Connexins: A Guide

Connexins Andrew L. Harris Editors Connexins A Guide 13 l Darren Locke Editors Andrew L. Harris New Jersey Medi...

Author: Andrew Harris | Darren Locke

227 downloads 786 Views 10MB 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

Connexins

Andrew L. Harris Editors

Connexins A Guide

13

l

Darren Locke

Editors Andrew L. Harris New Jersey Medical School UMDNJ Newark, NJ, USA [email protected]

ISBN: 978-1-934115-46-6 DOI 10.1007/978-1-59745-489-6

Darren Locke New Jersey Medical School UMDNJ Newark, NJ, USA [email protected]

e-ISBN: 978-1-59745-489-6

Library of Congress Control Number: 2008937815 # Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer ScienceþBusiness Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper springer.com

Editors’ Note

This monograph is intended to be a reference volume and a textbook on the connexin family of proteins. It is hoped that it will be informative to scientists wishing to learn about the field as well as to those who are active researchers in the field. It is organized and edited with the intent of addressing two sets of needs. Often scientists outside the field who are quite interested in specific aspects of connexin channel biology and function find it difficult to access the information they desire. Also, because the connexin field is uncommonly diverse, investigators within the field employ experimental methods and conceptual frameworks unfamiliar to other investigators, inside and outside the field. This book is designed to address the needs of both groups. The first section addresses the fundamentals of connexin biology, including genetics, structure, biophysics, electrophysiology, and several aspects of cell biology. The second section addresses the roles of connexin channels in specific organ systems or processes. Though each chapter is by different authors or sets of authors, the chapters have been edited and annotated to be cohesive, yet retain the perspective and tone of the original authors. There is some overlap of content, in places where it was integral to points being made by different author(s). In editing this volume for consistency, we have also had to make decisions regarding terminology and formatting: 1. There are contradictory, highly nuanced uses of the term hemichannel in the literature. Thus, for clarity in this volume we have chosen to use the term to refer to all connexin hexamers, functional or not, in junctional channels or not. The terms hemichannel and connexon are often used interchangeably. Some investigators use the term connexon to refer to the hexameric structure when it is part of a junctional channel, and hemichannel when it is not. Some use the same terms with the meanings reversed, arguing that a channel in a plasma membrane is not a half-channel and should therefore not be called a hemichannel. Others, with greater historical justification, make a distinction based on function, with connexon referring to all hexameric structures (as it was

v

vi

Editors’ Note

originally defined structurally [1,2]), and a hemichannel being a hexamer that is functional, whether or not it is part of a junctional channel. A variant of this is to refer to all hexameric forms as connexons, except when they are functional and unapposed in the plasma membrane. In this view, an unapposed hexamer in the plasma membrane is a hemichannel, and as a trafficking intermediate or in a junctional channel it is a connexon. It remains for others to resolve these issues. 2. Throughout this volume we have chosen to use the format for genes and proteins that seems to be in the most common use, as described in Chapter 1. The format is to represent the genes in italics, human genes and proteins in all capital letters (e.g., GJA1 and CX43) and genes for other species with only the first letter capitalized (e.g., Gja1 and Cx43). Where the genes and proteins are referred to generically, the latter format is used. There has been an ongoing discussion in the connexin community regarding nomenclature and formatting for the connexin genes and proteins. The issues are described in Chapter 1 and the Appendix. Finally, a single volume can never cover all the topics in a field, and that is true here. Several topics have not received separate chapters in this volume, though they are occasionally referred to. These topics include the role of connexins in the immune system [3,4,5], in hematopoiesis [6,7,8], and the possible nonchannel functions of connexin protein [9,10].

References 1. Goodenough, DA. Methods for the isolation and structural characterization of hepatocyte gap junctions. In: Korn ED, editor. Methods in Membrane Biology, vol 3. New York: Plenum, 1975. 2. Caspar DL, Goodenough DA, Makowski L, Phillips WC. Gap junction structures I. Correlated electron microscopy and X-ray diffraction. J Cell Biol. 1977;74:605–28. 3. Neijssen J, Pang B, Neefjes J. Gap junction-mediated intercellular communication in the immune system. Prog Biophys Mol Biol. 2007;94:207–18. 4. Oviedo-Orta E, Howard Evans W. Gap junctions and connexin-mediated communication in the immune system. Biochim Biophys Acta. 2004;1662:102–12. 5. Sa´ez JC, Bran˜es MC, Corvala´n LA, Eugenı´ n EA, Gonza´lez H, Martı´ nez AD, Palisson F. Gap junctions in cells of the immune system: structure, regulation and possible functional roles. Braz J Med Biol Res. 2000;33: 447–55. 6. Montecino-Rodriguez E, Dorshkind K. Regulation of hematopoiesis by gap junctionmediated intercellular communication. J Leukoc Biol. 2001;70:341–7. 7. Ploemacher RE, Mayen AE, De Koning AE, Krenacs T, Rosendaal M. Hematopoiesis: gap junction intercellular communication is likely to be involved in regulation of stromadependent proliferation of hemopoietic stem cells. Hematology 2000;5:133–47. 8. Rosendaal M, Krena´cs TT. Regulatory pathways in blood-forming tissue with particular reference to gap junctional communication. Pathol Oncol Res. 2000;6:243–249. 9. Wei CJ, Xu X, Lo CW. Connexins and cell signaling in development and disease. Annu Rev Cell Dev Biol. 2004;20:811–838. 10. Kalra J, Shao Q, Qin H, Thomas T, Alaoui-Jamali MA, Laird DW. Cx26 inhibits breast MDA-MB-435 cell tumorigenic properties by a gap junctional intercellular communication-independent mechanism. Carcinogenesis. 2006;27:2528–37.

Contents

Editor’s Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

Fundamentals of Connexin Biology . . . . . . . . . . . . . . . . . . . . . . .

1

1

The Family of Connexin Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eric C. Beyer and Viviana M. Berthoud

3

2

Gap Junction Channel Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mark Yeager

27

3

The Connexin Channel Pore: Pore-Lining Segments and Residues . . Vytas K. Verselis

77

4

Voltage-Gating Mechanisms of Connexin Channels. . . . . . . . . . . . . . Thaddeus Bargiello and Peter Brink

103

5

Chemical Gating of Connexin Channels . . . . . . . . . . . . . . . . . . . . . . . Rebecca Lewandowski, Junko Shibayama, Eva M. Oxford, Rosy Joshi-Mukherjee, Wanda Coombs, Paul L. Sorgen, Steven M. Taffet and Mario Delmar

129

6

Electrical Signaling with Neuronal Gap Junctions . . . . . . . . . . . . . . . Barry W. Connors

143

7

Permeability of Connexin Channels . . . . . . . . . . . . . . . . . . . . . . . . . . Andrew L. Harris and Darren Locke

165

8

Pharmacology of Connexin Channels. . . . . . . . . . . . . . . . . . . . . . . . . Miduturu Srinivas

207

Part I

vii

viii

Contents

9

Biogenesis and Degradation of Gap Junctions . . . . . . . . . . . . . . . . . . Linda S. Musil

225

10

Gap Junction Morphology and Dynamics in Situ . . . . . . . . . . . . . . . . Gina E. Sosinsky, Guido M. Gaietta and Ben N.G. Giepmans

241

11

Biochemistry of Connexins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Joell L. Solan and Paul D. Lampe

263

12

Pannexins or Connexins? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gerhard Dahl and Andrew L. Harris

287

Connexins in Organ Systems and Processes . . . . . . . . . . . . . . . .

303

13

Foreword: Gap Junctions and Emergent Organ Properties . . . . . . . . Daniel Goodenough

305

14

Connexins in Skin Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trond Aasen and David P. Kelsell

307

15

Connexins in the Nervous System. . . . . . . . . . . . . . . . . . . . . . . . . . . . Charles K. Abrams and John E. Rash

323

16

Connexins in the Respiratory Epithelium . . . . . . . . . . . . . . . . . . . . . . Bernard Foglia, Isabelle Scerri, Tecla Dudez and Marc Chanson

359

17

Connexins in Skeletal Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Roberto Civitelli and Henry J. Donahue

371

18

Connexins in Lens Development and Disease . . . . . . . . . . . . . . . . . . . Teresa I. Shakespeare, Richard T. Mathias and Thomas W. White

387

19

Connexins in the Mammalian Retina . . . . . . . . . . . . . . . . . . . . . . . . . Stephen C. Massey

397

20

Connexins in the Inner Ear. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regina Nickel, Andrew Forge and Daniel Jagger

419

21

Connexins in the Heart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nicholas J. Severs

435

22

Connexins in the Vasculature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cor de Wit and Stephanie E. Wolfle ¨

457

Part II

Contents

ix

23

Connexins and Atherosclerosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anna Pfenniger, Isabelle Roth and Brenda R. Kwak

469

24

Connexins in the Female Reproductive System. . . . . . . . . . . . . . . . . . Gerald M. Kidder and Elke Winterhager

481

25

Connexins in the Male Reproductive System . . . . . . . . . . . . . . . . . . . 495 Georges Pointis, Ce´line Fiorini, Je´rome Gilleron, Diane Carette and Dominique Segretain

26

Connexins and Secretion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sabine Bavamian, Philippe Klee, Florent Allagnat, Jacques-Antoine Haefliger and Paolo Meda

27

Connexins and Carcinogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529 Sophie Crespin, Norah Defamie, Laurent Cronier and Marc Mesnil

511

Appendix: Toward a New Nomenclature for Connexin Genes . . . . . . . . . . Gerald M. Kidder

543

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

547

Contributors

Trond Aasen, Ph.D. Center of Regenerative Medicine of Barcelona, Barcelona, Spain. Charles K. Abrams, M.D., Ph.D. Departments of Neurology and Physiology and Pharmacology & Program in Neurosciences, State University Of New York Downstate Medical Center, Brooklyn, United States. Florent Allagnat, Ph.D. Department of Internal Medicine, University Hospital, Lausanne, Switzerland. Thaddeus Bargiello, Ph.D. Department of Neuroscience, Albert Einstein College of Medicine, Bronx, United States. Sabine Bavamian, Ph.D. Department of Cell Physiology and Metabolism, University of Geneva, Geneva, Switzerland. Viviana M. Berthoud, Ph.D. Department of Pediatrics, University of Chicago, Chicago, United States. Eric C. Beyer, M.D., Ph.D. Department of Pediatrics, University of Chicago, Chicago, United States. Peter Brink, Ph.D. Department of Physiology and Biophysics, State University Of New York, Stony Brook, United States. Diane Carette, Ph.D. Institut National de la Sante´ et de la Recherche Me´dicale (INSERM), Nice, France.

xi

xii

Contributors

Marc Chanson, Ph.D. Department of Pediatrics, Geneva University Hospitals, Geneva, Switzerland. Roberto Civitelli, M.D. Division of Bone and Mineral Diseases & Departments of Internal Medicine and Cell Biology and Physiology, Washington University, St Louis, United States. Barry W. Connors, Ph.D. Department of Neuroscience, Brown University, Providence, United States. Wanda Coombs, Ph.D. Department of Pharmacology, State University Of New York Upstate Medical University, Syracuse, United States. Sophie Crespin, Ph.D. Institute of Physiology and Cell Biology, University of Poitiers, Poitiers, France. Laurent Cronier, Ph.D. Institute of Physiology and Cell Biology, University of Poitiers, Poitiers, France. Gerhard Dahl, M.D. Department of Physiology and Biophysics, University of Miami School of Medicine, Miami, United States. Norah Defamie, Ph.D. Institute of Physiology and Cell Biology, University of Poitiers, Poitiers, France. Mario Delmar, M.D., Ph.D. Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, United States. Henry J. Donahue, Ph.D. Division of Musculoskeletal Sciences, Department of Orthopaedics and Rehabilitation, Pennsylvania State University College of Medicine, Hershey, United States. Tecla Dudez, Ph.D. Department of Pediatrics, Geneva University Hospitals, Geneva, Switzerland. Ce´line Fiorini, Ph.D. Institut National de la Sante´ et de la Recherche Me´dicale (INSERM), Paris, France. Bernard Foglia, Ph.D. Department of Pediatrics, Geneva University Hospitals, Geneva, Switzerland. Andrew Forge, Ph.D. Centre for Auditory Research, University College London Ear Institute, London, United Kingdom.

Contributors

xiii

Guido M. Gaietta, Ph.D. National Center for Microscopy and Imaging Research, University of California San Diego, La Jolla, United States. Ben N.G. Giepmans, Ph.D. Molecular Imaging and Electron Microscopy, Department of Cell Biology, University Medical Center, Groningen, The Netherlands. Je´rome Gilleron, Ph.D. Institut National de la Sante´ et de la Recherche Me´dicale (INSERM), Paris, France. Daniel Goodenough, Ph.D. Department of Cell Biology, Harvard Medical School, Boston, United States. Jacques-Antoine Haefliger, Ph.D. Department of Internal Medicine, University Hospital, Lausanne, Switzerland. Andrew L. Harris, Ph.D. Department of Pharmacology and Physiology, New Jersey Medical School of the University of Medicine and Dentistry of New Jersey, Newark, New Jersey, United States. Daniel Jagger, Ph.D. Centre for Auditory Research, University College London Ear Institute, London, United Kingdom. Rosy Joshi-Mukherjee, Ph.D. Department of Pharmacology, State University Of New York Upstate Medical University, Syracuse, United States. David P. Kelsell, Ph.D. Centre for Cutaneous Research, Barts and the London, Queen Mary University of London, London, United Kingdom. Gerald M. Kidder, Ph.D. Department of Physiology and Pharmacology, University of Western Ontario & Children’s Health Research Institute, London, Canada. Philippe Klee, Ph.D. Department of Cell Physiology and Metabolism, University of Geneva, Geneva, Switzerland. Brenda R. Kwak, Ph.D. Division of Cardiology, Geneva University Hospitals, Geneva, Switzerland. Paul D. Lampe, Ph.D. Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, United States.

xiv

Contributors

Rebecca Lewandowski, M.D., Ph.D. Department of Pharmacology, State University Of New York Upstate Medical University, Syracuse, United States. Darren Locke, Ph.D. Department of Pharmacology and Physiology, New Jersey Medical School of the University of Medicine and Dentistry of New Jersey, Newark, New Jersey, United States Stephen C. Massey, Ph.D. Ophthalmology and Visual Science, University of Texas Medical School, Houston, United States. Richard T. Mathias, Ph.D. Department of Physiology and Biophysics, State University Of New York, Stony Brook, United States. Paolo Meda, M.D. Department of Cell Physiology and Metabolism, University of Geneva, Geneva, Switzerland. Marc Mesnil, Ph.D. Institute of Physiology and Cell Biology, University of Poitiers, Poitiers, France. Linda S. Musil, Ph.D. Department of Biochemistry and Molecular Biology, Oregon Health and Science University, Portland, United States. Regina Nickel, Ph.D. Centre for Auditory Research, University College London Ear Institute, London, United Kingdom. Eva M. Oxford, Ph.D. Department of Pharmacology, State University Of New York Upstate Medical University, Syracuse, United States. Anna Pfenniger, Ph.D. Division of Cardiology, Geneva University Hospitals, Geneva, Switzerland. Georges Pointis, Ph.D. Institut National de la Sante´ et de la Recherche Me´dicale (INSERM), Nice, France. John E. Rash, Ph.D. Department of Biomedical Sciences & Program in Molecular, Cellular and Integrative Neurosciences, Colorado State University, Fort Collins, United States.

Contributors

xv

Isabelle Roth, Ph.D. Division of Cardiology, Geneva University Hospitals, Geneva, Switzerland. Isabelle Scerri, Ph.D. Department of Pediatrics, Geneva University Hospitals, Geneva, Switzerland. Dominique Segretain, Ph.D. Institut National de la Sante´ et de la Recherche Me´dicale (INSERM), Paris, France. Nicholas J. Severs, Ph.D., D.Sc. National Heart and Lung Institute, Imperial College London, United Kingdom. Teresa I. Shakespeare, Ph.D. Department of Physiology and Biophysics, State University Of New York, Stony Brook, United States. Junko Shibayama, Ph.D. Department of Pharmacology, State University Of New York Upstate Medical University, Syracuse, United States. Joell L. Solan, Ph.D. Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, United States. Paul L. Sorgen, Ph.D. Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, United States. Gina E. Sosinsky, Ph.D. National Center for Microscopy and Imaging Research & Department of Neurosciences, University of California San Diego, La Jolla, United States. Miduturu Srinivas, Ph.D. Department of Biological Sciences, State University Of New York College of Optometry, New York, United States. Steven M. Taffet, Ph.D. Department of Microbiology and Immunology, State University Of New York Upstate Medical University, Syracuse, United States. Vytas K. Verselis, Ph.D. Department of Neuroscience, Albert Einstein College of Medicine, Bronx, United States. Thomas W. White, Ph.D. Department of Physiology and Biophysics, State University Of New York, Stony Brook, United States.

xvi

Contributors

Elke Winterhager, Ph.D. Institute of Anatomy, University Duisburg-Essen, Essen, Germany. Cor de Wit, Ph.D. Institute for Physiology, University of Lu¨beck, Lu¨beck, Germany. Stephanie E. Wolfle, Ph.D. ¨ Institute for Physiology, University of Lu¨beck, Lu¨beck, Germany. Mark Yeager, M.D., Ph.D. Department of Cell Biology, The Scripps Research Institute & Division of Cardiovascular Diseases, The Scripps Clinic, La Jolla, United States & Department of Molecular Physiology and Biological Physics, University of Virginia Health Sciences Center, Charlottesville, United States.

Part I

Fundamentals of Connexin Biology

Chapter 1

The Family of Connexin Genes Eric C. Beyer and Viviana M. Berthoud

Abstract The connexin genes code for a family of proteins that form intercellular gap junction channels. There are 21 connexin genes in the human genome and 20 in the mouse genome. Most connexin genes contain the protein coding region in a single exon. Variations in promoter usage and splicing of 5’-untranslated exons contribute to regulation of connexin expression. Connexin gene promoters contain binding sites for both cell type– independent and cell type–specific transcription factors that regulate connexin transcription and intercellular communication in many different cell types. Connexin transcription is also influenced by epigenetic mechanisms such as DNA methylation and histone acetylation. RNA stability and translation are influenced by binding of microRNAs and by utilization of internal ribosome entry sites. Mutations in connexin genes have been identified as associated with a variety of inherited diseases. These mutations can cause alterations in intercellular communication by affecting various processes of the connexin life cycle or channel function. The recent progress in connexin genetics has had a significant impact in elucidating the mechanisms leading to mutant connexin-associated diseases. These advances will provide the basis for future therapeutic interventions. Keywords Gene structure Disease mutations Nomenclature

E.C. Beyer (*) Department of Pediatrics, Section of Hematology/Oncology and Stem Cell Transplantation, University of Chicago, 5841 South Maryland Avenue, MC 4060, Chicago, IL 60637, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_1, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

3

4

E.C. Beyer and V.M. Berthoud

1.1 Introduction The connexins are a family of vertebrate proteins that form gap junction channels.1,2,3,4 The first members of this family were isolated by screening of complementary DNA (cDNA) libraries using antibodies directed against purified gap junctions to detect expressed fusion proteins [1] or by hybridization with oligonucleotides corresponding to amino acid sequences derived from the purified proteins [2,3]. Subsequently, additional members were identified from cDNA and genomic DNA based on their sequence similarities to those of known connexins [4,5,6]. As the genomes of various species were sequenced and became publicly available, other members were identified through computer screens that identified homologous nucleotide (or deduced amino acid) sequences. The full sequences of the connexin genes are now determined, and it is possible to establish their structures (including promoters, exons, and introns), to investigate their transcriptional and translational regulation and to examine mechanisms by which mutations within these genes affect intercellular communication between vertebrate cells. While intercellular channels and gap junction structures are found in many lower multicellular organisms, connexin genes have been identified only in deuterostomes. Gap junctions in most invertebrates, including Drosophila melanogaster and Caenorhabditis elegans, are formed by proteins termed innexins, which exhibit membrane topology similar to that of the connexins, but lack sequence homology [7,8,9]. Three genes in the genomes of higher vertebrates show substantial identity of their predicted amino acid sequences to the innexins; these have been designated pannexins [10]. While pannexins can form channels connecting the cytoplasm and extracellular space analogous to connexin hemichannels [11,12,13], it is not clear that they form intercellular channels except in certain overexpression systems [11,14,15] (see Chapter 12). Therefore, the current discussion is restricted to the connexins.

1.2 Nomenclature There are 21 connexin genes within the human genome and 20 connexin genes within the mouse genome (Table 1.1).5 Nineteen of the connexins have clearly identifiable orthologs in both mouse and human. However, there is no 1

Pfam accession number PF00029. http://pfam.sanger.ac.uk/family?acc=PF00029. PROSITE documentation PDOC00341. http://www.expasy.ch/cgi-bin/nicedoc.pl?PDOC00341. 3 ProDom family PD001135. http://prodomweb.univ-lyon1.fr/prodom/current/cgi-bin/ request.pl?db_ent1=PD001135&wanted=align&SSID=1184264083_31867. 4 SYSTERS protein family 152414. http://systers.molgen.mpg.de/cgi-bin/nph-fetchcluster. pl?PFAM=Connexin. 5 The guidelines used for designating human and mouse genes are described at http://www. bioscience.org/services/genenome.htm and http://www.informatics.jax.org/mgihome/nomen/. 2

1 Connexin Genes

5

Table 1.1 The connexin family.The connexin gene names have been adopted from those suggested by Sohl ¨ and Willecke [18], taking into consideration the phylogenetic tree from Cruciani and Mikalsen [30]. Groups IIIa and IIIb in Cruciani and Mikalsen largely correspond to the groups GJD and GJC in Sohl ¨ and Willecke with the exception of the gene for CX30.2/CX31.3. Therefore, this gene has been renamed GJC3. The CX23 gene has been assigned to a new connexin group (GJE). For simplicity, pseudogenes have not been included in this table. There is ongoing discussion within the gap junction research community about whether to change the system of naming connexins in order to eliminate discrepancies between the names of orthologous connexins in humans and other vertebrates. (Former names are indicated in parentheses) Human Mouse Protein name Gene name Protein name Gene name CX43 CX46 CX37 CX40 -------CX50 CX59 CX62

GJA1 GJA3 GJA4 GJA5 -------GJA8 GJA9 (GJA10) GJA10

Cx43 Cx46 Cx37 Cx40 Cx33 Cx50 -------Cx57

Gja1 Gja3 Gja4 Gja5 Gja6 Gja8 -------Gja10

CX32 CX26 CX31 CX30.3 CX31.1 CX30 CX25

GJB1 GJB2 GJB3 GJB4 GJB5 GJB6 GJB7

Cx32 Cx26 Cx31 Cx30.3 Cx31.1 Cx30 --------

Gjb1 Gjb2 Gjb3 Gjb4 Gjb5 Gjb6 --------

CX45 CX47 CX30.2/CX31.3

GJC1 (GJA7) GJC2 (GJA12) GJC3 (GJE1)

Cx45 Cx47 Cx29

Gjc1 Gjc2 Gjc3

CX36 CX31.9 CX40.1

GJD2 (GJA9) GJD3 (GJC1) GJD4

Cx36 Cx30.2 Cx39

Gjd2 Gjd3 Gjd4

CX23

GJE1

Cx23

Gje1

transcribed gene in the human genome orthologous to that coding for mouse Cx33, and two human connexin genes, coding for CX25 and CX59, do not have orthologs in the mouse genome. The human genomic database contains two In accordance with these conventions, human and rodent connexin DNA and RNA are indicated with italics, and proteins are designated by Roman font. Human connexin genes and proteins are designated using all upper-case letters (e.g., GJB1 and CX32, respectively). In keeping with widespread practice in the field, in this volume connexin genes and proteins from other species (e.g., mouse or rat) are identified using only initial upper case letters (e.g., GJb1 and Cx32, respectively).

6

E.C. Beyer and V.M. Berthoud

connexin pseudogenes related to the genes for GJA1 (CX43) [16] and CX31.9 [17]. So far, no connexin pseudogenes have been detected in the mouse genome. The inclusion of Cx23 as a ‘‘true’’ connexin is currently uncertain. Although the gene is present in both human and mouse genomes, its transcription and translation have not been demonstrated in humans, and the predicted protein contains only two of the characteristic three conserved cysteines in the putative extracellular loops [18]. The absence of genes for Cx25 and Cx59 in rodents is likely due to a rodentspecific loss of these sequences, since obvious orthologs to one or both have been identified in the chimpanzee, dog, and opossum genomes. Similarly, the Cx33 gene (Gja6) is not rodent-specific, since similar sequences are found on the human, chimpanzee, and dog X chromosome; the human and chimpanzee Gja6-like sequences have the characteristics of a pseudogene with several stop codons within their sequences, but the dog sequence encodes a 35.7 kilodalton (kDa) protein that is potentially functional. Members of the connexin protein family are denoted using a standard, operational nomenclature that utilizes the word connexin (abbreviated Cx) followed by a suffix indicating the predicted molecular mass of the polypeptide in kilodaltons [4]. As illustrative examples of this nomenclature, the 43 kDa protein originally identified in heart is called Cx43, and the 32 kDa protein originally identified in liver is called Cx32. The finding that different connexins have similar molecular masses has led to the use of a decimal point to distinguish them, for example, Cx30 versus Cx30.3 and Cx31 versus Cx31.1. Many of the orthologous human and mouse connexins show high levels of sequence identity. In addition, their molecular masses are often similar. Therefore, for many orthologs, the same abbreviation can be used, that is, CX43 and Cx43 for human and mouse orthologs, respectively. Unfortunately, there are several human and mouse connexin orthologs that have different predicted molecular masses and consequently different connexin protein names; pairs of such orthologs include mouse Cx29 (258 amino acids [aa]) and human CX30.2 (270 aa), mouse Cx30.2 (278 aa) and human CX31.9 (294 aa), and mouse Cx57 (505 aa) and human CX62 (543 aa). The confusing aspects of the nomenclature can be further illustrated by considering some orthologous connexins in more divergent species, for example, human/mouse Cx46, bovine Cx44, and chicken Cx56 are orthologous proteins. A second nomenclature for identification of connexins has been used by some investigators. This system was initially developed when vertebrate connexin genes were separated into two subgroups, -group and -group, on the basis of overall sequence similarities and length of the cytoplasmic loop [19,20]. Unfortunately, it is not clear that there are specific protein sequences or domains that reliably define either group, nor that there are functional correlates to the sequence differences. Moreover, analysis of several subsequently characterized connexin protein sequences (e.g., Cx36, Cx45, and Cx47) suggests that this simple categorization scheme needed refinement or at least addition of more subgroups [18].

1 Connexin Genes

7

Nevertheless, this Greek letter classification forms the basis for the current official nomenclature for identification of connexin genes. They are named starting with Gj for mouse or GJ for human genes as an abbreviation for gap junction, followed by a letter indicating the Greek letter subgroup, and then a number indicating the order of discovery. For example, the gene for Cx43 was the first identified connexin gene of the -group (Gja1) and the gene for Cx32 was the first identified connexin gene of the -group (Gjb1). Identification of connexin gene orthologs (or their absence) in different species raises interesting questions. For example, the zebrafish genome contains 37 connexin genes, and assignment of orthologs is complicated [21]; apparently, there are multiple orthologs of some mammalian connexins and other sequences without orthologs in higher vertebrates. A modified version of the Greek letter classification system for the human and rodent connexin genes was adopted at the International Gap Junction Conference in 2007. This gene nomenclature is shown in Table 1.1 (also see Appendix).

1.3 Domain Structure of the Connexins The relationships between members of the connexin family can be understood through examination of their sequences and the locations of different domains relative to the plasma membrane. After the initial cloning studies, hydropathy plots were generated and used to predict the locations of transmembrane helices and to produce a topological model for the connexins (Fig. 1.1). The model has been tested, and generally verified, through the use of immunocytochemistry to determine the cytoplasmic or extracellular reactivity of antibodies directed against peptides corresponding to subdomains within the connexin polypeptides [22,23,24,25,26,27]. The strategy has been applied to several connexins, but the

Fig. 1.1. Membrane topology of connexins. Diagram of a connexin polypeptide depicting its orientation relative to the plasma membrane. The protein contains four transmembrane domains (M1 to M4) with the amino-terminal and carboxyl-terminal domains (NT and CT, respectively) located on the cytoplasmic side of the membrane. The protein also contains two extracellular loops (E1 between M1 and M2, and E2 between M3 and M4) and a cytoplasmic loop (CL) between M2 and M3. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

8

E.C. Beyer and V.M. Berthoud

Fig. 1.2. Consensus sequence showing the conserved amino acids in the connexin polypeptides. This consensus was produced by aligning 20 human connexin proteins (excluding the putative connexin, CX23) using the AlignX component of Vector NTI Advance 10.3.0 (Invitrogen, Carlsbad, CA). Amino acid residues that are identical in all of these connexins are highlighted in bold; amino acids that are identical in at least 50% of these connexins are represented in black; amino acids that are less well conserved are represented in grey by X’s. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com)

model has been extrapolated to all family members. The topological model shows that the connexin polypeptide contains a short cytoplasmic amino-terminal domain (NT), four transmembrane domains (M1 to M4) separated by one cytoplasmic loop domain (CL) between M2 and M3, two extracellular loops (E1 and E2), and a carboxyl-terminal cytoplasmic domain (CT). Alignment of the amino acid sequences of family members reveals the locations of conserved and divergent regions within the connexin polypeptides. This analysis can also be used to generate a consensus sequence including residues that are highly conserved (or identical) among the family members (Fig. 1.2). The extracellular loops, E1 and E2, involved in docking together the hemichannels contributed by adjacent cells, are among the most conserved regions. Each extracellular loop contains three cysteines, and the number of amino acid residues separating them are conserved among connexins (E1 contains C–X6–C–X3–C and E2 contains C–X4–C–X5–C). These sequence similarities have been used to define two connexin consensus patterns or signatures: Connexin signature 16: C – [D/N/H] – [T/L] – X – [Q/T] – P – G – C – X – X – [V/A/I/L] – C – [F/Y] – D Connexin signature 27: C – X – X – X – X – P – C – X – X – X – [L/I/V/M/T/A] – [D/E/N/T] – C – [F/ Y/N] – [L/I/V/M/Q] – [S/A] – [K/R/H] – P There are two exceptions: Cx31 has the sequence C-X5-C-X5-C in E2 [28], and the putative connexin, Cx23, has only two cysteines in E1 and E2 [18]. The lengths and sequences of the NT, M1 through M4, E1 and E2 are relatively conserved among all connexins. The remaining cytoplasmic domains, the CL and the CT, have highly divergent amino acid sequences. The CT has the greatest variability in size among connexins, from as short as 10 to 12 aa in Cx26, Cx30.3, and Cx31.1 to more than 310 aa in CX59 and CX62. The length 6 7

http://ca.expasy.org/prosite/PS00407. http://ca.expasy.org/prosite/PS00408.

1 Connexin Genes

9

of the CL also exhibits substantial variability; Sohl and Willecke [17] have ¨ observed that connexins can be divided into three categories: those that have small (30 to 35 aa), medium (50 to 55 aa), or large (80 to 105 aa) CL domains.

1.4 Connexin Subfamilies As noted above, comparisons of the sequences of the connexin polypeptides from humans and rodents indicated the potential for grouping together sequences based on their similarities. A variety of dendrograms has been generated and published showing the relationships among the different connexin genes. Such studies have suggested that this gene family may have developed through duplication of ancestral connexin gene(s) [29]. These analyses have recently been expanded with the completion of sequencing of the genomes for additional vertebrate species, including chimpanzee, rat, dog, cow, elephant, opossum, chicken, Xenopus tropicalis, zebrafish, Fugu rubripes, and Tetraodon nigroviridis. Cruciani and Mikalsen [30] identified connexin genes from these species and included them with the human and mouse genes to analyze 303 sequences and thereby generate a phylogenetic tree of vertebrate connexins (Fig. 1.3). This analysis confirms that the connexin family contains the two subfamilies previously called group I or -group connexins and group II or -group connexins [20,29,31]. Their phylogenetic analysis confirms the identification of a third group of genes similar to that for Cx36 (sometimes referred to as the -group), which these authors have termed group IIIa. Cruciani and Mikalsen [30] have also clustered together the genes similar to that of Cx45 into group IIIb. Their analysis also confirms that only a few gene duplications or gene losses can explain all changes that have taken place along the main evolutionary branch from the pre-fish vertebrates to mammals and that two to six connexin genes were lost before the mammals emerged [30].

1.5 Genomic Organization Several years ago, it was observed that some connexin genes are clustered together within the genome [32]. This can be readily illustrated for the human chromosomes (Fig. 1.4). Human chromosome 1 codes for eight connexins, chromosome 6 for four connexins (including the putative connexin, CX23), chromosome 13 for three connexins, chromosome 17 for two connexins, and chromosomes 7, 10, 15, and the X chromosome code for one connexin each. On chromosomes 1 and 13 some of the connexin genes cluster closely together, within about 100 kilobases (kb). The closely linked sequences on human chromosome 1 are GJB5 (CX31.1), GJB4 (CX30.3), GJB3 (CX31), and GJA4 (CX37) in one cluster, and GJA5 (CX40) and GJA8 (CX50) in another cluster. On chromosome 13, GJB6 (CX30), GJB2 (CX26), and GJA3 (CX46) are closely

10

E.C. Beyer and V.M. Berthoud

Fig. 1.3. Phylogenetic tree of vertebrate connexins. The tree was generated based on analysis of 303 connexin gene sequences from human, chimpanzee, mouse, rat, dog, cow, elephant, opossum, chicken (Gg), Xenopus tropicalis (Xt), zebrafish (Dr), Fugu rubripes, and Tetraodon nigroviridis using the program MEGA3 (http://www.megasoftware.net). The numerical portion of the connexin name is at the base of the triangles. The widths of the triangles indicate the number of sequences in a group, and the lengths of the triangles indicate the variability of the included sequences. The evolutionary branches containing the included sequences are shown

1 Connexin Genes

11

Fig. 1.4. Chromosomal localization of connexin genes. Diagram of the human chromosomes in which the distribution and localization of connexin genes has been indicated. The genes are designated by their protein names. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

linked. In other vertebrates such as the mouse, the genes for the orthologs of these connexins are often found clustered in a similar manner (in syntenic regions) with their relative directions of transcription conserved. Whether clustered connexin genes are regulated coordinately remains to be determined.

1.6 Connexin Gene Structures The initially described structure of connexin genes was relatively simple (Fig. 1.5a). An untranslated 5’-exon (exon 1) is separated from a second exon (exon 2) by an intron of variable size. Exon 2 contains the complete protein coding sequence and the 3’-untranslated region (UTR). Subsequently, connexin gene structures that differed from this were described (Fig. 1.5b–d). For example, Gjb1 contains multiple 5’-untranslated exon 1 sequences (labeled Exon 1A and 1B on Fig. 1.5b) that are alternatively spliced to exon 2 after transcription driven from two tissue-specific promoters. In liver and pancreas, promoter P1 (located more than 8 kb upstream of the translation start codon) is used, and the transcript is processed to remove a large intron. In contrast, in Schwann cells, transcription is initiated from the nervous

Fig. 1.3 (continued) in parentheses. The scale bar indicates the length of a branch corresponding to a 10% difference in nucleotides. Interior branch statistics are shown at the nodes. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Cruciani and Mikalsen [30] with permission.)

12

E.C. Beyer and V.M. Berthoud

Fig. 1.5. Structure of the connexin genes. Diagrams depict the structures of different connexin genes with the coding regions in dark gray boxes and the noncoding regions in stippled boxes. Most connexin genes illustrated contain the complete, uninterrupted coding region in a single exon (a, b, and c). The variations occur in the 5’-untranslated region (UTR). (a) The initially described connexin gene structure contained only one 5’-UTR exon. (b) Some connexin genes contain two or more 5’-UTR exons (1A and 1B) that are alternatively utilized due to transcription from tissue-specific promoters. (c) Other connexin genes contain two or more 5’-UTR exons (1 and 2) that may be present with the coding exon (3) in the mature mRNA or that may be alternatively spliced to generate multiple mRNA variants. For these connexin genes, multiple mRNAs can be generated after transcription driven from a single promoter. (d) In a few connexin genes, the coding region is interrupted by an intron. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

system-specific promoter, P2, located 497 base pairs (bp) upstream from the translation start codon, and the transcript is processed to remove a small 355 bp intron. The downstream exon 2, which includes the entire coding sequence, is shared by both messenger RNAs (mRNAs). Gjc1 (Cx45) is an example of a connexin gene that contains three exons, including two 5’-UTR exons with the complete reading frame located in exon 3 (Fig. 1.5c) [33,34,35]. Gjc1 transcripts can be differentially spliced so that the 5’-UTR is generated either from two exons (1 and 2) or only from exon 2 [33,34,35]. In a few connexin genes (such as Gjd2, encoding Cx36), the coding region is located on both exon 1 and exon 2 and is interrupted by an intron (Fig. 1.5d) [36,37]. In the human connexin GJE1 gene (coding for CX23), the coding region is interrupted by two introns.

1 Connexin Genes

13

Recent investigations suggest that the use of multiple alternative exons to generate the 5’-UTR may be a feature of many connexin genes. Pfeifer et al. [38] reported that the mouse Cx43 gene contains six exons, five of which correspond to 5’-UTRs (exons 1A to 1E) and one containing the coding region (exon 2); it also has three alternate promoter regions (P1 to P3). These authors found nine different mouse Gja1 mRNAs generated through use of differential promoters and alternative splicing. Similarly, genes for Cx40, Cx45, and other connexins appear to contain multiple exons, allowing generation of alternatively spliced transcripts containing differing 5’-UTRs [33,34,39,40]. Aside from the tissue-specific transcription of Gjb1, the relative abundance of the different transcripts and the biological significance of their generation remain to be elucidated. However, in the case of Gjb1, it has been suggested that the transcripts differ in mRNA translation or stability due to their differing 5’-UTRs [41]. The spliced forms of transcripts of other connexins may confer similar regulation, since 5’-UTRs (especially ones that can form stable structures) may contribute to translational control [42].

1.7 Transcriptional and Translational Regulation of Connexin Expression The transcription of connexin DNAs is regulated by a variety of factors, and the results of investigations of this process have been reviewed previously [43,44]. Therefore, only highlights will be summarized here. In most connexin genes, the basal promoter is located 300 bp upstream from the transcriptional initiation site. This region contains binding sites for cell type–independent (ubiquitous) transcription factors including TATA boxbinding protein, Sp1/Sp3, and activator protein-1 (AP-1). Binding of these factors is not only important for basal expression of connexins, but may also be important for large changes in expression. For example, the dramatic increase in myometrial expression of Gja1 (Cx43) mRNA just prior to parturition may occur through production of c-fos, which interacts with the AP-1 sites [45]. Many studies have demonstrated the regulation of connexin transcription in response to treatment of cells with a variety of chemicals or biological substances (including cyclic adenosine monophosphate [cAMP], phorbol esters, and retinoids) or manipulation of signal transduction pathways (such as the Wnt/ -catenin/T-cell factor (TCF)/lymphocyte enhancer binding factor (LEF) pathway; reviewed elsewhere [44]). This transcriptional regulation has been most extensively analyzed for Gja1. The Gja1 promoter region (upstream from the major transcriptional start site) contains two potential TCF/LEF binding sites and one cAMP-responsive element [46]; however, there is little evidence to suggest that these substances directly activate Gja1 expression through binding to these regions [44]. Moreover, the Gja1 promoter does not contain clear retinoic acid responsive elements.

14

E.C. Beyer and V.M. Berthoud

The connexin gene promoters also contain binding sites for cell type–specific transcription factors. For example, multiple factors, including Nkx2.5, GATA4, and Tbx5, regulate Gja5 (Cx40) expression in the heart [47]. Indeed, Nkx2.5 is an important regulator of genes for three connexins (Cx40, Cx43, and Cx45) in the cardiovascular system, since transcripts of all three show reduced levels in Nkx2.5-null mouse embryos [48]. Tbx5 also plays a crucial role in regulating Gja5 expression, since levels of mRNA for Cx40 are markedly reduced in heterozygous Tbx5-null mice, a murine model for the Holt-Oram syndrome [49]. However, the reduction in Cx40 protein may not be the underlying cause of arrhythmias in these animals, since Tbx5 also affects the expression of several other genes [50]. The interplay between utilization of different transcription factors can be clearly illustrated for the differential expression of Gjb1 (Cx32). The P1 promoter contains binding sites for cell type–independent transcription factors, including Sp1/Sp3 and nuclear factor-1 (NF-1) [51,52,53] as well as for cell type–specific factors including hepatocyte nuclear factor-1 (HNF-1) [54]. The Schwann cell–specific Gjb1 promoter, P2, contains binding sites for cell type–specific transcription factors including the early growth response gene-2 (Egr2/Knox20) and SOX10, which also activates transcription of other myelin genes [55,56]. Connexin transcription can also be regulated through epigenetic mechanisms. Several studies have suggested that reduced connexin expression in tumors and transformed cell lines may be due, at least in part, to hypermethylation of the DNA in promoter regions or to histone acetylation. Methylation of the Gja1 and Gjb1 promoters at MspI/HpaII restriction sites correlates well with cell type–dependent decreased transcriptional activity in hepatic and cardiac cells [57]. Histone acetylation may regulate the transcription of genes for several connexins (including Cx32, Cx36, and Cx43) in a variety of cell types [58,59,60,61,62,63]. Indeed, treatment of some of these cells with histone deacetylase inhibitors (such as 4-phenylbutyrate, suberoylanilide hydroxamic acid, or trichostatin A) increases connexin expression and intercellular communication. The production of connexin protein can be affected by conditions that influence translational initiation. The 5’-UTR of the Gja1 mRNA contains a strong internal ribosome entry site (IRES) that confers increased translation [64]. Schiavi et al. [64] found that when the 5’-UTR of rat Gja1 was inserted between the two genes of a bicistronic vector and transfected into various cell lines, expression of the second gene was significantly increased. Hudder and Werner [41] have identified an IRES element in the 5’-UTR of the nerve-specific Gjb1 mRNA. IRES-mediated translation has previously been observed in key regulatory genes whose protein products are required during pathophysiological processes where cap-dependent translation is reduced or abolished [42]. Consistent with such a role in permitting connexin translation (under otherwise suppressive conditions), Lahlou et al. [65] have observed that the activities of

1 Connexin Genes

15

IRES elements in both the Gja1 and the Gjb2 mRNAs are enhanced in densityinhibited cells. Recently, studies have suggested the importance of microRNAs (miRNAs) in regulating connexin mRNAs. miRNAs are endogenous noncoding RNAs (22 nucleotides in length) that mediate posttranscriptional gene silencing by reducing translation or mRNA stability through annealing to complementary sequences in the 3’-UTRs of target mRNAs [66]. Two miRNAs, miR-206 and miR-1, bind to the Cx43 mRNA and decrease levels of Cx43 during muscle differentiation [67]. Moreover, miR-1 is overexpressed in humans with coronary artery disease and in rats with myocardial infarctions; this overexpression may exacerbate arrhythmogenesis by posttranscriptionally repressing Gja1 and KCNJ2, which encodes the K+ channel subunit Kir2.1 [68].

1.8 Connexin Mutations and Disease Germline mutations of several different connexins have been identified in association with a wide variety of inherited diseases, including neuropathies, deafness, epidermal diseases, cataracts, and oculodentodigital dysplasia (Table 1.2) (discussed in the chapters in Section II of this Volume). The inheritance of these diseases may be autosomal dominant, autosomal recessive, or X-linked. Some connexin mutation-disease associations are controversial. For example, GJA1 mutations were initially identified in cardiac tissue samples from patients with heterotaxy syndromes and congenital heart malformations, but this association was not subsequently confirmed in analysis of germline DNA. Mutations in GJD2 (CX36, which is expressed by neurons) have been excluded as a cause of schizophrenia [69,70], but a recent report suggests an association of schizophrenia with the GJA8 gene (encoding CX50) [71]; this is indeed surprising, since Gja8 mRNA is not detectable in tissues other than the ocular lens [72]. Disease-causing mutations can potentially occur anywhere in the connexin genes. They may cause disease through a variety of mechanisms, most of which alter intercellular communication. Mutations in noncoding regions might disrupt elements that regulate transcription or interfere with proper splicing to generate mature connexin mRNAs. A few such mutations have been identified in GJB1 associated with X-linked Charcot-Marie-Tooth disease that may alter the transcription of GJB1 in Schwann cells [56,73]. Similarly, a mutation that alters the IRES in untranslated regions of the mRNA may reduce translation of connexin protein. Hudder and Werner [41] studied a noncoding GJB1 mutation found in a family with Xlinked Charcot-Marie-Tooth disease (C to T transition at position –458 in relation to the start codon) and found that it abolishes function of the IRES in the 5’-UTR when studied in a bicistronic expression system.

16

Disease

E.C. Beyer and V.M. Berthoud Table 1.2 Hereditary diseases associated with germline connexin mutations Connexin

Neuropathy [Chapter 15]

Deafness (nonsyndromic) [Chapter 20]

Deafness (syndromic)8 [Chapter 20]

Skin disease (without deafness) [Chapter 14] Cataracts [Chapter 18]

Charcot-Marie-Tooth disease, X-linked, type 1 (CMTX1) Pelizaeus-Merzbacher-like disease-1 (PMLD1) DFNB1, DFNA3 DFNA2 DFNA3 Dominant deaf-mutism and palmoplantar keratoderma Ectodermal dysplasia keratitis-ichthyosisdeafness syndrome Sensorineural hearing loss and palmoplantar hyperkeratosis Mutilating keratoderma with sensorineural deafness (Vohwinkel syndrome) Deafness, peripheral neuropathy, and erythrokeratodermia variabilis Erythrokeratodermia variabilis9

CX321,2 [101] CX473 [102] CX264 CX315 CX306 CX437 [103] CX26 [104,105] CX26 [106] CX26 [107] CX26 [108]

CX31 [109]

CX31 [110,111] CX30.3 [112,113] Autosomal dominant hidrotic ectodermal CX30 [114] dysplasia (Clouston syndrome) Autosomal dominant zonular pulverulent CX5010 [115] cataract-1(CZP1) Autosomal dominant zonular pulverulent CX4611 [116] cataract-3 (CZP3) CX4312 [117]

Oculodentodigital dysplasia [Chapter 17] 1 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=302800. 2 http://www.molgen.ua.ac.be/CMTMutations/. 3 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=608804. 4 http://davinci.crg.es/deafness/index.php?seccion=mut_db&db=nonsynd& nonsynd=cx26mut. 5 http://davinci.crg.es/deafness/index.php?seccion=mut_db&db=nonsynd& nonsynd=cx31mut. 6 http://davinci.crg.es/deafness/index.php?seccion=mut_db&db=nonsynd& nonsynd=cx30mut. 7 http://davinci.crg.es/deafness/index.php?seccion=mut_db&db=nonsynd& nonsynd=cx43mut. While mutations of the gene for Cx43 (GJA1) associated with nonsyndromic deafness have been reported [103], they have not been confirmed. It has been suggested that the described mutations may have actually derived from amplification and sequencing of the pseudogene for CX43 [17]. 8 http://davinci.crg.es/deafness/index.php?seccion=mut_db&db=synd. 9 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=133200. 10 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=116200. 11 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=601885. 12 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=164200.

1 Connexin Genes

17

Fig. 1.6. Connexin life cycle and its relation to disease. Multiple steps may be impaired by connexin gene mutations that lead to alterations in intercellular communication and disease. These steps include (1) connexin mRNA transcription, splicing/processing, and stability; (2) translation and protein folding; (3) oligomerization and intracellular trafficking; (4) insertion of hemichannels into the plasma membrane; (5) docking of hemichannels and clustering of channels into gap junction plaques; and (6) degradation by the proteasome and lysosome. Connexin mutations can also lead to increased formation and opening of functional hemichannels or formation of gap junction channels that do not conduct or have altered gating/ permeability. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com)

The vast majority of identified connexin mutations are located within the protein coding region. The different protein mutations produce abnormalities at various steps in the connexin life cycle, including synthesis, assembly, channel function, and degradation (Fig. 1.6). Nonsense mutations or frameshift mutations leading to premature termination of translation result in loss of normal, full-length protein, for example, the 35delG mutation of CX26 [74]. Missense

18

E.C. Beyer and V.M. Berthoud

mutations can interfere with proper protein folding or oligomerization to form hexameric hemichannels resulting in mutant connexins that remain in the endoplasmic reticulum or Golgi apparatus [75] or are targeted for degradation. Fleishman et al. [76] have presented data suggesting that disease-associated mutations in the transmembrane regions of CX26 and CX32 may disrupt salt bridges involved in packing of the transmembrane helices (see Chapter 2). Among the cytoplasmically-retained mutants, many CX32 mutants are degraded through the proteasomal quality control system [77]; in contrast, other mutants such as CX50P88S [78] form degradation-resistant accumulations (see Chapter 9). Trafficking through intracellular compartments to the plasma membrane may be affected by mutations disrupting signals inherent within the wild-type connexin polypeptide or by generation of new retention/ retrieval signals, for example, CX46fs380 [79]. Hemichannels present within the plasma membrane must be kept closed; a mutation that increases the probability of hemichannel opening may cause cell death, as suggested for ‘‘leaky’’ CX32 mutants [80,81]. Hemichannels must ‘‘dock’’ with their counterparts from the apposing cell, and this process might also be disrupted by mutations. While the extracellular loops are involved in hemichannel docking and a number of mutations in these regions have been identified (especially in E1), none that specifically block hemichannel docking has yet been identified; indeed, the E1 mutants that have been studied accumulate intracellularly [82,83]. Mutations can also alter the properties of gap junction channels including permeability/selectivity [84,85] and gating [86]. Some loss-of-function mutants form gap junction channels that are nonfunctional or have very reduced function. Dominant-negative mutations inhibit the function of their wild-type counterparts [87,88] or of other coexpressed connexins [89], probably by forming mixed hemichannels; this mechanism may explain the autosomal dominant inheritance pattern of some connexin mutants. Mutations may alter the specificity and compatibilities of hetero-oligomerization; for example, wild-type Cx43 and Cx26 do not form heteromeric hemichannels, but several CX26 mutants associated with skin disease (where the two connexins are coexpressed) can inhibit the function of CX43 in a heterologous expression system [90] (see Chapter 14). Mutations may alter degradation of connexins from the plasma membrane. Moreover, mutations that alter phosphorylatable amino acid residues may also lead to perturbations of intercellular communication and disease, since phosphorylation is involved in several of the steps of the connexin life cycle, for example, synthesis, trafficking, gating, and internalization. A number of connexin mutations that alter putative phosphorylation sites have been described. Similarly, disease-associated mutations may alter sites that are subject to other posttranslational modifications such as ubiquitinylation, which is involved in protein degradation. Mutations may also alter interactions between connexins and other proteins, including chaperones, proteins involved in vesicular transport, and scaffolding proteins (see Chapters 9 and 11).

1 Connexin Genes

19

In addition to germline mutations, disease may be produced by the development of connexin mutations within somatic cells. DNA encoding several mutant forms of CX40 (including CX40P88S in multiple different samples) was isolated from the hearts of patients with atrial fibrillation [91] suggesting that somatic mutations of GJA5 (CX40) may be responsible for some cases of this arrhythmia. Besides proven mutations, there are a variety of polymorphic connexin variants. Many such variants have been identified for GJB1 (CX32) and GJB2 (CX26), and their designation as polymorphic variants (as opposed to disease-associated mutants) is clear due to the lack of disease association. Polymorphic variants within the GJA5 (CX40) promoter have been associated with atrial fibrillation [92]. Whether these are simply variants that cosegregate with the disease or whether they truly harbor abnormalities that alter normal physiology remains to be determined. The most provocative and potentially clinically important, but contradictory, data concern a single nucleotide polymorphism (SNP) that occurs in GJA4 (C1019T, which changes a proline to a serine at position 319 in CX37) (see Chapter 23). Both variants are very common in many populations. The GJA4-C1019 SNP has been associated with thickening of the carotid intima in Swedish men [93], with coronary artery disease in a Taiwanese population [94] and with myocardial infarction in Switzerland [95]. In other studies, the GJA4-T1019 SNP is a risk factor for acute myocardial infarction among Japanese [96,97] and Sicilian males [98]. Neither variant is associated with subclinical atherosclerosis in Finnish young adults [99]. The physiological basis of these observations remains elusive, since double whole cell patch clamp recordings of CX37S319 and CX37P319 expressed in Neuro2A cells have not revealed functional differences between channels formed by the two variants [100].

1.9 Conclusion Substantial progress has been made in our understanding of the genetics of the connexin family of gap junction proteins. Connexin genes from humans (and many other organisms) have been identified and fully sequenced. Substantial information has been obtained regarding the regulation of their expression. A wide variety of diseases have been associated with mutations in connexin genes. One goal for the future should be to continue to identify connexin mutants/ polymorphisms associated with disease and to elucidate the mechanisms by which they contribute to pathology. Most disease-associated connexin mutations characterized to date cause severe alterations of connexin function and are inherited as simple Mendelian traits, for example, autosomal dominant and recessive; in the future, additional mutations (potentially including ones that cause mild alterations of intercellular communication) may be identified that

20

E.C. Beyer and V.M. Berthoud

contribute to pathologies of multigenic etiology. It is very possible that additional somatic mutations or acquired abnormalities of connexin expression will be found to have important roles in disease. For connexin variants that have been termed disease-associated polymorphisms, it will be critical to determine whether they indeed directly cause cellular or physiological abnormalities. Alternatively, polymorphisms of some connexins may serve as markers of disease (perhaps due to linkage to other mutated genes that are the actual cause of the pathology); the mutated genes will need to be identified. The second major goal for the future of connexin genetics will be to develop translational approaches to manipulate connexin expression for therapeutic benefit and to continue to improve our understanding of the regulation of connexin expression (both genetic regulation of connexin transcription and epigenetic regulation of transcription/translation). Such information will be invaluable for the development of strategies for manipulating connexin levels (both increases and decreases) either through specific manipulation of cell type–specific transcription factors or through administration of drugs that modulate general processes such as histone deacetylation or DNA methylation. Utilization of these strategies may lead to new therapeutic approaches for diseases caused by connexin mutations and for common pathologies of complex etiologies in which connexin expression is altered or reduced, such as cancer (e.g., to facilitate delivery or efficacy of chemotherapy and gene therapy) and cardiac arrhythmias (e.g., to augment or block electrical conduction). Acknowledgment The research in the authors’ laboratories is supported by National Institutes of Health (NIH) grants HL59199, HD09402, and EY08368. We would like to thank Dr. Svein-Ole Mikalsen for performing further analysis to assign the gene encoding CX23 to a connexin subfamily.

References 1. Paul DL. Molecular cloning of cDNA for rat liver gap junction protein. J Cell Biol. 1986;103:123–34. 2. Kumar NM, Gilula NB. Cloning and characterization of human and rat liver cDNAs coding for a gap junction protein. J Cell Biol. 1986;103:767–76. 3. Zhang JT, Nicholson BJ. Sequence and tissue distribution of a second protein of hepatic gap junctions, Cx26, as deduced from its cDNA. J Cell Biol. 1989;109:3391–401. 4. Beyer EC, Paul DL, Goodenough DA. Connexin43: a protein from rat heart homologous to a gap junction protein from liver. J Cell Biol. 1987;105:2621–9. 5. Ebihara L, Beyer EC, Swenson KI, Paul DL, Goodenough DA. Cloning and expression of a Xenopus embryonic gap junction protein. Science. 1989;243:1194–5. 6. Willecke K, Jungbluth S, Dahl E, Hennemann H, Heynkes R, Grzeschik KH. Mouse connexin37: cloning and functional expression of a gap junction gene highly expressed in lung. J Cell Biol. 1991;114:1049–57. 7. Starich TA, Lee RY, Panzarella C, Avery L, Shaw JE. eat-5 and unc-7 represent a multigene family in Caenorhabditis elegans involved in cell-cell coupling. J Cell Biol. 1996;134:537–48.

1 Connexin Genes

21

8. Ganfornina MD, Sa´nchez D, Herrera M, Bastiani MJ. Developmental expression and molecular characterization of two gap junction channel proteins expressed during embryogenesis in the grasshopper Schistocerca americana. Dev Genet. 1999;24:137–50. 9. Bauer R, Loer ¨ B, Ostrowski K, Martini J, Weimbs A, Lechner H, Hoch M. Intercellular communication: the Drosophila innexin multiprotein family of gap junction proteins. Chem Biol. 2005;12:515–26. 10. Panchin Y, Kelmanson I, Matz M, Lukyanov K, Usman N, Lukyanov S. A ubiquitous family of putative gap junction molecules. Curr Biol. 2000;10:R473–4. 11. Huang Y, Grinspan JB, Abrams CK, Scherer SS. Pannexin1 is expressed by neurons and glia but does not form functional gap junctions. Glia. 2007;55:46–56. 12. Locovei S, Bao L, Dahl G. Pannexin 1 in erythrocytes: function without a gap. Proc Natl Acad Sci USA. 2006;103:7655–9. 13. Pelegrin P, Surprenant A. Pannexin-1 mediates large pore formation and interleukin-1 release by the ATP-gated P2X7 receptor. EMBO J. 2006;25:5071–82. 14. Bruzzone R, Hormuzdi SG, Barbe MT, Herb A, Monyer H. Pannexins, a family of gap junction proteins expressed in brain. Proc Natl Acad Sci USA. 2003;100:13644–9. 15. Vanden Abeele F, Bidaux G, Gordienko D, Beck B, Panchin YV, Baranova AV, Ivanov DV, Skryma R, Prevarskaya N. Functional implications of calcium permeability of the channel formed by pannexin 1. J Cell Biol. 2006;174:535–46. 16. Fishman GI, Spray DC, Leinwand LA. Molecular characterization and functional expression of the human cardiac gap junction channel. J Cell Biol. 1990;111:589–98. 17. Sohl ¨ G, Willecke K. Gap junctions and the connexin protein family. Cardiovasc Res. 2004;62:228–32. 18. Sohl ¨ G, Willecke K. An update on connexin genes and their nomenclature in mouse and man. Cell Commun Adhes. 2003;10:173–80. 19. Gimlich RL, Kumar NM, Gilula NB. Differential regulation of the levels of three gap junction mRNAs in Xenopus embryos. J Cell Biol. 1990;110:597–605. 20. Kumar NM, Gilula NB. Molecular biology and genetics of gap junction channels. Semin Cell Biol. 1992;3:3–16. 21. Eastman SD, Chen TH, Falk MM, Mendelson TC, Iovine MK. Phylogenetic analysis of three complete gap junction gene families reveals lineage-specific duplications and highly supported gene classes. Genomics. 2006;87:265–74. 22. Milks LC, Kumar NM, Houghten R, Unwin N, Gilula NB. Topology of the 32-kd liver gap junction protein determined by site-directed antibody localizations. EMBO J. 1988;7:2967–75. 23. Goodenough DA, Paul DL, Jesaitis L. Topological distribution of two connexin32 antigenic sites in intact and split rodent hepatocyte gap junctions. J Cell Biol. 1988;107:1817–24. 24. Beyer EC, Kistler J, Paul DL, Goodenough DA. Antisera directed against connexin43 peptides react with a 43-kD protein localized to gap junctions in myocardium and other tissues. J Cell Biol. 1989;108:595–605. 25. Yancey SB, John SA, Lal R, Austin BJ, Revel JP. The 43-kD polypeptide of heart gap junctions: immunolocalization, topology, and functional domains. J Cell Biol. 1989;108:2241–54. 26. Yeager M, Gilula NB. Membrane topology and quaternary structure of cardiac gap junction ion channels. J Mol Biol. 1992;223:929–48. 27. Zhang JT, Nicholson BJ. The topological structure of connexin 26 and its distribution compared to connexin 32 in hepatic gap junctions. J Membr Biol. 1994;139:15–29. 28. Hennemann H, Schwarz HJ, Willecke K. Characterization of gap junction genes expressed in F9 embryonic carcinoma cells: molecular cloning of mouse connexin31 and -45 cDNAs. Eur J Cell Biol. 1992;57:51–8. 29. Bennett MVL, Barrio LC, Bargiello TA, Spray DC, Hertzberg E, Sa´ez JC. Gap junctions: new tools, new answers, new questions. Neuron. 1991;6:305–20.

22

E.C. Beyer and V.M. Berthoud

30. Cruciani V, Mikalsen SO. The vertebrate connexin family. Cell Mol Life Sci. 2006;63:1125–40. 31. Haefliger JA, Bruzzone R, Jenkins NA, Gilbert DJ, Copeland NG, Paul DL. Four novel members of the connexin family of gap junction proteins. Molecular cloning, expression, and chromosome mapping. J Biol Chem. 1992;267:2057–64. 32. Schwarz HJ, Chang YS, Hennemann H, Dahl E, Lalley PA, Willecke K. Chromosomal assignments of mouse connexin genes, coding for gap junctional proteins, by somatic cell hybridization. Somat Cell Mol Genet. 1992;18:351–9. 33. Kru¨ger O, Plum A, Kim JS, Winterhager E, Maxeiner S, Hallas G, Kirchhoff S, Traub O, Lamers WH, Willecke K. Defective vascular development in connexin 45-deficient mice. Development. 2000;127:4179–93. 34. Jacob A, Beyer EC. Mouse connexin 45: genomic cloning and exon usage. DNA Cell Biol. 2001;20:11–19. 35. Baldridge D, Lecanda F, Shin CS, Stains J, Civitelli R. Sequence and structure of the mouse connexin45 gene. Biosci Rep. 2001;21:683–9. 36. Condorelli DF, Parenti R, Spinella F, Trovato Salinaro A, Belluardo N, Cardile V, Cicirata F. Cloning of a new gap junction gene (Cx36) highly expressed in mammalian brain neurons. Eur J Neurosci. 1998;10:1202–8. 37. Sohl ¨ G, Degen J, Teubner B, Willecke K. The murine gap junction gene connexin36 is highly expressed in mouse retina and regulated during brain development. FEBS Lett. 1998;428:27–31. 38. Pfeifer I, Anderson C, Werner R, Oltra E. Redefining the structure of the mouse connexin43 gene: selective promotor usage and alternative splicing mechanisms yield transcripts with different translational efficiencies. Nucleic Acids Res. 2004;32:4550–62. 39. Dupays L, Mazurais D, Ru¨cker-Martin C, Calmels T, Bernot D, Cronier L, Malassine´ A, Gros D, The´veniau-Ruissy M. Genomic organization and alternative transcripts of the human connexin40 gene. Gene. 2003;305:79–90. 40. Anderson CL, Zundel MA, Werner R. Variable promotor usage and alternative splicing in five mouse connexin genes. Genomics. 2005;85:238–44. 41. Hudder A, Werner R. Analysis of a Charcot-Marie-Tooth disease mutation reveals an essential internal ribosome entry site element in the connexin-32 gene. J Biol Chem. 2000;275: 34586–91. 42. Pickering BM, Willis AE. The implications of structured 5’ untranslated regions on translation and disease. Sem Cell Dev Biol. 2005;16:39–47. 43. Sa´ez JC, Berthoud VM, Bran˜es MC, Martı´ nez AD, Beyer EC. Plasma membrane channels formed by connexins: their regulation and functions. Physiol Rev. 2003;83: 1359–400. 44. Oyamada M, Oyamada Y, Takamatsu T. Regulation of connexin expression. Biochim Biophys Acta. 2005;1719:6–23. 45. Piersanti M, Lye SJ. Increase in messenger ribonucleic acid encoding the myometrial gap junction protein, connexin-43, requires protein synthesis and is associated with increased expression of the activator protein-1, c-fos. Endocrinology. 1995;136:3571–8. 46. Sullivan R, Ruangvoravat C, Joo D, Morgan J, Wang BL, Wang XK, Lo CW. Structure, sequence and expression of the mouse Cx43 gene encoding connexin 43. Gene. 1993;130:191–9. 47. Linhares VL, Almeida NA, Menezes DC, Elliott DA, Lai D, Beyer EC, Campos de Carvalho AC, Costa MW. Transcriptional regulation of the murine Connexin40 promotor by cardiac factors Nkx2-5, GATA4 and Tbx5. Cardiovasc Res. 2004;64:402–11. 48. Dupays L, Jarry-Guichard T, Mazurais D, Calmels T, Izumo S, Gros D, The´veniauRuissy M. Dysregulation of connexins and inactivation of NFATc1 in the cardiovascular system of Nkx2-5 null mutants. J Mol Cell Cardiol. 2005;38:787–98. 49. Bruneau BG, Nemer G, Schmitt JP, Charron F, Robitaille L, Caron S, Conner DA, Gessler M, Nemer M, Seidman CE, Seidman JG. A murine model of Holt-Oram

1 Connexin Genes

50.

51. 52. 53. 54.

55.

56.

57. 58.

59.

60.

61.

62.

63.

64. 65.

66. 67.

23

syndrome defines roles of the T-box transcription factor Tbx5 in cardiogenesis and disease. Cell. 2001;106:709–21. Moskowitz IP, Pizard A, Patel VV, Bruneau BG, Kim JB, Kupershmidt S, Roden D, Berul CI, Seidman CE, Seidman JG. The T-Box transcription factor Tbx5 is required for the patterning and maturation of the murine cardiac conduction system. Development. 2004;131:4107–16. Bai S, Spray DC, Burk RD. Identification of proximal and distal regulatory elements of the rat connexin32 gene. Biochim Biophys Acta. 1993;1216:197–204. Piechocki MP, Toti RM, Fernstrom MJ, Burk RD, Ruch RJ. Liver cell-specific transcriptional regulation of connexin32. Biochim Biophys Acta. 2000;1491:107–22. Field JM, Tate LA, Chipman JK, Minchin SD. Identification of functional regulatory regions of the connexin32 gene promotor. Biochim Biophys Acta. 2003;1628:22–9. Koffler LD, Fernstrom MJ, Akiyama TE, Gonzalez FJ, Ruch RJ. Positive regulation of connexin32 transcription by hepatocyte nuclear factor-1. Arch Biochem Biophys. 2002;407:160–7. Houlden H, Girard M, Cockerell C, Ingram D, Wood NW, Goossens M, Walker RW, Reilly MM. Connexin 32 promotor P2 mutations: a mechanism of peripheral nerve dysfunction. Ann Neurol. 2004;56:730–4. Bondurand N, Girard M, Pingault V, Lemort N, Dubourg O, Goossens M. Human connexin 32, a gap junction protein altered in the X-linked form of Charcot-Marie-Tooth disease, is directly regulated by the transcription factor SOX10. Hum Mol Genet. 2001;10:2783–95. Piechocki MP, Burk RD, Ruch RJ. Regulation of connexin32 and connexin43 gene expression by DNA methylation in rat liver cells. Carcinogenesis. 1999;20:401–6. Martin D, Tawadros T, Meylan L, Abderrahmani A, Condorelli DF, Waeber G, Haefliger JA. Critical role of the transcriptional repressor neuron-restrictive silencer factor in the specific control of connexin36 in insulin-producing cell lines. J Biol Chem. 2003;278:53082–9. Asklund T, Appelskog IB, Ammerpohl O, Ekstrom ¨ TJ, Almqvist PM. Histone deacetylase inhibitor 4-phenylbutyrate modulates glial fibrillary acidic protein and connexin 43 expression, and enhances gap-junction communication, in human glioblastoma cells. Eur J Cancer. 2004;40:1073–81. Jung JW, Cho SD, Ahn NS, Yang SR, Park JS, Jo EH, Hwang JW, Aruoma OI, Lee YS, Kang KS. Effects of the histone deacetylases inhibitors sodium butyrate and trichostatin A on the inhibition of gap junctional intercellular communication by H2O2- and 12-Otetradecanoylphorbol-13-acetate in rat liver epithelial cells. Cancer Lett. 2005;241:301–8. Ogawa T, Hayashi T, Tokunou M, Nakachi K, Trosko JE, Chang CC, Yorioka N. Suberoylanilide hydroxamic acid enhances gap junctional intercellular communication via acetylation of histone containing connexin 43 gene locus. Cancer Res. 2005;65:9771–8. Hernandez M, Shao Q, Yang XJ, Luh SP, Kandouz M, Batist G, Laird DW, AlaouiJamali MA. A histone deacetylation-dependent mechanism for transcriptional repression of the gap junction gene cx43 in prostate cancer cells. Prostate. 2006;66:1151–61. Vinken M, Henkens T, Vanhaecke T, Papeleu P, Geerts A, Van Rossen E, Chipman JK, Meda P, Rogiers V. Trichostatin A enhances gap junctional intercellular communication in primary cultures of adult rat hepatocytes. Toxicol Sci. 2006;91:484–92. Schiavi A, Hudder A, Werner R. Connexin43 mRNA contains a functional internal ribosome entry site. FEBS Lett. 1999;464:118–22. Lahlou H, Fanjul M, Pradayrol L, Susini C, Pyronnet S. Restoration of functional gap junctions through internal ribosome entry site-dependent synthesis of endogenous connexins in density-inhibited cancer cells. Mol Cell Biol. 2005;25:4034–45. Bushati N, Cohen SM. microRNA functions. Ann Rev Cell Dev Biol. 2007;23:175–205. Anderson C, Catoe H, Werner R. MIR-206 regulates connexin43 expression during skeletal muscle development. Nucleic Acids Res. 2006;34:5863–71.

24

E.C. Beyer and V.M. Berthoud

68. Yang B, Lin H, Xiao J, Lu Y, Luo X, Li B, Zhang Y, Xu C, Bai Y, Wang H, Chen G, Wang Z. The muscle-specific microRNA miR-1 regulates cardiac arrhythmogenic potential by targeting GJA1 and KCNJ2. Nat Med. 2007;13:486–91. 69. Meyer J, Mai M, Ortega G, Mossner R, Lesch KP. Mutational analysis of the connexin 36 gene (CX36) and exclusion of the coding sequence as a candidate region for catatonic schizophrenia in a large pedigree. Schizoph Res. 2002;58:87–91. 70. Aleksic B, Ishihara R, Takahashi N, Maeno N, Ji X, Saito S, Inada T, Ozaki N. Gap junction coding genes and schizophrenia: a genetic association study. J Hum Genet. 2007;52:498–501. 71. Ni X, Valente J, Azevedo M, Pato M, Pato C, Kennedy JL. Connexin50 gene on human chromosome 1q21 is associated with schizophrenia in matched case-control and familybased studies. J Med Genet. 2007;44:532–6 72. White TW, Bruzzone R, Goodenough DA, Paul DL. Mouse Cx50, a functional member of the connexin family of gap junction proteins, is the lens fiber protein MP70. Mol Biol Cell. 1992;3:711–20. 73. Musso M, Balestra P, Bellone E, Cassandrini D, Di Maria E, Doria LL, Grandis M, Mancardi GL, Schenone A, Levi G, Ajmar F, Mandich P. The D355V mutation decreases EGR2 binding to an element within the Cx32 promotor. Neurobiol Dis. 2001;8:700–6. 74. Carrasquillo MM, Zlotogora J, Barges S, Chakravarti A. Two different connexin 26 mutations in an inbred kindred segregating non-syndromic recessive deafness: implications for genetic studies in isolated populations. Hum Mol Genet. 1997;6:2163–72. 75. Deschenes SM, Walcott JL, Wexler TL, Scherer SS, Fischbeck KH. Altered trafficking of mutant connexin32. J Neurosci. 1997;17:9077–84. 76. Fleishman SJ, Sabag AD, Ophir E, Avraham KB, Ben Tal N. The structural context of disease-causing mutations in gap junctions. J Biol Chem. 2006;281:28958–63. 77. VanSlyke JK, Deschenes SM, Musil LS. Intracellular transport, assembly, and degradation of wildtype and disease-linked mutant gap junction proteins. Mol Biol Cell. 2000;11:1933–46. 78. Berthoud VM, Minogue PJ, Guo J, Williamson EK, Xu X, Ebihara L, Beyer EC. Loss of function and impaired degradation of a cataract-associated mutant connexin50. Eur J Cell Biol. 2003;82:209–21. 79. Minogue PJ, Liu X, Ebihara L, Beyer EC, Berthoud VM. An aberrant sequence in a connexin46 mutant underlies congenital cataracts. J Biol Chem. 2005;280:40788–95. 80. Abrams CK, Bennett MVL, Verselis VK, Bargiello TA. Voltage opens unopposed gap junction hemichannels formed by a connexin 32 mutant associated with X-linked Charcot-Marie-Tooth disease. Proc Nat Acad Sci USA. 2002;99:3980–4. ´ ´ ndez JM, Glass JD, Scherer SS, Mintz M, Barrio 81. Liang GS, de Miguel M, Gomez-Herna LC, Fischbeck KH. Severe neuropathy with leaky connexin32 hemichannels. Ann Neurol. 2005;57:749–54. 82. Beyer EC, Rodriguez J, Minogue PJ, Pal J, Berthoud VM, Ebihara L. Molecular mechanism underlying a congenital ‘zonular nuclear’ pulverulent cataract associated with a connexin50 mutation. ARVO Abstracts 2007;3635. 83. Gong X, Li L, Cheng C, White T, Mathias R, Chang B, Xia C-H, Gap junction mediated cataract formation and prevention. ARVO Abstracts 2007;3182. 84. Oh S, Ri Y, Bennett MVL, Trexler EB, Verselis VK, Bargiello TA. Changes in permeability caused by connexin 32 mutations underlie X-linked Charcot-Marie-Tooth disease. Neuron. 1997;19:927–38. 85. Bruzzone R, Cohen-Salmon M. Hearing the messenger: Ins(1,4,5)P3 and deafness. Nat Cell Biol. 2005;7:14–6. 86. Abrams CK, Freidin M, Bukauskas F, Dobrenis K, Bargiello TA, Verselis VK, Bennett MV, Chen L, Sahenk Z. Pathogenesis of X-linked Charcot-Marie-Tooth disease: differential effects of two mutations in connexin 32. J Neurosci. 2003;23:10548–58.

1 Connexin Genes

25

87. Pal JD, Berthoud VM, Beyer EC, Mackay D, Shiels A, Ebihara L. Molecular mechanism underlying a Cx50-linked congenital cataract. Am J Physiol. 1999;276: C1443–6. 88. Bruzzone R, Gome`s D, Denoyelle E, Duval N, Perea J, Veronesi V, Weil D, Petit C, Gabellec MM, D’Andrea P, White TW. Functional analysis of a dominant mutation of human connexin26 associated with nonsyndromic deafness. Cell Commun Adhes. 2001;8:425–31. 89. Marziano NK, Casalotti SO, Portelli AE, Becker DL, Forge A. Mutations in the gene for connexin 26 (GJB2) that cause hearing loss have a dominant-negative effect on connexin 30. Hum Mol Genet. 2003;12:805–12. 90. Rouan F, White TW, Brown N, Taylor AM, Lucke TW, Paul DL, Munro CS, Uitto J, Hodgins MB, Richard G. Trans-dominant inhibition of connexin-43 by mutant connexin-26: implications for dominant connexin disorders affecting epidermal differentiation. J Cell Sci. 2001;114:2105–13. 91. Gollob MH, Jones DL, Krahn AD, Danis L, Gong XQ, Shao Q, Liu X, Veinot JP, Tang AS, Stewart AF, Tesson F, Klein GJ, Yee R, Skanes AC, Guiraudon GM, Ebihara L, Bai D. Somatic mutations in the connexin 40 gene (GJA5) in atrial fibrillation. N Engl J Med. 2006;354:2677–88. 92. Firouzi M, Ramanna H, Kok B, Jongsma HJ, Koeleman BP, Doevendans PA, Groenewegen WA, Hauer RN. Association of human connexin40 gene polymorphisms with atrial vulnerability as a risk factor for idiopathic atrial fibrillation. Circ Res. 2004;95:e29–33. 93. Boerma M, Forsberg L, Van Zeijl L, Morgenstern R, De Faire U, Lemne C, Erlinge D, Thulin T, Hong Y, Cotgreave IA. A genetic polymorphism in connexin 37 as a prognostic marker for atherosclerotic plaque development. J Intern Med. 1999;246:211–8. 94. Yeh HI, Chou Y, Liu HF, Chang SC, Tsai CH. Connexin37 gene polymorphism and coronary artery disease in Taiwan. Int J Cardiol. 2001;81:251–5. 95. Wong CW, Christen T, Pfenniger A, James RW, Kwak BR. Do allelic variants of the connexin37 1019 gene polymorphism differentially predict for coronary artery disease and myocardial infarction? Atherosclerosis. 2007;191:355–61. 96. Yamada Y, Izawa H, Ichihara S, Takatsu F, Ishihara H, Hirayama H, Sone T, Tanaka M, Yokota M. Prediction of the risk of myocardial infarction from polymorphisms in candidate genes. N Engl J Med. 2002;347:1916–23. 97. Hirashiki A, Yamada Y, Murase Y, Suzuki Y, Kataoka H, Morimoto Y, Tajika T, Murohara T, Yokota M. Association of gene polymorphisms with coronary artery disease in low- or high-risk subjects defined by conventional risk factors. J Am Coll Cardiol. 2003;42:1429–37. 98. Listı` F, Candore G, Lio D, Russo M, Colonna-Romano G, Caruso M, Hoffmann E, Caruso C. Association between C1019T polymorphism of connexin37 and acute myocardial infarction: a study in patients from Sicily. Int J Cardiol. 2005;102(2):269–71. 99. Collings A, Islam MS, Juonala M, Rontu R, Ka¨honen M, Hutri-Ka¨honen N, Laitinen ¨ ¨ T, Marniemi J, Viikari JS, Raitakari OT, Lehtima¨ki TJ. Associations between connexin37 gene polymorphism and markers of subclinical atherosclerosis: the Cardiovascular Risk in Young Finns study. Atherosclerosis 2007;195:379–84. 100. Kumari SS, Varadaraj K, Valiunas V, Ramanan SV, Christensen EA, Beyer EC, Brink PR. Functional expression and biophysical properties of polymorphic variants of the human gap junction protein connexin37. Biochem Biophys Res Commun. 2000;274:216–24. 101. Bergoffen J, Scherer SS, Wang S, Scott MO, Bone LJ, Paul DL, Chen K, Lensch MW, Chance PF, Fischbeck KH. Connexin mutations in X-linked Charcot-Marie-Tooth disease. Science. 1993;262:2039–42. 102. Uhlenberg B, Schuelke M, Ru¨schendorf F, Ruf N, Kaindl AM, Henneke M, Thiele H, Stoltenburg-Didinger G, Aksu F, Topalolu H, Nu¨rnberg P, Hu¨bner C, Weschke B, Grtner J. Mutations in the gene encoding gap junction protein 12 (Connexin 46.6) cause Pelizaeus-Merzbacher-like disease. Am J Hum Genet. 2004;75:251–60.

26

E.C. Beyer and V.M. Berthoud

103. Liu XZ, Xia XJ, Adams J, Chen ZY, Welch KO, Tekin M, Ouyang XM, Kristiansen A, Pandya A, Balkany T, Arnos KS, Nance WE. Mutations in GJA1 (connexin 43) are associated with nonsyndromic autosomal recessive deafness. Hum Mol Genet. 2001;10:2945–51. 104. Richard G, White TW, Smith LE, Bailey RA, Compton JG, Paul DL, Bale SJ. Functional defects of Cx26 resulting from a heterozygous missense mutation in a family with dominant deaf-mutism and palmoplantar keratoderma. Hum Genet. 1998;103:393–9. 105. Uyguner O, Tukel T, Baykal C, Eris H, Emiroglu M, Hafiz G, Ghanbari A, Baserer N, Yuksel-Apak M, Wollnik B. The novel R75Q mutation in the GJB2 gene causes autosomal dominant hearing loss and palmoplantar keratoderma in a Turkish family. Clin Genet. 2002;62:306–9. 106. Richard G, Rouan F, Willoughby CE, Brown N, Chung P, Ryynnen M, Jabs EW, Bale SJ, DiGiovanna JJ, Uitto J, Russell L. Missense mutations in GJB2 encoding connexin26 cause the ectodermal dysplasia keratitis-ichthyosis-deafness syndrome. Am J Hum Genet. 2002;70:1341–8. 107. Heathcote K, Syrris P, Carter ND, Patton MA. A connexin 26 mutation causes a syndrome of sensorineural hearing loss and palmoplantar hyperkeratosis (MIM 148350). J Med Genet. 2000;37:50–1. 108. Maestrini E, Korge BP, Ocan˜a-Sierra J, Calzolari E, Cambiaghi S, Scudder PM, Hovnanian A, Monaco AP, Munro CS. A missense mutation in connexin26, D66H, causes mutilating keratoderma with sensorineural deafness (Vohwinkel’s syndrome) in three unrelated families. Hum Mol Genet. 1999;8:1237–43. ´ 109. Lopez-Bigas N, Olive´ M, Rabionet R, Ben-David O, Martı´ nez-Matos JA, Bravo O, Banchs I, Volpini V, Gasparini P, Avraham KB, Ferrer I, Arbone´s ML, Estivill X. Connexin 31 (GJB3) is expressed in the peripheral and auditory nerves and causes neuropathy and hearing impairment. Hum Mol Genet. 2001;10:947–52. 110. Richard G, Smith LE, Bailey RA, Itin P, Hohl D, Epstein EH, DiGiovanna JJ, Compton JG, Bale SJ. Mutations in the human connexin gene GJB3 cause erythrokeratodermia variabilis. Nat Genet. 1998;20:366–9. 111. Richard G, Brown N, Smith LE, Terrinoni A, Melino G, Mackie RM, Bale SJ, Uitto J. The spectrum of mutations in erythrokeratodermias-novel and de novo mutations in GJB3. Hum Genet. 2000;106:321–9. 112. Macari F, Landau M, Cousin P, Mevorah B, Brenner S, Panizzon R, Schorderet DF, Hohl D, Huber M. Mutation in the gene for connexin 30.3 in a family with erythrokeratodermia variabilis. Am J Hum Genet. 2000;67:1296–301. 113. Richard G, Brown N, Rouan F, Van der Schroeff JG, Bijlsma E, Eichenfield LF, Sybert VP, Greer KE, Hogan P, Campanelli C, Compton JG, Bale SJ, DiGiovanna JJ, Uitto J. Genetic heterogeneity in erythrokeratodermia variabilis: Novel mutations in the connexin gene GJB4 (Cx30.3) and genotype-phenotype correlations. J Invest Dermatol. 2003;120:601–9. 114. Lamartine J, Munhoz Essenfelder G, Kibar Z, Lanneluc I, Callouet E, Laoudj D, LemaY¨tre G, Hand C, Hayflick SJ, Zonana J, Antonarakis S, Radhakrishna U, Kelsell DP, Christianson AL, Pitaval A, Der Kaloustian V, Fraser C, Blanchet-Bardon C, Rouleau GA, Waksman G. Mutations in GJB6 cause hidrotic ectodermal dysplasia. Nat Genet. 2000;26:142–4. 115. Shiels A, Mackay D, Ionides A, Berry V, Moore A, Bhattacharya S. A missense mutation in the human connexin50 gene (GJA8) underlies autosomal dominant ‘zonular pulverulent’ cataract, on chromosome 1q. Am J Hum Genet. 1998;62:526–32. 116. Mackay D, Ionides A, Kibar Z, Rouleau G, Berry V, Moore A, Shiels A, Bhattacharya S. Connexin46 mutations in autosomal dominant congenital cataract. Am J Hum Genet. 1999;64:1357–64. 117. Paznekas WA, Boyadjiev SA, Shapiro RE, Daniels O, Wollnik B, Keegan CE, Innis JW, Dinulos MB, Christian C, Hannibal MC, Jabs EW. Connexin 43 (GJA1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet. 2003;72:408–18.

Chapter 2

Gap Junction Channel Structure Mark Yeager

Abstract Gap junction channels connect the cytoplasms of adjacent cells by the end-to-end docking of single-membrane hemichannels, each formed by a sixfold symmetric ring of connexin monomers. The connexins constitute a multigene family of polytopic membrane proteins that have four transmembrane hydrophobic domains, M1 to M4, two extracellular loops, E1 and E2, with the amino terminus and carboxyl terminus located cytoplasmically. There is a single cytoplasmic loop between M2 and M3. The different connexin isoforms can interact structurally in various ways: homomeric hemichannels are composed of a single connexin isoform; heteromeric hemichannels are composed by at least two different isoforms; homotypic junctional channels are formed by twelve identical connexin subunits; heterotypic channels are formed by two hemichannels that are each homomeric for different isoforms. The expression of multiple connexin isoforms in the same cell type, the multiplicity of isoforms, as well as their different structural combinations, likely provide exquisite functional tuning of this unique family of membrane channels. In spite of the diverse subunit compositions, the fundamental structure of the hemichannel is probably similar in unpaired hemichannels and junctional channels, and for channels formed by different connexin isoforms. The hexameric hemichannel with a central pore is clearly a conserved motif of gap junction channels that can be viewed as modular in design. Electron cryo-crystallography shows that the transmembrane region of each hemichannel is formed by a bundle of 24 a-helices that are staggered with respect to those in the apposed hemichannel. This stagger may be required for interdigitation of the structures formed by the extracellular loops across the gap, which fold at least in part as antiparallel b-strands. Sequence homology in the extracellular domains suggests common mechanisms for hemichannel interactions. The second extracellular loop, E2, guides selectivity in docking between M. Yeager (*) Department of Cell Biology, Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, United States & Department of Molecular Physiology and Biological Physics, University of Virginia Health Sciences Center, Charlottesville, VA 22908, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_2, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

27

28

M. Yeager

hemichannels formed by different isoforms. However, there is considerably more sequence variability of the portion (i.e., second half) of E2, suggesting that this region dictates specificity of hemichannel docking. The cytoplasmic loop, and especially the carboxyl-terminal domain, are the most divergent regions between different connexins, and confer unique functional or regulatory properties for channels formed by different connexins. The pore itself can accommodate molecules up to 1 kDa, which allows the cell–cell exchange of ions, metabolites, and signaling molecules to coordinate the metabolic and electrical activities of tissues. The extracellular surface of the pore is bounded by a continuous wall of protein that forms a tight seal to prevent the loss of permeants to the extracellular space. Mutagenesis, biochemical, dye transfer, and electrophysiological data, combined with computational studies, have suggested possible assignments for the four transmembrane a-helices within each subunit. The map derived by electron cryocrystallography shows that the transmembrane region of the pore within each hemichannel is bounded by twelve a-helices, two contributed by each connexin subunit. Most current models assign either M1 or M3 as pore-lining a-helices and M4 is agreed to be on the perimeter of the channel. Mapping of human mutations onto a suggested Ca model predicts that mutations that disrupt helix–helix packing impair channel function. In spite of this substantial progress in understanding the structural biology of gap junction channels, an experimentally determined structure at atomic resolution will be essential to confirm and clarify this working model. Keywords Membrane proteins Membrane channels Gap junctions Intercellular communication Electron microscopy Electron cryo-microscopy Atomic force microscopy X-ray diffraction Circular dichroism spectroscopy Substituted cysteine accessibility method Hemichannel Connexin Homomeric Heteromeric Homotypic Heterotypic Cx26 Cx32 Cx43

2.1 Introduction Cells of multicellular organisms interact with their immediate neighbors and with cells at a distance through a variety of modes of transmembrane signaling. Evidence for direct electrical coupling in cardiac muscle [1,2,3] and some neurons [4,5] has been known for decades. In addition to communication of ions between cells [6,7], there was evidence in nonexcitable tissues for intercellular passage of larger metabolites [8,9] and dyes [10]. Clearly, this process required the kinds of intimate contacts between cells, such as tight junctions and desmosomes, which had been described earlier in other contexts. The challenge, however, was to identify a specific structure responsible for the phenomena of electrical and molecular coupling. This was complicated by the inadequate distinction at that time between different types of intercellular contacts [11,12,13]. A final resolution was achieved using a membrane-impermeable colloidal form of lanthanum that remained in the extracellular spaces of en bloc fixed

2 Channel Structure

29

tissue. This electron-dense stain provided definition of a narrow gap between membranes of apposed cells at focal plaque-like regions [14] from which the term gap junction is derived. This morphology is the most consistent and recognizable feature of gap junction membranes, whether observed in situ or after biochemical isolation (Fig. 2.1). En face views of these domains revealed a closely packed array of stain-excluding particles, often showing a stain-filled center [14], which had been found previously using analogous, but less controlled, conditions of permanganate precipitation [4]. The nature of these particles was clarified with the advent of freeze-fracture techniques that provided internal views of membrane structure [15,16,17]. Closely packed arrays of intramembranous particles in the cytoplasmic half of the membrane were aligned with corresponding pits in the extracellular half of the bilayer of the apposing cell in specialized domains of the membrane where the extracellular space narrowed (Fig. 2.2). By utilizing rotary rather than unidirectional shadowing of the fracture replicas, central depressions in the particles could be detected [18]. This led to the model of the gap junction, which persists to date, as an array of aqueous channels connecting the cytoplasms of adjacent cells. The oligomeric channels are formed by the end-to-end docking of two hemichannels, also termed connexons, each of which is a hexameric cluster of protein subunits, termed connexins.

Fig. 2.1. Conventional thin-section electron microscopy has served as a keystone for defining the morphological signature of gap junctions. The characteristic appearance shows stain exclusion in the hydrophobic domains of the lipid bilayer and stain accumulation in the extracellular gap and on the cytoplasmic faces [19]. The sample enriched for cardiac gap junctions has been labeled by site-specific peptide antibodies against a synthetic peptide spanning residues 131 to 143 in Cx43. Such experiments using immunogold electron microscopy have been used to confirm the folding models shown in Fig. 2.7. The absence of gold label on single bilayer membranes and amorphous material indicates the specificity of the antibodies for cardiac gap junctions. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Yeager and Gilula [68] with permission.)

30

M. Yeager

Fig. 2.2. Electron micrographs of freeze-fractured BHK cells expressing Cx43. BHK cells express either full-length Cx43 (left) or a mutant in which the majority of the carboxyl-terminal domain has been deleted (right, L263, see Fig. 2.6a). Both replicas display plaques that exhibit particles on the P-fracture face and corresponding depressions on the E-fracture face that are typical of mammalian gap junctions. Scale bar 200 nm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Unger et al. [123] with permission.)

Conventional thin-section electron microscopy has served as a keystone for defining the morphological signature of gap junctions. The double-membrane appearance of gap junctions arises from stain exclusion in the hydrophobic domains of the lipid bilayer and stain accumulation in the extracellular gap and on the cytoplasmic faces (Fig. 2.1). Using classical staining techniques, crosssectional views of gap junctions appeared to be pentalaminar. Newer staining techniques distinguished the lipid head groups from the aqueous regions, so that the gap junction membranes appeared to be septalaminar (Fig. 2.3) [18,19]. This staining distribution distinguishes gap junctions from tight junctions, which appear pentalaminar by thin-section electron microscopy [14,19,20]. With the morphology thus defined, similar structures were identified in many tissues [20,21] of virtually every metazoan phylum studied [20,22,23,24]. As previously noted, functional demonstrations of cell–cell coupling were established in a wide range of systems, using both electrical coupling and the intercellular transfer of low molecular weight membrane-impermeable fluorescent dyes or cellular metabolites. Several studies, in particular by Flagg-Newton et al. [25], utilized an extensive array of fluorescent probes to establish an upper permeability limit for these channels of 1 kDa [25], or 2 kDa in the case of insect gap junctions [26]. In most cases the presence of morphological gap junctions could account for the coupling of cells, and in some cases represented the only identifiable intimate contact between coupled cells [27]. The current working assumption of the field is that structures defined by

2 Channel Structure

31

Fig. 2.3. Orthogonal edge view of a cardiac gap junction membrane two-dimensional (2D) crystal negatively stained with uranyl acetate. The cytoplasmic, lipid membrane, and extracellular gap regions are indicated by the letters C, M, and G, respectively. The repeating densities in the extracellular gap region have a periodicity of 42 A˚, which is half the unit cell spacing. This view is therefore down the [1,1] lattice plane. The arrows indicate repeating stain excluding areas in the cytoplasmic membrane face, which appear to suggest crystalline packing of the protein in this region. The total thickness is 250 A˚. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Yeager [70] with permission.)

morphology as gap junctions are the only mediators of the direct exchange of ions and small metabolites between cells (except for the rare occurrence of intercellular cytoplasmic bridges [see Chapter 6]).

2.2 Biochemical Isolation and Characterization After the initial characterization of gap junction morphology, identification of the biochemical composition was difficult since morphological criteria were the only means of assessing purity. Only a fraction of the cellular connexin is present as assembled gap junctions in the plasma membrane. Indeed, immunoprecipitation experiments have suggested that the pool of intracellular protein (distributed between the Golgi, endoplasmic reticulum, and lysosomes) is equal to or greater than that in the plasma membrane [28]. Even the surface membrane population may not be entirely assembled into junctional structures that are morphologically identifiable and resistant to detergent or alkali extractions used to isolate plaques (see below) [29]. In fact, functional unapposed hemichannels may be a fairly common feature of cells [30]. In the presence of low Ca2+, a variety of cultured cells take up extracellular Lucifer yellow dye in a pattern that closely correlates with the expression of connexins. These experiments are consistent with earlier claims that large open channels in the membranes of retinal horizontal cells are connexin hemichannels [31,32]. Nevertheless, it has been recently suggested that pannexin channels may be responsible for such dye uptake (see Chapter 12).

32

M. Yeager

Procedures to isolate gap junction plaques have relied on the resistance of these closely packed arrays of proteins to treatments with various nonionic detergents or with highly alkaline solutions that solubilize most cell membranes [33,34]. Not only do these treatments leave the protein arrays intact, but they also do not disrupt the pairing of the membranes (Fig. 2.4). The second most important step in all isolation procedures has been separation on the basis of density. Most procedures have utilized discontinuous sucrose gradients to isolate plasma membranes prior to detergent or alkali extraction [35,36,37]. The exact conditions may vary with the tissue, but plasma membranes typically sediment with a density between 1.15 and 1.20 g/cm3. Of more specific relevance to gap junctions, density gradients have been employed following the

Fig. 2.4. Transmission electron micrograph of a cardiac gap junction membrane plaque negatively stained with uranyl acetate. The membrane sheet is formed by a mosaic of crystalline domains, the largest of which is 0.5 mm2. This gap junction preparation has undergone protease cleavage in the region of residues 252 to 271 [68] to release the cytoplasmic carboxyl-terminal domain. This presumably accounts for the clarity of the hemichannels compared with images of gap junctions containing uncleaved Cx43. Such 2D crystals contain structural information to 15 A˚ resolution, which is sufficient to define the molecular boundary and quaternary structure of the hemichannels. Scale bar 1000 mm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Yeager [69] with permission.)

2 Channel Structure

33

solubilization and washing of the membranes to separate the gap junctions from the less dense lipid residue and more dense matrix and cytoskeletal material. The junctions usually float at an interface spanning the densities of 1.127 (30% weight per volume [w/v] sucrose) to 1.200 g/cm3 (54% w/v sucrose), consistent with an estimated density of 1.165 g/cm3 for gap junction plaques, based on equilibrium centrifugation on a continuous gradient [38]. Proteases were used for the initial isolations of gap junctions, since they had no obvious effect on junctional morphology [33]. In fact, the first application of the term connexin referred to proteolytic fragments of what we now recognize as the principal protein component of liver gap junctions [39], emphasizing the precarious nature of a purely morphological assay. In fact, this problem reemerged in the early protocols to isolate cardiac gap junctions, in which the principal protein component was thought to have a molecular weight comparable to the liver gap junction protein, that is, 28 kDa [40,41]. Only when stringent efforts were applied to prevent proteolysis (e.g., addition of phenylmethylsulfonylfluoride) during isolation was the full-length 43 kDa cardiac gap junction protein identified [42].

2.3 The Connexin Gene Family The characterization and sequencing of the first connexins heralded the use of antibodies and degenerate oligonucleotides to isolate complementary DNA (cDNA) clones for other members of the connexin family. Both approaches proved successful [43,44,45], which then led to low stringency hybridization and polymerase chain reaction screens of genomic and cDNA libraries that yielded a plethora of family members. More than 20 unique connexins have been identified to date in virtually all multicellular organisms, from mesozoa to humans [46,47], including a putative invertebrate connexin [48,49]. These clones and their relationships have been reviewed by Beyer [50], Willecke et al. [51], Kumar and Gilula [52,53], Goodenough et al. [54] and Nicholson and Bruzzone [55] (see Chapter 1). In the absence of a clear understanding of the functional differences between connexins, it has been difficult to establish a logical nomenclature for distinguishing the various members. For example, the acetylcholine receptor displayed a clear distinction between subunits required for functional assembly; consequently, an a, b, g, d terminology was utilized. Since gap junctions can clearly assemble from a single type of subunit, such a functional distinction is not applicable. Therefore, one practical approach was to add the molecular weight of each variant as a suffix to the descriptor Cx [56]. Hence, the proteins that compose gap junctions in liver that run at 28 kDa and 21 kDa by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) are now known as Cx32 and Cx26, respectively, based on their predicted formula weights established from the deduced protein sequence. Although this

34

M. Yeager

terminology carries the advantage of making no functional inferences, it can lead to confusion in comparisons between species. For example, chick Cx46 is orthologous to rat Cx50, and chick Cx56 is orthologous to rat Cx46 based on physiological properties and sequence [57] (see Chapter 1). As our functional understanding of these proteins increases, refinements in the nomenclature may emerge (see Appendix). On the basis of the broad division of the currently known connexin gene family into two more closely related subgroups, using a and b terminology [53,58] and, subsequently, an analogous group II and I classification by Bennett et al. [59,60] was proposed, in which subscripts are arbitrarily used to distinguish group members. The electrophysiological analysis of paired Xenopus oocytes suggests that, in general, connexins within each group tend to pair with each other [61]. The extrapolation of these comparisons of connexin sequences lead to the surprising conclusion that the a-group and b-group diverged at some time near the prokaryotic/eukaryotic split, well before multicellularity arose [62].

2.4 Lipid Bilayer Structure Revealed by X-Ray Diffraction Analysis Low-angle X-ray scattering of gap junctions in aqueous buffers [63] allowed more precise examination of the distribution of lipid and protein than traditional staining patterns of thin sections (Fig. 2.1). Specimens were prepared by pelleting gap junctions into stacks to separate the equatorial diffraction arising from the hexagonal packing of the channels from the meridional diffraction arising from the electron density fluctuations perpendicular to the plane of the gap junction membranes. The double-membrane structure of gap junctions gives rise to a center of symmetry in the extracellular gap. In this case, the one-dimensional electron density profile perpendicular to the membrane reduces to the sum of a series of cosine functions given by the amplitudes of the diffraction fringes along the meridian. The center of symmetry restricts the phases of the cosine functions to either 0 or 180 degrees. Phase assignments were based on the minimum wavelength principle [64]; since the thickness of liver gap junctions is 150 A˚, the phases for adjacent fringes in the meridional diffraction pattern must be of opposite sign if the peak separation in the diffraction pattern is < 2/150 A˚. In addition, density maps not consistent with the low-resolution images provided by thin-section electron microscopy could be excluded. The computed one-dimensional electron density maps were dominated by four peaks that corresponded to the location of the electron-dense phospholipid head groups in the two lipid bilayers composing the gap junction (Fig. 2.5). This map provided precise boundaries for the lipid bilayer domains of liver gap junctions; the head groups within each bilayer were separated by 42 A˚, the

2 Channel Structure

35

Fig. 2.5. One-dimensional electron density profile perpendicular to the plane of the gap junction. The high-density peaks correspond to the polar head groups of the bilayer lipids. The lowdensity minima at 43 A˚ are occupied by the lipid hydrocarbon chains and transmembrane protein domains, and solvent in the aqueous pore. The elevated density in the vicinity of the origin arises from protein domains in the extracellular gap. The continuous curve has been corrected for partial stacking of the membranes in the specimen. The dashed curve is the profile before correction. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Makowski et al. [63] with permission.)

hydrophobic portion of the bilayers was 32 A˚ thick, and the extracellular gap had a thickness of 35 A˚. The electron density in the aqueous extracellular gap and cytoplasmic regions could be adjusted by varying the amount of sucrose in the buffer [65]. In this way the distribution of protein in gap junctions could be examined. The analysis compared the predicted electron density for different models with the experimental density profiles derived for gap junctions suspended in aqueous buffers with varying electron densities. Although the transmembrane distribution of connexins is now an accepted fact, the X-ray scattering analysis provided the first compelling evidence for it. The electron density within the center of the hydrophobic region was much higher than expected for pure lipid, indicating the transmembrane disposition of the protein. The higher electron density in the extracellular gap compared with the electron density of the buffer confirmed that the channels span the gap, as expected from thin-section electron micrographs.

2.5 Hexameric Hemichannel Structure Revealed by Low-Resolution Electron Microscopy The paracrystalline, hexagonal packing of the annular channels in liver gap junctions is a critical feature that allowed the application of electron crystallography. Electron microscopy followed by digital image processing remains,

36

M. Yeager

to date, the approach that has enriched our knowledge of gap junction structure more than any other method. The first studies focused on the analysis of mouse [66] and rat liver [67] gap junctions. Thereafter, negatively stained heart [68,69,70], lens [71], liver [72,73], and arthropod [74] gap

Fig. 2.6. Computed diffraction patterns and projection density maps for gap junctions containing a recombinant form of rat heart Cx43 in which the majority of the carboxyl-terminal domain has been deleted. Cx43 is truncated at L263 [111]. (a) Negatively stained crystals display diffraction to 15 A˚ resolution. (b) When unstained crystals are examined using electron cryomicroscopy (cryoEM), the diffraction patterns extend to 7 A˚ resolution. (c) The projection density map shows the hexameric hemichannels, similar to maps at comparable resolution from rat heart gap junctions [68,69]. (d) The projection density map at 7 A˚ resolution [111] displays three major features: a ring of circular densities centered at a radius of 17 A˚ interpreted as a-helices that line the channel, a ring of densities centered at a radius of 33 A˚ interpreted as a-helices that are most exposed to the lipid, and a continuous band of density at 25 A˚ radius separating the two groups of helices. The hexagonal lattice had parameters a = b = 79 A˚ and a = 120 A˚. The spacing between grid bars is 40 A˚. (A highresolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Modified from Yeager and Nicholson [112] with permission.)

2 Channel Structure

37

junctions were manipulated to generate two-dimensional (2D) crystals amenable to electron microscopy and image analysis. Conserved features of all of the different gap junction structures were the annular appearance of the channels surrounding a central region of negative stain [75]. Based on connexin molecular weight and protein density, the annular appearance supports a model in which the channel is formed by an oligomeric radial cluster of connexin subunits. The diffraction patterns of negatively stained specimens extended to a nominal resolution of 15 to 20 A˚, and the computed phases to this level of resolution are consistent with plane group symmetry p6 (Fig. 2.6a). Therefore, in the published density maps, sixfold symmetry has been enforced, and the six domains correspond to a superposition of the density within the two hexameric hemichannels (Fig. 2.6c). Several caveats regarding the method of electron microscopy and image analysis should be noted. The method is an averaging technique, so any molecular heterogeneity will not be revealed in an averaged map. For instance, liver gap junctions contain not only Cx32 but also Cx26. At 15 to 20 A˚ resolution, subtle differences in the 2D projection density maps of channels formed by Cx32 or Cx26 hemichannels would be obscured. In addition, an averaged map at 15 A˚ resolution would also probably obscure hemichannels in which oligomers are formed by both Cx32 and Cx26 connexins. An example is provided by the acetylcholine receptor, which is formed by a heteropentamer of 2a, b, g, and d subunits. Projection density maps at 17 A˚ resolution did not reveal major differences between the homologous but different subunits [76].

2.6 Dodecameric Channel Structure Revealed by Low-Resolution Three-Dimensional Electron Crystallography In spite of the limiting resolution, electron microscopy and image analysis have been powerful methods for defining the three-dimensional (3D) molecular envelope of the hemichannels in liver gap junctions. As pioneered by Henderson and Unwin [77] in their studies of bacteriorhodopsin, the approach relies on the ability to tilt the membrane crystals via the goniometer stage of the electron microscope [78]. The projected views represent planar sections through the 3D density map. The 2D sections are combined in Fourier space, and the 3D Fourier transform is interpolated in the regions for which sections were not obtained. These continuous lattice lines are sampled as if the specimen is a 3D crystal, and the 3D density map is then obtained by back Fourier transformation. By this approach, low-resolution 3D density maps have been derived for rat liver gap junctions either stained with uranyl acetate [72,79] or in the frozenhydrated state [80]. Specimens imaged by negative stain microscopy are sensitive to beam-induced shrinkage, and the thickness of the liver gap junction was adjusted to match the membrane thickness measured from X-ray scattering of hydrated specimens. In addition, the stain puddles around the perimeter of the

38

M. Yeager

specimen so that for gap junctions the images reveal the molecular boundary of the channel. (Interestingly, the cationic stain uranyl acetate was able to penetrate the pore of the channel, whereas the anionic stain phosphotungstic acid was excluded [68]. This may imply a concentration of negative charges in the cytoplasmic vestibule of the channel.) In the case of electron cryomicroscopy (cryoEM) of unstained frozen-hydrated specimens, the derived maps reveal not only the surface contours of a structure but also the internal organization. CryoEM of unstained samples also minimizes the potential specimen shrinkage and distortion encountered with negative stain microscopy. Nonetheless, the low-resolution 3D maps by negative stain and by cryoEM were quite similar. The basic design of the liver gap junction hemichannel is a cluster of six rodlike subunits oriented roughly perpendicular to the membrane plane. The subunits have a diameter of 25 A˚ and surround a central channel that has a diameter of 20 A˚. At a resolution of 18 A˚, the contours of the channel within the hydrophobic interior of the bilayer were not revealed, though the hydrophobic core itself is 30 A˚. The hemichannel extends 20 A˚ into the extracellular space so that the gap is 35 to 40 A˚ thick. The thickness of the extracellular gap is larger than observed by conventional thin-section electron microscopy, possibly due to shrinkage during dehydration and embedding. Examinations of the contrast perpendicular to the membranes’ plane reveals that the negative stain in the center of the hemichannel in en face views primarily resides in an extracellular vestibule with less stain accumulation on the cytoplasmic face of the channel. Interestingly, the 3D maps only resolved density extending 10 A˚ into the cytoplasmic space. This was much less than expected, knowing that the aminoterminal domain (NT), the substantially longer carboxyl-terminal domain (CT), as well as the cytoplasmic loop (CL) between the second (M2) and third (M3) membrane-spanning domains should all reside on the cytoplasmic face and account for more than a third of the mass of Cx32, the predominant component in rat liver gap junctions. The CT contributes the majority of cytoplasmic residues, and it is likely that the cytoplasmic domains exhibit conformational flexibility so that their density would be smeared out during crystallographic averaging. Another explanation is that the map is less well resolved in the z direction perpendicular to the membrane because the specimens can only be tilted up to 60 degrees. The resulting missing cone of data tends to smear the map in this direction. A comparison of the X-ray scattering density profiles of native and trypsintreated mouse liver gap junctions showed that protein extends 85 to 90 A˚ from the center of the gap [65]. The lipid head group region on the cytoplasmic leaflet of the bilayer is centered 64 A˚ from the center of the gap. Assuming a head group region 10 A˚ thick, the cytoplasmic protein extends 15 to 20 A˚ beyond the membrane surface, which should be sufficient to accommodate the connexin domains predicted to reside on the cytoplasmic face, particularly given the higher complement of Cx26 in mouse liver gap junctions; Cx26 has a much shorter CT than Cx32. Image analysis of edge views suggested that cardiac gap

2 Channel Structure

39

junctions have a thickness of 250 A˚ so that the protein on the cytoplasmic face could extend 40 to 50 A˚ from the membrane surface, consistent with the larger size of the Cx43 CT compared with the Cx32 CT [81].

2.7 Topological Analysis Identified Four Hydrophobic Domains per Connexin Subunit The 3D density map determined by electron microscopy and image analysis at 15 to 20 A˚ resolution defined the molecular boundary of the cylindrical connexin subunits and the hexameric quaternary arrangement in the hemichannel [72,80]. However, the low-resolution of the maps did not resolve the folding of the polypeptide chain. Hydropathy analyses of the complete sequences of the various connexin family members revealed four conserved stretches of 20 to 28 hydrophobic residues suggesting four transmembrane regions (Fig. 2.7). These domains, M1 to M4, and particularly the cysteine-rich hydrophilic loops connecting the first and second spans (E1) and the third and fourth spans (E2), were the most conserved regions of the protein family. The most variable domains, both in length and sequence, were the CT and the CL, whose cytoplasmic locations were determined by protease protection assays of isolated gap junctions [40,82,83]. Direct biochemical confirmation of the connexin membrane topology suggested by hydropathy analysis was greatly aided by the close association of the paired membranes in isolated gap junctions. That is, the narrow extracellular gap of 20 A˚ did not provide sufficient space for penetration by proteases and antibodies. In contrast, the cytoplasmic surfaces were readily accessible, as exemplified by the decoration of gold-labeled secondary antibodies bound to site-specific peptide antibodies directed against a sequence predicted to reside on the cytoplasmic surface (Fig. 2.1). Accessibility to the extracellular surfaces required strongly denaturing conditions (6 M urea at pH 12) to split the membranes [84,85], which could alter protein conformation. Nevertheless, detailed topological studies of three members of the connexin family (Cx32, Cx43, and Cx26; see references in legend to Fig. 2.7) have utilized a combination of site-specific, peptide antibodies and treatments with highly specific proteases. In addition to the broad view of folding provided by these studies, a number of more specific conclusions about interactions between subdomains can be gleaned. Limited accessibility of the NT (20 amino acids [aa]) to both proteases [86] and many antibodies suggests that residues eleven to 22 are protected by either folding of the polypeptide or close association with the membrane. Protection of the CT of Cx26 was also demonstrated by a proteolytic time course, which showed that the binding of an antibody directed against the CT required CL cleavage. After exposure, this domain could also be cleaved by proteases [87].

40

M. Yeager

Fig. 2.7. Folding models for connexin proteins. The amino acid sequences of (a) Cx43 [56], (b) Cx32 [43,44], and (c) Cx26 [45] were deduced from complementary DNA (cDNA) analysis, and the residues are coded as follows: hydrophobic in yellow, acidic in red, basic in blue, and cysteine in green. Hydropathy analysis predicts four membrane-spanning domains, referred to as M1, M2, M3, and M4, proceeding from the amino terminus to the carboxyl terminus. The predicted locations of the extracellular and cytoplasmic regions were confirmed with sitedirected antibodies (blue and yellow bars indicate cytoplasmic and extracellular epitopes, respectively) and proteases (solid arrowheads indicate sites accessible from the cytoplasmic face; open arrowheads indicate sites only accessible after separation of the membranes). CT, chymotrypsin; PA, proteinase A; V8, staphylococcal V8 protease; TH, thrombin; LC, Lys-C protease; AN, Asp-N protease. Note the three conserved cysteine residues (shown in green) located in each of the extracellular loops (designated E1 and E2) of the three connexins. Cx43 (a) and Cx32 (b) indicate the locations of various functionally important residues (H95 in Cx43 and D2 and P87 in Cx32) and domains 1 (yellow) and 2 (blue) in Cx43 and 3 (orange), 4 (green), and 5 (violet) in Cx32 as determined by mutagenesis and chimera studies. Indicated sites of covalent modification are based on consensus sequences, modification of synthetic

2 Channel Structure

41

2.8 Evidence for Transmembrane a-Helical Structure Based on Membrane Topology, X-Ray Diffraction, and Circular Dichroism Spectroscopy Prior to 1997, the available density maps provided by X-ray diffraction and electron microscopy of gap junctions were not of sufficient resolution to resolve elements of secondary structure. Therefore, hypothetical working models were based on inferences from (1) the topological mapping of the primary sequence (Fig. 2.7) combined with the 3D density maps provided by electron microscopy and image analysis [88], (2) analysis of X-ray diffraction patterns of isolated gap junctions that had been oriented by centrifugation (Fig. 2.5) [65,89], and (3) circular dichroism (CD) spectroscopy of gap junction plaques (Fig. 2.8) [90,91]: 1. The topological studies described above support a folding model with four transmembrane domains of sufficient length so that they could be folded as ahelices. By analogy with soluble proteins having four antiparallel a-helices, a model was proposed by Milks et al. [88] in which the four transmembrane domains of Cx32 were folded as a four-helix bundle, which could be accommodated within the cross-sectional area of each connexin, estimated as 500 A˚2 from 3D density maps of liver gap junctions. M3 contains a series of polar amino acids spaced so that they could form an amphipathic a-helix; therefore, it was proposed that six M3 transmembrane domains from each of the six subunits form the innermost boundary of the transmembrane pore. 2. X-ray diffraction patterns recorded from oriented gap junction membranes displayed sharp fringes centered at 1/4.7 A˚ on the meridian and diffraction

Fig. 2.7. (continued) peptides in vitro, and mutagenesis studies. The significance of the functional domains (circled numbers) and specifically mutated residues are as follows: Domain 1 (yellow) is the predominant determinant for specificity of heterotypic interactions between Cx43, Cx50, and Cx46. This result was based on chimeras of Cx50 and Cx46. Domains 2 (blue) impart high sensitivity to pH gating of Cx43. Gating by pH is also influenced by the charge on H95. The SH3 domains of v-src bind to proline-rich motifs and a phosphorylated tyrosine at position 265 of Cx43. Domain 3 (orange) determines the polarity of the voltage sensor where a N2 to D2 change from Cx32 to Cx26 changes the gating polarity. Domains 3 (orange) and 4 (green) determine the voltage-gating parameters; the first extracellular loop, E1, is important in determining the transjunctional voltage required for closure; P87 serves as a critical feature of M2 involved in the transduction of the voltage response. This observation was based on studies in Cx26 but has not been confirmed for other connexins. Domain 5 (violet) constitutes potential calmodulin binding sites. Arginine residues have an inhibitory effect on the gating sensitivity of Cx32 to cytosolic acidification by CO2 exposure. The numbered references are as follows: 1, [85]; 2, [188]; 3, [86]; 4, [88]; 5, [189]; 6, [190]; 7, [191]; 8, [192]; 9, [193]; 10, [194]; 11, [195]; 12, [196]; 13, [163]; 14, [197]; 15, [68]; 16, [198]; 17, [199]; 18, [200]; 19, [201]; 20, [146]; 21, [177]; 22, [147]; [23], [143]; 24, [87]; 25, [202]; 26, [104]; 27, [203]; 28, [179]; 29, [204]; 30, [205]; 31, [206]. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (Modified from Yeager and Nicholson [112] and reproduced from Yeager [70] with permission.)

42

M. Yeager

Fig. 2.8. Circular dichroism (CD) spectroscopy of isolated cardiac gap junctions indicates substantial a-helical structure. (a) CD spectrum of rat cardiac gap junctions in which the carboxyl-terminal domain of Cx43 has been removed by protease cleavage to increase the fraction of the polypeptide contributed by the transmembrane domains. (b) CD spectra for polypeptides in a-helical (A), random coil (B), b-sheet (C), and b-turn (D) conformations [207]. Note the resemblance between the spectrum recorded from cardiac gap junctions and that for a polypeptide in an a-helical conformation. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Yeager [70] with permission.)

centered at 1/11 A˚ on the equator. These patterns were initially interpreted as b-sheets with the strands running more parallel than perpendicular to the surface. To test this hypothesis, hemichannel models were built that contained a transmembrane core based on known soluble protein structures that exhibited a-helical bundle and b-sheet conformations. In fact, the predicted diffraction patterns for purely a-helical conformations were in closer agreement with the X-ray diffraction data. For a-helices packed perpendicular to the membrane plane, a diffraction fringe at 1/5 A˚–1 would have been predicted. However, in four-helix bundle proteins, the a-helices are tilted with respect to each other. This tilting causes a shift in the fringe to 1/ 4.7 A˚–1. 3. From 190 to 240 nm, the CD spectra of proteins are sensitive to molecular geometry, and are therefore quite useful for determining protein secondary

2 Channel Structure

43

structure and monitoring conformational changes (Fig. 2.8) [92,93,94,95]. Circular dichroism spectroscopy measures the difference in absorption of left and right circularly polarized light. The magnitude of these difference measurements is quite small, so that the signal-to-noise ratio is typically <<1. Therefore, a protocol is typically used in which spectra are recorded over several hours by accumulating repetitive scans. In this way, each scan can be statistically compared with all others to detect instrumental variations during recording. A second relevant technical aspect is that the signal is dramatically affected by light scattering and results in absorption flattening in the far ultraviolet region, which will confound the prediction of secondary structure [96]. Light scattering can certainly occur with membrane specimens, and gap junction samples were sonicated and centrifuged so that only the smallest vesicles are used for CD measurements. The CD spectra of suspensions of rat liver [90] and heart gap junction vesicles [91] displayed considerable similarity to the spectra for a polypeptide with an a-helical conformation, and the estimated content of a-helix (40 to 50%) was sufficient for the four transmembrane domains of each subunit to be folded as a-helices.

2.9 Higher Resolution Density Maps Derived by Electron Cryomicroscopy Confirmed that the Transmembrane Domains Are a-Helices To resolve elements of secondary structure such as a-helices and b-sheets, the diffraction patterns from 2D crystals have to extend beyond 10 A˚ resolution. On the basis of the experience with bacteriorhodopsin [97], porin [98], plant light-harvesting complex II [99], and aquaporin-1 [100,101,102], a chemically pure protein is required to grow high-resolution 2D crystals. However, prior 2D crystallization studies of gap junctions used preparations isolated from heart and liver tissue. For example, isolated cardiac gap junctions possess very little inherent crystallinity compared with liver gap junctions. A strategy for in situ crystallization was to expose the isolated junctions to low concentrations of nonionic detergents in order to extract lipid and concentrate the protein in the membrane plane. Examination of biological specimens in the frozen-hydrated state offers the possibility of higher resolution structure analysis. Nevertheless, cryoEM studies of native gap junctions isolated from tissues such as the liver were limited to 15 to 20 A˚ resolution [80,103]. This limiting resolution was presumably based on molecular heterogeneity that could arise from (1) expression of multiple connexins within cells of the same tissue that could lead to formation of gap junctions that are assembled from different homomeric hemichannels, formation of heterotypic channels in which homomeric hemichannels are composed of different connexins, and formation of heteromeric hemichannels in which the oligomers contain multiple connexin types

44

M. Yeager

[104,105,106,107,108,109]; (2) inherent flexibility in the native protein; (3) partial denaturation during specimen preparation; (4) the presence of nonconnexin proteins that may disorder the lattice; (5) partial proteolysis; (6) different degrees of posttranslational modification; (7) rotational flexibility of the channels in the lattice; and (8) lipid heterogeneity that may prevent precise chemical interactions at the protein–lipid interface, which may be necessary for 2D crystallization of the channels. Several of these problems were overcome by expression of a single recombinant connexin in a stably transfected baby hamster kidney cell line under control of the inducible metallothionein promoter [110]. Ultrastructural studies demonstrated that a truncated form of rat Cx43 that lacked most of its CT (by truncation at L263; Cx43L263) assembled into gap junctions having the characteristic double-membrane morphology. Freeze-fracture images revealed that the Cx43L263 gap junctions formed small 2D crystals (Fig. 2.2), and their crystallinity and purity could be improved by extraction with nonionic detergents such as Tween20 and DHPC (1,2-diheptanoyl-sn-phosphocholine) [111], an approach similar to that taken for crystallization of native rat heart Cx43 gap junctions [69].

2.9.1 Two-Dimensional Projection Density Map by Electron Cryomicroscopy A projection density map derived from negatively-stained 2D crystals of Cx43L263 (Fig. 2.6c) closely resembled the maps for native Cx43 cardiac gap junctions [68,112]. A projection density cryoEM map based on the analysis of unstained, frozen-hydrated 2D crystals [111] showed that the recombinant channel had sixfold symmetry with a diameter of 65 A˚ (Fig. 2.6d). At 7 A˚ resolution, the map of the recombinant channel showed substantially more detail than the map of rat heart gap junctions at 16 A˚ resolution [68]. In particular, at 17 A˚ radius, the channel was lined by circular densities with the characteristic appearance of transmembrane a-helices that are oriented roughly perpendicular to the membrane plane [113,114,115,116,117,118,119,120,121]. A similar appearance for densities at 33 A˚ radius suggested the presence of a-helices at the interface with the membrane lipids. The two rings of a-helices were separated by a continuous band of density at a radius of 25 A˚, which arose from the superposition of projections of additional transmembrane a-helices and polypeptide density arising from the extracellular and intracellular loops within each connexin subunit. A notable feature of the projection density map was the 30 degree displacement between the rings of a-helices at 17 and 33 A˚ radius, which places constraints on possible structural models for the dodecameric channel (Fig. 2.9) [111]. As a consequence of the 30 degree displacement between the a-helices that line the channel (at 17 A˚ radius) and the a-helices that are most exposed to

2 Channel Structure

45

Fig. 2.9. Schematic models for the packing of helices, connexins, and hemichannels in the gap junction intercellular channel. Each connexin subunit is represented by a rectangle, and the transmembrane a-helices are depicted as circles. The twofold symmetry axes located in the extracellular gap generate the views of the apposed hemichannels in cell 1 (red) and cell 2 (blue) that form the intercellular channel. The model in (a) is in accordance with the observed projection density map (Fig. 2.6d) and predicts that the hemichannels within the junctional channel are rotationally staggered, as shown by the dashed lines and the arc. The superposition of the helices and this rotation dictate that the a-helices within one connexin will be superimposed with helices within two connexin subunits in the apposed hemichannel. The models shown in (b), (c), and (d) are inconsistent with the projection map shown in Fig. 2.6d. The model in (b) is representative of a class of structures that involve variable degrees of rotational stagger between the hemichannels and was generated by a rotation around a twofold symmetry axis that would give rise to strict p622 symmetry. A characteristic feature of such models is twelve density peaks at high radius. Depending on the rotational stagger (indicated by the dashed lines and the arc), the six peaks at low radius may be broadened or resolved into twelve separate peaks. The models in (c) and (d) also display p622 symmetry, but the hemichannels are not rotationally staggered. Note that the a-helices are colinear through the center of the channel, as shown by the dotted lines. In model (d) there is superposition of the connexin subunits of apposed hemichannels. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Yeager [70] with permission.)

46

M. Yeager

the lipid (at 33 A˚ radius), the two hemichannels forming the intercellular channel are rotationally staggered with respect to each other (as shown by the arcs in Figure 2.9a,b). The amount of rotational stagger dictates whether six or twelve peaks are resolved in the outer ring of a-helices at 33 A˚ radius. That is, models with less than 30 degrees of rotational stagger between the hemichannels are not consistent with the map in Figure 2.6d, since there would be twelve rather than six peaks in the outer ring of a-helices (Fig. 2.9b). In addition, the 30 degree displacement between the rings of helices is not consistent with models in which the a-helices are colinear through the center of the channel (Fig. 2.9c,d), as was suggested [88]. The model shown in Fig. 2.9a is in best agreement with the projection density map (Fig. 2.6d) that shows superposition of the resolved ahelices between the apposed connexins. Note that this model predicts that each subunit in one hemichannel will interact with two connexin subunits in the apposing hemichannel. Such an arrangement may confer stability in the docking of the hemichannels. Electron microscopy and image analysis of negatively stained gap junction plaques that had been split in the extracellular gap by urea treatment showed that the extracellular surface of each hemichannel contains six protrusions [72]. To form a tight seal with the apposed hemichannel, computer modeling of the map at 18 A˚ resolution predicted that these protrusions interdigitate in such a way that requires a 30 degree stagger [73], as was predicted from the projection cryoEM density map at 7 A˚ resolution [111].

2.9.2 Three-Dimensional Projection Density Map by Electron Cryomicroscopy To further explore the a-helical folding of the connexin polypeptide, a 3D cryoEM density map was determined at resolutions of 7.5 A˚ in the membrane plane and 21 A˚ in the vertical direction by recording and analyzing images of frozen-hydrated, tilted 2D gap junction crystals formed by Cx43L263 [122,123]. A side view of the 3D map (Fig. 2.10) showed that the recombinant gap junction channel had a thickness of 150 A˚. This reduced thickness, in comparison with 250 A˚ for native cardiac gap junction channels [70], was consistent with the lack of the 13 kDa CT in the Cx43L263 truncation mutant. Consistent with the projection map (Fig. 2.6d) [111], the outer diameter within the membrane region was 70 A˚, but the 3D map also revealed that the diameter decreased to 50 A˚ in the extracellular portion of the channel, giving the outer boundary an hourglass shape. A vertically sectioned view of the 3D map (Fig. 2.10b) revealed that the inner diameter of the channel narrowed from 40 A˚ to 15 A˚ (neglecting the contributions of amino acid side chains) in proceeding from the cytoplasmic to the extracellular side of the bilayer. The aqueous pathway then widened again to a diameter of 25 A˚ within the extracellular vestibule. Cross sections of the map within the hydrophobic region of the bilayers (Fig. 2.10c, top and bottom) displayed roughly circular contours of density

2 Channel Structure

47

Fig. 2.10. Three-dimensional structure of a recombinant Cx43 cardiac gap junction channel in which the majority of the carboxyl-terminal domain has been deleted. (a) A full side view of Cx43 truncated at L263 is shown. (b) The density has been cropped to show the channel interior. The approximate boundaries for the membrane bilayers (M), extracellular gap (E), and the cytoplasmic space (C) are indicated. The white arrows identify the locations of the cross sections that are parallel to the membrane bilayers (c). The red contours in (c) are at 1 A˚ above the mean density and include data to a resolution of 15 A˚ resolution. These contours define the boundary of the hemichannel and can be compared to previous low-resolution structural studies of liver [63,80] and heart [68] gap junctions. The yellow contours at 1.5 A˚ above the mean density include data to a resolution of 7.5 A˚. The roughly circular shape of these contours within the hydrophobic region of the bilayers is consistent with 24 transmembrane a-helices per hemichannel. The red asterisk in (b) marks the narrowest part of the channel where the aqueous pore is 15 A˚ in diameter, not accounting for the contribution of amino acid side chains, which are not resolved at the current limit of resolution. The noncrystallographic twofold symmetry that relates the two hemichannels of a gap junction channel has not been applied to the map. Hence, the similarity of the two hemichannels provides an independent measure on the quality of the reconstruction. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Unger et al. [136] with permission.)

that were typical, at the resolution of 7.5 A˚, for cross sections of a-helices. Therefore, each hemichannel contained 24 circular densities, consistent with the topological model in which each connexin subunit has four transmembrane a-helices. In contrast to previous cryoEM studies of 2D crystals derived from native plasma membranes [111,123,124], a recent cryoEM study [125] examined 2D crystals that were generated by reconstituting purified detergent-solubilized recombinant Cx26 into lipid bilayers (a M34A mutant). Surprisingly, the hemichannels appeared to re-dock during the reconstitution, thereby forming a triple bilayer structure with an up-down packing of adjacent

48

M. Yeager

Fig. 2.11. CryoEM structure of 2D crystals of reconstituted Cx26 hemichannels. The 3D map is contoured at 1 (light blue) and 2.4 (yellow) above the mean density. The inset in the upper left shows a 20 A˚ thick section perpendicular to the membrane plane through the density map of a hemichannel. This section corresponds to the region enclosed by the white lines shown in A. The arrowhead points to the large plug of density within the pore. The inner cytoplasmic protrusions (white arrows) extend from the cytoplasmic ends of helices B and C. (a–c) 30 A˚ thick slabs through the density map corresponding to the position of the lines shown in the Inset. The four a-helices are labeled A (cyan, A’), B (green, B’), C (yellow), and D (pink) as in the original Cx43 structure (Fig. 2.10) [136]. The arrowhead and white arrows represent the plug and the inner cytoplasmic protrusions, respectively. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Oshima et al. [125] with permission.)

dodecameric channels (Fig. 2.11). Adjacent, inverted hemichannels were packed tightly in the central bilayer, which enabled determination of a 3D map with an in-plane resolution of 9 A˚. The 3D map of Cx26M34A displayed a-helical packing that was very similar to the pattern in Cx43 (Fig. 2.8) [123,124]. Given the sequence conservation of M1 to M4 in connexin sequences and the similarity in structures of the 3D density maps of Cx43 and Cx26, it is likely that all gap junction channels will recapitulate this a-helical design (as depicted in Fig. 2.12).

2 Channel Structure Fig. 2.12. Molecular design of gap junction channels. (a) Top view showing the 30degree rotational stagger between docked hemichannels. Two subunits of the top hemichannel (in blue) are above one subunit of the bottom hemichannel (in red). The other subunits have been colored gray for clarity. The molecular boundary is depicted as a four-helix bundle, but there are other possibilities (Fig. 2.13). (b) Side view showing the top hemichannel in blue and the bottom one in red. Grayed areas denote parts of the structure that are most uncertain, especially the folding within the density at the boundary between the transmembrane assembly and the extracellular space. Putative b-sheets corresponding to E1 (on the perimeter of the extracellular gap) are drawn with thin lines to emphasize this ambiguity. The E2 loops are depicted as an interdigitating b-barrel [165]. Refer to Figs. 2 and 3 in Unger et al. [136] for the corresponding views of the 3D density map derived by cryoEM. The dimensions for the transmembrane ahelical domain and the extracellular gap are approximate. (A highresolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Kovacs et al. [186] with permission.)

49

50

M. Yeager

2.10 A Ca Molecular Model of the 24 a-Helices in the Gap Junction Hemichannel Given the availability of substantial mutagenesis, physiological and amino acid sequence data, and higher resolution cryoEM density maps, it was reasonable to attempt to build a working atomic model for the Ca positions of the amino acids in the 24 transmembrane a-helices within a hemichannel [124]. The key difficulties were that the map was of necessity a snapshot of a single structural state, and there was no guarantee that the structure corresponded to the dominant state probed by the mutagenesis/physiological studies. For these reasons, it is perhaps unrealistic to expect the two sets of data to be entirely reconciled; the functional state of the channels yielding the 3D map is uncharacterized, and substituted cysteine accessibility method (SCAM) studies used to identify pore-lining residues have their own potential ambiguities of interpretation (see below and Chapter 3). With these caveats in mind, an improved 3D cryoEM map (with in-plane resolution of 5.7 A˚ and vertical resolution of 19.8 A˚) was used as a basis for building the model [124]. The two most important new features of this model were proposals for the orientation and sequence identity of the transmembrane helices. For membrane proteins, evolutionarily conserved amino acids are more likely to mediate protein-packing interactions, and variable residues are more

Fig. 2.13. Possible molecular boundaries for the connexin subunit. Figure shows possible molecular boundaries for the connexin subunit including a helical bundle (left) and a checkmark (right) following the naming of Unger et al. [136]. Each shows two assignments for the four transmembrane a-helices within each subunit, according to [124] (top, blue) and [128] (bottom, green). Dashed lines denote the extracellular loops, E1 and E2, and the solid lines denote the M2 to M3 cytoplasmic loops. The a-helical rods designated A, B, C, and D in the 3D density map derived by electron cryocrystallography [136] are also indicated. Both assignments shown designate M3 as the major pore lining helix, but other studies suggest M1, as discussed in the text. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Kovacs et al. [186] with permission.)

2 Channel Structure

51

likely to face the lipid [126]. On the basis of the relative spatial locations of conserved and variable residues within the connexin family, as well as some of the SCAM data, the primary sequence of transmembrane segments M1 to M4 was assigned to the observed a-helices in the map (A is M2, B is M1, C is M3, D is M4) (Fig. 2.13, blue). This assignment predicted that M4 was the helix on the perimeter of the hemichannel. Support for this inference is suggested by experiments in which M4 of Cx43 was replaced with polyalanine without interfering with gap junctional communication [127]. The relative rotation angles of the ahelices fitted into the density map were estimated by analysis of evolutionary conservation and hydrophobicity of amino acid residues. Although this is the best defined model for the transmembrane domains of gap junctions as of this writing, the conformations of the amino acid side chains remain undetermined. In addition, the a-helical rods in the cryoEM density map display curvature that is not reflected in the idealized Ca model of Fleishman et al. [124]. With these provisos, the location of Cx26 and Cx32 mutations, respectively, causing human diseases such as nonsyndromic deafness (see Chapter 20) and Charcot-Marie-Tooth disease (see Chapter 15) could be mapped onto the Ca model

Fig. 2.14. A Ca model for the membrane spanning a-helices of a hemichannel. The Ca model (yellow ribbons) for the membrane spanning a-helices of a hemichannel was derived by combining the information from a computational analysis of connexin sequences, the results of a number of biochemical studies, and the constraints provided by a 3D cryoEM map (blue) [124]. While individually, none of these approaches provided high-resolution information, their sum yielded an atomic model that predicts how connexin mutations (red spheres) that result in diseases such as nonsyndromic deafness and Charcot-Marie-Tooth disease may interfere with formation of functional channels, by their disruption of helix–helix packing. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (Adapted from Fleishman et al. [124] with permission.)

52

M. Yeager

(Fig. 2.14). There was a surprising concentration of mutations at helix–helix interfaces, suggesting that disruption of helix packing interferes with channel function. The a-helical assignment of Cx32 in Fleishman et al. [124] differs from that deduced from Cx32 oocyte SCAM experiments [128] in which M1 and M2 were reversed (i.e., A is M1, B is M2) (Fig. 2.13, green), and from that suggested by the single-channel SCAM and domain swap studies (see below), which implicate M1 as pore-lining. As discussed above, these discrepancies might be attributed to methodological differences or to possible differences in conformation, such as the latter representing an open conformation and the former a closed conformation. Another possibility is the presence of conformational flexibility or ‘‘breathing’’ that would create transient solvent crevices between a-helices that would allow labeling of residues that do not line the pore.

2.11 Chemical Labeling Studies Suggest that the Amino Terminus, and the First or Third Transmembrane Domains Are Accessible to the Aqueous Pore Identification of the region of any channel protein that contributes directly to the lining of the pore is clearly essential to understanding the structural basis for its properties. In the case of connexin channels, M3 has been the de facto choice based on its amphipathic character when modeled as an a-helix (Fig. 2.7). Some patients with Charcot-Marie-Tooth disease have a point mutation in Cx32 at residue 26, from serine to the larger, hydrophobic residue leucine, in M1. Cx32S26L (Fig. 2.7b) was associated with a decrease in the permeability of the channel to large neutral polyethyleneglycol compounds without any change in single-channel conductance [129]. These results are consistent with a role for M1 in lining the channel, but as with many mutations in the absence of structure, it remains possible that effects could be exerted through conformational change from a distance. Starting with the experiments of Zhou et al. [130], a number of groups have used the SCAM method to probe the pores of several channels [131,132,133,134]. Skerrett et al. [128] continued these SCAM experiments for Cx32 using the paired oocyte system. Kronengold et al. [135] studied the porelining residues in Cx46 hemichannels by SCAM and patch clamp electrophysiology (see Chapter 3). In general, the SCAM results lead to a model in which multiple a-helices contribute to the wall of the pore, in particular M1, M3, and possibly M2. Notably, the 3D map at 7.5 A˚ resolution (Fig. 2.10) [136] and at 5.7 A˚ resolution (Fig. 2.14) [124], as well as the recent structure of Cx26 at 9 A˚ resolution (Fig. 2.11) [125], show that two a-helices contributed by each subunit line the pore of the channel. Domain swap studies have also shown that the charge selectivity of connexin pores can be controlled by E1 (Cx46, Cx32; [137]), suggesting that E1 contributes to the pore wall. Point mutations in the NT of Cx32 produced changes in

2 Channel Structure

53

the single-channel current-voltage relations consistent with electrostatic effects on the permeating ions [138]. Mutations at two positions in the NT of Cx40 showed that they were essential for spermine block of these channels, the block thought to be at the cytoplasmic vestibule [139]. Taken together, these data suggest that the NT, the second half of M1, and at least the initial part of E1 are directly involved in defining the conductance properties of connexin pores. As noted above, there are two phylogenetic groups of connexins, with Cx26 and Cx32, members of the b-group and Cx43 a member of the a-group [52,60] (see Chapter 1). While there must be some structural differences to account for different limiting pore diameters and charge selectivities, it would be truly remarkable if the fundamental organization and packing of the transmembrane helices were different. More to the point, the transmembrane densities derived from cryoEM of Cx26 [125] are virtually identical to those derived from cryoEM of Cx43 [136]. By the same token, there must be some differences in the pore-lining structures between unpaired hemichannels and hemichannels in junctional channels, simply by virtue of the docking interactions at the extracellular end of the hemichannel. Again, it would be remarkable if this resulted in wholesale differences in transmembrane packing. In fact, a host of data from measurements of unitary conductances, voltage sensitivities, pharmacological sensitivities, and other functional properties of single hemichannels and junctional channels suggest that this does not occur [135,140] (see Chapter 3). Since the differences cannot be readily explained by the considerations above, they may arise from some combination of the differences in the thiol-reactive reagents used, the different physical configurations of the experiments, and the relative reliabilities, sources of artifact, and constraints inherent in the two experimental protocols. Simply put, these different experiments may be revealing different kinds of information about the channels.

2.12 Heterotypic and Heteromeric Interactions Between Hemichannels and Connexins, Respectively If multiple connexins can be coexpressed in a single cell type, to what extent do different connexins interact within a gap junction plaque? Interactions between connexins can be divided into four types (Fig. 2.15): homomeric hemichannels are composed of a single connexin isoform; heteromeric hemichannels are composed by at least two different isoforms; homotypic junctional channels are formed by twelve identical connexin subunits; heterotypic channels are formed by two hemichannels that are each homomeric for different isoforms [53]. Evidence for heterotypic channels has been largely based on the electrophysiological analysis of paired Xenopus oocytes [104,105,141,142,143,144,145] or transfected cell lines [106,145,146] expressing different connexins. Through the efforts of several labs, a mapping of allowed and disallowed heterotypic

54

M. Yeager

Fig. 2.15. Schematic models for the possible arrangements of hemichannels to form gap junction channels. Hemichannels, which consist of six connexin subunits (red and blue), may be homomeric (composed of one connexin isoform) or heteromeric (composed of more than one connexin isoform). Hemichannels associate end-to-end to form a double-membrane gap junction intercellular channel. The channel may be homotypic (if hemichannels are identical) or heterotypic (if the two hemichannels are different). (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Kumar and Gilula [53] with permission.)

interactions within the connexin family has been established (Table 2.1) [61]. Overall, these results reveal that some connexins are highly selective in their communication patterns (e.g., Cx31 and Cx40), while others are rather promiscuous in heterotypic coupling (e.g., Cx45). Most connexins, however, show an intermediate behavior, typically interacting with more closely related connexins, e.g., Cx32, Cx26, Cx37, and Cx43. Thus, it appears that connexins within a homology class (i.e., a-group and b-group, or groups II and I, respectively) couple more frequently than those between classes. The two notable exceptions to this rule appear to be Cx30.3 (a b-group connexin that preferentially pairs with a-group connexins), and Cx50 (an a-group connexin that frequently pairs with b-group connexins). Analysis of intact and split liver gap junctions by scanning transmission electron microscopy revealed mass distributions consistent with heterotypic Cx

26

50 46 45 43 40 37 33 32 31.1 31 30.3 30 26

+ + – – – – + – – – + +

30

Table 2.1 Heterotypic coupling of connexins 30.3 31 31.1 32 33 37 40 43

– + + +

– – – –

– – – –

+

– –

–

– –

– +

+

–

+

+ + – – – – – +

– – –

+ + + +

– – + – +

– + + +

45

+

46

50

+ +

+

2 Channel Structure

55

and homotypic, but not heteromeric junctional channels [107]. Liver gap junction samples were mixed with a standard such as tobacco mosaic virus, and the specimens freeze-dried. The tobacco mosaic virus served as an internal mass standard to calibrate the mass measurements of hemichannels within a gap junction. Histograms for the mass measurements of liver gap junctions yielded three types of distributions: narrow and unimodal, and broad and unimodal or bimodal. The narrow unimodal distributions and the bimodal distributions were consistent with homotypic channels formed by Cx32-Cx32 or Cx26Cx26 pairings. Broad unimodal distributions could arise from heterotypic channels, homomeric channels, or pairing of heteromeric hemichannels. This ambiguity was resolved by examining the mass distributions for urea–split liver gap junction plaques. Since the average masses for the split junctions were always half of the mass for the intact junction, the hemichannels were homomeric. Hence, the broad unimodal distributions were consistent with heterotypic channels formed by the pairing of homomeric Cx26 hemichannels and homomeric Cx32 hemichannels. Measurements of electrical coupling between pairs of hepatocytes have also, in rare cases, revealed rectifying behavior [147] that is characteristic of heterotypic channels formed by Cx32 and Cx26 in paired oocytes [105]. Immunoelectron microscopy has demonstrated the existence of asymmetrically labeled gap junctions, for example between astrocytes, which express Cx43 [148,149,150] and Cx30 [151], and oligodendrocytes, which express Cx32 [152] and Cx45 [153,154]. The issue of heteromeric interactions (i.e., between different connexin subunits within a hemichannel) has been more difficult to assess since the properties of hemichannels have not been extensively examined and any studies on intact channels or junctions are complicated by potential heterotypic interactions (i.e., between hemichannels). Hepatic gap junctions are the best characterized case of connexins that are colocalized in situ. Immunoelectron microscopy demonstrated that Cx32 and Cx26 are intermingled, with no evidence of segregated domains [155,156]. However, limits on labeling densities using relatively large antibody probes did not allow resolution of the oligomeric structure of individual channels. Alternatively, the ability to examine the composition of detergentsolubilized and purified hemichannels has provided a strategy to examine heteromeric channels. Some protocols (e.g., 1% w/v Triton X-100 at 48C) isolate hemichannels by selecting for nonassembled hemichannels, rather than solubilizing hemichannels from assembled plaques [29]. Other detergent treatments are thought to disrupt the interactions between the two hemichannels within the gap junction, but not those between the subunits of the hemichannels [108]. Both approaches have been used to isolate hemichannels from animal tissues [109,157], exogenous expression systems [108,157], and even cell-free systems [158]. Immunoprecipitation or Western immunoblots of a 9S (Svedburg) species (compatible with connexin hexamers) suggested that heteromeric hemichannels can assemble between the

56

M. Yeager

two b-group connexins, Cx32 and Cx26, or the two a-group connexins, Cx46 and Cx50. On the basis of limited comparison, heteromeric interactions between a-group and b-group connexins (i.e., Cx43 and Cx32) appear to be disfavored [158]. Without in situ cross-linking experiments, it has been difficult to rule out the possibility of subunit exchange between oligomers, although this would seem to be unlikely. In addition, there is no evidence of more than two isoforms within a hexamer. The chemical properties that dictate heteromeric interactions between connexins are largely unexplored. In vitro model systems have been utilized to assess the possibility of heteromeric interactions. In one study, isolated connexins, which had been reconstituted into liposome membranes, suggested that heteromeric Cx32/Cx26 hemichannels might have different permeability properties from the homomeric forms [159]. In another approach, several connexins were coexpressed in transfected cell lines prior to pairing and measurements of coupling. When cells that expressed rat Cx43 and human CX37 were paired with each other, the single intercellular channel properties were more complex than could be accounted for by homomeric and heterotypic combinations, alone, and were most readily interpretable as indicating the presence of heteromeric channels [160].

2.13 Models for Hemichannel Docking 2.13.1 Conserved Cysteine Residues in the Extracellular Loops are Required for Docking The highly conserved extracellular loops, E1 and E2, presumably mediate hemichannel docking and thereby dictate the selectivity of heterotypic connexin interactions. The high degree of structural conservation is reflected by the invariant distribution of three cysteines in each loop (Fig. 2.7). Since mutation of any of these cysteines abrogates functional expression in the oocyte system, they are all essential for normal folding or channel function [161,162]. The Xenopus oocyte system has also been used to show that the assembly of functional gap junctions is enhanced under conditions that promote an exchange of disulfides [161], raising the intriguing possibility that different covalent linkage patterns may distinguish the hemichannel from the mature channel of the plaque. Of great value in building models for the interactions between these extracellular loops has been the elegant demonstration, made independently by two groups [163,164], that the disulfide bridges are exclusively intramolecular, with at least one linking E1 and E2 of a single connexin. Nevertheless, there are clear sequence differences between connexins that are critical. However, in the absence of an atomic resolution model for the conformation of, and interactions between, these loops, it is very difficult to discern which specific residues these may be.

2 Channel Structure

57

2.13.2 b-Sheet Structure in the Extracellular Loops In the oxidizing extracellular environment, it is likely that all three possible disulfides will form. A mutagenesis strategy has provided an initial mapping of the disulfide bridges [165]. Rather than deleting the cysteines, the first and third cysteine of each loop were moved within the sequence, both individually and in pairwise combinations, to identify the likely partners involved in disulfide bonds. Only certain paired movements of the cysteine residues were compatible with the assembly of functional channels. The pattern of pairings indicated that the first cysteine of each loop paired with the third cysteine of the opposite loop, suggesting that all three disulfide bonds form between the loops, and that E1 and E2 are arranged in an antiparallel configuration. Movements two or four residues away from the wild-type position were more effective than one or three in the restoration of functional channels. This periodicity suggested a b-sheet conformation for E1 and E2 that may be stabilized in an antiparallel configuration by three disulfides bridges. Testing a complete set of combinations of the six extracellular cysteine residues should lead to a complete mapping of the disulfide connections within the docking loops of connexins. Nevertheless, Foote et al. [165] have proposed that the interdigitation of protruding extracellular loops could form 24 stranded inner and outer antiparallel b-barrels. Such a model is similar to bacterial porins [98,166], but the barrels would be formed from b-strands contributed by different subunits. In contrast, each monomer within a porin trimer folds as a b-barrel. Direct structural information is clearly needed. Although the precise secondary structure within the gap was not revealed in the higher resolution 3D map [136], it was nevertheless clear that the interior band of density provided a continuous wall of protein (Fig. 2.10b), which functions as a tight electrical and chemical seal to exclude the exchange of substances with the extracellular milieu. Data are sparse as to the topological location of E1 and E2 in the extracellular gap. A careful electrophysiological and SCAM study of Cx46 hemichannels suggested that at least a portion of E1 is accessible for labeling [135]. If M3 is the major pore-lining helix, then one would expect that at least the aminoterminal portion of E2 is accessible to the pore. The cryoEM density map displays a continuous ring of density centered at a diameter of 34 A˚ and six arcs of density at a diameter of 52 A˚. These features suggest that the polypeptide in the gap is highly ordered and that the apposed connexins have extensive contact area, which is needed to form a tight seal separating the channel from the extracellular environment. However, these features do not allow confident assignment of the secondary structure of the protein within the extracellular gap. On the basis of the pattern of disulfide bonds between E1 and E2, Foote et al. [165] proposed a double b-barrel model for the extracellular region of the channel. Each b-barrel would be formed by a dodecameric, antiparallel interdigitation of b-turn-b projections from each hemichannel. To assess the plausibility of this proposition, a double b-barrel model (of which Fig. 2.12

58

M. Yeager

shows the inner layer) was built using the diameter of 34 A˚ in the cryoEM density map [136]. Although this may appear to be reasonable, the distance between loops belonging to adjacent connexins is shorter than ideal for a bbarrel, and as a consequence, only some of the interloop hydrogen bonds would be geometrically favored in this computational model. Also, note that this was also the case even for a virtually vertical orientation of the b-strands (Fig. 2.16b). Usually, b-strands in known b-barrel structures are more tilted with respect to the membrane plane [167], which would worsen the packing. It was estimated that a tilt of about 50 degrees would be needed in order for the full b-turn-b loops (about 35 aa) to pack within the accepted vertical gap space of 40 A˚ (Fig. 2.16b). It was therefore hypothesized that the extracellular loops near the surface of the lipid bilayer break from a b-conformation. Presumably this region of the loops (indicated with a shaded horizontal box in Fig. 2.12) has a more complicated conformation so as to fit the 40 A˚ gap. A plausible variation

Fig. 2.16. A statistical secondary-structure prediction with ICM software. (a) Sequences of loops E1 (top) and E2 (bottom). Tubes and arrows are predicted to be transmembrane helices and b-sheet secondary structure, respectively. Note that the prediction for E2 shows partial b-strand structure. The residue coloring comes from the alignment, and depends on both the conservation and the residue type as in Fig. 2.17. (b) Loops E1 (red) and E2 (green) belong to a single connexin subunit. The left side shows a side view, while the middle and right sides are pore views. The separation between the loops is 7 A˚. The right side shows that the 35 amino acids of the complete loops would have to be tilted by 50 degrees to be accommodated within the 40 A˚ extracellular gap (represented by horizontal lines), if one assumes that they form extended b-strands. The cysteines forming disulfide bonds are indicated by gold sticks. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Kovacs et al. [186] with permission.)

2 Channel Structure

59

could be that the b-sheets of each connexin do not lie with their planes perpendicular to the radius vectors, but are somewhat rotated, which would also help accommodate the inner barrel within the 34 A˚ diameter seen in the cryoEM density map. The outer arcs of density in the extracellular gap, observed at a diameter of 52 A˚, imply that the b-strands cannot be packed as a continuous b-barrel, since the same number of strands as in the inner barrel would span a larger diameter. Therefore, the continuous cylinder of density adjacent to the pore likely functions as the primary seal from the extracellular space. A statistical secondary structure prediction of the E1 and E2 extracellular loops was also performed using ICMsoftware (Fig. 2.16a) [168]. For E1, there was no preferred secondary structure; that is, the random coil was favored. For the E2 loop, the prediction yielded three segments of b-strand, which spanned about 40% of the length of the loop, consistent with the above model. By assuming a b-sheet secondary structure and carefully examining the sequence of the extracellular loops, particularly the positions of conserved prolines and cysteines, a potential spatial relationship between E1 and E2 can be defined. A stacked b-sheet model of E1 and E2 was built to see if the proposed disulfide bonding pattern of Foote et al. [165] could be constructed with correct covalent geometry and stereochemistry. This was actually the case (Fig. 2.16b). This model was built using the sequences of the Cx43 loops and placing the centrally located proline of each loop at the center of the b-hairpin. The cysteine residues align very well with good geometry, allowing the disulfide bonds to form correctly, even though the central one is shifted by one residue between E1 and E2. This model was then inserted into that of the whole channel shown in Fig. 2.12. However, the distance between the loops provided by this model, about 7 A˚, is less than that seen between the inner ring and outer arcs in the cryoEM map (9 to 10 A˚). A possible explanation, as suggested by others [165,169], could be that the 7 A˚ separation holds only for undocked hemichannels, but upon docking together, the disulfide bonds are exchanged to form intermolecular cross-links between the cysteines that reside in the middle of the b-strands, that is, not the b-hairpin cysteines. In order for these disulfides to form, the distance between the b-hairpins would have to be 30 A˚, as opposed to the 40 A˚ as suggested in Figure 2.16. The resulting gap of about 10 A˚ might presumably be occupied by the winding portion of the loops of the neighboring connexin.

2.13.3 Selectivity of Heterotypic Interactions May Be Guided by the Amino-Terminal Portion of the Second Extracellular Loop Several studies have attempted to determine which domains confer the specificity for heterotypic gap junction formation through the exchange of extracellular loops between connexins (Table 2.1). In one case, the E2 of Cx46, a

60

M. Yeager

connexin capable of heterotypic coupling with both Cx43 and Cx50, was sufficient to confer on Cx50 the ability to couple with Cx43, a normally forbidden combination [143]. While this points to a dominant role for E2 in defining heterotypic specificity, it was emphasized that in this particular combination, the E1 of Cx46 and Cx50 share 92% sequence identity. Nonetheless, further support for the dominant role of E2 was provided by a chimera of Cx32 (amino-terminal half, including E1) and Cx43 (carboxyl-terminal half, including E2), which showed Cx43-like interactions in pairing with Xenopus Cx38 [170] and Cx46, but not Cx50 [104]. Notwithstanding the consistent implication of E2 in defining the specificity of heterotypic interactions between connexins, recent evidence indicates that not only primary sequence, but also tertiary structure of these domains are critical determinants of docking specificity. Strikingly, an unpaired cysteine movement in E2 that disrupted the normal homotypic pairing of Cx32, induced a novel ability of Cx32 to pair with Xenopus Cx38 [165]. Since no significant change in primary sequence was involved, this change in specificity of heterotypic interactions apparently resulted from a distortion in the folding of E2. The same general conclusion can be deduced from the work of Haubrich et al. [171] employing a chimera of Cx43 and Cx40. In this case, the replacement of the extracellular domains of Cx40 with those of Cx43 only partially imparted Cx43 pairing characteristics. In some respects this chimera retained the properties of Cx40, presumably due to the constraints imposed by the transmembrane domains. These results underscore the limitations of the molecular biological approaches for defining functional domains in the absence of an atomic resolution structure. Most homotypic couplings between hemichannels result in functional channels, but many heterotypic channels do not. An extensive summary of these compatibilities is presented by Yeager and Nicholson [61]. In particular, it has been shown that Cx46 may be functionally paired with either Cx43 or Cx50, whereas Cx50 does not form functional channels with Cx43 [143]. These electrophysiological experiments in paired oocytes suggested that the E2 loop was a major determinant of isoform selectivity. This particular triplet (Cx46, Cx43, and Cx50) was analyzed in an attempt to elucidate a plausible set of rules dictating this selectivity behavior based on the E2 sequence difference between these isoforms (Fig. 2.17). Nonconserved residues are labeled and colored by polarity. Note that there is substantially more conservation in sequence in the carboxyl-terminal halves of the E2 loops, suggesting that selectivity is dictated by the amino-terminal regions, which are more variable in sequence. Notable differences were observed in certain key positions (arbitrarily numbered 8, 10, 16, 18, and 23), which could potentially participate in hydrogen bonds or salt bridges (indicated by small spheres) that are quite different between the isoforms. The only disallowed heterotypic pairing of this set, according to White et al. [143], is Cx43-Cx50. This incompatibility may result from the arginine residue at the eighth position in Cx50, conflicting with the nature of side chains near the b-hairpin (residues 16 to 23) of the paired hemichannel. For Cx43,

2 Channel Structure

61

Fig. 2.17. The E2 loop is a primary determinant of heterotypic compatibility. (a) Schematic key to show the compatibilities in pairing of heterotypic channels formed by Cx46, Cx43, and Cx50. (b–d) Comparison of the E2 loops of Cx46, Cx43, and Cx50, respectively. Only amino acids that differ in at least one of the three isoforms are shown in a stick representation. Cysteines are indicated with C and colored green on the ribbon. Residue coloring is as follows: yellow, hydrophobic; pink, polar; red, acidic; blue, basic. Note that the carboxyl-terminal halves of the loops are highly conserved, suggesting that selectivity is dictated by the aminoterminal regions of the loops. As discussed in the text, the hydrogen-bonding pattern near the b-hairpin of Cx43, along with its H22, may account for its incompatibility with Cx50. This analysis was based on the three localized regions that are predicted to be b-strands and did not depend on the complete b-turn-b fold depicted in the figure. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Kovacs et al. [186] with permission.)

these residues could form hydrogen bonds among themselves, leaving the side chain of H22 free, which together with the free side chain of R8 of Cx50, might obstruct the channel, or result in a steric clash. In contrast, when Cx50 is paired with Cx46, residues D16, N22, or T23 of the latter would be available to form hydrogen bonds with the side chain of R8 of Cx50, constraining its conformation so as to avoid an obstruction or clash. Homotypic Cx50 channels would therefore be allowed, for the same reason (N22 hydrogen-bonded to R8). By these criteria, the pairs Cx43-Cx46 (heterotypic), Cx46-Cx46 and Cx43-Cx43 (both homotypic) present no obstacle. Of course, this most likely represents an overly simplistic view based on a very limited sample size. Nevertheless, it appears that the potential packing of a simple secondary-structure model of E2 might begin to provide a working hypothesis for understanding the origin of pairing preferences amongst connexins.

62

M. Yeager

2.14 Gating Models Based on Electron Microscopy, Atomic Force Microscopy and Nuclear Magnetic Resonance Spectroscopy 2.14.1 Models for Calcium-Induced Closure By comparing changes in the 3D structure in the presence and absence of Ca2+, a model was proposed for the gating of liver gap junctions [79,172]. In the transition from the open to the closed configuration, the protein subunits decrease their angle of tilt by about 50 degrees tangential to the channel axis and slide with respect to one another along their long axis. Such a mechanism is appealing because it would involve sliding of adjacent protein surfaces to accomplish channel closure that would be more energetically favorable than dramatic conformational changes involving refolding of polypeptide chains. Although X-ray diffraction studies by Unwin and Ennis [173] provided evidence favoring the Ca2+-mediated structural transition, Caspar et al. [174] maintained that possible Ca2+-mediated changes in the X-ray diffraction patterns were masked by changes that resulted from variations in lattice constants and channel packing related to different preparations. A vertical section through the electron density map derived from X-ray patterns showed density on the cytoplasmic face not seen in the map obtained by electron microscopy [175]. This density in the X-ray map was termed a gating structure and was located outside the bilayer in the cytoplasmic domain of the hemichannel [174,175]. Gating was postulated to involve some kind of displacement of the gating structure. However, this change has not been delineated in separate specimens containing open and closed channels. A higher resolution structure analysis would presumably resolve the discrepancies between the maps provided by electron crystallography and X-ray diffraction. Atomic force microscopy (AFM) has provided insight into gap junction channel gating. In this method, an extremely fine needle is dragged along the surface of the specimen. Deflections of the needle on the order of angstroms can be detected by observing the reflection of laser light from the needle. An important advantage of AFM is that fully hydrated specimens can be examined in physiologic buffers. Although individual atoms in inorganic materials can be visualized by AFM, images of biological specimens are limited by the inherent deformability of biological specimens. The AFM images of the cytoplasmic surfaces of gap junctions have not displayed any molecular detail, presumably due to conformational flexibility of the protein. For instance, AFM studies showed that the CT of Cx43 is more distensible than E1 [176]. By the use of antibodies against the CT, the stretch length and energy required for stretching the CT supported a particle-receptor model for gating [177,178,179]. By increasing the applied force on the AFM tip, Hoh et al. [180] were able to dissect one membrane in the gap junction to reveal the six subunits in the hemichannels in the exposed extracellular surface. More recently, Mu¨ller et al. [181] visualized a Ca2+-induced decrease in the extracellular pore diameter of Cx26 hemichannels (Fig. 2.18). Recently, pH-induced closure of

2 Channel Structure

63

Fig. 2.18. Atomic force microscopy of the extracellular surface of Cx26 gap junction plaques show a decrease in the pore diameter in the presence of Ca2+. Applied force on the atomic force microscopy (AFM) tip was used to remove one membrane in the plaque so that the extracellular surface could be visualized. Images of the extracellular surfaces displayed in an idealized 2D lattice were recorded in the absence (a) and presence (b) of Ca2+. Correlation averages shown were calculated from several topographs recorded under equivalent conditions. The lateral resolution of the averages was limited to 15 A˚. To visualize their differences, a difference map was calculated (c). (d) Superimposed contours of the two conformations (red, no Ca2+; green, with Ca2+). All presentations show that the most significant change is observed at the pore. The contour plots, however, show that in the presence of Ca2+ the connexin surface protrusions move toward the pore center. Note that the greatest difference is in the size of the pore. The images exhibiting a vertical gray level range of 2 nm (a,b) and 0.8 nm (d) were displayed as relief tilted by 5 degrees. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Mu¨ller et al. [181] with permission.)

Cx26 channels was observed in aminosulfonate buffers such as 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES), but not in non-aminosulfonate buffers [182] (see Chapter 5).

2.14.2 Is the Amino-Terminal Domain of Connexin26 a Gating Particle? Nuclear magnetic resonance (NMR) spectroscopy of a 13 aa peptide corresponding to the NT of Cx26 displayed a two-turn a-helix, which then unraveled into a flexible loop-like structure [183]. It was hypothesized that this short NT

64

M. Yeager

helix is oriented parallel to the transmembrane helices lining the entrance to the pore, thus forming part of the conduction path and contributing to the voltage dependence of the channel. Support for this model is suggested by recent cryoEM studies of the M34A mutant of Cx26 (Fig. 2.11) [125]. A surprising and unexpected feature of the map was a plug of density in the cytoplasmic mouth of the pore. Several experimental conditions favored a closed conformation of Cx26, for example, use of the M34A mutant, low pH, aminosulfonate buffer, carbenoxolone, and high Ca2+ and Mg2+. The simplest interpretation is that the plug represents an aggregate of the NT, suggesting a simple mechanism for pore gating. The CT mediates low pH gating by interacting with the CL between M2 and M3 [178] (see Chapter 5). However, this mechanism has not been demonstrated for Cx26, and the CT of Cx26 does not contain the required segment. Confirmation of the chemical identity of the plug requires a bona fide high-resolution map so that the residues comprising the cytoplasmic plug can be identified. Even at an intermediate resolution such as 7 A˚, a difference 3D map between wildtype Cx26 and Cx26 without the NT would confirm whether or not the plug is formed by interacting NTs.

2.15 Conclusion In spite of detailed differences in the projection density maps of gap junctions isolated from different tissues, there are important conserved features. For instance, the hexameric hemichannel with a central channel is clearly a conserved motif in the design of gap junction membrane channels. The conduit for intercellular communication is formed by hemichannels in the register between closely apposed cells. Gap junction channels can be viewed as being modular in design. Homology in the transmembrane sequences suggests that the bundle of 24 a-helices that form each hemichannel will likely be a common architecture for the transmembrane channel. Similarly, sequence homology in the extracellular domains suggests similar mechanisms for hemichannel–hemichannel interactions. However, it is clear that differences in the extracellular loops appear to limit the promiscuity allowed in the pairing of heterotypic hemichannels. The CL between M2 and M3 and especially the CT are the most divergent regions between different connexins and presumably confer unique functional and regulatory properties for different connexins. However, these regions may be recalcitrant to structural determination, as a consequence of polypeptide flexibility that would lead to loss of resolution in any method utilizing averaging. The identification of specific functional sites or domains is critical for understanding the regulation of gap junction channels. Instead of blindly accumulating site mutations, a better strategy may be the generation of chimeric connexins where a loop domain of one connexin can be substituted for another.

2 Channel Structure

65

Unfortunately, with some notable exceptions of a series of swaps between Cx32 and Cx26, such chimeric connexins have proven distressingly uninstructive since they fail to form functional channels [170]. This may reflect a requirement for strict interactions of different connexins that are only satisfied when compatible swaps are made between connexins that interact in vivo. A cautionary note for all structural/functional studies on gap junctions is that by definition they are performed in exogenous systems, and at times the results may vary from that which might be seen in the native environment of the channel. Interpretation of the results will also remain limited as long as we lack an experimentally determined atomic resolution structure of the channel. The isolation of apparently intact hemichannels using appropriate detergents [184], and their expression in large quantities in the baculovirus system [184], raises hopes for the eventual generation of highly ordered 2D or 3D hemichannel crystals. To date, 2D crystals that are ordered to 6 A˚ resolution have been generated from a recombinant Cx43 with CT truncation [122,123,124,185]. CryoEM has for the first time confirmed the widely accepted model that the transmembrane domains of each connexin are folded as four a-helices. However, crystals that are ordered to better than 4 A˚ resolution would be required to visualize amino acid side chains. Currently, there are numerous areas for continued exploration of the structurefunction properties of gap junction channels: delineation of the precise interactions between different regions in the channel (e.g., the CT, CL, and NT), the precise folding of the extracellular loops that confers stability but also specificity in hemichannel–hemichannel pairing, the identification and orientation of the transmembrane a-helices in the 3D map, and specific mechanisms for gating. The regulation of gap junction channels is particularly fascinating since they manifest properties of both ligand-gated channels and voltage-gated channels. In addition, the lipid environment can likely regulate the gating state. One can imagine that gating involves multiple conformational switches such as movements of the CT during pH gating (see Chapter 5), movements of the NT during voltage-gating (see Chapter 4), or changes in a-helical packing resulting from lipophilic molecules that affect the a-helices on the perimeter of the channel. Such multiple triggers may converge on a final common conformational change that presumably involves a site in the bilayer near the extracellular gap since this is the narrowest portion of the channel (Fig. 2.8). The exploration of these mechanisms requires a strategy that integrates molecular biology, crystallography, and spectroscopic methods with functional studies using electrophysiology and transport assays. Acknowledgments The foundation for this chapter was based on some of our previous reviews [70,112,186,187], and I am especially grateful to Bruce Nicholson and Andrew Harris for their contributions. This work was supported by National Institutes of Health (NIH) grant RO1HL48908. We thank Julio Kovacs for preparation of Figures 2.12, 2.16, and 2.17; Kenton Baker for preparation of Figure 2.13; and Michael E. Pique for preparation of Figure 2.14.

66

M. Yeager

References 1. Weidmann S. The electrical constants of Purkinje fibres. J Physiol. 1952;118:348–60. 2. Barr L, Dewey MM, Berger W. Propagation of action potentials and the structure of the nexus in cardiac muscle. J Gen Physiol. 1965;48:797–823. 3. Weidmann S. The diffusion of radiopotassium across intercalated disks of mammalian cardiac muscle. J Physiol. 1966;187:323–42. 4. Robertson JD. The occurrence of a subunit pattern in the unit membranes of club endings in Mauthner cell synapses in goldfish brains. J Cell Biol. 1963;19:201–21. 5. Furshpan EJ, Potter DD. Transmission at the giant motor synapses of the crayfish. J Physiol. 1959;145:289–325. 6. Potter DD, Furshpan EJ, Lennox ES. Connections between cells of the developing squid as revealed by electrophysiological methods. Proc Natl Acad Sci USA. 1966;55:328–36. 7. Furshpan EJ, Potter DD. Low-resistance junctions between cells in embryos and tissue culture. Curr Top Dev Biol. 1968;3:95–127. 8. Subak-Sharpe H, Bu¨ rk RR, Pitts JD. Metabolic co-operation by cell to cell transfer between genetically different mammalian cells in tissue culture. Heredity. 1966;21:342–3. 9. Subak-Sharpe H, Bu¨rk RR, Pitts JD. Metabolic co-operation between biochemically marked mammalian cells in tissue culture. J Cell Sci. 1969;4:353–67. 10. Loewenstein WR, Kanno Y. Studies on an epithelial (gland) cell junction. J Cell Biol. 1964;22:565–86. 11. Karrer HE, Cox J. The striated musculature of blood vessels. J Biophys Biochem Cytol. 1960;8:135–50. 12. Dewey MM, Barr L. Intercellular connection between smooth muscle cells: the Nexus. Science. 1962;137:670–2. 13. Trelstad RL, Revel J-P, Hay ED. Tight junctions between cells in the early chick embryo as visualized with the electron microscope. J Cell Biol. 1966;31:C6–C10. 14. Revel JP, Karnovsky MJ. Hexagonal array of subunits in intercellular junctions of the mouse heart and liver. J Cell Biol. 1967;33:C7–C12. 15. Kreutziger GO. Freeze-etching of intercellular junctions of mouse liver. In: 26th Proceedings of the Electron Microscopy Society of America. Baton Rouge, Claitor’s, 1968:138. 16. Chalcroft JP, Bullivant S. An interpretation of liver cell membrane and junction structure based on observation of freeze-fracture replicas of both sides of the fracture. J Cell Biol. 1970;47:49–60. 17. Goodenough DA, Revel JP. A fine structural analysis of intercellular junctions in the mouse liver. J Cell Biol. 1970;45:272–90. 18. Hanna RB, Ornberg RL, Reese TS. Structural details of rapidly frozen gap junctions. In: Spray DC; editors. Gap Junctions. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1985. pp. 23–32. 19. Brightman MW, Reese TS. Junctions between intimately apposed cell membranes in the vertebrate brain. J Cell Biol. 1969;40:648–77. 20. Friend DS, Gilula NB. Variations in tight and gap junctions in mammalian tissues. J Cell Biol. 1972;53:758–76. 21. Gilula NB, Reeves OR, Steinbach A. Metabolic coupling, ionic coupling and cell contacts. Nature. 1972;235:262–5. 22. Loewenstein WR. On the genesis of cellular communication. Dev Biol. 1967;15:503–20. 23. Flower NE. Invertebrate gap junctions. J Cell Sci. 1977;25:163–71. 24. Peracchia C. Structural correlates of gap junction permeation. Int Rev Cytol. 1980;66:81–146. 25. Flagg-Newton J, Simpson I, Loewenstein WR. Permeability of the cell-to-cell membrane channels in mammalian cell junction. Science. 1979;205:404–7.

2 Channel Structure

67

26. Schwarzmann G, Wiegandt H, Rose B, Zimmerman A, Ben-Haim D, Loewenstein WR. Diameter of the cell-to-cell junctional membrane channels as probed with neutral molecules. Science. 1981;213:551–3. 27. Revel JP, Yee AG, Hudspeth AJ. Gap junctions between electrotonically coupled cells in tissue culture and in brown fat. Proc Natl Acad Sci USA. 1971;68:2924–7. 28. Musil LS, Cunningham BA, edelman GM, Goodenough DA. Differential phosphorylation of the gap junction protein connexin43 in junctional communication-competent and -deficient cell lines. J Cell Biol. 1990;111:2077–88. 29. Musil LS, Goodenough DA. Multisubunit assembly of an integral plasma membrane channel protein, gap junction connexin43, occurs after exit from the ER. Cell. 1993;74:1065–77. 30. Li H, Liu TF, Lazrak A, Peracchia C, Goldberg GS, Lampe PD, Johnson RG. Properties and regulation of gap junctional hemichannels in the plasma membranes of cultured cells. J Cell Biol. 1996;134:1019–30. 31. DeVries SH, Schwartz EA. Hemi-gap-junction channels in solitary horizontal cells of the catfish retina. J Physiol. 1992;445:201–30. 32. Malchow RP, Qian H, Ripps H. Evidence for hemi-gap junctional channels in isolated horizontal cells of the skate retina. J Neurosci Res. 1993;35:237–45. 33. Goodenough DA, Stoeckenius W. The isolation of mouse hepatocyte gap junctions. Preliminary chemical characterization and X-ray diffraction. J Cell Biol. 1972;54:646–56. 34. Evans WH, Gurd JW. Preparation and properties of nexuses and lipid-enriched vesicles from mouse liver plasma membranes. Biochem J. 1972;128:691–700. 35. Culvenor JG, Evans WH. Preparation of hepatic gap (communicating) junctions. Identification of the constituent polypeptide subunits. Biochem J. 1977;168:475–81. 36. Hertzberg EL. Isolation and characterization of liver gap junctions. Methods Enzymol. 1983;98:501–10. 37. Nicholson BJ, Revel J-P. Gap junctions in liver: isolation, morphological analysis, and quantitation. Methods Enzymol. 1983;98:519–37. 38. Hertzberg EL, Gilula NB. Isolation and characterization of gap junctions from rat liver. J Biol Chem. 1979;254:2138–47. 39. Goodenough DA. Bulk isolation of mouse hepatocyte gap junctions. Characterization of the principal protein, connexin. J Cell Biol. 1974;61:557–63. 40. Gros DB, Nicholson BJ, Revel J-P. Comparative analysis of the gap junction protein from rat heart and liver: is there a tissue specificity of gap junctions? Cell. 1983;35:539–49. 41. Nicholson BJ, Gros DB, Kent SBH, Hood LE, Revel J-P. The Mr 28,000 gap junction proteins from rat heart and liver are different but related. J Biol Chem. 1985;260:6514–7. 42. Manjunath CK, Goings GE, Page E. Proteolysis of cardiac gap junctions during their isolation from rat hearts. J Membr Biol. 1985;85:159–68. 43. Kumar NM, Gilula NB. Cloning and characterization of human and rat liver cDNAs coding for a gap junction protein. J Cell Biol. 1986;103:767–76. 44. Paul DL. Molecular cloning of cDNA for rat liver gap junction protein. J Cell Biol. 1986;103:123–34. 45. Zhang JT, Nicholson BJ. Sequence and tissue distribution of a second protein of hepatic gap junctions, Cx26, as deduced from its cDNA. J Cell Biol. 1989;109:3391–401. 46. Bruzzone R, White TW, Paul DL. Connections with connexins: the molecular basis of direct intercellular signaling. Eur J Biochem. 1996;238:1–27. 47. Gerido DA, White TW. Connexin disorders of the ear, skin, and lens. Biochim Biophys Acta. 2004;1662:159–70. 48. Phelan P, Stebbings LA, Baines RA, Bacon JP, Davies JA, Ford C. Drosophila ShakingB protein forms gap junctions in paired Xenopus oocytes. Nature. 1998;391:181–4. 49. White TW, Wang H, Mui R, Litteral J, Brink PR. Cloning and functional expression of invertebrate connexins from Halocynthia pyriformis. FEBS Lett. 2004;577:42–8. 50. Beyer EC. Molecular cloning and developmental expression of two chick embryo gap junction proteins. J Biol Chem. 1990;265:14439–43.

68

M. Yeager

51. Willecke K, Hennemann H, Dahl E, Jungbluth S, Heynkes R. The diversity of connexin genes encoding gap junctional proteins. Eur J Cell Biol. 1991;56:1–7. 52. Kumar NM, Gilula NB. Molecular biology and genetics of gap junction channels. Semin Cell Biol. 1992;3:3–16. 53. Kumar NM, Gilula NB. The gap junction communication channel. Cell. 1996;84:381–8. 54. Goodenough DA, Goliger JA, Paul DL. Connexins, connexons, and intercellular communication. Annu Rev Biochem. 1996;65:475–502. 55. Nicholson SM, Bruzzone R. Gap junctions: getting the message through. Curr Biol. 1997;7:R340–R4. 56. Beyer EC, Paul DL, Goodenough DA. Connexin43: a protein from rat heart homologous to a gap junction protein from liver. J Cell Biol. 1987;105:2621–9. 57. White TW, Bruzzone R, Goodenough DA, Paul DL. Mouse Cx50, a functional member of the connexin family of gap junction proteins, is the lens fiber protein MP70. Mol Biol Cell. 1992;3:711–20. 58. Gimlich RL, Kumar NM, Gilula NB. Differential regulation of the levels of three gap junction mRNAs in Xenopus embryos. J Cell Biol. 1990;110:597–605. 59. Bennett MVL, Barrio LC, Bargiello TA, Spray DC, Hertzberg E, Sa´ez JC. Gap junctions: new tools, new answers, new questions. Neuron. 1991;6:305–20. 60. Bennett MVL, Zheng X, Sogin ML. The connexins and their family tree. Soc Gen Physiol Ser. 1994;49:223–33. 61. Yeager M, Nicholson B. Structure and biochemistry of gap junctions. In: Hertzberg E, editor. Gap junctions, advances in molecular and cell biology. Stamford, CT: JAI Press; 2000. pp. 31–98. 62. Hoh JH, John SA, Revel J-P. Molecular cloning and characterization of a new member of the gap junction gene family, connexin-31. J Biol Chem. 1991;266:6524–31. 63. Makowski L, Caspar DLD, Phillips WC, Goodenough DA. Gap junction structures. II. Analysis of the X-ray diffraction data. J Cell Biol. 1977;74:629–45. 64. Sayre D. Some implications of a theorem due to Shannon. Acta Cryst. 1952;5:843. 65. Makowski L, Caspar DL, Phillips WC, Goodenough DA. Gap junction structures. V. Structural chemistry inferred from X-ray diffraction measurements on sucrose accessibility and trypsin susceptibility. J Mol Biol. 1984;174:449–81. 66. Caspar DLD, Goodenough DA, Makowski L, Phillips WC. Gap junction structures. I. Correlated electron microscopy and X-ray diffraction. J Cell Biol. 1977;74:605–28. 67. Zampighi G, Unwin PNT. Two forms of isolated gap junctions. J Mol Biol. 1979;135:451–64. 68. Yeager M, Gilula NB. Membrane topology and quaternary structure of cardiac gap junction ion channels. J Mol Biol. 1992;223:929–48. 69. Yeager M. In situ two-dimensional crystallization of a polytopic membrane protein: the cardiac gap junction channel. Acta Cryst. 1994;D50:632–8. 70. Yeager M. Structure of cardiac gap junction intercellular channels. J Struct Biol. 1998;121:231–45. 71. Lampe PD, Kistler J, Hefti A, Bond J, Mu¨ller S, Johnson RG, Engel A. In vitro assembly of gap junctions. J Struct Biol. 1991;107:281–90. 72. Perkins G, Goodenough D, Sosinsky G. Three-dimensional structure of the gap junction connexon. Biophys J. 1997;72:533–44. 73. Perkins GA, Goodenough DA, Sosinsky GE. Formation of the gap junction intercellular channel requires a 308 rotation for interdigitating two apposing connexons. J Mol Biol. 1998;277:171–7. 74. Sikerwar SS, Downing KH, Glaeser RM. Three-dimensional structure of an invertebrate intercellular communicating junction. J Struct Biol. 1991;106:255–63. 75. Yeager M. Electron microscopic image analysis of cardiac gap junction membrane crystals. Microsc Res Tech. 1995;31:452–66.

2 Channel Structure

69

76. Toyoshima C, Unwin N. Ion channel of acetylcholine receptor reconstructed from images of postsynaptic membranes. Nature. 1988;336:247–50. 77. Henderson R, Unwin PNT. Three-dimensional model of purple membrane obtained by electron microscopy. Nature. 1975;257:28–32. 78. Amos LA, Henderson R, Unwin PNT. Three-dimensional structure determination by electron microscopy of two-dimensional crystals. Prog Biophys Mol Biol. 1982;39:183–231. 79. Unwin PNT, Zampighi G. Structure of the junction between communicating cells. Nature. 1980;283:545–9. 80. Unwin PN, Ennis PD. Two configurations of a channel-forming membrane protein. Nature. 1984;307:609–13. 81. Yeager M, Unger VM, Falk MM. Synthesis, assembly and structure of gap junction intercellular channels. Curr Opin Struct Biol. 1998;8:517–24. 82. Nicholson BJ, Hunkapiller MW, Grim LB, Hood LE, Revel J-P. Rat liver gap junction protein: properties and partial sequence. Proc Natl Acad Sci USA. 1981;78:7594–8. 83. Manjunath CK, Nicholson BJ, Teplow D, Hood L, Page E, Revel J-P. The cardiac gap junction protein (Mr 47,000) has a tissue-specific cytoplasmic domain of Mr 17,000 at its carboxy-terminus. Biochem Biophys Res Commun. 1987;142:228–34. 84. Manjunath CK, Goings GE, Page E. Detergent sensitivity and splitting of isolated liver gap junctions. J Membr Biol. 1984;78:147–55. 85. Zimmer DB, Green CR, Evans WH, Gilula NB. Topological analysis of the major protein in isolated intact rat liver gap junctions and gap junction-derived single membrane structures. J Biol Chem. 1987;262:7751–63. 86. Hertzberg EL, Disher RM, Tiller AA, Zhou Y, Cook RG. Topology of the Mr 27,000 liver gap junction protein. Cytoplasmic localization of amino- and carboxyl termini and a hydrophilic domain which is protease-hypersensitive. J Biol Chem. 1988;263:19105–11. 87. Zhang J-T, Nicholson BJ. The topological structure of connexin 26 and its distribution compared to connexin 32 in hepatic gap junctions. J Membr Biol. 1994;139:15–29. 88. Milks LC, Kumar NM, Houghten R, Unwin N, Gilula NB. Topology of the 32-kd liver gap junction protein determined by site-directed antibody localizations. EMBO J. 1988;7:2967–75. 89. Tibbitts TT, Caspar DLD, Phillips WC, Goodenough DA. Diffraction diagnosis of protein folding in gap junction connexons. Biophys J. 1990;57:1025–36. 90. Cascio M, Gogol E, Wallace BA. The secondary structure of gap junctions. Influence of isolation methods and proteolysis. J Biol Chem. 1990;265:2358–64. 91. Yeager M. Structure and design of cardiac gap-junction membrane channels. In: Hall JE, Zampighi GA, Davis RM, editors. Progress in cell research, vol 3: Gap junctions. Amsterdam: Elsevier Science Publishers; 1993. pp. 47–55. 92. Johnson WC, Jr. Secondary structure of proteins through circular dichroism spectroscopy. Annu Rev Biophys Biophys Chem. 1988;17:145–66. 93. Fasman GD. Distinguishing transmembrane helices from peripheral helices by circular dichroism. Biotechnol Appl Biochem. 1993;18:111–38. 94. Bloemendal M, Johnson WC, Jr. Structural information on proteins from circular dichroism spectroscopy possibilities and limitations. Pharm Biotechnol. 1995;7:65–100. 95. Woody RW. Circular dichroism. Methods Enzymol. 1995;246:34–71. 96. Glaeser RM, Jap BK. Absorption flattening in the circular dichroism spectra of small membrane fragments. Biochemistry. 1985;24:6398–401. 97. Henderson R, Baldwin JM, Ceska TA, Zemlin F, Beckmann E, Downing KH. Model for the structure of bacteriorhodopsin based on high-resolution electron cryo-microscopy. J Mol Biol. 1990;213:899–929. 98. Jap BK, Walian PJ, Gehring K. Structural architecture of an outer membrane channel as determined by electron crystallography. Nature. 1991;350:167–70.

70

M. Yeager

99. Ku¨hlbrandt W, Wang DN, Fujiyoshi Y. Atomic model of plant light-harvesting complex by electron crystallography. Nature. 1994;367:614–21. 100. Cheng A, van Hoek AN, Yeager M, Verkman AS, Mitra AK. Three-dimensional organization of a human water channel. Nature. 1997;387:627–30. 101. Walz T, Hirai T, Murata K, Heymann JB, Mitsuoka K, Fujiyoshi Y, Smith BL, Agre P, Engel A. The three-dimensional structure of aquaporin-1. Nature. 1997;387:624–7. 102. Li H, Lee S, Jap BK. Molecular design of aquaporin-1 water channel as revealed by electron crystallography. Nat Struct Biol. 1997;4:263–5. 103. Gogol E, Unwin N. Organization of connexons in isolated rat liver gap junctions. Biophys J. 1988;54:105–12. 104. White TW, Paul DL, Goodenough DA, Bruzzone R. Functional analysis of selective interactions among rodent connexins. Mol Biol Cell. 1995;6:459–70. 105. Barrio LC, Suchyna T, Bargiello T, Xu LX, Roginski RS, Bennett MV, Nicholson BJ. Gap junctions formed by connexins 26 and 32 alone and in combination are differently affected by applied voltage. Proc Natl Acad Sci USA. 1991;88:8410–4. 106. Elfgang C, Eckert R, Lichtenberg-Frate´ H, Butterweck A, Traub O, Klein RA, Hu¨ lser DF, Willecke K. Permeability and selective formation of gap junction channels in connexin-transfected HeLa cells. J Cell Biol. 1995;129:805–17. 107. Sosinsky G. Mixing of connexins in gap junction membrane channels. Proc Natl Acad Sci USA. 1995;92:9210–4. 108. Stauffer KA. The gap junction proteins b1-connexin (connexin-32) and b2-connexin (connexin-26) can form heteromeric hemichannels. J Biol Chem. 1995;270:6768–72. 109. Jiang JX, Goodenough DA. Heteromeric connexons in lens gap junction channels. Proc Natl Acad Sci USA. 1996;93:1287–91. 110. Kumar NM, Friend DS, Gilula NB. Synthesis and assembly of human b1 gap junctions in BHK cells by DNA transfection with the human b1 cDNA. J Cell Sci. 1995;108:3725–34. 111. Unger VM, Kumar NM, Gilula NB, Yeager M. Projection structure of a gap junction membrane channel at 7 A˚ resolution. Nat Struct Biol. 1997;4:39–43. 112. Yeager M, Nicholson BJ. Structure of gap junction intercellular channels. Curr Opin Struct Biol. 1996;6:183–92. 113. Unwin PN, Henderson, R. Molecular structure determination by electron microscopy of unstained crystalline specimens. J Mol Biol. 1975;94:425–40. 114. Ku¨hlbrandt W, Downing KH. Two-dimensional structure of plant light-harvesting complex at 3.7 A˚ resolution by electron crystallography. J Mol Biol. 1989;207:823–8. 115. Havelka WA, Henderson R, Heymann JAW, Oesterhelt D. Projection structure of halorhodopsin from Halobacterium halobium at 6 A˚ resolution obtained by electron cryo-microscopy. J Mol Biol. 1993;234:837–46. 116. Schertler GFX, Villa C, Henderson R. Projection structure of rhodopsin. Nature. 1993;362:770–2. 117. Jap BK, Li H. Structure of the osmo-regulated H2O-channel, AQP-CHIP, in projection at 3.5 A˚ resolution. J Mol Biol. 1995;251:413–20. 118. Karrasch S, Bullough PA, Ghosh R. The 8.5 A˚ projection map of the light-harvesting complex I from Rhodospirillum rubrum reveals a ring composed of 16 subunits. EMBO J. 1995;14:631–8. 119. Mitra AK, Van Hoek AN, Wiener MC, Verkman AS, Yeager M. The CHIP28 water channel visualized in ice by electron crystallography. Nat Struct Biol. 1995;2:726–9. 120. Walz T, Typke D, Smith BL, Agre P, Engel A. Projection map of aquaporin-1 determined by electron crystallography. Nat Struct Biol. 1995;2:730–2. 121. Hebert H, Schmidt-Krey I, Morgenstern R, Murata K, Hirai T, Mitsuoka K, Fujiyoshi Y. The 3.0 A˚ projection structure of microsomal glutathione transferase as determined by electron crystallography of p 21212 two-dimensional crystals. J Mol Biol. 1997;271:751–8.

2 Channel Structure

71

122. Unger VM, Kumar NM, Gilula NB, Yeager M. Expression, two-dimensional crystallization, and electron cryo-crystallography of recombinant gap junction membrane channels. J Struct Biol. 1999;128:98–105. 123. Unger V, Kumar N, Gilula N, Yeager M. Electron Cryo-Crystallography of a Recombinant Gap Junction Channel at 7 A˚ Resolution. In: Gilula NB, editor. Gap junctionmediated intercellular signalling in health and disease. New York: John Wiley and Sons, 1999. 124. Fleishman SJ, Unger VM, Yeager M, Ben-Tal N. A Ca model for the transmembrane alpha helices of gap junction intercellular channels. Mol Cell. 2004;15:879–88. 125. Oshima A, Tani K, Hiroaki Y, Fujiyoshi Y, Sosinsky GE. Three-dimensional structure of a human connexin26 gap junction channel reveals a plug in the vestibule. Proc Natl Acad Sci USA. 2007;104:10034–9. 126. Baldwin JM. The probable arrangement of the helices in G protein-coupled receptors. EMBO J. 1993;12:1693–703. 127. Bao X, Chen Y, Lee SH, Lee SC, Reuss L, Altenberg GA. Membrane transport proteins with complete replacement of transmembrane helices with polyalanine sequences remain functional. J Biol Chem. 2005;280:8647–50. 128. Skerrett IM, Aronowitz J, Shin JH, Cymes G, Kasperek E, Cao FL, Nicholson BJ. Identification of amino acid residues lining the pore of a gap junction channel. J Cell Biol. 2002;159:349–60. 129. Oh S, Ri Y, Bennett MVLL, Trexler EB, Verselis VK, Bargiello TA. Changes in permeability caused by connexin 32 mutations underlie X-linked Charcot-MarieTooth disease. Neuron. 1997;19:927–38. 130. Zhou XW, Pfahnl A, Werner R, Hudder A, Llanes A, Luebke A, Dahl G. Identification of a pore lining segment in gap junction hemichannels. Biophys J. 1997;72:1946–53. 131. Akabas MH, Kaufmann C, Archdeacon P, Karlin A. Identification of acetylcholine receptor channel-lining residues in the entire M2 segment of the a subunit. Neuron. 1994;13:919–27. 132. Akabas MH, Kaufmann C, Cook TA, Archdeacon P. Amino acid residues lining the chloride channel of the cystic fibrosis transmembrane conductance regulator. J Biol Chem. 1994;269:14865–8. 133. Xu M, Akabas MH. Amino acids lining the channel of the g-aminobutyric acid type a receptor identified by cysteine substitution. J Biol Chem. 1993;268:21505–8. 134. Ku¨rz LL, Zu¨hlke RD, Zhang H-J, Joho RH. Side-chain accessibilities in the pore of a K+ channel probed by sulfhydryl-specific reagents after cysteine-scanning mutagenesis. Biophys J. 1995;68:900–5. 135. Kronengold J, Trexler EB, Bukauskas FF, Bargiello TA, Verselis VK. Single-channel SCAM identifies pore-lining residues in the first extracellular loop and first transmembrane domains of Cx46 hemichannels. J Gen Physiol. 2003; 122:389–405. 136. Unger VM, Kumar NM, Gilula NB, Yeager M. Three-dimensional structure of a recombinant gap junction membrane channel. Science. 1999;283:1176–80. 137. Trexler EB, Bukauskas FF, Kronengold J, Bargiello TA, Verselis VK. The first extracellular loop domain is a major determinant of charge selectivity in connexin46 channels. Biophys J. 2000;79:3036–51. 138. Oh S, Rivkin S, Tang Q, Verselis VK, Bargiello TA. Determinants of gating polarity of a connexin 32 hemichannel. Biophys J. 2004;87:912–28. 139. Musa H, Fenn E, Crye M, Gemel J, Beyer EC, Veenstra RD. Amino terminal glutamate residues confer spermine sensitivity and affect voltage gating and channel conductance of rat connexin40 gap junctions. J Physiol. 2004;557:863–78. 140. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys. 2001;34:325–472. 141. Werner R, Levine E, Rabadan-Diehl C, Dahl G. Formation of hybrid cell-cell channels. Proc Natl Acad Sci USA. 1989;86:5380–4.

72

M. Yeager

142. Hennemann H, Suchyna T, Lichtenberg-Frate´ H, Jungbluth S, Dahl E, Schwarz J, Nicholson BJ, Willecke K. Molecular cloning and functional expression of mouse connexin40, a second gap junction gene preferentially expressed in lung. J Cell Biol. 1992;117:1299–310. 143. White TW, Bruzzone R, Wolfram S, Paul DL, Goodenough DA. Selective interactions among the multiple connexin proteins expressed in the vertebrate lens: the second extracellular domain is a determinant of compatibility between connexins. J Cell Biol. 1994;125:879–92. 144. Dahl E, Manthey D, Chen Y, Schwarz HJ, Chang YS, Lalley PA, Nicholson BJ, Willecke K. Molecular cloning and functional expression of mouse connexin-30, a gap junction gene highly expressed in adult brain and skin. J Biol Chem. 1996;2714:17903–10. 145. Cao F, Eckert R, Elfgang C, Nitsche, J. M., Snyder, S. A., Hu¨lser, D.F., Willecke, K., Nicholson, B. J. A quantitative analysis of connexin-specific permeability differences of gap junctions expressed in HeLa transfectants and Xenopus oocytes. J Cell Sci. 1998;111:31–43. 146. Suchyna TM, Xu LX, Gao F, Fourtner CR, Nicholson BJ. Identification of a proline residue as a transduction element involved in voltage gating of gap junctions. Nature. 1993;365:847–9. 147. Verselis VK, Ginter CS, Bargiello TA. Opposite voltage gating polarities of two closely related connexins. Nature. 1994;368:348–54. 148. Dermietzel R, Volker M, Hwang T-K, Berzborn RJ, Meyer HE. A 16 kDa protein ¨ co-isolating with gap junctions from brain tissue belonging to the class of proteolipids of the vacuolar H+-ATPases. FEBS Lett. 1989;253:1–5. 149. Yamamoto T, Ochalski A, Hertzberg EL, Nagy JI. LM and EM immunolocalization of the gap junctional protein connexin 43 in rat brain. Brain Res. 1990;508:313–9. 150. Nagy JI, Yamamoto T, Sawchuk MA, Nance DM, Hertzberg EL. Quantitative immunohistochemical and biochemical correlates of connexin43 localization in rat brain. Glia. 1992;5:1–9. 151. Nagy JI, Ochalski PAY, Li J, Hertzberg EL. Evidence for the co-localization of another connexin with connexin 43 at astrocytic gap junctions in rat brain. Neuroscience. 1997;78:533–48. 152. Dermietzel R, Traub O, Hwang TK, Beyer, E., Bennett, M. V. L., Spray, D. C., Willecke, K. Differential expression of three gap junction proteins in developing and mature brain tissues. Proc Natl Acad Sci USA. 1989;86:10148–52. 153. Dermietzel R, Farooq M, Kessler JA, Althaus H, Hertzberg EL, Spray DC. Oligodendrocytes express gap junction proteins connexin32 and connexin45. Glia. 1997;20:101–14. 154. Kunzelmann P, Blu¨mcke I, Traub O, Dermietzel R, Willecke K. Coexpression of connexin45 and -32 in oligodendrocytes of rat brain. J Neurocytol. 1997;26:17–22. 155. Kuraoka A, Iida H, Hatae T, Shibata Y, Itoh M, Kurita T. Localization of gap junction proteins, connexins 32 and 26, in rat and guinea pig liver as revealed by quick-freeze, deep-etch immunoelectron microscopy. J Histochem Cytochem. 1993;4:971–80. 156. Zhang J-F, Ellinor PT, Aldrich RW, Tsien RW. Molecular determinants of voltagedependent inactivation in calcium channels. Nature. 1994;372:97–100. 157. Ghosh S, Safarik R, Klier G, Monosov E, Gilula NB, Kumar NM. Evidence for gap junction hetero-oligomers. Mol Biol Cell Supp. 1995;6:189a. 158. Falk MM, Buehler LK, Kumar NM, Gilula NB. Cell-free synthesis and assembly of connexins into functional gap junction membrane channels. EMBO J. 1997;16:2703–16. 159. Bevans CG, Kordel M, Rhee SK, Harris AL. Isoform composition of connexin channels determines selectivity among second messengers and uncharged molecules. J Biol Chem. 1998;273:2808–16.

2 Channel Structure

73

160. Brink PR, Cronin K, Banach K, Peterson E, Westphale EM, Seul KH, Ramanan SV, Beyer EC. Evidence for heteromeric gap junction channels formed from rat connexin43 and human connexin37. Am J Physiol. 1997;273: C1386-C96. 161. Dahl G, Levine E, Rabadan-Diehl C, Werner R. Cell/cell channel formation involves disulfide exchange. Eur J Biochem. 1991;197:141–4. 162. Dahl G, Werner R, Levine E, Rabadan-Diehl C. Mutational analysis of gap junction formation. Biophys J. 1992;62:172–82. 163. Rahman S, Evans WH. Topography of connexin32 in rat liver gap junctions. Evidence for an intramolecular disulphide linkage connecting the two extracellular peptide loops. J Cell Sci. 1991;100:567–78. 164. John SA, Revel J-P. Connexon integrity is maintained by non-covalent bonds: intramolecular disulfide bonds link the extracellular domains in rat connexin-43. Biochem Biophys Res Commun. 1991;178:1312–8. 165. Foote CI, Zhou L, Zhu X, Nicholson BJ. The pattern of disulfide linkages in the extracellular loop regions of connexin 32 suggests a model for the docking interface of gap junctions. J Cell Biol. 1998;140:1187–97. 166. Weiss MS, Abele U, Weckesser J, Welte W, Schiltz E, Schulz GE. Molecular architecture and electrostatic properties of a bacterial porin. Science. 1991;254:1627–30. 167. Schirmer T. General and specific porins from bacterial outer membranes. J Struct Biol. 1998;121:101–9. 168. Abagyan R, Totrov M, Kuznetsov D. ICM: a new method for structure modeling and design: Applications to docking and structure prediction from the distorted native conformation. J Comp Chem. 1994;15:488–506. 169. Sosinsky GE, Nicholson BJ. Structural organization of gap junction channels. Biochim Biophys Acta. 2005;1711:99–125. 170. Bruzzone R, White TW, Paul DL. Expression of chimeric connexins reveals new properties of the formation and gating behavior of gap junction channels. J Cell Sci. 1994;107:955–67. 171. Haubrich S, Schwarz HJ, Bukauskas F, Lichtenberg-Frate´ H, Traub O, Weingart R, Willecke K. Incompatibility of connexin 40 and 43 Hemichannels in gap junctions between mammalian cells is determined by intracellular domains. Mol Biol Cell. 1996;7:1995–2006. 172. Unwin PNT, Ennis PD. Two configurations of a channel-forming membrane protein. Nature. 1984;307:609–13. 173. Unwin PNT, Ennis PD. Calcium-mediated changes in gap junction structure: evidence from the low angle X-ray pattern. J Cell Biol. 1983;97:1459–66. 174. Caspar DLD, Sosinsky GE, Tibbitts TT, Phillips WC, Goodenough DA. Gap junction structure. In: Hertzberg EL, Johnson RG, editors. Modern cell biology, gap junctions. New York: Alan R. Liss; 1988. pp. 117–33. 175. Makowski L. X-ray diffraction studies of gap junction structure. In: Advances in Cell Biology. Greenwich: JAI Press; 1988. pp. 119–58. 176. Liu F, Arce FT, Ramachandran S, Lal R. Nanomechanics of hemichannel conformations: connexin flexibility underlying channel opening and closing. J Biol Chem. 2006;281:23207–17. 177. Ek JF, Delmar M, Perzova R, Taffet SM. Role of histidine 95 on pH gating of the cardiac gap junction protein connexin43. Circ Res. 1994;74:1058–64. 178. Ek-Vitorin JF, Calero G, Morley GE, Coombs W, Taffet SM, Delmar M. pH regulation of connexin43: molecular analysis of the gating particle. Biophys J. 1996; 71:1273–84. 179. Morley GE, Taffet SM, Delmar M. Intramolecular interactions mediate pH regulation of connexin43 channels. Biophys J. 1996;70:1294–302. 180. Hoh JH, Sosinsky GE, Revel J-P, Hansma PK. Structure of the extracellular surface of the gap junction by atomic force microscopy. Biophys J. 1993;65:149–63.

74

M. Yeager

181. Mu¨ller DJ, Hand GM, Engel A, Sosinsky GE. Conformational changes in surface structures of isolated connexin 26 gap junctions. EMBO J. 2002;21:3598–607. 182. Yu J, Bippes CA, Hand GM, Mu¨ller DJ, Sosinsky GE. Aminosulfonate modulated pH-induced conformational changes in connexin26 hemichannels. J Biol Chem. 2007;282:8895–904. 183. Purnick PE, Benjamin DC, Verselis VK, Bargiello TA, Dowd TL. Structure of the amino terminus of a gap junction protein. Arch Biochem Biophys. 2000;381:181–90. 184. Stauffer KA, Kumar NM, Gilula NB, Unwin N. Isolation and purification of gap junction channels. J Cell Biol. 1991;115:141–50. 185. Unger VM, Kumar NM, Gilula NB, Yeager M. Projection structure of a gap junction membrane channel at 7 A resolution. Nat Struct Biol. 1997;4:39–43. 186. Kovacs JA, Baker KA, Altenberg GA, Abagyan R, Yeager M. Molecular modeling and mutagenesis of gap junction channels. Prog Biophys Mol Biol. 2007;94:15–28. 187. Yeager M., Harris AL. Gap junction channel structure in the early twenty-first century: facts and fantasies. Curr Opin Cell Biol. 2007;19:521–528. 188. Goodenough DA, Paul DL, Jesaitis L. Topological distribution of two connexin32 antigenic sites in intact and split rodent hepatocyte gap junctions. J Cell Biol. 1988;107:1817–24. 189. Willecke K, Traub O, Look J, Stutenkemper R, Dermietzel R. Different protein components contribute to the structure and function of hepatic gap junctions. In: Hertzberg EL, Johnson RG, editors. Modern cell biology. New York: Alan R. Liss; 1988. pp. 41–52. 190. Beyer EC, Kistler J, Paul DL, Goodenough DA. Antisera directed against connexin43 peptides react with a 43-kD protein localized to gap junctions in myocardium and other tissues. J Cell Biol. 1989;108:595–605. 191. Yancey SB, John SA, Lal R, Austin BJ, Revel J-P. The 43-kD polypeptide of heart gap junctions: immunolocalization, topology, and functional domains. J Cell Biol. 1989;108:2241–54. 192. el Aoumari A, Fromaget C, Dupont E, Reggio H, Durbec P, Briand JP, Boller K, ¨ Kreitman B, Gros D. Conservation of a cytoplasmic carboxy-terminal domain of connexin 43, a gap junctional protein, in mammal heart and brain. J Membr Biol. 1990;115:229–40. 193. Laird DW, Revel J-P. Biochemical and immunochemical analysis of the arrangement of connexin43 in rat heart gap junction membranes. J Cell Sci. 1990;97:109–17. 194. Sa´ez JC, Nairn AC, Czernik AJ, Spray DC, Hertzberg EL, Greengard P, Bennett MVL. Phosphorylation of connexin 32, a hepatocyte gap-junction protein, by cAMP-dependent protein kinase, protein kinase C and Ca2+/calmodulin-dependent protein kinase II. Eur J Biochem. 1990;192:263–73. 195. Swenson KI, Piwnica-Worms H, McNamee H, Paul DL. Tyrosine phosphorylation of the gap junction protein connexin43 is required for the pp60v-src-induced inhibition of communication. Cell Regul. 1990;1:989–1002. 196. Kennelly PJ, Krebs EG. Consensus sequences as substrate specificity determinants for protein kinases and protein phosphatases. J Biol Chem. 1991;266:15555–8. 197. Rubin JB, Verselis VK, Bennett MVLL, Bargiello TA. A domain substitution procedure and its use to analyze voltage dependence of homotypic gap junctions formed by connexins 26 and 32. Proc Natl Acad Sci USA. 1992;89:3820–4. 198. Goldberg GS, Lau AF. Dynamics of connexin43 phosphorylation in pp60v-src-transformed cells. Biochem J. 1993;295:735–42. 199. Kanemitsu MY, Lau AF. Epidermal growth factor stimulates the disruption of gap junctional communication and connexin43 phosphorylation independent of 12-0-tetradecanoylphorbol 13-acetate-sensitive protein kinase C: the possible involvement of mitogen-activated protein kinase. Mol Biol Cell. 1993;4:837–48. 200. Liu S, Taffet S, Stoner L, Delmar M, Vallano ML, Jalife J. A structural basis for the unequal sensitivity of the major cardiac and liver gap junctions to intracellular acidification: the carboxyl tail length. Biophys J. 1993;64:1422–33.

2 Channel Structure

75

201. Sa´ez JC, Nairn AC, Czernik AJ, Spray DC, Hertzberg EL. Rat connexin43: regulation by phosphorylation in heart. In: Hall JE, Zampighi GA, Davis RM, editors. Progress in Cell Research, Gap Junctions. Amsterdam: Elsevier Science Publishers; 1993. pp. 275–81. 202. Loo LW, Berestecky JM, Kanemitsu MY, Lau AF. pp60src-mediated phosphorylation of connexin 43, a gap junction protein. J Biol Chem. 1995;270:12751–61. 203. Kwak BR, Sa´ez JC, Wilders R, Chanson M, Fishman GI, Hertzberg EL, Spray DC, Jongsma HJ. Effects of cGMP-dependent phosphorylation on rat and human connexin43 gap junction channels. Pflu¨gers Arch Eur J Physiol. 1995;430:770–8. 204. Warn-Cramer BJ, Lampe PD, Kurata WE, Kanemitsu MY, Loo LW, Eckhart W, Lau AF. Characterization of the mitogen-activated protein kinase phosphorylation sites on the connexin-43 gap junction protein. J Biol Chem. 1996;271:3779–86. 205. Kanemitsu MY, Loo LW, Simon S, Lau AF, Eckhart W. Tyrosine phosphorylation of connexin 43 by v-Src is mediated by SH2 and SH3 domain interactions. J Biol Chem. 1997;272:22824–31. 206. Wang XG, Peracchia C. Positive charges of the initial C-terminus domain of Cx32 inhibit gap junction gating sensitivity to CO2. Biophys J. 1997;73:798–806. 207. Chang CT, Wu C-SC, Yang JT. Circular dichroic analysis of protein conformation: inclusion of the b-turns. Anal Biochem. 1978;91:13–31.

Chapter 3

The Connexin Channel Pore: Pore-Lining Segments and Residues Vytas K. Verselis

Abstract Connexins form channels with large aqueous pores that mediate flux of ions and biological signaling molecules between cells and across the plasma membrane. Conductance and selectivity properties of connexin channels are diverse. A number of studies have attempted to define the connexin pore in molecular terms and to reveal the key elements of the pore structure that may differ across connexins. These studies include point mutagenesis, domain exchange, scanning cysteine accessibility mutagenesis, and electron cryomicroscopy. Each type of study has provided information that points to specific segments of connexin protein as contributing to the pore lining. However, despite this work, and significant advances in understanding channel structure, gating, and permeation, current views have not reached a consensus regarding the principal domains contributing to the pore. Keywords Gap junction Hemichannel Cysteine scanning accessibility mutagenesis Conductance Selectivity Pore Cx26 Cx32 Cx37 Cx40 Cx43 Cx46

3.1 Introduction Connexins comprise a class of ion channels with characteristically wide aqueous pores. In addition to mediating the flux of ions such as Kþ and Ca2þ, connexin channels have been shown to be permeable to a variety of metabolites and signaling molecules such as amino acids, nucleotides, small peptides, and sugars [1,2] (see Chapter 7). Recently, Cx43 gap junction channels were shown to mediate antigen cross-presentation through peptide transfer [3] and to be permeable to some small interfering RNAs [4]. Connexin channels, however, are not simply conduits for any molecule up to a certain cutoff size, which is often estimated to be 1 kDa; the permeability properties of different connexin V.K. Verselis (*) Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_3, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

77

78

V.K. Verselis

channels differ dramatically [1,5,6]. For example, some connexin channels prefer cations while others prefer anions, some restrict fluxes of molecules above 400 Da, and others, specifically heteromeric hemichannels, can even discriminate among purine cyclic nucleotides and isomers of inositol triphosphate [7,8]. Unitary conductances range from less than 10 picoSiemens (pS) to greater than 300 pS. It has been presumed, and is becoming increasingly evident, that connexin channel properties play important roles in tissue function, and in the case of permeability, in determining which biological signals are transmitted between cells. Central to understanding the underlying basis for the diverse permeability characteristics of connexin channels is to define the aqueous pore in molecular terms. Over the years, numerous biophysical studies led to inferences about the structures of ion channel pores and the basis for ionic selectivity, particularly in the exquisitely selective voltage-dependent Naþ, Kþ, and Ca2þ channels. A significant leap forward was made upon obtaining a crystal structure of the bacterial KcsA Kþ channel [9] and subsequently voltage-dependent bacterial and mammalian Kþ channels [10,11]. Reassuringly, structural inferences from biophysical studies regarding permeation and selectivity in Kþ channels are in reasonable agreement with the general features of the pore revealed by the crystal structure (reviewed in [12]). For connexins and a host of other ion channels, high-resolution crystal structures have not been obtained and descriptions of pore structure remain at the level of inference gathered from a combination of biochemical, biophysical, and, where available, lower-resolution structural data. There is no consensus, as yet, as to which domains principally line the pore in the transmembrane span, and the issue of pore structure remains very much a work in progress. This review summarizes current knowledge of pore-lining domains of connexins derived from a convergence of biophysical and structural studies.

3.2 Gap Junction Channel Structure and the Pore In gross structural terms, gap junction cell–cell channels are composed of two docked hemichannels, one contributed by each of two apposed cells. Biochemical and structural studies indicate that each hemichannel is a hexamer of connexin subunits, each having an accepted membrane topology of four transmembrane domains (M1 to M4) with the amino-terminal domain (NT) and the carboxyl-terminal domain (CT) located intracellularly (reviewed in [1]).

3.2.1 General Features: A Large Aqueous Pore Throughout In 1999, a structure of a gap junction channel was obtained at 7.5 A˚ resolution in the plane of the membrane from a three-dimensional (3D) projection density map derived by electron crystallography of frozen, hydrated two-dimensional (2D) crystals of recombinant Cx43 protein [13] (see Chapter 2). Since the CT of

3 Pore-Lining Segments and Residues

79

Cx43 is sensitive to proteolysis and thus could introduce inhomogeneity in the 2D crystal, it was truncated at amino acid L263 to remove the protease-sensitive site, along with much of the CT (Cx43L263). Views along the gap junction channel axis reveal 24 -helical bundles, suggestive of four transmembrane helices per connexin subunit, consistent with accepted membrane topology. The structure contains a central aqueous cavity throughout, spanning both apposing membranes and the intervening gap, which constitutes an overall length of 160 A˚

Fig. 3.1. Structure of a Cx43 gap junction channel and putative helix assignments to the pore. (A) Shown are side views of the channel; the right view is cropped to show the channel interior. Indicated are the approximate boundaries for the membrane (M), extracellular gap (E), and the cytoplasm. The red asterisk marks the narrowest part of the channel where the diameter of the aqueous pore is estimated to be 15 A˚ without taking amino acid side chains into consideration. (B) Side view of part of the 3D map shown in (a). Only the cytoplasmic and most of the membrane spanning regions of one hemichannel are shown. The red asterisk designates the most visible extension into the cytoplasm. The four transmembrane helices are labeled A, B, C, and D. The white and red lines follow the center of gravity for the four helices. There are segments of helix C and B that appear to line the pore (designated with red lines). (A,B: Adapted from Unger et al. [13] with permission.) (C) A model for the structure of the connexin hemichannel. Putative helical assignments are as indicated. Yellow spheres indicate residues predicted to map to the pore. These span five helical turns on the M3 segment and also include residues in M1 at the extracellular end. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com) (From Fleishman et al. [18] with permission.)

80

V.K. Verselis

(Fig. 3.1a). General features of the pore include wide vestibules at the cytoplasmic ends, pronounced narrowing at the outer membrane borders, and a widening in the region of the extracellular gap. Neglecting contributions of side chains, the Cx43 pore is 40 A˚ wide at the cytoplasmic vestibules, narrows to 15 A˚ at the external boundaries of each bilayer, and widens again to 25 A˚ in the intercellular gap. The pore should be narrowed considerably by side chains extending into the pore lumen. A 3D-reconstructed image of a Cx26 gap junction channel containing an M34A mutation, which is associated with hereditary deafness, was recently reported [14] and shows similarities to the Cx43L263 3D structure in terms of shape, size, and arrangements of the transmembrane helices. A distinguishing feature was the presence of a cytoplasmic density in the Cx26 pore. It is unclear if this density is a consequence of the M34A mutation or of the crystallization methods that were used. What about the pore in an unapposed hemichannel? At low (14 A˚) resolution, electron cryomicroscopy (cryoEM) and image analysis of gap junctions that are intact or split show similar diffraction patterns [15]. Using atomic force microscopy (AFM), unapposed Cx26 hemichannels were examined by removing one bilayer of a gap junction plaque through force dissection [16]. Images of the extracellular surface of a hemichannel show six subunits protruding 16 A˚ from the membrane arranged in a doughnut-shaped structure surrounding a central pore. The average diameter of the pore was 13 A˚. The cytoplasmic surface similarly formed a doughnut-shaped structure protruding 17 A˚ from the membrane. Average pore diameter was considerably wider, 28 A˚. The overall height of a Cx26 hemichannel, including the 16 and 17 A˚ protrusions, is 80 A˚, which is approximately half the length of the full gap junction channel revealed in both AFM and cryoEM images. In another AFM study, Cx43 hemichannels were examined after reconstitution into lipid bilayers [17]. Structural features of the pore on the extracellular face of Cx43 hemichannels were similar to those of Cx26 hemichannels exposed after force dissection; the height of the cytoplasmic protrusion differed presumably because of the differences in the lengths of the Cx26 and Cx43 CT domains. There were some differences in the structural features of the extracellular protrusions, which may reflect differences in hemichannels that were reconstituted and never docked versus those that were docked and force-dissected. In general, however, there appears to be gross structural conservation of hemichannels in docked and undocked forms, notwithstanding some structural variation in the extracellular domains. Pore dimensions obtained with AFM differ somewhat from the cryoEM model of Unger et al. [13], which may reflect the differences in the technologies as well as the conditions of the isolated junctional preparations.

3.2.2 Mapping Connexin Sequence to the Structure In the original structure of Unger et al. [13], distinct molecular boundaries of the subunits could not be assigned to the 24 -helical bundles. Nonetheless, two

3 Pore-Lining Segments and Residues

81

different helices, designated helix B and helix C, appear to line the pore in this model (Fig. 3.1b). Helix C has a substantial tilt, which accounts for the narrowing of the pore at the external boundary of the membrane. Likewise, the tilt takes helix C away from the pore proceeding toward the cytoplasm, exposing the second helix, B. A map of Cx32 superimposed onto that of Unger et al. [13] with improved in-plane resolution down to 5.7 A˚ allowed the helices, their positions, and tilt angles to be discerned within the transmembrane span. The connecting loops, however, were still undefined and thus the helices still could not be unambiguously assigned [18]. Combining this improved map with computational analyses using evolutionary conservation and hydrophobicity among connexin sequences, Fleishman et al. [18] generated putative assignments of the helices in the map to the predicted connexin transmembrane domains and proposed a number of potential packing interfaces. Pertaining to the aqueous pore, Fleishman et al. [18] suggest that helix C, which is more exposed to the aqueous pore, likely corresponds to M3, and that helix B corresponds to M1. Optimal helix orientations were computed based on these assignments, which provided residues predicted to line the aqueous pore. This model placed three of four charged residues in M1 and M3 with their side chains possibly extending into the lumen of the pore. Using the Cx32 sequence as a working example, residues R142 and E146 in M3 were designated as partly pore-lining, interacting, to some extent, with R32 in M1. These three charged positions could form a thin 4 to 5 A˚ polar belt around the pore lumen near the middle of the transmembrane span that could have a significant impact on ion conduction and charge selectivity. Other residues predicted to be exposed in the pore include W132, V139, and A147 in M3 and A35 and A39 in M1 (Fig. 3.1c). In the extracellular gap, details of secondary structure were not revealed, but this region exhibits a double-layered appearance, with the interior layer of protein forming a continuous wall, which suggests a design that could function to prevent leak of ions and other permeant molecules to the extracellular space where the hemichannels come together and dock. Given that the extracellular loop domains, E1 and E2, constitute the extracellular part of the channel, one of the extracellular loops likely forms the interior wall that lines the pore. In unapposed hemichannels, one might expect that the same extracellular loop segment constitutes the extracellular vestibule of the pore, even though there are undoubtedly some changes in structure associated with docking. The cytoplasmic aspect of the gap junction channel was not revealed in the 3D map, indicative of a flexible and poorly ordered region. The cytoplasmic domains are extensive and variable in sequence, particularly in the CT domain and the loop domain, CL, which connects M2 and M3.

3.3 Structure-Function Studies and the Connexin Channel Pore Structure-function approaches that combine mutagenesis and electrophysiological recording have an advantage over crystallography in that they probe functioning ion channels, often in their native environment. However, they

82

V.K. Verselis

rely on several types of experiments to deduce structure. A complicating factor, both for structure-function and crystallographic approaches, is that structure is not static. When deducing the structure of an ion channel pore, mutagenic studies have generally focused on changes in characteristics associated with an open channel such as unitary conductance and selectivity. The substituted cysteine accessibility method (SCAM), which combines chemical and mutagenesis approaches, is, perhaps, a more definitive means of identifying pore-lining residues in functioning channels. In addition, voltage-gating and the pore appear to have a unique relationship in connexin channels. Results from a variety of structure-function studies are described below.

3.3.1 Voltage-Gating and the Connexin Pore 3.3.1.1 Connexin Voltage Sensors and the Pore Likely Share Common Elements In the large family of voltage-dependent ion channels that includes Kþ, Naþ, and Ca2þ channels, the voltage sensor forms what appears to be a discrete domain, termed the voltage-sensing domain (VSD). When viewed in the plane of the membrane, four VSDs, one associated with each subunit, stick out from the corners of a central pore structure that is assembled from four pore-lining domains [19,20]. Linker regions translate the motions of the VSDs to the channel pore. Interestingly, the design of a VSD suggests that it could function independently of the pore-forming domains as a general membrane voltagesensing device. Such a design is not utilized in connexin channels, where the pore and the voltage sensor appear to share common elements. This design likely reflects a functional requirement, namely that gap junction channels sense the voltage difference between cells, that is, the transjunctional voltage (Vj), and not the voltage across a single membrane (Vm) [21] (reviewed in [22]). This means that coupling between two cells is modulated when the voltage of one cell changes relative to the other and that changes in Vm have no effect as long as the same Vj is maintained. To explain this property, it has been reasoned that the voltage-sensing elements are positioned in or near the pore where the transjunctional field can remain relatively constant for any given Vj, regardless of the absolute values of Vm [23,24] (see Chapter 4). Before describing molecular studies that identified residues contributing to connexin voltage sensing, and by inference the pore, it should be mentioned that each connexin hemichannel contains two distinct voltage-gating mechanisms [22]. One mechanism is characterized by gating transitions to a stable subconducting state, termed the residual conductance state, which has a greatly reduced but still significant channel conductance, that is, the pore is not completely occluded to ion conduction [25,26]. This mechanism is generally referred to as Vj/fast-gating, as it was the first mechanism described and believed at the time to be the sole mechanism that gated the gap junction channels in response to Vj. The second voltage-gating mechanism closes channels fully, that is, ion

3 Pore-Lining Segments and Residues

83

conduction is abolished. The signature of this mechanism is that the opening and closing gating transitions are composed of a series of transient subconductance states, which gives them the appearance of being slow when recorded at typical filtering frequencies. This mechanism has been provisionally termed loop/slow-gating, for it is believed to involve closure mediated by conformational changes in the extracellular loops [27]. Both of these gating mechanisms are sensitive to Vj in junctional channels and both mechanisms are intrinsic to hemichannels. 3.3.1.2 Voltage-Gating Studies Implicate the Amino-Terminal Domain and the First Extracellular Loop as Contributing to the Pore Within a typical family of voltage-dependent ion channels, gating polarity is fixed. For example, all Naþ channels open (activate) with membrane depolarization and close with hyperpolarization. In connexin channels, however, the polarity of Vj/fast-gating does not appear to be fixed. The Vj/fast gate in some connexin channels closes when the cytoplasmic side of the hemichannels is made positive, and in others when it is made negative. Molecular studies have shown that the NT contains the Vj/fast-gating voltage sensor and that the sign of the charge of the sensor confers the gating polarity [28,29]. Thus, it was proposed that the NT loops back into the membrane to form the cytoplasmic vestibule of the pore where residues can sense the local field created by Vj [30]. Which residues in the NT are involved? A negative charge substitution at the second position in the NT of Cx32 reversed the polarity of the Vj/fast-gating mechanism [28]. Because a hemichannel is hexameric, a single substitution replaces six charges. Using gating polarity as an assay, individually replacing the uncharged residues at the fifth, eighth, ninth, and tenth position in the NT with a negatively charged residue was also reported to reverse gating polarity; beyond position ten, substitutions had no effect [31]. These data indicate that residues in the NT, up to and including the tenth residue, can sense the transjunctional field and thus could contribute to the channel pore (Fig. 3.2). Subsequent studies at the single-channel level using a chimeric Cx32 hemichannel in which the E1 domain was substituted with that of the Cx43 sequence, Cx32*43E1, confirmed that the gating mechanism affected by these NT substitutions was, in fact, Vj/fast-gating [32]. A nuclear magnetic resonance (NMR) study of a Cx26 NT peptide, residues 1 to 15, showed a domain-hinge-domain structure with the hinge or open turn around a flexible glycine at position twelve that could potentially allow the amino terminus to loop back into the plane of the membrane and form the cytoplasmic vestibule of the pore [30]. The NMR structure of this putative pore forming part suggests some helical character, particularly at the amino terminus, and less ordered structure proceeding toward the hinge region (Fig. 3.2). In Cx40, Musa et al. [33] reported that substitutions of charged glutamic acid residues in the NT significantly altered voltage-dependent gating, consistent with the studies on Cx32. Likewise, Tong and Ebihara [34] reported that the charge at the ninth position in the NT of Cx56, the chicken ortholog of

84

V.K. Verselis

Fig. 3.2. Molecular model deduced from a nuclear magnetic resonance (NMR)-derived structure of the first 15 amino acids of Cx26 attached to M1. The model represents an axial projection viewed from the cytoplasm with six NTs forming the cytoplasmic vestibule of a channel. M1 is shown as a blue helix. The fifth residue is shaded yellow and the eighth in green. The turn at the glycine residue (G12) is indicated. The helical structure formed by residues two to ten is modeled to form the channel vestibule. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com). (Adapted from Purnick et al. [30] with permission.)

Cx46, contributed significantly to the difference in voltage-dependent gating between hemichannels formed of it and those formed by the closely related Cx45.6, the chicken ortholog of Cx50. A suggested mechanism involves residence of the ninth residue in the NT in the pore in the open state; movement of it in response to voltage leads to closure. Substitution of the ninth residue in the NT also affected single-channel conductance, as expected if it were pore-lining [34]. As with the NT, charge substitutions at the border of M1 and E1 in Cx32 also reversed gating polarity or suppressed the effects of charge substitutions in the NT domain [28]. Thus, like residues in the NT, those at the M1/E1 border were suggested to reside in the pore and contribute to voltage sensing. These experiments predated knowledge of the two distinct voltage-gating mechanisms, and it is now unclear which mechanism was affected by these charge substitutions. Similarly, Tong and Ebihara [34] reported that position 43 at the M1/E1 border of Cx56 or Cx45.6 significantly contributed to voltage-gating, like the ninth residue in the NT. In this case, the presence of a positively charged glutamic acid, E43, was associated with robust and rapid hemichannel closure. In the absence of divalent cations, however, the hemichannels did not appear to close.

3 Pore-Lining Segments and Residues

85

The simplest interpretation offered was that position E43 contributes to a divalent cation-binding site in the channel pore, which when occupied produces closure at negative potentials. Depolarization would thus represent unbinding and channel opening. As will be described later, E43 was identified as porelining in SCAM studies of Cx46 hemichannels.

3.3.2 Conductance and Selectivity Studies There have been a number of reports of mutations in connexins that affect single-channel conductance or selectivity, whether the mutations are experimentally introduced substitutions or ones that are naturally occurring in connexin-based diseases. These sites are scattered throughout the connexin sequence and obvious caveats abound, since point mutations can cause effects on conductance or selectivity even if they are remote from the pore through indirect (allosteric) effects. A number of studies moderate this concern, however, by more extensive mutagenesis and, where possible, by fitting together information derived from several independent approaches. 3.3.2.1 Open Channel Rectification and Selectivity Suggest the Amino-Terminal Domain and First Extracellular Loop Contribute to the Pore at Cytoplasmic and Extracellular Ends, Respectively As previously indicated, the NT domain has been suggested to form the cytoplasmic vestibule of the connexin channel pore, inferred from charge substitutions at the second, fifth, eighth, ninth, and tenth position in Cx32 that can reverse Vj/fastgating polarity. Using the Cx32*43E1 chimera that functions as an unapposed hemichannel [35], the current-voltage (I-V) relations of single Cx32*43E1 hemichannels were shown to have a modest inward rectification of the open state current, that is, lower conductance for current entering the cell compared with current leaving the cell [32]. Positive charge substitutions of NT residues at the second, fifth, and eighth positions increased inward rectification and substantially reduced unitary conductance, whereas negative charge substitutions linearized the open hemichannel I-V relation and increased unitary conductance. These data are consistent with electrostatic effects of charges at these locations in NT that are positioned in the pore, thereby influencing ion fluxes accordingly. Thus, effects of charge substitutions on open channel current rectification provide supportive evidence that some of the NT residues, identified in gating experiments, affect ion permeation electrostatically and therefore reside in the pore. Current through a single open unapposed Cx46 hemichannel was shown to inwardly rectify in symmetric salts, and reversal potentials under bi-ionic conditions showed Cx46 hemichannels to substantially prefer cations [27]. Taken together, these findings point to fixed negative charges in the Cx46 hemichannel pore located at the extracellular end that accumulate cations and thus produce larger currents flowing in the inward direction. In support, charge screening

86

V.K. Verselis

with high ionic strength solutions applied to the extracellular side substantially reduced the ratio of cation to anion permeability [36]. Furthermore, substitution of E1 of Cx32, an anion-preferring channel, into Cx46 reduced its unitary conductance, converted Cx46 from cation-preferring to anion-preferring, and changed the I-V relation form inwardly to outwardly rectifying.

3.3.2.2 Block by Polyamines Implicates the Amino-Terminal Domain A useful property for study of the pore of an ion channel is open channel block in which a blocker occludes ion flux by entry into the pore. Block can be voltage-dependent if the binding site lies within the electric field of the membrane. Spermine causes voltage-dependent block of Cx40 gap junction channels [37]. Addition of spermine to one cell of a Cx40 pair produced block when the cell containing spermine was made relatively positive, which favored movement of the positively charged polyamine into the channel. Block was found to be connexin-specific, and substitution of positively charged glutamic acid residues at the ninth and 13th positions in the NT of Cx40 with positively charged lysine residues that are present in spermine-insensitive Cx43 rendered Cx40 insensitive to block [33]. Although these results suggest that positively charged residues in the NT play a role in spermine block, substitution of glutamic acid residues at these positions into Cx43 did not make Cx43 sensitive to polyamines. Thus, it is not clear whether E9 and E13 are directly involved in binding spermine or whether mutating these positions produces structural alterations in the NT that disrupts binding. Nonetheless, these data suggest that the NT contains the receptor for voltage-dependent polyamine block, again providing evidence that the NT contributes to the cytoplasmic vestibule of the pore.

3.3.2.3 Exchange of Unitary Conductances Implicates the First Transmembrane Domain An approach that has been applied to a number of ion channels has been to construct chimeras using different isoforms in hopes of exchanging functional properties. Such a chimeric approach has several advantages over point mutations. One is that there is an expected phenotype, which is that a specific property of the donor channel, in this case unitary conductance or selectivity, is transferred to the recipient channel. Another is that large segments of sequence can be substituted, which is likely to be more successful at transferring intact a given property by minimizing anomalous structural alterations that can accompany point mutations. Finally, there is the capability of examining reciprocity, which minimizes the possibility that substitution in one direction produces the expected result by chance. Considering that there are 21 members in the connexin family, with substantial differences in conductance and selectivity of the channels formed by them, a chimeric approach is potentially useful for identifying domains involved in formation of the pore.

3 Pore-Lining Segments and Residues

87

Hu and Dahl [38] initially utilized Cx46 and chimeric Cx32*43E1 hemichannels to exchange domains. These hemichannels exhibit substantially different unitary conductances, channel kinetics, and preference to dwell in subconductance states. Exchange of the first transmembrane segment, M1, between these connexins resulted in hemichannels that exhibited most of the properties of the M1, donor. The importance of M1 in determining permeability was further supported by replacement of leucine at position 35 in M1 with glycine, which increased single-channel conductance and increased permeability to larger solutes. Rectification of the open state was not reproduced upon exchange of M1, suggesting that while this domain is an important determinant of conductance, there are additional elements outside this region. Hu et al. [39] extended the chimeric approach to Cx32 and Cx37, narrowing the relevant region to the external (carboxyl-terminal) half of M1. Replacing the carboxyl-terminal half of M1 of Cx46 with Cx32 or Cx37 resulted in channels with conductances similar to those of the Cx32*43E1 chimera and Cx37, respectively. Chimeras with only the cytoplasmic (amino-terminal) half of M1 retained the unitary conductance of wild-type Cx46. Thus, the external half of M1 is an important determinant of conductance, which as described below, is in agreement with SCAM studies of Cx46 hemichannels that identified M1 as contributing to the pore [40,41].

3.3.3 The SCAM Method Applied to Connexins SCAM is the most utilized structure-function approach to identify pore-lining residues in functioning ion channels and has been extraordinarily informative. In SCAM, individual residues are replaced with cysteine, one at a time, and accessibility to modification is evaluated using thiol-modifying reagents [42]. To date, there have been three primary SCAM studies of connexins. Results are summarized below. 3.3.3.1 SCAM of Connexin46 and Connexin32*43E1 Chimeric Hemichannels Identifies Residues in the First Transmembrane Domain The first SCAM study of connexins utilized maleimidobutyryl biocytin (MBB) as the thiol-modifying reagent to probe Cx46 and chimeric Cx32*43E1 hemichannels expressed in Xenopus oocytes [40]. Hemichannels were considered to be advantageous in that, like conventional ion channels, they are amenable to SCAM studies in which currents associated with channels expressed in the plasma membrane can be easily assayed electrophysiologically and changes in currents readily assessed upon bath application of reagent. Examining subsets of residues within M1 and M3, Zhou et al. [40] reported that two residues in M1 appeared to be modified based on an observed reduction in hemichannel current upon application of MBB to the bath. These residues were I34C and L35C in Cx46 hemichannels and the equivalent positions I33C and M34C in chimeric Cx32*43E1 hemichannels.

88

V.K. Verselis

In both cases, addition of MBB produced a 40 to 50% reduction in macroscopic current. At the level of a single hemichannel, MBB was reported to reduce unitary current of Cx46L35C hemichannels by 80% [43]. The discrepancy in the magnitude of the effect examined macroscopically and at the single-channel level was not explained. A possibility is that this modification increases open probability, which would be reflected in the magnitude of the macroscopic current. In Cx32*43E1 chimeric hemichannels, SCAM results of M3 were inconclusive. M3 of Cx46 was not tested. 3.3.3.2 SCAM of Connexin32 Gap Junction Channels Identifies the Third Transmembrane Domain as the Major Pore-Lining Helix Throughout the Membrane Span Concerned that results obtained from hemichannels may not be applicable to gap junctional channels, Skerrett et al. [44] undertook a SCAM study of Cx32 using a cut-open paired Xenopus oocyte technique whereby two oocytes are each placed in separate compartments, separated by a septum, and are permitted to establish contact and hence form gap junctions through a hole in the septum. To gain access to the channel pore, which is not exposed to the surface membrane, one oocyte was cut open exposing the cytoplasm to the solution in the compartment to which thiolmodifying reagent was added. The electrical coupling was measured as the conductance between the two compartments. In this study, all four transmembrane domains were probed using MBB as the thiol-modifying reagent. In total, a number of sites in all four transmembrane domains were reported to show an electrophysiological effect upon addition of MBB, namely a decrease in junctional conductance. The reductions in conductance, in general, were surprisingly small, ranging from 10 to 20%. Sites reactive in M1 and M4 were deemed to be outside the pore as MBB applied extracellularly in intact oocyte pairs showed the same reductions in conductance. Most of the sites ascribed to the pore were in M3 and a few were in M2. From the cytoplasmic end of M3, residues Y135C, V139C, F141C, and L144C were deemed reactive, forming a pattern of reactivity consistent with -helical structure. S138C was reported to significantly modify gating and could be tested for accessibility in the closed state only. It was concluded that S138C reacted in the closed state and that it would likely react in the open state because it lies on the reactive side of the helix. Continuing toward the extracellular gap, three additional residues were deemed reactive: A147C, F149C, and Y151C. These residues apparently deviate from helical periodicity, adopting a pattern consistent with a b-sheet structure. In M2, L80C, V84C, and L89C were reactive sites. Thus, Skerrett et al. [44] concluded that in the open channel, M3 is the major pore-lining helix throughout the transmembrane span, but that M2 also contributes to the pore at the cytoplasmic end. All residues in M1 found to be reactive to MBB in Cx46 hemichannels by Zhou et al. [40] were reactive in Cx32 gap junction channels. However, cysteine substitution of these residues produced a reverse gating phenotype and it was concluded that the residues were exposed in the closed state.

3 Pore-Lining Segments and Residues

89

3.3.3.3 Single-Channel SCAM of Connexin46 Hemichannels Identifies the First Extracellular Loop and the First Transmembrane Domain as Major Pore-Lining Domains A study by Kronengold et al. [41] utilized Cx46 hemichannels expressed in Xenopus oocytes, as in Zhou et al. [40], but differed in using the much smaller methanethiosulfonate (MTS) reagents and in restricting studies to the level of single hemichannels in excised patches. Concerned that water-accessible surfaces within the transmembrane span of a channel can be in the form of aqueous crevices and modifications of cysteine-substituted residues in these crevices can affect current allosterically even if they are remote from the pore, Kronengold et al. [41] took advantage of the fact that the pore is a unique aqueous surface that is confluent with the aqueous compartments on both sides of the membrane. Thus, residues in the pore of an open hemichannel should be accessible to reagent added from either side of the membrane (Fig. 3.3). This type of SCAM

Fig. 3.3. Single-channel substituted cysteine accessibility method (SCAM) can distinguish porelining residues from those in other aqueous compartments. Schematic of a hemichannel with cysteine-substituted residues illustrated at three locations, on the extracellular surface, in a crevice facing the extracellular space, and in the channel pore. In an open channel, cysteine residues on the surface or in crevices are accessible to modification by thiol-reactive reagents from only one side of the membrane whereas cysteine residues in the pore are uniquely accessible from either side due to confluence with extracellular and cytoplasmic aqueous compartments. MTS, methanethiosulfonate. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com). (Adapted from Kronengold et al. [41] with permission.)

90

V.K. Verselis

experiment can be accomplished in excised patches containing single hemichannels where the state of an individual channel can be monitored continuously and changes in single open channel properties after modification can be assessed. Also, availability of a variety of MTS reagents allowed introduction of side chains differing in size and charge and assessment of whether the effects on open channel properties were appropriately changed. Based on sidedness of accessibility and effects of modification on open single hemichannel currents, pore-lining residues were identified within a 17 amino acid segment from D51C in the E1 domain extending to L35C, which is predicted to be midway in M1. Starting with the outermost residue in E1, the reactive sites included D51C, G46C, E43C, A39C, and L35C. Positional effects of modification and opposing effects of the oppositely charged MTS reagents MTS-ETþ and MTS-ES– on single hemichannel conductance and open hemichannel current rectification were consistent with a location of these residues within the path of permeating ions. For example, reaction of D51C with MTS-ETþ reduced unitary conductance by 40% and abolished inward rectification, whereas MTS-ES– slightly increased conductance as well as open hemichannel rectification. Multiple reactions, or ‘‘hits,’’ presumably representing reactions on individual subunits, were visible. Occlusion studies and the similarity of the effects of positively charged MTS reagents and substitutions with positively charged amino acids, where all six subunits were sure to be replaced, indicated that all six subunits were likely modified by MTS-ETþ and MTS-ES–. Application of single hemichannel SCAM further toward the cytoplasmic end of M1 was inconclusive because cysteine substitution alone significantly affected the hemichannels, often causing rapid flickery gating that precluded reliable assessment of single-channel conductance and subsequent effects of MTS modification [41]. Therefore, the identity of the pore cytoplasmic to L35C was not elucidated in this study. However, SCAM over a span of ten residues in M3, E166 to F175, which is complementary to the reactive segment in M1, showed no evidence of reactivity with any of the MTS reagents [41]. Thus, these studies suggest that M3 does not contribute to the pore, at least in the extracellular half of the transmembrane span, and that it is formed of M1 continuing into E1 as the hemichannel emerges extracellularly.

3.4 Summary and Critical Evaluation of Data Studies of voltage-gating, selectivity, block, domain exchange, and SCAM implicate several regions within the connexin molecule as potential contributors to the pore, including the NT, M1, M2, M3, and E1. Although there is some agreement obtained from independent approaches, fundamentally there is a lack of a consensus about the identity of the major pore-lining helix within the transmembrane span of the channel as identified by SCAM. This lack of consensus merits critical evaluation of the possibilities that have been raised to explain the different assignments, which include connexin-specific

3 Pore-Lining Segments and Residues

91

differences and use of hemichannels versus cell–cell channels. In addition, the SCAM studies differ in the choice of thiol reagents used and in the use of macroscopic versus single-channel recordings to assay accessibility. These issues are addressed below.

3.4.1 Connexin-Specificity Phylogenetic studies suggest that there are two major groups of connexins, group II or -group, and group I or b-group [45,46] (see Chapter 1). Given that Cx32 belongs to group I, and Cx46 to group II, can the contrasting assignments in these two connexins be the result of connexin-specific differences in structure? Although there may be some structural motifs characteristic of each phylogenetic group, a gross difference in overall structure such that different helices contribute to pore formation appears to be incompatible with the similarities in the 3D structures derived for Cx43 and Cx26 [13,14]; the latter is a b-group connexin closely related to Cx32. A gross difference in overall structure is also incompatible with the degree of sequence conservation among all connexins in the transmembrane and extracellular loop domains as well as in the overall similarities in channel properties among all connexins; that is, they all form large channels permeable to hydrophilic dyes, they all gate in response to Vj, and most respond in the same way to the same modulators, for example, acidification and alkanols. Finally, the SCAM study of Zhou et al. [40] showed the same results for Cx46 (-group) and the chimera Cx32*43E1, which is mostly of Cx32 sequence and thus a b-group connexin. That is not to say that there will not turn out to be some differences between groups or even among connexin isoforms, but they will likely amount to subtle differences in helix packing or rotation or channel flexibility that either locally shifts the pattern of accessibility or exposes adjacent residues in one connexin and not another. Such differences may be most evident in a domain such as the NT, where NMR and biophysical studies indicate a flexible structure [30]. The GVN motif at position twelve to 14 in the b-group connexins appears to be involved in forming a tight bend in the NT. The absence of this motif in -group connexins may result in a sufficiently different conformation and consequently a different subset of NT residues contributing to the pore. Substituted cysteine accessibility method of this domain may prove problematic as cysteine substitutions alone at many positions may cause substantial alterations in conformation.

3.4.2 Hemichannels Versus Gap Junction Channels Perhaps the most notable difference in the studies that have yielded contrasting views is the use of hemichannels versus gap junction channels. Although some structural changes undoubtedly occur upon hemichannel docking, can both hemichannel configurations differ sufficiently such that different helices line the

92

V.K. Verselis

pore? Although this possibility has not been rigorously tested, comparison of a number of biophysical properties of unapposed hemichannels and corresponding cell–cell channels strongly suggests overall conservation of structure. First, in most instances where unitary conductances for a given connexin have been assessed in both configurations, the conductance of an undocked hemichannel is approximately twice that of the gap junction channel, consistent with a view that an undocked hemichannel essentially is half of a gap junction channel [27,47,48]. Regarding selectivity, a gap junction channel that is cationselective remains cation-selective as a hemichannel [36], and a gap junction channel that restricts passage of larger solutes has a similar profile as a hemichannel [48]. As far as gating is concerned, hemichannels have been shown to possess the same two principal voltage-gating mechanisms found in gap junction channels, one showing gating to the residual (long-lasting) substate and the other full closures via a series of transient substates [27,47,49,50]. In the case of Cx43, tagging the carboxyl terminus with enhanced green fluorescent protein (EGFP), which selectively abolishes gating to the residual substate in Cx43 gap junction channels [51], was subsequently shown to do the same in Cx43 hemichannels [52]. Finally, hemichannels and cell–cell channels have the same pharmacological profiles, exemplified by sensitivity to pH, alkanols and the quinine derivative mefloquine [47,50,53]. To be sure, there are differences between undocked hemichannels and gap junction channels, but these appear to be principally regulatory in nature, exemplified by the high sensitivity of undocked hemichannels to external divalent cations [54] and, in the case of Cx50, to external Kþ [55]. Thus, not surprisingly, regulatory sites may be exposed when hemichannels are undocked, but this does not appear to translate into a fundamental rearrangement of channel architecture. There have been scattered reports of differences in hemichannel versus gap junction channel properties [56], but the overwhelming body of evidence suggests that the two configurations are minimally different. Definitive studies await a systematic comparison of pore-lining residues in both configurations in the same connexin, whether by SCAM or by higher resolution structural studies.

3.4.3 Comparison of Thiol Reagents Used in SCAM Studies To what extent can the choice of thiol-reactive reagent affect the results of SCAM? The SCAM studies in connexins have used a maleimide (MBB), in the case of Zhou et al. [40] and Skerrett et al. [44], and a series of thiosulfonates (MTS reagents), in the case of Kronengold et al. [41]. Both MBB and MTS reagents are highly reactive with thiol groups. However, maleimides can react with amines, although this usually requires pH >7.5. Perhaps the most important consideration is the reaction rate, because cysteine residues in ion channels can be located in one of three general environments: on hydrophilic surfaces, on surfaces exposed to the lipid, and on surfaces that contact other protein surfaces, whether they reside on the same protein or an associated (accessory) protein. By using a hydrophilic thiol

3 Pore-Lining Segments and Residues

93

reagent, reactions on hydrophilic surfaces, such as the pore, should be considerably faster than on those buried in the protein core or exposed to lipid. However, cysteine residues anywhere in the protein can be reactive over time. Reductions in current with MBB added extracellularly to oocytes occurred over a course of 30 minutes. A similarly slow time course was observed with MBB in the cut-open oocyte preparation, although there is the added complication of access to the junction. In optimal conditions, the cytoplasm of a cutopen oocyte can be exchanged in 1 minute [57]. Thus, such slow kinetics of modification raises the concern that over that time, cysteine residues buried in the protein core are also being accessed. On the other hand, MTS reagents exhibit very rapid reaction rates in solution [42] and have the additional advantage in that they are electrophilic and react considerably faster with the ionized form, S–, than with the un-ionized form, by a factor of 109, which reduces the likelihood of reaction on lipid-facing or protein-facing surfaces where thiols typically remain un-ionized due to the low-dielectric environment. In the study by Kronengold et al. [41], modifications occurred within seconds of MTS application, indicative of reaction at a hydrophilic, readily accessible surface. Another consideration in SCAM studies is whether the magnitude of the effect produced by modification is in accordance with expectations. MBB is a fairly large molecule with a molecular mass of 537 Da and an abaxial diameter of 11 A˚. This large size should not present a problem of permeability as connexins are large channels, but the logical expectation is that modification with MBB should produce a large reduction in current, if not complete block, considering the possibility that multiple, if not all six, subunits can be modified. The reductions reported in the case of M1 residues in Cx46 hemichannels [40] were 40 to 50%, and in the case of M3 in Cx32 gap junction channels were surprisingly small, 10 to 20% [44]. A possibility is that the size of MBB may limit reactivity to fewer or even a single subunit, as suggested by Pfahnl and Dahl [43], who observed a single step change in current in single L35C hemichannels upon application of MBB. However, the reduction in single-channel conductance was 80%, considerably greater than the reduction observed macroscopically. In the case of Cx32, such steric hindrance would be difficult to reconcile with the small reduction in current observed; that is, ion conduction is not significantly impeded, suggesting that the pore remains substantially unoccluded after modification. The effects of MBB modification on singlechannel conductance of Cx32 gap junction channels were not examined. In comparison, MTS reagents are available with a number of different head groups, large and small. MTS-EAþ, MTS-ETþ, and MTS-ES– would fit in a cylinder 6 A˚ in diameter [42,58] and are certain to be permeable to connexin channels. One can introduce charged side chains of either sign, in the case of MTS-ETþ and MTS-ES–, or introduce a large, bulky side chain similar in size to MBB using MTSEA-biotin or with fluorophore head groups, such as fluorescein and rhodamine. Kronengold et al. [41] utilized this diversity and generally observed effects of modification in accordance with expectations. Modification by oppositely charged reagents produced substantially different results. With

94

V.K. Verselis

Cx46, which prefers cations and is hence biased with negative charge, modification with MTS-ETþ produced reductions in single-channel conductance and with MTS-ES– produced increases or little effect. Modification with MTSbiotin, which introduced a large side chain, produced large reductions in single-channel conductance, exceeding 60% in some cases and yielding noisy, flickery channels consistent with mobile, bulky side chains residing in the pore that significantly impede ion fluxes. There was also evidence that steric hindrance may have limited the number of reactions to fewer than six, that is, one per connexin subunit [59]. Overall, the differing effects of the differing MTS reagents and their qualitative accordance with expectations, raise confidence that the modified residues reside in the pore.

3.4.4 Electrophysiological Assays Recordings of macroscopic current provide a convenient means to assess whether modification has occurred upon addition of thiol reagent to the bath. It is also convenient to assess reaction kinetics, which can provide information about the environment of the cysteine residue. However, macroscopic current is a product of n, I, and Po, where n is the number of channels, i is the singlechannel current, and Po is open probability. Thus, a change in macroscopic current can occur by altering any of these parameters, or combinations thereof, most notably i and Po, and need not result from modification of a pore-lining residue. This can lead to false positives if a residue remote from the pore alters gating and hence Po. It can also lead to false negatives, if the effects on gating offset changes in single-channel conductance. Using single-channel recording as an electrophysiological assay, Kronengold et al. [41] could distinguish effects on i from effects on Po. A change in i is expected for a residue on the pore and the change, as described in the previous section, should depend on the thiol reagent and the resulting side chain. However, a change in i could also occur allosterically, that is, modification of a residue accessible from the bath that leads to a conformational change in the pore. Here, sidedness of accessibility provides a means of distinguishing such a residue from one that is in the pore; the latter should be accessible from both sides in an open channel. Excised patches containing hemichannels allow examination of sidedness of reactivity. A caveat is whether gating translocates a residue from one side of the membrane to the other as described for a number of ion channels. Kronengold et al. [41] combined sidedness of reactivity with single-channel recording, which allowed monitoring the state of the channel during modification. For gap junction channels, the cut-open oocyte technique utilized by Skerrett et al. [44] provides a means of addressing some of these same issues through the use of heterotypic gap junction channels where one hemichannel contains the substituted cysteine site and the other does not. Reagent added to the unsubstituted side should be able to access the cysteine-substituted

3 Pore-Lining Segments and Residues

95

residue in the other hemichannel only by traversing the pore. Skerrett et al. reported that cysteine-substituted mutant channels were generally studied using both homotypic and heterotypic pairs and yielded the same results. In the example given, MBB applied to either side of a homotypic L144C pair or the wild-type side of a heterotypic pair reduced conductance by similar amounts. A related issue is state-dependent accessibility. Skerrett et al. [44] reported that M1, which had sites reported to be accessible to modification in Cx46 hemichannels, was accessible in Cx32 gap junction channels, but only in the closed state, not the open state. Identified sites in M1 of Cx32 modified in the closed state were I30C, F31C, M34C, and V35C. Each of these displayed a ‘‘reverse gating’’ phenotype characterized by no measurable currents in homotypic channels and activation with positive transjunctional voltages relative to the mutant side when paired heterotypically with a wild-type hemichannel. Although largely closed in the absence of a transjunctional voltage, the channels are presumably gating and sampling the open state, particularly over a course of 30 minute exposure. Thus, state-dependence is strictly not tested by examining macroscopic currents, although an argument can be made using kinetics of accessibility if time spent in open versus closed states is sufficiently disparate. Although technically demanding, state-dependence of accessibility can be directly examined in single channels. Kronengold et al. [41] demonstrated access to modification of E1/M1 residues in the open state, but did not examine access when the hemichannels were closed.

3.4.5 Endogenous Cysteines A central assumption of SCAM is that of a clean background, that is, the thiolreactive reagents used should have no effect on the wild-type channel and the cysteine-substituted mutants should retain wild-type or near wild-type function. In the case of Cx46 hemichannels and Cx32 gap junction channels, MBB and MTS were reported to have no obvious electrophysiological effects on the wildtype channels, despite the presence of a number of native cysteine residues. All connexins have six conserved cysteine residues, three in each extracellular loop, which are believed to form disulfide bonds important for gap junction channel formation [60,61]. Thus, cysteine residues in the extracellular loop domains may be unavailable for reaction. However, a cysteine-less mutant of Cx43 was reported to function as a hemichannel; assayed by dye uptake, gap junction channel function was abolished [62]. Therefore, it may be that the cysteine residues in the extracellular loops are important for docking, but not for channel function, per se. If the absence of cysteine residues in the extracellular loops indeed does not affect hemichannel function, then the lack of effect of thiolreactive reagents could still signify unavailability, that is, that they are involved in disulfide bonds, or that they are available and react with no perceivable effect—a silent reaction. The same holds for other native cysteine residues that are scattered in Cx32 and Cx46, mostly in the cytoplasmic CT, although both connexins have a cysteine in M4, C201 in Cx32, and C218 in Cx46.

96

V.K. Verselis

The presence of native cysteine residues is not problematic unless a substituted cysteine renders one of the native cysteine residues reactive, due to structural alterations caused by the substitution, and its modification causes an electrophysiological effect. If that were the case, however, the cysteine substitution alone would likely produce substantial effects on channel function, such as gating and permeability. In the Cx46 hemichannel SCAM study, no substantial effects on channel function were reported for the reactive cysteine substitutions [41]. Also, charged substitutions at these positions that electrostatically mimicked the modifications with MTS reagents produced similar effects, indicating that the substituted cysteine was the one that was modified. Likewise, most of the cysteine substitutions in the Cx32 SCAM study produced channels with near wild-type gating behavior. However, a number were reported with severe alterations in gating, the so-called reverse-gating mutants [44]. The SCAM results of such mutants should be interpreted with caution as they do not meet the general criterion that cysteine-substituted mutants retain near wild-type function. Altered function, as indicated, likely reflects altered conformation that could signify that the cysteine side chain is not located in its normal environment.

3.5 The Connexin Pore: Where Do We Stand? Assignment of the major pore-lining helix in connexin channels remains in disagreement. Connexin-specific differences and hemichannel versus gap junction channel conformations do not appear to be viable explanations. One view, based largely on SCAM of Cx46 hemichannels, designates a segment from L35 to D51, which spans from about midway in M1 and extends into E1 10 residues from the M1/E1 border (Fig. 3.4a). If the pore proper extends sufficiently beyond D51, it is natural to expect that E1 would continue to line the pore, but this has not yet been examined. The structures of the extracellular loops may be sufficiently flexible to allow residues far along E1 or in E2 to contribute. In the other direction, the cytoplasmic half of M1 was sensitive to cysteine substitution, which precluded unambiguous assignment by SCAM. These data suggest that, perhaps, M1 comes into contact with another domain, possibly even deviating away from the pore allowing another domain to take over as a pore-lining segment. The other view, based on SCAM of Cx32 gap junction channels using the cut-open paired oocyte technique, designates a segment from Y135 to Y151, which essentially spans the length of M3, and a small segment of M2 at the cytoplasmic end (Fig. 3.4b). The pattern of accessibility at the extracellular end of the identified segment indicates a transition to a b-sheet and suggests a transition into E2; the E1 and E2 loops have been proposed to adopt an antiparallel b-sheet conformation [61]. Thus, E2 would be a natural candidate to continue as a pore-lining domain, although this has not been examined. A point of agreement between SCAM studies of Cx46 and Cx32 is that the cytoplasmic half of M1 is sensitive to mutagenesis.

3 Pore-Lining Segments and Residues

97

Fig. 3.4. Two current views of the connexin pore based on SCAM studies. (A) View derived from Cx46 hemichannel SCAM studies. Black depicts domains contributing to the pore, which includes the extracellular half of M1 and E1 extending extracellularly. The outermost residue D51 is 8 amino acids from the putative membrane boundary at E43. Pore-lining residues cytoplasmic to L35 were not identified. Based on biophysical studies, the NT is presumed to form the cytoplasmic vestibule. (B) View derived from Cx32 gap junction channel SCAM studies. Black depicts the entirety of M3 contributing to the pore and dark gray depicts a short segment of M2 contributing to the pore at the cytoplasmic end. The outermost residue Y151 is at the M3/E2 border. Pore-lining residues extracellular to Y151 were not identified. As in (A), NTs are presumed to form the cytoplasmic vestibule. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com).

Do other mutagenesis experiments support either one of these assignments? Point mutations of residues scattered throughout the connexin molecule have shown affects on unitary conductance, but many of these must be interpreted with caution as they may be due to indirect structural alterations. Studies using domain exchange generally have shown that the amino-terminal half of the connexin molecule (NT through the CL connecting M2 to M3) and particularly the NT, M1, and E1 domains, largely account for the differences in unitary conductance of channels formed of the various members of the connexin gene family. Reciprocity supports the contention that a significant part of the conduction pathway is exchanged and is in agreement with the view that M1 and E1 contribute to the pore.

98

V.K. Verselis

Regardless of these views, there is close to a consensus based on multiple lines of evidence, albeit indirect, that the NT contributes to the cytoplasmic vestibule of the pore. Confirmation by SCAM has not yet come and the structure of the cytoplasmic region of the channel remains unresolved in the cryoEM studies indicative of a less ordered or flexible domain; the latter is suggested by NMR and gating studies. The NT domain, 22 residues long, is suggested to bend so that the amino terminus loops back toward the membrane. How far it loops back and whether it extends part way into the membrane is not known. Extension into the membrane could mean that no transmembrane helix would contribute to the pore along its entire length. How do the biophysical studies fit with putative helical assignments in the cryoEM structure? No set of structure-function data is in total agreement with the predicted assignments to the pore. In the cryoEM map of Cx43, M3 was designated as the major pore-lining helix with some contribution of M1 toward the extracellular end. This assignment accommodates some identified residues in the contrasting studies, namely V139 and A147 in M3 in the case of SCAM studies in Cx32 gap junction channels, and M1 residues A35 and A39 in the case of SCAM studies in Cx46 hemichannels. In the electron crystallographic structure of Cx26, a cytoplasmic protrusion was suggested to possibly represent the NT because it caps the tops of helices B and C; these helices are equivalent to those designated in the Cx43 structure [14]. However, resolution remains insufficient to unequivocally assign the helices to specific sequences. A central issue, however, is the state of the channel represented in any of the electron crystallographic structures, specifically whether any native or functional state had actually been captured. Here lessons can be learned from crystallographic studies of Kþ channels. Although the features of the pore in the crystal structure of the Kv1.2 channel, particularly the narrow selectivity filter, are in good agreement with biophysical studies, structural features associated with gating are more complex and not so easily mapped to the structure. Gating models include various movements of the S4 voltage sensor and biophysical data regarding movement exhibit fundamentally different views, with movements ranging from as little as 2 A˚ to as much as 20 A˚ (reviewed in [20]). Relating these discrepancies to the crystal structure depends on the state of the channel, which as previously indicated, can adopt various physiological states, such as open, closed, or inactivated, as well as nonphysiological states possibly caused by crystallization techniques. In Kþ channels, the VSDs and the pore-forming domains are separate, and opening of the pore may involve more subtle structural movements than occur in the VSD. In gap junctions, however, the pore and gating are intimately associated, which means that pore structure may change significantly with the state of the channel. This means that cysteine mutagenesis of any given porelining residue that points into an aqueous environment need not result in a channel with near wild-type properties. Gating that involves the pore may mean that some of these residues rotate into a different environment, which can lead to altered or loss of function upon substitution. To occlude a large pore such as

3 Pore-Lining Segments and Residues

99

a connexin channel may involve fairly large movements and helix rotations. In this regard, gap junction channels may have more in common with mechanosensitive channels belonging to the MscL family. In the crystal structure of a MscL homolog from Mycobacterium tuberculosis [63], the pore is generally wide and has a strongly polar character except for some hydrophobic residues at the cytoplasmic end where the pore constricts. Interestingly, mutations of residues that face the lumen of the pore also significantly alter gating, which appears to involve a structural rearrangement such that pore-lining helices are pulled away from each other, thereby relaxing the constriction. There are similarities of the MscL pore and the putative gating mechanism with that of the acetylcholine receptor, another channel that has a fairly large pore [64]. In connexins, a complicating factor is the presence of two gating mechanisms that appear to adopt substantially different conformations, one leading to a closed state that still conducts ions and the other a nonconducting closed state. Thus, connexin channel pores may contain two regions that form constrictions and undergo more complex structural changes.

3.6 Conclusion Although there is considerable heterogeneity in conductance and selectivity among connexin channels, biophysical studies have shown a number of common properties indicative of an overall conservation in structure-function. The expectation is that the principal pore-lining helices will turn out to be the same in all connexin channels, and differences in specific residues, that either point into the pore lumen itself or contact other helices, will provide sufficient perturbations in pore structure to account for the connexin-specific differences in conductance and selectivity. Of interest is whether connexin channels possess a distinct region that serves as a molecular filter that sorts among the various potential biological signaling molecules, whether by size, charge, or some hydrophobic interaction. Electron crystallographic structures of gap junction channels formed of different connexins show similar positioning of helices in the transmembrane span and support the view that there is, indeed, strong structural conservation among even distant members of the connexin gene family. Which physiological state, if any, the structures represent remains unknown. Putative helical assignments to the structure have been made, but are not in agreement with all the available biophysical data. The value of biophysical studies is their independence from any crystal structure or model. Reconciliation between the current views of the pore will come by continued biophysical studies concomitant with higher resolution structures crystallized in various states. Acknowledgments The author would like to thank Drs. Miduturu Srinivas, Thaddeus Bargiello, and Myles Akabas for helpful discussions.

100

V.K. Verselis

References 1. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys. 2001;34:325–472. 2. Alexander DB, Goldberg GS. Transfer of biologically important molecules between cells through gap junction channels. Curr Med Chem. 2003;10:2045–58. 3. Neijssen J, Herberts C, Drijfhout JW, Reits E, Janssen L, Neefjes J. Cross-presentation by intercellular peptide transfer through gap junctions. Nature. 2005;434:83–8. 4. Valiunas V, Polosina YY, Miller H, Potapova IA, Valiuniene L, Doronin S, Mathias RT, Robinson RB, Rosen MR, Cohen IS, Brink PR. Connexin-specific cell-to-cell transfer of short interfering RNA by gap junctions. J Physiol. 2005;568:459–68. 5. Goldberg GS, Valiunas V, Brink PR. Selective permeability of gap junction channels. Biochim Biophys Acta. 2004;1662:96–101. 6. Weber PA, Chang HC, Spaeth KE, Nitsche JM, Nicholson BJ. The permeability of gap junction channels to probes of different size is dependent on connexin composition and permeant-pore affinities. Biophys J. 2004;87:958–73. 7. Bevans CG, Kordel M, Rhee SK, Harris AL. Isoform composition of connexin channels determines selectivity among second messengers and uncharged molecules. J Biol Chem. 1998;273:2808–16. 8. Ayad WA, Locke D, Koreen IV, Harris AL. Heteromeric, but not homomeric, connexin channels are selectively permeable to inositol phosphates. J Biol Chem. 2006;281:16727–39. 9. Doyle DA, Morais Cabral J, Pfuetzner RA, Kuo A, Gulbis JM, Cohen SL, Chait BT, MacKinnon R. The structure of the potassium channel: molecular basis of K+ conduction and selectivity. Science. 1998;280:69–77. 10. Jiang Y, Lee A, Chen J, Ruta V, Cadene M, Chait BT, MacKinnon R. X-ray structure of a voltage-dependent Kþ channel. Nature. 2003;423:33–41. 11. Long SB, Campbell EB, Mackinnon R. Crystal structure of a mammalian voltagedependent Shaker family Kþ channel. Science. 2005;309:897–903. 12. Noskov SY, Roux B. Ion selectivity in potassium channels. Biophys Chem. 2006;124:279–91. 13. Unger VM, Kumar NM, Gilula NB, Yeager M. Three-dimensional structure of a recombinant gap junction membrane channel. Science. 1999;283:1176–80. 14. Oshima A, Tani K, Hiroaki Y, Fujiyoshi Y, Sosinsky GE. Three-dimensional structure of a human connexin26 gap junction channel reveals a plug in the vestibule. Proc Natl Acad Sci USA. 2007;104:10034–9 15. Ghoshroy S, Goodenough DA, Sosinsky GE. Preparation, characterization, and structure of half gap junctional layers split with urea and EGTA. J Membr Biol. 1995;146:15–28. 16. Mu¨ller DJ, Hand GM, Engel A, Sosinsky GE. Conformational changes in surface structures of isolated connexin 26 gap junctions. EMBO J. 2002;21:3598–607. 17. Thimm J, Mechler A, Lin H, Rhee S, Lal R. Calcium-dependent open/closed conformations and interfacial energy maps of reconstituted hemichannels. J Biol Chem. 2005;280:10646–54. 18. Fleishman SJ, Unger VM, Yeager M, Ben-Tal N. A C- model for the transmembrane helices of gap junction intercellular channels. Mol Cell. 2004;15:879–88. 19. Long SB, Campbell EB, Mackinnon R. Voltage sensor of Kv1.2: structural basis of electromechanical coupling. Science. 2005;309:903–8. 20. Tombola F, Pathak MM, Isacoff EY. How does voltage open an ion channel? Annu Rev Cell Dev Biol. 2006;22:23–52. 21. Spray DC, Harris AL, Bennett MVL. Equilibrium properties of a voltage-dependent junctional conductance. J Gen Physiol. 1981;77:77–93. 22. Bukauskas FF, Verselis VK. Gap junction channel gating. Biochim Biophys Acta. 2004;1662:42–60.

3 Pore-Lining Segments and Residues

101

23. Harris AL, Spray DC, Bennett MVL. Kinetic properties of a voltage-dependent junctional conductance. J Gen Physiol. 1981;77:95–117. 24. Verselis VK, Bennett MVL, Bargiello TA. A voltage-dependent gap junction in Drosophila melanogaster. Biophys J. 1991;59:114–26. 25. Weingart R, Bukauskas FF. Gap junction channels of insects exhibit a residual conductance. Pflu¨gers Arch. 1993;424:192–4. 26. Moreno AP, Rook MB, Fishman GI, Spray DC. Gap junction channels: distinct voltagesensitive and – insensitive conductance states. Biophys J. 1994;67:113–9. 27. Trexler EB, Bennett MVL, Bargiello TA, Verselis VK. Voltage-gating and permeation in a gap junction hemichannel. Proc Natl Acad Sci USA. 1996;93:5836–41. 28. Verselis VK, Ginter CS, Bargiello TA. Opposite voltage-gating polarities of two closely related connexins. Nature. 1994;368:348–51. 29. Oh S, Abrams CK, Verselis VK, Bargiello TA. Stoichiometry of transjunctional voltagegating polarity reversal by a negative charge substitution in the amino terminus of a connexin32 chimera. J Gen Physiol. 2000;116:13–31. 30. Purnick PE, Benjamin DC, Verselis VK, Bargiello TA, Dowd TL. Structure of the amino terminus of a gap junction protein. Arch Biochem Biophys. 2000;381:181–90. 31. Purnick PE, Oh S, Abrams CK, Verselis VK, Bargiello TA. Reversal of the gating polarity of gap junctions by negative charge substitutions in the N-terminus of connexin 32. Biophys J. 2000;79:2403–15. 32. Oh S, Rivkin S, Tang Q, Verselis VK, Bargiello TA. Determinants of gating polarity of a connexin 32 hemichannel. Biophys J. 2004;87:912–28. 33. Musa H, Fenn E, Crye M, Gemel J, Beyer EC, Veenstra RD. Amino terminal glutamate residues confer spermine sensitivity and affect voltage-gating and channel conductance of rat connexin40 gap junctions. J Physiol. 2004;557:863–78. 34. Tong JJ, Ebihara L. Structural determinants for the differences in voltage-gating of chicken Cx56 and Cx45.6 gap-junctional hemichannels. Biophys J. 2006;91:2142–54. 35. Pfahnl A, Zhou XW, Werner R, Dahl G. A chimeric connexin forming gap junction hemichannels. Pflu¨gers Arch. 1997;433:773–9. 36. Trexler EB, Bukauskas FF, Kronengold J, Bargiello TA, Verselis VK. The first extracellular loop domain is a major determinant of charge selectivity in connexin46 channels. Biophys J. 2000;79:3036–51. 37. Musa H, Veenstra RD. Voltage-dependent blockade of connexin40 gap junctions by spermine. Biophys J. 2003;84:205–19. 38. Hu X, Dahl G. Exchange of conductance and gating properties between gap junction hemichannels. FEBS Lett. 1999;45:113–7. 39. Hu X, Ma M, Dahl G. Conductance of connexin hemichannels segregates with the first transmembrane segment. Biophys J. 2006;90:140–50. 40. Zhou XW, Pfahnl A, Werner R, Hudder A, Llanes A, Luebke A, Dahl G. Identification of a pore-lining segment in gap junction hemichannels. Biophys J. 1997; 72:1946–53. 41. Kronengold J, Trexler EB, Bukauskas FF, Bargiello TA, Verselis VK. Single-channel SCAM identifies pore-lining residues in the first extracellular loop and first transmembrane domains of Cx46 hemichannels. J Gen Physiol. 2003;122:389–405. 42. Karlin A, Akabas MH. Substituted-cysteine accessibility method. Meth Enzymol. 1998;293:123–45. 43. Pfahnl A, Dahl G. Localization of a voltage gate in connexin46 gap junction hemichannels. Biophys J. 1998;75:2323–31. 44. Skerrett IM, Aronowitz J, Shin JH, Cymes G, Kasperek E, Cao FL, Nicholson BJ. Identification of amino acid residues lining the pore of a gap junction channel. J Cell Biol. 2002;159:349–60. 45. Bennett MVL, Zheng X, Sogin ML. The connexins and their family tree. Soc Gen Physiol Ser. 1994;49:223–33.

102

V.K. Verselis

46. Cruciani V, Mikalsen SO. Evolutionary selection pressure and family relationships among connexin genes. Biol Chem. 2007;388:253–64. 47. Valiunas V, Weingart R. Electrical properties of gap junction hemichannels identified in transfected HeLa cells. Pflu¨gers Arch. 2000;440:366–79. 48. Bukauskas FF, Kreuzberg MM, Rackauskas M, Bukauskiene A, Bennett MV, Verselis VK, Willecke K. Properties of mouse connexin 30.2 and human connexin 31.9 hemichannels: implications for atrioventricular conduction in the heart. Proc Natl Acad Sci USA. 2006;103:9726–31. 49. Beahm DL, Hall JE. Hemichannel and junctional properties of connexin 50. Biophys J. 2002;82:2016–31. 50. Srinivas M, Kronengold J, Bukauskas FF, Bargiello TA, Verselis VK. Correlative studies of gating in Cx46 and Cx50 hemichannels and gap junction channels. Biophys J. 2005;88:1725–39. 51. Bukauskas FF, Bukauskiene A, Bennett MVL, Verselis VK. Gating properties of gap junction channels assembled from connexin43 and connexin43 fused with green fluorescent protein. Biophys J. 2001;81:137–52. 52. Contreras JE, Sa´ez JC, Bukauskas FF, Bennett MVL. Functioning of cx43 hemichannels demonstrated by single channel properties. Cell Commun Adhes. 2003;10:245–9. 53. Trexler EB, Bukauskas FF, Bennett MVL, Bargiello TA, Verselis VK. Rapid and direct effects of pH on connexins revealed by the connexin46 hemichannel preparation. J Gen Physiol. 1999;113:721–42. 54. Ebihara L, Steiner E. Properties of a nonjunctional current expressed from a rat connexin46 cDNA in Xenopus oocytes. J Gen Physiol. 1993;102:59–74. 55. Srinivas M, Calderon DP, Kronengold J, Verselis VK. Regulation of connexin hemichannels by monovalent cations. J Gen Physiol. 2006;127:67–75. 56. Valiunas V. Biophysical properties of connexin-45 gap junction hemichannels studied in vertebrate cells. J Gen Physiol. 2002;119:147–64. 57. Kaneko S, Akaike A, Satoh M. Cut-open recording techniques. Meth Enzymol. 1998;293:319–31. 58. Bera AK, Chatav M, Akabas MH. GABA(A) receptor M2-M3 loop secondary structure and changes in accessibility during channel gating. J Biol Chem. 2002;277:43002–10. 59. Kronengold J, Trexler EB, Bukauskas FF, Bargiello TA, Verselis VK. Pore-lining residues identified by single channel SCAM studies in Cx46 hemichannels. Cell Commun Adhes. 2003;10:193–9. 60. Dahl G, Levine E, Rabadan Diehl C, Werner R. Cell/cell channel formation involves disulfide exchange. Eur J Biochem. 1991;197:141–4. 61. Foote CI, Zhou L, Zhu X, Nicholson BJ. The pattern of disulfide linkages in the extracellular loop regions of connexin 32 suggests a model for the docking interface of gap junctions. J Cell Biol. 1998;140:1187–97. 62. Bao X, Chen Y, Reuss L, Altenberg GA. Functional expression in Xenopus oocytes of gap-junctional hemichannels formed by a cysteine-less connexin 43. J Biol Chem. 2004;279:9689–92. 63. Chang G, Spencer RH, Lee AT, Barclay MT, Rees DC. Structure of the MscL homolog from Mycobacterium tuberculosis: a gated mechanosensitive ion channel. Science. 1998;282:2220–6. 64. Miyazawa A, Fujiyoshi Y, Unwin N. Structure and gating mechanism of the acetylcholine receptor pore. Nature. 2003;423:949–55.

Chapter 4

Voltage-Gating Mechanisms of Connexin Channels Thaddeus Bargiello and Peter Brink

Abstract Channels formed by the connexin family of proteins display multiple forms of voltage dependence that have different sensitivities and time courses. Intercellular (junctional) channels are sensitive to two orientations of applied voltage: inside-out or transmembrane voltage (Vi-o or Vm), and transjunctional voltage (Vj). At least two voltage-gating mechanisms operate in both intercellular channels and in unapposed hemichannels. The Vj or fast-gating mechanism is sensitive to Vj. The loop or slow-gating mechanism is also sensitive to Vj and may also be sensitive to Vi–o/Vm and to channel closure by chemical agents. Both types of voltage-gating are intrinsic to hemichannels; in an intercellular channel, each apposed hemichannel contains separate gating structures arranged in series. The molecular determinants and the mechanism of polarity determination of Vj/fast-gating have been studied extensively for homomeric Cx26 and Cx32 hemichannels. These studies have shown that difference in gating polarity of Cx26 and Cx32 hemichannels results from a difference in the charge of the second amino acid residue. The voltage polarity to which Cx32 hemichannels close can be reversed by negative charge substitutions up to the tenth but not the eleventh residue. This has led to a structural model in which the first ten amino acid residues of the amino-terminal domain form the entry of the channel pore by virtue of a turn initiated by the flexibility of a glycine residue at the twelveth position. This model is supported by nuclear magnetic resonance (NMR)-derived structures of a peptide of the amino-terminal domain of Cx26 and permeation studies demonstrating that charged residues in the amino-terminal domain are determinants of unitary conductance and contribute to the rectification of current through the open channel. Initially, it was proposed that the inward movement of the charges in the amino-terminal domain initiates Vj/fast-gating. This view was complicated by the discovery of homomeric channels that display bipolar Vj/fast-gating, leading to the T. Bargiello (*) Dominick P. Purpura Department of Neuroscience, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_4, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

103

104

T. Bargiello and P. Brink

possibility that the voltage sensor functions as a center-open toggle switch. The subsequent conformational changes that result in channel closure to a substate may involve the actions of a proline kink in the second transmembrane domain, an interaction between the cytoplasmic loop and carboxyl-terminal domain, or could result from the movement of the amino-terminal domain as a gating particle. Less is known about the molecular determinants and mechanisms of loop/slow-gating. The process is mechanistically distinct from Vj/fast-gating, in at least its initiation. The two processes may share structural elements, although it is unlikely that the conformational changes with gating will be identical. Keywords Voltage-gating Voltage sensors Cx26 Cx32 Cx31.9 Cx37 Cx43 Cx46 Cx50

4.1 Introduction Gap junctions are aggregates of intercellular channels formed by the head-tohead union of two hemichannels, each spanning a plasma membrane of closely apposed cells. This distinctive structure allows sensitivity to two different voltages: (1) the voltage between the cytoplasm and the extracellular space, the inside-outside voltage, termed Vi-o or Vm (Fig. 4.1a); and (2) the voltage

Fig. 4.1. Diagram of a gap junction channel showing presumed isopotential lines resulting from the application of Vi–o and Vj. (a) Vi–o was established by hyperpolarizing both cells to a common holding potential of –50 mV. There is no Vj in this case. (b) Vj was established by hyperpolarizing the left cell to a holding potential of –100 mV while holding the right cell at 0 mV. Vi–o is also generated in regions of the channel with this voltage paradigm. The same Vj could be generated by holding the left cell at –50 mV and the right cell at +50 mV, that is, the voltage difference across the channel is –100 mV relative to the left cell; however, Vi–o would differ in this case. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

4 Voltage-Gating Mechanisms

105

difference between the cytoplasms of the two cells, the transjunctional voltage, termed Vj (Fig. 4.1b). Operationally, sensitivity to Vi-o is determined by imposing a series of equal voltages simultaneously in two coupled cells, so that Vj remains zero, and monitoring the junctional currents [1,2]. Sensitivity to Vj is determined by imposing Vj differences by making the voltages of the two cells unequal. This imposes a Vi-o gradient as well, but the magnitude of Vi-o depends on the potential of each cell, whereas Vj depends solely on the difference in membrane potentials. If the intercellular channels are sensitive to Vj only, the junctional currents will be mirror images of each other for Vj steps of equal amplitude but opposite sign. All gap junction channels encoded by the connexin gene family display various degrees of sensitivity to Vj, and some also display weak sensitivity to Vi–o. In addition to forming intercellular channels, hemichannels formed by several connexins, notably Cx46, Cx50, human CX26, and a Cx32 chimera in which the first extracellular loop is that of Cx43 (Cx32*43E1), form conductive channels even when they are not apposed by another hemichannel. In unapposed hemichannels, Vj is equivalent to Vi–o, and consequently, the two voltagedependent processes are elicited simultaneously by a change in membrane potential. Conductive hemichannels are more accessible to electrophysiological recording than intercellular channels and consequently provide a means to study the molecular mechanisms of voltage dependence at both the macroscopic and single-channel levels.

4.2 Structure-Function Relationships for Vi-o and Vj Sensitivities: Where Are the Voltage Sensors? The structure-function requirement of Vi–o sensitivity of an intercellular channel is that the voltage sensor (defined as the set of charged amino acids that move in response to a change in an applied electric field) must reside in regions of the channel that experience a change in the electric field due to a change in Vi–o. In a junctional channel the voltage field created by Vi–o falls along the length of the channel pore within the membrane spanning region (Fig. 4.1a). A Vi–o sensor could be located anywhere in this region. However, the most likely location for a Vi–o sensor is near the channel lumen in the vicinity of the extracellular gap, since the electric field due to Vi–o is strongest in this region, that is, where the isopotential lines are most spatially compressed. A Vi–o sensor located in this region would move in a direction perpendicular to the long axis of the channel, across the isopotential lines. For the case where there is also a nonzero Vj (as in Fig. 4.1b), charged residues located in this region would also experience a field generated by Vj (Fig. 4.1b), although it would be oriented perpendicular to that generated by Vi–o (compare Fig. 4.1a and b). Thus, for the channel to be sensitive only to Vi–o, the movement of the voltage sensor must be constrained so that it cannot move in response to the electric field created by Vj along the long axis of the channel.

106

T. Bargiello and P. Brink

In contrast, for intercellular channels displaying Vj sensitivity the voltage sensor would most likely reside in or near the channel lumen close to an intracellular end of the pore, as the electric field created by Vj is much larger than that generated by Vi–o in this region. Because the electric field due to Vj is oriented parallel to the long axis of the channel (Fig. 4.1b), the voltage sensor would likely move into or out of the channel pore, depending on its charge relative to the field. For example, a positively charged voltage sensor in the left hemichannel in Fig. 4.1b would move out of the channel pore (an inward movement with respect to the cytoplasm) when this cell is relatively negative. The possibility that a Vj-sensor could be located more toward the center of the extracellular gap is not precluded, but if it were located in this position its movement would have to be constrained such that it cannot move in response to the orientation of the field created by Vi–o.

4.3 Voltage Dependence of Homotypic and Heterotypic Junctional Channels Voltage-dependent changes in conductance of intercellular channels are exemplified by the conductance-voltage relations (G-V) shown in Fig. 4.2. These relations were obtained from macroscopic records of pairs of Xenopus oocytes expressing rat Cx32 or rat Cx26 in homotypic and heterotypic pairing configurations. In homotypic Cx32 junctions both initial and steady-state currents display voltage-dependent changes in conductance (Fig. 4.2a). These changes are symmetric about Vj = 0 mV and depend solely on Vj. Initial conductance is maximal at Vj = 0 mV and reduced by 30% at large Vj. At large Vjs, the steady-state G-V relation plateaus to a minimal conductance, Gmin, of 0.3. In Cx26 homotypic junctions (Fig. 4.2b), the initial conductance displays some dependence on Vi–o, detected as asymmetry between the effects of depolarizing one cell and hyperpolarizing the other to the same extent [3,4,5]. Junctional currents decrease slowly for either polarity of Vj and conductance decreases to a minimum at higher Vjs, corresponding to a minimal junctional conductance or Gmin of 0.2 at positive Vjs. There is also some asymmetry in the time course of the junctional currents depending on whether one cell is hyperpolarized or depolarized. This asymmetry is reflected in the steady-state G-V relation and is indicative of Vi–o dependence. Heterotypic junctions formed by pairing oocytes expressing Cx32 with those expressing Cx26 are characterized by an asymmetric G-V relation (Fig. 4.2c) that depends on Vj, that is, junctional currents elicited by hyperpolarization or depolarization of either cell are not identical. Evidently, the heterotypic Cx26–Cx32 pairing eliminates the Vi–o sensitivity of the Cx26. These channels have a pronounced rectification of initial currents, with a much greater initial conductance when the cell expressing Cx26 is made relatively positive and much less initial conductance when this cell is made relatively negative. The steady-state

4 Voltage-Gating Mechanisms

107

Fig. 4.2. Representative traces of conductance-voltage relations and initial and steady-state junctional currents obtained in pairs of Xenopus oocytes. (a) Homotypic Cx32-Cx32 junctions. (b) Homotypic Cx26-Cx26 junctions. (c) Heterotypic Cx26-Cx32 junctions. In the current traces, the time at which the initial conductance is measured is denoted by *, the time corresponding to steady-state conductance by #. In the conductance-voltage relations, initial conductance is depicted by t, and steady-state conductance by r. In all cases, positive Vj is relative positivity at the cytoplasmic mouth of the hemichannel appearing on the right side of the channel designation. For example, in the heterotypic junction, the x axis gives the Vj of the Cx26 side of the junction. The conductance-voltage relations of each junction are the average of at least ten separate oocyte pairs. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Adapted from Oh et al. [7] with permission.)

108

T. Bargiello and P. Brink

conductance decreases from the initial conductance only when the Cx26 side of the heterotypic junction is stepped to positive Vjs exceeding 40 mV (Fig. 4.2c). Similar Vj-dependence is seen in heterotypic gap junctions composed of Cx37 and Cx43 [6]. The mechanism and molecular determinants of the rectification of initial conductance of heterotypic Cx26–Cx32 junctions were reported by Oh et al. [7] to involve asymmetry in the flux of permeant ions through the open channel rather than rapid voltage-dependent gating processes. This rectification of currents through open channels results from an asymmetry in the distribution of fixed charges along the channel pore as predicted by the Poisson-NernstPlanck (PNP) equations [8]. This asymmetry is present only in the heterotypic junctional channels, and is thus a consequence of junctional channel formation. The rectification of initial conductance of Cx32 homotypic junctions can also be explained by the presence of fixed charge located near the ends of the intercellular channel, but in this case the charges would be distributed symmetrically and produce a sigmoidal current-voltage (I–V) relation in single-channel records [8]. In contrast, the changes observed in the steady-state conductance almost certainly result from voltage-dependent conformational gating changes. The decrease in steady-state conductance with voltage observed in heterotypic Cx26–Cx32 junctions is a consequence of the presence of separate voltagedependent gates in each hemichannel and their opposite gating polarities. The open probability of the Cx32 hemichannel of the heterotypic Cx26–Cx32 junctional channel decreases when the cytoplasmic face of the Cx32 hemichannel is negative, that is, Cx(–), relative to the Cx26 hemichannel, while the open probability of the Cx26 hemichannel decreases when its intracellular face is positive, that is, Cx(+), relative to the Cx32 hemichannel. The consequences of the head-to-head orientation of hemichannels, each containing separate voltage sensors and gates and opposite gating polarities, are illustrated in Fig. 4.3. Figure 4.3a shows the G-V relations for an idealized case of a homotypic channel that closes upon hyperpolarization of either cell. Figure 4.3b shows the case for a homotypic channel that closes upon depolarization of either cell. Figure 4.3c shows the idealized case for a heterotypic channel composed of the hemichannels illustrated in Fig. 4.3a,b. The solid and dashed lines represent the G-V relations arising from the gating of hemichannels on the right and left sides of the intercellular channel, respectively. Note that Fig. 4.3 depicts the serial arrangement of two gates, one in each hemichannel, which close fully. To account for the existence of a minimal conductance at high Vjs, the gates are thought not to close fully. The G-V relations of homotypic channels are symmetric about Vj = 0 mV, with each oppositely oriented hemichannel closing for a given polarity of Vj. In the case of Cx(–) homotypic junctional channels (Fig. 4.3a), each hemichannel closes when its intracellular face is relatively negative and opens when its intracellular face is relatively positive. The opposite is true for junctional channels in which each hemichannel closes when its intracellular face is positive

4 Voltage-Gating Mechanisms

109

Fig. 4.3 . Diagrammatic representations of gap junction hemichannel open/closed configurations at large Vjs of either sign in which the component hemichannels gate on relative negativity Cx(–) or relative positivity Cx(+). (a) Homotypic gap junction formed by negative gating hemichannels. (b) Homotypic gap junction formed by positive gating hemichannels. (c) Heterotypic channel formed with a negative and positive gating hemichannel. The expected conductance-voltage relations of the component hemichannels (lower panels). Positive values on the x axis refer to the sign of Vj relative to the hemichannel on the right side of any given pairing. The solid line depicts the conductance-voltage relation of the hemichannel on the right side of the pair, the dashed line the conductance voltage relation of the hemichannel on the left side of the pair. (d) Junctional currents elicited from a Cx43 gap junction with the voltage protocol shown in the top panel. Junctional currents increase then decrease to steady-state following the reversal of the voltage from +100 mV to –100 mV illustrating that closed hemichannels must open prior to closing of gates in the apposed hemichannel. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

110

T. Bargiello and P. Brink

(Fig. 4.3a). Since in both cases the G-V relation is similar, the G-V relations of homotypic channels preclude assignment of the Vj gating polarity of the component hemichannels. On the other hand, the G-V relation of the heterotypic Cx(–)–Cx(+) junction is asymmetric because both hemichannels close for the same Vj, which is relatively positive to Cx(+) and relatively negative to Cx(–). Neither hemichannel closes for the opposite polarity of Vj. The gating polarity of the component hemichannels can be inferred in this case, if one assumes that the polarity of voltage sensitivity of each hemichannel does not change with heterotypic pairing (there is no evidence for such a change). It should be stressed that the voltage-dependent changes in conductance described in this chapter do not result from ionic blockade by divalent cations as proposed by Puljung et al. [9] for human CX37 intercellular channels. Voltage-dependent transitions to substates and to fully closed states are observed in single-channel records of intercellular Cx32 and Cx26 channels and of excised patches of unapposed hemichannels under conditions where divalent cations are chelated. In many cases, Ca2+ and other divalent ions can influence voltage-gating by causing shifts in voltage dependence. Unapposed Cx50 hemichannels can also be regulated by monovalent metal cations, apparently by modulating the ability of Ca2+ to close the channel [10].

4.4 Contingent Gating Figure 4.3 illustrates an additional feature of voltage-gating of intercellular channels, termed contingent gating, first described experimentally by Harris et al. [11]. It was observed that the response of one of the hemichannels to a voltage change was contingent upon the status of the gate in the apposed hemichannel. It was proposed that this arose from the voltage sensor of each hemichannel being within the pore lumen. The reasoning is as follows: the closure of a gate in one hemichannel requires that the voltage drop within the pore occur entirely across the closed gate. Consequently, there would be no voltage drop across the voltage sensor in the apposed hemichannel and the voltage gate in this hemichannel would be open. Assuming the open probability of the hemichannel is 1.0, at Vj = 0 mV most homotypic gap junction channels display this property [12,13,14]. If the polarity of Vj is then reversed, the closed hemichannel must open before the voltage sensor in the apposed hemichannel would sense the change in Vj and initiate closure of its hemichannel. With this voltage paradigm, one usually observes an increase in junctional currents from the initial values due to the opening of the closed hemichannels, followed by a reduction in junctional currents due to closure of the voltage gates in the apposed hemichannels. Figure 4.3d illustrates an example for Cx43 gap junctions in which a Vj step was applied for 400 ms and then reversed in polarity. Note that the junctional current after the onset of polarity reversal shows an

4 Voltage-Gating Mechanisms

111

increase in current with a time course of many milliseconds and then declines with a time course that is slower than that elicited by the preceding polarization. The rates and shape of the junctional currents elicited by polarity reversal depend on the relative rates of opening and closure of the voltage gates in the two hemichannels. A major implication of contingent gating is that each hemichannel contains separate gating elements arranged in series. That is, the voltage-gating elements are not formed by the union of two hemichannels. Consequently, voltage-gating is a property intrinsic to hemichannels. This view was confirmed by studies of voltage-gating of unapposed hemichannels described later in this chapter. Contingent gating does not require that the actions of the hemichannel gates be independent, only that the gate of each hemichannel respond individually to the sensed voltage, as it is likely that interactions among the extracellular loops of two linked hemichannels that form a gap junction channel will change the conformation and hence the free energy of the conductance states of each hemichannel. It should also be noted that the field experienced by the Vj voltage sensor in a hemichannel depends on the electrical properties of the apposed hemichannel. For example, in heterotypic pairings of hemichannels with different electrical conductances, the voltage drop would be smaller across the hemichannel with greater conductance [15] resulting in an apparent shift in the voltage dependence of the gate in a given hemichannel.

4.5 Multiplicity of Vj-Gating Mechanisms Segments of a record of a single homomeric Cx32 intercellular channel expressed in transfected Neuro2A cells are presented in Fig. 4.4. The traces illustrate the presence of two different gating mechanisms, both of which are elicited only by Vj in these channels. The upper traces (Fig. 4.4a) illustrate transitions between the fully open state and several subconductance states, while the lower traces (Fig. 4.4b) illustrate gating that involves a series of small transitions between a fully open and a fully closed state that give the appearance of a slow event (positions 2 and 3 — see [16] — the partial closures of the channel observed between positions 1 and 2 in Fig. 4.4b likely correspond to the slow-gating mechanism that did not result in complete channel closure). Although not demonstrated in this record, complete closure of the channel by the slow mechanism can also occur from the subconductance states. This suggests that the two processes arise from different mechanisms, as closure to a substate does not preclude subsequent full closure. The term loop-gating arose from the suggestion that the slow transitions from the fully open to fully closed state involve the extracellular loops of the Cx46 hemichannel since the transitions are similar to those observed during initial hemichannel docking and formation [17,18]. Others term this slow-gating, reflecting the slower time course of the set of transitions. Owing to the

112

T. Bargiello and P. Brink

Fig. 4.4. Single-channel records of rat homotypic Cx32 expressed in pairs of Neuro2A cells. (a) A segment of a continuous record of junctional currents elicited by a Vj of +60 mV applied at position 1 in the record. Numerals 2 and 3 illustrate brief and long-lived closing to a substate. (b) A segment of the same single-channel record shown in (a), but after the application of Vj of –60 mV at position 1. All gating transitions shown in this record correspond to loop/slow-gating events. In other records, Vj/fast-gating transitions to substates comparable to that shown in (a) predominate. The traces shown are digitally filtered at 250 Hz for presentation. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Adapted from Oh et al. [7] with permission.)

similar appearance and inferred polarity of closure, it is generally assumed that the same Vj-dependent mechanism underlies loop/slow-gating transitions observed in unapposed hemichannels and intercellular channels, but this assumption needs to be verified. The relationships between loop/slow-gating transitions seen in response to Vj and the similar transitions initiated by Vi–o or that result from chemical gating have also not yet been established. Revilla et al. [19] have suggested that the Vi–o gate in Cx43 junctions differs from both the Vj/ fast-gate and the loop/slow-gate. In contrast, the transitions from the fully open state to substates for junctional channels (Fig. 4.4a) occur on a faster time scale (submillisecond) that cannot be resolved by the voltage clamp. This process was termed Vj-gating because the entry into substates can account for the Gmin seen in macroscopic recordings of most gap junction channels that are sensitive to Vj. Others use the term fast-gating or substate-gating to describe this mechanism. It should be noted that in macroscopic recordings, such as those of Cx32 shown in Fig. 4.2a, junctional conductance may decrease from the plateau if Vjs greater than 120 mV are applied. Presumably, this decrease in junctional conductance

4 Voltage-Gating Mechanisms

113

represents closure of loop/slow-gates and suggests that loop/slow-gating is less sensitive to voltage than Vj/fast-gating in Cx32 intercellular channels. Both loop/slow-gating and Vj/fast-gating transitions are also observed in single-channel records of Cx26. The single-channel current-voltage relation is linear and junctional currents are the same at large potentials for either polarity of Vj [7]. The reason for the lack of evidence for Vi–o dependence of initial currents at the single-channel level is unknown. In records of heterotypic Cx26–Cx32 single intercellular channels, transitions to substates (that is, Vj/fast transitions) are seen only when the Cx26 side of the channel is relatively positive. This closure corresponds to the reduction in junctional conductance observed in macroscopic recordings (Fig. 4.2c). Loop/ slow-gating transitions are not observed at either polarity of applied Vj in Cx26–Cx32 channels. This result is expected, as the initial and steady-state conductances are the same when the Cx26 side of the junction is relatively negative. Taken together, these observations suggested that the difference in the gating polarity of Cx32 and Cx26 hemichannels is due to the difference in the polarity of the Vj/fast-gate. As previously mentioned, this interpretation assumes that the Vj/fast-gates are operational in both Cx32 and Cx26 hemichannels when they are paired heterotypically. Thus, the simplest model is that the Vj/fast-gate in Cx26 hemichannels closes when the intracellular surface is positive, while the Vj/fast-gate of Cx32 hemichannels closes when the intracellular surface is negative, as illustrated in Fig. 4.3c. The reversal of the polarity of Vj/fast-gates by negative charge substitutions in Cx32 was confirmed by single-channel records of unapposed hemichannels (see below). The presence of loop/slow-gating transitions with either polarity of Vj in homotypic Cx32-Cx32 and Cx26-Cx26 intercellular channels preclude the assignment of a gating polarity to this process. The polarity of the loop/slow-gate was first established in studies of unapposed Cx46 hemichannels. In these channels the polarity of the Vj/fast-gate is positive, as for Cx26, while the loop/slow gate has negative polarity [17]. The polarity of loop/ slow-gating is negative in all unapposed channels examined to date. It should be noted that the determination of Vj/fast-gating polarity from macroscopic recordings can be complicated by the contribution of loop/slowgating to the observed current relaxations. Heterotypic channels formed by the union of hemichannels that display moderately sensitive loop/slow-gates closing with the same polarity are expected to show current relaxations for both polarities of applied Vj irrespective of the polarity of their Vj/fast-gates. Definitive assignment of polarity to Vj/fast-gating and loop/slow-gating requires single-channel recordings, preferably from unapposed hemichannels, as in this configuration the two gating mechanisms can be distinguished. In fact, Verselis et al. [20] were fortunate in their original analyses of Cx32 and Cx26 junctions in that the loop/slow-gates of heterotypic Cx26–Cx32 channels displayed weak or no sensitivity to Vj at the voltages examined. The existence of two separate gating mechanisms was not known at that time.

114

T. Bargiello and P. Brink

Gating polarity was ultimately established by single-channel studies of wildtype and mutant unapposed Cx32*43E1 hemichannels.

4.6 Molecular Determinants and Mechanisms of Vj/Fast-Gating The apparent difference in the polarity of Vj sensitivity of Cx32 and Cx26 hemichannels provided a means to probe the molecular determinants and mechanism of Vj/fast-gating. Following extensive analyses of the voltage dependence of chimeras of Cx26 and Cx32 in heterotypic and homotypic pairing configurations, the difference in Vj/fast-gating polarity was localized to a difference in the charge of the second amino acid residue located in the amino-terminal domain (NT) [20]. In Cx26, this residue is the positively charged aspartic acid, while in Cx32 it is the neutral amino acid asparagine. Negative charge substitutions of the asparagine residue in Cx32 (Cx32N2E or Cx32N2D) reversed the polarity of Cx32 hemichannel gating from negative to positive, while the substitution of positive and neutral amino acids reversed the gating polarity of Cx26 hemichannels in macroscopic recordings of heterotypic channels. Verselis et al. [20] proposed that charged residues in the NT of Cx32 and Cx26 formed a least part of a voltage sensor, net negative in Cx26 and net positive in Cx32, such that the application of a Vj, positive to the cytoplasmic face of Cx26 and negative to the cytoplasmic face of Cx32 would result in the movement of at least part of the NT toward the cytoplasm. It was suggested that the amino-terminal methionine residue contributed to the positive charge of this domain. The hypothesized gating mechanism was attractive in that it localized the Vj-sensor to a position where only an electric field resulting from Vj would develop (Fig. 4.1) and the conformational changes leading to channel closure would be identical in the two hemichannels despite their opposite gating polarities. Tong et al. [21] and Dong et al. [22] have reported that the NT plays an important role in voltage-gating and permeability of channels formed by chicken Cx45.6, an ortholog of Cx50. Gemel et al. [23] report that the NTs of Cx40 and Cx43 are determinants of the physiological properties of channels formed by these connexins. Gating data on Cx37 [24,25] derived from macroscopic and single-channel data are also consistent with a role for the NT. However, Gonzalez et al. [26] report that the correlation between Vj/fast-gating polarity and charge in the NT is not perfect. Clearly this issue warrants further investigation.

4.6.1 Stoichiometry of Polarity Reversal The experiments described above substituted or removed charged residues in all six connexin subunits. Oh et al. [27] investigated whether reversal of the voltage polarity to which a hemichannel closes required the substitution of charges in all six subunits. This was done by examining the polarity of Vj/fast-gating in

4 Voltage-Gating Mechanisms

115

unapposed hemichannels containing connexin monomers with the polarityreversing substitution Cx32N2E. The mutations were created on a background of Cx32*43E1, which forms functional unapposed hemichannels [28]. Unapposed hemichannels of Cx32*43E1 were shown by single-channel recordings to display both Vj/fast-gating and loop/slow-gating when the intracellular face of the hemichannel is sufficiently hyperpolarized, that is, both gates tend to close at negative potentials. The open probability of the channel increases when the cell is depolarized. The N2E substitution reverses the polarity of Vj/fast-gating; rapid closures to substates are only observed at depolarizing potentials greater than 20 mV while the polarity of loop/slow-gating events is unchanged. These single-channel studies further supported the view that the two processes, Vj/ fast-gating and loop/slow-gating, are functionally distinct. Heteromeric unapposed hemichannels composed of both Cx32*43E1 and Cx32N2E*43E1 subunits display bipolar Vj/fast-gating, closing to substates at both positive and negative potentials, with maximal open probability at a holding potential of 0 mV. Loop/slow-gating transitions are observed only with substantial hyperpolarization in these heteromeric channels. The number of hemichannels displaying bipolar Vj/fast-gating in cells expressing mixtures of the two connexin subunits coincides with the number of hemichannels expected to contain a single oppositely charged subunit. For example, injection of RNA at a 20:1 ratio of wild-type Cx32 to Cx32N2E is expected to yield 26.5% of the hemichannels containing at least one N2E subunit. Experimentally, 39 of 121 (32.2%) single channels displayed bipolar Vj/fast-gating. If two subunits were required, only 3.25% of channels would be expected to display bipolar gating. Similar results were obtained when the ratio of injected RNA was reversed, 1:20 ratio of wild-type Cx32 to Cx32N2E. There was also a correlation between the voltage sensitivity of the channel and the number of subunits available to close for a given polarity. For example, Vj/fastgating transitions at a given positive potential occurred more frequently in unapposed homomeric Cx32N2E hemichannels than heteromeric unapposed hemichannels containing a single Cx32N2E subunit. The simplest interpretation of these results is that the movement of the voltage sensor in a single connexin subunit is sufficient to initiate Vj/fast-gating and that it results from conformational changes in individual connexin subunits rather than by a concerted change in the conformation of all six subunits. If gating were to occur by a concerted movement, the voltage sensors in the five wild-type subunits would be improperly oriented with respect to the electric field in the heteromeric channel. It is likely that these sensors would immediately respond to the orientation of the electric field to favor channel opening. Consequently, one would predict that the duration of channel closure would be brief in heteromeric channels containing a single oppositely charged subunit. This did not appear to be the case [27], disfavoring a concerted conformational change underlying Vj/fast-gating. It should be noted that the individual gating model interpretation of the data is based on the assumption that channel closure is initiated by inward

116

T. Bargiello and P. Brink

movement of a voltage sensor contained in the NT and that the voltage profile across the voltage sensor produces a force that favors the inward movement of positively charged subunits at positive potentials and the inward movement of positively charged subunits at negative potentials. It also assumes that the voltage gradient across the voltage sensing residues is linear.

4.6.2 Range of Charge Substitutions that Cause Polarity Reversal The ability of charge substitutions at the second amino acid position to reverse the polarity of Vj/fast-gating of Cx26 and Cx32 hemichannels suggests that part of the NT of Cx32 and Cx26 resides within the channel pore, as charged residues could sense Vj in this location. The reversal of the polarity of Vj/fastgating provided a means to map the extent to which the NT senses the Vj field and hence define the physical structure of the NT relative to the channel entrance. In a series of papers, it was shown that negative substitutions up to and including the tenth but not the eleventh position could reverse the polarity of Vj/fast-gating of Cx32 and Cx32*43E1 hemichannels [29,30]. As expected, positive charge substitutions at the second, fifth, and eighth residues maintained the negative gating polarity of Cx32*43E1 hemichannels. These results led to the view that the first ten amino acids of NT of Cx32 formed the cytoplasmic entry of the channel pore. This view was further supported by the solution of the structure of an NT peptide of Cx26 [31] by nuclear magnetic resonance (NMR). Based on sequence homology [31] it is likely that all group I (b-group) connexins have a similar motif of two structured domains connected by a structured turn formed by residues twelve to 15. It was proposed that the glycine residue at the twelfth position is a critical determinant of the turn, as the substitution of this residue with proline did not alter channel function, whereas the substitution of amino acids, serine, valine, or tyrosine, which decrease flexibility, led to loss of function [31]. The turn provides a substantial degree of flexibility to the NT and should allow the first ten amino acids to be positioned in the channel pore. The view that the NT contributes to the formation of the Cx32 channel pore is further supported by studies of permeation demonstrating that charge substitutions at the second, fifth, and eighth positions substantially alter unitary conductance and the direction and amount of single-channel current rectification [30]. Note that if the voltage sensor resides in the channel pore, voltage sensitivity should depend on ionic strength of the surrounding solutions, becoming less sensitive as the ionic strength is increased and more sensitive as ionic strength is decreased. This prediction is a consequence of charge screening that would change the voltage profile of the channel and has yet to be examined in detail in Cx32 channels. Banach et al. [32] provided data for homotypic Cx43 channels that showed an effect opposite to the prediction, that is, decreasing salt

4 Voltage-Gating Mechanisms

117

concentration diminished Vj/fast-gating. Further investigation of the effect of charge screening on Vj/fast-gating is needed. The data described so far support the simple interpretation that the NT contains a voltage sensor for Vj/fast-gating and that charged residues in the NT initiate channel closure by moving toward the cytoplasm. Oh et al. [30] further explored the mechanism of polarity determination by utilizing the permeation model of Chen and Eisenberg [8] that solves the PNP equations in one dimension. The solution of the PNP equations at different membrane potentials provides a set of voltage profiles for a given channel model containing regions of fixed permanent charge in the channel pore. The first derivative of the voltage profile/distance relationship provides a measure of the voltage dependence of the force acting on the voltage sensor. Using this approach, Oh et al. [30] demonstrated that a positive charge located near the intracellular end of the channel (i.e., for a positively charged NT voltage sensor) would distort the voltage profile along the axis of the pore and provide a force favoring the inward movement of a voltage sensor at negative potentials and outward movement at positive potentials. An asymmetric energy barrier was assumed to exist (Fig. 4.5) that would permit only the inward movement of the voltage sensor. A second unanswered question addressed in this study was how the addition of negative charge to the NT causes the reversal of Vj/fast-gating polarity. Recall that the mutations that reverse gating polarity add negative charge to the positively charged NT of Cx32. If there are six positively charged residues in the wild-type channel (presumed to correspond to the positively charged amino-

Fig. 4.5. A schematic representation of the energy profile of a model channel at 0 mV corresponding to the gating model. Proposed by Verselis et al. [20], the inward movement of a voltage sensor (+ at negative voltages, – at positive voltages) would require the crossing of the energy barrier depicted by the peak A. The outward movement of a positively charged voltage sensor at negative holding potentials from the open state would be prohibited by the size of the energy barrier depicted by the peak B.

118

T. Bargiello and P. Brink

terminal methionines), it is unlikely that the addition of six negative charges would reverse, rather than neutralize, the valence of the voltage sensor. The PNP models considering two adjacent regions of fixed positive and negative charge produce voltage profiles characterized by regions of negative slope that reverse the orientation of the voltage gradient across the region of fixed charge. As a consequence, positive charges would experience a negative voltage gradient oriented toward the cytoplasmic surface of the channel at all potentials that would favor their inward movement, whereas the negative charges lie within a positive voltage gradient and would tend to move outward, away from the cytoplasm at all voltages. If the valence of the charges is equal, then the net force acting on the two regions of charge is close to zero at all potentials and consequently, the channel would not display any voltage dependence. The simple model fails to explain the reversal of gating polarity of Cx32 by negative charge substitutions such as N2E and G5D because the voltage profiles are symmetric over the two charges and the resulting forces acting on each charge are equal and opposite. Evidently, the reversal of gating polarity requires an asymmetry in the voltage profile such that the net force acting on both sets of charges is not zero. The required asymmetry could be provided by partial charges present in the NT or in other regions of the channel. Possible candidates include K22 at the NT/first transmembrane segment (M1) border; E208 at the border of the fourth transmembrane segment (M4) and the carboxyl-terminal domain (CT) (M4/CT border); the highly conserved charges in transmembrane segments, R32 in M1 and R142 and E146 in the third transmembrane segment (M3); and charged residues located at the M1 border with the first extracellular loop (E1) border. As the charges at these positions may not move in response to changes in applied voltage, they would not be considered to be part of the voltage sensor. However, they could influence the movements of the voltage sensor in the NT by causing distortions in the voltage profile across the channel pore.

4.6.3 Bipolar Vj/Fast-Gating The simplest model, in which Vj/fast-gating is initiated by inward movement of the voltage sensor, is challenged by the existence of homomeric hemichannels that display bipolar Vj/fast-gating. These include intercellular channels formed by Cx32T8D [29] and unapposed chimeric Cx32T8D*43E1 hemichannels, by unapposed chimeric Cx32*43E1 hemichannels that contain double point mutations N2R+G5D and N2E+G5K [30], and Cx50 unapposed hemichannels [33]. Single-channel records illustrating the bipolarity of Cx32N2R+G5D*43E1 unapposed hemichannels are presented in Fig. 4.6. The open probability of this channel is maximal at a holding potential of 0 mV and decreases with either hyperpolarization or depolarization. An interesting feature of unapposed hemichannels containing permutations of positive and negative charges at the second

4 Voltage-Gating Mechanisms

119

Fig. 4.6. Bipolar Vj/fast-gating of Cx32N2R+G5D*43E1 hemichannels. (a) A cell-attached record of a single Cx32N2R+G5D*43E1 channel at voltages ranging over 100 mV. Closures to substates by Vj/fast-gating are observed at both positive and negative potentials. (b) The current-voltage relation of a single Cx32N2R+G5D*43E1 channel obtained in an outside-out patch configuration illustrating the bipolarity and increased sensitivity of Vj/fast-gating at negative potentials. The current-voltage relation of the open state is linear, similar to that of Cx32G5D*43E1 hemichannels. (c) A plot of the Popen-voltage relation illustrating the bipolarity of Vj/fast-gating. The open probability of the channel is reduced at both positive and negative potentials and maximal at intermediate potentials. Vj/ fast-gating is more sensitive to negative potentials suggesting a dominant role of the charge at the second position. Data points were obtained from all point histograms after concatenation of voltage traces obtained at each voltage from at least three separate patches. Records were obtained with pipette and bath solutions composed of 140 mM KCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM ethyleneglycoltetraacetic acid (EGTA), and 10 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES), pH 7.6. (A highresolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Oh et al. [30] with permission.)

and fifth positions is that the degree of voltage sensitivity correlates with the sign of the charge at the second position. For example, Vj/fast-gating of Cx32N2R+G5D*43E1 is more sensitive to negative polarization, while Vj/ fast-gating of Cx32N2E+G5K*43E1 is more sensitive to positive polarizations. This result suggests that there may be a constriction in the vicinity of the second

120

T. Bargiello and P. Brink

residue, causing a greater proportion of the electric field to drop over this position. However, the apparent dominance of charges at the second residue could also be explained by invoking changes in the voltage profile of the channel such that a larger proportion of the applied voltage drops over the second residue specifically when charged residues are present at positions closer to the entrance of the channel pore. Oh et al. [30] could not create any realistic charge distributions with the PNP model that would create a voltage profile that would allow bipolarity of gating. This would require a voltage profile that produces a U-shaped force-distance relation. At this time, the simplest model to explain bipolarity is to allow the voltage sensor to behave as a center-open toggle switch, responding with an outward movement to one polarity and an inward movement to the opposite polarity of voltage. In this case, mutations that display bipolarity would have to alter the energy profile of the channel such that opposite movements of the voltage sensor would be energetically feasible, that is, that both movements reduce the height of energy barrier B in Fig. 4.5. The model does not require the existence of distinct gating mechanisms, as the opposite movement of the voltage sensor could be coupled to the same downstream mechanism that leads to channel closure by destabilizing the open state. Bipolarity could also arise if one invoked the existence of two open channel states each with different charge distributions, such that the voltage profile of one open state of the channel favors the inward movement of the voltage sensor at positive potentials, whereas the voltage profiles of the other open state favors the inward movement of the voltage sensor at negative potentials. This possibility is suggested by the presence of two open states in channels formed by positive charge substitutions, but it is unclear if these states are created or unmasked by these substitutions and there is no independent evidence to support this possibility.

4.7 Molecular Determinants and Mechanisms of Loop/Slow Gating At present, there is no published information regarding the molecular determinants and conformational changes underlying loop/slow-gating in connexin channels. In all cases examined, loop/slow-gates close with sufficient negative polarization [17,27,33]. No mutations have been reported that reverse the gating polarity of this process. The presence of a plateau in the G-V relation of most intercellular channels and further relaxation at higher Vjs is a good indication that the Vj/fast-gate is more sensitive to voltage than the loop/slowgate, but it is not clear if this relation will apply to all channels and there is no reason to believe that it will. Interestingly, the loop/slow-gates of Cx46 unapposed hemichannels appear to be mechanosensitive [34].

4 Voltage-Gating Mechanisms

121

4.8 Conformational Changes Associated with Voltage-Gating There are no published studies that directly demonstrate changes in conformation accompanying voltage-gating of connexin channels. In other channels, conformational changes during gating have been demonstrated directly using fluorescence spectroscopy [35], indirectly using state-dependent disulfide trapping [36] or differential accessibility to thiol-modifying reagents [37]. In connexin channels, conformational changes have been inferred from either structural or electrophysiological analyses of mutations. In a general framework, voltage-dependent processes are believed to result from the initial movement of a voltage sensor that is coupled to subsequent changes in conformation that alter ion passage through the permeation pathway. In the case of connexin channels, it is possible that the voltage sensor and gate are the same structure. For example, if charged residues in the NT comprise the voltage sensor for Vj/fast-gating, then by virtue of residing in the channel pore, the movement of the NT could occlude the ion permeation pathway, that is, form the physical gate. However, this does not need to be the case. Two studies have suggested that conformational changes involving a highly conserved proline residue in M2 of all connexins are involved in voltage-gating [38,39], while a third proposes that voltage-gating results from an interaction between the cytoplasmic loop (CL) and CT of Cx43 [40]. A fourth study proposes that the NT forms a gating particle in Cx26 hemichannels [41]. These are discussed in the following sections.

4.8.1 Role of Proline Kink in Voltage-Gating The intrinsic flexibility and the structural characteristics of proline-induced kinks in transmembrane helices allow conformational responses to specific interactions. This makes possible mechanisms in which the proline kinks function in the propagation of conformational changes from one protein domain to another [42]. Ri et al. [38] reported that Monte Carlo simulations of M2 in Cx32 predicted the existence of a significant bend in M2 (37 degrees) that resulted from the formation of a hydrogen bond between the side-chain hydroxyl of T86 and the backbone carbonyl of I82. The formation of this hydrogen bond is facilitated by a disruption of the helical character of M2 by a highly conserved proline at the 87th position in Cx32. The structural characteristics of the sequence ‘‘T86 P87’’ are termed a TP motif. Simulations indicated that these structural characteristics are retained with mutation of the T86 residue to serine or cysteine, which preserve the hydrogen bond potential, whereas mutations of T86 to alanine, valine, asparagine, and leucine, which remove the hydrogen bonding potential of the side chain of T86 with the backbone carbonyl of I82, substantially decreased the bend angle from that of wild-type, averaging 20 degrees. Ri et al. [38] observed that the macroscopic G-V relations of the T86

122

T. Bargiello and P. Brink

mutant gap junction channels were shifted according to the predicted bend angle such that the open probabilities of the mutant channels were progressively reduced, in the order L > V > N A C > S wild-type. These data led to a structural model in which the open conformation of the Cx32 channel corresponds to a M2 helix with a greater bend, and the closed conformation to a less bent helix. A gating model was proposed in which the application of a Vj would initiate a set of conformational changes that include breaking the proposed hydrogen bond between the side chain of the T86 residue and the backbone carbonyl of residue I82. The loss of this hydrogen bond would favor a reduction in the bend angle of the proline kink, which would mediate the conformational transition from the open state to the closed state. The predicted 17 degree decrease in the bend angle could displace the cytoplasmic end of M2 by as much as 4.4 A˚ toward the center of the pore lumen, resulting in a decrease of the radius at the intracellular end of the channel [38]. The model is supported by the observed increase in rectification of Vj-induced subconductance states, which suggests that the cytoplasmic end of Cx32 channels narrows during Vjmediated closure [7]. A similar narrowing has been proposed for Cx43 gap junctions by Bukauskas et al. [43]. Interestingly, 14 of 17 connexins contain a threonine, serine, or cysteine residue just upstream of the conserved proline residue in the TP motif, preserving the hydrogen bond potential and bend angle. Only the sequences in Cx43 (VP), Cx31.9 (AP), and Cx40.1 (LP) deviate from this motif. The conservation of sequence suggests that a similar mechanism could operate in most connexin channels. The difficulty with this study is that it relied exclusively on macroscopic current recordings of heterotypic intercellular channels rather than singlechannel recordings. In fact, several of the mutations caused substantial reductions in Gmin, suggesting that the voltage dependence of the loop/slow-gate was also shifted. Single-channel recordings are needed to clarify the effect of these mutations on Vj/fast-gating. The proposed mechanism could be tested further by the creation of mutations in the TP motif that would selectively lock the channel in the open or closed state. Suchyna et al. [39] also reported that mutations of P87 altered the expression of Vj-dependent gating but not pH gating of Cx26. On this basis, they concluded that residue P87 functions as a transduction element in voltage-dependent gating of Cx26. Although Suchyna et al. [39] did not explicitly define their use of the term transduction, it is likely that they were referring to the mechanism by which the mechanical movement of one protein domain could be propagated to another, rather than the means by which electrical energy is converted into mechanical energy, that is, the movement of the voltage sensor. They suggested that changes in conformation of the open and closed states resulted from cistrans isomerizations of the proline residue. However, cis-proline isomers have not been observed in -helices, most likely because this conformation would produce steric clashes with the backbone. Furthermore, this mechanism is unlikely, as the energy given for cis-trans isomerization, 13.5 kcal/mol by

4 Voltage-Gating Mechanisms

123

Suchyna et al. [39] is substantially greater than the energy required to close Cx26 hemichannels, approximately 4.4 kcal/mol (see [38]).

4.8.2 Role of the Carboxyl-Terminal Domain in Vj-Gating of Connexin43 Shibayama et al. [40] have proposed a different model of Vj/fast-gating for Cx43 hemichannels. The model is based on the observation of the loss of Vj/fast-gating but retention of loop/slow-gating in Cx43 mutations that truncate the CT. Shibayama et al. proposed that both Vj/fast-gating and pH gating of Cx43 channels result in part from a receptor-particle interaction. In this model of voltage-gating, the CT of Cx43 acts as a particle that interacts with a receptor contained within the CL, termed the L2 region. The L2 region contains two histidine residues located at residues 126 and 142. Mutation H142E reduces the frequency of Vj/fast-gating events, alters the structure of the L2 domain, and reduces the interaction between L2 and CT [40]. The authors suggest that voltage induces a conformational change or series of conformational changes in the open state that expose the receptor allowing the binding of the CT to L2. The binding is postulated to result in the entry of the open channel into the substate. A similar effect, loss of Vj/fast-gating and retention of loop/slow-gating, has been reported for carboxyl-terminal fusions of enhanced green fluorescent protein (EGFP) [44] (see Chapter 5). In this case, the loss of Vj/fastgating could reflect a steric restriction preventing the interaction of the CT with L2. Alternatively, the effects of this mutation could be to shift the voltage dependence of the Vj/fast-gating to large Vjs independently of that of loop/slow-gating. This would require that the free energy of the Vjinduced substate be increased while preserving the free energy of the open and fully closed states. The mechanism envisioned by Shibayama et al. [40] can be described by the state diagram shown in Fig. 4.7. Transitions between the two open states, O (receptor unexposed) and OR (receptor exposed) are voltage-dependent. The rate of entry of the OR state into the substate (ORP) depends on the probability that the receptor resides in a permissive state (OR) and the concentration of the particle P contained in the CT. The dwell time of the channel in the ORP depends on Koff, the rate of unbinding the particle from the receptor. A potential difficulty with the proposed model is that the opening rate (that is, the dwell

Fig. 4.7. Proposed Vj/fast-gating scheme for Cx43. A state diagram illustrating the proposed mechanism of Vj/fast-gating based on Shibayama et al. [40]

124

T. Bargiello and P. Brink

time in ORP) of most connexin channels is substantially voltage-dependent. For the dwell time of the substate to be voltage-dependent, Koff must be voltagedependent. This could only be accomplished if the bound particle-receptor lay in the Vj-field. It is not clear if the particle-receptor mechanism underlies Vj/fast-gating in other connexins, as the CL and CT segments of the protein sequence are among the least conserved across the connexin family. Cx43, Cx31.9, and Cx40.1 do not contain a TP motif in M2 but, as noted, rather have the sequences VP, AP, and LP, respectively, at this position. It would be informative to determine the functional properties of these motifs in the M2 of Cx32. Revilla et al. [45] have suggested that deletions of the CT of human CX32 (R220 mutation) abolish Vj/fast-gating of intercellular channels and have proposed that Vj/fast-gating might arise from the movement of the CT into the channel pore. The loss of Vj/fast-gating was inferred from the kinetic analysis of macroscopic currents. No single-channel data were presented to directly demonstrate the loss of Vj/fast-gating. The fast component described by Revilla et al. [46] was surmised to correspond to the closure of Vj/fast-gates but appeared to comprise only a small component of the current relaxations of the wild-type channel, which would not appear to account for the observed Gmin. The truncation shifted the voltage dependence of mutant channels to smaller Vjs and markedly decreased the time constant of current relaxations relative to wild-type. The shift in voltage dependence may have obscured the contribution of the fast component in the truncated channel and complicated the interpretation of kinetic analyses. The presence of a plateau in the G-V relation resulting in a substantial minimal conductance in the truncation mutation (Gmin 0.2 versus Gmin 0.3 in wild-type) would suggest the continued operation of the Vj/ fast-gates. It would be interesting to examine the gating of comparable truncations of Cx32*43E1 hemichannels that contain negative charge substitutions in the NT. Loss of Vj/fast-gating would be evidenced by the absence of Vj/fastgating transitions at positive potentials and the presence of loop/slow-gating transitions at negative potentials.

4.8.3 Connexin26 Gating Plug Model Recently, Oshima et al. [41] reported that the three-dimensional structure of human CX26 channels contains a plug in the cytoplasmic entry (see Chapter 2). They hypothesize that the plug is formed by all six adjacent NTs and propose that the plug acts as a gating particle that occludes the channel pore by moving away from the cytoplasmic surface deeper into the channel lumen, that is, an outward movement with respect to the cytoplasm. There are several difficulties with this model in terms of Vj/fast-gating. First, the direction of movement of the proposed NT plug is opposite to that predicted for the movement of an NT voltage sensor for Cx26 hemichannels. The electrophysiological studies described above suggest that an inward movement of

4 Voltage-Gating Mechanisms

125

negative charged sensor at positive potentials initiates channel closure, while Oshima et al. [41] suggest that the outward movement of the plug (into the channel pore) results in channel closure. Second, the structure shown appears to place the entire NT of all six connexins within the channel pore, whereas the electrophysiological data indicate that only the first ten residues in the NT lie within the channel pore. Third, if the plug, as shown, is a complex of six NTs, then all six NTs would be expected to move as a single unit, that is, gating would involve a concerted movement. Oh et al. [27] maintain that the NTs move individually to initiate Vj/fast-gating. Even if this interpretation by Oh et al. is incorrect, it is difficult to envision how heteromeric channels containing a single oppositely charged subunit could give rise to bipolar gating if they combined to form a gating plug. The simple expectation would be that the valence of the gating particle would be reduced and consequently, the voltage sensitivity would be reduced for a given polarization. Based on these considerations, it seems unlikely that gating by an NT plug can account for the observed properties of Vj/fast-gating in Cx26 channels. It is more difficult to assess if the plug gating model corresponds to loop/ slow-gating, as there is little published information detailing the molecular determinants and conformational changes underlying this gating mechanism. It is possible that the structure represents a channel closed by the loop/slowgating mechanism, in that the negative polarity of closure by the loop/slowgates would be correct. Alternatively, the structure presented by Oshima et al. [41] may represent a nonfunctional channel state reflecting a conformation resulting from the isolation and crystallization procedure.

4.9 Molecular Determinants of the Vi-o Sensor Revilla et al. [46] reported the presence of Vi–o dependence in macroscopic recordings of Cx43 intercellular channels expressed in paired Xenopus oocytes. They report that Vi–o dependence is lost in a CT truncation mutation at residue 243 (R243) but present in the truncation mutation at residue 257 (M257). There are only two charged residues in the 242 to 257 segment, the positively charged R243 residue and the positively charged D245 residue. The neutral substitution, R243Q reduced Vi–o dependence and the calculated gating charge, while the neutral substitution D245Q increased Vi–o dependence and the calculated gating charge. Neutralization of both charges removed Vi–o dependence. The mutations were reported to have no effect on the expression of Vj-dependence. Based on this evidence, the authors concluded that these charges form part of the Vi–o voltage sensor, proposing that the Vi–o sensor has a net positive charge and its outward movement initiates channel closure in response to depolarizing Vi–o. However, they did not address how charges contained near the border of M4 and CT would be in a position to sense changes in the electric field created by the changes in Vi–o and be insensitive to changes in Vj. This is problematic as the membrane

126

T. Bargiello and P. Brink

topology predicts that charges in this position would be located in the cytoplasm and consequently lie outside of any electric field (Fig. 4.1a)

4.10 Conclusion The principal feature of voltage-gating of connexin channels is the complexity and variability of the process across members of the protein family. At the present time, most information relates to the mechanisms and molecular determinants of Vj/fast-gating, although there is as yet no general consensus and the mechanism may differ among members of the gene family. Our understanding of the mechanisms and molecular determinants of loop/slow gating is incomplete, owing largely to lack of mutations that can lead to or be interpreted in terms of a gating model. It has been proposed that loop/slow-gating also underlies chemical gating and sensitivity to Vi–o, but this is based on the similar appearance of transitions among open and closed states and needs to be examined further. It should be noted that a major problem with structure-function studies that utilize mutations to dissect voltage-gating is that the approach is largely inferential and model dependent. For example, while it has been shown that charges in the NT determine the polarity of Vj/fast-gating, the movement of this region in response to voltage has not been demonstrated directly. Thus, it cannot be stated unequivocally that this region is the voltage sensor, although no other region of the protein has been shown to possess the required properties. Fluorescence spectroscopy provides the most promising approach toward understanding the mechanisms underlying channel gating in that the method provides a realtime assay of the conformational changes that occur during gating [35]. It should be possible to use this approach to describe the movements of the NT in response to voltage and to determine which domains are responsible for Vj/fast-gating and loop/slow-gating. Another approach, although indirect, utilizes state-dependent lock of the channel by the formation of disulfide bonds or coordination of divalent cations to infer the relative motions of segments of the channel. These approaches have proven valuable in elucidating the gating mechanisms of other ion channels and should provide further insights into the molecular determinants and mechanisms of voltage-gating of connexin channels. Acknowledgments We thank our colleagues at Einstein and Stony Brook for helpful discussions.

References 1. Verselis VK, Bennett MVL, Bargiello TA. A voltage-dependent gap junction in Drosophila melanogaster. Biophys J. 1991;59:114–26. 2. Bukauskas FF, Weingart R. Voltage-dependent gating of single gap junction channels in an insect cell line. Biophys J. 1994;67:613–25.

4 Voltage-Gating Mechanisms

127

3. Barrio LC, Suchyna T, Bargiello T, Xu LX, Roginski RS, Bennett MV, Nicholson BJ. Gap junctions formed by connexins 26 and 32 alone and in combination are differently affected by applied voltage. Proc Natl Acad Sci USA. 1991;88:8410–4. 4. Rubin JB, Verselis VK, Bennett MVL, Bargiello TA. A domain substitution procedure and its use to analyze voltage dependence of homotypic gap junctions formed by connexins 26 and 32. Proc Natl Acad Sci USA. 1992;89:3820–4. 5. Rubin JB, Verselis VK, Bennett MVL, Bargiello TA. Molecular analysis of voltage dependence of heterotypic gap junctions formed by connexins 26 and 32. Biophys J. 1992;62:183–93. 6. Brink PR, Cronin K, Banach K, Peterson E, Westphale EM, Seul KH, Ramanan SV, Beyer EC. Evidence for heteromeric gap junction channels formed from rat connexin43 and human connexin37. Am J Physiol. 1997;273: C1386–96. 7. Oh S, Rubin JB, Bennett MVL, Verselis VK, Bargiello TA. Molecular determinants of electrical rectification of single channel conductance in gap junctions formed by connexins 26 and 32. J Gen Physiol. 1999;114:339–64. 8. Chen D, Eisenberg R. Charges, currents and potentials in ionic channels of one conformation. Biophys J. 1993;64:1405–21. 9. Puljung MC, Berthoud VM, Beyer EC, Hanck DA. Polyvalent cations constitute the voltage-gating particle in human connexin37 hemichannels. J Gen Physiol. 2004; 124:587–603. 10. Srinivas M, Calderon DP, Kronengold J, Verselis VK. Regulation of connexin hemichannels by monovalent cations. J Gen Physiol. 2006;127:67–75. 11. Harris AL, Spray DC, Bennett MVL. Kinetic properties of a voltage-dependent junctional conductance. J Gen Physiol. 1981;77:95–117. 12. Brink PR, Ramanan SV, Christ GJ. Human connexin 43 gap junction channel gating: evidence for mode shifts and/or heterogeneity. Am J Physiol. 1996;271:C321–31. 13. Chen-Izu Y, Moreno AP. Spangleer SR. Opposing gates model for voltage-gating of gap junction channels Am J Physiol. 2001;281:C1604-C1613 14. Ramanan SV, Valiunas V, Brink PR. Non-stationary fluctuation analysis of macroscopic gap junction channel records. J Membr Biol. 2005;205:81–8. 15. Rackauskas M, Kreuzberg MM, Pranevicius M, Willecke K, Verselis VK, Bukauskas FF. Gating properties of heterotypic gap junction channels formed of connexins 40, 43, and 45. Biophys J. 2007;92:1952–65. 16. Oh S, Ri Y, Bennett MVL, Trexler EB, Verselis VK, Bargiello TA. Changes in permeability caused by connexin 32 mutations underlie X-linked Charcot-Marie-Tooth disease. Neuron. 1997;19:927–38. 17. Trexler EB, Bennett MVL, Bargiello TA, Verselis VK. Voltage-gating and permeation in a gap junction hemichannel. Proc Natl Acad Sci USA. 1996;93:5836–41. 18. Bukauskas FF, Elfgang C, Willecke K, Weingart R. Biophysical properties of gap junction channels formed by mouse connexin40 in induced pairs of transfected human HeLa cells. Biophys J. 1995;68:2289–98 19. Revilla A, Bennett MVL, Barrio LC. Molecular determinants of membrane potential dependence in vertebrate gap junction channels. Proc Natl Acad Sci USA. 2000;97:14760–5. 20. Verselis VK, Ginter CS, Bargiello TA. Opposite voltage-gating polarities of two closely related connexins. Nature. 1994;368:348–51. 21. Tong JJ, Liu X, Dong L, Ebihara L. Exchange of gating properties between rat Cx46 and chicken Cx45.6. Biophys J. 2004;87:2397–406. 22. Dong L, Liu X, Li H, Vertel BM, Ebihara L. Role of the N-terminus in permeability of chicken connexin45.6 gap junctional channels. J Physiol. 2006;576:787–99. 23. Gemel J, Lin X, Veenstra RD, Beyer EC. N-terminal residues in Cx43 and Cx40 determine physiological properties of gap junction channels, but do not influence heteromeric assembly with each other or with Cx26. J Cell Sci. 2006;119:2258–68. 24. Ramanan SV, Brink PR, Varadaraj K, Peterson E, Schirrmacher K, Banach K. A threestate model for connexin37 gating kinetics. Biophys J. 1999;76:2520–9.

128

T. Bargiello and P. Brink

25. Banach K, Ramanan SV, Brink PR. The influence of surface charges on the conductance of the human connexin37 gap junction channel Biophys J. 2000;78:752–60. 26. Gonzalez D, Gomez-Hernandez JM, Barrio LC. Molecular basis of voltage dependence of connexin channels: an integrative appraisal. Prog Biophys Mol Biol. 2007;94:66–106. 27. Oh S, Abrams CK, Verselis VK, Bargiello TA. Stoichiometry of transjunctional voltagegating polarity reversal by a negative charge substitution in the amino terminus of a connexin32 chimera. J Gen Physiol. 2000;116:13–31. 28. Pfahnl A, Zhou XW, Werner R, Dahl G. A chimeric connexin forming gap junction hemichannels. Pflu¨gers Arch. 1997;433:773–9 29. Purnick PE, Oh S, Abrams CK, Verselis VK, Bargiello TA. Reversal of the gating polarity of gap junctions by negative charge substitutions in the N-terminus of connexin 32. Biophys J. 2000;79:2403–15. 30. Oh S, Rivkin S, Tang Q, Verselis VK, Bargiello TA. Determinants of gating polarity of a connexin 32 hemichannel. Biophys J. 2004;87:912–28. 31. Purnick PE, Benjamin DC, Verselis VK, Bargiello TA, Dowd TL. Structure of the amino terminus of a gap junction protein. Arch Biochem Biophys. 2000;381:181–90. 32. Banach K, Ramanan SV, Brink PR. Homotypic hCx37 and rCx43 and their heterotypic form. In: Werner R, editor. Gap Junctions: IOS Press; 1998; 76–80. 33. Srinivas M, Kronengold J, Bukauskas FF, Bargiello TA, Verselis VK. Correlative studies of gating in Cx46 and Cx50 hemichannels and gap junction channels. Biophys J. 2005;88:1725–39. 34. Bao L, Sachs F, Dahl G. Connexins are mechanosensitive. Am J Physiol Cell Physiol. 2004;287:C1389–95. 35. Cha A, Bezanilla F. Structural implications of fluorescence quenching in the Shaker K+ channel. J Gen Physiol. 1998;112:391–408. 36. Horenstein J, Wagner DA, Czajkowski C, Akabas MH. Protein mobility and GABAinduced conformational changes in GABA(A) receptor pore-lining M2 segment. Nat Neurosci. 2001;4:477–85. 37. Yang N, George AL Jr, Horn R. Molecular basis of charge movement in voltagedependent sodium channels. Neuron. 1996;16:113–122. 38. Ri Y, Ballesteros JA, Abrams CK, Oh S, Verselis VK, Weinstein H, Bargiello TA. The role of a conserved proline residue in mediating conformational changes associated with voltage-gating of Cx32 gap junctions. Biophys J. 1999;76:2887–98. 39. Suchyna TM, Xu LX, Gao F, Fourtner CR, Nicholson BJ. Identification of a proline residue as a transduction element involved in voltage-gating of gap junctions. Nature. 1993;365:847–9. 40. Shibayama J, Gutierrez C, Gonzalez D, Kieken F, Seki A, Carrion JR, Sorgen PL, Taffet SM, Barrio LC, Delmar M. Effect of charge substitutions at residue his-142 on voltagegating of connexin43 channels. Biophys J. 2006;91:4054–63. 41. Oshima A, Tani K, Hiroaki Y, Fujiyoshi Y, Sosinsky GE. Three-dimensional structure of a human connexin26 gap junction channel reveals a plug in the vestibule. Proc Natl Acad Sci USA. 2007;104:10034–9. 42. Ballesteros JA, Weinstein H. Integrated methods for modeling G-protein coupled receptors. Methods Neurosci. 1995;25:366–428. 43. Bukauskas FF, Bukauskiene A, Verselis VK. Conductance and permeability of the residual state of connexin43 gap junction channels. J Gen Physiol. 2002;119:171–85. 44. Bukauskas FF, Bukauskiene A, Bennett MVL, Verselis VK. Gating properties of gap junction channels assembled from connexin43 and connexin43 fused with green fluorescent protein. Biophys J. 2001;81:137–52. 45. Revilla A, Castro C, Barrio LC. Molecular dissection of transjunctional voltage dependence in the connexin-32 and connexin-43 junctions. Biophys J. 1999;77:1374–83. 46. Revilla A, Bennett MVL, Barrio LC. Molecular determinants of membrane potential dependence in vertebrate gap junction channels. Proc Natl Acad Sci USA. 2000;97: 14760–5.

Chapter 5

Chemical Gating of Connexin Channels Rebecca Lewandowski, Junko Shibayama, Eva M. Oxford, Rosy Joshi-Mukherjee, Wanda Coombs, Paul L. Sorgen, Steven M. Taffet and Mario Delmar

Abstract The physiology of gap junctions is regulated by changes in the microenvironment of the cell. The most defined mechanism for such regulation is the closure of gap junction channels in response to acidification of the intracellular space, that is, pH gating or chemical gating. In the best studied case, that of Cx43, chemical gating involves interactions between a specific segment of the cytoplasmic carboxyl-terminal domain with a specific segment of the cytoplasmic loop domain. These interactions have been characterized at the molecular and structural levels. Additionally, there is evidence that interaction of cytoplasmic aminosulfonates and peptide reagents can also affect the gating mechanisms controlled by these inter-domain interactions. Keywords pH gating Acidification-induced uncoupling Particle-receptor model Cx26 Cx32 Cx40 Cx43 Cx45 Cx46 Cx50

5.1 Introduction Gap junctions facilitate the formation of functional networks. As such, the presence of an intercellular communication pathway is fundamental to synchronous tissue function. Yet, as for other ion channels, the existence of an open avenue for communication is as important as the possibility for proper isolation in case of need. It is therefore not surprising that, like other channels, gap junctions operate not as rigid pipes that put two cells in contact regardless of conditions, but instead as gated flexible elements that modify their conductive properties according to cellular needs. The heart is a good example with which to illustrate the physiological importance of gap junctions, both in health and in disease. The cardiac organ M. Delmar (*) Center for Arrhythmia Research, Division of Cardiovascular Medicine, University of Michigan, 5025 Venture Drive, Ann Arbor, MI 48108, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_5, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

129

130

R. Lewandowski et al.

is composed of millions of myocytes. For the heart to pump efficiently, the mechanical activity of each myocyte needs to synchronize with that of the rest, so that their work integrates into an effective contraction. This synchronization in mechanical function is mediated by the propagation of electrical signals through gap junctions. In the absence of open gap junction channels, ventricular fibrillation and death ensues [1,2,3,4,5,6]. Yet, while open gap junctions are fundamental to normal function, timely gap junction closure may revive (or kill) a heart. Consider the case in which coronary blood flow is abruptly restricted to a given area of the ventricular tissue. Under this condition, the affected myocardial cells attempt to preserve contractile activity by adenosine triphosphate production via the anaerobic pathway. The resulting intracellular acidification may lead to gap junction closure [7,8]. In itself, this process can be cardioprotective, since the closure of gap junctions would create a barrier to insulate the unaffected neighboring cells from the intercellular diffusion of ions, metabolites, and signaling molecules emanating from the injured tissue. Yet, this phenomenon could also be deleterious to the heart, since the electrical barriers created by the closed gap junctions can become an obstacle to the propagation of the action potential and as such, become a substrate for malignant ventricular arrhythmias [1,2,4,6,9]. From this standpoint, understanding the molecular mechanisms of pH gating of cardiac gap junctions is relevant to the implementation of balance between two physiological consequences of direct medical significance: cardioprotection and ischemia-induced arrhythmias (see Chapter 21). This chapter deals mostly with the molecular mechanisms of pH gating of connexin channels. The emphasis is on the chemical gating of Cx43 (the major cardiac connexin), simply because much more is known about the molecular basis of its regulation than for other connexins, though information about other connexins is discussed when available.

5.2 Regulation of Connexin43 Channels: The Particle-Receptor Hypothesis All Cx43 protein–protein interactions described so far (and where structurefunction has been studied) require the integrity of the carboxyl-terminal domain (CT) (see Chapter 11). The mechanism by which the CT regulates the channel is still under study, but all data thus far suggest it requires an interaction of the CT with a separate intracellular domain. This chapter summarizes some of the data that led to the hypothesis that the regulation of Cx43 channels follows a particle-receptor model similar, but not necessarily identical, to the ball and chain mechanism of voltage-gating of K+ channels [10,11]. This model proposes that the CT of Cx43 acts as a gating particle that under the appropriate conditions binds to a separate region of the protein, acting as a receptor. The

5 Chemical Gating

131

receptor would be affiliated with the pore, and the particle-receptor interaction would cause channel closure. Several disclaimers are appropriate at this point. The terms chemical gating and pH gating, which are used throughout this chapter, are often used in the gap junction literature to describe the regulation of gap junctions by chemical factors, or the regulation by intracellular pH. Use of these terms is not intended to imply that these regulators modulate the connexin channels by interposition of a channel gate, as conceptualized in the Hodgkin-Huxley-type models of gating. Similarly, when pH gating is described as a particle-receptor interaction, it is also not intended to imply that the closure of the channel is by direct particle block. Though that possibility exists, other mechanisms cannot be excluded. The first identification of the CT as a regulatory domain for pH gating came from studies showing that truncation of Cx43 (Cx43M257) prevented acidification-induced uncoupling [12]. Uncoupling could be restored if the truncated portion of the CT was coexpressed as a separate peptide [8]. These data led the authors to propose that pH gating of Cx43 functions as a particle-receptor interaction, where the CT domain acts as a gating particle. At normal pH, the gating particle would be away from the pore and the channel would be open. Upon acidification, the particle would bind to a separate region of the protein (a receptor) affiliated structurally or functionally with the pore. This particlereceptor interaction would lead to channel closure. It should be noted that this particle-receptor model is consistent with results regarding the regulation of Cx43 by insulin and insulin-like growth factor [13], platelet-derived growth factor [14], as well as src [15].

5.3 Structural Bases for the Particle-Receptor Interaction 5.3.1 Binding of the Carboxyl-Terminal Domain to the Cytoplasmic Loop Domain of Connexin43 The studies mentioned above led to the identification of the CT as a gating particle. However, the location of the receptor within the primary sequence of Cx43 remained elusive. Recently, a variety of spectroscopic (surface plasmon resonance, nuclear magnetic resonance [NMR]) and biochemical (enzymelinked sorbent assays, cross-linking) methods were used to show that the CT binds, selectively and specifically, to a peptide corresponding to the second half of the cytoplasmic loop (CL) in a pH-dependent manner [16]. The binding is greater, and the kinetics faster, at low pH. The results led to the hypothesis that the second half of the CL acts as the receptor for the gating particle. Further experiments of Hirst-Jensen et al. [17] show that amino acids M100 to Y155 of Cx43 (most of the CL) interact, directly or indirectly, with amino acids N343 to K346 and R376 to D379 of the CT.

132

R. Lewandowski et al.

5.3.2 High-Order Structures in the Carboxyl-Terminal and Cytoplasmic Loop Domains Additional studies have focused on identifying the resonance assignments that change with binding and, ultimately, spatial organization of each molecular fragment when isolated in solution. As shown in Fig. 5.1, the second half of the CL, L2 (residues 119 to 144), is organized as two -helical segments separated by a random coil. Each helix contains a histidine residue (positions 126 and 142), both of which are protonated upon acidification. Mutations in residue H142 can affect channel function [18]. Separate studies have demonstrated that the Cx43CT sequence is mostly configured as a random coil (Fig. 5.2). Interestingly, analyses of eukaryotic genomes estimate that 35 to 51% of all proteins contain disordered regions greater than 40 amino acids (aa) in length [19]. Protein database analysis (http://www.rcsb.org/pdb/home/home.do) determined that there are four broad categories of functions pertaining to disordered proteins: molecular recognition, molecular assembly, protein modifications, and entropic chains [19]. The major functions of such disordered segments, include protein–protein binding, flexibility, and phosphorylation, all apply to the Cx43CT. Thus, a disordered Cx43CT would be ideal for signaling by allowing different binding partners with both high specificity and low affinity to interact and to rapidly switch between molecular partners, thus activating alternative signaling paths [20].

5.3.3 Dimerization of Connexin43 Carboxyl-Terminal Domains Binding of the CT to the CL may not be the only pH-dependent interaction that occurs within the microenvironment of a Cx43 channel that leads to

Fig. 5.1 Solution structure of the L2 segment of the Cx43 CT. The peptide was diluted in phosphate buffered saline (PBS) pH 5.8 and maintained at 78C. Cx43L2 (amino acids 119 to 144) has two -helical domains, each containing a histidine residue (positions 126 and 142), separated by a random coil. Helical regions for the lowest energy structure are indicated in red/ yellow. Histidine residues in L2 (H126 and H142), both protonated upon acidification, are indicated in bold. (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com) (Modified from Duffy et al. [16] with permission.)

5 Chemical Gating

133

Fig. 5.2 Solution structure of Cx43CT. The peptide was diluted in PBS pH 5.8 and maintained at 78C. Cx43CT (amino acids 255 to 382) is mostly configured as a random coil, with two helical domains. Helical regions for the lowest energy structure are indicated in red/yellow. (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com) (Modified from Sorgen et al. [21] with permission.)

pH-induced closure. NMR translational diffusion experiments indicated reduced mobility of the CT fragment at low pH, which could be explained by the formation of a higher molecular weight species, consequent to oligomerization of the fragment [16]. Further studies, applying a variety of biophysical methods, demonstrated that the Cx43CT dimerized in a pH-dependent manner [21]. These results led to the suggestion that dimerization of the CT may be one of the structural changes involved in the pH regulation of Cx43. In this view, dimerization would increase the CT-CL binding affinity to bring the channel to a closed state. Additionally, the change in oligomeric state of the CT may play a role in modulating the molecular partners that associate with Cx43 under a given condition, thus acting as a switch for modifications in channel function. The relevance of this intermolecular interaction to the actual process of acidification-induced closure in Cx43 channels remains to be determined. It is interesting to note that the phenomenon of dimerization is not limited to Cx43. Tetrameric channel proteins such as hyperpolarization-activated cyclic nucleotide-gated K+ channels (HCN), the inositol triphosphate (IP3) receptor, and the N-methyl-D-aspartate (NMDA) inotropic receptor for glutamate have been dubbed ‘‘dimers of dimers’’ because of the self-association of specific subunit domains [22,23,24]. Cx43 would be considered a ‘‘trimer of dimers’’ once the CT domains dimerize. In some cases, dimerization has been shown in vitro using isolated protein fragments and then corroborated in functional channels [24,25]. Dimerization in other channel proteins has functional consequences. Indeed, in the case of cyclic nucleotide-gated HCN channels, dimerization of a regulatory domain of the protein substantially modifies channel function [26,27]. For -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors (non–NMDA-type receptors for glutamate), dimer formation by the ligand binding cores is required for activation of ion channel gating [23,28,29]. Directly analogous to the hypothesized functions for the Cx43CT domain, dimerization of

134

R. Lewandowski et al.

the calmodulin-binding domains of Ca2+-activated K+ channels has been suggested to physically alter the pore-forming helices to drive channel gating [30] and dimerization of the CT of the Na+/H+ exchanger NHE1 has been implicated in pH regulation [31]. It is indeed possible that oligomerization of regulatory domains is a common structural modification leading to modifications in channel conductivity. Further functional studies will be needed to assess this hypothesis.

5.3.4 Functional Studies: Modification of Connexin43 Gating by a Cytoplasmic Loop Peptide Biochemical and spectroscopic studies indicated that the Cx43CT can bind to a peptidic sequence corresponding to the second half of the CL of Cx43. Additional studies provided experimental evidence to show that the CT-CL interaction occurs within a functional Cx43 channel [32]. Cx43 channel currents were recorded in the presence of a peptide corresponding to the second half of the CL region (L2), delivered via the patch pipette. This manipulation did not modify unitary conductance, but decreased the frequency of transitions into a subconductance state (the residual state), prolonged open time, and altered the voltage dependence of the channel in a manner analogous to that observed with truncation of the CT (M257). More recent studies showed that mutation H142E in the L2 region had a similar effect on Cx43 single-channel activity [18] (see Chapter 4); this correlated with a loss of Cx43CT-L2 binding (by translational diffusion analysis) and a disruption of the secondary structure of L2 (determined by NMR). Overall, these studies support the hypothesis that the L2 segment acts as a receptor that interacts with a flexible intracellular gating element (Cx43CT) during channel gating.

5.3.5 Changes in Single-Channel Properties Consequent to pH Gating of Connexin43 Early studies revealed that acidification of the intracellular space did not modify the unitary conductance of the cardiac gap junction channels [33]. Follow-up studies by Bukauskas et al. [34] confirmed this observation for the Cx43 channels, and further showed that upon acidification, single-channels transit from the residual to a highly stable closed state. Furthermore, additional studies conducted using enhanced green fluorescent protein (EGFP)-tagged proteins show that loss of electrical coupling occurs without noticeable changes in the size of the gap junction plaque [35]. These results supported the notion that in the presence of low intracellular pH, gap junction channels close but remain in the membrane, available for reestablishment of junctional communication if intracellular pH returns to normal values.

5 Chemical Gating

135

5.4 Connexin43 Regulation: Intermolecular Interactions Of all intermolecular interactions described for regulation of Cx43 channels so far, none occurs in a domain other than the CT (this of course excludes the interaction of a hemichannel with its partner in the other cell). Several of these intermolecular interactions are modulated by pH. Previous studies have shown that the association of the src SH3 domain with Cx43 is tighter at low pH, which may cause the separation of Cx43 from the tight junction-associated protein zona occludens-1 (ZO-1; [36]). Furthermore, NMR studies led to the characterization of the regions of Cx43CT that are structurally involved in the Cx43-ZO1 and the Cx43-src interactions. The NMR pH titration experiments determined that the second PDZ domain of ZO-1 (PDZ2) affected the structure of the last 19 Cx43CT residues, a region larger than that reported to be required for Cx43CT-ZO-1 binding [37]. Interestingly, the src SH3 domain could partially displace the Cx43CT-PDZ2 complex, thus indicating that pH gating of Cx43 is not only subject to intramolecular interactions, but also complex interactions involving other molecules. These studies represent the first structural characterization of a connexin domain when integrated in a multimolecular complex.

5.5 Modification of the Particle-Receptor Interaction by Small Peptides: The Experience with RXP-E As a key player in the regulation of gap junctions, the CT presents itself as a target for chemical or genetic manipulation intended to modify function [32,38]. Peptidic molecules have emerged as an interesting approach for development of a gap junction pharmacology. It is known that Cx43CT is capable of interacting with other proteins. This ‘‘stickiness’’ of Cx43CT could be used to adhere peptidic sequences to it, and the interaction of Cx43CT with small peptides could modify the behavior of the gap junction channel. This rationale has been applied to peptides that can modify both the chemical and voltage-gating behavior of Cx43 [39,40]. Recent studies involving high-throughput phage display screening identified a peptidic sequence that binds Cx43CT to prevent octanol-induced and acidification-induced uncoupling. This 34 aa peptide was dubbed RXP-E [41]. NMR experiments demonstrated that peptide RXP-E binds to amino acids R376, D378, and D379 of the Cx43CT. In addition, there was a minor resonance shift in amino acids 343 to 346. Interestingly, the resonance of these amino acids also shift in the presence of a peptide corresponding to the CL [17], thus suggesting that RXP-E may compete with the Cx43CL region for binding to the gating particle. Peptide-based strategies may lead to a new pharmacological approach to the study of gap junctions, based on lessons learned through structure-function studies on pH gating of Cx43.

136

R. Lewandowski et al.

The use of peptides as pharmaceutical agents is a developing field. Zestril1/Prinivil1 (an angiotensin-converting enzyme inhibitor), Zoladex1 (a gonadotropin-releasing hormone agonist), Sandostatin1 (a somatostatin mimic), Accolate1 (a leukotriene receptor antagonist), Integrilin1 (an anticoagulant that selectively blocks the platelet glycoprotein IIb/IIIa receptor), Hirudin1 (a thrombin inhibitor), and Fuzeon1 (an HIV fusion inhibitor) are all peptide-based pharmaceuticals representing a rapidly growing portion of that market. As with RXP-E, several of these were first identified by phage display. The research and therapeutic utility of peptidic modifiers of the cytoplasmic domains of proteins, such as the connexin CT, relies on the ability to deliver exogenous peptides to intracellular targets. This has been seriously limited by the inability of peptidic molecules to cross the plasma membrane barrier. Recent advances, however, have made it possible to translocate peptides into the intracellular space. Key to this strategy has been the ability to fuse the peptide of interest with a cell-penetrating peptide (CPP; [42]). In a recent study, Kim et al. [43] reported a new generation of molecules with enhanced cytoplasmic localization (hence the term cytoplasmic transduction peptide [CTP]). Overall, these data show it is possible to transduce biologically active macromolecules into living cells using CPP or CTP technology. This approach opens the door for the pharmacological manipulation of a vast number of potential molecular targets previously protected by the lipid membrane of the cell. Among these targets is the regulatory domain(s) of Cx43 and among the cargoes, RXP-E. These and other techniques will be useful to determine whether modulation of the pH regulatory mechanism of Cx43 can affect, for example, the likelihood of cardiac arrhythmias under conditions that cause acidification of the intracellular space.

5.6 Mild Acidification Increases Coupling in Cardiac Preparations That gap junctions close in response to low pH is well established. Yet, Swietach and Vaughan-Jones [44,45] recently presented elegant studies demonstrating that mild acidification can actually increase gap junction coupling. Indeed, experiments performed in myocyte pairs isolated from mammalian ventricles showed a bell-shaped pH-dependence of gap junction permeability. Using flash photolysis, the authors showed that protons (H+) that accumulate in the intracellular space are dispersed by diffusion into neighboring territories by mobility through gap junction channels. These authors put forth the hypothesis that gap junction–mediated intercellular diffusion of H+provides for spatial dissipation of moderate local rises of intracellular H+concentration into neighboring cells, thus preventing acidification of the intracellular space in a localized, purportedly afflicted, area. The more familiar acidification-induced closure of gap junctions was observed when larger acid loads were used [45]. In that case, intercellular

5 Chemical Gating

137

movement of H+ was inhibited and membrane H+-transporters acted to extrude H+ from the cell. The authors then proposed a model whereby tissues would use gap junctions to dissipate excess H+ into neighboring cells in response to mild acidification, yet, in response to a further drop in intracellular pH, gap junctions would close and isolate the stressed area from healthy cells. Further studies are likely to yield a more complete picture of the functional implications of pHdependent regulation of Cx43.

5.7 pH Regulation of Other Connexins Stergiopoulos et al. [46] used the paired Xenopus oocyte system to show that pH sensitivity varies among connexins. Differences were attributed to the diversity of the primary sequence, particularly in regulatory domain regions like the CT. Further studies have shown that some (but not all) connexins follow the basic particle-receptor model of pH gating outlined above [46]. Examples of connexins that are not affected by truncation of the CT include Cx26 (there is a very short CT in that isoform), Cx32, and Cx46. Moreover, heterodomain interactions occur between Cx40 and Cx43: the CT of Cx40 can regulate a truncated Cx43 channel, and the CT of Cx43 can also rescue the pH sensitivity of a truncated Cx40 [46]. On the other hand, Cx45 does not require the integrity of its CT for pH regulation [47]. Studies on Cx46 suggest that pH gating of that connexin involves direct protonation of the connexin molecule. Indeed, Trexler et al. [48] examined the effects of acidification on Cx46 cell–cell channels expressed in Neuro2A cells and Cx46 hemichannels expressed in Xenopus oocytes. Both hemichannels and junctional channels were sensitive to pH, and two effects were observed: (1) rapid and reversible closure was reproducibly elicited by short exposure to low pH, and (2) poorly reversible or irreversible loss of channel activity occurred with longer exposures. The former was attributed to pH gating and the latter to pH-induced inactivation. Closure by pH was found to be voltagedependent and sensitive to the same polarity with low pH applied to either side. Trexler et al. thus proposed a pH sensor was located directly on Cx46 near the pore entrance on the cytoplasmic side. This study was also significant for its identification of two different mechanisms for pH regulation directly attributable to the connexin channel. The studies of Bevans et al. [49] further elaborated on the isoform specificity of connexin pH gating, and the involvement of other molecules in the process. These authors used a liposome permeability assay to study the pH sensitivity of Cx26 or Cx32 hemichannels. pH-dependent closure was shown when hemichannels were exposed to Good’s pH buffers (e.g., 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES), 2-(N-morpholino) ethanesulfonic acid (MES)), which have an aminosulfonate moiety in common. However, pH-dependent channel gating was not seen when non-aminosulfonate buffers were used (e.g., maleate, Tris,

138

R. Lewandowski et al.

bicarbonate). At constant pH, increased aminosulfonate concentration caused hemichannel closure. Similar effects were seen with taurine, the most common cytoplasmic aminosulfonate, at physiological concentrations. Thus, the possibility arises that acidification-induced effects on these channels may actually be mediated through protonation of cytoplasmic regulators. Modulation by aminosulfonates was found to be specific for connexin channels containing Cx26, and did not occur for homomeric Cx32 channels. The identification of taurine as a cytoplasmic compound that directly interacts with connexins to modulate channel activity could lead to the development of reagents for use in structure-function studies of connexin protein [49,50,51]. More recently, the structural basis of pH gating of Cx26 has been described. Yu et al. [52] used high-resolution atomic force microscopy to force-dissect Cx26 gap junctions and image the extracellular surface of the component hemichannels to show that in HEPES buffer the pore was closed at pH <6.5 and opened reversibly by increasing the pH to pH 7.6. Corresponding to the liposome work, this effect was not observed in non-aminosulfonate buffers. Analysis of the extracellular surface topographs revealed a gradual, pH-dependent increase in the pore diameter with increase in pH over this range. The outer hemichannel diameter remained unchanged, and there was an approximately 6.5 degree rotation in lobes at the extracellular surface of the hemichannels. These transitions suggest a modification of the Cx26 pore structure during pH gating. Furthermore, the change in the structure of the Cx26 pore in response to acidification seems consistent with that previously observed in relation to Ca2+-induced closure, where Mu¨ller et al. [53] observed a reduction in the diameter of the extracellular channel entrance, as well as the formation of microdomains in the cytoplasmic side. The modifications in pore structure described by Yu et al. [52] have not yet been studied in Cx43 channels. Similar observations may result. Indeed, the particle-receptor model postulates an interaction of intracellular domains leading to channel closure, though not necessarily a particle blockade of the pore in the absence of realignment in the transmembrane domains. It is possible that the observations of Yu et al. [52] and those proposed by the NMR studies on Cx43 [21] may converge into a common model, whereby cytoplasmic domains sense and transduce a reaction that is eventually effected by pore-forming domains of the molecule. Similarly, the nature of the intracellular mediators may vary among connexins, though the ultimate consequence of pH gating is a point of convergence that extends even into the invertebrate, innexin-based gap junctions. The functional relevance of pH gating in various systems is still a matter of investigation. It is worth noting the importance that protein processing plays in the maintenance of gap junction communication in the lens (see Chapter 18). There are important spatial differences in pH gating of gap junctions in the lens, where those in the cortical fibers close upon acidification, but those in the core region do not. This difference may be the result of the endogenous cleavage of the CT of Cx50, one of the connexins forming gap junction channels in the lens. Consistent with this observation, the truncated form of

5 Chemical Gating

139

Cx50 is less pH sensitive than the full-length protein [57,58]. As such, the studies of Lin et al. [54] provide evidence linking the nonuniform gating of gap junction channels in the lens with connexin cleavage, and reveal how cells in the core region remain connected despite the acidic environment caused by elevated lactate levels in this area [54,55,56].

5.8 Conclusion Much has been learned in recent years about the pH gating of connexins. As with any other field, new explorations have yielded a few answers and numerous questions. It is interesting that amid the large variations in primary sequence, structural constraints, and even the demands for intracellular modulators, the final product (pH gating) remains a constant. Is there something particularly important to this property? So important in fact that, through evolution, it needed to be preserved even if adapted to the physical properties of the various gap junction proteins? Or, conversely, is pH gating simply a by-product of the fact that a given protein makes a gap junction? These and other questions remain to be addressed in the future. Genetic and pharmacological tools will seek to determine the physiological relevance of this process and, if possible, the consequences of preventing it. To this point, studies on pH regulation have also served as a vehicle to learn about the structure of connexins and the general way in which they modulate intercellular communication in the organism.

References 1. Gutstein DE, Morley GE, Tamaddon H, Vaidya D, Schneider MD, Chen J, Chien KR, Stuhlmann H, Fishman GI. Conduction slowing and sudden arrhythmic death in mice with cardiac-restricted inactivation of connexin43. Circ Res. 2001;88:333–9. 2. Gutstein DE, Morley GE, Vaidya D, Liu F, Chen FL, Stuhlmann H, Fishman GI. Heterogeneous expression of gap junction channels in the heart leads to conduction defects and ventricular dysfunction. Circulation. 2001;104:1194–9. 3. Danik SB, Liu F, Zhang J, Suk HJ, Morley GE, Fishman GI, Gutstein DE. Modulation of cardiac gap junction expression and arrhythmic susceptibility. Circ Res. 2004;95:1035–41. 4. Eckardt D, Theis M, Degen J, Ott T, van Rijen HV, Kirchhoff S, Kim JS, de Bakker JM, Willecke K. Functional role of connexin43 gap junction channels in adult mouse heart assessed by inducible gene deletion. J Mol Cell Cardiol. 2004;36:101–10. 5. Guerrero PA, Schuessler RB, Davis LM, Beyer EC, Johnson CM, Yamada KA, Saffitz JE. Slow ventricular conduction in mice heterozygous for a connexin43 null mutation. J Clin Invest. 1997;99:1991–8. 6. van Rijen HV, Eckardt D, Degen J, Theis M, Ott T, Willecke K, Jongsma HJ, Opthof T, de Bakker JM. Slow conduction and enhanced anisotropy increase the propensity for ventricular tachyarrhythmias in adult mice with induced deletion of connexin43. Circulation. 2004;109:1048–55. 7. Spray DC, Harris AL, Bennett MVL. Gap junctional conductance is a simple and sensitive function of intracellular pH. Science. 1981;211:712–5.

140

R. Lewandowski et al.

8. Morley GE, Taffet SM, Delmar M. Intramolecular interactions mediate pH regulation of connexin43 channels. Biophys J. 1996;70:1294–302. 9. Danik SB, Liu F, Zhang J, Suk HJ, Morley GE, Fishman GI, Gutstein DE. Modulation of cardiac gap junction expression and arrhythmic susceptibility. Circ Res. 2004;95:1035–41. 10. Armstrong CM, Bezanilla F. Inactivation of the sodium channel. II. Gating current experiments. J Gen Physiol. 1977;70:567–90. 11. Zagotta WN, Hoshi T, Aldrich RW. Restoration of inactivation in mutants of Shaker potassium channels by a peptide derived from ShB. Science. 1990;250:568–71. 12. Liu S, Taffet SM, Stoner L, Delmar M, Vallano ML, Jalife J. A structural basis for the unequal sensitivity of the major cardiac and liver gap junctions to intracellular acidification; the carboxyl tail length. Biophys J. 1993;64:1422–33. 13. Homma N, Alvarado JL, Coombs W, Stergiopoulos K, Taffet SM, Lau AF, Delmar M. A particle-receptor model for the insulin-induced closure of connexin43 channels. Circ Res. 1998;83:27–32. 14. Moorby CD, Gherardi E. Expression of a Cx43 deletion mutant in 3T3 A31 fibroblasts prevents PDGF-induced inhibition of cell communication and suppresses cell growth. Exp Cell Res. 1999;249:367–76. 15. Zhou L, Kasperek EM, Nicholson BJ. Dissection of the molecular basis of pp60(v-src) induced gating of connexin 43 gap junction channels. J Cell Biol. 1999;144:1033–45. 16. Duffy HS, Sorgen PL, Girvin ME, O’Donnell P, Coombs W, Taffet SM, Delmar M, Spray DC. pH-dependent intramolecular binding and structure involving Cx43 cytoplasmic domains. J Biol Chem. 2002;277: 36706–14. 17. Hirst-Jensen BJ, Sahoo P, Kieken F, Delmar M, Sorgen PL. Characterization of the pH-dependent interaction between the gap junction protein connexin43 carboxyl terminus and cytoplasmic loop domains. J Biol Chem. 2007;282:5801–13. 18. Shibayama J, Gutierrez C, Gonzalez D, Kieken F, Seki A, Carrion JR, Sorgen PL, Taffet SM, Barrio LC, Delmar M. Effect of charge substitutions at residue his-142 on voltagegating of connexin43 channels. Biophys J. 2006;91:4054–63. 19. Dunker AK, Brown CJ, Lawson JD, Iakoucheva LM, Obradovi Z. Intrinsic disorder and protein function. Biochemistry. 2002;41:6573–82. 20. Dunker AK, Obradovic Z. The protein trinity-linking function and disorder. Nat Biotechnol. 2001;19:805–6. 21. Sorgen PL, Duffy HS, Spray DC, Delmar M. pH-dependent dimerization of the carboxyl-terminal domain of Cx43. Biophys J. 2004;87:574–81. 22. Schorge S, Colquhoun D. Studies of NMDA receptor function and stoichiometry with truncated and tandem subunits. J Neurosci. 2003;23:1151–8. 23. Sun Y, Olson R, Horning M, Armstrong N, Mayer M, Gouaux E. Mechanism of glutamate receptor desensitization. Nature. 2002;417:245–53. 24. Galvan DL, Mignery GA. Carboxyl-terminal sequences critical for inositol 1,4,5-trisphosphate receptor subunit assembly. J Biol Chem. 2002;277: 48248–60. 25. Leach RN, Boyett MR, Findlay JB. Expression, purification and spectroscopic studies of full-length Kir3.1 channel C-terminus. Biochim Biophys Acta. 2003;1652:83–90. 26. Matulef K, Zagotta W. Multimerization of the ligand binding domains of cyclic nucleotide-gated channels. Neuron. 2002;36:93–103. 27. Ulens C, Siegelbaum SA. Regulation of hyperpolarization-activated HCN channels by cAMP through a gating switch in binding domain symmetry. Neuron. 2003;40:959–70. 28. Jin R, Gouaux E. Probing the function, conformational plasticity, and dimer-dimer contacts of the GluR2 ligand-binding core: studies of 5-substituted willardiines and GluR2 S1S2 in the crystal. Biochemistry. 2003;42:5201–13. 29. Mayer ML, Olson R, Gouaux E. Mechanisms for ligand binding to GluR0 ion channels: crystal structures of the glutamate and serine complexes and a closed apo state. J Mol Biol. 2001;311:815–36.

5 Chemical Gating

141

30. Schumacher MA, Rivard AF, Bachinger HP, Adelman JP. Structure of the gating domain of a Ca2+-activated K+ channel complexed with Ca2+/calmodulin. Nature. 2001;410:1120–4. 31. Hisamitsu T, Pang T, Shigekawa M, Wakabayashi S. Dimeric interaction between the cytoplasmic domains of the Na+/H+ exchanger NHE1 revealed by symmetrical intermolecular cross-linking and selective co-immunoprecipitation. Biochemistry. 2004;43: 11135–43. 32. Seki A, Coombs W, Taffet SM, Delmar M. Loss of electrical communication, but not plaque formation, after mutations in the cytoplasmic loop of connexin43. Heart Rhythm. 2004;1:227–33. 33. Burt JM, Spray DC. Single-channel events and gating behavior of the cardiac gap junction channel. Proc Natl Acad Sci USA. 1988;85:3431–4. 34. Bukauskas FF, Peracchia C. Two distinct gating mechanisms in gap junction channels: CO2-sensitive and voltage-sensitive. Biophys J. 1997;72:2137–42. 35. Bukauskas FF, Jordan K, Bukauskiene A, Bennett MV, Lampe PD, Laird DW, Verselis VK. Clustering of connexin 43-enhanced green fluorescent protein gap junction channels and functional coupling in living cells. Proc Natl Acad Sci USA. 2000;97:2556–61. 36. Duffy HS, Ashton AW, O’Donnell P, Coombs W, Taffet SM, Delmar M, Spray DC. Regulation of connexin43 protein complexes by intracellular acidification. Circ Res. 2004;94:215–22. 37. Sorgen PL, Duffy HS, Sahoo P, Coombs W, Delmar M, Spray DC. Structural changes in the carboxyl terminus of the gap junction protein connexin43 indicates signaling between binding domains for c-src and zonula occludens-1. J Biol Chem. 2004;279: 54695–701. 38. Maass K, Ghanem A, Kim JS, Saathoff M, Urschel S, Kirfel G, Grummer R, Kretz M, Lewalter T, Tiemann K, Winterhager E, Herzog V, Willecke K. Defective epidermal barrier in neonatal mice lacking the C-terminal region of connexin43. Mol Biol Cell. 2004;15:4597–608. 39. Calero G, Kanemitsu M, Taffet SM, Lau AF, Delmar M. A 17mer peptide interferes with acidification-induced uncoupling of connexin43. Circ Res. 1998;82:929–35. 40. Seki A, Duffy HS, Coombs W, Spray DC, Taffet SM, Delmar M. Modifications in the biophysical properties of connexin43 channels by a peptide of the cytoplasmic loop region. Circ Res. 2004;95: e22–8. 41. Shibayama J, Lewandowski R, Kieken F, Coombs W, Shah S, Sorgen PL, Taffet SM, Delmar M. Identification of a novel peptide that interferes with the chemical regulation of connexin43. Circ Res. 2006;98:1365–72. 42. Harada H, Kizaka-Kondoh S, Hiraoka M. Antitumor protein therapy; application of the protein transduction domain to the development of a protein drug for cancer treatment. Breast Cancer. 2006;13:16–26. 43. Kim D, Jeon C, Kim JH, Kim MS, Yoon CH, Choi IS, Kim SH, Bae YS. Cytoplasmic transduction peptide (CTP): new approach for the delivery of biomolecules into cytoplasm in vitro and in vivo. Exp Cell Res. 2006;312:1277–88. 44. Swietach P, Vaughan-Jones RD. Spatial regulation of intracellular pH in the ventricular myocyte. Ann NY Acad Sci. 2005;1047:271–82. 45. Vaughan-Jones RD, Spitzer KW, Swietach P. Spatial aspects of intracellular pH regulation in heart muscle. Prog Biophys Mol Biol. 2006;90:207–24. 46. Stergiopoulos K, Alvarado JL, Mastroianni M, Ek-Vitorin JF, Taffet SM, Delmar M. Hetero-domain interactions as a mechanism for the regulation of connexin channels. Circ Res. 1999;84:1144–55. 47. Francis D, Stergiopoulos K, Ek-Vitorin JF, Cao FL, Taffet SM, Delmar M. Connexin diversity and gap junction regulation by pHi. Dev Genet. 1999;24:123–36. 48. Trexler EB, Bukauskas FF, Bennett MVL, Bargiello TA, Verselis VK. Rapid and direct effects of pH on connexins revealed by the connexin46 hemichannel preparation. J Gen Physiol. 1999;113:721–42.

142

R. Lewandowski et al.

49. Bevans CG, Kordel M, Rhee SK, Harris AL. Isoform composition of connexin channels determines selectivity among second messengers and uncharged molecules. J Biol Chem. 1998;273:2808–16. 50. Bevans CG, Harris AL. Regulation of connexin channels by pH. Direct action of the protonated form of taurine and other aminosulfonates. J Biol Chem. 1999;274:3711–9. 51. Tao L, Harris AL. Biochemical requirements for inhibition of connexin26-containing channels by natural and synthetic taurine analogs. J Biol Chem. 2004;279:38544–54. 52. Yu J, Bippes CA, Hand GM, Mu¨ller DJ, Sosinsky GE. Aminosulfonate modulated pH-induced conformational changes in connexin26 hemichannels. J Biol Chem. 2007;282:8895–904. 53. Mu¨ller DJ, Hand GM, Engel A, Sosinsky GE. Conformational changes in surface structures of isolated connexin 26 gap junctions. EMBO J. 2002;21:3598–607. 54. Lin JS, Eckert R, Kistler J, Donaldson P. Spatial differences in gap junction gating in the lens are a consequence of connexin cleavage. Eur J Cell Biol. 1998;76:246–50. 55. Baldo GJ, Mathias RT. Spatial variations in membrane properties in the intact rat lens. Biophys J. 1992;63:518–29. 56. Emptage NJ, Duncan G, Croghan PC. Internal acidification modulates membrane and junctional resistance in the isolated lens of the frog Rana pipiens. Exp Eye Res. 1992;54:33–9. 57. Mathias RT, Riquelme G, Rae JL. Cell to cell communication and pH in the frog lens. J Gen Physiol. 1991;98:1085–103. 58. DeRosa AM, Mui R, Srinivas M, White TW. Functional characterization of a naturally occurring Cx50 truncation. Invest Ophthalmol Vis Sci. 2006;47:4474–81.

Chapter 6

Electrical Signaling with Neuronal Gap Junctions Barry W. Connors

Abstract Neuronal gap junctions — electrical synapses — are ubiquitous across species and brain regions. Electrical synapses typically mediate rapid, symmetrical, bidirectional flow of ionic current. Their functions can be modulated, although the mechanisms of their regulation are poorly understood. Neuronal gap junctions allow robust synchronization of subthreshold and suprathreshold electrical activity among groups of neurons, and this may be their most common function. Keywords Electrical synapses Electrical coupling Synchrony Modulation Cx35 Cx36 Cx37 Cx40 Cx45 Cx50

6.1 Introduction An excitable cell generates signals by rapidly regulating the flow of ionic current across its outer membrane. Excitable cells such as neurons and muscle fibers often need to communicate with one another, and it is hard to imagine a simpler way for them to do this than by diverting a fraction of transmembrane current directly into a neighboring cell. That is precisely what can happen when cells are coupled by gap junctions. Gap junctions interconnect the cytoplasms of two adjacent cells, and their channels are permeable to the cells’ most common small ions. When two coupled cells are in different signaling states, such that a voltage is generated across a shared gap junction, current will flow — in accordance with Ohm’s law — from one cell to the other through gap junction channels. This electrical signaling mechanism is elegant for its simplicity, speed, efficiency, and flexibility. That is probably why gap junctions are ubiquitous among species and tissues that comprise multiple excitable cells. B.W. Connors (*) Department of Neuroscience, Box G-LN, Brown University, Providence, RI 02912, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_6, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

143

144

B.W. Connors

The idea that excitable cells might communicate rapidly via direct electrical connections is almost as old as the notion of bioelectricity itself [1,2,3]. It was particularly easy to accept such a mechanism for tissues that have an obvious need for swift, powerful, and reliable excitatory signaling between cells. Cardiac muscle cells comprise such a tissue, and the concept (if not the precise molecular substrate) of intercellular electrical coupling in the heart was always fairly easy to believe [4,5]. Electrical coupling between neurons has had a rockier history. Before compelling evidence was available to settle the issue, some of the most influential voices in neuroscience wrangled over the mechanistic nature of synapses between neurons and muscles [6,7,8,9]. As the dust settled at the close of the 1950s, it was clear that some synapses use chemical neurotransmitters while some synapses are purely electrical. We are now passing the golden anniversaries of the first observations of electrical synapses between neurons in invertebrates [10,11,12] and vertebrates [13]. Since those discoveries, we have seen spectacular advances of general biological thought and technology, and our understanding of the molecular and biophysical properties of gap junctions has benefited incalculably from that progress [14,15,16,17]. Despite early evidence for electrical synapses in a few corners of the mammalian brain, the challenges of recording intracellularly from multiple central neurons placed a serious limitation on the detection of electrical synapses in mammalian brains. Only over the last decade has it become clear that gap junctional signaling among neurons is a significant feature of circuits in nearly every nucleus and division of the vertebrate central nervous system [17,18,19,20,21,22,23]. Electrical synapses may appear on dendrites, somata, axons, and terminals of specific neurons, and they may engage in a wide variety of intercellular appositions. Because neuronal gap junctions are often relatively small, located on tiny cellular structures, or remote from typical sites of microelectrode recording, they can be difficult to observe and characterize. This chapter summarizes the basic mechanisms of electrical signaling through neuronal gap junctions, describes some of the techniques for recognizing gap junctional coupling among neurons, reviews a few of the more well-studied systems of electrically coupled cells in the nervous system, and speculates a bit about the electrical functions of gap junctions. Here the focus is on electrically coupled systems of neurons in vertebrate central nervous systems, particularly those in mammalian brains.

6.2 Electrical Coupling Between Neurons The biophysical properties of electrical synapses have been extensively reviewed and discussed [18,24,25,26]. Most early studies of electrical coupling between neurons focused on systems with clear technical advantages — large cells that could be easily visualized, and that could survive impalement with multiple intracellular microelectrodes or even wires without sustaining undue damage. Such neurons were found most easily in invertebrates, although some neurons

6 Electrical Synapses

145

in the brains of fish offered similar advantages [27,28,29,30]. As investigators’ interests turned to less accommodating types of coupled neurons, such as those of the mammalian central nervous system, they relied more often on indirect and less well-controlled methods.

6.2.1 Detecting Electrical Synapses The endless variety of neuronal shapes, sizes, and anatomical locales has inspired a assortment of methods for detecting electrical synapses. Some techniques are more compelling than others, although, as always, practicalities often dictate what can and cannot be done. It is worth summarizing some of the important features of electrical synapses and general principles relevant to their detection: 1. Simultaneous recordings are best. Because the essence of an electrical synapse is that ionic current can flow from one cell to an adjoining cell through a gap junction, the most compelling means of detection is to monitor the transmembrane potential of the two putatively coupled cells simultaneously under current-clamp or voltage-clamp. The most sensitive and reliable way to do this is to use two or more intracellular (sharp or whole-cell) electrodes (Fig. 6.1), which allow measurements of both cells’ membrane potentials (with respect to the extracellular space) and the transjunctional potential. In amenable preparations it may also be possible to use other means for recording the membrane

Neuron 1 30 mV V1 gap junction

I1

4 mV Neuron 2

V2 50 ms

Fig. 6.1 Detecting electrical synapses. Paired whole-cell current clamp recordings were obtained from two low threshold-spiking inhibitory interneurons in a slice of rat neocortex. Square, depolarizing and hyperpolarizing current steps were injected into cell 1, inducing attenuated, low-pass filtered signals in cell 2. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

146

B.W. Connors

potential from multiple cells, for example, voltage-sensitive dyes. When neurons are in poorly accessible locations, particularly in vivo, simultaneous recordings may be impractical or impossible. 2. Two cells or one? Once paired-cell recordings are established, it is essential to ascertain the obvious—that the two electrodes are not monitoring the membrane of the same cell. In some systems this may not be as easy as it sounds. Exceptionally high coupling strengths may suggest that the same cell is being recorded in two places, but some cell pairs are in fact very strongly coupled by gap junctions. On the other hand, gap junctional connections with low coupling strength can be mimicked by recordings from the same cell in two electrotonically distant compartments, for example, soma and a distal dendrite or axon. Dye injections to visualize the two cells and the positions of the electrodes can be very helpful in resolving these alternatives. Pharmacological or genetic treatments known to reduce gap junctional conductance are also diagnostic [31,32] (see Chapter 8). 3. Gap junction or cytoplasmic bridge? Some cells may be connected by intercellular bridges of membrane [33], and cell damage may lead to membrane fusion. This is a particular concern when working with traumatized in vitro preparations such as brain slices. Thin membrane-bound links between neurons, in principle, can have electronic behavior that is nearly identical to that of gap junctions interconnecting thin dendrites or axons. Again, gap junction–specific treatments such as antagonists and connexin knockdown strategies can help to rule out spurious intercellular connections. Testing the permeability of cell–cell connections is also compelling. Cytoplasmic bridges should allow even very large markers (e.g., proteins and large labeled dextrans) to diffuse freely between cells. On the other hand, the permeability of gap junction channels varies widely depending on connexin composition, but most molecules larger than 1 to 2 kDa cannot pass through gap junction channels [16] (see Chapter 7). 4. Electrical synapse or chemical synapse? This is usually easy to determine in isolated preparations, where chemical synapses can be blocked selectively with a variety of manipulations, for example, antagonizing presynaptic calcium currents or postsynaptic transmitter receptors. However, this may be very difficult to achieve in vivo or with cells where access to extracellular fluids is limited. Nevertheless, most postsynaptic potentials (PSPs) and currents (PSCs) have a well-defined reversal potential. Chemical excitatory PSPs should reverse around 0 mV. In contrast, the postsynaptic events of electrical synapses should be relatively insensitive to membrane potential. Exceptions include unusual rectifying gap junctions [34], which have never been described between mammalian neurons, or situations where electrotonic coupling strength from the postsynaptic to the presynaptic cell is so high that polarizing the postsynaptic membrane influences the amplitude of the presynaptic spike. Electrical synapses in vertebrates are also bidirectional; indeed, most electrical synapses that have been carefully measured in the mammalian brain have closely symmetrical strengths [19]. The vast majority of chemical synapses are

6 Electrical Synapses

147

unidirectional, although some synapses in the olfactory bulb and retina are exceptions (though there the transmission in each direction is mediated by distinct chemical synapses) [35]. Most chemical synapses also display substantial short-term dynamics — either depression, facilitation, or a combination of both — when activated repetitively [36]; electrical synapses tend to be far less dynamic in this sense. Finally, chemical synapses generally have longer delays to the onset of PSPs compared to electrical synapses. In practice this difference is not always detectable since the delay of a chemical synapse at 378C may be only 150 msec [37], whereas the instantaneous onset of an electrical synapse may remain immeasurably small long enough that it effectively has quite a significant delay [24]. 5. Electrical synapse or ephaptic coupling? Very closely apposed membranes with no specialized contacts may interact through electrical field effects or ephaptic transmission, that is, activity-induced changes of extracellular voltage [38]. Paired-cell recordings can easily rule out an ephaptic effect since its bandpass filtering characteristic eliminates steady or low-frequency transmission entirely. Ephaptic interactions generated by activity in single cells are also likely to be exceedingly weak, and large ephaptic effects usually require precisely synchronous activity in many neighboring neurons. When it is impractical to make paired recordings, a convincing case for the presence of electrical synapses can sometimes be made with a combination of alternative techniques. Suspected electrical PSPs (ePSPs) can be recorded from single cells and assessments of their voltage-dependence, frequency-sensitivity, and susceptibility to collision tests can be carried out. In the collision test, an antidromic action potential and a suspected ePSP are evoked at close intervals to deduce whether the ePSP originated from somadendritic membrane (where most neuronal gap junctions reside) or from the axon (axonal action potentials that partially invade the soma are sometimes mistaken for ePSPs) [26]. Dye-coupling should be attempted with careful injections of low molecular weight tracers, such as neurobiotin or small fluorescent markers [39,40]. Both indirect electrophysiology and dye-coupling approaches, whenever possible, should be tested for gap junction specificity by using the most selective blockers of connexin channels (see Chapter 8), or molecular genetic manipulations to build a strong circumstantial case for bona fide electrical synapses. Of course, electron microscopic identification of gap junctions connecting the neurons in question helps considerably. Dye-coupling experiments, properly applied to well-coupled sets of cells, can provide information about spatial patterns of large neuronal networks that even the most heroic electrophysiological approach could never hope to reveal. The effectiveness of dye-coupling depends on many factors: the concentration of dye injected into the donor cell, the total number and permeability of gap junction channels, the permeability of dye across the nonjunctional membranes, the length and caliber of cytoplasmic intercellular pathways along which the dye must diffuse, the volume of multicellular cytoplasm into which the dye becomes

148

B.W. Connors

diluted, and the sensitivity of the dye visualization methods [41]. When gap junctional pathways are relatively restricted, standard dye-coupling attempts may fail even when electrophysiological methods suggest robust connections [42]. The physical factors governing dye-coupling and electrical coupling are quite different, and the abilities of each to detect gap junctional networks are not always well correlated. At its best, dye-coupling can expose the identities of the neurons engaged in gap junctional relationships, the size of cell networks, and their three-dimensional (3D) architecture in dramatic fashion [43,44,45,46].

6.2.2 General Properties of Electrical Synapses If intracellular recordings can be obtained from two cells with even modest coupling strength, then it is usually possible to estimate the biophysical properties of their shared gap junction(s) with reasonable accuracy. The most common measure of the strength of coupling is the coupling coefficient (C ), also known as the coupling ratio [24]. In practice, a long, steady, subthreshold current pulse is usually delivered to each cell in turn (Fig. 6.1), and C is defined as the ratio between steady-state voltage deflection (V) of the postjunctional cell and that of the prejunctional cell: C12 ¼ V2 =V1 C21 ¼ V1 =V2 For a pair of cells, C can be separately defined for current passing from cell 1 into cell 2 (C12) and vice versa (C21). Because the V for the noninjected cell can never be larger than the V of the injected cell (assuming the membranes are electrically passive), C will always be less than one. Some authors report C as a percentage, for example, C = 0.10 = 10%. C for the two directions across an electrical synapse need not be identical. Even if the resistance of the gap junction (Rj) is symmetrical (i.e., the same for each direction of current flow), if the input resistances (R) of the two cells differ, then C will be largest when current is injected into the cell with the lower resistance. This follows from the analysis of the paths of current flow [24], which yield the following definitions: C12 ¼ R2 =ðRj þ R2 Þ C21 ¼ R1 =ðRj þ R1 Þ Notice that R of the injected cell does not affect C, but that the R of the cell to which it is coupled does affect C. Two cells with identical input resistances and a symmetric junction, of course, will yield identical coupling coefficients. Coupling coefficient is a useful empirical measure of the functional strength of an electrical connection, but it has its shortcomings due to its dependence on

6 Electrical Synapses

149

the properties of the nonjunctional membrane [26]. Estimating the conductance of the junction depends on estimates of nonjunctional membrane properties, and it can be difficult to measure these independently. If nonjunctional membranes are electrically active (as all neuronal membranes are, in various ways), then estimates of junctional properties can be seriously distorted. When junctions are electrotonically distant from microelectrode recording sites, the cable properties of intervening cell compartments will attenuate and filter electrical signals and complicate analysis of the junction per se. A voltage clamp, properly applied, can usually yield more direct and reliable estimates of junctional properties than it is possible to obtain with current-clamp methods. Finally, if more than two cells are coupled together, then the relationship between gap junctional and input resistances is more complex than defined above. In large networks of coupled cells, current can pass from one cell to another either directly, via their shared junctions, or indirectly, via parallel pathways through other coupled cells. It is very common for single neurons to form gap junctions with multiple neighbors. For example, each interneuron in the neocortex may couple directly to dozens of other interneurons, each of which couples to dozens of others, and so on [47,48]. Estimating the electrical properties of gap junctions in such large syncytial networks is very complex, analytically and experimentally [48]. Coupling coefficients for electrical synapses can be wildly different. Some invertebrate connections resemble those between cardiac muscle cells, and seem to be adapted for reliable transmission of action potentials from one cell to another — septal junctions between giant axons in crayfish have C 0.50 [49], and similar junctions in the earthworm have C 0.90 [29]. Estimates of C for a variety of coupled neuronal systems in vertebrates tend to be much smaller. The mean C for electrical synapses among closely spaced neocortical interneurons [42,50], thalamic reticular cells [51], inferior olivary neurons [52,53], and brainstem neurons of puffer fish [27] varied from about 0.04 to 0.10, with exceptionally strong connections ranging as high as 0.40. In each of these cases, cell properties were relatively uniform, and estimates of C were similar for each direction. It is worth noting that rectifying gap junctions have not been observed between any mammalian neurons. However, asymmetric coupling coefficients are observed, as predicted, between distinctly heterologous neuron pairs with different nonjunctional electrical properties. For example, in the retina, estimates across electrical synapses in the direction from AII amacrine cells to ON bipolar cells yielded C 0.60, whereas estimates across the same junction in the opposite direction were C 0.30 [54] (see Chapter 19). The conductance of the junction itself was symmetrical. The same authors found that C was identical (0.30) when measured bidirectionally between coupled pairs of AII amacrine cells [55]. The classical coupling coefficient is a steady-state measure. Electrical transmission across gap junction–coupled neurons, however, is strongly dependent on the kinetic aspects of the signals. This is generally not due to the properties of the gap junction channels per se. Although rare types of gap junctions can have extremely rapid voltage-dependent gating kinetics [28], the vast majority, including all types described so far in mammalian systems, have very slow

150

B.W. Connors

voltage-gating. Frequency-dependent transmission comes instead from the neurons’ nonjunctional membranes, which, like all biological membranes, have high specific capacitance and variable resting conductance that combine to produce significant electrical time constants. Passive cell membranes behave as first-order low-pass filters, and the addition of nonlinear (voltage-dependent) membrane conductances can generate bandpass filtering characteristics as well [56]. The frequency-dependence of transmission across electrical synapses has been measured carefully in a few mammalian systems. Two examples are shown in Fig. 6.2, where subthreshold sine-wave currents were applied at various frequencies to pairs of neocortical interneurons of two different types [57]. Electrical coupling was strongest at low frequencies, and signals above about 10 Hz were progressively attenuated (corner frequencies were 20 to 30 Hz). In other words, the effective coupling coefficient decreased as frequency increased. Transmitted signals were also progressively phase-lagged as their frequencies increased (not shown in Fig. 6.2). The consequences of low-pass filtering for biological signals that traverse electrical synapses can be profound. Under natural conditions, the most rapid signals that most neurons experience are action potentials. In typical pairs of mammalian central neurons for which the low frequency C 0.10, the attenuation of action potentials is about tenfold greater than slow events, that is, for action potentials, C 0.01 [57]. Thus, an action potential spike in one neuron generates an ePSP of about 1 mV in its coupled neighbors, as illustrated for two types of neocortical interneurons in Fig. 6.3 [42,58]. Low-pass filtering has other effects that are less obvious yet entirely predictable. When the shape of

Fig. 6.2 Low-pass filtering across electrical synapses. Sinusoidal stimulus currents were injected into neuron 1 while recording from neurons 1 and 2. Signal attenuation fell sharply with increasing frequency. Plot of attenuation as a function of frequency (Bode plot) shows averaged data from eight pairs of fast-spiking (FS) and five pairs of low threshold-spiking (LTS) interneurons. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Adapted from Gibson et al. [57] with permission.)

6 Electrical Synapses

151

Fig. 6.3 Transformation of action potential signals across electrical synapses. Traces show presynaptic spikes (upper) and electrical postsynaptic potentials (ePSPs; middle) for two types of neocortical interneurons. ePSPs are averaged data from ten and eight neuron pairs, respectively. Lower traces show normalized action potentials and ePSPs to illustrate phase lags between them. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Adapted from Gibson et al. [57] with permission.)

an action potential is monophasic and exclusively depolarizing, it generates a monophasic, depolarizing postsynaptic event (Fig. 6.3; LTS cells). By contrast, neurons whose action potentials include exceptionally rapid depolarizing phases (with high-frequency content) and relatively large and long-lasting (low-frequency content) afterhyperpolarizations generate distinctly biphasic postsynaptic events (Fig. 6.3; FS cells) [42,50]. In these cells, the hyperpolarizing phase of the action potential is transmitted more effectively than the depolarizing phase, and repetitive presynaptic action potential spikes can lead to steady postsynaptic hyperpolarization. Clearly, it is not accurate to think of electrical synapses as simply excitatory connections. Figure 6.3 (bottom panels) also illustrates the phase lags generated between presynaptic action potential spikes and postsynaptic events. Peak-to-peak delays are about 0.5 to 1.0 ms. Some neurons have distinctly different states of excitability, and their ability to transmit signals across electrical synapses can vary accordingly. Thalamic neurons famously exhibit both a tonic-spiking mode and an intrinsic bursting mode; each spike burst rides on a relatively slow underlying voltage envelope [59]. The low-frequency content of a single intrinsic burst envelope leads to an ePSP that is fivefold larger in amplitude and 25-fold larger in integral than that generated by an individual tonic spike (Fig. 6.4) [51]. Low-pass filtering also implies that the transmission of subthreshold events, which are generally much slower than action potentials, are relatively favored by electrical synapses. A striking example of this occurs in inferior olivary neurons, which generate large, intrinsic oscillations of their subthreshold membrane potential. These excitatory cells, whose axons form the climbing fibers of the cerebellar cortex, are extensively coupled by gap junctions [60,61]. Their

152

B.W. Connors A

C

2 B

20 mV 5 40 ms

ePSP Amp(mV)

10

1

Burstlets Spikelets

5

0

1

0.0 2

0.1 0.2 0.3 0.4 Coupling Coefficient

Fig. 6.4 Differential transmission of single action potentials and bursts across electrical synapses. Recordings of single action potentials (a) and bursts (b) made from pairs of thalamic reticular neurons; note that gains of presynaptic and postsynaptic cells are different, and that presynaptic spikes are truncated. (c) Coupling coefficient and sizes of electrical postsynaptic potentials (ePSPs) are closely correlated, and amplitude of ePSPs generated by presynaptic bursts are about fivefold larger than those generated by single presynaptic spikes. (A highresolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Adapted from Long et al. [51] with permission.)

intrinsic oscillations are well transmitted from cell to cell, leading to tight subthreshold synchrony across large networks of coupled olivary neurons (Fig. 6.5, top two traces) [52,62,63]. Synaptic potentials that are generated by chemical synapses in one neuron can also be transmitted to electrically coupled neurons, with a degree of attenuation dictated mainly by their frequency content [24,64,65,66]. WT

50 mV Cx36 KO

500 ms

Fig. 6.5 Electrical synapses mediate strong synchrony of both subthreshold and suprathreshold activity. Data recorded from a pair of electrically coupled inferior olivary neurons obtained from a wild-type (WT) mouse (top), and from a Cx36 knockout (KO) mouse (bottom). (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Adapted from Long et al. [52] with permission.)

6 Electrical Synapses

153

6.3 Neuronal Gap Junction Channels and Their Potential Differences 6.3.1 Neuronal Gap Junctions and Connexin36 Cells in the vertebrate central nervous system express about half of the 21 connexin subtypes. Glia seem to be the most indiscriminate in this regard; however, neurons are apparently much more selective. Several connexins that are strongly expressed in astrocytes and oligodendrocytes are almost never detected in neurons [67,68,69] (see Chapter 15). Among neurons, there are several reasons to think that Cx36 is the most widely expressed connexin that is necessary for functional electrical coupling. First, messenger RNA (mRNA) for Cx36 is widely expressed in the mammalian brain [70]. Second, Cx36 immunoreactivity is common in gap junctions that connect neurons but not glia [71]. Single-cell reverse-transcriptase polymerase chain reaction (RT-PCR) reveals that Cx36 mRNA is often present in interneurons of hippocampus and neocortex [72]. Third, and perhaps most notably, mice with null mutations for Cx36 have little or no electrical coupling between varieties of neuron types that are normally well coupled in wild-type mice. Among the neurons that apparently depend on Cx36 for electrical coupling are several types of inhibitory interneurons in the neocortex [73,74] and hippocampus [75], various neurons of the retina [76], inhibitory cells of the thalamic reticular nucleus [77], excitatory neurons of the inferior olive [52,78], dopaminergic cells of the substantia nigra pars compacta [79], g-aminobutyric acid (GABA)ergic cells (neurons that produce GABA) of the ventral tegmental area [80], and neurons of the suprachiasmatic nucleus [81,82]. It seems likely that Cx36 channels function homotypically [83,84], although it remains possible that they can form heterotypic channels, perhaps with Cx45 [85,86].

6.3.2 Other Neuronal Gap Junction Proteins It seems increasingly probable that connexins other than Cx36 contribute to electrical synapses in the central nervous system. Among the other connexins that may account for at least some neuronal coupling are Cx40, Cx37, and Cx45 in spinal motor neurons [87,88,89,90,91,92], and Cx50 in horizontal cells of the retina [46]. Cx45, which is important for coupling some cardiac muscle cells, is also tantalizingly expressed by a wide variety of neurons in the brain [93]. Indeed, the gross anatomical patterns of expression for Cx36 and Cx45 are remarkably similar though not identical [21]. Physiological evidence that neuronal Cx45 mediates electrical signaling in neurons remains indirect and is limited to the retina. Mice with a neuron-specific deletion of Cx45 displayed a significant reduction in rod-based (scotopic) vision and a loss of dye-coupling between AII amacrine cells and ON cone bipolar cells [86] (see Chapter 19). The

154

B.W. Connors

similarity of this retinal phenotype to that of Cx36 knockout (KO) mice [76,94] hints at the possibility of Cx36-Cx45 heterotypic channels; knockout of either connexin subtype would lead to loss of function. However, it is worth emphasizing that neither Cx45-dependent electrical coupling nor functioning Cx36Cx45 heterotypic junctions have been demonstrated with direct microelectrode recordings in any system of cells. Considering the current dearth of evidence, the function of Cx45 in central neurons remains highly ambiguous. If Cx45 does form working gap junction channels between neurons, it could have interesting biophysical implications. Gap junctions composed of Cx36 are, at least in expression systems, very poorly sensitive to transjunctional voltage. With voltage deflections of about 100 mV, their macroscopic conductance is only reduced by about half, suggesting that this trait of the channel has no physiological significance [83,95]. This is consistent with observations on native junctions between central neurons, where no voltage-dependence has yet been detected [42,57]. Channels comprised of Cx45, in contrast, are among the most voltage-dependent of all connexin channels [96]. If they do form neuronal junctions, Cx45 channels may turn out to be modulated by the differential physiological states of coupled cells, and the state of an electrically coupled network could depend on the heterogeneity of its constituent cells’ activity. Connexins are not the only family of gap junction proteins expressed in neurons. Pannexins have some sequence similarity to innexins [97,98], the invertebrate gap junction proteins, but pannexins are expressed in deuterostomes, including vertebrates. Connexins, interestingly enough, bear no sequence similarity to innexins [99]. There are three known pannexin genes, and two of them are expressed in the mammalian brain, apparently by neurons [97]. So far, it has not been possible to demonstrate that pannexins can form functioning gap junctions between neurons. Pannexin1, however, can form Ca2+-permeable hemichannels on the surface membrane of some cells [97,100] (see Chapter 12). With respect to the potential roles of Cx45, pannexins and other gap junction proteins in electrical signaling among neurons, the jury is decidedly still out.

6.4 Regulation and Modulation of Neuronal Gap Junctions Although studies of the plasticity of chemical synapses are legion [101], very little is known about the regulation of electrical synapses. Available evidence implies that a variety of activity-dependent mechanisms may modulate electrical synapses, including an influence of the neurotransmitter glutamate via neighboring ionotropic or metabotropic receptors (mGluRs), effects of a variety of other metabotropic transmitter and signaling systems, activity-dependent or pathology-dependent changes in cellular pHi or intracellular Ca2+ concentration ([Ca2+]i), and alterations in connexin expression and turnover rates. Some of these topics have been reviewed recently [21,31,102,103,104].

6 Electrical Synapses

155

The most extensively studied electrical synapses in a vertebrate are those between auditory nerve endings and large reticulospinal neurons of teleost fish, which are subject to both potentiation and depression [105,106]. Potentiation can last for hours. While the molecular mechanisms of this plasticity are not entirely clear, potentiation seems to depend on the mixed nature of these synapses; Cx35containing gap junctions lie closely adjacent to presynaptic glutamatergic active zones and a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor-containing and N-methyl-D-aspartate (NMDA) receptor-containing postsynaptic densities [102,107,108]. When the synapse is activated, NMDAgated channels allow the postsynaptic entry of Ca2+, postsynaptic [Ca2+]i increases, and Ca2+/calmodulin-dependent protein kinase II is activated, which is obligatory for the simultaneous potentiation of both the electrical and chemical components of the synaptic response [109]. Interestingly, activation of dopamine receptors also enhances both electrical and chemical components of this mixed synapse [110]. Cx35 and its mammalian ortholog, Cx36, share phosphorylation consensus sites that modulate their conductance in expression systems [111], so teleost-like mechanisms may also operate at some mammalian electrical synapses. Glutamate also seems to be a key mediator of the long-lasting reduction of electrical coupling between neurons of the thalamic reticular nucleus (TRN) [112]. Brief tetanic stimulation of glutamatergic corticothalamic fibers triggers strong excitation of TRN neurons that is mediated by mGluRs [51,113]. This excitation fades within a few seconds, but in its wake the strength of electrical coupling between pairs of TRN neurons is reduced by about 20%. Synaptic depression lasts for at least 30 minutes, and it is blocked by mGluR antagonists, mimicked by mGluR agonists, and reduces the ability of coupled TRN cells to synchronize their spiking. The signaling pathways that mediate mGluR-dependent depression of thalamic electrical synapses are unknown. Many of the common central neuromodulators, including dopamine and serotonin, have been implicated in the regulation of neuronal coupling in at least some neurons of the mammalian central nervous system [114,115,116]. Very few studies have used direct measures of electrical synapses to assess their modulation, however. Instead, most have relied on the neuron-to-neuron spread of dyes injected with sharp microelectrodes, or other indirect techniques, as measures of gap junctional connections. Dye-coupling can suffer from both false positives (due to cell damage, for example) and false negatives (if dye diffusing into coupled cells cannot be detected) [117,118,119], and controlling for these can be quite difficult (see Chapter 7). Nevertheless, some studies suggest quite interesting modulatory possibilities. For example, dopamine may mediate a bidirectional regulation of coupling in the nucleus accumbens and striatum, where actions of D1 receptors decreased and D2 receptors enhanced dye-coupling [120]. In the ventral tegmental area, inhibitory neurons seem to couple when either dopamine is administered or internal capsule fibers are stimulated; the authors concluded that the threshold for this effect corresponded to that of a stimulus-induced reward mechanism [121]. Dye-coupling among neurons of the supraoptic nucleus, which is particularly

156

B.W. Connors

robust, is influenced by a range of signaling molecules, neurotransmitter receptors, and general physiological state [116]. Studies of retinal neurons offer some of the most compelling evidence for transmitter-dependent modulation of neuronal gap junctions. It has long been known that dopamine is released within the retina as ambient light levels rise, and dopamine’s activation of D1 receptors on both horizontal cells and AII amacrine cells stimulates adenylyl cyclase and the activation of cyclic adenosine monophosphate (cAMP)-dependent protein kinase A (PKA) [115]. PKA-dependent phosphorylation leads to reduced gap junctional conductance by changing its channel gating kinetics [122]. Retinal levels of nitric oxide (NO) are also modulated by light, and NO regulates gap junctions interconnecting AII amacrine cells and cone bipolar cells [123]. As ambient light increases, gap junctional coupling is reduced, and the visual receptive field of each neuron diminishes because its responsiveness to stimuli becomes more independent of neighboring cells. This is one mechanism by which the retina neatly trades off spatial acuity for sensitivity as environmental light conditions vary (see Chapter 19). Gap junctions in nonneural systems can be strongly modified by a variety of physiological mechanisms, including intracellular H+ concentration (pHi) and [Ca2+]i [124]. pHi in particular can be a potent and reversible regulator of gap junction conductance [31,125], and it is well known that intense neural activity can lead to rapid, sizable changes in pHi, in both directions [126]. There is also some suggestion that Cx36-containing channels are closed by strong intracellular acidification [84]. Despite this promising set of facts, very few data suggest that vertebrate electrical synapses are regulated by physiological alterations in pHi. Dye-coupling and electrical coupling between horizontal cells in the retina are indeed reduced when pHi is decreased and increased when pHi is raised [127,128], although pH can also have strong effects on nonjunctional membrane properties of these cells [129]. Tests of pHi on nonretinal mammalian neurons are limited to a small number of dye-coupling studies, and results are inconsistent [118,130,131]. Direct measurements of the pHi or [Ca2+]i sensitivity of vertebrate neuronal gap junctions have not been published.

6.5 Functions of Electrical Synapses Electrical synapses are ubiquitous. It now seems a good bet that they occur in large numbers in nearly all regions of the central nervous systems of all species, including mammals. Electrical synapses certainly have many distinct functions, which have been extensively discussed [18,19,20,21,23,24,26,104,114,132,133,134]. In this sense they are not much different from glutamatergic and GABAergic synapses, which perform almost as many duties as there are regions of the brain [35]. For various reasons, the roles of neuronal gap junctions are still not as well understood as those of chemical synapses. Some clues, albeit weak ones, come from the complex behavioral phenotype of Cx36KO mice. The mutants are

6 Electrical Synapses

157

missing the lion’s share of electrical coupling among neurons in nearly all parts of the brain that normally have it, and that have been examined [21,52,73,75,76,81]. They also have greatly impaired rod-based vision [76], and some less than dramatic deficits on circadian behavior [81], motor control [78,135,136], some forms of memory [137], and a brain reward system [121]. Whether compensations occur during the development of Cx36KO mice remains to be seen. It has always been helpful to highlight the ways in which electrical and chemical synapses differ [24]. An important property of any synapse is the way it transforms signals as they pass from a presynaptic to a postsynaptic cell. All electrical synapses transmit subthreshold signals in a graded, continuous manner; that is, they have analog signaling properties (Figs. 6.1, 6.2, and 6.5). Chemical synapses are more diverse in this regard. Most chemical synapses require a full presynaptic spike to trigger transmitter liberation, although the release process itself is stochastic. Graded chemical synapses, where transmitter release is a continuous function of presynaptic voltage, are well known in certain sensory neural circuits [138]. Some chemical synapses combine elements of spike-triggered and graded release processes [139]. In pyramidal cells of the hippocampus and neocortex, for example, changes in the subthreshold membrane potential of the soma can propagate electrotonically down the axon and into presynaptic terminals, where they can modulate the probability of spiketriggered transmitter release [140,141]. The recent suggestion that gap junctions couple some mammalian central axons further blurs the functional distinctions between electrical and chemical synapses [142]. Axonal gap junctions imply that subthreshold and suprathreshold signals might be able to propagate along an axon and affect not only its own presynaptic terminals, but those of nearby axons as well. The most exclusive advantage of electrical synapses is their rapid reciprocity (bidirectionality). Signal direction can switch as quickly as neurons can change their membrane potentials. When the properties of coupled cells are uniform, transmission can also be highly symmetrical. These traits make electrical synapses uniquely suitable for coordinating neural activity across its entire dynamic range. For example, the coupled pair of olivary neurons shown in Fig. 6.5 (top) not only have highly correlated subthreshold oscillations, they also have exquisitely synchronized action potentials. When their electrical synapses are absent, in this case because the gene coding for Cx36 has been knocked out, olivary cell pairs continue to generate normal oscillations and action potentials yet each cell marches to its own drummer (Fig. 6.5, bottom) [52]. Spike-synchronizing effects of gap junctions have been widely demonstrated, in a diverse variety of neurons [50,55,73,75,77,81,143,144]. Even relatively modest electrical coupling strength can be enough to mediate robust spiking synchrony [145]. These properties imply that we should not think of electrical synapses as excitatory or inhibitory, although indeed they can at times be either. It may be more appropriate to consider them as synchronizing [18], a function they do so very well. In fact, the presence of electrical synapses among

158

B.W. Connors

a population of neurons should prompt us to ask when and why they need to synchronize their activity.

6.6 Conclusion Electrical synapses are a major feature of neural circuits throughout the central nervous systems of vertebrates. That much is beyond dispute. It is now clear that Cx36 is an important component of many electrical synapses in mammals, but the roles of other connexins and pannexins are ambiguous or unknown. The mechanisms by which most electrical synapses are regulated are also obscure. Judging from what is known about nonneuronal gap junction proteins, it is likely that electrical synapses are subject to a wide assortment of controls at the levels of gene expression, activity-dependent plasticity, and neurotransmitterinduced modulation. The most interesting unanswered questions ask what roles electrical synapses play in brain function and the neural basis of behavior. Electrical synapses are particularly effective at synchronizing subthreshold activity and action potentials. They can coordinate activity across single pairs of neurons and within large local networks. Electrical synapses undoubtedly serve other roles, including, perhaps, the transmission of chemical (nonelectrical) signals. Understanding the molecular, functional, and behavioral features of electrical synapses all remain major challenges. Happily, technical innovations in neuroscience are developing apace. A broad appreciation of electrical synapses will require eclectic systems of investigation. Acknowledgments I thank my colleagues Yael Amitai, Misha Beierlein, Rebecca Burwell, Scott Cruikshank, Erika Fanselow, Jay Gibson, David Golomb, Mike Jutras, Seung-Chan Lee, Tim Lewis, Carole Landisman, Mike Long, Jaime Mancilla, Saundy Patrick, David Paul, David Pinto, and Kris Richardson for their invaluable contributions to the research described here. The research in my laboratory is supported by National Institutes of Health (NIH) grants NS25983 and NS050434.

References 1. Brazier MAB. The historical development of neurophysiology. In: Field J, editor. Handbook of Physiology Section I: Neurophysiology. 1959;1:1–58. 2. Cowan WM, Kandel ER. A brief history of synapses and synaptic transmission. In: Cowan M, Su¨dhof TC, Stevens CF, editors. Synapses. Johns Hopkins University Press, Baltimore, 2001:1–87. 3. Valenstein ES. The war of the soups and the sparks: the discovery of neurotransmitters and the dispute over how nerves communicate. New York: Columbia University Press, 2005. 4. Page E, Shibata Y. Permeable junctions between cardiac cells. Annu Rev Physiol. 1981;43:431–41. 5. Sperelakis N. Propagation mechanisms in heart. Annu Rev Physiol. 1979;41:441–57. 6. Eccles JC. Conduction and synaptic transmission in the nervous system. Annu Rev Physiol. 1948;10:93–116.

6 Electrical Synapses

159

7. Dale SHH. Transmission of effects from nerve-endings; a survey of the significance of present knowledge. New York: Oxford University Press, 1952. 8. Eccles JC. The synapse: from electrical to chemical transmission. Annu Rev Neurosci. 1982;5:325–39. 9. Fatt P. Biophysics of junctional transmission. Physiol Rev. 1954;34:674–710. 10. Furshpan EJ, Potter DD. Mechanism of nerve-impulse transmission at a crayfish synapse. Nature. 1957;180:342–3. 11. Furshpan EJ, Potter DD. Transmission at the giant motor synapses of the crayfish. J Physiol. 1959;145:289–325. 12. Watanabe A. The interaction of electrical activity among neurons of lobster cardiac ganglion. Jpn J Physiol. 1958;8:305–18. 13. Bennett MVL, Crain SM, Grundfest H. Electrophysiology of supramedullary neurons in Spheroides maculatus. Orthodromic and antidromic responses. J Gen Physiol. 1959;43:159–88. 14. Wei CJ, Xu X, Lo CW. Connexins and cell signaling in development and disease. Annu Rev Cell Dev Biol. 2004;20:811–38. 15. White TW, Paul DL. Genetic diseases and gene knockouts reveal diverse connexin functions. Annu Rev Physiol. 1999;61:283–310. 16. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys. 2001;34:325–472. 17. Barbe MT, Monyer H, Bruzzone R. Cell-cell communication beyond connexins: the pannexin channels. Physiology. 2006;21:103–14. 18. Bennett MVL, Zukin RS. Electrical coupling and neuronal synchronization in the mammalian brain. Neuron. 2004;41:495–511. 19. Connors BW, Long MA. Electrical synapses in the mammalian brain. Annu Rev Neurosci. 2004;27:393–418. 20. Hestrin S, Galarreta M. Electrical synapses define networks of neocortical GABAergic neurons. Trends Neurosci. 2005;28:304–9. 21. Sohl ¨ G, Maxeiner S, Willecke K. Expression and functions of neuronal gap junctions. Nat Rev Neurosci. 2005;6:191–200. 22. McCracken CB, Roberts DC. Neuronal gap junctions: expression, function, and implications for behavior. Int Rev Neurobiol. 2006;73:125–51. 23. Cruikshank SJ, Landisman CE, Mancilla JG, Connors BW. Connexon connexions in the thalamocortical system. Prog Brain Res. 2005;149:41–57. 24. Bennett MVLL. Electrical transmission: a functional analysis and comparison to chemical transmission. Handbook of Physiology. 1977;1:357–416. 25. Spray DC, Bennett MVL. Physiology and pharmacology of gap junctions. Annu Rev Physiol. 1985;47:281–303. 26. Jaslove SW, Brink PR. In: De Mello WC, editor. Electronic coupling in the nervous system: cell-to-cell communication. New York: Plenum Publishing; 1987. pp. 103–48. 27. Bennett MVL, Nakajima Y, Pappas GD. Physiology and ultrastructure of electrotonic junctions. Supramedullary neurons. J Neurophysiol. 1967;30:161–79. 28. Jaslove SW, Brink PR. The mechanism of rectification at the electrotonic motor giant synapse of the crayfish. Nature. 1986;323:63–5. 29. Brink P, Barr L. The resistance of the septum of the median giant axon of the earthworm. J Gen Physiol. 1977;69:517–36. 30. Faber DS, Korn H. Neurobiology of the mauthner cell. New York: Raven Press; 1978. 31. Rozental R, Srinivas M, Spray DC. How to close a gap junction channel. Efficacies and potencies of uncoupling agents. Methods Mol Biol. 2001;154:447–76. 32. Pan F, Mills SL, Massey SC. Screening of gap junction antagonists on dye coupling in the rabbit retina. Vis Neurosci. 2007;24:609–18. 33. Fawcett DW. Intercellular bridges. Exp Cell Res. 1961;8:174–87. 34. Auerbach AA, Bennett MVL. A rectifying electrotonic synapse in the central nervous system of a vertebrate. J Gen Physiol. 1969;53:211–37. 35. Shepherd GM. The synaptic organization of the brain. New York: Oxford University Press; 2004.

160

B.W. Connors

36. Zucker RS, Regehr WG. Short-term synaptic plasticity. Annu Rev Physiol. 2002;64:355–405. 37. Sabatini BL, Regehr WG. Timing of synaptic transmission. Annu Rev Physiol. 1999;61:521–42. 38. Jefferys JG. Nonsynaptic modulation of neuronal activity in the brain: electric currents and extracellular ions. Physiol Rev. 1995;75:689–723. 39. Stewart WW. Functional connections between cells as revealed by dye-coupling with a highly fluorescent naphthalimide tracer. Cell. 1978;14:741–59. 40. Hoshi H, O’Brien J, Mills SL. A novel fluorescent tracer for visualizing coupled cells in neural circuits of living tissue. J Histochem Cytochem. 2006;54:1169–76. 41. Mills SL, Massey SC. The kinetics of tracer movement through homologous gap junctions in the rabbit retina. Vis Neurosci. 1998;15:765–77. 42. Gibson JR, Beierlein M, Connors BW. Two networks of electrically-coupled inhibitory neurons in neocortex. Nature. 1999;402:75–9. 43. Gutnick MJ, Connors BW, Ransom BR. Dye-coupling between glial cells in the guinea pig neocortical slice. Brain Res. 1981;213:486–92. 44. Lo Turco JJ, Kriegstein AR. Clusters of coupled neuroblasts in embryonic neocortex. Science. 1991;252:563–6. 45. Vaney DI. Retinal neurons: cell types and coupled networks. Prog Brain Res. 2002;136:239–54. 46. O’Brien JJ, Li W, Pan F, Keung J, O’Brien J, Massey SC. Coupling between A-type horizontal cells is mediated by connexin 50 gap junctions in the rabbit retina. J Neurosci. 2006;26:11624–36. 47. Fukuda T, Kosaka T, Singer W, Galuske RA. Gap junctions among dendrites of cortical GABAergic neurons establish a dense and widespread intercolumnar network. J Neurosci. 2006;26:3434–43. 48. Amitai Y, Gibson JR, Beierlein M, Patrick SL, Ho AM, Connors BW, Golomb D. The spatial dimensions of electrically-coupled networks of interneurons in the neocortex. J Neurosci. 2002;22:4142–52. 49. Johnston MF, Ramon F. Electrotonic coupling in internally perfused crayfish segmented axons. J Physiol. 1981;317:509–18. 50. Galarreta M, Hestrin S. A network of fast-spiking cells in the neocortex connected by electrical synapses. Nature. 1999;402:72–5. 51. Long MA, Landisman CE, Connors BW. Small clusters of electrically-coupled neurons generate synchronous rhythms in the thalamic reticular nucleus. J Neurosci. 2004;24:341–9. 52. Long MA, Deans MR, Paul DL, Connors BW. Rhythmicity without synchrony in the electrically uncoupled inferior olive. J Neurosci. 2002;22:10898–905. 53. Devor A, Yarom Y. Electrotonic coupling in the inferior olivary nucleus revealed by simultaneous double patch recordings. J Neurophysiol. 2002;87:3048–58. 54. Veruki ML, Hartveit E. Electrical synapses mediate signal transmission in the rod pathway of the mammalian retina. J Neurosci. 2002;22:10558–66. 55. Veruki ML, Hartveit E. AII (Rod) amacrine cells form a network of electrically-coupled interneurons in the mammalian retina. Neuron. 2002;33:935–46. 56. Hutcheon B, Yarom Y. Resonance, oscillation and the intrinsic frequency preferences of neurons. Trends Neurosci. 2000;23:216–22. 57. Gibson JR, Beierlein M, Connors BW. Functional properties of electrical synapses between inhibitory interneurons of neocortical layer 4. J Neurophysiol. 2005;93:467–80. 58. Baker R, Llinas R. Electrotonic coupling between neurones in the rat mesencephalic nucleus. J Physiol. 1971;212:45–63. 59. Steriade M, Llinas RR. The functional states of the thalamus and the associated neuronal interplay. Physiol Rev. 1988;68:649–742. 60. Llinas R, Baker R, Sotelo C. Electrotonic coupling between neurons in cat inferior olive. J Neurophysiol. 1974;37:560–71. 61. Sotelo C, Llinas R, Baker R. Structural study of inferior olivary nucleus of the cat: morphological correlates of electrotonic coupling. J Neurophysiol. 1974;37:541–59.

6 Electrical Synapses

161

62. Leznik E, Makarenko V, Llinas R. Electrotonically mediated oscillatory patterns in neuronal ensembles: an in vitro voltage-dependent dye-imaging study in the inferior olive. J Neurosci. 2002;22:2804–15. 63. Devor A, Yarom Y. Generation and propagation of subthreshold waves in a network of inferior olivary neurons. J Neurophysiol. 2002;87:3059–69. 64. Zsiros V, Aradi I, Maccaferri G. Propagation of postsynaptic currents and potentials via gap junctions in GABAergic networks of the rat hippocampus. J Physiol. 2007;578:527–44. 65. Zsiros V, Maccaferri G. Electrical coupling between interneurons with different excitable properties in the stratum lacunosum-moleculare of the juvenile CA1 rat hippocampus. J Neurosci. 2005;25:8686–95. 66. Pereda AE, Bell TD, Faber DS. Retrograde synaptic communication via gap junctions coupling auditory afferents to the Mauthner cell. J Neurosci. 1995;15:5943–55. 67. Willecke K, Eiberger J, Degen J, Eckardt D, Romualdi A, Guldenagel M, Deutsch U, Sohl ¨ G. Structural and functional diversity of connexin genes in the mouse and human genome. Biol Chem. 2002;383:725–37. 68. Nagy JI, Rash JE. Connexins and gap junctions of astrocytes and oligodendrocytes in the CNS. Brain Res Brain Res Rev. 2000;32:29–44. 69. Nagy JI, Li X, Rempel J, Stelmack G, Patel D, Staines WA, Yasumura T, Rash JE. Connexin26 in adult rodent central nervous system: demonstration at astrocytic gap junctions and colocalization with connexin30 and connexin43. J Comp Neurol.2001;441:302–23. 70. Condorelli DF, Belluardo N, Trovato-Salinaro A, Mudo G. Expression of Cx36 in mammalian neurons. Brain Res Brain Res Rev. 2000;32:72–85. 71. Rash JE, Yasumura T, Davidson KG, Furman CS, Dudek FE, Nagy JI. Identification of cells expressing Cx43, Cx30, Cx26, Cx32 and Cx36 in gap junctions of rat brain and spinal cord. Cell Commun Adhes. 2001;8:315–20. 72. Venance L, Rozov A, Blatow M, Burnashev N, Feldmeyer D, Monyer H. Connexin expression in electrically-coupled postnatal rat brain neurons. Proc Natl Acad Sci USA. 2000;97:10260–5. 73. Deans MR, Gibson JR, Sellitto C, Connors BW, Paul DL. Synchronous activity of inhibitory networks in neocortex requires electrical synapses containing connexin36. Neuron. 2001;31:477–85. 74. Blatow M, Rozov A, Katona I, Hormuzdi SG, Meyer AH, Whittington MA, Caputi A, Monyer H. A novel network of multipolar bursting interneurons generates theta frequency oscillations in neocortex. Neuron. 2003;38:805–17. 75. Hormuzdi SG, Pais I, LeBeau FE, Towers SK, Rozov A, Buhl EH, Whittington MA, Monyer H. Impaired electrical signaling disrupts g frequency oscillations in connexin 36deficient mice. Neuron. 2001;31:487–95. 76. Deans MR, Volgyi B, Goodenough DA, Bloomfield SA, Paul DL. Connexin36 is essential for transmission of rod-mediated visual signals in the mammalian retina. Neuron. 2002;36:703–12. 77. Landisman CE, Long MA, Beierlein M, Deans MR, Paul DL, Connors BW. Electrical synapses in the thalamic reticular nucleus. J Neurosci. 2002;22:1002–9. 78. De Zeeuw CI, Chorev E, Devor A, Manor Y, Van Der Giessen RS, De Jeu MT, Hoogenraad CC, Bijman J, Ruigrok TJ, French P, Jaarsma D, Kistler WM, Meier C, Petrasch-Parwez E, Dermietzel R, Sohl G, Gueldenagel M, Willecke K, Yarom Y. ¨ Deformation of network connectivity in the inferior olive of connexin 36-deficient mice is compensated by morphological and electrophysiological changes at the single neuron level. J Neurosci. 2003;23:4700–11. 79. Vandecasteele M, Glowinski J, Venance L. Electrical synapses between dopaminergic neurons of the substantia nigra pars compacta. J Neurosci. 2005;25:291–8. 80. Allison DW, Ohran AJ, Stobbs SH, Mameli M, Valenzuela CF, Sudweeks SN, Ray AP, Henriksen SJ, Steffensen SC. Connexin-36 gap junctions mediate electrical coupling between ventral tegmental area GABA neurons. Synapse. 2006;60:20–31. 81. Long MA, Jutras MJ, Connors BW, Burwell RD. Electrical synapses coordinate activity in the suprachiasmatic nucleus. Nat Neurosci. 2005;8:61–6.

162

B.W. Connors

82. Rash JE, Olson CO, Pouliot WA, Davidson KG, Yasumura T, Furman CS, Royer S, Kamasawa N, Nagy JI, Dudek FE. Connexin36 vs. connexin32, ‘miniature’ neuronal gap junctions, and limited electrotonic coupling in rodent suprachiasmatic nucleus. Neuroscience. 2007;149:350–371. 83. Al-Ubaidi MR, White TW, Ripps H, Poras I, Avner P, Gomes D, Bruzzone R. Functional properties, developmental regulation, and chromosomal localization of murine connexin36, a gap-junctional protein expressed preferentially in retina and brain. J Neurosci Res. 2000;59:813–26. 84. Teubner B, Degen J, Sohl ¨ G, Guldenagel M, Bukauskas FF, Trexler EB, Verselis VK, De Zeeuw CI, Lee CG, Kozak CA, Petrasch-Parwez E, Dermietzel R, Willecke K. Functional expression of the murine connexin 36 gene coding for a neuron-specific gap junctional protein. J Membr Biol. 2000;176:249–62. 85. Dedek K, Schultz K, Pieper M, Dirks P, Maxeiner S, Willecke K, Weiler R, Janssen-Bienhold U. Localization of heterotypic gap junctions composed of connexin45 and connexin36 in the rod pathway of the mouse retina. Eur J Neurosci. 2006;24:1675–86. 86. Maxeiner S, Dedek K, Janssen-Bienhold U, Ammermuller J, Brune H, Kirsch T, Pieper M, Degen J, Kruger O, Willecke K, Weiler R. Deletion of connexin45 in mouse retinal neurons disrupts the rod/cone signaling pathway between AII amacrine and ON cone bipolar cells and leads to impaired visual transmission. J Neurosci. 2005;25:566-76. 87. Bittman KS, Panzer JA, Balice-Gordon RJ. Patterns of cell-cell coupling in embryonic spinal cord studied via ballistic delivery of gap-junction-permeable dyes. J Comp Neurol. 2004;477:273–85. 88. Chang Q, Balice-Gordon RJ. Gap junctional communication among developing and injured motor neurons. Brain Res Brain Res Rev. 2000;32:242–9. 89. Chang Q, Gonzalez M, Pinter MJ, Balice-Gordon RJ. Gap junctional coupling and patterns of connexin expression among neonatal rat lumbar spinal motor neurons. J Neurosci. 1999;19:10813–28. 90. Chang Q, Pereda A, Pinter MJ, Balice-Gordon RJ. Nerve injury induces gap junctional coupling among axotomized adult motor neurons. J Neurosci. 2000;20:674–84. 91. Personius K, Chang Q, Bittman K, Panzer J, Balice-Gordon R. Gap junctional communication among motor and other neurons shapes patterns of neural activity and synaptic connectivity during development. Cell Commun Adhes. 2001;8:329–33. 92. Personius KE, Chang Q, Mentis GZ, O’Donovan MJ, Balice-Gordon RJ. Reduced gap junctional coupling leads to uncorrelated motor neuron firing and precocious neuromuscular synapse elimination. Proc Natl Acad Sci USA. 2007;104:11808–13. 93. Maxeiner S, Kruger O, Schilling K, Traub O, Urschel S, Willecke K. Spatiotemporal transcription of connexin45 during brain development results in neuronal expression in adult mice. Neuroscience. 2003;119:689–700. 94. Guldenagel M, Ammermuller J, Feigenspan A, Teubner B, Degen J, Sohl ¨ G, Willecke K, Weiler R. Visual transmission deficits in mice with targeted disruption of the gap junction gene connexin36. J Neurosci. 2001;21:6036–44. 95. Srinivas M, Rozental R, Kojima T, Dermietzel R, Mehler M, Condorelli DF, Kessler JA, Spray DC. Functional properties of channels formed by the neuronal gap junction protein connexin36. J Neurosci. 1999;19:9848–55. 96. Moreno AP, Laing JG, Beyer EC, Spray DC. Properties of gap junction channels formed of connexin 45 endogenously expressed in human hepatoma (SKHep1) cells. Am J Physiol. 1995;268:C356–65. 97. Bruzzone R, Hormuzdi SG, Barbe MT, Herb A, Monyer H. Pannexins, a family of gap junction proteins expressed in brain. Proc Natl Acad Sci USA. 2003;100:13644–9. 98. Yen MR, Saier MH Jr. Gap junctional proteins of animals: the innexin/pannexin superfamily. Prog Biophys Mol Biol. 2007;94:5–14. 99. Phelan P. Innexins: members of an evolutionarily conserved family of gap-junction proteins. Biochim Biophys Acta. 2005;1711:225–45.

6 Electrical Synapses

163

100. Huang Y, Grinspan JB, Abrams CK, Scherer SS. Pannexin1 is expressed by neurons and glia but does not form functional gap junctions. Glia. 2007;55:46–56. 101. Malenka RC, Bear MF. LTP and LTD: an embarrassment of riches. Neuron. 2004;44:5–21. 102. Pereda AE, Rash JE, Nagy JI, Bennett MVL. Dynamics of electrical transmission at club endings on the Mauthner cells. Brain Res Brain Res Rev. 2004;47:227–44. 103. Axelsen LN, Haugan K, Stahlhut M, Kjolbye AL, Hennan JK, Holstein-Rathlou NH, Petersen JS, Nielsen MS. Increasing gap junctional coupling: a tool for dissecting the role of gap junctions. J Membr Biol. 2007;216:23–35. 104. Hormuzdi SG, Filippov MA, Mitropoulou G, Monyer H, Bruzzone R. Electrical synapses: a dynamic signaling system that shapes the activity of neuronal networks. Biochim Biophys Acta. 2004;1662:113–37. 105. Yang XD, Connor JA, Faber DS. Weak excitation and simultaneous inhibition induce long-term depression in hippocampal CA1 neurons. J Neurophysiol. 1994;71:1586–90. 106. Pereda AE, Faber DS. Activity-dependent short-term enhancement of intercellular coupling. J Neurosci. 1996;16:983–92. 107. Rash JE, Pereda A, Kamasawa N, Furman CS, Yasumura T, Davidson KG, Dudek FE, Olson C, Li X, Nagy JI. High-resolution proteomic mapping in the vertebrate central nervous system: close proximity of connexin35 to NMDA glutamate receptor clusters and co-localization of connexin36 with immunoreactivity for zonula occludens protein-1 (ZO-1). J Neurocytol. 2004;33:131–51. 108. Smith M, Pereda AE. Chemical synaptic activity modulates nearby electrical synapses. Proc Natl Acad Sci USA. 2003;100:4849–54. 109. Pereda AE, Bell TD, Chang BH, Czernik AJ, Nairn AC, Soderling TR, Faber DS. Ca2+/calmodulin-dependent kinase II mediates simultaneous enhancement of gapjunctional conductance and glutamatergic transmission. Proc Natl Acad Sci USA. 1998;95:13272–7. 110. Pereda A, Triller A, Korn H, Faber DS. Dopamine enhances both electrotonic coupling and chemical excitatory postsynaptic potentials at mixed synapses. Proc Natl Acad Sci USA. 1992;89:12088–92. 111. Mitropoulou G, Bruzzone R. Modulation of perch connexin35 hemi-channels by cyclic AMP requires a protein kinase A phosphorylation site. J Neurosci Res. 2003;72:147–57. 112. Landisman CE, Connors BW. Long-term modulation of electrical synapses in the mammalian thalamus. Science. 2005;310:1809–13. 113. McCormick DA, von Krosigk M. Corticothalamic activation modulates thalamic firing through glutamate ‘metabotropic’ receptors. Proc Natl Acad Sci USA. 1992;89:2774–8. 114. Roerig B, Feller MB. Neurotransmitters and gap junctions in developing neural circuits. Brain Res Brain Res Rev. 2000;32:86–114. 115. Weiler R, Pottek M, He S, Vaney DI. Modulation of coupling between retinal horizontal cells by retinoic acid and endogenous dopamine. Brain Res Brain Res Rev. 2000;32:121–9. 116. Hatton GI. Synaptic modulation of neuronal coupling. Cell Biol Int. 1998;22:765–80. 117. Knowles WD, Funch PG, Schwartzkroin PA. Electrotonic and dye coupling in hippocampal CA1 pyramidal cells in vitro. Neuroscience. 1982;7:1713–22. 118. Connors BW, Benardo LS, Prince DA. Carbon dioxide sensitivity of dye coupling among glia and neurons of the neocortex. J Neurosci. 1984;4:1324–30. 119. Gutnick MJ, Lobel-Yaakov R, Rimon G. Incidence of neuronal dye-coupling in neocortical slices depends on the plane of section. Neuroscience. 1985;15:659–66. 120. O’Donnell P, Grace AA. Dopaminergic modulation of dye coupling between neurons in the core and shell regions of the nucleus accumbens. J Neurosci. 1993;13:3456–71. 121. Lassen MB, Brown JE, Stobbs SH, Gunderson SH, Maes L, Valenzuela CF, Ray AP, Henriksen SJ, Steffensen SC. Brain stimulation reward is integrated by a network of electrically-coupled GABA neurons. Brain Res. 2007;1156:46–58. 122. McMahon DG, Knapp AG, Dowling JE. Horizontal cell gap junctions: single-channel conductance and modulation by dopamine. Proc Natl Acad Sci USA. 1989;86:7639–43.

164

B.W. Connors

123. Bloomfield SA. Plasticity of AII amacrine cell circuitry in the mammalian retina. Prog Brain Res. 2001;131:185–200. 124. Spray DC, Stern JH, Harris AL, Bennett MVL. Gap junctional conductance: comparison of sensitivities to H and Ca ions. Proc Natl Acad Sci USA. 1982;79:441–5. 125. Spray DC, Harris AL, Bennett MVL. Gap junctional conductance is a simple and sensitive function of intracellular pH. Science. 1981;211:712–5. 126. Chesler M. Regulation and modulation of pH in the brain. Physiol Rev. 2003; 83:1183–221. 127. Qian H, Malchow RP, Ripps H. Gap-junctional properties of electrically-coupled skate horizontal cells in culture. Vis Neurosci. 1993;10:287–95. 128. DeVries SH, Schwartz EA. Modulation of an electrical synapse between solitary pairs of catfish horizontal cells by dopamine and second messengers. J Physiol. 1989;414:351–75. 129. Jouhou H, Yamamoto K, Iwasaki M, Yamada M. Acidification decouples gap junctions but enlarges the receptive field size of horizontal cells in carp retina. Neurosci Res. 2007;57:203–9. 130. Church J, Baimbridge KG. Exposure to high-pH medium increases the incidence and extent of dye coupling between rat hippocampal CA1 pyramidal neurons in vitro. J Neurosci. 1991;11:3289–95. 131. Rorig B, Klausa G, Sutor B. Intracellular acidification reduced gap junction coupling between immature rat neocortical pyramidal neurones. J Physiol. 1996;490:31–49. 132. Bennett MVL. Electrical synapses, a personal perspective (or history). Brain Res Brain Res Rev. 2000;32:16–28. 133. Bruzzone R, Dermietzel R. Structure and function of gap junctions in the developing brain. Cell Tissue Res. 2006;326:239–48. 134. Marder E. Electrical synapses: beyond speed and synchrony to computation. Curr Biol. 1998;8:R795–7. 135. Placantonakis DG, Bukovsky AA, Aicher SA, Kiem HP, Welsh JP. Continuous electrical oscillations emerge from a coupled network: a study of the inferior olive using lentiviral knockdown of connexin36. J Neurosci. 2006;26:5008–16. 136. Placantonakis DG, Bukovsky AA, Zeng XH, Kiem HP, Welsh JP. Fundamental role of inferior olive connexin 36 in muscle coherence during tremor. Proc Natl Acad Sci USA. 2004;101:7164–9. 137. Frisch C, De Souza-Silva MA, Sohl ¨ G, Guldenagel M, Willecke K, Huston JP, Dere E. Stimulus complexity dependent memory impairment and changes in motor performance after deletion of the neuronal gap junction protein connexin36 in mice. Behav Brain Res. 2005;157:177–85. 138. Sterling P, Matthews G. Structure and function of ribbon synapses. Trends Neurosci. 2005;28:20–9. 139. Marder E. Neurobiology: extending influence. Nature. 2006;441:702–3. 140. Hamzei-Sichani F, Kamasawa N, Janssen WG, Yasumura T, Davidson KG, Hof PR, Wearne SL, Stewart MG, Young SR, Whittington MA, Rash JE, Traub RD. Gap junctions on hippocampal mossy fiber axons demonstrated by thin-section electron microscopy and freeze-fracture replica immunogold labeling. Proc Natl Acad Sci USA. 2007;104:12548–53. 141. Alle H, Geiger JR. Combined analog and action potential coding in hippocampal mossy fibers. Science. 2006;311:1290–3. 142. Shu Y, Hasenstaub A, Duque A, Yu Y, McCormick DA. Modulation of intracortical synaptic potentials by presynaptic somatic membrane potential. Nature. 2006;441:761–5. 143. Beierlein M, Gibson JR, Connors BW. A network of electrically-coupled interneurons drives synchronized inhibition in neocortex. Nat Neurosci. 2000;3:904–10. 144. Christie JM, Bark C, Hormuzdi SG, Helbig I, Monyer H, Westbrook GL. Connexin36 mediates spike synchrony in olfactory bulb glomeruli. Neuron. 2005;46:761–72. 145. Mancilla JG, Lewis TJ, Pinto DJ, Rinzel J, Connors BW. Synchronization of electrically-coupled pairs of inhibitory interneurons in neocortex. J Neurosci. 2007;27: 2058–73.

Chapter 7

Permeability of Connexin Channels Andrew L. Harris and Darren Locke

Abstract Because of the diversity of connexin isoforms and their combinations that can compose connexin channels, there is large diversity in the permeability properties of the channels. Unlike most ion channels, the relevant permeabilities extend from charge selectivity among atomic ions such as Kþ and Ca2þ, through the size and charge selectivity among nonbiological tracer molecules, to highly specific selectivities among cytoplasmic molecules. Distinct experimental approaches are used to define each of these types of selectivity. In general, permeability to current-carrying atomic ions is high, but is substantially charge-selective in some cases. In general, there is permeability to molecular tracers up to 800 Da, with some channels having significantly lower cutoff thresholds, influenced by charge. Permeability to cytoplasmic molecules is less well explored, but seems to be highly variable and specific, often violating the size and charge selectivities suggested by studies using nonbiological tracers. The unitary channel conductances and the molecular permeabilities to tracer molecules and to cytoplasmic molecules do not correlate well with each other, and do not allow easy inferences or extrapolations about what biological molecules can permeate any particular form of connexin channel, and how well. In spite of this, even the sparse data obtained to date on cytoplasmic permeants gives clues as to the roles that connexin-specific permeabilities may play in biology. Keywords Connexin Gap junction Unitary conductance Pore width Charge selectivity Tracer permeability Molecular permeability Second messengers

A.L. Harris (*) Department of Pharmacology and Physiology, New Jersey Medical School of the University of Medicine and Dentistry of New Jersey, 185 South Orange Avenue, Newark, NJ 07103, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_7, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

165

166

A.L. Harris and D. Locke

7.1 Introduction Connexin channels are like other channels in having an aqueous pore through which ions can flow. However, they differ from most other membrane channels in not being uniquely defined by the selectivity of the pore, as are Kþ channels and Ca2þ channels, for example. This has created a puzzle for the study of the connexin channels: the ability to form a pore is obviously a key function of the protein, yet, aside from mediating electrical coupling, the connexin pores’ properties and precisely what goes through them remain largely cryptic. In addition, connexin channels exist in a multitude of compositions and several structural forms, and the junctional form presents significant challenges for experimentation. In spite of this, there has been a rich and inventive exploration of the properties of connexin channels, often drawing on electrophysiological techniques and supplementing them with novel and creatively adapted approaches. A variety of approaches have been applied to investigation of the permeability properties of connexin channels, well reviewed in Verselis and Veenstra [1]. Part of the intrigue about connexin channel permeability arises from the multiple perspectives and interests that can be brought to bear on the pore properties. The pores carry current, are permeable to ions, can be occupied by multiple ions of similar and different charge, and are selectively permeable to nonbiological and biological molecules. Key questions relate to the properties and mechanisms of the electrical conductance, ionic selectivity, size-selectivity, and selectivity among tracer and cytoplasmic molecules. This chapter reviews the fundamental issues and literature on these topics.

7.2 Permeability to Atomic Ions: Electrical Conductance The most easily quantifiable measure of the permeability of connexin channels is electrical conductance. Its measurement involves standard electrophysiological techniques, often applied in novel configurations. At its most basic level, the measurement involves the imposition of a known voltage (V) across the channel and a measurement of the current (I; rate of ion flux) that flows through the channel in response to that voltage. The conductance (G) is the proportionality between the voltage and the current (I = VG); the greater the conductance, the greater the rate of ion flux through the pore at a given voltage. In most situations, the ions that carry the current are Kþ and Cl–. Therefore, the electrical conductance reflects the ease or difficulty with which these small atomic ions can pass through the lumen of the channel, driven by the imposed voltage. The unit of channel conductance is typically picoSiemens (pS; 10–12 Siemens, where a Siemens is the conductance through a one Ohm resistor). Several factors can affect the rate of ion flow through a pore. In general, the wider a pore, the greater the conductance, since the ions can pass through with less interaction with each other or with the walls of the pore. In general, the

7 Permeability

167

longer a pore, the lower the conductance, since more energy is required to travel the longer distance. If the shape of the pore is a featureless cylinder, the conductance will increase as a function of the square of the radius (that is, as a function of the cross-sectional area of the lumen) and will decrease as a linear function of the length. For example, a short and wide cylindrical pore will have a higher conductance than one that is narrow and long, but a pore that is narrow and short may have a conductance similar to that of a pore that is wide and long. The lumens of most biological channels, including connexin channels, are not featureless cylinders, but have a complex internal topography determined by the various structural elements of the proteins and the amino acid side-chains that protrude into the lumen. For these reasons, it is important to recognize that electrical conductance tells us only about the overall ability the ions to move through the pore, and cannot be used to infer specific aspects of the shape or dimensions of the pore, in the absence of other information. A common inference is that a high conductance implies a wide pore and therefore permeability to large molecules. From the above example, one can see that high conductance could arise from a narrow pore that is very short, or from a pore that is wide over almost all of its length, but with a very short region that is quite narrow (and would therefore have little effect on the overall conductance). In spite of the inability to infer structural properties, electrical conductance is important in the context of electrical signaling; fewer connexin channels with high conductance can mediate the same degree of electrical coupling as can more channels with lesser conductance. Another factor that can affect the conductance of a channel is the availability of the permeant ions; if the concentration of the ions is greater or smaller, the measured current will be proportionally greater or smaller, and thus can skew the calculated conductance unless normalized to a standard ionic composition. Fortunately, almost all measurements of connexin channels are made under similar, somewhat physiological ionic conditions, with typically 120 to 150 mM salt, and so are largely comparable. In cases where the salt concentration is experimentally manipulated to be higher or lower, or the ions have different diffusional mobilities (e.g., when Cl– is replaced by glutamate), the resulting conductance values must be corrected if they are to be compared with other data. As a corollary to this, if a channel is permeable, for example, to Kþ but not to Cl–, then for a given KCl concentration, the concentration of ions available to carry the current through the pore is half of what would be the case if the channel were permeable to both ions. These kinds of considerations hint at the complexities that can be involved in comparisons of the electrical conductance of connexin channels. The conductance of single connexin channels, referred to as unitary conductance, is measured by the voltage clamp technique, in which a specified voltage is imposed across a channel and the resulting current is measured. For junctional channels, it is necessary to control the voltage in both of the coupled cells. Usually one cell is kept near its native membrane potential and the voltage in the other cell is stepped to a different value (see Chapter 4). In most cases, cells

168

A.L. Harris and D. Locke

are coupled by many hundreds of junctional channels, so to record the unitary conductance one must reduce the number of functional channels so that the opening and closing of single-channels can be resolved. This is usually done by exposure to the gap junction inhibitor halothane or other agents, which affect gating of the channels (e.g., the frequency with which they open) but not their permeability properties when open (see Chapter 8). In some cases, the level of coupling is so low that a pharmacological treatment is not required, as if often the case for Neuro2A cells transfected to express connexins. The unitary conductance of hemichannels can be recorded by electrically isolating a small patch of plasma membrane, using a glass pipette, which contains one or a few functional channels. A voltage is imposed across the patch and the change in current measured as the channels open and close. The actual value of the voltage across the membrane patch and the channel(s) is known only if a separate electrode is used to monitor the membrane potential of the cell. In other cases, the patch of membrane can be excised from the cell into the extracellular medium while adhering tightly across the opening of the patch pipette. In this case, the voltage control is more certain, as are the concentrations of the permeant ions. A sampling of unitary conductances of homomeric connexin channels, measured in approximately physiological salines, is shown in Table 7.1. Table 7.1 Unitary conductances of homomeric connexin channels. The data are from human or rodent orthologs unless otherwise noted, and obtained from cells transfected to express a single connexin. The conductances of subconductance states are noted in parentheses. Some of the variability is due to use of different cations (e.g., Csþ versus Kþ) in the different studies Unitary conductance (pS) Junctional channel conductance followed by substate conductance in parentheses. ‘‘hemi’’ indicates hemichannel Human/mouse conductance. Cx23 Cx25 Cx26

Cx30

Cx30.2-Cx31.3/ Cx29 Cx30.3 Cx31 Cx31.1 Cx31.9/Cx30.2 Cx32

n.a. n.a. 140(25) 135 102(17) 95 179(48) 141(21) 135(20) 352 hemi n.a. n.a. 85 n.a. 9 20 hemi 45(19)

[24] [25] [210] [64] [55] [211] [64] [212]

[213] [214] [215] [24]

7 Permeability

169

Table 7.1 (continued) Unitary conductance (pS) Junctional channel conductance followed by substate conductance in parentheses. ‘‘hemi’’ indicates hemichannel Human/mouse conductance.

Cx36 Cx37

Cx40

Cx40.1/Cx39 Cx43

Cx45

Cx46

Cx47 Cx50

Cx59 Cx62/Cx57

53 70(10–25) 10–15 14–15 300(63) 347 300 180 198(36) 175(30) 142 n.a. 60 100(60) 90(60) 90(60) 60(40) 80 120(30) 26/19 32 32 37(23) 57 hemi 140 140 140(10–60) 152(28) 300 hemi 300 hemi 250 hemi 55(8) 220(43) 200 212 352 hemi n.a. 27

[25] [20] [216] [217] [218] [28] [219] [28] [220] [221] [6] [222] [223] [224] [225] [226] [28] [66] [227] [28] [228] [229] [230] [231] [232] [233] [234] [4] [235] [212] [236] [237] [238] [239] [212] [240]

7.3 Permeability to Atomic Ions: Charge Selectivity Since by definition the current carrying ions are charged, their ability to enter and traverse the pore will be affected by the electrical forces they experience — the electrical field, as mentioned above, as well as any local electrostatic effects

170

A.L. Harris and D. Locke

that arise from partial and formal charges in the protein segments that form the pore or that impinge on the lumen from elsewhere in the protein. In broad strokes, these can be defined as charge experienced near the mouth of the pore, which affect ion local concentration and entry into the pore, and those within the pore itself, which affect the ability to permeate once inside. A different sort of selectivity among permeants occurs for classical ion-selective channels, in which there is a discrete thermodynamic binding site within the pore (the selectivity filter) at which the ion must bind in order to permeate; ions that do not satisfy the specific electrostatic and spatial constraints of that site do not occupy it and therefore cannot pass that point in the pore. For this to occur, the channel must narrow at the selectivity filter to the extent that the permeating ion comes into very close apposition with the walls of the pore. Because connexin channels are generally permeable to molecules at least 8 to 10 A˚ in minimal diameter, it is unlikely that the permeability to atomic ions through connexin channels is controlled by this type of mechanism (however, such a mechanism is possible for molecular permeants; see below). Measurement of connexin channel selectivity among current-carrying ions has been largely focused on the degree of charge selectivity — the relative ease with which positively charged or negatively charged atomic ions permeate the channels. This information can be obtained by several experimental approaches, all of which require some degree of control over the ionic species that are available to pass through the channels or of their concentrations. For junctional channels one must control the ionic environment inside the coupled cells. This can be achieved by the use of patch electrodes that allow the intraelectrode solution to exchange with the cytoplasms of the two cells. Once this is achieved, one can measure the junctional conductance under the altered ionic conditions and make inferences about the selectivity from the difference from that observed under normal conditions. Such measurements must take into account the different mobilities of the various ions. These measurements are technically difficult and a variety of controls and details must be incorporated into the analysis (an excellent description of this kind of experiment and the factors that must be taken into account is in [2]). For hemichannels, the ability to excise patches of membrane means that the ionic composition can be directly controlled. In this configuration, selectivity can be inferred from conductance measurements as well as reversal potential (Erev) measurements (permeability ratio). Of these two measures, Erev, being an equilibrium measurement, in theory gives the most accurate reflection of the interactions of the ions with the interior of the pore, subject to certain assumptions. Both measurements can be affected by interactions between the permeating ions and by Donnan potentials produced by the interactions of fixed charges in the pore with the permeants [3]. In most cases, the relative permeability of Kþ to Cl– is taken as the index of charge selectivity. Prior to such measurements, it was presumed that connexin channels would not have significant charge selectivity, due to the pores being so wide. The data

7 Permeability

Human/ mouse Cx26 Cx32

171 Table 7.2 Charge selectivity to atomic ions Conductance Permeability ratio cation/anion ratio cation/anion 2.6 2.6 0.94 0.94 0.77

Cx37 Cx40

2.3 3.4 4.5 6.9 6.2

Cx43

1.3 3.1 (chick) 1.0 1.0

Cx45 Cx46

10.0 (chick)

Cx50

2.4

7.0 10.1 hemi 10.3 hemi

[25] [70] [25] [70] [20] [218] [70] [70] [6] [241] [70] [70] [232] [62] [70] [232] [231] [4] [237]

show a surprisingly wide range of charge selectivities. Table 7.2 is a sampling of the relevant data. For a few connexin channels, there are detailed data regarding the relative permeabilities to various atomic ions of the same charge, for example, Kþ versus Naþ. This information can be informative regarding the energetics that influence an ion as it permeates. In some cases the selectivity sequences of connexin channels are unremarkable, accounted for by aqueous mobility or by simple interaction with a charged site within the pore, for example, as for Cx46 [4]. However, in other cases, such as Cx43 and Cx40, the selectivity sequences suggest that selectivity is influenced by interaction between anions and cations with each other within the pore or with the pore walls [5,6,7]. It has been suggested in these cases that anion permeation reduces the cation permeation. These kinds of effects violate the idea of independence of movement of each ionic species that underlies most of the commonly applied analytic formalisms for ion channel permeation, and may be a consequence of the fact that both anions and cations can enter the pore, unlike for most other ion channels. In fact, these sorts of effects have been noted in certain other channels that allow both cation and anion entry [8,9,10]. It is clear from Tables 7.1 and 7.2 that the degree of charge selectivity has little correlation with the unitary conductance (Fig. 7.1). The vastly different unitary conductances and charge selectivities, as well as their lack of correlation, suggest that the pores of connexin channels have diverse properties — structural and electrostatic.

172

A.L. Harris and D. Locke

10.0

10.3 10.1

conductance - cation/anion ratio permeability - cation/anion ratio

7.00

6.90 6.20

4.50 3.40

3.10

2.60

2.40

2.30

1.30

1.00

1.00

~1.00

1.06 1.30 0

30

50

Cx45 Cx32

90 Cx43

140 140

180

Cx26 Cx46

Cx40 Cx50

200

300

pS

Cx37

Fig. 7.1 Charge selectivity among atomic ions as a function of unitary conductance. The selectivity data in Table 7.2 are plotted against the corresponding unitary conductances. Bars above the 1.00 line represent cation selectivity, and the bars below the line represent anion selectivity. Open bars are selectivities derived from conductance ratios. Filled bars are selectivities derived from reversal potential measurements (permeability ratios). (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

7.4 Pore Width The role of connexin channels in molecular signaling between cells depends crucially on the limiting pore width of the channel — the diameter of the pore lumen at its narrowest point. We have seen above how the unitary conductance cannot be used as a reliable indicator of pore width. The simplest strategy to determine limiting pore width is to assess the ability of molecules of different sizes to permeate. However, since permeant charge can play a role, pore width is best assessed using uncharged molecules. In addition, since the pore lumens are unlikely to be featureless cylinders, molecular shape and the chemical interactions of the molecular probes with the pore walls are also likely to affect permeation. For this reason, it is difficult to obtain sizing information that is not contaminated by effects of permeant charge, shape, and chemical interaction. A few studies have used sets of highly homologous uncharged, relatively unreactive molecules (e.g., sugars, polyethyleneglycols) to directly assess the limiting pore width of connexin channels. However, even for these, one cannot

7 Permeability

173

exclude the possibility that the molecules have interactions with the walls of the pore that affect permeation, as occurs between maltose and the maltoporin channel [11,12]. However, with this caveat aside, the use of uncharged permeants has been informative (for a summary see [13]). In these studies, one cannot determine the actual distance between the van der Waals surfaces at the narrowed point, due to hydration, coordination, molecular flexibility, and so on, but one can say that a molecule of a particular van der Waals dimension or hydrodynamic radius can permeate or not. One set of studies utilized a set of 1!4-linked glucose sugars (maltosaccharides) derivatized with a small uncharged fluorescent moiety [14,15]. The 1!4 linkages between the saccharide units produce a rigid helical structure so that each additional saccharide (up to six) increases the minimal cross-sectional diameter. Use of these agents showed that homomeric Cx26 and heteromeric Cx26/Cx32 hemichannels were narrower than homomeric Cx32 channels, which were narrower than homomeric Cx43 channels. For these channels, the largest sugar permeable was the disaccharide, trisaccharide, and tetrasaccharide, respectively. This characterization of relative widths was confirmed and extended for the Cx26 and/or Cx32 channels by studies using cyclodextrins, cyclized 1!4-linked glucose sugars [16]. Another class of studies assessed the ability of polyethyleneglycols (PEGs) of different sizes to affect channel conductance. The idea is that PEGs that are too large to enter the pore will have minimal effect on channel conductance, but PEGs that are small enough to enter the pore will impede the flux of ions and thus reduce conductance [17,18,19]. Application of this technique showed that PEGs 400 Da and larger could not enter the pores, but PEGs 300 Da and smaller were able to enter the pores [20]. A similar study also showed that Cx32 channels have a size cutoff between PEG400 and PEG300 [21]. It also showed that Cx26 channels have a size exclusion limit around PEG200, while Cx37 channels have a size cutoff between PEG200 and triethyleneglycol (150 Da). Analogous work with unlabeled sugars indicated that mannitol (a linear monosaccharide alcohol) and stachyose (a branched tetrasaccharide) cannot permeate Cx40 channels [6], and that mannitol, but not raffinose (a branched trisaccharide) or stachyose, can permeate Cx43 channels [5]. These data suggest that Cx40 channels have a narrower limiting diameter than Cx43 channels. Other studies indicate that sorbitol (also linear monosaccharide alcohol), glucose (a monosaccharide), and sucrose (a disaccharide), but not stachyose, can enter Cx46 hemichannels [22,23]. Taken together, these studies suggest the following ranking of limiting pore diameter, in order of decreasing limiting pore diameter. The corresponding approximate unitary conductances of junctional channels are given in parentheses (heterotypic Cx26-Cx32 channels have a strongly rectifying unitary conductance; the values given are the conductances near V = 0 mV in two reports [24,25]): Cx43 (90 pS) > Cx32 (50 pS) > Cx26 (130 pS) = Cx26/Cx32 (48 to 89 pS) > Cx37 (300 pS) and Cx43 (90 pS) Cx46 (140 pS) > Cx40 (180 pS).

174

A.L. Harris and D. Locke

It is noteworthy that Cx37 channels seem to have the narrowest pore in spite of having the largest unitary conductance by far, and Cx32 channels, with the smallest unitary conductance in this series, have one of the widest pores. As noted previously, the disconnect between conductance and pore width can be explained by geometric considerations. For example, the high conductance-narrow limiting width case can be accounted for by a pore that is wide except for a very short constriction that would restrict molecular flux yet have minimal effect on electrical conductance. The low conductance-wide limiting width case can be accounted for by a pore that is just wide enough over its entire length to allow a given permeant to pass, or more complex properties, such as permeant-induced flexibility of the pore. To the extent that the limited data allow comparison between pore width and charge selectivity, there is a trend for the narrower channels to be more chargeselective, for example, the two channels characterized as the narrowest, Cx37 and Cx40, are significantly more charge-selective than the two channel characterized as being the widest, Cx43 and Cx32.

7.5 Molecular Permeability: Nonbiological Molecules The molecular permeabilities of connexin channels have been assessed using several classes of nonbiological tracer molecules, usually fluorescent, of widely divergent sizes, charges, and chemistries. Each type of connexin channel has characteristic permeabilities to these tracers (e.g., [26,27]). The literature documenting permeabilities to the various tracers of the various types of connexin channels is vast, and is not comprehensively reviewed here (a selective review can be found in [13]). This section presents a selection of information that allows empirical comparison of tracer permeabilities of various types of connexin channels. In the literature there is a tendency to use molecular weight as an index of tracer size, but a more appropriate parameter is the largest dimension of the minimal cross section of the molecule, that is, the largest dimension presented to the pore when oriented optimally for permeation. Where the tracers are largely planar and composed of conjugated aromatic rings, molecular weight does roughly correlate with this measure. Molecular modeling can give approximate van der Waals dimensions, but this neglects the contributions of hydration (which can be substantial), opportunities for hydrogen bonding and other electrostatic factors likely to affect the effective physical size as well as the charge envelope of the molecule (see [28]). Tracer molecules have been used successfully and extensively to report the existence and extent of junctional communication. Because of their diversity, they are less informative as tools with which to investigate the nature of the permeability pathway itself. At the most basic level, they reveal some information about the relative abilities of diverse, large permeants to pass through the pores. Almost all the tracers are charged, and, given their size, it is often

7 Permeability

175

presumed that their permeability would be affected by their charge to a greater degree than that of atomic ions, due to closer approach to charge on the walls of the pores. For this reason, one would expect that charge selectivity among these tracers is more profound than for atomic ions. Permeability to tracers has been assessed in various ways (see [13]). The least quantitative assesses the number of cells to which dye spreads from a donor cell loaded with tracer. This type of measurement can provide an all-or-none answer about permeability. Qualitative clues about relative permeabilities for different tracers can be inferred from this kind of study if differences in diffusional mobility are taken into account. Junctional permeability has been quantitatively assessed using fluorescence recovery after photobleaching (FRAP) [29,30,31,32,33]. The basic paradigm is to load cultured cells with dye, typically by use of acetoxymethyl (AM) ester derivatives of fluorescent tracers [34] or 5,6-carboxyfluorescein diacetate [35,36], which are then removed from the extracellular medium. The fluorescence signal in a cell is rapidly photobleached with a laser. The rate at which fluorescence returns to the bleached cell (i.e., the rate that dye diffuses into that cell from its neighbors though junctional channels) is measured. An effective transfer constant can be determined from this rate. This technique is called Gap-FRAP [33,37,38,39,40]. It has been adapted to three-dimensional (3D) structures (e.g., lens) by application of confocal microscopy [40,41]. Potential sources of error include assumptions of invariant number junctional channels and cell volumes across the population, and the possibility of photodynamic damage to the cells. To obtain information about relative permeability to a tracer through different types of connexin channels, the extent of dye spread must be normalized to the number of channels (usually taken as the junctional conductance divided by the unitary conductance) or to the permeability of a second tracer that is coapplied. The former index of channel number can be compromised by the existence of subconductance states, channel open probabilities that are less than unity, and the fact that for large plaques of active channels, the macroscopic conductance does not linearly reflect channel number, due to the overlapping access resistance domains of closely packed channels [42,43,44]. Ability to quantitate measurement of junctional flux is improved when the system is restricted to two cells and the rate of dye flux into the receiver cell is measured, and variations in intracellular distribution, cytoplasmic/nuclear binding, cell volume, and concentration in the donor cell are incorporated into the analysis, as well as number of junctional channels. In some cases, such studies allow determination of the absolute permeabilities to the tracers. Several studies of fluorescent tracer permeability that approach this level of rigor have been carried out [27,41,45,46,47,48,49,50,51,52]. Most of these studies include detailed, informed discussion of the parameters listed above and various strategies used to control for them. To calculate the per-channel permeabilities, several assumptions are typically made, including a presumption of a linear relation between the macroscopically observed rate of tracer flux and the junctional conductance to atomic

176

A.L. Harris and D. Locke

ions. This is known as the Pj/gj ratio, and has indeed been shown to be linear in most cases [1,53]. However, some of the calculated permeabilities have variability that could be accounted for by nonlinearity of this relation, suggesting that within a population of junctional channels the permeability to tracer molecules can vary independently of unitary conductance [45,46,52]. Cytoplasmic bridges can occur between cultured mammalian cells [54,55] and can mediate intercellular tracer transfer (see Chapter 6). Their contribution to intercellular transfer can be controlled for by conditions that close gap junction channels, such as low pH or use of any of the gap junction blockers (see Chapter 8). Demonstration of lack of transfer of gap junction–impermeable compounds, such as dextrans, serves the same purpose. Several attempts have been made to quantitatively model the calculated permeabilities of connexin channels to various tracers, incorporating factors such as frictional forces, various incarnations of ‘‘affinity’’, and their relations to the average pore width [1,47,48,49,52]. A limitation of these calculations is the necessary but clearly inappropriate simplifying assumption that the pore is a featureless cylinder of constant width. A common inference from these calculations is that there is substantial interaction between the permeating tracers and the interior of the pore that enhance permeability over what would be expected from purely physical considerations. Many tracers have been used (Table 7.3a). Two commonly used size-pairings are Lucifer yellow (–2; LY) and either neurobiotin (þ1; NB) or 4,6-diamidino2-phenyl-indole dihydrochloride (DAPI). Lucifer yellow is a highly planar, strongly anionic molecule whose smallest cross section is near the upper size limit of many connexin channels. Neurobiotin is among the smallest dyes used, and is cationic. It has the disadvantage of not being fluorescent, being detected after fixation with a fluorescent streptavidin derivative. DAPI (þ2) is fluorescent and a little smaller. There are a number of cases in which permeability to LY is much greater than that to NB or DAPI, and this is often taken to indicate discrimination on the basis of size, in spite of the charge difference. The Alexa series of tracers have also been used to discriminate pore width [27]. These tracers were designed to fluoresce at different wavelengths. To achieve this, the fluorochrome moiety was conjugated to different aromatic carrier structures, which fortuitously resulted in a somewhat homologous series of molecules of different dimensions. However, their molecular structures do not vary systematically and are chemically distinct. In addition, they have different numbers of discrete charges at different positions, which arise from different ionized moieties. Therefore, there are opportunities for substantially different chemical and electrostatic interactions with the pore. However, the permeability data so far are largely consistent with their permeation being determined by their sizes. There is a series of cationic tracers that is much more homologous, but has not been widely used. These are NB derivatives (þ1) with spacer linkages of different lengths [56]. They, like NB, are not fluorescent and are imaged only after fixation.

7 Permeability

177

Table 7.3 Molecular tracers used to study connexin channel permeability: tracers organized by mass (a) and by charge and mass (b) (a) Molecular Tracer: ionic form weight (Da) Size (A˚) Charge Hydroxycoumarin carboxylic acid (HCCA) 4,6-diamidino-2-phenylindole dihydrochloride (DAPI) N,N,N-trimethyl-2-[methyl(7-nitro-2,1,3benzoxadiazol-4yl)amino]ethanaminium (NBDT-M-TMA; NMT) Neurobiotin (NB) Ethidium bromide (EB) Po-pro-1(PP1) Alexa350 (A350) 6-carboxyfluorescein (CF) Dichlorofluorescein (DCF) Propidium iodide (PI) Lucifer yellow (LY)

206

286 314 325 326 376 401 414 443

Alexa488 (A488) Calcein (CA) Alexa594 (A594)

524 623 735

279

–6 15.4 6.0 [26]

þ2

þ1

280

12.7 5.4 [26] 11.6 9.3 [26] 13.0 5.2 3.2 [47] 12.6 12.7 8.5 [242] 12.3 12.7 5.5 [242] 12.9 9.3 [26] 10.6 9.5 [26] 12.6 14 5.5 [242] 11.3.1 10.5 9.0 [47] 16.6 13.8 9.3 [47]

þ1 þ1 þ2 –1 –2 –1 þ2 –2 –2 (–3, þ1) –4 (–6, þ2) –1 (–2, þ1)

(b) Anionic by increasing mass (charge)

All –1 –2 –4 –6

Cationic by increasing mass (charge)

all þ1 þ2

HCCA(–6), A350(–1), CF(–2), DCF(–1), LY(–2), A488(–2), CA(–4), A594(–1) A350, DCF, A594 CF, A488, LY CA HCCA DAPI(þ2), NBD(þ1), NB(þ1), EB(þ1), PP1(þ2), PI(þ2) NMT, NB, EB DAPI, PP1, PI

The most commonly used fluorescent cationic tracers are DAPI (þ2), ethidium bromide (þ1; EB) and propidium iodide (+2; PI). These molecules bind to nucleic acids, so obtaining quantitative information about their permeability is complicated by substantial cytoplasmic/nuclear binding. Two other cationic dyes

178

A.L. Harris and D. Locke

have been recently applied to studies of connexin channel permeability: the nucleic acid stain Po-Pro-1 (þ2), which has molecular weight near that of EB but two positive charges [57], and NBD-M-TMA (þ1), which is a bit smaller and has one positive charge [45]. The latter is a substrate for organic cationic transporters, so its mechanism of entry into cells must be controlled for. Assessment of charge selectivity using tracers is problematic, since there is no series of chemically homologous tracers of the same size and different charges. For this determination, investigators typically apply a set of dyes of positive and negative charge and see if the results can be explained by differences in charge. In one study, two structurally similar molecules with similar dimensions but different charge were used (6-carboxyfluorescein [CF] and dichlorofluorescein [DCF]) to help distinguish between the physical and electrostatic determinants of permeability of several connexins [28]. Table 7.3b organizes the tracers by charge and size. There are too many individual studies bearing on connexin channel permeability to tracers to be reported and discussed individually here (discussion of some of the relative permeabilities of tracers through specific connexin channels may be found elsewhere [46,48,51,52,58,59]). Table 7.4 presents a sampling of the data on the relative (1) permeability to different tracers of a single type of connexin channel, and (2) permeability to a single tracer through different types of connexin channels, where there was a reasonable attempt to normalize the data to allow comparisons across connexins. The information in the former case is the more reliable, since presumably the degree of junctional connexin expression and function is unchanged for the various tracers. In the latter case, the normalizations to number of functional junctional channels can introduce error. One striking feature of the aggregate data is that there is a reasonably high degree of consistency across the many experimental paradigms employed. Table 7.4 does not include data on the permeability of conductance substates [60,61,62] or that of various types of heteromeric and heterotypic channels [47,49,63,64,65,66]. Table 7.4a shows the relative selectivities of each type of channel to various tracers. Cation/anion selectivity is inferred below only when both types of tracers are directly compared. The table indicates the following findings: 1. Cx30 is highly selective for cationic tracers, as large and small ones are permeable, while anionic tracers are not. 2. Cx26 appears to be permeable to both cationic and anionic tracers, with greater permeability to the larger cationic over larger anionic tracers. 3. Cx36, Cx45, Cx46, and Cx57 may have a cation selectivity, but the data do not distinguish this from simply having a narrow pore. For example, Cx36 is highly permeable to the small cation EB and has much less permeability to larger anionic and cationic tracers. Cx45, Cx46, and Cx57 are similarly ambiguous, due to absence of comparison of cationic and anionic tracers of similar size.

7 Permeability

179

4. Cx32 seems to be permeable to large anions and small cations, but not to large cations, suggesting a wide, anionic selective pore. 5. There is no information on charge selectivity of Cx37, Cx40, or Cx45.6, but there is a hint about pore width due to lack of permeability to the largest Alexa tracer, but permeability to the smaller ones.

Table 7.4 Relative tracer permeabilities organized by connexin. Tracer permeabilities are organized by connexin isoform (a) and tracer (b). Some of the relevant references supporting each statement are provided, though many cases other studies contributed as well. Reports without comparative information for either tracers or connexins are not included. The symbol f indicates no detectable permeability. The prefix Cx has been omitted and the connexins identified only by the numerical portion of their names. Cx30.2 is the rodent ortholog of human CX31.9 and Cx45.6 is the chick ortholog of rodent and human CX50. Cx57 is the rodent ortholog of human CX62. Cationic dyes are in italics and anionic dyes in Roman point. (a) Connexin Relative tracer permeation 26

30

30.2 31 32

36 37 40 43

45

45.6 46 57

DAPI > LY NB > A488 > LY = CA >> fA594 NB > EB = PI > LY, CF A350 = A488 > A594 NB > LY NB > EB = PI >> fCF, LY PI >> fA488 DAPI, EB >> fPI A3501 >> fLY DAPI = LY >> fEB, PI A350 = A488 > A594 LY > DAPI LY >> fDAPI DAPI = LY >> fEB, PI EB >> CF = LY > PI A350 > A488 >> fA594 A350 > LY A350 > A488 >> A594 NMT = A350 > LY > A488 >> A594 A350 > LY LY = CA A488 > A350 = A594 A488 > A350 > A594 CF = LY >> EB > PI A350 > LY NB > LY A350 > A488 > A594 DAPI > LY A350 > A488 >> fA594 DAPI >> fLY Cal > LY NB>>LY

[27] [243] [64] [47] [243] [64] [65] [63] [63] [26] [47] [27] [232] [26] [162] [47] [63] [47] [52] [63] [46] [244] [47] [162] [63] [66] [47] [27] [244] [232] [46] [245, 246]

180

A.L. Harris and D. Locke Table 7.4 (continued) (b) Tracer Connexin permeability DAPI NMT NB EB

PI A350

CF LY

A488

CA A594

45 > 26 > 32 45 > 26 > 32 43 > 37 26 > 30 26 = 30 26, 37, 40 >> f31, 32 26 > 30 36 > 43 26, 37, 40 >> f31, 32 26 > 30 45 > 40 > 32 > 26 = 43 > 37 43 > 40 > 45 >> 30.2 45.6 > 43 26 >> f30 43 > 36 43 > 40 26 >> f30 43 > 40 > 45 >> f30.2 43 > 45.6 32 > 26 > 45 26 >> f30 43 > 46 43 > 45 26 > 30 32 > 26 > 45 43 > 36 26 >> f30 26, 26/30 >> f30 43 > 40 > 45 = 32 > 26 >> 37 43 > 45.6 43 = 32 43 >46 32 = 43 > 45 > 26 >> f37 43 >> f45.6

[53] [27] [45] [64] [243] [26] [64] [162] [26] [64] [47] [63] [244] [64] [162] [49] [247] [63] [244] [53] [64] [46] [66] [243] [27] [162] [65] [65] [47] [244] [110] [46] [47] [244]

6. Cx43 appears to be a relatively wide pore, permeable to large cations but only to small cations, similar to Cx32. 7. Cx30.2 seems to be quite narrow, being almost impermeable to A350, and impermeable to larger tracers. The permeability to PI may indicate a cation preference. It is intriguing to note that for Cx43 the permeability to A488 is greater than that to A350 (see below). Table 7.4b shows the relative permeabilities of different connexin channels to each tracer, allowing comparison of connexin channel properties. As mentioned above, these relations are more prone to error, as there is need for normalization

7 Permeability

181

to the number of functional junctional channels across connexins. In each study listed, there was some attempt to do this, but the levels of rigor and resolution vary greatly. The data permit the following rank orderings, accompanied by the approximate unitary conductances: 1. Permeability to large cationic tracers: Cx45 (30 pS) = Cx26 (130 pS) > Cx30 (150 pS) and Cx36 (15 pS) > Cx43 (90 pS) > Cx37 (300 pS). 2. Impermeability to large cationic tracers: Cx31 (85 pS) and Cx32 (50 pS) 3. Permeability to large anionic tracers: Cx32 (50 pS) = Cx43 (90 pS) > Cx40 (180 pS) > Cx45 (30 pS) > Cx26 (130 pS) > Cx30 (150 pS) >> Cx37 (300 pS), Cx30.2 (10 pS) and Cx43 (90 pS) > Cx46 (140 pS), Cx45.6 (200 pS) and Cx36 (15 pS). Several data points conflict with the above rank orderings. The relation between the permeabilities of Cx32 and Cx43 is ambiguous, as is the relation between the permeabilities of Cx45 and Cx26. There is a reversal of permeability of Cx45.6, relative to that of Cx43; Cx45.6 is essentially impermeable to the large anion Alexa594 (A594), but more permeable to smaller anion A350 than is the otherwise much more permeable Cx43. This, along with discrepancy noted above for Table 7.4a, suggests that movement of Alexa350 (A350) through Cx43 channels is specifically and anomalously reduced, presumably by unique interactions between A350 and the Cx43 pore that do not occur in other connexin channels. As mentioned, the above inferences about charge selectivity are uncertain due to the inability to rigorously differentiate the roles of permeant size and charge. Nevertheless, a few observations may be made: 1. From conductance and reversal potential measurements, Cx32 is only slightly anion selective and Cx43 essentially nonselective. Yet in both cases, permeability to the larger cations is greatly reduced relative to that of the large anions, to which they are equally permeable. This striking deviation could result from closer electrostatic interaction of the permeants with charges within the pores. 2. The most unambiguously cation-selective channel, Cx30, is among the narrowest, as inferred from its impermeability by large cationic tracers. 3. The magnitudes of channels’ unitary conductances do not correlate with the sizes of the permeant tracer molecules. In fact, there is almost an inverse correlation between permeation of large tracers and unitary conductance. Some of the larger tracer molecules permeate channels with low conductances (e.g., Cx45 and Cx32), and are unable to permeate channels with much higher conductances (e.g., Cx37, Cx26, Cx40, Cx30, Cx45.6). This dramatically confirms that a simplistic view of the lumen of the connexin channel as a ‘‘right cylinder’’ is incorrect, and suggests that the pore has the complex structure of most other eukaryotic membrane channels. The inverse permeability of A350 and A488 through Cx43 channels suggests that specific permeant-pore interactions, beyond size and charge, play roles in permselectivity of connexin channels, a theme that is further amplified below.

182

A.L. Harris and D. Locke

7.6 Molecular Permeability: Biological Molecules The first observation of second messenger junctional permeability, made 30 years ago [67], sparked broad interest in gap junction channels as mediators of intercellular molecular signaling and as key players in development, physiology, and disease. Only recently have investigators been able to directly address the key question: What specific biological molecules, second messengers and others, are able to permeate the various types of connexin channels, and how well? As can be inferred from the previous section, the pores of most, if not all, connexin channels are sufficiently wide to allow permeation by many cytoplasmic molecules. There is evidence for permeation through at least some types of connexin channels of virtually all soluble second messengers, amino acids, nucleotides, Ca2þ, and glucose and its metabolites [13,68,69]. Until a few years ago, the expectation was that any cytoplasmic molecule of appropriate size would permeate connexin channels, and that the limiting pore diameter would be the primary determinant of molecular permeation. It had also been widely presumed that, due to the relatively large pore diameter, there would not be significant charge selectivity among permeants. As can be seen from the data presented above, both expectations are demonstrably incorrect with regard to molecular tracer permeability, and therefore likely for biological molecule permeability as well. In addition, the lack of correlation between the simple indices of channel permeability (i.e., unitary conductance, charge selectivity, permeability to tracers, e.g., [28,70]) suggests that simple extrapolations from any of these indices to permeability of cytoplasmic molecules is unwarranted [47,51,71,72,73]. To date, only a few studies reveal differences in permeation by different biological molecules through a particular type of connexin channel, or differences in permeation by a particular biological molecule through different types of connexin channels. Even with this small data set, the results make clear that different connexin channels can have highly distinct and differential permeabilities among cytoplasmic molecules, and that these bear little discernible relation to the atomic ion charge selectivities or to the permeabilities to nonbiological fluorescent tracers. This suggests that there are highly specific permeant-channel interactions for biological permeants. Elucidation of the specific roles that gap junctions play in development, physiology, and pathology will require an understanding of this issue, empirically, if not mechanistically (comprehensive review of this topic can be found in [74], and other relevant reviews include [51] and [71]). The same considerations that apply to assessment of relative tracer permeabilities apply to assessment of relative cytoplasmic molecule permeabilities, but there are additional concerns. Introduction of arbitrary biologically active compounds into cell cytoplasms may have direct or downstream effects on junctional channels. Biological compounds can be degraded, a factor that does not typically affect assessment of permeability to nonbiological molecules.

7 Permeability

183

Consider the case where a biological compound is degraded at a rate within an order of magnitude of the rate of junctional flux, or faster. The detection of its rate of accumulation in the receiving cell will not accurately reflect the rate of junctional flux, giving an artifactually low apparent junctional permeability. Rates of degradation can differ between cells, and be dependent upon metabolic state. Most critically, the ability to quantitatively detect flux into a coupled cell is constrained by available detection methods, as chemical modification of the molecules (e.g., addition of fluorescent moieties) could affect permeability. Given these issues, it is no surprise that few studies yield rigorous information about relative permeabilities to different biological molecules for a given connexin channel, or about the relative permeability to a given cytoplasmic molecule for different connexin channels. The data on permeation of cytoplasmic molecules through connexin channels come from a variety of experimental approaches. Some involve direct measurements of molecular flux through unambiguously identified connexin channels. Other studies rely on indirect methods to infer the identity of the transferred compounds or that the transfer is through connexin channels (detailed discussion of the methods is in [74]). For junctional channels, determining the permeant species is problematic when intercellular Ca2þ signaling is taken as evidence of gap junction–mediated molecular communication. Such signaling is now well established to occur by two distinct pathways — gap junctional and paracrine — and both may operate in a given situation [75,76,77,78,79,80,81,82,83,84,85]. The former involves intercellular diffusion through gap junction channels of a molecular signal, usually considered to be inositol triphosphate (IP3) or Ca2þ. This signal elicits intracellular Ca2þ release, which regenerates the signal. The paracrine pathway involves extracellular release of adenosine triphosphate (ATP), via any of several connexin and non-connexin candidate mechanisms, which interacts with purinergic receptors on a neighboring cell, in turn inducing Ca2þ entry. Therefore, intercellular Ca2þ signaling may be due to gap junctional permeation by IP3 or Ca2þ, or to ATP release, with connexin hemichannels being one of the possible pathways for the last mechanism. This means that observations of intercellular Ca2þ signaling cannot be designated as junctional unless the paracellular pathway is excluded. This may be done by the use of suramin, oxidized ATP or other agents that either block postsynaptic purinergic receptors or that degrade ATP. Where established to be junctional, as opposed to paracrine, the Ca2þ signaling is usually considered as evidence that IP3 permeates the junctional channels. However, junctional Ca2þ flux may also be involved. In some cases, it is clear that regenerative Ca2þ signaling is mediated by IP3 flux and that the contribution of junctional Ca2þ flux is minor [86,87]. In other cases it appears that junctional Ca2þ flux is a major factor [78,88,89] and in still others that both are involved [90]. Computational modeling has supported each scenario [91,92,93,94,95]. Because of the difficulty of distinguishing the two junctional

184

A.L. Harris and D. Locke

permeants, this chapter does not distinguish IP3 from Ca2þ permeability when it is indicated by intercellular Ca2þ signaling. There is the possibility that intercellular molecular transfer can be mediated by junctional channels formed by pannexin [96,97,98]. Pannexin1 (Panx1) has been reported to form junctional channels when overexpressed in the Xenopus oocyte system [97] and in at least one cell line [99] but not in others [100,101]. Pannexin gene and protein expression have been demonstrated in a wide variety of tissues [96,97,102,103,104,105], but expression of Panx1 in most cell types used in permeability studies to date has not been directly assessed. This raises the possibility that junctional permeability mediated by pannexin channels could contribute to permeability that has been attributed to connexin channels, particularly if expression of the two proteins is positively correlated. However, the weight of current evidence is that Panx1 does not normally form junctional channels [100,101,106] (see Chapter 12). The most common method for detecting junctional transfer of specific cytoplasmic compounds is by monitoring the appearance of a specific compound in one cell in response to increase in the concentration of that compound in a coupled cell. This can be via a fluorescent sensor for the compound itself (e.g., FURA-2 for Ca2þ [88], H30 for cyclic adenosine monophosphate [cAMP] [107,108], LIBRA for IP3 [107]), or by the activity of channels that are regulated by a compound (e.g., cyclic nucleotide-gated (CNG) [72] or cystic fibrosis transmembrane conductance regulator (CFTR) chloride [60] channels for cAMP). One must consider that the elevation of a compound may be a secondary consequence of intercellular movement of a distinct, un-‘‘sensed’’ compound. Another approach involves metabolite capture, in which radiolabeled compounds are harvested specifically from recipient cells (cells not radiolabeled, but cocultured with metabolically radiolabeled donor cells) and then identified biochemically [109,110,111,112,113]. The caveat here is that the compounds thus identified may not be those that passed through the junctions, but rather their metabolic products. Studies involving hemichannels have much greater potential for artifactual findings, except in the cases where purified connexin protein is used. Most such studies involve assessment of the release of cytoplasmic compounds from cells into the extracellular medium. Proof that connexin hemichannels are the pathway for release is highly problematic because of other plasma membrane channels that can be permeable to large molecules, whose expression or regulation may be affected by the treatments that affect connexin channels. The difficulties and criteria for identifying hemichannel function in cells are comprehensively reviewed in Spray et al. [114]. The most prominent non-connexin pathway is that formed by Panx1 channels. These channels can exist on their own or activated by purinergic P2X7 receptors [115,116]. The Panx1 pore is permeable to Lucifer yellow, ATP, and glutamate [117,118,119,120,121,122,123,124,125,126,127]. When in association with P2X7 receptors its activity is enhanced by low extracellular Ca2þ [120], which is a common method of activating connexin hemichannels

7 Permeability

185

[128,129,130,131]. A recent report showed that in glial cells the ATP release that occurs during paracrine intercellular propagation of Ca2þ signaling is through P2X7-activated channels, and not through Cx43 hemichannels [124]. P2X7 receptor-associated and P2X7 receptor-unassociated Panx1 channels are blocked by several of the pharmacological agents commonly used to inhibit connexin channels. These include glycerrhetinic acid, carbenoxolone, mefloquine, octanol, and heptanol [102,115,124,132]. In addition, connexin mimetic peptides (peptides whose sequences correspond to segments of the extracellular loops of connexins) have been used to block connexin channels, and presumed to be specific for them, but now have been shown to substantially reduce currents though Panx1 channels at the same concentrations that block connexin channels [133]. These data make a strong case for not relying on these agents as the sole means to identify release of cytoplasmic compounds via connexin hemichannels. As pannexin channels are widespread, release of ATP and other cytoplasmic molecules through them must be considered a significant possibility that is not controlled for most of the studies on this issue; these recent findings call into question many of the previous conclusions regarding connexin channel properties that are based on molecular leakage studies. In some studies, correlation of release of cytoplasmic compounds with connexin expression is taken to indicate that the release is through connexin hemichannels. In several systems, there is clear and close correlation between connexin expression and ATP release [134,135,136]. However, a large literature documents the multifaceted biological effects of induced connexin expression. Altered connexin expression has been shown to have dramatic effects on the expression of hundreds of other gene products [137,138,139,140], which may have downstream effects on non-connexin plasma membrane proteins and other mechanisms of release. Such changes could alter expression or regulation of the other candidate channels. In several cases there is a notable lack of correlation between the conditions under which ATP release occurs and the conditions required for significant opening of the relevant connexin hemichannels. This important issue has been discussed in detail in several excellent review articles [98,106,141,142,143,144,145]. Other channels that can be permeable to large molecules include maxi-anion channels [146,147], the CFTR channel [148], plasma membrane voltage-dependent anion channels (VDACs) [149,150,151] and non-connexin mechanosensitive or volume-sensitive channels [152,153,154,155]. Several methods can be used to eliminate the potential contributions of these channels to observed release of cytoplasmic molecules; however, these controls are not applied in many of the published studies. It stands to reason that if connexin hemichannels mediate the flux into the medium of ATP (or glutamate, as has been reported), other cytoplasmic molecules ought to be released into the extracellular medium as well, but this has not been systematically investigated. Each study must be evaluated on its own merits. As indicated above, the vast majority of work documents all-or-none permeability (these data are summarized in [74]). From the all-or-none studies, with a

186

A.L. Harris and D. Locke

few exceptions, the most commonly studied second messengers — ATP, IP3/ Ca2þ, cAMP, cyclic guanosine monophosphate (cGMP) — have been found to be permeable to some degree through every type of connexin channel studied. The more interesting data are those that allow comparison of the ability of a given compound to permeate channels formed by different connexins, or how well different cytoplasmic molecules permeate a given type of connexin channel. These differences are undoubtedly the major determinants of the unique signaling functions that each type of connexin channel seems to serve. As for the tracer studies, obtaining information on the relative magnitudes of molecular permeability requires consideration of a multitude of factors, but rests on the ability to quantitatively determine rates of flux and normalization to number of channels. The ideal type of measurement is difficult to obtain. Monitoring or determining the concentration of compound in the donor cell and the rate of appearance of the compound in the receiver cell can be complicated if the concentration of free compound in the donor cell is not constant over the time course of the measurement, due to leakage, metabolism, or junctional transfer itself. In addition, calculation of junctional permeability can be compromised by cytoplasmic binding or spatial buffering. Measurement of junctional flux into the receiver cell also requires that attention be paid to the potential for leakage and metabolism over the time course of the measurement. Unless the total amount of compound entering the receiver cell is measured, variation in cell volume and inhomogeneities of compound distribution can introduce errors. Two recent studies show how most of these concerns can be successfully addressed: 1. In the work of Bedner et al. [72,156], cAMP was released photolytically in the donor cell, and activity of CNG channels in the receiver cell reported the transjunctional accumulation of cAMP. Activity of CNG channels was monitored by the Ca2þ signal of Fluo-4. The raw data showed no correlation between Ca2þ signal in the receiver cell with the number of gap junction channels as assessed by junctional conductance. This demonstrated the need for strict normalization for cell volume, expression of the sensor (in this case, the CNG channels), and nonlinearity of response of the sensor (the response of the CNG channels to increased cAMP, and the response of Fluo-4 to the increased Ca2þ signal). The correction for these factors was to use only data from experiments in which the cAMP to Fluo-4 signal in the receiver cell due to junctional transfer was equal to that obtained in the same cell by direct photolysis with variable flash duration. Once equivalence was obtained, the cAMP flux could be compared for cells expressing different connexins, when normalized to the number of channels. The use of such an internal calibration for each cell pair avoids issues of cell-specific variation in cAMP hydrolysis or CNG channel kinetics. The remaining potential sources of error relate to the possibility of different substate occupancies of the different

7 Permeability

187

connexin channels, and the access resistance issue mentioned previously, which varies with the size of the junctional plaque, both of which could affect calculation of the number of junctional channels. 2. Hernandez et al. [107] used novel ratiometric fluorescent sensors to directly assess junctional flux of cAMP and of IP3. Use of such sensors simplifies the measurement in eliminating the intermediate transduction step in the previous case from change in second messenger level to change in Ca2þ influx. The sensors used were fluorescence resonance energy transfer (FRET)-conjugates of proteins that are sensitive to the respective second messengers. For cAMP, the sensor was based on Epac, a guanine nucleotide exchange factor [157] and for IP3 was based on the ligand-binding domain of an IP3 receptor [158]. Each sensor was genetically expressed in the cells, and the responses calibrated in those cells via patch electrode dialysis with solutions of known concentrations. The donor cell was patch clamped with a pipette containing a known amount of compound (cAMP or IP3). The total ratiometric FRET fluorescence was determined for each cell, and the volume of each cell determined from through-focus z-axis sequences of confocal images. The junctional transfer rate was calculated from the FRET signals in the donor and receiver cells only during the interval when the FRET signal was known to linearly reflect the concentration of the compound being sensed. Immediately afterward, junctional conductance was determined by patch clamp of the receiver cell, to permit estimation of the number of functional junctional channels. The possibility of the exogenous compounds affecting junctional conductance during the experiment was controlled for by maintaining the dual-cell patch for long enough to see if there was a tendency for the junctional conductance to change. The two considerations mentioned above regarding possible error in estimation of number of junctional channels due to gating or subconductance states, and due to access resistance, apply here as well. Posttranslational modifications may affect permeability, either by direct modification of the permeation pathway or by altered gating leading to occupation of conductance substates. To date, this has been explicitly demonstrated only for Cx43 [60,61,159,160]. However, the role of posttranslational modifications on permeability is relatively unexplored; this leaves open the possibility that the rank orders of metabolite permeation may vary depending on the cell system used, due to different posttranslational modifications of the connexin. There is some evidence for expression-system specific differences in posttranslational modification of connexins (e.g., [161]) and in selectivity (e.g., [162]). The existing data that relates to connexin-specific differences in permeability to cytoplasmic molecules is given in Table 7.5. Before discussing the data in the table, it is important to note two remarkable and unexpected findings involving connexin channel permeation by biological molecules. One is that single-stranded interfering RNAs, which one would expect to be too large to permeate junctional channels, can in fact do so [163]. This

188

A.L. Harris and D. Locke

Table 7.5 Relative biological molecule permeabilities. Molecular permeabilities are organized by connexin isoform (a) and by permeant (b). As in Table 7.4, the Cx prefix is omitted. Data from junctional channels are indicated by ‘‘j’’ and from hemichannels by ‘‘h’’. (a) Connexin Relative permeability 26/32 h

26j 26/32j 32j

43j

(b) Permeant Adenosine AMP ADP/ATP cAMP

cAMP or cAMP depending on connexin isoform stoichiometry cGMP or cGMP depending on connexin isoform stoichiometry IP3 or fIP3 depending on stoichiometry and on IP3 isoform IP3 = cAMP siRNA 22mer >> fdsRNA 12mer Adenosine > glucose > ADP/ATP > glutathione> AMP > glutamate glutathione = glutamate > ADP/ATP ADP/ATP >> fadenosine ADP /ATP = glutathione = glutamate ADP/ATP = glucose > glutathione > AMP > glutamate >>f adenosine ssRNA 12mer > 16mer > 24mer >> fdsRNA 12mer siRNA 22mer >> fdsRNA 12mer peptides (aa) 4 > 5 > 6 > 7 > 8 > 9 > 10

[14,166] [14,166] [166] [73] [163] [112] [110] [110] [110] [112] [163] [163] [164]

Relative permeability

32j >> f43j 43j >> 32j 43j >> 32j 26/32h depending on stoichiometry 43j > 26j > 45j = 32j > 47j >> 36j cGMP 26/32h depending on stoichiometry Glucose 43j > 32j Glutamate 43j > 32j 43j >> 32j Glutathione 43j > 32j 43j >> 32j 26/32h depending on stoichiometry 26/32 h depending on IP3 stoichiometry and IP3 isoform 32j > 43j > 26j RNA 12mer 43j >> f26/32j siRNA 22 mer 43j >> f26/32j 1 Very low permeability.

[112] [112] [110] [14,15] [72,156] [14,15] [112] [110,112] [110,112] [15,166,167] [163] [163]

property is connexin-specific, with permeability through homotypic Cx43 channels and not through heterotypic Cx26-Cx32 channels (double-stranded RNA 12mers were not permeable through either type of channel). This provides proofof-principle for an exciting and potentially very important avenue of gap junction–mediated intercellular regulation. The other finding is that linear/unstructured peptides up to at least ten amino acids can permeate Cx43 junctional channels [164]. The importance of this is the potential role of cell–cell transfer of such peptides in antigen presentation and cross-presentation in the immune system. That this actually occurs has received recent experimental support [165].

7 Permeability

189

The mechanisms by which these molecules thread their way through junctional channels is unknown, but it seems likely that interaction with the pore facilitates the structural changes that must accompany permeation. As for the tracers, when one compares the pore diameters assessed by uncharged molecules with the biological permeants that have been shown to pass through them, one is struck by the absence of correlation. There is also deviation from expectations based on charge selectivity and the tracer permeability data. On the basis of the relative permeability data for Cx32 and Cx43 channels, one would not conclude that Cx32 is the narrower pore, which is clearly the case. For Cx32, since permeation of adenosine is greater than that of the anions adenosine diphosphate (ADP)/ATP [112], one might infer that increased negative charge and size of the permeant decreases the permeation. However, increased negative charge should favor permeation through this anion-preferring pore, and adenosine monophosphate (AMP) is smaller and less negative, yet permeates less well. In the case of Cx43, ADP/ATP permeate better than adenosine [112], making the largest and most negatively charged molecules the most permeant through this channel. Furthermore, the tracer data suggest that Cx32 and Cx43 are equally permeable to large anions, yet they display a remarkable and unpredictable divergence with regard to permeation by large, anionic biological molecules, with Cx43 favoring AMP, ADP/ATP, and cAMP, and Cx32 favoring adenosine and IP3. The relative permeabilities of the weakly anion selective channel (Cx32) do not correlate with the magnitude of negative charge. The anion cAMP permeates best through the wide Cx43 channel [5], next best through the narrow, cation-preferring Cx26 channel, followed by both the most highly cation-selective channel (Cx45) and the wide anion-preferring channel (Cx32) [72,156], which are equally permeable to it. In addition, the fact that some of the heteromeric Cx26/Cx32 channels are impermeable to some inositol triphosphates and permeable to others [166] makes clear that pore diameter and permeant charge are not the key determinants of the selective permeability. Cx26 is fairly cation-selective to atomic ions, and much less permeable to large tracer anions than Cx32 and Cx43, intermediate in permeability to cAMP, and yet more permeable than either to IP3. Cx26 is equally permeable to cAMP and IP3 [107], even though the latter has six times the negative charge. This is a striking example of the inability of charge selectivity derived from fluxes of atomic ions to predict charge selectivity for biological molecules. The exception is that the relative efficacy permeation of the highly anionic IP3 through Cx32, Cx26, and Cx43 channels does correlate to some extent with the degree of cation selectivity, with Cx32 being the most permeable, Cx43 less permeable, and Cx26 being the least permeable (presumably due to its narrower pore; Cx43 and Cx26 being equally and only moderately cation selective) [167]. From a purely biological perspective, it is worth noting that the data show that Cx26 is more permeable to cAMP than is Cx32, and that Cx32 is more

190

A.L. Harris and D. Locke

permeable to IP3 than is Cx26 (speculations on the consequences of this inverse permeability preference are in [72]). The work with reconstituted heteromeric Cx26/Cx32 hemichannels shows that some of these channels were permeable to IP3 and some were not, whereas all the inositol triphosphates tested permeated the corresponding homomeric hemichannels [166]. This suggests that heteromeric channels can have higher degrees of selectivity than homomeric channels. Given the ubiquitous distribution of heteromeric channels in vivo, this may be biologically important, and offers a unique role for heteromeric connexin channels. The significant differences in permeability cannot be readily accounted for by differences in permeant size or charge; it is difficult to account for the observed selectivity without invoking some kind of specific affinity between the channels and specific biological permeants. Based on computations, such interactions have been proposed for nonbiological permeants as well [47,48].

7.6.1 Signaling Consequences of Different Molecular Permeabilities A discussion of the potential biological importance of the specific relative permeabilities discovered to date is in [72]. From the papers cited in Table 7.5, the largest permeability differences are as follows: 1. For Cx32 junctional channels, a 8.3-fold difference in permeability between adenosine and glutamate [112] 2. For Cx43 junctional channels, a 3.4-fold difference between ADP/ATP and glutamate [112] 3. For ADP/ATP, a 100-fold and a 7.1-fold difference between Cx43 and Cx32 junctional channels [110,112] 4. For cAMP, a 33-fold difference between Cx43 and Cx36 junctional channels [72,156] 5. For glucose, a 4.3-fold difference between Cx43 and Cx32 junctional channels [112] 6. For glutamate, a 33.3-fold and a 8.3-fold difference between Cx43 and Cx32 junctional channels [110,112] 7. For glutathione, a 33.3-fold and a 4.3-fold difference between Cx43 and Cx32 junctional channels [110,112] If one presumes that a later measurement of the same permeabilities by the same research group is the more accurate one, the largest fold-differences range from 33.0 to 3.4. What is the biological utility of an intercellular pathway with differences in permeability to second messengers within this range or even less? For example, what difference would it make whether cAMP permeates a junctional channel

7 Permeability

191

several times better than IP3? The naive view would be that these molecules will easily equilibrate within a population of well-coupled cells, so relative differences in junctional permeability are unimportant, and all that matters is whether the molecules can permeate at all. This view neglects several factors. The profile of steady-state levels of a compound within a population of coupled cells critically depends on the specific relations among the rates of junctional flux, synthesis, degradation, and diffusion in cytoplasm of the compound. If the relations between these rates are different for another compound, its profile will be different. Second messenger molecules typically have constrained lifetimes and diffusional persistence. The average lifetimes of cAMP and IP3 in cytoplasm are 60 seconds [168] and <10 seconds [169,170,171], respectively. This fivefold difference in lifetime is a major determinant of effective range and steady-state concentration of these molecules. For a signaling molecule with restricted lifetime and diffusion, a relative increase in junctional permeability could substantially affect the level reached at any particular location within a population of receiver cells, and how fast (see below). It has been estimated that the effective range of IP3, a function of its cytoplasmic lifetime and diffusion constant, is 24 mM [172] and for cAMP is 81 mM [168,173]. If their diffusion is constrained by junctional transfer, their effective ranges are less. Thus, the relative junctional permeabilities to cAMP and IP3 determine their relative ranges of action. Computational modeling shows that differences in levels of junctional molecular permeability over a ten fold range — within that seen in the published data — can produce profound differences in tissue response to periodic release of second messengers [174]. Essentially, for two compounds with different junctional permeabilities the system behaves as if it is better coupled with regard to one compound than the other. Whenever there is a difference or a change in any of the relevant rates (junctional flux, synthesis, degradation; e.g., in response to hormone receptor activation or difference/change in connexin channel composition) at a particular location, the profile of concentration of a compound will change with kinetics that are functions of all the rates, including junctional permeability. If two compounds have different junctional permeabilities, that is, if the rates of junctional flux differ, the rates of change will be accordingly different. The consequence is that not only are the steady-state profiles of the compounds different, but the kinetics of change in those levels are different for each compound. Thus, the relative junctional permeabilities are key factors in any signaling or regulatory system that has a kinetic component, where the rate of change or oscillation in the level of a compound is important, for which there are many examples. A large literature supports the idea that oscillatory changes in signaling molecules (e.g., Ca2þ, cAMP) convey information distinct from changes in steady-state levels [175,176,177,178], and that such oscillations occur across systems of coupled cells [179,180,181,182,183,184,185,186]. In many cases,

192

A.L. Harris and D. Locke

intercellular spread of the oscillatory signal has been shown to be dependent on junctional communication. The rates at which the signaling molecules permeate junctions have a defining effect whether the oscillatory effects are transmitted and the nature of the oscillations themselves. Oscillation frequency can be a function of agonist dose (e.g., vasopressin in liver; [187]) and therefore the ability to follow or appropriately propagate the oscillatory signal depends on the kinetics of junctional flux, as shown computationally for differences in junctional permeability less than sevenfold [95]. Other computational work shows that the range of propagated Ca2þ waves varies strongly with changes in junctional permeability of less than fivefold [94,188]. The roles of signal kinetics and of the degree of junctional molecular flux in defining intercellular oscillatory signaling has been described computationally in other work as well [82,189,190,191,192] (see Chapter 22). These studies show that even modest differences or changes in selectivity at cellular junctions, within the range reported in the published work, can have major impact on the strength, character, and location of the intercellular signaling, under both steady-state and kinetic conditions.

7.6.2 Specific Affinities in Other Large Pores How can pores as wide as connexin channels be so highly selective? In fact, other wide pores have metabolite-specific permeabilities. Like connexin channels, these other channels conduct ions and a variety of small molecules. However, in each case, permeability to certain biological molecules is specifically enhanced, or reduced for homologous nonbiological permeants. The maltoporins (LamB channels) favor the flux of maltose over that of other oligosaccharides [12,193,194]. In these channels, the spatial configuration of luminal hydrophobic residues that spiral down the length of the pore form a greasy slide whose structure is optimized for maltose. The voltage-dependent anion channel is the primary pathway for metabolite flux through the outer membrane of mitochondria. It is conductive to ATP in its open state, but not the closed state, which is really just a lower conductance state, even though the hydrodynamic size of ATP is only about one-half that of the closed state diameter [151,195,196]. Single-channel studies show that a variety of nucleotides enter the pore equally well, but they interact with an intrapore binding site with affinities that range over a factor of 40, in the following order: reduced nicotinamide adenine dinucleotide phosphate (NADPH) > ATP/guanosine triphosphate (GTP) > reduced nicotinamide adenine dinucleotide (NADH) > nicotinamide adenine dinucleotide (NAD) = ADP > AMP >> uridine triphosphate (UTP). The ability to distinguish ATP from UTP suggests specificity for purine bases [196,197]. Genetic removal of a set of positive charges within the pore resulted in a channel that was cation rather than anion selective, and that had only about

7 Permeability

193

half the unitary conductance. In spite of this change, the permeability to ATP was unaltered [198]; one would have expected the permeability to ATP to decrease with this change in charge selectivity. It was concluded that ATP flux through VDAC channels is determined by specific sites within the channel, not by the overall charge density of the pore, reflected in the cation/anion selectivity among atomic ions. Thus, the general electrostatic interactions inferred from conductance measurements and reversal potentials of atomic ions do not play a crucial role in the molecular flux through this channel. This has direct implications for similar inferences about connexin channel permeability, as suggested by the discussion of Cx43 and Cx32 permeability above. In a different way to probe the basis of the selectivity, the ability of synthetic molecules of similar size and charge as ATP to permeate VDAC channels was assessed [196]. The anions tetraglutamate and 1-hydropene-3,6,8-trisulfonate were excluded from the pore, again making the point that highly specific interactions — in this case those of natural metabolites — are key determinants of selectivity. A third example is provided by the interaction of penicillin antibiotics with the bacterial porin OmpF. This channel is the primary pathway by which b-lactam antibiotics, including penicillin, gain entry into bacteria [194,199]. A series of penicillins, selected on the basis of size and charge, were assessed for ability to permeate the OmpF pore [200,201]. The results from experiments and molecular dynamics simulations show that among this set of highly similar molecules the ability to permeate the pore is crucially dependent on specific charge distributions and molecular flexibilities that permit complementation of the charge distribution at the narrowest part of the pore. In other words, a successful permeant must interact in a highly specific manner with the selectivity filter. In close analogy to interaction of a Kþ ion with the selectivity filter of Kþ channels, it was found that the strength of the interactions of the permeants at this site correlated with enhanced flux; if the interactions are weak, the molecule does not permeate. From these studies, it is clear that wide channels can have selectivity mechanisms that are highly specific for certain permeants, and whose properties cannot be extrapolated in a straightforward way from estimates of pore size, conductance or charge selectivity, or even from known permeants.

7.6.3 Where Is the Molecular Selectivity Filter? None of the channels mentioned above, or connexins, has canonical binding sites for the indicated molecules. The efficacy of noncanonical sites is demonstrated by the intriguing finding that providing a ring of arginines in the lumen of the -hemolysin pore confers upon it the ability to bind IP3 with low nanomolar affinity, but not another negatively charged second messenger, cAMP [202].

194

A.L. Harris and D. Locke

The lack of canonical sites is understandable because the canonical sites for these molecules have affinities too great to allow rapid flux. The issue of the relationship between intrapore molecular affinities and efficiency of transfer has been investigated computationally [203,204,205,206]. To localize the key sites of interaction with cytoplasmic permeants within connexin channels, one might look to examples where point mutations have affected the permeability to second messengers without appearing to affect other permeability properties. In Cx26, the V84L mutation was found to affect the ability to propagate intercellular Ca2þ signaling (e.g., altered permeability to IP3) without affecting either the unitary conductance or the permeability to LY [207]. This striking result suggests that V84 is pivotal for IP3 permeation, specifically. However, this residue is in the middle of the M2 transmembrane helix, and a V84C mutation shows only weak and inconsistent reactivity to thiol-reactive reagents [208] (see Chapter 3). If V84 is not pore lining, perhaps the modest difference in volume involved in the V84L substitution results in a change that propagates sterically or allosterically to alter topography or charge exposure in the pore lumen. Also in Cx26, in another study, V84L, V95M, and A88S (all in M2), and a T5M mutation of Cx30 (in the amino-terminal domain), affected the ability to support intercellular Ca2þ signaling without effects on junctional conductance [209]. For these mutants, the junctional permeability to IP3 was apparently affected, but single-channel measurements were not made. In spite of the intriguing effects of these mutants, there is not much of a basis, yet, upon which to speculate regarding the structures, connexin segments, or specific sites of connexin channels that define the observed molecular permeabilities.

7.7 Conclusion While many aspects of connexin pores have not yet been elucidated, there is a wealth of empirical information that is intriguing. Perhaps the most important point is the most obvious — there is a remarkable diversity of unitary conductances of the channels, which says that there is a remarkable diversity of internal pore structures. It follows that the permeability properties of connexins channels are both diverse and likely to be highly specialized. That is, each channel type is likely to be finely tuned regarding the cytoplasmic components that can pass through it and how well. We simply do not yet know the specifics. It may be that each type of connexin channel is as finely turned regarding the spectrum of cytoplasmic components that can permeate, and how well relative to each other, as are the more widely studied ion-specific channels. This implies that while probing the channels with atomic ions and tracer molecules will yield some information about the pore, it will require study of endogenous permeants to reveal the functions for which the channels are optimized. From a

7 Permeability

195

biophysical perspective, the challenge is to determine the precise molecular determinants of the selectivity, and how they are affected by mutations that have biological or medical consequences. From a biological perspective, the challenge is to determine the role that the specific permeabilities discovered play in given contexts — cell biological, developmental, physiological, and pathological. Acknowledgments Work in the authors’ lab included in this chapter was supported by National Institutes of Health (NIH) grants RO1 GM36044, R21 DC07470, R01 NS056509, T32 HL069752, and National Aeronautics and Space Administration (NASA) grant NNJ06HD91G. The authors regret not being able to cite all of the published work that bears on connexin channel permeability.

References 1. Verselis VK, Veenstra RD. Gap junction channels: permeability and voltage gating. In: Hertzberg EL, editor. Gap Junctions. Stamford, CT: JAI Press, 2000:129–92. 2. Veenstra RD. Determining ionic permeabilities of gap junction channels. In: Bruzzone R, Giaume C, editors. Methods in Molecular Biology: Connexin Methods and Protocols. Totowa, NJ: Humana Press, 2001:293–311. 3. Veenstra RD. Ion permeation through connexin gap junction channels: effects on conductance and selectivity. In: Peracchia C, editor. Gap Junctions: Molecular Basis of Cell Communication in Health and Disease. San Diego: Academic Press, 2000:95–129. 4. Trexler EB, Bennett MVL, Bargiello TA, Verselis VK. Voltage gating and permeation in a gap junction hemichannel. Proc Natl Acad Sci USA. 1996;93:5836–41. 5. Wang HZ, Veenstra RD. Monovalent ion selectivity sequences of the rat connexin43 gap junction channel. J Gen Physiol. 1997;109:491–507. 6. Beblo DA, Veenstra RD. Monovalent cation permeation through the connexin40 gap junction channel. J Gen Physiol. 1997;109:509–22. 7. Hu X, Ma M, Dahl G. Conductance of connexin hemichannels segregates with the first transmembrane segment. Biophys J. 2006;90:140–50. 8. Borisova MP, Brutyan RA, Ermishkin LN. Mechanism of anion-cation selectivity of amphotericin B channels. J Membr Biol. 1986;90:13–20. 9. Franciolini F, Nonner W. Anion-cation interactions in the pore of neuronal background chloride channels. J Gen Physiol. 1994;104:711–23. 10. Franciolini F, Nonner W. A multi-ion permeation mechanism in neuronal background chloride channels. J Gen Physiol. 1994;104:725–46. 11. Schirmer T, Keller TA, Wang Y-F, Rosenbusch JP. Structural basis for sugar translocation through maltoporin channels at 3.1 A˚ resolution. Science. 1995;267:512–14. 12. Wang Y-F, Dutzler R, Rizkallah PJ, Rosenbusch JP, Schirmer T. Channel specificity: structural basis for sugar discrimination and differential flux rates in maltoporin. J Mol Biol. 1997;272:56–63. 13. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys. 2001;34:325–472. 14. Bevans CG, Kordel M, Rhee SK, Harris AL. Isoform composition of connexin channels determines selectivity among second messengers and uncharged molecules. J Biol Chem. 1998;273:2808–16. 15. Locke D, Stein T, Davies C, Morris J, Harris AL, Evans WH, Monaghan P, Gusterson B. Altered permeability and modulatory character of connexin channels during mammary gland development. Exp Cell Res. 2004;298: 643–60.

196

A.L. Harris and D. Locke

16. Locke D, Koreen IV, Liu JL, Harris AL. Reversible pore block of connexin channels by cyclodextrins. J Biol Chem. 2004;279:22883–92. 17. Sabirov RZ, Krasilnikov OV, Ternovsky VI, Merzliak PG. Relation between ionic channel conductance and conductivity of media containing different nonelectrolytes. A novel method of pore size determination. Gen Physiol Biophys. 1993;12:95–111. 18. Krasilnikov OV, Sabirov RZ, Ternovsky VI, Merzliak PG, Muratkhodjaev JN. A simple method for the determination of the pore radius of ion channels in planar lipid bilayer membranes. FEMS Microbiol Immunol. 1992;5:93–100. 19. Bezrukov SM, Vodyanoy I. Probing alamethicin channels with water-soluble polymers. Effect on conductance of channel states. Biophys J. 1993;64:16–25. 20. Oh S, Ri Y, Bennett MVL, Trexler EB, Verselis VK, Bargiello TA. Changes in permeability caused by connexin32 mutations underlie X-linked Charcot-Marie-Tooth disease. Neuron. 1997;19:927–38. 21. Gong XQ, Nicholson BJ. Size selectivity between gap junction channels composed of different connexins. Cell Commun Adhes. 2001;8:187–92. 22. Qu Y, Dahl G. Accessibility of Cx46 hemichannels for uncharged molecules and its modulation by voltage. Biophys J. 2004;86:1502–9. 23. Ma M, Dahl G. Cosegregation of permeability and single-channel conductance in chimeric connexins. Biophys J. 2006;90:151–63. 24. Bukauskas FF, Elfgang C, Willecke K, Weingart R. Heterotypic gap junction channels (connexin26–connexin32) violate the paradigm of unitary conductance. Pflu¨gers Arch. 1995;429:870–2. 25. Suchyna TM, Nitsche JM, Chilton MG, Harris AL, Veenstra RD, Nicholson BJ. Different ionic selectivities for connexins 26 and 32 produce rectifying gap junction channels. Biophys J. 1999;77:2968–87. 26. Elfgang C, Eckert R, Lichtenbergfrate H, Butterweck A, Traub O, Klein RA, Hulser DF, Willecke K. Specific permeability and selective formation of gap junction channels in connexin-transfected HeLa cells. J Cell Biol. 1995;129:805–17. 27. Cao FL, Eckert R, Elfgang C, Nitsche JM, Snyder SA, Hulser DF, Willecke K, Nicholson BJ. A quantitative analysis of connexin-specific permeability differences of gap junctions expressed in HeLa transfectants and Xenopus oocytes. J Cell Sci. 1998;111:31–43. 28. Veenstra RD, Wang H-Z, Beblo DA, Chilton MG, Harris AL, Beyer EC, Brink PR. Selectivity of connexin-specific gap junctions does not correlate with channel conductance. Circ Res. 1995;77:1156–65. 29. Edidin M, Zagyansky Y, Lardner TJ. Measurement of membrane protein lateral diffusion in single cells. Science. 1976;191:466–68. 30. Peters R, Peters J, Tews KH, Bahr W. A microfluorometric study of translational diffusion in erythrocyte membranes. Biochim Biophys Acta. 1974;367:282–94. 31. Axelrod D, Koppel DE, Schlessinger J, Elson E, Webb WW. Mobility measurement by analysis of fluorescence photobleaching recovery kinetics. Biophys J. 1976;16: 1055–67. 32. Koppel DE. Fluorescence redistribution after photobleaching: a new multipoint analysis of membrane translational dynamics. Biophys J. 1979;28:281–92. 33. Deleze J, Delage B, Hentai-Ksibi O, Verrecchia F, Herve JC. Fluorescence recovery after photobleaching. In: Bruzzone R, Giaume C, editors. Methods in Molecular Biology: Connexin Methods and Protocols. Totowa, NJ: Humana Press, 2001:313–27. 34. Tsien RY, Pozzan T, Rink TJ. Calcium homeostasis in intact lymphocytes: cytoplasmic free calcium monitored with a new, intracellularly trapped fluorescent indicator. J Cell Biol. 1982;94:325–34. 35. Goodall H, Johnson MH. Use of carboxyfluorescein diacetate to study formation of permeable channels between mouse blastomeres. Nature. 1982;292:524–6. 36. Rotman B, Papermaster BW. Membrane properties of living mammalian cells as studied by enzymatic hydrolysis of fluorogenic esters. Proc Natl Acad Sci USA. 1966;55:134–41.

7 Permeability

197

37. Abbaci M, Barberi-Heyob M, Stines JR, Blondel W, Dumas D, Guillemin F, Didelon J. Gap junctional intercellular communication capacity by gap-FRAP technique: a comparative study. Biotechnol J. 2007;2:50–61. 38. Wade MH, Trosko JE, Schindler M. A fluorescence photobleaching assay of gap junctionmediated communication between human cells. Science. 1986;232:525–28. 39. Suter S, Trosko JE, el-Fouly MH, Lockwood LR, Koestner A. Dieldrin inhibition of gap junctional intercellular communication in rat glial cells as measured by the fluorescence photobleaching and scrape loading/dye transfer assays. Fundam Appl Toxicol. 1987;9:785–94. 40. Miller A. Quantitative junctional permeability measurements using the confocal microscope. Microscop Res Tech. 1995;31:387–95. 41. Miller AG, Hall JE. Junctional permeability measurements in the embryonic chick lens. Exp Eye Res. 1996;62:339–49. 42. Wilders R, Jongsma HJ. Limitations of the dual voltage clamp method in assaying conductance and kinetics of gap junction channels. Biophys J. 1992;63:942–53. 43. Ramanan SV, Mesimeris V, Brink PR. Ion flow in the bath and flux interactions between channels. Biophys J. 1994;66:989–95. 44. Hall JE, Gourdie RG. Spatial organization of cardiac gap junctions can affect access resistance. Microscop Res Tech. 1995;31:446–51. 45. Ek-Vitorin JF, Burt JM. Quantification of gap junction selectivity. Am J Physiol Cell Physiol. 2005;289:C1535–46. 46. Eckert R. Gap-junctional single-channel permeability for fluorescent tracers in mammalian cell cultures. Biophys J. 2006;91:565–79. 47. Weber PA, Chang H-C, Spaeth KE, Nitsche JM, Nicholson BJ. The permeability of gap junction channels to probes of different size is dependent on connexin composition and permeant-pore affinities. Biophys J. 2004;87:958–73. 48. Nitsche JM, Chang HC, Weber PA, Nicholson BJ. A transient diffusion model yields unitary gap junctional permeabilities from images of cell-to-cell fluorescent dye transfer between Xenopus oocytes. Biophys J. 2004;86:2058–77. 49. Valiunas V, Beyer EC, Brink PR. Cardiac gap junction channels show quantitative differences in selectivity. Circ Res. 2002;91:104–11. 50. Eckert R, Adams B, Kistler J, Donaldson P. Quantitative determination of gap junctional permeability in the lens cortex. J Membr Biol. 1999;169:91–102. 51. Nicholson BJ, Weber PA, Cao F, Chang HC, Lampe P, Goldberg GS. The molecular basis of selective permeability of connexins is complex and includes both size and charge. Braz J Med Biol Res. 2000;33:369–78. 52. Heyman NS, Burt JM. Hindered diffusion through an aqueous pore describes invariant dye selectivity of Cx43 junctions. Biophys J. 2008;94:840–54. 53. Verselis VK, White RL, Spray DC, Bennett MVL. Gap junctional conductance and permeability are linearly related. Science. 1986;234:461–4. 54. Bukauskas FF, Kempf C, Weingart R. Cytoplasmic bridges and gap junctions in an insect cell line (Aedes albopictus). Exp Physiol. 1992;77:903–11. 55. Valiunas V, Manthey D, Vogel R, Willecke K, Weingart R. Biophysical properties of mouse connexin30 gap junction channels studied in transfected human HeLa cells. J Physiol. 1999;519:631–44. 56. Mills SL, Massey SC. A series of biotinylated tracers distinguishes three types of gap junction in retina. J Neurosci. 2000;20:8629–36. 57. Hoshi H, O’Brien J, Mills SL. A novel fluorescent tracer for visualizing coupled cells in neural circuits of living tissue. J Histochem Cytochem. 2006;54:1169–76. 58. Sotelo C, Alvarado-Mallart RM. Embryonic and adult neurons interact to allow Purkinje cell replacement in mutant cerebellum. Nature. 1987;327:421–3. 59. Lewis CA, Stevens CF. Acetylcholine receptor channel ionic sensitivity: ions experience an aqueous environment. Proc Natl Acad Sci USA. 1983;80:6110–3.

198

A.L. Harris and D. Locke

60. Qu Y, Dahl G. Function of the voltage gate of gap junction channels: selective exclusion of molecules. Proc Natl Acad Sci USA. 2002;99:697–702. 61. Ek-Vitorin JF, King TJ, Heyman NS, Lampe PD, Burt JM. Selectivity of connexin 43 channels is regulated through protein kinase C-dependent phosphorylation. Circ Res. 2006;98:1498–505. 62. Bukauskas FF, Bukauskiene A, Verselis VK. Conductance and permeability of the residual state of connexin43 gap junction channels. J Gen Physiol. 2002;119:171–85. 63. Rackauskas M, Verselis VK, Bukauskas FF. Permeability of homotypic and heterotypic gap junction channels formed of cardiac connexins mCx30.2, Cx40, Cx43 and Cx45. Am J Physiol Heart Circ Physiol. 2007;293:H1729–36. 64. Yum SW, Zhang J, Valiunas V, Kanaporis G, Brink PR, White TW, Scherer SS. Human connexin26 and connexin30 form functional heteromeric and heterotypic channels. Am J Physiol Cell Physiol. 2007;293: C1032–48. 65. Sun J, Ahmad S, Chen S, Tang W, Zhang Y, Chen P, Lin X. Cochlear gap junctions coassembled from Cx26 and Cx30 show faster intercellular Ca2þ signaling than homomeric counterparts. Am J Physiol Cell Physiol. 2005;288:C613–23. 66. Martı´ nez AD, Hayrapetyan V, Moreno AP, Beyer EC. Connexin43 and connexin45 form heteromeric gap junction channels in which individual components determine permeability and regulation. Circ Res. 2002;90:1100–7. 67. Lawrence TS, Beers WH, Gilula NB. Transmission of hormonal stimulation by cell-tocell communication. Nature. 1978;272:501–6. 68. Harris AL. Connexin channel permeability to cytoplasmic molecules. Prog Biophys Mol Biol. 2007;94:120–43. 69. Pearson RA, Luneborg NL, Becker DL, Mobbs P. Gap junctions modulate interkinetic nuclear movement in retinal progenitor cells. J Neurosci. 2005;25:10803–14. 70. Veenstra RD. Size and selectivity of gap junction channels formed from different connexins. J Bioenerg Biomembr. 1996;28:327–37. 71. Goldberg GS, Valiunas V, Brink PR. Selective permeability of gap junction channels. Biochim Biophys Acta. 2004;1662:96–101. 72. Bedner P, Niessen H, Odermatt B, Kretz M, Willecke K, Harz H. Selective permeability of different connexin channels to the second messenger cyclic AMP. J Biol Chem. 2006;281:6673–81. 73. Hernandez VH, Bortolozzi M, Pertegato V, Beltramello M, Giarin M, Zaccolo M, Pantano S, Mammano F. Unitary permeability of gap junction channels to second messengers measured by FRET microscopy. Nat Methods. 2007;4:353–58. 74. Harris AL. Connexin channel permeability to cytoplasmic molecules. Prog Biophys Mol Biol. 2007;94:120–43. 75. Henriksen Z, Hiken JF, Steinberg TH, Jorgensen NR. The predominant mechanism of intercellular calcium wave propagation changes during long-term culture of human osteoblast-like cells. Cell Calcium. 2006;39:435–44. 76. Jorgensen NR, Henriksen Z, Sorensen OH, Eriksen EF, Civitelli R, Steinberg TH. Intercellular calcium signaling occurs between human osteoblasts and osteoclasts and requires activation of osteoclast P2X7 receptors. J Biol Chem. 2002;277:7574–80. 77. Braet K, Paemeleire K, D’Herde K, Sanderson MJ, Leybaert L. Astrocyte-endothelial cell calcium signals conveyed by two signalling pathways. Eur J Neurosci. 2001;13:79–91. 78. Boitano S, Dirksen ER, Sanderson MJ. Intercellular propagation of calcium waves mediated by inositol trisphosphate. Science. 1992;258:292–5. 79. Homolya L, Steinberg TH, Boucher RC. Cell to cell communication in response to mechanical stress via bilateral release of ATP and UTP in polarized epithelia. J Cell Biol. 2000;150:1349–59. 80. Sanderson MJ. Intercellular waves of communication. News Physiol Sci. 1996;11:262–9. 81. Giaume C, Venance L. Intercellular calcium signaling and gap junctional communication in astrocytes. Glia. 1998;24:50–64.

7 Permeability

199

82. Schuster S, Marhl M, Hofer T. Modelling of simple and complex calcium oscillations— from single-cell responses to intercellular signalling. Eur J Biochem. 2002;269:1333–55. 83. Bennett MR, Buljan V, Farnell L, Gibson WG. Purinergic junctional transmission and propagation of calcium waves in spinal cord astrocyte networks. Biophys J. 2006;91: 3560–71. 84. Scemes E, Giaume C. Astrocyte calcium waves: what they are and what they do. Glia. 2006;54:716–25. 85. Bowser DN, Khakh BS. Vesicular ATP is the predominant cause of intercellular calcium waves in astrocytes. J Gen Physiol. 2007;129:485–91. 86. Leybaert L, Paemeleire K, Strahonja A, Sanderson MJ. Inositol trisphosphate-dependent intercellular calcium signaling in and between astrocytes and endothelial cells. Glia. 1998;24:398–407. 87. Venance L, Stella N, Glowinski J, Giaume C. Mechanism involved in initiation and propagation of receptor-induced intercellular calcium signaling in cultured rat astrocytes. J Neurosci. 1997;17:1981–92. 88. Sa´ez JC, Connor JA, Spray DC, Bennett MVL. Hepatocyte gap junctions are permeable to the second messenger, inositol 1,4,5-trisphosphate, and to calcium ions. Proc Natl Acad Sci USA. 1989;86:2708–12. 89. Christ GJ, Moreno AP, Melman A, Spray DC. Gap junction-mediated intercellular diffusion of Ca2þ in cultured human corporal smooth muscle cells. Am J Physiol. 1992;263:C373–83. 90. Isakson BE, Ramos SI, Duling BR. Ca2þ and inositol 1,4,5-trisphosphate-mediated signaling across the myoendothelial junction. Circ Res. 2007;100:246–54. 91. Sneyd J, Charles AC, Sanderson MJ. A model for the propagation of intercellular calcium waves. Am J Physiol. 1994;266:C293–302. 92. Sneyd J, Wilkins M, Strahonja A, Sanderson MJ. Calcium waves and oscillations driven by an intercellular gradient of inositol (1,4,5)-triphosphate. Biophys Chem. 1998;72:101–9. 93. Iacobas DA, Suadicani SO, Spray DC, Scemes E. A stochastic two-dimensional model of intercellular Ca2þ wave spread in glia. Biophys J. 2006;90:24–41. 94. Hofer T, Venance L, Giaume C. Control and plasticity of intercellular calcium waves in astrocytes: a modeling approach. J Neurosci. 2002;22:4850–9. 95. Hofer T. Model of intercellular calcium oscillations in hepatocytes: synchronization of heterogeneous cells. Biophys J. 1999;77:1244–56. 96. Baranova A, Ivanov D, Petrash N, Pestova A, Skoblov M, Kelmanson I, Shagin D, Nazarenko S, Geraymovych E, Litvin O, Tiunova A, Born TL, Usman N, Staroverov D, Lukyanov S, Panchin Y. The mammalian pannexin family is homologous to the invertebrate innexin gap junction proteins. Genomics. 2004;83:706–16. 97. Bruzzone R, Hormuzdi SG, Barbe MT, Herb A, Monyer H. Pannexins, a family of gap junction proteins expressed in brain. Proc Natl Acad Sci USA. 2003;100:13644–49. 98. Barbe MT, Monyer H, Bruzzone R. Cell-cell communication beyond connexins: the pannexin channels. Physiology. 2006;21:103–14. 99. Vanden Abeele F, Bidaux G, Gordienko D, Beck B, Panchin YV, Baranova AV, Ivanov DV, Skryma R, Prevarskaya N. Functional implications of calcium permeability of the channel formed by pannexin 1. J Cell Biol. 2006;174:535–46. 100. Huang Y, Grinspan JB, Abrams CK, Scherer SS. Pannexin1 is expressed by neurons and glia but does not form gap functional gap junctions. Glia. 2006;55:46–56. 101. Penuela S, Bhalla R, Gong XQ, Cowan KN, Celetti SJ, Cowan BJ, Bai D, Shao Q, Laird DW. Pannexin 1 and pannexin 3 are glycoproteins that exhibit many distinct characteristics from the connexin family of gap junction proteins. J Cell Sci. 2007;120:3772–83. 102. Locovei S, Bao L, Dahl G. Pannexin 1 in erythrocytes: function without a gap. Proc Natl Acad Sci USA. 2006;103:7655–9. 103. Litvin O, Tiunova A, Connel-Alberts Y, Panchin Y, Baranova A. What is hidden in the pannexin treasure trove: the sneak peek and the guesswork. J Cell Mol Med. 2006;10:613–34.

200

A.L. Harris and D. Locke

104. Dvoriantchikova G, Ivanov D, Pestova A, Shestopalov V. Molecular characterization of pannexins in the lens. Mol Vis. 2006;12:1417–26. 105. Ray A, Zoidl G, Wahle P, Dermietzel R. Pannexin expression in the cerebellum. Cerebellum. 2006;5:189–92. 106. Dahl G, Locovei S. Pannexin: to gap or not to gap, is that the question? IUBMB Life. 2006;58:409–19. 107. Hernandez VH, Bortolozzi M, Pertegato V, Beltramello M, Giarin M, Pantano S, Mammano F. Unitary permeability of gap junction channels to second messengers measured by FRET microscopy and dual whole-cell recordings. Nat Methods. 2007;4:353–8. 108. Ponsioen B, Van Zeijl L, Moolenaar WH, Jalink K. Direct measurement of cyclic AMP diffusion and signaling through connexin43 gap junctional channels. Exp Cell Res. 2007;313:415–23. 109. Goldberg GS, Lampe PD, Sheedy D, Stewart CC, Nicholson BJ, Naus CCG. Direct isolation and analysis of endogenous transjunctional ADP from Cx43 transfected C6 glioma cells. Exp Cell Res. 1998;239:82–92. 110. Goldberg GS, Lampe PD, Nicholson BJ. Selective transfer of endogenous metabolites through gap junctions composed of different connexins. Nat Cell Biol. 1999;1:457–9. 111. Goldberg GS, Lampe P. Capture of transjunctional metabolites. In: Bruzzone R, Giaume C, eds. Methods in Molecular Biology: Connexin Methods and Protocols. Totowa, NJ: Humana Press, 2001:329–40. 112. Goldberg GS, Moreno AP, Lampe PD. Gap junctions between cells expressing connexin 43 or 32 show inverse permselectivity to adenosine and ATP. J Biol Chem. 2002;277:36725–30. 113. Alexander DB, Goldberg GS. Transfer of biologically important molecules between cells through gap junction channels. Curr Med Chem. 2003;10:2045–58. 114. Spray DC, Ye Z-C, Ransom BR. Functional connexin ‘hemichannels’: a critical appraisal. Glia. 2006;54:758–73. 115. Pelegrin P, Surprenant A. Pannexin-1 mediates large pore formation and interleukin-1b release by the ATP-gated P2X7 receptor. EMBO J. 2006;25:5071–82. 116. Locovei S, Scemes E, Qiu F, Spray DC, Dahl G. Pannexin1 is part of the pore forming unit of the P2X(7) receptor death complex. FEBS Letts. 2007;581:483–8. 117. Alves LA, Coutinho-Silva R, Persechini PM, Spray DC, Savino W, Campos de Carvalho AC. Are there functional gap junctions or junctional hemichannels in macrophages? Blood. 1996;88:328–34. 118. Surprenant A, Rassendren F, Kawashiima E, North RA, Buell G. The cytolytic P2z receptor for extracellular ATP identified as a P2x receptor (P2x7). Science. 1996;272: 735–8. 119. Rassendren F, Buell GN, Virginio C, Collo G, North RA, Surprenant A. The permeabilizing ATP receptor, P2X7. J Biol Chem. 1997;272:5482–6. 120. Virginio C, Church D, North RA, Surprenant A. Effects of divalent cations, protons and calmidazolium at the rat P2X7 receptor. Neuropharmacol. 1997;36:1285–94. 121. Duan S, Anderson CM, Keung EC, Chen Y, Swanson RA. P2X7 receptor-mediated release of excitatory amino acids from astrocytes. J Neurosci. 2003;23:1320–8. 122. Bisaggio RD, Nihei OK, Persechini PM, Savino W, Alves LA. Characterization of P2 receptors in thymic epithelial cells. Cell Mol Biol. 2001;47:19–31. 123. Ballerini P, Rathbone MP, Di Ioro P, et al. Rat astroglial P2Z (P2X7) receptors regulate intracellular calcium and purine release. Neuroreport. 1996;7:2533–37. 124. Suadicani SO, Brosnan CF, Scemes E. P2X7 receptors mediate ATP release and amplification of astrocytic intercellular Ca2þ signaling. J Neurosci. 2006;26:1378–85.

7 Permeability

201

125. Parpura V, Scemes E, Spray DC. Mechanisms of glutamate release from astrocytes: gap junction ‘hemichannels’, purinergic receptors and exocytotic release. Neurochem Int. 2004;45:259–64. 126. North RA. Molecular physiology of P2X receptors. Physiol Rev. 2002;82:1013–67. 127. Bao L, Locovei S, Dahl G. Pannexin membrane channels are mechanosensitive conduits for ATP. FEBS Letts. 2004;572:65–8. 128. Ebihara L, Steiner E. Properties of a nonjunctional current expressed from a rat connexin46 cDNA in Xenopus oocytes. J Gen Physiol. 1993;102:59–74. 129. Ebihara L. Xenopus connexin38 forms hemi-gap junctional channels in the nonjunctional plasma membrane of Xenopus oocytes. Biophys J. 1996;71:742–8. 130. Ebihara L, Liu X, Pal JD. Effect of external magnesium and calcium on human connexin46 hemichannels. Biophys J. 2003;84:277–86. 131. Beahm DL, Hall JE. Hemichannel and junctional properties of connexin 50. Biophys J. 2002;82:2016–31. 132. Bruzzone R, Barbe MT, Jakob NJ, Monyer H. Pharmacological properties of homomeric and heteromeric pannexin hemichannels expressed in Xenopus oocytes. J Neurochem. 2005;92:1033–43. 133. Wang J, Ma M, Locovei S, Keane RW, Dahl G. Modulation of membrane channel currents by gap junction protein mimetic peptides: size matters. Am J Physiol Cell Physiol. 2007;293:C1112–9. 134. Cotrina ML, Lin JHC, Alves-Rodrigues A, Liu S, Li J, Azmi-Ghadimi H, Kang J, Naus CCG, Nedergaard M. Connexins regulate calcium signaling by controlling ATP release. Proc Natl Acad Sci USA. 1998;95:15735–40. 135. Cotrina ML, Lin JHC, Lopez-Garcia JC, Naus CCG, Nedergaard M. ATP-mediated glia signaling. J Neurosci. 2000;20:2835–44. 136. Genetos DC, Kephart CJ, Zhang Y, Yellowley CE, Donahue HJ. Oscillating fluid flow activation of gap junction hemichannels induces ATP release from MLO-Y4 osteocytes. J Cell Physiol. 2007;212:207–14. 137. Naus CCG, Bond SL, Bechberger JF, Rushlow W. Identification of genes differentially expressed in C6 glioma cells transfected with connexin43. Brain Res Rev. 2000;32: 259–66. 138. Iacobas DA, Urban-Maldonado M, Iacobas S, Scemes E, Spray DC. Array analysis of gene expression in connexin-43 null astrocytes. Physiol Genomics. 2003;15:177–90. 139. Iacobas DA, Scemes E, Spray DC. Gene expression alterations in connexin null mice extend beyond the gap junction. Neurochem Int. 2004;45:243–50. 140. Iacobas DA, Iacobas S, Li WE, Zoidl G, Dermietzel R, Spray DC. Genes controlling multiple functional pathways are transcriptionally regulated in connexin43 null mouse heart. Physiol Genomics. 2005;20:211–23. 141. Sa´ez JC, Contreras JE, Bukauskas FF, Retamal MA, Bennett MV. Gap junction hemichannels in astrocytes of the CNS. Acta Physiol Scand. 2003;179:9–22. 142. Goodenough DA, Paul DL. Beyond the gap: functions of unpaired connexon channels. Nat Rev Mol Cell Biol. 2003;4:285–94. 143. Bennett MV, Contreras JE, Bukauskas FF, Sa´ez JC. New roles for astrocytes: gap junction hemichannels have something to communicate. Trends Neurosci. 2003;26: 610–17. 144. Contreras JE, Sanchez HA, Veliz LP, Bukauskas FF, Bennett MV, Sa´ez JC. Role of connexin-based gap junction channels and hemichannels in ischemia-induced cell death in nervous tissue. Brain Res Brain Res Rev. 2004;47:290–303. 145. Evans WH, De Vuyst E, Leybaert L. The gap junction cellular internet: connexin hemichannels enter the signalling limelight. Biochem J. 2006;397:1–14. 146. Bosma MM. Anion channels with multiple conductance levels in a mouse B lymphocyte cell line. J Physiol. 1989;410:67–90. 147. Sabirov RZ, Okada Y. Wide nanoscopic pore of maxi-anion channel suits its function as an ATP-conductive pathway. Biophys J. 2004;87:1672–85.

202

A.L. Harris and D. Locke

148. Reisin IL, Prat AG, Abraham EH, Amara JF, Gregory RJ, Ausiello DA, Cantiello HF. The cystic fibrosis transmembrane conductance regulator is a dual ATP and chloride channel. J Biol Chem. 1994;269:20584–91. 149. Dermietzel R, Hwang TK, Buettner R, Hofer A, Dotzler E, Kremer M, Deutzmann R, Thinnes FP, Fishman GI, Spray DC, Siemen D. Cloning and in situ localization of a brain-derived porin that constitutes a large-conductance anion channel in astrocytic plasma membrane. Proc Natl Acad Sci USA. 1994;91:499–503. 150. Okada SF, O’Neal WK, Huang P, Nicholas RA, Ostrowski LE, Craigen WJ, Lazarowski ER, Boucher RC. Voltage-dependent anion channel-1 (VDAC-1) contributes to ATP release and cell volume regulation in murine cells. J Gen Physiol. 2004;124:513–26. 151. Rostovtseva TK, Colombini M. ATP flux is controlled by a voltage-gated channel from the mitochondrial outer membrane. J Biol Chem. 1996;271:28006–8. 152. Takano T, Kang J, Jaiswal JK, Simon SM, Lin JH, Yu Y, Li Y, Yang J, Dienel G, Zielke HR, Nedergaard M. Receptor-mediated glutamate release from volume sensitive channels in astrocytes. Proc Natl Acad Sci USA. 2005;102:16466–71. 153. Hisadome K, Koyama T, Kimura C, Droogmans G, Ito Y, Oike M. Volume-regulated anion channels serve as an auto/paracrine nucleotide release pathway in aortic endothelial cells. J Gen Physiol. 2002;119:511–20. 154. Rutledge EM, Mongin AA, Kimelberg HK. Intracellular ATP depletion inhibits swelling-induced D-[3H]aspartate release from primary astrocyte cultures. Brain Res. 1999;842:39–45. 155. Bader P, Weingart R. Pitfalls when examining gap junction hemichannels: interference from volume-regulated anion channels. Pflu¨gers Arch. 2006;452:396–406. 156. Bedner P, Niessen H, Odermatt B, Willecke K, Harz H. A method to determine the relative cAMP permeability of connexin channels. Exp Cell Res. 2003;291:25–35. 157. Ponsioen B, Zhao J, Riedl J, Zwartkruis F, van der Krogt G, Zaccolo M, Moolenaar WH, Bos JL, Jalink K. Detecting cAMP-induced Epac activation by fluorescence resonance energy transfer: Epac as a novel cAMP sensor. EMBO Rep. 2004;5: 1176–80. 158. Tanimura A, Nezu A, Morita T, Turner RJ, Tojyo Y. Fluorescent biosensor for quantitative real-time measurements of inositol 1,4,5-trisphosphate in single living cells. J Biol Chem. 2004;279:38095–8. 159. Moreno AP, Fishman GI, Spray DC. Phosphorylation shifts unitary conductance and modifies voltage dependent kinetics of human connexin43 gap junction channels. Biophys J. 1992;62:51–3. 160. Bao X, Lee SC, Reuss L, Altenberg GA. Change in permeant size selectivity by phosphorylation of connexin 43 gap-junctional hemichannels by PKC. Proc Natl Acad Sci USA. 2007;104:4919–24. 161. Locke D, Koreen IV, Harris AL. Isoelectric points and post-translational modifications of connexin26 and connexin32. FASEB J. 2006;20:1221–3. 162. Charpantier E, Cancela J, Meda P. b cells preferentially exchange cationic molecules via connexin 36 gap junction channels. Diabetologia. 2007;50:2332–41. 163. Valiunas V, Polosina YY, Miller H, Potapova IA, Valiuniene L, Doronin S, Mathias RT, Robinson RB, Rosen MR, Cohen IS, Brink PR. Connexin-specific cell-to-cell transfer of short interfering RNA by gap junctions. J Physiol. 2005;568:459–68. 164. Neijssen J, Herberts C, Drijfhout JW, Reits E, Janssen L, Neefjes J. Cross-presentation by intercellular peptide transfer through gap junctions. Nature. 2005;434:83–8. 165. Mendoza-Naranjo A, Sa´ez PJ, Johansson CC, Ramirez M, Mandakovic D, Pereda C, Lopez MN, Kiessling R, Saez JC, Salazar-Onfray F. Functional gap junctions facilitate melanoma antigen transfer and cross-presentation between human dendritic cells. J Immunol. 2007;178:6949–57. 166. Ayad WA, Locke D, Koreen IV, Harris AL. Heteromeric, but not homomeric, connexin channels are selectively permeable to inositol phosphates. J Biol Chem. 2006;281: 16727–39.

7 Permeability

203

167. Niessen H, Harz H, Bedner P, Kramer K, Willecke K. Selective permeability of different connexin channels to the second messenger inositol 1,4,5-trisphosphate. J Cell Sci. 2000;113:1365–72. 168. Bacskai BJ, Hochner B, Mahaut-Smith M, Adams SR, Kaang BK, Kandel ER, Tsien RY. Spatially resolved dynamics of cAMP and protein kinase A subunits in Aplysia sensory neurons. Science. 1993;260:222–6. 169. Kasai H, Petersen OH. Spatial dynamics of second messengers: IP3 and cAMP as longrange and associative messengers. Trends Neurosci. 1994;17:95–101. 170. Wang SS-H, Alousi AA, Thompson SH. The lifetime of inositol 1,4,5-trisphosphate in single cells. J Gen Physiol. 1995;105:149–71. 171. Fink CC, Slepchenko B, Moraru II, Schaff J, Watras J, Loew LM. Morphological control of inositol-1,4,5-trisphosphate-dependent signals. J Cell Biol. 1999;147: 929–36. 172. Allbritton NL, Meyer T, Stryer L. Range of messenger action of calcium and inositol 1,4,5-trisphosphate. Science. 1992;258:1812–5. 173. Huang R-C, Gillette R. Kinetic analysis of cAMP-activated Naþ current in the molluscan neuron. J Gen Physiol. 1991;98:835–48. 174. Ramanan SV, Brink PR, Christ GJ. Neuronal innervation, intracellular signal transduction and intercellular coupling: a model for syncytial tissue responses in the steady state. J Theor Biol. 1998;193:69–84. 175. Hajnuczky G, Robb-Gaspers LD, Seitz MB, Thomas AP. Decoding of cytosolic calcium oscillations in the mitochondria. Cell. 1995;82:415–24. 176. Jafri MS, Keizer J. On the roles of Ca2þ diffusion, Ca2þ buffers, and the endoplasmic reticulum in IP3-induced Ca2þ waves. Biophys J. 1995;69:2139–53. 177. Dupont G, Swillens S, Clair C, Tordjmann T, Combettes L. Hierarchical organization of calcium signals in hepatocytes: from experiments to models. Biochim Biophys Acta Mol Cell Res. 2000;1498:134–52. 178. Breitwieser GE. Calcium sensing receptors and calcium oscillations: calcium as a first messenger. Curr Top Devel Biol. 2006;73:85–114. 179. Nathanson MH, Burgstahler AD, Mennone A, Fallon MB, Gonzalez CB, Sa´ez JC. Ca2þ waves are organized among hepatocytes in the intact organ. Am J Physiol. 1995;32:G167–71. 180. Robb-Gaspers LD, Thomas AP. Coordination of Ca2þ signaling by intercellular propagation of Ca2þ waves in the intact liver. J Biol Chem. 1995;270:8102–7. 181. D’Andrea P, Vittur F. Gap junctions mediate intercellular calcium signaling in cultured articular chondrocytes. Cell Calcium. 1996;20:389–97. 182. Rottingen JA, Camerer E, Mathiesen I, Prydz H, Iversen JG. Synchronized Ca2þ oscillations induced in Madin-Darby canine kidney cells by bradykinin and thrombin but not by ATP. Cell Calcium. 1997;21:195–211. 183. Tordjmann T, Berthon B, Claret M, Combettes L. Coordinated intercellular calcium waves induced by noradrenaline in rat hepatocytes: dual control by gap junction permeability and agonist. EMBO J. 1997;16:5398–407. 184. Dupont G, Tordjmann T, Clair C, Swillens S, Claret M, Combettes L. Mechanism of receptor-oriented intercellular calcium wave propagation in hepatocytes. FASEB J. 2000;14:279–89. 185. Ravier MA, Guldenagel M, Charollais A, Gjinovci A, Caille D, Sohl ¨ G, Wollheim CB, Willecke K, Henquin JC, Meda P. Loss of connexin36 channels alters b-cell coupling, islet synchronization of glucose-induced Ca2þ and insulin oscillations, and basal insulin release. Diabetes. 2005;54:1798–807. 186. Haddock RE, Grayson TH, Brackenbury TD, Meaney KR, Neylon CB, Sandow SL, Hill CE. Endothelial coordination of cerebral vasomotion via myoendothelial gap junctions containing connexins 37 and 40. Am J Physiol Heart Circ Physiol. 2006;291:H2047–56.

204

A.L. Harris and D. Locke

187. Rooney T, Sass E, Thomas AP. Characterization of cytosolic calcium oscillations induced by phenylephrine and vasopressin in single fura-2 loaded hepatocytes. J Biol Chem. 1989;264:17131–41. 188. Hofer T, Politi A, Heinrich R. Intercellular Ca2þ wave propagation through gapjunctional Ca2þ diffusion: a theoretical study. Biophys J. 2001;80:75–87. 189. De Blasio BF, Iversen JG, Rottingen JA. Intercellular calcium signalling in cultured renal epithelia: a theoretical study of synchronization mode and pacemaker activity. Eur Biophys J. 2004;33:657–70. 190. Perc M, Marhl M. Local dissipation and coupling properties of cellular oscillators: a case study on calcium oscillations. Bioelectrochem. 2004;62:1–10. 191. Tsaneva-Atanasova K, Zimliki CL, Bertram R, Sherman A. Diffusion of calcium and metabolites in pancreatic islets: killing oscillations with a pitchfork. Biophys J. 2006;90:3434–46. 192. Ullah G, Jung P, Cornell-Bell AH. Anti-phase calcium oscillations in astrocytes via inositol (1,4,5)-trisphosphate regeneration. Cell Calcium. 2006;39:197–208. 193. Dutzler R, Wang Y-F, Rizkallah PJ, Rosenbusch JP, Schirmer T. Crystal structures of various maltooligosaccharides bound to maltoporin reveal a specific sugar translocation pathway. Structure. 1996;4:127–34. 194. Nikaido H. Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev. 2003;67:593–656. 195. Rostovtseva TK, Bezrukov SM. ATP transport though a single mitochondrial channel, VDAC, studied by current fluctuation analysis. Biophys J. 1998;74:2365–73. 196. Rostovtseva TK, Komarov A, Bezrukov SM, Colombini M. VDAC channels differentiate between natural metabolites and synthetic molecules. J Membr Biol. 2002;187: 147–56. 197. Rostovtseva TK, Komarov A, Bezrukov SM, Colombini M. Dynamics of nucleotides in VDAC channels: structure-specific noise generation. Biophys J. 2002;82:193–205. 198. Komarov A, Deng D, Craigen WJ, Colombini M. New insights into the mechanism of permeation through large channels. Biophys J. 2005;89:3950–9. 199. Delcour AH. Solute uptake through general porins. Front Biosci. 2003;8:d1055–71. 200. Danelon C, Nestorovich EM, Winterhalter M, Ceccarelli M, Bezrukov SM. Interaction of zwitterionic penicillins with the OmpF channel facilitates their translocation. Biophys J. 2006;90:1617–27. 201. Nestorovich EM, Danelon C, Winterhalter M, Bezrukov SM. Designed to penetrate: time-resolved interaction of single antibiotic molecules with bacterial pores. Proc Natl Acad Sci USA. 2002;99:9789–94. 202. Cheley S, Gu L-Q, Bayley H. Stochastic sensing of nanomolar inositol 1,4,5-trisphosphate with an engineered pore. Chem Biol. 2002;9:829–38. 203. Berezhkovskii AM, Bezrukov SM. Optimizing transport of metabolites through large channels: molecular sieves with and without binding. Biophys J. 2005;88:L17–9. 204. Bauer WR, Nadler W. Molecular transport through channels and pores: effects of inchannel interactions and blocking. Proc Natl Acad Sci USA. 2006;103:11446–51. 205. Schwarz G, Danelon C, Winterhalter M. On translocation through a membrane channel via an internal binding site: kinetics and voltage dependence. Biophys J. 2003;84:2990–8. 206. Berezhkovskii AM, Bezrukov SM. Channel-facilitated membrane transport: constructive role of particle attraction to the channel pore. Chem Phys. 2005;319:343–9. 207. Beltramello M, Piazza V, Bukauskas FF, Pozzan T, Mammano F. Impaired permeability to Ins(1,4,5)P3 in a mutant connexin underlies recessive hereditary deafness. Nature Cell Biol. 2005;7:63–9. 208. Skerrett IM, Aronowitz J, Shin JH, Cymes G, Kasperek E, Cao FL, Nicholson BJ. Identification of amino acid residues lining the pore of a gap junction channel. J Cell Biol. 2002;159:349–59.

7 Permeability

205

209. Zhang Y, Tang W, Ahmad S, Sipp JA, Chen P, Lin X. Gap junction-mediated intercellular biochemical coupling in cochlear supporting cells is required for normal cochlear functions. Proc Natl Acad Sci USA. 2005;102:15201–6. 210. Valiunas V, Niessen H, Willecke K, Weingart R. Electrophysiological properties of gap junction channels in hepatocytes isolated from connexin32-deficient and wild-type mice. Pflu¨gers Arch. 1999;437:846–56. 211. Vogel R, Valiunas V, Weingart R. Subconductance states of Cx30 gap junction channels: data from transfected HeLa cells versus data from a mathematical model. Biophys J. 2006;91:2337–48. 212. Valiunas V, Weingart R. Electrical properties of gap junction hemichannels identified in transfected HeLa cells. Pflu¨gers Arch. 2000;440:366–79. 213. Abrams CK, Freidin MM, Verselis VK, Bargiello TA, Kelsell DP, Richard G, Bennett MV, Bukauskas FF. Properties of human connexin 31, which is implicated in hereditary dermatological disease and deafness. Proc Natl Acad Sci USA. 2006;103:5213–18. 214. Kreuzberg MM, Sohl ¨ G, Kim JS, Verselis VK, Willecke K, Bukauskas FF. Functional properties of mouse connexin30.2 expressed in the conduction system of the heart. Circ Res. 2005;96:1169–77. 215. Bukauskas FF, Kreuzberg MM, Rackauskas M, Bukauskiene A, Bennett MV, Verselis VK, Willecke K. Properties of mouse connexin 30.2 and human connexin 31.9 hemichannels: implications for atrioventricular conduction in the heart. Proc Natl Acad Sci USA. 2006;103:9726–31. 216. Srinivas M, Rozental R, Kojima T, Dermietzel R, Mehler M, Condorelli DF, Kessler JA, Spray DC. Functional properties of channels formed by the neuronal gap junction protein connexin36. J Neurosci. 1999;19:9848–55. 217. Teubner B, Degen J, Sohl ¨ G, Guldenagel M, Bukauskas FF, Trexler EB, Verselis VK, De Zeeuw CI, Lee CG, Kozak CA, Petrasch-Parwez E, Dermietzel R, Willecke K. Functional expression of the murine connexin 36 gene coding for a neuron-specific gap junctional protein. J Membr Biol. 2000;176:249–62. 218. Veenstra RD, Wang HZ, Beyer EC, Ramanan SV, Brink PR. Connexin37 forms high conductance gap junction channels with subconductance state activity and selective dye and ionic permeabilities. Biophys J. 1994;66:1915–28. 219. Kumari SS, Varadaraj K, Valiunas V, Ramanan SV, Christensen EA, Beyer EC, Brink PR. Functional expression and biophysical properties of polymorphic variants of the human gap junction protein connexin37. Biochem Biophys Res Commun. 2000;274:216–24. 220. Bukauskas FF, Elfgang C, Willecke K, Weingart R. Biophysical properties of gap junction channels formed by mouse connexin40 in induced pairs of transfected HeLa cells. Biophys J. 1995;68:2289–98. 221. Hellmann P, Winterhager E, Spray DC. Properties of connexin40 gap junction channels endogenously expressed and exogenously overexpressed in human choriocarcinoma cell lines. Pflu¨gers Arch. 1996;432:501–9. 222. Fishman GI, Spray DC, Leinwand LA. Molecular characterization and functional expression of the human cardiac gap junction channel. J Cell Biol. 1990;111:589–98. 223. Fishman GI, Moreno AP, Spray DC, Leinwand LA. Functional analysis of human cardiac gap junction channel mutants. Proc Natl Acad Sci USA. 1991;88:3525–9. 224. Moreno AP, Sa´ez JC, Fishman GI, Spray DC. Human connexin43 gap junction channels. Regulation of unitary conductances by phosphorylation. Circ Res. 1994;74:1050–7. 225. Spray DC, Moreno AP, Campos de Carvalho AC. Biophysical properties of the human cardiac gap junction channel. Braz J Med Biol Res. 1993;26:541–52. 226. Traub O, Eckert R, Lichtenberg-Frate H, Elfgang C, Bastide B, Scheidtmann KH, Hu¨lser DF, Willecke K. Immunochemical and electrophysiological characterization of murine connexin40 and -43 in mouse tissues and transfected human cells. Eur J Cell Biol. 1994;64:101–12.

206

A.L. Harris and D. Locke

227. Veenstra RD, Wang HZ, Beyer EC, Brink PR. Selective dye and ionic permeability of gap junction channels formed by connexin45. Circ Res. 1994;75:483–90. 228. Moreno AP, Laing JG, Beyer EC, Spray DC. Properties of gap junction channels formed of connexin 45 endogenously expressed in human hepatoma (SKHep1) cells. Am J Physiol. 1995;37:C356–65. 229. Kwak BR, Hermans MMP, deJonge HR, Lohmann SM, Jongsma HJ, Chanson M. Differential regulation of distinct types of gap junction channels by similar phosphorylating conditions. Mol Biol Cell. 1995;6:1707–19. 230. Valiunas V. Biophysical properties of connexin-45 gap junction hemichannels studied in vertebrate cells. J Gen Physiol. 2002;119:147–64. 231. Verselis VK, Trexler EB, Bukauskas FF. Connexin hemichannels and cell-cell channels: comparison of properties. Braz J Med Biol Res. 2000;33:379–89. 232. Trexler EB, Bukauskas FF, Kronengold J, Bargiello TA, Verselis VK. The first extracellular loop domain is a major determinant of charge selectivity in connexin46 channels. Biophys J. 2000;79:3036–51. 233. Hopperstad MG, Srinivas M, Spray DC. Properties of gap junction channels formed by Cx46 alone and in combination with Cx50. Biophys J. 2000;79:1954–66. 234. Sakai R, Elfgang C, Vogel R, Willecke K, Weingart R. The electrical behaviour of rat connexin46 gap junction channels expressed in transfected HeLa cells. Pflu¨gers Arch. 2003;446:714–27. 235. Pfahnl A, Dahl G. Localization of a voltage gate in connexin46 gap junction hemichannels. Biophys J. 1998;75:2323–31. 236. Teubner B, Odermatt B, Guldenagel M, Sohl ¨ G, Degen J, Bukauskas FF, Kronengold J, Verselis VK, Jung YT, Kozak CA, Schilling K, Willecke K. Functional expression of the new gap junction gene connexin47 transcribed in mouse brain and spinal cord neurons. J Neurosci. 2001;21:1117–26. 237. Srinivas M, Costa M, Gao Y, Fort A, Fishman GL, Spray DC. Voltage dependence of macroscopic and unitary currents of gap junction channels formed by mouse connexin50 expressed in rat neuroblastoma cells. J Physiol. 1999;517:673–89. 238. Srinivas M, Hopperstad MG, Spray DC. Quinine blocks specific gap junction channel subtypes. Proc Natl Acad Sci USA. 2001;98:10942–7. 239. Xu X, Berthoud VM, Beyer EC, Ebihara L. Functional role of the carboxyl terminal domain of human connexin 50 in gap junctional channels. J Membr Biol. 2002;186:101–12. 240. Manthey D, Bukauskas FF, Lee CG, Kozak CA, Willecke K. Molecular cloning and functional expression of the mouse gap junction gene connexin-57 in human HeLa cells. J Biol Chem. 1999;274:14716–23. 241. Lin X, Fenn E, Veenstra RD. An amino-terminal lysine residue of rat connexin40 that is required for spermine block. J Physiol. 2006;570:251–69. 242. Brink PR, Ramanan SV. A model for the diffusion of fluorescent probes in the septate giant axon of earthworm. Biophys J. 1985;48:299–309. 243. Beltramello M, Bicego M, Piazza V, Ciubotaru CD, Mammano F, D’Andrea P. Permeability and gating properties of human connexins 26 and 30 expressed in HeLa cells. Biochem Biophys Res Commun. 2003;305:1024–33. 244. Dong L, Liu X, Li H, Vertel BM, Ebihara L. Role of the N-terminus in permeability of chicken connexin45.6 gap junctional channels. J Physiol. 2006;576:787–99. 245. Becker DL, Mobbs P. Connexin 1 and cell proliferation in the developing chick retina. Exp Neurol. 1999;158:261–3. 246. Heinz WF, Hoh JH. Relative surface charge density mapping with the atomic force microscope. Biophys J. 1999;76:528–38. 247. Manthey D, Banach K, Desplantez T, Lee CG, Kozak CA, Traub O, Weingart R, Willecke K. Intracellular domains of mouse connexin26 and-30 affect diffusional and electrical properties of gap junction channels. J Membr Biol. 2001;181:137–48.

Chapter 8

Pharmacology of Connexin Channels Miduturu Srinivas

Abstract Connexin channels play a wide variety of roles in different cell types and tissues. Genetic and molecular approaches have proven useful for understanding the roles of these proteins in tissue function. Identification and characterization of specific and high-affinity inhibitors of these channels would greatly assist investigation of their physiological function; however, progress in this area has been slow. Nevertheless, recent studies have identified a number of small molecules and peptides that inhibit connexin channels. Although the specificity of these new drugs for connexin channels remains problematic, several of these reagents inhibit channels in an isoform-specific manner and do so with reasonable potency. These reagents are likely to be useful for acute studies that investigate the physiological roles of different connexin channels. In addition, some agents appear to bind within the permeability pathway and may be useful in structure-function studies of the pore. Keywords Blockers Carbenoxolone Polyamines Fenamates 2-Aminoethoxydiphenyl borate Octanol Oleamide Halothane Cyclodextrin Glycyrrhetinic acid Quinine Mefloquine Cx26 Cx32 Cx35 Cx36 Cx37 Cx40 Cx43 Cx46 Cx50

8.1 Introduction A long-standing problem in the gap junction field is the absence of specific highaffinity reagents or blockers for connexin channels. Such agents will be vital for investigating the physiological roles of connexin channels, especially those that affect channels in a connexin-subtype specific manner. High-affinity reagents will also be useful tools for structure-function studies of connexin channels. M. Srinivas (*) Department of Biological Sciences, State University of New York, College of Optometry, Biological Sciences, 33 West 42nd Street, New York, NY 10036, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_8, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

207

208

M. Srinivas

Specific inhibitors of connexin channels may have therapeutic utility, as connexin channels have been proposed as new and promising pharmacological targets in the treatment of epilepsy, cardiac arrhythmia, cancer, stroke, and essential tremor [1,2,3]. Finally, such reagents can also be used to eliminate or reduce coupling of cultured cells, thereby allowing the study of cellular physiologies or processes complicated by junctional coupling. A major reason for the paucity of specific high-affinity inhibitors is the intercellular location of the gap junction channel. This has made it difficult to design high-throughput screening assays, which have been instrumental in the identification of new classes of blockers for other ion channels. Also, the intracellular location of the pore mouth makes the junctional channel an unattractive target for biological toxins. In addition, because junctional channels cannot be studied in isolation, such as in excised patches, detailed studies of the mechanisms of action of drugs and of the structural determinants of inhibition have proven difficult. Knowledge of the structure of a binding site is important because it could facilitate structure-based drug design and lead to synthesis of additional compounds that have potent effects on coupling. Similarly, a description of the physicochemical features of the drug–receptor interaction, which would lead to the generation of pharmacophore models of drugs, is unavailable for many of the current inhibitors. Despite these drawbacks, use of both conventional electrophysiological techniques and novel ex vivo assays utilizing purified protein [4,5] have led to identification of several new classes of drugs that act directly on connexin channels to modulate their function [4,5,6,7,8,9,10]. Structure-activity studies for some of these compounds are in progress, and it is hoped that application of a classic medicinal chemistry approach will lead to new, more specific and higher affinity pharmacological agents. This chapter reviews the current state of knowledge with regard to exogenous pharmacological agents that act on connexin channels. Particular attention is paid to the characteristics of those agents that bear on their experimental and investigational utility. Many agents that affect gap junction coupling do so via downstream signaling mechanisms. This chapter considers only agents that have been shown to, or may, interact directly with connexin channels. In addition, it does not discuss recently identified modifiers of intrinsic connexin gating sensitivities, such as the RXP-E peptide [11] (see Chapter 5).

8.2 Agents that Affect Connexin Channels A number of pharmacological agents inhibit gap junctional communication [12,13], and in some cases also inhibit hemichannel currents in oocytes [14,15]. Uncoupling agents include glycyrrhetinic acid and its derivatives, polyamines and tetraalkylammonium ions [16,17], antimalarial drugs such as quinine and mefloquine [6,8], fenamates [18,19], 2-aminophenoxyborate (2-APB) and derivatives [7,9,10], volatile anesthetics (e.g., halothane and ethrane) [20,21],

8 Pharmacology

209

lipophilic compounds such as long-chain alcohols (e.g., heptanol and octanol) [22,23,24], fatty acid amides including oleamide [25,26], and cyclodextrins [5]. In addition to these compounds, peptides targeting extracellular loops of connexins inhibit junctional coupling [27,28,29]. With the exception of mefloquine and connexin-mimetic peptides, almost all uncoupling agents produce reversible blockade of connexin channels as assessed by electrophysiological, dyecoupling, or other techniques. Many of these agents affect connexin channels only when applied extracellularly and are far less effective when applied intracellularly via patch pipettes. This result has been frequently interpreted as indicating that the binding site is extracellular [7,18]. However, most gap junction inhibitors are membrane-permeant and lack of effect with intracellular application may be explained by rapid diffusion of a hydrophobic drug out of the cell, resulting in ineffective concentrations near an intracellular binding site. This pattern is similar to the block of K+ channels by tertiary amines such as verapamil [30] and methadone [31]. Many of the active drugs are unlikely to directly occlude connexin channels. Analysis of single-channel currents shows that none of these reagents produce fast flickery block, as expected if they bind and unbind in the open channel with rapid kinetics. Single-channel conductance is not reduced in the presence of these drugs; the main change observed at the single-channel level is a reduction in the apparent open probability (Po) [6,17,19,20,32,33]. Even those agents that appear to inhibit connexin channels in a voltage-dependent manner (e.g., polyamines, tetraalkylammonium ions, and possibly quinine), consistent with binding in the pore, predominantly affect the observed Po [16,17]. If the agent acts by direct pore block, this reduction in apparent Po might be explained by slow rates of binding and unbinding. However, detailed studies of the ‘‘on’’ and ‘‘off’’ rates and their dependence on drug concentration are not available for many connexin channel inhibitors. Nevertheless, the fact that many of the active drugs are small compared to the dimensions of the connexin pore makes it less likely that they directly occlude the open channel. In most cases, reduction of coupling or hemichannel macroscopic conductance is likely to be due to an effect on gating. Consistent with the notion of an effect on channel gating, many uncoupling agents induce a particular form of closure, characterized by slow transitions between open and fully closed states [6,19,33]. Such slow transitions can require tens of milliseconds for completion, during which channels might reside in multiple, partially conducting states. Slow transitions were initially described in studies of newly paired insect and mammalian cells, and were associated with channel formation or docking of apposed hemichannels [34,35]. Subsequently, they were described in Cx46 hemichannels where they were attributed to a distinct form of closure involving conformational rearrangements of the extracellular loops, termed slow/loop gating by Verselis et al. [36] (see Chapter 4). Almost all gap junction channel inhibitors and modulators, including low pH, halothane, n-alkanols, quinine, and fenamates, produce these slow transitions, suggesting that structural elements involved in closure are shared.

210

M. Srinivas

8.2.1 Glycyrrhetinic Acid and Derivatives The inhibitory effects of a-glycyrrhetinic acid, b-glycyrrhetinic acid and carbenoxolone on gap junction coupling have been known for over 20 years. Davidson and collaborators [37,38] demonstrated that these agents reversibly inhibited coupling in long-term cultures of human fibroblasts at concentrations as low as 20 mM. Subsequently, they have been shown to inhibit gap junctions in a wide variety of cell types and connexin hemichannels in Xenopus oocytes [14,15]. These agents also block dye uptake in mammalian cells, which is presumably mediated by connexin hemichannels that open as extracellular Ca2+ is lowered. However, recent studies indicate that pannexin channels and P2X7 receptors, which are also permeable to large molecular weight dyes, can be inhibited by derivatives of glycyrrhetinic acid and other such uncoupling agents [14,39,40] (see Chapter 12). Therefore, sole reliance on these agents to demonstrate the existence of functional hemichannels in mammalian cells should be avoided. Glycyrrhetinic acid and derivatives are triterpinoid saponins, naturally found in licorice, which are potent inhibitors of the enzyme 11-hydroxysteroid dehydrogenase that degrades cortisol and cortisone [41]. The uncoupling effect is unlikely to be related to the inhibition of this enzyme because glycyrrhizic acid, an even more potent dehydrogenase inhibitor, has no effect on gap junctions and, for this reason, is frequently used as a negative control. Detailed studies of the connexin-selectivity of glycyrrhetinic acid derivatives are lacking, but their widespread use and efficacy in reducing junctional coupling in a wide variety of cells suggests that large differences in sensitivity among connexin channel subtypes do not exist. However, the magnitude of decrease in junctional current caused by carbenoxolone may be connexin-dependent. In some Cx43-expressing cells, the drug does not inhibit junctional currents completely, i.e., a significant fraction of ionic coupling remains even at the highest concentrations [42,43]. In contrast, junctional coupling in Cx50-expressing cells is completely eliminated by carbenoxolone [12]. Despite the efficacy of these agents, the mechanism of action is largely unexplored. In some studies, carbenoxolone decreases the phosphorylation state of Cx43, leading to plaque disassembly, internalization, and reduction in the expression of Cx43, whereas in other studies there was little change in expression or phosphorylation [42,43,44,45]. Part of the variability may be due to differences in exposure times to the drug and its concentration. Gap junction channel blockade by carbenoxolone generally requires longer exposure times than for other uncoupling agents, but a marked reduction in coupling is observed within ten to 15 minutes of continuous perfusion at 100 to 200 mM. These exposure times do not appear to cause changes in the packing of connexin channels within gap junction plaques [45]. In contrast, studies evaluating connexin expression and phosphorylation typically expose cells to the drug for extended periods (>30 minutes) [44,45]. Thus, it is unclear whether the

8 Pharmacology

211

reduction in phosphorylation state of connexin subunits is required for the uncoupling effect of carbenoxolone. It is worth noting that connexin channels that lack classical phosphorylation sites (e.g., Cx26, or mutants of other connexins that lack the carboxyl-terminal domain) are sensitive to carbenoxolone. Carbenoxolone and its derivatives have nonspecific effects on a number of ion channels and cellular processes, although conflicting reports exist in the literature. Some studies indicate that carbenoxolone has no effect on evoked synaptic responses, intrinsic neuronal properties, or neuronal excitability, suggesting that the drug is largely without effects on other ion channels [46,47]. However, other studies report significant effects of carbenoxolone on intrinsic membrane properties and on antidromic responses of hippocampal CA3 pyramidal cells, indicating blockade of other ion channels [48,49]. In cultured neuronal cells, carbenoxolone was shown to alter intrinsic neuronal properties and decrease the input resistance and firing rate in response to depolarizing stimuli [50]. Carbenoxolone also blocks Ca2+ currents at concentrations similar to or lower than those that block connexin channels [51].

8.2.2 Polyamines and Other Ions Gap junctions formed by some connexins (e.g., Cx40) are blocked by tetraalkylammonium ions (TAA+) and polyamines such as spermine and spermidine when applied via patch pipettes [16,17]. Larger tetraalkylammonium cations, such as tetrabutylammonium and tetrapentylammonium, reduced junctional currents in cells expressing Cx40 or Cx43 [16]. Inhibition was voltage-dependent in a manner consistent with the movement of the blocking ions into the pore [16]. However, reduction of Cx40 channel activity was incomplete even at high voltages for the highest concentration tested (10 mM). Their utility is also limited by poor permeability to the binding site. In addition, large TAA+ ions block Ca2+ release from sarcoplasmic reticulum, and Cl– and K+ channels expressed in both heart and brain, and therefore cannot be used to block connexin channels without eliciting nonspecific effects [52,53]. Similarly, the effects of polyamines, which specifically block Cx40 but not Cx43 gap junctions, are voltage-dependent, but their low affinity (Kd 5 to 25 mM near 0 mV) and their effects on a number of other ion channels, including inward rectifier K+ channels, limit their utility as specific gap junction inhibitors [17,54]. Despite these limitations, these agents may be useful for the structure-function study of specific connexin subtypes.

8.2.3 Antimalarial Drugs It has been known for some time that quinine activates hemichannels in fish retinal horizontal cells and potentiates hemichannel currents in oocytes

212

M. Srinivas

expressing certain connexins (e.g., Cx35) but not others (e.g., Cx43) [55,56]. Surprisingly, quinine and derivatives have an inhibitory effect on gap junctions formed by Cx36 and by Cx50 (the concentration that inhibits 50% [IC50] values are 32 and 73 mM, respectively; [6]). The inhibition by quinine is connexinselective. Gap junctions formed by several other connexins are markedly less sensitive to quinine, in that 300 mM quinine produced little or no inhibition of channels formed by Cx32, Cx43, Cx26, or Cx40 [6]. Subsequent studies evaluating other quinine derivatives revealed that mefloquine, a widely prescribed antimalarial agent, is even more potent in reducing junctional conductance formed by Cx36 and Cx50 gap junctions [6]. Mefloquine inhibits Cx36 and Cx50 gap junction channels with IC50 values of 0.3 and 1.1 mM, respectively, and exhibits tenfold to 100-fold selectivity for these channels over Cx43, Cx32, Cx46, and Cx26 [8]. Quinine and derivatives are weak bases with pKa 8 to 9, and thus at physiological pH exist in both the uncharged free amine form and the cationic protonated form. The location of the binding site of quinine was investigated by determining the effect of quinine at a range of intracellular and extracellular pH values [6]. These results indicated that uncharged drug crosses the membrane and causes blockade by binding to the receptor in its protonated form. Additional evidence for an intracellular binding site was provided by experiments with a quaternary derivative of quinine, benzylquininium, which produced significant blockade of the Cx50 channels when applied internally but not when applied externally [6]. The ability of Cx50 to form hemichannels in Xenopus oocytes was exploited to further study the mechanism of inhibition [57]. Junctional channels behave as two hemichannels in series, making studies on hemichannels directly relevant to junctional channels. Initial experiments indicated that mefloquine inhibits Cx50 hemichannels but not Cx46, Cx26, or Cx32 hemichannels, selectivity similar to that observed for junctional channels, indicating the binding site is intrinsic to the hemichannel [57,58]. In addition, mefloquine, quinine, and its quaternary derivative benzylquininium, rapidly inhibited Cx50 hemichannels when applied to excised patches containing a single hemichannel, indicating a direct action on the channel [58]. These results indicate that quinine and derivatives reduce connexin channel activity most likely by binding to an intracellular site, possibly within the pore. Quinine blocks many channels and transporters and is therefore unlikely to be useful for long-term studies. However, the rapid onset and reversibility of inhibition by quinine has proven useful in acute studies. For example, quinine was used to assess the contribution of Cx50 to total coupling in lens epithelial cells, where it coexpressed with Cx43 [59]. Similarly, mefloquine was used to study the contribution of Cx50 to coupling in lens fiber cells [60]. The utility of mefloquine for assessing the role of Cx36 in the brain is limited by its poor permeability through tissue slices. The specificity of the drug for connexin channels is also limited [8]. Although results indicated that concentrations of mefloquine that reduce Cx36 coupling between neocortical interneurons

8 Pharmacology

213

exerted only moderate effects on chemical synapses or intrinsic cell properties, mefloquine caused increases in spontaneous synaptic activity and affected spiking during long high-frequency trains. Mefloquine also blocks P2X7 receptors with high-affinity [40], voltage-dependent L-type Ca2+ channels, Kir6.2 and KvLQT1 K+ channels at concentrations of 3 to 15 mM [61,62,63,64], which are only threefold to tenfold higher than those required to reduce coupling. Because of these effects, use of mefloquine to study roles of Cx36 and Cx50 should be accompanied by proper controls, for example, transgenic mice lacking the connexin of interest. Structure-activity studies of mefloquine are in progress and are expected to lead to the development of more potent and specific Cx36 and Cx50 inhibitors.

8.2.4 Fenamates Fenamates are nonsteroidal antiinflammatory drugs that inhibit cyclooxygenase when applied at nanomolar concentrations. They are highly lipophilic molecules that were first systematically studied as blockers of Cl- currents [65]. Fenamates were shown to reduce capacitative transients in monolayer cultures of a renal cell line (NRK cells), indicating inhibition of gap junctions. Uncoupling by flufenamic acid and derivatives was rapid, reaching steady-state within 30 seconds and quickly reversible [18,19]. Similarly, fenamates also rapidly and reversibly reduce connexin hemichannel currents [15]. The IC50 values for inhibition of connexin channels by fenamates ranges from 40 to 300 mM, depending on the derivative. The Hill coefficient for inhibition is 3.0, suggesting that binding of more than one molecule of the drug is required. Additional studies indicated that inhibition of gap junction channels by flufenamic acid is only marginally connexin-selective [19]. The site of action of fenamates is not known. Because fenamates are weak acids (pKa 6), it was expected that modifying extracellular and intracellular pH would affect the potency of blockade [19]. The efficacy of inhibition was indeed markedly enhanced at low external pH and decreased at high external pH, indicating that these drugs access their binding site via their uncharged membrane-permeable forms. However, modifying internal pH did not have a dramatic effect on the potency of these drugs. Additional experiments indicate that the effects of fenamates are not caused by binding within the pore, but rather by binding to a modulatory site, which leads to channel closure [19]. However, the precise location of the binding site remains ambiguous because the uncharged active form could act within either hydrophobic or hydrophilic connexin domains. Fenamates inhibit currents through a wide variety of channels, including nonselective cation channels, voltage-dependent K+ channels, L-type Ca2+ channels, and Cl– channels (e.g., the cystic fibrosis transmembrane conductance regulator) at concentrations ranging from 1 to 100 mM [66–69]. In addition,

214

M. Srinivas

these drugs potentiate the large-conductance Ca2+-activated K+ channel current [70,71]. Effects of fenamates on ligand-gated ion channels have also been reported [72]. Therefore, the usefulness of fenamates and derivatives as specific gap junction channel blockers is limited.

8.2.5 2-Aminophenoxyborate and Derivatives 2-Aminophenoxyborate (2-APB) is a well-known modulator of inositol 1,4,5triphosphate receptors [73]. It reversibly inhibits capacitative current transients in normal rat kidney and human embryonic kidney HEK293 cells with IC50 values in the low mM range [7]. 2-APB exhibits a strong selectivity for Cx36, Cx50, and Cx40 with surprisingly low IC50 values (1 to 3 mM), making it one of the most potent uncoupling agents identified to date [9]. Inhibition of other connexins require tenfold to 50-fold higher concentrations [9]. The actions of 2-APB are mediated by a direct binding of the drug to the connexin channel [10]. Transport-specific fractionation (TSF) assays, which provide an unambiguous demonstration of direct binding of a drug to connexin channels, indicate that the protonated form of 2-APB (100 mM) reduces the permeability of Cx32containing hemichannels reconstituted into liposomes [10]. Additional TSF studies by Tao and Harris [10] further indicate that an unmodified carboxylterminal domain (CT) is necessary for the effects of 2-APB on homomeric Cx26 channels, although direct proof requires exchange of domains between 2-APB– sensitive and 2-APB–insensitive connexin isoforms. The high potency and reversibility of 2-APB inhibition of some connexin channels make it a potentially useful drug for the delineation of their physiological roles. However, 2-APB is a notoriously nonspecific agent. Although concentrations of 2-APB that inhibit Cx36 or Cx50 coupling are three- to tenfold lower than those required to reduce inositol triphosphate (IP3)-induced calcium release and currents through store-operated Ca2+ channels [73,74], the drug affects many other membrane channels. Low concentrations of 2-APB (5 mM) activate Ca2+-release–activated Ca2+ currents and increase the Ca2+dependent inactivation of these channels [75]. At concentrations greater than 15 mM, 2-APB affects many other channels and transporters, including various transient receptor potential (TRP) channels [76,77]. Regardless, there is reason to believe that structure-activity studies of 2-APB may lead to the identification of connexin-specific drugs. Preliminary studies are promising in that they indicate that the structural determinants of block of Ca2+ release and of gap junction channels are different [10].

8.2.6 Lipophiles A large number of hydrophobic compounds inhibit currents through connexin channels. These include long-chain alkanols (e.g., octanol and

8 Pharmacology

215

heptanol), general anesthetics, and fatty acids and their amides, such as oleamide. The mechanism by which they uncouple cells appears to require partitioning into lipid bilayers and subsequent alteration of membrane fluidity or properties of lipids surrounding connexin channels [23,78,79,80] (for thorough review see [13]). Long-chain alkanols have long been known to reduce junctional coupling [22,24]. Both octanol and heptanol (1 to 2 mM) produce a rapid and reversible decrease in junctional conductance, and this has made them very useful for evaluation of effects of acute gap junction blockade. The potency of n-alkanols is inversely related to chain length. Thus, hexanol is ineffective at concentrations at which heptanol/octanol causes a complete reduction in junctional conductance and is therefore used as a control. There is no evidence for strong selectivity for block of gap junction channels formed by different connexins in mammalian cells, although it was reported that octanol selectively inhibits Cx50 but not Cx46 hemichannels in Xenopus oocytes [15]. Similarly, Cx56 (the avian ortholog of Cx46) is insensitive to heptanol [81]. Fluorescence anisotropy and electrophysiological experiments suggest that n-alkanols decrease junctional conductance by specifically decreasing the fluidity of cholesterol-rich domains that surround gap junction plaques. Increasing the bulk membrane fluidity of noncholesterol regions had no effect on junctional conductance [78]. Thus, as hypothesized [13], actions of n-alkanols may depend on the cholesterol content of the membranes and on the localization of gap junctions in such membranes [13]. This may explain why low concentrations of n-alkanols are protective against cell death induced by ischemia in brain [82] and heart [83,84], an effect that has been directly attributed to gap junction blockade. Similarly, the finding that low concentrations of heptanol (200 mM) selectively reduces gap junctional coupling without effects on nonjunctional processes in arteries [85] may be due to the localization of gap junctions in cholesterol-rich domains in this tissue [13]. Nevertheless, in addition to its effects on gap junctions, heptanol reduces other nonjunctional membrane ionic currents, including Na+ and Ca2+ inward currents and affects excitatory synaptic transmission [86,87,88]. Inhibition of these ion channels and a number of other proteins by n-alkanols occurs at concentrations typically used to reduce connexin channel currents, and therefore they should be used with caution when ascribing specific physiological processes to gap junctions. General anesthetics such as halothane and isoflurane rapidly and reversibly uncouple cells expressing gap junctions at concentrations used for inhalation anesthesia [20]. The rapidity and the reversibility of the effect have both been useful for study of the properties of single gap junction channels. The mechanism and site of action on the connexin molecules is unknown, although the high lipid solubility suggests that the mechanism involves effects at the hydrophobic lipid–protein interface, for example, disruption of acyl chain packing surrounding connexin channels [13]. Only a few studies have examined the connexinspecificity of halothane; however, the differences thus far demonstrated have

216

M. Srinivas

been quite modest [89]. For example, the difference in the IC50 values for closure of Cx40 and Cx43 channels by halothane was less than twofold [89]. Halothane also affects several other cellular processes (see [21]), so its use for assessing the physiological roles of gap junctions is limited. Oleamide (cis-9-octadecenamide), the primary amide of oleic acid, was first isolated from sleep-deprived cats and, subsequently, shown to induce physiological sleep when injected in animals [90]. Oleamide completely and reversibly reduces electrical and dye-coupling (IC50 20 mM) in cultured glial cells, which predominantly express Cx43 [25,26]. Reduction of coupling occurs within 1 to 2 minutes of perfusion of the drug [25]. Similarly, oleamide reduces dye-coupling in Cx32-expressing cells with similar IC50 values [25]. This result indicates that effects of oleamide are unlikely to be dependent on connexin isoform. The effects on gap junctions are stereoisomer-specific; trans-9-octadecenamide has no effect on Cx43-mediated coupling in glial cells [25]. Moreover, oleic acid, a hydrophobic congener of oleamide, has no effect on glial gap junctions [25]. Surprisingly, oleic acid (1 to 10 mM) inhibits coupling in cardiac myocytes, which also express Cx43 [91]. It is unclear why the uncoupling effect is cell-dependent, but as with n-alkanols, the variability may be related to the cholesterol content of plasma membranes of different tissues. Similar to nalkanols, oleamide has been proposed to induce closure of gap junctions by disrupting the fluidity of cholesterol-rich microdomains in which junctional channels reside [92]. The structure of oleamide — a long alkane chain with a centrally placed cis double bond and a primary amide that can form hydrogen bonds — also allows it to interact with membrane lipids [26]. Oleamide and oleic acid also block a number of ion channels and membrane receptors at concentrations well below those required to reduce junctional coupling [93,94,95,96]. Oleamide interacts with serotonin receptors with very high-affinity and blocks ligand-gated g-aminobutyric acid (GABA) receptor channels and Naþ channels at concentrations below 10 mM, reducing its utility as a specific gap junction blocking agent [93,94,95,96].

8.2.7 Cyclodextrins Cyclodextrins (CDs) are naturally occurring cyclic oligosaccharides of a-Dglucopyranose, with a relatively hydrophobic central cavity and a hydrophilic outer surface. This unique structure makes them capable of forming complexes with a wide variety of hydrophobic molecules and they have attracted considerable attention from pharmaceutical industries as drug delivery agents. The TSF studies indicated that aCD, bCD, and gCD, which consist of six, seven, or eight glucopyranose units linked a1!4, respectively, block the permeability of sucrose and urea through Cx32 hemichannels [5]. The two smaller CDs (aCD and bCD) showed complete block of homomeric Cx32 channels at concentrations of 6 and 15 mM, respectively, in a manner consistent with simple pore

8 Pharmacology

217

block [5]. Block by gCD was more complex, probably as a result of its larger size relative to the dimension of channel pore. Additional studies further indicated that the access of aCD or bCD to their binding site occurs from the cytoplasmic side, suggesting that the narrowest part of the pore is located extracellular to the site of block [5]. A simple pore block was also consistent with the effects of the three cyclodextrins on Cx26 homomeric and Cx26/Cx32 heteromeric channels, which have narrower pores compared to homomeric Cx32 channels. Block of sucrose and urea permeability occurred only with the smallest of the three cyclodextrins, aCD, whereas block by bCD appeared to be partial such that the smaller molecule, urea, but not sucrose, still permeated through the blocked channels [5] (see Chapter 7). These results are promising for they allow for a certain degree of connexinselectivity depending on the diameters of the pore of individual connexins. In addition, the fact that they bind in the permeation pathway makes them very useful for mapping the pore-lining domains of connexin channels, especially near the cytoplasmic end of the channel. It remains to be determined whether cyclodextrins block Cx26 or Cx32 channels in cells expressing these connexins in a similar fashion, although there is little reason to believe otherwise. Other drugs, such as 2-APB, which blocks permeability in TSF assays, remain just as effective in electrophysiological studies. Nevertheless, the recent demonstration that Cx26 forms functional hemichannels in Xenopus oocytes will help to further characterize the effects of cyclodextrins [97].

8.2.8 Connexin-Mimetic Peptides Extracellular domains of connexins are important for the docking and subsequent formation of gap junction channels. Therefore, it was reasoned that reagents that target these domains might reduce intercellular communication. Initial studies focused on antibodies directed against extracellular domains of connexins [98–100]. However, the high antibody concentrations and long exposure times required for efficacious blockade meant that they were of limited utility as specific blockers of junctional channels. An alternative approach has been the use of small peptides that correspond to specific sequences within extracellular loops E1 and E2 [27,101,102]. Such peptides targeting extracellular loop domains of Cx32 were shown to effectively reduce the formation of gap junction channels [27,101,102]. Subsequent studies indicated that peptides corresponding to extracellular sequences in Cx43, involving the conserved QPG and SHVR motifs of E1 (Gap26 peptide) and the SRPTEK motif in E2 (Gap27 peptide) of connexins, are most effective in reducing junctional communication [28,29,103,104,105]. The potential utility of this approach is highlighted by the ability to inhibit specific connexin subtypes simply by designing peptides unique to individual connexin isoforms. Such peptides corresponding to extracellular loops of Cx43, Cx37, and Cx40 were shown to act in a connexin-selective

218

M. Srinivas

fashion, and have been used to inhibit gap junctions in smooth muscle where these connexins are expressed [106,107,108,109]. There are several drawbacks associated with the use of these reagents. Inhibition of junctional communication by connexin-mimetic peptides typically requires incubation of cells for long durations (1 to 2 hours) and with high concentrations (typically 300 to 600 mM). Even at these concentrations, connexin-mimetic peptides do not reduce coupling completely [28,104,110]. In addition to inhibiting gap junctions, connexin-mimetic peptides also reduce dye uptake presumably mediated by connexin hemichannels. The effects of peptides on hemichannel permeability are relatively rapid (10 to 20 minutes), compared to their effects on gap junctions [29,111,112]. This result is expected; extracellular loop domains are likely to be freely accessible in the undocked hemichannel. Subsequent studies indicated that a peptide targeting the cytoplasmic loop (CL) of connexin was even more effective at inhibiting hemichannel-mediated dye uptake, suggesting to the authors that the peptide enters the cell [113]. Surprisingly, the CL peptide had no effect on gap junctions in the same preparation [113]. These results are difficult to explain, and lead to questions regarding the mechanism and specificity of action of these peptides. Disruption of coupling may be due to reduced docking of hemichannels or an ‘‘unzipping’’ of existing gap junction channels [28,29,104]. But in a recent study, connexin-mimetic peptides were shown to have no effect on number or size of Cx43-EGFP (enhanced green fluorescent protein) gap junction plaques, leading to speculation that they uncouple cells by inducing a conformational change, that is, by affecting gating [107]. It must be noted that only a fraction of channels in a plaque are actively open at any given time [114], and small changes in plaque size that may be sufficient to uncouple cells may not be easily assessed by immunocytochemical methods. A change in gating can be studied by determining effects on hemichannel currents in Xenopus oocytes, but to date there have been no such studies. In addition, there are no studies documenting whether they directly bind to extracellular loops. Quantitative estimates of the binding affinity of different connexin-mimetic peptides to loop regions are also unavailable. The specificity of these reagents also remains questionable. In rat mesenteric arteries, connexin mimetic peptides caused a reduction in electrotonic coupling without exerting major nonjunctional effects [115]. However, a recent report indicates that these peptides strongly reduce membrane currents in Xenopus oocytes expressing pannexin1 at concentrations similar to those that inhibit connexin channels [116] (see Chapter 12).

8.3 Conclusion Although a number of inhibitors of connexin channels have been identified in recent years using conventional and novel assays of channel function, the potency and specificity of these reagents are not optimal for long-term studies.

8 Pharmacology

219

None of the agents that have been identified inhibits connexin channels with nanomolar affinity. More troublesome is the fact that no identified compound inhibits connexin channels in a highly specific fashion. Nevertheless, application of classic medicinal chemistry principles for existing reagents is essential for improving their specificity and potency. It is also clear that new approaches are essential for the identification of high-affinity reagents for connexin channels. High-throughput screening assays, which have greatly facilitated the discovery of new lead compounds for drug development for other ion channels, must be developed for connexin channels. Similarly, determination of high-resolution structures of connexin channels and in silico drug binding sites is also of paramount importance for the identification of new compounds based on principles of rational drug design. Acknowledgments Work in the author’s laboratory related to this topic was supported by National Eye Institute (NEI) grant EY13869.

References 1. Salameh A, Dhein S. Pharmacology of gap junctions. New pharmacological targets for treatment of arrhythmia, seizure and cancer? Biochim Biophys Acta. 2005;1719:36–58. 2. Herve JC. Gap junction channels: from protein genes to diseases. Prog Biophys Mol Biol. 2007;94:1–4. 3. Martin FC, Handforth A. Carbenoxolone and mefloquine suppress tremor in the harmaline mouse model of essential tremor. Mov Disord. 2006;21:1641–9. 4. Bevans CG, Harris AL. Regulation of connexin channels by pH. Direct action of the protonated form of taurine and other aminosulfonates. J Biol Chem. 1999;274: 3711–9. 5. Locke D, Koreen IV, Liu JY, Harris AL. Reversible pore block of connexin channels by cyclodextrins. J Biol Chem. 2004;279:22883–92. 6. Srinivas M, Hopperstad MG, Spray DC. Quinine blocks specific gap junction channel subtypes. Proc Natl Acad Sci USA. 2001;98:10942–7. 7. Harks EG, Camina JP, Peters PH, Ypey DL, Scheenen WJ, van Zoelen EJ, Theuvenet A. Besides affecting intracellular calcium signaling, 2-APB reversibly blocks gap junctional coupling in confluent monolayers, thereby allowing measurement of single-cell membrane currents in undissociated cells. FASEB J. 2003;17:941–3. 8. Cruikshank SJ, Hopperstad M, Younger M, Connors BW, Spray DC, Srinivas M. Potent block of Cx36 and Cx50 gap junction channels by mefloquine. Proc Natl Acad Sci USA. 2004;101:12364–9. 9. Bai D, del Corsso C, Srinivas M, Spray DC. Block of specific gap junction channel subtypes by 2-aminoethoxydiphenyl borate (2-APB). J Pharmacol Exp Ther. 2006;319:1452–8. 10. Tao L, Harris AL. 2-aminoethoxydiphenyl borate directly inhibits channels composed of connexin26 and/or connexin32. Mol Pharmacol. 2007;71:570–9. 11. Shibayama J, Lewandowski R, Kieken F, Coombs W, Shah S, Sorgen PL, Taffet SM, Delmar M. Identification of a novel peptide that interferes with the chemical regulation of connexin43. Circ Res. 2006;98:1365–72. 12. Spray DC, Rozental R, Srinivas M. Prospects for rational development of pharmacological gap junction channel blockers. Curr Drug Targets. 2002;3:455–64. 13. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys. 2001;34:325–472.

220

M. Srinivas

14. Bruzzone R, Barbe MT, Jakob NJ, Monyer H. Pharmacological properties of homomeric and heteromeric pannexin hemichannels expressed in Xenopus oocytes. J Neurochem. 2005;92:1033–43. 15. Eskandari S, Zampighi GA, Leung DW, Wright EM, Loo DD. Inhibition of gap junction hemichannels by chloride channel blockers. J Membr Biol. 2002;185:93–102. 16. Musa H, Gough JD, Lees WJ, Veenstra RD. Ionic blockade of the rat connexin40 gap junction channel by large tetraalkylammonium ions. Biophys J. 2001;81:3253–74. 17. Musa H, Veenstra RD. Voltage-dependent blockade of connexin40 gap junctions by spermine. Biophys J. 2003;84:205–19. 18. Harks EG, de Roos AD, Peters PH, de Haan LH, Brouwer A, Ypey DL, van Zoelen EJ, Theuvenet AP. Fenamates: a novel class of reversible gap junction blockers. J Pharmacol Exp Ther. 2001;298:1033–41. 19. Srinivas M, Spray DC. Closure of gap junction channels by arylaminobenzoates. Mol Pharmacol. 2003;63:1389–97. 20. Burt JM, Spray DC. Volatile anesthetics block intercellular communication between neonatal rat myocardial cells. Circ Res. 1989;65:829–37. 21. Rozental R, Srinivas M, Spray DC. How to close a gap junction channel. Efficacies and potencies of uncoupling agents. Methods Mol Biol. 2001;154:447–76. 22. Johnston MF, Simon SA, Ramon F. Interaction of anaesthetics with electrical synapses. Nature. 1980;286:498–500. 23. Spray DC, Burt JM. Structure-activity relations of the cardiac gap junction channel. Am J Physiol. 1990;258:C195–205. 24. Deleze J, Herve JC. Effect of several uncouplers of cell-to-cell communication on gap junction morphology in mammalian heart. J Membr Biol. 1983;74:203–15. 25. Guan X, Cravatt BF, Ehring GR, Hall JE, Boger DL, Lerner RA, Gilula NB. The sleepinducing lipid oleamide deconvolutes gap junction communication and calcium wave transmission in glial cells. J Cell Biol. 1997;139:1785–92. 26. Boger DL, Patterson JE, Guan X, Cravatt BF, Lerner RA, Gilula NB. Chemical requirements for inhibition of gap junction communication by the biologically active lipid oleamide. Proc Natl Acad Sci USA. 1998;95:4810–5. 27. Warner A, Clements DK, Parikh S, Evans WH, DeHaan RL. Specific motifs in the external loops of connexin proteins can determine gap junction formation between chick heart myocytes. J Physiol. 1995;488:721–8. 28. Chaytor AT, Evans WH, Griffith TM. Peptides homologous to extracellular loop motifs of connexin 43 reversibly abolish rhythmic contractile activity in rabbit arteries. J Physiol. 1997;503:99–110. 29. Evans WH, De Vuyst E, Leybaert L. The gap junction cellular internet: connexin hemichannels enter the signaling limelight. Biochem J. 2006;397:1–14. 30. DeCoursey TE. Mechanism of K+ channel block by verapamil and related compounds in rat alveolar epithelial cells. J Gen Physiol. 1995;106:745–79. 31. Horrigan FT, Gilly WF. Methadone block of K+ current in squid giant fiber lobe neurons. J Gen Physiol. 1996;107:243–60. 32. Burt JM, Spray DC. Single-channel events and gating behavior of the cardiac gap junction channel. Proc Natl Acad Sci USA. 1988;85:3431–4. 33. Weingart R, Bukauskas FF. Long-chain n-alkanols and arachidonic acid interfere with the Vm-sensitive gating mechanism of gap junction channels. Pflu¨gers Arch. 1998;435:310–9. 34. Bukauskas FF, Peracchia C. Two distinct gating mechanisms in gap junction channels: CO2-sensitive and voltage-sensitive. Biophys J. 1997;72:2137–42. 35. Bukauskas FF, Weingart R. Voltage-dependent gating of single gap junction channels in an insect cell line. Biophys J. 1994;67:613–25. 36. Trexler EB, Bennett MVL, Bargiello TA, Verselis VK. Voltage-gating and permeation in a gap junction hemichannel. Proc Natl Acad Sci USA. 1996;93:5836–41. 37. Davidson JS, Baumgarten IM. Glycyrrhetinic acid derivatives: a novel class of inhibitors of gap-junctional intercellular communication. Structure-activity relationships. J Pharmacol Exp Ther. 1988;246:1104–7.

8 Pharmacology

221

38. Davidson JS, Baumgarten IM, Harley EH. Reversible inhibition of intercellular junctional communication by glycyrrhetinic acid. Biochem Biophys Res Commun. 1986;134:29–36. 39. Locovei S, Scemes E, Qiu F, Spray DC, Dahl G. Pannexin1 is part of the pore forming unit of the P2X(7) receptor death complex. FEBS Lett. 2007;581:483–8. 40. Suadicani SO, Brosnan CF, Scemes E. P2X7 receptors mediate ATP release and amplification of astrocytic intercellular Ca2+ signaling. J Neurosci. 2006;26:1378–85. 41. Monder C, Stewart PM, Lakshmi V, Valentino R, Burt D, Edwards CR. Licorice inhibits corticosteroid 11 -dehydrogenase of rat kidney and liver: in vivo and in vitro studies. Endocrinology. 1989;125:1046–53. 42. Martin W, Zempel G, Hu¨lser D, Willecke K. Growth inhibition of oncogene-transformed rat fibroblasts by cocultured normal cells: relevance of metabolic cooperation mediated by gap junctions. Cancer Res. 1991;51:5348–51. 43. Goldberg GS, Moreno AP, Bechberger JF, Hearn SS, Shivers RR, MacPhee DJ, Zhang YC, Naus CC. Evidence that disruption of connexon particle arrangements in gap junction plaques is associated with inhibition of gap junctional communication by a glycyrrhetinic acid derivative. Exp Cell Res 1996;222:48–53. 44. Guan X, Wilson S, Schlender KK, Ruch RJ. Gap-junction disassembly and connexin 43 dephosphorylation induced by 18 -glycyrrhetinic acid. Mol Carcinog. 1996;16:157–64. 45. Guo Y, Martı´ nez-Williams C, Gilbert KA, Rannels DE. Inhibition of gap junction communication in alveolar epithelial cells by 18 -glycyrrhetinic acid. Am J Physiol. 1999;276:L1018–26. 46. Yang Q, Michelson HB. Gap junctions synchronize the firing of inhibitory interneurons in guinea pig hippocampus. Brain Res. 2001;907:139–43. 47. Kohling R, Gladwell SJ, Bracci E, Vreugdenhil M, Jefferys JG. Prolonged epileptiform bursting induced by 0-Mg(2+) in rat hippocampal slices depends on gap junctional coupling. Neuroscience. 2001;105:579–87. 48. Gladwell SJ, Jefferys JG. Second messenger modulation of electrotonic coupling between region CA3 pyramidal cell axons in the rat hippocampus. Neurosci Lett. 2001;300:1–4. 49. Rekling JC, Shao XM, Feldman JL. Electrical coupling and excitatory synaptic transmission between rhythmogenic respiratory neurons in the preBotzinger complex. J Neurosci. 2000;20:RC113:1–5. 50. Rouach N, Segal M, Koulakoff A, Giaume C, Avignone E. Carbenoxolone blockade of neuronal network activity in culture is not mediated by an action on gap junctions. J Physiol. 2003;553:729–45. 51. Vessey JP, Lalonde MR, Mizan HA, Welch NC, Kelly ME, Barnes S. Carbenoxolone inhibition of voltage-dependent Ca channels and synaptic transmission in the retina. J Neurophysiol. 2004;92:1252–6. 52. Sanchez DY, Blatz AL. Block of neuronal chloride channels by tetraethylammonium ion derivatives. J Gen Physiol. 1995;106:1031–46. 53. French RJ, Shoukimas JJ. Blockage of squid axon potassium conductance by internal tetra-N-alkylammonium ions of various sizes. Biophys J. 1981;34:271–91. 54. Williams K. Interactions of polyamines with ion channels. Biochem J. 1997;325:289–97. 55. Malchow RP, Qian H, Ripps H. A novel action of quinine and quinidine on the membrane conductance of neurons from the vertebrate retina. J Gen Physiol. 1994;104: 1039–55. 56. White TW, Deans MR, O’Brien J, Al-Ubaidi MR, Goodenough DA, Ripps H, Bruzzone R. Functional characteristics of skate connexin35, a member of the g subfamily of connexins expressed in the vertebrate retina. Eur J Neurosci. 1999;11:1883–90. 57. Srinivas M, Kronengold J, Bukauskas FF, Bargiello TA, Verselis VK. Correlative studies of gating in Cx46 and Cx50 hemichannels and gap junction channels. Biophys J. 2005;88:1725–39. 58. Trelles M, Srinivas M. Molecular basis for the block of connexin channels by quinine and its derivatives. Biophysical Society Meeting 2007;442a.

222

M. Srinivas

59. White TW, Gao Y, Leping L, Sellitto C, Srinivas M. Optimal lens proliferation is dependent on the connexin isoform providing gap junctional coupling. Invest Ophthalmol Vis Sci. 2007;48:5630–7. 60. Martı´ nez-Wittinghan FJ, Srinivas M, Sellitto C, White TW, Mathias RT. Mefloquine effects on the lens suggest cooperative gating of gap junction channels. J Membr Biol. 2006;211:163–71. 61. Kang J, Chen XL, Wang L, Rampe D. Interactions of the antimalarial drug mefloquine with the human cardiac potassium channels KvLQT1/minK and HERG. J Pharmacol Exp Ther. 2001;299:290–6. 62. Gribble FM, Davis TM, Higham CE, Clark A, Ashcroft FM. The antimalarial agent mefloquine inhibits ATP-sensitive K-channels. Br J Pharmacol. 2000;131:756–60. 63. Traebert M, Dumotier B, Meister L, Hoffmann P, Dominguez-Estevez M, Suter W. Inhibition of hERG K+ currents by antimalarial drugs in stably transfected HEK293 cells. Eur J Pharmacol. 2004;484:41–8. 64. Maertens C, Wei L, Droogmans G, Nilius B. Inhibition of volume-regulated and calciumactivated chloride channels by the antimalarial mefloquine. J Pharmacol Exp Ther. 2000;295:29–36. 65. Wangemann P, Wittner M, Di Stefano A, Englert HC, Lang HJ, Schlatter E, Greger R. Cl-channel blockers in the thick ascending limb of the loop of Henle. Structure activity relationship. Pflu¨gers Arch. 1986;407:S128–41. 66. Doughty JM, Miller AL, Langton PD. Nonspecificity of chloride channel blockers in rat cerebral arteries: block of the L-type calcium channel. J Physiol. 1998;507:433–9. 67. McCarty NA, McDonough S, Cohen BN, Riordan JR, Davidson N, Lester HA. Voltagedependent block of the cystic fibrosis transmembrane conductance regulator Cl- channel by two closely related arylaminobenzoates. J Gen Physiol. 1993;102:1–23. 68. Gogelein H, Dahlem D, Englert HC, Lang HJ. Flufenamic acid, mefenamic acid and niflumic acid inhibit single nonselective cation channels in the rat exocrine pancreas. FEBS Lett. 1990;268: 79–82. 69. Farrugia G, Rae JL, Szurszewski JH. Characterization of an outward potassium current in canine jejunal circular smooth muscle and its activation by fenamates. J Physiol. 1993;468:297–310. 70. Ottolia M, Toro L. Potentiation of large conductance KCa channels by niflumic, flufenamic, and mefenamic acids. Biophys J. 1994;67:2272–9. 71. Greenwood IA, Large WA. Comparison of the effects of fenamates on Ca-activated chloride and potassium currents in rabbit portal vein smooth muscle cells. Br J Pharmacol. 1995;116:2939–48. 72. Woodward RM, Polenzani L, Miledi R. Effects of fenamates and other nonsteroidal antiinflammatory drugs on rat brain GABAA receptors expressed in Xenopus oocytes. J Pharmacol Exp Ther. 1994;268:806–17. 73. Maruyama T, Kanaji T, Nakade S, Kanno T, Mikoshiba K. 2APB, 2-aminoethoxydiphenyl borate, a membrane-penetrable modulator of Ins(1,4,5)P3-induced Ca2+ release. J Biochem. 1997;122:498–505. 74. Dobrydneva Y, Blackmore P. 2-Aminoethoxydiphenyl borate directly inhibits store-operated calcium entry channels in human platelets. Mol Pharmacol. 2001;60:541–52. 75. Prakriya M, Lewis RS. Potentiation and inhibition of Ca(2+) release-activated Ca(2+) channels by 2-aminoethyldiphenyl borate (2-APB) occurs independently of IP(3) receptors. J Physiol. 2001;536:3–19. 76. Hu HZ, Gu Q, Wang C, Colton CK, Tang J, Kinoshita-Kawada M, Lee LY, Wood JD, Zhu MX. 2-aminoethoxydiphenyl borate is a common activator of TRPV1, TRPV2, and TRPV3. J Biol Chem. 2004;279:35741–8. 77. Ma HT, Venkatachalam K, Li HS, Montell C, Kurosaki T, Patterson RL, Gill DL. Assessment of the role of the inositol 1,4,5-trisphosphate receptor in the activation of transient receptor potential channels and store-operated Ca2+ entry channels. J Biol Chem. 2001;276:18888–96.

8 Pharmacology

223

78. Bastiaanse EM, Jongsma HJ, van der Laarse A, Takens-Kwak BR. Heptanol-induced decrease in cardiac gap junctional conductance is mediated by a decrease in the fluidity of membranous cholesterol-rich domains. J Membr Biol. 1993;136:135–45. 79. Burt JM. Uncoupling of cardiac cells by doxyl stearic acids specificity and mechanism of action. Am J Physiol. 1989;256:C913–24. 80. Takens-Kwak BR, Jongsma HJ, Rook MB, Van Ginneken AC. Mechanism of heptanolinduced uncoupling of cardiac gap junctions: a perforated patch-clamp study. Am J Physiol. 1992;262:C1531–8. 81. Rup DM, Veenstra RD, Wang HZ, Brink PR, Beyer EC. Chick connexin-56, a novel lens gap junction protein. Molecular cloning and functional expression. J Biol Chem. 1993;268:706–12. 82. Rawanduzy A, Hansen A, Hansen TW, Nedergaard M. Effective reduction of infarct volume by gap junction blockade in a rodent model of stroke. J Neurosurg. 1997;87:916–20. 83. Garcia-Dorado D, Inserte J, Ruiz-Meana M, Gonzalez MA, Solares J, Julia M, Barrabes JA, Soler-Soler J. Gap junction uncoupler heptanol prevents cell-to-cell progression of hypercontracture and limits necrosis during myocardial reperfusion. Circulation. 1997;96:3579–86. 84. Saltman AE, Aksehirli TO, Valiunas V, Gaudette GR, Matsuyama N, Brink P, Krukenkamp IB. Gap junction uncoupling protects the heart against ischemia. J Thorac Cardiovasc Surg. 2002;124:371–6. 85. Christ GJ, Spektor M, Brink PR, Barr L. Further evidence for the selective disruption of intercellular communication by heptanol. Am J Physiol. 1999;276:H1911–7. 86. Quastel DM, Saint DA. Modification of motor nerve terminal excitability by alkanols and volatile anaesthetics. Br J Pharmacol. 1986;88:747–56. 87. McLarnon JG, Quastel DM. Thermodynamic parameters of end-plate channel blockade. J Neurosci. 1984;4:939–44. 88. Hirche G. Blocking and modifying actions of octanol on Na channels in frog myelinated nerve. Pflu¨gers Arch. 1985;405:180–7. 89. He DS, Burt JM. Mechanism and selectivity of the effects of halothane on gap junction channel function. Circ Res. 2000;86:E104–9. 90. Cravatt BF, Prospero-Garcia O, Siuzdak G, Gilula NB, Henriksen SJ, Boger DL, Lerner RA. Chemical characterization of a family of brain lipids that induce sleep. Science. 1995;268:1506–9. 91. Hirschi KK, Minnich BN, Moore LK, Burt JM. Oleic acid differentially affects gap junction-mediated communication in heart and vascular smooth muscle cells. Am J Physiol. 1993;265:C1517–26. 92. Lerner RA. A hypothesis about the endogenous analogue of general anesthesia. Proc Natl Acad Sci USA. 1997;94:13375–7. 93. Boger DL, Patterson JE, Jin Q. Structural requirements for 5-HT2A and 5-HT1A serotonin receptor potentiation by the biologically active lipid oleamide. Proc Natl Acad Sci USA. 1998;95:4102–7. 94. Lees G, Edwards MD, Hassoni AA, Ganellin CR, Galanakis D. Modulation of GABA(A) receptors and inhibitory synaptic currents by the endogenous CNS sleep regulator cis-9,10-octadecenoamide (cOA). Br J Pharmacol. 1998;124:873–82. 95. Thomas EA, Carson MJ, Sutcliffe JG. Oleamide-induced modulation of 5-hydroxytryptamine receptor-mediated signaling. Ann NY Acad Sci. 1998;861:183–9. 96. Verdon B, Zheng J, Nicholson RA, Ganelli CR, Lees G. Stereoselective modulatory actions of oleamide on GABA(A) receptors and voltage-dependent Na+ channels in vitro: a putative endogenous ligand for depressant drug sites in CNS. Br J Pharmacol. 2000;129:283–90. 97. Gonzalez D, Gomez-Hernandez JM, Barrio LC. Species-specificity of mammalian connexin-26 to form open voltage-dependent hemichannels. FASEB J. 2006;20:2329–38. 98. Hertzberg EL, Spray DC, Bennett MVL. Reduction of gap junctional conductance by microinjection of antibodies against the 27-kDa liver gap junction polypeptide. Proc Natl Acad Sci USA. 1985;82:2412–6.

224

M. Srinivas

99. Hofer A, Dermietzel R. Visualization and functional blocking of gap junction hemichannels (connexons) with antibodies against external loop domains in astrocytes. Glia. 1998;24:141–54. 100. Meyer RA, Laird DW, Revel JP, Johnson RG. Inhibition of gap junction and adherens junction assembly by connexin and A-CAM antibodies. J Cell Biol. 1992;119:179–89. 101. Dahl G, Nonner W, Werner R. Attempts to define functional domains of gap junction proteins with synthetic peptides. Biophys J. 1994;67:1816–22. 102. Dahl G, Werner R, Levine E, Rabadan-Diehl C. Mutational analysis of gap junction formation. Biophys J. 1992;62:172–80. 103. Boitano S, Evans WH. Connexin mimetic peptides reversibly inhibit Ca(2+) signaling through gap junctions in airway cells. Am J Physiol Lung Cell Mol Physiol. 2000;279:L623–30. 104. Berthoud VM, Beyer EC, Seul KH. Peptide inhibitors of intercellular communication. Am J Physiol Lung Cell Mol Physiol. 2000;279:L619–22. 105. Isakson BE, Seedorf GJ, Lubman RL, Evans WH, Boitano S. Cell-cell communication in heterocellular cultures of alveolar epithelial cells. Am J Respir Cell Mol Biol. 2003;29:552–61. 106. Dora KA, Martin PE, Chaytor AT, Evans WH, Garland CJ, Griffith TM. Role of heterocellular Gap junctional communication in endothelium-dependent smooth muscle hyperpolarization: inhibition by a connexin-mimetic peptide. Biochem Biophys Res Commun. 1999;254:27–31. 107. Martin PE, Wall C, Griffith TM. Effects of connexin-mimetic peptides on gap junction functionality and connexin expression in cultured vascular cells. Br J Pharmacol. 2005;144:617–27. 108. Griffith TM, Chaytor AT, Edwards DH. The obligatory link: role of gap junctional communication in endothelium-dependent smooth muscle hyperpolarization. Pharmacol Res. 2004;49:551–64. 109. Isakson BE, Duling BR. Heterocellular contact at the myoendothelial junction influences gap junction organization. Circ Res. 2005;97:44–51. 110. Kwak BR, Jongsma HJ. Selective inhibition of gap junction channel activity by synthetic peptides. J Physiol. 1999;516:679–85. 111. Braet K, Aspeslagh S, Vandamme W, Willecke K, Martin PE, Evans WH, Leybaert L. Pharmacological sensitivity of ATP release triggered by photoliberation of inositol1,4,5-trisphosphate and zero extracellular calcium in brain endothelial cells. J Cell Physiol. 2003;197:205–13. 112. Braet K, Vandamme W, Martin PE, Evans WH, Leybaert L. Photoliberating inositol1,4,5-trisphosphate triggers ATP release that is blocked by the connexin mimetic peptide gap 26. Cell Calcium. 2003;33:37–48. 113. De Vuyst E, Decrock E, Cabooter L, Dubyak GR, Naus CC, Evans WH, Leybaert L. Intracellular calcium changes trigger connexin 32 hemichannel opening. EMBO J. 2006;25:34–44. 114. Bukauskas FF, Jordan K, Bukauskiene A, Bennett MVL, Lampe PD, Laird DW, Verselis VK. Clustering of connexin 43-enhanced green fluorescent protein gap junction channels and functional coupling in living cells. Proc Natl Acad Sci USA. 2000;97:2556–61. 115. Matchkov VV, Rahman A, Bakker LM, Griffith TM, Nilsson H, Aalkjaer C. Analysis of effects of connexin-mimetic peptides in rat mesenteric small arteries. Am J Physiol Heart Circ Physiol. 2006;291:H357–67. 116. Wang J, Ma M, Locovei S, Keane R, Dahl GP. Modulation of membrane channel currents by gap junction protein mimetic peptides: size matters. Am J Physiol Cell Physiol. 2007;293:C1112–9.

Chapter 9

Biogenesis and Degradation of Gap Junctions Linda S. Musil

Abstract The dynamic regulation of gap junction biogenesis and degradation is a key element in the control of intercellular communication. This regulation starts in the endoplasmic reticulum, in which a large fraction of newly synthesized connexin molecules can be degraded. Multisubunit assembly of endogenously expressed connexins is first detected in the trans-Golgi network, as opposed to the endoplasmic reticulum wherein most other integral membrane proteins oligomerize. Gap junction plaques grow, at least in part, by lateral diffusion of plasma membrane connexin hemichannels to the periphery of the plaques. Connexins have a half-life of a few hours, being turned over within the endoplasmic reticulum by the ubiquitin/proteasome system and after transport to the cell surface via the lysosome. Although some aspects of gap junction biogenesis and degradation are relatively well understood, several intriguing unanswered questions remain. Keywords ER-associated degradation Assembly Hemichannel Endoplasmic reticulum Golgi trans-Golgi network Degradation Proteasome Heat-shock Lysosome Ubiquitin Cx26 Cx32 Cx43 Cx46

9.1 Introduction Compared to other cell–cell junctions (e.g., tight junctions, desmosomes, adherens junctions), the composition of gap junctions is deceptively simple. Biochemical, morphological, and physiological analyses have failed to identify an obligatory requirement for proteins other than connexins in gap junctional plaques. Moreover, exogenous expression of connexins in a wide variety of model systems is sufficient to result in the formation of gap junctions that are functionally and L.S. Musil (*) Department of Biochemistry and Molecular Biology L224, Oregon Health & Science University, 3181 Southwest Sam Jackson Park Road, Portland, OR 97239, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_9, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

225

226

L.S. Musil

ultrastructurally indistinguishable from those in native tissues. It is therefore perhaps surprising that there are many unanswered questions and controversies about how gap junctions are assembled and degraded. This chapter discusses key issues in gap junction biosynthesis and turnover, with a particular emphasis on subjects most relevant to scientists outside of, or new to, the field of gap junction research. Topics reviewed include how newly synthesized connexins are inserted into the endoplasmic reticulum (ER), degradation of connexins at the ER, site of hemichannel assembly, transport of hemichannels to the cell surface, and how gap junction plaques are degraded. Knowledge of these processes and their regulation is essential for understanding how gap junctions carry out their myriad physiological roles and how mutations in connexin genes cause human disease.

9.2 Agreed-Upon Truths of Gap Junction Biogenesis Connexins are synthesized within the ER as four-transmembrane integral membrane proteins that are not glycosylated. They are posttranslationally assembled into hexameric structures known as hemichannels or connexons. Once at the cell surface, a hemichannel embedded in the plasma membrane of one cell docks head-to-head with a hemichannel in the plasma membrane of an apposing cell to form an intercellular channel. Remarkably, this entire series of events is accomplished without any covalent bonds between connexin molecules. Gap junction plaques are clusters of cell–cell channels, which can be packed at densities of upward of 10,000/mm2. Gap junctional plaques can cover more than 50% of the cell surface in mature lens fibers, although in other cell types this fraction is usually considerably less. Electrophysiological analysis of the rate at which functional gap junctional channels accumulate at cell–cell interfaces indicates that plaque formation is a cooperative self-assembly process [1,2].

9.3 Key Issues in Gap Junction Biosynthesis and Turnover 9.3.1 Insertion and Degradation of Connexins at the Endoplasmic Reticulum With very few exceptions, translocation of newly synthesized secretory pathway proteins into the ER is a strictly cotranslational process in higher eukaryotes. It was therefore remarkable when Zhang et al. [3] first reported that Cx26 translated in vitro, unlike Cx32, could become associated with pancreatic ER microsomes either cotranslationally or well after protein synthesis had been terminated, and was integrated into the ER in the correct topological orientation. Currently, ‘‘there is no evidence for or against a significant role for posttranslational insertion [of Cx26] in situ’’ [3]. With this exception, it is generally considered that insertion of connexins into the ER is a cotranslational process.

9 Biogenesis and Degradation

227

Connexins do not have classical cleaved amino-terminal domain (NT) signal sequences. Another unusual, but not unique, feature of the connexin family is that the NT of nascent Cx43, Cx32, and Cx26 molecules can be clipped while in the ER. As characterized by Falk et al. [4,5], this is likely due to recognition by signal peptidase of a cryptic signal sequence within the ER lumen. This region becomes accessible to the protease only when the connexin polypeptide fails to fold properly, notably after extreme overexpression in tissue culture cells [5] or when disulfide bond formation is prevented by reducing agents [6]. Connexins cotranslationally inserted into microsomes in vitro also undergo this cleavage, which may be an indication of imperfect folding. A role for this cleavage in ER quality control is possible, but has not been verified. It is clear that a major component of ‘‘quality control’’ of newly synthesized proteins occurs at the level of the ER. Once thought to be confined to cytosolic and nuclear proteins, the ubiquitin/proteasome system is now recognized to play a major role in the degradation of membrane and soluble proteins synthesized within the ER [7]. The process by which a protein in the ER is degraded by proteasomes is referred to as ERAD (ER-associated degradation) and can be conceptualized as consisting of four tightly linked steps: 1. Recognition, mediated by proteins (e.g., protein chaperones and E3 ligases) that bind to the ERAD substrate because of a feature of the latter that destines it to be destroyed, this being typically incorrect or incomplete folding. 2. Polyubiquitination, in which the first ubiquitin moiety becomes linked through an isopeptide bond to (most commonly) the -amino group of a lysine residue within the ERAD substrate, and is then extended to form a K48-linked polyubiquitinated chain that targets the substrate for degradation by the 26S proteasome [8]. 3. Dislocation, in which the polyubiquitinated ERAD substrate is extracted from the ER into the cytosol [9]. For most proteins, polyubiquitination is a prerequisite for complete extraction from the ER. Subsequent association of the ERAD substrate with the 26S proteasome is also ubiquitin-dependent and may require additional proteins that serve as bridging or shuttle factors. 4. Proteolysis, in which the ERAD substrate is degraded into small peptides within the 20S proteasome core. One or more of the 20S peptidase activities are blocked by proteasome inhibitors such as MG132, epoxomicin, lactacystin, Acetyl-L-Leucyl-L-Leucyl-L-Norleucinal (ALLN), or the chemotherapeutic bortezimide [10]. Hallmarks of an ERAD substrate are that it continues to be degraded if its transport from the ER to the Golgi is blocked (e.g., by brefeldin A) and that its turnover in the ER is sensitive to proteasome inhibitors. By these criteria, approximately 40 to 50% of newly synthesized wild-type Cx43 and Cx32, and up to 100% of certain disease-causing connexin mutants (e.g., the CharcotMarie-Tooth–X-linked CX32E208K mutant), can be degraded by ERAD [11]. The fraction of wild-type connexin turned over by this pathway, although large, is not as great as that reported for certain other wild-type plasma membrane proteins,

228

L.S. Musil

notably the cystic fibrosis transmembrane conductance regulator (CFTR), a chloride channel (75% in most cell types). Given that ERAD is an important mechanism of connexin turnover, it might be expected that either upregulation or downregulation of this process would have a major effect on connexin protein levels and therefore potentially on gap junction function. VanSlyke and Musil [6] reported that dislocation of Cx43 and Cx32 from the ER into the cytosol was reduced by up to 75% when cells were subjected to mild inducers of the heat shock response, for example, 10 to 30 minutes at 428C, or oxidative stress induced by 80 mM sodium arsenite. Importantly, wild-type connexin spared from ERAD by such cytosolic stress remains in a full-length, membrane-integrated form capable of folding, multisubunit assembly, and transport to the cell surface. In otherwise gap junction assembly-inefficient cells such as CHO cells, this was associated with a marked increase in the formation and function of gap junctions. These studies were the first to demonstrate that ERAD could be inhibited by physiologically relevant forms of stress to increase the functional pools of a protein on the cell surface. More recent findings suggest that cytosolic stress acts to specifically inhibit the polyubiquitination step of ERAD [12]. In analogy to other integral plasma membrane proteins such as the CFTR, it is presumed that proper folding of connexins within the ER is a slow process and that reducing the rate at which connexins are dislocated from the ER by inhibition of ERAD allows the nascent molecules more time to reach a transport-competent conformation and thus proceed to the trans-Golgi network (TGN) and eventually the plasma membrane. What might be the in vivo relevance of cytosolic stress-induced upregulation of gap junctions? Physiologically, tissues experience several types of heat shock response-inducing events, including oxidative stress. Enhanced formation of gap junctions under such conditions could serve as a means to increase the movement of reduced glutathione and other protective antioxidants into stressed cells from less affected surrounding cells. In keeping with this hypothetical scenario, reduced glutathione appears to be a major Cx43 gap junction permeant in vivo [13], and gap junction–mediated transfer of reduced glutathione has been reported to greatly improve the function of oxidatively stressed cultured cardiac myocytes [14]. Importantly, severe stresses that can compromise cell viability (e.g., hyperthermia at temperatures above 428C for extended periods) lead to a reduction instead of an increase in connexin levels and gap junction formation [15,16]. Teleologically, this would serve to prevent antioxidants and other potentially protective gap junction permeants from being wasted on irreparably damaged cells, and perhaps also reduce the transfer of toxic metabolites from such doomed cells to their neighbors. Can ERAD be upregulated to decrease gap junction formation and function? A recent study by Mitra et al. [17] using human prostate cancer cells transfected to express Cx32 indicates that it can. Withdrawal of androgens from these cells led to a precipitous drop in the level of Cx32 protein, gap junctions, and intercellular dye-coupling. Cx32 was still synthesized, but was turned over

9 Biogenesis and Degradation

229

more rapidly than normal by a process blocked by proteasome inhibitors but not by conditions that inhibit transport of Cx32 from the ER to Golgi. Although consistent with the authors’ conclusion that androgen removal enhances Cx32 ERAD, these results might also be attributable to decreased translocation [18]. Regardless of the exact mechanism involved, these findings provide another example of how gap junction formation and therefore function can be regulated at the level of the ER [17].

9.3.2 Site of Hemichannel Assembly The well-established dogma is that multisubunit oligomerization of integral membrane proteins takes place within the ER, with unassembled subunits being degraded (by ERAD) as part of the cell’s protein quality-control system. One of the most compelling lines of evidence in support of this contention is that treatments or mutations that block ER-to-Golgi transport do not interfere with oligomeric assembly. Using blockers of this transport and two independent assays for hemichannel formation (chemical cross-linking and sucrose gradient fractionation), it was shown that detectable oligomerization of endogenous wildtype Cx43 occurs exclusively after its exit from the ER, in an intracellular compartment with properties consistent with the TGN [19]. This novel site of multisubunit assembly was found for all cultured cells and tissues examined, and later confirmed by another group using similar methods [20]. In another approach to investigate the site of oligomerization, Sarma et al. [21] and Mitra et al. [22] engineered Cx32 and Cx43 to include a di-lysinebased carboxyl-terminal ER retention-retrieval sequence (HKKSL). The Cx43-HKKSL construct ran on sucrose gradients as a monomer, as might be expected from the aforementioned results obtained with endogenous, wild-type Cx43. In contrast, an analogous Cx32 construct sedimented to a position consistent with a hexamer. These findings were interpreted as indicating that Cx32 but not Cx43 is oligomerized into hemichannels within the ER or ER-Golgi intermediate compartment. These studies did not investigate the intracellular site at which endogenously expressed wild-type Cx32 is oligomerized. Using the same methodology as was employed for Cx43, it was found that oligomerization of wild-type [35S]methionine-labeled Cx32 natively expressed in either MH1C1 or HepG2 cells was prevented when the cells were metabolically labeled in the presence of the ER-to-Golgi transport inhibitor brefeldin A [23]. In contrast, oligomerization did take place in the presence of brefeldin A in several cell lines in which Cx32 was expressed exogenously. Additional experiments in which the levels of Cx43 and Cx26 were systematically altered demonstrated that the site of hemichannel assembly was dictated neither by the connexin or cell type, but by the level of connexin protein such that overexpression shifted the compartment in which assembly was initiated from the TGN to the ER.

230

L.S. Musil

These results have important implications for understanding how certain connexin mutants can cause dominant disease. It has been proposed that a transport-incompetent connexin mutant can interfere with the movement or function of a coexpressed wild-type connexin by coassembling with it into mixed (heteromeric) hemichannels [24]. Such a dominant-negative disease mechanism is unlikely to be true for a mutant that is confined to the ER given that formation of hemichannels does not take place within this compartment under physiologically relevant conditions. It remains possible, of course, that the mutant and wild-type proteins physically interact within the ER in a nonhemichannel complex or aggregate. It is for such reasons that it is essential to determine whether an association between a wild-type and mutant connexin is due to bona fide hemichannel assembly (in which case the interaction would be likely to take place within the TGN) or some other type of binding event (which could begin within the ER and contain non-connexin proteins that could influence its pathophysiological consequences). Co-immunoprecipitation and fluorescence resonance transfer experiments alone are unable to distinguish between these two possibilities. Why should hemichannels be assembled within the TGN? One suggestion emerges from studies by Kumar and Gilula [25], in which Cx32 was highly overexpressed in BHK cells. Double-membrane gap junctions formed not only at cell–cell interfaces, but also within intracellular compartments including the ER and Golgi. One could imagine that at physiologically relevant levels of connexin expression, such futile formation of intracellular gap junctions could be prevented by delaying hemichannel assembly to the last stage of the secretory pathway, namely the TGN. Once formed, these hemichannels would be rapidly transported to the cell surface in vesicles whose dimensions would hinder the head-to-head docking of hemichannels. The aforementioned results of Kumar and Gilula [25] have been used as evidence that hemichannel formation normally takes place within the ER. In light of current knowledge, it is more likely that assembly of hemichannels within the ER is an artifact of connexin overexpression. Connexins are not the only types of newly synthesized integral membrane proteins that have the potential to mediate adhesion within an ER cisterna. The classical cadherins of the adherens junction as well as the desmogleins and desmocollins of the desmosome solve this problem by being synthesized as proforms that acquire the ability to mediate homophilic trans-adhesion only after cleavage of their propieces within the TGN [26]. It will be most interesting to learn what mechanism the integral membrane adhesive proteins of the tight junction (claudins) use to avoid sealing the lumen of the ER.

9.3.3 Mechanism of Hemichannel Assembly The ability of different connexins to oligomerize, termed heteromeric compatibility, can in general be predicted from their classification in sequence-based

9 Biogenesis and Degradation

231

subgroups, the largest of which are a-group and b-group [4] (see Chapters 1 and 2). Mutagenesis experiments using Cx43 (a-group) and Cx32 (b-group) have shown that the inability of these two subgroups to coassemble is determined by at least four residues, two in the NT and two at the cytoplasmic end of the third transmembrane domain (M3) [27]. The a-group contains a positively charged amino acid followed by a polar amino acid at residues twelve and 13, and the non-aromatic residues lysine, methionine, arginine, glycine, or asparagine in M3, whereas the b-group contains two noncharged amino acids at residues eleven and twelve, and a pair of aromatic tryptophan residues in M3. It remains to be determined whether formation of heteromeric hemichannels between members of the same connexin subgroup is mechanistically different from that of homo-oligomers, and whether homo-oligomerization or heterooligomerization proceeds by stepwise addition of monomers or larger oligomers, or is more of a concerted process.

9.3.4 Transport of Hemichannels to the Cell Surface Connexins can be detected within the Golgi complex by both biochemical and morphological techniques. Delivery of newly synthesized connexin to the cell surface is prevented when ER-to-Golgi transport is blocked with brefeldin A or by expression of a dominant-negative mutant of a guanosine triphosphatase (GTPase) required for ER-to-Golgi transport vesicle formation, as would be expected for a plasma membrane protein in the constitutive secretory pathway [28]. This includes Cx26, which had been suggested by others [29] to bypass the Golgi. There is at present no evidence to suggest that the route by which connexins reach the cell surface under native conditions is anything other than the canonical ER-to-Golgi-to-TGN-to-plasma membrane pathway. If connexins are assembled into oligomers within the TGN and pass through the TGN en route to the cell surface, one would not expect to find monomeric connexin in the plasma membrane. Indeed, Koval et al. [20] reported that when Cx46 is endogenously expressed in a cell type that (for unknown reasons) is unable to assemble it into a hemichannel, ostensibly wild-type Cx46 accumulates within the TGN but is not detectable on the plasma membrane by immunofluorescence microscopy. Also consistent with a requirement for connexin oligomerization for transport to the cell surface, Lauf et al. [30] showed that Cx43 or Cx32 tagged with DsRed appears to be retained intracellularly (i.e., not in the plasma membrane) unless coexpressed with its untagged or green fluorescent protein (GFP) fusion counterpart, after which both species colocalize in gap junctional plaques, presumably as a result of coassembly into heteromeric hemichannels. For single (unassociated) cells, hemichannels in the plasma membrane must by definition be delivered to free (unapposed) regions of the cell surface. Indeed, the presence of hemichannels in the nonjunctional plasma membrane has been

232

L.S. Musil

established by morphological, biochemical, and physiological methodologies [31,32,33,34]. This leads to the question of how hemichannels become incorporated into gap junctional plaques in cells with preexisting cell–cell contacts. Two extreme models can be envisioned: 1. A lateral diffusion model, in which hemichannels are randomly delivered to the cell surface and then assembled into gap junctions by lateral diffusion and clustering in the plane of the plasma membrane. 2. A directed insertion model, in which hemichannels are vectorially inserted into the cell–cell border by fusion of connexin-containing intracellular transport vesicles with the plasma membrane at points of cell–cell contact. Hemichannels that are instead delivered to nonjunctional areas of the plasmalemma would generally not become incorporated into a plaque. Support for the former model include reports that Cx43 is frequently inserted at nonjunctional areas of the membrane surface, and that such ‘‘free’’ hemichannels have a high rate of lateral diffusion in the plane of the plasma membrane [28,35]. Moreover, photobleached gap junctions recover fluorescence predominantly at the perimeter of the plaque, as would be expected if nonbleached nonjunctional hemichannels diffuse into, and are trapped at, the external border of the bleached junction [35]. Using pulse-labeled, tetracysteine-tagged Cx43, Gaietta et al. [36] demonstrated that younger connexin molecules accumulate preferentially at the plaque periphery in an expanding ring-like pattern (see Chapter 10). For model 2 above to be correct, all of these data would have to be refuted. An alternative possibility is that gap junction formation involves both lateral diffusion and directed insertion. Evidence in support of directed insertion has recently been published by Shaw et al. [37], who reported partial recovery after photobleaching of GFPtagged Cx43 throughout the gap junctional plaque within 5 seconds but not in nonplaque areas. However, some important issues are not addressed in this study. For example, do the transport vesicles that are reported to preferentially fuse with the adherens junction in the immediate vicinity of gap junctional plaques contain predominantly connexins as their cargo? If so, what is the evidence for such ‘connexin-rich’ anterograde transport carriers, and does the type of connexin dictate with which region of the cell surface the carrier fuses? Given that cells can simultaneously express plaques containing Cx43 in one domain and Cx32 in another [38], this is more than a theoretical concern. Moreover, although a close correlation between cadherin-based cell–cell adhesion and gap junction formation has been described in certain cell types, in other cases cadherin expression has been linked to decreased gap junction formation [39]. It is worth noting the freeze-fracture electron microscopy studies of Johnson et al. [40] demonstrated that gap junction assembly can occur at 48C, a temperature at which vesicular transport (and thus directed insertion) is abolished but which for most proteins is compatible with (slowed) lateral transport within the plane of the plasma membrane.

9 Biogenesis and Degradation

233

9.3.5 What Is the Half-Life of Gap Junctional Plaques? Pulse-chase analysis of metabolically labeled connexins in tissue culture cells and whole organs has revealed that the pool of newly synthesized, immunoprecipitable connexin protein has a half-life of only 1.5 to 3.5 hours [41,42,43]. Importantly, this rapid rate cannot be attributed solely to ERAD given that rapid turnover has also been documented for connexins after their transport to the plasma membrane using cell surface biotinylation [44,45]. At least four approaches have been used to assess the rate at which plaque-associated connexins are degraded: (1) metabolic labeling with [14C]-bicarbonate in living mice followed by biochemical isolation of gap junctional plaques from liver and autoradiography [46]; (2) assessment of the time course with which immunofluorescently detectable gap junctions disappeared from cells in which further connexin expression had been terminated [47]; (3) [35S]-methionine-labeling followed by cell surface biotinylation to measure the rate at which radiolabeled Cx43 incorporated into Triton-X-100-insoluble, phosphorylated gap junctional plaques was degraded [44]; and (4) the determination of the time course of loss of tetracysteine-tagged Cx43 pulse-labeled with FlAsH (fluorescein arsenical helix binder) reagent from cell surface plaques [36]. In each case, the apparent half-life of assembled plaques was five hours or less. One well-documented exception is gap junctions in mature lens fiber cells, which persist throughout the lifetime of the organism in the absence of new protein synthesis. This, however, is a consequence of the global loss of protein degradation machinery in these highly specialized cells and does not reflect a specific property of plaque-associated connexins. A common teleological argument against rapid turnover of gap junctions is that it is perceived as wasteful. Why irreversibly destroy connexins when junctional function can be regulated by channel gating? Although highly unusual for plasma membrane proteins, which typically have turnover rates of greater than 24 hours, rapid degradation kinetics are common characteristics of cytosolic and nuclear proteins involved in signal transduction. For several such regulatory molecules, a decrease in their turnover rate is a physiologically important mechanism whereby their function is prolonged or enhanced. Because gap junction assembly appears to be a cooperative self-assembly process, reducing the rate of connexin degradation would lead to a large increase in gap junction formation and intercellular communication. Indeed, it has been found that treatments that decrease connexin turnover (e.g., protein synthesis inhibitors, cytosolic stress) rapidly induce the de novo assembly of cell surface Cx43 into long-lived, functional gap junctions [6,42].

9.3.6 How Are Gap Junction Plaques Degraded? Because gap junction channels have short half-lives, an efficient mechanism must exist to internalize and degrade them. Unlike for most other types of

234

L.S. Musil

protein assembly, turnover of a gap junctional plaque involves two cells. Previous studies have documented the formation of ‘‘annular’’ gap junctions, in which an intact gap junctional plaque is internalized as a double-membraned structure into the cytosol of one of the two partner cells [48]. Given that transient disruption of the integrity of the plasma membrane is a common occurrence in many cell types and that mechanisms for healing such wounds are widely operative [49], it is perhaps not surprising that annular gap junction formation does not appear to compromise cell viability. If annular gap junction formation were the sole mechanism by which plaques are destroyed, it would be expected that large numbers of these structures would accumulate in gap junction–rich cells in which degradation, but not uptake, of cell surface–derived connexins was inhibited. This, however, is not what has been reported [48]. The frequency with which annular gap junctions are formed is highly dependent on the type of cell and its physiological state and does not necessarily correspond to the rate at which connexins are turned over. It is therefore possible that formation of annular gap junctions is more of a regulated, stimulus-induced event than a constitutive ‘‘housekeeping’’ process. Annular gap junctions may also be preferentially used to clear very large gap junctional plaques in transfected cells. This interpretation implies that another mechanism must be responsible for much of the turnover of gap junctions under basal conditions. The most attractive and obvious alternative to en bloc degradation of gap junctions is some degree of plaque disassembly and connexin dispersal prior to internalization and destruction. Indeed, evidence for such processes has been reported in hepatocytes in regenerating mouse liver [50]. How plaques could be disassembled remains one of the major unanswered questions in gap junction dynamics. What is the proteolytic system that degrades connexins after they have been delivered to the cell surface? The main, albeit not sole, route by which plasma membrane-derived proteins are destroyed is endocytosis and subsequent proteolysis within lysosomes. Indeed, lysosomal inhibitors increase the half-life of pulselabeled connexins in many cell types, slow the loss of cell surface biotinylated Cx43, and lead to enhanced colocalization of connexins with lysosomal markers [42,45]. Annular gap junctions have been reported to be associated with both clathrin-coated vesicles [51] and lysosomes [52], within which they appear to be degraded. As first reported by Laing et al. [53], inhibitors of the proteasome that have no direct effect on lysosomal enzymes also appear to reduce the rate at which gap junctions disappear from the cell surface under conditions in which they cannot be replenished with connexins from intracellular pools. Three possibilities can be envisioned, with the last two being mutually exclusive: 1. Proteasome inhibitors could enhance cell–cell adhesion, thereby indirectly promoting formation of intercellular junctions. Indeed, Tsukamoto and Nigam [54] have reported that such compounds blocked cell–cell dissociation and stabilized tight and adherens junction proteins at cell interfaces in cultured epithelia subjected to scattering conditions.

9 Biogenesis and Degradation

235

2. Degradation of connexins from the cell surface could be mediated by both proteasomes and lysosomes, either in two independent parallel processes or in sequential steps in the same pathway. The former mechanism is highly unlikely, given that the lysosomal inhibitor chloroquine (which has no activity against the proteasome under the conditions used) can block the degradation of up to 95% of cell surface biotinylated Cx43 for over 6 hours [45]. Biotinylated Cx43 spared from turnover by chloroquine is in a fulllength state, ruling out a previously proposed model in which proteasomemediated clipping of Cx43 serves as a prerequisite for further destruction within the lysosome [53]. 3. Cell surface Cx43 is degraded within the lysosome, and proteasome blockers inhibit this process by interfering with the targeting or transport of Cx43 from the cell surface to this compartment. This mechanism is supported by the observation that incubation of various cell types with proteasome inhibitors results in the accumulation of plasma membrane pools of connexins in cell surface gap junctional plaques, whereas after chloroquine exposure connexins redistribute to the cell interior in vesicles likely to be part of the endosome/lysosome system. Moreover, sensitivity to both lysosomal and proteasomal inhibitors after transport to the cell surface has been described for a large number of non-connexin plasma membrane proteins now believed to be direct substrates of the lysosome, but not of the proteasome [55].

9.4 Major Unanswered Questions in Gap Junction Biosynthesis and Degradation 9.4.1 Why is a Large Fraction of Newly Synthesized Connexin Molecules Degraded by Endoplasmic Reticulum-Associated Degradation? In analogy to other integral plasma membrane protein ERAD substrates (e.g., the CFTR), the most likely explanation is that the folding of nascent connexin polypeptides is a slow process and that many molecules are recognized as unfolded and are degraded by ERAD before they can attain a conformation that permits them to exit the ER. If so, it might be expected that conditions that block ERAD prior to the dislocation step would increase the period of time in which newly synthesized molecules could become properly folded and competent to traverse the secretory pathway. This is precisely what is observed for the CFTR when its polyubiquitination (a prerequisite for dislocation) is blocked by expression of a catalytically inactive E2 enzyme [56], and for connexins after exposure to cytosolic stress [6]. For all other previously described multisubunit proteins, the ER quality control machinery recognizes unassembled monomers as transport-incompetent. Given that hemichannel assembly is normally initiated in the TGN, this cannot

236

L.S. Musil

be the case for connexins. The ER must instead be able to discriminate between two forms of connexin monomer: one that has attained an intermediate level of folding that is trafficked to the Golgi, and another that is recognized as being incapable of ever attaining a mature conformation and is retained, for example, the CMTX-linked CX32E208K mutant [11]. It will be important to develop assays to assess the conformational state of both wild-type and mutant connexins during their biogenesis.

9.4.2 How Are Gap Junction Plaques Nucleated? For two hemichannels to dock to form an intercellular channel, the distance between the plasma membrane bilayers of the adjoining cells must be 2 nm (20 A˚). For this to occur, large or heavily charged cell surface constituents (e.g., glycoproteins, glycolipids) must be absent from the region of hemichannel–hemichannel contact. Intramembranous particle-poor cell–cell contact areas have been observed by freeze-fracture electron microscopy at presumptive sites of gap junction formation [57]. How such ‘‘formation plaques’’ assemble, and whether they are an obligatory intermediate in the assembly of all gap junctions, is unclear [50]. The molecular basis for gap junction initiation, growth, and hemichannel packing within the plaque remains a ‘‘Holy Grail’’ in this field. Understanding these processes is also likely to shed light on how plaques are dismantled and degraded.

9.4.3 How Different Is Gap Junction Assembly and Degradation Among Connexin Family Members? As discussed above, all members of the connexin family studied to date appear to be synthesized within the ER, transported through the Golgi en route to the cell surface, oligomerize into hemichannels within the TGN, and have rapid rates of turnover both before and after plaque formation. This does not, however, preclude the real possibility that connexin-specific differences in gap junction assembly or degradation exist. It has been well established that connexins differ in their posttranslational modifications, notably phosphorylation events, and in the types of non-connexin molecules with which they physically interact. In the case of Cx43, both of these processes have been linked to one or more steps in gap junction formation, although the molecular mechanisms underlying these effects are not yet clear [58,59]. Another interesting possibility that remains to be fully explored is that the rate or regulation of degradation may be somewhat different among connexin family members. In this regard, it has recently been proposed that one essential role of Cx30 in hearing may be to coassemble with Cx26 in the cochlea, thereby stabilizing Cx26 from turnover [60] (see Chapter 20).

9 Biogenesis and Degradation

237

9.4.4 Why Are There Large Differences Between Cell Types in Their Ability to Assemble (Wild-Type) Connexins into Gap Junctions? As previously mentioned, expression of a complementary DNA (cDNA) encoding a connexin family member is sufficient to induce the formation of abundant gap junctions in many cell types. In addition to such ‘‘assembly-efficient’’ cells, two additional classes exist: assembly-inefficient cells, which are comparatively less proficient in gap junction assembly under basal conditions; and assemblyincompetent cells, defined as forming only very small and sparse gap junctions [42]. In most cases, the mechanistic basis for these phenotypes is unknown and may be cell type-dependent. Interestingly, many transformed human and other animal cells are gap junction assembly-inefficient/incompetent in that their levels of gap junctions, and therefore gap junction-mediated intercellular coupling, are low despite ongoing expression of wild-type connexins. A very large literature has shown that loss, either naturally or chemically induced, of gap junctional communication contributes to the transformed phenotype, and that restoring gap junction function controls cancer cell growth in vitro and in vivo (reviewed in [61,62]) (see Chapter 25). Increasing the ability of transformed cells to assemble gap junctions could therefore be of therapeutic value and enhance cancer therapies that are dependent on, or may be augmented by, the gap junction–mediated bystander effect [63].

9.5 Conclusion As for several other integral plasma membrane channels, connexin channel biogenesis involves noncovalent oligomerization of subunits into either homomeric or heterotypic forms whose proper assembly appears to be required for transport to the cell surface. Unique aspects are that multisubunit assembly of connexins takes place within the TGN instead of the ER, and that gap junction formation involves the coordinated action of two cells. The number and composition of gap junctions can be dynamically regulated at the level of the ER by either upregulating connexin biosynthesis or decreasing the rate of connexin degradation, and at the cell surface by enhancing gap junction assembly or reducing connexin degradation. Acknowledgments This work was supported by National Institutes of Health (NIH) grants R01 NS40740-01 and R01 EY014622.

References 1. Valiunas V, Bukauskas FF, Weingart R. Conductances and selective permeability of connexin43 gap junction channels examined in neonatal rat heart cells. Circ Res. 1997;80:708–19.

238

L.S. Musil

2. Castro C, Gomez-Hernandez JM, Silander K, Barrio LC. Altered formation of hemichannels and gap junction channels caused by C-terminal connexin-32 mutations. J Neurosci. 1999;19:3752–60. 3. Zhang JT, Chen M, Foote CI, Nicholson BJ. Membrane integration of in vitro-translated gap junctional proteins: co- and posttranslational mechanisms. Mol Biol Cell. 1996;7:471–82. 4. Falk MM, Buehler LK, Kumar NM, Gilula NB. Cell-free synthesis and assembly of connexins into functional gap junction membrane channels. EMBO J. 1997;16:2703–16. 5. Falk MM, Kumar NM, Gilula NB. Membrane insertion of gap junction connexins: polytopic channel-forming membrane proteins. J Cell Biol. 1994;127:343–55. 6. VanSlyke JK, Musil LS. Dislocation and degradation from the ER are regulated by cytosolic stress. J Cell Biol. 2002;157:381–94. 7. Meusser B, Hirsch C, Jarosch E, Sommer T. ERAD: the long road to destruction. Nat Cell Biol. 2005;7:766–72. 8. Glickman MH, Ciechanover A. The ubiquitin-proteasome proteolytic pathway: destruction for the sake of construction. Physiol Rev. 2002;82:373–428. 9. Jarosch E, Lenk U, Sommer T. Endoplasmic reticulum-associated protein degradation. Int Rev Cytol. 2003;223:39–81. 10. Groll M, Huber R. Inhibitors of the eukaryotic 20S proteasome core particle: a structural approach. Biochim Biophys Acta. 2004;1695:33–44. 11. VanSlyke JK, Deschenes SM, Musil LS. Intracellular transport, assembly, and degradation of wildtype and disease-linked mutant gap junction proteins. Mol Biol Cell. 2000;11:1933–46. 12. Kelly SM, Vanslyke JK, Musil LS. Regulation of ubiquitin-proteasome system-mediated degradation by cytosolic stress. Mol Biol Cell. 2007;18:4279–91. 13. Goldberg GS, Lampe PD, Nicholson BJ. Selective transfer of endogenous metabolites through gap junctions composed of different connexins. Nat Cell Biol. 1999;1:457–9. 14. Nakamura TY, Yamamoto I, Kanno Y, Shiba Y, Goshima K. Metabolic coupling of glutathione between mouse and quail cardiac myocytes and its protective role against oxidative stress. Circ Res. 1994;74:806–16. 15. Laing JG, Tadros PN, Green K, Saffitz JE, Beyer EC. Proteolysis of connexin43-containing gap junctions in normal and heat-stressed cardiac myocytes. Cardiovasc Res. 1998;38:711–8. 16. Hamada N, Kodama S, Suzuki K, Watanabe M. Gap junctional intercellular communication and cellular response to heat stress. Carcinogenesis. 2003;24:1723–8. 17. Mitra S, Annamalai L, Chakraborty S, Johnson K, Song X, Batra S, Mehta, P. Androgenregulated formation and degradation of gap junctions in androgen-responsive human prostate cancer cells. Mol Biol Cell. 2006;17:5400–16. 18. Kang SW, Rane NS, Kim SJ, Garrison JL, Taunton J, Hegde RS. Substrate-specific translocational attenuation during ER stress defines a pre-emptive quality control pathway. Cell. 2006;127:999–1013. 19. Musil LS, Goodenough DA. Multisubunit assembly of an integral plasma membrane channel protein, gap junction connexin43, occurs after exit from the ER. Cell. 1993;74:1065–77. 20. Koval M, Harley JE, Hick E, Steinberg TH. Connexin46 is retained as monomers in a trans-Golgi compartment of osteoblastic cells. J Cell Biol. 1997;137:847–57. 21. Sarma JD, Wang F, Koval M. Targeted gap junction protein constructs reveal connexinspecific differences in oligomerization. J Biol Chem. 2002;277:20911–8. 22. Maza J, Das Sarma J, Koval M. Defining a minimal motif required to prevent connexin oligomerization in the endoplasmic reticulum. J Biol Chem. 2005;280:21115–21. 23. VanSlyke JK, Musil LS. Exogenous overexpression of connexins induces premature assembly of connexons in the ER J. Cell Commun Adhes. 2005;12:86. 24. Omori Y, Mesnil M, Yamasaki H. Connexin 32 mutations from X-linked Charcot-MarieTooth disease patients: functional defects and dominant-negative effects. Mol Biol Cell. 1996;7:907–16.

9 Biogenesis and Degradation

239

25. Kumar NM, Gilula NB. Molecular biology and genetics of gap junction channels. Semin Cell Biol. 1992;3:3–16. 26. Koch AW, Farooq A, Shan W, Zeng L, Colman DR, Zhou MM. Structure of the neural (N-) cadherin prodomain reveals a cadherin extracellular domain-like fold without adhesive characteristics. Structure. 2004;12:793–805. 27. Lagree V, Brunschwig K, Lopez P, Gilula NB, Richard G, Falk MM. Specific amino-acid residues in the N-terminus and M3 implicated in channel function and oligomerization compatibility of connexin43. J Cell Sci. 2003;116:3189–201. 28. Thomas T, Jordan K, Simek J, Shao Q, Jedeszko C, Walton P, Laird DW. Mechanisms of Cx43 and Cx26 transport to the plasma membrane and gap junction regeneration. J Cell Sci. 2005;118:4451–62. 29. George CH, Kendall JM, Evans WH. Intracellular trafficking pathways in the assembly of connexins into gap junctions. J Biol Chem. 1999;274:8678–85. 30. Lauf U, Lopez P, Falk MM. Expression of fluorescently tagged connexins: a novel approach to rescue function of oligomeric DsRed-tagged proteins. FEBS Lett. 2001;498:11–5. 31. Sa´ez JC, Retamal MA, Basilio D, Bukauskas FF, Bennett MVL. Connexin-based gap junction hemichannels: gating mechanisms. Biochim Biophys Acta. 2005;1711:215–24. 32. Ebihara L. New roles for connexons. News Physiol Sci. 2003;18:100–3. 33. Goodenough DA, Paul DL. Beyond the gap: functions of unpaired connexon channels. Nat Rev Mol Cell Biol. 2003;4:285–94. 34. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Qtly Rev Biophys. 2001;34:325–472. 35. Lauf U, Giepmans BN, Lopez P, Braconnot S, Chen SC, Falk MM. Dynamic trafficking and delivery of connexons to the plasma membrane and accretion to gap junctions in living cells. Proc Natl Acad Sci USA. 2002;99:10446–51. 36. Gaietta G, Deerinck DT, Adams SR, Bouwer J, Tour O, Laird DW, Sosinsky GE, Tsien RY, Ellisman MH. Multicolor and electron microscopic imaging of connexin trafficking. Science. 2002;296:503–7. 37. Shaw RM, Fay AJ, Puthenveedu MA, von Zastrow M, Jan YN, Jan LY. Microtubule plus-end-tracking proteins target gap junctions directly from the cell interior to adherens junctions. Cell. 2007;128:547–60. 38. Guerrier A, Fonlupt P, Morand I, Rabilloud R, Audebet C, Krutovskikh V,Gros D, Rousset B, Munari-Silem Y. Gap junctions and cell polarity: connexin32 and connexin43 expressed in polarized thyroid epithelial cells assemble into separate gap junctions, which are located in distinct regions of the lateral plasma membrane domain. J Cell Sci. 1995;108:2609–17. 39. Wang Y, Rose B. An inhibition of gap-junctional communication by cadherins. J Cell Sci. 1997;110:301–9. 40. Johnson RG, Meyer RA, Lampe PD. Gap junction formation: a self-assembly model involving membrane domains of lipid and protein. In: Sperelakis N, Cole WC, editors. Cell interactions and gap junctions. Boca Raton, FL: CRC Press; 1989. pp. 159–79. 41. Traub O, Look J, Dermietzel R, Brummer F, Hu¨lser D, Willecke K. Comparative characterization of the 21-kD and 26-kD gap junction proteins in murine liver and cultured hepatocytes. J Cell Biol. 1989;108:1039–51. 42. Musil LS, Le AC, VanSlyke JK, Roberts LM. Regulation of connexin degradation as a mechanism to increase gap junction assembly and function. J Biol Chem. 2000;275:25207–15. 43. Beardslee MA, Laing JG, Beyer EC, Saffitz JE. Rapid turnover of connexin43 in the adult rat heart. Circ Res. 1998;83:629–35. 44. Musil LS, Goodenough DA. Biochemical analysis of connexin43 intracellular transport, phosphorylation, and assembly into gap junctional plaques. J Cell Biol. 1991;115:1357–74. 45. VanSlyke JK, Musil LS. Cytosolic stress reduces degradation of connexin43 internalized from the cell surface and enhances gap junction formation and function. Mol Biol Cell. 2005;16:5247–57.

240

L.S. Musil

46. Fallon RF, Goodenough DA. Five-hour half-life of mouse liver gap-junction protein. J Cell Biol. 1981;90:521–6. 47. Fishman GI, Gao Y, Hertzberg EL, Spray DC. Reversible intercellular coupling by regulated expression of a gap junction channel gene. Cell Adhes Commun. 1995;3:353–65. 48. Jordan K, Chodock R, Hand AR, Laird DW. The origin of annular junctions: a mechanism of gap junction internalization. J Cell Sci. 2001;114:763–73. 49. McNeil PL, Steinhardt RA. Plasma membrane disruption: repair, prevention, adaptation. Annu Rev Cell Dev Biol. 2003;19:697–731. 50. Fujimoto K, Nagafuchi A, Tsukita S, Kuraoka A, Ohokuma A, Shibata Y. Dynamics of connexins, E-cadherin and a-catenin on cell membranes during gap junction formation. J Cell Sci. 1997;110:311–22. 51. Larsen WJ, Tung HN. Origin and fate of cytoplasmic gap junction vesicles in rabbit granulosa cells. Tissue Cell. 1978;10:585–598. 52. Qin H, Shao Q, Igdoura SA, Alaoui-Jamali MA, Laird DW. Lysosomal and proteasomal degradation play distinct roles in the life cycle of Cx43 in gap junctional intercellular communication-deficient and -competent breast tumor cells. J Biol Chem. 2003;278:30005–14. 53. Laing JG, Tadros PN, Westphale EM, Beyer EC. Degradation of connexin43 gap junctions involves both the proteasome and the lysosome. Exp Cell Res. 1997;236:482–92. 54. Tsukamoto T, Nigam SK. Cell-cell dissociation upon epithelial cell scattering requires a step mediated by the proteasome. J Biol Chem. 1999;274:24579–84. 55. Longva KE, Blystad FD, Stang E, Larsen AM, Johannessen LE, Madshus IH. Ubiquitination and proteasomal activity is required for transport of the EGF receptor to inner membranes of multivesicular bodies. J Cell Biol. 2002;156:843–54. 56. Younger JM, Ren HY, Chen L, Fan CY, Fields A, Patterson C, Cyr DM. A foldable CFTRF508 biogenic intermediate accumulates upon inhibition of the Hsc70-CHIP E3 ubiquitin ligase. J Cell Biol. 2004;167:1075–85. 57. Johnson R, Hammer M, Sheridan J, Revel JP. Gap junction formation between reaggregated Novikoff hepatoma cells. Proc Natl Acad Sci USA. 1974;71:4536–40. 58. Solan JL, Lampe PD. Connexin phosphorylation as a regulatory event linked to gap junction channel assembly. Biochim Biophys Acta. 2005;1711:154–63. 59. Hunter AW, Barker RJ, Zhu C, Gourdie RG. Zonula occludens-1 alters connexin43 gap junction size and organization by influencing channel accretion. Mol Biol Cell. 2005;16:5686–98. 60. Ahmad S, Tang W, Chang Q, Qu Y, Hibshman J, Li Y, Sohl ¨ G, Willecke K, Chen P, Lin X. Restoration of connexin26 protein level in the cochlea completely rescues hearing in a mouse model of human connexin30-linked deafness. Proc Natl Acad Sci USA. 2007;104:1337–41. 61. Mesnil M, Crespin S, Avanzo JL, Zaidan-Dagli ML. Defective gap junctional intercellular communication in the carcinogenic process. Biochim Biophys Acta. 2005;1719:125–45. 62. Mesnil M. Connexins and cancer. Biol Cell. 2002;94:493–500. 63. Mesnil M, Yamasaki H. Bystander effect in herpes simplex virus-thymidine kinase/ ganciclovir cancer gene therapy: role of gap-junctional intercellular communication. Cancer Res. 2000;60:3989–99.

Chapter 10

Gap Junction Morphology and Dynamics in Situ Gina E. Sosinsky, Guido M. Gaietta and Ben N.G. Giepmans

Abstract Gap junctions serve important functions in direct intercellular communication in almost all vertebrate cell types. Cells dynamically modulate gap junctional communication by regulating the synthesis, transport, gating, and turnover of the constituent junctional channels. Since the discovery of gap junctions by electron microscopic techniques, much insight has been gained about their molecular composition and regulation. The term gap junction refers not only to the dodecameric connexin channels, but also to the two plasma membranes that they span. The plaques formed by clustered connexin channels have various packing arrangements that may be related to functionality or interaction with cytosolic binding partner proteins. The lipid composition within plaques contains a relatively high percentage of cholesterol, which confers rigidity on the structure and decreased lateral mobility to the intercellular channels. This is not the case for connexin hemichannels in nonjunctional membranes, which actively traffic laterally within the plasma membrane to plaques in cell–cell apposition areas. Moreover, connexininteracting proteins may contribute to regulation of how connexin channels pack in gap junctions. These partners are part of the ‘‘nexus’’, that is, a specialized and integrated area of the cell membrane involved in intercellular communication. Keywords Hemichannel Intercellular channel Intercellular communication Plaque Packing Electron microscopy Live cell imaging Membrane proteins Lipid bilayer Cx26 Cx30 Cx32 Cx36 Cx43 Cx45.6 Cx46 Cx50 Cx56

10.1 Introduction The gap junction is a discrete cellular structure recognizable in cross section in thin-section electron micrographs as a pentalaminar or septalaminar structure (see Chapter 2) linking two cells [1]. These morphological units G.E. Sosinsky (*) National Center for Microscopy and Imaging Research, Department of Neurosciences, University of California, San Diego, 1070 Basic Science Building, 9500 Gilman Drive, La Jolla, CA 92093-0608, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_10, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

241

242

G.E. Sosinsky et al.

contain channels used for intercellular communication [2]. Previously known as maculae communicans, gap junctions contain one plasma membrane from each cell and tens to thousands of macromolecular channels (gap junction channels) that span the two apposed plasma membranes (Fig. 10.1). Gap junction channels cluster in vivo and it was through this clustering that intramembranous particles (IMPs) were first identified in cleaved plasma membranes in freeze-fracture electron microscopy (EM) [3]. These clusters of membrane channels appear to exclude other integral membrane proteins, thereby minimizing the surface area required to bring the two plasma membranes into close apposition [4]. Repulsive and attractive forces, lipid and protein composition, and other interacting proteins play important roles in formation and organization of cell–cell junctions. This chapter discusses recent developments in understanding the dynamics of gap junction channel and plaque formation, including the correlation between packing and the open versus closed states of the channels, the lipid environment, and the role of interacting proteins in regulation of gap junction dynamics.

Fig. 10.1 Schematic of gap junction organization. (a) Illustration of the membrane topology of a connexin. (b) Gap junctions consist of two plasma membranes and the dodecameric channel (a dimer of hexamers) that spans both bilayers. (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com)

10 Junction Morphology and Dynamics

243

10.2 Plaque Organization and Packing in Situ The technique of choice for visualizing gap junction channel packing in situ has historically been freeze-fracture EM. This technique provides the resolution to visualize topological en face views of >20 A˚ (2 nm) macromolecular complexes in the plasma membranes by splitting the membrane bilayer [5]. Freeze-fracture EM shows that gap junctions in tissues have from tens to thousands of closely packed membrane channels [1,6]. In fact, the close packing of IMPs seen in these micrographs is a benchmark for identifying a gap junction plaque. In freeze-fracture EM, tight junctions, identified as continuous strands rather than particles, are often seen in close apposition to gap junctions [7]. While gap junctions were discovered because of their typical appearance in EM, recent advancements in light microscopy (LM) [8] have highlighted the dynamic behavior of gap junction plaques in living cells (see below). The arrangement of the IMPs can range from a close-packed, aggregated appearance to morphologies where they are condensed into a collection of microdomains or are arranged in rows separated by the lipid membrane areas [9]. A canonical EM image of a gap junction obtained from rat liver is shown in Fig. 10.2a, and the freeze-fracture EM in Fig. 10.2b shows a series of gap junctions found in an astrocyte. In an analysis of isolated rat liver gap junctions, the lattices were hexagonal, but imperfect, containing deviations from ideal lattice positions, disclinations (curves in the lattice lines), grain boundaries, and ‘‘vacancies’’ [10]. This strongly suggests that while the channels are closely packed, there are repulsive forces that dictate a minimum interchannel distance [4,10]. In an unusual gap junction morphology, freeze-fracture EM of hamster heart shows a close packing arrangement of intercellular channels in which narrow rows of IMPs are arrayed in circular patterns around uniformly sized areas of lipid [11]. Braun et al. [4] have proposed that the channels aggregate to minimize the area of cell–cell apposition, thereby excluding membrane proteins with larger extracellular domains or with carbohydrate groups that would not fit if interspersed with connexin channels. It is interesting to note that Windoffer et al. [12] observed heterogeneous levels of fluorescence in Cx32-GFP gap junction plaques that were suggested to be due to differences in packing. Since replica techniques cannot unambiguously identify IMPs of different protein composition, elegant adaptations of correlated LM and EM have been developed combining fluorescence microscopy and immunogold detection of connexins in freeze-fracture EM replicas. This technique, freeze-fracture replica immunolabeling (FRIL) [13], was used to reveal the distribution of Cx26 and Cx32 in livers, and determine the connexin composition in brain and lens tissue [14,15,16]. The immunogold labels in Fig. 10.2a,b illustrate identification of Cx32 and Cx43 hemichannels, respectively. In particular, FRIL and correlative LM have been used to identify connexins in various glial and neuronal cell types in the brain. Brain tissue represents a challenging specimen for studying gap

244

G.E. Sosinsky et al.

Fig. 10.2 Gap junctions display different packing morphologies. (a) Freezefracture replica immunolabeled (FRIL) rat liver gap junction containing Cx32 (gold labels). This junction displays the typical appearance of channel packing. (b) FRIL of a rat astrocyte showing a ‘‘daisy chain’’ of Cx43 (gold labels) gap junctions. Gap junctions contain a variety of channel packing ranging from (c) quasi-crystalline, (d) hexatic, (e) ‘‘strings’’, (f) ‘‘row’’, to (g) reticular. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (Adapted from FRIL images of retina [17] with permission.)

junctions at the single-channel level because it is a highly specialized organ with vastly different and complex cellular morphologies. Moreover, at least seven connexin isoforms are found in neurons and glia (see Chapter 15). Recently, Kamasawa et al. [17] classified Cx36 gap junction packing into the following categories based on FRIL studies: A, quasi-crystalline; B, noncrystalline, but closed packed (hexatic); C, ‘‘strings’’, rows of ten to 70 IMPs; D, ‘‘ribbons’’, which were similar to strings but had approximately two or three parallel rows of IMPs; E, ‘‘reticular’’ or ‘‘anastomosing’’ gap junctions (such as described above for the hamster heart), which had lipid rich central domains surrounded by approximately circular arrangements of IMPs. Illustrations of these packing morphologies are shown in Fig. 10.2c–g [17]. Areas with combinations of these morphologies were also observed: F, ‘‘meandering’’, or areas where the gap junctions were too small to be assigned to one of the above categories; and G, ‘‘fragmented’’.

10 Junction Morphology and Dynamics

245

What are the functional implications of these packing arrangements? Early studies indicated that gap junction channels in intact and physiologically coupled junctions were more widely dispersed than in uncoupled cells or in isolated gap junction plaques [6,18]. Freeze-fractured gap junctions obtained by cryo-fixation had a mean center-to-center spacing of 95 A˚ and these were of the closed packed arrangement (B-type) [6]. A rearrangement of the channels within the plaques was found when tissues were treated with either Ca2+ or H+ prior to preparation for freeze-fracture [19,20]. However, another study on cardiac tissue indicated that consistency in the freeze-fracture specimen preparation was more critical than the application of agents such as Ca2+ [21], which can decouple certain communicating cells [22]. Recently, using atomic force microscopy (AFM) imaging combined with a molecular modeling analysis, loosely packed hemichannels were identified at the edges of aquaporin-0 arrays in isolated sheep lens membranes [23,24]; however, the connexin species was not identified. The authors postulate that different local lipid compositions dictate the segregation of the two channels into microdomains and that colocalization in these microdomains serves an important function in ensuring sufficient exchange of solutes between cells through the surrounding gap junctions and water exchange through aquaporin-0. Tetragonal aquaporin-4 arrays are often observed close to gap junctions in freeze-fracture EM of astrocyte membranes in the central nervous system [25]. Because the retina is a fairly accessible tissue and can be influenced by light level, it is a good model system for trying to determine if channel packing is indicative of a functional state [17] (see Chapter 19). Over 1100 gap junctions were identified and classified in retinas. Under three different illumination conditions (phototopic, mesoscopic, and scotopic), string (C-type) and ribbon (D-type) gap junctions were found primarily and in significant percentages in the functionally defined and anatomically distinct OFF sublamina of the inner plexiform layer while larger quasi-crystalline (A-type) or hexatically packed plaques (B-type) were the predominant morphological species throughout the entire inner plexiform layer [17] (see Chapter 19). Changes in packing or shifts to different morphologies may take place in a time frame much shorter than the conditions used in Kamasawa et al. [17] to prepare tissues rather than being indicative of a particular tissue or cell origin [26]. Whether the different packing morphologies have a general function in regulating cell–cell communication, hemichannel docking selectivity, turnover, or functions other than gap junctional communication remains under investigation.

10.3 Homomeric Versus Heteromeric Connexin Channels Hemichannels (also known as connexons) can be assembled from six identical connexin isoforms (homomeric) or from more than one connexin isoform (heteromeric). Consequently, an intercellular channel can be composed of two

246

G.E. Sosinsky et al.

identical homomeric hemichannels (a homotypic junctional channel) or two hemichannels of different heteromeric or homomeric composition (a heterotypic junctional channel). Mixing of connexins within a hemichannel or junctional channel is thought to be allowed by the high conservation of primary sequence in the extracellular and transmembrane domains; however, there is some selectivity as to which connexins partner with others [27]. The functionality of heteromeric and heterotypic connexin channels form a large body of literature that was recently reviewed by Cottrell and Burt [28], who reported that there are advantages in combining isoforms to create channels with unique physiological properties or create unique coupling between different cell types. Colocalization of connexins both within plaques and tissues is of interest in discerning the functional roles that different isoforms play in vivo. Direct imaging of connexin mixing within an isolated gap junction plaque was first obtained using immunogold labeling of mouse liver plaques [29], followed by isolated gap junctions [13] and gap junctions in situ [13,25]. Mixing and segregation of Cx43 and Cx26 are found in epidermis tissue [30]. More recently, fluorescent proteins genetically tagged to connexins expressed in tissue culture cells also revealed mixing of isoforms [31,32]. Fluorescent protein tags such as green or yellow fluorescent proteins (GFP or YFP) allow for imaging of connexin in live cells whereby the ebb and flow of connexins and their gap junction structures can be monitored in real time. In particular, studies by Falk and Lauf [33] have shown that fluorescent tagging, in combination with three-dimensional (3D) deconvolution microscopy, pushes the resolution for live-cell fluorescence LM. Three-dimensional volumes of deconvolution light microscopic data demonstrate clear separation of Cx43 and Cx32 domains and Cx43 and Cx26 domains within a single gap junction plaque, as well as complete overlap of fluorescence arising from Cx32 and Cx26 within a large gap junction plaque. Fluorescence resonance energy transmission (FRET) analysis [34] of cyan fluorecent protein (CFP)-tagged Cx30 and YFP-tagged Cx26 in keratinocytes detected plasma membrane interactions as well as those in the endoplasmic reticulum (ER) where oligomerization was hypothesized to occur [35]. The FRET signal is a result of both homomeric and heteromeric associations between Cx26 and Cx30 hemichannels since the interaction distance is short enough for both to occur [35]. However, while FRET can detect close associations (<8 nm), the resolution is still insufficient for determining whether the constituent channels are homomeric, heteromeric, homotypic, or heterotypic in composition. The higher resolution technique of scanning transmission electron microscope mass analysis has also been employed to investigate the issue of mixing of connexins within gap junctions [36]. The mass of populations of intercellular channels and hemichannels in mouse liver gap junctions was measured at their lattice positions. Analysis of these histograms indicated that heterotypic channels of Cx32 and Cx26 occur in vivo in isolated gap junction plaques, but heteromeric hemichannels could not be detected. Clear segregations and mixing could be seen on the molecular level in the intact plaques, while if heteromeric hemichannels

10 Junction Morphology and Dynamics

247

existed in the split junctions, the populations would have to be a component below the detection limit. Heteromeric hemichannels have been isolated from whole lens and liver membrane homogenates in biochemical analyses [37,38] and the differences between these studies may reflect the samples being evaluated, that is, detergent purified crystals versus whole membrane fractions. However, whether heteromeric hemichannels in tissues are more the exception than the rule remains to be determined [28], but from immunofluorescence of epidermis tissue it is clear that homomeric channels composed of different isoforms often mix within plaques [30]. A systematic approach to address the presence of each family member and the ability of single connexin/hemichannel detection in these studies are required to further substantiate these observations. The diversity of connexin isoforms together with the heteromeric mixing capabilities of some of them leads to numerous possibilities of unique channels. The importance of this channel diversity might be in regulation of functional pore size (i.e., what size, shape, and charge molecules can permeate), electrophysiological properties (e.g., single-channel conductances), cell-type–dependent communication, and regulation of turnover or gating.

10.4 Lipid Composition of Gap Junction Plaques As highly studied as connexins are, not much is known about the lipid constituents in a gap junction, particularly whether specific lipids influence the functionality of the channels or if the bulk physical properties of the lipid bilayer affect channel properties (for an in-depth review on lipids and gap junction structure and function, see [39]). Characterization of the density and packing distributions of gap junctions in hydrated samples was made possible by analysis of X-ray diffraction data of pellets of purified rat or mouse liver gap junctions [39,40]. Alterations in molecular structure are reflected in differences in the meridional diffraction, whereas differences in the equatorial diffraction reflect changes in lattice packing (the meridian refers to the vertical axis, whereas the equator axis is the horizontal axis of the two-dimensional [2D] diffraction pattern). Comparison of several specimens of partially ordered, stacked, mouse liver gap junctions (containing Cx32 and Cx26) showed a basic, invariant structure in which the cytoplasmic portions of the connexin molecules extend out 90 A˚ (9 nm) from the center of the membrane channel [40]. Fiber diffraction analysis showed that the extracellular gap was 35 A˚ thick, the lipid head groups were separated by 45 to 50 A˚, and the hydrophobic core of each membrane was 32 A˚ thick. A central pore runs through the gap junction channel with an opening only at both cytoplasmic ends. Because phospholipid head groups strongly scatter X-rays and the meridional diffraction was at sufficiently high-affinity, this 40-year-old model still stands as the most detailed structural measurements of the relationship between the protein and the lipid domains in the gap junction [40].

248

G.E. Sosinsky et al.

Under physiological conditions, most mammalian biological membranes are in a liquid-like, fluid state. Plasma membranes cannot be thought of as a uniform, bulk lipid environment, but rather as microdomains of different lipid compositions. In addition, the two leaflets have different lipid compositions. Protein complexes, such as hemichannels and junctional channels, may have requirements for specific lipids to form a boundary layer around the hydrophobic outer portions of the transmembrane domains [39] and also order the bulk lipids in an analogous manner to hydration shells for soluble proteins [41]. Many channel proteins have specific requirements for lipid composition for maintaining or establishing function [42]. Assessing the lipid composition in intact plaques is difficult. The techniques performed to isolate native gap junctions often co-isolate membrane fragments from nonjunctional membranes. Gradient centrifugation, cell fractionation [3,43], highly alkaline conditions [44], and treatments with one or more detergents have all been used to purify gap junction plaques [45]. These protocols take advantage of the detergent insolubility of closely packed connexin channels. The effect of the detergents or alkali is to remove nonessential lipids from the gap junction plaques and solubilize nonjunctional plasma membrane fragments [46]. In this context, a nonessential lipid is defined as one that extracts easily with detergent or alkali treatments from gap junction plaques, yet plaque or hemichannel structure is still maintained. In general, as more lipids are extracted from the plaques, the channel-to-channel spacing decreases from 100 to 110 A˚ to 75 to 85 A˚ [47,48]; packing density is altered by the removal of lipids from the spaces between the membrane channels. However, it is important to note that the protein-to-lipid ratio and channel packing is dependent on the method of isolation and alkali-based or detergent-based crystallization [49,50]. While the long-range packing order cannot provide high-affinity electron or X-ray diffraction, the short-range order in detergent-extracted and detergent-purified plaques is sufficient to provide structural information in the resolution range of 15 to 20 A˚ with electron crystallographic techniques [10,51]. Two cases of higher resolution imaging (<10 A˚) of gap junction channels have been reported [52,53]. In these crystal forms, connexins with short cytoplasmic carboxyl-terminal domains (CT) either purposely truncated or native formed higher resolution crystals than previously with junctions from liver or heart tissue. Most lipid analysis of gap junctions was done over ten years ago (see [49]). Treatments with detergents have been shown to selectively remove the phospholipids, but the cholesterol composition remains about the same [54]. In fact, the cholesterol content of isolated gap junctions is quite high compared to other membranes [49]. Increasing the cholesterol content of Novikoff cells resulted in a dramatic and rapid increase in the assembly of functional gap junctions as measured by freeze-fracture EM and dye transfer assays [55]. Gap junctions contain phosphatidylcholine (PC), as the major phospholipid component with smaller quantities of phosophatidylethanolamine (PE), phosphatidylserine (PS), and phosphatidylinositol (PI) [49]. Variable amounts of sphingomyelin

10 Junction Morphology and Dynamics

249

have been reported, but these numbers are dependent on the method and tissue of origin of the gap junction preparation. Cholesterol has been shown to be the major neutral lipid in purified gap junctions and is found at cholesterol-to-phospholipid ratios twofold to fourfold higher than for nonjunctional plasma membrane fractions [49]. These data indicated that cholesterol might fulfill a structural role in its interactions with the connexin channels. The presence of high levels of cholesterol and sphingolipids suggests that lipids as well as the protein impart rigidity to these plaques. Cholesterol may be necessary for the ordering and anchoring of clusters of the membrane channels for proper functioning. However, these lipid compositions were determined when preparation and purification of gap junctions was not fully optimized and difficult to reproduce, and results vary with tissue or cell of origin and on the purification method used. So while relative measurements are useful, exact numbers for the lipid composition are probably not accurate. Current studies reinvestigating the structural and functional role that lipids play in gap junctions have shown that lipids that are tightly associated with connexin channels and are purified along with junctional structures; these measurements should be much more accurate using more modern spectroscopic methods [56]. Recently, Biswas and Lo [16] demonstrated with FRIL and filipin, a molecule that acts as a detector for cholesterol in freeze-fracture EM, that the majority of gap junctions in the embryonic chick lens outer cortex are cholesterol-rich. Conversely, the majority of Cx56 and Cx45.6 gap junctions in the embryonic chicken lens inner cortex are cholesterol-poor or cholesterol-deficient. Endocytotic vesicles were observed in the outer and inner cortices during fiber differentiation and maturation, leading the authors to speculate that internalization of cholesterol vesicles is a step in the aging of gap junctions in fiber cells. Two-dimensional reconstituted crystals containing two symmetry-related sets of Cx26 channels and three membrane bilayers were formed without cholesterol, using dioleoyl-PC at a lipid-to-protein ratio of 1:1. In addition, PC, PS, and PE were used to reconstitute Cx32 channels into lipid vesicles that were determined to be functionally permselective [57]; thus it appears that bulk cholesterol is not required for channel activity or for 2D crystal formation. This does not preclude that there may be constitutive cholesterol bound to the channel that is not removed by detergents. Within the past ten years, domain organization studies of the plasma membrane have focused on areas known as lipid rafts. Lipid rafts were originally defined to be areas of the plasma membrane that are insoluble when extracted by cold nonionic detergents [58], indicative of an area of differential segregation of certain membrane proteins [59]. Lipid rafts are cholesterol-rich and sphingolipid-rich microdomains in the plasma membrane [60] and contain specific marker proteins that function as part of membrane trafficking and signal transduction [61]. Another characteristic of lipid rafts is a low buoyant density (especially compared to a purified gap junction fraction) when isolated from purified plasma membranes [62]. Due to the high percentage of cholesterol, lipid rafts are rigid structures and thicker than the canonical plasma membrane.

250

G.E. Sosinsky et al.

Caveolae are specialized lipid raft domains that contain the structural proteins known as the caveolins, and thus the most effective way of disrupting caveolar function is with sterol-binding drugs that sequester cholesterol—a prominent component of lipid rafts involved in caveolae formation. Such drugs include filipin and methyl-b-cyclodextrin [63,64,65], none of which typically affect clathrin-mediated endocytosis. Lipid rafts may move laterally and cluster with other rafts to create a scaffold upon which signaling networks are composed [66,67,68]. Lin et al. [69] demonstrated that activation of protein kinase C (PKC) increases the association between Cx43 and caveolin-1, and results in a marked decrease in gap junction plaques, suggesting that PKC activity may cause gap junction Cx43 channel redistribution to lipid rafts, possibly through a phosphorylation of Cx43. The cellular src proto-oncogene product, c-src, has been found in lipid rafts [70] and c-src also can phosphorylate Cx43 to regulate its function [71]. A few studies have examined whether gap junctions are associated with or can be classified as lipid rafts. When transfected in fibroblasts, certain connexins (Cx32, Cx36, Cx43, and Cx46), but not others (Cx26 and Cx50) colocalize to some extent with caveolin-1 [72]. However, gap junctions do not qualify as lipid rafts in themselves. These colocalizations of connexins with lipid raft markers might be due to both the biochemical isolation procedures that purify heterogeneous populations of lipid rafts and the association of lipid microdomains with nonjunctional hemichannels trafficking to the plasma membrane [73]. Therefore, Locke et al. [73] suggest that hemichannels within the plasma membrane may be transported in lipid rafts to gap junctions. Interestingly, gap junctions become internalized during methyl-b-cyclodextrin treatment, indicating that there is some relationship between plasma membrane cholesterol content and gap junction structures. How this observation relates to gap junction–lipid raft interactions remains to be determined.

10.5 Mobility of Channels and Hemichannels Studied in Live Cells The development of connexins tagged with aequorin [74], fluorescence, or tetracysteine-based tags (reviewed in [8]) has allowed the study of dynamic processes using live cell imaging [32,75,76]. However, tagging and/or overexpression of connexins often results in abnormally large gap junction plaques in cultured cells [76,77,78]. Moreover, GFP tagging of the amino terminus results in nonfunctional channels, while connexin channels with GFP at the carboxyl terminus are functional, but single-channel properties may be modified [79,80]. Differences between typical punctate and large plaques could reflect the redistribution of integrins, adhesion molecules, and other plasma membrane proteins from areas normally close to gap junctions to other parts to the plasma membrane as well as overexpression artifacts from using high levels of exogenous protein. Despite the above limitations, fluorescent tags are highly useful for obtaining information for

10 Junction Morphology and Dynamics

251

dissecting trafficking, signaling, and interaction pathways in living cells, when used at moderate levels and verified in tissue and in cells with endogenous proteins. Connexins have such a short half-life (1.5 to 5.0 hours, with specific times depending on the connexin and cell type; [81]) that turnover may be a key element of coupling regulation. It has been hypothesized that this high turnover rate [81] in combination with a low percentage of functional channels (10% within a gap junction plaque [80]) is a way that cells regulate their degree of coupling [80,82] (see Chapter 9). Rapid gap junction turnover is a characteristic of both immortalized tissue culture cells as well as native cells, including hepatocytes in liver [83], cardiac cells [84], and lens cells [85]. A counterexample to this is the two to three day turnover found in lens fiber connexins that may reflect the high degree of structural stability of this tissue [86,87]. In other cells, fast moving populations of connexin-tagged transport vesicles can be seen in live cell imaging movies coming to and from junctional and nonjunctional plasma membrane [12,32,75,76]. Plaques often move significantly during long observation times, fusing and separating and changing their shapes. Environmental factors such as cytosolic stress can affect gap junction plaque dynamics by reducing degradation from the cell surface [88]. However, whether gap junctions are larger or increased in number under conditions of stress remains to be determined. In addition, plaque remodeling can occur quickly as when cells lose their adherency or when they go through mitosis. In these instances, gap junctions are quickly internalized [89]. While most vital imaging studies have focused on qualitative assessment of connexin trafficking, a small number of studies have used even more sophisticated methods of fluorescence imaging. Diffusion rates were calculated from fluorescence recovery after photobleaching of Cx43-GFP plaques expressed in HeLa cells and also from the rate of movement of lipid areas that are often found in very large exogenous gap junctions. The rate of movement of the plaque components is intermediate between cytoskeletal-linked plasma membrane (slow at 10–7 to 10–8 cm2/seconds) and free lipid diffusion (fast at 10–9 cm2/seconds; [78]). This reflects the lateral immobility of the membrane channels, the close appositional area necessary for gap junction formation, and also the cholesterolrich plasma membrane that has increased stiffness. Application of fluorescence cross-correlation spectroscopy to fluorescent-tagged Cx46 hemichannels in HeLa cells demonstrated codiffusion rates of a slow component (8.9 10–9 cm2/seconds and 2 10–9 cm2/seconds), which was attributed to differential membrane curvature across the plasma membrane [90]. The order of magnitude of these measurements is consistent with hemichannels being a faster moving species than connexin channels in gap junction plaques that may lead to a fast incorporation at the edge of gap junction plaques [76] (Fig. 10.3). Rapid lateral movement of hemichannels and gap junction formation might be critical in contact formation between monocytes and tissue-specific antigen presenting cells that may result in peptide cross-presentation [91]. Using optical pulse-chase with fluorescent tags, new hemichannels are known to be added to the periphery and channels removed from the central

252

G.E. Sosinsky et al.

Fig. 10.3 Fluorescence recovery after photobleaching reveals that gap junction accretion occurs by addition to the edges of plaques. A large Cx43-GFP gap junction plaque is shown just prior to bleaching (pre-bleach, left column), one minute after bleaching (center) and 50 minutes after bleaching (right column). En face and cross-sectional views (rows a, b, respectively) from the three-dimensional reconstruction reveal that after photobleaching of a central area of the plaque, fluorescence is detected at edges 50 minutes post-bleach, indicating that new channels accrue at the plaque perimeter. (c) Illustration of the interpretation of these photobleaching experiments. The photobleached region is represented by the white area in the gap junction (GJ). The black regions represent fluorescence from portions of the gap junction outside of the photobleaching light source and the newly added connexin channels at the edge of the photobleached area. Vesicles containing hexamers (small gray circles) move from the Golgi apparatus (GA) to and fuse with the plasma membrane (PM) where, at some point, they dock with partner hexamers from an apposing cell to form channels, and then the connexin channels move to the perimeter of the plaque. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

region of plaques in cells [76,92] (Fig. 10.4), although a recent study challenged this mechanism [93]. Moreover, pulse-chase labeling suggested that newly synthesized Cx43 was transported predominantly in 100 to 150 nm vesicles to the plasma membrane and incorporated at the periphery of existing gap junctions. However, these vesicles are much larger in size than ones that sit underneath the plasma membrane and are likely part of the degradation cycle. Eventually, these vesicles fuse with lysosomes for protein breakdown. Older populations of connexins were removed from the center of the plaques into pleiomorphic vesicles of widely varying sizes as well as internalized gap junctions (annular junctions) and connexin-bearing lysosomes [32]. In rare instances, exocytotic vesicles were observed fusing with the plasma membrane, but no distinct hemichannel domains separate from the edges of gap junction plaques were detected. Endocytotic vesiculation of older populations of gap junction channels sitting underneath the plaque as seen in electron micrographs are consistent with the conclusion from the red/green bull’s-eye images seen in 3D reconstructions from confocal microscopy optical stacks. Another study of the closed channel mutant Cx43T154A contains data suggestive of a

10 Junction Morphology and Dynamics

253

Fig. 10.4 Connexin channels are added to the outside of the plaques. (a) Illustration of the optical pulse-chase experiments using FlAsH (green fluorescence) and ReAsH (red fluorescence) that distinguish between older (green) and newly (red) synthesized proteins. (b) Tetracysteine-tagged Cx43 gap junctions reveal that new channels (red) have been added to the outside of the plaque while older ones (green) are found in the center of the plaque. (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com)

hemichannel plaque next to a gap junction, consistent with the idea that hemichannels dock and are added to the outer edges of the plaque [76,92]. In summary, gap junctions are dynamic structures with channels being continually added and removed at the plasma membrane level. While these structures are thought to have great rigidity, large gap junctions can ebb and flow. It is believed that individual hemichannels have great mobility in the plasma membrane whereby they can move to the plaque edges.

10.6 Hemichannels in the Plasma Membrane Pools or domains of hemichannels must exist in the plasma membrane in order to quickly build gap junction plaques. Whether these are permanently in their closed state until incorporated into gap junction channels, and thus just serve as a trafficking intermediate [94,95], or exist as functional channels [96,97,98,99,100] under (patho)physiological conditions is still controversial especially in light of data showing that pannexins form functional single channels [101,102,103]. Preus et al. [104] postulated the existence of formation plaques from hexamers in the plasma membrane such that once two cells adhere, they produce regions with clustered 90 to 100 A˚ IMPs that interact across an initially wide extracellular space. This could bring the plasma membrane closer together, and when a critical particle density or proximity is reached, small gap junctions can form. Although

254

G.E. Sosinsky et al.

biochemical studies have shown indirectly that unpaired Cx43 hemichannels are in the plasma membrane, direct hemichannel visualization has not been demonstrated to date. Advancing new imaging technologies, single molecule visualization, and correlated LM and EM [8] might help to answer their existence in vivo or in situ. Musil and Goodenough [95] have shown that Cx43 oligomerization occurs in the trans-Golgi network, and the hexamer is transported to the plasma membrane. While the mechanisms of trafficking and oligomerization of other connexin isoforms may be different from Cx43, in general oligomerization into hexamers occurs prior to plasma membrane delivery and insertion [105,106,107] (see Chapter 9). Hemichannel plaques have been imaged with AFM in isolated preparations [108] and using freeze-fracture EM in Xenopus oocytes expressing exogenous Cx50 [109]. Endogenous Cx43 in the plasma membrane of cells might not only be in gap junctions, but also in small ‘‘hemi-junctions’’ below or just at the microscopic detection level that add to the edges of the plaque. In situ hemichannel domains at the EM level remain to be further explored [76]. At the LM level, Lauf et al. [92] showed that vesicles containing hemichannels fusing with the plasma membrane represent rare events during normal trafficking and are evidence of nonjunctional hemichannels in Cx43-GFP expressing HeLa cells [78,92]. However, these hemichannel trafficking vesicles could be sitting just underneath the plasma membrane, which cannot be resolved with light microscopy. Moreover, it remains to be established if this artificial situation represents the localization of endogenous hemichannels. Future studies are necessary to further address the localization of hemichannels in situ and in live cells and the mechanism of hemichannel delivery, which might reveal connexin-isotype specific plaque formation.

10.7 Connexin-Binding Proteins and Plaque Dynamics The most in-depth studies on the binding of connexins to cytosolic proteins have focused on Cx43 (see Chapter 11). It is well established that certain cytosolic proteins bind to Cx43 and affect Cx43 channel gating properties [110]. Cx43 can interact with several different signaling, scaffolding proteins, cytoskeletal and other cell–cell junction proteins, including the tight junction-associated proteins, zonula occludins 1 and 2 (ZO-1 and ZO-2). In addition, cell adhesion proteins were colocalized at cell–cell contact sites during the reappearance of gap junction plaques in regenerating hepatocytes [7], while Xu et al. [111] present evidence for the involvement of N-cadherin and b-catenins. The topic of how interaction partners feed back onto connexins is most recently reviewed in Giepmans [112]. This section discusses the potential impact of these binding partners on plaque formation or remodeling. Of particular interest in gap junction plaque dynamics is the interaction of ZO-1 with Cx43. Several other connexin-isoforms have also been shown to interact or colocalize with ZO-1 (Fig. 10.5) [113]. ZO-1 contains three PDZ

10 Junction Morphology and Dynamics

255

Fig. 10.5 Colocalization of Cx43 and ZO-1 in rat fibroblasts. (a) Cx43 immunofluorescence image (green). (b) Same area also immunostained for ZO-1 (red). (c) The merged image shows distinct overlap (yellow) at the plasma membrane of Cx43 and ZO-1. (A highresolution color version of this figure is available on the accompanying CD and online at www.springerlink.com)

domains, the second of which binds to the Cx43 CT domain [114]. In fact, modifications or additions to certain CT residues interfere with ZO-1 binding [114,115]. Most proteins with PDZ domains are membrane-associated and are found at specialized membrane domains such as synapses, junctions, and apical-basolateral interface regions. Because Cx43-GFP does not bind ZO-1 and forms large gap junctions, but can be rescued by wild-type Cx43, Hunter and colleagues [116] proposed that ZO-1 controls the rate of Cx43 channel accretion at the periphery of gap junction and thus regulates gap junction size and distribution. Consistent with the idea that ZO-1 plays a regulatory role is the observation that Cx43 plaques grow due to channel accumulation when Cx43 loses it ability to interact with ZO-1 [76,117]. Of interest, ZO-1 has also been implicated by direct regulation of gap junction gating, serving as a scaffold that connects a phospholipase directly to gap junctions and thereby controlling local Pl levels at the gap junction [118]. When gap junctions were first discovered, the idea was put forth that gap junction plaques were part of a higher order cellular organization. Recent data reveal binding of cytoskeletal proteins such as tubulins [119] and actin linker proteins [120] to Cx43. These underlying microtubules seem to form a 3D net underlying gap junction plaques [93]. Cytoplasmic trafficking of nonjunctional Cx43 hemichannels to the plasma membrane has been shown to be microtubule-dependent [75,92,93]. The functional significance of interactions with the actin cytoskeleton is much less clear. As discussed by Duffy et al. [121], the function of gap junctions as part of a larger organization, referred to as the nexus, may serve to provide stability for cytoskeletal networks, act as a signaling subdomain, interact with other junctional complexes, or regulate the passage of molecules between cells for other signaling pathways. Future work should include not only determining the binding site and kinetics of connexin-interacting partners, but also the supramolecular structures that the connexins and interacting proteins form at electron microscopic resolution.

256

G.E. Sosinsky et al.

10.8 Conclusion Current textbook illustrations depicting gap junctions as static structures have now been supplanted with a new paradigm of connexins, hemichannels, and gap junctions as highly mobile, dynamic, and interactive structures. Continuing questions remain: What kind of functions or functional changes give rise to the various morphologies? How do junctional and nonjunctional lipids influence plaque dynamics? How do connexin channels move through the plasma membrane? Is there a higher order organization involving gap junctions and cytosolic signaling or cytoskeletal networks? As imaging technology pushes toward single molecule detection, the results that will be obtained will have high value in several fields of cell biology, neurobiology, and biophysics and will aid in our understanding of the importance of cell–cell communication. Acknowledgments The authors are grateful to Dr. John Rash for generously providing the unpublished micrographs in Fig. 10.2 and for valuable discussions on channel packing. We also thank Tom Deerinck for supplying the image for Fig. 10.4. Support was contributed by National Science Foundation (NSF) grant MCB0543934, GM072881, and GM065937 (all to GES). Some of the work included here was conducted at the National Center for Microscopy and Imaging Research at San Diego, which is supported by National Institutes of Health (NIH) grant RR04050 awarded to Dr. Mark Ellisman.

References 1. Goodenough DA, Revel J-P. A fine structural analysis of intercellular gap junctions in the mouse liver. J Cell Biol. 1970;45:272–90. 2. Gilula NB, Reeves OR, Steinbach A. Metabolic coupling, ionic coupling, and cell contacts. Nature. 1972;235:262–5. 3. Goodenough DA. Bulk isolation of mouse hepatocyte gap junctions. Characterization of the principle protein, connexin. J Cell Biol. 1974;61:557–63. 4. Braun J, Abney JR, Owicki JC. How a gap junction maintains its structure. Nature. 1984;310:316–8. 5. Severs NJ, Shotton DM, editors. Rapid freezing, freeze-fracture, and deep etching. New York: Wiley-Liss; 1995. 6. Hirokawa N, Heuser J. The inside and outside of gap junction membranes visualized by deep etching. Cell. 1982;30:395–406. 7. Fujimoto K, Nagafuchi A, Tsukita S, Kuraoka A, Ohokuma A, Shibata Y. Dynamics of connexins, E-cadherin and -catenin on cell membranes duing gap junction formation. J Cell Sci. 1997;110:311–322. 8. Giepmans BN, Adams SR, Ellisman MH, Tsien RY. The fluorescent toolbox for assessing protein location and function. Science. 2006;312:217–24. 9. Shivers RR, McVicar LK. Gap junctions revealed by freeze-fracture electron microscopy. Microsc Res Tech. 1995;31:437–45. 10. Sosinsky GE, Baker TS, Caspar DLD, Goodenough DA. Correlation analysis of gap junction lattices. Biophys J. 1990;58:1213–26. 11. Skepper JN, Navarantnam V. Variations in the structure of nexuses in the myocardium of the golden hamster Mesocrietus auratus. J Anat. 1986;149:143–55. 12. Windoffer R, Beile B, Leibold A, Thomas S, Wilhelm U, Leube RE. Visualization of gap junction mobility in living cells. Cell Tissue Res. 2000;299:347–62.

10 Junction Morphology and Dynamics

257

13. Kuraoka A, Iida H, Hatae T, Shibata Y, Itoh M, Kurita T. Location of gap junction proteins, connexins 32 and 26, in rat and guinea pig liver as revealed by quick-freeze, deep-etch immunoelectron microscopy. J Histochem Cytochem. 1993;41:971–80. 14. Rash JE, Dillman RK, Morita M, Whalen LR, Guthrie PB, Fay-Guthrie D, Wheeler DW. Grid-mapped freeze-fracture: correlative confocal laser scanning microscopy and freezefracture electron microscopy of preselected cells in tissue slices. In: Severs NJ, Shotton DM, editors. Rapid Freezing, Freeze-fracture and Deep Etching. New York, Wiley-Liss, 1995. 15. Zampighi GA, Planells AM, Lin D, Takemoto D. Regulation of lens cell-to-cell communication by activation of PKCg and disassembly of Cx50 channels. Invest Ophthalmol Vis Sci. 2005;46:3247–55. 16. Biswas SK, Lo WK. Gap junctions contain different amounts of cholesterol which undergo unique sequestering processes during fiber cell differentiation in the embryonic chicken lens. Mol Vis. 2007;13:345–59. 17. Kamasawa N, Furman CS, Davidson KG, Rash JE, Sampson JA, Magnie AR, Gebhardt BR, Kamasawa M, Yasumura T, Zumbrunnen JR, Pickard GE, Nagy JI. Abundance and ultrastructural diversity of neuronal gap junctions in the OFF and ON sublaminae of the inner plexiform layer of rat and mouse retina. Neuroscience. 2006;142:1093–117. 18. Raviola E, Goodenough DA, Raviola G. Structure of rapidly frozen gap junctions. J Cell Biol. 1980;87:273–9. 19. Peracchia C, Peracchia LL. Gap junction dynamics: reversible effect of hydrogen ions. J Cell Biol. 1980;87:719–27. 20. Peracchia C, Peracchia LL. Gap junction dynamics: reversible effects of divalent cations. J Cell Biol. 1980;87:708–18. 21. Page E, Karrison T, Upshaw-Earley J. Freeze-fractured cardiac gap junctions: structural analysis by three methods. Am J Physiol. 1983;244:H525–39. 22. Peracchia C. Calcium effects on gap junction structure and cell coupling. Nature. 1978;271:669–71. 23. Buzhynskyy N, Hite RK, Walz T, Scheuring S. The supramolecular architecture of junctional microdomains in native lens membranes. EMBO Rep. 2007;8:51–5. 24. Scheuring S, Buzhynskyy N, Jaroslawski S, Goncalves RP, Hite RK, Walz T. Structural models of the supramolecular organization of AQP0 and connexons in junctional microdomains. J Struct Biol. 2007;160:385–94. 25. Rash JE, Yasumura T, Dudek FE, Nagy JI. Cell-specific expression of connexins and evidence of restricted gap junctional coupling between glial cells and between neurons. J Neurosci. 2001;21:1983–2000. 26. Rash J, Pickard G, Davidson K, O’Brien J, Hartwick A, Kamasawa N, Yasumura T, Nagy J. Exposure to dopamine and its D1 receptor antagonist SCH23390 produces largescale ultrastructural plasticity and changes in phosphorylation of connexin36 in neuronal gap junctions of adult rat retina. ASCB Annual Meeting. 2007;910. 27. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys. 2001;34:325–472. 28. Cottrell GT, Burt JM. Functional consequences of heterogeneous gap junction channel formation and its influence in health and disease. Biochim Biophys Acta. 2005;1711:126–41. 29. Zhang J-T, Nicholson BJ. The topological structure of connexin 26 and its distribution compared to connexin 32 in hepatic gap junctions. J Membr Biol. 1994;139:15–29. 30. Risek B, Klier FG, Gilula NB. Developmental regulation and structural organization of connexins in epidermal gap junctions. Dev Biol. 1994;164:183–96. 31. Laird DW, Jordan K, Thomas T, Qin H, Fistouris P, Shao Q. Comparative analysis and application of fluorescent protein-tagged connexins. Microsc Res Tech. 2001;52:263–72. 32. Falk M. Connexin-specific distribution within gap junctions revealed in living cells. J Cell Sci. 2000;113:4109–20. 33. Falk MM, Lauf U. High-affinity, fluorescence deconvolution microscopy and tagging with the autofluorescent tracers CFP, GFP, and YFP to study the structural composition of gap junctions in living cells. Microsc Res Tech. 2001;52:251–62.

258

G.E. Sosinsky et al.

34. Gu Y, Di WL, Kelsell DP, Zicha D. Quantitative fluorescence resonance energy transfer (FRET) measurement with acceptor photobleaching and spectral unmixing. J Microsc. 2004;215:162–73. 35. Di WL, Gu Y, Common JE, Aasen T, O’Toole EA, Kelsell DP, Zicha D. Connexin interaction patterns in keratinocytes revealed morphologically and by FRET analysis. J Cell Sci. 2005;118:1505–14 36. Sosinsky GE. Mixing of connexins in gap junction membrane channels. Proc Natl Acad Sci USA. 1995;92:9210–14. 37. Harris AL, Walter A, Paul D, Goodenough DA, Zimmerberg J. Ion channels in single bilayers induced by rat connexin32. Molec Brain Res. 1992;15:269–80. 38. Jiang JX, Goodenough DA. Heteromeric connexons in lens gap junction channels. Proc Natl Acad Sci USA. 1996;3:1287–91. 39. Cascio M. Connexins and their environment: effects of lipids composition on ion channels. Biochim Biophys Acta. 2005;1711:142–53. 40. Makowski L, Caspar DLD, Phillips WC, Goodenough DA. Gap junction structure II. Analysis of the X-ray diffraction data. J Cell Biol. 1977;74:629–45. 41. Badger J, Caspar DL. Water structure in cubic insulin crystals. Proc Natl Acad Sci USA. 1991;88:622–6. 42. Valiyaveetil FI, Zhou Y, MacKinnon R. Lipids in the structure, folding, and function of the KcsA Kþ channel. Biochem. 2002;41:10771–7. 43. Evans WH, Gurd JW. Preparation and properties of nexus and lipid enriched vesicles from mouse liver plasma membranes. Biochem J. 1972;128:691–700. 44. Hertzberg EL. A detergent independent procedure for the isolation of gap junctions from rat liver. J Biol Chem. 1984;259:9936–43. 45. Hand GM, Mu¨ller DJ, Nicholson BJ, Engel A, Sosinsky GE. Isolation and characterization of gap junctions from tissue culture cells. J Mol Biol. 2002;315:587–600. 46. Gogol E, Unwin N. Organization of connexons in isolated rat liver gap junctions. Biophys J. 1988;54:105–12. 47. Makowski L, Caspar DLD, Goodenough DA, Phillips WC. Gap junction structures III. The effect of variations in isolation procedures. Biophys J. 1982;37:189–91. 48. Sosinsky GE, Nicholson BJ. Structural organization of gap junction channels. Biochim Biophys Acta. 2005;1711:99–125. 49. Malewicz B, Kumar VV, Johnson RG, Baumann WJ. Lipids in gap junction assembly and function. Lipids. 1990;25:419–27. 50. Ghoshroy S, Goodenough DA, Sosinsky GE. Preparation, characterization, and structure of half gap junctional layers split with urea and EGTA. J Membr Biol. 1995;146:15–28. 51. Caspar DLD, Goodenough DA, Makowski L, Phillips WC. Gap junctions structures I. correlated electron microscopy and X-ray diffraction. J Cell Biol. 1977;74:605–28. 52. Unger VM, Kumar NM, Gilula NB, Yeager M. Three-dimensional structure of a recombinant gap junction membrane channel. Science. 1999;283:1176–80. 53. Oshima A, Tani K, Hiroaki Y, Fujiyoshi Y, Sosinsky GE. Three-dimensional structure of a human connexin26 gap junction channel reveals a plug in the vestibule. Proc Natl Acad Sci USA. 2007;104:10034–9. 54. Henderson D, Eibl H, Weber K. Structure and biochemistry of mouse hepatic gap junctions. J Mol Biol. 1979;132:193–218. 55. Meyer R, Malewicz B, Baumann WJ, Johnson RG. Increased gap junction assembly between cultured cells upon cholesterol supplementation. J Cell Sci. 1990;96:231–8. 56. Locke D, Harris AL. Interactions of hemi- and junctional channels with different phospholipids. Biophys J. 2007;92:443a. 57. Rhee SK, Bevans CG, Harris AL. Channel-forming activity of immunoaffinity-purified connexin32 in single phospholipid membranes. Biochem. 1996;35:9212–23.

10 Junction Morphology and Dynamics

259

58. Brown DA, London E. Functions of lipid rafts in biological membranes. Annu Rev Cell Dev Biol. 1998;14:111–36. 59. Simons K, Ikonen E. Functional rafts in cell membranes. Nature. 1997;387:569–72. 60. Anderson RG, Jacobson K. A role for lipid shells in targeting proteins to caveolae, rafts, and other lipid domains. Science. 2002;296:1821–5. 61. Parton RG, Richards AA. Lipid rafts and caveolae as portals for endocytosis: new insights and common mechanisms. Traffic. 2003;4:724–38. 62. Shogomori H, Brown DA. Use of detergents to study membrane rafts: the good, the bad, and the ugly. Biol Chem. 2003;384:1259–63. 63. Rothberg KG, Heuser JE, Donzell WC, Ying YS, Glenney JR, Anderson RG. Caveolin, a protein component of caveolae membrane coats. Cell. 1992;68:673–82. 64. Neufeld EB, Cooney AM, Pitha J, Dawidowicz EA, Dwyer NK, Pentchev PG, BlanchetteMackie EJ. Intracellular trafficking of cholesterol monitored with a cyclodextrin. J Biol Chem. 1996;271:21604–13. 65. Orlandi PA, Fishman PH. Filipin-dependent inhibition of cholera toxin: evidence for toxin internalization and activation through caveolae-like domains. J Cell Biol. 1998;141:905–15. 66. Allen JA, Halverson-Tamboli RA, Rasenick MM. Lipid raft microdomains and neurotransmitter signaling. Nat Rev Neurosci. 2007;8:128–40. 67. Brown DA. Lipid rafts, detergent-resistant membranes, and raft targeting signals. Physiology. 2006;21:430–9. 68. Michel V, Bakovic M. Lipid rafts in health and disease. Biol Cell. 2007;99:129–40. 69. Lin D, Zhou J, Zelenka PS, Takemoto DJ. Protein kinase Cg regulation of gap junction activity through caveolin-1-containing lipid rafts. Invest Ophthalmol Vis Sci. 2003;44:5259–68. 70. Hur EM, Park YS, Lee BD, Jang IH, Kim HS, Kim TD, Suh PG, Ryu SH, Kim KT. Sensitization of epidermal growth factor-induced signaling by bradykinin is mediated by c-Src. Implications for a role of lipid microdomains. J Biol Chem. 2004;279:5852–60. 71. Giepmans BN, Hengeveld T, Postma FR, Moolenaar WH. Interaction of c-Src with gap junction protein connexin-43. Role in the regulation of cell-cell communication. J Biol Chem. 2001;276:8544–9. 72. Schubert AL, Schubert W, Spray DC, Lisanti MP. Connexin family members target to lipid raft domains and interact with caveolin-1. Biochem. 2002;41:5754–64. 73. Locke D, Liu J, Harris AL. Lipid rafts prepared by different methods contain different connexin channels, but gap junctions are not lipid rafts. Biochem. 2005;44:13027–42. 74. Martin PE, George CH, Castro C, Kendall JM, Capel J, Campbell AK, Revilla A, Barrio LC, Evans WH. Assembly of chimeric connexin-aequorin proteins into functional gap junction channels: reporting intracellular and plasma membrane calcium environments. J Biol Chem. 1998;273:1719–26. 75. Jordan K, Solan JL, Dominguez M, Sia M, Hand A, Lampe PD, Laird DW. Trafficking, assembly, and function of a connexin43-green fluorescent protein chimera in live mammalian cells. Molec Biol Cell. 1999;10:2033–50. 76. Gaietta G, Deerinck TJ, Adams SR, Bouwer J, Tour O, Laird DW, Sosinsky GE, Tsien RY, Ellisman MH. Multicolor and electron microscopic imaging of connexin trafficking. Science. 2002;296:503–7. 77. Hunter AW, Jourdan J, Gourdie RG. Fusion of GFP to the carboxyl terminus of connexin43 increases gap junction size in HeLa cells. cell Commun Adhes. 2003;10:211–14. 78. Lopez P, Balicki D, Buehler LK, Falk MM, Chen SC. Distribution and dynamics of gap junction channels revealed in living cells. Cell Commun Adhes. 2001;8:237–42. 79. Contreras JE, Sa´ez JC, Bukauskas FF, Bennett MVL. Gating and regulation of connexin 43 (Cx43) hemichannels. Proc Natl Acad Sci USA. 2003;100:11388–93. 80. Bukauskas FF, Jordan K, Bukauskiene A, Bennett MV, Lampe PD, Laird DW, Verselis VK. Clustering of connexin43-enhanced green fluorescent protein gap junction channels and functional coupling in living cells. Proc Natl Acad Sci USA. 2000;97:2556–61.

260

G.E. Sosinsky et al.

81. Laird DW. The life cycle of a connexin: gap junction formation, removal, and degradation. J Bioener Biomembr. 1996;28:311–8. 82. Musil LS, Le A-CN, VanSlyke JK, Roberts LM. Regulation of connexin degradation as a mechanism to increase gap junction assembly and function. J Biol Chem. 2000;275:25207–15. 83. Fallon RF, Goodenough DA. Five-hrs half-life of mouse liver gap junction protein. J Cell Biol. 1981;90:521–6. 84. Beardslee MA, Laing JG, Beyer EC, Saffitz JE. Rapid turnover of connexin43 in the adult rat heart. Circ Res. 1998;83:629–35. 85. Musil LS, Beyer EC, Goodenough DA. Expression of the gap junction protein connexin43 in embryonic chick lens: molecular cloning, ultrastructural localization, and posttranslational phosphorylation. J Membr Biol. 1990;116:163–75. 86. Jiang JX, Goodenough DA. Phosphorylation of lens-fiber connexins in lens organ cultures. Eur J Biochem. 1998;255:37–44. 87. Berthoud VM, Bassnett S, Beyer EC. Cultured chicken embryo lens cells resemble differentiating fiber cells in vivo and contain two kinetic pools of connexin56. Exp Eye Res. 1999;68:475–84. 88. VanSlyke JK, Musil LS. Cytosolic stress reduces degradation of connexin43 internalized from the cell surface and enhances gap junction formation and function. Mol Biol Cell. 2005;16:5247–57. 89. Solan JL, Fry MD, TenBroek EM, Lampe PD. Connexin43 phosphorylation at S368 is acute during S and G2/M and in response to protein kinase C activation. J Cell Sci. 2003;116:2203–11. 90. Gerken M, Thews E, Tietz C, Wrachtrup J, Eckert R. Diffusion behavior of gap junction hemichannels in living cells. Curr Pharm Biotechnol. 2005;6:151–8. 91. Neijssen J, Herberts C, Drijfhout JW, Reits E, Janssen L, Neefjes J. Cross-presentation by intercellular peptide transfer through gap junctions. Nature. 2005;434:83–8. 92. Lauf U, Giepmans BN, Lopez P, Braconnot S, Chen SC, Falk MM. Dynamic trafficking and delivery of connexons to the plasma membrane and accretion to gap junctions in living cells. Proc Natl Acad Sci USA. 2002;99:10446–51. 93. Shaw RM, Fay AJ, Puthenveedu MA, von Zastrow M, Jan YN, Jan LY. Microtubule plus-end-tracking proteins target gap junctions directly from the cell interior to adherens junctions. Cell. 2007;128:547–60. 94. Evans WH. Assembly of gap junction intercellular communication channels. Biochem Soc Trans. 1994;22:788–92. 95. Musil LM, Goodenough DA. Multisubunit assembly of an integral plasma membrane channel protein, gap junction connexin43, occurs after exit from the ER. Cell. 1993;74:1065–77. 96. DeVries SH, Schwartz EA. Hemi-gap-junction channels in solitary horizontal cells of the catfish retina. J Physiol. 1992;445:201–30. 97. Spray DC, Ye ZC, Ransom BR. Functional connexin ‘hemichannels’: a critical appraisal. Glia. 2006;54:758–73. 98. Quist AP, Rhee SK, Lin H, Lal R. Physiological role of gap-junctional hemichannels: extracellular calcium-dependent isomotic volume regulation. J Cell Biol. 2000;148:1063–10744. 99. Bruzzone S, Franco L, Guida L, Zocchi E, Contini P, Bisso A, Usai C, De Flora A. A selfrestricted CD38-connexin 43 crosstalk affects NAD+ and cyclic ADP-ribose metabolism and regulates intracellular calcium in 3T3 fibroblasts. J Biol Chem. 2001;276:48300–8. 100. Bruzzone S, Guida L, Zocchi E, Franco L, De Flora A. Connexin 43 hemichannels mediate Ca2+-regulated transmembrane NAD+ fluxes in intact cells. FASEB J. 2001;15:10–12. 101. Boassa D, Ambrosi C, Qiu F, Dahl G, Gaietta G, Sosinsky G. Pannexin1 channels contain a glycosylation site that targets the hexamer to the plasma membrane. J Biol Chem. 2007;282:31733–43.

10 Junction Morphology and Dynamics

261

102. Penuela S, Bhalla R, Gong X-Q, Cowan K, Celetti S, BJ C, Bai D, Shao Q, Laird D. Pannexin1 and pannexin3 are glycoproteins that exhibit many distinct characteristics from the connexin family of gap junction proteins. J Cell Sci. 2007;120:3772–83. 103. Locovei S, Bao L, Dahl G. Pannexin 1 in erythrocytes: function without a gap. Proc Natl Acad Sci USA. 2006;103:7655–9. 104. Preus D, Johnson R, Sheridan J, Meyer R. Analysis of gap junctions and formation plaques between reaggregating Novikoff hepatoma cells. J Ultrastr Res. 1981;77:263–76. 105. Martin PE, Blundell G, Ahmad S, Errington RJ, Evans WH. Multiple pathways in the trafficking and assembly of connexin 26, 32 and 43 into gap junction intercellular communication channels. J Cell Sci. 2001;114:3845–55. 106. Koval M, Harley JE, Hick E, Steinberg TH. Connexin46 is retained as monomers in a trans-Golgi compartment of osteoblastic cells. J Cell Biol. 1997;137:847–57. 107. Maza J, Mateescu M, Sarma JD, Koval M. Differential oligomerization of endoplasmic reticulum-retained connexin43/connexin32 chimeras. Cell Commun Adhes. 2003;10:319–22. 108. Lal R, John SA, Laird DW, Arnsdorf MF. Heart gap junction preparations reveal hemiplaques by atomic force microscopy. Am J Physiol. 1995;268:C968-C977. 109. Zampighi GA, Loo DD, Kreman M, Eskandari S, Wright EM. Functional and morphological correlates of connexin50 expressed in Xenopus laevis oocytes. J Gen Physiol. 1999;113:507–24. 110. Thomas MA, Huang S, Cokoja A, Riccio O, Staub O, Suter S, Chanson M. Interaction of connexins with protein partners in the control of channel turnover and gating. Biol Cell. 2002;94:445–56. 111. Xu X, Li WE, Huang GY, Meyer R, Chen T, Luo Y, Thomas MP, Radice GL, Lo CW. Modulation of mouse neural crest cell motility by N-cadherin and connexin 43 gap junctions. J Cell Biol. 2001;154:217–30. 112. Giepmans BN. Role of connexin43-interacting proteins at gap junctions. Adv Cardiol. 2006;42:41–56. 113. Giepmans BN. Gap junctions and connexin-interacting proteins. Cardiovasc Res. 2004;62:233–45. 114. Giepmans BNG, Moolenar WH. The gap junction protein connexin43 interacts with the second PDZ domains of the zona occludins-1 protein. Curr Biol. 1998;8:931–4. 115. Sosinsky GE, Solan JL, Gaietta GM, Ngan L, Lee GJ, Mackey MR, Lampe PD. The C-terminus of connexin43 adopts different conformations in the Golgi and gap junction as detected with structure-specific antibodies. Biochem J 2007 Dec 15;408:375-85. 116. Hunter AW, Barker RJ, Zhu C, Gourdie RG. Zonula occludens-1 alters connexin43 gap junction size and organization by influencing channel accretion. Mol Biol Cell. 2005;16:5686–98. 117. Toyofuku T, Akamatsu Y, Zhang H, Kuzuya T, Tada M, Hori M. c-Src regulates the interaction between connexin-43 and ZO-1 in cardiac myocytes. J Biol Chem. 2001;276:1780–8. 118. van Zeijl L, Ponsioen B, Giepmans BN, Ariaens A, Postma FR, Varnai P, Balla T, Divecha N, Jalink K, Moolenaar WH. Regulation of connexin43 gap junctional communication by phosphatidylinositol 4,5-bisphosphate. J Cell Biol. 2007;177:881–91. 119. Giepmans BN, Verlaan I, Hengeveld T, Janssen H, Calafat J, Falk MM, Moolenaar WH. Gap junction protein connexin-43 interacts directly with microtubules. Curr Biol. 2001;11:1364–8. 120. Butkevich E, Hulsmann S, Wenzel D, Shirao T, Duden R, Majoul I. Drebrin is a novel connexin-43 binding partner that links gap junctions to the submembrane cytoskeleton. Curr Biol. 2004;14:650–8. 121. Duffy HS, Delmar M, Spray DC. Formation of the gap junction nexus: binding partners for connexins. J Physiol Paris. 2002;96:243–9.

Chapter 11

Biochemistry of Connexins Joell L. Solan and Paul D. Lampe

Abstract Vertebrate gap junctions, composed of integral membrane proteins encoded by the connexin gene family, are critically important in regulation of embryonic development, coordinated contraction of excitable cells, tissue homeostasis, normal cell growth, and differentiation. Connexin proteins typically have short half-lives and detergent-specific solubilities, and interact with a variety of protein-binding partners during the assembly, trafficking, assembly/ disassembly, and degradation of the oligomeric forms that constitute hemichannels and junctional channels. Phosphorylation has been implicated in the regulation of gap junctional communication at several stages of the connexin life cycle including hemichannel oligomerization, export of the protein to the plasma membrane, hemichannel activity, gap junction assembly, gap junction channel gating, and connexin degradation. Keywords Phosphorylation Kinase Phosphatase Cell signaling Posttranslational modifications Detergent solubility Binding partners Cytoskeleton Adherens junctions Channel biogenesis Cx26 Cx31.9 Cx31 Cx32 Cx36 Cx37 Cx40 Cx43 Cx45 Cx45.6 Cx46 Cx47 Cx50 Cx56

11.1 Introduction This chapter focuses on how the biochemical properties of connexins change during their maturation into gap junctions and eventual degradation. It first summarizes what is known about connexin posttranslational modification, detergent solubility, and interaction with other proteins. It then discusses in detail the stages of the life cycle of connexin proteins and the role that these biochemical factors play at each stage, during which dynamic events in gap junction assembly and disassembly occur. P.D. Lampe (*) Public Health Sciences Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, M5C800 Box 19024, Seattle, WA 98109, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_11, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

263

264

J.L. Solan and P.D. Lampe

More is known about the biochemical properties and the posttranslational modifications of Cx43 than of any other connexin; Cx43 is the most widely expressed connexin, and the predominant connexin endogenously expressed in most cell lines. In fact, even cell lines derived from tissues that do not normally express Cx43 often express it extensively and in some cases exclusively. For these reasons, many antibodies and mutant complementary DNA (cDNA) constructs exist for Cx43. Comparison of the number of Pubmed-listed reports (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db ¼ PubMed) indicates that more reports focus on Cx43 than on the other 20 connexins combined. Therefore, in this review, the biochemistry of Cx43 is used as the primary framework for discussion, with information about other connexins included when there is specific knowledge about the topic being discussed.

11.2 Connexin Posttranslational Modifications The most well known posttranslational modification of connexins is phosphorylation. Many connexins (e.g., Cx31, Cx32, Cx36, Cx37, Cx40, Cx43, Cx45, Cx46, Cx50, and Cx56) are phosphoproteins as shown by either a phosphatase-sensitive shift in their electrophoretic mobility, direct incorporation of [32P]-phosphate, or mass spectrometry [1,2,3]. Not all connexin family members have been investigated. Connexins do not appear to be glycosylated. A few reports suggest lipid modifications of Cx32, such as prenylation and acylation [4,5]. Furthermore, mass changes consistent with amino acid acetylation and hydroxylation, and g-carboxylation of glutamic acid residues, have been reported [4]. Although several reports have indicated that Cx43 can be ubiquitinylated, the role of the proteasome in degradation of plasma membrane Cx43 is much more controversial. However, it is clear that degradation of misfolded connexins via the proteasome (i.e., endoplasmic reticulum-associated degradation, ERAD) commonly occurs [6] (see Chapter 9).

11.2.1 Phosphorylation of Connexin43 Recent investigations have shown that activation of several kinases, including protein kinase A (PKA), protein kinase C (PKC), p34cdc2/cyclin B kinase, casein kinase I (CK1), mitogen-activated protein kinase (MAPK), and pp60src kinase (src), can lead to phosphorylation of the majority of the 21 serine and two tyrosine residues in the cytoplasmic carboxyl-terminal domain (CT) of Cx43. While many studies have correlated changes in kinase activity with changes in gap junctional communication, further research is needed to directly link specific phosphorylation events with changes in connexin oligomerization into hemichannels and gap junction channel assembly.

11 Biochemistry

265

Formation and degradation of gap junctions is a dynamic process with reports of half-lives of less than 1.5 to 5.0 hours in cultured cells and in tissues [7,8,9,10,11]. Therefore, regulation of gap junction assembly and turnover is likely to be critical in the control of intercellular communication. The work of Musil and Goodenough, Lau and collaborators, and several other investigators has shown that Cx43 is differentially phosphorylated throughout its life cycle in untreated cells [8,9,11,12,13,14,15,16]. Cx43 has multiple electrophoretic isoforms when analyzed by sodium dodecyl sulfate–polyacrylamide gel (SDSPAGE), including a faster migrating form that includes nonphosphorylated (P0 or NP) Cx43, and at least two slower migrating forms, commonly termed P1 and P2. Both P1 and P2 co-migrate with P0 following alkaline phosphatase treatment, suggesting that phosphorylation is the primary covalent modification reflected in their differences in electrophoretic mobility in SDS-PAGE analysis [9,15]. ‘‘It is emphasized that the P1/P2 nomenclature does not indicate the number of phosphate residues attached to each form since the stoichiometry of phosphorylation is unknown’’ [15]. In addition, some phosphorylated species migrate with the P0 band in SDS-PAGE [17]. The CT of Cx43 appears to be the primary region that becomes phosphorylated, but Cx36 and Cx56 can also be phosphorylated within the cytoplasmic loop domain (CL) [18,19]. Cx43 does not contain serine residues in its CL, and there are no reports of phosphorylation of the amino-terminal domain (NT) of connexins.

11.3 Detergent Solubility of Connexins and Membrane Domains Biochemical analyses and subcellular localization of gap junctions have often relied on the use of detergents to solubilize specific fractions of connexins. Connexins can be solubilized from cells and tissue using ionic detergents, such as sodium dodecyl sulfate or N-lauryl sarcosine. Musil and Goodenough [16] showed that the P2 gap junctional form of Cx43 was insoluble in the nonionic detergent Triton X-100 when extracted at 48C. While the P2 form is always enriched in the Triton X-100–insoluble fraction, substantial amounts of the P0 and P1 forms can also be found in this fraction in some cell types. This protocol has been used extensively to study differences between junctional and nonjunctional pools of Cx43; other ionic and nonionic detergents have also been used to immunoprecipitate Cx43 with its interacting partners. Selective detergent solubilization has also been used to study the membrane localizations of Cx26, Cx32 [20,21], and Cx37, which, like Cx43, exhibits a phosphorylation-dependent mobility shift in SDS-PAGE of the Triton X-100–insoluble fraction [22]. The physical basis of Cx43 Triton X-100 insolubility is not entirely clear. Several reports indicate that connexins can co-isolate or interact with Triton X-100–insoluble lipid raft components including caveolins and other proteins involved in raft-associated signaling pathways. Caveolin-1 interacts and cofractionates with at least a subset of Cx43 present in a cell [23,24,25]. Cx26 and Cx32

266

J.L. Solan and P.D. Lampe

were also found in membrane fractions enriched in lipid raft components [21]. However, connexin separated into gap junctional fractions is distinct from that found associated with lipid rafts [21] (see Chapter 10). Since lipid rafts are thought to be sites of protein sorting and coordination of signal transduction pathways, they could play multiple roles in the regulation of the connexin life cycle. However, the lack of clear definition of what constitutes a lipid raft and the apparent lack of reports of raft disruption having a dramatic effect on gap junctions make a specific role for the interaction harder to pinpoint. The lipid content of gap junctions has not been well defined. Many studies have indicated that gap junctions are likely enriched in cholesterol and sphingomyelin. These include electron microscopic techniques using filipin that indicate different junctional structures might contain different levels of cholesterol [26] and several analyses of membranes enriched in gap junctions that indicate that gap junctions are enriched for cholesterol [27,28]. However, it should be noted that the latter studies depend on the assumption that the gap junction purification protocol is not selectively removing specific lipids from the gap junction (see Chapter 10).

11.4 Connexin-Interacting Proteins Many connexin-interacting proteins have been discovered. For example, Cx43 has been reported to interact with zona occludens-1 (ZO-1), ZO-2, Src, PKA, PKC, protein kinase G (PKG), MAPK, the cyclin-dependent kinase cdc2, CK1, dystrophia myotonica protein kinase (DMPK), receptor protein tyrosine phosphatase- (RPTP), -catenin, -catenin, p120-catenin, cadherin, the cysteinerich 61/connective tissue growth factor/nephroblastoma-overexpressed family of growth regulators (NOV/CNN3), caveolin-1, / -tubulin, and drebrin (for reviews see [29,30,31]). Cx32 interacts with disk-large homologue-1 (Dlgh1), calmodulin, E-cadherin, occludin, and claudin [30,31,32]. In many cases, it is not clear whether these interactions are direct or indirect. Obviously, phosphatases and kinases interact with the connexins they de/phosphorylate, but it also appears that some of these substrate pairs are relatively stable and the interacting proteins can be ‘‘pulled down’’ with the connexins. Some of these interacting proteins likely influence the connexin life cycle and are discussed in detail in the next section.

11.5 Changes in Connexin Association, Interacting Proteins, and Phosphorylation During the Connexin Life Cycle The biochemical modifications and protein-connexin interactions presented above are discussed in the rest of this chapter in terms of their functional importance and interrelationships during the connexin life cycle (see Chapter 9).

11 Biochemistry

267

11.5.1 Intracellular Trafficking Properly folded and formed connexin hexamers are assembled within cytoplasmic membrane compartments. Cells contain quality control mechanisms that allow transit to the plasma membrane of properly folded, oligomerized proteins. Studies examining Human disease-associated CX32 mutants [33] or Cx43 from cells that are unable to assemble junctions [16,34] show that connexin can be targeted for degradation early in the secretory pathway in response to quality control checkpoints [6]. While many studies have focused on the amino acids within the transmembrane regions during oligomerization and their effect on quality control, the CT of Cx43 also appears to be involved in quality control in the Golgi apparatus (Table 11.1). Quality control pathways often rely on chaperone proteins to recognize conformational information in target proteins. Recent data indicate that the exit of Cx43 from the Golgi apparatus involves a quality control step that includes a conformational change upon phosphorylation near the carboxyl terminus. These studies rely on a monoclonal antibody, termed CT1 that recognizes Cx43 when it is in the Golgi and not when it is in the plasma membrane [35]. The epitope for this antibody requires that neither S364 nor S365 be phosphorylated or mutated to alanine or aspartic acid. Consistent with the data showing that intracellular Cx43 mostly migrates in the P0 position, the CT1 antibody recognizes almost exclusively the P0 isoform in resting cells. An exception to this is in cells treated with the PKC-activator phorbol 12-myristate 13-acetate (PMA), where CT1 recognizes several migratory isoforms except P1, that is, P0 and multiple bands around P2. PMA treatment also results in reduced conductance of gap junction channels [36], including inhibition of gap junction assembly from intracellular pools of Cx43 [7], consistent with the CT1 epitope being a Golgi-bound conformation. Additionally, the CT1 antibody does not recognize the P1 isoform, indicating that phosphorylation on S364 or S365 leads to formation of the P1 isoform, which is generally found in the plasma membrane [16]. Taken together, these data are consistent with the idea that phosphorylation at S364 or S365 mediates a conformational change that leads to the P1 isoform and to exit from the Golgi compartment. Although no clear chaperone protein has been shown to mediate Cx43 trafficking, it is easy to imagine a Golgi-resident chaperone that, like the CT1 antibody, can distinguish the conformational status of Cx43.

11.5.2 Trafficking to the Plasma Membrane 11.5.2.1 Cyclic Adenosine Monophosphate Elevated levels of cyclic adenosine monophosphate (cAMP) have been shown to upregulate gap junctional intercellular communication (GJIC) in cells expressing Cx43 [37,38], Cx40 [39,40] or Cx32 [41], but to downregulate Cx36

Endocytosis/degradation

Trafficking to plasma membrane/enhanced assembly Channel selectivity Channel gating

Heterotypic compatibility

Heteromeric compatibility

S279/S282 Y247/Y265 P274–P284 SH3-binding domain CT Cx32 residues 216–231 Cx36 residues 276–283 Cx50 CL, CT P253–P256 SH3-binding domain Y286–V289 tyrosine-based sorting motif

S368 CL, H142 S255

NT: -group residues 12–13, -group residues 11–12, M3/CL residues 152–153 (Cx43) E1, E2 intra-monomer disulfide bonds S364, S365, S368, S369, S373

src cAMP, PKA cAMP, PKA cAMP, PKA

ERK1/2, BMK, p34cdc2/ cyclin B MAPK src src

PKC, mitosis

cAMP, PKA, Epac

-group charged polar, group two uncharged -group LMRGN -group WW domain

CIP85

rPTP Calmodulin

c-src, v-src v-src

CT

[103]

[123] [86,124] [88] [87] [105]

[97] [90,91,92,122] [122]

[36] [121] [100]

[45,46,55]

[117,118,119,120]

[116]

Table 11.1 Connexin signaling and interacting domains: summary of the different regulatory domains of connexin, their function, involvement with signaling pathways, and sites for interaction of other proteins Regulatory residues or Function connexins1 Interacting pathways Interacting proteins Reference

268 J.L. Solan and P.D. Lampe

Interacting pathways

Interacting proteins

Caveolin 1, PKCg

F-actin, cortactin NOV/CNN3 Dlgh1 ZONAB

S262 S255, S262, S368 CL residues 130–136 CT Cx32 CL, CT Cx43, Cx32, Cx47

All residues refer to Cx43, unless noted.

1

Myoblast differentiation Growth control

Lipid raft association, Cx43, Cx32, Cx36, Cx46 (not Cx26, Cx50) Cell cycle regulation PKC, S-phase entry PKC, p34cdc2/cyclin B

Ubiquitin ligase

[81] [17,107,108] [112] [113] [32] [115]

[111] [23,24]

[76] [68] [104] [109]

Drebrin / -tubulin Nedd4

[70,71] [72,125,126]

K234–K243 S282–Y286, PY-motif Cx45.6 residues 356–380, S363; PEST domain Cx45 S381, S382 Cx43

Reference [79] [62,63,69]

ZO-1, ZO-2

Occludin, Claudin

Casein kinase 1 N-cadherin, -catenin, catenin, ZO-1, E-cadherin,

Microtubule association Gap junction turnover

Casein kinase 1 Cadherin cross-linking, adherens junction formation Tight junction formation, maintenance c-src, cell cycle

L380-I382 PDZ-binding domain

Cx32

S325–S330

Regulatory residues or connexins1

Actin association

Gap junction formation/ maintenance

Function

Table 11.1 (continued)

11 Biochemistry 269

270

J.L. Solan and P.D. Lampe

and Cx45 channels [42]. In the case of Cx43, these effects have been examined extensively under various conditions/cell types and ascribed to changes in connexin phosphorylation, localization, protein synthesis, transcriptional activation via the cAMP response element-binding protein, and cellular differentiation. For the purposes of this review, only acute effects, on the order of minutes to hours, are addressed, as these more likely reflect biochemical changes directly involving the connexin protein. Upregulation of GJIC in Cx43-expressing cells has been shown to be the result of increased trafficking from intracellular locations to plasma membrane [37,38,43]. In the case of other connexins, only a functional difference in GJIC has been demonstrated, so it is not clear whether these effects are on trafficking or gating. Connexin43 Cells expressing Cx43 respond to elevated cAMP with both an increase in GJIC and an increase in gap junction plaque size [37,38], a process termed enhanced assembly [43]. The process is sensitive to blockage of anterograde trafficking via brefeldin A or monensin [38,43], to microtubule inhibition via nocodazole [44], and to PKA inhibition via H89 [43]. These trafficking studies have shown that the cAMP-mediated increase in GJIC is a result of increased transport of Cx43 from intracellular locations to plaques. Phosphorylation at S364 has been implicated in regulating this process, as fibroblasts expressing the Cx43 mutants S364A or S364P did not respond to increased cAMP [45]. Interestingly, S364E mutants, wherein the glutamate is intended to mimic serine phosphorylation, did respond to cAMP, indicating that a negative charge at position 364 may be a prerequisite for enhanced assembly. Additionally, S364 was phosphorylated in cells. S365, S368, S369, and S373 are phosphorylated in rat primary granulosa cells in response to the follicle-stimulating hormone [46,47]. This process could be inhibited by H89, indicating that PKA is involved in this event. Although most evidence indicates that Cx43 is a poor substrate for PKA [46,48,49], purified PKA was able to phosphorylate wild-type Cx43 in vitro but not a construct in which S365, S368, S369, and S373 had been mutated to alanine [47]. Given the apparent role of S364 and S365 in both Golgi exit and cAMPmediated mobilization of Cx43 from intracellular locations, it is interesting to speculate that these processes are linked. cAMP has also been shown to activate a PKA-independent signaling pathway involving Epac (exchange protein directly activated by cAMP) that leads to activation of Rap, a small molecular weight guanosine triphosphatase (GTPase) of the Ras family [50,51]. Activation of this pathway has been linked to both integrin-based [52] and cadherin-based [53,54] cell adhesion, as well as mediating cAMP-induced secretion. Studies using cardiac myocytes showed that cAMP-mediated increases in GJIC and Cx43 accumulation in plaques are synergistically regulated by PKA and Epac activation [55]. Using chemical activators specific for PKA or Epac, it was shown that activation of PKA alone could moderately increase GJIC but did not result in accumulation of

11 Biochemistry

271

Cx43 in gap junctions, while Epac activation alone increased gap junctions with only a slight elevation in GJIC. Thus it appears that the cAMP stimulation of GJIC involves both PKA-mediated and Epac-mediated pathways. Interestingly, Epac activation also led to accumulation of adherens junctions; the relationship between adherens and gap junctions is discussed below. Connexin40 The evidence for positive regulation of Cx40 via cAMP comes from two types of experiments. In one type, Cx40 was transfected into SKHep1 human hepatoma cells, which do not endogenously express connexins. These cells responded to 8-BrcAMP, a cell-permeable cAMP analogue that is resistant to phosphodiesterases, with increased GJIC and mobility shift from 40 to 42 kDa when examined by SDS-PAGE [56]. In the other type, endothelial cells derived from a Cx40 knockout (KO) mouse were compared with wild-type endothelial cells. The latter responded to lipopolysaccharide [39] or hypoxia/reoxygenation [40] with reduced coupling. This reduction could be inhibited by treatment with 8-Br-cAMP. However, the Cx40KO endothelial cells did not reduce coupling in response to these insults, arguing that Cx40 channels were specifically targeted by these treatments. Note that Cx40 contains two serines within a PKA consensus sequence, S120 and S345, though the role of phosphorylation at these or other sites has not been determined. Connexin45 HeLa cells expressing Cx45 show 46 and 48 kDa bands by Western blot, and like Cx43, alkaline phosphatase treatment prior to SDS-PAGE collapses the slower migrating 48 kDa band to the faster migrating 46 kDa position. Unlike Cx43, treatment of these cells with 8-Br-cAMP results in reduced GJIC as well as an increase in the slower migrating phosphoisoform [42]. Connexin36 Cx36 is expressed in neural and retinal cells. In amacrine cells of the retina, cAMP activation results in a decrease in GJIC as well as an increase in Cx36 phosphorylation [57] (see Chapter 19). Phosphorylation of Cx36 fusion proteins consisting of the CL or the CT with purified PKA resulted in phosphorylation at S110 of the CL and S293 of the CT [18,58].

11.5.3 Docking and Assembly 11.5.3.1 Innate Regulation To form a junctional channel, a hemichannel must dock with a hemichannel in a neighboring cell. The ability of hemichannels composed of different connexins to

272

J.L. Solan and P.D. Lampe

dock with one another and form functional channels, termed heterotypic compatibility, resides in the second extracellular loop, between the third and fourth transmembrane domains, which varies across the connexins (see Chapter 2). What triggers the docking step is not entirely clear, but formation of gap junctions has been intimately associated with the ability of cells to make adherens junctions [15,59,60,61].

11.5.3.2 Adherens Junctions, Tight Junctions and the Cytoskeleton Adherens junctions provide strong mechanical attachments between cells, mediated by cadherins. Cadherins contain an extracellular domain that can homodimerize, thereby bringing cells into close contact, and are anchored to the actin cytoskeleton by interaction of their intracellular domains with catenins. Tight junctions perform a barrier function in epithelial cells, creating a tight seal between cells, and also define and restrict movement of integral membrane proteins between the apical and basolateral surfaces. The relationship among adherens junctions, tight junctions, and connexins has been explored in many cell types with generally consistent, but occasionally conflicting, results. The molecular basis of this relationship is still being investigated. In many, although not all, cases, formation of adherens junctions has been shown to be necessary for and to precede gap junction formation [62,63]. This association is generally thought to reflect the physical need for membranes to come into very close proximity for hemichannels in apposing cells to dock. In addition, adherens junction proteins are likely involved in signaling pathways that promote gap junction assembly. This is consistent with the fact that the ability of cadherins to enhance gap junction formation can be dependent on cell type. For example, in cells generated from an Ncadherin knockout mouse, neural crest cells demonstrated migration defects but no obvious change in Cx43 expression [61], while in cardiomyocytes, the ability to form gap junctions was greatly diminished [64]. There are also several examples of a reciprocal relationship where disruption of gap junctions can disrupt adherens junctions [60,65,66], but again this varies across cell systems. Consistent with much of the data is a model in which adherens junction formation or cross-linking of cadherins activates a signaling pathway that promotes delivery of connexin-containing vesicles to the plasma membrane, so that gap junctions form in the vicinity of adherens junctions. This is supported by time-lapse studies in cardiomyocytes, where a Ca2+ switch was used to induce junction formation. In this study, accumulation of the adherens junctional components, -catenin, -catenin, and ZO-1, occurred at the plasma membrane before Cx43 accumulation at these same sites [63]. Another time-lapse study utilizing live cell imaging and total internal reflection fluorescence microscopy on transiently transfected HeLa cells showed increased delivery of Cx43-containing vesicles to the plasma membrane in cells in which N-cadherin was cross-linked [67].

11 Biochemistry

273

This study also showed evidence that both gap and adherens junctions may provide tethering sites for plus-end microtubules carrying Cx43 cargo, allowing for directed delivery of Cx43 to sites of cell–cell contact. While this study focused on Cx43 as cargo on microtubules, a plus-end microtubule tethering function for gap junctions has been suggested previously [68] via direct binding of tubulin to a motif present at residues 234 to 243 (Table 11.1). Note that in this study, which used Rat1 fibroblasts, microtubules did not play a critical role in creating or maintaining plaques. Other studies using biochemical techniques and immunofluorescence microscopy show that the interaction between adherens junction components and Cx43 occurs intracellularly and that catenins, cadherins, connexins, and ZO-1 form complexes in the secretory pathway that are delivered to the plasma membrane together [63,69]. Interestingly, one of these studies [69] also showed that N-cadherin was quickly internalized after plasma membrane arrival, leading to a dramatic decrease in adherens junctions, whereas gap junctions were not affected, arguing that the important interactions are occurring before connexins reach the plasma membrane. Gap junctions, especially those containing Cx32, have also been convincingly shown to be associated and intermingled in tight junction strands via freeze-fracture electron microscopy [62,70] and can be co-immunoprecipitated with several tight junction proteins, including occludin and claudins [70,71]. Thus, there are substantial and convincing data showing that adherens and tight junction components interact with and affect the connexin life cycle both at the plasma membrane and in secretory compartments. Interestingly, there are common proteins found in cell–cell junctions, the scaffolding/ MAGUK (membrane-associated guanylate kinase) proteins ZO-1 and ZO-2 (Table 11.1). ZO-1, which has been shown to be a bona fide binding partner of Cx43 in many studies, interacts with all connexins that contain a PDZ binding motif in their CT, including Cx43, Cx45, Cx46, Cx47, Cx50, Cx36, and Cx31.9 [29]. Note that both studies that reported intracellular complexes between Cx43 and adherens junction components also identified ZO-1 in these complexes [63,69]. As a scaffolding protein, ZO-1 provides a linkage between the different cell–cell junctions, junctional components, and the cytoskeleton. Additionally, ZO-1 interaction with Cx43 is cell cycle regulated. In NRK cells, G0-phase cells show increased colocalization and interaction of ZO-1 and Cx43 by immunofluorescence and Far Western analysis [72]. This correlates with the report that G0 is the phase in which NRK cells are most efficient at gap junction assembly [17]. Surprisingly, in spite of being the best described connexin interacting protein, the functional significance of the interaction of connexins with ZO-1 has remained elusive. While the data presented in this section indicate that ZO-1 and the connexin PDZ binding domain may be important in the docking or assembly of gap junctions, many have speculated that ZO-1 plays a role in plaque maintenance or removal.

274

J.L. Solan and P.D. Lampe

11.5.4 Plaque Formation and Maintenance 11.5.4.1 Zona Occludens-1 The tight junction protein, ZO-1, is a scaffold protein containing multiple PDZ domains, a common structural domain of 80 to 90 amino acids found in the signaling proteins of bacteria, yeast, plants, and animals. A region near the carboxyl terminus of Cx43 has been shown to interact with the second PDZ domain of ZO-1. Microscopy-based analyses examining colocalization of ZO-1 and Cx43 in the plasma membrane indicate that ZO-1 is found predominantly at the outer edges of gap junction plaques [73,74,75]. Furthermore, when interaction between these two proteins is inhibited, by various methods, plaque size is increased [73,75]. Hunter et al. [73] proposed that ZO-1 interaction at the periphery of the plaque limits the rate of free hemichannel addition resulting in smaller plaques. Consistent with this idea are data indicating that cells in G0 phase of the cell cycle have increased interaction with ZO-1 and smaller gap junctions than S phase cells [72]. However, it is important to note that these studies manipulated the ZO-1-Cx43 interaction by interfering with the Cx43 PDZ binding domain, which could be interacting with other PDZ domaincontaining proteins in addition to ZO-1. In fact, it has been shown that the related ZO-2 protein can bind to the Cx43 CT via its PDZ domain [72]. The mechanism by which ZO-1, ZO-2 or other PDZ domain containing proteins can regulate gap junction size is not yet clear, but it seems reasonable to consider that they do, in light of the relationship between adherens junctions and gap junction formation discussed above. 11.5.4.2 Drebrin Drebrin, an actin-binding protein found in brain, has been shown to interact with Cx43 in astrocytes and Vero kidney epithelial cells [76]. Depletion of drebrin by interfering antisense oligodeoxyribonucleotides (RNAi) led to a dramatic decrease in gap junctions via endocytic degradation. Since drebrin was present in association with Cx43 exclusively at the plasma membrane and can bind actin, it was suggested that drebrin could link plaques to the submembrane cytoskeleton and thus stabilize them (Table 11.1). It has also been shown that immunoprecipitation of Cx43 from astrocytes can pull down -actin and that this interaction could be disrupted by chemical ischemia, a treatment that also resulted in apparent dephosphorylation [77] and internalization [78] of Cx43. Whether this association with actin is mediated by drebrin has not been shown. 11.5.4.3 Phosphorylation by Casein Kinase I In the case of Cx43, movement of hemichannels from the plasma membrane to the gap junction plaque has been shown to involve CK1 activity [79]. Inhibition of CK1 leads to a marked increase in nonjunctional Cx43 and a decrease in

11 Biochemistry

275

Triton X-100–insoluble gap junctions. CK1 consensus sites are found around S325/S328/S330 (Table 11.1), and CK1 has been shown to phosphorylate Cx43 at some combination of these residues, although which residues are critical is not entirely clear. Expression of site-directed mutants, in which these serines are converted to alanines, results in cells that have little to no ability to form gap junctions [25]. In addition, S325/S328/S330 have been shown to be important in vivo, as use of an antibody specific for S325/S328/S330 phosphorylation has shown that they are phosphorylated in gap junctions in the intercalated disk region of the heart, and that this phosphorylation is lost upon ischemia, when Cx43 localization at the intercalated disk is also lost [80]. 11.5.4.4 Phosphorylation by Protein Kinase C Treatment with PMA has been shown to dramatically decrease gap junction assembly. Time course, pulse-chase, and cell surface biotinylation experiments indicate that PMA acts by destabilizing newly forming gap junctions while not affecting the ones already present [7]. However, PMA effects on assembly have not as yet been linked to phosphorylation at a specific site in Cx43. The kinetics of PMA action can be complex since PMA can affect channel gating, gap junction assembly, and connexin half-life [7,36]. In addition, PKC becomes downregulated over time in the presence of PMA, thereby reversing many of these effects. Thus, reports that show differences in the effects of PMA or PKC activation in disparate or even the same cell types need to be compared with caution. Recent reports have begun to dissect the molecular basis of the regulation of Cx43 by PKC by defining which specific residues can be phosphorylated by PKC, which residues show increased phosphorylation in response to PMA treatment, and which specific PKC isozymes might be involved in this process (Table 11.1). Phosphorylation of S262 and S368 is increased in response to PMA. The former has been linked to increased cellular proliferation through an unknown mechanism [81], while the latter has been shown to underlie the PMA-induced reduction in intercellular communication and alteration of single-channel behavior, discussed below [82]. However, these or other potential PKC sites might be involved in regulating assembly or degradation of gap junctions and channels.

11.5.5 Gating The following review of gating is limited to changes in gating due to interaction with other proteins. 11.5.5.1 Protein Kinase C Following treatment with PMA, Cx43 phosphorylation is increased and gap junctional communication is decreased in a number of different cell types

276

J.L. Solan and P.D. Lampe

[7,12,13,83,84]. Phosphorylation of Cx43 on S368 increases upon treatment with PMA, increases through the cell cycle, and results in a reduction in unitary channel conductance (50 picoSiemens [pS] channels are favored over 100 pS channels) and a decrease in permeability to larger cations [17,36,82].

11.5.5.2 Calmodulin Changes in intracellular Ca2+ have been shown to be involved in chemical gating of both gap junctions and hemichannels. Cx32 [85,86], Cx50 [87], and Cx36 [88] have been reported to bind to and be regulated by calmodulin, an important Ca2+-binding protein. A fluorescent derivative of calmodulin was reported to bind to two peptides derived from Cx32, corresponding to the first 21 amino acids and residues 216 to 231 of the CT [86]. Cx32 and calmodulin colocalization was shown in HeLa cells transfected with wild-type Cx32 or a Cx32-GFP (green fluorescent protein) fusion protein [85]. Studies utilizing surface plasmon resonance and peptide competition experiments showed that calmodulin binds Cx36 at the juxtamembrane region, in a motif containing residues 276 to 283 [88]. A study utilizing a combination of confocal microscopy and immunoprecipitation of Cx50 fusion proteins including a truncation mutant lacking a CT, the CT alone, and the CL alone showed that calmodulin binding required an intramolecular interaction between the two regions [87].

11.5.5.3 Src The viral-src oncogene (v-src) [89] and the cellular-src proto-oncogene (c-src) [90,91] have been shown to interact with Cx43 and mediate phosphorylation of S247 and S265. These interactions occur via binding of the src SH3 and SH2 domains (Table 11.1). Pull-down experiments using lysates from cells expressing v-src or Cx43 mutants indicate that the proline-rich segment P253LSP256 of Cx43 is critical for SH3 binding, which promotes phosphorylation on S265, creating a potential SH2 binding site that allows phosphorylation on S247, culminating in channel closure [92]. Src binding to Cx43 has also been shown to inhibit the ability of ZO-1 and Cx43 to interact [90,91]. Using nuclear magnetic resonance (NMR) to examine the structure of the Cx43 CT in the presence of the interacting ZO-1 PDZ domain and Src SH3 domain, it was shown that Src binding to Cx43 could displace ZO-1, though the binding sites for these domains are quite distant [93]. Consistent with this, in cardiac myocytes, expression of constitutively active c-src inhibited the interaction between ZO-1 and wild-type Cx43 but not the Cx43 mutant Y265F [91]. Additionally, Cx43 from cultured astrocytes exposed to chemical ischemia had an increased association with c-src, extracellular signal-regulated kinase 1/2 (ERK1/2), and MAPK phosphatase-1 [77], and a decreased association with ZO-1 [94].

11 Biochemistry

277

11.5.5.4 Mitogen-Activated Protein Kinase Growth factors rapidly downregulate GJIC in various cell systems [1]. While dissecting the mechanisms of this downregulation is complicated by pleiotropic effects of growth factors on signaling, experiments using kinase inhibitors and Cx43 site-directed mutants indicate that at least part of this effect is a result of channel closure [95], although MAPK phosphorylation may also be involved in channel closure in at least some cell types [96]. Inhibitor data indicate that reduced GJIC is at least partially mediated via MAPK activation, although which MAPK form is the main actor appears to vary depending on cell type and conditions. ERK1/2 [97], p38 [98], c-Jun N-terminal kinase (JNK) [99], and Big MAPK1/ERK5 (BMK1) [100] have all been shown to downregulate GJIC in different systems (Table 11.1). Studies utilizing serine to alanine mutants of Cx43 have implicated S279, S282, and S255 as MAPK substrates [97], while BMK1 appeared to only target S255 [100]. Examples of p38-mediated inhibition of GJIC via Cx40 [39] and Cx32 [101] have also been described. However, it is not clear that all of these effects on GJIC can be attributed to effects on channel gating per se.

11.5.6 Internalization/Degradation 11.5.6.1 Proteasome and Lysosome As integral membrane proteins with multiple membrane spanning domains, it is expected that connexins are removed from the membrane via endocytosis and then degraded in the lysosome. In fact, studies have shown Cx43 preservation upon inhibition of lysosomal degradation. However, proteasomal inhibitors also stabilize some forms of Cx43. Generally, proteasomal inhibition appears to preferentially stabilize Cx43 found in plaques and migrating as the P2 form [102]. As stated at the beginning of this chapter, how the proteasome affects Cx43 degradation from the plasma membrane is not clear. What seems to be most consistent with the data is that this effect is indirect, with a separate, proteasome-sensitive protein affecting Cx43 residence in the plasma membrane. However, no such candidate proteins have been identified. In any case, the proteasome is clearly involved in ERAD degradation of misfolded Cx43 [6] (see Chapter 9).

11.5.6.2 PY-Motif and Tyrosine-Based Sorting Signals Residues 282 to 289 on Cx43 contain an overlapping MAPK target (S282), a proline-rich PY-motif protein interaction domain (XPPXY) and a tyrosinebased sorting signal (Yxxf, where f is hydrophobic). One study utilizing Cx43 mutants indicated that elimination of the tyrosine-based signal tripled the Cx43 half-life [103]. In another study, the PY-motif was shown to bind WW

278

J.L. Solan and P.D. Lampe

domains of the ubiquitin ligase Nedd4, and that this binding may be modulated by phosphorylation at S279 and S282 [104]. The WW domains are protein modules that mediate protein–protein interactions through recognition of proline-rich motifs and phosphorylated serine/threonine-proline sites. 11.5.6.3 Connexin43-Interacting Protein of 85 kDa A novel protein, CIP85, utilizes a SH3 domain to interact with the Cx43 prolinerich segment P253LSP256 (Table 11.1). This protein colocalized with Cx43 at gap junction plaques in HeLa cells. Use of a CIP85 mutant, which could not bind Cx43, indicated that CIP85 could mediate lysosomal degradation of Cx43 [105]. 11.5.6.4 Mitosis During mitosis, GJIC is shut down [106] and Cx43 is mostly found intracellularly, in what appear to be clusters of vesicles [106,107]. Whether these vesicles contain internalized or newly synthesized Cx43 has not been clearly demonstrated. In addition, a unique migratory phosphoform is produced, which migrates even slower than P2 by SDS-PAGE and is termed Pm or P3 [107,108]. Formation of this isoform is dependent on p34cdc2/cyclin B kinase activity and includes phosphorylation on S255 [107] and possibly S262 [108]. In addition, phosphorylation at S368 is increased during mitosis [17]. How these sites or others are involved in regulating the intracellular localization of Cx43 during mitosis is not presently clear. 11.5.6.5 Casein Kinase II and Chick Connexin45.6 Chick Cx45.6, whose mammalian ortholog is Cx50, is expressed in the lens and appears to be involved in lens development. Part of the developmental program appears to involve truncation of Cx45.6 by caspase3. This truncation can be inhibited when Cx45.6 is phosphorylated at S363 by CK2 [109]. Additionally, cells expressing mutant Cx45.6 containing a S363A substitution have a longer half-life than the wild-type version. S363 lies within a potential PEST domain, a region rich in proline-glutamic acid-serine-threonine, comprising residues 356 to 380. PEST domains are typically found in proteins that undergo proteasomal degradation, indicating that Cx45.6 phosphorylated on S363 is targeted by the proteasome [110]. 11.5.6.6 Serine Phosphorylation of Connexin45 Cx45 expressed in HeLa cells is phosphorylated on serine residues. When S381/ S382 or S384/S385 are mutated to other amino acids, the half-life of Cx45 decreased from 4.2 to 2.3 and 2.6 hours, respectively. Immunoprecipitation of wild-type or mutant Cx45 from [32P]-phosphate–labeled cells indicated that S381 and S382 can be targets of phosphorylation, suggesting that these sites may play a role in the Cx45 life cycle [111].

11 Biochemistry

279

11.6 Regulation of Growth and Differentiation There has been long-standing interest in the observation that GJIC and connexin expression can regulate cell growth and differentiation. For example, in cardiac myocytes, overexpression of wild-type Cx43 or a S262A mutant resulted in decreased DNA synthesis, and this effect did not require cell–cell coupling. A S262D mutant did not inhibit DNA synthesis, indicating that this PKC-sensitive site may play a role in cell cycle progression [81]. While the exact mechanism by which this occurs is not clear, there are examples in which specific mutants or interactions with specific proteins modulate growth control. A common theme that appears to be emerging is a role for connexins in sequestering potential signaling proteins to the plasma membrane as detailed below.

11.6.1 Myoblast Differentiation In a cell culture model, cellular differentiation induced by sphingosine 1-phosphate was shown to depend on Cx43 expression and involve p38-MAPK–dependent interaction with F-actin and cortactin. A Cx43 deletion construct, in which residues 130 to 136 in the CL were eliminated, did not mediate this effect nor did it interact with F-actin and cortactin [112]. However, in this case, the deletion did reduce GJIC as well.

11.6.2 Nephroblastoma-Overexpressed Family of Growth Regulators In a cell culture model for the early placental trophoblast (the malignant trophoblast cell line, Jeg3), Cx43 but not Cx40 or a CT truncated Cx43 construct was able to reduce cell growth in culture and tumor growth in nude mice. Expression of Cx43 was accompanied by increased expression of and interaction with NOV/CNN3, a protein that has growth suppressive effects. Cx43 expression also resulted in a change in NOV/CNN3 localization, from the nucleus to the plasma membrane where it colocalized with Cx43 [113].

11.6.3 Disk Large Homolog-1 and Connexin32 Using cells with Cx32 under the control of a tetracycline-off promoter, it was shown that expression of Cx32 led to increased expression of and interaction with Dlgh1. This was accompanied by translocation of Dlgh1 from the nucleus to the plasma membrane where it colocalized with Cx32. Interestingly, Dlgh1 is a classic PDZ domain-containing scaffold protein, but Cx32 does not contain a CT PDZ

280

J.L. Solan and P.D. Lampe

binding domain. Instead, yeast-two-hybrid assays showed that the SH3/Hook domain of Dlgh1 could interact with both the CL and CT of Cx32 [32].

11.6.4 ZO-1-Associated Nucleic Acid-Binding Protein ZO-1 interacts with a transcription factor ZONAB that can regulate cell proliferation [114]. Immunofluorescence analysis of mouse brain and spinal cord showed that both ZONAB and ZO-1 colocalized at gap junctions containing Cx47 and Cx32 in oligodendrocytes as well as Cx43 gap junctions in astrocytes. Similarly, immunofluorescence and freeze-fracture immunogold labeling showed colocalization among Cx36, ZO-1, ZO-2, and ZONAB in specific substructures in the retina [115].

11.7 Conclusion Connexins are highly regulated proteins. Even though they have atypically short half-lives, their export to the plasma membrane, assembly into gap junctions, gating, and degradation all appear to be regulated via posttranslational modification and interaction with other cellular proteins. Furthermore, as they proceed through their life cycle, interaction with lipid components, stability in the plasma membrane, detergent solubility, and other biochemical properties change. The complexity of these interactions and their full import are not fully understood, but the groundwork that has already been laid indicates that the future will hold many more exciting discoveries. Acknowledgments The work performed in the authors’ lab reviewed here was supported by National Institutes of Health (NIH) grant GM55632.

References 1. Lampe PD, Lau AF. The effects of connexin phosphorylation on gap junctional communication. Int J Biochem Cell Biol. 2004;36:1171–86. 2. Sa´ez JC, Martı´ nez AD, Bran˜es MC, Gonzalez HE. Regulation of gap junctions by protein phosphorylation. Braz J Med Biol Res. 1998;31:593–600. 3. Solan JL, Lampe PD. Connexin phosphorylation as a regulatory event linked to gap junction channel assembly. Biochim Biophys Acta. 2005;1711:154–63. 4. Locke D, Koreen IV, Harris AL. Isoelectric points and posttranslational modifications of connexin26 and connexin32. FASEB J. 2006;20:1221–3. 5. Huang Y, Sirkowski EE, Stickney JT, Scherer SS. Prenylation-defective human connexin32 mutants are normally localized and function equivalently to wildtype connexin32 in myelinating Schwann cells. J Neurosci. 2005;25:7111–20. 6. VanSlyke JK, Musil LS. Dislocation and degradation from the ER are regulated by cytosolic stress. J Cell Biol. 2002;157:381–94.

11 Biochemistry

281

7. Lampe PD. Analyzing phorbol ester effects on gap junction communication: a dramatic inhibition of assembly. J Cell Biol. 1994;127:1895–905. 8. Laird DW, Puranam KL, Revel JP. Turnover and phosphorylation dynamics of connexin43 gap junction protein in cultured cardiac myocytes. Biochem J. 1991;273:67–72. 9. Crow DS, Beyer EC, Paul DL, Kobe SS, Lau AF. Phosphorylation of connexin43 gap junction protein in uninfected and Rous sarcoma virus-transformed mammalian fibroblasts. Mol Cell Biol. 1990;10:1754–63. 10. Beardslee M, Laing J, Beyer E, Saffitz J. Rapid turnover of connexin43 in the adult rat heart. Circ Res. 1998;83:629–35. 11. Musil LS, Beyer EC, Goodenough DA. Expression of the gap junction protein connexin43 in embryonic chick lens: molecular cloning, ultrastructural localization, and posttranslational phosphorylation. J Membr Biol. 1990;116:163–75. 12. Berthoud VM, Ledbetter MLS, Hertzberg EL, Sa´ez JC. Connexin43 in MDCK cells: regulation by a tumor-promoting phorbol ester and calcium. Eur J Cell Biol. 1992;57:40–50. 13. Brissette JL, Kumar NM, Gilula NB, Dotto GP. The tumor promotor 12-O-tetradecanoylphorbol-13-acetate and the ras oncogene modulate expression and phosphorylation of gap junction proteins. Mol Cell Biol. 1991;11:5364–71. 14. Kadle R, Zhang JT, Nicholson BJ. Tissue-specific distribution of differentially phosphorylated forms of Cx43. Mol Cell Biol. 1991;11:363–9. 15. Musil LS, Cunningham BA, Edelman GM, Goodenough DA. Differential phosphorylation of gap junction protein connexin43 in junctional communication-competent and deficient cell lines. J Cell Biol. 1990;111:2077–88. 16. Musil LS, Goodenough DA. Biochemical analysis of connexin43 intracellular transport, phosphorylation and assembly into gap junctional plaques. J Cell Biol. 1991;115:1357–74. 17. Solan JL, Fry MD, TenBroek EM, Lampe PD. Connexin43 phosphorylation at S368 is acute during S and G2/M and in response to protein kinase C activation. J Cell Sci. 2003;116:2203–11. 18. Urschel S, Hoher T, Schubert T, Alev C, Sohl ¨ G, Worsdorfer P, Asahara T, Dermietzel R, Weiler R, Willecke K. Protein kinase A-mediated phosphorylation of connexin36 in mouse retina results in decreased gap junctional communication between AII amacrine cells. J Biol Chem. 2006;281:33163–71. 19. Berthoud VM, Beyer EC, Kurata WE, Lau AF, Lampe PD. The gap junction protein connexin 56 is phosphorylated in the intracellular loop and the carboxy-terminal region. Eur J Biochem. 1997;244:89–97. 20. VanSlyke JK, Musil LS. Analysis of connexin intracellular transport and assembly. Methods. 2000;20:156–64. 21. Locke D, Liu J, Harris AL. Lipid rafts prepared by different methods contain different connexin channels, but gap junctions are not lipid rafts. Biochemistry. 2005;44:13027–42. 22. Larson DM, Seul KH, Berthoud VM, Lau AF, Sagar GD, Beyer EC. Functional expression and biochemical characterization of an epitope-tagged connexin37. Mol Cell Biol Res Commun. 2000;3:115–21. 23. Lin D, Zhou J, Zelenka PS, Takemoto DJ. Protein kinase Cg regulation of gap junction activity through caveolin-1-containing lipid rafts. Invest Ophthalmol Vis Sci. 2003;44:5259–68. 24. Schubert AL, Schubert W, Spray DC, Lisanti MP. Connexin family members target to lipid raft domains and interact with caveolin-1. Biochemistry. 2002;41:5754–64. 25. Barth K, Gentsch M, Blasche R, Pfuller A, Parshyna I, Koslowski R, Barth G, Kasper M. Distribution of caveolin-1 and connexin43 in normal and injured alveolar epithelial R3/1 cells. Histochem Cell Biol. 2005;123:239–47. 26. Biswas SK, Lo WK. Gap junctions contain different amounts of cholesterol which undergo unique sequestering processes during fiber cell differentiation in the embryonic chicken lens. Mol Vis. 2007;13:345–59. 27. Alcala J, Katar M, Maisel H. Lipid composition of chick lens fiber cell gap junctions. Curr Eye Res. 1982;2:569–78.

282

J.L. Solan and P.D. Lampe

28. Henderson D, Eibl H, Weber K. Structure and biochemistry of mouse hepatic gap junctions. J Mol Biol. 1979;132:193–218. 29. Giepmans BN. Gap junctions and connexin-interacting proteins. Cardiovasc Res. 2004;62:233–45. 30. Giepmans BN. Role of connexin43-interacting proteins at gap junctions. Adv Cardiol. 2006;42:41–56. 31. Herve JC, Bourmeyster N, Sarrouilhe D. Diversity in protein-protein interactions of connexins: emerging roles. Biochim Biophys Acta. 2004;1662:22–41. 32. Duffy HS, Iacobas I, Hotchkiss K, Hirst-Jensen BJ, Bosco A, Dandachi N, Dermietzel R, Sorgen PL, Spray DC. The gap junction protein connexin32 interacts with the Src homology 3/Hook domain of discs large homolog 1. J Biol Chem. 2007;282:9789–96. 33. VanSlyke JK, Deschenes SM, Musil LS. Intracellular transport, assembly, and degradation of wildtype and disease-linked mutant gap junction proteins. Mol Biol Cell. 2000;11:1933–46. 34. VanSlyke JK, Musil LS. Cytosolic stress reduces degradation of connexin43 internalized from the cell surface and enhances gap junction formation and function. Mol Biol Cell. 2005;16:5247–57. 35. Sosinsky GE, Solan JL, Gaietta GM, Ngan L, Mackey M, Lampe PD. The C-terminus of Connexin43 present in the Golgi and in gap junctions is conformationally distinct: specific serines and prolines influence subcellular localization and ZO-1 interaction. Biochem J. 2007;408:375–85. 36. Ek-Vitorin JF, King TJ, Heyman NS, Lampe PD, Burt JM. Selectivity of connexin 43 channels is regulated through protein kinase C-dependent phosphorylation. Circ Res. 2006;98:1498–505. 37. Atkinson MM, Lampe PD, Lin HH, Kollander R, Li X-R, Kiang DT. Cyclic AMP modifies the cellular distribution of connexin 43 and induces a persistent increase in the junctional permeability of mouse mammary tumor cells. J Cell Sci. 1995;108:3079–90. 38. Burghardt RC, Barhoumi R, Sewall TC, Bowen JA. Cyclic AMP induces rapid increases in gap junction permeability and changes in the cellular distribution of connexin43. J Membr Biol. 1995;148:243–53. 39. Bolon ML, Kidder GM, Simon AM, Tyml K. Lipopolysaccharide reduces electrical coupling in microvascular endothelial cells by targeting connexin40 in a tyrosine-, ERK1/2-, PKA-, and PKC-dependent manner. J Cell Physiol. 2007;211:159–66. 40. Bolon ML, Ouellette Y, Li F, Tyml K. Abrupt reoxygenation following hypoxia reduces electrical coupling between endothelial cells of wildtype but not connexin40 null mice in oxidant- and PKA-dependent manner. FASEB J. 2005;19:1725–7. 41. Chanson M, White MM, Garber SS. cAMP promotes gap junctional coupling in T84 cells. Am J Physiol. 1996;271:C533–C9. 42. van Veen TAB, van Rijen HvM, Jongsma HJ. Electrical conductance of mouse connexin45 gap junction channels is modulated by phosphorylation. Cardiovasc Res. 2000;46:496–510. 43. Paulson AF, Lampe PD, Meyer RA, Atkinson MA, Walseth TF, Johnson RG. Cyclic AMP and LDL trigger a rapid enhancement in gap junction assembly through a stimulation of connexin trafficking. J Cell Sci. 2000;113:3037–49. 44. Johnson RG, Meyer RA, Li XR, Preus DM, Tan L, Grunenwald H, Paulson AF, Laird DW, Sheridan JD. Gap junctions assemble in the presence of cytoskeletal inhibitors, but enhanced assembly requires microtubules. Exp Cell Res. 2002;275:67–80. 45. TenBroek EM, Lampe PD, Solan JL, Reynhout JK, Johnson RG. Ser364 of connexin43 and the upregulation of gap junction assembly by cAMP. J Cell Biol. 2001;155:1307–18. 46. Yogo K, Ogawa T, Akiyama M, Ishida N, Takeya T. Identification and functional analysis of novel phosphorylation sites in Cx43 in rat primary granulosa cells. FEBS Lett. 2002;531:132–6. 47. Yogo K, Ogawa T, Akiyama M, Ishida-Kitagawa N, Sasada H, Sato E, Takeya T. PKA implicated in the phosphorylation of Cx43 induced by stimulation with FSH in rat granulosa cells. J Reprod Dev. 2006;52:321–8. 48. Shah MM, Martı´ nez A-M, Fletcher WH. The connexin43 gap junction protein is phosphorylated by protein kinase A and protein kinase C: in vivo and in vtitro studies. Mol Cell Biochem. 2002;238:57–68.

11 Biochemistry

283

49. TenBroek E, Lampe PD, Taffet S, Reynhout J, Martyn K, Kurata WE, Lau AF, Johnson RG. Enhanced gap junction assembly following the elevation of cAMP requires full-length Cx43. In: Werner R, editor. Gap Junctions. Amsterdam: IOS Press; 1998. pp.215–9. 50. Kawasaki H, Springett GM, Mochizuki N, Toki S, Nakaya M, Matsuda M, Housman DE, Graybiel AM. A family of cAMP-binding proteins that directly activate Rap1. Science. 1998;282:2275–9. 51. de Rooij J, Zwartkruis FJ, Verheijen MH, Cool RH, Nijman SM, Wittinghofer A, Bos JL. Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP. Nature. 1998;396:474–7. 52. Rangarajan S, Enserink JM, Kuiperij HB, de Rooij J, Price LS, Schwede F, Bos JL. Cyclic AMP induces integrin-mediated cell adhesion through Epac and Rap1 upon stimulation of the 2-adrenergic receptor. J Cell Biol. 2003;160:487–93. 53. Cullere X, Shaw SK, Andersson L, Hirahashi J, Luscinskas FW, Mayadas TN. Regulation of vascular endothelial barrier function by Epac, a cAMP-activated exchange factor for Rap GTPase. Blood. 2005;105:1950–5. 54. Fukuhara S, Sakurai A, Sano H, Yamagishi A, Somekawa S, Takakura N, Saito Y, Kangawa K, Mochizuki N. Cyclic AMP potentiates vascular endothelial cadherinmediated cell-cell contact to enhance endothelial barrier function through an EpacRap1 signaling pathway. Mol Cell Biol. 2005;25:136–46. 55. Somekawa S, Fukuhara S, Nakaoka Y, Fujita H, Saito Y, Mochizuki N. Enhanced functional gap junction neoformation by protein kinase A-dependent and Epac-dependent signals downstream of cAMP in cardiac myocytes. Circ Res. 2005;97:655–62. 56. Van Rijen HV, Van Veen TA, Hermans MM, Jongsma HJ. Human connexin40 gap junction channels are modulated by cAMP. Cardiovasc Res. 2000;45:941–51. 57. Mitropoulou G, Bruzzone R. Modulation of perch connexin35 hemi-channels by cyclic AMP requires a protein kinase A phosphorylation site. J Neurosci Res. 2003;72:147–57. 58. Ouyang X, Winbow VM, Patel LS, Burr GS, Mitchell CK, O’Brien J. Protein kinase A mediates regulation of gap junctions containing connexin35 through a complex pathway. Brain Res Mol Brain Res. 2005;135:1–11. 59. Keane RW, Mehta P, Rose B, Honig LS, Loewenstein WR, Rutishauser U. Neural differentiation, NCAM-mediated adhesion and gap junctional communication in neuroectoderm. A study in vitro. J Cell Biol. 1988;106:1307–19. 60. Meyer RA, Laird DW, Revel J-P, Johnson RG. Inhibition of gap junction and adherens junction assembly by connexin and A-CAM antibodies. J Cell Biol. 1992;119:179–89. 61. Xu X, Li WE, Huang GY, Meyer R, Chen T, Luo Y, Thomas MP, Radice GL, Lo CW. Modulation of mouse neural crest cell motility by N-cadherin and connexin 43 gap junctions. J Cell Biol. 2001;154:217–30. 62. Fujimoto K, Nagafuchi A, Tsukita S, Kuraoka A, Ohokuma A, Shibata Y. Dynamics of connexins, E-cadherin and -catenin on cell membranes during gap junction formation. J Cell Sci. 1997;110:311–22. 63. Wu JC, Tsai RY, Chung TH. Role of catenins in the development of gap junctions in rat cardiomyocytes. J Cell Biochem. 2003;88:823–35. 64. Luo Y, Radice GL. Cadherin-mediated adhesion is essential for myofibril continuity across the plasma membrane but not for assembly of the contractile apparatus. J Cell Sci. 2003;116:1471–9. 65. Paul DL, Yu K, Bruzzone R, Gimlich RL, Goodenough DA. Expression of a dominantnegative inhibitor of intercellular communication in the early Xenopus embryo causes delamination and extrusion of cells. Development. 1995;121:371–81. 66. Lee S, Gilula NB, Warner AE. Gap junctional communication and compaction during preimplantation stages of mouse development. Cell. 1987;51:851–80. 67. Shaw RM, Fay AJ, Puthenveedu MA, von Zastrow M, Jan YN, Jan LY. Microtubule plus-end-tracking proteins target gap junctions directly from the cell interior to adherens junctions. Cell. 2007;128:547–60.

284

J.L. Solan and P.D. Lampe

68. Giepmans BN,Verlaan I, Hengeveld T, Janssen H,Calafat J, Falk MM,MoolenaarWH.Gap junction protein connexin-43 interacts directly with microtubules. Curr Biol. 2001;11:1364–8. 69. Wei CJ, Francis R, Xu X, Lo CW. Connexin43 associated with an N-cadherin-containing multiprotein complex is required for gap junction formation in NIH3T3 cells. J Biol Chem. 2005;280: 19925–36. 70. Kojima T, Kokai Y, Chiba H, Yamamoto M, Mochizuki Y, Sawada N. Cx32 but not Cx26 is associated with tight junctions in primary cultures of rat hepatocytes. Exp Cell Res. 2001;263:193–201. 71. Nagasawa K, Chiba H, Fujita H, Kojima T, Saito T, Endo T, Sawada N. Possible involvement of gap junctions in the barrier function of tight junctions of brain and lung endothelial cells. J Cell Physiol. 2006;208:123–32. 72. Singh D, Solan JL, Taffet SM, Javier R, Lampe PD. Connexin 43 interacts with zona occludens-1 and -2 proteins in a cell cycle stage-specific manner. J Biol Chem. 2005;280:30416–21. 73. Hunter AW, Barker RJ, Zhu C, Gourdie RG. Zonula occludens-1 alters connexin43 gap junction size and organization by influencing channel accretion. Mol Biol Cell. 2005;16:5686–98. 74. Zhu C, Barker RJ, Hunter AW, Zhang Y, Jourdan J, Gourdie RG. Quantitative analysis of ZO-1 co-localization with Cx43 gap junction plaques in cultures of rat neonatal cardiomyocytes. Microsc Microanal. 2005;11:244–8. 75. Akoyev V, Takemoto DJ. ZO-1 is required for protein kinase C g-driven disassembly of connexin 43. Cell Signal. 2007;19:958–67. 76. Butkevich E, Hulsmann S, Wenzel D, Shirao T, Duden R, Majoul I. Drebrin is a novel connexin-43 binding partner that links gap junctions to the submembrane cytoskeleton. Curr Biol. 2004;14:650–8. 77. Li W, Nagy JI. Connexin43 phosphorylation state and intercellular communication in cultured astrocytes following hypoxia and protein phosphatase inhibition. Eur J Neurosci. 2000;12:2644–50. 78. Li WE, Ochalski PA, Hertzberg EL, Nagy JI. Immunorecognition, ultrastructure and phosphorylation status of astrocytic gap junctions and connexin43 in rat brain after cerebral focal ischaemia. Eur J Neurosci. 1998;10:2444–63. 79. Cooper CD, Lampe PD. Casein kinase 1 regulates connexin43 gap junction assembly. J Biol Chem. 2002;277:44962–8. 80. Lampe PD, Cooper CD, King TJ, Burt JM. Analysis of connexin43 phosphorylated at S325, S328 and S330 in normoxic and ischemic heart. J Cell Sci. 2006;119:3435–42. 81. Doble BW, Dang X, Ping P, Fandrich RR, Nickel BE, Jin Y, Cattini PA, Kardami E. Phosphorylation of serine 262 in the gap junction protein connexin-43 regulates DNA synthesis in cell-cell contact forming cardiomyocytes. J Cell Sci. 2004;117:507–14. 82. Lampe PD, TenBroek EM, Burt JM, Kurata WE, Johnson RG, Lau AF. Phosphorylation of connexin43 on serine368 by protein kinase C regulates gap junctional communication. J Cell Biol. 2000;126:1503–12. 83. Berthoud VM, Rook M, Hertzberg EL, Sa´ez JC. On the mechanism of cell uncoupling induced by a tumor promotor phorbol ester in clone 9 cells, a rat liver epithelial cell line. Eur J Cell Biol. 1993;62:384–96. 84. Reynhout JK, Lampe PD, Johnson RG. An activator of protein kinase C inhibits gap junction communication between cultured bovine lens cells. Exp Cell Res. 1992;198:337–42. 85. Peracchia C, Sotkis A, Wang XG, Peracchia LL, Persechini A. Calmodulin directly gates gap junction channels. J Biol Chem. 2000;275:26220–4. 86. Torok K, Stauffer K, Evans WH. Connexin 32 of gap junctions contains two cytoplasmic calmodulin-binding domains. Biochem J. 1997;326:479–83. 87. Zhang X, Qi Y. Role of intramolecular interaction in connexin50: mediating the Ca2+ dependent binding of calmodulin to gap junction. Arch Biochem Biophys. 2005;440:111–7. 88. Burr GS, Mitchell CK, Keflemariam YJ, Heidelberger R, O’Brien J. Calcium-dependent binding of calmodulin to neuronal gap junction proteins. Biochem Biophys Res Commun. 2005;335:1191–8.

11 Biochemistry

285

89. Loo LW, Kanemitsu MY, Lau AF. In vivo association of pp60v-src and the gap-junction protein connexin 43 in v-src-transformed fibroblasts. Mol Carcinogen. 1999;25:187–95. 90. Giepmans BN, Hengeveld T, Postma FR, Moolenaar WH. Interaction of c-Src with gap junction protein connexin-43. Role in the regulation of cell-cell communication. J Biol Chem. 2001;276:8544–9. 91. Toyofuku T, Akamatsu Y, Zhang H, Kuzuya T, Tada M, Hori M. c-Src regulates the interaction between connexin-43 and ZO-1 in cardiac myocytes. J Biol Chem. 2001;276:1780–8. 92. Lin R, Warn-Cramer BJ, Kurata WE, Lau AF. v-Src phosphorylation of connexin 43 on Tyr247 and Tyr265 disrupts gap junctional communication. J Cell Biol. 2001;154:815–27. 93. Sorgen PL, Duffy HS, Sahoo P, Coombs W, DelmarM, Spray DC. Structural changes in the carboxyl terminus of the gap junction protein connexin43 indicates signaling between binding domains for c-Src and zonula occludens-1. J Biol Chem. 2004;279:54695–701. 94. Duffy HS, Ashton AW, O’Donnell P, Coombs W, Taffet SM, Delmar M, Spray DC. Regulation of connexin43 protein complexes by intracellular acidification. Circ Res. 2004;94:215–22. 95. Cottrell GT, Lin R, Warn-Cramer BJ, Lau AF, Burt JM. Mechanism of v-Src- and mitogen-activated protein kinase-induced reduction of gap junction communication. Am J Physiol Cell Physiol. 2003;284:C511–20. 96. Zhou L, Kasperek EM, Nicholson BJ. Dissection of the molecular basis of pp60(v-src) induced gating of connexin 43 gap junction channels. J Cell Biol. 1999;144:1033–45. 97. Warn-Cramer BJ, Cottrell GT, Burt JM, Lau AF. Regulation of connexin-43 gap junctional intercellular communication by mitogen-activated protein kinase. J Biol Chem. 1998;273:9188–96. 98. Polontchouk L, Ebelt B, Jackels M, Dhein S. Chronic effects of endothelin 1 and angiotensin II on gap junctions and intercellular communication in cardiac cells. FASEB J. 2002;16:87–9. 99. Petrich BG, Gong X, Lerner DL, Wang X, Brown JH, Saffitz JE, Wang Y. c-Jun Nterminal kinase activation mediates downregulation of connexin43 in cardiomyocytes. Circ Res. 2002;91:640–7. 100. Cameron SJ, Malik S, Akaike M, Lerner-Marmarosh N, Yan C, Lee JD, Abe J, Yang J. Regulation of epidermal growth factor-induced connexin 43 gap junction communication by big mitogen-activated protein kinase1/ERK5 but not ERK1/2 kinase activation. J Biol Chem. 2003;278:18682–8. 101. Yamamoto T, Kojima T, Murata M, Takano K, Go M, Hatakeyama N, Chiba H, Sawada N. p38 MAP-kinase regulates function of gap and tight junctions during regeneration of rat hepatocytes. J Hepatol. 2005;42:707–18. 102. Qin H, Shao Q, Igdoura SA, Alaoui-Jamali MA, Laird DW. Lysosomal and proteasomal degradation play distinct roles in the life cycle of Cx43 in gap junctional intercellular communication-deficient and -competent breast tumor cells. J Biol Chem. 2003;278:30005–14. 103. Thomas MA, Zosso N, Scerri I, Demaurex N, Chanson M, Staub O. A tyrosine-based sorting signal is involved in connexin43 stability and gap junction turnover. J Cell Sci. 2003;116:2213–22. 104. Leykauf K, Salek M, Bomke J, Frech M, Lehmann WD, Durst M, Alonso A. Ubiquitin protein ligase Nedd4 binds to connexin43 by a phosphorylation-modulated process. J Cell Sci. 2006;119:3634–42. 105. Lan Z, Kurata WE, Martyn KD, Jin C, Lau AF. Novel rab GAP-like protein, CIP85, interacts with connexin43 and induces its degradation. Biochemistry 2005;44:2385–96. 106. Xie H, Laird DW, Chang T-H, Hu VW. A mitosis-specific phosphorylation of the gap junction protein connexin43 in human vascular cells: biochemical characterization and localization. J Cell Biol. 1997;137:203–10. 107. Lampe PD, Kurata WE, Warn-Cramer B, Lau AF. Formation of a distinct connexin43 phosphoisoform in mitotic cells is dependent upon p34cdc2 kinase. J Cell Sci. 1998;111:833–41.

286

J.L. Solan and P.D. Lampe

108. Kanemitsu MY, Jiang W, Eckhart W. Cdc2-mediated phosphorylation of the gap junction protein, connexin43, during mitosis. Cell Growth Differ. 1998;9:13–21. 109. Yin X, Gu S, Jiang JX. The development-associated cleavage of lens connexin 45.6 by caspase-3-like protease is regulated by casein kinase II-mediated phosphorylation. J Biol Chem. 2001;276:34567–72. 110. Yin X, Jedrzejewski PT, Jiang JX. Casein kinase II phosphorylates lens connexin 45.6 and is involved in its degradation. J Biol Chem. 2000;275:6850–6. 111. Hertlein B, Butterweck A, Haubrich S, Willecke K, Traub O. Phosphorylated carboxyterminal serine residues stabilize the mouse gap junction protein connexin45 against degradation. J Membrane Biol. 1998;162:247–57. 112. Squecco R, Sassoli C, Nuti F, Martinesi M, Chellini F, Nosi D, Zecchi-Orlandini S, Francini F, Formigli L, Meacci E. Sphingosine 1-phosphate induces myoblast differentiation through Cx43 protein expression: a role for a gap junction-dependent and independent function. Mol Biol Cell. 2006;17:4896–910. 113. Gellhaus A, Dong X, Propson S, Maass K, Klein-Hitpass L, Kibschull M, Traub O, Willecke K, Perbal B, Lye SJ, Winterhager E. Connexin43 interacts with NOV: a possible mechanism for negative regulation of cell growth in choriocarcinoma cells. J Biol Chem. 2004;279:36931–42. 114. Balda MS, Garrett MD, Matter K. The ZO-1-associated Y-box factor ZONAB regulates epithelial cell proliferation and cell density. J Cell Biol. 2003;160:423–32. 115. Ciolofan C, Li XB, Olson C, Kamasawa N, Gebhardt BR, Yasumura T, Morita M, Rash JE, Nagy JI. Association of connexin36 and zonula occludens-1 with zonula occludens-2 and the transcription factor zonula occludens-1-associated nucleic acid-binding protein at neuronal gap junctions in rodent retina. Neuroscience. 2006;140:433–51. 116. Lagree V, Brunschwig K, Lopez P, Gilula NB, Richard G, Falk MM. Specific aminoacid residues in the N-terminus and M3 implicated in channel function and oligomerization compatibility of connexin43. J Cell Sci. 2003;116:3189–201. 117. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys. 2001;34:325–472. 118. White TW, Paul DL, Goodenough DA, Bruzzone R. Functional analysis of selective interactions among rodent connexins. Mol Biol Cell. 1995;6:459–70. 119. Foote CI, Zhou L, Zhu X, Nicholson BJ. The pattern of disulfide linkages in the extracellular loop regions of connexin 32 suggests a model for the docking interface of gap junctions. J Cell Biol. 1998;140:1187–97. 120. Bao X, Chen Y, Reuss L, Altenberg GA. Functional expression in Xenopus oocytes of gap-junctional hemichannels formed by a cysteine-less connexin 43. J Biol Chem. 2004;279:9689–92. 121. Shibayama J, Gutie´rrez C, Gonzalez D, Kieken F, Seki A, Carrion JR, Sorgen PL, Taffet SM, Barrio LC, Delmar M. Effect of charge substitutions at residue his-142 on voltage-gating of connexin43 channels. Biophys J. 2006;91:4054–63. 122. Kanemitsu MY, Loo LW, Simon S, Lau AF, Eckhart W. Tyrosine phosphorylation of connexin 43 by v-src is mediated by SH2 and SH3 domain interactions. J Biol Chem. 1997;272:22824–31. 123. Giepmans BN, Feiken E, Gebbink MF, Moolenaar WH. Association of connexin43 with a receptor protein tyrosine phosphatase. Cell Commun Adhes. 2003;10:201–5. 124. Wang XG, Peracchia C. Positive charges of the initial C-terminus domain of Cx32 inhibit gap junction gating sensitivity to CO2. Biophys J. 1997;73:798–806. 125. Giepmans BN, Moolenaar WH. The gap junction protein connexin43 interacts with the second PDZ domain of the zona occludens-1 protein. Curr Biol. 1998;8:931–4. 126. Toyofuku T, Yabuki M, Otsu K, Kuzuya T, Hori M, Tada M. Direct association of the gap junction protein connexin-43 with ZO-1 in cardiac myocytes. J Biol Chem. 1998;273:12725–31.

Chapter 12

Pannexins or Connexins? Gerhard Dahl and Andrew L. Harris

Abstract Pannexins are a family of three vertebrate proteins that have moderate sequence homology with the innexin proteins, which compose gap junction channels in protostomes, including most invertebrates. However, it appears that in contrast to innexins, pannexins do not have the ability to form gap junction channels. They do, however, form nonjunctional plasma membrane channels (pannexons) that mediate regulated flux of molecules in the size range of second messengers between cytoplasm and the extracellular space. The dye permeability and pharmacological sensitivities of pannexin channels overlap those of connexin hemichannels, so it is possible that many of the phenomena that have been attributed to connexin hemichannels are in fact mediated by pannexons. For this reason, identifying which protein is involved in a particular cellular physiology requires careful evaluation of the specific conditions and requirements in each case. Several lines of experimentation have led to the suggestion that the adenosine triphosphate (ATP) release channel, a crucial element in the initiation and propagation of intercellular Ca2+ waves, is a connexin hemichannel. However, based on recently revealed properties of pannexin channels, which include mechanosensitivity and activation by cytoplasmic Ca2+, the pannexon must be considered a prime candidate for the ATP release channel. Pannexons also appear to form the large ATP-permeable pore that is activated by the purinergic P2X7 receptor complex, which is involved in inflammation. Keywords Pannexin Innexin Connexin Gap junction Hemichannel Pannexon ATP release Calcium wave P2Y receptor P2X7 receptor Cx32 Cx38 Cx43

G. Dahl (*) Department of Physiology and Biophysics, University of Miami School of Medicine, PO Box 016430, Miami, FL 33101, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_12, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

287

288

G. Dahl and A.L. Harris

12.1 Introduction Connexins form gap junction channels in deuterostomes, which include all vertebrates. A distinct family of proteins, the innexins, forms gap junction channels in protostomes, which include most invertebrates [1,2]. There is no sequence homology between these two families of gap junction proteins. A search of the human genome identified three innexin-related genes [3,4]. Because of the occurrence of homologous genes in both vertebrates and invertebrates, the corresponding proteins were termed pannexins. However, because subsequent studies indicate that the vertebrate homologs are functionally distinct from the innexins, they are treated here as a separate family, and the term refers only to the three vertebrate proteins, denoted as pannexin1 (Panx1), pannexin2 (Panx2), and pannexin3 (Panx3). Pannexins appear to have the same transmembrane topology as connexins and innexins, with four transmembrane domains and cytoplasmic carboxyl-terminal and amino-terminal domains [5,6]. Early functional studies indicated that Panx1 could form gap junction channels in paired oocytes, as well as produce a nonjunctional conductance [7]. Panx1 channels are larger than the largest connexin channels, with regard to both conductance and apparent pore width [8]. The cellular and tissue expression of pannexins overlaps with that of connexins. This chapter summarizes current knowledge about pannexin channels, with particular attention to the criteria for resolving their cellular functions as distinct from those of connexin channels. It illustrates the major criteria for making this distinction by discussing in depth the degree of correspondence between the characteristics of the adenosine triphosphate (ATP) release channel and those of plasma membrane pannexin and connexin channels. The preponderance of data argues strongly for direct pannexin, as opposed to connexin, involvement in nonvesicular ATP release.

12.2 Do Pannexins Form Gap Junction Channels? Despite the reports of the ability of pannexin channels to form gap junction channels in paired oocytes, they may not do so under conditions of normal expression, either in vivo or in vitro (Table 12.1). Functional assays in transfected cultured cells have failed to provide evidence of gap junctions formed by Table 12.1 Evidence relating to gap junction function of pannexin Evidence for gap junction Junctional conductance in paired oocytes [7]. function Dye transfer through unspecified channels in pannexin transfected cells [9,10] Evidence against gap junction No dye and electrical coupling in cell cultures expressing function Panx1 [6,11] No punctate staining typical for gap junctions [11,14,15] Pannexin is glycosylated [6,17]

12 Pannexin Channels

289

pannexins, with two exceptions [9,10]. No pannexin-induced dye-coupling or electrical coupling has been observed in a variety of pannexin-expressing host cells [6,11]. Thus, the only cases in which pannexins may form gap junctions are in conditions of overexpression, and in only a subset of those. Because in the overexpression studies using tissue culture cells no diagnostic test that would discriminate between connexin and pannexin channels was performed, the channel identity even in the apparent exceptions is not clear, as connexin expression could have been upregulated in the pannexin transfected cells. Gap junctions are characterized by punctate immunohistochemical staining at regions of intercellular contact [12,13]. The puncta correspond to gap junction plaques. Immunostaining for Panx1 does not show punctate staining [11,14], but rather the diffuse staining of the cell surface expected of a nonclustered protein distribution. In addition, in the epithelia tested to date, pannexin expression in the plasma membrane is limited to luminal rather than the basolateral surfaces where gap junctions are typically located [15]. At neuronal synapses, Panx1 is found only in the postsynaptic, and not the presynaptic, membrane, excluding a junctional role [16]. Unlike connexins, which are not glycosylated, Panx1 is glycosylated in the second extracellular loop [6,17]. This modification adds substantial bulk to the extracellular-facing aspect of the protein. Insertion of glycosylation sites into the extracellular loop domains of connexins blocks formation of junctional channels when these sites are glycosylated [18]. Therefore, it appears that bulky carbohydrates on the extracellular loops prevent the tight docking interactions between connexin hemichannels required for formation of a patent cell–cell channel. On this basis, it is difficult to imagine that glycosylated pannexins can form intercellular channels. How, then, can the formation of gap junctions by pannexin be explained, in the few cases where it occurs when (over)expressed? Xenopus oocytes are prone to faulty glycosylation [19,20,21]. Furthermore, in any transfected cell overexpression of exogenous protein may overwhelm the glycosylation machinery in the Golgi apparatus and allow an appreciable amount of aberrantly unglycosylated protein to proceed through the synthesis and insertion pathways. Thus, the reports of gap junction channel formation by Panx1 in oocytes and perhaps some other expression systems may be an expression system artifact and not reflect the ability of the protein to form an intercellular channel in vivo or under conditions of endogenous expression levels. The ability of Panx2 or Panx3 to form gap junctions, or any type of functional channel, has not been demonstrated. Bruzzone et al. [7] found that expression of Panx2 or Panx3 does not produce channel activity under normal test conditions. Panx2, however, did modulate Panx1 channel activity when the two pannexins were coexpressed. The gene structures predict that both Panx2 and Panx3 can be expressed as different isoforms due to alternative RNA splicing [22]. It is therefore possible that the channel-forming isoform(s) have not yet been studied, or that the experimental conditions favorable for channel activity of these pannexins have not yet been found.

290

G. Dahl and A.L. Harris

12.3 Properties of Pannexin Plasma Membrane Channels Cells that never form gap junctions, such as mature erythrocytes, express Panx1 [5]. Panx1 expression in oocytes leads to a nonjunctional membrane conductance [7]. The channels have a large unitary conductance (475 picoSiemens [pS]) and exhibit numerous subconductance states (Fig. 12.1) [8]. Panx1 channels are highly permeable to ATP, to the extent that current carried through the channel

Fig. 12.1 Single-channel currents in an inside-out membrane patch excised from an oocyte expressing Panx1. An uninterrupted recording segment of 140 seconds is shown together with an all point histogram of the entire segment. Characteristic of pannexin channels, several subconductance levels can be discerned (indicated by dotted lines). When actively gating, fully closed and fully open states are rare events. The unitary conductance of the full open state is 475 pS. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Bao et al. [8] with permission.)

12 Pannexin Channels

291

by ATP itself can be discerned. The channel is permeable to dyes that permeate connexin hemichannels and is accessible to polyethylene glycols up to 1.5 kDa [23]. Panx1 channels are closed at negative potentials, open at positive potentials, and inactivate over time. Panx1 channels can be activated at normal resting membrane potentials by mechanical stress [8] or by increases of cytoplasmic Ca2+ concentration to the micromolar range [24]. These properties suggest that pannexin plasma membrane channels have a physiological role: to regulate the flux of molecules in the size range of second messengers between cytoplasm and extracellular space [14]. The nonplaque, diffuse plasma membrane localization of pannexin protein seen immunohistochemically (mentioned above) is consistent with such a role.

12.4 The Adenosine Triphosphate Release Channel Intercellular Ca2+ waves are widespread. They serve diverse functions, including control of ciliary beat in airway epithelia, control of peripheral vascular perfusion, modulation of synaptic transmission by glia, and ossification [25,26,27,28]. Ca2+ waves are initiated by various physiological stimuli, including extracellular ATP and mechanical stress. Ca2+ wave propagation can involve two pathways: direct intercellular flux of inositol triphosphate (IP3) through gap junction channels and an extracellular pathway involving ATP release and purinergic receptors [25,29,30]. In the latter, ATP is released from the initiating cell (in response to mechanical stress, for example) and from cells in the wavefront in response to activation of purinergic receptors. The most compelling evidence for channel-mediated as opposed to vesicular release is that ATP can be released from cells that do not contain vesicles, such as erythrocytes [31]. On this basis, a channel for ATP release would be expected to be mechanosensitive and to open upon activation of purinergic receptors. On the basis of several experimental findings, connexin channels have been proposed to mediate ATP release. Gap junction inhibitors interfere with ATP release [32,33], and, under conditions of ATP release, cells take up extracellular dyes known to permeate connexin channels [34,35]. Conventional gap junction channel blockers also inhibit the dye uptake. In addition, there is a notable correlation between induced expression of connexin and ATP release [36]. Thus, it was reasonable to infer that connexins, in addition to forming gap junctions, provided the ATP release pathway in the form of connexin hemichannels in the nonjunctional membrane. The mechanosensitivity of the ATP release channel combined with its ATP permeability raise the possibility that pannexons play a role in the initiation of Ca2+ waves. Wave propagation may involve pannexons as well, since when coexpressed with either P2Y or P2X7 receptors [24,37] they open in the presence of extracellular ATP. Activation of the metabotropic P2Y receptors leads to the release of Ca2+ from intracellular stores, which in turn can activate pannexin

292

G. Dahl and A.L. Harris

channels. In the case of P2X7 receptors, activation of pannexin channels does not require influx of extracellular Ca2+ through the inotropic receptor [37]. Instead, pannexin activation may occur by protein–protein interaction. A P2X7-Panx1 interaction is indicated by co-immunoprecipitation of the two proteins [38].

12.4.1 Which of the Candidate Channels Are Active Under the Required Conditions? Under physiological conditions most connexins form hemichannels in the plasma membrane [39,40] that are closed until they dock during gap junction formation to form cell–cell channels. With some exceptions, connexin hemichannel currents tend to be activated by strong depolarization or reduction of extracellular Ca2+ below 0.5 mM [41,42]. Thus, the activity of endogenous connexin hemichannels is unlikely to be significant under normal physiological conditions, such as those under which Ca2+ waves occur. The situation for pannexin channels is quite different. The physiological stimuli for Ca2+ wave initiation and propagation — mechanical stress and activation of purinergic receptors — can induce pannexin channel activity at the resting membrane potential in a normal ionic environment (Fig. 12.2) [8,24,37]. By definition, during Ca2+ waves there is a propagated increase in cytoplasmic Ca2+. Gap junction channels are either closed by increased cytoplasmic Ca2+ or unaffected by it, depending on the particular connexin [43,44]. Opening of connexin hemichannels by cytoplasmic Ca2+ would require a reversal of Ca2+ sensitivity between single connexin hemichannels and hemichannels that are in gap junction channels. Such a reversal of gating sensitivity is not only conceptually unappealing but also has not been observed in connexins that form functional single hemichannels [45,46,47,48]. Pannexons, on the other

Fig. 12.2 Scheme depicting possible involvement of Panx1 channels in calcium wave initiation and propagation. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Locovei et al. [24] with permission.)

12 Pannexin Channels

293

hand, can be activated at the resting membrane potential by increases in Ca2+ concentration comparable to those in Ca2+ waves [24]. ATP release and Ca2+ wave propagation are facilitated by reduction of extracellular Ca2+ to the micromolar range [32,49,50]. This treatment also can open connexin hemichannels [48,51]. However, reduction of extracellular Ca2+ typically leads to a rise in cytoplasmic Ca2+ [52], which could activate pannexin channels. Therefore, the sensitivity of ATP release to low extracellular Ca2+ cannot be considered as unambiguous evidence for connexin as opposed to pannexin involvement.

12.4.2 Which of the Candidate Proteins Are Expressed in the Right Places? A candidate ATP release channel must exhibit an expression pattern consistent with its function. Many tissues express both connexins and pannexins and therefore cannot help to discriminate between the contributions of the two candidates. However, there are cases where ATP is released from cells that express pannexin but not connexin. A prime example is the erythrocyte [5], which releases ATP in a low oxygen environment or in response to shear stress [31,53]. ATP release in the presence of pannexin and absence of connexin also occurs in the receptor cells of the taste bud [54]. Although the gustatory epithelium expresses both Panx1 and several connexins [55], analysis of the expression patterns of individual cell types within taste buds revealed that the cells that release ATP express Panx1 but no connexins [54]. Conversely, there are no known cases where there is nonvesicular ATP release in the absence of pannexin, whether or not connexin is present. A candidate ATP release channel also must be expressed at the appropriate cellular sites. In the airway epithelium, ATP is released from the apical aspect of the cell. Panx1 is expressed at high concentration in these cells at the apical membrane, and not at all at the basolateral membrane [15]. Though not established for these particular cells, connexin expression in polarized cells is typically in the basolateral membrane and not in the apical membrane [12].

12.4.3 The Dilemma of Genetic Manipulation The strongest evidence for connexins serving as ATP release channels comes from studies where forced expression of connexins is correlated with increased ATP release [32,36]. However, this inference is directly contradicted by findings from cells derived from Cx43 knockout (KO) mice in which Ca2+ wave propagation proceeds at approximately the same speed in astrocytes from Cx43KO mice as in those from wild-type mice [56]. Because Cx43 is the major connexin expressed in these cells, both pathways for wave propagation, gap junctional

294

G. Dahl and A.L. Harris

and extracellular, ought to be eliminated in the Cx43KO cells. Electrical coupling was greatly attenuated, indicating that the persisting Ca2+ waves propagated via the extracellular pathway. Clearly, the ATP release required for this could not occur through Cx43 channels in these Cx43KO cells. What, then, can explain the experiments in which presence of ATP release channels seems to correlate with connexin expression? Compensatory mechanisms have been recognized to be common phenomena in both knockout and forced expression studies. Specifically, gene knockout or overexpression can cause widespread changes in expression of other genes, particularly in the case of connexin genes, which affect a key mechanism of intercellular signaling. For example, Iacobas et al. [57,58,59] have shown that up to 10% of the proteome is altered in Cx43KO mice; some proteins are upregulated and others downregulated. The consequence is a reversal of conventional scientific wisdom; in this case, only a negative result is interpretable: persistence of propagated Ca2+ in the absence of Cx43 indicates lack of involvement of Cx43 [56]. Although not established, abolition of Ca2+ waves could be due to downstream effects of the genetic manipulation. In this case, to draw a conclusion about the role of Cx43, it would be necessary to document whether the expression of pannexin, for example, was also downregulated. Similarly, ATP release studies in which connexin is upregulated consequently must also assess changes in pannexin expression, when the connexin manipulation has an effect. Acute expression in oocytes by injection of specific messenger RNA is less likely to suffer from compensatory complications. Panx1 expressed in oocytes results in ATP release in response to depolarization [8]. Under the same conditions, Cx43 expression did not lead to ATP release by these cells [8]. In 0 mM extracellular Ca2+, Cx38 may do so [60], although ATP release in these cells has been shown to be sensitive to short exposure to brefeldin, indicating a vesicular release mechanism [61]. Consistent with an ATP release function of Panx1, its knockdown by interfering antisense oligodeoxyribonucleotides (RNAi) in an astrocytoma cell line attenuates ATP-induced dye uptake [37].

12.4.4 The Dilemma of Pharmacology Reagents commonly used to block connexin channels are not specific for those channels; they also affect other membrane channels [14,62,63,64] (see Chapter 8). Importantly, carbenoxolone and niflumic acid block both connexin and pannexin channels. More relevant to ATP release, some compounds, such as octanol and 18 -glycyrrhetinic acid, block gap junction channels at concentrations that do not affect Ca2+ wave propagation [36]. The lipid oleamide has been suggested to discriminate between gap junction communication and Ca2+ wave propagation in glial cells [65], an apparent contradiction if connexin channels are involved in the latter. It was therefore seen as a major step toward connexin-specific agents when connexin mimetic peptides were found to attenuate ATP release and Ca2+ wave

12 Pannexin Channels

295

propagation [66]. Over two dozen publications have appeared in which connexin mimetic peptides were used to assess the contribution of connexin channels to Ca2+ waves. However, the specificity of these peptides for connexin channels has never been assessed. The currents through hemichannels formed by the connexin chimera Cx32*43E1 [47,67] are not acutely altered by the connexin mimetic peptides [23]. On the other hand, these peptides, termed GAP24 and GAP27, rapidly inhibit pannexin channel currents [23] (see Chapter 8). Pannexin mimetic peptides, based on the extracellular loop domains of Panx1, acutely inhibit both pannexin and connexin channels [23]. These data suggest that the effects of these peptides are not sequence-specific but arise from steric block of these large channels. No matter what the mechanism, the above findings argue that acute effects of the connexin peptides on a process, such as dye uptake, ATP release, and Ca2+ waves, indicate the involvement of pannexin rather than connexin channels. This point has been experimentally validated by two recent publications that together show that connexin mimetic peptides acutely inhibit ATP release (interpreted to indicate connexin involvement [55]) and that the cells so affected express pannexin but not connexin [54]. In summary, prior to knowledge about pannexin channel properties, connexin channel activity best explained some of the phenomena associated with ATP release in some tissues. However, as indicated above and summarized in Table 12.2, pannexons are now the stronger candidates for ATP release channels.

Table 12.2 Comparison of properties of connexin and pannexin channels Connexin Pannexin Channel opens under physiological conditions Channel is mechanosensitive

No Demonstrated only for Cx46 [72] Not demonstrated No No [23]

Yes [5,8,24] Yes [8]

Channel can be opened by ATP through P2 receptors Yes [24,37] Channel is activated by cytoplasmic calcium Yes [24] Channel activity is rapidly attenuated by connexin Yes [23] mimetic peptides Yes [5,15,54] Channel expression pattern matches ATP release Poorly1 Localization is consistent with sites of ATP release No Yes [15]. No [56] Yes [37,38] Protein knockout or knock-down decreases ATP release or surrogate measurements2 Yes [36,73] Yes [8] Exogenous expression increases ATP release or surrogate measurements2 Channel is inhibited by gap junction blockers Yes Yes [70] Channel mediates dye uptake Yes [74] Yes [37,38] 1 ATP release and Ca2+ waves are found in cells in the absence of connexin including invertebrates and vertebrate erythrocytes and taste receptor cells. 2 Typical surrogate measures for ATP release are dye uptake and Ca2+ wave propagation measurements.

296

G. Dahl and A.L. Harris

12.5 Pannexins, Purinergic Receptors, and Cell Death A second large, dye-permeable pore that has sometimes been confused with connexin channels is that activated by stimulation of the purinergic receptor P2X7R (also known as the P2Z receptor; [68]). This receptor is a member of the ionotropic P2X family of proteins. This receptor behaves like the other P2X receptors when exposed briefly to the agonists ATP or benoylbenzoyl-ATP by opening a relatively small 20 pS cation pore. With prolonged or repetitive exposure, it induces opening of a large, lytic, dye-permeable, and ATP-permeable pore. The properties of the P2X7R large pore are similar to those of Panx1 and connexin channels. Common features include unitary conductance, permeability to dyes typically used to study gap junction channels, sensitivity to cytoplasmic acidification, and sensitivity to gap junction channel blockers [8,24,69,70,71]. Several lines of experimentation indicate that P2X7R activation of the large pore requires Panx1 channels. It has been recently shown that P2X7R-mediated cell death involves Panx1 (Fig. 12.3) [37]. ATP-induced dye uptake in macrophages and in astrocytes is attenuated by RNAi targeting Panx1 [37,38]. Coexpression of P2X7R with Panx1 results in ATP-induced dye uptake in Xenopus oocytes, while neither of the two proteins alone mediates this phenomenon [37]. The activation of pannexon currents by ATP through P2X7R does not require influx of Ca2+ [37] but may occur through protein–protein interaction [38].

Fig. 12.3 Membrane currents and morphology of oocytes coexpressing P2X7R and Panx1 with exposure to ATP. (a) Brief exposure to 200 mM ATP resulted in a reversible inward current. Longer exposure to 500 mM ATP resulted in a current that initially reversed partially but subsequently increased to levels that escaped voltage clamp, indicating membrane breakdown. (b) Noninjected control oocytes (upper left group) and oocytes coexpressing P2X7R and Panx1 (lower right group) were exposed to 300 mM ATP. (c) Within three minutes a dramatic change of pigmentation was observed in the coexpressing but not in the control oocytes. (d) After fiveminutes exposure to ATP, all coexpressing oocytes lost cellular integrity, indicated by yolk oozing out of the cells. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Locovei et al. [37] with permission.)

12 Pannexin Channels

297

There is also evidence that activation of metabotropic P2Y receptors also results in Panx1 currents [24]. In this case, the activation is most likely via the increase in cytoplasmic Ca2+ that follows ligand binding to the receptor.

12.6 Conclusion Pannexins were discovered only recently and knowledge about their functional roles is just emerging. Although they were discovered based on their limited sequence homology with the invertebrate gap junction innexin proteins, pannexins do not form cell–cell channels. Instead their physiological role may be exclusively to provide a highly regulated pathway for molecules in the size range of second messengers to cross the plasma membrane. Considering the amino acid sequence relationship between innexins and pannexins, the following evolutionary scenario becomes plausible: An invertebrate ur-channel acquired the ability to dock to its counterpart in another cell to form a cell–cell channel. Gene duplications occurred in the innexin family. Some or all of the innexins formed dual function channels; in addition to gap junction function they retained the ability to act as nonjunctional channels, allowing the exchange of molecules between cytoplasm and extracellular space. With the advent of connexins the gap junction function was usurped by these proteins, while the modern innexins, the pannexins, were retained for nonjunctional purposes. Acknowledgments The authors thank Drs. Nirupa Chaudhari, Silviu Locovei, Ken Muller, and William Silverman for helpful discussions and reading the manuscript. Work in the lab of G.D. on pannexin channels is supported by National Institutes of Health (NIH) grant GM48610. Note added in proof: A reagent known to attenuate nucleotide release has been recently reported to inhibit pannexin but not connexin channels; Silverman W, Locovei S, Dahl GP. Proenecid, a gout remedy, inhibits pannexin 1 channels. Am J Physiol. 2008;295:C761-7.

References 1. Sohl ¨ G, Willecke K. Gap junctions and the connexin protein family. Cardiovasc Res. 2004;62:228–32. 2. Phelan P, Bacon JP, Davies JA, Phelan P, Bacon JP, Davies JA, Davies JA, Stebbings LA, Todman MG, Avery L, Baines RA, Barnes TM, Ford C, Hekimi S, Lee R, Shaw JE, Starich TA, Curtin KD, Sun YA and Wyman RJ. Innexins: a family of invertebrate gapjunction proteins. Trends Genet. 1998;14:348–9. 3. Panchin Y, Kelmanson I, Matz M, Lukyanov K, Usman N, Lukyanov S. A ubiquitous family of putative gap junction molecules. Curr Biol. 2000;10:R473–4. 4. Panchin YV. Evolution of gap junction proteins-the pannexin alternative. J Exp Biol. 2005;208:1415–9. 5. Locovei S, Bao L, Dahl G. Pannexin 1 in erythrocytes: function without a gap. Proc Natl Acad Sci USA. 2006;103:7655–9.

298

G. Dahl and A.L. Harris

6. Boassa D, Ambrosi C, Qiu F, Dahl G, Gaietta G, Sosinsky G. Pannexin1 channels contain a glycosylation site that targets the hexamer to the plasma membrane. J Biol Chem. 2007;282:31733–43. 7. Bruzzone R, Hormuzdi SG, Barbe MT, Herb A, Monyer H. Pannexins, a family of gap junction proteins expressed in brain. Proc Natl Acad Sci USA. 2003;100:13644–9. 8. Bao L, Locovei S, Dahl G. Pannexin membrane channels are mechanosensitive conduits for ATP. FEBS Lett. 2004;572:65–8. 9. Vanden Abeele F, Bidaux G, Gordienko D, Beck, B, Panchin, YV, Baranova, AV, Ivanov, DV, Skryma, R, Prevarskaya, N. Functional implications of calcium permeability of the channel formed by pannexin 1. J Cell Biol. 2006;174:535–46. 10. Lai CP, Bechberger JF, Thompson RJ, MacVicar BA, Bruzzone R, Naus CC. Tumorsuppressive effects of pannexin 1 in C6 glioma cells. Cancer Res. 2007;67:1545–54. 11. Huang Y, Grinspan JB, Abrams CK, Scherer SS. Pannexin1 is expressed by neurons and glia but does not form functional gap junctions. Glia. 2007;55:46–56. 12. Paul DL. Molecular cloning of cDNA for rat liver gap junction protein. J Cell Biol. 1986;103:123–34. 13. Bauer R, Lehmann C, Martini J, Eckardt F, Hoch M. Gap junction channel protein innexin 2 is essential for epithelial morphogenesis in the Drosophila embryo. Mol Biol Cell. 2004;15:2992–3004. 14. Dahl G, Locovei S. Pannexin: to gap or not to gap, is that a question? IUBMB Life. 2006;58:409–19. 15. Ransford GA, Dahl G, Fregien N, Conner GE, Salathe M. Pannexin 1 expression in airway epithelia. Am J Respir Crit Care Med. 2007;175:A887. 16. Zoidl G, Petrasch-Parwez E, Ray A, Meier C, Bunse S, Habbes H-W, Dahl G and Dermietzel R. Localization of the pannexin1 protein at postsynaptic sites in the cerebral cortex and hippocampus. Neuroscience. 2007;146:9–16. 17. Penuela S, Bhalla R, Gong XQ, Cowan KN, Celetti SJ, Cowan BJ, Bai D, Shao Q, Laird DW. Pannexin 1 and pannexin 3 are glycoproteins that exhibit many distinct characteristics from the connexin family of gap junction proteins. J Cell Sci. 2007;120:3772–83. 18. Dahl G, Nonner W, Werner R. Attempts to define functional domains of gap junction proteins with synthetic peptides. Biophys J. 1994;67:1816–22. 19. Thornhill WB, Levinson SR. Biosynthesis of electroplax sodium channels in Electrophorus electrocytes and Xenopus oocytes. Biochemistry. 1987;26:4381–8. 20. Goldberger G, Arnaout MA, Aden D, Kay R, Rits M, Colten HR. Biosynthesis and postsynthetic processing of human C3b/C4b inactivator (factor I) in three hepatoma cell lines. J Biol Chem. 1984;259:6492–7. 21. Mous JM, Peeters BL, Heyns WJ, Rombauts WA. Assembly, glycosylation, and secretion of the oligomeric rat prostatic binding protein in Xenopus oocytes. J Biol Chem. 1982;257:11822–8. 22. Baranova A, Ivanov D, Petrash N, Pestova A, Skoblov M, Kelmanson I. Shagin D, Nazarenko S, Geraymovych E, Litvin O, Tiunova A, Born T L, Usman N, Staroverov D, Lukyanov S and Panchin Y. The mammalian pannexin family is homologous to the invertebrate innexin gap junction proteins. Genomics. 2004;83:706–16. 23. Wang J, Ma M, Locovei S, Keane R, Dahl GP. Modulation of membrane channel currents by gap junction protein mimetic peptides: size matters. Am J Physiol Cell Physiol. 2007;293:C1112–9. 24. Locovei S, Wang J, Dahl G. Activation of pannexin 1 channels by ATP through P2Y receptors and by cytoplasmic calcium. FEBS Lett. 2006;580:239–44. 25. Sanderson MJ, Charles AC, Dirksen ER. Mechanical stimulation and intercellular communication increases intracellular Ca2+ in epithelial cells. Cell Regul. 1990;1:585–96. 26. Sigurdson WJ, Sachs F, Diamond SL. Mechanical perturbation of cultured human endothelial cells causes rapid increases of intracellular calcium. Am J Physiol. 1993;264:H1745–52.

12 Pannexin Channels

299

27. Newman EA. Propagation of intercellular calcium waves in retinal astrocytes and Mu¨ller cells. J Neurosci. 2001;21:2215–23. 28. Jorgensen NR, Geist ST, Civitelli R, Steinberg TH. ATP- and gap junction-dependent intercellular calcium signaling in osteoblastic cells. J Cell Biol. 1997;139:497–506. 29. Osipchuk Y, Cahalan M. Cell-to-cell spread of calcium signals mediated by ATP receptors in mast cells. Nature. 1992;359:241–44. 30. Hassinger TD, Guthrie PB, Atkinson PB, Bennett MVL, Kater SB. An extracellular signaling component in propagation of astrocytic calcium waves. Proc Natl Acad Sci USA. 1996;93:13268–73. 31. Bergfeld GR, Forrester T. Release of ATP from human erythrocytes in response to a brief period of hypoxia and hypercapnia. Cardiovasc Res. 1992;26:40–7. 32. Stout CE, Costantin JL, Naus CC, Charles AC. Intercellular calcium signaling in astrocytes via ATP release through connexin hemichannels. J Biol Chem. 2002;277:10482–8. 33. Zhao HB, Yu N, Fleming CR. Gap junctional hemichannel-mediated ATP release and hearing controls in the inner ear. Proc Natl Acad Sci USA. 2005;102:18724–9. 34. Hofer A, Dermietzel R. Visualization and functional blocking of gap junction hemichannels (connexons) with antibodies against external loop domains in astrocytes. Glia. 1998;24:141–154. 35. Leybaert L, Braet K, Vandamme W, Cabooter L, Martin PE, Evans WH. Connexin channels, connexin mimetic peptides and ATP release. Cell Commun Adhes. 2003;10:251–7. 36. Cotrina ML, Lin JH, Alves-Rodrigues A, Liu S. Li J, Azmi-Ghadimi H, Kang J, Naus CC and Nedergaard M. Connexins regulate calcium signaling by controlling ATP release. Proc Natl Acad Sci USA. 1998;95:15735–40. 37. Locovei S, Scemes E, Qiu F, Spray DC, Dahl G. Pannexin1 is part of the pore forming unit of the P2X7 receptor death complex. FEBS Lett. 2007;581:483–8. 38. Pelegrin P, Surprenant A. Pannexin-1 mediates large pore formation and interleukin-1b release by the ATP-gated P2X7 receptor. EMBO J. 2006;25:5071–82. 39. Musil LS, Goodenough DA. Biochemical analysis of connexin43 intracellular transport, phosphorylation, and assembly into gap junctional plaques. J Cell Biol. 1991;115:1357–74. 40. Musil LS, Goodenough DA. Multisubunit assembly of an integral plasma membrane channel protein, gap junction connexin43, occurs after exit from the ER. Cell. 1993;74:1065–77. 41. Ebihara L, Liu X, Pal JD. Effect of external magnesium and calcium on human connexin46 hemichannels. Biophys J. 2003;84:277–86. 42. Ebihara L. Physiology and biophysics of hemi-gap-junctional channels expressed in Xenopus oocytes. Acta Physiol Scand. 2003;179:5–8. 43. Peracchia C, Wang X, Li L, Peracchia LL. Inhibition of calmodulin expression prevents low-pH-induced gap junction uncoupling in Xenopus oocytes. Pflu¨gers Arch Eur J Physiol. 1996;431:379–87. 44. Spray DC, Harris AL, Bennett MVL. Comparison of pH and calcium dependence of gap junctional conductance. Kroc Found Ser. 1981;15:445–61. 45. Trexler EB, Bennett MVL, Bargiello TA, Verselis VK. Voltage-gating and permeation in a gap junction hemichannel. Proc Natl Acad Sci USA. 1996;93:5836–41. 46. Trexler EB, Verselis VK. The study of connexin hemichannels (connexons) in Xenopus oocytes. Methods Mol Biol. 2001;154:341–55. 47. Pfahnl A, Zhou XW, Werner R, Dahl G. A chimeric connexin forming gap junction hemichannels. Pflu¨gers Arch. 1997;433:773–9. 48. Pfahnl A, Dahl G. Gating of Cx46 gap junction hemichannels by calcium and voltage. Pflu¨gers Arch. 1999;437:345–53. 49. Li H, Liu TF, Lazrak A, Peracchia C, Goldberg GS, Lampe PD and Johnson RG. Properties and regulation of gap junctional hemichannels in the plasma membranes of cultured cells. J Cell Biol. 1996;134:1019–30.

300

G. Dahl and A.L. Harris

50. Stout C, Charles A. Modulation of intercellular calcium signaling in astrocytes by extracellular calcium and magnesium. Glia. 2003;43:265–73. 51. Ebihara L, Steiner E. Properties of a nonjunctional current expressed from a rat connexin46 cDNA in Xenopus oocytes. J Gen Physiol. 1993;102:59–74. 52. Zanotti S, Charles A. Extracellular calcium sensing by glial cells: low extracellular calcium induces intracellular calcium release and intercellular signaling. J Neurochem. 1997;69:594–602. 53. Sprague RS, Ellsworth ML, Stephenson AH, Kleinhenz ME, Lonigro AJ. Deformationinduced ATP release from red blood cells requires CFTR activity. Am J Physiol. 1998;275:H1726–32. 54. Huang YJ, Maruyama Y, Dvoryanchikov G, Pereira E, Chaudhari N, Roper SD. The role of pannexin 1 hemichannels in ATP release and cell-cell communication in mouse taste buds. Proc Natl Acad Sci USA. 2007;104:6436–41. 55. Romanov RA, Rogachevskaja OA, Bystrova MF, Jiang P, Margolskee RF, Kolesnikov SS. Afferent neurotransmission mediated by hemichannels in mammalian taste cells. EMBO J. 2007;26:657–67. 56. Scemes E, Dermietzel R, Spray DC. Calcium waves between astrocytes from Cx43 knockout mice. Glia. 1998;24:65–73. 57. Iacobas DA, Scemes E, Spray DC. Gene expression alterations in connexin null mice extend beyond the gap junction. Neurochem Int. 2004;45:243–50. 58. Iacobas DA, Iacobas S, Urban-Maldonado M, Spray DC. Sensitivity of the brain transcriptome to connexin ablation. Biochim Biophys Acta. 2005;1711:183–96. 59. Iacobas DA, Iacobas S, Spray DC. Connexin43 and the brain transcriptome of newborn mice. Genomics. 2007;89:113–23. 60. Bahima L, Aleu J, Elias M, Martin-Satue M, Muhaisen A, Blasi J, Marsal J and Solsona C. Endogenous hemichannels play a role in the release of ATP from Xenopus oocytes. J Cell Physiol. 2006;206:95–102. 61. Maroto R, Hamill OP. Brefeldin A block of integrin-dependent mechanosensitive ATP release from Xenopus oocytes reveals a novel mechanism of mechanotransduction. J Biol Chem. 2001;276:23867–72. 62. Brule G, Haudecoeur G, Jdaiaa H, Guilbault P. Inhibition, by an amphiphilic substance, niflumic acid, of the inward rectification of the crustacean muscle fiber. Arch Int Physiol Biochim. 1983;91:269–77. 63. Benoit E, Corbier A, Dubois JM. Evidence for two transient sodium currents in the frog node of Ranvier. J Physiol. 1985;361:339–60. 64. Vessey JP, Lalonde MR, Mizan HA, Welch NC, Kelly ME, Barnes S. Carbenoxolone inhibition of voltage-dependent Ca channels and synaptic transmission in the retina. J Neurophysiol. 2004;92:1252–56. 65. Guan X, Cravatt BF, Ehring GR,, Hall JE, Boger DL, Lerner RA, Gilula NB. The sleepinducing lipid oleamide deconvolutes gap junction communication and calcium wave transmission in glial cells. J Cell Biol. 1997;139:1785–92. 66. Braet K, Vandamme W, Martin PE, Evans WH, Leybaert L. Photoliberating inositol1,4,5-trisphosphate triggers ATP release that is blocked by the connexin mimetic peptide gap 26. Cell Calcium. 2003;33:37–48. 67. Purnick PE, Oh S, Abrams CK, Verselis VK, Bargiello TA. Reversal of the gating polarity of gap junctions by negative charge substitutions in the N-terminus of connexin 32. Biophys J. 2000;79:2403–15. 68. Surprenant A, Rassendren F, Kawashima E, North RA, Buell G. The cytolytic P2Z receptor for extracellular ATP identified as a P2X receptor (P2X7). Science. 1996;272:735–8. 69. Suadicani SO, Brosnan CF, Scemes E. P2X7 receptors mediate ATP release and amplification of astrocytic intercellular Ca2+ signaling. J Neurosci. 2006;26:1378–85.

12 Pannexin Channels

301

70. Bruzzone R, Barbe MT, Jakob NJ, Monyer H. Pharmacological properties of homomeric and heteromeric pannexin hemichannels expressed in Xenopus oocytes. J Neurochem. 2005;92:1033–43. 71. Coutinho-Silva R, Persechini PM. P2Z purinoceptor-associated pores induced by extracellular ATP in macrophages and J774 cells. Am J Physiol. 1997;273:C1793–C1800. 72. Bao L, Sachs F, Dahl G. Connexins are mechanosensitive. Am J Physiol Cell Physiol. 2004;287:89–95. 73. De Vuyst E, Decrock E, Cabooter L, Dubyak GR, Naus CC, Evans WH and Leybaert L. Intracellular calcium changes trigger connexin 32 hemichannel opening. EMBO J. 2006;25:34–44. 74. Paul DL, Ebihara L, Takemoto LJ, Swenson KI, Goodenough DA. Connexin46, a novel lens gap junction protein, induces voltage-dependent currents in nonjunctional plasma membrane of Xenopus oocytes. J Cell Biol. 1991;115:1077–89.

Part II

Connexins in Organ Systems and Processes

Chapter 13

Foreword: Gap Junctions and Emergent Organ Properties Daniel Goodenough

Keywords Emergent properties Tissue-specific processes Unanswered questions

Gap junctional intercellular communication (GJIC) is virtually universal between adjacent cells in multicellular organisms. GJIC involves gated intercellular channels that restrict cell–cell exchange of ions and small molecules, preserving the macromolecular identity of each cell. Other examples of direct intercellular exchange are the cytoplasmic continuity among plant cells, plasmodesmata, the retention of midbodies in developing spermatozoa as a result of incomplete cytokinesis, and the fusion of cells seen in skeletal muscle, syncytial trophoblast, fertilization, osteoclasts, and foreign-body giant cells. Since connexins are involved in direct cell–cell interaction, they may also have signaling functions beyond their channel-forming properties. For example, docking of hemichannels into an intercellular channel may result in transient interactions of intracellular proteins with the connexin cytoplasmic domains. This may result in signaling to the cytoskeleton or to the nucleus that another cell has been joined, thus participating in the phenomena of contact-inhibition of cell movement and growth. Cell–cell communication provided by gap junctions is very old in the evolutionary history of metazoans, and has provided speed, synchrony, switching, symbiosis, and stimulus/suppression to diverse functions of cell collectives. The discovery of the connexin gene family and subsequent mapping to the different cell types of different organs has opened experimental avenues into the elucidation of gap junction function in the physiology of the whole organism. Targeted and conditional gene ‘‘knockouts’’ and ‘‘knock-ins’’ have provided key information about the essential nature of connexin function in many organs. In the D. Goodenough (*) Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston MA, 02445, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_13, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

305

306

D. Goodenough

chapters in this section, there are links to disease, development and ionic homeostasis in the integument, bone, special senses, central nervous system, and exocrine glands; in cardiovascular, respiratory, and reproductive systems; and in carcinogenesis. In all cases, connexins are clearly essential for selected functions, although other functions are undoubtedly masked by redundancy. While each chapter brings us to the edge of current knowledge, most authors acknowledge, where appropriate, that the specific molecular ‘‘mechanism’’ for the connexin-influenced diseases or developmental events is still lacking. This conceptual and experimental challenge is one of systems complexity, where the biological function is not cell autonomous. Gap junction functions are often involved in emergent properties, aspects of tissue and organ biology that cannot be readily extrapolated from the properties or responses of individual cells. Tantalizing examples of these functions include ionic currents that result from interconnecting chains or sheets of cells with different or changing resting potentials, clearly seen in myocardium and the nervous system, but also evident in the eye lens and developing embryos. Does the direction of these currents permit a collective sense of left/right symmetry? Do adjacent tissues sense the fields generated by this current and derive positional information? Less clearly defined is the ability of stem cells to sense the volume of the whole organ and thus adjust their mitotic rate. Could this be achieved by the degree of dissipation of an intracellular junction-permeable small molecule signal? Are there gradients of nutrient pools along the villus/crypt axis in the intestine, or along the cords of hepatocytes in a liver lobule that can result in altered gene expression? These properties of collectives of cells constitute organ function at a supracellular level, not easily explained by identification of particular amino acids at the mouth of the pore, or even by studying signaling cascades within a single cell. The following chapters set the stage for this next level of understanding. The future challenges to students of GJIC are indeed interesting and complex.

Chapter 14

Connexins in Skin Biology Trond Aasen and David P. Kelsell

Abstract The skin is the largest organ of the body and exerts a variety of functions ranging from barrier homeostasis to the sense of touch. A variety of inherited skin diseases are associated with mutation in connexin genes. In addition, there is strong connexin involvement in skin cancer and in wound healing. Our understanding of the role of connexins in epidermal biology is rapidly developing through the use of several in vitro cell culture systems and of transgenic mouse models, in concert with the expression of disease-associated mutant connexin proteins. Keywords Skin Epidermis Gap junction Keratinocytes Keratitisichthyosis deafness Hystrix-like ichthyosis-deafness Clouston syndrome Vohwinkel0 s syndrome Erythrokeratodermia variabilis Hidrotic ectodermal dysplasia Oculodentodigital dysplasia Basal cell carcinoma Squamous cell carcinoma Cx26 Cx30 Cx30.3 Cx31 Cx31.1 Cx32 Cx37 Cx40 Cx43 Cx45

14.1 Introduction The skin is a complex organ arranged in three layers; the mesodermal-derived hypodermis and dermis, which are highly collagenous and fibroblast rich, and the ectodermal-derived outer covering of skin called the epidermis (and its associated eccrine sweat glands and hair follicles), which consists mainly of stratifying keratinocytes. The epidermis also includes other cells such as Langerhans’ cells, melanocytes and Merkel cells (for excellent recent reviews see [1,2]). D.P. Kelsell (*) Centre for Cutaneous Research, Institute of Cell and Molecular Science, Barts and the London, Queen Mary University of London, 4 Newark Street, Whitechapel, London E1 4AT, United Kingdom e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_14, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

307

308

T. Aasen and D.P. Kelsell

Early electron microscopy studies demonstrated the existence of extensive gap junctions in skin. The expression pattern of connexins in the epidermis has high temporal and spatial specificity, consistent with studies showing extensive but compartmentalized dye transfer between keratinocytes in murine epidermis [3,4]. To date, at least ten connexin genes (those coding for CX26, CX30, CX30.3, CX31, CX31.1, CX32, CX37, CX40, CX43, and CX45) have been shown to be expressed in human skin [5], whereas nine connexins appear to be expressed in murine epidermis (same connexins as in human except Cx32 [4,6,7,8,9,10,11]. Connexins, in particular CX43, are also expressed in the dermis of the skin [8,12,13], but their role in fibroblasts is not discussed further here. This chapter reviews the role of connexins in skin biology, focusing on epidermal connexins and the keratinocyte disorders associated with inherited connexin mutation.

14.2 Connexins and Skin There are both unique and extensive overlapping expressions of multiple connexins in the human epidermis (Fig. 14.1). In general, connexin expression is low in the basal layer, with potential stem cells reported to lack gap junctions [14], but it increases as keratinocytes differentiate into the spinous and granular layers. The dead, flattened, and anucleated corneocytes that shed from the skin surface are gap junction–deficient. The expression pattern varies dramatically depending on a number of factors such as site and previous exposure to trauma (see below). For example, CX26 and CX30 are expressed at low levels in interfollicular basal keratinocytes and in keratinocytes of the hair follicle [5,15], whereas in glabrous skin (hairless skin of the palms and soles) CX26 and CX30 are found at high levels in all the living layers of the epidermis [15,16,17]. CX43 is expressed at particularly high levels in the spinous and granular layers (Fig. 14.2a) [18,19]. However, there are clear differences in connexin expression between species; CX43 is expressed mainly in the basal layers of murine epidermis, whereas CX43 is expressed primarily in differentiating keratinocytes in human epidermis [6,20,21,22]. Therefore, one must be cautious in assessing specific connexin functions in mouse skin in relation to human epidermal biology. A more clearly demarcated expression in human epidermis is seen with CX30.3 and CX31, which are found only in late spinous and (predominantly) the granular layer (Fig. 14.2b) [5], as well as in the keratinocytes of the hair follicle. Thus, the expression pattern of connexins in skin is well defined, but the functions of this plethora of gene products remain largely unknown, although their significance is obvious given the number of connexin genodermatoses described below.

14 Skin

309

Fig. 14.1 Schematic drawing of human interfollicular skin. The typical structural units of upper skin with the epidermis subdivided into four differentiating layers of keratinocytes. The level of connexin expression is indicated in the different layers, with ‘‘+’’ indicating increasing levels of expression. Expression of connexins varies among body sites and upon exposure to the physical environment. For example, CX26 and CX30 are expressed at much higher levels in all layers in palmoplantar epidermis or in epidermis exposed to trauma. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

Fig. 14.2 Connexin expression in human interfollicular epidermis. Examples of CX31 (a) and CX43 (b) expression in human interfollicular epidermis. Note in (a) that some CX31 gap junctions are found in the basal and spinous layers (particularly with CX43) but the major portion of expression is seen in the granular layer. Some diffuse nonspecific staining of basement membrane and cornified layer is seen in (b). (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

310

T. Aasen and D.P. Kelsell

14.3 Connexin Genodermatoses Five connexin isoforms are currently associated with genodermatoses (inherited skin diseases): CX26, CX30, CX30.3, CX31, and CX43 [23,24]. In general, most mutations are autosomal dominant, but recessive CX31 mutations have also been described [25]. All connexin genodermatoses have an underlying hyperkeratosis (thickening) of the skin, but other epidermal phenotypes vary greatly even among mutations within one connexin gene. An example of the complexity of these connexin genodermatoses is seen with mutations in the GJB2 gene, which encodes CX26. Recessive GJB2 mutations are the most common cause of nonsyndromic deafness, but dominant GJB2 mutations can cause syndromic diseases involving the skin, including simple palmoplantar keratoderma (thickening of the skin, primarily the palms of the hands and soles of the feet) with various degrees of hearing impairment, keratitis-ichthyosis-deafness (KID) syndrome [26,27,28,29] with profound hearing loss, and the similar (yet genetically identical) hystrix-like ichthyosisdeafness (HID [30]), which in addition to deafness and severe eye complications has a more generalized skin disorder. Both KID and HID patients have an increased risk of skin cancer. Another CX26 mutation (N14K) causes hypotrichosis (less than normal amount of hair) and nail dystrophy [31] that closely resembles Clouston syndrome (see below). The D66H mutation [32] and more recently the G59S mutation [33] in CX26 have been identified in patients with Vohwinkel´s syndrome, a ‘‘starfish’’ pattern of keratoderma associated with mild to moderate hearing impairment and pseudoainhum (annular constrictions of the digits), which can lead to autoamputation. Mutations in GJB6 encoding CX30 are also associated with Clouston syndrome (hidrotic ectodermal dysplasia, HED), an autosomal dominant skin disorder characterized by palmoplantar hyperkeratosis, hair defects (partial to total hair loss), incomplete or arrested nail development (hypoplasia), and nail deformities [34,35]. However, it is evident that considerable phenotypic variability occurs here as well, with the CX30G11R and CX30A88V mutations also found in patients with pachyonychia congenita (a rare form of palmoplantar keratoderma characterized by abnormally thickened nails [36]). In addition, GJB6 mutations have been found in KID patients [37]. Mutations in the similar CX30.3 and CX31 genes cause erythrokeratodermia variabilis (EKV), a disease with persistent hyperkeratosis and an extensive inflammatory response resulting in migratory erythema, an abnormal redness of the skin [38,39,40,41,42,43]. Thus, the interplay between connexin isoforms is clearly important, as one connexin can cause several disorders and mutations in two different connexins can underlie the same disease. The involvement of one connexin in the function of many organs, and thus syndromic disease if mutated, is perhaps best illustrated by mutations in GJA1, which encodes CX43. Certain dominant GJA1 mutations cause oculodentodigital dysplasia (ODDD [44,45,46]), which is a pleiotropic congenital disorder characterized by abnormalities of the face, eyes,

14 Skin

311

dentition, and limbs. Occasionally, palmoplantar keratoderma has been reported as well as curly hair and hypotrichosis, but only in patients with mutations in the carboxyl-terminal domain (CT) of CX43 [47,48,49]. The infrequent and mild involvement of skin disease in ODDD is perhaps surprising considering the extensive expression pattern of CX43 in both the dermis and epidermis of skin.

14.4 Role of Connexins in Other Skin Conditions Connexins are clearly important in skin tissue homeostasis as demonstrated by the numerous inherited genotype-phenotype correlations described above. Thus, perhaps not surprisingly, their expression pattern is deregulated in hyperproliferative skin diseases including psoriasis and other pathological conditions such as skin cancer and wound healing.

14.4.1 Skin Cancer The two most common types of skin cancer, basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), show a dramatic upregulation of CX26 and CX30 expression [50]. A structural loss of CX43 gap junctions in BCC has also been reported [19]. Furthermore, there are distinct differences in the connexin expression pattern between the tumor subtypes. CX30.3 is usually expressed in differentiating keratinocytes of skin and is lost in undifferentiated BCCs but not in more differentiated SCC samples. CX32 is also induced in BCCs but not in SCC. Although the significance remains unclear, these findings indicate that connexins play a role in the differentiation status of the tumor. In many cases it has been demonstrated that connexins can act as tumor suppressors (for examples, see [51,52,53,54,55,56,57,58]; see also Chapter 27), and for skin cancer this notion may (or may not, depending on theories of activating mutations, see below) be supported by the fact that KID patients have a high chance of developing SCCs [23,59]. This may be a consequence of the repeated trauma and scarring of the skin in these patients. Recent in vitro data also suggest that CX43 reduces proliferation in basaloid SCC cells [60]. Data should be interpreted carefully, as there are a number of conflicting reports on connexins in cancer biology (see Chapter 27). For example, there is evidence of connexins facilitating aspects of tumorigenesis, including tissue invasion [61,62]. For skin this is particularly relevant to the highly metastatic melanoma tumors. In vitro models have suggested that expression of Cx26 in melanoma cells can aid metastatic tumor invasion in mice [63]. Intriguingly, expression of wild-type Cx26 in differentiating keratinocytes under the regulation of the involucrin promoter in a transgenic mouse model leads to persistent epidermal hyperproliferation in wounded epidermis [64]. This also links with

312

T. Aasen and D.P. Kelsell

the observation that hyperproliferative disorders such as psoriasis and warts have very high levels of CX26 and CX30 [16].

14.4.2 Wound Healing Connexins are regulated in a highly specific spatial and temporal pattern in skin during wound healing. Early studies in rats showed strong induction of Cx26 but downregulation of Cx31.1 and Cx43 as early as 2 hours post-wounding of the skin [21]. Coutinho et al. [8] noted downregulation of all connexins studied (Cx26, Cx30, Cx31.1, and Cx43) in murine epidermis and dermis as an immediate response (6 hours post-wounding), followed by dramatic changes in the connexin expression pattern. This suggested that Cx26 and Cx30 were associated with keratinocyte migration, Cx43 with proliferation, and Cx31.1 with differentiation, highlighting the spectrum of putative connexin functions in skin. Detailed work by Brandner et al. [65] in human skin (and in an in vitro skin model) demonstrated elegantly the initial loss of CX43 expression at the wound margin, whereas CX43 remained present in nonhealing wounds (chronic leg ulcers). They also observed subsequent induction of CX26 and CX30, but with negative staining of the wound margin until commencement of epidermal regeneration. Interestingly, CX26 and CX30 were present at the wound margins of most nonhealing wounds. Two distinct studies have now directly demonstrated the potential regulatory role that Cx43 may play during wound healing. Qiu et al. [66] demonstrated accelerated wound closure and healing in a mouse model using direct application of interfering antisense oligodeoxyribonucleotides (RNAi) to transiently downregulate Cx43 expression. Furthermore, earlier wound closure has also been observed in mice with conditional Cx43 knockout (KO) in skin [22]. These models suggest that downregulation of Cx43 can dramatically improve wound healing, thus offering tantalizing potential avenues for therapeutic research. It should be noted that no studies have been performed in humans, and the differences in Cx43 expression between murine and human skin may suggest different biological roles for Cx43 in the skin.

14.5 In Vitro and Animal Models In vivo observations of normal and pathological skin specimens have aided the descriptive analysis of connexins in skin. However, direct experimental models are required for the exact functional elucidation of different connexins in skin. Three main model systems have been utilized thus far: 1. Classical gap junction studies using in vitro models (mammalian cell culture and Xenopus oocyte assays) have been particularly informative. For example, cloning and expression of connexin mutants associated with skin disease

14 Skin

313

have revealed important information on protein localization including transinteractions with other proteins, as well as gap junction and hemichannel activities of mutant versus wild-type connexins [67]. Common et al. [68,69,70,71] noticed that dominant connexin mutations associated with skin disease caused an increase in cell death when expressed in keratinocytes, whereas those mutations in the same connexin that were associated only with deafness actually reduced keratinocyte cell death compared to wild-type connexins. Together with the observations of hemichannel activity of connexins, this has led to important steps toward explaining the observations of ‘‘one gene, several diseases.’’ 2. More specific to skin research is the employment of various human cell and organotypic models that aim to reconstitute the epidermis (and in some cases the dermis) in vitro. For an example of an organotypic model recapitulating human epidermis with differentiation-dependent expression of CX30.3 and CX31, see Fig. 14.3. These models have provided interesting findings on the function of connexins in a variety of important aspects of keratinocyte biology such as stratification, differentiation, invasion, and tumor biology [72]. In addition to investigating tumor-related functions of connexins, organotypic models have also shown promise for mutationspecific analysis [73] and indeed for tackling the unknown function of wild-type connexins [74]. 3. Several connexin mouse knockout models have been generated, although they are often either embryonically lethal or without obvious phenotype. Thus, simple knockout of the skin disease–associated connexins (Cx26, Cx30, and Cx31, or double knockouts of Cx30 and Cx31 or Cx31 and Cx43) result in no obvious skin phenotype [22,75,76]. It is thought that there is extensive functional redundancy among subclasses of connexins within the epidermis. Humans with recessive CX26 mutations, as well as having profound hearing loss, display a slightly thickened epidermis [77], a

Fig. 14.3 Histology and endogenous expression of CX31 and CX30.3 in organotypic raft cultures. Immortalized human keratinocytes grown for 14 days at the air–liquid interphase. Note that connexin expression is restricted to differentiated layers, similar to that observed in normal skin. (a) Haematoxylin and eosin stain. (b) CX31 expression. (c) CX30.3 expression (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

314

T. Aasen and D.P. Kelsell

feature that has been replicated in vitro using skin culture organotypic systems, whereas Cx26KO mice show embryonic lethality [78]. Transgenic expression using conditional or epidermal-specific promoters have led to interesting observations. Bakirtzis et al. [79,80] expressed the Vohwinkel´s syndrome–associated CX26 mutation, D66H, in mouse epidermis under control of a keratin promoter. This elegantly recapitulated pathological features of the human disease including hyperkeratosis and the formation of pseudoainhum of digits and tail. In another study, expression of a truncated version of Cx43 (M257) caused death in 97% of mice shortly after birth due to defective epidermal barrier function [81]. Ectopic expression of wild-type connexins has also been informative. Djalilian et al. [64] expressed wild-type Cx26 in mouse using the promoter of the epidermal-specific involucrin promoter. They demonstrated that downregulation of Cx26 is required for barrier acquisition during development, whereas persistent Cx26 expression kept wounded epidermis in a hyperproliferative state. The results were explained by the observation that ectopic expression of Cx26 in keratinocytes resulted in increased adenosine triphosphate release, which delayed epidermal barrier recovery and promoted an inflammatory response in resident immune cells. Again, an element of caution is required as this Cx26 expression was driven by the promoter of a different gene, with a different expression pattern.

14.6 Key Scientific Questions and Challenges Despite great progress, many questions remain regarding connexins in the skin. Further genotype-phenotype correlations are likely to be made in variants of disorders of keratinization already recognized, and perhaps more unexpected novel observations will be made. In addition, studies in other animal models may offer clues. For example, the spotted skin pigment pattern seen in the classic ‘‘leopard’’ mutant in zebrafish is caused by mutations in the zebrafish Cx41.8 gene, an ortholog of human CX40, which is expressed in skin [82]. No connexins have so far been associated with a human pigmentation disorder, but the zebrafish offers an excellent model for study of connexin function/ dysfunction. Key questions remain: What is the exact role of connexins in skin biology and why are such an array of connexins expressed in overlapping, but distinct, patterns? As discussed, a number of results demonstrate that connexins play key roles in the regulation of keratinocyte proliferation and differentiation. They may also have undiscovered roles in other processes, including stem cell regulation, immune defense recognition, compartmentalization, and skin barrier function. Elucidating these questions and hypotheses is an exciting and important challenge. There are a number of important but not mutually exclusive approaches. For example, obtaining information from mouse knockout models has been

14 Skin

315

hampered by connexin redundancy and embryonic lethality, but technical advances have made it relatively easy to generate tissue-specific knockout and double knockouts. The discovery of connexin mutations in a variety of tissues will also be useful for generating transgenic mice, as already nicely demonstrated with the CX26D66H Vohwinkel’s syndrome mouse [79,80]. As yet, no unifying vision of how connexin mutations cause skin disease has emerged. We still do not understand the functional utility of different permeability properties among different connexin isoforms (see Chapter 7), and what this means biologically in the epidermis where keratinocytes express multiple connexins. One aspect requiring further attention is the difference between loss-offunction versus gain-of-function mutations, as well as transdominant interactions, which might shed light on the wide variety of phenotypes within connexin-specific mutations. In this respect, the choice of model system used for investigating these mutations in vitro is also important. For example, it was reported that CX26D66H and CX2642E interact transdominantly with wildtype CX43 in Xenopus oocytes [17]; however, this does not appear to be the case when using human keratinocyte cell lines [80,83]. Furthermore, in syndromic skin disorders it is noted that the same mutation can have highly variable phenotype/penetrance, which indicates the importance of other genetic or environmental factors in determination of disease severity. For example, the CX26R75W mutation has been reported in several patients with profound hearing impairment but with varying degrees of skin symptoms. This implicates the potential involvement or interaction of other proteins. Indeed, in several cases, mutations discovered in other genes share phenotypes with connexin genodermatoses. For example, a variant of Vohwinkel´s syndrome (without hearing loss) is due to mutation in the loricrin gene, which has a function in late keratinocyte differentiation [84,85,86]. Indeed, there have been several cases of EKV and variations of HID/KID syndrome where no connexin mutations have been found. Identifying the genetic basis of the disease in these families could lead to the discovery of novel protein interactions or signaling pathways involving connexins. Direct searches for putative connexin interacting proteins may also be fruitful. One such protein reportedly interacting with Cx26 is YAF-2 (Y1-associated factor 2), which has been implicated in transforming growth factor-b (TGF-b) signaling and plays an important role in skin tissue homeostasis and disease [23]. Traditional cell culture as well as organotypic skin models and mouse transgenic studies will continue as useful models, particularly with the advances of RNAi and inducible constructs. However, it will become increasingly important to couple these techniques with more powerful genomic and proteomic tools using both wild-type and mutant constructs to elucidate the complex tissue-specific signaling pathways linked to connexins via either gap junction or non–gap junction functions [74]. Finally, as many genotype-phenotype correlations have already been made, the possibilities of skin therapy via connexin modulation should be considered. Indeed, due to connexin redundancy, eliminating or reducing the expression of

316

T. Aasen and D.P. Kelsell

mutant connexins such as Cx26 and Cx31 by RNAi or other techniques poses a tantalizing way of eliminating a disease-causing gene without having to consider replacement of a functional wild-type gene. With the easy access of skin (e.g., creams to knock down gene expression) and a possible direct and instant visible result, it is a field of high promise both for mutation-specific disorders such as EKV and as therapeutic approaches to much more common diseases such as psoriasis, where connexins may be significant modifiers.

14.7 Conclusion Multiple connexins are present in the human epidermis but their specific functions are still to be determined. However, their key roles in the skin are highlighted by the findings that their mutation causes inherited skin diseases. The investigation of specific connexin function in the epidermis using human disease-associated connexin mutants in conjunction with in vitro skin models and specific transgenic mouse models is revealing insights into the role connexins play in keratinocyte differentiation, proliferation, and epidermal tissue repair. Acknowledgment The authors acknowledge Dr. Wei Li Di (Centre for Cutaneous Research) for contributions to Fig. 14.1.

References 1. Fuchs E. Scratching the surface of skin development. Nature. 2007;445:834–842. 2. Kanitakis J. Anatomy, histology and immunohistochemistry of normal human skin. Eur J Dermatol. 2002;12:390–9. 3. Choudhry R, Pitts JD, Hodgins MB. Changing patterns of gap junctional intercellular communication and connexin distribution in mouse epidermis and hair follicles during embryonic development. Dev Dyn. 1997;210:417–30. 4. Brissette JL, Kumar NM, Gilula NB, Hall JE, Dotto GP. Switch in gap junction protein expression is associated with selective changes in junctional permeability during keratinocyte differentiation. Proc Natl Acad Sci USA. 1994;91:6453–57. 5. Di WL, Rugg EL, Leigh IM, Kelsell DP. Multiple epidermal connexins are expressed in different keratinocyte subpopulations including connexin 31. J Invest Dermatol. 2001;117:958–64. 6. Goliger JA, Paul DL. Expression of gap junction proteins Cx26, Cx31.1, Cx37, and Cx43 in developing and mature rat epidermis. Dev Dyn. 1994;200:1–13. 7. Risek B, Klier FG, Gilula NB. Multiple gap junction genes are utilized during rat skin and hair development. Development. 1992;116:639–51. 8. Coutinho P, Qiu C, Frank S, Tamber K, Becker D. Dynamic changes in connexin expression correlate with key events in the wound healing process. Cell Biol Int. 2003;27:525–41. 9. Butterweck A, Elfgang C, Willecke K, Traub O. Differential expression of the gap junction proteins connexin45, -43, -40, -31, and -26 in mouse skin. Eur J Cell Biol. 1994;65:152–63.

14 Skin

317

10. Hennemann H, Dahl E, White JB, Schwarz HJ, Lalley PA, Chang S, Nicholson BJ, Willecke K. Two gap junction genes, connexin 31.1 and 30.3, are closely linked on mouse chromosome 4 and preferentially expressed in skin. J Biol Chem. 1992;267:17225–33. 11. Willecke K, Heynkes R, Dahl E, Stutenkemper R, Hennemann H, Jungbluth S, Suchyna T, Nicholson BJ. Mouse connexin37: cloning and functional expression of a gap junction gene highly expressed in lung. J Cell Biol. 1991;114:1049–1057. 12. Langevin HM, Cornbrooks CJ, Taatjes DJ. Fibroblasts form a body-wide cellular network. Histochem Cell Biol. 2004;122:7–15. 13. Moyer KE, Ehrlich HP. Modulation of human fibroblast gap junction intercellular communication by hyaluronan. J Cell Physiol. 2003;196:165–70. 14. Matic M, Evans WH, Brink PR, Simon M. Epidermal stem cells do not communicate through gap junctions. J Invest Dermatol. 2002;118:110–6. 15. Di WL, Common JE, Kelsell DP. Connexin 26 expression and mutation analysis in epidermal disease. Cell Commun Adhes. 2001;8:415–8. 16. Lucke T, Choudhry R, Thom R, Selmer IS, Burden AD, Hodgins MB. Upregulation of connexin 26 is a feature of keratinocyte differentiation in hyperproliferative epidermis, vaginal epithelium, and buccal epithelium. J Invest Dermatol. 1999;112:354–61. 17. Rouan F, White TW, Brown N, Taylor AM, Lucke TW, Paul DL, Munro CS, Uitto J, Hodgins MB, Richard G. Trans-dominant inhibition of connexin-43 by mutant connexin-26: implications for dominant connexin disorders affecting epidermal differentiation. J Cell Sci. 2001;114:2105–13. 18. Salomon D, Masgrau E, Vischer S, Ullrich S, Dupont E, Sappino P, Saurat JH, Meda P. Topography of mammalian connexins in human skin. J Invest Dermatol. 1994;103:240–7. 19. Tada J, Hashimoto K. Ultrastructural localization of gap junction protein connexin 43 in normal human skin, basal cell carcinoma, and squamous cell carcinoma. J Cutan Pathol. 1997;24:628–35. 20. Kamibayashi Y, Oyamada M, Oyamada Y, Mori M. Expression of gap junction proteins connexin 26 and 43 is modulated during differentiation of keratinocytes in newborn mouse epidermis. J Invest Dermatol. 1993;101:773–8. 21. Goliger JA, Paul DL. Wounding alters epidermal connexin expression and gap junctionmediated intercellular communication. Mol Biol Cell. 1995;6:1491–501. 22. Kretz M, Euwens C, Hombach S, Eckardt D, Teubner B, Traub O, Willecke K, Ott T. Altered connexin expression and wound healing in the epidermis of connexin-deficient mice. J Cell Sci. 2003;116:3443–52. 23. van Steensel MA. Gap junction diseases of the skin. Am J Med Genet. 2004;131C:12–19. 24. Richard G. Connexin disorders of the skin. Clin Dermatol. 2005;23:23–32. 25. Gottfried I, Landau M, Glaser F, Di WL, Ophir J, Mevorah B, Ben-Tal N, Kelsell DP, Avraham KB. A mutation in GJB3 is associated with recessive erythrokeratodermia variabilis (EKV) and leads to defective trafficking of the connexin 31 protein. Hum Mol Genet. 2002;11:1311–16. 26. Yotsumoto S, Hashiguchi T, Chen X, Ohtake N, Tomitaka A, Akamatsu H, Matsunaga K, Shiraishi S, Miura H, Adachi J, Kanzaki T. Novel mutations in GJB2 encoding connexin26 in Japanese patients with keratitis-ichthyosis-deafness syndrome. Br J Dermatol. 2003;148:649–53. 27. Alvarez A, Del Castillo I, Villamar M, Moreno-Pelayo MA, Moreno F, Moreno R, Tapia MC. De novo mutation in the gene encoding connexin-26 (GJB2) in a sporadic case of keratitis-ichthyosis-deafness (KID) syndrome. Am J Med Genet A. 2003;117:89–91. 28. van Steensel MA, van Geel M, Nahuys M, Smitt JH, Steijlen PM. A novel connexin 26 mutation in a patient diagnosed with keratitis-ichthyosis-deafness syndrome. J Invest Dermatol. 2002;118:724–7. 29. Richard G, Rouan F, Willoughby CE, Brown N, Chung P, Ryynanen M, Jabs EW, Bale SJ, DiGiovanna JJ, Uitto J, Russell L. Missense mutations in GJB2 encoding

318

30.

31.

32.

33.

34. 35. 36.

37.

38.

39.

40. 41.

42.

43.

44.

45.

46.

T. Aasen and D.P. Kelsell connexin-26 cause the ectodermal dysplasia keratitis-ichthyosis-deafness syndrome. Am J Hum Genet. 2002;70:1341–48. van Geel M, van Steensel MA, Kuster W, Hennies HC, Happle R, Steijlen PM, Konig A. HID and KID syndromes are associated with the same connexin 26 mutation. Br J Dermatol. 2002;146:938–42. van Steensel MA, Steijlen PM, Bladergroen RS, Hoefsloot EH, van RavenswaaijArts CM, van Geel M. A phenotype resembling the Clouston syndrome with deafness is associated with a novel missense GJB2 mutation. J Invest Dermatol. 2004;123:291–3. Maestrini E, Korge BP, Ocana-Sierra J, Calzolari E, Cambiaghi S, Scudder PM, Hovnanian A, Monaco AP, Munro CS. A missense mutation in connexin26, D66H, causes mutilating keratoderma with sensorineural deafness (Vohwinkel’s syndrome) in three unrelated families. Hum Mol Genet. 1999;8:1237–43. Bondeson ML, Nystrom AM, Gunnarsson U, Vahlquist A. Connexin 26 (GJB2) Mutations in two Swedish patients with atypical Vohwinkel (mutilating keratoderma plus deafness) and KID syndrome both extensively treated with acitretin. Acta Derm Venereol. 2006;86:503–8. Smith FJ, Morley SM, McLean WH. A novel connexin 30 mutation in Clouston syndrome. J Invest Dermatol. 2002;118:530–2. Fraser FC, Der Kaloustian V. A man, a syndrome, a gene: Clouston’s hidrotic ectodermal dysplasia (HED). Am J Med Genet. 2001;100:164–8. van Steensel MA, Jonkman MF, van Geel M, Steijlen PM, McLean WH, Smith FJ. Clouston syndrome can mimic pachyonychia congenita. J Invest Dermatol. 2003;121:1035–38. Jan AY, Amin S, Ratajczak P, Richard G, Sybert VP. Genetic heterogeneity of KID syndrome: identification of a Cx30 gene (GJB6) mutation in a patient with KID syndrome and congenital atrichia. J Invest Dermatol. 2004;122:1108–13. Terrinoni A, Leta A, Pedicelli C, Candi E, Ranalli M, Puddu P, Paradis M, Angelo C, Bagetta G, Melino G. A novel recessive connexin 31 (GJB3) mutation in a case of erythrokeratodermia variabilis. J Invest Dermatol. 2004;12:837–9. Richard G, Brown N, Smith LE, Terrinoni A, Melino G, Mackie RM, Bale SJ, Uitto J. The spectrum of mutations in erythrokeratodermias—novel and de novo mutations in GJB3. Hum Genet. 2000;106:321–9. Strober BE. Erythrokeratodermia variabilis. Dermatol Online J. 2003;9:5. Macari F, Landau M, Cousin P, Mevorah B, Brenner S, Panizzon R, Schorderet DF, Hohl D, Huber M.. Mutation in the gene for connexin 30.3 in a family with erythrokeratodermia variabilis. Am J Hum Genet. 2000;67:1296–301. Wilgoss A, Leigh IM, Barnes MR, Dopping-Hepenstal P, Eady RA, Walter JM, Kennedy CT, Kelsell DP. Identification of a novel mutation R42P in the gap junction protein b-3 associated with autosomal dominant erythrokeratoderma variabilis. J Invest Dermatol. 1999;113:1119–22. Richard G, Smith LE, Bailey RA, Itin P, Hohl D, Epstein EH, Jr., DiGiovanna JJ, Compton JG, Bale SJ. Mutations in the human connexin gene GJB3 cause erythrokeratodermia variabilis. Nat Genet. 1998;20:366–9. Richardson RJ, Joss S, Tomkin S, Ahmed M, Sheridan E, Dixon MJ. A nonsense mutation in the first transmembrane domain of connexin 43 underlies autosomal recessive oculodentodigital syndrome. J Med Genet. 2006;43:e37. Lai A, Le DN, Paznekas WA, Gifford WD, Jabs EW, Charles AC. Oculodentodigital dysplasia connexin43 mutations result in non-functional connexin hemichannels and gap junctions in C6 glioma cells. J Cell Sci. 2006;119:532–41. Paznekas WA, Boyadjiev SA, Shapiro RE, Daniels O, Wollnik B, Keegan CE, Innis JW, Dinulos MB, Christian C, Hannibal MC, Jabs EW. Connexin 43 (GJA1) mutations cause

14 Skin

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57. 58.

59.

60. 61.

62.

319

the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet. 2003;72:408–18. Kelly SC, Ratajczak P, Keller M, Purcell SM, Griffin T, Richard G. A novel GJA 1 mutation in oculo-dento-digital dysplasia with curly hair and hyperkeratosis. Eur J Dermatol. 2006;16:241–5. van Steensel MA, Spruijt L, van der Burgt I, Bladergroen RS, Vermeer M, Steijlen PM, van Geel M. A 2-bp deletion in the GJA1 gene is associated with oculo-dento-digital dysplasia with palmoplantar keratoderma. Am J Med Genet A. 2005;132A:171–4. Kjaer KW, Hansen L, Eiberg H, Leicht P, Opitz JM, Tommerup N. Novel Connexin 43 (GJA1) mutation causes oculo-dento-digital dysplasia with curly hair. Am J Med Genet. 2004;127A:152–7. Haass NK, Wladykowski E, Kief S, Moll I, Brandner JM. Differential induction of connexins 26 and 30 in skin tumors and their adjacent epidermis. J Histochem Cytochem. 2006;54:171–82. Yano T, Fujimoto E, Hagiwara H, Sato H, Yamasaki H, Negishi E, Ueno K. Connexin 32 as an anti-invasive and anti-metastatic gene in renal cell carcinoma. Biol Pharm Bull. 2006;29:1991–4. Fujimoto E, Satoh H, Negishi E, Ueno K, Nagashima Y, Hagiwara K, Yamasaki H, Yano T. Negative growth control of renal cell carcinoma cell by connexin 32: possible involvement of Her-2. Mol Carcinog. 2004;40:135–42. McLachlan E, Shao Q, Wang HL, Langlois S, Laird DW. Connexins act as tumor suppressors in three-dimensional mammary cell organoids by regulating differentiation and angiogenesis. Cancer Res. 2006;66:9886–94. Kalra J, Shao Q, Qin H, Thomas T, Aoui-Jamali MA, Laird DW. Cx26 inhibits breast MDA-MB-435 cell tumorigenic properties by a gap junctional intercellular communication-independent mechanism. Carcinogenesis. 2006;27:2528–37. Qin H, Shao Q, Thomas T, Kalra J, Aoui-Jamali MA, Laird DW. Connexin26 regulates the expression of angiogenesis-related genes in human breast tumor cells by both GJICdependent and -independent mechanisms. Cell Commun Adhes. 2003;10:387–93. Qin H, Shao Q, Curtis H, Galipeau J, Belliveau DJ, Wang T, aoui-Jamali MA, Laird DW. Retroviral delivery of connexin genes to human breast tumor cells inhibits in vivo tumor growth by a mechanism that is independent of significant gap junctional intercellular communication. J Biol Chem. 2002;277:29132–8. Naus CC. Gap junctions and tumor progression. Can J Physiol Pharmacol. 2002;80:136–41. Yamasaki H, Omori Y, Krutovskikh V, Zhu W, Mironov N, Yamakage K, Mesnil M. Connexins in tumor suppression and cancer therapy. Novartis Found Symp. 1999;219:241–54. Mazereeuw-Hautier J, Bitoun E, Chevrant-Breton J, Man SY, Bodemer C, Prins C, Antille C, Saurat JH, Atherton D, Harper JI, Kelsell DP, Hovnanian A. Keratitisichthyosis-deafness syndrome: disease expression and spectrum of connexin 26 (GJB2) mutations in 14 patients. Br J Dermatol. 2007;156:1015–19. Shima K, Muramatsu T, Abiko Y, Yamaoka Y, Sasaki H, Shimono M. Connexin 43 transfection in basaloid squamous cell carcinoma cells. Oncol Rep. 2006;16:285–8. Oliveira R, Christov C, Guillamo JS, Debouard S, Palfi S, Venance L, Tardy M, Peschanski M. Contribution of gap junctional communication between tumor cells and astroglia to the invasion of the brain parenchyma by human glioblastomas. BMC Cell Biol. 2005;6:7. Lin JH, Takano T, Cotrina ML, Arcuino G, Kang J, Liu S, Gao Q, Jiang L, Li F, Lichtenberg-Frate H, Haubrich S, Willecke K, Goldman SA, Nedergaard M. Connexin 43 enhances the adhesivity and mediates the invasion of malignant glioma cells. J Neurosci. 2002;22:4302–11.

320

T. Aasen and D.P. Kelsell

63. Ito A, Katoh F, Kataoka TR, Okada M, Tsubota N, Asada H, Yoshikawa K, Maeda S, Kitamura Y, Yamasaki H, Nojima H. A role for heterologous gap junctions between melanoma and endothelial cells in metastasis. J Clin Invest. 2000;105:1189–97. 64. Djalilian AR, McGaughey D, Patel S, Seo EY, Yang C, Cheng J, Tomic M, Sinha S, Ishida-Yamamoto A, Segre JA. Connexin 26 regulates epidermal barrier and wound remodeling and promotes psoriasiform response. J Clin Invest. 2006;116:1243–53. 65. Brandner JM, Houdek P, Husing B, Kaiser C, Moll I. Connexins 26, 30, and 43: differences among spontaneous, chronic, and accelerated human wound healing. J Invest Dermatol. 2004;122:1310–20. 66. Qiu C, Coutinho P, Frank S, Franke S, Law LY, Martin P, Green CR, Becker DL. Targeting connexin43 expression accelerates the rate of wound repair. Curr Biol. 2003;13:1697–703. 67. Laird DW. Life cycle of connexins in health and disease. Biochem J. 2006;394:527–43. 68. Common JE, Di WL, Davies D, Kelsell DP. Further evidence for heterozygote advantage of GJB2 deafness mutations: a link with cell survival. J Med Genet. 2004;41:573–5. 69. Common JE, Di WL, Davies D, Galvin H, Leigh IM, O’Toole EA, Kelsell DP. Cellular mechanisms of mutant connexins in skin disease and hearing loss. Cell Commun Adhes. 2003;10:347–51. 70. Common JE, Becker D, Di WL, Leigh IM, O’Toole EA, Kelsell DP. Functional studies of human skin disease- and deafness-associated connexin 30 mutations. Biochem Biophys Res Commun. 2002;298:651–6. 71. Di WL, Monypenny J, Common JE, Kennedy CT, Holland KA, Leigh IM, Rugg EL, Zicha D, Kelsell DP. Defective trafficking and cell death is characteristic of skin diseaseassociated connexin 31 mutations. Hum Mol Genet. 2002;11:2005–14. 72. Aasen T, Hodgins MB, Edward M, Graham SV. The relationship between connexins, gap junctions, tissue architecture and tumor invasion, as studied in a novel in vitro model of HPV-16-associated cervical cancer progression. Oncogene 2003;22:7969–80. 73. Maher AC, Thomas T, Riley JL, Veitch G, Shao Q, Laird DW. Rat epidermal keratinocytes as an organotypic model for examining the role of Cx43 and Cx26 in skin differentiation. Cell Commun Adhes. 2005;12:219–30. 74. Unsworth HC, Aasen T, McElwaine S, Kelsell DP. Tissue-specific effects of wildtype and mutant connexin 31: a role in neurite outgrowth. Hum Mol Genet. 2007;16:165–72. 75. Kibschull M, Nassiry M, Dunk C, Gellhaus A, Quinn JA, Rossant J, Lye SJ, Winterhager E. Connexin31-deficient trophoblast stem cells: a model to analyze the role of gap junction communication in mouse placental development. Dev Biol. 2004;273:63–75. 76. Teubner B, Michel V, Pesch J, Lautermann J, Cohen-Salmon M, Sohl ¨ G, Jahnke K, Winterhager E, Herberhold C, Hardelin JP, Petit C, Willecke K. Connexin30 (Gjb6)deficiency causes severe hearing impairment and lack of endocochlear potential. Hum Mol Genet. 2003;12:13–21. 77. Meyer CG, Amedofu GK, Brandner JM, Pohland D, Timmann C, Horstmann RD. Selection for deafness? Nat Med. 2002;8:1332–3. 78. Gabriel HD, Jung D, Butzler C, Temme A, Traub O, Winterhager E, Willecke K. Transplacental uptake of glucose is decreased in embryonic lethal connexin26-deficient mice. J Cell Biol. 1998;140:1453–61. 79. Bakirtzis G, Jamieson S, Aasen T, Bryson S, Forrow S, Tetley L, Finbow M, Greenhalgh D, Hodgins M. The effects of a mutant connexin 26 on epidermal differentiation. Cell Commun Adhes. 2003;10:359–64. 80. Bakirtzis G, Choudhry R, Aasen T, Shore L, Brown K, Bryson S, Forrow S, Tetley L, Finbow M, Greenhalgh D, Hodgins M. Targeted epidermal expression of mutant Connexin 26(D66H) mimics true Vohwinkel syndrome and provides a model for the pathogenesis of dominant connexin disorders. Hum Mol Genet. 2003;12:1737–44.

14 Skin

321

81. Maass K, Ghanem A, Kim JS, Saathoff M, Urschel S, Kirfel G, Grummer R, Kretz M, Lewalter T, Tiemann K, Winterhager E, Herzog V, Willecke K. Defective epidermal barrier in neonatal mice lacking the C-terminal region of connexin43. Mol Biol Cell. 2004;15:4597–608. 82. Watanabe M, Iwashita M, Ishii M, Kurachi Y, Kawakami A, Kondo S, Okada N. Spot pattern of leopard Danio is caused by mutation in the zebrafish connexin41.8 gene. EMBO Rep. 2006;7:893–7. 83. Thomas T, Aasen T, Hodgins M, Laird DW. Transport and function of Cx26 mutants involved in skin and deafness disorders. Cell Commun Adhes. 2003;10:353–8. 84. Schmuth M, Fluhr JW, Crumrine DC, Uchida Y, Hachem JP, Behne M, Moskowitz DG, Christiano AM, Feingold KR, Elias PM. Structural and functional consequences of loricrin mutations in human loricrin keratoderma (Vohwinkel syndrome with ichthyosis). J Invest Dermatol. 2004;122:909–22. 85. O’Driscoll J, Muston GC, McGrath JA, Lam HM, Ashworth J, Christiano AM. A recurrent mutation in the loricrin gene underlies the ichthyotic variant of Vohwinkel syndrome. Clin Exp Dermatol. 2002;27:243–6. 86. Suga Y, Jarnik M, Attar PS, Longley MA, Bundman D, Steven AC, Koch PJ, Roop DR. Transgenic mice expressing a mutant form of loricrin reveal the molecular basis of the skin diseases, Vohwinkel syndrome and progressive symmetric erythrokeratoderma. J Cell Biol. 2000;15:401–12.

Chapter 15

Connexins in the Nervous System Charles K. Abrams and John E. Rash

Abstract This chapter reviews the localizations and physiological roles of connexins in neurons and glia of the central and peripheral nervous systems. Cx32 forms gap junctions in noncompact myelin in Schwann cells, which are thought to form a reflexive communication pathway connecting the outer and inner myelin layers. Cx29 is also expressed in myelinating Schwann cells, but does not appear to form gap junctions; its role remains to be elucidated. Mutations in CX32 cause an X-linked form of the inherited neuropathy Charcot-Marie-Tooth disease; most such mutations are likely to act through loss of function. Connexins may also play an important role in proliferating Schwann cells. Oligodendrocytes express Cx32, Cx47, and Cx29, while astrocytes express Cx43, Cx30, and possibly Cx26. Although astrocytes are extensively coupled to each other in vivo, oligodendrocyte coupling in vivo is demonstrable only to astrocytes, most via heterotypic Cx43–Cx47 or Cx30–Cx32 junctions. These junctions, along with those between astrocytes, may play a role in spatial buffering of Kþ ions and neurotransmitters, and may influence severity of tissue damage during ischemia. Mutations in CX47 cause Pelizaeus Merzbacher–like disease while mutations in CX43 cause oculodentodigital dysplasia. Only Cx36 and Cx45 have been definitively identified in nonretinal brain neurons, where they form electrical synapses; neuron–neuron coupling may play a role in the pathogenesis of epilepsy. Keywords Glia Oligodendrocyte Astrocyte Schwann cell Neuron X-linked Charcot-Marie-Tooth disease Pelizaeus Merzbacher–like disease Oculodentodigital dysplasia Tremor Epilepsy Ischemia Cx26 Cx29 Cx30 Cx32 Cx43 Cx47

C.K. Abrams (*) Departments of Neurology & Pharmacology and Physiology, State University of New York Downstate Medical Center, 450 Clarkson Avenue, Box 1213, Brooklyn, NY 11203, United States e-mail: [email protected]

A. Harris, D. Locke (eds), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_15, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

323

324

C.K. Abrams and J.E. Rash

15.1 Introduction The first half of this chapter reviews the localizations and physiological roles of connexins expressed in Schwann cells in the peripheral nervous system (PNS), and in astrocytes, oligodendrocytes, and neurons in the central nervous system (CNS). The second half of this chapter examines the roles that connexin channels play in both inherited and acquired disorders of the PNS and CNS.

15.2 Connexins in Peripheral Nervous System Glia Morphologic gap junctions were first identified in nonmyelinating Schwann cells about 40 years ago [1], but it was nearly 25 years later that Cx32 was identified in peripheral myelin [2]. More recently Cx29 has also been identified in myelinating Schwann cells [3,4,5]. This section discusses the localization and possible roles of connexins in myelinating and nonmyelinating Schwann cells.

15.2.1 Connexin Expression in Nonmyelinating Schwann Cells During peripheral nerve development [6] and after injury [7], Schwann cells proliferate in the early stages of developmental myelination and remyelination. Substantial evidence suggests that gap junction formation between proliferating Schwann cells plays a role in these processes. Four decades ago, Raine et al. [1] identified gap junction–like structures between proliferating Schwann cells during remyelination after experimental allergic neuritis. Later, Tetzlaff [8] showed the transient formation of gap junction–like structures between proliferating Schwann cells in regenerating chick nerve, and Konishi [9] demonstrated dye transfer between acutely dissociated or cultured mouse Schwann cells. In 1998, Dezawa et al. [10] showed ‘‘small-scale gap junction–like structures’’ between regenerating axons and adjacent Schwann cells, identified Cx32 at the axon–Schwann cell apposition and demonstrated transfer of biocytin between axons and Schwann cells. Other reports noted gap junction formation between Schwann cells transplanted into the CNS [11,12]. Gap junctions mediate communication between proliferating Schwann cells during development, remyelination, and nerve regeneration, and so may serve to distribute and synchronize signals important for proliferation and differentiation. Electrophysiological studies of primary proliferating Schwann cells in culture suggest that the gap junction channels coupling these cells are heterogeneous and that their characteristics are regulated by growth factors important in proliferation and differentiation [13,14,15]. However, the specific physiologic roles played by these gap junctions remain to be elucidated, and the identity of connexins that compose them is still unclear.

15 Nervous System

325

15.2.2 Connexin Expression in Myelinating Schwann Cells Schnapp and Mugnaini [16] and Sandri et al. [17] identified gap junctions in peripheral nerve myelin. Sandri et al. [17] also identified gap junctions in central myelin. However, the localization of Cx32 in the myelinating Schwann cell came about 15 years later, with the identification of mutations in CX32 as a cause of X-linked Charcot-Marie-Tooth disease (CMTX) (see below for further discussion of CMTX1) [2]. Subsequent immunohistochemical studies showed CX32 expression in the noncompact myelin of the paranodal loops and Schmidt-Lanterman (S-L) incisures. Figure 15.1 illustrates the terms used to describe regions of the myelinating Schwann cell) [4,18,19,20,21,22]. Freezefracture replica immunogold labeling (FRIL) [23] confirmed these localizations but also showed Cx32 gap junctions in internodes, between the outermost noncompact layer of myelin and the second outermost, partially compact layer. In addition, Cx32 is expressed in the regenerating nerve, likely within the myelin sheath, with a temporal expression pattern similar to that of other myelin genes such as P0 glycoprotein and myelin-associated glycoprotein [18,21,24].

Internode

Juxtaparanode

Paranode

Node

Periaxonal space Axon

SchmidtLanterman incisure

Inner mesaxon

Fig. 15.1 Components of the myelinating Schwann cell. This schematic shows the nodal, paranodal, juxtaparanodal, and internodal regions of the Schwann cell. For clarity, the drawing is not to scale and only four Schwann cell wraps are shown. Also shown is a Schmidt-Lanterman incisure, which, like the paranode, is a region of noncompact myelin. Black ovals show the locations of Cx29 immunoreactivity, while gray ovals show the locations of Cx32 immunoreactivity. See text for further details. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

326

C.K. Abrams and J.E. Rash

Cx29 is also expressed in peripheral nerve [3,4,5], but its expression pattern differs from that of Cx32 in important ways that suggest a functional distinction. Although Cx32 and Cx29 colocalize at S-L incisures, in other areas they do not; Cx29 is not found in the outermost paranodal areas or between the two outermost layers of myelin, the locations of Cx32 expression. In addition, Cx29 immunoreactivity is found adaxonally, along the inner mesaxon and at the inner paranodal and juxtaparanodal regions where Cx32 is not found [3,4]. Importantly, at these locations FRIL shows that Cx29 immunoreactivity is specifically associated with hexameric rosettes and plaque-like arrays of hexameric intramembrane particles [3], but without corresponding structures in the apposed axonal plasma membrane. Miller and da Silva [25] had previously found hexagonal rosettes in the Schwann cell plasma membrane, but were unable to determine if the rosettes were directly apposed to the particles in the axonal membrane. Stolinski et al. [26] showed that at least some Schwann cell particles are aligned with those on axons, suggesting the possibility that the particles of the Schwann cell and axon might form a functional unit. These findings suggest that Cx29 may not form junctional channels by itself. This idea is validated by studies showing that when expressed in transfected mammalian cells Cx29 fails to induce electrical coupling, either as homotypic channels or as heterotypic channels with Cx32 [4]. Based on recordings from cells expressing Cx32 paired heterotypically with cells expressing both Cx32 and Cx29, Altevogt et al. [4] suggest that Cx32 and Cx29 form functional heteromeric hemichannels. Yet, to the best of our knowledge, biochemical interactions between Cx29 and Cx32 have not been reported in situ, although such interactions could in theory occur at the S-L incisures. As mentioned, most Cx29 in peripheral nerve (and in the CNS as well; see below) is not colocalized with Cx32. For these reasons, interaction between Cx29 and Cx32 at locations other than S-L incisures in Schwann cells seems unlikely. This, along with evidence that Cx29 does not form homomeric gap junction channels, suggests that Cx29 may function in situ as nonjunctional hemichannels.

15.2.2.1 The Reflexive Gap Junction Pathway of Myelinating Schwann Cells Cx32 is required for normal function of peripheral nerve. This is evident from both the peripheral nerve pathology seen in mice with targeted ablation of the gene for Cx32 [27,28] and from the clinical manifestations in patients with CMTX who have clear loss-of-function mutations of CX32 (below). Gap junctions typically provide communication pathways between neighboring cells. However, in myelinating Schwann cells, the Cx32 gap junction channels are thought to form an intracellular reflexive pathway between adjacent layers of noncompacted myelin, and thus provide a radial diffusion pathway for signaling molecules, metabolites or ions from the inner (adaxonal) to the outer (abaxonal) cytoplasm of a myelinating Schwann cell. This pathway is calculated to be 350 to 1000 times shorter than the circumferential pathway

15 Nervous System

327

[18,29] and is likely to play an important role in normal Schwann cell function and axon–Schwann cell interactions. Balice-Gordon et al. [19] functionally demonstrated this pathway by injecting the junction-permeable fluorescent dye 5,6-carboxyfluorescein near the Schwann cell nucleus and tracking its transfer to the adaxonal region via radial paths thought to correspond to S-L incisures. However, they were also able to demonstrate similar dye transfer in nerve fibers from Cx32 knockout (KO) mice. One possible explanation is that a reflexive pathway still forms in the absence of Cx32, but that Cx32 is somehow required for its full function in myelin maintenance. In any case, the presence or identity of any other reflexive connexin is not clear. The failure to demonstrate Cx29 in gap junctions in peripheral nerve [3] or functional homotypic channels in exogenous expression systems [4] suggest that it does not participate in such a pathway. Cx47, a connexin expressed in oligodendrocytes (see below), might be an appealing candidate for the other connexin at S-L incisures, but most studies have failed to demonstrate Cx47 messenger RNA [30,31] or protein [32] in peripheral nerve. A single study showed amplification of Cx47 message by polymerase chain reaction (PCR) [33]. The signaling molecules, metabolites, or ions that travel through the reflexive pathway are unknown. One possibility is that this pathway participates in spatial buffering of extracellular Kþ ions. Another possibility is suggested by Toews et al. [34], who recently showed that expression of the IP3R3 isoform of the inositol triphosphate (IP3) receptor is enhanced in the regions of the myelinating Schwann cell where Cx32 is found; they suggest that Cx32 may mediate IP3 signaling, or vice versa. 15.2.2.2 Possible Roles of Connexin29 in Myelinating Schwann Cells The role of Cx29 in peripheral nerve is even less clear than that of Cx32, particularly given the points mentioned above. Its expression in glial precursors temporally coincides with specification of Schwann cell lineage, suggesting that it has a role in peripheral nerve development [35]. However, lack of Cx29 seems to have little functional consequence in most neural tissues. Eiberger et al. [36] examined mice in which the Cx29 gene was replaced by the LacZ gene. They found no adult CNS or PNS myelin abnormalities by electrophysiological or morphological assays in several neural components, including the auditory system. Tang et al. [37] noted that approximately 50% of mice with targeted ablation of Cx29 develop auditory neuropathy with delayed maturation of hearing and increased sensitivity to noise damage. The high-frequency hearing loss correlated with abnormalities in the myelin around the somata of spiral ganglion cells but not in myelin of the auditory nerve. Differences between the two studies may be due to differences in the age groups of the animals [36], but the reasons for 50% penetrance in the studies of Tang et al. [37] remain to be elucidated. The localization of Cx29 along the inner mesaxon by immunocytochemistry, and the identification of anti-Cx29

328

C.K. Abrams and J.E. Rash

labeled unapposed intramembranous hexameric particles by FRIL, suggest that Cx29 might function as a hemichannel involved in Kþ buffering at these locations. This hypothesis is supported by the work of Altevogt et al. [4], which showed that immunocytochemical staining of Schwann cell Cx29 coincided to a great degree with staining for the axonal Kþ channel Kv1.2. Twenty years earlier, Stolinski et al. [26] showed alignment between some hexameric Schwann cell membrane particles and those in the adjacent axonal membrane, leading those investigators to suggest that these structures might represent Kþ channels.

15.3 Connexins of Central Nervous System Glia Gap junctions between astrocytes (A-A junctions) were first identified in chicken, mouse, and goldfish almost 40 years ago [38]. Subsequently, investigators identified A-A and astrocyte-oligodendrocyte (A-O) junctions in several other species [39,40,41]. This section briefly reviews the localization of connexins in oligodendrocytes and astrocytes of mature vertebrates and comments on possible physiologic roles for gap junctions in these cell types. Connexins expressed in glial cells that make up the ependyma — the membrane that lines the ventricles of the brain and the central canal of the spinal cord [42,43] — are not considered in this chapter.

15.3.1 Astrocyte Connexins An expanding and consistent body of literature supports the localization of Cx30 and Cx43 to astrocytes. The relevant studies are summarized in Table 15.1. The expression of Cx26 in astrocytes remains unresolved. Several investigators have reported Cx26 expression in astrocytes [44,45,46,47]. On the other hand, Filippov et al. [48] examined the distribution of Cx26 using a LacZ reporter. They were unable to demonstrate expression of the reporter in astrocytes, though Cx26–LacZ expression was robust in other tissues such as liver, skin and meningeal cells, where Cx26 is expressed. Further study of the localization and role of Cx26 in astrocytes is required but any expression that occurs is likely to be at low levels and possibly in small region-specific subpopulations of astrocytes [46, 47].

15.3.2 Oligodendrocyte Connexins Cx29, Cx32, and Cx47 have been consistently identified in oligodendrocytes. Though Cx47 was first reported in neurons [30], it is now generally

Shows little overlap with Cx43 and Cx30, is found around oligo cell bodies in WT, and is absent in Cx32KO [44,47]. It is mostly subcortical [46]. LacZ reporter shows no astrocyte expression [48].

Levels of expression in GM are much higher than in WM [44].

General comments about localization

Gray matter (GM)

Levels of expression in GM are much higher than in WM. It extensively overlaps with Cx43 but

Level of expression varies by brain region, with high levels at astrocyte endfoot processes. It colocalizes with Cx43 [47,63,229].

Levels of expression in GM are much higher than in WM. Extensively overlaps with Cx30 but has

Found at astrocyte processes near oligo Cx32 and Cx47 [45,47,63,230]. Its expression persists in Cx32KO [49].

Levels of expression in GM are much higher than in WM. It is coexpressed with Cx32 in cell bodies [32,45].

Found in all oligos and in Bergmann glia (as assessed by reporter gene). It is more diffusely distributed than Cx32 or Cx47, and does not overlap with other connexins [36,44]. It is found at internode and juxtaparanode of small myelinated fibers by immunocytochemistry [4,47,230].

Highly localized with Cx47, especially on cell bodies [32,44,50]. It is dispersed throughout the neuropil [50].

Found in almost all oligos. Most Cx47-lacking cells are not associated with astrocyte Cx43 or Cx30, but are associated with astrocyte Cx26 [44]. It is found on somata and processes of intrafasicular oligos [4], and in paranodes and S-L incisures [18,49,69,230].

Found in almost all oligos, in cell bodies and processes. Most Cx32lacking cells are associated with astrocyte Cx43 and Cx30 [32,44]. Found between oligo somata and astrocytes and along the surface of internodal myelin. Linked to astrocytes (Cx47 >> Cx32) [49,230]. Coexpressed with Cx32 on cell body [32,44,50].

Table 15.1 Localization of glial connexins in the central nervous system: the cellular and subcellular localization of connexins of astrocytes and oligodendrocytes Astrocyte connexins Oligodendrocyte connexins Cx26 Cx30 Cx43 Cx29 Cx32 Cx47

15 Nervous System 329

little overlap with Cx26 [44].

Cx43

WT, wild-type; oligo, oligodendrocyte; S-L, Schimdt-Lantermann.

White matter (WM)

has little overlap with Cx26 [44].

Table 15.1 (continued) Astrocyte connexins Cx26 Cx30

In neocortex and other brain areas, found at juxtaparanodes and internode of small/ medium caliber fibers. No overlap with Cx32 is seen [4,32,45].

Found on the outer aspect of large myelinated fibers [18,32], but shows less localization with Cx47 than in GM [44], and is not found on cell bodies [32]. Others report variable colocalization with Cx47 [50].

Oligodendrocyte connexins Cx29 Cx32

Shows less colocalization with Cx32 in GM [44]. Plaques surround cell bodies of mostly Cx32negative oligos [32]. Others report colocalization with Cx32, probably associated with myelinating fibers [50].

Cx47

330 C.K. Abrams and J.E. Rash

15 Nervous System

331

acknowledged to be in oligodendrocytes instead [31,32,44,49,50]. On one hand, data on the expression of Cx45 in oligodendrocytes are conflicting. Cx45 staining has been reported in hippocampal [51] and spinal cord [52] oligodendrocytes. On the other hand, Kruger et al. [53] examined the expression of a Cx45–LacZ reporter in several brain regions and identified neuronal but not glial staining in thalamus and in the pyramidal cell layer of cortex. Expression was also seen in the CA3/CA4 area of hippocampus, but the specific cells were not identified. Furthermore, Kleopa et al. [32] failed to identify Cx45 by immunocytochemistry in spinal cord oligodendrocytes, although they did detect signal for this connexin in cerebral vasculature, consistent with prior reports [54,55]. By FRIL, Rash et al. [56] found Cx45 in neuronal gap junctions in olfactory bulb and in olfactory epithelium sustentacular cells but did not detect it in oligodendrocyte gap junctions. In summary, the weight of evidence suggests that Cx45 is not found in oligodendrocytes.

15.3.3 Astrocyte-Astrocyte Coupling There is substantial morphologic and functional evidence for gap junctions between astrocytes (for reviews see [57,58]). Here the focus is on recent developments. As noted in Table 15.1, astrocytes express Cx30, Cx43, and possibly Cx26. Cx26 and Cx30 are known to form heteromeric and heterotypic channels, but neither interacts with Cx43. Thus, A-A gap junctional channels could be composed of any of the allowed combinations of homomeric or heteromeric hemichannels in homotypic or heterotypic configurations. Cx30 and Cx43 show extensive colocalization (Table 15.1) [44,47]. However, some recent data [59] suggest that Cx43 and Cx30 do not form functional heterotypic channels. Wallraff et al. [60] suggest that at least in stratum radiatum astrocytes, Cx26 does not form homotypic channels, since astrocytes lacking Cx43 showed coupling reduced to approximately half of wild-type, while the additional deletion of Cx30 led to complete loss of A-A coupling. The simplest interpretation of their data is that homomeric Cx30 and Cx43 contribute roughly equally to A-A coupling; however, compensatory increases in expression of Cx30 may have overemphasized the contribution of Cx30 [61]. Wallraff et al. [62] showed that there are two distinct populations of astrocytes with regard to junctional coupling; GluT (glutamate transporter)–expressing astrocytes are extensively coupled only to other GluT-expressing astrocytes, while GluR (glutamate receptor)–expressing astrocytes are not coupled to other cells. Data regarding expression of specific connexins in these cells were not provided, making it difficult to correlate their findings [62] with expression data provided by other investigators.

332

C.K. Abrams and J.E. Rash

15.3.4 Astrocyte-Oligodendrocyte Coupling Nagy et al. [47] examined the connexins present at A-O interfaces. In wild-type mice, they found astrocytic Cx43, Cx30, and Cx26 staining at apposed oligodendrocyte somata. When Cx32, present only in oligodendrocytes, is knocked out, at A-O junctions Cx30 staining disappears, Cx26 staining is reduced 50%, and Cx43 staining remains robust. These findings suggest Cx43–Cx47 and Cx30–Cx32, and perhaps Cx26–Cx32 heterotypic A-O coupling. Altevogt et al. [44] found that although most oligodendrocytes express both Cx32 and Cx47, in those expressing only Cx32, the association was primarily with astrocyte Cx26, while in those expressing only Cx47, the association was with Cx43 and Cx30; since recent data [59] suggest that Cx30 is not heterotypically compatible with Cx47, A-O heterotypic configurations could include Cx43–Cx47, Cx30–Cx32, and Cx26–Cx32, and heteromeric interactions between Cx43 and Cx30 might play a role as well. Kamasawa et al. [49] showed that greater abundance of Cx47 than Cx32 on the oligodendrocyte side of A-O gap junctions corresponded to more Cx43 than Cx30 on the astrocyte side, and suggested that this supports the identification of Cx43–Cx47 and Cx30–Cx32 heterotypic channels as the primary components of A-O gap junctions. Table 15.2 summarizes the potentially compatible oligodendrocytes and astrocyte connexins. In conclusion, it is likely that A-O gap junctions consist predominantly of Cx43–Cx47 and Cx30–Cx32 heterotypic channels, although Cx26–Cx32 channels may play a role as well.

15.3.5 Oligodendrocyte-Oligodendrocyte Coupling Most studies find minimal or no evidence of gap junction coupling, either physiological or morphological, between adult oligodendrocytes when assayed in situ [41,63,64]. However, morphologically sparse [65] and functionally weak [66,67,68,69] gap junction coupling has been noted in cultures of oligodendrocytes. Pastor et al. [52] demonstrated Lucifer yellow and neurobiotin dye-coupling between gray matter oligodendrocytes but saw no such coupling in the white matter. Their results could be accounted for by regional differences in O-O coupling or by coupling via an astrocyte intermediary (OA-O coupling) as originally described by Mugnaini [70]. Colocalization of Cx32 and Cx47, especially in somata [32,44,49,50], suggests that O-O gap junctions, if present, may be composed of homomeric or heteromeric Cx32 or Cx47 hemichannels, though Cx32 appears heterotypically incompatible with Cx47 [59]. There is little evidence of gap junctional coupling between oligodendrocytes, but Kamasawa et al. [49] did show Cx32 in gap junctions between successive paranodal loops of oligodendrocyte myelin; these autocellular junctions may form a reflexive oligodendrocyte pathway similar to that seen in Schwann cells.

Cx43 Cx29 Cx32 Cx47

Cx30

Cx26

Compatible [223]

Compatible [224]

Compatible [224]

Compatible [227]

Not compatible [59]

Not compatible [225]

Compatibility unknown but highly unlikely because incompatible with related Cx32 [4] Compatibility unknown but highly unlikely because incompatible with related Cx32 [4] Not determined Incompatible [4]

Incompatible [228] Incompatible [4] Compatible [223]

Compatible [224]

Compatible [226]

Compatible [59] Not determined Incompatible [59] Compatible [30]

Incompatible [59]

Not determined

Table 15.2 Functional compatibility of glial connexins: studies of the heterotypic compatibility between connexins expressed in glial cells Cx26 Cx30 Cx43 Cx29 Cx32 Cx47

15 Nervous System 333

334

C.K. Abrams and J.E. Rash

15.3.6 Roles of Central Nervous System Glial Connexins A likely major role of A-A and A-O coupling is creation of a glial syncytium [70]. One function of such a syncytium, propagation of Ca2þ waves (see [71] for a recent review), has received much interest. Other functions may include other forms of intercellular signaling, transfer of metabolites and spatial buffering of increases in extracellular Kþ or neurotransmitters due to neuronal or synaptic activity. Examination of mice with targeted deletion of one or both oligodendrocyte connexins has provided insight into their roles. Mice with targeted ablation of Cx32 [20,27,28] or Cx47 [31,50] have only minimal physiological or pathological changes in the CNS and no observable CNS phenotype; however, animals lacking both connexins have abnormal movements and seizures, followed by death at 6 weeks. Pathology includes ‘‘axons with abnormal, thinly myelinated sheaths’’, ‘‘myelinated axons with markedly enlarged extracellular spaces separating the axon from its myelin sheath’’, and ‘‘occasional myelinated axons with enlarged collars of periaxonal oligodendrocyte cytoplasm’’ [50]. Targeted deletion of Kir4.1 [72], an inwardly rectifying Kþ channel found at astrocyte perivascular endfeet [73,74], leads to an unusual CNS pathology [72] similar to that of the Cx32/Cx47 double knockout. As noted above, Kamasawa et al. [49] identified Cx32 in central paranodal myelin and suggested that it may have a role in a reflexive pathway similar to that in the Schwann cell. They also presented a model for voltage-augmented concentration-dependent Kþ siphoning from the periaxonal space, through the paranodal Cx32 pathway and A-O gap junctions, to astrocyte endfeet. They suggested that failure of this pathway could explain the myelin swelling seen in mice lacking Kir4.1 or Cx47 and Cx32. Subsequent work by Menichella et al. [75] supports the hypothesis that oligodendrocyte Cx32 and Cx47 are involved in spatial buffering of Kþ during neuronal activity. They showed that severity of pathology in the Cx32/Cx47 double knockout is dependent on neuronal activity, and demonstrated that heterozygosity for either Cx47 or Kir4.1 alone causes little pathology in a Cx32KO background, while the double heterozygotes show striking pathology. In contrast to the cited progress in understanding of the roles of Cx32 and Cx47 in oligodendrocytes, there are still few clues to the role of Cx29 in the CNS. Studies of mouse models have also elucidated the roles of astrocyte connexins. Wallraff et al. [60] used mice expressing no astrocyte Cx43 and no Cx30 in any cell to demonstrate that Kþ buffering in hippocampus is at least partially dependent on Cx43-mediated and Cx30-mediated gap junctional coupling. In contrast to the Cx32/Cx47 double knockout [50], these mice had no gross behavioral abnormalities. This suggests that the effect of loss of the oligodendrocyte connexins (Cx32 and Cx47) is not simply disruption of Cx30–Cx32 and Cx43–Cx47 heterotypic coupling. Alternatively, compensation for loss of Cx43 and Cx30 may be more effective than for loss of Cx47 and Cx32. In a mouse with targeted ablation of astrocyte Cx43, Theis et al. [61] found a decrease of

15 Nervous System

335

roughly 50% in the amount of A-A coupling (similar to that reported in [60]), increased locomotor activity, and an increase in the velocity of spreading depression. The authors suggest that under normal conditions, astrocyte gap junctional communication attenuates spreading depression by reducing extracellular Kþ or glutamate concentrations. Roles for glial connexins independent of their channel function [76] and for plasma membrane hemichannels in glia [77,80,81] have been considered, but are not discussed further here.

15.4 Connexins in Central and Peripheral Neurons Although vertebrate electrical synapses were first described by Bennett et al. [82] in 1959, morphologic gap junctions between mammalian neurons were not identified until ten years later [38]; the demonstration of electrical synapses in mammalian brain soon followed (see [83]). This section reviews the localization of connexins in CNS and PNS neurons.

15.4.1 Expression of Connexins in Neurons Although expression of neuronal connexin proteins may be widespread, functional coupling between mammalian neurons appears to be predominantly restricted to inhibitory interneurons [83]. However, some experimental evidence [84,85] and theoretical modeling [86] suggest that pyramidal cells may be electrically coupled, possibly via axo-axonal connections. In addition, FRIL and thin-sections of dentate granule cells show axonal gap junctions [87]. The first neuronal connexins, Cx35 and Cx34.7, were cloned from a perch retina complementary DNA (cDNA) library in 1999. Immunostaining identified these proteins on bipolar cells and in unidentified processes in the inner plexiform and nerve fiber layer [88]. Shortly thereafter, Cx36 was identified as the rodent ortholog of Cx35 [89,90]. High levels of Cx36 mRNA were found in olfactory bulb, pineal gland, inferior olive, CA3/ CA4, and brainstem nuclei, while moderate levels were seen in many brain and spinal cord regions. No thalamic labeling was seen, except in reticular nuclei [89]. Immunocytochemical reactivity was especially intense in inferior olive and retina, but was also seen in many regions throughout the brain and spinal cord including the olfactory bulb, CA1 to CA3 regions of hippocampus, striatum, globus pallidus, dentate gyrus, and primary dendrites of Purkinje cells [42,91]. Cx45 has been localized to neurons of cerebral cortex and claustrum [92,93], as well as at mixed synapses of olfactory bulb glomeruli [56,94]. It is also expressed in subpopulations of retinal bipolar and amacrine cells [95,96,97,98,99].

336

C.K. Abrams and J.E. Rash

Other connexins are reportedly expressed in restricted populations of neurons. A single report [100] suggests that mature lower motor neurons in spinal cord (anterior horn cells, AHC) express RNA for Cx36, Cx37, Cx40, Cx43, and Cx45, and express Cx40, Cx43, and Cx45 protein, although no dye-coupling between these cells can be demonstrated 7 days postnatally. Lack of appropriate antibodies prevented evaluation of Cx36 and Cx37 protein. Dye and electrical coupling can be demonstrated earlier in development between AHC (for review see [101]). Cx57 immunofluorescence is found only in retinal horizontal cells [102,103] and Cx50 immunofluorescence was also found to be colocalized with it [104]. Cx26, Cx32, and Cx43 mRNA or protein have been reported in neuronal gap junctions by in situ hybridization or immunocytochemistry [105,106,107,108,109,110,111]. However, using immunocytochemistry and extensive examination by FRIL [42,58,63] and in situ hybridization [88], other investigators have failed to identify these connexins in CNS neurons. Likewise, although an initial report identified Cx47 as a neuronal connexin, more recent work suggests that it is found exclusively in oligodendrocytes [30, 31].

15.4.2 Neuron-Neuron Coupling The data summarized above suggest that the predominant connexin expression in neurons of the brain and spinal cord is restricted to Cx36 and Cx45. Dedek et al. [98] have recently argued for heterotypic interactions between Cx36 and Cx45 in mouse retina based on overlap of Cx36 and Cx45 immunostaining; however, other investigators found no overlap [97], and such overlap, even when present, does not necessarily indicate functional interaction.

15.4.3 Neuron-Glia Coupling Several investigators have provided morphologic or functional evidence for neuron–glia gap junctions in brain tissue [108,112] or in culture [113]. Dezawa et al. [10] reported ‘‘small-scale gap junction–like structures’’ between regenerating peripheral axons and adjacent Schwann cells. However, Rash et al. [63] used FRIL to examine 229 replicas from brain that were labeled for combinations of Cx30, Cx32, and Cx43, but found that no gap junctions in neurons or their coupling partners were labeled for any of these glial connexins. An earlier FRIL study from this group on more than 300 replicas also failed to identify neuronal/glial gap junctions [114]. In contrast to the earlier reports of morphologic and functional neuron–glia coupling in locus ceruleus [112], a recent FRIL study of gap junction coupling in this region found no evidence for gap junctions between neurons and glia [115].

15 Nervous System

337

15.4.4 Roles of Neuronal Connexins The primary physiologic role of neuronal gap junctions in mature organisms is likely to be the establishment of electrotonic synapses and neuronal synchronization [83,116,117,118] (see Chapter 6).

15.5 Diseases Caused by Connexin Mutations Mutations in five glial connexins cause heritable disorders. Interestingly, mutations in two connexins of CNS glia do not lead to disease of the CNS. Deletion of CX26 typically leads to nonsyndromic deafness due to cochlear dysfunction (see Chapter 20). The lack of a CNS phenotype in patients homozygous for complete loss-of-function CX26 mutations [119] may be explained if CX26 is either not expressed in astrocytes or not critical for astrocyte function. CX30 mutation generally causes autosomal dominant skin disease (see Chapter 14) or deafness [120,121,122,123,124] or contributes to digenically inherited deafness [125,126], but there is no overt CNS manifestation. All identified CX47 mutations lead to CNS dysfunction without extraneural phenotype. CX43 mutations lead to a pleiotropic disorder that may involve the CNS. On the other hand, CX32 mutations consistently lead to PNS abnormalities but are only infrequently associated with clinically evident CNS dysfunction (which may be due to compensation by the often colocalized Cx47 [49]).

15.5.1 X-Linked Charcot-Marie-Tooth Disease Charcot-Marie-Tooth (CMT) disease is a group of inherited diseases of the PNS. The most common form of X-linked CMT (CMTX1) is due to mutation in the gene encoding CX32 [2]. Over 290 different mutations in CX32 have been associated with this disorder (see http://www.molgen.ua.ac.be/CMTMutations/). The clinical course is characterized by slowly progressive symptoms, including distal greater than proximal muscle atrophy and weakness, loss of deep tendon reflexes, and sensory loss. However, in some cases the clinical phenotype differs substantially [127,128]. Many patients exhibit subclinical abnormalities of visual and brainstem auditory evoked responses [129,130], and a small but significant number of patients have clinically significant CNS deficits including transient encephalopathy associated with white matter abnormalities on magnetic resonance imaging (MRI) [131,132], corticospinal tract dysfunction [131,133], or deafness [134,135]. CMTX patients have nerve conduction deficits [136], and peripheral nerve specimens from them are characterized by axonal loss, partially failed regeneration, and myelin abnormalities [137,138,139,140]. The axonal nature of these abnormalities, in a disorder where the mutant protein is expressed in Schwann cells but not axons, suggests that

338

C.K. Abrams and J.E. Rash

CMTX develops due to failure of the Schwann cell to provide metabolic or trophic support for normal axonal function. The absence of overt CNS manifestations in most patients and in the Cx32KO mouse suggests that Cx32 is not critical for CNS functioning. 15.5.1.1 Loss of Function Connexin32 Mutants Patients with (1) deletion of the entire coding sequence for the CX32 gene [141,142,143], (2) promoter mutations [144,145], or (3) nonsense mutations in the upstream part of the coding sequence [146] are all likely to represent loss-offunction mutations. A recent study by Shy et al. [147] shows that the phenotype of most patients with CMTX1 resembles that of patients lacking the entire coding region of CX32, also suggesting that most cases of CMTX are due to complete loss of function of CX32. The loss-of-function model is also supported by the identification of mutations that induce changes in cell–cell channel gating and permeability [29,148,149,150] consistent with loss of CX32 function. Several mutations in the coding region of the CX32 gene lead to a complete or nearly complete loss of channel-forming ability [149,151,152,153]. In addition, some mutant forms of CX32 traffic inappropriately [154,155,156] and are eliminated by the cellular quality control machinery [157]. Careful examination of the clinical phenotype of patients with specific CMTX mutations may shed light on the roles of CX32 in the human nervous system. 15.5.1.2 Gain of Function Connexin32 Mutants Some CMTX mutations may lead to (toxic) gain of function. First, transgenic mice expressing the CX32R142W mutant on a background of wild-type CX32 develop a demyelinating neuropathy not explicable by CX32 overexpression [158]. Second, co-injection of mRNA for CX32R142W and wild-type CX26 in paired Xenopus oocytes leads to lower levels of coupling compared to paired oocytes injected with CX26 alone or in combination with wild-type CX32 [151]. Third, the S85C [159] and the F235C [160] mutants of CX32 form abnormally active functional hemichannels in the plasma membrane of Xenopus oocytes; a few such open hemichannels in the Schwann cell plasma membrane may cause loss of metabolites or dissipation of ionic gradients. Fourth, CNS manifestations, seen in only a subset of patients, may be due to gain of function, though they may also be due to differences in environmental exposure or genetic background.

15.5.2 Oculodentodigital Dysplasia Oculodentodigital dysplasia (ODDD) is an autosomal dominant syndrome due to mutations in CX43. It is defined by oculofacial abnormalities,

15 Nervous System

339

syndactyly, and hypoplastic tooth enamel. Many patients also suffer from a wide range of neurologic manifestations including cognitive deficiencies, deafness, disorders of extraocular motility, leg or generalized muscle weakness, ataxia, and bladder disturbances [161]. There are also abnormalities on MRI of the brain, including hypointensity of the deep gray matter, thought to be due to iron deposition, and hyperintensity of the occipital and periventricular white matter [161,162,163]. Interestingly, the neurological abnormalities seem at least in part, to be mutation-specific, though the paucity of independent families sharing the same mutations makes it impossible to exclude other genetic or environmental contributions to the phenotypes. Unfortunately, many of the clinical descriptions predate the identification of CX43 mutations in ODDD patients [164], making genotype-phenotype correlation difficult. In the ODDD mouse model described by Flenniken et al. [165], detailed neurological evaluation was not performed, but MRI of the brain was normal and no obvious neurological abnormalities were seen. Functional analyses of CX43 ODDD mutants in model systems show that most lead to partial or complete loss of channel-forming ability [166,167,168,169,170], and many have dominant-negative effects on wild-type Cx43 [167,168,169,170]. However, several lines of evidence suggest that most if not all of the observed symptoms are not due to simple loss of function. First, the phenotype of the Cx43KO [171] (for details see [164]) and of the astrocyte-specific Cx43 [61] differ phenotypically in many respects from that seen in ODDD patients. Second, the wide phenotypic variability (much wider than CMTX) described in ODDD is unlikely to be due to varying degrees of haploinsufficiency of the gene for CX43. Supporting this notion, mutations associated with neurological dysfunction and those not causing neurological symptoms have shown similar degrees of loss of function and dominant-negative effects in several different studies [166,167,168]. Third, no human recessive mutation in CX43 predicted to lead to simple loss of function has been identified in ODDD. At this point, the general mechanism of disease in ODDD (loss-of-function due to haploinsufficiency, dominant/ transdominant-negative effects, or gain-of-function) remains to be elucidated, and may be different in different tissues. Similarly, the specific roles of ODDD mutations in CNS dysfunction remain obscure.

15.5.3 Pelizaeus Merzbacher–Like Disease Pelizaeus Merzbacher–like disease (PMLD) is an autosomal recessive disease associated with defects in CX47. It is named for its phenotypic resemblance to the X-linked disorder Pelizaeus Merzbacher disease, which is associated with mutations in proteolipoprotein, the major protein of CNS myelin [172]. The clinical course of PMLD is characterized by nystagmus (involuntary eye movements) within the first six months of life, cerebellar ataxia by about four years, and spasticity by six years of age [33,173,174,175]. All patients studied have

340

C.K. Abrams and J.E. Rash

extensive white matter disease with relative sparing of the corticospinal tracts on MRI, consistent with abnormal myelination [33,173,174,175], while MRI spectroscopy studies in two patients show normal or near normal choline, Nacetylaspartate and creatine levels (normal N-acetylaspartate levels indicate lack of significant axonal damage) [175]. The pathogenesis of PMLD remains to be clarified. One explanation, consistent with the recessive inheritance pattern and relatively uniform phenotype, is that PMLD arises due to loss of function of CX47 in oligodendrocytes. A recent study showed that each of three point mutations in CX47 known to cause PMLD leads to at least partial accumulation of CX47 in the endoplasmic reticulum [176]. In addition, none of these mutants was able to form functional gap junctions when assayed by electrical or dye-coupling in transfected mammalian cells. However, these studies do not rule out the possibility of gain of function for one or more of these mutants, since loss of normal function does not rule out the possibility of gain of a nonchannel abnormal function. The finding that mice with targeted ablation of the gene for Cx47 [50] show minimal CNS pathology and no phenotype is inconsistent with a pure loss-of-function model for PMLD, although mouse oligodendrocytes may be less dependent on expression of Cx47 for normal function than are human oligodendrocytes. As noted above, O-O coupling is likely to be minimal in the normal brain, with most oligodendrocyte connexins participating in heterotypic A-O gap junctions; thus, loss of Cx47 function would likely lead to reduced A-O coupling. It is possible that the A-O pathway mediated by Cx43–Cx47 junctions differs in some physiologically significant way from that of Cx30–Cx32 junctions, meaning that loss of the former may not be fully compensated (at least in humans) by the latter.

15.6 Other Nervous System Diseases with Connexin Involvement Although only mutations in glial connexins have been clearly implicated in the pathogenesis of inherited human nervous system disorders, both glial and neuronal gap junctions have been proposed to play a role in the pathogenesis of several human neurological diseases including epilepsy, tremor, and cerebral ischemia.

15.6.1 Seizures and Epilepsy Although often used interchangeably, the terms epilepsy and seizure are not synonymous. An electrographic seizure is an abnormal (excessive) electrical discharge, while clinical seizures are additionally associated with abnormal behavior or sensation. On the other hand, epilepsy is a chronic disease state characterized by repeated seizures. Thus acute, short-term experiments tend to

15 Nervous System

341

address the pathophysiology of seizures, while longer term, chronic experiments tend to be more relevant to the pathogenesis of epilepsy. Seizures result from pathological synchronous discharges of groups of neurons; since mammalian neuron–neuron gap junctions have generally been identified between inhibitory neurons, changes in the synchrony of inhibitory tone may underlie seizures and epilepsy. In fact, Traub et al. [177] have suggested that axo-axonal gap junctions may be important in pathogenesis of seizures and epilepsy. Several types of experiments have investigated the effects of modulation of gap junctional activity on seizure activity. These data have been recently reviewed [178,179]. The first type of study consists of comparisons of the effects of seizure provoking stimuli on preparations from wild-type mice or mice with targeted deletion for a particular connexin. For example, hippocampal slice preparations from Cx36KO mice show alterations in rhythmic oscillations and reduced 4-aminopyridine–induced seizure activity compared with slices prepared from wild-type mice [180]. These results suggest that reduced coherence of activity or connectivity among interneurons attenuates epileptic potential. However, it remains to be shown how this loss of inhibitory synchrony might lead to epilepsy, and using the Cx36KO mouse to examine the role of Cx36 in seizures may be complicated by chronic compensatory changes in that mouse CNS [181]. Deletion of Cx32 leads to both neuronal hyperexcitability and subtle changes in the structure of CNS myelin [20]. As outlined above, Cx32 has not been identified in neurons; thus, the changes in excitability may be secondary to changes in the myelinating oligodendrocyte. A second type of experiment, examining the effects of a mutation of the Drosophila gap junction protein shakB2 (an innexin), reveals that this loss of function mutation confers an unusually high threshold for evoked seizures [182]. Because the shakB2 mutant confers loss of synaptic electrical transmission, the authors argue that the antiepileptic effect results from impairment of synchronous neuronal activity mediated by gap junctions. The third type of experiment examines the effects of gap junction blockers and activators on seizure activity induced by a wide range of stimuli. These reports show a consistent antiepileptic effect of the blockers when applied in slice preparations [183,184,185,186] or in vivo [184,187,188,189,190,191]. However, these pharmacologic agents are generally nonspecific, having both junctional and nonjunctional effects [83] (see Chapter 8). Srinivas et al. [192] have shown that quinine shows relative selectivity for Cx36 over several other connexins. Gajda et al. [189] used this property to examine the effect of blockade of Cx36 on 4-aminopyridine–evoked seizure activity. When quinine was applied after onset of seizures the number of discharges increased, the frequency components of the discharges were altered, and the durations were reduced, leading to an overall reduction in seizure activity. However, quinine also has many effects on intrinsic neuronal properties that may reduce epileptogenic potential [193]. Yang and Ling [183] recently showed that carbenoxolone reduces the amplitude and frequency of inhibitory postsynaptic currents in rat

342

C.K. Abrams and J.E. Rash

somatosensory cortex slices, suggesting that the antiepileptic effect of carbenoxolone may be due to uncoupling of inhibitory interneurons.

15.6.2 Tremor Palatal tremor (PT, sometimes called palatal myoclonus) is a rare disorder associated with rhythmic movements of the soft palate and sometimes other muscles [194]. Essential tremor (ET) is a common, often familial disorder, characterized by postural and kinetic tremors of the head limbs and voice. Abnormal rhythmic output from the inferior olivary (IO) nucleus and olivocerebellar circuitry has been implicated in both disorders. In some cases of PT, olivary morphology is abnormal. In animals, harmaline produces enhanced neural synchrony in the olivocerebellar pathway and tremor similar to that of ET [195,196]. A related compound, harmine, causes tremor in humans. In the classical model first put forth by Llinas et al. [197] in 1974, IO coupling (now known to be due to Cx36; [181]) is regulated by g-aminobutyric acid (GABA)induced changes in membrane conductance, which lead to greater or lesser synchrony of olivocerebellar output. Similarly, harmaline increases rhythm generation in the IO, through alteration of rhythm generating ionic conductances [198]. It also has species-specific effects of olivary gliosis (in mice) and Purkinje cell death (in rats) [199]. Interestingly, a subset of brains from patients with ET shows evidence of Purkinje cell degeneration [200,201]. Although it seems logical to ascribe IO synchrony to gap junctional coupling, experimental data challenge that inference. Long and colleagues [202] showed that Cx36KO mice exhibit almost no coupling of IO neurons but still generate 5 to 10 Hz rhythmic membrane oscillations. Furthermore, harmaline induces tremor in Cx36KO mice that is indistinguishable from that in wild-type mice. De Zeeuw et al. [181] showed that the IO neurons in the Cx36KO mice show a number of changes in intrinsic neuronal properties that may compensate for loss of Cx36. This raises the possibility that IO gap junction coupling is required for generation of tremor in wild-type, and that the compensatory changes that occur in the Cx36KO are also sufficient to generate tremor. Experiments by Placantonakis et al. [203] utilized lentivirus-mediated expression of a Cx36 dominant-negative mutant to demonstrate that Cx36-mediated coupling is not required for generation of harmaline-induced tremor, although gap junction mediated IO coupling does contribute to coherence of muscle firing during tremor. Thus, available evidence suggests that harmaline-induced tremor and possibly ET and PT do not require IO junctional coupling. In spite of the evidence that Cx36-containing IO gap junctions are not required for generation of harmaline tremor, the gap junction blockers octanol, carbenoxolone, and mefloquine all suppress this tremor in the mouse model, while related compounds that do not block gap junctions, glycyrrhizic acid (related to carbenoxolone) and chloroquine (related to mefloquine), do not [204].

15 Nervous System

343

Table 15.3 Effects of connexin expression or gap junction modulation in ischemic injury: studies examining the role of connexins in ischemia Preparation Technique Experimental design Key findings Whole animal

Slice

Knockout mice

Compare WT with Cx43+/– heterozygotes in MCAO Compare WT with astrocyte-specific Cx43KO in MCAO Compare WT with Cx43+/– heterozygotes in MCAO Compare WT with Cx32KO in brief global ischemia

Blockers

Compare effects of global ischemia in rats with or without direct hippocampal injection carbenoxolone, glycyrrhizic acid, endothelin, or quinine Compare effect of MCAO in octanol pretreated and nontreated rats Compare effects of transient forebrain ischemia in octanol pretreated and nontreated rats Compare effect of MCAO in halothane treated and nontreated rats Compare effects of glucose/ O2 deprivation with or without carbenoxolone Compare effects of exposure to kainite and FeSO4 with and without glycyrrhizic acid Compare effects of glucose/ O2 deprivation with or without carbenoxolone

Blockers

Antisense

Increased apoptosis in Cx43þ/– [215] Increased apoptosis and inflammation in Cx43KO [209] Increased infarct volume in knockout [216] Increased neurodegeneration in Cx32KO hippocampus [217] Pretreatment with any but quinine reduces apoptotic neuronal death; early postischemic treatment also somewhat protective [214] Reduced infarct volume in pretreated mice [210] Reduced hippocampal neuronal damage in octanol pretreated rats [218] Reduced infarct volume in treated mice [219] Carbenoxolone decreases cell death [220] GA exacerbated kainateinduced and FeSO4induced injury [221] Knockdown of Cx26 and Cx32 or Cx43 decrease cell death from glucose/O2 deprivation [220] Death signals are propagated through gap junctions [222]

Test for effects of Ca2+ ionophoresis, metabolic inhibition, free radical formation Blockers Compare effects of exposure Glycyrrhizic acid increases cell death [221] to FeSO4 and 4hydroxynonenal with and without glycyrrhizic acid MCAO, middle cerebral artery occlusion; WT, wild-type; KO, knockout. Culture

Exogenous expression of Cx43

344

C.K. Abrams and J.E. Rash

As noted above, these blockers are not specific, but the result raises the possibility that the pathogenesis of this tremor involves gap junctional coupling at some level.

15.6.3 Ischemia Effects of gap junctions in experimental ischemia depend on whether one uses a genetic or pharmacologic approach to modulate the level of gap junctional or connexin activity. As shown in Table 15.3, blockade of gap junction channels by a variety of agents appears neuroprotective, while in genetic models reduction or deletion of connexin expression lead to greater damage in ischemic insult. From a theoretical standpoint either conclusion makes sense, since gap junctions may increase the clearance of harmful substances or the delivery of protective ones [205,206]. As noted above, gap junction blockers lack specificity, and animal models may have secondary changes that are not easily predictable; on the other hand, connexins may have biological roles independent of gap junction formation that may not be affected by gap junction blockers [207]. Spreading depression has been shown to increase the damage that occurs as a result of focal ischemia [208]. One consistent result between the data from knockout mice and blocker experiments is that reductions in spreading depression correlate with protection from ischemia. Nakase et al. [209] observed that mice lacking Cx43 in astrocytes showed increased infarct volume and enhanced apoptosis after middle cerebral artery occlusion. Increased velocity of spreading depression was also seen in these animals. In a separate experiment, octanol treatment, which reduced stroke infarct volume in rats, also reduced spreading depression [210]. The action of connexins in ischemia may relate to opening of hemichannels in the plasma membrane and thus contribute to ischemic damage [77,211,212], either through release of excitatory neurotransmitter [213] or through disruption of cellular ionic gradients or loss of metabolites. Recently, it has been shown that carbenoxolone reduces lipid peroxide formation; this may influence the beneficial effect of carbenoxolone in ischemia [214]. However, it is unclear whether this benefit is mediated through modulation of gap junctional communication.

15.7 Conclusion This chapter summarized some of what is and is not known about glial and neuronal connexins, especially as pertaining to pathogenesis of human disease. In the last 15 years, specific connexins have been definitively identified in neurons, Schwann cells, oligodendrocytes, and astrocytes. Mutations in connexin genes or altered expression of wild-type gap junction proteins may play

15 Nervous System

345

important roles in pathogenesis of diseases such as epilepsy. The major challenge going forward is to definitively identify the physiological roles of these molecules and the cells in which they are normally expressed; this in turn will allow for rational therapeutic approaches to treat connexin-related diseases of the nervous system. Acknowledgments This work is supported by National Institutes of Health (NIH) grants 1K02NS50345 and 1R01NS050705 to CKA.

References 1. Raine CS, Wisniewski H, Prineas J. An ultrastructural study of experimental demyelination and remyelination. II. Chronic experimental allergic encephalomyelitis in the peripheral nervous system. Lab Invest. 1969;21:316–27. 2. Bergoffen J, Scherer SS, Wang S, Scott MO, Bone LJ, Paul DL, Chen K, Lensch MW, Chance PF, Fischbeck KH. Connexin mutations in X-linked Charcot-Marie-Tooth disease. Science 1993;262:2039–42. 3. Li X, Lynn BD, Olson C, Meier C, Davidson KG, Yasumura T, Rash JE, Nagy JI. Connexin29 expression, immunocytochemistry and freeze-fracture replica immunogold labeling (FRIL) in sciatic nerve. Eur J Neurosci. 2002;16:795–806. 4. Altevogt BM, Kleopa KA, Postma FR, Scherer SS, Paul DL. Connexin29 is uniquely distributed within myelinating glial cells of the central and peripheral nervous systems. J Neurosci. 2002;22:6458–70. 5. Sohl ¨ G, Eiberger J, Jung YT, Kozak CA, Willecke K. The mouse gap junction gene connexin29 is highly expressed in sciatic nerve and regulated during brain development. Biol Chem. 2001;382:973–8. 6. Jessen KR, Mirsky R. The origin and development of glial cells in peripheral nerves. Nat Rev Neurosci. 2005;6:671–82. 7. Carroll SL, Miller ML, Frohnert PW, Kim SS, Corbett JA. Expression of neuregulins and their putative receptors, ErbB2 and ErbB3, is induced during Wallerian degeneration. J Neurosci. 1997;17:1642–59. 8. Tetzlaff W. Tight junction contact events and temporary gap junctions in the sciatic nerve fibers of the chicken during Wallerian degeneration and subsequent regeneration. J Neurocytol. 1982;11:839–58. 9. Konishi T. Dye coupling between mouse Schwann cells. Brain Res. 1990;508:85–92. 10. Dezawa M, Mutoh T, Dezawa A, Adachi-Usami E. Putative gap junctional communication between axon and regenerating Schwann cells during mammalian peripheral nerve regeneration. Neuroscience 1998;85:663–7. 11. Dezawa M, Nagano T. Contacts between regenerating axons and the Schwann cells of sciatic nerve segments grafted to the optic nerve of adult rats. J Neurocytol. 1993;22:1103–12. 12. Rosenstein JM, Brightman MW. Regeneration and myelination in autonomic ganglia transplanted to intact brain surfaces. J Neurocytol. 1979;8:359–79. 13. Chanson M, Fanjul M, Bosco D, Nelles E, Suter S, Willecke K, Meda P. Enhanced secretion of amylase from exocrine pancreas of connexin32-deficient mice. J Cell Biol. 1998;141:1267–75. 14. Chandross KJ, Spray DC, Cohen RI, Kumar NM, Kremer M, Dermietzel R, Kessler JA. TNF inhibits Schwann cell proliferation, connexin46 expression, and gap junctional communication. Mol Cell Neurosci. 1996;7:479–500.

346

C.K. Abrams and J.E. Rash

15. Chandross KJ, Chanson M, Spray DC, Kessler JA. Transforming growth factor-b 1 and forskolin modulate gap junctional communication and cellular phenotype of cultured Schwann cells. J Neurosci. 1995;15:262–73. 16. Schnapp BJ, Mugnaini E. Membrane architecture of myelinated fibers as seen by freezefracture. In: Waxman SG, editor. Physiology and pathobiology of axons. New York: Raven Press; 1978. pp. 82–123. 17. Sandri C, Van Buren JM, Akert K. Membrane morphology of the vertebrate nervous system. A study with freeze-etch technique. Prog Brain Res. 1977;46:1–384. 18. Scherer SS, Deschenes SM, Xu YT, Grinspan JB, Fischbeck KH, Paul DL. Connexin32 is a myelin-related protein in the PNS and CNS. J Neurosci. 1995;15:8281–94. 19. Balice-Gordon RJ, Bone LJ, Scherer SS. Functional gap junctions in the Schwann cell myelin sheath. J Cell Biol. 1998;142:1095–104. 20. Sutor B, Schmolke C, Teubner B, Schirmer C, Willecke K. Myelination defects and neuronal hyperexcitability in the neocortex of connexin 32-deficient mice. Cereb Cortex. 2000;10:684–97. 21. Chandross KJ, Kessler JA, Cohen RI, Simburger E, Spray DC, Bieri P, Dermietzel R. Altered connexin expression after peripheral nerve injury. Mol Cell Neurosci. 1996;7:501–18. 22. Mambetisaeva ET, Gire V, Evans WH. Multiple connexin expression in peripheral nerve, Schwann cells, and Schwannoma cells. J Neurosci Res. 1999;57:166–75. 23. Meier C, Dermietzel R, Davidson KG, Yasumura T, Rash JE. Connexin32-containing gap junctions in Schwann cells at the internodal zone of partial myelin compaction and in Schmidt-Lanterman incisures. J Neurosci. 2004;24:3186–98. 24. Satake M, Yoshimura T, Ohnishi A, Kobayashi T. Connexin32 gene expression in rat sciatic nerves and cultured Schwann cells. Dev Neurosci. 1997;19:189–95. 25. Miller RG, da Silva PP. Particle rosettes in the periaxonal Schwann cell membrane and particle clusters in the axolemma of rat sciatic nerve. Brain Res. 1977;130:135–41. 26. Stolinski C, Breathnach AS, Martin B, Thomas PK, King RH, Gabriel G. Associated particle aggregates in juxtaparanodal axolemma and adaxonal Schwann cell membrane of rat peripheral nerve. J Neurocytol. 1981;10:679–91. 27. Anzini P, Neuberg DH, Schachner M, Nelles E, Willecke K, Zielasek J, Toyka KV, Suter U, Martini R. Structural abnormalities and deficient maintenance of peripheral nerve myelin in mice lacking the gap junction protein connexin 32. J Neurosci. 1997;17:4545–51. 28. Scherer SS, Xu YT, Nelles E, Fischbeck K, Willecke K, Bone LJ. Connexin32-null mice develop demyelinating peripheral neuropathy. Glia. 1998;24:8–20. 29. Oh S, Ri Y, Bennett MVL, Trexler EB, Verselis VK, Bargiello TA. Changes in permeability caused by connexin 32 mutations underlie X-linked Charcot-Marie-Tooth disease. Neuron 1997;19:927–38. 30. Teubner B, Odermatt B, Guldenagel M, Sohl ¨ G, Degen J, Bukauskas F, Kronengold J, Verselis VK, Jung YT, Kozak CA, Schilling K, Willecke K. Functional expression of the new gap junction gene connexin47 transcribed in mouse brain and spinal cord neurons. J Neurosci. 2001;21:1117–26. 31. Odermatt B, Wellershaus K, Wallraff A, Seifert G, Degen J, Euwens C, Fuss B, Bussow H, Schilling K, Steinhauser C, Willecke K. Connexin 47 (Cx47)-deficient mice with enhanced green fluorescent protein reporter gene reveal predominant oligodendrocytic expression of Cx47 and display vacuolized myelin in the CNS. J Neurosci. 2003;23:4549–59. 32. Kleopa KA, Orthmann JL, Enriquez A, Paul DL, Scherer SS. Unique distributions of the gap junction proteins connexin29, connexin32, and connexin47 in oligodendrocytes. Glia. 2004;47:346–57. 33. Uhlenberg B, Schuelke M, Ruschendorf F, Ruf N, Kaindl AM, Henneke M, Thiele H, Stoltenburg-Didinger G, Aksu F, Topaloglu H, Nurnberg P, Hubner C, Weschke B, Gartner J. Mutations in the gene encoding gap junction protein 12 (connexin 46.6) cause Pelizaeus-Merzbacher-like disease. Am J Hum Genet. 2004;75:251–60.

15 Nervous System

347

34. Toews JC, Schram V, Weerth SH, Mignery GA, Russell JT. Signaling proteins in the axoglial apparatus of sciatic nerve nodes of Ranvier. Glia. 2007;55:202–13. 35. Li J, Habbes HW, Eiberger J, Willecke K, Dermietzel R, Meier C. Analysis of connexin expression during mouse Schwann cell development identifies connexin29 as a novel marker for the transition of neural crest to precursor cells. Glia. 2007;55:93–103. 36. Eiberger J, Kibschull M, Strenzke N, Schober A, Bussow H, Wessig C, Djahed S, Reucher H, Koch DA, Lautermann J, Moser T, Winterhager E, Willecke K. Expression pattern and functional characterization of connexin29 in transgenic mice. Glia. 2006;53:601–11. 37. Tang W, Zhang Y, Chang Q, Ahmad S, Dahlke I, Yi H, Chen P, Paul DL, Lin X. Connexin29 is highly expressed in cochlear Schwann cells, and it is required for the normal development and function of the auditory nerve of mice. J Neurosci. 2006;26:1991–9. 38. Brightman MW, Reese TS. Junctions between intimately apposed cell membranes in the vertebrate brain. J Cell Biol. 1969;40:648–77. 39. Quigley HA. Gap junctions between optic nerve head astrocytes. Invest Ophthalmol Visual Sci. 1977;16:582–5. 40. Dermietzel R. Junctions in the central nervous system of the cat. 3. Gap junctions and membrane-associated orthogonal particle complexes (MOPC) in astrocytic membranes. Cell Tissue Res. 1974;149:121–35. 41. Massa PT, Mugnaini E. Cell junctions and intramembrane particles of astrocytes and oligodendrocytes: a freeze-fracture study. Neuroscience 1982;7:523–38. 42. Rash JE, Staines WA, Yasumura T, Patel D, Furman CS, Stelmack GL, Nagy JI. Immunogold evidence that neuronal gap junctions in adult rat brain and spinal cord contain connexin-36 but not connexin-32 or connexin-43. Proc Natl Acad Sci USA. 2000;97:7573–8. 43. Rash JE, Olson CO, Pouliot WA, Davidson KG, Yasumura T, Furman CS, Royer S, Kamasawa N, Nagy JI, Dudek FE. Connexin36 vs. connexin32, ‘miniature’ neuronal gap junctions, and limited electrotonic coupling in rodent suprachiasmatic nucleus. Neuroscience 2007;149:350–371. 44. Altevogt BM, Paul DL. Four classes of intercellular channels between glial cells in the CNS. J Neurosci. 2004;24:4313–23. 45. Nagy JI, Ionescu AV, Lynn BD, Rash JE. Connexin29 and connexin32 at oligodendrocyte and astrocyte gap junctions and in myelin of the mouse central nervous system. J Comp Neurol. 2003;464:356–70. 46. Nagy JI, Li X, Rempel J, Stelmack G, Patel D, Staines WA, Yasumura T, Rash JE. Connexin26 in adult rodent central nervous system: demonstration at astrocytic gap junctions and colocalization with connexin30 and connexin43. J Comp Neurol. 2001;441:302–23. 47. Nagy JI, Ionescu AV, Lynn BD, Rash JE. Coupling of astrocyte connexins Cx26, Cx30, Cx43 to oligodendrocyte Cx29, Cx32, Cx47: implications from normal and connexin32 knockout mice. Glia. 2003;44:205–18. 48. Filippov MA, Hormuzdi SG, Fuchs EC, Monyer H. A reporter allele for investigating connexin 26 gene expression in the mouse brain. Eur J Neurosci. 2003;18:3183–92. 49. Kamasawa N, Sik A, Morita M, Yasumura T, Davidson KG, Nagy JI, Rash JE. Connexin-47 and connexin-32 in gap junctions of oligodendrocyte somata, myelin sheaths, paranodal loops and Schmidt-Lanterman incisures: implications for ionic homeostasis and potassium siphoning. Neuroscience 2005;136:65–86. 50. Menichella DM, Goodenough DA, Sirkowski E, Scherer SS, Paul DL. Connexins are critical for normal myelination in the CNS. J Neurosci. 2003;23:5963–73. 51. Kunzelmann P, Blumcke I, Traub O, Dermietzel R, Willecke K. Coexpression of connexin45 and -32 in oligodendrocytes of rat brain. J Neurocytol. 1997;26:17–22. 52. Pastor A, Kremer M, Moller T, Kettenmann H, Dermietzel R. Dye coupling between spinal cord oligodendrocytes: differences in coupling efficiency between gray and white matter. Glia. 1998;24:108–20.

348

C.K. Abrams and J.E. Rash

53. Kruger O, Plum A, Kim JS, Winterhager E, Maxeiner S, Hallas G, Kirchhoff S, Traub O, Lamers WH, Willecke K. Defective vascular development in connexin 45-deficient mice. Development 2000;127:4179–93. 54. Li X, Simard JM. Connexin45 gap junction channels in rat cerebral vascular smooth muscle cells. Am J Physiol Heart Circ Physiol. 2001;281:H1890–8. 55. Li X, Simard JM. Increase in Cx45 gap junction channels in cerebral smooth muscle cells from SHR. Hypertension 2002;40:940–6. 56. Rash JE, Davidson KG, Kamasawa N, Yasumura T, Kamasawa M, Zhang C, Michaels R, Restrepo D, Ottersen OP, Olson CO, Nagy JI. Ultrastructural localization of connexins (Cx36, Cx43, Cx45), glutamate receptors and aquaporin-4 in rodent olfactory mucosa, olfactory nerve and olfactory bulb. J Neurocytol. 2005;34:307–41. 57. Nagy JI, Rash JE. Connexins and gap junctions of astrocytes and oligodendrocytes in the CNS. Brain Res Brain Res Rev. 2000;32:29–44. 58. Nagy JI, Dudek FE, Rash JE. Update on connexins and gap junctions in neurons and glia in the mammalian nervous system. Brain Res Brain Res Rev. 2004;47:191–215. 59. Orthmann-Murphy JL, Freidin M, Fischer E, Scherer SS, Abrams CK. Two distinct heterotypic channels mediate gap junction coupling between astrocyte and oligodendrocyte connexions. J Neurosci. 2007;27:13949–57. 60. Wallraff A, Kohling R, Heinemann U, Theis M, Willecke K, Steinhauser C. The impact of astrocytic gap junctional coupling on potassium buffering in the hippocampus. J Neurosci. 2006; 26:5438–47. 61. Theis M, Jauch R, Zhuo L, Speidel D, Wallraff A, Doring B, Frisch C, Sohl ¨ G, Teubner B, Euwens C, Huston J, Steinhauser C, Messing A, Heinemann U, Willecke K. Accelerated hippocampal spreading depression and enhanced locomotory activity in mice with astrocyte-directed inactivation of connexin43. J Neurosci. 2003;23:766–76. 62. Wallraff A, Odermatt B, Willecke K, Steinhauser C. Distinct types of astroglial cells in the hippocampus differ in gap junction coupling. Glia. 2004;48:36–43. 63. Rash JE, Yasumura T, Dudek FE, Nagy JI. Cell-specific expression of connexins and evidence of restricted gap junctional coupling between glial cells and between neurons. J Neurosci. 2001;21:1983–2000. 64. Rash JE, Duffy HS, Dudek FE, Bilhartz BL, Whalen LR, Yasumura T. Grid-mapped freeze-fracture analysis of gap junctions in gray and white matter of adult rat central nervous system, with evidence for a ‘panglial syncytium’ that is not coupled to neurons. J Comp Neurol. 1997;388:265–92. 65. Massa PT, Szuchet S, Mugnaini E. Cell-cell interactions of isolated and cultured oligodendrocytes: formation of linear occluding junctions and expression of peculiar intramembrane particles. J Neurosci. 1984;4:3128–39. 66. Kettenmann H, Orkand RK, Schachner M. Coupling among identified cells in mammalian nervous system cultures. J Neurosci. 1983;3:506–16. 67. Kettenmann H, Ransom BR. Electrical coupling between astrocytes and between oligodendrocytes studied in mammalian cell cultures. Glia. 1988;1:64–73. 68. Ransom BR, Kettenmann H. Electrical coupling, without dye coupling, between mammalian astrocytes and oligodendrocytes in cell culture. Glia. 1990;3:258–66. 69. Dermietzel R, Farooq M, Kessler JA, Althaus H, Hertzberg EL, Spray DC. Oligodendrocytes express gap junction proteins connexin32 and connexin45. Glia. 1997;20:101–14. 70. Mugnaini E. Cell junctions of astrocytes, ependyma, and related cells in the mammalian central nervous system, with emphasis on the hypothesis of a generalized functional syncytium of supporting cells. In: Fedoroff S, Vernadakis A, editors. Astrocytes, vol I. New York: Academic Press, 1986. pp. 329–71. 71. Scemes E, Giaume C. Astrocyte calcium waves: what they are and what they do. Glia. 2006; 54:716–25.

15 Nervous System

349

72. Neusch C, Rozengurt N, Jacobs RE, Lester HA, Kofuji P. Kir4.1 potassium channel subunit is crucial for oligodendrocyte development and in vivo myelination. J Neurosci. 2001;21:5429–38. 73. Higashi K, Fujita A, Inanobe A, Tanemoto M, Doi K, Kubo T, Kurachi Y. An inwardly rectifying K+ channel, Kir4.1, expressed in astrocytes surrounds synapses and blood vessels in brain. Am J Physiol Cell Physiol. 2001;281:C922–31. 74. Li L, Head V, Timpe LC. Identification of an inward rectifier potassium channel gene expressed in mouse cortical astrocytes. Glia. 2001;33:57–71. 75. Menichella DM, Majdan M, Awatramani R, Goodenough DA, Sirkowski E, Scherer SS, Paul DL. Genetic and physiological evidence that oligodendrocyte gap junctions contribute to spatial buffering of potassium released during neuronal activity. J Neurosci. 2006;26:10984–91. 76. Jiang JX, Gu S. Gap junction- and hemichannel-independent actions of connexins. Biochim Biophys Acta. 2005;1711:208–14. 77. Contreras JE, Sanchez HA, Eugenin EA, Speidel D, Theis M, Willecke K, Bukauskas FF, Bennett MV, Saez JC. Metabolic inhibition induces opening of unapposed connexin 43 gap junction hemichannels and reduces gap junctional communication in cortical astrocytes in culture. Proc Natl Acad Sci USA. 2002;99:495–500. 78. Dermietzel R, Meier C, Bukauskas F, Spray DC. Following tracks of hemichannels. Cell Commun Adhes. 2003;10:335–40. 79. Sa´ez JC, Contreras JE, Bukauskas FF, Retamal MA, Bennett MVL. Gap junction hemichannels in astrocytes of the CNS. Acta Physiol Scand. 2003;179:9–22. 80. Retamal MA, Cortes CJ, Reuss L, Bennett MV, Sa´ez JC. S-nitrosylation and permeation through connexin 43 hemichannels in astrocytes: induction by oxidant stress and reversal by reducing agents. Proc Natl Acad Sci USA. 2006;103:4475–80. 81. Stout CE, Costantin JL, Naus CC, Charles AC. Intercellular calcium signaling in astrocytes via ATP release through connexin hemichannels. J Biol Chem. 2002;277:10482–8. 82. Bennett M, Crain S, Grundfest H. Electrophysiology of supramedullary neuronsin Spheroides maculates: I. Orthodromic and antidromic responses. J Gen Physiol. 1959;43:159–88. 83. Connors BW, Long MA. Electrical synapses in the mammalian brain. Annu Rev Neurosci. 2004;27:393–418. 84. Schmitz D, Schuchmann S, Fisahn A, Draguhn A, Buhl EH, Petrasch-Parwez E, Dermietzel R, Heinemann U, Traub RD. Axo-axonal coupling. A novel mechanism for ultrafast neuronal communication. Neuron 2001;31:831–40. 85. Draguhn A, Traub RD, Schmitz D, Jefferys JG. Electrical coupling underlies high frequency oscillations in the hippocampus in vitro. Nature 1998;394:189–92. 86. Traub RD, Bibbig R, Piechotta A, Draguhn R, Schmitz D. Synaptic and nonsynaptic contributions to giant ipsps and ectopic spikes induced by 4-aminopyridine in the hippocampus in vitro. J Neurophysiol. 2001;85:1246–56. 87. Hamzei-Sichani F, Kamasawa N, Janssen WGM, Yasumura T, Hof PR, Wearne SL, Stewart MG, Young SR, Whittington MA, Rash JE, Traub RD. Evidence for gap junctions on hippocampal mossy fiber axons using thin-section and freeze-fracture replica immunogold labeling electron microscopy. Proc Nat Acad Sci USA. 2007;104:12548–53. 88. O’Brien J, Bruzzone R, White TW, Al-Ubaidi MR, Ripps H. Cloning and expression of two related connexins from the perch retina define a distinct subgroup of the connexin family. J Neurosci. 1998;18:7625–37. 89. Condorelli DF, Parenti R, Spinella F, Trovato Salinaro A, Belluardo N, Cardile V, Cicirata F. Cloning of a new gap junction gene (Cx36) highly expressed in mammalian brain neurons. Eur J Neurosci. 1998;10:1202–8. 90. Sohl ¨ G, Degen J, Teubner B, Willecke K. The murine gap junction gene connexin36 is highly expressed in mouse retina and regulated during brain development. FEBS Lett. 1998;428:27–31.

350

C.K. Abrams and J.E. Rash

91. Teubner B, Degen J, Sohl ¨ G, Guldenagel M, Bukauskas FF, Trexler EB, Verselis VK, De Zeeuw CI, Lee CG, Kozak CA, Petrasch-Parwez E, Dermietzel R, Willecke K. Functional expression of the murine connexin 36 gene coding for a neuron-specific gap junctional protein. J Membr Biol. 2000;176:249–62. 92. Condorelli DF, Trovato-Salinaro A, Mudo G, Mirone MB, Belluardo N. Cellular expression of connexins in the rat brain: neuronal localization, effects of kainateinduced seizures and expression in apoptotic neuronal cells. Eur J Neurosci. 2003;18:1807–27. 93. Maxeiner S, Kruger O, Schilling K, Traub O, Urschel S, Willecke K. Spatiotemporal transcription of connexin45 during brain development results in neuronal expression in adult mice. Neuroscience 2003;119:689–700. 94. Zhang C, Restrepo D. Expression of connexin 45 in the olfactory system. Brain Res. 2002;929:37–47. 95. Guldenagel M, Sohl ¨ G, Plum A, Traub O, Teubner B, Weiler R, Willecke K. Expression patterns of connexin genes in mouse retina. J Comp Neurol. 2000;425:193–201. 96. Sohl ¨ G, Guldenagel M, Traub O, Willecke K. Connexin expression in the retina. Brain Res Brain Res Rev. 2000;32:138–45. 97. Han Y, Massey SC. Electrical synapses in retinal ON cone bipolar cells: subtype-specific expression of connexins. Proc Natl Acad Sci USA. 2005;102:13313–8. 98. Dedek K, Schultz K, Pieper M, Dirks P, Maxeiner S, Willecke K, Weiler R, JanssenBienhold U. Localization of heterotypic gap junctions composed of connexin45 and connexin36 in the rod pathway of the mouse retina. Eur J Neurosci. 2006;24:1675–86. 99. Maxeiner S, Dedek K, Janssen-Bienhold U, Ammermuller J, Brune H, Kirsch T, Pieper M, Degen J, Kruger O, Willecke K, Weiler R. Deletion of connexin45 in mouse retinal neurons disrupts the rod/cone signaling pathway between AII amacrine and ON cone bipolar cells and leads to impaired visual transmission. J Neurosci. 2005; 25:566–76. 100. Chang Q, Gonzalez M, Pinter MJ, Balice-Gordon RJ. Gap junctional coupling and patterns of connexin expression among neonatal rat lumbar spinal motor neurons. J Neurosci. 1999;19:10813–28. 101. Chang Q, Balice-Gordon RJ. Gap junctional communication among developing and injured motor neurons. Brain Res Brain Res Rev. 2000;32:242–9. 102. Eiberger J, Degen J, Romualdi A, Deutsch U, Willecke K, Sohl ¨ G. Connexin genes in the mouse and human genome. Cell Commun Adhes. 2001;8:163–5. 103. Hombach S, Janssen-Bienhold U, Sohl ¨ G, Schubert T, Bussow H, Ott T, Weiler R, Willecke K. Functional expression of connexin57 in horizontal cells of the mouse retina. Eur J Neurosci. 2004;19:2633–40. 104. O’Brien JJ, Li W, Pan F, Keung J, O’Brien J, Massey SC. Coupling between A-type horizontal cells is mediated by connexin 50 gap junctions in the rabbit retina. J Neurosci. 2006;26:11624–36. 105. Belliveau DJ, Naus CC. Cellular localization of gap junction mRNAs in developing rat brain. Dev Neurosci. 1995;17:81–96. 106. Zhang C, Finger TE, Restrepo D. Mature olfactory receptor neurons express connexin 43. J Comp Neurol. 2000;426:1–12. 107. Priest CA, Thompson AJ, Keller A. Gap junction proteins in inhibitory neurons of the adult barrel neocortex. Somatosens Motor Res. 2001;18:245–52. 108. Nadarajah B, Thomaidou D, Evans WH, Parnavelas JG. Gap junctions in the adult cerebral cortex: regional differences in their distribution and cellular expression of connexins. J Comp Neurol. 1996;376:326–42. 109. Micevych PE, Popper P, Hatton GI. Connexin 32 mRNA levels in the rat supraoptic nucleus: upregulation prior to parturition and during lactation. Neuroendocrinology 1996;63:39–45.

15 Nervous System

351

110. Micevych PE, Abelson L. Distribution of mRNAs coding for liver and heart gap junction proteins in the rat central nervous system. J Comp Neurol. 1991;305:96–118. 111. Dermietzel R, Traub O, Hwang TK, Beyer E, Bennett MV, Spray DC, Willecke K. Differential expression of three gap junction proteins in developing and mature brain tissues. Proc Natl Acad Sci USA. 1989;86:10148–52. 112. Alvarez-Maubecin V, Garcia-Hernandez F, Williams JT, Van Bockstaele EJ. Functional coupling between neurons and glia. J Neurosci. 2000;20:4091–8. 113. Froes MM, Correia AH, Garcia-Abreu J, Spray DC, Campos de Carvalho AC, Neto MVL. Gap-junctional coupling between neurons and astrocytes in primary central nervous system cultures. Proc Natl Acad Sci USA. 1999;96:7541–6. 114. Rash JE, Yasumura T, Hudson CS, Agre P, Nielsen S. Direct immunogold labeling of aquaporin-4 in square arrays of astrocyte and ependymocyte plasma membranes in rat brain and spinal cord. Proc Natl Acad Sci USA. 1998;95:11981–6. 115. Rash JE, Olson CO, Davidson KGV, Yasumura T, Kamasawa N, Nagy JI. Identification of connexin36 in gap junctions between neurons in rodent locus coeruleus. Neuroscience 2007;147:938–56. 116. Meier C, Dermietzel R. Electrical synapses-gap junctions in the brain. Res Problems Cell Different. 2006;43:99–128. 117. Sohl ¨ G, Maxeiner S, Willecke K. Expression and functions of neuronal gap junctions. Nat Rev Neurosci. 2005;6:191–200. 118. Bennett MVL, Zukin RS. Electrical coupling and neuronal synchronization in the mammalian brain. Neuron 2004;41:495–511. 119. Kelsell DP, Dunlop J, Stevens HP, Lench NJ, Liang JN, Parry G, Mueller RF, Leigh IM. Connexin 26 mutations in hereditary nonsyndromic sensorineural deafness. Nature 1997;387:80–3. 120. Common JE, Becker D, Di WL, Leigh IM, O’Toole EA, Kelsell DP. Functional studies of human skin disease- and deafness-associated connexin 30 mutations. Biochem Biophys Res Commun. 2002;298:651–6. 121. Common JE, O’Toole EA, Leigh IM, Thomas A, Griffiths WA, Venning V, Grabczynska S, Peris Z, Kansky A, Kelsell DP. Clinical and genetic heterogeneity of erythrokeratoderma variabilis. J Invest Dermatol. 2005;125:920–7. 122. Essenfelder GM, Bruzzone R, Lamartine J, Charollais A, Blanchet-Bardon C, Barbe MT, Meda P, Waksman G. Connexin30 mutations responsible for hidrotic ectodermal dysplasia cause abnormal hemichannel activity. Hum Mol Genet. 2004;13:1703–14. 123. Jan AY, Amin S, Ratajczak P, Richard G, Sybert VP. Genetic heterogeneity of KID syndrome: identification of a Cx30 gene (GJB6) mutation in a patient with KID syndrome and congenital atrichia. J Invest Dermatol. 2004;122:1108–13. 124. Palmada M, Schmalisch K, Bohmer C, Schug N, Pfister M, Lang F, Blin N. Loss of function mutations of the GJB2 gene detected in patients with DFNB1-associated hearing impairment. Neurobiol Dis. 2006;22:112–8. 125. Del Castillo I, Moreno-Pelayo MA, Del Castillo FJ, Brownstein Z, Marlin S, Adina Q, Cockburn DJ, Pandya A, Siemering KR, Chamberlin GP, Ballana E, Wuyts W, Maciel-Guerra AT, Alvarez A, Villamar M, Shohat M, Abeliovich D, Dahl HH, Estivill X, Gasparini P, Hutchin T, Nance WE, Sartorato EL, Smith RJ, Van Camp G, Avraham KB, Petit C, Moreno F. Prevalence and evolutionary origins of the del(GJB6D13S1830) mutation in the DFNB1 locus in hearing-impaired subjects: a multicenter study. Am J Hum Genet. 2003;73:1452–8. 126. del Castillo FJ, Rodriguez-Ballesteros M, Alvarez A, Hutchin T, Leonardi E, de Oliveira CA, Azaiez H, Brownstein Z, Avenarius MR, Marlin S, Pandya A, Shahin H, Siemering KR, Weil D, Wuyts W, Aguirre LA, Martin Y, MorenoPelayo MA, Villamar M, Avraham KB, Dahl HH, Kanaan M, Nance WE, Petit C, Smith RJ, Van Camp G, Sartorato EL, Murgia A, Moreno F, del Castillo I. A novel

352

127. 128.

129.

130. 131.

132.

133.

134.

135.

136. 137.

138.

139.

140.

141.

142.

C.K. Abrams and J.E. Rash deletion involving the connexin-30 gene, del(GJB6-d13s1854), found in trans with mutations in the GJB2 gene (connexin-26) in subjects with DFNB1 nonsyndromic hearing impairment. J Med Genet. 2005;42:588–94. Lin G, Glass J, Scherer S, Fischbeck K. A unique mutation in Cx32 associated with severe, early onset CMTX in a heterozygous female. Ann NY Acad Sci. 1999;14:481–4. Felice KJ, Seltzer WK. Severe X-linked Charcot-Marie-Tooth neuropathy due to new mutations G59R(G>C), W44X(G>A) in the connexin 32 gene. Eur Neurol. 2000;44:61–3. Nicholson G, Corbett A. Slowing of central conduction in X-linked Charcot-MarieTooth neuropathy shown by brain stem auditory evoked responses. J Neurol Neurosurg Psychiatr. 1996;61:43–6. Nicholson GA, Yeung L, Corbett A. Efficient neurophysiologic selection of X-linked Charcot-Marie-Tooth families: ten novel mutations. Neurology 1998;51:1412–6. Hanemann CO, Bergmann C, Senderek J, Zerres K, Sperfeld AD. Transient, recurrent, white matter lesions in X-linked Charcot-Marie-Tooth disease with novel connexin 32 mutation. Arch Neurol. 2003;60:605–9. Paulson HL, Garbern JY, Hoban TF, Krajewski KM, Lewis RA, Fischbeck KH, Grossman RI, Lenkinski R, Kamholz JA, Shy ME. Transient central nervous system white matter abnormality in X-linked Charcot-Marie-Tooth disease. Ann Neurol. 2002;52:429–34. Marques W Jr, Sweeney JG, Wood NW, Wroe SJ, Marques W. Central nervous system involvement in a novel connexin 32 mutation affecting identical twins. J Neurol Neurosurg Psychiatr. 1999;66:803–4. Bahr M, Andres F, Timmerman V, Nelis ME, Van Broeckhoven C, Dichgans J. Central visual, acoustic, and motor pathway involvement in a Charcot-Marie-Tooth family with an Asn205Ser mutation in the connexin 32 gene. J Neurol Neurosurg Psychiatr. 1999;66:202–6. Takashima H, Nakagawa M, Umehara F, Hirata K, Suehara M, Mayumi H, Yoshishige K, Matsuyama W, Saito M, Jonosono M, Arimura K, Osame M. Gap junction protein b 1 (GJB1) mutations and central nervous system symptoms in X-linked Charcot-Marie-Tooth disease. Acta Neurol Scand. 2003;107:31–7. Nicholson G, Nash J. Intermediate nerve conduction velocities define X-linked CharcotMarie-Tooth neuropathy families. Neurology 1993;43:2558–64. Tabaraud F, Lagrange E, Sindou P, Vandenberghe A, Levy N, Vallat JM. Demyelinating X-linked Charcot-Marie-Tooth disease: unusual electrophysiological findings. Muscle Nerve. 1999;22:1442–7. Senderek J, Bergmann C, Quasthoff S, Ramaekers VT, Schroder JM. X-linked dominant Charcot-Marie-Tooth disease: nerve biopsies allow morphological evaluation and detection of connexin32 mutations (Arg15Trp, Arg22Gln). Acta Neuropathol. 1998;95:443–9. Senderek J, Hermanns B, Bergmann C, Boroojerdi B, Bajbouj M, Hungs M, Ramaekers VT, Quasthoff S, Karch D, Schroder JM. X-linked dominant CharcotMarie-Tooth neuropathy: clinical, electrophysiological, and morphological phenotype in four families with different connexin32 mutations. J Neurol Sci. 1999;167:90–101. Hahn AF, Ainsworth PJ, Bolton CF, Bilbao JM, Vallat JM. Pathological findings in the X-linked form of Charcot-Marie-Tooth disease: a morphometric and ultrastructural analysis. Acta Neuropathol. 2001;101:129–39. Ainsworth PJ, Bolton CF, Murphy BC, Stuart JA, Hahn AF. Genotype/phenotype correlation in affected individuals of a family with a deletion of the entire coding sequence of the connexin 32 gene. Hum Genet. 1998;103:242–4. Lin C, Numakura C, Ikegami T, Shizuka M, Shoji M, Nicholson G, Hayasaka K. Deletion and nonsense mutations of the connexin 32 gene associated with CharcotMarie-Tooth disease. Tohoku J Exp Med. 1999;188:239–44.

15 Nervous System

353

143. Nakagawa M, Takashima H, Umehara F, Arimura K, Miyashita F, Takenouchi N, Matsuyama W, Osame M Clinical phenotype in X-linked Charcot-Marie-Tooth disease with an entire deletion of the connexin 32 coding sequence. J Neurol Sci. 2001;185:31–7. 144. Ionasescu VV, Searby C, Ionasescu R, Neuhaus IM, Werner R. Mutations of the noncoding region of the connexin32 gene in X-linked dominant Charcot-Marie-Tooth neuropathy. Neurology 1996;47:541–4. 145. Flagiello L, Cirigliano V, Strazzullo M, Cappa V, Ciccodicola A, D’Esposito M, Torrente I, Werner R, Di Iorio G, Rinaldi M, Dallapiccola A, Forabosco A, Ventruto V, D’Urso M. Mutation in the nerve-specific 5’noncoding region of Cx32 gene and absence of specific mRNA in a CMTX1 Italian family. Hum Mutat. 1998;12:361. 146. Ionasescu V, Ionasescu R, Searby C. Correlation between connexin 32 gene mutations and clinical phenotype in X-linked dominant Charcot-Marie-Tooth neuropathy. Am J Med Genet. 1996;63:486–91. 147. Shy ME, Siskind C, Swan ER, Krajewski KM, Doherty T, Fuerst DR, Ainsworth PJ, Lewis RA, Scherer SS, Hahn AF. CMT1X phenotypes represent loss of GJB1 gene function. Neurology 2007;68:849–55. 148. Abrams CK, Oh S, Ri Y, Bargiello TA. Mutations in connexin 32: the molecular and biophysical bases for the X-linked form of Charcot-Marie-Tooth disease. Brain Res Brain Res Rev. 2000;32:203–14. 149. Ressot C, Gomes D, Dautigny A, Pham-Dinh D, Bruzzone R. Connexin32 mutations associated with X-linked Charcot-Marie-Tooth disease show two distinct behaviors: loss of function and altered gating properties. J Neurosci. 1998;18:4063–75. 150. Wang HL, Chang WT, Yeh TH, Wu T, Chen MS, Wu CY. Functional analysis of connexin-32 mutants associated with X-linked dominant Charcot-Marie-Tooth disease. Neurobiol Dis. 2004;15:361–70. 151. Bruzzone R, White TW, Scherer SS, Fischbeck KH, Paul DL. Null mutations of connexin32 in patients with X-linked Charcot-Marie-Tooth disease. Neuron 1994;13:1253–60. 152. Omori Y, Mesnil M, Yamasaki H. Connexin 32 mutations from X-linked CharcotMarie-Tooth disease patients: functional defects and dominant-negative effects. Mol Biol Cell. 1996;7:907–16. 153. Yoshimura T, Satake M, Ohnishi A, Tsutsumi Y, Fujikura Y. Mutations of connexin32 in Charcot-Marie-Tooth disease type X interfere with cell-to-cell communication but not cell proliferation and myelin- specific gene expression. J Neurosci Res. 1998;51:154–61. 154. Deschenes SM, Walcott JL, Wexler TL, Scherer SS, Fischbeck KH. Altered trafficking of mutant connexin32. J Neurosci. 1997;17:9077–84. 155. Yum SW, Kleopa KA, Shumas S, Scherer SS. Diverse trafficking abnormalities of connexin32 mutants causing CMTX. Neurobiol Dis. 2002;11:43–52. 156. Kleopa KA, Yum SW, Scherer SS. Cellular mechanisms of connexin32 mutations associated with CNS manifestations. J Neurosci Res. 2002;68:522–34. 157. VanSlyke JK, Deschenes SM, Musil LS. Intracellular transport, assembly, and degradation of wildtype and disease-linked mutant gap junction proteins. Mol Biol Cell. 2000;11:1933–46. 158. Jeng LJ, Balice-Gordon RJ, Messing A, Fischbeck KH, Scherer SS. The effects of a dominant connexin32 mutant in myelinating Schwann cells. Mol Cell Neurosci. 2006;32:283–98. 159. Abrams CK, Bennett MVL, Verselis VK, Bargiello TA. Voltage opens unopposed gap junction hemichannels formed by a connexin 32 mutant associated with X-linked Charcot-Marie-Tooth disease. Proc Natl Acad Sci USA. 2002;99:3980–4. 160. Liang GS, de Miguel M, Gomez-Hernandez JM, Glass JD, Scherer SS, Mintz M, Barrio LC, Fischbeck KH. Severe neuropathy with leaky connexin32 hemichannels. Ann Neurol. 2005;57:749–54. 161. Loddenkemper T, Grote K, Evers S, Oelerich M, Stogbauer F. Neurological manifestations of the oculodentodigital dysplasia syndrome. J Neurol. 2002;249:584–95.

354

C.K. Abrams and J.E. Rash

162. Ginsberg LE, Jewett T, Grub R, McLean WT. Oculodental digital dysplasia: neuroimaging in a kindred. Neuroradiology 1996;38:84–6. 163. Gutmann DH, Zackai EH, McDonald-McGinn DM, Fischbeck KH, Kamholz J. Oculodentodigital dysplasia syndrome associated with abnormal cerebral white matter. Am J Med Gen. 1991;41:18–20. 164. Paznekas WA, Boyadjiev SA, Shapiro RE, Daniels O, Wollnik B, Keegan CE, Innis JW, Dinulos MB, Christian C, Hannibal MC, Jabs EW. Connexin 43 (GJA1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet. 2003;72:408–18. 165. Flenniken AM, Osborne LR, Anderson N, Ciliberti N, Fleming C, Gittens JE, Gong XQ, Kelsey LB, Lounsbury C, Moreno L, Nieman BJ, Peterson K, Qu D, Roscoe W, Shao Q, Tong D, Veitch GI, Voronina I, Vukobradovic I, Wood GA, Zhu Y, Zirngibl RA, Aubin JE, Bai D, Bruneau BG, Grynpas M, Henderson JE, Henkelman RM, McKerlie C, Sled JG, Stanford WL, Laird DW, Kidder GM, Adamson SL, Rossant J. A Gja1 missense mutation in a mouse model of oculodentodigital dysplasia. Development 2005;132:4375–86. 166. Lai A, Le DN, Paznekas WA, Gifford WD, Jabs EW, Charles AC. Oculodentodigital dysplasia connexin43 mutations result in non-functional connexin hemichannels and gap junctions in C6 glioma cells. J Cell Sci. 2006;119:532–41. 167. McLachlan E, Manias JL, Gong XQ, Lounsbury CS, Shao Q, Bernier SM, Bai D, Laird DW. Functional characterization of oculodentodigital dysplasia-associated Cx43 mutants. Cell Commun Adhes. 2005;12:279–92. 168. Shibayama J, Paznekas W, Seki A, Taffet S, Jabs EW, Delmar M, Musa H. Functional characterization of connexin43 mutations found in patients with oculodentodigital dysplasia. Circ Res. 2005;96:e83–91. 169. Gong XQ, Shao Q, Lounsbury CS, Bai D, Laird DW. Functional characterization of a GJA1 frameshift mutation causing oculodentodigital dysplasia and palmoplantar keratoderma. J Biol Chem. 2006;281:31801–11. 170. Roscoe W, Veitch GI, Gong XQ, Pellegrino E, Bai D, McLachlan E, Shao Q, Kidder GM, Laird DW. Oculodentodigital dysplasia-causing connexin43 mutants are non-functional and exhibit dominant effects on wildtype connexin43. J Biol Chem. 2005;280:11458–66. 171. Reaume AG, De Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science 1995;267:1831–4. 172. Garbern JY. Pelizaeus-Merzbacher disease: genetic and cellular pathogenesis. Cell Mol Life Sci. 2007;64:50–65. 173. Salviati L, Trevisson E, Baldoin MC, Toldo I, Sartori S, Calderone M, Tenconi R, Laverda A. A novel deletion in the GJA12 gene causes Pelizaeus-Merzbacher-like disease. Neurogenetics 2007;8:57–60. 174. Wolf NI, Cundall M, Rutland P, Rosser E, Surtees R, Benton S, Chong WK, Malcolm S, Ebinger F, Bitner-Glindzicz M, Woodward KJ. Frameshift mutation in GJA12 leading to nystagmus, spastic ataxia and CNS dys-/demyelination. Neurogenetics 2007;8:39–44. 175. Bugiani M, Al Shahwan S, Lamantea E, Bizzi A, Bakhsh E, Moroni I, Balestrini MR, Uziel G, Zeviani M. GJA12 mutations in children with recessive hypomyelinating leukoencephalopathy. Neurology 2006;67:273–9. 176. Orthmann-Murphy JL, Enriquez AD, Abrams CK, Scherer SS. Loss-of-function GJA12/Connexin47 mutations cause Pelizaeus-Merzbacher-like disease. Mol Cell Neurosci. 2007;34:629–41. 177. Traub RD, Draguhn A, Whittington MA, Baldeweg T, Bibbig A, Buhl EH, Schmitz D. Axonal gap junctions between principal neurons: a novel source of network oscillations, and perhaps epileptogenesis. Rev Neurosci. 2002;13:1–30.

15 Nervous System

355

178. Rouach N, Avignone E, Meme W, Koulakoff A, Venance L, Blomstrand F, Giaume C. Gap junctions and connexin expression in the normal and pathological central nervous system. Biol Cell. 2002;94:457–75. 179. Nemani VM, Binder DK. Emerging role of gap junctions in epilepsy. Histol Histopathol. 2005;20:253–9. 180. Maier N, Guldenagel M, Sohl ¨ G, Siegmund H, Willecke K, Draguhn A. Reduction of high frequency network oscillations (ripples) and pathological network discharges in hippocampal slices from connexin 36-deficient mice. J Physiol. 2002;541:521–8. 181. De Zeeuw CI, Chorev E, Devor A, Manor Y, Van Der Giessen RS, De Jeu MT, Hoogenraad CC, Bijman J, Ruigrok TJ, French P, Jaarsma D, Kistler WM, Meier C, Petrasch-Parwez E, Dermietzel R, Sohl G, Gueldenagel M, Willecke K, Yarom Y. ¨ Deformation of network connectivity in the inferior olive of connexin 36-deficient mice is compensated by morphological and electrophysiological changes at the single neuron level. J Neurosci. 2003;23:4700–11. 182. Song J, Tanouye MA. Seizure suppression by shakB2, a gap junction mutation in Drosophila. J Neurophysiol. 2006;95:627–35. 183. Yang L, Ling DS. Carbenoxolone modifies spontaneous inhibitory and excitatory synaptic transmission in rat somatosensory cortex. Neurosci Lett. 2007;416:221–6. 184. Gigout S, Louvel J, Pumain R. Effects in vitro and in vivo of a gap junction blocker on epileptiform activities in a genetic model of absence epilepsy. Epilepsy Res. 2006;69:15–29. 185. Samoilova M, Li J, Pelletier MR, Wentlandt K, Adamchik Y, Naus CC, Carlen PL. Epileptiform activity in hippocampal slice cultures exposed chronically to bicuculline: increased gap junctional function and expression. J Neurochem. 2003;86:687–99. 186. Ross FM, Gwyn P, Spanswick D, Davies SN. Carbenoxolone depresses spontaneous epileptiform activity in the CA1 region of rat hippocampal slices. Neuroscience 2000;100:789–96. 187. Bostanci MO, Bagirici F. Anticonvulsive effects of carbenoxolone on penicillin-induced epileptiform activity: an in vivo study. Neuropharmacology 2007;52:362–7. 188. Bostanci MO, Bagirici F. The effects of octanol on penicillin induced epileptiform activity in rats: an in vivo study. Epilepsy Res. 2006;71:188–94. 189. Gajda Z, Szupera Z, Blazso G, Szente M. Quinine, a blocker of neuronal Cx36 channels, suppresses seizure activity in rat neocortex in vivo. Epilepsia 2005;46:1581–91. 190. Szente M, Gajda Z, Said Ali K, Hermesz E. Involvement of electrical coupling in the in vivo ictal epileptiform activity induced by 4-aminopyridine in the neocortex. Neuroscience 2002;115:1067–78. 191. Proulx E, Leshchenko Y, Kokarovtseva L, Khokhotva V, El-Beheiry M, Snead OC, 3rd, Perez Velazquez JL. Functional contribution of specific brain areas to absence seizures: role of thalamic gap-junctional coupling. Eur J Neurosci. 2006;23:489–96. 192. Srinivas M, Hopperstad MG, Spray DC. Quinine blocks specific gap junction channel subtypes. Proc Natl Acad Sci USA. 2001;98:10942–7. 193. Bikson M, Id Bihi R, Vreugdenhil M, Kohling R, Fox JE, Jefferys JG. Quinine suppresses extracellular potassium transients and ictal epileptiform activity without decreasing neuronal excitability in vitro. Neuroscience 2002;115:251–61. 194. Deuschl G, Raethjen J, Lindemann M, Krack P. The pathophysiology of tremor. Muscle Nerve. 2001;24:716–35. 195. Llinas R, Volkind RA. The olivo-cerebellar system: functional properties as revealed by harmaline-induced tremor. Exp Brain Res. 1973;18:69–87. 196. de Montigny C, Lamarre Y. Rhythmic activity induced by harmaline in the olivocerebello-bulbar system of the cat. Brain Res. 1973;53:81–95. 197. Llinas R, Baker R, Sotelo C. Electrotonic coupling between neurons in cat inferior olive. J Neurophysiol. 1974;37:560–71.

356

C.K. Abrams and J.E. Rash

198. Llinas R, Yarom Y. Oscillatory properties of guinea-pig inferior olivary neurones and their pharmacological modulation: an in vitro study. J Physiol. 1986;376:163–82. 199. Miwa H, Kubo T, Suzuki A, Kihira T, Kondo T. A species-specific difference in the effects of harmaline on the rodent olivocerebellar system. Brain Res. 2006;1068:94–101. 200. Louis ED, Vonsattel JP, Honig LS, Lawton A, Moskowitz C, Ford B, Frucht S. Essential tremor associated with pathologic changes in the cerebellum. Arch Neurol. 2006;63:1189–93. 201. Louis ED, Vonsattel JP, Honig LS, Ross GW, Lyons KE, Pahwa R. Neuropathologic findings in essential tremor. Neurology 2006;66:1756–9. 202. Long MA, Deans MR, Paul DL, Connors BW. Rhythmicity without synchrony in the electrically uncoupled inferior olive. J Neurosci. 2002;22:10898–905. 203. Placantonakis DG, Bukovsky AA, Zeng XH, Kiem HP, Welsh JP. Fundamental role of inferior olive connexin 36 in muscle coherence during tremor. Proc Natl Acad Sci USA. 2004;101:7164–9. 204. Martin FC, Handforth A. Carbenoxolone and mefloquine suppress tremor in the harmaline mouse model of essential tremor. Mov Disord. 2006;21:1641–9. 205. Perez Velazquez JL, Frantseva MV, Naus CC. Gap junctions and neuronal injury: protectants or executioners? Neuroscientist 2003;9:5–9. 206. Farahani R, Pina-Benabou MH, Kyrozis A, Siddiq A, Barradas PC, Chiu FC, Cavalcante LA, Lai JC, Stanton PK, Rozental R. Alterations in metabolism and gap junction expression may determine the role of astrocytes as ‘good samaritans’ or executioners. Glia. 2005;50:351–61. 207. Lin JH, Yang J, Liu S, Takano T, Wang X, Gao Q, Willecke K, Nedergaard M. Connexin mediates gap junction-independent resistance to cellular injury. J Neurosci. 2003;23:430–41. 208. Nedergaard M, Astrup J. Infarct rim: effect of hyperglycemia on direct current potential and 14C-2-deoxyglucose phosphorylation. J Cereb Blood Flow Metab. 1986;6:607–15. 209. Nakase T, Sohl ¨ G, Theis M, Willecke K, Naus CC. Increased apoptosis and inflammation after focal brain ischemia in mice lacking connexin43 in astrocytes. Am J Pathol. 2004;164:2067–75. 210. Rawanduzy A, Hansen A, Hansen TW, Nedergaard M. Effective reduction of infarct volume by gap junction blockade in a rodent model of stroke. J Neurosurg. 1997;87:916–20. 211. John SA, Kondo R, Wang SY, Goldhaber JI, Weiss JN. Connexin-43 hemichannels opened by metabolic inhibition. J Biol Chem. 1999;274:236–40. 212. Thompson RJ, Zhou N, MacVicar BA. Ischemia opens neuronal gap junction hemichannels. Science 2006;312:924–7. 213. Ye ZC, Wyeth MS, Baltan-Tekkok S, Ransom BR. Functional hemichannels in astrocytes: a novel mechanism of glutamate release. J Neurosci. 2003;23:3588–96. 214. Perez Velazquez JL, Kokarovtseva L, Sarbaziha R, Jeyapalan Z, Leshchenko Y. Role of gap junctional coupling in astrocytic networks in the determination of global ischaemia-induced oxidative stress and hippocampal damage. Eur J Neurosci. 2006; 23:1–10. 215. Nakase T, Fushiki S, Naus CC. Astrocytic gap junctions composed of connexin 43 reduce apoptotic neuronal damage in cerebral ischemia. Stroke 2003;34:1987–93. 216. Siushansian R, Bechberger JF, Cechetto DF, Hachinski VC, Naus CC. Connexin43 null mutation increases infarct size after stroke. J Comp Neurol. 2001;440:387–94. 217. Oguro K, Jover T, Tanaka H, Lin Y, Kojima T, Oguro N, Grooms SY, Bennett MV, Zukin RS. Global ischemia-induced increases in the gap junctional proteins connexin 32 (Cx32) and Cx36 in hippocampus and enhanced vulnerability of Cx32 knock-out mice. J Neurosci. 2001;21:7534–42.

15 Nervous System

357

218. Rami A, Volkmann T, Winckler J. Effective reduction of neuronal death by inhibiting gap junctional intercellular communication in a rodent model of global transient cerebral ischemia. Exp Neurol. 2001;170:297–304. 219. Warner DS, Ludwig PS, Pearlstein R, Brinkhous AD. Halothane reduces focal ischemic injury in the rat when brain temperature is controlled. Anesthesiology 1995;82:1237–45. 220. Frantseva MV, Kokarovtseva L, Perez Velazquez JL. Ischemia-induced brain damage depends on specific gap-junctional coupling. J Cereb Blood Flow Metab. 2002;22:453–62. 221. Blanc EM, Bruce-Keller AJ, Mattson MP. Astrocytic gap junctional communication decreases neuronal vulnerability to oxidative stress-induced disruption of Ca2+ homeostasis and cell death. J Neurochem. 1998;70:958–70. 222. Lin JH, Weigel H, Cotrina ML, Liu S, Bueno E, Hansen AJ, Hansen TW, Goldman S, Nedergaard M. Gap-junction-mediated propagation and amplification of cell injury. Nat Neurosci. 1998;1:494–500. 223. Barrio LC, Suchyna T, Bargiello T, Xu LX, Roginski RS, Bennett MV, Nicholson BJ. Gap junctions formed by connexins 26 and 32 alone and in combination are differently affected by applied voltage. Proc Natl Acad Sci USA. 1991;88:8410–4. 224. Dahl E, Manthey D, Chen Y, Schwarz HJ, Chang YS, Lalley PA, Nicholson BJ, Willecke K. Molecular cloning and functional expression of mouse connexin-30, a gap junction gene highly expressed in adult brain and skin. J Biol Chem. 1996;271:17903–10. 225. White TW, Paul DL, Goodenough DA, Bruzzone R. Functional analysis of selective interactions among rodent connexins. Mol Biol Cell. 1995;6:459–70. 226. Bukauskas FF, Elfgang C, Willecke K, Weingart R. Heterotypic gap junction channels (connexin26-connexin32) violate the paradigm of unitary conductance. Pflu¨gers Arch. 1995;429:870–2. 227. Werner R, Levine E, Rabadan-Diehl C, Dahl G. Formation of hybrid cell-cell channels. Proc Natl Acad Sci USA. 1989;86:5380–4. 228. Bruzzone R, White TW, Paul DL. Expression of chimeric connexins reveals new properties of the formation and gating behavior of gap junction channels. J Cell Sci. 1994;107:955–67. 229. Nagy JI, Patel D, Ochalski PA, Stelmack GL. Connexin30 in rodent, cat and human brain: selective expression in gray matter astrocytes, colocalization with connexin43 at gap junctions and late developmental appearance. Neuroscience 1999;88:447–68. 230. Li X, Ionescu AV, Lynn BD, Lu S, Kamasawa N, Morita M, Davidson KG, Yasumura T, Rash JE, Nagy JI. Connexin47, connexin29 and connexin32 coexpression in oligodendrocytes and Cx47 association with zonula occludens-1 (ZO-1) in mouse brain. Neuroscience 2004;126:611–30.

Chapter 16

Connexins in the Respiratory Epithelium Bernard Foglia, Isabelle Scerri, Tecla Dudez and Marc Chanson

Abstract The respiratory epithelium is positioned at the interface between the body and the environment. This highly differentiated epithelium plays a crucial role in maintaining the sterility of pulmonary tissues by orchestrating mechanical, innate, and acquired host defense systems. The functional integrity of the airway epithelium depends on expression and assembly of specific proteins into specialized junctional structures. The profile of connexin expression changes dramatically during organogenesis, and shortly after birth. Surprisingly, there is little information on the role of gap junctions and their constituent proteins, connexins, in the adult respiratory epithelium. Although it has been assumed that gap junctions are rare in the air-conducting and respiratory epithelia, recent data point to the roles of connexin channels in lung function and pathology. Keywords Airway epithelium Alveolar epithelium Calcium signaling Differentiation Inflammation Acute lung injury Human airway cell model Cx26 Cx30 Cx31 Cx31.1 Cx32 Cx37 Cx40 Cx43 Cx46

16.1 Introduction This chapter summarizes current knowledge on the profiles of the connexins expressed during the development of lung and airways, and discusses the roles fulfilled by gap junctions in the pathophysiology of airways and lung, especially in the context of redifferentiation after injury and propagation of proinflammatory signals.

M. Chanson (*) Laboratory of Clinical Investigation III, HUG. PO Box 14, 24 Micheli-du-Crest, 1211 Geneva 14, Switzerland e-mail: [email protected]

A. Harris, D. Locke (eds), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_16, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

359

360

B. Foglia et al.

16.2 Respiratory Epithelium: The Interface Between Innate and Acquired Immunity The respiratory tract epithelium is positioned at the interface between the body and the environment. It not only conditions incoming air with moisture but also orchestrates the pulmonary defense system. Indeed, the airway epithelium is exposed to a spectrum of inhaled toxins and chemical and mechanical stresses. The sterility of pulmonary tissues, and the host, is maintained by the concerted effects of mechanical, innate, and acquired host defense systems that recognize, localize, kill, and remove pathogens. Local injury induces a complex response that stimulates healing of injured tissues, cellular regeneration, and phagocytosis. The respiratory epithelium plays a crucial role in these processes by secreting cytokines and chemokines to recruit specialized phagocytes and immune cells. Migration of these cells into sites of infection and their interaction with injured respiratory epithelial cells orchestrate inflammation and tissue repair. In the respiratory tract, the innate host defense system is complemented by several other mechanisms, including the production of antimicrobial polypeptides, surfactants, the cough reflex, and mucociliary clearance [1,2,3,4,5]. Failure in these mechanisms allows local or systemic infection and destruction of lung tissue, which are frequently the basis of morbidity and mortality in a variety of disorders. Acute lung injury and acute respiratory distress syndrome are life-threatening clinical problems, particularly in sepsis, pneumonia and various lung diseases, including cystic fibrosis. The vital roles assumed by the respiratory epithelium require maintenance of a highly specialized tissue architecture that can be restored following injury. The functional integrity of the respiratory epithelium depends on expression and assembly of specific proteins into specialized junctional structures. These structures, in turn, define tissue compartments and facilitate interactions between the extracellular matrix (ECM) and epithelial cells as well as cell–cell communication. Gap junctions play crucial roles in these interactions by enabling cells to share signaling factors, a process referred to as gap junctional intercellular communication (GJIC). The diversity and cell-specific expression of connexins are thought to provide unique GJIC properties essential for tissue function [6]. Thus, understanding changes in connexin expression profiles during differentiation/repair, and characterization of the triggers for such differential expression are likely to help us understand how respiratory cells coordinate their functions and respond to injury.

16.3 The Respiratory Epithelium The mature respiratory system consists of two main portions with distinct functions: the air-conducting portion and the respiratory portion. The airconducting portion, which extends from the nasal cavities to the bronchioles,

16 Respiratory Epithelium

361

provides a passage for inhaled and exhaled air in and out of the respiratory portion. The human conducting airways are lined by a tall pseudostratified epithelium that includes several cell types, among which basal, ciliated, and secretory cells are the most abundant. This structure is hereafter referred to as the airway epithelium. Additional structures of the airway epithelium such as the submucosal seromucus glands are found mostly in the air-conducting portion. The respiratory portion, which is the site for gas exchange between blood and air, is composed of the respiratory bronchioles, alveolar ducts, and alveoli. The latter structures, which is referred to here as part of the alveolar epithelium, are lined by a thin epithelium composed of type I (ATI) and II (ATII) pneumocytes. Whereas ATI cells mediate gas exchange, ATII cells produce surfactants and are progenitors of ATI pneumocytes [7]. The organization of the human fetal respiratory epithelium is rudimentary during the first and second trimester of life. It is characterized by the presence of numerous cells undergoing division, and a paucity of ciliated cells. By the fifth month of development, the fetal lung is viable with the presence of small collapsed alveoli. Early electron microscopy studies detected the presence of gap junctions in the developing lung, which decreased in amount and size as the epithelial cells differentiate [8,9,10,11]. However, gap junctions could still be resolved between ATI and ATII pneumocytes at this stage by freeze-fracture [12]. Studying the development of ferret and human airways, Carson et al. [11,13] found transient expression of Cx26 and Cx32. While Cx26 and Cx32 are prominent in fetal airway epithelial cells, expression of these connexins declines with ciliation of the superficial airway epithelium and rapidly disappears after birth. Similar observations were made by Traub et al. [14], who reported more abundant Cx37 expression in embryonic mouse lung than in adult tissue. In adult animals, weak Cx37 expression was observed in bronchioles and alveolar epithelial cells, although expression of this connexin was not confirmed in adult rat alveolar epithelial cells [15]. In contrast to the superficial respiratory epithelium, gap junctions were observed between serous cells and between mucous cells in the adult rat submucosal glands [16]. In the mouse, Cx26 and Cx32 were detected between the epithelial cells, whereas Cx43 was observed between myoepithelial cells surrounding the glands [17].

16.4 Adult Airway Epithelium In apparent contradiction with electron microscope studies of adult lung tissue, there is a large body of biochemical and functional evidence of the presence of gap junctions and GJIC in cultured adult airway cells, and of their involvement in coordination of ciliary beating. Thus, extensive GJIC has been detected in explant outgrowths or cultures of tracheal epithelial cells [18,19,20,21]. It has been proposed that the intercellular propagation of Ca2þ waves is mediated by diffusion of inositol triphosphate (IP3) through gap junctions, resulting in the

362

B. Foglia et al.

release of Ca2þ from internal stores [18,19]. Changes in cytosolic Ca2þ concentration may in turn regulate ciliary beating [22]. Depending on the method of culture and the species studied, Cx32 and Cx43 have been detected [20,23,24]. The connexin expression profile has also been shown to be dependent on the composition of the ECM. Thus, growing rabbit tracheal epithelial cells on collagen and laminin-3 induced a shift from perinuclear to plasma membrane localization of Cx26 and Cx43 [22]. Interestingly, Cx46 showed an opposite expression pattern, being localized at the plasma membrane in airway epithelial cells grown on collagen only [21]. In all of these studies, however, airway cells were maintained under submerged culture conditions, which may not recapitulate the properties of a wellpolarized respiratory epithelium that is exposed to air. Cx32, as mentioned above, is detected only in fetal respiratory epithelial cells, whereas CX43 is prominent in poorly polarized human airway epithelial cell cultures and cell lines [25,26,27]. In this regard, Ehrhardt et al. [26] showed that CX43 expression in the 16HBE14o– human airway epithelial cell line declined with partial differentiation of the cells when cultured at the air–liquid interface. These observations suggest that CX43 is primarily expressed in undifferentiated airway epithelial cells, with perhaps the exception of the nasal mucosa [28], and raise the question of which connexin isoforms are expressed, if any, in a wellpolarized adult airway epithelium. Indeed, little or no GJIC was detected in well-polarized cultures of human airway epithelium exposed to air and in small clusters of mouse airway cells shortly after enzymatic dissociation [29,30]. In addition, the hypothesis of Ca2þ wave propagation via IP3 permeation through gap junctions was not supported in these air–liquid interface airway cultures. Homolya et al. [29] proposed instead a model based on the release of extracellular nucleotides and subsequent paracrine activation of purinergic receptors at the surface of airway epithelial cells. Recently, these questions have been addressed in mouse upper airways and in a human airway culture model. In adult mouse airway epithelium, Cx32, Cx43, Cx37, and Cx40 were not detected. Cx26, in contrast, was observed, although its detection was discontinuous [17]. A new finding reported by this study is that Cx30 and Cx31 are expressed in the adult mouse airway epithelium but exhibit different localization. Cx31 was clearly restricted to the basal side of the epithelium, whereas Cx30 showed a broader distribution (Fig. 16.1). In the human airway culture model system, human airway epithelial cells exposed to air differentiate in vitro into a pseudostratified mucociliated epithelium that can preserve its integrity for up to six to nine months [31]. Interestingly, the human airway epithelium model exhibits similar expression profiles for connexins as found in the intact adult mouse upper airway, validating its use, and in contrast to findings from submerged culture systems. The spatially distinct cellular expression patterns of Cx30 and Cx31 in adult epithelia imply that gap junctions produce distinct compartments of intercellular communication. This was confirmed by intracellular microinjection of gap junction–permeant tracers of different sizes, such as Lucifer yellow (LY) and

16 Respiratory Epithelium

363

Fig. 16.1. Connexin profile in the airway epithelium. A cryostat section of the mouse upper airways counterstained with Evans blue illustrates the surface airway epithelium and submucosal glands (top panel). Myoepithelial cells surrounding the glands are also indicated. A schematic view of the airway epithelium with tentative indication of the cellular expression for Cx26, Cx30, Cx31, Cx32, and Cx43 is shown (bottom panel). Cx26 is indicated in parentheses where its expression is discontinuous, as observed in the adult surface epithelium. Scale bar 10 mm. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com)

neurobiotin (NB), in basal, mucous, or ciliated cells of polarized human airway epithelium cultures. These experiments revealed that NB, the smaller of the two tracers, diffused mostly between neighboring ciliated cells but that LY did not. This observation is in keeping with a previous study reporting lack of LY diffusion in well-polarized cultures of human airway epithelium [29]. When injected into basal cells, LY was weakly transferred to other basal cells only, whereas NB diffused well from basal to ciliated cells. It is known that Cx30 is not permeable to LY [32] (see Chapter 7). These results suggest, therefore, that few basal cells are in contact with each other, which is to be expected for a pseudostratified architecture. The prominent GJIC observed between basal and ciliated cells, and between ciliated cells, could be mediated by Cx30, whereas Cx31 preferentially connects basal cells (Fig. 16.1). Finally, LY and NB never diffused to or from mucous cells, suggesting that this cell type is devoid of gap junctions or expresses connexin channels impermeable to these tracers. Thus, these results support the existence of signal-selective gap junctions in the adult polarized airway epithelium. They do not exclude the possibility that other

364

B. Foglia et al.

connexins, such as CX30.3 and CX31.1, for which messenger RNAs (mRNAs) are expressed in the human culture model and in airway epithelial cells obtained by nasal brushing, may contribute to airway epithelium function/GJIC pathways [30].

16.5 The Alveolar Epithelium Identification of the connexins expressed in the adult alveolar epithelium has been difficult due to the complex architecture of the tissue and because most connexin antibodies are ineffective on aldehyde-fixed tissues. Thus, immunodetection of connexins in this tissue is typically performed on methanol/acetone-fixed frozen sections, a procedure that poorly preserves tissue architecture. The expression and role of connexins in the alveolar epithelium has been recently reviewed [7,33]. At least five connexins have been detected by immunocytochemistry in adult rat alveoli. Cx26, Cx32, and Cx40 showed moderate staining, whereas Cx43 and Cx46 showed stronger immunoreactivity, but this could not be specifically attributed to the junctional membrane of a particular cell type [15,34]. With the exception of Cx32, which seems to be associated with ATII cells, from their location in the corners of alveoli, the cellular (ATI, ATII, endothelial cells, or macrophages) pattern of expression for these connexin within the alveolar septa is not known (Fig. 16.2).

Fig. 16.2. Connexin profile in the alveolar epithelium. The complex interaction between ATI and ATII pneumocytes, endothelial cells (endo), macrophages (macro), and blood leukocytes (leuko) is illustrated on a semi-thin-section of a perfused rat lung. Also indicated is the tentative cellular expression of Cx26, Cx32, Cx40, Cx43, and Cx46 according to in vivo and in vitro reports. Some connexins are indicated in parentheses when their specific cellular localization needs further in vivo confirmation. Scale bar 20 mm. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com)

16 Respiratory Epithelium

365

Most of our knowledge of connexin expression and function in alveolar epithelial cells comes from isolated ATII cells cultured under submerged conditions. Because primary ATII cells exhibit ATI-like phenotype within days of culture [35]), connexins were mostly compared between young (<2 days) and old (>6 days) cultures. These cultures roughly showed an inverse pattern of abundance with days in culture characterized by the decline of Cx26 and Cx32 and the increase of Cx43 and Cx46 [36,37,38]. These changes in connexin expression are likely to be reciprocally regulated with those of ECM constituents, as shown by the effects of culturing ATII cells on fibronectin-containing and/or laminin-5–containing matrices [37,39,40,41]. These culture conditions also affected GJIC between ATII-like and ATI-like pneumocytes. ATII cells are seldom adjacent to each other in adult lung tissue, so GJIC between these cells is anticipated to be rare. There are about twice as many ATII as ATI cells, suggesting that ATI and ATII cells are likely to be in contact with each other, reflecting the possibility for signaling networks within the alveolar environment. This notion was confirmed in vitro by Abraham et al. [34], who reported GJIC consistent with Cx43 junctions in heterocellular cultures of ATIlike and ATII-like cells. Interestingly, the transmission of Ca2þ waves between alveolar cells seems to differ according to the proportion of ATI-like and ATII-like cells present in the cultures and on the mode of cell stimulation [38,41]. Using a well-controlled heterocellular culture model of rat ATI and ATII cells, Isakson et al. [15] reported distinct patterns of Ca2þ communication depending on the cell origin of the Ca2þ signal. Ca2þ signals elicited in ATII cells are propagated to neighboring ATI cells via release of extracellular nucleotides, and subsequent paracrine activation of purinergic receptors located on ATI cells. In contrast, stimulation of ATI cells induced Ca2þ waves between ATI and ATII cells that required functional GJIC. Ca2þ communication between ATII cells was also dependent on GJIC but not on the extracellular nucleotide–dependent pathway. Altogether, these data are suggestive of distinct and well-defined patterns of molecular transfer between and among ATI and ATII pneumocytes [33], including both gap junction–dependent and gap junction–independent mechanisms. Whether the paracrine signaling pathway involves release of nucleotides via unpaired connexin hemichannels is a matter of debate (see Chapters 7 and 12).

16.6 Roles for Connexins in Lung and Airways A change in intracellular Ca2þ concentration is a common trigger for cellular processes in a variety of cells, including respiratory epithelial cells. Ca2þ signaling in airway and alveolar epithelial cells is involved in diverse functions such as ciliary beating, transepithelial ion transport, cytokine and chemokine secretion, surfactant secretion, cell differentiation, ECM protein synthesis, cell growth,

366

B. Foglia et al.

and cell death [7,33]. During development, repair after injury or in the course of various pathologies, there is extensive remodeling of the ECM that may change the profile of connexin expression. In view of the importance of appropriate connexin expression for selective Ca2þ signaling, and likely for intercellular diffusion of other vital metabolites, it is thought that GJIC contributes to most of these airway and lung functions. A role for Cx43-dependent interalveolar communication has been recently demonstrated in the intact, blood perfused rat lung. In this study, Ichimura et al. [42] applied realtime digital imaging techniques and photoexcited Ca2þ uncaging to monitor intracellular Ca2þ changes and ATII cell secretion. Interestingly, they report that local uncaging of Ca2þ resulted in secretory response in both the excited alveolus as well as neighboring alveoli. The latter response was blocked by Cx43 extracellular loop mimetic peptides, indicating that stimulation of distant ATII cells may occur through heterocellular connexindependent pathways. In these experiments, however, the contribution of other cell types cannot be excluded. In the intact lung, photoexcitation can also uncage Ca2þ loaded in endothelial cells, allowing the Ca2þ signal to propagate through endothelial gap junctions. Activation of distant endothelial cells may in turn affect the secretion of ATII cells. This scenario also recently received support. Using a similar technical approach, Parthasarathi et al. [43] showed that the propagation of Ca2þ waves between alveolar endothelial cells was abolished in mice with endothelialspecific deletion of the Cx43 gene as compared to wild-type animals. In addition, Ca2þ signaling was associated with changes in adhesion receptors for leukocytes and in microvascular permeability. In vitro studies on primary cultures of porcine microvascular endothelial cells and rat endothelial cell lines, suggest that Cx43 and Cx40 are involved in maintenance of the barrier function [44], and that GJIC may serve as a pathway for the spread of proinflammatory signals in the lung capillary bed. There are few data on the relationship between connexins and lung inflammation. On the one hand, expression of Cx43 protein increases in rat lungs that are injected with bacterial endotoxin [45]. On the other hand, endotoxin and proinflammatory mediators rapidly reduce GJIC in CX43-expressing human airway cell lines [25,46]. Although in apparent contradiction, these data suggest that increased Cx43 protein expression is a consequence of inflammation while its channel activity is decreased. Interestingly, CX43 channel activity is not affected by endotoxin and tumor necrosis factor-a (TNF-a) in airway cell lines derived from patients with cystic fibrosis [25,27,46], a human disease characterized by chronic lung infection and inflammation. Defects in the downregulation of CX43 channel activity, which would thus result in enhanced GJIC, may contribute to the perpetuation of proinflammatory signals in cystic fibrosis. Although confirmation in vivo is awaited, these reports suggest that Cx43 may play pivotal roles in mediating lung inflammation. The roles of CX26, CX30, and CX31 in the polarized human respiratory epithelium are informative as to their roles in other epithelial tissues, and vice

16 Respiratory Epithelium

367

versa. For example, Cx30 may contribute to the regulation of ciliary beating by an unusual mechanism that may help to reconcile previously conflicting reports [22,29]. Specific mutations in CX30, which are responsible for the skin disease hidrotic ectodermal dysplasia (HED, also called Clouston syndrome), cause abnormal CX30 hemichannel activity [47]. Interestingly, these mutations induced a gain of function of CX30 hemichannels characterized by increased leakage of adenosine triphosphate (ATP) to the extracellular space. Thus, release of ATP through Cx30 hemichannels may act as a paracrine messenger that not only can modulate ciliary beating but also alter epithelial factors involved in the control of airway cell proliferation and differentiation. Mutations in genes encoding CX26, CX30, or CX31 are responsible for epidermal disorders or hereditary hearing impairment. In skin, Cx26 is expressed in proliferative epidermis during early embryonic development or reepithelialization but inhibited at terminal differentiation [48,49]. Recent evidence indicates that Cx26 plays a role in maintenance of a hyperproliferative state in wounded skin [50]. Thus, by analogy with other epidermal tissues, Cx26 may play a key role in airway differentiation and the process of barrier acquisition during development or repair after injury. The functional role of connexins in acute lung injury and acute respiratory distress syndrome remains to be fully explored. Animal models with genes deleted for most of the connexins that are expressed in airways and lung have been generated, but the single knockout (KO) phenotype that escapes early lethality (Cx32KO) does not exhibit obvious pulmonary alterations. However, these mice may reveal specific airway phenotypes under stress conditions, as it has been recently reported that Cx32KO mice have increased incidence of bronchioalveolar lung tumors when exposed to carcinogens [51]. These studies, which may prove to be difficult due to the redundancy of connexin expression in this tissue, may require more complex strategies, including cell-targeted or double gene deletion or the use of a controlled model of lung inflammation.

16.7 Conclusion Shortly after birth, the lung undergoes a dramatic decrease in connexin expression in both the air-conducting and respiratory portions. It is now clear that the mature lung expresses multiple connexins and exhibits a complex pattern of GJIC pathways. The identity of the connexins expressed in mature alveolar cells is poorly known, but recent work suggests the involvement of Cx43-dependent GJIC in the propagation of inflammatory signals within the respiratory portion of the lung. The expression of connexins in the air-conducting portion has recently been described in detail, namely Cx26 and Cx31 between basal cells and Cx30 between ciliated cells. It is likely that these connexins play pivotal roles in differentiation and repair after injury of the airway epithelium. The identification of these connexins allows investigation of crucial questions such

368

B. Foglia et al.

as whether they contribute to the modulation of the inflammatory and immune response, an aspect of particular importance with respect to life-threatening respiratory disorders like acute respiratory distress syndrome and cystic fibrosis. Acknowledgments The authors thank Dr. Brenda Kwak for critical reading of the manuscript, and Suzanne Duperret for secretarial assistance. This work was supported by grants from the Swiss National Science Foundation, the French Association Vaincre la Mucoviscidose, and the Swiss Cystic Fibrosis Foundation.

References 1. Strieter RM, Belperio JA, Keane MP. Host innate defenses in the lung: the role of cytokines. Curr Opin Infect Dis. 2003;16:193. 2. Tiddens H, Silverman M, Bush A. The role of inflammation in airway disease. Remodelling. Am J Respir Crit Care Med. 2000;162:S7. 3. Travis SM, Singh PK, Welsh MJ. Antimicrobial peptides and proteins in the innate defense of the airway surface. Curr Opin Immunol. 2001;13:89–95. 4. Knowles MR, Boucher RC. Mucus clearance as a primary innate defense mechanism for mammalian airways. J Clin Invest. 2002;109:571–7. 5. Whitsett JA. Intrinsic and innate defenses in the lung: intersection of pathways regulating lung: morphogenesis, host defense, and repair. J Clin Invest. 2002;109:565–9. 6. Sa´ez JC, Berthoud VM, Bran˜es MC, Martı´ nez AD, Beyer EC. Plasma membrane channels formed by connexins: their regulation and functions. Physiol Rev. 1003;83: 1359–400. 7. Koval M. Sharing signals: connecting lung epithelial cells with gap junction channels. Am J Physiol Lung Cell Mol Physiol. 2002;283:L875–93. 8. Inoue S, Hogg JC. Intercellular junctions of the tracheal epithelium in guinea pigs. Lab Invest. 1974;31:68–74. 9. Inoue S, Hogg JC. Freeze-etch study of the tracheal epithelium of normal guinea pigs with particular reference to intercellular junctions. J Ultrastruct Res. 1977;61:89–99. 10. Schneeberger EE, Walters DV, Olver RE. Development of intercellular junctions in the pulmonary epithelium of the foetal lamb. J Cell Sci. 1978;32:307–24. 11. Carson JL, Reed W, Moats-Staats BM, Brighton LE, Gambling TM, Hu SC, Collier AM. Connexin 26 expression in human and ferret airways and lung during development. Am J Respir Cell Mol Biol. 1998;18:111–9. 12. Bartels H. The air-blood barrier in the human lung. A freeze-fracture study. Cell Tissue Res. 1979;198:269–85. 13. Carson JL, Willumsen NJ, Gambling TM, Hu SC, Collier AM. Dynamics of intercellular communication and differentiation in a rapidly developing mammalian airway epithelium. Am J Respir Cell Mol Biol. 1989;1:385–90. 14. Traub O, Hertlein B, Kasper M, Eckert R, Krisciukaitis A, Hu¨lser D, Willecke K. Characterization of the gap junction protein connexin37 in murine endothelium, respiratory epithelium, and after transfection in human HeLa cells. Eur J Cell Biol. 1998;77:313–22. 15. Isakson BE, Seedorf GJ, Lubman RL, Evans WH, Boitano S. Cell-cell communication in heterocellular cultures of alveolar epithelial cells. Am J Respir Cell Mol Biol. 2003;29:552–61. 16. Schneeberger EE, McCormack JM. Intercellular junctions in upper airway submucosal glands of the rat: a tracer and freeze-fracture study. Anat Rec. 1984;210:421–33.

16 Respiratory Epithelium

369

17. Wiszniewski L, Sanz J, Scerri I, Gasparotto E, Dudez T, Lacroix JS, Suter S, Gallati S, Chanson M. Functional expression of connexin30 and connexin31 in the polarized human airway epithelium. Differentiation. 2007;75:382–92. 18. Sanderson MJ, Chow I, Dirksen ER. Intercellular communication between ciliated cells in culture. Am J Physiol. 1988;254:C63–C74. 19. Boitano S, Dirksen ER, Sanderson MJ. Intercellular propagation of calcium waves mediated by inositol trisphosphate. Science. 1992;258:292–5. 20. Brezillon S, Zahm JM, Pierrot D, Gaillard D, Hinnrasky J, Millart H, Klossek JM, Tummler B, Puchelle E. ATP depletion induces a loss of respiratory epithelium functional integrity and down-regulates CFTR (cystic fibrosis transmembrane conductance regulator) expression. J Biol Chem. 1997;272:27830–8. 21. Isakson BE, Olsen CE, Boitano SB. Laminin-332 alters connexin profile, dye coupling and intercellular Ca2þ waves in ciliated tracheal epithelial cells. Resp Res. 2006;7:105. 22. Evans JH, Sanderson MJ. Intracellular calcium oscillations regulate ciliary beat frequency of airway epithelial cells. Cell Calcium. 1999;26:103–10. 23. Boitano S, Dirksen ER, Evans WH. Sequence specific antibodies to connexins block intercellular calcium signaling through gap junctions. Cell Calcium. 1998;23:1–9. 24. Boitano S, Evans, WH. Connexin mimetic peptides reversibly inhibit Ca2þ signaling through gap junctions in airway cells. Am J Physiol Lung Cell Mol Physiol. 2000;279: L623–30. 25. Chanson M, Berclaz PY, Scerri I, Dudez T, Wernke-Dollries K, Pizurki L, Pavirani A, Fiedler MA, Suter S. Regulation of gap junctional communication by a proinflammatory cytokine in cystic fibrosis transmembrane conductance regulator-expressing but not cystic fibrosis airway cells. Am J Pathol. 2001;158:1775–84. 26. Ehrhardt C, Kneuer C, Fiegel J, Hanes J, Schaefer UF, Kim KJ, Lehr CM. Influence of apical fluid volume on the development of functional intercellular junctions in the human epithelial cell line 16HBE14o-: implications for the use of this cell line as an in vitro model for bronchial drug absorption studies. Cell Tissue Res. 2002;308:391–400. 27. Huang S, Jornot L, Wiszniewski L, Rochat T, Suter S, Lacroix JS, Chanson M. Src signaling links mediators of inflammation to Cx43 gap junction channels in primary and transformed CFTR-expressing airway epithelial cells. Cell Commun Adhes. 2003;10:279–85. 28. Yeh TH, Su MC, Hsu CJ, Chen YH, Lee SY. Epithelial cells of nasal mucosa express functional gap junctions of connexin43. Acta Otolaryngol. 2003;123:314–20. 29. Homolya L, Steinberg TH, Boucher RC. Cell to cell communication in response to mechanical stress via bilateral release of ATP and UTP in polarized epithelia. J Cell Biol. 2000;18:349–60. 30. Tarran R, Gray MA, Evans MJ, Colledge WH, Ratcliff R, Argent BE. Basal chloride currents in murine airway epithelial cells: modulation by CFTR. Am J Physiol Cell Physiol. 1998;274:C904–13. 31. Wiszniewski L, Jornot L, Dudez T, Pagano A, Rochat T, Lacroix JS, Suter S, Chanson M. Long-term cultures of polarized airway epithelial cells from patients with cystic fibrosis. Am J Respir Cell Mol Biol. 2006;34: 1439–48. 32. Manthey D, Banach K, Desplantez T, Lee CG, Kozak CA, Traub O, Weingart R, Willecke K. Intracellular domains of mouse connexin26 and -30 affect diffusional and electrical properties of gap junction channels. J Membr Biol. 2001;181:137–48. 33. Boitano S, Safdar Z, Welsh DG, Bhattacharya J, Koval M. Cell-cell interactions in regulating lung function. Am J Physiol Lung Cell Mol Physiol. 2004;287:L455–59. 34. Abraham V, Chou ML, George P, Pooler P, Zaman A, Savani RC, Koval M. Heterocellular gap junctional communication between alveolar epithelial cells. Am J Physiol Lung Cell Mol Physiol. 2001;280: L1085–92. 35. Dunsmore SE, Rannels DE. Extracellular matrix biology in the lung. Am J Physiol. 1996;270:L3–27.

370

B. Foglia et al.

36. Lee YC, Yellowley CE, Li Z, Donahue HJ, Rannels DE. Expressions of functional gap junctions in cultured pulmonary alveolar epithelial cells. Am J Physiol. 1997;272: L1105–14. 37. Abraham V, Chou ML, DeBolt KM, Koval M. Phenotypic control of gap junctional communication by cultured alveolar epithelial cells. Am J Physiol. 1999;276: L825–34. 38. Isakson BE, Evans WH, Boitano S. Intercellular Ca2þ signaling in alveolar epithelial cells through gap junctions and by extracellular ATP. Am J Physiol Lung Cell Mol Physiol. 2001;280:L221–8. 39. Alford AI, Rannels DE. Extracellular matrix fibronectin alters connexin43 expression by alveolar epithelial cells. Am J Physiol Lung Cell Mol Physiol. 2001;280:L680–8. 40. Guo Y, Martı´ nez-Williams C, Yellowley CE, Donahue HJ, Rannels DE. Connexin expression by alveolar epithelial cells is regulated by extracellular matrix. Am J Physiol Lung Cell Mol Physiol. 2001;280:L191–202. 41. Isakson BE, Lubman RL, Seedorf GJ, Boitano S. Modulation of pulmonary alveolar type II cell phenotype and communication by extracellular matrix and KGF. Am J Physiol Cell Physiol. 2001;281:C1291–9. 42. Ichimura H, Parthasarathi K, Lindert J, Bhattacharya J. Lung surfactant secretion by interalveolar Ca2þ signaling. Am J Physiol Lung Cell Mol Physiol. 2006;291:L596–601. 43. Parthasarathi K, Ichimura H, Monma E, Lindert J, Quadri S, Issekutz A, Bhattacharya J. Connexin 43 mediates spread of Ca2þ dependent proinflammatory responses in lung capillaries. J Clin Invest. 2006;116:2193–2200. 44. Nagasawa K, Chiba H, Fujita H, Kojima T, Saito T, Endo T, Sawada N. Possible involvement of gap junctions in the barrier function of tight junctions of brain and lung endothelial cells. J Cell Physiol. 2006;208:123–32. 45. Fernandez-Cobo M, Gingalewski C, De Maio A. Expression of the connexin43 gene is increased in the kidneys and the lungs of rats injected with bacterial lipopolysaccharide. Shock 1998;10:97–102. 46. Huang S, Dudez T, Scerri I, Thomas MA, Giepmans BN, Suter S and Chanson M. Defective activation of c-Src in cystic fibrosis airway epithelial cells results in loss of tumor necrosis factor-a-induced gap junction regulation. J Biol Chem. 2003;278:8326–32. 47. Essenfelder GM, Bruzzone R, Lamartine J, Charollais A, Blanchet-Bardon C, Barbe MT, Meda P and Waksman G. Connexin30 mutations responsible for hidrotic ectodermal dysplasia cause abnormal hemichannel activity. Hum Mol Genet. 2004;13:1703–14. 48. Goliger JA, Paul DH. Expression of gap junction proteins Cx26, Cx31.1, Cx37, and Cx43 in developing and mature rat epidermis. Dev Dyn. 1994;200:1–13. 49. Choudrhy R, Pitts JD, Hodgins MB. Changing patterns of gap junctional intercellular communication and connexin distribution in mouse epidermis and hair follicles during embryonic development. Dev Dyn. 1997;210:417–30. 50. Djalilian AR, McGaughey D, Patel S, Seo EY, Yang C, Cheng J, Tomic M, Sinha S, Ishida-Yamamoto A and Segre JA. Connexin26 regulates epidermal barrier and wound remodeling and promotes psoriasiform response. J Clin Invest. 2006;116:1243–53. 51. King TJ, Lampe PD. The gap junction protein connexin32 is a mouse lung tumor suppressor. Cancer Res. 2004;64:7191–96.

Chapter 17

Connexins in Skeletal Biology Roberto Civitelli and Henry J. Donahue

Abstract The skeleton is a dynamic structure that constantly remodels in response to local and systemic stimuli to meet the needs of structural integrity, mechanical competence, and maintenance of mineral homeostasis. Control of bone remodeling requires coordinated activity among osteoblasts, osteocytes, and osteoclasts. In recent years, knowledge about the biological role of connexins in the skeletal system has significantly advanced, primarily as a result of studies involving mouse and human connexin genetics. Cx43, the most abundant gap junction protein in the skeleton, is required for normal skeletal development (bone modeling) and for its maintenance in postnatal life (bone remodeling). These biological functions are underscored by the skeletal malformations and severe osteopenia present in oculodentodigital dysplasia, a disease linked to CX43 gene (GJA1) mutations, and by Gja1 ablation in mice. Cx43 modulates osteoblast differentiation and function by allowing full responses to extracellular cues via upregulation of specific signaling pathways converging on connexinsensitive transcriptional units. Other connexins are present in the skeletal tissue, but their function is only partially understood. Gap junctional intercellular communication and gap junction hemichannels are also critical in mechanostransduction, functioning to integrate and amplify mechanical signals throughout bone cell networks. Keywords Bone Cartilage Mechanotransduction Oculodentodigital dysplasia Skeletal malformation ERK signaling Gene expression Cx40 Cx43 Cx45 Cx46

R. Civitelli (*) Division of Bone and Mineral Diseases, Department of Internal Medicine, 660 South Euclid Avenue, PO Box 8301, St. Louis, MO 63110, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_17, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

371

372

R. Civitelli and H.J. Donahue

17.1 Introduction Connexins and gap junctional communication serve critical functions in many aspects of skeletal development (modeling) and in adult bone maintenance (remodeling). This chapter summarizes the essential, current connexin biology as it applies to skeletal development, growth, maintenance, and response to stimuli, with emphasis on mechanotransduction.

17.2 Bone Development All the cells of the skeleton, including chondrocytes (cells producing and maintaining cartilage), osteoblasts (bone-forming cells), osteocytes (cells embedded in bone), and osteoclasts (bone-resorbing cells) express gap junction proteins, primarily Cx43 [1,2,3,4,5]. Cx45 and Cx46 are also present in osteoblasts, though their functions in these cells remain unknown [1,2]. Cx40 is also expressed in the embryonic skeleton, primarily in developing limbs, ribs, and sternum [6]. Expression levels of connexins change during differentiation and tissue morphogenesis. For example, in developing chick limb buds, Cx43 expression is strong in the apical ectodermal ridge and almost absent in nonridge ectoderm, suggesting that it may compartmentalize ridge cells and direct limb outgrowth [7,8]. Cx43 is also transiently expressed in precartilage condensation of the carpal and metacarpal bones [8]. The functional importance of Cx43 for limb development was initially demonstrated by the severe limb malformations that occur upon inhibition of Cx43 gene (Gja1) expression by antisense oligonucleotides, which include truncation of the limb bud, and its fragmentation or complete splitting into two or three branches [9]. Work in mouse and human genetics has further established the importance of Cx43 in skeletal development. Germline null mutations of Cx43 (Gja1–/–) result in hypomineralization of craniofacial bones, with a severe delay in ossification of the axial and appendicular skeleton [10]. The ossification defects affect skeletal elements of both neural crest and mesoderm origin, including the cranial vault, clavicles, ribs, vertebrae, and limbs. In addition, osteoblasts isolated from calvaria (i.e., roof of the parietal and occipital skull bones) of Gja1–/– mice exhibit reduced osteogenic differentiation and mineralization potential, as well as reduced expression of osteocalcin, a1(I) collagen and bone sialoprotein—three critically important bone extracellular matrix proteins [10]. These cellular abnormalities are reproduced in mice with an osteoblast-specific deletion of Gja1 [11], although these animals do not exhibit the craniofacial malformations nor the ossification defects of the germline Gja1–/– mutation. Since in this model gene ablation was obtained using a fragment of the a1(I) collagen promoter [11,12], these differences are most likely related to the timing of conditional gene inactivation, which occurs around birth, therefore after most of the skeleton has been modeled.

17 Skeletal Biology

373

As noted, malformations of facial bone development leading to aberrant maxillary and mandibular primordium and nasal pit defects have been described in chick embryos exposed to Cx43 antisense oligonucleotides [13,14]. Interestingly, these malformations are associated with downregulation of the homeobox transcription factor Msx1 [14], an important modulator of cranial development and patterning [15,16]. Work in zebrafish further establishes the important role of Cx43 in skeletal development across different species. The zebrafish fin is composed of bony segments that determine the length of the fin skeleton [17]. Zebrafish homozygous for the short fin (sof) mutation have tail segments that are approximately one-third the length of wild-type bony segments. Though the molecular details are still unclear, mutations of Gja1 result in defects of cell proliferation and osteogenic differentiation in zebrafish [18]. Considering the substantial corpus of data that has emerged from different species and genetic models pointing to a critical role of Cx43 in skeletal growth and development, it is surprising that germline deletion of Gja1 in mice does not severely affect skeletal patterning. This might be due to compensatory mechanisms, at least in the mouse, and perhaps involve other connexins, most likely Cx45 (encoded by Gjc1). Cx40 (encoded by Gja5) may provide another compensatory mechanism in early skeletogenesis and patterning. Studies in a mouse model of Holt-Oram syndrome, a human disorder characterized by limb malformations and heart disease caused by haploinsufficiency of T-box transcription factor, Tbx5 [19], reveal that many of the skeletal abnormalities present in Tbx5þ/– mice are shared by Gja5þ/– and Gja5–/– mice [6], and that Tbx5 regulates expression of Gja5 [6]. Interestingly, while dysmorphisms of phalanges, carpal bones, and sternum are common to Tbx5þ/– and Gja5 mutant mice, other features, such as rib and hindlimb defects were observed only in Gja5–/– mice, suggesting that Cx40 is involved in skeletal patterning during development (see below). However, there is no evidence that Gja5 is expressed in the postnatal skeleton. Perhaps the strongest evidence for a critical role of CX43 in skeletal development emerged from the demonstration of a linkage of the human disease oculodentodigital dysplasia (ODDD) to the GJA1 locus. To date, at least 24 distinct point mutations of GJA1 have been identified in subjects with this disease [20,21,22]. ODDD affects primarily the skeleton, with widened alveolar ridge in the mandible, dentition abnormalities including microdontia, anodontia, and enamel hypoplasia, but also cranial hyperostosis [23,24]. Some affected patients also have syndactyly of hands and feet, or hypoplasia or aplasia of the middle phalanges and broad tubular bones [24]. In vitro functional characterization of GJA1 mutants found in patients with ODDD has demonstrated that while most of the mutant proteins can assemble into gap junction plaques, they do not form functional channels [25]. In fact, some of them function as dominant-negatives for CX43-mediated gap junctional intercellular communication (GJIC) [26]. A mouse model derived from random N-ethyl-N-nitrosourea mutagenesis carrying a point mutation of Gja1

374

R. Civitelli and H.J. Donahue

Table 17.1 Comparison of phenotypic features among different GJA1 mutations 1 2 3 4 ODDD Gja1jrt/þ Gja1–/– cKO Mutation Craniofacial malformations þ þ þ Enamel hypoplasia þ þ ? Tooth growth defects ? þ ? Syndactyly þ þ Delayed ossification þ þ þ Osteopenia (low bone mass) ? þ n.a þ Osteoblast dysfunction ? þ þ Hematopoietic defects ? þ þ þ 1 ODDD: oculodentodigital dysplasia, human, common features [23–25]. 2 Gja1Jrt/þ: G60S mutation, mouse [28]. 3 Gja1–/–: germline null mutation, mouse [9]. 4 cKO: conditional Gja1 knockout using a fragment of the a1(I) collagen promoter [10].

(Gja1Jrt/þ) exhibits a phenotype closely resembling that of ODDD, including syndactyly, enamel hypoplasia, craniofacial abnormalities, and cardiac dysfunction [27]. The Gja1 mutation present in these mice, which produces a G60S substitution, has never been described in patients with ODDD, but it clearly results in a mutant protein with dominant-negative properties, as do most of the ODDD mutations. Intriguingly, the Gja1Jrt/þ mouse is only a partial phenocopy of human ODDD, the most notable difference being the absence of thickened bones of the cranial vault, frequently reported in ODDD patients [24]. On the other hand, abnormalities not typically described in ODDD are present in Gja1Jrt/þ and in Gja1–/– mice, as summarized in Table 17.1. Some of these discrepancies might be related to species differences, others to differences between recessive and dominant-negative mutations. Nonetheless, the fact that skeletal abnormalities represent the major phenotypic feature of ODDD, in humans and in a mouse model, provides genetic proof that skeletal development is one of the major sites of action of Cx43. The role of other connexins in the skeleton remains to be determined. Cx40 (Gja5) null mutants die very early in embryogenesis, before condensation of skeletal elements occurs [28]. Cx46 (Gja3) is present in osteoblastic cells, but is retained in trans-Golgi compartments and does not form functional membrane channels [29]. Of note, Gja3 null mice are viable and no major skeletal phenotype has been reported [30].

17.3 Postnatal Bone Maintenance While abundant in vitro data suggest that Cx43-mediated GJIC or hemichannel activity are important for osteoblast differentiated function (reviewed in [31,32]), only with the development of conditional, osteoblast-specific Gja1 deletion animal models, which overcome the perinatal lethality of germline

17 Skeletal Biology

375

Gja1 ablation [33], has the role of Cx43 in adult bone homeostasis been brought to the fore. As noted above, Gja1 ablation in cells controlled by the a1(I) collagen promoter — active in differentiated osteoblasts — does not cause skeletal malformations, but results in reduced bone mass (osteopenia) at maturity and throughout adult life [11]. A similar degree of osteopenia is present in Gja1Jrt/þ mice [27]; however, while Gja1 ablation severely impairs osteoblast differentiation and mineralization potential, resulting in decreased new bone formation in vivo [11], the cellular bases of the low bone mass in Gja1Jrt/þ mice remain to be clarified. In contrast to Gja1 deletion, expression of ODDD mutants in osteoblast-enriched calvaria cells does not lead to major functional abnormalities [34], suggesting that the dominant-negative effect of the mutant CX43 may not be sufficient to disrupt differentiation in committed osteoblasts. However, it may be sufficient to interfere with earlier steps of osteogenesis, perhaps hindering commitment from stromal cell precursors. In fact, preliminary results from a new mouse model in which Gja1 is deleted embryonically in cells that give rise to chondro-osteoprogenitors exhibit much more severe osteogenic defects than osteoblast-specific Gja1 ablated mice [35]. Intriguingly, hematopoietic precursor number and adipogenesis are increased in Gja1Jrt/þ mice, in spite of normal peripheral blood cell numbers [27], and mice with conditional deletion of Gja1 in bone marrow cells have severely impaired recovery of hematopoiesis after cytoablative treatment, even with bone marrow transplant of wild-type cells [36]. These emerging data expand the role of Cx43 to the bone marrow microenvironment, suggesting that Cx43 serves critical functions in bone marrow stem cell (both stromal and hematopoietic) commitment and differentiation, and that osteoblasts may support hematopoiesis in a Cx43-dependent manner. The availability of new in vivo models of tissue-specific Gja1 ablation should help address these novel and exciting actions of connexins in bone marrow biology. Notably, osteocytes as well as osteoblasts are deleted when Cx43 is suppressed using either the a1(I) collagen or osteocalcin promoters [11,12]. Osteocytes represent terminal differentiation of the osteoblast lineage, and as discussed later, they are intimately involved both in sensing mechanical signals applied to bone and in mediating the action of pharmacologic agents active on bone remodeling. Thus, it seems likely that Cx43 gap junctions or hemichannels in osteocytes may have a more critical role in the elaboration of adaptive responses of bone cells to mechanical, hormonal, or pharmacologic stimuli, as has been previously suggested [37,38]. Osteocyte-specific gene ablation models, currently under development, will certainly help clarify the role of connexins in osteocytes, the most abundant in the skeleton. Recent data offer proof that Cx43 is involved in conditioning bone cell responses to hormonal stimulation. Earlier in vitro data demonstrated that interference with Gja1 expression by antisense nucleotides diminishes cyclic adenosine monophosphate (cAMP) production in response to parathyroid hormone (PTH) [39], and that PTH upregulates Cx43 protein and gap junctional intercellular communication (GJIC) in osteoblastic cell networks [3,40,41].

376

R. Civitelli and H.J. Donahue

Furthermore, the ability of PTH to induce matrix mineralization in mature osteoblasts is markedly reduced when gap junctions are inhibited [42]. These data strongly suggest that the stimulatory effect of PTH on bone formation, which is observed in vivo with intermittent, low-dose administration of PTH [43,44], could be dependent on Cx43. In direct support of this hypothesis, treatment of osteoblast Gja1-deleted mice with intermittent daily doses of PTH results in severely attenuated increments in bone mass relative to wild-type mice, with reduced activation of bone formation rates, the consequence of an inability of Cx43 deficient osteoblasts to mount a full response to the hormone [11]. These results have important ramifications for translational research, as PTH analogues are currently used as anabolic therapy for osteoporosis [45]. It is tempting to speculate whether other bone anabolic stimuli, such as mechanical loading, would also be attenuated in conditions of Gja1 deficiency. Involvement of Cx43 in the skeletal response to pharmacologic agents may also include the response to inhibitors of bone resorption. In vitro data have demonstrated that the bisphosphonate alendronate can prevent pharmacologically induced apoptosis in osteoblasts and osteocyte-like cells, and that this effect requires Cx43 [46]. The antiapoptotic action of bisphosphonates seems to be mediated not by effects on GJIC, but by alendronate-induced opening of Cx43 hemichannels, in a src-extracellular signal-regulated kinase (ERK)-dependent fashion [47]. The intriguing role of Cx43 in the mechanism of action of bisphosphonates can now be tested in vivo.

17.4 Regulation of Gene Expression by Connexins The molecular mechanisms by which alteration in gap junction or hemichannel function affects gene expression have received recent attention. As noted earlier, primary osteoblasts isolated from Gja1–/– mice express lower levels of many markers of osteogenic differentiation [10]. Building on the observation that the biophysical properties of Cx45 channels are dominant in heteromeric Cx43/ Cx45 channels, and thus that Cx45 diminishes permeability and electrical conductance of Cx43 channels [48,49,50], it has been shown that expression of chick Gjc1 (encoding Cx45) downregulates transcription of osteoblast genes, primarily osteocalcin, and a1(I) collagen [51]. Conversely, expression of Gja1 in poorly coupled cells upregulates osteoblast gene transcription [51,52]. Attenuation of gene transcription upon interference with Cx43 function was reproduced in other osteoblast-like cell lines [53,54]. These data suggest that the permeability properties of homomeric Cx43 junctional channels interact with signaling pathways involved in regulation of gene activity. Recent studies have provided an in-depth analysis of gene transcription regulation by connexin, and disclosed mechanisms that might not be restricted to bone cells. Again, using chick Gjc1 overexpression to modulate Cx43-containing channels for studies on protein-DNA binding and promoter transcriptional

17 Skeletal Biology

377

activity, a DNA element that confers connexin sensitivity was identified in the osteocalcin promoter. This element, named connexin response element (CxRE), binds transcriptional complexes containing Sp1/Sp3 factors, and it is necessary and sufficient to confer connexin sensitivity to both the osteocalcin and a1(I) collagen promoters [55]. Subsequent studies have shown that either Gja1 deficiency or interference with Cx43 function (via Gjc1 overexpression and formation of heteromeric Cx43/Cx45 channels) alter ERK signaling, and this in turn modulates gene transcription from several osteoblast gene promoters. In fact, the low transcriptional activity of osteocalcin and a1(I) collagen promoters observed in cells with Cx43 deficiency can be mimicked by inhibition of ERK signaling and rescued by overexpression of constitutively active members of the ERK signaling cascade [56]. Furthermore, in conditions of high Cx43 abundance, both the activator Sp1 and the repressor Sp3 can occupy the promoter at a nearly 1:1 ratio, slightly favoring occupancy by Sp1. As a result, transcription from these promoters is high when Cx43 is highly expressed. In contrast, when Cx43 function is inhibited, Sp3 almost exclusively occupies the CxRE, resulting in low transcriptional activity. Thus, ERK cascade-dependent phosphorylation of Sp1 mediates preferential recruitment of Sp1 over Sp3 in well-coupled cells, and loss of Sp1 phosphorylation results in the preferential recruitment of Sp3 [56]. Based on these data, a model of connexin-dependent transcriptional regulation has been proposed, whereby activation of signaling cascades by extracellular ligand-receptor binding triggers a primary response in cells in which the receptor is expressed, whose magnitude is dependent primarily on ligand availability and receptor abundance. Signals generated by this primary response, for example, cAMP, inositol triphosphate (IP3), or cyclic adenosine diphosphate (cADP)-ribose, which can permeate gap junction channels formed by Cx43, are propagated to adjacent cells, where a secondary signaling cascade (e.g., ERK cascade in the above example) can be initiated. Thus, GJIC permits a propagated secondary response, which can potentiate the primary response by allowing a larger number of cells than those expressing the receptor to participate to the response [31,56]. It would be interesting to determine whether this mechanism, most likely not the only one by which connexins regulate gene transcription, also operates in other cell systems or other gene promoters. Cell–cell diffusion of signaling molecules may also amplify this secondary response by triggering autoregenerative mechanisms (i.e., capacitive Ca2þ release), or intercellular Ca2þ waves (see below), thus potentiating and coordinating the response within a coupled cell network. In addition, diffusion of second messengers from cell to cell may serve to equalize a response to a signal that is received by only a fraction of cells in a population. In fact, it is likely that only a small percentage of cells can respond to an extracellular signal, because not all cells express receptors, channels, or enzymes appropriate to detect a particular extracellular cue. Signal amplification via a well-coupled cell population would allow the cell ensemble to build a greater response than could be accomplished by a fraction of individual cells.

378

R. Civitelli and H.J. Donahue

17.5 Mechanotransduction Mechanical loading is anabolic to bone while disuse is catabolic. Substantial inroads have been made in understanding the mechanism by which bone tissue adapts to mechanical load, and a critical role for gap junctions, and perhaps hemichannels, has emerged [57,58]. Load-induced bone matrix deformation leads to a complex, nonuniform physical environment within the tissue consisting of fluid flow, direct mechanical strain, and electrokinetic effects on bone cells. Each of these biophysical signals affects the level or activity of various markers of bone cell activity or differentiation, thus contributing to the anabolic effect of mechanical load. In particular, electric fields and substrate deformation increase alkaline phosphatase activity and osteopontin messenger RNA, respectively [59,60]. Furthermore, application of fluid flow results in cytosolic Ca2þ mobilization, nuclear factor kB (NF-kB) activation, connexin protein expression and phosphorylation, osteopontin mRNA upregulation, and prostaglandin E2 (PGE2) release [61,62,63,64]. A critical role of GJIC in bone cells’ response to mechanical stimuli is further underlined by observations that gap junction–deficient cells are dramatically less responsive to fluid flow [63,65] and electric fields [66] than are normal osteoblastic cells. Thus, as in the case of PTH [11,39], GJIC increases the sensitivity of bone cells to mechanical signals [67]. There is abundant evidence suggesting that GJIC itself is regulated by mechanical signals. Application of equibiaxial strain by substrate deformation increases GJIC and Cx43 abundance at cell–cell contact sites in osteoblastic cells [41] and upregulates Gja1 expression in tendon cells [68]. Further, fluid flow increases gap junction expression and function in osteocytic cells [5,64,69,70], although this is not universally found and even opposite effects have been noted [71]. More consistent effects have been observed with mechanical stimulation in vivo, using different approaches. Application of mechanical load to rat metatarsal bones increases the incidence of osteocytic gap junctions [72], and expression of Cx43 by osteocytes is increased in areas of bone exposed to mechanical load during experimental tooth movement [73,74]. Thus, there is evidence that gap junctions contribute to the integration and amplification of biophysical signals within osteocytic-osteoblastic networks. As alluded to in the previous section, gap junctions may contribute to the integration and amplification of mechanical signals by facilitating signal propagation from cells that directly detect them to those that do not. As in many other cell systems, gap junctions among cultured chondrocytes and osteoblastic cells contribute to propagation of membrane deformation-induced Ca2þ waves [75–78]. Interestingly, the mechanisms by which connexins mediate this form of intercellular signaling are not limited to providing GJIC. Propagation of mechanically induced Ca2þ waves not only occurs via gap junctions but also through release of adenosine triphosphate (ATP), which, via activation of P2Y purinergic receptors, stimulates mobilization of intracellular Ca2þ [77,78].

17 Skeletal Biology

379

The possibility that ATP release in response to mechanical perturbation is through gap junction hemichannels in astrocytes and glioma cells [79,80] raises the intriguing prospect that connexins contribute to bone cell mechanotransduction both through GJIC and through ATP release via gap junction hemichannels. However, it is still unclear whether the latter mechanism operates in bone cells. Functional hemichannels were demonstrated in a human osteoblastic cell line, but shown not to be involved in ATP release [81]. Others were unable to detect functional hemichannels in rat osteoblastic cell lines, though a small percentage of primary culture human osteoblastic cells did express functional hemichannels, but only in low extracellular Ca2þ [82]. Fluid flow also stimulates the release of ATP from osteoblastic cells but this is not inhibited by blocking Cx43 hemichannels [83]. On the other hand, Cx43 hemichannels appear to be active in osteocytic cells, where they are reported to mediate fluid flow-induced PGE2 [84] and ATP [85] release. It should be noted that these discrepancies, in part, may be related to the difficulty in rigorously distinguishing hemichannel-mediated diffusion from exocytosis or other related events. According to recently proposed criteria for demonstration of hemichannel-mediated events [86], the existing experimental evidence in bone cell mechanotransduction would fulfill all the criteria suggested for minimal evidence of hemichannel involvement, but only a few criteria for strong evidence. Clearly, this is an area where additional research efforts are required (see Chapters 7 and 12). While accumulating evidence suggests that GJIC and ATP may have important roles in transducing mechanical signals throughout bone cell networks, most of this evidence has been obtained from experiments that examined osteoblastic cells in monoculture. However, the majority of cells in bone are osteocytes, which are the best candidates for detecting and coordinating responsiveness to mechanical signals and communicating these signals to osteoblasts and perhaps other skeletal cells to affect their behavior. While GJIC contributes to mechanically induced Ca2þ wave propagation from osteocytic to osteoblastic cells [87], surprisingly few studies have examined the physiological consequence of mechanical signals detected by osteocytes and communicated to osteoblasts. Using a novel coculture system whereby osteocytic cells are in physical contact and communicating via gap junctions with osteoblastic cells, it was recently found that when only osteocytes but not osteoblasts are exposed to fluid flow, alkaline phosphatase activity increases in the former cells, an indication of differentiated function. This does not occur when fluid flow is directly applied to the osteoblastic cells. Importantly, this ability of osteocytic cells to communicate the detection of a mechanical signal to osteoblastic cells can be blocked by pharmacologic inhibition of GJIC, or by removal of physical contact between osteocytic and osteoblastic cells [88]. Interestingly, neither fibroblastic nor osteoblastic cells are able to communicate mechanical signals to osteoblasts, suggesting that this may be a unique attribute of osteocytic cells. Such findings would be consistent with a model by which osteocytes function as the mechanical sensors, and then transmit mechanically generated signals

380

R. Civitelli and H.J. Donahue

throughout the osteocytic network and to cells on the bone surface, to effect a bone remodeling response. The next challenge will be to transfer these intriguing new in vitro data to in vivo settings, a challenge that could theoretically be tackled by conditional or inducible gene recombination approaches.

Fig. 17.1. Gap junctional integration and amplification of biophysical signals within bone cell networks. The upper panel represents the bone itself, and the lower panel cells at the bone surface. Osteocytic cells (uppermost, spiky) are the best situated to detect mechanical signals including endogenous electric fields (lightning bolt), substrate deformation or fluid flow induced shear stress. Gap junctions (ellipsoids) allow osteocytic cells that detect the signal to communicate that detection to other osteocytes as well as osteoblasts (middle, dark blue). Gap junctions also allow communication among osteoblasts, osteoblasts, and bone marrow stromal cells (lower, light blue), assuring a net cell ensemble response greater than would have occurred if only the cells directly detecting the mechanical signal responded. Hemichannels (half circles) may allow cells to release autocrine factors (e.g., ATP), which also contributes to cell–cell communication. (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com)

17 Skeletal Biology

381

17.6 Conclusion In the past few years, substantial progress has been made in our knowledge of connexin biology in the skeleton (Fig. 17.1). It is now understood that Cx43 is a fundamental component of the osteoblast and osteocyte phenotypes, and that loss of Cx43 function leads to skeletal malformations, inability in maintaining a normal bone mass in adult life, and impairment of bone cell responses to hormonal and mechanical stimuli. Furthermore, gap junctions and hemichannels enable mechanosensitive cells, such as osteocytes, to elaborate and communicate mechanical signals through the osteocytic network and to osteoblasts on the bone surface. Many questions remain unanswered, including, among others, the role and potential redundance of other connexins expressed in skeletal cells, the role of connexins in mediating cell–cell interactions within the bone marrow microenvironment and in hematopoiesis, the mechanisms by which GJIC or hemichannel activity modulate intracellular signaling to affect cell response to extracellular cues, and whether pharmacologic modulation of GJIC can effectively modify such responses. As we unlock these functions and mechanisms, the potential for therapeutic strategies may emerge. Acknowledgments This work was supported by National Institutes of Health (NIH) grants AR41255 (to RC) and AG13087-11 (to HD). We also thank Amanda Taylor, Ryan Riddle, and Jung Lim for help in preparing the manuscript.

References 1. Civitelli R, Beyer EC, Warlow PM, Robertson AJ, Geist ST, Steinberg TH. Connexin43 mediates direct intercellular communication in human osteoblastic cell networks. J Clin Invest. 1993;91:1888–96. 2. Steinberg TH, Civitelli R, Geist ST, Robertson AJ, Hick E, Veenstra RD, Wang H-Z, Warlow PM, Westphale EM, Laing JG, Beyer EC. Connexin43 and connexin45 form gap junctions with different molecular permeabilities in osteoblastic cells. EMBO J. 1994;13:744–50. 3. Donahue HJ, McLeod KJ, Rubin CT, Andersen J, Grine EA, Hertzberg EL, Brink PR. Cell-to-cell communication in osteoblastic networks: cell line-dependent hormonal regulation of gap junction function. J Bone Miner Res. 1995;10:881–9. 4. Ilvesaro J, Vaananen K, Tuukkanen J. Bone-resorbing osteoclasts contain gap-junctional connexin-43. J Bone Miner Res. 2000;15:919–26. 5. Cheng B, Zhao S, Luo J, Sprague E, Bonewald LF, Jiang JX. Expression of functional gap junctions and regulation by fluid flow in osteocyte-like MLO-Y4 cells. J Bone Miner Res. 2001;16:249–59. 6. Pizard A, Burgon PG, Paul DL, Bruneau BG, Seidman CE, Seidman JG. Connexin 40, a target of transcription factor Tbx5, patterns wrist, digits, and sternum. Mol Cell Biol. 2005;25:5073–83. 7. Coelho CN, Kosher RA. A gradient of gap junctional communication along the anteriorposterior axis of the developing chick limb bud. Dev Biol. 1991;148:529–35. 8. Dealy CN, Beyer EC, Kosher RA. Expression patterns of mRNAs for the gap junction proteins connexin43 and connexin42 suggest their involvement in chick limb morphogenesis and specification of the arterial vasculature. Dev Dyn. 1994;199:156–67.

382

R. Civitelli and H.J. Donahue

9. Green CR, Bowles L, Crawley A, Tickle C. Expression of the connexin43 gap junctional protein in tissues at the tip of the chick limb bud is related to the epithelial-mesenchymal interactions that mediate morphogenesis. Dev Biol. 1994;161:12–21. 10. Lecanda F, Warlow PM, Sheikh S, Furlan F, Steinberg TH, Civitelli R. Connexin43 deficiency causes delayed ossification, craniofacial abnormalities, and osteoblast dysfunction. J Cell Biol. 2000;151:931–44. 11. Chung DJ, Castro CH, Watkins M, Stains JP, Chung MY, Szejnfeld VL, Willecke K, Theis M, Civitelli R. Low peak bone mass and attenuated anabolic response to parathyroid hormone in mice with an osteoblast-specific deletion of connexin43. J Cell Sci. 2006;119:4187–98. 12. Castro CH, Stains JP, Sheikh S, Szejnfeld VL, Willecke K, Theis M, Civitelli R. Development of mice with osteoblast-specific connexin43 gene deletion. Cell Commun Adhes. 2003;10:445–50. 13. Becker DL, McGonnell I, Makarenkova HP, Patel K, Tickle C, Lorimer J, Green CR. Roles for a1 connexin in morphogenesis of chick embryos revealed using a novel antisense approach. Dev Genet. 1999;24:33–42. 14. McGonnell IM, Green CR, Tickle C, Becker DL. Connexin43 gap junction protein plays an essential role in morphogenesis of the embryonic chick face. Dev Dyn. 2001;222:420–38. 15. Alappat S, Zhang ZLY, Chen YP. Analysis of Msx1; Msx2 double mutants reveals multiple roles for Msx homeobox gene family and craniofacial genes in limb development. Cell Res. 2003;13:429–42. 16. Cohen MM Jr. Craniofacial disorders caused by mutations in homeobox genes MSX1 and MSX2. J Craniofac Genet Dev Biol. 2000;20:19–25. 17. Iovine MK, Johnson SL. Genetic analysis of isometric growth control mechanisms in the zebrafish caudal Fin. Genetics 2000;155:1321–29. 18. Iovine MK, Higgins EP, Hindes A, Coblitz B, Johnson SL. Mutations in connexin43 (GJA1) perturb bone growth in zebrafish fins. Dev Biol. 2005;278:208–19. 19. Basson CT, Huang T, Lin RC, Bachinsky DR, Weremowicz S, Vaglio A, Bruzzone R, Quadrelli R, Lerone M, Romeo G, Silengo M, Pereira A, Krieger J, Mesquita SF, Kamisago M, Morton CC, Pierpont ME, Mu¨ller CW, Seidman JG, Seidman CE. Mutations in human TBX5 cause limb and cardiac malformation in Holt-Oram syndrome. Nat Genet. 1997;15:30–5. 20. Richardson R, Donnai D, Meire F, Dixon MJ. Expression of Gja1 correlates with the phenotype observed in oculodentodigital syndrome/type III syndactyly. J Med Genet. 2004;41:60–7. 21. Kjaer KW, Hansen L, Eiberg H, Leicht P, Opitz JM, Tommerup N. Novel Connexin 43 (GJA1) mutation causes oculo-dento-digital dysplasia with curly hair. Am J Med Genet. 2004;127A:152–7. 22. Paznekas WA, Boyadjiev SA, Shapiro RE, Daniels O, Wollnik B, Keegan CE, Innis JW, Dinulos MB, Christian C, Hannibal MC, Jabs EW. Connexin 43 (GJA1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet. 2003;72:408–18. 23. Loddenkemper T, Grote K, Evers S, Oelerich M, Stogbauer F. Neurological manifestations of the oculodentodigital dysplasia syndrome. J Neurol. 2002;249:584–95. 24. Schrander-Stumpel CT, Groot-Wijnands JB, Die-Smulders C, Fryns JP. Type III syndactyly and oculodentodigital dysplasia: a clinical spectrum. Genet Couns. 1993;4:271–6. 25. Shibayama J, Paznekas W, Seki A, Taffet S, Jabs EW, Delmar M, Musa H. Functional characterization of connexin43 mutations found in patients with oculodentodigital dysplasia. Circ Res. 2005;96:e83–91. 26. Roscoe W, Veitch GI, Gong XQ, Pellegrino E, Bai D, McLachlan E, Shao Q, Kidder GM, Laird DW. Oculodentodigital dysplasia-causing connexin43 mutants are nonfunctional and exhibit dominant effects on wildtype connexin43. J Biol Chem. 2005;280:11458–66.

17 Skeletal Biology

383

27. Flenniken AM, Osborne LR, Anderson N, Ciliberti N, Fleming C, Gittens JE, Gong XQ, Kelsey LB, Lounsbury C, Moreno L, Nieman BJ, Peterson K, Qu D, Roscoe W, Shao Q, Tong D, Veitch GI, Voronina I, Vukobradovic I, Wood GA, Zhu Y, Zirngibl RA, Aubin JE, Bai D, Bruneau BG, Grynpas M, Henderson JE, Henkelman RM, McKerlie C, Sled JG, Stanford WL, Laird DW, Kidder GM, Adamson SL, Rossant J. A Gja1 missense mutation in a mouse model of oculodentodigital dysplasia. Development. 2005;132:4375–86. 28. Kumai M, Nishii K, Nakamura K, Takeda N, Suzuki M, Shibata Y. Loss of connexin45 causes a cushion defect in early cardiogenesis. Development. 2000;127:3501–12. 29. Koval M, Harley JE, Hick E, Steinberg TH. Connexin46 is retained as monomers in a trans-Golgi compartment of osteoblastic cells. J Cell Biol. 1997;137:847–57. 30. Gong X-H, Li E, Klier G, Kumar NM, Gilula NB. Disruption of a3 connexin gene leads to age-related cataract formation in mice. Mol Biol Cell. 1996;7:509a. 31. Stains JP, Civitelli R. Gap junctions in skeletal development and function. Biochim Biophys Acta. 2005;1719:69–81. 32. Stains JP, Civitelli R. Cell-cell interactions in regulating osteogenesis and osteoblast function. Birth Defects Res C Embryo Today. 2005;75:72–80. 33. Reaume AG, De Sousa PA, Kulkarmi S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science 1995;267:1831–4. 34. McLachlan E, Manias JL, Gong XQ, Lounsbury CS, Shao Q, Bernier SM, Bai D, Laird DW. Functional characterization of oculodentodigital dysplasia-associated Cx43 mutants. Cell Commun Adhes. 2005;12:279–92. 35. Watkins M, Ornitz D, Willecke K, Civitelli R. Connexin43 is required for normal skeletal development and bone mass acquisition. J Bone Miner Res. 2006;21:S56. 36. Presley CA, Lee AW, Kastl B, Igbinosa I, Yamada Y, Fishman GI, Gutstein DE, Cancelas JA. Bone marrow connexin-43 expression is critical for hematopoietic regeneration after chemotherapy. Cell Commun Adhes. 2005; 12:307–17. 37. Duncan RL, Turner CH. Mechanotransduction and the functional response of bone to mechanical strain. Calcif Tissue Int. 1995;57:344–58. 38. Donahue HJ. Gap junctional intercellular communication in bone: a cellular basis for the mechanostat set point. Calcif Tissue Int. 1998;62:85–8. 39. Van der Molen MA, Rubin CT, McLeod KJ, McCauley LK, Donahue HJ. Gap junctional intercellular communication contributes to hormonal responsiveness in osteoblastic networks. J Biol Chem. 1996;271:12165–71. 40. Schiller PC, Mehta PP, Roos BA, Howard GA. Hormonal regulation of intercellular communication: parathyroid hormone increases connexin43 gene expression and gapjunctional communication in osteoblastic cells. Mol Endocrinol. 1992;6:1433–40. 41. Ziambaras K, Lecanda F, Steinberg TH, Civitelli R. Cyclic stretch enhances gap junctional communication between osteoblastic cells. J Bone Miner Res. 1998;13:218–28. 42. Schiller PC, D’Ippolito G, Balkan W, Roos BA, Howard GA. Gap-junctional communication mediates parathyroid hormone stimulation of mineralization in osteoblastic cultures. Bone 2001;28:38–44. 43. Dobnig H, Turner RT. Evidence that intermittent treatment with parathyroid hormone increases bone formation in adult rats by activation of bone lining cells. Endocrinology 1995;136:3632–8. 44. Iida-klein A, Zhou H, Lu SS, Levine LR, Ducayen-Knowles M, Dempster DW, Nieves J, Lindsay R. Anabolic action of parathyroid hormone is skeletal site specific at the tissue and cellular levels in mice. J Bone Miner Res. 2002;17:808–16. 45. Neer RM, Arnaud CD, Zanchetta JR, Prince R, Gaich GA, Reginster JY, Hodsman AB, Eriksen EF, Ish-Shalom S, Genant HK, Wang O, Mitlak BH. Effect of parathyroid hormone (1-34) on fractures and bone mineral density in postmenopausal women with osteoporosis. N Engl J Med. 2001;344:1434–41.

384

R. Civitelli and H.J. Donahue

46. Plotkin LI, Manolagas SC, Bellido T. Transduction of cell survival signals by connexin-43 hemichannels. J Biol Chem. 2002;277:8648–57. 47. Plotkin LI, Aguirre JI, Kousteni S, Manolagas SC, Bellido T. Bisphosphonates and estrogens inhibit osteocyte apoptosis via distinct molecular mechanisms downstream of extracellular signal-regulated kinase activation. J Biol Chem. 2005;280:7317–25. 48. Moreno AP, Fishman GI, Beyer EC, Spray DC. Voltage-dependent gating and single channel analysis of heterotypic gap junction channels formed of Cx45 and Cx43. In: Progress in Cell Research. Y. Kanno, editor. Amsterdam, The Netherlands. Elsevier Science B.V.; 1995. pp. 405–8. 49. Koval M, Geist ST, Westphale EM, Kemendy AE, Civitelli R, Beyer EC, Steinberg TH. Transfected connexin45 alters gap junction permeability in cells expressing endogenous connexin43. J Cell Biol. 1995;130:987–95. 50. Martı´ nez AD, Hayrapetyan V, Moreno AP, Beyer EC. Connexin43 and connexin45 form heteromeric gap junction channels in which individual components determine permeability and regulation. Circ Res. 2002;90:1100–7. 51. Lecanda F, Towler DA, Ziambaras K, Cheng S-L, Koval M, Steinberg TH, Civitelli R. Gap junctional communication modulates gene expression in osteoblastic cells. Mol Biol Cell. 1998;9:2249–58. 52. Gramsch B, Gabriel HD, Wiemann M, Grummer R, Winterhager E, Bingmann D, Schirrmacher K. Enhancement of connexin 43 expression increases proliferation and differentiation of an osteoblast-like cell line. Exp Cell Res. 2001;264:397–407. 53. Schiller PC, D’Ippolito G, Brambilla R, Roos BA, Howard GA. Inhibition of gapjunctional communication induces the trans-differentiation of osteoblasts to an adipocytic phenotype in vitro. J Biol Chem. 2001;276:14133–8. 54. Upham BL, Suzuki J, Chen G, Wang Y, McCabe LR, Chang CC, Krutovskikh VA, Yamasaki H, Trosko JE. Reduced gap junctional intercellular communication and altered biological effects in mouse osteoblast and rat liver oval cell lines transfected with dominant-negative connexin 43. Mol Carcinog. 2003;37:192–201. 55. Stains JP, Lecanda F, Screen J, Towler DA, Civitelli R. Gap junctional communication modulates gene transcription by altering the recruitment of Sp1 and Sp3 to connexinresponse elements in osteoblast promotors. J Biol Chem. 2003;278:24377–87. 56. Stains JP, Civitelli R. Gap junctions regulate extracellular signal-regulated kinase signaling to affect gene transcription. Mol Biol Cell. 2005;16:64–72. 57. Hughes-Fulford M. Signal transduction and mechanical stress. Sci STKE. 2004;2004:RE12. 58. Robling AG, Castillo AB, Turner CH. Biomechanical and molecular regulation of bone remodeling. Annu Rev Biomed Eng. 2006;8:455–98. 59. McLeod KJ, Donahue HJ, Levin PE, Fontaine MA, Rubin CT. Electric fields modulate bone cell function in a density-dependent manner. J Bone Miner Res. 1993;8: 977–84. 60. You J, Reilly GC, Zhen X, Yellowley CE, Chen Q, Donahue HJ, Jacobs CR. Osteopontin gene regulation by oscillatory fluid flow via intracellular calcium mobilization and activation of mitogen-activated protein kinase in MC3T3-E1 osteoblasts. J Biol Chem. 2001;276:13365–13371. 61. Jacobs CR, Yellowley CE, Davis BR, Zhou Z, Cimbala JM, Donahue HJ. Differential effect of steady versus oscillating flow on bone cells. J Biomech. 1998;31:969–76. 62. Kurokouchi K, Jacobs CR, Donahue HJ. Oscillating fluid flow inhibits TNF-a-induced NF-kappa B activation via an Ikappa B kinase pathway in osteoblast-like UMR106 cells. J Biol Chem. 2001;276:13499–504. 63. Saunders MM, You J, Zhou Z, Li Z, yellowley CE, Kunze EL, Jacobs CR, Donahue HJ. Fluid flow-induced prostaglandin E2 response of osteoblastic ROS 17/2.8 cells is gap junction-mediated and independent of cytosolic calcium. Bone 2003;32:350–6. 64. Alford AI, Jacobs CR, Donahue HJ. Oscillating fluid flow regulates gap junction communication in osteocytic MLO-Y4 cells by an ERK1/2 MAP kinase-dependent mechanism. Bone 2003;33:64–70.

17 Skeletal Biology

385

65. Saunders MM, You J, Trosko JE, Yamasaki H, Li Z, Donahue HJ, Jacobs CR. Gap junctions and fluid flow response in MC3T3-E1 cells. Am J Physiol Cell Physiol. 2001;281:C1917–25. 66. Van der Molen MA, Donahue HJ, Rubin CT, McLeod KJ. Osteoblastic networks with deficient coupling: differential effects of magnetic and electric field exposure. Bone 2000;27:227–31. 67. Donahue HJ. Gap junctions and biophysical regulation of bone cell differentiation. Bone 2000;26:417–22. 68. Wall ME, Banes AJ. Early responses to mechanical load in tendon: role for calcium signaling, gap junctions and intercellular communication. J Musculoskelet Neuronal Interact. 2005;5:70–84. 69. Cheng B, Kato Y, Zhao S, Luo J, Sprague E, Bonewald LF, Jiang JX. PGE2 is essential for gap junction-mediated intercellular communication between osteocyte-like MLO-Y4 cells in response to mechanical strain. Endocrinology 2001;142:3464–73. 70. Cherian PP, Cheng B, Gu S, Sprague E, Bonewald LF, Jiang JX. Effects of mechanical strain on the function of gap junctions in osteocytes are mediated through the prostaglandin EP2 receptor. J Biol Chem. 2003;278:43146–56. 71. Thi MM, Kojima T, Cowin SC, Weinbaum S, Spray DC. Fluid shear stress remodels expression and function of junctional proteins in cultured bone cells. Am J Physiol Cell Physiol. 2003;284:C389–403. 72. Lozupone E, Palumbo C, Favia A, Ferretti M, Palazzini S, Cantatore FP. Intermittent compressive load stimulates osteogenesis and improves osteocyte viability in bones cultured ‘in vitro’. Clin Rheumatol. 1996;15:563–572. 73. Su M, Borke JL, Donahue HJ, Li Z, Warshawsky NM, Russell CM, Lewis JE. Expression of connexin 43 in rat mandibular bone and periodontal ligament (PDL) cells during experimental tooth movement. J Dent Res. 1997;76:1357–66. 74. Gluhak-Heinrich J, Gu S, Pavlin D, Jiang JX. Mechanical loading stimulates expression of connexin 43 in alveolar bone cells in the tooth movement model. Cell Commun Adhes. 2006;13:115–25. 75. D’Andrea P, Vittur F. Gap junctions mediate intercellular calcium signaling in cultured articular chondrocytes. Cell Calcium. 1996;20:389–97. 76. Donahue HJ, Guilak F, Van der Molen MA, McLeod KJ, Rubin CT, Grande DA, Brink PR. Chondrocytes isolated from mature articular cartilage retain the capacity to from functional gap junctions. J Bone Miner Res. 1995;10:1359–64. 77. Jørgensen NR, Geist ST, Civitelli R, Steinberg TH. ATP- and gap junction-dependent intercellular calcium signaling in osteoblastic cells. J Cell Biol. 1997;139:497–506. 78. Jørgensen NR, Henriksen Z, Brot C, Eriksen EF, Sorensen OH, Civitelli R, Steinberg TH. Human osteoblastic cells propagate intercellular calcium signals by two different mechanisms. J Bone Miner Res. 2000;15:1024–32. 79. Cotrina ML, Lin JH, Alves-Rodrigues A, Liu S, Li J, Azmi-Ghadimi H, Kang J, Naus CC, Nedergaard M. Connexins regulate calcium signaling by controlling ATP release. Proc Natl Acad Sci USA. 1998;95:15735–40. 80. Stout CE, Costantin JL, Naus CC, Charles AC. Intercellular calcium signaling in astrocytes via ATP release through connexin hemichannels. J Biol Chem. 2002;277:10482–8. 81. Romanello M, Pani B, Bicego M, D’Andrea P. Mechanically induced ATP release from human osteoblastic cells. Biochem Biophys Res Commun. 2001;289:1275–81. 82. Jørgensen NR, Teilmann SC, Henriksen Z, Civitelli R, Sorensen OH, Steinberg TH. Activation of L-type calcium channels is required for gap junction-mediated intercellular calcium signaling in osteoblastic cells. J Biol Chem. 2003;278:4082–86. 83. Genetos DC, Geist DJ, Liu D, Donahue HJ, Duncan RL. Fluid shear-induced ATP secretion mediates prostaglandin release in MC3T3-E1 osteoblasts. J Bone Miner Res. 2005;20:41–9.

386

R. Civitelli and H.J. Donahue

84. Cherian PP, Siller-Jackson AJ, Gu S, Wang X, Bonewald LF, Sprague E, Jiang JX. Mechanical strain opens connexin 43 hemichannels in osteocytes: a novel mechanism for the release of prostaglandin. Mol Biol Cell. 2005;16:3100–6. 85. Genetos DC, Kephart CJ, Zhang Y, Yellowley CE, Donahue HJ. Oscillating fluid flow activation of gap junction hemichannels induces ATP release from MLO-Y4 osteocytes. J Cell Physiol. 2007;212:207–14. 86. Spray DC, Ye ZC, Ransom BR. Functional connexin ‘hemichannels’: a critical appraisal. Glia 2006;54:758–73. 87. Yellowley CE, Li Z, Zhou Z, Jacobs CR, Donahue HJ. Functional gap junctions between osteocytic and osteoblastic cells. J Bone Miner Res. 2000;15:209–17. 88. Taylor AF, Saunders MM, Shingle DL, Cimbala JM, Zhou Z, Donahue HJ. Mechanically stimulated osteocytes regulate osteoblastic activity via gap junctions. Am J Physiol Cell Physiol. 2007;292:C545–52.

Chapter 18

Connexins in Lens Development and Disease Teresa I. Shakespeare, Richard T. Mathias, and Thomas W. White

Abstract Gap junctions are responsible for the coupling of cells that enables intercellular exchange of ions, small metabolites, and nutrients; lens homeostasis depends on these intercellular connections. Aberrant expression of the genes encoding lens connexins (Cx43, Cx46, and Cx50), and function of the connexins themselves, have been linked to cataractogenesis and ocular growth defects. The use of in vivo and in vitro experimental models has provided significant advances in understanding the function of gap junctions in the lens. Current data suggest that Cx46 is required for maintenance of lens clarity via its effects on Ca2þ homeostasis, and that Cx50 is required for proper lens growth. The substitution of either of these connexins for the other only partially ameliorates the pathological effects. In spite of these advances, questions about the mechanisms by which connexin-related pathologies occur in the mammalian lens remain largely unanswered. Keywords Lens Cataract Growth factor Development Cx43 Cx46 Cx50

18.1 Introduction Three connexin genes have distinct spatial and temporal expression in the vertebrate lens (Fig. 18.1). Cx43, Cx46, and Cx50 share the responsibility to promote proper development and maintain lens homeostasis [1]. Gap junction channels composed of Cx43 and Cx50 are found between lens epithelial cells on the anterior surface. However, as these cells differentiate into the fiber cells that make up the bulk of the lens volume, Cx43 is significantly downregulated. In contrast, Cx50 and Cx46 are upregulated to form the channels that couple fiber cells [2,3,4,5]. This developmentally regulated change in the repertoire of T.W. White (*) Department of Physiology and Biophysics, State University of New York, T5-147, Basic Science Tower, Stony Brook, NY 11794, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_18, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

387

388

T.I. Shakespeare et al.

Fig. 18.1 Role of connexins in lens internal circulation. (a) A sketch of the cellular organization of the vertebrate lens cut in cross section along the anterior-posterior axis showing the lens internal circulation. There are three physiologically distinct zones. The anterior surface is covered by a simple epithelium expressing Cx43 and Cx50 (light gray). Below the epithelium, the peripheral 20% of the lens is made up of a shell of differentiating fibers that make Cx46 and Cx50 (white). The central 80% of the lens contains the mature fibers lacking organelles and continuing to express Cx46 and Cx50 (dark gray). (b) A model of how the internal circulation of the lens is generated. In an intact lens, current flows in the pattern indicated by the arrows shown in panel (a). The major ion carrier of the circulating current appears to be Naþ, which enters the lens along the extracellular spaces between cells, then moves down its electrochemical gradient into fiber cells, where it returns to the surface via gap junctions. The pattern of gap junction coupling in the differentiating fibers directs the intracellular current flow to the lens equator, where the epithelial cells transport it out of the lens using Naþ/KþATPase activity. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

connexin subunits is critical for fulfilling the physiological role of connexinmediated communication in the lens. This chapter discusses the importance of lens connexins during postnatal growth as well as recent data characterizing connexin-related lens pathologies in experimental animal models.

18.1.1 Lens Circulating Current System Although gap junctions play a key role in the maintenance of lens homeostasis, the lens also depends on Naþ, Kþ, and Cl–conductances to create an internal circulating current, with Naþ being the primary current carrier [6]. Naþ, coupled with water, enters the lens at the anterior and posterior poles and flows inward along the extracellular spaces. Naþ is driven by its electrochemical

18 Lens

389

gradient to move into the fiber cells, where the direction of flow is reversed and the current flows back to the lens surface through gap junctions (Fig. 18.1). The coupling conductance of peripheral fiber cell gap junctions is highest at the equator, directing the intracellular current to the equatorial epithelium. Naþ/ Kþ–adenosine triphosphatase (ATPase) activity is concentrated in the equatorial epithelial cells where the intracellular flux is transported out of the lens to complete the circulatory loop [7,8,9]. The lens lacks blood vessels, so the fiber cells are faced with the challenge of how to exchange metabolites and waste products. The circulating current system contributes to the solution, since water and dissolved metabolites follow the Naþ current to create a microcirculatory system for the avascular lens that carries nutrients into the fiber cells and allows removal of metabolic waste. Ca2þ also flows through the lens in a manner similar to Naþ; hence, Ca2þ homeostasis depends on gap junction coupling and the circulating current (see below) [10,11].

18.2 Genetic Alterations of Lens Connexins Suggest Connexin Specialization Genetically altered mouse models have been used to investigate the importance of the lens connexin genes in vivo (Fig. 18.2). Examination of Cx43 knockout (KO) mice revealed that prenatal lens development was largely normal; however, further studies were impeded because of neonatal lethality [12,13,14]. Deletion of Cx46 resulted in a severe, senile-type cataract, with normal ocular development and growth [2]. The cataracts observed were initiated in the nuclear region of the lens, and were the result of cleavage and precipitation of g-crystallin proteins due to the activation of the Ca2þ-activated calpain protease Lp82, a lens-specific splice variant of calpain protease p94 (calpain3) [15]. In contrast, significantly decreased lens growth with only mild cataracts was reported for Cx50KO animals, indicating divergent roles for the two principal lens connexins [16,17]. The unique contribution of Cx50 to lens homeostasis was analyzed by targeting the replacement (knock-in, KI) of Cx50 with Cx46, thus generating the Cx50KI46 model [18]. In these mice, lenses maintained clarity, but continued to display a growth deficit. Deletion of both Cx46 and Cx50 produced additive effects of a severe cataract and reduced lens growth [19]. The extension of these mechanisms to general cataract is still unclear.

18.2.1 Connexin46 Is Important for Calcium Homeostasis in the Lens Cx46 is instrumental in maintaining a low Ca2þ concentration within the fiber cells. Gao et al. [11] proposed that Cx46 in the mature fibers was responsible for the cycling of Ca2þ ions to the epithelium, where the Ca2þ-ATPase and

390

T.I. Shakespeare et al.

Fig. 18.2 Lens phenotypes of genetically modified mice. Photographs of the anterior surface of wild-type, Cx50KO, Cx46KO, and Cx46/Cx50 double knockout lenses emphasize the importance of connexin genes for lens homeostasis. (a) Wild-type lenses display normal size and transparency. (b) Cx50KO results in a mild nuclear cataract and a 50% reduction in lens size. (c) Cx46KO lenses are normal in size with a dense nuclear cataract in addition to cortical opacities. (d) Double knockout of Cx46 and Cx50 causes a severe cataract and reduced lens size. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

Naþ/Ca2þ-exchange proteins were located. In this model, when Cx46 is absent, intracellular Ca2þ would accumulate, leading to cataractogenesis. This model was supported by measurements of intracellular Ca2þ concentration and coupling conductance in wild-type, Cx46KO and Cx50KI46 lenses. Intracellular Ca2þ was highest in Cx46KO lenses, consistent with the elimination of coupling caused by the ablation of Cx46 [11]. In addition, Cx50KI46 lenses had a higher coupling conductance and lower concentrations of intracellular Ca2þ when compared to wild-type lenses. Thus, there was a strong inverse correlation between the magnitude of Cx46-mediated coupling and the intracellular Ca2þ concentration, suggesting that Cx46 was the rate-limiting factor in Ca2þ efflux. These results were consistent with the hypothesis that Ca2þ-activated proteases initiate the Cx46KO cataract [15], and were further supported by recent studies of a Cx46/calpain3 double knockout mouse model showing that Cx46KO cataracts were significantly delayed when g-crystallin cleavage was absent [20].

18 Lens

391

18.2.2 Connexin50 and Lens Growth Cx50 seems to be influential in lens development, and is expressed in both the epithelial cells and fibers. Deletion of Cx50 causes a significant lens growth defect, but only mild nuclear cataracts [16,18]. Crystallin precipitation produces the cataract, while the growth defect results from fewer fiber cells. Loss or replacement of Cx50 (as in Cx50KO or Cx50KI46) caused a decrease in the number of dividing cells during the first postnatal week, suggesting that Cx50-mediated communication is essential for peak mitosis to occur [21]. This hypothesis has been supported by analysis of Cx50/calpain3 double knockout mice. Calpain3 deletion delayed cataract formation, but not the growth defect. In addition, calpain3 deletion altered the solubility of crystallin proteins and revealed a previously undetected aB-crystallin cleavage event in Cx50KO lenses [22]. These results suggest that Cx46 is required for the maintenance of lens clarity, whereas Cx50 is mainly required for proper lens growth.

18.3 Connexins and Growth Factors Influence Lens Development While much is separately known about how connexins and growth factors influence the lens, much less is known about how they interact. Members of many growth factor families are expressed in the lens, including fibroblast growth factor (FGF), insulin-like growth factor (IGF), and lens epithelialderived growth factor (LEDGF) [23,24]. The assembly and activity of gap junctions can be regulated by growth factors and cellular stress (Table 18.1). For example, IGF-I causes the disassembly of Cx43 plaques, whereas hydrogen peroxide-induced oxidative stress activates protein kinase C (PKC) leading to connexin phosphorylation [25,26]. Activation of kinases can alter the formation of gap junction plaques by phosphorylation on serine, threonine, or tyrosine in the carboxyl-terminal domain of connexin proteins. Cx43 can be phosphorylated by PKC, protein kinase A, casein kinase I, or mitogen-activated protein kinase (MAPK) [27,28,29,30] (see Chapter 11). LEDGF is a growth, adhesive, differentiation, and antiapoptotic factor expressed in lens epithelial cells [24]. LEDGF activation of PKCg produces gap junction disassembly through phosphorylation of Cx43 [31,32]. Regulation of gap junction communication is not solely mediated by PKCg; PKCa also has been reported to increase gap junction activity in lens epithelial cells [33]. To date, studies of connexin phosphorylation in epithelial cells have focused exclusively on Cx43, and relatively little is known about the regulation of Cx50 in this cell type. Fibroblast growth factors are expressed throughout the lens, and regulate both lens cell proliferation and fiber differentiation. Fibroblast growth factor is responsible for increasing gap junction–mediated dye transfer without altering junctional protein synthesis or assembly [34]. It has been hypothesized that FGF-induced activation of extracellular signal-regulated kinases (ERK) is

392

T.I. Shakespeare et al.

Table 18.1 Growth factors that influence lens connexins Growth factors and Eye expression Function/involvement in gap junctional kinases pattern intercellular communication Fibroblast growth factor (FGF)/ FGFR1–4

Entire eye; FGF1–3, 5, 7–13, 15, 19

Lens epithelialderived growth factor (LEDGF) Insulin-like growth factor-I (IGF)/ IGFR Epidermal growth factor (EGF)/ EGFR PKCa, PKCg isoforms

Lens epithelial cells

Protein kinase A (PKA) Mitogen-activated protein kinase (MAPK)

FGF increases gap junctional intercellular communication (GJIC) without altering connexin protein synthesis or assembly [34] Activates phosphorylation of Cx43 via PKCg [24,31]

Lens epithelial cells

Phosphorylates Cx43 via PKCg to decrease GJIC [27]

Lens epithelial cells

Activates phosphorylation of Cx43 and enhances GJIC [27]

Lens epithelial cells, the g isoform is more active Lens epithelial cells

Inhibits GJIC via phosphorylation of Cx43 [29,32]

Lens epithelial cells

Activator of GJIC via phosphorylation of Cx43 [27] Closes hemichannels via phosphorylation of Cx43 [27]

highest in the equatorial region of the lenticular cells and lower in the poles. As mentioned, impedance studies have shown that junctional coupling is higher in the equatorial region than at either pole [35]; thus regional differences in FGF signaling via the ERK pathway could create the observed distribution of intercellular coupling in the lens [34].

18.4 Connexin Mutations and Disease Cataracts are one of the leading causes of blindness in humans, and mutations in CX46 and CX50 cause congenital cataract [36,37]. One form of cataract (Online Mendelian Inheritance in Man [OMIM] number 116200) was mapped to the region of chromosome 1 containing the GJA8 gene (encoding Cx50) [38,39], and mutations within the coding region have been linked to cataract in several families [40,41,42]. Similarly, mutations in the GJA3 gene (encoding CX46) also cause cataract (OMIM 0601885), and six separate mutations have been identified in the coding region from affected patients [37,43,44,45,46]. Mouse models have provided insight into the mechanistic relationship between gap junction impairment and cataract formation. In one such model, a G22R point mutation in Cx50 causes a severe cataract and lens rupture that is modulated by its interaction with Cx46. In vitro studies showed that

18 Lens

393

coexpression of mutant Cx50G22R and wild-type Cx46 subunits formed functional gap junction channels with reduced conductance and altered voltagegating. In vivo studies demonstrated that knocked-in Cx46 could rescue lens rupture as mutant Cx50G22R forms heteromeric gap junction channels with Cx46 in lens epithelial cells [47,48]. A second dominant cataractous mouse line (L1) containing a missense Cx50 mutation (S50P) has also been characterized [19]. The combination of mutant Cx50S50P and wild-type Cx50 subunits inhibited the elongation of primary fiber cells, while the combination of Cx50S50P and wild-type Cx46 subunits disrupted the formation of secondary fiber cells. These findings provide novel in vivo evidence that intercellular communication modulated by distinct gap junctions can influence the fiber cell formation during development. Further study of additional animal models harboring Cx50 mutations [47,49,50] will help elucidate the impact of junctional communication on lens homeostasis. To date, no mouse models of dominant cataracts due to mutations in Cx46 or Cx43 have been identified.

18.5 Conclusion Expression of Cx46 and Cx50 is essential for lens growth and homeostasis. The underlying mechanisms that influence lens growth and clarity, however, are not completely understood. Future questions to consider include what effect does connexin expression have on cellular mitosis. Perhaps connexin expression is synchronized with the cell cycle to ensure that specific molecules are passed between cells. If this is true, why is Cx50 more effective than Cx46 at mediating the signals involved in lens growth? It may be useful to determine the exact molecules that these intercellular channels pass between cells to explain the relationship between connexins and growth regulating factors in the ocular environment. These questions may be answered by studying the permeation properties of both connexins. Conversely, questions about Cx46 and lens clarity are also of great interest. Cx46 and Cx50 are coexpressed in the fiber cells, yet why cannot Cx50 alone maintain lens clarity? What special properties of Cx46 are required for the circulating lens current to function? Would the circulating current still flow in a Cx46KI50 lens? Future research will use genetic animal models along with in vitro assays to further elucidate the mechanisms that control lens development and disease.

References 1. Gerido DA, White TW. Connexin disorders of the ear, skin, and lens. Biochim Biophys Acta. 2004;1662:159–70. 2. Gong X, Li E, Klier G, Huang Q, Wu Y, Lei H, Kumar NM, Horwitz J, Gilula NB. Disruption of a3 connexin gene leads to proteolysis and cataractogenesis in mice. Cell. 1997;91:833–43.

394

T.I. Shakespeare et al.

3. Beyer EC, Kistler J, Paul DL, Goodenough DA. Antisera directed against connexin43 peptides react with a 43-kD protein localized to gap junctions in myocardium and other tissues. J Cell Biol. 1989;108:595–605. 4. Musil LS, Beyer EC, Goodenough DA. Expression of the gap junction protein connexin43 in embryonic chick lens: molecular cloning, ultrastructural localization, and posttranslational phosphorylation. J Membr Biol. 1990;116:163–75. 5. White TW, Bruzzone R, Goodenough DA, Paul DL. Mouse Cx50, a functional member of the connexin family of gap junction proteins, is the lens fiber protein MP70. Mol Biol Cell. 1992;3:711–20. 6. Mathias RT, Rae JL, Baldo GJ. Physiological properties of the normal lens. Physiol Rev. 1997;77:21–50. 7. Gao J, Sun X, Yatsula V, Wymore RS, Mathias RT. Isoform-specific function and distribution of Na/K pumps in the frog lens epithelium. J Membr Biol. 2000;178:89–101. 8. Candia OA, Zamudio AC. Regional distribution of the Naþ and Kþ currents around the crystalline lens of rabbit. Am J Physiol Cell Physiol. 2002;282:C252–62. 9. Tamiya S, Dean WL, Paterson CA, Delamere NA. Regional distribution of Na,K-ATPase activity in porcine lens epithelium. Invest Ophthalmol Vis Sci. 2003;44:4395–99. 10. Donaldson P, Kistler J, Mathias RT. Molecular solutions to mammalian lens transparency. News Physiol Sci. 2001;16:118–23. 11. Gao J, Sun X, Martı´ nez-Wittinghan FJ, Gong X, White TW, Mathias RT. Connections between connexins, calcium, and cataracts in the lens. J Gen Physiol. 2004;124:289–300. 12. Reaume AG, de Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science 1995;267:1831–34. 13. Gao Y, Spray DC. Structural changes in lenses of mice lacking the gap junction protein connexin43. Invest Ophthalmol Vis Sci. 1998;39:1198–209. 14. White TW, Sellitto C, Paul DL, Goodenough DA. Prenatal lens development in connexin43 and connexin50 double knockout mice. Invest Ophthalmol Vis Sci. 2001;42:2916–23. 15. Baruch A, Greenbaum D, Levy ET, Nielsen PA, Gilula NB, Kumar NM, Bogyo M. Defining a link between gap junction communication, proteolysis, and cataract formation. J Biol Chem. 2001;276:28999–9006. 16. Rong P, Wang X, Niesman I, Wu Y, Benedetti LE, Dunia I, Levy E, Gong X. Disruption of Gja8 (a8 connexin) in mice leads to microphthalmia associated with retardation of lens growth and lens fiber maturation. Development 2002;129:167–74. 17. White TW, Goodenough DA, Paul DL. Targeted ablation of connexin50 in mice results in microphthalmia and zonular pulverulent cataracts. J Cell Biol. 1998;143:815–25. 18. White TW. Unique and redundant connexin contributions to lens development. Science 2002;295:319–20. 19. Xia CH, Cheng C, Huang Q, Cheung D, Li L, Dunia I, Benedetti LE, Horwitz J, Gong X. Absence of a3 (Cx46) and a8 (Cx50) connexins leads to cataracts by affecting lens inner fiber cells. Exp Eye Res. 2006;83:688–96. 20. Tang Y, Liu X, Zoltoski RK, Novak LA, Herrera RA, Richard I, Kuzack JR, Kumar NM. Calpain 3 disruption delays cataract formation in the 129 Cx46KO mouse lens. Invest Ophthalmol Vis Sci. 2006;47:4107. 21. Sellitto C, Li L, White TW. Connexin50 is essential for normal postnatal lens cell proliferation. Invest Ophthalmol Vis Sci. 2004;45:3196–202. 22. Tang Y, Zoltoski RK, Kuszak JR, Kumar NM. Disruption of calpain 3 gene prevents cataracts, but doesn’t correct lens growth deficiency in Cx50-knockout mice. Invest Ophthalmol Vis Sci. 2007;48:3636. 23. Lang RA. Pathways regulating lens induction in the mouse. Int J Dev Biol. 2004;48:783–91.

18 Lens

395

24. Singh DP, Ohguro N, Chylack LT Jr, Shinohara T. Lens epithelium-derived growth factor: increased resistance to thermal and oxidative stresses. Invest Ophthalmol Vis Sci. 1999;40:1444–51. 25. Lin D, Zhou J, Zelenka PS, Takemoto DJ. Protein kinase Cg regulation of gap junction activity through caveolin-1-containing lipid rafts. Invest Ophthalmol Vis Sci. 2003;44:5259–68. 26. Hossain MZ, Jagdale AB, Ao P, Kazlauskas A, Boynton AL. Disruption of gap junctional communication by the platelet-derived growth factor is mediated via multiple signaling pathways. J Biol Chem. 1999;274:10489–96. 27. Lampe PD, Lau AF. Regulation of gap junctions by phosphorylation of connexins. Arch Biochem Biophys. 2000;384:205–15. 28. Berthoud VM, Westphale EM, Grigoryeva A, Beyer EC. PKC isoenzymes in the chicken lens and TPA-induced effects on intercellular communication. Invest Ophthalmol Vis Sci. 2000;41:850–8. 29. Saleh SM, Takemoto DJ. Overexpression of protein kinase Cg inhibits gap junctional intercellular communication in the lens epithelial cells. Exp Eye Res. 2000;71:99–102. 30. Sa´ez JC, Martı´ nez AD, Bran˜es MC, Gonzalez HE. Regulation of gap junctions by protein phosphorylation. Braz J Med Biol Res. 1998;31:593–600. 31. Nguyen TA, Boyle DL, Wagner LM, Shinohara T, Takemoto DJ. LEDGF activation of PKCg and gap junction disassembly in lens epithelial cells. Exp Eye Res. 2003;76:565–72. 32. Lin D, Boyle DL, Takemoto DJ. IGF-I-induced phosphorylation of connexin 43 by PKCg: regulation of gap junctions in rabbit lens epithelial cells. Invest Ophthalmol Vis Sci. 2003;44:1160–8. 33. Wagner LM, Saleh SM, Boyle DJ, Takemoto DJ. Effect of protein kinase Cg on gap junction disassembly in lens epithelial cells and retinal cells in culture. Mol Vis. 2002;8:59–66. 34. Le AC, Musil LS. A novel role for FGF and extracellular signal-regulated kinase in gap junction-mediated intercellular communication in the lens. J Cell Biol. 2001;154: 197–216. 35. Baldo GJ, Mathias RT. Spatial variations in membrane properties in the intact rat lens. Biophys J. 1992;63:518–29. 36. Shiels A, Mackay D, Ionides A, Berry V, Moore A, Bhattacharya S. A missense mutation in the human connexin50 gene (GJA8) underlies autosomal dominant ‘zonular pulverulent’ cataract, on chromosome 1q. Am J Hum Genet. 1998;62:526–32. 37. Mackay D, Ionides A, Kibar Z, Rouleau G, Berry V, Moore A, Shiels A, Bhattacharya. Connexin46 mutations in autosomal dominant congenital cataract. Am J Hum Genet. 1999;64:1357–64. 38. Renwick JH, Lawler SD. Probable linkage between a congenital cataract locus and the Duffy blood group locus. Ann Hum Genet. 1963;27:67–84. 39. Church RL, Wang JH, Steele E. The human lens intrinsic membrane protein MP70 (Cx50) gene: clonal analysis and chromosome mapping. Curr Eye Res. 1995;14:215–21. 40. Berry V, Mackay D, Khaliq S, Francis PJ, Hameed A, Anwar K, Mehdi SQ, Newbold RJ, Ionides A, Shiels A, Moore T, Bhattacharya SS. Connexin 50 mutation in a family with congenital ‘zonular nuclear’ pulverulent cataract of Pakistani origin. Hum Genet. 1999;105:168–70. 41. Willoughby CE, Arab S, Gandhi R, Zeinali S, Arab S, Luk D, Billingsley G, Munier FL, Heon E. A novel GJA8 mutation in an Iranian family with progressive autosomal dominant congenital nuclear cataract. J Med Genet. 2003;40:e124. 42. Polyakov AV, Shagina IA, Khlebnikova OV, Evgrafov OV. Mutation in the connexin 50 gene (GJA8) in a Russian family with zonular pulverulent cataract. Clin Genet. 2001;60:476–8.

396

T.I. Shakespeare et al.

43. Rees MI, Watts P, Fenton I, Clarke A, Snell RG, Owen MJ, Gray J. Further evidence of autosomal dominant congenital zonular pulverulent cataracts linked to 13q11 (CZP3) and a novel mutation in connexin 46 (GJA3). Hum Genet. 2000;106:206–9. 44. Jiang H, Jin Y, Bu L, Zhang W, Liu J, Cui B, Kong X, Hu L. A novel mutation in GJA3 (connexin46) for autosomal dominant congenital nuclear pulverulent cataract. Mol Vis. 2003;9:579–83. 45. Bennett TM, Mackay DS, Knopf HL, Shiels A. A novel missense mutation in the gene for gap-junction protein a3 (GJA3) associated with autosomal dominant ‘nuclear punctate’ cataracts linked to chromosome 13q. Mol Vis. 2004;10:376–82. 46. Li Y, Wang J, Dong B, Man H. A novel connexin46 (GJA3) mutation in autosomal dominant congenital nuclear pulverulent cataract. Mol Vis. 2004;10:668–71. 47. Chang B, Wang X, Hawes NL, Ojakian R, Davisson MT, Lo WK, Gong X. A Gja8 (Cx50) point mutation causes an alteration of a3 connexin (Cx46) in semi-dominant cataracts of Lop10 mice. Hum Mol Genet. 2002;11:507–13. 48. Xia CH, Cheung D, DeRosa AM, Chang B, Lo WK, White TW, Gong X. Knock-in of a3 connexin prevents severe cataracts caused by an a8 point mutation. J Cell Sci. 2006;119:2138–44. 49. Steele EC Jr, Lyon MF, Favor J, Guillot PV, Boyd Y, Church RL. A mutation in the connexin 50 (Cx50) gene is a candidate for the No2 mouse cataract. Curr Eye Res. 1998;17:883–9. 50. Graw J, Loster J, Soewarto D, Fuchs H, Meyer B, Reis A, Wolf E, Balling R, Hrabe de Ang. Characterization of a mutation in the lens-specific MP70 encoding gene of the mouse leading to a dominant cataract. Exp Eye Res. 2001;73:867–76.

Chapter 19

Connexins in the Mammalian Retina Stephen C. Massey

Abstract Gap junctions are particularly numerous in the retina and found in every major retinal cell type. They provide the primary connections in certain retinal pathways and form the substrate for signal averaging in others. At least four neuronal connexins are found in the mammalian retina, and different cell types express specific connexins with distinct properties. Cones are coupled via Cx36 gap junctions, which are thought to improve the cone signal-to-noise ratio. In contrast, rod-cone coupling, perhaps via heterotypic gap junctions, forms the basis for an alternate processing pathway that is active at intermediate light intensities. Horizontal cells are extensively coupled to provide spatial averaging over a wide area. Modulation of these junctions changes the spatial profile of horizontal cell feedback to photoreceptors. In rabbit and cat retina, A-type horizontal cells are coupled via massive Cx50 gap junctions, whereas the axon-bearing or B-type horizontal cells of the mouse and rabbit retina have different coupling properties and may express Cx57. In the inner retina, the primary output of the rod bipolar cell is to AII amacrine cells, which form a well-coupled network via Cx36 gap junctions. The AII network is prominent in all mammalian retinas and seems to provide for signal averaging in the noisy high-gain rod pathway. In addition, the signaling pathway from AII amacrine cells to ON cone bipolar cells passes via gap junctions, some of which may be heterotypic Cx36–Cx45 gap junctions. Many other types of amacrine cells and ganglion cells are also coupled via Cx36 or Cx45 gap junctions. Ganglion cell coupling produces synchronized spike activity between neighboring cells of the same type. The prevalence of gap junctions in the retina may occur because signal averaging and noise reduction are important strategies in the early stages of visual processing. Finally, while several retinal connexins have now been described, the connexins expressed in some coupled cells have not yet been identified. This suggests there are additional neuronal connexins (or perhaps pannexins) still to be identified in the mammalian retina. S.C. Massey (*) Ophthalmology and Visual Science, University of Texas Medical School at Houston, 6431 Fannin, Houston, TX 77030, United States e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_19, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

397

398

S.C. Massey

Keywords Retina Photoreceptors Rod pathway Amacrine cell Ganglion cells Cx36 Cx45 Cx50 Cx55.5 Cx57

19.1 Introduction We live in a sea of photons and, for this reason, vision is almost ubiquitous among animals. Several engineering problems must be solved to make vision a useful and effective means of sensory input. One particularly desirable attribute is the ability to see in the dark, in the near absence of photons. Undoubtedly, in a world of prey and predators, nighttime vision has a high survival value. For a rod, the evolutionary conclusion and physical limit of this survival-driven race for visual sensitivity is the ability to detect a single photon. To achieve this level of sensitivity, rods contain millions of copies of rhodopsin to maximize the chance of capturing a single photon. Nevertheless, singlephoton responses in rods generate very small electrical signals against a background of noise produced by the phototransduction machinery. Later in the rod pathway, there is a second stage of high amplification followed by signal averaging in a coupled network of AII amacrine cells. In this system, the random noise in the rod pathway tends to cancel out. On the other hand, the light-driven signals are correlated and therefore additive. The fundamental properties of visual stimuli reinforce this capability. Most objects of visual interest stimulate multiple photoreceptors simultaneously, generating correlated responses that preferentially pass through the network. It may be inferred that one reason for the abundance of gap junctions in the retina is that signal averaging and noise reduction are important strategies in the early steps of visual processing. A second engineering problem arises from the vast range of light intensities in everyday life. The stimulus may vary from pitch black or starlight to bright open sunlight, an intensity range of more than ten log units. While the average visual scene has a range of intensities covering only two to three log units, the continual adaptation of retinal sensitivity slides this operating range through the entire range of light intensities. This is a critical function of the retina because, outside the normal operating range, we are functionally blind. Common examples include the blindness experienced when entering a dark cinema or driving into the setting sun. The intensity/saturation problem has been solved by several distinct mechanisms. An inherent part of the response to light is to modulate the gain of the phototransduction cascade, a process known as light adaptation. In addition, there is a supplementary gap junction pathway between rods and cones that is thought to operate at intermediate light intensities. Gap junctions are versatile circuit components that can be modulated quickly by adjustment of channel activity. While chemical neurotransmission is also adaptable, there is direct evidence that light and dopamine modulate key gap junction connections in the retina. Gap junctions may also be controlled by circadian mechanisms, which preset retinal circuits according to the time of day or night. Hence, in a

19 Retina

399

self-optimizing neural system such as the retina, gap junctions play an important role in adjusting local circuit parameters. Finally, in addition to addressing these engineering problems, gap junctions in the retina make primary connections between neurons. This means that gap junctions function as electrical synapses in the retina and are required in certain retinal pathways. They do not just modulate the circuits; gap junctions transfer the visual signal in certain retinal pathways and, if the connexin responsible is knocked out, that pathway is missing. Examples are rod-cone coupling and the gap junctions between AII amacrine cells and ON cone bipolar cells.

19.2 A Retinal Sandwich The retina is a layered structure composed of approximately 60 different types of neurons (Fig. 19.1) [1,2]. The somas of photoreceptors, 95% rods and 5% cones, are located in the outer nuclear layer (ONL), and there are separate rod and cone pathways throughout the retina. The inner nuclear layer (INL) contains the somas of horizontal cells, bipolar cells, and amacrine cells. Finally, the ganglion cell layer (GCL) contains ganglion cells, the output neurons of the retina, and some displaced amacrine cells. The three somatic layers sandwich two plexiform or dendritic layers, where most of the dendrites and synaptic interactions, both chemical and electrical, occur.

Fig. 19.1 Vertical section of the primate retina showing the layered structure. In the outer plexiform layer, photoreceptors (PR; rods and cones) contact horizontal cells (HC), and bipolar cells (BC). In the inner plexiform layer, bipolar cells (BC) contact amacrine cells (AC) and ganglion cells (GC). ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. (A high-resolution version of this figure is available on accompanying CD and online at www. springerlink.com)

400

S.C. Massey

The outer plexiform layer (OPL) contains connections between photoreceptors, which release glutamate at ribbon synapses, and the second order neurons, that is, horizontal cells and bipolar cells. Most mammalian species (except rodents) have two kinds of horizontal cell. They are large, laterally extensive neurons connected by gap junctions. The role of horizontal cells is to provide a negative feedback signal to photoreceptors. Bipolar cells are more diverse and transmit signals vertically from the outer retina to the inner retina where they make synaptic contact with amacrine cells and ganglion cells. There are approximately ten kinds of cone bipolar cell, evenly split between ON bipolar cells and OFF bipolar cells [3,4,5]. The opposite responses of ON and OFF cone bipolar cells are mediated by different glutamate receptors such that OFF bipolar cells have -amino-3-hydroxy-5methyl-4-isoxazolepropionic acid (AMPA)/kainate receptors [6] while ON bipolar cells express the sign-inverting metabotropic receptor mGluR6 [7]. The split of the cone signal into separate ON and OFF pathways is fundamental. This division persists throughout the retina and higher visual centers. The layered structure of the retina is also present in the inner plexiform layer (IPL): OFF bipolar cells stratify in the top half of the IPL, sublamina a, while ON bipolar cells project to the bottom half of the IPL, sublamina b (Figs. 19.1 and 19.2). The stratification of the IPL appears to function as a simple addressing system to determine which third-order neurons (i.e., amacrine and ganglion cells) are in synaptic contact with the incoming bipolar cell terminals. In turn, the cone bipolar cells make synaptic connections with amacrine cells and ganglion cells, OFF cells in sublamina a, ON cells in sublamina b. Ganglion cells are the output neurons of the retina. Many retinal cells including photoreceptors, horizontal cells, and bipolar cells respond with slow graded potentials but the ganglion cells generate full-blown 100 mV action potentials for transmission down the optic nerve. There are approximately 15 different types of ganglion cells [8] that are thought to produce independent channels in the visual system, each tuned to different aspects of the visual input such as contrast, acuity, color, movement, and temporal frequency. Many project to different regions of higher visual areas. Thus, the retina is thought to transmit twelve to 15 different aspects of the visual input. Ganglion cell axons run in the nerve fiber layer, close to the vitreal surface of the retina, and unite to form the optic nerve, which carries all information between the retina and the brain. Many ganglion cell subtypes are coupled via gap junctions, or sometimes via intermediary amacrine cells, but never between different ganglion cell types.

19.3 The High-Gain Rod Pathway Depending on the species, 30 to 100 rods are connected to each rod bipolar cell, the most numerous of all bipolar cell types, accounting for a third of the total. The high convergence contributes to the sensitivity of the rod pathway [9]. Rod

19 Retina

401

bipolar cells express mGluR6 receptors and give ON responses to light. They project to the bottom layer of the IPL (sublamina b) and thus conform to the functional stratification of the IPL. In contrast to cone bipolar cells, rod bipolar cells rarely contact ganglion cells directly. Instead, they make synaptic contact with an intermediary amacrine cell, the AII or rod amacrine cell, so called because of its obligatory role in the rod pathway (Fig. 19.2) [10,11]. In essence, the small, low threshold rod responses require high amplification. This produces a noisy output that is then filtered via the coupled network of AII amacrine cells so that correlated light-driven signals summate while the asynchronous noise tends to cancel. AII amacrine cells use glycine as a neurotransmitter and they make conventional inhibitory glycinergic synapses with OFF cone bipolar terminals in the top layer of the IPL (sublamina a). This is the route for rod-driven OFF signals to enter the cone pathways (Fig. 19.2). In contrast, AII amacrine cells make gap junctions with ON cone bipolar cells that serve as sign-conserving synapses [12].

Fig. 19.2 Schematic showing rod and cone pathways through the mammalian retina. Cone–cone (C) coupling is mediated by Cx36 gap junctions. The connexin expressed by rods is unknown. Cone signals pass directly via ON and OFF cone bipolar cells (CB), which ramify at different depths in the inner plexiform layer to contact ON and OFF ganglion cells, respectively. In contrast, rod signals pass via a single type of rod bipolar cell (RB), which synapses with AII amacrine cells. AII amacrine cells are connected via Cx36 gap junctions into a wellcoupled network. AII amacrine cells make inhibitory glycinergic synapses with OFF cone bipolar cells and sign-conserving electrical synapses with ON cone bipolar cells via heterotypic Cx36–Cx45 gap junctions. By these two pathways (shaded), rod signals enter the cone-driven circuits of the retina. Horizontal cells (HCs) are large, laterally extensive neurons that provide inhibitory feedback to photoreceptors in the OPL. A-type HCs (AH) are extensively coupled via Cx50 gap junctions. B-type HCs (BH) express Cx57. The B-type axon terminals (BHATs) contact rods and may be coupled via Cx57 gap junctions. (A high-resolution version of this figure is available on accompanying CD and online at www.springerlink.com)

402

S.C. Massey

Thus, the connections with ON bipolar terminals have the opposite sign compared to the glycinergic input to OFF bipolar cells. While cone pathways split at the first retinal synapses in the OPL, due to the expression of different glutamate receptors by ON and OFF cone bipolar cells, the rod pathways bifurcate in the IPL, due to the chemical and electrical connections made by AII amacrine cells [9]. In this way, the rod signals are said to piggy-back on the cone pathways (Fig. 19.2).

19.4 Retinal Connexins In terms of retinal circuitry, Cx36 is the dominant neuronal connexin of four (Cx36, Cx45, Cx50, Cx57) in the mammalian retina [13]. Cx36 is used for cone–cone coupling and in the AII amacrine cell network [14,15,16]. These are both good examples of a signal-averaging network. Horizontal cells were the first described and perhaps the best known coupled cells in the retina. In species where there are two kinds of horizontal cell with different coupling properties, each type is served by a different connexin. In the rabbit retina, axonless A-type horizontal cells are coupled via Cx50 gap junctions [17]. Some of these are assembled into giant plaques, which may contain hundreds of thousands of gap junction channels. In contrast, axon-bearing horizontal cells, like the B-type of the rabbit retina and those found in the mouse retina, may express Cx57 [18]. In the inner retina, certain amacrine and ganglion cell types are coupled by either Cx36 or Cx45 gap junctions. It has also been proposed that heterologous gap junctions between AII amacrine cells and ON cone bipolar cells utilize heterotypic Cx36–Cx45 gap junctions [19]. This provides a good explanation for the different properties of AII-bipolar gap junctions compared with AII-AII gap junctions [20]. Retinal locations of the different neuronal connexins are summarized in Fig. 19.2. In fish retina, a number of novel connexins have been identified, some of which are retina-specific [21,22]. Finally, the use of phosphospecific antibodies suggests that connexin phosphorylation may modulate coupling in some retinal networks [23].

19.5 Photoreceptor Coupling 19.5.1 Cone–Cone Coupling Most Cx36 gap junctions are found in the inner retina where Cx36 is particularly prominent in the AII amacrine cell network. However, Cx36 labeling is also present in the OPL. In the first reports, the Cx36 antibodies had a relatively low affinity and the OPL was not well labeled. However, use of improved antibodies revealed that Cx36 gap junctions connect many cone pedicles in the OPL [14].

19 Retina

403

The cell bodies of cones are located in the top row of the ONL. From here cone axons descend through the massed ranks of rod somas to form synaptic terminals, known as cone pedicles, in the OPL. Cone pedicles contain synaptic machinery such as synaptic ribbons, synaptic vesicles, and calcium channels required to release glutamate onto the dendrites of horizontal cells and bipolar cells, whose dendrites gather at each cone pedicle. These are complex synaptic structures; each pedicle may contain 20 to 40 synaptic ribbons and there may be several hundred postsynaptic processes from two kinds of horizontal cell and approximately ten cone bipolar cells [24]. In addition, there are six to ten fine processes, known as telodendria, which extend laterally from each cone pedicle to reach the adjacent members of the cone mosaic [25]. There are Cx36 plaques precisely at the contact points between telodendria and the ring of adjacent cone pedicles. This kind of pattern does not arise by chance. When Cx36 immunolabeling is observed exactly where two processes touch, there is a high degree of confidence that these are gap junction sites. Thus, Cx36 gap junctions are responsible for cone–cone coupling, and the network of telodendria forms the substrate for this coupling. This nicely confirms earlier ultrastructural studies that showed cone–cone gap junctions in freeze-fracture preparations or electron microscopy (EM) sections [26,27,28]. Recent physiological work has shown that adjacent cones are dye-coupled and electrically connected. Technically impressive recordings from pairs of cones have shown a bidirectional conductance of a few hundred picoSiemens (pS) [29,30,31]. This is consistent with the presence of ten to 100 junctional Cx36 channels. At first, the idea that cones should be connected by gap junctions seems counterintuitive, as it would seem to degrade visual acuity. However, the density of the cone mosaic slightly exceeds the resolution limits of the optics of the eye. It is the spherical aberrations in the lens and cornea that impose a limit on the maximum resolution. Phototransduction is inherently noisy, and calculations suggest that cone–cone coupling reduces the random noise while preserving correlated light driven signals. Thus, it has been estimated that cone–cone coupling increases the signal-to-noise ratio by as much as 70% with a very minor loss of resolution, minimized by the small excess of cones. The net result is a substantial increase in the high-frequency response curve [30].

19.5.2 Blue Cones Most mammals have only two kinds of cones, green and blue. Primates, including humans, are trichromatic with red, green, and blue cones. In both cases, blue cones account for a small fraction of the cone population, around 10%, and they have subtle morphological differences including a smaller cone pedicle [32]. In contrast to the coupling between green cones in the ground

404

S.C. Massey

squirrel (used experimentally for its large cones) or red/green cones in the primate retina, blue cones are not coupled to the ring of adjacent cones. Of course, because blue cones account for such a minority of cones, essentially all their neighbors are green or red/green. There are three lines of evidence for the lack of coupling between blue cones and their neighbors: First, blue cones are not dye-coupled to the surrounding cones. Injecting neurobiotin into green cones fills a small patch except for blue cones [33]. Second, most of the blue cone telodendria are too short to reach the adjacent cone pedicles [32]. Third, recordings from pairs of cones where one cone is blue show no evidence of electrical coupling [33]. One explanation for the apparent lack of blue cone coupling is that the penalty for color discrimination would be too high. The absorption curve for blue cones is distant from the red/green curves, so electrical coupling would lead to mixing of the color pathways. In contrast, the red/green absorption curves are close together. This means that red and green cones have similar responses to a given stimulus and thus their coupling improves the signal-to-noise ratio at a minimum expense of color discrimination, whereas blue cone coupling would decrease it substantially. Calculations for the primate retina suggest that red/ green coupling produces an increase in luminance or brightness detection, due to the decreased noise, while there is a minor decrease in color discrimination [31,34].

19.5.3 Rod-Cone Coupling In addition to cone–cone coupling, there is ultrastructural evidence for small gap junctions between rods and cones [26,35]. Furthermore, electrical responses with the physiological signature of rod input have been recorded in cones and in second-order neurons such as horizontal cells that are exclusively connected to cones [36,37,38,39]. Neurobiotin injections into primate cones revealed more than 20 coupled rods, in addition to the coupled cones. Finally, in immunolabeled material, there are many small Cx36 plaques surrounding the perimeter of cone pedicles, many on fine telodendria that are too short to reach neighboring cones. These appear to make gap junctions with neighboring rod synaptic terminals (known as spherules), which outnumber the cones by 20:1, and thus surround every cone pedicle. As mentioned above, cones express Cx36 and the contacts between rods and cones are immunolabeled for Cx36. However, the connexin expressed by the rods has not been identified. In Cx36 knockout (KO) mice, a -galactosidase reporter was heavily expressed in the outer nuclear layer [40]. Due to the number of rods relative to the number of cones, this was taken to indicate that rods express Cx36. However, in later experiments, Cx36 labeling was seen to be restricted to cones [14]. Furthermore, immunolabeled EM material showed that gap junctions between rods and cones were labeled for Cx36 but only on the

19 Retina

405

cone side of the gap junction [41]. Thus, it appears that rod–cone gap junctions are heterotypic. Rods apparently do not express Cx36. Cx50 and Cx57 appear to be restricted to horizontal cells and Cx45 is not found in the OPL. Thus, none of the four known neuronal connexins seem to be present in rods. Therefore, an unidentified connexin may be expressed by rods. It is common to think of rod and cone pathways through the retina as separate circuits that only merge in the inner retina. However, this is an oversimplification. In fact, rod–cone coupling provides an alternative pathway for rod signals to enter the cone pathways at intermediate light intensities [35]. In this pathway, gap junctions form the primary connections between rods and cones. To understand the function of rod–cone coupling, it is necessary to consider the different intensity ranges detected by rods and cones. The sensitivity threshold for a single rod is the absorption of one photon. However, as many as 100 rods converge to each rod bipolar cell and 100 rod bipolar cells may converge to a single ganglion cell. The ultimate visual threshold is determined by the product of these two convergence numbers and is usually given as one photon per 10,000 rods for an additional four log units of sensitivity. These small threshold signals are then amplified via a high-gain pathway through the rod bipolar cell and the AII network. At these low light levels, where only a few rods absorb a photon, the gap junctions between rods and cones will provide almost no signal to the cone. However, the high gain of this pathway is a mixed blessing: at around 2.5 photons/rod/second the rod bipolar cell becomes saturated [42]. Yet, this is still substantially below the cone threshold, which is approximately 30 photons/second [40]. However, at this intensity, there are enough photons to produce a response in every rod, and because rods outnumber cones by a factor of 20 and nearly every rod is coupled to a nearby cone, transmission via rod–cone gap junctions will produce a rod-driven response in the cone pedicle. In this way, rod–cone coupling is thought to provide an alternative pathway at intermediate light intensities, between rod bipolar saturation and cone threshold. It should be noted that this pathway is entirely dependent on gap junctions. Finally, it has long been proposed that this gap junction pathway is modulated by light so that rod–cone gap junctions are open at low light intensity and closed as the intensity increases and cones take over [35]. This would be a good example of gap junction plasticity whereby a neural circuit is optimized for the prevailing conditions.

19.5.4 Rod–Rod Coupling There are small gap junctions between rods in the mouse retina and, most recently, neurobiotin coupling has been reported between rods in the primate retina [27,36]. Rods are also coupled in other vertebrate species such as salamander [43]. As before, coupling between rods will decrease the uncorrelated noise. In addition, rod coupling will tend to spread the light-driven signal within

406

S.C. Massey

small groups of rods and, by delaying saturation in the rod bipolar pathway, increase the range of the rod-driven pathway. However, there is a snag: due to convergence, the rod bipolar cell threshold occurs when approximately one in 100 rods absorbs a photon. If a single rod absorbs a single photon but it is coupled to a surrounding patch of rods that receive none, then the threshold sensitivity will be reduced [36]. However, if rod coupling can be modulated, then perhaps the gap junctions would close under these conditions to maximize the threshold response in a single rod. As more light is available, averaging between rods via open gap junctions would then be useful to extend the range of the rod pathway by preventing saturation of the rod bipolar response. There is no evidence yet to support this idea but it serves as a good example of how gap junction plasticity could provide the best of both worlds in retinal circuits.

19.6 Horizontal Cell Coupling Horizontal cells are laterally extensive second-order neurons that provide a negative feedback signal to photoreceptors and thus play a critical role in the early steps of visual processing [1]. Like many sensory systems, vision starts with center/surround activity. While a small spot generates a strong response, stimulating the surrounding area inhibits the center response. Thus, center responses have an antagonistic surround that serves to sharpen the center response. The fundamental reason for center surround antagonism is that the average intensity of the visual input is of little value. Consider viewing this page on a computer screen or its reading as a page of text. The average intensity of the page has little value except to optimize the operating range of the retina. The real information lies in the contrast between the individual words or letters and the background. The visual system is designed to extract the details carried by local contrast. In large part, this is accomplished by horizontal cell feedback. Horizontal cells are large neurons but gap junction coupling dramatically enhances the size of the receptive field. Thus, horizontal cells carry a large, slow, average of the visual input, and modulation of the coupling will change the spatial properties of the feedback signal. In contrast, photoreceptors respond to local changes in contrast without regard to the average intensity. In some successful computer models of the retina, the bipolar cells transmit a signal derived from the difference between the local photoreceptor response and the wide-field response of horizontal cells [44]. The extent of the antagonistic surround is controlled in part by the strength of coupling in the network of horizontal cells. The mechanism has not been worked out in detail but modulation of horizontal cell coupling may optimize the center/surround balance of retinal circuits for different light levels. In most mammalian species there are two kinds of horizontal cell, but in rodents only one. In the rabbit retina, these are called A-type and B-type horizontal cells and they have dramatically different dye-coupling properties. The axonless A-type horizontal cells contact cones exclusively. They are the

19 Retina

407

Fig. 19.3 A-type horizontal cells are coupled via Cx50 gap junctions. (a) The A-type horizontal cell (HC) network was filled with neurobiotin (red). Cx50 labeling is green but appears yellow because it is colocalized with A-type HCs. The background mosaic of cone pedicles is shown by GluR5 receptor labeling of OFF cone bipolar dendrites where they converge at each cone pedicle. (b) High-affinity view shows many small Cx50 gap junctions between fine dendrites and giant Cx50 plaques between major dendrites of A-type HCs. Scale bars: (a) 20 mm; (b) 10 mm. (A high-resolution color version of this figure is available on accompanying CD and online at www.springerlink.com)

best-coupled cells in the retina and the only retinal neurons to pass Lucifer yellow [45,46,47]. A-type horizontal cells are coupled via Cx50 gap junctions (Fig. 19.3) [17]. These gap junction plaques can be huge, up to 100 mm2, and they must contain hundreds of thousands of gap junction channels. The unitary conductance of Cx50 channels is relatively high, so consequently the coupling resistance of A-type horizontal cells is very low, and it is impossible to voltage clamp these cells. In contrast, B-type horizontal cells in the rabbit retina and the single horizontal cell type of the mouse have axons that expand into a complex terminal structure. The somatic dendrites contact cones while the axon terminal contacts rods [45]. In effect, the axon terminal acts as a rod horizontal cell and presumably provides feedback to rods. Both ends of the axon-bearing horizontal cells are independently coupled [48], so, in fact, there are two separate coupled horizontal cell networks in the mouse OPL and three in the rabbit (A-type, B-type, and B-type axon terminal). B-type horizontal cells do not pass Lucifer yellow, nor do they express Cx50. In the mouse retina, the axon-bearing horizontal cells express Cx57 and in the Cx57KO coupling is dramatically reduced [18]. Thus, it seems that the expression of a different connexin, Cx57 rather than Cx50, confers dramatically different dye-coupling properties on axon-bearing horizontal cells. In the rabbit retina, Cx57 labeling was selectively located on the B-type axon terminal structures. This surprising result suggests that Cx57 may be selectively trafficked along the axon and perhaps an unknown

408

S.C. Massey

connexin is responsible for coupling between the somatic dendrites of B-type horizontal cells. While there is general agreement that horizontal cells provide feedback to photoreceptors, the exact synaptic mechanism is unknown. Early work suggesting a role for g-aminobutyric acid (GABA) release has not been confirmed [49], although some of the molecular components required to support vesicular release have been identified in horizontal cell dendrites [50]. In the absence of a conventional synaptic mechanism, two novel mechanisms have been proposed. Horizontal cells are well coupled via gap junctions (see above), and it has been suggested that unapposed hemichannels in the dendritic terminals could provide an ephaptic signal that modulates voltage-dependent Ca2+ channels in the photoreceptor terminal [51]. A change in Ca2+ flux would alter glutamate release from the cone terminal and thus complete a negative feedback loop. In support of this idea, immunolabeling of Cx55.5 has been found on horizontal cell dendrites adjacent to the cone terminals in fish retina [21]. However, the block of horizontal cell feedback by the gap junction antagonist carbenoxolone has also been attributed to a direct effect on Ca2+ channels [52]. A second alternative mechanism suggesting that protons mediate horizontal cell feedback has also been proposed. This is supported by the buffering action of 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES), which blocks feedback as well as the antagonistic surround signals found in ganglion cells that are probably generated, at least in part, by horizontal cell feedback [53]. However, proton-mediated feedback is complicated by another source of protons due to vesicular release from cones [54]. Finally, it is also possible that HEPES may have direct effects on hemichannels. Horizontal cell feedback is important, and although the mechanism remains controversial, it does seem likely that a novel process is involved.

19.7 AII Amacrine Cells–Cone Bipolar Cells: A Complex Heterocellular Network Rods are specialized to detect single photons under very low light conditions. The small signals they produce are amplified by rod bipolar cells, which are connected to AII amacrine cells. In turn, the AII splits the signals into ON and OFF pathways by connecting to OFF bipolar cells via glycinergic synapses and to ON bipolar cells via gap junctions (Fig. 19.2). Four or five different ON cone bipolar cells are dyecoupled to AII amacrine cells so this is truly a complex heterocellular network.

19.7.1 The AII Amacrine Cell Network AII amacrine cells are well coupled by prominent gap junctions that were described in the first morphological reports of this cell type [55]. The extent of coupling was dramatically confirmed by neurobiotin injections, which yield a

19 Retina

409

Fig. 19.4 The heterocellular network of AII amacrine cells and ON cone bipolar cells in the rabbit retina. The three-dimensional (3D) reconstruction from a confocal series to show the heterocellular network. A single AII amacrine cell was filled with neurobiotin. AII amacrine cells lie adjacent to the inner plexiform layer and overlap extensively. ON cone bipolar cells (three to five types) have cell bodies higher in the inner nuclear layer and they produce descending axons that are lost in the matrix of AII dendrites. (A high-resolution version of this figure is available on accompanying CD and online at www.springerlink.com)

mosaic of AII amacrine cells and a large population of overlying ON cone bipolar cells (Fig. 19.4) [56,57]. Cx36 is highly expressed by AII amacrine cells at dendritic crossings in the AII matrix and we have learned that this type of labeling pattern is diagnostic for gap junction proteins (Fig. 19.5) [15,16]. This suggests that coupling in the

Fig. 19.5 Cx36 in the AII amacrine cell network. AII amacrine cells were labeled with an antibody against calretinin (CR; gray). Cx36 labeling (white) is concentrated almost exclusively in the AII matrix. Note how the Cx36 labeling occurs at dendritic crossings in the matrix. This is the signature for gap junction labeling. Scale bar 5 mm. (A high-resolution version of this figure is available on accompanying CD and online at www.springerlink.com)

410

S.C. Massey

AII network is homotypic, via Cx36–Cx36 gap junctions. Paired recordings from adjacent AII amacrine cells show that the gap junctions are electrically bidirectional with a conductance in the range of 700 pS [58]. These recordings are noisy with a lot of spontaneous activity, as might be expected for a system that sits close to threshold so as to impart maximum sensitivity in low light conditions. Importantly, the gap junctions can synchronize action potentials between adjacent AII amacrine cells [58,59]. This is consistent with the idea that gap junctions in the AII network serve the purpose of signal averaging and noise reduction in the high-gain rod bipolar pathway [60,61]. Dopaminergic amacrine cells make a dense plexus of fine dendrites in the IPL that blanket the entire retina. There are prominent rings in the dopamine plexus and, in every ring, there is an AII amacrine cell [62]. This is the morphological signpost of a major synaptic input to AII amacrine cells. Dopamine itself dramatically reduces coupling in the AII network via cAMP-dependent phosphorylation of Cx36 gap junctions [56,63]. In addition, AII coupling is modulated by light, with reduced coupling under very bright or very dark conditions and maximum coupling at intermediate intensities [64]. Dopamine appears to function as a global mediator of light adaptation and a circadian regulator in the retina. Although the details have still to be determined, this seems to be a prime example of gap junction modulation to optimize retinal circuits according to the prevailing light intensity or the time of day.

19.7.2 AII-Bipolar Cell Coupling AII amacrine cells are coupled to ON cone bipolar cells via gap junctions. However, the AII-bipolar gap junctions have different properties than the AII-AII gap junctions. For example, in early electron micrographs, it was noticed that the AII-bipolar gap junctions look asymmetrical [10,65]. When a series of biotinylated tracers was dye injected into AII amacrine cells, the largest members of the series spread through the AII network but were poorly accumulated in the coupled bipolar cells [57,66]. All else being equal, the size exclusion probably indicates that AII–bipolar gap junctions have a narrower pore than AII–AII gap junctions. Furthermore, the two types of gap junction have different pharmacological properties. AII–AII gap junctions are closed by dopamine and cAMP while coupling between AII amacrine cells and ON bipolar cells is reduced by nitric oxide and cyclic guanosine monophosphate (cGMP) analogues [57,64]. AII–AII gap junctions are thought to be homotypic, that is, Cx36–Cx36; however, Cx36 labeling is also found at AII-bipolar contacts [16]. One possible explanation is that AII-bipolar gap junctions are heterotypic (see below). AII amacrine cells are glycinergic interneurons. In contrast, bipolar cells are excitatory neurons using glutamate as the neurotransmitter. However,

19 Retina

411

surprisingly, ON cone bipolar cells also contain glycine. Gap junction antagonists, such as carbenoxolone, prevent this accumulation, which indicates that the source of glycine is by metabolic coupling with AII amacrine cells [67]. In Cx45KO mice, the marker indicating deletion of the gene is expressed in certain bipolar cells. Furthermore, in these mice, the distribution pattern of glycine is changed such that bipolar cells no longer contain glycine [68]. This implicates Cx45 at AII-bipolar gap junctions. In double-labeled material, Cx36 and Cx45 sometimes appear as adjacent pairs located at AII-bipolar contacts [19,69]. This suggests that some AII-ON bipolar gap junctions may be heterotypic with Cx36 on the AII side and Cx45 on the bipolar side. Paired recordings between AII amacrine cells and ON cone bipolar cells show a large bidirectional conductance (1.2 nanoSiemens [59]), and neurobiotin can pass in both directions [70]. While the gap junction itself is electrically linear, there is some functional rectification that makes transmission more effective from AII to bipolar cell, due to the higher input resistance of the bipolar cells [59]. Because AII-bipolar gap junctions are bidirectional, input from the ON cone bipolar cell may also drive AII amacrine cells; this effect is absent in the Cx36KO mouse [71,72]. One usually thinks of AII to bipolar cell transmission, but in bright light, signaling in the opposite direction may mediate ON inhibition of OFF pathways via the AII amacrine cell, an effect known as crossover inhibition. Given the distribution of Cx36, Cx36KO mice might be predicted to have functional deficits in the rod pathways and indeed, this is the case. In an elegant series of experiments, Deans et al. [73] examined the physiological consequences of the Cx36KO. In these animals, neurobiotin injections into AII amacrine cells yield only a single cell because there is no gap junction coupling. Rod-driven ON signals were completely absent in the Cx36KO and ON responses could only be obtained above cone threshold. This indicates that Cx36 gap junctions are essential for the transmission of rod-driven ON signals through the retina. Both rod-cone gap junctions in the OPL and AIIbipolar gap junctions in the IPL are absent in the Cx36KO. However, there is general agreement that the most sensitive rod-driven signals use the high-gain rod bipolar pathway, and these signals are blocked in the knockout due to the absence of AII-bipolar gap junctions. ON responses with a slightly higher threshold may be blocked due to the absence of rod–cone coupling in the Cx36KO. In these animals, rod-driven OFF signals persist because there are no gap junctions along this pathway. Instead, rod-driven OFF signals pass from AII amacrine cells to OFF cone bipolar cells via conventional glycinergic synapses. In the Cx36KO, OFF signals have a reduced sensitivity and this may occur due to the absence of AII–AII coupling [74]. This suggests that signal averaging in the AII network is required to provide maximum sensitivity in rod-driven circuits. Finally, AII coupling is modulated by ambient light to optimize network properties under a variety of different light intensities from starlight, through twilight, to sunlight [74].

412

S.C. Massey

19.8 Ganglion Cell Coupling Many, if not most, ganglion cell types are coupled, either directly or via intermediary amacrine cells. For example, ganglion cells in many species are dyecoupled, and Cx36 is located at the dendritic crossings between certain wide field amacrine cells and the ganglion cell dendrites [75]. Again, this labeling pattern, at dendritic crossings, is diagnostic for gap junctions. It should be stressed that there is no coupling between different ganglion cell types. Coupling occurs exclusively between adjacent ganglion cells belonging to the same class. When there are intermediary amacrine cells, these must also be dedicated to a specific ganglion cell type. Coupling between certain bistratified ganglion cells is eliminated in the Cx45KO mouse, and this implies that some ganglion cells are coupled via Cx45 gap junctions [76]. Convincingly, many of the ganglion cell types that are dye-coupled also have a biochemical signature that indicates gap junction coupling. As noted above for AII-bipolar coupling, gap junctions can readily pass small molecules such as the inhibitory neurotransmitters GABA and glycine, which are commonly found in amacrine cells. Ganglion cells are excitatory neurons that use glutamate, so the presence of inhibitory transmitters indicates gap junction coupling to amacrine cells [77]. Furthermore, extracellular recording with a multielectrode array, which can sample hundreds of ganglion cells simultaneously, showed synchronous firing occurred in ganglion cell types that are coupled [78]. Correlated firing was originally proposed as a mechanism to carry additional information as a multineuronal code [79]. However, an alternative explanation is that synchronized firing is more effective at passing the visual signal to higher centers. Two closely timed spikes from adjacent ganglion cells have a much higher probability of producing a postsynaptic response in the lateral geniculate nucleus [80]. This is really another form of signal averaging because correlated signals produce a strong response while uncorrelated asynchronous noise is more likely to fail. As an example, adjacent ON directionally selective ganglion cells in the rabbit retina, which are well coupled via intermediary amacrine cells, generate synchronized spikes to stimulation in the preferred direction [81]. In the null direction, these cells produce fewer spikes that are desynchronized. ON directionally selective ganglion cells do not transmit a high-resolution image. Rather, they project to the accessory optic system where they initiate an optokinetic response of eye movements concerned with image tracking and stabilization. The synchronized spikes may form a more effective way to signal the direction of stimulus motion to the brain [81].

19.9 Coupling in Retinal Glial Cells While the primary focus of this review has been on neuronal circuits, there is also evidence for glial cell coupling in the retina. The primary glial cell types in the retina are vertically oriented, columnar cells known as Mu¨ller cells. Their

19 Retina

413

overlapping endfeet contain high levels of K+ channels and form the inner limiting membrane, sealing the retina from the vitreous humor filling the eye. Fine extensions of Mu¨ller cell processes envelop neuronal dendrites throughout the retina. Mu¨ller cells are intimately involved in clearing extracellular neurotransmitters, such as glutamate, and siphoning excess K+, generated by neuronal activity, to the vitreous space [82]. Mu¨ller cells are coupled and this may improve their spatial buffering properties [83]. Astrocytes are also located in the nerve fiber layer of the retina, adjacent to the ganglion cell layer. There is strong coupling between astrocytes, less between astrocyte–Mu¨ller cell pairs, and weak coupling between Mu¨ller cell pairs [84]. Dye-coupling between these cell types may appear asymmetrical with tracer passing from astrocytes to Mu¨ller cell but not in the reverse direction [85], perhaps due to differences in coupling strength or in kinetic factors. Ca2+ waves pass through this coupled network but are mediated by ATP release rather than gap junction coupling. Cx30 has been associated with astrocytes, while Cx43 is expressed by Mu¨ller cells [86,87]. These isoforms add to the complement of connexins expressed in the retina. Finally, it should be noted that pericytes, which surround retinal blood vessels, are also coupled via gap junctions [88].

19.10 Conclusion The retina is richly endowed with a wide variety of neuronal gap junctions. In many cases, the primary function of electrical coupling seems to be signal averaging. This may take the form of reinforcing correlated signals and reducing asynchronous noise (photoreceptors, AII amacrine cells), spatial averaging (horizontal cells), or synchronizing spiking cells (ganglion cells). Fundamentally, these are all variations of the same basic mechanism of electrical coupling via gap junctions. In addition, gap junctions are used as primary components to make direct connections in some retinal circuits. In the example of rod–cone coupling, this function includes averaging between the more numerous rods. AII-bipolar cell coupling provides a sign conserving input of opposite sign to the pathway served by glycine, an inhibitory neurotransmitter. The high level of circuit analysis, including the identification of many distinct cell types and their specific connections, makes the retina a particularly suitable preparation to investigate the functions of neuronal gap junctions. We can look forward to further progress as new connexins and pannexins are identified. In particular, the plasticity of gap junctions provides the potential to tune specific retinal circuits during light or circadian adaptation. Uncovering the intracellular and extracellular pathways which modulate gap junction permeability will be essential to understand how retinal sensitivity is optimized over such a wide range of different conditions. Acknowledgments This research was supported by the National Eye Institute (NEI) grants EY 06515 (to SCM) and EY 10608 (Vision Core Grant). Additional support was provided by an unrestricted grant from Research to Prevent Blindness to the Department of

414

S.C. Massey

Ophthalmology and Visual Science. SCM is the Elizabeth Morford Professor of Ophthalmology and Visual Science. Thanks to past and present members of the Massey lab who contributed time and figures, including Feng Pan, Jennifer O’Brien (Fig. 19.3), In-Beom Kim, Wei Li, and Brady Trexler (Fig. 19.4). Thanks to Steve Mills and John O’Brien for many thoughtful discussions. The author regrets that much high-quality work in this area could not be included or referenced due to space limitations.

References 1. Wassle H. Parallel processing in the mammalian retina. Nat Rev Neurosci. 2004; 5:747–57. 2. Shlens J, Field GD, Gauthier JL, Grivich MI, Petrusca D, Sher A, Litke AM, Chichilnisky EJ. The structure of multi-neuron firing patterns in primate retina. J Neurosci. 2006;26:8254–66. 3. MacNeil MA, Heussy JK, Dacheux RF, Raviola E, Masland RH. The population of bipolar cells in the rabbit retina. J Comp Neurol. 2004;472:73–86. 4. Hartveit E. Functional organization of cone bipolar cells in the rat retina. J Neurophysiol. 1997;77:1716–30. 5. Ghosh KK, Bujan S, Haverkamp S, Feigenspan A, Wassle H. Types of bipolar cells in the mouse retina. J Comp Neurol. 2004;469:70–82. 6. DeVries SH. Bipolar cells use kainate and AMPA receptors to filter visual information into separate channels. Neuron. 2000;28:847–56. 7. Slaughter MM, Miller RF. 2-amino-4-phosphonobutyric acid: a new pharmacological tool for retina research. Science. 1981;211:182–5. 8. Rockhill RL, Daly FJ, MacNeil MA, Brown SP, Masland RH. The diversity of ganglion cells in a mammalian retina. J Neurosci. 2002;22:3831–43. 9. Bloomfield SA, Dacheux RF. Rod vision: pathways and processing in the mammalian retina. Prog Retin Eye Res. 2001;20:351–84. 10. Strettoi E, Raviola E, Dacheux RF. Synaptic connections of the narrow-field, bistratified rod amacrine cell (AII) in the rabbit retina. J Comp Neurol. 1992;325:152–68. 11. Strettoi E, Dacheux RF, Raviola E. Synaptic connections of rod bipolar cells in the inner plexiform layer of the rabbit retina. J Comp Neurol. 1990;295:449–66. 12. Strettoi E, Dacheux RF, Raviola E. Cone bipolar cells as interneurons in the rod pathway of the rabbit retina. J Comp Neurol. 1994;347:139–49. 13. Massey SC, O’Brien JJ, Trexler EB, Li W, Keung JW, Mills SL, O’Brien J. Multiple neuronal connexins in the mammalian retina. Cell Commun Adhes. 2003;10:425–30. 14. Feigenspan A, Janssen-Bienhold U, Hormuzdi S, Monyer H, Degen J, Sohl ¨ G, Willecke K, Ammermu¨ller J, Weiler R. Expression of connexin36 in cone pedicles and off-cone bipolar cells of the mouse retina. J Neurosci. 2004;24:3325–34. 15. Feigenspan A, Teubner B, Willecke K, Weiler R. Expression of neuronal connexin36 in AII Amacrine Cells of the mammalian retina. J Neurosci. 2001;21:230–9. 16. Mills SL, O’Brien JJ, Li W, O’Brien J, Massey SC. Rod pathways in the mammalian retina use connexin36. J Comp Neurol. 2001;436:336–50. 17. O’Brien JJ, Li W, Pan F, Keung J, O’Brien J, Massey SC. Coupling between A-type horizontal cells is mediated by connexin 50 gap junctions in the rabbit retina. J Neurosci. 2006;26:11624–36. 18. Hombach S, Janssen-Bienhold U, Sohl G, Schubert T, Bu¨ssow H, Ott T, Weiler R, ¨ Willecke K. Functional expression of connexin57 in horizontal cells of the mouse retina. Eur J Neurosci. 2004;19:2633–40. 19. Dedek K, Schultz K, Pieper M, Dirks P, Maxeiner S, Willecke K, Weiler R, JanssenBienhold U. Localization of heterotypic gap junctions composed of connexin45 and connexin36 in the rod pathway of the mouse retina. Eur J Neurosci. 2006;24:1675–86.

19 Retina

415

20. O’Brien J, Nguyen HB, Mills SL. Cone photoreceptors in bass retina use two connexins to mediate electrical coupling. J Neurosci. 2004;24:5632–42. 21. Shields CR, Klooster J, Claassen Y, Ul-Hussain M, Zoidl G, Dermietzel R, Kamermans M. Retinal horizontal cell-specific promotor activity and protein expression of zebrafish connexin 52.6 and connexin 55.5. J Comp Neurol. 2007;501:765–79. 22. O’Brien J, Bruzzone R, White TW, Al-Ubaidi MR, Ripps H. Cloning and expression of two related connexins from the perch retina define a distinct subgroup of the connexin family. J Neurosci. 1998;18:7625–37. 23. Kothmann WW, Li X, Burr GS, O’Brien J. Connexin 35/36 is phosphorylated at regulatory sites in the retina. Vis Neurosci. 2007;24:363–75. 24. Haverkamp S, Grunert U, Wassle H. The cone pedicle, a complex synapse in the retina. Neuron 2000;27:85–95. 25. Ahnelt PK, Kolb H. The mammalian photoreceptor mosaic-adaptive design. Prog Retin Eye Res. 2000;19:711–77. 26. Raviola E, Gilula NB. Gap junctions between photoreceptor cells in the vertebrate retina. Proc Natl Acad Sci USA. 1973;70:1677–81. 27. Tsukamoto Y, Morigiwa K, Ueda M, Sterling P. Microcircuits for night vision in mouse retina. J Neurosci. 2001;21:8616–23. 28. Tsukamoto Y, Masarachia P, Schein SJ, Sterling P. Gap junctions between the pedicles of macaque foveal cones. Vision Res. 1992;32:1809–15. 29. Li W, DeVries SH. Separate blue and green cone networks in the mammalian retina. Nat Neurosci. 2004;7:751–6. 30. DeVries SH, Qi X, Smith R, Makous W, Sterling P. Electrical coupling between mammalian cones. Curr Biol. 2002;12:1900–7. 31. Hornstein EP, Verweij J, Schnapf JL. Electrical coupling between red and green cones in primate retina. Nat Neurosci. 2004;7:745–50. 32. Ahnelt P, Keri C, Kolb H. Identification of pedicles of putative blue-sensitive cones in the human retina. J Comp Neurol. 1990;293:39–53. 33. Zoidl G, Bruzzone R, Weickert S, Kremer M, Zoidl C, Mitropoulou G, Srinivas M, Spray DC, Dermietzel R. Molecular cloning and functional expression of ZfCx52.6: a novel connexin with hemichannel-forming properties expressed in horizontal cells of the zebrafish retina. J Biol Chem. 2004;279:2913–21. 34. Hsu A, Smith RG, Buchsbaum G, Sterling P. Cost of cone coupling to trichromacy in primate fovea. J Opt Soc Am A Opt Image Sci Vis. 2000;17:635–40. 35. Smith RG, Freed MA, Sterling P. Microcircuitry of the dark-adapted cat retina: functional architecture of the rod-cone network. J Neurosci. 1986;6:3505–17. 36. Hornstein EP, Verweij J, Li PH, Schnapf JL. Gap-junctional coupling and absolute sensitivity of photoreceptors in macaque retina. J Neurosci. 2005;25:11201–9. 37. Schneeweis DM, Schnapf JL. The photovoltage of macaque cone photoreceptors: adaptation, noise, and kinetics. J Neurosci. 1999;19:1203–16. 38. Schneeweis DM, Schnapf JL. Photovoltage of rods and cones in the macaque retina. Science 1995;268:1053–6. 39. Nelson R. Cat cones have rod input: a comparison of the response properties of cones and horizontal cell bodies in the retina of the cat. J Comp Neurol. 1977; 172:109–35. 40. Deans MR, Volgyi B, Goodenough DA, Bloomfield SA, Paul DL. Connexin36 is essential for transmission of rod-mediated visual signals in the mammalian retina. Neuron. 2002;36:703–12. 41. Lee EJ, Han JW, Kim HJ, Kim IB, Lee MY, Oh SJ, Chung JW, Chun MH. The immunocytochemical localization of connexin 36 at rod and cone gap junctions in the guinea pig retina. Eur J Neurosci. 2003; 18:2925–34. 42. Dunn FA, Rieke F. The impact of photoreceptor noise on retinal gain controls. Curr Opin Neurobiol. 2006;16:363–70.

416

S.C. Massey

43. Zhang J, Wu SM. Connexin35/36 gap junction proteins are expressed in photoreceptors of the tiger salamander retina. J Comp Neurol. 2004;470:1–12. 44. Zaghloul KA, Boahen K. Optic nerve signals in a neuromorphic chip I: outer and inner retina models. IEEE Trans Biomed Eng. 2004;51:657–66. 45. Pan F, Massey SC. Rod and cone input to horizontal cells in the rabbit retina. J Comp Neurol. 2007;500:815–31. 46. Mills SL, Massey SC. Distribution and coverage of A- and B-type horizontal cells stained with neurobiotin in the rabbit retina. Vis Neurosci. 1994;11:549–60. 47. Dacheux RF, Raviola E. Horizontal cells in the retina of the rabbit. J Neurosci. 1982;2:1486–93. 48. Vaney DI. The coupling pattern of axon-bearing horizontal cells in the mammalian retina. Proc Biol Sci. 1993;252:93–101. 49. Verweij J, Hornstein EP, Schnapf JL. Surround antagonism in macaque cone photoreceptors. J Neurosci. 2003;23:10249–57. 50. Cueva JG, Haverkamp S, Reimer RJ, Edwards R, Wassle H, Brecha NC. Vesicular g-aminobutyric acid transporter expression in amacrine and horizontal cells. J Comp Neurol. 2002;445:227–37. 51. Kamermans M, Fahrenfort I, Schultz K, Janssen-Bienhold U, Sjoerdsma T, Weiler R. Hemichannel-mediated inhibition in the outer retina. Science 2001;292: 1178–80. 52. Vessey JP, Lalonde MR, Mizan HA, Welch NC, Kelly ME, Barnes S. Carbenoxolone inhibition of voltage-dependent Ca channels and synaptic transmission in the retina. J Neurophysiol. 2004;92:1252–6. 53. Cadetti L, Thoreson WB. Feedback effects of horizontal cell membrane potential on cone calcium currents studied with simultaneous recordings. J Neurophysiol. 2006;95: 1992–5. 54. DeVries SH. Exocytosed protons feedback to suppress the Ca2+ current in mammalian cone photoreceptors. Neuron. 2001;32:1107–17. 55. Famiglietti EV Jr, Kolb H. A bistratified amacrine cell and synaptic cirucitry in the inner plexiform layer of the retina. Brain Res. 1975;84:293–300. 56. Hampson EC, Vaney DI, Weiler R. Dopaminergic modulation of gap junction permeability between amacrine cells in mammalian retina. J Neurosci. 1992;12:4911–22. 57. Mills SL, Massey SC. Differential properties of two gap junctional pathways made by AII amacrine cells. Nature. 1995;377:734–7. 58. Veruki ML, Hartveit E. AII (Rod) amacrine cells form a network of electrically-coupled interneurons in the mammalian retina. Neuron. 2002;33:935–46. 59. Veruki ML, Hartveit E. Electrical synapses mediate signal transmission in the rod pathway of the mammalian retina. J Neurosci. 2002;22:10558–66. 60. Vardi N, Smith RG. The AII amacrine network: coupling can increase correlated activity. Vision Res. 1996;36:3743–57. 61. Smith RG, Vardi N. Simulation of the AII amacrine cell of mammalian retina: functional consequences of electrical coupling and regenerative membrane properties. Vis Neurosci. 1995;12:851–60. 62. Massey SC, Mills SL. Antibody to calretinin stains AII amacrine cells in the rabbit retina: double-label and confocal analyses. J Comp Neurol. 1999;411:3–18. 63. Urschel S, Hoher T, Schubert T, Alev C, Sohl orfer P, Asahara T, Dermietzel R, ¨ ¨ G, Worsd ¨ ¨ Weiler R, Willecke K. Protein kinase A-mediated phosphorylation of connexin36 in mouse retina results in decreased gap junctional communication between AII amacrine cells. J Biol Chem. 2006;281:33163–71. 64. Xin D, Bloomfield SA. Comparison of the responses of AII amacrine cells in the darkand light-adapted rabbit retina. Vis Neurosci. 1999;16:653–65. 65. Wassle H, Grunert U, Chun MH, Boycott BB. The rod pathway of the macaque monkey retina: identification of AII-amacrine cells with antibodies against calretinin. J Comp Neurol. 1995;361:537–51.

19 Retina

417

66. Mills SL, Massey SC. A series of biotinylated tracers distinguishes three types of gap junction in retina. J Neurosci. 2000;20:8629–36. 67. Vaney DI, Nelson JC, Pow DV. Neurotransmitter coupling through gap junctions in the retina. J Neurosci. 1998;18:10594–602. 68. Maxeiner S, Dedek K, Janssen-Bienhold U, Ammermuller J, Brune H, Kirsch T, Pieper M, Degen J, Kruger O, Willecke K, Weiler R. Deletion of connexin45 in mouse retinal neurons disrupts the rod/cone signaling pathway between AII amacrine and ON cone bipolar cells and leads to impaired visual transmission. J Neurosci. 2005;25:566–76. 69. Han Y, Massey SC. Electrical synapses in retinal ON cone bipolar cells: subtype-specific expression of connexins. Proc Natl Acad Sci USA. 2005;102:13313–8. 70. Trexler EB, Li W, Mills SL, Massey SC. Coupling from AII amacrine cells to ON cone bipolar cells is bidirectional. J Comp Neurol. 2001;437:408–22. 71. Trexler EB, Li W, Massey SC. Simultaneous contribution of two rod pathways to AII amacrine and cone bipolar cell light responses. J Neurophysiol. 2005;93:1476–85. 72. Pang JJ, Abd-El-Barr MM, Gao F, Bramblett DE, Paul DL, Wu SM. Relative contributions of rod and cone bipolar cell inputs to AII amacrine cell light responses in the mouse retina. J Physiol. 2007;580:397–410. 73. Deans MR, Volgyi B, Goodenough DA, Bloomfield SA, Paul DL. Connexin36 is essential for transmission of rod-mediated visual signals in the mammalian retina. Neuron 2002;36:703–12. 74. Chow RL, Volgyi B, Szilard RK, Ng D, McKerlie C, Bloomfield SA, Birch DG, McInnes RR. Control of late off-center cone bipolar cell differentiation and visual signaling by the homeobox gene Vsx1. Proc Natl Acad Sci USA. 2004;101:1754–9. 75. Hu EH, Bloomfield SA. Gap junctional coupling underlies the short-latency spike synchrony of retinal ganglion cells. J Neurosci. 2003;23:6768–77. 76. Schubert T, Maxeiner S, Kruger O, Willecke K, Weiler R. Connexin45 mediates gap junctional coupling of bistratified ganglion cells in the mouse retina. J Comp Neurol. 2005;490:29–39. 77. Marc RE, Jones BW. Molecular phenotyping of retinal ganglion cells. J Neurosci. 2002;22:413–27. 78. DeVries SH. Correlated firing in rabbit retinal ganglion cells. J Neurophysiol. 1999;81:908–20. 79. Meister M. Multineuronal codes in retinal signaling. Proc Natl Acad Sci USA. 1996;93:609–14. 80. Usrey WM, Reppas JB, Reid RC. Paired-spike interactions and synaptic efficacy of retinal inputs to the thalamus. Nature. 1998;395:384–7. 81. Ackert JM, Wu SH, Lee JC, Abrams J, Hu EH, Perlman I, Bloomfield SA. Light-induced changes in spike synchronization between coupled ON direction selective ganglion cells in the mammalian retina. J Neurosci. 2006;26:4206–15. 82. Kofuji P, Newman EA. Potassium buffering in the central nervous system. Neuroscience. 2004;129:1045–56. 83. Mobbs P, Brew H, Attwell D. A quantitative analysis of glial cell coupling in the retina of the axolotl (Ambystoma mexicanum). Brain Res. 1988;460:235–45. 84. Ceelen PW, Lockridge A, Newman EA. Electrical coupling between glial cells in the rat retina. Glia. 2001;35:1–13. 85. Zahs KR, Newman EA. Asymmetric gap junctional coupling between glial cells in the rat retina. Glia. 1997;20:10–22. 86. Zahs KR, Kofuji P, Meier C, Dermietzel R. Connexin immunoreactivity in glial cells of the rat retina. J Comp Neurol. 2003;455:531–46. 87. Zahs KR, Ceelen PW. Gap junctional coupling and connexin immunoreactivity in rabbit retinal glia. Vis Neurosci. 2006;23:1–10. 88. Wu DM, Minami M, Kawamura H, Puro DG. Electrotonic transmission within pericytecontaining retinal microvessels. Microcirculation. 2006;13:353–63.

Chapter 20

Connexins in the Inner Ear Regina Nickel, Andrew Forge and Daniel Jagger

Abstract Intercellular communication via gap junctions is crucial for auditory function. This has been emphasized by the findings that mutations in certain connexin genes, in particular GJB2 and GJB6 (encoding CX26 and CX30), cause sensorineural deafness. Cx26 and Cx30 proteins are widely expressed in the epithelial and connective tissues of the cochlea and vestibular system, where they likely form heteromeric gap junction channels. Despite the study of mutant channels and of mouse models for both recessive and dominant autosomal deafness, it is still unclear why gap junctions are essential for auditory function and why Cx26 and Cx30 cannot compensate for loss of each other. It is generally thought that gap junctions play a role in the maintenance of ionic and metabolic homeostasis in the inner ear. Recent studies highlight the possible involvement of gap junctions in intercellular signaling via second messengers between the nonsensory cells. This chapter summarizes current knowledge about the molecular and functional properties of inner ear gap junctions and the inner ear pathologies associated with connexin mutations. Keywords Inner ear Cochlea Deafness Cx26 Cx29 Cx30 Cx31 Cx43

20.1 Introduction The vertebrate inner ear consists of the hearing organ, the cochlea, and the vestibular system, which detects movement of the head, thereby contributing to maintenance of balance. The functions of the organs of the inner ear are dependent on tightly controlled ionic environments, in particular for K+ ions, the main charge carrier for sensory transduction. The inner ear contains two major fluid spaces, which are separated by tight junction barriers. The scala media (Fig. 20.1a) is filled with endolymph, which is rich in K+ and poor in A. Forge (*) Centre for Auditory Research, UCL Ear Institute, University College London, 332 Gray’s Inn Road, London, WC1X 8EE, United Kingdom e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_20, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

419

420

R. Nickel et al.

Fig. 20.1 Expression of Cx26 and Cx30 in the mature rat cochlea. (a) Photomicrograph of a mid-turn region of a cochlear slice. (b) Cx26 expression in epithelial cells, and in connective tissues in the lateral wall. (c) Photomicrograph showing detail of outer hair cells (ohc), Deiters’ cells (Dc; immediately adjacent to ohc) and Hensen’s cells (Hc; located laterally). Cx26 (d) and Cx30 (e) are coexpressed in the supporting cells. Note the large plaque-like accumulations of Cx26 and Cx30 at the base of Deiters’ cells (arrows). (f) Photomicrograph of the connective tissues in the lateral wall that contribute to the generation of the endocochlear potential. Cx26 (g) and Cx30 (h) are coexpressed extensively in the fibrocytes of the spiral ligament, but to a lesser degree in the stria vascularis. Cx26 is detected alone in endothelial cells lining blood vessels in the stria vascularis (g, arrow). (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

Na+, thus resembling an intracellular fluid. The scala tympani, scala vestibuli, and the extracellular spaces within the inner ear contain perilymph, which has high Na+ and low K+ concentration, more typical of extracellular fluid. The

20 Inner Ear

421

essential driving force for sensory transduction in the mammalian cochlea is provided by a high positive potential (80 mV), the endocochlear potential (EP; for review see [1,2]). The ionic composition of the endolymph and the EP are generated and maintained by the stria vascularis, a stratified ion-transporting tissue that lines the lateral wall of the cochlear duct (Fig. 20.1f). The existence of the EP requires electrochemical isolation of the stria vascularis from endolymph and perilymph. This is achieved by tight junction networks between marginal cells, which line the scala media, and between basal cells, which separate the stria vascularis from the underlying spiral ligament as well as tight junctions between endothelial cells of the intrastrial capillaries that isolate the intrastrial spaces. The EP is generated across the membrane of the third cell type in the stria, the intermediate cells, which are enclosed between the marginal and basal cells. The sensory epithelia of the inner ear contain mechanosensory receptors, known as hair cells, each surrounded by nonsensory supporting cells. Across all vertebrate classes, supporting cells are interconnected by large (>10 mm2) gap junctions [3], which contain several hundred thousand junctional channels—

Fig. 20.2 Gap junctions in the sensory epithelia of the inner ear. (a) Freeze-fracture micrograph of the chicken auditory sensory epithelium showing hair cells (hc) and surrounding supporting cells (sc). Three large gap junction plaques are indicated by arrows 1, 2, 3. Scale bar 2 mm. (b) Gap junction plaque indicated at arrow 3 in panel (a) to reveal extent and size of a single junction. Scale bar 0.5 mm. (c) Region of junction in panel (b) to demonstrate that the morphology of the entire structure shown in panel (b) is formed of closely packed particles (or depressions) consistent with that of the freeze-fracture appearance of a gap junction plaque. Scale bar 0.5 mm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

422

R. Nickel et al.

among the largest gap junctions in the body (Fig. 20.2). The extensive coupling of the majority of cells within the inner ear via gap junctions is thought to play a role in fluid homeostasis or intercellular signaling (for review see [4,5]).

20.2 Gap Junction Networks in the Inner Ear Based on ultrastructural studies, two gap junction networks — that of the epithelial tissue and that of the connective tissue — have been described within the mammalian cochlear duct [6] and the vestibular system [7]. The epithelial gap junction system connects the supporting cells of the sensory epithelia and bordering epithelial cells. In the cochlea, gap junctions are present between the interdental cells of the spiral limbus, the supporting cells of the organ of Corti, and the cells within the root processes of the spiral ligament (Fig. 20.1b) [3,6,8]. No gap junctions have been identified between hair cells and supporting cells. The syncytial nature of supporting cells, and the segregation of hair cells from this functional unit, has been confirmed by electrophysiology [9,10] and by dye tracer studies in cochlear explants [11,12]. The gap junction system of inner ear connective tissues comprises the cells of the ion-transporting epithelium, various types of fibrocytes, and mesenchymal cells that line the scala vestibule. In the cochlear duct, gap junctions are present between the fibrocytes of the spiral limbus, the spiral ligament, and the suprastrial zone (Fig. 20.1b) [3,6,13]. Numerous gap junctions have been found between adjacent basal cells of the stria vascularis, between basal cells and intermediate cells, and between basal cells and fibrocytes of the spiral ligament (Fig. 20.1 g,h). No gap junctions have been identified between the intermediate and the marginal cells of the stria vascularis, or between adjacent marginal cells [3,6], suggesting that these cells are isolated from a functional unit formed by intermediate cells, basal cells and fibrocytes of the spiral ligament. However, as yet there are no functional studies confirming these proposed patterns of interconnection based on anatomical data.

20.2.1 Connexin Expression in the Inner Ear The predominant connexins in the mammalian inner ear are Cx26 and Cx30 (Fig. 20.1), which are present in all cells comprising the epithelial and connective tissue gap junction systems of the cochlea and vestibule [3,6,7,11,14,15,16]. Furthermore, double immunofluorescence revealed overlapping staining patterns for Cx26 and Cx30, suggesting colocalization of the two connexin isoforms within the gap junction plaques in the inner ear [3,14,16] that has been verified by co-immunoprecipitation of Cx26 and Cx30 [3,15] and by immunogold labeling of thin sections, showing Cx26 and Cx30 evenly distributed within the gap junction plaques [3].

20 Inner Ear

423

Another connexin isoform present in the mammalian cochlea is Cx31, which is confined to the spiral limbus and a particular population of fibrocytes in the spiral ligament and suprastrial zone [3,17,18]. In the rat cochlea, weak immunostaining of Cx43 has been detected between supporting cells in the organ of Corti, the stria vascularis, and spiral limbus [19]. This is contrary to findings in the postnatal mouse, where Cx43 is confined to cells lining the inside of the bony wall and the bone of the otic capsule, with no expression in the organ of Corti [3,20]. Cx29 is expressed in the Schwann cells surrounding the auditory and vestibular fibers of spiral ganglion neurons [21,22]. In the avian inner ear, the chicken orthologs to mammalian Cx30, Cx26, and Cx43 have been identified as major connexin isoforms [12]. Of note is that chick Cx30 is expressed exclusively in inner ear tissues [12,23]. Analogous to the distribution of mammalian Cx26/Cx30 gap junctions, chick Cx30 is also present in all cells comprising the gap junction networks of the sensory and ion transporting epithelia of the cochlear duct and vestibular sensory epithelium in the chicken [12].

20.3 Function of Inner Ear Gap Junctions The serial arrangement of gap junction networks within two distinct cochlear tissues has prompted the proposal that gap junctions form the structural pathways for K+ recirculation within the cochlea. The mechanistic details of this model have been reviewed extensively elsewhere [1,2,4,24,25]. Briefly, the model proposes that K+ ions exiting hair cells during auditory transduction are siphoned from the extracellular perilymph by supporting cells immediately adjacent to hair cells. They are then relayed radially via the epithelial gap junction network to the extracellular space of the lateral wall, where they are taken up by fibrocytes of the spiral ligament via the Na+/K+–adenosine triphosphatase (ATPase) and Na+/2Cl–/K+-cotransporter. Through the connective tissue gap junction system, K+ ions bypass the tight junction barrier of the basal cells and are released by intermediate cells into the extracellular space of the stria vascularis. K+ is finally taken up by marginal cells through Na+/K+-ATPase and Na+/2Cl–/K+-cotransporter and secreted back into the endolymph. This model is derived largely from indirect evidence, mostly in the form of protein expression data, and requires a substantial amount of support from direct physiological evidence. The importance of buffering K+ in perilymph is clear; depolarization-induced damage to hair cells and neurons would lead to permanent hearing impairment. A parallel model can be extended from the spatial K+ buffering observed in glial cell networks in the brain [26], where the extensive cell syncytium acts as a K+ sink, allowing redistribution of the K+ flux to regions of inactivity. In the inner ear, the extensive connexin expression in the epithelial and connective tissue networks along the cochlear partition (Fig. 20.1b) creates a substantial buffering volume.

424

R. Nickel et al.

The K+-buffering theories have a number of potential problems, not least the inability of individual connexins (either Cx26 or Cx30) to compensate for the loss of the other [27,28]. Both Cx26 and Cx30 homotypic channels can transfer K+ ions [29,30], and so should ably manage K+ when expressed alone in the cochlea. Furthermore, in the mammalian vestibular system and the avian inner ear, the ion-transporting epithelia do not generate an EP [31,32] and thus would not appear to require a gap junction–mediated bypass for K+ to the ion transporting epithelial cells. Nevertheless, extensive gap junction networks are present in these tissues [3,12], suggesting roles for gap junctions beyond K+ circulation.

20.3.1 Biophysical Characteristics of Cochlear Gap Junctions The properties of cochlear gap junctions have been determined in native cochlear tissue, using combinations of in vivo and ex vivo preparations, and in transfected cell lines expressing Cx26 and/or Cx30. Homotypic Cx26 and homotypic Cx30 channels have been studied extensively in cell culture systems using the double patch clamp technique [30,33,34,35]. Considering the high sequence homology between Cx26 and Cx30 [34], the properties of these channel types are surprisingly distinct, particularly in terms of conductance and voltage-gating. Furthermore, exogenously expressed heterotypic and heteromeric Cx26/Cx30 channels have been recently shown to differ in their biophysical properties from their homotypic counterparts [33]. The voltage-dependence and gating properties of gap junction channels have been determined in small groups of supporting cells. The voltage-dependence of gap junctions in Hensen’s cells (supporting cells at lateral border of organ of Corti; Fig. 20.1c) has been grouped into four distinct types [36]. These gating responses show various degrees of polarity-dependent or polarity-independent rectification. The observation of extensive asymmetric voltage-gating points to a complexity in gap junctional coupling, and argues in favor of nonhomotypic gap junction channel types. Even within single recordings it is possible that multiple channel types contribute to complex responses. Intercellular channels between adjacent Deiters’ cells (supporting cells immediately adjacent to outer hair cells; Fig. 20.1c) may also allow rectification of current flow [36]. In addition, there is evidence from patch clamp recordings that the gap junctional conductance between cochlear supporting cells can be modulated by intrinsic factors such as nitric oxide [37] or turgor pressure [38].

20.3.2 Molecular Selectivity of Cochlear Gap Junctions The molecular selectivity of cochlear gap junctions will be of equal importance to their electrophysiological properties. There is increasing evidence that intercellular fluxes of second messengers such as inositol phosphates and Ca2+ ions regulate cochlear physiology. Indeed, the impaired transfer of the Ca2+-

20 Inner Ear

425

mobilizing molecule inositol triphosphate (IP3) has been suggested as a cause of recessive deafness due to Cx26 mutation [39]. The molecular selectivity of cochlear gap junctions has been explored extensively using dye transfer techniques. As in other tissues, these channels have the ability to select among molecules of different molecular weight and/or charge. Native cochlear gap junctions, in common with other Cx30-containing junctions [14,29,33,35], often show a preference for positively charged species over negatively charged ones, and are relatively restrictive to movement of molecules with higher molecular weight. In recent studies utilizing a cochlear slice [11,40], mature Deiters’ cells (older than postnatal day twelve, P12) allowed intercellular transfer of neurobiotin (287 Da, charge +1) but resisted the passage of Lucifer yellow (443 Da, charge –2). However, in Hensen’s cells the passage of Lucifer yellow was less restricted. The variable expression of Cx26 and Cx30 in the organ of Corti is supported by differential permeability to Lucifer yellow, and by immunolocalization (Fig. 20.1d,e) [3,14,16]. In various species there is a higher expression of Cx30 than Cx26 in Deiters’ cells compared to Hensen’s cells [11,14,41]. In addition, there are dramatic postnatal changes in the permeability properties of gap junctions in the organ of Corti. There is a rapid increase of dye transfer between Deiters’ cells in rats during the first postnatal week, concomitant with increasing connexin immunoreactivity [11]. Though there is free transfer of Lucifer yellow between Deiters’ cells at P8, this likely reflects an immature phenotype, as this transfer is no longer apparent after the onset of hearing (>P12, see above). Contrary to the restricted transfer of Lucifer yellow between supporting cells of the organ of Corti, gap junctions between supporting cells in the avian inner ear are highly permeable to large anionic dyes [12]. Furthermore, asymmetric dye transfer between auditory supporting cells was observed, suggesting that directional signaling pathways exist within the avian cochlea. Although dye transfer experiments provide useful descriptions of gap junction characteristics it is perhaps more pertinent to study the movement of endogenous messengers such as Ca2+ and inositol phosphates. These agents present a greater technical challenge as they in themselves influence the physiology of the cells being studied. Gap junction–mediated Ca2+ waves have been monitored in HEK293 cells expressing Cx26 or Cx30 [14]. Ca2+ waves appear to spread significantly faster through groups of cells expressing heteromeric Cx26/Cx30 channels, compared with groups of cells expressing homomeric Cx26 or homomeric Cx30 channels. Such differences have yet to be confirmed in functionally mature native cochlear tissue. In supporting cells of immature cochlear cultures, injection of IP3 into a single cell initiated Ca2+ waves through neighboring cells [39], suggesting an innate permeability of these gap junctions to this molecule. Of note is the recent observation that homomeric Cx26 hemichannels allow the passage of myoinositol and various inositol polyphosphates. In contrast, heteromeric Cx26/Cx32 channels are much more selective, even distinguishing between isoforms of IP3 [42]. Comparable preliminary results have been observed for heteromeric Cx26/Cx30 channels [43] (see Chapter 7).

426

R. Nickel et al.

Future studies must consider variations in connexin composition of channels throughout the mature cochlea, as these are all likely to have quite distinct properties. While much energy has been spent defining the physiology of the supporting cells in the organ of Corti, it is just as important to understand the contribution of the connective tissue gap junction network in the lateral wall to normal hearing, and how its dysfunction leads to cochlear pathology. An emerging idea is a role for unmatched hemichannels in the membrane of supporting cells in the organ of Corti that may mediate the release of ATP [44,45]. It has been suggested that Ca2+-activated ATP release via hemichannels may underlie the propagation of Ca2+ waves between supporting cells in response to hair cell damage [44]. A role for Cx30 hemichannels in controlling osmotic balance in intermediate cells of the stria vascularis has also been proposed [46]. However, direct physiological evidence for functional hemichannels in vivo in the mature inner ear is still lacking.

20.4 Connexin Mutations and Deafness The importance of gap junctional communication for auditory function has been highlighted by the discovery that mutations in genes coding for CX26 [47], CX30 [48], and CX31 [49] cause hereditary hearing loss. There has also been a report of deafness arising from mutations in CX43 [50], but this was later attributed to mutations in the CX43 pseudogene [51]. Connexin mutations are associated with autosomal recessive and dominant hearing loss, whose phenotypes are mostly confined to the inner ear (nonsyndromic) but can occur with other clinical features (syndromic), in particular skin disorders (for review see [52]). Despite the genetic heterogeneity of nonsyndromic autosomal recessive deafness, a single locus on chromosome 13q11-12 accounts for up to 50% of this type of hearing loss [53,54]. The responsible gene has been identified as GJB2, which encodes CX26 [47]. More than 90, mostly recessive, mutations have been characterized in the GJB2 gene, including splice, nonsense, missense, and frameshift mutations (see http://www.davinci.crg.es/deafness). With a carrier frequency of 2 to 4%, the most common mutation in European and North American populations is a deletion of a single guanine nucleotide, known as 35delG, which results in a frameshift and the subsequent premature termination of protein translation. The majority of recessive CX26 mutations studied to date do not form functional channels in recombinant expression systems owing to impaired assembly of hemichannels, impaired targeting to the plasma membrane, or reduced protein stability [55,56,57,58]. However, several CX26 mutations form homotypic gap junction channels, albeit with reduced electrical coupling and impaired permeability to dye tracers [57,59,60,61]. A subset of CX26 mutant channels with amino acid substitutions in the second transmembrane domain (V84L, A88S, and V95M) did not significantly affect electrical

20 Inner Ear

427

coupling, but impaired the transfer of larger molecules such as IP3 [39,62]. Other CX26 mutants are able to contribute to gap junctions only when coexpressed with either their wild-type form (as in heterozygotic expression), or with wild-type CX30. Such heteromeric rescue has been observed for several connexin mutants [63,64]. Altered properties of such channels may also contribute to the various CX26-based deafness phenotypes [65]. Several rare missense mutations in GJB2 have been detected in families with autosomal dominant inherited deafness [66,67]. These mutations primarily affect amino acids within the extracellular loops (e.g., W44S, R75Q, and R75W), and result in impaired electrical coupling and dye transfer [68,69,70]. In addition, dominant CX26 mutations, in particular those that interfere with intracellular trafficking [71], may be associated with various skin disorders [72,73,74,75,76]. A 342 kilobase pair deletion in the gene coding for CX30, GJB6, has been linked to non–GJB2-related cases of nonsyndromic recessive deafness [77]. However, it is possible that this large deletion also has an effect on the regulation of GJB2 [78]. In addition, a missense mutation in GJB6 affecting the amino-terminal domain of CX30 (T5M) is associated with nonsyndromic autosomal dominant hearing impairment with late onset [48]. Unlike CX30 mutations associated with skin disease, T5M mutants formed electrically coupled channels but showed impaired permeability to dye tracer and IP3 [62,79]. Mutations in the gene encoding CX31 (GJB3) have been detected in two different disorders, deafness [49,80] and erythrokeratodermia variabilis [52], a skin disorder. Mutations linked to nonsyndromic, dominant deafness are concentrated at the second extracellular loop, and these proteins do not form functional channels [81]. The mild to moderate hearing loss is of late onset and, consistent with the CX31 expression pattern [18], affects high frequencies preferentially [49]. Autosomal recessive mutations located within the third transmembrane domain may be associated with nonsyndromic moderate to profound hearing loss [80].

20.4.1 Cochlear Pathology of Transgenic Mouse Mutants In agreement with the clinical phenotype of connexin mutations in humans, knockout (KO) of Cx26 [27,82] or Cx30 [28] in the inner ear of mice results in severe hearing impairment shortly after the onset of hearing, but no discernible vestibular dysfunction. Ablation of Cx31 does not result in morphological or functional defects in the inner ear of mice [83]. Differences between murine [84] and human Cx31 [85] in the ability to form heterotypic gap junction channels may contribute to the disparities in the phenotype between mice and humans. In addition, while normal auditory function may not be compromised by the loss of Cx31, it would be interesting to study the susceptibility to noise and aging of

428

R. Nickel et al.

Cx31KO mice, especially as the absence of functional CX31 channels is associated with progressive deafness, a characteristic of age-related hearing loss. Conflicting reports have been presented for Cx29KO mice. While one study found no effect of the deletion of Cx29 on auditory function [21], another group detected demyelination of spiral ganglion neurons in these mice, which may be responsible for the prolonged latency in auditory brainstem responses and higher sensitivity to noise damage [22]. In a mouse model for recessive Cx26-related deafness, the neonatal lethality of Cx26KO mice was overcome by restricting the deletion of Cx26 to the epithelial gap junction system [27]. The inner ear of homozygous mice developed normally, but by P14, soon after hearing onset, apoptosis of supporting cells of the inner hair cells was observed, extending later to the outer hair cells and their supporting cells. It has been suggested that the death of supporting cells surrounding the inner hair cells is caused by oxidative stress owing to the interference of accumulated K+ with the removal of the neurotransmitter glutamate from the extracellular space [27]. In another mouse model, the function of Cx26 was inhibited using the R75W mutation, which is associated with dominantly inherited hearing loss [82]. Its dominant-negative effect on the function of gap junction channels was confirmed by electrophysiology and dye transfer analyses that demonstrated that gap junctional communication was not only inhibited through homotypic Cx26R75W channels but also through heteromeric Cx26R75W channels that also contain wild-type Cx26 or Cx30 [65,70,72]. At two weeks of age, Cx26R75W mutant animals displayed deformities of both the tunnel of Corti and the supporting cells. By seven weeks, the outer hair cells had degenerated, whereas the inner hair cells were still present. No apparent effect was detected in the stria vascularis, which was confirmed by a normal EP. The confinement of the pathogenicity of Cx26R75W to the epithelial gap junction system of the inner ear may point to impaired K+ buffering within the organ of Corti, with subsequent adverse effects on cochlear homeostasis to which hair cells are particularly sensitive. The knockout of Cx30 in both the sensory epithelium and the connective tissues of the inner ear resulted in failure to generate an EP in 2-week-old homozygous mutant mice despite normal inner ear development and endolymphatic K+ concentrations [28]. From P18 onward, apoptosis of mainly hair cells was observed. In adult mice the endolymphatic K+ concentration was decreased, contributing to the profound deafness of adult mutant mice. Recent work has suggested that the failure of Cx30KO animals to generate an EP arises from breakdown of the permeability barriers around strial capillaries, thereby dissipating the electrical potential [46]. It is speculated that this results from disturbance of osmotic balance caused by loss of functional Cx30 hemichannels from the membranes of strial intermediate cells. The phenotypes of these transgenic mice demonstrate that Cx26 and Cx30 are both required for auditory function (but not vestibular function), and cannot compensate for each other. This would account for the nonsyndromic

20 Inner Ear

429

manifestations of CX26 and CX30 mutations, despite their expression in other tissues. It would also support the contention that Cx26 and Cx30 proteins may specifically act in concert to provide heteromeric hemichannels with the necessary properties that support auditory function [14,15,40,65]. In addition, it may be that the deletion of one of these connexin genes decreases expression of the other, resulting in an insufficient number of functional gap junctions in the cochlea. Although no change in the expression patterns of Cx30 and Cx26 was originally reported for Cx26KO and Cx30KO mice [27,28,82], respectively, Cx30KO mice have reduced Cx26 protein levels, despite normal messenger RNA levels. This suggests that homotypic Cx26 channels are relatively unstable, which may lead to an insufficient number of functional gap junction channels. Lending support to this hypothesis is the fact that the overexpression of Cx26 restores hearing in Cx30KO mice [86]. While the proposed rapid degradation of homotypic connexin channels may contribute to hearing impairment in cases where mutations result in the inability to form heteromeric gap junction channels, it does not account for deafness associated with connexin mutants that oligomerize into hemichannels and form functional junctional channels. Studies of those mutations highlighted the fact that inner ear gap junctions, in addition to their proposed participation in ionic homeostasis, may also play a role in intercellular signaling via second messengers [39,62].

20.5 Conclusion Building on the insights into functional properties of inner ear gap junctions and their mutations, largely gained from studies in expression systems, the focus must now be turned to native tissues and the junctional permeabilities to endogenous metabolites and second messengers. This will enable us to unravel the complexities of gap junctional communication in the inner ear and to better understand cochlear pathologies that result from different connexin mutations. Acknowledgments The authors are supported by the Biotechnology and Biological Sciences Research Council (BBSRC), Deafness Research UK, The Royal National Institute for Deaf People (RNID), and the Royal Society. We apologize to colleagues whose work could not be cited owing to space restrictions.

References 1. Wangemann P. Supporting sensory transduction: cochlear fluid homeostasis and the endocochlear potential. J Physiol. 2006;576:11–21. 2. Wangemann P. K+ cycling and the endocochlear potential. Hear Res. 2002;165:1–9. 3. Forge A, Becker D, Casalotti S, Edwards J, Marziano N, Nevill G. Gap junctions in the inner ear: comparison of distribution patterns in different vertebrates and assessment of connexin composition in mammals J Comp Neurol. 2003;467:207–31.

430

R. Nickel et al.

4. Zhao HB, Kikuchi T, Ngezahayo A, White TW. Gap junctions and cochlear homeostasis. J Membr Biol. 2006;209:177–86. 5. Kikuchi T, Kimura RS, Paul DL, Takasaka T, Adams JC. Gap junction systems in the mammalian cochlea. Brain Res. 2000;32:163–6. 6. Kikuchi T, Kimura RS, Paul DL, Adams JC. Gap junctions in the rat cochlea: immunohistochemical and ultrastructural analysis. Anat Embryol. 1995;191:101–18. 7. Kikuchi T, Adams JC, Paul DL, Kimura RS. Gap junction systems in the rat vestibular labyrinth: immunohistochemical and ultrastructural analysis. Acta Otolaryngol. 1994;114:520–8. 8. Jahnke K. The fine structure of freeze-fractured intercellular junctions in the guinea pig inner ear. Acta Otolaryngol Suppl. 1975;336:1–40. 9. Oesterle EC, Dallos P. Intracellular recordings from supporting cells in the guinea pig cochlea: DC potentials. J Neurophysiol. 1990;64:617–36. 10. Santos-Sacchi J, Dallos P. Intercellular communication in the supporting cells of the organ of Corti. Hear Res. 1983;9:317–26. 11. Jagger DJ, Forge A. Compartmentalized and signal-selective gap junctional coupling in the hearing cochlea. J Neurosci. 2006;26:1260–8. 12. Nickel R, Becker D, Forge A. Molecular and functional characterization of gap junctions in the avian inner ear. J Neurosci. 2006;26:6190–9. 13. Reale E, Luciano L, Franke K, Pannese E, Wermbter G, Iurato S. Intercellular junctions in the vascular stria and spiral ligament. J Ultrastruct Res. 1975;53:284–97. 14. Sun J, Ahmad S, Chen S, Tang W, Zhang Y, Chen P, Lin X. Cochlear gap junctions coassembled from Cx26 and 30 show faster intercellular Ca2+ signaling than homomeric counterparts. Am J Physiol. 2005;288:C613–23. 15. Ahmad S, Chen S, Sun J, Lin X. Connexins 26 and 30 are co-assembled to form gap junctions in the cochlea of mice. Biochem Biophys Res Commun. 2003;307:362–8. 16. Lautermann J, Ten Cate WJ, Altenhoff P, Gru¨mmer R, Traub O, Frank H-G, Jahnke K, Winterhager E. Expression of the gap-junction connexins 26 and 30 in the rat cochlea. Cell Tissue Res. 1998;294:415–20. 17. Lopez-Bigas N, Arbones ML, Estivill X, Simonneau L. Expression profiles of the connexin genes, Gjb1 and Gjb3, in the developing mouse cochlea. Mech Dev. 2002;119 Suppl 1:S111–5. 18. Xia AP, Ikeda K, Katori Y, Oshima T, Kikuchi T, Takasaka T. Expression of connexin 31 in the developing mouse cochlea. Neuroreport. 2000;11:2449–53. 19. Suzuki T, Takamatsu T, Oyamada M. Expression of gap junction protein connexin43 in the adult rat cochlea: comparison with connexin26. J Histochem Cytochem. 2003;51:903–12. 20. Cohen-Salmon M, Maxeiner S, Kruger O, Theis M, Willecke K, Petit C. Expression of the connexin43- and connexin45-encoding genes in the developing and mature mouse inner ear. Cell Tissue Res. 2004;316:15–22. 21. Eiberger J, Kibschull M, Strenzke N, Schober A, Bu¨ssow H, Wessig C, Djahed S, Reucher H, Koch DA, Lautermann J, Moser T, Winterhager E, Willecke K. Expression pattern and functional characterization of connexin29 in transgenic mice. Glia 2006; 53:601–11. 22. Tang W, Zhang Y, Chang Q, Ahmad S, Dahlke I, Yi H, Chen P, Paul DL, Lin X. Connexin29 is highly expressed in cochlear Schwann cells, and it is required for the normal development and function of the auditory nerve of mice. J Neurosci. 2006;26:1991–9. 23. Heller S, Sheane CA, Javed Z, Hudspeth AJ. Molecular markers for cell types of the inner ear and candidate genes for hearing disorders. Proc Natl Acad Sci USA. 1998;95:11400–5. 24. Kikuchi T, Adams JC, Miyabe Y, So E, Kobayashi T. Potassium ion recycling pathway via gap junction systems in the mammalian cochlea and its interruption in hereditary nonsyndromic deafness. Med Electron Microsc. 2000;33:51–6.

20 Inner Ear

431

25. Wangemann P. K+ cycling and its regulation in the cochlea and the vestibular labyrinth. Audiol Neurootol. 2002;7:199–205. 26. Kofuji P, Newman EA. Potassium buffering in the central nervous system. Neuroscience. 2004;129:1045–56. 27. Cohen-Salmon M, Ott T, Michel V, Hardelin JP, Perfettini I, Eybalin M, Wu T, Marcus DC, Wangemann P, Willecke K, Petit C. Targeted ablation of connexin26 in the inner ear epithelial gap junction network causes hearing impairment and cell death. Curr Biol. 2002;12:1106–11. 28. Teubner B, Michel V, Pesch J, Lautermann J, Cohen-Salmon M, So¨hl G, Jahnke K, Winterhager E, Herberhold C, Hardelin JP, Petit C, Willecke K. Connexin30 (Gjb6)deficiency causes severe hearing impairment and lack of endocochlear potential. Hum Mol Gen. 2003;12:13–21. 29. Manthey D, Banach K, Desplantez T, Lee CG, Kozak CA, Traub O, Weingart R, Willecke K. Intracellular domains of mouse connexin26 and -30 affect diffusional and electrical properties of gap junction channels. J Membr Biol. 2001;181:137–48. 30. Valiunas V, Manthey D, Vogel R, Willecke K, Weingart R. Biophysical properties of mouse connexin30 gap junction channels studied in transfected human HeLa cells. J Physiol. 1999;519:631–44. 31. Ninoyu O, Hommerich C, Morgenstern C. Endolymph formation in the inner ear of pigeons. ORL J Otorhinolaryngol Relat Spec. 1987;49:1–8. 32. Wangemann P. Comparison of ion transport mechanisms between vestibular dark cells and strial marginal cells. Hear Res. 1995;90:149–57. 33. Yum SW, Zhang J, Valiunas V, Kanaporis G, Brink PR, White TW, Scherer SS. Human connexin26 and connexin30 form functional heteromeric and heterotypic channels. Am J Physiol Cell Physiol. 2007;293:C1032–48. 34. Dahl E, Manthey D, Chen Y, Schwarz HJ, Chang YS, Lalley PA, Nicholson BJ, Willecke K. Molecular cloning and functional expression of mouse connexin-30,a gap junction gene highly expressed in adult brain and skin. J Biol Chem. 1996;271:17903–10. 35. Beltramello M, Bicego M, Piazza V, Ciubotaru CD, Mammano F, D’Andrea P. Permeability and gating properties of human connexins 26 and 30 expressed in HeLa cells. Biochem Biophys Res Commun. 2003;305:1024–33. 36. Zhao HB. Directional rectification of gap junctional voltage-gating between Deiters cells in the inner ear of guinea pig. Neurosci Lett. 2000;296:105–8. 37. Blasits S, Maune S, Santos-Sacchi J. Nitric oxide uncouples gap junctions of supporting Deiters’ cells from Corti’s organ. Pflu¨gers Arch. 2000;440:710–2. 38. Zhao HB, Santos-Sacchi J. Effect of membrane tension on gap junctional conductance of supporting cells in Corti’s organ. J Genet Physiol. 1998;112:447–55. 39. Beltramello M, Piazza V, Bukauskas FF, Pozzan T, Mammano F. Impaired permeability to Ins(1,4,5)P3 in a mutant connexin underlies recessive hereditary deafness. Nat Cell Biol. 2005;7:63–9. 40. Forge A, Marziano NK, Casalotti SO, Becker DL, Jagger D. The inner ear contains heteromeric channels composed of Cx26 and Cx30 and deafness-related mutations in Cx26 have a dominant-negative effect on Cx30. Cell Commun Adhes. 2003;10:341–6. 41. Zhao HB, Yu N. Distinct and gradient distributions of connexin26 and connexin30 in the cochlear sensory epithelium of guinea pigs. J Comp Neurol. 2006;499:506–18. 42. Ayad WA, Locke D, Koreen IV, Harris AL. Heteromeric, but not homomeric, connexin channels are selectively permeable to inositol phosphates. J Biol Chem. 2006;281:16727–39. 43. He L, Ayad WA, Harris AL. Expression and functional analysis of deafness-associated heteromeric connexin channels. ASCB Meet Abstr. 2006;358. 44. Gale JE, Piazza V, Ciubotaru CD, Mammano F. A mechanism for sensing noise damage in the inner ear. Curr Biol. 2004;14:526–9. 45. Zhao HB, Yu N, Fleming CR. Gap junctional hemichannel-mediated ATP release and hearing controls in the inner ear. Proc Natl Acad Sci USA. 2005;102:18724–9.

432

R. Nickel et al.

46. Cohen-Salmon M, Regnault B, Cayet N, Caille D, Demuth K, Hardelin JP, Janel N, Meda P, Petit C. Connexin30 deficiency causes instrastrial fluid-blood barrier disruption within the cochlear stria vascularis. Proc Natl Acad Sci USA. 2007;104:6229–6234. 47. Kelsell DP, Dunlop J, Stevens HP, Lench NJ, Liang JN, Parry G, Mueller RF, Leigh IM. Connexin 26 mutations in hereditary nonsyndromic sensorineural deafness. Nature 1997;387:80–3. 48. Grifa A, Wagner CA, D’Ambrosio L, Melchionda S, Bernardi F Lopez-Bigas N, Rabionet R, Arbones M, Della M, Estivill X, Zelante L, Lang F, Gasparini P. Mutations in GJB6 cause nonsyndromic autosomal dominant deafness at DFNA3 locus. Nat Genet. 1999;23:16–8. 49. Xia JH, Liu CY, Tang BS, Pan Q, Huang L, Dai HP, Zhang BR, Xie W, Hu DX, Zheng D, Shi XL, Wang DA, Xia K, Yu KP, Liao XD, Feng Y, Yang YF, Xiao JY, Xie DH, Huang JZ. Mutations in the gene encoding gap junction protein b-3 associated with autosomal dominant hearing impairment. Nat Genet. 1998;20:370–3. 50. Liu XZ, Xia XJ, Adams J, Chen ZY, Welch KO, Tekin M, Ouyang XM, Kristiansen A, Pandya A, Balkany T, Arnos KS, Nance WE. Mutations in GJA1 (connexin 43) are associated with nonsyndromic autosomal recessive deafness. Hum Mol Genet. 2001;10:2945–51. 51. Paznekas WA, Boyadjiev SA, Shapiro RE, Daniels O, Wollnik B, Keegan CE, Innis JW, Dinulos MB Christian C, Hannibal MC, Jabs EW. Connexin 43 (GJA1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet. 2003;72:408–18. 52. Rabionet R, Lopez-Bigas N, Arbones ML, Estivill X. Connexin mutations in hearing loss, dermatological and neurological disorders. Trends Mol Med. 2002;8:205–12. 53. Kenneson A, Van Naarden Braun K, Boyle C. GJB2 (connexin 26) variants and nonsyndromic sensorineural hearing loss: a HuGE review. Genet Med. 2002;4:258–74. 54. Snoeckx RL, Huygen PL, Feldmann D, Marlin S, Denoyelle F, Waligora J, MuellerMalesinska M, Pollak A, Ploski R, Murgia A, Orzan E, Castorina P, Ambrosetti U, Nowakowska-Szyrwinska E, Bal J, Wiszniewski W, Janecke AR, Nekahm-Heis D, Seeman P, Bendova O, Kenna MA, Frangulov A, Rehm HL, Tekin M, Incesulu A, Dahl HH, du Sart D, Jenkins L, Lucas D, Bitner-Glindzicz M, Avraham KB, Brownstein Z, del Castillo I, Moreno F, Blin N, Pfister M, Sziklai I, Toth T, Kelley PM, Cohn ES, Van Maldergem L, Hilbert P, Roux AF, Mondain M, Hoefsloot LH, Cremers CW, Lopponen T, Lopponen H, Parving A, Gronskov K, Schrijver I, Roberson J, Gualandi F, Martini A, Lina-Granade G, Pallares-Ruiz N, Correia C, Fialho G, Cryns K, Hilgert N, Van de Heyning P, Nishimura CJ, Smith RJ, Van Camp G. GJB2 mutations and degree of hearing loss: a multicenter study. Am J Hum Genet. 2005;77:945–57. 55. Thonnissen E, Rabionet R, Arbones ML, Estivill X, Willecke K, Ott T. Human connexin26 (GJB2) deafness mutations affect the function of gap junction channels at different levels of protein expression. Hum Genet. 2002;111:190–7. 56. Oshima A, Doi T, Mitsuoka K, Maeda S, Fujiyoshi Y. Roles of Met-34, Cys-64, and Arg75 in the assembly of human connexin 26. Implication for key amino acid residues for channel formation and function. J Biol Chem. 2003;278:1807–16. 57. D’Andrea P, Veronesi V, Bicego M, Melchionda S, Zelante L, Di Iorio E, Bruzzone R, Gasparini P. Hearing loss: frequency and functional studies of the most common connexin26 alleles. Biochem Biophys Res Commun. 2002;296:685–91. 58. Martin PE, Coleman SL, Casalotti SO, Forge A, Evans WH. Properties of connexin26 gap junctional proteins derived from mutations associated with non-syndromal heriditary deafness. Hum Mol Genet. 1999;8:2369–76. 59. Bicego M, Beltramello M, Melchionda S, Carella M, Piazza V, Zelante L, Bukauskas FF, Arslan E, Cama E, Pantano S, Bruzzone R, D’Andrea P, Mammano F. Pathogenetic role of the deafness-related M34T mutation of Cx26. Hum Mol Genet. 2006;15:2569–87.

20 Inner Ear

433

60. Skerrett IM, Di WL, Kasperek EM, Kelsell DP, Nicholson BJ. Aberrant gating, but a normal expression pattern, underlies the recessive phenotype of the deafness mutant Connexin26M34T. FASEB J. 2004;18:860–2. 61. Wang HL, Chang WT, Li AH, Wu CY, Chen MS, Huang PC. Functional analysis of connexin-26 mutants associated with hereditary recessive deafness. J Neurochem. 2003;8:735–42. 62. Zhang Y, Tang W, Ahmad S, Sipp JA, Chen P, Lin X. Gap junction-mediated intercellular biochemical coupling in cochlear supporting cells is required for normal cochlear functions. Proc Natl Acad Sci USA. 2005;102:15201–6. 63. Xia CH, Cheung D, DeRosa AM, Chang B, Lo WK, White TW, Gong X. Knock-in of 3 connexin prevents severe cataracts caused by an 8 point mutation. J Cell Sci. 2006;119:2138–44. 64. Hu¨lser DF, Rutz ML, Eckert R, Traub O. Functional rescue of defective mutant connexons by pairing with wildtype connexons. Pflu¨gers Arch. 2001;441:521–8. 65. Marziano NK, Casalotti SO, Portelli AE, Becker DL, Forge A. Mutations in the gene for connexin 26 (GJB2) that cause hearing loss have a dominant-negative effect on connexin 30. Hum Mol Genet. 2003;12:805–12. 66. Feldmann D, Denoyelle F, Blons H, Lyonnet S, Loundon N, Rouillon I, Hadj-Rabia S, Petit C, Couderc R, Garabedian EN, Marlin S. The GJB2 mutation R75Q can cause nonsyndromic hearing loss DFNA3 or hereditary palmoplantar keratoderma with deafness. Am J Med Genet A. 2005;137:225–7. 67. Denoyelle F, Lina-Granade G, Plauchu H, Bruzzone R, Chaib H, Levi-Acobas F, Weil D, Petit C. Connexin 26 gene linked to a dominant deafness. Nature 1998;393:319–20. 68. Piazza V, Beltramello M, Menniti M, Colao E, Malatesta P, Argento R, Chiarella G, Gallo LV, Catalano M, Perrotti N, Mammano F, Cassandro E. Functional analysis of R75Q mutation in the gene coding for Connexin 26 identified in a family with nonsyndromic hearing loss. Clin Genet. 2005;68:161–6. 69. Deng Y, Chen Y, Reuss L, Altenberg GA. Mutations of connexin 26 at position 75 and dominant deafness: essential role of arginine for the generation of functional gapjunctional channels. Hear Res. 2006;220:87–94. 70. Chen Y, Deng Y, Bao X, Reuss L, Altenberg GA. Mechanism of the defect in gapjunctional communication by expression of a connexin 26 mutant associated with dominant deafness. FASEB J. 2005;19:1516–8. 71. Thomas T, Telford D, Laird DW. Functional domain mapping and selective transdominant effects exhibited by Cx26 disease-causing mutations. J Biol Chem. 2004;279:19157–68. 72. Richard G, White TW, Smith LE, Bailey RA, Compton JG, Paul DL, Bale SJ. Functional defects of Cx26 resulting from a heterozygous missense mutation in a family with dominant deaf-mutism and palmoplantar keratoderma. Hum Genet. 1998;103:393–9. 73. Maestrini E, Korge BP, Ocana-Sierra J, Calzolari E, Cambiaghi S, Scudder PM, Hovnanian A, Monaco AP, Munro CS. A missense mutation in connexin26, D66H, causes mutilating keratoderma with sensorineural deafness (Vohwinkel’s syndrome) in three unrelated families. Hum Mol Genet. 1999;8:1237–43. 74. Richard G, Rouan F, Willoughby CE, Brown N, Chung P, Ryynanen M, Jabs EW, Bale SJ, DiGiovanna JJ, Uitto J, Russell L. Missense mutations in GJB2 encoding connexin-26 cause the ectodermal dysplasia keratitis-ichthyosis-deafness syndrome. Am J Hum Genet. 2002;70:1341–8. 75. Heathcote K, Syrris P, Carter ND, Patton MA. A connexin 26 mutation causes a syndrome of sensorineural hearing loss and palmoplantar hyperkeratosis (MIM 148350). J Med Genet. 2000;37:50–1. 76. Uyguner O, Tukel T, Baykal C, Eris H, Emiroglu M, Hafiz G, Ghanbari A, Baserer N, Yuksel-Apak M, Wollnik B. The novel R75Q mutation in the GJB2 gene causes

434

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

R. Nickel et al. autosomal dominant hearing loss and palmoplantar keratoderma in a Turkish family. Clin Genet. 2002;62:306–9. del Castillo I, Villamar M, Moreno-Pelayo MA, del Castillo FJ, Alvarez A, Telleria D, Menendez I, Moreno F. A deletion involving the connexin 30 gene in nonsyndromic hearing impairment. N Engl J Med. 2002;346:243–9. Common JE, Bitner-Glindzicz M, O’Toole EA, Barnes MR, Jenkins L, Forge A, Kelsell DP. Specific loss of connexin 26 expression in ductal sweat gland epithelium associated with the deletion mutation del(GJB6-D13S1830). Clin Exp Dermatol. 2005;30:688–93. Common JE, Di WL, Davies D, Galvin H, Leigh IM, O’Toole EA, Kelsell DP. Cellular mechanisms of mutant connexins in skin disease and hearing loss. Cell Comm Adhes. 2003;10:347–51. Liu XZ, Xia XJ, Xu LR, Pandya A, Liang CY, Blanton SH, Brown SD, Steel KP, Nance WE. Mutations in connexin31 underlie recessive as well as dominant nonsyndromic hearing loss. Hum Mol Genet. 2000;9:63–7. He LQ, Liu Y, Cai F, Tan ZP, Pan Q, Liang DS, Long ZG, Wu LQ, Huang LQ, Dai HP, Xia K, Xia JH, Zhang ZH. Intracellular distribution, assembly and effect of diseaseassociated connexin 31 mutants in HeLa cells. Acta Biochim Biophys Sin. 2005;37:547–54. Kudo T, Kure S, Ikeda K, Xia AP, Katori Y, Suzuki M, Kojima K, Ichinohe A, Suzuki Y, Aoki Y, Kobayashi T, Matsubara Y. Transgenic expression of a dominant-negative connexin26 causes degeneration of the organ of Corti and nonsyndromic deafness. Hum Mol Genet. 2003;12:995–1004. Plum A, Winterhager E, Pesch J, Hallas G, Rosentreter B, Traub O, Herberhold C, Willecke K. Connexin31-deficiency in mice causes transient placental dysmorphogenesis but does not impair hearing and skin differentiation. Dev Biol. 2001;231:334–47. Elfgang C, Eckert R, Lichtenberg-Frate H, Butterweck A, Traub O, Klein RA, Hu¨lser DF, Willecke K. Specific permeability and selective formation of gap junction channels in connexin-transfected HeLa cells. J Cell Biol. 1995;129:805–17. Abrams CK, Freidin MM, Verselis VK, Bargiello TA, Kelsell DP, Richard G, Bennett MV, Bukauskas FF. Properties of human connexin 31, which is implicated in hereditary dermatological disease and deafness. Proc Nat Acad Sci USA. 2006;103:5213–8. Ahmad S, Tang W, Chang Q, Qu Y, Hibshman J, Li Y, So¨hl G, Willecke K, Chen P, Lin X. Restoration of connexin26 protein level in the cochlea completely rescues hearing in a mouse model of human connexin30-linked deafness. Proc Nat Acad Sci USA. 2007;104:1337–41.

Chapter 21

Connexins in the Heart Nicholas J. Severs

Abstract Gap junctions form the cell–cell pathways for propagation of the precisely orchestrated patterns of current flow that govern the regular rhythm of the healthy heart. As in most tissues and organs, multiple connexin types are expressed in the heart; Cx43, Cx40, and Cx45 are found in distinctive combinations and relative quantities in different, functionally specialized subsets of cardiac myocyte. Cardiac development is marked by major alterations in the spatiotemporal patterns of expression of these connexins, and studies on transgenic animals, in which specific connexins are ablated or substituted, help differentiate their roles in regional tuning of impulse propagation and in morphogenesis. Mutations in genes that encode connexins have only rarely been identified as a cause of human cardiac disease, but remodeling of connexin expression and gap junction organization are well documented in acquired adult heart disease, notably coronary heart disease. Remodeling may take the form of alterations in the distribution of gap junctions and the amount and type of connexin(s) expressed. The evidence that these alterations can contribute to the development of arrhythmia in the diseased human heart is now substantial. It is well established that heterogeneous reduction in Cx43 expression and disordering in gap junction distribution occur in defined forms of ventricular disease and correlate with electrophysiologically identified arrhythmic changes and contractile dysfunction in animal models. Features of gap junction organization and connexin expression may also contribute to the most common form of atrial arrhythmia, atrial fibrillation, though conclusions in this area are less well founded. A major task ahead is to define the precise functional properties conferred by the distinctive connexin coexpression patterns of different myocyte types in health and disease. Keywords Heart Myocytes Arrhythmia Coronary heart disease Cx40 Cx43 Cx45 N.J. Severs (*) Cardiac Medicine, National Heart and Lung Institute, Imperial College, Guy Scadding Building, Dovehouse Street, London SW3 6LY, United Kingdom e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_21, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

435

436

N.J. Severs

21.1 Introduction The heart contracts constantly, without conscious effort, three billion times or more in a typical human life span. For the left ventricle, this requires synchronization of the beating activity of some three billion myocytes. Contraction of these myocytes is triggered by action potentials that originate from within the heart, in a region of tissue called the sinoatrial (SA) node located in the right atrium near the entry point of the vena cava. The gap junctions and their component connexins, which link together all the individual myocytes of the heart, form the pathways that carry these electrical signals in a precisely orchestrated manner, ensuring synchronization of myocyte contraction at the level of the cardiac chamber and coordination of chamber contraction at the level of the organ. After its initiation in the SA node, the action potential travels from myocyte to myocyte across the atria, causing their contraction and expulsion of blood into the ventricles. The action potential is then conveyed through a specialized conduction system, comprising the atrioventricular (AV) node, the His bundle, the right and left main bundle branches and Purkinje fibers, to the contractile working ventricular myocardium. Slowing of the action potential upon its arrival at the AV node ensures that the ventricles are timed to contract at precisely the right time after they have been filled with blood by the action of the atria. Apart from propagation of the electrical signal for contraction, cardiac myocyte gap junctions have the potential to act as pathways for the direct passage of signaling molecules and ions from cell to cell, as in nonexcitable tissues. That they may do so during cardiac morphogenesis is suggested by the appearance of developmental malformations when certain connexin types are lacking, though the identity of such putative signaling molecules is unknown. The nature and extent of gap junctional nonelectrical signaling in the mature heart, and whether isolated gap junction hemichannels in the plasma membrane play significant functional roles, are also unresolved.

21.2 Expression in Cardiomyocytes of the Normal Adult Heart That three principal connexins — Cx43, Cx40, and Cx45 — are expressed in cardiac myocytes is thoroughly documented. Although Cx43 predominates in the heart as a whole, it is typically coexpressed in characteristic combinations and relative quantities with Cx40 or Cx45 in a chamber-related, myocyte typespecific and developmentally regulated manner (Fig. 21.1) [1,2,3,4]. The working (contractile) myocytes of the ventricle are irregular, rod-shaped cells that are extensively interconnected by clusters of Cx43-containing gap junctions (Fig. 21.2). The gap junctions vary in size (Fig. 21.3) and are organized, together with two types of mechanical junction, fascia adherens

21 Cardiac Tissue

437

Fig. 21.1 Typical patterns of connexin expression in the different chambers and myocyte types of the adult mammalian heart. This is a generalized summary, and species-related variations occur. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

Fig. 21.2 Features of Cx43 gap junction organization in ventricular myocytes, as seen by immunoconfocal microscopy. (a) Isolated ventricular myocyte; note clusters of immunostained gap junctions at the ends of the main body of the cell and at the end of side branches that terminate at various points along the cell length. These are the sites of the intercalated disks. (b) In longitudinal sections of intact ventricular myocardium, the clusters of Cx43 gap junctions are seen as series of linear arrays, marking the sites of the intercalated disks connecting the adjoining cells. (c) In transverse sections of myocardium, the intercalated disk clusters of CX43 gap junctions are seen in en face view. The gap junctions at the periphery of the intercalated disk are often larger than those in the interior region. (a,b) From rat myocardium. (c) From human myocardium. Scale bar 25 mm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

438

N.J. Severs

Fig. 21.3 Typical views of cardiac gap junctions as seen by freeze-fracture electron microscopy. These views show the E-face of the uppermost membrane of the gap junction; the upper cell has been fractured away together with the half-membrane leaflet nearest the viewer, to leave for viewing the apposing half-membrane leaflet attached to the extracellular space, that is, the gap between the junctional membranes. Hemichannels have been plucked out of this half-membrane leaflet, leaving quasi-crystalline arrays of pits. Gap junctions vary enormously in size, from tens of thousands of hemichannels to fewer than ten. (a) A gap junction of average size from a longitudinal segment of disk membrane (see Fig. 21.4). (b) A small gap junction, typical of those intercalated within fascia adherens-containing zones. Both these examples come from glutaraldehyde-fixed and glycerol-cryoprotected samples in which the hemichannels (and their corresponding pits) appear as multiple small hexagonal arrays separated by raised lipid islands. (c) Structure of the gap junction in a beating heart, frozen directly from the living state. The individual arrays tend to merge and intervening lipid areas are not raised (from rat ventricular myocardium). Scale bars 100 nm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

junctions and desmosomes, at structures called intercalated disks (Fig. 21.4). The intercalated disks join adjacent myocytes in a step-like structure so that the myofibrils (contractile apparatus) of end-to-end abutting cells are mechanically coupled via the fascia adherens junctions. The gap junctions typically occupy intervening portions of the disk steps that lie parallel to the long axis of the cell (Fig. 21.4), with larger gap junctions at the disk periphery

21 Cardiac Tissue

439

Fig. 21.4 Thin-section electron microscopy illustrating the organization and structure of gap junctions at the intercalated disk. (a) The paired membranes of the disk present the appearance of a set of irregular steps (dashes), with the fascia adherens (fa) junctions occupying the vertical segments, and the gap junctions (gj) in the longitudinal zones. (b) The gap junctions in the longitudinal zones are more clearly discernible at higher magnification, alternating with fascia adherens junctions. Inset: A pentalaminar structure of the gap junction, representing the two closely apposed plasma membrane domains, is visible at high magnification (from rat ventricle). Scale bars 1 mm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

(Fig. 21.2c) [5,6,7]. Gap junction organization, together with features of tissue architecture such as the size and shape of the cells, combine to determine the pattern of impulse propagation such that it is faster in the direction of the long axis of the cells than in the transverse axis (anisotropic propagation). Atrial myocytes are slender cells compared with their ventricular counterparts, and thus have shorter, less elaborate intercalated disks. In most mammalian species, including humans, atrial myocytes coexpress Cx43 with Cx40 (Fig. 21.5), colocalized within the same gap junctional plaque [2,3]. Adult working ventricular myocytes, by contrast, lack detectable Cx40. In both ventricular and atrial working myocardium, Cx45 is typically present in low quantities, with higher levels in the atria than the ventricles [3].

440

N.J. Severs

Fig. 21.5 Immunogold thin-section electron microscopy demonstrating presence of both CX40 and CX43 in human atrial myocyte gap junctions, and CX40 in Purkinje myocyte gap junctions. Scale bar 100 nm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

The myocytes of the impulse generation and conduction systems are quite distinct from those of the contractile ventricular and atrial cells, both morphologically [8] and with respect to their connexin expression profiles. Sinoatrial and AV nodal myocytes have small, dispersed gap junctions composed of Cx45 and, in the case of the mouse, also Cx30.2 [9,10,11,12], each of which (in homomeric, homotypic configuration in vitro) forms low conductance channels [13,14]. These gap junction features of nodal myocytes suggest relatively poor coupling, which in the AV node may contribute to slowing of impulse propagation and thus ensure sequential contraction of the cardiac chambers. In the rabbit SA node, a limited zone in which Cx45 and Cx43 are coexpressed has been postulated as the exit route for the impulse into the atrial tissue [9]. In the AV nodal region of this species, three-dimensional reconstructions reveal complex compartmentalized connexin expression patterns, with the compact node and transitional cells predominantly expressing Cx45, and the His bundle, lower nodal cells, and posterior nodal extension coexpressing Cx45 with Cx43 [15]. Downstream from the His bundle, the conduction system myocytes of most mammals, including humans, prominently express Cx40 (Fig. 21.5c) [16,17,18,19,20]. In rodents, Cx45 is coexpressed with Cx40 in a central zone of the bundle branches and Purkinje fibers, enveloped by an outer zone in which only Cx45 is found [16,17,18,19,20,21]. Apart from Cx40 and Cx45, distal parts of the conduction system also express Cx43 [1,16]. In general, different mammalian species share the key features of connexin expression described above. It should be noted, however, that not all reports are consistent. A number of connexins other than those discussed above have been attributed to cardiac myocytes, but details remain fragmentary and unconfirmed. Some differences, such as apparent lack of Cx40 expression in rat atrial muscle

21 Cardiac Tissue

441

and in guinea pig conduction system [1, 22], appear to be due to species variation. Others, for example, those relating to connexin patterns in subregions of the conduction system, may arise, in some instances, from misinterpretation of this complex structure. Whether CX31.9 occurs in the human AV node, in a similar manner to its mouse ortholog, Cx30.2, is unknown. It has been speculated that species variations in patterns of coexpression within the AV node between large and small mammals may be related to the greater need for impulse delay in the latter [23]. Valid extrapolation of functional data from transgenic mice to the human requires a more detailed comparative knowledge of species-specific patterns of cardiac connexin expression than currently available.

21.3 Cardiac Development Marked alterations in the spatiotemporal patterns of expression of Cx43, Cx40, and Cx45 take place during cardiac development [24,25,26]. These have been most comprehensively documented in the mouse. The first connexin to be detected during murine development is Cx45; its appearance coincides with the onset of cardiac contractions at embryonic day (ED) 8.5. Cx45 shows widespread expression throughout the heart in the early stages of development, but is particularly prominent (and initially the sole connexin) in the early stages of formation of the conduction system [19]. From about ED 11 or 12, the Cx45 level declines except in the conduction system. Cx43 and Cx40 are first detectable about one day after the initial appearance of Cx45. Cx43 is detected first in the ventricles and, by about ED 12.5, in the atria. Concurrently, Cx40 becomes prominent in the atria and the inner trabeculated parts of the ventricle. From about ED 14 or 15, Cx40 levels decline in the ventricles except in the forming conduction system, while the expression of Cx43 progressively increases during the course of embryonic development and into early postnatal development, remaining conspicuously absent from the proximal regions of the conduction system. Only limited information is available on the developmental aspects of connexins in other species; from the fragmentary evidence available, some of the expression patterns in human development parallel those in mouse [27]. The changing patterns of connexin expression during embryonic development appear to be related to optimization of contraction during a period in which major alterations in structure are taking place, the laying down and patterning of the electrical coupling pathways required for the mature heart, and, more speculatively, the formation of communication compartments and morphogen gradients related to the developmental process itself. After birth, there is a progressive reorganization of ventricular gap junction distribution from a dispersed distribution into clusters at the intercalated disks [28,29,30] as the myocytes adopt mature morphological characteristics. In this process, polarization of the mechanical junctions to the disk precedes that of the gap junctions [28].

442

N.J. Severs

21.4 Transgenic Mice: Essentials Numerous studies have investigated the role of connexins in cardiac function using transgenic animals in which expression of specific connexins is ablated, or one connexin type is substituted for another. These studies highlight roles for connexins both in impulse propagation and cardiac morphogenesis [31]. The presence of Cx45 is critical for cardiac development. Knockout (KO) of this connexin in mice (i.e., germline ablation of both alleles of the Cx45 gene) results in death at about ED 10 from contractile failure attributed to conduction block [32,33]. There are also developmental abnormalities in these animals, notably defects of the endocardial cushion and looping, as well as in vascular development. The cushion defects are reportedly avoided when Cx45KO is restricted to the myocardium [34]. Cx40KO can also lead to developmental malformations (atrioventricular septation defects, double outflow right ventricle, tetralogy of Fallot) in a proportion of the mice, though many survive to birth [35,36,37,38]. The surviving animals show impairment of impulse propagation through the conduction system [38,39,40,41]. Transgenic mice in which Cx40 is replaced by Cx45 show slow conduction in the right bundle branch of the conduction system, but normal conduction velocities elsewhere [42]. Knockout of Cx30.2, by contrast, leads to more rapid conduction through the AV node, indicating that though Cx45 can maintain AV nodal conduction, expression of both these connexins in the mouse AV node appears necessary to achieve the correct delay between atrial and ventricular activation [12]. Knockout of Cx43 demonstrates that this connexin is not required by myocytes for cardiac development. The mice survive to term but die shortly after birth owing to obstruction of the right ventricular outflow tract [43], a nonmyocyte developmental abnormality attributed to defective migration of the cardiac neural crest cells destined to form the outflow tract septation complex [44,45,46]. Recent studies suggest that rather than an effect on cardiac neural crest cells, it is lack of Cx43 in noncrest neuroepithelial cells that gives rise to the developmental defect [47,48]. Replacement of Cx43 by Cx40 or Cx32 rescues the otherwise postnatal lethality of Cx43KO. The overall features of the electrocardiogram in these knockout/knock-in mice after birth are largely similar to those of wild-type animals, though there is an increased arrhythmic tendency [49]. These findings indicate that one connexin type may substitute functionally for another to a quite remarkable extent, though not perfectly; specific connexins confer subtly distinctive properties [49]. The effects of lack of Cx43 in the myocardium of the normally formed heart after birth have been determined by cardiac-restricted Cx43KO [50]; these mice demonstrate marked reduction of ventricular conduction velocity and lethal spontaneous ventricular arrhythmia by two months of age. Thus, Cx43 is essential for normal impulse propagation in the more mature ventricle but not in the early neonatal ventricle where low amounts of Cx45 are still present. Cardiac-

21 Cardiac Tissue

443

specific knockout of mechanical junction proteins emphasizes functional interdependence between junction types; for example, depletion of N-cadherin, the adhesive protein that binds adjoining cells at fascia adherens junctions, leads to reduced Cx43 levels and decreased conduction velocity [51].

21.5 Heart Disease 21.5.1 Genetic Defects Mutations of the CX43 gene involving base substitutions corresponding to sites in the carboxyl-terminal domain that can be phosphorylated were reported in a subset of patients with complex cardiac malformations (outflow tract anomalies and visceroatrial heterotaxia; failure to establish normal left/right asymmetry [52]). Other reports examining different patient groups failed to confirm this link [53,54,55]. However, these studies did not examine cardiac tissue, suggesting somatic mutation may have been responsible. Dominant mutations in other parts of the CX43 gene are associated with oculodentodigital dysplasia (ODDD), a developmental abnormality affecting the limbs, teeth, face, and eyes but only rarely the heart [56,57,58]. Chromosomal deletion that includes the CX40 gene has been reported in a minority of cases (three of 505) of congenital heart disease involving anomalies of the aortic arch [59]. In addition, the atrial arrhythmia, atrial fibrillation, appears in some instances to be associated with heterozygous somatic missense mutations in the CX40 gene and polymorphisms within the gene’s regulatory region [60,61].

21.5.2 Acquired Adult Heart Disease Coronary heart disease is the leading cause of death and disability in most industrialized countries of the developed and developing world [62], and disturbances of the cardiac rhythm (arrhythmias) are a common, serious, and often fatal complication of this and other forms of heart disease. In view of the role of gap junctions in mediating the patterns of impulse flow that govern orderly contraction of the healthy heart, a key question is whether alterations in connexin expression or gap junction organization (remodeling) contribute to abnormal impulse propagation and the arrhythmia in human heart disease [63,64,65,66]. The underlying cause of coronary heart disease is atherosclerosis of the coronary arteries. The resultant narrowing of the arteries may give rise to mild transient cardiac ischemia on exercise (stable angina), or, where an arterial lesion ruptures and thrombosis ensues, occlusion of the artery, severe cardiac ischemia, and myocardial infarction. In those patients who survive, extensive or repeated myocardial infarction may lead to congestive heart failure.

444

N.J. Severs

21.5.2.1 Ventricular Disease Effects of Cardiac Ischemia and Infarction in Animal Models The sudden occlusion of a coronary artery causes acute focal ischemia, and the ensuing ionic and metabolic alterations result in changes to impulse formation and propagation. In particular, a decrease in the action potential upstroke velocity, a transient increase followed by slowing of conduction and alterations to the refractory period, will occur, events that lead to localized conduction block and reentry arrhythmia (self-perpetuating ‘‘circus’’ movements of the impulse). These changes are not uniform across the ischemic zone, and so ectopic foci, in which the impulse is generated at abnormal sites within the ventricle itself, emerge. Although alterations to the action potential play a large part in these processes, the drop in pH brought about by ischemia contributes to electrical uncoupling of gap junctions. Prior to uncoupling, Cx43 is dephosphorylated. The adverse effects of ischemia on Cx43 gap junctions are substantially reduced by ischemic preconditioning, that is, administration of brief, repetitive ischemic episodes that help protect the heart from a subsequent more severe ischemic insult [67,68,69,70]. The benefit of preconditioning is reportedly abolished in transgenic mice expressing half the normal level of Cx43 [71], yet coronary occlusion in these mice reportedly leads to smaller infarcts than in their wild-type counterparts [72]. The cell swelling that results from ischemia has been attributed to opening of nonjunctional Cx43 hemichannels [73], but this requires further substantiation. Spread of ischemic injury from myocyte to myocyte (rigor contracture and cell death) appears to involve gap junction–mediated passage of chemical signals [74]. Gap junction disarray with lateralization (i.e., dispersed distribution, distant from the intercalated disks) occurs rapidly in response to experimentally induced ventricular ischemia and infarction in the rat [75]. It is also conspicuous in some rat models of ventricular hypertrophy [76,77], where it correlates with reduced longitudinal conduction velocity [77]. Longer term changes reported in myocardium distant from the infarct in the canine ventricle include reduction in the size and the number of gap junctions per unit length of intercalated disk, and fewer side-to-side connections between myocytes [78]. Altered Distribution of Connexin43 Gap Junctions in Diseased Human Ventricle As in animal models of infarction, disturbed arrangements of gap junctions are a prominent feature of the border zone of surviving cells around zones of myocardial infarct scar tissue in the human ventricle (Fig. 21.6). Both laterally disposed gap junctions connecting adjacent myocytes, and internalized (nonfunctional) gap junctional membrane, contribute to this abnormal pattern [79]. Gap junction remodeling around regions of focal ischemia may allow current to bypass the ischemic region effectively, thus helping to preserve cardiac function.

21 Cardiac Tissue

445

Fig. 21.6 Immunoconfocal microscopy of human ventricular myocardium from a patient with end-stage ischemic heart disease. CX43 gap junctions are seen in the surviving myocardium around the infarct scar, but the infarct scar itself is immunonegative for CX43. Inset: Detail of gap junction staining from myocytes bordering the infarct scar; note that the immunolabel is dispersed over the lateral borders of the cells, rather than showing the normal polarization at intercalated disks. Scale bars 100 mm. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

However, the conduction delays resulting from rerouting of the current may increase the risk of reentry circuits, leading to a lower threshold for arrhythmia in the longer term. Smaller areas of gap junction disarray than those associated with infarcts are also found in end-stage heart failure due to idiopathic dilated cardiomyopathy and myocarditis [80], and in the ventricles of patients with compensated hypertrophy due to valvular aortic stenosis [81]. In decompensated hypertrophy from the same cause, a different abnormality is apparent; patches in which the otherwise normally arrayed gap junctions are fewer or absent [82]. Particularly disordered CX43 gap junction arrangements go hand in hand with the haphazard myocyte organization that typifies human hypertrophic cardiomyopathy [83]. Yet another form of structural remodeling is associated with a condition termed hibernating myocardium in patients with ischemic heart disease [83]. Hibernating myocardium essentially means ventricular myocardium that does not contract properly but which recovers its contractile function after normal blood flow is restored by coronary artery bypass surgery. In human hibernating myocardium, the large CX43 gap junctions typically found at the periphery of the intercalated disk are smaller in size, and the overall amount of immunodetectable CX43 per intercalated disk is reduced, compared with normally perfused myocardial regions of the same heart [83]. These observations emphasize how CX43 gap junction remodeling may be linked to impaired ventricular contraction (in addition to arrhythmia) in the heart afflicted with coronary artery disease [83].

446

N.J. Severs

Reduction in Connexin43 Levels in Diseased Human Ventricle Apart from disturbances in the distribution of gap junctions, reduction in CX43 transcript and protein levels have been consistently reported in the left ventricles of transplant patients with end-stage congestive heart failure regardless of whether the condition is due to ischemic (i.e., coronary) heart disease, idiopathic dilated cardiomyopathy, valvular aortic stenosis, or other etiologies [79,80,83,84,85]. The reduction in CX43 is spatially heterogeneous and develops progressively during the course of disease, as indicated by the pattern of change observed in pressure-overloaded hearts with valvular aortic stenosis [81] and its presence in nonfailing ventricles of patients with ischemic heart disease [86].

Alteration in Expression of Connexin45 and Connexin40 in Diseased Human Ventricle Apart from reduced CX43, elevated levels of CX45 [87] and CX40 (specifically in ischemic heart disease [84]) have been reported in the failing human ventricle. These findings require further verification in view of the potentially profound functional effects mediated by altered ratios of coexpression (see below). Major questions concern the mechanisms by which CX43 is downregulated in the diseased heart, whether this and other features of gap junction distribution hold any adverse functional consequence, and how, in the face of changing expression levels, the presence of more than one connexin influences function.

Regulation of Connexin Expression Levels Elucidation of the structure of the genes encoding Cx43, Cx40 and Cx45 forms the basis for understanding how transcription factors interact with their target elements to control transcript expression during cardiac disease and development [88]. Posttranscriptional regulatory mechanisms operate at the levels of connexin synthesis, assembly into hemichannels and gap junction plaques, and degradation. The extracellular signaling pathways leading to altered connexin expression in disease may be triggered in part by mechanical forces, and involve cyclic adenosine monophosphate (cAMP), angiotensin II, and growth factors such as vascular endothelial growth factor, activated via protein kinases such as focal adhesion kinase and c-Jun N-terminal kinase (JNK) [89,90]. JNK activation in a transgenic mouse model leads to an impressive combination of loss of Cx43, slowing of ventricular conduction, contractile dysfunction, and congestive heart failure [91]. The involvement of connexin-interacting proteins as regulatory elements is currently attracting interest [92]. For example, binding of zonula occludens-1 to Cx43 limits the size of gap junctions [93], while deficiencies of plakoglobin, desmin, and N-cadherin (all adherens junction proteins) result in reduced connexin expression, gap junction remodeling and arrhythmia in mouse and humans [94,95,96,97,98,99] (see Chapter 11).

21 Cardiac Tissue

447

Does Reduced Connexin43 Contribute to Development of Arrhythmia? Whether the reduced ventricular CX43 levels found in human heart disease contribute to arrhythmia has been controversial. As noted earlier, key conditions that lead to reentry arrhythmia are slowed conduction and conduction block. In assessing the possible significance of changes in overall connexin levels, it needs to be borne in mind that the fraction of connexin assembled into gap junction plaques, the types of connexin coexpressed, and regulatory factors affecting channel gating will all contribute to gap junction coupling. It is simplistic and plainly erroneous to assume that a reduction in Cx43 will automatically lead to slowing of impulse propagation and thereby inexorably precipitate reentry arrhythmia. The mammalian heart has a considerable surfeit of gap junctions; nonmammalian chordates sustain cardiac function with just a tiny quantity of gap junctions by comparison. Computer modeling predicts that even substantial reductions in gap junction content in mammalian heart would have no effect on propagation velocity [100,101]. In the cardiac restricted Cx43KO mouse, a Cx43 reduction in the order of 90% is required to precipitate acute lethal spontaneous ventricular arrhythmia [50]. Disease-related alterations to gap junctions and connexins represent one facet of a constellation of interacting factors that may influence susceptibility to rhythm disturbance. Arrhythmias are multifactorial in origin, involving an interplay between gap junction coupling, membrane excitability, and cell and tissue architecture [102,103,104,105]. The importance of this interplay is emphasized by observations that action potential propagation can fail in wellcoupled cells if these form a large mass (‘‘sink’’) receiving a limited amount of depolarizing current from a smaller source (i.e., a source/sink mismatch) but the conduction block can be overcome by reducing rather than increasing coupling in the sink [106,107]. In theory, then, a drastic reduction of Cx43 could, by reducing coupling in the diseased ventricle, form part of a protective response that increases the safety of conduction. With all these points in mind, a strong case that disease-related gap junction and connexin remodeling contributes to the arrhythmic substrate can nevertheless be made. In intact isolated hearts of transgenic mice expressing half the normal level of Cx43, experimental ischemia reportedly leads to a marked increase in incidence, frequency, and duration of ventricular tachycardias [108]. Importantly, the CX43 reduction seen in the failing human ventricle is not uniform; a considerable variation in the level of CX43 is found, with some regions of some diseased hearts reaching a reduction of >90% of control values [84] (similar to the levels at which fatal cardiac arrhythmia occurs in the cardiac-restricted Cx43KO mouse). Thus, average values for the overall reduction in ventricular CX43 in the diseased human heart disguise considerable spatial heterogeneity in the extent of the reduction. The critical importance of heterogeneity in Cx43 expression to disturbances in electromechanical function has been emphasized in a mouse model generated to give patches of myocardium lacking Cx43 [109]. These experimental mice show

448

N.J. Severs

abnormal impulse propagation and contractile dysfunction [109], in keeping with hypotheses generated from the human studies [78,82,83]. A further refinement has been the development of outbred cardiac-restricted knockouts, which show a more gradual decline in Cx43 levels than the original cardiac-restricted knockout animals [110]. In these outbred animals, reduction in Cx43 to 18% of control levels is associated with a 50% slowing of propagation velocity, and 80% of the animals are inducible into lethal ventricular arrhythmias [110]. A key factor in this arrhythmic susceptibility is that deficiency of ventricular Cx43 and reduced coupling in the ventricle alters the source/sink relationship at the junction with the Purkinje fibers (which, with unaltered levels of Cx40 and Cx45 themselves may maintain normal coupling); junctions that are normally quiescent become activated, leading to wave-front collisions in the ventricular muscle [111]. Taking all these findings together, what can be stated with certainty? Remodeling of ventricular myocyte gap junctions and connexins, notably disordering in the pattern of junctional distribution and heterogeneously reduced levels of Cx43, do occur in certain categories of human heart disease and have been correlated with electrophysiologically identified proarrhythmic changes in animal models. Definitive proof of direct cause and effect in the human heart is, for methodological reasons, difficult to obtain, but the notion that gap junction/ connexin abnormalities act as one of a set of interacting factors contributing to arrhythmogenic substrates has gained wide currency.

21.5.2.2 Atrial Disease Atrial fibrillation, an arrhythmia in which wavelets of electrical activity propagate in multiple directions, leads to progressive electrical and structural remodeling of the atrium, exacerbating the condition such that it often becomes chronic. There is now a considerable literature on connexins and gap junctions in atrial fibrillation but, in contrast to the data on ventricular dysfunction, the picture remains somewhat confusing, with a range of disparate and seemingly contradictory findings [112,113,114,115,116,117,118,119,120,121]. A variety of factors underlie the reported disparities, including species and age differences, manner of induction of atrial fibrillation, different clinical subsets of patients and associated pathological factors, and flawed techniques or interpretation. A separate aspect from established atrial fibrillation is whether preexisting features of gap junction organization or connexin expression predispose to initiation of the arrhythmia. The initiating foci are in some instances situated in proximal portions of the thoracic veins that have a sleeve of myocardium continuous with that of the atria, and heterogeneous patterns of connexin expression and gap junction organization have been identified at these sites [122]. Atrial fibrillation is also common after coronary bypass surgery, and risk of developing this complication is positively correlated with the level of CX40 prior to onset [123].

21 Cardiac Tissue

449

21.6 Functional Significance of Connexin Coexpression in Health and Disease The existence of distinctive connexin coexpression patterns in the healthy heart opens the possibility of a wide range of different molecular stoichiometries within heteromeric hemichannels and heterotypic channels. A major unanswered question is how these stoichiometries translate into differentiation in electrophysiological (and permeability) properties between the different myocyte types of the normal adult heart and in the changing expression patterns found in development and disease. Apart from data on transgenic animals, extrapolation of functional properties to intact hearts in vivo is based predominantly on in vitro studies of cell pairs in which gap junction channels are made from a single connexin type. For example, because Cx40 channels have high unitary conductance, the presence of large amounts of Cx40 in the conduction system is supposed to facilitate rapid propagation of the impulse to the Purkinje fibers and into the ventricular myocardium. The reality is that, because Cx40 is coexpressed with Cx45, and in the distal fibers also with Cx43, a range of possible mixes of molecular makeup are created, the properties of which cannot reliably be predicted from data on single connexins expressed in vitro. The potentially profound effects of coexpression of Cx43 and Cx40 are highlighted by studies using cultured atrial cells from homozygous and heterozygous knockout mice. In these coexpressing cells, increasing the ratio of Cx40 to Cx43 has the counterintuitive effect of reducing propagation velocity [124]. This starts to make sense of otherwise surprising findings such as an inverse relationship between Cx40 level and propagation velocity in the human atrium [125], and that higher than average levels of Cx40 places patients at risk of developing postoperative atrial fibrillation [123]. New cell models in which coexpression patterns can be manipulated at will to mimic those in vivo, and which are amenable to analysis of function and connexin make-up of the gap junction channel, promise further progress on this front in the future [65].

21.7 Conclusion We now have substantial knowledge of the role of gap junctions and their component connexins in the heart. We know which connexins are expressed, and where, in the mature heart; how their expression patterns evolve during development; and the nature and functional consequences of remodeling of gap junctions and connexin expression in ventricular disease. We have a good grasp of the function of connexins and gap junctions in the heart from electrophysiological and biophysical studies, and studies on transgenic animals. Many questions remain unanswered, in particular, do gap junctions mediate direct molecular signaling of functional significance between cardiac myocytes of the developing, adult, and diseased

450

N.J. Severs

heart, and, if so, what is the nature and action of the messenger molecules or ions involved? What are the precise functional properties dictated by the distinctive connexin coexpression patterns of different myocyte types in health and disease? Other areas in which our knowledge remains far from complete are the regulatory mechanisms involved in remodeling of gap junctions and connexin expression in disease, whether gap junctions and connexins contribute to chronic atrial fibrillation, and the part played by connexin gene mutations in congenital abnormalities. As a final comment, there is a common tendency to portray results from limited studies as offering the promise of new therapies for cardiac disease. The findings that reduced coupling can increase the safety of conduction, and that increasing the ratio of Cx40 to Cx43 reduces rather than increases conduction velocity, are among many illustrating that, even if proposed therapeutic interventions were to achieve the cellular alteration intended, the functional consequences to the patient may be the opposite of those expected. More sophisticated thinking is needed to underpin translational research. Acknowledgments The author wishes to thank Stephen Rothery for his help in preparing the figures, and the British Heart Foundation for support (grants PG/05/003/18157, PG/05/111 and FS/07/012).

References 1. Van Kempen MJA, Ten Velde I, Wessels A, Oosthoek PW, Gros D, Jongsma HJ, Moorman AF, Lamers WH. Differential connexin distribution accommodates cardiac function in different species. Microsc Res Tech. 1995;31:420–36. 2. Severs NJ, Rothery S, Dupont E, Coppen SR, Yeh HI, Ko YS, Matsushita T, Kaba R, Halliday D. Immunocytochemical analysis of connexin expression in the healthy and diseased cardiovascular system. Microsc Res Tech. 2001;52:301–22. 3. Vozzi C, Dupont E, Coppen SR, Yeh H-I, Severs NJ. Chamber-related differences in connexin expression in the human heart. J Mol Cell Cardiol. 1999;31:991–1003. 4. Gros DB, Alcole´a S, Dupays L, Meysen S, The´veniau-Ruissy M, Teunissen BEJ. Connexins in cardiac development: expression, role and transcriptional control. In: Winterhager E, editor. Gap junctions in development and disease. Berlin: Springer-Verlag; 2005. pp. 29–55. 5. Gourdie RG, Green CR, Severs NJ. Gap junction distribution in adult mammalian myocardium revealed by an antipeptide antibody and laser scanning confocal microscopy. J Cell Sci. 1991;99:41–55. 6. Hoyt RH, Cohen ML, Saffitz JE. Distribution and three-dimensional structure of intercellular junctions in canine myocardium. Circ Res. 1989;64:563–74. 7. Severs NJ. Gap junction shape and orientation at the cardiac intercalated disk. Circ Res. 1989;65:1458–61. 8. Severs NJ. Constituent cells of the heart and isolated cell models in cardiovascular research. In: Piper HM, Isenberg G, editors. Isolated adult cardiomyocytes. Boca Raton, FL: CRC Press; 1989. pp. 3–41. 9. Coppen SR, Kodama I, Boyett MR, Dobrzynski H, Takagishi Y, Honjo H, Yeh HI, Severs NJ. Connexin45, a major connexin of the rabbit sinoatrial node, is co-expressed with connexin43 in a restricted zone at the nodal-crista terminalis border. J Histochem Cytochem. 1999;47:907–18. 10. Coppen SR, Severs NJ, Gourdie RG. Connexin45 (6) expression delineates an extended conduction system in the embryonic and mature rodent heart. Dev Genet. 1999;24:82–90.

21 Cardiac Tissue

451

11. Honjo H, Boyett MR, Coppen SR, Takagishi Y, Severs NJ, Kodama I. Heterogeneous expression of connexins in rabbit sinoatrial node cells: correlation between connexin isoform and cell size. Cardiovasc Res. 2002;50:89–96. 12. Kreuzberg MM, Schrickel JW, Ghanem A, Kim JS, Degen J, Janssen-Bienhold U, Lewalter T, Tiemann K, Willecke K. Connexin30.2 containing gap junction channels decelerate impulse propagation through the atrioventricular node. Proc Natl Acad Sci USA. 2006;103:5959–64. 13. van Veen TA, van Rijen HV, Jongsma HJ. Electrical conductance of mouse connexin45 gap junction channels is modulated by phosphorylation. Cardiovasc Res. 2000;46:496–510. 14. Rackauskas M, Kreuzberg MM, Pranevicius M, Willecke K, Verselis VK, Bukauskas FF. Gating properties of heterotypic gap junction channels formed of connexins 40, 43 and 45. Biophys J. 2007;92:1952–65. 15. Ko YS, Yeh HI, Ko YL, Hsu YC, Chen CF, Wu S, Lee YS, Severs NJ. Three-dimensional reconstruction of the rabbit atrioventricular conduction axis by combining histological, desmin, and connexin mapping data. Circulation. 2004;109:1172–9. 16. Gourdie RG, Severs NJ, Green CR, Rothery S, Germroth P, Thompson RP. The spatial distribution and relative abundance of gap-junctional connexin40 and connexin43 correlate to functional properties of the cardiac atrioventricular conduction system. J Cell Sci. 1993;105:985–91. 17. Bastide B, Neyses L, Ganten D, Paul M, Willecke K, Traub O. Gap junction protein connexin40 is preferentially expressed in vascular endothelium and conductive bundles of rat myocardium and is increased under hypertensive conditions. Circ Res. 1993;73:1138–49. 18. Davis LM, Rodefeld ME, Green K, Beyer EC, Saffitz JE. Gap junction protein phenotypes of the human heart and conduction system. J Cardiovasc Electrophysiol. 1995;6:813–22. 19. Coppen SR, Gourdie RG, Severs NJ. Connexin45 is the first connexin to be expressed in the central conduction system of the mouse heart. Exp Clin Cardiol. 2001;6:17–23. 20. Miquerol L, Meysen S, Mangoni M, Bois P, van Rijen HV, Abran P, Jongsma H, Nargeot J, Gros D. Architectural and functional asymmetry of the His-Purkinje system of the murine heart. Cardiovasc Res. 2004;63:77–86. 21. Coppen SR, Dupont E, Rothery S, Severs NJ. Connexin45 expression is preferentially associated with the ventricular conduction system in mouse and rat heart. Circ Res. 1998;82:232–43. 22. Gros D, Jarry-Guichard T, Ten Velde I, de Maziere A, van Kempen MJ, Davoust J, Briand JP, Moorman AF, Jongsma HJ. Restricted distribution of connexin40, a gap junctional protein, in mammalian heart. Circ Res. 1994;74:839–51. 23. Coppen SR, Severs NJ. Diversity of connexin expression patterns in the atrioventricular node: vestigial consequence or functional specialization? J Cardiovasc Electrophysiol. 2002;13:625–6. 24. Van Kempen MJA, Vermeulen JLM, Moorman AFM, Gros D, Paul DL, Lamers WH. Developmental changes of connexin40 and connexin43 messenger RNA. Cardiovasc Res. 1996;32:886–900. 25. Alcolea S, Theveniau-Ruissy M, Jarry-Guichard T, Marics I, Tzouanacou E, Chauvin JP, Briand JP, Moorman AF, Lamers WH, Gros DB. Downregulation of connexin 45 gene products during mouse heart development. Circ Res. 1999;84:1365–79. 26. Gourdie RG, Green CR, Severs NJ, Thompson RP. Immunolabeling patterns of gap junction connexins in the developing and mature rat heart. Anat Embryol. 1992;185:363–78. 27. Kaba RA, Coppen SR, Dupont E, Skepper JN, Elneil S, Haw MP, Pepper JR, Yacoub MH, Rothery S, Severs NJ. Comparison of connexin 43, 40 and 45 expression patterns in the developing human and mouse hearts. Cell Commun Adhes. 2001;8:339–43. 28. Angst BD, Khan LU, Severs NJ, Whitely K, Rothery S, Thompson RP, Magee AI, Gourdie RG. Dissociated spatial patterning of gap junctions and cell adhesion junctions during postnatal differentiation of ventricular myocardium. Circ Res. 1997;80:88–94. 29. Peters NS, Severs NJ, Rothery SM, Lincoln C, Yacoub MH, Green CR. Spatiotemporal relation between gap junctions and fascia adherens junctions during postnatal development of human ventricular myocardium. Circulation. 1994;90:713–25.

452

N.J. Severs

30. Fromaget C, El Aoumari A, Gros D. Distribution pattern of connexin43, a gap-junctional protein, during the differentiation of mouse heart myocytes. Differentiation. 1992;51:9–20. 31. Gros D, Dupays L, Alcolea S, Meysen S, Miquerol L, Theveniau-Ruissy M. Genetically modified mice: tools to decode the functions of connexins in the heart-new models for cardiovascular research. Cardiovasc Res. 2004;62:299–308. 32. Kumai M, Nishii K, Nakamura K, Takeda N, Suzuki M, Shibata Y. Loss of connexin45 causes a cushion defect in early cardiogenesis. Development 2000;127:3501–512. 33. Kruger O, Plum A, Kim JS, Winterhager E, Maxeiner S, Hallas G, Kirchhoff S, Traub O, Lamers WH, Willecke K. Defective vascular development in connexin 45-deficient mice. Development. 2000;127:4179–93. 34. Nishii K, Kumai M, Egashira K, Miwa T, Hashizume K, Miyano Y, Shibata Y. Mice lacking connexin45 conditionally in cardiac myocytes display embryonic lethality similar to that of germline knockout mice without endocardial cushion defect. Cell Commun Adhes. 2003;10:365–9. 35. Simon AM, Goodenough DA, Paul DL. Mice lacking connexin40 have cardiac conduction abnormalities characteristic of atrioventricular block and bundle branch block. Curr Biol. 1998;8:295–8. 36. Gu H, Smith FC, Taffet SM, Delmar M. High incidence of cardiac malformations in connexin40-deficient mice. Circ Res. 2003;93(3):201–6. 37. Kirchhoff S, Kim JS, Hagendorff A, Thonnissen E, Kruger O, Lamers WH, Willecke K. Abnormal cardiac conduction and morphogenesis in connexin40 and connexin43 doubledeficient mice. Circ Res. 2000;87:399–405. 38. Kirchhoff S, Nelles E, Hagendorff A, Kru¨ger O, Traub O, Willecke K. Reduced cardiac conduction velocity and predisposition to arrhythmias in connexin40-deficient mice. Curr Biol. 1998;8:299–302. 39. Hagendorff A, Schumacher B, Kirchhoff S, Lu¨deritz B, Willecke K. Conduction disturbances and increased atrial vulnerability in connexin40-deficient mice analyzed by transesophageal stimulation. Circulation. 1999;99:1508–15. 40. Tamaddon HS, Vaidya D, Simon AM, Paul DL, Jalife J, Morley GE. High-resolution optical mapping of the right bundle branch in connexin40 knockout mice reveals slow conduction in the specialized conduction system. Circ Res. 2000;87:929–36. 41. van Rijen HV, van Veen TA, van Kempen MJ, Wilms-Schopman FJ, Potse M, Krueger O, Willecke K, Opthof T, Jongsma HJ, de Bakker JM. Impaired conduction in the bundle branches of mouse hearts lacking the gap junction protein connexin40. Circulation. 2001;103:1591–8. 42. Alcolea S, Jarry-Guichard T, de Bakker J, Gonzalez D, Lamers W, Coppen S, Barrio L, Jongsma H, Gros D, van Rijen H. Replacement of connexin40 by connexin45 in the mouse: impact on cardiac electrical conduction. Circ Res. 2004;94:100–9. 43. Reaume AG, de Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science. 1995;267:1831–4. 44. Huang GY, Cooper ES, Waldo K, Kirby ML, Gilula NB, Lo CW. Gap junctionmediated cell-cell communication modulates mouse neural crest migration. J Cell Biol. 1998;143:1725–34. 45. Sullivan R, Huang GY, Meyer RA, Wessels A, Linask KK, Lo CW. Heart malformations in transgenic mice exhibiting dominant-negative inhibition of gap junctional communication in neural crest cells. Dev Biol. 1998;204:224–34. 46. Wei CJ, Xu X, Lo CW. Connexins and cell signaling in development and disease. Annu Rev Cell Dev Biol. 2004;20:811–38. 47. Kretz M, Eckardt D, Kruger O, Kim JS, Maurer J, Theis M, van Rijen HV, Schorle H, Willecke K. Normal embryonic development and cardiac morphogenesis in mice with Wnt1-Cre-mediated deletion of connexin43. Genesis. 2006;44:269–76. 48. Liu S, Liu F, Schneider AE, St Amand T, Epstein JA, Gutstein DE. Distinct cardiac malformations caused by absence of connexin 43 in the neural crest and in the non-crest neural tube. Development. 2006;133:2063–73.

21 Cardiac Tissue

453

49. Plum A, Hallas G, Magin T, Dombrowski F, Hagendorff A, Schumacher B, Wolpert C, Kim J, Lamers WH, Evert M, Meda P, Traub O, Willecke K. Unique and shared functions of different connexins in mice. Curr Biol. 2000;10:1083–91. 50. Gutstein DE, Morley GE, Tamaddon H, Vaidya D, Schneider MD, Chen J, Chien KR, Stuhlmann H, Fishman GI. Conduction slowing and sudden arrhythmic death in mice with cardiac-restricted inactivation of connexin43. Circ Res. 2001;88:333–9. 51. Li J, Patel VV, Kostetskii I, Xiong Y, Chu AF, Jacobson JT, Yu C, Morley GE, Molkentin JD, Radice GL. Cardiac-specific loss of N-cadherin leads to alteration in connexins with conduction slowing and arrhythmogenesis. Circ Res. 2005;97:474–81. 52. Britz-Cunningham SH, Shah MM, Zuppan CW, Fletcher WH. Mutations of the connexin43 gap-junction gene in patients with heart malformations and defects of laterality. N Engl J Med. 1995;332:1323–9. 53. Casey B, Ballabio A. Connexin43 mutations in sporadic and familial defects of laterality. N Engl J Med. 1995;333:941. 54. Debrus S, Tuffery S, Matsuoka R, Galal O, Sarda P, Sauer U, Bozio A, Tanman B, Toutain A, Claustres M, Le Paslier D, Bouvagnet P. Lack of evidence for connexin 43 gene mutations in human autosomal recessive lateralization defects. J Mol Cell Cardiol. 1997;29:1423–31. 55. Gebbia M, Towbin JA, Casey B. Failure to detect connexin43 mutations in 38 cases of sporadic and familial heterotaxy. Circulation. 1996;94:1909–12. 56. Paznekas WA, Boyadjiev SA, Shapiro RE, Daniels O, Wollnik B, Keegan CE, Innis JW, Dinulos MB, Christian C, Hannibal MC, Jabs EW. Connexin 43 (GJA1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet. 2003;72: 408–18. 57. Flenniken AM, Osborne LR, Anderson N, Ciliberti N, Fleming C, Gittens JE, Gong XQ, Kelsey LB, Lounsbury C, Moreno L, Nieman BJ, Peterson K, Qu D, Roscoe W, Shao Q, Tong D, Veitch GI, Voronina I, Vukobradovic I, Wood GA, Zhu Y, Zirngibl RA, Aubin JE, Bai D, Bruneau BG, Grynpas M, Henderson JE, Henkelman RM, McKerlie C, Sled JG, Stanford WL, Laird DW, Kidder GM, Adamson SL, Rossant J. A Gja1 missense mutation in a mouse model of oculodentodigital dysplasia. Development 2005;132:4375–86. 58. Gong XQ, Shao Q, Lounsbury CS, Bai D, Laird DW. Functional characterization of a GJA1 frameshift mutation causing oculodentodigital dysplasia and palmoplantar keratoderma. J Biol Chem. 2006;281:31801–11. 59. Christiansen J, Dyck JD, Elyas BG, Lilley M, Bamforth JS, Hicks M, Sprysak KA, Tomaszewski R, Haase SM, Vicen-Wyhony LM, Somerville MJ. Chromosome 1q21.1 contiguous gene deletion is associated with congenital heart disease. Circ Res. 2004;94:1429–35. 60. Hauer RN, Groenewegen WA, Firouzi M, Ramanna H, Jongsma HJ. Cx40 polymorphism in human atrial fibrillation. Adv Cardiol. 2006;42:284–91. 61. Gollob MH, Jones DL, Krahn AD, Danis L, Gong XQ, Shao Q, Liu X, Veinot JP, Tang AS, Stewart AF, Tesson F, Klein GJ, Yee R, Skanes AC, Guiraudon GM, Ebihara L, Bai D. Somatic mutations in the connexin 40 gene (GJA5) in atrial fibrillation. N Engl J Med. 2006;354:2677–88. 62. Robenek H, Severs NJ. Cell interactions in atherosclerosis. Boca Raton, FL: CRC Press; 1992. 63. Severs NJ, Coppen SR, Dupont E, Yeh HI, Ko YS, Matsushita T. Gap junction alterations in human cardiac disease. Cardiovasc Res. 2004;62(2):368–77. 64. Severs NJ, Dupont E, Coppen SR, Halliday D, Inett E, Baylis D, Rothery S. Remodelling of gap junctions and connexin expression in heart disease. Biochim Biophys Acta. 2004;1662:138–48. 65. Severs NJ, Dupont E, Kaba RA, Thomas N. Gap junction and connexin remodeling in human heart disease. In: Winterhager E, editor. Gap junctions in development and disease. Berlin-Heidelberg: Springer-Verlag; 2005. pp. 57–82. 66. Severs NJ, Dupont E, Thomas N, Kaba R, Rothery S, Jain R, Sharpey K, Fry CH Alterations in cardiac connexin expression in cardiomyopathies. Adv Cardiol. 2006;42:228–42.

454

N.J. Severs

67. Beardslee MA, Lerner DL, Tadros PN, Laing JG, Beyer EC, Yamada KA, Kleber AG, Schuessler RB, Saffitz JE. Dephosphorylation and intracellular redistribution of ventricular connexin43 during electrical uncoupling induced by ischemia. Circ Res. 2000;87:656–62. 68. Jain SK, Schuessler RB, Saffitz JE. Mechanisms of delayed electrical uncoupling induced by ischemic preconditioning. Circ Res. 2003;92:1138–44. 69. Schulz R, Gres P, Skyschally A, Duschin A, Belosjorow S, Konietzka I, Heusch G. Ischemic preconditioning preserves connexin 43 phosphorylation during sustained ischemia in pig hearts in vivo. FASEB J. 2003;17: 1355–57. 70. Schulz R, Heusch G. Connexin43 and ischemic preconditioning. Adv Cardiol. 2006;42:213–27. 71. Schwanke U, Konietzka I, Duschin A, Li X, Schulz R, Heusch G. No ischemic preconditioning in heterozygous connexin43-deficient mice. Am J Physiol. 2002;283:H1740–2. 72. Kanno S, Kovacs A, Yamada KA, Saffitz JE. Connexin43 as a determinant of myocardial infarct size following coronary occlusion in mice. J Am Coll Cardiol. 2003;41: 681–6. 73. Li F, Sugishita K, Su Z, Ueda I, Barry WH. Activation of connexin-43 hemichannels can elevate [Ca2þ]i and [Naþ]i in rabbit ventricular myocytes during metabolic inhibition. J Mol Cell Cardiol. 2001;33:2145–55. 74. Garcia-Dorado D, Rodriguez-Sinovas A, Ruiz-Meana M. Gap junction-mediated spread of cell injury and death during myocardial ischemia-reperfusion. Cardiovasc Res. 2004;61:386–401. 75. Matsushita T, Oyamada M, Fujimoto K, Yasuda Y, Masuda S, Wada Y, Oka T, Takamatsu T. Remodeling of cell-cell and cell-extracellular matrix interactions at the border zone of rat myocardial infarcts. Circ Res. 1999;85:1046–55. 76. Emdad L, Uzzaman M, Takagishi Y, Honjo H, Uchida T, Severs NJ, Kodama I, Murata Y. Gap junction remodelling in hypertrophied left ventricles of aortic-banded rats: prevention by angiotensin II type1 receptor blockade. J Mol Cell Cardiol. 2001;33:219–31. 77. Uzzaman M, Honjo H, Takagishi Y, Emdad L, Magee AI, Severs NJ, Kodama I. Remodeling of gap-junctional coupling in hypertrophied right ventricles of rats with monocrotaline-induced pulmonary hypertension. Circ Res. 2000;86:871–8. 78. Luke RA, Saffitz JE. Remodeling of ventricular conduction pathways in healed canine infarct border zones. J Clin Invest. 1991;87:1594–602. 79. Smith JH, Green CR, Peters NS, Rothery S, Severs NJ. Altered patterns of gap junction distribution in ischemic heart disease. An immunohistochemical study of human myocardium using laser scanning confocal microscopy. Am J Pathol. 1991;139:801–21. 80. Kostin S, Rieger M, Dammer S, Hein S, Richter M, Klovekorn WP, Bauer EP, Schaper J. Gap junction remodeling and altered connexin43 expression in the failing human heart. Mol Cell Biochem. 2003;242:135–44. 81. Kostin S, Dammer S, Hein S, Klovekorn WP, Bauer EP, Schaper J. Connexin 43 expression and distribution in compensated and decompensated cardiac hypertrophy in patients with aortic stenosis. Cardiovasc Res. 2004;62:426–36. 82. Sepp R, Severs NJ, Gourdie RG. Altered patterns of cardiac intercellular junction distribution in hypertrophic cardiomyopathy. Heart 1996;76:412–17. 83. Kaprielian RR, Gunning M, Dupont E, Sheppard MN, Rothery SM, Underwood R, Pennell DJ, Fox K, Pepper J, Poole-Wilson PA, Severs NJ. Downregulation of immunodetectable connexin43 and decreased gap junction size in the pathogenesis of chronic hibernation in the human left ventricle. Circulation. 1998;97:651–60. 84. Dupont E, Matsushita T, Kaba RA, Vozzi C, Coppen SR, Khan N, Kaprielian R, Yacoub MH, Severs NJ. Altered connexin expression in human congestive heart failure. J Mol Cell Cardiol. 2001;33:359–71. 85. Kitamura H, Ohnishi Y, Yoshida A, Okajima K, Azumi H, Ishida A, Galeano EJ, Kubo S, Hayashi Y, Itoh H, Yokoyama M. Heterogeneous loss of connexin43 protein in nonischemic dilated cardiomyopathy with ventricular tachycardia. J Cardiovasc Electrophysiol. 2002;13:865–70.

21 Cardiac Tissue

455

86. Peters NS, Green CR, Poole-Wilson PA, Severs NJ. Reduced content of connexin43 gap junctions in ventricular myocardium from hypertrophied and ischaemic human hearts. Circulation. 1993;88:864–75. 87. Yamada KA, Rogers JG, Sundset R, Steinberg TH, Saffitz JE. Upregulation of connexin45 in heart failure. J Cardiovasc Electrophysiol. 2003;14:1205–12. 88. Teunissen BE, Bierhuizen MF. Transcriptional control of myocardial connexins. Cardiovasc Res. 2004;62:246–55. 89. Saffitz JE, Kleber AG. Effects of mechanical forces and mediators of hypertrophy on remodeling of gap junctions in the heart. Circ Res. 2004;94:585–91. 90. Yamada K, Green KG, Samarel AM, Saffitz JE. Distinct pathways regulate expression of cardiac electrical and mechanical junction proteins in response to stretch. Circ Res. 2005;97:346–53. 91. Petrich BG, Eloff BC, Lerner DL, Kovacs A, Saffitz JE, Rosenbaum DS, Wang Y. Targeted activation of c-Jun N-terminal kinase in vivo induces restrictive cardiomyopathy and conduction defects. J Biol Chem. 2004;279: 15330–8. 92. Giepmans BN. Role of connexin43-interacting proteins at gap junctions. Adv Cardiol. 2006;42:41–56. 93. Hunter AW, Barker RJ, Zhu C, Gourdie RG. Zonula occludens-1 alters connexin43 gap junction size and organization by influencing channel accretion. Mol Biol Cell. 2005;16:5686–98. 94. Kaplan SR, Gard JJ, Carvajal-Huerta L, Ruiz-Cabezas JC, Thiene G, Saffitz JE. Structural and molecular pathology of the heart in Carvajal syndrome. Cardiovasc Pathol. 2004;13:26–32. 95. Gard JJ, Yamada K, Green KG, Eloff BC, Rosenbaum DS, Wang X, Robbins J, Schuessler RB, Yamada KA, Saffitz JE. Remodeling of gap junctions and slow conduction in a mouse model of desmin-related cardiomyopathy. Cardiovasc Res. 2005;67: 539–47. 96. Kostetskii I, Li J, Xiong Y, Zhou R, Ferrari VA, Patel VV, Molkentin JD, Radice GL. Induced deletion of the N-cadherin gene in the heart leads to dissolution of the intercalated disc structure. Circ Res. 2005;96:346–54. 97. Li J, Patel VV, Radice GL.Dysregulation of cell adhesion proteins and cardiac arrhythmogenesis. Clin Med Res. 2006;4:42–52. 98. Uzumcu A, Norgett EE, Dindar A, Uyguner O, Nisli K, Kayserili H, Sahin SE, Dupont E, Severs NJ, Leigh IM, Yuksel-Apak M, Kelsell DP, Wollnik B. Loss of desmoplakin isoform I causes early onset cardiomyopathy and heart failure in a Naxos-like syndrome. J Med Genet. 2006;43:e5. 99. Saffitz JE. Dependence of electrical coupling on mechanical coupling in cardiac myocytes: insights gained from cardiomyopathies caused by defects in cell-cell connections. Ann NY Acad Sci. 2005;1047:336–44. 100. Jongsma HJ, Wilders R. Gap junctions in cardiovascular disease. Circ Res. 2000;86:1193–7. 101. van Rijen HV, van Veen TA, Gros D, Wilders R, de Bakker JM. Connexins and cardiac arrhythmias. Adv Cardiol. 2006;42:150–60. 102. Rohr S, Kucera JP, Fast VG, Kleber AG. Paradoxical improvement of impulse conduction in cardiac tissue by partial cellular uncoupling. Science 1997;275:841–4. 103. Shaw RM, Rudy Y. Ionic mechanisms of propagation in cardiac tissue—roles of the sodium and L-type calcium currents during reduced excitability and decreased gap junction coupling. Circ Res. 1997;81:727–41. 104. Spach MS, Heidlage JF, Dolber PC, Barr RC. Electrophysiological effects of remodeling cardiac gap junctions and cell size. Circ Res. 2000;86:302–11. 105. Bernstein SA, Morley GE. Gap junctions and propagation of the cardiac action potential. Adv Cardiol. 2006;42:71–85. 106. Kle´ber AG, Rudy Y. Basic mechanisms of cardiac impulse propagation and associated arrhythmias. Physiol Rev. 2003;84:431–88. 107. Wang Y, Rudy Y. Action potential propagation in inhomogeneous cardiac tissue: safety factor considerations and ionic mechanisms. Am J Physiol Heart Circ Physiol. 2000;278:H1019–29.

456

N.J. Severs

108. Lerner DL, Yamada KA, Schuessler RB, Saffitz JE. Accelerated onset and increased incidence of ventricular arrhythmias induced by ischemia in Cx43-deficient mice. Circulation. 2000;101:547–52. 109. Gutstein DE, Morley GE, Vaidya D, Liu F, Chen FL, Stuhlmann H, Fishman GI. Heterogeneous expression of gap junction channels in the heart leads to conduction defects and ventricular dysfunction. Circulation. 2001;104:1194–9. 110. Danik SB, Liu F, Zhang J, Suk HJ, Morley GE, Fishman GI, Gutstein DE. Modulation of cardiac gap junction expression and arrhythmic susceptibility. Circ Res. 2004;95:1035–41. 111. Morley GE, Danik SB, Bernstein S, Sun Y, Rosner G, Gutstein DE, Fishman GI. Reduced intercellular coupling leads to paradoxical propagation across the Purkinje-ventricular junction and aberrant myocardial activation. Proc Natl Acad Sci USA. 2005;102:4126–9. 112. Elvan A, Huang X, Pressler ML, Zipes DP. Radiofrequency catheter ablation of the atria eliminates pacing-induced sustained a trial fibrillation and reduces connexin43 in dogs. Circulation. 1997;96:1675–85. 113. Sakabe M, Fujiki A, Nishida K, Sugao M, Nagasawa H, Tsuneda T, Mizumaki K, Inoue H. Enalapril prevents perpetuation of atrial fibrillation by suppressing atrial fibrosis and overexpression of connexin43 in a canine model of atrial pacing-induced left ventricular dysfunction. J Cardiovasc Pharmacol. 2004;43:851–9. 114. Ausma J, van der Velden HM, Lenders MH, van Ankeren EP, Jongsma HJ, Ramaekers FC, Borgers M, Allessie MA. Reverse structural and gap-junctional remodeling after prolonged atrial fibrillation in the goat. Circulation. 2003;107: 2051–8. 115. van der Velden HM, Ausma J, Rook MB, Hellemons AJ, van Veen TA, Allessie MA, Jongsma HJ. Gap junctional remodeling in relation to stabilization of atrial fibrillation in the goat. Cardiovasc Res. 2000;46:476–86. 116. Polontchouk L, Haefliger JA, Ebelt B, Schaefer T, Stuhlmann D, Mehlhorn U, KuhnRegnier F, De Vivie ER, Dhein S. Effects of chronic atrial fibrillation on gap junction distribution in human and rat atria. J Am Coll Cardiol. 2001;38:883–91. 117. Kostin S, Klein G, Szalay Z, Hein S, Bauer EP, Schaper J. Structural correlate of atrial fibrillation in human patients. Cardiovasc Res. 2002;54:361–79. 118. Nao T, Ohkusa T, Hisamatsu Y, Inoue N, Matsumoto T, Yamada J, Shimizu A, Yoshiga Y, Yamagata T, Kobayashi S, Yano M, Hamano K, Matsuzaki M. Comparison of expression of connexin in right atrial myocardium in patients with chronic atrial fibrillation versus those in sinus rhythm. Am J Cardiol. 2003;91:678–83. 119. Kanagaratnam P, Cherian A, Stanbridge RD, Glenville B, Severs NJ, Peters NS. Relationship between connexins and atrial activation during human atrial fibrillation. J Cardiovasc Electrophysiol. 2004;15:206–16. 120. Takeuchi S, Akita T, Takagishi Y, Watanabe E, Sasano C, Honjo H, Kodama I. Disorganization of gap junction distribution in dilated atria of patients with chronic atrial fibrillation. Circ J. 2006;70:575–82. 121. Kanagaratnam P, Dupont E, Rothery S, Coppen S, Severs NJ, Peters NS. Human atrial conduction and arrhythmogenesis correlates with conformational exposure of specific epitopes on the connexin40 carboxyl tail. J Mol Cell Cardiol. 2006;40:675–87. 122. Yeh HI, Lai YJ, Lee SH, Lee YN, Ko YS, Chen SA, Severs NJ, Tsai CH. Heterogeneity of myocardial sleeve morphology and gap junctions in canine superior vena cava. Circulation. 2001;104:3152–7. 123. Dupont E, Ko Y, Rothery S, Coppen SR, Baghai M, Haw M, Severs NJ. The gapjunctional protein, connexin40, is elevated in patients susceptible to post-operative atrial fibrillation. Circulation. 2001;103:842–9. 124. Beauchamp P, Yamada KA, Baertschi AJ, Green K, Kanter EM, Saffitz JE, Kleber AG. Relative contributions of connexins 40 and 43 to atrial impulse propagation in synthetic strands of neonatal and fetal murine cardiomyocytes. Circ Res. 2006;99:1216–24. 125. Kanagaratnam P, Rothery S, Patel P, Severs NJ, Peters NS. Relative expression of immunolocalized connexins 40 and 43 correlates with human atrial conduction properties. J Am Coll Cardiol. 2002;39:116–23.

Chapter 22

Connexins in the Vasculature ¨ Cor de Wit and Stephanie E. Wolfle

Abstract Vascular function requires the highly coordinated behavior of individual cells. The modulation of vascular resistance and blood flow required to match oxygen delivery to a wide dynamic range of tissue needs can be achieved only by coordinated diameter changes over large distances along the vessel. This is accomplished by longitudinal long-distance communication through the vessel wall by gap junctions. The vascular endothelial cells are structurally suited for this task and are coupled extraordinarily well to form a functional unit within the vessel wall. Cx40 is the most abundant connexin in the vascular endothelium. Its loss results in functional deficits, such as hypertension and a lack of coordination of vascular responses. Gap junctions also couple the vascular smooth muscle cells. These intercellular junctions are formed by Cx43, though evidence of Cx45 expression has been provided recently. In addition, gap junctions interconnect endothelial and smooth muscle cells heterocellularly to create short-distance transverse pathways. This theoretically allows reciprocal direct communication pathways between the cells of these two vascular compartments and could potentially provide for the function of the elusive endothelium-derived hyperpolarizing factor. Keywords Conducted response Endothelium-derived hyperpolarizing factor Myoendothelial junction Microcirculation Hypertension Vascular tone Cx37 Cx40 Cx43

22.1 Introduction Oxygen delivery and blood flow to organs must match tissue needs to avoid insufficient or excessive perfusion. Because tissue oxygen demand varies substantially in response to changes in tissue function, blood flow C.de Wit (*) Institut fu¨r Physiologie, Universita¨t zu Lu¨beck, Ratzeburger Allee 160, 23538 Lu¨beck, Germany e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_22, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

457

458

C. de Wit and S.E. Wolfle ¨

regulation requires changes of vascular diameter over a wide dynamic range. Vascular caliber is the most important parameter controlling resistance, and thereby flow [1], and is determined by the local interaction of endothelial and smooth muscle cells. The endothelium exerts control of smooth muscle tone and diameter by the release of substances (e.g., nitric oxide and prostaglandins) that diffuse to the adjacent vascular smooth muscle (VSM) and affect its tone. In addition to the extracellular release of substances, endothelial cells (ECs) also communicate directly with VSM through myoendothelial gap junctions (MEGJs). This establishes a transverse communication pathway that allows electrical communication by direct transfer of charge. Moreover, MEGJs also provide a pathway for direct diffusional exchange of molecules (e.g., cyclic adenosine monophosphate [cAMP], inositol triphosphate [IP3], Ca2þ) without the dilutional effect of transfer through the extracellular space. Notably, this also enables VSM to deliver signaling molecules back to the endothelium, establishing direct feedback. Another functionally important communication pathway orchestrates the behavior of the cells of the vascular wall longitudinally along the vessel, connecting and forcing the cells to act as a single unit. The need for such a communication pathway arises from the structural design of the vascular tree and the considerable distances that are spanned by arteries and arterioles within the microcirculatory network, which is the main source of vascular resistance. If vascular dilations were restricted to the arterioles in the vicinity of the capillary bed where a metabolic stimulus arises, the resulting flow increase would be constrained by flow-limiting upstream resistance. To achieve sufficient blood flow and oxygen delivery to the tissue, an overall conductance increase from the larger arteries down the capillaries is necessary. Thus, coordinated dilation along the length of the vascular tree is a prerequisite for the drastic increases in flow that are required (up to 25-fold in working skeletal muscle [2]). Because dilation ascends the arteriolar tree, it is termed ascending or conducted dilation. It occurs over a distance of several millimeters in the microcirculation and may also affect diameters of vessels outside of the skeletal muscle (conducting vessels). Ascending dilations therefore arise by the communication of signals along the vessel wall that enable minute changes in vascular diameter along the vessel to serve tissue needs. One coordinating signal is the physical force generated by the flowing blood acting on the EC layer, eliciting an upstream, flow-induced dilation [3]. Other communication pathways may contribute. Gap junctional signaling is an attractive mechanism as it connects a large number of cells, enabling them to act as a functional unit. For this purpose, signaling within each cell layer, EC or VSM, is required and is achieved via homocellular coupling forming a long-distance, longitudinal signal transfer that synchronizes cellular behavior within a segment along the vessel.

22 Vasculature

459

22.2 Transverse Signaling: Myoendothelial Coupling Gap junctions between ECs and VSM provide direct pathways for shortdistance communication. Morphological heterocellular junctions are typically found on projections arising from ECs through the internal elastic lamina that brings the cell membranes into close apposition [4]. The MEGJs are small in size and number compared to endothelial junctions. They have been identified in selected vascular beds [5] and also form in a coculture system of ECs and VSM [6]. For many years research has focused on the identification of a mediator that is released from the endothelium and induces smooth muscle hyperpolarization (known as endothelium-derived hyperpolarizing factor, EDHF). The contribution of gap junctions to EDHF-type responses has been recognized only recently. Gap junctions could provide a pathway for direct current transfer from EC to VSM, obviating the need for a distinct molecular factor [7]. If tight myoendothelial coupling exists, the membrane hyperpolarizations induced in the EC by acetylcholine should spread to VSM. Indeed, recordings obtained in the two cell types were synchronous and indistinguishable in isolated hamster arteries [8]. Similarly, in mesenteric arterioles in vitro, current injected into an EC changed the membrane potential of VSM and vice versa, showing bidirectional current transfer [9]. In contrast, in in vivo experiments, responses to acetylcholine as well as spontaneous fluctuations of the membrane potential varied between ECs and VSM [10]. Nevertheless, gap junctions seem to be crucial to achieve the full dilator potency of endothelium-dependent dilator mechanisms in some preparations; that is, dilation is reduced by inhibition of gap junctions [7]. This suggests that a dilatory principle is transferred from ECs to VSM through heterocellular gap junction channels. This could be a dilatory signaling molecule or simply the direct transfer of charge [11]. In addition to electrical signaling, Ca2þ and the second messenger IP3 can be transferred via the MEGJ [12]. The role of myoendothelial coupling in EDHF-type dilation is currently heavily debated, an ongoing area of intensive research, and many controversies remain. Some of the controversy may be due to variations in MEGJ composition in different vessel types. Others may depend on the different experimental conditions (in vivo versus in vitro) that may alter the role of MEGJ in EDHF-type dilations.

22.3 Longitudinal Signaling: Conducted Vascular Responses Discrete stimulation of arterioles within the vascular network induces vasomotor responses not only at the site of stimulation but also in upstream vessels through multiple branches; that is, the response ascends the vascular tree [13]. This can be studied experimentally in vitro in isolated vessels or in vivo using intravital microcirculatory approaches that allow locally confined stimulation of arterioles with vasoactive substances [14,15]. The dilatory responses also

460

C. de Wit and S.E. Wolfle ¨

conduct to downstream locations. The most common tissues for study of longitudinal signaling are thin skeletal muscles (cremaster muscle, cheek pouch, retractor muscle), but this phenomenon is not restricted to vessels residing in skeletal muscle and can also be observed in mesenteric, kidney, cerebral, and coronary arterioles [16,17]. Although local stimulation with many vasodilators induces a response at the application site only, some elicit conducted dilations [18]. Thus, the local dilation itself does not necessarily initiate a conducting signal, and therefore perturbation of blood flow (e.g., enhanced flow) is an unlikely initiator of a conducted response. Common properties of substances that initiate conducted responses are that they act on ECs (e.g., acetylcholine, bradykinin) and they induce changes of the membrane potential (see below). In contrast, substances like nitric oxide (NO) donors, which act directly on smooth muscle, elicit dilation at the application site only [19]. The dilation travels along the vessel wall at a very high speed (>1 mm/ second), which excludes the transmission of a dilatory molecular signal by diffusion through gap junctions or the extracellular space. Such transmission velocities can be achieved by electrical signals, which may conduct along the vessel wall or possibly through accompanying nerves. However, the latter mechanism was excluded by the blockade of fast Naþ-channels via tetrodotoxin, which did not affect the conducted responses. Thus, it is generally accepted that a mechanism contained within the vascular wall itself underlies conducted dilations. Conducting signals are not limited to dilations, as locally evoked constrictions also conduct along the vascular wall [20].

22.4 Conducted Responses Require Changes of the Membrane Potential In most tissues, a locally initiated depolarization results in local and conducted vasoconstriction, and, conversely, locally initiated hyperpolarization gives rise to a dilatory conducted response. Although the electrical signals that produce constriction and dilation are opposite in polarity, they both require signal conduction along the vascular wall made possible by low-resistance channels, gap junctions, that connect the cells.

22.4.1 Conducting Dilations The transmission velocity of conducted responses suggests that electrotonic signals travel along the wall to induce remote diameter changes. Indeed, dilators that are able to initiate remote dilations hyperpolarize ECs and VSM in vitro and in vivo. Although these stimulators concomitantly release NO and prostaglandins from the endothelium, blockade of NO synthase and cyclooxygenase

22 Vasculature

461

does not affect remote dilation induced by acetylcholine application [19]. In contrast, preventing the hyperpolarization abrogates local and conducted responses. Further evidence of electrical coupling mediating the conducted response is provided by the measurement of membrane potential along the vessel, revealing large changes of the membrane potential at substantial distances from the stimulation site [9,10,21,22]. Thus, the initiation of conducted dilations requires an endothelial hyperpolarization at the site of stimulation. This hyperpolarization depends on the activation of Ca2þ-dependent Kþ channels (KCa). The different subtypes of KCa channels are divided according to their conductance into small, intermediate, and large conductance channels [23]. In ECs, small and intermediate KCa channels seem to be of special importance [24,25]. The Ca2þ-dependency suggests that a rise of intracellular Ca2þ in ECs closely related to the hyperpolarization is a second important signal that follows endothelial stimulation. In vivo, a Ca2þ rise seems to be a prerequisite for the endothelial hyperpolarization because chelation of intracellular Ca2þ prevents the agonist-induced hyperpolarization and a subsequent conducted dilation [21,26]. However, an increase of Ca2þ does not seem to be required to initiate the response if the hyperpolarization is induced by other means, for example, by current injection [27]. Furthermore, prevention of the hyperpolarization by inhibition of KCa (without altering the rise in Ca2þ) abrogates local and conducted responses. Thus, it seems that the hyperpolarization is necessary and sufficient to initiate a conducted dilation, and a rise in Ca2þ is required only to launch this hyperpolarization. This view is supported by the fact that local increases in Ca2þ do not travel along the vascular wall [21]. However, very recently this has been challenged because Ca2þ waves were found to propagate at a lower speed for considerable distances along the EC layer [28]. This again raises the question of which signal, Ca2þ or hyperpolarization, is required to elicit the remote dilation. Spatially confined blockade of KCa channels at conducted sites did not prevent the remote dilation in response to acetylcholine despite abrogation of the dilation after local application [19]. This suggests that activation of KCa channels is required at the stimulation site but is not essential at the remote site. While this scenario does not support a role for a spreading Ca2þ wave in acetylcholine-induced dilations, other substances may act by different mechanisms. In fact, conducted dilations initiated by bradykinin (also an endotheliumdependent dilator) seem to rely on the release of NO along the vessel [20]. Because the endothelial NO synthase is Ca2þ-dependent, a NO wave may be due to an underlying Ca2þ wave acting as a second, slowly propagating signal [29]. Taken together, different mechanisms are able to deliver a message along the vessel wall, depending on the stimulus and possibly also on the tissue studied. The large distance over which diameter changes can be observed suggests a regenerative, amplifying mechanism. It may consist of voltage-dependent or inward rectifier Kþ channels that amplify the initial hyperpolarization as it travels along the vascular wall and regenerates the membrane potential change.

462

C. de Wit and S.E. Wolfle ¨

Likewise, Ca2þ-induced Ca2þ-release may amplify a Ca2þ wave. At present, little is known about alterations of conducted responses under pathophysiological conditions. In hypertensive animals conducted dilations are diminished [30] but they are preserved in hypercholesterolemic mice [31].

22.4.2 Conducting Constrictions Conducted vasoconstrictions seem to rely on depolarization. A locally induced depolarization spreads along the vessel wall and induces remote depolarization and constriction. Other vasoconstrictors (e.g., norepinephrine) may induce depolarizations of various amplitudes depending on species and tissue [32,33]. If the depolarization is sufficient, it gives rise to a conducted vasoconstriction. However, in some tissues smooth muscle cells are poorly coupled (e.g., retractor muscle of the hamster) and conducted vasoconstrictions are rarely observed [34]. This may be related to a dense sympathetic nerve innervation that may be able to induce constrictions along the overall vessel length. Therefore, tight gap junctional coupling is not necessary in this case. However, in the majority of the vessels studied, conducted constrictions are readily observed.

22.5 Pathway for the Longitudinally Conducted Signal Gap junctions are found in both ECs and VSM, but ECs seem to be especially capable of transmitting a long-distance signal because of their anatomical shape and the vessel length that one EC extends (80 to 140 mm). In elegant experiments, Looft-Wilson et al. [35] destroyed each cell layer separately by a lightdye technique and evaluated how responses along the conducting pathway were affected by the selective impairment of each cell layer. These experiments demonstrated that the EC layer is crucial for transmitting a dilatory signal in the murine microcirculation. This was also the case for a hamster feeding vessel residing outside the skeletal muscle [22]. However, within the skeletal muscle in hamsters, dilations still traversed sections of injured EC in vessels. Thus, in some vessel types a gap junction–coupled VSM is sufficient to transmit a conducted dilation that is abolished only after additional destruction of the smooth muscle layer [36]. Unexpectedly, the destruction of each cell layer separately at different locations along the vessel prevented the conduction through the second site of damage in vessels [20], which suggests conduction along two parallel cables. This implies that these cables act as separate conduction pathways unconnected to each other, and that a signal is unable to jump between the layers via MEGJ when traveling along the wall. In contrast to dilatory signals, the disruption of the VSM prevented the conduction of a locally initiated vasoconstriction independent of the initiating stimulus. Thus, signals to induce conducted

22 Vasculature

463

constrictions are transmitted through the VSM layer, whereas dilatory signals are conducted along the EC and in some preparations additionally along the VSM layer.

22.6 Connexins Have Distinct Roles in the Vascular Wall Of the large family of connexins, only Cx37, Cx40, Cx43, and Cx45 are expressed in the vasculature. Most data are from large vessels, in which Cx43 is the connexin predominantly expressed in VSM; its expression in ECs is reported to be scarce. Cx40 and Cx37 prevail in ECs throughout the vascular tree [37,38,39]. Expression of Cx45 has been only recently demonstrated in the vascular wall, most likely in the smooth muscle. Although this pattern is a general feature, the distribution varies between species and more importantly between vessel types (Fig. 22.1). This is especially true for small arteries and arterioles, in which quite distinct expression levels have been found [40,41,42]. Very recently, it has been shown in a coculture system that MEGJs are composed of Cx40 (expressed in ECs) and Cx43 (expressed in ECs and VSM). Cx37 is found in both cells but excluded from the MEGJ [6]. However, which connexins contribute to MEGJ is presently still a matter of debate; for example, Cx37 is reported to be present in MEGJ in cerebral vessels [43]. Strikingly, little is known about connexin expression in capillaries and veins.

Fig. 22.1 Connexin distribution in the vasculature. Three connexins (Cx37, Cx40, and Cx43) have been located at the borders of endothelial cells throughout the vasculature (dominant connexins in bold). Cx40 is the predominant connexin in most vessels, especially in arteries and arterioles. In smooth muscle cells (SMC), Cx37, Cx43, and Cx45 are expressed, but subtypes vary with vessel size. Whereas Cx43 is highly expressed in the aortic arch and ascending aorta, its expression decreases in the abdominal aorta and is not detected in the SMC of many other arteries and arterioles. Instead, Cx37 prevails in these vessels in SMC. As an exception, Cx40 is the dominating connexin in SMC of preglomerular afferent arterioles at the juxtaglomerular apparatus of the kidney (not shown). Myoendothelial gap junctions (MEGJs) have been found in most vessels, but their connexin composition is still a matter of debate. (A highresolution version of this figure is available on the accompanying CD and online at www. springerlink.com)

464

C. de Wit and S.E. Wolfle ¨

The analysis of specific roles for certain connexins has been hampered by the lack of specificity of gap junction blockers, and the fact that they are difficult to deliver at the concentrations required (see Chapter 8). Recently, small peptides that interfere with the extracellular domain of connexins have been used in vitro to evaluate myoendothelial gap junctional communication [7,44]. These peptides abrogate the EDHF-type dilation and hyperpolarization in VSM in response to acetylcholine treatment of arteries in vitro. Therefore, it is suggested that MEGJs allow charge transfer from ECs to VSM, thus challenging the view that a transferable chemical factor is released from ECs (as outlined above). A different approach is the use of mice that are deficient (knockout, KO) for specific connexins. Whereas Cx40KO and Cx37KO animals are viable, Cx43KO and Cx45KO animals die perinatally, and therefore for these connexins the studies are limited to mice with cell-specific gene disruption. Study of these mice has revealed that specific connexins serve specific functions in the microcirculation. Cx40, which is mainly expressed in the endothelium in these vessels, is required for the conduction of vasodilator responses along the arteriolar wall, whereas the conduction of constrictions remained unaffected by Cx40KO [33,45]. This is consistent with the different pathways for signal conduction outlined above, that is, dilations being conducted in a Cx40-dependent manner along the endothelium and constrictions being conducted through the smooth muscle cell layer. Cx37KO has not been reported to produce a vascular defect. However, a critical role of Cx37 is revealed in double Cx37/ Cx40 knockout animals. Mice with either Cx40KO or Cx37KO are viable, but animals lacking both connexins die perinatally, with vascular abnormalities suggesting that Cx37 or heteromeric Cx37/Cx40 or heterotypic Cx37-Cx40 channels serve important functions [46]. Together, these results demonstrate that EC coupling is indispensable. Interestingly, Cx40KO animals are hypertensive [47]. This hypertension is not caused by alterations of the important myoendothelial dilator NO, and is coincident with an altered vasomotion pattern consisting of brief spontaneous vessel closures. Together with a defect of conduction along EC, it may result in increased peripheral resistance and thus contribute to the observed hypertension. One may expect that the hypertension would be counterbalanced by compensatory renal excretion mechanisms, for example, by a suppression of the renin-angiotensin system. However, compensatory renal mechanisms may also depend on Cx40 because this is expressed in the renin-secreting cells in the kidney [48]. In fact, the negative feedback control of renin secretion is abrogated in Cx40KO mice and plasma renin is elevated fivefold in these mice despite hypertension. The high plasma renin level itself is not the cause of arterial hypertension because the difference in pressure between wild-type and Cx40KO mice persists after blockade of the renin-angiotensin-system [49]. This suggests that deletion of Cx40 compromises both vascular and renal function, which together cause considerable hypertension. Further research is needed to identify the mechanism that initiates hypertension in these animals.

22 Vasculature

465

In contrast, endothelial-specific knockout of Cx43 (by Cre/loxP-system using the endothelial promoter TIE2, an angiopoietin-1 receptor tyrosine kinase with immunoglobulin and epidermal growth factor homology domains) results in hypotension and concurrent bradycardia [50]. The authors reported enhanced levels of NO, but also increased plasma levels of angiotensin II, leaving the cause of hypotension unclear. Others have also studied endothelial-specific Cx43KO mice using the same procedure but have not observed any alterations in blood pressure or heart rate [51]. The reasons for these divergent results are still unclear, and require further investigation. However, what is clear is that connexins are crucial for the function of arterioles, regulation of vascular tone, and possibly contribute to the physiological control of peripheral vascular resistance and arterial pressure.

22.7 Concluding Remarks Vascular function requires the coordination of the behavior of the cells of the vessel wall. This is achieved by the coupling of ECs and VSM through gap junctions into two functional units. Coupling of the EC layer through Cx40dependent gap junctions provides a pathway for fast signal transduction along the vessel wall itself. The deletion of genes encoding vascular connexins in mice demonstrates a critical role of Cx40 in the maintenance of conduction and, interestingly, control of arterial pressure. This hypertension related to the deletion of Cx40 is caused by abnormalities in vascular and renal function. Future research using cell-targeted disruption of connexins will shed important light on the function of other connexins (Cx37, Cx43, and Cx45) in the vascular wall, including their contributions to MEGJs, which provide an important transverse signaling pathway from ECs to VSM and which contribute to endothelium-dependent dilation. Acknowledgment Studies performed in the authors’ laboratory were supported by the Deutsche Forschungsgemeinschaft (WI 2071/1-1).

References 1. Pohl U, Wagner K, de Wit C. Endothelium-derived nitric oxide in the control of tissue perfusion and oxygen supply: physiologic and pathophysiologic implications. Eur Heart J. 1993;14:93–8. 2. de Wit C. Connexins pave the way for vascular communication. News Physiol Sci. 2004;19:148–53. 3. Pohl U, de Wit C. A unique role of NO in the control of blood flow. News Physiol Sci. 1999;14:74–80. 4. Dora KA, Sandow SL, Gallagher NT, Takano H, Rummery NM, Hill CE, Garland CJ. Myoendothelial gap junctions may provide the pathway for EDHF in mouse mesenteric artery. J Vasc Res. 2003;40:480–90.

466

C. de Wit and S.E. Wolfle ¨

5. Sandow SL, Looft-Wilson R, Doran B, Grayson TH, Segal SS, Hill CE. Expression of homocellular and heterocellular gap junctions in hamster arterioles and feed arteries. Cardiovasc Res. 2003;60:643–53. 6. Isakson BE, Duling BR. Heterocellular contact at the myoendothelial junction influences gap junction organization. Circ Res. 2005;97:44–51. 7. Griffith TM. Endothelium-dependent smooth muscle hyperpolarization: do gap junctions provide a unifying hypothesis? Br J Pharmacol. 2004;141:881–903. 8. Emerson GG, Segal SS. Electrical coupling between endothelial cells and smooth muscle cells in hamster feed arteries—role in vasomotor control. Circ Res. 2000;87:474–9. 9. Yamamoto Y, Klemm MF, Edwards FR, Suzuki H. Intercellular electrical communication among smooth muscle and endothelial cells in guinea-pig mesenteric arterioles. J Physiol. 2001;535:181–95. 10. Siegl D, Koeppen M, Wolfle SE, Pohl U, de Wit C. Myoendothelial coupling is not ¨ prominent in arterioles within the mouse cremaster microcirculation in vivo. Circ Res. 2005;97:781–8. 11. de Wit C, Wolfle SE. EDHF and gap junctions: important regulators of vascular tone ¨ within the microcirculation. Curr Pharm Biotechnol. 2007;8:11–25. 12. Isakson BE, Ramos SI, Duling BR. Ca2þ and inositol 1,4,5-trisphosphate-mediated signaling across the myoendothelial junction. Circ Res. 2007;100:246–54. 13. Segal SS, Duling BR. Flow control among microvessels coordinated by intercellular conduction. Science 1986;234:868–70. 14. Delashaw JB, Duling BR. Heterogeneity in conducted arteriolar vasomotor response is agonist dependent. Am J Physiol. 1991;260:H1276–82. 15. Rivers RJ. Pharmacologic study of muscarinic receptor subtypes and arteriolar dilations: a comparison of conducted and local responses. J Cardiovasc Pharmacol. 1999;33:388–93. 16. Steinhausen M, Endlich K, Nobiling R, Parekh N, Schutt F. Electrically induced vasomotor responses and their propagation in rat renal vessels in vivo. J Physiol. 1997;505:493–501. 17. Rivers RJ, Hein TW, Zhang C, Kuo L. Activation of barium-sensitive inward rectifier potassium channels mediates remote dilation of coronary arterioles. Circulation 2001;104:1749–53. 18. Doyle MP, Duling BR. Acetylcholine induces conducted vasodilation by nitric oxidedependent and -independent mechanisms. Am J Physiol. 1997;272:H1364–71. 19. Hoepfl B, Rodenwaldt B, Pohl U, de Wit C. EDHF, but not NO or prostaglandins, is critical to evoke a conducted dilation upon ACh in hamster arterioles. Am J Physiol Heart Circ Physiol. 2002;283:H996–1004. 20. Budel S, Bartlett IS, Segal SS. Homocellular conduction along endothelium and smooth muscle of arterioles in hamster cheek pouch: unmasking an NO wave. Circ Res. 2003;93:61–8. 21. Dora KA, Xia J, Duling BR. Endothelial cell signaling during conducted vasomotor responses. Am J Physiol Heart Circ Physiol. 2003;285:H119–26. 22. Emerson GG, Segal SS. Endothelial cell pathway for conduction of hyperpolarization and vasodilation along hamster feed artery. Circ Res. 2000;86:94–100. 23. Ledoux J, Werner ME, Brayden JE, Nelson MT. Calcium-activated potassium channels and the regulation of vascular tone. Physiology 2006;21:69–78. 24. Jackson WF. Potassium channels in the peripheral microcirculation. Microcirculation 2005;12:113–27. 25. Si H, Heyken WT, Wolfle SE, Tysiac M, Schubert R, Grgic I, Vilianovich L, Giebing G, ¨ Maier T, Gross V, Bader M, de Wit C, Hoyer J, Kohler R. Impaired endothelium-derived hyperpolarizing factor-mediated dilations and increased blood pressure in mice deficient of the intermediate-conductance Ca2þ-activated Kþ channel. Circ Res. 2006;99:537–44. 26. Duza T, Sarelius IH. Conducted dilations initiated by purines in arterioles are endothelium dependent and require endothelial Ca2þ. Am J Physiol Heart Circ Physiol. 2003;285:H26–37.

22 Vasculature

467

27. Emerson GG, Segal SS. Electrical activation of endothelium evokes vasodilation and hyperpolarization along hamster feed arteries. Am J Physiol Heart Circ Physiol. 2001;280:H160–67. 28. Uhrenholt TR, Domeier TL, Segal SS. Propagation of calcium waves along endothelium of hamster feed arteries. Am J Physiol Heart Circ Physiol. 2007;292:H1634–40. 29. Domeier TL, Segal SS. Electromechanical and pharmacomechanical signaling pathways for conducted vasodilatation along endothelium of hamster feed arteries. J Physiol. 2007;579:175–86. 30. Kurjiaka DT. The conduction of dilation along an arteriole is diminished in the cremaster muscle of hypertensive hamsters. J Vasc Res. 2004;41:517–24. 31. Wolfle SE, de Wit C. Intact endothelium-dependent dilation and conducted responses in ¨ resistance vessels of hypercholesterolemic mice in vivo. J Vasc Res. 2005;42:475–82. 32. Hungerford JE, Sessa WC, Segal SS. Vasomotor control in arterioles of the mouse cremaster muscle. FASEB J. 2000;14:197–207. 33. de Wit C, Roos F, Bolz SS, Kirchhoff S, Kruger O, Willecke K, Pohl U. Impaired conduction of vasodilation along arterioles in connexin40 deficient mice. Circ Res. 2000;86:649–55. 34. Segal SS, Welsh DG, Kurjiaka DT. Spread of vasodilatation and vasoconstriction along feed arteries and arterioles of hamster skeletal muscle. J Physiol. 1999;516:283–91. 35. Looft-Wilson RC, Payne GW, Segal SS. Connexin expression and conducted vasodilation along arteriolar endothelium in mouse skeletal muscle. J Appl Physiol. 2004;97:1152–58. 36. Bartlett IS, Segal SS. Resolution of smooth muscle and endothelial pathways for conduction along hamster cheek pouch arterioles. Am J Physiol Heart Circ Physiol. 2000;278:H604–12. 37. Yeh HI, Rothery S, Dupont E, Coppen SR, Severs NJ. Individual gap junction plaques contain multiple connexins in arterial endothelium. Circ Res. 1998;83:1248–63. 38. Severs NJ, Rothery S, Dupont E, Coppen SR, Yeh HI, Ko YS, Matsushita T, Kaba R, Halliday D. Immunocytochemical analysis of connexin expression in the healthy and diseased cardiovascular system. Microsc Res Tech. 2001;52:301–22. 39. Haefliger JA, Nicod P, Meda P. Contribution of connexins to the function of the vascular wall. Cardiovasc Res. 2004;62:345–56. 40. Hill CE, Rummery N, Hickey H, Sandow SL. Heterogeneity in the distribution of vascular gap junctions and connexins: implications for function. Clin Exp Pharmacol Physiol. 2002;29:620–25. 41. Gustafsson F, Mikkelsen HB, Arensbak B, Thuneberg L, Neve S, Jensen LJ, HolsteinRathlou NH. Expression of connexin 37, 40 and 43 in rat mesenteric arterioles and resistance arteries. Histochem Cell Biol. 2003;119:139–48. 42. Looft-Wilson RC, Payne GW, Segal SS. Connexin expression and conducted vasodilation along arteriolar endothelium in mouse skeletal muscle. J Appl Physiol. 2004;97: 1152–58. 43. Haddock RE, Grayson TH, Brackenbury TD, Meaney KR, Neylon CB, Sandow SL, Hill CE. Endothelial coordination of cerebral vasomotion via myoendothelial gap junctions containing connexins 37 and 40. Am J Physiol Heart Circ Physiol. 2006;291: H2047–56. 44. Chaytor AT, Bakker LM, Edwards DH, Griffith TM. Connexin-mimetic peptides dissociate electrotonic EDHF-type signaling via myoendothelial and smooth muscle gap junctions in the rabbit iliac artery. Br J Pharmacol. 2005;144:108–14. 45. Figueroa XF, Paul DL, Simon AM, Goodenough DA, Day KH, Damon DN, Duling BR. Central role of connexin40 in the propagation of electrically activated vasodilation in mouse cremasteric arterioles in vivo. Circ Res. 2003;92:793–800. 46. Simon AM, McWhorter AR. Vascular abnormalities in mice lacking the endothelial gap junction proteins connexin37 and connexin40. Dev Biol. 2002;251:206–220.

468

C. de Wit and S.E. Wolfle ¨

47. de Wit C, Roos F, Bolz SS, Pohl U. Lack of vascular connexin 40 is associated with hypertension and irregular arteriolar vasomotion. Physiol Genomics. 2003;13: 169–77. 48. Haefliger JA, Demotz S, Braissant O, Suter E, Waeber B, Nicod P, Meda P. Connexins 40 and 43 are differentially regulated within the kidneys of rats with renovascular hypertension. Kidney Int. 2001;60:190–201. 49. Wagner C, de Wit C, Kurtz L, Grunberger C, Kurtz A, Schweda F. Connexin40 is essential for the pressure control of renin synthesis and secretion. Circ Res. 2007;100:556–63. 50. Liao Y, Day KH, Damon DN, Duling BR. Endothelial cell-specific knockout of connexin 43 causes hypotension and bradycardia in mice. Proc Natl Acad Sci USA. 2001;98:9989–94. 51. Theis M, de Wit C, Schlaeger TM, Eckardt D, Kruger O, Doring B, Risau W, Deutsch U, Pohl U, Willecke K. Endothelium-specific replacement of the connexin43 coding region by a lacZ reporter gene. Genesis 2001;29:1–13.

Chapter 23

Connexins and Atherosclerosis Anna Pfenniger, Isabelle Roth and Brenda R. Kwak

Abstract Studies on human blood vessels and on mouse models, as well as population studies, have led to the hypothesis that connexins play an important role in the pathogenesis of atherosclerosis. This inflammatory disease in the vascular wall involves three main cell types that closely interact with each other — monocytes/macrophages, endothelial cells, and smooth muscle cells — which each express a distinct pattern of connexins. Expression of the three major vascular connexins, Cx37, Cx40, and Cx43, is differentially modified in atheroma-associated cells during atherosclerotic plaque development. Mouse models have shown that Cx43 has an atherogenic effect, whereas Cx37 and Cx40 seem to be atheroprotective. Several mechanisms underlying these respective effects have been proposed. Keywords Atherosclerosis Endothelium Macrophage Smooth muscle Blood vessels Cx31.9 Cx37 Cx40 Cx43 Cx45

23.1 Introduction Cardiovascular diseases currently constitute the major cause of death in developed countries, and are becoming increasingly prevalent in developing countries. Atherosclerosis, an inflammatory disease of medium and large arteries, is the most important cause of cardiovascular disease [1]. The main consequences of atherosclerosis are myocardial infarction, cerebral infarction, and aortic aneurysm, making atherosclerosis a major threat to human health worldwide. This chapter reviews current knowledge regarding the involvement of the three major vascular connexins in atherosclerosis. B.R. Kwak (*) Division of Cardiology, Department of Medicine, Geneva University Hospitals, Foundation for Medical Research, 64 Avenue de la Roseraie, CH-1211 Geneva 4, Switzerland e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_23, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

469

470

A. Pfenniger et al.

23.2 Pathogenesis of Atherosclerosis Specific lesions, called atheromas, whereby lipids and inflammatory cells accumulate in the walls of the arteries over time, characterize atherosclerosis. The distribution of atheromas is highly characteristic in humans; the lesions tend to develop around the origins of arteries, which are regions of turbulent blood flow [2]. This suggests a role for hemodynamics in the pathogenesis of atherosclerosis. The lower abdominal aorta is most commonly affected, followed by the coronary arteries, the popliteal arteries (behind the knee), the descending thoracic aorta, the internal carotid arteries, and the vessels of the circle of Willis at the base of the brain. This distribution correlates with the clinical impact of atherosclerosis. Atherosclerosis is a progressive disease. The first steps of atherosclerotic lesions occur during childhood, and the disease usually becomes symptomatic after the fourth decade. Atherosclerosis is considered a chronic inflammatory response to some form of endothelial injury [2,3]. In early lesions, called fatty streaks, the endothelial cells become activated, which is manifested by increased endothelial permeability and leukocyte adhesion. Various factors seem to initiate this endothelial dysfunction, including turbulent blood flow, hypertension, cigarette-derived products, microorganisms, diabetes, and hypercholesterolemia. It has been shown that the transition from laminar to turbulent flow induces the expression of many proinflammatory genes in endothelial cells, including cytokines and adhesion molecules [4]. At the same time, hypercholesterolemia (more precisely, the excess of low-density lipoprotein [LDL]-cholesterol) promotes the accumulation of lipids in the intima, the innermost layer of the arterial wall [5]. The expression of endothelial adhesion molecules results in the transmigration of monocytes and T lymphocytes into the intima [6,7]. Free radicals modify the lipids and yield oxidized LDL, which has two negative consequences. First, macrophages ingest the oxidized lipoproteins through the scavenger receptor and evolve into the characteristic foam cells [2]. Second, the oxidized lipoproteins have chemotactic and immunogenic properties [8,9] that enhance the inflammatory response. As the hypercholesterolemia and the inflammatory response persist, the fatty streaks develop over the years into more complex atheromas. Medial smooth muscle cells migrate into the intima, where they proliferate and synthesize extracellular matrix components such as collagen and elastin [2]. They form an organized layer beneath the endothelium, a fibrous cap, and thus increase the mass of the plaque. As the lesion progresses, the increased lipid mass accumulates within the extracellular space. These lipid pools, foam cells, and necrotic debris form a necrotic core under the fibrous cap in advanced atheromas. The fibrous cap plays an important role in the stability of the plaque; most clinical events related to atherosclerosis, such as myocardial and cerebral infarction, are caused by rupture of the cap, which exposes highly thrombogenic material beneath it [6,7]. This in turn leads to the formation of a blood clot, which may occlude the vessel and cause ischemia in the tissue irrigated by it.

23 Atherosclerosis

471

23.3 Mouse Models of Atherogenesis As there is a lack of noninvasive methods to accurately detect and characterize atherosclerotic lesions in humans, mouse models with well-defined genetic modifications have been widely used to study the mechanisms involved in atherogenesis. Currently, two mouse models are most commonly used to study atherosclerosis: LDL receptor (LDLR)-deficient mice (LDLR–/–) and apolipoprotein E-deficient mice (ApoE–/–). The products of both deleted genes are involved in lipoprotein trafficking. Moreover, mutations in these two genes have been linked with familial hypercholesterolemia syndromes in humans [2], a condition associated with accelerated development of atherosclerosis, frequently resulting in early myocardial infarction. The LDL receptor is necessary for the hepatic uptake of LDL particles and the recirculation of cholesterol. LDLR–/– mice have serum cholesterol values that are only two times higher than those in normal mice. They do not develop significant atherosclerotic plaques when fed a regular diet. However, when fed a high-cholesterol diet, they rapidly develop prominent hypercholesterolemia, and numerous atherosclerotic lesions arise in their arteries [10]. Apolipoprotein E is also involved in the clearance of lipoprotein particles; patients bearing mutations in this gene have elevated levels of chylomicron (large lipoprotein particles) remnants and intermediate density lipoprotein-cholesterol [2]. ApoE–/– mice have serum cholesterol values that are four to five times higher than those in normal mice. They develop atheromas even when fed a regular diet [11], and this tendency is enhanced when fed a high-cholesterol diet.

Fig. 23.1 Induction of atherosclerosis at predilection sites. Advanced atherosclerotic lesions (arrowheads) in low-density lipoprotein (LDL) receptor-deficient mice after 14 weeks of cholesterol-rich diet are observed at locations exposed to turbulent shear stress. Whole heart-aortic arch preparation (a) and Sudan-IV–colored cryosections of aortic arch (b) and sinus (c) are shown. The aortic arch and aortic sinus (AS) as well as the brachiocephalic (BC), left common carotid (CCA) and subclavian artery (SA) are indicated. (A high-resolution version of this figure is available on the accompanying CD and online at www. springerlink.com)

472

A. Pfenniger et al.

In both models, the time course and distribution of plaques in the arterial tree are reproducible. Moreover, the distribution pattern shows strong similarity to that of the atherosclerotic lesions found in humans (Fig. 23.1). The main limitation of the mouse models is that even though they develop extensive atheromas, these lesions rarely rupture spontaneously. Since plaque rupture is the cause of all the major complications of atherosclerosis in humans [6,7], an appropriate model to study this last step of the human pathogenesis is currently lacking.

23.4 Connexins and Atherosclerosis The development of atherosclerotic plaques implies complex communication pathways among the endothelial cells, smooth muscle cells, and inflammatory cells involved in this process. Most of the suggested signaling pathways consist of paracrine communication via cytokines, chemokines, and various growth factors. However, one could imagine that a more direct and precise pathway, such as gap junction communication, could also play a role in the development of atherosclerosis. In fact, over the last 15 years there has been increasing evidence that connexins participate in atherogenesis. The first experimental support of this hypothesis was obtained by Polacek et al. [12] in 1993; studying atherosclerotic plaques in human carotid arteries, they showed that macrophage foam cells in the plaques expressed high amounts of messenger RNA for CX43, even though peripheral blood monocytes and tissue macrophages, from which these arise, did not express this connexin. They subsequently confirmed this result in a rabbit model, at the same time discovering that Cx43 was downregulated in smooth muscle cells from the media underneath the plaque [13]. These experiments suggested that the expression of connexins is altered in cells involved in atheroma formation. In 1995, Blackburn et al. [14] showed that in the smooth muscle cells of human coronary artery plaques, CX43 was first upregulated, and then downregulated in more advanced lesions. This brought forth the idea that connexin expression evolves in atherosclerotic plaques over time, depending on the stage of the lesion. More recently, CX37, another connexin found in the arterial wall, has also been investigated; a polymorphism in the CX37 gene has been linked to arterial stenosis and myocardial infarction in humans [15,16,17,18,19]. Therefore, CX37 has been proposed as a potential marker for atherosclerosis development. This indirect evidence of the involvement of connexins has greatly stimulated research in this field.

23.4.1 Connexin Expression in Atherosclerotic Plaques In a healthy vascular wall, three main connexins are expressed. Endothelial cells express both Cx37 and Cx40, and medial smooth muscle cells strongly express

23 Atherosclerosis

473

Cx43 [20,21,22,23]. Some endothelial cells, in particular portions of the vascular tree in the regions that experience highly turbulent flow, express Cx43 in addition to Cx37 and Cx40 [24]. Some studies suggest that smooth muscle cells might also express Cx40 or Cx37 [25,26]. Thus, the connexin expression patterns of endothelial and smooth muscle cells are different, even if they overlap in part. In addition, Cx45 and Cx31.9 have been found in vascular smooth muscle cells [27,28]. Using the atherosclerosis model of LDLR–/– mice fed a high-cholesterol diet, it has been shown that the pattern of connexin expression is highly regulated during the process of atherosclerotic development [29]. Cx37, which is characteristic of endothelial cells in healthy arteries, seems to be downregulated in endothelial cells during atherogenesis. After 14 weeks of high-cholesterol diet, this connexin cannot be found in the endothelium overlaying advanced atheromas, but it is still present in the endothelium of the nondiseased areas surrounding the plaques (Fig. 23.2). Surprisingly, Cx37 becomes expressed, or upregulated, in other cell types in the lesions. After ten weeks of atherogenic diet, Cx37 is present in the intima and at 14 weeks it is expressed close to the lipid core as well. At these particular sites it colocalizes with macrophage markers. This indicates that macrophages in atherosclerotic lesions express Cx37. After 14 weeks of cholesterol-rich diet, the medial smooth muscle cells underneath advanced atheromas also start to express Cx37. It is noteworthy that this connexin is not expressed in smooth muscle cells forming the fibrous cap in the intima. It therefore seems that connexin expression in those two populations of smooth muscle cells is differentially regulated.

Fig. 23.2 Connexin expression in atherosclerotic lesions. A schematic drawing of an atherosclerotic lesion in the artery wall. Expression patterns of the three vascular connexins are indicated for the nondiseased part of the artery, the shoulder, and the center of the advanced atheroma. The endothelial cells (EC), the smooth muscle cells (SMC), and the extracellular matrix (ECM) are indicated. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

474

A. Pfenniger et al.

Cx40 has a fate similar to Cx37 in the endothelium: after 14 weeks of atherogenic diet its expression is lost in the endothelium overlaying advanced atheromas (Fig. 23.2). One can therefore imagine that some of the regulatory pathways might be similar for those two connexins. However, unlike Cx37, Cx40 is not expressed by other types of cells in advanced atheromas. Cx43 is first upregulated and then downregulated in smooth muscle cells of atherosclerotic lesions, as mentioned earlier. Unlike the observations for Cx37, Cx43 is present in the smooth muscle cells in the fibrous cap as well as in the media (Fig. 23.2). Both smooth muscle cell populations seem to express less Cx43 in advanced atheroma after 14 weeks of high-cholesterol diet. Macrophages in advanced lesions start to express Cx43 as well. Interestingly, a subpopulation of endothelial cells in advanced atheromas, at the shoulder region of the plaque, also expresses Cx43. In contrast, endothelial cells covering the center of the plaque do not express any connexins. This has a particular significance, because the shoulder region is subjected to turbulent flow. Moreover, it is generally considered the weak point of the plaque, the place where it is most prone to rupture. If we focus on the different cell types, endothelial cells normally express Cx37 and Cx40. As the lesion progresses, the endothelium covering the whole plaque loses expression of both these connexins, and the endothelial cells at the shoulder region start to express Cx43. Smooth muscle cells normally express Cx43. As atherosclerosis progresses, a subset of smooth muscle cells migrates and proliferates in the intima; in those cells, Cx43 is first upregulated and then downregulated. In the smooth muscle cells remaining in the media, Cx43 is also downregulated, but these cells start to express Cx37 as well. Monocytes, which express small amounts of Cx37, migrate into the plaques and take up oxidized phospholipids to become macrophage foam cells that show enhanced Cx37 expression. Macrophage foam cells in advanced lesions coexpress Cx37 and Cx43. Each connexin forms junctional channels with different permeability and regulation properties. Therefore, the modification of the expression pattern of these three connexins might differentially regulate the communication among the different cell types involved.

23.4.2 The Role of Connexins in Atherosclerosis The study of connexins in atherosclerotic plaques of various stages has established that the expression of these proteins is modified during the progression of the disease. The data suggest that connexins play a role in the pathogenesis, but do not establish causality between connexins and atherosclerosis. More extensive research on the vascular connexins has been conducted in recent years to further investigate this association. For this purpose, mouse models have been created through interbreeding connexin knockout mice and atherosclerosisprone mice (LDLR–/– or ApoE–/– mice).

23 Atherosclerosis

475

23.4.2.1 Connexin43 This connexin is widely expressed throughout the entire body and plays a major role in cardiac function. In fact, homozygous Cx43KO mice die at birth because of severe cardiac malformations [30]. Therefore, Gja1þ/ mice bred with LDLR–/– mice, being LDLR-deficient and expressing half the normal amount of Cx43, are used as a model to study the effect of this connexin on atherosclerosis. They have been compared to Gja1þ/þLDLR–/– mice after both groups were subjected to a high-cholesterol diet [31,32]. Importantly, LDLR–/– mice homozygous and heterozygous for Cx43 have comparable lipid profiles, leukocyte counts and body weight, and have a similar response to a high-cholesterol diet [31]. Thus, Gja1þ/ mice do not have enhanced risk factors for atherosclerosis (such as hypercholesterolemia) when compared to their Gja1þ/þ littermates. The aortas and aortic roots of both groups displayed atherosclerotic plaques. Interestingly, in spite of an absence of enhanced risk factors, the surface area of the atherosclerotic lesions is reduced by 50% in Gja1þ/– mice. The composition of the atherosclerotic plaques is altered as well: there is a smaller lipid core, a reduced amount of inflammatory cells, and a thicker fibrous cap containing more smooth muscle cells and collagen. In fact, the plaques are globally more stable in Cx43 heterozygous mice than in homozygous mice. Thus, the reduction of Cx43 expression by half seems to reduce the development of atherosclerotic plaques, as well as to render them less prone to rupture. In other words, these results suggest that Cx43 has an overall atherogenic effect. Several mechanisms have been suggested to explain these results [32]. First, Cx43 may participate in the endothelial dysfunction at the shoulder regions of the plaque. Expression of Cx43 in those cells might increase their expression of adhesion molecules and cytokines, which would promote leukocyte transmigration at this location. Cx43 is also expressed in macrophages of advanced lesions, and one could imagine that their proliferation would be promoted, or that they might secrete more metalloproteinases and cytokines, which in turn would lead to an unstable plaque. Finally, Cx43 is expressed in smooth muscle cells; the presence of this connexin could decrease the migration and proliferation of those cells in the intima, and inhibit the production of collagen, which would result in a weakening of the fibrous cap. All these effects would enhance conditions that render the atherosclerotic lesion more prone to rupture. The results suggest that it could be therapeutically useful to inhibit Cx43-mediated intercellular communication or to decrease Cx43 expression in atherosclerotic vessels. In this respect, an interesting observation has been made upon statin treatment, a well-known cholesterol-lowering class of drugs. When Gja1þ/þLDLR–/– mice were treated with pravastatin concomitantly with a high-cholesterol diet, their atherosclerotic lesions exhibited properties similar to the plaques found in Gja1þ/LDLR–/– mice (plaques were smaller, contained less inflammatory cells and lipids, and more fibrous material [31]). Moreover, it has been shown that

476

A. Pfenniger et al.

statins reduce the expression of Cx43 in human endothelial and smooth muscle cells in vitro as well as in mouse atheroma in vivo. Of note, unlike humans, mice treated with statins do not have lower cholesterol serum values, because the targeted enzyme, hepatic hydroxymethylglutaryl coenzyme A reductase, is upregulated. Therefore, it seems likely that a part of the beneficial effect of statin treatment on atherosclerosis might be mediated by the downregulation of Cx43 in vessels rather than its effect on cholesterol levels. Moreover, simvastatin has been shown to reverse the hyperlipidemia-induced downregulation of endothelial Cx37 in mice, thus possibly yielding an additional connexinmediated benefit of this treatment [33]. 23.4.2.2 Connexin37 A polymorphism of the CX37 gene resulting in a proline to serine amino acid substitution at position 319 has been discovered in the human population. Several population studies on patients suffering from arterial stenosis or myocardial infarction suggested that the CX37 polymorphism affects the development of atherosclerosis in humans [15,16,17,18,19]. This implies a role for CX37 in atherogenesis. To study this connexin, Cx37KO mice were interbred with ApoE-deficient mice. When fed a high-cholesterol diet, those Gja4–/–ApoE–/– mice displayed more or larger atheromas than did their Gja4þ/þApoE–/– littermates [34]. Thus, the presence of Cx37 seems to limit the extent of atherosclerosis, and this protein might play a protective role. Further, it was demonstrated that Cx37 decreases leukocyte transmigration into the intima, one of the prerequisite steps of atherogenesis. Adoptive transfers, when macrophages extracted from one mouse are injected into another, showed that absence of Cx37 in macrophages, but not in endothelial cells, increased the number of macrophages subsequently found in atherosclerotic plaques. In vitro studies on isolated macrophages suggest that the presence or absence of Cx37 expression inversely modulates adhesion, one of the early steps of leukocyte recruitment. In fact, Gja4–/– macrophages adhere more than Gja4þ/þ macrophages on nonbiological surfaces such as glass coverslips or culture dishes as well as onto a monolayer of endothelial cells. Because macrophages are individual cells and do not form gap junctions, Cx37 hemichannels might play a role in the regulation of cell adhesion. A potential mediator of this regulation pathway is adenosine triphosphate (ATP), because Cx37-expressing macrophages release more ATP into the extracellular space than do Cx37-deficient cells. Moreover, when extracellular ATP is degraded by the enzyme apyrase, an increase in adhesion of Cx37-expressing macrophages is observed. This suggests that Cx37 mediates a decrease in leukocyte adhesion through an autocrine ATP-dependent mechanism, which in term would decrease the extent of atherosclerotic plaques [34]. To determine whether the Cx37 polymorphism had an effect on this antiadhesive property, macrophage-derived cell lines were transfected with each isoform of Cx37 (namely Cx37P319 and Cx37S319) and subjected to

23 Atherosclerosis

477

similar adhesion experiments. There was a significant difference in adhesion between these two transfectants, with the cells expressing Cx37P319 adhering less than the cells expressing the Cx37S319 polymorphism, which is proposed as a marker for atherosclerotic plaque development [34]. These results correlate well with clinical observations attributing a protective effect to Cx37P319 against myocardial infarction [17,18]. These experiments point to a specific relation between Cx37 and early atherogenesis. This connexin may also play a role in the involvement of other cell types (such as endothelial cells) and at later stages of atherogenesis.

23.4.2.3 Connexin40 This connexin has been much less studied in relation to atherosclerosis than Cx43 or Cx37. Most of the data are very recent. A problem for controlled study is that homozygous Cx40KO mice (Gja5–/–) suffer from hypertension [35,36,37], a major independent risk factor for atherosclerosis. To overcome this limitation, a mouse model with a selective deletion of Cx40 in endothelial cells has been designed using the Cre-LoxP system for tissue-specific deletion using the TIE2 promoter (see Chapter 22). TIE2-Cre+Gja5fl/flApoE–/– mice (i.e., mice lacking Cx40 in the endothelium) were compared to Gja5fl/flApoE–/– control mice to confirm that the selective deletion of Cx40 in the endothelium did not affect the mean arterial pressure or the heart frequency [38]. When both groups were fed a high-cholesterol diet for ten weeks, the surface of atherosclerotic plaques was significantly increased in the mice lacking Cx40 in the endothelium [38]. These data suggest a potential atheroprotective role for Cx40. One could imagine that this protection is lost in the arterial regions that become atheromatous, as this connexin disappears from the endothelium covering atheromas. The mechanisms underlying the atheroprotection remain to be explained.

23.5 Conclusion Three connexins are involved in the development of atherosclerosis. Cx43 seems to have an overall atherogenic effect, and Cx37 and Cx40 display atheroprotective properties. So far, one potential mechanism has been clearly demonstrated, namely the inhibitory effect of Cx37 on monocyte adhesion. Thus, the next challenge is to identify the mechanisms that involve Cx40 and Cx43. Another remaining question is whether the results obtained in mouse models can be applied to humans. Although mouse models closely mimic human pathology, a few differences remain, such as the absence of plaque rupture in mice, which could result from differences in the expression or activity of connexins in these two species. This problem awaits the development of new animal models as well as development of minimally invasive methods that would enable the findings of animal models to be confirmed in humans.

478

A. Pfenniger et al.

Acknowledgments The authors thank Cindy Wong, Christos Chadjichristos, Jean-Paul Derouette, Sandrine Morel, Bernard Foglia, Esther Sutter, and Marc Chanson for helpful discussions. This work was supported by the Swiss National Science Foundation (grant PPOOA-116897).

References 1. Murray CJ, Lopez AD. Global mortality, disability, and the contribution of risk factors: global burden of disease study. Lancet. 1997;349:1436–42. 2. Schoen FJ, Cotran RS. Atherosclerosis. In: Cotran RS, Kumar V, Collins T, editors. Robbins pathologic basis of disease, 6th ed. Philadelphia: WB Saunders; 1999. pp. 498–510. 3. Ross R. Atherosclerosis: an inflammatory disease. N Engl J Med. 1999;340:115–26. 4. Davies PF, Spaan JA, Krams R. Shear stress biology of the endothelium. Ann Biomed Eng. 2005;33:1714–8. 5. Glass CK, Witztum JL. Atherosclerosis: the road ahead. Cell. 2001;104:503–16. 6. Libby P. Inflammation in atherosclerosis. Nature. 2002;420:868–74. 7. Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med. 2005;352:1685–95. 8. Libby P, Hansson GK. The immune response in atherosclerosis: a double-edged sword. Nat Rev Immunol. 2006;6:508–19. 9. Tedgui A, Mallat Z. Cytokines in atherosclerosis: pathogenic and regulatory pathways. Physiol Rev. 2006;86:515–81. 10. Ishibashi S, Goldstein JL, Brown MS, Herz J, Burns DK. Massive xanthomatosis and atherosclerosis in cholesterol-fed low density lipoprotein receptor-negative mice. J Clin Invest. 1994;93:1885–93. 11. Zhang SH, Reddick RL, Burkey B, Maeda N. Diet-induced atherosclerosis in mice heterozygous and homozygous for apolipoprotein E gene disruption. J Clin Invest. 1994;94:937–45. 12. Polacek D, Lal R, Volin MV, Davies PF. Gap junctional communication between vascular cells. Induction of connexin43 messenger RNA in macrophage foam cells of atherosclerotic lesions. Am J Pathol. 1993;142:593–606. 13. Polacek D, Bech F, McKinsey JF, Davies PF. Connexin43 gene expression in the rabbit arterial wall: effects of hypercholesterolemia, balloon injury and their combination. J Vasc Res. 1997;34:19–30. 14. Blackburn JP, Peters NS, Yeh HI, Rothery S, Green CR, Severs NJ. Upregulation of connexin43 gap junctions during early stages of human coronary atherosclerosis. Arterioscler Thromb Vasc Biol. 1995;15:1219–28. 15. Boerma M, Forsberg L, Van Zeijl L, L Morgenstern R, De Faire U, Lemne C, Erlinge D, Thulin T, Hong Y, Cotgreave IA. A genetic polymorphism in connexin 37 as a prognostic marker for atherosclerotic plaque development. J Intern Med. 1999;246: 211–18. 16. Yeh H-I, Chou Y, Liu H-F, Chang S-C, Tsai C-H. Connexin37 gene polymorphism and coronary artery disease in Taiwan. Int J Cardiol. 2001;81:251–5. 17. Yamada Y, Izawa H, Ichihara S, Takatsu F, Ishihara H, Hirayama H, Sone T, Tanaka M, Yokota M. Prediction of the risk of myocardial infarction from polymorphisms in candidate genes. N Engl J Med. 2002;347:1916–23. 18. Listi F, Candore G, Lio D, Russo M, Colonna-Romano G, Caruso M, Hoffmann E, Caruso C. Association between C1019T polymorphism of connexin37 and acute myocardial infarction: study in patients from Sicily. Int J Cardiol. 2005;102:269–71. 19. Wong CW, Christen T, Pfenniger A, James RW, Kwak BR. Do allelic variants of the connexin37 1019 gene polymorphism differentially predict for coronary artery disease and myocardial infarction? Atherosclerosis. 2007;191:355–61.

23 Atherosclerosis

479

20. van Kempen MJ, Jongsma HJ. Distribution of connexin37, connexin40 and connexin43 in the aorta and coronary artery of several mammals. Histochem Cell Biol. 1999;112: 479–86. 21. Severs NJ, Rothery S, Dupont E, Coppen SR, Yeh HI, Ko YS, Matsushita T, Kaba R, Halliday D. Immunocytochemical analysis of connexin expression in the healthy and diseased cardiovascular system. Microsc Res Tech. 2001;52:301–22. 22. Hill CE, Phillips JK, Sandow SL. Heterogeneous control of blood flow amongst different vascular beds. Med Res Rev. 2001;21:1–60. 23. Figueroa XF, Isakson BE, Duling BR. Connexins: gaps in our knowledge of vascular function. Physiology. 2004;19:277–84. 24. Gabriels JE, Paul DL. Connexin43 is highly localized to sites of disturbed flow in rat aortic endothelium but connexin37 and connexin40 are more uniformly distributed. Circ Res. 1998;83:636–43. 25. Little TL, Beyer EC, Duling BR. Connexin43 and connexin40 gap junctional proteins are present in arteriolar smooth muscle and endothelium in vivo. Am J Physiol. 1995;268:H729–39. 26. Nakamura K, Inai T, Nakamura K, Shibata Y. Distribution of gap junction protein connexin37 in smooth muscle cells of the rat trachea and pulmonary artery. Arch Histol Cytol. 1999;62:27–37. 27. Li X, Simard JM. Multiple connexins form gap junction channels in rat basilar artery smooth muscle cells. Circ Res. 1999;84:1277–84. 28. Nielsen PA, Beahm DL, Giepmans BN, Baruch A, Hall JE, Kumar NM. Molecular cloning, functional expression, and tissue distribution of a novel human gap junctionforming protein, connexin-31.9. Interaction with zona occludens protein-1. J Biol Chem. 2002;277:38272–83. 29. Kwak BR, Mulhaupt F, Veillard N, Gros DB, Mach F. Altered pattern of connexin expression in atherosclerotic plaques. Arterioscler Thromb Vasc Biol. 2002;22:225–30. 30. Reaume AG, de Sousa PA, Kulkarni S, Lanqille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science. 1995;267:1831–4. 31. Kwak BR, Veillard N, Pelli G, Mulhaupt F, James RW, Chanson M, Mach F. Reduced connexin43 expression inhibits atherosclerotic lesion formation in low-density lipoprotein receptor-deficient mice. Circulation 2003;107:1033–9. 32. Wong CW, Burger F, Pelli G, Mach F, Kwak BR. Dual benefit of reduced Cx43 expression on atherosclerosis in LDL receptor-deficient mice. Cell Commun Adhes. 2003;10:395–400. 33. Yeh HI, Lu CS, Wu YJ, Chen CC, Hong RC, Ko YS, Shiao MS, Severs NJ, Tsai CH. Reduced expression of endothelial connexin37 and connexin40 in hyperlipidemic mice: recovery of connexin37 after 7-day simvastatin treatment. Arterioscler Thromb Vasc Biol. 2003;23:1391–7. 34. Wong CW, Christen T, Roth I, Chadjichristos CE, Derouette JP, Foglia BF, Chanson M, Goodenough DA, Kwak BR. Connexin37 protects against atherosclerosis by regulating monocyte adhesion. Nature Med. 2006;12:950–4. 35. De Wit C, Roos F, Bolz SS, Kirchhoff S, Kruger O, Willecke K, Pohl U. Impaired conduction of vasodilation along arterioles in connexin40-deficient mice. Circ Res. 2000;86:649–55. 36. De Wit C, Roos F, Bolz SS, Pohl U. Lack of vascular connexin40 is associated with hypertension and irregular arteriolar vasomotion. Physiol Genomics. 2003;13:169–77. 37. Wagner C, de Wit C, Kurtz L, Grunberger C, Kurtz A, Schweda F. Connexin40 is essential for the pressure control of renin synthesis and secretion. Circ Res. 2007;100:556–63. 38. Chadjichristos CE, Roth I, Hoepfl B, van Veen TA, Deutsch U, van Kempen MJ, de Wit C, Kwak BR. Increased development of atherosclerosis in mice with endothelial-specific deletion of connexin40. Circulation. 2005;112:II-142.

Chapter 24

Connexins in the Female Reproductive System Gerald M. Kidder and Elke Winterhager

Abstract Connexins play diverse roles in the reproductive biology of females. In the ovary, connexin expression is regulated by pituitary gonadotropins. In developing ovarian follicles, gap junctions couple the growing oocyte and its surrounding follicle (granulosa) cells into a functional syncytium allowing nutrients and signaling molecules to pass between the two cell types. In the mouse, the gap junctions between granulosa cells are composed of Cx43, whereas those coupling the granulosa cells with the growing oocyte are composed of Cx37. In the absence of either connexin, oocyte and follicle development are impaired. Additional connexins have been identified in antral and preovulatory follicles of mice and other animals, but their roles have not been elucidated. Although human follicle cells express CX43, females with dominant inactivating mutations in the gene GJA1 (encoding CX43) are generally fertile, suggesting that this connexin may play a less critical role in the human ovary than in the murine ovary. In the uterus, gap junction channels composed of Cx26 (in the endometrial epithelium) and Cx43 (in the endometrial stroma) are temporally and spatially regulated by both ovarian hormones and by embryonic signals. Cx26 and Cx43 are inducible by estrogen via its receptor ER, and complete suppression of both connexin genes in the preimplantation phase is produced by progesterone. At implantation, connexins are induced specifically in the epithelium of the implantation chamber due to embryo (blastocyst) recognition, an event mediated by a signaling cascade independent of estrogen receptors. The physiological role of this highly regulated expression of endometrial connexins is not known. In the rodent myometrium, Cx43 is the predominant connexin and is strongly hormonally regulated. It is downregulated by progesterone, leading to a quiescent uterus during pregnancy, and is upregulated by an increased estrogen/progesterone ratio. Experiments with myometrium-specific knockout mice have proved that Cx43 induction at term is important for contraction during labor. G.M. Kidder (*) Department of Physiology and Pharmacology, Schulich School of Medicine and Dentistry, University of Western Ontario, London, Ontario N6A 5C1, Canada e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_24, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

481

482

G.M. Kidder and E. Winterhager

Keywords Ovary Oogenesis Folliculogenesis Uterus Endometrium Myometrium Implantation Cx37 Cx43

24.1 Introduction Gap junctions play a variety of developmental and physiological roles in the male and female reproductive system. This chapter summarizes the current knowledge of connexin expression and gap junction function in the ovary and uterus, two well-studied components of the female reproductive system. Research on these organs has focused on identifying the connexins involved, the regulation of their expression by hormones and other influences, and their specific roles as revealed using connexin mutant mice. One complicating factor has been that, since reproductive function is gained relatively late (i.e., puberty), analysis of reproductive deficiencies in connexin knockout (KO) mice is sometimes precluded by fetal or postnatal lethality due to disruption of essential developmental events or physiological functions. Hence fertility, the ultimate test of reproductive function, cannot always be assessed.

24.2 The Ovary Connexin expression in the ovary is subject to temporal variation due to the changing hormonal environment during each cycle of follicular growth and ovulation. In adult rodents, where most investigations have been carried out, the ovaries contain a range of follicles in different stages of development. Hence, some studies have used pubertal females, in which the first wave of follicular development is relatively synchronous. These studies have identified Cx43 and Cx37 as playing essential roles in oogenesis.

24.2.1 Ovarian Follicle Development The ovarian follicle consists of a single oocyte, layers of surrounding somatic cells (follicle or granulosa cells), and an outer rim of theca cells (Fig. 24.1) [1]. The follicle grows through proliferation of granulosa cells, adding additional layers until a fluid-filled cavity, the antrum, develops. The antrum separates the granulosa cells into two subpopulations, the cumulus granulosa surrounding the oocyte and the mural granulosa adjacent to the theca layer. Within the follicle, the oocyte grows and develops competence to undergo meiosis and be fertilized. During each ovarian cycle, depending on the species, one or more growing follicles become mature, whereas others undergo atresia (death by apoptosis). During their development, follicles pass from the primordial (nongrowing) stage through primary (unilaminar) and secondary (multilaminar) growing stages until becoming tertiary

24 Female Reproduction

483

Fig. 24.1 Architecture of the ovarian follicle. The developing oocyte is surrounded by layers of granulosa cells, which are metabolically coupled with the oocyte via gap junctions. In mice, the gap junctions connecting the granulosa cells with each other are composed of Cx43, whereas those connecting the oocyte with the granulosa cells are composed of Cx37. Both types of gap junctions are essential for oogenesis. (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com) (From Simon and Goodenough [1] with permission.)

(antral) follicles. A fully-grown antral follicle, ready for ovulation, is called a preovulatory or Graafian follicle. After ovulation, the follicle remnant develops into a highly vascularized endocrine organ, the corpus luteum. During the growth of mammalian oocytes there is continuous coupling with the surrounding granulosa cells via gap junctions, and the granulosa cells themselves are coupled to each other. A range of molecules including amino acids, glucose metabolites, and nucleotides are transferred to the growing oocyte via the gap junctional pathway (reviewed in [2,3]). In addition, signals that regulate meiotic maturation of fully grown oocytes pass through the oocyte-granulosa cell gap junctions (reviewed in [4]).

24.2.2 Ovarian Follicles Numerous connexins have been detected within the follicles of various species (summarized in [5]). However, the mouse is the only experimental model in which roles for connexins in oogenesis have been unequivocally demonstrated, based largely on mutational analysis. Cx43 is continuously expressed in fetal mouse ovaries from the onset of ovarian differentiation [6,7]. In growing follicles its expression increases in response to rising follicle-stimulating hormone (FSH) levels (reviewed in [8]). Cx43 forms numerous large gap junctions between granulosa cells. Since its ablation by gene targeting completely abolishes coupling among those cells, it was concluded that Cx43 alone contributes to gap junction channels in preantral granulosa cells [9,10]. Cx32 and

484

G.M. Kidder and E. Winterhager

Cx45 are also present, at least in gap junctions of antral follicles [5], but their functions have not yet been identified: Cx32KO mice are fully fertile [11], whereas Cx45KO mice die in utero, precluding studies of ovarian function [12,13]. In contrast with gap junctions linking granulosa cells, junctions linking the oocyte with surrounding granulosa cells in growing mouse follicles are composed exclusively of Cx37. Ablation of this connexin by gene targeting removes all gap junctions from the oocyte surface and disrupts oocyte-granulosa cell coupling [14,15]. Ablation of Cx43, however, does not disrupt coupling between oocytes and granulosa cells, indicating that Cx43, though present at or very near the oocyte surface, does not contribute significantly to those gap junctions [14,15,16,17]. Figure 24.1 illustrates the current view of the defined contributions of Cx37 and Cx43 to gap junctions in growing mouse follicles. The distinct roles of Cx37 and Cx43 in ovarian development and folliculogenesis have been elucidated through analysis of knockout mice. Mice lacking Cx37 are viable. In Cx37KO females, folliculogenesis proceeds until a late preantral stage before oocyte growth ceases prematurely, before acquisition of full meiotic competence [14,18]. Eventually, the oocytes degenerate and the mutant ovaries become filled with structures resembling corpora lutea, as though the granulosa cells had differentiated prematurely to become luteal cells. Thus, it is likely that signals passing from the oocyte to the granulosa cells via Cx37 channels are required to maintain the differentiated state of the granulosa cells, preventing them from luteinizing before ovulation. Folliculogenesis was also disrupted by the targeted ablation of the gene encoding Cx43 (Gja1) [6,19]. In these mice, in addition to a reduction in the number of germ cells in the fetal gonads, oocyte growth was retarded and folliculogenesis was either arrested in the primary stage or retarded, depending on the parental mouse strain. These results were interpreted as indicating a requirement for coupling among granulosa cells via Cx43 channels to sustain granulosa cell proliferation [19]. Subsequent work showed that the proliferation deficit of Cx43KO granulosa cells is due, at least in part, to reduced responsiveness to an oocyte-derived mitogen, growth differentiation factor-9 (GDF9; [20]). Oocytes obtained from Cx43KO follicles are morphologically abnormal, meiotically incompetent, and could not be fertilized [19]. Release of the oocyte from the mature follicle (ovulation) is triggered by a surge of luteinizing hormone (LH) and involves downregulation of Cx43 expression and loss of intercellular coupling within the granulosa cell layers [8]. This loss of coupling is hypothesized to restrict meiosis-arresting signals from reaching the oocyte, thus triggering the first meiotic division.

24.2.3 Corpora Lutea The corpus luteum is a complex, highly vascularized endocrine structure whose major product is progesterone (reviewed in [21]). As for the preovulatory follicle,

24 Female Reproduction

485

several connexins or their messenger RNAs have been identified in corpora lutea of various species; the list includes Cx26, Cx32, and Cx43 in sheep [22,23,24], Cx30.3 and Cx43 in pig [25], and Cx43 in cow, baboon, human, and rat [26,27,28]. Their expression changes with luteal growth and regression. However, to date there are no definitive data indicating what roles these connexins play in luteal function. Given the neonatal death of mice lacking Cx43 [29], conditional knockout of Gja1 is required to allow experimental studies of corpora lutea in that mammal.

24.2.4 The Human Ovary Given the importance of connexins in mouse ovarian function, it is of interest to explore their roles in human female fertility. CX43 is expressed in granulosa cells of human ovaries [30] and there is evidence of a correlation between levels of mRNA for CX43 in human granulosa cells and the developmental competence of the associated oocytes [31]. It has been suggested that an important function of gap junctional coupling during folliculogenesis is to supply the growing oocyte with glucose to support glycolysis in the hypoxic environment of the follicle [32]. If this supposition is correct, then an impairment of oocyte quality would be an expected outcome if gap junctional coupling within the follicle were impaired. However, women carrying one of the mutant alleles of GJA1 that cause the dominant oculodentodigital dysplasia (ODDD) syndrome are generally fertile [33,34], although detailed tests of their ovarian function have not been reported. The ODDD-causing mutations severely reduce CX43mediated gap junctional coupling [35,36]; hence, it is curious that these women do not appear to have reproductive problems. Cx43 may play a less critical role in the fertility of human females than it does in the mouse. If so, one may infer either that gap junctional coupling between granulosa cells in the human is less important, or that another connexin is involved along with CX43.

24.3 The Uterus Expression of connexins in the uterus is correlated with different physiological events such as transformation of the uterus during the preimplantation phase, embryo implantation, and parturition. Endometrial connexins are precisely regulated during preimplantation and periimplantation phases, whereas at birth, the connexins of the myometrial compartment become prominent in facilitating parturition.

24.3.1 The Endometrium In preparation for embryo implantation, the endometrium differentiates into a receptive state that allows adhesion and invasion of the trophoblast [37]. This

486

G.M. Kidder and E. Winterhager

endometrial differentiation is primarily coordinated by the ovarian hormones, progesterone and estrogen, which modulate uterine events in a spatiotemporal manner. This hormonal modulation is a prerequisite for the endometrium to interact with blastocyst signals [38], and together these two types of signaling influence connexin gene expression. The main connexin isoforms expressed in the rodent endometrium have been identified as Cx26 and Cx43 [39]. 24.3.1.1 Expression During the Preimplantation Phase In nonpregnant rats only the uterine epithelium of the estrous phase shows Cx26 immunoreactivity, whereas Cx43 is found in the stromal compartment during cycling. In pregnancy, the connexin expression pattern changes dramatically. During the first 3 days, when rising progesterone levels drive endometrial differentiation to the receptive phase, both connexins — Cx26 in the uterine epithelium and Cx43 in the stromal compartment — are rapidly downregulated before implantation [40]. A search for compensating upregulation of other connexins, including Cx30, Cx32, Cx37, Cx40, and Cx45, gave negative results. Thus in the preimplantation period, it is likely that cells in the rodent endometrium lack direct junctional communication, as already shown in an earlier study in the rabbit [41]. This phenomenon of a noncommunicating epithelium in very early pregnancy is most unusual since cells of most epithelial tissues are connected by connexin channels. Endometrial connexins in rats are under the control of maternal progesterone during the preimplantation phase, since treatment with antiprogestin prevented the physiological suppression of connexin transcripts encoding both Cx26 and Cx43 [42]. Only supraphysiological estradiol concentrations during the receptive phase led to an induction of Cx26 in the uterine epithelium, overriding the suppressive effect of progesterone. Compared to the gene encoding Cx43, expression of the Cx26 gene, Gjb2, is much more responsive to estrogen. As Cx26 expression was inhibited by simultaneous application of an antiestrogen, estrogen receptor–mediated regulation is likely. Since a similar spatiotemporal pattern of connexin expression occurs in the mouse endometrium, mice were used to clarify the roles of the different estrogen receptors. Experiments with knockout mice lacking estrogen receptor (ERKO) or b (ERbKO) reveal that induction of connexins is mediated via ER but not ERb [43]. In summary, the expression of Cx26 and Cx43 in the rat endometrium is sensitively and differentially regulated by estrogen and progesterone. Transcription of the gene encoding Cx26 is more sensitive to estrogen than that encoding Cx43, and only a hormonal milieu similar to that in early pregnancy is able to suppress transcription of both genes (Fig. 24.2). 24.3.1.2 Expression at Implantation With the onset of the implantation reaction there is a strong induction of connexin expression in the epithelial cells of the implantation chamber.

24 Female Reproduction

487

Fig. 24.2 Scheme depicting the regulation of epithelial Cx26 expression during preimplantation and periimplantation stages of pregnancy in rodents. P, progesterone; E2, estradiol. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

Transcripts for Cx26 are already expressed in both rat and mouse by day four postcoitum, followed one day later by expression of the Cx26 protein restricted to the luminal epithelium of the implantation chamber. This seems to be a local effect due to the presence of a blastocyst because it does not occur between implantation sites or in pseudopregnant animals, where a blastocyst is lacking [39]. This local restriction of Cx26 induction has been confirmed by comparison of mouse uterine tissue of implantation and interimplantation sites by microarray analysis, which demonstrates that upregulation of Cx26 occurs only in the implantation sites [44]. As in rodents, in rabbits a strong expression of functional gap junction channels is induced by the blastocyst in the uterine epithelium of the implantation chamber as a response to embryo recognition, as evidenced by tubal ligation experiments. In contrast to rodents, however, in rabbits a different connexin, Cx32, connects the epithelial cells of the implantation chamber [41]. The progesterone-mediated suppression of connexins outside the implantation chamber can be abolished not only by elevated estrogen via an estrogen-receptor mediated pathway, but also by an unknown signal from the implanting embryo (Fig. 24.2). Blastocyst signaling is mediated by the estrogen receptor–independent signaling cascade, since application of a pure antiestrogen to pregnant rats had no effect on epithelial Cx26 induction [43]. Cx43 is also involved in implantation. In rodents, the implantation process is accompanied by a transformation of the endometrial stroma into a compact decidual tissue. Concomitant with the decidualization of the stromal cells surrounding the implantation chamber, intense staining for Cx43 is seen in this primary decidual zone [45]. Cx43 expression increases and spreads out with ongoing decidualization in both rodent species. In contrast to mice [46], in the rat Cx26 is coexpressed with Cx43 in the primary decidual zone [39].

488

G.M. Kidder and E. Winterhager

It remains to be clarified if the tremendous increase in Cx43 mRNA and protein expression in the decidualizing stromal cells is regulated via ER or is induced by the implanting blastocyst. It was demonstrated that artificially induced decidualization in pseudopregnant rats leads to an induction of Cx43 expression in the deciduoma. As in the case of the epithelial Cx26 response, the expression of Cx43 in the decidual cells is not changed by application of an antiestrogen. Furthermore, ERKO mice decidualize and express high amounts of Cx43 [43]. Taken together, these experiments demonstrate that Cx43 upregulation concomitant with the decidualization process is independent of ER action. Until recently, it was not possible to demonstrate the physiological importance either of the local induction of epithelial Cx26 at implantation sites or the enhanced expression of Cx43 during decidualization because loss of Cx26 causes early gestational lethality in knockout mice due to an impairment in glucose transport across the placental barrier [47], and Cx43 deficiency results in death directly after birth [29]. Experiments with appropriate tissue-specific knockouts for the two connexins in the uterine epithelium as well as in the decidua are still lacking. One may hypothesize that the epithelial gap junction channels help to spatially restrict local apoptosis as the invading trophoblasts replace epithelial cells during rodent implantation. In support of this idea, a correlation between cells expressing Cx26 and those undergoing apoptosis has been observed [48].

24.3.1.3 The Human Endometrium Corresponding to the situation in rodents, CX26 and CX43 are the main connexin isoforms expressed in the human endometrium and their expression appears to be hormonally regulated during the menstrual cycle. Three gap junction isoforms, CX26, CX43, and to a lesser extent CX32, have been detected, each with distinct distribution and spatial and temporal regulation [49,50]. CX26 is expressed with increasing intensity in the uterine epithelial cells during the course of the proliferative phase. However, CX26 could hardly be detected in the late proliferative phase and the secretory phase. CX43 expression characterizes the stromal compartment and shows a similar regulation with a decrease during the secretory phase. In summary, connexins are localized to different tissue compartments in the human endometrium and are likely to be hormonally regulated in the course of the menstrual cycle. Cell–cell communication could serve as a general control mechanism for coordinated tissue differentiation upon sex hormone stimulation.

24.3.2 The Myometrium The importance of gap junction induction in the myometrium in preparation for parturition was documented in 1977 by Garfield et al. [51]. Since then, numerous

24 Female Reproduction

489

publications have investigated the electrophysiological propagation of contraction waves via gap junction channels during the birth process and the regulation of myometrial preparation by steroid hormones. Meanwhile, four different connexins, Cx26, Cx40, Cx43, and Cx45, have been identified in the myometrium, the most prominent being Cx43. Well-defined hormonal regulation of gap junctional communication via Cx43 channels has been observed in the myometrium. This myometrial Cx43 expression is suppressed by progesterone [52] and elevated by estrogen [53]. Downregulation of Cx43 expression in the myometrium by progesterone maintains the smooth muscle cells in a quiescent state, disabling synchronous muscular contraction that could lead to premature labor. In normal pregnancies, the onset of labor is indicated by a tremendous increase in gap junctional communication that leads to coordinated electrical activity in the myometrium prior to parturition [54,55]. At least in rodents, this phenomenon has been correlated with an increase in the estrogen/progesterone ratio [56]. Myometrial-specific ablation of Cx43 in the mouse has further clarified the role of this connexin in parturition [57]. Interestingly, most of these mice suffered a prolongation of pregnancy, but parturition was not inhibited. Furthermore, all other term-associated markers, such as the oxytocin receptor, prostaglandin F receptor, and progesterone serum levels remained unchanged, pointing to a predominantly electrical coupling function of the Cx43 channel at parturition. In contrast to Cx43, Cx26 expression is highest during late pregnancy but decreases at term. Its upregulation in the endometrium is progesterone dependent [58]. Cx45 is present in the nonpregnant uterus and during early pregnancy but declines thereafter [59]. No functional role has been suggested yet for Cx26, Cx40, or Cx45. Cx40KO female mice, which are viable and fertile [60,61], are not reported to have altered parturition.

24.3.2.1 The Human Myometrium As in rodents, the main connexin isoform expressed in the human myometrium is CX43, with CX40 and CX45 having been detected as well [62]. However, these channels may not be under the control of maternal steroid hormone changes during pregnancy. An upregulation of CX43 transcripts similar to that in other species does occur in women shortly before and during labor [63], but the hormonal control of this in human myometrium is likely more complicated than in rodents because the estrogen/progesterone ratio does not change before parturition. Interestingly, lower expression of CX43 is associated with prolonged labor in women [64], consistent with the findings from myometrium-specific knockout mice. It is also worth noting that women with the ODDD-causing GJA1 mutations, where assembly of CX43 gap junctions is impaired, have not been reported to suffer serious complications of pregnancy or birth [33,34].

490

G.M. Kidder and E. Winterhager

24.4 Conclusion Connexins play diverse roles in the reproductive biology of females. The research reviewed here only begins to hint at the diversity of developmental and physiological functions of gap junctional communication in this context; many more examples will surely come to light as additional studies are performed with mutant mice and more detailed clinical tests are applied to human females with connexin gene mutations. Acknowledgments Work from the Kidder laboratory was funded by the Canadian Institutes of Health Research. Work from the Winterhager laboratory was funded by the Deutsche Forschungsgemeinschaft.

References 1. Simon AM, Goodenough DA. Diverse functions of vertebrate gap junctions. Trends Cell Biol. 1998;8:477–83. 2. Eppig JJ. Intercommunication between mammalian oocytes and companion follicle cells. BioEssays. 1991;13:569–74. 3. Eppig JJ, O’Brien M, Wigglesworth K. Mammalian oocyte growth and development in vitro. Mol Reprod Dev. 1996;44:260–73. 4. Edry I, Sela-Abramovich S, Dekel N. Meiotic arrest of oocytes depends on cell-to-cell communication in the ovarian follicle. Mol Cell Endocrinol. 2006;252:102–06. 5. Kidder GM. Roles of gap junctions in ovarian folliculogenesis: implications for female infertility. In: Winterhager E, editor. Gap junctions in development and disease. Berlin: Springer-Verlag; 2005. pp. 223–37. 6. Juneja SC, Barr KJ, Enders GC, Kidder GM. Defects in the germ line and gonads of mice lacking connexin43. Biol Reprod. 1999;60:1263–70. ´ 7. Pe´rez-Armendariz EM, Sa´ez JC, Bravo-Moreno JF, Lopez-Olmos V, Enders GC, Villalpando I. Connexin43 is expressed in mouse fetal ovary. Anat Rec Part A. 2003;271A:360–7. 8. Granot I, Dekel N. The ovarian gap junction protein connexin43: regulation by gonadotropins. Trends Endocrinol Metab. 2002;13:310–3. 9. Gittens JEI, Mhawi AA, Lidington D, Ouellette Y, Kidder GM. Functional analysis of gap junctions in ovarian granulosa cells: a distinct role for connexin43 in early stages of folliculogenesis. Am J Physiol Cell Physiol. 2003;284:C880–7. 10. Tong D, Gittens JEI, Kidder GM, Bai D. Patch clamp study reveals that the importance of connexin43-mediated gap junctional communication for ovarian folliculogenesis is strain-specific in the mouse. Am J Physiol Cell Physiol. 2006;290:C290–7. 11. Nelles E, Bu¨tzler C, Jung D, Temme A, Gabriel H-D, Dahl U, Traub O, Stu¨mpel F, Jungermann K, Zielasek J, Toyka KV, Dermietzel R, Willecke K. Defective propagation of signals generated by sympathetic nerve stimulation in the liver of connexin32-deficient mice. Proc Natl Acad Sci USA. 1996;93:9565–70. 12. Kru¨ger O, Plum A, Kim J-S, Winterhager E, Maxeiner S, Hallas G, Kirchhoff S, Traub O, Lamers WH, Willecke K. Defective vascular development in connexin45-deficient mice. Development 2000;127:4179–93. 13. Kumai M, Nishii K, Nakamura K, Takeda N, Suzuki M, Shibata Y. Loss of connexin45 causes a cushion defect in early cardiogenesis. Development. 2000;127:3501–12. 14. Simon AM, Goodenough DA, Li E, Paul DL. Female infertility in mice lacking connexin37. Nature. 1997;385:525–9.

24 Female Reproduction

491

15. Veitch GI, Gittens JEI, Shao Q, Laird DW, Kidder GM. Selective assembly of connexin37 into gap junctions at the oocyte/granulosa cell interface. J Cell Sci. 2004;117:2699–707. 16. Kidder GM, Mhawi AA. Gap junctions and ovarian folliculogenesis. Reproduction. 2002;123:613–20. 17. Gittens JEI, Kidder GM. Differential contributions of connexin37 and connexin43 to oogenesis revealed in chimeric reaggregated mouse ovaries. J Cell Sci. 2005;118:5071–8. 18. Carabatsos MJ, Sellitto C, Goodenough DA, Albertini DF. Oocyte-granulosa cell heterologous gap junctions are required for the coordination of nuclear and cytoplasmic meiotic competence. Dev Biol. 2000;226:167–79. 19. Ackert CL, Gittens JEI, O’Brien MJ, Eppig JJ, Kidder GM. Intercellular communication via connexin43 gap junctions is required for ovarian folliculogenesis in the mouse. Dev Biol. 2001;233:258–70. 20. Gittens JEI, Barr KJ, Vanderhyden BC, Kidder GM. Interplay between paracrine signaling and gap junctional coupling in ovarian follicles. J Cell Sci. 2005;118:113–22. 21. Reynolds LP, Redmer DA. Growth and development of the corpus luteum. J Reprod Fertil Suppl. 1999;54:181–91. 22. Grazul-Bilska AT, Redmer DA, Bilski JJ, Jablonka-Shariff A, Doraiswamy V, Reynolds LP. Gap junctional proteins, connexin 26, 32, and 43 in sheep ovaries throughout the estrous cycle. Endocrine. 1998;8:269–79. 23. Grazul-Bilska AT, Reynolds LP, Bilski JJ, Redmer DA. Effects of second messengers on gap junctional intercellular communication of ovine luteal cells throughout the estrous cycle. Biol Reprod. 2001;65:777–83. 24. Borowczyk E, Johnson ML, Bilski JJ, Borowicz PP, Redmer DA, Reynolds LP, Grazul-Bilska AT. Expression of gap junctional connexins 26, 32, and 43 mRNA in ovarian preovulatory follicles and corpora lutea in sheep. Can J Physiol Pharmacol. 2006;84:1011–20. 25. Itahana K, Morikazu Y, Takeya T. Differential expression of four connexin genes, Cx26, Cx30.3, Cx32, and Cx43, in the porcine ovarian follicle. Endocrinology. 1996;137:5036–44. 26. Grazul-Bilska AT, Redmer DA, Johnson ML, Jablonka-Shariff A, Bilski JJ, Reynolds LP. Gap junctional protein connexin43 in bovine corpora lutea throughout the estrous cycle. Biol Reprod. 1996;54:1279–87. 27. Khan-Dawood FS, Yang J, Dawood MY. Expression of gap junction protein connexin43 in the human and baboon (Papio anubis) corpus luteum. J Clin Endocrinol Metab. 1996;81:835–42. 28. Risek B, Guthrie S, Kumar NM, Gilula NB. Modulation of gap junction transcript and protein expression during pregnancy in the rat. J Cell Biol. 1990;110:269–82. 29. Reaume A, De Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science. 1995;267:1831–4. 30. Furger C, Cronier L, Poirot C, Pouchelet M. Human granulosa cells in culture exhibit functional cyclic AMP-regulated gap junctions. Mol Human Reprod. 1996;2:541–8. 31. Tsai M-Y, Lan K-C, Huang K-E, Huang F-J, Kung F-T, Chang S-H. Significance of mRNA levels of connexin37, connexin43, and connexin45 in luteinized granulosa cells of controlled hyperstimulated follicles. Fert Steril. 2003;80:1437–43. 32. Gregory L, Leese HJ. Determinants of oocyte and preimplantation embryo quality: metabolic requirements and the potential role of cumulus cells. Hum Reprod. 1996;11:96–102. 33. Paznekas WA, Boyadjiev SA, Shapiro RE, Daniels O, Wollnik B, Keegan CE, Innis JW, Dinulos MB, Christian C, Hannibal MC, Jabs EW. Connexin43 (GJA1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am J Hum Genet. 2003;72:408–18. 34. Richardson RR, Donnai D, Meire F, Dixon MJ. Expression of Gja1 correlates with the phenotype observed in oculodentodigital syndrome/type III syndactyly. J Med Genet. 2004;41:60–7.

492

G.M. Kidder and E. Winterhager

35. Roscoe W, Veitch GIL, Gong X-Q, Pellegrino E, Bai D, McLachlan E, Shao Q, Kidder GM, Laird DW. Oculodentodigital dysplasia-causing connexin43 mutants are non-functional and exhibit dominant effects on wildtype connexin43. J Biol Chem. 2005;280:11458–66. 36. Flenniken AM, Osborne LR, Anderson N, Ciliberti N, Fleming C, Gittens JEI, Gong X-Q, Lounsbury C, Moreno L, Nieman BJ, Peterson K, Qu D, Roscoe W, Shao Q, Tong D, Veitch GIL, Voronina I,Vukobradovic I, Wood G, Zhu Y, Zirngibl R, Aubin JE, Bai D, Bruneau B, Grynpas M, Henderson J, Henkelman RM, Laird DW, Sled JG, Stanford WL, Kidder GM, Adamson SL, Rossant J. A Gja1 missense mutation in a mouse model for oculodentodigital dysplasia (ODDD). Development. 2005;132:4375–86. 37. Schlafke S, Welsh AO, Enders AC. Penetration of the basal lamina of the uterine luminal epithelium during implantation in the rat. Anat Rec. 1985;212:47–56. 38. Carson DD, Bagchi I, Dey SK, Enders AC, Fazleabas AT, Lessey BA, Yoshinaga K. Embryo implantation. Dev Biol. 2000;223:217–37. 39. Winterhager E, Gru¨mmer R, Jahn E, Willecke K, Traub O. Spatial and temporal expression of connexin26 and connexin43 in rat endometrium during trophoblast invasion. Dev Biol. 1993;157:399–409. 40. Gru¨mmer R, Chwalisz K, Mulholland J, Traub O, Winterhager E. Regulation of connexin26 and connexin43 expression in rat endometrium by ovarian steroid hormones. Biol Reprod. 1994;51:1109–16. 41. Winterhager E, Bru¨mmer F, Dermietzel R, Hu¨lser DF, Denker HW. Gap junction formation in rabbit uterine epithelium in response to embryo recognition. Dev Biol. 1988;126:203–11. 42. Gru¨mmer R, Traub O, Winterhager E. Gap junction connexin genes Cx26 and Cx43 are differentially regulated by ovarian steroid hormones in rat endometrium. Endocrinology. 1999;140:2509–2516. 43. Gru¨mmer R, Hewitt SW, Traub O, Korach KS, Winterhager E. Different regulatory pathways of endometrial connexin expression: preimplantation hormonal-mediated pathway versus embryo implantation-initiated pathway. Biol Reprod. 2004;71:273–81. 44. Reese J, Das SK, Paria BC, Lim H, Song H, Matsumoto H, Knudtson KL, DuBois RN, Dey SK. Global gene expression analysis to identify molecular markers of uterine receptivity and embryo implantation. J Biol Chem. 2001;276:44137–45. 45. Winterhager E, Stutenkemper R, Traub O, Beyer E, Willecke K. Expression of different connexin genes in rat uterus during decidualization and at term. Eur J Cell Biol. 1991;55:133–42. 46. Pauken CM, Lo CW. Non-overlapping expression of Cx43 and Cx26 in the mouse placenta and decidua: a pattern of gap junction gene expression differing from that in the rat. Mol Reprod Dev. 1995;41:195–203. 47. Gabriel HD, Jung D, Bu¨tzler C, Temme A, Traub O, Winterhager E, Willecke K. Transplacental uptake of glucose is decreased in embryonic lethal connexin26-deficient mice. J Cell Biol. 1998;140:1453–61. 48. Joswig A, Gabriel HD, Kibschull M, Winterhager E. Apoptosis in uterine epithelium and decidua in response to implantation: evidence for two different pathways. Reprod Biol Endocrinol. 2003;1:44. 49. Jahn E, Classen-Linke I, Kusche M, Beier HM, Traub O, Gru¨mmer R, Winterhager E. Expression of gap junction connexins in the human endometrium throughout the menstrual cycle. Hum Reprod. 1995;10:2666–70. 50. Granot I, Dekel N, Bechor E, Segal I, Fieldust S, Barash A. Temporal analysis of connexin43 protein and gene expression throughout the menstrual cycle in human endometrium. Fertil Steril. 2000;73:381–6. 51. Garfield RE, Sims S, Daniel EE. Gap junctions: their presence and necessity in myometrium during parturition. Science 1977;198:958–60.

24 Female Reproduction

493

52. Chwalisz K, Fahrenholz F, Hackenberg M, Garfield R, Elger W. The progesterone antagonist onapristone increases the effectiveness of oxytocin to produce delivery without changing the myometrial oxytocin receptor concentrations. Am J Obstet Gynecol. 1991;165:1760–70. 53. Di WL, Lachelin GC, McGarrigle HH, Thomas NS, Becker DL. Oestriol and oestradiol increase cell to cell communication and connexin43 protein expression in human myometrium. Mol Hum Reprod. 2001;7:671–9. 54. MacKenzie LW, Garfield RE. Hormonal control of gap junctions in the myometrium. Am J Physiol. 1985;248:C296–308. 55. Ou CW, Orsino A, Lye SJ. Expression of connexin-43 and connexin-26 in the rat myometrium during pregnancy and labor is differentially regulated by mechanical and hormonal signals. Endocrinology 1997;138:5398–407. 56. Lye SJ, Nicholson BJ, Mascarenhas M, MacKenzie L, Petrocelli T. Increased expression of connexin-43 in the rat myometrium during labor is associated with an increase in the plasma estrogen: progesterone ratio. Endocrinology 1993;132:2380–6. 57. Doring B, Shynlova O, Tsui P, Eckardt D, Janssen-Bienhold U, Hofmann F, Feil S, Feil ¨ R, Lye SJ, Willecke K. Ablation of connexin43 in uterine smooth muscle cells of the mouse causes delayed parturition. J Cell Sci. 2006;119:1715–22. 58. Orsino A, Taylor CV, Lye SJ. Connexin-26 and connexin-43 are differentially expressed and regulated in the rat myometrium throughout late pregnancy and with the onset of labor. Endocrinology 1996;137:1545–53. 59. Albrecht JL, Atal NS, Tadros PN, Orsino A, Lye SJ, Sadovsky Y, Beyer EC. Rat uterine myometrium contains the gap junction protein connexin45, which has differing temporal expression from connexin43. Am J Obstet Gynecol. 1996;175:853–58. 60. Kirchhoff S, Nelles E, Hagendorff A, Kru¨ger O, Traub O, Willecke K. Reduced cardiac conduction velocity and predisposition to arrhythmias in connexin40-deficient mice. Curr Biol. 1998;8:299–302. 61. Simon AM, Goodenough DA, Paul DL. Mice lacking connexin40 have cardiac conduction abnormalities characteristic of atrioventricular block and bundle branch block. Curr Biol. 1998;8:295–8. 62. Kilarski WM, Dupont E, Coppen S, Yeh HI, Vozzi C, Gourdie RG, Rezapour M, Ulmsten U, Roomans GM, Severs NJ. Identification of two further gap-junctional proteins, connexin40 and connexin45, in human myometrial smooth muscle cells at term. Eur J Cell Biol. 1998;75:1–8. 63. Chow L, Lye SJ. Expression of the gap junction protein connexin-43 is increased in the human myometrium toward term and with the onset of labor. Am J Obstet Gynecol. 1994;170:788–795. 64. Cluff AH, Bystrom B, Klimaviciute A, Dahlqvist C, Cebers G, Malmstrom A, EkmanOrdeberg G. Prolonged labor associated with lower expression of syndecan 3 and connexin43 in human uterine tissue. Reprod Biol Endocrinol. 2006;4:4–24.

Chapter 25

Connexins in the Male Reproductive System Georges Pointis, Ce´line Fiorini, Je´rome Gilleron, Diane Carette and Dominique Segretain

Abstract Male fertility is a highly controlled process that requires proliferation, meiosis, and differentiation of male germ cells in the testis, maturation of spermatozoa in the epididymis, and functional male accessory glands. In addition to endocrine and paracrine controls, gap junctions and their constitutive proteins, connexins, are essential for male fertility. The predominant testicular connexin, Cx43, is distributed in the interstitial compartment, between Leydig cells and in the seminiferous tubule compartment. Experiments in which the Cx43 gene is deleted show that in the seminiferous epithelium, Cx43 participates in Sertoli cell functional synchronization, Sertoli to germ cell dialogue, and germ cell proliferation and differentiation. Connexins are also present in other structures of the male genital tract where they participate in coordinating cell-specific activities and in modulating initiation and maintenance of smooth muscle tone. In addition, connexins appear to play important roles in pathology of the testis and prostate. Keywords Male reproductive tract Male infertility Testis Testicular cancer Erectile dysfunction Cx26 Cx30.2 Cx31 Cx31.1 Cx32 Cx33 Cx37 Cx40 Cx43 Cx46 Cx50

25.1 Introduction Testes perform two major functions: an endocrine function characterized by the synthesis of androgens by Leydig cells, and an exocrine function via spermatogenesis. Spermatogenesis, which occurs within seminiferous tubules, is a highly controlled process in which Sertoli cells provide structural and nutritional support for developing germinal cells (reviewed in [1]). Spermatogenesis requires a functional hypothalamo-pituitary system involving luteinizing hormone (LH) and follicle-stimulating hormone (FSH), as well as androgens that G. Pointis (*) INSERM U 670, 28 Avenue de Valombrose, 06107 Nice Cedex 02, France e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_25, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

495

496

G. Pointis et al.

act as paracrine signals. Androgens control not only spermatogenesis but also differentiation of the urogenital sinus and genital tubercle into the prostate, urethra, penis, and scrotum (reviewed in [2]). Gap junctional intercellular communication (GJIC) occurs in the testis and throughout the male genital tract. This chapter focuses on the identification of connexins in the testis, epididymis, seminal vesicle, prostate, and corpus cavernosum, and their roles in the control of spermatogenesis and male genital tract activity. Their potential dysfunctions in testis (spermatogenic arrest, carcinoma-in-situ, seminoma) and prostate (benign hyperplasia, adenocarcinoma) diseases are also discussed.

25.2 The Testis In the mature rat testis, messenger RNA for at least 11 different connexins (Cx26, Cx30.2, Cx31, Cx31.1, Cx32, Cx33, Cx37, Cx40, Cx43, Cx46, and Cx50) has been detected [3,4,5,6,7,8]. All but Cx46 are present in polysomes and presumably translated in adult seminiferous tubules [7]. Figure 25.1 summarizes the current view of connexin mRNA and protein expression in different testicular cell types. Transcripts for Cx26, Cx30.3, Cx31, Cx31.1, Cx32, Cx37, Cx40, Cx43, Cx45, and Cx46 are found in the fetal mouse testis [9]. Cx43 is so far the predominant connexin in the testis and is present in both interstitial tissue and seminiferous epithelium.

25.2.1 The Interstitial Compartment Cx43 is the only connexin detected in Leydig cells of many different species including rat, mouse, guinea pig, mink, frog, and human [10,11,12,13,14,15].

Fig. 25.1 Distribution of connexin RNA and protein in different testicular cell types. Connexin protein distribution is in boldface. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

25 Male Reproduction

497

Cx37 is also found in the interstitial compartment but is specifically located in the endothelial cells of the blood vessels [10].

25.2.2 The Seminiferous Epithelium Within the seminiferous tubules, Cx43 is located in the basal compartment of the seminiferous tubules of rat, guinea pig, mink, and human [10,11,12,13, 16,17]. High-resolution microscopy shows that Cx43 is mainly present at the base of Sertoli cells in close association with adjacent Sertoli cells and with germ cells (Fig. 25.2). This specific distribution of Cx43 within the seminiferous epithelium is supported by in situ hybridization demonstrating that Cx43 mRNAs are present in Sertoli cells and in basally located germ cells such as spermatogonia and spermatocytes in testes of rat [14] and human [17,18]. Dye-coupling experiments performed in situ, using a variety of low molecular weight tracers, reveal multiple routes of GJIC in rat seminiferous tubules that differ in permeability properties and show alternative gating properties [19]. Using a sophisticated in situ assay that allows correlation of dye transfer,

Fig. 25.2 Localization of Cx43 in a cross-section of the adult rat testis. Left panels: Cx43 immunolabeling in the basal compartment (arrows) of the seminiferous tubules and in the interstitial compartment between Leydig cells (arrowheads). Nuclei of Leydig cells, germ and Sertoli cells were identified by 40 , 6-diamidino-2-phenylindole (DAPI) staining (magnification 60). Right panels: High-resolution deconvolution microscopy analysis of the basal compartment of a rat seminiferous tubule (magnification 600). Cx43 staining is located at the base of the Sertoli cell in close contact with germ cells (arrows). Nuclei of a germ cell and of two Sertoli cells (S1 and S2) were identified by DAPI staining. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

498

G. Pointis et al.

cell type identification, and Cx43 localization, it has been demonstrated that Cx43 coupling is much more pronounced from Sertoli cells to germ cells than in the reverse direction, suggesting that Cx43 gap junctions participate in the control exerted by Sertoli cells on germ cells [20]. Cx26 and Cx32 are found only in the apical region of the rat seminiferous tubules, but the specific cell types that express these connexins have not been identified [7]. Cx30.2, the mouse ortholog of human CX31.9, is detected at high levels in vascular smooth muscle cells and at a low level between Sertoli and germ cells [8]. Cx31 mRNA and protein are specifically localized in spermatogonia and early spermatocytes (leptotene and zygotene) and absent in Sertoli cells [21]. Cx50 transcripts appear to be limited to germ cells, pachytene spermatocyte, and round spermatids but the presence of the protein was not investigated [7]. The presence of connexins has been also reported in the testes of other species such as frog [15] and fish [22,23]. In rat and mouse, Cx33 protein is mainly restricted to the testis [5,24], in the basal compartment of seminiferous tubules [24,25]. Its presence has been clearly demonstrated in situ in Sertoli cells [26] and in a Sertoli cell line [25], and Cx33 mRNAs have been detected by reverse-transcriptase polymerase chain reaction (RT-PCR) in germ cells [27]. Cx33, in contrast to other connexins, has been reported to exert a connexin-specific inhibitory effect on GJIC when mRNAs are microinjected into paired Xenopus oocytes [28]. In Sertoli cells, formation of a heterotypic Cx33-Cx43 complex is associated with total inhibition of GJIC between adjacent cells [25]. These findings provide the first evidence of a novel mechanistic model by which a native connexin, exerting a dominant-negative effect, can inhibit GJIC. In human testis, CX43 mRNA and protein are present within the seminiferous tubules and in Leydig cells [16,17]. CX26 has been found within seminiferous tubules [29] and CX31.9 in testicular vascular smooth muscle [8,30]. CX40 transcripts were also detected in the testis [31]. Although no ortholog of the rat-specific testis Cx33 is in the human genome [26,32], it is possible that a different member of the connexin family exerts similar dominant-negative effects in the human testis.

25.2.3 Control of Connexin Expression in the Testis Despite the well-documented presence of several connexins in the testis, little is known about the physiological regulators of GJIC, or when they may be active. Leydig cells respond in vitro to human chorionic gonadotropin exposure by reducing Cx43 mRNA and protein levels and inducing a cellular redistribution of the protein [33]. Activators of protein kinase A (PKA) and protein kinase C (PKC) similarly affect Cx43 expression in the M3 Leydig cell line [34]. In contrast, an inhibitor of protein kinases, staurosporine, prevents and reverses the uncoupling effect of cyclic adenosine monophosphate (cAMP) and 12-Otetradecanoyl-phorbol-13-acetate, suggesting that GJIC between Leydig cells is downregulated by both PKA and PKC, interacting in a complex manner [35].

25 Male Reproduction

499

Cx43 expression in Leydig cells appears to be regulated during fetal and postnatal development in an age-dependent and functional manner [36,37]. The relative temporal correlation among the increase in the number, the size of Leydig cells, and their Cx43 expression during neonatal life is consistent with the possibility that Cx43 participates in the control of developmental processes involved in testosterone production and secretion [37]. Between mature Leydig cells, Cx43 gap junctions may locally coordinate the androgenic activity in relation with the stages of spermatogenesis [10]. In rat and human seminiferous epithelium, Cx43 mRNA and protein expression vary with the stage of spermatogenic cycle [10,11,12,13,14,16,17,24,38]. Although the mechanisms that trigger this stage-dependent expression are unknown, the possibility that a specific germ cell population (perhaps late spermatids) influences Cx43 expression in Sertoli cells has been suggested [13]. That steroids control GJIC between Sertoli cells is shown by the FSHdependent upregulation of GJIC [39]. In vitro, esterified forms of testosterone and of 17 -estradiol markedly downregulate the electrical and diffusional coupling of rat Sertoli cells [40,41]. Thyroid hormones, specifically triiodothyronine (T3), are known to control neonatal Sertoli cell proliferation and differentiation [42]. Treatment of neonatal rats with propylthiouracil, a potent inducer of neonatal hypothyroidism, led to a delocalization of Cx43 from the cell plasma membrane to the cytoplasm of Sertoli cells [43] and increased Sertoli cell growth associated with reduced levels of Cx43 [44]. As expected, the inhibitory effect of T3 on Sertoli cell proliferation was reversed by two inhibitors of gap junction coupling, -glycyrrhetinic acid and oleamide, which suggests that the Cx43-based gap junctions participate in this neonatal process. The possibility that retinoids may also control Sertoli cell Cx43 gene expression is supported by observations in transgenic mice not expressing the retinoid X receptor that show reduced Cx43 expression [14]. This hypothesis is supported by findings indicating that vitamin A treatment restores Cx43 expression in a vitamin A–deficient rat model, which otherwise exhibits impairment of both spermatogenesis and testicular Cx43 gene expression [45]. Sertoli cell phagocytosis of residual bodies could also participate in the control of germ cell proliferation by stimulating interleukin-1 and controlling positively Cx33 and negatively Cx43 gene expression. It has been suggested that the induction of dominant inhibition by Cx33 could overcome the control exerted by Sertoli cells on germ cell proliferation through Cx43-based gap junctions (Fig. 25.3) [46].

25.2.4 Physiological Role of Connexins in the Testis During testicular development, connexins participate in germ cell and Sertoli cell proliferation and differentiation, and ensure metabolic and signal coupling from mature Sertoli cells to germ cells and the synchronization of male germ cell

500

G. Pointis et al.

Fig. 25.3 Schematic representation of the effect of residual body phagocytosis on Sertoli cell Cx33 and Cx43 gene expression. 1, residual bodies (RB), the cytoplasmic portion of the elongated spermatid, are phagocytosed by Sertoli cells; 2, phagocytosis stimulates interleukin-1 production; 3, interleukin-1 enhances Cx33 and inhibits Cx43 gene expressions; 4, Cx33 by associating specifically with Cx43 [47] could block the control exerted by Sertoli cells on germ cell proliferation through Cx43-based gap junctions. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

maturation [19,20]. During fetal testicular development, the hypothesis that Cx43 is required for germ cell migration and targeting to the gonads has been suggested in light of data demonstrating a role of Cx43 in the migration of other cell types [47,48,48]. Whether such specific effects occur by GJIC-dependent or GJIC-independent means is unclear at this time [49,49,51]. During the first weeks of postnatal life, as discussed above, Cx43 is a target gene for T3 and thus probably participates in the control of Sertoli cell proliferation during this developmental period [44]. During adulthood, direct demonstration of Cx43 involvement in spermatogenesis comes from data in transgenic mice in which the Cx43 gene is inactivated (Table 25.1). Although Cx43 knockout (KO) mice die at birth of cardiac malformation [52], testes from Cx43KO fetuses exhibit a 50% depletion in primordial germ cells, indicating that Cx43 participates in the control of germ cell proliferation during fetal development as well [53]. Similarly, testes from Cx43KO fetuses do not allow normal proliferation and differentiation of germ cells when grafted under the kidney capsules of adult males [54]. Similar results have been reported in Drosophila males carrying a null mutation in the zero population growth (zpg) locus that encodes the germline-specific gap junction protein, innexin 4 [55].

25 Male Reproduction

501

Table 25.1 Analysis of male fertility in connexin knockout and connexin knock-in mice. Seminiferous tubules showing Sertoli-cell–only syndrome are totally devoid of germ cells. In mice Cx43KI32 and Cx43KI40 the coding sequence of the Cx43 gene is substituted by the coding sequences of Cx32 or Cx40, respectively. The Sertoli cell-specific Cx43 knockout (SCCx43KO) mouse was generated using Cre-lox technology Knockout (KO)/knock-in Connexin Location (KI) Effect on spermatogenesis Cx26

Seminiferous tubule

KO

Cx31 Cx32 Cx33 Cx37

Seminiferous tubule Seminiferous tubule Seminiferous tubule Blood vessels

KO KO KO KO

Cx40

Seminiferous tubule

Cx43

Seminiferous tubule, Leydig cells, blood vessels

KO Cx37/Cx40 double KO KO

KO Cx43KI32 Cx43KI40 SC-Cx43KO

Cx46 Cx47 Cx50 Innexin 4

Seminiferous tubule Not identified Seminiferous tubule Spermatogonia, early spermatocytes, somatic cyst cells

KO KO KO KO

Unknown since embryonic lethality [59] Normal spermatogenesis [60] Normal spermatogenesis [61] Unknown (embryonic lethality?) Normal spermatogenesis but female are sterile [62] Normal spermatogenesis [63–65] Impaired testicular function due to testicular vascular anomalies [66] Germ cell deficiency in fetal testis [54]

Germ cell deficiency in host natally grafted testis [54] Alteration of spermatogenesis (Sertoli-cell–only syndrome) [56] Aberrant Sertoli cell proliferation and delayed maturation in adult (spermatogonia only) [57] Normal spermatogenesis [58] Normal spermatogenesis [59] Normal spermatogenesis [60] Reduced number of spermatogonia in Drosophila [55]

Transgenic mice, in which the coding sequence of the Cx43 gene is substituted (knock-in, KI) by the coding sequences of Cx32 or Cx40, are viable. Males of Cx43KI32 and Cx43KI40 lineages are sterile with seminiferous tubules depleted in germ cells exhibiting a Sertoli cell-only (SCO) phenotype [56]. Recent data, using SCO-Cx43KO mice, suggest that Cx43 is essential for spermatogenesis but not for spermatogonial maintenance/proliferation [57]. The detailed mechanism(s) that produce sterility in these animal models are unknown. The possibility that altered connexin expression induces apoptotic signaling in the testis via the caspase3 and the nuclear factor B (NF-B) pathway has been suggested [58]. In contrast to Cx43KO mice, animals homozygous for targeted genetic deletion of Cx26, Cx31, Cx32, Cx40, Cx46, Cx47, or Cx50 show normal fertility [58,59,60,61,62,63,64,65,66].

502

G. Pointis et al.

25.2.5 Pathophysiological Aspects Previous morphological studies reported severe alteration of testicular gap junctions in human pathological testes with impaired spermatogenesis [67,68,69,70,71]. In azoospermic patients with SCO syndrome, CX43 mRNA and protein are undetectable within seminiferous tubules [16] but are maintained in Leydig cells [17]. Thus, it has been suggested that this altered distribution of CX43, which is probably related to a defect in the maturation of Sertoli cells, could be used as a biological marker of Sertoli cell dysfunction [17]. Impaired GJIC is associated with the effects of carcinogens and oncogenes [72] and is a typical feature of numerous neoplastic tissues (reviewed in [73]). CX43 is not found in human testis infiltrated with carcinoma-in-situ (CIS) or in seminoma [29]. Other studies report, however, that CX43 is normal in pure testicular seminoma and in a seminoma cell line but that the protein is aberrantly localized within the cytoplasm of tumor cells [18,74]. Analysis of the kinetics of Cx43 distribution during early and advanced stages of tumorigenesis in transgenic mice demonstrated that aberrant cytoplasmic Cx43 localization precedes the reduction in Cx43 protein levels, allowing the authors to propose Cx43 delocalization as an early event of uncontrolled cell proliferation in the pathological testis [75]. In contrast to CX43, CX26, which is not expressed by normal human testis, is upregulated and displays a strong cytoplasmic Sertoli cell staining in infiltrated tubules with spermatogonial arrest or CIS only [29]. This has an interesting parallel with previous data from human breast cancer indicating that CX26 expression is strengthened in proportion to the grade of malignancy [76]. Other studies showed that CX40 mRNA levels increase in testicular seminoma as compared to normal human testis [31]. Although the reasons for the alterations of CX26 and CX40 expression in tumoral testes are presently unknown, it has been suggested that these two connexins should be considered as potential new diagnostic markers for testicular germ cell tumors.

25.3 The Male Reproductive Tract 25.3.1 Epididymis and Seminal Vesicles The formation of mature germ cells not only involves their proliferation, meiosis, and differentiation during spermatogenesis within the seminiferous tubules, but also their maturation during transit in the initial segment, the caput, the corpus, and the cauda of the epididymis, where they acquire both fertilizing capacity and progressive motility. The presence of multiple connexin mRNAs (for Cx26, Cx30.3, Cx31.1, and Cx32), whose expression is segmentspecific and age-dependent, has been reported in the rat epididymis [77]. The developmental patterns of connexin mRNA levels suggest that some connexins

25 Male Reproduction

503

(Cx26 and Cx43) play roles in the differentiation of the epididymal epithelium and that there is a switch from Cx26 to Cx31.1 that is crucial for epididymal function [78]. Differential regulation of mRNA or protein levels occurs along the epididymis. Cx43 expression appears androgen-dependent in the initial segment of the epididymis [79] and thyroid hormone–dependent in the proximal regions of the epididymis, but not in the cauda epididymis [43]. The reasons for such control of connexin expression in different regions of the epididymis are presently unknown. Whether spermatozoa-epididymis GJIC is required for sperm to become mature is presently not established. Seminal vesicles secrete a large variety of products essential for male fertility. Cx32-based gap junctions in epithelial cells of seminal vesicles can probably participate in coordination of function and communication not only between epithelial cells but also with nearby vascular and mesenchymal cells [80].

25.3.2 Prostate The prostate is an exocrine gland predominantly formed by two cell types: epithelial cells, which line the ducts and acini, and mesenchymal cells, which form the stroma. Gap junctions, which are present between secretory luminal and basal cells of the prostate, can coordinate the activities of the two cell types [80]. In normal prostate, Cx43 is expressed in undifferentiated and mature basal cells as well as urogenital sinus mesenchymal cells, whereas Cx32 is found in luminal cells. Expression of Cx32 in the acinar epithelial cells is concomitant with the acquisition of the differentiated state [81,82]. Cx43 gap junctions are also present in the interstitial cells of the prostate, which function as pacemaker cells for the smooth muscle tone of the prostate [83]. Castration of male rats is associated with both a marked increase in Cx43 expression and an induction of apoptosis of prostate gland cells, whereas androgen treatment abolishes the castration-induced Cx43 expression and prevents apoptosis [84]. The findings that Cx32 is not affected in the same way by androgens suggest that Cx43 may negatively affect specific aspects of prostate epithelial cell survival and proliferation. Estrogens also influence connexin expression in prostate tissue. Both in vivo neonatal estrogenization and in vitro estrone exposure of nontumorigenic human prostate epithelial cell line increase Cx43 and decrease Cx32 expression and may result in defective cell–cell communication [81,85]. This altered Cx43/Cx32 ratio probably reflects a switch from differentiation to dedifferentiation of prostate epithelial cells. Whether these effects are classically mediated through an estrogen response element or activator protein-1 (AP-1) transcription factor site, or imply activation of cAMP-dependent signal transduction pathways, remains to be elucidated.

504

G. Pointis et al.

25.3.3 Prostatic Diseases and Connexins Like the majority of neoplastic cells, prostate carcinoma cells have reduced expression of CX43 and CX32 [82,86,87,88]. In vitro and in vivo studies have shown that the expression, trafficking, and assembly of CX43 and CX32, the two connexins expressed by well-differentiated epithelial cells of the normal prostate, are impaired during prostate cancer progression [88,89]. This may lead to aberrant epithelial growth, dedifferentiation, and tumor progression. Studies using nontumorigenic and tumorigenic human prostate epithelial cell lines revealed that an increase in the CX43/CX32 ratio may restore GJIC in junctionally deficient cells, providing a basis for the development of new strategies for the prevention and treatment of human prostate cancer [85,90,91]. Similarly, forced expression of CX43 or CX26 in vitro or in vivo leads to inhibition of tumoral cell growth, which may be reinforced by combining gene therapy with chemotherapy [92,93].

25.3.4 Corpus Cavernosum The major connexin in the corpus cavernosum smooth muscle is Cx43, which has been characterized in situ and in cultured cells. Relaxation of the smooth muscle in the helicine arteries and the trabeculae allows the accumulation of blood in the cavernous sinusoids that produces penile erection. Cx43-based gap junctions establish a syncytial cellular network allowing spread of electrical current and of signaling molecules throughout the corpus cavernosum, ensuring coordination of smooth muscle relaxation that leads to penile erection [94,95]. Connexin dysfunction is associated in some cases with impaired corpus cavernosum function. Reduced electrical signaling due to CX43 gap junction impairment could be responsible for erectile dysfunction in some cases (reviewed in [96]). This hypothesis is supported by observations of a negative correlation between CX43 levels measured in corporal tissue strips and patient age [97,98]. There is also intriguing evidence that in diabetic rats impaired penile erection is accompanied by enhanced molecular, but not electrical, communication through Cx43 gap junctions [99], which may increase the intercellular spread of 1-adrenergic initiated contractile signals. This situation may contribute to the increased contractility and decreased relaxation of corporeal myocytes thought to be characteristic of erectile dysfunction in diabetic men.

25.4 Conclusion Gap junctions are widespread in the male reproductive tract and play essential roles in several functions including germ cell proliferation and differentiation, and control of smooth muscle tone in different structures. CX43 appears as to

25 Male Reproduction

505

be a good marker of Sertoli cell function since alteration of its expression and aberrant localization within the cytoplasm is associated with human testicular diseases such as infertility and cancer. Connexin gene mutations lead to a large variety of human diseases [100], yet no alteration of male fertility due to spermatogenic arrest or to seminoma have been reported, suggesting that other mechanisms may be responsible for testicular pathologies. Epigenetic inactivation of connexin genes, the deleterious influence of testicular toxicants, altered trafficking of the protein, and the interaction of connexins with other protein partners are important areas for future investigation [101]. One suspects impaired connexin expression in prostatic adenocarcinoma and in the physiopathology of the human erectile response. The development of strategies to restore normal connexin expression in affected cells of the male reproductive system will be required, once the role of specific connexins in the processes is precisely determined. Effective pharmacological agents and strategies for connexin gene transfer to these cells with high efficiency and without deleterious side effects may be possible. Such strategies have been already initiated in different domains of male reproductive pathology such as prostate disease [85] and erectile dysfunction [99]. Acknowledgments Preparation of this chapter and our studies reported herein were supported by grants from the Institut National de la Sante´ et de la Recherche Me´dicale (INSERM), the French Ministry of Research and Technology, and Pfizer, and by European Union grant 81 672 E.

References 1. Griswold M. Protein secretion by Sertoli cells: General consideration. In: Russell LD, Griswold MD, editors. Clearwater: Cache River Press; 1993. pp. 195–200. 2. Hughes IA, Lim HN, Martin H, Mongan NP, Dovey L, Ahmed SF, Hawkins JR. Developmental aspects of androgen action. Mol Cell Endocrinol.2001;185:33–41. 3. Zhang JT, Nicholson BJ. Sequence and tissue distribution of a second protein of hepatic gap junctions, Cx26, as deduced from its cDNA. J Cell Biol. 1989;109:3391–401. 4. Kadle R, Zhang JT, Nicholson BJ. Tissue-specific distribution of differentially phosphorylated forms of Cx43. Mol Cell Biol. 1991;11:363–9. 5. Haefliger JA, Bruzzone R, Jenkins NA, Gilbert DJ, Copeland NG, Paul DL. Four novel members of the connexin family of gap junction proteins. Molecular cloning, expression, and chromosome mapping. J Biol Chem. 1992;267:2057–64. 6. Hennemann H, Dahl E, White JB, Schwarz HJ, Lalley PA, Chang S, Nicholson BJ, Willecke K. Two gap junction genes, connexin 31.1 and 30.3, are closely linked on mouse chromosome 4 and preferentially expressed in skin. J Biol Chem. 1992;267:17225–33. 7. Risley MS. Connexin gene expression in seminiferous tubules of the Sprague-Dawley rat. Biol Reprod. 2000;62:748–54. 8. Nielsen PA, Kumar NM. Differences in expression patterns between mouse connexin-30.2 (Cx30.2) and its putative human orthologue, connexin-31.9. FEBS Lett. 2003;540:151–6. 9. Juneja SC. mRNA expression pattern of multiple members of connexin gene family in normal and abnormal fetal gonads in mouse. Indian J Physiol Pharmacol. 2003;47:147–56.

506

G. Pointis et al.

10. Risley MS, Tan IP, Roy C, Sa´ez JC. Cell-, age- and stage-dependent distribution of connexin43 gap junctions in testes. J Cell Sci. 1992;103:81–96. 11. Varanda WA, de Carvalho AC. Intercellular communication between mouse Leydig cells. Am J Physiol. 1994;267:C563–9. 12. Pelletier RM. The distribution of connexin 43 is associated with the germ cell differentiation and with the modulation of the Sertoli cell junctional barrier in continual (guinea pig) and seasonal breeders’ (mink) testes. J Androl. 1995;16:400–9. 13. Batias C, Defamie N, Lablack A, Thepot D, Fenichel P, Segretain D, Pointis G. Modified expression of testicular gap-junction connexin 43 during normal spermatogenic cycle and in altered spermatogenesis. Cell Tissue Res. 1999;298:113–21. 14. Batias C, Siffroi JP, Fenichel P, Pointis G, Segretain D. Connexin43 gene expression and regulation in the rodent seminiferous epithelium. J Histochem Cytochem. 2000;48:793–805. 15. Palmiero C, Ferrara D, De Rienzo G, d’Istria M, Minucci S. Ethane 1,2-dimethane sulphonate is a useful tool for studying cell-to-cell interactions in the testis of the frog, Rana esculenta. Gen Comp Endocrinol. 2003;131:38–47. 16. Steger K, Tetens F, Bergmann M. Expression of connexin 43 in human testis. Histochem Cell Biol. 1999;112:215–20. 17. Defamie N, Berthaut I, Mograbi B, Chevallier D, Dadoune JP, Fenichel P, Segretain D, Pointis G. Impaired gap junction connexin43 in Sertoli cells of patients with secretory azoospermia: a marker of undifferentiated Sertoli cells. Lab Invest. 2003;83:449–56. 18. Segretain D, Chevallier D, Fenichel P, Pointis G. Gap junction connexin 43 expression in human normal germ cells and testicular seminoma cells. Trends Reprod Biol. 2005;1:49–55. 19. Risley MS, Tan IP, Farrell J. Gap junctions with varied permeability properties establish cell-type specific communication pathways in the rat seminiferous epithelium. Biol Reprod. 2002;67:945–52. 20. Decrouy X, Gasc JM, Pointis G, Segretain D. Functional characterization of Cx43 based gap junctions during spermatogenesis. J Cell Physiol. 2004;200:146–54. 21. Mok BW, Yeung WS, Luk JM. Differential expression of gap-junction gene connexin 31 in seminiferous epithelium of rat testes. FEBS Lett. 1999;453:243–48. 22. Cheng S, Shakespeare T, Mui R, White TW, Valdimarsson G. Connexin 48.5 is required for normal cardiovascular function and lens development in zebrafish embryos. J Biol Chem. 2004;279:36993–37003. 23. Batlouni SR, Carreno FR, Romagosa E, Borella MI. Cell junctions in the germinal epithelium may play an important role in spermatogenesis of the catfish P. fasciatum (Pisces, Siluriformes). J Mol Histol. 2005;36:97–110. 24. Tan IP, Roy C, Sa´ez JC, Sa´ez CG, Paul DL, Risley MS. Regulated assembly of connexin33 and connexin43 into rat Sertoli cell gap junctions. Biol Reprod. 1996;54:1300–10. 25. Fiorini C, Mograbi B, Cronier L, Bourget I, Decrouy X, Nebout M, Ferrua B, Malassine A, Samson M, Fenichel P, Segretain D, Pointis G. Dominant-negative effect of connexin33 on gap junctional communication is mediated by connexin43 sequestration. J Cell Sci. 2004;117:4665–72. 26. Fischer P, Brehm R, Konrad L, Hartmann S, Kliesch S, Bohle RM, Bergmann M. Connexin 33: a rodent-specific member of the gap junction protein family? J Androl. 2005;26:75–84. 27. Chung SS, Lee WM, Cheng CY. Study on the formation of specialized inter-Sertoli cell junctions in vitro. J Cell Physiol. 1999;181:258–72. 28. Chang M, Werner R, Dahl G. A role for an inhibitory connexin in testis? Dev Biol. 1996;175:50–6. 29. Brehm R, Marks A, Rey R, Kliesch S, Bergmann M, Steger K. Altered expression of connexins 26 and 43 in Sertoli cells in seminiferous tubules infiltrated with carcinoma-insitu or seminoma. J Pathol. 2002;197:647–53.

25 Male Reproduction

507

30. White TW, Srinivas M, Ripps H, Trovato-Salinaro A, Condorelli DF, Bruzzone R. Virtual cloning, functional expression, and gating analysis of human connexin31.9. Am J Physiol Cell Physiol. 2002;283:C960–70. 31. Okada K, Katagiri T, Tsunoda T, Mizutani Y, Suzuki Y, Kamada M, Fujioka T, Shuin T, Miki T, Nakamura Y. Analysis of gene-expression profiles in testicular seminomas using a genome-wide cDNA microarray. Int J Oncol. 2003;23:1615–35. 32. Willecke K, Eiberger J, Degen J, Eckardt D, Romualdi A, Guldenagel M, Deutsch U, So¨hl G. Structural and functional diversity of connexin genes in the mouse and human genome. Biol Chem. 2002;383:725–37. 33. You S, Li W, Lin T. Expression and regulation of connexin43 in rat Leydig cells. J Endocrinol. 2000;166:447–53. 34. Goldenberg RC, Fortes FS, Cristancho JM, Morales MM, Franci CR, Varanda WA, Campos de Carvalho AC. Modulation of gap junction mediated intercellular communication in M3 Leydig cells. J Endocrinol. 2003;177:327–35. 35. Cristancho JM, Campos de Carvalho AC, Varanda WA. Short term regulation of cell-cell communication in M3 Leydig cells. A perforated patch study. Biochim Biophys Acta. 2000;1496:325–32. 36. Bravo-Moreno JF, Diaz-Sanchez V, Montoya-Flores JG, Lamoyi E, Sa´ez JC, PerezArmendariz EM. Expression of connexin43 in mouse Leydig, Sertoli, and germinal cells at different stages of postnatal development. Anat Rec. 2001;264:13–24. 37. Perez-Armendariz EM, Lamoyi E, Mason JI, Cisneros-Armas D, Luu-The V, Bravo Moreno JF. Developmental regulation of connexin 43 expression in fetal mouse testicular cells. Anat Rec. 2001;264:237–46. 38. Lablack A, Bourdon V, Defamie N, Batias C, Mesnil M, Fenichel P, Pointis G, Segretain D. Ultrastructural and biochemical evidence for gap junction and connexin 43 expression in a clonal Sertoli cell line: a potential model in the study of junctional complex formation. Cell Tissue Res. 1998; 294:279–87. 39. Pluciennik F, Joffre M, Deleze J. Follicle-stimulating hormone increases gap junction communication in Sertoli cells from immature rat testis in primary culture. J Membr Biol. 1994;139:81–96. 40. Herve JC, Pluciennik F, Bastide B, Cronier L, Verrecchia F, Malassine A, Joffre M, Deleze J. Contraceptive gossypol blocks cell-to-cell communication in human and rat cells. Eur J Pharmacol. 1996;313:243–55. 41. Pluciennik F, Verrecchia F, Bastide B, Herve JC, Joffre M, Deleze J. Reversible interruption of gap junctional communication by testosterone propionate in cultured Sertoli cells and cardiac myocytes. J Membr Biol. 1996;149:169–77. 42. Hess RA, Cooke PS, Bunick D, Kirby JD. Adult testicular enlargement induced by neonatal hypothyroidism is accompanied by increased Sertoli and germ cell numbers. Endocrinology 1993;132:2607–13. 43. St-Pierre N, Dufresne J, Rooney AA, Cyr DG. Neonatal hypothyroidism alters the localization of gap junctional protein connexin 43 in the testis and messenger RNA levels in the epididymis of the rat. Biol Reprod. 2003;68:1232–40. 44. Gilleron J, Nebout M, Scarabelli L, Senegas-Balas F, Palmero S, Segretain D, Pointis G. A potential novel mechanism involving connexin 43 gap junction for control of Sertoli cell proliferation by thyroid hormones. J Cell Physiol. 2006;209:153–61. 45. Luk JM, Mok BW, Shum CK, Yeung WS, Tam PC, Tse JY, Chow JF, Woo J, Kam K, Lee KF. Identification of novel genes expressed during spermatogenesis in stagesynchronized rat testes by differential display. Biochem Biophys Res Commun. 2003;307:782–90. 46. Fiorini C, Decrouy X, Defamie N, Segretain D, Pointis G. Opposite regulation of connexin33 and connexin43 by LPS and IL-1 in spermatogenesis. Am J Physiol Cell Physiol. 2006;290:C733–40.

508

G. Pointis et al.

47. Xu X, Li WE, Huang GY, Meyer R, Chen T, Luo Y, Thomas MP, Radice GL, Lo CW. Modulation of mouse neural crest cell motility by N-cadherin and connexin 43 gap junctions. J Cell Biol. 2001;154:217–30. 48. Wei CJ, Xu X, Lo CW. Connexins and cell signaling in development and disease. Annu Rev Cell Dev Biol. 2004;20:811–38. 49. Plotkin LI, Manolagas SC, Bellido T. Transduction of cell survival signals by connexin-43 hemichannels. J Biol Chem. 2002;277:8648–57. 50. Qin H, Shao Q, Curtis H, Galipeau J, Belliveau DJ, Wang T, Alaoui-Jamali MA, Laird DW. Retroviral delivery of connexin genes to human breast tumor cells inhibits in vivo tumor growth by a mechanism that is independent of significant gap junctional intercellular communication. J Biol Chem. 2002; 277:29132–8. 51. Qin H, Shao Q, Thomas T, Kalra J, Alaoui-Jamali MA, Laird DW. Connexin26 regulates the expression of angiogenesis-related genes in human breast tumor cells by both GJICdependent and -independent mechanisms. Cell Commun Adhes. 2003;10:387–93. 52. Reaume AG, de Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC, Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice lacking connexin43. Science. 1995;267:1831–4. 53. Juneja SC, Barr KJ, Enders GC, Kidder GM. Defects in the germ line and gonads of mice lacking connexin43. Biol Reprod. 1999;60:1263–70. 54. Roscoe WA, Barr KJ, Mhawi AA, Pomerantz DK, Kidder GM. Failure of spermatogenesis in mice lacking connexin43. Biol Reprod. 2001;65:829–38. 55. Tazuke SI, Schulz C, Gilboa L, Fogarty M, Mahowald AP, Guichet A, Ephrussi A, Wood CG, Lehmann R, Fuller MT. A germline-specific gap junction protein required for survival of differentiating early germ cells. Development 2002;129:2529–39. 56. Plum A, Hallas G, Magin T, Dombrowski F, Hagendorff A, Schumacher B, Wolpert C, Kim J, Lamers WH, Evert M, Meda P, Traub O, Willecke K. Unique and shared functions of different connexins in mice. Curr Biol. 2000;10:1083–91. 57. Sridharan S, Simon L, Meling DD, Cyr DG, Gutstein DE, Fishman GI, Guillou F, Cooke PS. Proliferation of adult Sertoli cells following conditional knockout of the gap junctional protein GJA1 (Connexin 43). Biol Reprod. 2007;76:804–12. 58. Lee NP, Leung KW, Wo JY, Tam PC, Yeung WS, Luk JM. Blockage of testicular connexins induced apoptosis in rat seminiferous epithelium. Apoptosis. 2006;11:1215–29. 59. Nelles E, Butzler C, Jung D, Temme A, Gabriel HD, Dahl U, Traub O, Stumpel F, Jungermann K, Zielasek J, Toyka KV, Dermietzel R, Willecke K. Defective propagation of signals generated by sympathetic nerve stimulation in the liver of connexin32-deficient mice. Proc Natl Acad Sci USA. 1996;93:9565–70. 60. Simon AM, Goodenough DA, Li E, Paul DL. Female infertility in mice lacking connexin 37. Nature. 1997;385:525–9. 61. Gabriel HD, Jung D, Butzler C, Temme A, Traub O, Winterhager E, Willecke K. Transplacental uptake of glucose is decreased in embryonic lethal connexin26-deficient mice. J Cell Biol. 1998;140:1453–61. 62. Kirchhoff S, Nelles E, Hagendorff A, Kruger O, Traub O, Willecke K. Reduced cardiac conduction velocity and predisposition to arrhythmias in connexin40-deficient mice. Curr Biol. 1998;8:299–302. 63. Simon AM, Goodenough DA, Paul DL. Mice lacking connexin40 have cardiac conduction abnormalities characteristic of atrioventricular block and bundle branch block. Curr Biol. 1998;8:295–8. 64. Plum A, Winterhager E, Pesch J, Lautermann J, Hallas G, Rosentreter B, Traub O, Herberhold C, Willecke K. Connexin31-deficiency in mice causes transient placental dysmorphogenesis but does not impair hearing and skin differentiation. Dev Biol. 2001;231:334–7. 65. van Rijen HV, van Veen TA, van Kempen MJ, Wilms-Schopman FJ, Potse M, Krueger O, Willecke K, Opthof T, Jongsma HJ, de Bakker JM. Impaired conduction in the bundle

25 Male Reproduction

66. 67. 68. 69. 70. 71. 72. 73. 74.

75.

76.

77.

78. 79. 80.

81.

82. 83. 84.

85.

86.

509

branches of mouse hearts lacking the gap junction protein connexin40. Circulation. 2001;103:1591–98. Simon AM, McWhorter AR. Vascular abnormalities in mice lacking the endothelial gap junction proteins connexin37 and connexin40. Dev Biol. 2002;251:206–20. Nagano T, Suzuki F. Freeze-fracture observations on the intercellular junctions of Sertoli cells and of Leydig cells in the human testis. Cell Tissue Res. 1976;166:37–48. Bigliardi E, Vegni-Talluri M. Gap junctions between Sertoli cells in the infertile human testis. Fertil Steril. 1977;28:755–8. Schleiermacher E. Ultrastructural changes of the intercellular relationship in impaired human spermatogenesis. Hum Genet. 1980;54:391–404. Camatini M, Franchi E, de Curtis I, Anelli G, Masera G. Chemotherapy does not affect the development of inter-Sertoli junctions in childhood leukaemia. Anat Rec. 1982;203:353–63. Cavicchia JC, Sacerdote FL, Ortiz L. The human blood-testis barrier in impaired spermatogenesis. Ultrastruct Pathol. 1996;20:211–18. Trosko JE. The role of stem cells and gap junctional intercellular communication in carcinogenesis. J Biochem Mol Biol. 2003;36:43–8. Mesnil M. Connexins and cancer. Biol Cell. 2002;94:493–500. Roger C, Mograbi B, Chevallier D, Michiels JF, Tanaka H, Segretain D, Pointis G, Fenichel P. Disrupted traffic of connexin 43 in human testicular seminoma cells: overexpression of Cx43 induces membrane location and cell proliferation decrease. J Pathol. 2004;202:241–6. Segretain D, Decrouy X, Dompierre J, Escalier D, Rahman N, Fiorini C, Mograbi B, Siffroi JP, Huhtaniemi I, Fenichel P, Pointis G. Sequestration of connexin43 in the early endosomes: an early event of Leydig cell tumor progression. Mol Carcinog. 2003;38:179–87. Jamieson S, Going JJ, D’Arcy R, George WD. Expression of gap junction proteins connexin 26 and connexin 43 in normal human breast and in breast tumors. J Pathol. 1998;184:37–43. Dufresne J, Finnson KW, Gregory M, Cyr DG. Expression of multiple connexins in the rat epididymis indicates a complex regulation of gap junctional communication. Am J Physiol Cell Physiol. 2003;284:C33–43. Cyr DG, Finnson KW, Dufresne J, Gregory M. Cellular interaction and the bloodepididymal barrier. New York: Plenum; 2002.pp. 103–118. Cyr DG, Hermo L, Laird DW. Immunocytochemical localization and regulation of connexin43 in the adult rat epididymis. Endocrinology 1996;137:1474–84. Meda P, Pepper MS, Traub O, Willecke K, Gros D, Beyer E, Nicholson B, Paul D, Orci L. Differential expression of gap junction connexins in endocrine and exocrine glands. Endocrinology. 1993;133:2371–8. Habermann H, Chang WY, Birch L, Mehta P, Prins GS. Developmental exposure to estrogens alters epithelial cell adhesion and gap junction proteins in the adult rat prostate. Endocrinology. 2001;142:359–69. Habermann H, Ray V, Habermann W, Prins GS. Alterations in gap junction protein expression in human benign prostatic hyperplasia and prostate cancer. J Urol. 2002;167:655–60. Van der Aa F, Roskams T, Blyweert W, De Ridder D. Interstitial cells in the human prostate: a new therapeutic target? Prostate 2003;56:250–5. Huynh HT, Alpert L, Laird DW, Batist G, Chalifour L, Alaoui-Jamali MA. Regulation of the gap junction connexin 43 gene by androgens in the prostate. J Mol Endocrinol. 2001;26:1–10. Carruba G, Webber MM, Quader ST, Amoroso M, Cocciadiferro L, Saladino F, Trosko JE, Castagnetta LA. Regulation of cell-to-cell communication in nontumorigenic and malignant human prostate epithelial cells. Prostate. 2002; 50:73–82. Tsai H, Werber J, Davia MO, Edelman M, Tanaka KE, Melman A, Christ GJ, Geliebter J. Reduced connexin 43 expression in high grade, human prostatic adenocarcinoma cells. Biochem Biophys Res Commun. 1996;227:64–69.

510

G. Pointis et al.

87. Hossain MZ, Jagdale AB, Ao P, LeCiel C, Huang RP, Boynton AL. Impaired expression and posttranslational processing of connexin43 and downregulation of gap junctional communication in neoplastic human prostate cells. Prostate 1999;38:55–59. 88. Mehta PP, Perez-Stable C, Nadji M, Mian M, Asotra K, Roos BA. Suppression of human prostate cancer cell growth by forced expression of connexin genes. Dev Genet. 1999;24:91–110. 89. Govindarajan R, Zhao S, Song XH, Guo RJ, Wheelock M, Johnson KR, Mehta PP. Impaired trafficking of connexins in androgen-independent human prostate cancer cell lines and its mitigation by -catenin. J Biol Chem. 2002;277:50087–97. 90. Carruba G, Stefano R, Cocciadiferro L, Saladino F, Di Cristina A, Tokar E, Quader ST, Webber MM, Castagnetta L. Intercellular communication and human prostate carcinogenesis. Ann NY Acad Sci. 2002;963:156–68. 91. Saladino F, Carruba G, Quader ST, Amoroso M, Di Cristina A, Webber MM, Castagnetta LA. Connexin expression in nonneoplastic human prostate epithelial cells. Ann NY Acad Sci. 2002;963:213–7. 92. Tanaka M, Grossman HB. Connexin 26 induces growth suppression, apoptosis and increased efficacy of doxorubicin in prostate cancer cells. Oncol Rep. 2004;11:537–41. 93. Fukushima M, Hattori Y, Yoshizawa T, Maitani Y. Combination of non-viral connexin 43 gene therapy and docetaxel inhibits the growth of human prostate cancer in mice. Int J Oncol. 2007;30:225–31. 94. Christ GJ, Moreno AP, Parker ME, Gondre CM, Valcic M, Melman A, Spray DC. Intercellular communication through gap junctions: a potential role in pharmacomechanical coupling and syncytial tissue contraction in vascular smooth muscle isolated from the human corpus cavernosum. Life Sci. 1991;49:PL195–L200. 95. Campos de Carvalho AC, Roy C, Moreno AP, Melman A, Hertzberg EL, Christ GJ, Spray DC. Gap junctions formed of connexin43 are found between smooth muscle cells of human corpus cavernosum. J Urol. 1993;149:1568–75. 96. Pointis G. Connexin43: emerging role in erectile function. Int J Biochem Cell Biol. 2006;38:1642–6. 97. Serels S, Day NS, Wen YP, Giraldi A, Lee SW, Melman A, Christ GJ. Molecular studies of human connexin 43 (Cx43) expression in isolated corporal tissue strips and cultured corporal smooth muscle cells. Int J Impot Res. 1998;10:135–143. 98. Karicheti V, Christ GJ. Physiological roles for K+ channels and gap junctions in urogenital smooth muscle: implications for improved understanding of urogenital function, disease and therapy. Curr Drug Targets. 2001;2:1–20. 99. Brink PR, Valiunas V, Wang HZ, Zhao W, Davies K, Christ GJ. Experimental diabetes alters connexin43 derived gap junction permeability in short-term cultures of rat corporeal vascular smooth muscle cells. J Urol. 2006;175:381–6. 100. Gerido DA, White TW. Connexin disorders of the ear, skin, and lens. Biochim Biophys Acta. 2004;1662:159–70. 101. Pointis G, Segretain D. Role of connexin-based gap junction channels in testis. Trends Endocrinol Metab. 2005;16:300–6.

Chapter 26

Connexins and Secretion Sabine Bavamian, Philippe Klee, Florent Allagnat, Jacques-Antoine Haefliger and Paolo Meda

Abstract Four decades ago, a serendipitous finding revealed that a membraneimpermeable tracer could be rapidly exchanged between salivary gland cells. Since this first demonstration of direct cell–cell communication in a secretory system, gap junctions, connexins, and coupling have been shown to be obligatory attributes of all multicellular glands. The distribution of various connexin isoforms has now been mapped in most of these organs and shown to differ in endocrine and exocrine systems, because of a specific transcriptional control of the connexin genes. Many studies have provided further circumstantial evidence of a role for connexin-dependent signaling in the secretory function of a variety of glands. This function has been directly demonstrated in vivo in only a few systems, in which the mechanism linking connexin signaling to secretion has started to be elucidated. However, this mechanism, and implications for the physiological and pathophysiological function of secretory cells remains to be unraveled. Keywords Exocrine Endocrine Glands Enzymes Hormones Calcium signaling Coupling Secretion Cx26 Cx30 Cx32 Cx36 Cx40 Cx43

26.1 Introduction Secretion is an ubiquitous process whereby highly differentiated cells synthesize, store, and release signaling, structural, nutritional, enzymatic, or ionic products of various chemical composition, to either the circulation (endocrine secretion) or the external milieu (exocrine secretion). Secretion is accomplished by individual epithelial, connective, muscle, or nerve cells, as well as by glands, that is, larger assemblies of specialized epithelial cells. The latter mechanism, which developed soon after the appearance of multicellular organisms [1], was P. Meda (*) Department of Cell Physiology and Metabolism, University of Geneva, C.M.U., 1 rue Michel Servet, 1211 Geneva 4, Switzerland e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_26, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

511

512

S. Bavamian et al.

presumably selected and maintained throughout evolution by the necessity of vital secretions required in amounts exceeding the capacities of single cells. In turn, the assembly of many secretory cells, often of different types within the same gland, required the development of mechanisms to coordinate their functioning, together with that of gland vascular and mesenchymal cells, to ensure the proper production and release of the secretory products. Indeed, both the lack and the excess of these products result in deleterious effects. With evolution, these mechanisms were integrated into a complex network of pathways for indirect (e.g., mediated by neurotransmitters, hormones, ions, or nucleotides) and direct cell–cell communication (e.g., mediated by cell adhesion molecules, receptors, or gap junctions) that ensure both redundancy and feedback of the control mechanisms [1,2]. An obligatory feature of this network is the participation of connexin channels that ensure gap junctional communication between secretory cells [3,4,5,6]. Previous reviews have provided systematic coverage of the distribution of connexins in different types of endocrine and exocrine glands [3,4,5,6]. This chapter reviews what is known about gland connexins, and the numerous central questions on this topic that remain to be addressed. We also outline the therapeutic perspectives that are opened by our increasing knowledge of connexin biology.

26.2 What Is Known About Gland Connexins? During the last 40 years, following the innovative report that membraneimpermeable molecules may be rapidly exchanged between gland cells [7], the following information has been conclusively validated in secretory systems.

26.2.1 Connexins Are Obligatory Attributes of Secretory Cells of Multicellular Glands Connexins are expressed by most, if not all, types of secretory cells that compose exocrine, endocrine, and pheromone glands [3,4,5,6,8]. Thus, it is now established that connexins connect cells that synthesize and store a large variety of protein (e.g., endocrine pancreas), glycoprotein (e.g., exocrine pancreas), and lipidic molecules (e.g., adrenal cortex), as well as mixtures of the above (e.g., mammary gland) and catecholamines (e.g., adrenal medulla). These cells also release these products, as well as water and ions (e.g., salivary glands), salts (e.g., liver), acid (e.g., gastric glands), and sugars (e.g., liver) by either a constitutive (exocytosis independent; e.g., liver) or regulated (exocytosis dependent; e.g., pancreas) process [3,4,8]. Connexins are also largely expressed in glands that release whole living (e.g., ovary) or dead cells (e.g., sebaceous glands) by the holocrine process [3,4,8].

26 Glands

513

26.2.2 Secretory Cells Express Multiple Connexin Isoforms Of the 21 connexins that are expressed in mammals, six (Cx26, Cx30, Cx32, Cx36, Cx40, and Cx43) have been documented in primary glands [3,4,8]. While in many cases the same secretory cell simultaneously expresses several connexin isoforms (e.g., Cx26, Cx32, and Cx43 in the follicular cells of thyroid), in others a single connexin type is expressed in vivo (e.g., Cx36 in the -cells of the pancreas) [3,4,8]. The distribution of these different isoforms is not a function of the biochemical nature of the secretory product. For example, Cx43 is expressed by the peptideproducing principal cells of the parathyroid, the glycoprotein-producing cells of the pituitary, and the steroid-producing cells of the adrenal cortex [3,4,8]. Connexin distribution is also not determined by the rate and mechanism of these secretory processes. For example, Cx32 is expressed by the acinar cells of the exocrine pancreas, which periodically release large amounts of digestive enzymes by exocytosis, as well as by the hepatocytes of liver, which continuously release smaller amounts of bile salts and glucose by a constitutive route [3,4,8]. Notably, the type of connexin isoform is not linked to the embryological origin of the glands. For example, Cx26 is found in the ectoderm-derived alveolar cells of the mammary gland, in the endoderm-derived acinar cells of pancreas, and in the mesoderm-derived Sertoli cells of testis [3,4,8].

26.2.3 Exocrine and Endocrine Glands Express Incompatible Connexin Isoforms In most exocrine glands, secretory cells coexpress multiple connexins, usually of the -group (e.g., the acinar cells of the pancreas coexpress Cx26 and Cx32; Figs. 26.1 and 26.2), whereas in endocrine and pheromone glands secretory cells typically express only one isoform, often of the -group or g-group (e.g., the -cells of pancreas express Cx36; Fig. 26.2) [3,4,8] (see Chapter 1). Strikingly, the few secretory cells that have both endocrine and exocrine features coexpress connexins of different groups. An example of this is the follicular cells of thyroid, which express Cx26 and Cx32 ( -group) as well as Cx43 (-group) [8]. This differential distribution is regulated by specific transcription factors. For example, the promoter region of GJD2 (coding for CX36) comprises a cisrepressor element of 23 base pairs (bp) (RE-1), also referred to as neuronrestrictive silencer element (NRSE), which, upon binding of the cognate RE-1 silencing transcription factor (REST), prevents the expression of Cx36 except in pancreatic -cells and neurons. Alternatively, the pancreatic acinar cells of mice lacking the basic helix-loop-helix transcription factor Mist1 feature reduced levels of both Cx32 and Cx26 [9,10]. The cell-specific pattern of gland connexins appears essential for proper in vivo secretion. An example of this is the ectopic expression of Cx32 in pancreatic -cells, which natively express Cx36, alters their insulin secretion [11,12].

514

S. Bavamian et al.

Fig. 26.1 Endocrine and exocrine glands coexists in the pancreas. (a) Section through a mouse pancreas reveals three islets of Langerhans, which collectively form the endocrine portion (endo) of the gland, dispersed within the acinar tissue, which forms the exocrine portion of the organ (exo). (b) Immunostaining for amylase reveals that the enzyme (yellow) is stored at the apical pole of the exocrine acinar cells and is not found in a nearby endocrine islet. (c) Immunostaining of a consecutive section for insulin shows that the hormone (red) is restricted to the cells of the endocrine islet and is absent in the exocrine acini. Scale bar 60 mm. (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com)

26.2.4 The Native Pattern of Gland Connexins Can Be Highly Modulated Gland connexins may change as a function of the developmental or differentiation state of secretory cells. For example, Cx26 is barely detectable in the mammary glands of virgin animals, increases during the early stages of pregnancy, and reaches a maximal, steady level during lactation [13]. Gland connexins are also altered under pathophysiological conditions (e.g., the prolonged exposure to supraphysiological levels of glucose reversibly downregulates the expression of Cx36 in cells producing insulin) [14], pathological conditions (e.g., rat pancreatic -cells express Cx36, whereas the cells of a rat insulinoma express Cx43) [15,16], and pharmacological conditions (e.g., exposure to trichostatin A promotes the expression of Cx43 by primary hepatocytes) [17]. Further alterations are rapidly induced by in vitro conditions; primary liver hepatocytes,

26 Glands

515

Fig. 26.2 The endocrine and exocrine pancreas feature different gap junctions and connexins. (a) Electron microscopy reveals electron-dense amylase-containing granules in the cytoplasm of the exocrine pancreatic acinar cells (exo), and the much smaller insulincontaining granules, with a peripheral electrolucent halo, in the -cells of endocrine islets (endo). (b) This difference in granule size allows exocrine and endocrine pancreatic cells in freeze-fracture replicas to be distinguished. (c) The typical gap junction plaque found between pancreatic acinar cells is formed by hundreds to thousands of channels and feature a smooth rounded profile. (d) The typical gap junction plaque (arrowheads) found between pancreatic islet -cells is formed by a few dozen hemichannels only, features a variable, polygonal profile, and usually occurs in clusters associated with short tight junction strands.(e) Gap junctions are formed by large amounts of Cx32 (and Cx26), a connexin that is not found in pancreatic islets (endo). (f) These gap junctions are formed exclusively by Cx36, which is not found in the surrounding pancreatic acini. Scale bars 800 nm in (a) and (c), 70 nm in (d) and (e), 28 mm in (b), and 14 mm in (f). (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com)

516

S. Bavamian et al.

which natively express Cx32 with or without Cx26, almost consistently express only Cx43 after short periods of culture. Other changes occur after cell transformation; the renin-producing cells of kidney express Cx40 in situ, whereas the derived As4.1 tumor cell line coexpresses Cx37, Cx40, Cx43, and Cx45 [18,19]. These changes correlate with major qualitative secretory alterations (e.g., contrary to native hepatocytes coupled by Cx26 and Cx32, which express high levels of albumin messenger RNA [mRNA], cultured hepatocytes coupled by Cx43 have minimal levels of its transcript) and quantitative secretory alterations (e.g., As4.1 cells produce much less renin than the primary renin-producing cells of kidney) [18,19,20].

26.2.5 Connexins Form Gap Junction Plaques Between Secretory Cells Electron microscopy and immunofluorescence have consistently revealed gap junction plaques in the membranes of gland cells [3,4,5,6,7,8, 21]. Often, the junctions of exocrine secretory cells comprise much larger numbers of channels than those found in the cognate surface epithelium; for example, the gap junctions of adult hepatocytes are orders of magnitude larger than those connecting the absorptive cells of the intestinal epithelium, from which liver develops [21]. In contrast, the gap junctions of endocrine gland cells are usually significantly smaller, for example, most of the Cx36 gap junctions of pancreatic -cells comprise less than 20 channels [3,4,5,6,22]. In some glands (e.g., thyroid follicles), gap junction plaques are seen both unassociated with other junctional structures, and closely associated with tight junctions. In other secretory cells (e.g., pancreatic -cells) the gap–tight junction association is almost obligatory [3,4,5,6,7,8, 22,23].

26.2.6 Gap Junction Channels Couple Secretory Cells Microelectrode or patch clamp recordings have documented that most types of secretory cells are ionically coupled in vitro [3,4,5,6,7,8]. The existence of functional connexin channels has been further validated in vitro and in vivo by microinjection of membrane-impermeable tracer molecules, as well as by loading of secretory cells with membrane-permeable esters of other tracers, whose intercellular diffusion can be monitored by either fluorescence recovery after photobleaching or Mn2+-induced quenching of fluorescence [3,4,5,6,7,8]. Fewer studies have documented the cell–cell transfer of endogenous molecules, such as phosphorylated glucose metabolites and nucleotides [24,25]. The characteristics of both ionic and dye-coupling vary in different types of secretory cells as a function of the connexin isoforms that compose the junctional channels, for example, the homotypic Cx36 channels of pancreatic -cells have a small unitary conductance, favor the exchange of small positively charged

26 Glands

517

molecules, and are not voltage sensitive, whereas the Cx26 and Cx32 channels of pancreatic acinar cells have much larger conductance, are much more permeable to negatively charged molecules, and are voltage sensitive [26,27,28]. The spatial extent of coupling also varies from one gland to another, as a function of the size and number of gap junction plaques. For example, coupling extends throughout each acinus of the exocrine pancreas, whereas it is restricted to small groups of -cells in the endocrine pancreatic islets. The connexin dependence of both electrical and dye-coupling has been established in vivo in only a few glands, after either deletion of specific connexin genes, silencing, or overexpression of the cognate transcripts [29,30]. Depending on whether the targeted secretory cell type expresses one or multiple connexins, gap junction plaques, dye-coupling, electrical coupling, and cell–cell synchronization were abolished (e.g., in pancreatic -cells of Cx36 knockout [KO] mice), decreased (e.g., in pancreatic acinar cells of Cx32KO mice), or increased (e.g., in pancreatic -cells of transgenic mice in which the rat insulin promoter [RIP] drives the -cell–specific expression of Cx32; RIP-Cx32 mice; Fig. 26.3). In the latter two cases, the characteristics of coupling differed from those of the wild-type, in agreement with the imposed changes in the connexin expression [31].

26.2.7 Connexins Contribute to Control of Secretion The main function of glands is the regulated biosynthesis, storage, and release of secretory products. Increasing circumstantial evidence indicates that connexins are required for the fine regulation of these functions in many glands, a requirement that has been experimentally demonstrated, in vivo, in only a few cases [3,4,5].

26.2.7.1 Pancreatic Acinar Cells Acinar cells form the exocrine portion of the pancreas (99% of the adult gland volume) and secrete a mixture of about 20 (pro)enzymes. These cells are extensively coupled through Cx32 and Cx26 junctional channels [27,32,33,34]. Cx32KO mice display a significant decrease in acinar cell coupling, with the remaining coupling being provided by persistent, though reduced, expression of Cx26. This change in connexin expression produced an increase in the nonstimulated secretion of amylase and in the circulating levels of this digestive enzyme, but did not affect the maximal stimulation of amylase secretion induced by acetylcholine [32,33]. These findings show that Cx32-dependent coupling is required for maintenance of the low levels of basal pancreatic secretion that occur between meal periods, by a mechanism that cannot be mediated by the remaining, reduced levels of Cx26. While the mechanism for this is not understood, a likely hypothesis is that the widespread coupling of acinar cells, which usually extends throughout each pancreatic acinus, maintains the signal (i.e., Ca2+) for stimulation of secretion below the threshold level required for substantial activation

518

S. Bavamian et al.

Fig. 26.3 The extent and regulation of dye-coupling differs in the endocrine and exocrine pancreatic cells. (a) Microinjection of Lucifer yellow into individual acinar cells reveals that these amylase-producing cells are coupled to each other within a resting pancreatic acinus. (b) After a few minutes stimulation by acetylcholine, the same procedure reveals the uncoupling of most acinar cells. (c) Microinjection of Lucifer yellow into individual -cells reveals that restricted groups of these insulin-producing cells are coupled within a resting pancreatic islet. (d) No -to- cell diffusion of the same gap junction tracer is observed within a pancreatic islet isolated from a Cx36KO mouse. (e) In contrast, when Lucifer yellow was injected in an islet isolated from a RIPCx32 mouse, in which the rat insulin promoter drives -cell–specific expression of Cx32, significantly larger groups of coupled -cells were readily observed under resting conditions. Scale bar 30 mm. in (a) and (b), and 15 mm in (c–e). (A high-resolution color version of this figure is available on the accompanying CD and online at www.springerlink.com)

of the secretory machinery [3,4,5]. In other words, the signal is diluted by widespread junctional coupling. This implies that some of the Cx32-coupled cells do not spontaneously generate as much of the signal as others and serve as ‘‘sinks’’ to keep the basal level low. On the other hand, the finding that Cx32KO mice do not display defects in the stimulated secretion of pancreatic enzymes, indicates that either Cx32 is not

26 Glands

519

required for this postprandial function, or that its functioning is turned off to properly stimulate acinar cells, or that it can be compensated in this function by the residual Cx26. This presents an apparent paradox: stimulated secretion requires the contact-dependent recruitment and synchronized function of many secretory cells [35,36], but the phenotype of Cx32KO mice is similar to that achieved by the uncoupling effect of acetylcholine in wild-type animals [3,6,27]. Thus, acetylcholine, the main natural stimulus for the release of pancreatic (pro)enzymes, induces the uncoupling of acinar cells in both the pancreas and the physiologically related salivary glands [3,4,6], yet maximally stimulates their cognate secretions, in vitro and in vivo [6,27]. Studies of the temporal sequence of the acetylcholine effects support the view that the timing and extent of this uncoupling must be selectively controlled and finely tuned to efficiently achieve maximal stimulation of the pancreas. Immediately after binding to acinar cell receptors, acetylcholine induces in pancreatic acini waves of cytosolic free Ca2+ that propagate from cell to cell by a mechanism partly dependent on junctional channels [34] and thereby recruit large numbers of cells for active secretion [35]. Thus, coupling is initially required for spreading across pancreatic acini of the secretion-promoting second messengers generated by agonist stimulation. This is physiologically important, since only some acinar cells directly contact cholinergic terminals in situ [36]. In this way, the coupling in this initial phase compensates for this native functional heterogeneity [35,36] and ensures stimulation of sufficiently large numbers of cells when needed, while presumably retaining the low basal level of nonstimulated secretion. Subsequently, however, when the Ca2+ increase induced by acetylcholine is maximal [6,34], the acinar cells become uncoupled. This will eliminate the dilution effect mentioned above, and allow the now stimulated acinar cells to experience maximal levels of the secretory signals. The finding that the secretory alterations observed in Cx32KO mice (i.e., increased basal release of amylase and unchanged maximal release of the enzyme during stimulation) are mimicked by acute exposure of acini to uncoupling drugs, indicates that the physiological, acetylcholine-induced uncoupling depends on the gating of existing connexin channels, rather than on changes in the levels of either the Cx32 or Cx26 proteins [32]. Strikingly, whereas the proper control of basal secretion can be achieved only by Cx32, since loss of this protein causes excessive nonstimulated secretion in spite of the persistence of Cx26, the recruitment of increasing numbers of cells into active secretion is still achieved when only the latter protein persists in Cx32KO mice. Together, these studies provide direct evidence that Cx32-dependent coupling is important for a normal basal secretion of pancreatic enzymes, and that this function cannot be supported by Cx26 channels. However, the remaining Cx26 channels seem capable of providing the intercellular pathway required for recruitment and initial stimulation of acetylcholine-induced secretion. They further indicate that the coupling signaling is sufficiently important to affect the in vivo functioning of the exocrine pancreas [32].

520

S. Bavamian et al.

26.2.7.2 Pancreatic Islet Cells Each of the five secretory cell types of pancreatic islets, which collectively form the endocrine portion of the pancreas (1% of the adult gland volume), secretes a different peptide hormone (insulin, glucagon, somatostatin, pancreatic polypeptide, or ghrelin) involved in the regulation of blood glucose and metabolism [2]. Among these cells, the -cells producing insulin are the most abundant (80% of the islet volume), and are coupled by Cx36 channels [15,29]. The finding that -cells of Cx36KO mice do not have gap junction plaques and are electrically uncoupled and dye-uncoupled supports the view that -cells do not express other connexin isoforms, at least in vivo [29]. Uncoupled -cells show an increased basal release of insulin, since they cannot be restrained by the cell–cell diffusion of hyperpolarizing currents generated in the adjacent cells [29,37]. As a result of this large, nonstimulated release of the hormone, they also cannot increase their insulin secretion when challenged with physiologically relevant glucose concentrations [29]. Uncoupled -cells further show a reduced response to higher concentrations of glucose, due to their dependence on Cx36 to synchronize throughout each pancreatic islet both the glucose-induced oscillations in cytosolic Ca2+, which drive parallel oscillations in insulin output [38], and the electrotonic currents that control membrane potential and, hence, cell excitability [39]. Accordingly, Cx36KO mice do not display a pulsatile release of insulin [29], an alteration that is typical of prediabetic states and of the type II, non–insulin-dependent form of diabetes [40]. Similar observations were made in vitro, after either pharmacological blockade of connexin channels or antisense interference of Cx36 mRNA [30,41,42]. Independent experiments further indicate that substantial overexpression of Cx36 reduces glucose-induced insulin secretion, but does not affect the basal release of the hormone [16,30]. Thus, both excess of Cx36 signaling and the lack of this connexin appear to be similarly deleterious for nutrient-stimulated insulin secretion [30]. The reason why adequate levels of Cx36 signaling are required for proper insulin secretion in response to glucose stimulation is not fully understood. The natural stimulation of this secretion by glucose induces larger and more steadystate secretory and metabolic responses from intact pancreatic islets than from single cells. When tested individually, normal -cells show a large functional heterogeneity, as revealed by the intrinsic quite different biosynthetic and secretory heterogeneity of individual capabilities in vitro and in vivo -cells [2,3,43,44]. Thus, when simultaneously exposed to stimulatory levels of glucose, some -cells secrete and biosynthesize substantial amounts of insulin, whereas other apparently healthy -cells, located close to but not contacting other cells, do not. In clusters of cells coupled by Cx36 gap junctions, increasing the stimulus for secretion allows for the progressive recruitment of actively secreting cells [2,3,43,44,45,46], presumably because coupling results in an equilibration of their intrinsic differences and synchronizes their function [3,6]. While several aspects of this model remain to be experimentally validated, the

26 Glands

521

biological result is that glucose induces larger, more steady secretory and metabolic responses from intact pancreatic islets, or clusters of -cells, than from single cells [2,3,43,44,45,46]. Together, these experiments show that Cx36 is essential for the proper control of insulin secretion, via mechanisms that cannot be mediated by other connexin isoforms [12].

26.3 What We Do Not Know About Gland Connexins If compelling evidence supports a physiologically relevant role of connexins in the secretion of several glands, the following main questions remain open, if not unexplored.

26.3.1 What Is the Molecular Mechanism Whereby Connexins Control Gland Cell Function? Only in a few glands has this mechanism started to be elucidated. Many of the endogenous molecules that permeate connexin channels (e.g., Ca2+, K+, inositol triphosphate [IP3], glycolytic intermediates, nucleotides) are also implicated in the control of secretion [2,3,4,5,6], complicating identification of the signals that link changes in connexin-dependent communication to changes in gland function. The most direct evidence on this point is for the endocrine pancreas, in which stimuli-induced Ca2+ transients that require connexin channels to become synchronized in different cells play a central role [6,29,41]. Why an intercellular asynchrony of such transients is deleterious remains to be established. Conceivably, irregular Ca2+ oscillations could alter many cell events, including gene expression and apoptotic pathways [29,47,48,49]. Furthermore, secretory systems are just beginning to be evaluated for connexin functions that may not be mediated by cell coupling, but depend on the interactions of the junctional proteins with other membrane or cytosolic partners, on their control of gene expression or on the formation of hemichannels contributing to the permeability of the nonjunctional domains of the cell membrane [50,51].

26.3.2 What Is the Cellular Mechanism Whereby the Connexin-Dependent Signaling Affects Secretion? Connexin-dependent coupling is driven by diffusion, thereby achieving a rapid equilibration of ionic and molecular electrochemical gradients between coupled cells. It may be anticipated that, whenever the concentration of critical signal molecules reaches a threshold level for activation/inhibition of an effector mechanism, the secretory function of gland cells is modified not only in the cell in which the ionic and molecular change first occurred, but also in all cells coupled to it [2,3,4,5]. The coupling-induced equilibration of cytoplasmic

522

S. Bavamian et al.

molecules would also be expected to synchronize the secretory functions modulated by gap junction–permeant molecules in tissues composed of functionally heterogeneous cells. Conversely, uncoupling or the expression of connexins resulting in the establishment of electrochemical gradients could prevent an excessive dilution of critical signals, thus permitting proper activation of effector mechanisms. Several of these expectations have been experimentally verified in the pancreas, whose main secretory cells are functionally heterogeneous, and require coupling for recruitment, synchronization, and control of basal as well as stimulated secretion, but remain to be validated in most other secretory systems [35,36,43,44].

26.3.3 What Is the Hierarchical Position of the Connexin-Dependent Signaling? Vital functions, including insulin and catecholamine secretion, rely on signal pathway crosstalk and several redundant control mechanisms, which ensure both the persistence of function when one control is defective, and fine tuning of its regulation to adapt to the ever-changing needs of the organism. The finding that chronic alterations of -cell connexins are sufficient to reproduce in vivo defects of insulin secretion that are observed in type II diabetes [12,29], indicates that connexin signaling plays a significant role in the network mechanism that controls the endocrine pancreas. The reason why the connexin-dependent signaling contributes to this signaling network more prominently than other mechanisms remains to be understood, as does its hierarchical position among the pathways that allow individual cells to become integrated in a functionally coherent tissue. Also, interactions between connexin-dependent and connexin-independent pathways should be investigated. For example, the Cx36 signaling of pancreatic -cells interacts with the signaling pathway that is dependent on the K+ ATP channels [45], as well as with that activated by the ephrin ligand-receptor system [46], raising the question of which of these mechanisms actually initiates, sustains, and effects the changes in secretion. In view of the similar, but not identical, characteristics of the connexin, innexin, and pannexin proteins [52,53], future studies should also investigate whether and how the pathways controlled by these distinct families of gap junction–forming proteins crosstalk within glands.

26.3.4 Are Connexins Required for Nonsecretory Functions of Gland Cells? The involvement of connexins or coupling in secretion does not rule out roles in the developmental growth and morphogenesis of glands, the differentiation of secretory cells, and their renewal after birth. Few studies have addressed these

26 Glands

523

questions. A significant role for connexins in postnatal morphogenesis has been shown in vivo only for the mammary gland. Participation in the growth of differentiated secretory cells has been documented in the endocrine pancreas, liver, and testis [3,11,12,13]. Furthermore, it is now necessary to assess whether secretory cells interact via connexins with the vessels and ducts, which are indispensable for the proper function of glands. In the kidneys, the renin-producing cells are joined to each other and to nearby endothelial cells by Cx40, and these endothelial cells are themselves joined to companion cells by Cx43. Replacement of Cx43 by Cx32 (knock-in, KI) protects Cx43KI32 mice from an experimentally induced hypertension due to decreased renin production [54], whereas loss of Cx40 makes the animals hypertensive, due to excessive stimulation of renin biosynthesis and release [55,56]. Thus, renin secretion is controlled both by Cx40 signaling in the cells that produce the hormone, and by Cx43-dependent signaling that is generated in the nearby vascular cells [54,55]. Since Cx43 and Cx40 cannot form heterotypic channels, the data open the intriguing possibility that specific effects of connexins be mediated at short distance by the paracrine action of signal molecules, whose release is somehow controlled by specific connexins, as recently documented in the inner ear [57]. The nature of these putative signals is not determined.

26.3.5 Do Connexins Contribute to Secretory Diseases? The established participation of connexins in the physiological functions of glands, and the secretory alterations observed after the loss or blockade of these channels, raise the possibility that perturbed connexin signaling is involved in the pathogenesis of secretory diseases. While there is no direct support at this time, increasing circumstantial evidence in animal models and human tissues calls for careful investigation of this possibility. In the type I form of diabetes, an autoimmune attack kills most pancreatic -cells, leading to residual mass insufficient to sustain the insulin demand of the organism. Recent experiments on a variety of transgenic mice expressing different levels of Cx36 have shown that Cx36 protects -cells in vivo against molecules that experimentally reproduce the autoimmune attack and, conversely, that loss of Cx36 significantly sensitizes -cells to these aggressive conditions, making the mice overtly diabetic. In the type II form of diabetes, the -cell mass is less severely reduced, but the residual insulin-producing cells can no longer release the hormone in response to increase in the level of blood glucose. In humans, this defect is first heralded by the loss of the oscillatory release of insulin and, once fully established, is characterized by increased basal levels of the hormone and the inability of the pancreas to further increase its insulin output in response to a glucose challenge, leading to a sustained hyperglycemia. Most of these alterations were observed in rodents lacking Cx36 [29], whereas sulfonylurea

524

S. Bavamian et al.

treatments (that stimulate the insulin release from the glucose-unresponsive diabetic -cells) increase -cell gap junctions and coupling [22]. Together with the recent findings that prolonged exposure of insulin-producing cells to high glucose concentrations downregulates Cx36 expression [14], and that the transcript levels of GJD2 (coding for CX36) be markedly reduced in the pancreatic islets of type II diabetics, the data raise the intriguing possibility that reduced levels of Cx36 are implicated in the pathogenesis and maintenance of the disease.

26.3.6 Could We Use Connexins for Therapeutic Approaches? In view of the above, it remains to be investigated whether one could take advantage of connexins to develop innovative therapeutic approaches to diseases due to primary or secondary gland dysfunctions. This may involve identification of drugs targeted to specific connexins, which in turn implies the development of novel models for the high-throughput screening of candidate molecules. The unavailability of such molecules is presently a major limitation in testing the functions of connexins. Connexins will also be essential for the forthcoming implementation of cell therapies in which surrogate cells generated in vitro are used for the in vivo replacement of damaged cells. This replacement implies that the transplanted cells functionally integrate into the host tissue, which in glands would imply the development of appropriate connexin-dependent cell interactions. Strikingly, the embryonic stem cells and progenitor cells that are the basis of many such trials do not express many of the connexins found in adult differentiated secretory cells [58], raising the exciting prospect that proper expression of specific connexin isoforms may be feasible and instrumental for the future development of innovative, and most needed cell therapies.

26.4 Conclusion Secretion, as for other multicellular processes, implies the coordinated function of many cells, which is controlled by a variety of mechanisms for cell–cell communication. Recent studies, which have much benefited from the knowledge on connexins gained in nonsecretory systems, have shown that connexindependent mechanisms have a central role in glands. Conversely, the study of secretory systems has provided novel insights into the cell biology of connexins that were not suspected, and are hardly approachable in most other cell systems. Acknowledgments This work is supported by grants from the Swiss National Science Foundation (310000-109402), the Juvenile Diabetes Research Foundation (1-2005-46 and 1-2007158), Novo Nordisk, and the Geneva Program for Metabolic Disorders (GeMet).

26 Glands

525

References 1. LeRoith D. Are all cells ‘‘endocrine’’? In: Becker KL, editor. Principles and practice of endocrinology and metabolism. Philadelphia: JB Lippincott; 1990. pp. 10–3. 2. Meda P. Intercellular communication and insulin secretion. In: Zahnd GR, Wollheim CB, editors. Contributions of physiology to the understanding of diabetes. Berlin: Springer; 1997:24–42. 3. Michon L, Nlend R, Bavamian S, Bischoff L, Boucard N, Caille D, Cancela J, Charollais A, Charpantier E, Klee P, Peyrou M, Populaire C, Zulianello L, Meda P. Involvement of gap junctional communication in secretion. Biochim Biophys Acta. 2005;1719:82–101. 4. Serre-Beinier V, Mas C, Calabrese A, Caton D, Bauquis J, Caille D, Charollais A, Cirulli V, Meda P. Connexins and secretion. Biol Cell. 2002;94:477–92. 5. Meda P. The role of gap junction membrane channels in secretion and hormonal action. J Bioenerg Biomembr. 1996;28:369–77. 6. Petersen OH. The electrophysiology of gland cells. London: Academic Press; 1980. 7. Kanno Y, Loewenstein WR. Low-resistance coupling between gland cells. Some observations on intercellular contact membranes and intercellular space. Nature. 1964;201:194–5. 8. Meda P, Pepper MS, Traub O, Willecke K, Gros D, Beyer EC, Nicholson B, Paul D, Orci L. Differential expression of gap junction connexins in endocrine and exocrine glands. Endocrinology. 1993;133:2371–8. 9. Martin D, Tawadros T, Meylan L, Abderrahmani A, Condorelli DF, Waeber G, Haefliger JA. Critical role of the transcriptional repressor neuron-restrictive silencer factor in the specific control of connexin36 in insulin-producing cell lines. J Biol Chem. 2003;278:53082–9. 10. Rukstalis JM, Kowalic A, Zhu L, Lidington D, Pin CL, Konieczny SF. Exocrine specific expression of Connexin32 is dependent on the basic helix-loop-helix transcription factor Mist1. J Cell Sci. 2003;116:3315–25. 11. Plum A, Hallas G, Magnin T, Dombrowski F, Hagendorff A, Schumacher B, Wolpert C, Kim J, Lamers WH, Evert M, Meda P, Traub O, Willecke K. Unique and shared functions of different connexins in mice. Curr Biol. 2000;10:1083–1091. 12. Charollais A, Gjinovci A, Huarte J, Bauquis J, Nadal A, Martin F, Andreu E, Sanchez-Andres JV, Calabrese A, Bosco D, Soria B, Wollheim CB, Herrera PL, Meda P. Junctional communication of pancreatic cells contributes to the control of insulin secretion and glucose tolerance. J Clin Invest. 2000;106:235–43. 13. Locke D, Stein D, Davies C, Morris G, Harris AL, Evans WH, Monaghan P, Gusterson B. Altered permeability and modulatory character of connexin channels during mammary gland development. Exp Cell Res. 2004;298:643–60. 14. Allagnat F, Martin D, Condorelli DF, Haefliger JA. Glucose represses connexin36 in insulin-secreting cells. J Cell Sci. 2005;118:5335–44. 15. Serre-Beinier V, LeGurun S, Belluardo N, Serre-Beinier V, LeGurun S, Belluardo N, Trovato-Salinaro A, Charollias A, Haefliger JA, Condorelli DF, Meda P. Cx36 preferentially connects -cells within pancreatic islets. Diabetes. 2000;49:727–34. 16. Vozzi C, Ullrich S, Charollais A, Philippe J, Orci L, Meda P. Adequate connexin-mediated coupling is required for proper insulin production. J Cell Biol. 1995;131:1561–72. 17. Vinken M, Henkens T, Vanhaecke T, Papeleu T, Geerts A, Van Rossen E, Chipman JK, Meda P and Rogiers V. Trichostatin a enhances gap junctional intercellular communication in primary cultures of adult rat hepatocytes. Toxicol Sci. 2006;91:484–92. 18. Stutenkemper R, Geisse S, Schwarz HJ, Look J, Traub O, Nicholson BJ, Willecke K. The hepatocyte-specific phenotype of murine liver cells correlates with high expression of connexin32 and connexin26 but very low expression of connexin43. Exp Cell Res. 1992;201:43–54. 19. Haefliger JA, Demotz S, Braissant O, Haefliger JA, Demotz S, Braissant O, Suter E, Waeber G, Nicod B. Meda, P. Connexins 40 and 43 are differentially regulated within the kidneys of rats with renovascular hypertension. Kidney Int. 2001;60:190–201.

526

S. Bavamian et al.

20. Ryan MJ, Liu B, Herbowy MT, Gross KW, Hajduczok J. Intercellular communication between renin expressing As4.1 cells, endothelial cells and smooth muscle cells. Life Sci. 2003;72:1289–301. 21. Friend DS, Gilula NB. Variations in tight and gap junctions in mammalian tissues. J Cell Biol. 1972;53:758–76. 22. Meda P, Halban P, Perrelet A, Renold AE, Orci L. Gap junction development is correlated with insulin content in the pancreatic cell. Science. 1980;209:1026–28. 23. Li X, Olson C, Lu S, Nagy JI. Association of connexin36 with zonula occludens-1 in HeLa cells, TC-3 cells, pancreas, and adrenal gland. Histochem Cell Biol. 2004;122: 485–98. 24. Meda P, Amherdt M, Perrelet A, Orci L. Metabolic coupling between cultured pancreatic -cells. Exp Cell Res. 1981;133:421–30. 25. Kohen E, Kohen C, Thorell B, Mintz DH, Rabinovitch A. Intercellular communication in pancreatic islet monolayer cultures: a microfluorometric study. Science. 1979;204:862–5. 26. Moreno AP, Berthoud VM, Perez-Palacios G, Perez-Armendarig EM. Biophysical evidence that connexin-36 forms functional gap junction channels between pancreatic mouse -cells. Am J Physiol Endocrinol Metab. 2005;288:E948–56. 27. Chanson M, Orci L, Meda P. Extent and modulation of junctional communication between pancreatic acinar cells in vivo. Am J Physiol. 1991;261:G28–36. 28. Meda P, Michaels RL, Halban PA, Orci L, Sheridan JD. In vivo modulation of gap junctions and dye coupling between -cells of the intact pancreatic islet. Diabetes 1983;32:858–68. 29. Ravier MA, Guldenagel M, Charollais A, Gjinovci A, Caille D, Sohl ¨ G, Wollheim CB, Willecke K, Henquin JC, Meda P. Loss of connexin36 channels alters -cell Coupling, islet synchronization of glucose-induced Ca2+ and insulin oscillations, and basal insulin release. Diabetes 2005;54:1798–807. 30. Le Gurun S, Martin D, Formenton A, Le Gurun S, Martin D, Formenton A, Maechler P, Caille D, Waeber G, Meda P, Haefliger JA. Connexin-36 contributes to control function of insulin-producing cells. J Biol Chem. 2003;278:37690–7. 31. Harris AL. Emerging issues of connexin channels: biophysics fills the gap. Q Rev Biophys 2001;34:325–472. 32. Chanson M, Fanjul M, Bosco D, Nelles E, Suter S, Willecke K, Meda P. Enhanced secretion of amylase from exocrine pancreas of connexin32-deficient mice. J Cell Biol. 1998;141:1267–75. 33. Meda P, Bruzzone R, Chanson M, Bosco D, Orci L. Gap junctional coupling modulates secretion of exocrine pancreas. Proc Natl Acad Sci USA. 1987;84:4901–4. 34. Chanson M, Molard P, Meda S, Sutter S, Jongsma JH. Modulation of pancreatic acinar cell to cell coupling during ACh-evoked changes in cytosolic Ca2+. J Biol Chem. 1999;274:282–7. 35. Bosco D, Chanson M, Bruzzone R, Meda P. Visualization of amylase secretion from individual pancreatic acini. Am J Physiol. 1988;254:G664–70. 36. Bosco D, Soriano JV, Chanson M, Meda P. Heterogeneity and contact-dependent regulation of amylase release from individual acinar cells. J Cell Sci. 1994;160:378–388. 37. Speier S, Gjinovci A, Charollais A, Meda P, Rupnik M. Cx36-mediated coupling reduces -cell heterogeneity, confines the stimulating glucose concentration range and affects insulin release kinetics. Diabetes 2007;56:1078–86. 38. Atwater I, Rojas E, Scott A. Simultaneous measurements of insulin release and electrical activity from single microdissected mouse Islets of Langerhans. J Physiol. 1979;291:57P. 39. Valdeolmillos M, Gomis A, Sanchez-Andres JV. In vivo synchronous membrane potential oscillations in mouse pancreatic -cells: lack of coordination between islets. J Physiol. 1996;493:9–18. 40. O’Rahilly S, Turner RC, Matthews DR. Impaired pulsatile secretion of insulin in relatives of patients with non-insulin-dependent diabetes. N Engl J Med. 1988;318:1225–30.

26 Glands

527

41. Calabrese A, Zhang M, Serre-Beinier V, Caton D, Mas C, Satin LS, Meda P. Connexin 36 controls synchronization of Ca2+ oscillations and insulin secretion of MIN6 cells. Diabetes 2003;52:417–24. 42. Meda P, Bosco D, Chanosmn M, Giordano E, Vallar L, Wollheim CB, Orci L. Rapid and reversible secretion changes during uncoupling of rat insulin-producing cells. J. Clin Invest. 1990;86:759–68. 43. Salomon D, Meda P. Heterogeneity and contact-dependent regulation of hormone secretion by individual cells. Exp Cell Res. 1986;162:507–20. 44. Bosco D, Meda P. Actively synthesizing -cells secrete preferentially after glucose stimulation. Endocrinology. 1991;129:3157–66. 45. Schuster S, Mahrl M, Hofer T. Modelling of simple and complex calcium oscillations. From single-cell responses to intercellular signaling. Eur J Biochem. 2002;269:1333–55. 46. Spray DC, Ye ZC, Ransom BR. Functional connexin ‘‘hemichannels’’: a critical appraisal. Glia. 2006;54:758–73. 47. Iacobas DA, Iacobas S, Spray DC. Connexin43 and the brain transcriptome of newborn mice. Genomics. 2007;89:113–23. 48. Rocheleau JV, Remedi MS, Granada B, Rocheleau JV, Remedi MS, Granada B, Haed WS, Koster JC, Nichols CG, Piston DW. Critical role of gap junction coupled KATP channel activity for regulated insulin secretion. PLoS Biol. 2006;4:e26. 49. Konstantionova I, Nikolova G, Ohara-Imaizumi M, Meda P, Kucera T, Zarbalis K, Wurst W, Nakamatsu S, Lammert E. EphA-ephrin-A mediated cell communication regulates insulin secretion from pancreatic islets. Cell. 2007;129: 359–70. 50. Harris AL. Connexin channel permeability to cytoplasmic molecules. Prog Biophys Mol Biol. 2007;94:120–43. 51. Breitwieser GE. Calcium sensing receptors and calcium oscillations: calcium as a first messenger. Curr Top Dev Biol. 2006;73:85–114. 52. Phelan P. Innexins: members of an evolutionarily-conserved gap-junction proteins. Biochim Biophys Acta. 2005;1711:225–45. 53. Panchin YV. Evolution of gap junction proteins-the pannexin alternative. J Exp Biol. 2005;208:1415–19. 54. Haefliger JA, Krattinger N, Martin D, Pedrazzini T, Capponi A, Doring B, Plum A, Charollais A, Willecke K, Meda P. Connexin43-dependent mechanism modulates renin secretion and hypertension. J Clin Invest. 2006;116:405–13. 55. Krattinger N, Capponi A, Mazzolai L, Aubert JF, Caille D, Nicod P, Waeber G, Meda P, Haefliger JA. Loss of Connexin40 alters renin production, causing hypertension. Kidney Int. 2007;72:814–22. 56. Wagner C, de Wit C, Kurtz L, Grunberger C, Kurtz A, Schweda F. Connexin40 is essential for the pressure control of renin synthesis and secretion. Circ Res. 2007;100:556–63. 57. Cohen-Salmon M, Regnault B, Cayet N, Caille D, Demuth K, Hardelin JP, Janel N, Meda P, Petit C. Connexin30 deficiency causes intrastrial fluid-blood barrier disruption within the cochlear stria vascularis. Proc Natl Acad Sci USA. 2007;104:6229–34. 58. Oyamada M, Oyamada Y, Kaneko T, Takamatsu T. Regulation of gap junction protein (connexin) genes and functions in differentiating ES cells. Meth Mol Biol. 2002;185:63–9.

Chapter 27

Connexins and Carcinogenesis Sophie Crespin, Norah Defamie, Laurent Cronier and Marc Mesnil

Abstract Cancer was the first pathology to be associated with a dysfunction of gap junctions, over 40 years ago. Since then, data supporting this association have accumulated without explaining clearly the molecular events that enable connexins, the structural proteins of gap junctions, to control cell proliferation. It appears that one of the key determinants of the role of gap junctional intercellular communication in tumor progression is whether tumor cells form gap junctions with each other or with surrounding normal cells. Furthermore, differences in this ability may have different consequences at different stages of tumor progression. In some cases, it appears that control of the cell cycle progression by connexins may be independent of the establishment of gap junctional intercellular communication. Recent data suggest that connexins act on gene expression through pathways that have not yet been clearly elucidated. Moreover, other recent work suggests that connexins are involved not only in cell proliferation but also in other characteristics of the cancer phenotype such as invasion and metastasis. Interestingly, connexin expression is positively correlated with the latter processes and negatively correlated with the former. Keywords Tumor suppression Cell proliferation Cell cycle Gene expression Invasion Metastasis Cx26 Cx32 Cx43

27.1 Introduction Cancer was the first pathology to be associated with dysfunction of gap junctional intercellular communication (GJIC). This association, which mostly originated 40 years ago from the work of Loewenstein [1], was regarded as the unique pathological consequence of a gap junction deficiency. This early M. Mesnil (*) Institut de Physiologie et Biologie Cellulaires, CNRS-UMR 6187 Poˆle Biologie Sante´, Universite´ de Poitiers, 40 Avenue du Recteur Pineau, 86022 Poitiers Cedex, France e-mail: [email protected]

A. Harris, D. Locke (eds.), Connexins: A Guide DOI 10.1007/978-1-59745-489-6_27, Ó Humana Press, a part of Springer ScienceþBusiness Media, LLC 2009

529

530

S. Crespin et al.

idea was based on the observation that some tumor cells were not coupled by gap junctions, in contrast to their normal counterparts [2,3,4]. Since tumor cells are characterized by loss of contact inhibition, gap junctions, as key mediators of direct cell–cell contact and communication, were an obvious factor to investigate, either causative or permissive, in the progression to a cancerous state. Since the initial observations, data linking uncontrolled cell proliferation (a major characteristic of the cancer phenotype) and loss of gap junctions have accumulated. The idea that gap junctions are involved in cell growth control is strengthened by the fact that many potent tumor-promoting agents are inhibitors of GJIC [5,6]. This chapter reviews the most recent data concerning involvement of connexins in these fundamental aspects of cancer biology — cell growth regulation, invasion, and metastasis. It is still too early to propose a clear mechanistic picture for the roles of connexins in carcinogenesis but it is possible to suggest that, depending on the connexin type and the cell type in which they are expressed, connexins play different roles at different steps of cancer progression.

27.2 An Evolving Link Between Connexins and Cancer From the 1960s to the 1980s, various sets of data were gathered that led to the following summative observations: 1. Nonmutagenic carcinogens (tumor-promoting agents) and mitogens can decrease GJIC at subtoxic doses [7]. 2. Inversely, so-called antineoplastic chemicals promote cell coupling [8]. 3. Tumor-derived cells are often gap junction deficient [9]. 4. Re-induction of connexin expression (either by diffusible factors or transfection of connexin complementary DNAs [cDNAs]) produces a decrease of cell proliferation [10]. Each of these observations verified, in one way or another, the original hypothesis of Loewenstein [1], in that decreased GJIC capacity is synonymous with increased cell proliferation. From these observations, it was possible to propose a scenario in which gap junctions were actively involved in tumor progression. This scenario was most prominently put forward by Yamasaki [11], who proposed that the inhibition of GJIC by tumor promoters could elicit the clonal expansion of genetically altered cells during the promotion phase of carcinogenesis. Thereafter, during the tumor progression phase, these tumor-promoted cells, unable to communicate with their normal counterparts (either by the inability to form gap junctions at all, or the specific inability to form gap junctions with the surrounding normal cells), could continue to expand because of their isolation from normalizing growth factors that could have diffused from normal cells through a gap junctional pathway [12,13].

27 Cancer

531

During a second period (from the mid-1990s to now), it emerged that the involvement of gap junctions in cancer was more complex than previously supposed. One reason was new data from a variety of cancer models that seemed to contradict the above scenario. Some model systems, such as rat liver cancer, showed an interesting correlation between the loss of GJIC and tumor progression [14], but this correlation was not so evident in other systems [15]. In addition, the situation in vitro did not always reflect what was observed in vivo [16]; for an extensive review of this literature see [9]. Another factor added complexity to the general hypothesis: connexins are members of a multigene family [17,18], which suggests that their involvement in cell growth regulation may be specific to the type of connexin and the type of cells in which they are expressed [19]. Thus, the inhibition of cell growth could depend less on the induction of GJIC per se than on the type of connexin that is expressed and where [9]. Perhaps the most surprising recent finding is that connexin expression may drive phenotypic changes independently of GJIC. For example, normalization of human glioblastoma cell lines by induced expression of CX43 is associated with a cytoplasmic localization of the connexin but not with recovery of GJIC capacity [20]. This supports a new view that connexins (or at least some of them) might serve two functions that are not necessarily linked: modulation of direct intercellular communication and direct modulation of cell growth control. The challenge presented by these emerging data is to elucidate the molecular mechanisms permitting connexins to regulate the gene expression involved in cell growth control. Thus, connexins have been long associated with regulation of cell growth. The next section briefly discusses the underlying molecular aspects of such regulation. Their involvement has been extended unexpectedly to other aspects of the carcinogenesis process. For instance, recent data suggest that they are implicated actively, but independently from cell growth regulation, in the critical steps of invasion and metastasis. This emerging involvement appears paradoxical since it suggests that lack of connexin expression would favor the growth of cancer cells while its expression increases the aggressive phenotype of cancer cells. This aspect is extensively reviewed in later sections.

27.3 Connexins, Cell Proliferation, and Tumor Suppressor Effect This section focuses on how gap junctions and connexins may control cell proliferation and the molecular events that suggest that connexins are tumor suppressors (for an overview of the association between GJIC, connexins and cancer see [9]). The mechanisms described below are, for the most part, derived from in vitro studies where the expression of connexins has been exogenously induced.

532

S. Crespin et al.

It is generally acknowledged that GJIC is reduced during cellular growth and proliferation. There are two hypotheses for this phenomenon. The first proposes that downregulation of GJIC is the indirect result of activation of serine/ threonine-specific protein kinase cascades, involving mitogen-activated protein kinase (MAPK) and Akt (a phosphoinositide-activated kinase), which negatively regulate gap junctions [21,22]. However, most investigators believe that gap junctions are actively involved in cell proliferation control, and that their downregulation contributes to loss of growth control rather than being a consequence of it. Therefore, the second hypothesis proposes that gap junctions are regulators of cell cycle progression, that is, a ‘‘gatekeeper’’ [23] that controls exchange of cytoplasmic molecules that have regulatory effects. Consistent with this idea, in a model of regenerating rodent liver, GJIC increases during the G1 phase and decreases during the S phase. The reduction in expression of the two major liver connexins, Cx32 and Cx26, in the S phase results from decreased messenger RNA stability [24]. In another rat liver model [25], phosphorylation of Cx43 by protein kinase C induces the disruption of GJIC during the progression from G0 to the S phase (Fig. 27.1). In normal rat kidney epithelial cells, Cx43 is phosphorylated during the S phase and the G2/ M transition [26]. This state of phosphorylation is associated with a translocation of Cx43 from plasma membrane to cytoplasm (Fig. 27.1). A hyperphosphorylated form of Cx43, associated with a cytosolic localization and disruption of coupling, appears to be specific for the G2/M transition (Fig. 27.1) [27,28,29]. Moreover, it has been proposed that Cdc25A phosphatase could upregulate phosphorylation of Cx43 inducing a loss of the GJIC capacity during the G1/S transition (Fig. 27.1) [30]. Many studies have shown that expression of connexins induces upregulation of growth repressors and downregulation of growth enhancers [31,32,33,34, 35,36,37,38,39,40]. In the majority of these cases, the signaling pathway that is induced by the connexin expression is unknown (Fig. 27.2). However some evidence suggests that connexins could have a direct effect on gene expression. This idea is supported by data from an osteoblastic model in which a connexin responsive element (CxRE), a cytosine/thymine (CT)-rich region, has been identified in the promoter of the osteocalcin and 1(I) collagen genes [41]. In this model, extracellular growth stimulation induces the synthesis of second messengers that transit through gap junctions to activate ERK (extracellular signalregulated kinases, a type of MAPK) and phosphatidylinositol 3 (PI3)/Akt pathways. The translocation of ERK into the nucleus activates transcription factors that recognize CxRE and induce osteocalcin and 1(I) collagen transcription (Fig. 27.2). Another example supporting this hypothesis is provided by HEK293 cells where the effect of N-cadherin on cell proliferation and p21 expression (a cyclin-dependent kinase inhibitor, which functions as a regulator of cell cycle progression at G1 phase) are both increased by Cx43 coexpression (Fig. 27.1) [42]. Recent studies suggest that the effect of connexins on cell proliferation may be independent of their role in GJIC. One detailed study supporting this idea comes from human osteosarcoma cells, in which forced expression of Cx43

27 Cancer

533

Fig. 27.1 Connexins and cell growth regulation. In this hypothetical cell, some of the molecular events associating connexins and the regulation of the cell cycle progression are presented. The phases of the cell cycle are noted from G1 to M. In squares, the different molecular processes implicate connexins (references noted as numbers in one corner of each square). The bullets (.) indicate phosphorylation sites of the carboxyl-terminal part of Cx43. cAMP, cyclic adenosine monophosphate; cdc25A, cell division cycle 25 homolog A; Cdk1, cyclin-dependent kinase 1; Cdk2, cyclin-dependent kinase 2; Cx32, connexin32; Cx43, connexin43; Cx43P3, hyperphosphorylated form of Cx43; cycB, cyclin B; cycE, cyclin E; Her-2, human epidermal growth factor receptor-2; p21, cyclin-dependent kinase inhibitor p21; p27, cyclindependent kinase inhibitor p27; PKC, protein kinase C; skp2, S-phase kinase-associated protein 2; TCF, Tr-Cx43, truncated carboxyl-terminal domain of Cx43; Ub, ubiquitination of p27. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

inhibits cell proliferation without restoring GJIC. In this model, Cx43 increases expression of p27/Kip1 (a cyclin-dependent kinase inhibitor) via increased degradation of ubiquitin ligase subunit Skp2, responsible for the ubiquitination of p27 (Figs. 27.1 and 27.2) [31,32,33]. Finally, it is interesting to note that Cx43, or at least a cytoplasmic part of the protein, has been localized inside the nucleus by immunocytochemistry (Fig. 27.2) [43,44,45].

534

S. Crespin et al.

Fig. 27.2 Connexins and regulation of gene expression. In this hypothetical cell the signaling pathways associating connexins with the regulation of gene expressions are presented (references noted as numbers in one corner of each square). The bullets (.) indicate phosphorylation sites. AKT, v-akt murine thymoma viral oncogene homolog; Col I, 1(I) collagen; Cx32, connexin32; Cx43, connexin43; CxRE, connexin-responsive element; Cyr61, cysteine-rich angiogenic inducer 61; ERK, extracellular regulated mitogen-activated protein (MAP) kinase; FGFR3, fibroblast growth factor receptor 3; Her-2, human epidermal growth factor receptor-2; IGF-1, insulin-like growth factor 1; IGFRBP-4, insulin-like growth factor binding protein 4; MEK, MAP kinase-ERK kinase; MFG-E8, milk fat globule–epidermal growth factor (EGF) 8; nov, nephroblastoma overexpressed gene; p21, cyclin-dependent kinase inhibitor p21; p27, cyclin-dependent kinase inhibitor p27; PI3K, phosphatidylinositol 3kinase; Sp1, Sp1 transcription factor; Sp3, Sp3 transcription factor; Tr-Cx43, truncated carboxyl-terminal domain of Cx43. (A high-resolution version of this figure is available on the accompanying CD and online at www.springerlink.com)

All these elements argue for the involvement of connexins in the regulation of cellular proliferation in two ways: the first due to the capacity of cells to communicate through gap junctions, and the second due to the expression of connexin without the establishment of GJIC. During the progression toward cancer, alteration of GJIC, aberrant localization of connexins, or their lack of expression could be part of the deregulation of the cell cycle that leads to the uncontrolled proliferation of cancer cells. The regulatory effect of connexins on

27 Cancer

535

cell proliferation and the fact that connexin deficiency is known to result in increased susceptibility to carcinogenesis explains why connexins have been proposed as class II tumor suppressors [46,47].

27.4 Connexins, Cell Migration, and Invasion As mentioned previously, the correlation between reduced GJIC and neoplastic transformation has led to the general hypothesis that reduced cell–cell communication is a critical step in carcinogenesis. However, the roles that connexins play in the acquisition of a malignant phenotype remain unclear, and some of the data appear contradictory. In some cases, breakdown of GJIC in cancer cells is correlated with metastatic capacity, and in others it seems that connexin expression supports invasion and metastasis (for a review of this literature see [9]). One of the most aggressive tumors, malignant glioma, is distinguished by rapid and widespread migration of glioma cells into surrounding brain tissue. Since glial tumors are heterogeneous, the pattern of invasion is nonrandom, and the invasive margin of these tumors represents a small portion of the entire tumor, it has been hypothesized that invasive glioma cells would show phenotypic differences when compared with cells isolated from the tumor core [48]. In fact, cell lines established from different regions of a glioma (white matter, gray matter, and tumor core) differed in the relation among growth, migration, and gap junctions. Cell lines from invaded white matter showed stimulated migration on collagen and variable migration on merosin (a group of laminins that share the 2 chain), whereas migration of cell lines derived from invaded gray matter showed the reciprocal responses. In each case, there was an inverse relationship between the numbers of cells demonstrating GJIC and the migration rate of the cells; decreased expression of the Cx43 gene was associated with accelerated motility of the glioma cells. Another set of studies investigated the roles of homocellular and heterocellular GJIC on the effects of connexin expression. A study of the impact of Cx43 expression on the adhesive and invasive properties of malignant gliomas concluded that Cx43-expressing glioma cells establish functional heterocellular gap junctions with the host astrocytes [49]. This observation is key to understanding the cellular interactions between tumor cells and their surroundings. A recent study observed a pronounced lack of Cx43 protein in a well-described genetically engineered murine (GEM) glioma that exhibited little tumor cell spreading [50]. On the contrary, Cx43 protein was present in the biopsies from diffuse infiltrating GEM lesions [50]. Moreover, an in vitro approach, analyzing the participation of CX43-mediated GJIC in human glioblastoma invasion, has shown that homocellular interaction between tumor cells supports intercellular adhesion, whereas heterocellular glioblastoma-astrocyte interactions (through functional GJIC) support tumor cell migration [51]. It seems that CX43 is responsible for the heterocellular functional coupling since its levels of

536

S. Crespin et al.

expression correlated positively with invasiveness. Cx43 expression was also associated with invasiveness in a rat prostate cancer model [52]. In other words, it seems that the loss of adhesion among tumor cells correlates with their loss of homocellular coupling, a prerequisite for cell invasion; cell invasion is apparently favored (at least in the models studied) by the establishment of heterocellular communication between tumor cells and normal surrounding cells. Cx43 may not be the only connexin associated with motility in prostate cancer since a correlation between increased CX26 expression and prostate cancer progression has been observed [53]. Using three-dimensional (3D) organoid cell culture, when CX26-mediated GJIC was inhibited by blockers, invasion and migration were strikingly decreased. These results tend to confirm that GJIC is important for inducing the invasiveness phenotype, particularly if it promotes heterocellular interactions.

27.5 Connexins and Metastasis Reflecting the complexity of cancer progression, metastatic cells must overcome numerous physical obstacles that could bar metastasis. Indeed, several steps are discernible in the pathophysiological cascade of metastasis: cell dissociation, loss of cellular adhesion, tissue invasion, increased motility, entry and survival in the bloodstream, extravasation by diapedesis, and formation of secondary tumors at a distant site [54]. During this dramatic progression, cell-intrinsic mechanisms together with environmental pressures are probably involved in the selection leading to metastasis. However, no clear evidence permits the conclusion that GJIC is required for this process even if it is involved during cellular dissociation leading to invasion and in the interaction between tumor cells and endothelial cells.

27.5.1 Cell Dissociation and Cellular Heterogeneity The first description of a difference in gap junctional specialization between abdominal metastases and nontumorous cells noted that the presence of internalized gap junctions was limited to estrogen-induced renal carcinoma cells [55]. Several studies report data suggesting that the loss of GJIC correlates with the metastatic ability in human tumor cell lines [56]. Among ultrastructural differences investigated between clonal cell lines with various metastatic potentials, gap junctions and desmosomes were found only between weakly metastatic cells and between normal cells [57]. Similarly, loss of GJIC correlated with metastatic potential in mammary adenocarcinoma cells [58]). Similar results were demonstrated between mouse skin carcinoma cells [59], between human lung carcinoma cells [60], and between breast cancer cells [61]. Moreover, in the latter study, transfection with breast metastasis suppressor I cDNA leads to a CX43

27 Cancer

537

expression level and GJIC similar to that in normal breast cells. The loss of homocellular communication observed in some cancers could lead to greater cell diversity and heterogeneity within the tumor cell population [62]. Recently, in a metastatic model in mouse, Cx32 expression in renal cell carcinoma abolished the metastatic potential of these cells in lung and liver [63]. Relevant to this finding, expression of CX32 reduced the levels of soluble factors — plasminogen activator inhibitor-1 and vascular endothelial growth factor — currently associated with the process of invasion and metastasis [64].

27.5.2 Transendothelial Migration and Secondary Tumor Site To metastasize, cancer cells must invade tumor-associated vasculature (intravasation) to gain access to distant sites in the body and also to invade secondary sites (extravasation) by a transendothelial migration process termed diapedesis. Little is known about GJIC involvement in tumor cell diapedesis, and in turn, possible heterocellular communication between cancer and endothelial cells. Most observations in this context were obtained for breast tumor cells and melanoma cells. For instance, by performing a cDNA library subtraction, it was found that increased expression of Cx26 was responsible for the high spontaneous metastasis of the B16 mouse melanoma cell subline [65]. Furthermore, gap junctional uncoupling by transfection with a dominant-negative variant [65], or by a gap junction uncoupler [66,67], led to in vivo reduction of the number of melanoma metastatic nodes, illustrating the correlation between GJIC and spontaneous metastasis. In GJIC-deficient mammary epithelial tumor cells, transfection of CX43 induced the rapid formation of functional heterocellular Cx43-gap junction channels together with a twofold increase in diapedesis efficiency of human microvascular endothelial cells [68]. From this study, it was concluded that heterocellular GJIC may be an important factor in enhancing metastasis. If gap junctions are involved in extravasation, it is probably a part of a more complex phenomenon in which paracrine communication, endothelial cell adhesion, and gap junctions are involved. In any case, a clear relationship has been observed between endothelial cell adhesion and communication of lungmetastatic cancer cells. It was shown that the level of coupling at focal adhesion contacts depends on Cx43 in both cell types and on the expression level of the receptor/ligand pair that mediates adhesion between tumor cells and endothelium. Significantly increased adhesion and communication levels in highly metastatic lung carcinoma cells imply a role of GJIC in cancer metastasis, presumably by facilitating extravasation [69]. Heterocellular contacts between cancer cells and endothelial cells might also induce a rapid and transient decrease of homocellular GJIC of the latter cell type [70]. Finally, it was postulated that connexin expression could elicit specific tissue targeting of metastatic cells, particularly to bones. Indeed, heterocellular GJIC

538

S. Crespin et al.

can occur between metastatic breast carcinoma cells and the human osteoblastic cell line (hFOB1.19) at a greater functional level than the homocellular coupling between osteoblastic cells [71]. According to these studies, the heterocellular GJIC capability of cell lines with bone cells positively modulate their metastatic ability. In summary, no simple picture of GJIC in relation to metastasis emerges. However, during this complex phenomenon, spatiotemporal expression or function of connexins seems to occur, with decreased homocellular GJIC playing a role in cell dissociation and invasion from the primary site, whereas increased heterocellular GJIC is seen during intravasation or extravasation.

27.6 Conclusion Despite a long history, the association between GJIC and cancer has not yet been mechanistically explained. If the fundamental questions — Is the progression of solid tumors due to the progressive loss of GJIC? Is the loss of GJIC only a consequence of the cancer progression? — are still waiting for clear answers, new questions have been added to the previous ones: Why do some cancer cells communicate? Why could connexin expression affect the cell growth rate without necessarily involving GJIC? Paradoxically, these more recent and disturbing questions may bring interesting answers that would help in our understanding the real involvement of gap junctions in cancer. At least it is clear that connexins, contrary to what was originally thought, are not classical tumor suppressors, as mutation of their genes are rarely associated with carcinogenesis. Moreover, they seem to have different effects on different steps of carcinogenesis as, depending on the connexin isoform or the cell type in which it is expressed, connexins seem to favor proliferation when they are downregulated and increase invasion and metastasis when they are overexpressed. Overexpression of a classical tumor suppressor gene would never enhance critical steps of the carcinogenesis process. One way to approach the real involvement of connexins in the cancer process would be to study carefully their behavior, in terms of level of expression and subcellular localization, in the complex tumoral microcosm (in particular, in tumor cell subpopulations such as cancer stem cells). Then, depending on the cancer model, a new picture could emerge showing the connexins differently involved in the different steps of cancer progression: downregulation or aberrant localization in proliferating cancer cells and reexpression (and possibly GJIC establishment with other cell types such as endothelial cells) during invasion and metastasis. Acknowledgments This work was supported by La Ligue Contre le Cancer (Comite´s de la Vienne, de la Charente et de la Charente-Maritime).

27 Cancer

539

References 1. Loewenstein WR. Junctional intercellular communication and the control of growth. Biophys Biochim Acta. 1979;560:1–65. 2. Loewenstein WR, Kanno Y. Intercellular communication and the control of tissue growth: lack of communication between cancer cells. Nature. 1966;209:1248–1249. 3. Loewenstein WR, Kanno Y. Intercellular communication and tissue growth. I. Cancerous growth. J Cell Biol. 1967;33:225–34. 4. Penn RD. Ionic communication between liver cells. J Cell Biol. 1966;29:171–4. 5. Murray AW, Fitzgerald DJ. Tumor promoters inhibit metabolic cooperation in cocultures of epidermal and 3T3 cells. Biochem Biophys Res Commun. 1979;91:395–401. 6. Yotti LP, Chang CC, Trosko JE. Elimination of metabolic cooperation in Chinese hamster cells by a tumor promoter. Science. 1979;206:1089–91. 7. Budunova IV, Williams GM. Cell culture assays for chemicals with tumor-promoting or tumor-inhibiting activity based on the modulation of intercellular communication. Cell Biol Toxicol. 1994;10:71–116. 8. King TJ, Bertram JS. Connexins as targets for cancer chemoprevention and chemotherapy. Biophys Biochim Acta. 2005;1719:146–60. 9. Mesnil M, Crespin S, Avanzo JL, Zaidan-Dagli ML. Defective gap junctional intercellular communication in the carcinogenic process. Biophys Biochim Acta. 2005;1719:125–45. 10. Yamasaki H, Naus CC. Role of connexin genes in growth control. Carcinogenesis. 1996;17:1199–213. 11. Yamasaki H. Gap junctional intercellular communication and carcinogenesis. Carcinogenesis. 1990;11:1051–58. 12. Yamasaki H. Aberrant expression and function of gap junctions during carcinogenesis. Environ Health Perspect. 1991;93:191–7. 13. Mesnil M, Yamasaki H. Cell-cell communication and growth control of normal and cancer cells: evidence and hypothesis. Mol Carcinog. 1993;7:14–7. 14. Krutovskikh VA, Oyamada H, Yamasaki H. Sequential changes of gap-junctional intercellular communications during multistages rat liver carcinogenesis: direct measurement of communication in vivo. Carcinogenesis. 1991;12:1701–6. 15. Wilgenbus KK, Kirkpatrick CJ, Knuechel R, Willecke K, Traub O. Expression of Cx26, Cx32, Cx43 gap junction proteins in normal and neoplastic human tissues. Int J Cancer. 1992;51:522–9. 16. Jamieson S, Going JJ, D’Arcy R, George WD. Expression of gap junction proteins connexin 26 and connexin 43 in normal human breast and in breast tumors. J Pathol. 1998;184:37–43. 17. Kumar NM, Gilula NB. The gap junction communication channel. Cell. 1996;84:381–8. 18. Willecke K, Eiberger J, Degen J, Eckardt D, Romualdi A, Guldenagel M, Deutsch U, So¨hl G. Structural and functional diversity of connexin genes in the mouse and human genome. Biol Chem. 2002;383:725–37. 19. Mesnil M, Krutovskikh V, Piccoli C, Elfgang C, Traub O, Willecke K, Yamasaki H. Negative growth control of HeLa cells by connexin genes: connexin species-specificity. Cancer Res. 1995;55:629–39. 20. Huang RP, Fan Y, Hossain MZ, Peng A, Zeng ZL, Boynton AL. Reversion of the neoplastic phenotype of human glioblastoma cells by connexin 43 (Cx43). Cancer Res. 1998;58:5089–96. 21. Kojima T, Yamamoto T, Murata M, Lan M, Takano K, Go M, Ichimiya S, Chiba H, Sawada N. Role of the p38 MAP-kinase signaling pathway for Cx32 and claudin-1 in the rat liver. Cell Commun Adhes. 2003;10:437–43. 22. Kojima T, Yamamoto T, Lan M, Murata M, Takano K, Go M, Ichimiya S, Chiba H, Sawada N. Inhibition of MAP kinase activity moderates changes in expression and

540

23. 24.

25. 26.

27.

28. 29.

30.

31.

32. 33.

34.

35.

36.

37.

38.

39.

S. Crespin et al. function of Cx32 but not claudin-1 during DNA synthesis in primary cultures of rat hepatocytes. Med Electron Microsc. 2004;37:101–13. Vinken M, Vanhaecke T, Papeleu P, Snykers S, Henkens T, Rogiers V. Connexins and their channels in cell growth and cell death. Cell Signal. 2006;18:592–600. Kren BT, Kumar NM, Wang SQ, Gilula NB, Steer CJ. Differential regulation of multiple gap junction transcripts and proteins during rat liver regeneration. J Cell Biol. 1993;123:707–18. Lampe PD, Lau AF. The effects of connexin phosphorylation on gap junctional communication. Int J Biochem Cell Biol. 2004;36:1171–86. Solan JL, Fry MD, TenBroek EM, Lampe PD. Connexin43 phosphorylation at S368 is acute during S and G2/M and in response to protein kinase C activation. J Cell Sci. 2003;116:2203–11. Xie H, Laird DW, Chang TH, Hu VW. A mitosis-specific phosphorylation of the gap junction protein connexin43 in human vascular cells: biochemical characterization and localization. J Cell Biol. 1997;137:203–10. Kanemitsu MY, Jiang W, Eckhart W. Cdc2-mediated phosphorylation of the gap junction protein, connexin43, during mitosis. Cell Growth Differ. 1998;9:13–21. Lampe PD, Kurata WE, Warn-Cramer BJ, Lau AF. Formation of a distinct connexin43 phosphoisoform in mitotic cells is dependent upon p34cdc2 kinase. J Cell Sci. 1998;111:833–41. Melchheier I, von Montfort C, Stuhlmann D, Sies H, Klotz LO. Quinone-induced Cdc25A inhibition causes ERK-dependent connexin phosphorylation. Biochem Biophys Res Commun. 2005;327:1016–23. Zhang YW, Morita I, Ikeda M, Ma KW, Murota S. Connexin43 suppresses proliferation of osteosarcoma U2OS cells through post-transcriptional regulation of p27. Oncogene. 2001;20:4138–49. Zhang YW, Kaneda M, Morita I. The gap junction-independent tumor-suppressing effect of connexin 43. J Biol Chem. 2003;278:44852–6. Zhang YW, Nakayama K, Nakayama KI, Morita I. A novel route for connexin 43 to inhibit cell proliferation: negative regulation of S-phase kinase-associated protein (Skp2). Cancer Res. 2003;63:1623–30. Koffler L, Roshong S, Kyu-Park I, Cesen-Cummings K, Thomspson DC, Dwyer-Nield LD, Rice P, Mamay C, Malkinson AM, Ruch, RG. Growth inhibition in G(1) and altered expression of cyclin D1 and p27(kip-1) after forced connexin expression in lung and liver carcinoma cells. J Cell Biochem. 2000;79:347–54. Bradshaw SL, Naus CC, Zhu D, Kidder GM, D’Ercole AJ, Han VK. Alterations in the synthesis of insulin-like growth factor binding proteins and insulin-like growth factors in rat C6 glioma cells transfected with a gap junction connexin43 cDNA. Regul Pept. 1993;48:99–112. Qin H, Shao Q, Belliveau DJ, Laird DW. Aggregated DsRed-tagged Cx43 and overexpressed Cx43 are targeted to lysosomes in human breast cancer cells. Cell Commun Adhes. 2002;8:433–9. Goldberg GS, Bechberger JF, Tajima Y, Merritt M, Omori Y, Gawinowicz MA, Narayanan R, Tan Y, Sanai Y, Yamasaki H, Naus CC, Tsuda H, Nicholson BJ. Connexin43 suppresses MFG-E8 while inducing contact growth inhibition of glioma cells. Cancer Res. 2000;60:6018–26. Flachon V, Tonoli H, Selmi-Ruby S, Durand C, Rabilloud R, Rousset B, MunariSilem Y. Thyroid cell proliferation in response to forced expression of gap junction proteins. Eur J Cell Biol. 2002;81:243–52. Fujimoto E, Satoh H, Negishi E, Ueno K, Nagashima Y, Hagiwara K, Yamasaki H, Yano T. Negative growth control of renal cell carcinoma cell by connexin 32: possible involvement of Her-2. Mol Carcinog. 2004;40:135–142.

27 Cancer

541

40. Gupta N, Wang H, McLeod TL, Naus CC, Kyurkchiev S, Advani S, Yu J, Perbal B, Weichselbaum RR. Inhibition of glioma cell growth and tumorigenic potential by CCN3 (NOV). Mol Pathol. 2001;54:293–9. 41. Stains JP, Civitelli R. Cell-to-cell interactions in bone. Biochem Biophys Res Commun. 2005;328:721–7. 42. Kamei J, Toyofuku T, Hori M. Negative regulation of p21 by b-catenin/TCF signaling: a novel mechanism by which cell adhesion molecules regulate cell proliferation. Biochem Biophys Res Commun. 2003;312:380–7. 43. de Feijter AW, Matesic DF, Ruch RJ, Guan X, Chang CC, Trosko JE. Localization and function of the connexin 43 gap-junction protein in normal and various oncogeneexpressing rat liver epithelial cells. Mol Carcinog. 1996;16:203–12. 44. Moorby C, Patel M. Dual functions for connexins: Cx43 regulates growth independently of gap junction formation. Exp Cell Res. 2001;271:238–48. 45. Dang X, Doble BW, Kardami E. The carboxy-tail of connexin-43 localizes to the nucleus and inhibits cell growth. Mol Cell Biochem. 2003;242:35–8. 46. Trosko JE. The role of stem cells and gap junctional intercellular communication in carcinogenesis. J Biochem Mol Biol. 2003;36:43–8. 47. Lee SW, Tomasetto C, Sager R. Positive selection of candidate tumor-suppressor genes by subtractive hybridization. Proc Natl Acad Sci USA. 1991;88:2825–9. 48. McDonough WS, Johansson A, Joffee H, Giese A, Berens ME. Gap junction intercellular communication in gliomas is inversely related to cell motility. Int J Dev Neurosci. 1999;17:601–11. 49. Lin JHC, Takano T, Cotrina ML, Arcuino G, Kang J, Liu S, Gao Q, Jiang L, Li F, Lichtenberg-Frate H, Haubrich S, Willecke K, Goldman SA, Nedergaard M. Connexin 43 enhances the adhesivity and mediates the invasion of malignant glioma cells. J Neurosci. 2002;22:4302–11. 50. Jensen NA, Pedersen KM, Lihme F, Rask L, Nielsen JV, Rasmussen TE, Mitchelmore C. Astroglial c-Myc overexpression predisposes mice to primary malignant gliomas. J Biol Chem. 2003;278:8300–8. 51. Oliveira R, Christov C, Guillamo JS, de Bouard S, Palfi S, Venance L, Tardy M, Peschanski M. Contribution of gap junctional communication between tumor cells and astroglia to the invasion of the brain parenchyma by human glioblastomas. BMC Cell Biol. 2005;6:7. 52. Miekus K, Czernik M, Sroka J, Czyz J, Madeja Z. Contact stimulation of prostate cancer cell migration: the role of gap junctional coupling and migration stimulated by heterotypic cell-to-cell contacts in determination of the metastatic phenotype of Dunning rat prostate cancer cells. Biol Cell. 2005;97:893–903. 53. Tate AW, Lung T, Radhakrishnan A, Lim SD, Lin X, Edlund M. Changes in gap junctional connexin isoforms during prostate cancer progression. Prostate. 2006;66:19–31. 54. Gupta GP, Massague´ J. Cancer metastasis: building a framework. Cell. 2006;127: 679–95. 55. Letourneau RJ, Li JJ, Rosen S, Villee CA. Junctional specialization in estrogen-induced renal adenocarcinomas of the golden hamster. Cancer Res. 1975;35:6–10. 56. Navolotski A, Rumjnzev A, Lu H, Proft D, Bartholmes P, Zanker KS. Migration and gap junctional intercellular communication determine the metastatic phenotype of human tumor cell lines. Cancer Lett. 1997;118:181–7. 57. Ren J, Hamada J, Takeichi N, Fujikawa S, Kobayashi H. Ultrastructural differences in junctional intercellular communication between highly and weakly metastatic clones derived from rat mammary carcinoma. Cancer Res. 1990;50:358–62. 58. Nicolson GL, Dulski KM, Trosko JE. Loss of intercellular junctional communication correlates with metastatic potential in mammary adenocarcinoma cells. Proc Natl Acad Sci USA. 1988;85:473–6. 59. Kamibayashi Y, Oyamada Y, Mori M, Oyamada M. Aberrant expression of gap junction proteins (connexins) is associated with tumor progression during multistage mouse skin carcinogenesis in vivo. Carcinogenesis. 1995;16:1287–97.

542

S. Crespin et al.

60. Zhang ZQ, Zhang W, Wang NQ, Bani-Yaghoub M, Lin ZX, Naus CC. Suppression of tumorigenicity of human lung carcinoma cells after transfection with connexin43. Carcinogenesis. 1998;19:1889–94. 61. Saunders MM, Seraj MJ, Li Z, Zhou Z, Winter CR, Welch DR, Donahue HJ. Breast cancer metastatic potential correlates with a breakdown in homospecific and heterospecific gap junctional intercellular communication. Cancer Res 2001;61:1765–7. 62. Carystinos GD, Bier A, Batist G. The role of connexin-mediated cell-cell communication in breast cancer metastasis. J Mammary Gland Biol Neoplasia. 2001;6:431–40. 63. Fujimoto E, Sato H, Shirai S, Nagashima Y, Fukumoto K, Hagiwara H, Negishi E, Ueno K, Omori Y, Yamasaki H, Hagiwara K, Yano T. Connexin32 as a tumor suppressor gene in a metastatic renal cell carcinoma cell line. Oncogene 2005;24:3684–90. 64. Hagiwara H, Sato H, Shirai S, Kobayashi S, Fukumoto K, Ishida T, Seki T, Ariga T, Yano T. Connexin 32 down-regulates the fibrinolytic factors in metastatic renal cell carcinoma cells. Life Sci. 2006;78:2249–54. 65. Ito A, Katoh F, Kataoka TR, Okada M, Tsubota N, Asada H, Yoshikawa K, Maeda S, Kitamura Y, Yamasaki H, Nojima H. A role for heterologous gap junctions between melanoma and endothelial cells in metastasis. J Clin Invest. 2000;105:1189–97. 66. Ito A, Morita N, Miura D, Koma Y, Kataoka TR, Yamasaki H, Kitamura Y, Kita Y, Nojima H. A derivative of oleamide potently inhibits the spontaneous metastasis of mouse melanoma BL6 cells. Carcinogenesis 2004;25:2015–22. 67. Ohba Y, Kanao Y, Morita N, Fujii E, Hohrai M, Takatsuji M, Hirose H, Miura D, Watari A, Yutsudo M, Zhao H, Yabuta N, Ito A, Kita Y, Nojima H. Oleamide derivatives suppress the spontaneous metastasis by inhibiting connexin 26. Int J Cancer. 2007;121:47–54. 68. Pollmann MA, Shao Q, Laird DW, Sandig M. Connexin 43 mediated gap junctional communication enhances breast tumor cell diapedesis in culture. Breast Cancer Res. 2005;7:R522–34. 69. el-Sabban ME, Pauli BU. Adhesion-mediated gap junctional communication between lung-metastatic cancer cells and endothelium. Invasion Metastasis. 1994–1995;14:164–76. 70. Cai J, Jiang WG, Mansel RE. Gap junctional communication and the tyrosine phosphorylation of connexin 43 in interaction between breast cancer and endothelial cells. Int J Mol Med. 1998;1:273–78. 71. Kapoor P, Saunders MM, Li Z, Zhou Z, Sheaffer N, Kunze EL, Samant RS, Welch DR, Donahue HJ. Breast cancer metastatic potential: correlation with increased heterotypic gap junctional intercellular communication between breast cancer cells and osteoblastic cells. Int J Cancer 2004;111:693–7.

Appendix: Toward a New Nomenclature for Connexin Genes Gerald M. Kidder

Abstract Beginning with the 2005 International Gap Junction Conference, an international working group has been reexamining the nomenclature for the connexin genes of humans and other mammals. The objective was to eliminate the confusion that has been caused in the past from the use of multiple connexin gene symbols and different symbols for orthologous connexin genes in different mammals. At the 2007 International Gap Junction Conference, it was decided to endorse and update the naming system that has been considered the official nomenclature for the past decade; this nomenclature recognizes subdivisions of the gene family based on DNA sequence comparisons. Orthologous genes in other mammalian species will have the same symbols as the cognate human genes, whereas nonorthologous genes will be assigned unique numbers. This appendix summarizes the progress to date in establishing a generally accepted connexin nomenclature. Keywords Genes Nomenclature Proteins

History For more than a decade, the official names of connexin genes have included the root symbol GJ (for gap junction) combined with (Greek) letters designating subgroups based on degrees of sequence identity, e.g., GJA1 for CX43, GJB1 for CX32, etc. (see Chapter 1). This system had been adopted when only a few of the genes had been cloned and without general discussion within the gap junction research community. Furthermore, its adoption preceded the discovery of a second, unrelated gene family in vertebrates, the pannexins (PANX), encoding proteins structurally similar to connexins that may also form gap G.M. Kidder Department of Physiology and Pharmacology, Schulich School of Medicine and Dentistry, University of Western Ontario, London, Ontario N6A 5C1, Canada e-mail: [email protected]

543

544

G.M. Kidder

junction–like channels [1], although this possibility is now being met with skepticism (see Chapter 12). With the completion of the human and other mammalian genome sequences and realizing that no clear functional rationale has yet emerged for partitioning connexins into subgroups, it was considered by many in the community that the time was right for reexamination of the nomenclature issue. It had become common in the literature for connexin genes to be named using a variety of symbols, and connexins encoded by orthologous genes in different species had been referred to by different names while some encoded by nonorthologous genes in different species had been referred to by the same name [2]. A connexin gene nomenclature was needed that would be widely accepted and would eliminate such confusion. At the 2005 International Gap Junction Conference held in British Columbia, Canada, the decision was made to devise a way of naming connexin genes that does not depend on nucleotide sequence comparisons. Subsequently, the working group proposed a new nomenclature, endorsed by the Human and Mouse Gene Nomenclature Committees, which assigns arbitrary numbers (from 1 to 22) to individual mammalian connexin genes. A key feature of this system was that orthologous connexin genes in all mammals would have the same name. This system could easily be harmonized with the connexin genes of other vertebrates such as the zebrafish, where 37 putative connexin genes have been identified in the genome [3], by assigning additional numbers to those genes considered not orthologous to mammalian genes. CXN was proposed as the root gene symbol (i.e., CXN1, CXN2) to distinguish the new system from the old.

Current Status The proposed new connexin gene nomenclature was submitted for publication so that, once in print, it would become official. However, in response to comments from anonymous reviewers and members of the gap junction research community, the working group decided to withdraw the manuscript until the issue of protein names could also be addressed. Mechanisms for aligning gene and protein names were the subject of discussions within the gap junction research community leading up to and at the International Gap Junction Conference in Denmark in August 2007. At the 2007 conference, the decision made at the previous meeting to abandon the sequence-based subdivisions of the family as an element of the gene nomenclature was rescinded by an overwhelming vote. The system that was endorsed is based on the existing official human connexin gene names (GJA1, GJB2, etc.) with currently unassigned genes being named according to their orthologous relationships to the other genes, and with some recognized errors in subgroup designation being corrected. Where connexin genes in other

Proposed Connexin Gene Nomenclature

545

animals are clearly orthologous with particular human genes, they assume the human gene names; where they are not orthologous, new names will be assigned that do not duplicate existing names in other species. As the process of defining orthologous relationships for different species is completed, the official gene names are being posted on the respective species nomenclature Web sites. (for human genes go to http://www.genenames.org/genefamily/gj.php and for mouse genes go to http://www.informatics.jax.org/ to search for the term ‘connexin’)

References 1. Panchin Y, Kelmanson I, Matz M, Lukyanov K, Usman N, Lukyanov S. A ubiquitous family of putative gap junction molecules. Curr Biol. 2000;10:R473–4. 2. Sohl ¨ G, Willecke K. An update on connexin genes and their nomenclature in mouse and man. Cell Commun Adhes. 2003;10:173–80. 3. Eastman SD, Chen TH, Falk MM, Mendelson TC, Iovine MK. Phylogenetic analysis of three complete gap junction gene families reveals lineage-specific duplications and highly supported gene classes. Genomics. 2006;87:265–74.

Index

A a-Group connexins, See Connexin subgroups Actin, 255, 269, 272, 274, 279 Acute respiratory distress syndrome (ARDS), 360, 367 See also Respiratory epithelium Adenosine diphosphate (ADP), 188–190, 192 Adenosine monophosphate (AMP), 188–189, 192 Adenosine triphosphate (ATP), 186, 188–189, 192–193 Adenosine triphosphate (ATP) release by glia, 185, 379, 413 in inner ear, 426 by keratinocytes, 314 by macrophages, 476 in respiratory epithelium, 362, 365, 367 in retina, 413 in skeletal biology, 378–379 Adenosine triphosphate (ATP) release channel, identity of, 184–185, 291–295 See also Ca2+ wave propagation Adherens junction, 225, 230, 232–234, 269, 271–274, 436, 438–439, 443, 446 Adrenal gland, see Secretory glands Airway epithelium, 291, 293, 359–368 See also Respiratory epithelium Akt kinase, 532, 534 Alexa 350 (A350), Alexa 488 (A488), Alexa 594 (A594), see Tracers Alkanols, 91–92, 209, 214–216 heptanol, 185, 209, 215 hexanol, 215 octanol, 135, 185, 209, 214–215, 294 See also Pharmacology of connexin channels

Alveolar epithelium, 361–368 See also Respiratory epithelium Amacrine cells, 149, 153, 156, 271, 335, 398–402, 408–413 AII amacrine–bipolar cell coupling, 149, 410–412 AII amacrine cell network, 408–410 See also Retina 2-aminophenoxyborate (2-APB), 208, 214, 217 See also Pharmacology of connexin channels Aminosulfonate hydroxyethyl-piperazine ethanesulfonic acid (HEPES), 63–64, 137–138 and pH gating, 62–64, 137–138 taurine, 138 See also Chemical gating Amino-terminal domain (NT), 7–8, 38, 265, 268 contribution to pore aqueous pore accessibility studies, 52–53, 63–64, 78, 91, 97 open channel rectification and selectivity studies, 52–53, 85–86 polyamine blockade implications for, 53, 86 voltage-gating studies, 83–84, 114–118, 121 as gating particle, 63–64, 124–126 and heterotypic compatibility, 265, 268 structure of, 63–65, 91, 98 See also Connexin domain organization Androgen, 228–229, 499, 503 Anesthetics, see General anesthetics Annular gap junctions, see Gap junction plaque biogenesis, maintenance and degradation

547

548 Antigen presentation, 188, 251 Antimalarial drugs, 211–213 benzylquininium, 212 mefloquine, 92, 185, 208–209, 212–213 quinine, 92, 208–209, 211–212 See also Pharmacology of connexin channels Antisense RNA knockdown of Cx26, 343 of Cx32, 343 of Cx36, 520 of Cx43, 312, 343, 372–373, 375 of pannexin, 294 Apolipoprotein E-deficient mice (ApoE / ), see Atherosclerosis Apoptosis and cell death, 296, 313, 343–344, 366, 376, 391, 428, 444, 488, 501, 503, 521 Aquaporin, 245 Arrhythmia, 14–15, 19, 130, 136, 442–448 atrial fibrillation, 19, 443, 448–449 Cx40 and, 14, 19, 442–443, 448 Cx43 and, 15, 442, 444–445, 447–448 ventricular fibrillation, 130 See also Heart; Ventricular heart disease Astrocytes, 55, 153, 244, 328–335, 337, 343, 413 astrocyte-astrocyte (A-A) coupling, 55, 328, 331, 334–335 astrocyte-oligodendrocyte (A-O) coupling, 328, 332, 334, 340 See also Nervous System Atherosclerosis, 469–479 atheromas, 471–474, 476–477 connexins and, 19, 472–477 Cx37 polymorphism, 19, 476–477 inflammation in, 469–470, 472, 475 mouse models ApoE / (apolipoprotein E-deficient mice), 471, 476–477 LDLR / (LDL receptor-deficient mice), 471, 473–475 pathogenesis, 470 statins, 475–476 Atomic force microscopy (AFM), 62–63, 80, 138, 245, 254 Atrial fibrillation, see Arrhythmia Atrioventricular (AV) node, 436, 440–442 See also Heart Auditory system, 419–429 adenosine triphosphate (ATP) release, 426 anatomy and physiology, 419–422 Ca2+ waves, 425–426

Index deafness, connexin mutants causing, 16, 51–52, 80, 310, 314–315, 426–429 mouse models of, 426–429 nonsyndromic, 16, 51, 310, 426–427 syndromic, 16, 310–311, 314–315, 367, 427 See also Connexin diseases gap junctions, 419, 422–426 expression, 422–423 functions, 423–424 molecular selectivity, 424–426 properties, 424–426 B b-Group connexins, see Connexin subgroups Benzylquininium, see Antimalarial drugs; Pharmacology of connexin channels Biocytin, see Tracers Biological molecules, permeability of, see Connexin pore - molecular permeability Bipolar cells, 149–153, 156, 335, 399–411 See also Retina Bisphosphonate, 376 Bone development, see Skeletal biology Brefeldin A, 227, 229, 270, 294 See also Gap junction plaque biogenesis, maintenance and degradation C Ca Molecular model of transmembrane domains, 50–52 See also Connexin channel structure; Transmembrane domains Ca2+-calmodulin protein kinase II, 155 Ca2+ waves connexin versus pannexin involvement, 184–185, 291–295 in glia, 334 in inner ear, 425–426 in respiratory epithelium, 361–362, 365–366 in retina, 413 in secretory glands, 519 in skeletal biology, 377–379 in vasculature, 461–462 See also Adenosine triphosphate (ATP) release channel Cadherins, see Connexin interacting proteins Calcein, see Tracers

Index Calmodulin, see Connexin interacting proteins Calpain protease, see Lens Cancer, 529–538 breast, 502 bystander effects, 237, 537 cell migration, and invasion, 535–536 cell proliferation, and tumor suppressor effect, 531–535 kidney, 516, 536–537 liver, 271, 531, 537 lung, 367 metastasis, 536–538 cell dissociation and cellular heterogeneity, 536–537 transendothelial migration, 537–538 prostate, 504 skin basal cell carcinoma (BCC), 311 squamous cell carcinoma (SCC), 311 testicular, 502 Carbenoxolone, see Glycyrretinic acid and derivatives; Pharmacology of connexin channels Carboxyfluorescein, see Tracers Carboxyl-terminal (CT) domain, 7–8, 38–39, 64–65, 78, 80–81, 95 and 2-APB effects, 214 in chemical gating, 64, 123–124, 130–134, 137–138, 268–269 See also Chemical gating; Particle-receptor hypothesis dimerization, 132–134 effects of tagging, 92, 214 flexibility, 62 interactions with cytoplasmic loop (CL), 64, 123, 133–134 interactions with other proteins, 135 calmodulin, 276 disk large homolog-1 (Dlgh1), 280 src, 135, 276 ZO-1, 135, 255, 273–274, 276 ZO-2, 274 See also Connexin interacting proteins interactions with RXP-E peptide, 135 phosphorylation of, 264–265, 267, 271, 391 and disease, 311, 443 and trafficking, 267 and residual state, 92 solution structure of, 132–133, 276 and trafficking, 267

549 in Vi-o gating, 125 in Vj/fast-gating, 121, 123–124 Cardiomyocytes, 136, 216, 228, 270, 272, 276, 279, 436–442, 444–445, 488 Casein kinase (CK) CK1, 264, 269, 274–275, 391 CK2, 278 Cataracts, 16, 389–393 See also Connexin diseases; Lens Catenin, see Connexin interacting proteins Caveolin, see Connexin interacting proteins Cell adhesion and migration, 232–234, 253, 270, 272, 312, 442–443, 475–477, 500, 535–537 Cell cycle and progression of, 269, 273–274, 276, 279, 532–534 Cell lines and cultured cells airway epithelial, 361–363, 366 astrocytes, 276 astrocytoma, 294 ATI and ATII pneumocytes, 365 cardiac atrial cells, 449 cardiac myocytes, 228 chondrocytes, 378 cochlear supporting cells, 425 endothelial, 366, 459, 463 epidermis, 313 epithelial (unspecified), 234 fibroblasts (transfected), 210, 250, 255, 270, 273 glial, 216, 336 glioblastoma, 531, 535 HeLa (transfected), 251, 254, 271–272, 276, 278 hepatocytes, 516 hepatoma, 229, 271 keratinocyte, 313, 315 kidney, 44, 213 Leydig cell, 498 macrophage-derived, 476 melanoma, 537 Neuro2A cells, 19, 111–112, 137, 168 neurons, 211, 336 oligodendrocytes, 332 osteoblast-like, 376, 379, 538 osteoblasts, 378–379 prostate epithelial, 228, 503–504 Schwann cell, 324 seminoma, 502 Sertoli cell, 498 smooth muscle, 459, 463, 504 tracheal epithelial, 361

550 Cell lines and cultured cells (cont.) transfected (unspecified), 14, 31, 53, 56, 175–176, 184, 208, 227, 229, 231, 234, 246, 250–251, 265, 279, 288–289, 311, 326, 340, 424, 516, 536 trophoblast, 279 tumor cell (unspecified), 535–536 Cell-penetrating peptide (CPP), 136 Cell proliferation, 275, 311–312, 314, 324, 367, 373, 380, 391, 475, 484, 499–503, 530–535 Charcot-Marie-Tooth disease, X-linked (CMTX), 15–16, 51–52, 227, 236, 325–328, 337–338 See also Connexin diseases; Nervous system Charge selectivity, see Connexin pore - charge selectivity Chemical gating (pH gating), 41, 62–64, 123–239, 156 non-Cx43 connexins, 137–139 particle-receptor mechanism, 123, 130–136 and RXP-E peptide, 135–136 Chemical synapses, 146–147, 157 differences from electrical synapses, 146–147, 157 mixed chemical and electrical synapses, 155, 335 See also Electrical synapses; Nervous system Chimeric connexins Cx26*Cx32, 11, 65 Cx32*Cx37, 87 Cx32*Cx43, 60 Cx32*43E1, 83, 85–87, 91, 105, 114–115, 295 Note: for point mutants made on Cx32*43E1 background, see Connexin mutants - cysteine substitution; Connexin mutants point mutants and deletions Cx32*Cx43*Cx46, 87 Cx32*Cx46, 52 Cx40*Cx43, 60 Cx46*Cx50, 59–60 Cholesterol and connexin channel inhibitors, 215–216 hypercholesterolemia, 470–471 in junctional plaques, 248–251, 266 low-density lipoprotein (LDL)-cholesterol, 470

Index statins, 475–476 See also Atherosclerosis Chondrocytes, see Skeletal biology CIP85, see Connexin interacting proteins Circular dichroism (CD) spectroscopy a-helical structure of transmembrane domains, 42–43 Claudin, see Connexin interacting proteins Clouston syndrome, see Connexin diseases; Hidrotic ectodermal dysplasia (HED); Skin Conductance, see Connexin pore - conductance Conductance substate (residual state), 82, 92, 104, 110–113, 115, 119, 123–124, 134, 187 conductances of, 168–169 permeability, 178 phosphorylation and, 275–276 rectification of, 122 relation to Vj/fast gating, 82, 92, 110–112 See also Connexin pore - conductance; Vj/fast-gating Conductance-voltage (G-V) relations, 106–110, 120–121, 124 See also Voltage sensitivity Conducted vascular responses, see Vasculature Cones, 398–405, 411 blue, 403–404 cone–cone coupling, 401–403 rod–cone coupling, 398–399, 404– 405, 411, 413 See also Retina Connexin channel structure, 35–53, 56–64, 78–81 See also Amino-terminal domain (NT); Carboxyl-terminal domain (CT); Connexin domain organization; Connexin pore - porelining domains; Cytoplasmic loop domain (CL); Extracellular loops; Transmembrane domains Connexin consensus/signature sequence, 8 Connexin diseases - by affected organ auditory system deafness, nonsyndromic DFNA2, 16, 427 DFNB1 and DFNA3, 16, 51, 310, 426–427 deafness, syndromic, 427 deafness, peripheral neuropathy and erythrokeratoderma variabilis (EKV), 16, 310, 315, 427

Index

551 hystrix-like ichthyosis deafness (HID), 16, 310, 315 keratitis-ichthyosis deafness (KID) syndrome, 16, 310–311, 315 mutilating keratoderma with sensorineural deafness (Vohwinkel’s Syndrome), 16, 310, 314–315 palmoplantar keratoderma, 16, 310–311 sensorineural hearing loss and palmoplantar hyperkeratosis, 16, 310

bone oculodentodigital dysplasia (ODDD), 15, 16, 310–311, 338–339, 373–375, 443, 485, 489 heart visceroatrial heterotaxia, 15, 443 lens, 16, 389–393 autosomal dominant zonular pulverulent cataract-1 (CZP1), 16 autosomal dominant zonular pulverulent cataract-3 (CZP3), 16 nervous system Charcot-Marie-Tooth disease, Xlinked (CMTX), 15–16, 51–52, 227, 236, 325–328, 337–338 deafness, peripheral neuropathy and erythrokeratoderma variabilis (EKV), 16, 310, 315, 427 oculodentodigital dysplasia (ODDD), 15, 16, 310–311, 338–339, 373–375, 443, 485, 489 Pelizaeus Merzbacher-like disease (PMLD), 16, 339–340 skin skin only erythrokeratoderma variabilis (EKV), 16, 310, 315, 427 hidrotic ectodermal dysplasia (Clouston syndrome, HED), 16, 310, 367 hypotrichosis, 310–311 syndromic deafness, peripheral neuropathy and erythrokeratoderma variabilis (EKV), 16, 310, 315, 427

hystrix-like ichthyosis deafness (HID), 16, 310, 315 keratitis-ichthyosis deafness (KID) syndrome, 16, 310–311, 315 mutilating keratoderma with sensorineural deafness (Vohwinkel’s Syndrome), 16, 310, 314–315 palmoplantar keratoderma, 16, 310–311 sensorineural hearing loss and palmoplantar hyperkeratosis, 16, 310 Connexin diseases - by connexin Cx26 deafness, nonsyndromic DFNB1 and DFNA3, 16, 51, 310, 426, See also Auditory system deafness, syndromic, 427 hystrix-like ichthyosis deafness (HID), 16, 310, 315, See also Auditory system; Skin keratitis-ichthyosis deafness (KID) syndrome, 16, 310–311, 315, See also Auditory system; Skin mutilating keratoderma with sensorineural deafness (Vohwinkel’s Syndrome), 16, 310, 314–315, See also Auditory system; Skin Palmoplantar keratoderma, 16, 310–311, See also Auditory system; Skin sensorineural hearing loss and palmoplantar hyperkeratosis, 16, 310, See also Auditory system; Skin hypotrichosis, 310–311 See also Skin Cx30 deafness, nonsyndromic DFNA3, 16, 427, See also Auditory system skin disease (without deafness) hidrotic ectodermal dysplasia (Clouston syndrome, HED), 16, 310, 367, See also Skin

552 Connexin diseases - by connexin (cont.) Cx30.3 skin disease (without deafness) erythrokeratoderma variabilis (EKV), 16, 310, 315, 427, See also Skin Cx31 deafness, nonsyndromic DFNA2, 16, 427, See also Auditory system deafness, syndromic deafness, peripheral neuropathy and erythrokeratoderma variabilis (EKV), 16, 310, 315, 427, See also Auditory system; Nervous system; Skin skin disease (without deafness) erythrokeratoderma variabilis (EKV), 16, 310, 315, 427, See also Skin Cx32 Charcot-Marie-Tooth disease, X-linked (CMTX), 15–16, 51–52, 227, 236, 325–328, 337–338 See also Nervous system Cx43 oculodentodigital dysplasia (ODDD), 15, 16, 310–311, 338–339, 373–375, 443, 485, 489 See also Nervous system; Skeletal biology visceroatrial heterotaxia, 15, 443 See also Heart Cx46 autosomal dominant zonular pulverulent cataract-3 (CZP3), 16 See also Lens Cx47 Pelizaeus Merzbacher-like disease (PMLD), 16, 339–340 See also Nervous system Cx50 autosomal dominant zonular pulverulent cataract-1 (CZP1), 16 See also Lens Connexin domain organization, 7–9, 38–41, 43 See also Amino-terminal domain (NT); Carboxyl-terminal domain (CT); Cytoplasmic loop

Index domain (CL); Extracellular loops; Transmembrane domains Connexin evolutionary relationships, 9–10, 34 Connexin expression, control of, see Tissuespecific connexin promoters; Transcription of connexin genes; Translation of connexin RNA Connexin functions in specific tissues and processes, see Organ systems and processes Connexin gene family, 4–7 gene structure, 11–13 genomic localization, 9–11 polymorphisms (single nucleotide, SNPs), 18–20, 443, 472, 476–477 pseudogenes, 5–6, 16, 426 See also Connexin orthologs; Connexin subgroups Connexin interacting proteins, 266–280 cadherin, 34, 254, 269, 270, 272–273, 443, 446, 532 calmodulin, 41, 268, 276 catenin a-catenin, 269, 272 b-catenin, 13, 254, 266, 269, 272 caveolins, 250, 265–266, 269 CIP85, 268, 278 claudin, 250, 255–266, 269, 273 disk large homolog-1 (Dlgh1), 269, 279–280 drebrin, 269, 274 nephroblastoma-overexpressed family of growth regulators (NOV/ CNN3), 269, 279 and plaque dynamics, 254–255 Y1-associated factor 2 (YAF-2), 315 zonula occludens ZO-1, 135, 254–255, 269, 272–274, 276, 280 associated nucleic acid-binding protein (ZONAB), 269, 280 and plaque formation and maintenance, 254–255, 274 ZO-2, 254, 269, 273–274 See also Tight junctions Connexin-mimetic peptides, 185, 209, 217–218, 294–295 See also Pharmacology of connexin channels

Index Connexin mutants - cysteine substitution Note: as in text, D indicates deletion, * indicates chimeric protein Cx32A147C, 88 Cx32F141C, 88 Cx32F149C, 88 Cx32F235C, 338 Cx32F31C, 95 Cx32I30C, 95 Cx32I33C*43E1, 87 Cx32L144C, 88, 95 Cx32L80C, 88 Cx32L89C, 88 Cx32M34C, 95 Cx32M34C*43E1, 87 Cx32S138C, 88 Cx32S85C, 338 Cx32T86C, 121 Cx32V139C, 88 Cx32V35C, 95 Cx32V84C, 88, 194 Cx32Y135C, 88 Cx32Y151C, 88 Cx46A39C, 90 Cx46D51C, 90 Cx46E43C, 90 Cx46G46C, 90 Cx46I34C, 87 Cx46L35C, 87, 90, 93 Connexin mutants - point mutants and deletions Cx26A88S, 194, 427 Cx26D66H, 310, 314–315 Cx26G59S, 310 Cx26M34A, 47–48, 64, 80 Cx26N14K, 310 Cx26R75Q, 427 Cx26R75W, 315, 427–428 Cx26T5M, 194 Cx26V84L, 194, 427 Cx26V95M, 194, 427 Cx26W44S, 427 Cx2635G, 15, 426 Cx2642E, 315 Cx30A88V, 310 Cx30G11R, 310 Cx30T5M, 427 Cx32*43E1, 83, 85–88, 91, 105, 114–116, 118, 124, 295 Cx32E208K, 227, 236 Cx32G5D, 118 Cx32G5D*43E1, 119 Cx32L35G, 86

553 Cx32N2D, 114 Cx32N2E, 114–115, 118 Cx32N2E*43E1, 115 Cx32N2E+G5K*43E1, 119 Cx32N2R+G5D*43E1, 118–119 Cx32R142W, 338 Cx32R220, 124 Cx32S26L, 52 Cx32T8D, 118 Cx32T8D*43E1, 118 Cx32T86A, 121 Cx32T86L, 121 Cx32T86N, 121 Cx32T86S, 121 Cx32T86V, 121 Cx40E13K, 86 Cx40E9K, 86 Cx40P88S, 15, 18 Cx43D245Q, 125 Cx43G60S, 374–375 Cx43H142E, 123, 134 Cx43L263, 30, 36, 44, 46–47, 79–80 Cx43M257, 125, 131, 134, 314 Cx43R243Q, 125 Cx34R234, 125 Cx43S262A, 279 Cx43S262D, 279 Cx43S364A, 270 Cx43S364E, 270 Cx43S364P, 270 Cx43T154A, 252 Cx43Y265F, 276 Cx45, 6S363A, 278 Cx46L35C, 87 Cx50G22R, 392–393 Cx50P88S, 15 Cx50S50P, 393 Connexin orthologs, 4–7, 11, 544–545 murine orthologs, correspondence with human, 5, 179 non-murine orthologs, correspondence with human, 6 avian Cx45.6, 83–84, 114, 179, 278 avian Cx46, 34 avian Cx56, 34, 215 teleost Cx35, 155, 335 zebrafish Cx41.8, 314 See also Connexin gene family Connexin pore - charge selectivity among atomic ions, 169–175, 178–179, 181–182, 189, 192–193 measurement of, 170–171

554 Connexin pore - charge selectivity (cont.) mutation and modification affecting, 85–90 among tracer molecules, 177–181 Connexin pore - conductance rectification of, 85–87, 90, 108, 116, 122, 173 See also Current-voltage (I-V) relation, single channel; Single channel rectification substate (residual state), 82, 92, 104, 110–113, 115, 119, 123–124, 134, 187 conductances of, 168–169 permeability, 178 rectification of, 122 relation to Vj/fast gating, 82, 92, 110–112 See also Vj/fast-gating unitary conductances, 78, 92, 167–169, 171–172, 181 Connexin pore - molecular permeability biological molecules, 77, 182–194, 424–425, 521 comparison with other channels, 192–193 effects of mutation on, 427 signaling consequences, 190–192 See also Oscillatory signaling nonbiological molecules (tracers), 174–181 uncharged molecules (assessment of limiting pore width), 52, 172–174 Connexin pore - pore-lining domains, 96–99 amino-terminal domain (NT), 52–53, 63–64, 78, 83–86, 91, 97, 114–118, 121 first extracellular loop (E1), 52–53, 81, 83–87, 89–90, 96–97, 118 transmembrane domains (M1, M2, M3, M4), 50–53, 65, 79–91, 96–99 evaluation of SCAM studies of, 90–96 See also Connexin channel structure Connexin response element (CxRE), 377, 532, 534 Connexin somatic mutation, 19–20, 443 Connexin subgroups, 5–7, 9–10, 34, 54, 91, 513, 543–544 compatibility of, 34, 54, 56, 230–231, 268 See also Heterotypic specificity

Index Contingent gating, 110–111 See also Voltage-gating Corpus cavernosum, see Male reproductive system Coupling coefficient, 148–150, 152 See also Electrical synapses CryoEM, see Electron cryomicroscopy (cryoEM) Current-voltage (I-V) relations, single channel, 53, 85–86, 108, 113, 119 See also Connexin pore - conductance; Single channel rectification Note: In the listing of individual connexins below, human and non-human orthologs are grouped together; corresponding gene names are in italics; for more detailed information, see Subject entries Cx23 (CX23, Gje1, GJE1), 5–6, 8, 12 Cx25 (Gjb7), 5–6 Cx26 (CX26, Gjb2, GJB2) aminosulfonate sensitivity, 62–64, 137–138 amino-terminal (NT) domain gating model, 63–64, 124–125 antisense RNA knockdown, 343 in astrocytes, 328, 331–332, 337 in cancer, 536–537 channel structure, 80 chemical gating, pH gating, 137–138 chimera Cx26*Cx32, 11, 65 in CNS dysfunction, 337 conductance, 168 in deafness and, 16, 51, 310–311, 314–315, 336–337, 426–429 expression levels in Cx32 knockout mice, 519 in female reproductive system, 486–488 gene knockout, 313–314, 427–429, 488, 501 hemichannels, 54–55, 80, 103, 105, 108, 113–114 heteromeric with Cx32, 54–55, 189–190 in inner ear, 419, 422–426 in male reproductive system, 496, 498, 501–504 mutants Cx26A88S, 194, 427 Cx26D66H, 310, 314–315

Index Cx26G59S, 310 Cx26M34A, 47–48, 64, 80 Cx26N14K, 310 Cx26R75Q, 427 Cx26R75W, 315, 427–428 Cx26T5M, 194 Cx26V84L, 194, 427 Cx26V95M, 194, 427 Cx26W44S, 427 Cx2635G, 15, 426 Cx2642E, 315 in neurons, 336 permeability, 173, 178, 181, 188–190, 194, 217–218 heteromeric with Cx32, 54–55, 189–190 in respiratory epithelium, 362, 364–367 in secretory glands, 513–514, 516–517, 519, 532 in skin, 308, 311–312, 315–316 hypotrichosis, 310–311 hystrix-like ichthyosis deafness (HID), 16, 310, 315 keratitis-ichthyosis deafness (KID) syndrome, 16, 310–311, 315 mutilating keratoderma with sensorineural deafness (Vohwinkel’s Syndrome), 16, 310, 314–315 palmoplantar keratoderma, 16, 310–311 voltage dependence, 103, 106–108, 110, 113–114, 116, 122, 124–125 Cx29 gene knockout, 327, 428 See also Cx30.2 Cx30 (CX30, Gjb6, GJB6) in astrocytes, 328, 331–332 conductance, 168 deafness and, 16, 310, 367, 426–429 in female reproduction, 486 gene knockout, 313, 334, 427–429 hemichannels, 428 in inner ear, 419, 422–426 mutants Cx30A88V, 310 Cx30G11R, 310 Cx30T5M, 427 neuron-glia coupling, 336 in oligodendrocyte, 334 permeability, 178, 181, 194 in respiratory epithelium, 362–363, 367 in retina, 413

555 in skin, 16, 308, 310–312, 367 hidrotic ectodermal dysplasia (Clouston syndrome, HED), 16, 310, 367 skin cancer and, 311–312 Cx30.2 (Cx29, CX30.2, CX31.3, Gjc3, GJC3) in CNS, 327, 334 deafness and, 327, 428 gene knockout, 427–428, 442 gene nomenclature, 5 in heart, 440–442 hemichannels, 326 in inner ear, 423 in myelinating Schwann cell, 324–330 in oligodendrocyte, 328–330 permeability, 180 in testis, 328, 496, 498 Cx30.3 (CX30.3, Gjb4, GJB4) in airway epithelium, 363–364 in female reproductive system, 485 in male reproductive system, 496, 502 in skin, 54, 308, 311, 313 erythrokeratoderma variabilis (EKV), 16, 310, 315, 427 Cx31 (CX31, Gjb3, GJB3) in airway epithelium, 362–363 conductance, 168 deafness and, 16, 310, 315, 426–428 extracellular loop domain structure, 8 gene knockout, 313, 427–428, 501 in inner ear, 423 permeability, 181 in respiratory epithelium, 363, 366–367 in skin, 308, 313, 316 erythrokeratoderma variabilis (EKV), 16, 310, 315, 427 in testis, 498, 501 Cx31.1 (CX31.1, Gjb5, GJB5) in airway epithelium, 364 in male reproductive system, 496, 502–503 in skin, 308, 312 Cx31.3 see Cx30.2 Cx31.9 (CX31.9, Gjd3, GJD3) in atherosclerosis, 473 conductance, 168 in heart, 441 interacting proteins, 273 permeability, 179 pseudogene, 6 in vasculature, 473, 498 voltage-gating, 121

556 Cx32 (CX32, Gjb1, GJB1) in airway epithelium, 362, 363–365 in astrocytes, 328, 331–332 in cancer, 228, 311, 367, 504, 537 cancer and, 311 chemical gating, pH gating, 137–138 chimeras Cx32*Cx37, 87 Cx32*Cx43, 60 Cx32*Cx43*Cx46, 87 Cx32*Cx46, 52 Cx32*43E1, 83, 85–87, 91, 105, 114–115, 295 in CNS, 332, 334, 336, 341 conductance, 168–169 cysteine substitution Cx32A147C, 88 Cx32F141C, 88 Cx32F149C, 88 Cx32F235C, 338 Cx32F31C, 95 Cx32I30C, 95 Cx32I33C*43E1, 87 Cx32L144C, 88, 95 Cx32L80C, 88 Cx32L89C, 88 Cx32M34C, 95 Cx32M34C*43E1, 87 Cx32S138C, 88 Cx32S85C, 338 Cx32T86C, 121 Cx32V139C, 88 Cx32V35C, 95 Cx32V84C, 88, 194 Cx32Y135C, 88 Cx32Y151C, 88 in female reproductive system, 483–485, 487–488 gap junctions, 226–232, 273 gating movements, 137–138 gene knockout, 327, 329, 332, 334, 338, 341, 343, 367, 484, 501, 517–519 in heart, 442 hemichannels, 18, 54–55, 105 in hypertension, 523 interacting proteins, 265–266, 273, 276, 280 in male reproductive system, 228–229, 498, 501–504 mutants Cx32*43E1, 83, 85–88, 91, 105, 114–116, 118, 124, 295 Cx32E208K, 227, 236

Index Cx32G5D, 118 Cx32G5D*43E1, 119 Cx32L35G, 86 Cx32N2D, 114 Cx32N2E, 114–115, 118 Cx32N2E*43E1, 115 Cx32N2E+G5K*43E1, 119 Cx32N2R+G5D*43E1, 118–119 Cx32R142W, 338 Cx32R220, 124 Cx32S26L, 52 Cx32T8D, 118 Cx32T8D*43E1, 118 Cx32T86A, 121 Cx32T86L, 121 Cx32T86N, 121 Cx32T86S, 121 Cx32T86V, 121 in ovarian follicles, 483 permeability, 173–174, 179–181, 189–190, 193, 216–217 heteromeric with Cx26, 54–55, 189–190 pharmacology, 214, 216–217, 295 phosphorylation, 277 in PNS, 324–327 Charcot-Marie-Tooth disease, X-linked (CMTX), 15–16, 51–52, 227, 236, 325–328, 337–338 SCAM, 52, 87–88, 91, 95–96 in secretory glands, 513, 516–519, 523 in skin, 308 trafficking, 229, 231, 267, 338 transcriptional regulation, 13, 14, 513 voltage dependence, 103, 106–108, 110, 113–114, 116–120, 124 Cx33 (Gja6) dominant-negative effects on trafficking, 498 gene knockout, 501 in male reproductive system, 496, 498 orthologous proteins, 5–6 Cx34.7, 335 Cx35, see Cx36 Cx36 (CX36, Cx35, Gjd2, GJD2) autoimmune disease, 523 chemical gating, pH gating, 156 conductance, 169 gene knockout, 152, 154, 156–157, 341, 405, 411, 517–518, 520, 523 gene structure, 12 interacting proteins, 80, 250, 276 in neurons, 153, 335–336, 341–342

Index permeability, 178, 181, 190, 516 pharmacology, 212, 214, 341 phosphorylation, 155, 264, 265, 267, 271, 410 in retina, 154, 401–405, 409–412 schizophrenia, 15 in secretory glands, 513–514, 519–524 trafficking, 267, 271 transcription, 14, 513 voltage sensitivity, 154 Cx37 (CX37, Gja4, GJA4) in airway epithelium, 362 in atherosclerosis, 13, 19, 472–474, 476–477 conductance, 169 in female reproductive system, 483–484 gene knockout, 404, 464, 476, 484, 501 in male reproductive system, 496, 501 in neurons, 153, 336 permeability, 171, 173–174, 179, 181 pharmacology, 217 in secretory glands, 516 in skin, 308 in vasculature, 463–464 voltage dependence, 110, 114 Cx40 (CX40, Cx41.8, Gja5, GJA5) alternate gene splicing, 13 in atherosclerosis, 472–474, 477 channel structure, 86 chemical gating, pH gating, 137 chimera Cx40*Cx43, 60 conductance, 169 in female reproductive system, 489 gene knockout, 271, 373–374, 442, 464, 477, 489, 501, 523 tissue restricted, 477 in heart, 14, 19, 436–443, 446–449 in hypertension, 464, 477, 516, 523 interacting proteins, 279 in male reproductive system, 496, 502 mutants Cx40E13K, 86 Cx40E9K, 86 Cx40P88S, 15, 18 in neurons, 153, 335 permeability, 171, 173, 179, 181 pharmacology, 86, 211–212, 214, 216–218, 366 phosphorylation, 277 in respiratory epithelium, 362, 364, 366 in secretory glands, 516

557 nonsecretory functions of glands and, 523 in skeletal biology, 372–374 Holt-Oram syndrome, 14, 373 in skin, 308, 314 in testis, 498, 501–502 trafficking, 267, 271 transcription, 10, 14, 19, 373, 446 promotor polymorphism, 19 in vasculature, 463, 464, 472–474, 477, 523 Cx40.1 (CX40.1, Cx39, Gjd4, GJD4), 5, 122, 124 Cx41.8, 314 See also Cx40 Cx43 (CX43, Gja1, GJA1) alternate gene splicing, 13 carboxyl terminal (CT) domain, 92, 123, 130–136 cell migration, 272, 442, 500, 535–536 channel structure, 42, 44–50, 62, 65, 78–80, 86, 92 chemical gating, pH gating, 131, 137 conductance, 169 cytoplasmic loop (CL) domain, 121, 123, 130–136 deafness and, 426 disease oculodentodigital dysplasia (ODDD), 15, 16, 310–311, 338–339, 373–375, 443, 485, 489 visceroatrial heterotaxia, 15, 443 erectile dysfunction, 504 extracellular loop domain structure, 59–60 in astrocytes, 328, 331–332, 334, 535 in atherosclerosis, 472, 474–476 in cancer, 502, 504, 531–537 in CNS dysfunction, 337 in female reproductive system, 13, 14, 481, 483–489 gap junction biogenesis, 227–229, 231–232, 234–236 gene knock-in Cx43KI32, 442, 501, 523 Cx43KI40, 442, 501 gene knockout, 293–294, 312, 334, 339, 343, 372, 375, 376–377, 389, 442, 447–448, 464–465, 475, 481, 484, 488–489, 500–501 tissue restricted, 339, 343, 372, 374–375, 442, 447–448, 465, 481, 489, 501 in heart, 15, 272, 275, 344, 436–451

558 Cx43 (CX43, Gja1, GJA1) (cont.) hemichannels, 80, 92 assembly, 228–230 bisphosphonate opening, 376 nonjunctional/unapposed, 444 heterotypic compatibility, 60–61 in hypertension, 465, 523 interacting proteins, 135, 250, 254–255, 265–266, 268, 273–276, 278–280, 391, 446 junctional morphology, 246, 250–255 in lens, 387, 389, 391, 393 in male reproductive system, 496–501, 503–504, 536 mechanotransduction, 378 metastasis, 537 miRNA and, 15 mitosis, 278 mutants Cx43D245Q, 125 Cx43G60S, 374, 375 Cx43H142E, 123, 134 Cx43L263, 30, 36, 44, 46–47, 79–80 Cx43M257, 125, 131, 134, 314 Cx43R234, 125 Cx43R243Q, 125 Cx43S262A, 279 Cx43S262D, 279 Cx43S364A, 270 Cx43S364E, 270 Cx43S364P, 270 Cx43T154A, 252 Cx43Y265F, 276 myoblast differentiation and, 15, 279 in neurons, 279, 336 nuclear localization of, 533 in oligodendrocyte, 334 permeability, 171, 173–174, 180–181, 187–190 pharmacology, 86, 135–136, 210–212, 216–218, 499 phosphorylation, 210, 264–265, 267, 270, 274–278, 391–392, 444–445, 498 MAPK activity and, 277 src activity and, 135, 250, 276, 376 posttranslational modification, 187, 236, 264 pseudogene, 6, 16, 426 in respiratory epithelium, 362, 364–366 in retina, 413 SCAM, 95 in secretory glands, 513–514, 516, 523

Index in skeletal biology, 310, 372–378, 532 in skin, 308, 310–312, 314 trafficking, 13, 229, 231, 251–252, 255, 265, 267, 270–272, 277, 373, 498, 532 transcriptional regulation, 13, 14, 376–377, 481, 486, 503, 532 transmembrane replacement by polyalanine, 51 in vasculature, 374, 463–465, 523 atherosclerosis, 469, 472, 474–476 conditional deletion with TIE–2, 366, 465 voltage-gating, 80, 92, 110, 112, 114, 116, 121–125, 130–136, 187 Cx44, see Cx46 Cx45 (CX45, Gjc1, GJC1) alternate RNA splicing, 13 in atherosclerosis, 473 chemical gating, pH gating, 137 conductance, 169 in female reproductive system, 484, 489 gene knock-in Cx45KI40, 442 gene knockout, 153, 411, 442, 464, 484 tissue restricted, 153, 442 gene structure, 13 in heart, 436–437, 439–442, 441, 446, 449 interacting proteins, 273 in male reproductive system, 496 mutants Cx45.6S363A, 278 in neurons, 153–154, 335–336 in oligodendrocytes, 331 permeability, 171, 178, 181, 189 phosphorylation, 271, 278 in retina, 153, 402, 405, 411–412 in skeletal biology, 372–373, 376–377 in skin, 308 trafficking, 270–271 in vasculature, 463–464 voltage dependence, 154 Cx45.6, see Cx50 Cx46 (CX46, Cx44, Cx56, Gja3, GJA3) cataracts, 16, 389, 392 chemical gating, pH gating, 137 chimera Cx46*Cx50, 59–60 conductance, 169 cysteine substitution Cx46A39C, 90 Cx46D51C, 90

Index Cx46E43C, 90 Cx46G46C, 90 Cx46I34C, 87 Cx46L35C, 87, 90, 93 gene knockout, 389–390, 501 hemichannels loop/slow gating in, 111, 209 mechanosensitive nature, 120, 295 nonjunctional/unapposed, 105 open channel rectification, 85 heterotypic compatibility, 61 interacting proteins, 250, 269, 273, 280 in lens, 16, 389–390, 392 in male reproductive system, 496 mutants Cx46L35C, 87 permeability, 85, 94, 171, 173, 178, 181 pharmacology, 212, 215 in respiratory epithelium, 364–365 SCAM, 52, 56, 59, 87–99 in skeletal biology, 372, 374 trafficking, 231, 251, 374 voltage-gating, 87, 93, 95–96, 98, 105, 123–124 Cx47 (CX47, Gjc2, GJC2) in CNS, 328, 332, 334, 336 conductance, 169 gene knockout, 334, 501 Pelizaeus Merzbacher-like disease (PMLD), 16, 339–340 in PNS, 327 Cx50 (CX50, Cx45.6, Gja8, GJA8) cataract, 16, 389 chemical gating, pH gating, 138 conductance, 169 gene knock-in Cx50KI46, 389–391, 393 gene knockout, 389–391, 501 hemichannels freeze-fracture EM, 254 heteromeric compatibility, 54 heterotypic compatibility, 60 K+ effects on hemichannel versus junctional channel, 92 interacting proteins, 250, 273, 276 in lens, 387, 389, 391–393 in male reproductive system, 498 mutants Cx50G22R, 392–393 Cx50P88S, 15 Cx50S50P, 393 in neurons, 153, 336

559 permeability, 171, 181 pharmacology, 210, 212, 214, 215 phosphorylation, 278 in retina, 153, 335, 402, 405, 407 schizophrenia, 15 voltage-gating, 87, 93, 95–96, 98, 105, 110, 114 Cx55.5, 408 Cx56, see Cx46 Cx57 gene knockout, 407 See also CX62 CX59 (GJA9) orthologs, 5, 6 CX62 (Cx57, Gja10, GJA10) conductance, 169 permeability, 178 Cyclic adenosine monophosphate (cAMP) permeability, 184, 186–191, 377, 458 effects on gap junction communication, 267–278, 375, 410, 446, 498, 533 Cyclodextrins (CDs), 173, 216–217 See also Pharmacology of connexin channels Cytoplasmic bridges, 146, 176 Cytoplasmic loop (CL) domain, 6–9, 38–39, 50, 64–65 and calmodulin binding, 276 in chemical gating, 64, 123–124, 131–133, 268–269 See also Chemical gating; Particlereceptor hypothesis CL peptide effects, 123, 134, 218 interaction with carboxyl-terminal (CT) domain, 64, 123, 131, 133 interaction with Dlgh1, 269, 279 interaction with RXP-E peptide, 135 phosphorylation, 265, 271 and residual state, 134 solution structure, 132 in Vj gating, 123–124 See also Connexin domain organization Cytoplasmic transduction peptide (CTP), 136 Cytoskeleton, see Actin; Connexin interacting proteins; Microtubules Cytosolic stress, 228–229, 233, 235, 251 D Deafness, see Auditory system; Connexin diseases Deiters’ cells, 420, 424–425 See also Auditory system

560 Detergent solubility, 55, 233, 265–266, 274–275 See also Gap junction plaque biogenesis, maintenance and degradation Diabetes, 504, 520, 522–524 4,6-Diamidino-2-phenyl-indole dihydrochloride (DAPI), see Tracers Dicholorofluorescein, see Tracers Disk large homolog-1 (Dlgh1), see Connexin interacting proteins Docking, hemichannel, see Extracellular loops; Hemichannels; Heterotypic specificity Dominant-negative interactions and mutations, 18, 315, 339, 342, 373–375, 428, 498, 537 Drebrin, see Connexin interacting proteins Dye coupling, cautions and pitfalls, 147–148, 155, 175–176 Dystrophia myotonica protein kinase (DMPK), 266 E Ear, see Auditory system Electrical coupling, see Electrical synapses Electrical synapses, 143–158, 335–336, 342 cytoplasmic bridges and, 146 detection, 145–147 differences from chemical synapses, 146–147, 157 ephaptic coupling, 147, 408 functions in nervous system, 156–157 mixed chemical and electrical synapses, 155, 335 neuronal connexins, 153–154 neurotransmitter modulation of, 154–156 by g-aminobutyric acid (GABA), 342 by dopamine, 155–156, 398, 410 by glutamate, 155 by nitric oxide (NO), 156, 410 by serotonin, 155 properties coupling coefficient, 148–150, 152 low-pass filtering and, 150–151 subthreshold synchrony, 152 suprathreshold synchrony, 152 rectifying gap junctions, 55, 146, 149, 411, 424 See also Nervous system Electron cryomicroscopy (cryoEM) 2D projection map of connexin channels, high resolution, 36, 43–44, 80

Index 3D density map of connexin channels, high resolution, 46–49, 64, 78–80, 98 See also Connexin channel structure Electron microscopy and image analysis, 35–39 2D projection map of connexin channels, low resolution, 36–37, 62 3D density map of connexin channels, low resolution, 37–39 negative stain versus frozen-hydrated, 37–38 See also Connexin channel structure Emergent organ properties, 305–306 Endocrine glands, see Secretory glands Endometrium, 485–488 See also Female reproductive system Endothelium-derived hyperpolarizing factor (EDHF), 459, 464 See also Vasculature Ephaptic coupling, 147, 408 Epilepsy and seizure, 334, 340–342 See also Nervous system ER-associated degradation (ERAD), see Gap junction plaque biogenesis, maintenance and degradation Erectile dysfunction, 504–505 Erythrokeratodermia variabilis (EKV), 16, 310, 315, 427 See also Connexin diseases; Skin Estrogen, 481, 486–489, 503, 536 Ethidium bromide, see Tracers Exocrine glands, see Secretory glands Extracellular loops (E1, E2), 7–8, 40, 50, 65 conserved cysteine residues in, 6, 8, 56 E1 and contribution to pore, 81, 96–97 from domain swap studies, 52 from open channel rectification and selectivity studies, 52–53, 85–86 from SCAM studies, 89–90 from unitary conductance studies, 86–87 from voltage-gating studies, 83–85, 118 flexibility of, 62 and gating, 83, 111 See also Loop/slow-gating hemichannel docking, 18, 56–59 E2 and specificity of heterotypic docking, 59–61 See also Heterotypic specificity

Index peptides corresponding to, see Connexinmimetic peptides b-sheet structure of, 49, 57–61 See also Connexin domain organization Extracellular signal regulated kinase (ERK), 268, 276–277, 376–377, 391–392, 532, 534 F Fast gating, see Vj/fast-gating Fatty acid amide, see Oleamide Female reproductive system, 481 connexins in, 483–489 ovary and oogenesis, 482–485 corpora lutea, 484–485 uterus endometrium, 485–488 implantation, 486–488 myometrium, 13, 488–489 Fenamates, 109, 213–214, 294 See also Pharmacology of connexin channels Fertility, 485, 489, 501–503 See also Female Reproductive System; Male Reproductive System Fibroblast growth factor (FGF), 391–392 First extracellular loop domain (E1), see Extracellular loops First transmembrane domain (M1), see Transmembrane domains Flufenamic acid, see Fenamates Fluorescence/Fo¨rster resonance energy transfer (FRET), 187, 246 Fluorescence recovery after photobleaching (FRAP), 175, 232, 251–252 See also Gap-FRAP technique Fourth transmembrane domain (M4), see Transmembrane domains Freeze-fracture of gap junctions, 29–30, 232, 236, 242–245, 248–249, 254, 273, 280 Freeze-fracture replica immunolabeling (FRIL), 243–244, 249, 325–326, 328, 331, 335–336 G g-Group connexins, see Connexin subgroups Ganglion cells, retinal, 399–402, 405, 408, 412–413 See also Retina Gap-FRAP technique, 175, 516 Gap junction plaque biogenesis, maintenance and degradation, 225–237, 250–255, 266–278

561 channel assembly, quality control and trafficking to plasma membrane connexin insertion at ER, 226 effects of cytosolic stress, 228–229, 233, 235, 251 ER-associated degradation (ERAD), 227–229, 235–236, 267 Golgi apparatus and quality control, 267 hemichannel assembly site and mechanism, 229–231 proteasomal quality control, 17–18, 227, 229, 264, 267, 277–278 trans-Golgi network (TGN) and oligomerization, 228–230, 254 transport to cell surface, 231–232, 270–271 ubiquitination, 227–228, 269, 533 degradation, 233–235, 252, 277–278 internalization, 234, 277–279 annular gap junctions, 234, 252 lysosomal degradation, 17, 234–235, 252, 277–278 proteasomal degradation issues, 234–235, 264 dynamic aspects connexin-binding proteins and plaque dynamics, 254–255 half-life of gap junctional plaques, 233, 251, 265 junction assembly and degradation among connexins, differences, 236 plaque formation, maintenance and growth, 251–253 mechanisms of plaque growth directed insertion, 232 lateral diffusion, 232, 252–253 plaque formation and maintenance, 253, 272–275 plaque nucleation, 236 Gap junction plaques biochemical isolation and characterization, 31–33 biogenesis, see Gap junction plaque biogenesis, maintenance and degradation lateral mobility of channels within plaques and plasma membrane, 250–253 lipid composition of, 247–250

562 Gap junction plaques (cont.) mixing of connexin isoforms within plaques, 55, 245–247 organization and packing, 241–245 split, 39, 46, 54–55, 80, 243, 247 structure of, 28–35 Gating calmodulin interaction and, 276 chemical gating, see Chemical gating conformational changes of amino-terminal (NT) domain based, 63–64, 124–126 Ca2+-induced closure, 62–64 carboxyl-terminal (CT) domain, 123–125 particle-receptor mechanism, 123, 130–136 proline kink in M2, 104, 121–123 protein kinase C (PKC), and, 275 src interaction and, 276 voltage-gating, see Loop/slow-gating; Vi-o (inside-out voltage) gating; Vj/fast-gating; Voltage sensitivity General anesthetics ethrane, 208 halothane, 168, 208–209, 215–216 isoflurane, 215 See also Pharmacology of connexin channels Genodermatoses (inherited skin diseases), see Connexin diseases; Skin Glutathione, 188, 190, 228 Glycosylation, 289 of connexins, absence of, 226, 264, 289 of pannexins, 288–289 Glycyrretinic acid and derivatives, pharmacology of, 210–211 carbenoxolone, 185, 210–211, 294, 408 glycyrrhetinic acid, 185, 210–211 glycyrrhizic acid, 210 See also Pharmacology of connexin channels Golgi apparatus and quality control, 267 trans-Golgi network (TGN) and oligomerization, 228–230, 254 See also Gap junction plaque biogenesis, maintenance and degradation Growth factors fibroblast growth factor (FGF), 391–392 insulin-like growth factor (IGF), 131, 391–392, 534

Index lens epithelial derived growth factor (LEDGF), 391–392 H Halothane, see General anesthetics; Pharmacology of connexin channels Harmaline, 342 Heart, 435–450 acquired adult heart disease, 443–448 arrhythmia, 14–15, 19, 130, 136, 442–448 Cx40 and, 14, 19, 442–443, 448 Cx43 and, 15, 442, 444–445, 447–448 atrial fibrillation, 19, 443, 448–449 atrioventricular (AV) node, 436, 440–442 cardiac development, 441–442 cardiomyocytes, 136, 216, 228, 270, 272, 276, 279, 436–442, 444–445, 488 conduction system, 440–441 connexin knock-ins, 442 connexin knockouts, 442, 447–448 gap junction remodeling, 443–450 ischemia, 130, 215, 275–276, 443–447 sinoatrial (SA) node, 436, 440 transgenic mice, 442–444, 446–447 ventricular fibrillation, 130 ventricular heart disease, 444–447 visceroatrial heterotaxia, 15, 443 Hemichannels AFM of, 62–63, 80, 138, 245 assembly, 229–231 See also Gap junction plaque biogenesis, maintenance and degradation ATP release channel, pannexins and, 184–185, 291–295 boundaries of connexins within, 50 conductance of, 168–169 conductive, 105 Cx32*43E1 hemichannel, 83 See also Connexin mutants - point mutations and deletions definition, v–vi, 29 dimensions of, 38, 46–48 in disease and pathology, 338, 344, 367, 428, 444 docking of, 18 docking structure, 56–62 rotational stagger in junctional channels, 44–46

Index versus gap junction channel properties and structure, 53, 81, 91–92 heteromeric, 55–56, 246–247 definition, 53–54 and dominant-negative mutation, 18 heteromeric Cx26/Cx32, 54–55, 189–190 molecular selectivity of, 78, 188–190 homomeric definition, 53–54, 188, 190 intrinsic chemical and aminosulfonate gating of, 137–138 See also Aminosulfonate; Chemical gating intrinsic voltage-gating of, 83, 106–114, 118–120 See also Vi-o (inside-out voltage) gating; Vj/fast-gating; Voltage sensitivity molecular permeability, 188, 190 nonjunctional/unapposed, 18, 31, 231–232, 250–251, 253–255, 274, 291, 326, 444 pharmacology of, 210–216, 218 pore width, 173 See also Connexin mutants - point mutants and deletions Hensen’s cells, 425 See also Auditory system Heterotypic junctional channels in auditory system, 424, 427 definition, 53–54 in nervous system, 153–154, 326, 331–332, 334, 336, 340 permeability of, 188–190 rectifying conductance of, 108 in retina, 401–402, 405, 410–411 in vasculature, 464 voltage-dependence of, 106–111, 113–114 Heterotypic specificity and determinants of, 18, 41, 53–56, 59–61, 268, 333 See also Connexin subgroups Hidrotic ectodermal dysplasia (Clouston Syndrome, HED), 16, 310, 367 See also Connexin diseases; Skin Hippocampus, 153, 157, 331, 334–335, 343 Holt-Oram syndrome, 14, 373 Homotypic junctional channels definition, 53–56 permeability of, 179–181, 186–190 voltage-dependence of, 106–110 Horizontal cells, 153, 156, 336, 399–408

563 See also Retina Hydropathy analysis, 39–40 See also Connexin domain structure Hydroxycoumarin carboxylic acid (HCCA), see Tracers Hypercholesterolemia, 470–471, 475 See also Atherosclerosis; Cholesterol Hypertension, 464–465, 477, 523 See also Vasculature Hystrix-like ichthyosis deafness (HID), 16, 310, 315 See also Auditory system; Connexin diseases; Skin I Image analysis, see Electron microscopy and image analysis Immune system, 188, 251, 310, 314, 343, 360, 366–368, 469–470, 472, 475–476, 523 Implantation, see Female reproductive system Inferior olive (IO) coupling, 152 See also Tremor Inflammation in atherosclerosis, 469–470, 472, 475 in nervous system, 343 in respiratory epithelium, 360, 366–368 in skin, 310, 314 Inner ear, see Auditory system Innexins, 4, 138, 154, 287–288 See also Pannexins Inositol triphosphate (IP3), 183, 187, 189–191 permeability, 191 See also Connexin pore - molecular permeability Inside-out voltage, see Vi-o (inside-out voltage) Insulin-like growth factor (IGF), 131, 391–392, 534 Interfering RNA (RNAi), permeation of, 77, 187–188 Internal ribosome entry site (IRES), 14–15 See also Translation of connexin RNA Invertebrate gap junctions, 4, 138, 144, 149, 154, 288, 295, 297 Ischemia cardiac, 130, 215, 275–276, 443–447 nervous system, 215, 274, 343–344 Islet cells, pancreatic, 514–515, 517–518, 520–521, 524 See also Diabetes; Secretory glands

564 Isoflurane, see General anesthetics; Pharmacology of connexin channels K+ K buffering, 327–328, 334, 413, 423, 424, 428 Keratitis-ichthyosis-deafness (KID) syndrome, 16, 310, 315 See also Auditory system; Connexin diseases; Skin Kidney cancer, 516, 536–537 hypertension, 464–465, 477, 523 Knock-ins Cx43KI32, 442, 501, 523 Cx43KI40, 442, 501 Cx45KI40, 442 Cx50KI46, 389–391, 393 Knockouts Cx26KO, 313–314, 427–429, 488, 501 Cx29KO, 327, 428 Cx30KO, 313, 334, 427–429 Cx30.2KO, 442 Cx31KO, 313, 427–428, 501 Cx32KO, 327, 329, 332, 334, 338, 341, 343, 367, 484, 501, 517–519 Cx33KO, 501 Cx36KO, 152, 154, 156–157, 341–405, 411, 517–518, 520, 523 Cx37KO, 404, 464, 476, 484, 501 Cx40KO, 271, 373–374, 442, 464, 477, 489, 501, 523 Cx43KO, 293–294, 312, 334, 339, 343, 372, 375, 376–377, 389, 442, 447–448, 464–465, 475, 481, 484, 488–489, 500–501 Cx45KO, 153, 411, 442, 464, 484 Cx46KO, 389–390, 501 Cx47KO, 334, 501 Cx50KO, 389–391, 501 Cx57KO, 407 tissue-restricted Cx40, 477 Cx43, 339, 343, 372, 374–375, 442, 447–448, 465, 481, 489, 501 Cx45, 153, 442 L LDL receptor (LDLR)-deficient mice (LDLR / ), see Atherosclerosis Lens, 387–393

Index anatomy, physiology and connexin distribution, 387–389 Ca2+ homeostasis in, 389–390 calpain protease, 389–391 cataracts and connexins, 16, 389–393 See also Connexin diseases connexin mutations, 392–393 development, 391–392 lens epithelial derived growth factor (LEDGF), 391–392 Limb development, see Skeletal biology Lipid rafts, 249–250, 265–266, 269 See also Gap junction plaque biogenesis, maintenance and degradation Live cell imaging of connexins, 246, 250–253, 272 Liver cancer, 271, 531, 537 gap junction structure, 28–35, 37–38 liver Cx32 promoter, 11 mixing of connexin isoforms within plaques, 37, 55, 243, 245–247 plaque organization, 241–245 regeneration, 532 tissue-specific promoters, 11 Loop/slow-gating, 82–83, 111–113, 115, 120, 122–123, 125, 209 and amino-terminal (NT) domain model of gating, 125 conductance transitions of, 111–113 and connexin channel inhibitors, 209 definition, 82–83, 111–112 distinguished from Vj/fast-gating, 113, 115, 123 mechanisms and determinants of, 120 polarity of, 113 shifts in voltage dependence of, 122 See also Gating; Vj/fast-gating; Voltage sensitivity Low-density lipoprotein (LDL)-cholesterol, 470 See also Atherosclerosis Low-pass filtering, see Electrical synapses Lucifer yellow (LY), see Tracers Lung, see Respiratory epithelium Lysosomes, see Gap junction plaque biogenesis, maintenance and degradation M Macrophages, 472–476 See also Atherosclerosis

Index Maleimidobutyryl biocytin (MBB), 87–88, 92–93, 95 See also SCAM Male reproductive system, 495–505 corpus cavernosum, 504 effects of connexin knockouts, 501 epididymis, 502–503 prostate, 503–504 seminal vesicles, 502–503 spermatogenesis, 495, 499–502 testis control of connexin expression, 498–499 interstitial compartment, 496 pathophysiological aspects, 502 role of connexins, 499–501 seminiferous epithelium, 497–498 Mammary gland, see Secretory glands Mechanosensitivity of connexin channels, 120, 295 Mechanotransduction in bone, 378–380 See also Skeletal Biology Mefloquine, see Antimalarial drugs; Pharmacology of connexin channels Metastasis, see Cancer Methanethiosulfonate (MTS), see SCAM MicroRNAs (miRNAs), 15 Microtubules, 255, 269–270, 273 Mimetic peptides connexin-mimetic peptides, 185, 209, 217–218, 294–295 pannexin-mimetic peptides, 295 Mist1, see Transcription factors Mitogen-activated protein kinase (MAPK), 264, 266, 268, 276–277, 279, 391–392, 532, 534 Mitosis, 251, 268, 278, 391 Molecular permeability, see Connexin pore - molecular permeability Mu¨ller cells, 412–413 See also Retina Mutations, see Connexin mutations Mutilating keratoderma with sensorineural deafness (Vohwinkel’s syndrome), 310, 314–315 See also Auditory system; Connexin diseases Myoblasts, 269, 279 Myocytes, see Cardiomyocytes Myoendothelial gap junctions (MEGJs), 458–459, 462–465 See also Vasculature

565 Myometrium, 13, 488–489 See also Female reproductive system N Nephroblastoma-overexpressed family of growth regulators (NOV/ CNN3), see Connexin interacting proteins Nervous system, 143–158, 323–345 adenosine triphosphate (ATP) release by glia, 185, 379, 413 astrocytes, 55, 153, 244, 328–335, 337, 343, 413 astrocyte-astrocyte (A-A) coupling, 55, 328, 331, 334–335 astrocyte-oligodendrocyte (A-O) coupling, 328, 332, 334, 340 Ca2+ waves in glia, 344 chemical synapses, 146–147, 157 differences from electrical synapses, 146–147, 157 mixed chemical and electrical synapses, 155, 335 CNS neurons, 143–148, 153–156, 335–336, 399–413 connexin diseases of Charcot-Marie-Tooth disease, Xlinked (CMTX), 15–16, 51–52, 227, 236, 325–328, 337–338 deafness, peripheral neuropathy and erythrokeratoderma viaiabilis (EKV), 16, 310, 315, 427, See also Auditory system; Skin oculodentodigital dysplasia (ODDD), 15–16, 310–311, 338–339, 373–375, 443, 485, 489, See also Skeletal biology Pelizaeus Merzbacher-like disease (PMLD), 16, 339–340, See also Connexin diseases electrical synapses, see Electrical synapses epilepsy and seizure, 334, 340–342 hippocampus, 153, 157, 331, 334–335, 343 inferior olive (IO), 152 See also tremor (below) inflammation in, 343 ischemia, 215, 274, 343–344 neurotransmitter modulation of coupling, 154–156 by g-aminobutyric acid (GABA), 342 by dopamine, 155–156, 398, 410 by glutamate, 155

566 Nervous system (cont.) by nitric oxide (NO), 156, 410, 424 by serotonin, 155 oligodendrocytes, 153, 327–329, 331, 334, 336, 340–341 astrocyte-oligodendrocyte (A-O) coupling, 332 oligodendrocyte-oligodendrocyte (O-O) coupling, 55, 332, 340 pannexins in, 154, 289, 296 reflexive coupling in glia, 326–327, 332, 334 Schmidt-Lanterman (S-L) incisures, 325–327, 329–330 Schwann cells, 11, 14–15, 324–328, 334, 336–338 myelinating, 325–328 nonmyelinating, 324 tissue-specific promoters, 11–12, 14 tremor, 342, 344 N-ethyl-N-nitrosourea mutagenesis of Cx43, 373 Neurobiotin (NB), see Tracers Nitric oxide (NO) neuronal gap junctions and, 156, 410 vasculature and, 460–461, 464–465 N, N, N-trimethyl-2-[methyl-(7-nitro-2, 1, 3-benzoxadiazol-4yl)amino]ethanaminium (NBDT-M-TMA, NMT), see Tracers Nomenclature for connexin genes and proteins, vi, 4–7, 9–10, 33, 543–545 See also Connexin subgroups Nonbiological molecules (tracers), permeability of, see Connexin pore - molecular permeability Nuclear magnetic resonance (NMR) studies of Cx43 CT, 63–64, 83–84, 91, 98, 116, 131–135, 276 See also Carboxyl-terminal (CT) domain O Oculodentodigital dysplasia (ODDD), 15–16, 310–311, 338–339, 373–375, 443, 485, 489 See also Connexin diseases; Nervous system; Skeletal biology Oleamide (cis-9-octadecenamide; fatty acid amide), 209, 215–216, 294 See also Pharmacology of connexin channels

Index Oleic acid, 216 See also Pharmacology of connexin channels Oligodendrocytes, 153, 327–329, 331, 334, 336, 340–341 astrocyte-oligodendrocyte (A-O) coupling, 332 oligodendrocyte-oligodendrocyte (O-O) coupling, 55, 332, 340 See also Nervous system OmpF, 193 Organ systems and processes atherosclerosis, 469–477 carcinogenesis, 529–538 female reproductive system, 481–490 heart, 435–451 inner ear, 419–429 lens, 387–393 male reproductive system, 495–505 nervous system, 323–345 overview, 305–306 respiratory epithelium, 359–368 retina, 397–414 secretion, 511–524 skeletal biology, 371–381 skin, 307–316 vasculature, 457–465 Oscillatory signaling, 191–192, 520–521, 523 Ovary, see Female reproductive system P Palmoplantar hyperkeratosis, 16, 310 See also Auditory system; Connexin diseases; Skin Palmoplantar keratoderma, 16, 310–311 See also Auditory system; Connexin diseases; Skin Pancreas, see Secretory glands Pannexin-mimetic peptides, 295 Pannexins (pannexin1, Panx1, unless noted), 287–297 ATP release channel/Ca2+ waves and, 185, 291–295 conductance and pore width, 290–291 definition, 4, 154, 288–289 dye uptake, 31 gap junction channel formation, evidence concerning, 184, 288–289 gene name, 543 in gustatory epithelium, 293 interaction with purinergic receptors, 291–292, 296–297 membrane distribution, 289

Index mimetic peptides, 295 in neurons, 154, 289, 296 pannexin2 (Panx2), 288–289 pannexin3 (Panx3), 288–289 permeability, 154, 210 pharmacology of, 185, 210, 218, 294–295 Parathyroid hormone (PTH), 375–376, 378 See also Skeletal biology Particle-receptor hypothesis, 123, 130–136 and Vj/fast gating, 123–124 See also Chemical gating PDZ binding motif and interactions, 135, 254–255, 269, 273–274, 276, 279–280 See also Carboxyl-terminal (CT) domain; Connexin interacting proteins Pelizaeus Merzbacher-like disease (PMLD), 16, 339–340 See also Connexin diseases; Nervous system Peptides amino terminal (NT) domain, 83, 103, 116, 276 carboxyl-terminal (CT) domain, 131–133, 135, 276 cell-penetrating peptide (CPP), 136 cytoplasmic loop (CL) domain, 123, 134, 218 cytoplasmic transduction peptide (CTP), 136 extracellular loop (E1, E2) domain connexin-mimetic peptides, 185, 209, 217–218, 294–295, 464 pannexin-mimetic peptides, 295 peptide-based pharmaceuticals, 136 RXP-E chemical gating (pH gating) and, of Cx43, 135–136 Permeability, see Connexin pore - charge selectivity; Connexin pore conductance; Connexin pore molecular permeability Pharmacology of connexin channels, 207–219 2-aminophenoxyborate (2-APB), 208, 214, 217 alkanols, 91–92, 209, 214–216 heptanol, 185, 209, 215 hexanol, 215 octanol, 135, 185, 209, 214–215, 294 antimalarial drugs, 211–213 benzylquininium, 212

567 mefloquine, 92, 185, 208–209, 212–213 quinine, 92, 208–209, 211–212 connexin-mimetic peptides, 185, 209, 217–218, 294–295 cyclodextrins (CDs), 173, 216–217 fenamates, 109, 213–214, 294 general anesthetics, 215–216 ethrane, 208 halothane, 168, 208–209, 215–216 isoflurane, 215 glycyrrhetinic acid and derivatives, 210–211 carbenoxolone, 185, 210–211, 294, 408 glycyrrhetinic acid, 185, 210–211 glycyrrhizic acid, 210 oleamide (cis-9-octadecenamide; fatty acid amide), 209, 215–216, 294 oleic acid, 216 pannexin-mimetic peptides, 295 polyamines, 53, 86, 209, 211 tetraalkylammonium ions (TAA+), 208–209, 211 See also Chemical gating pH gating, see Chemical gating Phospholipids composition of gap junction plaques, 247–250 lipid rafts, 249–250, 265–266, 269 See also Gap junction plaque biogenesis, maintenance and degradation phospholipase and Cx43, 255 Phosphorylation (of Cx43 unless otherwise noted) cell cycle, changes during, 532 connexins known to be phosphoproteins, 264 Cx36 (Cx35), 265 effects on channel conductance, 153–156, 271, 402, 410 Cx37, 265 Cx40, 271 Cx45, 271 Cx50 (chick Cx45.6), 278 Cx56, 265 detergent solubility and, 233, 265 effects of carbenoxolone on, 210–211 fluid flow and, 378 gap junction plaque life-cycle, 264–278 See also Gap junction plaque biogenesis, maintenance and degradation

568 Phosphorylation (of Cx43 unless otherwise noted) (cont.) ischemia and, 274, 444 interaction with growth factors, 391–392 interaction with SP1/SP2 transcription factors, 377 lipid rafts and, 250 mutation at sites of, 443 oxidative stress, 391 SDS-PAGE gel mobility shift, 265 SH3 binding domain and, 40–41 See also Akt kinase; Ca2+/calmodulin protein kinase II; Casein kinase (CK); Dystrophia myotonica protein kinase (DMPK); Extracellular signal regulated kinase (ERK); Mitogenactivated protein kinase (MAPK); Protein kinase A (PKA); Protein kinase C (PKC); Protein kinase G (PKG); src Plaque formation and maintenance, see Gap junction plaque biogenesis, maintenance and degradation Plaques, see Gap junction plaques Poisson-Nernst-Planck (PNP) equations, 108, 117–118, 120 See also Vj/fast-gating Polarity reversal of Vj/fast-gating, see Vj/ fast-gating Polyamines, 209, 211 See also Pharmacology of connexin channels Polyethyleneglycols (PEGs), 52, 172–173, 291 See also Connexin pore - molecular permeability Po-pro-1, see Tracers Pore-lining domains, see Connexin pore pore-lining domains Posttranslational modifications, 264–265 acetylation, 264 acylation, 264 g-carboxylation, 264 glycosylation, 289 of connexins, absence of, 226, 264, 289 of pannexins, 288–289 hydroxylation, 264 phosphorylation, see Phosphorylation prenylation, 264 ubiquitination, see Ubiquitination

Index Projection density maps of connexin channels 2D high resolution, 36, 43–44, 80 low resolution, 36–37, 62 3D high resolution, 46–49, 64, 78–80, 98 low resolution, 37–39 Proline kink in M2, 104, 121–123 See also Transmembrane domains; Vj/fast-gating Propidium iodide, see Tracers Prostaglandin E2 (PGE2) release, 379 Prostate, see Male reproductive system Proteasome, 17, 227, 229, 234–235, 264, 277–278 See also Gap junction plaque biogenesis, maintenance and degradation Protein kinase A (PKA), 156, 181, 264, 266, 268, 270–271, 392, 498 Protein kinase C (PKC), 250, 264, 266–269, 275–276, 279, 391–392, 498, 532–533 Protein kinase G (PKG), 266 Pseudogenes, see Connexin gene family Purinergic receptors P2X7 receptor, 184–185, 210, 213, 291–292, 296 P2Y receptor, 291, 297, 378 P2Z receptor, see P2X7 receptor pannexin, ATP release and, 184–185, 291–292, 296 PY-motif and interactions, 269, 277–278 See also Carboxyl-terminal (CT) domain; Connexin interacting proteins; Ubiquitination Q Quinine, see Antimalarial drugs; Pharmacology of connexin channels R Rectification, single channel, see Single channel rectification Rectifying gap junctions, 55, 146, 149, 411, 424 See also Electrical synapses Residual conductance state, see Conductance substate; Vj/fast-gating Respiratory epithelium, 359–368 acute lung injury, 360, 367 acute respiratory distress syndrome (ARDS), 360, 367

Index adenosine triphosphate (ATP) release, 362, 365, 367 airway epithelium, 291, 293, 359–368 alveolar epithelium, 361–368 ATI and ATII pneumocytes, 361, 364–366 Ca2+ signaling and, 361–362, 365–366 inflammation, 360, 366–368 innate and acquired immunity, 360 roles of connexins in, 365–367 Retina, 153–156, 335–336, 397–414 adenosine triphosphate (ATP) release in, 413 amacrine cells, 149, 153, 156, 271, 335, 398–402, 408–413 AII amacrine-bipolar cell coupling, 149, 410–412 AII amacrine cell network, 408–410 bipolar cells, 149–153, 156, 335, 399–411 Ca2+ waves in, 413 cones, 398–405, 411 blue, 403–404 cone–cone coupling, 401–403 rod-cone coupling, 398–399, 404–405, 411, 413 ganglion cells, 399–402, 405, 408, 412–413 high-gain rod pathway, 398, 400–402, 410 horizontal cells, 153, 156, 336, 399–408 Mu¨ller cells, 412–413 photoreceptors, 399–406 photoreceptor coupling, 399, 402–406 rods, 398–400, 411, 413 rod–cone coupling, 399, 404–405 rod–rod coupling, 405–406 RNA splicing, 11–13, 15, 17, 289 See also Translation of connexin RNA Rods, 398–400, 411, 413 rod–cone coupling, 399, 404–405 rod–rod coupling, 405–406 See also Retina RXP-E peptide, 135–136 S Salivary glands, see Secretory glands SCAM (substituted cysteine accessibility method) effects on macroscopic conductance, 52, 87–88 effects on single channel conductance, 52, 89–90 endogenous cysteines, 95–96 evaluation of conflicting data, 90–99

569 identification of first extracellular loop (E1) and first transmembrane domain (M1) as major pore-lining domains, 52, 57, 87–90 identification of third transmembrane domain (M3) as major porelining domain, 52, 88 thiol reagents maleimidobutyryl biocytin (MBB), 87–88, 92–93, 95 methanethiosulfonate (MTS), 89–90, 92–96 state-dependent accessibility, 95 See also Connexin pore - pore-lining domains Schizophrenia, 15 Schmidt-Lanterman (S-L) incisures, 325–327, 329–330 See also Nervous system Schwann cells, 11, 14–15, 324–328, 334, 336–338 myelinating, 325–328 nonmyelinating, 324 See also Nervous system Second extracellular loop (E2), see Extracellular loops Second transmembrane domain (M2), see Transmembrane domains Secretory glands, 511–524 adrenal gland, 513 endocrine and exocrine differences in connexin expression, 513–514 gap junctions and secretion control, 517–521 gap junctions between secretory cells, 516–517 mammary gland, 513–514, 523 nonsecretory functions and connexins, 522–523 pancreas acinar cells, 513–514, 517–519 Ca2+ waves in, 519 Cx43 promoter, 11 islet cells, 513–515, 517–518, 520–521, 524 b-cells and insulin secretion, 513–515, 518, 520–524 See also Diabetes parathyroid, 513 salivary gland, 512, 519 secretory diseases, 523–524 thyroid, 513

570 Seizures, see Epilepsy and seizure Seminal vesicles, see Male reproductive system Single channel rectification, 85–87, 90, 108, 116, 122, 173 See also Connexin pore - conductance; Current-voltage (I-V) relation Single nucleotide polymorphism (SNP), see Connexin gene family Sinoatrial (SA) node, 436, 440 See also Heart Skeletal biology, 371–381 adenosine triphosphate (ATP) release in, 378–379 bone development, 372–374 bone maintenance, 374–376 Ca2+ waves in, 377–379 chondrocytes, 372–378 connexin regulation of gene expression in, 376–377 connexin response element (CxRE), 377, 532 mechanotransduction, 378–380 oculodentodigital dysplasia (ODDD), 15–16, 310–311, 338–339, 373–375, 443, 485, 489 See also Connexin diseases; Nervous system osteoblasts, 372, 374–376, 378 osteoclasts, 372 osteocytes, 372, 375, 378–380 osteopenia, 375 parathyroid hormone (PTH), 375–376, 378 Skin, 307–316 adenosine triphosphate (ATP) release by keratinocytes, 314 anatomy and connexin distribution, 307–309 animal models, 312–314 cancer basal cell carcinoma (BCC), 311 squamous cell carcinoma (SCC), 311 connexin diseases of with auditory system involvement deafness, peripheral neuropathy and erythrokeratoderma variabilis (EKV), 16, 310, 315, 427, See also Nervous system

Index hystrix-like ichthyosis deafness (HID), 16, 310, 315 keratitis-ichthyosis deafness (KID) syndrome, 16, 310–311, 315 mutilating keratoderma with sensorineural deafness (Vohwinkel’s Syndrome), 16, 310, 314–315 palmoplantar keratoderma, 16, 310–311 sensorineural hearing loss and palmoplantar hyperkeratosis, 16, 310 See also Auditory system erythrokeratoderma variabilis (EKV), 16, 310, 315, 427 hidrotic ectodermal dysplasia (Clouston syndrome, HED), 16, 310, 367 hypotrichosis, 310–311 See also Connexin diseases differentiation, 312 hair follicle, 307–308 inflammation, 310, 314 mutations, 18, 315 psoriasis, 311 in vitro models, 312–313 wound healing, 312 See also Cancer Slow gating, see Loop/slow-gating SP1/SP3 (specificity proteins 1 and 3), see Transcription factors Spermatogenesis, 495, 499–502 See also Male reproductive system Spermidine, see Pharmacology of connexin channels; Polyamines Spermine, see Pharmacology of connexin channels; Polyamines Src, 131, 135, 264, 266, 376 cellular- (c-src), 250, 268–269, 276 gating and, 276 viral- (v-src), 41, 268, 276 src homology (SH) motifs and interactions, 41, 135, 268, 276, 278, 280 See also Carboxyl-terminal (CT) domain; Connexin interacting proteins; src Surface plasmon resonance, 131, 276 Syndromic deafness, see Auditory system; Connexin diseases

Index T Testis, see Male reproductive system Tetraalkylammonium ions (TAA+), 208–209, 211 See also Pharmacology of connexin channels Tetracysteine-based tags, 232, 233, 250, 253 Thiol reagents maleimidobutyryl biocytin (MBB), 87–88, 92–93, 95 methanethiosulfonate (MTS), 89–90, 92–96 See also SCAM Third transmembrane domain (M3), see Transmembrane domains Tight junctions, 41, 268, 272–276, 515–516 See also Gap junction plaque biogenesis, maintenance and degradation; Zonula occludens Tissue-restricted knockouts, see Knockouts Tissue-specific connexin promoters liver, 11, 14 nervous system, 11–12, 14 pancreas, 11 Topological analysis, 7–8, 39–40 See also Connexin channel domain structure Tracers, 174–181 4, 6-diamidino-2-phenyl-indole dihydrochloride (DAPI), 176–177, 179–180 Alexa dyes, 177, 179–181 biocytin, 324 calcein (CA), 177, 179–180 carboxyfluorescein (CF), 175, 177–178, 327 dichlorofluorescein (DCF), 177–178 ethidium bromide (EB), 177–180 hydroxcoumarin carboxylic acid (HCCA), 177 Lucifer yellow (LY), 176–177, 179–180, 184, 194, 332, 362–363, 407, 425, 518 neurobiotin (NB), 147, 176–177, 332, 363, 404–406, 407–409, 411, 425 N, N, N-trimethyl-2-[methyl-(7-nitro-2, 1, 3-benzoxadiazol-4yl)amino]ethanaminium (NBDT-M-TMA; NMT), 177–178 permeability to, 174–181 po-pro-1 (PP1), 177–178 propidium iodide (PI), 177, 179–180

571 Transdominant interactions, see Dominant-negative interactions and mutations Transcription of connexin genes effects of mutation on, 15 epigenetic mechanisms, 14, 505 microRNAs (miRNAs), 15 regulation, 11–15, 486, 489, 513 See also Tissue-specific connexin promoters; Transcription factors Transcription factors activator protein-1 (AP-1), 13, 503 cAMP response element-binding protein (CREB), 270 c-fos, 13 early growth response gene-2 (Egr2/Knox20), 14 GATA binding protein 4 (GATA4), 14 hepatocyte nuclear factor-1 (HNF-1), 14 Mist1, 513 neuron-restrictive silencer element (NRSE), 513 NK2 transcription factor related, locus 5 (Nkx2, 5), 14 nuclear factor-1 (NF-1), 14 RE-1 silencing transcription factor (REST), 513 RY (sex determining region Y)-box 10 (SOX10), 14 SP1/SP3 (specificity proteins 1 and 3), 13–14, 377 TATA box binding protein, SP1/SP3, 13–14, 377 T-box transcription factor (Tbx5), 14, 373 T-cell factor (TCF)/lymphocyte enhancer binding factor (LEF), 13 Y1-associated factor 2 (YAF-2), 315 See also Transcription of connexin genes Transendothelial migration, see Cancer Transgenic mice (connexin, excluding knock-ins) Cx26, 311, 314–315, 428–429 Cx30, 428–429 Cx32, 338, 517 Cx36, 523 Cx43, 444, 447 Transjunctional voltage (Vj) sensitivity, see Vj/fast-gating Translation of connexin RNA effect of mutation on, 15, 17, 426 initiation and regulation, 13–15

572 Translation of connexin RNA (cont.) internal ribosome entry site (IRES), 14–15 See also RNA splicing Transmembrane domains (M1, M2, M3, M4), 7–8, 18, 40, 48, 51–52, 81 a-helical structure, 41–52 pore-lining, 52–53, 65, 79–91, 96–99 replacement by polyalanine, 51 and voltage sensing, 82–85, 118 See also Connexin domain organization Transmembrane voltage (Vm) sensitivity, see Vi-o (inside-out voltage) Transport-specific fractionation, 214, 216–217 Tremor, see Nervous system Tumor suppressor effect of connexins, 311, 531–535 See also Cancer Turnover, see Gap junction plaque biogenesis, maintenance and degradation Tyrosine-based sorting signals, see Gap junction plaque biogenesis, maintenance and degradation U Ubiquitination, 227–228, 269, 533 See also Gap junction plaque biogenesis, maintenance and degradation Uncoupling agents, see Pharmacology of connexin channels Unitary conductances of connexin channels, 78, 92, 167–169, 171–172, 181 See also Connexin pore - conductance Uterus endometrium, 485–488 implantation, 486–488 myometrium, 13, 488–489 See also Female reproductive system V Vasculature, 457–465 Ca2+ waves, 461–462 conducted vascular responses, 457–463 conducted constrictions, 462 conducted dilations, 460–462 connexin knockouts in, 264 connexins in, 457–465 endothelium-derived hyperpolarizing factor (EDHF), 459, 464 hypertension, 464–465, 477, 523 hypotension, 465 microcirculation, 458, 462, 464

Index myoendothelial gap junctions (MEGJs), 458–459, 462–465 nitric oxide (NO), 460–461, 464–465 vascular endothelial cells (ECs), 458–465 in atherosclerosis, 472–477 vascular smooth muscle (VSM) in atherosclerosis, 458–460 in vascular function, 457–465 See also Atherosclerosis Ventricular fibrillation, see Arrhythmia Ventricular heart disease, 444–448 Cx40 gap junctions, 446 Cx43 gap junctions altered distribution of, 444–445 arrhythmia development, 447–448 cardiac ischemia and infarction, 175, 344, 444–446 reduction in, 446–448 regulation of, 446 Cx45 gap junctions, 446 See also Arrhythmia; Heart Vestibular system, 422–424, 427–428 Vi-o (inside-out voltage, Vm) gating, 103–107, 112–113, 125–126 definition, 104–105 lack of evidence in single channels, 113 location of sensor relative to electric field, 105–106 in macroscopic currents, 106–107 molecular determinants of, 125–126 relation to loop/slow and chemical gating, 112 See also Voltage sensitivity Vj/fast-gating, 82–83, 112–120 application of PNP theory to, 117–118 conductance transitions of, 112–113 conformational changes associated with, 121–125 amino-terminal (NT) domain, 124–125 carboxyl-terminal (CT) domain, 123–124 proline kink in M2, 121–123 control of polarity, 83–85, 113–118 definition, 82–83, 112 distinguished from loop/slow-gating, 113, 115, 123 mechanisms and determinants of, 114–120 mutations affecting polarity, 83–85, 114–118

Index voltage sensor, 83–85 See also Loop/slow-gating; Voltage sensitivity Vm, see Vi-o (inside-out voltage) gating Vohwinkel’s syndrome, see Mutilating keratoderma with sensorineural deafness Voltage sensitivity (macroscopic, Vj), 103–111 conductance-voltage (G-V) relation, 106–110, 120–121, 124 contingent gating and Vj sensitivity, 110–111 location of Vj sensor relative to electric field, 105–106 loop/slow-gating, see Loop/slow-gating mechanisms of, 104–105, 111–114 Vi-o (inside-out voltage) gating, see Vi-o (inside-out voltage) gating Vj/fast-gating, see Vj/fast-gating

573 W Wound healing, 312, 314, 367 See also Skin X Xenopus oocyte prep, cut-open, 88, 93–96 X-ray diffraction of gap junction plaques, 34–35, 38, 247 Z Zebrafish connexins, 7–10, 314, 373, 544 Zonula occludens ZO-1, 135, 254–255, 269, 272–274, 2 76, 280 associated nucleic acid-binding protein (ZONAB), 269, 280 plaque formation and maintenance and, 254–255, 274 ZO-2, 254, 269, 273–274 See also Tight junctions