Na+ Channel MDEG1. J. Biol. Chem., 1998. 273(25): p. 15418–15422. Chen, C.-C., et al., A sensory neuron-specific, proton-gated ion channel. PNAS, 1998. 95(17): p. 10240–10245. Bassler, E.L., et al., Molecular and functional characterization of acid-sensing ion channel (ASIC) 1b. J Biol Chem, 2001. 276(36): p. 33782–7. Ugawa, S., et al., Cloning and functional expression of ASIC-b2, a splice variant of ASIC- b. Neuroreport, 2001. 12: p. 2865–2869. Grunder, S., et al., A new member of acid-sensing ion channels from pituitary gland. Neuroreport, 2000. 11(8): p. 1607–11. Akopian, A., et al., A new member of the acidsensing ion channel family. Neuroreport, 2000. 11(10): p. 2217–2222. Babinski, K., K.-T. Le, and P. Seguela, Molecular Cloning and Regional Distribution of a Human Proton Receptor Subunit with Biphasic Functional Properties. J Neurochem, 1999. 72(1): p. 51–57. Firsov, D., et al., Mutational Analysis of Cysteine-rich Domains of the Epithelium Sodium Channel (ENaC). IDENTIFICATION OF CYSTEINES ESSENTIAL FOR CHANNEL EXPRESSION AT THE CELL SURFACE. J. Biol. Chem., 1999. 274(5): p. 2743–2749.

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

North, R.A., Families of ion channels with two hydrophobic segments. Curr Opin Cell Biol, 1996. 8(4): p. 474–83. Doyle, D.A., et al., The Structure of the Potassium Channel: Molecular Basis of K+ Conduction and Selectivity. Science, 1998. 280(5360): p. 69–77. Sheng, S., et al., Characterization of the Selectivity Filter of the Epithelial Sodium Channel. J. Biol. Chem., 2000. 275(12): p. 8572–8581. Sheng, S., et al., Epithelial Sodium Channel Pore Region. STRUCTURE AND ROLE IN GATING. J. Biol. Chem., 2001. 276(2): p. 1326–1334. Coscoy, S., et al., The Pre-transmembrane 1 Domain of Acid-sensing Ion Channels Participates in the Ion Pore. J. Biol. Chem., 1999. 274(15): p. 10129–10132. Snyder, P.M., et al., Electrophysiological and biochemical evidence that DEG/ENaC cation channels are composed of nine subunits. J Biol Chem, 1998. 273(2): p. 681–4. Coscoy, S., et al., The Phe-Met-Arg-Phe-amideactivated Sodium Channel Is a Tetramer. J. Biol. Chem., 1998. 273(14): p. 8317–8322. Firsov, D., et al., The heterotetrameric architecture of the epithelial sodium channel (ENaC). EMBO J., 1998. 17(2): p. 344–352. Eskandari, S., et al., Number of Subunits Comprising the Epithelial Sodium Channel. J. Biol. Chem., 1999. 274(38): p. 27281–27286. Canessa, C.M., et al., Amiloride-sensitive epithelial Na+ channel is made of three homologous subunits. Nature, 1994. 367: p. 463–467. Bassilana, F., et al., The acid-sensitive ionic channel subunit ASIC and the mammalian degenerin MDEG form a heteromultimeric H+gated Na+ channel with novel properties. J Biol Chem, 1997. 272(46): p. 28819–22. Baron, A., R. Waldmann, and M. Lazdunski, ASIC-like, proton-activated currents in rat hippocampal neurons. J Physiol, 2002. 539(Pt 2): p. 485–94. Ugawa, S., et al., Amiloride-Insensitive Currents of the Acid-Sensing Ion Channel-2a (ASIC2a) / ASIC2b Heteromeric Sour-Taste Receptor Channel. J. Neurosci., 2003. 23(9): p. 3616–3622. Babinski, K., et al., Mammalian ASIC2a and ASIC3 Subunits Co-assemble into Heteromeric Proton-gated Channels Sensitive to Gd3+. J. Biol. Chem., 2000. 275(37): p. 28519–28525. Benson, C.J., et al., Heteromultimers of DEG/ ENaC subunits form H+-gated channels in mouse sensory neurons. PNAS, 2002. 99(4): p. 2338–2343.

3.10 Other pH-activated Channels 47

48

49

50

51

52

53

54

55

56

57

58

59

Adams, C.M., P.M. Snyder, and M.J. Welsh, Paradoxical stimulation of a DEG/ENaC channel by amiloride. J Biol Chem, 1999. 274(22): p. 15500–4. McNicholas, C.M. and C.M. Canessa, Diversity of Channels Generated by Different Combinations of Epithelial Sodium Channel Subunits. J. Gen. Physiol., 1997. 109(6): p. 681–692. Garcia, M., et al., Amiloride analogs inhibit Ltype calcium channels and display calcium entry blocker activity. J. Biol. Chem., 1990. 265(7): p. 3763–3771. Tang, C.M., F. Presser, and M. Morad, Amiloride selectively blocks the low threshold (T) calcium channel. Science, 1988. 240(4849): p. 213–215. Murata, Y., et al., Non-selective effects of amiloride and its analogues on ion transport systems and their cytotoxicities in cardiac myocytes. Japanese Journal of Pharmacology, 1995. 68(3): p. 279–85. Ismailov, I.I., et al., Identification of an Amiloride Binding Domain within the alpha -Subunit of the Epithelial Na+ Channel. J. Biol. Chem., 1997. 272(34): p. 21075–21083. Schild, L., et al., Identification of Amino Acid Residues in the alpha , beta , and gamma Subunits of the Epithelial Sodium Channel (ENaC) Involved in Amiloride Block and Ion Permeation. J. Gen. Physiol., 1997. 109(1): p. 15–26. Escoubas, P., et al., Isolation of a tarantula toxin specific for a class of proton-gated Na+ channels. J Biol Chem, 2000. 275(33): p. 25116–21. Escoubas, P., et al., Recombinant production and solution structure of PcTx1, the specific peptide inhibitor of ASIC1a proton-gated cation channels. Protein Sci, 2003. 12(7): p. 1332–1343. Baron, A., et al., Zn2+ and H+ are coactivators of acid-sensing ion channels. J Biol Chem, 2001. 276(38): p. 35361–7. de Weille, J. and F. Bassilana, Dependence of the acid-sensitive ion channel, ASIC1a, on extracellular Ca(2+) ions. Brain Res, 2001. 900(2): p. 277–81. Korkushco, A.O., O.A. Krishtal, and N.I. Chernevskaya, Steady-state characteristics of the proton receptor in the somatic membrane of rat sensory neurons. Neirofiziologiia, 1983. 15: p. 632–638. Immke, D.C. and E.W. McCleskey, Lactate enhances the acid-sensing Na+ channel on ischemia-sensing neurons. Nat Neurosci, 2001. 4(9): p. 869-70.

60

61

62

63

64

65

66

67

68

69

70

71

72

Roumy, M. and J.-M. Zajac, Neuropeptide FF, pain and analgesia. European Journal of Pharmacology, 1998. 345(1): p. 1–11. Askwith, C.C., et al., Neuropeptide FF and FMRFamide potentiate acid-evoked currents from sensory neurons and proton-gated DEG/ ENaC channels. Neuron, 2000. 26(1): p. 133–41. Catarsi, S., K. Babinski, and P. Seguela, Selective modulation of heteromeric ASIC protongated channels by neuropeptide FF. Neuropharmacology, 2001. 41(5): p. 592–600. Deval, E., et al., Effects of neuropeptide SF and related peptides on acid sensing ion channel 3 and sensory neuron excitability. Neuropharmacology, 2003. 44(5): p. 662–671. Xie, J., et al., ASIC3 and ASIC1 mediate FMRFamide-related peptide enhancement of H+-gated currents in cultured dorsal root ganglion neurons. J Neurophysiol, 2003: p. 00707.2002. Immke, D.C. and E.W. McCleskey, Protons open Acid-sensing ion channels by catalyzing relief of ca(2+) blockade. Neuron, 2003. 37(1): p. 75–84. Askwith, C.C., et al., DEG/ENaC ion channels involved in sensory transduction are modulated by cold temperature. PNAS, 2001. 98(11): p. 6459–6463. Adams, C.M., et al., Protons activate brain Na+ channel 1 by inducing a conformational change that exposes a residue associated with neurodegeneration. J Biol Chem, 1998. 273(46): p. 30204–7. Berdiev, B.K., et al., Actin modifies Ca2+ block of epithelial Na+ channels in planar lipid bilayers. Biophys J, 2001. 80(5): p. 2176–86. Ismailov, II, et al., Mechanosensitivity of an epithelial Na+ channel in planar lipid bilayers: release from Ca2+ block. Biophys J, 1997. 72(3): p. 1182–92. Waddell, P.J. and S.N. Lawson, Electrophysiological properties of subpopulations of rat dorsal root ganglion neurons in vitro. Neuroscience, 1990. 36(3): p. 811–822. Olson, T.H., et al., An acid sensing ion channel (ASIC) localizes to small primary afferent neurons in rats. Neuroreport, 1998. 9(6): p. 1109–1113. Ugawa, S., et al., Amiloride-blockable acid-sensing ion channels are leading acid sensors expressed in human nociceptors. J. Clin. Invest., 2002. 110(8): p. 1185–1190.

71

72

3 Acid-sensing Ion Channels 73

74

75

76

77

78

79

80

81

82

83 84

85

Voilley, N., et al., Nonsteroid anti-inflammatory drugs inhibit both the activity and the inflammation-induced expression of acid-sensing ion channels in nociceptors. J Neurosci, 2001. 21(20): p. 8026–33. Mamet, J., et al., ProInflammatory Mediators, Stimulators of Sensory Neuron Excitability via the Expression of Acid-Sensing Ion Channels. J Neurosci, 2002. 22(24): p. 10662–70. Benson, C.J., S.P. Eckert, and E.W. McCleskey, Acid-Evoked Currents in Cardiac Sensory Neurons : A Possible Mediator of Myocardial Ischemic Sensation. Circ Res, 1999. 84(8): p. 921–928. Sluka, K.A., A. Kalra, and S.A. Moore, Unilateral intramuscular injections of acidic saline produce a bilateral, long-lasting hyperalgesia. Muscle Nerve, 2001. 24(1): p. 37–46. Sluka, K.A., et al., Chronic hyperalgesia induced by repeated acid injections in muscle is abolished by the loss of ASIC3 but not ASIC1. Pain, . In Press. Chen, C.C., et al., A role for ASIC3 in the modulation of high-intensity pain stimuli. Proc Natl Acad Sci U S A, 2002. 99(13): p. 8992–7. Price, M.P., et al., The DRASIC cation channel contributes to the detection of cutaneous touch and acid stimuli in mice. Neuron, 2001. 32(6): p. 1071–83. Garcia-Anoveros, J., et al., Transport and localization of the DEG/ENaC ion channel BNaC1alpha to peripheral mechanosensory terminals of dorsal root ganglia neurons. J Neurosci, 2001. 21(8): p. 2678–86. Price, M.P., et al., The mammalian sodium channel BNC1 is required for normal touch sensation. Nature, 2000. 407(6807): p. 1007–11. Anzai, N., et al., The multivalent PDZ domaincontaining protein CIPP is a partner of acidsensing ion channel 3 in sensory neurons. J Biol Chem, 2002. 277(19): p. 16655–61. Ugawa, S., et al., Receptor that leaves a sour taste in the mouth. Nature, 1998. 395: p. 555–556. Lin, W., T. Ogura, and S.C. Kinnamon, AcidActivated Cation Currents in Rat Vallate Taste Receptor Cells. J Neurophysiol, 2002. 88(1): p. 133–141. Wemmie, J.A., et al., Acid-Sensing Ion Channel 1 Is Localized in Brain Regions with High Synaptic Density and Contributes to Fear Conditioning. J. Neurosci., 2003. 23(13): p. 5496–5502.

86

87

88

89

90

91

92

93

94

95

96

97

98

99

de la Rosa, D.A., et al., Distribution, subcellular localization and ontogeny of ASIC1 in the mammalian central nervous system. J Physiol (Lond), 2003. 546(1): p. 77–87. Krishtal, O.A., et al., Rapid extracellular pH transients related to synaptic transmission in rat hippocampal slices. Brain Research, 1987. 436(2): p. 352–356. Wemmie, J.A., et al., The acid-activated ion channel ASIC contributes to synaptic plasticity, learning, and memory. Neuron, 2002. 34(3): p. 463–77. Jovov, B., et al., Immunolocalization of the acidsensing ion channel 2a in the rat cerebellum. Histochemistry and Cell Biology, 2003. 119(6): p. 437–46. Nedergaard, M., et al., Dynamics of interstitial and intracellular pH in evolving brain infarct. Am J Physiol, 1991. 260(3 Pt 2): p. R581–8. Kristian, T., et al., Calcium metabolism of focal and penumbral tissues in rats subjected to transient middle cerebral artery occlusion. Exp Brain Res, 1998. 120(4): p. 503–9. Allen, N.J. and D. Attwell, Modulation of ASIC channels in rat cerebellar Purkinje neurons by ischaemia-related signals. J Physiol (Lond), 2002. 543(2): p. 521–529. Johnson, M.B., et al., Global ischemia induces expression of acid-sensing ion channel 2a in rat brain. Journal of Cerebral Blood Flow and Metabolism, 2001. 21: p. 734–740. Biagini, G., et al., Regional and Subunit-Specific Downregulation of Acid-Sensing Ion Channels in the Pilocarpine Model of Epilepsy. Neurobiology of Disease, 2001. 8(1): p. 45–58. Tominaga, M., et al., The cloned capsaicin receptor integrates multiple pain-producing stimuli. Neuron, 1998. 21(3): p. 531–43. Caterina, M.J., et al., Impaired Nociception and Pain Sensation in Mice Lacking the Capsaicin Receptor. Science, 2000. 288(5464): p. 306–313. Chen, S., J. Wang, and S.A. Siegelbaum, Properties of Hyperpolarization-Activated Pacemaker Current Defined by Coassembly of HCN1 and HCN2 Subunits and Basal Modulation by Cyclic Nucleotide. J. Gen. Physiol., 2001. 117(5): p. 491–504. Stevens, D.R., et al., Hyperpolarization-activated channels HCN1 and HCN4 mediate responses to sour stimuli. Nature, 2001. 413(6856): p. 631–635. Lesage, F., Pharmacology of neuronal background potassium channels. Neuropharmacology, 2003. 44(1): p. 1–7.

73

4

Chemosensory Transduction in Caenorhabditis elegans Noelle L’Etoile

4.1

Introduction 4.1.1

The organism C. elegans

In 1963 Sydney Brenner decided to “tame” the worm Caenorhabditis elegans. He envisioned the worm, a small (1.5 mm long), transparent, free-living soil nematode, to be the metazoan equivalent of the bacteriophage, the perfect beast with which to “dissect the genetic specification of the nervous system.” His stated goal was to understand, in molecular detail, how a nervous system is designed to generate behavior [1]. It is not hard to see why he chose C. elegans: this small beast is an appealing genetic organism; hundreds of animals can inhabit one petri dish, dining on standard laboratory strains of E. coli. With a generation time of only three days, genetic manipulations are just a bit slower than in yeast and large populations can be grown in a matter of days. Since it is a hermaphrodite under standard cultivation conditions, the populations can be essentially cloned and recessive mutations are easily uncovered. The hermaphrodite can also be mated into by males, which allows for mapping of mutations. Males are easily produced by high temperatures or ethanol exposure, and they are maintained by mating. But the key quality that lured Sydney Brenner to the worm, Caeno (recent) rhabditis(rod) elegans (nice), was its photogenic aspect: high-quality electron micrographs could be obtained and used to reconstruct its entire nervous system. In 1986 the hermaphrodite’s 302-neuron nervous system was reconstructed by White and Sulston using serial electron micrographs of two worms [2]. Each chemical synapse and gap junction was noted and the communicating cells were identified. When the worms were compared, relatively little variation of either cell body position or synapse number was observed. In this way, a potential circuit diagram of the worm’s nervous system was described. To create a “functional” circuit diagram, two criteria had to be met: first, an assay for neuronal function had to be established, and second, the contribution of each neuron to the behavior had to be assessed. To this end, each sensory neuron and many of the Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

74

4 Chemosensory Transduction in Caenorhabditis elegans

interneurons were individually ablated by laser. The laser-operated animals were subjected to behavioral studies and their deficits noted. This led to a rough description of some sensory circuits [3]. The behavioral repertoire of such a small nervous system might have proven too restricted to be interesting. The invariance of cell body position and cell number also gives rise to the speculation that C. elegans behavior might prove to be completely “hard-wired,” leaving little room for plasticity. Though C. elegans has neither eyes nor ears, the range of behaviors generated by this simple nervous system is proving to be quite astonishing. The number of “new” behaviors grows as the number of researchers examining the link between genetics and behavior grows. Behavioral responses to external stimuli such as temperature, volatile chemicals, water-soluble chemicals, pheromones, light and harsh touch, and bacteria, which are ill-defined components of other worms, can all be observed reproducibly in the laboratory. The behaviors that arise in response to simple stimuli can be modulated, often by the pairing of food or starvation with a specific second stimulus, to give rise to complex behaviors. For example, “social feeding” or “swarming” arises from the pairing of food with other worms [4]. Pairing of a specific temperature with food results in a preference for or “memory” of the food-associated temperature and aversion to a temperature that is paired with starvation [5–7]. Pairing of starvation with a salt or an odor will cause aversion to the paired compound [8, 9]. A rather bizarre behavior is induced by crowding and starvation: nictation. Nictating worms writhe about each other and form long strands, perhaps in an attempt to catch a ride to greener pastures on a passing insect. Nothing is known about genes that regulate or produce this behavior. Though not a behavior, the decision to go into the alternate, dauer larval stage is a sensory-driven decision that relies on some of the same molecules as the afore-mentioned behaviors. The larval C. elegans can, if starved, crowded, and placed at high enough temperatures, decide to enter the dauer larval state. In this state, it becomes desiccation resistant by sealing off its buccal cavity and growing a thick cuticle. Its metabolism drops, it stores fat, moves little, and pumps not at all. Though the lifespan of a worm in good conditions is a little less than three weeks (20–28 days), as dauers, worms can live for more than six months [3]. Perhaps in an effort to maximize its ability to sense favorable conditions, the dauer’s olfactory sensory cilia expand to twice the size of the equivalent non-dauer larva’s cilia [10]. Once conditions become more hospitable, the dauer will shed its cuticle, resume gonad development, and become a fertile adult. The sensory signal generated by some of these disparate chemical stimuli will be the subject of this chapter. The lens through which we will view these behaviors is the channel that is most directly stimulated by sensory stimulus. The primary signal transduction channels expressed in C. elegans sensory neurons are of only two classes, the tax channels and the osm-related channels.

4.2 The Chemosensory Organs

4.1.2

Introduction to the Channels

Though many behavioral screens have yielded mutations in the CNG TAX channel, the tax channels’ discovery was initiated by the genetic screens for salt chemotaxisdefective mutants carried out in Dr. Richard Russel’s lab [11]. The screens, designed to probe the molecular basis for chemosensation, identified two mutants, tax-2 and tax-4. These were later mapped and cloned by the Bargmann and Ohshima labs and identified as the beta and alpha subunits of the C. elegans cyclic nucleotide-gated channel [12, 13]. The sequence of the C. elegans genome contains four additional subunits that have yet to be characterized. Another primary sensory transduction channel was identified as a result of screens for osmotic (osm) avoidance-defective mutants (Thomas lab), light nose-touch avoidance-defective mutants (Kaplan lab), and volatile chemotaxis-defective mutants (Bargmann lab). This channel consists of the transient receptor potential channel type V (TRPV; see Tab. 8.3) subunit OSM-9 [14] in complex with at least one other TRPV subunit, OCR-2 (OSM-9 capsaicin receptor related). The OCR-1, -2, -3, and -4 subunits were identified by sequence homology to OSM-9 [15–18], and null mutants of ocr-1 and ocr-2 were generated by imprecise excision of a TC1 transposon [16]. This chapter will describe (1) the organs the worm uses to sense chemical and thermal stimuli; (2) the assays employed to monitor and quantify the sensory response; (3) the cells and molecules responsible for generating the behavioral response; and (4) the channels’ structure and how their activity is modulated. A subject that will be touched on only briefly is the role of the sensory channel in development.

4.2

The Chemosensory Organs

Though a terrestrial nematode, C. elegans lives at the air-water interface and senses its chemical environment by taste and smell. C. elegans has three organs devoted to sensing its external chemical environment: the amphid and inner labial organs in the anterior, and the phasmid organ in the posterior. The amphids and phasmids are bilaterally symmetric organs that lie on the left and right sides of the worm, while there are six radially symmetric inner labials. These “organs” are defined by the presence of an opening to the exterior of the animal that is supported by structural epithelial cells, such as the socket cell, that hold the exposed sensory cilia in place [19, 20].

4.2.1

The Amphid Organ

The amphid (anterior) organ contains 12 pairs of sensory neurons, eight of which are directly exposed to the environment and four of which are embedded within a glial-like sheath cell (see Fig. 4.1). The exposed amphid neurons (ADF, ADL, ASE, ASG, ASH,

75

76


Fig. 4.1 The sensory organs of C. elegans. Top: schematic of the worm with amphid, inner labial, and phasmid organs labeled. Bottom left: cartoon representation of the sensory endings of the amphid neurons. ASE, ASG, ASH, ASI, ASK, ASJ, AFD,

and ADL cilia are bounded by the socket cell and exposed to the environment (Adapted from [75]). Structure of the sensory endings of these amphid neurons. Adapted from [21]

ASI, ASJ, and ASK) mostly sense micromolar concentrations of water-soluble compounds, while the sheathed neurons (AWA, AWB, and AWC) sense volatile compounds in the picomolar range while AFD senses heat [21, 22]. Using these cells, the worm responds to both attractive and repulsive stimuli. Some amino acids, cyclic nucleotides, basic pH, salts, and a host of volatile compounds (at low concentration) are actively sought out by the worm, while D-tryptophan, acid pH, high osmotic strength, heavy metals, detergents, and high concentrations of volatiles are actively avoided. Although a single amphid sensory neuron may be able to sense a number of compounds, the behavioral response generated by that neuron will lead to either attraction or repulsion, never both [23]. Cells within the amphid organ can be grouped into two categories: those that probably use the cyclic nucleotide-gated channel comprised of the alpha and beta subunits TAX-4 and TAX-2 as their primary signal transduction channel [12, 13], and those that are predicted to use both the OSM-9 and OCR-2 subunits of the transient receptor potential (TRP) channel [16] (see Tab. 4.1). ASH and AWA express both TRP channel subunits, and mutation of either subunit abolishes both ASH-mediated repulsion and AWA-mediated chemotaxis [14]. AWC, AFD, ASE, and ASK all express TAX- 2 and TAX-4, and responses generated by these

4.2 The Chemosensory Organs Tab. 4.1

C. elegans chemosensory organs, neurons, stimuli, and channels

Organ

Neuron

Class of stimulus sensed

Channel used

Channels expressed

Amphid

AWA

Attractive-volatile

TRP

AWC

Attractive-volatile

TAX

AWB ADL

Repulsive-volatile Repulsive-volatile, water-soluble Repulsive osmolarity, nose-touch, water-soluble Attractive salts

TAX ?

OSM-9, OCR-1, OCR-2 TAX-2, TAX-4, OSM-9 TAX-2, TAX-4 OSM-9, OCR-1, OCR-2 OSM-9, OCR-2

TAX

AFD PHA

Repulsive water-soluble/ Attractive amino acid Dauer pheromone/ Attractive water-soluble (minor) Dauer pheromone (minor)/ Attractive water-soluble (minor) Dauer pheromone/ Attractive water-soluble (minor) Dauer recovery promoting factors, e.g., food Temperature Repulsive water-soluble

TAX ?

PHB

Repulsive water-soluble

?

IL1 and IL2 (six each) URX/ BAG

IL1 may be mechanosensory IL2 may be chemosensory Internal cues that promote “social” feeding

?

TAX-2, TAX-4, OSM-9 TAX-2, TAX-4, OSM-9 TAX-2, TAX-4, OSM-9 TAX -2, TAX-4 TAX-2(?), TAX-4, OSM-9(?) TAX-2(?), OCR-2 TAX-4, OSM-9(?) OSM-9, OCR-2

TAX

TAX-2, TAX-4

ASH ASE ASK ADF ASG ASI ASJ

Phasmid

Inner labials Cilia exposed to worm’s interior

TRP TAX

? ? ? TAX

TAX-2, TAX-4, OSM-9 TAX-2, TAX-4, OSM-9 OSM-9, OCR-2

cells are defective in tax-2 and tax-4 mutants [12, 13]. AWC and ASE express the CNG channel subunits along with one TRP channel subunit, OSM-9. osm-9 mutants chemotax well to AWC and ASE-sensed compounds; therefore, OSM-9 is not the primary sensory signaling channel [8, 24, 25]. In fact, though hour-long exposures to these compounds will adapt the wild-type worm’s responses, osm-9 mutants never adapt; their response to AWC and ASE-sensed compounds is as robust after exposure as before [8, 25]. It is important to note that these adaptation defects appear after prolonged exposure and do not interfere with the rapid desensitization that is required for normal sensation of a gradient. Thus, expression of both channel subunits predicts which channel is used as the primary signal transduction channel [16], and OSM-9 would seem to promote sensory adaptation when it is expressed by itself in the context of TAX-2 and TAX-4 expression. With a very limited set of at most 32 sensory neurons, C. elegans is able to respond to a wide array of stimuli. The strategy by which the nematode diversifies its ability to

77

78


respond to its environment without expanding its array of sensory neurons is to have multiple receptors on each neuron. Interestingly, there seems to be bit of redundancy: at least two pairs of neurons respond to most classes of stimuli. In general one pair will express the TAX CNG channel and the other will express the TRP channel (Tab. 4.1). Two pairs of cells sense volatile compounds and cause a chemotaxis response. One pair, the AWCs, uses TAX-4/2 [12, 13], and the other, the AWAs, uses the TRP channels [24]. Three pairs detect repulsive volatiles: the AWB pair uses TAX channels [26], while ASH [14] and ADL [14] use the TRP channel [16]. Though ASH, ADL, and ASK each detect water-soluble repellents, ASH and ADL express the TRP channel, while ASK uses the TAX channel [27]. Four pairs of neurons seem to be important for regulating the dauer response: ASG, ASI, ASJ, and ADF. Of these, ADF expresses the TRP channel and the others use the TAX channel [14]. What the worm achieves by this division of labor is unclear. Independent modulation of cells that utilize either class of channel may allow for a greater regulation of behavior by either developmental or environmental influences.

4.2.2

Phasmid Organ

The phasmid organ contains two sensory neurons, PHA and PHB (Fig. 4.1). Until recently, little was known about the function of the phasmid organ. Using a new assay, the “dry drop test,” Hilliard et al. [28] examined the phasmids’ function. In this assay, a repellant is allowed to soak into the agar immediately in front of a moving worm. The worm runs into the spot containing a high concentration of the repellant and backs away from it. If an additional spot of repellant is placed behind the worm, it will back into the second spot and move forward. Thus, the “dry drop test” separates amphid from phasmid responses. The duration of the backing response is 3 s when a repellent compound is encountered solely by the head of the worm. The second dry drop encountered by the tail decreases the duration of the backing response to 2 s. The decrease indicates that neurons in the tail promote the forward movement of the worm. This makes sense if the phasmids respond to repellant stimuli by making the worm move away from an encounter with a noxious compound that is localized at the rear of the worm. The contribution of the phasmids to this behavior was uncovered first by weakening the repellent-induced backing response. This was accomplished by laser-ablating a subset of the amphid neurons responsible for sensing repellents. This allowed Hilliard et al. [28] to observe a weakened backing response that was dependent upon an intact phasmid support cell. This was mimicked by a TAX-4 mutation. The phasmids are interesting in that they express both TRP subunits (OCR-2 and OSM-9) [16] and require both TAX channel subunits for proper axon outgrowth [12].

4.3 Behavioral Assays

4.2.3

Inner Labial

The six inner labials each contain two types of neurons, IL1 and IL2. Not much is known about the function of the inner labials. They have sensory cilia that are exposed to the environment. It is hard to predict which channels these neurons use because they seem to express only the OSM-9 subunit of the TRP channel [16, 24].

4.2.4

The Sensory Signaling Circuit

Sensory neurons relay their information to each other via chemical synapses and gap junctions. They also synapse onto interneurons that, in turn, integrate and relay sensory information to the head motor neurons. It is the head motor neurons that allow the animal to swim in response to a sensory cue [2]. The disparity between the resolution with which we can describe the physical circuit and the resolution with which we describe the actual flow of information is appalling. The traditional approaches of laser-ablation studies may need to be combined with cellspecific reporters of neuronal function such as the cameleon GFP pioneered by the Schafer lab [29] or the pericam GFP [30] used by the Axel lab to study olfactory processing in Drosophila in order to achieve an understanding of the path that information takes through the worm’s “brain.” Perhaps, too, the Mariqc lab’s use of electrophysiology [31] will allow a description of the circuits that generate chemosensory behavior. Another approach is to reconstitute the circuit to determine the minimal number and type of connection that would allow the worm to perform a specific behavior. This could be accomplished in the glutamate transporter minus, mutant eat-4, background [32] by placing the cDNA for the transporter, EAT-4, downstream of promoters specific for particular sensory neurons. For example, to reconstitute the olfactory chemosensory circuit, AWCs, specific interneurons and specific head motor neurons may all have to express the glutamate transporter to enable the eat-4 null worm to chemotax toward an AWC-sensed odor. The minimal number would give a clue as to the true working circuit diagram.

4.3

Behavioral Assays

In order to create a circuit diagram and to understand the molecular underpinnings of behavior, robust behavioral assays had to be developed. Initial investigations into C. elegans chemosensory behaviors focused on chemotaxis towards food or specific chemicals and thermotaxis towards its preferred temperature [5, 11, 21]. The following descriptions of the assays used to probe the role of the TAX and TRP channels in generating a behavioral response are meant to allow the reader both to appreciate

79

80


the limitations of these assays and to evaluate some of the data presented in the subsequent sections.

4.3.1

Chemotaxis

To assess chemotaxis of a single worm, a slightly dry square plate is used. The worm is placed onto the center of the plate, the diluted compound of interest is placed at one end, and the diluent is placed at the other end [33, 34] (Fig. 4.2, top left). The worm is allowed to roam the plate for one hour. A grid is drawn on the plate and numbered such that the center, vertical line has the value 0 and the next flanking lines are 2 in the direction of the odor and –2 in the direction of the diluent. The tracks left by the worm as it navigated the chemical gradient are noted, and each time a track crosses a line, the value of the line is added. Thus, a journey that takes the worm straight to the chemical would accrue a value of 6 and one that takes it directly away from the odor –6. In the absence of a chemical cue, the value is generally 0. Such assays are generally carried out on laser-operated individuals.

Fig. 4.2 Behavioral assays for sensory signaling. Top left: single-worm attraction or repulsion assay. Top right: population-attraction assay. Bottom right: population-repulsion assay. Bottom right: thermosensation assay

4.3 Behavioral Assays

Population assays (Fig. 4.2, top right) are more statistically powerful and can be used to compare animals with different genotypes. These assays are useful in genetic mapping studies and when subtle changes in behavior need to be assessed. Population assays are carried out by placing a large number of animals (100) onto a 10-cm plate containing a thin film of agar and then placing a point source of the compound of interest and its diluent equidistant from the initial placement of the worms. An anesthetic is placed at the point source of the compound and at the counter–point source of the diluent control. After one hour, the worms at each point and throughout the plate are counted and a chemotaxis index is derived [35]. The point source sets up a gradient whose shape can be measured by conductivity for a soluble compound [36] but is really unknowable for a volatile one. The assays, however, are quite robust; mutants have been isolated in behavioral screens where those animals that fail to seek out the point source of the attractive chemical are isolated. In a careful study, Pierce-Shimomura et al. [36] concluded that C. elegans exhibits a biased random walk towards the attractive point source. This is a slight modification of the bacterial run and tumble where tumbles are initiated by a decrease in attractant concentration and runs are promoted by increases in attractant concentration. These authors found a slight bias in how a worm reoriented itself after the equivalent of a tumble (coined a pirouette). In the absence of a gradient, the reorientation of the worm after a spontaneous pirouette was 1808 from the initial direction of the worm’s movement. However, if the worm was traveling up a concentration gradient, the worm’s head was more likely to be pointed into the gradient than away from it, regardless of the direction in which it was initially traveling.

4.3.2

Repulsion

Repulsion from volatile chemicals has been tested using a plate assay as described for single-worm assays ([33], Fig. 4.2, top left). The negative chemotaxis examined in this assay requires sensation of an extended spatial gradient. This is less important for the response to soluble repellents. To assay repulsion from soluble repellents such as high osmolarity (8 M glycerol) noxious chemicals (SDS), the “drop test” has been used [16, 28]. In this assay, a droplet of chemical is placed just ahead of a moving worm so that the worm enters the drop of its own accord. Repellants such as quinine cause rapid backing of the worm once it enters the drop. The worm then changes its direction completely. Water-soluble repellents are sensed by the head and tail of the animal. The animal compares the stimulation at the head to that at the tail in deciding in which direction to move. Thus, a repellant at the head causes the worm to back. If the same repellent is sensed by the tail, the worm “realizes” it is surrounded and that moving backward may be worse than forward; thus, the response is somewhat randomized. A modification of this assay, the “dry drop” test, is described along with the phasmid organ. To assay populations of worms for their ability to respond to a water-soluble repellent such as 4 M fructose, a 2.5-cm diameter ring of the compound is placed onto a 6

81

82


cm normal growth plate and the population of worms is placed at the center of the ring (Fig. 4.2, bottom left). The number of worms that cross the ring in 10 minutes is counted ([37, 38]). Wild-type worms will usually back in response to the ring of noxious compound, but osm mutants will ignore it.

4.3.3

Thermotaxis

Temperature is of great importance to a small, poikilothermic animal such as C. elegans. To analyze individuals and populations, the radial thermal gradient assay can be performed [5, 6, 39]. If it is fed at a specific temperature and presented with a radial temperature gradient, the worm will seek out the thermocline that matches the temperature at which it last dined. The animal will then track the thermocline. The radial temperature gradient (Fig. 4.2, bottom right) is produced by placing a small (closed) vial of frozen ammonium acetate (the melting temperature is 17 8C) at the center of a petri dish containing a thin film of agar and then moving the dish to 25 8C. To assess the thermal preference of a population of worms, the percent of worms at the center (17 8C), the middle (20 8C), and the edge (25 8C) of the dish is determined. Wild-type worms will accumulate in the region that most closely matches their cultivation temperature [5, 6, 39]. The precision with which the worm tracks the thermocline indicates that C. elegans can detect a thermal gradient of less than 0.1 8C. Once a worm is starved at a given temperature, it will avoid that temperature. The “memory” of temperature can be dissected from the ability to sense temperatures. If the major thermosensory neuron, AFD, is ablated, most animals become athermotactic, that is, they will move randomly on a temperature gradient, while some will be cryophilic. If the AIZ interneuron is ablated, they become thermophilic but will still track the new hotter isotherm [6]. This indicates that interneuron-ablated animals can sense temperature gradients; they just fail to determine what the “right” temperature is. Other population assays employ an extended gradient [5, 40]. These will not be examined here.

4.3.4

“Social Feeding” or Bordering

To quantify the gathering of worms at the borders of the lawn, a large number of animals (100) are placed onto a 2.5-cm diameter lawn of bacteria that has been grown for two days so that it has a defined border. After three hours, the fraction of animals at the border of the lawn is calculated [41]. To establish whether a single worm will feed socially, that worm is “marked” by a GFP transgene and mixed with non-fluorescent “social” animals. If it joins the clumps of social animals, it is deemed “social” [4].

4.4 How Is The Response to Each Stimulus Generated?

4.4

How Is The Response to Each Stimulus Generated?

This section will describe what is known about the signal transduction pathway from receptor-stimulus interaction to the channel opening. 4.4.1

The Chemotaxis Olfactory Response

Together, the AWC and AWA olfactory neuron pairs are responsible for chemotaxis towards more than 60 known single compounds [35]. AWA expresses ODR-10, the receptor for diacetyl, the odor of buttery popcorn [42]. Thus far, of the 1000 or so GPCRs predicted by the C. elegans genome (Genome Sequencing Consortium [43–45]), no other GPCR has been identified as an odor receptor. One other GPCR, STR-2, is expressed in a suggestive pattern: it is expressed in one of the two AWC neurons. It is randomly expressed in either the left or right AWC, never both [46]. Each olfactory neuron is thought to respond to at least 15 volatile chemicals [46]. One model for olfactory signaling starts with odor binding to a GPCR on an AWA or AWC (Fig. 4.3) neuron and stimulating the G alpha ODR-3 [42, 47]. There are 20 G-alphas, 2 G-betas, and 2 G-gammas expressed in the worm [48]. Of these, the G-alpha ODR-3 seems to be primarily responsible for olfactory signal transduction. An odr-3 null is unable to respond to any volatile compound, nor can it detect repellents [47]. Thus, ODR-3 would seem to be a central player in C. elegans chemosensation. There is redundancy of signaling through the G-alphas, however, as some responses are blunted but not eliminated in the odr-3 null and other responses can be restored to an odr-3 null by loss of a second G-alpha such as GPA-5. In AWC, ODR-3 is expressed along with GPA-2, GPA-13, and GOA-1. In AWA, ODR-3, GPA-5, and GPA-6 are all expressed [48]. The target of ODR-3 activity is not known, as its sequence does not predict whether it is a Gi or a Gs [47]. Downstream of ODR-3, differences between signal transduction in AWA and AWC become apparent. While in AWC, cGMP production via the guanylyl cyclases ODR-1 and DAF-11 is required for response to all odors sensed by these neurons, no guanylyl cyclase is required within AWA [49–51]. Since there is no precedent for a G-alpha stimulating guanylyl cyclases, it is hard to predict whether changes in cGMP levels are produced by activation of the guanylyl cyclases in an undiscovered way or by the inhibition of a phosphodiesterase. In either case, cGMP levels are predicted to rise in response to GPCR signaling, and the increased cGMP is thought to bind and increase the open-probability of the cGMP-gated channel. TAX-2 and TAX-4 are both required for response to all odors sensed by AWC [12, 13]. The cGMP-dependent protein kinase (PKG) EGL-4 is also thought to become activated by odor signaling. The active PKG then downregulates the AWC neuron. One target for its activity may be TAX-2, the beta subunit of the cyclic nucleotide-gated channel. Mutation of a serine residue at the C-terminus to the non-phosphorylatable alanine (S727A) renders the tax channel unable to adapt to short odor exposures [52].

83

84

4 Chemosensory Transduction in Caenorhabditis elegans Fig. 4.3 Models for sensory signal transduction within the olfactory AWC (top), AWA (second from top), gustatory ASE (third from top), and thermosensory AFD (bottom) neurons


In the AWA neuron (Fig. 4.3), odor binding stimulates the G-alpha ODR-3 [47] and this is thought to open, by some unknown mechanism, the TRP-like channel comprised of OSM-9 and OCR-2 [8, 16]. There are no known mutants that are defective for AWA adaptation. Many mysteries remain: each neuron responds to at least 15 odors and each odor response is independently adapted. How is this achieved? What are the multiple guanylyl cyclases doing? What are the multiple G-alphas doing?

4.4.2

Chemotaxis to Water-soluble Compounds

The ASE pair of neurons is responsible for attraction to many salts [50]. ASER allows chemotaxis towards chloride and potassium, while ASEL provides the worm with the ability to sense sodium [36]. Though the ASE right and left neurons respond to different attractants, they both require theTAX-2/4 channel to generate their response [12, 13]. They both express the G-alpha GPA-3 (and no other) [48]. While ASEL expresses two guanylyl cyclases, gcy-6 and gcy-7, ASER expresses only gcy-5 [36, 53]. Electrophysiological studies on the ASE neuron indicated that ASE depolarizes in response to salt exposure, probably as a result of the TAX-2/4 channel opening [54]. Adaptation to salt was shown to require the G-gamma GPC-1 [25]. The involvement of the Ggamma argues that salt sensation requires a GPCR. The primary signal transduction channel could be an ENaC/Deg-type channel and the G-alpha, G-gamma guanylyl cyclases, and the cyclic nucleotide-gated channel might act peripherally. One model for salt sensation (Fig. 4.3) is proposed to start with salt binding to a GPCR and stimulating the G-alpha GPA-3. Once again, no one knows exactly how the levels of cGMP are increased in response to salt stimulation, but the guanylyl cyclases GCY-6 and 7 in ASEL and GCY-5 in ASER may provide the cGMP. The increased cGMP levels lead to channel opening, depolarizing the neuron and leading to chemotaxis. G-gammas are postulated to attenuate GPCR signaling by stimulating beta arrestin kinases to phosphorylate the ligand-bound GPCR, which leads to beta-arrestin binding and inactivation of the receptor [55]. Interestingly, the TRP channel OSM-9 is also required for salt adaptation [25].

4.4.3

Repellents The ASH Polymodal Sensory Neuron When a worm encounters a noxious stimulus such as a water-soluble repellent (bitter compounds, heavy metals, or SDS) [28], high osmotic strength [50], a repulsive volatile (octanol) [26], or a touch on the nose [56], it rapidly backs away and often reverses its course. ASH is the single neuron pair responsible for most of the worm’s escape behavior from these disparate noxious stimuli [28, 50, 56]. How all these disparate cues are sensed by ASH is still a mystery, as no receptor has been identified. Though 4.4.3.1

85

86


mechano-, osmo-, and chemosensation would seem very different, ASH requires one channel comprised of OSM-9 and OCR-2 to respond to each stimulus [16]. The Galpha ODR-3 is also required for all ASH responses [16]. The requirement for ODR-3 makes it seem likely that GPCRs initiate the signaling response to most if not all ASH-sensed stimuli. Worms that express the VR1 capsaicin receptor in ASH respond to capsaicin by backing and reversing. Tobin et al. found that this response did not require ODR-3, indicating that the VR1 channel is directly gated by capsaicin while the native C. elegans TRP channel is probably not directly gated by noxious stimuli. Instead, the G-alpha ODR-3 may initiate DAG signaling and the IP2 pathway may be required to gate this channel. Calcium entry, once again, seems to dampen the response to external stimuli sensed by ASH. The calcineurin A, tax-6 mutant worms reverse in response to concentrations of glycerol (2 M) that wild-type worms cannot sense [39]. The ASH-mediated responses require EAT-4, the glutamate transporter [16]. This indicates that signaling from ASH (or an intermediary neuron) to the backward command interneurons requires glutamate release. Hart et al. [57] and Mariq et al. [58] showed that mechano- and osmosensation could be differentially affected by mutation of a single non-NMDA-type glutamate receptor, GLR-1. Mechanosensation required GLR-1 while osmosensation did not. Mellem et al. [31] suggested that osmosensation activates both NMDA (NMR-1) and non-NMDA glutamate receptor (GLR-1 and GLR2), while mechanosensation activates only the non-NMDA receptor GLR-1/2. They postulate that osmostimulation evokes a larger glutamate release than does mechanostimulation. The larger release may activate both the synaptic GLR-1/2 and the extra synaptic NMR-1, while the smaller mechano-induced release activates only the synaptic GLR-1/2.

4.4.4

Thermotaxis

The ability to find the correct temperature is essential to the worm. At least two sensory neurons are required for the worm to locate the isotherm that matches the temperature at which it was raised (Fig. 4.3). Laser-ablation studies identified AFD as the major neuron required to drive the worm towards the hotter climes. Thus, ablation of AFD makes some wild type worms athermotaxic and others cryophilic [6]. Ablation of the AIY interneuron caused cryophilic behavior, while ablation of the interneuron AIZ caused thermophilic behavior. Thus, it is the interplay between the two interneurons that results in the worm’s ability to seek the appropriate temperature. Laser ablations, however, could not identify the neuron “X,” postulated to signal to AIZ. How temperature generates a signal within AFD is really unknown, though we have hints at the identity of some players. cGMP and calcium are probably major second messengers within AFD. cGMP is implicated because tax-2/4 mutants are athermotactic, meaning that they seek neither hot nor cold [13]. It really is not known whether TAX-2/4 opens or closes in response to heat. In support of heat causing channel opening and calcium entry, the action of the calcineurin A is to dampen the response of


AFD [39]. This is what would be postulated to happen if stimulation is to cause the neuron to become less sensitive to the stimulus and allow it to react to a wider range of temperatures. In support of a major role for cGMP in AFD signaling, Yu et al. [53] found that the guanylyl cyclase GCY-8 is expressed in AFD and may provide the cGMP to open the channel in response to changes in temperature. Several key players are noticeable for their absence from AFD. Though expression studies using engineered GFP reporters must be taken with a grain of salt, such studies failed to observe G-alpha or TRP-like channel expression in AFD. The absence of a G-alpha might indicate that the temperature receptor is not a GPCR. Mori [7] postulated that a receptor guanylyl cyclase such as GCY-8 might be the temperature receptor. Since the TRP channels provide the sensory signal for temperature in mammals, they would seem to be the most obvious candidates for an AFD receptor, but to date they seem neither to be expressed in AFD nor involved in temperature-sensing behavior. Killing AFD caused cryophilic phenotypes, while a tax-2/4 mutant is athermotaxic. This leads to the hypothesis that both AFD and the cryotaxic sensory neuron “X” use the tax channels to signal temperature. Not much more is known about neuron X. 4.4.5

Feeding Behavior

Worms exhibit two strategies when they feed: in a pattern reminiscent of the binary feeding strategies of Drosophila larvae [59], they dine either alone or en masse. The strain adopted by Sydney Brenner, N2, isolated from a compost heap in Bristol, England, dines alone. This makes them easy to pick from a plate and they behave nicely, i.e., they do not tear into the agar or “burrow.” About half of the wild strains actually are much messier [41]; they aggregate at the lawn’s edge and burrow beneath the agar, making them harder to capture with a “pick” (the platinum wire we use to move individuals about). This intriguing behavior spurred identification of its genetic basis. A single gene, originally called bor-1 for its bordering defects (Randy Cassada, personal communication), was shown to be responsible for the phenotype. An inactivating mutation within the neuropeptide Y receptor was shown to cause the bordering, burrowing, and “social feeding” behavior [41]. Half the wild-type strains examined had a polymorphism at a critical residue of the neuropeptide Y receptor that was predicted to decrease its function. Thus, the “solitary” N2 strain has a more active form of the receptor and the “social” strain has a less active form [41]. In C. elegans biology, as is soon to be true of biology in general in the post-genomic era, the genetic basis for behavior is often known and guesses can be made about how the gene is involved with the behavior, but there is a huge gap between the gene and the behavior. What does the gene do? Which neurons require the gene, and what do these neurons actually do for the animal? In the case of the “social feeding” phenotype, Mario de Bono and Cori Bargmann decided to determine which other genes were responsible for the bordering behavior; thus, they looked for suppressors of the bordering and burrowing phenotype. It turned

87

88


out that both the tax and osm channel mutants suppressed this phenotype. That is, TAX-2/4 and OSM-9/OCR-2 are required to interpret the environmental cues that promote bordering, burrowing, and swarming [4]. Juliet Coates found that the TAX channels were required in neurons that had their sensory cilia exposed to the interior of the worm. Furthermore, these cells (AQR, PQR, and URX) express the soluble guanylyl cyclase GCY-32 [53, 60]. Other soluble guanylyl cyclases have been shown to dimerize and produce cGMP via the binding of the gas NO. The worm genome does not have a plausible NO synthase, and the gas that binds GCY-32 therefore remains a mystery. One idea is that the concentration of gas depends on the ratio of bacteria to oxygen in the environment. In the moist environment of the soil, gasses are trapped in specific microenvironments and can indicate the bacterial “load” of that space. Thus, the CO2 and O2 content of a space can indicate a source of food. Worms that are in an airtight environment with an abundance of bacteria can “suffocate” in the anoxic conditions produced by the bacteria. One way for a group of worms to get out of an environment that is teetering on the brink of having too little oxygen is to swarm together to burrow their way out. It would be helpful if worms that liked such environments would enter them together so that they could help each other exit as well. The idea is that they are truly social animals and can go about almost like a living “drill bit” that can bore into and out of high-risk, high-gain environments. de Bono et al. [4] determined that the TRP channel comprised of OSM-9 and OCR-2 functions in the ASH and ADL neurons that respond to water-soluble repellents to promote burrowing behavior in the “social feeding,” npr-1 mutant strain. The environmental cues that ASH and ADL respond to by promoting burrowing are unknown but may be byproducts of bacterial metabolism [4]. Though the decision to enter the alternative, dauer larva stage is not a behavior per se, but rather a developmental decision, it is regulated in part by signals generated by the TAX channels. This is not surprising since many of the cues that lead to the dauer decision are food related and the tax channels are primarily used for food location and sensation. Tax channels are also required for the dauer promoting pathways; thus, channel mutants are potent inhibitors of the dauer decision in some genetic backgrounds.

4.5

Structure of the TAX, Cyclic Nucleotide-gated Channels of the Worm

The alignment of the TAX channel subunits with several representative mammalian channels is shown in Fig. 4.4A [61]. Several key features are revealed in the alignment. The TAX-2 and TAX-4 sequences share key motifs with the vertebrate CNG channels: six transmembrane domains (S1–6), one pore-forming loop (P), and a cyclic nucleotide-binding domain. Examination of the selectivity filter [62] of TAX-4 and TAX-2 indicated that TAX-4 is an alpha subunit and TAX-2 a beta subunit (see Fig. 4.4B). Coburn and Bargmann [12] showed that tax-2 defective mutants can be rescued by over-expression of the TAX-4 cDNA but not the TAX-2 cDNA. This is consistent with the evidence that homomeric alpha subunits can reconstitute channel activity

4.5 Structure of the TAX, Cyclic Nucleotide-gated Channels of the Worm

when expressed heterologously, but beta subunits cannot [63–66]. Corroborating evidence in Komatsu et al. showed that HEK cells transfected with TAX-4 alone were able to generate a cGMP-induced current, but cells transfected with TAX-2 alone were not [67]. The cyclic nucleotide-binding domains of TAX-2 and TAX-4 were predicted to bind cGMP. The cyclic nucleotide-binding domain consists of two alpha helices followed by an anti-parallel beta roll, which is followed by a third alpha helix (the C helix). Aspartic acid 604 within the bovine alpha subunit seems to be responsible for cGMP selectivity in some cases [68]. The corresponding residue within TAX-2 and TAX-4 is also an aspartic acid. The specificity for cGMP was confirmed by Komatsu et al. [67], who showed that HEK-expressed TAX-2/4 channels are half-maximally activated by about 10 lm cGMP, while it took two orders of magnitude more cAMP to reach EC50. cGMP’s EC50 for the TAX-2/4 channel was in the same range as cGMP and cAMP are for the mammalian olfactory channels; this is 10-fold less than cGMP’s EC50 for the rod channels [64, 67]. This may support the idea that in C. elegans, the TAX channels open and the sensory neurons depolarize in response to stimulation. Thus far, the evidence for this is based on electrophysiology of a few neurons including ASE [54], AWC, and AWA [69]. There are four additional predicted cyclic nucleotide-gated channel subunits in the C. elegans genome. F14H8.6 (this channel is the fourth gene on the cosmid F14H8), C23H5.7, Y76B12C.1, and F38E11.12 have been added to the alignment in Fig. 4.4A. cDNA has been identified for F14H8.6, Y76B12C.1, and F38E11.12 (http:// www.wormbase.org). By sequence analysis, F14H8.6 and F38E11.12 seem to be alpha subunits (Fig. 4.4B). F14H8.6 is predicted to be liganded by cGMP, but it is hard to predict the ligand specificity of F38E11.12. Y76B12C.1 and C23H5.7 are predicted beta subunits that could be equally responsive to cAMP or cGMP. Four kilobases of sequence upstream of the start site of the alpha subunit F38E11.12 (formerly known as F38E11.8) were used to drive a GFP expression reporter. Expression was observed in the AWC, AFD, ASE, AWB, and ASI sensory neurons ([27], Tab. 4.2). TAX-2 and TAX-4 are expressed in these same neurons, as well as ASJ, ASK, ASG, URX, and BAG (Tab. 4.2). What might these additional subunits be doing? Recent data from Zhong et al. [70] indicate that a leucine zipper within the C-terminus of the alpha subunits of the rod and olfactory channels trimerizes to give a channel stoichiometry of three alphas to one beta. An examination of the alignment in Fig. 4.4C shows that TAX-4 has the same trimeric leucine zipper (also note Fig. 2a in [70]). TAX-2, like vertebrate beta subunits, does not share the amino acids shown to be critical for co-immunoprecipitation of the trimeric leucine zipper found in the rod alpha subunits. This might indicate that the TAX channels share the mammalian cyclic nucleotide-gated channel’s A3:B1 stoichiometry by containing three subunits of TAX-4 and one of TAX-2 per functional channel. The other alpha subunits, F14H8.6 and F38E11.12, both share the conserved leucine zipper, while the predicted beta subunits do not. Since TAX-4 and F38E11.12 are expressed together in AWC, AFD, ASE, AWB, and ASI, they could form alternate channels [27]. F38E11.12 may act like CNGA4 [71] and change the sensitivity of the TAX channels for factors such as calmodulin that modulate their activity. The fact that these

89

90


Fig. 4.4 (A) Alpha and beta subunit alignments. (B) Selectivity filter. (C) Leucine zipper alignment

channel subunits have not been identified in behavioral screens leads to the speculation that their function is subtle and may be required only for modulating sensory behavior. The alignment of the “new” subunits reveals another interesting feature: each subunit has an extra 46 amino acids between S5 and the pore-forming helix. It has been suggested that this region could form part of the selectivity filter for the channel

4.5 Structure of the TAX, Cyclic Nucleotide-gated Channels of the Worm

Fig. 4.4

continued

91

92

4 Chemosensory Transduction in Caenorhabditis elegans Tab. 4.2

C. elegans cyclic nucleotide-gated channel subunits

Subunit

Alpha/beta

Cyclic nucleotide

Expression

Mutant defects

TAX-4

Alpha (confirmed)

CGMP (confirmed)

Confirmed by cDNA

Chemotaxis-defective: AWC-sensed odors

GFP reporters show: Repulsion: AWB-sensed AWC, AWB, ASE, ASK, odors ASI, ASJ, AFD, ASG, Chemotaxis defective: URX, BAG, PHA ASE-sensed salts Thermotaxis defective (AFD) Suppresses daf-11 dauer constitutivity (multiple neurons) ASJ and ASE axons show aberrant morphology Suppresses npr-1 “social feeding” (URX) Phasmid axons show aberrant morphology TAX-2

Beta (confirmed)

CGMP (confirmed)

Confirmed by cDNA GFP reporters show: AWC, AWB, ASE, ASK, ASI, ASJ, AFD, ASG, URX, BAG, PQR

Chemotaxis-defective: AWC-sensed odors. Repulsion: AWB-sensed odors. Chemotaxis defective: ASE-sensed salts. Thermotaxis defective (AFD). Suppresses daf-11 dauer constitutivity (multiple neurons). ASJ and ASE axons show aberrant morphology. Suppresses npr-1 “social feeding” (URX). Phasmid axons show aberrant morphology

4.6 Channel Regulation Tab. 4.2

C. elegans cyclic nucleotide-gated channel subunits

Subunit

Alpha/beta

Cyclic nucleotide

Expression

Mutant defects

F14H8.6

Beta (proposed)

cGMP (proposed)

Confirmed by cDNA

No known mutations in this gene

F38E11.12

Beta (proposed)

Unable to predict

Y76B12C.1

Alpha (proposed)

Unable to predict

C23H5.7

Alpha (proposed)

Unable to predict

No expression data Confirmed by cDNA GFP reporters show: AWC, AWB, AFD, ASI expression Confirmed by cDNA



No expression data No cDNA confirmation No known mutations in available this gene

(Tsung-Yu Chen and Chul Sun Park, personal communication). The extra sequence has a net negative charge of 6, which might increase the affinity for cations, allow Mg to block the pore more completely, and increase the voltage dependence of the channel.

4.6

Channel Regulation

The classical modulator of olfactory channels is the calcium-binding protein calmodulin [68, 72–74]. The TAX channels could be subject to regulation by calmodulin, but it is difficult to predict whether any of the subunits do, in fact, contain a calmodulinbinding site. There are, however, multiple predicted phosphorylation sites for both cA/GMP-dependent protein kinases and calcium-dependent protein kinases. As discussed in the olfactory signaling section of this chapter, the cGMP-dependent protein kinase, EGL-4, was shown to be required for adaptation of the AWC neuron to odors [52]. Phosphorylation of serine 727 within TAX-2’s C-terminus was shown to be required for effective adaptation to 30 minutes or less of AWC-sensed odor exposure [52]. Longer odor exposures led to adaptation of the olfactory response that was the same as wild-type. This indicates that after brief exposures to odor, a kinase may phosphorylate TAX-2, making it less responsive to stimulation. The position of the phosphorylation site, in the Cterminus but after the cN-binding domain, makes it hard to predict how phosphorylation might regulate channel activity. This phosphorylation site is not conserved in the other beta subunits (Fig. 4.4A).

93

94


Analysis of the phosphatase calcineurin A (tax-6) mutants indicates that the AWC response in tax-6 mutants is hyperadapted [39]. tax-6 mutants are unable to chemotax toward AWC-sensed odors. Kuhara et al. were able to show that the adaptation-defective mutant osm-9 is able to suppress the chemotaxis defects of a tax-6 mutant [39]. Thus, brief exposures to odor may cause rapid down-regulation of AWC’s response to that odor via OSM-9. Usually, calcineurin keeps this rampant adaptation in check [39]. Thus, the kinase EGL-4 and the phosphatase TAX-6 may work in opposition to each other. If they work on the same residue, cGMP produced in response to odor may decrease the sensitivity of the channel to cGMP (making the cell less responsive to odor), while calcium entry through this or another channel may negate this and reset the sensitivity of the channel (and make the cell responsive after a lag). It would be the ratio between the activities of these molecules that would give the AWC neuron its dynamic range. The fact that the double tax-6 osm-9 mutant can chemotax indicates that the phosphatase is not required for the dynamic alterations within AWC that allow for gradient sensation. Prolonged exposures to odor must drive additional changes, perhaps by the action of the PKG in the nucleus [52]. Regulation of the TAX channel in other neurons has not been examined directly. Regulation of the TRP channels may occur by alteration of the subunit composition of the channel: neurons that express both OSM-9 and OCR-2 use this channel for neuronal signaling, while neurons that express only the OSM-9 subunit use it for adaptation of the sensory signal. In both AWC and ASE, the primary chemosensory signal is provided by TAX-2/4, while OSM-9 allows for neuronal adaptation to odor and salt, respectively [8, 14, 25].

References 1 2

3

4

5

6

Brenner, S., The genetics of Caenorhabditis elegans. Genetics, 1974. 77(1): p. 71–94. White, J.G., et al., The structure of the ventral nerve cord of Caenorhabditis elegans. Philos Trans R Soc Lond B Biol Sci, 1976. 275(938): p. 327–48. Sulston, J. and J. Hodgkin, Methods: The nematode Caenorabditis elegans. AKA C. elegans I. Cold Spring Harbor Laboratory Press, 1988: p. 587–606. de Bono, M., et al., Social feeding in Caenorhabditis elegans is induced by neurons that detect aversive stimuli. Nature, 2002. 419(6910): p. 899–903. Hedgecock, E.M. and R.L. Russell, Normal and mutant thermotaxis in the nematode Caenorhabditis elegans. Proc Natl Acad Sci U S A, 1975. 72(10): p. 4061–5. Mori, I. and Y. Ohshima, Neural regulation of thermotaxis in Caenorhabditis elegans. Nature, 1995. 376(6538): p. 344–8.

7

8

9

10

11

Mori, I., Genetics of chemotaxis and thermotaxis in the nematode Caenorhabditis elegans. Annu Rev Genet, 1999. 33: p. 399–422. Colbert, H.A. and C.I. Bargmann, Odorantspecific adaptation pathways generate olfactory plasticity in C. elegans. Neuron, 1995. 14(4): p. 803–12. Saeki, S., M. Yamamoto, and Y. Iino, Plasticity of chemotaxis revealed by paired presentation of a chemoattractant and starvation in the nematode Caenorhabditis elegans. J Exp Biol, 2001. 204(Pt 10): p. 1757–64. Albert, P.S. and D.L. Riddle, Developmental alterations in sensory neuroanatomy of the Caenorhabditis elegans dauer larva. J Comp Neurol, 1983. 219(4): p. 461–81. Dusenbery, D.B., R.E. Sheridan, and R.L. Russell, Chemotaxis-defective mutants of the nematode Caenorhabditis elegans. Genetics, 1975. 80(2): p. 297–309.

4.6 Channel Regulation 12

13

14

15

16

17

18

19

20

21

22

23

24

25

Coburn, C.M. and C.I. Bargmann, A putative cyclic nucleotide-gated channel is required for sensory development and function in C. elegans. Neuron, 1996. 17(4): p. 695–706. Komatsu, H., et al., Mutations in a cyclic nucleotide-gated channel lead to abnormal thermosensation and chemosensation in C. elegans. Neuron, 1996. 17(4): p. 707–18. Colbert, H.A., T.L. Smith, and C.I. Bargmann, OSM-9, a novel protein with structural similarity to channels, is required for olfaction, mechanosensation, and olfactory adaptation in Caenorhabditis elegans. J Neurosci, 1997. 17(21): p. 8259–69. Harteneck, C., T.D. Plant, and G. Schultz, From worm to man: three subfamilies of TRP channels. Trends Neurosci, 2000. 23(4): p. 159–66. Tobin, D., et al., Combinatorial expression of TRPV channel proteins defines their sensory functions and subcellular localization in C. elegans neurons. Neuron, 2002. 35(2): p. 307–18. Caterina, M.J., et al., The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature, 1997. 389(6653): p. 816–24. Bargmann, C.I., Neurobiology of the Caenorhabditis elegans genome. Science, 1998. 282(5396): p. 2028–33. Ward, S., et al., Electron microscopical reconstruction of the anterior sensory anatomy of the nematode Caenorhabditis elegans.?2UU. J Comp Neurol, 1975. 160(3): p. 313–37. Ware, R., et al., Caenorhabditis elegans: sensory input and motor output. J. Comp. Neurol., 1975. 162: p. 71–110. Ward, S., Chemotaxis by the nematode Caenorhabditis elegans: identification of attractants and analysis of the response by use of mutants. Proc Natl Acad Sci U S A, 1973. 70(3): p. 817–21. Dusenbery, D.B., Behavior of free-living nematodes., in Nematodes as Biological Models, B. Zuckerman, Editor. 1980, Academic Press: New York. p. 127–158. Riddle, D.L., C. elegans II. Cold Spring Harbor monograph series 33. 1997, Plainview, N.Y.: Cold Spring Harbor Laboratory Press. xvii, 1222. Colbert, H.A. and C.I. Bargmann, Environmental signals modulate olfactory acuity, discrimination, and memory in Caenorhabditis elegans. Learn Mem, 1997. 4(2): p. 179–91. Jansen, G., D. Weinkove, and R.H. Plasterk, The G-protein gamma subunit gpc-1 of the nematode C.elegans is involved in taste adaptation. Embo J, 2002. 21(5): p. 986–94.

26

27

28

29

30

31

32

33

34

35

36

37

38

39

Troemel, E.R., et al., Divergent seven transmembrane receptors are candidate chemosensory receptors in C. elegans. Cell, 1995. 83(2): p. 207–18. Coburn, C.M., Cyclic Nucleotide Gated Channels in C. elegans, in Thesis. 1996, Univ. California, San Francisco: San Francisco. Hilliard, M.A., C.I. Bargmann, and P. Bazzicalupo, C. elegans responds to chemical repellents by integrating sensory inputs from the head and the tail. Curr Biol, 2002. 12(9): p. 730–4. Suzuki, H., et al., 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, 2003. 39(6): p. 1005–17. Wang, J.W., et al., Two-photon calcium imaging reveals an odor-evoked map of activity in the fly brain. Cell, 2003. 112(2): p. 271–82. Mellem, J.E., et al., Decoding of polymodal sensory stimuli by postsynaptic glutamate receptors in C. elegans. Neuron, 2002. 36(5): p. 933–44. Lee, R.Y., et al., EAT-4, a homolog of a mammalian sodium-dependent inorganic phosphate cotransporter, is necessary for glutamatergic neurotransmission in caenorhabditis elegans. J Neurosci, 1999. 19(1): p. 159–67. Troemel, E.R., B.E. Kimmel, and C.I. Bargmann, Reprogramming chemotaxis responses: sensory neurons define olfactory preferences in C. elegans. Cell, 1997. 91(2): p. 161–9. Sagasti, A., et al., Alternative olfactory neuron fates are specified by the LIM homeobox gene lim4. Genes Dev, 1999. 13(14): p. 1794–806. Bargmann, C.I., E. Hartwieg, and H.R. Horvitz, Odorant-selective genes and neurons mediate olfaction in C. elegans. Cell, 1993. 74(3): p. 515–27. Pierce-Shimomura, J.T., T.M. Morse, and S.R. Lockery, The fundamental role of pirouettes in Caenorhabditis elegans chemotaxis. J Neurosci, 1999. 19(21): p. 9557–69. Culotti, J.G. and R.L. Russell, Osmotic avoidance defective mutants of the nematode Caenorhabditis elegans. Genetics, 1978. 90(2): p. 243–56. Vowels, J.J. and J.H. Thomas, Multiple chemosensory defects in daf-11 and daf-21 mutants of Caenorhabditis elegans. Genetics, 1994. 138(2): p. 303–16. Kuhara, A., et al., Negative regulation and gain control of sensory neurons by the C. elegans calcineurin TAX-6. Neuron, 2002. 33(5): p. 751–63.

95

96

4 Chemosensory Transduction in Caenorhabditis elegans 40

41

42

43

44

45

46

47

48

49

50

51

52

Yamada, Y. and Y. Ohshima, Distribution and movement of Caenorhabditis elegans on a thermal gradient. J Exp Biol, 2003. 206(Pt 15): p. 2581–93. de Bono, M. and C.I. Bargmann, Natural variation in a neuropeptide Y receptor homolog modifies social behavior and food response in C. elegans. Cell, 1998. 94(5): p. 679–89. Sengupta, P., J.H. Chou, and C.I. Bargmann, odr-10 encodes a seven transmembrane domain olfactory receptor required for responses to the odorant diacetyl. Cell, 1996. 84(6): p. 899–909. Genome sequence of the nematode C. elegans: a platform for investigating biology. The C. elegans Sequencing Consortium. Science, 1998. 282(5396): p. 2012–8. Robertson, H.M., Two large families of chemoreceptor genes in the nematodes Caenorhabditis elegans and Caenorhabditis briggsae reveal extensive gene duplication, diversification, movement, and intron loss. Genome Res, 1998. 8(5): p. 449–63. Robertson, H.M., Updating the str and srj (stl) families of chemoreceptors in Caenorhabditis nematodes reveals frequent gene movement within and between chromosomes. Chem Senses, 2001. 26(2): p. 151–9. Troemel, E.R., A. Sagasti, and C.I. Bargmann, Lateral signaling mediated by axon contact and calcium entry regulates asymmetric odorant receptor expression in C. elegans. Cell, 1999. 99(4): p. 387–98. Roayaie, K., et al., The G alpha protein ODR-3 mediates olfactory and nociceptive function and controls cilium morphogenesis in C. elegans olfactory neurons. Neuron, 1998. 20(1): p. 55–67. Jansen, G., et al., The complete family of genes encoding G proteins of Caenorhabditis elegans. Nat Genet, 1999. 21(4): p. 414–9. L’Etoile, N.D. and C.I. Bargmann, Olfaction and odor discrimination are mediated by the C. elegans guanylyl cyclase ODR-1. Neuron, 2000. 25(3): p. 575–86. Bargmann, C.I. and H.R. Horvitz, Chemosensory neurons with overlapping functions direct chemotaxis to multiple chemicals in C. elegans. Neuron, 1991. 7(5): p. 729–42. Birnby, D.A., et al., A transmembrane guanylyl cyclase (DAF-11) and Hsp90 (DAF-21) regulate a common set of chemosensory behaviors in caenorhabditis elegans. Genetics, 2000. 155(1): p. 85–104. L’Etoile, N.D., et al., The cyclic GMP-dependent protein kinase EGL-4 regulates olfactory adaptation in C. elegans. Neuron, 2002. 36(6): p. 1079–89.

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

Yu, S., et al., Guanylyl cyclase expression in specific sensory neurons: a new family of chemosensory receptors. Proc Natl Acad Sci U S A, 1997. 94(7): p. 3384–7. Goodman, M.B., et al., Active currents regulate sensitivity and dynamic range in C. elegans neurons. Neuron, 1998. 20(4): p. 763–72. Kameyama, K., et al., Activation by G protein beta gamma subunits of beta-adrenergic and muscarinic receptor kinase. J Biol Chem, 1993. 268(11): p. 7753–8. Kaplan, J.M. and H.R. Horvitz, A dual mechanosensory and chemosensory neuron in Caenorhabditis elegans. Proc Natl Acad Sci U S A, 1993. 90(6): p. 2227–31. Hart, A.C., S. Sims, and J.M. Kaplan, Synaptic code for sensory modalities revealed by C. elegans GLR-1 glutamate receptor. Nature, 1995. 378(6552): p. 82–5. Maricq, A.V., et al., Mechanosensory signalling in C. elegans mediated by the GLR-1 glutamate receptor. Nature, 1995. 378(6552): p. 78–81. Osborne, K.A., et al., Natural behavior polymorphism due to a cGMP-dependent protein kinase of Drosophila. Science, 1997. 277(5327): p. 834–6. Coates, J.C. and M. de Bono, Antagonistic pathways in neurons exposed to body fluid regulate social feeding in Caenorhabditis elegans. Nature, 2002. 419(6910): p. 925–9. Combet, C., et al., NPS@: network protein sequence analysis. Trends Biochem Sci, 2000. 25(3): p. 147–50. Flynn, G.E. and W.N. Zagotta, A cysteine scan of the inner vestibule of cyclic nucleotide-gated channels reveals architecture and rearrangement of the pore. J Gen Physiol, 2003. 121(6): p. 563–82. Finn, J.T., M.E. Grunwald, and K.W. Yau, Cyclic nucleotide-gated ion channels: an extended family with diverse functions. Annu Rev Physiol, 1996. 58: p. 395–426. Zagotta, W.N. and S.A. Siegelbaum, Structure and function of cyclic nucleotide-gated channels. Annu Rev Neurosci, 1996. 19: p. 235–63. Biel, M., X. Zong, and F. Hofmann, Cyclic nucleotide gated channels. Adv Second Messenger Phosphoprotein Res, 1999. 33: p. 231–50. Kaupp, U.B., Family of cyclic nucleotide gated ion channels. Curr Opin Neurobiol, 1995. 5(4): p. 434–42. Komatsu, H., et al., Functional reconstitution of a heteromeric cyclic nucleotide-gated channel of Caenorhabditis elegans in cultured cells. Brain Res, 1999. 821(1): p. 160–8.

4.6 Channel Regulation 68

69

70

71

Varnum, M.D., K.D. Black, and W.N. Zagotta, Molecular mechanism for ligand discrimination of cyclic nucleotide-gated channels. Neuron, 1995. 15(3): p. 619–25. Nickell, W.T., et al., Single ionic channels of two Caenorhabditis elegans chemosensory neurons in native membrane. J Membr Biol, 2002. 189(1): p. 55–66. Zhong, H., et al., The heteromeric cyclic nucleotide-gated channel adopts a 3A:1B stoichiometry. Nature, 2002. 420(6912): p. 193–8. Bradley, J., D. Reuter, and S. Frings, Facilitation of calmodulin-mediated odor adaptation by cAMP-gated channel subunits. Science, 2001. 294(5549): p. 2176–8.

72

73

74

75

Chen, T.Y. and K.W. Yau, Direct modulation by Ca(2+)-calmodulin of cyclic nucleotide-activated channel of rat olfactory receptor neurons. Nature, 1994. 368(6471): p. 545–8. Kurahashi, T. and A. Menini, Mechanism of odorant adaptation in the olfactory receptor cell. Nature, 1997. 385(6618): p. 725–9. Leinders-Zufall, T., M. Ma, and F. Zufall, Impaired odor adaptation in olfactory receptor neurons after inhibition of Ca2+/calmodulin kinase II. J Neurosci, 1999. 19(14): p. RC19. Perkins, L. A. et al., Mutant sensory cilia in the nemat ode Caenorhabditis elegans. Dev Piol, 1986. 117 (2): p. 465–87.

97

99

5

Vertebrate Olfactory Signal Transduction and the Interplay of Excitatory Anionic and Cationic Currents Johannes Reisert and Jonathan Bradley

Abstract

Over the past 40 years much data have been generated from studies on olfaction, starting with experiments using electrophysiology and, over the last decade, molecular biology. From these works a picture is beginning to emerge as to how odorant binding to olfactory receptor neurons (ORNs) is transduced into an electrical signal. Volatile chemicals (on the order of 300 Da or less) act as stimulants in olfaction by recognition and binding to receptor proteins expressed on the surface of ORNs. ORNs are bipolar cells that extend single dendrites to an epithelial border from which cilia protrude into a nasal mucus, covering the olfactory epithelium in the nasal cavity. Therefore, unlike rods and cones in the eye, or hair cells in the ear, ORNs are in direct contact with the external environment. Olfactory signal transduction and the generation of the depolarizing receptor current occur in the cilia, and hence are a function of the unique extraciliary environment of the nasal mucus. ORNs employ a mechanism that allows them to cope with this particular ionic composition and still provide reliable electrical readout of odor binding. That is, the receptor potential combines ion conduction of two distinct channels: a cAMP-gated Ca2+ channel and a Ca2+-activated Cl– channel, both of which contribute to excitation. Concentrating on vertebrates, we will introduce the olfactory system and then describe the characteristics of the two channels in question and their interaction to generate the odor-induced receptor current.

5.1

Introduction 5.1.1

Tissue

In vertebrates the sensory tissue, or olfactory epithelium (OE), is located in the nasal cavity. In mammals the inhaled air first passes over a respiratory epithelium (RE) where it is warmed and moistened. The lateral walls of the posterior portion of the Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

100

5 Vertebrate Olfactory Signal Transduction and the Interplay of Excitatory Anionic and Cationic Currents

nasal cavity

basal cell

Fig. 5.1 Schematic diagram of the olfactory epithelium. Illustration of the three basic cell types in the olfactory epithelium. Basal cells are located along the mucus basement membrane. Soma of the olfactory sensory neurons are located in the central portion of the epithelium and have a bipolar morphology with a single dendrite extending to the lumen and an axon supporting extending to the olfactory bulb. The tip of the dendrite cell forms a knob from which emanate the olfactory cilia. Supporting cells constitute the third type of cell in the epithelium

olfactory receptor cell

cavity are elaborated into a series of complex folds called turbinal extensions (Fig. 6.1A). These have a surface area of a few square centimeters in man and more than 100 cm2 in dogs. The OE is located on these ecto- and endoturbinals and is divided into two histologically distinct layers isolated by a basement membrane. The deeper layer, or lamina propria, is glandular and contains vascular and connective tissue. The Bowman’s glands have ducts that traverse the basement membrane and neuroepithelium to the nasal lumen and secrete components of a protective mucus layer. These include secretory forms of the immunoglobulin types A and M, bacteriostatic and bactericidal proteins including lactoferrin and lysozyme, and detoxifying enzymes [1]. The lamina propria layer also contains the nerve fascicles of the receptor neurons, located in the upper neuroepithelial layer. The neuroepithelium is 100–200 lm thick and is pseudostratified, consisting of three cell types with the nuclei of each cell type in fairly discrete layers (see Fig. 5.1).

5.1.2

Olfactory Receptor Neurons

The olfactory receptor neurons (ORNs) constitute the predominant cell type in the OE and have nuclei scattered in a band six or eight nuclei wide through the central portion of the epithelium. There are about 106 ORNs in the rat, and they constitute 75–80 % of all the epithelial cells [2]. ORNs are bipolar with unmyelinated axons [3] that group into fascicles in the lamina propria surrounded by a single Schwann cell and project directly to the CNS, signaling in the olfactory bulb. A single dendrite terminates at the nasal lumen and ends in a structure referred to as a dendritic knob. Emanating from each dendritic knob and extending into the lumen of the nasal cavity are 5–20 immotile cilia 50–200 lm in length (in mammals) that taper from 2–3 lm at their base to about 0.1 lm at their tip. ORNs are a unique neuronal cell type in that they continually turn over on average every 30–60 days and are replenished from a population of mitotic basal cells [4]. Although the ORNs are derived from the basal cells and differenti-

5.1 Introduction

ate into mature sensory neurons, several features classify them as “juvenile” in comparison to neurons elsewhere. First, the time course of death of the ORNs in response to axotomy is rapid in comparison to other neurons [5]. Neurons in adult animals typically survive axotomy or die over the course of days or weeks, even when the site of axotomy is relatively close to the soma [6]. In contrast, ORNs, like neuronal populations in embryos and neonates [7], respond to axotomy with profound and rapid death. Immunohistochemically, ORNs also resemble “juvenile” neurons elsewhere in that they retain a pattern of intermediate filament- and microtubule-associated protein expression, which is characteristic of immature neurons [8–10]. Also resembling “juvenile” neurons elsewhere, ORNs retain a high concentration of intracellular chloride [11], an important feature of olfactory signal transduction (see below).

5.1.3

Sustentacular Cells

Sustentacular, or supporting cells, span the width of the epithelium and have their nuclei located most apically, forming a single layer. In the rat, sustentacular cells constitute 10-12 % of all the epithelial cells [2]. These cells act to physically protect and electrically isolate the receptor neurons, and also contribute secretions to the mucus [12]. Each sustentacular cell is surrounded by the dendrites and somata of 2–8 sensory neurons [4]. Sustentacular cells have a relatively hyperpolarized resting membrane potential (–80 to –120 mV) and a low membrane impedance (10 M) [13]. Injection of current into these cells fails to generate an action potential, and studies with injection of fluorescent dye indicate that they are not electrically coupled [12, 13]. In addition to physically protecting the underlying sensory neurons, sustentacular cells contain detoxification enzymes including an olfactory-specific isoform of cytochrome P450 (P-450olf1) (phase I biotransformation enzyme) and UDP-glucuronyl transferase (UGTolf) (phase II biotransformation enzyme) at levels that rival their concentration in the liver [14, 15]. In phase I biotransformation, cytochrome P-450 catalyzes the hydroxylation of the substrate, and in phase II UGT catalyzes transfer of glucuronic acid from UDP-glucuronate to hydroxyl groups, converting a hydrophobic molecule into a hydrophilic, membrane-impenetrable one that is readily excreted. Beyond their obvious role in the destruction of inhaled toxins, it has been suggested that these enzymes play a role in the metabolism of odorants. Hydroxyl-containing odorants are substrates for UGTolf, and UGTolf is two- to fivefold more active than the liver enzyme on odorants [14]. Moreover, glucuronidation abolishes the activity of odorants in an in vitro assay for olfactory signal transduction [14]. Although to date no direct role in olfaction has been demonstrated for these enzymes, these data perhaps suggest a role in signal termination by shortening the tissue residence time of odorants and facilitating their removal through chemical modification. Mucus by itself has been proposed to concentrate the odorants, which is favored by their air-mucus partition coefficients [16].

101

102


5.1.4

Basal Cells

The third basic cell type in the OE is the stem cells or basal cells located at the base of the olfactory epithelium along the basement membrane. In contrast to mature neurons, basal cells do not express the cytoplasmic olfactory marker protein (OMP) [17] but do express N-CAM (a general marker for mature CNS neurons) [18]. Two types of basal cells have been described, which are referred to as globose basal cells and horizontal basal cells. Globose and horizontal cells can be distinguished ultrastructurally and immunohistochemically. Globose cells, as the name implies, have rounded nuclei. These cells stain lightly with toluidine blue and are located just one cell layer above the basal lamina. Horizontal cells can be distinguished from globose cells because they have elongated nuclei, stain darkly with toluidine blue, lie adjacent to the basal lamina, and are also the only cells in the OE that express cytokeratin [19]. Evidence that the basal cells are a self-renewing source of receptor neurons comes from the observations that both globose and horizontal cells undergo mitosis and incorporate 3H-thymydine, with 90 % of the labeled cells ending up in the receptor cell zone [4]. Bulbectomy (removal of the olfactory bulb, which is the target of the receptor neurons) increases the production of neurons and specifically enhances mitotic activity of the globose cells [19]. These data suggest that the globose cells are the direct precursors of new neurons. Using a replication-incompetent retrovirus that expresses human placental alkaline phosphatase, the fate of cells from both types of basal cells was determined to high resolution [20]. The results show that the globose cells are the major source of new olfactory receptor neurons and that some progeny of the globose cells divide, or transiently amplify, in the globose cell zone of the OE before migrating up to the receptor cell zone and differentiating into mature receptor neurons. This is consistent with observations in vitro of an immediate neuronal precursor that is keratin- and N-CAM negative, migratory, and non-neurite bearing [21], characterized as a transient amplifying cell dependent on fibroblast growth factors [22].

5.2

Recording Odor-induced Electrical Activity 5.2.1

The Electroolfactogram

The electroolfactogram (EOG) was one of the early electrophysiological experiments done with the OE [23]. An EOG is essentially a field recording with an electrode placed on the mucus near the cilia during the application of odorants. It is thought to represent a summation of the receptor neuron generator potentials and shows a monophasic negative-voltage transient involving sodium (but is unaffected by tetrodotoxin) and a requirement for extracellular calcium [24]. An EOG cannot be recorded from respiratory epithelium and is abolished when the epithelium is lesioned with zinc sulfate or Triton X-100 to remove the cilia [25]. Killing the receptor neurons by transection of the

5.3 Odorant Responses of Single Isolated Olfactory Receptor Neurons

olfactory nerve also abolishes the EOG [26]. Recovery of the EOG is accompanied by repopulation of the epithelium with receptor neurons and re-growth of cilia. The EOG amplitude is dependent on the concentration of odorant applied, with larger amplitude for higher concentrations. However, when odorant is repeatedly applied, there is a reduction in the relative amplitude. Later, recording from isolated ORNs, this was shown to involve Ca2+ influx reflecting adaptation (see below). Analysis of EOG responses indicates that many regions of the epithelium can respond to a given odorant, although some areas are activated to a greater extent than others [27–31]. Thus, it seems likely that some functional heterogeneity exists across the OE. How does the biochemical process of an odorant interaction on the surface of the cilia change the membrane potential of the sensory neuron and get converted into an electrical signal sent to the brain? Several events came together to explain this. Electrophysiologically, it had been observed that cAMP played a role in olfaction [32]. In these early experiments, Minor and Sakina recorded EOGs from explants of frog OE that they could place in a tube or gutter and over which Ringer solution flowed. They were able to record slow (odorant-like) depolarizations of the ORNs when membrane-permeable dibutyryl derivatives of cAMP were introduced into the flow of the Ringer. In addition, they could see odorant-potentiating effects with phosphodiesterase inhibitors and odorant-depressing effects with phosphodiesterase activators. Minor and Sakina even proposed what we now call a “signal transduction pathway” for odorant detection [32]: “Cyclic 3’, 5’- AMP is considered to play the role of mediator in the mechanism of excitation of the olfactory receptor; during interaction between an odiferous substance and the receptor, adenylyl cyclase is activated and the concentration of cyclic 3’, 5’-AMP increases; this, in turn, causes depolarization of the receptor cell membrane.”

5.3

Odorant Responses of Single Isolated Olfactory Receptor Neurons

A more detailed study of the olfactory response was made possible with the use of metal- or KCl-filled electrodes [29, 33, 34]. Action potentials, fired by individual olfactory receptor cells, could now be recorded extracellularly. A comprehensive study of odor-induced action potential responses was performed by Getchell and Shepherd [30, 35]. They reported that salamander olfactory receptor cells generate spikes in response to odor presentations; furthermore, a higher stimulus concentration led to an increase in spike firing frequency. Similar observations were made by many others [30, 36–39]. The maximal odor-induced spike firing frequency ranges from 20 to 40 Hz. But while the frequency monotonically increases with higher odor concentrations, the number of fired spikes does not show such uniform changes. At high odor concentrations, the spike train is often shorter than at lower concentrations, and fewer spikes are fired in response to higher odor concentrations [33, 39–41]. At these high concentrations the amplitude of individual spikes in an odor-induced train progressively decreases with time, until the spikes are not detectable anymore. This phenomenon was observed as soon as odor-induced spike trains could be recorded [33, 36, 37, 42–44] and has been

103

104


suggested to arise from progressive inactivation of voltage-activated Na+ channels during prolonged depolarization [39]. Not only can the spike amplitude decline during odor presentation, but also in some cases spikes reemerge with progressively increasing spike amplitude [30, 40, 42, 45]. For a 10-s odor exposure, an initial high-frequency burst is often followed by spiking of lower frequency, with the spiking ceasing temporarily or completely during the stimulation [30, 33, 46]. For stimulation durations lasting over 30 s, repetitive bursts of spikes have been reported that are driven by underlying oscillations of the intracellular voltage [47]. Seventy percent of frog ORNs respond with an oscillatory response pattern, with bursts of action potentials appearing around every 6 s [45], and every 1 s in mouse ORNs [48]. In addition to excitatory responses, odor-induced inhibitory responses in the form of a reduction of the basal spike frequency were recorded by Getchell and Shepherd (see also [29, 35, 38, 49]). Resting potentials of olfactory receptor cells were shown to be in the range of –40 to –70 mV by recording the intracellular potential either with intracellular electrodes or by means of the whole-cell current-clamp configuration [42, 43, 50–53]. Vertebrate olfactory receptor cells depolarize upon stimulation in a graded manner with associated generation of action potentials [42, 51, 54–56]. Alternatively, odor-induced hyperpolarization has been reported for the toad and the mudpuppy, which causes the above-mentioned reduction in basal spike rate [57, 58]. ORNs are electronically very compact cells and have a high input resistance of several G [47, 52] Consequently, even a small depolarizing current can lead to a depolarization sufficiently large to generate action potentials [39, 47, 59–62]. Still, olfactory receptor cells are surprisingly quiet at rest, with a low basal spike rate of between 0.05 and 3 Hz [30, 36, 38, 42, 63, 64], an indication that the transduction mechanism itself is quiet when not stimulated. A more precise study of how the odor-induced receptor current depends on stimulus concentration was made possible in the late 1980s with the introduction of whole-cell patch-clamp techniques into olfactory research [51, 54, 65]. A detailed analysis has been done for olfactory receptor cells isolated from the tiger salamander [66] and exposed to odor for 1.2 s (see Fig. 5.2). When stimulated at low concentrations, the receptor cell responded with a transient inward receptor current, which terminated shortly after the end of the stimulation. An increase in stimulus concentration increased the response in a graded manner, and at high concentrations the receptor current outlasted the stimulation by several seconds. The amphibian dose-response relationship is narrow, and saturation of the receptor current occurs within a 10fold increase of stimulus concentration (see Fig. 5.2Ab and 5.2Bb). This is reflected by the average Hill coefficient of 3–4. The sigmoidal shape of the dose-response curves implies a high level of cooperativity in the response. Especially at low odor concentrations, olfactory receptor cells integrate over the stimulus presentation for up to 1.2 s; hence, longer stimulations lead to larger receptor currents [66, 67].

5.3 Odorant Responses of Single Isolated Olfactory Receptor Neurons

Fig. 5.2 Odor-induced responses in isolated salamander ORNs. Increasing concentration of acetophenone (A) or amyl acetate (B) were applied for 1.2 s as indicated by the solution monitor on top (S), and response families were recorded using the whole-cell voltage-clamp technique. The holding potential was –55 mV. The peak receptor current is plotted against the acetophenone (Ab) or amyl acetate (Bb) concentration and fitted with the Hill function (Hill coefficient n = 3, K1/2 = 17 lM and n = 4.2, K1/2 = 53 lM). Modified from [66] with permission from the Journal of Physiology, London

The simultaneous recording of receptor current and spike firing is possible using the suction pipette technique [68], where the cell body of an ORN is sucked into the tip of a recording pipette. The current recorded in this configuration has two components, the slow transduction current and the fast biphasic current spikes that represent action potentials. This technique was used to record odor-induced responses from mouse ORNs (Fig. 5.3). Compared with the responses of frog ORNs, the spike firing rate of mouse ORNs depends even more steeply on odor concentration, and the maximal spike firing rate is reached with only a threefold increase in odor concentration (Fig. 5.3C). Interestingly, this is an increase too small to saturate the receptor current, which still continues to rise even when spike firing has saturated (Fig. 5.3C) [48]. During longer odor exposures, the receptor current response is transient and begins to decline even in the presence of odor. At low to medium concentrations the receptor current can actually be reduced to values near zero [45]. For high concentrations, the receptor current increases quickly at the onset of stimulation, declines slowly thereafter, and remains present for the duration of the stimulus [45, 69, 70].

105

106


Fig. 5.3 The dose-response relation of a mouse olfactory receptor cell. (A) The suction pipette technique was used to record responses to the odor cineole (15) at the indicated concentration. (B) Rising phase of the receptor current during which action potentials are fired. (C) Dose dependence of the spike frequency, number of spikes fired, and the peak receptor current. Modified from [48] with permission of the Journal of Physiology, London

Olfactory receptor cells respond only when odor is delivered to the cilia [51, 71], and the response increases linearly with the length of the ciliary bundle exposed to the odor [68]. These findings demonstrate that the odor-sensing part of olfactory receptor cells is restricted to the cilia. A delay between odorant delivery and electrical response has been reported to be as short as 50 ms when measured with the suction pipette technique [68]. However, a range of 140 to 660 ms has been measured for different cells [41, 51, 65]. This time delay is consistent with the notion that the generation of the receptor current comprises several biochemical steps and that a second messenger system is involved in olfactory transduction.

5.4

Components of the Transduction Pathway

G-proteins are substrates for ADP-ribosylation. The adenylyl cyclase (AC) stimulating G-protein Gas is ADP-ribosylated by cholera toxin and the AC inhibiting G-protein Gai is ADP-ribosylated by pertussis toxin [72], as is the phospholipase C activating G-protein Gao. Pace and Lancet [73] were able to biochemically identify a Gas like

5.4 Components of the Transduction Pathway

Ca2+

Na+

R

Gα β γ

AC ATP cAMP CaMKII Ca2+

Ca2+/CaM

Ca2+ Cl-

Na+ Fig. 5.4 Schematic diagram of the olfactory signal transduction pathway. Odorants interact with a seven-helix receptor (R) activating a G-protein (G), which stimulates adenylyl cyclase (AC) to produce cAMP. cAMP opens the CNG channel, allowing Ca2+ influx. Ca2+ activates a Cl– channel, and Cl– efflux results in amplification of the initial depolarization. A Na+/Ca2+ exchanger uses the Na+ gradient to extrude Ca2+, helping to terminate signaling. Ca2+ also feeds back negatively to downregulate the CNG channel’s sensitivity to cAMP and the activity of AC by interaction with calmodulin (CaM) and CaM-kinaseII (CamKII), respectively

G-protein in OE cilia membranes 0.5 kDa larger than the liver Gas using cholera toxin and 32P labeled ADP-ribose. Pretreatment of the cilia membranes with cholera toxin decreased the odorant stimulation of AC in a dose dependent manner, establishing a relation between a Gas like G-protein and odorant stimulation of AC. Unlike the cholera toxin catalyzed ADP-ribosylated protein, the three proteins ADP-ribosylated by pertussis toxin were not OE specific. These data, and the previously described electrophysiological data, directed molecular studies to clone and characterize nearly all the members of a signal transduction pathway in which receptors of the seven helix class activate a Gas like G-protein which would stimulate an AC, increasing cAMP levels. Inturn the increase in cAMP concentration would then directly gate open a cyclic nucleotide-gated (CNG) channel, initiating depolarization of the ORN and the firing of action potentials. The receptor current later was shown to comprise a second inward conductance; the initial influx of Ca2+ through the CNG channel triggers a Cl– conductance, which amplifies the original predominantly Ca2+ driven depolarization. This pathway is diagrammed in Fig. 5.4.

107

108


5.5

Cloning of G Proteins Expressed in the OE 5.5.1

Gaolf

To clone cDNAs for G-proteins expressed in the OE, a library was screened with an oligonucleotide against a portion of an 18-amino-acid sequence common to all known G-proteins at the time. Clones for five different G-proteins were isolated. The majority of clones were Gas; the rest were Gao Gai1, Gai2, and the new species Gai3 [74]. In order to determine whether these cDNA clones represented mRNAs expressed in the ORNs, Northern blots were done with RNA from rats that had or did not have a bulbectomy. The results showed that none of the G-proteins had their signal reduced due to bulbectomy, which results in death of the ORNs by removal of their target, suggesting non-sensory cell expression of these genes in the OE [75]. A second screen done at low stringency was initiated using as a probe a mixture of the five G-proteins plus a degenerate sequence against a conserved region of the GTP-binding domain. One of five clones identified encoded a class that weakly hybridized to Gas. This new G-protein, named Gaolf, is 88 % identical to Gas at the amino acid level and shows robust expression in the OE. Gaolf can activate AC in S49 cyc– kin– cells, a cell line devoid of Gas, by incubation with GTP-c-S or AlF4–, potent activators of G-proteins. Not surprisingly, Gaolf is also activated as potently as Gas by cholera toxin. Gaolf can also couple to the b2 adrenergic receptor in transfected S49 cyc– kin– cells about as well as Gas [76]. Bulbectomy results in loss of Gaolf message, and peptide-specific antisera against Gaolf stain the OE cilia and axon bundles [77]. Ultrastructural/immunohistochemical electron microscopy studies using rapid freezing followed by rapid freeze-etching or freeze-substitution localize both Gaolf and Gas to the long, thin, distal portions of the cilia [78]. These data suggested that Gaolf and possibly Gas are the AC-stimulating Gproteins that mediate olfaction. The functional significance of Gaolf in olfactory transduction was convincingly demonstrated using mice that had disrupted expression of the protein. These animals lacked odorant-induced olfactory EOGs [79]. The exclusive expression of Gaolf in the OE, however, has been challenged with the report of Gaolf expression in the basal ganglia [80] and its implication in dopamine D1 receptor signaling through AC [81].

5.5.2

Adenylyl Cyclase

A wide variety of odorants were shown to stimulate cAMP production in the OE [82], and these increases were linearly correlated to the EOG amplitude evoked in response to odorant stimulation, an indication that signal transduction utilizes mainly cAMP as the second messenger [83]. In addition, adenylyl cyclase activity was shown to be highly enriched in olfactory cilia [84], which led to immunoidentification [85] and finally cloning of an adenylyl cyclase from the rat OE [86]. Bakalyar and Reed screened a

5.6 Odorant Receptors

rat olfactory cDNA library at low stringency with a mixture of coding sequence from the type I and type II adenylyl cyclase enzymes. Two overlapping clones were combined to generate a cDNA with an open reading frame that encodes 1144 amino acids. This enzyme is referred to as adenylyl cyclase type III (AC III). The deduced amino acid sequence appears topologically similar to the 1134-amino-acid type I enzyme. Both proteins have two hydrophobic regions: one near the NH2 terminus and the other between amino acid residues 600 and 850, each containing six potential transmembrane regions. Northern blot analysis indicated that AC III mRNA is enriched in the OE and that AC III mRNA disappeared after bulbectomy. Similarly, protein expression on the cilia reduced concomitantly with bulbectomy as detected by staining with a polyclonal antibody. Immunoelectron microscopy also localized ACIII to the cilia [87]. When expressed stably in HEK293 cells, AC III had almost no basal activity (4.70.1 pmol min–1 mg protein–1), as compared to control (4.00.3 pmol min–1 mg protein–1). In contrast, AC I has a relatively high basal activity (125.68.8 pmol min–1 mg protein–1). A similar level of basal activity is observed for AC II under these conditions. When the various AC isoforms were stimulated with forskolin in vitro, AC III proved to be most active of the three. The functional basal concentration of cAMP in ORNs has been shown to be just below that required for opening of the CNG channels [88]. This is probably an equilibrium set at rest in the cells between the activities of AC III and phosphodiesterases. It has been proposed that this can act to set the cAMP concentration at or very near threshold for activation of the transduction pathway [88]. During long odor exposures lasting several seconds, AC III activity can be downregulated by Ca2+/calmodulin kinase II (CaMK II)-dependent phosphorylation [89, 90]. Inhibition of CaMK II prevents adaptation of the odor-induced response [91]. A Ca2+/calmodulin-dependent phosphodiesterase is also present in olfactory cilia and is proposed to be involved in response termination [92, 93]. As with Gaolf, animals engineered to be deficient in expression of AC III lack odorant-induced olfactory EOGs and nurse poorly or fail olfaction-based behavioral tests [94].

5.6

Odorant Receptors

Single-cell spike recordings [27] demonstrate that response selectivity of ORNs is broad. Whole-cell recording experiments indicate that there is roughly a 50 % chance of finding a cell responsive to a mix of three odorants, and 50 % of those cells will respond to more than one of the three odorants [66]. However, although they appear broadly tuned, they do show specificity even when tested with high or low concentrations of an ineffective odorant. These characteristics, of being responsive to more than one stimulus, and having a relatively low sensitivity differ from those of photoreceptors and hormone receptors. Such distinctions likely reflect fundamental differences between these superficially similar systems, suggesting that the OE acts as a low-specificity detector array. In fact, a controversy has developed [69, 95, 96] over the suggestion that, as photoreceptors can signal detection of single photons [97], ORNs can

109

110


signal binding of single odorant molecules, i.e., odorants can induce quantal events (currents) in the ORNs [70]. Menini et al. recorded in the whole-cell configuration from salamander ORNs with low concentrations of odorant (0.5 lM cineole) for long periods of time (>30 min). Their claim that small current responses seen during these recordings correspond to quantal evoked-odor activation of the above-outlined transduction cascade is not substantiated by the data. The biggest problem with their interpretation of this data as representing quantal responses is that their type of analysis assumes a linear relationship between current and the number of quantal events. As outlined above, the transduction current is highly nonlinear, which critically undermines an estimation of quantal-event amplitudes. A very large gene family of closely related olfactory-specific, seven-helix receptors has been identified using a polymerase chain reaction (PCR) probe. The probe was generated by amplification of olfactory epithelial cDNA with oligonucleotides based on seven-helix receptor sequences from transmembrane regions (TM) II and VII from hormone receptors [98]. The original 10 complete cDNA sequences and 8 partial sequences cloned from rat have been expanded through genome-mining techniques such that the full complement of receptor genes from mouse (approximately 1300, of which 20–30 % are pseudogenes) [99] and human (approximately 900, of which 60–70 % are pseudogenes) [100, 101] are known and available on hybridization matrix chips (at least for mouse).

5.7

Cyclic Nucleotide-gated Channel in OE

More than 10 years after the EOG studies of Minor, a cGMP-gated conductance mediating phototransduction in rod and cone outer segments was discovered [102, 103]. These data led Nakamura and Gold [104] to do patch-clamp experiments on the cilia of toad olfactory neurons and look for a cAMP-gated conductance. In excised patches, their results directly demonstrated an outwardly rectifying, nonselective cation conductance with little sensitivity to voltage (in the absence of divalent cations) that could be gated by cAMP (concentration for half maximal activation, K1/2, of 2.4 lM). Unexpectedly, all the patches also showed a 1.7 higher sensitivity to cGMP (K1/21.6 lM). The Hill coefficients were 1.8 for cAMP and 1.7 for cGMP, indicating cooperativity of ligand binding. They proposed that an odorant-stimulated increase in cyclic nucleotide concentration has a direct activating effect on a cation conductance that initiates a depolarizing response to odorants. Nakamura and Gold added the caveat that this model does not apply to odorants that do not appear to stimulate adenylyl cyclase [82] or to those that stimulate polyphosphoinositide turnover [105]. Besides cAMP, IP3 and cGMP have been implicated as second messengers in olfactory transduction. For further discussion on these other pathways, we would like to point the interested reader towards two comprehensive reviews [95, 106]. Following the pioneering work of Nakamura and Gold, several whole-cell and excised patch recording studies were conducted [107–110], further characterizing the cAMP-gated channel in terms of its pharmacology and activation behavior. The chan-

5.7 Cyclic Nucleotide-gated Channel in OE

nel densities have been determined to be 8 per lm2 [111] or 70 per lm2 [112] in rat and frog, respectively, although much higher values have been reported (see [113, 114]). One of the most detailed studies [107] characterized the CNG channels in the dendritic knob, dendritic stalk, and soma of ORNs from rat and frog. The characteristics determined were rectification, activation by cyclic nucleotides, selectivity for monovalent cations, inhibition by cytosolic acidification, and inhibition by organic blockers. Summarizing their results, it was determined that the single-channel conductance, in the absence of divalent cations, was 8–35 pS for rat and frog (see also [111, 112, 114, 115]). There was a graded increase in patch current with the application of cAMP or cGMP to the cytosolic face of the membrane, saturating at 10–30 lM. In agreement with Nakamura and Gold [104], the K1/2 for cAMP was 2.5 lM (+50 mV) and 4 lM (–50 mV), with a Hill coefficient (n) of 1.8 for rat ORNs. For cGMP the values were (rat) K1/2 1.0 lM (+50 mV), 1.8 lM (–50 mV), n 1.3 (Fig. 5.5). There was weak outward rectification that was strongly dependent on cyclic nucleotide concentration and influenced by membrane voltage. There was no rundown of the current, and the conductance remained constant over the course of 30 min. The selectivity for monovalent cations was determined by applying voltage ramps in the presence or absence of saturating concentrations of cAMP (40 lM) with an extracellular concentration of 150 mM Na+ in the pipette and successive exposure of the intracellular membrane to similar concentrations of Na+, K+, Li+, Rb+, or Cs+. The different reversal potentials revealed by the I-V relationships under each biionic condition yielded the following sequence of permeation ratios relative to Na+: Na+(1):K+(0.81):Li+(0.74):Rb+(0.74):Cs+(0.52). The values from rat channels differ slightly form those measured for newt: Na+(1): Li+(0.93): K+(0.93): Rb+(0.91): Cs+(0.72) [116]. The fact that the current carried by Li+ is smaller than the Na+ current, despite the equilibrium selectivity for Li+, is an anomalous permeation property that has also been observed for the cGMP-gated channel in rod photoreceptors [117]. Reducing the pH of the solution on the cytosolic side of the inside-out patch from 7.0 to 5.0 inhibited the cAMP-induced current by 60 % and completely blocked it at pH 4 in 18 s (solution change 70 % of the fractional current (Pf) being carried by Ca2+ at (Ca2+)o = 3 mM. Thus, much of the CNG channel’s contribution to the receptor current is the divalent cation Ca2+ and not monovalent Na+ flux. The reduction of Na+ flux through the CNG channel also prevents Na+ from accumulating in the small ciliary space, which is important for maintaining the Na+ gradient across the cilia [122]. When the effects of several organic blockers (amiloride, the phenylalkylamine D600 [methoxyverapamil], and the benzothiazepine diltiazem) were assessed, it was found that all could block to some degree from the inside. Some data suggest that amiloride can also block from the outside [47]. There is a question about whether this effect might actually be due to amiloride getting to the inside via the basolateral membrane of the cell, based on the long latency of effect. Similarly, data on block of the retinal rod and cone CNG channels by extracellular and intracellular l-cis-diltiazem are consistent with a single binding site on the cytoplasmic side of the channel [123]. There is no data on external block of the olfactory CNG channel by phenylalkylamines or benzothiazepines. In excised patches from rat or frog ORNs, the compounds showed a voltage-dependent block; there was a 10–12 higher sensitivity at more positive membrane potentials, presumably due to the positively charged compound being driven into the channel. In patches from frog ORNs, the Ki for amiloride decreased from 400 lM at –100 mV to 17 lM at +100 mV. The Ki at +60 mV for amiloride on the rat channels was 71 lM. This is two orders of magnitude higher than that of Na+ epithelial channels [124] (Ki 0.1–0.5 lM) but is similar to Na+/H+ exchangers [125] and T-type Ca2+ channels [126] (3–80 lM). On the other hand, the olfactory CNG channels are more sensitive to amiloride than are other Na+ transporters. For instance, the Na+/Ca2+ exchanger Ki is 1 mM [127], Na/K-ATPase Ki is 3 mM [128], Na-glucose transporter Ki is 2 mM [129], and voltage-sensitive Na+ channels Ki is 0.6 mM [129]. In parallel with their similar sensitivity to amiloride, both the CNG channel [130] and the T-type Ca2+ channel [136] have an even higher sensitivity to the amiloride derivative 3’-4’-dichlorobenzamil, suggesting a possible relationship between these channels. D600 and diltiazem, other Ca2+ channel blockers, also block the olfactory CNG channels in a voltage-dependent manner but have Ki values up to two orders of magnitude higher than on Ca2+ channels. Frings et al. [107] determined Kis for D600 of 200 lM at –20 mV and 12 lM at +100 mV for the CNG channels from frog ORNs. The ability of diltiazem to block the CNG channels showed stereospecificity, with the l-cis enantiomer being more potent (Ki 70 lM) than d-cis diltiazem (Ki 200 lM) at +50 mV. This same stereospecificity is also true for the photoreceptor CNG channel [131]. Ca2+channels also show stereospecific diltiazem block, but the d-cis enantiomer is more potent than the l-cis one. The diltiazem stereoselectivity implies that this block of the CNG channels is not “nonspecific” but rather involves

5.7 Cyclic Nucleotide-gated Channel in OE

Fig. 5.5 Activation by cyclic nucleotides of native rat CNG channels. I-V curves from rat knob membrane patches with different concentrations of (top left) cAMP or (botton left) cGMP, with symmetrical divalent-free solutions. (top right) dose-response relationships of activation by cyclic nucleotide derived from I-V curves. Modified from [107] with permission of the Rockefeller University Press

interaction with a binding site on the cytoplasmic side of the channel. For photoreceptor CNG channels, diltiazem binding and interaction with cGMP appear to involve separate domains; when purified photoreceptor channels incorporated into artificial membranes are digested with trypsin (producing a 63-kDa product), they retain activation by cGMP but lose sensitivity to l-cis diltiazem [132].

113

114


5.8

Cloning of a CNG Channel Expressed in the OE

Cloning of the rat olfactory CNG channel (rOCNC1, CNG2, and CNGA2) [133, 134] was facilitated by the earlier cloning of the bovine retinal rod CNG channel (bRCNC1, CNG1, and CNGA1), which followed a biochemical approach of purification and amino acid sequencing of protein from bovine retina [135] (a confusing use of pet names for these subunits has been clarified to an accepted nomenclature [136]). To isolate clones of the olfactory CNG channel, Dhallan et al. screened an olfactory cDNA library at low stringency with the retinal bCNGA1. The resultant rat clone, rCNGA2, was shown to be a nonspecific cation channel gated by cAMP or cGMP. cDNAs from bovine and catfish OE encoding a homologous channel with similar function were also cloned [134, 137]. Topologically, the channel resembles the pore-forming subunit of voltage-gated potassium channels, with six transmembrane domains, a signature pore-forming sequence, and an 80- to 100-residue intracellular carboxy terminal domain resembling the cyclic nucleotide-binding domain of the regulatory subunit of protein kinase A and the catabolite activator protein (CAP) of bacteria (for a more detailed presentation of CNG channel structure function analysis, see [138]). Although CNGA2 formed functional CNG channels when expressed heterologously, there was a discrepancy in the rectification properties and sensitivity to cAMP relative to the native channel. Subsequent cloning and expression analysis demonstrated the existence of a “modulatory” subunit of the rat olfactory CNG channel, rCNGA4 [139, 140] (see alignment in Fig. 4.4A). Modulatory is a functional label because when expressed alone, CNGA4 does not form functional CNG channels (but see [141]). Heterologous expression of CNGA2 together with CNGA4 produced heteromeric channels more similar in cAMP sensitivity to native channels but still about threefold less sensitive. Sometime later, yet a third subunit, rCNGB1b [142, 143], was definitively characterized to be a component of the native channel [142]. Heterologous expression of rCBGB1b together with rCNGA2 and rCNGA4 further increased cAMP sensitivity, essentially mimicking that of the native channel. In addition rectification ratios, the presence of a subconductance state, and co-localization of expression in situ all argue strongly that the native channel is a heteromeric complex composed of a combination of CNGA2, CNGA4, and CNGB1b [124] (see Fig. 5.6). Unclear is the stoichiometric relationship of the subunits in the supposed tetrameric channel, although it has been suggested that the stoichiometry is (CNGA2)2-(CNGA4)(CNGB1b) [144].

5.9 Negative Feedback by Ca on the CNG Channel Fig. 5.6 Ligand sensitivity of native and heterologously expressed olfactory CNG channels. Dose-response relations fit to the Hill equation for the activation of macroscopic currents by cAMP at +40 mV. Reproduced from [142] with permission of the Society of Neuroscience

5.9

Negative Feedback by Ca2+ on the CNG Channel

Ca2+ not only is a component of the receptor current but also acts to terminate the response and control adaptation. An odor-induced receptor current, which is transient and adapts in Ringer, fails to adapt when Ca2+ influx through the CNG channels is minimized by removing external Ca2+ [145, 146] or when internal Ca2+ buffering is increased using BAPTA-AM [91]. Olfactory cilia contain the Ca2+-binding protein calmodulin at a concentration of 1 lM [147]. In retinal rod photoreceptors, it has been shown that Ca2+ bound to calmodulin (Ca-CaM) reduces the cGMP sensitivity of the native rod CNG channel from 19 lM to 32 lM, effectively reducing the cation influx by two- to sixfold [148]. Therefore, when light stimulates hydrolysis of cGMP and the channel closes, Ca2+ influx is reduced, while the Na+/Ca2+/K+ co-transporter continues to extrude Ca2+ and internal Ca2+ drops. This relieves the inhibitory binding of Ca-CaM and can promote recovery by stimulating the channels to rebind cGMP. The prominence of this feedback in phototransduction, however, is not considered to be major [149]. In olfaction, feedback inhibition by Ca-CaM on the CNG channel is more important. It was shown that Ca2+ could act directly on the native olfactory CNG channel in excised patches to reduce the open probability, with a threshold for inhibition

115

116


Fig. 5.7 Effect of Ca-CaM on activation by cyclic nucleotides of native rat CNG channels. I-V curves from rat knob membrane patches with different concentrations of cAMP in symmetrical divalent-free solution in the (A) absence or (B) presence of Ca-CaM. (C) Doseresponse relationships (fit to the Hill equation) of activation by cyclic nucleotide derived from I-V curves. From [205] with permission

around 100 nM Ca2+ [146]. Later, using a two-pulse odor protocol or with flash photolysis of “caged cAMP” (the latter generates a receptor current directly by gating the CNG channel without activation of the G-protein-coupled cascade), Kurahashi and Menini [67] showed that for short (100 ms) activations, adaptation can be explained solely by negative feedback by Ca2+ on the CNG channel. Thus, the odorant-induced intracellular elevation of Ca2+ is thought to promote adaptation because Ca-CaM can directly reduce the sensitivity of the native CNG channel for cAMP (Fig. 5.7) [150–152]. The principal subunit CNGA2 has been extensively characterized as a binding target of Ca-CaM. Expressed homomeric CNGA2 channels are inhibited by cytoplasmic application of Ca-CaM [150, 153–156] (see [157] for review) and the rate of channel inhibition by Ca-CaM is limited by the rate of Ca-CaM binding [151]. Co-expression of the modulatory subunits CNGA4 and CNGB1b together with CNGA2, however, increases the rate of inhibition by Ca-CaM by two orders of magnitude [151]. These data beg the question of how it is that CNGA2 should

5.9 Negative Feedback by Ca on the CNG Channel

be the binding target of Ca-CaM [150, 153–156] yet CNGA4 and CNGB1b control the rate of inhibition/binding of Ca-CaM [151]. The identified calmodulin-binding site on CNGA2 [153] is a classic basic amphiphilic alpha helix (Baa motif) [158] regulating channel inhibition in a strictly Ca2+-dependent manner by binding only a complex of Ca-CaM [153–155]. Bradley et al. [159] tested the relevance of this site in mediating Ca-CaM inhibition of CNGA2-A4-B1b channels. Quite surprisingly, they found that this site has no role in this process (Fig. 5.8, top). In the absence of a role for CNGA2 in inhibition of native channels by Ca-CaM, the question then becomes how does calmodulin associate with the native channels, and by what mechanism does calmodulin facilitate Ca2+-dependent feedback inhibition? To answer these questions, CNGB1b was first examined. The gene cngb1 is expressed as a number of splice variants in retina, sperm, and olfactory epithelium (see [138] and references therein). Variant CNGB1a is expressed in retinal rod photoreceptors and contains two calmodulin-binding sites (CaM1 and CaM2). CaM1, termed an “unconventional” CaM-binding site, is located in the NH2-terminus of CNGB1a and confers a weak ability of Ca-CaM to alter cGMP sensitivity of native heteromeric rod CNGA1B1a channels [160, 161]. CaM2, a Baa motif, resides in the COOH-terminus of CNGB1a and has not been ascribed any function so far [160, 161]. In CNGB1b, the variant of cngb1 expressed in ORNs, a large NH2-terminal glutamic-acid-rich domain of CNGB1a, is replaced by 74 unique amino acids [142, 143]. Otherwise, CNGB1a and CNGB1b are identical. Accordingly, the CaM1 and CaM2 sites de-

Ca-CaM

Inorm

1

wt A2∆CaM-A4-B1b

0 0

5

10 time (s)

15

20

Ca-CaM A2-A4(L292E)-B1b A2-A4-B1b(L183E)

1

Inorm

Fig. 5.8 Analysis of heteromeric channel inhibition by Ca-CaM in excised inside-out patches. Top: deletion of the Baa-type Ca-CaM binding site in CNGA2 has no effect on heteromeric channel inhibition by Ca-CaM. Comparison of A2-A4-B1b channels (black trace) and A2(CaM)A4-B1b channels (gray trace). Bottom: only the IQ-type CaM binding sites in CNGA4 and CNGB1b are necessary for Ca-CaM inhibition of heteromeric channels. Deletion of the Baa Ca-CaM binding site, CaM2, in CNGB1b has no effect on heteromeric channel inhibition by Ca-CaM (black trace). Singlepoint mutations in the IQ-type CaM binding sites in CNGA4 or CNGB1b completely abolish heteromeric channel inhibition by CaCaM (dark gray and light gray traces, respectively)

0

A2-A4-B1b∆CaM2 0

20

40 time (s)

60

80

117

118


scribed for CNGB1a are also part of CNGB1b. Co-expression of CNGA2 and CNGA4, with a CNGB1b mutant lacking CaM2, did not alter the inhibitory effect of Ca-CaM (Fig. 5.8, bottom, black trace, A2-A4-B1b D CaM2). Between residues 183 and 193 of CNGB1b (LQELVKMEKER), CaM1 resembles a “generalized” IQ-type CaM-binding motif ({I,L,V}QxxxRxxxx{R,K}) [161], understood to mediate binding of Ca2+-free calmodulin [162, 163] (apocalmodulin). Mutation of a single key residue [164] (L183E) in this site specifically eliminated Ca-CaM inhibition of heteromeric CNG channels (Fig. 5.8, bottom, light gray trace, A2-A4-B1b(L183E)). Thus, the integrity of a binding site for apocalmodulin in CNGB1b is necessary for Ca-CaM inhibition of native channels. What about CNGA4, which had also been shown to be necessary for fast inhibition of native channel by Ca-CaM [151]? An IQ-type motif (LQHVNKRLERR), very similar in sequence to that of the CaM1 site in CNGB1a/b, located in CNGA4 between the sixth transmembrane domain (S6) and the cAMP-binding site in the cytoplasmic COOH-terminus was found. This region, termed the “C-linker,” is thought to be an integral part of the gating machinery of CNG channels (for review, see [138]). A single amino acid substitution (L292E) in this IQ site rendered heteromeric channels completely insensitive to inhibition by Ca-CaM, (Fig. 5.8, bottom, medium gray trace, A2-A4(L292E)-B1b). Heteromeric channels with mutated IQ motifs in both CNGA4 and CNGB1b, but with the CNGA2 Baa motif intact (CNGA2-A4(L292E)B1b(L183E)), also failed to respond at all to Ca-CaM, supporting the conclusion that CNGA2 has no function in channel inhibition [159]. In olfactory CNG channels, CNGA4 and CNGB1b serve to increase cAMP sensitivity, so that a few lM cAMP are enough to open the heteromeric channels [142, 143]. Calmodulin appears to function as a cAMP sensitivity switch for the channel, blocking the effects of either CNGA4 or CNGB1b, or both, when triggered by Ca2+.

5.10

The Olfactory Ca2+-activated Cl– Channel

From a historical perspective, olfactory transduction was seen as a system that worked analogously to vertebrate phototransduction: activation of a G protein-coupled receptor triggers an electrical response by generating a current through a CNG channel. This view was altered when Kleene and Gesteland [165] reported the presence of a Ca2+activated Cl– conductance in olfactory cilia. This finding provided an explanation for a long-standing observation that the olfactory response persists even in the absence of external Na+ [166–169]. Three questions were raised: what is the Ca2+ source to gate the Cl– channel, when is the conductance opened, and is the ensuing Cl– current excitatory or inhibitory? Apparently olfactory cilia do not contain Ca2+ stores because application of pharmacological agents that cause Ca2+ release from intracellular stores fail to generate ciliary Ca2+ transients [170]. Additionally, Ca2+ signaling in olfactory receptor cells is regulated separately in the cilia and in the cell body, which allows the generation of compartmentalized Ca2+ transients [171]. This eliminates the cell body as a source of Ca2+

5.11 Activation of the Cl– Conductance

in the cilia. The origin of ciliary Ca2+ has been shown to be exclusively the CNG channel. Odor stimulation increases the ciliary Ca2+ concentration by opening CNG channels [171, 172], and this primary Ca2+ signal generates an additional secondary Cl– current. This current is inward, and therefore excitatory, and carries 80 % of the odor-induced current in rodents [173] and 36–65 % in amphibians [174, 175]. Two functions for this unusual Cl– current have been proposed. First, it provides a nonlinear, low-noise amplification of the small primary Ca2+ current [173, 176]. Second, it reduces the dependency of the receptor current on external mucosal Na+, which might vary depending on the environment, a situation more relevant for amphibians [174] and fish [177] than for terrestrial mammals. Many of the characteristics of the odor-induced response therefore depend on the interplay between the CNG channel (the Ca2+ source) and the Cl– channel and on their regulation. Furthermore, since spike frequency is a function of the receptor current, the interaction of these two channels will affect action potential generation (see above). To support the excitatory Cl– current across the ciliary membrane, the Cl– distribution in the cilia and the mucus has to be such as to allow the Cl– reversal potential to be positive with respect to the resting potential of the cell. Using electrophysiological methods, the Cl– reversal potential has been shown to be around 0 mV in Xenopus [175] and –45 mV in mudpuppy [178]. Alternative approaches are to measure the Cl– distribution directly using fluorescent Cl– sensitive dyes, energy-dispersive xray microanalysis, or ion-selective microelectrodes. With these methods intracellular Cl– concentrations were found to be high and to range from 40 to 80 mM [179–181]. Furthermore, the Cl– concentration in the mucus is lower than in interstitial fluid (93 mM in toad mucus and 55 mM in rat mucus [179, 182]), thereby aiding Cl– efflux from the cilia. After being first discovered in frog cilia [165], the Ca2+-activated Cl– current in olfactory receptor cells has since been found in newt [174], Xenopus [175], salamander [183], trout [177], mouse [48], and rat [111, 173, 184]. In all species the Cl– current is excitatory, underscoring its importance in olfactory transduction. Interestingly, since the first (electrophysiological) description of the Cl– channel in 1991, its molecular identity has remained elusive.

5.11

Activation of the Cl– Conductance

In their instrumental description of the Ca2+-activated Cl– conductance, Kleene and Gesteland [165] sealed whole frog olfactory cilia into patch pipettes and obtained current-voltage relationships (Fig. 5.9, top). Increasing Ca2+ concentrations were applied to the cytoplasmic side of the cilia, which gave rise to increasing currents with a K1/2 for Ca2+ of 4.8 lM and a Hill coefficient of 2 (Fig. 5.9, bottom). On average the maximal Cl– current of a single cilium was –50 pA at –50 mV [176], with average ciliary lengths of 47 lm. Surprisingly, the maximal current is not correlated with ciliary length [185]. In symmetrical NaCl solution, the current reversed close to 0 mV and showed slight inward rectification, as is the case in the rat Cl– channel [111]. When internal Cl– was

119


+400

300 150 4,10 3 2 1 0

pA 100

mV +100

0-2 3 4 10 150 300

400

100 Normalized conductance

120

80 60 40 20 0 0.1

1

10

100

2+

[ Ca ] i , uM Fig. 5.9 Activation of the frog ciliary Cl– conductance by Ca2+. Top: a frog ORN cilium was sealed inside a patch pipette and excised from the dendritic knob. The cytoplasmic side was exposed to increasing Ca2+ concentrations, and current-voltage relations were recorded (Ca2+ concentration in lM next to each I-V relation). Bottom: Ca2+ dependence of the Cl– conductance. The conductance was measured between 0 and –50 mV and normalized to the value obtained at 300 lM Ca2+ (data of seven cilia). A Hill fit (solid line) yielded a Hill coefficient of 2 and a K1/2 of 4.8 lM. Reproduced from [165] with permission of the Society of Neuroscience

5.12 Single Channel Properties and Channel Densities

increasingly replaced by gluconate–, the reversal potential shifted to values that closely followed the predictions of the Nernst equation for a Cl– channel [165], confirming Cl– as the charge carrier. Replacing Na+ with choline+ or Tris+ did not alter the Cl– current magnitude [165], and Na+ hardly permeates the channel (PNa/PCl = 0.034 [184]). Membrane patches excised from dendritic knobs of rat olfactory receptor cells were used to study mammalian Cl– channels [111, 184]. When patches were held at –40 mV and rapidly exposed to Ca2+ for 10 s, Cl– currents quickly activated then slowly inactivated by 40 % over 10 s. The current magnitude depended steeply on the Ca2+ concentration, and fitting the Hill function to the peak current yielded a K1/2 of 2.2 lM and a Hill coefficient of 2.8 [111], values close to the ones found in the frog [165]. Nonneuronal Ca2+-activated Cl– channels display similar Ca2+ sensitivities (for review, see [186]), while a higher K1/2 of 26 lM with a low Hill coefficient (n) of 1 has previously been reported for the rat Cl– channel [196]. External Ca2+ does not seem to affect the Cl– current magnitude [111]. Activation by Ca2+ was voltage dependent, and at +40 mV a lower K1/2 (1.5lM, n = 3) [111] was observed, as has been the case for Ca2+-activated Cl– channels in, e.g., Xenopus oocytes [187, 188], rat parotid acinar cells [189], or smooth muscle cells [190]. Interestingly, sensitivity of the frog Cl– channel seems not be voltage dependent [165], nor has it been demonstrated previously in the rat [184]. Mg2+, which is found at high concentrations inside cells, cannot substitute for Ca2+ as the activating ion. Concentrations as high as 2 mM, or even 10 mM, did not elicit a Cl– current in patches excised from frog or rat ORNs [111, 186], which indicated that Mg2+ does not play a physiological role in gating the Cl– channel. Ba2+ activated the channel only poorly, while 1 mM Sr2+ evoked 80 % of the maximal Ca2+-activated current [111, 191].

5.12

Single Channel Properties and Channel Densities

Both the frog and the rat Cl– channel have a conductance too small to be observed directly as single channels in excised patches; therefore, noise analysis was used as the investigative tool. Larsson et al. [112] excised whole frog olfactory cilia and monitored the current noise associated with the macroscopic currents elicited by increasing Ca2+ concentrations applied to the cytoplasmic side of the cilium. Conscientiously, they took into consideration that, since they recorded from a long slender cilium, voltage clamp will be imperfect. They obtained a single-channel conductance of just 0.8 pS, with a maximal open-probability of 1 [112, 176]. With the known length and diameter of the cilium, the channel density was calculated to be around 80 channels per lm2. The single-channel conductance in the rat was determined by making use of the observation that in some excised patches the Cl– current declined slowly after removal of cytoplasmic Ca2+ application, probably due to slow solution exchange between the patch and the superfusate. Current and noise levels during the slow falling phase could therefore be determined, and noise analysis yielded a single-channel conductance of

121

122


1.3 pS with a maximal open-probability of nearly 1 [111]. This range of single-channel conductance from 0.8 to 1.3 pS for the frog and rat channels is similar to values reported in smooth muscle [190, 192–194]. The Cl– channel density in rat cilia was not determined by dividing by the plasma membrane area. Instead, the theory of buffered Ca2+ diffusion was used. Here, an excised patch was held in symmetrical cholineCl solution with 1 mM CaCl2 in the pipette solution. Application of cAMP to the cytoplasmic side of the patch yielded opening of the CNG channel and to Ca2+, but not to choline+, influx, since the latter is impermeant. The incoming Ca2+ will radially diffuse away, and the Ca2+ concentration will decrease with increasing distance from the source (the CNG channel). The speed of decline is determined by the cytoplasmic Ca2+ buffer in the superfusate of the patch and can be calculated using the theory of buffered Ca2+ diffusion [195, 196] (which describes the diffusion of Ca2+ in the presence of buffers). Depending on the distance between the CNG and a Cl– channel, both of which are present in the excised ciliary membrane, the Cl– channel will “see” a certain Ca2+ concentration and generate a Cl– current governed by its Ca2+ sensitivity. With a previously calculated Ca2+ doseresponse curve as a standard, the level of Ca2+ can be determined, which in turn allows calculation of the distance between the CNG and the Cl– channels. With this method the rat Cl– channel density was found to be 62 channels per lm2 [111], which is similar to the frog [112].

5.13

Regulation of Cl– Channel Activity

Some ion channels permeable to Cl– have their open probability and sensitivity to Ca2+ up- or downregulated by various kinases such as Ca2+/calmodulin-dependent protein kinase II (CaMK II) or PKC (for a comprehensive review, see [186]). Little is know, however, about such regulation of the olfactory Ca2+-activated Cl– channel. In excised membrane patches from rat ORNs the Cl– current reduces with time, or “runs down,” and has been attributed to the progressive loss of functional channels [111], which has been described in a variety of other tissues (for review, see [197]). The cause of the rundown is not understood, but it could be the loss of a soluble factor or dephosphorylation. On the other hand, rundown has not been observed for the frog Cl– channel recorded in excised cilia [165, 198] Ca2+/calmodulin plays an important role in regulating the sensitivity of the CNG channel (see 5.9). It is therefore interesting to investigate whether calmodulin affects the Cl– channel. Neither the frog nor the rat Cl– channel in an excised cilium patch showed any change in current level when calmodulin was added to the cytoplasmic solution [111, 199]. This indicates that cytosolic calmodulin does not regulate the channel directly but does not rule out that it might do so indirectly via, e.g., CaMK II, as is the case for AC III. Termination of the odor-induced response requires the shutoff of all steps in the transduction cascade to allow the Cl– channel, the final step, to close. Moreover, the Ca2+ that accumulates in the cilia and keeps the Cl– channel open has to be removed. A

5.14 Amplification of the CNG Current and Generation of the Cl– Current

Na+/Ca2+ exchanger is present in olfactory receptor cells [200, 201] and is responsible for ciliary Ca2+ removal and response termination [202]. In Na+-free solutions (i.e., Na+ replaced by Li+) Na+/Ca2+ exchange cannot function and the return to pre-stimulus Ca2+ levels is slowed. Consequently, a prolonged current is generated which is carried by Cl– [45, 202]. This demonstrates an interesting role reversal for Na+ relative to its “normal” electrophysiological function; Na+ is not essential for the generation of the receptor current but rather is necessary for the termination of the response. A Ca2+ATPase has also been described in a plasma membrane-rich fraction of olfactory epithelium from salmon [203].

5.14

Amplification of the CNG Current and Generation of the Cl– Current

The gating of Cl– channels is dependent on the intraciliary Ca2+ concentration, which in turn depends on the opening of CNG channels and intraciliary cAMP concentration. The interplay of these two channels in the ciliary membrane (in conjunction with the Na+/Ca2+ exchanger) greatly determines the shape of the odor-induced response and, ultimately, the cells’ firing pattern to the brain. Kleene [198] investigated how cAMP can lead to a Cl– current in ORNs. Application of cAMP to the cytoplasmic side of an excised cilium sealed inside a patch pipette led to

)

*

F)

Fig. 5.10 Origin of the Cl– current. (A) A frog cilium was sealed inside a patch pipette and excised. The pipette solution contained 1 mM Ca2+, while the bath solution contained 100 lM cAMP; both solutions had similar Cl– concentrations. The membrane potential was stepped to potentials from +80 mV to –60 mV in 20-mV increments. Only at negative potentials was a biphasic current observed. (B) The secondary current component could be suppressed by applying the Cl– channel blocker niflumic acid Modified from [198] with permission from Elsevier

IA?

123

124


the generation of a biphasic current at negative, but not at positive, holding potentials (Fig. 5.10A). The additional current component observed at negative potentials could be blocked by the Cl– channel blocker niflumic acid (Fig. 5.10B), required the presence of Ca2+ in the extracellular solution, and increased when the cytoplasmic Ca2+ buffering capacity was reduced [198]. Thus, at negative potentials opening of the CNG channel first caused a mixed Na+ and Ca2+ influx with the accumulating Ca2+ activating the secondary Cl– current. What is the relative contribution of these two current components to the overall current? In the experiment in Fig. 5.10, the two current phases were of equal magnitude [176], which indicates that the CNG current is amplified twofold by the Cl– channel. This is consistent with the observation that during the odor response around half of the receptor current is carried by Cl–. Similarly, in an excised frog cilium the saturating CNG and Cl– currents are of comparable magnitude in the presence of physiological concentrations of Ca2+ and Mg2+ [176]. While extracellular divalents do not seem to affect the Cl– channel [111], they greatly reduce the single-channel conductance of the CNG channel to 0.56 pS (compared to 15 pS in divalent free conditions; see 5.7), a value close to the one for the Cl– channel of 0.5 pS [176]. The observation that both single-channel conductances are small has an important consequence. Individual channels contribute only small currents to the overall response, a mechanism that reduces noise due to the opening of individual channels. But do the sequential activation of both channels and the twofold amplification of the primary CNG channel by the Cl– channel lead to an increase in noise? Careful individual evaluation of the noise levels during the two phases of inward current recorded from frog cilia (see Fig. 5.10) demonstrated that little noise is added due to the activation of Cl– channels [176]. The main source of noise seems to be the CNG channel (Fig. 5.10B), with its maximal open-probability reaching only 0.68 in contrast to the maximal open-probability of 1 for the Cl– channel. Taken together, extracellular Ca2+ blocks the CNG channel and hence reduces noise, but influx of Ca2+ amplifies the CNG current without a further increase in noise [176]. In rat ORNs Cl– channels carry around 80 % of the odor-induced current [173], and in the presence of the divalent cations Ca2+ and Mg2+, membrane patches excised from rat ORNs have a Cl– current 30 times larger than the CNG current [111], quite different from frog olfactory cilia, where both channels conduct similar amounts of current. While the Cl– channel densities in frog and rat are comparable, rat CNG channels seem to be present at a 10-fold lower density [111, 112]. With this lower density of CNG channels in the rat ciliary membrane, it is interesting to ask how many CNG channels are responsible for contributing the Ca2+ to activate an individual Cl– channel. Or, put differently, how far does Ca2+ travel once it enters the cilia through a single CNG channel before it is bound to a Ca2+ buffer? Within 20 nm, Ca2+ entering through a channel will diffuse freely before it can be bound to a buffer; therefore, Ca2+ buffers do not affect the Ca2+ concentrations near Ca2+ channels [204]. With this precedent, the magnitude of the Ca2+-activated Cl– current should not be affected by Ca2+ buffers if the Cl- channels are within 20 nm of the Ca2+-permeable CNG channels. Membrane patches excised from the dendritic knob of rat ORNs were used to investigate this question. A patch was held in symmetrical

5.14 Amplification of the CNG Current and Generation of the Cl– Current

125

A CholCl 1mM HEDTA

CholCl 1mM HEDTA

Test solution

Current (pA)

0

-10

-20

-30

0

2

4

6

8

10

12

Time (s)

B

1

2 3 45

6

r1

C % contribution to [Ca2+]local free

Fig. 5.11 Interaction between the CNG and the Cl– channel. (A) A patch excised from a rat dendritic knob was held in symmetrical cholineCl solutions with 1 mM Ca2+ added to the pipette solution. Application of 67 lM Ca2+ to the cytoplasmic side activated a saturated Cl– current (lower trace), while 100 lM cAMP in 1 mM HEDTA (middle trace) or 0.2 mM HEDTA (upper trace) generated currents of smaller magnitude. This current could be suppressed by niflumic acid, identifying it as a Cl– current (not shown). (B) Model for the arrangement of Cl– (squares) and CNG (circles) channels in the ciliary membrane. CNG channels at increasing distances (r1–r6) will contribute Ca2+ to the Cl– channel in the center. (C) Level of Ca2+ contribution from CNG channels positioned in circles 1–6. Modified from [111] with permission from the Rockefeller University Press

100

2

1

3

4

5

circle 6

8

9

90

1 mM HEDTA 80 70 60

0.2 mM HEDTA

50 1

2

3

4

5

6

7

10 11

number of channels contributing to [Ca2+]local free

126


choline Cl solution with the inclusion of 1 mM Ca2+ in the pipette. Under these ionic conditions, any observed current may originate only from flux of Cl– or Ca2+. In Fig. 5.11A, Cl– currents were evoked by applying either a saturating concentration of Ca2+ (lower trace) or cAMP in 0.2 mM (middle trace) or 1 mM (upper trace) of the Ca2+ buffer HEDTA. With 1 mM of HEDTA, a smaller Cl– current was observed than with 0.2 mM of HEDTA; thus, the increased Ca2+ buffer concentration can reduce the effective Ca2+ concentration at the Cl– channel once it enters through the CNG channel. From this observation that Cl– currents with different magnitudes were observed at two Ca2+ buffer concentrations, it was concluded that the CNG and the Cl– channel do not form transduction complexes. Therefore, instead of a configuration in which a CNG and a Cl– channel are very close to each other, the known ratio of Cl– to CNG channels of 4:1 (based on Cl– channel rundown; see 5.13) was used to construct an equidistant matrix for the distribution of Cl– and CNG channels in the patch membrane (Fig. 5.11B) [111]. The theory of buffered Ca2+ diffusion [195, 196] yielded two parameters: (1) the channel densities of the CNG and the Cl– channel and (2) the Ca2+ diffusion coefficient of 90 lm2 s–1 [111] for the cytosolic side of the patch. Depending on the internal Ca2+ buffer concentration, Ca2+ can diffuse considerable distances, and up to 50 % of the Ca2+ at a given Cl– channel originates not from the nearest CNG channel but from channels farther way (Fig. 5.11C). Thus, a Cl– channel sums the Ca2+ that enters through multiple CNG channels, a mechanism that can reduce noise originating from fluctuations of individual CNG channels, further enforcing the notion that these channels form a functional network with an amplification component. On the other hand, the number of CNG channels contributing to the Ca2+ concentration at a given Cl– channel is limited to around 10, which indicates that within a long, slender cilia, the spread of a Ca2+ signal might be spatially limited. Interestingly, odor-induced localized increases in intraciliary Ca2+ have been reported [171].

5.15

Open Questions

With Ca2+ playing such a critical role in olfactory transduction, many questions remain to be addressed. For instance, while the Cl– channel samples Ca2+ from multiple CNG channels, does the Ca2+ that leads to desensitization of the CNG channel via Ca2+/CaM originate from multiple CNG channels or a single CNG channel? How does the additional depolarization due to the Cl– current affect the Ca2+ current through the CNG channel and, therefore, intracellular Ca2+ levels? On the other hand, an increased level of depolarization could reduce Na+/Ca2+ exchange, maintain Ca2+ at an increased level, and therefore prolong the odor response. ORNs convey their action potentials directly to mitral/tufted cells in the olfactory bulb. But how do the CNG and the Cl– channels contribute to spike generation? Is the small Ca2+ current through the CNG channel sufficiently large to trigger action potentials, and how does the Cl– channel affect spike firing, which is, after all, the way the nose communicates with the brain?

5.15 Open Questions

Although much is known about individual components of the olfactory signal transduction, we are only beginning to understand the intricacies of the complex interaction within the transduction cascade and how this shapes the odor-induced response. Acknowledgement We thank the Howard Hughes Medical Institute and King-Wai Yau for their support, and Stephan Frings for reading the manuscript.

References 1

2

3 4

5

6

7

8

9

Getchell, M.L., Getchell, T.V., Immunohistochemical localization of components of the immune barrier in the olfactory mucosae of salamanders and rats. Anat. Rec., 1991. 231: p. 358–374. Farbman, A.I., ed. Cellular interactions in the development of the vertebrate olfactory system. Molecular Neurobiology of the Olfactory System, ed. F.L. Margolis, Getchell, T.V. 1988, Plenum, N.Y. 319–332. Gasser, H.S., Olfactory nerve fibres. Journal of General Physiology, 1956. 39: p. 473–496. Graziadei, P.P.C., and Monti Graziadei, G.A., Neurogenesis and neuron regeneration in the olfactory system of mammals. I. Morphological aspects of differentiation and structural organization of the olfactory sensory neurons. J. Neurocytol., 1979. 8: p. 1–18. Masukawa, L.M., B. Hedlund, and G.M. Shepherd, Changes in the Electrical-Properties of Olfactory Epithelial-Cells in the Tiger Salamander After Olfactory Nerve Transection. Journal of Neuroscience, 1985. 5(1): p. 136–141. Berkelaar, M., Clarke, D.B., Wang, Y.C., Bray, G.M., Aguayo, A.J., Axotomy results in delayed death and apoptosis of retinal ganglion-cells in adult-rats. J. Neurosci., 1994. 14: p. 4368–4374. Snider, W.D., Elliott, J.L., Yan, Q., Axotomyinduced neuronal death during development. J. Neurobiol., 1992. 23: p. 1231–1246. Schwob, J.E., Farber, N.B., Gottlieb, D.I., Neurons of the olfactory epithelium in adult-rats contain vimentin. J. Neurosci., 1986. 6: p. 208–217. Ophir, D., Lancet, D., Expression of intermediate filaments and desmoplakin in vertebrate olfactory mucosa. Anat. Rec., 1988. 221: p. 754–760.

10

11

12

13

14

15

16

17

18

19

Viereck, C., Tucker, R.P., Matus, A., The adultrat olfactory system expresses microtubule-associated proteins found in the developing brain. J. Neurosci., 1989. 9: p. 3547–3557. Ben-Ari, Y., Excitatory actions of gaba during development: the nature of the nurture. Nat Rev Neurosci, 2002. 3(9): p. 728–39. Okano, M., Tagaki, S.F., Secreation and electrogenesis of the supporting cell in the olfactory epithelium. J. Physiol. Lond., 1974. 242: p. 353–370. Masukawa, L.M., Hedlund, B. , Shepherd, G.M., Electrophysiologicl properties of identified cells in the in vitro olfactory epithelium of the tiger salamander. J. Neurosci., 1985. 5: p. 128–135. Lazard, D., Zupko, K., Poria, Y., Nef, P., Lazarovits, J., Lazarovits, J., Horn, S. and M. Kehn, Lancet, D., Odorant signal termination by olfactory UDP-glucuronosyl transferase. Nature, 1991. 349: p. 790–793. Nef, P., Heldman, J., Lazard, D., Margalit, T., Jaye, M., Hanukoglu, I., Lancet, D., Olfactoryspecific cytochrome-p-450: cDNA cloning of a novel neuroepithelial enzyme possibly involved in chemoreception. J. Biol. Chem., 1989. 264: p. 6780–6785. Amoore, J.E., Buttery, R.G., Partition coefficients and comparative olfactomery. Chem. Sens. Flavor, 1978. 3: p. 57–71. Margolis, F.L., A brain protein unique to the olfactory bulb. Proc. Natl. Acad. Sci., 1972. 69: p. 1221–1224. Key, B., Akeson, R.A., Olfactory neurons express a unique glycosylated form of the neural adhesion molecule (N-CAM). J. Cell Biol., 1990. 110: p. 1729–1743. Levey, M.S., Chikaraishi, D.M., Kauer, J.S., Characterization of potential precursor populations in the mouse olfactory epithelium using immunocytochemistry and autoradiography. J. Neurosci., 1991. 11: p. 3556–3564.

127

128

5 Vertebrate Olfactory Signal Transduction and the Interplay of Excitatory Anionic and Cationic Currents 20

21

22

23

24

25

26

27

28

29

30

31

32

33

Caggiano, M., Kauer, J.S., Hunter, D.D., Globose basal cells are neuronal progenitors in the olfactory epithelium: a lineage analysis using a replication-incompetent retrovirus. Neuron, 1994. 13: p. 339–352. Gordon, M.K., Mumm, J.S., Davis, R.A., Holcomb, J.D., Calof, A.L., Dynamics of mash1 expression in-vitro and in-vivo suggest a non-stem cell site of mash1 action in the olfactory receptor neuron lineage. Mol. Cell Neurosci., 1995. 6: p. 363–379. Dehamer, M.K., Guevara, J.L., Hannon, K., Olwin, B.B., Calof, A.L., Genesis of olfactory receptor neurons in vitro : Regulation of progenitor cell divisions by fibroblast growth-factors. Neuron, 1994. 13: p. 1083–1097. Ottoson, D., Analysis of the electrical activity of the olfactory epithelium. Acta Physiol. Scand., 1956. 122: p. 1–83. Leveteau, J., et al., Role of divalent cations in EOG generation. Chemical Senses, 1989. 14(5): p. 611–620. Bronshtein, A.A. and A.V. Minor, Regeneration of olfactory flagella and restoration of the electroolfactogram following application of Triton X-100 to the olfactory muscosa of frogs. Tsitologiya, 1977. 19: p. 33–39. Simmons, P.A., Getchell, T.V., Neurogenesis in the olfactory epithelium: loss and recovery of transepithelial voltage transients following olfactory nerve section. J. Neurophysiol., 1981. 45: p. 516–528. Sicard, G. and A. Holley, Receptor cell responses to odorants: similarities and differences among odorants. Brain Res, 1984. 292(2): p. 283–96. Duchamp, A., Revial, M.F., Holley, A., MacLeod. P., Odor discrimination by frog olfactory receptors. Chem.Senses Flavor, 1974. 1: p. 213–233. Gesteland, R.C., Lettvin, J.Y., Pitts, W.H., Chemical transmission in the nose of the frog. J. Physiol. London., 1965. 181: p. 525–559. Getchell, T.V., Shepherd, G.M., Responses of olfactory receptor cells to step pulses of odor at different concentrations in the salamander. J. Physiol. London, 1978. 282: p. 521–540. Kauer, J.S. and D.G. Moulton, Responses of olfactory bulb neurones to odour stimulation of small nasal areas in the salamander. Journal of Physiology, 1974. 243(3): p. 717–737. Minor, A.V., Sakina, N.L., Role of cyclic adenosine-3’,5’ -monophosphate in olfactory reception. Neurofysiologiya, 1973. 5: p. 415–422. Shibuya, T. and S. Shibuya, Olfactory epithelium: unitary responses in the tortoise. Science, 1963. 140: p. 495–496.

34

35

36

37

38

39 40

41

42

43

44

45

46

47

48

Getchell, T.V., Analysis of unitary spikes recorded extracellularly from frog olfactory cells and axons. Journal of Physiology-London, 1973. 234: p. 533–551. Getchell, T.V. and G.M. Shepherd, Adaptive properties of olfactory receptors analysed with odour pulses of varying durations. Journal of Physiology-London, 1978. 282: p. 541–560. Mathews, D.F., Response pattern of single neurons in the tortoise olfactory epithelium and olfactory bulb. Journal of General Physiology, 1972. 60: p. 166–180. Baylin, F., Temporal pattern and selectivity in the unitary responses of olfactory receptors in the tiger salamander to odor stimulation. Journal of General Physiology, 1979. 74: p. 17–36. O’Connell, R.J. and M.M. Mozell, Quantitative stimulation of frog olfactory receptors. Journal of Neurophysiology, 1969. 32: p. 51–63. Trotier, D., Intensity coding in olfactory receptor cells. Seminars in Cell Biology, 1994. 5: p. 47–54. van Drongelen, W., Unitary responses of near threshold responses of receptor cells in the olfactory mucosa of the frog. Journal of PhysiologyLondon, 1978. 277: p. 423–435. Reisert, J. and H.R. Matthews, Adaptation of the odour-induced response in frog olfactory receptor cells. Journal of Physiology-London, 1999. 519(Sep): p. 801–813. Trotier, D. and P. MacLeod, Intracellular recordings from salamander olfactory receptor cells. Brain Research, 1983. 268(2): p. 225–237. Masukawa, L.M., B. Hedlund, and G.M. Shepherd, Electrophysiological Properties of Identified Cells in the Invitro Olfactory Epithelium of the Tiger Salamander. Journal of Neuroscience, 1985. 5(1): p. 128–135. Trotier, D., Physiology of transduction in olfaction and taste. Seminars in the Neurosciences, 1990. 2: p. 69–76. Reisert, J. and H.R. Matthews, Responses to prolonged odour stimulation in frog olfactory receptor cells. Journal of Physiology, 2001. 534(Pt 1): p. 179–191. Baylin, F. and D.G. Moulton, Adaptation and cross-adaptation to odor stimulation of olfactory receptors in the tiger salamander. Journal of General Physiology, 1979. 74: p. 37–55. Frings, S. and B. Lindemann, Odorant response of isolated olfactory receptor cells is blocked by amiloride. Journal of Membrane Biology, 1988. 105(3): p. 233–243. Reisert, J. and H.R. Matthews, Response properties of isolated mouse olfactory receptor cells. Journal of Physiology-London, 2001. 530: p. 113–122.

5.15 Open Questions 49

50

51

52

53

54

55

56

57

58

59

60

61

Morales, B., et al., Inhibitory K+ current activated by odorants in toad olfactory neurons. Proceedings of the Royal Society of London Series B-Biological Sciences, 1994. 257(1350): p. 235–242. Pun, R.Y.K. and R.C. Gesteland, Somatic Sodium Channels of Frog Olfactory Receptor Neurons Are Inactivated At Rest. Pflugers Archiv-European Journal of Physiology, 1991. 418(5): p. 504–511. Firestein, S., G.M. Shepherd, and F.S. Werblin, Time course of the membrane current underlying sensory transduction in salamander olfactory receptor neurons. Journal of Physiology-London, 1990. 430(V): p. 135–158. Trotier, D., A patch-clamp analysis of membrane currents in salamander olfactory receptor cells. Pflugers Archiv-European Journal of Physiology, 1986. 407(6): p. 589–595. Kawai, F., T. Kurahashi, and A. Kaneko, T-type Ca2+ channel lowers the threshold of spike generation in the newt olfactory receptor cell. Journal of General Physiology, 1996. 108(6): p. 525–535. Anderson, P.A.V. and B.W. Ache, Voltage and current-clamp recordings of the receptor potential in olfactory receptor cells In situ. Brain Research, 1985. 338(2): p. 273–280. Anderson, P.A.V. and K.A. Hamilton, Intracellular recordings from isolated salamander olfactory receptor neurons. Neuroscience, 1987. 21(1): p. 167–173. Getchell, T.V., Analysis of intracellular recordings from salamander olfactory epithelium. Brain Research, 1977. 123: p. 275–286. Dionne, V.E., Chemosensory responses in isolated olfactory receptor neurons from Necturus maculosus. Journal of General Physiology, 1992. 99(3): p. 415–433. Morales, B., P. Labarca, and J. Bacigalupo, A Ciliary K+ Conductance Sensitive to Charibdotoxin Underlies Inhibitory Responses in Toad Olfactory Receptor Neurons. FEBS Letters, 1995. 359(1): p. 41–44. Leinders-Zufall, T., G.M. Shepherd, and F. Zufall, Regulation of cyclic nucleotide-gated channels and membrane excitability in olfactory receptor cells by carbon monoxide. Journal of Neurophysiology, 1995. 74(4): p. 1498–1508. Maue, R.A. and V.E. Dionne, Patch-clamp studies of isolated mouse olfactory receptor neurons. Journal of General Physiology, 1987. 90(1): p. 95–125. Lynch, J.W. and P.H. Barry, Action potentials initiated by single channels opening in a small neuron (rat olfactory receptor). Biophysical Journal, 1989. 55(4): p. 755–768.

62

63

64

65

66

67

68

69

70

71

72 73

74

75

Madrid, R., et al., Tonic and phasic receptor neurons in the vertebrate olfactory epithelium. Biophysical Journal, 2003. 84(6): p. 4167–4181. Getchell, T.V., Unitary responses in frog olfactory epithelium to sterically related molecules at low concentrations. Journal of General Physiology, 1974. 64: p. 241–261. Frings, S., S. Benz, and B. Lindemann, Current recording from sensory cilia of olfactory receptor cells insitu. 2. Role of mucosal Na+, K+, and Ca2+ ions. Journal of General Physiology, 1991. 97(4): p. 725–747. Firestein, S. and F. Werblin, Odor-induced membrane currents in vertebrate olfactory receptor neurons. Science, 1989. 244(4900): p. 79–82. Firestein, S., C. Picco, and A. Menini, The relation between stimulus and response in olfactory receptor cells of the tiger salamander. Journal of Physiology-London, 1993. 468: p. 1–10. Kurahashi, T. and A. Menini, Mechanism of odorant adaptation in the olfactory receptor cell. Nature, 1997. 385(6618): p. 725–729. Lowe, G. and G.H. Gold, The spatial distributions of odorant sensitivity and odorant-induced currents in salamander olfactory receptor cells. Journal of Physiology-London, 1991. 442(OCT): p. 147–168. Lowe, G. and G.H. Gold, Olfactory transduction is intrinsically noisy. Proceedings of the National Academy of Sciences of the United States of America, 1995. 92(17): p. 7864–7868. Menini, A., C. Picco, and S. Firestein, Quantal-like current fluctuations induced by odorants in olfactory receptor cells. Nature, 1995. 373(6513): p. 435–437. Kurahashi, T., Activation by odorants of cationselective conductance in the olfactory receptor cell isolated from the newt. Journal of PhysiologyLondon, 1989. 419(DEC): p. 177–192. Gilman, A.G., G-proteins and duel control of adenylate cyclase. Cell, 1984. 36: p. 577–579. Pace, U., Lancet, D., Olfactory GTP-binding protein: Signal-transducing polypepide of vertebrate chemosensory neurons. Proc. Natl. Acad. Sci. USA, 1986. 83: p. 4947–4951. Jones, D.T., Reed, R.R., Molecular cloning of five G-bindin protein cDNA species form rat olfactory neuroepithelium. J. Biol. Chem., 1987. 262: p. 14241–14249. Jones, D.T., Barbosa, E., Reed, R.R., Expression of G-protein a subunits in rat olfactory neuroepithelium: Candidates for olfactory signal transduction. Cold Spring Harbor Symp. Quan. Bio., 1988. LIII: p. 349–353.

129

130

5 Vertebrate Olfactory Signal Transduction and the Interplay of Excitatory Anionic and Cationic Currents 76

77

78

79 80

81

82

83

84

85

86 87

88

89

Jones, D.T., Masters, S.B., Bourne, H.R., Reed, R.R., Biochemical-characterization of 3 stimulatory GTP binding proteins: The large and small forms of Gs and the olfactory-specific Gprotein, Golf. J. Biol. Chem., 1990. 265: p. 2671–2676. Jones, D.T., and Reed, R.R., G-olf: an olfactory neuron-specific G-protein involved in odorant signal transduction. Science, 1989. 244: p. 790–795. Menco, P.M.B., Ultrastructural aspects of olfactory transduction and perireceptor events. Seminars Cell Bio., 1994. 5: p. 11–24. Belluscio, L., et al., Mice deficient in G(olf) are anosmic. Neuron, 1998. 20(1): p. 69–81. Drinnan, S.L., Hope, B.T., Snutch, T.P., Vincent, S.R., Gaolf in the basal ganglia. Mol. Cell Neurosci., 1991. 2: p. 66–70. Herve, D., Levistrauss, M., Mareysemper, I., Verney, C., Tassin, J.P., Glowinski, J., Girault, J.A., Golf and Gs in rat basal ganglia: Possible involvement of Golf in the coupling of dopamineD1 receptor with adenylyl cyclase. J. Neurosci., 1993. 13: p. 2237–2248. Sklar, P.B., Anholt, R.R.H., and Snyder, S.H., The odorant-sensitive adenylate cyclase of olfactory receptor cells: differential stimulation by different classes of odorants. J. Biol. Chem., 1986. 261: p. 15538–15543. Lowe, G., T. Nakamura, and G.H. Gold, Adenylate cyclase mediates olfactory transduction for a wide variety of odorants. Proceedings of the National Academy of Sciences of the United States of America, 1989. 86(14): p. 5641–5645. Pace, U., Hanski, E., Salomon, Y., and Lancet, D., Odorant-sensitive adenylate cyclase may mediate olfactory reception. Nature, 1985. 316: p. 255–258. Pfeuffer, E., et al., Olfactory adenylyl cyclase. Identification and purification of a novel enzyme form. J Biol Chem, 1989. 264(31): p. 18803–7. Bakalyar, H.A. and R.R. Reed, Science, 1990. 250: p. 1403–1406. Menco, B.P.M., et al., Ultrastructural localization of olfactory transduction components: the G-protein subunit Golf-Alpha and type III adenylyl cyclase. Neuron, 1992. 8(3): p. 441–453. Pun, R.Y. and S.J. Kleene, Contribution of cyclic-nucleotide-gated channels to the resting conductance of olfactory receptor neurons. Biophysical Journal, 2003. 84(5): p. 3425–3435. Wei, J., et al., Phosphorylation and inhibition of olfactory adenylyl cyclase by CaM kinase II in neurons: a mechanism for attenuation of olfactory signals. Neuron, 1998. 21(3): p. 495–504.

90

91

92

93

94

95

96 97

98

99

100

101

102

103

104

Wei, J., G. Wayman, and D.R. Storm, Phosphorylation and inhibition of type III adenylyl cyclase by calmodulin-dependent protein kinase II in vivo. Journal of Biological Chemistry, 1996. 271(39): p. 24231–24235. Leinders-Zufall, T., M. Ma, and F. Zufall, Impaired odor adaptation in olfactory receptor neurons after inhibition of Ca2+/calmodulin kinase II. Journal of Neuroscience, 1999: p. 19:RC19 (1–6). Borisy, F.F., et al., Calcium/calmodulin-activated phosphodiesterase expressed in olfactory receptor neurons. Journal of Neuroscience, 1992. 12(3): p. 915–923. Yan, C., et al., Molecular cloning and characterization of a calmodulin-dependent phosphodiesterase enriched in olfactory sensory neurons. Proceedings of the National Academy of Sciences of the United States of America, 1995. 92(21): p. 9677–9681. Wong, S.T., et al., Disruption of the type III adenylyl cyclase gene leads to peripheral and behavioral anosmia in transgenic mice. Neuron, 2000. 27(3): p. 487–497. Gold, G.H., Controversial issues in vertebrate olfactory transduction. Annu Rev Physiol, 1999. 61: p. 857–71. Gold, G., Lowe, G., Single odorant molecules? Nature, 1995. 376: p. 27. Baylor, D.A., T.D. Lamb, and K.W. Yau, Responses of retinal rods to single photons. Journal of Physiology, 1979. 288: p. 613–634. Buck, L., and Axel, R., A novel multigene family may encode odorant receptors: a molecular basis for odor recognition. Cell, 1991. 65: p. 175–187. Zhang, X. and S. Firestein, The olfactory receptor gene superfamily of the mouse. Nat Neurosci, 2002. 5(2): p. 124–33. Glusman, G., et al., The complete human olfactory subgenome. Genome Res, 2001. 11(5): p. 685–702. Rouquier, S., et al., Distribution of olfactory receptor genes in the human genome. Nat Genet, 1998. 18(3): p. 243–50. Fesenko, E.E., Koleniskov, S.S., Lyubarsky, A.L., Induction by cGMP of cationic conductance in plasma membrane of retinal rod outer segment. Nature, 1985. 313: p. 310–313. Haynes, L.W., Yau, K.W., Cyclic GMP-sensitive conductance in outer segment membrane of catfish cones. Nature, 1985. 317: p. 61–64. Nakamura, T., and Gold, G.H., A cyclic nucleotide-gated conductance in olfactory receptor cilia. Nature, 1987. 325: p. 342–344.

5.15 Open Questions 105 Huque, T., Bruch, R.C., Odorant and guanine

118 Yoshikami, S., Hagins, W.A., Cytoplasmic pH

nucleotide-stimulated phosphoinositide turnover in olfactory cilia. Biochem. Biophys. Res. Commun., 1986. 137: p. 36–42. Schild, D. and D. Restrepo, Transduction mechanisms in vertebrate olfactory receptor cells. Physiological Reviews, 1998. 78(2): p. 429–466. Frings, S., Lynch, J.W., Lindemann, B., Properties of cyclic nucleotide-gated channels mediating olfactory transduction: activation, selectivity, and blockage. J. Gen. Physiol., 1992. 100: p. 45–67. Firestein, S., B. Darrow, and G.M. Shepherd, Activation of the sensory current in salamander olfactory receptor neurons depends on a G-protein-mediated cAMP second messenger system. Neuron, 1991. 6: p. 825–835. Firestein, S., F. Zufall, and G.M. Shepherd, Single odor-sensitive channels in olfactory receptor neurons are also gated by cyclic nucleotides. J. Neurosci., 1991. 11: p. 3565–3572. Zufall, F., S. Firestein, and G.M. Shepherd, Analysis of single cyclic nucleotide-gated channels in olfactory receptor cells. J. Neurosci., 1991. 11: p. 3573–3580. Reisert, J., et al., The Ca-activated Cl Channel and its control in rat olfactory receptor neurons. Journal of General Physiology, 2003. 122(3): p. 349–364. Larsson, H.P., S.J. Kleene, and H. Lecar, Noise analysis of ion channels in non-space-clamped cables: Estimates of channel parameters in olfactory cilia. Biophysical Journal, 1997. 72(3): p. 1193–1203. Kurahashi, T., Kaneko, A., High-density campgated channels at the ciliary membrane in the olfactory receptor cell. Neuroreport, 1991. 2: p. 5–8. Kaur, R., et al., IP3-gated channels and their occurrence relative to CNG channels in the soma and dendritic knob of rat olfactory receptor neurons. Journal of Membrane Biology, 2001. 181(2): p. 91–105. Kolesnikov, S.S., A.B. Zhainazarov, and A.V. Kosolapov, Cyclic nucleotide-activated channels in the frog olfactory plasma membrane. FEBS Letters, 1990. 266: p. 96–98. Kurahashi, T., The response induced by intracellular cyclic-AMP in isolated olfactory receptor cells of the newt. Journal of Physiology-London, 1990. 430(V): p. 355–371. Menini, A., Currents carried by monovalent cations through cyclic GMP-activated channels in excised patches from salamander rods. J. Physiol., 1990. 424: p. 167–185.

in rod outer segments and high energy phosphate metabolism during phototransduction. Bio. Phys. J., 1985. 47: p. 101a (Abstr). Zufall, F. and S. Firestein, Divalent cations block the cyclic nucleotide-gated channel of olfactory receptor neurons. J Neurophysiol, 1993. 69(5): p. 1758–68. Dzeja, C., et al., Ca2+ permeation in cyclic nucleotide-gated channels. Embo Journal, 1999. 18(1): p. 131–144. Frings, S., Seifert, R., Godde, M., Kaupp, U.P., Profoundly different calcium permeation and blockage determine the specific function of distinct cyclic nucleotide-gated channels. Neuron, 1995. 15: p. 169–179. Lindemann, B., Predicted profiles of ion concentrations in olfactory cilia in the steady state. Biophysical Journal, 2001. 80(4): p. 1712–21. Haynes, L.W., Block of the cyclic GMP-gated channel of vertebrate rod and cone photoreceptors by l-cis-diltiazem. J. Gen. Physiol., 1992. 100: p. 783–801. Benos, D.J., Amiloride: a molecular probe of sodium transport in tissues and cells. Amer. J. Physiol., 1982. 242: p. C131–C145. Zhuang, Y.X., Cragoe, E.J., Shaikewitz, T., Glaser, L., Cassel, D., Characterization of potent Na/H exchange inhibitors from the amiloride series in a431 cells. Biochem., 1984. 23: p. 4481–4488. Tang, C.M., Presser, F., Morad, M., Amiloride selectively blocks the low threshold (T) calciumchannel. Science, 1988. 240: p. 213–215. Bielefeld, D.R., Hadley, R.W., Vassilev, P.M., Hume, J.R., Membrane electrical-properties of vesicular Na-Ca exchange inhibitors in single atrial myocytes. Circul. Res., 1986. 59: p. 381–389. Soltoff, S.P., Mandel, L.J., Amiloride directly inhibits the Na,K-ATPase activity of rabbit kidney proximal tubules. Science, 1983. 220: p. 957–959. Kleyman, T.R., Cragoe, E.J., Amiloride and its analogs as tools in the study of ion-transport. J. Membr. Bio., 1988. 105: p. 1–21. Nicol, G.D., Schnetkamp, P.P.M., Saimi, Y., Cragoe, E.J., Bownds, M.D., A derivative of amiloride blocks both the light-regulated and cyclic gmp-regulated conductances in rod photoreceptors. J. Gen. Physiol., 1987. 90: p. 651–669. Stern, J.H., Kaupp, U.B., Macleish, P.R., Control of the light-regulated current in rod photoreceptors by cyclic-GMP, calcium, and l-cis-diltiazem. Proc. Natl. Acad. Sci., 1986. 83: p. 1163–1167.

106

107

108

109

110

111

112

113

114

115

116

117

119

120

121

122

123

124

125

126

127

128

129

130

131

131

132

5 Vertebrate Olfactory Signal Transduction and the Interplay of Excitatory Anionic and Cationic Currents 132 Hurwitz, R., Holcombe, V., Affinity purifica-

133

134

135

136 137

138

139

140

141

142

143

144

145

tion of the photoreceptor GMP-gated cation channel. J. Biol. Chem., 1991. 226: p. 7975–7977. Dhallan, R.S., Yau, K-W., Schrader, K.A., Reed, R.R., Primary structure and functional expression of a cyclic nucleotide-activated channel from olfactory neurons. Nature, 1990. 347: p. 184–187. Ludwig, J., et al., Primary structure of cAMPgated channel from bovine olfactory epithelium. FEBS Lett., 1990. 270: p. 24–29. Kaupp, U.B., Niidome, T., Tanabe, T., Terada, S., Bonigk, W., Stuhmer, W., Cook, W.J., Kangawa, K., Matsuko, H., Hirose, T., Miyata, T., Numa, S., Primary structure and functional expression from complementary cDNA of the rod photoreceptor cyclic GMP-gated channel. Nature, 1989. 342: p. 762–766. Bradley, J., et al., Nomenclature for ion channel subunits. Science, 2001. 294(5549): p. 2095–6. Goulding, E.H., et al., Molecular cloning and single-channel properties of the cyclic nucleotidegated channel from catfish olfactory neurons. Neuron, 1992. 8: p. 45–58. Kaupp, U.B. and R. Seifert, Cyclic nucleotidegated ion channels. Physiol Rev, 2002. 82(3): p. 769–824. Bradley, J., Li, J., Davidson, N., Lester, H.A., Zinn, K, Heteromeric olfactory cyclic nucleotidegated channels: A subunit that confers increased sensitivity to cAMP. PNAS, 1994. 91: p. 8890–8894. Liman, E.R., Buck, L.B., A second subunit of the olfactory cyclic nucleotide-gated channel confers high sensitivity to cAMP. Neuron, 1994. 13: p. 611–621. Broillet, M.C. and S. Firestein, b subunits of the olfactory cyclic nucleotide-gated channel form a nitric oxide activated Ca2+ channel. Neuron, 1997. 18(6): p. 951–958. B€onigk, W., et al., The native rat olfactory cyclic nucleotide-gated channel is composed of three distinct subunits. J Neurosci, 1999. 19(13): p. 5332–47. Sautter, A., et al., An isoform of the rod photoreceptor cyclic nucleotide-gated channel beta subunit expressed in olfactory neurons. Proc Natl Acad Sci U S A, 1998. 95(8): p. 4696–701. Zheng, J. and W.N. Zagotta, Determine cyclic nucleotide-gated channel stiochiometry with FRET. Biophys J, 2003. 84(2): p. 138A. Kurahashi, T. and T. Shibuya, Ca2+-dependent adaptive properties in the solitary olfactory receptor cell of the newt. Brain Research, 1990. 515(1–2): p. 261–268.

146 Zufall, F., G.M. Shepherd, and S. Firestein,

147

148

149

150

151

152

153

154

155

156

157

158

159

Inhibition of the olfactory cyclic nucleotide gated ion channel by intracellular calcium. Proc R Soc Lond B Biol Sci, 1991. 246(1317): p. 225–30. Anholt, R.R.H., Rivers, A.M., Olfactory transduction - cross-talk between 2nd-messenger systems. Biochem., 1990. 29: p. 4409–4054. Hsu, Y.-T. and R.S. Molday, Modulation of the cGMP-gated channel of rod photoreceptor cells by calmodulin. Nature, 1993. 361: p. 76–79. Fain, G.L., et al., Adaptation in vertebrate photoreceptors. Physiol Rev, 2001. 81(1): p. 117–151. Chen, T.-Y., Yau, K-W., Direct modulation by Ca2+-calmodulin of cyclic nucleotide-activated channel of rat olfactory receptor neurons. Nature, 1994. 368: p. 545–548. Bradley, J., D. Reuter, and S. Frings, Facilitation of calmodulin-mediated odor adaptation by cAMP-gated channel subunits. Science, 2001. 294(5549): p. 2176–8. Munger, S.D., et al., Central role of the CNGA4 channel subunit in Ca2+-calmodulin-dependent odor adaptation. Science, 2001. 294(5549): p. 2172–5. Liu, M., et al., Calcium-calmodulin modulation of the olfactory cyclic nucleotide-gated cation channel. Science, 1994. 266(5189): p. 1348–54. Varnum, M.D. and W.N. Zagotta, Interdomain interactions underlying activation of cyclic nucleotide- gated channels. Science, 1997. 278(5335): p. 110–3. Grunwald, M.E., et al., Molecular determinants of the modulation of cyclic nucleotide-activated channels by calmodulin. Proc Natl Acad Sci U S A, 1999. 96(23): p. 13444–9. Zheng, J., M.D. Varnum, and W.N. Zagotta, Disruption of an Intersubunit Interaction Underlies Ca2+-Calmodulin Modulation of Cyclic Nucleotide-Gated Channels. J Neurosci, 2003. 23(22): p. 8167–8175. Trudeau, M.C. and W.N. Zagotta, Calcium/ calmodulin modulation of olfactory and rod cyclic nucleotide-gated ion channels. J Biol Chem, 2003. 7: p. 7. O’Neil, K.T. and W.F. DeGrado, How calmodulin binds its targets: sequence independent recognition of amphiphilic alpha-helices. Trends Biochem Sci, 1990. 15(2): p. 59–64. Bradley, J., Bonigk, W., Yau, K. W., Frings, S., Calmodulin permanently associates with rat olfactory CNG channels under native conditions. Nat Neurosci, 2004 7: p. 705–710.

5.15 Open Questions 160 Grunwald, M.E., et al., Identification of a do-

161

162

163

164

165

166

167

168

169

170

171

172

main on the beta-subunit of the rod cGMP-gated cation channel that mediates inhibition by calcium-calmodulin. J Biol Chem, 1998. 273(15): p. 9148–57. Weitz, D., et al., Calmodulin controls the rod photoreceptor CNG channel through an unconventional binding site in the N-terminus of the beta-subunit. Embo J, 1998. 17(8): p. 2273–84. Bahler, M. and A. Rhoads, Calmodulin signaling via the IQ motif. FEBS Lett, 2002. 513(1): p. 107–13. Erickson, M.G., et al., FRET two-hybrid mapping reveals function and location of L-type Ca2+ channel CaM preassociation. Neuron, 2003. 39(1): p. 97–107. Zuhlke, R.D., et al., Calmodulin supports both inactivation and facilitation of L-type calcium channels. Nature, 1999. 399(6732): p. 159–62. Kleene, S.J. and R.C. Gesteland, Calcium-activated chloride conductance in frog olfactory cilia. Journal of Neuroscience, 1991. 11(11): p. 3624–3629. Yoshii, K. and K. Kurihara, Role of cations in olfactory reception. Brain Research, 1983. 274: p. 239–248. Suzuki, N., Effects of different ionic invironments on the responses of single olfactory responses in the lamprey. Comperative Biochemistry and Physiology. A Comperative Physiology, 1978. 61: p. 461–467. Kleene, S.J. and R.Y.K. Pun, Persistence of the olfactory receptor current in a wide variety of extracellular environments. Journal of Neurophysiology, 1996. 75(4): p. 1386–1391. Tucker, D. and T. Shibuya, A physiologic and pharmacologic study of olfactory receptors. Cold Spring Harb Symp Quant Biol, 1965. 30: p. 207–15. Zufall, F., T. Leinders-Zufall, and C.A. Greer, Amplification of odor-induced Ca2+ transients by store-operated Ca2+ release and its role in olfactory signal transduction. Journal of Neurophysiology, 2000. 83(1): p. 501–512. Leinders-Zufall, T., et al., Imaging odor-induced calcium transients in single olfactory cilia: Specificity of activation and role in transduction. Journal of Neuroscience, 1998. 18(15): p. 5630–5639. Leinders-Zufall, T., et al., Calcium entry through cyclic nucleotide-gated channels in individual cilia of olfactory receptor cells: Spatiotemporal dynamics. Journal of Neuroscience, 1997. 17(11): p. 4136–4148.

173 Lowe, G. and G.H. Gold, Nonlinear amplifi-

174

175

176

177

178

179

180

181

182

183

184

cation by calcium-dependent chloride channels in olfactory receptor cells. Nature, 1993. 366(6452): p. 283–286. Kurahashi, T. and K.-W. Yau, Co-existence of cationic and chloride components in odorant-induced current of vertebrate olfactory receptor cells. Nature, 1993. 363(6424): p. 71–74. Zhainazarov, A.B. and B.W. Ache, Odor-induced currents in Xenopus olfactory receptor cells measured with perforated-patch recording. Journal of Neurophysiology, 1995. 74(1): p. 479–483. Kleene, S.J., High-gain, low-noise amplification in olfactory transduction. Biophysical Journal, 1997. 73(2): p. 1110–1117. Sato, K. and N. Suzuki, The contribution of a Ca2+-activated Cl– conductance to amino-acidinduced inward current responses of ciliated olfactory neurons of the rainbow trout. Journal of Experimental Biology, 2000. 203(2): p. 253–262. Dubin, A.E. and V.E. Dionne, Action potentials and chemosensitive conductances in the dendrites of olfactory neurons suggest new features for odor transduction. Journal of General Physiology, 1994. 103(2): p. 181–201. Reuter, D., et al., A depolarizing chloride current contributes to chemoelectrical transduction in olfactory sensory neurons in situ. Journal of Neuroscience, 1998. 18(17): p. 6623–6630. Nakamura, T., H. Kaneko, and N. Nishida, Direct measurement of the chloride concentration in newt olfactory receptors with the fluorescent probe. Neuroscience Letters, 1997. 237(1): p. 5–8. Kaneko, H., T. Nakamura, and B. Lindemann, Noninvasive measurement of chloride concentration in rat olfactory receptor cells with use of a fluorescent dye. American Journal of Physiology, 2001. 280(6): p. C1387–1393. Chiu, D., T. Nakamura, and G.H. Gold, Ionic Composition of Toad Olfactory Mucus Measured With Ion-Selective Microelectrodes. Chemical Senses, 1988. 13(4): p. 677–678. Firestein, S. and G.M. Shepherd, Interaction of anionic and cationic currents leads to a voltage dependence in the odor response of olfactory receptor neurons. Journal of Neurophysiology, 1995. 73(2): p. 562–567. Hallani, M., J.W. Lynch, and P.H. Barry, Characterization of calcium-activated chloride channels in patches excised from the dendritic knob of mammalian olfactory receptor neurons. Journal of Membrane Biology, 1998. 161(2): p. 163–171.

133

134

5 Vertebrate Olfactory Signal Transduction and the Interplay of Excitatory Anionic and Cationic Currents 185 Kleene, S.J., R.C. Gesteland, and S.H. Bryant,

195 Neher, E., Concentration profiles of intracellular

An electrophysiological survey of frog olfactory cilia. Journal of Experimental Biology, 1994. 195: p. 307–328. Frings, S., D. Reuter, and S.J. Kleene, Neuronal Ca2+-activated Cl– channels - homing in on an elusive channel species. Progress in Neurobiology, 2000. 60(3): p. 247–289. Kuruma, A. and H.C. Hartzell, Bimodal control of a Ca2+-activated Cl– channel by different Ca2+ signals. Journal of General Physiology, 2000. 115(1): p. 59–80. Callamaras, N. and I. Parker, Ca2+-dependent activation of Cl– currents in Xenopus oocytes is modulated by voltage. American Journal of Physiology Cell Physiology, 2000. 278(4): p. C667–675. Arreola, J., J.E. Melvin, and T. Begenisich, Activation of calcium-dependent chloride channels in rat parotid acinar cells. Journal of General Physiology, 1996. 108(1): p. 35–47. Piper, A.S. and W.A. Large, Multiple conductance states of single Ca2+-activated Cl– channels in rabbit pulmonary artery smooth muscle cells. Journal of Physiology, 2003. 547(Pt 1): p. 181–196. Nakamura, T., et al., Gated Conductances in Native and Reconstituted Membranes From Frog Olfactory Cilia. Biophysical Journal, 1996. 70(2): p. 813–817. Hirakawa, Y., et al., Ca2+-dependent Cl– channels in mouse and rabbit aortic smooth muscle cells: regulation by intracellular Ca2+ and NO. American Journal of Physiology, 1999. 277: p. H1732–1744. Van Renterghem, C. and M. Lazdunski, Endothelin and vasopressin activate low conductance chloride channels in aortic smooth muscle cells. Pflugers Archiv, 1993. 425(1–2): p. 156–163. Kl€ockner, U., Intracellular calcium ions activate a low-conductance chloride channel in smoothmuscle cells isolated from human mesenteric artery. Pflugers Archiv, 1993. 424(3–4): p. 231–237.

calcium in the presence of a diffusible chelator, in Experimental Brain Research, U. Heinemann, et al., Editors. 1986, Springer: Heidelberg. p. 80–96. Bauer, P.J., The local Ca concentration profile in the vicinity of a Ca channel. Cell Biochemistry and Biophysics, 2001. 35(1): p. 49–61. Becq, F., Ionic channel rundown in excised membrane patches. Biochimica et Biophysica Acta, 1996. 1286: p. 53–63. Kleene, S.J., Origin of the chloride current in olfactory transduction. Neuron, 1993. 11(1): p. 123–132. Kleene, S.J., Both external and internal calcium reduce the sensitivity of the olfactory cyclic-nucleotide-gated channel to cAMP. Journal of Neurophysiology, 1999. 81(6): p. 2675–2682. Noe, J., et al., Sodium/calcium exchanger in rat olfactory neurons. Neurochemistry International, 1997. 30(6): p. 523–531. Jung, A., et al., Sodium/calcium exchanger in olfactory receptor neurons of Xenopus laevis. Neuroreport, 1994. 5(14): p. 1741–1744. Reisert, J. and H.R. Matthews, Na+-dependent Ca2+ extrusion governs response recovery in frog olfactory receptor cells. Journal of General Physiology, 1998. 112(5): p. 529–535. Lo, Y.H., T.M. Bradley, and D.E. Rhoads, High-Affinity Ca2+,Mg2+-ATPase in plasma membrane-rich preparations from olfactory epithelium of atlantic salmon. Biochimica Et Biophysica Acta-Biomembranes, 1994. 1192(2): p. 153–158. Neher, E., Vesicle pools and Ca2+ microdomains: new tools for understanding their roles in neurotransmitter release. Neuron, 1998. 20(3): p. 389–399. Reuter, D., Ph. D. Thesis. 2000.

186

187

188

189

190

191

192

193

194

196

197

198

199

200

201

202

203

204

205

135

6

Transduction Channels in the Vomeronasal Organ Emily R. Liman and Frank Zufall

6.1

Introduction

The vomeronasal organ (VNO) is a sensory structure found in the nasal cavity of most vertebrate animals that detects chemosignals, including some pheromones. Pheromones are defined as chemicals released by an animal that elicit stereotyped behavioral or neuroendocrine responses in other animals of the same species [1]. In mammals, pheromones play an important role in regulating reproductive and social behaviors. For example, pheromones can induce suppression of estrus, aggressive behavior, suckling, and mating, effects that depend on the nature of the pheromones released as well as the gender, age, and neuroendocrine status of the recipient. The VNO mediates many, but not all, of the effects of pheromones in vertebrates, and it may also play a role in detecting chemosignals that are not pheromones [2, 3]. For example, in the snake the VNO plays a role in prey detection [4]. In recent years, great progress has been made in understanding mechanisms by which chemosensory stimuli are detected and transduced by the VNO (reviewed in [5]). We now know the identity of most of the elements of the transduction cascade, including the identity of at least one component of the transduction channel, the ion channel TRPC2 [45]. The identification of TRPC2 in the VNO and the subsequent generation of TRPC2 knockout mice have not only provided insight into mechanisms of sensory transduction but also have allowed further study into the functional significance of the vomeronasal system. Thus, at the end of this chapter, information from TRPC2 knockout mice is used to evaluate the function of the vomeronasal system in the mouse and its evolution in humans.

Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

136

6 Transduction Channels in the Vomeronasal Organ

6.2

Anatomy of the Vomeronasal System

Two anatomically distinct organs sense pheromones: the main olfactory epithelium (MOE) and the VNO. The VNO is a paired tubular structure at the rostral end of the nasal cavity (see Fig. 6.1) that receives input either from the nasal or oral cavity, depending on species. Because the VNO is a closed tube, non-volatile chemicals are introduced into the lumen of the VNO by an active pumping mechanism [6]. Recent in vivo recording shows that in mice the VNO is most active when an animal is in direct contact with a conspecific and is most sensitive to chemicals present in the oral or anogenital region [7]. In contrast, the main olfactory system responds to odorants and pheromones that are volatile and are passed over the epithelium during the respiratory cycle. The VNO sends projections to the accessory olfactory bulb (AOB), which lies adjacent to the main olfactory bulb (MOB). However, projections from the AOB, unlike those from the MOB, do not reach cortical areas; instead, signals are sent to hypothalamic regions involved in the control of stereotyped behavior and neuroendocrine responses. In cross-section the VNO appears as a crescent-shaped structure (see Fig. 6.1), and within this crescent two broad zones can be distinguished. The apical region contains sensory neurons that express the signaling molecular Gai2 and a set of putative receptor molecules (V1Rs; discussed below), whereas basal neurons express Gao and V2R receptors. Projections from the apical neurons are to the anterior part of the AOB, whereas basal neurons project to the posterior part of the AOB [5]. It is likely that some of the projections remain segregated at higher brain regions [8, 9] and that these two zones play differential roles in regulating behavior. The VNO sensory neurons are bipolar cells that send a single dendrite to the lumen of the vomeronasal epithelium and a single axon to the AOB (see Fig. 6.1). The den-

Fig. 6.1 Anatomy of the mouse vomeronasal system. (A) Schematic drawing of a parasagittal section shows the VNO and the turbinates of the main olfactory epithelium (OE). The VNO projects to the accessory olfactory bulb (AOB), whereas the OE projects to the olfactory bulb (OB). (B) In a section taken at the position indicated in (A), both

the OE on the roof of the nasal cavity and the VNO at the base (depicted as crescents) can be seen. (C) A close-up of the epithelium of the VNO shows bipolar vomeronasal sensory neurons, which send dendrites containing microvilli to the lumen of the epithelium. From [69]

6.3 Sensory Responses Involve Generation of Action Potentials and Ca2+ Entry

drite ends in a tuft of microvilli, which serve to increase the surface area over which pheromone receptors are expressed. The observation that VNO neurons contain microvilli, whereas olfactory neurons are ciliated, suggests a distinct evolutionary origin for the two sets of cells. This is supported by the recent molecular data showing different sets of signaling molecules in the two cell types. As will be discussed later, a major difference between sensory transduction in the VNO and that in the MOE is that VNO sensory transduction is mediated by a PLC signaling pathway leading to the opening of an ion channel in the transient receptor potential (TRP) family, whereas transduction in the MOE is mediated by cyclic nucleotides that act directly on a cyclic nucleotide-gated (CNG) channel. It may not be a coincidence that other microvillous sensory neurons, such as the fly photoreceptor (see Chapter 8), the vertebrate hair cell, and the vertebrate taste cell, also express TRP channels, whereas ciliated neurons such as the vertebrate photoreceptor express CNG channels. The reason for this correlation between structure and signaling is not known, and it may have a functional significance or instead be an evolutionary vestige.

6.3

Sensory Responses Involve Generation of Action Potentials and Ca2+ Entry

An important question that was until recently unresolved is whether sensory transduction in the VNO leads to an excitatory response followed by the induction of action potentials, as in the main olfactory system, or to the suppression of activity, as in the visual system. This is a key issue because some papers claimed that sensory transduction led to suppression of activity [10, 11] and only recently have these results been refuted. Given the extreme diversity of receptors expressed by VNO sensory neurons, finding the answer to this question required that methods be devised to record simultaneously from large populations of VNO sensory neurons [12, 13]. Summed field potentials recorded from the microvillous layer of intact mouse VNO sensory epithelium was first employed by Leinders-Zufall et al. [12] to identify chemically defined ligands acting on vomeronasal neurons and their receptors. These studies clearly showed that pheromonal ligands generate negative deflections in the VNO field potential, consistent with an excitation of the VNO neurons. Subsequently, this method proved highly useful for the functional phenotyping of mice with targeted mutations in genes that encode the TRPC2 channel, V1r receptors, or adenylyl cyclase III [3, 14, 15] (see below). A 64-microelectrode array was used by Holy et al. [13] to record extracellular action potential activity from large subsets of sensory neurons in an explant of mouse VNO epithelium. This study demonstrated that as many as 40 % of the recorded neurons responded by increasing their spike frequency rate to the complex pheromonal stimuli that are present in urine. Moreover, subsets of VNO neurons were activated selectively by either male or female urine, whereas the response of other cells was independent of the sex of the donor animal [13]. This method has also proven useful for the characterization of mice with specific defects in VNO transduction [50].

137

138


Comparable results were also obtained by other researchers using extracellular single-electrode recording in tissue slices from rat VNO [16]. To examine responses of individual vomeronasal neurons to chemically defined ligands, Leinders-Zufall et al. [12] used extracellular loose-patch recordings to register activity from single, optically identified neurons in mouse VNO tissue slices. These experiments clearly demonstrated that VNO neurons respond to sensory stimulation with an increase in the rate of action potential firing. Therefore, it is now clear that VNO neurons in several species generate an excitation in response to chemostimulation. To record from a large number of spatially defined cells simultaneously, LeindersZufall et al. [12] used confocal imaging to measure Ca2+ responses of VNO slices. They showed for the first time that sensory neurons exhibit an increase in intracellular Ca2+ in response to chemostimulation and that these Ca2+ signals are eliminated after removal of extracellular Ca2+, indicating that they were triggered primarily by Ca2+ entry. These experiments also showed that the detection thresholds for pheromones by individual VNO neurons are at remarkably low concentrations, near 10–11 M, and that the individual cells are selectively tuned to detect only one or a few chemicals [12]. Subsequent studies using snake and rat VNO neurons have confirmed that chemostimulation produces an increase in intracellular Ca2+ in these cells [17, 18]. However, in the snake VNO, removal of extracellular Ca2+ did not fully suppress the chemoresponse [17]. Recent patch-clamp recordings from single mouse VNO neurons have revealed the ionic basis underlying this excitatory response (Fig. 6.2) [19]. These findings show that mouse VNO neurons generate a depolarizing receptor potential that, in turn, causes the cells to discharge action potentials (Fig. 6.2). Underlying this graded depolarization is a transient inward current, the sensory current, which represents the earliest electrical event in the transduction process. Since the input resistance of VNO neurons is

Fig. 6.2 Sensory responses of single mouse VNO neurons to focal stimulation of the dendritic tip with dilute urine (DU, 1/100), a rich source of natural pheromones. (A) Under current clamp, sensory stimulation produces a depolarizing receptor potential leading to robust action potential discharges. (B) Under voltage clamp, stimulation generates a rapidly activating and then deactivating inward current, the sensory current (holding potential, –70 mV). Recordings were obtained from visually identified neurons using an acute VNO tissue slice preparation. From [19]

6.4 Two Families of G-protein-coupled Receptors Mediate VNO Transduction

very high, only a few picoamperes of inward current are needed to generate action potentials [20]. What is the nature of the signaling cascade that leads to the generation of this sensory current? This question has been approached with both molecular and physiological methods.

6.4

Two Families of G-protein-coupled Receptors Mediate VNO Transduction

An early insight into VNO transduction came from the observation that components of MOE transduction were absent from the VNO [21–23]. Thus, a whole new set of signaling molecules remained to be found. Most elusive of these, perhaps, were the receptor molecules. A priori there was little information concerning the number or chemical makeup of these receptors. Thus, a major breakthrough in our understanding came with the cloning of the first family of VNO receptor genes by Dulac and Axel in 1995 [21]. In cloning the VNO receptor genes, Dulac and Axel made only two assumptions: that cells would differ in the receptor transcripts that they expressed and that the expression of these receptors would be restricted to the VNO. By cloning mRNAs that were differentially expressed between two single VNO neurons, Dulac and Axel identified a putative seven-transmembrane protein which, at the time, showed little homology to known G-protein-coupled receptors and was uniquely expressed in the VNO [21]. It is now known that there are 137 related genes with intact reading frames in the mouse genome [24], including the original receptor family (originally named VNs and subsequently named V1Rs) and a second group of receptors (originally named V3Rs and subsequently named V1Rds), [25] and that these receptors show structural similarity to a class of taste receptors (T2Rs). V1R receptors are restricted in expression to the apical region of the VNO sensory epithelium [21]. A distinct family of receptors (V2Rs) that are expressed in the basal region of the VNO sensory epithelium was subsequently identified [26–28]; these receptors, of which there are 100 in the mouse, contain a large amino terminus and show structural similarity to metabotropic glutamate receptors and T1Rs taste receptors. How these receptors interact with their ligands is not known. Recently it was shown that non-classical major histocompatibility complex (MHC) class Ib molecules interact with V2Rs and participate in receptor trafficking and possibly ligand binding [29]. Evidence that either V1Rs or V2Rs are involved in chemosensory transduction remains sparse. V2R receptor protein has been detected in sensory microvilli [29], and a mouse in which a large cluster of V1Rs was deleted shows some behavioral abnormalities consistent with a partial loss of pheromone transduction [15]. Given the large repertoire of VNO receptors, linking specific receptors with ligands and with behavioral responses remains an important challenge for the future. A step in this direction was made with a report that VNO neurons expressing the V1Rb2 receptor respond specifically to nanomolar concentrations of 2-heptanone [30].

139

140


6.5

Signaling Downstream of G Proteins May Involve a PLC

What is the nature of the signaling cascade that mediates VNO sensory transduction? Several lines of investigation have yielded insight into this question. Immunohistochemistical studies have shown that two G proteins, Gao and Gai2, are expressed in distinct and non-overlapping zones of the VNO epithelium and that their distribution coincides exactly with that of the two classes of receptors [31, 32]. This suggests that V1Rs may couple to Gai2 and V2Rs to Gao. Biochemical experiments show activation of these G proteins to pheromonal stimuli [33], and the observation that Gai2 and Gao proteins are present in microvilli provides further support for this hypothesis [34]. More recently, mice that carry a targeted deletion of Gai2 have been found to show defects in pheromone-mediated behaviors, consistent with a role for Gai2 in VNO transduction [35]. The coupling between G protein and downstream effectors is perhaps the least understood step in the signal transduction cascade. There is general consensus that a phospholipase C (PLC) isoform plays a critical role in sensory transduction. Phospholipase C hydrolyzes PIP2 into IP3 and DAG. Biochemical studies have shown that IP3 is generated in response to sensory activation of the VNO in snake [36], hamster [37], and pig [38] and that physiological responses of VNO receptor neurons to urine stimuli are blocked by a selective inhibitor of PLCs (U71332) [13, 18, 19]. However molecular studies have mysteriously failed to reveal the identity of the putative PLC. As the G alpha subunits Gi and Go are not typically involved in signaling through a PLC, it has been suggested that bc subunits, specifically Gc2 and Gc8, are responsible for activation of the PLC signaling pathway. Evidence to support this comes from the observation that antibodies directed against Gc8 and Gc2 block the generation of IP3 in response to urinary pheromones and a2 globulin, respectively [39].

6.6

Second Messengers for VNO Transduction: Functional Studies

To determine mechanisms of sensory transduction, a number of laboratories have studied functional responses of VNO sensory neurons to putative second messengers. No responses to cAMP or cGMP were detected in mouse VNO neurons [20], indicating that transduction is distinct from that of either the main olfactory system or the visual system of vertebrates. Responses to a wide range of second messengers that are generated downstream of PLC (Ca2+, IP3, DAG, and PUFAs) have been reported, but a connection to sensory transduction as described below has been made for only two of them: DAG and its metabolite arachidonic acid. In particular the following second-messenger responses have been reported: 1. Ca2+: Ca2+ has been shown to activate a 25-pS cation channel in excised patches from hamster [40] and mouse VNO neurons (Liu and Liman, unpublished results). This channel is equally permeable to Na+, K+, and Cs+ but is not detectably

6.7 Identification of the TRPC2 Ion Channel as a Candidate Transduction Channel

permeable to Ca2+. The channel is activated by only high concentrations of intracellular Ca2+ (K1/2 0.5 mM) and it is blocked by micromolar levels of adenine nucleotides (ATP and cAMP). The molecular identity of this channel is not known, although it may belong to the TRPM class of ion channels, several of which are Ca2+ activated [70–73]. This conductance may play a role in amplifying the primary sensory response but is unlikely to mediate primary sensory transduction. 2. IP3: Dialysis of IP3 through a pipette has been shown to activate a conductance in VNO neurons from turtle [41], rat [42], and snake [43]; however, none of these reports showed that IP3 directly activates a membrane channel in these cells. Furthermore, other researchers have observed only small and slow responses to intracellular dialysis of mouse VNO neurons with IP3 ([19] and Liman, unpublished observations), inconsistent with the previous studies. In addition, pharmacological block of the IP3 receptor does not interfere with the response of mouse neurons to pheromones in urine [18]. 3. Arachidonic acid (AA): AA, a polyunsaturated fatty acid that is generated as a metabolite of DAG, induced an increase in Ca2+ and a slowly activating inward current in rat VNO neurons, although direct activation of a membrane channel by AA has not been demonstrated [18]. Application of the DAG lipase inhibitor RHC-80267, which inhibits the synthesis of AA from DAG, blocked Ca2+ responses to urinary pheromones in the experiments by Spehr et al. [18], leading the authors to conclude that AA plays a role as a second messenger for VNO transduction. However, RHC80267 failed to block activation of the sensory current in mouse VNO neurons [19]. Thus, it is still unclear whether an AA-activated conductance plays a role in VNO sensory transduction. 4. Diacylglycerol (DAG): Analogues of DAG activate a 42-pS channel in excised insideout patches from the dendritic tip of VNO sensory neurons [19] (Fig. 6.3). The channel is permeable to Na+, Cs+, and Ca2+ but not to N-methyl-D-glucamine (with relative permeabilities of PCa/PNa = 2.7 and PCs/PNa = 1.5). A current with similar properties is activated by DAG in whole-cell recordings, showing a linear I-V relation with a reversal potential close to 0 mV [19]. These currents are blocked by external Ca2+ and by 2-aminoethoxydiphenyl borate (2-APB), which blocks some TRP channels [44]. These properties match those of the pheromone-induced current, providing support that the DAG-activated channel mediates pheromone transduction [19]. This notion is further supported by the result that both the DAG-activated channel and the sensory response are significantly impaired in mice with a targeted deletion in the TRPC2 gene (see below).

6.7

Identification of the TRPC2 Ion Channel as a Candidate Transduction Channel for VNO Sensory Signaling

An essential element of the signal transduction cascade is the ion channel that ultimately converts the chemical signaling into an electrical response. At present, the best candidate for this channel is the molecule TRPC2 (originally called TRP2), which was

141

142


Fig. 6.3 Identification of a diacylglycerol-activated cation channel in inside-out membrane patches from the dendritic tip of VNO neurons. (A) Addition of the endogenous DAG analogue 1-stearoyl-2-arachidonoyl-sn-glycerol (SAG, 100 lM) to the bathing medium produces an increase in channel activity that gives rise to a sustained inward current. This effect is accompanied by an increased noise level that reflects fluctuations in the activity of single channels. The presence of ATP, GTP, Ca2+, or Mg2+ in the bath is not required for activation of this current (holding potential, –80 mV). (B, C) Examples of single-channel events before (B) and after (C) application of SAG (10 lM). The SAGactivated channel has a unitary conductance of 42 pS in divalent cationfree solutions with Na+ as the sole cation and shows none of the hallmarks of capacitative Ca2+ entry channels (holding potential, –80 mV). From [19]

identified by Liman et al. in 1999 [45]. TRPC2 was identified in a PCR screen for ion channels in the VNO that were related to dTRP, the ion channel that mediates phototransduction in the fly (see Chapter 8). Because dTRP acts downstream of a PLC signaling cascade, it was hypothesized that a related channel would mediate PLC-dependent signaling in the VNO. As will be discussed below, this hypothesis has now been confirmed. The TRPC2 message encodes a protein of 885 amino acids in the rat and 890 amino acids in the mouse. It is distantly related to other mammalian TRPC channels (30 % amino acid identity), and, like other TRP channels, contains six trans-

6.7 Identification of the TRPC2 Ion Channel as a Candidate Transduction Channel

Fig. 6.4 Structure of the TRPC2 ion channel. (A) The rat TRPC2 protein contains 885 amino acids with an ankyrin repeat domain in the N-terminus and a coiled-coil domain in the C-terminus. The six transmembrane (tm) domains are thought to fold like those of a K+ channel, with a proposed pore region between the fifth and sixth tm domains. TRPC2 is likely to form a tetramer, like the K+ channel. (B) Phylogenetic tree of mammalian TRPC channels. TRPC2 is the most divergent of the mammalian TRPC channels

membrane domains (Fig. 6.4). Like voltage-gated K+ channels [46, 47], it is assumed to assemble as a tetramer. TRPC2 is expressed uniquely and abundantly in the VNO, with expression restricted to VNO sensory neurons [45] (Fig. 6.5). Since the original cloning of TRPC2 from the VNO, additional, longer splice isoforms have been identified [48, 49]. At present the significance of these is under dispute, and additional reports have confirmed that the isoform originally identified in the VNO is specifically expressed only in the VNO [49]. Further evidence that the short isoform is indeed the only TRPC2 protein expressed in the VNO comes from an examination of the protein by Western blot analysis using an antibody raised against the C-terminus of the TRPC2 protein. This shows that a

Fig. 6.5 Expression of TRPC2 in the VNO. (A) Labeling of a section of VNO from an adult rat with a digoxigenic antisense probe directed against TRPC2 reveals strong expression (blue reaction product) of the TRPC2 mRNA in VNO sensory neurons (N). (B) TRPC2 protein, detected by labeling with an antiTRPC2 antibody (red), is localized to the luminal surface of the epithelium. (C) In singly dissociated VNO sensory neurons, TRPC2 immunoreactivity (red) is clearly seen in the tuft of microvilli at the distal end of the dendrite. Scale bars: (B) 100 lm (C) 5 lm From [45]

143

144


protein of identical molecular weight is expressed in the VNO and in HEK cells transfected with the short isoform of TRPC2 [45]. To determine whether TRPC2 is likely to mediate VNO transduction, Liman et al. [45] localized the protein in sections of VNO and in singly dissociated VNO neurons (Fig. 6.5). These experiments showed a striking restriction of the TRPC2 protein to sensory microvilli, strongly supporting the view that the channel plays a role in sensory transduction [45]. This is further supported by immuno-EM localization of TRPC2 [34].

6.8

TRPC2 Is Essential for Pheromone Transduction

If TRPC2 is a major component of the VNO transduction channel, then targeted deletion of the channel should impair VNO sensory transduction. Indeed, two groups have independently generated TRPC2 -/- mice and shown that in these mice there is little or no electrical response to putative pheromones [14, 50]. Stowers et al. tested TRPC2 -/mice with a 64-microelectrode array and found a complete absence of electrical responses to dilute urine. Leypold et al. measured field potentials in response to dilute urine and reported a severe reduction in the amplitude and sensitivity of the response, although a small residual field potential was observed. Furthermore, as described below, TRPC2 -/- mice show severe behavioral defects in response to pheromones, further supporting the notion that TRPC2 is essential for pheromone transduction.

6.9

Mechanism of TRPC2 Activation

Identification of the mechanism by which TRPC2 is gated has been impeded by the fact that, in heterologous cell types, TRPC2 is retained in the endoplasmic reticulum [49]. TRPC2 is structurally related to ion channels that mediate Drosophila phototransduction. Considerable literature has provided a detailed understanding of phototransduction in Drosophila, which therefore serves as a model for understanding VNO transduction [51] (see Chapter 8). In the Drosophila rhabdomere, light-induced activation of a PLC leads to the opening of a light-activated conductance that is composed of three types of TRP channels, dTRP, dTRP-like, and TRPc. Recent evidence suggests that these TRP channels are activated by DAG or its metabolites (polyunsaturated fatty acids), possibly in combination with the reduction in phosphatidyl inositol 4,5 bisphosphate (PIP2) [52, 53]. Other mammalian homologues of the Drosophila TRPs are also activated by downstream products of PLC. Two of these (hTRPC3 and hTRPC6) appear to be activated directly by DAG [54], while others may be activated by association with an IP3 receptor [55, 56]. A major insight into the mechanism of activation of TRPC2 channels has come from comparing second messenger-gated ion channels in VNO sensory neurons from wild-

6.9 Mechanism of TRPC2 Activation

type (wt) and TRPC2 mutant mice [19]. This work revealed a novel Ca2+-permeable cation channel that exists in high density in the plasma membrane at the dendritic tips of wt VNO neurons and that is defective in TRPC2 -/- mice (Fig. 6.6). The channel can be gated effectively by the endogenous DAG analogue 1-stearoyl-2-arachidonoylsn-glycerol (SAG), as well as the other DAG analogues including OAG and DOG, but not by a monoacylglycerol or by IP3. Channel activation by DAG is independent of protein kinase C, Ca2+, and Mg2+, suggesting that DAG exerts a direct effect on the channel, not unlike its effects on hTRPC3 and hTRPC6 [54] . In TRPC2 -/- mice, the DAG-gated current is conspicuously absent, with a reduction in the maximum current in response to SAG of 8 % of wt levels. Thus, it is very likely that the TRPC2 is a major component of the DAG-activated channel. The presence of a small but significant residual conductance in TRPC2-/- sensory neurons suggests that other DAG-activated channel subunits may exist in these cells. One possibility is that TRPC2 forms a heteromultimeric channel complex with these predicted channel subunits, which otherwise function only poorly on their own. The mechanism of activation of TRPC2 deduced from the work of Lucas et al. [19] is inconsistent with the proposal that TRPC2 is gated by store-dependent Ca2+ mobilization or by a possible postulated interaction of TRPC2 with an IP3 receptor [48, 57]. These mechanisms are further unlikely given the strict localization of TRPC2 to sensory microvilli that are devoid of membranous compartments that could serve as Ca2+ stores [34, 45].

Fig. 6.6 TRPC2-/- VNO neurons display a striking defect in the activation of the DAG-gated channel. (A) Representative families of whole-cell currents to a series of depolarizing and hyperpolarizing voltage steps (as indicated in the figure) recorded from isolated wild-type VNO neurons under control conditions. Experiments are performed in the presence of 1 lM tetrodotoxin to block voltage-gated Na+ channels; voltage-activated K+ channels are blocked by using a Cs+-based pipette solution. Dotted line, zero current level. (B) A prominent DAG-gated conductance is observed in these cells following the addition of 100 lM SAG to the bath solution. (C, D) In VNO neurons from TRPC2-/- mice, SAG application fails to activate a large conductance. However, a drastically diminished residual response to SAG still exists in these cells (D). From [19]

145

146


6.10

TRPC2 Knockout Mice: Behavioral Defects

Previous understanding of the function of the VNO in animal behavior came primarily from studies in which the VNO was surgically ablated in animals and the effects of the ablation were assessed by behavioral assays after the animals recovered [58]. However, this approach is limited because few researchers are experienced in the surgical procedures and because the animals must be exposed to major surgery, which might mask subtle effects of the VNO ablation. The genetic ablation of TRPC2, which severely impairs the VNO sensory response, thus provides an excellent model system in which to determine the biological role of the VNO. These animals have a nonfunctional VNO, although in other ways they appear to be normal. For example, contrary to a report that the TRPC2 plays a role in sperm acrosome reaction [59], these mice are fertile and indistinguishable from wild-type littermates in the number of offspring. The behavior of the TRPC2 knockout mouse has been examined independently by two labs that have reached broadly similar conclusions [14, 50]. Two key results have emerged. First, TRPC2 is essential for pheromone-evoked male-male aggression (Fig. 6.7). In a resident-intruder assay, which tests for male-male aggression, TRPC2-/males fail to initiate attack behavior, although they are physically and neurologically capable of displaying aggressive interactions [14, 50]. Interestingly, presumably as a result of this deficit in displaying aggressive behavior in response to male pheromones, TRPC2-/- males usually fail to establish dominance and instead display urine-marking behavior typical of subordinate males [14]. Aggressive behavior is also severely attenuated in lactating female TRPC2-/- mice that are confronted with a

Fig. 6.7 TRPC2 knockout mice show behavioral abnormalities. (A) Average attack frequency of a resident male mouse of the genotype indicated in response to a castrated male intruder swabbed or not swabbed with pheromones from a sexually mature mouse. Note that male pheromones elicit attack behavior in the wt but not the TRPC-/mouse. (B) Mounting behavior of wt and TRPC2 -/- mice to same intruder mice. There is a dramatic increase in mounting behavior in the TRPC2-/- mouse From [50]

6.11 Loss of VNO Signaling Components in Human Evolution

male intruder, indicating that signals transduced by the VNO initiate aggressive behavior in both males and females [14]. Second, a striking defect is seen in the sexual behavior of TRPC2-/- males (Fig. 6.7). Although TRPC2-/- males mate normally with females, they display increased sexual behavior towards other males, i.e., mounting other males at a much higher rate [14, 50]. This behavior has not been observed in animals in which the VNO was surgically ablated, possibly due to secondary effects of the surgery [58]. This unexpected result has been interpreted as evidence that TRPC2-mediated signaling may be essential for gender discrimination [50]. One possible model consistent with these data is that mounting is an innate behavior that is inhibited by male pheromones acting through the VNO. TRPC2-/- males, therefore, persist in mounting other males. Because of the absence of major defects in male-female sexual behavior in TRPC2-/- mice, pheromones or other sensory cues essential for mating may not be detected by the VNO, but rather by other sensory systems such as the main olfactory epithelium. On the basis of these results, it is now clear that TRPC2 is essential for the detection of male-specific cues in the VNO that, in turn, regulate the expression of complex behavioral repertoires including aggressive and sexual behaviors.

6.11

Loss of VNO Signaling Components in Human Evolution

Whether humans have a functional VNO has until recently been a contentious question; a small pit is found in the nasal septum of most humans, but it has been difficult if not impossible to determine whether this pit contains functional VNO sensory neurons [60]. Recent evidence from molecular studies now strongly suggests that the human VNO is vestigial. Notably, of the several hundred VNO receptor genes in the human genome, nearly all are pseudogenes and only five contain intact reading frames [61]. Moreover, the TRPC2 gene, which is essential for VNO function in the mouse, is a pseudogene in humans [45, 62]. Thus, unless one is to hypothesize that a whole new set of genes is used for VNO transduction in humans, it is hard to escape the conclusion that the human VNO is vestigial. When did the human VNO become vestigial and why? Because the TRPC2 gene is expressed uniquely in the VNO and is essential for VNO function, the loss of a TRPC2 gene can serve as a marker for the loss of VNO function. Based on this reasoning, Liman and Innan [63] and Zhang and Webb [64] examined sequences of the TRPC2 gene from a large number of primate species. The human TRPC2 gene has six mutations that generate premature stop codons, resulting in a severely truncated protein. The earliest mutation is a nonsense mutation that is shared by all old world (OW) monkeys and apes and that is predicted to generate a protein that is missing much of its C-terminus [63, 64] (Fig. 6.8). Because this mutation occurs in a wellconserved region of the protein, it is likely to impair functioning. Thus, based on the observation that this mutation is found in all OW monkeys and apes but not in new world (NW) monkeys, we can date the loss of a functional TRPC2 gene in the human

147

148

6 Transduction Channels in the Vomeronasal Organ Fig. 6.8 The earliest mutation in the TRPC2 gene occurred in the common ancestor of OW monkeys and apes. (A) Sequence of the TRPC2 from monkeys was examined for the presence of stop codons or frameshift mutations. Mutations are indicated by the numbers 1–9, placed at the point in the phylogenetic tree where the mutation is inferred to have occurred. Note that mutation 9 was found in OW monkeys and gibbon but not in other apes, indicating that it either arose twice or that there was a reversion event (indicated by a white 9 on a black background). (B) Schematic representation of the TRPC2 gene, indicating the position of each mutation. Black bars represent transmembrane domains (From [63]).

lineage to 25–40 million years ago. This dating is further supported by an examination of selective pressure on the TRPC2 gene, which was also relaxed at this time [63]. If humans have lost a functioning VNO, how can we explain the presence of V1R genes with intact open reading frames (ORFs) in the human genome? Zhang and Webb [64] have considered the possibility that these genes are relics of an incomplete process of pseudogenization of this large family of receptors. Indeed, mathematical models show that approximately five receptors are expected to be intact if the initial repertoire was similar in size to that of the mouse (140 intact ORFs) and selective pressure was relaxed 23 million years ago. This timing is consistent with previous anatomical studies showing that a well-developed VNO and a morphologically distinct AOB were also lost in the common ancestor of OW monkeys and apes [65, 66]. Why did the ancestor of apes and OW monkeys lose a functioning VNO, while NW monkeys retained a functioning VNO? Interestingly, around the same time that the VNO was lost, the common ancestor of OW monkeys and apes acquired trichromatic vision through a duplication of the green opsin gene [67, 68]. Trichromatic vision, which employs three cone types, provides improved discrimination among colors from green to red and has therefore been postulated to have arisen as an adaptation to foraging for fruits or plants among a dense green canopy [68]. This enhanced visual capacity may also have proven useful for discriminating reproductive and social status

6.12 Summary: Is TRPC2 the VNO Transduction Channel?

among our distant ancestor, as seen in the highly colorful sexual skin of many OW monkeys and apes. Thus, vision that allowed discrimination at a distance may have replaced VNO-dependent pheromone signaling in mediating social interactions in humans.

6.12

Summary: Is TRPC2 the VNO Transduction Channel?

The evidence reviewed here suggests a model for VNO sensory transduction (Fig. 6.9) whereby binding of pheromones or other signaling molecules to V1R or V2R receptors leads to activation of a PLC that hydrolyzes PIP2 to IP3 and DAG. DAG then acts directly on the ion channel TRPC2, which allows an influx of Ca2+ and Na+ into the cell. Ca2+ may further act on Ca2+-activated nonselective channels that further amplify the electrical signal. The sensory current leads to a depolarization and the generation of actions potentials. The following evidence supports that TRPC2 is a major component of the VNO transduction channel: 1. It is in the right place. TRPC2 appears to be exclusively expressed in the VNO and the protein is restricted to the sensory microvilli [45]. 2. It is structurally related to other channels known to mediate sensory transduction in other systems. 3. Targeted deletion of TRPC2 largely eliminates sensory responses to pheromones [14, 50]. A possible concern with these experiments is that the loss of sensory responses is due to a nonspecific effect, such as cell death or developmental arrest. This does not appear to be the case, as although there is some loss of neurons in TRPC2-/- mice, the remaining neurons appear healthy and can fire action potentials in response to depolarization. 4. Targeted deletion of TRPC2 leads to the specific absence of a DAG-activated current [19]. With properties similar to that of the current activated in response to activation of the sensory signaling pathway.

2+

Na+

Ca

Pheromone

Na+

Ca2+

TRPC2 Fig. 6.9 Proposed mechanism of VNO sensory transduction. V1R and V2R receptors couple through Gai2 and Gao to a PLC, leading to hydrolysis of PIP2 into IP3 and DAG. DAG acts directly on the TRPC2 ion channel to allow an influx of Ca2+ and Na+

VR G

PLC

2

+++

DAG +

Na

IP3

Ca2+

149

150


To further confirm the role of TRPC2 in VNO sensory transduction, it will be important to see that sensory responses in TRPC2-/- mice can be rescued by expression of TRPC2 and to determine whether in a heterologous cell type, TRPC2 makes a DAGactivated channel.

Acknowledgements

We thank C. Dulac for critical reading of this chapter. The authors’ research is supported by grants from the National Institutes of Health/National Institute on Deafness and other Communication Disorders to E.R.L. and F.Z. References 1

2

3

4

5

6

7

8

9

10

Karlson, P. and M. Luscher, ’Pheromones’: a new term for a class of biologically active substances. Nature, 1959. 183: p. 55–56. Sam, M., et al., Neuropharmacology. Odorants may arouse instinctive behaviours. Nature, 2001. 412(6843): p. 142. Trinh, K. and D.R. Storm, Vomeronasal organ detects odorants in absence of signaling through main olfactory epithelium. Nat Neurosci, 2003. 6(5): p. 519–25. Halpern, M., The organization and function of the vomeronasal system. Annual review of Neuroscience, 1987. 10: p. 325–62. Dulac, C. and A.T. Torello, Molecular detection of pheromone signals in mammals: from genes to behaviour. Nat Rev Neurosci, 2003. 4(7): p. 551–62. Meredith, M., et al., Vomeronasal pump: significance for male hamster sexual behavior. Science, 1980. 207(4436): p. 1224–6. Luo, M., M.S. Fee, and L.C. Katz, Encoding pheromonal signals in the accessory olfactory bulb of behaving mice. Science, 2003. 299(5610): p. 1196–201. Martinez-Marcos, A. and M. Halpern, Differential projections from the anterior and posterior divisions of the accessory olfactory bulb to the medial amygdala in the opossum, Monodelphis domestica. Eur J Neurosci, 1999. 11(11): p. 3789–99. von Campenhausen, H. and K. Mori, Convergence of segregated pheromonal pathways from the accessory olfactory bulb to the cortex in the mouse. Eur J Neurosci, 2000. 12(1): p. 33–46. Moss, R.L., et al., Urine-derived compound evokes membrane responses in mouse vomeronasal receptor neurons. J Neurophysiol, 1997. 77(5): p. 2856–62.

11

12

13

14

15

16

17

18

19

20

Moss, R.L., et al., Electrophysiological and biochemical responses of mouse vomeronasal receptor cells to urine-derived compounds: possible mechanism of action. Chem Senses, 1998. 23(4): p. 483–9. Leinders-Zufall, T., et al., Ultrasensitive pheromone detection by mammalian vomeronasal neurons. Nature, 2000. 405(6788): p. 792–6. Holy, T.E., C. Dulac, and M. Meister, Responses of vomeronasal neurons to natural stimuli. Science, 2000. 289(5484): p. 1569–72. Leypold, B.G., et al., Altered sexual and social behaviors in trp2 mutant mice. Proc Natl Acad Sci U S A, 2002. 99(9): p. 6376–81. Del Punta, K., et al., Deficient pheromone responses in mice lacking a cluster of vomeronasal receptor genes. Nature, 2002. 419(6902): p. 70–4. Inamura, K., et al., Laminar distribution of pheromone-receptive neurons in rat vomeronasal epithelium. J Physiol (Lond), 1999. 517(Pt 3): p. 731–9. Cinelli, A.R., et al., Calcium transients in the garter snake vomeronasal organ. J Neurophysiol, 2002. 87(3): p. 1449–72. Spehr, M., H. Hatt, and C.H. Wetzel, Arachidonic acid plays a role in rat vomeronasal signal transduction. J Neurosci, 2002. 22(19): p. 8429–37. Lucas, P., et al., A Diacylglycerol-Gated Cation Channel in Vomeronasal Neuron Dendrites Is Impaired in TRPC2 Mutant Mice: Mechanism of Pheromone Transduction. Neuron, 2003. 40: p. 551–561. Liman, E.R. and D.P. Corey, Electrophysiological characterization of chemosensory neurons from the mouse vomeronasal organ. J Neurosci, 1996. 16(15): p. 4625–4637.

6.12 Summary: Is TRPC2 the VNO Transduction Channel? 21

22

23

24

25

26

27

28

29

30

31

32

33

34

Dulac, C. and R. Axel, A novel family of genes encoding putative pheromone receptors in mammals. Cell, 1995. 83(2): p. 195–206. Berghard, A., L.B. Buck, and E.R. Liman, Evidence for distinct signaling mechanisms in two mammalian olfactory sense organs. Proc Natl Acad Sci U S A, 1996. 93(6): p. 2365–9. Wu, Y., R. Tirindelli, and N.J. Ryba, Evidence for different chemosensory signal transduction pathways in olfactory and vomeronasal neurons. Biochem Biophys Res Commun, 1996. 220(3): p. 900–4. Rodriguez, I., et al., Multiple new and isolated families within the mouse superfamily of V1r vomeronasal receptors. Nat Neurosci, 2002. 5(2): p. 134–40. Pantages, E. and C. Dulac, A novel family of candidate pheromone receptors in mammals. Neuron, 2000. 28(3): p. 835–45. Herrada, G. and C. Dulac, A novel family of putative pheromone receptors in mammals with a topographically organized and sexually dimorphic distribution. Cell, 1997. 90(4): p. 763–73. Matsunami, H. and L.B. Buck, A multigene family encoding a diverse array of putative pheromone receptors in mammals. Cell, 1997. 90: p. 775–784. Ryba, N.J. and R. Tirindelli, A new multigene family of putative pheromone receptors. Neuron, 1997. 19(2): p. 371–9. Loconto, J., et al., Functional expression of murine V2R pheromone receptors involves selective association with the M10 and M1 families of MHC class Ib molecules. Cell, 2003. 112(5): p. 607–18. Boschat, C., et al., Pheromone detection mediated by a V1r vomeronasal receptor. 2002. 5(12): p. 1261–1262. Berghard, A. and L. Buck, Sensory transduction in vomeronasal neurons: Evidence for Gao, Gai, and adenylyl cyclase II as major components of a pheromone signalling cascade. J Neurosci, 1996. 16: p. 909–918. Halpern, M., L.S. Shapiro, and C. Jia, Differential localization of G proteins in the opossum vomeronasal system. Brain Research, 1995. 677(1): p. 157–61. Krieger, J., et al., Selective activation of G protein subtypes in the vomeronasal organ upon stimulation with urine-derived compounds. J Biol Chem, 1999. 274(8): p. 4655–62. Menco, B.P., et al., Ultrastructural localization of G-proteins and the channel protein TRP2 to microvilli of rat vomeronasal receptor cells. J Comp Neurol, 2001. 438(4): p. 468–89.

35

36

37

38

39

40

41

42

43

44

45

46

47

Norlin, E.M., F. Gussing, and A. Berghard, Vomeronasal phenotype and behavioral alterations in G alpha i2 mutant mice. Curr Biol, 2003. 13(14): p. 1214–9. Luo, Y., et al., Identification of chemoattractant receptors and G-proteins in the vomeronasal system of garter snakes. J Biol Chem, 1994. 269(24): p. 16867–77. Kroner, C., et al., Pheromone-Induced Second Messenger Signaling In the Hamster Vomeronasal Organ. Neuroreport, 1996. 7(18): p. 2989–2992. Wekesa, K.S. and R.R. Anholt, Pheromone regulated production of inositol-(1, 4, 5)-trisphosphate in the mammalian vomeronasal organ. Endocrinology, 1997. 138(8): p. 3497–504. Runnenburger, K., H. Breer, and I. Boekhoff, Selective G protein beta gamma-subunit compositions mediate phospholipase C activation in the vomeronasal organ. Eur J Cell Biol, 2002. 81(10): p. 539–47. Liman, E.R., Regulation by voltage and adenine nucleotides of a Ca2+-activated cation channel from hamster vomeronasal sensory neurons. J Physiol, 2003. 548(Pt 3): p. 777–87. Taniguchi, M., M. Kashiwayanagi, and K. Kurihara, Intracellular injection of inositol 1,4,5trisphosphate increases a conductance in membranes of turtle vomeronasal receptor neurons in the slice preparation. Neuroscience Letters, 1995. 188(1): p. 5–8. Inamura, K., M. Kashiwayanagi, and K. Kurihara, Inositol-1,4,5-trisphosphate induces responses in receptor neurons in rat vomeronasal sensory slices. Chemical Senses, 1997. 22(1): p. 93–103. Taniguchi, M., D. Wang, and M. Halpern, Chemosensitive conductance and inositol 1,4,5trisphosphate-induced conductance in snake vomeronasal receptor neurons. Chem Senses, 2000. 25(1): p. 67–76. Clapham, D.E., L.W. Runnels, and C. Strubing, The TRP ion channel family. Nat Rev Neurosci, 2001. 2(6): p. 387–96. Liman, E.R., D.P. Corey, and C. Dulac, TRP2: a candidate transduction channel for mammalian pheromone sensory signaling. Proc Natl Acad Sci U S A, 1999. 96(10): p. 5791–6. MacKinnon, R., Determination of the subunit stoichiometry of a voltage-activated potassium channel. Nature, 1991. 350(6315): p. 232–5. Liman, E.R., J. Tytgat, and P. Hess, Subunit stoichiometry of a mammalian K+ channel determined by construction of multimeric cDNAs. Neuron, 1992. 9(5): p. 861–71.

151

152

6 Transduction Channels in the Vomeronasal Organ 48

49

50

51

52

53

54

55

56

57

58

59

60

61

Vannier, B., et al., Mouse trp2, the homologue of the human trpc2 pseudogene, encodes mTrp2, a store depletion-activated capacitative Ca2+ entry channel. Proc Natl Acad Sci U S A, 1999. 96(5): p. 2060–4. Hofmann, T., et al., Cloning, expression and subcellular localization of two novel splice variants of mouse transient receptor potential channel 2. Biochem J, 2000. 351(Pt 1): p. 115–22. Stowers, L., et al., Loss of sex discrimination and male-male aggression in mice deficient for TRP2. Science, 2002. 295(5559): p. 1493–500. Hardie, R.C. and P. Raghu, Visual transduction in Drosophila. Nature, 2001. 413(6852): p. 186–93. Chyb, S., P. Raghu, and R.C. Hardie, Polyunsaturated fatty acids activate the Drosophila light-sensitive channels TRP and TRPL. Nature, 1999. 397(6716): p. 255–9. Hardie, R.C., Regulation of TRP channels via lipid second messengers. Annu Rev Physiol, 2003. 65: p. 735–59. Hofmann, T., et al., Direct activation of human TRPC6 and TRPC3 channels by diacylglycerol. Nature, 1999. 397(6716): p. 259–63. Kiselyov, K., et al., Functional interaction between InsP3 receptors and store-operated Htrp3 channels. Nature, 1998. 396(6710): p. 478–82. Yuan, J.P., et al., Homer binds TRPC family channels and is required for gating of TRPC1 by IP3 receptors. Cell, 2003. 114(6): p. 777–89. Brann, J.H., et al., Type-specific inositol 1,4,5trisphosphate receptor localization in the vomeronasal organ and its interaction with a transient receptor potential channel, TRPC2. J Neurochem, 2002. 83(6): p. 1452–60. Wysocki, C.J. and J.J. Lepri, Consequences of removing the vomeronasal organ. J Steroid Biochem Mol Biol, 1991. 39(4B): p. 661–9. Jungnickel, M.K., et al., Trp2 regulates entry of Ca2+ into mouse sperm triggered by egg ZP3. Nat Cell Biol, 2001. 3(5): p. 499–502. Meredith, M., Human vomeronasal organ function: a critical review of best and worst cases. Chem Senses, 2001. 26(4): p. 433–45. Rodriguez, I. and P. Mombaerts, Novel human vomeronasal receptor-like genes reveal speciesspecific families. Curr Biol, 2002. 12(12): p. R409–11.

62

63

64

65

66

67

68

69

70

71

72

73

Wes, P.D., et al., TRPC1, a human homolog of a Drosophila store-operated channel. Proc Natl Acad Sci U S A, 1995. 92: p. 9652–9656. Liman, E.R. and H. Innan, Relaxed selective pressure on an essential component of pheromone transduction in primate evolution. Proc Natl Acad Sci U S A, 2003. 100(6): p. 3328–32. Zhang, J. and D.M. Webb, Evolutionary deterioration of the vomeronasal pheromone transduction pathway in catarrhine primates. Proc Natl Acad Sci U S A, 2003. 100(14): p. 8337–41. Bhatnagar, K.P. and E. Meisami, Vomeronasal organ in bats and primates: extremes of structural variability and its phylogenetic implications. Microsc Res Tech, 1998. 43(6): p. 465–75. Meisami, E. and K.P. Bhatnagar, Structure and diversity in mammalian accessory olfactory bulb. Microsc Res Tech, 1998. 43(6): p. 476–99. Hunt, D.M., et al., Molecular evolution of trichromacy in primates. Vision Res, 1998. 38(21): p. 3299–306. Surridge, A.K., D. Osorio, and N.I. Mundy, Evolution and selection of trichromatic vision in primate. Trends in Ecology and Evolution, 2003. 18(4): p. 198–205. Liman, E.R., Pheromone transduction in the vomeronasal organ. Current Opinion in Neurobiology, 1996. 6: p. 487–493. Liu, D., and E.R. Liman, Intracellular Ca2+ and the phospholipid PIP2 regulate the taste transduction ion channel TRPM5. Proc Natl Acad Sci USA, 2003. 100: p. 15160–15165. Prawitt, D. et al., TRPM5 is a transient Ca2+activated cation channel responding to rapid changes in [Ca2+]i. Proc Natl Acad Sci USA, 2003. 100: p. 15166–15171. Launay, P. et al., TRPM4 is a Ca2+-activated nonselective cation channel mediating cell membrane depolarization. Cell, 2002. 109: p. 397–407. Hofmann, T., et al., TRPM5 is a voltage-modulated and Ca(2+)-activated monovalent selective cation channel. Curr Biol, 2003. 13: p. 1153–1158.

153

7

Transduction Mechanisms in Taste Cells Kathryn Medler and Sue C. Kinnamon

7.1

Introduction

All organisms respond to chemicals found in their external environment. These responses detect nutrients, conspecifics, predators, or potentially harmful conditions. The ability to discriminate items with nutritional value from items that are harmful depends on chemical detectors called taste receptor cells. Taste receptor cells are specialized neuroepithelial cells that are housed as taste buds in the lingual epithelium of the oral cavity. Taste buds consist of 50 to 150 taste receptor cells that extend apical processes into the oral cavity to detect stimuli and form synapses with gustatory neurons to send information to the brain (Fig. 7.1). Taste cells have properties of both neurons and epithelial cells and are responsible for converting chemical stimuli into electrical signals. Like neurons, they are capable of generating receptor and/or action potentials and can form classical chemical synapses with gustatory neurons [1–4]. Taste cells express TTX-sensitive voltage-gated Na+ channels and multiple types of voltage-gated K+ and Ca2+ channels [5–7]. However, like epithelial cells, taste cells express several types of epithelial ion channels and have a limited life span of about two weeks. As a result of turnover, gustatory neurons must disconnect from dying taste cells to form new synaptic contacts with emerging taste cells [1, 4]. In mammals, taste buds are found in specialized protrusions called papillae. These papillae consist of fungiform, foliate, and circumvallate on the tongue, with additional taste buds present on the soft palate, scattered throughout the oral cavity and in the “geschmacksstreifen,” a row of taste buds between the hard and soft palate. Taste buds are innervated by the facial, (VII), glossopharyngeal (IX), or vagus (X) cranial nerves, depending on their location in the tongue. A branch of the facial nerve, the chorda tympani, innervates the fungiform papillae and the anterior portion of the foliate papillae, while the glossopharyngeal nerve innervates the rest of the foliate and the circumvallate papillae. The greater superficial petrosal branch of the facial nerve innervates the taste buds found in the soft palate and geschmacksstreifen [8, 9]. Taste receptor cells detect a wide range of chemically diverse stimuli that comprise the different taste qualities: sweet, salty, sour, bitter, and umami, the detection of Transduction Channels in Sensory Cells. Edited by S. Frings and J. Bradley Copyright ª 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 3-527-30836-9

154

7 Transduction Mechanisms in Taste Cells

Fig. 7.1 Morphology of the peripheral taste system. The upper panel shows a schematic of a mammalian tongue, showing fungiform, circumvallate, and foliate papillae and their innervation. A taste bud is illustrated in the bottom panel. The taste pore is the site of interaction with taste stimuli, while afferent gustatory neurons carry information from the taste cells to the brain

amino acids, primarily glutamate. These taste qualities can be divided into two broader categories, ionic and complex, based on the chemical structure of the taste stimulus. Ionic stimuli include Na+ and H+, which elicit salty and sour tastes, respectively. These ions can interact directly with apically located ion channels to depolarize taste cells, by either permeating or blocking the channel, or they can diffuse through tight junctions to interact with basolateral ion channels. More chemically complex stimuli, including sugars, amino acids, and most bitter stimuli, interact with apically located G-proteincoupled receptors (GPCR) or, in some cases, ligand-gated channels. Because taste is tied to the nutritional needs of an organism, there is considerable species diversity in transduction mechanisms. This necessitates multiple mechanisms to detect different stimuli. In this chapter, we will describe both ion channels directly modulated by taste stimuli and ion-channel targets of second messengers generated as a result of GPCR signaling. When possible, we will discuss their role in taste trans-

7.2 Ionic Stimuli

duction. Describing all the channels that have been proposed to mediate transduction in different species is beyond the scope of this review. Instead, we will focus on the best-characterized channels, for which there is either direct molecular or physiological evidence for their role in transduction.

7.2

Ionic Stimuli

Two taste qualities, salty and sour, detect the presence of ions in the oral cavity. Due to their simple structure and charge, ions are capable of passing through ion channels on the apical membrane of taste cells to directly depolarize taste cells. Since ions can also pass through tight junctions between taste cells, they are able to interact at the basolateral membrane in addition to the apical pore.

7.2.1

Salt

Salt taste is the detection of cations, such as Na+, K+, and Li+. This leads to the ingestion of NaCl and other required minerals, thus helping to maintain ion and water homeostasis. The specific ionic needs vary across species depending on the ionic content of their diet and surrounding environment. When NaCl is present in sufficient levels in the oral cavity, Na+ permeates apically located ion channels, causing membrane depolarization and activation of voltage-gated Na+ and K+ channels. In addition to interacting with apically located channels, NaCl and other monovalent salts diffuse passively across tight junctions, where they can potentially interact with basolateral channels. In addition, this paracellular shunt pathway generates a hyperpolarizing field potential around the taste cells that depresses the receptor potential and makes it less likely to fire action potentials. The field potential is influenced by the anion present. For example, the chloride ion (Cl–) of NaCl is very permeable and thus provides a shunting current that minimizes the hyperpolarizing field potential, while less conductive anions do not. This causes a difference in their saltiness perception as compared to NaCl [10, 11]. Studies to determine the role of the anion in salt transduction have found no evidence of Cl– co-transporters, exchangers, or channels at the apical membrane, further supporting the idea that anions influence salt perception via the paracellular pathway [12]. Epithelial Sodium Channel Salt taste transduction can be divided into an amiloride-sensitive component and an amiloride-insensitive component. Amiloride is a diuretic that blocks a resting Na+ conductance found in many species. When the channel was first characterized in taste cells, it was initially called the amiloride-sensitive sodium channel (ASSC), but it is now generally accepted that this channel is the epithelial sodium channel (ENaC), 7.2.1.1

155

156


a member of the degenerin family of channels that is widely expressed in epithelial tissues throughout the body [13]. The ENaC is a heteromultimer consisting of a-, b-, and c-ENaC subunits (Fig. 7.2) [14]. In expression systems, the a subunits are sufficient to induce channel activity, but the b and c subunits are needed to confer maximal Na+ current [15]. However, in taste cells, it is thought that all three subunits are needed to form a functional channel. Immunocytochemical analyses in rat taste cells demonstrate labeling in nearly all the fungiform papillae and in about half of the foliate and vallate papillae. The labeling intensity is significantly lower in the vallate papillae compared to the fungiform, especially for the b and c subunits [16–18]. Interestingly, treatment with aldosterone increases the apical immunoreactivity of the b and c ENaC subunits in all papillae and increases both the number of amiloride-sensitive taste cells and the amplitude of the response. Additionally, aldosterone treatment induces an amiloride-sensitive current in about half the vallate taste cells, which are normally unresponsive to amiloride. The upregulation of b and c ENaC subunits by aldosterone seems to induce or increase the channel activity [16], which is consistent with its role in regulating ENaCs in other tissues. Several hormones that are involved in osmotic regulation and Na+ balance have been found to regulate ENaC expression. Like aldosterone, arginine vasopressin (AVP) increases Na+ transport in taste tissue. AVP is thought to act on ENaCs via the V2-type vasopressin receptors, as its effects are mimicked by cAMP. These hormones increase the number of ENaCs expressed at the apical membrane, presumably through a receptor-mediated increase in intracellular cAMP levels [19]. Meanwhile, both atrial natriuretic peptide and oxytocin cause decreases in Na+ transport in lingual epithelia. These data suggest that hormones may regulate the sensitivity of salt detection [20]. In taste cells, the ENaC is constitutively active and has a very high affinity for amiloride, with a Ki in the sub-micromolar concentration range. The amiloride dose-response curve can be fitted with a single binding isotherm, indicating a possible 1:1 stoichiometry for the drug-channel interaction [18]. In rat taste cells, it has been shown that

Fig. 7.2 Membrane topology of each subunit (a, b, and c) ENaC. M1 and M2 indicate helical transmembrane domains; CRD1 and CRD2 indicate cysteine-rich domains; glycosylation sites are indicated (6 on a-ENaC, 12 on b-ENaC; and 5 on

c-ENaC). The filled bar within CRD1 of a-ENaC indicates the relative position of a six-amino-acid region known to bind amiloride. Reprinted from [14] with permission

7.2 Ionic Stimuli

Na+ inhibits its own influx through the ENaC. This Na+ self-inhibition is restricted to the amiloride-sensitive response and is slow, taking up to 15 s to become maximally inhibited. It is thought that Na+ binds the ENaC receptor on its extracellular face and reduces the Na+ permeability of the channel. While its role is not completely understood, Na+ self-inhibition may be involved in adaptation to Na+ salts or in the regulation of salt intake [21, 22]. The amiloride-sensitive conductance is believed to be the main contributor to salt transduction in animals that have inherently low salt diets. Studies in hamsters found that over half of all taste cells had some ENaC activity and that within the amiloridesensitive cells, ENaCs accounted for an average of 48 % of the total current generated (Fig. 7.3) [23]. This is true for most rodent species, while in humans the amilorideinsensitive response is predominant, with a relatively minor amiloride-sensitive response. However, to some degree, the salt response in all species studied is comprised of both amiloride-sensitive and -insensitive components. Aquatic animals, especially freshwater varieties, have a very strong driving force to maintain ionic homeostasis. Despite the different environmental pressures on the maintenance of water and ion balance in these animals, there are similarities in their salt transduction mechanisms, specifically the role of the ENaC. In frogs, more than 50 % of taste cells had a stationary inward Na+ current at –80 mV that was partially or fully blocked by amiloride. This conductance was slightly modulated by voltage but was not voltage gated [24]. However, there are some differences in the ENaC expressed in frog taste cells when compared with ENaCs expressed in other systems. The unitary conductance of the ENaC in frog skin epithelia is 6–10 pS compared to less than 2 pS for the taste ENaC. Additionally, the selectivity ratio for Na+ to K+ is less for the frog

Fig. 7.3 Amiloride inhibits the NaCl-induced currents. (A) Short circuit current resulting from NaCl solutions presented to rat tongue epithelium in the absence and presence of 10–4 M amiloride. (B) Time course of the open-circuit potential after

0.001 M NaCl was replaced with 1.5 M NaCl in the presence and absence of 10–4 M amiloride. Reprinted Fig. 1 with permission from Heck et al. (1984) Science 223:403–405 (Copyright 1984, AAAS)

157

158


taste channel than for ENaCs in mammalian taste cells and ENaCs in other tissues. Based on reversal potentials, the selectivity of this channel in frogs was K+>Na+>Li+>Cs+ [25]. Amiloride-insensitive Pathway The amiloride-insensitive pathway is not well characterized and is likely comprised of multiple mechanisms. One proposed mechanism is that the ENaC and/or nonselective cation channels on the basolateral membrane are permeable to Na+ that has diffused through tight junctions. Another possible mechanism is that an amiloride-insensitive Na+ channel is expressed on the apical membrane of taste cells. Physiological evidence supports both mechanisms, although mechanisms vary considerably in different species. Although the ENaC is normally restricted to the apical membrane in most transporting epithelia, basolateral expression has been proposed to mediate a portion of the amiloride-insensitive salt response in taste cells. Since amiloride cannot pass through tight junctions, it would not inhibit these channels. Voltage-clamp studies in rat found amiloride-sensitive channels on the basolateral membrane, though their properties differ from the apically located channels. Basolateral channels are less sensitive to amiloride, with a Ki of 52 lM compared with a Ki of 0.2 lM for apically located channels. ENaCs at the apical membrane are highly selective for Na+, with a Na+/ Cs+ permeability ratio greater than 10, while the channels at the basolateral membrane are not as selective for Na+ (PNa/PK =3.7) [26]. Two types of amiloride-insensitive responses have been characterized in mice in addition to the amiloride-sensitive response. The first response has a slight inward rectification, has a reversal potential close to the equilibrium potential for nonselective cation channels, and is blocked by Cd2+. The other response shows no rectification and has a reversal potential close to the equilibrium potential for Cl– at the basolateral 7.2.1.2

Fig. 7.4 The amiloride-insensitive response can be either enhanced or inhibited by cetylpyridinium chloride (CPC). (A) Effect of 250 lM CPC on the integrated chorda tympani response to 300 mM NaCl and 300 mM NaCl + 100 lM amiloride. CPC enhanced the response by the same magni-

tude in each case. (B) Effect of 2 mM CPC on the integrated chorda tympani response to 300 mM NaCl and 300 mM NaCl + 100 lM amiloride. CPC suppressed the entire amiloride-insensitive part of the response. Reprinted from [28] with permission

7.2 Ionic Stimuli

membrane, but not at the apical membrane. The Cl– channel blocker, NPPB, potentiates the salt response and abolishes the low reversal potential, indicating that NPPB blocks a Cl– channel. These data suggest that at least three components contribute to the salt response in mice: the amiloride-sensitive Na+ channel, an NPPB-sensitive Cl– channel, and a Cd2+-sensitive nonselective cation channel. The amiloride-sensitive Na+ channel and the nonselective cation channel are located at the apical membrane, while the Cl– channel is located on the basolateral membrane [27]. Evidence for apically located, amiloride-insensitive Na+ channels in rats comes from studies using a modulator, cetylpyridium chloride (CPC). At low concentrations, CPC enhances the amiloride-insensitive Na+ response, but at high concentrations the response is suppressed (Fig. 7.4). CPC also modulates responses to KCl and NH4Cl, suggesting that the channel may be nonselective for cations. Since this response is modulated by voltage fields applied across the epithelium, it is likely mediated by an apical conductance [28]. Further work will be required to characterize the properties of this conductance. In the frog, channels other than the ENaC are involved in transducing salt stimuli. Both Na+ and Cl– permeate apical channels believed to be gated by NaCl, leading to the generation of receptor potentials. This is in addition to cationic channels present on the basolateral membrane that also contribute to generation of the response to salty stimuli [29]. There is also evidence for two Na+-dependent K+ conductances in frog taste cells. Two different conductances, 35.8 pS and 9.4 pS, are activated by high internal NaCl. It is thought that these channels contribute to the stabilization of the resting potential and that they may be involved in salt transduction [30].

7.2.2

Sour

Sour taste is the detection of protons in the oral cavity. This taste quality serves to detect unripe or spoiled food in order to avoid ingestion of potentially harmful levels of acids. Multiple receptors have been proposed to serve as sour receptors, but to date, none have emerged as definitive. Because protons modulate most ionic conductances, nearly all taste cells respond to acid stimulation with changes in membrane conductance [31] and intracellular acidification [32]. However, it is not clear that all of these cells communicate this information to the nervous system. In slice preparations from mice, focal applications of acidic taste stimuli result in the acidification of the entire taste bud, but only a few taste cells also respond with increases in intracellular Ca2+. Cells showing acid-induced increases in intracellular Ca2+ also respond to KCl depolarization, and responses are blocked by Ba2+ and Cd2+, suggesting influx via voltagegated Ca2+ channels (Fig. 7.5) [33]. It seems probable that the taste cells expressing voltage-gated Ca2+ channels are the taste cells that generate signals to the nervous system in response to sour stimuli. Sour perception is affected by both extracellular pH (pHo), and the accompanying anions. At the same pH, weak acids produce larger nerve responses and a greater sour taste sensation when compared to strong acids [34]. Lyall et al. [32] hypothesize that

159

160

7 Transduction Mechanisms in Taste Cells Fig. 7.5 Ca2+ response to citric acid depends on Ca2+ influx through voltage-gated Ca2+ channels. (A) Superimposed Ca2+ responses in a mouse taste cell stimulated repeatedly with citric acid (pH 3, arrowheads) either before or after 2 mM Ba2+ in the bath. Mean (SEM) amplitudes of Ca2+ responses to citric acid shown in the bar graphs (*P

Transduction Channels in Sensory Cells

TRP Ion Channel Function in Sensory Transduction and Cellular Signaling Cascades (Frontiers in Neuroscience)

Histidine Kinases in Signal Transduction

Sensory Marketing

Signal Transduction

Sensory Worlds in Early America

Signal Transduction

Molecular Mechanisms in Visual Transduction

Signal Transduction

Magnetofluiddynamics in channels and containers

Signal Transduction in the Retina

Frank Signal Transduction in Cancer

Magnetofluiddynamics in Channels and Containers

Sensory Processing in Aquatic Environments

Sensory Processing in Aquatic Environments

Sensory Changes in Later Life

Sensory Mechanisms

Dendritic Cells in Clinics

Dendritic Cells in Cancer

Methods in Insect Sensory Neuroscience

Sensory Systems

Channels Destiny

Drosophila Cells in Culture

Receptor Signal Transduction Protocols

Receptor Signal Transduction Protocols

Signal Transduction, Second Edition

Signal Transduction Protocols

Receptor Signal Transduction Protocols

Signal Transduction Protocols

Extra-Sensory Perception

Doing Sensory Ethnography

Transduction Channels in Sensory Cells

TRP Ion Channel Function in Sensory Transduction and Cellular Signaling Cascades (Frontiers in Neuroscience)

Histidine Kinases in Signal Transduction

Sensory Marketing

Signal Transduction

Sensory Worlds in Early America

Signal Transduction

Molecular Mechanisms in Visual Transduction

Signal Transduction

Magnetofluiddynamics in channels and containers

Signal Transduction in the Retina

Frank Signal Transduction in Cancer

Magnetofluiddynamics in Channels and Containers

Sensory Processing in Aquatic Environments

Sensory Processing in Aquatic Environments

Sensory Changes in Later Life

Sensory Mechanisms

Dendritic Cells in Clinics

Dendritic Cells in Cancer

Methods in Insect Sensory Neuroscience

Sensory Systems

Channels Destiny

Drosophila Cells in Culture

Receptor Signal Transduction Protocols

Receptor Signal Transduction Protocols

Signal Transduction, Second Edition

Signal Transduction Protocols

Receptor Signal Transduction Protocols

Signal Transduction Protocols

Extra-Sensory Perception

Doing Sensory Ethnography

Recommend Documents