Ion Channel Factsbook, Volume 4: Voltage-Gated Channels

THE ION CHANNEL FactsBook IV Voltage-Gated Channels Other books in the FactsBook Series: Edward C. Conley The Ion Ch...

Author: William J. Brammar

55 downloads 1350 Views 62MB 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

THE ION CHANNEL FactsBook IV

Voltage-Gated Channels

Other books in the FactsBook Series: Edward C. Conley The Ion Channel Factsbook I: Extracellular Ligand-Gated Channels Edward C. Conley The Ion Channel Factsbook H: Intracellular Ligand-Gated Channels A. Neil Barclay, Albertus D. Beyers, Marian L. Birkeland, Marion H. Brown, Simon J. Davis, Chamorro Somoza and Alan F. Williams The Leucocyte Antigen FactsBook, 1st edn Robin Callard and Andy Gearing The Cytokine FactsBook Steve Watson and Steve Arkinstall The G-Protein Linked Receptor FactsBook Rod Pigott and Christine Power The Adhesion Molecule FactsBook Grahame Hardie and Steven Hanks

The Protein Kinase FactsBook The Protein Kinase FactsBook CD-Rom Kris Vaddi, Margaret Keller and Robert Newton The Chemokine FactsBook Marion E. Reid and Christine Lomas-Francis The Blood Group Antigen FactsBook A. Neil Barclay, Marion H. Brown, S.K. Alex Law, Andrew J. McKnight, Michael G. Tomlinson and P. Anton van der Merwe The Leucocyte Antigen FactsBook, 2nd edn Jeff Griffiths and Clare Sansom The Transporter FactsBook Robin Hesketh The Oncogene and Thmour Suppressor Gene FactsBook, 2nd edn Shirley Ayad, Ray P. Boot-Handford, Martin J. Humphries, Karl E. Kadler and C. Adrian Shuttleworth The Extracellular Matrix FactsBook, 2nd edn TakW. Mak The Gene Knockout FactsBook

THE ION CHANNEL

FactsBook IV Voltage-Gated Channels Edward C. Conley Molecular Pathology, c/o Ion Channel/Gene Expression University of Leicester/Medical Research Council Centre for Mechanisms of Human Toxicity, U.K.

William

J. Brammar

Department of Biochemistry, University of Leicester, UK

Academic Press SAN DIEGO LONDON BOSTON NEW YORK SYDNEY TOKYO TORONTO

This book is printed on acid-free paper Copyright © 1999 by ACADEMIC PRESS

All rights reserved No part of this publication may be reproduced or transmitted in any form by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Academic Press 24-28 Oval Road, London NWI 7DX, UK http://www.hbuk.co.uk/ap/ Academic Press 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.apnet.com ISBN 0-12-184453-6 A catalogue record for this book is available from the British Library

Library of Congress Catalogue Card Number: 98-86472

Typeset in Great Britain by Alden Bookset, Oxford Printed in Great Britain by WBC, Bridgend, Mid Glamorgan 99 00 01 02 03 04 WB 9 8 7 6 5 4 3 2 1

Contents Cumulative table of contents for Volumes I to IV (entry 01 resume) Acknowledgements Introduction &. layout of entries (entry 02 resume) How to use The Ion Channel FactsBook Guide to the placement criteria for each field Abbreviations (entry 03 resume)

_

VIII

_ _ _ _ _

xn xm

xv

XVII XXXIX

VOLUME IV VOLTAGE-GATED CHANNELS VLG Key facts (entry 41) Voltage-gated channel families - key facts References

3 21

VLG Ca (entry 42) Voltage-gated calcium channels Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References Note Added in Proof

22 22 39 55 69 84 101 131 140 153

VLG CI (entry 43) Voltage-gated chloride channels Nomenclatures Expression Sequence Analysis Structure and Functions Electrophysiology Pharmacology Information Retrieval References

154 154 157 166 173 177 186 189 193

VLG K A-T [native] (entry 44) Rapidly inactivating, transient outward i A-type' K+ currents in native cell types of vertebrates Nomenclatures Expression Electrophysiology Pharmacology Information Retrieval References

196 196 201 206 211 219 223

VLG K DR [native] (entry 45) 'Delayed rectifier'-type K+ currents in native cell types of vertebrates Nomenclatures Expression Structure and Functions Electrophysiology Pharmacology Information Retrieval References

226 226 231 246 249 252 268 269

II

_1......--

_

VLG K eag/elk/erg (entry 46) K+ channels encoded by genes related to Drosophila eag (ether-a-go-go) (gene subfamilies eag, elk, erg) Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

275 275 284 291 298 306 313 321 323

VLG K Kv-beta (entry 47) Cytoplasmic (Kv,B) subunits co-assembling with pore-forming (Kvo) voltage-gated potassium channel subunits Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

327 327 332 346 352 359 363 366 372

VLG K Kvl-Shak (entry 48) Vertebrate K+ channels related to Drosophila Shaker (Kvo subunits encoded by gene subfamily Kvl) Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

374 374 383 417 436 460

484 507 513

VLG K Kv2-Shab (entry 49) Vertebrate K+ channel subunits related to Drosophila Shab (Kvo subunits encoded by gene subfamily Kv2) Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

524 524 528 537 541 545 550 554 556

VLG K Kv3-Shaw (entry 50) Vertebrate K+ channels related to Drosophila Shaw (Kvo subunits encoded by gene subfamily Kv3) Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

559 559 565 582 591 599 607 610 614

II

"------------VLG K Kv4-Shal (entry 51) Vertebrate K+ channel subunits related to Drosophila Shal (Kvo: subunits encoded by gene subfamily Kv4) Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

617 617 621 629 632 634 638 642 644

VLG K Kvx [unassigned] (entry 52) Listing of cDNA clones encoding Kv channels with unassigned gene family relationships Nomenclatures Expression Sequence Analyses Structure and Functions Information Retrieval References

647 647 651 653 654 655 656

VLG K M-i [native] (entry 53) 'Muscarinic-inhibited' K+ channels underlying 1M (M-current in native cell types) Nomenclatures Expression Structure and Functions Electrophysiology Pharmacology Information Retrieval References

657 657 661 665 669 679 699 699

VLG (K) minK (entry 54) 'Minimal' protein subunits (minK, IsK) eliciting 'slow-activating' voltage-gated currents in oocytes Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

703 703 708 723 728 743 750 760 765

VLG Na (entry 55) Voltage-gated sodium channels Nomenclatures Expression Sequence Analyses Structure and Functions Electrophysiology Pharmacology Information Retrieval References

768 768 772 788 796 806 816 827 831

Rubrics (entry 13 resume)

839

Index

842

III

Cumulative table of contents for Volumes I to IV Contents Cumulative table of contents for Volumes I to IV (entry 01) Acknowledgements Introduction and layout of entries (entry 02) How to use The Ion Channel FactsBook Guide to the placement criteria for each field Abbreviations (entry 03)

VOLUME I

EXTRACELLULAR LIGAND-GATED CHANNELS

ELG Key facts (entry 04)

Extracellular ligand-gated receptorchannels - key facts

ELG CAT 5-HTa (entry 05) Extracellular 5-hydroxytryptaminegated integral receptor-channels ELG CAT ATP (entry 06) Extracellular ATP-gated receptorchannels (P2xR) ELGCATGLUAMPA!KAIN(entry07) AMPA / kainate-selective (nonNMDA) glutamate receptor-channels ELG CAT GLU NMDA (entry 08) N-Methyl-D-aspartate (NMDA)selective glutamate receptor-channels

VOLUME II

ELG CAT nAChR (entry 09) Nicotinic acetylcholine-gated integral receptor-channels ELG Cl GABAA (entry 10) Inhibitory receptor-channels gated by extracellular gamma-aminobutyric acid ELG Cl GLY (entry 11) Inhibitory receptor-channels gated by extracellular glycine Feedback and access to the CellSignalling Network (entry 12) Rubrics (entry 13) Entry and field number rubrics

INTRACELLULAR LIGAND-GATED CHANNELS

ILG Key facts (entry 14) The intracellular ligand-gated channel group - key facts ILG Ca AA-LTC4 [native] (entry 15) Native Ca2 + channels gated by the arachidonic acid metabolite leukotriene C4 incorporating general properties of ion channel regulation by arachidonate metabolites ILG Ca Ca InsP4 S [native] (entry 16) Native Ca2 + channels sensitive to inositol l,3,4,5-tetrakisphosphate (InsP4)

ILG Ca Ca RyR-Caf (entry 17) Caffeine-sensitive Ca2 + -release channels (ryanodine receptors, RyR) ILG Ca CSRC [native] (entry 18) Candidate native intracellularligand-gated Ca2 + -store repletion channels ILG Ca InsPa (entry 19) Inositol l,4,5-trisphosphatesensitive Ca2 + -release channels (InsPaR)

1...----

_

ILG CAT Ca [native] (entry 20) Native calcium-activated nonselective cation channels (NS ca ) ILG (CAT) cAMP (entry 21) Cation channels activated in situ by intracellular cAMP ILG CAT cGMP (entry 22) Cation channels activated in situ by intracellular cGMP

ILG Cl Ca [native] (entry 25) Native calcium-activated chloride channels (Clca) ILG K AA [native] (entry 26) Native potassium channels activated by arachidonic acid (KAA ) incorporating general properties of ion channel regulation by free fatty acids

ILG Cl ABC-CF (entry 23) ATP-binding and phosphorylationdependent CI- channels (CFTR)

ILG K Ca (entry 27) Intracellular calcium-activated K+ channels (Kca )

ILG Cl ABC-MDR/PG (entry 24) Volume-regulated CI- channels (multidmg-resistance P-glycoprotein)

ILG K Na [native] (entry 28) Native intracellular sodium-activated K+ channels (KNa )

VOLUME III

INWARD RECTIFIER AND INTERCELLULAR CHANNELS

INR K Key facts (entry 29) Inwardly-rectifying K+ channels key facts INR K ATP-i [native] (entry 30) Properties of intracellular ATPinhibited K+ channels in native cells INR K G/ACh [native] (entry 31) Properties of muscarinic-activated K+ channels underlying l KAch in native cells INR K [native] (entry 32) Properties of 'classical' inward rectifier K+ channels in native cells (excluding types covered in entries 30 & 31) INR K [subunits] (entry 33) Comparative properties of protein subunits forming inwardlyrectifying K+ channels (heterologously-expressed cDNAs of the KIR family) INR (KINa) IfhQ (entry 34) Hyperpolarization-activated cation channels underlying the inward currents if, ih, i Q

TUN [connexins] (entry 35) Intercellular gap junction channels formed by connexin proteins

MEC [mechanosensitive] (entry 36) Survey of ion channel types activated by mechanical stimuli MIT [mitochondrial] (entry 37) Survey of ion channel types expressed in mitochondrial membranes NUC [nuclear] (entry 38) Survey of ion channel types expressed in nuclear membranes OSM [aquaporins] (entry 39) The vertebrate aquaporin (water channel) family

SYN [vesicular] (entry 40) Channel-forming proteins expressed in synaptic vesicle membranes (synaptophysin)

II

_L.-...-VOLUME IV

_

VOLTAGE-GATED CHANNELS

VLG Key facts (entry 41) Voltage-gated channels - key facts VLG Ca (entry 42) Voltage-gated calcium channels VLG Cl (entry 43) Voltage-gated chloride channels VLG K A-T (entry 44) Properties of native'A-type' (transient outward) potassium channels in native cells VLG K DR (entry 45) Properties of native delayed rectifier potassium channels in native cells VLG K eag/elk/erg (entry 46) K+ channels related to Drosophila gene subfamilies eag,elk,erg VLG K Kv-beta (entry 47) Beta subunits associated with Kv (alpha subunit) channel complexes VLG K Kvl-Shak (entry 48) Vertebrate K+ channel subunits related to Drosophila Shaker (Kv subfamily I) incorporating general features of Kv channel expression in heterologous cells

VLG K Kv2-Shab (entry 49) Vertebrate K+ channel subunits related to Drosophila Shab (Kv subfamily 2) VLG K Kv3-Shaw (entry 50) Vertebrate K+ channel subunits related to Drosophila Shaw (Kv subfamily 3) VLG K Kv4-Shal (entry 51) Vertebrate K+ channel subunits related to Drosophila Shal (Kv subfamily 4) VLG K Kvx (Kv5.1/Kv6.1) (entry 52) Features of the 'non-expressible' cDNAs IK8 and KI3 VLG K M-i [native] (entry 53) Properties of native 'muscarinicinhibited' K+ channels underlying 1M

VLG (K) minK (entry 54) 'Minimal' protein subunits inducing 'slow-activating' voltagegated K+ currents VLG Na (entry 55) Voltage-gated sodium channels

ION CHANNEL RESOURCES

Resource documents and/or links supporting their scope will appear on the Ion Channel Network web site (www.le.ac.uk/csn/ ) from January 1999. Resource A Resource C G protein-linked receptors Compounds and proteins used in ion regulating ion channel activities channel research (alphabetical listing) Resource D Resource B 'Generalized' electrical effects of endogeneous receptor agonists

II

'Diagnostic'tests

Resource E Ion channel book references (sorted by year of publication)

"-----------Resource F Supplementary ion channel reviews (listed by subject)

Resource I Framework of cell-signalling molecule types (preliminary listing)

Resource G Reported 'consensus sites' and 'motifs' in primary sequences of ion channels

Resource' Search criteria &. CSN development

Resource H Listings of cell types

Resource K Framework for a multidisciplinary glossary

Feedback: Comments and suggestions regarding the scope, arrangement and other matters relating to the coverage/contents can be sent to the e-mail [email protected]. (see field 57 of most entries for further details)

II

Acknowledgements Thanks are due to the following people for their time and help during compilation of the manuscripts: Professors Peter Stanfield, Nick Standen and Gordon Roberts (Leicester), and Ole Petersen (Liverpool) for advice; to Chris Hankins and Richard Mobbs of the Leicester University Computer Centre, and to Dr Tessa Picknett and Chris Gibson of Academic Press for their enthusiasm and patience. Gratitude is also expressed to all of the anonymous manuscript readers who supplied much constructive feedback, as well as the following who provided advice, information and encouragement: Ihab Awad (Center for Neuroscientific Databases, Minnesota), Jonathan Bard (Edinburgh), Mark Boyett (Leeds), David Brown (London), Cecilia Canessa (Yale), Marty Chalfie (Columbia), K. George Chandy (UC Irvine), David Clapham (Mayo Foundation), Noel Davies (Leicester), Dario DiFrancesco (Milano), Ian Forsythe (Leicester), Harry Fozzard (Chicago), Klaus Groschner (Graz), George Gutman (UC Irvine), Mike Huerta (NIMH, HBP), Rolf Joho (Texas SMC), Benjamin Kaupp (Julich), Steve Kozlow (NIMH, HBP), Jeremy Lambert (Dundee), Neil Marrion (OHSU), Shigetada Nakanishi (Kyoto), Jitendra Patel (Zeneca USA), Martin Ringwald (Jackson Labs), Gordon Shepherd (Yale), David Spray (Yeshiva), Zhong-ping Sun (Columbiail, Steve Watson (Oxford) and George Wilcox (Center for Neuroscientific Databases, Minnesota). Thanks are also due to the Department of Pathology at the University of Leicester, Harcourt Brace, the Medical Research Council, the British Heart Foundation and Zeneca Pharmaceuticals, for providing generous sponsorship, equipment and facilities. We would like to acknowledge the authors of all those papers and reviews

which in the interest of completeness we have quoted, but have not had space to cite directly. ECC would like to thank Professors Denis Noble in Oxford, Anthony Campbell in Cardiff and Richard Gregory in Bristol for help and inspiration, and would like to dedicate his contributions to Paula, Rebecca and Katherine for all their love and support.

Left: Edward Conley, Right: William Brammar

Introduction & layout of entries Edward C. Conley

Entry 02 resume

The Ion Channel FactsBook is intended to provide a 'summary of molecular properties' for all known types of ion channel protein in a cross-referenced and 'computer-updatable' format. Today, the subject of ion channel biology is an extraordinarily complex one, linking several disciplines and technologies, each adding its own contribution to the knowledge base. This diversity of approaches has left a need for accessible information sources, especially for those reading outside their own field. By presenting 'facts' within a systematic framework, the FactsBook aims to provide a 'logical place to look' for specific information when the need arises. For students and researchers entering the field, the weight of the existing literature, and the rate of new discoveries, makes it difficult to gain an overview. For these readers, The Ion Channel FactsBook is written as a directory, designed to identify similarities and differences between ion channel types, while being able to accommodate new types of data within the framework. The main advantages of a systematic format is that it can speed up identification of functional links between any 'facts' already in the database and maybe provide a raison d'etre for specific experiments where information is not known. Although such 'facts' may not go out-of-date, interpretations based on them may change considerably in the light of additional, more direct evidence. This is particularly true for the explosion of new information that is occurring as a direct consequence of the molecular cloning of ion channel genes. It can be anticipated that many more ion channel genes will be cloned in the near future, and it is also likely that their functional diversity will continue to exceed expectations based on pharmacological or physiological criteria alone.

An emphasis on properties emergent from ion channel molecular functions Understanding how the interplay of currents through many specific ion channel molecules determines complex electrophysiological behaviour of cells remains a significant scientific challenge. The approach of the FactsBook is to associate and relate this complex cell phenotypic behaviour (e.g. its physiology and pharmacology) to ion channel gene expression-control wherever possible even where the specific gene has not yet been cloned. Thus the ion channel molecule becomes the central organizer, and accordingly arbitrates whether information or topics are included, emphasized, sketched-over or excluded. In keeping with this, ion channel characteristics are described in relation to known structural or genetic features wherever possible (or where they are ultimately molecular characteristics). Invariably, this relies on the availability of sequence data for a given channel or group of channels. However, a number of channel types exist which have not yet been sequenced, or display characteristics in the native form which are not precisely matched by existing clones expressed in heterologous cells (or are otherwise ambiguously classified). To accommodate these channel types, summaries of characteristics are included in the standard entry field format, with inappropriate fieldnames omitted. Thus the present 'working arrangement' of entries and fields is broad enough to include both the 'cloned' and 'uncloned' channel types, but in due course will be gradually supplanted by a comprehensive classification based on gene locus, structure, and relatedness of primary sequences. In all cases, the scope of the FactsBook entries is limited to those proteins forming (or predicted to form) membrane-bound, integral ionic channels

-

_1...-.---

-

_

entry 02 resume

by folding and aSSocIatIon of their primary protein sequences. Activation or suppression of the channel current by a specified ligand or voltage step is generally included as part of the channel description or name (see below). Thus an emphasis is made throughout the book on intrinsic features of channel molecule itself and not on those of separately encoded, co-expressed proteins. In the present edition, there is a bias towards descriptions of vertebrate ion channels as they express the full range of channel types which resemble characteristics found in most eukaryotes.

Anticipated development of the dataset - Integration of functional information around molecular types Further understanding of complex cellular electrical and pharmacological behaviour will not come from a mere catalogue of protein properties alone. This book therefore begins a process of specific cross-referencing of molecular properties within a functional framework. This process can be extended to the interrelationships of ion channels and other classes of cell-signalling molecules and their functional properties. Retaining protein molecules (i.e. gene products) as 'fundamental units of classification' should also provide a framework for understanding complex physiological behaviour resulting from co-expressed sets of proteins. Significantly, many pathophysiological phenotypes can also be linked to selective molecular 'dysfunction' within this type of framework. Finally, the anticipated growth of raw sequence information from the human genome project may reveal hitherto unexpected classes and subtypes of cell-signalling components - in this case the task then will be to integrate these into what is already known (see also description of Field number 06: Subtype c.lassifications and Field number 05: Gene family).

The Cell-Signalling Network (CSN) From the foregoing discussion, it can be seen that establishment and consolidation of an integrated 'consensus database' for the many diverse classes of cell signalling molecules (including, for example, receptors, G proteins, ion channels, ion pumps, etc.) remains a worthwhile goal. Such a resource would provide a focus for identifying unresolved issues and may avoid unnecessary duplication of research effort. Work has begun on a prototype cell-signalling molecule database cooperatively maintained and supported by contributions from specialist groups: The Cell-Signalling Network (CSN) in mid-1996 has been designed to disseminate consensus properties of a wide range of molecules involved in cell signal transduction. While it will take some time (and much good-will) to establish a comprehensive network, the many advantages of such a co-operative structure are already apparent. Immediately, these include an 'open' mechanism for consolidation and verification of the dataset, so that it holds a 'consensus' or 'validated' set of information about what is known about each molecule and practical considerations such as nomenclature recommendations (see, for example, the IUPHAR nomenclature sections under the CSN 'home page'). The CSN also allows unlimited cross-referencing by pointing to related information sets, even where these are held in multiple centres. On-line descriptions of technical terms (glossary items, indicated by dagger symbols (t) throughout the text) and reference to explanatory references (e.g. on associated signalling components such

-

l'---e_n_t_ry_O_2_r_e_s_u_m_e

_

as G protein t -linked receptors t ) are being written for use with this book. Eventually, applications could include (for instance) direct 'look-up' of graphical resources for protein structure, in situ and developmental gene expression atlases t, interactive molecular models for structure/function analysis, DNA/protein sequences linked to feature tables, gene mapping resources and other pictorial data. These developments (not presently supported) will use interactive electronic media for efficient browsing and maintenance. For a brief account of the Cell-Signalling Network, see Feedback eiJ CSN access, entry 12. For a full specification, see Resource T- Search criteria eiJ CSN development.

HOW TO USE THE ION CHANNEL FACTSBOOK

Common formats within the entries A proposed organizational hierarchy for information about ion channel

molecules Information on named channel types is grouped in entries under common headings which repeat in a fixed order - e.g. for ion channel molecules which have been sequenced, there are broad sections entitled NOMENCLATURES, EXPRESSION, SEQUENCE ANALYSES, STRUCTURE &. FUNCTIONS, ELECTROPHYSIOLOGY, PHARMACOLOGY, INFORMATION RETRIEVAL and REFERENCES, in that order. Within each section, related fieldnames are listed, always in alphabetical order and indexed by a field number (see below), which makes electronic cross-referencing and 'manual' comparisons easier. While the sections and fields are not rigid categories, an attempt has been made to remain consistent, so that corresponding information for two different channels can be looked up and compared directly. If a field does not appear, either the information was not known or was not found during the compilation period. Pertinent information which has been published but is absent from entries would be gratefully received and will be added to the 'entry updates' sections within the CSN (see Feedback eiJ CSN access, entry 12). Establishment of this 'field' format has been designed so that most 'facts' should have their logical 'place'. In the future, this arrangement may help to establish 'consensus' properties of any given ion channel or other cell-signalling molecule. This validation process critically depends on user feedback to contributing authors. The CSN (above) establishes an efficient electronic mechanism to do this, for continual refinement of entry contents.

Independent presentation of lacts' and conventions for cross-referencing The FactsBook departs from a traditional review format by presenting its information in related groups, each under a broader heading. Entries are not designed or intended to be read lfrom beginning to end', but each Ifact' is presented independently under the most pertinent fieldname. Independent citation of 'facts' may sometimes result in some repetition (redundancy) of general principles between fields, but if this is the case some effort has been made to 'rephrase' these for clarity (suggested improvements for presentation of any 'fact' are welcome - see Field number 57: Feedback).

-

_ _~

e_n_try __ 02_r_es_u_m_e_

For readers unfamiliar with the more general aspects of ion channel biology, some introductory information applicable to whole groups of ion channel molecules is needed, and this is incorporated into the 'key facts' sections preceding the relevant set of entries. These sections/. coupled with the glossary items (available on-line, and indicated by the daggert symbol, see below) provide a basic overview of principles associated with detailed information in the main entries of the book. Extensive cross-referencing is a feature of the book. For example, cross-references between fields of the same entry are of the format (see Fieldname, xx-yy). Crossreferences between fields of different channel type entries are generally of the format see fieldname under SORTCODE, xx-yy; for example - see mRNA distribution, under ELG Cl GABAA , 10-13. This alphabetical 'sortcode' and numerical 'entry numbers' (printed in the header to each page) are simply devices to make crossreferencing more compact and to arrange the entries in an afproximate running order based on physiological features such as mode of gating, ionic selectivityi, and agonist t specificity. A 'sort order' based on physiological features was judged to be more intuitive for a wider readership than one based on gene structure alone, and enables 'cloned' and 'uncloned' ion channel types to be listed together. The use and criteria for sortcode designations are described under the subheading Derivation of the sortcode (see Field number 02: Category (sortcode)). Entry 'running order' is mainly of importance in book-form publications. New entries (or mergers/subdivisions between existing entries) will probably use different serial entry numbers as 'electronic pointers' to appropriate files. Cross-references are frequently made to an on-line index of glossary items by dagger symbols t wherever they might assist someone with technical terms and concepts when reading outside their own field. The glossary is designed to be used side-byside with the FactsBook entries and will be accessible in updated form over the Internet with suitable software (for details, see Feedback eiJ CSN access, entry 12).

Contextual markers and styles employed within the entries Throughout the books, a six-figure index number (xx-yy-zz, e.g. 19-44-01: ) separates groups of facts about different aspects of the channel molecule, and carries information about channel type/entry number (e.g. 19InsP3 receptorchannels), information type/field number (e.g. -44-, Channel modulation) and running paragraph number (datatype) (e.g. -01). This simple 'punctate' style has been adopted for maximum flexibility of updating (both error-correction and consolidation with new information), cross-referencing and multi-authoring. The CSN specification includes longer term plans to further structure field-based information into convenient data-types which will be indexed by a zz numerical designation. r-..I

Italicized subheadings are employed to organize the facts into related topics where a field has a lot of information associated with it:. Specific illustrated points or features within a field are referenced to adjacent figures. Usage of abbreviations and common symbols are defined in context and/or within the main abbreviations index at the front of each book. Abbreviated chemical names and those of proprietary pharmaceutical compounds are listed within Resource C - Compounds eiJ proteins. Generally, highlighting of related subtopics emergent from the molecular properties ('facts') associated with the ion channel under description are indicated within a field

-

l_e_n_t_ry_O_2_r_e_s_u_m_e

__

by lettering in bold. Throughout the main text, italics draw attention to special cases, caveats, hypotheses and exceptions. The 'Note: ' prefix has been used to indicate supplemental or comparative information of significance to the quoted data in context.

Special considerations for integrating properties derived from tcloned' and tnative' channels While a certain amount of introductory material is given to set the context, the emphasis on molecular properties means the treatment of many important biological processes or phenomena is reduced to a bare outline. References given in the Related sources and reviews field and Resource F - Supplementary ion channel reviews are intended to address this imbalance. For summaries of key molecular features, a central channel 'protein domain topography model' is presented. Individual features that are illustrated on the protein domain topography model are identified within the text by the symbol [PDTM]. Wherever molecular subtype-specific data are quoted (such as the particular behaviour of a ion channel gene familyt member or isoformt) a convention of using the underlined trivial or systematic name as a prefix has been adopted - e.g. mIRKl: i RCKl: i Kv3.1: etc.

GUIDE TO THE PLACEMENT CRITERIA FOR EACH FIELD

Criteria for NOMENCLATURES sections This section should bring together for comparison present and previous names of ion channels or currents, with brief distinctions between similar terms. Where systematic names have already been suggested or adopted by published convention, they should be included and used in parallel to trivial names. Field number 01: Abstract/general description: This field should provide a summary of the most important functional characteristics associated with the channel type. Field number 02: Category (sortcode): The alphabetical1sortcode' should be used for providing a logical running order for the individual entries which make up the book. It is not intended to be a rigorous channel classification, which is under discussion, but rather a practical index for finding and cross-referencing information, in conjunction with the six-figure index number (see above). The Category (sortcode) field also lists a designated electronic retrieval code (unique embedded identifier or VEl) for 'tagging' of new articles of relevance to the contents of the entry. For further details on the use and implementation of UEIs, see the description for Resource T(in this entry) and for a full description, see Resource T- Search criteria eiJ CSN development.

Derivation of the sortcode: Although we do not yet have a complete knowledge of all ion channel primaryt structures, knowledge of ion channel gene familyt and superfamilyt structure allows a working sort order to be established. To take an

-

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try __ 02_r_es_u_m_e------J

example, the extracellular ligand-gated (ELG) receptor-channels share many structural features, which reflects the likely duplication and divergent evolution of an ancestral gene. The present-day forms of such channels reflect the changes that have occurred through adaptive radiation t of the ancestral type, particularly for gatingt mechanism and ionic selectivityt determinants. Thus, the entry running order (alphabetical, via the sortcode) of the FactsBook entries should depend primarily on these two features. The sortcode therefore consists of several groups of letters, each denoting a characteristic of the channel molecule: Entries are sorted first on the principal means for channel gatingt (first three letters), whether this is by an extracellular ligandt (ELG), small intracellular ligandt (ILG) or transmembrane voltage (VLG). For convenience, the ILG entries also include certain channels which are obligately dependent on both ligand binding and hydrolysis for their activation e.g. channels of the ATP-binding cassette (ABC) superfamily. Other channel types may be subject to direct mechanical gating (MEC) or sensitive to changes in osmolarity (OSM) - see the Cumulative tables of contents and the first page of each entry for descriptions and scope. Due to their unusual gating characteristics, a separate category (INR) has been created for inward rectifier-type channels. The second sort (the next three letters of the sortcode) should be on the basis of the principal permeant ions, and may therefore indicate high selectivity for single ions (e.g. Ca, Cl, K, Na) or multiple ions of a specified charge (e.g. cations - CAT). Indefinite sortcode extensions can be assigned to the sortcode if it is necessary to distinguish similar but separately encoded groups of channels (e.g. compare ELG Cl GABAA , entry 10 and ELG Cl GLY, entry 11). Field number 03: Channel designation: This field should contain a shorthand designation for the ion channel molecule - mostly of the form X y or X(y) where X denotes the major ionic permeabilities t (e.g. K, Ca, cation) and Y denotes the principal mechanism of gatingt where this acts directly on the channel molecule itself (e.g. cGMP, voltage, calcium, etc.). Otherwise, this field contains a shorthand designation for the channel which is used in the entry itself. Field number 04: Current designation: This field should contain a shorthand designation for ionic currents conducted by the channel molecule, which is mostly of the form IXI Y ), Ix,Y or I x -y where X and Yare. defined as above. Field number 05: Gene family: This field should indicate the known molecular relationships to other ion channels or groups of ion channels at the level of amino acid primary sequence homologyt, within gene familiest or gene superfamilies t. Where multiple channel subunits are encoded by separate genes, a summary of their principal features should be tabulated for comparison. Where the gene family is particularly large, or cannot be easily described by functional variation, a gene family tree t derived by a primary sequence alignment algorithm t (see Resource D - IDiagnostic' tests) may be included as a figure in this field. Field number 06: Subtype classifications: This field should include supplementary information about any schemes of classification that have been suggested in the literature. Generally, the most robust schemes are those based on complete knowledge of gene familyt relationships (see above) and this method can identify

-

1~_e_n_t_ry_O_2_r_e_s_u_m_e

-----"_

similarities that are not easily discernible by pharmacological or electrophysiological criteria alone - see, for example, the entries TUN (connexins), entry 35, and JNR K (subunits), entry 33. Note, however, that some native t channel types are more conveniently 'classified' by functional or cell-type expression parameters which take into account interactions of channels with other co-expressed proteins (see, for example, discussion pertaining to the cyclic nucleotide-gated (CNG-) channel family in the entries JLG Key facts, entry 14, JLG CAT cAM~ entry 21, and JLG CAT cGM~ entry 22. Debate on the 'best' or 'most appropriate' channel classification schemes is likely to continue for some time, and it is reasonable to suppose that alternative subtype classifications may be applied and used by different workers for different purposes. Since the 'running order' of the FactsBook categories depends on inherent molecular properties of channel cDNAs t, genes t or the expressed proteins, future editions will gradually move to classification on the basis of separable gene locit. Thus multiple channel protein variants resulting from processes of alternative RNA splicingt but encoded by a single gene locust will only ever warrant one 'channel-type' entry (e.g. see BKci variants under JLG K Ca, entry 27). Distinct proteins resulting from transcription T of separable gene loci, for example in the case of different gene family members, will (ultimately) warrant separate entries. For the time being, there is insufficient knowledge about the precise phenotypic t roles of many 'separable' gene family members to justify separate entries (as in the case of the VLG K Kv series entries). Classification by gene locus designation (see Field number 18: Chromosomal location) can encompass all structural and functional variation, while being 'compatible' with efforts directed to identifying phenotypic and pathophysiologicali roles of individual gene products (e.g. by gene-knockout t, locus replacement t or disease-linked gene mapping t procedures - see Resource D - tDiagnostic' tests). Subtype classifications based on gene locus control can also incorporate the marked developmental changes which pertain to many ion channel genes (see Field number 11: Developmental regulation) and can be implemented when the 'logic' underlying gene expression-control t for each family member is fully appreciated. A 'genome-based' classification of FactsBook entries may also help comprehend and integrate equivalent information based for other ('non-channel') cell-signalling molecules (see Resources G, Hand J). Field number 07: Trivial names: This field should list commonly used names for the ion channel (or its conductance t ). Often a channel will be (unsystematically) named by its tissue location or unusual pharmacological/physiological properties, and these are also listed in this field. While unsystematic names do not indicate molecular relatedness, they are often more useful for comparative/descriptive purposes. For these and historical reasons, trivial names (e.g. clone/isolate names for K+ channel isoforms) are used side-by-side with systematic names, where these exist. A standardized nomenclature for ion channels is under discussion, e.g. see the series of articles by Pongs, Edwards, Weston, Chandy, Gutman, Spedding and Vanhoutte in Trends Pharmacol Sci (1993) 14: 433-6. Future recommendations on standardized nomenclature will appear in files accessible under the IUPHAR entry of the CellSignalling Network (see Feedback etJ CSN access, entry 12).

-

_

•

entry 02 resume - - - - - - - -

Criteria for EXPRESSION sections This section should bring together information on expression patterns of the ion channel gene, indicating functional roles of specific channels in the cell type or organism. The complex and profound roles of ionic currents in vertebrate development (linking plasma membrane signalling and genome activation) are also emphasized within the fields of this section. Field number 08: Cell-type expression index: Comprehensive systems relating the expression of specified molecular components to specified anatomical and developmentalloci ('expression atlases') are being developed in a number of centres and in due course will form a superior organizational framework for this type information (see discussion below). In the meantime, the range of cell-type expression should be indicated in this field in the form of alphabetized listings. Notably, there is a substantial literature concerned with the electrophysiology of ion channels where the tissue or cell type forms the main focus of the work. In some cases, this has resulted in detailed lexpression surveys', revealing properties of interacting sets of ion channels, pumps, transporters and associated receptors. Such review-type information is of importanee when discussing the contribution of individual ion channel molecules to a complex electrophysiological phenotype t and/or overall function of the cell. For further references to 'cell-type-selective' reviews, see Resource H - Listings of cell types.

Problems and opportunities in listing ion channel molecules by cell type: Understanding the roles which individual ionic chcmnels play in the complex electrophysiological phenotypes of native t cells re:mains a significant challenge. The overwhelming range of studies covering aspects of ion channel expression in

vertebrate cells offers unique problems when eompiling a representative overview. Certainly the linking of specific ion channel gene expression to cell type is a first step towards a more comprehensive indexing, and towards this goal, cell-typeselective studies are useful for a number of reasons. First, they can help visualize the whole range of channel expression by providing an inventory of conductances t observed. Secondly, these studies generally define the experimental conditions required to observe a given conductance. Thirdly, they include much information directly relating specified ionic conductances to the functions of the cell type concerned. Collated information such as this should be of increasing utility in showing the relationship of electrophysiological phenotype to mechanistic information on their gene structure and expression-control (which largely correlates with cell-type lineage). At this time it is diffic.ult to build a definitive catalogue of ion channel gene expression patterns mapped. to cell type, not only because the determinants of gene expression are scarcely explored, but also because there remain many unavoidable ambiguities in phenotype dt~finition. Some of these problems are discussed below.

Problems of uneven coverage/omissions: Certain cell preparations have been intensely studied for ion channel expression while others have received very little attention for technical, anatomical or other reasons. Furthermore, a large number of native t ionic currents can be induced or inhibited by agonists t that bind to co-expressed G protein t -coupled receptors t. Thus a difficulty arises in deciding whether channel currents can be unambiguously defined in tenns of action at a separately encoded

-

1L..-.-_ en_t_ry_0_2_re_s_u_m_e

_

receptor protein. While it is valid to report that an agonist-sensitive current is expressed in a defined cell type, the factors of crosstalkr and receptor-transducert subtype specificities in signalling systems are complex and may produce an ambiguous classification. Receptor-coupled agonist-sensitivities are an important factor contributing to cell-pharmacological and -electrical phenotypet, but the treatment here has been limited to a number of tabular summaries of ion channel regulation through coupling to G protein-linked effectort molecules (see Resource A - G protein-linked receptors). As stated earlier, the entries are not sorted on agonist specificity except where the underlying ion channel protein sequence would be expected to form an integral ionic channel whose gating t mechanism is also part of the assembled protein complex.

Cell preparation methods are variable: A further problem inherent in classifying ion channels by their patterns of expression is that the choice of tissue or cell preparation method may influence phenotype t. The behaviour of channel-mediated ionic currents can be measured in native t cells, e.g. in the tissue slice, which has the advantages of extracellular ionic control, mechanical stability, preserved anatomical location, lack of requirement for anaesthetics and largely undisturbed intercellular communication. Cell-culture techniques show similar advantages, with the important exceptions that normal developmental context, anatomical organization and synaptic arrangements are lost and (possibly as a consequence) the 'expression profile' of receptor and channel types might change. Cultured cell preparations may also be affected by 'de-differentiationt, processes and (by definition) cell lines t are uncoupled from normal processes of cell proliferationt, differentiation t and apoptosis t. Acutely dissociated cells from native t tissue may provide cell-typespecific expression data without anomalies introduced by intercellular (gap junctional) conductances, but the enzymatic or dispersive treatments used may also affect responses in an unknown way. Verbal descriptions of cell-type expression divisions are arbitrary and are not rigorous: Definitive mapping of specific ion channel subtype expression patterns has many variables. Localization of specific gene products are most informative when in situ localizations are linked to the regulatory factors controlling their expression (see glossary entry on Gene expression-control t ). The complexity of this task can extend to processes controlling, for example, developmental regulation, co-expressed protein subunit stoichiometries and subcellular localizations. Complete integration of all structural, anatomical, co-expression and modulatory data for ion channels could eventually be accommodated within interactive graphical databases which are capable of providing 'overlays' of separately collected in situ expression data linked to functional properties of the molecules. By these methods, new data can be mathematically transformed to superimpose on fixed tissue or cell co-ordinates for comparison with existing database information. Software development efforts focused on the acquisition, analysis and exchange of complex datasets in neuroscience and mouse development have been described, and the next few years should hopefully see their implementation. For further information, see Baldock, R., Bard, J., Kaufman, M. and Davidson, D. (1992) A real mouse for your computer, Bioessays 14: 501-2

-

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try __ 02_r_es_u_m_e_

Bloom, F. (1992) Brain Browser, v 2.0. Academic Press (Software). Kaufman, M. (1992) The Atlas of Mouse Development, Academic Press Wertheim, S. and Sidman, R. (1991) Databases for Neuroscience, Nature 354: 88-9 To help rationalize the choices available for selection of these 'prototype' classifications, see Resource H - Listings of cell types. These listings may also have some practical use for sorting the subject matter of journal articles into functionally related groups. A proposed integration of information resources relating different aspects of cell-signalling molecule gene expression is illustrated in Fig. 4 of the section headed Feedback eiJ CSN access, entry 12. Field number 09: Channel density: This field should contain information about estimated numbers of channel molecules per unit area of membrane in a specified preparation. This field lists information derived from local patch-clamp 'sampling' or autoradiographic detection in membranes using anti-channel antibodies. The field should also describe unusually high densities of ion channels ('clustering') in specified membranes where these are of functional interest. Field number 10: Cloning resource: This field should refer to cell preparations relatively 'rich' in channel-specific mRNA (although it should be noted that many ion channel mRNAs are of low abundance t ). Otherwise this field defines a 'positive control' preparation likely to contain messenger T RNA t encoding the channel. Preparations may express only specific subtypes of the channel and therefore related probes (especially peRl probes) may not work. Alternatively, a genomic t cloning resource may be cited. Field number 11: Developmental regulation: This field should contain descriptions of ion channel genes demonstrated (or expected to be) subject to developmental gene regulation - e.g. where hormonal, chemical, second messengert or other environmental stimuli appear to induce (or repress) ion channel mRNA or protein expression in native t tissues (or by other experimental interventions). Protein factors in trans t or DNA structural motifs t in cis t which influence transcriptional activation t, transcriptional enhancement t or transcriptional silencing t should also be listed under this fieldname. Information about the timing of onset for expression should also be included if available, together with evidence for ion channel activity influencing gene activation t or patterningt during vertebrate development. Field number 12: Isolation probe: This field should include information on probes used to relate distinct gene products by isolation of novel clones following lowstringency cross-hybridization screens t. The development of oligonucleotidet sets which have been used to unambiguously detect subtype-specific sequences by PCRt, RT-PCRt or in situ hybridizationt should be identified with source publication. Both types of sequence may be able to serve as unique gene isolation probes, dependent upon the library t size, target abundance t, screening stringency t and other factors. Field number 13: mRNA distribution: This field should report either quantitative/ semi-quantitative or presence/absence (±) descriptions of specific channel mRNAs in defined tissues or cell types. This type of information is generally derived from

-

1L....._e _n_t_ry_O_2_r_e_s_u_m_e

_

Northern hybridizationt, RNAase protectiont analysis, RT-peRt or in situ expression assays. See also notes on expression atlases under Field number 08: Cell-type expression index. Field number 14: Phenotypic expression: This field should include information on the proposed phenotypet or biological roles of specified ion channels where these are discernible from expression studies of native t (wild-type) genes. Phenotypic t consequences of naturally occurring (spontaneous) mutationst in ion channel genes are included where these have been defined, predicted or interpreted (see also Fields 26-32 of the STR UCTURE etJ FUNCTIONS section for interpretation of site-directed mutagenesis t procedures as well as Resource D - IDiagnostic' tests). Associations of ion channels with pathological states, or where molecular 'defects' could be 'causatory' or contribute to the progression of disease should be listed in this field (for links with established cellular and molecular pathology databases, see Fig. 4 of Feedback etJ CSN access, entry 12). The Phenotypic expression field may include references to mutations in other ('nonchannel') genes which affect channel function when the proteins are co-expressed. It is also used to link descriptions of specific (cloned) molecular components to native cell-electrophysiological phenotypes. In due course, this field will be used to hold information on phenotypict effects of transgenict manipulations of ion channel genes including those based on gene knockoutt or gene locus I" replacementt protocols. Field number 15: Protein distribution: This field should report results of expression patterns determined with probes such as antibodies raised to channel primaryt sequences or radiolabelled affinity ligandst . Field number 16: Subcellular locations: This field should describe any notable arrangements or intracellular locations related to the functional role of the channel molecule, e.g. when the channel is inserted into a specified subcellular membrane system or is expressed on one pole of the cell only (e.g. the basolateral t or apical t face). Field number 17: Transcript size: This field should list the main RNA transcriptt sizes estimated (in numbers of ribonucleotides) by Northern t hybridization analysis. Multiple transcript sizes may indicate (i) alternative processing ('splicingt ') of a primary transcriptt, (ii) the use of alternative transcriptional start sitest, or (iii) the presence of 'pre-spliced' or lincompletely spliced' transcripts identified with homologous nucleotide probest in total cell mRNAt populations. Note that probes can be chosen selectively to identify each of these categories; 'full-length' coding sequencet (exonict) probes are the most likely to identify all variants, while probes based on intronic t sequences (where appropriate) will identify 'pre-splice' variants.

Criteria for SEQUENCE ANALYSES sections This section should bring together data and interpretations derived from the nucleic acid or protein sequence of the channel molecule. The symbol {PDTM} denotes an

-

_"--


illustrated feature on the channel monomer protein domain topography model, which is presented as a central figure in some entries for sequenced ion channels. These models are only intended to visualize the relative lengths and positions of features on the whole molecule (see the description for field number 30, Predicted protein topography). The PDTMs as presented are highly diagrammatic - the actual protein structure will depend on patterns of folding, compact packing and multi-subunit associations. In particular, the relative positions of motifs, domain shapes and sizes are subject to re-interpretation in the light of better structural data. Links to information resources for protein and nucleic acid sequence data are described in the Database listings field towards the end of each entry. Field number 18: Chromosomal location: This field should provide a chromosomal locust designation (chromosome number, arm, position) for channel gene(s) in specified organisms, where this is known. N'otes on interactive linking to gene mapping database resources appear under an option of the Cell-Signalling Network 'home page' (see Feedback eiJ CSN access, entry 12). Field number 19: Encoding: This field should report open reading framet lengths as numbers of nucleotides or amino acid residues encoding monomeric channel proteins (i.e. spanning the first A of the ATG translational start codont to the last base of the translational termination codont). The field should report and compare any channel protein length variants in different tissues or organisms. If considered especially relevant or informative, selected primaryt sequence alignments of different gene family members may appear under this field. Field number 2.0: Gene organization: This field should describe known intront and exon t junctions within or outside the protein coding sequence, together with positional information on gene expression-control t elements and polyadenylationt sites where known. Note: Functional changes as a result of gene expressioncontrol should be listed under the Developmental regulation field. Field number 2.1: Homologous isoforms: This field should indicate independently isolated and sequenced forms of entire channels which either show virtual identity or of such high homologyt that they can be considered equivalent should also appear in this field (but see note on percentage conservation values under Field number 28: Domain conservation). Isoformsr * of a channel protein can exist between closely related species or between different tissues of the same species (i.e. the same gene may be expressed in two or more different tissues, sequenced by two groups but named independently). Some tissue-specific variation may also result from alternative splicingt, yielding subtly distinct forms of channel protein. Since small numbers of amino acid changes may exist from individual-toindividual (as a result of normal sequence polymorphism t in populations) separate isolates may yield sequence isoforms which can be shown to be 'equivalent' by Southern hybridizationt procedures (see Field number 25: Southerns). * Note: In the entries of this book a restrictive definition of molecular identity (or

near identity) is used to define an isoformt. In this restricted sense, 'isoforms' would be expected to be the product of the same genet (or gene variant produced by, for example, alternative splicingt), and therefore have very similar or identical

11II

1'--_e_n_t_ry_O_2_r_e_s_u_m_e

-----'_

molecular constitutions and functional roles in specified cell types of closely related species. Comparative information on different gene familyt members or multiple variants affecting particular protein domainst may also be included under the Gene family and Domain conservation fields respectively. Field number 22: Protein molecular weight (purified): This field should state reported molecular weights estimated from relative protein mobilities using SDSPAGEt methods (e.g. following affinityt purification from native t or heterologous t cell membranes). Data derived from native t preparations generally includes the weight contribution from oligosaccharide t chains added during post-translational protein glycosylationt . In general, extracellular saccharider components of glycoproteinst may contribute 1-85 % by weight, ranging from a few to several hundred oligosaccharide chains per glycoprotein molecule. Field number 23: Protein molecular weight (calc.): This field should list the molecular weight of monomeric channel proteins equivalent to the summated (calculated) molecular weights of constituent amino acids in the reported sequence (e.g. derived from open reading frames t of cDNAt sequences). If 'calculated' molecular weights are less than 'purified' molecular weights (previous field) this may indicate the existence of post-translational glycosylationt on nativet expressed protein subunits in vivo. Field number 24: Sequence motifs: This field should report the position of putative regulatory sites as deduced from the protein or nucleic acid primaryt sequence (with the exception of potential phosphorylation sites for protein kinasest, which are listed under Field number 32: Protein phosphorylation). Positions of sequence motifst illustrated on the monomer protein domain topography model are denoted by the symbol [PDTM]. Typical consensust sites include those for enzymes such as glycosyl transferasest, ligand t -binding sites, transcription factort -binding sitest etc. N-glycosylationt motifs are sometimes indicated using the shorthand designation Ngly:. Signal peptide cleavage sites (sometimes designated by Sig:) can be derived by comparing sizes of the signal peptidet and the mature chaint. Field number 25: Southerns: This field should include information which reports the existence of closely related DNA sequences in the genomet or reports the copy numbert of individual genes via Southern hybridization t procedures. Note that nativet diploid somaticr cells will generally maintain two copies of a given ion channel gene locust, but stable t heterologoust expression procedures may result in mUltifle locus insertiont. Multiple locus insertion can be quantitated in Southern hybridization procedures using two probes of similar length and hybridization affinityt, one specific for a native locus (which will identify two copies) and one for the heterologous gene (which will yield a hybridization signal proportional to the copy number). Note also that the copy number parameter can not be equated to the physiological expression level of the recombinantt protein unless locus control regions are incorporated as part of the channel expression construct (for details, see the section entitled Gene copy number under Resource D - IDiagnostic' tests, and the section describing heterologous ion channel gene expression under Resource H - Listings of cell types).

-

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try __ 02_r_es_u_m_e_

Criteria for STRUCTURE

&,

FUNCTIONS sections

This section should bring together information based on functional analysis or interpretation of ion channel structural elements. This section includes data derived from functional studies following site-directed mutagenesis t of ion channel genes and molecular modelling studies at atomic scale. Future developments linking on-line information resources for protein structure to 'functional datasets' are illustrated in Fig. 5 of Feedback etJ CSN access, entry 12, and in Resource J Search criteria etJ CSN development. Field number 26: Amino acid composition: This field should include information on channel protein hydrophilicityt or hydrophobicityt where this is of structural or functional significance. Similarities to other related proteins should be emphasized. Field number 27: Domain arrangement: This field should describe the predicted number and arrangement of protein domainst when folded in the membrane as determined by hydropathicity analysist of the primaryt sequence. Note that structural predictions of transmembrane domainst on the basis of hydrophobicityt plots may be misleading and prematurely conclusive. For example, high resolution ( rv 9 A) structural studies of the nicotinic acetylcholine receptor (nAChR, see ELG CAT nAChR, entry 09) predict that only one membrane-spanning a-helixt (likely to be M2, a pore-lining domain) is present per subunit, with the other hydrophobic regions being present as /1-sheets t (see Unwin, J Mol Biol (1993) 229: 1101-24). By contrast, extracellular ligand-gated (ELG) channels such as the nAChR display four predicted membrane-spanning regions (MI-M4) on the basis of hydrophobicity plots. From the foregoing it must be emphasized that all assignments given for

the number or arrangement of 'predicted' domains in this field are tentative. Field number 28: Domain conservation: This field should point out known structural and/or functional motift sequences which have been conserved as protein subregions of ion channel primaryt sequences during their evolution (such as those encoding a particular type of protein domain t ). Cross-references should be made to functionally related domains conserved in different proteins including 'non ion channel' proteins. Note that 'percentage conservation' values are not absolute as they depend on which particular subregions of channel sequences are aligned, the numbers and availability of samples, and/or which sequence alignment algorithms t are used. Field number 29: Domain functions (predicted): This field should indicate predicted functions of channel molecular subregions based on structural or functional data e.g. regions affecting properties such as voltage-sensitivity, ionic selectivityt, channel gating t or agonisti binding. Field number 30: Predicted protein topography: This field should include information on the stoichiometric t assemblyT patterns of protein subunits derived from the same or different genes. This field indicates whether channel monomers are likely to form homomultimerst, heteromultimerst or both, and lists estimated physical dimensions of the protein if these have been published. Note: 'topography' is a convenient term borrowed from cartography which when applied to proteins, implies a 'map' at a level of detail or scale intermediate between that

--

l_e_n_t_ry_O_2_re_s_u_m_e_"

_

of an amino acid sequence and a larger-scale representation such as a protein multimeric complex. Topographic maps (or 'models') are therefore particularly useful for displaying selected sets of (inter-related) datatypes within a single 'visual framework'. The protein domain topography models (symbolized by [PDTM] throughout the entries) provide prototypes for this form of data representation. The considerable scope for further development of 'shared' topographical models which interactively report and illustrate many different features in the text are described in Search Criteria etJ CSN Development (Resource J). The terms 'protein topography' and 'protein topology' are often used interchangeably (sic), but the latter should be reserved for those physical or abstract properties of a molecule which are retained when it is subjected to 'deformation'. Field number 31: Protein interactions: This field should report well-documented examples of the channel protein working directly in consort with separate proteins in its normal cellular role(s). The 'protein interactions' described need not involve physical contact between the proteins (generally referred to as 'protein-protein' interactions), but may involve a messenger t molecule. The scope of this field therefore includes notable examples of protein co-localization or functional interaction. For instance, reproduction of nativet channel properties in heterologous t cell expression systems may require accessory subunit expression (e.g. see VLG K Kv-beta, entry 47). Common channel-receptor or G protein-channel interactions are described in principle under Resource A - G protein-linked receptors, and Field number 49: Receptor/transducer interactions. Field number 32: Protein phosphorylation: This field should describe examples of experimentally determined 'phosphomodulation' of ion channel proteins, and if possible list sites and positions of phosphorylation motifst within the channel sequence. Only those consensus sitest explicitly reported in the literature are shown, and these may not be a complete description and may not be based on functional studies. Examples of primaryt sequence motifst for in vitro phosphorylation by several kinasest are listed in Resource C - Compounds etJ proteins and Resource G - Reported tConsensus sites' and tmotifs'. Abbreviations used within this field for various enzyme motifst (e.g. Phos/PKA) are listed in Abbreviations, entry 03. Electrophysiological or pharmacological effects of channel protein phosphorylation in vitro by use of purified protein kinasest should also be described or cross-referenced in this field.

Criteria for ELECTROPHYSIOLOGY sections This section should bring together information concerning the electrical characteristics of ion channel molecules - how currents are turned on and off, which ions carry them, their sensitivity to applied membrane voltage or agonists, and how individual molecules contribute to total membrane conductance in specified cell types. Field number 33: Activation: This field should contain information on experimental conditions or factors which activate (open) the channel, such as the binding of ligandst, membrane potential changes or mechanical stimulation. Descriptions of characteristic gating f behaviour such as flickering t, bursting t, activation latency t

_'---


or thresholdt of opening are also included. Applicable models of activation and the time course of current flow are briefly described here or referred to Field number 38: Kinetic model. Field number 34: Current type: Where clarification is required, this field should contain general descriptive information on the type, shape, size and direction of ionic current. Field number 35: Current-voltage relation: This field should report the behaviour of the channel current passed in response to a series of specified membrane potential shifts from a holding potentialt under a specified recording configurationt. For ligand t -gated channels (i.e. those with sortcodes beginning ELG and ILG) entries should report the current evoked by specific concentrations of agonistt applied at various holding potentials. This field should attempt to illustrate channel behaviour by listing a range of parameters such as slope conductancet, reversal potentialst and steepnesst of rectifyingt (non-ohmict) behaviour. The conventions used for labelling the axes of I-V relations for different charge carrierstare outlined in the on-line glossary. Field number 36: Dose-response: This field should contain information relating activator 'dose' (e.g. concentration) to channel 'response' parameters (e.g. open timet , open probabilityt) and whether there are maxima or minima in the response. Agonistt dose-response experiments are used to derive parameters such as the Hill coefficient t and Equilibrium dissociation constant t . Field number 37: Inactivation: This field should describe any inactivationt behaviour of the channel in the continued presence of activating stimulus. The

field includes information on voltage- and agonistt -dependence, with indications of time course and treatments which extend or remove the inactivation response. Where known, this field will distinguish channel inactivation from receptor desensitizationt processes, which are of partic.ular significance for the extracellular ligandt -gated (ELG) channel types (see ELG Key facts, entry 04). Field number 38: Kinetic model: This field should contain references to major theoretical and functional studies on the kinetic behaviour of selected ion channels. The field contents is limited to a simple description of parameters, terms and fundamental equations. Field number 39: Rundown: This field should collate information on channel 'rundownt, ('washout') phenomena observed during whole-cellt voltage clampt / cytoplasm dialysist or patch-clampt experiments. Conditions known to accelerate or retard the development of rundown should also be listed. Field number 40: Selectivity: This field should report data on relative ionic permeabilities t under stated conditions by means of permeability ratio t and/or selectivity ratio t parameters. The field may also compare measured reversal potentials t in response to ionic equilibrium potentials t with specified charge carriers under physiological conditions. This field also lists estimated physical dimensions of ionic selectivity filterst where derived from ion permeationt or electron micrographic studies.

e_n_t_ry_O_2_r_e_s_u_m_e

1__

_

Field number 41: Single-channel data: This field should report examples of singlechannel current amplitudes and single-channel conductances t measured under stated conditions. In the absence of authentic single-channel data, estimates of channel conductances t derived from whole-cell recording t and fluctuation analysis t may be listed.

Field number 42: Voltage sensitivity: This field should describe the behaviour of the channel in terms of parameters (e.g. Popen t) which are directly dependent upon applied membrane voltage. A distinction should be made between 'voltage sensitivity' resulting from intrinsic voltage-gating t phenomena (i.e. applicable to channels possessing integral voltage sensors r) and indirect effects of applied membrane voltage influencing general physical parameters such as electrochemical driving force t .

Criteria for PHARMACOLOGY sections This section should bring together information concerning pharmacological or endogenous modulators of ion channel molecule activity. Regulatory cascades in cells may simultaneously activate or inhibit many different effector proteins, including ion channels. Analysis of patterns of sensitivity to messengers t and exogenous compounds can help elucidate the molecular signalling pathway in the context of defined cell types.

Field number 43: Blockers: This field should list compounds which reduce or eliminate an ionic current by physical blockade of the conductance t pathway. The field should include notes on specificity, sidedness and/or voltage sensitivity of block, together with effective concentrations and resistance to classes of blockers where approrriate. Where sites of block have been determined by site-directed mutagenesis , these should be cross-referenced to Domain functions, field 29.

Field number 44: Channel modulation: This field should summarize information on effects of important pharmacological or endogenous modulators, including descriptions of extracellular or intracellular processes known to modify channel behaviour. Loci of modulatory sites on the channel protein primary t sequence (as determined by site-directed mutagenesis t procedures) should be cross-referenced to

Domain functions, field 29.

Field number 45: Equilibrium dissociation constant: This field should list published values of Kd for agents whose concentration affects the rate of a specified process. See also on-line glossary entry for equilibrium dissociation constant t .

Field number 46: Hill coefficient: This field records calculated Hill coefficients t of ligand t -activated processes. The Hill coefficient (n) generally estimates the minimum number of binding/activating ligands although the actual number could be larger. For example, a Hill coefficient reported as n ~ 3 suggests that complete channel activation requires co-operative binding of at least four ligand molecules (e.g. see ILG CAT cGMp, entry 22). See also Field number 36: Dose-response. Field number 47: Ligands: This field should include principal high-affinity radioligands t which have been used to investigate receptor-channel function and that

--

_L...-

e_n_t _ry_O_2_r_e_su_m_e_

are commercially available. Note that numbers of ligand t -binding sites cannot be equated to functional receptors because they only indicate the presence of a ligand-binding entity that may not necessarily be linked to an effectort moietyt. Field number 48: Openers: This field should list compounds (or other factors) which increase the open probabilityt (Popen ) or open timet of the channel in native t tissues. Field number 49: Receptor/transducer interactions: This field should briefly discuss known links to discrete (i.e. separately encoded) receptor and G protein molecules (see also Resource A - G protein-linked receptors, accessible via the CSN). Types of lreceptor/transducer/channel' interactions account for many of the physiological responses of ion channel molecules within complex signalling systems. Note: Many pharmacological agents acting at receptor or transducer proteins (beyond the scope of these entries, but see Watson, S. and Arkinstall, S. (1994) The GProtein Linked Receptor FactsBook. Academic Press, London) partially exert their biological effects because these receptor/transducers have ion channel molecules as an ultimate effectort protein. Field number 50: Receptor agonists (selective): For the extracellular ligandt -gated (ELG) receptor-channels, this field should list compounds which selectively bind to the ligand receptor fortion of the molecule and thereby increase the open timet, open probability or conductancet of the integral channel. Antagonistst should be categorized as competitivet, non-competitiver or uncompetitivel where this has been determined. Field number 51: Receptor antagonists (selective): This field should list agents that selectively bind to the ligand t receptor portion of integral receptor-channel molecules but do not activate a response. Field number 52: Receptor inverse agonists (selective): This field should list compounds which selectively bind (extracellular ligand-gated) receptor-channels but which initiate an opposite response to that of an agonistt, i.e. tending to reduce the open timet, open probabilityt or conductancet of the integral channel.

Criteria for INFORMATION RETRIEVAL sections This section should provide links to other sources of information about the ion channel type, particularly accession to sequence database, gene expression, structure-function and bibliographic resources operating over the Internet or available on CD-ROM. A full discussion of the potential scope for integration of these resources with molecular-based entries appears in Resource T- Search criteria etJ CSN development. Brief details are given in Feedback etJ CSN access, entry 12. Field number 53: Database listings/primary sequence discussion: This field should tabulate separately listed items of relevance to the channel type and may include 'retrieval strings' such as locus names, accession numbers, keyword-containing identifiers and other miscellaneous information. Note that terms used by databases are often abbreviated (e.g. K for potassium, Na for sodium etc., therefore only specific identifiers (such as the accession numbers, locus and author names)

-

en_t_ry_O_2_re_s_u_m_e

L.....-_

_

should be used for retrieval. The actual names and numbers quoted have been sourced from NCBI-GenBank® (prefixed gb:) or EMBL (prefixed em:). Since there is now a high concordance between the contents of the EMBL and NCBI-GenBank® nucleic acid databases, the NCBI-GenBank® accession numbers given should retrieve the information from either database. Note that in all of the Database listings sections, the lower case prefixes are not part of the locus name or accession number, but merely indicate the relevant database. Sources of pre-translated protein sequences are indicated by references to the following databases (given in alphabetical order following the NCBI-GenBank® nucleic acid reference): SWISSPROT (prefixed sp:), Protein Identification Resource (prefixed pir:). The journal-scanning component of GenBank uses the NCBI 'Backbone' database (prefixed bbs: for backbone sequence, composed of several individual sequence segments; bbm: for backbone molecule) - these are maintained by the NCBIT (National Center of Biotechnology Information).

General notes on sequence retrievals: Updating and error-correction procedures for public domain databases may modify a protein or nucleic acid sequence (retrievable by a given accession number) between releases of a database. Thus, two users performing an analysis on a given database record may come to different conclusions depending upon which release was used. Note also that (i) accession numbers sometimes disappear with no indication of whether a new record has replaced the old one, (ii) multiple databases sometimes each give a different accession number to a single record, and (iii) some databases do not respect the ranges of accession numbers 'reserved' by other databases. Although the 'traditional' format of accession numbers has been a letter followed by five digits (with a maximum space of 2.6 million identifiers), the rapid rate of sequence accumulation will eventually force a different format to be used. Because of these problems, the NCBI now uses unique integer identifiers (UIDs) to identify sequence records and encourages their use as the 'real' accession numbers for sequence records. Reference numbers prefixed 'gim' can be read from CD-ROM media, but only refer to a 'GenInfo Import ill' - a temporary identifier unique only to a given release of the CD-ROM compilation (such as a numbered release of Entrez - see below). Should a sequence supplied by a database change, the record will usually be allocated a new 'gim' number, but the old one will still be available under its UID from the ill database. Because of the transient nature of 'gim' identifiers, they are not recommended as search/retrieval parameters and are generally not listed in the Database listings field (except where an accession number proper has not been found). In compiling The Ion Channels FactsBook, extensive listings of aligned protein or nucleic acids to show sequence relatedness have been avoided in some entries (as these were judged to be best served by development of on-line data resources specializing in sequence alignments - for a prototype, see Hardison et al. (1994) Genomics 21: 344-53). Alternatively or in addition, alignments can be performed according to need by dedicated sequence-manipulation software. Presently available compilations of sequences (e.g. Entrez can perform powerful'neighbouring t analyses' based on pre-computed alignments of any sequence against the remainder of the existing database. Establishment of homologous alignments t can proceed by finding a match between the query sequence and any member of the 'neighbouring set'. In practice, comprehensive retrievals can be performed interactively by just

-

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try __ 02_r_es_u_m_e_

one or two rounds of neighbouring analysis. As indicated at the beginning of each Database listings field, the range of accession numbers provided can be used to initiate relevant searches, but following on from this, neighbouring analysis is strongly recommended to identify newly reported and related sequences. Descriptions of features based on primaryt sequence data listed within fields of the SEQUENCE ANALYSES or STRUCTURE &. FUNCTIONS sections can be more readily interpreted if an interactive sequence analysis program is available. Electronic mail servers t at the NCBI can receive specially formatted e-mailt queries, process these queries, and return the search results to the address from which the message was sent out. No specific password or account is needed for these, only the ability to send e-mail to an Internet t site. For local searches, alignment programs such as BLAST can also be retrieved by anonymous file-transfer protocol l or FTP. Detailed information on interactive linking to remote nucleic acid and protein database resources will appear under an option of the Cell-Signalling Network 'home page' (see Feedback eiJ CSN access, entry 12). Accession numbers can be issued for newly submitted sequences (normally within 24 hours) by remote Internet connection or by formatting/submission software (e.g. Seqwin, obtainable from the NCBI using an anonymous l FTPt). NCBI-GenBank® can also be accessed over the World Wide Webt (http://www.ncbLnlm.nih.gov). Sample retrievals in the absence of a CD-ROM resource: For a nucleic acid sequence from the EMBL database, use the e-mailt address below exactly as shown, specifying the appropriate accession number (nnnnnn) by the GET NUC command. For example, a database entry can be automatically e-mailed to you by the EMBL servert : [email protected] GET NUC:nnnnnn An analogous procedure can be used to retrieve protein sequences from the

SWISSPROT database, substituting the GET NUC: command with GET PROT:. Nucleic acid sequences from NCBI-GenBank® can be retrieved using the servert at the NCBI. In this case, send an e-mailt message to the service (address below) specifying the name of the database, the command BEGIN and. the accession numbers or key words. A sample request is shown below for an accession number nnnnnn: [email protected] DATALIB genbank BEGINnnnnnn Protein sequences from the Protein Identification Resource (pir:) can be obtained using an e-mailt request containing the command GET followed by the database code. The database code is distinct from the accession number but can be obtained by typing the command ACCESSION and then the number. For example, to specify a request for an entry of database code XXXX containing the accession number nnnnnn, you would send an e-mail IIlessage as follows: [email protected] GET XXXX ACCESSION nnnnnn

'---_en_t_ry_O_2_re_s_u_m_e

_

General information on using these file serverst can be obtained using the above e-mailt addresses followed by the single command HELP. The Database listings tables contain short-form references to original research articles which have discussed features of the channel protein and/or nucleic acid primaryt sequence(s). Sequences are retrievable with the specified accession number or the author name shown in the short-form reference.

Field number 54: Gene mapping locus designation: This field should list references to human gene mapping locit using terms defined by a human genome mapping workshop (HGMW)I convention where possible. Notes on interactive linking to gene mapping database resources appears under an option of the Cell-Signalling Network 'home page' (see Fig. 4 of Feedback etJ CSN access, entry 12). The opportunities for linking to a wide range of genetic information resources are discussed in Resource T- Search criteria etJ CSN development. Field number 55: Miscellaneous information: This is a 'catch-all' field used within the entry to reference relevant peripheral information or perspectives on the channel molecule or its function. This field also should be used to contain information about ion channels showing partial functional relatedness to those in the main entry, but which also possess some features indicating the expression of a distinct genet (for example, description of potassium-selective ligand-gated t channels within an entry describing non-selective cation channels gated by the same ligand t , or vice versa). Normally, ion channels with distinct properties are covered in 'their own' entry whenever there is sufficient information available to make a clear set of 'defining characteristics'; the Miscellaneous information field therefore encompasses those channels which either have been infrequently reported, show only minor variations with the channel type under description, or are otherwise beyond the scope of the (present) collection of (largely) vertebrate channel-type entries. Field number 56: Related sources eY reviews: For reasons of space, the FactsBook cannot provide citations for every 'fact' within individual entries. Citations within this field should provide a starting point for locating key data through major reviews and other primaryt sources where these have been quoted extensively within the entry. A full discussion of how future entries will be linked to established on-line bibliographic resources appears in entry 12 and Resource J. Field number 57: Feedback: Information supplementary to the entries will be accessible from the CMHT server following publication of the complete book series (see below). An aim in compiling this book is that the scope and arrangement of the information should, in time, be refined towards containing what is most useful, authoritative and up-to-date: Feedback from individual users is an essential part of this process. The Feedback field identifies the appropriate address for e-mail feedback of significant corrections, omissions and updates for the contents of a specified entry and fieldname. Comments regarding new or modified field categories (or supplementary reference-type material for incorporation into entries and appendices) would also be most welcome from users (for details on accessing entry updates via the Cell-Signalling Network, see Feedback etJ CSN access, entry 12).

-

_ entry 02 resume ~ - - - - - - -

In-press updates: (inserted at appropriate points): This field has been used occasionally (at the most relevant points in the printed versions of the book) to index publications containing important (direct) evidence which may significantly alter several statements or conclusions in the 'finalized' entry as sent to the publishers. It is acknowledged that no 'book-form' information index can ever be completely up-to-date, and it is in the nature of scientific progress that 'interpretations' based on reported 'facts' may change considerably in the light of additional or more direct experimental approaches to a problem. Using literature 'neighbouring' techniques, on-line companion entries enable users to be directed towards citations containing the 'latest' interpretations (or important 'additional facts'). The pace of change across all of the fields touched-on by the FactsBook means that 'specialists' in a given area can help 'speed-up' this indexing process by e-mailtnotificationwhere.reinterpretation' is justified (see Feedback etJ CSN access, entry 12, and Resource J). According to the original aims and 'philosophy' of the project, the entries will probably never be 'complete' as such. More appropriately, the framework will continue to evolve towards one which is hopefully more useful, authoritative, and able to comprehensively relate tconsensus' knowledge on ion channel molecular signalling.

Criteria for REFERENCES sections This section should contain 'short-form' references for numbered citations within the entry. For textbook coverage, refer to the Book references listed under Related sources and Reviews (field 56), Resource E - Ion channel book references, Resource F Supplementary ion channel reviews and Resource H - Listings of cell types. Plans for 'hyperlinking' to full bibliographic databases within the CSN framework are

described in Resource J - Search criteria etJ CSN development.

Criteria used for compilation of supporting computer-updatable resources The following reference appendices are referred to within the text and figures of the main entries. Updated versions of these files will be accessible via the thorne page' of the Cell-Signalling Network from January 1999 - for further details, see Feedback etJ CSN access, entry 12 and Resource J - Search criteria etJ CSN development. Resource A - G protein-linked receptors: A large number of ion channels are regulated as part of signalling cascades initiated by activation of G protein-coupled receptor proteins. This appendix should describe the basic principles associated with this type of regulation, limiting descriptions to those most relevant to ion channels. Tabulations of known receptort and G proteint molecules should form a framework of possible regulatory mechanism.s based on specific protein subtypes. The entry may clarify or suggest likely interactions between receptors, transducerst (e.g. G proteins) and ion channel molecules described under the fieldnames Developmental regulation, field 11, Protein interactions, field 31, Protein phosphorylation, field 32, Channel modulation, field 44, and Receptor/transducer interactions, field 49.

-

_en_t_ry_O_2_re_s_u_m_e

1.....

_

Resource B - 'Generalized' electrical effects of endogenous receptor agonists: This resource should present a 'visual summary' of general patterns of agonistt induced ionic current fluxes that have been reported across a large number of studies, predominantly in the central nervous system. By summarizing patterns of channel modulation and gating, Resource B may help to indicate whether receptort agonists tend to act in an excitatoryt or inhibitoryt fashion 'or both'. Resource B forms a prototype for 'mapping' textual and bibliographic information to hyperlinked t visual frameworks. Resource C - Compounds etJ proteins: Compounds and proteins mentioned in the entries which are commonly used to investigate ion channel function and modulation should be listed, including those used to analyse interactions with other cell-signalling molecules. In general, only frequently reported compounds which are commercially available are described in this appendix. Resource D - 'Diagnostic' tests: This appendix is intended to be a 'field-referenced' listing of common experimental manipulations used to 'implicate or exclude' the contribution of a given signalling component or phenomenon associated with ion channel signalling. For the most part, these approaches use the pharmacological tools listed under Resource C, but may also include sections describing common molecular biological and electrophysiological 'diagnostic' procedures. Resource E - Ion channel book references: This appendix should list details of published books which have addressed themes in ion channel biology or closely related topics. These references complement those of the main entries, which are almost entirely based on citations from scientific journals. Resource F - Supplementary ion channel reviews: The ion channel literature contains a large number of useful 'minireviews' which summarize the development of defined subjects and which do not necessarily fall into a single channel 'molecular type' category. This appendix should therefore list these 'supplementary' sources, indexed by topic. Updated Ire-writes' of subject reviews covering similar areas may replace earlier listings. Note: Subject reviews dedicated to aspects of an ion channel type or family can usually be found under the Related sources eiJ reviews field of appropriate entries. 'Topic-based' reviews making reference to the basic properties in the 'molecular type' entries are planned for expansion within the CSN framework. Resource G - Reported 'Consensus sites' and 'motifs': Based on extensive analysis of primaryt sequences and determination of substrate specificities for various enzymes1 a number of 'consensus' recognition sequences for post-translational modification I of proteins (including ion channels) have been determined. While these sites are not absolute, they can be highly conserved across whole families of ion channel proteins and in many cases (e.g. following phosphorylation) can lead to profound changes in ion channel function. However, the presence of 'consensus' sites or motifst (or even demonstrations of substrate specificity in vitro) does not necessarily prove that such modifications operate in vivo. This appendix should list 'consensus' motifs that are well-characterized, giving examples of 'authentic' sites for comparison. This appendix also contains references to subsets of

--

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_O_2_r_es_u_m_e

----J

consensust sites from genomic DNA sequences associated with mechanisms of ion channel gene expression-control t (e.g. in trans t protein factors which act at DNA structural motifs t in cis t, influencing transcriptional activationt, transcriptional enhancementt or transcriptional silencing t of ion channel genes t ).

Resource H - Listings of cell types: Studies of ion channels within the context of cell-type function often reflect 'recruitment' of selected genes from the genomet in a cell-developmentallineaget . Because of this, similar 'sets' of ion channel molecules can often be observed in cell types with broadly similar functions. This appendix should describe a framework for describing how integrated sets of ion channel molecules (and their associated signalling components) have co-evolved for specific functions in terminally differentiatedf cell-types. To begin with, a tentative classification of functional cell types should be employed, used to cross-reference 'surveys' of ion channel expression wherever possible. This appendix should also contain available information pertaining to efficient and appropriate heterologous expression of ion channel genes in selected cell types, as this is often a limiting factor in biophysical characterization of clonedt ion channel cDNAt or gene products. Resource I - Framework of cell-signalling molecule types: The flow of information into, within and between cells (signal transduction) generally depends on a multiplicity of co-expressed cell-signalling molecules which provide 'measured' responses to stimuli. Communication between different cellular compartments (e.g. between the cytoplasm and the nucleus) often requires 'interconversion' or 'transduction' of chemical, electrical (ionic), metabolic and enzymatic signals, with receptors and ion channels playing key roles in transducing such stimuli. For example, the 'activation' of signal transduction molecules such as kinases t or transcription factors t appear to 'sense' 'activated' conditions which resembles Ca2 +-, voltage- or ligandt -gating phenomena commonly observed for ion channels. These modes of protein activation t probably have rnany features in common, and understanding their interrelationship has important consequences for comprehending fundamental links between receptor signalling, cell activation and gene expression. To facilitate integration of information between these diverse fields of study, this appendix should provide a preliminary listing of signal transduction molecules, with some consideration of their inter-dependency in the 'activated' state. By making a rational 'connection' between activation of receptors, ion channels, enzymes and other effector t proteins, it is hoped that SOlne general principles will emerge on the electrical- and ligand t -control of complex cell phenotyrest (such as those affecting the cell cycle t, cell proliferationt, cell differentiation and apoptosist). The importance of ion channel activation (and activation of receptor/G protein transducerst which modulate ion channel activity) in other fundamental cell processes such as signal transmission/amplification, secretion (multiple forms), muscular contraction, endocytosist (and other cellular 'uptake' phenomena), sensory transduction (all types), cell volume control/osmotic responses, mechanotransduction (various forms), membrane potential control (multiple modes) and developmental compartment formation are well-documented and multiple examples appear in several fields, notably Developmental regulation, field 11, Phenotypic expression, field 14, Domain functions, field 29, Protein interactions, field 31, Protein phosphorylation, field 32 and Channel modulation, field 44.

--

1....._e_n_t_ry_O_2_re_s_u_m_e

_

Resource T- Search criteria &J CSN development. The framework of database entries which form the basis of The Ion Channel FactsBook were derived by 'scanning' primary research articles and reviews appearing in a set list of 'principal' journals dealing with ion channel and receptor signalling. A disadvantage of 'journal scanning' by 'keyword' is that search terms used are often ambiguous, and contextual or unconventional grammatical usage of keyword terms within articles often results in failure of specific retrieval. To circumvent this problem, this appendix should suggest new unique embedded identifiers (UEIs) which when specified by authors in the keywords section of submitted articles should ensure appropriate electronic retrieval from the primary literature. The adoption of finalized 'UEI' codes should be open to debate, but may eventually incorporate ion channel gene locus names (see field 54) where these are established by international convention. In the mean time, prototype UEIs will be used in the CSN pages to 'tag' new/backlogged article citation lists that contain information 'affecting' an existing or new entry. This also affords a simple 'queuing' mechanism for 'future coverage' of articles that have appeared or have been 'missed' during the main entry compilations (e.g. as notified in the feedback files, field 57). In due course, individual articles will be analysed and 'field referenced' within entry/fieldname update documents for downloading/printing. Prototype UEIs for each entry (distinguishing studies on native versus heterologous cells) are indicated in Category (sortcode), field 02. Updating by 'journal scanning' will proceed retrospectively from the 'most recent' articles to eventually 'overlap' with articles cited in the printed entries. In general, UEIs will be used to attempt systematic keywording/retrieval based on molecular criteria: the central principle of unique embedded identifiers is that they can 'automatically' find articles on topics of interest (in for example literature scans). Coupling to an 'expansion' section with further search terms in a conventional order will help enormously in data compilation/consolidation processes on strictly defined subjects within 'validated' databases. These issues will be discussed in the Resource 1 documents. Finally, Resource J should act as a forum for discussing limitations of data representation when comparing ion channel properties and suggest improved methods for facilitating information exchange (including graphical resources), diagnostic conventions, resolution of lcontroversial' results, and identification of areas or highly focused topics requiring consolidation/extension of knowledge. The importance of standardized computer software compatible with Internet t -mediated communication should be emphasized (see also Feedback eiJ CSN access, entry 12). Contents organization within each 'specialist' field of the FactsBook gives further opportunities for comparative data analysis. In due course, the -zz term of the xx-yy-zz index number will be used to indicate such structured information.

Criteria used for selection of on-line glossary and index items Consolidated versions of the FactsBook support glossary (i.e. extensions, updates and corrected items) are accessible from the Cell-Signalling Network 'home page' (see Feedback eiJ CSN access, entry 12). Index of on-line glossary items [ t]: To avoid unnecessary duplication of definitions within the text and to provide assistance to readers unfamiliar with a field, the

-

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_t_ry_O_2_r_es_u_m_ e

-----l

on-line glossary should provide short introductions to technical terms and concepts. Throughout the text, cross-references to the on-line glossary items are shown by means of a dagger symbol t .

'Self-indexing' in The Ion Channel FactsBook, volumes I to IV For the most part, The Ion Channels FactsBook should be 'self-indexing': 1. Locate the channel 'molecular type' by sortcode, table of contents or the

Rubrics at the rear of the book. 2. Go to the appropriate section (NOMENCLATURES, EXPRESSION, SEQUENCE ANALYSES, STRUCTURE & FUNCTIONS, ELECTROPHYSIOLOGY, PHARMACOLOGY, INFORMATION RETRIEVAL or REFERENCES). 3. Look under the most appropriate fieldname (as described by the criteria above). Further 'structuring' will arise in due course, when more data are entered (see previous sections). The subject index should also allow the initial location of entries through alternative names of channels, associated signalling phenomena or commonly reported properties. Electronic cross-relation of topics is intended to be a development focus of the CSN, exploiting the principle of hyperlinking between database files stored in 'addressible' loci. For further details on how this might be achieved, see Resource T- Search criteria eiJ CSN development.

Feedback: Comments and suggestions regarding the scope, arrangement and other matters relating to this introduction can be sent to the e-mail feedback file [email protected]. (see field 57 of most entries for further details)

Abbreviations For most abbreviations of compound names in use, refer to the Resource C Compounds etJ proteins, entry 58, as well as the FactsBook entries. Abbreviations for ion channel currents are listed under the Current designation field of each entry. Terms marked with a dagger symbol appear in the on-line glossary section OCa2+ 5-HT 7TD

Ca 2+-free solution 5-hydroxytryptamine; serotonin 7 transmembrane domains

A

AV AVN

ampere t amino acidt after-hyperpolarizationt 4-aminopyridine action potential duration t atrio-ventricular atrio-ventricular nodet (of heart)

BKCa BP bp

large ('big/)-conductance calcium-activated K+ channels blood pressure base pairs t

C C-terminal CjA or C-A Ca(mech) or Camech CAv cds CF CICR CI(Ca) or Clca CNG CNS COOH CRC cRNA CTK Cx or Cxn

coulombt carboxylt terminalt (of protein) cell-attachedt (recording configuration) mechanosensitivet Ca2+ channel voltage-gated t Ca2+ channels codingt sequence (used in GenBankt® entries) cystic fibrosis t calcium-induced calcium-release calcium-activated chloride channel cyclic-nucleotide-gated (channels) central nervous system t carboxyl groupt calcium release channels complementaryt RNA cytoplasmic tyrosine kinase t (cf. RTK) connexin

Da Dephosjenzyme

daltons putative (consensus t ) site for dephosphorylation t by a specified enzyme, e.g. DephosjPP-I: endogenous protein phosphatase-I; DephosjPP-2A: protein phosphatase-2A dihydropyridine receptor Duchenne muscular dystrophyt depolarizing post-synaptic potentialt

aa AHP 4-AP APD

DHPR DMD DPSP

E EAA

E-C

potential differencet, inside relative to outside excitatory amino acidt excitation-contractiont

_L..-ELG Em EMBL EMF EPP EPSP ER Erev F

F fS

G g

G/Gmax gb: gj, Gj or GO) HGMW HH h.p. HVA

e_n_try_0_3_re_s_u_m_e_'_

500/0 effective concentration equilibrium potential t for K+ ions (analogous nomenclature for other ions) extracellular ligandt -gated (as used in FactsBook sortcode) membrane potential t European Molecular Biology Laboratoryt electromotive forcet endplate t fotentialt excitatory post-synaptic potentialt endoplasmic reticulumt reversal potentialt faradt Faraday's constantt femtosiemens (10- 15 Siemenst) conductancet conductance (unit - Siemenst, formerly reciprocal ohmst or mho t ) peak conductancet designation for GenBank® accession numbert gap-junctional conductancet Human Gene Mapping Workshopt after Hodgint -Huxleyr holding potential high voltage activated

InsP3R or IP3R IPSC IPSP ISH

current t subscript abbreviation for intracellular peakt currentt inside-outt (patcht, recording configurationt) concentration which gives 500/0 of maximal inhibition effect in a dose-inhibition response curvet intracellular ligandt -gated (as used in FactsBook sortcodest) maximal currentt collective abbreviation for inositol polyphosphatest - e.g. InsP3,InsP4 inositol l,4,5-trisphosphate-sensitive receptor-channel inhibitoryt post-synaptic currentt inhibitoryt post-synaptic potentialt in situ hybridization

JCC

junctional channel complex

k KA or K(A) kb KCa, Kca or K(Ca)

Boltzmann's constantt A-typet K+ channels kilobasest (kbp - kilobase pairs or bp x 103) calcium-activated K+ channels

I i

I/Imax I/O or 1-0 ICso

ILG I max

InsP(x) or InsPx

-

1__e_n_t_ry_0_3_r_e_su_m_e

K1(0) KIR or K(IR) K(mech) or Kmech

knt

Kv LTD

LTP LVA mAChR MARCKS Mb MEPC MDa MH Mr

mRNA mV

_

equilibrium dissociation constantt kilodaltonst (daltons x 103 ) equilibrium dissociation constantt for an inhibitor for KATP channels, the ATP concentration (J.LM) that produces half-maximal inhibition of channel activity inhibition constantt at zero voltaget inward rectifier t -type K+ channels shorthand designation for mechanosensitive K+ channels kilonucleotides voltage-gated K+ channels (generally delayed rectifierst) long-term depressiont long-term potentiationt low voltaget activated muscarinic acetylcholine receptor myristoylatedt , alanine-rich C-kinase substrate megabases t (Mbp - megabase pairs) miniature endplate currents megadaltonst (daltons x 106 ) malignant hyperthermia relative molecular masst messenger RNAt millivolt (10- 3 V)

NH2 NSA NSC NSC(Ca) nt N-terminal

number of functional channels also - Avogadro's number t Hill coefficientt nicotinic acetylcholine receptor-channel shorthand designation for voltage-gated Na+ channels predicted sites for N-linked glycosylationt (e.g. N-gly: aa122, specifying amino acid number 122 from known glycosylaset substrates) aminot group non-selective anion (channel) non-selective cation (channel) non-selective cation channels (calcium-activated) nucleotides amino-terminal (of protein)

o O-gly OHC 0/0 or 0-0

subscript abbreviation for extracellular O-linked glycosylationt outer hair cells outside-outt (patcht, recording configurationt)

P

prefix for post-natal day; used to describe developmental status (e.g. P5 == post-natal day 5) permeabilityt of Ca2 + ions (analogous nomenclature for other ions, e.g. PK , PNa , PRb etc.) pituitary adenylyl cyclase-activating polypeptide phosphodiesteraset

N

n nAChR Nav N-gly:

Pea PCAP PDE

-

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_O_3_re_s_u_m_e_ _

[PDTM]

pHi Phos/enzyme

pj. PIR pir: PKA PKC PNS poly(A) poly(A)+ P open or Po pS PSC PSP PSS

protein domain topography model; within the text, use of the abbreviation in square brackets denotes a positional feature illustrated on the model intracellular pHt Putative t (consensus t ) site for phosphorylation t by a specified enzyme, e.g. Phos/CaM kinase II - multi-functional (Ca2+/calmodulin)-dependent protein kinase II; Phos/CaseKII: casein kinase II; Phos/GPK: glycogen phosphorylase kinase; Phos/MLCK: myosin light-chain kinase; Phos/PKA: cAMP-dependent protein kinase (PKA); Phos/PKC: protein kinase C (PKC); Phos/PKG: cGMP-dependent protein kinase; Phos/TyrK: tyrosine kinase (TyrK) subtypes post-injection Protein Identification Resourcet (protein sequence database) designation for Protein Identification Resourcet accession numbers protein kinase A protein kinase C peripheral nervous systemt polyadenylationt (site) polyadenylatedt (mRNA) fraction of total cellular RNA channel open probabilityt picosiemens (10- 12 Siemenst) post-synaptic currentt post-synaptic potentialt porcine stress syndrome coefficientt for a ten-degree change in temperature

R R

r.p. rRNA RTK RyR S SAN SAPs s.c.a. s.c.c. s.c.p. SCR SD SDS-PAGE

SEM SFA

-

receptor resistance t (unit - ohm t ), reciprocal of conductance t resting potentialt ribosomal RNAt receptor tyrosine kinase t (at plasma membrane,cf. CTK) ryanodine receptor-channel Siemens t (unit of conductance t ; reciprocal ohm t or mho t ) sino-atrial nodet (of heart) signal-activated phospholipasest single-channel amplitudet single-channel conductancet (symbol, ,) single-channel permeabilityt single-channel recordingt standard deviation sodium dodecyl sulphate-polyacrylamide gel electrophoresist (i) standard errort of the means t or (ii) scanning electron microscopyt spike frequency adaptationt

1L-_e_n_t_ry_0_3_r_e_s_u_m_e

_

indicates the range of amino acids which form the signal peptide of a precursor protein (e.g. Sig: aal-26); alternatively, the abbreviation indicates the actual cleavage sitet forming the signal peptide t and mature chaint from the precursort protein substance P designation for SWISSPROT protein sequence database accession numbert sarcoplasmic reticulumt disulphide bond t ; in sequence database entries, the S-S: symbol is sometimes used to denote positions of a known disulphide bond linkaget or motift between two residues on a protein molecule, e.g. an experimentally determined link between residues 154 and 182 on the same chain would be written as S-S: 154-bond-182.

Sig:

SP sp: SR S-S:

TPeA+ TT

v

transmembrane melting temperaturet upper limit to the amount of material that carrier-mediated transport can move across a membrane tetrapentylammonium ions transverse tubulet

VDAC VDCC

voltt voltage t voltage-activated calcium channels; analogous nomenclature for other channelsl e.g. VACIC, VAKC, VANaC, VDAC voltage-dependent t anion channel voltage-dependentt calcium channel

W/C or W-C WCR

whole-cellt (recording configuration) whole-cellt recording

YAC

yeast artificial chromosomet

V VACaC

, ,j or

,(j)

IJ-A

n

w-CgTx

unitary (single-channel) conductance single-channel junctional conductancet microamp (10- 6 Amperes) ohmt, unit of electrical resistancet; reciprocal of conductance t omega-conotoxin

Feedback: Comments and suggestions regarding the scope, arrangement and other matters relating to the abbreviations section can be sent to the e-mail [email protected]. (see field 57 of most entries for further details)

lID

This Page Intentionally Left Blank

VOLTAGE-GATED CHANNELS

This Page Intentionally Left Blank

VLG Key facts Edward C. Conley

Voltage-gated channel families key facts Entry 41

Note: The ~key facts' sections are intended for readers unfamiliar with the more general aspects of ion channel biology, and contain selected introductory information applicable to whole groups of ion channel molecules. These sections, coupled with the on-line glossary items (indicated by daggert symbols attached to key terms in context) provide a basic overview of principles associated with more detailed information within the main entries of the book.

Membrane potential fluctuations represent a primary mechanism for cell signal transduction 41-01-01: A large diversity of voltage-sensitive responses in many cell types

support a central role for membrane potential t control in the evolution of complex cell signalling systems. Voltage-gated (viz. voltage-dependent, voltage-sensitive, voltage-activated) ion channels thus induce transmembrane ionic flow in response to sensed changes in transmembrane potential t . These proteins have permanently charged or dipolart regions which are forced to move by fluctuations in the local electric Heldt. Physical movement of the voltage sensor thereby couples to conformational changes associated with channel gating. The initiation of voltage gating has (by definition) little or no dependence on extracellular/intracellular ligands or protein phosphorylation (compare ELG key facts, entry 04, ILG key facts, entry 14 and INR key facts, entry 29) although these factors commonly modulate responses by altering (for example) voltage dependence of gating, current duration/amplitude, activation kinetics or channel protein interactions. During evolution, some channels have 'retained' voltage sensitivity of gating while being 'obligately' modulated by ligands (e.g. Kca channels, see ILG K Ca, entry 27; see also Voltage sensitivity under ELG CAT NMDA, 08-42). By contrast, other channels appear to have 'lost' voltage sensitivity in favour of 'obligate' ligand gating functions (e.g. cGMP-gated cation channels, see ILG CAT cGMp, entry 22). Comparative structural and functional aspects of the voltage-gated cationselective channel superfamily (Table 1) have been the subject of several reviews 1 - 10. The evolutionary origin of voltage-gated and other channels has been specifically discussed11,12.

Voltage-gated channels in excitable versus non-excitable cells 41-01-02.: Voltage-gated channels are abundant in excitable cells that pass

action potentials (such as nerve and muscle) and most information on how the different classes of voltage-gated channel interact have thus been determined in excitable cells. Although present at relatively low densities, several subtypes of voltage-gated channels have also been characterized in non-excitable cells (such as exocrine/endocrine secretory and blood cells). In any cell type, a given class of voltage-gated ion channel may have

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _en_t_ry_4_1_

predictable roles associated with propagation and control of that cell type's 'excitability cycle' (for details, see specific entries). For instance, some voltage-operated Ca2+ channels (VOCC, VLC Ca, entry 42) have major roles in triggering secretory vesicular release and muscle contraction, while voltage-gated Na+ channels (VLC Na, entry 55) have received most attention for their role in propagation of regenerative action potentials. In general, K+ channels stabilize cell voltages by counteracting the depolarizing effect of Ca2+ and Na+ channels - they repolarize cells following action potentials and thus modify excitability and firing patterns. Together with several classes of inwardly rectifying K+ channel (INR-series entries, Volume ill), some classes of voltage-gated K+ channel contribute to the membrane potential of resting cells and can therefore markedly influence properties such as excitability threshold t, basal secretion rate, maintenance of vascular/muscular 'tone' and cell volume. Comprehending the molecular diversity and patterns of receptor-coupled modulation of voltage-gated Ca2+ and K+ channel effectors t (in particular) has also been of significance for those interested in molecular mechanisms of signal integration and plasticityt.

Co-evolved 'integrative' properties of voltage-gated channels in shaping action potentials 41-01-03: Patterns of propagation for 'all-or-none' signals in the nervous

system are heavily influenced by the intrinsic functional properties of the channel proteins carrying the component ionic currents which summate to action potentials. For example, a voltage-gated Na+ channel subtype able to turn on and off rapidly would promote a high frequency of firing. Alternatively, a more slowly inactivating Na+ channel subtype would broaden an action potential. Co-expression with specific subtypes of K+ channel might also limit the maximum firing frequency. A striking demonstration of the roles of specific interacting sets of voltage-gated ion channel subtypes in 'shaping' action potentials is seen in the cardiac conduction system: action potentials recorded directly from each anatomical subregion of heart (broadly, sinoatrial node, atrial, atrio-ventricular node, His-Purkinje and ventricular tissue) display robust action potential shapes 'characteristic' of their constituent ionic components/channel subtypes (see also INR key facts, entry 29). Voltage-gated ion channels possess characteristic singlechannel conductances and maintain constant ionic selectivities depending on both specificity of ion binding and pore size. Thus 'selection' of channel components to 'fit' a functional role is dependent upon their integrative properties within a cellular compartment. In consequence, the interactive properties of multiple channel subtypes are of critical importance in computational approaches to understanding neuronal function 13 . Co-ordinate regulation (cell-specific expression) of voltage-gated ion channel genes and those encoding other signalling components is ultimately determined by genetic control elements t that chiefly reside 'upstream' of protein structural genes. These may include various transcriptional activator t and transcriptional silencert elements. For example, the latter have been characterized as being responsible for 'restrictive' expression of certain voltage-gated Na+ channels to neuronal but not non-neuronal cells (for an example, see Cene organization under VLC Na, 55-20).

II

1'--_e_n_t_ry_4_1

_

Sequence relatedness defines voltage-gated cation channel families and superfamilies 41-01-04: The functional diversity of mammalian voltage-gated cationselective channels appears to have been generated via primordial gene duplication and selection processes, which have given rise to the coexistence of multiple related-sequence genes within the genome t (Le. gene familyt members; for a detailed discussion of possible gene family evolutionary mechanisms, see Chromosomal location under VLC K Kv1Shak, 48-18). Complete sequencing of prokaryote genomes such as Escherichia coli has uncovered important candidates for ancestral K+ channel genes with homology to voltage-gated channels of higher organisms (e.g. kch, see Miscellaneous information under VLC K Kv1-Shak, 48-55). Thus homologues of voltage-gated K+ channel genes are now known to exist in an extraordinarily wide evolutionary range of organisms from Escherichia through Streptomyces, ciliate protists (e.g. Paramecium), Arabidopsis, jellyfish, worms, squid and (as most extensively characterized) flies and vertebrates. Amino acid sequence homology comparisons have revealed the protein domain arrangements of voltage-gated and cyclic nucleotide-gated (CNG) channel families to show many gross similarities14 (see also ILC CAT cAMP and ILC CAT cCMp, entries 20 and 21). These groups can be incorporated into a larger gene superfamilyt that includes the voltage-gated Na+, Ca+ and K+ channels, cyclic nucleotide-gated cation channels, hyperpolarization-activated cation channels (INR (K/Na)IfhQ' entry 34) and Ca2+-activated K+ channels (Table 1; for the distinct domain arrangements of voltage-gated Cl- channels, see next paragraph). In general, all channels in the voltage-gated superfamily listed in Table 1 show a similar protein domain arrangement (the 'SI - S6 + P arrangement') consisting of a set of six predicted membrane-spanning segments (hydrophobic regions SI to S6) plus a pore-forming region (P-region, formerly the H5 domain between 85 and 86 - see {PDTM} Fig. 6 under VLC K Kv1 Shak, entry 48). Additionally, multiple tissue-specific variants of some voltage-gated ion channels can arise through alternative splicingt of primary RNA transcripts (e.g. see VLC K Kv3-Shaw, entry 50).

tMatching' native channel currents with those conducted by heterologously expressed channel proteins 41-01-05: Several factors underlie the difficulty of 'matching' properties of native cell voltage-gated channel currents to the 'contribution' of individual channel protein subunits. With regard to Kv channels, for example, heterologous t expression of single channel subunits (Le. in the specified

II

II Table 1. Comparative protein domain arrangements within the cation-selective voltagelsecond-messenger-gated channel gene superfamily (Sortcodes beginning VLG- are covered in this volume) (From 41-01-01) Sortcodes/class subunit (example)

Descriptive notes/cross-references/typical protein domain arrangement

Gene family designation/ selectivity (see entries)

Refs

VLGCa Voltage-gated Ca2 + channel, a: subunit (see note 3)

In Cava subunits, the Sl-S6+P arrangement is

Cav (in keeping with Kv channels, see note 2)

e.g. 15?16

present in each of foul internal repeats as shown. In Cav (and Nav) channels, the a subunit is a long, single, folded polypeptide - (cf. the extracellular ligand-gated (ELG) channel superfamily (entries 04 to 11), which are composed of multiple subunits)

Generally highly Ca2+ selective

The Cava-subunit is co-expressed with a2/8, {3 and, subunits. Many structurally and functionally distinct subtypes exist. In general, Cav channels help conduct Ca2+ influx activated by depolarization, initiating a large number of cellular functions including muscular contraction and neurotransmitter release. Key: +, positively charged voltage sensor element (S4 domain); P, pore-forming domain.

g ~ ~ ~

VLG K series

(except VLG K Kv-beta and VLG K minK) Voltage-gated K+ channel, a subunit

In Kv 0: subunits, the Sl-S6+P arrangement is present only once per K+ channel a subunit. Functional channels in vivo have been shown to be formed by association of four Kvo: subunits and four Kv{3 subunits (for details, see VLC K Kv-beta, entry 47).

Kv (as first established by Chandy and colleagues, see note 2)

e.g. 17

t"1"

~ ~

Generally highly K+ selective

In general, Kv channels have cellular repolarization functions (Le. depolarized membrane potentials opening K+ channels leading to hyperpolarization and return to rest). Differential expression and modulation of Kv channel subtypes can radically alter cell excitability, Le. the firing patterns and duration of action potentials (see various VLC K series entries). Key: +, positively charged voltage sensor element (S4 domain}j P, pore-forming domain. VLGNa

Voltage-gated Na+ channel, subunit (see note 3)

a

In Nav 0: subunits, the Sl-S6+P arrangement is Nav present in each of four internal repeats as shown. Like (in keeping with Kv channels, Cay channels, the 0: subunit is a long, single, folded note 2) polypeptide. Generally highly Na+ selective

II

g ~

e.g. 16,18

II

Table 1. Continued Sortcodes/class subunit (example)



Refs

Mammalian Nav Q subunits are co-expressed with two smaller f3 subunits. Functionally and structurally distinct Nav channel subtypes exist (see VLC Na, entry 55), but most are steeply voltage dependent. In general, Nav channels are activated by depolarization and induce depolarization; responsible for action potential propagation. Key: +, positively charged voltage sensor element (S4 domain); P, pore-forming domain. ILG K Ca (Volume II) Ca2 +-activated K+ channel, Q

subunit

In KCa Q subunits, the SI-S6 + P arrangement is KCa present only once per K+ channel Q subunit. Species (in keeping with Kv channels, homologues of the genes encoding vertebrate and fly note 2) large-conductance Ca2+-activated K+ channels are described in ILC K KCa, entry 27.

19

Generally highly K+ selective

Key: Grey bar, Ca2+-binding site; Ca, calcium. P, poreforming domain. The Ca2+-activated K+ channel of Drosophila is encoded by slopoke; vertebrate homologues are generally named 'slo' with a species prefix; the 'SI-S6 + P' arrangement and subunit stoichiometry has some similarity to Kv channels described above.

g ~ ~ .......

ILG CAT cAMP (Volume IT) ILG CAT cGMP (Volume IT) Cyclic nucleotide-gated cation channels, a subunits

In CNG a subunits, the 81-86 + P regions are present

only once per subunit with subunit stoichiometries and assemblies analogous to Kv channels. CNG-a/ f3 subunit co-associations described in functional channels (see entries 21 and 22).

Key: Hatched bar, cyclic nucleotide t binding site; cNUC, cyclic nucleotide; P, pore-forming domain.

Ca2 + channel for inositidemediated Ca2 + entry {putative}

81-86 + P present once per subunit. An example is the putative Ca2 + channel for phosphoinositide tmediated Ca2+ entry encoded by the trp (transient receptor potential) gene, which probably mediates invertebrate photoreception. For brief discussion of trp see Table 1 under ILG Ca CSRC (entry 18, Volume II) and Miscellaneous information under ILG CAT cGMp, 22-55.

CNG (cyclic nucleotide-gated, see entries 21 and 22)

II

S M"

~ ~

.......

Mixed cation Na+/K+ /Ca+ selective dependent upon subtype (see fields 21-40 and 22-40) In press update: Hyperpolarization-activated cation channels also conform to this pattern - for refs, INR (K/ Na)IfhQ' entry 34. (-)

(non-vertebrate, but compare vertebrate channel showing M1- M6 + P arrangement in Fig. 1 of ILG Ca InsPa, entry 19)

Generally highly Ca2+ selective

Key: Black bar, calmodulint -binding site; CAM, calmodulin; P, pore-forming domain.

1,20

21,22

II

Table 1. Continued Sortcodes/class subunit (example)


Plant K+ channel/transporter SI-S6 + P present once per subunit. Sequence similarity between these molecules (expressed in plants) and the Shaker-related channels indicates a common ancestor gene prior to the separation of the plant and animal kingdoms (ca. 700 million years ago for background to K+ channel evolution see Cmomosomallocation under VLC K Kv1-Shak, 48-18).


Refs

(-)

23,24

(non-vertebrate; see also E. coli kch channel gene described under field 48-55).

K+ selective

Key: Light grey bar, cyclic nucleotide t -binding site; cNUC, cyclic nucleotide; P, pore-forming domain.

Notes: 1. Comparison of protein domain topography in the voltage/second messenger-gated superfamily of ion channels and molecules of the ATP·binding cassette (ABC) superfamily reveal distant structural relationships (see also Table 1 under ILC key facts, entry 14, and also PDTM of InsPg-gated Channels, entry 19 (Fig. 1).). 2. Whereas the gene family prefix 'Kv' is well-established, analogous universal designations have not appeared for other families; the designation shown is used on the ICN web site (www.le.ac.uk/csn/ ) to stay consistent with the Kv channel nomenclature. 3. Both voltage-gated sodium channels (see VLC Na, entry 55) and calcium channels (see VLC Ca, entry 42) carry inward currents which depolarize t the membrane through flow of positive charges into the cell. Conversely, outward currents which flow following activation of potassium channels (see VLC K-series, entries 44 to 54) and inward currents which flow following activation of Clchannels (see VLC Cl, entry 43) tend to hyperpolarizet the membrane by making the interior of the cell more negative.

Ig ~ ~ ..-

~e_n_t_ry_4_1

_

absence of endogenous channel subunits capable of co-assembling with the heterologous component) is generally assumed to occur via formation of homomultimeric t (i.e. homotetrameric t) channels. However, co-assembly of 'accessory' subunits with pore-forming (alpha) subunits can significantly influence the inactivation, modulation and expression properties of voltagegated channels formed from Q; subunits alone (for further examples, see VLC K Kv-beta, entry 47 and VLC K Kv1-Shak, entry 48). Heterologous expression systems may also not 'reproduce' significant modulatory conditions found in native cells. Note: These difficulties in matching 'native' with 'cloned' channel properties have necessitated separate treatment of their literatures within this volume. For example, while it is likely that many of the native cell currents extensively described as 'A-type' (entry 44) or 'delayed rectifier-type' (entry 45) arise from the proteins described in the VLG K Kv series entries, there is often relatively little direct evidence to 'implicate' specific protein subtype(s). Applications of new computer-based analytical techniques 13 and consideration of multiple functional properties may help to improve 'concordance' between 'native' and 'cloned' channels. In general, however, data derived from native versus heterologous cell preparations have been kept separate. There is still much to learn about the regulation of expression patterns for multiple voltage-gated channel subunits in the eNS (and elsewhere) and the specific subunit compositions of heteromultimeric channels in vivo.

Minor differences in voltage-gated channel sequences can underlie major functional changes 41-01-06: Introduction of relatively minor changes in channel primary DNA sequences (e.g. by site-directed mutagenesis) can radically alter functional properties of the heterologously t expressed protein. Such structurefunction t analyses have been enormously valuable in defining protein domain 'functions' acquired through processes of divergent t evolution (where the protein encoded by a single 'progenitor' gene may be 'adapted' to perform multiple functions, for example in different cell types). Alteration (mutation) of DNA sequences encoding 'key' amino acid residues participating in highly specific protein movements or interactions (including the sensitivity to 'external' modulators) may thus markedly alter phenotypic t properties in cells. Most generally, mutations affecting native channel protein conformation t, subunit assembly t properties or equilibrium (allosteric t) behaviour can be predicted to produce functional changes. Structurefunction analyses have helped to localize 'critical' residues for channel functions such as ion selectivity, permeation and block (for examples, see Selectivity, field 40 and Blockers, field 43 in various entries), inter- and intra-channel protein domain interactions (see Protein interactions, field 31), intracellular 'targeting'/localization mechanisms (see Subcellular localization, field 16) and various phosphomodulatory mechanisms (see Protein phosphorylation, field 32). A large number of studies have used mutagenesis to generate channel variants that show altered activation properties (see Activation, field 33), rectification properties (see Current-voltage relation, field 35), inactivation mechanisms (see Inactivation, field 37), voltage-gating or the voltage-dependence t of various processes (see Voltage sensitivity, field

III

_L.-

en _ t_ry_4_1_

42). Notwithstanding these striking functional effects, many mutations

introduced experimentally (or 'accumulated' during evolution) may be 'silent' or produce a 'non-functional' protein according to the criteria employed. Judgements as to the 'functional' effects of ion channel gene mutations may thus be subjective: effects may have some dependence on the heterologous t or native cell 'background' (and other, co-expressed, interacting proteins) (see also notes on selection of heterologous gene expression systems in the footnote to Table 1 under ILG key facts, entry 14).

Non-exhaustive examples of mutations in voltage-gated channel genes implicated in genetic disorders 41-01-07: As further described in Phenotypic expression (field 14) of relevant

entries, mutations in genes encoding voltage-gated Ca2+, CI-, K+ and Na+ channels have been shown to be responsible for a number of human and animal genetic disorders: Ca2+ channels: Mice with mdg mutations have muscular dysgenesis, where lack of the 0lS subunit of the skeletal muscle L-type Ca2+ channel affects excitation-contraction coupling. Mutations at the mouse tottering (tg) locus affect the 0lA subunit of a P/Q-type Ca2 + channel and induce a neurological disorder resembling petit mal epilepsy in humans. The human autosomal t dominant t disease hypokalaemic periodic paralysis (hypoPP) is caused by mutations in the CACNL1A3 gene that encodes the 0lS subunit of the skeletal muscle L-type Ca2 + channel. Mutations in the human CACNL1A4 gene encoding an 0IA subunit of a P /Q-type Ca2+ channel have been identified in patients with familial hemiplegic migraine (FHM). Genetic alterations affecting the channel Oli} subunit have also been shown to be the cause of autosomal t dominantT cerebellar ataxia. CI- channels: Mutations in the C1CN1 gene encoding the skeletal muscle CI- channel are responsible for dominant myotonia congenita (Thomsen's disease) and for recessive generalized myotonia (Becker's disease). Dent's disease, an X-linked hypercalciuric nephrolithiasis, is caused by mutations in the kidney-specific C1CN5 gene. K+ channels: Point mutations in the hKvl.l gene have been associated with one form of familial episodic ataxia, characterized by brief episodes of ataxia t with myokymiat (continuous movement or 'rippling' of muscles) evident between attacks consistent with reduced capacity for repolarization in affected nerve cells. Two forms of the familial cardiac long QT (LQT) syndrome have been associated with mutations in two distinct genes encoding potassiumselective voltage-gated K+ channels (KvLQTl underlying LQT1, see VLG K minK, entry 54 and HERG underlying LQT2 - for further details, see Phenotypic expression, 46-14 and Chromosome location, 46-18). Na+ channels: Missense t mutations in the human gene (SCN4A) specifying the adult skeletal muscle voltage-gated Na+ channel 0 subunit are associated with two hereditary disorders of sarcolemmal excitation, hyperkalaemic periodic paralysis (HYPP or hyperPP) and paramyotonia congenita (PC).

II

l_e_n_t_ry_4_1

_

Mutations in the seNSA gene encoding a cardiac sodium channel a subunit are known to underlie the LQT3 form of LQT syndrome (see cross-references above). Mutations of the med locus on mouse chromosome 15 (the ScnBa gene) produce a recessive t neurological disorder, motor endplate disease, with symptoms ranging from mild ataxia t to juvenile lethality. Neuronal defects, including lack of signal transmission at the neuromuscular junctiont, excess pre-terminal arborization t and degeneration of cerebellar Purkinje cells, are apparent in the disorder.

Distinct mechanisms underlying cellular versus subcellular distributions of channel subtypes

41-01-08: Virtually every excitable t cell type expresses multiple subtypes of voltage-dependent channels specific to the 'functions' of that cell type. Functionally specialized cells may exhibit a wide variety of 'current types' each displaying individual patterns of modulation. For example, olfactory receptor neurones display six classes of voltage-dependent ion currents including Na+ currents, L-type Ca2 + currents, Ca2 +-activated K+ current, a transient K+ current, a delayed rectifier K+ current, and an inward rectifier K+ current25 . Notably, multiple protein subtypes may support each class of current. For example, techniques such as single cell RT-PCRt have revealed co-expression of multiple subtypes of Kv channel within individual neurones (see also INR key facts, entry 29). Furthermore, as described in field 13 (mRNA distribution) and field 15 (Protein distribution) many studies have employed in situ hybridization and immunological techniques for localization of voltage-gated channel mRNAs and proteins across broad areas of tissue. In general, these studies have predicted the existence of elaborate developmental control mechanisms for appropriate 'placement' of ion channels in various developmental compartments (see Developmental regulation under TUN [connexins), 35-11). Once expressed within a developmental compartment or cell type, however, subcellular targeting mechanisms (normally involving discrete protein interactions) become important for assembling the 'multiprotein machines' that constitute functional ion channels in vivo (for examples of these principles, see Subcellular locations, field 16 and Protein interactions, field 31).

Structural elements determine compatibility of subunit associations 41-01-09: Full 'reconstitution' of the native properties of voltage-gated ion channels may require complex interactions of multiple protein subunits (for examples within the superfamily, see Protein interactions, field 31). Many examples exist where 'accessory' protein components other than those which form the main voltage-sensitive/ion-selective component interact to form the native channel (ibid.). Within some gene families, structural elements or regions have been defined that appear to determine compatibility of protein subunit associations. A well-characterized example of this is the region within the N-terminal domain of Kv (voltage-gated K+) channels that has been shown to mediate subunit-subunit interactions in homomultimeric t channels as well as determining compatibility of co-assembly between different K+ channel subunits26 . Translocation of this region to recipient channel monomers promotes co-assemblies characteristic of the

II

_L.....

e_n_t _ry_4_1_

donor channel subunit (see Protein interactions under VLC K Kv1-Shak, 4831 and VLC K Kv2-Shab, 49-31). These studies provided a mechanistic basis for understanding why different members of the same subfamily of Kv channels were able to co-assemble and form functional heteromultimeric t channels while subunits of different subfamilies do not co-assemble (for details see Protein interactions under VLC K Kv1-Shak, 48-31).

The fourth hydrophobic segment (S4) functions as a tvoltage-sensor' in voltage-gated cation channels 41-01-10: A common feature of voltage-gated potassium, sodium and calcium channels is that the 84 transmembrane segment contains basic t amino acid residues (e.g. Lys or Arg) at every third or fourth position (for examples, see the protein domain topography mode1s, PDTM, of various entries). This arrangement confers an intrinsic voltage sensitivity such that changes of membrane potential exert an electrostatic force on the charges or dipoles of the voltage sensor. This force induces protein conformational changes (including outwardly directed movement of the 54 element) that lead to channel opening (contrast the arrangement for voltage-gated chloride channels, next paragraph). Movements of voltage sensors can be detected as gating current t, and these currents show significant differences between inactivated and non-inactivated channels. The number of charges that are translocated during channel opening (the gating charge t ) can be estimated by examining the steepness of the voltage dependence of gating, as reviewed27 (see also panel D under Figure 1). In the original 84 hypothesis, four 54 sequences corresponded to voltage-sensitive gating particles as proposed by Hodgkin and Huxler8 . Classical measurements of Na+ channel gating current (i.e. movement of charges intrinsic to the channel) estimated that four to six charges move from one side of the membrane to another as the channel opens29. This was accounted for in the 54 model by depolarization causing each of the 54 sequences to move by approximately one helical turn. In general, voltage-gated channels differ in (i) the voltage activation range over which membrane potential changes can open the ion pore, and (ii) in the transition rates t between the open (conducting) and the various non-conducting states of the channel (see Activation, field 33, Kinetic model, field 38; also compare t.he various domain arrangements in the inset figures to Table 1). Note: During action potentials, membranes may typically undergo voltage changes on the order of "-JO.I V ("-JI06 VIm).

A distinct mechanism for voltage sensing in voltage-gated chloride channels 41-01-11: The 'voltage-gated' CI- channels lack an obvious structural motif analogous to the 54 domain which functions as a voltage sensor in the voltage-gated cation channels (see paragraph 41-01-10). Analysis of the voltage and pH dependence of channel gating suggests that a single acidic residue is primarily responsible for voltage sensing: an aspartate in the first

II

1

e_n_t_ry_4_1

_

transmembrane segment is a prime candidate as the voltage sensor, since its replacement by a glycine in a recessive form of myotonia congenita greatly affects voltage-dependent gating of the hCIC-l CI- channel. The gating behaviour of CIC-l channels is consistent with a slow blocking/unblocking event mediated by a cytoplasmic 'gate' that is postulated to behave as an open-channel blocker, analogous to the 'ball and chain' model for the inactivation of voltage-gated Na+ and K+ channels. The voltage-induced conformational change is postulated to alter the 'on-rate' of the blocking particle (see note) (for further details, see the VLG Cl fields Phenotypic expression, 43-14, Domain functions, 43-29 and Kinetic model, 43-38); also compare the description of the voltage-gated anion channel (VDAC) characterized in mitochondrial membranes under the MIT [mitochondrial, native}, entry 37. Note: Not all members of the CIC family are voltage gated; the CIC-3 channel, for example, is voltage independent and remains open at all membrane potentials.

Distinctions between voltage-gated and inwardly rectifying K+ channels 41-01-12: By definition, voltage-gated K+ channels have their ionic conductance (g) regulated as a function (f) of membrane potential alone (V) - i.e. g == f(V). This is to be contrasted with inward rectifier K+ channels (see the INR K series entries, Volume III) whose conductances depend on the electrochemical gradient or driving force t on the ions (Le. gK versus voltage (V) varies according to a function (f) of gK == f(V - EK), where EK is the equilibrium potential l for potassium ions. In general, therefore, the current-voltage relation t for an inward rectifier will also depend on factors like [K+]o that can influence EK and in consequence, the prevailing driving force on potassium ions (compare also the effects of [K+}o on HERG channels under VLC K eag/elk/erg, entry 46). Figure 1 illustrates the origin of certain conductance parameters for voltage-gated channels that are cited in the main entries.

Many voltage-gated channels automatically inactivate with time

41-01-13: In comparison to some ligand t -gated channels (e.g. see the ELG and ILG series entries in Volumes I and II) most voltage-gatedt channels conduct relatively short-duration currents, before intrinsic channel inactivationt or intrinsic gating t processes sensitive to changes in membrane potential 'cut off' the signal under physiological conditions (see below). Whereas the opening and closing of some voltage-gated channels appear to depend entirely on the prevailing membrane potential, many show lprogressive' inactivation (even under maintained depolarizing conditions that would initially cause the channels to open). Inactivated channels are thus refractory to further depolarization-induced openings. Agents that inhibit Na+ channel inactivation in nerve axons (e.g. the pyrethrins, natural insecticides isolated from the flowers of the genus Chrysanthemum, and their synthetic analogues, the pyrethroids) cause repetitive firing and depolarizing after-potentials t. The synthetic insecticide DDT (dichlorodiphenyltrichloroethane) acts in a similar fashion. A large body of work has analysed different inactivation mechanisms, particularly in cloned voltage-gated potassium and sodium channels

II

II . . - - - - - - - - - Panel A: Generation of current families [2] Larger depolarizations generate larger currents which activate more rapidly

[1 ] Recording of the potassium current that flows by changing the membrane potential in a stepwise manner

c

Faster

I I I

Exponential tail current } -

[3]

lO

~

vm:J--'--------

-ve

Tail currents start with an abrupt drop as the channels close In response to the change in Vm. The degree of change in current depends on where the membrane voltage is relative to the equilibrium potential for potassium ions (E K>

Figure 1. Origin of conductance parameters for voltage-gated channels as mentioned in text, exemplified for non-inactivating voltage-gated K+ channels. In these examples K+ current would be isolated (i.e. measured exclusively) from other membrane currents present (either using pharmacological blockers, appropriate voltage protocols and/or by heterologous expression a cloned K+ channel gene of interest).

g ~ .,J:::l.

......

(1)

=

~

~ I-'

Panel B: Generation of current-voltage (I-V) relations

[2] [3]

The I-V relation in this example shows step depolarizations to have little effect near EK, but having increasing effect with increasing depolarization - see panel C.

IK

[1 ] From the types of record shown in panel A, it is possible to plot the membrane current (at its peak) against the voltage at which it was recorded.

-ve

I

EK Figure 1. Continued

II

~

o

Vm

+ve

At the membrane potential defined by EK, potassium ions will move neither inwards nor outwards across the membrane - as defined by the Nernst equation. EK is the membrane voltage at which the concentration gradient is exactly balanced by the electrical gradient working on potassium ions.

II ,.....-.----- Panel C: Chord conductance and slope conductance [2] The chord conductance for potassium (gK) is higher at more positive potentials - i.e. it increases with increasing depolarization

[1 ] Conversion of potassium current into a chord conductance is obtained* as the slope of a line (dotted) joining a point on the I-V relation to the equilibrium potential for potassium ions.

*assuming Ohms law, I = VIR I K = gK (Vm-E K )

[3]

IK Chord conductance gK =

I K

, ,,""

(Vm-EK )

,,"

-ve

t~--'"

EK

,",,"

o

,,;.

More rarely, the slope __ . __ . __ . __ . __ . __ . __ .•_. __ conductance parameter is used as the simple slope of the I-V relation (~II~V). The slope conductance, unlike the chord conductance, is dependent on ionic conditions.

I

Vm

+ve

Figure 1. Continued ~

= ~ ~ ......

Panel D: Normalized conductance-voltage relations

, [1 ] The conductancevoltage curve shows how the probabi Iity of a channel being open varies with voltage.

Conversion of currents to conductances (as described in panels A-C) and plotting against membrane potential produces a sigmoid relation with a near-zero conductance at -ve membrane potentials and a maximum at +ve potentials (see note 1).

For voltage-activated channels, g-V relations can be fitted with Boltzmann functions of the form:

Relative 9

K

Relative gK

1

=

1 1 + exp [(Vh-Vm)/k]

[3] from above Vh voltage at which gK is half-maximal; k a constant related to the number of equivalent charges involved in channel gating (and that describes the steepness of the voltage-dependence). Together, Vh and k specify the voltage-dependence of the channel (see note 2).

= =

o

-ve

Vm

EK Notes: 1: The maximum level of the plateau will depend on the concentration of permeant ions, so the curves are normalized to reach the same maximum. 2: For a purely voltage-gated channel, alteration of EK to zero (Le. by raising the extracellular [K+]) should not alter the position of the g-V curve along the voltage axis - compare with inward rectifiers described under INR {key facts}, entry 29. If g-V curves do not shift with EK set to zero, the potassium conductance varies as a function of voltage alone.

Figure 1. Continued

II

!

~ ..-

[2]

_'--

e_n_t_ry_4_1_

(for details, see Inactivation, field 37, under the VLC Na and VLC K series entries). The 'hinged-lid' and 'ball-and chain' mechanisms (ibid.) have become dominant models for fast inactivation of voltage-gated Na+ and K+ channels respectively. Note: Prior to the molecular cloning of large numbers of K+ channel genes, K+ current inactivation behaviour was a principal means of channel classification (see discussion under VLC K A-T [native], entry 44 and VLC K DR [native], entry 45).

Some blockers of voltage-dependent channels act from the cytoplasmic side on open channels 41-01-14: Several chemical blockers of voltage-gated ion channels appear to bind at a site close to the vestibulet near the cytoplasmic opening of the poret - e.g. local anaesthetics for Na.+ channels (see below), quaternary D600 for some Ca2+ channels and tetraethylammonium ion (TEA) and its analogues for some K+ channels. Historically, blocking behaviour like this was used to gain a picture of channel structure in the absence of any direct molecular information. Blocking agents which became 'lodged' within the channel as it closed could be rendered less effective (i.e. 'dislodged') by increasing concentrations of the permeant ions on the extracellular side. These and other observations indicate:d that the 'gate' for closure of the channel was on the cytoplasmic side and that the pore itself did not change its shape significantly during gating. Local anaesthetics, such as lidocaine and procaine, are generally lipid-soluble tertiaryt amine compounds that inhibit propagated action potentials by blocking Na+ channels. Note: Blockade of voltage-gated Na+ channel function is of potential clinical use to augment neuroprotection during ischaemic or traumatic injury.

Voltage-gated channels are generally modulated by activation of separate receptor proteins 41-01-15: The majority of channels coupled to discrete receptor and transducert proteins can be gated by membrane potential changes even in the absence of receptor agonist. These ehannels can therefore be considered receptor modulated, where neurotransmitters enhance or suppress the primary, voltage-dependent responses. Although some neurotransmitters may have effects on voltage-dependent Na+ and CI- channels, the majority of neurotransmitter actions appear to modulate the function of Ca2+ and K+ channels via phosphorylation/dephosphorylation (see Protein phosphorylation, field 32 and Receptor/transducer interactions, field 49, under various entries). Ca2+ channel modulation reflects the role of calcium ion not only as a principal depolarizing charge carrier, but also a ubiquitous second messengert (i.e. modulating neurotransmitter release, enzyme activation and other channels). Expression of a great diversity of voltagegated channel subtypes (most capable of independent modulation) permits 'fine tuning' of functions such as firing pattern, Ca2 + influx and resting potential. The precise effect of modulation is primarily determined by characteristics of voltage gating for each channel type. In general, activation of any K+ current will bring the cell closer to the K+ reversal potential t , which in most cell types is more negative than the resting potential. Thus activation of a K+ current (or its enhancement through neurotransmitter action) will

1'--_e_n_t_ry_4_1

I

----J_

depress excitabilityt, while interactions which decrease K+ currents will enhance cell excitability (see also INR key facts, entry 29).

Feedback Error-corrections, enhancements and extensions 41-57-01: Please notify specific errors, omissions, updates and comments on this 'key facts' entry by contributing to its e-mail feedback file. For this entry, send e-mail messagesTo:[email protected]. indicating the appropriate paragraph by entering its six-figure index number (xx-yy-zz or other identifier) into the Subject: field of the message (e.g. Subject: 41-01-03). Please feedback on only one specified paragraph or figure per message, normally by sending a corrected"replacement according to the guidelines in Feedback etJ CSN Access. Notified changes will be indexed from within the CSN website (www.le.ac.uk/csn/).

REFERENCES 1 2 3 4

5 6 7

8 9 10

11 12 13 14

15 16 17

18 19 20 21

22 23

24

25 26 27

28 29

Jan, Cell (1989) 56: 13-25. Catterall, Trends Neurosci (1993) 16: 500-6. Honore, Fund Clin Pharmacol (1994) 8: 108-16. Isom, Neuron (1994) 12: 1183-94. Pusch, Physiol Rev (1994) 74: 813-27. Ertel, Drug Develop Res (1994) 33: 203-13. Jentsch, Chloride Channels (1994) 42: 35-57. Keynes, Q Rev Biophys (1994) 27: 339-434. Perez-Reyes, Drug Develop Res (1994) 33: 295-318. Catterall, Annu Rev Biochem (1995) 64: 493-531. Hille, Q TExp Physiol (1989) 74: 785-804. Salkoff, Trends Neurosci (1992) 15: 161-6. Conley, Meth Enzymol (1998) In press. Jan, Cell (1992) 69: 715-18. Miller, TBioi Chem (1992) 267: 1403-6. Catterall, Science (1988) 242: 50-61. Jan, Annu Rev Physiol (1992) 54: 535-55. Trimmer, Annu Rev Physiol (1989) 51: 401-18. Atkinson, Science (1991) 253: 551-5. Jan, Nature (1990) 345: 672. Hardie, Neuron (1992) 8: 643-51. Phillips, Neuron (1992) 8: 631-42. Sentenac, Science (1992) 256: 663-5. Anderson, Proc Natl Acad Sci USA (1992) 89: 3736-40. Miyamoto, T Gen Physiol (1992) 99: 505-30. Li, Science (1992) 257: 1225-30. Hille, Ionic Channels of Excitable Membranes, 2nd edn. (1992) Sinauer Associates, Sunderland, MA. Hodgkin, TPhysiol (1952) 117: 500-44. Almers, Rev Physiol Biochem Pharmacol (1978) 82: 96-190.

II

. . . Voltage-gated calcium channels

William

J. Brammar

Entry 42

NOMENCLATURES

Abstract/general description 42-01-01: Calcium-selective voltage-gated channel types are ubiquitous in excitable membranes. In general, unlike Na+ channels, voltage-gated Ca2+ channels are not rapidly inactivating, so that they are able to maintain inward currents for longer depolarizing responses. This electrical role of Ca2+ channels is important in secretory glands and endocrine organs, where Ca2 + channels dominate the electrical response and produce a prolonged depolarization to drive the maintained secretion. It is also important in cardiac and smooth muscle, where the longer depolarization is required to maintain the contraction. 42-01-02: Calcium channels regulate a wide range of cellular events as they can convert electrical into chemical signals (e.g. intracellular calcium flux and neurotransmitter release). Since resting intracellular Ca2+ concentration is low ("-110- 7 M, see ILG key facts, entry 14), only a small amount of Ca2 + influx is required to initiate signalling. In conjunction with cell-surface receptors and the intracellular Ca2 +-release channels (e.g. see ILG Ca InsPa, entry 19 and ILG Ca Ca RyR-Cal, entry 17), voltage-gated calcium channels initiate and control diverse cellular responses. These responses are often cell-type-specific and include secretion, metabolic adjustments, cell proliferation, contraction and control of gene expression. 42-01-03: Voltage-gated Ca2+ channels are of many types, including L-, T-, N-, P-, Q- and R-types, distinguished by biophysical and pharmacological criteria (see Gene family, 42-05-03). Although sequences encoding many different Ca2 + channel proteins have been cloned and characterized, the structural basis of Ca2+ channel diversity is still not fully understood, and the relationship between native currents and expressed genes is still being established. 42-01-04: Purification and immunoprecipitationt have shown that voltagegated Ca2 + channels are hetero-oligomerict protein complexes. The L-type Ca2 + channel of skeletal muscle, for example, consists of five subunits, 01, 02, {3, 8 and '"Y. The 01 subunit contains the pore t of the channel and is the main site of action of the L-type-specific Ca2 + channel agonistst and antagonistst. There are also various isoformst of the {3 subunit which differentially affect the electrophysiological behaviour of the channels produced by heterologoust expression of 0 subunit eDNAs. 42-01-05: L-type calcium channels, which are found in virtually all excitable and many non-excitable tissues, are re~ldily blocked by 1,4-dihydropyridines (e.g. nifedipine), phenylalkylamines (e.g. verapamil) and benzothiazepines (e.g. diltiazem). Three known 01 subunits, 0lS, 0lC and OlD, encoded by separate genes, are involved in different L-type Ca2 + channels: 0lS in skeletal muscle and 0lC and OlD in brain and heart.

II

I

entry42

1..-.---

_ _

42-01-06: N-type channels, expressed in neuronal tissue and associated with neurotransmitter release, are characterized by their high voltage activated (HVA) currents and their sensitivity to block by the cone snail neuropeptide omega-conotoxin GVIA (w-conotoxin GVIA or w-CTx). The activity of N-type channels is inhibited by the action of noradrenaline at pre-synaptic 02 receptors, via a G protein/second messenger system. The 01 subunit of Ntype Ca2+ channels has been identified as 0lB. 42-01-07: P-type channels, particularly prevalent in cerebellar Purkinje cells, activate over a range of potentials positive to -50mV and inactivate very slowly (tlj2 rv I s). They are sensitive to block by nanomolar concentrations of the peptide toxin w-agatoxin IVA from the funnel web spider, Agelenopsis aperta, and to micromolar concentrations of the polyamine FTx isolated from the same source. 42-01-08: T-type (low voltage activated, LVA) channels constitute a broad class of Ca2+ channels that transiently activate at negative potentials, typically in the range -70 to -50 mV, and are relatively sensitive to changes in resting membrane potential. T-type channels have been called 'fast' because of their rapid, voltage-dependent inactivation, but they close (deactivate) 10-100-fold more slowly than other voltage-gated Ca2+ channels. Their contribution to total Ca2+ current can be assessed from the amplitude of the slow component of deactivation ('tail') currents. It is a characteristic of T-type channels that they are less sensitive to classical Ca2+ channel agonists and antagonists than the high-threshold channel types. 42-01-09: The skeletal muscle and cardiac L-type Ca2+ channels are both essential components of excitation-contraction (E-C) coupling. In skeletal muscle, the L-type channel acts as the voltage-sensor that controls release of Ca2+ from the sarcoplasmic reticulumt in response to depolarization of the membrane of the transverse tubules, without needing Ca2+ entry across the sarcolemma. In contrast, the cardiac L-type channel functions only to permit entry of Ca2+ in response to depolarization, and the elevated Ca2+ concentration subsequently triggers the release of Ca2+ from the sarcoplasmic reticulum. Dihydropyridine-sensitive L-type channels are also present in brain, where the Ole subunit is the pore-forming component. 42-01-10: Mutations affecting Ca2+ channel subunits have been implicated in a number of genetic disorders. Mice with mdg mutations have muscular dysgenesis, in which lack of the 0lS subunit of the skeletal muscle L-type Ca2+ channel affects excitation-contraction coupling. Mutations at the mouse tottering (tg) locus affect the alA subunit of a P/Q-type Ca2+ channel and cause a neurological disorder resembling petit mal epilepsy in humans. The human autosomalt dominant t disease hypokalaemic periodic paralysis (hypoPP) is caused by mutations in the CACNL1A3 gene that encodes the alS subunit of the skeletal muscle L-type Ca2 + channel. Mutations in the human CACNL1A4 gene encoding an alA subunit of a P/Q-type Ca2+ channel have been identified in patients with familial hemiplegic migraine (FHM). Genetic alterations affecting the channel OIl} subunit have also been shown to be the cause of autosomal t dominantT cerebellar ataxia (see Phenotypic expression, 42-14-01 to 42-14-09).

II

_'--

e_n_try_4_2_1

42-01-11: Ca2+ channels are modulated by a variety of neurotransmitters and hormones, acting via G protein-linked receptors. In some cases the G protein interacts directly with the Ca2+ channel, without involvement of a diffusible second messenger. The inhibitory action of G proteins is selective for some channel types, the N-type and P/Q-type Ca2+ channels being the principal targets. This modulation of Ca2+ channels is important in pre-synaptic inhibition, because the inhibition of Ca2+ entry to pre-synaptic neurones blocks transmitter release. The channels are also modulated by phosphorylation by protein kinases A, C and G (PKA, PKC and PKG) and by phosphatase action. 42-01-12: The different voltage-activated Ca2+ channels are all activated by

depolarization, but they show marked differences in their sensitivity. Currents are potentiated by depolarizing pre-pulses, and this 'facilitation' involves voltage-dependent phosphorylation by PKA. 42-01-12: The binding sites for several different classes of Ca2+ channel

agonists and antagonists have been defined, and naturally occurring peptide toxins have been used to distinguish channel subtypes. Calcium channel antagonists cause vascular relaxation and are used in the treatment of hypertension, angina and cardiac and eerebral ischaemia.

Category (sortcode) 42-02-01: VLG Ca, Le. voltage-gated calcium channels.

Channel designation A unified nomenclature for voltage-gated Ca 2+ channels 42-03-01: A unified nomenclature for voltage-gated Ca2+ channels, based on rules that allow the description of the component subunits, has been proposed1 . The heteromerict channel is described as an a1x!3n'n.a28n complex, where X is a capital letter (5, A, B, C, D, E, etc.) that identifies the gene encoding the 0.1 subunit and n is a number (I, 2, 3, etc.) that identifies the gene encoding the other subunits. Note that the 0.2 and 8 polypeptides are encoded as a single polypeptide chain that is subsequently cleaved to give the two polypeptides that are disulphide-linked. Splice t variants are denoted by y, a lower case letter (Le. alA-a, a1A-h, !31a, !31h etc.). Where no second gene is known, such as for the 0.28 subunit, the capital letter or numerical subsc.ript is omitted. In this system, the skeletal muscle L-type Ca2+ channel, for example, has the composition a1s!31a,a2 8a1.

Current designation 42-04-01: ICa; I(Ca). Currents through the different types of Ca2+ channels are

distinguished by use of subscripts designating the channel-type: for example, current carried by Ca2+ through L-type channels is designated lCa,L.

II

II...-_e_ _ _ry_42 nt

_

Gene family The voltage-gated channel superfamily 42-05-01: Calcium channels form part of the voltage-gated channel superfamilyt, which also includes most sodium channels and some potassium channels (see VLG key facts, entry 41 and compare with VLG Na and VLG K series, entries 44 to 55).

The skeletal muscle Ca 2+ channel contains four protein subunits 42-05-02: Much information about Ca2+ channel structure has come from studies on the skeletal muscle isoform, which consists of a complex of four distinct protein subunits - 01, 02/8, (3, and, (reduction of disulphide bonds can release the 8 'subunit' from the 0.26 protein (see {PDTM}, Fig. 4).

Molecular characteristics of Ca 2+ channel subunits 42-05-03: A wide range of tissue-specific cDNAs encoding calcium channel subunits have now been cloned and sequenced (see Database listings, 4253-01). Basic characteristics of the distinct gene products are given in Table 1. More detailed information specific to each subunit under each field in this entry is indicated by the prefixes 0.1 subunit:, 0.2/6 subunit:, {3 subunit:, and 1 subunit: Because of the large number of cloned variants, these prefixes are sometimes extended with a gene isolate reference name and/or a bibliographic reference where appropriate, e.g. 0.1 subunit, gene rbC2 : (see also Homologous is oform s, 42-21). Further structure - function information for cloned subunits can also be found by reference to the 'primary sequence discussion' papers cited in the Database listings, 42-53.

Subtype classifications 42-06-01: See also Gene family, 42-05 and Homologous isoforms, 42-21. By electrophysiological and pharmacological criteria, there are several subtypes of voltage-dependent Ca2+ channels, distinguishable by their voltage dependence, time course of inactivation, and susceptibility to antagonists and toxins. Briefly, these have been described as L-type, T-type, N-type, Ptype60, Q-type and R_type61 , which for comparative purposes only are listed together within the fields of this entry. Where L, T, N, P, Q or R-channel type-specific data are cited, this is made clear by using an underlined subheading (e.g. 'L-type:') and subunit-specific data are indicated in context. The defining features of these channel types, based on electrophysiological/ pharmacological criteria are given in Table 2. Independent structural and genetic data will provide a more complete classification of these broad functional categories. Table 2 is annotated with cross-references to individual fieldnames which give more detailed information.

Trivial names 42-07-01: T-type Ca2+ channels can be activated by small depolarizations t from the resting potential t, so they are known as 'low voltage activated' (LVA) or 'low-threshold' Ca2+ channels. They are also called 'fast' channels

II

_L...-

e_n_try_4_2_

Table 1. General molecular characteristics of protein subunits encoded by distinct genes forming voltage-gated calcium channels - relationship of different nomenclatures (from 42-05-03) 01 subunit, class A, e.g. encoded by rat gene rbA 2 and human gene CACNL1A4

42-05-04: Brain class A and B al subunits resemble each other, but are less closely related to the al subunits forming putative L-channels. Fulllength class A al subunits do not show homology with class C: (/L-type') al subunits in the putative DHP-binding domain, indicating that alA subunits contribute to a DHP-resistant type of Ca2+ channeI3 - s . 42-05-05: Variants of class A sequences arising from alternative splicing t have been reporteds,6. Two alA cDNA sequences cloned from rabbit brain represent splice variants t that encode channel subunits differing in the Cterminalt sequence, beginning at amino acid residue 2230. The BI-l variant has 2273 amino acids, while the BI-2 isoform has 2424 residues. These two isoforms of the alA subunit, expressed alone or in conjunction with a2, f3 and, subunits, produce functional channels with no significant differences in electrophysiological properties5 . 42-05-06: The rat brain QlA subunit cDNA expressed in Xenopus oocytes results in rapidly activating Ba2+ currents (t==3.3ms) that peak after about 20ms and inactivate slowly (43.6% current remained after 400ms). The half-point for current activation (Vl/2act) is -l.OmV. The current is insensitive to Bay K8644 (10 J.lM) and wconotoxin GVIA (1 J.lM), but is partially (70%) blocked by w-conotoxin MVIIC (5 J.lM) and wagatoxin IVA (23% at 200nM). The alA current is also sensitive to blockade by Cd2+ (760/0 blocked at 10 J.lM). Co-expression t of the alA cDNA with cDNAs encoding f3 subunits increased the average size of the whole-cell Ba2+ current, without changing the rate of activation. While the presence of the f3lb and f33 subunits increased the inactivation rate of the Ba2+ currents, the f32a subunit decreased the inactivation, so that> 70% remained after a 400 ms pulse. In addition, all f3 subunits shifted the I-V relations of the QlA subunit to more hyperpolarized potentials 7 . The electrophysiological characteristics of the channels containing the alA subunit vary with the

II

1i--_e_n_t _ry_4_2

---'_

Table 1. Continued subunit, class A, e.g. encoded by rat gene rbA 2 and human gene CACNL1A4 a1

type of the associated (3- subunit. The channels formed by co-expression of cDNAs encoding Q1A and (32a have the characteristics of P-type currents, while the co-expression of Q1A and (31b or (33 cDNAs produces currents similar to Q-type currents of cerebellar granule cells 7. 42-05-07: 0.2/8 subunits can enhance class A 0.1 subunit activity following co-expression in Xenopus oocytes, which indicates the 0.2/8 subunit may be a structural component of DHPinsensitive Ca2 + channel class, and that the subunit structure of these may resemble that of L-type channels5 . at subunit: rv190kDa (skeletal muscle, low MW form, see Protein M~ below): derived from the 210 kDa precursor by proteolytic removal of a C-terminal peptide8 .

42-05-08: at subunit: rv2l0 kDa (skeletal muscle, high MW form, see Protein M~ below) at subunit: Shows striking homology to other cloned voltage-gated channels, including Na+ (300/0 homology) and K+ - see VLG key facts. at subunit: The skeletal muscle 0.1 subunit comprises four homologous internal repeats, each containing six predicted membrane spanning sequences (see PDTM). The fourth of each of these helices contains positively charged amino acids every 3 or 4 residues and is likely to represent the voltage sensor (see VLC key facts). 42-05-09: at subunit: The Q1 subunit contains binding sites for drugs such as dihydropyridines and phenylalkylamines that have been used to define L-channels pharmacologically (see Receptor antagonists, below). Electrophysiological studies have suggested that the DHP-binding site is localized at the outer surface of the 0.1 -subunit because DHP derivatives which carry net charge (i.e. are membraneimpermeable) only have access to the 'receptor' when applied extracellularly - e.g. 9 • Affinitylabeling experiments have indicated that the DHP-binding site is formed by a combination of the extracellular portions of the S6 helices of the third and fourth repeating domains, as well as the loop linking S5 and S6 of the third domain10,11.

II

_

entry 42

L...---

_

Table 1. Continued subunit, class A, e.g. encoded by rat gene rbA2 and human gene CACNL1A4 (}:1

subunit: stable heterologous expression in mouse L-cells results in formation of DHPsensitive Ca2+ channels12 in the absence of 0.2/8, {3 or , subunits. (}:1

42-05-10: (}:1 subunit: Electrically-stimulated contraction of skeletal muscle can occur in the absence of extracellular Ca2+ (due to the voltage sensor role in this tissue - see ILG Ca Ca RyrCaf). By contrast, cardiac 0.1 subunits exhibit functional L-type Ca2+ channel properties and are dependent upon extracellular Ca2+ for initating contractile events. 42-05-11: (}:1 subunit: Electrophysiological and pharmacological studies of the Ca2+ channels produced by co-expression of rabbit cDNAs encoding o.1A, {31 and skeletal muscle 0.2/8 subunits in Xenopus oocytes 13 show characteristics consistent with those of the Q-type channels of cerebellar granule neurones 14. (see Subtype classifications, below).

subunits, class B, e.g. encoded by rat gene rbB 2 (human gene CACNL1A5)

42-05-12: See also notes on class A 0.1 subunits. The o.IB subunit is a component of N-type Ca2+

subunits, class C, e.g. encoded by gene rbC (rat)2 (human gene CACNL1Al)

42-05-13: A full-length cDNA sequence cloned from rat brain cDNA2 encodes an 0.1 subunit (rbC-I) of 2140 amino acids that is highly homologous (95% identical) to the 0.1 subunit from rabbit cardiac muscle19 and also to other sequences from rabbit lunio. cDNAs encoding a variant 0. subunit, rbC.II, with three extra amino acids in the cytoplasmic loop between domains II and ID and 13 amino acid substitutions in a stretch of 28 amino acids in the S3 region of domain IV, were shown to arise by alternative splicing of transcripts from the same gene2. The same 13 amino acid changes in the rbC-II domain IV S3 segment are also found in the rabbit lun~o and rat aorta21 L-type channel 0.1 subunit.

(}:1

(}:1

II

channels that are sensitive to the cone snail toxin

w-conotoxin GVIA15,16. Transient co-expression of cDNA sequences encoding o.1B (BID), {31a and 0.2 subunits in a human embryonic kidney (HEK-293) cell line produced characteristic N-type Ca2+ channel activity17. N-type channels were also produced by expression of o.1B cDNA in myotubes from mdg mutant mice18.

entry42

I-

_

- - - - - - -

Table 1. Continued subunits, class C, e.g. encoded by gene rbC (rat)2 (human gene CACNL1Al) al

42-05-14: The basic electrophysiological and pharmacological properties of the cardiac and smooth muscle isoforms are similar2, except that Ba2+ currents through the smooth muscle al subunit are IO-times more sensitive to the DHP nisoldipine than those through the cardiac muscle 23 al subunit . Heterologous t expression of 'CaCh2B' cDNA encoding the alC 'smooth muscle' subunit in stably transformed Chinese hamster ovary cells yielded Ca2+ channels with kinetic and pharmacological properties very close to those of native channels in smooth muscle24. 42-05-15: Bmax values are enhanced for the skeletal muscle al subunits when co-expressed with a2/8 subunits following liposome reconstitution25. All expressed alC coding sequences result in highvoltage-activated Ca2+ currents showing minimal inactivation during a depolarizing pulse and sensitive to dihydropyridines.

al

subunits, class D, e.g. encoded by gene rbD (rat)2 (human gene CACNL1A2)

42-05-16: Full-length cDNA clones corresponding to class D have been isolated from rat brain26, human neuroblastoma15, human pancreatic (3 cells27 and a hamster insulin-secreting cellline28 . Expression of the human aID coding sequence, together with cDNAs encoding (32 and a2b subunits, results in dihydropyridine-sensitive L-type Ca2 + channels15, and it is presumed that class D al subunits may be archetypes of a 'neuroendocrine' form of L-type channel. (Expression of the sequence encoding aID alone, or co-expression with cRNA encoding a2b, did not generate functional Ca2+ channels in Xenopus oocytes15.)

subunits, class E, e.g. encoded by rat gene rbE29 (human gene CACNL1A6).

42-05-17: Full-length cDNA sequences encoding a class E al subunit have been isolated from rat (rbE-IT)29, rabbit (BIT-I, BII_2)3o, mouse and human brain31 cDNA libraries. The rbE-IT sequence encodes an al subunit of 2222 amino acids with a predicted molecular weight of 252 kDa with 53 % homology to class A and class B al subunits and 230/0 identity to class C and class D proteins. I Ba carried by the channels produced by transient expression of rbE-II eDNA in Xenopus oocytes first activated at -SOmV and peaked at around -IOmV. The currents were insensitive to Bay K

al

II

_L..-

,

e_n_try_4_2_1

Table 1. Continued 01 subunits, class E, e.g. encoded by rat gene rbE29 (human gene

CACNL1A6).

8644 (10 JlM), nifedipine (10 JlM) and w-conotoxin GVIA (1 JlM), but were partially blocked by wagatoxin IVll (33% block at 200 JlM) and sensitive to block by Ni2+ (IC so = 28 JlM) and Cd2+ (>80% block at 10 JlM Cd2+)29. The presence of the rat brain (31 subunit did not affect the whole-cell current or the rate of activation, but shifted the I-V relation and the voltage dependence of inactivation to more negative potentials29. Expression of the human Q1E coding sequence in HEK-293 cells and Xenopus oocytes produced rapidly inactivating (7 ~20ms at OmV) whole-cell currents that were enhanced about 40-fold by co-expression with sequences encoding human neuroneal Q2 and (3 subunits31 .

II

Ql subunits, class S, e.g. encoded by gene CaChl or CACNL1A3 (human)

42-05-18: The cDNA encoding the Q1 subunit ('Q1S') from rabbit skeletal muscle was the first sequence corresponding to a Ca2+ channel subunit to be cloned32 • The deduced protein of 1873 amino acids has a calculated Mr of 212018 and is encoded by an mRNA of 6.5 kb in rabbit skeletal muscle32 • The mRNA encoding the Q1S subunit has been detected in kidney and brain by sensitive peR-based techniques33 • The Q1 subunit is present in skeletal muscle as two size-variants, a full-length, minor (~S%) form of ~212kDa and a major (~9S%) species of ~ 190 kDa, derived from the longer protein by post-translationalt cleavage close to amino acid residue 16908 .

llccessory/regulatory subunits (intro)

42-05-19: Although a1 subunits are capable of forming Ca2+ channels by themselves, consistent co-purification of other, distinct proteins with a1 has determined they form part of a larger multisubunit native protein complex in cells (see [PDTM), Fig. 4). The origin and arrangement of these aI-associated subunits (a2/ 8, (3, ')') exemplified by the skeletal muscle Ca2+ channel is shown in Fig. 1.

02/6 subunit

42-05-20: Co-purifies with a1 subunit. From skeletal muscle, the ~17S kDa a2/8 subunit shifts to ~ 150 kDa (= a2) upon reduction t of disulphide bonds together with the appearance of the 8 proteins of ~2S kDa, 22 kDa and 17kDa35,36.

entry42

I-

_

- - - - - - -

Table 1. Continued

Figure 1. Subunits of the skeletal muscle L-type calcium channel protein complex. All subunits except f3 have hydrophobic domains and are predicted to be transmembranal or membrane associated. The Q;2 and 8 peptides are linked by disulphide bonds (S-S) and are cleaved from a common precursor. The Q;l, Q;2, 8 and ~ subunits are glycosylated. The molecular weights of the subunits of the L-type channel in skeletal muscle are Q;l = 175kDa; Q;2 = 143kDa; f3 = 54kDa; ~ = 30kDa and 8 = 27kDa34• (Redrawn from Dunlap (1995) Trends Neurosci 18: 89-98.) (From 42-05-13) 0.2/6 subunit

42-05-21: The Q;2/8 protein is the product of a single gene, with the Q;2 portion forming the Nterminal sequence (amino acids 1-934) and the 8 portion forming the C-terminal sequence (amino acids 935-1080)35-37 with a disulphide bridget linking them. 42-05-22: Both the Q;2 and the 8 portion of the Q;2/8 subunit are heavily glycosylated 35-37, e.g. the rabbit Q;2/8 ( rv 125 kDa) contains 18 consensus glycosylation sites and two cAMP-dependent phosphorylation sites38 (see positions under the Database listings, 42-53). Although the Q;2 subunit contains consensus phosphorylation sites, this subunit has not been shown to be phosphorylated in vivo. 42-05-23: mRNA sequences related to the skeletal muscle Q;2/8 subunit are expressed in cardiac and smooth muscle and the eNS, and antiskeletal

II

_L-

e_n_try_4_2_

Table 1. Continued a2/6 subunit

muscle a2/8 antibodies precipitate the DHPbinding complex from both skeletal muscle and brain39. (Note: Channel proteins purified on the basis of high affinity w-CgTx (Le. those of the N-type) are inefficiently precipitated ("-1100-fold slower than those of other L-type Ca2+ channels and would not open during the 2 ms action potential of a skeletal muscle. Extracellular Ca2+ ions are not needed for excitation-contraction coupling in skeletal muscle, because the function of the L-type channel is to act as the voltage sensor for the ryanodine receptor in the SR membrane 77. Skeletal muscle from mouse embryos homozygous for mdg (muscular dysgenesis) mutations lack the al subunit of the skeletal muscle L-type Ca2 + channel 78 and are defective in excitation-contraction coupling: they make action potentials but cannot twitch. Expression of cDNA encoding wild-type t a1 subunits restores dysgenic myotubes t to the normal phenotype 79,8o (See Phenotypic expression, 42-14-01).

Several types of voltage-gated Ca 2+ channel govern exocytosis from presynaptic terminals 42-08-02: Neurotransmitter release involves exocytosis t of vesicles in presynaptic nerve terminals, a process that requires Ca2+ influx through voltage-dependent Ca2+ channels. Exocytosis can be studied directly by fluorescence-imaging techniques following the introduction of the fluorescent styryl dye FMI-43 into pre-synaptic vesicles 81,82. Unloading of the dye from single synapses during electrical stimulation depends on the propagation of action potentials and is completely blocked by tetrodotoxin (1 JlM). The effects of various Ca2+ channel blockers on release of FMl-43 have been studied with rat hippocampal neurones in culture. Individual synapses respond differently to the blockers, implying considerable heterogeneity of Ca2+ channel distribution. Although blocking N-type channels inhibited about 60% of overall FMI-43 release, the data indicate

II

_"--

e_n_try_4_2------1

that L-, N-, P- and other channel types participate in exocytosis in synaptic boutons83 . The effects of w-conotoxin GVIA and w-agatoxin VIA on postsynaptic currents also demonstrate the presence of at least N-type and Ptype Ca2+ channels in pre-synaptic nerve terminals in the rat central nervous system84 .

Synaptic transmission at hippocampal CA3-CAl synapses involves Q-type channels 42-08-03: Selective blockade of voltage-gated Ca2+ channels has been used to determine the channel types involved in synaptic transmission between rat hippocampal CA3 and CAl neurones in brain slices. The use of wconotoxin GVIA (w-CTx-GVIA) (1 f.lM) to block N-type channels caused a rapid but incomplete depression of synaptic transmission (average inhibition of 46%). Specific blockers of L-type (nimodipine, 5 f.lM) and Ptype (w-agatoxin IVA, 30nM) channels had no effect, but transmission was reversibly eliminated by removal of external Ca2+ ions. Application of wconotoxin MVIIC (5 f.lM), a blocker of N-type and Q-type channels, eliminated synaptic transmission, both in the presence and the absence of w-CTx-GVIA, supporting the contention that N-type and Q-type channels entirely account for the Ca2+ currents required for synaptic transmission in these neurones 61 . Note that this conclusion has been both criticized as premature85 and further supported86 • (See Channel modulation, 42-44, for effects of neuromodulators on Q-type channel function).

L-type channels in cochlear hair cells 42-08-04: Voltage-gated calcium channels from chick cochlear hair cells have

the characteristics of L-type Ca2+ channels, including sensitivity to dihydropyridines and insensitivity to the peptide toxins w-Aga-IVA, w-CTx-GVIA and w-CTx-MVTIC. No differences in kinetics or voltage dependence of activation of I Ba were found between tall and short hair cells. These L-type channels are responsible for processes requiring voltage-dependent calcium entry through the basolateralt cell membrane, such as transmitter release and activation of Ca2+ -dependent K+ channels87•

P-type channels are expressed in small cell lung carcinoma cells

42-08-05: A partial cDNA clone t sharing extensive nucleotide identity with sequences encoding a P-type calcium channel 01 subunit has been isolated from a small cell lung carcinoma (SCLC) cell line. Anti-peptide antibodies generated to a unique acidic region in the IVS5-S6 linker of the putative SCLC P-type channel reacted specifically with a cell surface molecule in SCLC cells and inhibited calcium currents in SCLC cells, measured by whole-cell patch clamp t. Similar calcium currents were also inhibited by the P-type channel-specific toxin, w-Aga-IVA. The inhibitory effects of the antibody and the toxin were not additive, consistent with their acting on the same type of channel88 . (The expression of neurone-related P-type channels by SCLC cells is of interest because small-cell lung carcinoma is frequently associated with paraneoplastic t disorders affecting the nervous system.)

1L.-_e_n_t _ry_4_2

----'_

Channel density L-type channels are more abundant than T-type channels in ventricular myocytes 42-09-01: The functional density of L-type channels in guinea-pig ventricular myocytes89, canine cardiac Purkinje cells 90 and smooth muscle myocytes 91 is in the range 1-5/J.1m2. This density is approximately ten times higher than that of T-type channels in guinea-pig ventricular myocytes (0.1-0.3/J.1m2)92. In a rat adrenal medullary tumour cell line, PC12, estimates of the number of L-type channels from agonist- or antagonist-binding assays, 1200-6000 binding sites per ce11 93,94, are in good agreement with those from electrophysiological measurements of peak Ca2+ currents (2500 channels/ceI195 ).

Relative densities of L-type and T-type channels vary with cell type 42-09-02: The relative densities of L-type and T-type channels can be assessed from the maximum current amplitudes. In guinea-pig ventricular myocytes the T-type current amplitude is generally 100 Jlm from the soma, while the L-type currents were primarily recorded in patches within 100 Jlm of the soma149. Whole-cell voltageclamp measurements with isolated dendritic segments ('dendrosomes') from rat hippocampal neurones confirmed these data and showed that T-type channels provide a larger fraction of the Ca2+ influx in dendrites than in cell bodies 15o. At 2 mM Ca2 +, approximating physiological ionic conditions, the maximal amplitude of the T-type tail currents exceeded the combined amplitude of all high voltage activated (HVA) channels. The use of specific channel blockers allowed the high voltage activated Ca2+ currents to be dissected into L-type (1"V20%), N-type (1"V39%), P-type (1"V28%) and Q-type (1"V13%). A low contribution from R-type currents could be detected after inhibiting L-type, N-type and P/Q-type currents with w-conotoxin MVIIC (5 JlM) plus nimodipine (10 JlM)150. (For observations on the activation of dendritic Ca2+ channels by excitatory post-synaptic potentials, see Activation, 42-33.)

L-type Ca 2+ channels are localized in neuronal cell bodies and bases of proximal dendrites 42-16-03: Immunocytochemical studies have shown that L-type calcium channels are localized in neuronal somatat and the bases of proximal dendrites t in rat brain, spinal cord and retina, leading to the suggestion that the neuronal L-type channels may link summed electrical activity in dendrites with biochemical regulatory processes governed by intracellular Ca2+ in the soma39,142.

N-type channels are clustered in tactive zones' on peripheral nerve terminals 42-16-04: N-type Ca2+ channels, labelled with fluorescently tagged w-conotoxin GVIA, were revealed by confocal microscopy to be localized exclusively at the 'active zones' of the frog neuromuscular junction. (They co-localized with the nicotinic acetylcholine receptor, detected with fluorescent a-bungarotoxin.) Cross-sections of the junctions showed that the N-type channels are clustered on the pre-synaptic membrane adjacent to the post-synaptic membrane151 . Note that autoantibodies from patients with the myasthenic disorder Lambert-Eaton syndrome show similar staining patterns152.

Transcript size 42-17-01: The sizes of the mRNA species corresponding to the various Ca2+ channel subunits are summarized in Table 4.

SEQUENCE ANALYSES

11

The symbol {PDTM} denotes an illustrated feature on the channel protein domain topology model (Fig. 4).

II

_'---

e_n_try_4_2_1

Chromosomal location Summary of locations of human genes encoding Ca 2+ channel subunits 42-18-01: The chromosomal locations of human genes encoding subunits of the voltage-gated Ca2+ channels are shown in Table 5.

The gene encoding the 0.2/8 subunit is linked to the malignant hyperthermia locus 42-18-02: Malignant hyperthermia susceptibility (MHS) is an autosomal t dominantt disorder affecting skeletal muscle, manifested as potentially fatal hypermetabolic crises triggered by commonly used anaesthetics and involving a breakdown in the mechanisms regulating sarcoplasmic Ca2+ fluxes. The gene (CACNL2A) encoding the human 0.2/6 subunit of voltagegated Ca2+ channels co-segregatest with the MHS locus in some families and remains a candidate for one of the genes affected in disease families 162. The CACNL2A gene is linked to a polymorphic t dinucleotide-repeat t marker on chromosome 7q162 and has been mapped to 7q21-q22 by somatic cell hybridt analysis 161 .

The human gene encoding

o.1S

maps to a disease locus, HypoPP

42-18-03: Hypokalaemic periodic paralysis (hypoPP) is an autosomal t dominantt disorder involving muscular weakness and paralysis that can be triggered by insulin, adrenaline and ingestion of carbohydrate. Patients have low levels of serum K+ «3 mM) during an episode. The hypoPP locus has been mapped in three independent disease families to chromosome 1q3132. The gene (CACNL1A3) encoding the muscle calcium channel o.lS subunit maps to the same region, and co-segregrates with hypoPP without recombinants in two informative £amilies125. Point mutations in the CACNL1A3 gene have been identified in hypoPP patients (see paragraph 42-14-03).

Some Ca 2+ channel subunit genes are mapped in the mouse 42-18-04: The mouse chromosomal locations of genes encoding the N-type Ca2+ channel al subunit (Cchnal gene) and three neuroneal /3 subunits (the Cchb2, Cchb3 and Cchb4 genes) have been determined by linkage analysis. The N-type aI, /32 and /34 subunits are specified by genes at separate sites on proximal chromosome 2 and the /.~3 subunit gene maps to chromosome IS, within 1.3 cMt of the Wntl marker167. Note that this location places the mouse Cchb3 gene within a large block of syntenyt to human chromosome 12q, and that the human /33 subunit gene maps to 12q13165. The Cch11a2 gene encoding the L-type Ca2+ channel al subunit from mouse brain is located on chromosome 14157 and the Cch11a3 gene specifying the skeletal muscle type alS subunit maps to mouse chromosome 1168 .

I II

Encoding 42-19-01: For subunit-specific data, see Gene family, 42-05 and Domain functions, 42-29.

g t"'t-

~ ~

t:--J

Table 5. Cmomosomallocations of human genes encoding subunits of voltage-gated Ca 2+ channels (From 42-18-01)

Gene CaChl (CACNL1A3) CaCh2 (CACNL1Al) CaCh3 (CACNL1A2) CaCh4 (CACNL1A4) CaCh5 (CACNL1A5) CaCh6 (CACNL1A6) CaAl (CACNL2A) CaBl CaB3 CaGl

II

Subunit encoded

Channel type

Location

Reference

°IB

N

lq32 12p13.3 3p14.3. 19p13.1-p13.2 9q32-q34 lq25-q31 7q21-q22 17q21-q22 12q13 17q24

1257 1557 156

°IA

L L L P/Q

°IS Ole OlD

OlE

02/8 {31 {33

,

R/(T?} all all all all

154 1577 158 1057 159 1597 160 159 1617 162 1637 164 165 587 1637 166

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_2_

Gene organization Multiple genes and alternative splicing generates variety of structure and function 42-20-01: The existence of different classes of calcium channel subunits, encoded by separate genes, together with alternatively spliced variants of those classes, may help account for the wide range of kinetic and pharmacological properties observed for Ca2+ currents in different types of native cells.

Variants of rbC

Q1C

subunit arise by alternative splicing

42-20-02: The rat rbC gene encodes two variants of the alC Ca2+ channel subunit, rbC-I and rbC-II, by alternative splicingt. The major sequence difference between the two, lying within a region of 28 amino acids representing the S3 segment of domain ~ arises from alternative exonst. The isolation and sequencing of rat genomict DNA clones covering this region of the rbC gene shows the 84 bf rbC-I exont lies upstream of a 84 bp rbC-II exon, separated by an intron of 632 bp2. Although the mRNAs encoding the rbC-I and rbC-II variants are found in all regions of the CNS, the relative amounts of the two mature mRNA species vary between brain regions 2 . (See mRNA distribution, 42-13, for more detail).

Alternative Q1C transcription starts in mouse erythroid leukaemia cells 42-20-03: The mouse gene specifying the cardiac ale subunit is transcribed from a different transcription start in the mouse erythroleukaemia cell line, MELC l53 . The MELC transcription-start site is within an intron of the alC gene, and gives rise to a shortened mRNA that encodes a protein lacking the first four transmembrane segments of the alC subunit. Transcription factor t -recogn,ition elements upstream of the MELC start site include three 'GATA boxes t , at -81, -350 and -434 and two 'CACCC boxes t , at -58 and -127 (with reference to the putative cap-site t of the MELC mRNA}l53. (The protein translated from the MELC alC mRNA in Xenopus oocytes does not result in active Ca2+ channels l53 .)

The human CACNL1A4 gene has 47 exons

42-20-04: The exon-intron t structure of the human CACNL1A4 gene, encoding the alA subunit of the P/Q-type Ca2+ channel, has been determined by sequencing all the DNA fragments (average size rv2 kb) containing exons that were subclonedt from 10 cosmidt clones t constituting a 'contigl ' covering the gene. The gene contains 47 exons t, with the untranslated 5' region of the mRNA and the translation-start codon in exon 1 and the stop codon in exon 47. The exons are distributed over rv300 kb of genomic DNA of human chromosome 19. Intron 7 contains a highly polymorphic t (CA}-repeat t sequence (D19S1150) with an observed heterozygosityt of 0.82. The exon-intron organization of the CACNL1A4 gene and a list of exon-specific primert pairs are given in ref. los .

II

l_e_n_t_ry_42

The

_

02-8

transcript is alternatively spliced

42-20-05: The 02 -6 subunit is encoded by a single gene in several mammalian species, but different isoforms of the protein are produced by alternative splicingt of the primary transcript (see Domain functions, 42-29).

Skeletal muscle and brain isoforms of human {3l subunit are encoded by a single gene 42-20-06: Two cDNAs (,81Bl and ,81B2) isolated from human hippocampal libraries and the human skeletal muscle ,81M cDNA are derived from splicet variants of the transcript from a single gene51 . The {lIM and {liBI cDNA sequences are identical except for a region of 154np encoding 52 amino acids in {lIM that is replaced by 19 np encoding seven amino acids in both {liBI and {lIB2. The sequence of {lIB2 is identical to that of {liBI for the first 1482 nucleotides, but {lIB2 contains an elongated 3' end encoding 152 amino acids. The corresponding genomic sequence contains two exons, the 5' one of 154 np specific to {lIM and the 3' exon of 19 np that is present in both {liBI and {lIB2. The,81 transcript undergoes a novel splicing event in {lIB2 that excises a translation termination codon and a poly(A)-additiont signal that are used in both {lIM and {liBI. The mature {lIB2 mRNA is 1.98 kb longer than the {liBI mRNA and encodes an alternative C-terminus t that differs from that of {liBI over the last 34 amino acids and is elongated by a further 118 amino acids.

The human gene encoding the 'Y subunit has four coding exons 42-20-07: The human gene encoding the skeletal muscle Ca2+ channel 'r subunit contains four coding exons, separated by introns of 9.4, 1.0 and 1.3 kb 58 .

Homologous isoforms 0lS

subunits make functional Ca 2 + channels in dysgenic myotubes

42-21-01: The alS subunit of skeletal muscle does not give active Ca2+ channels when produced by heterologous expression in Xenopus oocytes169. Skeletal muscle myotubes from mice with muscular dysgenesis (mdg) have provided a valuable host cell type for heterologous expression of cDNAs encoding (}IS subunits 80. Production of the (}IS subunit by heterologous expression in mdg myotubes results in restoration of dihydropyridinesensitive L-type Ca2+ channels and excitation-contraction (E-C) coupling8o. The ale subunit of cardiac muscle L-type Ca2+ channel also restores E-C coupling, but in this case it does not require Ca2+ entry, reflecting the situation in cardiac cells and contrasting with the (}ISdependent coupling121 . Heterologous expression in mouse L-cells of sequences encoding the skeletal muscle QIS subunit also produces functional L-type Ca2 + channels 170.

Subtypes of the rat brain splicing

OlD

subunit are formed by alternative

42-21-02: The rat brain aID subunit isoform RBol' which contains 1634 amino acids, has 71 % and 76% identity with the rabbit skeletal muscle

II

_L...-

.

e_n_try_4_2_

(aIS) and cardiac (ale) isoforms, respective1r6. Variant cDNA clones isolated from a rat brain library reveal RBal subtypes due to alternative splicingt. Two versions of a 28 amino acid region spanning the S3 segment of domain IV differ in 11 amino acids, all conservative t replacements26 . Note that this same region is alternatively spliced in the rbC-I and rbC-II variants expressed in the rat CNS2 (see 42-20-02). In addition, three variations of the sequence encoding the intracellular region between domains 1 and II were isolated. One predicted product lacks 12 and a second lacks an adjacent 20 amino acids in this region compared with the longest of the three variants26 .

Alternative splicing of the

Q1B

transcript in rat brain

42-21-03: Detailed analysis of the alB sequences expressed in rat superior cervical ganglia (rSCG) revealed four discrete sites where individual cDNA sequences differed from the rat brain-derived clone, rbB-I: (i) a single base change (A to G) resulting in the amino acid change E177G in IS3; (ii) a three-base deletion removing amino acid A415 from the intracellular IS6IISI loop; (iii) a 12-base deletion of sequences encoding a tetrapeptide, SFMG, in the putative extracellular IIIS3-IIIS4 loop, and (iv) a six-base insertion creating an insertion of two amino acids (ET) in the putative extracellular IVS3-IVS4100p171. The G177 form was invariant amongst 11 clones analysed, but both alternatives at the other three positions were found. There was a marked difference in the frequency of both of the SFMG- and ETcontaining variants between brain (+SFMG == 76 %, + ET == 5%) and rSCG (+SFMG==39%, +ET==87%). The dominant forms of alB in ganglia, rnalB-b (LlSFMG, +ET), and brain, rnalB-d (+SFMG, LlET), when produced by heterologous t expression in Xenopus oocytes in the presence of rat brain /33 subunit, show significant differences in the macroscopic rates of channel activation and inactivation and in the voltage dependence of activation. The rna 1B-d currents, on average, activate 1.7-fold faster, inactivate fourfold faster and activate at potentials 5 mV more negative than the rnalB-b currents171 . The injection of rnalB-d eRNA into Xenopus oocytes, in the absence of co-expressed /3 subunit cRNA, produced Ba2+ currents with gating kinetics 'similar to those of the native N channel'. Co-expression of the rba3 subunit enhanced the peak current amplitude by rv3.5-fold, without significantly affecting the gating kinetics 171 .

Marine ray proteins are highly honl0logous to mammalian brain subunits

0'.1

42-21-04: The doe-I cDNA isolated from the forebrain of the marine ray, Discopyge ommata, encodes an al subunit with the following amino acid identities to mammalian Ca2+ channel al subunits: BII (alE subunit of

rabbit brain), 68%; rbB (alB subunit of rat brain), 600/0; BI (alA subunit rabbit brain), 63%. Heterologous co-expressiont of doe-l coding sequences with those encoding mammalian a2 and /3 subunits produced high voltage activated, rapidly inactivating Ba2+ currents that were very sensitive to block by Ni2+ 14. The doe-4 sequence, which is strongly expressed in the electric lobe of D. ommata, encodes a protein with 61 % identity to doe-I, but more closely related to rbB (72~1 identity) and BI (68% identity)172. D. ommata also encodes a Ca2+ channel, doe-2, with high homology to the

II

entry 42

_

A B C (a)

SETLKPD 627 ..

a2a

., 50oHPNLQPKPIGVGIPTINLRKRRPrNQNPKSQEPvTLDFLD539 . . . 610KLEETITQAR~

a2b

.. 50oHPNLQPK

NPKSQEPVTLDFLD 520

591 KLEETITQARSKKGKMKDSETLKPD 615 •.

a2c

.. 50oHPNLQPK

EPVTLDFLD 515

586KLEETITQARSKKGKMKDSETLKPD610 ••

a2d

.. 50oHPNLQPK . . . • . . . . . . . . . . . . . . . . . . . . EPVTLDFLD 515 . . . 586 KLEETITQARY . . . . • . . SETLKPD 603 ..

a2e

.. 500HPNLQPK . • . . . . . . . . . . . . . . . . . NPKSQEPVTLDFLD 520

(b)

.. CCA

591KLEETITQARY

SETLKPD 608 ..

AAG CCT ATT GGT GTA GGT ATA CCG ACA ATT AAT TTA AGG

P

K

PIG

V

G

I

P

TIN

L

R

AAA AGG AGA CCC AAC GTT CAG AAC CCC AAA TCT CAG GAG CCA .. K

(c)

R

a2a/d/e

R

P

V

Q

N

P

K

S

Q

E

P

.. CAG GCC AGA TAT TCA GAA ACC .. Q

a2b/ c

N

A

R

Y

SET

.. CAG GCC AGA TCA AAA AAG GGA AAA ATG AAG GAT TCA GAA ACC

Q

A

R

S

K

KG

K

M

K

D

SET

Figure 2.#Variants of the mouse 02/8 subunit arising from alternative splicingt. (a) Amino acid sequences of the five variants: the alternatively spliced regions (A, B and C) are shown at the top. (b) cDNA sequence and encoded amino acids in regions A and B, with alternative splice acceptort sites (CAG) underlined. (c) cDNA sequence and encoded amino acids in region C. Alternatively spliced sequences and encoded amino acids are shown in bold type. (Reproduced with permission from Angelotti (1996) FEBS Lett 397: 331-7.) (From 42-21-07) mammalian L-type channels172. The separation of Ca2+ channels into L-type and 'non-L-type' must therefore predate the divergence of the cartilaginous marine rays from bony animals >4 x 108 years ago.

Homology between {3 subunit isoforms 42-21-05: The four isoforms of /3 subunit (rabbit) show the following sequence identities: CaBl/CaB2a, 71.0%; CaBl/CaB2b, 71.50/0; CaBl/CaB3, 66.60/0; CaB2a/CaB2b, 97.2%; CaB2a/CaB3, 64.7%; CaB2b/CaB3, 64.9%52.

Mouse and human a2/8 variants arising from alternative splicing (Fig. 2) 42-21-06: A single gene encoding 02/8 subunits in the mouse gives rise to five variants of the Ca2+ channel component by alternative splicing t of the premRNAt. RNAase-protection t assays show that the 02a variant is present in skeletal muscle and aorta, 02b is specific to brain, 02c is specific to heart, where 02d is also found at a lower level, and smooth muscle expresses the 02d and 02e mRNAs l18 . Corresponding human splice variants have been located to skeletal muscle (02a), brain (02b) and aorta (02c)15. The sizes of the various isoforms of the Ca2+ channel subunits and the corresponding mRNA species are shown in Table 6.

II

11

Table 6. Sizes of the different Ca 2 + channel subunits and their mRNAs (From 42-21-07) Class

cDNA

CSkm

alB

ale

aID

Number of aa

1873 1191 a Carp (Cyprinus carpio) skeletal muscle 1852

Rabbit skeletal muscle

alS

alA

Source

BI-l BI-2 rbA-I hBI-l malA

Rabbit brain Rat brain Human brain Mouse brain

2273 2424 2212

Molecular mass (kDa) mRNA size (kb) 6.4 4.4

257 273 252

9.4

5

8.8 and 8.3 8.5 8.6 and 8.2

4

2164 262 252 262 261 265

Rat aorta Mouse brain Mouse erythroleukaemic cell line

2171 2166 2140 2143 2169 2139 1864

243 242 240 240 244 240 211

Human neuroblastoma cell line Human pancreatic islet Rat hippocampus Hamster insulin-secreting cell line

2161 2181 1634 1610

245 248 187 182

Human neuroblastoma cell line

pCARD3 pSCaL rbC-I rbC-IT VSMal mbC MELC

Rabbit heart Rabbit lung Rat brain

aID CACN4 RBal HCa3a

Rat brain Rabbit brain Marine ray, Discopyge ommata

32

212 147 210

2239 2237 2336 2339 2326b

alB-I alB-2 rbB-I Bill doe-4

Reference 109

173

130

116 15 16 18

12-13

174

19 20

ca. 12 and 8

2 21 104

22,9.5 and 7.5

153

ca. 11 8.6 and 6.5 c

27

15 26 28

(1)

~

~ ~ ~

OlE

Rat brain Marine ray, Discopyge ommata

2259 2178 2222 2223

254 245 252 252

a2a a2a a2b a2b a2c a2d a2e rB-a 2

Rabbit skeletal muscle Mouse skeletal muscle Human brain Mouse brain Mouse heart Mouse heart Mouse, smooth muscle Rat brain

1080 + 26 d 1067 + 24 d 1072 + 24 d 1060 + 24 d 1055 + 24d 1048 + 24 d 1053 + 24 d 1080 + 24d

CaB-I (3c (3IM

Rabbit skeletal muscle Human heart Human skeletal muscle

(3lb (3a (3IB2

Rabbit brain

11 and 10.5

30 (t)

~

t'+

29

8

174

124 123 123 122 122 121 122 124

/"V8 and 3.8 e

38

524 522 523

58 58 58

1.9,1.61

Rat brain Human heart Human hippocampus

597 597 596

65 65 65

(32 (3b (3IBI

Rat brain Human heart Human hippocampus

478 477 478

53 53 53

{32a

CaB2a

Rabbit heart

606

68

52

{32b

CaB2b

Rabbit heart

632

71

52

02

{31a

{31b

{31c

II

BIT-1 BIT-2 rbE-IT doe-1

118 15 118 118 118 118 40

48 50 51

49

3.4 3.0g

50 51

15 50 51

~ ~ ~

II

Table 6. Continued Class

cDNA

Source

/32c /32d

CaB2c CaB2d

Rabbit brain Rat brain

{33

{33

{34 f

Number of aa

Molecular mass (kDa) mRNA size (kb)

Reference

604

68

46

CaB3

Rat brain Rabbit brain

484 477

55 54

52

{34

Rat brain

519

58

Rabbit skeletal muscle Human skeletal muscle

222

25

52

53

54

1.2

57 58

Notes: 1. For use of rbA-rbD nomenclature see, for example, refs. 2,4,175,176. 2. The L-type Ca2+ channel is composed of five subunits (aI, a2, (3, 8 and f). a The mRNA encoding this variant of the skeletal muscle Ca2 + channel al subunit is 2 kb shorter than the alS mRNA that predominates in adult tissue, due to an internal deletion109. b A variant doe-4 cDNA encodes a product with a block of 20 extra amino acids inserted after residue 406, due to alternative splicing172. C The rat brain aID sequence is expressed via an 8.6 kb transcript in brain, heart, aorta, uterus and lung, and as a 6.5 kb mRNA in aorta and skeletal muscle26 . dThe rabbit skeletal muscle a2 subunit is encoded as a protein of 1106 amino acids that includes a signal sequence of 26 residues38 . The human15 and mouse l18 brain homologues have similar structures, but with 24 amino acid leader peptides. e The 8 kb mRNA species was present in rat brain, skeletal muscle, heart and lung; the 3.8 kb mRNA was present as a minor species in rat skeletal muscle40. fThe 1.9 and 1.6kb species were present in rabbit skeletal muscle, but the {31a mRNA in rabbit brain was a 3.0kb species48 . 48 g 3.0 kb is the size of the cognate mRNA species in rabbit brain .

(I)

Ia ~ ~ ~

II...-__ _ _ry_42

_

en t

.....

... u

1S

(Carp Sk)

.....- - - - -.. u 1s(Sk)

DHP-sensitive

. - - - - - -.. u 1c(rbC-II)

....- - - - - - u 1A(rbA-1) ........- - - - - u 1s (rbB-1) ....- - - - - u 1s(doe-4) , . . . - - - - - u 1E(doe-1)

DHP-resistant ....-

I

20

I

40

....- - - - - - u1E(rbE-II)

i

60

80

i

100

Percentage identity

Figure 3. Similarity tree of primary sequences of cloned voltage-gated Ca 2+ channel al subunits. The predicted amino acid sequences of the following major classes of al subunits are compared: Carp Sk 173, Sk32, rbc-II2 , alD 15, rbA-I4 , rbB-I16, doe-1, doe_4 172, rbE-II29• (Reproduced with permission from Stea et al. (1995) In Handbook of Receptors and Channels (ed. R. A. North), pp.113-51. CRC Press, Cleveland, Ohio.) (From 42-21-08)

Comparison of 0'.1 subunit sequences 42-21-07: Sequences encoding six distinguishable classes of voltage-activated Ca2+ channel al subunits have been characterized by molecular cloning. The primary sequences of these subunits have been compared by means of a 'similarity tree', as shown in Fig. 3.

An alternative O'.le transcript produced in a mouse erythroleukaemic cell line 42-21-08: Molecular cloning of cDNAs derived from mRNA of a mouse erythroleukaemia cell line (MELe) revealed a truncated ale transcript, produced by transcription of the mouse alC gene from a start site downstream of the one used in heart. The MELC mRNA encodes an alC subunit variant that is missing the first four putative transmembrane regions of motif I of the cardiac alC subunit. The AUG codon that specifies Met299 of the cardiac al subunit is used as the translation-initiation signal in the MELC mRNA, which encodes a polypeptide of 1864 amino acids. Messenger RNA species containing the 5' sequence of the mouse cardiac alC mRNA were not detectable in MELC cells. Injection of MELC alC cRNA into Xenopus oocytes failed to produce functional Ca2+ channels153 .

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_2__1

The trp gene of Drosophila has homology with

al

subunit genes

42-21-09: The trp ('!ransient !eceptor I!otential') gene of Drosophila melanogaster, mutations in which are affected in phototransduction such that the light response quickly declines to baseline during prolonged intense illumination, encodes a light-activated Ca2+ channel subunit177 with sequence homology to the al subunit of vertebrate Ca2+ channels, particularly in the membrane-spanning regions S3_S6 178 . Three of the four basic amino acids in S4 that constitute the voltage sensor of the al subunits are absent from the trp S4 region, though the overall homology can still be discerned178. A cDNA isolated from a Drosophila head cDNA library encodes a calmodulint -binding protein that has 39% overall identity with the trp protein, including 76% identity in the region spanning S3_S6 178. The gene encoding this protein has been named the !.ransient receptor Qotential-like (trpl) gene178.

A modulator of pre-synaptic N-type Ca 2+ channel activity identified

by tsuppression cloning' 42-21-10: On injection into Xenopus oocytes, the mRNA from the electric lobe of the electric ray, Torpedo californica, gives rise to a characteristic Ca2+ current that resembles N-type currents of mammalian neurones. A novel cDNA sequence has been isolated on the basis that its in vitro transcript inhibits the ability of the electric lobe mRNA to generate Ca2+ currents in Xenopus oocytes. (This approach, involving screening for clones via the inhibition of expression of known mRNAs, has been called 'suppression cloning'.) This in vitro transcript is an antisense RNA that is complementary to the sequence of part of the mRNA encoding a subunit of the Torpedo Ca2+ channel, or an associated protein that stimulates channel activity. The expression of the corresponding sense RNA selectively potentiates the N-type Ca2+ currents obtained from translation of electric lobe mRNA in Xenopus oocytes. The full-length cDNA encodes a predicted ~a2+ ~hannel ~ubunit l' (CCCS1), containing 195 amino protein, '~andidate acids, that is cysteine-rich and has "V70% identity to 'cysteine string proteins' (CSP) of Drosophila 179 . These CSP proteins have been immunochemically detected at nerve-ending membranes at synapses throughout the Drosophila nervous system (quoted in ref.179).

Protein molecular weight (purified) Two isoforms of a subunit in skeletal muscle L-type channels 42-22-01: The al subunits of the L-type Ca2+ channel of rabbit skeletal muscle occur as two isoforms t : a 212kI)a species that comprises about 5% of the total, and a major (95%) 190kI)a form, derived from the 212kDa form by proteolytict cleavage near amino acid residue 16908 .

Complexes containing the shorteneti form of the functional channels

alS

subunit are

42-22-02: It was proposed that the smaller form of alS could act as the 'voltage sensor' working in conjunction with the ryanodine receptor (see ILG Ca

II

i_e_n_ _ry_42

_

t

Ca RyR-Caf, entry 17), whereas the larger form acts as an L-channel. Countering this suggestion, L-type channel activity has been observed following exclusive reconstitution of solubilized low molecular weight form into phospholipid bilayers25 . In addition, a shortened form of the rabbit skeletal muscle Ql subunit, missing the final 211 amino acids from the C-terminust and therefore similar to the naturally occurring truncated form, obtained by expression of an appropriately deleted eDNA, has been shown to function as both voltage sensor and calcium channel in dysgenic myotubes 180. The L-type calcium channel purified from skeletal muscle also contains the {3 subunit of 57 kDa, the '"Y subunit of 25 kDa and the disulphide-linked Q2/8 subunits of 125 kDa 589 .

L-type Ca 2+ channels can be purified as an w-conotoxin GVIA-binding protein

42-22-03: L-type Ca2+ channels from rat brain have been purified by

monitoring the binding of 12sI-labelled w-conotoxin GVIA and photoincorporation of N-hydroxysuccinimidyl-4-azidobenzoate-[12SI]-w-conotoxin GVIA. The purified w-conotoxin receptor complex comprised a 230 kDa Ql subunit, a 140 kDa Q2 subunit, and additional proteins of apparent molecular masses of 110, 70 and 60kDa62 . Photoaffinityt labelling of partially purified protein fractions from bovine 181 , porcine182, guinea pig183 and chick184 heart showed that the Ql subunit of the cardiac L-type channel has an apparent molecular mass of 195 kDa. I"V

Disulphide linkage to the {; subunit associates the the membrane-bound channel

Q2

subunit with

42-22-04: The 0.2/6 subunit from skeletal muscle is a glycoproteint with an

observed molecular weight of 175000 under non-reducing conditions. After reduction, the Q2 subunit has an observed size of 150 kDa, and 8-related species of 25 kDa, 22 kDa and 17 kDa are produced. The stoichiometric ratios of these species in the Q2/8 complex are 1.0 (Q2), 0.31, 0.47 and 0.08 in order of decreasing size. All three 8-related species have the same Nterminust, starting at Ala935 of the encoded propeptide, and differ from each other only in the degree of N-linked glycosylation t . After deglycosylationt with endoglycosylasest the Q2-peptide has a molecular weight of 105kDa and the 8 peptide one of 17kDa. Under reducing conditions, the Q2-peptide can be solubilized from skeletal muscle membrane preparations by extraction at pH II, a procedure that removes peripheral but not integral membrane proteins. The 8 subunit remains in the membrane fraction under these conditions. The Q2 subunit is thus associated with the Ca2+ channel in the membrane through S-S linkage to the 8 subunit36 .

Purified N-type channels contain an additional glycoprotein subunit

42-22-05: A functional N-type Ca2+ channel complex purified from digitonin-

solubilized rabbit brain membranes consisted of a 230 kDa subunit, QIB, tightly associated with a 160kDa subunit, Q2/8, a 57kDa subunit, {33, and an additional 95 kDa glycoprotein subunit. Affinity-purified antibodies prepared against the 95 kDa subunit precipitated 740/0 of the 12sI-labelled

II

_L-...

e_n_try_4_2_1

w-conotoxin-binding sites from rabbit brain, without precipitating the brain L-type channels. The purified complex formed a functional Ca2+ channel with the same pharmacological and electrophysiological properties as those of the native w-conotoxin-sensitive N-type channel in neurones 63 .

Molecular weights of native (3 subunits from Western blots

42-22-06: Antibodies against specific Ca2+ channel /3 subunit subtypes recognize {31 subunits of 78 and 80 kDa, a {32 subunit of 74 kDa, a {33 subunit of 58 kDa and {34 subunits of 55 and 59 kDa in extracts from rat brain membranes. The /32 antibody detected subunits of 70, 74 and 87kDa in rat cardiac microsome membranes. The rat phaeochromocytoma cell line, PC12, produced readily detectable /33 subunit, and a /32 subunit of 70kDa could be detected only after "J10-fold enrichment on an immunoaffinity column106. Immunoprecipitation with anti-/3 subunit antibodies shows that the 0lB subunits of N-type channels are associated with /32 and /33 subunits in PC12 cells 106 (see Protein interactions, 42-31). The /3 subunit from skeletal muscle Ca2+ channels has a reduction-insensitive apparent molecular mass of 52-65 kDa as measured on SDS_PAGEt185-187.

The skeletal muscle channel, subunit 42-22-07: The purified, subunit from the skeletal muscle dihydropyridine (DHP) receptor is a heavily glycosylated protein with an apparent molecular mass of 30-33 kDa as measured on SDS_PAGEt186-188.

Protein molecular weight (calc.) 42-23-01: See Table 6, 42-21-07.

Sequence motifs 42-24-01: For subunit-specific data, see Gene family, 42-05. All six of the conserved consensus sites for cAMP-dependent protein phosphorylation on the 01(212) subunit are located in the C-terminal tail (see {PDTM}, Fig. 4). Note: The truncated 01(175) form (see paragraph 42-29-01) lacks at least three of these motifs.

N-glycosylation (Asn-X-Ser/Thr) balD: 44 (NSS), 96 (NSS), 155 (NST), 225 (NHS), 329 (NGT), 463 (NTS), 478 (NVS), 1547 (NAT), 1635 (NTT), 1705 (NTT), 1762 (NMS), 2013 (NGS)15. PKA {K/R-K/R-X-S/T} biD: 464 (KRNT), 687 (KRST), 1700 (RRDS), 1773 (KRPS), 1922 (RRSS), 1932 (RRQS)15. PKC {S/T-X-K/R} balD: 45 (SSK), 81 (SQR), 91 (SKK), 228 (SGK), 502 (SRR), 646 (SMK), 683 (TKR), 913 (SFR), 1310 (SNR), 1670 (TKR), 1695 (SDR), 1707 (THR), 1724 (TEK), 1788 (SHK), 1878 (SER), 1902 (SRR), 1905 (SPR), 1917 (SHR), 1966 (SSK), 1977 (STR), 2032 (SYR), 2046 (SFR), 2052 (SDK), 2123 (SHR)15.

II

l_e_n_t_ry_42

----'_

Southerns The

QIS

subunit is encoded by a single gene in the mouse

42-25-01: Southern blots of mouse (strain 129/ReJ) liver DNA, cut with restriction enzymes Apal, BamHI, EcoR! and Kpnl and probed with different regions of the rabbit skeletal muscle DHP receptor cDNA, are consistent with a single gene encoding the mouse DHP receptor80. Comparison of the blots of DNAs from wild-type and mutant (mdg) mice with muscular dysgenesis, probed with cDNA segments corresponding to different sections of the rabbit coding sequence, revealed differences between +/+ and mdg/mdg DNAs in at least two regions of the mouse gene encoding the skeletal muscle DHP receptor80 . Southern blots of rabbit and human DNAs cut with the restriction enzymes EcoR! and BamHI and probed with a rabbit alS cDNA at high stringencyt are consistent with a single gene. Probing at reduced stringency reveals other sequences related to the alS gene38 .

A single gene encodes the rbC-I and rbC-II Ql subunits in rat 42-25-02: Southern blots of digests of rat genomic DNA with the EcoRI, Pstl, HindllI and EcoRV restrictiont endonucleases gives identical patterns of single bands when probed with oligonucleotides specific for the rbC-I and rbC-II al subunit coding sequences. The conclusion that a single gene encodes the two subunits by differential splicing was confirmed by PCR analysis and by direct sequencing of genomic clones2 . The exons t encoding the rbC-I and rbC-II domain IV S3 segments are separated by an intront of 632 bp, with the rbC-I-specific exon located upstreamt of the intron2 .

Restriction map of the human CACNLIA4 gene 42-25-03: The human CACNL1A4 gene that encodes the alA subunit of the P/Q-type Ca2 + channel has been isolated as a set of ten overlapping cosmidt clonest. The DNA cloned in each cosmid has been restriction mapped and the positions of the BcoRI sites presented105. The collection of ten cosmids, covering >300 kb of genomic DNA, is available on request (see 42-10-02).

Southerns of the mouse

Qle

gene

42-25-04: Southern blots of mouse (DBA/2) genomic DNA, digested with BgID, Ndel, Pstl and Xhol and developed with an alC cDNA probe, are shown in ref. 153 . These blots also show the genomic rearrangement in both alleles of the alC locus in the mouse MELC erythroleukaemic cellline153 (see paragraph 42-21-09).

STRUCTURE AND FUNCTIONS

Domain arrangement 42-27-01: See Subtype classifications, 42-06.

_"--

e_n_try_4_2__1

Domain conservation Mutations affecting the pore region alter cation selectivity 42-28-01: Certain deletion mutants of the Shaker K+ channel display functional similarities to voltage-gated Ca2+ channels (see Selectivity, 42-40 and ref. 189). This may be due to a Gly-Asp pair in the K+ channel deletion mutant matching a Gly-Glu pair in three of the four pore-region domains of a Ca2 + channel. Mutation of the acidic Asp to Glu alters the Ca2+-blocking affinity rvl0-fold (see Selectivity, 42-40). Mutations changing the equivalent residues (K1422 and A1714) to glutamates in Na+ channels result in Ca2+ channel-like ion conduction properties190 (see Domain conservation under VLC Na, 55-28 and Selectivity under VLC Na, 55-40).

Domain functions (predicted) The multimeric nature of L-type channels 42-29-01: L-type channels: Native L-type channels are a complex of five protein subunits, aI, a2, {3, 'Y and 8 (for reviews, see refs. 34,35,191-193). The a1 subunit constitutes the central functional component of the complex, contains the receptor sites for calcium channel antagonists (see below) and can function independently as a voltage-gated ion channel in heterologous expression systems (see [PDTMj, Fig. 4). The alS subunit of skeletal muscle is present as a minor, full-length species of 214 kDa and a major, processed form sometimes referred to as ad 175), but shown to have an Mr of 190 kDa by Fergusont plot analysis8 .

The mode of excitation-contractiol1 coupling is determined by the subunit

0'.1

42-29-02: L-type channels: The putative intracellular loop connecting homologous domains IT and III (see [PDTM), Fig. 4) was found to be sufficient to induce direct calcium release through E-C coupling t in skeletal muscle 194 . The IT-III loop domain is likely to interact with the SR Ca2+ release channel or may be coupled through an additional 95 kDa protein195. Chimaerast of the skeletal and cardiac L-type channels expressed in dysgenic myotubes (see Phenotypic expression, 42-14) have shown that if the IT-III cytoplasmic loop region derives from skeletal muscle, induction of Ca2 + release does p.ot require extracellular Ca2 +194 .

The DHP-binding site on L-type channels is external 42-29-03: L-type channels: Most DHP derivatives used in electrophysiological experiments are uncharged molecules at physiological pH and their access to the DHP-binding site is not limited by the lipid bilayer of the cell membrane. When a fully charged, quaternary DHP derivative, SDZ 207-180 (1 ~M), was used, external but not internal application blocked L-type channel current in a voltage-dependent manner196. Note: that dihydropyridines bind strongly to lipid bilayers and concentrate along the surface of the bilayer197.

Sites of DHP and phenyalkylamine binding to 0'.1 subunits identified 42-29-04: L-type channels: The first repeat of the putative transmembrane helices is required for fast-activation kinetics of the Ca2 + current as seen in

II

l_e_n_t_ry_4_2

---'_

cardiac muscle198 (see Activation, 42-33). The DHP-binding site of the QlS subunit of L-type channels has been localized by photoaffinityt labelling followed by limited proteolysis and peptide mapping. Diazipine and azidopine label peptides spanning 989-1022, within the extracellular loop N-terminal to the sixth membrane-spanning segment of the third repeat (IIIS6)10,11. Isradipine, whose photoreactive group is intrinsic to the binding centre of the antagonist, covalently labels a peptide corresponding to residues 1023-1088, including transmembrane segment IIIS6 and adjacent extracellular and intracellular residues 199. Dihydropyridines also photoaffinity label a sequence in the region of IVS6 11,200. The latter sequence, extending from E1349 to W1391, is also involved in phenyalkylamine binding11 . A model for dihydropyridine binding to the L-type Ca2+ channels proposes a hydrophobic binding cleft at the extracellular end of the interface between domains III and ~ allowing DHP binding to affect domain interactions that are important in channel gating199. (See 42-29-05 for the determination of sites that are crucial for DHP action).

Mapping of critical sites for dihydropyridine action 42-29-05: L-type channels: Critical sites for the action of dihydropyridines (DHPs) on the Ql subunits of L-type Ca2+ channels have been mapped by construction of chimaerict subuits in which segments of the DHP-insensitive rat brain alA subunit, BI-2, were introduced into the cardiac ale subunit. Currents produced by chimaeric channel subunits were analysed after heterologous co-expression with Q2 and 13a subunits in Xenopus oocytes. Currents produced by chimaeric channels containing the S3-S6 region of motif IV of the BI-2 subunit were insensitive to the DHP agonist, (+)-(S)-202-791, and the antagonist, (-)-(S)-202-791 (both at 1 JlM). A chimaera containing the S3-S5 region of motif IV of the BI-2 subunit showed normal sensitivity to both DHPs, but one in which the substitution covered the region S2-S6 of region IV showed a dramatic decrease in the sensitivity to the DHP agonist, (+)-(8)-202-791 (ECso 213 OM, compared with ECso == 89 OM for the control QlC channel), and to other DHP agonists, while retaining normal sensitivity to antagonists (ECso 787 OM). These data show that the important area for DHP action on QlC subunits is the S5-S6linker of motif I~ and that there are distinct sites of interaction for the agonist and antagonist enantiomers of 202-791 201 • Note that these findings contradict the model based on photoaffinity labeling studies (see paragraph 42-29-04), suggesting that the sites labelled by photoaffinity ligands are not important for functional effects of DHPs.

The Ole subunit contains a Ca 2+ -binding motif 42-29-06: L-type channels: The site required for Ca2+-dependent inactivation of L-type cardiac Ca2+ channels is a Ca2+ -binding EF-handt motif at residues 1499-1533 in the C-terminus of the rat alC subunit202,203 (see Inactivation, 42-37). A conserved motif in the I-II linker of the

01

subunit binds to the {3

subunit 42-29-07: L-type channels: 3sS-labelled rat Ca2+ channel 13la, 13lh and 133 subunits prepared by coupled in vitro transcription and translationt of the

II

_ entry 42

1.-.---

_

cognate cDNAs bind to the QIS subunit of rabbit skeletal muscle after separation by SDS-PAGEt and transfer to nitrocelluloset 2 0 4 . Cloning of small fragments of the QIS eDNAt into an expression vectort allowed the identification of peptide 'epitopes' that were capable of interacting with the radiolabelled I3tb subunit probe. All interacting peptides shared a 45 amino acid sequence extending from amino acid 341 to 385 of the atS subunit204 • This region is located within the putative cytoplasmic linker between repeats I and II of the QIS subunit (see {PDTM}, Fig. 4). Similar regions of the rabbit QlC-a, rabbit QlA and rat QlB subunits also allowed in vitro interaction with the labelled ,BIb subunit, even though the I to II linker sequences of the four Ql subunits show only 19% overall identity. Sequence comparisons of the interacting Ql subunit peptides reveal an interaction motif that can be minimally described by QQ-E-L-GY-WI-E, which is always found 24 amino acids downstream of the S6 transmembrane element of repeat I. Site-directedt mutagenesis and directed deletion of the codonst for residues in this motif confirms their importance in the interaction with the,B subunit and in the in vivo modulation of channel properties204 . A glutamate in SS2 repeat III is a determinant of ion selectivity 42-29-08: L-type channels: The peptide loop joining transmembrane regions S5 and S6 in all four repeats of the Ql subunit is hypothesized to fold back into the membrane to form the lining of the channel. These SS2 segments of the four repeats of the at subunits each contain a similarly placed glutamic acid (E) residue. Mutational alteration of these E residues in repeats III and N changes the ion permeation properties of the channel. The channels containing the QlA subunit with the change E1469Q, produced by transient co-expression of cRNAs encoding QlA, Q2 and ,B subunits in Xenopus oocytes, shows a shift in reversal potential t from the 57.6mV of the wild-type channel to 37.5mV. Oocytes expressing E1469Q channels also pass much higher ratios of outward currents to inward currents than those expressing wild-type channels. The E1469Q change also reduced the sensitivity of the channel to block by Cd2+, changing the ICso from rv 1 J.lM to rv200 J.lM. The mutation had no significant effect on the sensitivity of the channel to C02+ or Ni2+, and marginally decreased the sensitivity to Li3+. The corresponding mutation in repeat ~ E1765Q, had a smaller effect on the reversal potential, changing it to 49.8 mV, but no significant effect on the sensitivity to the inorganic cations205 • (Note that the corresponding amino acids to the E1469 and E1765 of QlA in the rat Na+ channel II are K1422 and A1714: changing either of these Na+ channel residues to E altered the permeation properties of the Na+ channel to resemble those of a Ca2+ channel190 (see VLC Na, entry 55).

The C-terminal tail of the cardiac opening

Gle

subunit affects channel

42-29-09: L-type channels: The cardiac QlC subunit has an intracellular cterminal ltail' of 665 amino acids. The expression in Xenopus oocytes of cDNAs containing deletions, producing QlC subunits lacking from 307 to 472 amino acids at the C-terminus, led to Ba2+ currents that were 4- to 6fold higher than those obtained with the wild-type QlC. There was no

II

lL.....-.._ _ _ry_4_2

_

en t

change in the amount of charge movement during voltage-dependent gating, or in the unitary conductancer. Removal of up to 70% of the intracellular Cterminal sequence increased current density by facilitating the coupling between the voltage-dependent gating and channel opening, leading to an increased Popen for the mutant channels206 . The intracellular action of the proteolytic enzyme trypsin in ventricular myocytes can also cleave the Cterminal tail of the al subunit and generate increased currents207 (see Inactivation, 42-37-13).

Determinants of voltage-dependent inactivation kinetics 42-29-10: L-type channels: Analysis of the properties of channels containing

chimaeric t al subunits showed that the key determinants of the kinetic differences in voltage-dependent inactivation are localized to a region of 600/0) inhibition of wholecell Ba2+ currents and reduced the rate of inactivation in both cases. The inhibitory action of syntaxin 1A on ICa,L involves interaction with the alC subunit, since it obtains in the absence of the auxiliary a2/8 and 13 subunits. A truncated form of syntaxin, containing only amino acids 1-267 and missing the C-terminal membrane-anchor sequence, was ineffective in inhibiting currents generated by the full complement of L-type subunits, suggesting that the interaction depends on association of syntaxin with the membrane. Co-expression of cRNAs encoding SNAP25 and N-type channel subunits reduces the current amplitude (rv 29%) and shifts the steady-state voltage-dependence of inactivation by rv 10mV towards positive potentials. SNAP25 is also able to reverse the stronger, syntaxin-generated inhibition, without affecting the syntaxin-reduced rate of inactivation. SNAP25 also reduces (by rv21 0/0) the current amplitude obtained with L-type channels and partially reverses the strong inhibitory effect of syntaxin 1A. These data are consistent with the formation of complexes involving L-type or Ntype channels, syntaxin 1A and SNAP25 that are likely to be important in synaptic vesicle t docking and neurotransmitter release223 .

Interaction of N-type channel with synaptic core complex is important for neurotransmitter release 42-31-08: N-type Ca2+ channels can be extracted from rat brain membranes and partially purified in association with the synaptic core complex, containing syntaxin, SNAP25 and synaptobrevin. Syntaxin is co-immunoprecipitated from this complex by an antibody (CNB3) against a C-terminal peptide of the N-type Ca2+ channel alB subunit. This co-precipitation is blocked by incubation with recombinant peptide 'Ln_m(718-963)' (5.5 JlM), containing the synaptic protein interaction ('synprint') site from the intracellular loop connecting domains II and ill of alB. A control fusion protein, Ln_m(670-800) from the alS subunit of an L-type Ca2+ channel, has no blocking effect224 . Microinjection of Ln_m(718-963) (1.6 JlM in cell soma)

II

_ 1.....---

entry42

I ,

into pre-synaptic superior cervical ganglion neurones (SCGNs) reversibly inhibited synaptic transmission (",24% decrease in excitatory post-synaptic potentials by ",15 min, recovering to control level by 30-40 min). The Ln_m(718-963) peptide was not inhibitory to Ca2+ currents through N-type channels in SCGNs, measured by whole-cell patch-clamp recording. These results support the hypothesis that binding of the synaptic core complex to pre-synaptic N-type Ca2+ channels is necessary for Ca2+ influx to trigger the release of neurotransmitter224.

Protein phosphorylation Phosphorylation by protein kinase A activates L-type Ca 2+ channels 42-32-01: L-type: Phos/PKA: The activity of L-type Ca2+ channels can be regulated by cAMP-dependent protein phosphorylation. Skeletal muscle contractile force and L-type Ca2+ currents are increased by agents that increase intracellular cAMP. Phosphorylation of purified channels increases the number that are active in ion conductance186 (see also Sequence motifs, 42-24). Purified skeletal muscle Ca2+ channels do not form functional channels in lipid bilayers unless phosphorylated by PKA225 and their activity is linearly related to the extent of phosphorylation226 . Under basal conditions, 30-40% of the PKA phosphorylatable sites of rat skeletal muscle Ca2+ channels were phosphorylated: additional phosphorylation occurred when myocyte cAMP was elevated by 8-bromo-cAMP, forskolin, ,B-adrenergic agonists or calcitonin gene-related peptide227. Phosphorylation following stimulation by ,a-adrenergic agonists can increase macroscopic t lca by "'tenfold228, but is blocked by PKI (1 JlM), a specific inhibitor of PKA229 . Single-channel conductance is !lot affected by phosphorylation, but channels shift to gating modes with repeated or long channel openings, and more 'dormant-state' channels are activated230,231 (see also Rundown, 4239). Note that in intact cardiac myocytes, elevation of cAMP leads to phosphorylation primarily of the ,B subunit of the voltage-gated Ca2+ channel232 . The L-type channels of dorsal root ganglion233, hippocampal234 and cerebellar granule235 neurones are also activated by PKA. Facilitationt of Ltype Ca2+ channels by depolarizing voltage steps requires a protein phosphorylation event and is blocked by inhibitors of PKA236,237 (see Activation, 42-33).

The

alS

subunit is phosphorylated in vitro by PKA

42-32-02: Both the 'short' (190kDa) and the 'long' (212kDa) forms of the rabbit skeletal muscle (tIS subunit are rapidly phosphorylated in vitro by PKA, at Ser687, located in the intracellular loop between repeats IT and Ill, and slowly phosphorylated at Ser1617, within the intracellular C-terminal domain238 . The longer form is also phosphorylated by PKA on Ser1854239.

cAMP-dependent phosphorylation is ineffective following heterologous expression 42-32-03: Although the Ca2+ current (lca) of rabbit cardiac myocytes increases threefold during internal dialysis with 5 mM forskolin plus 50 mM ffiMX, Ca2+ channels obtained by expression of cDNAs encoding the QIC-a and QIC-b

II

lL....-e_n_t_ry_42

-----I_

subunits in stably transfected Chinese hamster ovary (CHO) and human embryonic kidney (HEK) 293 cell lines were not affected by treatments designed to stimulate cAMP-dependent phosphorylation24o . In similar experiments in stably transfected human embryonic kidney (HEK-293) cells producing cardiac alC and f32 subunits, although basal Ca2+ current was not increased by forskolin, inhibitors of PKA did decrease the basal current. This decrease was reversed by either forskolin or okadaic acid241 . It is therefore apparent that heterologously expressed Ca2+ channel subunits can be fully phosphorylated under basal conditions.

Phosphatase inhibitors increase L-type calcium currents in frog cardiac myocytes 42-32-04: The phosphatase inhibitors okadaic acid (OA) and microcystin (MC) caused large increases in L-type calcium currents (ICaL) in frog cardiac myocytes in the absence of f3-adrenergic agonists. The dose-response curves for ventricular cell lCaL fitted with a single-site relationship and K l / 2 values of 1.58 JlM for OA and 0.81 JlM for MC. These data suggest that the predominant form of phosphatase active on the L-type channels in this cell type is protein phosphatase 1. Inhibition of phosphatase 2B (calcineurin) was without appreciable effect. Reducing intracellular ATP concentrations did not affect the basal ICa, but ATP-depletion completely prevented the increase in lca induced by OA or MC, demonstrating that the stimulatory effects of OA and MC on lca depend on a phosphorylation reaction. Internal perfusion of PKI(S-22), a peptide inhibitor of PKA, was without effect on basal levels of ICa, suggesting that this kinase is not phosphorylating these channels under basal conditions. Although PKI is capable of completely blocking the response of lca to isoprenaline or forskolin, it does not affect the increase in ICa induced by MC or OA. The increased lca obtained in the presence of saturating concentrations of OA or MC could be further elevated by application of a l3-adrenergic t agonist, forskolin or cAMP, showing that PKA does not mediate the OA response and that phosphatase inhibition does not result in complete phosphorylation of PKA sites242 . Comparative note: The effects of phosphatase inhibitors on L-type Ca2+ channels are generally opposite to those (i.e. decreased current amplitude) on the delayed rectifier K+ channel, lK. (For details, see Protein phosphorylation under VLC K DR [native], 45-32).

A-kinase anchoring protein required for phosphorylation of L-type channel subunits by PKA

42-32-05: Voltage-dependent potentiation of skeletal muscle L-type Ca2+ channels requires phosphorylation by PKA. Potentiation by endogenous PKA is prevented by a peptide derived from the conserved kinase-binding domain of a PKA-anchoring protein (AKAP). This peptide does not inhibit potentiation in the presence of exogenous catalytic subunit of PKA (2 JlM), suggesting that kinase anchoring is specifically blocked by the AKAP peptide. The AKAP peptide did not affect the level or voltage dependence of basal Ca2+ channel activity before potentiation, suggesting that proximity between the skeletal muscle Ca2+ channel and PKA is critical

II

_L..-

e_n_try_4_2_1

for voltage-dependent potentiation of Ca2+ channel activity but not for basal activitr43,244. In HEK-293 cells transiently expressing cardiac L-type Ca2+ channel O:le and f32a subunits, phosphorylation of the 0:1 subunit by activated PKA, but not that of the f32a subunit, depended on co-expression of functional AKAP79244. The in vivo phosphorylation of the 0:1 subunit was reflected in increased currents in response to forskolin or cell-permeant cyclic AMP analogues244. The crucial site for PKA-dependent regulation of the cardiac L-type Ca2+ channel appears to be Serl928 in the 0:1 subunit: channels containing 0:1 subunits with an Sl928A substitution do not show increased phosphorylation or higher peak currents after PKA activation244. 42-32-06: N-type and P-type: Phos/PKA: The al subunit of N-type Ca2+ channels from rabbit brain is phosphorylated by PKA245, and cAMP enhances P-type currents obtained by expression of rat cerebellar mRNA in Xenopus oocytes246 .

Activation of PKC potentiates synaptic transmission 42-32-07: Phos/PKC: The activation of PKC by phorbol esters enhances neurotransmitter release from sympathetic neurones247 and potentiates synaptic transmission in the hippocampus248 and neuromuscular junctions249. These effects are associated with activation of voltagedependent Ca2+ channels and increased Ca2+ influx.

Activation of PKC inhibits some Ca 2+ currents but stimulates others 42-32-08: In some neuronal preparations, activation of PKC can lead to inhibition of N-type Ca2 + channel activity as well as occluding the inhibitory effect of neurotransmitters250-252. Observations with other neuronal preparations contrast with this and show stimulatory effects of PKC activation on Ca2+ currents in frog sympathetic253,254, rat255,256 and guinea-pig hippocampa1257 neurones. Activation of PKC by phorbol esters (0.2 JlM phorbol 12,13-dibutyrate) increased T-type Ca2+ currents by 30% in cultured neonatal rat ventricular cells l14.

Activation of PKC stimulates L-type channel activity in skeletal and smooth muscle cells 42-32-09: Activation of PKC with phorbol esters or diacylglycerol analogues stimulated DHP-sensitive Ca2+ currents in cultured embryonic chick muscle cells258. In vitro phosphorylation of purified rabbit T tubule membrane by PKC stimulated Ca2 + currents after reconstitution into liposomes259 and treatment of planar lipid bilayers containing T tubule membranes with purified PKC increased the average Popen of L-type Ca2+ channels by 50%260. Phorbol esters also stimulated L-type Ca2+ channel activity in rat aortic rings261, cultured aortic A7r5 cells262, rat portal vein cells263 and rabbit saphenous artery cells264.

Activation of PKC reduces G protein-mediated inhibition of Ca 2+ currents in a variety of neurones 42-32-10: External application of phorbol-12-myristate-13-acetate (PMA) (500 nM) enhances the basal IBa in cerebral cortical, CAl pyramidal, CA3

II

1'--_e_n_t_ry_42

_

pyramidal, superior cervical ganglion and dorsal root ganglion neurones of the rat and bullfrog sympathetic neurones265 . The enhancement is voltage dependent: in cerebral cortical pyramidal neurones, for example, maximal enhancement of IBa (1500/0) was obtained at a test potential of -25mV and the strength of the enhancement declined with increasing depolarization beyond this value. In the same set of neuronal preparations, the PMA application also greatly reduced the inhibitory effects of a neurotransmitter receptor agonists, including baclofen (a GABAB agonist), 2-chloroadenosine (an adenosine receptor agonist), oxotremorine methiodide (a muscarinic receptor agonist) and luteinizing hormone-releasing hormone, all at saturating concentrations. The disruptive effects of PMA were eliminated by including 200 JlM PKC(19-36), a pseudosubstrate t peptide inhibitor of PKC, in the patch pipette. Note that the PKC inhibitor did not block the inhibitory effects of the neurotransmitters themselves, only the ability of PMA to interfere with that inhibition, showing that PKC is not part of the pathway coupling transmitter receptors to Ca2+ channels, but has the ability to regulate these pathways265.

Activation of PKC in sympathetic neurones increases N-type Ca 2+ currents 42-32-11: The application of the phorbol ester PMA (500nM), to acutely dissociated adult rat superior cervical ganglion neurones increases the amplitude of voltage-gated Ca2+ currents. The stimulatory effect is voltage dependent, maximal stimulation being obtained at a test potential of ",0 mV, and is blocked by PKC(19-36) (200 JlM), a pseudosubstrate inhibitor of PKC activity. This activation of Ca2+ currents involves both N-type channels and a channel type that is insensitive to w-conotoxin GVIA. The phosphatase inhibitor okadaic acid (1 JlM) accelerates the activating effect of PKA266 .

The nature of the Ca 2 + channel a subunit is crucial in determining modulation by PKC 42-32-12: Following heterologous t co-expression of rat Ql and 13lB cDNAs in

Xenopus oocytes, whole-cell Ba2+ currents from channels containing alB and alE subunits were increased by treatment with phorbol-12-myristate-13acetate (PMA; 100nM), while those from cells expressing QlA and Qle subunits were unaffected267. Pre-incubation of the injected oocytes with the protein kinase inhibitor staurosporine (5-10 JlM), eliminated the inducing effect of PMA. The stimulation of PKC activity through the phospholipase C-dependent second messengert pathway, via activation of the metabotropic glutamate receptor, mGluR1a, also up-regulates the currents obtained from QlE + 13lb channels in Xenopus oocytes267. Staurosporine (5 JlM) blocks this glutamate-induced up-regulation267.

{3 subunits are required for PKC-dependent up-regulation 42-32-13: The IBa through Ca2+ channels produced in Xenopus oocytes by expression of QlB or QlE cRNA are not affected by the stimulation of PKC activity, unless a (3 subunit is present. Stimulation of currents by PKC activation occurred when sequences encoding (3lb (140%), (32a (126%), (33

III

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_2---J

(146%) or {34 (184%) were co-expressed with those encoding the alB subunit. These effects are unaltered by co-expression of sequences encoding the 02 subunit267.

The domain I-II linker is crucial for PKC-dependent up-regulation

42-32-14: The Ca2+ channel /3 subunit binds to the domain I-II linker of 01 subunits204 (see Domain functions, 42-29). This same region is also the

crucial determinant of PKC-dependent modulation of Ca2+ channels. A chimaerict 01 subunit, 0lA/B, consisting of the entire 0lA subunit except for substitution of the 0lB domain I-II linker, produces currents that are sensitive to PKC-dependent stimulation in the presence of /3lb subunits in Xenopus oocytes, in contrast to the currents through 0lA + {3lb channels267.

Phosphorylation by PKC affects voltage dependence and kinetics

42-32-15: The action of PKC in increasing whole-cell currents through Ca2+

channels containing 0lB and OlE subunits in Xenopus oocytes is accompanied by differential effects on the voltage dependence of the currents. The I-V relationship for currents elicited by alB//3lb channels are shifted about 4 mV to the left after treatment of the oocytes with the phorbol ester PMA (100nM), while those from alE//3lb channels show a 4 mV shift to the right. In both cases, the voltage dependence of inactivation was shifted about 10mV rightwards after PMA treatment267. The rates of activation and inactivation of the currents shown by oocytes producing the OIE//3lb subunits were also significantly slowed by incubation with PMA, but those of cells making OIB/{3lb channels were unaffected267.

PKC enhances L-type and N-type Ca 2+ currents in frog sympathetic neurones 42-32-16: Frog sympathetic neurones possess both w-conotoxin GVIA (w-

CTx)-sensitive N-type and dihydropyridine-sensitive L-type Ca2+ channels. Isolation of the individual types using 'v-CTx (3 JlM) and nimodipine (1 JlM) shows that 76% of the IBa in these preparations is carried by N-type channels254 . Treatment of the cells with phorbol dibutyrate (1 JlM) enhances whole-cell currents by an average of 35C}6, with roughly similar quantitative effects on both L-type and N-type channels. The stimulatory effects of the phorbol estert were eliminated by pretreatment with staurosporine (500nM), a protein kinase inhibitor, and with PKC(19-3l) (5 JlM), a specific pseudosubstratet inhibitor of PKC. When G protein activation was varied by the use of GTP analogues, the extent of enhancement of IBa by PKC activation was unaffected, indicating that stimulation of IBa was not the result of removing tonic G protein inhibition in this case (see Channel modulation, 42-44, for contrasting examples). Studies of single-channelt currents in isolated patches showed that both L-type and N-type channels opened more frequently after application of phorbol dibutyrate, with the increased Popen being due to a decrease in the closed-time interval between adjacent openings from 68 ms to 32 ms. 'The unitaryt currents at -10 mV of N-type channels (0.72pA) were not affected by stimulation of PKC activitr54.

II

l_e_n_t_ry_42

--'_

PKC enhances neurotransmitter release triggered by Q-type Ca 2+ channels

42-32-17: Synaptic transmissiont between hippocampal CA3 and CAl neurones relies on the activities of N-type and Q-type Ca2+ channels. In the presence of w-conotoxin GVIA (1 JlM) to block N-type channels, application of the phorbol ester phorbol-12, 13-dibutyrate (1 mM), an activator of PKC, produces a delayed but sustained potentiation (222 0/0) of the synaptic response. This potentiation was eliminated by w-conotoxin MVIIC (5 JlM), a blocker of Qtype Ca2+ channels61 . PKC activation also diminishes the inhibition of Q-type channel activity caused by the interaction of neuromodulators with their receptors61 (for details see Channel modulation, 42-44).

Caveat on Ca 2+ channels in Xenopus oocytes 42-32-18: Note that Xenopus oocytes possess a small (ca. -5 nA) endogenous Ca2+ current that is insensitive to dihydropyridines, w-conotoxin and Agelenopsis aperta venom, and is blocked by the divalent cations C02+, Cd2+ and Ni2+. These currents are increased by intracellular injection of cAMP and by application of phorbol ester, consistent with regulation by PKA and PKC268 . These endogenous currents are also enhanced about fourfold by the heterologous expression of cDNA encoding the rat ,BIb subunit267. cis-Fatty acids attenuate Ca 2+ currents via activation of PKC 42-32-19: cis-Fatty acids, which activate protein kinase C, attenuate Ca2+ currents in mouse neuroblastoma cells269 . Arachidonic acid (AA) depresses hippocampal Ca2+ current in a dose- and time-dependent manner, similar to the effects of phorbol esters. A similar depression of Ca2+ currents has been observed using a xanthine-based free-radical-generating system270. The specific PKC inhibitor PKCI19-36, the protein kinase inhibitor H-7, and the superoxide free-radical scavenger superoxide dismutase (SOD) each blocked ICa depression by 70-800/0. Complete block of the AA response occurred when SOD was used simultaneously with a PKC inhibitor, suggesting that both PKC and free radicals t playa role in AA-induced suppression of lca.

The L-type Ca 2+ current in rat pinealocytes is inhibited by cGMP 42-32-20: Phos/PKG: Administration of dibutyryl-cGMP (100 JlM), or elevation of cGMP by nitroprusside or noradrenaline, caused inhibition of the L-type Ca2+ channel current in rat pinealocytes, as measured in wholecell patch-clamp t determinations. This action of cGMP was independent of cAMP elevation, since cAMP antagonists had no effect on the inhibition. The protein kinase inhibitors (1-(S-isoquinolinesulphonyl}-2-methylpiperazine (H-7) and N(2-guanidinoethyl}-S-isoquinolinesulphonamide (HAI004), blocked the dibutyryl-cGMP effect on the L-type Ca2+ channel current, suggesting the involvement of cGMP-dependent protein kinase in the inhibition by cGMp271 .

PKA phosphorylates Ca 2+ channel Ql subunits in rat hippocampus 42-32-21: Antipeptide antibodies specific for the 01 subunits of the class B, C or E calcium channels from rat brain specifically recognize pairs of polypeptides of 220 and 240 kDa, 200 and 220 kDa, and 240 and 250 kDa,

II

_ entry42

I

'------------_.

respectively, in hippocampal slices in vitro. These calcium channels are localized predominantly on pre-synaptic and dendritic, somatic and dendritic, and somatic sites, respectively, in hippocampal neurones. Both size-forms of alB and alE and the full-length form of alC can be phosphorylated by PKA after solubilization and immunoprecipitationt. Stimulation of PKA in intact hippocampal slices also induced phosphorylation of 25-50% of the PKA sites on class B N-type channels, class C L-type channels and class E Ca2+ channels. Tetraethylammonium ion (TEA), which causes neuronal depolarization and promotes repetitive action potentials and neurotransmitter release by blocking voltage-gated K+ channels, also stimulated phosphorylation of class B, C and E al subunits, suggesting that phosphoryation of these three classes of channels by PKA is responsive to endogenous electrical activity in the hippocampus272.

Two

Q1B

species as substrates for pl'otein kinases

42-32-22: The rat brain N-type Ca2+ channel al subunit exists as two size variants, 220 and 250kDa143 . Immunoblotting with antibodies directed against C-terminal peptides of the longer species showed that the two proteins have different C-termini273 . Both the 220 and 250 kDa forms of the alB subunit acted as in vitro substrates for PKA, PKC and PKG. In contrast, calcium- and calmodulin-dependent protein kinase IT (CaM kinase II) phosphorylated only the long form of the alB subunit273 .

Two size variants of Qle subunits as substrates for protein kinases 42-32-23: The alC subunit of the neuroneal L-type Ca2+ channel from rat brain exists in two size forms, LC2, with an apparent molecular mass of 210-235 kDa, and a C-terminally truncated shorter form, LCl, with an

apparent molecular mass of 190-195 kI)a. The longer isoform, but not the shorter, can be phosphorylated in vitro by PKA274 . Both LCI and LC2 are substrates for PKC, CaM kinase II and PKG274 .

Phosphatases inhibit PKA-dependerzt stimulation of Ca 2+ channel activity 42-32-24: Dephosphorylation: Phosphatases I, 2A and 2B have all been shown to be capable of the in vitro dephosphorylation of L-type Ca2+ channels of skeletal muscle275. Phosphatase 1 (2 flM) completely inhibits the ,B-adrenoreceptormediated increase in cardiac Ca2+ currents276, and a specific peptide inhibitor of phosphatase 1 enhances Ca2+ currents in cerebellar granule neurones235 and dorsal root ganglia277. Note that the endogenous inhibitor of phosphatase 1 is itseH dependent on phosphorylation by PKA for its activity (reviewed in ref. 278) and that the calcium-activated phosphatase 2B ('calcineurin') is implicated in the Ca2+ -dependent inhibition of Ca2+ currents279,280

ELECTROPHYSIOLOGY Note on endogenous ion channels in Xenopus oocytes: Denudedt oocytes t from Xenopus laevis are frequently used as the host cells for transient t expression of cloned sequences encoding Ca 2 + channel subunits. The

II

1'--_e_n_t_ry_4_2

_

design and interpretation of such experiments is complicated by the presence of several voltage-dependent ion channels endogenous to the Xenopus oocyte. The most prominent such conductance is a Ca 2+-dependent Clcurrent. Interference from this is usually prevented by recording Ca 2+ channel activity with Ba 2+ as the charge carrier, in a barium-methanesulphonate solution: Ba 2+, in contrast with Ca 2+, does not activate the Clcurrent281 and methanesulphonate does not permeate the Cl- channe1282 . The use of Ba 2+ has the additional advantage that it tends to block endogenous K+ channels. Endogenous voltage-dependent Na+ channels can be blocked by including tetrodotoxin in the bath solution. Endogenous Ca2+ channels activating around -25 m V and giving peak currents at +10mV have also been described in non-injected oocytes282-284. Currents with these characteristics are activated when Xenopus oocytes are injected with human neuronal f32 subunit cRNA alone, or in combination with a2b cRNA 15. This behaviour is in contrast with that of oocytes injected with rabbit skeletal muscle a2a and f31 cRNAs, which apparently do not develop an IBa upon depolarization 5 . f",J

Activation Ca 2+ channels vary in their sensitivity to depolarization 42-33-01: The different voltage-activated Ca2+ channels show marked differences in their sensitivity to depolarization t . The Ca2+ channels that are activated by small depolarizations, termed 'low voltage activated (LVA) Ca2+ channels', also tend to show rapid, voltage-dependent inactivation. The 'high voltage activated (HVA) Ca2+ channels' often lack rapid inactivation, and can therefore be detected without interference from LVA currents by starting from depolarized holding potentials (e.g. -30 mV).

Sequences in repeat I of the L-type Ql subunit are critical in determining activation kinetics 42-33-02: Heterologous t expression of eDNA constructs encoding chimaeras t of the skeletal (CaChl) and cardiac (CaCh2a) muscle a1 subunits in dysgenic mouse myotubes shows that refeat I determines the activation kinetics of the channel following depolarizing steps from a holding potential of -80 mV. All functional chimaeric channels containing repeat I from skeletal muscle showed 'slow' activation (Tact 10-100ms), and all those with repeat I from the cardiac muscle subunit had 'rapid' activation (Tact 1-7ms)198. Analysis of tail currents t obtained after stepping from a strongly depolarizing potential gives the same conclusion: the region of repeat I determines the kinetics 198. More detailed mapping with chimaeric t channels shows that it is the structures of the 83 segment and the linker connecting 83 and 84 segments of repeat I that are critical for the difference in activation kinetics between cardiac and skeletal muscle L-type calcium channels285 .

Photolysis of caged InsP3 activates Ca 2+ channels in T cells 42-33-03: T-type: Flash photolysis of a caged InsP3 analogue which is not readily metabolized to InsP4 (l-(a-glycerophosphoryl) inositol 4,5bisphosphate) is sufficient to activate plasma membrane calcium current

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_2_

resembling T-type voltage-gated Ca2+ channels in human T cells286 . While elevation of internal Ca2+ inactivates the channel, internal perfusion with InsP4 does not affect it. This effect of caged InsPa may be indirect, through depletion of Ca2+ stores (see ILG Ca CSRC [native), entry 18), or direct, with a plasma membrane-associated InsPa receptor86 (see ILG Ca InsPa, entry 19).

L-type currents rapidly inactivate on repolarization 42-33-04: Repolarization of the membrane after brief activating pulses produces inward tailt currents that decay rapidly as the L-type Ca2+

channels close. The time constants for the L-type currents in various preparations, and for the more stable T-type currents, are shown in Table 7.

Depolarizing pre-pulses activate L·-type Ca 2+ channels 42-33-05: Currents through L-type Ca2+ channels can be reversibly enhanced

by manipulation of the membrane potential by depolarizing pre-pulses, a phenomenon known as 'facilitation'. This behaviour of L-type channels has been extensively studied in bovine chromaffin cells297-299, where wholecell currents tripled after facilitation by depolarizing pre-pulses. The increase in current was brought about by dramatically increased opening probabilities of an L-type channel, due to long-lived openings299 . The same channels can also be activated by by repetitive depolarizations in the physiological range, such as by increased splanchnic nerve activitr98. It has been suggested that the physiological role of this increase in Ca2+ current is the stimulation of rapid catecholamine secretion in response to danger or stress298 . Voltage-dependent facilitation of L-type Ca2+ currents has also been studied in skeletal, smooth and cardiac muscle (reviewed in ref. 300), and with 01 subunits produced by heterologous expression in mammalian cell lines or Xenopus oocytes.

Facilitation of L-type Ca 2+ currents is associated with long channel openings 42-33-06: The smooth muscle Ca2+ channel ale subunit, produced by

heterologoust expression in stably transformed Chinese hamster ovary cells, generated Ca2+ currents that were potentiated two- to three-fold by strongly depolarizing pre-pulses. Analysis of single-channel behaviour showed that pre-pulse facilitation involved the induction of 'mode 2-like gating', characterized by long openings and high Popen• In this system, there was no evidence for the involvement of channel phosphorylation in the facilitation process301 . The induction of mode-2 gating by depolarizing prepulses has also been observed in chromaffin302 and cardiac303 cells.

Facilitation of L-type currents involves voltage-dependent phosphorylation 42-33-07: Facilitationt of L-type currents by depolarization can be suppressed

by inhibitors of protein phosphorylation or by the intracellular activity of phosphatase 2A. The facilitation process is normally reversible, but reversibility is blocked by phosphatase inhibitors. The voltage- and

II

(D

=' ~ ~ ~

Table 7. Time constants for L-type and T-type currents (From 42-33-04) Current type

Cardiomyocytes

L-type

0.4-0.8 (bovine ventricular cardiomyocytes: -50 mV; 35°C)287 1.7-1.9 (rat and guinea-pig ventricular cardiomyocytes: -50mV; 22°C)288 0.2-0.4 (guinea pig atrial myocytes: -45mV; 21°C)289 1.5 (rat ventricular myocytes: -50mV; 22°C)290 5 (guinea-pig atrial myocytes: -45 mV; 21°C)289

T-type

II

t

(ms)

Smooth muscle

t

(ms)

Skeletal muscle

t

(ms)

0.4 (rat smooth muscle cell 9 (tfast) and 220 (tslow ) line, A7r5: -40mV; 22°C)291 (frog skeletal muscle 0.25 (guinea-pig basilar artery fibres: -90mV; 16°C)293 cells: -55 mV; 22°C)292 5-10 (rat skeletal muscle fibres: -60 to -90mV; 17°C)294

Other cell types

t

(ms)

0.2 (chick sensory neurones: -60mV; 20°C)295

1.1 (chick sensory neurones: -60 m V; 20°C)295 6 (mouse fibroblast 3T3 cell line: -60 m V; 22°C)296

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_2_1

frequency-dependent facilitation ('potentiation') of skeletal muscle L-type Ca2+ currents by high-frequency depolarizing pre-pulses requires phosphorylation by cAMP-dependent protein kinase (PKA) in a voltagedependent manner37• PKA also potentiates the voltage-dependent facilitation obtained when the cardiac al subunit is produced by heterologous t expression in Chinese hamster ovary cell lines304 and with the neuronal ale subunit in Xenopus oocytes305 . Reversal of the potentiated currents was blocked by the phosphatase inhibitor okadaic acid304. These findings suggest that facilitation of L-type Ca2 + currents by pre-pulses or repetitive depolarizations involves voltage-dependent phosphorylation of the L-type Ca2+ channel or a closely associated modulatory protein236 . Note that currents produced by co-expression of alA, alB or alE subunits with (3 subunits in oocytes do not show voltage-dependent facilitation305. Frequency-dependent facilitation of L-type Ca2 + currents in cardiac myocytes and skeletal muscle fibres and of L-type and T-type currents in smooth muscle myocytes have been reviewed98 .

Ca 2+ channels in dendrites activated by post-synaptic potentials

42-33-08: The dendrites t of CAl pyramidal neurones have been shown to contain T-type, R-type and L-type Ca2+ channels, in addition to voltageactivated Na+ channels149. The T-type channels and Na+ channels were opened by subthreshold excitatory post-synaptic potentials (EPSPs t) of 10mVat the site of recording, but opening of the high voltage activated (Rtype) channels required somatically generated action potentials or trains of suprathreshold synaptic stimulation. The EPSP-activated T-type channel activity was infrequent at holding potentials of 10-15 mV depolarized from resting potential, but a 4 s hyperpolarizing pre-pulse 400 ms before synaptic stimulation increased the fractional open time in a voltage-dependent manner. This suggests that the contribution of T-type channels to EPSP amplitude and Ca2+ influx would be maximal for EPSPs occurring after hyperpolarizing inhibitory post-synaptic potentials (IPSPs t) or spike-mediated after-hyperpolarizations t 149.

Current type L-type channels in transverse tubules as voltage sensors for E-C coupling 42-34-01: Transverse tubule Ca2+ currents are small and slow, underlying the role of L-type Ca2+ channel components as voltage sensors: E-C couplingt in skeletal muscle does not depend on extracellular Ca2+ influx, nor a diffusible second messenger (see Abstract/general description under ILG Ca Ca RyRCaf, 17-01).

Current-voltage relation Voltage-dependent opening involves the movement of three to five gating charges 42-35-01: The different types of voltage-dependent Ca2+ channels all open with depolarization in a highly voltage-dependent manner. Peak Popen rises

II

1'--_e_n_t_ry_42

----"_

to a maximum at large depolarizations and its dependence on voltage can be described by a Boltzmann relation t equivalent to three to five gating charges t moving across the membrane during the activation process 71 .

Voltage-dependent gating of skeletal muscle Ca 2+ channels 42-35-02: Native skeletal muscle SR channels in excised patches or incorporated into planar bilayers show a macroscopic I-V relation consistent with voltage-dependent gating. The macroscopic conductance is significantly increased at positive voltages, due to an increase in channel open time, with T remaining unchanged (linear between -lOOmV and +lOOmY, Popen rvO.l; dramatically increasing between +60mV and +120mV Popen --t rvl.0)306-309.

1- V relationships for L-type and T-type channels compared 42-35-03: The macroscopic I-V relation for both L-type and T-type Ca2+ channels are 'bell-shaped', with near zero amplitude at threshold t, maximal current at Vpeak and smaller amplitude at more positive potentials310. The I-V relation for T-type currents has its maximum about 30mV negative to that of L-type currents96, due to the higher Popen of T-type channels at the more negative potentials310.

The 1- V relations of L- and T-type currents are sensitive to shielding effects 42-35-04: The position of peak Ca2+ channel current on the voltage axis of the 1-V plot is dependent on the external cation species and its concentration,

due to shielding t effects of the cations. At concentrations in the range 2 to 25 mM, equimolar substitution of external Ca2+ by Ba2+ or Sr2+ shifts the 1-V relation of L-type currents towards more negative voltages by ca. 10mV in cardiac311 and smooth muscle 91,96 myocytes, and by as much as 20mV in frog skeletal muscle312 . Very much smaller shifts in the I-V relation are seen with T-type currents313 . Raising the external Ca2+ concentration shifts the I-V curve to the right by l-2mV/mM in cardiac, skeletal muscle312 and smooth muscle314 cells. The T-type 1-V relation shows a similar response 96 . The magnitude of the shifts caused by Ca2+, Ba2 +, Sr2 + and other divalent cations saturates at higher concentrations, with shifts of 30-40mV for both L_type 90,91,96,311,315 and T-type currents 96,313 when millimolar extracellular Ca2+ or Ba2+ is raised to 100mM. The I-V relationships for L-type and T-type Ca2+ channels are compared in Fig. 5.

Dose-response Current amplitudes plateau at high external ion concentrations 42-36-01: The amplitude of the whole-cell current through voltage-gated Ca2+ channels depends on the external Ca2+ concentration. With the latter in the range between 0.1 and 3 mM the current amplitude response is approximately linear in cardiac myocytes316 and frog skeletal muscle317, but the response decreases above this range and the current amplitude plateaus at Ca2+ concentrations of 30-50 mM. The concentration giving half-saturation was

II

_'--

e_n_try_4_2----1

5 Ca//145 Cs +100

Figure 5. I-V relationship for peak L-type and T-type currents in dog atrial myocytes bathed in 5 mM Ca 2+ solution. (Reproduced with permission from Bean (1985) J Gen Physiol 86: 1-30.) (From 42-35-04)

13.8 mM in frog atrial myocytes318, 20--30 mM in frog skeletal muscle317 and only 1.2mM in guinea-pig taenia caeci cells319• The currents obtained with monovalent cations as charge carriers also saturate, but at much higher ionic concentrations. Half-saturation of the Ca2+ channels derived from rat skeletal muscle incorporated into lipid bilayers was achieved at 200300mM Na+, when I Ba half-saturated at 40mM Ba2+32o.

Inactivation Inactivation of L-type currents is voltage dependent 42-37-01: The relation between membrane potential and steady-state inactivation of L-type Ca2+ channels can be assessed by applying short ( rv 300 ms) depolarizations t to a constant potential from varying holding potentials. The sigmoidal relation fits a Boltzmannt function: for cardiac tissue321 and myocytes322 bathed in millimolar concentrations of Ca2+, Vh is rv30mV and k is rv7mV (i.e. Vh is 10-15mV positive to threshold and rv30mV negative to Vpeak) (see {PDTM}, Fig. 4). In smooth muscle myocytes, the slope factor of the Boltzmann relation is again rv7mV but the Vh is commonly 35-50mV negative to Vpeak31S,323,324. The Vh in skeletal muscle cells can be even more negative than that in smooth muscle cells312. Subthreshold depolarizations can induce significant inactivation of L-type currents in cardiac42, smooth muscle 91 and skeletal muscle32s preparations, suggesting that channel opening is not necessary for inactivation.

II

1L....-__ _ _ry_4_2

en t

_

L-type currents decline during maintained depolarization 42-37-02: Macroscopic L-type currents activated by depolarizing steps decline during maintained depolarization of cells from cardiac tissue311, smooth muscle326-328 and skeletal muscle325 . The kinetics of the decline are complex and variable, with fits to one or two exponentials321 and a slow (seconds) component329 being described for cardiac tissue preparations. For guinea-pig ventricular myocytes, the faster exponential has a time constant of 3-7 ms and amplitude 0.6, and the slower component 30-80 ms and amplitude 0.4 at rv35°C 33o . Similarly complex time courses have been described for the decay of lea,L in smooth muscle cells328~331. For guinea-pig urinary bladder myocytes, the estimated time constants for lea,L inactivation near Vpeak at 35°C were 5, 40 and 200ms332. Estimates for the inactivation time constants for lea,L from rabbit portal vein myocytes at 22°C were 1525, 80-215 and> 1000 ms314.

L-type Ca 2+ currents in skeletal muscle decay slowly during maintained depolarization 42-37-03: The L-type currents in skeletal muscle decay very slowly, by a voltage-dependent inactivation, during maintained depolarizationt 312. Single channels from rabbit T tubular membrane, incorporated into lipid bilayers and activated by treatment with BAY K 8644, showed voltagedependent inactivation with T = 3.7s at OmV333 .

Inactivation of L-type channels in neuronal cells is Ca 2+ dependent 42-37-04: L-type Ca2+ channels in a wide variety of neuronal and muscle cell preparations undergo calcium-dependent inactivation, where Ca2+ ions entering the cell during depolarizing test pulses cause or accelerate inactivation. In neurones of the mollusc Aplysia califarnica injection of the Ca2+ chelator EGTA (2-10mM), slows the rate of inactivation of lea during a depolarizing pulse from a holding potential of -40 mV. The substitution of 100 mM Ba2+ for 100 mM Ca2 + in the extracellular solution also retards inactivation334 . Similar observations have been made with guinea-pig ventricular myocytes335, rat ventricular myocytes288 and guinea-pig atrial cardioballs336 . The extent of inactivation of lea in a pituitary tumour cell line, GH3, during a 60ms pre-pulse at various depolarizing voltages correlates well with the amount of Ca2+ entering the cell during the prepulse. The inactivating effect of the conditioning pulse is eliminated by replacement of the extracellular Ca2+ by .Ba2 +337 and reduced by concentrations of external C 0 2+ or Cd2+ that give partial block of lea,L 327. Intracellular Ca 2+ -release closes L-type Ca 2+ channels 42-37-05: Intracellular release of Ca2+ from the photolabilet Icaged,t Ca2+ chelator DM-nitrophen causes inactivation of the L-type Ca2 + current in isolated guinea-pig ventricular heart cells. This Ca2+ photorelease inactivates the Ca2+ current without affecting the inactivation-associated intramembrane charge movement, or 'gating current't, whereas voltage-induced inactivation reduces the Ca2 + current and the gating current proportionally. The Ca2 +dependent inactivation apparendy closes the L-type Ca2+ channel through a mechanism that does not involve its voltage-sensing region338 .

II

_'--

e_n_try_4_2-----J

Ca 2+ -dependent inactivation of L-type currents is a localized event 42-37-06: The kinetics of IBa,L and INa,L in embryonic chick ventricular cells are unaffected by simultaneous entry of Ca2+ into the cell through Ca2+ channels that are outside the membrane patch339. This observation supports the hypothesis that Ca2+ -dependent inactivation of L-type currents involves Ca2+ ions localized within or near those open channels that have transported the ions into the ce11340.

Ca 2+ -dependent inactivation requires a Ca 2+ -binding motif on the subunit

(}1

42-37-07: L-type Ca2+ channels produced by heterologous t expression in HEK-293 cells of cDNAs encoding alC and (32a subunits show Ca2+dependent inactivation. In contrast, channels produced by co-expression of cDNAs encoding (32a and alE, the pore-forming component of a mediumthreshold, neuronal Ca2+ channel, lack the Ca2+ -dependent process. Manipulation of the alC and alE cDNAs to produce chimaerict channels showed that Ca2+ -dependent inactivation depends on a Ca2+ -binding motif (an EF-hand t ) located at amino acids 1510-1560 on the alC subunit203 .

T-type channels do not undergo calcium-dependent inactivation 42-37-08: T-type Ca2 + channels do not undergo Ca2+ -dependent inactivation: the time course of their inactivation is insensitive to the replacement of Ca2+ by Ba2+ 96,313 or Sr+ 341, and to increases in external Ca2+ concentration322. The inactivation time constants at 22°C are 1".J50ms at 20mV negative to 96 Vpeak and 1".J20 ms at Vpeak and more positive potentials ,325,342.

T-type currents inactivate at more negative potentials than L-type currents 42-37-09: The steady-state inactivation of T-type currents in a variety of muscle cell types occurs at more negative potentials than that of L-type currents. In cardiomyocytes, for example, T-type current is fully inactivated at potentials positive to -45 mV when the bathing solution contains millimolar concentrations of divalent cations322. The inactivation-voltage relation fits with a Boltzmannnt distribution function with a slope of 46 mV in cardiomyocytes313,322,341, smooth muscle myocytes 96 and skeletal muscle myotubes325 . The Vh is 1".J-70mV at millimolar concentrations of external Ca2+ for the T-type channels in cardiomyocytes313,322 and 1".J-80mV in skeletal muscle cells325 and smooth muscle myocytes 96 .

Inactivation rates depend on (3-subunits 42-37-10: The kinetics of inactivation of Ca2+ channels formed by heterologousteo-expressiont of cRNAst depend on the identity of the (3 subunit. For channels containing the rabbit alA subunit produced in Xenopus oocytes, inactivation was :most rapid for the alA + a2/ 8 + (33 combination (7 = 23 S-l), less rapid for alA + a;/8 + (31 (7 = 16 S-l) and slowest for alA + a2/8 + (32b complexes (7 = 9 S-l )13. The same three (3 subunit isoforms impose the same relative order of inactivation rates on the co-expressed alC subunit52 . For any particular (3 subunit, the rate of decay of alA current was at least an order of magnitude faster than that of

II

1'--_e_n_t_ry_4_2

_

the corresponding ale current13. In other studies, the rat brain {32a subunit dramatically slowed the inactivation of the alA current obtained by transient expression in Xenopus oocytes 7 .

Molecular determinants of voltage-dependent inactivation 42-37-12: The difference in the rates of voltage-dependent inactivation of Ca2+ channels containing the alA and doe-1 (see paragraph 42-21-04) al subunits have been exploited to localize the determinants of the inactivation process. Following co-expression of cRNAs encoding the aI, a2 and {31 subunits in Xenopus oocytes, channels containing the doe-1 al subunit inactivated two- to threefold faster than those contining alA subunits. (These recordings were carried out with Ba2+ as the charge carrier, to minimize the effects of Ca2+ -dependent inactivation; see paragraphs 42-3705 to 42-37-07). Analysis of the properties of chimaeric t channels showed that the key determinants of the kinetic differences are localized to a region of 250 ms207. Carboxypeptidase At (1 mg/ml) had a similar action on the current, but leucineaminopeptidase t was without effect. Inactivation rates obtained with Ba2+ as charge carrier were not affected by proteolysis, suggesting that the slowing of Iea,L was due to loss of Ca2+ -dependent inactivation. In porcine coronary artery smooth muscle cells, trypsin (1 mg/ml) and carboxypeptidase A (1 mg/ml) produced a fourfold increase in Iea,L without affecting the kinetics344 . Lower concentrations of trypsin (0.05-0.1 mg/ml) dialysed into A7r5 smooth muscle cells increased Iea,L with no change of inactivation rate, but the enhanced IBa,L was associated with a marked slowing of inactivation345 •

Kinetic model Opening of L-type channels modelled to involve two voltagedependent steps 42-38-01: For the voltage-dependent activation of L-type Ca2+ channels, opening has been modelled as a two-step process involving the transition from one closed state to another before the open state is reached: both these steps are assumed to be voltage dependent297. Appropriate choices for the rate constants governing the reversible transitions between the closed

II

_ _ _ _ _ _ _ _ _ _ _ _ _.

en_t_ry_4_2----11

and open states describes the behaviour of single L-type channels during depolarization. The model has been embellished to describe the activation of each of the Ca2+ channel types in sensory neurones (see Kostyuk et al., 1988, under related sources and reviews, 42-56).

Rundown Rundown involves dephosphorylation

42-39-01: The HVA Ca2+ channels of vertebrates and molluscs display rapid rundown in the absence of an A kinase phosphorylation system346,347. The intracellular agents which slow this process, including ATP, Mg2+, the catalytic subunit of protein kinase At and cAMP, indicate that rundown, which has three distinct phases, is due in part to loss of phosphorylation348 . Channel inhibition could also be due to endogenous calcium-independent phosphatases349, calcium-dependent proteases350 and calcineurin, a calcium-dependent phosphatase that dephosphorylates the channel causing inactivation. The L-type Ca2+ channe:ls in patches excised from smooth muscle cells show rundown, but their activity can be retained for several hours by bathing the inside surfaces with buffer solutions containing mM EGTA351-353. f"'.J5

ATP or ADP plus Mg2+ inhibit rundown of L-type Ca 2+ channels 42-39-02: In isolated chromaffin cells, the rundown of HVA Ca2+ channels was abolished by ATP at concentrations above 0.4 mM. Inosine 5'trisphosphate (2mM) could not replace ATP, whereas ·GTP could, but at higher concentrations. The effect of .ATP in blocking rundown required MgCl2 and the liberation of a phosphate group, since the ATP analogue 5'adenylyl-imido-bisphosphate (AMP-PNP) could not substitute for ATP. ADP, in the presence of Mg2+ only, could replace ATP in the same concentration range, even when the pathways converting ADP into ATP were blocked; GDP was ineffective. The inhibition of rundown by ATP or ADP was abolished by increasing the internal Ca2+ concentration (from pCa?? to pCa6.0, where pCa=-loglo[Ca2+]). Mg-ADP (ImM) in the bathing solution did not prevent rundown of the Ca2+ channels in the inside-out patch recording configuration354.

Ca 2+ -dependent proteases can be a major cause of rundown 42-39-03: In molluscan (Helix) neurones, leupeptin, a general inhibitor of Ca2+ -dependent proteases, protects against rundown, especially in the presence of ATp355. The rundown of Ica,L in guinea-pig ventricular myocytes was accelerated by dialysis with calpain, a cysteine endopeptidase, and inhibited by dialysis of calpastatin, a specific inhibitor of calpain activity348.

Rundown is aggravated at depolarized holding potentials 42-39-04: The whole-cell Ica,L in frog ventricular myocytes declines in two phases during pulsing from a holding potential of -40 m V: a faster phase (t 40s), reversible by shifting the holding potential to -90 mY, and a slower (t 15 min), irreversible phase of rundown. The biphasic decay was f"'.J

f"'.J

f"'.J

II

i_

e_n_t _ry_4_2

-----'_

absent at a holding potential of -90mV, but strongly accelerated at holding potentials positive to -60mV356 . Enhancement of rundown at holding potentials positive to -80mV has also been seen with rat 357 and rabbit358 ventricular myocytes and with frog atrial359 and ventricular313 cells. Note that rundown results in functional uncoupling of voltage-dependent charge movement from ion-permeation, since the Ca2+ channel gating current remains unchanged as the ionic current runs down360-362.

Note: The LVA or T-type Ca2+ channels rundown much less rapidly under whole-cell recording conditions.

Selectivity Voltage-gated Ca 2+ channels are permeable to several divalent cations 42-40-01: The voltage-gated Ca2+ channels are generally highly permeable to Ca2+, Sr2+ and Ba2+, though the relative order of permeability depends on the channel type. For L-type and N-type channels, higher currents are obtained with Ba2+ than with Ca2+, whereas T-type channels give similar conductances for these two ions. In addition to giving elevated currents in L-type channels, Ba2+ ions also block K+ currents, making them preferable to Ca2+ ions for the electrophysiological study of L-type channels. Note that Mg2 + ions do not permeate the L-type Ca2 + channels of skeletal muscle fibres 363, cardiac tissue364 or smooth muscle myocytes365.

Ca 2+ channels show anomalous mole-fraction dependence 42-40-02: Like several other ion-channel types, including inwardly rectifying K+ channels, delayed rectifier K+ channels and KCa channels, voltage-gated Ca2+ channels show a behaviour called 'anomalous mole-fraction dependence'. The conductance of L-type Ca2+ channels, for example, is high with external Ca2 +, t increased with Ca2 +-free Ba2 + solutions, but decreased when Ba2 + ions are added to Ca2 +-containing solutions316,317. These observations have been interpreted to support a single-file, multi-ion pore modeI316,317,366,367, in which the Ca2+ channel pore is capable of holding two or more divalent cations at the same time. Ca2+ binds more tightly to the channel than Ba2+, and therefore dominates in the binding competion, but leaves the channel to pass through more slowly than Ba2+. A more extreme example of the same phenomenon can be seen when Ca2 + is added to divalent-free solutions containing monovalent cations. At very low Ca2+ concentrations «7 nM), L-type Ca2+ channels from frog skeletal muscle are highly permeable to Na+. As Ca2+ is added, the Na+ current is titrated away, Ca2+ behaving as a channel blocker with a K D of 0.7 JlM. Current carried by Ca2 + ions returns as the [Ca2+] rises into the millimolar range (see Hille, 1992, under Related sources and reviews, 42-56).

L-type channels are highly permeable to several monovalent cations 42-40-03: In the absence of divalent cations, external monovalent cations can carry inward currents through Ca2 + channels. External Na+ can support large currents in cardiac myocytes316, skeletal muscle fibres 317, urinary bladder

II

_ L-----

entry 42

-----J

smooth muscle368 and from single channels incorporated into lipid bilayers369. (Compared with the single-channel conductance of rv20pS obtained with 110mM Ba2 +, that measured in the presence of 110-150mM Na+ was 75-110pS in guinea-pig cardiac myocytes370.) External Li+, Cs+ and K+ can also carry inward currents in cardiac tissue371,372. Organic cations such as the tetramethylammonium ion can permeate through Ltype Ca2 + channels in skeletal muscle and cardiac ventricular myocytes373,374. Permeability to tetramethylammonium ion determines that the minimum pore diameter is 6A374, a value that was confirmed by studies of permeability to monovalent cations in Ca2 + channels from rat skeletal muscle transverse tubules incorporated into planar lipid bilayers375 . In these in vitro preparations, NHt ions produced a higher single-channel conductance than Na+, Li+, K+ or CS+ 375. The reconstituted T-tubule channels had a fourfold lower conductance for hydrazinium ions than for ammonium ions, although these ions are amost identical in size375.

T-type channels are permeable to monovalent cations 42-40-04: The T-type Ca2 + channels of chick dorsal root ganglion (DRG) cells are permeable to monovalent cations at low external Ca2+ concentrations «100 JlM). The permeability ratios for monovalent ions with reference to internal Na+(1.0) were Li+(1.0), K+(0.45), Rb+(0.45) and Cs+(0.33)376.

Reversal potential for Ca 2+ channels is strongly influenced by K+ permeability 42-40-05: Voltage-gated Ca2+ channels are highly selective for Ca2 +, i.e. they generally show rv1000-fold preference for Ca2 + over K+ and Na+ {e.g. see ref. 377 for cardiac Ca 2+ channels}. Note that the internal K+ concentration is usually 106 times higher than that of Ca2+ ions so that, despite their lower permeability, K+ ions still carry the majority of the outward current through Ca2 + channels under strong depolarization. This means that the reversal potentialt (Erev ) for Ca2 + channels is not at the thermodynamic Eca (+124mV), but at a much less positive value. The measured value of Erev in the presence of known extracellular and intracellular concentrations of two competing ions allows calculation of the relative permeabilities of the two ions. Using this approach, PCa/PNa and PCa/PK are in the range 1000-11000 in frog atrial myocytes318 and guinea-pig ventricular myocytes370,377. The value of PCa/PCs in guinea-pig ventricular myocytes has been estimated as 4200-10000370,378, and PBa/Pcs as 1356_1700311,370,379. Similar values of PCa/PCs and PBa/Pcs have been derived for skeletal muscle and smooth muscle Ca2 + channels. The general order of ion selectivity for L-type channels is Ca > Sr > Ba > Li > Na > K > Cs (reviewed in ref. 380). Note that this order is directly opposite that determined on the basis of currentcarrying ability, and reflects the relative mobilities of the ions in dilute solutions375 . Mutant K+ channels mimic Ca 2+ channels 42-40-06: Site-directed mutagenesist of cDNAs encoding the Shaker K+ channel189 has produced deletion mutant channels having biophysical properties reminiscent of voltage-gated Ca2+ channels. In the absence of

II

II...-_e_n_t _ry_4_2

---'_

divalent cations, the mutant K+ channels conduct monovalent cations nonselectively, and divalent cations inhibit this process. Divalent cations at low concentrations block conduction through Ca2+ channels: only at millimolar concentrations of Ca2+ does Ca2 + conduction occur (see also Domain conservation, 42-28).

Changes of glutamate to lysine affect cation selectivity of L-type channels 42-40-07: Replacement of conserved glutamic acid (E) residues in the SSI-SS2 (Ipore-lining') regions of each of the first three repeats of the L-type (}:1 subunits by positively charged lysine (K) residues results in channels that carry Li+ currents, but negligible Ba2+ currents. The mutant channels effectively become selective for monovalent cations. Current-voltage relationships are also affected by these mutations, and the permeability ratio PLi/PK increased about three-fold for E736K (repeat II) relative to wild-type. The potency of Ca2 +-block of Li+ currents is reduced from 100- (E1446K: repeat IV) to 1000-fold (El145K: repeat ill). The conserved glutamic acid residues in all four repeats each make significant but different contributions to ion binding, selectivity and permeation381 .

Model for Ca 2+ selectivity and permeation involves simultaneous binding of two Ca 2+ ions 42-40-08: Data obtained with (}:1 subunits carrying mutational replacements of the four conserved glutamic acid residues in the SSI-SS2 repeats suggest that all four glutamate carboxylatest contribute collectively but asymmetrically to high-affinity binding of a single divalent cation. This binding blocks the passage of monovalent cations in the narrow pore and generates strong selectivity for Ca2+ over Na+ or K+. It is suggested381 that the four glutamates can also simultaneously interact with two Ca2+ ions, but with a greatly reduced affinity for both. This doubly occupied state is attained when a Ca2 + ion enters from outside, to compete with a previously bound Ca2 + ion for the carboxylate groups. Sharing of the ligands and possible electrostatic repulsion between the two adjacent Ca2 + ions would reduce their binding affinity and promote net influx of Ca2+ ions into the cell.

Single-channel data Popen for L-type channels is voltage dependent 42-41-01: The single L-type channels of cardiac and smooth muscle cells are activated by depolarizationt, with Popen increasing to reach a maximum at about +40 mV. In cardiomyocytes, Popen has a sigmoidal dependence on voltage, with slopes of about 7mV382 . Slopes of 4-7mV have been reported for L-type channels from arterial383-385 and airway386,387 smooth muscle cells. Note that the current elicited on depolarization to a maximally activating potential does not arise from all the functional Ca2+ channels in the cell membrane: only a fraction of the channels which open will be open at any given time after depolarization, and functional channels rapidly enter and leave the activatable state380 .

II

_

......

e_n_try_4_2

-----J

L-type channels show two kinds of gating 42-41-02: L-type Ca2+ channels show sudden changes of gating kinetics at the single-channel level. Channels commonly show frequent short «1 ms) openings during a 200 ms depolarization. In a series of consecutive sweeps, clusters of prolonged openings, mimicking the single-channel behaviour in the presence of a Ca2+ channel agonist, are occasionally found. Kinetic description of these long openings requires a rate constant for the channel closing that is 100 times lower than that in the more typical short opening behaviour388,389. The two types of gating behaviour have been termed 'mode I' (short opening) and 'mode 2' (long opening), and the phenomenon has been called 'mode-switching'389. It has been suggested that Ca2+ channel agonists bind more strongly to channels in mode 2 and thereby favour that gating mode389. The stimulation of L-type currents by dihydropyridine agonists involves an increase in Popen. Mean open times increased from 1-2 to 5-20ms in cardiac myocytes, smooth muscle myocytes, skeletal muscle cells and non-muscle cells (reviewed in reference 98).

Popen of T-type channels increases "With depolarization 42-41-03: Single-channel T-type current has a threshold 20-30mY more negative than that of L-type currents in cardiomyocytes 90,92 and smooth muscle cells310. The Popen increases with depolarization, with the slopes of the Boltzmannt relationship in the range 4-11 my 92,310. The latency t of Ttype currents is voltage sensitive92,310: in guinea-pig ventricular cells, for example, mean first latency decreased from 11 ms at -35 mY to 3.5 ms at _10my39o.

Shielding by extracellular cations is evident in single-channel 1-V curves

42-41-04: The shielding t effects of extracellular cations (see Current-voltage relation, 42-35) are observable at the single-channel level. For example, the raising of the external Ba2+ concentration from 10 to 110mM causes a 20 mY positive shift in the single-channel I-V curves for both L-type and Ttype Ca2+ channels in guinea-pig coronary artery myocytes310.

L-type and T-type currents show multiple single-channel conductance levels 42-41-05: L-type channels in cell-attached patches can switch between different conductance levels. With cell-attached patches of guinea-pig ventricular myocytes, in addition to the common rv25pS (90-110mM Ba2+) conductance, transitions to and from a rv 17 pS subconductance are seen391,392. Five distinct sublevels, 8-9, 12-13, 16-18, 23-24 and 28 pS, have been detected in cell-attached patches of GH3 pituitary cells393 . Ttype currents in cell-attached patches of guinea-pig ventricular myocytes show a subconductance at about one-half the conductance of the predominant 7pS level92.

L-type channels show distinct modes of gating 42-41-06: Single-channel recordings of L-type currents from cardiomyocytes show two types of channel opening behaviour: 'short-opening', in which

II

l_e_n_t_ry_42

_

channel openings last about 1 ms, and 'long-opening', in which opening times are 5-20 times longer388 . Long-openings represent about 100/0 of the total Ltype openings in cardiac cells303,339,388. They are also apparent in smooth muscle cells from the pig urinary bladder394 and airway395 and in bovine chromaffin cells298,396. Channels produced by expression of cRNAs encoding rabbit Qle, Q2/8 and {31 subunits in Xenopus oocytes had mean opening times of rvO.5ms and a unitary conductance of 21.4pS13.

Unitary currents through L-type channels saturate at extreme membrane potentials 42-41-07: Unitary outward currents t through L-type Ca2+ channels in membrane patches from undifferentiated PC12 cells increase linearly with membrane potential in the range +20 to +220mV. The response at higher test potentials begins to diminish and the unitary current reaches a plateau at about 240mV. At this saturating level the current is diffusion-limited, as shown by the response to changes of viscosity t (produced by adding glycerol to the test solution) and changes in the concentration of the charge-carrying cation397. Unitary inward currents carried by 170 mM external K+, an effective permeant ion in the absence of divalent cations, are linearly related to voltage in the range -10 to -340mV. Decreasing the diffusion coefficientt, by using 20-40% glycerol solutions, produces saturation of the unitary inward current, with a linear relationship between relative diffusion coefficient and the current amplitude at saturation. The different saturation values of the outward and inward currents have been used as the basis of calculations demonstrating that the functionally defined external pore entrance is almost twice as wide as the internal mouth (radii of 5.29 and 2.70 A respectively)397. Under physiological conditions, where the external Ca2+ concentration is about 1 mM, inward flow of Ca2+ is not likely to be diffusion-limited397. alA

channels produced in Xenopus oocytes show two open times

42-41-08: The Ca2+ channels produced by heterologous t co-expressiont of cRNAs t encoding rabbit alA, a2/6 and {31 subunits in Xenopus oocytes open briefly in response to depolarization from a holding potential of -80mV. The distribution of opening times was fbi-exponential', with T values of rvO.S ms and rv2.6 ms 13 . The mean single-channel conductance t was 15.9pS, and the unitary current at OmV was 0.8pA13.

N-type channels show three gating modes 42-41-09: Single-channel recordings of N-type Ca2+ channels in frog sympathetic neurones show three distinct patterns of gating, termed high-, medium- and low-Po, at test potentials of -10mV. In the high-Po mode, openings are relatively long-lasting (mean open duration rv3 ms) and closings are short (mean closed time rv2 ms): Po is typically rvO.5. The medium-Po mode is characterized by briefer openings (mean open duration rv1.6ms) and longer closings (mean closed time rv10ms), giving a Po of rvO.1S. In the low-Po mode, openings are even shorter (rv O.5ms) and closings more prolonged (rv 50 ms). Within a single sweep, direct transitions between any two gating modes are found. Noradrenaline (100)lM) in the

II

_'--

.

e_n_try_4_2_

patch pipette inhibits the current by about 50% and markedly reduces the prevalence of high-Po gating39B .

Voltage sensitivity The rate of activation of L-type currents is voltage-dependent 42-42-01: The rate at which cardiac L-type Ca2+ currents reach peak activation following a depolarizing voltage-switch is increased at more positive potentials. For guinea-pig ventricular myocytes, the time to peak is but only about ams at +20m~ measured at 35°C in about 9ms at -20m~ 3.6 mM Ca2+ bath solutions 9B . The L-type currents of smooth muscle myocytes behave similarly314. The same voltage-sensitive phenomenon is seen with the L-type currents of skeletal muscle cells, but the times to reach full activation are about 10-fold elongated399.

The f3 subunit increases peak currents and shifts voltage sensitivity 42-42-02: Co-expressiont in Xenopus oocytes of cRNAs encoding /3 subunits with those specifying the QIC subunit leads to a large increase in peak currents, the magnitude of which is voltage dependent. This effect arises from a change in the voltage sensitivity of the QI subunits, with /32 causing a 9 mV displacement of the 1-V relationship towards more negative potentials46 . Co-expression of /32 coding sequences also led to a four-fold acceleration of channel activation kinetics and a 3.5-fold increase in specific binding of the DHP antagonist [3H] PN200-110. The latter observation argues for a role of the {3 subunit in increasing the number of channels at the cell surface46 .

The f3 subunit facilitates pore-opening in cardiac L-type channels 42-42-03: Gating currentst associated with the movement of charged elements in the ion channel protein have been measured for Ca2+ channels produced by expression of cDNAs encoding the rabbit cardiac QI subunit QIC-a in Xenopus oocytes. The co-expression of cDNA encoding the rabbit type 2a cardiac f3 subunit (f32a) had no effect on the magnitude or time course of the gating current, but did increase I Ba severalfold and shifted the G-V curves by 16mV towards more negative potentials. These observations imply that the f3 subunit does not modulate the movement of the voltage sensor, but interacts with the QI subunits to facilitate the opening of the poret after the voltage sensor charge has moved102.

The activation of T-type currents occurs at more negative potentials 42-42-04: The activation of T-type currents shows many similarities to that of L-type currents (see above), including the slope of the activation curve, time to peak current and increase of rate with progressive depolarization t. The activation thresholdt and Vh are generally 20-40mV more negative for Ttype than for L-type currents (reviewed in ref.9B). The rates of activation and inactivation of T-type channels are both voltage dependent over the range from -60 to -IOmV71 .

I

entry42

1....---

_ _

PHARMACOLOGY

Blockers Voltage-gated Ca 2+ channels show differential sensitivity to multivalent cations 42-43-01: The sensitivities of the various types of voltage-gated Ca2+ channels to divalent cations are shown in Table 8.

Lanthanum ion is a powerful blocker of L-type channels 42-43-02: The lanthanum ion is a powerful blocker of ICa and lBa in L-type Ca2+ channels. In frog atrial myocytes, lca in the presence of 2.5 mM external Ca2 + was completely blocked by 10 JlM La2+ 400,406. Assuming a 1: 1 binding of La2+ to a channel site, the K o was estimated as < 1 JlM400 . Binding of La2+ to the channel was inhibited by increasing concentrations of Ca2 + 400 . The major effect of La2+ at the single-channel level is to reduce Popen by prolonging interburst closures371 . La2+ ions are also powerful blockers of L-type Ca2+ channels in smooth muscle407,408.

Magnesium ions are weak blockers of L-type channels 42-43-03: M g2+ is a weak blocker of ICa through L-type Ca2+ channels. In frog atrial cells, for example, 5 mM Mg2+ failed to reduce lca in the presence of 2.5 mM external Ca2 + 406 . The EC so of Mg2 + for lca (2.5 mM Ca2 +) in guineapig taenia caeci myocytes is 10mM319, and that in skeletal muscle (10mM external Ca2 +) is ",33 mM317. The effects of Mg2 + on lBa are more striking: in frog atrial myocytes, lBa at 2.5 mM Ba2+ was almost completely inhibited by 5 roM Mg2 + 406 . The blocking effect of Mg2+ is increased at more negative membrane potentials371 . Given the weakness of the Mg2 +_ dependent block in the presence of Ca2+, it is unlikely that external Mg2+ gives any significant inhibition of Ca2+ currents under physiological conditions in vivo. Note that elevated intracellular Mg2+ concentrations obtainable during metabolic inhibition can significantly inhibit ICa,L in frog 409 and guinea_pig410 cardiomyocytes.

Three main structural classes of Ca 2+ channel blockers 42-43-04: Three structurally unrelated classes of drugs, the dihydropyridines, phenylalkylamines and the benzothiazepines, have been valuable in defining different types of Ca2+ channel and evaluating their physiological roles. The structures of representative members of the three classes are shown in Fig. 6.

Blockage of cardiac L-type channels by dihydropyridines is voltage sensitive 42-43-05: L-type Ca2+ channels are characterized and defined by their sensitivity to dihydropyridines, such as nifedipine, nitrendipine and nimodipine. These drugs block L-type channels more potently if the current is elicited from depolarizedl potentials, when the channels are partially inactivated, than from hyperpolarizing t holding potentials, when the channels are in the non-inactivated 'resting' state411 . Nitrendipine at ",0.4nM causes 500/0 block of cardiac L-type channels for currents elicited

II

iii Table 8. Sensitivities of various types of voltage-gated Ca 2+ channels to divalent cations (From 42-43-01) Channel type

ICso (J,1M) for block by

Tissue Cd2+

L-type

N-type P-type Q-type T-type

Atrial myocytes (frog) Dorsal root ganglia (chick) Skeletal muscle (frog) Ventricular myocytes (guinea-pig) Dorsal root ganglia (chick) Heterologous expression alA, a218, {3l, 'Y (rabbit) in . J. (enopus oocytes Cerebellar granule cells (rat) Pituitary cell line (GH3) Atria (canine) Sinoatrial node (rabbit) Dorsal root ganglia (chick)

C0 2+

Ni2 +

Reference La2+ 1

10 300 20 10 1

401

1280 2S0

402

680

371 401

13

1 188

66 777

14 2.4

403 404

2000 40 20

400

74 401

The N-type currents in the rat neuroblastoma x glioma hybrid cell line NGI08-1S are particularly sensitive to block by gadolinium (Gd3+) ions (Kn 0.7J,1M)405.

g t"+

~ ~ ~

---l_

II-._e_n_t_ry_42 DIHYDR()PYRIDINES

PHENYLALKYLAMINES

Nitrendipine

Verapamil

0600 Nifedipine BENZOTHIAZEPINES

~FJ O,N6~OOCHJ CH)

N H

CH)

Bay K 8644 Diltiazenl Figure 6. Structures of members of the three major classes of Ca 2+ channel antagonists dihydropyridines, phenylalkamines and benzothiazepines. (From 42-43-04)

from a depolarized holding potential, but rv 700 nM is required for the same inhibition of currents elicited from a negative holding potential411 . Nimodipine is more potent than nitrendipine or nifedipine in blocking at negative holding potentials: 1 mM nimodipine blocks cardiac L-type current completely even at negative holding potentials412 . At this concentration, nimodipine has very little effect on P-type Ca2+ channels in rat cerebellar Purkinje neurones or N-type Ca2+ channels in rat sympathetic neurones 68 • Note that nimodipine (5 ~M) augments the currents through Ca2+ channels containing QlA subunits produced by heterologous co-expression of cRNAs

_'--

e_n_try_4_2_1

in Xenopus oocytes13. Dihydropyridines, including nifedipine, are used clinically for the treatment of angina, hypertension, cerebral and cardiac ischaemia and Raynaud's phenomenon. Nimodipine has been approved for the treatment of ischaemic neurological deficits following subarachnoid haemorrhage.

High concentrations of dihydropyridines have non-specific blocking effects 42.-43-06: Although micromolar concentrations of dihydropyridines are specific blockers of L-type channels, concentrations above 10 JlM can result

in non-specific blocking effects on non-L-type Ca2+ channels and on Na+ and K+ channels. Nitrendipine at 10 JlM inhibits cardiac T-type current by rv50% 404, neuronal N-type current by 10_25%413 and cardiac Na+ current by 10_50%414. This concentration of nitrendipine is without effect on Ptype Ca2+ channels in cerebellar Purkinje neurones413 . (Beware that apparent blocking effects can be influenced by the solvent in which the dihydropyridine is dissolved. Bay K 8644 dissolved in dimethyl sulphoxide stimulated Ica,L and inhibited Ica,T in neuroblastoma cells, but the inhibitory effect disappeared when the solvent was ethanol or polyethylene glycol415.)

Phenylalkylamines and benzothiazepines are blockers of L-type Ca 2+ channels 42.-43-07: The clinically useful Ca2+ channel blockers, the phenylalkylamines,

such as verapamil, and the benzothiazepines, such as diltiazem, are effective blockers of L-type Ca2+ channels. Both classes of compound have highaffinity binding sites on L-type Ca2+ channels that are distinct from the high-affinity dihydropyridine-binding sites. Verapamil and diltiazem show use-dependent block, so that it is difficult to obtain steady-state block on a reasonable time scale. The phenylalkylamine-binding site in the L-type Ca2+ channel 01 subunit has been mapped by photoaffinity labellingt and immunoprecipitationt to helix S6 of domain IV and the beginning of the intracellular C-terminal region416. These findings are consistent with the observations that D800, a quaternary phenylalkylamine, blocked cardiac Ltype Ca2 + channels from the intracellular surface of the membrane417, and that inhibition of Ca2+ currents by phenylalkylamines is greatly accelerated by depolarizations that open the channels418 . A model placing the phenylalkylamine receptor site within the intracellular opening of the channel pore, accessible for high-affinity phenylalkylamine binding only when the pore is open, has been suggestedl99. Binding of phenylalkylamines at a receptor site in such a position would directly block ion movement through the pore. Note: Diltiazem at concentrations between ~ 12 and 200 JlM has been shown to give rise to non-specific increases in ionic permeability in biological membranes419.

The binding of phenylalkylamines to the L-type channel is inhibited by divalent cations 42.-43-08: Cd2 + and other divalent cations at concentrations > 100 JlM reversibly inhibit the binding of radiolabelled phenylalkylamines to L-type

II

l~e_n_t_ry_4_2

_

Ca2+ channels in skeletal muscle membranes42o. Note that this behaviour is in complete contrast to the high-affinity binding of dihydropyridines, which depends on the presence of divalent cations42o .

Different classes of blocker can bind co-operatively to L-type channels 42-43-09: Binding studies using homogenates of various tissues containing Ltype Ca2+ channels have shown that dihydropyridines, phenylalkylamines and benzothiazepines occupy distinct, allostericallyt coupled sites on the Ltype channel protein (reviewed in ref.421). These allosteric t interactions between different classes of blockers are voltage dependent422,423.

Diphenylalkylamines are non-specific blockers of L-type and T-type channels 42-43-10: A number of diphenylalkylamines, including flunarizine, cinnarizine, bepridil and fendiline, act as relatively non-specific Ca2+ channel blockers. (They are also calmodulin antagonists and can block other cation channels.) Fendiline block of ICa,L in guinea-pig ventricular myocytes involves acceleration of inactivation and a negative shift in steady-state inactivation424 . In the presence of fendiline, Bay K 8644 caused a further reduction of ICa,L, explained as involving an allosterict interaction between fendiline and Bay K 8644424. In guinea-pig atrial and ventricular myocytes, cinnarizine, flunarizine 425 and cinnarizine426 were more effective at blocking T-type (KD rv 1 JlM for inactivated channels) than L-type channels.

Some sodium channel blockers can also block L-type Ca 2+ channels 42-43-11: A number of lipophilic compounds that block Na+ channels, including quinidine, flecainide, propafenone, phenytoin, amiodarone, amiloride and prajmalium, can also block L-type and T-type Ca2 + channels when used at relatively high concentrations. The actions of these compounds on L-type and T-type Ca2+ channels have been reviewed98 •

Miscellaneous synthetic blockers of L-type Ca 2+ channels 42-43-12: Because of the potential value of pharmacological modulation of Ltype Ca2+ channel activity, a wide range of synthetic organic ligands have been investigated. Relevant observations on some of these L-type channel ligands are presented in Table 9.

L-type and T-type channels are blocked by extracellular protons 42-43-13: The current through L-type channels in cardiac ventricular and atrial tissue is blocked by reduction of the external pH by 1.5-3 units 433,434. Blockage of lBa and lsr was more severe than that of lCa 434. Protonationt of a site at the external surface of the L-type channel reduced single-channel currents threefold with Na+ as the charge carrier, without affecting currents obtained in the presence of 110 mM external Ba2+. The apparent pK of the external protonation site was affected by the nature of the charge-carrying ion: the pK of 8.6 obtained in the presence of Cs2+ was reduced to 8.2 and 7.4 with K+ and Na+, respectively, as charge carriers. It

II

_ entry42

I

"'-------------

Table 9. Miscellaneous antagonists of L-type Ca 2+ channel activity (From 42-43-11)

Compound (type)

Observations and references

McN5691 (ethynylbenzenealkanamine)

2 JlM McN5691 abolishes L-type Ca2+ currents in neonatal rat ventricular myocytes without affecting those in rat anterior pituitary (GH3) cells. The blockade is voltage-dependent, and binding of [3H]-DHP is competitively inhibited. The ICso for inhibition of Ca2+ influx into heart cells by McN569l is 7.6nM, 100-fold lower than that obtained. with GH3 cells427.

HOE166 (benzolactam)

Binds to skeletal muscle channels with a K D of 0.27 nM: gives voltage-dependent block of L-type current in smooth muscle cells (A7r5) at 10-lOOnM428 .

MDL1233A (lactamimide)

Inhibits L-type current in rat anterior pituitary (GH3) cells at micromolar concentrations, without affecting T-type currents429.

BMY20064 (DHP derivative)

A hybrid molecule that is as potent as nifedipine as an L-type Ca2+ channel antagonist in smooth muscle, but also acts as a potent Ql adrenoreceptor antagonist43o.

Niguldipine (DHPdiphenylpiperidyl hybrid)

Has a Ki of 0.1 nM for Ca2+ channels of skeletal muscle, heart and brain, with the (+)-enantiomer being 40 times more active than the (- )-enantiomer. Niguldipine also binds non-stereoselectively (Ki == 0.05 nM) to brain QlA adrenoreceptors431 .

Information taken from a review432 on Dliscellaneous ligands for L-type Ca2+ channels.

has been concluded that the interaction between protons and permeant cations is allosterict 372,435. T-type channels in guinea-pig ventricular patches has also been shown to be sensitive to external H+ ions: whole-cell T-type conductance was maximal at plfo 9.0, reduced to 35% maximal at pHo 7.4 and 'near zero' at pHo 6.0436 .

Taicatoxin blocks L-type channels 42-43-14: The peptide toxin taicatoxin, from the Australian taipan snake, is a voltage-dependent blocker of L-type Ca2+ channels effective at nanomolar

II

l_e_n_try_4_2

_

concentrations437. Binding of the toxin reduces the Popen of L-type channels from cardiac cells and influences drug-binding at the dihydropyridine site through an allosterict interaction437.

Blockage of L-type channels by an alkaloid 42-43-15: Tetrandine, an alkaloidt isolated from the Chinese medicinal herb, Stephania tetranda, is vasodilatory in vivo and reversibly inhibits L-type Ca2+ currents in cardiac membranes and GH3 cells, competing directly for the diltiazem-binding site438 .

The peptide toxin w-conotoxin is a selective blocker of N-type Ca 2+ channels 42-43-16: A peptide isolated from the venom of the sea snail, Conus geographicus, called w-conotoxin GVIA (w-CTx), is a highly selective blocker of N-type Ca2+ channels. The high voltage activated Ca2+ currents in chick sensory neurones have single-channel conductances of 13 pS and 25pS in 110mM BaCl2, characteristic of N-type and L-type currents respectively. The large-conductance channels are selectively eliminated by dihydropyridines, while w-CTx (5 JlM) irreversibly blocks only the smallconductance channels. In whole-cell recordings, the macroscopicthigh voltage activated Ca2+ current was completely blocked by w-CTx and insensitive to DHPs in 600/0 of the cells. The remaining cells expressed both a DHP-sensitive Ca2+ current and a DHP-insensitive, w-CTx-sensitive, N-type current439. In rat dorsal root ganglion neurones, saturating concentrations of w-CTx (>1 JlM) block about 50% of the high voltage activated Ca2+ current. The dose of toxin giving half-maximal block is ~30nM413. Note that w-CTx (2 JlM) does not block L-type channels in these preparations413 . The Ca2+ entry associated with neurotransmitter release from rat Purkinje cell neurones is 500/0 blocked by w-CTx, with an IC so of ~100nM84.

Heterologously expressed N-type channels are sensitive to w-conotoxin

42-43-17: When the cDNA encoding the human neuronal Ca2+ channel alB-I subunit was transiently co-expressed with sequences encoding human {32 and a2b subunits in the human embryonic kidney cell line HEK-293, an N-type Ca2+ current, sensitive to w-CTx but insensitive to dihydropyridines, was obtained17. The transfected cells showed a single class of saturable binding sites for w-CTx, with a KD of 55 pM 17.

Structural determinants of blockade of N-type channels by w-conotoxin GVIA 42-43-18: Following heterologous expression of cRNAs in Xenopus oocytes (N-type) calcium channels containing alB subunits, together with a2 and {31 subunits, are sensitive to w-conotoxin GVIA (w-CTxj 5 JlM), while those comprising alA, a2 and {31 subunits are unaffected by this peptide. Channels containing chimaerict subunits, in which single domains of the alB subunit have been replaced by the corresponding domain from the alA subunit, all show reduced sensitivity to w-CTx, with subunit ill from alA

_L...-

e_n_try_4_2_

having the largest effect (nine-fold reduction of rate of onset of blockade). Analysis of the effects of a series of mutations in the mSS-llIHS region of the alB subunit showed the crucial importance of residues immediately Nterminal to llIHS, consistent with blockage of the pore by binding of w-CTx at this position2 0 8 . At least some of the mutations that reduced sensitivity to w-CTx-GVIA also reduced the inhibitory effects of w-CTx-MVIIA, suggesting that the two w-conotoxins interact with similar sites on the alB subunit208 .

The peptide blocker w-agatoxin IVA is specific for P-type channels 42-43-19: The peptide toxin w-agatoxin IVA (w-Aga-IVA), a 48 amino acid peptide originally purified from Agelenopsis aperia spider venom, blocks Ptype Ca2+ channel current in rat Purkinje neurones (KD rv 2 nM), but has no effect on identified T-type, L-type or N-type currents in a variety of central and peripheral neurones68 . This block by w-Aga-IVA is reversible by a brief regime of strong depolarizations. The rate of dissociation of toxin from the channel is increased by a factor of about 104 at +90mV compared with -90mV68 . w-Aga-IVA and the related peptide w-Aga-IIIA block glutamate release from pre-synaptic nerve terminals prepared from rat frontal cortex with ICso values of 30nM and O.74nM, respectively, suggesting that P-type channels are important in excitation-secretion coupling in these synapses69. The inhibitory effect of w-.Aga-IVA in rat Purkinje cell axons, measured by depression of post-synaptic currents, had an ICso of rvlOnM 84 .

Funnel-web spider toxin (FTx) blocks P-type channels 42-43-20: The low molecular weight toxin FTx purified from the funnel-web spider, Agelenopsis aperta, specifically blocks the P-type Ca2+ channels in cerebellar Purkinje cells67. The naturally occurring FTx and a synthetic analogue, sFTx, inhibited pre-synaptic 'Ca2+ currents in nerves innervating the mouse levator auris muscle, and blocked muscle contraction and neurotransmitter release evoked by nerve stimulation in an isolated nervemuscle preparation, indicating that the: P-type channel is the predominant voltage-dependent Ca2 + channel in the :motor nerve terminals440.

Neuroleptic drugs can block L-type Ca 2+ channels

42-43-21: Neuroleptic t drugs of the diphenylbutylpiperidine series, usually considered as blockers of the dopamine receptor, also block L-type Ca2+ channels in neuronal and muscle cells. One such compound, fluspirilene, blocks L-type channels in skeletal muscle with an ICso of O.1-0.2nM, independently of membrane voltage441 . These compounds are much less potent in neuronal tissue and smooth muscle, where ICso values of rvlOOnM have been reported442,443. Diphenylbutylpiperidines have been shown to inhibit the binding of dihydropyridines and devapamil to L-type channels noncompetitively, probably via allosteric interactions444,445.

An oxime inhibitor of L-type channels

42-43-22: The nucleophilicr oxime, 2,3-butanedione monoxime (BDM), inhibits muscular contraction and depresses action potentials in muscle fibres. Voltage-clamp measurements show that BDM (rvlOmM) depresses

II

i_

e_n_t_ry_42

---l_

in guinea-pig taenia caeci cells446, rat447 and guinea-pig ventricular myocytes (IC so = 6 mM)448.

[Ca,L

Phenytoin blocks T-type but not L-type channels

42-43-23: Phenytoin (3-100 JlM) blocks T-type Ca2+ channels in NIE-II5 neuroblastoma cells without affecting the L-type channels in these cells. The blockage of T-type curents is use- and voltage-dependent, increasing at higher holding potentials449.

Calciseptine is a specific blocker of L-type Ca 2+ channels 42-43-24: Calciseptine, a 60 amino acid peptide isolated from the venom of the black mamba, Dendroasopis polylepis polylepis, reversibly inhibited contractile activity (IC so = 15 nM) and [Ca,L, but not [Ca,T, in A7R5 rat aortic smooth muscle cells and cardiac preparations45o . The L-type currents in rat insulinoma cell lines (RINm5F and HIT) and in chicken dorsal root ganglion cells were 500/0 inhibited by 1 JlM calciseptine. T-type calcium currents in undifferentiated rat NIE-II5 neuroblastoma cells, N-type currents in chick dorsal root ganglion cells and L-type currents in skeletal muscle cells were insensitive to the toxin (1 JlM)450. SKetJF 96365 is a selective Ca 2+ channel blocker 42-43-25: SK&.F 96365 {1-{,8-[3-{4-methoxy-phenyl}propoxy]-4-methoxyphenethyl}-lH-imidazole hydrochloride}, structurally distinct from the known 'calcium antagonists,' shows selectivity in blocking receptormediated calcium entry (RMCE) compared with receptor-mediated internal Ca2+ release. The IC so for inhibition of RMCE by SK&F 96365 in human platelets stimulated with ADP or thrombin was 8.5 JlM or 11.7 JlM respectively: such concentrations of SK&F 96365 did not affect internal Ca2+ release. Similar effects of SK&F 96365 were observed in suspensions of neutrophils and in single endothelial cells. The effects of SK&F 96365 were independent of cell type and of agonist, as would be expected for a compound that modulates post-receptor events. Voltage-gated Ca2+ entry in fura-2-loaded GH3 (pituitary) cells and rabbit ear-artery smooth muscle cells held under voltage-clamp was also inhibited by SK&F 96365, but the ATP-gated Ca2+-permeable channel of rabbit ear-artery smooth muscle cells was unaffected by SK&F 96365. Thus SK&F 96365 (unlike the 'organic Ca2+ antagonists') shows no selectivity between voltage-gated Ca2 + entry and RMCE, although its lack of effect on ATP-gated channels shows that it can discriminate different types of RMCE451 .

Dextromethorphan blocks L- and N-type Ca 2+ channels 42-43-26: The morphinan dextromethorphan blocks voltage-operated inward Ca2+ and Na+ currents and NMDA-induced currents in cultured cortical neurones and PCI2 cells452 . Dextromethorphan blocks Ba2 + current through L- and N-type Ca2+ channels with similar potencies (52-71 JlM) in both types of cells. The effect is not voltage-dependent, in contrast to that of amlodipine (a dihydropyridine). Dextromethorphan blocks Ba2 + current completely, unlike amlodipine and w-conotoxin, which produce only partial inhibition. Note: The voltage-activated Na+ and Ca2+ channels in cortical

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_2_1

neurones were inhibited by similar concentrations of dextromethorphan (ICso rv 80 JlM)452 but this is rv100-fold less potent than its effect on NMDA receptors (IC so 0.55 JlM) (see Blockers under ELG CAT GLU NMDA, 08-43).

T-type currents in pituitary GH3 cells are resistant to Ni 2+ and ethosuximide 42-43-27: T-type Ca2 + current in pituitary (GH3) cells is relatively resistant to blockade by Ni2+. The concentrations of inorganic cations producing 50% block of GH3 T-type currents are: La3+ 2.4 JlMj Cd2+, 188 JlMj Ni2+, 777 JlM. The T-type currents in GH3 pituita.ry cells are also blocked by the following compounds: nifedipine (rv 50 JlM), D600 (51 JlM), diltiazem (131 JlM), octanol (244 JlM), pentobarbital (985 JlM), methoxyflurane (1.41 mM) and amiloride (1.55mM). Phenytoin and ethosuximide produce 36 and 10% block at 100 JlM and 2.5 JlM, respectively403. This lack of sensitivity to ethosuximide contrasts with the block of T-type currents in thalamic relay neurones obtained with this petit mal anticonvulsant453 . J,

Diphenylbutylpiperidine-based antipsychotics preferentially block T-type currents 42-43-28: Diphenylbutylpiperidine (I)PBP)-based antipsychotic agents, including penfluridol and fluspirilene)' preferentially block T-type Ca2 + currents in the rat medullary thyroid carcinoma 6-23 (clone 6) cell line. Penfluridol inhibited T-type current, with 10mM Ca2 + as charge carrier, with an ICso of 224 OM. The high voltage activated Land N currents in these cells were less sensitive to inhibition by penfluridol. Block of T-type currents by penfluridol was enhanced by depolarizing test pulses. T-type Ca2+ currents in the human TT C cell line were also blocked by penfluridol, and the potency was enhanced by reduction of the extracellular Ca2+ concentration. Non-DPBP antipsychotics, including haloperidol, clozapine and thioridazine, also blocked T-type channels, but these were 20-100 times less potent than the DPBPS454.

Na+, Mg2+ and Ca 2+ compete for tIle same binding site in T-type channels 42-43-29: The T-type Ca2 + channels of chick dorsal root ganglion (DRG) cells are permeable to monovalent cations at low concentrations of external Ca2 + (see paragraph 42-40-04). External Ca2+ ions block Na+ currents with an apparent Kn of 1.8 JlM at -20 mV. Mg2+ ions are also potent blockers of the Na+ current, but much less effective against Ca2 + currents, consistent with the idea that Na+, Mg2+ and Ca2+ compete for the same binding site in Ttype channels376.

Channel modulation Ca 2+ channel activity is modulated via G protein-linked receptors 42-44-01: The activities of voltage-dependent Ca2 + channels are subject to modulation via a very wide range of neurotransmitter receptors. In many cell types different subtypes of channel are modulated independently (see Resource A-G protein-linked receptors). Modulation of Ca2 + channels by

II

1__e_n_t_ry_42

--'_

protein kinase C (PKC) activity in many different tissue preparations has been reviewed98,455. PKC activity results in either down- or up-modulation of Ca2 + channel currents262 with the precise effect showing a strong cell-type dependency. Most reports of PKC modulation show a stimulatory (e.g. smooth muscle) or inhibitory (e.g. sensory neurones, PC12 cells) response on Ca2+ currents (cf. the majority of observations on K+ and CI- channels, which show PKC phosphorylation to have an inhibitory effect, although the lK(v) current of cardiac ventricle is a notable exception to this 455 ). The inhibition of N-type and P/Q-type currents via G protein-linked- receptors usually involves direct, membrane-limited interaction of the activated G protein f3T subunits with the a1 subunit of the Ca2+ channel (reviewed in re!s456,457), though there are examples in which a diffusible second messenger is implicated458-46o.

Inhibition of Ca 2+ currents via G protein-linked receptors 42-44-02: Examples of inhibition of calcium currents by neurotransmitters and neuromodulators acting at G protein-linked receptors are shown in Table 10. Ginsenoside Rf inhibits Ca 2+ channels via a pertussis-sensitive G protein 42-44-03: A saponint, ginsenoside Rf (Rf), isolated from the oriental 'folk medicine' ginseng t , inhibits N-type and other unidentified types of highthreshold Ca2+ channels in rat dorsal root ganglion sensory neurones. Halfmaximal inhibition was achieved at 40 J.1M Rf and the inhibitory effect was eliminated by pre-treatment of the neurones with pertussis toxin (250 ng/ ml for 16h). The receptor responsible for mediating the effects of Rf has not been identified. Similar inhibition of lea by Rf was obtained with differentiated F-11 cells, a cell line that is a hybrid between rat dorsal root ganglia and mouse neuroblastoma cells467.

Purified

GsQ

subunit stimulates L-type channels in vitro

42-44-04: Purified Gsa subunit, activated with guanosine 5'-O-(3-thiotrisphosphate), increases the activity and delays the rundown of porcine cardiac L-type Ca2+ channels incorporated into planar lipid bilayers. This effect is specific for activated Gsa and is inhibited by activated Gil a. The actions of PKA and Gsa are additive, suggesting different pathways for channel stimulation468 . Similarly, the activated as subunit from G s purified from human erythrocytes stimulated Ca2+ currents through single L-type channels in membrane patches from guinea-pig ventricular myocytes469 and bovine cardiac sarcolemmal vesicles incorporated into lipid bilayers47o. In the presence of Bay K 8644 (1 J.1M), activated as increased Popen fivefold, via a decrease in activation, without changing the unitary conductance47o. Evidence for the direct interaction of G s and L-type channels is provided by their co-purification from skeletal muscle471 .

(3-adrenoreceptor agonists increase cardiac Ca 2+ currents 42-44-05: Noradrenaline increases cardiac Ca2+ currents via activation of the ,B-adrenoreceptor, linking to G s and the activation of adenylate cyclase to

II

II

Table 10. Examples of inhibition of Ca 2+ currents by neurotransmitters and neuromodulators acting at G protein-linked receptors

(From 42-44-02) Receptor agonist

Receptor

Tissue

Channel

Acetylcholine

MI

Sympathetic neurones (rat) Superior cervical ganglion neurones (rat) Neuroblastoma x glioma hybrid cell line (NG 108-15) Sensory neurones (rat) Superior cervical ganglion neurones (rat) Ciliary ganglion (chick) Dorsal root ganglion (rat) Hippocampal pyramidal neurones (rat) Adrenal cortical cell line, Y1 (mouse) Superior cervical ganglion neurones (rat) Neuroblastoma x glioma hybrid cell line (NG 108-15) Ventricular myocytes (rat) Sensory neurones (rat) Cerebellar Purkinje cells sympathetic ganglion neurones (bullfrog)

N-type

Acetylcholine

M2/~

Adenosine Al Adenosine Al Adenosine Angiotensin IT (see note 1) All AT-1 Cannabinoids Endothelin-1 GABA GABAB GABA GABAB Luteinizing hormonereleasing hormone (LHRH) Neuropeptide Y NPY receptor Dorsal root ganglion (rat) Myenteric plexus Cerebellar granule neurones Noradrenaline Sensory neurone (chick) 02 Sympathetic neurones (frog) Noradrenaline (see note 2) {3 Hippocampal granular neurone (rat) Opiates Dorsal root ganglion J-t (mouse) 8 Parathyroid hormone Tail artery smooth muscle (rat) (see note 3)

References 598 460 599 600 460

N-type N-type N-type L-type N-type N-type T-type N-type P-type N-type

591 602 603 461 458 502 114 596~597

601 509

N- and T-type

478

Q-type N-type

609

608~139

592 593

N-type N-type

234 479 594

'"

595

L-type

464

('b

~

f""t

~ ~

t:..J

Prostaglandin £2 Somatostatin Substance P Vasoactive intestinal polypeptide (VIP) (see note 4) Vasopressin

SSTR4

VI

Superior cervical ganglion neurones (rat) Pituitary cell line (AtT-20) Heterologous expression in Xenopus oocytes Superior cervical ganglion neurones (rat) Sympathetic neurones (rat)

N-type L- or N-type N-type N-type N-type

A7rS smooth muscle cells

L-type

604 605,606

552 607

466

610

Notes: 1. Angiotensin II stimulates Ca2+ currents in the adrenal cortical cellline461 and in bovine adrenal glomerulosa cells462,463. 2. The N-type currents in rat hippocampal granular neurones are elevated by noradrenaline234 . 3. Bovine PTH depresses L-type Ca2+ currents in smooth muscle cells from rat tail artery464, but synthetic rat PTH fragment (1-34) stimulates ICa,L in snail neurones, probably via activation of PKC465 . 4. The inhibitory action of VIP is pertussis toxin insensitive, but reduced by cholera toxin and anti-Gs antibodies, suggesting that it involves G s 466.

II

('b

= ~ t"t-

~ ~

_'--

e_n_try_4_2

----J

increase cAMP levels and potentiate protein kinase A (PKA}230,472. The effect of phosphorylation by PKA is to increase Popen 472 and the mean open time of the cardiac channel231 (see Protein phosphorylation, 42-32). The stimulation of cardiac L-type currents via the ,B-adrenoreceptor, Gs , PKA pathway has been reviewed98 . Studies of ,B-adrenergi.c stimulation of L-type channels in smooth muscle cells have produced highly variable results, with inhibition, stimulation and no effect all being reported for different preparations (reviewed in ref.98).

Facilitation of Ca 2+ channel activity involves relief of G proteinmediated inhibition 42-44-06: Depolarizing pre-pulses (e.g. +80 mV for SO ms) transiently increase, or 'facilitate l ', Ca2+ currents by reversing the G protein-mediated inhibition induced by a neurotransmitter, GTP,S473-475 or under basal conditions in the absence of a transmitter474,476. In rat cerebral cortical neurones, maximally effective pre-pulses (+80mV for SOms) facilitated Ca2+ current by 23% with internal GTP (300 JlM) and by SOC~ with internal GTP,S (300 IlM)265. The pre-pulse facilitation was greatly reduced (6%) in the presence of internal guanosine S'-O-(2-thiodiphosphate) (GDP,BS) (2IlM), an inhibitor of G protein activation. In some, but not all, cortical neurones, pre-pulse facilitation was inhibited by w-conotoxin GVIA (10 IlM) block of N-type Ca2+ channels. Similar results were obtained in parallel experiments with rat superior cervical ganglion neurones~!65. Pre-pulse facilitation in cerebral cortical and superior cervical ganglion neurones was strongly reduced or prevented by activation of PKC265.

Neuronal Ca 2+ channels are subject to tonic inhibition by G proteins 42-44-07: In several (but not all) neuronal cell types, the Ca2+ current is subject to 'tonic inhibition' by intrinsic G protein activity in the absence of added agonist474,476,477. This inhibition is relieved by intracellular guanosine S'-[,B-thio]diphosphate, an i.nhibitor of G protein activation. Activation of G protein via a receptor agonist further inhibits Ca2+ channels that are already tonically inhibited. Following removal of an inhibitory agonist by washout, the tonie inhibition is temporarily removed. This 'rebound facilitation', which fades within a few minutes, is prevented by the inhibitory action of pertussis toxin on the G protein474.

The inhibitory G protein is often Go 42-44-08: In sensory neurones, inhibition. of calcium currents by neuropeptide Y can be prevented by blocking ao, and inhibition reappears following reperfusion with activated ao478. The activated Gao subunit is also able to restore opiate-mediated inhibition of Ca2+ currents in pertussis toxintreated NGI08-IS cells479 and dopamine-induced inhibition of Ca2+ currents in snail neurones480. Such experiments can be criticized because of the use of artificially high concentrations of non-physiological G proteins, which might non-specifically dominate the signal transduction process 41 . The use of a mutant NGI08-IS cell line in which the Go al subunit is resistant to pertussis toxin overcomes this criticism. In this mutant line,

II

lL...--_ _ _ry_4_2 en t

_

inhibition of Ca2+ currents by opioids and noradrenaline are resistant to pertussis toxin, in contrast to its sensitivity in the wild-type parental line, demonstrating that the transduction pathway involves Go a1 481 . The inhibitory effect of somatostatin remained toxin-sensitive in the mutant cell line, consistent with its involving a different Ga subunit481 . The use of anti-G protein antibodies and antisense DNA oligonucleotides to identify the species of G protein involved in modulating Ca2+ currents has been reviewed in ref. 41 (see also multiple effects of Go on K+ currents 482 and Resource A-G protein-linked receptors). The Gao subunit co-purifies with the N-type voltage-dependent Ca2+ channels from rat sympathetic neurones 483 .

The Ca 2+ channel {3 subunit is involved in the interaction with Go 42-44-09: The DHP agonist Bay K 8644 enhances GTPase activity in neuronal membranes, via an effect on the V max of the enzyme. This stimulation is blocked by antibodies against Gao, but not by anti-Gai antibodies484, and by antibodies against the Ca2+ channel f3 subunit (quoted in ref. 41). Depletion of the Ca2+ channel f3 subunits using antisense r oligonucleotides in dorsal root ganglion neurones enhances the ability of the GABAB agonist (- )-baclofen to inhibit the Ca2+ current. These observations have been interpreted to suggest that Gao binds to a site on the channel al subunit in such a way as to inhibit, sterically or allosterically, the association of the channel al and f3 subunits 41 . The modulation of N-type channels involves G protein {3, subunits 42-44-10: The whole-cell ICa in adult rat sympathetic neurones, largely carried by N-type Ca2+ channels, is subject to inhibition known as 'kinetic slowing' by 10 JlM noradrenaline or direct G protein activation by intracellular guanylyl-imidodiphosphate (GppNHpi 500 JlM), a non-hydrolysable GTP analogue. Intranuclear injection of eDNA constructs t encoding G{31 and G1'2 subunits in sympathetic neurones gave rise to kinetic slowing of lca 'virtually identical' to that obtained with noradrenaline or GppNHp. Similar inhibitions were observed in neurones injected with G{31 + G1'3 or G{31 + G1'7 cDNAs, but not in those injected with either Gf31 or G1'2 eDNA alone. The injection of a eDNA encoding a constitutively active GaDA subunit (Q20SL mutant) produced no apparent effect. Application of noradrenaline (10 JlM) to neurones previously injected with G{31 + G1'2 cDNAs produced little additional effect. Titration of endogenous Gf3l' subunits by overexpression of injected eDNA encoding Gao attenuated the ability of noradrenaline to inhibit lca. Very similar observations have been made by injection of purified bovine brain Gf3l' protein or RNA encoding Gf321'3 into rat superior cervical ganglion neurones which express N-type Ca2+ channels485 . These findings demonstrate that the voltage-dependent inhibition of neuronal N-type Ca2+ channels by noradrenaline is mediated by the G protein {3, 1 subunits48s ,486.

G protein {3, subunits modulate PIQ-type Ca 2+ channels 42-44-11: Co-transfection of cDNAs encoding alA, ,BIb and a28 subunits of PI Q-type Ca2+ channels into tsA-201 cells gives rise to Ba2+ currents that

II

_'--

e_n_try_4_2_1

respond to activation of endogenous G proteins by GTP,S and were facilitatedt by a large depolarizingt pre-pulse in the presence of the GTP analogue. Co-transfection of tsA-201 cells with cDNAs encoding the Ca2+ channel subunits and G protein subunits, G{32. plus G-r3' mimicked G protein activation, shifting the voltage dependence of channel activation to more positive potentials and allowing facilitation by a conditioning prepulse. The G{32 subunit alone was nearly as effective as the {32,3 combination in modulating Ca2+ channel behaviour, but the G'3 subunit alone was without effect. The pertussis toxin-sensitive Ga subunits, Gail, Gai2, Gai3 and GaOA, had very little effect on voltage dependence and none on facilitation of the heterologously produced P/Q-type channels485 .

G protein (3"'( complex binds to the I-II cytoplasmic loop of the subunit

alA

42-44-12: Radiolabelledt G{3112. complex:, prepared by in vitro translationt in the presence of [35S]-methionine, binds directly and specifically to fusiont proteins containing the cytoplasmic loop connecting repeats I and IT (amino acids 360-486) of the Ca2.+ channel alA subunit487. Similar binding is found to the analogous I-IT loop regions of alB and alE subunits, but not with those from alS or ale subunits. The binding of the {3, complex involves two regions within the I-IT loop of alA, the al interaction domain (AID), a sequence of 18 amino acids that is necessary and sufficient for binding of the Ca2+ channel {3 subunit (see paragraph 42-29-06), and a second region (D2) within amino acids 402-487. The apparent Kd of the in vitro binding of the G{3I,2 complex to the AID sequence of alA is 63llM487, suggesting a 10- to 20-fold lower affinity than that of the Ca2+ channel {3 subunit488 . Binding of {3, to the D2 region of alA occurs with an apparent Kd of 24llM. Within the AID region of alA, which has the sequence QQIEBE1NGYMEWISKAE, alteration of the underlined residues strongly reduced or eliminated in vitro {3, binding. (Note that Y-WI (10-14) has been shown to be critical for binding of the Ca2+ channel f3 subunit489.) Calcium channels containing the mutant alA subunit with a R387E replacement within the AID sequence are insensitive to activation of G proteins in the Xenopus oocyte487.

Phosphorylation in the I-II loop of :the inhibition by G(3"'(

al

subunit antagonizes

42-44-13: The inhibition of Ca2+ currents by activated G proteins can be studied during transient expression of cl)NAs encoding alA or alB subunits, together with those encoding a2 and f3lb' in human embryonic kidney (HEK) cells. Stimulation of an endogenoust somatostatin receptor with 100llM somatostatin gives 650/0 inhibition of alB currents and 23% inhibition of alA currents. This inhibition can be relieved by a strong (150 mY) depolarizing pre-pulse ('pre-pulse facilitation'). Purified Gf3r subunits (10nM) in the patch pipette also mediate an inhibition of the Ntype currents from alB-containing cha.nnels that is subject to pre-pulse facilitation. Synthetic peptides (2 JlM) containing sequences from the alA or alB I-IT loop region applied via the patch pipette were able to relieve the G{3,-imposed inhibition: peptides containing aIB(353-371), aIB(372-389), aIB(410-428), aIA(384-403) and aIA(416-434) were all effective in this

II

1....._e_n_t_ry_4_2

-----I_

assay. In vitro phosphorylation of the aIB(410-428) and aIA(416-434) peptides by protein kinase C (PKC) eliminated their ability to quench the G{3'Y inhibition. These data support a model in which inhibitory G{3'Y interaction with I-II regions of alA or alB subunits is affected by PKC-dependent phosphorylation of residues within the Gf3'Y-binding site490 •

Adenosine receptors modulate Ca 2+ currents in hippocampal neurones 42-44-14: The selective activation of adenosine receptor subtypes has different effects on Ca2+ channels from acutely isolated pyramidal neurones from the CA3 region of guinea-pig hippocampus. Activation of adenosine Al receptors primarily inhibited N-tvre Ca2+ current, while activation of A2b receptors produced potentiation of P-type, but not N-type, Ca2+ current. The potentiation was blocked by inhibition of protein kinase A491.

Adenosine counteracts cAMP-dependent stimulation of cardiac L-type channels 42-44-15: The effect of PI purinoceptor (adenosine Al receptor) stimulation on L-type channels in cardiomyocytes depends on the whether the preparation is in a 'basal' state or in a cAMP-mediated state of activation. The Al receptors are coupled to inhibitory Gi, so that adenosine depresses f3-adrenergic stimulation of adenyl cyclase activity, cAMP formation and PKA-mediated stimulation of L-type Ca2+ currents. The isoproterenolstimulated L-type Ca2+ currents in guinea-pig ventricular492, guinea-pig atrial493, rabbit sinoatrial node494 and frog ventricular495 myocytes is markedly diminished by adenosine acting via a PTx-sensitive G protein. The suggestion that adenosine acts by activation of a phosphatase496, via Gi, rather than by inhibition of adenylate cyclase, remains plausible.

Dopamine D 1 receptors activate L-type channels in chromaffin cells 42-44-15: Stimulation of the dopamine D I receptors in bovine chromaffin cells activates Ca2+ currents in the absence of pre-depolarizations or repetitive activity. This activation by D I agonists, mediated by cyclic AMP and protein kinase A, results from prolonged openings of an otherwise quiescent 27pS L-type channel298 .

Angiotensin II stimulates L-type currents in an adrenal cortical cell line 42-44-16: The vasoactive octapeptide angiotensin IT is the major stimulator of aldosterone secretion from adrenocortical glomerulosa cells, an effect that is dependent on Ca2+ influx through voltage-dependent Ca2+ channels. Studies with the murine adrenocortical cell line, Y1, showed that angiotensin IT (1 nM to 1 J,lM) stimulates a slowly inactivating L-type current, on average 1.7-fold. A rapidly inactivating T-type current was not affected by angiotensin IT. The stimulatory effect of angiotensin II on L-type currents was blocked by pertussis toxin (100ng/ml) acting to modify the a subunit of a Gi-like G protein461 . A similar stimulatory effect of angiotensin II (1 J,lM) on L-type currents was shown in freshly isolated porcine glomerulosa cells461 .

II

_L.-.---------------------e-n-try-4-2--1

Q-type Ca 2+ channels respond to neuromodulators 42-44-17: The Ca2+ currents required to stimulate glutamatergic t synaptic transmission between hippocampal CA3 and CAl neurones are carried by N-type and Q-type channels61 . Blockage of the N-type channels with w-conotoxin GVIA (1 J.1M) allows the effects of neuromodulators on Q-type channel activity to be monitored. Stimulation of metabotropic glutamate (IS,3R-ACPD, 200J.1M)), 1'-aminobutyric acid type B ((-)-baclofen, 5 J.1M), adenosine (2-chloroadenosine, 5 J.1M) or acetylcholine (carbachol, 10 J.1M) receptors significantly depressed synaptic transmission under these conditions. Prior application of the phorbol ester phorbol 12,13-dibutyrate (1 J.1M) to stimulate PKC activity, reduced or abolished the inhibitory effects of the neuromodulators on Q-type channel activity61. An endogenous brain peptide modulates L-type and T-type currents 42-44-18: Calcium channels can be modulated following application of a low molecular weight endogenous peptide purified from rat brain: in cardiac cells L-type Ca2 + currents are enhanced while in neuronal cells both L-type and T-type currents are inhibited497.

Galanin inhibits Ca 2+ channels via two species of G protein 42-44-19: Galanin, a neuropeptide of 29 amino acids that is widely distributed throughout the central t and peripheralt nervous systems, inhibits voltagegated Ca2 + channels by interaction with a G proteint -linked receptor in rat insulinoma RINm5F cells and in rat pituitary GH3 cell line498-5oo. Microinjection of antisense t oligonucleotides t designed to inhibit expression of genes encoding different G protein subunits shows that both these cell types couple the galanin receptor to the Ca2+ channel mainly via the Go protein consisting of o.Ol{32,2, with the o.Ol{33,4 species also being used, but less efficiently501. Cannabinoids inhibit N-type Ca 2+ channels via a pertussis toxin-sensitive G protein 42-44-20: The cannabinomimetic aminoalkylindole WIN 55,212-2 reversibly inhibits ICa in the neuroblastoma-glioma cell line NGI0S-15, with an ICso of 'less than 10nM'. The effect is stereospecifict , the enantiomert WIN 55,212-3 being without effect at 1 J.1M, and is blocked by pertussis toxin (500 ng/ml). The inhibitory effect of the cannabinoid was eliminated in the presence of w-conotoxin (1 mM), a specific blocker of N-type channels, which decreased ICa by 380/0, but was insensitive to dihydropyridine antagonism that reduced ICa in NG108-15 cells by 27%502.

Fatty acids modulate lea in smooth muscle 42-44-21: Free fatty acids such as arachidonic acid (AA) have potential roles in the physiological and/or pathological modulation of lca in smooth muscle. In smooth muscle cells from rabbit ileum 10-30 J.1M AA causes a gradual depression of lca 503. This inhibitory effect is not prevented by the cyclooxygenase inhibitor indomethacin (10 J.1M) or the lipoxygenase inhibitor nordihydroguaiaretic acid (10J.1M) (see also lLG K AA, entry 26). The PKC inhibitors H-7 and staurosporine do not mimic this action of AA. Certain

II

1L...-__ _ _ry_42

en t

_

other cis-unsaturated fatty acids (palmitoleic, linoleic and oleic acids) can also depress ICa, while a trans-unsaturated fatty acid (linolelaidic acid) and saturated fatty acids (capric, lauric, myristic and palmitic acids) show no inhibitory effects on ICa. Myristic acid consistently increases the amplitude of lca at negative membrane potentials503 .

Ca 2+ influx into hypothalamic neurones can inhibit voltage-gated Ca 2+ currents 42-44-22: In cultured rat hypothalamic neurones both NMDA and nonNMDA receptor-channels (see ELG series, entries 04-11) are permeable for Ca2+. Calcium influx through these channels activates a calmodulindependent mechanism, which can lead to high voltage activated Ca2+ current inhibition504.

Inositol phosphates modify gating of Ca 2+ channels 42-44-23: Intracellular InSP3 and InsP4 have been shown capable of eliciting Ca2+ entry into rat cerebellar neurones by modifying the gating characteristics of voltage-dependent calcium channels. InsPa (ECso 0.5 JlM) shifted the steady-state inactivation curve towards more positive values 505.

Modulation of voltage-gated Ca 2 + channel function by ethanol 42-44-23: Ethanol inhibits voltage-dependent Ca2+ flux in vitro in synaptosomes and pre-synaptic nerve terminals (threshold, 25 mM; ICso > 150 mM)506 and in cultured phaeochromocytoma (PC12) cells 94,507. The electrophysiological and pharmacological characteristics of the channel involved in PC12 cells identify it as an L-type channel507. Chronic ethanol treatment increases dihydropyridine-binding sites in membranes from whole brain and cerebral cortex of mice and rats, an effect that is blocked by Ca2+ channel antagonists. Increases in dihydropyridine binding are also found in heart tissue of ethanol-dependent rats and in adrenal chromaffin cells and PC 12 cells after growth in the presence of 200 mM ethanol for 6 days 94. Menthol, a major constituent of peppermint oil, acts as an antagonist of L-type Ca2+ channels at concentrations of 10-100 JlM508 . A figure illustrating the relative potentiating effects of ethanol on GABAA mediated CI- flux and the depression of NMDA-, kainate- and voltagegated Ca2+ flux in synaptosomes from rat cortex appears as Fig. 6 in Channel modulation under ELG Cl GABAA, 10-44.

N-type Ca 2 + currents are inhibited by adenosine A 1 receptor activation 42-44-24: In acutely isolated pyramidal neurones from the CA3 region of guinea-pig hippocampus, the Al adenosine receptor agonist 2-chloroadenosine (2-CA) (100 JlM) inhibits ICa by 14%. The N-type Ca2+ channel blocker w-conotoxin GVIA (w-CTx) (5 JlM) removed most of the inhibition by 2-CA, confirming that the Al receptor agonist primarily inhibits N-type channels. The effect of 2-CA is voltage dependent, the I-V relationship being shifted toward a more negative potential in the presence of the agonist491 .

II

_L...-

e_n_try_4_2_1

Inhibition of N-type Ca 2+ currents by luteinizing hormone-releasing hormone 42-44-25: The chicken type IT luteinizing hormone-releasing hormone (LHRH) inhibited the N-type Ca2+ and Ba2+ currents of neurones from bullfrog paravertebral sympathetic ganglia with a half-maximally effective concentration of 20 nM509. There was no detectable effect on the nimodipine-sensitive L-type currents in these preparations. The inhibition of N-type currents was sensitive to the activation state of the channels: the hormone had little effect when applied during a long depolarization (-20mV) that opened the channels, but was effective when applied in the resting state (-90mV) and currents were elicited by short (5ms) depolarizations (-20mV) every 200ms. Inhibition was relieved when channels were activated by short (3 s) pre-pulses to membrane potentials above -30 mV. A kinetic model to fit the data proposes that inhibition results from the binding of activated G proteins to N-type channels to stabilize the closed state, and that activation of the channel complex destabilizes the binding of the G protein. The preferred version of the model suggests the binding of four G protein subunits per Ca2+ channel, one for each domain of the channeI509.

G protein-linked receptors can activate Ca 2+ currents 42-44-26: Activation of receptors in secretory cells can enhance Ca2+ currents via pertussis toxin-sensitive G proteins (reviewed in ref. 510). Thyrotropinreleasing hormone stimulation of Ca2+ currents in rat pituitary GH3 cells requires activation of both PKC and Gi2511.

P-type Ca 2+ currents are potentiated by activation of adenosine A 2b receptors 42-44-27: In the presence of the Al receptor antagonist 8-cyclopentyl-l,3dimethylxanthine (CPT), exposure of guinea-pig CA3 hippocampal neurones to adenosine caused a 63.5% increase in lca: w-CTx had no effect on the potentiated ICa, suggesting that N-type Ca2+ channels are not involved in this augmentation. The potentiationt was strongly inhibited by w-agatoxin NA (0.1 J.1M), a specific P-type Ca2+ channel blocker, establishing the involvement of P-type channels in activation via adenosine receptors. The specific adenosine A2 receptor agonist N6-[2-(3,5-dimethyloxyphenyl)-2-(2-methylincreased lca by 33% above the phenyl)ethyl]adenosine (DPMA) (0.1 ~i), control value. The highly selective adenosine A2a receptor agonist CGS 21680 (1 J.1M), had no effect on ICa, suggesting the conclusion that the A2b adenosine receptor is probably involved in the potentiation of lca by adenosine491 . The protein kinase A inhibitor peptide WIPTIDE (10 JlM) prevented ICa potentiation by CPT and adenosine, establishing that the potentiation via A2b activation involves cAMP-dependent protein kinase activity491.

T-type channels are modulated by G protein activation 42-44-28: The T-type Ca2+ channel currents from cultured rat dorsal root ganglion neurones can be modulated by G protein activation. Photorelease 5'-O(3-thio) trisphosphate (GTP,S) from a of intracellular ~anosine photolabile 'caged't precursor had dose-dependent effects on the T-type

1'--_e_n_try_4_2

_

current. At 6 J.1M, GTP,S enhanced the current, but higher concentrations (up to 20 J.1M) were inhibitory. The inhibitory response, but not the stimulation, was sensitive to pertussis toxin, suggesting the involvement of more than one G protein in T-type Ca2+ channel modulation. Low concentrations of the GABAB agonist (-)-baclofen (2 J.1M), potentiated the T-type current, but 100 J.1M(-)-baclofen was inhibitory512. Bradykinin (0.1 J.1M), baclofen (2 J.1M) and internal GTP,S (100J.1M) inhibit T-type currents in the dorsal root ganglion-neuroblastoma cell line, ND7_23 513 .

Endothelin enhances T-type Ca 2+ currents 42-44-29: Endothelin-l (ET-1), a 21 amino acid vasoconstrictive t peptide, increases intracellular Ca2+ levels and has hypertrophic action on ventricular myocytes. In cultured neonatal rat ventricular myocytes, ET-1 (10nM) increased the maximum current density of Ica,T from -3.0J.1A/cm2 to -4.4J,1A/cm2 . This enhancement by ET-1 was dose dependent, with the maximal response at approximately 10nM and a half-maximal dose of 1.3 nM. The stimulation of ICa,T was antagonized by protein kinase C inhibitors staurosporine (0.2 J.1M) and 1-(S-isoquinolinesulphonyl)-2-methylpiperazine (B-7, 20 J.1M) in the pipette solution. Extracellular application of phorbol esters, activators of protein kinase C, also increased the maximal current density of ICa,T' an effect that was blocked by H-7 (20J.1M) in the pipette solution. The application of ET-1 had no significant effect on ICa,L in this systeml14 .

Stimulation of the j-t-opioid receptor inhibits P/Q-type and N-type Ca 2+ channels 42-44-30: During heterologousteo-expressiont of cDNAst encoding the J..topioid receptor and Ca2+ channel subunits in Xenopus oocytes, activation of the receptor with the synthetic enkephalin [n-Ala2 , N-Me-Phe4 , GlyolS]enkephalin (DAMGO; 1 JlM) produced rv20% inhibition of Ba2+ currents from P/Q-type channels (alA, a2, ,84) -and rvS5% inhibition of those from Ntype channels (alB, a2, ,84). The inhibitory effect of DAMGO was reversible on washout and prevented by the opioid receptor antagonist naloxone (10 J.1M), and the G protein antagonist pertussis toxin (2 J.1g/ml). The inhibition was not obtained when the channels contained the alC or alE subunits. The currents obtained following expression of alA cRNA, in the absence of a2 and {3 subunits, were also sensitive to opioid inhibition, but the effect was enhanced about threefold in the absence of the ,8 subunit. The absence of the a2 subunit was without effect on the degree of inhibition. On the basis of these results, it has been suggested that the activated G protein, probably via the released {3" subunits (see paragraph 42-44-11), interferes with Ca2+ channel {3 subunit binding to the alA I-IT linker, or affects {3 subunit interactions in other regions of the al subunit514.

Equilibrium dissociation constant Dissociation constants for binding of drugs to L-type channels 42-45-01: The dissociation constants for the binding of several drugs to L-type Ca2+ channels are shown in Table 11. Note that the drugs are likely to have

II

II Table 11. Equilibrium dissociation constants for binding of various drugs to L-type Ca 2+ channels (From 42-45-01) Class

Dissociation constants (nM)

Ligand

Reference

Skeletal muscle

Heart

Brain

(-)-Azidopine (- )-Iodopine (- )-Sadopine (+ )-Sadopine

0.29-0.7 0.2 0.35 0.4 0.51 0.4

0.051 0.052 0.030 n.d. n.d.

0.075 0.044 0.096 0.06 n.d.

(- )-Desmethoxyverapamil (Devapamil) N-methyl-LU 49888

1.5-2.2 2.0

1.4-2.5 n.d.

1.6 1.4

620,621

Benzothiazepines

(+ )-cis-Diltiazem

39-50

40-80

37-50

622-624

Diphenylbutylpiperidines

Fluspirilene

0.100

0.070

n.d.

441,445,625

Benzothiazinones

HOE-166

0.100

n.d.

n.d.

428

1/4-Dihydropyridines

Phenylalkylamines

(+) PN 200-110

611 612 183,613-615 616 617

618,619

I

~

='

~

~ ~

1__e_n_t_ry_4_2

_

higher affinity for a particular state of the channel. For example, the apparent Ko for nitrendipine block of ICa,L in canine ventricular myocytes is 0.73 J.1M at a holding potential of -80 mY, but 0.36nM at about -15 my411 , reflecting the preferential binding to the open and/or inactivated states, compared to the resting state of the channel.

w-Conotoxin inhibits Ca 2+ channels in an electric ray 42-45-02: w-Conotoxin (w-CTx) gives dose-dependent inhibition of increases in free calcium concentration in, and acetylcholine release from synaptosomes isolated from a Japanese electric ray, Narke japonica, following

depolarization with a high concentration of potassium ions. Half-maximal inhibitions (IC so )of the increase in intrasynaptosomal Ca2+ and acetylcholine release were obtained using 8 and 7 J.1M w-CTx, respectively. Assay using radioiodinated toxin revealed a specific binding site with a Ko of 2.8 J.1M and a density of 45 pmol/mg synaptosomal protein. Binding assay with synaptosomal plasma membrane showed a K o = 7 J.1M and a density of 200pmol/mg protein. The radiolabelled toxin was covalently cross-linkedt (by disuccinimidyl suberate) to a protein that has a molecular weight of 170000 as determined by SDS-PAGEt and autoradiography515.

The peptide toxin w-agatoxin IVA is a potent blocker of P-type channels 42-45-03: The 48 amino acid peptide toxin w-agatoxin IVA (w-Aga-IVA) from the venom of the funnel web spider, Agelenopsis aperta, blocks P-type Ca2+

channel currents in rat Purkinje neurones with an ICso of "",2 nM68 • Note that this value differs markedly from the 200nM w-Aga-IVA that gives 50% block of currents obtained after co-expressiont of cRNAs encoding rabbit 0IA, 02/6 and {31 subunits in Xenopus oocytes 13 .

Cone snail toxin w-CTx-MVIIC is a potent blocker of a lA -containing channels 42-45-04: The 26 amino acid peptide toxin, w-conotoxin MVIIC (w-CTxMVIIC), from the piscivorous marine snail Conusrnagus, is a blocker of both N-type and P-type Ca2+ channels, with an ICso of 1-10 J.1M for P-type channels516 . In Xenopus oocytes expressing cRNAs encoding 0IA, 02/6 and {31 subunits, w-CTx-MVIIC blocked I Ba with an IC so of 60kb of X chromosome DNA 43-20-06: The human C]CN4 gene has been shown by long-range t mapping in

the Xp22 region to occupy 60-80 kb of chromosomal DNA and to contain at least ten exons11 .

The human G1GN5 gene has 12 exons 43-20-07: The coding region of the human C]CN5 gene has been shown

to consist of 12 exons l and the exon-intron t boundaries have been determined by a PCRt -based strategy on cDNA and human genomic DNA templates 1 . None of the introns t of theCICN5 gene correspond in position with any of the 22 introns of the CICNI gene1 .

II

-----l_

II--_e_n_t _ry_43

~

~

221 ~

1

===~I ==~::~~~_-_I_~_81

2

3

,_302_..~,.D.JIOo iI

.......4_34...-.....;,.m&&.

333 ~

4

6

;_! ~_63 ~1m1I~ ...........

8

:~_~4_ _.,lC&

.....'...,;;,;::_ .........'_7...-:m __'......

,~- '980

__........

ATG

1 ~

~

234 ~

~

220 ~

9

10

,."S~

,' l}(: ..

)6., ..

. ~

188 ~

12

11

14

11~83

, 111 •. . .1.6677 ' _115_2_ _ 11402 11472

"/~DJf~~

CL

l'2

~

bp

rl73

18

19

16

I>lti

S

CL

~

383

17

~

13

93 bp

20

22

21

~r--2~_96

....... ,

==

STOP

2967

Figure 1. Structure of the human CLCNI gene. The complete protein-coding region of the gene is shown as three open boxes, with the first base of the initiation codon (ATG) as nucleotide 1 and the last base of the termination codon as nucleotide 2967. The grey areas represent regions of exons that encode putative transmembrane domains (D1-D12). D13 is a highly conserved cytoplasmic segment near the C-terminus of the protein. Arrowheads indicate the positions of introns (1-22) that interrupt the coding region: the sizes of some introns are indicated above the arrows. The exons are drawn to scale, with the position of the first bases of the exons indicated by the numbers to the right of the arrowheads. (Reproduced with permission from Lorenz (1994) Human Mol Genet 3: 941-6.) (From 43-20-05)

Homologous isoforms 43-21-01: The homologies between the various CIC proteins are shown in

Table 4.

The tkidney G1G' subfamily 43-21-02: Several cDNAst with sequence homologyt to the CIC family have been cloned from rat2,17 and human2 kidney RNAs. The rat CIC-Kl cRNA

gives rise to 'slightly outwardly rectifying time- and voltage-independent chloride currents I in Xenopus oocytes 171 but there are reports that the kidney-specific channel cRNAs (ClC-Kl 1 ClC-K21 ClC-Ka and CIC-Kb)1 either singly or in pairwise combinationsl do not give rise to detectable currents in the Xenopus oocyte system2 . l

A short variant of G1G-2K in rat kidney 43-21-03: A cDNA isolated from rat kidney mRNA1 CIC-K2S1 encodes a

632 amino acid protein in which 55 amino acids containing the putative

II

II

Table 4. Identities (per cent) between the various CiC isoforms (From 43-21-01)

CIC-O CIC-O CIC-l CIC-2 CIC-3 CIC-4 CIC-5 CIC-Kl CIC-K2 CIC-Ka

CIC-l

CIC-2

CIC-3

55

49 55

20.9 24 26.3

CIC-4 1 s (in comparison to 'A-type' currents, which usually inactivate in 1 s (reviewed in ref.l and in Hille, 1992, see Related sources and reviews, 45-56); these include K(v), K(dr), K(DR) or DRK, the first three normally signifying the subscripted forms, Le. KV1 Km, KDR- The: terms K-DRI and K-DR2 channels have been used for multiple distinguishable subtypes of delayed rectifier in circular smooth muscle of canine colon2 . 'I-channels' or 'I-type' channels are used in ref. 3 to describe delayed rectifiers in peripheral myelinated axons from Xenopus laevis. Dorsal root ganglion neurone (DRG) channel subtypes distinguishable on the basis of their single-channel conductances, kinetics and sensitivity to external tetraethylammonium ion have been named DRF1 DR11 DR21 DR34. Note also the term 'delayed rectifier' has been applied to proposed channel activator proteins such as 1sK (minK) -

for clarification, see VLC (K) minK, entlY 54.

1__e_n_t_ry_45

_

Current designation Most designations in use do not make reference to the underlying subunit composition 45-04-01: Generally, I K; also variants IK(dr); IK(dK); IK(v) (i.e. the 'd' subscript representing gelayed outward K+current); Ix (in cardiac muscle; see next paragraph). Multiple components of the native delayed rectifier current 1dK in canine colonic myocytes have been described5 with protocols which separate the current into three distinct components that differ in their kinetics and pharmacology (I-dKft I-dKs and I-dKn ). In embryonic chick dorsal root ganglion (DRG) neurones 6 InRI< also passes Cs+ and Rb+; under conditions where intracellular [K+ h is reduced to 35 kb at the genomic level) included the mutant sequences associated with four different eag mutant alleles (see Phenotypic expression, 46-14). A hybridization t probe based on the hydrophobic core sequences of Drosophila eag was subsequently used to retrieve further cDNA clones encoding eag and those encoding elk, the eag-like K+ channel from Drosophila head-specific cDNA libraries 1 . The isolation of cDNA clones encoding the Drosophila Eag K+ channel polypeptide (see Isolation probe, 46-12) formed a prototype for an extended gene family as a distinct branch of the voltage-gated channel gene superfamily. Subsequently isolated members have been placed in eag, elk and erg subfamilies based on their sequence similarity. Isolation of other clones (or species homologues of existing clones) in the gene family may clarify and extend the present classification. The eag designation has been retained for mammalian species homologues (see Table 1 under Gene family, 46-05). 46-01-02: Presently known eag, elk and erg subfamily members share at least 470/0 amino acid identity in their hydrophobic cores with all sequences possessing a segment homologous to a cyclic nucleotide-binding domain. Additional localized sequence comparisons indicate that members of this family are distantly related to certain 'six transmembrane domain' inwardly rectifying K+ channels of plant origin (e.g. clones KAT1, AKT1 and KST1, see descriptions under Domain conservation, 46-28). 46-01-03: Higher mammalian erg genes have undergone much functional adaptation from their likely common ancestral origin with present-day voltage-gated channel superfamily genes. For example, an erg subfamily cDNA originally isolated from a hippocampal cDNA library and designated

_ _ _ _ _ _ _ _ _ _.

en_t_ry_4_6_

as the human ~ag-!elated gene (h-erg or HERG (see Table 1 and Isolation probe, 46-12) displays strong K+ inward rectification following heterologous expression in oocytes due to the operation of a novel inactivation mechanism which prevents K+ efflux during depolarization2 (see Fig. 6 under Activation, 46-33). Otherwise, the biophysical properties of human erg subfamily channels expressed in Xenopus oocytes are very similar to the rapidly activating IK,r component in native cardiac myocytes (associated with initiation of terminal repolarization in the cardiac action potential). Unique fast-inactivation properties of erg subfamily channels which result in their characteristic inwardly rectifying property are described under Inactivation, 46-37. 46-01-04: Mutations in HERG are associated with the chromosome 7-linked form of long QT (LQT) syndrome. Gene linkage and physical mapping analyses have co-localized the type 2 LQT syndrome pedigree marker LQT2 and HERG to chromosome locus 7q35-36 (ibid. 3 ). Single-strand conformation polymorphism t (SSCP) and direct sequence analyses in type 2 LQT families have revealed a spectrum of mutations in HERG, including, to date, two intragenic deletions, one splice-donor mutation and three missense mutations (for further details and description of delayed repolarization arrhythmias, see Phenotypic predisposing to life-threatening card~iac expression, 46-14 and Chromosome location, 46-18). 46-01-05: The beneficial effects of /3-blockade in preventing development of arrhythmias may be due to cAMP-dependent regulation of channel subunits other than HERG, important candidates being KvLQT1, voltagegated Na+ channels and voltage-gated Ca2 + channels. Different LQT2 mutations may affect HERG function in different ways, and there appears to be a role for additional subunits in native HERG function. 46-01-06: Rat eag mRNA is expressed at high levels in brain, with no apparent signal outside the CNS, at least by relatively low-sensitivity Northern blots. In situ hybridization studies indicate that rat eag mRNA is most prominently expressed in the hippocampal formation and cerebellum. The highest levels of eag transcripts were detected in the granular cells of the dentate gyrus, in the CA3 pyramidal cells of the hippocampal formation and in the cerebellar granule cells. Rat eag is expressed at relatively low levels in the hippocampal CAl field, the caudate putamen, in the granule cell layer of the olfactory bulb and some neocortical layers. By similar techniques, erg transcripts show a 'wide tissue distribution' and are 'abundant' in heart as well as brain, retina, thymus, and adrenal gland and detectable in skeletal muscle, lung, and cornea mRNA pools (see mRNA distribution, 46-13). 46-01-07: In neurones, integrin-mediated neurite outgrowth in neuroblastoma cells appears to depend on the activation of HERG channels in situ (see Receptor/transducer interactions, 46-49). SH-SY5Y human neuroblastoma cells exposed to substrate-bound laminin or soluble laminin induce an early (5-15 min) activation of the HERG-like 1:lR in these cells (for further details, see Developmental regulation, 46-11). Specific inhibition of heterologously expressed rat eag channels occurs at the onset of maturation in Xenopus oocytes. This inhibition, indicated by a marked reduction in r-eag current

l_e_n_t_ry_4_6

_

amplitude, is mimicked when the mitosis-promoting factor (a complex of cyclin Band p34 (cdc2)) is injected into oocytes. 46-01-08: Drosophila eag (1174 amino acids) and elk (1284 amino acids)

encode relatively large polypeptides in comparison to those encoded by the Shaker family (approx. double their 'typical' length of 500-600 aa, excepting those encoded by Shab, 924 aa and drk, 853 aa). The eag and elk open reading frames are also considerably longer than those 'typical' of cyclic nucleotide-gated channels. Within the h-erg locus, several intron/ exon junctions have been characterized (see Gene organization, 46-20). 46-01-09: Noted similarities of eag/elk/erg family members with CNG and plant inward rectifier family sequences are described in Table 4 under Domain conservation, 46-28. Hydropathy analyses of eag, elk and erg subfamily polypeptides suggest a '81-86 + H5' transmembrane domain arrangement typical of those found in the voltage-gated K+ channel superfamily. The eag/elk/erg family proteins contain a positively charged amphipathic segment S4 and a P-region (ibid.). An N-glycosylation motif NX[S/T] is shared by the eag/elk/erg subfamilies located 12-17 aa upstream of the respective pore regions (see Sequence motifs, 46-24). A cross-family consensus site for phosphorylation by protein kinase C exists 2-4 aa downstream of transmembrane domain S6 (see Protein phosphorylation, 46-32).

46-01-10: A subunit interaction domain designated as NAB (HERG) has been localized at the hydrophilic cytoplasmic N-terminus of HERG (analogous to similar domains in the Shaker subfamily - see entry 48). NAB (HERG) is able to form a tetramer in the absence of the remainder of the HERG protein. Significantly, the HERG Ll1261 mutation associated with one long Q- T syndrome kindred (see Table 2 under Phenotypic expression, 46-14) results in a truncated protein that contains the subunit interaction domain and can explain observed effects on decreased channel expression. The minK protein (entry 54, initially proposed to co-assemble with KvLQTl to conduct native cardiac IK,s, ibid.) has also been shown to enhance functional expression of HERG-K+ currents in CHO cell co-transfection studies. 46-01-11: Injection of either Drosophila or rat eag mRNA into Xenopus oocytes gives rise to voltage-activated, non-inactivating outward K+currents of the 'delayed rectifier' type. In comparison to Drosophila eag, rat eag currents show a marked variation in activation kinetics according to the voltage pulse protocol used to elicit outward current (see Activation, 4633). Differences and similarities between vertebrate eag and erg subfamily channel currents, together with distinctions between heterologously expressed HERG currents, native inwardly rectifying cardiac IK,r and native IK1 are also summarized under Activation, 46-33 and Current-voltage relation, 46-35 and Inactivation, 46-37. The resemblance of h-erg inactivation to IC-type' inactivation (characterized as a 'slow inactivation' mechanism in other K+ channels such as Kv1.3 and involving a 'conformational switch' in the outer mouth of the channel) is described under Inactivation, 46-37.

_1-...

e_n_try_4_6_

46-01-12: Despite marked differences in the characteristics of eag and erg channel currents, initial analysis of the voltage-dependent gating transitions in these subfamilies is characteristic of most channels in the 54containing superfamily of ion channels. A schematic model of HERG channel gating modified from those developed for Sh channels is cited under Kinetic model, 46-38. 46-01-13: Some variation in ionic selectivity characteristics have been reported for heterologously expressed IJrosophila and mammalian eag and erg channel homologues (summarized ~in Table 5 under Selectivity, 46-40). Alignment of sequences corresponding to the pore-forming (P-domain) regions of several K+ -selective channels has revealed some distinguishing features in the eag/elk/erg subfamilies (summarized in Fig. 5, this entry). 46-01-14: A large number of studies have described block of native IK,r and heterologously expressed h-erg channels by class ill antiarrhythmic drugs, that act by slowing cardiac repolarization (phase 3) and prolong the duration of action potentials (see Blockers, 46-43). Agents that have been extensively investigated include E-4031, MK-499, sotalol, dofetilide, clofilium, quinidine, sematilide, L691,121 and WIN 61773-2 (ibid.). [3H]Dofetilide binds with high affinity to sites associated with the guinea-pig cardiac IK,r channel (see Ligands, 46-47). Quinidine and sotalol have been associated with an lacquired' or ldrug-induced' form of long QT syndrome (see Blockers, 46-43). Many other non-selective blockers of IK,r have been described (ibid.). 46-01-15: Rat eag channels expressed within HEK-293 cells have been described as 'rapidly and reversibly inhibited' by rises in [Ca2+]i between 30 and 300nM (mean IC so of 67nM in an I/O patch)4. Generation of intracellular calcium oscillations following muscarinic receptor activation appeared to induce a synchronous inhibition of r-eag-mediated outward current. These and other data led to the conclusion that r-eag channels are 'voltageactivated, calcium-inhibitable' channels, with the Ca2 +-inhibitory effects independent of calcium-dependent kinases and phosphatases4. There is some discussion relating these properties in rat eag to similarities and differences to native M-current (see entry 53 and notes 4, 5 and 6 in Table 6, this entry). 46-01-16: In voltage-clamped oocytes expressing Drosophila eag, bath application of membrane-permeable 8-Br-cAMP (8-bromoadenosine 3'5'cyclic monophosphate, 1 mM) or 8-Br-cGMP (8-bromoguanosine 3'5'-cyclic monophosphate, 1 mM) increase the amplitude of eag outward currents, while thresholds of activation are shifted to more negative potentials. Rapid application of cAMP (2 mM) to inside-out patches produces significant increases in outward current amplitudeS. Similar effects cannot be induced by cGMP in inside-out patches, indicating a direct modulation' of Drosophila eag channels by binding of cAMP to the channel protein. Cyclic nucleotide binding (see entries 21 and 22) has been excluded from a role in Drosophila, mouse or rat eag channel gating although a role for cyclic nucleotide-dependent phosphomodulation appears likely (see Channel modulation, 46-44). Notably, heterologously expressed HERG channel activity is unaffected by cyclic nucleotides (ibid.). I

lL..--e_n_t_ry_4_6

_

46-01-17: Extracellular K+ has a profound effect on increasing the amplitude of HERG current; under conditions of lowered extracellular K+ (i.e. in modest serum hypokalaemia t ) the effects of IK,r blockers tend to be exaggerated, lengthening of cardiac action potentials, and inleading to ~excessive' duction of torsade de pointes. Elevation of serum [K+] in hypokalaemic patients receiving medications which block IK,r or in individuals with the chromosome 7-linked LQT syndrome has been demonstrated to correct abnormalities of repolarization duration, T-wave morphology, QT/RR slope (slope of the relation between QT interval and cycle length), and QT dispersion in patients with chromosome 7-linked LQT syndrome (see Channel modulation, 46-44). Extracellular Mg2+ (at physiological concentrations) 'dramatically slows' activation of r-eag channels expressed in oocytes in a dose- and voltage-dependent manner (ibid.).

Category (sortcode) 46-02-01: VLG K eag/elk/erg, i.e. voltage-gated potassium channels encoded by genes in the subfamilies eag, elk and erg. The eag-related family is a distinct branch of the voltage-gated channel gene superfamily (for clarification, see Fig. 1 under Gene family, 46-05).

Information sorting/retrieval aided by designated gene family nomenclatures 46-02-02: The gene product prefix (used as a 'unique embedded identifier' or VEl) for 'tagging' and retrieving information relevant to this entry on the CSN website will be of the form VEl: eag or elk or erg. If a systematic nomenclature is established for the eag-related gene family, than this may take preference (see this field in other entries for examples).

Channel designation Designations distinguishing eag family member properties 46-03-01: At the time of compilation, no universal nomenclature has been proposed for eag, elk or erg channels. For convenience, properties reported for different species homologues of these subfamilies are designated in this entry by an underlined prefix indicating the species/subfamily combination (e.g. d-eag:, m-eag:, r-eag:, d-elk:, h-erg: etc. (see Table 1 under Gene family, 46-05). Note: Strictly, the capitalized name 'HERG' has been used to specify the gene encoding human erg (ether-a-go-go-related) subfamily channels (human eag-related gene, ibid.); for this reason references to HERG are italicized within this entry.

Current designation 46-04-01: d-eag: Native Drosophila currents 'affected by' mutations in eag are briefly described under Protein interactions, 46-31. h-erg: Evidence suggesting the human eag-related gene (HERG) encodes the channel protein subunit underlying human cardiac rapidly-activating delayed rectifier ~+ current IK,r (designated with the comma between the K and r in these entries) is

_ entry 46

'------------

summarized under Phenotypic expression, 46-14). In native cardiac ventricular preparations, IK,r has also been designated as I-Kr and/or 'E4031-sensitive repolarization current' (e.g. ref. 76, see also Blockers, 46-43). Compare also the distinctive properties of the Islow-activating' delayed rectifier current component in heart designated as IK,s under KvLQT1 in VLC [K] minK, entry 54.

Gene family eag, elk and erg define an extended family of K+ channel genes 46-05-01: The isolation of eDNA clones encoding the Drosophila Eag K+ channel polypeptide (see Isolation probe, 46-12) formed a prototype for an extended gene family, with subsequently isolated members placed in eag, elk and erg subfamilies based on their sequence similarity. Isolation of other clones (or species homologues of existing clones) in the gene family may clarify and extend the present classification. Table 1 therefore lists a partial nomenclature for genes related to Eag. Presently known eag, elk and erg subfamily members share at least 470/0 amino acid identity in their Table 1. Incomplete gene family nomenclature and Inon-systematic' names for ion channel genes/protein subunits showing relatedness to eag. cDNA Iclone' names (isolates) in use are indicated in bold. (From 46-05-01)

Subfamily Drosophila (see note 2) isoforms Eag

eag6 Denoted in this entry as d-eag: (see note1)

Elk

elk1 Denoted in this entry as d-elk: (see note 1)

Erg

Human isoforms

Rat isoforms

Mouse isoforms

rat eag7 m_eag1?8 Denoted in this Denoted in entry as this entry as r-eag: (see m-eag: (see note 1) note 1)

h-erg1 also in capitals, i.e. HERG Denoted in this entry as h-erg: (see noteland Channel designation, 46-03)

Notes: 1. The designation used to indicate species and subfamily data is underlined, which mostly appears at the beginning of paragraphs within this entry. 2. Amino acid sequence alignments for these channels are given in Fig. 2.

entry46

_

I" " ' - - - - - - - - - -

hydrophobic cores with all sequences possessing a segment homologous to a cyclic nucleotide-binding domain (see ref. 1, Domain conservation, 46-28, ILG CAT cAM~ entry 22 and ILG CAT cGM~ entry 23). Additional localized sequence comparisons indicate that members of this family are distantly related to certain 'six transmembrane domain' inwardly rectifying K+ channels of plant origin (e.g. clones KAT1, AKT1 and KST1, see descriptions under Domain conservation, 46-28).

Functional diversity within the eag/elk/erg gene subfamilies 46-05-02: Both Drosophila eag and the mouse species homologue (m-eag) encode voltage-gated, outwardly rectifying K+ channelss"s. In Drosophila, the eag gene encodes a polypeptide homologous to, but distinct from, the Shaker K+ channel subunits 9; electrophysiological studies have revealed allele t -specific effects of eag on four identified K+ currents in Drosophila larval muscles lo (for details, see Phenotypic expression, 46-14 and Protein interactions, 46-31).

An extended gene family tree including the Eag, Elk and Erg channel polypeptides 46-05-03: Estimates of amino acid identities and a dendrogram t representing the gene family relationships between known polypeptide sequences of eag, elk and erg subfamily members appears in Fig. 1 compared to various voltage-gated and cyclic-nucleotide-gated channel families. The dendrogram depicted in Fig. Ib indicates a common ancestral origin for present-day voltage-gated channel superfamily genes (see ref. 11). Note: Sequence alignments of eag, elk and erg subfamily members predicting a similar overall structure including a central hydrophobic core and substantial similarity in a putative cyclic nucleotide-binding domain are shown under Encoding, 46-19 (for descriptions of these features, see Domain conservation, 46-28).

Adaptive radiation of human erg genes 46-05-04: An Erg subfamily cDNA isolated from a hippocampal cDNA library and designated as the human ~ag-!elated gene (h-erg or HERG (see Table 1 and Isolation probe, 46-12) displays -strong K+-inward rectification following heterologous expression in oocytes under certain conditions2 (see Fig. 6 under Activation, 46-33). This characteristic is due to the operation of a novel inactivation mechanism which prevents K+ efflux during depolarization (for details, see Inactivation, 46-37). h-erg does not appear to be the human homologue of Drosophila eag - the latter shares only about 500/0 amino acid identity with h-erg and lacks the inactivation mechanism described above. Note: Isolation of a mammalian elk homologue has been noted in ref. 1; however no reports on the functional expression of elk subfamily proteins were found during compilation of this entry.

Subtype classifications 46-06-01: d-eag: Drosophila eag channels in Xenopus oocytes display properties characteristic of delayed rectifier-type voltage-gated K+ channels

II

100

eag

slo Shab

Shal Shaw

11 18

"42 47 41 100

13 19 39

43 100

14 21

46 100

11 19

100

14

100

68

100

100

11

14

13

11

15

ShB

cGMP

cAMP

AKT1

14

14

14

11

18

24

25

100

KAT1

13

12

11

11

19

24

23

62

100

elk

12

13

11

14

17

20

25

26

24

100

h-erg

13

17

15

15

16

23

24

27

24

49

100

m-eag

12

12

13

16

13

25

27

27

26

47

49

100

eag

10

13

12

versus

Shaw

Shal

ShB

17

slo

Shab

20

24

26

15

KAT1 AKT1 cAMP cGMP

25

elk

49

49

65

m-eag h-erg

(a) Amino acid identities, percent

I

I

I

I

I

I

~

I

I....-

I

n

I (b) Dendrogram

I

Elk

H-Erg

M-Eag

Eag

KAT1

AKT1

cAMP

cGMP

Sio

Shab

ShB

Shal

Shaw

I

I

I~

0\

I~~

("I)

Figure 1. (a) Percentage amino acid identities limited to the hydrophobic core (Sl-S6 domains) in representative members of eag, elk and erg subfamilies are shown in comparison to voltage- and cyclic nucleotide-gated channels. Sequences used for computation of these values included the eag family members eag, m-eag, h-eag and elk 1, the plant inward rectifiers KAT1 and AKT1 (see Domain conservation, 46-28), the rat olfactory cGMP-gated cation channel (CNG-2, see ILG CAT cAM~ entry 21), the bovine retinal cGMP-gated cation channel (CNG-1, see ILG CAT cGM~ entry 22), the Drosophila slo Ca 2+-activated K+ channel (dslo, see ILG K Ca, entry 27) and the voltage-gated K+ channel family members Kv4, Shaw, RK5, Shal, RCK1, ShB, DRK1, Shab, IK8 and K13 (see Database listings, field 53, under the VLG K Kv series, entries 48 to 52). (b) Dendrogram showing relatedness between the representatives of the voltage-gated gene superfamily members described in (a). Horizontal branch lengths are inversely proportional to the similarity between the sequences. Note: Multiple sequence comparisons when limited to the hydrophobic core (Sl-S6 domains) of these channels (as used to calculate the similarities here) indicate that the plant inwardly rectifying K+ channels AKT1 and KAT1 are more closely related to the eag family than they are to the Shaker family (for details, see Domain conservation, 46-28). (All data taken from Warmke and Ganetzky (1994) Proc Natl Acad Sci USA 91: 3438-42 using sequence comparison algorithms defined therein.) (From 46-05-03)

II

~

M"

~ ~

0'\

_L--

e_n_try_4_6_1

(see Voltage sensitivity, 46-42). These channels also show some properties reminiscent of some intracellular ligand-gated channels in that current amplitudes are 'directly' modulated by intracellular cyclic AMP (see Channel modulation, 46-44, JLG key facts, entry 14 and JLG CAT cAM~ entry 21). 46-06-02: h-erg: The biophysical properties of human erg subfamily channels expressed in Xenopus oocytes are very similar to the rapidly activating delayed rectifier current IK,r in cardiac myocytes 12. The unique properties of erg subfamily channels, particularly those mechanisms that can give rise to inwardly rectifying K+ currents, are described in the ELECTROPHYSIOLOGY section (fields 33 to 42).

Trivial names Retention of trivial name 46-07-01: d-eag: Drosophila ether-a-go-go originally referred to behavioural phenotypes of eag mutants shown under ether anaesthesia; the eag designation has been retained for mammalian species homologues (see also Gene family, 46-05, Channel designation, 46-03 and Phenotypic expression 46-14).

EXPRESSION

Cell-type expression index See Isolation probe, 46-12 and mRNA distribution, 46-13.

Studies characterizing IK,r in native cells 46-08-01: h-erg: The rapidly activating 'delayed rectifier' current IK,r (thought to be dependent on channel protein subunits encoded by HERG 12 ) has been characterized in various cardiac preparations including ventricular myocytes of guinea-pig13 and cat14, cultured atrial myocytes (cell line AT1)15 and in nodal cells of rabbit heart16 (for further details on IK,r, see Phenotypic expression, 46-14).

Cloning resource Stable HERG-like channel currents in cardiac and neurone-derived cell lines 46-10-01: cDNA and genomic library resources used for isolation of presently known eagfelkferg sequences are referred to under Isolation probe, 46-12 and references under Database listings, 46-53.. Note that atrial tumour myocytes derived from transgenic mice (AT-l cells) possess a current with similar characteristics to native cardiac IK,r 17. A constitutively expressed inwardly rectifying current described as 'HERG-type' has been extensively characterized in a rat dorsal root ganglion (DRG) x mouse neuroblastoma hybrid cell line (F-ll) (see also Blockers, 46-43).

1'--_e_n_t_ry_4_6

_

Developmental regulation Mitosis-promoting factor-mediated suppression of rat eag current in oocytes 46-11-01: Specific inhibition of heterologously expressed rat eag channels occurs at the onset of maturation in Xenopus oocytes. This inhibition, indicated by a marked reduction in r-eag current amplitude, is mimicked when the mitosis-promoting factor (a complex of cyclin Band p34 (cdc2)) is injected into oocytes 18 .

Integrin-mediated activation of HERG-like current associated with neurite extension 46-11-02: SH-SY5Y human neuroblastoma cells exposed to substrate-bound laminin or soluble laminin induce an early (5-15 min) activation of the HERG-like hR in these cells 19. In cells adherent to substrate-bound laminin, 1IR potentiation can persist for 90-120min accompanied by a similar, but transient, increase in cell leak conductance (GL ), shifting the resting potential from -12 to nearly -30 m V over approx. 120 min. In cells adherent to polylysine, exposure to soluble laminin induces hyperpolarization through a transient potentiation in 1m, without any effect on leak conductance 19. Notably, these effects of substrate-bound and soluble laminin can be mimicked by antibodies raised against the integrin {31 subunit (i.e. culture-dish substrate-bound antibodies simultaneously activate the hR and GL , whereas in soluble form they only activate 1m). Significantly, cells adherent to substrate-bound laminin undergo neurite extension, which can be inhibited by E-4031 and Cs+ (see Blockers, 46-43). Notes: 1. Integrin-independent substrates such as polylysine or bovine serum albumin have no effect on these cells. 2. Earlier studies showed activation of a HERG-like 1IR upon integrin-mediated adhesion to fibronectin or vitronectin in a subclone of a murine neuroblastoma cellline2o.

Isolation probe De novo sequence cloning of Drosophila eag 46-12-01: d-eag: Chromosome walkingt and chromosome jumping t techniques were originally used to identify cDNAs that corresponded to the eag transcript21 . These cDNAs (spanning >35 kb at the genomic level) included the mutant sequences associated with four different eag mutant alleles t (see Phenotypic expression, 46-14). A hybridization t probe based on the hydrophobic core sequences of Drosophila eag was subsequently used to retrieve further cDNA clones encoding eag and those encoding elk, the eag-like K+ channel from Drosophila head-specific cDNA libraries 1 .

Probes used to retrieve mammalian eag/erg subfamily homologues 46-12-02: m-eag/h-erg/r-eag: Degeneratet primerst based on the Drosophila eag peptide sequences ACIWY, TYCDL, ILGKGD were initially used to PCRamplify sequences at a low annealing temperature (42°C) using randomprimed cDNA from mouse skeletal muscle as template; products corresponding to putative eag homologues were identified by cross-hybridization to a

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_6--'

Drosophila elk probe by low-stringency cross-hybridization. Fragments within the range 750-950 bp were recovered from agarose gels and were reamplified with sense and antisense degenerate primers based on ACIWY and ILGKGD (see above), subcloned and sequenced. Low-stringency hybridization using Drosophila elk-hybridizing fragments did not result in retrieval of elk homologues from a mouse brain cDNA libraryl. Additional PCR analysis and hybridization screening of the same library isolated a sequence representing mouse eag (m-eag)l, which was subsequently used to probe a human hippocampal cDNA library to isolate a related sequence defining a third branch of the eag family, designated human eag-related gene (h-erg) (see Gene family, 46-05). The extreme 5' end of the h-erg sequence reported in ref. 1 was deduced from a composite of genomic and cDNA sequences (see 'Incomplete' cDNA sequences under Resource D - Diagnostic tests). In separate studies, a Drosophila eag cDNA probe (nt 1155-2329) was used in low-stringency t hybridizations to isolate a rat genomic library clone encompassing the predicted 5' and 3' terminal sequences of rat eag (r-eag)7. PCR screening of a rat cerebellum cDNA libraries retrieved two overlapping cDNA clones, yielding a combined cDNA of 4.6 kb and predicting an open reading frame of 962aa 7. Rat eag sequences were also amplified from hippocampal cDNA libraries.

mRNA distribution High-level expression of rat eag mRNA in brain 46-13-01: r-eag: Rat eag mRNA has been reported to be 'predominantly expressed'inthe central nervous system 7 with no apparent signal on Northern t blots of poly(A)+ RNA derived from rat heart, spleen, lung, liver, skeletal muscle, kidney and testes. In situ hybridization using 47-mer (nt 1738-1691) and 49-mer (nt 3126-3077) oligonucleotide probes 7 show that rat eag mRNA is most prominently expressed in the hippocampal formation and cerebellum. The highest levels of eag transcripts were detected in the granular cells of the dentate gyms, in the CA3 pyramidal cells of the hippocampal formation and in the cerebellar granule cells. Rat eag is expressed at relatively lower levels in the hippocampal CAl field, the caudate putamen, in the granule cell layer of the olfa.:tory bulb and some neocortical layers7. tHigh abundance and wide distribution' of h-erg transcripts 46-13-02: h-erg: In Northern blot analyses, a PCR probe (HERG nt 679-2239) shows strongest hybridization to human heart mRNAs (two bands, see Transcript size, 46-17), with weak signals in brain, liver and pancreas3 . RNAase protection assays using eDNA fragment probes of erg homologues from guinea-pig, rabbit, human, dog and rat have confirmed that erg message is 'expressed uniformly' throughout the heart of all five species22 . By these techniques, erg transcripts show a 'wide tissue distribution' and are 'abundant' in heart (estimated at '50% more abundant' than Kv4.3 message - see entry 51) as well as brain" retina, thymus, and adrenal gland and detectable in skeletal muscle, lung, and cornea mRNA pools22.

l_e_n_t_ry_4_6

------'_

Phenotypic expression Background - ether-a-go-go (eag) mutants exhibit thyperexcitability' phenotypes 46-14-01: d-eag: Drosophila carrying the eag mutation display an abnormal legshaking phenotype23 . These effects were subsequently linked to increases in neuronal excitability and neurotransmitter release at the neuromuscular junction consistent wth defects in K+ channel conductances24~25. The mutations designated eag 1, eag4PM , eagX6 and eag24 reduce different currents in larval muscles, where they interact synergistically with Shaker mutations (see VLG K Kv1-Shak, entry 48) to enhance the behavioural and physiological phenotypes. Double mutants such as eaglSh S , eag 1Sh 120 , eag4PMSh120, eag24 Sh s and eag24 Sh 120 have a Ivigorous' leg-shaking phenotype under anaesthesia, sometimes coupled with a 'wings-down' phenotype26~27. These mutants are also characterized by high-frequency, spontaneous excitatory junction potentials (EJPs) of high amplitude at neuromuscular junctions24~28, with spontaneous discharges of EJPs or action potentials in adult flight muscles27. Transmitter release persists for an order of magnitude longer in eag Sh double mutants than in either single mutant, resulting in the characteristic 'plateau-shaped' synaptic potentials and long trains of action potentials in motor axons. For the types of current reduced in eag mutants, see Protein interactions, 46-31. Notes: 1. Roles for cAMP-modulation of eag-type channels in controlling synaptic efficiency in the central and peripheral nervous systems have been discussed5 (see Channel modulation, 46-44). 2. The possible role of calcium/calmodulin-dependent protein kinase II (CaMKII) in influencing synaptic plasticity phenotypes by functional modulation of Drosophila Eag!9 are summarized under Protein phosphorylation, 46-32. 3. The resemblance between the Drosophila EAG current and the mammalian M-current (entry 53) is briefly discussed under [Ca2+h ions under Blockers, 46-43.

HERG encodes the potassium channel underlying IK,r in heart 46-14-02: h-erg: Following its heterologous expression in Xenopus oocytes, the biophysical properties of the human eag-related gene product (HERG) have been reported12 as 'nearly identical' to the rapidly activating 'delayed rectifier' K+ current (IK,r) described in native in cardiac myocytes13- 16. IK,r is known to have an important role in the initiation of repolarization during cardiac action potentials (see also Phenotypic expression under INR K native, 32-14). Despite its name, the native cardiac delayed rectifier contributes exhibits profound inward rectification (see ref.3D and Tseng, 1995, under Related sources and reviews, 46-56). These similarities between HERG-induced current and IK,r extend to their strong inward rectification t properties due to operation of a novel inactivation mechanism limiting K+ efflux during depolarization (see Inactivation, 4637), activation by extracellular K+ (see Channel modulation, 46-44), blockage by lanthanum and cobalt ions (see Blockers, 46-43). Comparative note: The inward rectification displayed by IK,r is distinguishable from that induced by the'classical' I K1 inward rectifier, which contributes to the final phase of ventricular repolarization (for clarification, see Activation, 46-33 and associated figures under Phenotypic expression of INR K [native], 32-14).

II

_'--

e_n_t_ry_46_ _1

Mutations in HERG are associated with the chromosome 7-1inked form of long QT syndrome 46-14-03: h-erg: Many studies aimed at identifying the molecular bases of human cardiac arrhythmias t have focused on the incidence of long QT (LQT) syndrome t, a relatively rare disorder that is characterized by prolongation of the QT interval f on electrocardiograms, blackouts (syncopall attacks), seizures t and sudden death (typically following a ventricular arrhythmia, torsade de pointes t - see also next paragraph). Diagnostic criteria for the various forms of LQT syndrome have been summarized31 ; these make important distinctions between congenital (familial) and sporadic t forms of the disorder (see also the reviews/commentaries in refs32-34). Gene markers t which segregate t with various forms of the disorder (within LQT pedigrees) have been mapped to chromosomes Ilp15.5 (LQT1), 7q35-36 (LQT2) and 3p21-24 (LQT3), (for details, see Chromosomal location, 46-18). Subsequent gene linkage and physical mapping analyses co-localized LQT2 and HERC to chromosome locus 7q35-36 (ibid. 3). Single-strand conformation polymorphism t (SSCP) and direct sequence analyses in LQT2 families have revealed. a spectrum of mutations in HERC, including two intragenic deletions, one splice-donor mutation and three missense mutations (for details, see Table 2). Strong expression of the HERC mRNA transcript in human heart (see mRNA distribution, 46-13) and the the loss of function in mutant channels (some by a dominant negative t effect that varies in severity), have led to the proposal that HERC is LQT2 and its expression provides a cellular basis for torsade de pointes3,35 (see paragraph 46-14-13). Comparative note: In separate studies, LQT3 has been linked to a three amino acid deletion in the seNSA gene encoding the human heart voltage-gated sodium channel Q subunit hHI (~KPQ, affecting the cytoplasmic linker between domains m' and IV). Heterologously expressed ~KPQ mutant channels show a small sustained inward current compared to the wild type, reflecting a defect in c.hannel inactivation36 (summarized in Phenotypic expression, Chromosomal location and Inactivation under VLC Na, 55-14, 55-18 and 55-37 respectively).

-

Delayed repolarization predisposing to life-threatening arrhythmias 46-14-13: h-erg: Mechanistic links between inherited forms of LQTs and the (more common) acquired forms of LQTs have been proposed by Sanguinetti et a1. 12 (see also Blockers, 46-43 and Channel modulation, 46-44). Direct support for hypotheses invoking ion channel mutation to explain certain forms of congenital LQTs comes from the equivalence of LQT2/HERG and LQT3/SCN5A (see paragraph 46-14-03, Tables 2 and 3). This 'myocellular hypothesis' is also indirectly supported by QT prolongation induced following pharmacological block of K+ channels in human and animal models (see ref.41 and Blockers, 46-43). Delays in myocyte repolarization (indicated by prolonged QT intervals) are known to induce reactivation of cardiac L-type Ca2+ channels (see VLC (~a, entry 42) resulting in secondary depolarizations42,43. In turn, these events have been suggested as a likely cellular mechanism for torsade de pointes, a sinusoidal twisting of the QRS axis around the isoelectric line on electrocardiograms44- 46 • Torsade de pointes can degenerate into ventricular fibrillation t prior to sudden death.

entry46

I-

_

- - - - - -

Table 2. Summary of gene rearrangements and point mutations at the HERG locus associated with the incidence of long QT syndrome3 . For genetic linkages to the incidence of different LQT phenotypes, see Chromosome location, 46-18 (From 46-14-03) Rearrangement or mutation

Description, incidence and functional implication (where known)

HERG/LQT2 intragenic deletions

46-14-04: h-erg: A panel of PCR primers designed to the published eDNA sequence of HERG (ref. l , see Database listings, 46-53) were used to identify three intronic sequences within the protein-coding region (see Gene organization, 46-20). sscpt analyses using selected primers identified an A to G polymorphism t within HERG (nt 1692 on the eDNA) and analysis of one pedigree (kindred 2287 or K2287) produced results consistent with a null allele t; in repeat experiments with flanking PCR primers, unaffected members of K2287 (and >200 other unaffected individuals) showed a single product of 170 bpi in LQTs-affected individuals of K2287, two products of 170 and 143 bp were detected. This feature indicated the presence of a 27bp deletion beginning at HERG nt position 1498 (M500-F508, predicted to disrupt the S3 putative membrane-spanning

~I500-F508

domain as indicated on the [PDTMj, Fig. 4).

~1261

46-14-05: h-erg: SSCP analysis in a different family (K2595) identified an aberrant conformert in LQTsaffected individuals which was not present in >200 unaffected individuals; sequencing of normal and aberrant conformers revealed a single base deletion at nt position 1261 (~1261, predicted to introduce a frameshift

mutation resulting in the introduction of a stop codon within 12 aa residues, as indicated on the [PDTMj, Fig. 4; this deletion has been determined to affect HERG intersubunit association and expression level (see ref. 37

under Protein interactions, 46-31). HERG/LQT2 point mutations N470D

A561V

46-14-06: h-erg: In K2596 an A to G substitution was identified at the eDNA nt position 1408 (N470D, i.e.

substitution of an aspartate residue for a conserved asparagine at codon 470, predicted to alter the S2 putative membrane-spanning domain as indicated on the [PDTMj, Fig. 4). 46-14-07: h-erg: Three aberrant SSCP conformers were identified in LQT-affected members of K1956, K2596 and K2015. Cloning and sequencing of normal and aberrant conformers from K1956 identified a C to T substitution at position 1682 (A561\l; i.e. substitution of valine for a highly conserved valine at codon 561, predicted to alter

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_6_

Table 2. Continued Rearrangement or mutation

Description, incidence and functional implication (where known)

the S5 putative membrane-spanning domain as indicated on the {PDTM}, Fig. 4). A561T

46-14-08: A missense G-to-A transition at position 1681 resulting in Ala561Thr in domain 85 is associated with a prolonged T wave of low amplitude on the surface ECG. Specifically, a distinctive biphasic T-wave pattern in the left precordial leads of all affected subjects has been associated with patients expressing the Ala561Thr mutation (regardless of age, gender and beta-blocking therapy)38.

I593R

46-14-09: A single nucleotide substitution of thymidine to guanine (TI961G, Ile593Arg) in the channel pore region associated with LQT syndrome has been reported39.

splice site error

46-14-10: h-erg: In K2015, a G to C substitution was identified, predicted to disrupt the splice-donor sequence of intron III (see Gene organization, 46-20) and in consequence, affecting the cyclic nucleotide-binding domain. None of the aberrant conformers found in LQTaffected individuals were found in >200 unaffected individuals. In a separate study, a G to A substitution producing Va1822Met: in the cyclic nucleotide-binding domain of HERG in LQT syndrome has also been characterized40 .

V822M

HERG/LQT2 de novo mutation (sporadic case; one kindred)

46-14-11: h-erg: An aberrant SSCP conformer present in one individual of K2269 (which was not identified in either parent or in >200 unaffected individuals) revealed a G-to-A substitution at nt 1882 (G628S, i. e. substitution of

a glycine residue for a highly conserved serine at codon 628, predicted to alter the pore-forming domain as

indicated on the {PDTM}, Fig. 4). The identification of this de novo mutation provided further evidence for the equivalence of HERG and LQT2. Predicted neuronal phenotypes

46-14-12: ComparativB note: HERG is also expressed in brain (see Isolation probe, 46-12) implying that mutant HERG channels may exert some (presently undefined) effect on neuronal function in LQT2-affected individuals. Notably, none of the presently identified LQT families show signs of congenital neural hearing loss (a finding associated with the rare autosomal recessive form of LQT) or other phenotypic abnormalities3. The spectrum of presently known mutations in HERG (this table, see also {PDTM}, Fig. 4) nlay alter channel function in different ways.

e_n_t_ry_4_6

1_ _

_

Implications for tgene-specific' therapy of LQT syndrome 46-14-14: Initial studies showing differential responses of patients taking into account known LQT2 and LQT3 gene defects 47 have indicated that (i) LQT3 patients may be more likely to benefit from Na+ channel blockers and from cardiac pacing due to a higher risk of arrhythmia at slow heart rates; (ii) LQT2 patients appear to be at higher risk of developing syncope under stressful conditions because of the combined arrhythmogenic effect of catecholamines with the insufficient adaptation of Q-T interval when heart rate increases 47.

Earlier hypotheses for origin of LQT phenotypes: tautonomic imbalance' 46-14-15: Comparative note only: A second hypothesis for LQT phenotypes suggests that a predominance of left autonomic innervation causes abnormal cardiac repolarization and arrhythmias (see Schwartz et al., 1995, under Related sources and reviews, 46-56). This hypothesis is supported by (i) induction of arrhythmias in canine models following right stellate ganglionectomy and (ii) clinical observations which suggest that certain LQT patients respond favourably to ,B-adrenergic blockade and left stellate ganglionectomy procedures (ibid.). For further mechanisms linking altered autonomic activity and arrythmias in LQT, see Receptor/transducer interactions, 46-49 and Channel modulation, 46-44.

Transcript size Band doublets revealed by HERG probes may indicate further transcript diversity 46-17-01: h-erg: The relatively strong and selective expression of HERG in human heart poly(A)+ mRNA is detected as two hybridizing bands of 1"'V4.1 and 1"'V4.4kb by Northern t assays using a HERG peR probe (nt 679-2239)3. The size of either hybridizing band is consistent with the predicted size from the largest open reading frame of the HERG eDNA (see Encoding, 4619), but the different sizes may represent alternatively spliced t transcripts or the co-expression of closely related sequences3 .

SEQUENCE ANALYSES Note: The {PDTM} symbol denotes an illustrated feature on the channel protein domain topography model (Fig. 4).

Chromosomal location Mutations in HERG are associated with chromosome 7-linked forms of long QT syndrome 46-18-01: h-erg: The gene encoding the human ether-a-go-go-related gene (HERG) has been localized to human chromosome 7q35-36 by linkage analysist and physical mapping t3 . This finding placed HERG at the same position as LQT2, one of three genes associated with autosomal dominantt forms of the long QT (LQT) syndrome (see Phenotypic expression, 46-14). As summarized in Table 3 extensive genetic analysis in many LQTs

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_6_1

Table 3. Comparative summary of genetic linkages to forms of inherited long QT syndrome (From 46-18-01) LQT form/gene

Chromosome localization and gene product where known

LQTl (=KVLQTl, comparative note only, see also VLG K minK, 54-14 and 54-31)

46-18-02: The dominant t LQT1 gene has been linked by positional cloningt to human chromosome Ilp15.5 and encodes a protein (KVLQTl) with the structural features of a voltage-gated K+ channel with approx. 30% sequence identity to Drosophila Shaker48 (for summary, see Phenotypic expression under VLG K Kv1-Shak, 48-14). Earlier studies reported a 'tight' linkage to a polymorphism t at the Harvey-ras gene in several LQT families at IlpI5.5 49,5o. H-ras was excluded as a candidate for LQTl by linkage51 and sequence analyses (cited as unpublished in ref.3). The primary sequence of KVLQTl (above) and linkage analysis 52 have also excluded mutations in other K+ channel genes on IIp as LQT1 candidates (including the KCNA4 gene encoding hKvl.4 and the KCNC1 product encoding hKv3.1 - for refs, see fields 48-18 and 50-18). To March 1995, seven autosonlal dominant LQT families have been linked to the Ilp15.5 marker3,49,5o,53 and mutations in the LQTl gene are the most common cause of inherited LQT (>500/0 of known incidences). For updates, see OMIM Entry: 192500 IVentricular Fibrillation With Prolonged QT Interval [Ward-Romano Syndrome; WRS; Long QT Syndrome, Type 1; LQT; LQT1}'.

LQT2 (=HERG, this entry)

46-18-03: The dominant LQT2 gene has been localized to human chromosome 7q35-36. Identification of (i) new LQT families all linking to the 7q35-36 marker; (ii) mapping of HERG to the same locus; (iii) demonstration of strong expression of HERG in human heart (see mRNA distribution, 46-13) and (i'1~ characterization of six mutations in HERG within LQT subjects have provided strong evidence for the equivalence of LQT2 and HERG3. A summary of HERG rearrangements and point mutations reported in this study is given in Table 2 under Phenotypic expression, 46-14. To March 1995, 14 autosomal dominant LQT families have been linked to the 7q35-36 marker3,49,50,53. For updates, see OMIM Entry: 152427 ILong QT Syndrome, Type 2 {LQT2;HERGJ'. 46-18-04: Comparative note: Other genes co-localizing

to 7q35-36, including C1CN1, encoding a skeletal muscle chloride channel and CHRM2, encoding cardiac muscarinic M2 receptors were excluded as LQT2

entry46

_

I' - - - - - - - - - Table 3. Continued LQT form/gene

Chromosome localization and gene product where known candidates following linkage analysis (Wang et a1., cited in ref. 3 ).

LQT3 (== SeNSA; comparative note only - see VLG Na, entry 55)

46-18-05: The dominant LQT3 gene has been linked by positional cloning to human chromosome 3p21-24. The incidence of LQT3 has been associated with a three amino acid deletion in the SCN5A gene encoding the human heart type V voltage-gated sodium channel Q subunit hHI (LlKPQ, affecting the cytoplasmic linker between domains ITI and IV). Heterologously expressed LlKPQ mutant channels show a small sustained inward current compared to the wild type, reflecting a defect in channel inactivation36 (for further details see Phenotypic expression, Chromosomal location and Inactivation under the entry VLC Na, 55-14, 55-18 and 55-37 respectively). To March 1995, three autosomal dominant LQT families have been linked to the 3p21-24 marker3 ,49,5o,53. For updates, see OMIM Entry: 600163 'Long QT Syndrome, Type 3 [LQT3; SCN5A)'.

Other forms, including LQT4

46-18-06: Notes: 1. There is presently little information

on genetic linkage associated with the recessive form of LQT known as Jervell-Lange-Nielson (JLN) syndrome (see OMIM 220400). 2. Significantly, mutation of the HERC/LQT2 gene has been associated with one case of sporadic LQT (the commonest form, and by definition not heritable) (see Table 2 and the [PDTM), Fig. 4). 3. Hypothetically, distinct forms of LQT may be associated with mutations in genes encoding ion channels, carriers or pumps contributing to cardiac action potential shape/ duration and/or its modulation in situ. (see also Channel modulation, 46-44). To March 1995, three autosomal dominant LQT families remain unlinked to LQT1, LQT2 or LQT3 3 ,49,50,53. Schott et a1. 54 used linkage analysis to map the LQT4 locus in a 65-member family in which the LQT syndrome was associated with more marked sinus bradycardia than usual, leading to sinus node dysfunction. Positive linkage was obtained for markers located on 4q25-q27; maximum lod score t == 7.05 for marker D4S402. For updates, see OMIM Entry: 600919 'Long QT Syndrome, Type 4 [LQT4; Long QT Syndrome with Sinus Bradycardia}'.

Note: For general features of the long QT syndrome, see Phenotypic expression, 46-14 and references to OMIM (Online Mendelian Inheritance in Man) in table.

II

EAg II-EAg H-Erg Elk EAg II-EAg H-Erg Elk EAg II-EAg H-Erg Elk EAg II-EAg H-Erg Elk

108 10. 102 102

~~~~~~~~~~~~~~~~~~~~~~:HI

. . . . . . . . . . . . . . . GSCFLC~DV . . . . . . . . . . . . . . . O~~C~DI

VK

EiQ~;E~;~:~;~::::::::::::::::::::::::::::::::::::::::: E

GA RD

III

ILM~EVVIIEKDMVGSPAHDTMHRGPPTSWLAPGRAKTFRLKLPALLALTARESSVRSGGAGGA~APGAVVVD SHKQIEHT~IILEIIMVMEECD

: : : : : : : : ~ : : :Ii· GL::~::ffi:::~::

VDLTPAAPSSES · . . . . . . . . . . .

L

~

~

~

~

:8

: : : : : : : :: : : : : : : : : : : : : : : : : : :: : : : : : : : : : : : : : EVTAIIDMHVAGLGPAEERRALVGPGSPPRSAPGQLPSPRAHSLMPDASGSSCSLART G L P o(I;Jo G P A A S D G D T E. . . . . . . . . . . . . . . . . . . . . . . AGE G MN L D V P A G C MMG R R

~~~~~~~~~~~~~~~~~:~~~~~:~~~

.

185 162 307 210

S :: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

SR SCASVRRASSADDIEAMRAGVLPPPPRHASTGA SR..................................

- - - - - - - $1--I WD V~C

PT/ilQS. MrLlAH MM~SA~~

· iFSAHLPTLKD . . . . . . . . . . . . . . . . . . . . . . . GVLQ LAPSVQKGEM

168 U6 197 169

:~::~~~~~~@~:::::::

SVFALTAALLGARFRAGSNA~M

. M~Q~RQEA PKTPPHJilLmYCA~ Q KQEAPKTPPH~I HYCV

VH~HS.~AEVL~GS~.L

~:~:~~~i::~~~~~~~~~~~~~~~~~~~~~:~~:~:~:~~;:H~~~AP;

~ K~g:~

KT

I~:~m~~HY~~

WD

KOI:D

II I

H5 226 416 276

I

t t

EAg II-EAg H-Erg Elk

333 3H 513 363

EAg II-EAg H-Erg Elk

.22 .32 598 .51

EAg II-Eag H-Erg "Elk

532 5.2 703 5U

Ea; II-Eag H-Erg Elk

812 657

Eag M-Eag H-Erg Elk

639

6t9

· KDSAVGiAAM RA LTYCD LH ~K~~~ . . . . . . . . . . . . . . . KEATLA SCAMVRALTYCDLHVIK~LQKVL

EVL

ViTSNGQ"TiTTNSiGQ~~~~K~:~~~KA~:~~D~HC~:M~G~

:~~

F~S

F

~~NS~ARM~'

T~:HS

SRN~

~ :E~~~~

T mRH . vm .. I " LRK

::~~~HD~~~

RL I FRRVA DVK~:KfEJLA RIVFRKISDVKIiJ:~ERMK

ER

R/ilN!SP

RI!JN~A

~:~:YE:~~=;~~:~~~~~;~~~~~~N~i:~~~~~~~~~M~

717 727 901 767

Eag M-Eag H-Erg Elk

QLPQNQDHL~~IFSKFRR"PQVQAGSKELVGGSGQSDVEKGDGEVER~V

Eag M-Eag H-Erg Elk

SRVv GWAR

Eag M-Eag H-Erg Elk

. . . . . . . . LA Q R D~V A T vfL1t> MKV DV R L E L QR MQ Q R. . . . . . . . . . . . . . I G R lED LfLl 100 J1A inward currents in high K Ringer's (not shown)

V(mV) ·150 -125 -100

-75

-50

. . . ••. .

-25

o·

l2

25

:.1·· · · .'

50

-2

-4 -6

IblA)

-8 -10

·12 -14

-80 -Jiiiiiiiiiii-.:1Q

Voltage protocol: Depolarizing, as shown Vertical bar: 1 J.1A; horizontal bar: 100 ms

Filled squares: HERG inward current Filled circles: water-injected control

Figure 6. Comparison of mouse eag and human erg channel currents following heterologous expression in Xenopus oocytes. Under two-microelectrode voltage-clamp, oocytes were bathed in either lnormal' Ringer's solution or lhigh-K+' Ringer's solution with composition as indicated. Voltage-recording and current-passing electrodes (tip resistance 0.5-1 MO) were filled with 2 M KCl throughout. (a) Large outward current mediated by m-eag following a series of depolarizing voltage steps (from -80mV,

l_e_n_t_ry_46

----'_

stepping to +60mV as shown 2 ). (b) Large m-eag current in high K+ Ringer's. (c) Small inward current mediated by HERG under the same voltage protocol as in (b). This small current was not observed in water-injected oocyte controls under the same voltage protocol. (d) Large inward currents mediated by HERG in elevated [K+]o following hyperpolarizing steps with a depolarizing pre-pulse (stepping to -130mV with a +60mV pre-pulse for 200ms as shown. Note the rapid activation of inward current (e.g. for hyperpolarization to -105 m V, time-to-peak was 27 ± 4.6 ms; mean ± SD; n == 7). (e) In tnormal' Ringer's, when HERG expression is increased by t10-fold or more' over levels used to study the inward component, an temerging outward component', exhibiting kinetics consistent with native cardiac IK,r appears. (I) Current-voltage relation from data in (d) illustrating strong inward rectification due to attenuation of K+ efflux by channel inactivation following depolarization. For further details see ref. 2 , Phenotypic expression, 46-14; Inactivation, 46-37; Kinetic model, 46-38; Blockers, 46-43 and Openers, 46-48. (Data compiled from different experiments reported in Trudeau et al. (1995) Science 269: 92-5.) (From 46-33-02)

slow components, with time constants of lS.2ms and 470ms at a test potential of +40mV. The fast component represents approx 200/0 of total current amplitude. In direct comparison, d-eag and m-eag are both voltage dependent and outwardly rectifying, but m-eag currents are generally sustained for the duration of an activating voltage command 77. See also Rundown, 46-39 and Table 5 under Selectivity, 46-40.

Rapid channel inactivation reduces conductance at positive voltages 46-37-02: h-erg: Unlike other members of the Eag family of voltage-gated,

outwardly rectifying potassium channels, the channel encoded by the human eag-related gene (HERG) displays an inwardly rectifying potassium channel under specified conditions 2 (see Activation, 46-33). HERG channels display typical gating properties of eag-related and other outwardly rectifying, S4-containing potassium channels, but with the addition of a voltage-dependent inactivation process that attenuates potassium efflux during depolarization. This feature of HERG channel function appears critical to the maintenance of normal cardiac rhythmicity (i.e. in the physiological suppression of arrhythmias, specifically extra premature afterbeats)78 (see Phenotypic expression, 46-14).

Inactivating gating mechanisms 46-37-03: h-erg: Ie-type' inactivation, considered to be a 'slow inactivation'

mechanism in other K+ channels such as Kvl.3, appears to involve a 'conformational switch' in the outer mouth of a channel (see Inactivation under VLG K Kv1-Shak, 48-37). N-type inactivation (or any mechanism relying on an intracellular pore blocker) was 'ruled out' by studies on HERG channels using (i) intracellular TEA+ ion (which inhibits N-type inactivationt) and (ii) N-terminal deletions (HERG ~2-37379, but see notes 1 and 2 below). Since extracellular TEA+ did interfere with inactivation, some resemblance to

II

_'---

.

en_t_ry_4_6-----1

features of IC-type' inactivation was postulated for h_erg78~79. Measurements of the instantaneous t current-voltage relationship for h-erg channels (in the saponin-permeabilizedt cut-open oocyte clamp t preparation)8o have determined the rate of inactivation to be strongly voltage-dependent at depolarized potentials (i.e. in contrast to C-type inactivation in other voltage-gated K+ channels). The voltage-dependence could be modulated independently of activation by increasing [K+]o from 2 mM to 98 mM, suggesting that inactivation of h-erg has its own intrinsic voltage sensor80. Notes: 1. Ntype processes would compete with TEA+ binding to the intracellular mouth and hence predict a slower inactivation in the presence of TEA+, which is not observed. 2. Another study81 has reported that truncation of the Nterminal region of HERG did shift the voltage dependence of activation and inactivation by +20 to +30mV, with the rate of deactivation of the truncated channel being 'much faster' than that of wild-type HERG81 .

Kinetic model Gating models developed for other voltage-gated channels applied to HERG 46-38-01: h-erg: Despite marked differences in the characteristics of m-eag and h-erg channel currents (e.g. see Fig. 6 under Activation, 46-33) initial analysis of the voltage-dependent gating transitions in these subfamilies2 is characteristic of most channels in the 54-containing superfamily of ion channels. A schematic model of HERG c.hannel gating modified from those developed for Sh channels has been represented as follows: Cl f==! C2 f==! C n

f=::::!

0

f==!

I(B)

According to the model2, HERG channels are 'at rest' in the closed state (Cd. Following depolarization HERG channels undergo voltage-dependent transitions leading sequentially to the open state (0) and to the inactivated or blocked state, I(B). Since only very small outward currents pass through HERG channels at the onset of depolarization (see Fig. 6), the forward rate for the 0 ~ I(B) transition must be very fast 2. During (subsequent) hyperpolarization, the inactivation is removed and channels enter the 0 state, resulting in inward current flow. The HERG current is transient (see Fig. 6d) as channels make the 0 ~ C transition, albeit at a slower rate than m-eag channels (see panels A and B in Fig. 6). HERG rectification properties resulting from rapid inactivation (particularly in its comparison with native cardiac IK,r) are further analysed in ref. 12. Note: The 'HERGtype' hR described in the rat dorsal root ganglion (DRG) x mouse neuroblastoma hybrid cell line (F-ll) and native cardiac IK,r have been compared by means of a unique kinetic model 74 .

Kinetic modelling of external divalent modulation processes 46-38-02: r-eag: A kinetic model for rat eag activation has been developed from data indicating that all four r-eag channel subunits undergo extracellular Mg2+ -dependent conformational transitions prior to final channel activation82 (see Channel modulation, 46-44).

l'---e_n_t_ry_4_6

--'_

Rundown Bag channel rundown can be reversed by 'patch cramming' 46-39-01: As part of the comparative study of Drosophila versus mouse eag K+ current properties expressed in oocytes 77 (see Table 5 under Selectivity, 46-40) d-eag currents 'rundown' more rapidly than do m-eag currents in

excised macropatchest. Rundown is generally reversible by inserting the patch into the interior of the oocyte, a feature generally taken to indicate that an undetermined 'cytosolic factor' regulates channel activity or stability. Note: Cyclic nucleotide binding (see entries 21 and 22) has been excluded from a role in d-eag or m-eag channel gating 77 although a role for cyclic nucleotide-dependent phosphomodulation appears likely (see also Gene family, 46-05 and Channel modulation, 46-44).

Selectivity Selectivity comparisons between eag species homologues 46-40-01: A summary of selectivity characteristics for heterologously expressed Drosophila and mammalian eag and erg channel homologues (compiled from refs. s,7) appears in Table 5. The table records reported differences 7 between Drosophila, rat and mouse eag channel selectivity expressed in oocytes. Alignment of sequences corresponding to the pore-forming (P-domain) regions of several K+ -selective channels has revealed some distinguishing features in the eagJelkJerg subfamilies (summarized in Fig. 5, this entry).

Single-channel data 46-41-01: d-eag: A single-channel conductance of 4.9 pS determined by nonstationary noise analysist was reported for Drosophila eag heterologously expressed in oocytes s. h-erg: Elementaryt properties of HERG-induced current in oocytes and native IK,r have been directly compared83 : In common with other studies, single-channel conductance was dependent on the extracellular potassium concentration ([K+]o). At physiological [K+]o, 'Y = 2pS, and at 100mM [K+]o, 'Y = lOpS. Single-channel openings occur in bursts with mean duration 26ms at -100 mY; in oocytes, h-erg channel mean open time was 3.2ms and closed times were 1.0 and 26ms.

Voltage sensitivity Steep voltage dependence of heterologously expressed eag channels 46-42-01: d-eag: Drosophila eag mRNA injected into oocytes under conventional two-electrode voltage clamp supports the expression of depolarization-activated, non-inactivating outward currents with activation thresholdst in the range -40 to -30mYs . Outward currents in this recording configuration are biphasic, showing an initial fast-rising phase followed by a slower second component. I- V relationships for Drosophila eag measured in inside-out membrane patches show a broadly similar I- V relation, but (i) a 'partly inactivating' current is observed and (ii) the biphasic rising phase is absent. Notes: 1. A Ca2+ -activated CI- channel (endogenous to the oocyte, see ILG Cl Ca, 46-25) may contribute to this

_'--

e_n_try_4_6

--J

Table 5. Selectivity of eag and erg channels following heterologous expression (From 46-40-01) Species/subfamily Reported properties/comparative notes (data from references as indicated)

Drosophila eag

46-40-02: d-eag: Generally, selective for K+ over Na+, and 'unexpectedly' permeable to NHt (ref.5, see also note 1

for apparent Ca 2+ permeability). The relative permeabilities of a series of test ions (note 2) follows the series K+ (1) > Rb+ (0.75 ±0.11) > NHt (0.25 ±0.08) > Cs+ (0.42 ± 0.14»> Na+ (0.11 ± 0.06) > Li+ (0.08 ± 0.06) (in the ratios shown in brackets; corresponding to Eisenman's seriest IV)5. Rat eag

46-40-03: r-eag: Strongly selective for K+ over Na+. Permeability ratios (note 3) follow the series K+ (1) > Rb+ (0.72±0.26»NHt (0.13 ±0.04) »> Na+ «0.01), Li+

«0.01), Cs+ «0.01). Note: Rat eag channels are not permeable to Cs+ (compare with Drosophila eag, above). Human erg

46-40-04: h-erg: Nernstiant relations confirm h-erg

channels are potassium selective (compare the GFG motif conservation described in Fig. 5 under Domain conservation, 46-28).

Notes: 1. Results consistent with simultaneous K+ efflux and Ca 2+ influx through Drosophila eag channels expressed in oocytes was reported and discussed in ref. 5 . This unusual property may have suggested mechanisms for modulation of synaptic efficiency in vivo and an alternative explanation for effects of eag mutants on multiple native K+ currents (including IK,ca types - see Protein interactions, 46-31). Notably, however, a comparative study of Drosophila versus mouse eag K+ current properties expressed in oocytes has appeared 77; this countered the previous reportS by finding no evidence for Ca2+ flux through eag channels, concluding instead that both d-eag and m-eag channels were 'highly selective' for K+ over Na+ ions. See also Rundown, 46-39 and Selectivity, 46-40. 2. As determined by measuring reversal potentials under bi-ionic conditions, with 115 mM of the chloride salt of the test ion and 1.8 mM EGTA outside versus 100mM KCI and 10mM EGTA inside; outside-out patches, pH7.2 both sides). 3. As obtained from reversal potential determined in two-electrode voltage clamp of oocytes (bath solutions contained 115 mM of the test cation as the chloride salt, 1.8mM CaCl2 and 10mM HEPES, pH 7.2).

'biphasic' activation property. 2. Rat eag channels expressed in the same system do not display biphasic activation, which has been suggested to be due to inherent differences in Ca2+ permeability 7 (see note 1 in Table 5 under Selectivity, 46-40).

II

l_e_n_t_ry_46

_

PHARMACOLOGY

Blockers Block of native IK,r and HERG channels by class III antiarrhythmic drugs 46-43-01: A large literature exists describing patterns of IK,r block by drugs that act by slowing cardiac repolarization (phase 3) and prolong the duration of action potentials (described as 'class III' antiarrhythmic agents). Class ill drugs such as E-4031 (1-[2-(6-methyl-2-pyridyl)ethyl]-4-(4-methylsulfonylamidobenzoyl)piperidine, a sotalol derivative) and MK-499 are often described as potent and relatively specific blockers of IK,r in cardiac myocytes, although some studies report that these agents block other channel types. Many studies have described patterns of block of native IK,r or erg subfamily channels by dofetilide, clofilium, quinidine and sematilide (see also compounds listed in Table 6, this field).

Acquired LQT following therapy with cardiac K+ channel blockers 46-43-02: h-erg: Certain antiarrhymic drugs, such as quinidine and sotalol that include actions which block the cardiac rapidly activating delayed rectifier current (generally designated as IK,r) have been associated with an lacquired' or ldrug-induced' form of long QT (LQT) syndrome (see Phenotypic expression, 46-14). These similarities extend to the induction of torsade de pointes that is observed in familial LQT syndrome (ibid.). The cloning of HERG, its marked similarities of functional properties to IK,r following expression in oocytes (this entry) and the demonstration of mutations in HERG associated with the chromosome 7-linked LQT syndrome (see Chromosomal location, 46-18) thus provide a 'mechanistic link' between inherited and acquired forms of LQT syndrome 12. Pharmacological agents such as sotalol and dofetilide probably exert their antiarrhythmic effects by 'modest' lengthening of cardiac action potentials, thereby suppressing reentrant arrhythmias t (see also Channel modulation, 46-44 for application of elevated [K+]o for correction of LQT anomalies).

Channel modulation Heterologously expressed Drosophila eag ldirectly' modulated by cAMP 46-44-01: d-eag/r-eag: In voltage-clamped oocytes expressing Drosophila eag, bath application of membrane-permeable 8-Br-cAMP (8-bromoadenosine 3'5'cyclic monophosphate, 1 mM) or 8-Br-cGMP (1 mM) increase the amplitude of eag outward currents, while thresholds of activation are shifted to more negative potentials. These effects of 8-Br-cAMP (but not 8-Br-cGMP) persist in the presence of the non-specific protein kinase inhibitor H-7 (1-(5-isoquinolinylsulphonyl)-2-methylpiperazine)5. Following its heterologous expression in Xenopus oocytes, rapid application of cAMP (2 mM) to inside-out patches produces 'significant increases' (typically 10-15 0/0) in outward current amplitude5. This increase is rapidly reversible on perfusion with intracellular bathing solution lacking cAMP. Similar effects cannot be induced by cGMP

_

•

entry 46 '----------------

Table 6. Reported patterns of ionic and pharmacological block of native IK,r or heterologously expressed eag and erg subfamily channels (From 46-43-01) Blocker

Species/homologue, characteristics and references

Small ionic blockers Calcium ions (intracellular)

46-43-03: r-eag: Rat eag channels expressed within HEK-293 cells have been described as 'rapidly and reversibly inhibited' by rises in [Ca2+]i between 30 and 300nM (mean ICso of 67nM in an I/O patch)4. Generation of intracellular calcium oscillations following muscarinic receptor activation appeared to induce a synchronous inhibition of r-eag mediated outward current. These and other data led to the conclusion that r-eag channels are 'voltage-activated, calcium-inhibitable' channels, with the Ca2 +inhibitory effects independent of calcium-dependent kinases and phosphatases4. See also notes 4, 5 and 6 for refs. discussing similarities and differences to Mcurrent.

Cadmium ions

46-43-04: d-eag: C:admium ions at 1 mM do not affect amplitudes of Drosophila eag-mediated outward currents in Xenopus oocytes. Note: This concentration blocks voltage-gated Ca 2 + channels endogenous to the oocyte5 .

Lanthanum ions

46-43-05: h-erg: HERG currents in oocytes can be blocked by lanthanum ions 12.

Barium ions

46-43-06: h-erg: When applied in the bath solution, Ba2+ ions inhibit peak inward HERG current in twoelectrode voltage-elamped oocytes at an IC so (note 1) ~O.6 mM (Fig. 7a2 );: the block is not apparently voltage dependent as observed with 'classical' inward rectifiers (compare Blockers under INR K native, 33-43).

Caesium ions

46-43-07: h-erg: Inhibition of peak inward HERG current in oocytes by Cs+ ions (1 mM in bath solutions) is more effective at negative voltages (Fig. 7b) and appears to reflect Cs+ entering the pore from the outside and interfering with K+ permeation (ibid. 2 ). 46-43-08: r-eag: Relatively impermeable to Cs+ ions 7 (see Selectivity, 46-40).

Large organic ion blockers TEA+

46-43-09: d-eag: I(:so (note 1)5: TEA: 33 ± 11 mM; 4AP >100mM (i.e. resistant; compare to most Sh-type voltage-gated K+ channels under the VLG K entry series). 46-43-10: r-eag: IC so (note 1)7: TEA: 28 ± 13 mM; 4-AP >100mM.--

II

1'---_e_n_t_ry_4_6

_

Table 6. Continued Blocker

Species/homologue, characteristics and references

Pharmacological agents (specifically citing IK,r block) E-4031; voltagedependence

46-43-11: Open-channel blockt of IK,r and HERG by methanesulfonanilides such as dofetilide, E-4031 and MK-499 has been characterized in detail 84 . MK-499 preferentially blocks open HERG channels (steadystate block is half-maximal at 123 ± 1211M at a test potential of -20mV. MK-499 (150nM) does not affect the voltage dependence of activation and rectification nor the kinetics of activation and deactivation of HERG. Notably, 1 J.lM MK-499 and E-4031 have no effect on HERG when oocytes are voltage clamped to a negative potential and are not pulsed during equilibration with the drug. However, MK-499 does block HERG current if oocytes are repetitively pulsed, or clamped at a voltage positive to the threshold potential for channel activation84. Apparent discrepancies of this result with earlier studies showing significant block of IK,r in isolated myocytes by similar drugs, even in the absence of pulsing may be due to differences in species (HERG versus guinea-pig and mouse IK,r), tissues (oocytes versus myocytes) and/or specific drugs. 46-43-12: In some comparative studies 85, E-4031 in

MK-499

E-4031; lack of K+ channel target selectivity

the micromolar range inhibits several classes of K+selective channels (e.g. transient, sustained, and inwardly rectifying) in addition to a sodium-selective current in dissociated rat taste receptor cells. Other comparative studies have compared £-4031 with sematilide and d-sotaloI 86; (see paragraph 46-43-01).

E-4031; [Mg2 +]i dependence of selectivity

46-43-13: Under normal physiological conditions, E-4031 is a specific blocker of IK,r. However, in the absence of intracellular Mg2 +, E-4031 also partially blocks IK,s (see VLC {K} minK, entry 54). Block of IK,s can be prevented by prior treatment of cells with isoproterenol, suggesting that E-4031 only blocks nonphosphorylated IK,s channels in the absence of intracellular Mg2 +75 . Note: E-4031 (0.01-0.3J.lM) is a potent contractile agent in rat portal vein preparations.

Drug-induced arrhythmias

46-43-14: Block of I K,r has been associated with druginduced cardiac arrhythmias (see paragraph 46-43-02).

Dofetilide (also a 46-43-15: The methanesulfonanilide I Kr blocker Kir2.x family blocker) dofetilide acts as as slow-onset/slow-offset, highaffinity open channel blocker of HERG expressed in oocytes (EC so approx. 12 ± 211M)87. Dofetilide block has been extended to single-channel studies of HERG

_r.......-

e_n_try_4_6_


Species/homologue, characteristics and references in oocytes 83 . Notably, however, dofetilide has also been shown to potently block inwardly rectifying channels of the Kir2 family (see note 2).

Dofetilide interaction 46-43-16: Chimaeric constructs between hIRKl and ROMK1 (Kir 1.1, dofetilide resistant) have been used sites to show hydrophobic interactions are essential for dofetilide block in hIRK (and possibly HERG channels)88. L691,121 WIN 61773-2 High- and lowaffinity dofetilidebinding sites

II

46-43-17: Competitive binding studies indicate that E-4031, L-691,121 (3,4-dihydro-1'-[2-(benzofurazan-Syl)ethyl]-6-methanesulonamidospiro[(2H)-1benzopyran-2,4'-piperidin]-4-one), and WIN 61773-2 [(R)( + )-4,S-dihydro-4-methyl-1-phenyl-3(2phenylethyl)-( 1H)-2,4- benzodiazepine monohydrochloride] inhibit IK,r channels by interacting at sites distinct from the high affinity [3H]-dofetilide-binding site (see Ligands, 46-47). WIN 61773-2 binding suggests that it is an allosteric modulator of the dofetilide-binding site89. In separate studies, dofetilide has been shown to interact with high- and low-affinity sites in guinea-pig myocytes in a distinct manner, but it is likely that the highaffinity dofetilide-binding site is related to I K ,r 9o .

Dofetilide affecting cardiac pacemaker function

46-43-18: In rabbit sinoatrial node (SAN) cells, dofetilide can separate delayed rectifier current into 'drug-sensitive' current (IK,r) and 'drug-insensitive' current (I K,s)91. The dofetilide-sensitive current activates rapidly and it showed two components of deactivation, the larger of which is very slow, while the dofetilide-insensitive current activates more slowly and deactivated quickly. Dofetilide slows spontaneous activity, suggesting that IK,r contributes to the pacemaker activity of the SAN cell91 . It has been concluded that SAN IK,r plays an essential roles in (i) determining the maximum diastolic potential and (ii) ensuring the firing of the following action potential in SAN cells 92. Other results suggest that IK,r has 'much less effect' on atrioventricular nodal pacemaker activity than on Sl~ pacemaker activity93.

Non-specific blockers (but including I K,l) NE-I0064

46-43-19: The antiarrhythmic agent NE-I0064 (azimilide) has been reported as a 'selective blocker' of the slowly activating component of the delayed rectifier, IK,s (see KvLQT1/minK under VLC {K} minK, entry 54). In ferret cardiac papillary muscle

""---e_n_t_ry_4_6

_


Species/homologue, characteristics and references preparations, however, NE-I0064 blocks lK,r at an ICso of 0.4 JlM, nearly tenfold greater potency to lK,s (ICso approx. 3 JlM). Note: NE-I0064 also inhibits lea in a use-dependent t fashion 94 .

RP 58866/terikalant

46-43-20: The class III antiarrhythmic agent RP 58866 and its active enantiomer, terikalant, were originally reported to 'selectively block' the inward rectifier K+ current, lkl (see Blockers under lNR K [native), 32-43). However, these drugs also potently block lK,r with ICsos values of 22 and 31 mM, respectively95.

Berberine

46-43-21: The antiarrhythmic drug berberine prolongs action potential duration in cat ventricular myocytes without altering other variables of the action potential and appears to 'preferentially block' lK,r 96 .

Flecainide

46-43-22: The antiarrhythmic class IC drug flecainide also possesses class III effects. 10-30 JlM flecainide inhibits the lK,r but not lK,r in guinea-pig cardiac ventricular myocytes 97. 46-43-23: Along with Kvl.5 98, HERG has been

Terfenadine Desmethylastemizole described as a primary cardiac ventricular target of Erythromycin the non-sedating antihistamine terfenadine with a Ketoconazole Kd(apparent) of 350 nM in oocytes, approx. 10 times more sensitive than Kvl.S (Kd(apparent) of 2.7 J,lM)99. These findings may have relevance to the fact that administration of the antihistamine terfenadine to patients can result in acquired long QT syndrome and ventricular arrhythmias 99 (e.g. Seldane, whose clinical plasma concentration may reach 100nM). In addition to terfenadine, another histamine receptor antagonist (astemizole) has been shown to prolong the QT interval in electrocardiographic recordings in cases of overdose or inappropriate co-medications. Astemizole has been shown to block HERG in oocytes at nanomolar concentrations 100 in addition to its metabolite, desmethylastemizole101 . Other agents commonly assocated with cardiotoxicity (e.g. erythromycin102 and ketoconazole 103 ) also block HERG currents at high potencyl04. Cocaine

46-43-24: Cocaine non-selectively blocks a current identical to 'E-4031 sensitive' current lK,r in isolated guinea-pig ventricular myocytes (IC so rv 4 JlM) in

II

_L.-

e_n_try_4_6_


Species/homologue, characteristics and references addition to L-type calcium current and the TTxsensitive plateau current at higher concentrations (30-100 )lM)105.

BRL-32872

46-43-25: The novel antiarrhythmic agent BRL-32872 [N-(3,4-dimethoxyphenyl)-N-[3[[2-(3,4dimethoxypheny1)ethyI] propyl]-4-nitrobenzamide hydrochloride] inhibits the IK,r in guinea-pig cardiac preparations (EC so 2.8 )lM)106.

Combretastatin Bl

46-43-26: Combretastatin Bl, a polyhydroxybibenzyl compound extracted from the fruit of Combretum kraussii, (the source of 'hiccup nut' toxin) inhibits HERG-type native K+ channels (ICso 300 )lM)107.

Benzodiazepine derivatives

46-43-27: Characterization of a series of 4,5-dihydro3-[2-(methanesulfonamidophenyl)ethyl]-lH-2,4benzodiazepines as potential antiarrhythmic agents that interact at channels underlying IK,r has appeared108.

Other common blockers

46-43-28: d-eag: Quinine ICso 0.7 ± 0.3 mM (note 1)5; quinidine,0.4±0.15mM; quinine 0.9±0.3mM; quinidine 0.4 ± 0.2 mM (i.e. similar to d-eag).

Notes: 1. ICso indicates concentration of blocker required for half-maximal inhibition of the stated current determined under two-electrode voltageclamp. 2. hIRKI (hKir2.1, see entry 33) has been reported109 as a target for dofetilide (IC so 533nM at 40mV and 20°C) with 'no significant effects' on hKvl.2, hKvl.4, hKvl.5, or hKv2.l. 3. The 'HERG-like' current in F-ll neuroblastoma cells (see Cloning resource, 46-10) shares pharmacological features described for native cardiac IK,rincluding similar sensitivities to E-403l, WAY-123,398, Cs+, Ba2 + and La3 +74. 4. Functionally, a 'Ca2+-inhibitable' property might be expected to specifically amplify excitatory stimuli related to membrane depolarization and [Ca2 +h. 5. In control experiments, Kvl family channels such as Kvl.l and Kvl.2 (entry 48) were not inhibited by 400)lM [Ca2+h. 6. [Ca2+h has been suggested to transduce muscarinic block of M-channels, underlying native M-current (for details see VLC K M-i [native], entry 53). These and other kinetic similarities have been discussed in two serial commentaries 110,111. f".J

II

IL-e_n_ _ry_4_6

_

t

-2

/ (JlA)

-3

0 [Cs+]

(mM)

-4

0 -5

[Ba 2 +]

-5

-6

(mM)

Figure 7. (a) Cs+ inhibition of peak inward HERG current in voltage-clamped oocytes. The figure shows an I- V plot before (.) and after (e) application of 1 mM Cs+ in the bath solution, with inhibition being more effective at hyperpolarized potentials. (b) Ba 2+ inhibition of HERG current, with no apparent voltage-dependent component over the concentration range shown. Bath solution was 20mM KC1, 80mM N-methyl glucamine chloride, 1.8mM CaC12, 1.0mM MgC12, 5mM HEPES (pH 7.4). (Reproduced with permission from Trudeau et al. (1995) Science 269: 92-5.) (From 46-43-06)

in inside-out patches, indicating a 'direct modulation' of Drosophila eag channels by cAMps suggesting binding of cAMP to the channel protein. Comparative notes: 1. Rat eag channel currents are unaffected by 8-BrcAMP or 8-Br-cGMP (2mM) in voltage-clamped oocytes 7. 2. Unlike the vertebrate cyclic nucleotide-gated (CNG) channels, which are relatively voltage insensitive (see ILG CAT cAM~ entry 21 and ILG CAT cGMp, entry 22), activation of eag channels show a 'very steep' voltage dependences,8. 3. A possible role for cAMP-modulation of Drosophila eag-type channels in controlling synaptic efficiency in the central and peripheral nervous systems has been discusseds (see also Selectivity, 46-40 and Voltage sensitivity, 46-42).

Heterologously expressed HERG channel activity is unaffected by cyclic nucleotides 46-44-02: h-erg: The HERG channel contains a segment homologous to a cyclic nucleotide-binding domain near its C-terminus (see Domain conservation, 46-28). Following its heterologous expression of homomultimeric HERG subunits in Xenopus oocytes, application of membranepermeable analogues of cAMP and cGMP to bath solutions show 'no significant effects' on current magnitude or voltage-dependence of channel activation 12 (compare with previous paragraph).

Correction of LQT abnormalities by [K+ 10 modulation of HERG 46-44-03: The profound modulatory effects of extracellular K+ in increasing the amplitude of HERG current12 (see Single channel data, 46-41) implies

II

_'--

e_n_t_ry_4_6_

that under conditions of lowered extracellular K+ (i.e. in modest serum hypokalaemia t ) the effects of IK,r blockers would tend to be exaggerated, leading to iexcessive' lengthening of cardiac action potentials, and induction of torsade de pointes. Modest serum hypokalaemia is a commonly observed condition and its incidence has long been associated with ventricular arrythmias l12 and acquired LQTS (see Resen, 1988, under Related sources and reviews, 46-56). Elevation of serum [K+] in hypokalaemic patients receiving medications which block IK,r or in individuals with the chromosome 7-linked LQTS was originally suggested12 as a potential therapeutic intervention for prevention of torsade de pointes. In a later study l13, increase of serum [K+] has been demonstrated to correct abnormalities of repolarization duration, T-wave morphology, QT/RR slope (slope of the relation between Q-T interval and cycle length), and Q-T dispersion in patients with chromosome 7-linked LQTS.

Extracellular divalent cation modulation of rat eag 46-44-04: Extracellular Mg2 + at physiological concentrations 'dramatically slows' activation of r-eag channels expressed in oocytes in a dose- and voltage-dependent manner82 . Similar effects on r-eag activation kinetics by other divalent cations can be observed to an extent that correlates with the ions enthalpy of hydration t. Extracellular H+ ions have been shown to compete with [Mg2 +]0 as the modulatory effects of Mg2 + can be abolished at low pH (i.e. lowering the external p:H also result in a slowing of the activation). A strong dependence of rat eag activation on both [Mg2+]0 modulation and the resting potential have been postulated to constitute a system for fine-tuning iK+ channel availability' in neuronal cells (i.e. where all four r-eag subunits undergo a Mg2 +-dependent conformational transition prior to final channel activation82 .

Ligands Displacement of [3H}-dofetilide with other antiarrythmics 46-47-01: [3H]-Dofetilide binds with high affinity to sites associated with the guinea-pig cardiac IK,r channel (see Table 6; several class ill antiarrhythmic agents, including dofetilide, clofilium, quinidine, sotalol and sematilide, competitively displace [3H]-dofetilide with IC so values that correlate with those for blockade of the IK,r channel (cited in ref. 89) (see also Blockers, 46-43).

Openers 46-48-01: For 'opening' of HERG by extracellular K+ 1 see Channel modulation, 46-44.

Receptor/transducer interactions Mechanisms of arrhythmia suppression by beta-blockade 46-49-01: h-erg: The mechanisms of receptor-coupled (endogenous) control of HERG current in native cardiac cells are presently unclear. Arrhythmias and

1

e_n_t_ry_4_6

_

syncope in LQTS (see Phenotypic expression, 46-14) are induced by sympathetic nerve activity, and because of this, ,B-adrenergic antagonists are frequently prescribed (these agents also suppress the effect of defective KvLQTl channels, see entry 54). Since HERG current does not appear to be affected by cyclic nucleotides (in heterologous cells, see Channel modulation, 46-44) the beneficial effects of ,B-blockade in preventing development of arrhythmia may be due to cAMP-dependent regulation of channel subunits other than HERG, important candidates being voltagegated Na+ channels (see LQT3 in Table 3 under Chromosomal location, 46-18 and Protein phosphorylation under VLG Na, 55-32) and voltage-gated Ca2 + channels (i.e. reducing effects of cAMP-dependent phosphorylation suppressing 'secondary depolarizations' - see Phenotypic expression, 46-14 and Protein phosphorylation under VLG Ca, 42-32). Note however that different LQT2 mutations may affect HERG function in different ways, and there appears to be a role for additional subunits in native HERG function (see Protein interactions, 46-31). These factors make the 'precise' contribution for receptor-linked modulation of HERG difficult to analyse.

Integrin-mediated neurite outgrowth 46-49-02: h-erg: Integrin-mediated neurite outgrowth in neuroblastoma cells appears to depend on the activation of HERG channels in situ - for further details on possible receptor/effector coupling mechanisms, see refs 19,20,114,115 and Developmental regulation, 46-11.

INFORMATION RETRIEVAL

Database listings/primary sequence discussion 46-53-01: The relevant database is indicated by the lower case prefix (e.g. gb:) which should not be typed (see Introduction eiJ layout of entries, entry 02). Database locus names and accession numbers immediately follow the colon. Note that a comprehensive listing of all available accession numbers is superfluous for location of relevant sequences in GenBank® resources, which are now available with powerful in-built neighbouringt analysis routines (for description, see the Database listings field in Introduction etJ layout of entries, entry 02). For example, sequences of cross-species variants or related gene familyt members can be readily accessed by one or two rounds of neighbouring t analysis (which are based on pre-computed alignments performed using the BLASTt algorithm by the NCBIt). This feature is most useful for retrieval of sequence entries deposited in databases later than those listed below. Thus, representative members of known sequence homology groupings are listed to permit initial direct retrievals by accession number, unique sequence identifiers (Seq ID: numbers), author /reference or nomenclature. Following direct accession, however, neighbouringt analysis is strongly recommended to identify newly reported and related sequences.

II

-

entry 46

Eag/Elk/Erg subfamily members Listings are sorted first by subfamily designation then species name (equivalent sequences with different clone names are grouped together in rows). See also notes at foot of table. Type (see field 05)

Original description

Species, DNA source

d-eag

Drosophila eag cDNA sequence

Drosophila head-specific library

m-eag Mouse eag species homologue

Original isolate ORF

Accession Sequence/ discussion

Sequence of cDNA CH20 ORF: 1174aa Mouse brain Mouse brain cDNA library cDNA library ORF: 989aa

gb: M61157

Warmke, Science (1991) 252: 1560-2.

gb: U04294

Warmke, Proc Natl Acad Sci USA (1994) 91: 3438-42. Warmke, Proc Natl Acad Sci USA (1994) 91: 3438-42. Ludwig, EMBOJ (1994) 13: 4451-8.

d-elk

Drosophila elk cDNA sequence

Drosophila elk == ~aghead-specific like K+ library channel

gb: U04246

r-eag

Rat eag cDNA sequence

gb: Z34264

h-erg

Mouse erg species homologue

Rat Reag01/ cerebellum Reag02, cDNA library composite sequence, see Isolation probe, 4612.0RF: 962aa Human erg== ~aghippocampal related gene. cDNA library Isolated by highstringency screen with an m-eag probe:

gb: U04270

Warmke, Proc Natl Acad Sci USA (1994) 91: 3438-42.

Related sources and reviews 46-56-01: Major sources used for compilation of this entrT,7,12; long Q-T syndrome review/molecular mechanisms (1996)116; antiarrhythmic interventions; clinical aspects l17; (1992 review)118; debate relating to resemblance of EAG to native M-current (entry 53)110,111.

II

l_e_n_t_ry_4_6

----'_

Book references: Rosen, D.M. (1988) Arrhythmogenic potential of class III antiarrhythmic agents: comparison with class I agents. In Control of Cardiac Arrhythmias by Lengthening Repolarization (ed. B.N. Singh), pp.559-76. Futura Publishing, Mount Kisco, New York. Schwartz, P.J., Locatis, E.H., Napolitano, C. and Priori, S.C. (1995) The long QT syndrome. In Cardiac Electrophysiology: From Cell to Bedside (eds D. Zipes and J. Jalife), pp. 788-81. Saunders, Philadelphia. Tseng, C.-N. (1995) in Cardiac Electrophysiology: From Cell to Bedside (eds D. Zipes and J. Jalife), pp.260-8. Saunders, Philadelphia.

Feedback Error-corrections, enhancements and extensions 46-57-01: Please notify specific errors, omissions, updates and comments on this entry by contributing to its e-mail feedback file (for details, see Resource T- search criteria). For this entry, send e-mail messages To: [email protected], indicating the appropriate paragraph by entering its six-figure index number (xx-yy-zz or other identifier) into the Subject: field of the message (e.g. Subject: 46-32-02). Please feedback on only one specified paragraph or figure per message, normally by sending a corrected replacement according to the guidelines in Feedback etJ CSN Access. Enhancements and extensions can also be suggested by this route (ibid.). Notified changes will be indexed from within the CSN website (www.le.ac.uk/csn/).

REFERENCES 1

2

3 4

5 6 7

8 9

10 11

12 13 14 15 16 17 18

Warmke, Proc Natl Acad Sci USA (1994) 91: 3438-42. Trudeau, Science (1995) 269: 92-5. Curran, Cell (1995) 80: 795-803. Stansfeld, Proc Natl Acad Sci USA (1996) 93: 9910-14. Briiggemann, Nature (1993) 365: 445-8. Shuck, T Biol Chem (1994) 269: 24261-70. Ludwig, EMBO T (1994) 13: 4451-8. Robertson, Biophys T (1993) 64: A340 (Abstr.). Warmke, Science (1991) 252: 1560-2. Zhong, Science (1991) 252: 1562-4. Strong, TMol Biol Evol (1993) 10: 221-42. Sanguinetti, Cell (1995) 81: 299-307. Balser, T Gen Physiol (1990) 96: 835-63. Follmer, Am TPhysiol (1992) 262: C75-83. Yang, Circ Res (1994) 75: 870-8. Shibasaki, TPhysiol (1987) 387: 227-50. Yang, Circ Res (1994) 75: 870-8. Bruggemann, Proc Natl Acad Sci USA (1997) 94: 537-42.

II

_L..-

,

en_t_ry_4_6_1

Arcangeli, Cell Adhes Commun (1996) 4: 369-85. Arcangeli, TCell Biol (1993) 122: 1131-43. 21 Drysdale, Genetics (1991) 127: 497-505. 22 Wymore, Circ Res (1997) 80: 261-8. 23 Kaplan, Genetics (1969) 61: 399-409. 24 Ganetzky, TNeurogenet (1983) 1: 17--28. 25 Ganetzky, Trends Neurosci (1985) 8: 322-6. 26 Stern, Genetics (1990) 124: 133-43. 27 Engel, T Comp Physiol fA] (1992) 171: 93-104. 28 Budnik, TNeurosci (1990) 10: 3754-68. 29 Griffith, Proc Natl Acad Sci USA (1994) 91: 10044-8. 30 DiFrancesco, Philos Trans R Soc Lond Ser B (1985) 307: 353-63. 31 Schwartz, Circulation (1993) 88: 782--4. 32 Keating, TCardiovasc Electrophysiol (1994) 5: 146-53. 33 Zipes, Circulation (1991) 84: 1414-19. 34 Welsh, Nature (1995) 376: 640-1. 35 Sanguinetti, Proc Natl Acad Sci USA (1996) 93: 2208-12 (correction on p.8796). 36 Bennett, Nature (1995) 376: 683-5. 37 Li, T Biol Chem (1997) 272: 705-8. 38 Dausse, T Mol Cell Cardiol (1996) 28: 1609-15. 39 Benson, Circulation (1996) 93: 1791-5. 40 Satler, Am TMed Genet (1996) 65: 27-35. 41 Antzelevitch, TAm Coll Cardiol (1994) 23: 259-77. 42 January, Circ Res (1989) 64: 977-90. 43 January, T Cardiovasc Electrophysiol (1990) 1: 145-60. 44 Surawicz, TAm Coll Cardiol (1989) 14: 172-84. 45 Surawicz, Cardiovasc Res (1992) 26: 1021-9. 46 Surawicz, Experientia (1987) 43: 1061-8. 47 Schwartz, Circulation (1995) 92: 3381-6. 48 Wang, Nature Genet (1996) 12: 17-23. 49 Keating, Science (1991) 252: 704-6. 50 Keating, Am THum Genet (1991) 49: 1335-9. 51 Roy, Nature Genet (1994) 8: 113-14. 52 Russell, Am THum Genet (1995) 57: 503-7. 53 Jiang, Nature Genet (1994) 8: 141-7. 54 Schott, Am THum Genet (1995) 57: 1114-22. 55 Guy, Science (1991) 254: 730. 56 Schachtman, Science (1992) 258: 663-5. 57 Very, Pfliigers Arch-Eur TPhysiol (1994) 428: 422-4. 58 Anderson, Proc Natl Acad Sci USA (1992) 89: 3736-40. 59 BertI, Proc Natl Acad Sci USA (1995) 92: 2701-5. 60 Sentenac, Science (1992) 256: 663-5. 61 Muller-Rober, EMBO T (1995) 14: 2409-16. 62 Jan, Cell (1992) 69: 715-18. 63 Heginbotham, Science (1992) 258: 1152-5. 64 Hao, Biophys T(1997) 72: MP025. 65 Salkoff, Nature (1981) 293: 228-30. 66 Wu, T Neurosci (1985) 5: 2626-40. 19

20

III

lL......--_ _ _ry_4_6

en t

67

68 69

70

71 72 73

74 75 76 77

78

79 80 81 82

83 84

85

86 87

88 89

90 91

92 93 94 95 96

97 98 99

100 101 102 103 104 105 106 107 108 109 110

111 112 113

114 115

_

Wu, Science (1983) 220: 1076-8. Zhong, T Neurosci (1993) 13: 4669-79. Tinklenberg, Ann NY Acad Sci (1991) 625: 532-9. Chen, Biophys T(1996) 70: SUAM9. Tang, Biophys T (1997) 72: THP14. Yang, Circ Res (1995) 77: 1246-53. McDonald, Circulation (1996) 94: 948. Faravelli, T Physiol (1996) 496: 13-23. Sudo, T Pharmacol Exp Ther (1996) 276: 951-7. Clay, Biophys T (1995) 69: 1830-7. Robertson, Neuropharmacology (1996) 35: 841-50. Smith, Nature (1996) 379: 833-6. Schonherr, T Physiol (1996) 493: 635-42. Wang, FEBS Lett (1996) 389: 167-73. Spector, TCeil Physiol (1996) 107: 611-19. Terlau, Pfliigers Arch (1996) 432: 301-12. Kiehn, Circulation (1996) 94: 2572-9. Spector, Circ Res (1996) 78: 499-503. Sun, Neurosci Lett (1996) 204: 149-52. Lynch, T Cardiovasc Pharmacol (1995) 25: 336-40. Snyders, Mol Pharmacol (1996) 49: 949-55. Kiehn, Mol Pharmacol (1996) 50: 380-87. Chadwick, Drug Dev Res (1995) 34: 376-80. Fiset, TMol Cell Cardiol (1996) 28: 1085-96. Lei, Exp Physiol (1996) 81: 725-41. Ono, Am T Physiol (1995) 38: H 453-H 462. Yamazaki, Eur T Pharmacol (1996) 297: 233-9. Fermini, T Cardiovasc Pharmacol (1995) 26: 259-71. Jurkiewicz, Circulation (1996) 94: 2938-46. Sanchezchapula, Br TPharmacol (1996) 117: 1427-34. Wang, TMol Cell Cardiol (1996) 28: 893-903. Crumb, Mol Pharmacol (1995) 47: 181-90. Roy, Circulation (1996) 94: 817-23. Suessbrich, FEBS Lett (1996) 385: 77-80. Zhou, Biophys T (1997) 72: WAMD1. Daleau, Circulation (1995) 91: 3010-16. Dumaine, Biophys T (1997) 72: TUP15. Roy, Circulation (1996) 94: 3753. Clarkson, T Mol Cell Cardiol (1996) 28: 667-78. Bril, TPharmacol Exp Ther (1995) 273: 1264-72. Guatteo, Neuroreport (1996) 7: 2575-9. Johnson, TMed Chem (1995) 38: 2551-6. Kiehn, Circ Res (1995) 77: 1151-5. Stansfeld, Trends Neurosci (1997) 20: 13-14. Mathie, Trends Neurosci (1997) 20: 14. Curry, Lancet (1976) 2: 231-3. Compton, Circulation (1996) 94: 1018-22. Vannelli, TNeurosci (1995) 15: 4382-94. Bianchi, Biochem Biophys Res Commun (1995) 210: 823-9.

II

_L..116 117 118 119 120

II

Roden, Circulation (1996) 94: 1996-2012. Roden, Annu Rev Med (1996) 47: 135-48. Sanguinetti, Hypertension (1992) 19: 228-36. Frings, Neuron (1995) 15: 169-79. Doyle, Science (1998) 280: 69-77.

e_n_t _ry_46_1

VLG K Kv-beta Edward C. Conley

Cytoplasmic (Kv,B) subunits co-assembling with pore-forming (Kvo) voltage-gated potassium channel subunits Entry 47

NOMENCLATURES Abstract/general description 47-01-01: The existence of Kv-beta subunit polypeptides (Kv{3, this entry) were first detected in native membranes forming part of co-immunoprecipitated complexes using antibodies raised against various Kv alpha subunit polypeptides. In particular, affinity-purificationt and immunoprecipitation of the a-dendrotoxin (a-DTx) sensitive K+ channel complex from bovine brain initially revealed the presence of 38 kDa (major) and 41-42kDa (minor) {3 subunit polypeptides in association with the 'DTx acceptor complex' including various Kvl subfamily a subunit polypeptides. These early studies provided important direct evidence for native voltage-gated K+ channels consisting of an eight subunit hetero-oligomeric sialoglycoprotein complex ('4a4{3 octamers'). 47-01-02: Availability of cDNA clones encoding both Kva and Kv{3 subunits confirmed the presence of rv38-41 kDa {3 subunit polypeptides in 'tight association' with the pore-forming subunits within expressing cells (see Protein molecular weight (purified), 47-22.). Because Kv{3 subunits do not in themselves form integral, ion-selective pores, they have been termed lauxiliary' or laccessory' proteins associated with pore-forming (a subunit) complexes. Subsequently, heterologous co-expression of Kv{3 with Kva subunits has shown {3 subunits able to profoundly affect channel properties (including modulation of inactivation kinetics, voltage dependence of gating and current amplitudes - see Phenotypic expression, 47-14). Moreover, endogenous Kv{3 subunits (e.g. in cell lines) can affect heterologous expression properties of Kva subunits, and Kv{3-toKva-subunit mRNA ratios are known to influence phenotype in heterologous expression systems (ibid.). Many data are consistent with Kv{3 subunits acting as a 'scaffold' or 'adaptor' to bring about'appropriate' hetero-oligomer formation where cells co-express multiple Kva subtypes. 47-01-03: In vivo, certain Kv{3 subunits may have the capacity to modify neuronal output and firing patterns by regulating the 'phenotypic' expression of A-type t Kv channel activities (ibid., see also VLC K A- T [native], entry 44). Heterologous co-expression experiments have determined that Kv{31a subunits can accelerate rates of inactivation in expressed K+ currents (for details, see Inactivation, 47-37). In the heart, the alteration of several functional properties of hKvl.5 channel subunits by Kv{31b subunits offers a potential mechanism for Kva/{3 associations to generate/modulate regional variations in cardiac repolarizing current.

47-01-04: Hypothetically, Kv{31 subunits may act as sensors for 'oxidative stress' in neurones, by virtue of reversible oxidation/reduction of a 'critical cysteine' in the Kv{31 inactivating ball domain (hypothetically, increases in oxygen radicals t would be predicted to eliminate (3 subunit inactivating functions, and thereby enhance K+ efflux) (for details and an illustration of possible 'redox sensing' mechanisms, see the optimized alignments of Kv{3 N-termini (Domain functions, 47-29) and Channel modulation, 47-44).

II

_'--

e_n_t_ry_4_7--J

47-01-05: Extensive, comparative Kv,B subunit-specific mRNA hybridization studies (summarized under mRNA distribution, 47-13) have confirmed patterns of differential (heterogeneous) expression, particularly in brain. High densities of Kv,Bl mRNA are detectable in the striatum, the CAl subfield of the hippocampus, and in cerebellar Purkinje cells. High densities of Kv,B2 mRNA can be seen in the cerebral cortex, cerebellum and brainstem. Kv,B3 mRNA has a distinct distribution pattern, but is still predominantly expressed in brain. cRNA probes have characterized several Kv,B variants in various regions of heart and the vasculature. 47-01-06: Several studies have used imlnunocytochemical methods to colocalize K+ channel a and,B subunit polypeptides in brain slices (summarized under Protein distribution, 47-15). Co-localized Kv,B1 and Kv,B2 protein has been shown to be concentrated in neuronal perikarya, dendrites and terminal fields, and in the juxtaparanodal region of myelinated axons. Furthermore, immunoblot and reciprocal co-immunoprecipitation t analyses have predicted that most K+ channel complexes containing Kv,Bl also contain Kv,B2, which is the major Kv,B component in brain; individual Kv channels may contain 'two or more' biochemically and functionally distinct Kv,B subunit polypeptides. 47-01-07: A Kva-Kv,B interaction site motif in KvI subfamily channel N -termini has been described (see Sequence motifs, 47-24). Studies of cytoplasmic Nterminal domains of Kvl a subunits have determined that the Kv,131-binding site overlaps with the region known as the NAB(Kvl) tetramerization domain (for details, see Protein interactions under VLC K Kv1-Shak, 48-31). The NAB domain appears to mediate assembly of both Kva-Kva and Kva-Kv,B complexes. 47-01-08: A number of potential post-translatory modification t sites exist on Kv,B proteins including multiple phosphorylation consensus sites (see Protein phosphorylation, 47-32) and sites for amidation t . As predicted for proteins with cytoplasmic locations, Kv,B subunit sequences do not contain motifs for leader sequencest or N-glycosylation, but structural motifs thought to be involved in NADPHt binding and the hydrogen transfer mechanismt in aldo-keto reductase t superfamily members are conserved. 47-01-09: Kv,B2 subunits have major roles in promoting K+ channel protein surface expression (see Subcellular locations, 47-16). By analogy to an action of a chaperone t protein, Kv,B2 exerts effects on associated Kvl.2 which include promotion of co-translational N-linked glycosylation of the nascent Kvl.2 polypeptide, increased stability of Kv,B2 /Kval.2 complexes, and enhanced Kv1.2 protein turnover. 47-01-10: Stable association of Kv,132 and Kvl.2 a subunits occurs early in K+channel biosynthesis (5 min), most likely at the endoplasmic reticulum. These observations, together with determination of lKvo: lKv,13 subunit stoichiometric assemblies in co-transfected cells have been used to propose a K+channel oj,13 subunit interaction model (as illustrated in Fig. 1 under Subcellular locations, 47-16). 47-01-11: Kv,B subunit-related primary sequences have been identified in flies encoded by the Hyperkinetic (Hk) locus, plants and prokaryotes (see Database listings, 47-53 and Miscellaneous information, 47-55), perhaps

II

"---e_n_t_ry_4_7

---l_

suggesting that primordial t K+ channels were assembled from Kva:,B subunit combinations. On the basis of primary sequence comparisons, various Kv,B subunit proteins are likely to share structural relationships with members of the NAD(P)H-dependent aldo-keto reductase t superfamilyt (ibid.). Kv,B cDNA sequence variants described thus far show a general pattern of variable N-termini with constant I core' regions; the existence of alternative splicing of Kv,Bl primary transcripts is now well established (see Gene organization, 47-20) and can account for marked sequence variation at N-terminL Subunit genes belonging to different Kv,B subfamilies (e.g. Kv,Bl, Kv,B2 and Kv,B3, see Gene family, 47-05) show much amino acid sequence variability throughout their predicted open reading frames. 47-01-12: No single, systematic nomenclature for Kv,B subunits or genes has been adopted, and this has contributed to some confusion in subunit/gene naming within the early literature. In particular, the late recognition of alternative splicing t of the Kv,Bl primary transcriptt (see Gene organization, 47-20) led to the same name being used for different molecular entities (for co-listing of names in use, see Table 1 under Gene family, 47-05). Overall, studies of Kv,B subunits have suggested further possibilities for extending K+ current diversity, and in consequence, additional mechanisms for modulation of excitability t in native cells.

Category (sortcode) 47-02-01: VLG K Kv-beta, Le. auxiliary (beta) subunits associated with poreforming Kv (alpha) subunits. Note: Although Kv beta subunit proteins (Kv,B, this entry) do not form integral pores their profound significance as 'accessory' or 'auxiliary' subunits for generating functional variation in oligomeric complexes with certain Kv channel a subunit channels (e.g. see Protein interactions, field 31 in entries 48-51) warrants independent description. Different Kva/,B subunit compositions may generate voltagegated potassium channels with properties distinct from the ones expressed by a subunits alone (for review see ref. 2 and other articles listed under Related sources and reviews, 47-56).

Information sorting/retrieval aided by designated gene family nomenclatures 47-02-02: The gene product prefix (used as a unique embedded identifier or VEl) for 'tagging' and retrieving information relevant to the contents of this entry on the CSN website will be of the form VEl: Kvbns where n is a designated number in a [future] systematic nomenclature and (where appropriate) s is a designated ~lice variant letter (e.g. Kv,Bla or where Greek characters are not available, Kvbla). Within this entry, paragraph 'running orders' (sort orders) are largely determined alphanumerically by systematic nomenclatures - Le. denoting species and gene product prefix, sometimes combined with any trivial or clone name(s) where these have been used in the source reference (e.g. hKv,Blb(= clone hKv,B3):). Where properties are likely to apply to all or several subfamily proteins (Le. irrespective of species or isolate) the 'species' term may be omitted.

II

_L..-

e_n_t_ry_4_7_

Channel designation 47-03-01: Although no systematic designations have been proposed for specification of Kv channels comprising defined Kva/Kv{3 subunit complexes, these would normally make reference to gene family relationships and species of origin. For example, a complex formed between homomeric human Kvl.5a subunits with human Kv,Bl subunits could be designated as 'hKvl.5a:hKv,Bla' or hKvl.5a/,Bla. The known a/{3 subunit stoichiometry of 4a to 4,B subunits is generally assumed, unless the a subunit components are physically associated (tandemly linked at the DNA level) in a defined heteromultimeric complex (see Channel designation under VLG K Kvl-Shak, 48-03). More generally, Kva/,B complexes have been referred to as 'a4,B4 stoichiometric complexes' or '4a4{3 octamers' (see Predicted protein topography, 47-30).

Gene family Gene family assignments (in lieu of a systematic nomenclature) 47-05-01: As described in Subtype classifications (47-06) there was no single nomenclature for Kv,B subunits or their genes at the time of entry compilation, and this has contributed to some confusion in subunit/gene naming within the early literature. In particular, the late recognition of alternative splicing t of the Kv,Bl primary transcriptt (see Gene organization, 47-20) led to the same name being used for different molecular entities and conversely, different names being used for the same gene product. In due course, all nomenclature is likely to converge to gene family-based classifications, but for present purposes, alternative names in use are listed side-by side in Table 1.

Relationship of Kvj3 subunits to the NAD(P)H-dependent oxidoreductase superfamily 47-05-02: Like Kva subunits, ,B subunit-related primary sequences have been identified in flies encoded by the Hyperkinetic (Hk) locus, plants and prokaryotes (see Database listings, 47-58 and Miscellaneous information, 47-55). The representation of Kv,B-like subunits over such a wide 'evolutionary gap' has led to the suggestion that primordial t K+ channels were assembled from Kva:,B subunit combinations2 . Furthermore, sequence and secondary structure alignments have indicated a distant gene family relationship between genes encoding various Kv,B subunit proteins and members of the NAD(P)H-dependent aldo-keto reductase t superfamilyt, see Miscellaneous information, 47-55.

Subtype classifications Classifications/nomenclatures for entry

Kv~j

genes/subunits used in this

47-06-01: Kv,B gene/subunit names used in this entry are largely as they have appeared in the source references based on the initial nomenclature

1'--_e_n_t_ry_4_7

_

Table 1. Gene family nomenclature and Inon-systematic' names in use for {3 subunits associated with vertebrate voltage-dependent K+ channels. cDNA Iclone' names (isolates) in common use are indicated in bold. For background to modified nomenclatures used as underlined prefixes to entry paragraphs, see Subtype classifications, 47-06. (From 47-05-01) Paragraph Human Rat Mouse Other designation isoforms isoforms isoforms species Subfamily Kv{31 Kv,81a Names in use and refs (see notes 1 and 2) Kv,81b Names in use and refs Kv,81c Names in use and refs

hKv{31 transcripts hKv,81a: hKv,81 3

hKv,81b: hKv,83 5 hKv,83 6 hKv,83 7 hKv,81c: hKv,81.3 8

rKv{31 transcripts rKv,81a: rKv,81 4

mKv{31 transcripts

rKv,81b: rKv,83 7

Other

fKv,83: Ferret Kv,81b 7

Subfamily Kv{32 Names in use and refs

hKv{32

rKv{32 rKv,82: rKv,82 RCK,82 9

mKv{32

Other bKv,82: Bovine Kv,82 9

Subfamily Kv{33 Names in use and refs

hKv{33

rKv{33 rKv,83: rKvj33 10,11 (see note 7)

mKv{33

Other

Notes: 1. For reasons outlined in Subtype classifications, 47-06, the underlined prefix names are used to establish a 'running order' of entry paragraphs, with any 'clone' or 'isolate' names appended (e.g. hKv,81b (== clonehKv,8a). 2. rKv{31a, rKv{31b and hKv{32 subunits share >82 % identity in the Cterminal 329 aa and show 'low identity' in the N-terminal 79 aa, indicating the likely position of splicing - see the alignment under Encoding, 47-19. 3. Numerical subscripts have sometimes been used to designate Kv{3 subunit types in the literature, but designations in this entry have been largely cited as non-subscripted forms (in line with ref. 4 ). 4. All Kv{3 subunit designations are tentative (for updates on Isystematic' nomenclature of Kv beta subunits, see the IUPHAR pages on the CSN). 5. For convenience, references to the Drosophila Hyperkinetic subunit/gene are prefixed by dHK{3. 6. For sequence accession numbers, see Database listings, 47-53. 7. Rat Kv{33 (as designated in refs.l0,11, but see note 4) is a 403 amino acid residue protein with a 680/0 amino acid sequence homology to Kv{3 1.1.

III

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_t_ry_4_7_

established in ref. 4 (with occasional use of initial letters to designate species). Note, however, that two cDNAs t encoding voltage-gated K+ channel {3 subunits isolated from human heart, originally designated hKv {33 (in ref. s) and hKv,81.3 (in ref.8) represent alternative splice variants t arising from a single gene (i.e. encoding all human Kv,81 variants; as confirmed independently in ref. 12 ). Before this was established, England et a1. 8 suggested a simplified nomenclature for Kv{3 subunits encompassing all the then known gene and vertebrate species variants. To retain consistency with other entries, however, this nomenclature has not been adopted exactly as proposed. Designation of splice variants of the Kv{31 gene as Kv{3I.I, Kv{3I.2 and Kv{3I.3 (as proposed8 ) is inconsistent with designation of splice variants in other K+ channel systelnatic nomenclatures, which are normally designated by serial letters -- e.g. Kv3.2a, Kv3.2b, Kv3.2c and Kv3.2d (i.e. all products of the single Kv3.2 gene - see VLG K Kv3-Shaw, entry 50) or Kirl.la, Kirl.lb, Kirl.lc, Kirl.ld and Kirl.le (i.e. all products of the single Kirl.l gene - see INR K [subunits), entry 33). Adoption of Kv{3I.I, Kv{3I.2 and Kv{3I.3 (as proposed8 ) might also incorrectly imply these gene products arise from separate genes within a gene subfamily (as correctly implied for the established KVQ gene subfamily nomenclature series Kvl.l, Kvl.2, Kvl.3, etc. (see VLC;' K Kv1-Shak, entry 48). For the present purposes therefore, designation of alternative splice variants of the Kv{31 gene will follow the series Kv,81a, Kv,81b and Kv,81c. This nomenclature retains consistency with other entries, and helps indicate their possession of a common C-terminal exon (see Gene organization, 4720). The designations Kv{3la, Kv{3lb and Kv{3lc appear as underlined prefixes coupled to original clone names and have determined the 'running order' of entry paragraphs. All alternative names in use are listed and compared in Table I under Gene famil'y, 47-05. For updates on nomenclatures applicable to Kv{3 subunits and other K+ channel subunits (including newly isolated clones appearing after going to press) see the CSN website (www.le.ac.uk/csn/).

Trivial names 47-07-01: Because Kv{3 subunits do not in themselves form integral, ionselective pores, they have been termed lauxiliary' or laccessory' proteins associated with pore-forming (Q subunit) complexes. Selected crossreferences for {3 subunits associated with voltage-gated channels other than the Kv family are listed under Miscellaneous information, 47-55.

EXPRESSION

Cell-type expression index 47-08-01: See Isolation probe, 47-12, mRN~ distribution, 47-13, Subcellular locations, 47-16, Protein distribution, 47-15 and Protein molecular weight (purified) 47-22.

II

I

entry47

L..---,

_ _

Channel density 47-09-01: For notes on the capacity of the Kv,B2 subunit co-expression to alter cell surface expression characteristics of K+ channel complexes (and in consequence channel densities), see Phenotypic expression, 47-14.

Cloning resource 47-10-01: For sources of mRNA used to construct eDNA libraries for retrieval of presently known Kv,B subunit isolates, see Database listings, 47-53. Low molecular mass fractions of rat brain poly(A)+ mRNA have been shown capable of slowing13, accelerating or modifying the surface expression14 of outward K+ currents mediated by A-type channels (following co-injection with KVQ subunit mRNA in oocytes). The (unidentified) factors encoded by these fractions may include ,B subunit activities (for significance, see Phenotypic expression, 47-14 and other fields).

Developmental regulation 47-11-01: No specific examples of Kv,B gene or protein regulation in a developmental context were described at the time of compilation. From the phenotypic roles already established for Kv,B subunit variants, however, (see Phenotypic expression, 47-14) several known developmental 'cues' (e.g. phosphomodulation t or altered redox states t ) could be expected to exert profound changes on cellular excitability or Kv channel expression determinants.

Isolation probe Microsequencing procedures used to isolate Kv{3 subunit prototypes 47-12-01: bKv,81a: Affinity purificationt and immunoprecipitation of the adendrotoxin (a-DTx)-sensitive K+ channel complex from bovine brain1S- 18 revealed the presence of both 38 kDa (major) and 41-42 kDa (minor) ,B subunit polypeptides in association with the DTx acceptor complex including various Kvl subfamily Q subunit polypeptides (see Protein interactions, 47-31 and for background, see Blockers, 47-43). Direct microsequencing t of 'separated' bovine ,B subunits initially proved technically difficult, and it was first necessary to purify proteolytic fragments using trypsin t , V8t and Asp Nt proteases 9 . Microsequencing of 10 PAGEtpurified peptides yielded nine new amino acid sequences which did not match any known KVQ subunit in database searches 9 . The partial sequences were used to design oligonucleotide peR primer pairs to retrieve eDNA fragments, and later a full-length eDNA sequence from a Agtl0 bovine brain eDNA library (367 aa, see Database listings, 47-53). Using cloned eDNA encoding the bovine DTx acceptor ,B subunit 9 as a probe, cDNAs encoding two homologous polypeptides, initially designated Kv,Bl and Kv,B2 were isolated from rat brain cortex cDNA libraries by conventional low-stringency hybridization4 . Note: The inability to align one peptide sequence derived from microsequencing of ,B subunits associated with QDTx-sensitive K+ channels 9 appeared consistent with the presence of a

II

_'--

.

e_n_t_ry_4_7_

(minor) alternative {3 subunit isoform with a divergent N-terminus and 78% sequence identity to the (major) bovine Kv{32.

Early co-immunoprecipitation experiments suggesting the presence of (3 subunits 47-12-02: The first studies using toxin cross-linking and immunoaffinity procedures identifying multiple Kva/{3 subunit interactions in native K+ channel subunit assemblies (i.e. as part of large hetero-oligomeric t sialoglycoproteint complexes) are outlined in Table 5 under Protein molecular weight (purified), 47-22. Similarly, immunoprecipitation of K+ channel complexes from radioiodinated rat brain membranes using antibodies specific for the rKv2.1/drkl K+ channel a subunit revealed the presence of a 38 kDa {3 subunit polypeptide in 'tight association' with the 'delayed rectifier' complex19. (Note, however that the molecular species is likely to be distinct from Kv{31 or Kv,81, since the latter do not appear to co-precipitate with Kv2.1.)

Isolation of likely splice variant forms of Kv(31 47-12-03: hKv{31b (== clone h Kv,83): A peR-generated eDNA fragment corresponding to nt 435-1089 of rKv{32 was used to screen a A eDNA library derived from cardiomyopathic human heart left ventricular mRNA pools5. A 'partial' clone exhibiting predicted homology (but little identity) to rKv{31a and rKv{32 was used to retrieve a further clone from a eDNA library derived from healthy human cardiac left ventricular eDNA. This clone had a 3.2 kb insert including a large internal ORF predicting a 408 aa coding region for clone hKv{33 (hKv{31b, lor clarification, see Subtype classi-

fications, 47-06).

mRNA distribution General notes on Kv(3 mRNA distributions (see also this field under VLG K Kv1-Shak, 47-13) 47-13-01: mRNA distribution studies which take into account the likely alternative exon t usage from Kv,8 genes were incomplete at the time of compilation. Reported patterns may therefore be different according to which region(s) of Kv{3 sequences were used as a probe, and, in particular, some exon probe sequences may be shared between splice variants. Further investigation of Kv,8 gene organization and transcript mapping t is likely to define probes unique to each 'isoform'. Furthermore, although mRNA distribution patterns may indicate intracellular sites of channel protein biosynthesis (generally the endoplasmic reticulum t in 'expressing' cell types), they do not necessarily report the 'final' distribution(s) of subunits into mature assembled channel complexes, and post-translational transport of protein to distal sites are well-documented. This factor is of some significance for (at least) Kv{32 subunits, whose expression exerts marked changes in cellular localizations of Kva/{3 channel complexes (see Phenotypic expression, 47-14). With these caveats, the following paragraphs

summarize results of Kv{3 mRNA distribution studies, defining the probe in each case.

III

lL...--_ _ _ry_4_7 en t

_

Comparative Kv{31, Kv{32 mRNA versus protein expression patterns 47-13-02: rKv(31; rKv(32: ISH: A summary of expression patterns described for Kv(31 versus Kv(32 subunits in adult rat brain from the extensive study by Rhodes et al. (1996)20 is given in Table 2. This work directly compared subunit-specific mRNA hybridization to immunolocalization patterns especially in neocortex/hippocampus, striatum/basal forebrain, thalamus/ hypothalamus, midbrain and cerebellum/brainstem. In general, (3 subunit genes appeared heterogeneously expressed, with high densities of rKv(31 mRNA in the striatum, CAl subfield of the hippocampus, and cerebellar Purkinje cells. High densities of rKv(32 mRNA were observed in the cerebral cortex, cerebellum and brainstem (for details and original autoradiographs, see ref. 20).

Kv{31a mRNA is predominantly expressed in the rat nervous system 47-13-03: Kv(31a/other splice variants: ISH: Using antisense oligonucleotide probes complementary to the 3' untranslated sequence of Kv(31 (nt 15911542 and nt 1696-1647 in ref. 4 ) in situ hybridization of rat brain sections shows 'notably high' expression of Kv(31 mRNA in (i) the CA1/CA2 fields of the hippocampus; (ii) the caudate putamen. 'Moderate to high' expression is observed in (i) dentate gyrus; (ii) neocortical layers (pyramidal cells); (iii) cerebellum (Purkinje and granule cells); (iv) the CA3 fields of the hippocampus and in (v) thalamic nuclei4 . Comparative note: Sensitive RTPCRt analyses also do not detect Kv(31a mRNA expression in human cardiac ventricle or atriums. Note: Kv,83 (as designated in refs. 10,11, see Table 1 under Gene family, 47-05) is primarily expressed in brain but with a distinct distribution to those of Kv(31.1 and Kv(32.

Kva/Kv{3 mRNA co-distribution patterns (see also Protein distribution, 47-15) 47-13-04: ISH: Comparisons of KVQ and Kv(3 mRNA co-expression patterns may indirectly suggest in vivo co-assembly of their subunits (but see caveats in paragraph 47-13-01). In this regard, it is interesting that Kvl.l mRNA (e.g. RCK1) mRNA is expressed in Purkinje cells, but not in corpus striatum21 , while Kvl.4 mRNA (e.g. RCK4) is highly expressed in corpus striatum, but not in Purkinje cells 21 . Comparison with Kv,B1 mRNA expression (ref. 4, paragraph 47-13-03) may therefore suggest selective assembly of Kvl.l/Kv(31-type K+ channels in Purkinje cells as opposed to Kvl.4/Kv(31-type K+ channels in corpus striatum4 . Irrespective of these specific interpretations, it is clear that differential Kv(3 subunit gene expression can significantly contribute to the structural and functional diversity of Kv channels in the mammalian nervous system (see other fields). Independent in situ hybridization studies using cRNA probes for Kv(31 and Kv(32 20 also reveal heterogeneous expression patterns in adult rat brain, with high densities of Kv(31 mRNA in the striatum, CAl subfield of the hippocampus, and cerebellar Purkinje cells, and high densities of Kv(32 mRNA in the cerebral cortex, cerebellum and brainstem20 (see also Protein distribution, 47-15).

II

II

Table 2. Distribution of Kvf31 and Kvf32 mRNA and protein in rat brain (From 47-13-02) Kvf32 Kvf3l mRNA mRNA Subfield/lamina/sublamina Region

Kvf32 Kvf3l Immunoreactivity Immunoreactivity

++ ++ +++ +++ +++ + + ++ + + ++

+ +++ +++ ++ ++ + + ++ + + ++

+ ++ ++ +++ ++ + + ++ +++

+ +++ +++ +++ ++ + + ++ +++

Infragranular Granule cell Inner third Middle third Outer third

++ +++ + + +

+++ + + ++ + + +

+ ++ + ++ +

+ ++ + +++ ++

S.oriens S. pyramidale S. radiatum S. moleculare

++ + + ++ ++ +

++ +++++ ++ +

++ ++ ++ +

++ + + ++ +++ +

+++++ +++++ +++++ ++

++ ++ ++ ++

++ ++ ++ + + ++

++ ++ ++ + + ++

Cortex I II ill IV V VI

Hippocampus Dentate gyrus

CAl

Striatum Caudate Accumbens Olfactory tubercle Globus pallidus

I! ~

........

Basal forebrain

(b

Medial septal n. Lateral septal n. Diagonal band vert. Diagonal band hor. Nucleus basalis

+++ + +++ +++ +++

+++ + +++ +++ +++

+++ + +++ +++ +++

+++ + +++ +++ +++

Basolateral n.

+++

+ + ++

++

++

Anterior n. Lateral n. Laterodorsal n. VL VPM VPL Lateral geniculate n. Medial geniculate n.

+++ ++ ++ +++ +++ +++ ++ ++ ++

+ + ++ ++ ++ +++ ++ +++ +++ ++ ++

++ +++ ++ +++ +++ +++ +++ ++ ++

++ +++ ++ ++ ++ ++ ++ ++ ++

Medial n. Lateral n.

+ ++

++ ++

++ ++

+ +

Sup. colliculus Inf. colliculus Substantia nigra Pars compacta Pars reticulata Redn. nITI

+++ + + ++

++ +++

+++ ++

++ ++

++ + + + ++ +++ +++

++ + + + ++ +++ +++

++ + + ++ +++++ +++ + + ++

++ + + ++ +++++ +++ + + ++

Amygdala Thalamus

Hypothalamus Habenula

Midbrain

II

nIV

::s

r-t

~ ~

"""".J

II

Table 2. Continued

Region

Subfield/lamina/sublamina

Kv,B1 mRNA

Kv,B2 mRNA

Kv,B1 Kv,B2 Immunoreactivity Immunoreactivity

mnV nVI nVII nVIII

+++ +++ +++ ++

+++ +++ +++ ++

+ + ++ + + ++ + + ++ ++

+ + ++ + + ++ + + ++ ++

mnIX mnX mnXI mnXll

+++ +++ +++ +++

+++ +++ +++ +++

+ + + +

+ + + +

Purkinje cells Granule cells Interneurons Deep nuclei

+ + ++ ++ + +++

++ ++ ++ +++

+++ + + +++

Pons

Medulla

+ + + +

++ ++ ++ ++

+ + + +

++ ++ ++ ++

Cerebellum

+++ + ++ +++

Reproduced with permission from Rhodes et al. (1996) TNeurosci 16: 4846-60.

(t)

=

~

~ ~

l_e_n_try_4_7

---'_

Kvf31b subunit mRNA distribution in human heart 47-13-05: hKv{31b (== clone hKv{33): Northern t blotting analysis 5 using probes

specific for the unique N-terminal segment of hKv{31b shows its mRNA is (i) f".Jtwofold more abundant in human cardiac ventricle than in atrium and (ii) expressed in both 'healthy' and cardiomyopathic human heart tissue. Probes directed against conserved C-terminal segments (ibid.) are f".Jfourfold more abundant in left ventricle than in left atrium, suggesting further {3 subunit mRNAs are present in heart 5 . RT-PCRt analyses with Kv{31bspecific primer pairs detected mRNA expression in human brain cortex cDNA5 . Note: This study also reported Kv{32-specific cDNAs derived from heart tissues. Kv{31b == Kv{33: The Kv{31b transcript (designated Kv{33 in ref. 7) is detectable in multIple tissues, but is most abundant in aorta and left ventricle of the heart 7.

Comparative mRNA analyses using RNAase protection assays 47-13-06: A quantitative expression analysis of Kv{3 subunit mRNAs plus 16

different Kva channel mRNAs in rat sympathetic ganglia has been made using an RNAase protectiont assay22. Eleven a subunit genes and two {3 subunit genes were expressed in sympathetic ganglia; evidence for differential expression (between the superior cervical, coeliac and superior mesenteric ganglia) was obtained for Kv{31, Kval.2, Kval.4 and Kva2.2 22 . Notes: 1. Kv{31 transcripts (and those of Kval.2, Kval.4 and Kva2.2) were 'more abundant' in the pre-vertebral ganglia. 2. None of the distributions obtained in this study22 'matched' those of the M-current or the D2current, which are both prominent in electrophysiological studies of sympathetic neurones (see VLC K M-i, entry 53).

Phenotypic expression 47-14-01: General note: Functional changes 'conferred' by Kv{3 subunits when

co-expressed with Shaker subfamily (Kvla) subunits may have important implications for the 'reproduction' of native channel properties from 'cloned' channels in heterologous cell expression systems. For further discussion of this topic (including the importance of 'appropriate' heterooligomer formation, see also the following fields under under VLC K Kv1Shak (entry 48): Channel designation (48-03), Cell-type expression index (48-08), Developmental regulation (48-11), Phenotypic expression (48-14) and Protein interactions (48-31).

Regulation of native Kv channel properties by Kvf3 subunit expression 47-14-02: Kv{31a: In general, Kv{31 subunits can modify neuronal output and

firing patterns by regulating the 'phenotypic' expression of A-type t Kv channel activities4 : For example, co-expression of rKv{31a subunits with either the rKvl.l a subunit or rKvl.4 a subunit accelerates the rate of inactivation of the expressed K+ currents4 (compared to the rate of inactivation arising from these a subunit polypeptides expressed alone - for details, see Inactivation, 47-37). Association of rKv{31a with a subunits thus confers rapid 'A-type" inactivation on non-inactivating Kvl.l channels ('delayed

II

_L--

e_n_t_ry_4_7_

rectifiers', see also following paragraphs). This effect is probably mediated by an inactivating lball domain' within the variable Kv,Bla N-terminus (for further details, see Channel modulation, 47-44).

Possible functional role of Kv{3 subunits as sensors for toxidative stress' in neurones 47-14-03: Reversible oxidation/reduction of a critical cysteine in the Kv,Bl inactivating ball domain (see above) has been proposed as a potential sensor for oxidative stress in specific neurones 4 . According to this hypothesis, abnormal increases in oxygen radicals t (as in ischaemia t) would be predicted to eliminate ,B subunit inactivating functions, and thereby enhance K+ efflux. For putative mechanisms, see Channel modulation, 47-44.

tConversion' of hKvl.5 delayed rectifier channels to tpartially inactivating' channels 47-14-04: hKv,Bla ( == clone hKv,B3): The hKv,Blb subunit derived from human

heartS alters several functional properties of hKvl.5 channel subunits (which are also normally expressed in heart tissue) when heterologously coexpressed in Xenopus oocytes (see VL(; K Kvl-Shak, entry 48). hKv,Blb cDNA expression 'converts' hKvl.5 froln a delayed rectifier to a channel with rapid, but 'partial' inactivation (for details, see Inactivation, 47-37). hKvl.5/clone Kv,Blb channels also activate at more hyperpolarized voltages and show 'dramatically slower' deactivation (ibid.). Hypothetically, ldif_ ferential association' of channels such as hKv1.5 with ,B subunits in cardiac atrium and ventricle could explain rt~gional differences in repolarizing currents. Commonly encountered difficulties in 'matching' native channel currents to those obtained following heterologous expression of 'cloned' channels are outlined under Channel designation of VLC K Kvl-Shak, 4804. See also following paragraphs.

Kv{3 to KvO'. subunit mRNA ratios influence phenotype in heterologous expression systems (example) 47-14-05: hKv,Bla ( == clone hKv,B3): When cRNAs encoding hKvl.5 and hKv,Bl b subunits are mixed and co-injected into Xenopus oocytes the observed effects (see previous paragraph) depend on the ratio of the two mRNAs: higher Kv,B:KvQ subunit mRNA ratios inhibi1~ hKvl.5 current, whereas lower Kv,B:KvQ ratios may not influence hKvl.5 current (Kv,B alone does not express current)s. In practice, hKv,B mRNA is 'titrated-in' with fixed KVQ to maximize any observable phenotypic effects (see Inactivation, 47-37).

Endogenous Kv{3 subunits in cell lines can affect heterologous expression properties 47-14-06: hKvl.5 Q subunit expression properties differ when heterologously expressed in HEK-293 and mouse L-cells (see below). These differences can be accounted for by the presence of an endogenous Kv,B2.1 subunit coassembling with the transfected hKvl.5 protein in L-cells23, in contrast to HEK-293 cells where no Kv,B protein or mRNA is detectable. In the absence of Kv,B2.l (in HEK-293s), midpoints for activation and inactivation for hKvl.5 were -0.2 ± 2.0 and -9.6 ± 1.8 mY, respectively. Subsequently,

II

1'--_e_n_t_ry_4_7

---'_

the 'L-cell phenotype' (including the kinetics and voltage dependence of activation and slow inactivation) could be 'completely reconstituted' in HEK-293 cells following heterologous expression of Kv,82.1 23 • These observations further underlined the importance for characterization of endogenous ion channel components in heterologous expression systems prior to extensive biophysical characterizations or comparisons.

Role of Kvf32 subunits in K+ channel protein surface expression, stability and turnover 47-14-07: Kv{32: Within the Xenopus oocyte expression system, rKv,82 does not alter macroscopic t K+ channel current properties when co-expressed with rKvl.l or rKvl.44 . Notably, the rKv{32 subunit lacks an N-terminal inactivation 'ball domain' motif (compare previous paragraphs). Heterologous co-expression of Kvl.2 and Kv,82 has, however, been shown to (i) increase aKvl.2 protein turnover and (ii) promote the transport of Kvl.2 to the cell surface (for further details, see Subcellular locations, 47-16). For additional phenotypic features of Kv,82, e.g. modulation of the inactivation properties of Kvl.4 a subunit complexes, see Fig. 4 under Inactivation, 47-37.

Phenotypic functions of Drosophila Hyperkinetic 47-14-08: dHk,8: The Hyperkinetic (Hk) gene encodes a Drosophila homologue of mammalian Kv,8 subunits24 . Mutations in the Hk gene alter firing patterns and A-current properties in native and cell preparations, 'similar to but "milder" than' Sh mutations25 and epistasis t of Sh and Hk has been observed in double mutants. Genetic and physiological studies of the Drosophila Hyperkinetic (Hk) mutant reveals defects in the function or regulation of K+ channels encoded by the Shaker (Sh) locus. Co-expression of Hk with Sh in Xenopus oocytes increases current amplitudes and changes the voltage dependence and kinetics of activation and inactivation, consistent with predicted functions of Hk in vivo. Sequence and secondary structure alignments of Hk with mammalian Kv,8 sequences show they represent an additional branch of the aldo-keto reductase t superfamily (see Gene family, 47-05 and Miscellaneous information, 47-55).

In vivo effects of mutations in the Drosophila Hyperkinetic (Hk) gene (examples) 47-14-09: dHk,8: Drosophila Hk (Hyperkinetic) mutants were initially shown24 to induce a leg-shaking phenotype under anaesthesia (which showed similarities to mutational phenotypes at the Shaker locus, albeit with less 'severity'). Hk mutations were also associated with rhythmic bursts of spontaneous activity in motor neurones of thoracic ganglion in adult flies 25,26. Current-clamp t of cultured 'giant' neurones derived from Hk mutants can distinguish two types of spontaneous sustained firing (occurring over >5 min) which do not occur in wild-type t neurones. In the first type, action potentials display little undershoot t and pre-potential t , while the second is characterized by strong undershoot and pronounced pre-potentiaI27. Voltageclamp t studies in Hk giant neurones displaying abnormal spontaneous firing are associated with (i) approx. 50% reduced I A compared to wild-type neurones, with all-or-none action potentials under current injection t i (ii) slower recovery

11II

_1...-

en_t_ry_4_7_1

Table 3. In vivo functional effects of Hk beta subunit mutants in the Drosophila larval muscle preparation. (From 47-14-10)

Wild-type Hk (various alleles)

Wild-type Hk (various alleles)

IA (nA/nF)a

IK (nA/nF)a

tp

7

(ms)b

(ms)b

5.3±0.3 3.2±0.3

3.6±0.3 3.2±0.2

9.3±0.2 13.6±0.4

10.4±0.3 14.9±0.2

Vmli2 (mV)b

Vmsl ope (mV/e-fold)b

Vh 1/ 2 (mV)b

Vhslope (mV/e-fold)b

-13.5 -11.0

8.0 6.5

-39 -33

4.5 4.5

Values were derived following step-depolarization of larval muscle fibres (-80 to OmV) at 11°C (n == 9-15, mean ± SEM). b t p, time to peak; 7, inactivation time constant of IA; Vml/2, voltage for halfconductance of IA; Vmslope , limiting slope for activation curve; Vh 1/ 2 , voltage for half-inactivation of IA; Vhslope, limiting slope for inactivation curve. Data a

of I A from inactivation and (iii) a relative insensitivity to Q-dendrotoxin27. These findings are consistent with predictions24 that Sh-Hk heteromeric channels would activate earlier in the action potential (with greater magnitude) and hence would be more efficient at repolarization.

Effects of Hk mutations on A-current modulation in Drosophila larval muscle (examples) 47-14-10: dHk,B: The role of the Hyperkinetic gene product in the in vivo modulation of Drosophila larval muscle fA is supported by various Hk mutants displaying (i) reduced current amplitudes and (ii) substantially slower kinetics of channel activation, inactivation and recovery compared to wild-types28 (summarized in Table 3). Notably, no significant defects in Drosophila I K are detectable (in contrast to the marked effects on [A); furthermore, effects with Hk null alleles t are 'no more extreme' than with other Hk alleles, indicating a modulatory role for the ,B subunit in channel gating and conductance28 .

Effects of Drosophila Hyperkinetic following heterologous co-expression with Shaker (examples) 47-14-11: Co-expression of wild-type Drosophila Hyperkinetic (Hk) ,B subunits with Drosophila Shaker (Sh) Q subunits in voltage-clamped Xenopus oocytes24,29 show (i) ""twofold increases in current amplitude in Sh plus Hk co-expressors compared to when the Q subunit is expressed alone; (ii) acceleration of activation kinetics; (iii) acceleration of inactivation kinetics and (iv) an approx. 10mV shift of the voltage dependence of activation and inactivation in the hyperpolarizing direction. When co-expressed with Sh Q subunits lacking a fast inactivation domain, the effects on current amplitude and activation kinetics persist29, consistent with predicted functions of Hk in

vivo (see previous paragraphs).

1'--_e_n_t_ry_4_7

_

Protein distribution Co-localization of K+ channel a and f3 subunit polypeptides in rat brain 47-15-01: A summary of reported co-localizations patterns for o./{3 K+ channel

subunits and other notable {3 subunit protein distributions appears in Table 4. Note that a number of antibodies used in these studies have been reported as having both species and isoform cross-reactivity. This property is largely dependent upon which region of the {3 subunit polypeptide is used to design the peptide epitope t, with C-terminal-derived antibodies having notable cross-reactivity amongst presently known isoforms (see sequences under Encoding, 47-19 and Domain conservation, 47-18). Cross-reactivity has both advantages and disadvantages within comparative distribution studies (for a discussion, see ref. 30).

Subcellular locations Hypotheses accounting for observed a/f3 subunit distributions in neurones 47-16-01: Immunohistochemical staining and co-localization patterns of Kvo. and Kv{3 subunits (see Protein distribution, 47-15 and Protein interactions, 47-31) predict mechanisms that can 'direct' the selective interaction of K+ channel 0. and {3 subunit polypeptides (discussed in ref. 30).

These data are consistent with properties of K+ channels in specific subcellular domains regulated by the formation of heteromultimeric K+channel complexes containing limited combinations of 0. and {3 subunits. At the time of compilation, little direct information regarding specific mechanisms of subcellular 'targeting' of {3 subunits in native membranes was available (e.g. to somatic, dendritic, axonal or synaptic membranes). Evidence for o.4{34 stoichiometric complex assembly as an early event in channel biosynthesis within transfected cells has been used to propose a model of K+channel o./{3 subunit interaction (see next paragraph and Fig. 1).

Promotion of cell surface expression and stability of Kvl.2 by co-expression with Kvf32 47-16-02: Stable association of the Kv{32 and Kv1.2o. subunits occurs early in biosynthesis, presumably at the endoplasmic reticulum31 . In COS cells transfected with expression constructs encoding Kvl.2 alone, the majority of the expressed protein appears 'trapped' within the cell (as determined by immunocytochemistry). Although Kv{32 has no effect on the kinetics of 0. subunit inactivation (ref. 4, see Inactivation, 47-37) heterologous coexpression of Kv1.2 and Kv{32 promotes the transport of Kv1.2 to the cell surface. These o.j{3 interactions in COS cells exhibit a predicted stoichiometry of 1: 132, in agreement with studies in native cells (see Protein molecular weight (purified), 47-22). By analogy to an action of a chaperone t protein, Kv{32 exerts effects on associated Kv1.2 which include promotion of co-translational N-linked glycosylation of the nascent Kv1.2 polypeptide, increased stability of Kv{32jKvo.1.2 complexes, and increased efficiency of cell surface expression of Kv1.2 (as above). Furthermore, Kv1.2 protein turnover occurs

_'---

e_n_t_ry_4_7_

Table 4. Summary of Kv{3 subunit co-distributions detected in brain (see also VLC K Kv1-Shak entry 48) (From 47-15-01) Kv{3 Properties and cross-references Kv,81a Most K+ channel complexes containing Kv{31 also contain Kv{32 Kv,81b 47-15-02: Immunohistochemical staining using subunit-specific monoclonal and affinity-purified polyclonal antibodies has revealed that the Kv{31 and Kv{32 polypeptides frequently co-localize20. Immunoblot and reciprocal co-immunoprecipitationt analyses predict that most K+ channel complexes containing Kv{31 also contain Kv{32. Co-localized Kv{31 and Kv{32 protein has been shown to be concentrated in neuronal perikarya, dendrites and terminal fields, and in the juxtaparanodal region of myelinated axons. These data further suggested that individual Kv channels may contain 'two or more' biochemically and functionally distinct Kv{3 subunit polypeptides20. Table 2 (under mRNA distribution, 47-13) compares protein versus mRNA expression patterns described by Rhodes et al. (1996)20 in adult rat brain. Kv,82 47-15-03: Immunohistochemical staining30 using antibodies directed against the C-terminal18 aa residues of the bovine Kv{32 subunit (associated with the dendrotoxin-sensitive K+ channel complex - see Protein molecular weight (purified), 47-22 and Database listings, 4753) shows 'widely distributed' immunoreactivity in adult rat brain30 . The antibody was cross-reactive (see paragraph 47-15-01), detecting two major bands of 50 kDa (non-specific) and 38 kDa with a lessprominent band of 41 kDa in rat brain membrane preparations (ibid. 30 ). The cellular distribution of {3 subunit immunoreactivity corresponds closely with that for Kvl.2 (and to a lesser extent Kvl.4) but not with Kv2.1, consistent with other findings from the same study (see Protein interactions in this entry, 47-31 and under VLC K Kv-Shak, 48-31). Further detailed descriptions relating differential distributions of anti-Kv{32 immunoreactivities (e.g in neocortex, hippocampus, piriform cortex, striatum, cerebellum, cranial nerve nuclei and most major white matter pathways) are illustrated in ref. 30. 47-15-04: Using immunoblot and reciprocal co-immuno-precipitationt analyses an extensive independent studro concluded (i) that Kv,82 is the major {3 subunit present in rat brain membranes and (ii) Kv{32 is a component of 'almost all' K+ channel complexes containing Kv1Q subunits (see also mRNA distribution, 47-13).

Notes: 1. Determination of K+ channel Q subunit polypeptide associations with Kv{3 subunits have been performed using ~reciprocal co-immunoprecipitation' methodologies. In these assays, crude membrane preparations are subjected to immunoprecipitation using antibodies specific for Q or {3 subunit polypeptides; resultant immmunoprecipitation products are then analysed by immunoblotting, using either the Q or {3 subunit-specific antibodies. 2. In-situ protein localization studies for Kv{3 subunits (e.g. ref. 30) contain detailed descriptions relating {3 subunit expression to specific brain regions. To help consolidate all such data into a single reference source and make it available on the WWW, please forward citation details to the e-mail address shown under Feedback, 47-57.

(D

=

51

f""t'

~

Kv1.2a protein

~

""'" 4. Early KV1.2a:Kv~2 interaction promoting surface expression possibly via assisting correct folding or assembly

3

k--'

Cytoplasm

~

ER lumen

1

1. Exit of nascent polypeptide from ribosome complex

, 3. Post-translational glycosylation (oligosaccharide chain addition) Asn-207

IKeYI II

~

Direction of movement for newly-synthesized Kv1.2 polypeptide

'r

Consensus site for N-linked 9'ycosy'ation on Kv1.2 (a unique site between domains 51 and 52, see Sequence motifs, 47-24).

Figure 1. Model for association of Kv{32 subunits with partially synthesized Kv1.2a subunits (steps 1 to 4). The co-translational association (indicated at step 4) has been proposed to affect the efficiency of co-translational oligosaccharide addition (step 3) due to effects on Kv1.2 folding either before or during ER translocation (step 2) of the Sl and S2 transmembrane segments. (From 47-16-02)

_ 1 . . . . . - - - -_

_.

entry47

I ,

at a higher rate in Kvl.2/Kv,B2 co-transfectants, and these cells display increased binding of (l 25 I]-DTx (see Protein interactions under VLC K KvShak, 48-31) by increasing Bmax without affecting K0 31 . Additional features

of a/,B subunit interactions and a proposed model for these early biosynthetic events (occurring ~5 min of nascent protein synthesis) are outlined in Fig. 1, this field.

Transcript size 47-17-01: rKv,Bl/hKv,Blb (= clone hKv/3a): No extensive study of transcriptional control/mapping of Kv,B genes was published at the time of entry compilation, although the existence of alternative splicing of Kv,Bl primary transcripts is now well established (see Cene organization, 47-20). Using a Kv,Bl riboprobe t (nt 669-694) in Northern t blots of total cell RNA extracted from rat brain, two transcripts of approx. size 3.6 and 3.9 knt are observed4 . These signals are absent in mRNA pools from heart, skeletal muscle or kidney. The size of the hKv,Blb(=clonehKv,Ba) transcript in various human heart mRNA pools has been estimated as ",4 knt 5 .

SEQUENCE ANALYSES

Chromosomal location 47-18-01: hKv,Bl: The human Kv,Bl.l gene ('KCNAIB' - see note 1 below) has been mapped to human chromosome 3q26.1 by fluorescence in situ hybridization t (FISH) confirmed by PCR screening of the CEPH YACt library and a chromosome 3 position predicted by somatic cell hybrid mapping33 . hKv,Bb (= clone hKv,Ba): In separate studies, the gene encoding clone hKv,B3 was also localized to human chromosome 3 using a human/ rodent cell hybrid mapping panel and Southern blot analyses 5 . hKv,B2: The human Kv beta 1.2 gene ('KCNA2B' - see note 1 below) has been localized by FISH to human chromosome Ip36.3 consistent with a chromosome 1 location as indicated by somatic cell hybrid mapping33 . Notes: 1. These gene names and human chromosomal locus names appeared in ref. 33 and are subject to review (compare with Cene mapping locus designation under VLC K Kv1-Shak, 48-54 for established locus names for the Kv1a subunit genes, which are an unrelated gene family). 2. In Drosophila, Hk is located at polytene bands 9B7-8 (Schlimgen, A.K., cited in ref. 24 ).

Encoding Variable N-termini linked to a constant aida-keto reductase core 47-19-01: Kv,B subunit cDNAs (as described thus far) encode alternative Nterminal regions coupled to a conserved ('constant') region homologous to the aldo-keto reductases (the laldo-keto core', see also Domain conservation, 47-28). An illustration of this arrangement, together with an alignment of rKv,Bla, hKv,Blb and rKv,B2 amino acid sequences (modified from ref. 5) is shown in Fig. 2. Kv,Blb = Kv,B3: A cDN.A (designated Kv,B3 in ref. 7) has a

II

_ _ _ry_4_7 en t

_

L..--

(a)

....__....",, Alternative

-..-

Aldo-keto reductase-homologous 'core' region "

:~; • • •~ • •II-• • •H • • •_II• •I

t-

N-termini

....__.....r;'

;';';'

(b)

•

MHLYKPACAD 1PSPKLGLPK SSESALKCRW HLAVTKTQPQ AACKPVRPSG MQV S1ACTEHNL- -RNGEDRLLS KQSS-APNVV N-ARAKFRTV MYPESTTGS

hKv~3 rKv~l rKv~2

*.* 0 51 AAEQKYVEKF LRVHG1SLQE TTRAETGMAY RNLGKSGLRV SCLGLGTWVT 44 -11ARSLGT- TPQ-H---K- S-AKQ---K- ---------- ---------10 P-RLSLRQTG SPGM1Y-TRY GSPKRQLQF- ---------- ---------101 FGGQ1SDEVA ERLMT1AYES GVNLFDTAEV YAAGKAEV1L GS11KKKGWR 94 ---------- ---------- ---------- ---------- ---------60 -----T--M- -H---L--DN -1-------- --------V- -N--------

o.

••

151 RSSLV1TTKL YWGGKAETER GLSRKH11EG LKGSLQRLQL EYVDVVFANR 144 ---------- ---------- ---------- ---------- ---------110 ---------1 F--------- ---------- --A--E---- ----------

•

201 194 160

PDSNTPMEE1 VRAMTHV1NQ GMAMYWGTSR WSAME1MEAY SVARQFNM1P ---------- ---------- ---------- ---------- -----------P------T ---------- ---------- --S------- -------L--

251 244 210

PVCEQAEYHL FQREKVEVQL PELYHK1GVG AMTWSPLACG 11SGKYGNGV ---------- ---------- ---------- ---------- ----------1-------M ---------- ---F------ ---------- -V----DS-1

301 294 260

PESSRASLKC YQWLKER1VS EEGRKQQNKL KDLSP1AERL GCTLPQLAVA ---------- ---------- ---------- ---------- ----------PY------G -----DK-L- ----R--A-- -E-QA----- --------1-

•

..

--

351 WCLRNEGVSS VLLGSSTPEQ L1ENLGA1QV LPKMTSHVVN E1DN1LRNKP 344 ---------- ---------- ---------- ---------- ---------310 ---------- ----A-NA-- -M--1----- ---LS-S1-H ---S--G---

•

401 YSKKDYRS 394 -------360 --------

Figure 2. (a) General Kv{3 subunit arrangement encoding alternative Nterminal regions {white boxes} coupled to a conserved tconstant'} region {black boxes} homologous to the aldo-keto reductases {the aldo-keto reductase core} (Based on figure from ref.12.) (b) Alignment of rKv{31a, rKv{32 and hKv{31b {clone hKv{33} amino acid sequences, illustrating several features described in the text. Key: -, identical amino acids; *, consensus sites for phosphorylation by PKC common to rKv{31a and rKv{32 but not hKv{31b {clone hKv{33}; ., consensus sites for phosphorylation by PKC in clone hKv{33, some of which are conserved in other {3 subunits, with Ser13 and Thr71 being unique to clone hKv{33; 0, consensus site for phosphorylation by cAMP-dependent protein kinase {Ser153 in clone hKv{33} conserved across all Kv{3 subunits; D, location of a potential splice site predicted by the nucleic acid sequence {notably, this coincides with a point of divergence between the {3 subunit N-termini}. (Alignment modified from England et al. {1995} Proc Natl Acad Sci USA 92: 6309-13.) {From 47-19-01}

II

_ entry 47

"-----------

408 amino acid open reading frame, and possesses a unique 79 amino acid Nterminal leader (but is otherwise identical with rat Kv,B1a over the 329 Cterminal amino aCids 7). Kv,B3 (as designated in refs 10,11, see Table 1 under Gene Family, 47-05) also has a long N-terminal structure and induces inactivation in N-terminal deleted Kv1.4 (but not in other members of the Kv1 channel family).

Gene organization Alternative splice variants 47-20-01: 'Direct' sequence comparisons between genomic and cDNA sequences encoding Kv,B subunits were not published at the time of compilation. From cDNA sequence variants described thus far, there appears to be a general ,B subunit pattern of variable N-termini with constant 'core' regions (see also Domain conservation, 47-28 and Domain functions, 47-29). Analysis of nucleic acid sequences, suggests5 the point of amino acid sequence divergence between rKv,B1a and clone hKv,B3 is an alternative splice junction (illustrated in Fig. 2, under Encoding, 47-19). This is supported by the majority of the 'C-terminal 4/Sths' being identical between rKv,B1a and hKv,B3 (ibid.). In c.omparison, the first 72 N-terminal amino acids of rKv,B1a do not align with the Kv,B2 N-terminus; in this case, however, differences in the 'C-terminal 4/Sths' of the rKv,B1 and rKv,B2 sequences are 'interspersed' (albeit they represent largely conservative substitutions)4.

Homologous isoforms Species homologues of Kvf3 subunit proteins show notably high identities

47-21-01: rKv,B1/bKv,B2: The rat Kv,82 amino acid sequence4 shares approx. 990/0 amino acid identity with the previously obtained sequence for the bovine Kv,82 subunit9 (Le. they differ by only 4 residues out of 367). A 100% identity has been noted between the C-terminal 329 aa segment between the rat Kv,B1 sequence and the human clone Kv,83 (derived by alternative splicing from the Kv,81 gene in humans). These identities indicate that the respective proteins have 'equivalent' or at least similar physiological roles in the different species. General note: Subunit genes belonging to different Kv,B subfamilies (e.g. Kv,B1, Kv,B2 and Kv,B3, see Gene family, 47-05) show much antino acid sequence variability throughout that predicted by their open reading frames. For further interspecies comparisons, see also the sequences shown under Database listings, 47-53.

Protein molecular weight (purified) Kva/f3 subunit assemblies revealing the native constitution of voltage-gated K+ channel 47-22-01: A summary of protein Mr estimates for recombinant Kv,B subunits and native protein complexes containing Kv,B subunits appears in Table 5. Minor variations in gel mobilities/molecular weights for Kv,B subunits have

II

II...-_e_n_try_4_7

_

Table 5. Protein molecular weight estimates for recombinant Kv{3 subunits and native protein complexes containing Kv{3 subunits (From 47-22-01) Kv{3 Mr (purified), notes and cross-references designation/ complex 47-22-02: Antibodies raised against a C-terminal 18 aa peptide epitopet of the {3 subunit originally associated with the bovine dendrotoxin-sensitive K+ channel complex (bKv{32, see below and Database listings, 47-53) have been used to characterize the related {3 subunit polypeptides in rat brain (ref. 30, see Protein distribution, 47-15). These cross-reactive antibodies (ibid.) can detect a 'major' 38kDa polypeptide and a 'minor' 41 kDa polypeptide in rat brain membranes, possibly corresponding to the predicted sizes of the rKv{32 and rKv{31a respectively (compare this table, below and Protein molecular weight (calc.), 47-23). (See also notes 1-3). 47-22-03: Historically, use of 12sI-labelled a-DTx (see Early studies first defining Blockers, 47-43 and Ligands, 47-47) in immunoprecipitation native K+ experiments34,35 with rat cerebrocortical synaptosomes were important in establishing (i) that the toxin receptors were channels as large heterointegral membrane proteins, while subsequent cross-linking studies36,37 led to the first identification of K+ channel a oligomeric subunits with Mr estimates falling between 65 kDa (Nsialoglycoproteins deglycosylated, see below) and 78 kDa (glycosylated), depending on the electrophoretic conditions used 18 . Anomalous 47-22-04: Note: Channel proteins cross-linked to toxin molecules may exhibit lanomalous' migration in SDSmigration of a subunits but PAGEt analyses· digestion with sialidase t and peptide not {3 subunits N-glycosidase FT abolishes anomalous migration but bands may still be 'broad and diffuse' indicating subunit heterogeneity. Beta subunit bands do not exhibit anomalous migration behaviour since they are not glycosylated (see Sequence motifs, 47-24). Immuno47-22-05: Following a-DTx affinity purification from bovine precipitates of brain (see above), fast-activating, voltage-sensitive K+ intact K+ channels can be purified into distinct oligomer populations channel with (i) a major, 'large form' population (containing Mr complexes 39 kDa (3 subunits) being in the molecular weight range 370(a-DTx 420 kDa (after correction for detergent binding) with a minor acceptor 'small form' population (lacking 39 kDa (3 subunits) in the range 240-265 kDa 16 . Note: These estimates were for the populations) protein moiety of the hetero-oligomeric sialoglycoprotein complexes, and in order to correct for attached carbohydrate, the degree of glycosylation was assumed to be that for the Na+ channel (f'./30 0/0 w/v hexose; u = 0.61 ml/g). Beta subunits within DTx-sensitive complexes

47-22-06: On the basis of these ranges, {3 subunit-containing forms are typically cited as the '400 kDa complex', as

II

_L....-

en_t_ry_4_7_1

Table 5. Continued

Kv{3 designation/ complex

Mr (purified), notes and cross-references

determined from sedimentation analysest using the buoyant density method (giving 'large form' Svedberg unitt values of 11.2S and 9.9S in H20 and 2H20 respectively16). See consistency with Q;4{34 stoichiometric complexes under Predicted protein topography, 47-30). 47-22-07: Using gel filtrationt analysis, a Stoke's radiust of 8.6 nm has been determined for the large form of the oligomeric receptor 9,16. This estimate agrees with studies carried out on solubilized Q;-DTx acceptors. Kv,81a

Note: rKv{31a ORF: 401 aa

47-22-08: Kv{31a: An Mr value of 41 kDa has been determined for rat brain Kv{31a (Nakahira et. a1., cited in ref. 30 ). rKv{31a transiently expressed in COS-l cells has a 'mobility similar to, but slightly larger than' 41 kDa (ibid.). (See also note 3.)

47-22-09: hKv{31b (== clone hKv(33): Specific estimates of Note: hKvj31b purified Mr were not found during compilation, but may be expected to be close to those for Kvj31a (compare ORFs in ORF: 408aa first column and conserved regions between Kvj31a/Kvj31b under Domain conservation, 47-28).

Kv,81b

Kv,82

Note: bKvj32 ORF: 367aa

47-22-10: ~2: Minor variations in reported Mr also exist for Kvj32, most commonly estimated at 38 kDa30. (See also notes 2 and 3.)

Notes: 1. On the basis of similar gel mobilities under denaturingt (reducingt) or nondenaturing t conditions, the 31 kDa and 41 kDa subunits do not apparently contain extensive intra- or interchain disulphide bridges (cited in ref. 30). 2. In independent studies19, a 38 kDa polypeptide co-immunoprecipitating with Kv2.1/drkl subunits probably represent a 9istinct 13 subunit to the Kvj31a and Kv{32 proteins described above (for Kv2.1/d.rk1, see VLC K Kv2-Shab, entry 49). 3. Non-specific bands at rv50 kDa (sometimes observed in co-immunoprecipitations of Kv 13 subunits3o ) are likely to represent heavy chains of rabbit IgG reacting with anti-rabbit secondary antibodies. 4. Purification (detergent/buffer) and storage conditions enhancing stability of Q;-DTx receptor complexes are detailed in ref. 16 and references therein. 5. Protein kinase A appears to phosphorylate a 40 kDa protein which coimmunoprecipitates with the type n channel (hKvl.3) from membranes of Jurkat T cells38 (see Table 16 in Protein phosphorylation under VLC K Kv1Shak, 48-32 and also the description of Kvj3 subunit expression in activated T lymphocytes39 (abstract). 6. Compare protein mol. wt of (i) mammalian Kv{3 subunits (above) with that of the mammalian maxi (BKCa) channel (31 kDa (for refs, see ILC K Ca, entry 27, and Miscellaneous information, field 55, this entry) and (ii) an Arabidopsis thaliana 13 subunit (38.4 kDa) (see ref.40).

II

lL...--e_n_ _ry_4_7 t

_

been reported between studies, but these may be attributable to differences in electrophoretic conditions used. Molecular weight differences of Kv0/{3 complexes are probably due to differential glycosylation patterns and 0 subunit compositions. Table S also summarizes studies of a-dendrotoxin receptor complexes t, where the Kvo subunit components correspond to several Shaker-related Kv1 subfamily proteins tightly but non-covalently associated with smaller Kv{3 subunit variants. These studies provided important direct evidence for native voltage-gated K+ channels consisting of large eight-subunit hetero-oligomeric complexes ('404{3 octamers', see also Predicted protein topography, 47-30).

Protein molecular weight (calc.) 47-23-01: Kv{31a: Predicted from cDNA sequence (ORF 401 aa) rv44.7kDa 4. hKv{31b (~e hKv{33): Predicted rv4S.0kDa (for the full ORF of 408 aa); note that if cleavage by amidation t were to occur (see Sequence motifs, 4724 and Resource C - Consensus sites and motifs, entry 62) the predicted molecular mass for clone Kv{33 would be rv36 kDa 5 . Kv{32: Predicted from cDNA sequence (ORF 367 aa) rv41 kDa (40983 Da)4,9. - -

Sequence motifs A KVQ-Kv{3 interaction site motif in Kvl subfamily channel N-termini 47-24-01: The Kv1.S N-terminal region (90 aa residues, positions 112-201) has been determined sufficient for interactions of Kv1.So and Kv{31 subunits (this entry, see also Protein interactions under VLC K Kv1-Shak, 48-31). One study41 has further delineated an interaction site motif (FYE/ QLGE/DEAM/L, residues 193-201 in Kvo1.S) which has been (i) found only in N-termini of the Shaker-related Kv1 subfamily channel proteins and (ii) shown necessary for Kv{31-mediated rapid inactivation of Kv1.S currents. These results also indicated that hetero-oligomerization between 0 and Kv{31 subunits is restricted to Shaker-related potassium channel 0 subunits. (See also Protein interactions, 47-31 and next paragraph.)

The NAB domain mediates assembly of both KVQ-KVQ and KVQ-Kv{3 complexes 47-24-02: In an independent study, mapping of the Kv,L31-binding site to regions in cytoplasmic N-terminal domains of Kv10 subunits have determined that it overlaps with the region known as the NAB(Kvl) tetramerization domain1 (for details, see Protein interactions under VLC K Kv1Shak, 48-31). By means of channel chimera/deletion experiments, it has been shown that the NAB(Kv1) domain is essential for reproducing Kv{31mediated inactivation1. (See also Protein interactions, 47-31 and paragraph 47-24-01). 47-24-03: On the basis of primary sequence comparisons (see Fig. 2 under Encoding, 47-19) a number of potential post-translatory modification t sites exist on Kv{3 proteins including multiple phosphorylation consensus sites

II

_'--

e_nt_ry_4_7_1

(see Protein phosphorylation, 47-32) and for amidationt. The motif associated with protein amidation is conserved in all presently known Kv{3 subunits (indicated by a short horizontal line in the alignment shown in Fig. 2). For predicted consequence of subunit cleavage by amidation, see Protein molecular weight (calc.), 47-24. For background refs on peptide amidation, see Resource G - Consensus sites and motifs, entry 62.

Lack of leader and N-glycosylation motifs 47-24-04: As predicted for proteins with cytoplasmic locations, Kv{3 subunit sequences do not contain motifs for leader sequencest or N-glycosylationt (consistent with the deduced lack of added carbohydrate chains on Kv{3 subunits within experiments 9 described in Table 5). For conservation of structural motifs associated with Kv{3 sequences thought to be involved in NADPHf binding and the hydrogen transfer mechanism t in aldo-keto reductase t superfamily members, see also Channel modulation, 47-44 and Miscellaneous information, 47-55.


Amino acid composition 47-26-01: Hydropathyt analyses of Kv{3 subunits (e.g. in refs. 4,9) do not indicate the presence of hydrophobic stretches of sufficient length to act as transmembrane domains (i.e. at least 19 residues with an average hydropathy index t of >1.6 by the criteria of Kyte and Doolittle42 ). Kv{3 subunits appear to be peripheral membrane proteins with a 'dominant hydrophilic character' based on analyses by the method of Klein et. a1. 43 .

Domain arrangement 47-27-01: By the criteria of Garnier et a1. 44 the Kv,82 subunit appears to include four major a-helical t domains spanning residues 66-103, 126-140, 193-236 and 270-298 9 (for sequence, see Fig. 2 under Encoding, 47-19). Comparative note: This a-helical pattern has also been predicted for cloned ,8 subunits of voltage-gated calcium channels from skeletal muscle (see VLG Ca, entry 42).

Domain conservation Kv{31, Kv{32 and Kv{33 show high sequence conservation in their C-terminal portions 47-28-01: hKv,81b (== clone hKv{33): As illustrated in Fig. 2 (under Encoding, 47-19) homology between the presently known Kv,8 subunits is greater in their C-terminal regions, with 'significant differences' occurring in the Nterminal region. As pointed out in ref. 5, the C-terminal 329 aa of rKv{31a and hKv,81b (== clonehKv,83) are 1000/0 identical and share rv85% identity to rKv{32. Notably, however, the N-terminal 79 aa of rKv,81a and hKv,81b share only rv25 % identity, with this region also being 'difficult to align'

II

l'---e_n_t_ry_4_7

_

with the rKvj32 sequence5 (see Domain functions, 47-29). The point of divergence between Kvj3la and Kvj3lb (clone hKvj33 (see Fig. 2) has been suggested as a candidate alternative splice junction5 (see Subtype classifications, 47-06 and Gene organization, 47-20).

Molecular basis of selective interactions between Kva and Kv{3 subunits 47-28-02: Kvj31 and Kvj32 exhibit indistinguishable Kvla subunit selectivity in complexes containing Kval.l, Kval.2, Kval.3, Kval.5 or Kval.6 32. Reciprocal co-immunoprecipitationt studies have further indicated that selective interaction between 'compatible' Kva and Kvj3 subunits are mediated through conserved domains (analogous to the 'TI' or 'NAB' domain in Kva subunit N-termini - for details, see Protein intereactions under VLG K Kv1-Shak, 48-31). Notably, the interaction(s) of Kvj3 subunits with Kvl (Shaker-related) a subunits do not require the 13 subunit Nterminal domains (Le. mutants Kvj3l~N70, lacking 70 aa residues, and Kvj32~N22, lacking 22 aa residues)32. Note: These authors used these observations to argue that a previously observed failure of Kvj31 N-terminal mutants to modulate inactivation kinetics of Kvl family members could not be simply due to a 'lack of subunit interaction' (see also Protein interactions, 47-31).

Structural relationships between Kv{3 subunits and aldo-keto reductase superfamily members

47-28-03: Kvj3 subunits were originally described4 as showing 'no significant

homologies' to other proteins. Since their original description, however, sequence similarities to members of the NAD(P)H-dependent aldo-keto reductase t superfamilyt have been pointed out24,25 (see Miscellaneous information, 47-55). These analyses indicate that the 'most conserved' residues are those aligning to key secondary structure determinants (see Fig. 6, ibid.). Notably, many of the structural motifs known to be involved in NADPH binding and the hydrogen transfer mechanism in aldo-keto reductase t superfamily members are also conserved in mammalian Kvj3 and Drosophila Hk subunits 24 .

Conservation of primary sequences between Drosophila hyperkinetic and mammalian Kv{3 47-28-04: In general, the physiological effects seen when Hk is co-expressed

with Sh are distinct from those reported for either mammalian Kvj3la/b or Kvj32 subunits.

Domain functions The rKv{31a subunit N-terminal tball' domain confers inactivation properties upon rKvl.l 47-29-01: The inactivating 'ball' domain in the N-terminus of Kvj31 promotes

rapid closure of open Kv channels which cannot otherwise inactivate rapidly4 (see Phenotypic expression, 47-14). Strikingly, the oxidation or reduction of a ~critical cysteine' in the inactivation ball domain of Kvj3la reversibly

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_t_ry_4_7_

switches the channel inactivation mode from 'fast' to 'slow' and vice versa (for further details, see Channel modulation, 47-44, Fig. 3, this field, and ref. 4).

Comparison of rKv{31a and hKv{31b subunit N-termini 47-29-02: hKv,Bla; hKv,Blb (== clone hKvj33): hKv,Blb induces an inactivating property upon hKvl.5 delayed rectifier currents, but to a lesser extent than hKv,Bla on rKvl.l (see previous paragraph). Although the level of identity between rKv,Bla and hKv,Blb N-termini is low (ca. 250/0, see Domain conservation, 47-28) predictions based on the algorithms of Chou and Fasman suggest that rKv,Bla and hKv,Blb N-termini have 'nearly identical' secondary structures (cited in ref. 5). Furthermore, optimized N-terminal alignments (different to those shown in Fig. 2 under Encoding, 47-19) which introduce gaps and highlight conservative substitutions between rKv,Bla and hKv,Blb termini have helped identify residues that appear important for conferment of inactivating properties (as denoted and summarized in Fig. 3, this field).

Predicted protein topography Differences in mass between Kva and Kva/b complexes tbest fit' a a4{34 stoichiometry 47-30-01: Summations of 'accurate' protein molecular weight differences between a-DTx acceptor complex populations purified from bovine brain consisting of (i) major 'large' Mr range complexes of 370-420 kDa (containing 39 kDa ,B subunits) and (ii) minor 'small' range complexes of 240-265 kDa (lacking 39 kDa (3 subunits) are most consistent with native K+ channel oligomers containing four Kvn and four Kv,8 subunits (generally designated as 'Q4,B4 stoichiometric complexes' or '4a4,8 octamers'46). Note: Alternative combinations (such as Q4,B3, Q4,B2 or Q4,Bl) are incompatible with this data (see also Protein molecular weight (purified), 47-22). For the collective evidence supporting tetrameric assemblies of KVQ subunits, see the VLC K Kv series, entries 48 to 51 inclusive.

Protein interactions Significance of co-expressed Kv{3 subunits in accounting for tnative' K+ channel currents 47-31-01: The full 'variety' of protein subunit interactions which generate native cell K+ current diversity have only recently begun to be explored (see brief discussion within Channel designation of the entry VLC K Kv1Shak, 48-03). 'Correlation' or 'matching' of endogenous cell currents with heterologously expressed K+ channel Q subunit channel currents is not straightforward, since the latter show only a limited range of diversity (Atype r versus delayed rectifier t currents, often with similar pharmacological properties). The demonstration of functional heterotetrameric t Kv channel assemblies (ibid.) and the existence/association properties of Kv,8 subunits in vivo (this entry) can both profoundly alter functional properties and therefore must be taken into account in comparative studies. Most

1'--_e_n_t_ry_4_7

_

biochemical data are consistent with the {3 subunit being 'firmly associated' at intracellular sites with a subunits (as initially proposed9 ). There is no evidence for disulphide linkages between Kva/{3 subunits, but the Itight' non-covalent interactions in oligomeric subunit complexes are not disrupted by high salt concentrations or the vigorous procedures involved in co-immunoprecipitations using anti-a subunit antibodies (see below).

Direct evidence for the hetero-oligomeric constitution of Kv channel complexes in vivo 47-31-02: Characterization of seven monoclonal antibodies raised against adendrotoxin-sensitive K+ channels (purified from bovine cerebral cortex) first showed Kv{3 subunits to be distinct proteins (and not proteolytic fragments of the larger subunit)47. 'Authentic' subunit compositions of neuronal K+ channels purified from bovine brain have been analysed using monoclonal t antibodies reactive with Kva subunits (e.g. mAb 5 selective for Kv 1.2)46 or polyclonal t antibodies specific for fusion proteins containing C-terminal regions of several mammalian Kv proteins (ibid.). Western blottingt of K+ channel complexes from several brain regions using adendrotoxin (a-DTx, see Blockers, 47-43), show precipitation of several different Kva subunits in variable amounts according to brain region. Results from this type of investigation are summarized in Table 6, which also cites early descriptions of Kv{3 subunits co-purifying within native Kva/{3 complexes.

Kvla subfamily-selective interactions with Kvj3 subunits 47-31-09: Both the Kv{31 and Kv{32 have been shown to display 'robust and selective' interaction with five members of the Shaker-related (Kvl) a subunit subfamily when co-expressed in mammalian (COS) cells32 (Kvl.l, Kvl.2, Kvl.3, Kvl.5, and Kvl.6 - see VLC K Kvl-Shak, entry 48). Conversely, interaction of Kv,81 and Kv,82 with members of the Kv2 (Shabrelated) and Kv3 (Shaw-related) a subunit subfamilies could not be detected by immunoblot analysis. Note: Although a member of the Kv4, Shal-related subfamily, Kv4.2 was determined to interact with Kv{3 subunits in this study32. Notably, however, this interaction showed characteristics distinct from that typical between Kv{3-Kval family members: Co-immunoprecipitation t could not be disrupted by SDS at 0 mY; (ii) an 18 mV hyperpolarizing shift in the activation curve and (iii) slowed deactivation (7 == 8.0ms vs. 3S.4ms at -SOmV)5 (for further details, see legend to Fig. 4). Notes: 1. The hyperpolarizing shift described above would give rise to significant fractions of open channels at membrane potentials where hKvl.S alone would not normally be open (e.g. in the range -40 to -30mV).

II

l_e_n_t_ry_4_7

_

(a)

hKv 1.5

fr r - - - - -

-~--~/

~

40 msec (d)

(c)

C 1.0

~

Kvl.S + hKvP3~~-a

C 1.0

V I12=-24 mV

~

='

=' u

to)

+

+

~

]

.~ ~

~

0.5

~

0.5

.~

t;

E o

E o z

z

0.0 -80

-40

0

40

Membrane potential (m V)

o. 0

~

- - - - r - - , - - -__- - - .

r-j

0.0

0.5

1.0

1.5

2.0

Recovery Time (sec)

Figure 4. Functional expression ofhKv(31b : hKv1.5 channels. hKv(31b (= clone hKv(33): Voltage-dependent hKv1.5 current in oocytes in the absence (a) and in the presence (b) of hKv(31b shows a pronounced inactivating component induced by the (3 subunit. This effect is only observed at pulse potentials >OmV (currents shown were produced at a holding potential of -80mV with test pulse from -90 to +50mV in 10mV increments, followed by a return to -40mV). (c) The normalized current-voltage relation, calculated from deactivating tail currents at -40 m V and indicated by arrows in panels (a) and (b), illustrates the approx. 20mV hyperpolarizing shift in the activation curve observed in the presence (e) versus the absence (0) of hKvb1b. Differences in hKvl.5 recovery time following inactivation are also affected by the presence of hKv(31b: Following induction of inactivation (500ms pre-pulse to +50mV) currents were elicited by a step back to -BOmV for various time periods (d) followed by a test pulse to +50mV to test the degree of inactivation (measurement of peak normalized current flowing during the text pulse). As indicated in panel (d), recovery time courses were best fitted with an exponential function with a time constant of 51 ms. (Reproduced with permission from England et al. (1995) Proc Natl Acad Sci USA 92: 6309-13.) (From 47-37-02)

II

_

.

entry 47 "--------------

2. rKvl.l/hKv,Bla channels do not display any hyperpolarizing shift in the activation curve as seen with hKvl.S/hKv,Blb (ref. 4 and previous paragraph). 3. hKvl.S currents do not become fully inactivated by the presence of hKv,Blb (Fig. 4b) even after a 500 ms inactivating pre-pulse (Fig. 4d). Although non-stoichiometric expression ratio of a:,B subunits could be responsible for 'partial' inactivation (see paragraph on a:,B mRNA .ratios under Phenotypic expression, 47-14-05) there is presently no direct evidence to support this 5 . 4. Similar 'conversion' of a 'delayed rectifier type' to a 'rapidly inactivating type' channel has been described following co-expression of clone RBKVI.S (==rabbit Kv1.5, see VLG K Kv1-Shak, entry 48) with Kv,BI 5o. 5. Further functional characterization of Kvj31 subunits co-expressed with various K channel a subunits showed49 that (i) Kvj31 could induce inactivation in members of the Kvl subfamily with the exception of Kvl.6, while (ii) it could not induce inactivation of Kv2.I, Kv3.4-Ll2-28 and Kv4.1 channels49.

Kv{31b (clone Kv(33) alters Kvl.4 channel states which follow --activation 47-37-03: Kvj31b (== clone Kvj33): ferret Kval.4: When co-expressed in oocytes with a Kvl.4a subunit (clone FK1, see VLG K KV1-Shak, entry 48), clone Kvj33 has been shown51 to (i) accelerate both the fast and the slow component of Kvl.4 inactivation; (ii) alter the relative contributions of the two components of inactivation by increasing the contribution of a slow component to the inactivation process; (iii) slow recovery from inactivation for Kv1.4 (but not for a Kvl.4 deletion mutant lacking N-type inactivation) and (iv) slow deactivation kinetics. Steady-state activation (and the time course of Kvl.4 current activation) were not strongly influenced by co-expression of clone Kv,B3 51 .

Modulation of hKval.4 and Kval.5 current kinetics by clone hKv{33 in oocytes 47-37-04: Kvj31b (== clone hKvj33): hKvj33 subunits (cloned from human atrial mRNA, as designated in ref. 6 - see Table 1 under Gene family, 4705) also accelerate the inactivation of co-injected hKvl.4 currents and induce fast inactivation of non-inactivating co-injected hKvl.5 currents in oocytes 6 . Clone hKvj33 has no apparent effect on hKval.I, hKval.2 or hKva2.1 currents in this system. Note: Ra~ Kv,83 (as designated in refs 10,11, see Table 1 under Gene family, 47-05) also has a long N-terminal structure and induces inactivation in N-terminal deleted Kvl.4 (but not in other members of the Kvl channel family)ll. 47-37-05: Kvj3lb == Kvj33: Co-expression of Kvj3lb (designated Kvj33 in ref. 7) with K+ channel a subunits can increase the rate of inactivation from 4- to 7-fold in a Kvl.4 or Shaker B channels, with other kinetic parameters being unaffected 7. In contrast to Kvj3la (this field), Kvj3lb has no apparent effect on Kvl.l channel kinetics in oocytes 7 •

N-terminal peptide segments of Kv{31 alone cannot induce fast inactivation properties 47-37-06: Certain N-terminal peptides of the Kvj31 subunit alone are unable to mimic the increases in the rate of inactivation of whole-cell mKvl.1

II

_

1~_e_n_t_ry_4_7

currents (as observed with the entire Kv,81 subunit)52 (compare with (i) the ability of N-terminal pepides from fast-inactivating Kv channel a subunits to induce N-type inactivation in Inactivation under VLC K Kv1-Shak, 48-37 and (ii) the minimal interaction domain motif described under Sequence motifs, 47-24).

Kv(32 modulates the inactivation properties of Kvl.4a subunits 47-37-07: Kv,82: Although the Kv,82 subunit has been shown to increase

protein turnover and promote the transport of Kva to the cell-surface (see Subcellular locations, 47-16), it can also 'modulate' the inactivation properties (i.e. increase the inactivation rate) of Kval.4 when co-expressed in Xenopus oocytes 12. Notably, Kv,82 does not induce inactivation when coexpressed with certain non-inactivating (N-terminal truncated) Kvl.4 channel constructs or non-inactivating Kvl.l channels. Moreover both Kv,81 and Kv,82 are able to increase the amplitude of Kvl.4 currents expressed in Xenopus oocytes. Note: In a different study49, Kv,82 did not induce inactivation when co-expressed Kvla subunits; by means of Kv,81/ Kv,82 N-terminal chimaeras, however, it was shown that Kv,82 subunits do co-assemble with these Kva subunits. Notably, Kv,81/Kv,82 and Kv,83/Kv,82 chimaeras were able to induce fast inactivation of several Kvl channels, also indicating that Kv,82 associates with these a subunits49.

PHARMACOLOGY

Blockers High-affinity toxin-binding studies instrumental in the discovery of Kv(3 subunits 47-43-01: The use of a-dendrotoxin (a-DTx) and its homologue 6-dendrotoxin (8-DTx, exhibiting a subtly different specificity, see Dolly et al., 1994, under Related sources and reviews, 47-56) has been of central importance in

defining Kva/,8 subunit associations within native K+ channel complexes using immunoaffinity-binding methods (for description, see Isolation probe, 47-12). Historically, such approaches led to the first identification of K+ channel a subunits and were important for establishing K+ channels from large, hetero-oligomeric sialoglycoproteins t in native tissues (for references, see Protein molecular weight (purified), 47-22). In general, a-DTx efficiently inhibits fast-activating, voltage-dependent, aminopyridine-sensitive K+ currents that exhibit slow inactivation53,54. Related, but rapidly inactivating K+ channel variants in neuronal preparations have been described as 'less susceptible' or 'insensitive' to a_DTx36,55; see also Dolly et al., 1987, under Related sources and reviews, 47-56. Notes: 1. The presence of K+ channel,8 subunits may alter sensitivity to block by pyridines (see ref. 13). 2. {3Bungarotoxin binds with high affinity to a subpopulation of a-DTx sites56 . 3. As determined by site-directed mutagenesis t, a-DTx itself binds to residues close to the extracellular mouth of intact channel complexes, without any reported affinity for the constituent ,8 subunits themselves (see Protein interactions, 47-31 and refs in the VLC K Kv series, entries 47 to 52; also

II

_

...

e_n_t_ry---,-4_7_

compare to toxin binding properties of fJ subunits of calcium-activated K+ channels, under ILG K Ca, entry 27).

Utility of neurotoxin probes in channel expression/distribution and structure/function studies 47-43-02: In general, binding of the dendrotoxins and {3-bungarotoxin in brain

slice preparations reveal 'distinct, yet overlapping' patterns of distribution (e.g. as in rat, refs 57,58, see also Dolly et al., 1987, under Related sources and reviews, 47-56, and Protein distribution, 47-15). The ability to readily (i) modify rrimary structures of toxins for which the high-resolution crystal structures are known and (ii) purify large quantities of active recombinant toxins from bacterial hosts may help refine 'subtype-specific' immunoaffinity probes for native K+ channel complexes (see review; ref. 59).

Channel modulation tOxidoreductive' modulation of Kv{31 N-terminal tball' domain/ inactivation properties 47-44-01: Alignment of the N-terminal domain of Kv{31 subunits with those of Kvo subunits that undergo rapid N-terminal (N-type) inactivation demonstrates some broad structural conservation (illustrated in Fig. 3 under Domain functions, 47-29). The inactivating ball domain in the N-terminus of Kv{31 promotes rapid closure of open Kv channels which cannot otherwise inactivate rapidly (ref. 4 , see Phenotypic expression, 47-14 and Inactivation, 47-37). The redox state (Le. oxidation or reduction) of a Icritical cysteine' in the inactivation ball domain of Kv{31a reversibly switches the channel inactivation mode from 'fast' to 'slow' and vice versa4 : When serine is substituted for cysteine at position 7 (Kv{31C7S, see Fig. 3 under Domain functions, 4729) the sensitivity of (3 balls to oxidation is eliminated. This result is similar to that observed when 0 ball cysteines (ibid.) are replaced by serine60 and suggests that such substitutions 'destabilize' the interaction between the ball and its receptor bordering the channel pore. Figure 5 summarizes additional experiments which contributed to the proposal of inactivating {3 ball domains which are sensitive to redox modulation. Notes: 1. The term '{3 ball' was introduced to distinguish these domains from 'a ball' domains of Kvo subunit channels that undergo fast inactivation. 2. Inactivation induced by Kv{33 (as designated in refs 10,11, see Table 1 under Gene family, 47-05) is also regulated by the intracellular redox potential11 . 47-44-02: Comparative note: For further potential modulations of of{3 channel

complexes by oxidoreductive biochemical pathways, see Miscellaneous information, 47-55. For other examples of ion channel modulation in relation to cellular redox state, see refs 45,61 and Channel modulation under ELG CAT GLU NMDA, 08-44, ILG CI ABC-CF, 23-44, ILG K Ca, 27-44 and INR K ATP-i, 30-4462 .

Cytochrome P-450 inhibitors reduce native Kv currents in pulmonary arterial myocytes 47-44-03: Comparative note: Cytochrome P-450 (P-450) is an NADPH-

requiring and O2 -dependent mono-oxygenase system expressed in lung

II

Il..--e_n_t_ry_4_7

-----'_

(a)

(c)

(b)

red

RCK4

RCK4~1-110

RCK4~1-IIO

(e)

(d)

(f)

+ KvPI

200 ms ... ox -red

RCK4~1-110

RCK4~1-110

RCK4~1-110

+ KvPIC7S

+ Kv~1~1-34

+ peptide

Figure 5. 'Switching' of inactivation mode in Kvl.4 (RCK4) channels by oxidation or reduction of Kv{31 N-terminal ball domains. For background to experiments see paragraph 47-44-01. (a) For outward currents elicited by depolarizing pulses (to +50mV from an h.p. of -100mV in inside-out oocyte patches) N-type inactivation properties are lost following exposure of wild-type Kvl.4 channels to oxidative treatments (ox., by addition of 0.1 % H2 0 2 to the bath). N-type inactivation is restored following reducing treatments (red., by addition of 5 mM glutathione to the bath). (b) N-type inactivation and redox sensitivity are lost following removal of the Kvl.4 ball domain, as exemplified by the Kvl.4 deletion mutant RCK4~1-110, lacking the first 34 N-terminal amino acids 4 ,63. (c) Co-expression of RCK4~1-110 and Kv{31 subunits restores a rapidly inactivating current under reducing conditions (7[ == 4.2 ± 0.8ms). (d) Traces obtained when wild-type Kvl.4 is co-expressed with Kvf31 subunits are indistinguishable from those in (c). (e) Kv{3 subunits lacking their first N-terminal 34 amino acids (Kv{31~1-34) cannot confer fast-inactivating properties when coexpressed with RCK4~1-110. (f) 'Partial' restoration of inactivation to RCK4~1-110 channels is observed when a peptide comprising aa 1-24 of the Kv{31 N-terminus is applied to the bath (Kv{31-N wt ; 100 jjM). When the N-terminal peptide containing a CED ~ SCN substitution with increased net positive charge is similarly applied (Kvf31-N SGN ; 100 jjM; for residues see Fig. 3) a fast inactivating current results which is comparable to that seen with intact Kv{31 subunits. Note: The 'residual' slow current decay in the absence of N-type inactivation (e.g. under oxidizing conditions in panels a and c) is attributable to C-type inactivation involving the channel pore structure (see refs 64 ,65 and Inactivation under VLC K KvlShak, 48-37). (Reproduced with permission from Rettig et a1. (1995) Nature 369: 289-94.) (From 47-44-01)

II

_1....--

e_n_t _ry_4_7

-----J

which may act as an '02 sensor' in hypoxic pulmonary vasoconstriction. In pulmonary arterial myocytes, the P-450 inhibitors clotrimazole, miconazole, and 1-aminobenzotriazole (l-ABT) 'significantly and reversibly inhibit' native steady-state voltage-gated K+ currents and depolarize the cells bathed in either Ca2+-containing (1.8 mM) or Ca2+-free bath solution66 • These data may support a role for the P-450 system in linking the regulation of pulmonary vascular tone to the alteration of cellular redox status via its influence on Kv currents and sensitivity to O2 tension and NADPH66 . Notes: 1. Pre-treatment of pulmonary arterial smooth muscle with 4-aminopyridine (4-AP, 10 mM) but !lot tetraethylammonium prevents the subsequent inhibitory effect of P-450 inhibitors on Kv currents. 2. The effects of the P-450 inhibitors resemble those induced by hypoxia, reduced glutathione, and 4-aminopyridine.

Ligands Radioligands for co-precipitation of Kv{3 subunit acceptor complexes 47-47-01: Availability of radioiodinatedt derivatives of a-DTx, 8-DTx and {3bungarotoxin has allowed high-affinity acceptor complexes to be identified, purified and characterized in mammalian brain (for further details, see Isolation probe, 47-12 and Blockers, 47-43).

Receptor/transducer interactions 47-49-02: Although no studies have been reported regarding the functional regulation of Kv,8 subunit components coupled to specified receptor/second messenger systems, the known patterns of protein phosphorylation (see field 32, Protein phosphorylation) and possibly redox modulation (field 44, Channel modulation) might predict their existence. Hypothetically, receptor-coupled Kv{3 subunit modulation may influence in vivo phenotypes explicable in terms of those demonstrated for recombinant Kv{3 subunits in heterologous expression systems (e.g. see Phenotypic expression, 47-14, Subcellular locations, 47-16 and Inactivation, 47-37).

INFORMATION RETRIEVAL

Database listings/primary sequence discussion 47-53-01: The relevant database is indicated by the lower case prefix (e.g. gb:) which should not be typed (see Introduction etJ layout of entries, entry 02). Database locus names and accession numbers immediately follow the colon. Note that a comprehensive listing of all available accession numbers is superfluous for location of relevant sequences in GenBank® resources, which are now available with powerful in-built neighbouringt analysis routines (for description, see the Database listings field in Introduction etJ layout of entries, entry 02). For example, sequences of cross-species variants or related gene familyt members can be readily

II

"'--.e_n_t_ry_4_7

_

accessed by one or two rounds of neighbouring t analysis (which are based on pre-computed alignments performed using the BLASTt algorithm by the NCBlt ). This feature is most useful for retrieval of sequence entries deposited in databases later than those listed below. Thus, representative members of known sequence homology groupings are listed to permit initial direct retrievals by accession number, unique sequence identifiers (Seq ID: numbers) author /reference or nomenclature. Following direct accession, however, neighbouring t analysis is strongly recommended to identify newly reported and related sequences. Kv,B designation (this entry)

Species, original isolate name

ORF

cDNA source

Sequence/ discussion

rKv,81a

Rat Shaker ,81 subunit4

ORF: 401 aa rat gb: X70662 brain cortex cDNA library

Rettig, Nature (1994) 369: 289-94.

=

rKv,81

gb: not found McCormack, FEBS lett (1995) 370: 32-6.

Human brain cDNA Kv,81 clone3

hKv,81a

Accession

hKv,81b hKv,83

=

clone

Human, ,83 subunit clones

ORF: 408 aa cardiacgb: L39833 ventricular cDNA

England, Proc Nati Acad Sci USA (1995) 92: 6309-13.

hKv,81b hKv,83

=

clone

Human, ,83 subunit clone 7

ORF: 408 aa cardiac atrial cDNA

gb: U16953

Majumder, FEBS Lett (1995) 361: 1316.

gb: U17968

Morales, TBiol Chern (1995) 270: 6272-7.

ORF: 408 aa gb: U17966 (Musteia putorius)

Morales, TBioi Chern (1995) 270: 6272-7.

rKv,81b = clone Rat cDNA 7 rKv,83 partial seq.

399 nt segment

gb: U17967

Morales, TBioi Chern (1995) 270: 6272-7.

hKv,81c

ORF: 419 aa

gb: L47665

England, TBioi Chern (1995) 270: 28531-4.

ORF: 367 aa bovine Bas prirnigenius (aurochs) brain cDNA

gb: X70661 (as gb) cIte as X70662 in ref. shown at right

Scott, Proc Natl Acad Sci USA (1994) 91: 1637-41.

hKv,81b = clone Human, cDNA 7 400 nt segment hKv,83 partial seq. fKv,81b fKv,83

=

=

clone

Ferret heart, ,83 subunit 7

Kv,81.3 Human cardiac ventricular cDNA8

bKv,82

Bovine Shaker ,82; 'major' component of a-dendrotoxinsensitive K+ channels 9

hKv,82

Human brain cDNA Kv,82 clone3

Yi

gb: not found McCormack, FEBS Lett (1995) 370: 326.

II

_L..... Kv{3 designation (this entry)

e_n_t_ry_4_7_

Species, original isolate name

rKv,82 or RCK,82 Rat Shaker ,82 subunit 9

ORF

Accession

cDNA source ORF: 367 aa rat gb: X76724 brain cortex cDNA library

Sequence/ discussion Scott, Froc Natl Acad Sci USA (1994) 91: 1637-41.

rKv,83

Rat brain Induces partial gb: not found Heinemann, Kv,83 10,11. Shares inactivation in Biophys T(1995) 68% identity channels of the Kvl 68: A361. with rKv,81 subfamily Heinemann, FEBS Lett (1995) 377: 383-9.

dHk,8

Drosophila, ORF: 546 aa adult gb: U23545 hyperkinetic (Hk)head cDNA 24 gene locus

Chouinard, Froc Natl Acad Sci USA (1995) 92: 6763-7.

Note: As part of the,8 subunit homology searches described in ref. 45 (see Miscellaneous information, 47-55), several partial cDNA clones were identified from rice and Arabidopsis (gb: D24756 j gb:Z30863 j gb: Z18389 j gb: D24673, of unknown function). These clones showed >40-60% identity with the Shaker ,8 subunits over their entire length of up to 96 aa.

Gene mapping locus designation 47-54-01: For chromosomal locus names that have been used (KCNA1B and KCNA2B) but are under review, see the notes under the field Chromosomal location, 47-18 (see also online resources such as OMIM for confirmation of

locus designations).

Miscellaneous information Cross-references to studies of accessory ({3) subunits associated with other channels 47-55-01: Accessory proteins or {3 subunits have been shown to modulate voltage dependence of gating, current amplitudes and inactivation kinetics of voltage-gated Ca2 + and Na+ channels (see refs 67- 77 and also relevant fields of VLC Ca, entry 42 and VLC Na, entry 55). Structural and functional features of a non-pore forming, 'accessory' {3 subunit of 31 kDa associated with pore-forming (maxi) K ca channel a subunits (e.g. refs. 78-82) are described in Protein molecular weight and Predicted protein topography under ILC K Ca, 27-22 and 27-30 respectively. An Arabidopsis thaliana cDNA encoding a 38.4 kDa polypeptide with homology to the mammalian Kv{3 subunits has been described40 . Kv{3 subunits possess unrelated primary sequences to all of the 'accessory' subunits indicated above, but do exhibit general structural similarity in that they are predominantly basic t (pI ~ 9.5) and hydrophilic, reflecting their common location at the cytoplasmic face of the membrane 83 . (see also Related sources and reviews, 47-56).

II

----J_

l..--e_n_t_ry_4_7

A4 B 8 C 8

il

~1

L

H 41 I 75

S1 V T PNSSAQQ 1SIINIPFAE NLGKS RVSCL NLGKS RVSCL

J 1 K 2

SQARPATVLGAMEMlRRMDV----TSSSASVRA-----FLQ~TE1~If~GQS~1LIPLGLGLG

D 10 E 5 F 13 G 7

I

P3

f34

U3

loop A

~ A 63

~

B 70 C 70 D 72 E 64

74 66 108 142 1 K 63

F G H I J

-----W -----W

----------------------------------------

RSI-----CKVK1ATlAAp----MFGKTLKlADIRFQIET

------l I

U4

A B C D E F G H

125 DESGNVVPSDTN1L 132 DEHGKLLFETVD1C 132 DENGK1LFDTVDLC 134 DENGRV1YHKSNLC 127 NADGT1CYDSTHYK 136 EVEDLLPFDV---KG 117 -----------RWL 166 ------------ME I 200 ------------MEE1V J 24 ------------LEEOM K 117 ------------1ESDLQ

U6 A B C D E F G H

193 200 S 200 202 193 NE 199 172 223

A B C D E F

232 240 240 244 232 234 202 290 324 141 243

I

~ I

KS1 KS1 KS QA TKA1 RS1

~

I

I

E~!li

L

I EEF L NPF FSF I L L PSVG------FAN PNTI------FAN SNT ------1FYSIRVDATT ------FY~FIDHGT -------

Us

~

I

r-

HLQVEIKPG--LKY CRQLE KPG--LKY RRQLE KPG--LKY RRQLE KPG--LKY I SRQ1D SVAS----V SVKKLQN SVAT----1 TEPMLKTL1DETG----V WSSME1MEAYSVARNFNL1 F -QR WSAME1MEAYSVARQFNM1 CE F-QR SYSPELTQKAAA1LKEER-VPLF1H PNYNMF-NR SWEVAEDCTLCKKNGW1MlTVY MYNA1-TR

II~

.a

loop B

1QYCQSK----I1VVTIY SP-DRPWAKPEDI----------------ILIDP---------LDYCKSK----D11LVSYCT SSRDKTWVDQKS ---------------- LDDP---------DFCKSK----D1VL SHREEPWVDPNS ---------------- L DP---------EVSASSM---TSF1 TCRNPLWVNVSS P--------------L---------SS-DRAWRDPDE ---------------- L P---------1AHCQAR----IEVT EFCKEN---- 11VT RKGASRGPN---------------------E D---------FHDEH---- 1RTES RRS--------------------------~L Q---------EVNLPELFHK1 GAMT CG1VSGKYDSG~PYSRASLKGYNWLKDK1IS RRQQA---KL I 257 EVQLPELYHK1 GAMT CG11SGKYGNG ESSRASLKCYQWLKER1VS RKQQN---KL J 80 W1ENGLLDTLG-E1 GC1VF QGLLTSRYLNG1 GARANQGGS--LKASAQ LL------GR1 GGLLTGRYKYQDKDGKNPESRFFGNPFSQLYMDRYWKEEHFNG1A K 174 QVETELFPCLRH-F RFYlF

Us

G H I

J K

UH2 ~

EISQDMTTL Q SEDMKAL Q SEEMKA1 S KEEMKD1 TFSPEEMKQL EQDHHK1 W D ADQVDA1 1GA1QVLPKLSSS11H GA1QVLPKMTSHVVN ALVEEGPLEPAVVDA

loop C

---j ----------:....---------1 LSIRVCALLS---------CTSHKDlPFHEEF DG FRYNNAKY---------FDDHPNHPFTDE DG RYLTLD1---------FAGPP~PFSDEY EA RFVEMLM---------WSDHP PFHDEY NA RY1VPMLTVDGKRVPRDAGHP PFNDPY SQ1SQSRL1SGPTK---PQLADLWDDQ1 259 SGLERGRLWDGD-----------------PDTHEEM 350 E1DS1LGNKP--------------YSKKDlRS I 384 E1Ql1LRNKP--------------YSKKQlRS 1-1

A B C D E F G H

289 297 297 300 289 291

J 183

K 311 FDQAwlLVAHECP---------------NlFR

Figure 6. Amino acid alignment of selected putative NADPH-dependent aldo-keto reductase superfamily members including Kv{3 subunits. Alignments are related to known secondary structure elements of aldose reductase (top line A) as annotated.

II

_L.-

e_n_t_ry_4_7_

Key to Fig. 6 alignment: Sequence Protein (selected putative in NADPH-dependent aldo-keto alignment reductase superfamily members)

Species

Database reference (sp: Swissprot; gb: Genbank)

Aldose reductase 3-a-hydroxysteroid reductase Trans-1,2-dihydrobenzene-1,2diol dehydrogenase 3-oxo-5-~-steroid 4 dehydrogenase Aldehyde dehydrogenase NAD(P)H-dependent 6'deoxychalcone synthase 2,5-diketo-D-gluconic acid reductase Shaker ~2 subunit Shaker ~1 subunit igrA gene product (potential, partial sequence) Aflatoxin B1 aldehyde reductase

Human Rat Human

sp: P15121 sp: P23457 sp: Q04828

Rat

sp: 31210

Human Soybean

sp: P14550 sp: P26690

Corynebacterium Bovine Rat Pseudomonas

sp:P15339 gb: X70661 gb: X70662 gb: M37389

Rat

gb: X74673

A

B C D

E F G H I

T K

Annotations: ~1 to ~8: Segments forming ~ strands in aldose reductase. al to a8: Segments forming a helices in aldose reductase. Loops A to C: Loop regions in aldose reductase. Loops aHl and QH2: Two additional C-terminal helices. Numberings indicate portions of original amino acid sequences that were used in the alignment. Predominant amino acid identities are indicated as white lettering on black (see also Database listings, 47-53). Note: Other superfamily proteins with noted homologies24 to Kv~ and Hk proteins include bovine prostaglandin F synthase and lens crystallin from frog. Similar homology alignments (using a number of different algorithms) have resulted in significant matches with over 60 putative superfamily members45 . (Reproduced with permission from ref.45.) (From 47-55-02)

Structural relationships within the aldo-keto reductase superfamily 47-55-02: Structural analyses of Drosophila Hyperkinetic sequences (see Chouinard et al., 1993, under Related sources and reviews, 47-56 and ref.24) and amino acid homology searches for Kvf3 subunit homologues across many phyla45 have identified structural similarities of Kv~ subunits to members of the NAD(P)H-dependent aldo-keto reductase t superfamily t. This gene family encompasses a 'structurally similar but functionally diverse' group of cytosolic enzymes that all use NADPHt as a cofactor. Figure 6 shows an alignment of selected putative NADPH-dependent aldo-keto reductase superfamily members in relation to the known secondary structure of aldose reductase 84 . Similar homology alignments to Fig. 6 (using a number of different algorithms) have resulted in significant matches with over 60 putative superfamily

II

1__e_n_t_ry_4_7

-----..J_

members45 . Although percentage amino acid identities between Kv{3 and aldoketo reductase superfamily members are relatively low (typically 15-30%) this partly reflects the wide species coverage of sequences in alignments (identical residues or conservative substitutions occur within several key structural determinants - see Fig. 6). Notably, many of the residues known to be involved in NADPH binding and the hydrogen transfer mechanism are conserved in both mammalian Kv{3 and Drosophila Hk subunits24 . While it is premature to draw 'functional' conclusions on the basis of alignment data alone, it has been tentatively suggested that (3 subunits might be functional oxidoreductase enzymes. If this is the case, it might imply that (i) K+ channels are modulated through the redox state of NADP(H) or NAD(H) in vivo and/or that (ii) the {3 subunits themselves might be regulated by K+ channel activity45. For further discussion of these hypotheses, see refs24,25; see also Channel modulation, 47-44 and the annotated legend to Fig. 6.

Related sources and reviews 47-56-01: Reviews on structural and functional aspects of Kv{3 subunits2,59; minor part of cardiac potassium channel molecular physiology review85; see also the commentary on modification of channel inactivation properties86; auxiliary subunits of other voltage-gated ion channels83; commentary on calcium channel {3 subunits87 (see also VLC Ca, entry 42 and VLC Na, entry 55).

Book references: Chouinard, S.W., Schlimgen, A.K. and Ganetzky, B. (1993) Abstract. In Neurobiology of Drosophila, p.96. Cold Spring Harbor Laboratory Press, Plainview, NY. Dolly, J.D., Stansfeld, C.E., Breeze, A.L., Pelchen-Matthews, A., March, S.T. and Brown, D.A. (1987) In Neurotoxins and their Pharmacological Implications (ed. P. Jenner), pp. 114-16. Raven Press, New York. Dolly, J.D., Munitz, Z.M., Parcej, D.N., Hall, A.C., Scott, V.E.S., Awan, K.A. and Owen, D.G. (1994) In Neutrotoxins and Neurobiology (eds K.F. Tipton and F. Dajas), pp. 103-22. Ellis Horwood, Chichester.

Feedback Error-corrections, enhancements and extensions 47-57-01: Please notify specific errors, omissions, updates and comments on this entry by contributing to its e-mail feedback file (for details, see Resource T- Search criteria). For this entry, send e-mail messagesTo:[email protected]. indicating the appropriate paragraph by entering its six-figure index number (xx-yy-zz or other identifier) into the Subject: field of the message (e.g. Subject: 47-24-03). Please feedback on only one specified paragraph or figure per message, normally by sending a corrected replacement according to the guidelines in Feedback etJ CSN Access. Enhancements and extensions can also be suggested by this route (ibid.). Notified changes will be indexed from within the CSN website (www.le.ac.uk/csn/).

II

_L....REFERENCES Yu, Neuron (1996) 16: 441-53. Pongs, Semin Neurosci (1995) 7: 137-46. 3 McCormack, FEBS Lett (1995) 370: 32-6. 4 Rettig, Nature (1994) 369: 289-94. 5 England, Proc Natl Acad Sci USA (1995) 92: 6309-13. 6 Majumder, FEBS Lett (1995) 361: 13-16. 7 Morales, TBiol Chem (1995) 270: 6272-7. 8 England, TBiol Chem (1995) 270: 28531-4. 9 Scott, Proc Natl Acad Sci USA (1994) 91: 1637-41. 10 Heinemann, Biophys T (1995) 68: A361. 11 Heinemann, FEBS Lett (1995) 377: 38~3-9. 12 McCormack, FEBS Lett (1995) 370: 32-6. 13 Rudy, Neuron (1988) 1: 649-58. 14 Chabala, T Gen Physiol (1993) 102: 713-28. 15 Parcej, Biochem T(1989) 264: 623-4. 16 Parcej, Biochemistry (1992) 31: 11084·-8. 17 Muniz, Biochemistry (1992) 31: 12297-303. 18 Scott, T Biol Chem (1990) 265: 20094-·7. 19 Trimmer, Proc Natl Acad Sci USA (1991) 88: 10764-8. 20 Rhodes, TNeurosci (1996) 16: 4846-60. 21 Kues, Eur T Neurosci (1992) 4: 1296-308. 22 Dixon, Eur TNeurosci (1996) 8: 183-91. 23 Uebele, T Biol Chem (1996) 271: 2406--12. 24 Chouinard, Proe Natl Aead Sci USA (1995) 92: 6763-7. 25 Stern, TNeurogenet (1992) 8: 157-72. 26 Stern, T Neurogenet (1989) 5: 215-28. 27 Yao, Soc Neurosci Abstr (1995) Abstract 121.11. 28 Wang, Soc Neurosci Abstr (1995) Abstract 121.10. 29 Wilson, Soc Neurosci Abstr (1995) Abstract 121.8. 30 Rhodes, TNeurosci (1995) 15: 5360-71. 31 Shi, Neuron (1996) 16: 843-52. 32 Nakahira, T Biol Chem (1996) 271: 7084-9. 33 Schultz, Genomics (1996) 31: 389-91. 34 Black, Biochem T (1986) 237: 397-404. 35 Black, Biochemistry (1988) 27: 6814-20. 36 Dolly, T Physiol (Paris) (1984) 79: 280-303. 37 Mehraban, FEBS Lett (1984) 174: 116-22. 38 Cai, T Biol Chem (1993) 268: 23720-7. 39 Prystowsky, FASEB T(1996) 10: 2504. 40 Tang, Plant Physiol (1995) 109: 327-30. 41 Sewing, Neuron (1996) 16: 455-63. 42 Kyte, TMol Bioi (1992) 157: 105-32. 43 Klein, Biochim Biophys Acta (1985) 815: 468-76. 44 Garnier, TMol Biol (1978) 120: 97-120. 45 McCormack, Cell (1994) 79: 1133-5. 46 Scott, Biochemistry (1994) 33: 1617-23. 1

2

II

e_n_t _ry_4_7_1

1'--_e_n_try_4_7

47 48

49 50 51

52 53 54

55 56

57

58 59

60

61 62 63

64 65 66

67

68 69 70

71 72 73

74 75 76 77 78

79 80 81

82 83 84

85 86 87

---'_

Muniz, Biochemistry (1992) 31: 12297-303. Schuhmann, FEBS Lett (1994) 341: 208-12. Heinemann, TPhysiol (1996) 493: 625-33. Sasaki, FEBS Lett (1995) 372: 20-4 (correction in Vol. 379: 202). Castellino, Am T Physiol (1995) 38: H385-91. Stephens, FEBS Lett (1996) 378: 250-2. Stansfeld, Neurosci Lett (1986) 64: 299-304. Stansfeld, Neuroscience (1987) 23: 893-902. Halliwell, Proc Natl Acad Sci USA (1986) 83: 493-7. Breeze, Eur T Biochem (1989) 178: 771-8. Awan, Neuroscience (1991) 40: 29-39. Pelchen-Matthews, Brain Res (1988) 441: 127-38. Dolly, Biochem Soc Trans (1994) 22: 473-8. Ruppersberg, Nature (1991) 352: 711-14. Lee, FEBS Lett (1992) 311: 81-4. Islam, FEBS Lett (1993) 319: 128-32. Hoshi, Science (1990) 250: 533-8. Lopez-Barneo, Recept Channels (1993) 1: 61-71. Hoshi, Neuron (1991) 7: 547-56. Yuan, Am T Physiol (1995) 37: C259-70. Dewaard, T Physiol (1995) 485: 619-34. Dewaard, Neuron (1994) 13: 495-503. Isom, T BioI Chern (1995) 270: 3306-12. Patton, T BioI Chern (1994) 269: 17649-55. Isom, Science (1992) 256: 839-42. Messner, T BioI Chern (1985) 260: 10597-604. McHugh-Sutkowski, T BioI Chern (1990) 265: 12393-9. Makita, Genornics (1994) 23: 628-34. Makita, T BioI Chern (1994) 269: 7571-8. Yang, Neuron (1993) 11: 915-22. Bennett, FEBS Lett (1993) 326: 21-4. Garcia-Calvo, T BioI Chern (1994) 269: 676-82. Knaus, T BioI Chern (1994) 269: 3921-4. Knaus, T BioI Chern (1994) 269: 17274-8. McManus, Neuron (1995) 14: 645-50. Tseng-Crank, Proc Natl Acad Sci USA (1996) 93: 9200-5. Isom, Neuron (1994) 12: 1183-94. Wilson, Science (1992) 257: 81-4. Deal, Physiol Rev (1996) 76: 49-67. Aldrich, Curr BioI (1994) 4: 839-40. Bean, Nature (1994) 368: 15-16.

II

Vertebrate K+ channels related to Drosophila Shaker (Kva: subunits encoded by gene subfamily Kvl) Edward C. Conley

Entry 48

Note on coverage: Features of Kvo. subunits forming K+ -selective, voltage-gated channels in heterologous cell expression systems are listed within this and the following three entries according to their gene family relationships (VLG K Kv1Shak, entry 48; VLG K Kv2-Shab, entry 49; VLG K Kv1-Shaw, entry 50 and VLG K Kv1-Shal, entry 51). The roles of Kv(3 subunits in extending the structural and functional diversity of voltage-gated K+ channels by association with specified Kvo. subunits are described under VLG K Kv-beta, entry 47. General properties applicable to voltage-gated K+ channels in native cells (largely where the subunit composition has not been determined) are summarized under VLG K A-T native ('transient outward' K+ channels, entry 44) and VLG K DR native ('delayed rectifiers', entry 45). Inwardly rectifying, K+ -selective channels formed from subunits encoded by the Kir gene family are described under the entries beginning INR K (entries 29 to 33 inclusive). Because of their distinct functional or structural properties, separate entries have also been assigned to voltage-gated K+ channels in the eag family (VLG K eag/elk/erg, entry 46), calcium-activated K+ channels (ILG K Ca, entry 27), sodium-activated K+ channels (ILG K Na, entry 28), channels underlying native cell M-current (\lLG K M-i native, entry 53), and 'minimal' K proteins (VLG [K) minK, entry 54). In all entries, there is a bias of coverage to vertebrate K+ channel types - see special note under Category (sortcode), field 02.

NOMENCI-JATURES

Abstract/general description 48-01-01: Vertebrate K+ channel subunits whose primary amino acid sequences are most closely related to those of Drosophila Shaker are grouped in the voltage-gated K+ channel subfamily 1 (Kvl, this entry). In mammals, the Shaker-related gene subfamily consists of seven characterized isoforms (Kvl.l to Kvl.7); initial reports of further Shaker-related isoforms have appeared (see Gene family" 48-05). Shaker and Shaker-related channels probably represent the most intensively studied group of K+ channels. 48-01-02: Most Kv channels display marked heterogeneous expression patterns, many of which appear subject to developmental control (for specific examples, see Developmental regulation, 48-11). Spatial- and temporal-selective induction of Kv channel activities in development (developmental heterogeneity) presumably reflects functional specialization within particular 'developmental compartments' or terminally differentiated cell types (for discussion, see TUN [connexins), entry 35). Localization (segregation) of different Kv proteins to specific subcellular domains may also significantly affect native channel properties, particularly where they are co-localized with modulatory proteins. Much indirect evidence exists for developmental co-regulation of genes encoding multiple channel types channel and channel-modulatory proteins (i.e. 'co-ordinated' multigene expression).

II

1'--_e_n_t_ry_4_8

_

48-01-03: Kvl subfamily channel developmental ontogeny has been studied extensively in Xenopus, taking advantage of well-characterized morphological differentiation of identified neurone subsets (and the opportunity for comparing channel subtypes versus native currents this offers). For example, within the embryonic nervous system, transcripts encoding the Kvl.l homologues XShal and (at a lower level) XSha2 (a Kvl.2 homologue) are detectable. Furthermore, the onset of neurogenesis in Xenopus (stage 13) is associated with the induction of Kvl.2 gene expression (isolate XSha2) (see Developmental regulation, 48-11). Well-characterized development models involving Kvl subfamily channels also exist for non-excitable cells, for example Kvl.l and Kvl.3 in lymphoid thymocyte precursors (ibid.). 48-01-04: Several novel aspects of gene expression control have been described for Kvl subfamily members. These include (i) multiple hormonal regulation of Kvl.S gene transcription (and protein induction) in a tissue-specific manner (e.g. TRH or dexamethasone); (ii) sensitivity to 'cold stress' responses; regulation by cAMP and K+-induced depolarization (the former mediated by a cAMP-response element t (CRE) in the S/-non-coding region of the rKvl.S gene interacting with CRE-binding protein (CREB); (iii) regulation by chronic morphine administration; (iv) cell-specific negative regulation of rKvl.S by a gene silencer t element and (v) heterologous positive regulation of the native hKvl.S gene by a locus control region or 'gene activation element' (for details of each of these, see Developmental regulation, 48-11.) 48-01-05: A number of approaches have been used to deduce 'function' of a given K+ channel subunit in vivo, and in most cases this is not a simple task (see Phenotypic expression, 48-14). 'Clues' to function may be derived from, for example, naturally occurring mutant Kvl channels in populations, perhaps associated with an inherited disease (as for mutant Kvl.l/episodic ataxia, ibid.); non-random subcellular distributions or subunit co-assemblies (ibid.); specific functions suggested by a limited Kv1 subtype expression profile in a specified cell type (e.g. Kv1.3 in T cells, microglia and osteoclasts, ibid.); transgenic overexpression (i.e. 'forcing' expression of a Kv channel in a particular cell type, as in the case of Kv1.5 in pancreatic j3 cells, ibid.) or gene knockout t approaches. Much debate has centred upon the difficulties of 'predicting' specified Kv subunit 'contributions' to voltage-gated K+ currents observed in native cells (for discussion, see Channel designation, 48-03 and in VLC K Kv-beta, entry 47). Despite these difficulties, many authors have pointed out 'notable similarities' between native and heterologously expressed channel properties (see Table 4 under Phenotypic expression, 48-14). 48-01-06: Many cell types express voltage-gated K+ channel subpopulations that are characterized by different functional and pharmacological properties Generally, Kvl subunit genes show distinct but overlapping expression patterns within brain and other tissues. However, the expression pattern of some Kvl subunit mRNAs in native tissues appear ubiquitous. Tabular summaries of reported tissue distributions for Kvl subfamily mRNAs by techniques such as in situ hybridization, Northern hybridization, RNAase protection assay and RT-PCR appears in Table 3 under mRNA distribution, 48-13.

II

_ entry 48

'-------------

48-01-07: In general, Kv channel subunits in native cell types are expressed at 'low densities', with more than one subtype co-expressed per cell- factors that have made 'direct' purification methods for specific Kv channel proteins technically formidable without toxin immunoaffinity methods (see next paragraph, Protein molecular weight (purified), 48-22 and Blockers, 48-43). Notably, however, certain membrane-associated putative guanylate kinases such as PSD-95, SAP97, chapsyn-110 and hdlg have been shown to promote differential Kv channel clustering by associating their PDZ domains with Cterminal residues in Kvl.l, Kvl.2, Kvl.3 and Kvl.4. Chapsyn-IIO associates tightly with the post-synaptic density in brain and mediates the clustering of both Kv channels and NMDA receptors (entry 08) in heterologous cells (see Channel density, 48-09). 48-01-08: Native Kv channel assemblies immunoprecipitated with highaffinity toxins such as a-dendrotoxin (o-DTx) are octameric (a4,84) sialoprotein complexes (for details, see VLG K Kv-beta, entry 47). Subtypes of a-DTx acceptor complex in bovine brain using a monoclonal antibody selective for Kv1.2 (mAb 5) and polyclonal antibodies for other Kvo subunits indicate that Kv1.1, Kv1.4 and Kv1.6 can be components of the acceptor, while 'virtually all' DTx receptors contain the Kvl.2 0 subunit (see Protein molecular weight (purified), 48-22). Anti-Kv antibodies have been important tools for understanding events in early channel biosynthesis and assembly, defining specific subunit interactions (e.g. within heteromultimers) and in measuring channel Mr values (ibid.). 48-01-09: Mechanisms contributing to the origin of the Kv multigene family include many different types of genetic duplication, recombinational exchange and transfer mechanism that have operated during evolution. In particular, localized, tandem gene duplication (followed by sequence divergence) is associated with the generation of clusters of functionally related genes, and this appears typical for Kv channel gene arrangements (Fig. 1 and text under Chromosomallocat:ion, 48-18). Large-scale duplications and/or rearrangements (e.g. affecting whole chromosomes or substantial segments of them) can also generate contiguous segments of linked genes that have a corresponding (paralogous t) region of related, linked genes elsewhere in the genome. This 'clustered' distribution of several groups of Kv channel genes in the mouse (ibid.) has confirmed and extended the existence of such paralogous regions. Genomic studies such as these have allowed proposal of evolutionary time-scales for ]{v gene duplication and divergence from primordial types (Fig. 1). 48-01-10: In Drosophila, K+ channel subtype heterogeneity is determined by alternative splicingt of the limited set of K+ channel genes; in the vertebrate Shaker-related gene subfamily (all single-copy genes), most protein-coding regions are intronless (the exception being Kvl.?, which has a single intron interrupting the structural gene in the region encoding the Sl-S2 extracellular loop - see Gene organization, 48-20). A summary of intron sequences found in Kvl subfamily (within non-coding regions) is listed in Table 9. A potential role in control of channel mRNA stability conferred by ATTTA and ATTTG motifs in the 3' untranslated region of the Kvl.4 gene (KCNA4) has been described (ibid.).

II

1L....-__ _ _ry_4_8

en t

_

48-01-11: Prior to the availability of high-resolution K+ channel molecular structure information494, modelling approaches reliant on interpretation of structure-function t data (e.g. 'iterative' testing of structural predictions by means of site-directed mutagenesis t followed by functional analysis of mutant proteins) had been a dominant approach. These models are subject to continual refinement, and the numbers of protein motifs (field 24) involved with designated 'functions' are increasing, often following detection of motif homologies between protein families. The 'delineation' of named domains and their description within the Kv1 family are described under fields such as Domain conservation, 48-28, Predicted protein topography, 48-30 and Protein interactions, 48-31. Structural elements for homophilic interaction between Kv1 subunits have been extensively characterized (ibid.). 48-01-12: Phosphorylation of KVQ subunits (and/or their accessory protein components) can alter characteristics such as channel current amplitude, voltage-dependence of gating and the kinetics of activation (see Protein phosphorylation, 48-32). Phosphorylation of K+ channels is thus an important general mechanism for modulating calcium entry, action potential firing patterns (threshold, frequency, height, width) and 'effector' responses coupling membrane excitability to secretion, muscular contraction or gene transcription (see Phenotypic expression, 48-14). Activation of neurotransmitter receptors commonly alter excitability and synaptic efficacy by generating intracellular second messengers, with several second messengers acting through protein kinase and phosphatase proteins that alter properties of K+ channels. Specific examples of Kv1 subfamily phosphomodulation are described in Table 16 under Protein phosphorylation, 48-32. 48-01-13: The molecular mechanisms of channel activation and inactivation following displacement of voltage-sensing domains has been most extensively studied in Drosophila Shaker channels (see Activation, 48-33, Inactivation, 48-37 and Voltage sensitivity, 48-42). For example, there is an extensive literature on the origins of gating current t (i.e. movements of electronic charge across the membrane field during voltage activation), Ntype inactivation (field 37) and C-type inactivation (ibid.). These topics are further discussed within structured tables under these fields. 48-01-14: In physiological ionic gradients, Kv channels show high selectivity for K+ over Na+ ions. A large number of studies (Table 21 under Selectivity, 48-40) have determined that ion selectivity functions in voltage-dependent K+ channels are predominantly associated with a segment of 21 contiguous residues known as the P-region (i.e. the pore region between domains S5 and S6). The P-region, along with the S6 segment and the S4-S5 linker appear to contain most of the pore determinants. The P-region is the only recognizably conserved segment in all K+ -selective channels, irrespective of subfamily or species (ibid.). Kv1 subfamily channels have unitary conductances in the range 8-18pS (see Table 22 under Single-channel data, 48-41). 48-01-15: A large number of studies have reported properties of peptide toxin, ionic and pharmacological blockers of Kv 1 subfamily channels in heterologous expression systems (for overvie~ see Blockers, 48-43). High-throughput screening methods have begun to yield compounds with claimed high

II

_____________________ en_t_ry_4_8_

selectivity and potency against Kvl channel subfamily members. In addition, sophisticated 'toxin docking' approaches (based on assignment of 'pairwise' amino acid interactions between blocker and target) have helped define spatial arrangements of Kv channel residues (ibid.). Scanning mutagenesis t (see Glossary) can identify sets of inhibitor residues critical for making energetic contacts with the channel, using thermodynamic mutant cycle analysis. Issues of selectivity, relative potency and mechanisms of block (listed by blocker and by target) are discussed within structured tables under Blockers (field 43). 48-01-16: A number of physiological and pharmacological factors/agents that have been cited as modulating the electrophysiological characteristics of KvI subfamily channels, including arachidonate, Ca2 + ions, K+ ions and redox state are discussed under Channel modulation, 48-44. The large size of the Kv subunit family (and the considerable number of potential couplings to receptor/second messenger systems) predict a considerable number of receptor/transducer/effector combinations operating within native cells. Despite the large potential number of couplings, only relatively few have been demonstrated directly (see Table 28 under Receptor/transducer interactions, 48-49). 48-01-17: Homologues of voltage-gated K+ channel genes are now known to exist in an extraordinarily wide evolutionary range of organisms from Escherichia through Streptomyces, ciliate protists (e.g. Paramecium), Arabidopsis, jellyfish, worms, squid and (as most extensively characterized) flies and vertebrates. The Escherichia coli sequence predicts a protein 417 residues long with 'extensive similarity' to eukaryotic proteins in amino acid sequence, six apparent transmembrane regions with a P-region motif between S5 and S6. Given that there is lack of direct evidence that the Escherichia K+ channel gene was 'imported' from genomes of higher organisms, its existence supports the view that the Shaker K+ channel subfamily was 'functionally established' prior to the first major radiation of

metazoans.

Category (sortcode) 48-02-01: VLG K Kvl-Shak, Le. a subunits of vertebrate voltage-gated potassium channels encoded by Kv gene subfamily 1 whose primary (amino acid) sequences show highest identity to those of Drosophila Shaker (see Gene family, 48-05). Homomultimeric t channels formed from Kvla subunits are normally named in accordance with the cDNA name (see Channel designation, 48-03). Conventional names of Kv gene (chromosomal) loci are listed under Gene mapping locus designation, 48-54.

Special note: Bias of coverage to vertebrate K+ channels 48-02-02: The fundamental and continuing importance of studies in Drosophila for delineation of K+ channel gene family and structurefunction relationships cannot be underestimated, and several comparative features are listed in the entry. For reasons of space, however, this entry excludes many of the known properties specifically derived for Drosophila

II

lL....--_ _ _ry_4_8 en t

_

and other non-vertebrate Shaker-subfamily channels (but see references in Related sources and reviews, 48-56).

Problems of 'single entry' coverage for large Kv channel families 48-02-03: 'Grouping' of channels using the criterion of primary sequence relatedness (for the purposes of print publication, as in the Kvl-Kv4 subfamilies, entries 48 to 51) has both advantages and disadvantages: Advantages include the ability to compare homomeric channel properties applicable in heterologous t expression systems and 'common' subfamily features. Disadvantages include the large number of apparent 'variables' between studies and the lack of a simple method to index properties applicable to heteromultimeric subunit combinations in native and heterologous cells; these difficulties are not unique to K+ channel subunit families.

Information sorting/retrieval aided by designated gene family nomenclatures 48-02-04: The gene product prefix (used as a 'unique embedded identifier' or VEl) for 'tagging' and retrieving information relevant to this entry on the CSN website will be of the form VEl: Kvl.n where n is a designated !!umber in the systematic nomenclature (e.g. Kvl.4). Within this entry, paragraph 'running orders' (sort orders) are largely determined alphanumerically by systematic nomenclatures - i.e. denoting species and gene product prefix combined with any trivial or clone name{s) where these have been used in the source reference (e.g. rKvl.4/RCK4:). Where properties are likely to apply to all or several subfamily proteins (i.e. irrespective of species or isolate) the 'species' term may be omitted. Entry updates covering novel isoforms will index information using these conventions established from systematic nomenclatures based on gene family relationships.

Channel designation 48-03-01: Vertebrate voltage-dependent K+ channel genes/cDNAs/channels related to Drosophila Shaker are assigned to Shaker-related subfamily 1 (this entry), presently designated as Kvl.l to Kvl.7 by the published nomenclature 1,2. Species 'equivalents' (see next paragraph) are indicated in a Kv name by a species prefix (e.g. human Kvl.l, rat Kvl.l, mouse Kvl.l) or conventionally by their initial letter (e.g. hKvl.l, rKvl.l, mKvl.l). For descriptions of vertebrate K+ channel subunit genes segregating into homology groups defined by Drosophila Shab, see VLC K Kv2-Shab, entry 49; for Drosophila Shaw-related Kv genes, see VLC K Kv3-Shaw, entry 50 and for Drosophila Shal-related Kv genes, see VLC K Kv4-Shal, entry 51. Strictly, Kv genes t (i.e. chromosomal sequences) should be referred to by their gene locus designation (see Cene mapping locus designation, 48-54) although in practice 'Kv numbers' are used to conveniently 'specify' genes, mRNAs, cDNAs, cRNAs and protein subunit variants arising from or associated with expression of a given Kv gene product or locus. Note: Two additional mammalian cDNAs for which Drosophila homologues have not (presently) been identified, are tentatively classified {in the absence of

II

_L.-

e_n_try_4_8_

functional expression) as cDNA IK8 as K(v)5.1 and cDNA K13 as K(v)6.1 (for a brief account, see VLC K Kvx, entry 52). This nomenclature conflicts with a designation used for the Aplysia clone aKv5.1 (for a brief description of clone aKv5.1, see Abstract/general description under VLC K Kvx, 52-01).

Intended use of the term tisoform' within this entry

48-03-02: Properties of specific isoforms t with the same tKv number' are

intended to designate genes/protein subunits that are 'so closely related' at the primary sequence level that they might be judged 'equivalent' in different vertebrate species. The presumption of physiological 'equivalence' is open to debate, since channel function and modulation in situ (i.e. within the 'same' tissue of different species) may critically depend on the coexpression of other proteins which may have arisen since speciation (i.e. evolutionary separation) occurred. These complex issues, including the evolutionary antecedents of vertebrate K+ channels, assemblies of 'functional motifs' in primary sequences and the conservation of the Shaker, Shal, Shab and Shaw subfamilies as 'essential sets' of excitable t channels have been reviewed3,4 (see also Cene family, 48-05).

Parallel designation of tsystematic' and tnon-systematic' (isolate/ clone) names 48-03-03: In addition to the systematic nomenclatures based on sequence homology groupings (Table 1), non-systematic (original 'isolate' or 'clone')

names may persist in use for descriptive or 'discriminatory' purposes. Independently isolated (and named) subunit designations are listed in Table 1 under Gene family, 48-05. Subtype- or 'isoform-specific' information is indicated in this entry using an underlined prefix (e.g. RCKl:) within appropriate fields. In an attempt to preserve a 'logical' running order, properties are mostly (but not exclusively) listed in 'ascending' Kv number (coupled to the clone/isolate name, e.g. Kvl.l/RCKl:).

Other specific designations in occasional use 48-03-04: (i) In cases where it is important to distinguish Kv gene loci from Kv

gene products conventional gene locus designations can be used (see Gene mapping locus designation, 48-54). (ii) Specific designations of native and heterologously expressed K+ channel subunit complexes may have to take account of known subunit associations/stoichiometries that occur in cells (e.g. see Channel designation under VLG: K Kv-beta, 47-03 and paragraph 47-04-05). (iii) In the special case of tandem linkage of cDNAs encoding different Kv subunits (within a single open reading frame t construct) additional nomenclature may be needed to specify subunit order/position and predicted composition (i.e. A-B-A-B versus A-A-B-B etc.). Furthermore, phenotypic t expression of site-directed mutations t in one (or more) of the tandemly linked subunits may show some subunit 'position-dependence'. It should also be noted that tandem linkage of K+ channel subunits may not guarantee stoichiometric t relationships of the expressed channels5 (described under Protein interactions, 48-31).

II

1l.-_e_n_t_ry_4_8

_

tMatching' native channel currents with those of heterologously expressed Kv subunit channels 48-03-05: Several factors underline the difficulty of 'matching' properties of native cell voltage-gated K+ channels to the 'contribution' of individual Kv subunits: Heterologous t expression of single Kv subunits (i.e. in the specified absence of endogenous channel subunits capable of co-assembling with the heterologous component) is generally assumed to occur via formation of homomultimerict (i.e. homotetramerict) channels6 . Relatively few 'native' K+ channel currents can be 'accounted for' in this way (for

possible exceptions, see paragraph in Phenotypic expression, 48-14). Functional heteromultimerict channels can be assembled from subunit components belonging to the same Kv subfamily (e.g. between members of the Kvl subfamily) but not between members of different subfamilies within the same tetrameric complex (for details and mechanisms, see Protein interactions, 48-31). Despite this, 'closely related' channels from the same subfamily, even when expressed in the same cell, need not coassemble and as a result may be differentially localized within the cell (for an example, see Subcellular locations, 48-16). Furthermore, the presence of certain Kv,B subunits can significantly influence the inactivation, modulation and expression properties of voltage-gated K+ channels formed from Q subunits alone (for details, see VLC K Kv-beta, entry 47). Heterologous expression systems may not 'reproduce' significant modulatory conditions found in native cells (e.g. those dependent on co-expressed receptor/ transducer protein activities, cytosolic modulators of channel function or other 'critical' post-translatory modifications t). Finally, there are numerous examples of native/heterologous current 'matching' which have not anticipated 'contributions' from novel channel subunits that were 'uncloned' at the time of analysis.

Current designation 48-04-01: For currents conducted by homomultimeric t channel assemblies in heterologous cells, this could follow the shorthand form I Kvn .n , where n.n is the Kv subfamily number of the gene encoding the monomer, although this type of designation is rarely used.

Gene family An tessential set' of K+ channel subunit genes conserved from flies to humans 48-05-01: The Kv channel Q subunit gene family constitutes the most extensive and diverse group within the voltage-gated channel superfamilyt (see VLC key facts, entry 41). Drosophila Shaker, Shal, Shab and Shawrelated sequences 7- 9 define four subfamilies of voltage-gated K+ channels which are conserved in vertebrates 3,10 (subfamily Kv1, this entry, and subfamilies Kv2, Kv3 and Kv4, entries 49 to 51 inclusive). Conservation of this subfamily structure over such a wide 'evolutionary gap' suggests a common relationship to ancestral gene(s) in existence prior to the divergence of mammals and insects (see refs. 3 ,4 and paragraphs describing

III

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_8_

the putative K+ channel gene kch in the bacterium Escherichia coli, under Miscellaneous information, 48-55). The similarities between multiple genes in the Sh superfamily indicates that ancestral gene(s) encoding K+ channels underwent extensive duplication and subsequent variation (selection) in the chordate lineage (generating a large number of closely related genes). The availability of lcomplete' genomic sequences for several prokaryote organisms (together with increased knowledge on gene homologue relationships in lower and higher eukaryotes in coming years) may clarify these 'adaptive' functions of Kv and other gene products to their various physiological roles in differentiated cell types. Note: To indicate their relatedness to the Sh superfamily t of genes, some authors have used the designation ShI to designate mammalian homologues of Shaker. Selected references describing the original isolation and characterization of the Drosophila Shaker K+ channel gene products can be found under Related sources and reviews, 48-56. For a clarification of gene family relationships of the KATl and AKTl channels from Arabidopsis thaliana (which share some structural features with the Kv subfamilies) see Domain conservation under VLG K eag/elk/erg, 46-28).

The vertebrate Kvl gene subfamily: channel subtype diversity 48-05-02: Whereas in Drosophila K+ channel subtype heterogeneity is determined by alternative splicingt of the limited set of K+ channel genes, vertebrates primarily use multiple distinct genes to encode channel subtypes (although several alternative splicing events are known within the Kv gene family, affecting both non-coding and coding regions (for known examples, see Gene organization, field 20 of entries; particularly in VLG K Kv3-Shaw, 50-20). A listing of vertebrate Shaker-related genes/cDNAs encoding voltage-gated K+ channel subunits, with 'systematic' and 'isolate' (clone) names is given in Table 1.

A gene family tree for Kvl subfamily members 48-05-03: Figure 1 shows a Shaker subfamily phylogenetic tree including some non-mammalian channel sequences, based on the 1994 analysis of Chandy and Gutman (1994, see Related sources and reviews, 48-56) and an earlier analysis by the same authors 55 . The later analysis groups Xenopus Xshal with Kvl.l and Xsha2 with Kvl.2.

Basic criteria for Kv subfamily assignments 48-05-04: Generally, the central hydrophobic core of vertebrate Kv channels can be readily identified as homologous t to that of a channel encoded by one of the Drosophila genes, although there is less conservation of sequence in the C-terminal and N-terminal cytoplasmic regions. As a general guide, members within the same subfamily are ''"''700/0 identical' at the amino acid sequence level, whereas members among different subfamilies show '4050%' sequence identity8. Note that tpercentage homologies' or identities within a group of channel sequences may be influenced by the number of channel sequences included in the analysis, the regions of proteins chosen for alignment (i.e. total versus hydrophobic tcore' sequences), species of origin, and the sequence alignment algorithm employed (see Resource D

II

I

entry48

_

~------

- Diagnostic tests, entry 59). Conserved subfamily lcompatibility' determinants in the N-terminal region of Kv channel have been subjected to extensive structure/function analyses (for details, see Protein interactions, 48-31).

Subtype classifications Kv channel classifications/nomenclatures are subject to modification 48-06-01: Voltage-gated K+ channel genes reported after July 1991 usually have followed the 'Kv nomenclature' as an essential part of their description1 . Existing classifications of Kv channels may require 'ongoing refinement' (e.g. to accommodate new isolates which do not show sufficient amino acid homologyt for classification as members of the Shaker/Shab/Shaw/Shal subfamilies). This is exemplified by the putative K+ channel cDNA isolates K13 and IK8 (briefly described under VLC K Kvx, entry 52) which appear to be representative of novel K+ channel gene subfamilies. The 'relatedness' of presently known Kv1 subfamily isolates is indicated under Cene family, 48-05.

Earlier 'classifications' of voltage-gated K+ channels are inadequate 48-06-02: As further described in VLC K A-T [native], entry 44 and VLC K DR [native], entry 45, it is likely that the use of the terms 'A-type' and 'delayed

rectifier' will assume less significance for voltage-gated K+ channel classification: These divisions cannot account for the known subunit heterogeneity expressed in native cells and defined hetero-oligomeric arrangements of Kva subunits (channel-forming, this entry and entries 49 to 52 inclusive) complexed with Kv,8 subunits (see VLC K Kv-beta, entry 47) may be able to account for many observed inactivation properties at the level of specified native cell types (ibid.).

Trivial names 48-07-01: The Shaker-related voltage-gated K+ channel subfamily. Kv1a

subunits; pore-forming Kv1 subfamily channels. Other 'trivial names' of Kv channel subunits/genes/cDNAs are indicated in Table 1 under Gene family, 48-05.

EXPRESSION In order to compare 'similarities and differences' information has been collated in tabular format for several fields for each of the subtypes of K+ channels mostly by 'ascending' Kv number. Note, however, that many 'independently reported' results for Kv isoforms t show only minor or no differences, thus 'full listing' of data 'by isolate name' is superfluous. Where apparent discrepancies exist, or where specific information might be useful, data are listed under both the Kv nomenclature 1 and the original gene/subunit ('clone') name. Source references for 'sequence-related'

II

II

Table 1. Cene family nomenclature and original names in use for vertebrate voltage-dependent K+ channel cDNAs/subunits related to Drosophila Shaker, subfamily 1. cDNA 'clone' names (isolates) in common use are indicated in bold. See also VLC (key facts, entry 41) and Subtype classifications, 48-06. (From 48-05-02) Kv designation

Human isoforms

Rat isoforms

Mouse isoforms

Other species

Kvl.l Names in use and refs (notes 4 and 5)

hKvl.1 HuK(I)11-13

rKvl.1

mKvl.1 MBKl 18 MKl 16,19

Other Hamster Kvl.1 16 Rabbit Kvl.1 2o Xenopus Xsha1 21

Kvl.2 Names in use and refs

hKvl.2 HuK(rIV)11-13

rKvl.2 BK2 15 RCKS 22 NGK1 23 RK2 17 RAK24 RH1 25

mKvl.2 MK2 19 MKS 16

Other Bovine Kvl.2 (BGKS)26 Rabbit Kvl.2 2o Xenopus Xsha221 Canine Kvl.2 (CSMKl)27


hKvl.3 HuK(III)11,12 hPCN3 28 HLK3 29

rKvl.3 RCK3 22 RGKS 31 Isolate KV3 (see

mKvl.3 MK3 16,19,33

Other Rabbit Kvl.3 2o; see also Rabbit kidney I glibenclamide-sensitive' clone rabK (vl.3)34

mKvl.4 MK4 16

Other Bovine Kvl.2 (BAK4)39 Canine Kvl.2 (dKvl.2)27 Rabbit Kvl.4 2o (see note 4) Ferret Kvl.4 (FKl)4o

(see note 7)

RBKl 14 BKl 15 RKl 16,17

HGKS 30


RCKl 10

hKvl.4 HuK(II)11-13 hPCN228,35 HK1 36

note 2)32 rKvl.4 RCK4 22 RHK1 37 RK4 16,38 RKS 16

(tl

l:j t""t-

~ ~

00


II

hKvl.5 HuK(VI)11,12 HK241 hPCN1 28 HCK1 42 fHK 43

rKvl.5 Isolate KVl (see note 2)32 RK3 38 RMK2 44

mKvl.5 mKvl.5 45


hKvl.6 HBK248 HuK(V)11,12

rKvl.6 RCK2 48 Isolate KV2 (see note 2)32

mKvl.6 MK2 16 MK6 49


hKvl.7 Kvl.7 50,51 (abstracts) (diabetic (3 cells)

rKvl.7 RK6 16

mKvl.7 Other MK6 16 Hamster Kvl.7 16 52 MK4 mK(v)1.7 (see note 8)

K(v)1.8 (see note 6)

rK(v)1.8

rK(v)1.8

mK(v)1.8 Other mK(v)1.8 (see note 9)

Other Hamster Kvl.5 16 Bovine BAK5 46 Rabbit Kvl.5 (RBKVl.5)47

Other Hamster Kvl.6 16

Notes: 1. If the isolate (clone) name is used to identify a data source in this entry, the Kv nomenclature 1 also appears. Amino acid sequence alignments for these proteins are shown under Encoding, 48-19. 2. To ensure there is no confusion between Kv subfamily numbers and certain clone (isolate) names (KVl, KV2 and KV3)1,2, all occurrences of the latter (capitalized) are preceded by the word 'isolate'. 3. An 'RCK2' clone named in the original RCKI paper10 has now been re-named RCKlai RCK2 is defined as the rat brain 'equivalent' of the human brain K+ channel HBK2 (nomenclature Kvl.6). 4. A series of partial Shaker-related sequences have been amplified by RT-PCR t from rabbit kidney cortex and the renal epithelial cell line LLC-PK 1 20 (see Channel density, 48-09). A further partial Shaker-like cDNA (KC6, similar to Kvl.2/RBK2) from rabbit kidney cDNA was later reported53, bringing the total number of detectable Kv cDNAs in kidney to six. One of these isoforms (KC22) appeared to be a novel K-channel gene that is highly expressed in the rabbit kidney and in primary cell cultures of rabbit distal tubule and likely to be involved in epithelial K+ transport53 (see also clone 'rabK(v1.3)' described under Phenotypic expression, 48-14).

~

::s

t"1"

~ ~

00

n

5. Only supplementary or contrasting information for channel/equivalents' are listed in entries - these data are quoted under isolate names to enable identification and comparison of source data (see INFORMATION RETRIEVAL section). 6. Genes encoding K+channels which are expected to be voltage-gated (but which have not been experimentally confirmed) are generally designated by a K(v) prefix. 7. The tabulation is not exhaustive, and subunit genes not indicated or incorrectly designated can be communicated to the entry e-mail feedback file (see Feedback, 48-57). 8. Cited as mouse K(v)1.7, K. Kalman et al. unpublished data, in Chandy and Gutman, 1994, under Related sources and reviews, 48-56. 9. Cited as mouse K(v)1.8, B. Tempel (ibid.); a partial cDNA has been isolated from mouse cochlea by RT-PCR; chromosomal mapping (see field 48-18) and genomic sequencing of the mKvl.8 region is cited as in preparation within ref. 54).

(1)

::s

t"1"

~ ~

00

l_e_n_t_ry_48

_

mKv1.1 (MK1) rKv1.1 (RBK1) hKv1.1 (HUK[I)) ~---xKv1.1 (XSHA1) mKv1.2(MK2) rKv1.2(RGK5) hKv1.2(HUK[IV)) dKv1.2 bKv1.2(BGK5) ~---xKv1.2(XSha2) mKv1.3(MK3) rKv1.3(RGK5) hKv1.3(HPCN3) rKv1.5(KV1 ) hKv1.5(HPCN1 ) mKv1.6(MK6) rKv1.6(RCK2) hKv1.6(HBK2) ' - - - - - - - - mKv1.7 mKv1.4 rKv1.4(RCK4) hKv1.4(HPCN2) bKv1 .4(BAK4) ~-----APLK ~--------Shaker

Figure 1. Shaker subfamily phylogenetic tree. (Reproduced with permission from Chandy and Cutman (1994) In Handbook of Receptors and Channels (ed. R.A. North). CRC Press, Boca Raton.) (From 48-05-03)

features can be found under Database listings/primary sequence discussion, 48-53. In the case of missing or incorrect data, see the Feedback field, 48-57.

Cell-type expression index Cell-type expression patterns implied using molecular probes specific for Kv1 family members are described under Cloning resource, 48-10; Isolation probe, 48-12; mRNA distribution, 48-13 and Protein distribution, 48-15. Kv channels are ubiquitously expressed in excitable cells 48-08-01: K+ -selective channels are generally considered to be expressed in all vertebrate cells, although marked differential expression is observed at the cell type level (e.g. in terms of channel subtypes, co-assembly properties, relative densities and 'timing' of developmental onset). These variables in part account for the remarkable functional diversity of voltage-gated K+ channels in native vertebrate cells, including their major roles in regulation of membrane potential in excitable t cells such as nerve and muscle (see also VLC K A-'J: entry 44 and VLC K DR, entry 45). Within single cell types, there is direct evidence for large numbers of co-expressed Kv gene family members (e.g. in the PC12 phaeochromocytoma cell line model, where nine separate gene family products have been detected56; notably, this study could not use the full complement of presently known probes for Kv channels).

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_nt_ry_4_8_

Diversity of cardiac Kv channel expression 48-08-02: In the heart, voltage-gated K+ channels incorporating Kv subunits are likely to contribute to modulation of action potential frequency and duration (see also the INR K series, entries 29 to 33 inclusive, VLC K eag/ elk/erg, entry 46 and VLC K Kv-beta, entry 47). Molecular cloning t studies have shown multiple (>7) Kv subunit genes to be expressed in rat or human heart from the Kvl, Kv2 or Kv4 subfamilies (not including the KvLQTl- or erg-related families) (see mRNA distribution, 48-13 and under VLC K Kv2Shab, 49-13 and VLC K Kv4-Shal, 51-13).

Voltage-gated K+ channel expression is not restricted to excitable cells 48-08-03: Notably, a number of voltage-gated K+ channels are known to be expressed in non-excitablet cells (i.e. those cells which do not pass action potentials, such as lymphocytes) although no studies have reported Kv channel subunit expression in native differentiated erythrocytes. Voltagesensitive K+ channels are also generally pot observed in patch-clamp studies of native epithelial cells. Note, however, sensitive RT-PCRt techniques have revealed several Shaker subtype-specific mRNAs2o,53 in rabbit kidney cortex and the renal epithelial cell line LLC-PKl, implying such channels do exist in these tissues at very low densities (see also mRNA distribution, 48-13).

Heterologous cell expression of Kv channels 48-08-04: Functional studies of ion channels or other signalling proteins encoded by cDNAs or genes require that the proteins be correctly assembled in a transmembrane confi~ration. This is usually accomplished by the injection of messenger RNAt, cRNAt, or by stable transfection t of DNA into heterologous cells which do not normally express the protein(s). For stable expression, simple 'driving' of expression by strong tissue-specific promoterst and enhancerst does not guarantee optimal functional expression: 'Correct' folding and membrane insertion of channels may also require co-expression of accessory subunits or chaperonet functions in the heterologous system (e.g. see effects of Kvf32 in Phenotypic expression under VLC K Kv-beta, 47-14). Stably expressed channel activity may also be susceptible to endogenous kinase/phosphatase regulation, which in turn may be influenced by growth factor signalling, developlnental factors or cell cycle position. While these technical problems remain, heterologous cell expression methods (used in combination with voltage-clampt or patchclampt analysis of currents) have been especially powerful in the identification of structure-function relationships in ion channels. Other 'cell-free' approaches designed to overcome potentially 'confounding' influences of endogenous ion channels, neurotransmitter receptors and receptor-channel subunits have been described for Shaker channels57.

Limitations of Kv channel analysis by heterologous cell expression 48-08-05: The subunits of the Kv series probably underlie the properties of a 'significant proportion' of voltage-gated K+channels expressed in native tissues (see mRNA expression, 48-13, but see also VLC K eag/elk/erg, entry 46 and VLC {K} minK, entry 54). The 'spatial distribution' of channel

II

entry48

_

I" - - - - - - - - - -

subunit-specific mRNAs presumably reflects some 'functional specialization' in particular cell types, tissue regions or systems: Distribution patterns of 'kinetically similar' delayed rectifier K+ channels are distinct in brain58 with the implication that 'delayed rectifiers' and other channels possess functional differences that may be impossible to infer from heterologous expression studies in isolation (see also (i) note on 'matching native current properties' under Channel designation, 48-03 and (ii) the significant roles of accessory subunits in native channel complexes under VLG K Kv-beta, entry 47). Methodological note: Application of single-cell RT-PCRt eDNA amplification methods combined with patch-clamp methods (e.g. ref. 59, as applied to AMPA receptor-channels, see ELG CAT GLU AMPA/KAIN, entry 07) may help determine subunit 'contributions' to native cell currents. Products of at least nine distinct Shaker subfamily genes have been shown to express in single cells56, see also 60 . Alternatively, or in addition, use of clonal cell lines where known channel isoforms are expressed (e.g. hKvl.3 ~ type n channel of Jurkat T lymphocytes61 ) are also important models for channel genotype-phenotype studies.

Channel density Channel clustering at the plasma membrane by PDZ domain-Kv C-terminal interactions 8-09-01: Kv1 subfamily: The membrane-associated putative guanylate kinases PSD-95, SAP97, chapsyn-ll0 and hdlg have been shown to promote differential Kv channel clustering by associating their PDZ domains with Cterminal residues in Kvl.l, Kvl.2, Kvl.3 and Kvl.4 (for details and refs, see Protein interactions, 48-31 and Fig. 7). PSD-95 induces plaque-like clusters of K+ channels at the cell surface, while SAP97 co-expression results in the formation of large, round intracellular aggregates into which both SAP97 and the K+ channel proteins are co-localized. The 'efficiency' of surface clustering by PSD-95 has been shown to vary with different Kvl subunits: whereas 'striking' Kvl.4 clustering occurs in >600/0 of co-transfected cells, Kvl.l and Kvl.2 form clusters with PSD-95 at low efficiency (approx. 100/0 of cells)62. Chapsyn-ll0 associates tightly with the post-synaptic density in brain and mediates the clustering of both Kv channels and NMDA receptors (entry 08) in heterologous cells 63 . In native rat brain, chapsyn-ll0 protein shows a somatodendritic expression pattern (see note 1) that overlaps partly with PSD-95 but that contrasts with the axonal distribution of SAP97 (note 2). Chapsyn-ll0 and PSD-95 may form heteromultimers at post-synaptic sites to form a 'scaffold' for clustering of receptors, ion channels, and other signalling proteins 63 . Notes: 1. Compare Kv4.2's predominantly somatodentritic localization64 under Subcellular locations, 51-16. 2. Compare Kvl.4's predominantly axonallocalization64 .

Comparative difficulty of purifying K+ channel proteins to homogeneity from native tissues 48-09-02: In general, K+ channel subunits are expressed at Ilow densities', with more than one subtype co-expressed per cell. These factors make 'direct' purification methods for specific K+ channel proteins technically formidable:

II

____________________ en_t_ry_4_8-----J

Whereas 200- to SOO-fold purification can be sufficient to purify Na+ or Ca2+ channel protein to homogeneity from brain65,66, K+ channel subunits require a 10000- to 20 OOO-fold purification for equivalent purity67. In some cases, toxin immunoaffinity t methods have been successfully employed for purification of some classes of K+ channel protein (for reviews see refs 68,69 - see also VLG K Kv-beta, entry 47).

Novel methods for active K+ channel protein production and assay from heterologous cells 48-09-03: Although there is no 'notably rich' source of native vertebrate cell K+ channel protein (see previous paragraph), high-level expression (",,10 7 Shaker K+ channels/transfected cell) driven by an adenovirus promoter t has been suggested as a 'plausible route' for purification of 'milligram-level' amounts of functional K+ channels from heterologous cells 70 . Reconstitution of functional Shaker channels from an insect expression system following their immunoaffinity purification has also been described 71 . Note: This and a previous study 72 described the application of the proton pump bacteriorhodopsin as a light-driven current source for the development and control of transmembrane potentials in reconstituted vesicles. Bacteriorhodopsin is capable of depolarizing the membrane (to at least 0 mV level) within a few seconds, while more rapid depolarizations can be achieved by the application of a brief intense illumination preceding the preset illumination level (supercharging). This light-driven vesicular voltage-control system can be applied to functional assay of Kv (and other) channels, permitting optimization of 'active' protein production for structural studies.

Reported changes in channel densities accompanying thymocyte maturation processes 48-09-04: Kvl.3: The type n K+ channels of native T lymphocytes (encoded by Kvl.3, see Current type, 48-34) show variable densities of expression in parallel with the rate of cell division 73, 74. Immature proliferating thymocytes display ",,200-300 channels/cell, the number decreasing to ",,10-20 channels/cell during differentiation into mature, quiescent T cells 73, 75. Activation of mature T cells results in a 20-fold increase in K+ channel number, the increase being exclusively of type n 76 .

Other phenomena associated with tvariable' Kv channel expression in heterologous cells 48-09-05: Kvl.l/RCKl: Direct correlation between channel density and transcriptional rate of a metallothionein promoter fused to the RCKI K+ channel in stably transfected Sol 8 cells has been reported 77. Mouse L cells stably transfected with Kvl.l eDNA have been estimated to yield 10004000 functional surface channels/cell 78 (see also Protein interactions, 48-31). Note: A number of heterologous expression studies have specifically related variable mRNA expression levels to differences in inactivation kinetics, pharmacological profile and other functions. For example, 'variable' expression levels of a Shaker clone have also been reported to affect membrane potential in oocytes 79 : At high channel densities (Gmax > 10pA/mV) the mean membrane potential was stabilized at

II

l_e_n_t_ry_48

---'_

approximately -60 mV and independent of slow Ie-type' inactivation; at lower channel densities, membrane potential was 'very unstable', with its mean value (and amplitude of fluctuations) being strongly influenced by the process of C-type inactivation (see Inactivation, 48-3 7). For further descriptions of differentiated functional properties with 'variable' expression levels, see Current type, 48-34; Inactivation, 48-37 and Blockers, 48-43. See also the fields Activation and Kinetic model under VLG K minK, 54-33 and 54-38, respectively}.

Cloning resource 48-10-01: The ubiquity of Kvl series channel expression and their possession of intronless t coding regions (all except Kvl.7, see Gene organization, 48-20) has enabled retrieval of cDNA and genomic clones containing full-length ('expressible') sequences from a wide range of DNA libraries, as exemplified in Table 2.

Developmental regulation Multiple mechanisms for generating K+ channel diversity 48-11-01: Voltage-gated K+ channels display heterogeneous expression patterns, many of which appear subject to developmental control (see examples, below). Molecular heterogeneity may arise from differential expression of multiple K+ channel subunit genes (see mRNA distribution, 48-13), alternative splicing (see Gene organization, 48-20) and/or the formation of heteromultimers from different subunits (see Protein interactions, 48-31). Spatial- and temporal-selective induction of Kv channel activities in development (developmental heterogeneity) presumably reflects functional specialization within particular 'developmental compartments' or terminally differentiated cell types (for further background, see Developmentalregulation under TUN [connexinsj, 35-11). Localization (segregation) of different Kv proteins to specific subcellular domains may also significantly affect native channel properties, particularly where they are co-localized with modulatory proteins (for examples, see Subcellular locations, 48-16 and Protein interactions, 48-31).

Developmental heterogeneity of Kv channels in brain - early studies 48-11-02: Kvl.I/Kvl.2/Kvl.3/Kvl.4: mRNA transcript distribution patterns of 11 Kv channel genes (encoding both slow- and fast-inactivating K+ channels from four different gene families) have been systematically examined in a developmental context81 . This analysis confirmed that Kv subunit-specific mRNAs are independently expressed and can be characterized by individual but overlapping expression patterns (see mRNA distribution, 48-13). Evidence for heterogeneous expression and post-natal developmental regulation in the hippocampal formation were obtained for multiple transcripts RCK2, RCK3, RCK4, RCKS, Raw3 and Shall gene products. Transcripts present in the hippocampus throughout post-natal life included those encoding RCKI, Rawl, Raw2 and DRKI 82 .

II

_L..-

e_n_try_4_8_

XShal expression in Rohon-Beard cells, neural crest derivatives and glia 48-11-03: Xenopus-Kvl.l/XShal/Kvl.2/XSha2: In primary sensory neurones of Xenopus laevis, the functional differentiation of delayed rectifier potassium current regulates the waveform of the action potential, while acquisition of A-current (entry 44) also plays a major role in regulating excitability. Within the embryonic nervous system, transcripts encoding XShal and (at a lower level) XSha2 are detectable 83 . The onset of neurogenesis in Xenopus (stage 13) is associated with the induction of Kvl.2 gene expression (isolate XSha2)21. XShal mRNA is expressed in Rohon-Beard cells (primary sensory neurones which exhibit developmentally regulated action potentials - see paragraph 48-11-04) and several structures containing neural crest derivatives including spinal ganglia, the trigeminal ganglion, and branchial arches. Detection of XSha1 in motor nerves and lateral spinal tracts also suggests that both CNS and PNS glia express the mRNA83 . Although many different types of Xenopus spinal neurones exhibit homogeneous development of IKv both in vivo and in culture, transcripts of two genes encoding delayed rectifier current, Kvl.l (this entry) and Kv2.2, are expressed heterogeneously during the period in which the current develops: Kvl.l mRNA is detectable in a maximum of 300/0 of cells while I Kv is immature; Kv2.2 mRNA appears later in approx. 600/0 of mature neurones. For further details of Kv2.2 developmental expression in ventral spinal neurones (as opposed to dorsal spinal neurone distribution of Kvl.l transcripts)84, see Developmental regulation under VLG K Kv2-Shab, 49-11.

In vivo (endogenous) regulatory factors that can 'rescue' perturbed neuronal differentiation 48-11-04: Xenopus-Kvl.l/XSha1: Injection of Shaker-like potassium channel transcripts into two-cell stage Xenopus embryos correlate with larger delayed rectifier current amplitudes and shortened action potential durations 85. XSha1 'overexpression' is associated with reductions in the number of morphologically differentiated Rohon-Beard neurones that appear in cultures prepared from neural plate stage (17.5 h) embryos (see note 1 and paragraph 48-11-03). This suggested that 'premature' modulation of impulses suppresses normal developmental cues in this model in vitro system (ref. 85 ). In comparison, a later study86 showed that morphological differentiation of Rohon-Beard neurones in situ (in vivo) was only 'slightly affected' by overexpression of K+ channels (i.e. endogenous developmental factors (regulatory processes or interactions) appeared to compensate for the effect of channel overexpression). Notably, when cultures are prepared from older neural tube embryos (22-24 h), more neurones containing 'excess' K+ channel RNA differentiate morphologically in vitro. Recovery of differentiation capacity is possible if a minimum of 5 h of further development in vivo is allowed (under conditions in which (i) rapid elevations of [Ca2 +h are permitted and (ii) half of the nervous system has normal levels of potassium channel RNA (see note 1; for general significance of [Ca 2+h transients in development, see this field under ILG Ca Ca RyRCaf, 17-11 and under ILG Ca InsP3, 19-11). Notes: 1. Hypothetically, increased K+ current could oppose long-duration Ca2+ influx and nuclear Ca2+ elevations (see NUC [nuclear, native], entry 38) associated with differentiation

II

l_e_n_t_ry_4_8

_

Table 2. Source tissues for DNA library constructions (From 48-10-01) Kvl.l

MBKl: Mouse brain cDNA library. MKl: Mouse genomic DNA library. RBKl: Rat hippocampal cDNA library. RBKI cDNA and native channels with its properties have been systematically characterized in C6 astrocytoma cells and native astrocytes 80 • RCKl: Rat cerebral cortex cDNA library. RKl: Rat aorta cDNA library.

Kvl.2

RCKS: Rat brain cortex cDNA library. NGKl: Mouse neuroblastoma x rat glioma hybrid cell cDNA library. MK2: Mouse genomic DNA library. RK2: Rat cardiac cDNA library. BK2: Rat brain cDNA libraries. CMSKl: Canine colonic circular smooth muscle cDNA.

Kvl.3

hPCN3: Isolated from human insulinoma cDNA library and from genomic DNA library sources. MK3: Mouse genomic DNA library. RGKS: Isolated from a genomic DNA library. RCK3: Brain, cortex cDNA library. Isolate KV3: Rat brain cDNA library. HLK3: Human brain and T lymphocyte cDNA libraries. Isolate rabK(vl.3): Rabbit kidney medullary cell line GRB-PAPI (see Phenotypic expression, 48-14).

Kvl.4

HKl: Human left ventricular cDNA library. hPCN2: Human foetal skeletal muscle cDNA library. RCK4: Brain, cortex cDNA library. RHKl/RK3: Rat cardiac cDNA library.

Kvl.5

HK2: Human left ventricular cDNA library. hPCNl: Insulinoma cDNA library and genomic library. Isolate KVl: Rat brain cDNA library. RK4: Size-fractionated cardiac cDNA library. RMK2: Rat skeletal muscle cDNA library. BAKS: cDNA library from bovine adrenal medulla, and, subsequently, a bovine genomic library. Rabbit Kvl.S: Portal vein vascular smooth muscle cDNA.

Kvl.6

HBK2: Human foetal brain cortex cDNA library. Isolate KV2: Rat brain cDNA library.

phenotypes such as neurite outgrowth, neurone-myocyte contacts and acquisition of GABA-like immunoreactivity (see ELG Cl GABAA, entry 10). Significantly, the in vitro study described above 85 concluded that larger potassium currents were not compensated by changes in inward currents. 2. In these experiments RNA 'overexpression' is limited to half of the embryo in order to provide an internal control; when K+ channel RNA is overexpressed throughout the embryo, few neurones are observed to differentiate morphologically in vitro (even if cultures are prepared from older neural tube embryos).

Modified endogenous Xenopus Kv channel behaviour by exogenous Kv channel expression 48-11-05: rKvl.l: Rat Kvl.l channel currents have been analyzed in blastomeres during a 12h period prior to stage 15 (early-mid neurula) and in

II

_ 1......----

entry 48 _

differentiating muscle cells (post-stage 15) following injection of rKvl.l cRNA into fertilized eggs87. In rKvl.l-injected embryos, a high fraction of blastomeres express a delayed rectifier-type current distinguishable from the endogenous muscle delayed potassium current (IK, Ix) by its different voltage dependence. Notably, native IK, Ix appears much earlier in development than in control embryos; this and other pharmacological data suggests an interaction/induced modification between Kvl.l and endogenous channel subunits in the course of development87.

Functional expression of Kvl.3 in myotomal muscle of developing Xenopus embryos 48-11-06: Kvl.3: Injection of cRNA encoding Kvl.3 into fertilized eggs of Xenopus can give rise to functional Kvl.3 channels in dissociated, cultured myotomal muscle cells (i.e. from embryos of stage 20-22 about 40 h post-fertilization at 19°C, cultured in vitro for 2 days)88. Kvl.3 currents were distinguished from endogenous delayed rectifiers (sustained and transient) and an inward rectifier K+ current) by its inactivation properties and its high sensitivity to the non-selective blocker charybdotoxin88.

Kv mRNA repression associated with seizures

48-11-07: Kvl.2/Kv4.2: Seizure t activity induced by the convulsant drug pentylenetetrazole (Metrazole) has been reported to be followed by a reduction of Kvl.2 mRNA (~90min to 6h pj.) and Kv4.2 mRNA (3-6h pj.) in the dentate granule cell layer of the hippocampus, while Kvl.l mRNA levels remained unchanged58 . mRNA signals return to normal (control levels) within 12h. Notably, administration of the anticonvulsant diazepam prior to pentylenetetrazole protects animals from seizure and blocks the suppression of Kvl.2 and Kv4.2 mRNAs, suggesting the mRNA repression is due to induced neuronal activity evoked by seizure (rather than an unrelated side-effect of the drug).

Mechanism of the tantiproliferative' effect of K+ channel blockers in Tcells 48-11-08: Kvl.3: Charybdotoxin (see Blockers under ILG K Ca, 27-43) depolarizes human peripheral T lymphocytes and renders them insensitive to activation by mitogens t. The scorpion venom peptides noxiustoxin (NxTx) and margatoxin (MgTx) block only the voltage-activated channels, whereas charybdotoxin blocks three types of calcium-activated potassium channel and Kv channels in these cells89 . All three toxins induce 'equivalent' depolarization and block T cell activation in human T cells. On the basis of these and other results 89, it has been concluded that membrane potential of resting T cells is set by voltage-activated channels and that blockade of these channels is sufficient to produce depolarization of unstimulated human T cells, thereby preventing mitogenic activation (but see below). Notably, the potent Kvl.3 blocker CP-339,818 (which competitively inhibits [l 25 I]-charybdotoxin from binding to the external vestibule of Kvl.3) suppresses T cell activation in vitr0 90 (see Blockers, 48-43). Supplementary notes: Depolarization attenuates the increase in cytosolic Ca2+ that normally occurs during receptor-ligand coupling91 . Following

II

l_e_n_t_ry_48

--...._

elevation of [Ca2 +h, KCa channels may also contribute to the Tcell membrane potential, as supported by the ability of charybdotoxin to block the hyperpolarization that follows stimulation by mitogens or Ca2 + ionophores 92- 98 (for further details see ILG Ca InsPJ, entry 19 and ILG K Ca, entry 27). More recent work has established an important role for Kvl.3 in maintaining secretion of the lymphokine IL-2 in activated T cells: In this case, following T cell receptor ligation t , Kv1.3 opening prevents the influx of calcium into the (non-excitable t ) cell from 'collapsing' its own electrical gradient (thereby sustaining the IL-2 secretory response). These features have motivated development of blockers for Kvl.3 as potent immunosuppressors by their effects on IL-2 secretion (see Blockers, 48-43 and the 1996 review in ref. 99).

Kvl.l and Kvl.3 in a lymphoid/thymocyte development model 48-11-09: Kvl.l/Kvl.3: Murine foetal and adult CD4-CD8- (immature) thymocytesf display currents with the distinctive rroperties of Kvl.l and Kvl.3, and their mRNA can be detected by RT-PCR in these cells10o. When applied to a murine foetal thymic organ culture system, non-selective peptide blockers of Kvl.l (dendrotoxin, DTx - see Blockers, 48-43) and Kvl.3 (charybdotoxin, CTx - ibid.) decrease thymocyte yields in organ culture without affecting thymocyte viability: In comparison to untreated thymic lobes, DTxtreated thymi contain 56 ± 8% thymocytes (n == 8) while CTx-treated thymi contain 74 ± 4 % thymocytes (n == 7). DTx and CTx also alter the developmental progression of thymocytes in foetal organ culture, consistent with functions critical to thymocyte pre-clonal expansion and/or maturation1OO.

'Parallel developmental onset' of Kv channel expression and voltage-gated Na+ channels 48-11-10: Kvl.3/Kvl.4: There is an 'early onset' of Kvl.3 (RCK3) and Kvl.4 (RCK4) mRNA expression comparable to the temporal mRNA expression pattern of rat sodium channels II and ill in the CNS 101 ,102. Kvl.4 (isolate RCK4)-specific probes detect an mRNA species of "-I4300nt in heart which increases during development 101 .

'Monotonic' development of Kv protein expression patterns in developing hippocampus 48-11-11: Kvl.4/Kvl.5: The 'spatiotemporal' expression patterns of the Kvl.4, Kvl.5, Kv2.1, Kv2.2 and Kv4.2 polypeptides in rat hippocampal neurones developing in situ have been the subject of a comparative study l03. The development of protein expression patterns in situ (see Protein distribution, 48-15) has been described as Imonotonic', i.e. while the 'temporal' and 'spatial' development varies among Kv channels, each subtype initially appears in its adult pattern (taken to suggest that the mechanisms underlying spatial patterning operate through development). Immunoblotting techniques have confirmed the differential temporal expression of K+ channels in the developing hippocampus, and demonstrate developmentally regulated changes in the 'microheterogeneity' of some Kv subtypes 103 (see Subcellular locations, 48-16).

II

_1...-

e_n_try_4_8_

Hormonal regulation of pituitary excitability via K+channel gene expression 48-11-12: Kv1.5: Glucocorticoidt agonists such as the steroid dexamethasonet

(DEX) can specifically increase Kvl.S mRNA in normal and clonal (GH3) rat pituitary cells104,,105. Dexamethasone application can rapidly induce Kvl.S gene transcription, without affecting Kvl.S mRNA turnovert (tl/2 == O.Sh). These effects are correlated with 3-fold increased expression of a 76 kDa Kvl.S protein (see Protein molecular weight (purified), 48-22) within l2h without altering its half-life (tl/2 == 4h)105. Kvl.S protein induction is also associated with an increase in a non-inactivating component of the voltagegated K+ current. In parallel experiments, dexamethasone did not affect Kvl.4 transcript or protein expression105. DEX treatment has also been demonstrated to increase total cellular and GH3 surface Kvl.5 protein and, in consequence, to modify cell surface homomeric and Kvl.4/Kvl.S heteromeric subunit composition106 . The actions of DEX are markedly tissue-specific: For example, in adrenalectomized rats, DEX rapidly induces Kvl.S mRNA in cardiac ventricle, skeletal muscle and pituitary, but does not affect Kvl.S mRNA levels in hypothalamus or lung107. Methodological note: Dexamethasone induction of the slowly inactivating current in the clonal rat pituitary cell line GH4Cl can be blocked by antisense t phosphorothioate t deoxyoligonucleotides to the Kvl.SmRNA sequence 108. By contrast, antisense deoxyoligonucleotides against Kvl.4 mRNA specifically decrease the expression of the dexamethasone-insensitive rapidly inactivating current in these cells 108.

Induction of Kv channel mRNAs by tstress responses' 48-11-13: Mimicking the 'hormonal stress response' using adrenocorti-

cotrophic hormone has no apparent effect on several Kv mRNA levels58 but the cold stress response of rats has been shown to significantly increase cardiac Kvl.S mRNA expression107. Comparative note: Independent studies on mKvl.l (described in Table 16 under Protein phosphorylation, 48-32) have suggested that decreased basal activities of protein kinase A can 'upregulate' mKvl.l channel expression by changing steady-state levels of RNA and by other post-transcriptional mechanisms 109 (ibid.).

TRH enhances excitability of GH3 cells by inhibition of K+ channel gene expression 48-11-14: Kvl.S: Neuropeptide regulation of K+ channel gene expression is

capable of inducing 'long-term' changes in neuronal action potential activity and synaptic transmission. For example, thyrotrophin-releasing hormone (TRH) downregulates Kvl.S (and Kv2.l) K+ channel mRNAs in G03 pituitary cells by decreasing rates of transcription110. These changes can be correlated with significantly decreased immunoreactivities and K+ current expression within 12 h of TRH application.

Developmental regulation of rKvl.5 by cAMP and K+ -induced depolarization 48-11-15: rKvl.S/rKvl.l: A sequence conforming to a cAMP response elementt (CRE) has been identified in the S'-non-coding region of the rKvl.S

II

1L....-__ _ _ry_48

en t

_

gene at position +636 relative to a transcriptional start site111 (see also Gene organization, 48-20). In primary cardiac cells, cAMP induces a 6-fold increase in the steady-state levels of Kv1.S transcript. In GH3 pitutary cells, by contrast, cAMP induces a S-6-fold decrease in steady state levels of Kvl.S transcript by reduction in Kv1.S gene transcription rate. Progressive deletion of S' non-coding sequences (coupled to a reporter gene assay) have shown that (i) the consensus CRE can confer cAMP inducibility to Kvl.S and (ii) this element is capable of binding CRE-binding protein (CREB) and CRE modulator protein (CREM) in gel shift assays. KCI-induced depolarization has also been shown to increase steady-state levels of Kvl.S transcript in primary atrial cells and decrease it in GH3 clonal pituitary cells111,,112; the KCI treatment has no effect on Kvl.4 or Kv2.l transcription and NaCI cannot suppress Kvl.S transcription l12 .

A K+ channel gene silencer with a dinucleotide repetitive element 48-11-16: Kvl.S: A gene silencert element (alternatively named Kvl.S repressor element or KRE) has been identified by deletion analysis of the Kvl.S gene promoter195. In cell lines that do not express Kvl.S, cis constructs of KRE selectively decreases expression of Kvl.S and reporter genes. KRE contains a dinucleotide repetitive element (polyGT 19 (GAh(CAhs(GAh6h deletion of the repetitive element from reporter constructs abolishes an in vitro selfassociation property and the silencing activity. KRE appears to form a stable nucleoproteint complex with nuclear extracts from GH3 cells (which cannot be detected with CHO and and COS-7 cell extracts)195.

Quantitation of Kv current, mRNA and protein during post-natal heart development 48-11-17: Comparative developmental analysis of the native rat cardiac ventricular currents Ito (entry 44) and I K (entry 45) versus multiple Kva protein/mRNA expression levels (Kvl.2, Kvl.4, Kvl.S, Kv2.l and Kv4.2) has been performed l13 . In this study, mean Ito densities increased 4-fold between birth and P30, whereas IK densities varied only slightly. No variation in either the time- nor the voltage-dependent properties of Ito or IK could be detected over this period. The same study showed ventricular Kvl.4 mRNA levels were 'high' at birth, increased between PO and PIa, and subsequently decreased to 'very low' levels in adult rat hearts (decreases in message were accompanied by a marked reduction in Kvl.4 protein, see note). Notably, mRNA levels of all other Kv channels studied (Kvl.2, Kvl.S, Kv2.l and Kv4.2) increase (3- to S-fold) between birth and adult. Paradoxically however, Kvl.2 and Kv4.2 protein levels increase between PS and adult, whereas Kv1.S protein remained constant while Kv2.l decreased. Note: On the basis of this and earlier data, these authors suggested that Kv1.4 does not contribute to the formation of functional K+ channels in adult rat ventricular myocytes l13,,114. In earlier studies a 'marked increase' in the amount of mRNA encoding rKvl.S (clones KVl/RMK2) in cardiac tissue has been noted during development and a similar but less-pronounced increase of both Kvl.S and Kvl.6 (isolate KV2) transcript in brain32,,44 (see also Channel density, 48-09).

II

_L..-.-.

e_n_try_4_8__

Directed transcription of the native hKvl.5 gene within single chromosomal fragments 48-11-18: hKvl.5: A single excised human genomic DNA fragment encompassing the hKvl.5 gene promoter region contiguous to its coding sequence has been used to heterologously express Kvl.5 protein in a mammalian cell line lacking voltage-gated K+ channels l15 . In this study, the hKvl.5 gene fragment was (i) linked downstream of a human ,a-globin locus control region t (LCRt or 'gene-activation element') and (ii) stably incorporated into the host erythroleukaemia cell genomic DNA by recombination following transfection of the construct. hKvl.5 gene locus activation and hKvl.5 current can be induced following treatment with various polar/ apolart chemical inducers of differentiation (for details, see Developmental regulation and Voltage sensitivity under ILG K Ca, 27-11 and 27-42 respectively). Expression levels are characteristically independent of chromosomal integration position and are proportional to locus control region (LCR)-gene construct copy number, a parameter specific for each clonal cell line. These results demonstrate the general utility of LCRdirected gene-activation methods for ion channel gene or cDNA expression l15 .

Moderate induction of Kv gene transcription induced by chronic morphine administration 48-11-19: rKvl.5/rKvl.6: Prolonged opiate administration leads to the development of tolerancet and dependencet, while acute opiate administration differentially affects voltage-dependent K+ currents in vivo. While opiate activation of K+ channels is well established (see INR K G/ACh [native], entry 31), opioid-induced inhibi.tion of K+ conductances has also been described. In separate studies l16, 'significant increases' in abundance for mRNAs encoding Kvl.5 (2.1 ±O.15-fold by ISHt) and Kvl.6 (2.3 ±O.5fold) are induced in the spinal cord of rats following chronic morphine administration (compared with controls). Moreover, Kvl.5 protein level was also increased by 1.9-fold in the spinal cord of morphine-treated rats. These findings have been interpretedl16 as 'cornpensatory' for persistent opioidinduced inhibition of K+ channel activity in motor neurones, perhaps contributing to tolerance/dependence conditions.

Isolation probe tCore region' probes maximize retrieval of related sequences 48-12-01: Generally, vertebrate K+ channel isoforms have been isolated using single or mixed probest homologous to Drosophila Shaker locus or from existing gene family members. Homologous isoformst in different vertebrate species are generally highly conserved at the amino acid level, and have been isolated by low-stringency cross-hybridization. Because of the greater degree of conservation for sequences associated with 'core regions', low-stringencyt hybridization screens with these probes were better able to isolate related genes. Note: 'Core' regions e.ncompass the sequences encoding the transmembrane domains, ionic selectivity and voltage-sensitivity determinants (for further details see Domain functions, 48-29 and the [PDTM], Fig. 6).

II

entry 48 L......---

_ _

Probes for selective reporting or retrieval of individual family members 48-12-02: Channel 'isoform-specific' probes (either nucleic acid or antibody) are generally derived from C-terminal, N-terminal or untranslated sequences which are divergent amongst the gene family members. Between distantly related species (e.g. vertebrate and fly) K+ channel amino acid sequences can retain strong homology, although randomization of nucleotides at degenerate t positions and different codon usage t underlies divergence at the nucleotide level. Panels of PCR t primer pairs designed to selectively amplify known isoform sequences from genomic or cDNA sources are in common use (e.g. primer sets for Kvl.l, Kvl.2 and Kv4.2 subtype sequences are described in58 and for MK3 in33 ).

mRNA distribution Differential expression patterns of the Kv gene subfamily mRNAs 48-13-01: Many cell types express voltage-gated K+ channel subpopulations that are characterized by different functional and pharmacological properties l17,118. Generally, Kv subunit genes show distinct but overlapping expression patterns within brain and other tissues 58,101. However, the expression pattern of some Kv subunit mRNAs in native tissues appear ubiquitous. Partially overlapping expression patterns in defined regions (e.g. ref. 81, see Developmental regulation, 48-11) indirectly suggest the formation of different heteromultimeric K+ channel complexes; strictly, however mRNA distributions indicate sites of biosynthesis, and do not necessarily indicate the ultimate locus of protein expression (see Protein distribution, 48-15). Furthermore, the roles of different K+ channel proteins could be markedly influenced by their localization to specific subcellular domains (see Subcellular locations, 48-16). Table 3 lists some qualitative (verbal) summaries of Kv subunit mRNA expression patterns, as originally reported. Although of limited value, they can at least identify discrepancies in gross distribution patterns and can serve as a foundation for systematic comparative studies using genotype/phenotype database approaches (see footnotes to Table 3 and Resource H - Index of cell types).

Information derived from mRNA 'expression'surveys covering specified cell types (example) 48-13-02: Since molecular probes 'covering' complete (known) gene families or subfamilies are now available, some studies have attempted characterization of mRNA abundance or distribution within single cell or tissue types. For example, the quantitative assessment of eighteen different voltage-activated potassium channel genes in rat sympathetic ganglia has been described l19 . Comparative studies like this can reveal the range of subtype expression. Thus, sympathetic ganglia were shown to express eleven a subunit genes and two {3 subunit genes l19 . Relative levels of mRNA expression can also be compared between different functional 'lineages' of cells; for example, between the superior cervical ganglion (SCG) and two pre-vertebral sympathetic ganglia, the coeliac ganglion (CG) and the superior mesenteric ganglion (SMG). In these cases, only four mRNA subtypes were shown to be differentially expressed:

II

_""---

e_n_try_4_8------1

Table 3. Reported tissue distributions of Kvl subfamily mRNAs (From 48-13-01) Kv

Systematic/clone names, original descriptions, notes and references

Kvl.l

hKvl.2: RT-PCR: Human airway smooth muscle120 (see Phenotypic expression, 48-14). rKvl.I-RCKI: Northerns: Localized distribution suggests the channel mediates functions that are concentrated in distinct subregions. Detected in cortex and cerebellum but not in heart, kidney, spleen, lung, testis, liver and muscle10. RCKI mRNA is of generally low abundance (compared to actin mRNA probes which show rvxSOO the signal intensity of RCKI probes). Predominantly expressed in the adult nervous system (rvneurones and/or glial cells) at approx. equal mRNA abundance to RCKS; high mRNA levels found in sciatic and other peripheral nerves 101 . Maximal levels in caudal brain regions, relatively low levels in rostral regions and in retina: Cerebellum + + + +; cerebral cortex + + +; corpus striatum +; hippocampus ++; inferior colliculus + + + +; medulla-pons + + + + +; midbrain (-colliculi) -t- + + +; olfactory bulb +; spinal cord + + + +; superior cOlliculus + + + + + 101. Comparative note: RCKI closely resembles the temporal and regional expression pattern of rat sodium channel I in the CNS 101,102. mRNA not detected/low abundance in adrenal and submandibular glands, kidney, cardiac, skeletal and smooth muscle. ISH: High mRNA levels detected in granule cells of the dentate gyrus, the pyramidal cells of the Ammon horn (CA3 >. CAl), and in the cerebellum. Low expression levels were found in basal ganglia (caudate putamen, globus pallidus and ventral pallidum)121. rKvl.l-RKI: Northerns: Brain » cardiac atrium == skeletal muscle; no signal in liver and cardiac ventricle 17. rKvl.I-BKI: ISH: Expressed at relatively high levels in the hippocampus, thalamus, cerebral cortex and cerebellum. Not expressed in the medial habenula; hybridization relatively strong in the superficial and deep layers (cf. Kvl.2); in hippocampal subregions, the relative distribution of channel mRNA is CA3 ~ dentate gyrus> CAl. Concentrated expression in the cell bodies of the granule cells of the dentate gyrus and the pyramidal cells of CA3 and CAl - therefore expressed in neurones and possibly in CNS glial cells. Expressed in many hilar cells at much higher levels than in the neighbouring dentate granule or CA3 pyramidal cells. Expressed in Purkinje cells and granule cells but not detected in cells within the molecular layer (cf. Kvl.2 and Kv4.2). Seizure t activity does not affect Kvl.l mRNA levels (see Developmental regulation, 48-11) 58. mKvl.I-MBKI: Northerns: Brain +. + + +; heart +; skeletal muscle +. mKvl.l/mKvl.2/mKvl.3: RT-PCR: mRNA from enzymatically dissociated, isolated mouse rod bipolar cells show that these neurones co-express Kvl.l, Kvl.2 and Kvl.3 122 (see Subcellular locations, 48-16).

---J_

1L-_e_n_t_ry_4_8 Table 3. Continued Kv


Kvl.l

xKvl.l/XShal: RNAaseP: XShal mRNA is expressed in Rohon-Beard cells (primary sensory neurones which exhibit developmentally regulated action potentials) and several structures containing neural crest derivatives83 (see Developmental regulation, 48-11).

Kvl.2

hKvl.2: RT-PCR: Human airway smooth muscle120 (see Phenotypic expression, 48-14). rKvl.2/BK2: ISH: Generally no localized distribution observed, but resembles cell density patterns determined by Nissl stainingt; suggests the channel mediates functions that are generally distributed. 1s,s8 rKvl.2/RAK: Northerns: Northern blots show RAK mRNA in adult rat atrium (strong) and in rat ventricle (very weak)24 (see Phenotypic expression, 48-14). rKv1.2/RCK5: Northerns: Cerebellum + + + +; cerebral cortex + + +; corpus striatum +; hippocampus + +; inferior colliculus + + + +; medulla-pons + + + + +; midbrain (-colliculi) +; olfactory bulb ++; spinal cord + + + + +; superior colliculus ++101. rKv1.2/RCK5/RK2: Northerns: Relative mRNA abundance brain> cardiac atrium> aorta == ventricle. No signal in skeletal muscle 17. Expression restricted the adult nervous system (rvneurones and/or glial cells) at approx equal mRNA abundance to RCKl; high mRNA levels found in sciatic and other peripheral nerves 101 . Different probes show feint RCK2 signals in embryonal tissue and neonatal heart but not adult heart RNA118. rKvl.2/Kv4.2: ISH: Expressed in the medial habenula; hybridization relatively diffuse in the superficial and deep layers (cf. Kv1.1); in hippocampal subregions, the relative distribution of channel mRNA is CA3 > dentate gyrus ~ CAl. High dentate gyrus granule cell and hilar cell expression - as for Kvl.l. Most abundant in Purkinje cells and in a subset of neurones in the molecular layer, compared to signals in the neighbouring granule cell layer (cf. Kvl.l and Kv4.2). Seizure activity can reduce Kvl.2 and Kv4.2 mRNA levels s8 (see also Developmental regulation, 48-11). rKvl.2/Kvl.4/rKvl.5: Using a quantitative RNAase protectiont assay comparing the abundance of fifteen different potassium channel mRNAs in rat cardiac atrial and ventricular muscle123, only Kvl.2, Kvl.4 and Kvl.5 (plus Kv2.1 and Kv4.2) were judged to be expressed at I significant levels' (see also mRNA distribution under VLC K Kv4-Shal, 51-13). Canine Kv1.2/CSMK1: Northerns: CSMK1 is expressed in a wide variety of gastrointestinal smooth muscles27. Portal vein, renal artery, and uterus do not express CSMK1 mRNA, suggesting that (among smooth muscles) expression of this K+ channel may be restricted to gastrointestinal smooth muscles27 (see also Phenotypic expression, 48-14).

II

_'--

e_n_try_4_8__

Table 3. Continued Kv


Kvl.2

xKvl.2/XSha2: RNAaseP: Expressed in the nervous system but not detectable in skeletal muscle21 .

Kvl.3

rKvl.3/Isolate KV3/RGK5: RT-PCR: RNA transcripts detected in brain, lung and spleen; also cloned from rat peripheralleukocytes32 . rKvl.3/RCK3: Northerns: Cerebellum (-); cerebral cortex + +; corpus striatum + +; hippocampus + +; inferior colliculus + + + + +; medulla-pons + + + -t-; midbrain (-colliculi) + + +; olfactory bulb + + + +; spinal cord + + +; superior colliculus

+ + +101.

mKvl.3/MK3/HLK3: RT-PCR: MK3 mRNA is expressed in T lymphocytes. Encodes a channel with biophysical and pharmacological properties indistiguishable from those of voltage-gated type n K+ channels in T lymphocytes31 (see Current type, 48-34). Kvl.4

hKvl.4/HKl: Northerns: Atrium + + +; ventricle + + +. rKvl.4/RCK4: Northerns: Cerebellum (-); cerebral cortex + + +; corpus striatum + + + + +; hippocampus + + + +; inferior collicul\, us + + ++; medulla-pons +; midbrain (-colliculi) + + +; olfactory bulb + + + + +; spinal cord (-); superior colliculus + + + + +101 (see also Kv3.4, this table, under VLC K Kv3-Shaw, comparing distributions of RCK4 to Raw3, encoding another A-type K+channel. rKvl.4/RK3: Northerns: Relative mRNA abundance: Brain~cardiac atrium> aorta == skeletal muscle:> cardiac ventricle. 17 Transcripts also detected in skeletal muscle, tongue, stomach, small intestine, uterus and liver. rKvl.4/RGHK9: ISH: mRNA detectable in anterior pituitary; particularly abundant in the hippocampus 124.

Kvl.S

hKvl.5/HK2: Northerns: Atrium + + + + +; ventricle +. hKvl.5/hPCNl: Northerns: Brain (-); kidney (-); liver (-); lung (-); pancreas ++; skeletal muscle (-); ventricle +. RT-PCR: hPCNl is present in normal human pancreatic islets but is 'not detectable' in skin fibroblasts or HepG2 cells. RT-PCR: Human airway smooth muscle 120 (see Phenotypic expression, 48-14). rKv1.5/Isolate KV1: Northerns: Relatively widespread distribution: found in brain, heart, kidney, lung and skeletal muscle32 (see also Developmental regulation, 48-11). Comparative note: Isolate KVl mRNA is expressed at high levels in the GH3 pituitary cell line; although Kvl behaves as a non-inactivating delayed rectifier in oocytes, the voltage-dependent K+ current in GH.3 cells is a relatively rapidly inactivating current but has similar pharmacology to KVl current in oocytes). Rat spinal cord motor neurones are 'highly emiched' in the Kvl.5 and Kvl.6 nlRNAs (see l16 described under Developmental regulation, 48-11). Kvl.5 is also widely expressed in glial cells of brain and spinal cord125 (see Protein distribution, 48-15).

entry48

I-

_

- - - - - -



Kvl.5

rKvl.5/RK4: Northerns: Ubiquitous in excitable tissue, being present at comparable levels in atrium, ventricle, aorta, brain and skeletal muscle. Relative mRNA abundance: Cardiac atrium == cardiac ventricle == aorta == skeletal muscle> brain. No signal in liver. RK4 is the most abundant cardiac channel mRNA of the RKI to RKS group17. rKvl.5/RMK2: RNAaseP: The 3' non-coding regions of the brain, cardiac and skeletal muscle RMK2 transcripts are identical44 . rKvl.5/Kvl.5: RT-PCR: RNA transcripts for Kvl.l, Kvl.2, and Kvl.5 have been detected in adult rat sciatic nerve preparations (see also Kv1.1/Kv1.5 under Protein distribution, 48-15).

Kvl.6

hKvl.6/HBK2: Northerns: Not found in tissues other than brain. rKvl.6/RCK2: Northerns: Detected mainly in midbrain areas and brainstem. ISH: 'Homogeneous' expression levels in most brain regions; slightly elevated message levels in the piriform cortex, the olfactory tubercule, and the dorsal endopiriform nucleus. Low expression in the hippocampus, the central medial thalamic nucleus, the zona incerta, the medial amygdaloid nuclei, and the lateral amygdaloid area. Generally low expression in the cerebellum but 'significant' in the Purkinje celllayerl18 . Rat spinal cord motor neurones (see Kv1.5, above)l16. rKvl.6/Isolate KV2: Northerns: Transcripts detected in brain only32.

Notes: 1. 'Verbal' descriptions of expression patterns are included where they have been specifically reported, but these are a poor substitute for the original in situ hybridization data (ISH:). 2. Relative expression levels (mostly according to schemata in original references) have been indicated by a variable number of '+' signs within a group of tissues (e.g. + + +), where the comparison applies within that group only. 3. Mapping of expression patterns is a complex task and has to take many variables into account, such as in situ localization, developmental regulation, subunit stoichiometry, and factors regulating 'overlapping' or coexpression. For further details of annotated digitized image libraries which permit direct comparison of mRNA/protein expression patterns in relation to biological specialization of tissue regions (e.g. in brain) see Resource TSearch Criteria and Genotype/phenotype models under Resource H - Cell types. 4. To help consolidate multiple in situ hybridization data for Kv subunit expression patterns into a single reference source for availability on the WWW, please forward citation details as described under Feedback, 48-57. 5. Other designations used above include Northems: as determined by Northern blot r analysis (low sensitivity); RNAaseP: as determined by RNAase protectiont analysis (intermediate sensitivity); RT-PCR: as determined by reverse transcription-PCR amplificationt analysis (high sensitivity).

II

_'--

e_n_try_4_8_

Kv1.2 and Kv1.4 (this entry), Kv2.2 (entry 49) and Kv{j1 (entry 47); transcripts from all four genes were more abundant in the pre-vertebral ganglia. By comparison with native channel currents (see entries with '[native)' suffix in their sortcode) this type of study can 'propose or reject' candidate subunits underlying native currents. For example, on this basis, it was concluded that (i) members of the Kv4 family are likely to underlie low-threshold A-current in sympathetic neurones (as detailed in entry 51) while (ii) cDNAs encoding two currents prominent in native sympathetic neurones (M-current and D 2 -current, see entry 53) were 'yet to be identified,119.

Phenotypic expression Point mutations in the hKvl.l gene associated with one form of familial episodic ataxia

48-14-01: hKv1.1: One form of familial episodic ataxia t (EA) is characterized by brief episodes of ataxia t with myokymia t (continuous movement or 'rippling' of muscles) evident between attacks consistent with reduced capacity for repolarization in affected nerve cells (for disease phenotype description in relation to developmental onset, see AEMK under OMIM 160120). Using a group of Genethont markers t from a region of human chromosome 12p carrying a paralogous t cluster of K+ channel genes (see Chromosomal location, 48-18) Litt et. al. 126,127 initially described linkage t in four AEMK kindreds t. Chemical cleavage mismatchr mutational analysis of genomic PCR products specifying the coding region of KCNAlled to the description of several additional, different, missense point mutations t in hKv1.1 in these families12S-130. Direct sequencing of products showed that each of the four families had a different missenser mutation, each predicted to affect a highly conserved residue in the hKv1.1 protein (Va1174Phe in the S1transmembrane domain, Arg239Ser in S2, Phe249De in the S2/S3 intracellular loop, and Val408A1a at the border of S6 and the Cterminal domain). All these mutations were present in the heterozygous t state, suggesting that EA/myokymia can result from autosomal dominantr mutations in the KCNA1 gene on chromosome 12p13 (see Chromosomal location, 48-18). Mutant expression studies in Xenopus oocytes have demonstrated that two of the EA subunits form homomeric channels with altered gating properties. For example, Kv1.1 V408A channels (mutated in the region encoding the C-terminal) have voltage dependence similar to wild-type Kv1.1 channels, but have faster kinetics and increased C-type inactivation l29 (see Inactivation, 48-37). Alternatively, the voltage dependence of Kv1.1 F184C channels (mutated in the region encoding TM1) is shifted 20 mV positive. Four other EA subunits studied did not produce functional homomeric channels but reduced the potassium current when co-assembled with wild-type subunits l29.

Roles of Kv channels from subcellular heteromultimeric co-assemblies 48-14-02: Kv1.2; Kv1.4: Subcellular distributions of Kv1.2-containing K+

channels indicate they play diverse functional roles in several neuronal compartments, regulating pre-synaptic Q! post-synaptic membrane excitability, depending on the neuronal cell type (see Subcellular locations, 4816). Kv1.4 protein is targeted to axons and possibly terminals, suggesting a

II

l_e_n_t_ry_48

_

pre-synaptic role in synaptic transmission64 (for some exceptions, e.g. olfactory bulb mitral cell dendrites, see ref. 131 ). In contrast to Kvl.4, a second Q subunit forming A-type channel protein (Kv4.2, see VLC K Kv-Shal, entry 51) is concentrated in post-synaptic dendrites and somata, implying distinct roles for these channel types in vivo (see also Subcellular locations, 48-16). Evidence supporting heteromeric co-assembly of rKvl.4 and rKvl.2 in rat brain has been obtained (see Protein distribution, 48-15 and Protein interactions, 48-31). The Kv 1.4/Kv 1.2 heteromultimer localized in axons and nerve terminals combines features of both 'parent' subunits, resulting in an A-type K+ channeI132 ). At presynaptic loci, Kvl.4/Kvl.2 heteromultimers are likely to form native A-type K+ channels regulating neurotransmitter release. For evidence suggesting fast-inactivating channel components (such as Kv1.4) can mediate 'longlasting' after-hyperpolarizations (tending to reduce action potential frequency) during recovery from inactivation, see Table 19 under Inactivation, 48-37. For general phenotypic roles of A-type channels in native neurones, see VLC K A-T [native], entry 44.

Proposed role of Kvl.2 in smooth muscle excitation-contraction coupling 48-14-03: Canine Kvl.2/CMSK1: Kvl.2 is expressed in canine gastrointestinal muscles; its potential role in regulation of electrical slow wave activity has been described27 (see also mRNA distribution, 48-13). In Xenopus oocytes CSMKI cRNA induces the expression of low-conductance K+ channels displaying a linear current-voltage relation in inside-out patches with a slope conductance of 14pS. The channels were blocked in a concentrationdependent manner by 4-aminopyridine (IC so 74 JlM)27.

Kvl.3 in regulatory volume decrease (RVD) in response to hypotonic shock

48-14-04: mKvl.3: Transient transfection t of a Kvl.3 expression construct into cells of the mouse cytotoxic T lymphocyte line CTLL-2 (unable to volume regulate and 'devoid' of voltage-dependent K+ channels) reconstitutes their ability to volume regulate 133 . This property appears to 'critically depend' on (i) volume-induced membrane potential changes and (ii) the Kv subunit isoform used: When transfected with Kv3.1 transient expression constructs, CTLL-2 cells do not show RVD. According to a model for regulatory volume decrease (RVD, see Fig. 2 and ref. 133) the ability of a Kv channel to confer the capacity for RVD may be partially explicable in terms of different voltage dependence of activation for Kv1.3 and Kv3.1 channels (for background to Kv1.3 and Kv3.1 in native T lymphocytes, see Developmental regulation under VLC K DR [native], 45-11; for other properties, see VLC K Kv3-Shaw, entry 50).

No apparent role for Kvl.3 in CSF-l-stimulated proliferation of microglia 48-14-05: rKvl.3: In cultured microglia established from neopallia of newborn rats an anion current and a K+ current (resembling Kv1.3) are activated reversibly under hypo-osmotic (cell swelling) conditions l34 (see paragraph 48-14-04 and Fig. 3, p. 419). Although this study suggested that both Kir and CI- channels are necessary for proliferation of rat microglia stimulated by colony-stimulating

_'--

e_n_try_4_8_

in hypotonic medium (e.g. for CTLL, see text, 25% PBS containing r1L-2 at 8.3 ng/ml)

in hypotonic medium

Cell swelling and retu rn to isotonic volume

1. Volume expansion 2. CI channel opening (depolarizing cell towards Eel, e.g. between -10 and -35 mV)

wale/"

3. Opening of Kvl.3 rc~:~eIS [accompanied by

/

RVe Figure 2. Simplified model for crLL regulatory volume decrease (RVD) following hypotonic shock. (From 48-14-05) factor 1 (CSF-l), it found no evidence that 'even a transient activation' of Kv channels were necessary for the CSF-stimulated proliferation process l34 .

A Kvl.3 homologue with predicted roles in renal medullary K+

transport 48-14-06: Rabbit Kvl.3; Rabbit Kvl.2: The potassium conductance of the kidney medullary (papillary epithelial) cell line GRB-PAPI is composed of Shaker-like potassium channels. From RT-PCRt studies (see notes to Table 1 under Gene family, 48-05) candidate genes encoding these channels include Kvl.l, Kvl.2, Kvl.4 and latterly, the 'glibenclamide-sensitive' rabK(vl.3) clone34, which is also expressed in native renal medulla and brain. The latter encodes an ORF of 513 aa, and is related to but not identical to other Kvl.3 isolates (e.g. in having a different N-terminus and a single-channel conductance that saturates following expression in oocytes). Note: Independent electrophysiological studies 135 have shown that GRBPAP1 cells (grown on permeable supports and capable of developing electrogenic Na+ transport) display a 'dominant' K+current' that is slowly inactivating, time- and voltage-dependent; in this study, RT-PCR of GRBPAPI mRNA amplified a segment of Kvl.2.

l_e_n_t_ry_48

_

IDominant' effects of Kvl.3 on resting membrane potentials in native and heterologous cells 48-14-07: Kvl.3: An outwardly rectifying K+ channel in osteoclasts with many properties resembling those of Kvl.3 has been described136. Application of charybdotoxin (50nM, see Blockers, 48-43) changes osteoclast resting membrane potential (inducing a depolarizing shift of 5-10mV from a typical resting potential of -50mV). Note: RT-PCRt on osteoclast RNA pools has identified expression of both Kvl.3 and Kir2.1/IRKl (entry 33). rKvl.3: Sensitivity to the non-selective blocker charybdotoxin (CTx) (see Blockers, 48-43) has been used to argue a contribution of Kvl.3 to the repolarization of action potentials at pre-synaptic terminals of hippocampal inhibitory neurones 137j CTx-induced facilitation of transmission may be partly explained by its effects on Kv channels rather than large-conductance, calcium-activated channels (BKca, see ILG K Ca, entry 27). hKvl.3: Heterologous expression of human Kvl.3 'resets' the resting potential of chinese hamster ovary (CHO) cells, 'clamping' it within a narrow range close to the threshold of activation of Kvl.3 138. This property partly depends on Kvl.3's intrinsically steep voltage dependence of activation and slow/ incomplete inactivation. Inhibitors of Kvl.3 such as margatoxin depolarize transfected CHO cells to the potential of non-transfected cells 138 .

Membrane potential regulation in excitation-secretion coupling of pancreatic j3 cells 48-14-08: hKvl.5/hPCNl: Stable transgenic t overexpression of the tetraethylammonium (TEA)-insensitive hKvl.5 channel in pancreatic cells of transgenic mice attenuates glucose-activated increases in [Ca 2 +]i and prevents the induction of TEA-dependent [Ca2 +]i oscillations 139 (for further background, see ILG Ca InsP3, entry 19 and ILG K Ca, entry 27). Augmentation of expression in pancreatic /3 cells was confirmed by immunoblot studies of isolated islets and immunohistochemical analysis of pancreas sections. Whole-cell current recordings showed the presence of high-amplitude TEAresistant K+ currents in transgenic islet cells, whose expression correlated with hyperglycaemia and hypoinsulinaemia. Note: hPCNl was originally derived from a human insulinoma line28 .

Kv channel phenotypes in non-excitable Ltk- mouse fibroblasts 48-14-09: hKvl.5: Following stable t expression in Ltk- mouse fibroblasts (Lcells) the fast-activating, non-inactivating hKvl.5 channel (i) prevents dexamethasone-induced increases in intracellular volume and (ii) inhibits Na+ / K+ -ATPase activity by 25 % (as measured by 86Rb+ uptake)140. Independent measurements of alanine transport was also lower in Kv1.5-expressing cells, indicating that the expression of this channel modifies Na+-dependent amino acid transport. Expression of the rapidly inactivating subunit Kv1.4 did not alter alanine transport relative to wild-type or sham-transfectedt cells. The properties induced by Kv1.5 expression may be related to changes in the resting membrane potential induced by this channel (-30 mV) in contrast to that measured in wild-type sham-transfected, or Kv1.4transfected cells (-2-0mV). Quinidine block of hKv1.5 (60J,!M) negates the effects of Kv1.5 expression on intracellular volume, Na+ /K+ -ATPase, and

II

_ _ _ _ _ _ _ _ _ _.

e_n_try_4_8_

electrogenic Na+ -dependent alanine transport 140. Comparative note: Kvl.5: Although human airway smooth muscle cells express mRNA from several members of the Kvl gene subfamily (see mRNA distribution, 48-13) Kvl.5 appears to 'predominate' voltage-dependent current and the regulation of basal tone 120.

'Closely matching' heterologously expressed Kv subunit channels to 'native' types 48·14·10: The difficulties involved in 'predicting' specified Kv subunit 'contributions' to voltage-gated K+ currents observed in native cells is outlined under Channel designation, 48-03 and in VLG K Kv-beta, entry 47). In the brain, for example, multiple Kv gene products are likely to be involved in regulating excitability of any single neurone. Despite the existence of multiple mechanisms which generate K+ current diversity (see Developmental regulation, 48-11), several descriptions have appeared where native and heterologously expressed channel properties exhibit 'notable similarity' (see Table 4).

Physiological roles deduced from spontaneous mutants and gene knockout (background) 48·14·20: Heritable phenotypes associated with mutated or deleted gene loci provide a potentially powerful method for analysing specific contributions of Kv and other channels to higher organismal function. In humans, patterns of disease or symptomatic disorders (observed in families or populations) may follow laws of Mendelian inheritance t. As such, disease phenotypes may be linked to the specific inheritance of 'defective' gene products or closely linked markers! (ibid.). In this way, a number of physiological disorders have been associated with mutations in specific ion channel genes. The well-established roles of K+ channels in the nervous system makes them important candidates for heritable and sporadic forms of neurological disease both in humans and in animals. For this reason, much effort is being directed to gene mapping/disease linkage studies, and the production of transgenic t animal models of disease carrying targeted disruptions or specified mutations in cloned genes (see Gene knockout t in the on-line glossary). Phenotype analysis of gene knockout mice may not be straightforward, especially for gene products that have important roles in development or that commonly form complexes with other proteins in native cells. Furthermore, many null t mutants have no apparent phenotypic changes, or may require detailed long-term observation, behavioural analysis or extensive histological comparisons to reveal them. To complement these approaches, genetic analysis of several extant mouse mutants (including those exhibiting certain neurological disturbances) is also underway. As described above, these studies attempt to find evidence for linkage of disease markers to known or predicted chromosomal loci/regions encoding specified gene product(s). For comparative purposes, Table 7 (under Chromosomal location, 48-18) also includes some cross-referenced examples of these, relating known mutant phenotype/markers to regions including Kv channel gene loci.

lL...--_ _ _ry_4_8

_

en t

Table 4. Notable similarities between tc10ned' and tnative' voltage-gated K+ channel properties. See general tcautionary' notes in this field and under Channel designation, 48-03. For important contributions of Kv{3 subunits to native channel phenotypes, see VLC K Kv-beta, entry 47. (From 48-14-10) Kv

Properties and cross-references

Kvl.l

48-14-11: Kvl.l-RBKl: RBKI has been cited as contributing to the dominant native voltage-gated K+ current in C6 astrocytoma (glial) cells 80 (compare Kv1.5, below).

Kvl.2

48-14-12: rKvl.2/RAK: IKv 1. 2 in Xenopus oocytes (isolate RAK, one amino acid difference from isolate BK2) 'compare closely' to native rat neuronal delayed rectifier-type currents in cardiac atrium (see Activation, 48-33 and ref.24).

Kvl.3

48-14-13: Kvl.3/KV3: The properties of heterologously expressed Kvl.3 and Kv3.1 subunits correspond closely to the 'predominant' voltage-activated K+ channels of native T lymphocytes (n-type and 1type respectively; for further details see 141 , VLC K Kv3-Shaw, entry 50 and cross-references to Kv1.3 in this entry). 'Kvl.3-like' currents have also been described in a kidney medulla cell line and osteoclast preparations (see Phenotypic expression, 48-14).

Kvl.4

48-14-14: rKvl.4: For prominent immunostaining in axons and terminals, see also Subcellular locations, 48-16. 48-14-15: rKv1.4/RGHK9: The A-type current of rat pituitary tumour cell line GH3 /B-6 (as compared to the biophysical and pharmacological properties of the rKvl.4 isolate RGHK9 cDNA expressed in Xenopus oocytes following RT-PCRt from a GH3 /B-6 poly(A)+ template 124 ). Although the activation kinetics differ, the properties of 'cloned' and 'native' currents are indistinguishable with respect to (i) 4-aminopyridine block; (ii) voltage dependence and slope of steady-state activation and inactivation and (iii) slow recovery from inactivation and time constant values 124. 48-14-16: rKvl.4/RHKl: The isolates RHK1 37 (r-vKvl.4 isolate RCK4 22 ) the clone FKl (ferret Kvl.4)4o and clone HKl (hKvl.4)36 were originally described as displaying fast activation and inactivation 'similar' to the transient outward current (ITO, entry 44) described in native rat ventricular myocytes 142. Notably, however, Kvl.4 has more recently been shown to be a low-abundance protein in rat atrial and ventricular myocytes l14; in consequence, its role in myocyte excitability has been questioned (ibid.). For further determinants of inactivation behaviour in Kv channel complexes, see VLC K Kv-beta, entry 47.

Kvl.5

48-14-17: hKvl.5/fHK: Intact human cardiac atrial myocytes display a rapid delayed rectifying K+ current with properties and kinetics 'identical to' those expressed by a K+ channel isolate (fHK) cloned from human heart eDNA and stably expressed in a human-derived cellline143 (for comparative data, see references to Kv1.5 (isolate

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_8_



Kv1.5

fHK) under Activation, 48-33, Single-channel data, 48-41 and Blockers, 48-43). Although there is some 'correspondence' between the cloned and native channel properties, the 'precise' contribution of Kv1.5 channel current to the cardiac action potential remains unclear. See also/compare direct evidence for roles of erg subfamily channels (e.g. HERG) underlying native cardiac currents under Phenotypic expression of VLC K eag/elk/erg, 46-14. 48-14-18: hKv1.5/HK2: Although not conclusive (see above) Ltkcell line stably expressed t HK2 (cloned from human cardiac ventricular cDNA) displays has also been specifically stated to possess many similarities to the rapidly activating delayed rectifying currents described in adult rat atrial and neonatal canine epicardial myocytes l44 . If verified, this would suggest that human Kv 1.5 contributes to the initial fast repolarization and to the K+ conductance during the plateau phase of the cardiac action potential l44 . Note: HK2 current does not resemble either the rapid or the slow components of delayed rectification described in guinea-pig myocytes. 48-14-19: rKv1.5: A specific contribution of Kv1.5 to the delayed rectifying K+ current of spinal cord astrocytes (see Protein distribution, 48-15) has been proposed following incubation of these cells with antisense oligodeoxynucleotides 125 . Antisense treatment (i) reduces delayed rectifier current density and (ii) shifts the potassium current steady-state inactivation (without altering current activation, cell capacitance or cell resting potential) (compare Kvl.l, above).

Notes: Historically, cloning of the Drosophila Shaker (see Related sources and reviews, 48-56) established that altered neurological phenotypes coupled to locomotor dysfunctions could be directly related to mutations in genes encoding voltage-gated potassium channels. Through the extensive use of behavioural mutants, comparative genetic analyses of the major K+ currents in embryonic Drosophila neurones and muscle have been performed, and the 'relative contribution' of the Shaker, Shab, Shaw and Shal loci have been determined (for brief summary, see Phenotypic expression under VLC K Kv2-Shab, 49-14). Detailed comparisons of 'native' versus 'cloned' channels encoded by the Shaker locus in Drosophila are also included in references listed under Related sources and reviews, 48-56; see also properties of novel potassium channels in Drosophila photoreceptors described in ref. 145 . Additional references to Drosophila K+ channel genes in relation to their mammalian counterparts are described under entry 27 (ILC K Ca, Drosophila slo, mouse mslo, human hslo); entry 46 (VLC K eag/elk/ erg, Drosophila eag-related loci including HERG) and entry 47 (Drosophila hyperkinetic, related to mammalian Kv{3 subunits).

II

1'--_e_n_t_ry_4_8

_

Protein distribution KVQ subunit components of immunoprecipitated DTx acceptor complexes 48-15-01: Subtypes of a-dendrotoxint (o-DTx)-susceptible K+ channels are prevalent throughout mammalian brain and are composed of heterogeneous K+ channel protein assemblies l46 that also contain auxiliary Kv/J subunits147 (for details, see VLC K Kv-beta, entry 47). Immunoprecipitationt of the DTx acceptor complext in bovine brain using a monoclonal t antibody selective for Kvl.2 (mAb 5) and polyclonal antibodies for other Kvo subunits indicate that Kvl.l, Kvl.4 and Kvl.6 can be components of the acceptorl46 . While 'virtually all' DTx receptors contain the Kv1.20 subunit l46,148,149, the distribution of other o subunits in the complex can vary considerably across different brain regions, as summarized in Table 5. Autoradiographic co-distribution studies using affinitypurified anti-Kyo and anti-Kv,8 subunit antibodies and 125I-labeled DTx have shown that (i) the areal and laminar distribution of o-DTx acceptors correspond most closely to Kval.2 and Kv{32 immunoreactivity149; (ii) o-DTx binding also corresponds closely to Kvl.l and Kvl.6 immunoreactivity in the cerebral and cerebellar cortices, hippocampus and in fibre pathways and (iii) with the exception of the hippocampus and entorhinal cortex, [125 I]0-DTx binding does not correspond closely to Kvl.4 or Kv,8l staining patterns. Kvl.l, Kvl.4, Kvl.6 and Kv,8l components of o-DTx acceptors may be limited to specific projectiont systems. Note: o-DTx is a 56 amino acid peptide isolated from the venom of the African green mamba. It binds selectively and with high affinity to voltage-gated K+ channels in mammalian brain.

Comparative studies examining multiple Kv protein distribution in heart and brain 48-15-02: Westernt blotting of cardiac atrial versus ventricular membrane proteins with a panel of anti-Kv channel has confirmed the presence of Kvl.2, Kvl.S, Kv2.1 and Kv4.2 in heart and revealed differences in the relative abundances of these subunits in the two membrane preparations l14 . Kvl.5 levels are weaker, but comparable in the two preparations; Kv2.1 (entry 49) and Kv4.2 (entry 51) abundances appear higher in atrial membranes. Anti-Kvl.4 antibodies reveal Kvl.4 is not an abundant protein in adult rat atrial or ventricular myocytes l14 (for further details, see Table 5). In a separate extensive study 131, specific polyclonal antibodies for five Kvl subfamily 0 subunits (Kvl.l, Kvl.2, Kvl.3, Kvl.4, Kvl.6) were used to determine detailed distribution patterns in rat hippocampus, cerebellum, olfactory bulb and spinal cord. This study 131 extended but partly contradicted the previously reported64 'stereotypical' targeting of Kvl.4 channel 0 subunits to axonal compartments (e.g. in reporting a part-dendritic localization of Kvl.4 protein in olfactory bulb mitral cell in addition to axons).

Subcellular locations General significance of Kv protein subcellular locations in neurones 48-16-01: The particular 'functional role' of any given set of ion channels may vary depending on their 'precise location' on the cell surface or within

11II

_L...--

,

e_n_try_4_8_

Table 5. Summary of KvO'. subunit co-distributions as detected in brain by immunoprobes 146 . See also VLC K Kv-beta, entry 47, mRNA distribution, 48-13 and Subcellular locations, 48-16. (From 48-15-01) Kv


Kvl.l

48-15-03: Kvl.l: Variable amounts of Kvl.l subunits are observed in K+ channels purified from cerebellum, corpus striatum, hippocampus, cerebral cortex and brainstem, with a notably 'larger proportion' of Kvl.l in DTx acceptor complexes isolated from the cerebral cortex and brainstem. 48-15-04: Kvl.l/Kvl.2: In the hippocampus, both mKvl.l and mKvl.2 proteins are present in axons, often near or at synaptic terminals in the middle molecular layer of the dentate gyrus, while only mKv1.1 is detected in axons and synaptic terminals in the hilar/ CA3 region150. In the cerebellum, both mKvl.l and mKvl.2 are localized to axon terminals and specialized junctions among axons in the plexus region of basket cells. Strong differential staining is observed in the olfactory bulb, where mKvl.2 is localized to cell somata and axons, as well as to proximal dendrites of the mitral cells (see also Subcellular locations, 48-16). 48-15-05: Kvl.l/Kvl.2: Antibodies recognizing Kvl.l, Kvl.2 and synapse-associated protein 90 (SAP90) are predominantly localized (concentrated) in septate-like junctional regions, which connect the basket cell axonal branchlets in the Pinceau (a network of cerebellar basket cell axon branchlets surrounding the initial segment of the Purkinje cell axon)151. These results are consistent with SAP90 assisting the formation of Kvl.l/Kvl.2 heteromultimers in cerebellar Pinceaux junctions. Comparative note: By contrast; (i) Kv3.4 (entry 50) is uniformly distributed over the Pinceau and the pericellular basket surrounding the Purkinje cell soma (i.e. localized to the Purkinje cell axon initial segment) and (ii) voltage-gated Na+ channels (entry 55) are not detectable in the Pinceau, but are localized to the Purkinje cell axon initial segment 151 . 48-15-06: Schwann cell staining ~,.yith anti-Kvl.l antibodies reveals high concentrations of Kvl.l (i) in the axonal membrane at juxtaparanodal regions and (ii) intracellularly, within perinuclear compartments. Although it is likely that intracellular immature forms of these (compartmentalized) signals repn:~sent subunits, comparison with Kvl.5 distributions (this table) indicate that closely related Kv channels (from the same subfamily) need not co-assemble and can be localized differentially in the same cell152 (see also N-glycosylation under Sequence motifs, 48-24 and Protein interactions, 48-31).

Kvl.2

48-15-07: Kvl.2: Kvl.2 subunits are uniformly distributed in brain regions listed under Kvl.l. Kvl.2 co-immunoprecipitates with Kvl.l in rat brain membranes, and has been co-localized with Kvl.l in the juxtaparanodal regions at nodes of Ranvier in myelinated axons and

----l_

l"----e_n_t_ry_48



Kvl.3

in terminal fields of basket cells in mouse cerebellum 153,154 (see above and note below). Kv1.2 is co-localized with Kv1.4 in the dentate gyrus 132 and co-precipitates with Kv1.4 within rat brain membrane preparations. Note: A striking example of co-localization between Kv1.2 and PSD-95 occurs in the basket cells 155 (for significance see Fig. 7 under Protein interactions, 48-31).

Kvl.4

48-15-08: Kv1.4: See Subcellular locations, 48-16 for predominant axonal/terminal distribution of Kv1.4 protein. Variable amounts of Kv1.4 subunits are observed in K+ channels purified from the brain regions listed under Kv1.1, with a larger proportion of Kv1.4 being found in DTx acceptor complexes from the hippocampus. Note: Kv1.4 is not an abundant protein in adult rat atrial or ventricular myocytes l14 : A 'very faint' band was detected at 97 kDa in atrial and ventricular preparations when an anti-Kv1.4 antibody (that reveals intense Kv1.4 expression in brain) is used at a 5- to 10-fold higher concentration (see Table 4 under Phenotypic expression, 48-14).

Kvl.5

48-15-09: Kv1.5 and Kv1.1 proteins have distinct distributions in Schwann cells. Anti-Kv1.5 antibodies have localized subunit expression to (i) the Schwann cell membrane at the nodes of Ranvier; (ii) 'bands' that run along the outer surface of the myelin and (iii) intracellular distributions in the vicinity of the nucleus 152 (see also Kv1.1, this table). rKv1.5: By immunohistochemistry: Kv1.5 protein is abundant in glial cells of adult rat hippocampal and cerebellar slices, as well as in cultured spinal cord astrocytes. Note: Kv1.5 immunoreactivity was described as particularly intense in the endfoot processes of astrocytes surrounding the microvasculature of the hippocampus 125 (see also Phenotypic expression, 48-14).

Kvl.6

48-15-10: Kv1.6: Kv1.6 subunits are uniformly distributed in those brain regions listed under Kv1.1.

Notes: 1. In this study 146 Kv-specific antibodies generally precipitated a different proportion of the channels detectable with radioiodinatedt a-DTx in every brain region (anti-Kv1.2> 1.1» 1.6> 1.4), consistent with a widespread distribution of hetero-oligomeric subtypes (see Protein interactions, 48-31). 2. For examples of immunohistochemical localizations of Kv1 subfamily channels in single-cell types coupled with functional studies64,156 described under Subcellular locations, 48-16. internal membranes. As pointed out in refs. 64,157, this is of particular significance for K+ channels in neuronal membranes, which have the capacity to respond uniquely to a given input, and where neuronal integration depends on local responses of spatially segregated inputs to the

III

_ _ _ _ _ _ _ _ _ _.

e_nt_ry_4_8_

cell and the communication of these 'integration centres' via the axon. The subunit diversity shown by the Kv channel family is thus of major significance for supporting a multiplicity of functional roles (via formation of many multiple channel types) and subcellular localizations or 'targeting' of channels may extend this functional diversity still further. For example, differential subcellular and regional distributions of Kv subunits forming Atype t channels (rKv1.4 and rKv4.2) imply distinct roles for these channel proteins in vivo (this field, below, Protein distribution, 48-13 and corresponding entry/fieldnames under VLC K Kv4-Shal (i.e.. 51-13, 51-16). Kv1 subfamily Q subunits interact through their cytoplasmic C-terminus with a family of membrane-associated (putative) guanylate kinases, including PSD95 and SAP97. Heterologous co-expression of either SAP97 or PSD-9S with a number of Shaker-type subunits induces 'co-clustering' of the kinase homologues with the Kv channels 62 (for types of clustering induced by PSD-95 and SAP97, see Protein interactions, 48-31; see also effects following mutation of the C-terminal lEDTV' motif in Kv1.4 under Sequence motifs, 48-24).

Co-location of Kvl.l/Kvl.2 in ;uxtaparanodal regions of myelinated axons 48-16-02: Kv1.1/Kv1.2: mKv1.1 and mKv1.2 polypeptides (probably as heteromultimers) occur in subcellular regions where rapid membrane repolarization may be important (juxtaparanodal r regions of nodes of Ranvier in myelinated axons t and terminal fields t of basket cells t 153. In separate studies, mKv1.1 and mKv1.2 proteins have also been shown to be present in unmyelinated axons, specialized junctions between axons, and proximal dendrites 150 (for broader distributions, see Table 5 under Protein distribution, 48-15).

Kvl.2 subcellular locations appear distinct in different cell types 48-16-03: Kv1.2: Kv1.2 protein immunoreactivity shows a complex differential subcellular distribution within different neurones of rat brain 158 . At certain loci (e.g. hippocampal and cortical pyramidal cell and Purkinje cells) Kv1.2 appears concentrated in dendrites. In other neurones such as cerebellar basket cells, Kv1.2 is predominantly (if not exclusively) localized to nerve terminals 158, although some immunoreactivity is associated with certain axon tracts.

Comparison of temporal Kv protein expression patterns observed in situ versus in vitro 48-16-04: While the timing of expression for Kv1.4, Kvl.S, Kv2.1, Kv2.2 and Kv4.2 polypeptides observed in situ within the hippocampus appear to be retained in vitro, properties such as cellular and subcellular localization appear to be different 103 . While similarities in Kv protein expression in situ and in vitro indicate the same regulatory mechanisms control spatiotemporal patterning, observed differences between levels of expression for all subtypes studied (except Kv2.1) indicate additional mechanisms operating in situ which mediate Kv channel abundance 103 . Kv1.4 protein is targeted to axons and possibly terminals, suggesting a pre-synaptic role in synaptic transmission64 (for some exceptions, e.g. olfactory bulb mitral cell dendrites,

III

lL...--e_n_t_ry_4_8

----I_

see ref. 131 ) (see also the general contrasting subcellular locations between the A-type channel subunits Kv1.4 and Kv4.2 described under VLC K Kv4Shal, entry 51).

Distributions of Kvl subfamily channels in rod bipolar cells of mice 48-16-05: mKv1.1; mKv1.2; mKv1.3: Generally, immunohistochemical localizations in single-cell types show Kv channel subunits to have a unique subcellular distribution (see above). For example, in enzymatically dissociated isolated bipolar cells of mice Kv1.1 immunoreactivity is detected in the dendrites and axon terminals, whereas Kv1.2 and Kv1.3 subunits are localized to the axon and the post-synaptic membrane of the rod ribbon synapse, respectively122,159. Whole-cell patch-clamp studies on the same preparation indicate that the activation voltage of the native delayed rectifier current (IK ) of the isolated bipolar cell and the inhibitory constants for current blockade by TEA, 4-AP, and Ba2 + are 'most similar' to properties measured for Kv1.1 (as expressed in oocytes; TEA and 4-AP inhibitory constants for native I K differ from the inhibitory constants for Kv1.2 or Kv1.3 in oocytes). Despite these differences, all three channels are likely to function in the intact retina to allow complex modulation of retinal synaptic signals 122. On the basis of retinal Kv subunit distribution studies, it has been proposed that each Kv channel subtype is associated with a specific subcellular functional module, and each local K+ conductance responds uniquely to local voltage and second messenger signals 159. Notably, no single Kv channel subtype is expressed over the entire length of retinal neurones.

Role of Kvj32 subunits in promoting surface expression of Kv channel subunit complexes 48-16-06: Kv1.2: Heterologous co-expression of Kval.2 with Kv,B2 has been shown to promote the transport of aKv1.2 to the cell surface (for further details, see Subcellular locations under VLC K Kv-beta, 47-16). Comparative notes: 1. Amongst the Kv,B subunit variants initially isolated and characterized, Kv,B2 was unusual as it did not exert any marked changes on the inactivation properties of co-expressed Kv1 family subunit channels (ibid.). 2. Prolonged treatment of oocytes expressing Kv1.1 with (S-p)-8-BrcAMPS induces channel phosphorylation and can direct channel expression to the plasma membrane 160 (for further details, see Table 16 under Protein

phosphorylation, 48-32).

Polarized expression of Kvl.4 to basolateral membrane of MDCK cells 48-16-07: Kv1.4: Immunocytochemistry and confocal microscopy of the polarized epithelial cell line MDCK shows specific basolateral membrane localization of Kv1.4, detectable in two forms (glycosylated and nonglycosylated)161 (see also the Kv2.1 cytoplasmic domain associated with polarized (lateral membrane) expression and clustering in MDCK cells under Subcellular locations, 49-16).

III

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_8_

Table 6. Data pertaining to RNA transcripts from genes encoding Kv1 subfamily K+ channel proteins (From 48-17-01) Kv1.1

48-17-02: hKv1.1: 9.5kb 120 - human airway smooth muscle (see

Phenotypic expression, 48-14). 48-17-03: mKv1.1: MBK1/MK1: Single ~8kb transcript, probing brain poly (A)+. Note: Relatively large for the channel protein size. RCK1: 1.0-5.0kb detected, 4-5kb being consistent with RCK 1 cDNA size10. RK1: 8 kb, also hybridizes to less-intense bands of 3 and 4kb. Kv1.2

48-17-04: MK2/RK2/RCK5: Hybridizes to a ~10kb transcript (consistent with RCK5-probed blots). RCK2: Restricted expression, chiefly in brain; probes from 3' UTR t detects a 'high level' ~6 kb species in neonatal and adult brain RNA; short RNA species of ~0.6 and 0.8 kb can be seen in brain and adult skeletal muscle respectivelyl18. BK2: ~9.5 kb 15. hKv1.2: 11 kb 120 (human airway smooth muscle)

Kv1.3

48-17-05: RGK5/isolate KV3/MK3: 9.5 kb.

Kv1.4

48-17-06: HK1: Two transcripts of 3.5 kb and 8 kb (both poly(A)+ and total cardiac mRNAs). RCK4: ~4.3 kb in heart and other tissues (see Cell-type expression index, 48-08). RK3: strongest hybridization to a 4.5 kb transcript in heart RNA; 15 kb and 1.5 kb bands also reported. 48-17-07: rKv1.4: rKv1.4 (isolate RCK4)-specific probes detect an mRNA species of ~4300 nt in heart which increases during development 101 . The detection of 'long transcripts' (ca. 80009500nt) from the Kv1.1, Kv1.2 and Kv1.3 (intronless t ) genes indicates that a relatively large proportion of the mRNA is untranslated, e.g. in the isolate rnKv1.1/MBKl this would indicate >6 kb (from approx. 8 kb) is untranslated if the observed polyadenylation sites are used in vivo. 48-17-08: Methodological note: l'ranslation of Kv1.4 employing

an internal ribosome entry sequence (IRESt) has been noted162. IRES elements enable a single RNA transcript encoding two different proteins to be 'driven' from the same promoter (Le. on a single transcript) but translated independently. This has several applications in ion channel expression, for example, in coupling expression of an ion channel subunit to a selectable resistance protein product (e.g. neomycin for stable expression) or to an intrinsic expression reporter protein (e.g. green fluorescent protein t ).

Kv1.5

48-17-09: hKv1.5: 3.5 plus 4.4kb (human airway smooth muscle)120 (see Phenotypic expression, 48-14). HK2: In RNAase protection assays, the HK2 sequence protects 2.5 kb and 1.5 kb messages; neither is sufficient for full-length HK2. hPCN1: A single, low-abundance 4 kb transcript reported in insulinoma, RIN5FS and HIT m2.2 cell RNA. RK4: 3 kb transcript in heart RNA. Isolate KV1: 3.5 kb.

---J_

1L.-_e_n_t _ry_48

Table 6. Continued Kvl.5

48-17-10: mKvl.5: Two alternatively spliced mRNAs of Kvl.5 have been described in mouse heart45 : A long form, encoding a 602 aa protein, and a short form (Kvl-5Ll5'), where the first 200 N-terminal amino acids upstream of the transmembrane segment S1 is deleted. By RNAase protection t both variants are present in a wide range of tissues (e.g. heart, brain and thymus), with the long form being described as 'predominant'. It was deduced that Kvl-5Ll5' arises by an intra-exonic splicing event. An additional short cDNA clone designated as Kvl-5Ll3' encoding a non-functional C-terminal truncated protein (that inhibited expression of the long isoform) was also described in this study45.

Kvl.6

48-17-11: HBK2: Hybridization of an RCK2 probe with total rat brain

mRNA followed by RNAase protection t produced a resistant band of 6800 nt. Isolate KV2: 6.5 kb.

Transcript size Significance of reported transcript sizes 48-17-01: Transcript size ranges as reported for different Kvl channel gene

subfamily members are listed in Table 6. Amongst the subfamily, there are some notable instances of 'long transcripts' comprising much untranslated RNA (e.g. Kvl.l, Kvl.2, Kvl.3) and also evidence for alternative splicing of mRNA transcripts (see Table 9 under Gene organization, 48-20 and Kv1.5, Table 6). Reporting of similar-sized transcripts with small variations in predicted coding region lengths (rvl-lOaa) between separate studies may indicate the isolation of different alleles t of the same Kv gene (see Protein molecular weight (purified), 48-22).

SEQUENCE ANALYSES The symbol [PDTM] denotes an illustrated feature on the generalized Kv channel protein domain topography model (Fig. 6).

Chromosomal location Mechanisms contributing to the origin of the Kv multigene family 48-18-01: The complexity of higher eukaryotic genomes reflects contributions

from many different types of genetic duplication, recombinational exchange and transfer mechanism that have operated at several levels during evolution (for revie~ see Lewin, 1994, under Related sources and reviews, 48-56). As discussed in ref. 54, the 'modern' distribution of K+ channel genes can be partly explained by invoking basic mechanisms such as gene duplicatio~ chromosomal duplication and rearrangement, changes in whole ploidy I number (genome duplicationt) and/or differential gene silencingt . In

III

_1.......-

e_n_try_4_8-----1

particular, localized, tandem gene duplication (followed by sequence divergence) is associated with the generation of clusters of functionally related genes, and this appears typical for Kv channel gene arrangements (this field, below and Fig. 3). Large-scale duplications and/or rearrangements (e.g. affecting whole chromosomes or substantial segments of them) can also generate contiguous segments of linked genes that have a corresponding (paralogoust) region of related, linked genes elsewhere in the genome. The 'clustered' distribution of several groups of Kv channel genes in the mouse has confirmed and extended the existence of such paralogous regions54 (ibid.). Aberrant recombinationt events such as unequal crossing-overr and possibly excision-insertion events mediated by retrovirusest or retrotransposonst 163 may also 'isolate' (separate) genes from clusters and 'place' them at distal genomic loci. Initial analyses of known Kv channel gene chromosome localizations (Table 7) have identified several apparently 'isolated' genes, which may have undergone transplacement by these or other mechanisms.

Kv genes are clustered in paralogous regions of the mouse and human genomes 48-18-02: Striking evidence for Kv gene clustering t at chromosomal loci has been found in the mouse genome (i.e. where more than one K+ channel gene maps to the same chromosomal region)54. In this study, interspecific backcross (IB) analysist (this field, below) was used to identify 'tight' linkage groupst on mouse chromosome 3 (containing six linked Kv genes) and mouse chromosome 6 (containing four linked Kv genes). Some evidence for paralogy also exists for regions on mouse chromosome 7. Notably, Kv genes from distantly related subfamilies can be located within the same cluster (Table 7, below and Fig. 3). A separate analysis2 of Kv channel genes from the Shaker and Shaw subfamilies have been localized to probable clusters on human chromosomes 11 and 19q13, suggesting these regions may also be paralogous. Interspecies homologies between chromosomal segments encompassing Kv channel genes on mouse chromosomes 3, 6 and 7 suggests regions of human chromosomes Ip, 12p and 19q may also be paralogous54 (see also Table 8). These findings have important implications for deducing possible evolutionary mechanisms giving rise to clustering (see Figure 3). The chromosome 3 and 6 clusters are described as part of Table 7. Comparative note: In Drosophila, the genes encoding Shaker, Shab, Shaw, and Shal do not appear to be clustered in the genome l64 .

Supplementary note: Regions of synteny between mouse and human chromosomes 48-18-03: 'Absolute' chromosome locus mapping of all Kv channel gene family members is incomplete (at the time of compilation). In the absence of direct evidence (and because definition of 'exact' map positions in relation to multiple markers in humans is relatively difficult) a number of studies have 'inferred' or 'predicted' human chromosomal loci from known map positions determined in mouse (and vice versa). These predictions rely on maps of linkage t and synteny homologies t between mouse and human (e.g. see refs. 165-167). These maps have been applied in chromosome localization studies of a wide variety of cloned genes in mouse {for background, see

II

r--------------- Present-day 'isolated' Kv genes (unlinked) - ...I

I I

:

Kcn

I

:

I

Encoding primordial channel gene (ancestral)

(see text under Chromosomal location, 48-18)

Aberrant I segmental recombination events promoting gene dispersal

Ken

Yc.

Encoding precursor of the Shaker (Kena) & Shal (Kcnd) subfamilies

I

+

i-----------------~ Hypothetical

'Yc

I

~

""

I

h J :

(approx. 300 million years ago) I I

I I

Encoding precursor of the Shab (Kcnb) & Shaw (Kcnc) subfamilies

I I I I I I I I I

(examples, see text under Chromosomal location, 48-18) Locus: Kcnd1 rs Encoding:?

Kcna3

Kcna8

Kcnc4

mKv1.2

mKv1.3

mKv1.8

mKv3.4

Mouse chromosome 3 (synteny with human 1 p) Locus:

Multiple chromosomal rearrangements

divergence

Segregation to contemporary locations

Encoding:

Kcna1

Kcna6

Kcna5

mKv1.1

mKv1.6

mKv1.5

Mouse chromosome 6 (synteny with human 12p) Locus: Encoding:

·

Kcnb

Kcna

Kcnc

Shab

Shaker

Shaw

Locus:

·

KCNA7

KCNA3

hKv1.7

hKv1.3

Human chromosome 19q

Encoding:

Establishment of four different Kv gene subfamilies w:iQLjQ the divergence of flies and mammals (> 600 million years ago)

II

Kcna2

I

'"

Shal

I

Present-day Kv gene 'clusters' (linked)

Functional divergence

Kcnd

X

-------------------------------------------

t

Local gene duplications Sequence

X

I I I I I

genome duplication event

I

:

Ken

X

I L

hypothetical duplication

X

X X

I

KCNA4

KCNC1

hKv1.4

hKv3.1

Human chromosome 11

Figure 3. Hypothetical events in Kv channel gene evolution based on contemporary chromosomal distributions in mouse and humans. (After Lock et al. (1994) Genomics 20: 354-362.) (From 48-18-01)

('b

='

t"'t

~ ~

00

_L-..

e_n_try_4_8_

Table 7. Chromosome locations reported for Kvl subfamily genes (From 48-18-02) Consensus/summary Further details (see note 2) (for symbols see note 1) Kv1.1 hKvl.l, KCNA1, 12p13 mKvl.l, Kcnal, chromosome 6 [~]

hKvl.l/HuK(I): (Human locus name KCNA1): Human chromosome 12p1354,173-175. For evidence that mutations in KCNAl cause myokymia with periodic ataxia, see Phenotypic expression, 48-14 and OMIM entry #160120: Myokymia with Periodic Ataxia [Paroxysmal Ataxia with Neuromyotonia, Hereditary; Episodic Ataxia With Myokymia; EAM; Ataxia, episodic, with myokymia; AEM; AEMK; Episodic ataxia, Type 1; EA1]. mKvl.l: (Mouse locus name Kcnal): Mouse chromosome 6, part of a cluster within a paralogous t region (see Fig. 3).

'Kv1.1-related sequences' in the mouse genome 54 Kcnalrsl, chromosome 6 [~] Kcnalrs2, chromosome 2 Kcnalrs3, chromosome 3 [~] Kcnalrs4, chromosome 7 [~]

48-18-03: There are at least four sequences designated as 'mKvl.l-related' in the M. spretus genome (Kcnalrsl, Kcnalrs2, Kcnalrs3, Kcnalrs4,54, see below). It is possible these cross-hybridizing loci are identical to other loci identified with Kv-gene specific probes (ibid.); alternatively, these may represent unidentified K+ channel genes or pseudogenes t. Loci for these sequences have been reported as Kcnalrsl: Mouse chromosome 6, part of a cluster within a paralogous t region (see Fig. 1). Kcnalrs2: Mouse chromosome 2. Note: The locus for recessive anorexia (anx) mutation is in a similar region to that reported for Kcnalrs2 (and possibly mKvl.4) on mouse chromosome 2. Pre-weanling mice homozygous for anx are characterized by anorexia/ reduced body weight, uncoordinated gait, head weaving, hyperactivity and body tremors 176 . The gene encoding follicle-stimulating hormone B is also located in this region 175 . Kcnalrs3: Mouse chromosome 3, part of a cluster within a paralogous t region (see Fig. 1). Kcnalrs4: Mouse chromosome 7. Note: The recessive gene qv associated with the quivering phenotype is located in the same region as Kcnalrs4 and possibly mKvl.7/mKv3.3. Mice homozygous for qv are characterized by pronounced, incessant quivering, locomotor instability, partial hindleg paralysis and deafness 177 (see also mKv4.1-related sequence under Chromosomal location in VLC K Kv4-Shal, 51-18).

l_e_n_t_ry_4_8

_

Table 7. Continued Consensus/summary Further details (see note 2) (see note 1) Kvl.2 hKvl.2/HuK(IV): (Human locus name KCNA2): hKvl.2, KCNA1, [?] 12 First was tentatively assigned to human chromosome 1233; later analyses placed the gene on human or Ip mKvl.2, Kcna2, chromosome Ip54. chromosome 3 [~] mKvl.2: (Mouse locus name Kcna2): Mouse chromosome 3, part of a cluster within a paralogous t region (see Fig. 3). Kvl.3 hKvl.3, KCNA3, Ip13.3 mKvl.3, Kcna3, chromosome 3 [~] (For updates see OMIM 176263)

hKvl.3/HuK(Ill): (Human locus name KCNA3): Human chromosome Ip13.3 29,54. mKvl.3: (Mouse locus name Kcna3): Mouse chromosome 3, part of a cluster within a paralogous t region (see Fig. 3).

Kvl.4 hKvl.4, KCNA4, [?] Ilp14.1 or llq13.4q14.1 mKvl.4, Kcna4, chromosome 2

hKvl.4/hPCN2: (Human locus name KCNA4): Human chromosome Ilp14.1 175,178 or 11p14-p13 179, consistent with Kcna4 localized to the homologous segment on mouse chromosome 2 in ref. 180 (see below) or llq13.4-q14.1 181 . mKvl.4: (Mouse locus name Kcna4): Mouse chromosome 2 (compare Kcna1rs2, this table). Marker order determined as (proximal)-Acra-Kcna4Pax-6-a-Pck-l-Kras-3-Kcnbl-(distal)180.

Kvl.5 hKvl.5, KCNA5, 12p13 mKvl.5, Kcna5, chromosome 6 [~] (For updates see OMIM 176267)

hKvl.5/HuK(VI): (Human locus name KCNA5): Human chromosome 12p13 (i.e. distal short arm band 13)42,173. Note that trisomyt of the human 12p region is associated with an epileptiform disorder (trisomy 12p syndrome182) characterized by 3 Hz spike and wave discharges, but a specific link to altered K+ channel activities has not been reported to date. mKvl.5: (Mouse locus name Kcna5): Mouse chromosome 6, band F, part of a cluster within a paralogoust region (see Fig. 3).

Kvl.6 hKvl.6: (Human locus name KCNA6): Predicted to be hKvl.6, KCNA6, 12p on human chromosome 12p54,180. mKvl.6, Kcna6, mKvl.6: (Mouse locus name Kcna6): Mouse chromosome 6 [~] chromosome 6, part of a cluster within a paralogoust (For updates see region (see Fig. 3). OMIM 176257)

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _en_t_ry_4_8_

Table 7. Continued Consensus/summary Further details (see note 2) (see note 1)

II

Kvl.7 hKvl.7, KCNA7, 19q13.3 mKvl.6, Kcna6, chromosome 7 [~] (For updates see OMIM 176268)

hKvl.7: (Human locus name KCNA7): The mouse clone MK4 sequence has been used as probe to show the location of the human isoform to be located on chromosome 19q13.3 173 (K. Kalman, unpublished). mKvl.7: (Mouse locus name Kcna7): Mouse chromosome 7 (in close proximity to the gene encoding mKv3.3; K. Kalman, unpublished; see VLC K Kv3-Shaw, entry 50) (compare/see also Kcna1rs2, this table).

K(v)I.8 hKvl.8, KCNA8, Ip mKvl.6, Kcna6, chromosome 3 [~] (For updates see OMIM 176269)

hKvl.8: (Human locus name KCNA8): Predicted to be on human chromosome Ip54. mK(v)I.8: (Mouse locus name Kcna8): Mouse chromosome 3, part of a cluster within a parologous t region (see Fig. 3).

Other laltered neurological phenotype' mutants mapping to regions close to known Kcn loci in mice

Comparative notes: Chronic locomotor hyperactivity is observed in several mouse mutants (see paragraphs below and descriptions of anorexia (anx) and quivering (qv) this table; see a1s0 54). Notably, such phenotypes were used to initially identify each of the potassium channel defects defined by, for example, Drosophila Shaker, Shab, Shaw, Sha1, Slo, eag and Hk. For a consolidated, crossreferenced listing of these and other known ion channel defects to specified or candidate mutant loci, see resources such as Online Inheritance in Man (OMIM).

Opisthotonus

48-18-04: The mouse chromosome 6 region encompassing Kcnca1 (Kvl.l), Kcnca5 (Kvl.5), Kcnca6 (Kvl.6) and Kcnca1rs1 (this table) also encompasses the recessive opt (Opisthotonus) mutation54,183. Homozygous opt mice are characterized by loss of motor co-ordination at approx post-natal day 12 and subsequently develop severe seizures, leading to death at approx. 3-4 weeks 183 (but see next paragraph).

Deafwaddler

48-18-05: The mouse chromosome 6 region encompassing opt (previous paragraph) also carries the recessive dfw (deafwaddler) mutation54,184. Homozygous dfw mice are characterized by deafness, head bobbing behaviour and walking with a hesitant 'wobbly' gait l84. A detailed mapping study185 has since established the gene order on the distal arm of

l_e_n_t_ry_48

_

Table 7. Continued Consensus/summary Further details (see note 2) (see note 1) Deafwaddler

chromosome 6 to be cen-opt-dfw-Rho (D6Mit44)Kcna1, Kcna5, Kcna6, suggesting that the neurological mutants opt and dfw (above) affect two different genes, neither of which is caused by a mutation in the clustered K+ channel genes.

Other crossreferences

48-18-06: For known chromosomal locations of vertebrate Shab, Shaw and Shal subfamily genes, see FactsBook fields 49-18, 50-18 and 51-18 respectively. 48-18-07: For details on ion channel genes linking to familial human long QT syndrome, see Chromosomal location under VLC K eag/elk/erg, 46-18. 48-18-08: For trisomy 12p syndrome, see this table,

above. 48-18-09: For spinocerebellar ataxia type 2 (SCA2)

mapping to human 12q23-q24, see ref. 186 and compare to human K+ channel loci, this table. 48-18-10: For benign familial neonatal convulsions

Supplementary note on expanded trinucleotide repeat motifs (of general significance, no relationship to Kv or other ion channel gene loci reported to date).

(epilepsy) linked to genetic markers on chromosome 20q, see ref. 187 and compare to human K+ channel loci, this table. 48-18-11: 'Expansion' of the trinucleotide repeat CAG encoding polyglutamine have been associated with five hereditary neurodegenerative diseases including Huntington's disease (HD), spinocerebellar ataxia type 1 (SCA1), Machado-Joseph disease (SCA3), spinobulbar muscular atrophy (SBMA) and dentatorubal and pallidolyusian atrophy (DRPLA) (for minireview of

related references, see188 ). Each of these diseases involves a progressive loss of specific neuronal populations, and all are characterized by the phenomenon known as genetic anticipation, where symptoms appear at earlier ages and with greater severity in successive generations. Hypotheses relating expanded trinucleotide repeats to changes in signalling protein conformation have been discussed (ibid.).

Notes: 1. [?] symbol denotes possible conflict; [~] symbol denotes likely paralogous cluster. 2. This column list species equivalents (see also Table 8), notes on phenotypes (if known) and other cross-references. See also references to OMIM (Online Mendelian Inheritance in Man). OMIM entry 176260 presents a general discussion of voltage-gated K+ channels.

II

_L.-

e_n_try_4_8_

Table 8. Summary of mouse/human chromosomal syntenies or homologies used to infer or predict 'equivalence' of chromosomal segments containing Kv gene clusters. For further background, see refs165-167~169-172 (From 48-18-02) Affecting Kv localization

Mouse (known)

Kvl.l

Kcna1 on chromosome 6 Region homologous to human chromosome 12p (in paralogous cluster with Kcna5 and Kcna6) (location confirmed for KCNA5, see table 7) (see also ref. 189 in footnote)

Kvl.2

Kcna2 on chromosome 3 Between regions (in paralogous cluster) homologous to human chromosome lq and Ip (designated as 1p in54; see also Kvl.2 in Table 7 and Kv3.4, this table)

Kvl.4

Kcna4 on chromosome 2 Region homologous to human chromosome 11p

Kvl.5

Kcna50n chromosome 6 Region homologous to (in paralogous cluster) human chromosome 12p

Kvl.6

Kcna6 on chromosome 6 Region homologous to (in paralogous cluster) human chromosome 12p

Kvl.8

Kcna8 on chromosome 3 Region homologous to (in paralogous cluster) human chromosome Ip Kcnb1 on distal Region homologous to chromosome 2 human chromosome 20q Kcnc2 on chromosome 10 Region homologous to human chromosome 12q Kcnc4 on chromosome 3 Region known (human (in paralogous cluster) chromosome Ip21) Kcnd2 on chromosome 6 Region homologous to human chromosome 7q

Comparative note only Kv2.1, see entry 49 Comparative note only Kv3.2, see entry 50 Comparative note only Kv3.4, see entry 50 Comparative note only Kv3.4, see entry 50 Comparative note only Kv4.1, see entry 51 Comparative note only Kv4.1-related sequence see entry 51

Kcnd1 on proximal X chromosome Kcnd1rs on chromosome 3

Human (predicted)

Region homologous to the petite arm of human X Region homologous to human chromosome 1p

Note: Ref. 189 described yeast artificial chromosomes (YAC) propagating 1 Mb segments of human chromosome 12p13 with KCNA1, KCNA5, KCNA6 coclustered within a segment of approx. 300 kb size; these genes are organized 'head-to-tail' and are transcribed from the same DNA strand in the order KCNA6-KCNA1-KCNA5. Differential mRNA expression patterns of these genes (field 13) suggest they are transcribed independently of each other189.

II

l_e_n_t_ry_48

_

ref. 168). Patterns of co-segregation with specific linked genes within interspecific backcross (IB) mapping panels (e.g. the >1300 gene panel used for m analyses of between C57BL/6J x M. spretus strains) have major applications in deduction of syntenic t and paralogous t regions between the murine and human genomes 165,166,169-172. For further technical background and references to mapping of cloned DNAs in the mouse, see Chromosomal location under Resource D - Diagnostic tests.

Encoding A tglobal' sequence alignment of Kvl subfamily members 48-19-01: An amino acid sequence alignment, representing the majority of Kvl subfamily proteins is shown in Fig. 4. This figure denotes several general features of Kvl sequences described in other fields, and references to this figure are given under the most pertinent fieldname. Note that individual nucleotide and/or amino acid sequences may be retrieved for local analysis using the accession numbers listed under Database listings, 48-53.

Gene organization The majority of vertebrate Kvl genes are intronless in their proteincoding regions 48-20-01: The protein-coding regions of most Kvl subfamily genes are uninterrupted in the genome (Le. are intronless 19 ) according to presently recognized splice site t consensus sequences. Notably, the Kvl.7 gene possesses an intron t in the region of its sequence encoding the SI-S2 extracellular loop52. The lack of introns in coding sequences of the vertebrate Kv genes contrasts with Drosophila Shaker, where multiple exonst spanning > 120 kb can be alternatively spliced t to generate at least five functional transcripts 190,191. Different Drosophila Shaker channels inactivate over different time courses (see references in special section on Drosophila Shaker under Related sources and Reviews, 48-56). Most (but not all) members of the vertebrate Shaker, Shab, Shaw and Shal-related K+ channel gene subfamilies mostly yield single mRNA species (see field 17), each of which exhibits a distinct pattern of tissue-specific expression. Kv subunit genes are generally single copyt within the genome (see Southerns, 48-25). Related genes arise principally by mutation-selection t acting on duplicated genes (for proposed mechanisms, see Chromosomal location, 48-18). See also the 'head-to-tail' arrangements described for multiple Kvl subfamily genes in mammalian genomes (ibid.) which can appear in clusters (e.g. the KCNA6-KCNA1-KCNA5 cluster on human chromosome 12p189 (described in footnote to Table 8).

Summary of intron sequences found in Kvl subfamily non-coding regions 48-20-02: Kvl.l/Kvl.2/Kvl.3: The detection of 'long transcripts' (ca. 80009500nt) from the Kvl.l, Kvl.2 and Kvl.3 genes (see Transcript size, 48-17) together with several shorter ones with the same probe suggest that splicing of primary transcripts from these genes occurs in vivo. To date, the majority

II

II

mKv1. 1(MK1) rKv1.1 (RBK1) hKv1.1 (HUK( I» xKv1. 1()(Sha1) mKv1. 2(MK2) rKv1.2(RCK5) hKv1.2(HUK( IV» dKv1.2 bKv1. 2( BGK5) xKv1.2(XSha2) mKv1.3(MK3) rKv1.3(RGK5) hKv1.3(HPCN3) mKv1.4 rKv1.4(RCK4) hKv1.4( HPCN2) bKv1.4(BAK4) rKv1. 5(KV1) hKv1.5(HPCN1 ) mKv1.6(MK6) rKv1.6(RCK2) hKv1.6(HBK2) APLK Shaker

mKv1.1(MK1 ) rKv1.1CP.&K1) hKv1. 1(HUK( I» x~v1.1(XSha1 )

~~~ j~:~~~)

hKv1.2(HUK(lV» dKv1.2 bKv1.2(BGK5) xKv1.2(XSha2) mKv1.3(MK3) rKv1.3(RGKS) hKv1.3(HPCN3) mKv1.4 rKv1.4(RCK4) hKv1.4(HPCN2) bKv1.4(BAK4) rKv1.5(KV1 ) hKv1.5(HPCN1 ) mKv1.6(MK6) rKv1.6(RCK2) hKv1.6(HBK2) APLK Shaker

MTV

MSGEJA

oEAS~

-----V

----A----·----

APGHPQoGS

--·-A·-·------

I

YPRQA

DHDDHE

I

I

150

IA---M --T-VL--·--- • - HP -Q---AT-OPV -·-AAL-----· T -oPEAT-OPV ---AAl-·---- T -oPEAT-OP---AAl---··- T ·oPEAT--P· ---AAl····-- T -oPEAT-OP---AAl-·---- T -oPEAT-oLT -GSVGFA----- -oPEP VP-OHL lEPEA-GGGG-o-PQ-GCGSGGGGGG CDR-EPlPP AlPAAGEQo VP-OHl lEPEA-GGGG-O·PQ-GCVSG GG COR-EPlPP AlPAAGEQO VP-oHl ! LEPEV-OGGG- -PQ-GC GGGG CDR-EPYPP SlPAAGEQO - E- AMV- A- SSGCNSHMPYGYAAQARARERERlAHSR - AA· AAYAAATAAVEGTGGSGGGPHHH HQTRGAYSSHOp· GSRGSRRRRRQRTEKKK LHHRQSSFPHCSolMPSGSEEK I lRElSEEEED - EEEEEEEEEGR FYYSEEDHGD ·E·AMV·A-SSGCNSHMPYGYAAQARARERERlAHSR-AA-AAYAAATAAVEGTGGSGGGPHHHHQTRGAYSSHDP-GSRGSREEEATRTEKKKKLHHRQSSFPHCSOLMPSGSEEKIlRElSEEEED-EEEEEEEEEGRFYYSEEoHGD -E·AMv-X-SSGCASRMpTGTXXbARARERERLARSRRAA-RAYAAATAAVEGSGGSGGGSHHHHQSRGACTSHoP-SSRGSREEEATRSEKKKA HYRQSSFPHCSOlMPSGSEEKILRElSEEEED-EEEEEEEEEGRFYVSEooHGD - E- AMV- A- SSGCNSHMPYGYAAQARARERERlAHSR - AA - AAYAAAT AAVEGGGGSGGSQHHHHPSRGACTSHOP- SGRGSRRRRRPHPEKKK YHHRQSSFPHCSDlMPSGSEEK I lRolSEE· EE -OOEEDEEEEGR FYVSEEoHGE -EISlVPl- -GS AMTLRGGG EAG- -CYQTPRGECGCPPTSGLNNQSKETLlRGRTT LED-N QGGRPlPPMAQElPQPRRLSAEoEEGEGDPGlGTVEEDQAPQDAGS -E IAlVPl- -GG AMTVRGGD EAR-GCGQATGGELQCPPTAGLSDGPKEPAPKGRGA Q-o-DSGYRPLPPlPoPGYRPlPPlPEElPRPRRPPPEDEEEEGDPGlGTYE -Q AlGTAS RSEKSl TLAAPGEVRGPEGEQQDAGEF RSEKSL TLAAPGEYRGPEGEQQDAGEF RSEKSL TLAAPGEVRGPEGEQQDAGDF -E -AMAGI EGNGGPAGYRDSYHSS Q RPL LRSSNLPN SR S FPKLSEE-NANENGMGVPGS DYDCS - A AYAGL YGlGEoRQHRKKQQQQ QQHQKEQLEQKEEQKK I AERKlQLREQQlQR NSLDGYGS LPKlSSQ- EEGGAGHGFGGGPQHFEP I PHoHD

CC

ERJI NI

SGlRFE1QlKTlAQFP~TllGNPKKR.RYFDPlRNE~FFDRNIP~~D~Il1YVQSG~RlRRPYNvplDMFSEE

I

KF~ElGEEAMEK~REDEGF

I KE I EE

-------------------------------------------------------------------------------------------------------

RPlPE~

-. ----v· ----. - -. - -. --- -S-·· - T· - - ·1·· -A- - - - - - - -. - •• - -. - - - - - - -. -. - - - - - •••• - - - - - - - - - - - - - - - - - - - - - - •• - - - -Y- -

- - - -. -0-

--- - - - -- -. - - - - -. - - -. - - -E-· - -0-······ -- - - ••• -. -. - - - - - - - -. - - -. - - - - --- ••• - •• --I···· ·R- - - - - --. - -M- - - - - -Y- - -- .... - .•. -- ---- - - -.- -·E·- - -0--······-· ------- .... -- -- - --- _e. --.•• ----.- -··I·----R--- - --·-·-M-· ----Y-----·T--·-- -- ---- - - -- •. -E--- -0·-- -··F--·· -- -I--Y·- - -- -- - -. ---- -.- ·-----··---I··---R-·--·-·· --M------Y-- -

- - - - - --N -- ·---·N -- -·--·N

.-.------.-.- - -. - --- ---E---·D-·--···------- - --. -- .---- --.- -- .-- - --- •• -.- ·--1---· -R---- ----- -M- -- ---Y- --

300

- - • --·-N

---.- ••• ---- ---- -- - .-- -E----D-----·--···· -- - -- ------ -- - -. -- -- - - -- --- ------·I-----R-·-·· -----M-- ----Y--- -- ----·N - - - - - • - - - - • - - - - - - - •S- • - E- • - - 0 - - - - - - - • - - - - - - - - - •• - - - - - - - - - - F- - - - - - - - - - - - - - - - I - - - - - R- - - - - - - - - - I - - - - - - - - - - -. - - - - ON - • G - ••• - - - - - - - • - - - - - • - C- - - E- - • - 0 - - R- - • - - - - • - - - - - - - - •• - - • - - - - - - • - - - - - I • - • - - - - I - I - - - • - R- - Q- - - - - • - - - • - - - • -l R- - - - - - - RR - - G -. - - - - • - • - • - - • - - - • - C· - • E••• - 0 - - R- - - •• - - • - ••• - - - - - - • - - - • - - - - • - - - - - I - - - - - - - I - I - - •• - R- - Q- • - - - - - - - • - - - - ·lR- - - - - - - RR -·G -.------ ----- - - - ---Co. ·E-·· -D--R--··- - .y. -. --- - ---- - --.- - - -- --- --1- -- - ---1-1--- ·-R- -Q-- .---- - --- ----lR- -- - ---RR GCSYTDllPQEOGGGGG YSSVRYSD- - -. -- - ·Y-· - - - - - ·M- - - - - - ·E- - - -D-E- -TQ- - - - -. - - - - - - -. -. -. - - - - - - - - -- - - -K- -- - - - F-I-T- ·Y- - -Q-- - - -ll- - - - - - - -YR- - -O-A- --N GCSYTDllPQOOGGGGG YSSYRYSo- - -- - -- -Y-_· - - - - -M-· - -. --E-- - -D-E- -TQ- - - - -- - - -- -. - - - - - - - - - - - - - - -- - - -K- --.- -F-I-T- -V· - -Q-- - - -ll- - -- - - - -YR- - ·o-A---N ECSYTollPQOEGGGG YSSVRYSD- - --- - - -Y- - - -. - - -M-· - - - --E- - - -O-E- -TQ-- - - - - - - - ••• - - - - - - -.- - - - - - -- -. -K- -. -- -f-I-T- .y. - .Q-- - - -ll- - - - - - - -VR- - ·O-A-·-N ECSYTollAQODGGGGGGGSGGGGYSSVRYSD- - - - _•• ·Y- - - - - - - -M- - _. - - -E- -. -O-E- -TQ- - - - - - -. - - - - - - - - - - - - - - - -. - - - - --K- -. - - -F-I-T- ·Y- --Q-- - --ll- - -- - - - -VR- - ·O-A---N lHHQ - ·l-· -. - - - - - - - -G-· - - - - - -. - -D-A- -lH- - - - - - - -. -- - _. - - - - -G- - - - -. - - - -- - - -. --S- -V-Ao·-R- -Q- -0- - - -R- - - - - - -- --EK- --RN LHHQ - -H---- -- - - •• --G-Q----- ----D-A·-lP---· - ---.- -- -- - -- --G-· - - -- - -- ---- -- ·-S- -V-AD--R--Q--o- -. -R-G- -. - ----EK--YRN QEAE GGGG - - SS - -l· - - - - - • - - - - - - R- - Sl- ·0- - - - 0 - GR - Y- F- - • - - •• - - - - • - - - - - - - - - - - - - - - - - - - - - - - • - - - I -M - - - R- - Q. - • - -lAA - - - - - - ClP - GG - DE K- - - SQ QEAEGGGG • -SS- -l- - - - - - - -y. - - -R- -Sl- -D· - - -D-GR-Y- F- - - - - - - - - - - - - •• - - - - - - - - - - _. _. -. - - - - - - - -I-M-· -R- -Q- ·0- -lAA- - - - - ·ClP-GG-DEK- - -SQ PEAGGGGG - -SS- -l- - _. -- - - - -. - -R--Sl- -0-·· -D-GR-Y- F- -. - - - - - - - - - - - - - - - - - - - - - -. -. --. - -- - - - - ·I-l---R- -Q- -D· -lAA- - - -. -ClP-GG-DEK- --SQ - EK - • - QN C - - - - - - Y- - - - - •• - • - • - N- - - 0 • - - - - • Q- - N- - Y- - - - - • - - - • - - • - - - - - - - • F- - - - - • - • - • - - - - - - Y- - - • - - - - - - - - N- F- RY - - - - - - - - FC - - - - - - Y- - - - - - - - • R- - N· • - 0 - - - - 0 - AR -l- - - - - - - - - - - - - - S- - - - - - - - - - - - - - - • - - - - - - • - - - Y- - - - - - - - - - - DQ - I N- - - - - - - - - - - E• - - - ON

*

* *** **

*

** **** * *

* * ** * ***** ** ****** **. * .. * .* .*. ** **

****

*

Figure 4. Amino acid sequences of Shaker-related Kv1 subfamily isoforms together with that of Drosophila Shaker. Reported ORF lengths in the Kv1 subfamily include Kvl.l: MBK1MK1/RBK1/RCK1/RK1: 485 aa. Kvl.2: MK2/BK2: 499 aa; RCK5: 498 aa. Kvl.3: hPCN3: 524 aa. MK3: 530 aa. Isolate KV3/RGK5/RCK3: 525 aa. Kvl.4: hPCN2/HK1: 653 aa. RCK4: 655 aa. Kvl.5: hk2: 605 aa. hPCN1: 613 aa. Isolate/clone KV1/RK4: 602 aa. HCK1: 584 aa. Kvl.6: HBK2: 529 aa. Isolate/clone KV2: 530 aa. RCK2: 500 aa. (Alignment kindly provided by George Gutman, University of California at Irvine.) (From 48-19-01)

~

= ~ r l'

~

00

n«v1. 104K1) rKv1. 1(RBK1) hKvl.l (HUK( I» xKvl .1 (XShal ) n«vl.2(MK2) rKv1.2(RCK5) hKv1. 2( HUK(I V» dKvl.2 bKvl.2(BGK5) xKvl.2(XSha2) n«v1.3(MK3) rKv1.3(RGK5 ) hKvl.3(HPCN3) n«v1.4 rKv1. 4( RCK4) hKvl.4(HPCN2) bKvl.4(BAK4) rKv1. 5(KV1) hKvl.5(HPCNl ) n«vl.6(MK6) rKvl.6(RCK2) hKvl • 6( HBK2) APLK Shaker

n«vl.1(MK1) rKv1.1(RBK1 ) hKv1 • 1( HUK (I » xKvl.1 (XSha1) n«v1. 2(MK2) rKv1.2CRCKS) hKvl.2(HUK(lV» dKvl.2 bKvl.2(BGK5) xKvl.2CXSha2) n«vl.3(MK3) rKv1.3(RGKS) hKv1. 3( HPCN3) n«vl.4 rKvl.4(RCK4 ) hKvl.4(HPCN2) bKvl.4(BAK4) rKvl.S(KVl ) hKvl. 5(HPCN1) n«vl.6(MK6) rKvl.6(RCK2) hKvl •6( HBK2 ) APlK Shaker

EYQRQWLLIEYPESSGPA~vIA

E~

SI YI FClEfLpElK

DDKD

IFTG; I H; I;:;TVIY

lSNI

FTDP'FlVETlCII~~SFElWRFFA~PSKTDFFK~IANFIDIVA

450

N---

----V----~---

------Fl--- -- -Fl-- ----Fl--

----1---------

...** .** .* .** .** *. ** .** ** *******

i I ~~F

ITlGTE IIEQ E

-

-

PKC

I

~S4~

~S5------..

. .*** .. ** *. **. **. * I

~P~

*

*

** **

~S6-600

GNQ KGEQATSlAllRVIRlVRVFRI FKlSRHSItGlQI lGQflhlMlElGlll FFlF IGVllFSSAYTfAEAEEAESHfSSIPDAFWAWSMTTVGYGOMYPVT IGGKIVGSlCAIAGVl T

--- -------._-_._._-----------._----------------._----------------------------------------------------------------.-------.

-----------M--- -

-----------l--KP-----------L--KP-----------l--KP-----------l--KP-----T-----l--KP-----------l--KT-----------L--RQG -----------l--RQG -----------l--RQG -l- - - - - - - -Dl-Q-QG -l- - - - - - - -Dl-Q-QG -l - - - - - - - - Dl- Q- QG -l - - - - - - - -0 l- Q- QG

-F- --- ---- -l--F - - - - - - - - -l--F - - - - - - - - -lVQRH- f - - - - - - - - -l VQRH- F- - - - - - - - -l VQ - QQQPASG - GGQN - Q- - 114 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - K- - Q- - - - - - - - - - - - - • - - - - - - - - - - - - - - - DODD -l- P- - - - - - - - - - - T- - - - - - - - - - - M- Y- - - - - - - - - - - - - - - QSKSNN- -114- -- - - ---- ---- - - - - - - -- ---- - -- - -- --- ----- -------- -- - -- - - -- - - - -- - -DADQT- -K-- - - ---- - - -T--- - -- - --R- IGW--l- - ----- ----- - --- --- --VV-O - - -- - ---A-VV- -EEDTlNlPKAPVSPQOKSSN- -114- - - - -- - -- --- - - - - - -- - - -- -- - - - - -R- -- --- - - - - - - - - - - - - - -V- - - -- - - - - --GSEN-F-K- - - -- -- - - --T-- - - -- ---T--GF'I---- - -- -W- - ---

* .*** .* .* .

.

* .* **. ** .************* .****** * *****************. ***** * *

g t"'t

-f- -- ---- --f-Q- ---- -1- --I- - --- ------ - - - - ---- - -ERI - SR-WN~-fK- -S --- -- -Y- --- - - --- -- - -- ------ ----E- - -- -- ---- ---f- -- - - --- - - - -- - --- -1- -- - -- ---- ----S-- - - ---IfR -ENE DMHGGGV-F-TYs-=FIG-Q Q-TS --- -- - - - --- -- - - -- --Fl-- ---- ---AG--T-- - -1- - ---F- --- - -- -- - - -- --- --1- - - - ----------S- - - - - - - IfR -ENE DMHGGGV-f-TYS~IG-Q Q-TS -- - - - - - - --- - - - --- - -fL- - ---- ---SC--T-- -- I ----T -F- -- - - -- -- - - -- - ----I--------------S- - - - - --lfR -ENE DMHGSGY-F-TYS~IG-Q Q-TS -- - - - - - - - -- - - - --- - -Fl- -- - -- ---AG--T- ---1-- ---F- - - - - - - -- -- - ------I--------------S- -- - -- -I FR -ENE DMHGGGY-F-TYS~IG-Q Q-TS -- - -- -- - - -- - ---AG--T----I-- --_F- - -- - -- - - - - - ------S- -- - - - -IFR -ENE DMHGSGY-F-TYS~IG-Q Q-TS -- - - - -- - - -- - - - ------AG--T- - -- I -- --_f-K-- - -- -- - - - -- -- - - I --- I --T-- ---- -S- -- -- --VFR -ENE DMHGSG-NYYSYP~-RFQ K--T -- - - - - -- - -- - - - -- - -l-- -- -AV- -T-l--I-- --0f- - -- - -- - - - -- -- - - - -G--- - - -L-- -- -- -- - - - - - ---FR -E- -YPASPSQOVFEMN~SGAP SGASS-S- - - -Y- - - - - - - - - - - - -l- - - --- - - -AT-SR- - --l- - --0f- - -- - - - - - - - - --R- --G- - - - - -l-- -- ----- - - - - - --FR -E- -YPASPSQDVFEM~SGAS SGASS-S- -- -V- - -- - - - - - - - - -L- - - --- - --AT-SR-- - -l-- --Df-- -- - - - - -- -- - --- - -G- - ----L---- -- -------- - -FR -E- -YPASTSQOSFEM~SGSR AGASS-S-- - -V- - -- - - - -- - - - -l- - - - -- - --AT-SR- -- -l-- ---FKK- I - - - - -- - - --S---G- -----l--- --- - -- - -- - - - -FR - -R-llMAlSA-GHS-ll-=tr='SAPHlENSGHTI -N- - - - - - - -V- -V- -- - -F- - -C-- -1-QAl- - --- - - I - --OS -FKK- I - - -- - - - ---S---GR- ----l---- --- - - --- - - - -FR - -R-llMAlSA-GHS-ll:U::SAPHLENSGHTI -N-- - -- - - -V- -V--- - -F- - -C-- - --QAl- - -- - - - I - ---S -FKK- I - - -- -- - - - -0- - -G- - - - - -l- - - - --- - - - - - - - - -FR - -R-lVMAlSA-GHGGll:U::SAPHlENSGHTI-N- - - - - - - -V- -V- - - - - F- - -C- - - - -QAl- - - - - - - I - - --S -FKK- I --- - - - -- - -S---G-- - ---l- ----- --- ----- --fR --R-LIMAlS- -GHGGll:U::SAPHPENSGHTI-N- - - -- - - -V- -V- - - - -F-- -C- - - --QAl- - -- ---1- - --S -F- - - - - -1- - - - - -- -S- -A- - - - - -l- - -- --1 T- - - - - - - -FR DEREllRHPPVPPQPPAP APGINGSVSGmSGPTVAPlLPRTlA- - - - - - - -T-V- - -T- - -l- - - - - - - - -AE-SR- - - - I - -Y-_f- - - - - -1- - - - - - - -5- -A-- - - - -l- - -- - - IT- - - - - - - -FR DEREllRHPPAPHQPPAP APGAflC'SGVMAPPSGPIVAPllPRTlA- - - - - - - -T-V- - -T- - -l- - - - - - - --AG-SR- - - -1- -V-PF- - - - - - - - - - - - - - - - - -G- - - - - -l- - -- - - - - - - - -- - -QFRADGRGGSNEGSGTRlSPASR SHEEE-EDEDSYAFPGST"PSGGlG-GGTSSlSTlGGSF- - - - - -l- - - - - -V- -T- - -l- - -S- - - - -AA- -R- - - -1- -l-PF- - - - - - - - - - - - - - - - - -G-- - - - -l- - - - - - - - - - - - - - -QFRADGRGGSNEGSGIRMSPASRGSHEEE-EDEDSYAFPGSIPSGGlG-GGTSSFSTlGGSF- - - - - -l- - - - - -Y- -T- - -l- - -S- - -- -M- -R- - - - I - -l-PF- - - - - - - - - - - - - - - - - -G- - - - - -l- - - - - - -- - - - - - - -QFRVDGRGGNN GGYSRVSPVSRGSQEEEE-E-DSYTFHH-l TPGEMG-GGSSSlSTlGGSf- - - - - -l- - - - - -V- -T- - -l- - -S- - - - -PA- -R- - - -1- -l--F- -R- - - -- -- - - - -M- -lC- -F- -VI --l- ---- - - - - - -QFHYRWNSTANDSKESIEEO- IPK-NE- - - - I - -C- - - - -T- - -l- - -AS- -f-lG- - - - - - -C- - - --K--K----- -- - - - -QA- - -V- -I--F---l- - - - - - --- - - -FHYKVFmUlCTK IEEDEVPDI---- -L1- -- - - -- -T- - -T- --l-- -N-lN-CROV- -V-- - I-

Figure 4 Continued

II

I VSVMVI

. * *** ****** .********* *

* ** ** ****

~ ~

00

II

mlCv1. 1(MK1) rKv1.1(RBK1 ) hKv1.1(HUK(I » xKv1.1 (XSha1) mlCv1.2(MK2) rKv1.2(RCK5) hKv1.2(HUK(lV» dKv1.2 bKv1.2(BGK5) xKv1.2(XSha2) n«v1.3(MK3) rKv1.3(RGK5) hKv1.3(HPCN3) n«v1.4 rKv1.4(RCK4) hKv1.4(HPCN2) bKv1.4(BAK4) rKv1.5(KV1 ) hKv1.5(HPCN1 ) n«v1.6(MK6) rKv1.6(RCK2) hKv1.6(HBK2) APlK Shaker

• FNYFYHRElEGEE QAQllH VSS PNllsDSD

lSR

RSS~T1S

KSE~

IME I EEDMNN~

I AHYRQAN I

~TGN

I CTtA DQ --AT --

I NCV~KSKLL

748 TDV

---- N--- ---------- -----M-Y-----------------V----A--1- N- - - -. - - - N- - - - - - - - AM ---- - - - - -l- - - - ON F- E- - - - - - --L- NT -Y--IT-M----KKS- -A- - - - -0- - -Q-GV- - -NEDF-EE-lK-A- - -Y-Q -T-C-KIP-SP--l- NT -Y--IT-M---- - -Y-Q -T-C-KIP-SP-KKS- -A- - - - -0- - -Q-GV- - -NEDF-EE-lK-A--l- NT -Y--IT-M------Y-Q -T-C-KIP-SP-KKS--A-----0---Q-GV---NEDF-EE-lK-A--L- NT -Y--IT-M------Y-Q -T-C-KIP-SP-KKS--A-----0---Q-GV---NEDF-EE-lK-A---Y-Q -T-C-KIP-SP-KKS--A-----0---Q-GV---NEDF-EE- - -Y-Q -T-C-KIP-SP-QKS- -A- -l- -0- - -Q-GV-H-NEDF-EK-lK-A- -lG NT -Y- - IT -M- - -- - -YM- -G-CQH-S-SAE E-RKA- -N- -l----V- - -GGM- QSAFP- TPFK- - -STAT- - -N NNPNS- - - IK- IF - -- - -YM- -G-CQH-S-SAE E-RKA- -N- -l----V- - -GGM- HSAFP-TPFK- - -STAT - - -N NNPNS- - - IK- I F- --S-YM- -G-CQH-S-SAE E-RKA--N--l----V---GGM- HTAFP-TPFK---STAT---N NNPNS---IK-IF--- - - - - - - - - - - - - - - - - - - - -N- - -T- -TQNAV-C-Y-P-Nl -KKF- - - TSS-LGO- - - l-M- -GVKE -LCGKEEKCQGK-DD ET DKN- -S-AKAVE - -- - - - - - - - - - - - - - - - - - - - -N- - -T- -TQNAV-C-Y-P-Nl -KKF- - - TSS-lGO- - - l-M- -GVKE-LCGKEEKCQGK-DDSET DKN- -S-AKAVE- -- - - - - - - - - - - - - - - - - - - - -N- - -T- -TQNAV-C-Y-P-Nl -KKF- - - TSS-LGD- - - L -M- -GVKE -lCAKEEKCQAK-DDSET DKN- -S-AKAVE- -- - - - - - - - - - - - - - - - - - - - -N- - -T- -TQNAV-C-Y-P-Nl -KKF - - -TSS-LGO- - - L -M- -GVKE -LCAKE KCQGK-DDSET DKN-VS-AKAVE- -- - - - - - - - - - - - - - - - - - - -DH- - - -A-KEEQGNQRRESGlDT GGQ-KY-C-KA- FCKTGGS l-SSDSIRRGSCPLEKCHlKAKSNVDl RR Sl YAl-LDT-R E--l --------------------DH-- P-V-KEEQGTQSQGPGLDR GYQ-KY-G-RG-FCKAGGT l-NADSARRGSCPLEKC-VKAKSNVDl RR SLYAl-lDT-R E--l ---------------------Q-- -G-YT- VTCGQPTP D-K ATONGlG -PDF A-ASRERRPS-lPTPH-AYAE KRM--E---------------------Q-- -G-YT- VTCGQPTP D-K ATDNGlG -PDF A-ASRERRSS-lPTPH-AYAE KRM--E---------------------Q-- -G-YT- VTCGQP-P D-R ATDNGLG -PDF P-ANRERRPS-lPTPH-AYAE KRM--E- - - - - - - - - - - - - - - - - - -G-STDKG QYKH-Q-C- -YPE KKDSlDSECGSOIMEMEEGNH- T PLTEKVKE-HAIKANNPGSDYGL E- -- - - - - - - - - - - - - - - - - - -ADR- -M-S-NFNH-T-CSY-PGAlGQHlKKSSL SESSSDIMD-DDGIOAT-PGl TDHTGRHMVPFLRTQQSFEKQQlQlQLQLQQQSQSPHGQQMTQQQQl-QNGlRS-NSlQlRH-NAMAVSIE- --

* .****** .*** .******

Figure 4 Continued

(D

::s

c-t-

~ ~

00

l_e_n_t_ry_48

_

of these events have been related to the presence of introns in 5' untranslated (non-translated) regions t (designated as 5'-UTRs or 5'_NTRs)18,19 as described in Table 9. Alternative splicing t in untranslated regions could conceivably play a role in controlling tissue specificity or quantitative expression levels of Kv proteins. See also Kv3.1 (isolate KV4) and Kv3.3 under VLC K Kv3Shaw, entry 50.

Multiple polyadenylation signals in the Kvl.4 gene

48-20-07: Kv1.4: KCNA4 contains three conserved polyadenylation t signals in the 3' UTR. Although the 'precise' employment of these signals is unclear, their differential utilization may partly account for the different transcript sizes (2.4kb, 3.5kb, 4.5kb) detected in Northern t assays using Kv1.4 probes 175 (see also Transcript size, 48-17 and predicted relationship of transcript size to stability in paragraph 48-20-08).

Possible role of 3' UTR motifs in control of channel mRNA stability 48-20-08: Kv1.4: The 3' untranslated region of KCNA4 and the rat, bovine and human KCNA4/Kv1.4 cDNAs contain multiple ATTTA and ATTTG motifs 175 (e.g in rKv1.4 (RCK4) 5'-ATTTAATAG ATATAGGTCACAATTTAATCTTGGATTTA ATTAAA-3'). These sequences are reminiscent of conserved AU sequences from the 3' untranslated region of (cytokine) GMCSF mRNA that have been shown to 'destabilize' cytokine messages (via selective mRNA degradation) when present in 3' UTRs 193 . By analogy, message stability of K+ channels would be increased when the number of these repeats are minimized. In the case of the KCNA4 gene (see Table 9 and Transcript size, 48-17) the shortest detected messages (2.4 kb in length) are likely to contain no more than 400 nt of UTR (i.e. assuming the ",2 kb segment comprising the 654 aa coding region is intact). This implies that the shortest (2.4 kb) transcript would have the highest stability, because the ATTTA motifs would be excluded from the 2.4 kb transcript 175. In Xenopus oocytes, mKv1.4 equivalent amounts of 3.5 kb transcripts produce ",4 to 5fold larger currents than 4.5 kb mRNAs 192. Note: tUA-rich sequences' have also been shown to induce ttranslational blockade' independently from these effects on message stabili tY 94.

Comparative note: Alternative splicing in the Drosophila Shaker gene 48-20-09: For initial descriptions of alternative mRNA splicing producing K+ channel diversity from the Drosophila Shaker gene see refs 190,191,196-198. An immunological characterization of specific K+ channel components produced by alternative splicing from the Shaker locus has confirmed that different subtypes of A-channels are formed in different tissues 199: On immunoblots, the protein products from the locus are in the size range 65000-85000 Da with no smaller products being evident (compare Protein molecular weight (purified), 49-22). In agreement with in situ hybridization studies, immunocytochemical techniques reveal a 'non-uniform' distribution of Shaker products in the brain of the adult fly199.

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_nt_ry_4_8_

Table 9. Predicted introns in Kv1 subfamily non-coding regions (From 48-20-02) Kvl.l

Kvl.2

Kvl.4

Kvl.5

48-20-03: mKvl.l/MBKI genomic: Alternative splicing of a short intron in the 5' UTR of mKvl.l can give rise to both spliced and unspliced polyadenylated RNA 18 . The 5' intron (position -886 to -514 relative to the start codon of the protein product shown in Encoding, 48-19) includes three potential initiation codons conforming to the consensus sequence PNNATG (where P is A or G and N is any nucleotide). 48-20-04: mKvl.2/MK2 genomic: Comparison of mouse Kvl.2 genomic sequences with rat Kvl.2 cDNA sequences (see Methodological notes, this field) reveals (at least) two introns in the 5' UTR. According to the analysis of Chandy and Gutman (1994, see Related sources and reviews, 48-56), an intron closest to the 'authentic' translational start site (positions -607 to -136) contains six possible translational start sites producing ORFs ranging from 18 to 156 bp (two of which meet the preferred Kozak initiation sequence PNNATG, see above). Only the splice acceptor site t of the other (upstream) intron has been detected at position -920. 48-20-05: mKvl.4/genomic: The Kvl.4 basal promoter is GC-rich, contains three SPI repeats (CCGCCC, -65 to -35), lacks canonical TATAAA and GGCAATCT motifs, and has no apparent tissue specificity. One region enhances activity of this promoter192. Analysis of the mKvl.4 genomic sequence (KCNA4) has revealed a single large intron ("-I3.4kb) in its 5' UTR 175. A single exon contains the remaining "-10.8 kb of the 5' UTR, a large open reading frame t ("-12.0 kb encoding the 654aa mKvl.4 protein), and all of the known 3' UTR ("-11.1 kb). The 5' UTR contains eight sequences conforming to PNNATG (see above). The longest detectable transcript from this gene (4.5 kb, see Transcript size, 48-17) can be accounted for by the sequence of the genomic DNA, although the precise ends of this transcript have not been defined. Note: The shortest detectable Kv1.4 transcript (2.4 kb, ibid.) is predicted to contain (at most) 400bp of non-coding region and lacks any 3' ATTTA motifs (see paragraph 48-20-08). Furthermore, the majority of the 5' ATG motifs (ibid.) would not be present in a transcript of this size, and in consequence, the 2.4 kb transcript may be ex~ected to have both increased stability and translational efficiency l 5. 48-20-06: rKvl.5: The rKvl.5 gene lacks a canonical t TATA boxt , has several transcription start sites t, and the 5' non-coding sequence is intronless 111 . For details of the cAMP-response element (CRE) identified in the 5' UTR of rKv1.5, see Developmental regulation, 48-11.

Notes: 1. The presence of 'upstream' initiating codons (ATGs) or in 5' UTR sequences (see, for example, MBKI and mKv1.4 in Table) may inhibit translation of K+ channel and other cDNAs in heterologous cell expression systems (probably by interference with ribosome'scanning' processes for the'authentic' start site, the competing initiation sites 'delaying' authentic codon binding). 2. Comprehensive analysis of alternative splicing events generally requires direct comparison (alignment) of genomic t (chromosomal) sequences and all cDNA t sequences (for further background,. see Gene family and Gene organization under ILG K Ca, 27-05 and 27-20 respectively).

l_e_n_t_ry_4_8

_

Homologous isoforms Individual Kvl subfamily sequences are generally highly conserved between species 48-21-01: For a full listing of cDNA isolate names (i.e. clone names crossreferenced to species of origin) falling into the Kvl subfamily, see Cene family, 48-05. For further background to Kv channel interspecies identities and Kv gene evolution, see Domain conservation, 48-28 and Chromosomal location, 48-18.

Protein molecular weight (purified) Kv subunit M r values following in vitro versus in vivo translation 48-22-01: rKvl.4 j hKvl.3: Specific polyclonal antibodies raised to a synthetic peptide sequence (residues 13-37) from the N-terminal of rat Kvl.4 (antibody Kvl.4N) detects a full-length monomeric polypeptide of ",88000 Da when Kv 1.4 mRNA is translated in vitro and subjected to SDSPAGEt analysis 64 . On immunoblots of native rat brain membranes, the Kv 1.4N antibody precipitates a diffuse band of ",95000 Da. The 7 kDa difference between in vitro and in vivo translation products is probably due to additional post-translatory glycosylationt in the latter system (see Protein molecular weight (calc.), 48-23 and Sequence motifs, 48-24). Similarly, the apparent molecular mass of the immunoprecipitated type n channel from Jurkat cells (~ hKv 1.3) is approximately 65 000 Da (significantly greater than that of the 58 000 Da in vitro translated product, and suggestive of post-translational modification events)200. Specific immunoprecipitation in both native and cell-free preparations can be blocked by an excess of the Kvl.4 peptide immunogens but not by unrelated competitor pepides64 . For details of molecular weight variants of Kv1.1 during early biosynthesis and subunit assembly, see Protein interactions, 48-31. A 'very faint' band is detected at 97 kDa in cardiac atrial and ventricular preparations when an anti-Kvl.4 antibody (that reveals intense Kvl.4 expression in brain) is used at a 5- to la-fold higher concentration in Westernt blots l14 (see also Protein distribution, 48-15). Mrvalues for affinity-purified Kv channel complexes from native cells 48-22-02: Native Kv channel assemblies (e.g. those immunoprecipitated with high-affinity toxins) are octameric (04.84) complexes as first established by Oliver Dolly and co-workers (for details, see VLC K Kv-beta, entry 47). Protein assemblies such as the DMB K+ channel sialated glycoproteint complex (described under Blockers, 48-43) has a total molecular weight ",450 kDa and possesses toxin-binding subunits between 65 and 95 kDa. Rat brain DMB protein is composed of polypeptides of 80, 42 and 38 kDa, whereas in bovine brain peptides of 74, 42 and 38 kDa are found 201,202. The larger 80/74 kDa peptides generally form the toxin-binding subunits and the smaller subunits do not appear to be glycosylated (for significance, see Protein molecular weight (purified) under VLC K Kv-beta, 47-22). Comparative notes: 1. Kvl.l/RCKI expression in oocytes can be detected by immunoprecipitation as a 57 kDa polypeptide by SDS-PAGEt (representing

II

_L..-

.

e_n_try_4_8_

Table 10. Calculated molecular weights (From 48-23-01) Kvl.l

MBK1: 56.4kDa. RBK1: 54.6kDa. RCK1: 56.4kDa (56379). RK1: 54.6kDa

Kvl.2

BK2: 56.7 kDa (56706). RCK5: 56.8 kDa (56 768).

Kvl.3

Isolate KV3: rv58.4 kDa (rv 55 kDa when translated in vitro). RCK3: 58.4 kDa (58397) (see domain structure model [PDTMj, Fig. 6). HK1: 73.2kDa (73211). RCK~: 73.4kDa (73398)

Kvl.4 Kvl.5

HK2: 66.6kDa (66640). hPCN1: 67.1 kDa (67097). Isolate KV1: rv66.6 kDa (rv 67 kDa when translated in vitro). RK4: rv66.6 kDa. HCK1: rv64.3 kDa

Kvl.6

HBK2: 58.9 kDa (58891). Isolate KV2: rv58.8 kDa (rv 56 kDa when translated in vitro)

functional channel subunit monomers that are N-glycosylated)203. 2. Kv1.3 purified from African green monkey kidney cells (CV-1) using a vaccinia virus/T7 hybrid expression system has an estimated size of approx. 64 kDa by SDS-PAGE204 . By sucrose gradient sedimentation, purified Kv1.3 is approx. 270 kDa, consistent with it being a homotetrameric complex of 64 kDa subunits under these conditions 204 . Negative-staining electron microscopy reveals that Kv1.3 protein can form small crystalline domains consisting of tetramers (approx. dimensions 65 x 65 A) with the centre of each tetramer comprising a stained depression possibly representing the ion conduction pathway204.

Comparative note: Drosophila Shaker 48-22-03: Recombinant baculovirus t constructs expressing the Drosophila Shaker H4 cDNA induce a band migrating at approx. 75 kDa on SDS-PAGE gels stained with Coomassie blue205 . As indicated by the lack of requirement for radiolabelling, the 75 kDa species (confirmed to be Shaker protein by Western blotting l ) accounts for a 'substantial fraction' of the membrane protein in Shaker-infected Sf9 cells205 .

Protein molecular weight (calc.) 48-23-01: Voltage-gated K+ channel of mammalian brain are oligomers of glycosylated polypeptides of typical predicted molecular weight of 6595 kDa (monomeric). The molecular weight values quoted in Table 10 are generally derived from sum totals of constituent amino acids (i.e. from predicted amino acid sequences derived from cDNA sequences).

Sequence motifs 48-24-01: Entry cross-references to predicted 'functions' for a number of common sequence motifs described in the literature are listed in Table 11.

II

---J_

l_e_n_t_ry_48

Table 11. Cross-references for sequence motifs described in the Kv channel family (From 48-24-01) Motif, site or function

Described under

{E/L)TDV C-terminal motif

This field (Sequence motifs) and Fig. 7 under Protein interactions, 48-31

H5 (pore) motif (K+ channel signature sequence)

In Kv channels, see Selectivity, field 40 of Kventries. Variation in CNG channels, see entries ILG (CAT) cAMP, entry 21 and ILG CAT cGMP, entry 22. In calcium-activated K+ channels, see ILG K Ca, entry 27. Variation in eag-related channels, see VLG K eag/elk/erg, entry 46. Absence in minK channels, see VLK (K) minK, entry 54

HMMEEM~ALS motif (N-terminals of Kv2, Kv3, Kv4, not Kvl)

This field (Sequence motifs)

Intron/exon splice site motifs

Gene organization, field 20

IYLESCCQARY motif (N-terminals of Kv2, Kv3, Kv4, not Kvl)


N-Glycosylation motifs


NAB Kv assembly motifs

Protein interactions, field 31

O-Glycosylation motif (Kvl.6)


Phosphorylation motifs

Protein phosphorylation, 48-32

PNNATG motif (nucleotide)

Gene organization, 48-20

Proline-rich motif

Interacts with SH3 domains of Src (see Protein phosphorylation, 48-32)

Repetitive sequence motifs


Serine/cysteine motif in ball domain of A-type channels

Table 19 under Inactivation, 48-37; see also sequence alignments under VLG K Kv-beta, entry 47

Signal sequence motifs (absent)


Tl assembly motifs (domains A and B)

Protein interactions, field 31

II

_L-

e_n_try_4_8_

Table 11. Continued

Motif, site or function

Described under

Voltage-sensor motifs including: leucine heptad (zipper) motif conserved positively charged residues conserved negatively charged residues

Voltage sensitivity, field 42

YFFDR N-terminal motif

Protein interactions, 48-31

Other specific sites determined by structure/ function analysis

e.g. Other structural determinants contributing to Kv channel assembly, block, expression, inactivation, selectivity, voltage sensing and modulation; see crossreferences in Table 14 under Domain functions, 48-29

A number of other 'motifs' are described within the Kv1 subfamily sequence alignments in Fig. 4 under Encoding, 48-19, cross-references in Table 14 under Domain functions, 48-29 and listings in Resource G - Consensus sites and motifs. Phosphorylation motifs are described under Protein phosphorylation, 48-32.

C-terminal deletion/mutation including the conserved (E/L)TDV motif 48-24-02: hKv1.5: The C-terminal regions of Kv1 subfamily K+ channels show little conservation between isoforms (see Domain conservation, 48-28). An exception to this is the last four C-terminal residues {E/L)TDV, which are well-conserved from Drosophila to human. Deletions of 4, 16, and 57 Cterminal residues of hKv1.5 do not affect whole-cell current amplitude, midpoint of activation, degree of inactivation, or activation kinetics compared to wild-type hKvl.5 following heterologous expression in mouse L-cells 206 . Deletion experiments removing C-terminal residues of rKv1.1 channels have shown similar results 206 . Mutation of the C-terminal sequence (-ETDV) of Kv1.4 abolishes its binding to, and prevents its clustering with the membrane-associated (putative) guanylate kinases SAP97 and PSD-95 62 (see Subcellular locations, 48-16 and Fig. 7 under Protein interactions, 48-31).

Repetitive sequences in Kv subunits 48-24-03: Figure 4 (under Encoding, 48-19) shows several simple alanine (A), glutamine (Q), glycine (G), histidine (H) and lysine (K) repeats in several Kv protein-coding regions. This feature is most marked in the N-terminal of the Kv1.4 subfamily proteins which show rapid inactivation properties. In this

II

l_e_n_t_ry_4_8

_

Table 12. Sequence motifs specifically reported in Kvl subfamily proteins (see also Protein phosphorylation, 48-32) (From 48-24-04) Kvl.l

MBK1: In common with all known Kv channels, no N-terminal signal peptide sequence is apparent. RBK1: N-gly: Five potential Nglycosylation sites (Asn-X-Thr-Ser); glycosylation in vivo could potentially increase the Mr of RBK1 protein to values close to affinity-purified dendrotoxin receptor ('"'-'76kDa). RCK1: N-gly: A single N-glycosylation site at Asn207.

Kvl.2

RCK5: N-gly: N-glycosylation sites at Asn38, 207, 465, 480 and 490 conforming to NXT/S. Note: Most of these sites are at analogous positions in the RCK series proteins. BK2: N-gly: Five possible Nglycosylation sites t 15. t

Kvl.3

RCK3: N-gly: Two putative N-glycosylation sites at Asn59 and 229 (see [PDTM], Fig. 6). Isolate KV3: N-gly: Six potential Nglycosylation sites (one in the N-terminal region, two in the Sl-S2 loop, and three within the C-terminal domain32 .

Kvl.4

HK1: N-gly: Asn-X-Ser/Thr N-glycosylation t motif at aa 181 and 352 (Note: only the 352 position is extracellular). hPCN2: N-gly: See also Domain conservation, 48-28. RCK4: N-gly: Three putative Nglycosylation sites at Asn183, 354 and 644.

Kvl.S

HK2: N-gly: Putative N-glycosylation site ataa 124. hPCN1: N-gly: A single N-glycosylation motif conforming to Asn-Xaa-(Ser or Thr) at aa 121. A putative leucine zipper t sequence is conserved in the S4H4 region. hPCNl shows the Shaker-type (Arg or Lys-Xaa-Xaa) motif repeated seven times in the S4 transmembrane domain. Isolate KV1: Potential leucine zipper t sequence found between the S4 and S5 domains (Leu402-Leu430). N-gly: There are five potential sites for N-glycosylation (N-X-5/T); four are located in the N-termina1 region, and one (Asn290) is within the 51-S2 loop. HCK1: N-gly: Asn125 and 190.

Kvl.6

HBK2: This and other Kv1.6 isolates do not have an N-glycosylation site in the loop between the Sl and S2 transmembrane domains (as seen in the 10-30aa long SI-S21oop of other cloned K+ channels). The single N-glycosylation motif at Asn46 is intracellular, and therefore not N-glycosylated. Notably, the Sl-S2 bend is unusually long ('"'-'70 aa) and is rich in glycine/serine residues, reflecting a possible O-glycosylation t. Isolate KV2: Potential leucine zipper t (Leu356 to Leu384); N-gly: one (non-functional) N-glycosylation site in the N-terminal domain32 .

case, the repeats begin just after a stretch of positive charges thought to form part of the inactivating tbal}, domain, as underlined in Fig. 4 (ibid.). Note: Such simple repeated sequences may provide a 'spacer' function within channel tertiary structures and/or playa role in channel assembly.

II

_'--

e_n_t_ry_4_S______

Reported N-glycosylation motifs Kvl-Kv4 subfamilies 48-24-04: With the exception of mKvl.6 and hKvl.6, an extracellular N-glycosylation motif is present between the 51 and 52 transmembrane domains in all Kvl subfamily sequences (as indicated by asterisks in Fig. 4 under Encoding, 48-19; actual motif sequences are underlined). This conservation is notable considering the relative divergence of the 51-52 loop regional sequences compared to the remainder of the 'hydrophobic core' domains. Further reported sites are indicated in Table 12. Note, however, that motifs located intracellularly or in transmembrane domains are not utilized (see Resource G - Consensus sites and motifs). Assembly and maturation of Shaker K+ channels 48-24-05: Voltage-dependent K+ channel a subunits generally assemble to form tetrameric membrane protein complexes. Glycosylation occurs at two positions in the 51-52 loop: Shaker protein is made as a partially glycosylated precursor (immature form) that is converted to a fully glycosylated product207. Further studies208 have indicated that the immature protein is coreglycosylated in the endoplasmic reticulum (ER) whereas the mature protein is further modified in the Golgi apparatus. Shaker channels appear to assemble on the ER, while maturation can be inhibited by blockers of ERto-Golgi transport by (i) treatment with brefeldin A or (ii) incubation at Is o C208 . Glycosylation of Shaker channels in oocytes occurs in two stages, generating an immature and a mature form of Shaker protein. Notably, however, glycosylation does not appear essential either for assembly of functional channels or for their transport to the cell surface 209 .

Southerns 48-25-01: Generally, Kvl-subfamily genes are single-copy in the genome (see Gene organization, 48-25).

STRUCTURE AND FUNCTIONS Because of the large volume of structure/function data available for Kv channels, general similarities derived fronl multiple sequence alignments of all Kv family proteins are tabulated under Domain conservation, 48-28 (Table 13) and Domain functions, 48-29 (Table 14). Further information on each subtopic is then cross-referenced to relevant entry/field name loci as part of these tabulations. In-press update: The study of K+ channel structure-function relationships has been considerably advanced by the availability of the Streptomyces lividans crystal structure494 .

Amino acid composition Composition of K+ channel protein domains based on hydropathicity analyses 48-26-01: In general terms, protein subunits encoded by all Kv gene subfamilies can be divided into three broad regions of approximately equal

II

'---e_n_t_ry_4_8

_

size with regard to amino acid composition: The 'N-terminal third' [PDTM] and 'e-terminal third' [PDTM] are hydrophilic, whereas the 'middle third' (the core region, [PDTM]) incorporates six stretches of hydrophobic t residues forming a-helical transmembrane domains (SI-S6, [PDTM])210 plus an H5 sequence t (also designated as the P-region or pore-lining domain) between SS and S6 segments (for 'mapping' of 'functions' to K+ channel subunit domains, see the [PDTM), Fig. 6, Domain arrangement, 48-27, Domain conservation, 48-28 and Domain functions, 48-29).

Domain arrangement Continual refinement of K+ channel structural models 48-27-01: Prior to the availability of high-resolution K+ channel molecular

structure information494, modelling approaches reliant on interpretation of structure-function t data (e.g. 'iterative' testing of structural predictions by means of site-directed mutagenesis t followed by functional analysis of mutant proteins) had been a dominant approach. These models are subject to continual refinement; hence, where available on internet resources, contemporary models of K+ channel protein structure or related datasets will be cross-referenced from entry update pages via the CSN (www.le.ac.uk/csn/). The impact of cysteine scanning mutagenesis (note 1) and toxin-channel interaction site mapping (note 2) are reviewed in ref. 211 . These experimental approaches have been the key for proposing that ion conduction pathways in K+ channels are short, and that permeating ions appear to traverse channel proteins via short canals ('canaliculi') of 10 A or less (note 3). Notes: 1. Cysteine scanning mutagenesis employs substitution of a natural channel amino acid residue by cysteine to make it reactive with sulphydryl-specific probes. 2. Toxin-channel interaction site mapping employs peptide neurotoxins of known three-dimensional structure to map contours of contact sites in channel pores. 3. 'Canaliculi' are implied from identification of channel residues in pore regions that are a few residues apart and yet are on opposite sides of the membrane; similar structures have been proposed for the movement of gating charges in voltage-sensing domains (as reviewed in ref. 211).

Oxidative induction of intersubunit disulphide bonds 48-27-02: Several voltage-activated K+ channel primary sequences contain

two conserved cysteine residues in putative transmembrane segments S2 and S6. A proposal that these cysteines form an intrasubunit disulphide bond210 was originally tested207 using site-directed mutagenesis followed by electrophysiological and biochemical analysis of the Shaker B K+ channel. Each Shaker B subunit contains seven cysteine residues, including the conserved residues Cys286 and Cys462 and a less conserved cysteine, Cys24S. Each cysteine in the Shaker B protein can be mutated individually without eliminating functional activity, suggesting that the protein does not contain an 'essential' disulphide bond for protein folding or the assembly of active channels (at least under most conditions, see below). Furthermore, limited proteolysis, electrophoresis and immunoblotting under reducing and non-reducing conditions indicate (i) that the two conserved

II

_'--

e_n_try_4_8_

residues Cys286 and Cys462 do not form a lnon-essential' disulphide bond with each other or with Cys24S 207. Subsequently, however, it was shown that disulphide bonds are formed between Shaker subunits in intact cells exposed to oxidizing conditions (detectable by the presence of four high molecular weight adducts on Shaker protein purified under non-reducing conditions). In these cases, intersubunit disulphide bond formation can be eliminated upon serine substitution of either Cys96 in the N-terminal or CysSOS in the C-terminal of the protein212 . These patterns of intersubunit disulphide bond formation (between Cys96 and CysSOS) thus provide important direct biochemical evidence for (i) the N- and C-terminal regions of adjacent subunits being proximal within native channel structures and (ii) Shaker K+ channels contain four pore-forming subunits :S-S bond adducts representing dimers, trimers, and two forms of tetramer (one linear, one circular containing one, two, three or four disulphide bonds, respectively)212.

Domain conservation Interspecies identities (see also Gene family, 48-05) 48-28-01: In general, the mammalian Shaker-related isoforms show a high degree of interspecies identity at the amino acid level. For example, as derived from the sequence alignments in Fig. 4 (under Encoding, 48-19) human, rat, mouse, dog and bovine Kv1.2 amino acid sequences show >98% identity over their entire protein-coding regions. Xenopus Kv1.2 shows >91 % identity to the mouse equivalent. Notably, rKv1.5 and mKv1.5 show greater divergence (870/0 overall identity, discounting an II-residue size difference); this diversity increases in the first 100 N-terminal residues «640/0 identity, as compared to ",,99.40/0 identity between rKv1.2 and hKv1.2 in the same segment). Relatively high amino acid conservation is seen in Shaker-related channels across large evolutionary distances (e.g. aligned sequences of Drosophila Shaker and rKv1.1 (RCK1) still exhibit ",,82 % identityl0). For further background to Kv channel gene evolution, see Chromosomal location, 48-18.

Regions showing high conservation across all Kv subfamily members 48-28-02: The 'core region' sequences comprising a-helical domains 51-56 (see [PDTMj, Fig. 6) show ",,400/0 amino acid identity between any two subfamilies with ",,700/0 identity within each subfamily, irrespective of species. Within this region, conservation is generally higher for amino acid residues facing the intracellular side of the membrane and on transmembrane domains S4, S5 and S6 10 .

K+ selectivity determinants (the P-region) is highly conserved across all K+ channel subfamilies 48-28-03: The P-region (approximately 22 residues, as indicated in the sequence alignment under Encoding, 48-19) can be identified in all known K+ selective channel subunit sequences, including all Kv subunit channels (entries 48 to 51 inclusive), calcium-activated K+ channel subunits (see ILG K Ca, entry 27), inward rectifier K+ channel subunits (see INR K subunits,

II

e_n_t_ry_48

1_ _

_

entry 33) and evolutionary antecedents (e.g. Drosophila K+ channels, see Related sources and reviews, 48-56 and bacteria such as E.coli, see Miscellaneous information, 48-55). Figure 5 shows an enlarged alignment of the P-region with additional features marked including consensus residues across presently known subunit sequences (see also Selectivity, 48-40). In press update: see also the Streptomyces lividans K+ channel crystal structure study494.

tlnternal repeat' arrangements in voltage-gated channels 48-28-04: In comparison to voltage-sensitive sodium and calcium channels which show four internal repeats (each forming six transmembrane segments, see VLC Ca, entry 42 and VLG Na, entry 55), voltage-gated potassium channels consist of one internal Irepeat', with the functional channel arising from the assembly of four monomers and in some demonstrated cases, association of Kv,8 subunits (for details see Predicted protein topography, 48-30, Protein interactions, 48-31, VLC K Kv-beta, entry 47 and VLG Key facts, entry 41). The 'single-repeat arrangement' of K+ channel subunit genes (domains Sl-S6 + HS loop, see VLG key facts, entry 41) probably formed the prototype for other members of the voltagegated channel family via gene duplication events.

Cross-species conservation - overview 48-28-05: Voltage-dependent potassium channels are generally highly conserved from Drosophila to vertebrate central nervous systems (see Gene family, 48-05). Mechanisms partly explaining known patterns of Kv channel gene loci in the murine and human genomes are outlined under Chromosomal location, 48-18. Hypotheses which invoke chromosomal or genomic duplication affecting primordial (ancient) forms of K+ channel genes (see ref.54 and Fig. 1, under Abstract/general description, 48-01) predict that murine Shaker-like genes on mouse chromosomes 3 and 6 have a Ipairwise' parology t (ibid.). This hypothesis is supported by the sequence relatedness of (i) Kcna1 (encoding mKvl.l on chromosome 6) with Kcna2 (encoding mKvl.2 on chromosome 3) and (ii) Kcnca5 (encoding mKvl.S on chromosome 6) with Kcnca8 (encoding mKvl.8 on chromosome 3). Intragenic recombination processes (i.e. exchanging limited stretches of protein-coding regions between highly related genes during homologous pairing t ) would be expected to have important roles in promoting 'functional specialization' of Kv channels during genome evolution. These processes (and larger scale rearrangements) may also act in non-coding regions to alter features such as cell-type expression range (i.e. susceptibility to cis-acting gene regulatory factors t including second messengers and transcriptional activator t /silencer t complexes), subcellular targeting (see Subcellular locations, 48-16) and intron/exon splice site choice (see Gene organization, 48-20).

Poorly conserved sequences in the extracellular loop 51-52 48-28-06: The loop between putative transmembrane domains Sl and S2 is the most variable within the core region of the RCK proteins22 . It has been shown for an Aplysia K+ channel APLK that the 'Sl-S2 loop' is glycosylated t ,

II

II [2] * symbols indicate identical amino acid residues to Drosophila Shaker (top line)

[1] Alignment of putative P-regions of Kv subunit family members (as indicated) against that of Drosophila Shaker (top line) illustrates the high degree of conservation within this region.

r---- P-region

~

[3] ~ symbols indicate conservative substitutions from Drosophila Shaker

Drosophila Shaker NSFFKSIPDAFWWA~

~1.1

E*H*S***********~****~**~*Y**TIG****

Drosophila Shab

~K*V***~*****~****

r~2.1

~K*****AS****~****

Drosophila Shaw

HND*N***LGL***~*****

r~3.1

H~H**N**IG**********

Drosophila Shal

A*K*T***A***~*****

r~4.1

K~*T***A***~*****

rK(v)S.l (:IK8) rK(v)6.1 (K13)

~L*****QS****~****

All Irv-series All K channels

F sIP

} - - See VLG K Kv2-Shab, entry 49

} - - See VLG K Kv3-Shaw, entry 50

} - - See VLG K Kv4-Shal, entry 51

} - - See VLG K Kvx, entry 52

SPE*T***~****~****

fWwa vt:MTTvGYCnn britt

G

~

[4] In consensus sequences, capitals indicate residues that are completely conserved; lower case residues ·predominate '

[5] The IGYG 1 motif (boxed, see text) is 'near-universally I conserved in potassium-selective channels. Notable exceptions are minK channels (see VLG K minK, entry 54) and mammalian K channels related to Drosophila eag (possessing a GFG sequence at the positions homologous to GYG - for details see VLG K eag, entry 46) ("D

Figure 5. Alignment of P (pore) regions from multiple K+ channel subtypes. (From 48-28-03)

=

~

,J::l..

00

1'--_e_n_t_ry_4_8

---'_

providing direct evidence for its extracellular location t (see Fig. 4 under Encoding, 48-19 and ref.213).

IIsoform-specific' properties 48-28-07: The N- and C-terminal regions may vary considerably, generating isoform-specific properties as for K+ channel variants in Drosophila. Further notable examples of domain conservation are given in Table 13.

Domain functions (predicted) Structure-function studies for Shaker-type channels 48-29-01: Special note: The extraordinary range and depth of studies

concerning structure-function analysis of known voltage-gated K+ channels makes it difficult to compile comprehensive summaries without extensive substructuring of the entry, hence the following is a limited overview. Protein 'domain functions' which show similarities across the entire superfamily of voltage-gated channels are also described in VLC key facts, entry 41, VLC Ca, entry 42 and VLC Na, entry 55. Table 14 lists studies which have employed mutational analysis to test hypotheses concerning various functional attributes of Shaker-related K+ channels, subject to the limitations described above. Where applicable, the list is indexed by defined structural loci, cross-referenced to supplementary information in specified paragraphs or entries. See also the additional conclusions from the Streptomyces lividans K+ channel crystal structure study494.

Predicted protein topography Predicted, observed and modelled voltage-gated K channel structures and assemblies 48-30-01: As introduced in VLC key facts, entry 41, protein domain models for Kv channel subunits are broadly similar to those described for voltagedependent Ca 2 + and Na+ channels (see VLC Ca, entry 42 and VLC Na, entry 55). Voltage-gated K+ channel genes represent a single repeat compared to the four repeats typical of the Ca2 + and Na+ channels. Early expectations that four interacting Kvo: subunits form the ion-conducting pathway have largely been borne out by (i) studies on native immunoprecipitated O:4{34 channel complexes (see entry 47 and Protein molecular weight (purified), 48-22); (ii) visualization of tetrameric subunit assemblies in plan view by electron microscopy228 (65 x 65A for Kv1.3)204 and (iii) production of homomultimeric t or heteromultimeric t aggregates with defined composition6,229 (see also ref.s and oxidative induction of intersubunit disulphide bonds under Domain arrangement, 48-27). Heteromultimeric associations contribute to the functional diversity of voltagegated potassium channels by generating distinct properties from the 'parent' homomultimers (e.g. RCK1/RMK2 heteromultimers 44 ). A predicted topography of the Kv1.3 pore region using NMR t -derived structures of scorpion toxins that interact with the pore and complementary toxin and channel mutagenic cycles has also appeared230 . This study predicted that

II

_ entry 48 - - - - - - - Table 13. Cross-references for patterns of domain conservation/divergence in Kv1 subfamily K+ channel proteins derived from amino acid sequence alignments (From 48-28-07) Domain, motif or defined region (ordered by position in primary sequence)

Summary, cross-reference for further details (see note 1)

Intracellular NVariability in Kv1 subfamily (see Fig. 4 under terminus (N) Encoding, 48-19). Kv1 subfamily, specific Role in rapid (N-type) inactivation (see Table 19 examples under Inactivation, 48-37). mKv1.1/MBK1: Residues 71-123 and 287-349 show >95 % identical residues to the Drosophila Shaker protein sequence; nucleotide sequences are divergent (see Isolation probe, 48-12). rKv1.2/BK2: A 136 residue hydrophilic region leading into the first hydrophobic domain contains 94 % identical residues with Drosophila Shaker. Extracellular loop SI/S2 Generally poorly conserved between different (eI2) isoforms, e.g. hKv1.6/HBK2: Currents expressed in oocytes 'closely resemble' those of RCK1/RCK3/ RCK5 despite large variation in the extracellular loop domain between domains 51 and 52. rKvl.6/clone KV2: Homologous to rKvl.5/clone KVl, differing predominantly in the N- and Cterminal regions and in the loop linking the 51 and 52 domains. In rKvl.6/clone KV2, this region is highly acidic, whereas in isolate KVl it is enriched in proline. Transmembrane domain S4 (S4)

Role in voltage-sensing (see Voltage sensitivity, 4842).

rKv1.2/BK2: (example) An uninterrupted sequence of 63 amino acids corresponding to the 54 sequence and the adjacent 55 domain are completely conserved between rKv1.2 and Drosophila Shaker. hKv1.4/hPCN2: (example) Shows conservation of the positively charged (Arg or Lys-Xaa-Xaa) 7-repeat motif of Shaker-type K+ channels in the 54 transmembrane segment. Pore-lining region H5 High conservation in Kv and other K+ -selective (P) and flanking regions subfamilies (see Fig. 4 under Encoding, 48-19). Variations within other ion channel families (see cross-references within Table 11 under Sequence motifs, 48-24).

IL-_e_n_t_ry_4_8

_

Table 13. Continued Transmembrane domain S6 (S6)

High conservation in Kvl subfamily; contributes to part of the pore (see Fig. 4 under Encoding, 48-19).

Hydrophobic core, spanning S1 to S6 (above rows)

Conservation in Kvl subfamily; in particular, all transmembrane regions are (relatively) highly conserved between different Kv1 channels and Drosophila Shaker (see Fig. 4 under Encoding, 48-19).

Intracellular Cterminus (C)

Variability in Kvl subfamily (see Fig. 4 under Encoding, 48-19). Involvement in 'C-type' inactivation (see Inactivation, 48-37). hKvl.5: Stably expressed 32 amino acid C-terminal extensions of hPCNl/Kvl.5 channels by an epitope-fusion tag t can be made without any significant effects on channel activity, monomeric structure or N-glycosylation214 .

Note: 1. Examples only, as taken from the literature; compare alignment in Fig. 4 under Encoding, 48-19 and the PDTM, Fig. 6. For evolutionary antecedents in flies and bacteria see Gene family, 48-05, Chromosomal location, 48-18 and Miscellaneous information, 48-55. Kvl.3 has a shallow vestibule (4-8A deep, ,,-,28-32 A outer dimension, ,,-,2834A at its base) at the external entrance to the pore. The pore (9-14A external entrance) tapers to a width of 4-5 A at a depth of ,,-,5-7 A from the vestibule. High-resolution NMR has also been used to determine structure of the N-terminal inactivation ('ball') domain peptides from rKvl.4 (this entry) and rKv3.4 (entry 50). The compact rKv3.4 ball folds to permit the formation of an intramolecular disulphide bridge structure and exposure of two phosphorylation sites. Both ball domains possess surface regions that are positively charged, hydrophobic and negatively charged. Other atomic scale structural models for Kv channels231-236 are subject to continual refinement, as supported by crystal structures (e.g. ref, 494). Where available on internet resources, contemporary models of K+ channel protein structure or related datasets will be cross-referenced from entry update pages via the CSN website (www.1e.ac.uk/csn/).

'Outer shell' plus 'inner core' structure 48-30-02: A typical monomeric protein domain topography model [PDTM] for Kv 1 subfamily channels is shown in Fig. 6. This broad'S I-S6 + H5' arrangement is preserved in other Kv subfamilies (Kv2, Kv3, Kv4 - entries 49 to 51). Molecular cloning of Kir channel family members led to the distinction of an tinner core structure' necessary for K+ -selective permeation where Ml, H5 and M2 in Kirs are 'equivalent' to the S5, H5 and S6 domains in Kv family channels (compare the [PDTM) in INR K [subunits}, entry 33). Thus Kir channels lack the touter shell' of SI, S2 and

_'--

e_n_try_4_8_

Table 14. Cross-references for structure/function studies in Kv1 subfamily channels based on analysis of defined mutant proteins (From 48-29-01) Function/feature/motif (alphabetical order)

Summary of subtopics and cross-references for further details (see note 1)

Assembly mechanisms

Listed under Protein interactions, 48-31, including: Expression-suppression by truncated channels Kv subfamily-specific assembly Channel assembly as an early event in channel biosynthesis Shaker K+ channels folding and assembling in the endoplasmic reticulum208 (see also review in Isacoff et al. (1993) under Related sources and reviews, 48-56).

Block, determinants

Listed under Blockers, 48-43, including: Clofilium215 Charybdotoxin receptor mappint16 Toxin receptor transfer17 External TEA +- and K+ -binding site Internal TEA+-binding site Dendrotoxin-binding site Mast cell degranulation peptide-binding site 4-Aminopyridine-binding site

Domain functions, comparative

Cross-referenced from Table 13 under Domain conservation, 48-28.

Expression, determinants

Listed under Protein interactions, 48-31, including: Expression-suppression by truncated channels218 Role of certain Kv{3 subunits in modifying expression properties (see also VLC K Kv-beta, entry 47 and Subcellular locations, 48-16).

Motifs, phosphorylation

Listed under Protein phosphorylation, 48-32.

Motifs, N-glycosylation

Listed under Sequence motifs, 48-24.

Motifs, leucine zipper

See Voltage sensitivity, 48-42.

Inactivation, determinants

Listed under Inactivation, 48-37, including: Fast, N-type inactivation; N-terminal'ball' domains 219 'Ball' domain receptor Slow, C-type inactivation Role of Kv{3 subunits in inactivation (see also VLC K Kv-beta, entry 47).

Intersubunit disulphide bond formation

Intersubunit S-S adducts from oxidized (neighbouring) cysteines in assembled native channels (see Domain arrangement, 48-27).

1'---_e_n_t_ry_4_8

_

Table 14. Continued Function/feature/motif (alphabetical order)

Summary of subtopics and cross-references for further details (see note 1)

In vivo mutant analysis

Biological functions inferred from phenotypes of (i) spontaneous or transgenic mutant constructs in Kv channel genes or (ii) transgenic Kv gene deletions ('knockouts') are summarized under Phenotypic expression, 48-14. Mutant gene locit which are associated with (candidates for) the incidence of transmissible phenotypes consistent with defects in Kv channel function are summarized in Chromosomal location, field 18 of relevant entries.

Selectivity, determinants

Listed under Selectivity, 48-40, including: Pore region/ion selectivity determinants (Pregion)22o-223; K+ channel signature sequence mutations224 .

Structural models

Contemporary K+ channel protein structure models or related datasets will be cross-referenced from entry update pages via the CSN (for details, see Feedback &. CSN access, entry 12).

Rectification properties

Listed under Current-voltage relation, 48-35, including: Modification of rectification by domain deletions225,226.

Voltage sensing

Listed under Voltage sensitivity, 48-42, including: S4 voltage sensor mutations Electrostatic interactions in S4 Recovery of 'non-expressible' S4 charge neutralization mutants227.

Voltage-activation

For 'conversion' of depolarization-activated, outwardly rectifying channels into hyperpolarization-activated, inwardly rectifying channels225, see Current-voltage relations (this table).

Zn2+modulation

Listed under Channel modulation, 48-44.

Note: 1. Examples only, as taken from the literature; compare alignment in Fig. 4 under Encoding, 48-19 and the PDTM, Fig. 6.

II

II I (a) Monomeric domains ('unfolded') I Voltage-sensor domain (one of four in assembled multimer)

l

I (b) Monomeric subunit, schematic ('folded')

N-glycosylation site (motif NXT/5) Approximate position of N-glycosylation site in all Kv subfamily members (except Kv1.6)

Extracellular

!I l ~ i!lI lt !~

54-55 loop forming 'receptor' for 'ball' domain

~1[s]]1[Sj

Intracellular

I

?1; )

+++

N H2

Amphipathic 'ball' domain at N-termlnus of fast-Inactivating Kv channel subumts

P-region

~

~

Part of 'ball' receptor

(H5)

Approximate position of consensus PKC site in all known Kv subunits Part of N-terminus , (contiguous With 51 ' \ / not shown; relatlvei y +++ long in Kv1.4; ~ : C-termlnal not shown)

..;

Approximate position of homophilic interaction domain (see Protein interactions, 48-31)

For locations of phosphomodulation motifs see Protein phosphorylation, 48·32; For regions affecting block. see Blockers. 48-43

Amphipathic 'ball' domain at N-terminus of fast-inactivating Kv channel subunits

K

::::::(-----1) :+-J:::':' ::.:::::.: Kv1 : + :::::::.

(c) Tetrameric a-subunit assembly, schematic, in plan see also Protein interactions, 48-31 and VLG K Kv-beta. entry 47

::.::::._'~:::::::

Channel symbol ('b

~

t"'t"

Figure 6. Protein domain topography model [PDTM] for a Kv series potassium channel

Q

subunit monomer. (From 48-30-01)

~ ~

00

1L.-_e_n_t_ry_4_8

_

Table 15. Summary of interactions characterized within Shaker-related channel domains and interactions between channels and other classes of protein. For earlier functional studies on incidence of Kv channel heteromultimer formation, see field text. (From 48-31-01) Interacting domains or discrete proteins

Description of interaction and cross-references

Kvl subfamily a subunits interacting with Kv{3 (accessory) subunits

48-31-02: Native a-DTx-sensitive Kv channels occur as tightly associated oligomeric structures in a stoichiometry t of four a to four {3 subunits (for details, see VLC K Kv-beta, entry 47). Although the large numbers of potential combinations of such a4{34 complexes have not been systematically described at the level of single neurones, they are likely to contain different a subunit and {3 subunit isoforms. The subfamily-specific binding site of K v{31 has been mapped to a region overlapping NABKvl, a N-terminal domain (this table, below); NAB Kv1 is essential for the KV{31-mediated inactivation, as well as a-a and a-{3 interactions 237. Further details on Kva/{3 interactions, physiological roles, modulatory properties, gene family relationships, and other features of Kv{3 subunits are also discussed in entry 47 under appropriate fields.

Kv{31-binding site

Kv1 subfamily a 48-31-03: Kvl.4: Selective protein interactions subunits interacting between the discs large (dIg) family of with SAP97 and PSD-95 membrane-associated guanylate kinases with a Kvl.4 channel C-terminal segment were originally detected following screening of a human brain eDNA-derived yeast two-hybrid t system library155. The key features of these domain interactions are illustrated in Fig. 7 (see also chapsyn-ll0 under Channel density, 48-09). The PDZ 1 +2 structural module

48-31-04: The PDZ domains of the human dIg product (a tumour suppressor t ) are organized into two conformationally stable modules, one consisting of PDZ domains '1 + 2' which binds K+ channels, and the other (PDZ3) corresponding to the third PDZ domain (see Fig. 7). The purified PDZ(1 + 2) module, but not the PDZ3 domain can also specifically bind ATp238 (see also Subcellular locations, 48-16 and Sequence motifs,

See also legend to Fig. 7 48-24). a-Dendrotoxin

a-DTx complexing with Kv subunits, see Protein distribution, 48-15.

III

_~

e_n_try_4_8_

Table 15. Continued Interacting domains or discrete proteins

Description of interaction and cross-references

48-31-05: Associations of different N-terminal domains of Kva subunits are strictly 'subfamily specific' (see Domain functions, 48-29). As shown in Fig. 8, the amino acid sequences of known Shaker subfamily genes from different vertebrate species (short names at left, defined in Gene family, 48-05) and the corresponding region of Drosophila Shaker B (ShB) channels show high sequence conservation in this region. The hydrophilic Nterminal domains of Shaker Band RCKI expressed as bacterial fusion proteins t show homophilic binding visualized by Western blots t of crude protein239. Essentially, this region determines Designation of molecular determinants subunit compatibility for co-assembly. of Kv-subfamily-specific Subsequently, conserved structural motifs in each of the four subfamilies (following detailed subunit associations deletion analysis) have been named NABKvl, NABKv2, NABKv3 and NABKv4240. Note, however, NAB domain some studies242l'243 have found nomenclature electrophysiolog.ical data consistent with heteromultimeric assemblies (i.e. producing 'novel currents') following injection of Xenopus oocytes with a 1: 1 mixture of mRNAs encoding channels from distinct subfamilies (e.g. NGK1/Kvl.2 and NGK2/Kv3.1a)242-244.

N-terminal homophilic interaction domains Selected independent studies defining tetrameric assembly domains of Kva subunits

Tl domain nomenclature

Tl subdomains

48-31-06: In independent studies which assayed post-translational processing and assembly of Aplysia AKOI (Akvl.la) K+ channel subunit proteins through deletion mutagenesis t, only deletions retaining the SI domain [PDTM] (Fig. 6) were inserted into the membrane213 . By means of these deletion analyses, subunit assembly has been shown to be 'critically driven' by a conserved sequence in the N-terminal cytoplasmic region, designated the tetramerization 1 or Tl assembly domain213l'245. Subsequent studies determined that if a subunit protein was to heteromultimerize with a Shaker subunit protein, two regions within the Tl domain, Tl subdomains A and B, must be of the Shaker subtype245 . Moreover, incompatibility of a Shaw A region for assembly with a Shaker protein was shown to depend upon the composition of a 30 amino acid conserved sequence in the A region245 .

"---e_n_t_ry_4_8

--'_


Description of interaction and cross-references Functional Shaker AKvl.la channels are not formed when a glycine is substituted for a serine in the seventh position of subdomain A, indicating a critical role for the serine in wild-type channel assembly245.

Stable, selftetramerization of Tl domain in isolation

48-31-07: Purified T 1 domain self-assembles to form a highly stable tetramer, most probably a closed ring structure246i no separate dimeric or trimeric components can be observed following the assembly process.

Evidence for 'core' assembly domains

48-31-08: Note: A deletion mutant of Kvl.3lacking the first 141 amino acids (Kvl.3, tetramerization domain deleted, Tl-) can form functional channels. A further deletion analytical study has suggested that additional association sites in the central core of Kvl.3 can also mediate oligomerization247. 48-31-09: Deletion of 255 amino acids in the N-

terminal domain of hKvl.4 prevents the formation of hybrid channels within the subfamily but has no effect on homomultimerization or voltagedependent gating. Co-expression of N-terminal deletion mutants of hKvl.4 and Kv2.l result in the formation of functional hybrid channels, suggesting that the N -terminus serves as a recognition site necessary for hetero- but not homomultimeric channel assembly within a subfamily248. Role of Sl segment in assembly

48-31-10: Note: Data maintaining an important role for the 51 segment in the co-assembly of homo- and heterotetrameric K+ channels has been presented, together with a discussion/erratum249 .

Early events in Kv channel biosynthesis and subunit assembly

48-31-11: When Kvl.l and Kvl.4 are co-translated in vitro, isoform-specific antisera co-purifies both proteins even at early time points, suggesting rapid subunit assembly78. Kv subunits belonging to a different subfamily (Kv2.1, see VLC K Kv2-Shab, entry 49) do not assemble with Kvl.l or Kvl.4 using similar co-translation assays. Immune purification with Kv1.1 antisera of surface channels on mouse L-cells transfected with Kvl.l eDNA (see Channel density, 48-09) identify a two-species 57 kDa and 59 kDa doublet band as analysed by SDS-PAGEt (which is absent in precipitates from

Cotranslational assays

II

_L...--

e_n_try_4_8----J


Lack of effect of Nglycosylation on assembly

II

Description of interaction and cross-references sham-transfected cells). Pulse-chase t metabolic labelling with [35 S]3-Met supported the interpretation that the (precursor) 57 kDa species gave rise to the 59 kDa (product) protein within several minutes of initiation synthesis. At longer chase t times, a 57 kDa species reappeared, - indicating both the early precursor and a mature protein had identical electrophoretic mobilities t. Notably, mutation of the consensus extracellular glycosylation t site (N207, see alignment under Encoding, 48-19 and Sequence motifs, 48-24) yielded two proteins at steady state: (i) a 55 kDa core peptide and a 57 kDa species. Loss of the ability to N-glycosylate at N207 had little effect on channel synthesis, turnover, or function. In summary, these and other results 78 suggested (i) heteromeric assembly of Shaker-like channels is co-translational, and (ii) N207 glycosylation of Kv1.1 occurs but is not required for subunit assembly, transport, or function.

Subcellular targeting dependent upon Kv protein interactions

48-31-12: Differential 'sorting' of Kv1.2 in different cell types and subcellular locations has been hypothesized 158 to occur following its association with varying K+ channel subunits, implying that Kv 1.2 may participate in distinct heteromultimeric K+ channels within different subcellular domains (see Subcellular locations, 48-16).

Co-assembly with dominant negative mutants

48-31-13: Some studies250 have shown that, at least in heterologous systems like the Xenopus oocyte, co-expression of wild-type and dominant negative mutant Kv1.1 subunits mainly results in loss-ofactive channels rather than producing channels with altered conductance25o .

Comparative note: Common co-existence of multiple current systems in single Drosophila cells

48-31-14: In Drosophila, Shaker, Shal, Shab and Shaw express independent K+ current systems, and a 'molecular barrier' to heteropolymerization is present 9 . Co-expression of all four K+ channel products does not alter their individual properties. The ability to express multiple, independent 'Acurrent' types together with multiple, independent delayed rectifier types is thus common to both lower and higher eukaryotic cells 9 .

_ _ _ry_4_8 en t

_

L...--

(a)

Kv1.4 C-terminal bait See note 3 and interaction of SH3 domains with proline-rich

.,'.,'

(b)

PSD~5 ~SAP9~~~~~~~~-m~~ld/l:i:i:i:J:i~13_~ I :

I I

I I

I I

I I

I I

I aa 224·405

I

I

I

I

I

I I

aa 736-914

I

hdlg Guanylate kinase domain (puatatlve)

shaded box alternative splice variants of hdlg

Note: A similar domain alignment can be made for SAP-97 and other PDZ-containing famify members

(c)

I

Kv1.4 56

I

Kv1.4 C-terminal

_

intermediate deletions - - - - - - - - - - - - - -

In vitro binding of PSD-95

Yeast 2hybrid assay

+

+

+

+

+

+

CSNAKAVETDV 11

terminal aa - 'sufficient'

TOV 3

terminal aa • 'essential'

+

Figure 7. Clustering of Kvl.4 channels via C-terminal interaction with PDZ domains within PSD-95 and hdlg, membrane-associated putative guanylate kinases. (a) Using Kvl.4 C-terminal as Ibait', screening of a human brain cDNA-derived yeast two-hybrid library identified selective protein interactions with several overlapping clones. [b] Kvl.4C-interacting proteins included PSD-95 and hdlg (== SAP97, human homologue of the Drosophila discs large), both members of a membrane-associated guanylate kinase family, shown aligned here. By two-hybrid assay, binding affinity to Kvl.4C was determined to be in the order PDZl + PDZ2 > PDZ2 alone » PDZl alone. PDZ3 domain binding was undetectable. [c] Progressive deletion analysis revealed that the C-terminalll residues of Kv1.4C (645655) are sufficient, and the last four amino acids are essential for PSD-95 interaction. The C-terminal three amino acids of Kv1.4 (TDV) is highlyconserved in the Kv1 but not Kv2, Kv3 or Kv4 subfamilies (see Encoding, 48-19). Notes: 1. PDZ domains were named from the conjunction of membrane-associated proteins they were first described in: fSD-95, 12iscslarge, the tight junction proteins ZO-l and syntrophin. 2. Mutations of Drosophila Discs large result in profound disorganization of synaptic structure, thus supporting a central role for the PDZ-containing families of

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_8_

S3 segments, which may function as an electrostatic shield between the hydrophobic membrane interior and the. highly charged S4 voltage-sensor sequence.

Protein interactions See also fields of VLG K Kv-beta, entry 47.

Multimeric assemblies and subcellular targetting of Kv channel subunits 48-31-01: A large number of protein-protein and multiprotein functional interactions have been studied with respect to Shaker channel and Kvl family members. These studies have unveiled structural bases for homomultimerism and heteromultimerism, showing their dependence upon interactions within Kv channel molecular assemblies. These studies have also shown that targeting of channels within the cell can critically depend upon interactions between Kv channel domains and other classes of protein (see Subcellular locations, 48-16). Table 15 summarizes some key findings related to intra- and intersubunit interactions which have included references to Shaker-related channels.

Incidence and properties of heteromultimeric channels 48-31-15: The formation of hetero-oligomeric channels containing multiple Kvo: subunit and Kvj1 subunit components (see VLC K Kv-beta, entry 47) has been proposed as an important mechanism for generating functional diversity among voltage-gated K+ channels in situ 132,146,153. Formation of heteromultimeric K+ channels by eDNA co-transfection t and cRNA t coinjection has been demonstrated experimentally for Drosophila 253 and mammalian254,255 Shaker-related proteins. Studies of Shaker channels in Drosophila photoreceptor cells suggests that different members of the Kvl subfamily assemble heteromultimeric channels in ViV0 145 . The generation of channel diversity in vertebrates by this mechanism appears to be restricted (see VLC key facts, entry 41 and Domain arrangement under VLC K Kv3-

Shaw, 50-27).

Properties of heteromultimers compared to constituent homomultimers (examples) 48-31-16: RCKl!RCK4: Heterologously expressed RCKI and RCK4 subunits can co-assemble to form RCK1,4 channels with 'intermediate' properties: Currents mediated by RCKI channels (IKv 1.1) do not inactivate in the

proteins in clustering (see Channel density~ 48-09). 3. The SH3 domain (Src homology domain 3) has been shown to bind to a proline-rich motif in Kv1.5 - see Protein phosphorylation, 48-32. 4. The ESDV motif is found at the C-terminus of the NMDA receptor subunits NR2A and NR2B (see entry 08, Vol. I); C-terminal baits of these proteins (116 aa/122 aa respectively) also bind multiple members of the PSD-95 family. (Figures based on data in Kim et al. (1995) Nature 378: 85-88.) (From 48-31-03)

II

(l)

=

t"1"

~ 90 I

ShB Ak01 XSha2 RCK1 RCK2 RCK3

RCK4 RCK5 Kv1 HBK1

100 I

110 I

120 I

130 I

140 I

150 I

160 I

170 I

180 I

190 I

PQHFEPI PRDHDFCERVVINVSGLRFETQLRTLNQFPDTLLGDPARRLRYFDPLRNEYFFDRSRPSFDAILYYYQSGGRLRRPVNVPLDVFSEEIKFYELGDQAINKFREDEGF NGMGV-GSDYDCS-----------------K-----------N-QK-N--Y---N F EN_FERY_------DSYDPEP--EC-------I---------K--S---E------KK-M------N F I R EE_MEI ---SYPRQADHD--EC-------I---------K--A---H----N-KK-M------N M EE_ME__Y_---EFQEAEGGGGCCSS--L---I------------SL---------G--V-F------------N I_M R__Q E_LAA C LPPAL-AAGEQ-C-------I---------K--C---E------K--M--------N L I I_I R__Q-_EE_ME__y _ GGGGYSSVRyS-C----------------MK--A---E------EK-TQ-----------__N K F_I_T__V Q-_EE_LL__y _ --DTYDPEA--EC-------I---------K--A---E------KK-M--------N I R EE_MEM y EEDQA-QDAGSLHHQ--L--I---------G--A---N-------K--R-----N G S AD__R_-Q-__E_MER ---

SYPRQADH---EC-------I---------K--A---N----N-KK-M--------------D-------------------F------M-----------EE-ME-P-----hPCN1 TVEDQALGTASLHHC--H--I---------G--A---N-------K-LP-------------N-----G-----------------S----AD--R--Q---E-MER------hPCN2 GGGGYSSVRYS-S----------------MK--A---E------EK-TQ-------------N------------------------F-I-T--V---Q--EE-LL-------hPCN3 -PSLPAAGEQDCCG------I--I------K--C---E------K--M-----V--------N------------------------I-I-----R--Q--EE-ME----- _ DFPEAGGGGGCCSS--L---I----------S-SL---------G--V-F------------N--------------------------I-L---R--Q---E-LAA------C HBK2 DRK1

RRVRLNVGGLAREVLWRTLDRLPRTRLGK--DC-TB -S-- QVCDDYSLE-

~

00

196 168 (77) 134 (78)

137(n)

140 (73) 154 (75) 278 (72)

133 (78) 210 (72) 176 (n) 199 (70) 275 (74) 251 (74) 140 (72)

-------RPGA-TS--NF-RT ---BMMEEMCALS--Q-LDYWGIDEIYLESCCQARYB 134(19)

Figure 8. Initial identification of a structural element for homophilic interaction in Shaker-subfamily K+ channels. The homophilic interaction domain was originally defined as a 114 amino acid fragment in the N-terminal (aa 83-196 in the ShE sequence as shown). This sequence is shared by all known Drosophila Shaker alternative splice variants 190?191?197?251 and is highly conserved (> 70% identity) within the Kv1-Shaker-related subunit gene subfamily. Included for comparison in the alignment are the Shaker-related K+ channel gene AK01a from Aplysia252 and the rat Shah-subfamily gene DRK1 (see VLC K Kv2). Dashes (- - - -) indicate runs of amino acids identical to the ShE sequence at equivalent positions. The position of the last amino acid is given by the number on the right? with percentage identities to the ShE sequence in parentheses. The demonstration of a subunit compatibility region is consistent with the ability of ShE and RCK1 (for example) to co-assemble and form heteromultimeric channels253 in contrast to subunits from different gene subfamilies (cf. DRK1 sequence from subfamily Kv2.1 in alignment). (Alignment from Li et a1. (1992) Science 257: 1225-1230.) (From 48-31-05)

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_8_

millisecond time range and are sensitive to DTx and TEA. Currents mediated by RCK4 channels (JKv 1.4 ) inactivate rapidly, but are 'practically insensitive' to DTx and TEA. The RCK1,4 channels inherit properties from both RCK1 and RCK4 subunits mediating transient currents that are sensitive to DTx and TEA. The RCK1,4 channels have a single-channel conductance that resembles RCK1 channels but their gating resembles RCK4 channels255 (see also the heteromultimeric channel RCK1,4 under Current-voltage relation, 48-35).

tExpression suppression' by interaction of functional with truncated channel proteins 48-31-17: Certain deletion mutants of Kv1.3 have been shown to specifically suppress Kv1.3 currents (arising from full-length sequences) following their heterologous co-expression218 . Suppression of 'full-length' Kv1.3 currents requires the truncated Kv1.3 sequence to contain both the N-terminal and 81 domains. In general, N-terminal-truncated DNA sequences from a given K+ channel subfamily can suppress expression of members of (only) the same subfamilr18. Deletion of the first 141 amino acids of wild-type Kv1.3 yields a current identical to that of full-length Kv1.3, but it cannot be suppressed by a truncated Kv1.3 containing the N-terminus and Sl. Furthermore, suppression of native cell (endogenous) currents appears conditional (e.g. constitutively expressed K+ currents in Jurkat or GH3 cell lines are not suppressible; upregulated Shaker-like K+ currents in GH3 cells are suppressible)218.

Protein phosphorylation General perspective on phosphomodulation of K+ channel activities 48-32-01: Phosphorylation of K+ channel Q subunits (and/or their accessory protein components) can alter characteristics such as channel current amplitude, voltage-dependence of gating and the kinetics of activation. Phosphorylation of K+ channels is thus an important general mechanism for modulating calcium entry, action potential firing patterns (threshold, frequency, height, width) and 'effector' responses coupling membrane excitability to secretion, muscular contraction or gene transcription (see Phenotypic expression, 48-14). Activation of neurotransmitter receptors commonly alters excitability and synaptic efficacyt by generating intracellular second messengers t, with several second messengers acting through protein kinaset and phosphataset proteins that alter properties of K+ channels. Specific examples of Kv1 subfamily phosphomodulation are described in the paragraphs below and in Table 16. For further examples and general perspectives, see the review refs256-259, JLG key facts, entry 14, Resource A - G protein-linked receptors and supplementary notes under Receptor/transducer interactions, 48-49. Comparative note: The location of sites for phosphomodulation motifs may vary between studies of the same gene, reflecting slight differences in residue numbering or specification of motif Iconsensus'. In general, 'functional' motifs are highly conserved across species in multiple amino acid sequence alignments; some motif positions are conserved within subfamilies or even families (see following paragraphs).

II

lL....-_ _ _ry_4_8 en t

_

Table 16. Potential phosphorylation motifs and regulatory mechanisms applicable to Kv1 gene subfamily members (for abbreviations see note 1) (From 48-32-03) Kvl.l

48-32-05: mKvl.l/MBKl/rKvl.lRBKl: PKA: Motif at aa 443-446 (cAMP-dependent protein kinase).

'Basal' phosphorylation state - multiple effects of prolonged cAMP analogues 48-32-06: rKvl.l/RCKl: PKA: Ser445; one motif (Ser322) is extracellular and therefore non-functional. RCKI is a substrate for in vitro phosphorylation by the catalytic subunit of protein kinase A 203; Kv1.1 appears to be partially phosphorylated in its basal state in Xenopus oocytes and can be further phosphorylated upon treatment for a short time with a cAMP analogue203 . Site-directed mutagenesis studies demonstrate that phosphorylation of a single site on the C-terminus of Kvl.l can 'fully account' for these phosphorylations 2 0 3 . Further studies 160 showed that although treatment for a short time with various cAMP analogues do not markedly affect the channel function in oocytes, prolonged treatment (lO-16h) with the membrane-permeant cAMP analogue S-{p)-8-Br-cAMPS (8-bromoadenosine 3',5'-cyclic monophosphorothioate) enhances current amplitude through (i) wild-type Kvl.l channels and (ii) mutant Kvl.l channels that cannot be phosphorylated by pkA activation. This enhancement can be inhibited in the presence of the membrane-permeant protein kinase A inhibitor R-(p)-8-Br-cAMPS. These and other experiments support the interpretation that prolonged treatment with S-(p)-8-Br-cAMPS regulates RCKI function via two mechanisms: (i) a pathway leading to enhanced channel synthesis (inhibited by cycloheximide) and (ii) a pathway involving channel phosphorylation that directs channels to the plasma membrane 160 (see also Subcellular locations, 48-16). mKv1.1: PKA: Long-term effects of PKA on mKv1.1 channels have been studied using Chinese hamster ovary (CHO) cell lines stablyt transfected with (i) mKvl.l alone and (ii) co-transfected with both mKv1.l and a plasmid construct expressing a dominant negative t mutation in the regulatory subunit of PKA (PKA-unresponsive, inducing a chronic reduction in the basal PKA activity, following binding to endogenous catalytic subunits of PKA)l09. In the cell lines expressing mutant PKA regulatory subunit (condition ii, reducing basal PKA activity) mKvl.l current density is 3.4-fold higher under condition ii, although current kinetics are unaltered. RNAase protection t assays indicate that levels mKvl.l RNA are marginally increased (by approx. 2-fold) under condition ii while protein levels are increased by approx. 3-fold. These and other results were taken to suggest that PKA can regulate mKvl.l channel expression by changing steady-state levels of RNA and by other post-transcriptional mechanisms 109.

II

_'--

e_n_try_4_8----J


48-32-07: rKvl.2/RCK5: An N-terminal mutation (Thr46Val) eliminates stimulatory effects of injected PKA (see below); motif at aa Ser448.

Mechanism of PKA stimulation of Kvl.2 in oocytes 48-32-08: Kvl.2: PKA: Activation of PKA by (i) application of isoproterenol to oocytes expressing a ,B-adrenergic receptor or (ii) direct injection of the catalytic subunit of PKA induces an increase in Kvl.2 current261 . The effect has been localized to phosphorylation at an N-terminal residue (Thr46, see above); single-channel recordings show that of three comnductance levels of Kvl.2, PKA modulation increases the time spent in the higher conductance states (converting previously 'silent' channel proteins to activate on depolarization).

Tyrosine phosphorylation 'downstream' of 7-TD receptor activation 48-32-09: rKvl.2/RAK: TyrK/PKC: Activation of Ml muscarinic acetylcholine receptors in Xenopus oocytes potently and acutely suppresses co-expressed Kvl.2 current amplitude through a pathway involving [Ca2 +Ji elevation, phospholipase C activation, and direct tyrosine phosphorylation of the channel262 (see note 2). N-terminal Tyr132Phe mutant channels display qualitatively less suppression than wild-type Kvl.2. Further analysis of neuroblastoma cells showed that a similar tyrosine kinase-dependent pathway links endogenous G protein-coupled receptors to suppression of the RAK channel native to these cells262 . The suppression effect of Kvl.2 could be mimicked by activation of PKC. Note: Identification of a novel tyrosine kinase (PYK2) was followed by a direct demonstration of Kvl.2 phosphorylation and rapid suppression of current263 . Notably the PYK2 kinase is activated either by PKC or by [Ca2+]i elevation (see above) and also regulates MAP kinase t functions.

Comparative note on native channel complexes usually containing Kvl.2 (see 48-22) 48-32-10: Patch-clamp analysis of liposomes containing dendrotoxin-binding proteins have been reported to yield K+ channels (Mr rv80 kDa, Shaker-like) whose activity is enhanced by (i) cAMP-dependent protein kinase (see Activation, 48-33) and by (ii) and an 'endogenous kinase' that co-purifies with the channel complex (Mr rv38 kDa, which was not phosphorylated itself in this study)264. Note: This is similar in molecular weight to a Kv,B subunit, see Protein molecular weight (purified), 48-22; no kinase activity has since been reported for Kv,B subunits, but they have been determined to act as protein kinase substrates (see Protein phosphorylation under VLC K Kv-beta, 47-32).

II

1L...-__ _ _ry_4_8

_

en t


(Cross-modulation' of Kvl.3 by different protein kinases

48-32-11: hKvl.3: PKA/PKC: Phosphoamino acid analysis t of the 32P-metabolically labelled Jurkat Tcell type n channel (~hKvl.3) has revealed exclusive phosphorylation of serine residues20o . In vitro phosphorylation studies have shown that the Jurkat type n channel can act as a substrate for both PKA and PKC. Note: PKA also phosphorylates the 40 kDa (,8 subunit) protein which co-immunoprecipitates with the type n channel20o, suggesting that Kv,8 forms part of a regulatory mechanism for Kvl.3 in vivo (see VLC K Kv-beta, 47-22 and 47-32 respectively). 48-32-12: HKvl.3/HLK3; rKvl.3/RCK3: PKA/PKC: Application of 5-HT/cloned 5-HT2 receptor suppresses Kvl.3 265 (for further details see Receptor/transducer interactions, 48-49). Suggestions of 'cross-modulation' by PKA and PKC have appeared266 : Phorbol ester and PKA suppression267 are reversed by PKC inhibitors (H7, staurosporine, polymixin Band anti-PKC antibody). AlkPhos: Application of alkaline phosphatase t (via the patch pipette) increases n type channel conductance under basal conditions and reverses the inhibition produced by PKA266 . 48-32-13: mKvl.3: PKA/PKC: In oocytes co-expressing the mouse 5-HT 1c receptor and mKvl.3 channel, addition of 5-HT (serotonin, 100nM) causes a complete and sustained suppression of Kvl.3 currents in approx. 20min268 . The 5-HT-mediated suppression of Kv1.3 currents proceeds via activation of a pertussis toxin-sensitive G protein and a subsequent rise in intracellular Ca2+, but Ca2+ does not directly block the channel. The mechanism of this suppression is not clear - deletion of the first 146 amino acids from the N -terminal, PKC, calmodulin or phosphatase inhibition do not alter the effect268 . Comparative note: 5-HT has no effect on mKv3.1 currents when co-expressed with 5-HT 1c recepto~68.

C-type inactivation modulated by phosphorylation 48-32-14: There is evidence that the rate of C-type inactivation of Kvl.3 is markedly slowed by simultaneous mutation of three putative phosphorylation sites 269 (see Inactivation, 48-37).

Kvl.3 as a potential regulator in T lymphocyte apoptosis 48-32-15: rKvl.3/RCK3: TyrK: Activation of the Jurkat T lymphocyte Fas receptor associated with induction of apoptosis inhibits Kvl.3 current in Tcells. Inhibition of Kvl.3 current has been correlated with tyrosine phosphorylation of immunoprecipitated and blotted Kvl.3 (see note 1). The TyrK inhibitor herbimycin A (which prevents Fas-induced apoptosis) and p56(lck) tyrosine kinase-minus Jurkat cells abolish Kvl.3 inhibition and phosphorylation by anti-Fas antibody (restoration of the p56(lck) kinase partly restores the inhibitory effect270 .

II

_'--

e_n_try_4_8_

Table 16. Continued

A constitutive phosphorylation/dephosphorylation cycle profoundly influencing Kvl.3 activity 48-32-16: Kvl.3: Co-expression of Kvl.3 with (i) a constitutively active tyrosine kinase (v-src) or (ii) a receptor tyrosine kinase (human epidermal growth factor receptor) in HEK-293 cells results in a large increase in the tyrosine phosphorylation of the channel protein (reversible by alkaline phosphatase treatment before Western blot analysis )271. Inhibition of native tyrosine phosphatases with the membrane-permeant agent pervanadate induces a 'time- and concentration-dependent increase in the tyrosine phosphorylation of Kvl.3, associated with a time-dependent decrease in Kvl.3 current'. TyrK phosphomodulations are eliminated in a Tyr449Phe mutant Kvl.3 271 . Notes: 1. Tyrosine kinases have been shown critical in Fas-induced cell death. 2. Notably, an earlier study reported that deletion of the consensus TyrK region from the mouse Kvl.3 cDNA had no apparent effect on channel expression in heterologous cells (for details, see ref. 268, below and Receptor/transducer interactions, 48-49). Kvl.4

48-32-17: rKvl.4/RCK4: A single putative phosphorylation site at Ser601. PKC: hKvl.4 activity is depressed by stimulation of PKC and hKvl.4 subunits can be phosphorylated by PKC in vitr0 272 .

Kvl.5

48-32-18: hKvl.5/HK2: Putative phosphorylation site at aa 547-550; hKvl.5/HCKI: PKA: cAMP-dependent PKA at Ser556 (HCKI). 48-32-19: hKvl.5/hPCNl: CaMKII: Ca2+ /calmodulin protein kinase II motif at Thrl33 conforming to Arg-Xaa-Xaa-(Ser or Thr). PKA: PKA phosphorylation motif at Ser557 and Ser580 conforming to (Arg or Lys)-(Arg or Lys)-Xaa-(Xaa)-(Ser or Thr)28. 48-32-20: rKvl.5/Isolate KVI: PKA: Four potential sites for phosphorylation by cAMP-dependent protein kinase (PKA); three are in the carboxy-terminal region and one is in the N-terminal region32. CaMKII: The N-terminal phosphorylation site (Ser81) also lies within a casein kinase IT recognition sequence32 .

Tyrosine kinase interactions with Kvl.5 48-32-21: hKvl.5/hPCNI: TyrK: Qirect association of the Src tyrosine kinase with cloned hKvl.5 and native hKvl.5 in human myocardium has been demonstrated273, mediated by an interaction/ binding between the proline-rich motif of hKvl.5and the SH3 domain of Src (see SH3 in Fig. 7 under Protein interactions, 48-31). Tyrosine phosphorylation of Kvl.5 suppresses channel current in cells co-expressing v-Src in a manner similar to Kvl.3 273 (see this table, above). Note: Modulation (suppression) of Kvl.5 by coexpression with receptors for the peptide growth factors FGF (fibroblast growth factor) and PDGF (platelet-derived growth factor) have been reported274 . Similar suppression was observed following co-expression of thrombin or rat 5-HTlc receptors.

II

l_e_n_t_ry_4_8

_


48-32-22: hKvl.6/HBK2: A single intracellular phosphorylation motif at SerSII. 48-32-23: rKvl.6/Isolate KV2: PKA: Two potential sites for cAMPdependent protein kinase (both in the C-terminal region); CaMKII: One potential casein kinase II phosphorylation site within the SI-S2 loop (Ser222 )32.

Shaker 48-32-24: Comparative note: Cytoplasmic application of phosphatases in excised inside-out patches of oocytes expressing Drosophila Shaker channels slows N-type inactivationt gating (see Inactivation, 47-37). Subsequent application of purified catalytic subunit of protein kinase A and ATP reverses the effect, accelerating N-type inactivation (back to its initial rapid rate)275. A C-terminal consensus site for PKA phosphorylation appears responsible for this modulation (for further details see 275). Notes: 1. Information applicable to a given subunit is denoted by field tagging in bold according to the following convention: PKA: protein kinase A; PKC: protein kinase C; CamKII: Ca2 + -dependent calmodulin kinase II; TyrK: tyrosine kinase; AlkPhos: alkaline phosphatase. 2. This study did not identify the tyrosine kinase that was phosphorylating the channel following M 1 receptor activation, but some kinases can be activated by intracellular calcium elevations or protein kinase C (e.g. PYK2).

Conserved protein kinase C modification sites in the S4-S5 cytoplasmic loop 48-32-02: As exemplified by the listings in Table 16, activation of protein kinase C is associated with suppression of Kvl subfamily channel activities. All Kvl subfamily protein sequences possess one or two consensus t sites for modification by protein kinase C (PKC consensus S/T-X-R/L, see Resource G - Consensus sites and motifs). Moreover, such a site is present in the S4SS loop of all known Kv subunits (in a region proposed in some models236,26o to form part of the internal surface of the ion-conducting pathway). Several Kv channels have been shown to act as PKC phosphorylation substrates t in vitro.

Phosphorylation of Kvl subfamily members by protein kinase A isoforms 48-32-03: A number of consensus sites for phosphorylation by cAMPdependent protein kinase (PKA) and other kinases occur in the Kv series (see Table 16 and Resource G - Consensus sites and motifs), mainly in the Nand C-termini. The position of one of these sites in the C-terminus (approx. 30 aa distant from domain S6) is conserved in the Shaker subfamily, while the others appear to be specific to individual channel proteins. In Drosophila Shaker, phosphatase treatment slows the rate of N-type inactivation, and can be reversed by applying the catalytic subunit of PKA; elimination of the C-

II

_____________________ en_t_ry_4_8__

terminal consensus phosphorylation site (KKKS to KKAA) did not alter normal N-type inactivation, but did prevent the action of the phosphatase. By analogy, the possibility that C-terminal PKA phosphorylation sites may regulate mammalian Kv{3 subunit-Kvo subunit associations (and hence (paradoxically) N-type inactivation and channel membrane targeting functions) has been discussed256 (see also entry 47).

A conserved consensus site for tyrosine kinase phosphorylation in

Kvl subfamily proteins 48-32-04: The motif BPSFQAILY, which conforms to a 'consensus' for phosphorylation by protein tyrosine kinase ([R or K or L]-gap of 2 or 3 aa-[D or E]gap of 2 or 3 aa-[Y*]) is present in the N-terminus of all Kvl subfamily channels (for a 'local' sequence alignment of this region see 'Tyrosine kinase' under Resource G - Consensus sites and motifs; for a 'global' sequence alignment, see Fig. 4 under Encoding, 48-19). The high degree of motif conservation strongly suggests a role for protein tyrosine kinase modulation of Kvl subunits in native cells (for examples, see 'TyrK:' headings in Table 16).

ELECTROPHYSIOLOGY

Activation Studies of Drosophila Shaker channels 48-33-01: As further described under Voltage sensitivity, 48-42, the molecular mechanism of channel activation following displacement of voltage-sensing domains has been most extensively studied in Drosophila Shaker channels. There is a large literature on the origins of gating current t (Le. movements of electronic charge across the membrane field during voltage activation e.g. refs276-282). For an overview of these topics, see ref. 283 and other reviews cited under field 56; see also Kinetic model, 48-38. Note: Similarities have been discussed284 between gating mechanisms of Shaker-type channels and ligand-activated ion channels, comparing opening of the NMDA receptorchannel (entry 08) arising from repulsion between negatively charged W590s (analogous to W435s of the Shaker K+ channel).

Voltage activation of Kv subfamily channels 48-33-02: Comparative studies of voltage activation of vertebrate Kvl subfamily currents in oocytes (e.g. rKvl.l, rKvl.4, rKvl.5 and rKvl.6285 ) show a broadly similar voltage dependence of activation, with halfactivation voltages ranging between -50 and -11 mV and maximum steepness (yielding an e-fold change for voltage increments between 3.8 and 7.0mV)285. Kvl.4 has a shallower activation curve, most likely due to coupling with its fast-inactivation process. Table 17 compares some published values of Kvl channel activation parameters, although 'absolute' values may vary between expression sytems or species, and/or show some temperature dependence (see note 2 in Table 17).

II

1

e_n_t_ry_4_8

_

Table 17. Activation properties of Kv1 subfamily homomultimeric channels (From 48-33-02) Kvl.l

48-33-03: rKv1.1/RCK1: Activates at potentials positive to -30mV; for Kv1.1/RCK1 in oocytes, V n,1/2 - 29.7 ± 7mV; an - 6.5 ± 1.8 mV; 22 t n 15.5 ± 4.4 ms (see note 1; other published values for V n include mKv1.l: -27mV, -34mV; hKvl.l: -30mV, see note 2). For mKvl.l in L929 cells V 1/ 2 was determined as -32±2mV286 . Gating behaviour of single channels is 'not homogeneous': channels either occupy open, conducting states (with only brief closures at irregular intervals) or a gating behaviour which consists of a rapid succession of brief openings and closures, both with average durations of milliseconds22 .

Kvl.2

48-33-04: rKvl.2/RAK: Activates with time constants ranging from 58ms at -20mV to 6ms at +60mV; does not show significant inactivation over 800ms. For Kvl.2/RCK5 in oocytes, V n,1/2 - 34.3 ± 9.7mV; an - 4.5 ± 1.3 mV; t n 6.3 ± 1.8 ms22 (see note 1; other published values for V n include rKvl.2: +3mV and +4.8mV; hKvl.2: -5mV, see note 2). For rKvl.2 in B82 cells V 1/ 2 was determined as -27 ± 6mV286 .

Kvl.3

48-33-05: rKvl.3/Isolate KV3: For Kvl.3/RCK3 in oocytes, V n,1/2 - 25.2 ± 7mV; an - 6.6 ± 1.9mV; t n 13.7 ± 6.5ms22 (see note 1; other published values for V n include mKvl.3: -23mV and -35mV; rKvl.3: -23mVand -14mV; hKvl.3: -13 mV and -20mV, see note 2). For mKvl.3 in L929 cellsV l / 2 was determined as - 26 ± 8 m V 286 .

Kvl.4

48-33-06: rKvl.4/RCK 4: Activates in the subthreshold range 22 (~-60mV) similar to native K+ channels l17; for Kvl.4/RCK4 in oocytes, V n,1/2 - 21.7 ± 7mV; an - 16.9 ±3mV; t n 3.2 ± O.8ms22 (see note 1; other published values for V n include mKv1.4: -27mV and -35mV; rKv1.4: -22mV; hKv1.4: -5mV and -34mV, see note 2).

Kvl.5

48-33-07: hKv1.5/hPCN1: -25mV activation threshold; overall, similar to isolate KV1. Isolate KV1: Activation threshold of ~ -40 mV. Half-maximal activation occurred at -3 mV. 48-33-08: hKv1.5/HK2: Ltk- cell line stably expressed t HK2 (cloned from human cardiac ventricular eDNA) has an activation time course that is fast and sigmoidal (time constants declining from 10ms to 500 ms versus ",50 ms typical of many A-currents. Similar to (but not identical to) the T lymphocyte n channel (the mouse n channel 7b is 107 ms, cf. 7b of RGK5 of 612 ms); these differences are attributed to the differences between oocyte and native cell expression31 . HLK3: Inactivation kinetics (inactivating or non-inactivating) are dependent upon mRNA expression level (summarized under Current type, 48-34). Kvl.3 (isolate/clone KV3): A model for state-dependent inactivation properties applicable to Kv 1.3 channels is described under Kinetic model, 48-38.

Kvl.4

Kv1.4/RCK4Kv1.4/RCK4: V h1 / 2 -73.6±5.4mV; ah 12.8±2.2mV (note 1)22. Inactivating, transient outward K+ currents. RCKl,4 heteromultimeric channels recover more than 2 times faster from inactivation caused by depolarizing pulses than those mediated by RCK4 homomultimeric channels 255 . RCK4 and other cloned IK(A) channels have been reported to Ire-open' after repolarization of the membrane (see ref. 307 and Inactivation under VLC K Kv3-Shaw, 50-37).

_e_n_t_ry_4_8

_

Table 20. Inactivation properties listed by subunit (From 48-37-01)

Kvl.5

Kvl.5/HK2/hPCNI/isolate KVl: Vhl/2 - 44.8 ± 1.4mV; ah 5.6mV (note 1)22; delayed rectifiers, time-dependent outward K+ currents. Inactivation closely resembles RCKI (mean time constant for inactivation 1155 ± 54 ms at 30 mV and 690 ± 17 ms at 60 mV)28 (see note 2). I".J

Kvl.6

Kvl.6/HBK2/RCK2: Resemble Kvl.l/RCKI and Kvl.2/RCK5 channels 48 .

Notes: 1. Vh1 / 2 is the pre-pulse membrane potential at which the current response to a step to 0 mV is 500/0 of its maximal value (pre-pulse duration 25 s, holding potential -80 mV; note that RCKI and RCK5 channels do inactivate over a period of several minutes - therefore the value for a 25 s pre-pulse may not reflect the true steady-state equilibrium potential for inactivation). ah is the slope of the steady-state inactivation (h) curve. It is the change in pre-pulse membrane potential (in mV) necessary to cause an e-fold reduction in the size of the response to a test pulse of 0 mV. Inactivation parameter definitions, conditions and comparative data from ref. 22 . 2. Co-expression of hKvl.5 'n subunits' with hKv{j3 induces inactivation behaviour, hyperpolarizing shifts in the activation curve and slowing of deactivation kinetics (see Fig. 4 in Inactivation under VLC K Kv-beta, 47-37). 3. A study of [K+]o elevations on three inactivating channels (rKvl.4/RHKl, Shaker H4 and H37, the latter mediated by C-type inactivation) has appeared317. For all three channels, elevating [K+]o caused an increase in the channels' chord conductances and a negative shift in the calculated activation curves. Several differences related to the channels' inactivation processes were determined317. 4. Recovery from inactivation of ShakerB K+ channels occurs in two phases, a fast phase (lasting for approx. 200 ms) followed by a slow phase (often requiring several seconds for completion). With Na+, choline+, or Tris+ outside, approx. 150/0 of the channels recover in the fast phase (-80mV), and the other 850/0 enter a second inactivated state from which recovery is very slow. Recovery in the slow phase is not influenced by external ions, but is speeded by hyperpolarization331 . 5. The Drosophila Shaker channel lacks a redox-sensitive cysteine at an equivalent position to the mammalian A-type Kv channels and fast inactivation of this channel does not disappear in excized patches308 .

novel 'short-term memory effect' and firing delays similar to those seen in native hippocampal neurones345 . 48-38-02: A review on the kinetics of voltage-gated ion channels has appeared346; see also evaluations of kinetic models for activation of the Shaker channel347, approaches to parameter optimization of gating models for Shaker channels (based on non-ideal voltage-clamp data)348 and citations under Related sources and reviews, 48-56.

II

_ entry 48

- - - - - - - -

Selectivity See note on atomic scale structural models for Kv channels (including pore domains) under Predicted protein topography, 48-30. In-press update: The availability of crystal structure information for the Streptomyces lividans K+ channel (ref. 494 ) has clarified many of the K+ ion selectivity mechanisms described here and in Table 21.

The S4-S5 linker, P-region and S6 segments 48-40-01: A large number of studies (Table 21) have determined that ion selectivity functions in voltage-dependent K+ channels are predominantly associated with a segment of 21 contiguous residues known as the P-region (i.e. the pore region between domains SS and S6, see [PDTM), Fig. 6). The P-region, along with the 86 segment and the 84-85 linker appear to contain most of the pore determinants (see also Table 21). The great majority of pore structure-function studies have used the Drosophila Shaker channel as their experimental system, although parallel studies of Kv channels have provided new insights on pore architecture (particularly by means of channel-toxin interactions - see Blockers, 48-43). A number of biochemical approaches have been key in studying selectivity determinants (e.g. the use of silver ion22o or membrane-impermeant methanethiosulphonate349,35o to probe reactivities with reporter cysteines placed at specified positions in the pore). The topography of the external pore and vestibule of the Shaker K+ channel as initially predicted by silver ion probing in ref. 220 is shown in Fig. 9. Note: Many of the detailed predictions of these approaches are subject to continual refinement, and so are not described in detail in this field. Where they exist, pointers to contemporary

pore (structural) models on the Internet may be available from the CSN website (see entry 12).

A highly conserved amino acid sequence segment across all known K+ -selective channels 48-40-11: In physiological ionic gradients, Kv channels show high selectivity for K+ over Na+ ions. No exceptions to the expected dependence of reversal potentials t on the external K+ concentration have been reported (55 m V/ decade t from tail current analysis t ). As illustrated on the [PDTM] (Fig. 6), all potassium-selective channel proteins comprise external and internal mouths leading to a narrow transmembrane pore that may be simultaneously occupied by several K+ ions. Ion conduction pathways in K+ channels are short (relative to those of ELC channels, Volume I). Permeating ions appear to traverse channel proteins via short 'canals' ('canaliculi') of 10 A or less. For links to structural models for K+ channels, see Domain arrangement, 48-27. Amino acids in the P-region, the only recognizably conserved segment in all K+ -selective channels, irrespective of subfamily or species (see previous paragraph and Miscellaneous information, 48-55) influence ion permeation, selectivity and sensitivity to pore blockers. Drugs and toxins that occlude the channel pore (e.g. extracellular/intracellular TEA+, DTx and CTx, see Blockers, 48-43) interact with residues in the loop linking the S5 and S6 segments and the P-region (see also Table 21).

II

I

entry48

_

~------

Table 21. Protein domains affecting ionic selectivity in Kv1 channels (From 48-40-01) In-press update: descriptions of selectivity determinants from K+ channels protein crystallography 494.

Feature

Description and cross-references

Principal determinants 48-40-02: Experimental approaches using point mutations352,353 or chimaeric constructs354 have of ionic selectivity provided evidence that the ion-selective pore of voltage-sensitive K+ channels is chiefly determined by a sequence of ",20 amino acids located between the highly conserved S5-S6linker (the H5 sequence or The P-region latterly P-region, see the [PDTMj, Fig. 6). This sequence contains amino acids that determine properties such as organic ion and toxin blocker sensitivitr24,303,355,356, selectivity towards NHt and Rb+ 353 and single-channel conductance354.

The P-region 'invagination'

48-40-03: In many K+ channel models (see the [PDTMj, Fig. 6) the polypeptide segment between S5 and S6 forms a deep invagination into the membrane, entering and exiting the lipid layer at the same (extracellular) site210 . 48-40-04: A single-site mutation (T441S) in the H5 region of Shaker has been shown to increase the apparent relative permeability of the channel to NHt, an effect that is sensitive to small changes in external K+ 357. This behaviour is consistent with an anomalous mole fraction t effect which is not apparent in the wild-type channel, and supports the view that T441S alters the affinity of a putative ion-binding site for NHt and ammonium derivatives.

The 'YG' motif

Interconversion of ionic selectivity by point mutations

48-40-05: The property of K+ selectivity in homomeric voltage-gated K+ channels is related to the presence of two 'extra' amino acids, typically YG (Tyr-Gly) that are absent from the pore-forming region of cation-non-selective cyclic nucleotide-gated channels358 (see feature labelled K+ selectivity determinant in the sequence alignment under Encoding, 48-19 and compare the aligned sequence figure in Domain conservation under ILG CAT cAM~ 21-28). Mutations have been introduced into the pore region of voltage-activated K+ channel coding sequences which confer the essential features of ion conduction in the cyclic nucleotide-gated ion channels, i.e. poor selectivity among monovalent cations and divalent cation block358 .

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_8_

Table 21. Continued Feature

Description and cross-references 48-40-06: Comparative note: Ca2 + channel

characteristics can be 'transferred' to voltage-gated Na+ channels when residues Lys (in domain In) and/or Ala (in domain IV) in the P-region of the Na+ channel are replaced by Glu residues (cited in ref.221, see also VLC Na, entry 55). When Shaker P-region residues are mutated in order to 'match' the P-region of the Na+ channel, the mutants can be expressed, whereas when analogous mutants are made in Kvl.l channels, the mutants failed to express221 . 'Non-P-region' domains in ion permeation

48-40-07: Single amino acid substitutions in the 84-85 loop alter Rb+ selectivity, single-channel K+ and Rb+ conductances, and the sensitivity to open channel block produced by intracellular TEA+, Ba2 + and Mg2 + have been described359 . Note: For residues that 'switch' the 'preferred' ionic conductances of K+ to Rb+ in Kv channel pores, see ref. 360 in Selectivity under VLC K Kv-Shab, 49-40.

K+ /Rb+ 'switching'

48-40-08: Evidence supporting a role for the 86 segment in K+ ion permeation (and in governing the sensitivity to internal TEA+ and Ba2 +) has been presented following studies of Shaker-NGK2 (Kv3.1a, entry 50) chimaeras. Transfer of the Kv3.1a S6 segment into Shaker induces the S6 chimaera to adopt the single-channel conductance and blocker sensitivity of the Kv3.1a channel223 .

'Post-S6' region

48-40-09: In ShakerB, exchanging a short stretch of nine cytoplasmic amino acids located just past the S6 domain, ('post-86'), produces a large increase in potassium conductance with no effect on internal or external TEA blockade222 .

Inconsistency of silver ion probe data with eight-stranded {3-barrel models of pores

48-40-10: Early proposals of a ,a-barrel model for a K+ channel pore structure composed of four {3 hairpins 6,352-354 have been contradicted by results derived from reactivities of mutated Cys residues with silver ion22o (see also the [PDTMj, Fig. 6).

Depolarization-induced assay for Kv channels

86

Rb+ efflux as a prototype high-throughput

48-40-12: CHO cells expressing Kvl.5 have been used in a prototype highthroughput assay that measures depolarization-stimulated 86Rb+ efflux as an indicator of K+ channel activation351 . Importantly, there is a high

II

l_e_n_t_ry_4_8

---I_

Figure 9. Topography of the external pore and vestibule of the Shaker K+ channel. A hypothetical single Shaker subunit, a 90° sector of the outer pore region and vestibule is displayed. Residues studied here are indicated by P-region numbering; residues F425 and K427 are also indicated by Shaker numbering. Arrows indicate residues whose positions relative to the pore axis were previously established by mapping with reversible pore blockers. Shaded residues have side-chains that are proposed to project into the aqueous pore of vestibule. Unshaded residues are Ag+ -unresponsive positions that are proposed to project away from the aqueous phase. Diagonal stripes indicate residues whose insensitivity to Ag+ cannot be clearly interpreted in terms of side-chain projection. The reactivity of residue A2 (horizontal stripes) is slower than that of the other Ag+responsive residues. (Reproduced with permission from Lii, Science (1995) 268: 304-7.) (From 48-40-01). In-press update: Compare with details of the P-region/selectivity filter of the Streptomyces lividans K+ channel revealed by protein crystallography 494.

signal:noise ratio in this system, as non-transfected or vector-transfected control cells do not display measurable 86Rb+ efflux under depolarizing conditions. Applications of the assay in discriminating isoform-specific K+ channel modulators has been discussed351 .

I

Single-channel data 48-41-01: A summary of single-channel properties derived for Kvl subfamily channels appears in Table 22.

II

_ entry 48 "------------

Table 22. Single-channel properties of Kv1 subfamily members (From 48-41-01) Kvl.l

RCK1: Slope conductances t of RCKI in its main open state over -60mV to 40mV were 8.9 and 9.3pS with 100mM K+ (cytoplasmic) and normal Ringers (pipette), rising to a 22 pS main state with 100 mM extracellular K+ 343 . RCKl: Chord conductance t (see note 1) assuming a reversal potential of -lOOmV was rv8.7pS22. Stably expressed RCKI channels in Sol-8 cells display a unitary conductance of 14pS77. Mouse Kvl.l stably expressed in L929 cells rvl0pS286.

Kvl.2

RCKS: Rat Kvl.2 stably expressed in B82 cells rv18pS286.

Kvl.3

MK3: In oocytes, mouse Kvl.3/MK3 unitary conductance rv13pS33 (cf. to human lymphocyte n-type K+ channel (13 pS)). RCK3: Chord conductance (see note 1) rv9.6pS22. Mouse Kvl.3 stably expressed in L929 cells rv 14 pS286.

Kvl.4

RCK4: RCK4 chord conductance (see note 1) rv4.7pS22.

Kvl.5

RMK2/isolate Kvl: Channel activity occurs as bursts of fast openings and closings with frequent openings to subconductance states. The main conductance state has a unitary conductance of 7.9 pS. Unitary events activate and exhibit holding potentialdependent inactivation44 . Human Kvl.S stably expressed in MEL cells rv8 pS286.

Kvl.6

HBK2/RCK2: HBK2: 8.7 pSi RCK2: 9.1 pS48.

Notes: 1. For derivation of chord conductance, see Glossary; values listed here all assume a reversal potential of -100 mV. 2. See also unitary conductance comparisons across Kv channels in Chandy and Gutman (1994, see Related sources and reviews, 48-56).

Voltage sensitivity IIntrinsic' voltage sensitivity 48-42-01: A defining function of Kv series of channels is their intrinsic voltage sensitivity, mainly (but not exclusively, see footnote in Table 23) associated with movement of charges within the S4 transmembrane domain upon depolarization, triggering protein conformational changes that open the channel (introduced in VLC key facts, entry 41; see also Sequence motifs, 48-24).

Segment S4 conveys the voltage dependence of the channels as well as the gating charges t which move across the membrane when the channel is activated277,361-363. The equivalent of 12-14 electronic charges are estimated to be transferred across the membrane during the activation of a single Shaker channel364 (also as cited in ref.365). Within Kvo. subunit tetramers, this

corresponds to approximately three charges per subunit. Steady-state measurements on Kvl and Shaker channels obtained by site-directed

II

l_e_n_t_ry_48

---'_

mutagenesist show that stepwise reductions of positive charge within the S4 region correlate with a progressive decrease in the channels' overall gating valence t (gating charge)366 and hence alter tertiary structure367 and voltage dependence of gating368 . Charges that are 'buried' within transmembrane domains will interact more effectively with the transmembrane electric field than those at the surface365 . Note: This field provides only limited coverage of structure-function studies of voltage sensing in Drosophila Shaker channels. For supplementary information, see ref. 283 and those listed in Related sources and reviews, 48-56.

The tArg/Lys-X-X-Arg/Lys' and tleucine zipper' motifs 48-42-02: Although the overall predicted 'domain arrangement' of the proteins encoded by Shaker/Shab/Shaw/Shal gene locit are similar, each subfamily possesses a variable number of repeats of the' Arg/Lys-X-X-Arg/Lys motif' in the S4 putative transmembrane domain (see Sequence motifs, 48-24). A protein sequence motif reminiscent of the leucine zipper t is located immediately adjacent to the S4 transmembrane domain (the S4-S5 loop) of Shaker-, Shab- and Shal-related K+ channels321,369. Leucine zipper motifs (4 to 5 leucines repeated every seventh amino acid residue) are also found in S4 domains of voltage-gated Ca 2 + and Na+ channels and in M4 transmembrane domains of some ligand-gated channels - see also Voltage

sensitivity under INR K/Na IfhQ , 34-42.

Structural implications accompanying S4 movement from lipid to aqueous media 48-42-03: Using Fourier transform infrared (FTIR) spectroscopyt in aqueous solution, Drosophila Shaker S4 peptide in trifluoroethanol adopts an ahelical t conformation370 (in good agreement with the results of 2D NMRt studies of S4 peptide based on rat brain sodium channels - see VLC Na, entry 55). A predominantly a-helical structure is also observed when the S4 peptide is present in aqueous lysophosphatidylcholine micelles t in dimyristoyl phosphatidylcholine acid dimyristoyl phosphatidylglycerol lipid bilayers37o . In contrast to this, the S4 peptide in aqueous solution is in a random coil conformation. The coil-to-helix transition observed for the S4 peptide (upon its transfer from aqueous solution to lipid membrane) suggests the segment has a high degree of conformational flexibility and can undergo large changes in its structure in response to its environment370 (Le. as proposed to occur during voltage activation). Comparative note: The transition may also have some significance for gating mechanisms of ligandgated channels that retain an S4-like motif but are voltage-insensitive (e.g. the CNG channels, see entries 21 and 22).

Reporters for S4 movement prior to and during gating 48-42-04: Single cysteine substitutions have been introduced into the Shaker S4 segment, followed by voltage activation of the variants (in oocytes) while adding the membrane-impermeable cysteine-modifying reagent PCMBS (para-chloromercuribenzenesulphonate, 100 ~M)371. In these studies, PCMBS inhibited K+ currents elicited by mutants L358C, L361C, V363C and L366C, but not those by V367C and S376C, suggesting that the exposure of

II

_'--

en_t_ry_4_8_

84 during gating terminates at L366. By the criterion of cysteine modification, 54 domain movement occurs below the resting potential of the cell (i.e. when the channel would be non-conducting)371. In separate studies, internal and external accessibility of 54 residues in open and closed cysteine-substituted Shaker K+ channels to the membrane-impermeant thiol reagent methanethiosulfonate-ethyltrimethylammonium (MT8ET) indicated that the distribution of buried (voltage-sensing) residues changed (move outward) when channels open372. Furthermore, by combining site-specific fluorescent labelling of the Shaker channel protein with voltage clamping, the conformational change occurring during voltage-gating has been measured in real time373 . By this approach, channel activation was predicted to involve movement of at least seven amino acids of 54 from a 'buried' position, correlated with the displacement of the gating charge. Other work374 has concluded that movement of the 54 domain's N-terminal half (but not the C-terminal end) underlies gating charge, and that this portion of the 54 segment appears to move across the entire transmembrane voltage difference in association with channel activation374 . Note: Opening of a Shaker K+ channel is associated with a displacement of 13.6 electron charge units. Chargeneutralizing mutations of the first four positive charges in the Shaker 54 segment lead to large decreases (approximately 4 electron charge units each) in the gating charge. However, the gating charge of Shaker Ll10 (possessing 10 altered non-basic residues in 54) is identical to the wild-type channel374.

A mutant confirming 'separation' of voltage-sensing and ion-conduction functions 48-42-05: Mutant, non-conducting Shaker channels can still undergo the closed-open conformation in response to voltage changes: A mutation in the pore region of Shaker (W434F) completely abolishes ion conduction without affecting the gating charge of the channel. Gating currents in the non-conductive mutant are identical in their kinetic and steady-state properties to those in conductive channels375 . Extensive mutational analyses of the segment encoding domain 54 in different voltage-gated K+channels have indicated that basic t residues at distinct positions in 54 encounter different structural interactions or charge environments that determine their role in gating. Table 23 summarizes properties of several Shaker channel mutants that have helped identify specified electrostatic interactions that may be important in the mechanism of voltage-dependent activation.

'On' and 'off' gating currents (see glossary)

48-42-11: Gating current t lon' transients for Kv-type and native potassium channels are fast «SO Jls) and have amplitudes up to several tens of picoamps. Upon repolarization to -100 mV following small depolarizations, loff' gating currents are observed, which rapidly «1 ms) reverse most of the 'on' charge displacement. However this fast recovery of gating charge is markedly reduced upon increasing the amplitude of the depolarizing pulse, and the temporary charge immobilization is complete within a few milliseconds at positive membrane potentials363 (cf. Voltage sensitivity under VLC Na, 55-42). This phenomenon occurs for both non-inactivating

II

_ _ _ry_4_8 en t

l....--

_

Table 23. Electrostatic interactions in the Shaker S4 voltage sensor as predicted by structure-function analysis (From 48-42-05) Double mutation

Observed or inferred property

K374Q (S4 'neutralization' mutation - see note 1)

48-42-06: Blocks maturation of the protein, possibly interfering with folding into the native conformation.

R377Q (S4 'neutralization' mutation)

48-42-07: Blocks maturation of the protein (as above).

E293Q (specific 'second-site' rescue mutation in S2)

48-42-08: 'Specifically and efficiently' rescues K374Q. Suggests that K374 and £293 form a strong, local, electrostatic interaction that stabilizes the structure of the channel365 .

D316N (specific 'second-site' rescue mutation in S3)

48-42-09: 'Specifically and efficiently' rescues K374Q. Suggests that K374 and D316 form a strong, local, electrostatic interaction (as above}365 (see also note 2).

R368Q (S4 'neutralization' mutation)

48-42-10: Greatly decreases the valence t of the second component (q2) of charge movement and the proportion of the total charge it carries. R368Q (and R377K) decrease the voltage dependence of the whole-cell current and alter voltage-dependent gating at the single-channelleveI376 . Compared with the wild-type channel, they increase the latency to first opening, destabilize the open state, and alter the equilibria of voltage-dependent transitions, so that some of the charge movement occurs after the first opening376 .

Notes: 1. Although several charge neutralization mutants in S4 could not be functionally expressed, construction of multimeric Kvl.l/RCKI cDNAs has enabled functional expression of all charge neutralizations in the S4 segment227. 2. The S4 segment is not solely responsible for gating charge movement in Shaker K+ channels377. In channels containing neutralization mutations, four positions contribute significantly to the gating charge: £293, an acidic residue in S2 (see above) and three basic residues in the S4 segment: R365, R368 (see above) and R371. and inactivating K+ channels (compare to gating currents measured from Na+ channels, where a maximal two-thirds of the total gating charge mobilizes, and this immobilization is correlated with inactivation)305,378.

Sialidation affects voltage dependence of activation 48-42-12: In native cells, K+ channels form large, octameric sialoglycoproteins t (for details see VLC K Kv-beta, entry 47). Normal post-translational

II

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_4_ 8

--J

addition of negatively charged sialic acids (i.e. glycosylation) can be prevented by transfection of Kv channel cDNAs into Chinese hamster ovary cell lines deficient in glycosylation (Lec mutants)379. Kvl.l channels in Lec mutant lines show a dependence of activation (V 1/ 2 ) that is shifted to more positive voltages with slower activation kinetics compared with controls. A similar positive shift can be recorded in Kvl.ltransfected control cells following treatment of with sialidase t or by raising extracellular Ca2+ (without effect on the Lec mutants). These and other results have been taken to indicate that the presence of negatively charged sialic acid influences the local electric field detected by the Kv 1.1 voltage sensor.

Effects of proline substitutions in S4 segments 48-42-13: The steep voltage dependence of Kvl.l (RCKl) channel opening resides in transitions between closed states, whereas the direct transitions into and out of the open state are very rapid and not markedly voltage dependent. Introduction of a proline residue into the Kvl.l S4 domain (L30SP, at a position corresponding to that in which a Pro is found in voltage-dependent sodium and calcium channels) at anyone of four domains in a concatenated tetrameric channel construct leads to currents with a Ishallower' activation curve380. Additionally, a fast component of deactivation (from strongly positive potentials) is detectable in these mutants. Notably, L30SP substitution in two domains is only tolerated if the domains were non-adjacent38o .

PHARMACOLOGY

Blockers Reported sensitivities of Kv channels to 'classical' blockers (caveats) 48-43-01: A large number of studies have reported properties of peptide toxin, ionic and pharmacological blockers of Kv1 subfamily channels in heterologous expression systems. In attempting to provide an overview of this literature, information types from different studies (pertaining to specified subunits or subunit combinations) have been placed under several italicized subheaders for each blocker or inhibitor (see footnotes to tables). These headers list general information prior to semiquantitative data - some lack of 'concordance' between independent studies make detailed quantitative comparisons difficult. Comparative data (indexed by Kv subunit) is limited to a single compilation table (Table 26). Several 'uncontrolled' variables influencing potency t and selectivity t of block may exist between studies in native and different heterologous expression systems (e.g. heteromultimer formations, association with Kvj3 subunits, variability in expression levels or complex mechanisms of block).

Steps towards rational drug design for K+ channel targets 48-43-02: Problems inherent in rational drug design for ion channel targets have been addressed in a meeting review38l . It is generally acknowledged

II

e_n_t_ry_48

1_ _

---'_

that K+ channels are 'difficult' drug targets, but that drug modulation of K+ channel gating (i.e. including facilitation of opening) has important therapeutic applications. Progress to date has largely depended on understanding in detail molecular interactions between K+ channel-selective peptide toxin blockers (or modulators) and their receptors in the P-region. To this end, a number of pioneering studies (e.g. ref. 23o,382-384) have used assignment of 'pairwise' amino acid interactions between blocker and target defining the spatial arrangement of ion channel residues. Scanning mutagenesis t approaches can identify sets of inhibitor residues critical for making energetic contacts with the channel; using thermodynamic mutant cycle analysis t channel residues relative to the known (rigid) toxin structure can be inferred (for further examples of this approach, see Table 24 and reviews on K+ channel-blocking peptide families 385 and K+ channel inhibitors386). Further properties of ionic and pharmacological blockers citing Kv1 subfamily channels are listed in Table 25.

Perspectives on 'specificities' of K+ channel-blocking toxins 48-43-03: As emphasized by Miller385, individual toxins falling into the charybdotoxin-, noxiustoxin-, kaliotoxin- and tytiustoxin-type subfamilies of K+ channel blockers cannot be relied upon to 'dissect' or selectively inhibit the numerous K+ currents found in native neurones. Natural venoms may typically comprise complex mixtures of "-150 distinct peptides, and only a small proportion (0.01-0.50/0) of these mixtures may have K+ channelblocking activity (the major proportion being represented by use-dependent t Na+ channel toxins which stabilize Na+ channel open states - see Blockers under VLC Na, 55-43). Epileptogenic activity induced by peptide toxins such as MCDP and DTx (see Table 24) may be due to block of K+ channels that are active at the resting membrane potential and control neuronal excitability387. K+ channel blockers are therefore likely to 'co-operate' with the Na+ channel toxins to initiate and maintain depolarization respectively, leading to, for example, disruption of neuronal activity and muscular tetany. Notably, the 'strategy' various venomous organisms use to immobilize their prey is remarkably similar despite the extreme biochemical diversity in venoms. An interesting example has been documented388 in the form of the fish-hunting marine snail Conus purpurascens, which combines neuromuscular block and excitotoxic shock to immobilize its prey rapidly. In this case, 'excitotoxic shock' is induced by the peptide kappa-conotoxin PVIIA (which inhibits Shaker channels) while delta-conotoxin PVIA delays Na+ channel inactivation388 . 48-43-14: Comparisons of toxin sequences that block Kv channels, highlighting the 'n-KTx' systematic nomenclature suggested by Miller, are shown in Fig. 11.

Summary of blocker data indexed by subunit/gene 48-43-30: Discrepancies in 'absolute' reported concentrations required for 500/0 block exist between studies, and this may reflect a dependence on the heterologous expression system employed. For example, compare (i) the higher potency of 4-aminopyridine sensitivity in mammalian cells such as

II

_1....-

e_n_try_4_8_

Table 24. Properties of peptide toxin blockers citing Kvl subfamily channels (see also Table 25, indexing blocker properties by Kv subunit) (From 48-43-02) Toxin/peptide inhibitor

Description, examples and references

Agitoxins I, 2, 3

48-43-04: Purified peptides agitoxins I, 2 and 3 derived from LeiuIUs quinquestriatus var. hebraeus venom are potent inhibitors of the Shaker K+ channel (Kd < 1 nM) and various Kv channels389. Thermodynamic mutant cycle analysis t has been used to map Shaker channel residues relative to the solved390 Agitoxin 2 structure. The K27M (inhibitor) - Y44SF (channel) interaction depends on the K+ ion concentration. These and other approaches384 including agitoxin footprinting391 predict a shallow K+ channel vestibule formed by the pore loops, with the selectivity filter located at the centre of the vestibule approx. sA from the extracellular solution384 .

Mutant cycle analysis

Agitoxin footprinting

Charybdotoxin (CTx)

Toxin-channel mapping

Transfer of toxin sensitivity

Pre-synaptic effects

II

48-43-05: Description: A peptide isolated from the

venom of the scorpion Leiurus quinquestriatus, originally reported as a specific blocker (sic.) of Kca channels (see entry 27). The large amount of information obtained on interactions of CTx with K+ channels has been reviewed385?392. Structural models of charybdotoxin (and other toxins) bound to Shaker and Kv channels have been developed through iterative structure-function analysis of toxin-channel interactions (e.g. refs 216?230?382?383?393-395) generally using the compact and rigid toxin molecular structural co-ordinates as templates for complementarity. This model has been subject to further testing by predicting throughspace electrostatic interactions between specific pairs of channel-toxin residues396 . Sites/mechanism: 'Transfer of scorpion toxin sensitivities from the 'highly sensitive' Kv 1.3 to the insensitive Kv2.1 potassium channel by exchange of a stretch of amino acids between SS and S6 ('the SS-S6 linker') helped confirm the P-region as a toxin receptor17 (see Blockers under VLG K Kv3-Shaw, 50-43). Native: rKvI.3: CTx-induced facilitation of transmission may be partly explained its effects on Kv channels in pre-synaptic terminals of hippocampal inhibitory neurones rather than BKca channels (see ILG K Ca, entry 27). In the presence of TEA+, application of CTx greatly increases the

lL...--e_n_t_ry_4_8

-----J_

Table 24. Continued Toxin/peptide inhibitor

Selectivity of CTx

Pore residues correlating with CTx sensitivity o:-Dendrotoxin (DTx)

Significance in first identifying native K+ channel complexes in brain

Inhibition of slowly inactivating neuronal currents

Description, examples and references amplitude of IPSCs t. In comparison, the specific BKca blocker iberiotoxin fails to augment IPSCs, whereas kaliotoxin and margatoxin (blocking Kv1.3»1.2) mimicks the facilitating effect of CTx137. Specificity: In a direct comparative study286, CTx blocked rKvl.2 and mKvl.3 channels, but had no effect on currents through mKv1.1, hKv1.5 and mKv3.1b channels. Kv1.1 and Kv1.5 have a phenylalanine (f"-J190A3) at a position on the outer mouth of the pore whereas Kv1.3 has a glycine (f"-J60 A3) at this position and was very sensitive to block by CTx. Kv1.2 has a glutamine (of intermediate size at f"-J144A3) at this position; this has been interpreted as permitting CTx to reach its binding site with some steric hindrance resulting in a weaker CTx block than that of Kv1.3. Note: The CTx resistance of mKv 1.1 is in conflict with earlier reports 22 in which ICso (rKv1.1, CTx) was determined as 22 nM. 48-43-06: Description: a-Dendrotoxin is a 59 aa

residue basic peptide from the snake Dendroaspis angusticeps (green mamba) venom. Used extensively to co-immunoprecipitate native K+ channel octomers (40:4,8) from mammalian brain (generally including Kvl.2, for other subunit compositions, see VLC K Kv-beta, entry 47 and text below). The use of dendrotoxins in K+ channel biology has been reviewed397. Sites/mechanism/specificity: DTx blocks current through mKv1.1 and rKv1.2 channels (see above) with high affinity (half-blocking concentration of f"-J20nM) and through mKv1.3 channels with lower affinity (half-blocking concentration of f"-J250 nM). In this comparative study286, there was no effect on current through hKv1.5 and mKv3.1b at 100nM. In Kv 1.1, three residues critical for DTx binding (A352P, E353S and Y379H) are located in the S5-S6 100p398. Native: Dendrotoxins act mainly on neuronal K+ channels. In general, DTx most effectively inhibits slowly inactivating neuronal K+ current. A-type Kv channels are blocked by higher DTx concentrations or are insensitive399 .

II

_L...-

en _ t_ry_4_8_



DMB protein complex DMB protein, a native brain membrane K+ channel complex, has been purified and named on the basis of possessing multiple toxin binding sites for QTx (dendrotoxin), MCDP (mast cell degranulating peptide) and ,B-~Tx (,B-bungarotoxin)67 (see also VLC K Kv-beta, entry 47). MCDP and DTx may block K+ channels active at the resting membrane potential and that control neuronal excitability, possibly underlying their known epileptogenic Epileptogenic activities387. Some DTx-sensitive clones are activities insensitive to ,B-BTx, although this is in contrast to native currents in motor nerve terminals 40o . 'DTx-sensitive' homomultimeric channels (e.g. Kvl.l, Kvl.2, Kvl.6, all exhibiting 'delayedrectifier type' currents in oocytes) generally require Correlation with 4-AP only 50-200 JlM 4-AP to induce block (see 4-Ap, Table 25). sensitivities Kaliotoxin (KTx)

'KTx docking' models of the Kvl.3 vestibule

Margatoxin (MgTx)

l1li

48-43-07: Description: KTx, a Androctonus

mauretanicus mauretanicus (scorpion) venom peptide, was originally described as specific Ca2+-channel blocker (sic.). Potency/specificity: In a single comparative study286, half-blocking concentration of KTx for Kvl.l was rv40nM.. Kvl.3 was almost 2 orders of magnitude more sensitive to block by KTx than was Kvl.l, with a half-blocking concentration of rvO.65nM. Kvl.2, Kvl.5 and Kv3.1 were insensitive to KTx. Sites/mechanism: KTx residue Lys27 interacts with residues in the K+ channel signature sequence (GYGD) in Kvl.3; these and other results from KTx-docking studies have predicted the signature sequence extends into a shallow trough at the centre of a wider external vestibule230,401 (for illustration, see Fig. 10). 48-43-08: Description: 39 residue scorpion toxin.

Sites/mechanism: The solution structure of MgTx402 shows it to be similar to the related toxins charybdotoxin and iberiotoxin; additional residues insert in a manner that extends the ,B-sheet by one residue. Notes: MgTx has been noted to inhibit immune responses in vivo403 .

l_e_n_t_ry_48

----J_



Mast cell degranulating peptide (MCDP)

48-43-09: Description: MCDP (22 aa residues, isolated from bee venom) blocks voltage-gated K+ channels and has potent convulsant activity. Native: MCDP affects both fast-inactivating (native A-type) and slow-inactivating (delayed rectifier) K+ channels. See also DTx (this table). Specificity/potency: In a comparative studr86, mKvl.l and rKvl.2 are half-blocked by "J450nM MCDP; four Kv channels (mKvl.3, hKvl.4, hKvl.5 and mKv3.l) were resistant. Variable Kd values for mKvl.l (e.g. 10-fold lower in ref. 22 ) have may be attributable to toxin puritr86.

Maurotoxin

48-43-10: Chemical synthesis and characterization of maurotoxin, present in the venom of the Tunisian chactoid scorpion Scorpio maurus has been described404 . Synthetic maurotoxin blocks Kvl.l, Kvl.2 and Kvl.3 currents with half-maximal blockage (ICso) at 37,0.8 and l50nM, respectively.

Noxiustoxin (NTx)

48-43-11: Description: A 39 amino acid polypeptide

isolated from the Mexican scorpion Centruroides noxius (see also Pil, Pi2, Pi3, this table). Native: Blocks native (lipid bilayer solubilized) BKca channels and squid axon K+channels. Specificity: Kvl.3 and Kvl.2 are highly NTxsensitive (Kd "J2 nM)286. Pandinotoxins Pi I, Pi2 and Pi3

48-43-12: Three novel 35-residue peptides, Pil, Pi2 (proline at aa 7) and Pi3 (Glu at aa 7), purified from venom of the scorpion Pandinus imperator reversibly block ShakerB channels from the outside with 1: 1 stoichiometry (Pil IC so ''JlOnM405 or Kd "J32 nM in zero external [K+ ]406; Pi2 Kd "J8.2 OM; Pi3 Kd "J140OM)407. Pi2 and Pi3 share approx. 500/0 identity to noxiustoxin (this table); all peptides can displace the binding of [12S I]noxiustoxin to brain synaptosome membranes. Note: An independent group 408 has purified and characterized pandinotoxin (PiTx)-Ka, PiTx-K,B, and PiTx-Kr from venom of Pandinus.

Tytiustoxin or tityustoxin (TsTx)

48-43-13: Tityustoxin-K-alpha interacts with the o:-dendrotoxin-binding site on Kvl.2 K+ channel by binding to the same or closely related sites409.

II

_ _ _ _ _ _ _ _ _ _ _.

en_t_ry_4_8_

Figure 10. Schematic model showing the sites of interation between KTx and the external vestibule of Kv1.3. (Reproduced with permission from Aiyar et a1. (1995) Neuron 15: 1169-81.) (From 48-43-07) CHO to results obtained in whole-cell Xenopus oocytes241,452, and (ii) the slower inactivation of some channels in whole oocytes versus isolated patches (cited in ref.286). The reasons for these discrepancies remain unclear, although they may also reflect an altered accessory subunit constitution between different expression systems (see VLC K Kv-beta, entry 47). Table 26 summarizes Kv1 subfamily blocker data listed by Kv subunit/gene and indicates some reported discrepancies.

Reported variabilities in toxin sensitivities with level of mRNA expression in oocytes 48-43-31: Kv1.2: Injection of 'low' Kv1.2 cRNA concentrations (1""-J0.2ng) into

oocytes have been reported to form Kv1.2 channels sensitive to dendrotoxin I at an IC so of approx. 2 nM344 . In this study, 'high' cRNA concentrations (1""-J20ng) generated Kv1.2 channels which were largely insensitive to dendrotoxin I (IC so approx. 200 nM). At 'low' cRNA concentrations, the expressed Kv1.2 channels were also blocked by other polypeptide toxins such as MCDP (ICso 20 nM), charybdotoxin (IC so 50 nM), and ,B-bungarotoxin (ICso 50 nM), by binding to distinct but allosterically related sites on the channel protein344 . Notably, the channels formed following injection of 20ng cRNA/oocyte were 'totally resistant' to 100nM MCDP and 'hardly altered' by charybdotoxin and ,B-bungarotoxin (1 JlM)344 (see also variabilities in inactivation behaviour with level of mRNA expression reported in the

II

i_e_n_t_ry_4_8

_

Table 25. Properties of ionic and pharmacological blockers citing Kvl subfamily channels (From 48-43-02) Ionic/pharmacological blocker


4-(Alkylamino)-1,4dihydroquinolines

48-43-15: Synthesis and structure-activity relationshipst of a series of 4-(alkylamino)-I,4dihydroquinolines have been described410 as potential anti-inflammatory agents and novel inhibitors of voltage-activated n-type (rvKv1.3) channels on human T cells (IC 5o values from 10-5 to 10- 7 M. The naphthyl analogue 7c exhibits > 1DO-fold selectivity for inhibition of 125 [ I]charybdotoxin binding to n-type channels compared with inhibition of [3H]dofetilide binding to cardiac K+ channels 410 .

4-Aminopyridine (4- AP) 48-43-16: Description: 'Classical' K+ channel blocker. For further information, see this field in other K+ channel entries, this volume. Sites/mechanisms: Block by 4-AP of several cloned Kv1 subfamily channels increases the apparent rate Open channel block of inactivation, suggesting open channel block, with the blocker trapped in the closed channel286 . Effects of 4-AP in modulating slow Ie-type' Interference with inactivation of stably expressed Kv1.5 channels (in 'C-type' inactivation part by inhibiting gating currentst) have been described411 (see also Inactivation, 48-37). The characterization of two functionally distinct Blocker subsites Isubsites' for binding of internal blockers412 may explain why some blockers interfere with C-type inactivation through an allosterict effect, whereas others do not. Experiments performed with chimaeric channels suggest 4-AP-binding sites can be formed from the association of the N-terminal end of S5 and C-terminal end of S6, which are both thought to lie in the inner vestibule413 . In Kvl.4, 4-AP block is potentiated by removing the fast inactivation gate of the channel414 . Potentiation of 4-AP Furthermore, short-pulse trains that activate rKv1.4 block by removal of without inactivation induced more block by 4-AP N-type inactivation than a long pulse that activated and then inactivated gates (see Inactivation, the channel. Binding site accessibility is controlled 48-37) by the channel gating apparatus (i.e. suggesting that both activation and inactivation gates limit the binding of 4-AP to the channel) and binding site affinity modulated by membrane voltage414 . Native: See study on 4-AP block on native lymphocyte Kv channels415 .

II

_'----

e_n_try_4_8_

Table 25. Continued Ionic/pharmacological blocker


Specificity: 'Dendrotoxin-sensitive' homomultimer channels (e.g. Kvl.l, Kvl.2, Kvl.6, all exhibiting 'delayed rectifier-type' currents in oocytes) generally require 50-300 JlM 4-AP to induce block (see Q-DTx, Table 24). Potency: In comparative studies (e.g. ref, 416) 4-AP half-blocking concentrations ranged between rv200 and rv600 JlM for Kvl.l, Kvl.2, Kvl.3 and Kvl.5. Comparative note: mKv3.1 (entry 50) was rvl0 times Relative resistance of more sensitive to block by 4-AP, with a half-blocking A-type channels (for concentration of rv30 JlM. Generally, rapidly exceptions, see entries inactivating channels require millimolar concentrations of 4-AP for complete block. 50 and 51) Anandamide 48-43-17: Kv1.2: Anandamide was identified in porcine brain as an endogenous cannabinoid receptor ligand and is a possible counterpart to the psychoactive component of marijuana (8(9)- tetrahydrocannabinol or 6(9~THC). It has been reported417 that anandamide directly inhibits Shaker-related Kv1.2 channels (IC so , 2.7 JlM)417. 8(9)-THC: also inhibits Kvl.2 channels with comparable potency (IC so 2.4 JlM), as well as several Nacyl-ethanolamides with cannabinoid receptorbinding activity. 48-43-18: Capsaicin, derived from capsicum, is used Capsaicin to define nociceptive t sensory neurones and provides models for deep hyperalgesia. Kvl.l, Kvl.2, Kvl.3, Kvl.5and Kv3.1 are blocked by capsaicin (half-blocking concentration range from 23 JlM (for hKvl.5) to 158 JlM (for mKv3.1b)286. Note: Capsaicin has been shown to block voltage-gated K+ currents in native rabbit Schwann cells, dorsal root ganglion cells, T cells (type .I K+ currents, see Kv3.1), and vertebrate axons (for refs, see286). 48-43-19: CP-339,818 (l-benzyl-4-pentylimino-l, CP-339,818 competitively inhibits 4-dihydroquinoline) and two analogues (CP-393,223 [l 2S I]CTx from binding and CP-394,322) potently block Kvl.3 from the extracellular side in T lymphocytes 90 (IC so rv200 OM to the external and at 1: 1 stoichiometry). Block is use-dependent, vestibule of Kvl.3 with a preference for the C-type inactivated state of the channel (see Inactivation, 48-37). CP-339,818 is also a potent blocker of Kvl.4 but not Kvl.l, Kvl.2, Kvl.5, Kvl.6, Kv3.1-4, or Kv4.2. Notably, CP339,818 suppresses T cell activation in vitr0 90 (see Developmental regulation, 48-11). Relative sensitivity of delayed rectifiers

II

1'--_e_n_t_ry_4_8

---'_



Clofilium

48-43-20: For hKvl.5 stably expressed in CHO cells, steady-state half-inhibition concentration for the class ill antiarrhythmic compound clofilium in inside-out patches is 140 ± 80 nM (cf. the value of 840 ± 390 nM determined in outside-out patches). Clofilium accelerates apparent current inactivation but does not influence the kinetics of current activation or deactivation, with rate of onset of channel block voltage-independent and rate of recovery from block slower at hyperpolarized potentials. Elevation of [K+]o accelerates recovery, suggesting an 'activation trap' mechanism, Le. where clofilium is trapped near the pore418 . Some discrepancy in cited inhibitory concentrations exist, which report block at concentrations higher than the therapeutic dose (e.g. clofilium blocking hKvl.3 at 60 JlM 29 and hKvl.5 at 50 JlM 143 ) See also Blockers under VLG K eag/elk/erg, 46-43.

Acceleration of apparent inactivation

Activation trap

Docosahexaenoic acid (DHA)

48-43-21: External blockade of Kvl.5 by the polyunsaturated fatty acid docosahexaenoic acid (DHA) has been reported419 . DHA accelerates of the apparent inactivation and decreases peak current in a manner similar to that produced by the class ill antiarrhythmic tedisamil.

Diacylglycerol (DAG)

48-43-22: Following activation of phospholipase C t , the common second messenger t intracellular I, 2-dioctanoyl-sn-glycerol (C8:0) (DOG) 'blocks' Kv1.3, Kv1.6 and Drosophila Shaker channels 420, appearing macroscopically as a large acceleration of inactivation rate (doubling the apparent inactivation rate at 162nM DOG). The action of DOG is independent of PKC activation, and appears to act by blocking channel open state rather than modulating gating (DOG and TEA+ ions interact at overlapping but non-identical sites when blocking Kvl.3)42o. Note: Longer carbon chain length DAGs (10 and 12 carbons) are less effective in producing this response.

Diltiazem (non-selective)

48-43-23: The non-selective Ca2 + and CNG channel blocker diltiazem (see Blockers under ILG CAT cGMp, 22-43) also inhibits various Kvl subfamily channels at micromolar concentrations 286 .

II

_L.-

en_t_ry_4_8_



Flecainide (non-selective)

48-43-24: Kvl.l, Kvl.2, Kvl.3, Kvl.5 and Kv3.1 are blocked by flecainide, a class Ic antiarrhythmic drug that inhibits IK but not ITO components in heart (half-blocking concentrations ranging from 53 ~M for mKvl.3 to 217~M for rKvl.2)286. Note: Since these concentrations are significantly higher than the therapeutic dose, it may indicate that homomultimers of these Kv channels are not the therapeutic targets of flecainide.

Nifedipine (non-selective)

48-43-25: Nifedipine blocks voltage-gated calcium channels at nanomolar concentrations (see Blockers under VLG Ca, 42-43); at micromolar concentrations, however, nifedipine also blocks types n (Kvl.3) and 1 (Kv3.1) voltage-gated K+ channels in lymphocytes. Current through various Kv channels is also inhibited by nifedipine at concentrations ranging from 5 ~M (for mKvl.3) to 131 ~M (for mKv3.1b)286.

Resiniferatoxin (capsaicin analogue)

48-43-26: The plant diterpene ester resiniferatoxin is an analogue of capsaicin and binds with high affinity to rat dorsal root ganglion and spinal cord. Resiniferatoxin blocks Kvl.l, Kvl.2, Kvl.3, Kvl.5 and Kv3.1 with half-blocking concentrations ranging from 3 ~M (for Kvl.3) to 46 ~M (for Kv3.1)286.

Terfenadine

48-43-27: Native cardiac atrial myocyte I Kur sensitive to 4-aminopyridine (this table) has similarities to current through Kvl.5o:. At therapeutic concentrations, the antihistamine terfenadine produces a time-dependent non-selective block in Kvl.5a current that is consistent with blockade from the cytoplasmic side of the channel421 . Note: Terfenadine also blocks cardiac current IK,r (entry 46) and IK,s (entry 54).

Tetraethylammonium ions (TEA or TEA+)

48-43-28: Description: 'Classical' K+ channel blocker. TEA+ ions were first used to selectively inhibit the potassium conductance in squid axons over 40 years ago. Sites/mechanism: Mapping of external and internal TEA-binding sites on Kv channels have been discussed in detail within several reviews (see Chandy and Gutman, 1994, and citations in Related sources and reviews, 48-56). External TEA interacts with a tyrosine at the C-terminal end of the

TEA+ site mapping

II

l_e_n_t_ry_4_8

---'_



P-region356 . Kv channels with tyrosines at this position, (e.g. mKv1.1 and mKv3.1) are half-blocked Interactions of TEA+ by rvO.3 mM external TEA286 . TEA-resistant with specified residues channels have a valine (rKv1.2) and an arginine (rKv1.5) in this position. Other published results 13 have stated lower sensitivity of human Kv1.1 to external TEA (Kd of 20 mM), i.e. unlike its mouse and rat homologues. Notably mKv1.3 (possessing a histidine in the homologous position to tyrosine, IC so rv10mM external TEA+ in L929, A4 fibroblasts and isolated oocyte patches33,286. Other whole-cell studies using two-electrode voltage clamp methods 22 report 50 mM external TEA+ as an ICso of rKv1.3. The Kv1.3 histidine residue (above) is also involved in slow 'C-type inactivation' (see Interference with Inactivation, 48-37); thus external TEA+ slows C-type inactivation down Kv1.3's apparent rate of inactivation (Kv1.3 blocked by TEA cannot inactivate332, and the titration of the histidine appears to change the rate of inactivation)34o. Other ions

48-43-29: S5-S6 segment mutations alter patterns of open channel block produced by intracellular Mg2 + (in the 'millimolar range'), Ba2 +359 and TEA+ (see above). 'Strong' inward rectifier K+ channels are several orders more sensitive to internal blockage by Mg2 + ions (i.e. in the 'micromolar- range' - for details, see INR K subunits, entry 33).

same study under Blockers, 48-43). For other variabilities in Kv current properties associated with expression level, see Current type, 48-34.

Channel modulation Listing of reported modulators for Kvl subfamily channels 48-44-01: A number of physiological and pharmacological agents that have been cited as influencing the electrophysiological characteristics of Kv1 subfamily channels appear in Table 27. While many of these may appear to have non-specific mechanisms of action, the listing includes several factors/ agents of potential physiological or pathophysiological significance, including modulatory effects by intracellular and extracellular ions, metabolites, signalling intermediates and redox modulators (compare with

similar listings in other Kv-series entries).

II

_____________________ en_t_ry_4_8.-

_ _ _u__

~1

Common name

0 -

Charybdotoxin-type: CTx lq2 IbTx LbTx

5

10

15

20

--1!L Turn .1L. 25

30

35

subfamily 1 ZFTNVSCTTSKE-eWSVCQRLHNTSRG-KCMNKKCRCYS ZFTQEsCTASNQ-CWSICKRLHNTNRG-KCMNKKCRCYS ZFTDVoCSVSKE-CWSVCKDLFGVDRG-KCMGKKCRCYQ

u-KTx Name

VFIDVS£SVSKE~WAP£KAAVGTDRG-K£MGKKCKCY?

1.1 1.2 1.3 1.4

Noxiustoxin-typ8: subfamily 2 NTx T I I NVKCT-SPKQCSKPCKELYGSSAGAKCMNGKCKCYNN MgTx T I I NVKCT-SPKQCLPPCKAQFGQSAGAKCMNGKCKCYP CITx1 I T I NVK£T-SPOO£LRP£KDRFGQHAGGK£ I NGK£K£YP

2.1 2.2 2.3

Kaliotoxin-type: subfamily 3 KITx GVEI NVKCSGSP-QCLKPCKDA-GMRFG-KCMNRKCHCTP? AgTx2 GVP I NVSCTGSP-QC I KPCKDA-GMRFG-KCMNRKCHCTPK GVP I NVPCTGSP-oe I KPCKDA-GMRFG-KCMNRKCHCTPK AgTx3 AgTx1 GVP I NVKCTGSP-QCLKPCKDA-GMRFG-KC I NGKCHCTPK VR I PVS£KHSG-o£LKP£KDA-GMRFG-K£MNGK£~TPK KITx2

3.1 3.2 3.3 3.4 3.5

Tytiustoxin-type: subfamily 4? TyKa VF I NAK£RGSPE-£LPK£KEA I GKAAG-K£MNGKCKCYP

4.1

Figure 11. Comparison of toxin sequences that block Kv channels, highlighting the a-KTx systematic nomenclature suggested by Mille?85. (Reproduced with permission from Miller (1995) Neuron 15: 5-10.) (From 48-43-14)

Ligands An ultra-high affinity, reversible radioligand for brain Kv channels 48-47-01: Kvl.2; Kvl.3: Monoiodotyrosine margatoxin ([ 125 I]_MgTx) has been described as an 'extraordinarily high-affinity ligand' for voltage-gated potassium channels in mammalian brain439. The [125 1]_MgTx K d of 0.1 pM to purified rat brain synaptic plasma membrane vesicle targets under equilibrium binding conditions has been confirmed by kinetic experiments (Kd 0.07 pM), competition assays employing native margatoxin (MgTx, K· 1"J0.15 pM), and receptor saturation studies (Kd 0.18 pM). Affinity labelling of the binding site in rat brain synaptic plasma membranes employing [125 1]_ MgTx and the bifunctional cross-linking reagent disnccinimidyl snberate induces specific and covalent incorporation of MgTx into a glycoprotein of Mr 74000, a value which reduces to 63 000 upon deglycosylation (compare Molecular weight (purified), 48-22). Antibody studies have suggested that at least Kvl.2 and Kvl.3 are integral constituents of the rat brain MgTx receptor complex. I"J

1

I"J

Receptor/transducer interactions Note: As further described below, some receptor/transducer interactions may be inferred from protein phosphorylation sensitivity (see Protein phosphorylation, 48-32) or from studies of channel modulation by receptor agonists in native cells (see this field in entries marked [native)}.

II

_e_n_t_ry_48

_

tPotential' versus tdemonstrated' couplings of Kv channels to receptor/transducers 48-49-01: This field is limited to studies which have reported heterologous coexpression of specific receptor subtypes with specified Kv channel types. Because of the large size of the Kv subunit family and the number of potential couplings to receptor/second messenger systems, there are a considerable number of predicted receptor/transducer/effector combinations operating within native cells. Some couplings can be inferred from the types of enzymatic effector protein(s) that a given receptor subtype can activate (e.g. phospholipase ct, adenylate cyclaser, guanylate cyclase t, etc.) although these are not equivalent to demonstrations of coupling. Many enzyme components may be part of a single receptor response, and studies often do not discriminate between (i) direct effects of an activated enzyme on a channel protein and (ii) an intermediate role for the same enzyme activating another enzyme within a signal amplification cascade (e.g. a tyrosine kinase isoform being activated by a protein kinase C subtype). Possession of substrate recognition motifs t within a channel's primary sequence (or even in vitro phosphorylation data) is not absolute proof of enzyme-channel interaction in vivo (although it might be evidence in support of it).

tReconstitution' of receptor/transducer/effector pathways 48-49-02: Relatively few studies have been published attempting to 'mimic' channel regulation as observed in native cells with 'entirely' cloned components. 'Partial' reconstitution of native properties may rely on ubiquitous expression of genes encoding other 'critical' components (e.g. those encoding G proteins and enzymes such as kinases or phosphatases). In isolation, 'heterologous reconstitution' data may be misleading without 'full' knowledge of the signal transduction pathway and some control over heteromultimer formation (especially those constituting the effector channel). With these caveats, however, heterologous reconstitution provides a useful route for determining stmcture-function relationships affecting receptor-linked channel modulation properties, particularly when these properties can be performed on native and heterologous cell preparations in parallel (see also next paragraph).

tPunctional consequences' of Kv channel modulation in situ 48-49-03: Receptor/transducer control of Kv (or other) channel activities can be interpreted in terms of the physiological effects which 'modulated' channel opening or closing would have in situ. Thus, the indirect effects of receptor agonists on properties such as voltage dependence of gating, kinetics of activation and current amplitude may be 'interpretable' in terms of effects on cellular function. 'Functional effects' may vary with cell type, but may include alterations in secretory behaviour, rates of action potential firing, action potential shape/duration or efficiency of coupling to muscular contraction or gene transcription events (for specific examples see Developmental regulation (field 11), Phenotypic expression (field 14), Protein phosphorylation (field 32), Channel modulation (field 44) and Receptor/transducer interactions (field 49) of most entries). In general, receptor-coupled modulation which opens K+ channels will hyperpolarize cells, inducing an 'inhibitory' phenotype (e.g. with respect to secretion or neurotransmission in excitable cells). Conversely, suppression of K+ channel current will assist depolarization of cells, inducing

II

II

Table 26. Summary of Kvl subfamily blocker data listed by subunit (From 48-43-30)

TEA (mM)

CTx (nM)

DTx (nM)

MCDP(nM)

4-AP (mM)

Other blockers/specific examples/additional notes (see also Tables 24 and 25)

mKvl.l

0.4 0.3

>1500

21

490

1.1 0.29

rKvl.l

0.6 0.8

22

12

45

1.0 0.16

hKvl.l

20

>100

ND

ND

1.1

rKvl.l/RBKl: TEA (or Et4N+) ion, IC so r-v0.5mM. Using concatenated DNAs to construct Kvl.l (RBKl)-based tetrameric channels containing four, two, or no TEA-sensitive subunits, it has been shown that bound TEA interacts simultaneously with all four subunits 422 . ,B-Bungarotoxin (r-v200 nM)i apamin-insensitive (r-v 1 J,1M).

rKvl.2

>560 129 10

17 6 1.7

24 4 2.8

440 180

0.6 0.8 0.2

hKvl.2

>50

10

ND

ND

0.8

>2000

0.20 0.4

rKvl.2/RCK2: TEA (and more potently, tetrapentylammonium) at the intracellular surface decreases channel open time and increases the duration of closed intervals. Extracellular TEA causes an apparent reduction in single-channel amplitude. Slower block at high-affinity internal site than at low-affinity external site. TEA is a voltage-dependent open channel blocker (see Table 25). The internal TEA-binding site is r-v25% into the membrane from the cytoplasmic margin. External TEA also requires an open channel, but block has less voltage sensitivity. External and internal TEA sites define the inner and outer margins of the aqueous pore 423 . (b

mKvl.3

11

2.6 0.5-20

250

rKvl.3/RCK3/MK3: Oocytes expressing MK3 are sensitive to a wide spectrum of pharmacological agents including (in increasing order of potency)

='

f"1"

~ ~

00

mKv1.4

rKv1.4

>100

>40

>200

>2000

hKv1.4

>50

ND

ND

ND

rKv1.5

>40

>200

>200

>600

hKv1.5

330 >40

>1000

>1000

>10000

0.27 >0.1

rKv1.6

4 7 1.7

1 >3000

20 25

10 200

1.5 0.3

hKv1.6

7

1

20

10

1.5

hKv1.3

II

13

>5003,12

503 11 12 >40 13 1414 30 15 >1600

rKv1.3

>10003

0.8

ND

ND

>1000

>200

>2000

1.5

0.19 0.3 ND

13 1.2 0.8 0.7 0.4

TEA, 4-AP, quinine, verapamil and CTx. Type n lymphocyte channels (",native Kv1.3) show the same order of sensitivity33. RCK3 and MK3 are not blocked by nanomolar concentrations of DTx and MCDp22,67. rKv1.4/RCK4: Transient currents can be recorded in the presence of non-inactivating currents by using TEA at 10-100mM (e.g. ref. 22,255 (ef. RCK1, above).

hKv1.5/hPCN1: Outward current is inhibited by 4-AP dependent on current activation and is enhanced by repetitive stimulation (in 50 JlM 4-AP ",30-38 % block; in 100 JlM 4-AP 54-62 % block; Ki ", 0....'S:-.'"::> #~'lf ~""rs-qj ~ ~ ~ ~ CO 200 nM) calcium-dependent dephosphorylation of M··channels is proposed to increase the short open time, low open probability gating mode 1 activity. The calcium/ calmodulin-dependent phosphatase calcineurin (protein phosphatase 2B, see Related sources and reviews, 53-56) has been implicated as a candidate dephosphorylating activity in M-channel responses. In bullfrog sympathetic neurones, a candidate kinase promoting nl0de 2 gating behaviour is myosin light chain kinase (ibid.). Independent proposals for 'tonic regulation' of M-channels by variations in resting intracellular [Ca2+] (i.e. without any role for a dephosphorylating activity) have appeared (see below). 53-01-06: The 'precise' mechanisms modulating M-channels in vivo are still unclear (see Channel modulation, 53-44)., Notably, additional observations

suggest that activation of calcineurin (see above) does not mediate receptorcoupled suppression of M-current (see Protein phosphorylation, 53-32). Furthermore, although several receptors that couple to phospholipase C and the production of InsPa and diacylglycerol may inhibit K M (see ILG Ca InsP;j, entry 19), the agonist-induced suppression of M-current in frog sympathetic neurones has been reported to be independent of the phospholipase C second messenger cascade. Much work has been done on specific modulatory mechanisms of M-currents (mostly in sympathetic neurones and NGI08-15 cells). In many of these studies (particularly those involving arachidonate and its metabolites and calcium ions) results have shown lack of consensus

l_e_n_t_ry_S_3

_

between studies or have otherwise had conflicting conclusions (or are incomparable due to different recording conditions or lack of controls). Several of these apparently disparate observations could be due to cell-type specificity and consequentially, effector-channel heterogeneity. Breakdowns of M-current modulator and agonist (receptor ligand) actions (with any special conditions or caveats for their interpretation) are tabulated in the fields Channel modulation, 53-44 and Receptor/transducer interactions, 53-49. 53-01-07: Much collated evidence indicates that M-current suppression involves diffusible (probably cytoplasmic) messenger(s) and there is general agreement on this, although not about its/their identity(ies), which remain unknown (uncertain) at the time of compilation (see Protein phosphorylation, 53-32, Rundown, 53-39, Channel modulation, 53-44 and Receptor/transducer interactions, 53-49). However, some authors have proposed that intracellular Ca2+ ions may fulfil all of the roles of the 'M-current messenger', based on reductions in M-channel open probability following application of Ca2+ ions to excised inside-out patches of rat sympathetic neurones (see Table 3 under Channel modulation, 53-44). Since this effect occurred in the absence of ATP, it further suggested independence from protein phosphorylation/dephosphorylation cycles (cf. this field, above) and was used as the basis for proposing a role for Ca2 + ions as a 'direct' inhibitor of M-channels following receptor activation (for further discussion and a perspective on these results, see Table 3). 53-01-08: While the 'molecular identities' of M-channel effectors presently remain as candidates (see Cloning resource, 53-10) some progress has been made defining the likely G protein transducers in the muscarinic (Md receptor and bradykinin (B2 ) receptor responses in rat ganglia (Goq /11, which also activates phospholipase C). Although several exogenous agents can suppress M-currents in a receptor-independent manner (e.g. phorbol esters, see discussion under Protein phosphorylation, 53-32), there are no well-characterized, highly selective pore blockers available for M-channels. M-channels are sensitive to block by external barium ions (e.g. K i == 300 ~M in sympathetic ganglion cells); external Ba2+ shows a preferential block of outward current (for details, see Blockers, 53-43).

Category (sortcode) 53-02-01: VLG K M-i [native]. i.e. Voltage-gated K+ channels (principally) inhibited by muscarinic agonists in native cells (for full receptor range, see Receptor/transducer interactions, 53-49). The tnon-inactivating' channel cDNAs aKv5. 1 and r-eag are candidates for M-channel effectors, as described under Cloning resource, 53-10. However, since no structural 'identity' has been confirmed to date, most properties described in this entry are limited to descriptions of native cell preparations and many are subject to reinterpretation. Although KM channels pass outward 'delayed rectifier-type' currents, their distinctive regulatory properties (ibid.) have not been reproduced in heterologous expression systems with any of the presently known 'delayed rectifier-type' channels formed from Kvl to Kv4 gene products (entries 48 to 51 inclusive).

_i...--

e_nt_ry_5_3_

Information sorting/retrieval aided by designated gene product nomenclatures 53-02-02: Should the gene(s) encoding the KM channel be cloned, a gene product (information retrieval) prefix (Unique Embedded Identifier or VEl) will 'tag' new articles of relevance to the contents of this entry in the CSN pages according to convention (see this field in other entries for details).

Channel designation 53-03-01: Channel proteins with regulatory and kinetic properties similar to the 'classically defined' 'M-channel' (see Abstract/general introduction, 5301 and Phenotypic expression, 53-14) have been designated as K(M)1 K M or K(M). Since the 'molecular basis' or subunit composition of channels underlying M-current is presently unclear, the generic term 'M-channel' is used in this entry. The designation Kx has been used for channels in the inner segment of rod photoreceptors which share the kinetic properties of the M-current observed in other cell types (see Phenotypic expression, 53-14).

Current designation 53-04-01: Usually designated as IK(M), I K .M or 1M . The designation I Kx , for the '1M -like' voltage-dependent K+ current in photoreceptors (see crossreferences in Phenotypic expression, 53-14). The designation IKM,ng has been used for M-current in differentiated NGI08-15 mouse neuroblastoma x rat glioma hybrid cells 1 (see Blockers, 53-43 and Channel modulation, 53-44). As outlined in the entry, 1M is generally designated as meeting criteria such as being a non-inactivating time- and voltage-dependent outward current which is suppressed by muscarinic agonists (e.g. oxotremorine> muscarine > bethanechol); 1M activates at membrane potentials more positive than -60 mV and is slowly turned off when the membrane is hyperpolarized back to -60mV (data for 1M in acutely dispersed coeliac-superior mesenteric ganglia (C-SMG) from adult rat2 ). An extensive comparative study of ion permeation, conduction and blocking properties of native M-current (1M ) versus delayed rectifier (I DR ) in isolated bullfrog sympathetic neurones has appeared3 .

Gene family Similarities and distinctions between K M and cloned K+ channels 53-05-01: Comparative note: The gene family relationships for gene(s)

encoding M-channel protein(s) are presently unknown (but see Cloning resource, 53-10 for candidates). From studies conducted in native cells (this entry), there are some grounds for expecting some structural distinctions between presently known K+ channels and those of M-channels. For example, several conduction properties of M-channels (summarized under Selectivity, 53-40) predict some novel structural features of M-channels that are not seen in the Kvl to Kv4 gene subfamilies. For similarities of KM to novel 'non-inactivating' channels cloned from Aplysia, see Cloning resource, 53-10. Further distinctions between 'classical' 1M , I K and IA

1'--_e_n_t_ry_S3

_

components are outlined/cross-referenced under Activation, 53-33. For example, there is a notable pattern of resistance to 'classical' K+ channel blockers (see Blockers, 53-43). M-channels also exhibit 'hallmark' patterns of modal gating (see Protein phosphorylation, 53-32 and Single-channel data, 53-41) and receptor-coupled modulation (see Receptor/transducer interactions, 53-49). Many other examples of native K+ channel inhibition coupled to ligand activation of 'separate' receptors exist, including that of ATP-inhibited K+ channels (KATP , described under INR K ATP-i [native], entry 30). In this case, functional 'reconstitution' of the native I KATP has been achieved following co-expression of an inward rectifier subunit (Kir 6.2, see INR K [subunits], entry 33) and the sulphonylurea receptor4 . Note that extracellular ATP has been reported to suppress (i) macroscopic McurrentS and (ii) the I KM .ng component in NCI08-IS cells following P2 receptor activation (see Receptor/transducer interactions, 53-49). Intracellular ATP can augment K M activity by promoting mode 2 gating behaviour (e.g. see Fig. 1 under Protein phosphorylation, 53-32).

Trivial names Generic tM' nomenclature denoting a wide range of specific receptor-channel interactions 53-07-01: The terms 'M-current' or 'M-channel' (used on the basis of suppression following muscarinic (M) receptor stimulation) has also been used to describe channel effectors coupled to receptors other than muscarinic types (for examples of these, see Receptor/transducer interactions, 53-49).

EXPRESSION

Cell-type expression index Original description and twell-characterized' preparations 53-08-01: The 'M-current' was originally identified in bullfrog sympathetic neurones 6 as a potassium current that is suppressed by the cholinergic t agonist muscarine. Although it has been described as 'widely distributed' in mammalian sympathetic t and central neurones, M-channels do not appear to be ubiquitous. As described elsewhere in the entry, M-current is particularly well-characterized in sympathetic and parasympathetic 7 ganglia of frog, rat superior cervical ganglion (SeC) neurones, guinea-pig coeliac neurones, rat hippocampus (CAl pyramidal neurones), olfactory cortical neurones8 , cultured spinal cord neurones 9 , NG108-15 differentiated mouse neuroblastoma x rat glioma hybrid cells1 (a 'neurone-like' cell line model expressing the '1M -like' conductance designated as IKM,ng 1 - see Channel modulation, 53-44), toad gastric smooth muscle cells ('1M -like'), pituitary lactotrophs, rat ('non-neuronal, M-like current')10 and rod photoreceptor inner segments (I KX , '1M -like', see Phenotypic expression, 53-14). For more detailed tissue distributions, see the references in Related sources and reviews, 53-56 (see also next paragraph).

II

_ entry 53 - - - - - - -

Comparative note on K+ channel mRNA analyses in sympathetic ganglia 53-08-02: A quantitative expression analysis of 18 different KVQ and Kv{3 subunit mRNAs in rat sympathetic ganglia has been made using an RNAase protection assayll. Eleven Q subunit genes and two {3 subunit genes were found to be expressed in sympathetic ganglia by this method, with evidence for differential expression (between the superior cervical, coeliac and superior mesenteric ganglia) being obtained for Kv{31, KVQl.2, KVQl.4 and KVQ2.2 11 . Notably, none of these distributions 'matched' those of the M-current, which is prominent in sympathetic neurones (this entry).

Cloning resource 53-10-01: Two candidates for M-channel-related cDNAs have appeared: First, when expressed in Xenopus oocytes, the 'non-inactivating' K+ channel aKv5.1 from Aplysia 12 (briefly described in VLG K Kvx, entry 52) has a similar activation slope to KM, activates from -60mV and can therefore contribute to the resting potential and firing patterns of neurones. Secondly, the rat eag-candidacy for an M-channel was mentioned in the perspective by Hille13 and discussed in the original paper14 . r-eag produces a non-inactivating K+ current that is suppressed by muscarine in HEK293 cells (see note). Further to this, two commentaries148,149 have specifically discussed the resemblance between the Drosophila eag current and the mammalian M-current. Note: Initial comparisons between r-eag and M-current (e.g. selectivity, channell current tissue distributions, pharmacology) did not show exact concordance at the time of compilation, and hence have been omitted. Further properties of r-eag are described under VLG K eag/elk/erg, entry 46. 53-10-02: Predicted 'close associations' of 'the K M channel' and kinase, phosphatase or anchor proteins of known sequence (for example, see Protein interactions, 53-31 and Protein phosphorylation, 53-32) could conceivably offer a potential cloning route from eNS cDNA libraries, either by employment of protein immunoaffinityt methods or yeast 2-hybridl technology (for technical background, see Resource D - Diagnostic tests).

Developmental regulation tDevelopmental consequences' of receptor-regulated KM channel activities are unknown 53-11-01: Although largely undocumented, receptor control of neuronal excitabilityt by M-current modulation may be expected to have regulatory roles in 'long-term' developmental expression phenotypes (e.g. involving patterns of de novo gene expression/suppression) in addition to 'short-term' adaptive effects (see Phenotypic expression, 53-14 and Receptor/transducer interactions, 53-49). The 'developmental consequences' of neuropeptide receptor stimulation affecting KM channel gating are largely unknown, but several of these affect neurosecretory and hormone-release phenotypes of significance in brain development {for further background, see the Developmental regulation fields of several ILG series entries (Volume II) in particular those of ILG Ca InsPJ, 19-11 and ILG K Ca, 27-11.

l_e_n_t_ry_S_3

_

Growth conditions affecting neuroblastoma cell morphology and selective expression of 1M 53-11-02: Differentiatedt mouse neuroblastoma x rat glioma (NGI08-15) hybrid cells exhibit both a 'classical' delayed rectifier potassium current (IK ) and an M-current-like component (IKM .ng ). Experimentally defined growth conditions have been shown to markedly affect cell morphology and electrical phenotype of these cells15 . NGI08-IS cells transfected with Ml muscarinic receptors grown with I % foetal bovine serum in the absence of PGE 1 (prostaglandin El ) and IBMX (isobutylmethylxanthine - see note 1) show abolition of the usual pleomorphism, leaving two populations of small cells with (i) stellate morphology and (ii) spherically symmetrical geometries. Whole-cell patch-clamp studies indicate that the two cell morphologies have identical electrophysiological properties, displaying I K , a small current through a 'T-like' Ca2 + channel, but no detectable M-current. Stimulation with the muscarinic agonist carbachol can shift the distribution of cells to a 'more stellate morphology' (within 24 hours) and later 'downregulate' 80K/MARCKSt by 22 ± 70/0 (after 48 hours, see note 2)15. Notes: 1. PGEl and IBMX are usually added in preparation for electrophysiological studies of NGI08-IS cells. 2. MARCKS is an acronym for myristoylated, alanine-rich C-kinase substrate (also known as 87K, pp80). MARCKS is a prominent calmodulin-binding PKC substrate. K+ -induced depolarization of rat hippocampal slices results in significant phosphorylation of this and other PKC substrates (see, for example ref16. and Resource G - Consensus sites and motifs).

Phenotypic expression General phenotypic functions/characteristics of M-channels 53-14-01: Open M-channels provide an outward, K+ -selective (hyperpolarizing) current under 'resting' conditions (i.e. in the absence of receptor agonists that induce its suppression) and thus plays an important role in stabilization of cell excitabilityf (see below). Suppresion of M-current following activation of several G protein-linked receptors (see Receptor/transducer interactions, 53-49) results in membrane depolarization and an increase in membrane input resistance (i.e. making it more likely that a cell will fire action potentialst). When neurones are depolarized (i.e. towards the 'threshold for firing' action potentials) 'slow activation' of M-channels tend to hyperpolarize the cell membrane back towards rest (see also Activation, 53-33). In the absence of agonist-induced inhibition, M-channels do not inactivatet under maintained depolarization, and thus are often described as being 'tonically active', remaining 'steadily activated' or 'persistent' at potentials positive to -70mV6 . Upon membrane hyperpolarization, M-channels are deactivatedt ('turned off'), exhibiting slow relaxationsf in response to voltage jumps (see also legend to Fig. 3, under Blockers, 53-43).

tClassical' observations defining M-current function in sympathetic neUlones 53-14-02: M-current plays a major role in spike adaptationt phenotypes. Synaptic suppression of M-current following application of muscarinic-cholinergic

_'---

en_t_ry_5_3_

agonists underlies 'slow' and 'late slow' excitatory post-synaptic potentials (slow EPPs) in sympathetic neurones17. Induction of slow EPPs reduces adaptationt and permits prolonged, repetitive firing of action potentials, for example in sympathetic ganglia18.

tOver-recovery' phenotypes in sympathetic neurones 53-14-03: M-current is transiently augmented following removal of receptor agonist in frog sympathetic neurones, a phenomenon termed 'over-recovery,t19. 'Over-recovery' is [Ca2+ h-dependent and a similar phenomenon has been shown following removal of agonists that induce suppression of voltage-gated calcium currents (for details, see Protein phosphorylation, 53-32 and Single-channel data, 53-41).

Phenotypic roles of M-like K+ channels in rod photoreceptor inner segments 53-14-04: An 'M-channel-like' activity designated as Kx in has been characterized within the inner segment of salamander rod photoreceptors20- 22 (for background, see Fig. 2 under Phenotypic expression of ILG CAT cGMp, 22-14). IKx has likely roles in (i) setting the dark resting potential; (ii) shaping small photovoltages (i.e. 'accelerating' rod responses to light of low intensity)21 and (iii) sensory adaptation: In rod cells, I Kx has been described as a 'standing' outward current (of about 40 pA at -30mV in the dark) that deactivates t slowly (Tmax == 0.25 S)21 following light-induced hyperpolarization (due to cGMP-gated channel closure - (see ILG CAT cGMp, entry 22). As IKx falls, Ih channels (described under INR K/Na IfhQ, entry 34) tend to open (particularly with large hyperpolarizations). Thus both Kx (closures) and hchannel (openings) tend to oppose or 'damp-out' the strong hyperpolarizing effects of light. Furthermore, since Kx and h-channels have relaxation times in the order of 100-200ms (at 22°C) they reduce the photoreceptor voltage response over this time course. These and other mechanisms of neural processing in the photoreceptor inner segment (including the role of Kx channels) have been reviewed22 . Note: Although the voltage and time dependences of IKx are similar to M-current, IKx can be distinguished from M-current because it is not suppressed by acetylcholine and external Ba2+ lblock' shifts the activation range of I Kx strongly in the positive direction21 (for further details, see Blockers, 53-43).

M-channels in other cell types 53-14-05: 1. Suppression of M-current by synaptic activation may underlie seizure-generation activity in cortical neurones23 . 2. 'KM -like' channel responses have also been characterized in gastric smooth muscle cells (see Receptor/transducer interactions, 53-49). 3. Changes in 1M -like expression coincident with phenotypic changes in neuroblastoma cell morphology are described under Developmental regulation, 53-11.

Functional distinctions between receptor-coupled KM and KCa channels 53-14-06: The opening of any K+ channel with depolarization tends to limit the depolarization itself, and in this regard, 1M has been referred to as a

1'--_e_n_t_ry_5_3

_

lbraking' current, effectively bringing action potential firing frequency under 'receptor control' (see previous paragraphs and Receptor/transducer interactions, 53-49). This type of function is broadly similar to receptormediated reductions in calcium-activated K+ currents (for details, see ILG K Ca, entry 27). Although KCa inhibition may occur over a wide range of membrane potentials, M-current inhibition can only occur when the channels are open (i.e. by depolarization from about -60 m V). In bullfrog sympathetic ganglion cells, muscarine reduces both I K.M and IK.Ca, while in other cells (e.g. hippocampal pyramidal neurones) approx 10-fold greater concentrations of muscarinic agonist are required to suppress IK.ca than for reduction in I K.M • Differential transducer t /effector t coupling in different cell types may be able to account for these observations (see Receptor/ transducer interactions, 53-49).

Protein distribution 'Differential distribution' of M-channels by functional criteria 53-15-01: Based on their response to a maintained depolarizing current stimulus, neurones in rat pre-vertebral and paravertebral sympathetic ganglia can be classified as lphasic' or ltonic' neurones. Notably, M-current has been characterized as present in all phasic neurones, but is 'very weak' or absent in 'tonic' neurones24 . Computer models (based on voltage-clamp data), suggest that different firing properties of phasic and tonic neurones can be accounted for by differential expression of the M-current/channels24 .


Protein interactions Predicted close associations of K M channel and kinase/phosphatase proteins 53-31-01: Patterns of 'reciprocal' phosphomodulation of M-channel gating modes25 (see Protein phosphorylation, 53-32) predict native M-channels to be closely associated with enzymes regulating both channel dephosphorylation (e.g. calcineurin, protein phosphatase 2B) and phosphorylation (putatively MLCK, ibid.). Note: Approximately 50-700/0 of rat brain calcineurin is membrane associated26 and is known to bind to membraneassociated 'anchoring protein' which may 'target' calcineurin to its protein substrates, such as ion channels. Single classes of anchoring protein are known to associate with both calcineurin and protein kinase A27. Observations of K M activities in sympathetic neurone preparations (see Protein phosphorylation, 53-32) are consistent with both kinase and phosphatase activities remaining associated with membrane patches following excision.

Functional interactions involved in 'neural processing' within photoreceptor inner segments 53-31-02: In the photoreceptor inner segment, Kx channels (sharing kinetic properties of 'classical' M-channels) appear to shape responses to dim light

_'--

e_n_try_5_3_

and set the dark resting potential. For functional interactions of Kx channels with cation-selective, hyperpolarization-activated h-channels (described under JNR KINa JfhQ , entry 34) and cGMP-gated channels (described under JLG CATcGMp, entry 22) see ref.22 and Phenotypic expression, 53-14).

Protein phosphorylation Ca 2+ -dependent phosphorylation/dephosphorylation underlying K M gating mode transitions 53-32-01: Calcium-dependent phosphorylation/dephosphorylation cycles have been shown to determine macroscopic M-channel amplitudes in bullfrog sympathetic neurones by controlling transitions between gating modes 1 and 2 (for illustration, see Fig. 1, this field and Single-channel data, 53-41). Agonists such as muscarine decrease M-channel activity by selective reduction of the long open time, high open probability gating mode 2 (ibid., ref. 28, but see also note 3). At moderate [Ca2+h (50-150nM) calcium-dependent phosphorylation appears to promote gating mode 2 (ibid.), producing characteristic macroscopic M-current relaxations t. At higher [Ca2+Ji (>200nM) calcium-dependent dephosphorylation of Mchannels increases the short open time, low open probability gating mode 1 activity. The calcium/calmodulin-dependent phosphatase calcineurin (protein phosphatase 2B, see Related sources and reviews, 53-56) has been implicated as a candidate dephosphorylating activity in M-channel gating mode shifts25 (Fig. 1): Addition of a 'pre-activated' form of calcineurin (CaN420, see note 1) to whole-cell pipette solutions inhibits the macroscopic M-current in sympathetic neurones. In excised, inside-out patch recordings, this latter effect has been shown to be associated with a selective loss of mode 2 M-channel activitr5 (see note 2). Notes: 1. The pre-activated, calcium-independent form of rat brain calcineurin (CaN420 ) contains a stop codon at residue 420 which lacks an autoinhibitory domain t that normally binds both calcium and calmodulin29 " 2. The action of CaN420 can sometimes be reversed by the inclusion of ATP in whole-cell pipette solutions or by application to the intracellular face of excised patches. Long Popen states restored by ATP are, howeve.r, distinguishable from 'authentic' mode 2 behaviou~5. 3. Additional observations25 suggest that activation of calcineurin does not mediate receptor-coupled suppression of M-current. For example (i) cells dialysed with calcineurin autoinhibitory peptide (in the electrode solution) retain sensitivity to suppression by muscarine and (ii) the kinetic mechanism for effects of CaN420 appears different from that of muscarine (i.e. kinetic models predict effects of muscarine being due to selective reduction in mode 2 with no effect on mode 128, while CaN420 produces a large increase in transition rate between modes25 ). For further details on calcineurin affecting M-current 'rundown', see Rundown, 53-39.

MLCK: A candidate kinase promoting mode 2 gating behaviour of K M 53-32-02: The effects of ATP in 'restoring' CaN42o-inhibited M-current (see paragraph 53-32-01) take several minutes to plateau, consistent with the activation of an ATP-dependent kinase (of unknown identity, see Fig. 1, this field). This kinase does not appear to be (i) protein kinase A (as addition of

('l)

(b) Modal gating shifts by phosphorylation/dephosphorylation

(a) Suppression of M-current by agonists

(Comparative note: muscarine induces selective reduction of mode 2, see this field)

(diffusible messenger; presently unclear)

(mode 2) K

e.g for ACh and BK in rat ganglia

Pi

II

long open time high open probability

open time behaviour shows no apparent voltage-dependence

open time and voltage-dependence similar to macroscopic M-current deactivation

_--- (dephosphorylated) ----

. . -- I

Unknown 'inhibitory' messengers (for distinctions between possible mechanisms, see also Single-channel data, 53-41 and Channel modulation, 53-44).

short open time low open probability

K

diffusible

Elevated 'local' increases in intracellular [Ca]? For reported 'direct' inhibitory effects of calcium (independent of phosphoregulation) See Channel modulation, 53-44; see also ILG Ca InsP3, entry 19, ILG K Ca, entry 27 and VLG Ca, entry 42

Gating mode 2

Gating mode 1

\

~

'"

I

.

'Modest' elevations of Ca (50-150 nM) ~ M-current • 'Larger' increases (>200 nM) can inhibit M-current See Channel modulation, 53-44.

_

I?-

ATP- and Cadependent kinase

I

I

,

,/ Putatively, myosin

" II I I I I I I

",/

'1

Phosphorylation-dephosphorylation cycle

I·"

light-chain kinase (MLCK, see text)

controlling transition between gating modes and controlling M-current amplitUde

•

Figure 1. Proposed Ireciprocal' regulation of KM channel modal gating behaviour by Ca 2+ -dependent phosphorylationdephosphorylation cycles. For background, see paragraphs 53-32-01 and 33-32-02. Based on data from N.V. Marrion (1996) Neuron 16: 163-173.

::st"'t'" ~ CJ1 CJ.J

_ entry 53

- - - - - - cAMP analogues or forskolin do not affect macroscopic M-current in sympathetic neurones30,31) or (ii) protein kinase C (as activation of PKC by phorbol esters or diacylglycerol analogues inhibit the M-current31 - 35 (see this field, below). A candidate for this kinase is myosin light chain kinase (MLCK) and a 'pre-activated' form of MLCK enhances M-current in bullfrog sympathetic neurones36 . MLCK has been shown to be expressed in these neurones and application of MLCK inhibitory peptide or wortmannin suppresses both (i) agonist-induced 'over-recovery' (see Phenotypic expression, 53-14) and (ii) 1M when measured at a intracellular calcium concentration of 100 nM36,37. Mechanisms and calcium dependency for wortmannin inhibition of 1M were further refined in ref. 38 .

Comparative note: Protein phosphatase inhibitors without apparent effect on 1M 53-32-03: Inclusion of diphosphoglyceric acid (DPG, 1-2.5 mM, a phosphatase inhibitor) or alkaline phosphatase (100 Jlg/ml) fail to affect amplitude of muscarinic responses of 1M channels in frog sympathetic ganglion cells to muscarine39 (see also Channel modulation, 53-44 and Rundown, 53-39). The protein phosphatase 1 (PP1) inhibitors okadaic acid (1 JlM) and microcystin LR (200nM) also do not have any effect on macroscopic M-current in bullfrog sympathetic neurones25 .

The mechanisms mediating agonist-induced suppression of K M are presently unclear 53-32-04: Several receptors that couple to phospholipase C and the production of InsP3 and diacylglycerol may inhibit KM (see ILG Ca InsP3, entry 19). However the agonist-induced suppression of M-current in frog sympathetic neurones has been reported to be independent of the phospholipase C second messenger cascade32 . Furthermore, LHRH receptor activation has been shown to inhibit M-current in a PKC-independent manner19. In hippocampal pyramidal cells, agonists that stimulate phosphatidylinositol (PI) turnover (as well as direct injection of InsP3) reduce M-current. Protein kinase C activators have no effect however, indicating that modulation is Ca2+ independent in these cells40 (see also next paragraph}. A specific role for diacylglycerol (DAG) in M-channel regulation in toad gastric smooth muscle cells has been proposed by some authors 41 : In support of this, extracellular application of 1,2-dioctanoyl-sn-glycerol (DiCS, a synthetic DAG that is a potent activator of PKC), reversibly suppresses M-current in these cells41 . In this study, DiC8 suppressed endogenous and isoproterenol-induced M-current without altering the time course of M-current deactivation, suggesting that it acts by decreasing the number of channels available to be opened41 .

Inhibition of M-current by phorbol esters (as distinct from agonist pathways) 53-32-05: Non-selective activation of PKC by phorbol esters is known to suppress M-current in bullfrog sympathetic neurones. For example, PDBu (4-,B-phorbol 12,13-dibutyrate) irreversibly suppresses M-current in a concentration-dependent manner (Ki approx. 38 nM)34. Common treatments that induce PKC inhibition (e.g. pseudo-substrate peptides like PKCI 19-31),

l

e_n_t_ry_53

_

staurosporine and H-7 (1-(5-isoquinolinylsulphonyl)2-methylpiperazine, see note, below) all antagonize PDBu-mediated suppression of M-current34. Inhibitors such as these have variable effects in suppressing M-current coupled to endogenous receptor/transducer systems: For example, suppression of Mcurrent in bullfrog sympathetic neurones by muscarine and luteinizing hormone-releasing hormone (LHRH) is unaffected by PKCII9-31 and H-7, but is antagonized by staurosporine. Overall, this has suggested that suppression of M-current by agonists is probably not mediated by activation of PKC34 (see also previous paragraph). Notably, addition and subsequent removal of PDBu to M-current previously suppressed by muscarine prevents the action of PDBu (closing M-channels by voltage or Ba2+ block cannot prevent PDBu suppression). To reconcile these results, it has been further proposed that two 'interconvertible' populations of M-channels exist in bullfrog sympathetic neurones, one that is sensitive to both agonist and PDBu and another that can only be suppressed by agonist. This invokes a 'protective' effect of muscarine channel closure to subsequent effects of PDBu. Partial suppression of M-current (by low concentrations of muscarine) antagonize the response to PDBu, where the magnitude of 1M suppression is equivalent to that induced with PDBu alone34 . Notes: 1. Other studies have found that H-7 did not prevent responses to phorbol 12-myristate 13-acetate (PMA), suggesting a PKC-independent mechanism for PMA modulation42 . 2. PKCII9-31 peptide (an inhibitor of protein kinase C) inhibits phorbol ester and arachidonate-induced decreases of 1M .ng in NGI08-15 cells43 .

ELECTROPHYSIOLOGY

Activation M-current possesses a unique characteristic of sustained activation 53-33-01: M-current is voltage dependent and is slowly activated in the 'subthreshold' range for action potential initiation, Le. at potentials 'close to rest' or at hyperpolarized potentials (but generally, positive to -65 mV, see below). Due to its 'non-inactivating' nature, M-current has been frequently described as 'persistent' at slightly depolarized membrane potentials. Thus, hyperpolarization induced by 1M is frequently described as 'stabilizing' cell excitability (in the absence of receptor agonists that induce its suppression) by contributing most of the sustained membrane current in the 'subthreshold' range (for details, see Phenotypic expression, 53-14 and Receptor/transducer interactions, 53-49). In general, the rate of KM opening is relatively slow for small depolarizations (e.g. to -40 or -30mV) but may be faster for large depolarizations occurring during action potentials t or strong excitatory post-synaptic potentials t (see also Phenotypic expression, 53-14).

Distinctions between 1M, 1K and 1A 53-33-02: In addition to its sustained activation (see previous paragraph) Mcurrent in rat sympathetic neurones can be distinguished from other potassium currents by four features. (i) Its activation range (V 1/ 2 == -45 mV)44 is hyperpolarized compared to the delayed rectifier (V 1/ 2 == -6mV)45 and the

_'---

e_n_try_5_3----J

fast-inactivating A-current (V1/ 2 = -30mV)46 in the same preparation. (ii) Activation/deactivation kinetics are 'substantially slower' for KM than for other voltage-activated potassium currents (e.g. deactivation rate constants are approx. 10 times slower at equivalent potentials47). (iii) Although Mchannels have similar selectivity characteristics to other voltage-gated K+ channels, other pore properties, for example the lack of anomalous molefraction behaviourt and the effects of external Ba2+ ions are unusual (see ref. 48, summarized under Selectivity, 53-40 and Blockers1 53-43). (iv) Mchannels also exhibit 'hallmark' patterns of modal gating I proposed to be regulated in part by calcineurin (see Protein phosphorylation, 53-32 and Single-channel data, 53-41) and receptor-coupled modulation (see Receptor/ transducer interactions, 53-49). Other distinctions between M-channel and other K+ channel pharmacology are described under Blockers, 53-43. See also the Methodological note describing M-current 'isolation' protocols employed when 1M is co-expressed with 1K and 1A (Blockers, 53-43).

Different external cations do not affect deactivation kinetics of 1M 53-33-03: With the exception of Na+, the nature of the permeant ion appears to have little effect on the gating properties of M-channels, i.e. the voltagedependence of the deactivation time constant t is similar when recorded in 15 mM K+, TI+, Rb+ andNHt48 (see also Selectivity, 53-40). The time constant decreases e-fold for a 60-80mV hyperpolarization48,49. Comparative note: Deactivation kinetics of cloned Shaker channels50 and several native delayed rectifiers are slowed when Rb+ is the permeating ion, an effect proposed to be due to impedance of channel closing when the pore contains an ion51 . In M-channels, ions other than K+ apparently have longer pore residency times, but do not impede channel closing48 .

Current-voltage relation 53-35-01: 1M channels generally display non-rectifying 1- V relations.

Inactivation 53-37-01: M-current is characteristically non-inactivating, although it does undergo slow calcium-dependent 'rundown', possibly mediated by calcineurin (see Rundown, 53-39). Sustained M-current ultimately depends on the period of time that M-channels reside in the long open time, high open probability mode 2 (for details, see Protein phosphorylation, 53-32 and Single channel data, 53-41).

Kinetic model Open and shut states of the KM are voltage sensitive but only shut states are muscarine sensitive

53-38-01: A kinetic analysis identifying M-channel states t sensitive to muscarine and membrane potential has been presented for KM channels in the cell-attachedt, dissociated rat superior cervical ganglion neurone preparation52. In this analysis, M-channel activities recorded at 30 mV positive

1'--_e_n_t_ry_53

_

to the resting membrane potential level (-60 m V) show three shut times t (7s 1 == 8.0 ± 2.2ms; 7 s2 == 71.3 ± 8.6ms and 7s3 == 740 ± 220ms) and two open times t (70 1 == 10.6 ± 1.9ms and 7 0 2 == 59.3 ± 8.7ms). When bursts t of Mchannel openings were determined as those including 7 s 1, two exponential components are evident in burst duration distributions (7bl == 11.0 ± 0.9 ms and 7b2 == 80.4 ± 11.0 ms). Membrane hyperpolarization significantly lengthens all three shut times and shortens both open times. Hyperpolarization also enhances the relative contribution of lhigh-conductance M-channels' (see Single-channel data, 53-41) and decreases the relative contribution of low-conductance M-channels to overall activity. Application of 10 ~M muscarine outside the patch (see Receptor/transducer interactions, 53-49) lengthens all three M-channel shut times without significantly affecting their open times 52 . Note: A kinetic model which permits interpretation of how the distribution of M-channel modal gating may be affected by both phosphorylation and dephosphorylation is presented in ref. 25 (see also Fig. 1 and associated text under Protein phosphorylation, 53-32).

Models for 'over-recovery' phenomena 53-38-02: A model has appeared fitting observed relations between extent and rate of transient enhancement or 'over-recovery' of 1M responses following agonist removal in bullfrog sympathetic neurones 53 (see Channel modulation, 53-44). This equilibrium model can be extended to account for actions of arachidonic acid metabolism inhibitors in this preparation (ibid.).

M-current kinetics can be modelled to produce resonant behaviour in neurones 53-38-03: Neurones are able to generate voltage signals at a certain frequency (as a result of the subthreshold oscillations) and to 'preferentially respond' to inputs arriving at the same frequency (resonance behaviour characterized by a peak in impedance magnitude)54. To aid understanding of subthreshold behaviour and resonance, an isopotential membrane model of guinea-pig cortical neurones was developed and compared to experimental observations 54 . The model consists of a leak current, a fast 'persistent' sodium current and a slow non-inactivating potassium current. The kinetics of the M-current and membrane capacitance are sufficient to produce both voltage oscillations and resonant behaviour. The kinetics of the K+ current by itself is sufficient to produce resonance behaviour (typically observed at depolarized levels). Within the model, the Na+ current amplifies the peak impedance magnitude and is essential for the generation of subthreshold oscillation54.

Kinetics of G protein-mediated modulation of M-current 53-38-04: The time course of 1M suppression and recovery has been shown to 'closely follow' the agonist-dependent kinetics of nucleotide interaction with G proteins, thereby suggesting that subsequent messengers were unnecessary55. In this analysis, M-current inhibition rates displayed an exponential time course, 'hyperbolically dependent' on GTP analogue concentration with a limiting value of 0.53 min-I. Muscarine induces a 'concentration-dependent acceleration' of the rate of nucleotide-induced inhibition, with a plateau of about 20 min- 1 and an exponential time course.

_"--

e_nt_ry_5_3_

In control (untreated) neurones, 1M recovery rate (following agonist removal) is approx.3-7min- 1 •

Rundown M-current rundown resulting from activation of a Ca 2+ -dependent phosphatase 53-39-01: In voltage-clamped bullirog sympathetic neurones recorded using a deactivating protocol t (see Blockers, 53-43) M-current reactivation (within the first minute of clamping to -38 mV) is observed as an increase in current amplitude ('run_up/)25. This run-up is calcium dependent (see note 1). During 20 min continuous recording periods, this augmented M-current can be observed to run-down25 . Inclusion of CaN420, a truncated, pre-activated form of calcineurin in the whole-cell pipette solution (for background, see Protein phosphorylation, 53-32) inhibits M-current to a much greater extent (rv 55%) after 20 min recording than rundown observed in control cells. This effect of CaN420 can be blocked by adding the 25 aa calcineurin inhibitory peptide (CaN2S ) corresponding to the autoinhibitory domain of wild-type calcineurin. Furthermore, inclusion of CaN2S or the calcineurin inhibitor cyclosporin A results in (i) greater augmentation of M-current during the first minutes of recording and (ii) a reduction in the amount of current lost during rundown. These and other results 25 suggest that rundown results from activation of endogenous calcineurin. Notes: 1. 'Run-up' is absent in cells bathed in Ca2+free/Mn 2+ replacement Ringer solution25 . 2. Voltage-clamp of neurones at -38 mV causes an accumulation of intracellular calcium19,25 due to sustained calcium channel activation57 which may modulate M-current. 3. A mildly acidic intracellular pH reduces 'rundown' of M-current31 . For further details on Phosphorylation/dephosphorylation cycles affecting M-current activity, see Protein phosphorylation, 53-32.

ATP-,-S substantially reduces M-current 53-39-02: Under whole-cell recording conditions, the steady-state 1M in frog sympathetic ganglion cells is maintained 'for at least 20 min/ when the patch pipette contains neither ATP nor C~9 (compare ATP additions reversing the effect of CaN420 under Protein phosphorylation, 53-32). This study39 further reported that inclusion of ATP or cAMP or the ATP antagonist, ,8,,-methyleneATP (rv l-2nM) failed to alter the rate of 1M 'run down/. By contrast, inclusion of ATP..,-S (1 or 2lllM) resulted in a rv60% reduction of the current within 18 min. Despite the inability of ATP-,-S to maintain steady-state 1M , it had no effect on the ability of muscarine (2-100 IlM) to suppress a 'constant fraction' of the available current39. Independent studies of frog sympathetic neurones have shown that stimulation of G proteins with high GTP-,-S/GTP ratios (inducing a net accumulation of the active G* state) can produce faster rates of 1M inhibition (see also Channel modulation, 53-44).

ATP hydrolysis is required for desensitization of M-channel responses to substance P

53-39-03: Desensitizationt of M-channel responses to the agonist substance P (SP) has been studied in dissociated sympathetic neurones from bullfrog58 .

E

l_e_n_t_ry_53

_

When ATP (in the recording pipette) is replaced with AMP-PNP, SP still inhibits 1M , but no desensitization is observed, indicating that ATP hydrolysis is required for desensitization (see also notes 1 and 2). When low doses of muscarine (sufficient to inhibit 1M , but not to elicit desensitization), are applied simultaneously with a 'desensitizing dose' of SP, 1M remains depressed and does not desensitize 58. Notes: 1. Classical neurotransmitter receptor protein desensitization t may also account for 1M desensitization of inhibition. Desensitization mechanisms may thus be specific for different agonists. 2. Hypothetically, the enzyme(s) which mediate different forms of desensitization may be 'compartmentalized' in cells58 .

Selectivity Permeation versus conduction properties of M-channels 53-40-01: In rat superior cervical ganglion neurones, reversal potential t (Erev ) measurements for macroscopic 1M yield a relative ionic permeabilityt (Px/ PK ) series of Tl+ (1.5»K+ (1.0»Rb+ (0.8»Cs+ (0.2»NHt (O.l»>Na+ (0.008)48, under bi-ionic t conditions, with 150mM internal and 15mM external monovalent cation. Although the permeability sequence is very similar to other K+ -selective channels (implying a high degree of conservation in the selectivity filter) M-channels show a number of features atypical of potassium channels, as summarized in Table 1. Relative conductance t measurements estimated by the slope of instantaneous current-voltage relationships indicate a different selectivity sequence to that measured by Erev : gX/gK in 15 mM external cation was K+ (1.00) > Tl+ (0.28) > NHt (0.23) > Rb+ (0.11) > Cs+ (0.10). Collectively, these may be indicative of a distinct pore structure outside of the selectivity filter as compared to previously cloned K+ channels 48 (see also Table 1). Comparative note: In general terms, other K+ channels, both voltage-dependent and inward rectifiers ('native' and 'cloned') pass another ion (Tl+ or NHt) better than K+.

Single-channel data Single M-channels display non-stationary kinetic behaviour in sympathetic neurones 53-41-01: In bullfrog neurones bathed in isotonic K+ solutionst, KM displays conductance levels of approx. lOpS and 15pS where each level exhibits similar modal gating t (for details, see paragraph 53-41-02 and Table 2). Other estimates of KM channel conductances (reviewed in ref.28) exhibit some variance. For example (i) an M-channel conductance of 1-3pS was estimated by noise analysist using physiological Ringer'st solution5,47,66 and (ii) sustained, depolarization-activated M-current with higher conductances was reported in cell-attachedt recordings of dissociated rat superior cervical ganglion neurones (three main levels of open-channel conductance t approx. 7pS, 12pS and 19p5 in physiological Ringer's)52 (see also Kinetic model, 53-38). It has been noted, however, that under conditions where Goldman rectificationt is removed (i.e. by recording macroscopict M-current in isotonic K+ solutions) M-channel conductances are increased by four- to sixfold, yielding an estimated unitary conductance of between 8 and 18 pS

_'---

e_n_try_5_3_

Table 1. IAtypical' conduction properties of M-channels in rat sympathetic neurones48 (From 53-40-01) Feature

Summary/comparison with other K+ -selective channels (see note 1)

Conduction of monovalent cations relative to K+ is Ivery low'

53-40-02: All permeant monovalent cations are 'much less conductive' than K+ through the M-channel (see paragraph 53-40-01). These results appear consistent with the M-channel pore possessing a large and non-selective outer vestibule (Le. like other K+ channels). In M-channels, however, only K+ ions may be able to permeate quickly, with other ions being relatively 'retarded'48.

The nature of the permeant ion does not affect deactivation kinetics

53-40-03: The time course of M-current deactivationt induced by hyperpolarization is unaffected by different external cations. In comparison, several potassium currents exhibit a slowing of deactivation when Rb+ is used instead of K+ as the permeant ion (e.g. Shaker B ~6-4650, inward rectifiers in bovine pulmonary artery endothelial cells59 and other native K+ channels60 ).

M-current does not exhibit anomalous mole fraction behaviour

53-40-04: K+ channels exhibit an anomalous mole fraction t effect, suggesting that they are multi-ion pores t. In two independent studies of KM channels (in bullfrog sympathetic ganglion cells61 and rat sympathetic neurones 48 ) no evidence for anomalous mole fraction behaviour is apparent. In the latter study, titration of K+ with increasing fractions of Rb+ causes the relative conductance to decrease, but it does not exhibit a conductance minimum as expected for anomalous mole fraction behaviour. The absence of the expected increase in conductance (Le. upon complete replacement of one ion for another at some point in the titration) may be related to the low conduction of ions other than K+ thorugh KM channels (this table).

IPreferential' block of 53-40-05: M-current is blocked by low concentrations of Ba2 + ions 62,63 acting at external outward current by sites1 and exhibiting a 'preferential' block for external Ba 2+ ions (mechanisms unclear) outward current48 . The mechanism(s) of Ba2+ block of M-channels appears distinct from Ba2+ block in calcium-activated K+ channels, K+ -selective inward rectifiers and other voltage-gated K+ channels (for details see comparative notes under Blockers, 53-43).

entry53 I-

_ - - - - -

Table 1. Continued Feature

Summary/comparison with other K+ -selective channels (see note 1)

Cs+ and Na+ permeability (comparative note)

53-40-06: KM channels are 'moderately permeable' to

Predicted conservation of selectivity filter region (comparative note)

Cs+ ions and are 'slightly permeable' to Na+ ions (see paragraph 53-40-01) but are 'essentially impermeable' to NMDG. In comparison, Cs+ and Na+ are not permeant in (i) skeletal muscle BKca channels (although Cs+ is permeable to SKca channels in chromaffin cells - see ILC K Ca, entry 27); (ii) delayed rectifiers in myelinated nerve or T lymphocytes (see VLC K DR [native], entry 45) or (iii) starfish egg inward rectifier channels (see Selectivity under INR K [native], 32-40). 53-40-07: KM channels share the permeation sequence K+ > Rb+ > NHt > Cs+ with several cloned K+ channels belonging to distinct gene families (for example Shaker64, see VLC K Kv1-Shak, entry 48) and the renal K+ channel ROMK2 65 (see INR K [subunits], entry 33). Although these cloned channel types have different gating mechanisms and physiological functions, their conservation of the TXGYG motift (see Selectivity under VLC K Kv1-Shak, 48-40) might also predict that KM channels possess this motif.

Notes: 1. See also properties of novel 'non-inactivating' K+ channels cloned from Aplysia under Miscellaneous information, 53-55. (W. Gruner, cited in ref. 28). A 21 pS channel conductance with similar properties to KM recorded in bullfrog sympathetic neurones (ibid.) is consistent with conductance values extrapolated from noise analysis and the quantitative descriptions in ref. 28 . See also influence of [Ca 2+h on switching of gating modes (cross-referenced from paragraph 53-41-03, below) which may also help to account for the difficulties in resolving KM activities between studies.

Control of modal gating as a mechanism for M-current neuromodulation 53-41-02: Evidence for selective reduction of a single modet of M-channel gating following application of muscarine has been reported28 . M-channels in dissociated bullfrog sympathetic neurones can be resolved in the cellattachedt configuration into two conductance states, which exhibit appropriate voltage-dependent kinetics and two modes of gating: mode 1 and mode 2. 'Mode I' openings comprise short open time, low open probability events; open time behaviour of mode 1 has no apparent voltage dependence.

II

Table 2. Summary of properties derived from single-channel recordings of M-channels in bullfrog sympathetic neurones (From 53-41-01)

Property/characteristic/preparation

Description

Illustration

Amplitude classes NB Cell-attached configuration, symmetrical K+ (see notes 1 and 2)

Two amplitude classes of channels can be recorded with properties consistent with macroscopic M-current (see panel (a)). At hyperpolarized potentials (e.g. - 70 mV) fast channel events of two amplitudes are discernible. Membrane depolarization (e.g. to -30mV) evokes longer duration openings of both amplitude classes.

(a)

-30~

~.~~I,~~~r'J"

~M~ILil ~

,~' '~'Ir~~o

Il~ 40ms

Unitary 1- V relations and estimated conductances of the two amplitude classes in symmetrical K+ (see notes 1 and 4) Channel open probability versus membrane potential

Panel (b) is the 1-V relation for the same patch as in panel (a). For n == 10 records (mean ± SD), , == 14.7 ± 3.2 pS and 10 ± 1.5pS. Membrane depolarization increases the frequency of channel opening (Popen ) from 0.03 (at -70 m V) to 0.3 (at -30 m V).

(b) -1 ()()

V (mV) -80

·60

·40

·20

-200

~ Q)

·400

~

"'C

~

a. E

g

-800

C Q)

~

-1000

~

CJ1

·600

ctS

(J

·1200

CJ,J

Ensemble t current characteristics

Open state kinetics for M-channel conductance classes

Holding a cell-attached patch (exhibiting both conductance classes) at -80 mV and stepping to -50mV (500ms pulse duration every 5 s) produces a time-dependent increased inward current as the channel opens slowly during pulses (see note 3).

Open time analysis (all events, not shown) in amplitude and open duration histograms could be best described by the sum of two exponentials with time constants (TO of 9.2 and 3.1 ms)

("t)

I:' t'+

~ (Jl

UJ

~I

200 ms

Exclusion of either conductance class (lOpS or 16pS, see above) from this analysis results in distributions identical to those seen with all events, indicating the two conductance classes have identical open state kinetics (see also note 4).

Notes: 1. Pipette solution was 90mM KMeS04' 20mM KCI, 5mM MgCI2, 0.1 mM CaCI2, 5mM Na-HEPES (pH 7.2); cells perfused in this solution have an m.p. of OmV. 2. Openings are downward deflections in single-channel records. 3. Deactivation rates of these ensemble currents resemble M-current, increasing e-fold in 22mV of hyperpolarization, from 56.2± 15.8ms at -60mV to 8.7±3.4ms at -80mV (as compared with e-fold in 23mV for macroscopic M-current 5 ). 4. The lower conductance openings (lOpS, see above) and the higher-conductance openings (16pS, see above) are consistent with subconductance states of the M-channel (i.e they are unlikely to represent different channels28 ). 5. Data and figures reproduced with permission from Marrion (1993) Neuron 11: 77-84.

11

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_t_ry_5_ 3

---l

'Mode 2' openings represent long open time, high open probability behaviour with open time and voltage dependence appropriate for that underlying Mcurrent deactivation5,28. Significantly, muscarine decreases M-channel activity by selective reduction of mode 2 through an unidentified Idiffusible' second messenger (see Protein phosphorylation, 53-32 but also Channel modulation, 53-44). Consistent with a diffusible messenger, KM channel activity can be reduced by application of agonists distal from patches of membrane containing K M channels. Notes: 1. In general, K M resides in both modes with the time spent in each varying between patches. When M-current is suppressed, its macroscopic kinetics do not change, consistent with the effect of muscarine being on the opening rate of the channel and not on the closing rate. The action of muscarine in reducing high Popen modal activity (by affecting the opening rate, see above) has the consequence that the closed times become longer. 2. Loss of mode 2 M-channel activity occurs within 1 min of patch excision25 (see also Rundown, 53-39). Application of ATP (lmM) can sometimes restore a long Popen state, but this state is distinFurther descriptions of M-channel guishable from mode 2 behaviou~5. unitary behaviour and gating modes are summarized and illustrated in Table 2. For the calcium-dependent phosphorylation/dephosphorylation mechanisms underlying transitions between gating modes, see Fig. 1 and associated text under Protein phosphorylation, 53-32.

Comparative notes: Modal gating in other channels

53-41-03: As discussed in ref. 28, several aspects of M-channel modal behaviour have similarities to both N- and L-type calcium currents (for further refs, see VLG Ca, entry 42). Reduction of mode 2 M-channel activity by muscarine (paragraph 53-41-02) is reminiscent of suppression of neuronal (N-)type Ca2+ current by noradrenaline, where similar reductions in long To, high Popen states have been described67. Modal gating changes can also account for the observed facilitation of N-type Ca2+ current that is evoked following removal of agonist 68 (compare M-channel lover-recovery,19 under Phenotypic expression, 53-14). Augmentation of cardiac (L-)type Ca2+ current by isoproterenol occurs by potentiation of high-activity (long To, high Popen ) gating modes of the L-channe169 . In comparison, augmentation of K M channel activities following removal of agonist19 can be mimicked by raising intracellular calcium levels 70 (since 'switching' of gating modes is influenced by [Ca2+h, 'low intracellular Ca2+' conditions may account for difficulties in maintaining the K M signal transduction pathway (for significance, see Fig. 1 under Protein phosphorylation, 53-32). Further examples of modal gating have been documented for (i) glutamate receptor-channels in locust muscle 71 (for background see the ELG CAT GLU series, entries 07 and DB); (ii) nicotinic acetylcholine receptorchannels 72 (see ELG CAT nAChR, entry 09); (iii) intracellular Ca2+-activated K+ channels (see ILG K Ca, entry 27); voltage-gated Na+ channels 74 (see VLG Na, entry 55) and possibly the S-channel l in sensory neurones of Aplysia 75 (see Miscellaneous information, 53-55). The functional diversity of these channel types suggests that modifications of modal gating behaviour by biological ligands is a significant mechanism for neuromodulation of channel function in vivo.

l"---e_n_t_ry_5_3

_

PHARMACOLOGY

Blockers M-channels are resistant to most tclassical' K+ channel blockers 53-43-01: There are no well-characterized, highly selective blockers available for M-channels. When applied to extracellular or electrode solutions at the stated concentrations, M-channel currents are 'relatively unaffected' by the blockers 4-aminopyridine (4-AP, 1 mM), caesium ions (2 mM), gadolinium ions (0.1 mM), 3,4-diaminopyridine (1 mM), zinc (0.1 mM) and d-tubocurarine (for further details, see ref.28 and Brown (1988) under Related sources and reviews, 53-56).

Effects of external Ba 2+ on M-current

53-43-02: M-channels are sensitive to block by external barium ions 62 (e.g. K i == 300 J.lM in sympathetic ganglion cells)63. Notably, external Ba2+ shows a preferential block of outward current48 : Although the muscarinic agonist oxotremorine-M (10 J.lM, see Receptor/transducer interactions, b3-49) has equal effects on both activation and deactivation relaxations (approx. 500/0), Ba2+ ions (1 mM) inhibit inward M-current by approx. 150/0 and outward M-current by approx. 500/0 48 (as illustrated in Fig. 2). Ba2+ block of M-channels does not appear to be voltage sensitive and does not significantly alter gating kinetics 48 . '1M -like' responses in some preparations are also weakly and non-selectively 'blocked' by Ba2+ ions (see the following paragraph).

Ba 2+ inhibition of the tIM-like' component I Kx in rod photoreceptor inner segments 53-43-03: The voltage-dependent K+ current I Kx in the inner segment of salamander rod photoreceptors (see Phenotypic expression, 53-14) is inhibited by external Ba2+ ions. Ba2 + ions are likely to interact with Kx channels at multiple sites, as indicated by distinguishable effects on (i) tail current voltage sensitivitYi (ii) voltage dependence and (iii) conductance properties2o . In (i), reductions in the voltage sensitivity of I Kx tail currents are induced by Ba2+ ions (with a Ko.s approx. 0.2 mM). In (ii), Ba2+ ions shift the voltage dependence over which I Kx appears (to more positive potentials, Ko. s approx. 2.4mM, see note 4). In high [K+]o (100mM) the voltage range of activation of I Kx is shifted 20 mV negative, as is the tau-voltage relation. In (iii), Ba2+ ions do reduce I Kx conductance but at low sensitivity (Ko. s approx. 76 mMi 'high [K+]o' does not prevent this effect, but does abolish barium's ability to affect voltage dependence and voltage sensitivity)2o. Notes: 1. Ca2+, Co 2+, Mn2+, Sr2+ and Zn2+ ions do not show comparable actions to Ba2+ on the voltage dependence or the voltage sensitivity of I Kx tail currents. 2. Ba2+ ions also alter apparent time courses of activation and deactivation of I Kx channels in a concentration-dependent manner, consistent with a slow but steeply voltage-dependent blocking and unblocking action. 3. These data can be accommodated in models which assume that Ba2+ has a voltageindependent and a voltage-dependent blocking action on open or closed I Kx , with the voltage-dependent component accounting for both (i) the reduction

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_t_ry_5_3_

(a)

Ba

OxoM

=-------- .

---------. ~-------v

~~_-----o

1200 PA

1200 pA lOOms

(b)

lOOms

Outward current

Inward current 60

.g=

50

j

40

u

I

30

~

20

u

&

10

WOXOM

~OXOM

Ba

N4yJBa i

-120 -100

-80

-60

-40

i

i

-20

0

voltage (mY) Figure 2. Comparison of M-current suppression by external Ba 2+ ions and the muscarinic agonist oxotremorine M in rat sympathetic neurones. (a) Mcurrent deactivations were evoked by a 1 s voltage step from -70 to

-30mV Activation was elicited by 30mV depolarizations from -50mV (for full recording conditions suppressing IA and IK, see ref.48 and methodological note under Activation, 53-33). M-current suppression following application of external Ba2 + ions (1 mM, is larger for inward current (0) compared to outward current (e). Ba 2+ reduces the instantaneous current with relatively little effect on the time-dependent relaxations evoked by negative voltage steps. In comparison, oxotremorine M (10mM) has approximately equal effects on inward current (V) and outward current (~). (b) Effect of external Ba 2+ ions and oxotremorine M on M-current relaxations evoked by the protocol (avel'aged from 4 or 5 cells for each data point). (Reproduced with permission from Cloues and Marrion (1996) Biophys J 70: 806-12 (From 53-43-02).

in voltage sensitivity of IKx tail currents and (ii) the Ba2+ -induced shift in voltage dependence20 . 4. The strong 'positive shift' induced in the activation range of IKx by Ba2+ (and the absence of muscarinic suppression) distinguish IKx from I M 21 (for further comparison and roles of llO(, see Phenotypic expression, 53-14).

-----J_

lL....--e_n_t_ry_53

Inhibition properties of the 1M-like

IKM,ng

53-43-04: The current designated IKM,ng in differentiated NGI08-15 (mouse neuroblastoma x rat glioma hybrid) cells has been shown to be inhibited by external divalent cations, comprising both depression of the maximum conductance and a positive shift of the activation curve1. Inhibition by these ions persists in Ca2+-free solutions and addition of Ca2+ (10 J.1M free [Ca2+]) or Ba2 + (1 mM total [Ba2+]) to the pipette solution under whole-cell patchclamp does not significantly change IKM,ng (for significance, see Table 3 under Channel modulation, 53-44 and this field, above). In this study, cation inhibition of IKM,ng showed the following decreasing order of potency (with millimolar IC so values as denoted): Zn2+ (0.011) > Cu2+ (0.018) > Cd2+ (0.070) > Ni2+ (0.44) > Ba2+ (0.47) > Fe2+ (0.69) > Mn2+ (0.86) > Co2+ (0.92) > Ca2+ (5.6) > Mg2+ (16) > Sr2+ (33). La3+ ions do not inhibit IKM,ng at concentrations which inhibited lea in the same cells and organic Ca2+ channel blockers are ineffective. IKM,ng is reduced by 9-amino-1,2,3,4-tetrahydroacridine (ICso 8 J.1M) and quinine (30 J.1M) but is insensitive to tetraethylammonium (IC so > 30 mM), 4-aminopyridine (> 10 mM), apamin (>3 J.1M) or dendrotoxin (>100nM)1. See also Receptor/transducer interactions, 53-49.

Weak KM-inhibiting activity of an SKca blocker 53-43-05: Dequalinium, a bisquaternary compound used as a non-peptide SKca blocker (see Blockers under ILC K Ca, 27-43), produces a 18% inhibition (at 10 J.1M) of the M-current in rat sympathetic neurones 76 .

Methodological note: M-current tisolation' protocols 53-43-06: In native cells, M-current is frequently co-expressed with other voltage-dependent K+ currents, particularly the delayed rectifier and Acurrent types. Deactivating protocols in combination with blockers can be employed to isolate M-current components (see also legend to Fig. 3 and Rundown, 53-39). For example, whole-cell voltage-clampt can be applied at depolarized levels (e.g. -30 to -38 mY, where most other voltage-gated K+ currents are inactivated); M-current can then be revealed by its deactivation t (relaxation t ) following hyperpolarizing steps (typically in 10mV increments of 1 s duration) which produce slow increases in outward current. 'Contaminating' delayed rectifier components are often blocked with tetraethylammonium ions (TEA+, 10mM, see Blockers under VLC K DR [native], 45-43); except where TI+ is the external cation (Ki = 1 mM48 ). 4-Aminopyridine (4-AP, 10 mM, see Blockers under VLC K A-T [native], 44-43) may also be included to reduce both A-current46 and delayed rectifier components (Ki = 0.3 mM48 ). Notes: I. Tail current t through KM channels is sensitive to oxotremorine M (see Table 4 under Receptor/transducer interactions, 53-49). Generally, oxotremorine M-sensitive components should display the same deactivation kinetics t as non-subtracted tail currents. 2. Since M-current is suppressed by muscarinic agonists, most biophysical parameters are measured in their absence. 3. Preliminary evidence for expression of M-channels can be obtained by monitoring membrane potential (i.e. observing agonist-induced depolarizations that fail to reverse when the membrane is hyperpolarized). However it should be noted that this behaviour may also be expected from a decrease in a Kca conductance (in which the

II

_'---

.

e_n_try_5_3_

voltage-sensitivity results from the voltage sensitivity of the 'persistent' or 'activating' calcium current). Thus voltage recordings alone cannot distinguish between 1M suppression and general reductions in 'background' K+ conductances coupled with increases in cation conductance 77 (see also ILG K Ca, entry 27 and the KM/Kca comparative note under Phenotypic expression, 53-14).

Distinguishing

K1eak

and K M responses in pre-vertebral neurones

53-43-07: Muscarinic responses in guinea-pig coeliac neurones are produced by suppression of two K+ currents: the M-current and a muscarine-sensitive leak current. These components can be distinguished by their differential sensitivity to Cs+ and Ba2+ ions. Briefly, barium (2 mM) reduces the M-current and the leak K+ current, whereas caesium (2 mM) reduces the M-current but does not affect leak current (for further details, see ref, 78).

Channel modulation For collated evidence that M-current suppression involves diffusible messenger(s) see also Protein phosphorylation, 53-32, Rundown, 53-39 and Receptor/transducer interactions, 53-49.

[Ca 2+ Ji modulation of K M responses (cross-references)

53-44-01: Receptor agonists that suppress the M-current in bullfrog 7o,79 (but not rat80 ) sympathetic neurones simultaneously increase the intracellular calcium concentration (for background and further cross-references see Table 3, this field and Receptor/transducer interactions, 53-49). The effects of [Ca2 + h chelation of intracellular [Ca2 +- h following receptor stimulation can block the occurrence of 'over-recovery' (ref. 70, see Phenotypic expression, 53-14 and Fig. 3 and associated text under Protein phosphorylation, 53-32). Variable effects on 1M observed at different [Ca2+h may be able to be reconciled by considering alternative sites of protein modulation, e.g. (i) on the channel itself; (ii) on associated transducer or effector molecules; (iii) by coinvolvement of additional messenger molecules or (iv) a combination of these. Higher levels of [Ca2 +Ji may activate membrane-bound phosphatase activities associated with KM channels and responsible for selective reductions in mode 2 gating behaviour (ibid.). Low intracellular Ca2+ conditions may account for difficulties in maintaining the KM signal transduction pathway (see discussion

in paragraph 53-41-03).

Common

tIM

suppression-uncoupling' effects of decreased cellular

NAD+ 53-44-02: NGl08-l5 cells pre-treated for 5-15 h with 5 mM streptozotocin can prevent (or 'uncouple') suppression of M-current induced by several G proteinlinked receptor agonists 81 . This effect can be shown to prevent suppression of 1M following bath application of ATP (100 JlM), bradykinin (lOnM), angiotensin II (lOOnM), endothelin 1 (lOOnM) and acetylcholine (lOJlM, in an NGl08-l5 cell line stably expressing M1-muscarinic receptors). Notably, suppression properties through all receptor types are restored by simultaneous incubation with nicotinamide (5 mM), suggesting that signal transduction

II.-._e_n_t _ry_53

_

from these five different receptors to M-channels shares a common pathway requiring NAD+ 81 .

Arachidonate and its metabolites may non-selectively modulate K(M) responses 53-44-03: Modulation of M-current responses by arachidonic acid metabolites has been characterized in several preparations, including bullfrog sympathetic neurones, hippocampal neurones and NGI08-15 neuroblastoma x glioma hybrid cells (see paragraphs below). The multiplicity of proteins susceptible to lipophilic 'modulation' (described in ILG Ca AA-LTC4 [native), entry 15, and ILG K AA [native}, entry 26) can often 'confound' analysis of these responses. In consequence, some of the literature is confusing or inconsistent and descriptions below are limited to the 'overall' reported effects on KM-like responses (i.e. some observations may not have been 'reproducible' between independent studies and may even be considered 'epiphenomena' by some authors. Please use 'feedback' (field 57) to clarify and update 'consensus' properties wherever possible.

Evidence for leukotriene C4 augmentation of M-current in hippocampus 53-44-04: In rat hippocampal pyramidal neurones, somatostatins (SS, somatotatin-14 and -28) partly exert hyperpolarizing effects by augmentation of muscarinic-inhibited channels, a process that can be mimicked in slice preparations by exogenous application of arachidonate82. The action of both arachidonate and SS can be blocked by (i) the general lipoxygenase inhibitor nordihydroguaiaretic acid (see note 4) or (ii) the specific 5-lipoxygenase inhibitors 5,6-methanoleukotriene ~ methylester and 5,6-dehydroarachidonic acid. The 1M -augmentation response is insensitive to (i) the cyclooxygenase inhibitor indomethacin; (ii) PGE2; (iii) PGF2a; (iv) PGI2; (v) the 12-lipoxygenase product 12-hydroperoxyeicosatetraenoic acid or (vi) the 12-lipoxygenase inhibitor baicalein, suggesting the involvement of 5-lipoxygenase metabolite(s) (probably leukotriene C4 ) as a mediator in the process82,83 (see also ILG Ca LTC4 [native), entry 15). Notes: 1. The studies reported in ref. 82 also found evidence for a 'direct' activating effect of arachidonate on a hyperpolarizing K+ channel distinct from KM (see also ILG K AA [native), entry 26), which was most effective at at membrane potentials 'near rest' (as compared to LTC4 modulation of KM , which was most effective at 'slightly depolarized' potentials). 2. In ref. 83, the 1M -augmenting effects of somatostatin were abolished by the phospholipase A2 inhibitors quinacrine and 4-bromophenacyl bromide. 3. The augmenting/ inhibiting effects of somatostatin/acetylcholine have been taken as evidence for lreciprocal' regulation of KM by two different receptor/transducer systems in hippocampal neurones. 4. Nordihydroguaiaretic acid can also prevent 1M enhancement by arachidonate in bullfrog sympathetic neurones 53,54 (see also following paragraphs).

Prevention of somatostatin-induced 1M augmentation by a FLAP inhibitor 53-44-05: Comparative note: 5-Lipoxygenase-activating protein (FLAP, an

18 kDa integral membrane protein required, in peripheral cells, for the

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_t_ry_5_3_

activation of 5-lipoxygenase, see paragraph 53-44-04) is expressed in various regions of the rat brain, including hippocampus, cerebellum, primary olfactory cortex, superficial neocortex, thalamus, hypothalamus and brainstem. Highest levels of expression are observed in cerebellum and hippocampus, and co-localization of 5-lipoxygenase and FLAP has been shown in CAl pyramidal neurones 85 . Inhibition of FLAP with the compound MK-886 (0.25-1 JlM) can prevent the somatostatin-induced augmentation of the hippocampal K+ M-current85 .

Proposed roles of arachidonate in M-current tover-recovery'in bullfrog sympathetic neurones 53-44-06: Exogenous application of arachidonate enhances 1M of whole-cell voltage-clamped bullfrog sympathetic neurones in a 'dose-dependent and reversible manner' (ICso rv2.8 JlM)53. Application of muscarine or LHRH in this preparation shows an initial phase of M-current reduction, followed by a transient enhancement or lover-recovery' when agonists are removed. Notably, the extent of 'over-recovery' increases with the extent of the preceding inhibition, while the rate and degree of inhibition increases with the concentration of agonist. By contrast, the rate of recovery and the extent of 'over-recovery' are independent. Inhibitors of phospholipase A2 (quinacrine and bromophenacyl bromide, BPB) and a lipoxygenase inhibitor (nordihydroguaiaretic acid) prevent over-recovery without significantly affecting the rate or extent of 1M inhibition (compare previous paragraph and subsequent proposals for Ca 2 + -dependent mechanisms of tover-recovery' as described under Protein phosphorylation, 53-32). Notes: 1. The 'inhibitory' process (typical half-life rv21 s) and 'enhancement l process (typical half-life rv53 s) as described above required fast solution exchange techniques to distinguish them kinetically. 2. Application of muscarine can also inhibit A-type K+ channels in several preparations (e.g. ref. 84, see also Channel modulation under VLC K A-T, 44-44). In the bullfrog sympathetic neurone preparation53, the effect of PLA2 inhibitors (above) appears to prevent over-recovery by extending the inhibitory process half-life to >80 s (compare note 1), and is inconsistent with 'simultaneous' 1M enhancement and l A inhibition (see also Kinetic model, 53-38). 3. Further work86 in bullfrog sympathetic neurones has suggested that intracellular Ca2+ ions enhance 1M by stimulating the arachidonate metabolic pathway86 (see also Channel modulation, 53-44). In the latter studies, dephosphorylation was interpreted to upregulate 1M , while 'over-recovery' was interpreted as a result of stimulation of the lipoxygenase pathway and phosphatases by increased [Ca2+h 86 •

Membrane potential of NG10B-15 cells (set principally by 1M ) is modulated by arachidonate 53-44-07: Arachidonic acid (AA) affects membrane potentialt (EM) in NGI0815 neuroblastoma x glioma hybrid cells) in a dose-dependent manner87. At relatively high concentrations (25-50 JlM), AA first increases and later decreases a current (designated as Ih, see note below) flowing mainly through open M-channels (which holds EM at -30mV). At lower concentrations (5-10 JlM) AA only decreases Ih. AA also accelerates the gating t kinetics t of 1M (for further descriptions, see ref. 87). AA may exert its effects

l_e_n_t_ry_53

_

following internal accumulation or by depletion of eicosanoidt products (for background, see ILG K AA [native], entry 26). Non-specific increases in membrane fluidityt by AA may also explain some effects on 1M gating t kineticst seen in this preparation. Notes: 1. The designation Ih is most often used to describe certain members of the hyperpolarization-activated current family whose collective properties are listed under INR K/Na IfhQ, entry 34.

Other tmodulators' of M-channel responses 53-44-08: The specific mechanism(s) of M-current modulation may be dependent on cell type, or may be conditional due to 'indirect' effects (e.g. dependent on selective co-expression of specific receptor/transducer components (see Receptor/transducer interactions, 53-49). With these caveats, a number of other (mostly non-specific) putative 'modulatory' effects on KM-like responses have been noted, which are summarized in Table 3.

Receptor/transducer interactions For further data suggesting that diffusible messengers are involved in Mcurrent suppression, see also Protein phosphorylation, 53-32 and Channel modulation, 53-44.

Cell-type-selective receptor/transducer control mechanisms of 1M 53-49-01: M-current suppression patterns have mostly been described as consistent with involvement of a diffusible second messenger (i.e. channel activity can be reduced by application of receptor agonists outside nonexcised membrane patches (see also Table 3 under Channel modulation, 5344). For discussion of the formal possibility that G proteins interact directly with M-channels, see refs. ss,99 and Kinetic model, 53-38. 'Direct' G protein interactions with effector t channels are characterized with relatively fast agonist response times (see, for example, the entry INR K G/ACh [native], entry 31). In this regard, it is pertinent to note the relatively slow agonist response times observed for many native KM channels and to cite evidence for involvement of 'extrinsic' proteins which are likely to modulate KM channels in situ (see Protein interactions, 53-31 and cross-references therein). In bullfrog neurones, receptor agonists that suppress the M-current simultaneously increase the intracellular calcium concentration 70,79 (for further background, see ILG Ca Ca RyR-Caf, entry 17 and ILG Ca InsPJ, entry 19). Effects of elevated [Ca2 +h in control of M-current are summarized under Protein phosphorylation, 53-32 and Channel modulation, 53-44.

tObligate dependence' of K M regulation on co-expressed protein components 53-49-02: Receptor/transducer control of M-current is a significant pathway for regulation of excitability in sympathetic neurones 17,100. For example, suppression of 1M by neurotransmitters such as acetylcholine result in membrane depolarization and increased input resistance (making cells more likely to fire action potentials) (for further details, see Phenotypic expression, 53-14). Regulation of KM gating in sympathetic neurones is closely associated with kinase and phosphatase effectorst, which themselves could be under

II

Table 3. M-current response modulators (mostly non-specific, not necessarily proving direct effects on the KM protein or pertinent to a physiological regulation of the channel - see Protein phosphorylation, 53-32 and Receptor/transducer interactions, 53-49) (From 53-44-08) Putative modulators

Description/cross-references

Effects of intracellular calcium on KM (see also effects of Ca2+-dependent calcineurin and MLCK under Protein phosphorylation, 53-32).

53-44-09: M-current modulation by intracellular Ca2 + has been described as 'biphasic', where small increases in [Ca2+h enhance 1M (SO-lS0nM, comparable to [Ca2+h produced during action potentials). By contrast, larger, sustained increases in [Ca2+h (>200nM) tend to reduce the sensitivity of muscarinic response and inhibit 1M (for further details see

refs 37,7o,88-90).

ATP-dependent and ATP-independent effects

53-44-10: Maintenance of 1M at 'resting' calcium levels, and the enhancement of 1M by 'modest' rises in [Ca2 +Ji both require ATP in whole-cell pipette solutions 19,9o (for likely significance, see Protein phosphorylation, 53-32). Reductions in M-channel open probability following application of Ca2+ ions to excised inside-out patches of rat sympathetic neurones have been described91 . This effect occurs in the absence of ATP, suggesting independence from protein phosphorylation and proposes the role of Ca2+ ions as a 'direct' inhibitor of M-channels following receptor activation. Perspective on 'direct binding' Note: Whether (i) 'direct' Ca 2 + modulation; (ii) the alternative interpretation, of Ca 2 + acting by changing activity of modulatory enzymes (e.g. kinases/phosphatases, see Protein versus 'indirect Ca 2 + modulation' of M-channel phosphorylation, 53-32) or (iii) both was able to account for all M-current modulations was effectors unclear at the time of compilation. Certainly 'direct inhibitory' effects occur at {Ca 2 +h much greater than that seen following application of receptor agonists. Furthermore, no calcium rise is seen in rat superior cervical ganglion (SCG) sympathetic neurones with agonists. However, it may be argued (see ref.91) that the effects of Ca 2 + are local to the plasma membrane, and not measurable in the bulk of the cytoplasm. Induced {Ca 2+Ji rises that inhibit M-current also activate BKca channels (see ILG K Ca, entry 27) but BKca activation is not normally observed with receptor agonists that modulate M-current. As described elsewhere in this entry (e.g. field 32), small increases in internal calcium (50-150nM) appear to increase M-current by phosphorylation by an unknown kinase (possibly MLCK in bullfrog neurones) to promote high Popen channel activity. Larger {Ca 2 + h loads would be predicted to decrease M-current by activating calcineurin and promoting low Popen channel activity. Notably, however, these modulations may have little in common with the pathways that agonists use, which are still uncertain.

9 f""t

~ CJ1

(J.J

Proposed ttonic regulation' of M-channels by variations in resting intracellular [Ca 2+J

Conflicting reports on effects of Ca 2+ chelation tRequirement' for low concentrations of Ca 2+ for 1M suppression

II

53-44-11: In rat superior cervical sympathetic neurones, application of Ca2+ to the internal face of inside-out patches produces two forms of unitary M-channel inhibition151 : (i) a slow, all-or-nothing suppression of activity, enhanced by patch depolarization and (ii) a fast block associated with a concentration-dependent shortening of open times (Le. compatible with open-channel block, also enhanced by patch depolarization). In this study, both forms of block appeared independent of dephosphorylation events 151 (cf. Protein phosphorylation, 53-32). In cell-attached patch recordings; (i) M-channel activity increases during exposure of the cell to Ca2+ -free solution and (ii) is rapidly reduced on applying 2 mM Ca2+ to the extra-patch solution, observations which have been used to suggest a 'tonic' regulation of M-channels by [Ca2+Ji in these cells 151 . 53-44-12: Chelation of intracellular [Ca2 +] following receptor stimulation can block the occurrence of 'over-recovery,70 (see Phenotypic expression, 53-14 and Protein phosphorylation, 53-32). Some reports state that high concentrations of intracellular Ca2 + buffers such as BAPTA (e.g. 20 mM) block80,92 or do not block actions of muscarine19,70,90. In ref. 80, muscarinic stimulation of rat sympathetic neurones was able to suppress 1M without apparently raising [Ca2+Ji, however, 150/0). This was taken as injection of BAPTA (r-v11-12mM) reduced suppression of 1M (82~ evidence for a 'minimal [Ca 2+]j requirement' for 'continued operation' of the pathway coupling muscarinic receptors to M-type K+ channels in frog neurones.

Effects of Ca 2 + store releasing agents

53-44-13: Elevation of intracellular [Ca2 +] with caffeine (see ILG Ca Ca RyR-Caf, entry 17) reduces IKM,ng 1. Other methylxanthines (isobutyl-methylxanthine, theophylline) in the millimolar range also reduce 1M 23,93.

Ca 2+ -dependent potentiation Internal Ca 2+ and tover-recovery' phenomena

53-44-14: Other work 94 in cultured bullfrog sympathetic neurones has described a calcium-dependent potentiation of M-current, showing a hyperpolarizing displacement of the steady-statet activation curvet with high internal calcium (suggesting Ca2+ ions modulate kinetics of M-current, thereby regulating the number of M-channels being open at given membrane potentials 94 ). In a separate paper95, these authors proposed a similar shift in the M-current kinetics/activation curve 'caused' over-recovery phenomena (washout of 20 JlM muscarine inducing an hyperpolarizing shift of approx. 13 mY). For background to tover-recovery', see cross-references, this table, above.

(t)

='

t""I"

~ c.n

VJ

II

Table 3. Continued Putative modulators

Description/cross-references

Ca 2+ -dependent AHP Heparin prevents M-current over-recovery but not M-current suppression

53-44-15: A calcium-dependent after-hyperpolarization t (AHP) in dissociated bullfrog sympathetic neurones has been ascribed not only to an IAHP component (see ILG K Ca, entry 27) but also to a calcium-dependent potentiation of M-current150. Other studies have shown intracellular application of the inositol1,4,5-trisphosphate antagonist, heparin (150M (see ILG Ca InsPs, entry 19), has little or no effect on muscarine-induced M-current suppression whilst the 'over-recovery' phase of the response was markedly attenuated, supporting the interpretation that agonist-induced elevation of intracellular Ca2 + may account for M-current 'over-recovery'.

Effects of varying extracellular 53-44-16: Steady-state IKM,ng in differentiated NG108-15 mouse neuroblastoma x rat glioma hybrid calcium on IKM,ng cells is increased on removing external Ca2 + 1 . In the presence of external Ca2 +, reactivation of IM,ng (after a hyperpolarizing step) is delayed. The delay in reactivation is preceded by an inward Ca2+ current, and coincides with an increase in intracellular [Ca2+] as measured with indo-1 fluorescence. IKM,ng response inhibition by

calmodulin antagonists

53-44-17: In NG 108-15 neuroblastoma x glioma hybrid cells, I KM .ng responses are inhibited by the calmodulin antagonist compound W7 (N-( 6-aminohexyl)-5-chloro-l-naphthalenesulphonamide )93.

Alteration of KM responses by 53-44-18: For effects of intracellular ATP on M-current phosphomodulation, see Fig. 3 and ATP and nucleotide analogues associated text under Protein phosphorylation, 53-32 53-44-19: ATP-,-S and ,B,,-methylene-ATP increase the rise time and duration of the response of 1M channels to muscarine in frog sympathetic ganglion cells39 (see also Rundown, 53-39 and Protein Phosphorylation, 53-32).

Modulation of 1M by arachidonate and its metabolites

53-44-20: Several studies have described modulation of M-current responses by arachidonic acid and its metabolites (see separate paragraphs, this field; for general background, see also ILG Ca AA-LTC4 [native], entry 15 and ILG K AA [native], entry 26).

('l)

::s

~ CJ1

W

Comparative effects of ethanol on 1M related to electrical changes in hippocampus

53-44-21: Ethanol enhances muscarinic excitatory responses in rat hippocampal neurones in vivo and, like muscarinic agonists, reduces the M-current in these neurones in vitro. Although superfusion of relatively low concentrations of ethanol (11-22mM alone) has 'little effect' on stratum radiatum (SR)-evoked field potentials t in hippocampal slices (for background, see Receptor/transducer interactions, 53-49), these concentrations enhance (by 10-90%) both the depressions of evoked field potentials and depolarizations elicited by muscarinic agonists in this preparation (ibid., ref. 97). At higher concentrations (22-44 mM) ethanol also enhances amplitudes and duration of muscarinic slow excitatory post-synaptic potentials t (sEPSPs) recorded intracellularly in CAl and CA3 neurones 97. These latter effects require muscarinic receptor stimulation, and can be enhanced by eserine and blocked by atropine.

1M modulation by linopirdine,

53-44-22: It has been hypothesized that the augmentation of neurotransmitter release induced by linopirdine (DuP 996) in rat hipfocampal CAl pyramidal neurones in vitro may be due in part to 'block' of 1M in these neurones 9 . Note: There is no evidence for direct physical blockade (Le. pore occlusion) of M-channels by linopirdine.

a neurotransmitter release enhancer

'Modulation' of M-current 53-44-23: For references describing the modulatory/blocking effects of external Ba2+ ions on 1M and responses by external Ba2 + ions '1M -like' component l Kx (see Blockers, 53-43).

It

(t

::s

M'

~ CJ1

W

_ entry 53 '----------receptor regulation (see Protein phosphorylation, 53-32). Different 'combinations' of receptor/transducer/effector pathways may be able to account for reported differences in M-channel regulation. The 'precise' mechanisms involved in coupling neurotransmitter receptors to KM channels may have a cell type dependence, and several examples of these are described in Table 4.

Up-regulation of tKM-like' channels in gastric smooth muscle 53-49-28: In freshly dissociated gastric smooth muscle cells of the toad Bufo marinus, an '1M-like' current (see notes 1 and 2) can be activated ('upregulated') following isoproterenol (ISO) acting at ,a-adrenoceptors126. The M-current activation with ,a-adrenergic agonists can be mimicked by the cell-permeant analogue 8_bromo_cAMp126,127, suggesting that channel phosphorylation by protein kinase A may activate the M-channel. In keeping with a 'dual regulation' hypothesis126, muscarine or acetylcholine antagonizes the increase in 1M induced by adrenaline or CAMP128. Notes: 1. Some differences were noted126 between 'endogenous' M-current and the 'ISO-induced' M current, namely that the latter usually exhibited slower relaxations on hyperpolarizing voltage commands and displayed a steadystate conductance/voltage relationship that was shifted (relatively) in the negative direction along the voltage axis. 2. The criteria used to identify the 'ISO-induced' component as 'M-current' included it being (i) outward and K+ -selective; (ii) 'suppressible' by muscarine or acetylcholine; (iii) remaining 'steadily activated' following depolarization and (iv) deactivatedt with hyperpolarization, exhibiting slow relaxations in response to voltage j umps126. 3. Substance P (see Table 4) also reduces the acetylcholine-sensitive KM-like conductance in freshly dissociated toad stomach smooth muscle cells 129,130.

Cell type selective effects of neuromodulators: 1M suppression in ganglia 53-49-29: The electrical effects of somatostatin-28 (SS-28 or 'whole somatostatin') and somatostatin-14 (SS-14 or 'cyclic somatostatin') on bullfrog paravertebral sympathetic (dorsal root) ganglion neurones are variable according to cell type 131,132. In C cells (relatively small cells of diameter ~20 JlM) muscarine produces hyperpolarization; SS-28 is also inhibitory and activates an inwardly entry 31); SS-14 is ineffective in rectifying K+ current, KG (see 1NR K G/A(~h, C cells. Note: In separate studies on C cells 133, substance P (0.1-1 JlM) was shown to inhibit an M-type potassiuln current while ATP (1-10 JlM) activated an Na+/K+ current (see 1h under 1NR K 1fhQ [native], entry 34 and also the comparative notes, below). In B cells (relatively large cells where muscarine produces excitatory effects), S8-14 is also excitatory and is more effective than SS-28 in suppressing 1M131 . Comparative notes: In A cells (~65 JlM in diameter) here ATP has been shown to inhibit M-current, substance P (0.1-1 JlM) also inhibited this potassium current without activating 1h133. A full analysis of the modulation and interactions between 1M and 1h in bullfrog sympathetic ganglia has appeared134, concluding that resting membrane conductance of these ganglion cells can be regulated by basal activities of calmodulin-dependent protein kinase (acting to increase amplitude of 1M, which was activated at potentials more positive than -65mV) and protein kinase A (acting to increase amplitude of 1h, which was

Table 4. Receptor/transducer and signalling pathways involved in M-channel regulation (From 53-49-02)

II

Receptor/transducer

Examples (non-exhaustive, notes and cross-references)

Receptors/ligands mediating 1M suppression (excitatory responses) Muscarinic Ml and M 3

53-49-03: In rat superior cervical ganglion (SCG) sympathetic neurones, muscarinic suppression of M current is slow, BAPTA sensitive, and mediated by receptors of the M 1 subtype (as distinct from the pertussis toxin-sensitive suppression of Ca2+ current in the same cells, mediated primarily by M 4 muscarinic receptors)101. For general effects of muscarinic agonists on hippocampal CAl field potentials, see separate paragraph, this field.

Note: 1M is sometimes designated 1KM .ng in NGI08-IS cells.

M 3 receptors couple 'most effectively' to 1M in transfected mouse NG108-15 models 53-49-04: Direct comparison of the ability of different muscarinic acetylcholine receptors (mAChR) to inhibit 1M , has been assayed in clones of NGI08-IS mouse neuroblastoma x rat glioma cells transfected cDNA encoding mAChRs M 1 -M4 using tight-seal, whole-cell patch clamp recordings. No significant inhibition of 1M is seen in non-transfected cells, or in M 2 or M 4 DNA-transfected cells 102 (for AChl mM; muscarine 100 J,1M). At maximally effective concentrations, ACh or muscarine produces complete inhibition of 1M in M 3 -transfected cells, but only partial (50-60%) inhibition in M 1 DNA-transfected cells 103 . --Comparative note: An earlier, detailed study of the mechanism of M-current inhibition by stably expressed muscarinic M 1 receptors in NGI08-IS cells suggested that activation of phospholipase C and inhibition of 1K .M represented parallel (Le. independent) pathway responses to acetylcholine104. This study also found that inhibitors of phospholipase A2 , lipoxygenase, cyclooxygenase or nitric oxide synthase (for example) had 'no significant effect' on ACh-induced inhibition of 1K .M (compare to other studies described under Channel modulation, 53-44). These findings may still be compatible with an interpretation that heterologously expressed receptors coupling to phospholipase C and the production of InsP3 and DAG can inhibit M-current. However, evidence that either InsP3 or DAG can mediate agonist-induced suppression of M-current is lacking.

P2U purinoceptors

53-49-05: Extracellular ATP suppresses macroscopic M-current100.

(l)

~

M'"

~ CJ1 ~

III

Table 4. Continued Receptor/transducer


M-current inhibition following activation of P2 receptors in NG10B-15 cells

53-49-06: Phospholipase C-linked nucleotide receptors (probably of the P2U subtype, see below) activated by both UTP and ATP (100 JiM) inhibit I KM .ng by 44% (UTP) and 42% (ATP) in NGI08-15 cells 10s . Mean ICso values derived in these cells were: UTP: 0.77 ±0.27 JiM; ATP: 1.81 ±0.82J,1M. The order of nucleotide t and nucleoside t activities for M-current inhibition in NG108-1S cells (100 JiM agonist concentrations) was determined as UTP ==ATP > adenosine Sf -O-3-thiotriphosphate (ATP-,S)==inosine Sf-triphosphate (ITP) > 2-methylthio ATP (2-Me-S-ATP) > ADP==GTP» G:,,B-methylene ATP (AMP-CPP), adenosine 1 0 5 . Notes: 1. Inhibitory effects of the P2 agonists on M-current can not inhibited by (i) suramin (500 JiM) or (ii) pre-incubation with pertussis toxin (12 h in 500 ng/ml PTx). 2. M-current inhibition is frequently preceded by a Ca2+-activated transient outward current, responding to intracellular Ca 2+ release (see ILG K Ca, entry 27) but no effects on the delayed rectifier K+ current in this preparation were noted105 . 3. The nucleotide/nucleoside-sensitivity series for M-current inhibition is consistent with agonist sensitivities for PLC-coupled P2U receptors. cDNAs encoding P2U receptors have been isolated from from a NG108-1S cell cDNA library.

Substance P (SP, NK1)

53-49-07: Substance P (SP, 3 nM to 1 JiM) suppresses M-current in neurones of bullfrog dorsal root ganglia (DRG) by reducing the maximum M-conductance in the voltage range of -10 to -130mV106. Notes: 1. Neurokinin A (NKA) and neurokinin B (NKB) also produces the inward current in DRG cells with the rank order of agonist potency being NKA == SP » NKB. 2. SP also decreases a voltageindependent background K+ current (IKB , regulated by intracellular ATP, see also INR K ATP-i [native], entry 30) at a holding potential more negative than -60 m V. K+ current suppression/ decreased membrane conductance is associated with development of an inward current following application of Sp 106 . 53-49-08: In coeliac ganglion neurones, muscarine (ICso 3 JiM) and substance P (SP, ICso 100nM) cause depolarizations or inward currents (under voltage clamp) at the resting potential (-SSmV) associated with a decreased membrane conductance107. These changes have been correlated with 'block' of M-like current, slow IAHP and a K+ 'leak' component (but not an apamin-sensitive IAHP nor the IA 107.

Bradykinin

53-49-09: Bradykinin (BK) is a peptide mediator released during inflammatory processes that has a potent excitatory action in sympathetic neurones, partly due to its inhibitory effects on M-channels

(t)

l:j t'+

~ CJ1

(J.J

(EC so ~ 1.9 nM)108. The BK response is mediated by a G protein pathway similar to that activated by muscarinic acetylcholine receptors - Le. is selectively inhibited by microinjection of antibodies to Go (aq/ld and has been shown to involve the B2 receptor subtype108. In a separate study109, bradykinin (10 JlM) applied outside cell-attachedt patches of differentiated NGI08-15 cells was able to reversibly reduce steady-state ('in-patch') M-channel activity to 28.8 ± 6.1 % of control value (more potently than muscarine - cf. 38.4 ± 11.7% with muscarine)109. Inhibition in NGI08-1S cells was accompanied by a lengthening of channel shut times without significant change in open times or distribution of conductance levels 109. 53-49-10: Bradykinin produces 'dual' opposing but sequential effects on separate outward and inward (opposing) currents in voltage-clamped NG-I08 cells at clamp potentials between -60 and -30 mV (designated IBK(out) and IBK(in), respectively)110. IBK(in) results primarily from inhibition of the Ca2 +-independent 1M (IBK(out) results from activation of a Ca2 +-dependent, voltage-insensitive K+ conductance (see ILG K Ca, entry 27). The effect on IBK(in) is not replicated by a rise of intracellular Ca2 + and must therefore be generated by another mechanism 110. 53-49-11: In NGI08-1S cells, M current modulation by bradykinin appears to involve a PTx-sensitive G protein110-112. Protein kinase C (PKC) activation by phorbol esters causes suppression of I KM .ng in these cells l12 . Note that data compatible with PKC mediating part of the bradykinin effect in NGI08-1S have been published (focal application of O.I-S JlM bradykinin inhibits 1M by about 60%; SnM bradykinin inhibits by about 40%)113. Bradykinin can produce an additional suppression to that induced by phorbol dibutyrate (PDBu). 53-49-12: Inhibition of the M-current suppression in PC12 cells by bradykinin is mediated by a PTx-insensitive G protein; some steps in the second messenger cascade are modulated by Ca2 +152.

TRH

II

53-49-13: Thyrotrophin-releasing hormone (TRH) has been shown to suppress M-current in dissociated rat hippocampal CAl pyramidal neurone preparations l14 . Similarly, TRH inhibits an 'non-neuronal, M-like current' in normal rat pituitary lactotrophs; since this current begins to activate at ~ -60 m V it is likely to be active under the normal resting potentials of lactotrophs (-3S to -4S mV)10. Functional note: The TRH-induced increase in firing frequency is dependent on extracellular calcium and contributes to prolactin secretion.

(l)

=:s

M-

~ c.n VJ

II



LHRH

53-49-14: In rat superior cervical ganglion neurones, tLHRH (luteinizing-hormone releasing hormone) receptors activate a 'PLC/DAG/PKC' pathway (similar to that used by muscarinic M}/Ma- or substance Preceptor-mediated M-current suppression, see above). 53-49-15: In frog sympathetic neurones at concentrations that inhibit M-current, LHRH (and substance P) 'strongly reduce' N-type Ca2+ current but induce a leak conductance that may contribute to slow EPSPs (muscarine produces little reduction of Ca2+ current under these conditions, even in cells in which it strongly suppress the M-current)115.

Glutamergic agonists

Angiotensin II

See also endothelin-l under ICommon IIA1-suppressionuncoupling' effects of decreased cellular NAD+' under Channel modulation, 53-44.

Inhibition of 1M by metabotropic glutamatergic agonists 53-49-16: In addition to muscarinic cholinergic agonists, the metabotropic glutamatergic receptor agonist ACPD (I-aminocyclopentane-ls,3R-dicarboxylic acid) and appears to activate separate intracellular transduction pathways having 'convergent' inhibitory effects onto 1M and I AHP in basolateral amygdala pyramidal neurones of rat. These modulations result in membrane depolarization and reductions in the amplitude and duration of the slow AHpl16 (see also ILG K Ca, entry 27). 53-49-17: The octapeptide angiotensin II inhibits 'approx. SO%' of KM channels in cultured rat superior cervical ganglion (SCG) neurones without significant elevations in [Ca2 +Ji concentrationl17 (for earlier work first suggesting a common inhibition of KM by muscarine and angiotensin II in SCG, see refs 44,118). This inhibition appears to use a slow, second messenger-dependent, pertussis toxin-insensitive signalling pathway (as used by muscarinic agonists, possibly involving G q , see this table). By the same (apparent) pathway, angiotensin II also inhibits an N-type calcium channel via AT} receptors and this can be attenuated by inclusion of GDP-,B-S (2 mM) in the pipette. M-like current in NGIOS-IS cells is also suppressed by angiotensin II (see ref. 1, this table). Supplementary note: Angiotensin II is best characterized in the renin-angiotensin-aldosterone t system regulating blood pressure, but it also acts as a neurotransmitter in the central and peripheral nervous systems. For references to G protein coupling mechanisms and signalling functions of angiotensin II, see listings under Related sources and reviews, 53-56, and Receptor/transducer interactions under (for example) VLG Ca, 42-49 and INR K [native], 32-49.

g ~ CJ1

(J.J

'Opposing' effects of opioids on 1M depending on receptor subtype coupling Receptors mediating 1M augmentation (inhibitory responses) Kappa opioid subtypes; {3-adrenoceptors (possibly, '1M-like')

For complex effects of somatostatin, see separate paragraphs, this field.

53-49-18: The opioid kappa receptor-selective agonist U-50,488H significantly augments 1M in most CA3 hippocampal neurones l19. Similarly, the kappa-selective agonists dynorphin A and dynorphin B (which exist in mossy fibre afferents to CA3 pyramidal neurones), also markedly augment 1M at low concentrations (20-100nM)119. Note: At relatively high concentrations (1-1.5 JlM) dynorphin A has an inhibitory effect on 1M • In contrast, the opioid delta receptor-selective opiate agonists DADL (D-Ala2,D-Leus-enkephalin) and DPDPE ([D-Pen(2,5)]-enkephalin) reduce the hippocampal CA3 M-current, with partial reversal of this effect by naloxone. 1M is 'not consistently altered' by mu-opioid receptor agonists such as DAMGO ([D-Ala2,NMe-Phe4,Gly-ol]enkephalin) or non-opioid fragments of dynorphin (e.g. des-Tyr-dynorphin)119. These and other results help to explain (i) the mixed effects of opiates seen in other studies and (ii) a potential post-synaptic function for the endogenous opiates contained in the CA3 mossy fibres l19. 53-49-19: In gastric smooth muscle cells of the toad, an '1M -like' current can be activated ('up-regulated') by activation of {3-adrenoceptors (see paragraph 53-49-29).

G protein transducers mediating 1M inhibition

PTx-insensitive (compare PTx-sensitive coupling in NC108 cells; this table, below) Analysis of G protein selectivity using anti-C protein antibodies

II

53-49-20: M-current inhibition mediated by G protein(s) 'coupling to' receptors for muscarinic agonists, LHRH, substance P and UTP (this table) have been described in sympathetic neurones 19,3l,129. In cultured rat sympathetic neurones, 1M modulation is independent of the cyclic nucleotides cAMP and cGMP (but appears to involve activation of a pertussis-toxin {PTx)-insensitive G protein, generation of diacylglycerol (DAG) and activation of protein kinase C 3l (see other paragraphs, this table). The 'precise' roles of DAG and PKC in KM modulation are unclear and may exhibit marked cell type dependence (see also Protein phosphorylation, 53-32). 53-49-21: Following their microinjection in rat superior cervical ganglion (SCG) sympathetic neurones, antibodies specific for Gag or Gall (but not those for Gao) reduce M-current inhibition by the muscarinic agonist oxotremorine-M (acting at M I muscarinic receptors, see this table). Immunoblotting experiments demonstrated the presence of both Gaq and Gall while GaOl (but virtually no G a02 ) was present in the preparation12l . Note: This study also established that anti-Gao antibodies (but not those specific for Gaqjll or Gall-3) reduce a-adrenoceptor-mediated inhibition of lea in SCG neurones (see also Receptor/transducer interactions under VLC Ca, 53-49).

(b

=:s

M-

~ CJ1 VJ

III



Signal transduction mechanisms

53-49-22: Generally, in the cell-attachedt recording configuration, sustained depolarization-activated

Muscarinic receptor agonists M-current can be inhibited by application of muscarine or muscarinic receptor agonists outside the-patch. In dissociated rat superior cervical sympathetic neurones, M-channel closure by muscarinic acetylcholine receptor stimulants is consistent with a requirement for a diffusible messenger (i.e. the inhibition is not due to a local direct 'membrane-delimited' interaction between the receptor, transducer and channel proteins)122 (compare Receptor/transducer interactions under 1NR K G/ACh, 31-49; for Ca 2 + as an candidate for a diffusible messenger, see Channel modulation, 53-44). 53-49-23: Comparative note: Models based on agonist-dependent kinetics of nucleotide interaction with G proteins (accounting for time course of slow synaptic potentials caused by 1M inhibition in sympathetic neurones) without requirement for subsequent messengers are briefly described under Kinetic model, 53-38. 53-49-24: Note: Involvement of multiple messengers may be implicated in M-current suppression in

view of the multiple receptor types with an 'appropriately' slow time course123.

Time course of M-current modulation by muscarinic agonists

53-49-25: The time course and latency of M-current inhibition has been studied in whole-cell dissociated bullfrog sympathetic neurones under conditions where fast inward currents (through nAChR, see entry 09) are abolished by voltage-clamping to -39mV124. Under these conditions; (i) ACh or muscarine induces a slower inward current developed after a latency of ~200 ms, due to M-current inhibition; (ii) the time elapsed from the termination of maximal agonist ~plication (10 JlM muscarine) to the initiation of recovery was 4.5 s ('tr - a '); (iii) the t r - a value is independent of the duration of the stimulus between 0.5 and 10 s, (iv) applications >4 s are necessary to reach ~95% of the maximal inhibition during agonist application and (v) saturating concentrations of the second messenger are produced in 700/0 in vertebrates, see below) although 'species-dependent' sequence variations 441'59 have been specifically correlated with functional differences in a number of cases (see Protein phosphorylation, 54-32; Activation, 54-33; Current-voltage relation, 54-35 and Blockers, 54-43). Mouse and rat IsK share 920/0 sequence identity; human isoforms are 76-770/0 identical to the rodent sequences.

Protein molecular weight (purified) 54-22-01: MinK proteins typically migrate as a band of approximately 15 kDa (e.g. from rat kidney) following denaturing PAGE.

Protein molecular weight (calc.) 54-23-01: The rat kidney minK ORFt predicts a protein molecular weight of 14.7kDa (14698Da).

Sequence motifs 54-24-01: minK proteins contain two conserved consensus sites for Nglycosylation in their N-termini (see Fig. 4 under Encoding, 54-19 and the

II

D

B

EXOn¥n 1B

A

;:1'= I:

U

La

I

I

Ie

..

I D

~

Exon 2

)1:;:

I

I

E H

H

H

B

),'=

KHB

B

B

~

IJ

K pK127,28 pKI 29, 30

E:~t

K

GTIIC

... actggattctgCCaCCCaCagagCaCagctccccatctgctttt9tcaaca99agtattgtct~gttgcaggtaggcccacg9a99gcct

Spl

tcagttacccccagaagcacacaggagctgtggtccttagaagtgaccagatgaggactaccatggtga99g999t99gggg999C99cagtgacccttg APl

gctgtgtcttttagctgtagctcagaagcctgtctctgtatcttatctactaggetgacggagccctttcaccagaactgcccctgctgiCtCicacctg tcaagcgqgtgggcaacgctatgaaattgaccaggaagtcacctgggccacaccagcgcttgctcctgctgggagagaaaggcggttcatacttgcctaa -143 Exon1A gaact~GGAGGTGGCGCCAGGGCTGGAGTTCCTCCAGCAACTGACTGCctGTAGCAGAGCCCCGACTGTTTAGCCACCTCTGCAC¢GTCCATCCAGGTC -40 ~~CCGT~C~taa9taacagctcaatatacccttgctgaata9ctgc9999ccacaggggtccagcaggatgtgaagacaaag9gagct9gagaagttc

tgaaccatctctgaeagtgtttggaactatttaacaaggtgtttgtccagtgctgacccagtcatctgtcctctttattctcccacagactgtctgccag GTI~

Sph

GTI

ccctaagcccacccgctggagtcaaagtgcttcctgccctt9999ct999tgag9catgcagcatgcagg9t99999tgcagggtgggttgggtaacctc ~ p U 1 Sp1 1 ~ 2 ~ -.. 2 cgagacctgggctgactgagccaagccctgcctgctggaagccccagggctctgaggctccgcccatccaaagctgcattgcttaggtgcctctgggatt

I

J .--.tIOF

I

- 9 109

. ~

.

Exon 18

-40

gtctgtcagtcttgtctg~1CtTAGAC¢CCAACACGGGCTCCAAGGATCCCGCTGCACACCAGGCTCCCTTGGCTTCTAGAC¢¢AOgtaagtcaggg 3 ~ 3 ~

l.

-39 agggc t cc tg t e t e ta ggaca gg t ace· •..•••• • ( I ntron - 9 • 5 kbp) .• • •.•..•. t tea t t t ca cG ....A. ,.G..T..T..T~T".G. ,.Cl"l lT. ,.C~C:"': A.,.C~A

-~F

..T~C:":"A.,.G.,.G.,.G-rAAA-r-r~Cr7l c~&r'IfI'f

Exon 2

GAAGCCCCAGGATGGCCCTGTCCAATTCC·······(366bp)·······CTGTCATGAACCCCATAGTTAATTAATAGACAAGTGATAAGTGGGTCTTT ~etAlaLeuSerAsnSer • • . (122aa) . . . LeuSerl 522 CTAGTCAAATGCCTGCCAGTCTTTATTGTAGAGGTACCCTTGAG~TAAGGGG~-~AATAACACCAGTTTTCTGAAATTG~ttctttctata

gtaatcaatttatt···3'

c

g ~ CJ1 ~

Figure 5. Organization of the gene encoding the rat IsK (minK) protein. Boxed text comprises exon lA (complete sequence, exon lB (complete sequence), exon 2 (beginning and ending of sequence, omitting 366 bp/122 aa as shown). In the upper part of the figure, the segment filled in black represents the protein-coding region, while the white segments represent the 5' and 3' untranslated regions. A B

C

D

E

G H I

J K L

II

Partial restriction map and derived gene structure produced from overlapping phage lambda genomic clones spanning the rat IsK region. Region common to two mRNA species designated IsK-A and IsK-B and encoded by a single exon (exon 2, filled black) consisting of 561 bp. Exon 2 thus contains the entire protein-coding region; this region and the 3' untranslated region are uninterrupted by intron insertion. The sequence AATAAC (double underlined) is located at 20 nucleotides upstream from the polyadenylation site. This AAUAAC variant of the Itypical' AAUAAA sequence can serve as a signal for polyadenylation of mRNA (for details on specificity, see Wickens and Stephenson (1984) Science 226: 1045-51). The 5' sequence specific for IsK-A and IsK-B mRNAs are encoded by two separate exons (exon lA, located rvl0.0kb upstream from exon 2 and exon lB, located rvl0.0kb upstream from exon 2). Nucleotide sequences of exons and flanking regions, with nucleotide numbering relative to the first residue of the translation initiation codon at F. Approximate nucleotide position denoting a 5' terminus deduced from primer extension analysis t Approximate nucleotide positions denoting 5' termini deduced from RNAase protection analysis t. The different locations predicted in G and H may be due to the presence of a palindromic structure in this region (see J). Single-headed numbered arrows indicate short direct repeat sequences Double-headed arrow indicating the location of a palindromic structure. Predicted alternative splice patterns as deduced from composite sequences of eDNA clones pK127 to pK130 as indicated. Ends of large intron of 9.5kb (internal sequence omitted). Other symbols Neither a TATA boxt nor a CAAT boxt is present in the 5' flanking regions of exon lA or exon lB. There are, however many potential promotert , enhancert and cis-regulatoryt element motifs present in the 5' flanking regions of both exon lA or exon lB (indicated by GTHC, Spl, APl, GT1, and Sph (for details on specificity, see Wingender, E. (1988) Nucleic Acids Res 16: 1879-902).

(Reproduced with permission from Iwai et al. (1990)

J Biochem Tokyo 108: 200-6.)

(From 54-20-01)

g t""t

~ c.n ~

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_t_ry_54---'

[PDTM}, Fig. 6). The role of N-glycosylation in 'polarized' expression of minK has been described60 . A single cysteine residue in the C-terminus (Cys106 residue in the C-terminus of the human/mouse IsK proteins, Cysl07 in rat) appears to act as an oxidizable site (ibid.; for details, see Table 7 under Channel modulation, 54-44 and the [PDTM), Fig. 6). There are no consensus N-terminal signals predicted from minK cDNA sequences, suggesting that they are not proteolyticallyt processed upon insertion in the membrane. For sequence motifs associated with modifications by protein kinases, see Protein phosphorylation, 54-32.

Southerns 54-25-01: Low-stringencyt hybridization of minK probes to rat genomic DNA digested with panels of restrictiont enzymes have shown 'not more than two bands,6. Such banding patterns are consistent with a single-copy gene (or limited number of minK-related sequences) being represented in the genome.


Domain arrangement Membrane orientation of minK subunits 54-27-01: MinK subunits possess a single hydrophobic transmembrane domain (MI, sometimes designated HI) of 23 amino acids (see Fig. 4 under Encoding, 54-19 and the [PDTM), Fig. 6). Immunolocalization of a nine amino acid 'epitope tag' fused to the N-terminus of minK have confirmed the N-terminal domain to be extracellular61 . Effects of Serl03Ala mutations in the rat kidney minK (which prevent inhibition by protein kinase C - see Protein phosphorylation, 54-32) indicate the residue to be intracellular (see

[PDTM), Fig. 6).

Alpha helical arrangement and functional effects of Cys substitutions 54-27-02: A spacing of 'critical' phenylalanines at every third residue have led to predictions of an a-helical conformation for the membrane-spanning domain of IsK (but see Protein topography, 54-30). For functional effects of MI Cys substitutions on activation and subunit association properties, see

Domain functions, 54-29 and Protein interactions, 54-31 respectively.

Domain conservation 54-28-01: The single predicted transmembrane domain (see [PDTM), Fig. 6, uncharged residues ",44-66 in the neonatal mouse heart sequence} shows 96% identity between mouse and human (residues 41-90). In common with ion channel proteins, the N- and C-termini show greater divergence, with the C-terminal being relatively 'more conserved': N-terminal domains share only 60-78 % identity amongst known isolates.

'-.--e_n_t_ry_5_4

_

For predicted N-terminal topography,

N H2

(aa 1) Approx position of 'extra' a.partate

~

se. legend.

~::~~~g(~~)

N- and C-terminal 'double· deletion' mu1ants (OAF's of 63 aa

1~O :~ ~~~~al: ~~~:red

are~(e.g.~10-39:~94-130.

to 129 aa for mouse, human and guinea-pig)

delineated by

~

symbols on figure)

Extracellular

Monomeric domains

Intracellular

(' ~

Cy.107

PKC-Inhlblton

_':!='~~.~-'~~--~ SH

I

PltC: • 10110210310.105106107108 Leu Qlu .IIJlX Phe Arg Ala eys Tyr

Rat

(S103 mutants are 'non-phosphorylatable' by PKC)

l

lIouse

Leu Qlu .IIJlX Phe ArIiJ Ala eys 'ryr Leu Qlu .IIJlX Tyr Arg Ser CYI Tyr

B\DII&D

", I

Note: IsK proteins also appear to form aggregate. within phospholipid membranes

COOH

Guinea-pig (ha.ologous reliJion) Leu alu Asn eys ArIiJ Ser Cys eys (Guinea-pig native IK,5 & recombinant 15K are 'PKC-enhanced') See Protein phosphoryfation. 44-32

(aa 130)

(

.

(C""'~lIn"n9

R

9~"

could Includ"

c-s-sij . other minK subunits

C-SH ~

- cytoskeletal proteins ) Suggested minK subunit stoichiometry by experiments of Wang and Goldstein (1995) Neuron 14: 1303-1309. See Predicted protein topography, 54-30.

~

Key • •

-~

- other membrane proteins

Likely site for minK chelation by mercury ions See Channel modulation, 54-44.

Dlsulphlde bond formation at the cysteine 'oxidizable site' (Cys 107)

1 - - - - - - See Channel modulation. 44-30.

to protein kinase C substrate Arg-Ser-Lys-Lys (aa 68-71)

- Posltlon.of CamKII

l' - Positions

)

m.o.1I! (extracellular, non-functional; aa 35-41)

of motifs lor N-glyc05ylation (ae 5 and ae 26)

NOTE: All relative positions of motifs. domain shapes and sizes are diagrammatic and are subject to re-interpretation.

Protein symbol (K)

· ·:· :,.'~ O

";:-:":::'::1.

: 1,~,

.J. . · · · ·

+..•....

1:'"

Figure 6. Protein domain topography model (PDTM) for MinK (lsK) protein monomers. Amino acid residue positions apply to the rat kidney isoform (Takumi et al. (1988) Science 242: 1042-45). Note: The subunit topography is shown tunfolded'. Some studies (see Predicted protein topography, 54-30) have suggested that both the N-terminal and transmembrane segments of 1sK adopt a-helical structures (with both regions being located within the lipid bilayer while their linking segment lies on the surface of the membrane - see also later paper predicting tilted f3-strand structures in transmembrane domain, ibid.). Note: For interpretation of minK expression data in Xenopus oocytes, see Protein interactions, 54-31 and Selectivity, 54-40. (From 54-28-01)

Domain functions (predicted) Note: For interpretation of minK expression data in Xenopus oocytes, see Protein interactions, 54-31 and Selectivity, 54-40.

1sK transmembrane domain Cys substitutions affecting activation properties 54-29-01: A series of six mutants of the rat kidney IsK protein have been constructed where consecutive individual amino acids (residues 53-58) of the transmembrane region have been substituted by cysteine62 . Oocyte expression of one mutant (Phe55Cys) displayed activation curves that shift in a hyperpolarizing direction, while for another mutant (Phe58Cys) activation curves shift in a depolarizing direction, suggesting that the (hydrophobic) phenylalanine residues may play a critical role in IsK activation gating. The spacing of 'critical' phenylalanines at every third residue may indicate

_ entry 54 ~-------

(a)

500

-20mv 400

...>'ii'

:>

300 t=~ 0Coo

u:i

zZO

200

c~ :::t-

o

100

~g~2~~~~~~~~~~~~~~g~~g2~~~~~~~g~~~~ ~~~:

2~

~

-2~

~~~~~~~~~~~~~b~~~

0)

M

C

(b)

300......-------------------------------,

Omv

(c)

200

100

o ~o~oo~~~N~~~~~~~~~o~~g2~~~~g~~g~~g~

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~bbffi~ zQ)

c(

C

Cl

M

CO') It)

EXTRACELLULAR

(d) Rat Human

CYTOPLASMIC 100

120

130

ARVLESFrr~I~A~V~THLPEL~LS ARVLESY~S~V~L~I~NTHLPET~SP

AT OR 00 01< N E

Figure 7. Summary of functional effects following amino acid substitutions in 1sK sequences. In (a) to (c), the columns show the mean percentage channel activity' for specified site-directed mutant 1sK proteins (as shown) relative to wild-type 1sK (WT) expressed in Xenopus oocytes*. The different histogram sets compare channel activities for the same set of mutants (measured 90 s after depolarization from a holding potential of -60mV stepped to -20mV (in (a)), to OmV (in (b)) I

l_e_n_t_ry_54

_

an a-helical conformation for the membrane-spanning domain of IsK. These results also indicate that one face of the helix may represent a region of subunit association (see also Protein interactions, 54-31). Note: Expression of the remaining four eys mutants (positions 53, 54, 56, 57) in Xenopus oocytes resulted in currents which were similar to wild-type62 . Mutations in the hydrophobic domain which do not significantly alter expression of the voltage-gated currents are described in refs S3,S4 • See also Fig. 7, Protein phosphorylation, 54-32; Selectivity, 54-40 and Blockers, 54-43.

Functional N- and C-terminal tdouble-deletion mutants' 54-29-02: Following oocyte expression, an N- and C-terminal deletion mutant designated ~10-39:~94-130 (see [PDTMj, Fig. 6) containing only 63 of the 130 residues of the rat IsK protein supports K+ currents that are indistinguishable from those expressed from wild-type proteins63 • This demonstrates that the internal (most highly conserved) sequences are 'sufficient' for the formation of functional channels (or, alternatively, do not impair putative 'activator' functions - see Protein interactions, 54-31). Comparative note: These conclusions should be compared with those describing the ability of the N34 and C-27 minK peptides alone to activate el- and K+ channel activities in oocytes (for details, see Selectivity, 54-40). The N-34 and C-27 peptide sequences are underlined on the sequence alignment (Fig. 4) shown under Encoding, 54-19.

Cross-references to other minK structure-function analyses 54-29-03: Numerous other point t and deletion/truncation t mutants of minK subunits have been expressed, some of which give rise to dominant-Iethal t phenotypes (see Predicted protein topography, 54-30), changes in I-V relationships (see Current-voltage relation, 54-35), subunit association and putative 'activator' functions on 'silent' channel activities (see Protein interactions, 54-31 and Selectivity, 54-40), susceptibility to modulation by

and to +20mV (in (c)). Unshaded columns denote substitutions outside the transmembrane domain M1, with shaded columns indicating mutants within the M1 region. In (d), the aligned sequences of rat and human IsK are annotated with positions of potential N-glycosylation sites (CHO), the domain positions, and the one-residue tdeletion'in the human IsK sequence (hyphen, see also the [PDTMj, Fig. 6). Locations of amino acid substitutions are mapped by reference to their tgross' effects: 'J, mutants showing tmoderate' reductions in channel activity; ., mutants showing tdrastic' reductions in channel activity; 0, tsilent' mutants showing no significant effects on IsK activity; e, positions of mutants which tenhance' IsK activity. The thick arrow denotes the direction of K+ current permeation. Note: The results shown here summarize effects within a single comparative study (Takumi eta al. (1991) J BioI Chern 266: 22192-8 from which the figures are taken with permission); other structure/function studies are indicated in the text. *Note: These and other results generated in the Xenopus oocyte expression system should be interpreted in the light of findings described under Protein interactions, 54-31 and Selectivity, 54-40. (From 54-29-03)

III

_'--

e_n_t_ry_54_

protein kinase C (see Protein phosphorylation, 54-32)J ionic selectivity functions (see Selectivity, 54-40), Hg2+ ion chelation patterns or sensitivity to oxidants (see Table 7 under Channel modulation, 54-44). In one extensive study 63J a series of 31 amino acid mutations were made in the IsK protein (mostly in the 'internal! regionJ see previous paragraph) which produced a range of effects on IsK channel activity (including 'silene or 'neutral! effects! through 'moderate! to 'drastic! reductions in activity! to those which conferred IsK current enhancement). A summary of these 'variable effects J by reference to specific mutations is shown in Fig. 7 (for further details, see ref. 63).

Predicted protein topography Note: For interpretation of minK expression in Xenopus oocytes, especially minK-KvLQTl co-assembly data, see Protein interactions, 54-31.

Subunit topography/stoichiometry/assembly patterns of t.junctional' minK protein complexes 54-30-01: Subunit stoichiometry of minK channels in Xenopus oocytes has been investigated by co-expression of wild-type t (wt) minK subunits and a point mutant t D77N of minK subunits64 . In this study! evidence in support of three key assumptions was reported! namely (i) wild-type and D77N monomers are expressed in an equal and independent manner (see note 1); (ii) they do not rreferentially self- or transassociate and (iii) D77N has a 'dominant lethal! effect on minK functional expression (see note 2)64. By means of kinetic and binomial distribution analyses (but not single-channel analyses)! these observations supported a model for 'complete! minK potassium channels in which just two minK monomers are co-assembled with other (unidentified) non-MinK subunits (see the PDTM, Fig. 6). Notes: 1. MinK D77N mutants pass no current but have been shown to be expressed in the plasma membrane (as determined by immunolocalization of an incorporated c-myc epitope-tag t QKLISEEDL between minK aa residues 22 and 23). 2. When wild-type and D77N minK cRNAs are coinjected in a 1: 1 ratio J only a 'small fraction! of the current seen with wildtype cRNA alone is observed; following co-injection of 2 ng wild-type plus 0.5 ng D77N cRNA! a fraction of 0.63 ± 0.04 of wild-type current is obtained. Since the latter currents have pharmacological! selectivity and gating properties very similar to those of wild-type alone! it suggests that only 'fully wild-type! channels can conduct current64 . Note: The 'heterotetramer! model suggested by the D77N/wild-type study is incompatible with 'homomeric! models (see next paragraph).

'Homomeric' models for minK oligomers 54-30-02: According to an independent study 65 minK channels were predicted to form 'by assembly of at least 14 monomeric subunits!65. These authors applied a binomial distribution t to the reduction in current amplitudes (at -20mV) observed in oocytes (see note 1) following co-injection of a constant amount of wild-type minK mRNA with increasing amounts of mutant S69A mRNA (see note 2). On the basis of the stoichiometric J

entry54

_

I" - - - - - - - - - -

analysis, a model was proposed for channel formation. Notes: 1. These and other results generated in the Xenopus oocyte expression system should be interpreted in the light of findings described under Protein interactions, 5431 and Selectivity, 54-40. 2. Mutant S69A (see also Fig. 7) was employed in these experiments to induce a I discernible shift' in the I-V relationship at -20mV (i.e. to more depolarized potentials. minK currents are not normally observed at potentials negative to OmV with this mutant alone. The relationship of these findings to those showing minK acting as a channel regulator (see Protein interactions, 54-31) is unclear.

Membrane-bound secondary structure prediction of a-helical conformation 54-30-03: Synthesis, fluorescent-labelling, structural and functional characterization using circular dichroism t of five polypeptides (comprising 20-63 amino acids within the rat IsK protein) have suggested that both the Nterminal and transmembrane segments of IsK adopt a-helical structures66 (compare next paragraph). This approach has also suggested that (i) both the a-helical transmembrane segment and the N-terminal of IsK are located within the lipid bilayerj (ii) the linking segment between the two segments lies on the surface of the membranej (iii) IsK proteins form aggregates within phospholipid membranes and (iv) certain truncated forms of IsK protein are unable to form functional K+ channels in planar lipid membranes 66 . Note: The latter observations were taken as evidence to support the view that the IsK protein alone cannot form K+ channels (see Protein interactions, 54-31).

Membrane-bound secondary structure prediction of {3-strand conformation 54-30-04: Twenty-seven residue peptides corresponding to the transmembrane domain of minK/lsK have been incorporated into synthetic saturated-chain phospholipid membranes prior to structural characterization by a number of spectroscopic techniques 67. Under defined conditions, the minK/lsK transmembrane peptide has been determined to adopt a predominantly ,a-strand secondary structure (compare previous paragraph). In this case, the peptide backbone is oriented at an approx. 56° angle relative to the membrane normal in dry films of phosphatidylcholines. Hydration of the dry film (in the gel phase) does not appear to affect the orientation of the peptide backbone, and a relatively small change in orientation occurs when the bilayer undergoes transition to the fluid phase67. ESR and NMR spectroscopic data indicate (i) that the incorporated peptide restricts rotational motion of lipids in a similar manner to that found for integral membrane proteins and (ii) a selectivity in the interaction of anionic phospholipids with the peptide. Other data indicate that (i) approximately two to three phospholipid molecules interact directly with each peptide monomer (consistent with a limited degree of aggregation of the ,B-sheet structuresh (ii) lipid-peptide complexes have a lamellar structure and (iii) the surface areas occupied by lipid molecules are "'-J30A2 per chain in the peptide complexes, while the additional membrane surface area contributed by the peptide is "'-J112A2 per monomer, consistent with the strong (56°) tilt in the long axis of the ,Bstrand in the dry film (as above)67.

II

II

Table 4. Summary of factors supporting the hypothesis that minK/1sK proteins act as activators of endogenous channel activities. The table lists factors which have been put forward which counter the proposition that minK forms independent channels. Features of minK regulation that are described in other fields of the entry are largely omitted; in consequence, the listing is not exhaustive, and is mainly intended to provide some perspective on the ongoing debate (see also paragraph 54-31-01 and Fig. 8). (From 54-31-01) Factor/effect/observation

Description, conditions, references (see note 1)

Sufficiency of synthetic hydrophilic peptides to induce Cl- or K+ channel activities (direct evidence for regulatory function)

54-31-02: Addition of synthetic IsK hydrophilic peptides derived from C- and N-termini (see next row) are sufficient to activate slow K+ and Cl- channels in untreated Xenopus oocytes. The peptide-induced biophysical and pharmacological characteristics are similar to those exhibited by the native IsK protein68 . For definition of peptides and selectivities of induced currents, see Fig. 10 under Selectivity, 54-40.

Designation of Icritical domains' for induction of Cl- or K+ channel activities

54-31-03: IsK mutagenesis has identified N- and C-terminal domains as critical for the induction of Cl- and K+ channel activities, respectively: The S68T mutant shows a 'complete loss' of K+ current activity with 'no alteration' of Cl- channel activityS (see also mutant S69A under Fig. 7). Truncation of the IsK C-terminus (in the Tr80 mutant) resulted in a 'complete loss' of the K+ channel activity with a 'striking increase' of Cl- current amplitudeS (compare to the reportedly functional 'double-deletion' mutant ~10-39:~94 130 described under Domain functions, 54-29). 54-31-04: Deletion of 28 residues of the IsK N-terminus (in the ~11,38 mutant) 'abolished' expression of the Cl- current and 'doubled' the amplitude of K+ currents.

KvLQT1/minK co-expression

See paragraph 54-31-01, Fig. 8 and 'lack of minK expression in mammalian expression systems' (this table, below).

Noted lack of K+ channel selectivity determinants

54-31-05: IsK/minK proteins do not exhibit 'signature sequences' associated with all other 'cloned' channels that are K+ -selective, Le. a conserved K+ -selective pore (P) region (latterly designated H5) flanked by either six or two membrane-spanning regions (compare entries such as 1NR K [subunits], entry 33; VLC K eag/elk/erg, entry 46 or VLC K Kv1-Shak, entry 48). The difficulty in accounting for a pore structure with both K+ and Cl- selectivity properties led to the suggestion that the IsK protein itself was an 'activator of endogenous and 'dormant' K+ and Cl- channels/so

B M-

~ CJl ~

54-31-06: Note: No structural model for K+ channel pores incorporating minK subunits has been suggested; existing models for Kv channel pores (e.g. see entries 48 and 49) do not appear directly applicable. Observation of CI- selective currents following heterologous expression in Xenopus oocytes

54-31-07: IsK/minK expression in Xenopus oocytes may induce a Cl- -selective current 5 'very similar to' or 'the same as' a Cl- current produced by phospholemman (PLM) expression in oocytes (for comparison, see below and description of phospholemman under Miscellaneous information, 54-55). These Cl- -selective currents were described as exhibiting biophysical, pharmacological and regulatory characteristics 'very different' from those of IsK-induced K+ channel activities 5 (see below).

Differential Ipermissive' expression conditions

54-31-08: IsK cRNA at low concentrations (threshold at 0.03 ng/Jll) induces slowly activating, K+ -selective currents upon depolarization5 . 54-31-09: IsK cRNA at high concentrations (typically 1 Jlg/Jll) induces slowly activating, Cl--selective currents upon hyperpolarization (amplitude -0.S±0.3JlA at -130mV; threshold potential -80mV)5. See also Activation, 54-33.

Differential pharmacology

54-31-10: The compound DIDS (see Openers, 54-48) was reported to 'potently inhibit' both IsK- and PLM-induced Cl- currents without affecting the K+ current (compare conflicting result, ibid.). 54-31-11: Ba2 + ions were reported to be more potent blockers of both IsK- and PLM-induced Cl- currents (IC so 0.4mM and 0.3mM respectively) than the K+ current (ICso 3mM) (compare Blockers, 54-43). 54-31-12: Clofilium was an activator of both IsK- and PLM-induced Cl- currents5 (compare Blockers, 54-43).

Differential modulation

II

54-31-13: While IsK-associated K+ levels were enhanced by elevating internal Ca2+, IsK-associated Cl- currents were enhanced by lowering external Ca2+S (see also Channel modulation, 54-44).

g t""I"

~ c.n ~

II

Table 4. Continued Factor/effect/observation

Description, conditions, references (see note 1)

54-31-14: Increasing osmolarity of the extracellular medium by 40 mOsM using saccharose or sorbitol69 decreases K+ channel activity by rv69% ±3% (the same treatment was 'without effect' on CI- channel activity following IsK or PLM expressions - compare Channel modulation, 54-44). Differential phosphoregulatory properties

54-31-15: The phorbol ester PMA decreased the IsK-associated K+ current (see Protein phosphorylation, 54-32) but strongly increased the IsK-associated CI- current and notably, did not affect PLM-induced CI- currentss. 54-31-16: The Serl02Ala mutant suppressed PKC inhibition of the K+ current (as in Protein phosphorylation, 54-32) but had no effect on PKC enhancement of the CI- current. This was taken as evidence for the existence of two different sites for PKC regulation on IsK, in support of a model in which IsK activates two distinct endogenous channels in oocytes S (see 'Alternative models', below).

Lack of minK expression in mammalian expression systems and IsK reconstitution studies in bilayers

54-31-17: Prior to the successful co-expression of minK and KvLQTl, numerous unsuccessful attempts were made to express minK/lsK alone in several eukaryotic cell types including the skeletal muscle cell line C2C12, the T lymphocyte cell line Jurkat, the vaccinia/T7, the fibroblastic cell line CHO and Sf9/baculovirus systems 70. In each case, although the transfection procedure appeared successful, no current with IsK characteristics could be recorded from these cells. Bilayer reconstitution experiments using membranes highly enriched in IsK protein were reported as inconsistent with IsK being a channel located in membranes of intracellular organelles 70. Other attempts to express the IsK protein as a K+ channel in lipid bilayers have also failed 71 . Note, however that recombinant phospholemman proteins can support lindependent' anion channel formation in lipid bilayers 72 (see Miscellaneous information, 54-55). Note: minK/KvLQTl co-expression yields large current in CHO, Sf9 and COS cells (see paragraph 54-31-01 and Fig 8).

B t""t'"

~ CJ1 ~

Co-existence of lfunctional' and Inon-functional' forms of IsK in the plasma membrane of oocytes

54-31-18: When Xenopus oocytes are injected with up to 50ng amounts of 'epitope-tagged' minK mRNA (see Predicted protein topography, 54-30 and the [PDTM], Fig. 6), the levels of surface protein are proportional to the amount of injected mRNA. Notably, however, the amplitude of the minK current recorded in the oocyte system saturates at 1 ng of injected mRNA. At mRNA levels above 1 ng, the kinetics of activation of the current differ in oocytes with high or low levels of minK RNA (activation is slower with higher levels of minK protein in the plasma membrane). Collectively, these and other results were taken to support a model in which minK protein forms functional potassium channels by association with a factor endogenous to the oocyte, with non-functional forms of minK also being present61 (see also Activation, 54-33, Kinetic model, 54-38 and Channel modulation, 54-44).

Alternative models

54-31-19: Essential features of alternative models in which the IsK protein (designated as IsK, Cl) acts as 'a potent activator of endogenous and otherwise silent K+ or Cl- channels' are illustrated in ref. 5 . These models invoke a low-affinity interaction between IsK, Cl and phospholemman (i.e dependent on high IsK cRNA concentrations - but note evidence for 'independent' phospholemman channel formation, this table, above). Conversely, these models invoke a high-affinity interaction between IsK subunits and endogenous K+ channels (i.e. sustainable with low IsK cRNA concentrations).

Note: 1. For comparative purposes, see also Phenotypic expression, 54-14.

II

~

= t '+

~

CJ1 ~

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_5_4_ _

Protein interactions minK as a regulatory protein underlying IK,s in association with KvLQTl protein 54-31-01: There has been much debate regarding the ability or inability of minK-like proteins to form integral, ion-selective channels when expressed alone. Evidence for the hypothesis stating that minK/lsK is a regulator (activator) of endogenous channels through protein interactions5 is summarized in Table 4. Significantly, it has been shown that short synthetic peptides from the cytoplasmic and extracellular domains of minK protein are sufficient to induce K+ and CI- channel activities in Xenopus oocytes 68 • Furthermore, the type of heterologous cell used for minK expression appears to be critical for 'latent channel' activation. To date, no mammalian heterologous cell type has been able to support expression of minK alone (ibid.), although Xenopus oocytes produce robust current when injected with minK cRNA (see various fields of this entry). In late 1996, two papers appeared simultaneouslr,3 supporting a hypothesis that minK and KvLQTl protein subunits co-assemble to form channels conducting current 'nearidentical' to native cardiac IK,s.. In keeping with previous observations (see above) minK could be expressed 'alone' in Xenopus oocytes but not in the mammalian cell lines COS2 or CH03. Furthermore, in keeping with the 'latent channel' hypothesis; (i) a eDNA encoding a Xenopus homologue of KvLQTl was retrieved from an oocyte c:DNA library3; (ii) minK produced a large potentiation in KvLQTl current following mammalian cell expression and (iii) co-expression of FLAG epitope-taggedt versions of KvLQTl coexpressed with minK in the Sf9 (insect epithelial) cellline2 showed associated molecular species consistent with minK-KvLQTl protein interactions in vivo. Together with earlier observations linking KvLQT1 mutations with the most common form of inherited long QT syndrome (LQT1, see Chromosomal location under VLC K eag/elk/erg, 46-18) the minK/LQTl studies2,,3 therefore implicated dysfunction of native IK,s in LQT1, and provided a model for IK,s as an important 'target' for class mt antiarrhythmic drugs. Representative traces showing the effect of minK/KvLQT1 co-

expression in mammalian cells (i.e. in th.e absence of tendogenous or latent' subunits) are shown in Fig. 8.

Functional interactions of minK with ion pump/transporter proteins 54-31-20: Xenopus oocytes expressing slowly activating IsK channels

superfused with the nitroso-donort S-nitrosocysteine (SNaC, see Resource C) result in an increase of IsK which is greatly enhanced when the amino acid exchanger rBAT 73 is co-expressed (for further details on SNaC modulation and the role of co-expressed rBAT facilitating SNac transport into the oocyte, see Channel modulation, 54-44). In kidney epithelial cells, possible functional interactions with the Na+/K+-ATPase and Na+-coupled transport systems have been described12 (see also Fig. 2 under Phenotypic expression, 54-14). Notes: 1. MinK expression in non-excitable epithelia is associated with permeation of K+ ions13 into both the interstitial and luminal spaces (see Fig. 2 under Phenotypic expression, 54-14 and Protein

l_e_n_try_5_4

_

(b) __ K V LQT1 + hminK C 1.0l-.-KvLQT1

~

o.sl

B 0.6

.

Q)

:E; 0.4 co

~ 0.2 04J=lJ:-+.......--.,-------r--~---.

-60 -40 -20 0 mV

20

40

(e) Kv LQT1 + hminK hminK

K V LQT1

~] 2s

Figure 8. KvLQTl and human minK co-expression in CHO cells induces a current lnearly identical' to cardiac IK,s. (a) Currents recorded from CHO

cells transfected with KvLQTl alone. (b) Normalized isochronal activation curves for KvLQTl and KvLQTl + minK. (c)-(e) Currents from cells transfected as shown, illustrating potentiating effect of minK + KvLQTl interaction. (Reproduced with permission from Sanguinetti (1996) Nature 384: 80-3.) (From 54-31-01)

distribution, 54-15). 2. Structure-function studies of the single putative (0helical) membrane-spanning domain of IsK have indicated that one face of the helix may represent a region of subunit association62 (see Domain functions, 54-29).

minK antisense oligonucleotides reduce I KR (rapid component) amplitudes in AT-l cells -54-31-21: In atrial tumour myocyte lines (AT-I cells) minK mRNA is detectable, but no IK,s can be recorded10 (see also Cloning resource, 54-10). AT-l cells exposed to antisense oligonucleotides (targeting the 5' translation start site of the minK cDNA cloned from an AT-l library) exhibit a Isignificantly reduced' I Kr amplitude (= rapid component) compared with 'sense' and 'medium-only' controls. I Kr activation, rectification, deactivation and sensitivity to the blocker dofetilide were reported as 'unaffected'. Note: 1. Two different antisense oligonucleotides produced the same effect on I Kr amplitude without effect on cell size or on L- or T-type Ca2+ currents native to AT-l cells10. 2. Compare 'equivalence' of I Kr to erg-subfamily currents (see VLC K eag/elk/erg, entry 46) and see effect of minK/LQTl co-expression, described under Phenotypic expression, 54-14 and this field, above.

II

Table 5. Regulatory properties of heterologously expressed minK (IsK) proteins potentially involving phosphorylation. For comparative purposes, phosphorylation properties of the native cardiac current IK,s are also listed, together with reported species differences, and notable effects in selected mutants (all denoted by underlined prefix:). MinK/IsK data are applicable to heterologous expression in Xenopus oocytes, unless otherwise indicated. (From 54-32-01) Effect/regulator/mechanism

Characteristics, species dependence, references (see note 1)

Inhibitory effects of PKC and PKC activators e.g PDD, PMA. (inhibition by protein phosphorylation)

Rat, kidney minK wild-type: Heterologously expressed minK protein activities are markedly inhibited following microinjection of purified protein kinase C (PKC). The PKC-activators PDD (phorboI12,13-didecanoate, a phorbol ester, rvSOnM) and OAG (1-oleoyl-2-acetyl-rac-glycerol, a diacylglycerol analogue, rv 10 JlM) also inhibit the current 74 . Protein kinase C phosphorylation of the minK channel shifts its voltage-dependence of activation 74 (see also Voltage sensitivity, 54-42). Mouse, cardiac minK wild-type: Inhibitory effects have been observed for PMA (12-phorbol myristate 13-acetate, a phorbol ester t ) showing a 70% inhibition (at 0.1 JlM) with OAG also inhibiting minKs. Rat, kidney minK Ser103Ala mutant: Inhibition of the current is not seen in S103A mutants (a marginal enhancement occurs with PDD, although other properties of the current remain unchanged). This has been taken to indicate that inhibition results from direct phosphorylation of the rat kidney protein at Ser103 74 (see the {PDTM}, Fig. 6). In the rat kidney minK, Ser103 lies within a sequence that resembles the PKC consensus site Arg-Val-Leu-Glu-Ser-Phe-Arg. Substitution of Ser103Ala also prevents the inhibition of minK by PDD and OAG, whereas Ser7SAla has no significant effect upon the IsK current. Guinea-pig, cardiac rninK/IK,s: In 'sharp contrast' to the rat and mouse IsK currents (above), activation of protein kinase C (PKC) with phorbol esters moderately increases the amplitude of the guinea-pig IsK current44, analogous to its effects on the endogenous IK,s current in guinea-pig cardiac myocytes. Guinea-pig, cardiac minK mutant: Mutagenesis of the guinea-pig IsK sequence at four residues in the cytoplasmic tail can alter the phenotype of the current response (i.e. from a small PKC-dependent enhancement to PKC inhibition, the latter being characteristic of rat and mouse IsK (e.g. PKC phosphorylation at Ser102 decreases IsK current amplitude)44.

Species difference: Wild-type guinea-pig cardiac IsK/IK,s enhanced by PKC activators. For an illustration of species sequence differences at this site, see the {PDTM}, Fig. 6.

('l)

= ~ c.n ~

Preliminary evidence for protein kinase C mediation of minK channel down-regulation following activation of apical P2U purinoceptors in K+ -secretory epithelial cells of the inner ear has appeared 75 • Stimulatory effects of PKC inhibitors, Rat, kidney minK: The protein kinase C inhibitor staurosporine (rv3 JlM) prevents minK e.g staurosporine inhibition under conditions similar to those described above (e.g. see refs8 47 74). 7

Stimulatory effects of PKA and PKA activators (mechanism unclear)

Single PKA consensus site eliminations have no apparent effect.

Small stimulatory effects of microinjected Ca2 +

(see Channel modulation, 54-44)

7

Rat, uterine minK wild-type: In a heterologous expression system, the amplitude of minK currents can be substantially increased by agents that raise cAMP levels and decreased by treatments that lower cAMP levels 76 (see also Channel modulation, 54-44). Pre-injection of a protein inhibitor of the cAMP-dependent protein kinase blocks the effect of increased cAMP levels, without any change in voltage dependence or kinetics of the channel 76. A form of parathyroid hormone (PTH) regulation in oocytes can be mimicked by activators of protein kinase A (PKA) (see Receptor/transducer interactions, 54-49). Mutations that eliminate the only PKA consensus site on the minK protein do not block the effects of kinase activation. The basic mechanism of PKA regulation of minK proteins thus remains unclear. Hypothetically, these effects might be due to actions at other proteins that regulate minK current amplitude; alternatively, minK proteins may be inserted into or removed from the membrane in 'minK-loaded' vesicles in response to changes in PKA activity 76. Note: PKA has been shown to increase the membrane surface area of oocytes, consistent with this hypothesis. Guinea-pig, cardiac 11K,s: 'Positively regulated' by PKA in response to ,a-adrenergic stimulation 77 (for a review, see ref.78, see also Receptor/transducer interactions, 54-48). Mouse, cardiac minK: minK cloned from neonatal mouse heart is enhanced twofold by microinjected Ca2 + and Ca2 +Icalmodulin-dependent protein kinase II; this stimulation is reversed by the calmodulin antagonist W_7 8 (but see possible blocking effects of

calmodulin antagonists, this table, below).

II

Elevation of minK current amplitude by raised [Ca2 +h possibly through the activation of an associated calcium-activated protein kinase21,47.

(D

l:S

M-

~ tTl ~

II

Table 5. Continued Effect/regulator/mechanism

Characteristics, species dependence, references (see note 1)

Supplementary note on CamKII motifs

The CAM-kinase II motif region in mouse and rat (residues 35-41, SQLRDDSK) differ from the human sequence (SPRSSDGK) and later analyses have indicated an extracellular location for the N-terminal. The apparent functional modulation by CaM kinase II would require its modulatory site to be intracellular which is incompatible with the established single-transmembrane model shown in the PDTM (Fig. 6).

Inhibitory effects of calmodulin antagonists, e.g trifluoperazine, chlorpromazine, W-7: hypothetical, probably by direct channel blockade (see ref. 79 and Blockers, 54-43)

Inhibition of human IsK channels in oocytes by calmodulin antagonists The calmodulint antagonists trifluoperazine (10-[3-(4-methyl-1-piperazinyl)-propyl]-2(trifluomethyl)-lOH-phenothiazine), chlorpromazine (2-chloro-1 0-(dimethylaminopropyl)phenothiazine) and W-7 (N-(6-aminohexyl)-5-chloro-1-naphthalen-sufonamide) inhibit depolarization-activated IsK channels (ECso between 70 and 100 JlM). Notes: (i) These calmodulin antagonists inhibit IsK at both physiological and enhanced [Ca2+Ji (see Channel modulation, 54-44). (ii) None of the antagonists abolish the inhibitory effects of A23187 (calcimycin) or hypotonic t extracellular fluid on human IsK (see Channel modulation, 54-44).

Note: 1. For potential phosphorylation-independent effects, see Channel modulation, 54-44.

g t'+

~ CI1 ~

1'--_e_n_t_ry_S4

_

Protein phosphorylation 'Down-regulation' of minK activities by PKC-dependent protein phosphorylation 54-32-01: A number of distinct mechanisms for second messenger-dependent phosphoregulation of minK proteins have been proposed (summarized in Table 5). Several of these collective properties show remarkable similarity to those observed for native IK,s components in heart (see Phenotypic expression, 54-14). While minK regulation (inhibition) by protein kinase C isoforms is established, the mechanisms underlying protein kinase A and Ca2+ modulation are still unclear. Analysis of phosphoregulatory patterns for minK proteins are complicated by various 'species-specific' effects, although a number of these have been explained by specific differences in primary structure amongst homologous isoforms (see the [PDTM), Fig. 6}. A further 'complicating' factor for analysis of minK phosphoregulatory properties is its proposed functional interactions with 'latent' channel proteins (summarized under Protein interactions, 54-31).

ELECTROPHYSIOLOGY

Activation Note: All results generated in the Xenopus oocyte expression system should be interpreted in the light of findings described under Protein interactions, 54-31 and Selectivity, 54-40.

Depolarization activates minK (IsK) channels very slowly, possibly by subunit aggregation 54-33-01: minK protein expression in Xenopus oocytes gives rise to K+ currents which slowly activate when the membrane is depolarized to potentials greater than -SOmV (activation occurs at least two orders of magnitude slower than any other ion channel, see example in Fig. 9). Activation thus takes place in the order of 'seconds-to-minutes' and often does not reach a steady state (even after several minutes, hence isochronal t currents, rather than steady-statet currents are generally stated in comparative studies42 ). Activation is sigmoidal in response to voltage steps, with a noticable delay occurring at higher depolarizations (see also next paragraph). Studies employing chemical cross-linking during activation80 (see also Table 7 under Channel modulation, 54-44) have suggested that a major conformational change occurs during minK channel gating (which can be stabilized by cross-linking agents). These findings are also consistent with models in which minK channels activate by voltage-dependent subunit aggregation (see also Protein interactions, 54-31). Alterations of gating parameters by neutral substitutions of Leu52 within the transmembrane domain of IsK are shown in Fig. 7 under Domain functions, 54-29.

Voltage-dependent activation of epithelial IsK proteins in situ 54-33-02: A proposal for activation of IsK-dependent K+ transport (efflux) in secretory epithelia due to the depolarizing effects of Na+ entry (following

111

_'---

e_n_try_5_4_

activation of Na+/sugar or amino acid co-transport systems) is outlined in Fig. 2 and associated text under Phenotypic expression, 54-14. See also Protein interactions, 54-31.

Gating of 1sK dependent on the amount of mRNA injected into Xenopus oocytes 54-33-03: Some studies (e.g. refs. 61 ,84) have shown variable activation kinetics of IsK dependent upon the amount of mRNA injected into Xenopus oocytes. Injection of larger amounts of IsK mRNA (e.g. in the 10-50ng range, versus the 100pg to 1ng range) result in slower activation kinetics (larger tau values) with a longer initial delays during activation. Similar variabilities in activation kinetics can occur with time following single injections of mRNA81 (see also Protein interactions, 54-31 and Kinetic model, 54-38). In contrast to activation, deactivation t of minK does not exhibit the same dependence upon the amount of RNA injected61 . Comparative note: Functional changes dependent upon expression of variable amounts of Kv channel cRNAs in oocytes have also been reported (for brief accounts, see the fields Current type, Inactivation and Blockers under VLC K Kv-Shak, 48-34, 48-37 and 48-43 respectively).

tSpecies differences' in activation/deactivation kinetics 54-33-04: The first 30 s of activation during depolarizations to potentials between -10 and +40mV are best described by a tri-exponential rise function for each of the human, mouse or rat IsK proteins in Xenopus oocytes59. Notably, however, the activation rates of human 18K channels have been reported as 'significantly faster' than those for either !nouse or rat IsK proteins following expression in Xenopus oocytes59 (examples of time constants required for adequate fitting of initial delay and onset: are shown in the legend to Fig 9). Deactivationt of IsK currents is also slow (requiring seconds for full relaxation t) and deactivation rates also show some 'species dependence'. Human IsK currents deactivate more rapidly than the rodent currents, although the deactivation kinetics for each of the species variants are best described by a biexponential decay function 59. Differences in deactivation rate may be attributable to acidic residues in the N-termini of the species variants82.

Electrophysiological parameters for 18K derived from isochronal currents Note: Half-points of steady-state (sic.) activation curves cannot be precisely determined since extremely long times (e.g. 200 s) are required for the current to saturate during step depolarizations42 . Using 30 s depolarizations, however, the slope factor (k) and midpoint (V 1/ 2 ) of the conductance-voltage curves have been estimated at k = 12, Vl~2 = -5 mV (for rat/mouse IsK) and k = 9, V 1/ 2 = -11 mV (for human IsK)8,21, "4.

Current-voltage relation Note: All results generated in the Xenopus oocyte expression system should be interpreted in the light of findings described under Protein interactions, 54-31 and Selectivity, 54-40.

III

1'--_e_n_t_ry_54

(a)

_

50~

(c)

m.v_---_

1600

.-60 ..

1200

(b)

280nAL 400ms

c(

c:

'E 800 CD

t: :::J 0

400

0 -60

280nAL 4s

·40

20 0 ·20 Voltage. mV

40

60

Figure 9. Typical activation of 1sK/minK currents following heterologous expression in Xenopus oocytes*. Records were obtained following microinjection of 100ng rat 1sK cRNA, analysed 3 days post-injection. (a) Rat 1sK currents elicited by 2 second depolarizations from a holding potential -60mV to +50mV in 15mV steps. (b) Rat 1sK currents elicited by 20 second depolarizations with the same voltage protocol (note different time scale). (c) Current-voltage relations for the currents in (a) • and currents in (b) •. Comparative notes: 1. Currents could be well-fit (r = 0.998) by an exponential model of activation t containing a minimum of three independent time constants (Tactl = 0.16 s, Tact2 = 2.1 s, Tact3 = 20 s) (see text). 2. Outward tail currents are generally observed upon repolarization (see (b)), which follow a similar time course for activation, requiring many seconds to decay to baseline. 3. Native 1K ,s recorded from guinea-pig ventricular myocytes6 show similar activation and deactivation kinetics to those described here. Under the same conditions as above, guinea-pig 1K ,s can also be well-fit (r = 0.996) by the triexponential model yielding similar time constants (Tactl = 0.36 s, Tac t2 = 1.5 s, Tact3 = 20 S)6. * Note: These and other results generated in the Xenopus oocyte expression system should be interpreted in the light of findings described under Protein interactions (field 54-31) and Selectivity (field 54-40). (Reproduced with permission from Folander (1990) Proc Natl Acad Sci USA 87: 2975-9.) (From 54-33-01)

54-35-01: I-V relations derived for rat IsK are shown in Fig. 9c under Activation, 54-33. Current/voltage relationships of the human and rat IsK currents may 'differ significantly' (e.g. with greater depolarizations required for activation of the human channe159). For other apparent cspecies differences', e.g. in phos-

phorylation patterns, activation/deactivation kinetics, voltage dependence, and La3+ block, see Protein phosphorylation, 54-32; Activation, 54-33 and Blockers, 54-43. For the effect of the minK mutant S69A inducing shifts in the 1-V relationship, see Predicted protein topography, 54-30 and also Fig. 7.

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _e_n_try_5_4------1

54-35-02: Open channel current-voltage relationships determined from macroscopic T minK currents in Xenopus oocytes show rat and human minK current has a mild inwardly rectifyingt characteristic, passing inward current 'at least 20-fold' more readily than outward current83 . Note: A similar rectification property has been reported for the slow (IK,s) component of cardiac delayed rectifier (see also Phenotypic expression, 54-14).

Inactivation 54-37-01: Following their slow, time-dependent activation, currents associated with heterologously expressed minK protein are characteristically non-inactivatingt under maintained depolarization. For deactivationt characteristics, see Activation, 54-33.

Kinetic model 54-38-01: Two kinetic schemes accounting for variable activation kinetics of IsK dependent upon the amount of mRNA injected into Xenopus oocytes have been presented84 (see also Activation, 54-33).

Selectivity Note: Results generated in the Xenopus oocyte expression system should also be interpreted in the light of findings described in paragraph 54-40-01 and those under Protein interactions, 54-31.

Distinct selectivities induced by minK peptides - direct evidence for regulatory functions 54-40-01: Short (27-34 aa) synthetic peptides derived from C- and N-termini of IsK/minK have been determined 'sufficient' to activate (respectively) slow K+ or Cl--selective currents in Xenopus oocytes68 . 'Peptide-induced' currents display biophysical and pharmacological characteristics similar to those exhibited by the native IsK protein. Together with previous (less direct) observations (Table 4 under Protein interactions, 54-31) these data were taken to provide further evidence that IsK/minK represents a regulatory subunit of pre-existing 'silent' (latent) channel complexes rather than a channel per see Figure 10 illustrates typical current traces, I-V relations and selectivity data derived for oocytes microinjected with C-27 IsK peptide (inducing a depolarization-activated K+ current) and perfused with N-34 IsK peptide (inducing a hyperpolarization-activated CI- current). The N-34 and C-27 peptide sequences are also underlined on the sequence alignment (Fig. 4) shown under Encoding, 54-19.

Comparison of IsK-induced channel selectivities with native cardiac l K ,s components 54-40-02: Expression of wild-type minK cDNA in oocytes gives rise to pores exhibiting a permeability sequence typical of 'virtually all' known K+channels: K+ > Rb+ > NHt> Cs+» Na+, Li+ 54,85,86 with tailt current reversal potentialst shifting approx. 58 mV per decade change in external

l_e_n_t_ry_5_4

----'_

K+ 53 (but see notable absence of K+ channel Isignature sequences' and effects of endogenous Cl- current activation under Protein interactions, 54-31). Some variability in permeability ratios to permeant cations have been reported between heterologously expressed IsK and native IK,s (e.g. compare the PK/Pcs of 16.4 determined for native guinea-pig IK,s 87 to 6.24 determined for rat IsK in oocytes; these may be associated with noted sequence differences between the species). Note: Hypo-osmotically induced K+ transport pathways in vestibular dark cells from the gerbil inner ear have ion selectivity and multiple blocker insensitivity consistent with involvement of minK (for details, see ref.88; see also Phenotypic expression, 54-14, especially Table 2).

Apparent MinK selectivity functions investigated by site-directed mutagenesis 54-40-03: Site-directed mutagenesis studies in oocytes with a synthetic minK sequence54 show that K+ selectivity can be altered by specific mutations at Phe55 (for location, see the [PDTM), Fig. 6 and Fig. 7 under Domain functions, 54-29). For example, the mutant F55T yields minK channels that are rv3.5-fold more permeable to Cs+ and NHt (reducing PK/Pcs to 4.6 - see previous paragraph and Domain functions, 54_29)53,54. Note: Despite the indication that minK subunits contribute 'directly' to selectivity properties54, it is still unclear whether these effects are due to formation of homomultimeric minK ion channels alone or necessarily require interaction with one or more endogenous ion channels (see Protein interactions, 54-31). In a later study89 also in oocytes, mutations in a specific minK region (spanning approx. £44 to 148) altering sensitivity influencing external blockade by tetraethylammonium and methanethiosulfonate ethylsulfonate were used to support the view that minK is directly involved in forming a K+ -selective ion conduction pathway89.

Single-channel data 54-41-01: Unitary conductances associated with minK/lsK proteins in oocytes (see Note under Activation, 54-33) are relatively low and have sometimes been described as 'below the limits of detection' (i.e.

Ion Channel Factsbook, Volume 4: Voltage-Gated Channels

Ion Channel Factsbook I: Extracellular Ligand-Gated Channels (Factsbook)

Ion Channel Factsbook I: Extracellular Ligand-Gated Channels (Factsbook)

Mechanosensitive Ion Channels

Sensing with Ion Channels

Ion Channels: Methods and Protocols

Ion channels Methods and Protocols

Pumps Transporters and Ion Channels

Ion Channels and Disease: Channelopathies

Ion Channels and Their Inhibitors

Ion Channel Diseases, Volume 63 (Advances in Genetics)

The HLA FactsBook (Factsbook)

Ion Channels (Topics in Medicinal Chemistry, 3)

Voltage-Gated Ion Channels as Drug Targets

Ion Channels and Plant Stress Responses

The Nuclear Receptor FactsBook (Factsbook)

T-Cell Receptor FactsBook (FACTSBOOK SERIES)

Channel Surfing: Riding the Waves of Channels to Profitable Trading

Channel

Voltage-Sensitive Ion Channels: Biophysics of Molecular Excitability

Biological Membrane Ion Channels. Dynamics, Structure and Applns

Sensing with Ion Channels (Springer Series in Biophysics)

Ion Channel Localization (Methods in Pharmacology and Toxicology)

Reality Show, Tome 4 : Reconquista Channel

Advances in Gas Phase Ion Chemistry, Volume 4 (Advances in Gas Phase Ion Chemistry)

Ion Channel Regulation (Advances in Second Messenger and Phosphoprotein Research)

Molecular and Cellular Insights to Ion Channel Biology

Ion

ion

Ion

Ion

Ion Channel Factsbook, Volume 4: Voltage-Gated Channels

Ion Channel Factsbook I: Extracellular Ligand-Gated Channels (Factsbook)

Ion Channel Factsbook I: Extracellular Ligand-Gated Channels (Factsbook)

Mechanosensitive Ion Channels

Sensing with Ion Channels

Ion Channels: Methods and Protocols

Ion channels Methods and Protocols

Pumps Transporters and Ion Channels

Ion Channels and Disease: Channelopathies

Ion Channels and Their Inhibitors

Ion Channel Diseases, Volume 63 (Advances in Genetics)

The HLA FactsBook (Factsbook)

Ion Channels (Topics in Medicinal Chemistry, 3)

Voltage-Gated Ion Channels as Drug Targets

Ion Channels and Plant Stress Responses

The Nuclear Receptor FactsBook (Factsbook)

T-Cell Receptor FactsBook (FACTSBOOK SERIES)

Channel Surfing: Riding the Waves of Channels to Profitable Trading

Channel

Voltage-Sensitive Ion Channels: Biophysics of Molecular Excitability

Biological Membrane Ion Channels. Dynamics, Structure and Applns

Sensing with Ion Channels (Springer Series in Biophysics)

Ion Channel Localization (Methods in Pharmacology and Toxicology)

Reality Show, Tome 4 : Reconquista Channel

Advances in Gas Phase Ion Chemistry, Volume 4 (Advances in Gas Phase Ion Chemistry)

Ion Channel Regulation (Advances in Second Messenger and Phosphoprotein Research)

Molecular and Cellular Insights to Ion Channel Biology

Ion

ion

Ion

Ion

Recommend Documents