rectifying potassium channels

A thesis submitted to the University of Leicester for the deg Doctor of Philosophy 2003

Gayle Martine Passmore BSc (Leicester) Department of Cell Physiology & Pharmacology University of Leicester UMI Number: U4955B9

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion.

Dissertation Publishing

UMI U495539 Published by ProQuest LLC 2013. Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under 17, Code.

ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 For Nan Abstract

The patch clamp technique was used to investigate the effect of mutations in the P-region on permeation characteristics and ionic selectivity in members of the Kir2.0 subfamily. Kjr2.2 exhibits electrophysiological characteristics similar to those of Kir2.1 though there are differences in both unitary conductance and the kinetics of Ba2+ blockage. The P-region of these channels is virtually identical with the exception that leucine (L) at position 148 in Kjr2.2 is replaced by phenylalanine (F) in Kir2.1. The effects of mutating L148 to phenylalanine in Kjr2.2 and F147 to leucine in Kir2.1 on unitary conductance and channel sensitivity to Ba2+ were investigated. Neither mutation altered unitary conductance from that seen in the wild-type channel. However, mutation L148F in Kjr2.2 reduced the association rate constant for Ba2+ blockage without affecting affinity. In contrast, mutation F147L in Kjr2.1 increased channel affinity for Ba2+ without affecting the association rate constant. Thus, residues outside the P-region are responsible for the difference in unitary conductance and some of the differences in Ba2+ block in Kjr2.2 and Kir2.1.

The effects on ionic selectivity of substituting single amino acid residues at position 143 in the P-region of Kir2.1 were also investigated. Substitution of isoleucine by hydrophobic residues such as valine (I143V) and leucine (I143L) raised the relative Rb+ permeability, whilst substitution of the more hydrophilic residue threonine (I143T) enhanced K+ selectivity. Two further mutants, I143C and I143S, failed to yield currents, but could be rescued by bathing cells in extracellular solution containing lOmM dithiothreitol (DTT). The permeability ratios were then similar to wild-type. The rescue of mutant channels by DTT suggests that the pore of Kjr channels may undergo conformational changes. In vitro translation studies suggest that channel function is rescued through the disruption of an intra­ subunit disulphide bond. Acknowledgements

First, and foremost, I would like to express my sincere thanks to my supervisor Peter Stanfield for his advice and assistance throughout the duration of my PhD, and also for his patience during the latter three years. Also, thanks to Dr Noel Davies for his help with Tracan and general problem solving, Dr Ian Ashmole for his help and ideas with the molecular biology and biochemistry, Dr Mike Sutcliffe for his Kn channel modelling and Drs Blair Grubb and Robert Norman for their help with Western blotting. I would also like to thank Drs Chris Abrams, Gareth Thompson, Michael Holland and Christina Skipper who gave me much needed advice in the laboratory on a daily basis.

The members of my new lab at UCL have provided me with much encouragement and I would particularly like to thank Professor David Brown and Dr Steve Marsh for their continued support over the last three years.

Last, but by no means least, I would like to thank my parents for all of their loving support over the years, without which it would have been impossible to achieve this goal.

iv Table of Contents

Title Page i Dedication ii Abstract iii Acknowledgements iv

Table of Contents V List of Figures xi List of Tables xiii Abbreviations ?

Chapter 1 - Introduction 1

1.0. General Introduction 1 1.1. Ion Channels 4 1.1.1. A historical perspective of ion channels 4 1.1.2. Classification of ion channels 7 1.1.3. A role for potassium channels 7 1.2. diversity 9 1.2.1. Early potassium channel cloning studies 11 1.2.2. The three classes of potassium channel 13 1.2.3. Six-transmembrane domain potassium channels 13 1.2.3a. -related potassium channels 13 Gating 14 Inactivation 15 13-subunits 20 1.2.3b. Calcium-activated potassium channels (KCa) 20 BKca channels 20 SKca channels 22 IKca channels 22 1.2.3c. Eag-like potassium channels 23 1.2.3d. KCNQ potassium channel 24 1.2.4. Two-transmembrane domain potassium channels 25 1.2.4a. Kir 1.0 subfamily 26 Tissue distribution of Kir1.0 subfamily channels 26 Properties of Kir1.0 subfamily channels 27 Role o f Kirl . 0 subfamily channels 28 1.2.4b. Kir2.0 subfamily 28 Kir2J 28 Kir2.2 30 Kir23 31 Kir2A 32 Kir2.0 channels interact with anchoring 32 Modulation of Kir2.0 subfamily channels 33 1.2.4c. Kir3.0 subfamily 34 Tissue distribution 34

Kaoi is a heteromultimer o f Kir3.1 and Ktr3.4 35 Neuronal G--gated inward rectifier potassium Channels 35 Modulation ofKir3.0 subfamily channels 36 1.2.4d. Kir4.0 subfamily 37 Biophysical properties of KiA.O subfamily channels 37 Distribution and role of Ki/l-0 subfamily channels 38 KiA-0 channels interact with anchoring proteins 39 Heteromeric Ki/l.x channels display distinct biophysical properties 39 1.2.4e. Kir5.0 subfamily 40 Distribution and role o f Kir5.1 /PSD-95 40 Kir5.1 heteromers 40 Distribution and role of heteromeric Ki/l.x-Kir5.1 Channels 41 1.2.4f. Kir6.0 subfamily 41

Role o f K a t p channels 42 Properties ofKi,6.0 channels 42 Sulphonylurea receptors 43 Structure and properties ofKATP channels 44 1.2.4g. Kir7.0 subfamily 44 Biophysical properties ofKir7.1 45 Distribution and role of Kir7.1 45 1.2.4h. Drosophila Kirs 46 1.2.4i. A family of Kir channels in prokaryotes 46 1.2.5. Two-pore domain potassium channels 46 1.3. Assembly and stoichiometry of potassium channels 48

VI 1.3.1. Kv channels assemble as tetramers 48 1.3.2. The T1 domain is required for tetramerization of Kv channels 49 1.3.3. Kir channels assemble as tetramers 50 1.3.4. channels can form heteromers 50 1.3.5. Structural determinants of Kjr channel assembly 51 1.4. Inward rectification 52 1.4.1. Properties of inward rectifier potassium channels 53 1.4.1a. Inward rectification depends on Vm and the external K+ concentration 53 1.4.1b. The conductance is proportional to the square root of [K+]0 55 1.4.1c. Activation 55 1.4.Id. Inactivation 55 1.4. le. Block by external cations 56 1.4. If. Anomalous mole fraction effects 57 1.4.2. Mechanism of inward rectification 57 1.4.2a. The K-activated K-channel model 58 1.4.2b. The blocking particle model 59 1.4.2c. M g 2+block 60 1.4.2d. Polyamine block 60 1.4.2e. Intrinsic gating 63 1.4.2f. Do Kir channels exhibit intrinsic rectification? 66 1.4.3. Structural determinants of inward rectification 67 1.5. Physiological importance of Kir channels 69 1.5.1. K^ channels underlie K+ homeostasis in the kidney 69 1.5.2. K^ channels underlie K+-induced vasodilation 70 1.5.3. K^ channels regulate cardiac action potentials 70 1.5.4. Kir channels help regulate the heartbeat 71 1.5.5. K^ channels regulate blood glucose levels 71 1.5.6. Kir channels are important in development 73 1.6. Selectivity 75 1.6.1. Structural determinants of K+ selectivity in Kv channels 75 1.6.2. Theories of selectivity 77 1.6.3. Molecular basis of selectivity 78 1.7. Aims 81

vii Chapter 2 - Materials & Methods 82

2.1. Molecular biology 82 2.1.1. Cloning and site-directed mutagenesis 82 2.1.2. Purification of plasmid DNA 83 2.2. Cell culture 83 2.2.1. Cell culture reagents and plastic-ware 83 2.2.2. Incubator 83 2.2.3. Cell lines 84 2.2.4. Growth of cells in culture 84 2.2.5. Chinese hamster ovary (CHO) cells 84 2.2.5a. Routine growth and maintenance of CHO cells 85 2.2.5b. Subculture of CHO cells 85 2.2.5c. Preparation of CHO cells for electrophysiological recording 86 2.2.6. Murine erythroleukaemia (MEL) cells 86 2.2.6a. Routine growth and maintenance of MEL cells 87 2.2.6b. Subculture of MEL cells 87 2.2.6c. Induction of MEL cells 87 2.2.6d. Preparation of MEL cells for electrophysiological recording 87 2.2.7. Storage of cell lines 88 2.2.8. Haemocytometer cell counts 89 2.3. Transfection Methods 89 2.3.1. Lipid-mediated transfection 90 2.3.2. Transient transfection of CHO cells using PerFect Lipid™ 90 2.3.3. Transient transfection of CHO cells using FuGENE™ 6 91 2.3.4. Levels of expression 91 2.3.4a.The amount of cDNA used for PerFect Lipid transfections was reduced 92 2.3.4b.The incubation time for Perfect Lipid transfections was reduced 92 2.3.4c. A weaker promoter was used to drive expression 92 2.3.4d. Different methods for the purification of cDNA were utilized 94 2.4.3e. Cells were patched at different intervals post transfection 94 2.3.5. Stable transfection of MEL cells using electroporation 94 2.4. Electrophysiology 96 2.4.1. Principles of patch clamp 96 2.4.2. Gigaohm seal 97

viii 2.4.3. Setup and recording chamber 98 2.4.4. Pipette and reference electrodes 98 2.4.5. Solutions 100 2.4.6. Perfusion system 101 2.4.7. Manufacture of pipettes 101 2.4.8. Pipette diameter and resistance 102 2.4.9. Patch clamp amplifier 102 2.4.10. Procedure for obtaining a giga-ohm seal 106 2.4.11. Capacity transients 107 2.4.12. Recording configurations 107 2.4.13. Whole cell capacitance 110 2.4.14. Series resistance compensation 110 2.4.15. Junction potentials 111 2.4.16. Data acquisition 114 2.4.17. Data Analysis 115 2.5. Detection Methods For Protein Expression 117 2.5.1. Western blotting 117 2.5.1a. Sample preparation 118 2.5.1b. Gel electrophoresis 118 2.5. lc. Transfer of proteins 119 2.5.Id. Western blotting procedure 120 2.5.le. Protein detection 120 2.5.2. In vitro transcription translation 121 2.512a. The coupled transcription-translation assay 121 2.5.2b. Gel electrophoresis 122 2.5.2c. Analysis of translation products 123

Chapter 3 - An electrophysiological characterization of the strong inwardly rectifying K+ channel Ku-2.2 124

3.1. Introduction 124 3.2. Methods 125 3.3. Results 126 3.3.1. Kir2.2 expresses inward rectifier potassium channels in MEL and CHO cells 126 3.3.2. Dependence of rectification on Vm and [K+]0 128 3.3.3. The conductance of Kir2.2 is dependent on [K+]0°38 130 3.3.4. Dependence of activation gating on Vm and [K+]0 133 3.3.5. Block of Kir2.2 by extracellular cations 135 3.3.6. Single channel current properties of Kir2.2 135 3.4. Discussion 138

Chapter 4 - A comparison of Ba2+ blockage in the murine K+ channels Kb.2.1 and Kir2.2 144

4.1. Introduction 144 4.2. Methods 146 4.3. Results 146 4.3.1. Barium blocks IQr2.2 in a concentration-, time- and voltage-dependent manner 146 4.3.2. Steady-state analysis of blockade of Kir2.2 by Ba2+ 148 4.3.3. Kinetic analysis of the macroscopic Kir2.2 current blockade by [Ba2+]Q 149 4.3.4. Block of single Kir2.2 currents by Ba2+ 153 4.3.5. Effects of mutation L148F in Kjr2.2 and mutation F147L in Kir2.1 on Ba2+ blockade 161 4.3.6. Effect of mutation L148F in Kir2.2 and mutation F147L in Kir2.1 on unitary conductance 166 4.4. Discussion 169

Chapter 5 - Mutation of Isoleucine 143 affects ionic selectivity and permeation in Kir2.1 177

5.1. Introduction 177 5.2. Methods 180 5.3. Results 180 5.3.1. Ionic selectivity of Kir2.1 wild-type 180 5.3.2. Ionic selectivity of 1143 mutants 182 5.3.3.1143C and I143S mutants are non-functional 182 5.3.4. Correlation between side chain properties and Rb+/K+ selectivity 184 5.3.5. Mutation of 1143 does not alter Rb+ blockage 187 5.3.6. Mutations of 1143 have little effect on channel macroscopic kinetics 187 5.3.7. Hypotheses for the mechanism underlying the loss of function in I143C and I143S 189 5.3.8. Mutation of 1143 to alanine gives functional channels but with a reduced conductance 191 5.3.9. I143C C149S and I143S C149S double mutants do not form functional channels 191 5.3.10. Evidence for an intra-subunit disulphide bond 194 5.4. Discussion 198

Chapter 6 - Summary 206

Chapter 7 - Bibliography 20?

xi List of Figures

Figure 1.1 Molecular architecture of potassium channels 10 Figure 1.2 Open and closed conformations of potassium channels 16 Figure 1.3 Inactivation of potassium channels 17 Figure 1.4 The ‘Cross-over’ effect 54 Figure 1.5 Mechanism of inward rectification 62 Figure 1.6 IQr channels regulate blood glucose levels 72 Figure 1.7 Features of ion permeation in potassium channels revealed by the crystal structure of KcsA 80

Figure 2.1 Levels of channel expression 93 Figure 2.2 Diagram to illustrate the recording chamber 99 Figure 2.3 Diagram of resistive headstage 104 Figure 2.4 Diagram of capacitive headstage 105 Figure 2.5 Diagram illustrating the different configurations of patch-clamping 109 Figure 2.6 Liquid junction potentials 112

Figure 3.1 Kjr2.2 expresses inward rectifier K+ channels in MEL and CHO cells 127 Figure 3.2 Conductance-voltage relationship for Kjr2.2 129 Figure 3.3 Kir2.2 wild-type currents are sensitive to changes in external K+ concentration 131 Figure 3.4 Kjr2.2 wild-type current are dependent on [K+]0. 132 Figure 3.5 Activation of Kir2.2 wild-type currents under hyperpolarization and dependence on [K+]0. 134 Figure 3.6 Block of Kir2.2 wild-type currents by Ba2+ and Cs+. 136 Figure 3.7 Single channel current properties of Kir2.2 137

Figure 4.1 Block of Kir2.2 wild-type currents by Ba2+ 147 Figure 4.2 Voltage- and concentration-dependence of the steady-state block of Kir2.2 by Ba2+ 151 in Kir2.2 wild-type 151 Figure 4.3 Concentration and voltage dependence of Ba 2+ block in Kir2.2 wild-type 152

xii Figure 4.4 Voltage-dependency of the association (k+b) and dissociation (&_b) rate constants for block of Kir2.2 by Ba2+. 154 Figure 4.5 Block of single Kir2.2 wild-type currents by Ba2+ is voltage-dependent 156 Figure 4.6 Block of single Kjr2.2 wild-type currents by Ba2+ is concentration dependent 157 Figure 4.7 Concentration and voltage-dependence of mean open time 159 Figure 4.8 Voltage-dependency of the association rate constant (£+b) forblockofKir2.2byBa2+. 160 Figure 4.9 Block of wild-type and mutant Kir2.2/Kir2.1 channels by Ba2+ 162 Figure 4.10 Voltage-and concentration-dependence of the steady-state block of wild-type and mutant Kir2.2/Kir2.1 currents by Ba2+ 163 Figure 4.11 Voltage dependency of the time constant (xblock) and association (&+b) and dissociation (&_b) rate constants for block of Kir2.2/Kir2.1 wild-type and mutant channels by Ba2+ 165 Figure 4.12 Comparison of unitary conductance between wild-type and mutant Kir2.2/Kir2.1 channels 167 Figure 4.13 Current-voltage relationship for wild-type and mutant Kir2.2/Kir2.1 channels 168 Figure 4.14 Energetic profile of Ba2+ blocking interaction with the pore of Kir2.2/Kir2.1 wild-type and mutant channels 171 Figure 4.15 Schematic representation of the tetrameric structure of Kir2.2 viewed from the extracellular side 173

Figure 5.1 Ionic selectivity of Kir2.1 181 Figure 5.2 Ionic selectivity of Kir2.1 1143V, I143L and I143T 183 Figure 5.3 Ionic selectivity of Kir2.11143C and I143S 185

Figure 5.4 Correlation between P ri/P k and hydrophilicity 186 Figure 5.5 Expression of Kir2.11143A gives functional channels 192 Figure 5.6 Expression of Kjr2.11143C C149S, I143S C149S and 1143 C149S 193 Figure 5.7 In vitro translation of Kjr2.1 wild-type and mutant channels using the reticulocyte system 195 Figure 5.8 In vitro translation of Kir2.1 wild-type in the presence of microsomal membranes 197 Figure 5.9 In vitro translation of Kjr2.1 1143S in the presence of membranes 198

xiii Figure 5.10 Schematic representation of the tetrameric structure of Kjr2.1 viewed from the extracellular side 202

List of Tables

Table 4.1 Rate constants for block and unblock of Kjr2.2 by Ba2+ 150 Table 4.2 Voltage dependence of the mean open time for single fQr2.2 currents in presence of 200/tM Ba2+ 155 Table 4.3 Concentration dependence of the mean open time for single Kir2.2 currents at -97mV 158 Table 5.1 Alignment of P-region sequences of Kir2.1, KcsA and several Kv channels 179 Table 5.2 Properties of isoleucine 143 mutants 188

xiv Abbreviations

4-AP 4-aminopyridine AA Arachidonic acid ABC ATP binding cassette ADP 5'-diphosphate

Ag+ Silver ion AHP Afterhyperpolarization AKAP The A kinase anchoring protein ATP Adenosine 5' triphosphate Ba2+ Barium ion BFNC Benign familial neonatal convulsions BKCa High conductance calcium-activated potassium channels Ca2+ Calcium ion cDNA complementary deoxyribonucleic acid CCD Cortical collecting duct CFTR transmembrane regulator CHO Chinese hamster ovary cr Chloride ion co2 Carbon dioxide Cs+ Caesium ion CTAL Cortical thick ascending limb CTX C-terminus Carboxyl (COOH) terminus d 2 Dopamine receptors DCT Distal convoluted tubule DMEM Dulbecco’s modified eagle’s medium DMSO Dimethyl sulphoxide Em Resting

Ek Potassium ion equilibrium potential

Etc\ Reversal potential EDTA Ethylenediaminetetraacetic acid EGFP Enhanced green fluorescent protein F Faraday’s constant FBS Foetal bovine serum

GAGAb y-aminobutyric receptors XV G418 Geneticin® antibiotic GSH Reduced glutathione GSSG Oxidized glutathione HEPES V-[2-hydroyethyl]piperazine-vV'-[2-ethanesulphonic acid] IBX Iberiotoxin

/ k Potassium current IKca Intermediate conductance calcium-activated potassium channels /Na Sodium current K+ Potassium ion [K+]i Intracellular potassium ion concentration [K+]0 Extracellular potassium ion concentration Kir Inward rectifier potassium channel Kv Voltage-gated potassium channel

K2P Two pore potassium channel LCR Locus control region MAGUK Membrane associated guanylate kinase Mg2+ Magnesium ion Mg2+i Intracellular magnesium Mg-ADP Magnesium adenosine 5' diphosphate MEL Murine erythroleukaemia MEM-a Alpha minimal essential medium MB-IRK2 Mouse brain IRK2 mRNA Messenger ribonucleic acid MTAL Medullary thick ascending limb MTSEA Ethyl-ammonium-methanethiosulphonate MTSET 2-trimethylammoniumethyl-methanethiosulponate bromide Na+ Sodium ion nAChR Nicotinic acetylcholine receptor NH4+ Ammonium ion NMDG A-methyl-D-glucamine OMCD Outer medullary collecting duct I-V Current-voltage Popen Probability of opening Px Permeability coefficient of ion X PBS Phosphate buffered saline PHHI Familial persistent hyperinsulinemic hypoglycaemia of infancy pHi Intracellular pH

PIP2 Phosphatidylinositol-4, 5-bisphosphate PKA Protein kinase A PKC Protein kinase C PSD Postsynaptic density R Gas constant Rb+ Rubidium ion RB-IRK2 Rat brain IRK2 RNA Ribonucleic acid SAP Synapse-associated protein SEM Standard Error Mean

SKca Low conductance calcium activated potassium channels STX SUR Sulphonylurea receptor T Absolute temperature T1 Tetramerization domain 1

^act Activation time constant TAL Thick ascending limb TEA+ ion TTX

Vm Membrane potential v/v volume per volume wv weaver [X]i Intracellular ion concentration [X]o Extracellular ion concentration z Ion valency

xvii Table 0.1. Abbreviations for Amino acids

Side-chain Name Triple-letter Single-letter type code code

Aliphatic Glycine Gly G Alanine Ala A Valine Val V Leucine Leu L Isoleucine lie I

Aromatic Phenylalanine Phe F Tyrosine Tyr Y Trytophan Trp W

Sulphur-group Cysteine Cys C containing Methionine Met M

Hydroxyl- Ser S group Threonine Thr T Containing

Basic Lysine Lys K Arginine Arg R Histidine His H

Acidic Aspartate Asp D Glutamate Glu E

Amide Glutamine Gin Q Asparagine Asn N

The table is divided into two halves. The top half lists those residues that are hydrophobic whilst the bottom half lists those that are hydrophilic. Chapter 1

Introduction

1.0. General introduction

In recent years it has become evident that numerous diseases are caused by mutations in a variety of that encode ion channels. Termed ‘’, these conditions include fatal disorders such as cystic fibrosis, the long-QT syndrome and familial persistent hyperinsulinemic hypoglycaemia of infancy (PHHI), and less serious disorders such as benign familial neonatal convulsions (BFNC). Numerous drugs and therapies exist for the treatment of such diseases. However, these medical treatments have not typically been discovered by way of molecular insight but from either clinical use for other disorders or by functional screening, and many have serious side effects (see for example Goldstein & Colatsky, 1996). Identification of the underlying molecular bases of these diseases and a more comprehensive understanding of how mutations alter the normal function of ion channels should facilitate the development of novel drugs or alternative therapies for the treatment of such conditions.

1 The importance of functional studies in the development of alternative therapies is well illustrated by research into cystic fibrosis (see Greger et a l, 2001; Gelman & Kopito, 2002). This fatal disease, which affects 1 in 2500 people, is caused by mutations in the encoding the cystic fibrosis transmembrane conductance regulator (CFTR). CFTR is a member of the (ATP)-binding cassette (ABC) protein superfamily and functions as a involved in the regulation of chloride fluxes in the lung and exocrine tissues.

More than 1000 mutations have been identified in CFTR. However, deletion of a phenylalanine residue at position 508 (F508), located in the first ATP binding domain, accounts for the vast majority of cases of this disease (for reviews see Aidley & Stanfield, 1996; Ackerman & Clapham, 1997; Gelman & Kopito, 2002). Structure-function studies have shown that deletion of F508 leads to defects in folding, protein stability, activation kinetics and trafficking to the membrane, which in turn result in chloride transport failure (see for example Cheng et a l, 1990; for review see Gelman & Kopito, 2002). The transport of chloride through CFTR channels is accompanied by the movement of water, thus the failure of chloride transport results in a reduction in the amount of fluid produced by the epithelial cells of the lungs and exocrine tissues. This leads to the accumulation of thick mucus in the lungs and pancreas, which in turn causes difficulties with breathing and digestion. Cystic fibrosis patients suffer frequent pulmonary infections, and damage to epithelial cells of the lungs and pancreas ultimately has fatal consequences.

Traditional therapies include chest percussion and postural drainage to relieve obstruction of the small airways and the use of antibiotics to treat respiratory infections, recombinant human DNase to decrease the viscosity of the secretions and anti-inflammatory drugs to reduce the inflammatory response. However, an understanding of how defects in the CFTR protein cause cystic fibrosis suggests the possibility of novel therapeutic approaches. For example, one investigation has successfully looked at relocation of AF508CFTR proteins at the cell surface using glycerol as a chemical chaperone (Sato et al., 1996). Thus, drugs that rescue or correct folding and/or trafficking of this CFTR mutant may be beneficial in the treatment of most forms of cystic fibrosis (see Gelman & Kopito, 2002).

Clearly, structure-function studies of ion channels are very important. However, to understand how gene mutations affect ion channels requires knowledge about how these channels function normally. In this study, I have investigated two members of the strong

2 inwardly rectifying potassium channel (Kir) family, Kir2.1 and Kir2.2. Until recently, mutations in genes encoding inward rectifier potassium channels had been linked to three disease states: Bartter’s syndrome, PHHI and the weaver mouse (for reviews see Sanguinetti & Spector, 1997; Curran, 1998; Abraham et a l , 1999; Reimann & Ashcroft, 1999).

Bartter's syndrome is a rare, autosomal recessive renal tubular disorder characterized by salt wasting and hypokalemic alkalosis. Mutations in several different ion channels and transporters have been linked with this disorder (see for example Simon et al., 1996a, b; 1997; Peters et al., 2002). Mutations in ROMK channels underlie type II Bartter’s syndrome (see section 1.2.4a/1.5.1 for more detailed discussion).

Mutations in Kir6.2 and its associated protein SUR1, which result in defective K a t p channels, underlie the hereditary disorder persistent hyperinsulinemic hypoglycaemia of infancy (PHHI), a condition characterized by unregulated insulin secretion and severe hypoglycaemia (for review see Aguilar-Bryan et a l, 2001; see section 1.2.4f for further discussion).

The weaver (wv) mutation in mice is caused by a single mutation in the pore region of Kir3.2, which results in the loss of K+ selectivity (see for example Patil et a l, 1995; Slesinger et al, 1996; Navarro et a l, 1996). This neurological disorder has been attributed to a reduced number of cerebellar granule cells during development and loss of dopaminergic cells in the substantia nigra. Owing to the degeneration of dopaminergic neurones exhibited by wv mice, these animals provide an interesting model for Parkinson’s disease (see sections 1.2.4c. and 1.5.6 for further discussion).

Recently, animal knockout studies of Kjr2.1 and Kir2.2 have highlighted the importance of the strong inward rectifiers in the heart and in vascular smooth muscle (see sections 1.5.2 and 1.5.3; Zaritsky et a l, 2000; 2001). Kir2.1 knockout mice die within hours of birth from a complete cleft of the secondary palate suggesting that Kjr2.1 also plays an important role in development. Furthermore, mutations in Kjr2.1 have been linked with Andersen’s syndrome, a rare disorder characterized by periodic paralysis, cardiac arrhythmias and dysmorphic features (Plaster et a l, 2001). In their elegant study, Plaster et a l (2001) demonstrated a link between muscle and cardiac electrical phenotypes demonstrating the importance of potassium channels in modulating excitability. In addition, they also provided evidence for an important role of Kjr2.1 in development.

3 In this introductory chapter I will review previous studies of structure and function in IQr channels. However, my review will not focus on the Kirs alone, since studies of structure and function in voltage-gated potassium channels (Kv) have provided many clues to which structural elements of Kjr channels might determine their functional properties. Hence, this introductory chapter will also consider some of the previous structure-function studies undertaken in Kv channels. Particular emphasis is placed on those studies that investigated the molecular basis of K+ selectivity. Before reviewing these studies, I will outline the earlier functional studies of inward rectification in native tissues. However, I should like to begin by introducing the concept of ion channels.

1.1. Ion channels

In this first section I will describe some of the major experiments that have led to the discovery of ion channels. I will then briefly consider some of the properties of ion channels and the membrane of excitable cells.

1.1.1. A historical perspective of ion channels

Ion channels are generally thought to have evolved millions of years ago, around the same time that cells began to develop membranes. The cell membrane, which consists of a lipid bilayer, serves as a relatively impermeable barrier to the passage of most water-soluble molecules. To achieve transport of molecules such as ions, sugars, amino acids, and numerous cell metabolites across the lipid membrane, cells developed specialized transmembrane proteins of which there are two main classes: carrier proteins and ion channels.

Briefly, carrier proteins (also known as ion pumps or ion transporters) are large proteins, which move ions using free energy stored in an electrochemical gradient or released by hydrolysis of ATP. Carrier proteins transport ions relatively slowly, passing 102 ions per second, and were initially thought to achieve transport of small molecules back and forth across the membrane via ‘a ferryboat-like’ mechanism (see Hille, 2001). However, carrier proteins are now considered to be like ion channels (see below), but with two gates, one external to and one internal to the ion-binding cavity. The two gates have been proposed to open alternatively, exposing the extracellular and intracellular media to the ion binding sites. These gating movements lead to the entrapment of ions within the interior of the pump and the subsequent release to the opposite side (see for example Artigas & Gadsby, 2002).

4 In contrast, ion channels form narrow, hydrophilic (water-filled) pores, which when open allow the passage of ions down their electrochemical gradient. Although the existence of ‘membrane pores’ was recognized as early as the 1840s (for review see Hille et al., 1999), the concept of discrete proteins selective for the transport of individual ions was not proposed until over a century later. Instead, early research began from a different perspective when, in the 1880s, Sydney Ringer discovered that the solution used to bathe an isolated frog heart must contain precise ratios of sodium (Na+), potassium (K+) and calcium (Ca2+) ions in to maintain its electrical excitability (see Hille, 2001).

At the beginning of the 20th century Julius Berstein introduced the concept that the membrane of excitable cells is selectively permeable to K+ at rest, but increases its permeability to other ions during excitation. His hypothesis proposed the ‘ ’ of membrane and nerve to be a ‘diffusion potential’, which arises as a result of the passive movement of K+ ions from the cytoplasm to the extracellular solution down their concentration gradient. He also used the term ‘membrane breakdown’ to describe the transient increase in permeability to other ions during excitation (for review see Hille, 2001).

Around about the same time, Hermann proposed that the cell membrane of excitable cells was itself electrically excitable and that the propagation of a nervous impulse occurred by way of electrical stimulation of unexcited membrane by the already active regions (see Hille, 2001). Later, Hodgkin (1937a, b) confirmed that nervous impulses, commonly known as action potentials, propagate electrically as originally hypothesized by Hermann, and Cole and Curtis (1938; 1939) confirmed that membrane permeability does increase during excitation, as first proposed by Bernstein.

Following the measurement of a full action potential from an axon using an intracellular micropipette (Hodgkin & Huxley, 1939), it was suggested that the cell membrane becomes selectively permeable to sodium (Na+) during excitation. Further experiments using low-Na+ external solutions and 24Na as a tracer confirmed the role of Na+ in the rising phase of the action potential (Hodgkin & Katz, 1949; Keynes, 1951).

In 1952, Hodgkin and Huxley published a series of papers (1952a, b, c, d) describing the changes in membrane Na+ and K+ permeability during action potential propagation in the squid axon, formulating equations that described the manner in which these currents varied

5 with membrane potential and time. Hodgkin and Huxley (1952d) proposed that sodium current (7 n 3) and potassium current (7 k ) were localized at particular sites in the membrane called ‘active patches’.

Later, following experiments in which they measured unidirectional fluxes using isotopes of K+, Hodgkin and Keynes (1955) proposed that ions could cross the membrane through narrow tubes or channels, with several ions in the channel at any given moment but moving in single file. A picture of individual proteins responsible for the transport of ions was finally beginning to emerge.

The idea that separate pathways for Na+ and K+ may exist was further supported by the discovery of a number of specific blocking agents several years later. Tetrodotoxin (TTX) and Saxitoxin (STX) were shown to selectively block 7 n 3, leaving 7 k untouched, whilst in contrast, the tetraethylammonium ion (TEA+) was shown to block Ik but not 7n3 (reviewed in Aidley & Stanfield, 1996; Hille, 2001).

Estimates of the single channel conductance of the conventional delayed rectifier in squid axon (~12pS) and frog nerve (~4pS) from current fluctuations provided further evidence for the existence of ion channels. The single channel conductance of the delayed rectifier implied that one potassium channel could pass a K+ ion at least every microsecond, a turnover number too high for a carrier mechanism (see for example Conti et al., 1976; Begenesich & Stevens, 1975; for review see Hille & Schwarz, 1978).

Perhaps the greatest breakthrough for research came in 1976 with the introduction of the patch clamp technique (Neher & Sakmann, 1976). Developed by Erwin Neher and Bert Sakmann, this technique allows the direct measurement of current flowing through individual ion channels. Following refinement (Hamill, 1981), ‘patch-clamping’ today remains a very useful tool for the investigation of ion channels.

More recently, the use of recombinant DNA technology has revolutionized the study of ion channels. By the early 1990s, numerous ion channels had been cloned and sequenced. Using site-directed mutagenesis combined with electrophysiological recording, many of the functionally important amino acids of ion channels have been identified. However, one drawback with such an approach is that a small change in one part of the protein may affect

6 the overall protein conformation, and in turn its function, even if the altered part is not directly involved in the function of interest.

By the late 1990s many hypotheses about the molecular mechanisms underlying ion channel function had been proposed, but fundamental questions remained owing to the absence of a true physical picture of channel structure (see for example Armstrong & Hille, 1998). However, the first three-dimensional structure of an ion channel was soon to be reported. Doyle et al. (1998b) used crystallographic analysis to reveal the structure of the pore of a potassium channel (KcsA) from the bacterium Streptomyces lividans. For the first time it was possible to understand, at a molecular level, the mechanisms that control ion selectivity and conduction in potassium channels.

1.1.2. Classification of ion channels

Since the findings of Hodgkin and Huxley in the 1950s, ion channel research has mainly focused on two channel properties: gating and selectivity.

Ion flow through channels is controlled by a process known as gating, during which the channel is either open or closed (see section 1.2.3a). Channel gating is usually in response to a stimulus such as a change in membrane potential or the binding of a neurotransmitter. When the channel is open, ions are free to move down their electrochemical gradient. Channel closure prevents such movement. Open channels are highly permeable to some but not all ions. This property is known as selectivity (see section 1.6). Pharmacological experiments have demonstrated the existence of channels selective for individual ions including K+, Na+, Ca2+ and chloride (Cl'). Together, the properties of selectivity and gating are used to classify channels. For example, a channel that is permeable to K+, and opens and closes in response to a change in membrane potential is classified as a voltage-gated potassium (Kv) channel.

1.1.3. A role for potassium channels

As we will see later, a huge variety of potassium channels exist (see section 1.2), and potassium channel dysfunction is implicated in a number of genetic and acquired neurological disorders (Shieh et al., 2000). But, what are the functional roles of potassium channels? To answer this, let us first consider one of the essential properties of excitable cells.

7 All cells have a resting membrane potential (.Em), a negative electrical charge across the membrane, typically in the order of -70mV for excitable cells. The resting membrane potential plays a central role in numerous cellular responses including the excitability of nerve and muscle cells. Its maintenance is therefore crucial to cell function. How does the resting membrane potential arise?

In living cells, the concentration of K+ inside the cell is much greater than outside, the reverse being true for Na+. These concentration differences are maintained partly by a Na+, K f- ATPase, which pumps two K+ ions into the cell, for every three Na+ ions pumped out. Although this creates a potential difference across the cell membrane, so that the inside is negative with respect to the outside, the activity of the Na+-K+ ATPase is not the main contributor to the resting membrane potential. Instead, it is the K+ gradient across the plasma membrane that predominantly determines Em.

At rest the plasma membrane is selectively permeable to K+, owing to the presence of ‘X* leak channels \ K+ exits the cell by moving down its concentration gradient, creating a negative potential across the cell membrane, which in turn opposes any further efflux of K+ from the cell. When the concentration gradient driving K+ out of the cell equals the opposing electrical gradient, the net flow of K+ ceases. The membrane potential at which this occurs is known as the equilibrium potential for K+ (£k)- A similar balance for other major ions including Na+ and Cl" also exists. The equilibrium potential for an ion, X, can be calculated from the Nemst equation:

whereR is the gas constant, T is the absolute temperature (°K), z the valency of the ion, F the Faraday constant and [X]0 and [X]i are the extracellular and intracellular concentrations of the ion.

Although Ek predominantly determines Em, living cells are not exclusively permeable to K+ but have a small resting permeability to other ions including Na+. This is commonly reflected in an Em slightly positive to Ek. The involvement of potassium channels in cell excitability underlies a number of their important physiological functions. Hyperpolarization of the membrane, which arises as a consequence of K+ moving down its concentration gradient, underlies the termination of action potentials in electrically excitable cells such as nerve and muscle. In addition, potassium channels play essential roles in, for example, the regulation of the heartbeat, the release of neurotransmitters, muscle contractility and hormone secretion (for review see Miller, 2000; Hille, 2001). Some of the physiological roles of Kjr channels will be considered in more detail later (see section 1.5).

1.2. Potassium channel diversity

Through electrophysiological studies a large diversity of potassium channels, which differ in their biophysical properties, pharmacology, modulation and tissue distribution, have been identified. In addition to uncovering the mechanisms responsible for this functional diversity, cloning efforts have revealed new superfamilies of potassium channels. The diversity of potassium channels is well illustrated by the genome of the simple nematode Caenorhabditis elegans in which over 80 genes encoding potassium channels were identified (Bargmann, 1998; Wei et a l, 1996). Alternative splicing, heteromeric assembly of different subunits and assembly of potassium channel subunits with various auxiliary subunits further increases this diversity.

Why is there such a huge diversity of potassium channels? It is thought that potassium channels may provide the modulatory functions that make excitable cells different from one another, whilst other channels such as calcium channels provide a function that is perhaps regulated in similar ways in all cells.

Based on structure, potassium channels are currently classified into one of three major categories (see figure 1.1): six-transmembrane domain potassium channels such as the Kv channels, two-transmembrane domain potassium channels including the Kjr channels and two- pore potassium channels (K2p). Potassium channels belonging to each of these categories will be considered in more detail below. However, I would like to begin this second section of my introduction by reviewing the early studies on the cloning of ion channels, with particular emphasis on those selective for K+.

9 Figure 1.1.

Molecular architecture of potassium channels. Predicted membrane topology of 6 transmembrane (TM) domain potassium channels, which include Kv, Eag-\\ke and KCNQ potassium channels ( top panel), 2 TM domain potassium channels such as the Kir channels {middle panel) and two-pore potassium channels {bottom panel). P denotes the highly conserved pore region(s), S4 the voltage-sensor and N- and C- the amino and carboxyl termini, respectively.

outside

\J \J

outside

C

N

1 0 1.2.1. Early potassium channel cloning studies

During the early years of ion channel studies, research was restricted to looking at the functional properties of native ion channels. However, the presence of multiple, overlapping conductances in native cells complicates the study of individual channels. In the last 20 years, the application of molecular cloning technology to ion channel research has identified the primary structure of many ion channels, and in combination with the use of heterologous expression systems, has enabled detailed analysis of the molecular basis for the function of individual channels in the absence of contaminating currents.

Biochemical purification of the nicotinic acetylcholine receptor (nAChR) channel and the voltage-gated was relatively simple, owing to the availability of tissues containing large quantities of these proteins and specific high-affinity ligands that bind to these channels (see for example Vandlen et a l, 1979; Conti-Tronconi & Raftery, 1982; Agnew et a l, 1978; Barchi, 1982). The subsequent cloning of these channels quickly followed (Noda et a l, 1982; Noda et a l, 1984). However, the cloning of potassium channel genes required a slightly different approach since it was not possible to purify potassium channel proteins, owing to an inability to identify any tissues rich in potassium channels. Furthermore, few specific high-affinity ligands for potassium channels were known. The alternative approach, described below, made use of genetic mutations in the Shaker locus of the fruit fly Drosophila melanogaster.

Flies carrying the Shaker mutation were originally identified because they violently shake their legs when under ether anaesthesia (Kaplan & Trout, 1969). The idea thatShaker may encode a potassium channel gene was first considered in 1977 during experiments on the neuromuscular junction in Drosophila larvae (Jan et a l, 1977). Following nerve stimulation, larvae carrying one of two Shaker mutations, ShKS133 or ShrK012°, displayed abnormal twitching and enhanced neurotransmitter release from the neuromuscular junctions. Jan et al (1977) found that this could be mimicked in wild-type larvae following treatment with the potassium channel blockers 4-aminopyridine (4-AP) and TEA+ ions, respectively. Based on these observations, it was proposed that the two Drosophila mutations might cause defects in potassium channels, which fail to repolarize the nerve terminal properly.

Tanouye et a l (1981) provided further evidence to strengthen the hypothesis that this defect was due to an abnormal K+ conductance. A delay in the repolarization of action potentials, observed in recordings from the cervical giant fibre axons of Drosophila adults carrying

11 Shaker mutation ShKSm, could be mimicked in wild-type fibres following application of the potassium 4-AP.

It was later shown, using the voltage clamp technique, that a rapidly activating, inactivating K+ current known as the A-type current (see section 1.2.3a), present in normal pupal and adult flight muscle, was completely absent from the flight muscle of Drosophila carrying the Shaker mutation ShKSU3. As in previous studies, this K+ current defect could be mimicked by application of 4-AP to normal pupal or adult flight muscle (see for example Salkoff & Wyman, 1981; for review see Tanouye et al., 1986). Similar results have been described in larval muscle (see for example Wu & Haugland, 1985). It was proposed that the previously reported enhanced neurotransmitter release at the neuromuscular junction, and the abnormally long delays in repolarization of the cervical giant axon action potentials in Drosophila mutants, were also likely to be the result of abnormal ‘A’ currents.

In the late 1980s several groups (Kamb et a l, 1987; Papazian et al., 1987; Baumann et a l, 1987) isolated and characterized genomic DNA and several complementary DNA (cDNA) clones from the Shaker locus. From the sequencing results of two cDNA clones, shAl and shA2 (Tempel et al., 1987), it was shown that the predicted protein contained six transmembrane segments characteristic of ion channels and an arginine-rich sequence similar to a region found in voltage-gated sodium channels (see Noda et a l, 1986). These results further supported the hypothesis thatShaker codes for a structural component of a voltage- gated potassium channel.

In 1988, it was confirmed that the Shaker gene does indeed encode a K+-selective channel following expression studies in Xenopus oocytes (Timpe et al., 1988). Oocytes injected with Shaker mRNA displayed transient outward currents, which showed similar voltage sensitivity, inactivation kinetics, ionic selectivity and sensitivity to 4-AP with native A-current.

It is now known that the Shaker locus encodes several different protein products, a consequence of alternative splicing of a large primary transcript in which different combinations of exons are selected as the primary Shaker transcript and processed into mature messenger RNA (see for example Baumann et a l, 1987; Pongs et a l, 1988; Schwarz et al., 1988). These different classes of cDNA clones represent different potassium channel subtypes, which differ greatly in their kinetics and voltage-sensitivity (Butler et al., 1989; Wei et al., 1990) and are termed Shab (Kv2), Shaw (Kv3) and Shal (Kv4). Like Shaker, Shal

12 encodes an A-type K+ current, whilst Shab and Shaw encode delayed rectifier-type K+ currents, which either inactivate very slowly or display no measurable inactivation. Mammalian homologues for each member of the Drosophila gene family have been described (see for example Temple et al., 1988; for summary see Salkoff et al., 1992).

1.2.2. The three classes of potassium channels

To discuss the numerous types of potassium channel in detail is beyond the scope of this thesis. Therefore, members of the six-transmembrane domain and two-pore domain structural classes of potassium channels are considered only briefly. The two-transmembrane domain potassium channels, to which the inward rectifier potassium channels belong, will be discussed in more detail, though the emphasis will be on the strong inward rectifiers, which are members of the second subfamily Kir2.0.

1.2.3. Six-transmembrane domain potassium channels

Potassium channels containing six transmembrane domains (S1-S6; see top panel of figure 1.1) represent by far the largest and most diverse group of potassium channels. This class comprises both voltage-gated (Kv) and calcium-activated (Kca) potassium channels, which include among their many members the Shaker- related, ether-a-go-go- (EAG) related and KCNQ potassium channel subunits. A functional channel comprises four a-subunits (see section 1.3.1), and for some channels, four accessory P-subunits (see below), which may dramatically or subtly alter the physiological properties of the tetrameric a-subunit complex. The region between S5 and S6, known as the P-region, constitutes part of the ion conduction pore (see section 1.6), whilst the S4 segment, in which lysine or arginine is repeated every third position, is thought to form the voltage sensor.

1.2.3a. Shaker-related potassium channels

In addition to the four subfamilies originally isolated (Kvl-4, see section 1.2.1), an additional five subfamilies (Kv5-9) have been identified (for review see Coetzee et al., 1999; Kaczorowski & Garcia, 1999). When expressed alone Kv5, Kv6, Kv8 and Kv9 subfamily members do not generate functional channels but are capable of modulating the activity of members of the Kv2 subfamily (for review see Song, 2002). Whilst some of these subfamilies contain multiple subtypes (e.g. KV1 is comprised of Kvl.1-1.8), others contain only one (e.g. Kv5.1). The members of this superfamily are thought to underlie the classically defined delayed-rectifier and A-type K+ currents, which are activated on depolarization but differ in

13 their activation and inactivation kinetics (see Mathie et al., 1998). These currents are essential for shaping action potential waveforms and their repolarization (for review see example Song, 2002).

A-type currents (7 a ), which were first characterized in molluscan neurones (Connor & Stevens, 1971; Neher, 1971) are typically activated at subthreshold membrane potentials and rapidly inactivate upon depolarization. Five members (Kv1.4, 3.4, 4.1 - 4.3) of the Kv family exhibit similar properties to 7a when heterologously expressed (for review see Song, 2002). Coexpression of the Kv(3l subunit (see section on P-subunits below) with members of the KV1 subfamily also yields currents with similar properties to 7a.

In contrast, delayed-rectifier currents (7 d r ), like those originally studied by Hodgkin and Huxley (1952a-d), exhibit only mild inactivation over a 100-200ms depolarization. Members of at least five potassium channel subfamilies (Kvl, 2 and 3, eag and KCNQ) produce delayed-rectifier like currents when heterologously expressed.

Gating

The manner in which Kv channels couple changes in membrane potential to channel opening and closure has remained elusive for sometime despite being the subject of intensive investigation. The results obtained from a combination of approaches including mutagenesis studies, accessibility measurements, fluorescence labelling and electrophysiology suggest that channel gating occurs in association with the movement of charged amino acids in the S4 and S2 segments within the cell membrane (see for example Papazian et al., 1991; Aggarwal & MacKinnon, 1996; Cha & Bezanilla, 1997; Starace & Bezanilla, 2001; for reviews see Bezanilla, 2000; Fedida & Hesketh, 2001).

More recent studies in KcsA have begun to shed light on the molecular mechanism of voltage-dependent gating. In KcsA, it is thought that the gate of the channel is formed at the site where the four M2 helices intersect (‘the inner helix bundle’; see Figure 1.2A) and that channel gating occurs by a rotation of the inner helices that line the pore (Perozo et a l, 1999). The channel opens following a counterclockwise rotation of the Ml and M2 helices, which occurs in response to changes in pH. Gating of KcsA has also been shown to involve small movements of the C-terminus region and movement of the internal side, but not the external side, of the selectivity filter (Perozo et al., 1999). Whilst the results of Perozo et al. (1999) suggested that there is only a subtle increase in diameter at the inner helix bundle, a more

14 recent study has suggested that the pore of KcsA could open to a much wider diameter (Zhou et al., 2001a).

Further clues to how potassium channels may gate have been provided by comparing the crystal structures of MthK (Jiang et al., 2002a), a channel from the archeon Methanobacterium thermoautotrophicum, and KcsA (Doyle et al., 1998b). These potassium channels were crystallized in their open and closed states, respectively. Jiang et al. (2002b) found that there were very large differences in the arrangement of the inner helices between MthK and KcsA. Whilst the helices were almost straight in KcsA, forming a bundle near the intracellular solution (i.e the inner helix bundle), they were bent and splayed open in MthK creating a wide (12A) entrance (see Figure 1.2B).

In MthK, the inner helices bend at a hinge point termed the gating hinge, which corresponds to a glycine residue in the amino acid sequence. Jiang et al. (2002b) proposed that opening might occur following the application of lateral force to the C-terminal half of the inner helices, causing the gating hinge to bend approximately 30° and the bundle to come apart. The glycine residue is highly conserved among potassium channels suggesting that such conformational changes may underlie gating in most potassium channels. Indeed, the results from a recent mutagenesis study in Shaker support this view (Yifrach & MacKinnon, 2002).

Inactivation

Kv channels also close spontaneously, typically in response to maintained depolarizations. Kv channels display a wide variety of inactivation time constants ranging from a few milliseconds to a few seconds, which have very different effects on the repetitive firing properties of individual neurones. Channel inactivation can be brought about in two distinct ways known as ‘N- or C-type inactivation’ (Hoshi et al., 1991; for review see Kukuljan et al., 1995).

N-type inactivation of potassium channels is typically regarded as fast inactivation and may be compared to the inactivation of sodium channels, which was proposed to occur by way of a ‘ball and chain’ type mechanism. Armstrong and Bezanilla (1977) proposed that during inactivation of sodium channels a cytoplasmic particle that forms part of the channel blocks the passage of ions when bound to the internal mouth of the pore (see Figure 1.3).

15 A. B.

Bundle

Figure 1.2.

Open and closed conformations of potassium channels. A , Schematic representation illustrating two subunits of the KcsA potassium channel. The red asterisk marks the central cavity, whilst the selectivity filter is highlighted in orange. The gate is formed by the inner helix bundle, where the M2 helices cross. B, Stereo diagrams illustrating the open and closed states of two potassium channels viewed from within the membrane (upper panel) and from the intracellular solution (lower panel). The inner helices of the closed KcsA pore (red) are almost straight, forming a bundle near the intracellular solution. In contrast, the inner helices of the open MthK pore (black) are bent and splayed open creating a wide entrance. (From Jiang et al., 2002b). Closed Closed Open C-type N-type inactivation inactivation

Figure 1.3. Inactivation o f potassium channels. Schematic representation illustrating C- and N-type inactivation of potassium channels. C-type inactivation involves a constriction of the selectivity filter, whilst N-type inactivation involves a ‘ball and chain’ type mechanism as described in the text. Adapted from Yellen (1998) according to the results of Zhou et al. (2001a), which suggested that the hydrophobic central cavity forms the receptor site for the inactivation gate in N-type inactivation. Using mutagenesis approaches, Iverson & Rudy (1990), Hoshi et al. (1990) and Zagotta et al. (1990) demonstrated that the N-terminus was important for N-type inactivation. The first 22 amino acids are thought to form the ‘inactivation ball’, which is tethered to the rest of the channel via the remainder of the N-terminal domain (for review see Kukuljan et al., 1995). Both charged and hydrophobic amino acid residues contribute to the interaction between the inactivation ball and the channel (Murrell-Lagnado & Aldrich, 1993a, b).

In contrast, C-type inactivation can be seen as a slow decay in the current, which remains after the rapid phase of inactivation. The time constants for this slower inactivation vary among alternatively spliced Shaker channel variants that differ in their C-termini (Hoshi et al., 1991). The inhibition of C-type inactivation by external TEA+ (Grissmer & Cahalan, 1989; Choi et al., 1991), external K+ and other monovalent cations (see for example Lopez- Bameo et al., 1993; Baukrowitz & Yellen, 1996) suggested that this type of inactivation involved a constriction or collapse of the outer mouth of the channel pore. The alteration of C-type inactivation by amino acid mutations in the pore region (position 449 in Shaker B) or the S6 transmembrane domain (position 463 in Shaker ) further supported this view (Hoshi et al., 1990; Lopez-Bareno et al., 1993; Schliefet al., 1996). Such conformational changes have been shown to involve a cooperative movement of all four subunits of the tetramer (see Panyi et al., 1995; Ogielska et al., 1995).

Additional evidence for a rearrangement of the outer mouth of the pore during C-type inactivation has been provided by cysteine substitution experiments. Shaker channels, in which threonine 449 is mutated to cysteine, are more sensitive to block by cadmium in the C- type inactivated state, suggesting that the spatial distribution of each of the four residues at position 449 changes during C-type inactivation (Yellen et al., 1994). Furthermore, substitution of cysteine for a methionine (position 448) located at the outer mouth of the Shaker channel pore resulted in the formation of disulphide bonds between subunits in the C- type inactivated state but not in the closed state (Liu et al., 1996). These studies indicated that the structural rearrangement changes the exposure pattern of at least three amino acids (448- 450 in Shaker) near the outer mouth of the pore

External cations and TEA+ have been proposed to compete with C-type inactivation by a ‘foot-in-the-door’ like mechanism whereby occupancy of a site within the pore by these ions prevents the closure of the channel at the outer mouth of the pore (see Lopez-Bareno et al., 1993; Baukrowitz & Yellen, 1996). However, using Shaker B mutants, Molina et al. (1997)

18 have shown that TEA+ can bind to residues located at the external entryway of the pore and block the channels without altering C-type inactivation. TEA appears to compete with C-type inactivation only when it penetrates deeply into the pore. This suggests that the conformational changes associated with C-type inactivation are local, occurring at a site deep within the outer mouth of the pore close to the selectivity filter. This hypothesis is further supported by the observation that the binding of a channel toxin to a site in the outer vestibule is relatively unaffected by C-type inactivation (Liu et al., 1996). Molina et al. (1997) speculated that the residue at position 449 in Shaker is not directly involved in inactivation, but regulates the occupancy of ions in deeper regions of the channel mouth.

Later, Starkus et al. (1997) found that Shaker channels were capable of conducting Na+ and Li+ in the C-type inactivated state provided that internal K+ was completely removed, and argued that the mechanism of C-type inactivation was unlikely to involve a structural collapse affecting the conduction pathway. Instead, it was suggested that channels appear inactivated in physiological solutions because internal K+ ions prevent Na+ ions from permeating through the channels and that K+ permeability relative to Na+ permeability is greatly reduced in the C- type inactivated state. Starkus et al. (1997) proposed that the selectivity filter changes its properties during C-type inactivation and that the change in relative permeabilities is the primary mechanism of the inactivation seen in C-type inactivated states. This is likely to be a consequence of a reduction in the outer pore diameter. K+ permeability is abolished because K+ ions are too large to fit the pore whilst Na+ ions may now experience a better low energy fit.

By constructing a chimeric K+ channel that conducted Na+ in the absence of K+, Kiss and Korn (1998) found that occupancy of the selectivity filter binding-site slowed inactivation. Furthermore, occupancy of an outer site by either K+ or Na+ influenced inactivation indirectly by trapping K+ at the selectivity filter binding-site. Kiss and Korn (1998) described the selectivity filter as a ‘dynamic aperture’, which in the open state is fully open. They proposed that the inner diameter of this aperture decreases during inactivation, which results in a reduced conductance through the channel. However, such a constriction could only occur when the binding site is unoccupied.

Together, the aforementioned studies suggest that the selectivity filter region of the pore undergoes a constriction during C-type inactivation (see figure 1.3), ‘pinching off’ the permeation pathway. Further support for this hypothesis has come from experiments

19 investigating pore accessibility using Ba2+ as a probe (see Harris et al., 1998; Basso et al., 1998). Channel inactivation occurs with Ba2+ bound at a high affinity site in the pore, but the conformational change in the outer pore region prevents dissociation of the blocker to the external solution.

P-subunits

To date three P subunits, which associate with members of the Kv channel family, have been cloned (Rettig et al., 1994; for review see Robertson, 1997; Pongs et al., 1999; Hanlon & Wallace, 2002). Designated Kvpi, Kvp2 and Kvp3, these proteins share a highly conserved core sequence lacking any membrane spanning segments, but display amino termini of varying lengths. Further diversity arises from the alternative splicing of the Kvpi subunit

(England et al., 1995). The assembly of P subunits with Kv a subunits may serve several functions. For example, p subunits can increase the cell surface expression of coexpressed a subunits, shift the voltage dependence of activation and confer the property of rapid inactivation to the channel (for review see Robertson, 1997; Martens et al., 1999; Korovkhina & England, 2002). X-ray crystallography studies suggest that the p subunit may interact either directly or indirectly with the channels voltage sensor (Gulbis et al., 1999).

1.2.3b. Calcium-activated potassium channels (Kca)

Gardos (1958) first reported the existence of a Ca2+-sensitive K+ conductance in red blood cells. Subsequently, Ca2+ -activated K+ currents have been identified in many cell types including snail neurones (Meech & Strumwasser, 1970), chromaffin cells (Marty, 1981) and skeletal muscle (Pallotta et al., 1981; Romey & Lazdunski, 1984). These channels are thought to play an important role in coupling changes in intracellular Ca2+ (Ca2+0 with membrane potential and are involved in determining the shape of action potentials and the duration of the interspike interval. On the basis of biophysical and pharmacological properties, Kca can be divided into three main types (see Sah, 1996; Vergara et al., 1998; Sah & Faber, 2002): high- (BKCa), intermediate- (IKca) and low- (SKca) conductance channels.

BKca channels

BKca channels exhibit a high unitary conductance, greater than lOOpS in symmetrical K+. These voltage-dependent channels are modulated by high concentrations of Ca2+ (l-10piM), and in neurones colocalize with calcium channels (for review see Toro et al., 1998; Vergara et al., 1998; Bissonnette, 2002; Choe, 2002; Orio et al., 2002). Channel opening occurs

20 independently of Ca2+ but when Ca2+ increases to the micromolar level the channel switches to a state in which less electrical energy is required to open the channel (Meera et al., 1996).

The first BKca channel to be cloned was from the Drosophila mutant Slowpoke (Atkinson et al., 1991; Adelman et al., 1992), though related channels have subsequently been isolated from other species including rat and mouse (Christie et al., 1989; Butler et al., 1993). Only one a subunit has been isolated but further diversity arises as a consequence of alternative splicing, phosphorylation of the a subunit and association with different P subunits, four of which have been cloned to date (for review see Sah & Faber, 2002; Orio et al., 2002).

Also known as maxi-K channels, BKca channels are classified as members of the Kv channel superfamily, which typically contain six transmembrane domains. However, these channels are unique in that they contain an additional membrane-spanning domain, SO, at the amino termini (for review see Toro et al., 1998; Kaczorowski & Garcia, 1999; Korovkina & England, 2002). For the purpose of this review they will be categorized in the same group as Kv channels owing to similarities in their amino acid sequences.

An additional four hydrophobic segments are found in the carboxyl terminus of the BKca channel, the last of which is in close proximity to a stretch of negatively charged amino acids termed the ‘calcium bowl’. This region is thought to form one of the binding sites for Ca2+ (for review see Orio et al., 2002; Sah & Faber, 2002). A tetramerization domain (see section 1.3.1) named BK-T1 has also been identified in the hydrophilic region between the S6 and S7 hydrophobic domains (Quirk & Reinhart, 2001).

As for Kv channels, BKca channels assemble as tetramers (see section 1.3.1) and associate with (3 subunits, though theBKca P subunit consists of two membrane-spanning domains (for

review see Kaczorowski & Garcia, 1999). The additional transmembrane segment of the a

subunit, SO, plays a crucial role in P-subunit modulation of the channel (Wallner et al., 1996).

Although sensitivity to Ca2+ is conferred by the a subunit, association of the a subunit with a

P subunit affects the Ca2+ sensitivity by shifting the voltage dependence of activation to more hyperpolarized membrane potentials (see Vergara et al., 1998; Sah & Faber, 2002). Several studies have demonstrated the P subunit to be essential for coupling localized increases in Ca2+i to BKca currents and the repolarization of action potentials (Brenner et al., 2000; Plliger et al., 2000).

21 Pharmacologically, BKca channels are sensitive to block by TEA+, charybdotoxin (CTX) and iberiotoxin (IBX), but are insensitive to apamin (for review see Sah & Faber, 2002). Functionally, activation of BKca channels is thought to underlie action-potential repolarization and the fast afterhyperpolarization (AHP) that follows (for review see Sah, 1996; Vergara et al., 1998).

SKca channels

In contrast, SKca channels display a much smaller unitary conductance (5-20pS), are voltage- insensitive and are activated by much lower concentrations of Ca2+ (100-400nM). To date, three genes, SKI-3, which encode SKca channels have been identified (Kohler et a l, 1996; Chandy et al., 1998). These subtypes differ in their tissue distribution and their sensitivity to apamin. Whilst SK2 and SK3 are blocked by apamin, SKI is insensitive. However, all SK subtypes are insensitive to block by low concentrations of TEA+, CTX and IBX (for review see Sah & Faber, 2002).

Structurally, SKca channels are similar to Kv channels in that they consist of six membrane- spanning domains, though little homology between Kv and SKca channels with the exception of the P-region is displayed at the amino acid level. Surprisingly, SKca channels are not gated by voltage despite the presence of an S4 domain containing three positively charged residues. Sensitivity to Ca2+ is conferred by calmodulin to which the channel is covalently linked. The binding of Ca2+ to calmodulin leads to a conformational change in the channel and its subsequent opening (for review see Sah & Faber, 2002).

SKca channels were thought to underlie the medium and slow AHPs (for review see Sah, 1996; Vergara et al., 1998), though recent evidence suggests that existing cloned SKCa channels are not responsible for the slow component (for review see Sah & Faber, 2002). SK3 channels have been linked to the hereditary disease myotonic muscular dystrophy (for review see Wickenden, 2002). lKca channels

The third type of Ca2+-activated potassium channel, which has an intermediate conductance, exhibits limited expression having only been found in peripheral tissue. Although the first observed calcium-activated K+ conductance identified by Gardos falls into this category (Christophersen, 1991), cDNA encoding an IKca channel (IK1) was not isolated until very recently (Ishii et al., 1997; Joiner et al., 1997; Logsdon et al., 1997; Ghanshani et al., 1998).

22 An IKca channel is also formed on coexpression of mslo, the BK channel cloned from brain, and a newly identified potassium channel termed Slack (Joiner et al., 1998). Like SKca channels, IKca channels are insensitive to voltage being gated by submicromolar concentrations of Ca2+i, which gates the channel by way of calmodulin (Xia et al., 1998; Fanger et al., 1999; for review see Kaczorowski & Garcia, 1999). However, IKca currents display inward rectification.

Pharmacologically, IK1 is sensitive to block by CTX and clotrimazole but is insensitive to IBX and apamin (for review see Sah & Faber, 2002). Very little is known about the function of IKca channels. However, the biophysical and pharmacological properties of IK1 are identical to those of native IKCas in lymphocytes (Khanna et al., 1999) and the block of native IKca channels in red blood cells may be of therapeutic value in the treatment of sickle cell anaemia (for review see Wickenden, 2002).

1.2.3c. Eag-like potassium channels

The first member of this family, Eag, was isolated from Drosophila in the early 1990s (Warmke et al., 1991), though its locus, which carries a mutation that causes an ether- sensitive leg-shaking phenotype was originally discovered in 1969 (Kaplan & Trout, 1969). Mutations of eag cause repetitive firing and enhanced transmitter release in motor neurones (Ganetzky & Wu, 1983).

Eag contains a cytoplasmic C-terminus segment that is homologous with a cyclic - binding domain (cNBD), and exhibits greater homology at the amino acid level with cyclic nucleotide-gated cation channels than with Shaker -related potassium channels. However, when heterologously expressed, Eag displays voltage-dependent, K+-selective currents (Robertson et al., 1996; Pond & Nerbonne, 2001).

Homology screening has subsequently identified several related channels (Warmke & Ganetzky, 1994; for review see Ganetzsky et al., 1999), which are currently divided into three subfamilies (for review see Schwarz & Bauer, 1999): Eag, Elk (Eag-like potassium channel) and Erg (Eag-related gene). In Drosophila, a single gene defines each subfamily, whilst in mammals each subfamily contains multiple members with distinct kinetic properties. These potassium channels form a diverse family, which includes the cardiac delayed rectifier current fK(Vr) encoded by the human ether-a-go-go related gene (HERG) and auxiliary MiRPl (KCNE2) subunits. HERG has been implicated in the long QT syndrome, a hereditary

23 condition characterized by excessively long cardiac action potentials (the QT interval is prolonged on the electrocardiogram), which can cause ventricular fibrillation and may result in sudden cardiac death (for reviews see Ganetzky et al., 1999; Sanguinetti, 2000; Vandenberg et al., 2001).

Owing to similarities in current kinetics, Eag-like currents were once considered as candidates for the molecular correlates of the M-current (see for review et al., 1997), though members of the KCNQ family are now considered more likely candidates (see section below).

1.2.3d. KCNQ potassium channel family

The KCNQ gene family represents the newest member of the six transmembrane domain potassium channel family. To date, five genes (KCNQ1-5) have been identified (for review see Robbins, 2001). KCNQ2 and KCNQ3 heteromultimers are thought to underlie the M- current (7 m ) (Wang et al., 1998), a slowly activating, non-inactivating K+ current that is responsible for damping neuronal excitability. However, homomeric KCNQ2-5 and heteromers of KCNQ3 with KCNQ5 or KCNQ4 also form functional channels with similar biophysical and pharmacological properties to 7m (see Selyanko et al., 2000; Wickenden et al., 2001; Jentsch, 2000).

Brown and Adams (1980) first discovered 7m during their studies of the effects of acetycholine (ACh) on sympathetic ganglion neurones, naming it the M-current because it is suppressed by muscarinic receptor activation. In many cell types 7M is the only current active at voltages near the threshold for action potential firing (for review see Marrion, 1997). Subsequently, it has been discovered that 7M is in fact suppressed by activation of many receptor types including substance P, luteinizing hormone releasing hormone (LHRH), purinergic and P-adrenergic (for review see Marrion, 1997). It is thought that a diffusible second messenger couples receptor activation to suppression of 7m, owing to the large number of receptor types that underlie its modulation. The mechanism by which 7M is modulated remains elusive, though a recent study points to the hydrolysis of phosphatidylinositol 4, 5- bisphosphate (PIP2) as a crucial determinant of M-channel modulation (Suh & Hille, 2002).

Mutations in four (KCNQ 1-4) out of the five KCNQ genes underlie hereditary diseases, causing electrical hyperexcitability in cardiac arrhythmia and epilepsy, and underlying defects in transepithelial transport in congenital deafness and cell degeneration in progressive hearing loss (for review see Jentsch, 2000).

24 1.2.4. Two-transmembrane domain potassium channels

Until 1993 all the potassium channels that had been cloned exhibited a similar structure. The molecular features of numerous other potassium channels including inward rectifiers remained unknown. Although several studies had reported the expression of inward rectifier K+ currents in Xenopus oocytes following injection of messenger ribonucleic acid (mRNA) isolated from various sources, expressed currents were small (see Yoshii & Kurihara, 1989; Lewis et al., 1991; Cui et al., 1992.

Inwardly rectifying K+ currents are the predominant conductance in J774.1 cells, a murine macrophage cell line (McKinney & Gallin, 1988). Injection of mRNA isolated from these cells into Xenopus oocytes resulted in the expression of large, K+-selective currents (Perier et al., 1992). This high level of expression of inward rectifier current and the absence of other expressed current made J774.1 mRNA highly suitable as a starting material for expression cloning of inward rectifier potassium channel cDNA. Although an inward rectifier potassium channel was eventually cloned from these cells (IRK1; Kubo et al., 1993a), the first inward rectifier was cloned from rat kidney (ROMK1; Ho et al., 1993).

Hydropathy analysis of ROMK1 (Ho et al., 1993) and IRK1 (Kubo et al., 1993a) suggested the presence of only two transmembrane domains termed M l and M2, in contrast to the six transmembane domains displayed by Kv channels. Although these two proteins were clearly different from Kv channels in their overall structure, they were found to contain an H5-like region (now termed the P-loop) similar to that found in Kv channels. In particular, the GYG motif (see section 1.6), which was subsequently shown to be critical for K+ selectivity in Kv channels (Heginbotham et al., 1994), was conserved. In their structural model, Ho et al. (1993) proposed that the M l and M2 segments span the membrane flanking the P-loop, which was likely to form the pore of the channel (see middle panel of figure 1.1). ROMK1 and IRK1 were proposed to resemble the ‘core region’ (S5, H5 and S6) of Kv channels, and were thus classified as a new superfamily (Kubo et al., 1993a).

The cloning of a new member of the inward rectifier potassium channel family, GIRK1 (Kubo et al., 1993b), quickly followed. GIRK1 was shown to share 43% and 39% identity in the total amino acid sequence with IRK1 and ROMK1, respectively. ROMK1, IRK1 and GIRK1 can be distinguished from each other based on their electrophysiological properties. Whilst all three clones encode inward rectifier potassium channels, ROMK1 and IRK1 are

25 constitutively active (see sections 1.2.4a and b) and display differences in their degree of rectification (see section 1.4.2), whilst GIRK1 currents require activation by G-proteins (see section 1.2.4c).

Since these preliminary studies, numerous cDNAs encoding inward rectifier potassium channels have been isolated (for review see Nichols & Lopatin, 1997). Because insight into the diversity of potassium channels has been gained using the combined approaches of physiology and molecular biology, a confusing array of terminology existed for the description of potassium channels. For inward rectifier potassium channels, the first channels to be cloned were named based on molecular cloning, i.e ROMK1, IRK1 and GIRK1. When Doupnik et a l (1995) divided inward rectifier potassium channels into subfamilies based on their amino acid identity and electrophysiological properties a standardized nomenclature was introduced. Members are termed Kir x.y, where ‘x’ designates the subfamily and ‘y’ designates the subtype within the subfamily. Thus, ROMK1, IRK1 and GIRK1 are now known as Kirl.la, Kir2.1 and Kir3.1, respectively. Each of the seven subfamilies, which have been described to date, is considered further below.

1.2.4a. 1.0 subfamily

Shortly after the cloning of Kirl.la from rat (Ho et al., 1993), its homologue was isolated from human (Yano et a l, 1994; Shucket a l, 1994). Additional transcripts are formed by alternative splicing of Kirl.la and are found in both rat and human (Zhou et a l, 1994; Yano et a l, 1994; Shucket a l, 1994; Boim et a l, 1995; Kondo et a l, 1996). Originally named ROMK2-6, but renamed Kirl.lb-f using standardized nomenclature, these transcripts share a common 3' exon encoding the majority of the channel protein. However, they exhibit N- termini with variable lengths. For example, Kirl.lb lacks the first 19 amino acids of Kirl.la, whilst Kirl.lc contains an additional 7 amino acids.

Tissue distribution ofKir1.0 subfamily channels

Kjrl.l transcripts have been detected in a wide variety of tissues including brain, heart, liver, pancreas, skeletal muscle and spleen but are most abundant in kidney (see Yano et a l, 1994; Shuck et a l, 1994). The Kirl.l isoforms are differentially expressed along the nephron in the rat kidney (see Lee & Hebert, 1995; Boim et a l, 1995). Kirl.lb displays the widest distribution having been detected in the medullary thick ascending limb (MTAL), cortical thick ascending limb (CTAL), cortical collecting duct (CCD) and distal convoluted tubule (DCT). In contrast, Kirl.la is specifically expressed in the CCD and outer medullary

26 collecting duct (OMCD), whilst Kirl.lc is expressed in the MTAL, CTAL and the DCT (see Lee & Hebert, 1995; Boim et al., 1995; for review see Schulte & Fakler, 2000).

Properties ofKir1.0 subfamily channels

These low-conductance channels are characterized by a high, voltage-independent open probability (Popen), mild inward rectification and high K+ selectivity (Ho et al., 1993). Furthermore, they exhibit a marked sensitivity to intracellular pH (pHj), which may be important for K+ homeostasis during different metabolic states (Fakler et al., 1996b). Small reductions in pHi from, for example, 6.8 to 6.4 or 7.4 to 6.8 decrease the Popen of Kir1.0 channels (see Tsai et al., 1995; Fakler et al., 1996b; Doi et al., 1996; Choe et al., 1997; Shuck et al., 1997; McNicholas et al., 1998). The reduction in Popen has been shown to be a consequence of the appearance of a long closed state (Choe et al., 1991’, McNicholas et al., 1998).

The steep pH sensitivity of members of the Kir1.0 subfamily is conferred by a lysine (K80 on Kirl.la, K61 on Kirl.lb), located immediately before the first transmembrane domain Ml (Fakler et a l, 1996b; Choe et al., 1997). An additional residue (T70 on Kirl.la and T51 on Kirl.lb), located in a conserved hydrophobic region known as the ‘Q-region’ (known alternatively as MO; Ho et al., 1993), also plays an important role in the modulation of pH sensitivity (Choe et al., 1997). Furthermore, two arginines (R41 and R311) located in the N- and C-termini have also been identified as crucial determinants of gating of Kir1.0 subfamily channels by pHj (Schulte et al., 1999; for review see Schulte & Fakler, 2000).

Kirl.O subfamily channels are inhibited by Ba2+ (Ho et al., 1993; Zhou et al., 1996; Loffler & Hunter, 1997), but are insensitive to TEA+ (Ho et al., 1993). Kirl.lb has been shown to be modulated by ATP, low concentrations being required to prevent channel run-down, but higher concentrations resulting in channel inhibition (McNicholas et al., 1994). Kirl.lb is also modulated by protein kinase A (PKA) (McNicholas et al., 1994; Xu et al., 1996). PKA phosphorylation enhances the interaction of PIP 2 with Kir1.0 subfamily channels (Liou et al., 1999), which is required for constitutive activity (see section on modulation of Kir2.0 subfamily channels for further discussion of PIP 2). Channel regulation by PKA requires the involvement of an A kinase anchoring protein, AKAP75/79 (Ali et al., 2001. See section 1.4.2b for more detailed discussion of AKAPs).

27 Role of Kirl.0 subfamily channels

Members of the Kir1.0 subfamily play an important role in the recycling of K+ in the thick ascending limb (TAL) and secretion of K+ across the apical membrane of mammalian cortical collecting (CCD) tubules in the kidney (see section 1.5.1. for detailed discussion; for review see Wang et al., 2002).

When expressed alone, the properties of Kirl.l channels are not absolutely identical to those of the native small-conductance channels expressed in the kidney (see for example Ho et al., 1993). However, co-expression with members of the ABC family of proteins such as CFTR and sulphonylurea receptors (SUR, see section 1.4.2f), which are both expressed in the distal nephron, alters channel properties (see for example McNicholas et al., 1996; Ruknudin et al., 1998). For example, co-expression of Kirl.la with CFTR in oocytes increased the sulphonylurea sensitivity of Kirl.la to to a level similar to that of the native channels (McNicholas et al., 1996).

The recycling of K+ through Kirl.l channels is essential in the maintenance of the function of the Na+-K+-2Cl-cotransporter and mutations in Kirl.l channels causes type 13 Bartter’s syndrome, characterized by an impaired renal concentrating capacity (see section 1.5.1). Several mutations in Kirl.l have been linked with Bartter’s syndrome, which abolish channel function. These include mutations that introduce nonsense codons or frameshifts in the amino terminus (see for example Simon et al., 1996b), a mutation in the core (Ml-P-loop-M2) domain that alters the ion conduction pathway (see Derst et al., 1998), and a mutation that deletes the last 60 amino acids at the extreme carboxyl terminus (see Simon et al., 1996b; Flagg et al., 1999).

1.2.4b. Kir2.0 subfamily

The four members of this second subfamily are, like members of the Kir1.0 subfamily, constitutively active, but in contrast exhibit strong inward rectification (see Chapter 3). The currents carried by these channels display rapid activation upon hyperpolarization and are blocked by Ba2+ and Cs+ (see section 1.4.1).

Kir2.1

The first member, IRK1 (but now named Kir2.1) was originally isolated from murine J774.1 cells (Kubo et al., 1993a) but homologues have subsequently been isolated from various

28 sources including mouse and human brain (Morishige et al., 1993; Ashen et al., 1995; Tang et al., 1995), rabbit, human, mouse, rat and guinea-pig heart (Ishii et al., 1994; Ishihara & Hiraoka, 1994; Raab-Graham et al., 1994; Ashen et al., 1995; Wood et al., 1995; Nagashima et al., 1996; Liu et al., 2001a), a rat basophilic leukaemia (RBL-2H3) cell line (Wischmeyer et al., 1995), chick cochlear sensory epithelium (Navaratnam et al., 1995), bovine aortic and corneal endothelial cells (Forsyth et al., 1997; Yang et al., 2000b), human blood eosinophils (Tare et al., 1998), rat and human arterial smooth muscle cells (Horowitz, 1997; Bradley et al., 1999), human, chick and rabbit lens epithelium (Rae & Shepard, 1998), and a HeLa cell line (Klein et al., 1999).

Kubo et al. (1993) reported the unitary conductance of the Kir2.1 clone isolated from J774.1 cells to be ~21pS with 140mM K+in the patch pipette. The Kir2.1 clone isolated from mouse brain exhibits a similar conductance (~21pS, Morishige et al., 1993), but the Kjr2.1 clones isolated from rabbit, rat and human have a slightly higher conductance at ~23pS (Nagashima et al., 1996), 25pS (Wischmeyeret al., 1995) and ~30pS, respectively despite similar recording conditions (Raab-Graham et al., 1994; Tang et al., 1995; Wood et al., 1995).

Minor differences in the amino acid sequences between species may be responsible for the variation in unitary conductance. For example, the Kir2.1 clone isolated from chick cochlear sensory epithelium, which displays a unitary conductance of ~17pS, contains a glutamine (Q125) in the Ml-P-region linker, the corresponding residue of which is a negatively charged glutamate (E125) in the mouse clones. Navaratnam et al. (1995) found that mutation of this residue to a glutamate (Q125E) increased the single channel conductance of the channel to 28pS.

An alternative possibility is that species-specific regulatory proteins may associate with Kir2.1 and alter the single channel conductance. In rat, a ubiquitously expressed protein known as KCRF (for K channel regulatory factor), decreases the level of expression of Kjr2.1 in oocytes (Keren-Raifman et al., 2000). However it was not tested whether this was a consequence of a reduced unitary conductance, a reduced Popen, or simply a reduction in the number of channels being expressed in the membrane.

Variations in the single channel conductance have also been observed within the same species. Thus, whilst the Kir2.1 clone isolated from the chick cochlear sensory epithelium has a unitary conductance of ~17pS (Navaratnam et al., 1995), the clone isolated from lens

29 epithelium displays a conductance of ~30pS (Rae & Sheppard, 1998), similar to the value reported for human clones. A recent study has demonstrated that on expression in Xenopus oocytes and in HEK293 cells, Kir2.1 channels display a broad distribution of unitary conductance ranging from 2 to 30pS (Picones et al., 2001).

Kjr2.1 is expressed in a wide variety of tissues, as is apparent from the numerous sources from which it has been cloned. For example, the properties of the Kir currents in arterial smooth muscle most closely resemble those of expressed Kh-2.1 currents (Bradley et al., 1999). These Kir currents are thought to play a pivotal role in K+-mediated vasodilation (see section 1.5.2. for a more detailed discussion). Several studies (Brahmajothi et al., 1996; Nakamura et al., 1998; Zaritsky et al., 2001) have also shown that Kir2.1 is the major component of the strong inward rectifier in the heart (/Ki), which is involved in stabilizing the resting membrane potential and regulating the shape and duration of cardiac action potentials (see section 1.5.3 for more detailed discussion).

Kir2.2

The first evidence that the Kir2.0 subfamily consisted of multiple genes came in 1994 with the cloning of a novel inward rectifier potassium channel, RB-IRK2, from rat brain (Koyama et al., 1994). Homologues have subsequently been cloned from mouse brain (MB-IRK2, Takahashiet al., 1994), human atrium (hIRKI, Wibleet al., 1995), human genomic or cosmid libraries (hKir2.2, Namba et al., 1996; Hugnot et al., 1997) and guinea-pig heart (Liu et al., 2001a). Renamed Kir2.2 according to standardized nomenclature, this second member of the Kjr2.0 subfamily shares 70% homology with Kir2.1. Despite the high degree of homology that Kir2.2 shares with Kjr2.1, Kir2.2 exhibits a higher single channel conductance (~34pS-41pS in 140mM K+; Takahashiet al., 1994; Wible et al., 1995) under similar recording conditions. Preliminary studies have also suggested that it may also be more sensitive to block by Ba2+ (Thompson et al., 1999). Takahashiet al. (1994) have suggested that the difference in the single channel conductance between the two channels might be a consequence of a single difference in the amino acid sequences of the P-region (see Chapter 4).

A variant of Kir2.2 known as hKir2.2v, which differs by three amino acids in the P-region, does not express functional channels in Xenopus oocytes. Furthermore, when coexpressed with Kir2.2, hKir2.2v inhibited Kir2.2 currents (Namba et al., 1996). Mutational analysis indicated that the differences in the P-region between Kir2.2 and hKir2.2v were not

30 responsible for the loss of conductance in hKir2.2v, suggesting a role for the intracellular C- terminal region of hKir2.2v.

In contrast to Kir2.1, Kir2.2 is found predominantly in the cerebellum, but is also expressed in many other tissues including forebrain, skeletal muscle, kidney, uterus, heart and cochlea (Koyama et al., 1994; Takahashiet al., 1994; Wible et al., 1995; Hibino et al., 1997). Kir2.1 is virtually absent from the atrium (Ishii et al., 1994; Ishihara & Hiraoka, 1994; Karschin & Karschin, 1997; Melnyk et al., 2002), whilst Kir2.2 is found in both the atrium and ventricle (Koyama et al., 1994; Wible et al., 1995; Karschin & Karschin, 1997).

Despite its location on 17pl 1.1, to which the Smith-Magenis syndrome1 has also been mapped, mutations in the Kir2.2 gene have not been implicated in any disease states (Hugnot et al., 1997).

Kir2.3

A third member of the Kir2.0 subfamily, named Kjr2.3 (Morishige et al., 1994; Lesage et al., 1994), was cloned from mouse brain by several groups (see Morishige et al., 1993; Lesage et al., 1994; Bredt et al., 1995), and homologues have been isolated from human (Makhina et al., 1994; Perier et al., 1994; Tang & Yang, 1994), rat (Bond et al., 1994; Falk et al., 1995) and from hamster insulinoma and mouse collecting duct (Ml) cell lines (Collins et al., 1996; Welling, 1997).

At between 10-15pS, Kir2.3 displays a much lower single channel conductance than the first two members of the Kir2.0 subfamily under similar recording conditions (Makhina et al., 1994; Morishige et al., 1994; Collins et al., 1996; Welling, 1997). Although Kir2.3 differs from Kir2.1 and Kir2.2 by only one or two residues in the P-region, it contains a distinct stretch of amino acids in the extracellular loop linking the M l and P-regions, which may determine the lower conductance and sensitivity to Ba2+.

Kir2.3 is absent from the brain and body until embryonic day 21 (Karschin & Karschin, 1997) and thereafter displays a more restricted distribution in comparison with Kir2.1 and Kir2.2. Expression of Kir2.3 has been reported in the forebrain (see Lesage et al., 1994; Morishige et al., 1994; Bredt et al., 1995; Karschin et al., 1996; Stonehouse et al., 1999), atrium (see for

1 The phenotype of the Smith-Magenis includes midface hypoplasia, speech delay and growth retardation

31 example Melynk et al., 2002), astrocytes (Perillan et al., 2000), renal epithelial cells (Welling, 1997) and dorsal root ganglion neurones (Chopra et al., 1999).

Kir2.4

A fourth member of the Kir2.0 subfamily, Kir2.4, has been isolated from rat brain (Topert et al., 1998), human retina (Hughes et al., 2000) and guinea-pig heart (Liu et al., 2001a). Like the previously cloned members of this subfamily, Kir2.4 exhibits strong rectification and time- dependent kinetics, but displays a much lower affinity for the channel blocker Ba2+. The rat brain Kir2.4 also displays a lower affinity for block by Cs+, whilst human Kjr2.4 exhibits an affinity similar to Kh-2.1 channels. As for Kjr2.3, the single channel conductance of Kjr2.4 is lower (15pS in 140mM K+) than for Kirs2.1 and 2.2 (Topert et al., 1998). Sequence alignment suggests that rat Kir2.4 shares between 53-63% homology with Kjrs 2.1-2.3 (Topert et al., 1998) and less than 48% homology with other members of the Kir1.0, 3.0, 5.0 and 6.0 subfamilies. The human and rat Kir2.4 homologues share approximately 92% identity (Hughes et al., 2000).

Topert et al. (1998) found that rat Kir2.4 was preferentially expressed in the large choline acetyltransferase immunopositive neurones of cell motor nuclei, but failed to detect any message for Kir2.4 in the retina. In contrast, Hughes et al. (2000) observed predominant expression of human Kjr2.4 in the neural retina. The presence of Kjr2.4 in the retina has been confirmed recently with its detection in Muller cells (Raap et al., 2002). Strong inwardly rectifying potassium channels such as Kjr2.4 may be involved in the K+ influxes through membrane domains that envelope synapses, though this remains to be confirmed. Recently, Kir2.4 has also been found in guinea-pig atrium and ventricle, though its presence is restricted to neuronal cells (Liu et al., 2001a). The localization of Kir2.4 in the brain and in neuronal cells suggests that it may play a general role in the peripheral and central nervous system.

Kir2.0 channels interact with anchoring proteins

Kir2.0 channel subtypes have been shown to associate with members of the membrane associated guanylate kinase (MAGUK) protein family, which are thought to function as scaffolds for molecules such as PKA, facilitating neuronal regulation. Kir channels interact with regions of these proteins known as PDZ domains by way of a PDZ binding motif (Thr/Ser-X-Val), which is located at the C-terminus of the channel (Songyang et al., 1997). The MAGUK proteins with which Kir2.0 channel subtypes interact include postsynaptic density (PSD)-95/synapse-associated protein (SAP) 90 (Cohen et al., 1996a; Nehring et al., 2000) and SAP97 (Leonoudakis et al., 2001). In Kir2.3, phosphorylation by PKA disrupts its interaction with PSD-95 (Cohen et al., 1996a). PSD-95 family members are thought to control the cell surface expression of ion channels, influencing trafficking from the endoplasmic reticulum and inducing the clustering of channels in the plasma membrane (see for example Horio et al., 1997; Kurschner et al., 1998). However, Kjr2.1 also contains a cytoplasmic C-terminal motif that is necessary for its efficient export from the endoplasmic reticulum (Stockklausner et al., 2001).

Like Kirl.O subfamily channels, Kir2.1 has been shown to associate with another class of anchoring proteins known as AKAPs. Kjr2.1 associates with AKAP5, previously known as AKAP79, by way of its amino and carboxyl termini (Dart & Leyland, 2001). AKAP5 binds several enzymes including PKA and protein kinase C (PKC). However, whilst elevations in the intracellular cAMP level result in increased Kjr2.1 currents the activation of PKC has no effect.

Modulation ofKir2.0 subfamily channels

Members of the Kir2.0 subfamily are modulated by a variety of mechanisms (for review see Stanfield et al., 2002). For example, Kir2.1, Kjr2.3 and Kir2.4 possess consensus sites for phosphorylation by PKA and PKC. However, downregulation of channel activity in response to stimulation by PKC has been confirmed only for Kir2.3 (Henry et al., 1996), whilst evidence for the downregulation of Kjr2.1 appears to be conflicting (Fakler et al., 1994b; Henry et al., 1996; Wischmeyer & Karschin, 1996; Zhuet al., 1999b). Kir2.1 channels are, however, inhibited by tyrosine kinase phosphorylation (Wischmeyer et al., 1998). The modulation of Kir2.3 by PKC is mediated by direct phosphorylation of the channel protein and a threonine (T53) in the amino terminus has been identified as the PKC phosphorylation site (Zhu et al., 1999b).

Kir2.3 channels appear to be modulated by a variety of mechanisms that do not affect other members of the Kir2.0 subfamily. Thus, Kjr2.3 is modulated by arachidonic acid (AA), a cis- polyunsaturated fatty acid present in the plasma membrane (Liu et al., 2001b). The enhancement of Kir2.3 currents by AA is thought to occur by way of a direct action. Kir2.3 is also inhibited by G-protein (3y subunits, which appears to be mediated through a direct binding of the G(3y subunits to the N-terminus of Kir2.3 (Cohen et al., 1996b). Finally, Kir2.3 is inhibited by pH. A drop in pHc inhibits single channel conductance (see Coulter et al., 1995), whilst a reduction in pHj primarily inhibits Popen (see Zhu et al., 1999a). A short motif

33 in the N-terminus (in MO), which consists of approximately eight amino acids centred about threonine 53, is an important determinant of channel sensitivity to pHi (see Qu et al., 1999).

Kir2.0 subfamily channels also interact strongly with PIP 2 , which when bound to the channel stabilizes the open state (see Huang et al., 1998; see also for review Stanfield et al., 2002). A stretch of positively charged amino acids (PKKR, residues 186-189 in Kir2.1) located in the

C-terminus mediates channel binding of PIP 2 . The arginine at position 189 is highly conserved among K n channels. In addition to the PKKR motif, the C-terminus of Kir2.1 possesses multiple binding sites for PIP 2 , which are thought to confer direct gating of these channels by PIP2 without the need for additional gating molecules (see Zhang et al., 1999;

Soom et al., 2001; see also section 1.2.4.c). Mutations in these regions reduce binding of PIP 2 to the C-terminus, lowering channel open probability.

1.2.4c. Kir3.0 subfamily

This third subfamily comprises the G-protein regulated Kjrs (GIRKS) of which four members, designated Kjr3.1-3.4 (GIRKS 1-4), have been identified in mammals (see for example Kubo et al., 1993b; Dascal et al., 1993; Ashford et al., 1994; Lesage et al., 1994; Bond et al., 1995; Krapivinsky et al., 1995a; Lesage et al., 1995; Dissman et al., 1996; Spauschus et al., 1996). An additional member, Kir3.5 (GIRK5/XIR), has also been found in Xenopus oocytes (Hedin et al., 1996). Further diversity of the Kir3.0 subfamily arises from the alternative splicing of the Kir3.2 genes to give distinct mRNA products termed Kir3.2 a - d, which differ in the length and amino acid sequence of their amino and carboxyl termini (see for example Dascal et al., 1993; Lesage et al., 1995; Tsaur et al., 1995; Isomoto et al., 1996a; Nelson et al., 1997; Inanobe et al., 1999b; Zhuet al., 2001).

Tissue distribution

Northern analysis has shown that in the heart, mRNA for Kir3.1 is more abundant in the atrium than in the ventricle (Kubo et al., 1993b). Kjrs 3.1, 3.2 and 3.3 are widely expressed throughout the CNS (Karschin et al., 1994, 1996; Kobayashi et al., 1995), whilst expression of Kir3.4 is only moderate (Ashford et al., 1994; Spauschus et al., 1996; Karschin et al., 1996; Wickman et al., 2000). However, Kjr3.4 is abundantly expressed in the heart (see for example Ashford et al., 1994). Whilst Kjr3.2b is ubiquitously expressed (Isomoto et al., 1996a), the other isoforms of Kir3.2 appear to be expressed in a tissue-specific manner. Thus, Kjr3.2a is specifically expressed in brain (Lesage et al., 1994), Kir3.2c is found in both brain and

34 pancreas (Lesage et al., 1995; Tsauer et al., 1995) and Kir3.2d is expressed in the testis (Inanobe et al., 1999b).

K.ACh is a heteromultimerofKir3.1 and Kir3.4.

Although early studies demonstrated that Xenopus oocytes injected with Kir3.1 mRNA express inward rectifier potassium channels resembling the muscarinic acetylcholine potassium channel (KAch) found in the heart (Kubo et al., 1993b; Dascal et a l, 1993), subsequent studies have shown that Kjr3.1 cannot form functional channels by itself but must coassemble with Kirs 3.2, 3.3 or 3.4 (Krapivinsky et al., 1995a; Kofuji et al., 1995; Duprat et al., 1995). In Xenopus oocytes, Kir3.1 associates with an endogenous protein, Kjr3.5 (GIRK5/XIR), to form functional channels (Hedin et al., 1996).

Together, Kir3.1 and Kir3.4 (CIR) encode a heteromeric channel with similar electrophysiological properties to KAch (Krapivinsky et al., 1995a; Hedin et al., 1996; Corey et al., 1998), which is involved in regulating the heartbeat (see section 1.5.4). Association with Kjr3.4 is essential for the cell surface expression of Kir3.1 (Kennedy et al., 1999), though it is thought that homomeric Kir3.4 channels do not contribute significantly to the KAch current (/kaoi; Bettahi et al., 2002). I^ACh is severely reduced in atrial myocytes from Kir3.1 knock-out mice despite normal expression levels of the Kir3.4 protein. Furthermore, Kir3.1 knock-out mice display a modest resting tachycardia. Residues in the carboxyl terminus of Kjr3.4 are, however, important for Gpy binding and activation of /kaoi (Krapivinsky et al., 1998b; see below).

Neuronal G-protein-gated inward rectifier potassium channels

Early studies suggested that neuronal G-protein-gated inward rectifier potassium channels might consist of Kir3.1 and Kjr3.2 subunits (Kofuji et al., 1995; Velimirovic et al., 1996). Indeed, heteromeric channels consisting of Kir3.1 and Kir3.2 have been identified in the cortex, hippocampus and cerebellum (Liao et al., 1996). Furthermore, Signorini et al. (1997) have reported a reduction of Kir3.1 protein levels in the brains of Kir3.2 knockout mice suggesting that the majority of Kir3.1 in the brain interacts with Kir3.2.

Kir3.2 is the only member of the Kjr3.0 subfamily found in dopaminergic neurones (Karschin et al., 1996; Inanobe et al., 1999a), where stimulation of dopamine (D 2) or y-aminobutyric acid (GABAb) receptors generates slow inhibitory postsynaptic potentials by activating a G- protein-gated inward rectifier potassium channel (Innis & Aghajanian, 1987; Lacey et al., 1988). As discussed in the opening introduction, the weaver mouse phenotype, which is caused by a single point mutation (G156S) in the pore region of Kir3.2 (Patil et al., 1995), is associated with the loss of cerebellar granule cells and dopaminergic cells of the substantia nigra (see section 1.5.6 for further discussion), regions in which Kir3.2 is expressed.

Kir3.2 and Kir3.3 are also co-expressed in some brain regions (see for example Duprat et al., 1995; Lesage et al., 1995; Kofuji et al., 1995) and have been purified as a complex from native brain (Jelacic et al., 2000). Kofuji et al. (1995) reported that Kir3.3 had an inhibitory effect on Kir3.2 currents, whilst Jelacic et al. (2000) have shown that heteromeric Kir3.2/Kir3.3 channels display a smaller unitary conductance and a lower sensitivity to activation by G(3y than heteromeric channels containing Kir3.1 subunits.

Modulation of Kir3.0 subfamily channels

Hormones and neurotransmitters regulate the activity of Kir3.0 channels. Among the G- protein-linked receptor subtypes that activate members of the Kjr3.0 subfamily are muscarinic

(M2), cholinergic, GABAb, serotonin (5-HTiA), adenosine (PI), somatostatin, opioid (p, k ,

5), cx2-adrenoergic, and D2 receptors (see for example North, 1989; for review see Jan & Jan, 1997).

In the heart, stimulation of the muscarinic receptor leads to activation of the inward rectifier Kaoi (see section 1.5.4). This process occurs by way of a pertussis toxin-sensitive G protein (Pfaffinger et al., 1985; Breitweiser & Szabo, 1985), but does not involve a diffusible second messenger (Soejima & Noma, 1984). Receptor activation leads to the exchange of GTP for GDP and the subsequent dissociation of the G-protein into G a and Gpy subunits (for review see Yamada et al., 1998; Stanfield et al., 2002). Although both a and Py subunits of the G- protein have been suggested to activate /kaoi (see for example Logothetis et al., 1987; Codina et al., 1987; Yantani et al., 1988a, b), the majority of experimental evidence supports the hypothesis that activation of /kaoi occurs by way of the Py subunit (Logothetis et al., 1987; Kurachi et al., 1989; Wickman et al., 1994; Krapivinsky et al., 1995b; for review see Stanfield et al., 2002).

As for the native cardiac channel, members of the Kir3.0 subfamily (Reuveny et al., 1994; Kofuji et al., 1995) are gated by the py subunits of heterotrimeric G-proteins, which directly bind to the channel. Both the N- and C-terminal domains of GIRK1 have been demonstrated

36 to be important in mediating channel activation by Gpy subunits (Reuveny et a l, 1994; Takao et al., 1994; Dascal et a l, 1995; Huang et a l, 1995; Kunkel & Peralta, 1995; Slesinger et a l, 1995).

How does the binding of GPy subunits lead to channel activation? Recent studies suggest that the binding of Gpy subunits might lead to conformational rearrangements of the second transmembrane domain and the proximal C-terminal domain, which results in an increase in Popen owing to an increase in open time kinetics and in burst duration (see for example Sadja et a l, 2001; Yi et a l, 2001; for review see Stanfield et a l, 2002). Gpy subunits also stabilize the interaction of G-protein gated inward rectifiers with PIP 2, which as described in section 1.2.4b stabilizes the open state of inward rectifier potassium channels (see for example Huang et a l, 1998). Inhibition of heteromeric Kir3.1/Kir3.4 channels with a PIP 2 antibody suggested that the binding affinity of Kir3.0 subfamily channels for PEP 2 is weaker than those of Kirl.l and Kir2.1 (Huang et a l, 1998). However, the affinity of Kir3.0 channels for PIP 2 is enhanced by GPy subunits. In addition, intracellular Na+ has also been shown to stabilize the channel-

PIP2 interaction in Kir3.0 subfamily channels (see Ho & Lagnado, 1999a; Zhang et a l, 1999).

1.2.4d. Kir4.0 subfamily

The first member of the Kir4.0 subfamily was cloned from rat brain and named BIR10 (Bond et a l, 1994). Also known as BIRK1 and KA b - 2 (Bredt et a l, 1995; Takumi et a l, 1995), this channel is now termed Kir4.1 using standardized nomenclature and has since been isolated from human kidney and mouse brain (Shuck et a l, 1997; Kurschner et a l, 1998). The initial classification of the Kir4.1 clones caused confusion. Because these clones displayed greatest homology with Kirl.la, and contained both a Walker type-A motif and ATP-binding domain,

K ir 4 .1 was initially classified into a subfamily termed K a b , which also contained Kirl.la (Takumi et a l, 1995). Subsequently, related channels have been cloned from salmon brain (Kir4.3; Kubo et a l, 1996), human kidney and skeletal muscle (Kjr4.2; Ohira et a l, 1997; Gosset et a l, 1997; Shuck et a l, 1997), guinea-pig kidney (Kir4.2; Derst et a l, 1998) and mouse and rat liver (mKij4.2, Kjr4.2 and Kir4.2a; Pearson et a l, 1999; Hill et a l, 2002).

Biophysical propertiesKi/t.O of subfamily channels

KirS 4.1, 4.2 and 4.3 have all been demonstrated to form functional channels when expressed in heterologous systems (see for example Bond et a l, 1994; Bredt et a l, 1995; Takumi et a l, 1995; Tucker et a l, 1996; Kubo et a l, 1996; Pearson et a l, 1999; Hill et a l, 2002), though

37 the Kir4.2 clones isolated from human failed to give functional channels (Shuck et al., 1997). Kij4.1 displays a high Popen and has a single channel conductance between 20 - 27pS in 140mM symmetrical K+ (Takumi et al., 1995; Yang & Jiang, 1999; Tanemoto et al., 2000). Kir4.2 also displays a high Popen and a unitary conductance of approximately ~25pS under similar recording conditions (Derst et al., 1998; Pessia et al., 2001), whilst the conductance of Kir4.3 is higher at 37pS (Kubo et al., 1996).

Kir4.1 and Kir4.2 have been shown to be sensitive to changes in pH*, though Kh4.2 is more sensitive than Ki,4.1 (Shuck et al., 1997; Pearson et al., 1999; Yang & Jiang, 1999; Pessia et al., 2001). Both Kir4.1 and Kir4.2 contain a lysine residue immediately before the Ml transmembrane domain, which corresponds to a lysine previously reported to be an important determinant of pH* sensitivity in Kirl.l channels (see section 1.2.4a; see also Fakler et al., 1996b; Shucket al., 1997; Pearson et al., 1999; Xu et al., 2000a). The increased sensitivity of Kii4.2 to pHj is conferred by an additional pH-sensing mechanism involving the C-terminus (Pessia et al., 2001). In K„4.1, intracellular protons inhibit Popen while low pHj enhances the single channel conductance (Yang & Jiang, 1999).

Members of the K^.O subfamily display intermediate rectification (Bond et al., 1994; Takumi et al., 1995; Kubo et al., 1996; Shucket al., 1997; Pearson et al., 1999; Lourdel et al., 2002), which has been linked to the presence of a serine in the carboxyl terminus. In the strong inward rectifier Kir2.1 the analogous residue is a negatively charged glutamate (E299; Kubo & Murata, 2001; see section 1.4.3).

Distribution and role ofKi,4.0 subfamily channels

Kir4.1 was initially found to be predominantly expressed in specific brainstem nuclei (Bredt et al., 1995) and in glial cells of the brain (Takumi et al., 1995), but has also been detected in kidney (Ito et al., 1996; Shuck et al., 1997; Tanemoto et al., 2000). The function of Kjr4.1 in the kidney involves the recycling of K+ for the maintenance of the Na+-K+-2C1' co-transporter (see also section 1.2.4e).

Kij4.1 is also present in the marginal cells of the stria vascularis in the cochlea where it is thought to be critically involved in the generation of the positive endoccochlear potential, which is essential for hearing (Ito et al., 1996; Hibino et al., 1997). Hibino et al. (1997) have reported that a mutant strain of mice known as Wv/Wv, which are deaf and have an endoccochlear potential of OmV, lack Kir4.1 from their stria vascularis.

38 KiA-2 has been localized in kidney, lung, pancreas, heart and brain (Ohira et al., 1997; Gosset et al., 1997; Shuck et al., 1997; Derst et al., 1998). In hepatocytes, Kjr4.2 is localized to the basolateral membrane where it is thought to couple to anion secretion and transepithelial water flow during bile formation (Hill et al., 2002). In contrast, the expression of Kjr4.3 appears to be limited to brain (Kubo et al., 1996), but very little is known about the function of these channels.

Ki/i.O channels interact with anchoring proteins

Like members of the Kjr2.0 subfamily, Kir4.0 channels also interact with proteins of the PSD- 95 family (Horio et al., 1997; Kurschner et al., 1998; Connors & Kofuji, 2002), which appear to increase the efficiency of channel formation. For example, in Muller cells functional expression of the dystrophin isoform Dp71 is necessary for the clustered localization of Kir4.1 channels. Dp71 is a core protein in a PDZ-containing protein complex, which spans the cell membrane to bridge the extracellular matrix with the actin cytoskeleton (Connors & Kofuji, 2002). The localized distribution of Kjr4.1 channels in Muller cells is important for ‘K+ siphoning’, a specialized form of K+ spatial buffering essential for the regulation of extracellular K+ levels (see Newman, 1993).

Several studies have shown that interaction of K^.O channels with PDZ-containing proteins leads to an increase in current density, which was found to be a consequence of an increase in the number of functional channels expressed in the membrane (Horio et al., 1997; Kurschner et al., 1998). However, whilst Connors and Kofuji (2002) found that Dp71 was critical for the clustering of Kir4.1 in mouse Muller cells, the level of cell surface expression was unaffected.

Heteromeric Ki/t.x channels display distinct biophysical properties

Kir4.1 and Kir4.2 also form heteromultimeric channels. Kir4.1 forms heteromers with Kjr 1.1 (Glowatzki et al., 1995), Kir2.1 (Fakler et al., 1996a) and Kir5.1 (Pessia et al., 1996; Xu et al., 2000a), whilst Kir4.2 forms heteromers with Kir5.1 (Pearson et al., 1999; Pessia et al, 2001). Heteromeric Kir4.1-Kir5.1 or Kir4.2-Kir5.1 channels display distinct properties from their homomeric counterparts including an increase in macroscopic amplitude, a higher unitary conductance and a lower Popen (Pessia et al., 1996; Pearson et al., 1999; Tanemoto et al., 2000; Yang et al., 2000b; Pessia et al., 2001). Furthermore, heteromerization of Kir5.1 with Kjr4.1 enhances the pHj-sensitivity of Kir4.1 so that the heteromultimeric channel is sensitive to inhibition by protons within the physiological range (Tanemoto et al., 2000; Tucker et al., 2000; Xu et al., 2000a; Yang et al., 2000a; Pessia et al., 2001). This inhibition is mediated by a reduction in the Popen, which arises from a decrease in the mean open time and an increase in the mean closed time, and is modulated by PIP 2 (Yang et al., 2000a).

Ki,4.1 can also form heteromers with members of the Kjr3.0 subfamily. However, this leads to inhibition of the Kir4.1 current mediated by structural motifs residing in the transmembrane domains of Kir3.0 (Tucker et al., 1996).

1.2.4e. Kir5.0 subfamily

Kir5.1 resides in its own subfamily owing to significant structural differences. Originally cloned from rat brain (BIR9, Bond et al., 1994), novel sequences for mouse, rat and human homologues have also been reported (Mouri et al., 1998; Derst et al., 2001a). When expressed alone Kir5.1 fails to form functional channels (Bond et al., 1994; Pessia et al., 1996; Salvatore et al., 1999; Tanemoto et al., 2002). However, coexpression of Kir5.1 with the anchoring protein PSD-95 in HEK293T cells mediates the formation of functional homomeric channels. PSD-95 alters the intracellular location of Kir5.1 and increases cell surface expression by slowing the rate of internalisation (Tanemoto et al., 2002).

Distribution and role ofKirS.l/PSD-95

Kjr5.1 has been detected in the adrenal gland, spleen, liver, kidney, testis, and brain (Bond et al., 1994; Salvatore et al., 1999; Tanemoto et al., 2000; Derst et al., 2001). Immunoprecipitation results suggest that PSD-95 and Kir5.1 form a protein complex in the brain (Tanemoto et al., 2002). These PSD-95/Kir5.1 complexes have been detected on the somato-dendritic plasma membrane in primary cultured neurones. The activity of PSD- 95/Kir5.1 channel complexes following heterologous expression in HEK293T cells was reversibly inhibited by PKA activation. Thus, as for Kir2.3 channels (see section 1.2.4b), PKA activation disrupts the interaction of Kjr5.1 with PSD-95. It has been suggested that these complexes play an important role in synaptic transmission (Tanemoto et al., 2002).

Kir5,l heteromers

It has been suggested that Kir5.1 interacts exclusively with members of the Kj^.O subfamily (Konstas et al., 2002). However, one study has shown that Kir5.1 assembles with Kir2.1 to give electrically silent channels (Derst et al., 2001a). Interestingly, the genes for Kir5.1 and

4 0 Kir2.1 are co-localized on the same chromosome, 17q and 11 in human and mouse, respectively (Morishige et al., 1993; Mouri et al., 1998; Derst et al., 2001a). It is thought that Kir2.1 might have been copied early in the evolutionary process to encode a protein with high amino acid similarity, and that this protein, Kir5.1, acts as a negative regulator of the Kir2.1 conductance.

Distribution and role of heteromeric Kif4.x-Kir5.1 channels

The functional properties of heteromeric Ki,4.x-Kir5.1 channels closely resemble those of potassium channels present in the distal convoluted tube (DCT) of the kidney, where the recycling of K+ across the basolateral membrane is essential for the reabsorption of Na+ by way of the Na+-K+-2C1' co-transporter (Lourdel et al., 2002). Indeed, K ^ .l, Kh4.2 and Kir5.1 are present in the DCT of kidney (Ito et al., 1996; Tucker et al., 2000; Derst et al., 2001a; Lourdel et al., 2002), and Kir4.1 co-localizes with the Na+-K+-2C1' co-transporter (Lourdel et al., 2002).

Kjr5.1 and Kir4.1 are also both expressed in the brain, but their failure to coimmunoprecipitate suggests that they do not preferentially coassemble (Tanemoto et al., 2000). However, the co-localization of Kir4.1 and Kir5.1 in the brainstem and their high sensitivity to pHi suggest that they may be the molecular correlates of the native lQr channels involved in CO 2 chemosensitivity in brainstem neurones (Pineda & Aghajanian, 1997; Yang et al., 2000a). Inhibition of these channels results in depolarization and an increase in membrane excitability, which is thought to underlie the spread of excitation to other brainstem neurones such as those involved in cardio-respiratory control. Thus, CO 2 chemosensitivity might be coupled to a corresponding change in excitability of the cardio-respiratory system (see Yang & Jiang, 1999; Yang et al., 2000a).

Kir4.2 and Kir5.1 have been co-localized in pancreas (Pessia et al., 2001), where they may be involved in the pH-dependent secretion of K+ produced by the secretion of bicarbonate.

1.2.4f. Kir6.0 subfamily

Members of the Kir6.0 subfamily associate with sulphonylurea receptors (SURs) to form KA tp channels, a group of weakly inwardly rectifying potassium channels that are sensitive to ATP.

KA t p channels were first described in cardiac myocytes (Noma, 1983), but have subsequently been found in numerous tissue including pancreatic beta cells (Cook & Hales, 1984), skeletal muscle (Spruce et al., 1985) and smooth muscle (Standen et al., 1989). These channels are

41 inhibited by millimolar concentrations of ATP and activated by Mg2+-adenosine 5'-diphosphate (Mg-ADP; see for example Nichols et al., 1996). Channel activation occurs when the cytoplasmic ATP/ADP ratio declines, for example, in response to hypoxia or to inhibition of cell metabolism.

Role ofK a t p channels

KA t p channels play important roles in both physiological and pathophysiological conditions by coupling cell metabolism with membrane electrical excitability. For example, they control insulin secretion from pancreatic P-cells (see section 1.5.5) and regulate vascular smooth muscle tone, but are also involved in the response to cardiac and cerebral ischaemia (for review see Ashcroft & Gribble, 1998). The diversity of K a t p channels arises from the assembly of different Kir6.0 and SUR subtypes, which are distinguished by their sensitivity to inhibition by sulphonylureas such as and glibenclamide, and their activation by potassium channel openers such as (see Inagaki et al., 1996; for review Aguilar- Bryan et al., 1998; Babenko et al., 1998).

Properties of Kirf.O channels

The first member of the Kir6.0 subfamily, Kir6.1, was isolated from a rat pancreatic islet cDNA library (Inagaki et al., 1995a), a second member (Kir6.2) being isolated shortly after (Sakura et al., 1995). Subsequently, further homologues of Kir6.1 and Kjr6.2 have been isolated from various human, mouse, rabbit and rat tissues (Inagaki et al., 1995b, c; Takano et al., 1996; Tokuyama et al., 1996; Yamada et al., 1997; Brochiero et al., 2002).

Kjr6.1 was originally proposed to be a KA t p channel, being named uKA t p -1 accordingly (Inagaki et al., 1995a). However, it was later shown that members of the Kir6.0 subfamily must assemble with sulphonylurea receptors (SUR) to form functional K a t p channels (Inagaki et al., 1995c; Sakura et al., 1995; Gribble et al., 1997). Kir6.1 and truncated versions of Kjr6.2, in which an endoplasmic retention signal is deleted, can form functional channels in the absence of SURs (Tucker et al., 1997; Zerangue et al., 1999; Brochiero et al., 2002). However, these channels exhibit a reduced sensitivity to ATP.

Kir6.1 and Kir6.2 exhibit different unitary conductances (Kir6.1/SUR ~35pS; Kir6.2/SUR ~ 80pS in 150mM symmetrical K+) despite that the amino acid sequences of their pore forming regions are identical (see Inagaki et al., 1995a; 1996; Isomoto et al., 1996b; Yamada et al., 1997). Residues in the extracellular linkers between the P-region and transmembrane

42 domains have been shown to determine this difference in unitary conductance (Kondo et al., 1998; Repunte et al., 1999).

Sulphonylurea receptors

Sulphonylurea receptors (SUR) are members of the ATP-binding cassette (ABC) protein superfamily to which the CFTR protein implicated in cystic fibrosis (see section 1.0) also belongs. Encoded by two genes designated SUR1 and SUR2, these receptors have a predicted topology of 17 transmembrane-spanning helices, organized into three domains termed TMO, TM1 and TM2. Furthermore, two intracytoplasmic folds (NBF1 and NBF2) contain motifs for nucleotide binding (see Aguilar-Bryan et al., 1995).

SUR1 was originally cloned from a hamster insulin secreting cell line (Aguilar-Bryan et al., 1995) and is found in pancreatic (3 cells, brain and heart. SUR1 exhibits a high affinity for sulphonylureas and interacts with several inward rectifier potassium channels including Kjr6.2 and Kirl.l (Ammala et al., 1996). Kir6.2/SUR1 complexes are thought to represent the pancreatic P-cell K a t p channel (see Inagaki et al., 1995c). A recent study has reported the existence of two novel forms of SUR1 known as SUR1A2 and SUR1BA31 (Gros et al., 2002). When heterologously expressed with Kjr6.2, SUR1A2, which has a single amino acid change in the first nucleotide-binding domain, forms functional channels with characteristics similar to the wild-type K a t p channels of pancreatic P cells. In contrast, heteromers of Kir6.2 and SUR1BA31, which lacks transmembrane domains 16 and 17, fail to reach the cell membrane.

The second sulphonylurea receptor, SUR2, is ubiquitously expressed in a diverse range of tissues including the heart, skeletal muscle, ovary, brain, tongue, pancreatic islets, lung, testis, adrenal gland, stomach and colon. SUR2 is less sensitive to inhibition by both ATP and sulphonlyureas, and is not activated by diazoxide (see for example Inagaki et al., 1996). Two major splice variants of SUR2, known as SUR2A and SUR2B, and a number of minor splice variants have been described (see Inagaki et al., 1996; Isomoto et al., 1996b; Chutkow et al.,

1996; 1999; Bronchiero et al., 2002). Kjr6.2/SUR2A displays similar properties to the KA tp channel of cardiac and skeletal muscle (see for example Inagaki et al., 1996), whilst heteromers of Kir6.1 or Kir6.2 with SUR2B are thought to represent the KA t p channel found in vascular and nonvascular smooth muscle (see Isomoto et al., 1996b; Yamada et al., 1997).

43 Structure and properties ofK a t p channels

Functional K a t p channels are thought to be octameric complexes of Kir6.0 subfamily channels and SUR subunits, which combine in a 1:1 stoichiometry (Clement et al., 1997; Shyng & Nichols, 1997; Inagaki et al., 1997). Several lines of evidence suggest that the Kir6.1/Kir6.2 subunits form the conduction pathway of KA t p channels, conferring inhibition by ATP (see for example Tucker et al., 1997; Proks et al., 1997; for review see Ashcroft & Gribble, 1998). The SURs are thought to act as regulatory subunits, conferring sensitivity to both sulphonylureas and potassium channel openers, and enhancing ATP sensitivity (see Aguilar- Bryan, 1995; Inagaki et al., 1995c; Ammala et al., 1996).

Katp channels are also modulated by PIP 2, which has been shown to reduce the sensitivity of

K a t p channels to inhibition by ATP, thus increasing channel open probability (see for example Baukrowitz et al., 1998; for review see Baukrowitz & Fakler, 2000). PDP 2 reduces channel affinity for ATP by reducing the association rate constant for ATP inhibition without affecting the dissociation rate constant. PIP 2 interacts with positively charged residues in the C-terminus of the Kir6.2 subunit (see also section 1.4.2b) and SUR1 further stabilizes the PDVchannel interaction (Baukrowitz et al., 1998).

Mutations in both Kir6.2 and SUR1 subunits have been linked with persistent hyperinsulinemic hypoglycaemia of infancy (PHHI), an autosomal recessive disease characterized by excessive and unregulated insulin release (Thomas et al., 1995; Bryan & Aguilar-Bryan, 1997). Disease-causing mutations in Kir6.2 and SUR1 subunits have been found to result in either a loss of Mg-ADP regulation or to result in a total loss of KA tp channel activity (see for example Thomas et al., 1995; for review see Aguilar-Bryan et al., 2001; Houpio et al., 2002). Children bom with PHHI can develop life-threatening hypoglycaemia if left untreated and individuals with severe PHHI, who fail to respond to treatment with KA t p channel openers such as diazoxide, typically require a partial or subtotal pancreatectomy for survival.

1.2.4g. Kir7.0 subfamily

Unique biophysical properties led Krapivinsky et al. (1998a) to propose that the inward rectifier potassium channel they had isolated from human foetal brain, which they named Kir7.1, be assigned to a new subfamily. With a hydrophobicity profile typical of Kjr channel subunits, Kir7.1 displayed closest homology to human Kirl.lc , sharing 38% identity at the

44 amino acid level. However, lack of the putative ATP-binding site and no apparent sensitivity to intracellular ATP distinguished Kir7.1 from members of the Kir1.0 subfamily. Kir7.1 has subsequently been cloned from various rat, human and guinea-pig tissues (Partiseti et al, 1998; Dbring et al, 1998; Nakamura et al, 1999; Shimuraet al, 2001; Derst et al., 2001b).

Biophysical propertiesKir7.1 of

The most striking features of Kir7.1 include a shallow dependence on [K +] 0 (see section 1.4.1 for further discussion), a very low estimated single channel conductance (~ 50-200fS), and a low sensitivity to block by external Ba2+ and Cs+ (see for example Krapivinsky et al., 1998a; Doring et al., 1998). Several of these properties may be attributed to differences in the pore region of Kir7.1. Replacement of a methionine (Met-125) in the pore of Kir7.1 with arginine, as conserved throughout the rest of the inward rectifier potassium channel family, restored channel dependence on [K+]0, and dramatically increased single channel conductance and sensitivity to Ba2+ (Krapivinsky et al., 1998a; Doring et al., 1998).

Distribution and role ofKir7.1

Early distribution studies (Krapivinsky et a l, 1998a; Partiseti et al., 1998) showed that Kir7.1 was present in a variety of tissues including brain, kidney, small intestine and stomach. Although such non-specific tissue distribution offered no clues to the function of the channel, the presence of Kir7.1 in neurones but not glial cells suggested a neurone-specific role. Krapivinsky et al. (1998a) suggested that the main role of Ki/7.1 is in the reliable setting of the resting membrane potential, a low conductance enabling significant precision in such regulation.

In contrast, Doring et al. (1998) were unable to detect Kir7.1 in any neuronal, glial, or connective tissue. Instead, a strong signal for Kir7.1 was detected in the secretory epithelial cells of the choroid plexus, which forms the blood-cerebrospinal fluid barrier in the mammalian brain. The presence of Kir7.1 in these cells, in addition to epithelia of lung, kidney and testis, led Doring et al. (1998) to propose that Kir7.1 channels might be involved in the regulation of transepithelial K+ transport.

More recent studies have confirmed the expression of Kjr7.1 in choroid plexus (Nakamura et al., 1999) and reported a robust expression of Ki/7.1 in several other epithelial cell types. These include thyroid follicular and intestinal epithelial cells (Nakamura et al., 1999), rat pancreatic acini (Kim et al., 2000), retinal pigmented epithelial cells (Kusaka et a l, 2001;

45 Shimura et al., 2001) and proximal tubule epithelial cells from guinea-pig and human (Derst et al., 2001b). Such cell-specific expression provides further support for the role of Kjr7.1 in the regulation of transepithelial K+ transport. Furthermore, it has been shown that Kir7.1 is co-localized with the Na+, K+-ATPase in several of these tissues (Nakamura et al., 1999; Kusaka et al., 2001) suggesting a close functional coupling between the channel and ion pump. It is thought that recycling of K+ through Kjr7.1 channels is important for maintaining the activity of the Na+, K+-ATPase.

1.2.4h. Drosophila Kirs

By homology screening with conserved mammalian sequences three Kir genes designated dKirI, dKirH and dKirIH were identified in the Drosophila genome (Doring et al., 2002). Only dKirI and dKirII expressed functional channels in Drosophila S2 cells. Functional channel expression in Xenopus oocytes was also possible following introduction of mutations in the cytoplasmic domains. dKirI and dKirII are phylogenetically most closely related to the members of the human Kir2.0 channel subfamily. dKirI transcripts were absent from embryos, while dKirII and dKirIII were expressed in the embryonic hindgut and in Malpighian tubules. In the adult fly, dKirII was predominantly expressed in the head.

1.2.4i. A family of Kir channels in prokaryotes

Five prokaryotic sequences that code for proteins whose closest relatives are eukaryotic Kirs have been identified (Durell & Guy, 2001). Termed KirBacs, these prokaryotic proteins exhibit homology with eukaryotic Kjrs in their C-terminus portion, whilst the sequence similarity of the P-loop and M2 transmembrane segments is intermediate between those of eukaryotic Kirs and other bacterial potassium channels.

1.2.5. Two-pore domain potassium channels

The two-pore domain potassium (K 2p) channels represent the most recently discovered class of potassium channels. Characterised by a lack of voltage- and time-dependency, with linear or near-linear current-voltage relationships in symmetrical K+, these channels are thought to correlate to the background (or ‘leak’) potassium channels that play essential roles in setting the resting membrane potential (see section 1.1.3).

In C.elegans, these two-pore domain channels represent the largest group of potassium channels, with approximately 50 genes encoding them (Wei et al., 1996). However, some of

46 these channels have unusual pore domains and may not encode functional potassium channels

(Bargmann, 1998), as is the case for some of the mammalian K 2P channels recently isolated (e.g. TASKS, Kim & Gnatenco, 2001; Ashmole et a l, 2001; KCNK7, Salinas et a l, 1999).

The first potassium channels with two P domains (PI and P2) and either four or eight transmembrane domains were isolated from yeast (Ketchum et a l, 1995; Zhou et a l, 1995; Lesage et a l, 1996a; Reid et a l, 1996) and Drosophila (Goldstein et a l, 1996). The cloning of the first mammalian K 2P channel quickly followed (Lesage et al., 1996b). The authors called this channel TWIK-1 (for Tandem of P domains in a Weak Inward rectifying K+ channel) since it exhibited functional characteristics similar to a weak inward rectifier potassium channel including saturation of the current-voltage relationship at depolarized potentials in the presence of internal Mg2+(Mg2+i).

Based on the assumption that four P regions are required to form a K+-selective pore (see section 1.3) TWIK-1 was predicted to be a dimer, a functional channel arising from the homologous assembly of two subunits or from assembly with another subunit. More recent studies suggest that expression of TWIK-1 in Xenopus oocytes fails to give functional channels indicating that an unidentified accessory subunit may be required for its functional expression (Goldstein et al., 1997; Pountney et al., 1999).

Lesage et al. (1996b) proposed that the overall structural motif of TWIK-1 is similar to what one would obtain by making a tandem of two classical inward rectifier potassium channel subunits. However, unlike K n channels, TWIK-1 contains an unusually large loop of 59 amino acids between the first transmembrane domain and the first pore loop (see bottom panel of figure 1.1). A specific cysteine residue residing in this large M l-Pl loop has since been implicated in the formation of homomeric channels by extracellular disulphide bond formation (Lesage et al., 1996c). Furthermore, whilst the first pore domain contains the GYG motif typical of most other K+-selective channels, the second pore domain contains a leucine in place of the tyrosine residue.

TWIK-1 is expressed in a wide variety of human tissues including heart and brain. Pharmacologically, TWIK-1 is sensitive to some of the classical potassium channel blockers such as Ba2+, but is relatively insensitive to others including TEA and 4-AP. This lack of specific pharmacology is typical of all K 2P channels and has hampered the elucidation of their roles in vivo.

47 Subsequently, a further thirteen mammalian TWIK-1 related channels have been cloned, which can be tentatively classified into the following subgroups: the TWIK, TASK, TALK, TRAAK/TREK and THIK/KCNK7 subfamilies (for summary see Karschin et al., 2001). These K;p channels are regulated by diverse physical and chemical stimuli. For example, members of the TWIK subfamily are sensitive to pHj, whilst members of the TASK subfamily are regulated by pHo. Other modulators of K 2P channel activity include free fatty acids such as arachidonic acid, volatile anaesthetics such as chloroform and halothane, and PKA and PKC (for reviews see Lesage & Lazdunski, 2000; Patel & Honore, 2001).

1.3. Assembly and stoichiometry of potassium channels

By the end of the 1980s it was clear that the principal subunits of voltage-gated channels selective for K+, Na+ or Ca2+ had closely related molecular structures (Catterall, 1988). These channels were proposed to contain four homologous transmembrane domains, each domain consisting of six membrane-spanning a-helices surrounding a central ion pore (Noda et a l,

1984; Tanabe et al., 1987; Tempel et al., 1987). The a-subunit of sodium and calcium channels contains four homologous internal repeats, whilst the a-subunit of potassium channels comprises only one of these repeats. Potassium channels were therefore proposed to assemble as tetramers (Tempel et al., 1987).

13.1. Kv channels assemble as tetramers

Isacoff et al. (1990) first demonstrated, using electrophysiological analysis of covalently- linked dimers, trimers, tetramers and pentamers, that functional potassium channels contain an even number of subunits, which was later narrowed down to four (MacKinnon, 1991; Liman et al., 1992). Further evidence that Kv channels assemble as tetramers has been provided by studies using sucrose gradient centrifugation of in vitro translated and solubilized protein and electron-microscopy of detergent-solubilized and purified Shaker protein (Shen et a l, 1993; Li et al., 1994). The latter study provided clear images of Kv channels, revealing a marked four-fold symmetry consistent with a tetrameric subunit composition.

Direct biochemical evidence that Shaker potassium channels contained four pore-forming subunits came in 1996 (Schulteis et al., 1996). By exposing intact cells to oxidizing conditions, disulphide bonds were generated between adjacent subunits, resulting in the formation of dimers, trimers, linear tetramers and circular tetramers of Shaker protein. More

48 recently, it has been shown that tetramers preferentially assemble from dimerization of dimers (Tu & Deutsch, 1999).

1.3.2. The T1 domain is required for tetramerization of Kv channels

Kv channels consist of either the same subunit (homomeric) or different subunits (heteromeric). However, only polypeptides in the same subfamily can form heteromeric channels (Christieet al., 1990; Isacoff et al., 1990; Ruppersburg et al., 1990; Covarrubias et al., 1991).

In Shaker channels, a conserved cytoplasmic sequence preceding the first transmembrane domain is required for tetramerization (Li et al., 1992). This 114 amino-acid region known as Tl, for ‘Tetramerization domain 1’, self-associates to form tetramers (Shen et al., 1993). Deletion of the Tl domain results in the failure of Kv channels to form tetramers. It is thought that the Tl domain is necessary for efficient assembly and trafficking beyond the endoplasmic reticulum (Schulteis et al., 1998). However, Kv1.3 subunits can form functional channels following its deletion (Tu et al., 1996). The primary role of the Tl domain in some KV1 channels is thought to be the prevention of their assembly with incompatible subunits. Indeed, the Tl domain has been shown to prevent heteromer formation of some members of the KV1 subfamily with members of other Kv subfamilies (Li et al., 1992; Xu et al., 1995; for review see Papazian, 1999).

It has been suggested that two different faces of the Tl domain mediate the association of monomers to form dimers and the subsequent dimerization of dimers to give tetramers (Tu & Deutsch, 1999). The structures of the Tl domains from the Kv4 and Kv3 subfamilies of Aplysia califomica have been determined and have been shown to exhibit four-fold, rotational symmetry (Kreusch et al., 1998; Bixby et al., 1999). At the centre of the T l tetramer is a water-filled cavity. This water-filled cavity does not form part of the ion conduction pore, but may represent the cytoplasmic vestibule in the ion-conduction pathway (Kreusch et al., 1998). The subunit interface of the Tl domain is highly polar and for all ‘non-Shaker-type’ Kv channels, Zn2+ ions, which are present at the assembly interface, act as the main structural determinant.

In addition to the Tl domain, the first (SI) and sixth (S 6 ) transmembrane domains of Kv channels have also been shown to be important for channel assembly (Shen et al., 1993;

49 Babila et a l, 1994). The SI transmembrane segment enhances the binding of the Tl domain to the wild-type channel.

1.3.3. Kir channels assemble as tetramers

Less is known about how Kjr channels assemble. Owing to the similarity between Kjr channels and the core region (S5-H5-S6) of Kv channels, it was proposed that Kjr channels also assemble as tetramers (Kubo et a l, 1993a). This was soon confirmed for homomeric Kjr2.1 (Yang et al., 1995b), Kirl.l and Kir4.1 channels (Glowatzki et a l, 1995), and for the heteromerically assembled Kir3.1/Kir3.4 channels (Silverman et a l, 1996a; Corey et a l, 1998). However, one study suggested that functional Kir2.1 channels might consist of twelve subunits (Omori et a l, 1997) and others have suggested that GIRK1, in combination with either an unknown subunit or with Kir3.4, might contain between 3-5 (Inanobe et a l, 1995) or 8-12 subunits (Krapivinsky et a l, 1995a).

1.3.4. Kir channels can form heteromers

Like Kv channels, members of the Kjr family also form heteromers, but in contrast, Kjr subunits belonging to different subfamilies can also coassemble to form functional channels. Heteromeric assembly was first demonstrated for members of the Kir3.0 subfamily (Krapivinsky et a l, 1995a; Kofuji et a l, 1995), and later for several other Kjr channels including Kir4.1 and K|rl.l (Glowatzki et a l, 1995), Kjr4.1 and Kir5.1 (Pessia et a l, 1996), Kir2.1 and K j^.l (Fakler et a l, 1996a) and Kjr4.2 and Kir5.1 (Pearson et a l, 1999). It is likely that other members of the K;r family also form functional heteromers, owing to the coexpression of different members in the same tissue. For example, Kjr2.1 and Kir2.3 are co-localized in the same regions of the brain (Fink et a l, 1996), though until recently the issue of whether members of the Kjr2.0 subfamily form heteromers has remained controversial.

Tinker et a l (1996) suggested that heteromerization between different Kir subunits may be case specific. For example, tagged Kir2.1 and Kjr2.2 proteins expressed in HEK293 cells do not co-precipitate suggesting that they are unable to coassemble. However, results from several recent studies have suggested that Kir2.1 and Kir2.2 may co-assemble to form heteromeric channels (Zaritsky et a l, 2001; Preisig-Muller et a l, 2002).

Co-assembly of Kjr2 .1 and Kir2.2 has been suggested as a possible explanation for the 50% reduction of 7ki observed in myocytes from mice lacking the Kjr2.2 gene, despite evidence to

50 suggest that Kir2.1 accounts for the major component of this conductance (Zaritsky et al., 2001). These results imply that Kjr2.1 is essential for the function or translocation of the heteromeric complex, whilst Kir2.2 is less essential. Thus, in the absence of Kir2.1, Kir2.2 would not be able to form homomers, but in the absence of Kir2.2, Kir2.1 forms homomeric channels that generate a current 50% smaller than that observed in wild-type, owing either to the expression of fewer channels or the smaller unitary conductance carried by homomeric Kir2.1 channels.

More direct evidence for the coassembly of members of the Kir2.0 subfamily has been presented recently (Preisig-Muller et al., 2002). Preisig-Muller et al. (2002) used several approaches, including expression of concatenated Kir2.x-Kir2.y channels, coexpression of non-functional Kir2.x constructs with wild-type Kir2.x channels and co-expression of Kir2.1 mutants related to Andersen’s syndrome with wild-type Kir2.x channels to show that Kir2.1 can form heteromeric channels with Kir2.2 and Kir2.3. Co-immunoprecipitation of Kir2.1 with Kir2.3 channels provides further evidence for the formation of heteromers by members of the Kir2.0 subfamily (Preisig-Muller et al., 2002). The fourth member of the Kir2.0 subfamily, Kir2.4, has also been shown to form heteromers with Kir2.1 (Schram et al., 2002).

Although Kir channels can form heteromers between subfamilies the results from one study, which investigated coassembly of Kvl.l with Kir2.1, suggest that they do not form functional heteromers with Kv channels (Tygat et al., 1996).

1.3.5. Structural determinants of Kir channel assembly

If heteromerization between different Kjr channels is case specific, as proposed by Tinker et al. (1996), which regions of the channel determine subunit compatibility? In contrast to Kv channels, Kir channels lack a tetramerization domain. However, several regions of Kir channels have been implicated in channel assembly. These include the transmembrane domains (Tucker et al., 1996) and the N-terminus (Fink et a l, 1996). However, Tinker et al. (1996) found no effect of deletion of the N-terminus on Kir subunit assembly, but instead demonstrated the proximal C-terminus and second transmembrane domain (M2) to be important in determining the homo- and heteromeric assembly of Kjr channels. For members from the same subfamily, such as Kjr2.2 and Kir2.1, both the M2 and proximal C-terminus regions appear to influence assembly. In contrast, the compatibility of the C-terminus regions is the major influencing factor for heteromerization of members of different subfamilies such as Kjrl . 1 and Kir2.1. A recent study on the heteromeric assembly of Kir5.1 and Kir4.0

51 channels has confirmed the importance of the proximal C-terminus as a determinant of heteromerization (Konstas et al., 2002).

1.4. Inward rectification

The terms ‘anomalous’ or ‘inward-going’ rectification are used to describe the potassium conductance that allows the passage of large inward current but only small outward current. Katz (1949) first observed this conductance whilst studying the electrical properties of frog skeletal muscle fibres bathed in isotonic potassium sulphate solution. Katz (1949) observed a low membrane resistance for inward current but a high resistance for outward current, i.e. the membrane conductance was larger over hyperpolarizing potentials. Katz (1949) introduced the term proprietes ‘ detectrices anormales’ to describe this ‘asymmetrical’ K+ conductance, which displayed rectification in the opposite direction to ‘delayed rectification’, previously described by Hodgkin et al. (1949). Later, anomalous rectification was also reported in cardiac Purkinje fibres (see for example Hutter & Noble, 1960; Carmeliet, 1961; Hall et al., 1963; Noble, 1965). Potassium channels that display this type of conductance are now termed ‘inward rectifier’ or ‘inwardly-rectifying’ potassium channels.

Preliminary studies of inward rectification were carried out in skeletal muscle (see for example Hodgkin & Horowicz, 1959; Adrian & Freygang, 1962a, b; Nakajima et al., 1962; Adrian, 1964; Horowicz et al., 1968; Adrian, 1969; Adrian et al., 1970a, b; Stanfield, 1970; Aimers, 1972a, b). However, it was later found that egg cell membranes of tunicates (Takahashiet a l, 1971; Miyazaki et al., 1974a, b) and starfish (Hagiwara & Takahashi, 1974; Miyazaki et al., 1975; Hagiwara et al., 1976) also display the property of inward rectification. Fewer complications with the technique used to measure the membrane potential and better space clamp conditions made the membrane of egg cells a more suitable preparation for studying the phenomenon of inward rectification. Furthermore, unlike muscle fibres, egg cell membranes are virtually impermeable to chloride ions, which must be eliminated if studying the effects of K+ on membrane potential in muscle fibres. Thus, many of the more extensive studies of the electrophysiological properties of inward rectification were carried out in egg cell membranes (see Hagiwara & Takahashi, 1974; Miyazaki et al., 1974a, 1974b, 1975; Hagiwara et al., 1976; 1977; 1978; Hagiwara & Yoshii, 1979; Hagiwara & Jaffe, 1979; Hagiwara, 1983). However, several groups continued to investigate inward rectification in skeletal muscle (see for example Standen & Stanfield, 1978a, b; 1979; 1980; Leech & Stanfield, 1981).

52 1.4.1. Properties of inward rectifier potassium channels

Together, the aforementioned studies revealed several characteristic features of inward rectification, which are used to distinguish inward rectifier potassium channels from other potassium channel subtypes. These properties are considered in more detail below.

1.4.1a. Inward rectification depends on Vm and the external K+ concentration

Hodgkin and Horowicz (1959) first suggested that inward rectification did not depend on membrane potential alone, as in Kv channels, but on the difference between the membrane potential and the equilibrium potential (V - £k)- To explain why fairly large membrane depolarizations were initiated by a rise in [K+]0, but only small membrane hyperpolarizations were initiated by a fall in [K+]0, Hodgkin and Horowicz (1959) proposed that potassium channels in frog skeletal muscle could pass a large current in high K+ if the driving force was inward (i.e. V < £k), but only a small current if the driving force was outward (i.e. V > EK).

To determine whether inward rectification was dependent on the electrochemical potential, as opposed to membrane potential (Vm) and external potassium concentration ([K+]0), Hagiwara and Yoshii (1979) studied the effects of altering intracellular potassium concentration ([K+]i). They demonstrated that channel gating depended solely on Vm when [K+]0 was fixed and [K+]i altered. Studies in frog skeletal muscle have subsequently confirmed this observation (Leech & Stanfield, 1981; Hestrin et al., 1981). It is therefore more correct to describe inward rectification as being dependent on Vm and [K+]0 rather then the electrochemical potential for K+.

At potentials just positive to E&, outward currents flow through inward rectifier potassium channels. However, at much more positive potentials these channels switch off, which is

reflected by a region of negative slope conductance in the I-V relationship. As [K +] 0 increases, the outward current passing through these channels also increases causing the I-V relationships to cross (see Figure 1.4; see for example Adrian, 1969). This is often referred to as ‘cross-over’.

53 ’Cross-over'

-25

Vm (mV)

f [Kl„

Figure 1.4.

The ‘Cross-over’ effect As [K+] 0 increases, the I-V relationships of inwardly rectifying potassium channels shift to the right as shifts to more positive potentials. The outward current flowing through inward rectifiers at potentials just positive to EK increases with increasing [K+]0. The I-V relationships of strong inward rectifiers, including the native inward rectifiers of skeletal muscle and egg cell membranes, cross each other owing to the presence of a ‘negative slope conductance’ region.

54 1.4.1b. The conductance is proportional to the square root of [K+]0.

The conductance of native inward rectifier potassium channels of skeletal muscle and egg cell membranes is roughly proportional to the square root of [K+]0, a four fold increase in [K +] 0 approximately doubling the slope conductance (see Horowicz et al., 1968; Hagiwara & Takahashi, 1974; Miyazakiet al., 1974; Hagiwara, 1976). Cloned Kir channels also exhibit this steep dependence of conductance on [K +] 0 (see for example Ho et al., 1993; Kubo et al., 1993a; Makhinaet al., 1994), which is conferred by an arginine located at the outer mouth of the pore. Kir7.1 channels, in which arginine is replaced by methionine, lack this dependence.

Thus, K+-conductance increases only slightly with varying [K +] 0 in Kir7.1 channels (Doring et al., 1998; see also sections 1.2.4g and 3.4).

1.4.1c. Activation

For negative voltages below Ek, a step change in membrane potential elicits an instantaneous inward current through inward rectifier potassium channels, which increases in magnitude with time to reach a steady-state (see for example Hagiwara et al., 1976; Hagiwara & Jaffe, 1979; Hagiwara, 1983). The time-constant (xact) for the current to reach steady-state follows a single-exponential relationship, the value of Tact decreasing with increasing hyperpolarization. Like the chord conductance, activation of inward rectifiers was found to depend on both Vm and [K+]0 (see for example Leech & Stanfield, 1981). It is now known that the time- dependent activation of inward current represents unblock of the channel pore by polyamines (see Lopatin et al., 1995; see section 1.4.2d for detailed discussion).

1.4.1d. Inactivation

At very negative potentials, the inward current through inward rectifier channels often declines with time after reaching its maximum amplitude. In skeletal muscle, this ‘inactivation’ at very negative membrane potentials was initially attributed to depletion of K+ in the tranverse (‘T’) tubules system (see for example Adrian & Freygang, 1962b; Adrian et al., 1970a). However, further studies (Adrian et al., 1970b; Aimers, 1972a, b) suggested that the decline in conductance at very negative membrane potentials could not be explained wholly through K+ depletion. For example, Adrian et al. (1970b) found that there could be a large decrease in current without much alteration of EK- Aimers (1972a) found that the decline in potassium conductance in skeletal muscle fibres with hyperpolarization exhibited two components: a rapid component at very negative potentials and a slow component at less negative potentials. While the slow component was attributed to changes in tubular K+

55 concentration, the kinetics of the rapid component were affected by temperature, indicating that it represented a time-dependent change in membrane permeability. Aimers (1972a) proposed that membrane hyperpolarization led to the closure of a gate within the membrane, which in turn caused a reduction in the K+ permeability.

Later, Standen and Stanfield (1979) reported that a voltage-dependent block of the inward current by Na+, which was present in the external solution, was responsible for the permeability change under hyperpolarization. Similar effects have been reported for the inward rectifier of egg cell membranes (Ohmori, 1978, 1980; Fukushima, 1982) and for other native channels (see for example Biermans et al., 1987). However, inactivation can still occur in the absence of any Na+. In skeletal muscle this was, like the decay of currents at more positive membrane potentials, attributed to K+ depletion (Standen & Stanfield, 1979). However, studies in cloned Kir channels have provided evidence to suggest that other mechanisms are responsible for the hyperpolarization-induced inactivation, including channel block by extracellular cations and residual hydroxyethylpiperazine (HEP) present in HEPES buffered solutions (see for example Choe et al., 1999; Guo & Lu, 2002; see section 3.4 for detailed discussion of inactivation in the absence of Na+).

1.4.1e. Block by external cations

It is well established that the inward movement of K+ can be blocked by certain cations including Cs+, Ba2+ and Rb+ (see for example Hagiwara & Takahashi, 1974; Hagiwara et al., 1976; 1978; Gay & Stanfield, 1977; Standen & Stanfield, 1978a; 1980). Studies of potassium channel block by, for example, Ba2+ in starfish egg membrane (Hagiwara et al., 1978) and frog skeletal muscle (Standen & Stanfield, 1978a) have shown inward rectifier channels to be highly sensitivity to Ba2+, which blocks in a time-, concentration- and voltage-dependent manner. Indeed, the high sensitivity of inward rectifier potassium channels to Ba2+ is often used as a tool for identifying novel inward rectifier channels (see for example Kubo et al., 1993a; Takahashiet al., 1994; Topert et al., 1998). Kir7.1 appears to lack the property of high affinity blockage by Ba2+ and Cs+, but sensitivity to block by Ba2+ is increased by the methionine to arginine mutation, which also restores the steep dependence of rectification on [K+]0(see sections 1.2.4g and 1.4.1b; Krapivinsky et al., 1998a; Doring et al., 1998).

56 Although monovalent, the effective valence (z')f of the blocking reaction for Cs+ in inwardly rectifying potassium channels is greater than 1. Such high values of z' result from inwardly rectifying potassium channels being multi-ion pores (see Hille & Schwarz, 1978), with voltage moving permeant ions as well as the blocking ion to bring Cs+ to its site of action.

1.4.1f. Anomalous mole fraction effects

Inwardly rectifying potassium channels display a characteristic known as anomalous mole fraction effects, where the membrane conductance and resting membrane potential become smaller in the presence of a mixture of ions than in the presence of either ion alone (see Hagiwara & Takahashi, 1974; Hagiwara et al., 1977). For example, Hagiwara & Takahashi (1974) found that the membrane conductance of starfish eggs was greater with extracellular solutions containing TINO 3 than with solutions containing KNO 3 at the same concentration. However, when the Tl+ and K+ solutions were mixed, the membrane conductance became smaller and the reversal potential more negative than in the presence of either of the pure salts. This characteristic of inward rectifier potassium channels supports the hypothesis that potassium channels are multi-ion pores (see Hille & Schwarz, 1978) and can be explained by the fact that K+ binds more tightly in the pore than Tl+.

1.4.2. Mechanism of inward rectification

By the end of the 1970s, the phenomenon of inward rectification had been studied extensively in both skeletal muscle and egg cell membranes. Although several theories to account for the asymmetrical potassium conductance displayed by inward rectifiers had been proposed, the mechanism of inward rectification remained unknown (see for example Leech & Stanfield, 1981).

The theories, which have been proposed to account for the mechanism of rectification, can be divided into two major categories. The first category includes several different models in which the gating of the channel is regulated by the binding of external K+ to ‘sites’ on the channel (see for example Horowicz et al., 1968; Ciani et al., 1978). This has been referred to as the ‘K-activated K-channel model’ (Pennefather et al., 1992). The second category comprises models in which the channel is occluded by an internal blocking particle, which is

f z' = the effective valence of the blocking ion x the fraction of the total potential drop through which the ion moves.

57 driven out of the channel when K+ currents are driven inward (see for example Armstrong, 1969; 1975; Hille & Schwarz, 1978; Standen & Stanfield, 1978b).

1.4.2a. The K-activated K-channel model

Horowicz et al. (1968) first proposed a carrier model for the mechanism of inward rectification following a series of experiments in which they measured unidirectional K+ fluxes by studying tracer movement in frog skeletal muscle. They found that the ‘system responsible for K+ movement was gradually activated by increasing [K+]0\ Furthermore, it was shown that K+ influx was related to the square of the [K +] 0 at constant internal potential and [K+]j. These observations led Horowicz et al. (1968) to propose a model in which K+ ions were assumed to move through the membrane by way of association with a ‘mobile carrier system’, which was activated by the binding of K+ ions to a ‘nonmobile control site’.

Later, Ciani et al. (1978) proposed a model that also involved activation by binding of K+ ions to a ‘gating site’. They considered two mechanisms to explain the steady-state electrical properties of the egg cell membrane when K+ is the only permeant ion. Their first hypothesis, which they termed the ‘electro-chemical gating’ hypothesis, was based on a previously suggested mechanism used to explain the voltage-dependent conductance induced in artificial bilayers by various antibiotics such as alamethicin. Alamethicin molecules have been proposed to adsorb to membrane surfaces with which they have direct contact and, on application of a positive potential, orient themselves in such a manner to enable the formation of aggregates and hence conductive pores. The negative end of the molecule is thought to adsorb to the membrane so strongly that on application of a negative potential, the pore fails to conduct.

Ciani et al. (1978) noticed that the properties of inward rectification in egg cell membranes displayed similarities with the voltage-dependence and rectification of conductance in artificial bilayers following addition of antibiotics to one side of the membrane. Based on this observation, they argued that a similar model, in which charged molecules located at the inner side of the membrane are required to orient themselves and aggregate to form permeable pores, could explain the mechanism of inward rectification. However, the model for inward rectification needed to take into account the dependence of conductance on [K+]0. Thus, the binding of external cations was also assumed necessary for pore formation. Both membrane potential and [K +] 0 were therefore assumed to affect the density of open pores.

58 1.4.2b. The blocking particle model

Armstrong (1969) first proposed the ‘blocking particle model’ for inward rectification following a series of experiments in which he studied the effects of TEA+ ions and their analogues on the potassium current in the giant axon of squid (see Armstrong & Binstock, 1965; Armstrong, 1966; 1969; 1971; 1975). Armstrong noticed that the TEA+ induced ‘in­ going rectification’ in squid axons and the inward rectification of skeletal muscle displayed similarities. Based on these observations he suggested a model for inward rectification in which K+ moving through the membrane displaces an internal blocking particle, which is located outside the electrical field of the membrane. The blocking particle hypothesis has since provided the basis for several other models of inward rectification (see Hille & Schwartz, 1978; Standen & Stanfield, 1978b; Urban & Hladky, 1979; Cleeman & Morad, 1979).

Hille and Schwartz (1978) simulated such a system using the Eyring rate theory. Their pore model contained three binding sites, and at least two ions were assumed to be moving in single file. The blocking ion was assumed to be a monovalent ion, which was proposed to block the channel from the inner end, but could not pass through the pore to the external solution. Their model correctly predicted the dependence of rectification on membrane potential and [K+]0, and could also account for anomalous mole fraction effects. However, the steepness of rectification predicted by their model was too shallow.

In the same year, Standen and Stanfield (1978b) used a similar blocking particle model to accurately fit their experimental measurements of potassium currents in frog sartorius muscle. Their blocking particle model assumed that the potassium conductance depends on the K+ concentration within a channel and that this conductance is reduced by a blocking particle, which is driven into the channel by depolarization. The shape of the I-V relationship could be accurately predicted based on the assumption that two monovalent blocking ions compete with K+ for binding to a site within the ion conduction pore. Interestingly, Standen and Stanfield (1978b) demonstrated that identical predictions could be made by assuming that the blocking particle was divalent, but binds at only one site, providing the dissociation constant for the blocking ion was adjusted accordingly. However, curves of the correct shape could not be predicted if only one monovalent blocking particle was assumed to bind at only one site.

59 1.4.2c. Mg2* block

Evidence to support the blocking particle model came in 1987 following three independent studies of inward rectifier potassium channels in cardiac myocytes (Horie et al., 1987; Matsuda et al., 1987; Vandenberg, 1987; for review see Matsuda, 1991). These studies indicated that rectification was a consequence of a voltage-dependent block by intracellular magnesium (Mg 2+i) ions.

Horie et al. (1987) first demonstrated that physiological concentrations of Mg2+i blocked the outward current through inwardly rectifying ATP-sensitive potassium channels of the heart.

Following removal of Mg 2+i, the current-voltage relationship became virtually linear. This observation was also confirmed by Matsuda et al. (1987) and Vandenberg (1987), who reported the loss of rectification in inward rectifier K+ currents following excision of an inside-out patch into divalent free internal solution. On addition of Mg2+ (~lmM) to the internal solution, rectification was restored. However, rectification was not observed following addition of either Ba2+ or Ca2+.

1.4.2d. Polyamine block

In addition to the voltage-dependent block of these channels by Mg 2+i, Matsuda et al. (1987) also observed the rapid closure of inward rectifier potassium channels during depolarization in the absence of Mg2"1). This suggested the existence of a voltage-dependent gating mechanism, which occurs independently of Mg2+i block.

Further studies on native (Ishihara et al., 1989; Oliva et al., 1990; Silver & DeCoursey, 1990) and cloned channels (see for example Stanfield et al., 1994a) confirmed that a time-dependent component of the inward rectifier K+ current existed irrespective of the presence or absence of Mg2+i, thus supporting the idea of an ‘intrinsic gating’ mechanism (see Leech & Stanfield, 1981). The voltage-dependent block by Mg2* was proposed to be responsible for the instantaneous inward rectification on depolarization. Thus, the instantaneous current on hyperpolarization was thought to reflect relief of Mg2* block.

In their 1994 review, Jan and Jan compared the intrinsic gating mechanism of inward rectifier potassium channels with the ‘ball-and-chain’ inactivation gate of Kv channels (see section 1.2.3a). Jan and Jan (1994) proposed that inward rectifier potassium channels might have retained the pore and cytoplasmic gate of an ancestral channel from which other channels, such as the Kv channels, also arose. However, in the same year, three independent groups

60 published results showing that the gating mechanism believed to be intrinsic was in fact a consequence of a voltage-dependent block of the pore by polyamines (Ficker et al., 1994; Lopatin et al., 1994; Fakler et al., 1994).

In the first of these reports, Ficker et al. (1994) described the loss of rectification following the formation of inside-out patches, from Xenopus oocytes expressing Kir2.1 channels, into Mg2+-free internal solution. However, application of spermine, spermidine and putrescine to the internal solution resulted in a block of the outward currents through Kir2.1. Lopatin et al. (1994) and Fakler et al. (1994) reported similar observations. Subsequently, several other studies have confirmed that a voltage-dependent block of inward rectifiers by polyamines underlies the ‘intrinsic’ rectification (see for example Fakler et al., 1995; Yamada & Kurachi, 1995; Yang et a l, 1995a; Ishiharaet al., 1996).

Polyamines are present in almost all cells and play important roles in protein synthesis, cell division and cell growth. Spermine, spermidine and putrescine are low molecular weight organic cations and at physiological pH the primary and secondary amine moieties of these molecules carry a charge. Lopatin et al. (1994) suggested that at potentials positive to Ek, these highly charged linear molecules could pass deep into the narrow pore of the channel to reach the blocking site. In a model to account for polyamine-induced inward rectification, two molecules of spermine were proposed to sequentially block the inward rectifier channel (Lopatin et al., 1995; see figure 1.5).

Lopatin et al. (1995) showed that the time constants for current activation in cell-attached patches expressing Kir2.1 closely matched the measured time constants for unblock of the channel by spermine, spermidine and putrescine. Thus, the time-dependent activation of inward current (see section 1.4.1c) represents unblock of the channel pore by polyamines. The strong voltage dependence of rectification exhibited by Kir2.1 is predominantly due to the effect of intracellular spermine, which is tetravalent. Fakler et al. (1995) suggested that the higher voltage dependence of rectification induced by spermine compared with spermidine was most likely due to its higher valency. Indeed, both the voltage-dependence and affinity of channel block by polyamines increase with the increasing valency of the polyamine molecule (Ficker et al., 1994; Lopatin et a l, 1994; Fakler et a l, 1994). Thus, whilst only nanomolar concentrations of spermine (z = +4) and spermidine (z = +3) are required to block Kir2.1 at positive voltages, putrescine (z = +2) blocks with only micromolar affinity.

61 A. B. C. D.

Outside

\ r

Spermine

Figure 1.5.

Mechanism of Inward Rectification. Schematic representation to illustrate the mechanism of polyamine block in inward rectifier potassium channels as proposed by Lopatin et al. (1995). Two molecules of spermine block the pore in a sequential manner, binding initially to a shallow site to block (B) then binding to a second site located deep (D) within the pore (see Lopatin et a l 1995; see also section 1.2.4d). Further evidence that a voltage-dependent block of Kjr channels by polyamines was responsible for intrinsic rectification came from studies in which inward rectification was examined following pharmacological modification of the cellular polyamine content. Bianchi et a l (1996) studied the native inward rectifier of rat basophil leukaemia (RBL-1) cells following treatment with an inhibitor of polyamine synthesis. Similarly, Shyng et al. (1996) studied recombinant Kjr2.1 channels in oocytes, which had also been injected with inhibitors of polyamine synthesis. Both studies reported a decrease in rectification and an increase in outward current through Kjr channels. Shyng et a l (1996) also expressed Kir2.3 in Chinese hamster ovary (CHO) cells deficient in ornithine decarboxylase, the enzyme that catalyses the synthesis of putrescine from L-omithine. These cells require exogenous putrescine to synthesize spermidine and spermine, and to maintain normal growth. Withdrawal of exogenous putrescine from these cells led to a relief of rectification in these strong inward rectifier channels.

1.4.2e. Intrinsic gating

Together, the aforementioned studies provide substantial evidence that channel block by polyamines is responsible for the strong inward-rectification of Kjr channels. However, it has been proposed that there is another component of intrinsic gating that is enhanced by the action of intracellular Mg2+ and polyamines, but is not mediated by polyamine block.

Aleksandrov et a l (1996) incorporated Kir2.1 channels into lipid bilayers to test the relative contribution of various mechanisms to inward rectification. Although polyamines are presumably absent from lipid bilayers, Kir2.1 channels still displayed an intrinsic, voltage- dependent rectification. Aleksandrov et a l (1996) proposed that in addition to the voltage- dependent block of the channel by Mg2* and polyamines, there was a distinct, fast gating process amplified by the binding of Mg2* and/or polyamines to sites on the cytoplasmic face. Unlike the blocking mechanism, it was dependent on the membrane potential and the extracellular and intracellular K* concentration. These results could not be explained by a simple pore-blocking model as originally hypothesized by Armstrong (1969), but were instead consistent with the electrochemical regulation of the gating mechanism proposed by Ciani et a l (1978).

Several other studies have since suggested the existence of an additional intrinsic gating mechanism. Shiehet a l (1996) observed a residual time-dependent inactivation of the Kir2.1 outward current after excising patches into Mg2* and polyamine-free solution, and unlike the

63 long-pore plugging effect of spermine, this intrinsic gating mechanism was sensitive to pHj. A similar voltage-dependent deactivation of Kir2.0 currents had previously been observed, following the prolonged washing of inside-out membrane patches with Mg2+- and polyamine- free solution (Ficker et al., 1994; Lopatin et a l, 1994; Fakler et al., 1994), but had been attributed to an incomplete washout of polyamines.

In the absence of Mg2* and polyamines, raising the pHj increased the inactivation rate of the outward current, but in their presence either decreased or had no effect on the rate of inactivation. Conversely, reducing the pHi had no effect on channel kinetics but blocked inward and outward currents in a voltage-dependent manner. In Kir2.1 D172N, a mutant channel with a reduced affinity for polyamines (see section 1.4.3), raising pHj had a similar effect on channel kinetics as in the absence of polyamines, accelerating inactivation at depolarized potentials. These observations led Shieh et al. (1996) to propose that the inactivation of outward current that remained after 5 minutes in Mg2+ and polyamine-free solution was not due to the slow washout of these substances, but represented a third mechanism for rectification that was likely to be intrinsic.

Shiehet al. (1996) expressed their uncertainty about whether this mechanism makes an important contribution to the inward rectification of Kir2.1 under physiological conditions. However, they suggested that it might play an important role in regulating the magnitude of outward current through Kir2.1 by way of its interaction with Mg2+- and polyamine-bound substates. This could be important in, for example, the heart where the action potential is long and inward rectifier potassium channels play a key role in regulating the onset of rapid repolarization (see section 1.5.3). This component of gating might also be important during ischaemia, when marked changes in pH occur. Under such circumstances, the resting K+ conductance would decrease in response to the fall in pH and in turn contribute to cellular depolarization, abnormal excitability and altered repolarization.

A more recent study by Lee et al. (1999) supports the idea of an intrinsic gate. By studying interactions of the spider venom toxins, philanthotoxin and argiotoxin with spermine and spermidine, Lee et al. (1999) provided evidence that polyamine block is more complex than a simple direct open channel block. These toxins, which are structurally similar to spermine at one end of the molecule, but have a bulky hydrophobic aromatic group at the other end, interfere with the ability of spermine or spermidine to block Kir2.1 channels. This effect cannot be readily explained by a direct open channel block mechanism, but may be accounted

64 for by an intrinsic gate mechanism in which high affinity block is due to a ‘spermine- intrinsic’ gate complex. It is thought that the multiple positive charges on the polyamine molecule facilitate the voltage-dependent interaction of the intrinsic gate-polyamine complex with a pore-docking site. Conversely, low affinity block may be explained by direct channel block of untethered spermine molecules.

The authors of this study recognized that they could not rule out the possibility of a compartmentalized pool of polyamines, which remained following extensive washing of excised patches. Nonetheless, their model represented an attractive alternative to the traditional direct channel block model, especially in light of the available structural information on the pore dimensions of a K+ channel (Doyle et al., 1998), since in this model only one spermine molecule would be required to block the pore. The distance from the internal surface to the selectivity filter for KcsA was estimated at 3nm, which could easily accommodate one spermine molecule at 2 nm in length but not two stacked lengthwise on top as proposed by Lopatin et a l (1995).

In the absence of structural data showing spermine bound to the channel pore it was impossible to conclude whether either of the models described above represented the true mechanism of ‘intrinsic’ rectification in strong inward rectifier potassium channels. The drawback with the long-pore plugging model was that from available structural information, it appeared that the channel pore could not accommodate two spermine molecules stacked lengthwise on top of one another. With evidence that spermidine is unable to permeate the pore of Kir2.3 (Sha et al., 1996), a model in which only one spermine blocks the pore seemed favourable at the time (Lee et al., 1999).

However, a recent study suggests that naturally occurring polyamines may act as permeant blockers (Guo & Lu, 2000b). Though Guo and Lu (2000b) acknowledged the existence of intrinsic gating, they argued that the model described by Lee et al. (1999) failed to account for the characteristic ‘hump’ in the current-voltage curve of strong inward rectifier potassium channels. Neither did this model account for the residual current remaining at positive potentials. On investigating the complex blocking nature of polyamines, they noted that block of Kir2.1 by diamines tends to a non-zero level at positive membrane potentials. This is also the case for cGMP-gated channels, in which diamines act as permeant blockers (Guo & Lu, 2000a).

65 Furthermore, recent x-ray crystallography studies of the N- and C-termini of an inward rectifier suggest that at 30A in length, the cytoplasmic pore of inward rectifiers could accommodate more than one polyamine molecule (Nishida & MacKinnon, 2002).

1.2.4f. Do channels exhibit intrinsic rectification?

The results from two recent studies, described below, suggest that inward rectifier potassium channels do not exhibit intrinsic rectification.

Guo and Lu (2000c) noticed that the degree of residual rectification in the absence of Mg2+ and polyamines, attributed in previous studies to intrinsic rectification (see section 1.4.2e), varied between studies from different laboratories. Guo and Lu (2000c) examined the effects of using different pH buffers in their experimental solutions. Interestingly, when HEPES or MOPS were used as the pH buffer, Kir2.1 currents exhibited profound relaxation at positive membrane potentials. In contrast, when phosphate or borate solutions were used current relaxation was removed, the steady-state I-V relationship becoming virtually linear. Some inward rectification remained at very positive membrane potentials (i.e. +100mV), which became more prominent on lowering the intracellular concentration of EDTA from 5 to ImM. Thus, the slight relaxation of current at +100mV in the presence of phosphate and 5mM EDTA appeared to be a consequence of block by intracellular cations, which was subsequently confirmed by its removal in a mutant channel (Kir2.1 D172N; see section 1.4.3) with a reduced affinity for intracellular cations.

The degree of current relaxation at positive membrane potentials was also demonstrated to vary with different concentrations of HEPES or different commercial sources of HEPES suggesting that the apparent intrinsic rectification observed in previous studies (see Aleksandrov et al., 1996; Shiehet al., 1996; Lee et al., 1999; Guo & Lu, 2000b) may be a consequence of block by HEPES and some impurity (Guo & Lu, 2000c). It has subsequently been confirmed that the residual rectification of Kjr2.1 in the absence of intracellular Mg2+ and polyamines can be accounted for by impurities in chemicals routinely used in recording solutions (Guo & Lu, 2002). This includes, for example, traces of hyrdroxyethylpiperazine (HEP) and ethylenediamine present in commercially available HEPES and EDTA, respectively. Inward rectification (and also hyperpolarization induced inactivation; see section 1.4. Id) can be completely removed in solutions using phosphate as a pH buffer and an optimal concentration of EDTA (~ 5 mM) to prevent block by intracellular cations, without introducing block by contaminating levels of ethylenedi amine.

66 Although several complicated models have been proposed to explain the mechanism of rectification in strong inward rectifiers, the exact nature of this process remains elusive. Whilst the results of Guo and Lu (2000c; 2002) suggest that Kir channels do not exhibit intrinsic rectification, others disagree (see Matsuda et al., 2003). It is perhaps worthy to note, however, that the glycine residue implicated in the gating of KcsA/MthK and Shaker channels (see section 1.2.3a; Jiang et al., 2002b; Yifrach & MacKinnon, 2002) is conserved throughout the Kir channel family.

1.4.3. Structural determinants of inward rectification

Studies in cloned inward rectifier potassium channels have greatly advanced our knowledge of the mechanism of inward rectification and its structural determinants (for reviews see Doupnik et al., 1995; Fakler & Ruppersburg, 1996; Jan & Jan, 1997; Nichols & Lopatin, 1997). In fact, the residues that determine gating in inward rectifier potassium channels were identified before the voltage-dependent block of the pore by polyamines was implicated as the mechanism for intrinsic rectification.

Initial studies focused on a negatively charged amino acid in the M2 transmembrane domain (Stanfield et al., 1994b; Wible et al., 1994; Fakler et al., 1994a). For all cloned strong inward rectifier potassium channels this residue is an aspartate (single letter code D). In Kjr2.1, mutation of this residue to the uncharged amino acids glutamine (D172Q) or asparagine (D172N) produced channels, which displayed weaker rectification and no time-dependence on activation (Stanfield et al., 1994b; Wible et al., 1994). The affinity for Mg2+i by the D172Q mutant channels was also reduced (Stanfield et al., 1994b), though D172N mutant channels still retained a high affinity for Mg2+i (Wible et al., 1994).

The corresponding residue in the weak inward rectifier potassium channel Kirl.la is an asparagine (N). Mutagenesis studies (Lu & MacKinnon, 1994; Wible et al., 1994) showed that substitution of aspartate for asparagine in Kirl.la (N171D) converted the channel from a weak to a strong inward rectifier potassium channel with an increased affinity for Mg2+i and ‘Kir2.1-like’ gating. Thus, a negatively charged amino acid in M2 was shown to be essential for intrinsic gating and to form part of the binding site for Mg 2+i.

Taglialatela et al. (1994) used chimeric constructs of Kirl.la and Kir2.1 to search for the molecular determinants of inward rectification. They found that exchange of the carboxyl

67 terminus of Kirl.la for that of Kir2.1 produced a strongly rectifying channel with a higher affinity for Mg 2+i, which suggested that the structural determinants for rectification were located in the C-terminus. A residue in the C-terminus, a glutamate (E224) in Kir2.1, was later identified as being an important determinant of inward rectification (Yang et a l, 1995a).

Later, other unknown sites were proposed to play a key role in determining the extent of rectification. sWIRK (see section 1.2.4d) displays weak inward rectification despite the presence of a negatively charged glutamate (E179) at the site corresponding to D172 in Kjr2.1 (Kubo et a l, 1996). Although sWIRK contains a glycine (G231) at the analogous position to E224 of Kir2.1, no differences in rectification could be distinguished when this residue was mutated to a glutamate (G231E). Thus, the differences in rectification between sWIRK and Kir2.1 are not due to this difference in this C-terminus residue. Further evidence for the involvement of additional residues in determining the extent of rectification comes from the observation that a mutant of K ^ .l, G210E, displays weaker inward rectification than a mutant of Kirl.la, N171E, despite the presence of identical residues at the two ‘rectification- controlling sites’ (Xu et a l, 2000b).

Another cytoplasmic residue, E299 in Kir2.1, was recently identified as being an important determinant of strong rectification (Kubo & Murata, 2001). The corresponding residues in all other strong inward rectifier potassium channels and in Kirl.la, which may be converted into a strong inward rectifier on mutation of the asparagine at position 171 to aspartate (see above), are negatively charged. In contrast, in sWIRK, Kir4.1 and Kir7.1 (which display weak inward rectification) this residue is a hydrophobic serine. The mutant channel Kir2.1 E299S displayed less intense inward rectification, and the hyperpolarized-induced activation and decay of outward currents at depolarized potentials were slower. Furthermore, mutation E299S reduced both the sensitivity to block by spermine of the outward current and the single channel conductance (Kubo & Murata, 2001). Kubo and Murata (2001) proposed that E224 and E299 are located at the inner vestibule forming an intermediate site to which spermine binds but does not plug the pore. It was proposed that the purpose of this site is to facilitate the entry and exit of spermine to and from its final pore-plugging binding site, which is located deeper within the pore.

Although the three residues discussed above have been shown to be critical determinants of the intensity of inward rectification, it is thought that other residues remain to be identified. For example, preliminary studies have shown that mutation of a residue (T51E) in the Q-

68 region (also known as MO) affects the voltage-dependent gating of Kirl.lb (Choe et al., 1997). Furthermore, the recently resolved crystal structure of the cytoplasmic pore of Kir3.1 has revealed many pore-lining residues, which remain to be altered by mutagenesis (Nishida & MacKinnon, 2002). It is likely that some of these amino acids play an important role in polyamine-induced rectification.

1.5. Physiological importance of Kir channels

Although Kir channels carry large inward currents, under physiological conditions these channels work at membrane potentials just positive to Ek or the resting membrane potential, since the membrane potential of real cells rarely goes below E&. Kir channels are involved in a number of physiological processes, some of which are considered below.

1.5.1. Kir channels underlie K+ homeostasis in the kidney

As discussed briefly in section 1.4.2a, Kirl.l plays an important role in the recycling of K+ in the thick ascending limb (TAL) and secretion of K+ across the apical membrane of mammalian cortical collecting (CCD) tubules in the kidney (for review see Wang et al., 1992; 2002). K+ recycling in the TAL is essential for the maintenance of the activity of the Na+-K+- 2C1' co-transporter, which is partly responsible for reabsorption of the filtered NaCl load. K+ secretion across the CCD tubules occurs in two steps: K+ is first pumped into the cell via the basolateral Na-K-ATPase, then K+ passively exits into the urinary space via native, small- conductance Kirs in the apical membrane (see Wang et al., 1990; Wang & Giebisch, 1991).

The role of Kirl.l channels in K+ recycling in the TAL is highlighted by the renal salt-wasting disorder Bartter’s syndrome in which mutations in Kirl.l cause loss of function (see also sections 1.0 and 1.2.4a). Mutations in Kirl .l are thought to result in an inability to reabsorb K+ from the thick ascending loop to the renal tubule, causing an inhibition of the Na+-K+-2C1- transporter and salt wasting. The loss of salts causes an increase in aldosterone levels, resulting in an increased Na+ absorption in exchange for K+ and H*, which causes hypokalemic alkalosis (for review see Sanguinetti & Spector, 1997). Animal knockout studies provide further support for the role of Kjrl.l channels in K+ homeostasis in the kidney. Kirl.l-deficient mice, in which Kirl.l expression is absent from the TAL and CCD, exhibit some of the characteristics of human Bartter’s syndrome including polyuria and salt wasting (Lorenz et al., 2002; Lu et al., 2002).

69 1.5.2. channels underlie K+-induced vasodilation

An increase in the activity of Kir channels has been proposed as one of the mechanisms that may contribute to K+-induced hyperpolarization of vascular smooth muscle and vasodilation (see for example Edwards et al., 1988; Knot et al., 1996; for review see Sobey & Faraci, 2000). Modest increases in [K+]0in the brain or heart act as a metabolic signal causing vasodilation of small cerebral and coronary vessels, which in turn may lead to a selective increase in the perfusion of metabolically active tissue. These modest increases in [K +]0 cause a shift in the channel gating properties of Kir channels (see figure 1.4) and a subsequent increase in the resting outward K+ current. Efflux of K+ through Kjr channels leads to hyperpolarization of the membrane and vasorelaxation by a mechanism involving the closure of voltage-gated calcium channels, which in turn leads to a reduction in the intracellular Ca2+ levels (for review see Standen & Quayle, 1998).

Native Kjr channels in arterial smooth muscle display characteristics similar to members of the Kir2.0 subfamily (see for example Quayle et al., 1996). Bradley et al. (1999) have identified transcripts for Kjr2.1 in arterial smooth muscle and showed that the biophysical and pharmacological properties of the native Kjr currents most closely resemble those of cloned Kir2.1 currents. More recently, Kir2.1 gene knockout studies have shown that cerebral arteries from mice lacking the Kir2.1 gene fail to dilate in response to modest elevations in extracellular K+ (Zaritsky et al., 2000), thus confirming the role of Kir2.1 channels in K+- induced vasodilation in cerebral arteries.

1.5.3. Kir channels regulate cardiac action potentials

7k i, the strong inward rectifier in the heart, serves several functions. Firstly, 7ki is responsible for maintaining the resting membrane potential of ventricular myocytes (for review see Lopatin & Nichols, 2001). 7ki is also involved in maintaining the long plateau phase of the cardiac action potential, owing to the decrease in membrane conductance at depolarized potentials, which helps to conserve metabolic energy (Isomoto et al., 1997). Finally, the outward current that flows at membrane potentials just positive to the reversal potential facilitates the final repolarization of the action potential (see for example Ibarra et al., 1991; Shimoni et al., 1992; for review see Lopatin & Nichols, 2001).

Functional studies have revealed that 7ki exhibits strong inward rectification and is sensitive to block by Ba2+ and Cs+ (see for example Sakmann & Trube, 1984a, b; Imoto et al., 1987; Matsuda et al., 1989). In keeping with the functional properties of 7Ki, more recent studies have provided evidence to suggest that the molecular correlates of Zki are members of the Kir2.0 subfamily (Brahmajothi et al., 1996; Nakamura et al., 1998; Zaritsky et al., 2001). In situ hybridisation studies have indicated expression of Kjr2.0 subfamily mRNA in ventricular myocytes, though the exact identity could not be resolved owing to the inability to design unique probes for each member of the Kir2.0 subfamily (Brahmajothi et al., 1996). However, evidence that Kir2.1 contributes to 7Ki has subsequently been provided using Kjr2.1 antisense oligonucleotides, which partially reduced I k i in cultured rat ventricular rat myocytes (Nakamura et al., 1998). More recently, mice lacking the Kir2.1 gene have been shown to exhibit no detectable inwardly rectifying K+ current (Zaritsky et al., 2001). This loss of 7ki resulted in broader action potentials and an increase in the amount of spontaneous activity in isolated ventricular myocytes. Kir2.1 knockout mice displayed bradycardia. Thus, the Kir2.1 gene appears to encode the channels primarily responsible for the 7ki conductance.

1.5.4. Kir channels help regulate the heartbeat

Cardiac muscle contains a G-protein-activated inward rectifier, known as the muscarinic acetylcholine potassium channel (K aoO , which is involved in the cholinergic regulation of the heartbeat (see Kovoor & Lester, 2002 for review). Acetylcholine (ACh), released following parasympathetic stimulation of the vagus nerve, activates Gi-coupled M2 muscarinic receptors, which leads to replacement of Ga-bound GDP with GTP and dissociation of GPy from Ga. As discussed in section 1.2.4c, the Gpy subunits bind to Kacii channels, causing them to open. Activation of 7kach leads to hyperpolarization of the myocyte membrane and, in turn, a lengthening of the pacemaker cycle. After prolonged exposure to ACh, the current desensitises owing to stimulation of M3 receptors by ACh and the subsequent activation of

Gotq, which in turn activates PLCP, resulting in hydrolysis of PIP 2 and reduction of the G- protein-activated current (for review see Jan & Jan, 2000; Kovoor & Lester, 2002).

1.5.5. Kir channels regulate blood glucose levels

In the endocrine pancreas, K atp channels formed from Kir6.2 and S U R 1 (see section 1.2.4f) play a pivotal role in the regulation of blood glucose levels and insulin release (see figure

1.6). A rise in blood glucose levels leads to an increase in the uptake and metabolism of glucose, elevation of the intracellular ATP concentration, and a subsequent ATP- concentration dependent inhibition of the K atp channel, which controls the cells resting membrane potential. The resulting depolarization of the plasma membrane opens voltage- dependent calcium channels and the subsequent influx of extracellular Ca triggers2+ the fusion

71 Insulin Ca -channels

♦ Glucose

Metabolism

Depolarization

Figure 1.6.

channels regulate blood glucose levels. Schematic illustration of the glucose metabolism pathway in pancreatic (3-cells (adapted from Baukrowitz & Fakler, 2000). Glucose metabolism

decreases the ADPiA I P ratio, which, in turn, reduces the activity of K a t p channels leading to membrane depolarization and the opening of voltage-dependent Ca2+ channels. Ca2+ influx through these channels leads to an increase in [Ca24]i, which stimulates insulin exocytosis.

72 of secretory granules with the plasma membrane and the release of insulin (for review see Baukrowitz & Fakler, 2000; Aguilar-Bryan et al., 2001; Houpio et al., 2002).

Sulphonylureas, such as tolbutamide and glibenclamide, which inhibit KA t p channels, are used in the treatment of non-insulin-dependent (type II) diabetes mellitus owing to their

stimulatory effect on insulin secretion. Conversely, K a t p channel openers such as diazoxide are used to inhibit insulin release (for review see Baukrowitz & Fakler, 2000; Aguilar-Bryan et al., 2001; Houpio et al., 2002).

1.5.6. Kir channels are important in development

As discussed briefly in the opening introduction, mutations in the gene encoding Kir2.1 channels have been shown to underlie Andersen’s syndrome. This hereditary condition is characterized by periodic paralysis, cardiac arrhythmias and dysmorphic features including short stature, curvatures of the spine and changes in the shape of the face (Plaster et al., 2001), which suggests an important role for Kir2.1 in development. This view is further supported by the observation that mice lacking the Kir2.1 gene display a complete cleft of the secondary palate (Zaritsky et al., 2000). Furthermore, additional mutations (e.g. R67W in the N-terminus, T192A in the C-terminus) in the Kir2.1 gene have been linked with altered cardiac and skeletal muscle phenotypes (Ai et al., 2002; Andelfinger et al., 2002). Heterologous expression of these Kir2.1 mutants revealed loss of function and variable dominant-negative effects when co-expressed with wild-type Kir2.1.

It is not known exactly why Kir2.1 channel expression is so critical for normal development, though it appears to be related to the function of Kir2.1 in setting very negative resting potentials. This is highlighted, for example, by the involvement of Kir2.1 in the development of skeletal muscle. Skeletal muscle fibres are formed by the fusion of mononucleated myoblasts, which is essential for skeletal muscle development, growth and repair. This process is Ca2+-dependent, requiring an influx of Ca2+ (for review see Bemheim & Bader, 2002). It is also dependent on membrane hyperpolarization, which occurs initially through EAG currents and then through Kir currents (see for example Shin et al., 1997; Liu et al., 1998; Fischer-Lougheed et al., 2001; also see for review Bemheim & Bader, 2002). The Kjr channel responsible for the hyperpolarization preceding fusion is Kir2.1. Myoblast fusion can be prevented following block of Kir2.1 by Ba2+ or Cs+ (Liu et al., 1998) and following inhibition of Kh-2.1 using antisense oligonucleotides (Fischer-Lougheed et al., 2001). Membrane hyperpolarization sets the resting membrane potential of ‘fusion-competent’

73 myoblasts within a suitable voltage range for Ca2+ influx through T-type Ca2+ channels to occur.

Further evidence for the role of Kir channels in development is provided by the weaver (wv) mouse, which carries a single point mutation (G156S) in the pore region of Kir3.2. The wv mouse exhibits numerous abnormalities including ataxia, hyperactivity and male infertility (for review see Hess, 1996). These abnormalities are associated with the loss of cerebellar granule cells and the loss of dopaminergic cells in the substantia nigra (see for example Rakic & Sidman, 1973). In the wv mouse, cerebellar granule cells fail to migrate from the external to the internal granule layer, their final destination in the mature cerebellum, and as a result die in the postmitotic zone of the external granule layer (see for example Smeyne & Goldowitz, 1989).

The G156S mutation substitutes a serine for the first glycine of the highly conserved GYG motif, which is important for K+ selectivity (see section 1.6). Following heterologous expression, homomers or heteromers (with Kir3.1) of Kir3.2 carrying the wv mutation display a loss of K+ selectivity, abnormal permeability for Ca2+ and constitutive channel activity (see for example Slesinger et al., 1996; Navarro et al., 1996; Kofuji et al., 1996; Silverman et al., 1996b). Constitutive activation results in high basal levels of Na+ influx that cause chronic membrane depolarization, which in turn results in cell death (see for example Slesinger et al., 1996; Navarro et al., 1996). Similar observations have been reported in cultured granule cells from wv mice (Kofuji et al., 1996).

Why the cerebellar granule cells in wv mice fail to differentiate and subsequently die has been a topic of much debate. Kofuji et al. (1996) suggested that the Na+ influx through wv K;r3.2 channels was the underlying cause of the wv phenotype since block of the Na+ influx in mutant granule cells allowed granule cell differentiation to proceed. In contrast, Surmeier et al. (1996) suggested that the loss of Kir3.2 currents, rather than the gain of a non-specific cation conductance, was responsible for the death of cerebellar granule cell neurones. Surmeier et al. (1996) suggested that the death of cerebellar granule cells might be a consequence of a sustained membrane depolarization and excessive Ca entry2+ and accumulation, which results from a failure to regulate membrane potential owing to the loss of the Kir3.2 current. However, studies in Kir3.2 knockout mice suggest that the wv phenotype is not a consequence of the loss of homomeric/heteromeric Kir3.2 channel function since Kir3.2 knockout mice do not exhibit abnormalities similar to wv mice (Signorini et al., 1997),

74 favouring the ‘gain of function’ hypothesis. Rossi et al. (1998) argue that the loss of Kir3.2 current and its regulation of membrane potential play an important role in cerebellar granule cell degeneration in wv mice. Both constitutively active and Kir3.2 currents were absent from granule cells at the premigratory stage, the stage at which granule cell degeneration begins in wv mice.

What is the role of functional Kir3.2 channels in cerebellar development? When cerebellar granule cells become postmitotic they are exposed to elevated levels of extracellular glutamate, which activates N-methyl-D-aspartate (NMDA) receptors, leading to a rise in cytosolic Ca2+ levels that trigger migration. At the same time, metabotropic glutamate receptors are activated, which potentiate Ca2+ currents and activate Kir3.2 currents, which provide a hyperpolarizing current that is thought to counterbalance the depolarizing, ionotropic glutamate currents (see Surmeier et al., 1996).

1.6. Selectivity

All potassium channels exhibit very similar ion permeability characteristics, being at least 100 fold more permeable to K+ than to Na+ (Hille, 2001). Na+ ions are smaller than K+ ions, so how do potassium channels distinguish K+ ions from Na+ ions with such high fidelity without

o disrupting the flow of ions at a rate (10 ions per second) near the diffusion limit? In this final section of my introduction I will discuss what is known about the molecular basis of selectivity in potassium channels. I will begin by introducing some of the early mutagenesis studies that were undertaken in Kv channels to determine which region of the channel constitutes the ion conduction pore. I will then briefly consider two of the major theories for the molecular mechanism of selectivity. Finally, I will discuss the features of ion permeation in potassium channels revealed by the crystal structure of KcsA (Doyle et al., 1998b).

1.6.1. Structural determinants of K+ selectivity in Kv channels

Studies of the molecular basis of selectivity in potassium channels began in the early 1990s, shortly after the cloning of Shaker and its related potassium channel genes (see section 1.2.1). Mutagenesis studies initially concentrated on the S5-S6 loop, following the finding that mutations in this region disrupted the binding of CTX, a scorpion venom toxin that inhibits several types of potassium channel (MacKinnon & Miller, 1989). Because CTX is physically large, it was concluded that the amino acids with which it interacted must lie in the outer vestibule of the pore.

75 Later, through the use of the smaller ion TEA+, MacKinnon and Yellen (1990) were able to identify residues lining the pore. This approach led to the identification of two positions (431 and 449) in the primary structure of the Shaker potassium channel that seemed to be located in or near the outer mouth of the ion conduction pore.

Experiments to determine whether the S5-S6 loop of Kv channels constituted part of the pore continued, and the results from several independent studies soon established that the conduction pathway was indeed located in this region. First, Yellen et a l (1991) demonstrated that internal TEA+ blockade was altered by mutation of a residue (T441 in Shaker ) at a position almost midway between the two residues (431 and 449) previously implicated in external TEA+ blockade. Next, Yool and Schwarz (1991) demonstrated that mutation of single amino acids in the corresponding region of ShB channels, including threonine 441, altered channel selectivity, affecting N H / and Rb+ permeability. Shortly after, Brown and his colleagues (Hartmann et al., 1991; Brown, 1994; Brown et a l, 1994) substituted a stretch of 21 amino acids, proposed to form the pore of the Kv3.1 clone for the corresponding region in a phenotypically different Kv channel, Kv2.1. The resulting chimera (CHM) displayed a single channel conductance and TEA blocking profile characteristic of the donor channel, whilst its voltage sensitivity resembled that of the host channel. This suggested that the ion conduction properties were specified entirely by the donor peptide module without modification by the host channel.

Subsequent studies by the same group concentrated on the individual residues within this stretch of amino acids to establish exactly which residues determined the phenotypic differences between Kv2.1 and Kv3.1 (Kirsch et a l, 1992; Taglialatela et a l, 1993). Of the 21 residues, only 12 were identical. Mutations of the remaining 9 residues were made and the valine at position 374 of Kv2.1 (leucine 401 in Kv3.1) was soon identified as the residue responsible for the functional differences between Kv2.1 and Kv3.1.

Together, the results from the aforementioned studies implied that the stretch of amino acids between the S5 and S6 segments (now known as the P-region) formed the conduction pathway. However, if this region was a-helical in structure it could not possibly span the entire membrane. Yellen et a l (1991) considered this possible, suggesting that the channel could have an hourglass-like structure with a short, narrow pore and wider vestibules at one or both ends. However, a P-barrel structure for the pore was favoured at the time (Yellen et a l,

76 1991; Hartmann et a l, 1991; Brown et a l, 1993). The residue at position 374 in Kv2.1 (valine 443 in Shaker ) was considered to be positioned at an ‘’, hence its importance in determining ion conduction (Kirsch et a l, 1992).

These early studies painted a somewhat oversimplified picture of the channel pore, as it soon became evident that both the S4-S5 linker and S6 segment also contribute to the pore. Mutations in the S4-S5 linker were found to alter unitary conductance (Isacoff et a l, 1991), Rb+ selectivity and blockade of the channel by internal TEA+, Ba2+ and Mg2+ (Slesinger et a l, 1993). Mutations in S6 also had effects on single channel conductance and block by internal TEA+ and Ba2+ (Lopez et a l, 1994). Thus, the existing model of the potassium channel pore was modified from one in which the pore was comprised mainly of P-region segments to a more sophisticated model in which the long pore consisted of structural elements from the P- region, the S6 segment and the S4-S5 linker. The P-regions were thought to comprise the external region of the pore, in which the selectivity filter (see below) could be found, whilst the S6 segments and the S4-S5 loops were thought to form the inner part of the pore.

1.6.2. Theories of selectivity

Prior to the resolution of the crystal structure of KcsA, two major hypotheses had been offered to explain how potassium channels might select for K+ ions over Na+ ions. The first hypothesis, proposed by Bezanilla and Armstrong (1972) and Bertil Hille (1973), suggested that selection against Na+ occurred in the narrowest part of a potassium channel, which Bezanilla and Armstrong (1972) called ‘the tunnel’. Approximately 3.0-3.3A in diameter, the tunnel was considered to consist of several oxygen-containing groups fixed rigidly in position to provide a good fit for a K+ ion but not for a Na+ ion. A K+ ion would be required to shed its waters of hydration in order to permeate the pore. Thus, selectivity was suggested to depend on the crystal diameter of the ion. Furthermore, K+ ions would be kept from passing one another by mutual repulsion (Hille, 1973).

Kumpf and Dougherty (1993) proposed an alternative hypothesis for ion selectivity in potassium channels. Kumpf and Dougherty (1993) noticed that the sequence of the pore region of a number of potassium channels was surprisingly hydrophobic and contained a large number of aromatic residues. Their results from computational modelling suggested that cation-Tt interactions could be responsible for ionic selectivity, K+ in the pore being coordinated in a cage of n electrons generated at the face of a ring of aromatic residues (Kumpf & Dougherty, 1993). In support of this hypothesis, it had previously been shown that

77 K+ selectivity in Shaker was partly determined by the presence of the tyrosine (Y445) of the GYG motif (Heginbotham et al., 1992).

A more detailed study by Heginbotham and her colleagues later demonstrated that a highly conserved stretch of eight amino acids (single letter code - TXXTXGYG) known as the ‘potassium channel sequence’ was important for selectivity (Heginbotham et al., 1994). The hydroxyl groups of the three threonines (in Shaker there is also a threonine (T441) at position 3 of the signature sequence) and the aromatic ring of the tyrosine seemed prime candidates for interacting with and selecting for K+ ions. By mutating these eight amino acids and measuring permeability ratios, it was concluded that the hydroxyl groups at positions three and four of the signature sequence and the aromatic ring at position seven were not essential for K+ selectivity. Because the aromatic group was not required for K+ selectivity, it was suggested that K+ selectivity was not conferred by cation-7t interactions, but instead by interactions with oxygen atoms as first proposed some 20 years earlier (Bezanilla & Armstrong, 1972; Hille, 1973).

1.6.3. Molecular basis of selectivity

We now know from the crystal structure of the Streptomyces lividans potassium channel KcsA that the selectivity filter is formed by backbone carbonyl oxygen atoms from five amino acids (TVGYG), including the two conserved glycine residues in the signature GYG motif (see figure 1.7A; Doyle et al., 1998b). The carbonyl oxygen atoms of the second conserved glycine residue are thought to assist in the hydration and dehydration of a K+ ion at the extracellular entryway (Zhou et al., 2001b). Once dehydrated, the partially negatively charged carbonyl oxygens of the selectivity filter mimic the waters of hydration forming four potential ion-binding sites arranged so that a dehydrated K+ ion fits exactly, binding tightly to the pore (Doyle et al., 1998b; Morais-Cabral et al., 2001; Zhou et al., 2001b). In contrast, a Na+ ion would fit so loosely that it energetically favours to remain in aqueous solution.

Whilst this accounts for the high selectivity exhibited by potassium channels, such tight binding of a K+ ion to the pore would not allow for the high rates at which K+ moves through the channel. How is this achieved? The ion conduction pore contains three binding sites in single file, as originally hypothesized by Hille and Schwarz (1978). The hydrophobic tunnel of the selectivity filter, adjusted to a width of 3A by a sheet of aromatic amino acids in the surrounding pore helices, accommodates two of these K+ ions, which are located close enough (7.5A apart) that the ions electrostatically repel each other (see figure 1.7B). This

78 electrostatic repulsion lowers the binding affinity for a third K+ ion residing in the inner mouth of the channel, displacing it out of the selectivity filter and into a water-filled cavity, the second ion then taking its place.

Remarkably, the overall structure of this potassium channel validated many of the previous findings from functional studies with Kv channels. For example, the outer vestibule possesses the right shape to bind scorpion toxins related to CTX (MacKinnon et a l, 1998) and, as for other potassium channels, TEA and Ba2+ block both the inner and outer pores (Heginbotham et al., 1999). Thus, the overall architecture of the potassium channel pore may be conserved through evolution.

79 Figure 1.7. Features of ion permeation in potassium channels revealed by the crystal structure of KcsA. A, Ribbon representation of two of the four subunits of the KcsA channel viewed from the side. The selectivity filter (coloured red) is formed by backbone carbonyl oxygen atoms from five amino acids including the GYG triplet. B, The selectivity filter of KcsA contains two K+ ions (coloured green), located at opposite ends. Mutual electrostatic repulsion between adjacent K+ ions underlies rapid permeation. (From Doyle et al., 1998b) 1.7. Aims

Although the atomic structure of a bacterial potassium (KcsA) channel pore has been determined using X-ray crystallography (Doyle et al., 1998), a physical picture of a Kjr channel pore has yet to be resolved. With a two-transmembrane-helix topology, Kjr channels are similar in their overall structure to KcsA yet share little homology at the amino acid level. In contrast, the amino acid sequence that forms the pore is well conserved between Kv channels and KcsA. It has been suggested that the ion conduction pore is conserved among all potassium channels (Lu et al., 2001b). However, the M1-P-M2 region of eukaryotic Kir channels exhibits numerous features that distinguish them from other potassium channel (see for example Kubo et al., 1998; Durrell & Guy, 2001). Until three-dimensional pictures of Kjr and Kv channels are readily available, mutagenesis studies remain a powerful tool in determining the contribution of specific amino acids to channel function. Here, by combining site-directed mutagenesis with studies of ion block, conduction and ionic selectivity, I have investigated the contribution of residues in the P-region to the permeation properties of Kir2.1 and Kir2.2.

81 Chapter 2

Materials & Methods

2.1. Molecular biology

2.1.1. Cloning and site-directed mutagenesis

For expression in Chinese hamster ovary (CHO) cells, EcoRI/XhoI and Hindlll/Xhol fragments containing the entire coding region of the mouse brain Kir2.1 (Stanfield et a l, 1994a) and Kjr2.2 genes, respectively, were sub-cloned into the expression vector pcDNA3. Substitution mutations, with the exception of Kir2.1 F147L, were generated by oligonucleotide-directed in vitro mutagenesis either by the method of Taylor et al. (1985) using commercially available kits (Amersham Pharmacia Biotech), or by the method of Kunkel (1985). Mutation F147L was generated using the ‘Quik Change’ mutagenesis kit (Stratagene). For expression in murine erythroleukaemia (MEL) cells, a Hind IU/XhoI fragment containing the entire coding region of the mouse brain Kir2.2 gene, ligated to the P- globin locus control region (LCR)-promoter (see section 2.2.6), was sub-cloned into the MEL

82 expression vector pNV3 (Shelton et a l, 1993). DNA sequencing of the entire Kir cDNAs was performed to verify mutations.

2.1.2. Purification of plasmid DNA

It is important that the DNA used in transfections is not contaminated with RNA, bacterial chromosomal DNA and protein, and is generally in good condition (Ehlert et a l, 1993). To reduce such contaminants, plasmid DNA encoding wild-type and mutant Kir channels was purified using either commercially available kits, such as the Wizard® Plus SV Minipreps DNA Purification System (Promega Corporation) or the Qiagen Maxi Plasmid kit (Qiagen), or by banding on a caesium chloride gradient as described by Bimboim and Doly (1979).

2.2. Cell culture

In order to minimize the risk of infection of cell populations by bacteria and fungi, all cell culture was carried out in a class II laminar flow hood (Medical Air Technology Ltd) utilizing aseptic techniques. Surfaces were swabbed with 70% prior to use and latex gloves were worn at all times.

2.2.1. Cell culture reagents and plastic-ware

All cell culture reagents were purchased from Gibco BRL. Tissue culture grade flasks, 12- well plates and 35mm petri dishes were purchased from Nunclon™. All other plasticware was purchased from Falcon. The tissue culture flasks are made of sulphonated polystyrene, which has a negatively charged surface. Adhesion between this surface and the negatively charged surface of the cell membrane occurs by way of divalent cations and basic proteins, which form a layer between the surface of the plastic and the cells (Butler, 1991; 1996).

2.2.2. Incubator

Cells were incubated at 37°C in 10% carbon dioxide (CO 2) atmosphere at 95% humidity using a Galaxy CO 2 incubator (Scientific Lab Supplies, Ltd.). The high level of humidity in the incubator prevents excessive evaporation from the tissue culture flasks and plates whilst

10% CO2 is used to maintain the cultures at their optimal pH (6.9-7.4) by way of the bicarbonate-C02 buffering system. Whilst a difference of one or two degrees below 37°C simply slows cell growth, higher temperatures result in cell death. It is essential, therefore, that the temperature of the incubator is not allowed to rise above 37°C.

83 2.2.3. Cell lines

From their study of the growth potential of human embryonic cells, Hayflick and Moorhead (1961) concluded that ‘normal’ cells have a finite lifespan. Human embryonic cells could be repeatedly sub-cultured for approximately 50 generations, beyond which they were incapable of further growth. However, some cells may display an infinite growth capacity following a process known as transformation, in which they lose sensitivity to the stimuli that control their growth. These populations of transformed cells are often described as ‘continuous’ cells growing well in simple growth media, which has not been supplemented with growth factors (Butler, 1991; Butler, 1996). Two different cell lines have been used in the studies described within this thesis. Outlined below are the general properties, maintenance and preparation of CHO and MEL cells, respectively.

2.2.4. Growth of cells in culture

Cells in culture follow a pattern in which they go through several different phases of growth. These are described as lag-, growth-, stationary- and decline phases. The lag phase represents a period of time during which the synthesis of cellular growth factors occurs and there appears to be no increase in cell density. The duration of this phase is dependent on the viability of the cells and the density of cells used at inoculation, the lower the initial inoculation density the longer the phase. It is thought that the cells enter the growth phase once the concentration of growth factors reach their optimum levels. The number of cells initially increases exponentially during the growth phase, following which the cells enter into a stationary phase during which there is no further increase in cell density. Although the cells are no longer growing at this stage, they may remain metabolically active. Finally, cells enter into a decline phase during which cell death occurs. This is often in response to depletion of nutrients. In order to ensure continued growth of cells in a new culture, following sub­ culture, it is best to passage the cells whilst they are still growing. Cells split from the stationary phase will grow but the lag phase will be longer following inoculation (Butler, 1991; 1996).

2.2.5. Chinese hamster ovary (CHO) cells

CHO cells were originally derived from an ovarian biopsy of a female Chinese hamster in 1957 and have since been established as a very stable, ‘continuous’ cell line (for review Butler, 1991; Doyle et al., 1998a). The consistency and reproducibility of results that can be acquired by using a cell line comprised of a single cell type, such as CHO cells, in addition to the ease with which such cells may be maintained in culture, make them highly suitable for biological studies of channels and receptors. Furthermore, only three types of voltage- dependent ionic membrane conductances, a Ca2+-activated K+ current, a voltage-activated Na+ current and a voltage-activated Ca2+ current similar to the L-type Ca2+ current, have been recorded from untransfected CHO cells (Skryma et al., 1994). Since these epithelial-like cells appear to lack endogenous inward rectifier potassium channels, and may be easily transfected (see section 2.3) with cDNA encoding the protein of interest, they provide an ideal system for the expression of cloned Kir channels.

2.2.5a. Routine growth and maintenance of CHO cells

CHO cells were grown in 80cm3 tissue culture flasks and were incubated at 37°C in 10% CO 2 atmosphere at 95% humidity. Growth was found to be optimal in alpha minimal essential medium (MEM-a) with Glutamax-1 (L-alanyl- L-glutamine), without ribonucleotides and without deoxyribonucleosides. Standard incubation medium was supplemented with 10% (v/v) foetal bovine serum (FBS). Medium was stored at 4°C but was incubated at 37°C for 30 minutes prior to use.

2.2.5b. Subculture of CHO cells

Cells were grown to approximately 80% confluency and split twice weekly. The medium was removed and the cells washed with 10ml Ca2+/Mg2+-free phosphate buffered saline (PBS) to remove any traces of serum, which contains excess proteins that deactivate trypsin and glycoproteins such as fibronectin, which provide a surface coating aiding cell attachment. To dislodge cells from the substratum, approximately 1ml of (lx) trypsin-ethylenediaminetetraacetic acid (EDTA; TE) was added to the flask and left for 5 minutes. Trypsin is a proteolytic enzyme, which aids detachment of the adherent cells by breaking down the proteins that bind the cells to the substratum. In addition, the EDTA chelates any remaining divalent cations.

Once the cells had detached from the sides of the tissue culture flask, 9ml of standard incubation medium was added to inactivate the trypsin. The cell suspension was then transferred to a centrifuge tube and centrifuged for 2 minutes at lOOOg at room temperature. The supernatant was discarded under sterile conditions and the pellet resuspended in 10ml standard incubation medium. In general, 200-400pl of cell suspension was transferred into 15ml of fresh medium for the maintenance of the cell line and between 10-150pl per well of a

85 12 well plate for transient transfection. For consistency in experiments, CHO cells were not used beyond passage 45.

2.2.5c. Preparation of CHO cells for electrophysiological recording.

In general, the transfected CHO cells were ready for electrophysiological recording 12-60 hours after transfection. To dislodge cells from the bottom of the dish, the medium was removed and the cells washed with sterile Ca2+/Mg2+-free PBS. 300pl of TE were then added and left for approximately 5 minutes. To inactivate the trypsin, 1ml of standard incubation medium was added and the cell suspension transferred to a 15ml centrifuge tube and centrifuged for 2 minutes at lOOOg at room temperature. Following centrifugation, the supernatant was discarded and the cells were resuspended in approximately 1ml serum free medium. Cells were plated out into 35mm tissue culture dishes containing the appropriate bath solution. For single channel recording, dishes precoated with poly-L-lysine were used. Under sterile conditions, poly-L-lysine (Sigma) was dissolved in 50ml ‘milliQ’ water and 1ml was used to coat each 35mm tissue culture dish. After 5 minutes, the solution was aspirated from the dishes and the surfaces washed twice with sterile water. The dishes were left to dry overnight and stored at 4°C until required.

2.2.6. Murine erythroleukaemia (MEL) cells

MEL cells are erythroid progenitor cells, which when transformed by the Friend virus complex are arrested at the proerythroid stage of development. They were used in the initial stages of this study in conjunction with the (3-LCR promoter, which is erythroid specific. Using such a system can be beneficial since expression of ion channels can be achieved independently of the position of integration of the transfected gene within the host genome. Such ‘position effects’ are typically responsible for variations in the gene expression level (Shelton et al., 1993; the reader is referred to this reference for a more detailed description of the use of the MEL cell expression system).

MEL cells can be induced to undergo a pattern of terminal erythroid differentiation using chemical agents such as dimethyl sulphoxide (DMSO), during which they synthesize high levels of globin proteins. In this study, the Kir2.2 gene was ligated to the P-LCR promoter, thus functional expression of Kir2.2 correlated with the DMSO-induced production of globin proteins.

86 As with CHO cells, the low endogenous expression of ion channels in MEL cells make them suitable for expression of Kjr channels. Only three channels have been observed in untransfected MEL cells: a Ca2+-activated K+ channel, a voltage-independent, monovalent cation, stretch activated channel and a Cl" channel (Arcangeli et al., 1987).

2.2.6a. Routine growth and maintenance of MEL cells

MEL cells were grown in 25cm3 tissue culture flasks and were incubated at 37°C in 10% C 02 atmosphere at 95% humidity. Growth was found to be optimal in Dulbecco's Modified Eagle’s Medium (DMEM) with Glutamax-I, with 25mM HEPES (7V-[2-Hydroxyethyl] piperazine-A/-[2-ethanesulphonic acid]), and without sodium pyruvate. The medium was supplemented with 10% v/v FBS, and geneticin® antibiotic (G418; Sigma) was added to give a final concentration of 0.4mg.mrJ to select for transfected cells (see section 2.3.5). Under sterile conditions, G418 was dissolved in PBS to give a concentration of 200mg.ml'1. The G418 stock solution was then filtered using a 0.22/xm bottle top filter (Nunclon™), divided into 1ml aliquots and stored at -20°C until required. Medium was stored at 4°C but was incubated at 37°C for 30 minutes prior to use.

2.2.6b. Subculture of MEL cells

MEL cells were grown to approximately 90% confluency and split twice weekly. The cells were dislodged from the bottom of the flask by gently tapping the sides. In general, 30-50|il of cell suspension was transferred into 6ml of fresh medium for flasks to be induced 3-4 days after subculture and 60-100pl into flasks to maintain the cell line.

2.2.6c. Induction of MEL cells

Transfected MEL cells were induced to undergo differentiation, and hence potassium channel gene expression, by applying 2% v/v DMSO. This typically involved adding 120pl of DMSO to 6ml of cell suspension approximately 3-4 days after subculture when the cells were approximately 50% confluent.

2.2.6d. Preparation of MEL cells for electrophysiological recording

Typically, electrophysiological recordings can be made from MEL cells between 2-4 days post-induction. Induced cells were dislodged from the bottom of the flask by gently tapping the sides and a 200-500/U aliquot of cell suspension transferred into a 15ml centrifuge tube.

87 The cell suspension was then diluted to 5ml using extracellular recording solution (section 2.4.5) and centrifuged for 2 minutes at lOOOg at room temperature. The supernatant was then discarded and the cell pellet resuspended in 1ml extracellular solution. As for CHO cells, MEL cells were plated out in 35mm tissue culture dishes containing extracellular solution.

2.2.7. Storage of cell lines

In order to prevent loss of an established cell line by contamination, cells were stored cryogenically frozen under liquid nitrogen. Cells can be stored indefinitely in freezing medium containing a cryoprotectant such as DMSO, which protects cells from disruption by preventing build up of electrolytes during the freezing process (Lovelock & Bishop, 1959). For maximum cell survival it was important to freeze the cells slowly and thaw rapidly.

Cells to be frozen down (either untransfected CHO or MEL cells or stably transfected MEL cells) were grown to confluency in 30ml of growth medium in 175cm3 tissue culture flasks (Nunclon™). Cells were dislodged from the side of the flask as described above for each cell line. The density of cells was determined by performing a cell count using a haemocytometer (see section 2.2.8) and the appropriate volume of cell suspension containing 5 x 106 cells transferred into a sterile 15ml centrifuge tube and centrifuged for 2 minutes at lOOOg at room temperature. The supernatant was poured off under sterile conditions and the pellet resuspended in 1 ml of freezing medium containing 50% v/v FBS and 10% v/v DMSO.

200j l i 1 aliquots of this cell suspension were immediately transferred into previously labelled 1.25ml cryotubes and placed into a -80°C freezer overnight. The cryotubes were then transferred on dry ice to a liquid nitrogen refrigerator.

To restart a stock from frozen, a cryotube containing the required cell type or mutant was removed from the liquid nitrogen store and placed on dry ice for transport. The cell suspension was thawed by immersing the cryotube in a water bath at 37°C, and transferred to a 15ml centrifuge tube using a P200 Gilson pipette. The cells were gently resuspended by adding 10ml of prewarmed medium drop by drop and were then transferred to a 175cm3 tissue culture flask containing an additional 20ml of growth medium. Safety goggles and protective gloves were worn at all times when handling frozen cryotubes in case of explosion. 2.2.8. Haemocytometer cell counts

Haemocytometer cell counts were often performed to determine cell density per ml of medium and hence, the appropriate volume of cell suspension to use when freezing cells or plating cells out prior to transfection.

A haemocytometer consists of a thick glass slide, the counting chamber and coverslip, which when placed in the correct position gives a chamber depth of 0.1mm. The glass slide is typically engraved with a nine square counting grid, each square having dimensions of 1mm2, some of which are divided further into 16 or 25 smaller squares (see Doyle et al, 1998a).

Before use, the haemocytometer (Neubauer) was thoroughly cleaned using 70% IMS and the coverslip placed in the correct position on the glass slide. The appearance of Newton’s rings, ‘rainbow-like’ interference patterns, which are visible on holding the haemocytometer up to the light, indicated when the coverslip was in the correct position. The cell suspension was mixed and 50/xl removed and mixed with an equal volume of Trypan blue solution, which penetrates the membrane of non-viable cells, staining them blue. The haemocytometer chamber was filled by capillary action, a Gilson tip containing the cell suspension and Trypan blue mixture being placed at the junction between the coverslip and counting chamber. The haemocytometer was then placed on a microscope, the engraved area brought into focus and, under low magnification, the number of unstained cells in each of two 1mm2 areas counted. In general, cells in the left-hand and top grid markings were included whilst those in the right- hand and bottom markings were excluded (Doyle et al, 1998a). The total number of viable cells was then calculated as follows:

Total number of viable cells = Total cell volume (ml) x 2 (dilution factor in Trypan blue) x mean number of unstained cells (unstained count -s- 2) x 104.

If the cell count was low, more than two 1mm2 areas were counted and averaged.

2.3. Transfection methods

Transfection involves the transfer of nucleic acid into mammalian cells and can be used to produce cells yielding a large quantity of a single protein for the study of the regulation and function of genes. The incorporation of DNA into mammalian cells was first demonstrated in the 1960s. Using isolated human nuclear DNA, Szybalska and Szybalska introduced a

89 functional gene for the hypoxanthine-guanine phosphoribosyltransferase (HGPRT) enzyme into human cells lacking an endogenous form of the enzyme (for reviews see Butler, 1996; MacDonald, 1991).

A variety of methods are now employed for the transfection of DNA into mammalian cells. These include: calcium phosphate coprecipitation, DEAE-dextran mediated transfection, electroporation, lipid-mediated transfection and the use of viral vectors. Whilst some of the above methods are more suited for short-term expression others are more effective for the stable expression of DNA. The transfection methods used for establishing the transient and stable expression systems used throughout this study are described below.

2.3.1. Lipid-mediated transfection

Early methods for lipid-mediated transfection involved the incorporation of DNA into mammalian cells by way of liposome vesicles (Feigner et al., 1987), formed from positively charged lipid and negatively charged DNA, which bind to the cell and are then internalised by endocytosis (for reviews see Doyle et al., 1998a; Otter-Nilsson & Nilsson, 1999). A variety of cationic lipids are commercially available including Invitrogen’s PerFect Lipid™ range, with pFx™-8 giving optimum transfection efficiency in CHO cells. The incorporation of foreign DNA into mammalian cells by way of transfection is often a very inefficient process. It is, therefore, essential to have a method for selecting those cells that are expressing the gene of interest. In this study, enhanced green fluorescent protein (eGFP) was used as a marker for successful transfection. Expression of eGFP in combination with an ion channel produces a bright green fluorescent signal when viewed with conventional epi-fluorescence optics for fluorescein isothiocyanate. Fluorescent, cotransfected cells invariably express the introduced ion channel (Marshall et al., 1995).

2.3.2. Transient transfection of CHO cells using PerFect lipid™

CHO cells were plated out 24 hours prior to transfection at a density of 1 x 104 cells per well of a 12 well plate. Per reaction, 1ml of serum free medium was placed into each of two polystyrene universals (Bibby Sterilin Ltd.). Universals made from polystyrene were used because, unlike glass or polypropylene, polystyrene does not attract the concentration- and ionic strength-dependent aggregates which can form during the reaction between lipid and DNA (Doyle et al, 1998a).

90 Next, the lipid (pFx™-8) was removed from storage at -20°C and allowed to equilibrate to room temperature. Once warm, 12pl of the lipid were transferred into one of the two universals. Meanwhile, lpg of eGFP DNA and 4pg of channel DNA were added to the medium in the second universal. The contents of the two universals were then mixed and set aside. The medium was aspirated and the cells washed with sterile Mg2+/Ca2+-free PBS to remove any traces of serum. 1ml of the medium containing the DNA and lipid was then added to each well (2 wells per transfection). The cells were incubated for 4 hours following which the medium was removed and replaced with 1ml standard incubation medium, and the cells returned to the incubator for 12-60 hours.

2.3.3. Transient transfection of CHO cells using FuGENE™ 6

The non-liposomal transfection reagent FuGENE 6 (Roche Molecular Biochemicals) provides an alternative method for lipid-mediated transfection of eukaryotic cells. The method offers the advantage that transfection can be performed in the presence of serum. The need to remove any traces of serum prior to transfection and to replace the serum following incubation of the cells with the transfection reagent is therefore abolished. This offers the convenience of being able to perform transfections at the end of the day, aiding experimental design.

CHO cells were plated out 24 hours prior to transfection as described above. 100/xl of serum free medium was placed into a flip-top eppendorf to which 3/H of FuGENE™6 was added, being careful not to allow the lipid to come into contact with the sides of the eppendorf tube. The diluted transfection reagent was incubated at room temperature for 5 minutes. Meanwhile, I fig of channel cDNA and 0.5 fig of GFP were mixed together in a separate eppendorf and put aside. Following incubation, the diluted transfection reagent was added to the DNA in the second eppendorf, mixed and incubated for a further 15 minutes at room temperature. 50 fi\ of the transfection mixture were removed and added to cells in a single well of a 12 well plate. This was repeated for a second well using the remaining 50fi\ of transfection mixture. The plates were then gently swirled to ensure even dispersal and returned to the incubator.

2.3.4. Levels of expression

Following transient transfection using PerFect lipid or FuGENE™ 6, current sizes were typically large, often in excess of 5-10nA. Consequently, large series resistance errors (see

91 section 2.4.14) were frequently encountered. In an attempt to reduce current size and subsequent problems with series resistance errors, the transfection protocols described above were modified. Figure 2.1 shows the average current sizes in pA/pF following various changes made to the transfection protocols (see below). Since previous studies have shown a correlation between the intensity of fluorescence and level of channel expression (Marshall et al., 1995), both bright and faint cells (i.e barely fluorescing) were patched. The following changes were made:

2.3.4a. The amount of cDNA used for PerFect Lipid transfections was reduced

The total amount of cDNA used per transfection was normally 5/xg (see section 2.3.2), 4fig of which were channel cDNA and the remainder marker cDNA. To determine whether current density could be reduced, by using a lower amount of cDNA, the total cDNA used per transfection was halved, maintaining the ratio of channel to marker. Figure 2.1 A shows that using lower amounts of cDNA per transfection did not significantly alter current density (students unpaired t-test, P > 0.05). However, there was a significant difference in current density between bright and faint cells.

2.3.4b. The incubation time for PerFect Lipid transfections was reduced

According to the manufacturers instructions, the optimal incubation period for transfection with Perfect Lipid is 4 hours. Thus, to test whether sub-optimal incubation times gave smaller currents the incubation time for transfection was reduced to 3 hours. Figure 2. IB shows that whilst reducing the incubation time for transfection to 3 hours did not significantly alter current density in bright cells (P > 0.05), a significant reduction in current density was seen in faint cells (P < 0.05).

2.3.4c. A weaker promoter was used to drive expression

Recombinant protein production in mammalian cells can be enhanced by use of the appropriate transcriptional element. The strength of any given promoter/enhancer is often cell-type dependent. In CHO cells, the immediate-early promoter and enhancer of the human cytomegalovirus (CMV) has been shown to be more effective than the Rous sarcoma virus 3' LTR promoter (Liu et al., 1997). In general, the CMV promoter was used to drive expression of Kir channels throughout this study. Since this gave high current densities, expression driven by the RSV promoter was investigated to determine whether use of a weaker promoter could consistently give smaller currents. Figure 2.1C shows clearly that RSV driven

92 Figure 2.1.

Levels of channel expression. Graphs to show how current density changes following minor alterations to the method of transfection. A, the total amount of cDNA used was reduced from 5/rg to 2.5/xg. B, the incubation time for transfection was reduced from 4 to 3 hours. C, a weaker promoter (RSV) was used in place of CMV to drive expression. D, the cDNA was purified using a CsCl gradient instead of commercially available kits. E, the cells were patched at 48 rather than 24 hours post-transfection. The clear bars represent bright cells, whilst the shaded bars represent faint cells. Reductions in current density were tested for statistical significance using unpaired student t-tests (* P < 0.05, ** P < 0.01). A.

2.5/ig

4 h * 5/i g

* 2.5/ig

5/ig 4 h

-1500 -1000 -500 0 -1500 -1000 -500 0

Current Density (pA/pF) Current Density (pA/pF)

c. Z).

RSV CsCl

CMV Kit

RSV CsCl * CMV Kit

-1500 -1000 -500 0 -1500 -1000 -500 0

Current Density (pA/pF) Current Density (pA/pF)

48 h

24 h

48 h

24 h

-1500 -1000 -500 0

Current Density (pA/pF)

93 expression did result in significantly reduced current densities in both bright (P < 0.05) and faint ( P < 0.01) cells.

2.3.4d. Different methods for the purification of cDNA were utilized

In this study, plasmid DNA was initially purified using commercially available kits as described in section 2.1.2. Such kits have become the modem standard for large-scale plasmid purification, yielding DNA with a purity equal to, or even exceeding that of DNA purified using banding on a CsCl gradient. Since the purity of the DNA can affect the efficiency of transfection (Ehlert et al., 1993), it was investigated whether current density could be reduced using CsCl purified DNA in transfections. Figure 2. ID shows that using CsCl purified DNA significantly reduced current density in bright cells (P < 0.05) but not in faint cells. In fact, the mean current density in faint cells, which had been transfected with CsCl purified cDNA was larger than that in faint cells that had been transfected with kit purified cDNA.

2.3.4e. Cells were patched at different intervals post transfection

Cells were typically healthy for up to 72 hours post-transfection. Thus, current density was compared at 24 and 48 hours post-transfection. Figure 2. IE shows that the average current density was significantly lower in bright (P < 0.05) and faint (P <0.01) cells 48 hours post­ transfection.

Although some of the changes to the method of transfection as described above clearly reduced current density, current levels remained high under most conditions. However, under certain conditions currents were small enough to avoid problems with series resistance. Since bright cells typically displayed very high current densities, these cells were avoided. By reducing to 1/xg the amount of cDNA used per transfection, in addition to reducing the incubation time to 3 hours and patching faint cells 48-60 hours after transfection, expression levels were low enough to resolve single channels.

2.3.5. Stable transfection of MEL cells using electroporation

The stable transfection of MEL cells was performed by electroporation, which involves the exposure of a cell suspension to a brief electrical impulse resulting in the formation of transient pores in the cell membrane and DNA uptake (Doyle et al, 1998a).

94 1x10 cells in logarithmic phase growth were required per transfection. Untransfected MEL cells were grown to confluency in 30ml of G418-free medium in a 175cm3 tissue culture flask. To determine the cell density a cell count was performed using a haemocytometer (see section 2.2.8). The appropriate volume of cell suspension containing lxlO7 cells was transferred into a sterile 15ml centrifuge tube and spun at lOOOg for 2 minutes at room temperature. The supernatant was poured off under sterile conditions and the pellet washed twice by resuspending in an equal volume of PBS and centrifuging as before.

Following the second wash, the pellet was resuspended in 0.9ml of cold electroshock buffer containing (mM): NaCl, 140; HEPES, 25; NaLLPCU, 0.75, which was titrated to pH 7.5 with 1M NaOH, filtered through a 0.22/xm bottle top filter and stored at 4°C. The cell suspension was then transferred to a sterile, pre-chilled 0.4cm path-length electroporation cuvette to

which 50pl of linearized expression construct (lmg.ml"1) was added. The cuvette was capped and inserted into the holder of the electroporator. The conformation of the plasmid DNA affects the efficiency of the transfection, use of linearized DNA improving the yield of stable transfectants (Potter et a l , 1984; Toneguzzo et a l, 1986; Chu et a l, 1987). The electroporator (Bio-Rad Gene Pulser®) was set at 0.25kV and 960pF for the voltage and capacitance extender, respectively. Following electroporation, the cuvette was removed from the apparatus at which stage the cell suspension appeared frothy.

150 and 300pl aliquots of the transfected cell suspension were pipetted into 50ml centrifuge tubes containing 30ml of G418-free medium and plated out 1ml per well into separate 24 well

plates. The cells were then incubated at 37°C for 24-48 hours, during which period each well was overlaid with 1ml G418 free medium supplemented with G418 at 2.0mg.ml'1, giving a final concentration of l.Omg.ml'1, to select for transfectants.

In general, colonies of transfectants began to appear within 7-14 days of transfection. Using a P20 Gilson pipette the colonies were picked off and plated out into 6 well plates containing 2ml standard growth medium, containing O^mg.ml'1 G418, per well. Following subculture and induction using 2% v/v DMSO, the individual colonies were screened for ionic currents. Those colonies that expressed high levels of current were then sub-cultured into large tissue culture flasks (175cm3) and grown to confluency for the subsequent freezing process and storage in liquid nitrogen (see section 2.2.7).

95 2.4. Electrophysiology

Ionic currents were measured using the patch clamp technique. Using this electrophysiological method it is possible to record both macroscopic and single-channel currents flowing across biological membranes through ion channels. Neher and Sakmann (1976) first used the patch clamp technique to record single channel currents from the membrane of denervated frog muscle fibres. The technique was later refined (Hamill et al., 1981) and today is probably the most widely used method for the recording of macroscopic and single channel currents from small cells.

2.4.1. Principles of patch-clamp

The recording of ionic currents using the patch clamp technique is achieved by ‘clamping’ the membrane at a pre-selected value known as the ‘holding potential’ using a feedback amplifier. Membrane currents are then activated by sudden changes in the command potential, or in the composition of the solution bathing the membrane.

The total membrane current (7m) comprises both ionic (7i) and capacitive components (7C) as described by:

(Equation 2.1)

Since

(Equation 2.2) where Q is the store charge and is related to voltage (V) and membrane capacitance (C) according to the relation:

Q = CxV, (Equation 2.3)

(Equation 2.4)

96 Thus

(Equation 2.5)

To eliminate capacitive currents, the voltage change must be instantaneous. During patch clamp, the membrane potential is changed from the holding potential to the desired voltage as rapidly as possible. Thus, any capacity current is restricted to the edges of the square pulse. Once the voltage change is over, any remaining current may be attributed to current flowing through ion channels. The speed at which the membrane potential is changed is dependent on two parameters, the membrane capacitance (Cm) of the cell and the series resistance (. Rs), which will be considered later in more detail (see sections 2.4.13 and 2.4.14).

2.4.2. Gigaohm Seal

In order to make patch clamp recordings, it is first necessary to isolate a patch of membrane electrically from the external solution. This involves pressing a fire-polished pipette against the surface of the cell and applying gentle suction. This leads to the formation of a 'gigaohm' (G£2) seal, with a resistance in excess of 109£2, from which it is possible to proceed onto one of several different recording configurations (sections 2.4.12; figure 2.5).

A high resistance seal is important for the following reasons:

To reduce leak current - the lower the seal resistance, the larger the fraction of current through the patch of the membrane that will leak out through the seal and will thus not be measured.

To reduce thermal noise - in any electrical recording, the thermal noise in a current ‘source’, termed Johnson noise , is determined by the source resistance according to:

(Equation 2.6)

where an is the root mean square (r.m.s.) of the current flowing through the resistance ( R), k is the Boltzmann constant, T the absolute temperature and A f the bandwidth (Gibb, 1995). Thus, the current noise in any recording decreases with increasing source resistance. For the measurement of single channel currents, of the order of a pA, at a bandwidth of 1 kHz, R

97 needs to be approximately 2G£2 or higher. In a patch clamp setup, R consists of the headstage feedback resistor, the seal resistance and the patch resistance. The largest noise source tends to dominate because the noise contributions from each source add together according to the relation:

rmsT = -yjrmsf + rms\ + ...rms2n (Equation 2.7) where rm ^is the total noise variance observed from n independent noise sources. Thus, both a high resistance seal and the use of a low-noise current-to-voltage converter amplifier (see section 2.4.9) are important for low noise levels and essential for good resolution of single channel events.

2.4.3. Setup and recording chamber

The patch-clamp setup used throughout this study consisted of an inverted microscope (Nikon) with fluorescent attachment for viewing eGFP-expressing cells. This was placed on the base plate of an air-suspension vibration isolation table and surrounded by a Faraday cage. Whole-cell and single channel currents were amplified and filtered using either an Axopatch 200B (Axons Instruments) or EPC-7 (List Instruments) amplifier, and digitised using a TL-1 Labmaster interface (Axon Instruments). Whole-cell and single-channel currents were viewed on a cathode-ray oscilloscope (CRO) and stored directly to a computer hard disk or to digital analogue tape (D.A.T), respectively. The amplifier headstage was mounted to a hydraulic manipulator (Narashige Instruments) by way of a metal plate, whilst the tip of the perfusion system was attached to a coarse mechanical manipulator. The recording chamber (figure 2.2) consisted of a 35mm petri dish (Nunclon™) inserted into a sheet of perspex containing a separate hole for the immersion of a silver/silver chloride reference electrode (Ag-AgCl pellet) in bath solution. The chamber containing the Ag-AgCl pellet was connected to the chamber containing the cells by way of a ‘salt bridge’ (U-shaped capillary tube) filled with the extracellular solution.

2.4.4. Pipette and reference electrodes

The pipette and reference (bath) electrodes facilitate the detection of ionic currents in solution as electrical currents by the headstage of the patch clamp amplifier. Silver/silver chloride (Ag-AgCl) electrodes, which exchange electrons for chloride ions in solution, are typically used.

98 35mm Petri dish containing to circuit ground „ Salt bridge

O

perspex mount

silver/silver chloride pellet

Figure 2.2.

Diagram to illustrate the recording chamber. A 35mm Petri dish containing the cells was inserted into a sheet of perspex and connected via a ‘salt bridge’, filled with extracellular solution, to a separate chamber into which a silver/silver chloride pellet was inserted. This pellet was connected in turn to the circuit ground on the amplifier headstage. The pipette electrode consisted of a silver wire coated in silver chloride, which was inserted into the pipette holder being careful to ensure it was in contact with the pin that plugs into the headstage. This wire was often scratched when exchanging patch pipettes resulting in ‘drift’ of the pipette potential. To prevent significant drift, the silver wire was regularly ‘rechlorided’ by dipping into molten AgCl.

The reference (bath) electrode consisted of a pellet made from metallic silver (30%) and powdered AgCl (-70%), compressed around a length of silver wire and connected to the headstage. The circuit from bath solution to pipette solution, to headstage, and back to bath solution was completed by way of a ‘salt bridge’.

2.4.5. Solutions

All solutions used in the experiments described within this thesis were prepared by adding milliQ water to chemicals in their solid form, with the exception of CaCl2 and MgCl2, which were added in the appropriate volume from a 1M stock solution. The final pH of all solutions was adjusted to 7.2 using 1M HC1, KOH or RbOH as appropriate. Before use, all solutions were filtered using a 0.2/xm acrodisc filter (Pall Gelman Sciences).

For whole cell recording, the extracellular solution contained (mM): KC1, 70; A-methyl-D-

glucamine (NMDG), 70; HEPES, 10; CaCl2, 2; MgCl 2, 2; pH adjusted to 7.2 with 1M HC1. In experiments examining the effects of extracellular K+ concentration [K+]0, NMDG replaced K+ when [K+]c was reduced or was omitted when the K+ concentration was increased to 140mM. For ion substitution experiments, the extracellular solution was prepared as described above but with 70mM RbCl replacing the KC1. To examine the blocking effects of 2+ _j_ Ba or Cs , BaCl2 and CsCl were added to the extracellular solution at concentrations of 5, 10, 30, 50, 100 and 300/xM for Ba2+, and 50 and 500/iM for Cs+.

For whole cell recording, the intracellular solution contained (mM): KC1, 105; HEPES, 10; EDTA, 10; pH adjusted to 7.2 with 1M KOH and KC1 added to bring [K+] to 140mM.

For single channel recording, the intracellular solution contained (mM): KC1, 105; HEPES, 10; EDTA (free not disodium), 10; CaCl2, 2; pH adjusted to 7.2 with 1M KOH and KC1 added to bring [K+] to 140mM. The extracellular solution contained (mM): KC1, 200; HEPES, 10; CaCl2, 2; MgCl2, 2; pH with 1M HC1 to 7.2. To examine the blocking effects of Ba2+, BaCl2 was added to the extracellular solution at concentrations of 100, 200 and 300jaM.

100 2.4.6. Perfusion system

Solutions were delivered from a multi-line gravity-fed perfusion system, consisting of five 5ml syringes attached via polythene tubing to a common perfusion tip positioned in the vicinity of the cell being patched. Each syringe contained different solutions, the outflow from each syringe being controlled by a one-way tap. Construction of this perfusion system is described in more detail elsewhere (McKillen, 1994; Abrams, 2000).

2.4.7. Manufacture of pipettes

For whole cell recording, the use of low resistance pipettes is more important than low noise since the use of pipettes with high resistances can result in large voltage errors. Patch pipettes were manufactured using a vertical microelectrode puller (pp-83, Narishige Instruments). Thin walled (1.5mm x 1.17mm) borosilicate glass capillaries with inner filaments (Clark Electromedical Instruments) were pulled in two stages. To promote gigaohm sealing, pipettes were then fire-polished using a Narishige fire polisher to give resistances between 2-5 megohm (MQ). Heat polishing improves the success of sealing by smoothing the edges of the tip and removing dust particles.

Low background noise is crucial for the resolution of single channels (see section 2.4.2).

Thus, pipettes for single channel recording were manufactured from thick walled (1.5mm x 0.86mm) borosilicate glass capillaries with inner filaments (Clark Electromedical Instruments), pulled in two stages using a vertical microelectrode puller as described above. To reduce stray capacitance and the noise levels of recordings, the tapered region of the patch pipettes was coated with Sylgard® (Gibb, 1995; Ogden & Stanfield, 1994; Standen & Stanfield, 1992).

Sylgard® (Dow-Coming 184 transparent encapsulating resin) is a hydrophobic substance consisting of two separate components, a resin and a curing agent, which should be mixed together in 10:1 proportions before use. It is used to coat patch pipettes for single channel recording. Coating patch pipettes with Sylgard® reduces stray capacitance by preventing creep of the bath solution up the tip of the pipette and by increasing the size of insulation between the bath and pipette solutions.

101 In general, a volume of 5ml was mixed together and left at room temperature for 2-3 hours. The mixture, which was usually thicker and free of air bubbles by this stage, was then aliquoted into 1ml eppendorf tubes and stored at -20°C. Prior to fire polishing, Sylgard® was applied to the tapered end of the patch pipettes using a fine syringe needle, and cured using a heated coil. Patch pipettes were then fire-polished to give resistances of 8-10M£2.

2.4.8. Pipette diameter and resistance

Occasionally, the pipette size after pulling, or after fire polishing, was checked using a simple procedure to measure the ‘bubble number’ (see for example Standen & Stanfield, 1992). This involved connecting the back of an electrode to a polythene tube connected to a 10ml syringe with the plunger withdrawn to the 10ml mark. The tip of the pipette was dipped into methanol, contained in a vial with a dark background, and the plunger gently depressed until a stream of bubbles originating from the tip was observed. The volume at which this occurred was the ‘bubble number’. For whole cell electrodes the ideal bubble number is 6-7ml before and 5ml after fire polishing, whilst for single channel electrodes a bubble number between 4- 6ml before and 2-4ml after fire polishing was desirable.

The pipette resistance was often calculated once the pipette had been inserted into the bath solution prior to seal formation. At this point, pipettes with inappropriate resistances were discarded. To calculate pipette resistance, the gain of the patch clamp amplifier was set on lOmV/pA and a lmV command voltage of 10ms duration was applied to the amplifier to generate a current through the pipette. Pipette resistances were subsequently calculated from the size of the current trace in response to the lmV command voltage using Ohm’s Law:

V = IR (Equation 2.8)

where V is voltage, I is current and R is resistance.

2.4.9. Patch clamp amplifier

Although initial channel recordings were obtained using an EPC-7 (List Electronics) patch clamp amplifier, the majority of channel currents included in this thesis were recorded using an Axopatch 200B patch clamp amplifier (Axon Instruments). The headstage of the Axopatch 200B implements the conversion of current to voltage by way of either a ‘resistive-’ or ‘capacitive-feedback’ circuit depending on which mode (whole cell or patch, respectively) has been selected.

102 For whole-cell recording, the Axopatch 200B uses a high resistance-feedback circuit (figure 2.3) consisting of an operational amplifier (Al) with high input impedance, a feedback resistor (Rf, typically 50 or 500M£2) and a differential amplifier (A2). The negative feedback circuit from the output of the amplifier is used to force the voltage in the patch pipette (Vp), detected at the inverting (-) input, to follow the command voltage (Vcmd), applied to the non­ inverting input (+). Owing to the high input impedance of the operational amplifier, the current source being measured (ip) is forced to flow through the feedback resistor (Rf). The amplifier output (Vo) is therefore proportional to the input current as follows:

(y -V A O P (Equation 2.9) 1p = Rf since Vp = Vcmd then,

f yo ~ y cmd d ^ (Equation 2.10) 1p = R f when rearranged this gives

V o ~ ~ ip R f + V cm d (Equation 2.1l)

The current flow through the patch pipette is subsequently given by the output of the differential amplifier (A2), which measures the difference between the output (V0) of the operational amplifier (Al) and the command voltage (Vcmd). The output from the differential amplifier is then fed into a high frequency boost circuit for correction of stray capacitance.

For single channel recording the Axopatch 200B utilises a capacitive feedback circuit (figure 2.4). Also known as the integrating headstage, this uses a small capacitor (Cf) in the feedback path of the current-voltage converter. In response to a constant input current, the headstage generates a ramp, the slope of which is proportional to the current (Finkel, 1991). The slope of the ramp represents the rate of charging of the voltage across the capacitor and is measured by a differentiator to give the current in the patch pipette (ip).

103 Figure 2.3.

Diagram of the resistive headstage. The current-to-voltage converter consists of an operational amplifier, Al, and its feedback resistor, Rf. (-) and (+) represent the inverting and non-inverting inputs at which the pipette potential ( Vp) and command potential ( Vcmd) are detected, respectively. Cf is the stray capacitance across Rf. The differential amplifier, A2, measures the difference between the output (V0) of A l and Vcmd to give the current being measured in the patch pipette (ip).

Cf

A2 to ‘boost’ circuitry

J Figure 2.4.

Diagram o f the capacitive headstage. The current-to-voltage converter utilises a small capacitor, C f, in the feedback bath. The output of amplifier A l, Vo is a ramp whose slope is proportional to the current in the pipette ( ip). A second capacitor, Cd, resistor, Rd and the amplifier A2 form the differentiator circuit, the output of which is proportional to the pipette current.

Cf

A l A2

to offset and scaling The output of the integrating headstage, (Vo) is related to the input current according to the relation:

(Equation 2.12) so that the V0 is the time integral of ip as described by:

1 V0 (r) = jip d t (Equation 2.13) C /f n0

Correction of the frequency response is achieved by way of a differentiator circuit formed by a capacitor (Cd), resistor (Rd) and amplifier (A2). The final output voltage of the circuit is given by:

(Equation 2.14)

and is proportional to the current measured in the pipette.

2.4.10. Procedure for obtaining a gigaohm seal

The first step when patch clamping is to obtain a high resistance or ‘gigaohm’ seal (section 2.4.2). In their 1981 paper, Hamill et al. (1981) described several procedures, which increased the probability of obtaining a gigaohm seal. These included precautions to keep the pipette and cell surfaces clean and the application of gentle suction to the inside of the pipette once it was in contact with the cell. By filtering all solutions as described in section 2.4.5 and applying gentle positive pressure to the pipette before immersing its tip in the bath, it is possible to keep the tip of the pipette free of debris. The resuspension of the cultured cells in serum free media or filtered extracellular solution also aids the formation of a giga-seal. Finally, once the pipette is touching the surface of the cell, gentle suction may be applied to the inside of the pipette using a syringe attached to an outlet on the pipette holder by way of a length of polythene tubing. In general, if the above conditions were adhered to, gigaohm seals could be obtained as described below.

The patch pipette was back-filled with intracellular solution and inserted into the pipette holder so that the tip of the chlorided silver wire was in contact with the solution. The patch pipette and its holder were then inserted into the input connector of the patch clamp

106 headstage. To prevent damage to the headstage from static electricity, a grounded metal object, such as the Faraday cage, was touched prior to handling the headstage. Following application of slight positive pressure, the pipette was lowered into the bath solution using the coarse controls of the manipulator. At this point, any offset potentials (Vo) between the pipette and reference electrodes were cancelled by adjusting the 'pipette offset' control. The amplifier adds this variable offset to the command voltage (Vcmd) so that the current flow at Vcmd is zero. However, the situation is complicated by the existence of junction potentials (see section 2.4.15) during the reference measurement, which disappear on going whole cell. With the gain of the patch clamp amplifier set on- lOmV/pA, a lmV command voltage of 10ms duration was applied to the amplifier to generate a current through the pipette. At this point, the pipette resistance was calculated (see section 2.4.8).

The formation of a gigaohm seal represents a sudden increase in seal resistance and this process may be closely monitored, by observing the size of the pulse on the oscilloscope screen. According to Ohm’s Law, an increase in resistance leads to a reduction in current. Gigaohm seal formation is therefore reflected by a rapid decrease in the size of the visible current pulse, which eventually disappears. By using a combination of the coarse and fine movements of the manipulator, the pipette was brought close to the cell to be patched. Once the pipette and cell membrane were touching, gentle suction was applied to the pipette. A negative voltage (-17mV) was applied to the interior of the pipette to encourage the formation of a gigaohm seal.

2.4.11. Capacity transients

Once a gigaohm seal had been achieved, the command voltage was increased to lOmV and the capacity transients associated with pipette capacitance (Cp) minimized using the pipette capacitance compensation controls. This is necessary because, for resistive headstages, even small voltage steps can lead to saturation of the voltage output during charging of the fast capacity transient. For larger voltage steps, saturation can persist for a millisecond, or more, during which time information on, for example, fast-activating currents may be lost. Cancellation of pipette capacity transients is achieved by injection of a current, equal in size to the charging current, by way of a capacitor internally connected to the input.

2.4.12. Recording configurations

Cell-attached patches were obtained by following the protocol for seal formation as described above. Although the cell remains intact, the cell-attached configuration may be used to record

107 single channel activity. One problem with the configuration, however, is that the cytoplasmic side of the membrane cannot be easily manipulated. Another problem associated with the cell-attached configuration is that the membrane potential is unknown. For these reasons, excised patches may be the preferred method for obtaining single channel recordings. Furthermore, by positioning the patch pipette just below the surface of solution interface, it is possible to reduce capacitance with the excised patch configurations. However, it has been documented that channel properties can change with patch excision (Trautmann & Siegelbaum, 1983; Covarrubias & Steinbach, 1990). One advantage with the cell-attached configuration is, that the biological environment of the channel remains unperturbed. As illustrated in figure 2.5, the cell-attached patch configuration is the precursor to the other patch-clamp configurations described below.

Inside-out patches were formed by pulling the patch pipette away from the cell, whilst in the cell-attached mode (figure 2.5). In this configuration, the membrane patch is detached from the cell with the cytoplasmic membrane surface facing the bath solution. It is, therefore, possible to change the cytosolic environment. The major problem with this configuration, and of particular relevance to inward rectifier potassium channels, is the loss of cytosolic factors involved in normal channel behaviour. Furthermore, on excising a patch the membrane often 'seals over' to form a vesicle. ‘Vesicle formation’, which was evident as a reduction in the amplitude of the single channel openings and a 'rounding' in their appearance, could occasionally be rectified on brief exposure to air by gently moving the patch through the air-solution interface, although, more often than not, this resulted in loss of the patch. In the inside-out patch configuration, positive command voltages correspond to a hyperpolarization of the patch membrane. Pipette capacitance and background noise was reduced following movement of the patch pipette near the surface of the bath solution.

The whole-cell configuration , as its name implies, permits the recording of ionic currents associated with the entire cell membrane. It was achieved by applying suction through the pipette following gigaohm seal formation. This resulted in the rupture of the membrane and the subsequent exposure of the cell contents to the solution inside the electrode. Rupture of the membrane was indicated by the sudden appearance of large capacity transients at the leading and trailing edges of the voltage pulse, which were minimized by adjusting the ‘whole cell capacitance’ and ‘series resistance’ controls simultaneously as described in section 2.4.13. For large currents, series resistance compensation was also used as described in section 2.4.14.

108 Patch Pipette Cell Attached (Giga Ohm Seal)

Cell

Light Suction

More Pull Suction

Inside-Out Patch

Whole-Cell Configuration

Pull

Outside-Out Patch

Figure 2.5. Diagram illustrating the different configurations of patch-clamping. Once a gigaohm seal has been formed it is possible to use one of four different recording configurations. The cell-attached mode permits the recording of single-channels. However, disadvantages with this mode make the inside-out configuration more favourable for recording single-channels. The whole-cell configuration permits the recording of currents associated with the entire cell membrane. The outside-out patch, achieved by pulling the pipette away from the cell in whole-cell mode, may be advantageous if whole-cell currents are very large.

109 Outside-out patches were formed by slowly pulling the patch pipette away from the cell whilst in whole-cell configuration (figure 2.5) so that the outside of the membrane remains exposed to the bath solution. As with inside-out patches, the pipette was raised to just below the solution interface to reduce capacitance. Pipette capacitance was adjusted accordingly.

2.4.13. Whole cell capacitance

As described in section 2.4.12, a sudden increase in the size of the capacity transients was visualized on going whole-cell. Since any stray capacitance associated with the pipette was cancelled prior to rupturing the membrane (section 2.4.11), the remaining capacity transients could be attributed to the cell. As we saw in section 2.4.1, the total membrane current (7m) has two components, one ionic (70 and the other capacitive (7C). During patch clamp, the membrane potential is changed from the holding potential to the desired voltage as rapidly as possible. Thus, any capacity current is restricted to the edges of the square pulse. The speed at which the membrane potential is changed is dependent on two parameters, the membrane capacitance (Cm) of the cell and the series resistance (7?s), a resistance that exists between the cell interior and the pipette. For membrane resistances that are large in comparison with Rs, the time constant for charging the cell is

z = CmRs (Equation 2.15)

The currents associated with charging the cell capacitance can be compensated using the ‘whole cell capacitance’ and ‘series resistance’ controls on the amplifier, from which it is possible to estimate values for Cm and Rs, respectively. Occasionally, the duration of these capacity transients appeared quite long and on attempting to compensate, series resistance values were extremely high. In general, this was indicative of bad access to the cell. However, on applying additional suction access could usually be improved, indicated by an increase in the amplitude and a decrease in the duration of the capacity transients and a subsequently lower series resistance.

2.4.14. Series resistance compensation

One possible source of error during patch clamp arises as a result of the series resistance ( Rs), which affects the accuracy with which the membrane potential can be controlled. When a current, 7, flows across the membrane a voltage error, of size V = IRS, will arise leading to a discrepancy between the actual membrane potential and the potential measured by the amplifier. Such discrepancies will be larger for larger currents or series resistances. In

110 studies of channel selectivity, such voltage errors have little consequence since we are interested in measuring the reversal potential of the current, at which there is no net flow of current and therefore, no voltage error. However, series resistance errors do have serious implications for the measurement of channel properties that are voltage dependent, such as block by Cs+ or Ba2+ and channel activation.

In CHO cells, inward rectifier currents were often in excess of 5nA. For a current of 5nA and with a typical series resistance of 5MQ, a voltage error of 25mV will exist. Evidently, the currents activated by the command voltage will not reflect the ‘true’ currents that would be activated by the actual membrane potential. However, voltage errors due to current flow over Rs can be compensated using the amplifier series resistance cancellation circuitry, which works by a adding a percentage of the voltage error to the command voltage by way of a feedback circuit. Although in theory 100% compensation is possible, in practice, only 50- 80% was achieved due to the occurrence of oscillations at higher values. Thus, for a 5nA current and 5M£2 series resistance, a voltage error of 7.5mV would still remain following 70% compensation.

2.4.15. Junction potentials

During patch clamp recording a liquid junction potential will arise at the tip of the patch pipette when solutions of different ionic composition are used to fill the pipette and bath. Liquid junction potentials are the result of an imbalance in charge, which develops as a consequence of the different rates at which ions diffuse across the concentration gradient at the interface between two solutions. This occurs because each individual ion has a different mobility, the more mobile an ion the faster it diffuses.

Using the pipette offset dial on the patch clamp amplifier it is possible to compensate for the junction potential (V l ) that develops at the interface between the internal and external solutions, when the two solutions are of different ionic composition. The amplifier generates an equal and opposite potential (Vl1) when the pipette offset is used to zero the trace, which is added to Vcmd (Figure 2.6A).

Ill Figure 2.6.

Liquid junction potentials. A, Prior to sealing the liquid junction potential (V l) that arises at

the tip of the pipette is balanced by an equal and opposite 'offset potential' (V l1) using the pipette compensation on the amplifier. B, On sealing, Vl disappears, being replaced by Vm because the membrane forms a barrier between the pipette and bath solutions. The offset

potential, V l1, still exists. C, Following membrane-rupture, a new junction potential forms but disappears with dialysis of the cell contents by the pipette solution. Vl1 remains. The curly arrow represents the direction in which the circuit is traversed. V L1

Pipette

Cell Ag/AgCI Pellet

V p v L1

Pipette

Cfell Ag/AgCI fl Pellet

Vp V L1

P ipette

Cell

Ag/AgCI Q Pellet

112 However on sealing, the junction potential between the two solutions disappears because the membrane patch forms a barrier between the pipette and the bath solutions. However, the junction potential correction applied by way of the amplifier still exists (figure 2.6B). Thus, Vm becomes

Vm ={Em- V m i)+VL' (Equation 2.16)

On going whole-cell, a new junction potential arises between the intracellular solution and the internal contents of the cell, but disappears on effective dialysis of the cell contents by the pipette solution (figure 2.6C). Vmis then

Vm = V cm d - V L (Equation 2.17)

Liquid junction potentials were measured experimentally using a 3M KC1 electrode as the reference electrode, which was constructed from thin walled borosilicate glass and filled with 3M KC1 containing 0.5% agar. Patch pipettes were pulled as previously described and back­ filled with intracellular solution. With the patch clamp amplifier in current clamp mode, both the patch pipette and reference electrodes were dipped into a petri dish containing intracellular solution. With the amplifier in ‘track’ mode, the voltage reading, or tip potential, was adjusted to zero using the pipette offset. The bath solution was then exchanged for extracellular solution. The resulting junction potential between the pipette and the bath was measured as the voltage difference between the two solutions and was noted from the LED display once the reading had stabilized.

Alternatively, the magnitude of junction potentials can be determined by calculating them using the generalized Henderson Liquid Junction Potential Equation (Barry & Lynch, 1991). For A polyvalent ions, the junction potential (El) of the extracellular solution ( S ) with respect to the pipette solution (P), is given by:

N

VS-Vr= K SF ln« (Equation 2.18) F N

113 where

(Equation 2.19)

i= i where u, a and z represent the mobility, activity and valency of each ion species (z), respectively, and where R , T and F have their usual meanings. In this study, the molar concentration of the ion species was used in place of its activity in junction potential calculations.

Using the flow system, the bathing solution was exchanged several times throughout the course of many of the experiments described within this study. This would have led to additional junction potentials arising between the solution immediately in contact with the cell and the original bath solution. Since the solution in the recording dish was not continuously renewed, the ionic composition of the bath solution would have changed every time a cell was perfused with a different solution. The composition of the bath solution against which the pipette was nulled was therefore unknown, as was the junction potential correction, and for this reason, no attempt was made to correct for any errors arising from liquid junction potentials in this study.

2.4.16. Data acquisition

Whole-cell and single-channel data were initially filtered at 5 and 2kHz respectively using the amplifier’s built-in 4-pole lowpass Bessel filter. Single channel data were filtered further when sampling sections of single channel recordings from D.A.T. to the computer hard drive using TAPE. Analogue Bessel filters and digital Gaussian filters add together to give a final bandwidth, f c, according to the relationship:

(Equation 2.20)

Thus, to give a final bandwidth of 1kHz, single channel currents were digitally filtered at 1.15kHz.

114 Whole-cell and single-channel data were stored directly onto the computer hard disc or digital audio tape (D.A.T), respectively. However, this requires the output of the amplifier (Vo), which is an analogue signal, to be converted into a digital signal. This analogue-to-digital conversion (ADC) was achieved by way of an external interface, which was connected to both the amplifier and the computer (see section 2.4.3). The interface also converts the digital command voltage, generated by the computer, to an analogue signal for the amplifier (Digital- to-analogue conversion, DAC). Data were monitored on a cathode ray oscilloscope (CRO), which was connected to the amplifier and interface so that the display was synchronized with the triggering of the voltage command.

During digitisation the amplitude of the signal is measured at intervals determined by the sampling frequency, and stored as binary numbers. Data were sampled at 10kHz in accordance with the Nyquist Sampling Theorem, which states that the minimum sampling rate is twice the signal bandwidth (Ogden, 1994).

2.4.17. Data analysis

Electrophysiological data was acquired and analysed using AXGOX, a suite of programs written in Microsoft V.7.1.BASIC (Davies, 1993). STIMTOR was used to generate voltage commands, ramps and complex voltage sequences whilst data were manipulated and analysed in TRACAN. Selected sections of continuous, single channel recordings were transferred from D.A.T tapes to the computer hard disc by way of TAPE. Event detection was performed in TRACAN, then analysed and fitted in EVTANR. General data manipulation was undertaken using Microsoft Excel spreadsheet, statistical analysis was performed using

GraphPad Instat and graphs were created in Jandel SigmaPlot®.

For whole-cell data, preliminary analysis involved minimizing the capacity transients since a residual transient often remained following analogue capacitance compensation (section 2.4.13). This involved ‘back-subtracting’, digitally, the capacity transient at the end of the voltage pulse from the capacity transient at the beginning of the voltage pulse. In theory, this should result in the removal of any residual capacity transient. However, because the magnitude of the capacity transients was frequently asymmetrical, back-subtraction of the capacity transients was not always appropriate.

Although whole-cell current records with obvious leak components were not used for further analysis, in some experiments such as the measurement of Rb+/K+ permeability ratios in

115 channels exhibiting small currents, the use of records with leak components was unavoidable. Under such circumstances, leak subtraction was performed by subtracting a linear fraction obtained by fitting currents greater than 60mV and at £k- This procedure assumes that the K+ conductance is zero at these potentials. Further methods of analysis for whole-cell data are explained later in the relevant results chapters.

For single channel data, preliminary analysis involved producing an average leak current. Sections of continuous recording in the closed state were selected and averaged then subtracted from the file to give a zero current level for the current when in the closed state. More detailed analysis to determine the unitary current, mean open times (MOT) and probability of opening (Popen) were then carried out as described below.

Unitary current - where possible, the size of the unitary current for Kir channels was measured by constructing amplitude histograms from continuous recordings containing both open and closed events. Peaks corresponding to the channel closed and open levels were obtained, and fitted with gaussian distributions to give a mean channel opening. However, for single channels obtained in response to episodic stimulation, current levels were measured using cursors and a number of values averaged.

Popen - The probability of opening was determined by measuring the times (fj) spent at current levels corresponding toy = 0, 1, 2...N open channels. The overall Popen was given by:

7=1 (Equation 2.21)

where T is the duration of the recording and N, which is the maximum number of channels in the patch, was set as the maximum number of simultaneous openings observed during one recording.

MOT - Mean open times were analysed using the half-amplitude threshold crossing method described by Colquhoun and Sigworth (1983) in which an event is detected when it crosses the 50% threshold between open and closed levels of the channel. Events shorter than 0.18ms were not detected since:

deadtime = (Equation 2.22) fc

116 and single channel currents filtered at an overall bandwidth of 1kHz. Events were written to an idealized record from which open- and closed-time histograms were constructed. These distributions were plotted with log time on the abscissa and the square root of occurrence on the ordinate labelled events/bin. Plotting the data as described allows exponential components with time constants varying by many orders of magnitude to be displayed on the same graph. The individual exponential components appear as peaks, which correspond to the time constant of each component, the height of the peak corresponding to the number of events for each component. The data were fit to an exponential distribution of the form:

n ( a ■ ^ — exp(-,/^ (Equation 2.23)

j = i where and Tj are the relative area and the time constant of component j. The number of components, n, was set to the number of open or closed states and the data fit using the method of maximum likelihood (Colquhoun & Sigworth, 1983).

Mean open times were corrected for missed events (i.e short closures less than the minimum resolution, which were too brief to be detected) by multiplying the mean open time by the proportion of closed events detected, which was found by integrating the fitted closed time distribution between the minimum resolution and infinity (see Davies et al., 1992).

2.5. Detection methods for protein expression

Numerous methods for the detection of protein expression now exist. However, each method has its limitations. For example, whilst Western blotting is useful for determining whether expression of a certain type of protein occurs in a particular cell type or tissue, it cannot be used to determine the cellular origin of that protein (Coyler, 1999). Furthermore, immunological techniques such as Western blotting require the use of an antibody specific to the protein being investigated. In this study, two different techniques have been employed to determine the expression of wild-type and mutant channels as described below.

2.5.1. Western blotting

First described by Towbin et al. (1979), Western blotting couples the separation of proteins by polyacrylamide gel electrophoresis, protein transfer to nitrocellulose sheets and subsequent binding assays to give a method, which ultimately allows the detection of proteins by autoradiography. The electrophoretic step separates the proteins according to their apparent

117 molecular weight whilst an exact copy of the electrophoretic gel is transferred onto the nitrocellulose membrane. The proteins are then probed with a primary antibody followed by an additional ‘detecting protein’, such as a secondary antibody, which is conjugated to a marker such as the marker enzyme horseradish peroxidase (HRP), which can be detected using an enhanced Chemi-luminescence kit. Hames (1990), Appelbaum and Shatzman (1999) and Colyer (1999) have provided excellent descriptions of Western blotting.

2.5.1a. Sample preparation

CHO cells were transfected using FuGENE™ transfection reagent as described in section 2.3.3. Forty-eight hours after transfection, the cells were collected following removal using TE as described in section 2.2.5c. At this point, cells were incubated for 30 minutes in either lOmM dithiothreitol (DTT) or PBS as appropriate, following which they were centrifuged and washed twice as described previously. 400/xl homogenizing buffer (NaCl, 150mM; EDTA, ImM; EGTA, ImM; NEM, 20mM; Tris pH 7.4, 10; 1% Triton-X-100 [Iso- octylphenoxypolyethoxyethanol]; plus the addition of 5fi\ protease inhibitor cocktail) was added to each sample of cells and left on ice for 5 minutes. The samples were then triturated using a 1ml Gilson tip and aliquoted in 12.5/xl samples into 1ml Eppendorf tubes and stored at -20°C until required for gel electrophoresis.

2.5.1b. Gel electrophoresis

The separation of proteins according to their molecular weight was performed using Sodium Dodecyl Sulphate-Polyacrylamide Gel Electrophoresis (SDS-PAGE), using the Mini-Protean II electrophoresis system (Bio-Rad). To ensure good gel adhesion, the glass electrophoresis plates and spacers were washed prior to assembly with 70% ethanol, and allowed to air dry. The glass-plate ‘sandwich’, consisting of one large plate, one small plate and two spacers of uniform thickness, was then assembled using the gel cast apparatus. The resolving gel (acrylamide, 10%; Glycine, 213mM; Tris base, 27.5mM; SDS, 3.9mM; ammonium persulphate solution, 20fi\ of 10%; TEMED [N, N, N', N'-tetramethylethylenediamine], 20/xl) was then prepared and poured between the plates, taking care to avoid air bubble formation, to a previously marked level 5cm high. Acrylamide is highly toxic, therefore disposable plastic gloves were worn at all times when handling gel mixtures in either liquid or solid form. The ammonium persulphate solution is unstable, it was therefore made fresh just before use by dissolving O.lg ammonium persulphate in 1ml distilled water. The ammonium persulphate solution and TEMED were added just before pouring the gel mixture,

118 polymerisation of acrylamide being initiated by the addition of ammonium persulphate and accelerated by addition of TEMED (Hames, 1990). The resolving gel was immediately overlaid with 0.1% SDS and set aside for 20-30 minutes to allow for acrylamide polymerisation. Once polymerisation had occurred, an interface between the SDS and gel layers was clearly visible. At this point, the apparatus was tilted to discard the SDS layer and the gel rinsed with distilled water, any excess water being blotted from the gel surface using filter paper.

The stacking gel (acrylamide, 4%) was then prepared and layered on top of the resolving gel. A ten-well comb was immediately inserted into the stacking gel mixture, taking care not to trap any air bubbles beneath the comb. The stacking gel was then topped up with the remaining gel mixture and left to polymerise for 30-40 minutes.

Meanwhile, the protein samples were defrosted on ice, then mixed 1:1 with 2x laemmli denaturing sample buffer (Tris pH 6.8, 125mM; glycerol, 20% v/v; SDS, 4% w/v; bromophenol blue, 0.004% w/v) without DTT and heated at 100°C for 5 minutes. The protein samples and pre-stained markers were then centrifuged at 13,000rpm for approximately 4 minutes.

Once the stacking gel was set, the comb was carefully removed and the wells rinsed with distilled water. The glass plates were then removed from the setting frame, clipped into the reservoir frame and lowered into the electrophoresis tank. The inner and outer reservoirs of the tank were filled with running buffer (mM: glycine, 213; Tris base, 27.5; SDS, 3.9) so that the inner lower plate and the bottom of the outer plates were covered with buffer. Finally, the protein standards, pre-stained markers and samples were loaded into the appropriate wells and the gel run for approximately 1.5 hours at 200 V.

2.5.1c. Transfer of proteins

The transfer of proteins from the SDS-PAGE gel was performed by ‘semi-dry electrotransfer’ (for review see Colyer, 1999). On completion of electrophoresis, the SDS-PAGE gel was carefully removed from one of the glass electrophoresis plates. Using a scalpel blade, the stacking gel was removed and the upper left hand comer of the resolving gel marked to aid orientation. The gel was gently eased off the remaining glass plate into a plastic storage box containing transfer buffer (Tris base, 48mM; glycine, 39mM; SDS, 1.3mM; methanol, 20%) and left to soak for approximately 10-20 minutes.

119 Meanwhile, four pieces of blotting paper and one piece of nitrocellulose paper, cut to the same size as the gel, were also soaked in transfer buffer. Next, one piece of blotting paper was removed from the transfer buffer, drained and placed on the anode plate of a semi-dry electrophoretic transfer cell (Bio-Rad), being careful to remove any air bubbles using a glass rod. Using the same procedure, a second piece of blotting paper followed by the nitrocellulose membrane (Schleicher & Schuell) was superimposed upon the first piece of blotting paper. The SDS-PAGE gel was then carefully drained and positioned on top of the nitrocellulose membrane such that the first lane would transfer to the left-hand side of the membrane (i.e. marked comer to left). The stack was completed, by placing the remaining two pieces of blotting paper on top of the gel. Finally, the cathode plate was placed on top of the transfer stack and the proteins transferred for 30 minutes at 15 Volts.

The transfer of the pre-stained markers was checked after 30 minutes by carefully lifting the top layers of blotting paper and the gel away from the nitrocellulose membrane. Before discarding the gel, the nitrocellulose membrane was cut around the outline of the gel and the top left hand comer of the membrane marked.

2.5.1d. Western blotting procedure

Following overnight incubation of the nitrocellulose membrane in blocking buffer at 4°C, the membrane was washed three times for 5 minutes using Tris Buffered Saline (TBS; mM: Tris base, 50; NaCl, 150) containing 0.2% v/v Tween 20. The membrane was then incubated in primary antibody1 for at least one hour. The primary antibody was removed and retained for future use (primary antibody may be retained for up to 3 weeks after initial use) and the membrane washed a further three times for 10 minutes each wash with TBS containing 0.2% Tween 20. The membrane was then incubated in a peroxidase-conjugated secondary antibody for 1 hour followed by five washes with TBS for 10 minutes each wash. A plate shaker was used throughout the above procedures to ensure the even washing and incubation with antibody of the entire membrane.

2.5.1e. Protein detection

Protein bands were detected using the ECL Plus™ detection system (Amersham Pharmacia Biotech) according to the manufacturer instructions. The reagents were removed from the

1 Primary antibody was obtained from Dr R Norman, Dept. Medicine, University of Leicester.

120 fridge and allowed to equilibrate to room temperature. The solutions were mixed and poured onto the surface of a clean electrophoresis glass plate. The nitrocellulose membrane was removed from the buffer, drained of any excess solution and placed protein side down on the glass plate for 5 minutes. The nitrocellulose membrane was drained of any excess detection reagent and wrapped in cling film, being careful to remove any air bubbles. The wrapped blot was then placed, protein side up, in an x-ray film cassette. The cassette was then taken to a dark room where, under red safe lights, a sheet of autoradiography film (Hyperfilm™, Ammersham) was placed on top of the membrane, the cassette closed and the film exposed for approximately 12 seconds. The film was then developed using an Amersham Hyperdeveloper. The exposure time for further pieces of film was altered according to the appearance of the preceding exposures.

2.5.2. In vitro transcription translation

Over the last decade, the synthesis of proteins in vitro has become a popular tool in research. Today, in vitro systems utilising either DNA or RNA as initial templates are readily available from commercial suppliers (Clemens, 1999). These ‘cell-free synthesizing systems’ may, for example, be used in the characterization of plasmid clones and the study of structural mutations. In this study, a ‘coupled transcription-translation system’ has been used in an attempt to determine the effects of mutations I143C and I143S in Kir2.1 on channel structure, with particular emphasis on the formation of disulphide bonds. This system combines DNA transcription and mRNA translation to give a protein product. The proteins synthesized during coupled transcription-translation were then analysed by SDS-PAGE followed by autoradiography.

It is possible that disulphide formation in Kir2.1 mutants I143C and I143S occurs after translation. Canine pancreatic microsomal membranes were therefore included in some of the protein synthesis reactions described in this study. The use of microsomal membranes, together with coupled transcription-translation, facilitates the study of cotranslational and early post-translation events.

2.5.2a. The coupled transcription-translation assay

In vitro transcription-translation was carried out using the TNT® (T7) coupled reticulocyte lysate system (Promega Corporation) according to the manufacturer instructions. The reagents were removed from storage at -70°C and immediately placed on ice. For reactions to

121 be carried out in the absence of canine pancreatic microsomal membranes, the following components were added, step by step, to previously labelled 0.5ml micro-centrifuge tubes: 25/xl of rabbit reticulocyte lysate, 2^tl reaction buffer, 1/xl amino acid mixture minus methionine, 1/xl ribonuclease inhibitor, 1/xl RNA pdymerase, 4/xl (~40/xCi) L-[35S]methionine (specific activity > lOOOCi/mmol; Amersham Pharmacia Biotech), 2/xl DNA (0.5/xg//xl) and 14/xl nuclease-free water to give a final volume of 50/xl per reaction. The reagents were gently mixed and the reactions incubated in a water bath at 30°C for 90 minutes.

For reactions to be carried out in the presence of microsomal membranes the following components were added, step by step, to a previously labelled 0.5ml micro-centrifuge tube: 12.5/xl lysate, 0.5/xl reaction buffer, 0.5/xl amino acid mixture minus methionine, 0.5/xl ribonuclease inhibitor, 0.5/xl RNA polymerase, 1/xl (~10/xCi) I^[35S]methionine, 1/xl DNA and 2 canine microsomal membranes and nuclease-free water to give a final volume of 25/xl per reaction. Oxidised glutathione (GSSG) was substituted for water at concentrations of 0.5 and ImM. Reactions were incubated as described above.

2.5.2b. Gel electrophoresis

The proteins synthesized during coupled transcription-translation were separated by SDS- PAGE using a Hoefer Scientific gel electrophoresis kit. Prior to assembly, the glass plates were cleaned with detergent and swabbed with 70% IMS. With spacers of uniform thickness at each vertical edge, the apparatus was assembled so that the glass plates were level. In initial experiments, an 8% polyacrylamide gel (0.1% w/v SDS; Tris pH 8.3, lOOmM; Bicine, lOOmM; ammonium persulphate, 150/xl of 10%; TEMED, 150/xl) was prepared and poured, being careful to avoid bubble formation. Later experiments used 10% acrylamide gels (10% SDS; mM: tris, 100; Bicine, 100). A fifteen well comb was immediately inserted into the gel mixture, being careful not to trap any air bubbles under the comb, and the gel put aside to set.

Meanwhile, the assay samples were removed from the water bath and prepared for loading. For each translation reaction, 2.5/xl of translated product was mixed with 20/xl loading buffer (Tris pH 6.8, 50mM; Glycerol, 25% v/v; SDS, 6%, w/v; Bromophenol blue, 0.01% w/v). (3- mercaptoethanol (p-ME) was added to samples to a final concentration of 40mM. Samples were run in the presence of a reducing agent. To denature the proteins, the samples were boiled for 2 minutes at 100°C. The protein samples were then centrifuged at 13,000rpm for approximately 4 minutes. Once the gel was set, the comb was carefully removed and the wells rinsed with distilled water. The glass plates were clipped into the reservoir frame and lowered into the electrophoresis tank. The inner and outer reservoirs of the tank were filled with running buffer (Tris pH8.3, lOOmM; Bicine, lOOmM; SDS, 0.1% w/v; Glycerol, 25%) so that the inner lower plate and the bottom of the outer plates were covered with buffer. Finally, the protein standards, pre-stained markers and samples were loaded into the appropriate wells and the gel run at 15mA initially then overnight at 4mA.

2.5.2c. Analysis of translation products

Following overnight electrophoresis, the gel was removed from the tank and soaked in an organic scintillant (Amersham International’s Amplify™ reagent) for fluorographic enhancement of the signals. This dramatically increases the sensitivity of detection of 35S- labeled proteins by converting the emitted energy of the isotope to visible light, thus increasing the proportion of energy that may be detected by X-ray film.

Prior to film exposure, the gel was dried under vacuum. The gel was placed on a double layer of Whatmann® 3MM filter paper, which had been dampened with distilled water, wrapped in plastic wrap and placed in a conventional gel dryer at 80°C for 1-2 hours.

When dry, the gel was exposed on X-ray film (Hyperfilm™, Ammersham) for 24 hours at

-70°C. The film was then developed using an Amersham Hyperdeveloper. The exposure time for further pieces of film was altered according to the appearance of the preceding exposures.

123 Chapter 3

An electrophysiological characterization of the strong inwardly rectifying K+ channel

3.1. Introduction

Inwardly rectifying potassium channels are currently grouped into seven subfamilies based on similarities in amino acid sequences and physiological characteristics, such as their degree of rectification (see section 1.2.4; for reviews see Doupnik et al., 1995; Fakler & Ruppersburg, 1996; Nichols & Lopatin, 1997). The second of these subfamilies, known as the Kir2.0 subfamily, comprises the ‘strong inward rectifier potassium channels’, characterized by their steep inwardly rectifying properties and time-dependent gating kinetics (see section 1.2.4b for detailed description; for review see Stanfield et al., 2002).

124 To date, four members of the Kir2.0 subfamily have been isolated from various rat, mouse, rabbit and human sources (see section 1.2.4b). These channels exhibit electrophysiological properties typical of native inward rectifiers including an asymmetrical K+ conductance, a shift in activation potential in accordance with Vm and [K+]0, dependence of conductance on the square root of [K+]0 and block by extracellular cations such as Ba2+ and Cs+ (see section 1.4.1; see for example Hodgkin & Horowicz, 1959; Hagiwara & Takahashi, 1974; Hagiwara & Yoshii, 1979; Leech & Stanfield, 1981).

The members of the Kir2.0 subfamily can be distinguished by their different unitary conductances (see section 1.2.4b). Kir2.2 displays the highest conductance at around 34pS- 41pS in 140 mM symmetrical K+ (Takahashiet al., 1995; Wible et al., 1995) followed by Kir2.1 at 22pS. Kir2.3 and Kjr2.4 exhibit lower conductances between 11 and 15pS under similar recording conditions (Makhina et al., 1994; Morishige et al., 1994; Collins et al., 1996; Welling et al., 1997; Topert et al, 1998).

A voltage- and [K+]0-dependent block by intracellular polyamines and Mg2+ has been shown to be the underlying mechanism for the strong rectification exhibited by these channels (see sections 1.4.2c & d). Several independent studies have demonstrated the structural determinants of ‘strong’ inward rectification to be two negatively charged residues, an aspartate (D) in M2 and a glutamate (E) in the C-terminus (see section 1.4.3; see also Stanfield et al., 1994b; Wible et al., 1994; Fakler et al., 1994; Yang et al., 1995a). Both residues are conserved throughout the Kir2.0 subfamily. Recently, an additional residue (E299 in Kir2.1) also conserved amongst the strong inward rectifiers has also been shown to be important in gating by polyamines (Kubo & Murata, 2001; see section 1.2.4d).

This chapter is entirely dedicated to the properties of the second member of this subfamily, Kjr2.2. Although the electrophysiological properties of Kjr2.2 have been characterized previously (Takahashi et al., 1994; Koyama et al., 1994), these studies utilized expression in Xenopus oocytes. Here, I look at expression of Kjr2.2 in mammalian cell lines. The content of this first chapter is intended as a reference for the electrophysiological properties of a ‘wild-type’ channel before presenting my results from mutagenesis studies in the outer selectivity filter of Kjrs 2.2 and 2.1 in subsequent chapters.

125 3.2. Methods

The electrophysiological properties of Kjr2.2 were examined following expression of the Kir2.2 gene in either murine erythroleukaemia (MEL) cells or Chinese hamster ovary (CHO) cells as described in the methods chapter. Whole cell currents were recorded with 140mM K+ in the patch pipette and 70mM K+ in the bath. For experiments investigating the effect of external K+ concentration ([K+]0), the bath solution was exchanged for solutions containing 14, 35 and 140mM K+, respectively. For the block experiments, Ba2+ and Cs+ were added to the standard bath solution (70K/70NMDG) at the concentrations indicated in figure 3.6.

Single channel currents were recorded in the inside-out or cell-attached configurations (see section 2.4.12) from CHO cells that had been transiently transfected with cDNA encoding the channel 48-60 hours before recording (see sections 2.3.2 and 2.3.4). For single channel recording, the pipette and bath solutions contained 200mM and 140mM K+, respectively.

3.3. Results

3.3.1. Kir2.2 expresses inward rectifier potassium channels in MEL and CHO cells

Figures 3.1 A and B show representative whole cell recordings from MEL and CHO cells, respectively. Currents recorded from untransfected cells are illustrated in the left panel of the figures whilst the right panels illustrate currents recorded from cells transfected with Kir2.2 cDNA. Untransfected cells exhibited no significant inwardly rectifying K+ currents. In contrast, cells in which the Kir2.2 gene had been expressed displayed prominent inward-going currents, which averaged steady-state amplitudes o f-3.82 ± 0.58 nA (n = 14; Vm = -132mV) in MEL and -2.97 ± 0.38 nA (n = 17; Vm = -132mV) in CHO cells. Very little current was seen at depolarizing potentials. At extreme hyperpolarized potentials, Kjr2.2 currents exhibited a time-dependent decay, or ‘inactivation’, which varied between cells.

Mean current-voltage ( I-V) relationships from untransfected (triangles) and transfected (circles) MEL and CHO cells are shown in figures 3.1C and D, respectively. Steady state currents were measured 30-45ms after the onset of the voltage step. As expected for a strong inward rectifier potassium channel, the steady-state I-V relationships are steep and approximately linear for membrane potentials more negative than the K+ equilibrium potential, Ek (-17mV).

126 Figure 3.1.

Kjr2.2 expresses inward rectifier Kf channels in MEL and CHO Cells. A and B show representative current recordings, obtained with 70mM [K+]0, from MEL and CHO cells, respectively. Currents were recorded in response to 50ms voltage steps from a holding potential of -17mV, to test potentials ranging from +63 to -132mV, in 5mV decrements. Every third record is shown for clarity. The left panel of each figure shows currents recorded from untransfected cells whilst the right panels show currents recorded from cells transfected with Kir2.2 wild-type cDNA. C and D show steady-state current-voltage relationships. The symbols represent the mean ± SEM, where larger than the symbol, for n = 11 - 17 cells. Circles, transfected cells; triangles, untransfected cells. A. B.

—e L

0.25nA 0.25nA 10ms

C. D. V ] (mV)

-150 -100 -50 -150 -100 -50

b -2

b -2 b -3

b - 4

L -5 The deviation from this linear relationship at potentials positive to Ek clearly reflects the asymmetrical conductance exhibited by these channels.

Figure 3.2 illustrates the relationship between chord conductance and membrane potential for Kir2.2 wild-type currents expressed in CHO cells. The chord conductance was calculated from the I-V relationships (Figure 3. ID) using:

Sk - — (Equation 3.l) V-EK where gKis the chord conductance to K+ and V - Ek is the driving force on K+. Its relation to membrane potential is best fitted with a double Boltzmann expression of the form:

8k - + ------4 ------— (Equation 3.2) & max -i , (y-vj 1 + exp y-y. 1 + exp K k2

where b was allowed to vary between 0 and 1. Vi and V2 (mV) give the voltages at which the relative conductance gK = 0.5 and kj and k2 (mV) are the factors effecting the steepness of the relationship. In control experiments, V\ and V2 were -25.1 ± 0.8 and -1.8 ± 2.2 mV, respectively. k\ and ki were 4.6 ± 0.2 and 9.8 ± 0.6 mV, respectively. In 70mM [K+]0, the maximum gK expressed was 31.2±3.8nS(n=16; mean ± SEM).

3.3.2. Dependence of rectification on Vm and [K+]0

A characteristic feature of inwardly rectifying potassium channels is their dependence of rectification on Vm and [K+]0. The selectivity of Kir2.2 wild-type currents for K+ was investigated by measuring the zero current (reversal) potential in external solutions of varying K+ concentration. The reversal potential was measured by fitting a fifth order polynomial function of the form:

f(x)= a + bx + cx2 + dxl + exx + f x 5 (Equation 3.3) to a small region of the I-V relationship where the current intersects the abscissa(I = 0).

Figure 3.3A shows the effect of changing [K+]0 on Kir2.2 wild-type currents. As [K+]c was lowered from 140mM to 70, 35 and 14mM, external K+ being substituted by N-methyl-D- glucamine, the amplitude of inward currents decreased. Lowering [K+]0 also resulted in a

128 1.0

0.5

-150 -100 -50

Vm (rnV)

Figure 3.2.

Representative conductance-voltage relationship for Kir2.2 wild-type. Plot of the normalized chord conductance (g'i0 versus membrane potential for Kh-2.2 wild-type expressed in CHO cells. The chord conductance was calculated from the I-V relationship using equation 3.1. Values were normalized to the maximum chord conductance and are plotted against Vm. The data are fit with equation 3.2, and the fit shown is to voltages ranging from +63 mV to -67 mV, where the conductance starts to decline owing to inactivation of the current at negative membrane potentials. V\ = -19.5mV, V2 = -0.7mV, k\ = 4.7mV, &2 = 11.4mV and b = 0.3.

129 parallel shift in theI-V relationship, both the reversal and rectification potentials 1 shifting (figure 3.3B, C) from -3.1 ± 0.7 (140mM; n = 12) to -16.5 ± 0.6 (70mM; n = 14), -31.1 ± 2.1

(35mM; n = 12) and -52.1 ± 0.7 (14mM; n = 12). With an internal K+ concentration ([K+]0 of 140mM, these values are in close agreement with K+ equilibrium potentials (EK) predicted from the Nemst equation (0, -17, -35 and -58mV for 140, 70, 35 and 14mM [K+]0, respectively). The I-V relationships (Figure 3.3C) clearly display the ‘cross-over’ phenomenon typical of inward rectifier K+ conductances (see section 1.4.1a), where an increase in [K +] 0 increases the outward current rather than decrease it as one would predict from a simple change in EK.

The relationship between [K+] 0 and reversal potential is shown in figure 3.4A. As expected for a K+ selective ion channel, the reversal potential depends on [K +] 0 shifting 49.7 ± 0.7 mV

(mean ± SEM, n = 12 - 14) for a tenfold change in [K+]0. This is in close agreement to values obtained in previous studies on native channels (49mV), though NaCl was used to substitute for external K+ (Sakmann & Trube, 1984a). Together, these results show that Kjr2.2 is a K+ selective channel and that rectification depends on Vm and [K+]0.

3.3.3. The conductance of Kjr2.2 depends on [K+]o0 38

Figure 3.4B clearly illustrates that as [K+] 0 is lowered, the slope conductance of Kjr2.2 wild- type currents, measured from the straight portion of the I-V relationship, decreases. The K+ conductance of inward rectifier K+ channels in skeletal muscle (Aimers, 1971), starfish egg cell membranes (Hagiwara & Takahashi, 1974), ventricular cells (Sakmann & Trube, 1984a) and of cloned inward rectifier potassium channels (for example see Ho et a l, 1993; Kubo et al., 1993a; Makhinaet al., 1994) has been shown to depend, approximately, on the square root of the [K +] 0 (see also section 1.4.1b). Figure 3.4B shows a double logarithmic plot of the slope conductance as a function of the [K+]0. The data points have been fitted with a linear regression of the form:

G = a ([k +]0)x (Equation 3.4) where A is a constant and x represents the slope of the fit and has the value 0.38 ± 0.05 (n = 9), similar to the relationships for other inward rectifier potassium channels (0.46, Perrier et

1 The values quoted in the text were corrected for junction potentials predicted from the Junction Potential Programme ‘JPCalc’ copyright (c) 1992-1997 PH Barry. Figure 3.3B and C do not show this correction.

130 Figure 3.3.

Kir2.2 wild-type currents are sensitive to changes in external lC concentration. A, representative current recordings obtained with external solutions of various K+ concentrations as indicated in the figure. Currents were elicited in response to 50ms voltage steps, from a holding potential of -17mV to test potentials ranging from +63 to -132mV, in 5mV decrements. Every fifth record is shown for clarity. B shows steady-state current- voltage relationships for Kir2.2 wild-type currents recorded in external solutions of various K+ concentrations. Steady-state current amplitudes, normalized to the maximum current amplitude at -132mV with 70mM [K+]0, are plotted. The symbols represent the mean ± SEM, where larger than the symbol, for n = 12 - 14 cells. Triangles, 14mM [K+]0; squares, 35mM [K+]0; circles, 70mM [K+]0; inverted triangles, 140mM [K+]0. C, expansion of current- voltage relationships (shown in B) about the x-axis to show ‘cross-over’ phenomenon. Individual current-voltage relationships are colour coded for clarity. Symbols as in B. 14mM [K+]o

35mM [K+]o

70mM [K+]o

0.02 I__ b

140mM [K+]o

- - 0.01

L -0.02 B.

Figure 3.4. Kir2.2 wild-type currents are dependent on Vm and [JCjo. A , semilogarithmic plot of the external K+ concentration ([K+]0) versus membrane potential (Vm). The symbols represent the mean ± SEM, where larger than the symbol, for n = 12 cells. The data has been fitted with a linear regression (solid line), which has a slope of 48.6mV per decade. The dashed line represents the fit to reversal potentials predicted from the Nemst equation with a slope of 58mV per decade. B, double logarithmic plot of the external K+ concentration ([K+]0) versus slope conductance (gV)- Slope conductance was measured in the linear portion of the I-V curves at potentials 20mV, and more negative, to the predicted potassium equilibrium potential (£k). The symbols represent the mean ± SEM, where larger than the symbol, for n = 9 cells. The slope of the regression line fitting the data points is 0.39. al., 1992; 0.47, Kubo et al., 1993a; 0.33, Makhinaet al., 1994). Thus, K;r2.2 wild-type channel conductance is approximately proportional to the square root of the [K+]0.

3.3.4. Dependence of activation gating on Vm and [K+]0

Inwardly-rectifying potassium channel currents commonly display a property known as activation whereby currents initially increase, in an exponential manner, on hyperpolarization (see section 1.4.1c). Activation of Kir2.2 wild-type currents, under control conditions (70K/70NMDG) is illustrated in figure 3.5A. The time constants of activation (xact) were obtained by fitting the currents with a single exponential function of the form:

(Equation 3.5)

The results of the fitting procedure are shown superimposed on the activation traces in Figure 3.5A

The relationship between the activation time constants and membrane potential are shown in figures 3.5B and C, and are fitted with a linear regression of the form:

(Equation 3.6)

where a is xact (ms) at OmV, and b represents the voltage-dependence of activation gating.

The time constants for activation (xact) clearly decrease (figure3.5B, C), falling e-fold for a

17.1 ± 1.5mV hyperpolarization in MEL (n = 7) cells. A similar voltage-dependence of activation was seen in CHO cells at 70mM [K+]0; xact = 16.5 ± 1.5mV (n = 10), 35mM [K+]0; xact= 17.2 ± 0.2mV (n = 9) and 14mM [K+]o; xact = 15.6 ± 1.2mV (n = 12). These values agree closely with those reported by others for native inward rectifiers (e.g 18mV; Leech & Stanfield, 1981). Although the voltage-dependency of activation time constants did not change with varying [K+]0, the range over which the current activated shifted approximately with the change inEK. Thus, gating of Kjr2.2 wild-type currents depends both on membrane potential and [K +] 0 as has been previously described for native (see for example Hagiwara & Yoshii, 1979; Leech & Stanfield, 1981) and cloned (see for example Stanfield et al., 1994a) inward rectifiers.

133 Figure 3.5.

Activation ofKir2.2 wild-type currents under hyperpolarization and dependence on [if] 0. A, representative current recordings elicited by 50ms voltage steps of 20, 30 and 40 mV negative from £ K (-17mV). The currents were fitted with equation 3.5 and the fits are shown superimposed on the traces. B , semi logarithmic plot of time constants for activation (xact) plotted against voltage (Vm) for experiments with 70mM [K+]0. The symbols represent the mean ± SEM, where larger than the symbol, for n = 3-10 cells. Filled circles, MEL cells; open circles, CHO cells. The solid and dashed lines show the best fits to the mean data for Kjr2.2 wild-type channels expressed in MEL and CHO cells, respectively. C, semilogarithmic plot of time constants of activation (xact) plotted against voltage (Vm) for experiments, in CHO cells, with various external K+ concentrations. The symbols represent the mean ± SEM, where larger than the symbol, for n = 3 - 16 cells. Circles, 70mM [K+]0; triangles, 35 mM [K+]0; hexagons, 14mM [K+]0. Note that the scaling of the x axis differs from that in part b. The solid line shows a linear fit to the mean data for 70mM K+, whilst the dashed lines show the mean fit for the activation time constants obtained in 70mM [K +] 0 shifted -17.5mV and - 40.5mV, respectively. -17mV 3.3.5. Block of Ku-2.2 by extracellular cations

Another distinguishing feature of inwardly rectifying potassium channels is their susceptibility to block by extracellular cations such as Ba2+ and Cs+. Thus, the effect of Ba2+ and Cs+ on Kir2.2 wild-type currents was examined and the results are shown in figure 3.6. Kir2.2 wild-type currents were initially recorded in 70mM external K+ solution, then perfused with 70mM K+ external solution containing Ba2+ or Cs+ at the concentrations indicated in the figure. Currents were elicited in response to voltage steps, of 1000ms duration, from a holding potential of -17mV (equilibrium potential for K+; EK) to test potentials ranging from +80mV to -115mV in 5mV steps. Every fifth record is shown for clarity.

As is clearly illustrated, Ba2+ and Cs+ block Kir2.2 wild-type currents in a concentration-, time-, and voltage-dependent manner, block increasing as the membrane potential is hyperpolarized. Ba2+ and Cs+ blocked both the instantaneous and steady-state currents, though inhibition by Cs+ exhibited less time-dependency and had little effect on currents at potentials more positive than -70mV. The block of Kir2.2 wild-type currents by extracellular Ba2+ and Cs+ is characteristic of inward rectifiers, as previously shown for RB-IRK2 (Koyama et al., 1994) and MB-IRK2 (Takahashiet al., 1994). The inhibition of Kir2.2 currents by Ba2+ is described in more detail in chapter 4. An extensive description of the inhibition of a strong inward rectifier by Cs+ is described elsewhere (see Abrams, 2000).

3.3.6. Single channel current properties of Kjr2.2

Figure 3.7A shows representative, single channel current traces recorded using the cell- attached patch configuration, from CHO cells expressing Kir2.2 wild-type channels. As for macroscopic currents, single channel currents were observed predominantly in the inward- going direction thus exhibiting strong inward rectification. The patch from which example recordings are shown in figure 3.7A contained four channels. At potentials positive to -80mV events in which all four channels were closed were rare but as the membrane was hyperpolarized to -lOOmV and more negative, bursts of single channel openings could be resolved. Preliminary data (not shown) suggested that P0pen decreases with hyperpolarization, nPo decreasing from 0.85 at -40mV to 0.48 at -160mV.

135 Figure 3.6.

Block of Kir2.2 wild-type currents by Ba2+ and Cs+. and A B show representative current recordings elicited by voltage steps to +63, +38, +13, -12, -37, -62, -87 and -112mV for Is from a holding potential of -17mV. The left panel of each figure shows records obtained under control conditions (70K+/70NMDG) whilst the right panel shows the effect of external Ba2+ (A) or Cs+ (.B) at the concentrations indicated. C and D show steady-state current-voltage relationships under control conditions and in the presence of Ba2+ (C) or Cs+ ( D). Steady-state current amplitudes, normalized to the maximum current amplitude at -132mV under control conditions, are plotted. The symbols represent the mean ± SEM, where larger than the symbol, for n = 5 cells. Filled circles, control; filled triangles, 5/xM cation; open triangles, 50/tM cation ;filled squares, 500/tM cation. A 0|iM Ba2+ 5^M Ba2+ B OfiM Cs + 50fiM Cs+

2nA 2nA 2+ 50^iM Ba 500(iM Cs+ r 250ms 250ms

C. A

-150 -100 -50 -150 -100 -50

ss Figure 3.7.

Single channel current properties of Kir2.2. A, Representative single channel recordings obtained in 140mM [K+]i and 200mM [K+]0 from Kjr2.2 wild-type in response to various membrane potentials as noted in the figure. Channel openings are represented by pulses of inward current seen as downward deflections from the baseline (dashed line). The dotted lines represent the open levels for one, two and three channels. B, Plot of single channel current amplitude versus membrane potential for Kjr2.2 wild-type. The data are fit with a linear regression of the form i = y(V - Erew) where EKV is the predicted reversal potential, constrained to 9mV. The symbols represent the mean ± SEM, where larger than the symbol for n = 8 cells. Unitary Current Amplitude (pA) Figure 3.7B shows a plot of unitary conductance versus membrane potential for Kh-2.2. Slope conductances were determined by regression fits of the form:

i = y ( V - E rev) (Equation 3.7) where y is the unitary current and Zsrev is the predicted reversal potential, which in this case has been constrained to 9mV. The slope conductance (figure 3.7B) for Kir2.2 channels, with

200mM K+0 in the pipette and 140mM K* in the bath, was calculated as 38.6 ± l.lpS (mean ±

SEM; n = 8). This is in close agreement with previous measurements of unitary conductance for Kir2.2 expressed in HEK293 cells and Xenopus oocytes (38.0 ± 0.3pS, Yamashita et al.,

1996; 34.2 ± 2.1pS, Takahashiet a l, 1994; 1995).

3.4. Discussion

The electrophysiological basis for identification of a strong inward rectifier potassium channel rests on characteristics such as strong rectification of the macroscopic current, voltage-dependent block by Ba2+ and Cs+ and an appropriate single channel conductance. The results presented in this chapter clearly demonstrate that the whole cell and single channel properties of Kir2.2 heterologously expressed in both MEL and CHO cells are consistent with the electrophysiological properties of native strong inward rectifier potassium channels and Kir2.2 heterologously expressed in Xenopus oocytes (see section 1.4.1; see also Koyama et al., 1994; Takahashiet al., 1994, 1995), indicating that both MEL and CHO cells represent suitable expression systems for further mutagenesis studies of members of the Kjr2.0 channel subfamily.

As anticipated for Kir2.2, hyperpolarization from a holding potential of -17mV revealed large, rapidly activating, inward-going currents, whilst currents in response to depolarizations were small. At very negative membrane potentials the current exhibited a time-dependent decay, or ‘inactivation’. This has been previously observed for native (see for example Biermans et al., 1987; McKinney & Gallin, 1988) and cloned channels (see for example Kubo et al., 1993a; Takahashiet al, 1994; Shieh, 2000). Whilst Na-dependent inactivation of inward-rectifying currents has been well characterized in both frog skeletal muscle (Standen & Stanfield, 1979) and tunicate egg cell membranes (Ohmori, 1978; 1980; Fukushima, 1982), the mechanism underlying inactivation in Na-free solution has, until recently, remained unclear.

138 In their study of MB-IRK2 (Kir2.2), Takahashiet a l (1994) also observed a prominent, voltage-dependent inactivation of currents at negative membrane potentials. Consistent with their macroscopic data, they saw a steep decline in open probability of single Kir2.2 currents as the membrane was hyperpolarized. This was attributed to an increase in the occurrence of “long closed gaps”, channel closures of 200ms duration and more. In this study, preliminary data suggested that Kir2.2 currents expressed in CHO cells display a similar decline in open probability with hyperpolarization. However, owing to high expression levels (no single channel patches were obtained) analysis to predict whether this was a consequence of an increase in the number of long closed events could not be conducted.

A decrease in open probability with hyperpolarization has also been reported for Kirl.lb (Chepilko et a l, 1995; Choe et a l, 1998; 1999) and Kir2.1 (Choe et a l, 1999). Although the decline in open probability for Kirl.lb is less voltage dependent than for the strong inward rectifiers Kir2.1 and Kir2.2, it has been clearly associated with an increase in the number of long closures. Chepilko et a l (1995) suggested that the voltage dependence of the long closures is an intrinsic property of the channel. Although it was noted that the ‘putative blocked state’ of Kirl.lb channels in the presence of 0.5mM Ba2+ had a mean lifetime very similar to that of the long-lived closed state of the channel under control conditions, the similarity of these time constants was considered fortuitous. Further explanation for the possible mechanism underlying the decrease in open probability with hyperpolarization was not given.

Several hypotheses have been proposed to account for ‘inactivation’ of inward rectifier currents at negative membrane potentials in the absence of Na+. Early theories from studies in skeletal muscle attributed the decay of currents during hyperpolarization to depletion of K+ ions in the t-tubule system (Adrian & Freygang, 1962a; Adrian et a l, 1970a). However, as described in section 1.4.Id, the decline in conductance during a maintained hyperpolarization could not be entirely attributed to depletion of tubular K+ (Adrian et a l, 1970b; Aimers, 1972a, b), but also involved changes in membrane permeability. Interestingly, Kir2.2 channels are co-localized with SAP97 in the t-tubules of cardiac ventricular myocytes (Leonoudakis et a l, 2000). Thus, in their native environment, inactivation of Kjr2.2 channels at very negative membrane potentials is likely to involve depletion of tubular K+.

Could depletion of extracellular K+ be responsible for the inactivation under the experiment conditions described herein? One study has reported that ion depletion/accumulation can

139 occur following overexpression of cloned channels in mammalian expression systems (Frazier et al., 2000). Following expression of Kv2.1 channels in HEK293 cells, Frazier et al. (2000) found that long depolarizations in these small cells produced current-dependent changes in [K+]j that mimicked inactivation and changes in ion selectivity (Frazier et al., 2000). We have seen that hyperpolarization activates large (>lnA), inward-going currents in mammalian cells expressing Kir2.2. In such small cells (lOpF), this could lead to changes in the K+ gradient, particularly at very negative membrane potentials. If depletion of [K+] 0 or accumulation of [K+]i was the underlying cause for inactivation, we might expect to see the current reversing at potentials more negative than those predicted by the Nemst equation. However, this is not the case. In fact, currents are reversing at values more positive than predicted by the Nemst equation. Furthermore, [K+] 0 depletion is also an unlikely cause for inactivation in this study since high extracellular/intracellular [K+] has been used.

An alternative hypothesis is that block by external cations including Ca2+, Mg2+ and Ba2+ may be responsible for the hyperpolarization induced decay of inward rectifier K+ currents. Whilst the extracellular recording solution contained 2mM Ca2+ and Mg2+, Ba2+ was not added. 'j. However, it is possible that block by contaminating levels of Ba present in commercially available salts could be partly responsible for the inactivation of Kjr2.2 currents at hyperpolarized potentials since Kir2.2 appears to exhibit a high affinity for Ba2+ (see Chapter 4). Furthermore, several observations from previous studies on other members of the inward rectifier potassium channel family support the idea that block by trace amounts of Ba2+ may play a pivotal role in channel inactivation at very negative membrane potentials.

First, Kjr2.4, which has a low affinity for Ba2+ does not exhibit any voltage dependent inactivation during a 500ms voltage-pulse in Na+-free external solution (Topert et al., 1998). In contrast, Kir2.2, which is highly sensitive to Ba2+ (see chapter4), displays prominent inactivation at hyperpolarized potentials. Thus, channel sensitivity to Ba2+ appears to parallel voltage dependent inactivation.

Single channel studies of gating in Kirl.lb (ROMK2) provide further evidence that Ba2+ block may underlie channel inactivation in certain experimental situations. In the absence of divalent cations, Kirl.lb currents exhibit long closures of similar duration and voltage dependency to closures in the presence of Ba2+ (Choe et al., 1998). With 5mMEDTA, the long closed events were virtually eliminated indicating that contaminant divalent cations induce the long closed states. Further experiments using SO4 2', which forms an insoluble

140 'J, 'J. complex with Ba , implied Ba to be the main blocker, since long closures occurred less frequently on addition of K 2SO4 to the pipette solution (Choe et a l, 1998).

Similar observations have also been reported for Kir2.1 (Choe et a l, 1999; Guo & Lu, 2002) and for other potassium channels including calcium-activated potassium channels (Diaz et al., 1996). Choe et al. (1998) proposed that a contaminating Ba2+ concentration as low as 0.1/zM would be sufficient to promote the long closed periods observed in Kirl.lb at very negative membrane potentials.

Shieh (2000) showed that voltage-dependent processes controlled the inactivation of Kir2.1 channels, which was reduced by increases in external [K+]. This ‘protective function’ of K+ could provide further support for block of the channel by external Ba2+ at hyperpolarized potentials since K+ is known to compete with Ba2+ for the binding site in K+ channels (Standen & Stanfield, 1978a; Eaton & Brodwick, 1980; Armstrong & Taylor, 1980; Vergara & Latorre, 1983). However, Shieh (2000) carried out their experiments in the presence of 5mM EDTA, yet still observed inactivation implying that block by external Ba2+ was not the cause of inactivation. Instead, inactivation of Kir2.1 currents was compared with the C-type inactivation of Kv channels and it was proposed that K+ ions may protect the channel from inactivation by binding within the outer pore of Kir2.1, which becomes obstructed during voltage change thereby preventing inactivation.

An arginine (R) at position 148 in Kir2.1 was shown to influence this binding site since mutation of this residue to a tyrosine (Y) removed inactivation (Shieh, 2000). However, a previous study has shown that mutation of R148 in Kjr2.1 to histidine (H) alters the rate at which Ba2+ blocks the channel (Sabirov et al., 1997). Furthermore, Kir7.1, which contains a methionine at this position, does not exhibit inactivation at hyperpolarized potentials (Doring et al., 1998) and has a low affinity for Ba2+ (see Krapivinsky et al., 1998a; Doring et al., 1998).

Shieh and Lee (2001) continued to investigate whether inactivation of Kir2.1 channels at negative membrane potentials is a consequence of block by permeant ions or a conformational change induced by the binding of permeant ions. Their approach involved

substituting Ammonium (NH 4 +) or Thalium (Tl+) ions for K+ to examine whether other permeant ions protected against inactivation. In the presence of NHi+ or Tl+, Kir2.1 channels exhibited two types of current decay: a ‘permeant-ion-protected inactivation’ in which higher

141 concentrations of N H / or Tl+ protected Kir2.1 channels from inactivation and a ‘permeant- ion-enhanced inactivation’. At higher concentrations of N H /, only the permeant-ion-induced inactivation was observed. Shieh and Lee (2001) speculated that N H / (or Tl+) could be acting as a permeable blocker of Kir2.1, or alternatively, that binds directly to the channel, inducing inactivation by destabilizing the open state. However, the rate of decay in lOmM N H / did not correlate well with the effective rate coefficient for ion transfer suggesting that block by NfLf1- was not the cause of inactivation. In support of their hypothesis that permeant ions bind to the channel thereby causing a conformational change, Shieh and Lee (2000)

found that NH 4 +-induced inactivation was very sensitive to temperature, inactivation increased by a factor of 2.76 with a 10°C increase in temperature. Furthermore, the R148Y mutation used in their previous study (Shieh, 2000), which alters the dimensions of the outer

pore, abolished NH 4 +-induced inactivation.

However, a recent study suggests that in the absence of external cations and with low [K+]0, hyperpolarization-induced inactivation is a consequence of block by residual hydroxyethylpiperazine (HEP), present in HEPES buffered solutions (Guo & Lu, 2002). When the HEPES present in extracellular solutions was replaced with phosphate, inactivation at very negative membrane potentials was removed. Thus, it appears that the K+-sensitive hyperpolarization-induced inactivation exhibited by inward rectifiers is not an intrinsic gating property of these channels.

Based on the observations described above it seems likely that block by external cations, including trace amounts of Ba2+, and block by HEPES from the external solution may have contributed to the voltage-dependent inactivation of Kir2.2 currents at very negative membrane potentials reported herein. However, without further experimentation it is impossible to conclude which of the above mechanisms may underlie hyperpolarization- induced inactivation of Kir2.2 currents.

Kir2.2 also exhibited the characteristic dependency of rectification on [K+]0. It is now known that these characteristic properties of Kjr channels are conferred by an arginine, R149 in Kir2.2, conserved in the outer pore of most Kir channels. Kir7.1, which carries a methionine at this position, exhibits a shallow dependence of conductance on [K+]0 in addition to a non-saturating conductance-voltage relationship. Replacement with arginine, as conserved throughout the Kjr family, restored the steep dependence of rectification on [K+]0 (see Doring et al., 1998).

142 Finally, the conductance-voltage relationship clearly follows a typical Boltzmann function with saturation, which indicates a deviation from the independence principle by ion-ion or ion-pore interactions. Furthermore, in keeping with previous studies on strong inward rectifier potassium channels (see for example Leech & Stanfield, 1981;Lopatin et al., 1995; Lopatin & Nichols, 1996),the conductance-voltage relationship exhibited both shallow and steep components. Intracellular Mg2+ was omitted from recording solutions throughout this study, thus the shallow and steep components of the conductance-voltage relationship are likely to represent instantaneous and steady state components of polyamine block, respectively. Block of HRK1 (Kir2.3) and Kir2.1 by spermine (or spermidine) have previously been demonstrated to exhibit two components (Lopatin et al., 1995;Lopatin & Nichols, 1996). This has been explained by a model in which two molecules of spermine block the pore in a sequential manner, binding initially to a shallow site to block then binding to a second site located deep within the pore (see Lopatin et al., 1995; see also section 1.2.4d)

In conclusion, the whole cell and single channel properties of Kir2.2 heterologously expressed in both MEL and CHO cells are entirely consistent with the electrophysiological properties of native strong inward rectifier potassium channels and Kir2.2 heterologously expressed in Xenopus oocytes (see section 1.4.1; see also Koyama et al., 1994; Takahashiet al., 1994, 1995). Thus, both MEL and CHO cells represent suitable expression systems for further mutagenesis studies of members of the Kjr2.0 channel subfamily.

143 Chapter F our

A comparison of Ba blockage and unitary conductance in the murine K+ channels Kir2.1 and Kir2.2.

4.1. Introduction

A feature of inwardly rectifying potassium channels in different cell types is their high sensitivity to block by barium ions (Ba2+). The blocking action of Ba2+ has been demonstrated for a large number of Kjr channels, both native and cloned. These include Kn channels of frog skeletal muscle (Standen & Stanfield, 1978a), starfish eggs (Hagiwara et al., 1978), heart cell membrane (Sakmann & Trube, 1984a, b; DiFrancesco et al., 1984), rat corticotropes (Kuryshev et al., 1997) and the cloned channels Kirl.la (Loffler & Hunter, 1997) and Kirl.lb (Zhou et al., 1996).

144 Although Ba2+ has been shown to block all members of the Kir2.0 subfamily (Kubo et a l, 1993a; Takahashiet a l, 1994; Koyama et a l, 1994; Morishige et a l, 1994; Topert et a l, 1998;Hughes et a l, 2000), extensive characterization of Ba2+ blockade has been described only for Kir2.1 (Shiehet a l, 1998;Alagem et a l, 2001). Here, block of Kir2.2 currents by Ba2+ is investigated further.

The expression ‘open channel block’ is often used to describe the mechanism by which Ba2+ is believed to block potassium channels. With a similar crystal radius to K+, Ba2+ can enter the pore of the open channel, but because of its divalent charge blocks current by binding to K+ binding sites with a high affinity (see Standen & Stanfield, 1978a; Neyton & Miller, 1988; Sabirov et a l, 1997; Jiang & MacKinnon, 2000). Previous studies in inward rectifier (Standen & Stanfield, 1978a), delayed-rectifier (Eaton & Brodwick, 1980; Armstrong & Taylor, 1980) and calcium-activated (Vergara & Latorre, 1983) potassium channels have shown that K+ appears to compete with Ba2+ for the binding site. These findings suggest that Ba2+ ions interact with the amino acid residues lining the ion conduction pore. Thus, Ba2+ is a very useful tool for probing the molecular nature of potassium channel pores.

In general, the strong inward rectifier Kir2.2 exhibits electrophysiological characteristics similar to those of the closely related channel Kir2.1 (Takahashiet a l, 1994; Koyama et a l, 1994; Takahashiet a l, 1995). However, differences in unitary conductance (Takahashi et a l, 1995)and in sensitivity to Ba2+ (Thompson et a l, 1998)have been reported. The P-regions of Kir2.2 and Kir2.1 are virtually identical. However, leucine (L) at position 148 of Kjr2.2 is replaced by a phenylalanine (F) in Kir2.1, as illustrated in the partial sequence alignment of Kir2.1 and Kjr2.2 below:

K ir2.1 139 T QTTIGYGFR C U9

Kir2.2 140 T Q T TI G Y G L R C 150

A previous study has shown that a single amino acid difference between the P-regions of Kirl.lb and Kir2.1 accounts for the 10-fold difference in sensitivity to Ba2+ exhibited by these two channels (Zhou et al., 1996). Furthermore, single amino acid mutations in the P-regions of potassium channels have been shown to affect unitary conductance (see for example MacKinnon & Yellen, 1990; Taglialatela et a l, 1993; Choe et a l, 2000). In this study, the mutants L148F in Kir2.2 and F147L in Kjr2.1 were constructed to determine whether this

145 single difference in the P-region sequences of Kir2.2 and Kjr2.1 accounts for the differences in unitary conductance and Ba2+ block displayed by the two channels.

4.2. Methods

The electrophysiological properties of wild-type and mutant Kir2.2/Kir2.1 channels were examined following expression of the relevant gene in either CHO or MEL cells as described in the methods chapter. Whole cell currents were recorded from transiently transfected CHO cells, with 140mM K+ in the patch pipette and 70mM K+ in the bath. External barium

([Ba2+]0) was applied by changing the control bath solution for solutions containing 1 0 - 300jttM BaCl2.

For Ba2+ block experiments, single channel currents were recorded in the cell-attached patch configuration from stably transfected MEL cells with 70mM K+ in the patch pipette and 140mM K+ in the bath. Ba2+ was added to the pipette solution at either 100, 200 or 300jaM. The mean open time was calculated and corrected for missed events as described in section 4.3.4.

For comparison of unitary current between wild-type and mutant Kir2.2/Kir2.1 channels, single channel currents were recorded from transiently transfected CHO cells with 200mM K+ in the patch pipette and 140mM K+ in the bath.

4.3. Results

4.3.1. Ba2+ blocks Kir2.2 in a concentration-, time- and voltage-dependent manner

Figure 4.1 shows representative recordings from CHO cells expressing IQr2.2 channels in the presence of various concentrations of extracellular barium ([Ba2+]0). Currents are in response to 200ms hyperpolarizing voltage steps to -37, -57, -77, -97 and -117mV from a holding potential of -17mV. Control records are shown in figures 4.1 A, whilst figures 4. IB and C show currents obtained in the presence of 30/xM and 300/iM Ba2+, respectively. For clarity, figure 4. ID shows current recordings elicited in response to a single voltage step to -77mV, in the presence of 0, 30 and 300/tM Ba2+. Together, these traces clearly illustrate that Ba2+ blocks Kir2.2 currents in a concentration-, time- and voltage-dependent manner. Ba2+ inhibits both the instantaneous and steady-state current. However, only inhibition of the steady-state

146 Figure 4.1

Block of Kir2.2 wild-type currents by Ba2+. A, representative current recordings elicited in response to voltage steps to -37, -57, -77, -97 and -117mV for 200ms from a holding potential of -17mV under control conditions (70K/70NMDG). B and C, current recordings from the same cell following perfusion with 30 and 300/iM Ba2+, respectively. D, current recordings elicited in response to a single voltage step of -77mV, in the presence of 0, 30 and 300/iM Ba2+.

147 current by Ba2+ will be analysed here. For low [Ba2+]0, voltage steps of longer duration were used to ensure that currents reached steady-state.

4.3.2. Steady-state analysis of blockade of Kjr2.2 by Ba2+.

For a simple binding reaction where Ba2+ binds reversibly to the open channel, the kinetic equation can be expressed as follows:

*<, /c+b[Ba2+] C \ Ba2++ O s OBa2+ (Equation 4.1) closed kc open & b blocked

where kQ and kc are transition rate constants for channel opening and closing, k+\y is the blocking rate constant and &_b is the rate constant for unbinding. For such a reaction, the equilibrium dissociation constant (A'd) for binding of Ba2+ can be defined by:

k K d = ^ ~ (Equation 4.2) *+b

The Kd was obtained by examining the steady-state fractional current (iBa/Icontroi) at different membrane potentials (Vm). Steady-state currents obtained in the presence of Ba2+ were normalized to currents obtained in the absence of Ba2+ at the same potential. For example, at -77mV, and in the presence of lOfiM Ba2+, the steady-state fractional current for IQr2.2 wild-

type was 0 .1 0 ± 0 .0 1 (n = 6 ).

Figure 4.2A shows a plot of steady-state fractional currents versus extracellular Ba2+ concentration for different membrane potentials. Data for concentrations of Ba2+ lower than 10/zM were not included in the analysis because these results exhibited a large scatter, owing to series resistance errors arising as a result of large currents. Thus, the fits shown in the inset of figure 4.2A are simulated, based on fits to the data obtained at higher concentrations of [Ba2+]„, which are shown in the main figure. The data were fit using the Hill-Langmuir equation:

I ( ( [r,„2+M 1 [Ba2t] Ba _ 1 + (Equation 4.3) ^Control Kd

148 where the hill coefficient was constrained to 1 , based on an assumption of one-to-one binding of Ba2+ in the pore of Kjr2.2 as previously assumed for native strong inward rectifiers (see for example Standen & Stanfield, 1978a). At -77mV, Kjr2.2 was half-blocked by 1.13/iM.

Figure 4.2B shows the relationship between the Kd (derived from the plots in figure 4.2A) versus membrane potential (Vm). The data were fit to the Woodhull equation:

(Equation 4.4)

where Kd(0) is the concentration of blocking ion causing half-maximal block at OmV. <5 is a slope factor indicating the voltage sensitivity at Kd, and z, F, R and T have their usual meanings ( RT/F = 25mV at 20°C).

The voltage dependence of the Kd for Ba2+ binding in Kir2.2 wildtype decreases e-fold for a

32.5mV hyperpolarization. Kd(0) and 8 were 13.68pM and 0.40, respectively. In its simplest

terms, a 8 value of 0.40 suggests that the binding site for Ba2+ is located approximately 40% of the way across the membrane electrical field. However, the electrical distance is not necessarily the same as the physical distance because the membrane electric field may not be uniformly distributed along the pore. The Kd-Vm relationship for Kir2.2 became less steep at Vm more negative than -107mV. This may reflect Ba2+ dissociation into the intracellular space at very negative membrane potentials as previously described for Kir2.1 heterologously expressed in Xenopus oocytes (Shieh et al., 1998) and for the high-conductance Ca2+ - activated potassium channel (see Neyton & Miller, 1988).

4.3.3. Kinetic analysis of the macroscopic Kjr2.2 current blockade by [Ba2+]0

The voltage-dependent inactivation of whole cell currents in the presence of Ba2+ follows a monoexponential time course and the time constant of the decay can be found by fitting the data with a single exponential function of the form:

(Equation 4.5) block

where/(r) is a function of time, t is the time measured from the onset of the voltage protocol, Tbiock is the blocking time constant, A + c is the current amplitude at t = 0 and c is the steady state current. Figure 4.3A shows a semilogarithmic plot of the time constants of inactivation

149 (Tbiock) of the whole cell currents as a function of membrane potential (Vm) in the presence of various [Ba2+]G. The blocking time constants for Kir2.2 decreased exponentially with increasing hyperpolarization, falling e-fold for a 26.9 ± 1.08 mV (n = 7) hyperpolarization in the presence of 100/xM [Ba2+]D. Blockage by 100/xM Ba2+ at -97mV reached equilibrium in 7.5 ± 0.3ms (n = 6).

The time constant for blockade of the channel by Ba2+ (Tbiock) may be described by:

(Equation 4.6). block k+b[Ba2+] + k_b where k+b is the blocking rate constant, k.b is the rate constant for unbinding and [Ba 2+] 0 is the extracellular Ba2+ concentration.

The reciprocal of the measured time constant for onset of block (1/Tbiock) is a linear function of the blocker concentration [Ba2+], and is shown in figure 4.3B as a function of [Ba 2+] 0 at negative membrane potentials ranging from -67 to -107mV. k+b and k_b can be determined from the slope of the regression and from the y-axis intercept, respectively. Estimation of k.b by this method gave small negative values (not shown). However, given that k+b is so much greater than k.b, these negative values for k.b are within likely experimental error. Therefore, k.b values were calculated using k+b and K& according to equation 4.2 where K& is the equilibrium dissociation constant calculated by examining the fractional current (iBa/Icontroi) at different membrane potentials as described in section 4.3.2. k+b and k.b, calculated in this manner, are shown in numerical form in table 4.1 below.

Table 4.1. Rate constants for block and unblock of Kir2.2 by Ba2+

v m (mV) Mean k+b Mean k.b

(M 'V 1) (s'1)

-67 4.54 x 10s 0.71

-77 7.30 x 10s 0.83

-87 10.06 x 1 0 5 0.89

-97 13.48 x 10 5 0.94

-107 19.58x 10 5 1.08

150 Figure 4.2.

Voltage- and concentration-dependence o f the steady-state block o f Kir2.2 by Ba2+. A, semilogarithmic plot of the steady-state current, obtained in the presence of Ba2+, plotted as a fraction of that obtained under control conditions against Ba2+ concentration. The voltages are -67mV ( closed circles), -77mV ( open squares), -87mV ( closed triangles ), -97mV ( open inverted triangles) and -107mV ( closed diamonds). The symbols represent the mean ± SEM, where larger than the symbol, for n = 1 - 9 cells. The curves were fitted to the Hill-Langmuir equation (Equation 4.3) with the hill slope constrained to 1. Inset, simulated fits based on fits of actual data shown in main part of figure. B, semilogarithmic plot of the voltage-dependence of ATd(V). The data are fitted to the Woodhull equation (Equation 4.4). Fractional Current

0 1

Fractional Current o o o ©

o

o o

•p

o

K J V) (M) Figure 4.3

Concentration and voltage dependence of Ba2+ block in Kir2.2 wild-type. A , dependence of the time constant for the onset of blockade, Tbi 0Ck, upon membrane potential (Vm) at different [Ba2+]0. The symbols represent the mean ± SEM, where larger than the symbol, for n = 3 - 7 cells. Closed circles, \0fiM [Ba2+]0; open triangles, 30/rM [Ba2+]0; closed squares, lOOpM [Ba2+]0; open diamonds, 300/xM [Ba2+]0. B, dependence of l/Xbi0Ck upon [Ba 2+]0 at different membrane potentials (Vm). The symbols represent the mean ± SEM, where larger than the symbol, for n = 3 - 7 cells. Closed circles, -67mV; open squares, -77mV; closed triangles, -87mV; open inverted triangles, -97mV; closed diamonds, -107mV. The data are fit with linear regressions of the form 1/^block = ^+b[Ba ] + &-b j 00£ 0 0 ? oo r

o

oo?

OOP

009

008

F Figure 4.4 shows a semilogarithmic plot of k+b and k.b versus membrane potential (Vm). The data is fitted with equations modified from the Woodhull model (Woodhull, 1973):

zF3+b ^ £+b

(7fS 3 *_bOO = M 0) e x t | - ^ V n (Equation 4.8) where £+b(V) and k.b(V) are the association and dissociation rate constants, &+b(0) and ^_b(0) are the values of these rate constants at OmV, 8 +b is the electrical distance between the entrance of the channel pore and the energy barrier, 8 _b is the electrical distance between the Ba2+ binding site and its rate limiting barrier for exit, Vm is the membrane potential, z is the valency of the ion (z = 2 for Ba2+), and F, R and Thave their usual meanings.

Clearly, k+b is voltage dependent increasing e-fold for a 28.2mV hyperpolarization. £+b(0) and <5+b were 44.81 x 103 M '1 s'1 and 0.45, respectively. 5+b is comparable to the value of 8 (0.40) obtained from the analysis of steady-state fractional currents. Thus, the voltage- dependence of the Kd of Ba2+ block in Kir2.2 is due to the voltage-dependence of the association rate constant k+b. Indeed, k.b showed little voltage dependence (figure 4.4). In fact, k.b could be well fit by assuming that it is independent of voltage (dashed line, figure 4.4) giving &_b(0) as 0.89s’1. This may be because Ba2+ leaves the channel either to the outside, or in addition, to the inside. However, this was not investigated further.

4.3.4. Block of single Kir2.2 channel by Ba2+

The effects of Ba on Kir2.2 currents were investigated in more detail at the single channel level. Figure 4.5 shows representative single channel recordings, obtained in the cell-attached mode at the voltages indicated, with 200/xM [Ba2+] in the patch pipette. The corresponding open-time distributions are shown to the right of the traces. Open-time distributions were fit with a single exponential as described in section 2.4.17 and of the form:

n ( a ■'] f ( t )=^ — exp^'^ (Equation 4.9) j = i i T; where aj and Tj are the relative area and the time constant of component j, respectively. The number of components ( n) was set to the number of open or closed states and the data fit

153 / 105 / 101

C3- -B* - 10° ■*' & ■Q

L lO’1 1 I -110 -100 -90 -80 -70 -60

v m (mV)

Figure 4.4.

Voltage-dependency o f the association (k+\) and dissociation (k.tj rate constants for block o f Kir2.2 by Ba2+. Semilogarithmic plot of k+b (filled symbols) and k.b (open symbols) versus membrane potential (Vm) for K,r2.2. k+b was obtained from Xbiock by fitting linear regression lines to 1 /ibiock = ^+b[Ba2+] + k b, where k+b is the slope of the line. k.b was obtained from Kd analysis. The data are fit with equations modified from the Woodhull model (Equations 4.7 and 4.8), where for an e-fold change in k+b there is a 28.2mV change in membrane potential. k.b displayed little voltage dependence (solid line), and could be well fit by assuming that it is independent of voltage (red dashed line).

154 using the method of maximum likelihood (Colquhoun & Sigworth, 1983). Mean open times were corrected for missed events (i.e. short closures less than the minimum resolution, which were too brief to be detected) by multiplying the mean open time by the proportion of closed events detected, which was found by integrating the fitted closed time distribution between the minimum resolution and infinity (see Davies et al., 1992).

As the membrane was hyperpolarized, block of Kir2.2 currents was enhanced. This is reflected in the decrease in mean open time as illustrated in figure 4.5 and recorded in numerical form in table 4.2 below.

Table 4.2. Voltage-dependence of the mean open time for single Kir2.2 currents in the presence of 200/xM Ba2+.

Membrane Potential Mean Open Time (ms)*

(mV) *measured with 2 0 0 [Ba2+]0

-57 5.7 ± 2.0 (n = 2)

-77 2.8 ±0.9 (n = 3)

-97 1.8±0.3 (n = 3)

-117 1.610.3 (n = 2) o oo II -137 H

The concentration dependence of Ba2+ block at the single channel level is illustrated in figure 4.6. Single channel recordings were obtained in the cell-attached mode, at a holding potential of-97m V with 100, 200 or 300jiiM Ba2+ in the patch pipette. The mean open time decreases as the concentration of external Ba2+ is increased as indicated in the right hand panel of figure 4.6 and recorded in numerical form in table 4.3 below.

155 Figure 4.5.

Block of single Kir2.2 wild-type currents by Ba2+ is voltage-dependent. The left panel shows representative single channel recordings of Kir2.2 wild-type currents in response to 57mV (A), 77mV ( B), 97mV (C), 117mV ( D) and 137mV ( E) hyperpolarizing steps. Data are from cell- attached patches bathed in 140K+ with 70K+ and 200/iM Ba2+ in the patch pipette. The arrows indicate the level at which all channels are closed. The right panel shows open time distributions for the corresponding currents shown in the left panel. Events/bin Events/bin Events/bin Events/bin Events/bin

Time (ms) Ul ON Figure 4.6.

Block of single Kir2.2 wild-type currents by Ba2+ is concentration dependent. The left panel shows representative single channel recordings of Kir2.2 wild-type currents in response to a 97mV hyperpolarizing step. Data are from cell-attached patches bathed in 140mM K+ with 70K+ and 100/zM (A), 200/tM (B) or 300/iM (C) Ba2+ in the patch pipette. The arrows indicate the level at which all channels are closed. The right panel shows open time distributions for the corresponding currents shown in the left panel. Events/bin Events/bin Events/bin Table 4.3. Concentration-dependence of the mean open time for single Kir2.2 currents at -97mV.

[Ba2+]0 Mean Open Time (ms)*

0*M) *measured at -97mV

100 3.3 ± 0.3 (n = 4)

200 1.8 ±0.3 (n = 3)

300 2.0 ± 0.3 (n = 3)

Figure 4.7A shows a plot of the corrected mean open time (cmot) of the single channel currents as a function of membrane potential (Vm) in the presence of 100/tM [Ba2+]Q. The mean open time for Kir2.2 channel decreased exponentially with increasing hyperpolarization, falling e-fold for a 62.3mV hyperpolarization.

The simple model given for the blocking reaction given in equation 4.1 predicts that the mean open time (x0) in the presence of Ba2+ will be given by:

Tq = L (Equation 4.10) £+b[Ba2+] + £l where k+b is the association rate constant, kc is the transition rate constant for channel closure and [Ba2+] is the concentration of Ba2+ in the patch pipette. Thus, the reciprocal of t 0 is a linear function of the blocker concentration as shown in figure 4.7B for -97mV. k+b and kc can be determined from the slope of the regression and the y-axis intercept, respectively. The reciprocal of kc should predict the mean open time in the absence of Ba2+. At -97mV, kc was 0.2 ms'1, which predicts t 0 as 5ms in the absence of Ba2+. However, single channel currents were not recorded in control solutions, thus a comparison of the predicted xQ with x0 measured in the absence of Ba2+ cannot be made.

Figure 4.8 shows a plot of k+b versus membrane potential (Vm). The data is fitted with equation 4.7 modified from the Woodhull model. Clearly, k+b is voltage dependent increasing e-fold for a 30.0mV hyperpolarization. &+b(0) and 8+b were 66.46 x 103 s'1 M '1 and 0.44, respectively. Whilst the value for the association rate at OmV predicted from analysis of

158 Figure 4.7.

Concentration and voltage-dependence of mean open time. A, dependence of the mean open time (cmot) on membrane potential (Vm). The symbols represent the mean ± SEM for n = 3 - 4 cells. Data are from cell-attached patches bathed in 140K+ with 70K+ and lOO/tM Ba2+ in the patch pipette. B, plot of 1/cmot versus membrane potential (Vm). The symbols represent the mean ± SEM for n = 3 - 4 cells. Data are from cell-attached patches, bathed in 140K+ with 70K+ and 100, 200 or 300/iM Ba2+ in the patch pipette, in response to a 97mV hyperpolarizing step. The data were fitted with a linear regression of the form 1/cmot = fc+b[Ba2+] + kc. cmot (ms)

1/cmot (ms’ ) r- 107

r 106 ^

L 105

■160 -140 -120 -100 -80 -60 -40

Vm(mV)

Figure 4.8.

Voltage-dependency of the association rate constant (k+b) for block of Kir2.2 by Ba2+. Semilogarithmic plot of the association rate constant obtained from analysis of microsopic currents (filled symbols) and macroscopic currents (open symbols) versus membrane potential (Vm). The data are fitted with equation 4.7 modified from the Woodhull model, where for an e-fold change in kb, there are 30.0 and 28.2mV changes in membrane potential, respectively.

160 macroscopic currents is smaller (k+b(0) = 44.81 x 103 s'1 M '1) than that predicted from analysis of microscopic currents, the difference is only slight and well within experimental error. Thus, the values of £+b(0) and <5+b obtained from the analysis of single channel currents are in close agreement with those obtained from analysis of macroscopic currents (see section 4.3.3).

4.3.5. Effects of mutation L148F in Kir2.2 and mutation F147L in Kir2.1 on Ba2+ blockade.

9-i- TheA'd(O) for block of Kir2.1 by Ba has been reported by two separate groups as 62 (Shieh et al., 1998)and 131/zM (Alagem et al., 2001), respectively. With a ATd(0) of approximately 14/xM, Kjr2.2appears to have a higher affinity for Ba2+. However, Ba2+ block also depends on the extracellular K+ concentration and a meaningful comparison is possible only with identical K+ concentrations. Thus, block of Kir2.1 wild-type currents by Ba2+ was also investigated. Furthermore, block of the mutant channels Kir2.2 L148F and Kir2.1 F147L by Ba2+ was also investigated to determine whether the single difference in the P-region sequences of Kjr2.2 and Kir2.1 accounts for the difference in affinity for Ba2+.

Block of wild-type and mutant Kir2.2/Kir2.1 channels is shown in Figure 4.9. As for Kir2.2 wild-type, Ba2+ blocks Kir2.1 wild-type and mutant Kjr2.2/Kir2.1 channels in a concentration-, voltage- and time-dependent manner. The currents shown in figure 4.9 were obtained in the absence or presence of lOOjiiM [Ba ]0 in response to a single voltage step of -97mV.

Steady-state fractional currents were examined as described in section 4.3.2. In the presence of lOjiiM Ba2+, the steady-state fractional currents remaining at -77mV for Kir2.1 (0.44 ±

0.06; n = 6) were larger than for Kjr2.2 (0.10 ± 0.01; n = 6). Mutation F147L in Kir2.1 resulted in a significant reduction in the steady-state fractional currents remaining at -77mV in the presence of 10/xM Ba2+ (0.14 ± 0.03, n = 5; P < 0.05), while mutation L148F in Kjr2.2 had no significant effect (0.16 ± 0.03, n = 3; P > 0.05).

Figure 4.10A shows dose-response curves for Kir2.1/Kir2.2 wild-type and mutant channels at -77mV fitted with equation 4.3. At -77mV, Kir2.2 and Kir2.1 were half-blocked by 1.13/xM and 8.52jLiM, respectively. Kjr2.2 L148F required a similar concentration to Kir2.2 wild-type (2.34jLtM Ba2+) for half-blockage at -77mV, whilst mutation F147L in Kir2.1 reduced the concentration required for half-blockage to only 1.77jnM.

161 A. B. K. 2.2 WT K.2.1 WT r lOOpM [Ba2+]o lOOpM [Ba2+n ]o

OuM [Ba2+]o Ot-iM [Ba2+]o InA

100ms c. D. K 2.2 L148F K.2.1 F147L lOOuM [Ba2+]o lOOuM [Ba2+]o

Ol*M [Ba2+]o OuM (Ba2+]o

Figure 4.9.

Block o f wild-type and mutant Kir2.2/Kir2.1 channels by Ba2+. Representative current recordings from Kjr channels, elicited in response to a single voltage step of-97m V in the absence and presence of lOOpM [Ba2+]0. A, Kjr2.2 wild-type. B, Kjr2.1 wild-type. C, Kjr2.2 L148F. D, Kjj-2.1 F147L.

162 Figure 4.10.

Voltage- and concentration-dependence of the steady-state blockKw2.2IKir2.1 of wild-type and mutant currents by Ba2*. A, semilogarithmic plot of the steady-state current, obtained in the presence of Ba2+, plotted as a fraction of that obtained under control conditions against Ba2+ concentration for Kjr2.2/Kir2.1 wild-type and mutant channels at -77mV. The symbols represent the mean ± SEM, where larger than the symbol, for n = 1 - 7 cells. Solid red circles, Kjr2.2 wild- type; solid blue squares , Kjr2.1 wild-type; open red circles , Kjr2.2 L148F; open blue squares, Kir2.1 F147L. The curves were fitted to the Hill-Langmuir equation (Equation 4.3) with the hill slope constrained to 1. Inset, simulated fits based on fits of actual data shown in main part of figure. By semilogarithmic plot of the voltage-dependence of ATd(V) for Kir2.2/Kir2.1 wild-type and mutant currents. Symbols as in part A. The data are fitted to the Woodhull equation (Equation 4.4). Fractional Current Figure 4.10B shows the relationship between the Kd (derived in plots in figure 4.10A) versus membrane potential (Vm). The data were fitted with equation 4.4. The equilibrium dissociation constant (Kd(0)) for Kir2.1 was higher at 201/xM than for Kjr2.2 (13.68jitM) and 8 was 0.52 (8 for Kjr2.2 = 0.40). The values of £d(0) and 8 for Kir2.1 are in close agreement with previously reported values (Kd(0) = 165/tM, 5 = 0.54), where block of Kjr2.1 by Ba2+ was measured under similar conditions (see Thompson et a l, 2000a). Mutation F147L in Kir2.1 affected both the affinity and voltage dependence of the steady-state block by Ba2+ reducing ATd(0) and 8 to 15.54juM and 0.35, respectively. In contrast, mutation L148F in

Kir2.2 had no effect (/fd(0) = 20.20/tM, 8 = 0.33).

Figure 4.11A shows a plot of Tbiock versus membrane potential (Fm) in the presence of varying [Ba2+]0. The blocking time constants for Kir2.2/Kjr2.1 wild-type and mutant channels decreased exponentially with increasing hyperpolarization having similar voltage dependencies. Tbiock decreased e-fold for a 26.9 ± 1.08 mV (n = 7 ), 33.8 ± 3.3 mV (n = 13), 29.4 ± 0.5 mV (n = 7), and 29.0 ± 0.8 mV (n = 9) hyperpolarization for Kjr2.2, Kir2.1, Kir2.2 L148F and Kjr2.1 F147L respectively, in the presence of 100juM [Ba2+]0.

Blockage by 100/tM external Ba2+ at -97mV reached equilibrium faster (P < 0.05; ANOVA followed by Duncan’s Multiple Range Test) in Kjr2.2 (Tbiock = 7.5 ± 0.3ms; n = 6) than in

Kjr2.1 (Tbiock = 26.0 ± 1.9ms; n = 12). Mutation L148F in Kir2.2 resulted in a significant increase (P < 0.05) in Tbiock by 100/tM Ba2+ (Tbiock = 25.9 ± 2.6ms; n = 7) at -97mV.

However, the reverse mutation F147L in Kir2.1 had no significant effect on Tbiock (25.0 ± 1.8ms; n = 9).

Figure 4.1 IB shows a plot of the &+b and k.b versus membrane potential (Vm) for wild-type and mutant Kir2.2/Kir2.1 channels. The data were fit with equations 4.7 and 4.8, respectively. ^+b(0) was 45.13 x 103 M 'V 1 for Kir2.2 wild-type and 21.17 x 103 M '1 s’1 for Kir2.1 wild- type. Mutation L148F in Kir2.2 reduced &+b(0) to 16.99 x 103 M '1 s'1, mutation F147L in

Kir2.1 having no significant effect (&+b = 15.42 x 103 M’1 s'1). 5+b was similar for both wild- type and mutant channels (Kir2.2 = 0.45; Kir2.1 = 0.41; Kir2.2L148F = 0.37; Kir2.1 F147L = 0.42).

As for Kir2.2 wild-type, the voltage dependence of the Kd of Ba2+ block in K;r2.1 wild-type, Kir2.2 L148F and Kir2.1 F147L is predominantly due to the voltage-dependence of the

164 Figure 4.11

Voltage dependency of the time constant (Tbiock) and association (k+b) and dissociation (k.b) rate constants for block ofKir2.2/Kir2.1 wild-type and mutant channels by Ba2+. A, Voltage-dependency of the time constant for the onset of blockade, Tbiock* in the presence of

100/jM [Ba2+]0. The symbols represent the mean ± SEM, where larger than the symbol, for n = 3 - 13 cells. Solid red circles, Kir2.2 wild-type; solid blue squares, Kir2.1 wild-type; open red circles, Kjr2.2 L148F; open blue squares, Kir2.1 F147L. The data were fitted with a linear regression of the form 1/ibiock = &+b[Ba2+] + k.b. B, Semilogarithmic plot of k+b and k.b against membrane potential (Vm) for Kir 2.2/2.1 wild-type and mutant channels. Symbols as in part A. The data were fitted with equations modified from the Woodhull model (equations 4.7 and 4.8), though fc_b is fitted on the assumption that it is independent of voltage. . A B.

block 0 -i - 100 10 - -110 -110 a -100 -100 n r~ “n 9 -80 -90 Vm(mV) 9 -80 (mV) Vm -90 -70 7 -60 -70 _o □ -60 - 107 r- 101 / 105 / - - 106 io- 10 °

* b (s-') (M"‘s 165 association rate constant k+b. k.b showed little voltage dependence and could be well fit by assuming that it is independent of voltage (see figure 4.1 IB). Jc.b( 0) was then 0.89s'1 and 1.97s'1 for Kir2.2 and Kir2.1, respectively. Mutations L148F in Kir2.2 and F147L in Kir2.1 lowered &-b(0) to 0.46s'1 and 0.38s1, respectively.

5.4.6. Effect of mutation L 1 4 8 F in Ku-2.2 and mutation F 147L in Kir2.1 on unitary conductance

The results presented above show that the amino acid at position 148/147 in the pore of Kir2.2 ■j, and Kir2.1 partly controls the rate at which Ba ions enter the pore and also affects the channel affinity for Ba2+. Single channel currents were also measured for Kir2.2/Kir2.1 wild- type and mutant channels to examine whether this single difference between the pores of Kir2.2 and Kir2.1 has a similar effect on K+ permeation, and is therefore responsible for the differences in unitary conductance exhibited by the two channels.

Figure 4.12 shows representative, single channel current traces recorded using the inside-out patch configuration, from CHO cells expressing Kir2.2 wild-type, Kir2.1 F147L, Kir2.2 L148F and Kir2.1 wild-type channels at -160, -120, -80 and -40mV. As for macroscopic currents, single channel currents were observed predominantly in the inward-going direction exhibiting strong inward rectification.

Figure 4.13 shows a plot of unitary conductance versus membrane potential for wild-type and mutant Kir2.2/Kjr2.1 channels. Slope conductances were determined by regression fits of the form:

i = y(y - Erev) (Equation 4.11) where y is the unitary current and EKW is the predicted reversal potential, which in this case has been constrained to 9mV.

The slope conductance (figure 4.13) for Kir2.2 channels, with 200mM [K+]0 in the pipette and 140mM [K+]i in the bath, was higher (y = 36.4 ± 1.2pS; mean ± SEM; n = 3), than for Kir2.1 wild-type channels (y = 25.0 ± 0.3pS; mean ± SEM; n = 3). These values were not significantly altered (P > 0.05; ANOVA followed by Duncan’s Multiple Range Test) by the mutation F147L in Kir2.1 (y = 26.8 ± 0.4pS; mean ± SEM; n = 3) or by the mutation L148F in

Kir2.2 (y = 36.0 ± 0.8pS; mean ± SEM; n = 3).

166 Kir2.2 wild-type Kir2.1 F147L Kir2.1 wild-type Kir2.2 L148F ------A.

...I B. t

j ; u l «u«-«h

J 2pA 20ms Figure 4.12.

Comparison o f unitary conductance for wild-type and mutant Kjr channels . Single channel recordings obtained in 140mM [K+]i and 200mM [K+]0 at -160mV (A), -120mV (B), -80mV (C) and -40mV (Z)) from Kir2.2 wild-type, Kir2.1 F147L, Kir2.1 wild-type and Kir2.2 L148F. Channel openings are represented by pulses of inward current seen as downward deflections from the baseline (solid line). Vm (rnV)

-200 -150 -100 -50 I__

< B -o 3 o,

a 2 t-4 3 u .1 5

Figure 4.13.

Current-voltage relationship fo r Kir wild-type and mutant channels. Plot of single channel current amplitude versus membrane potential (Vm) for Kir2.2 wild-type {solid red circles), Kir 2.1 wild-type (solid blue squares), K|r2.2 L148F (open red circles ) and Kir 2.1 F147L (open blue squares ). The data are fit with a linear regression of the form i = y (V - £ rev) where Erey is the predicted reversal potential, constrained to 9mV. The symbols represent the mean ± SEM, where larger than the symbol, for n = 3 cells.

168 4.4. Discussion

The results presented in this chapter clearly show that block of Kir2.2 currents, expressed in CHO cells, by Ba2+ occurs in a concentration-, time- and voltage-dependent manner. Standen and Stanfield (1978a) were among the first to report the nature of Ba2+ block in inwardly rectifying potassium channels. They predicted that Ba2+ blocks the native inward rectifier in frog skeletal muscle by binding to a site within the membrane, fitting their results by assuming a one-to-one binding of Ba2+ to the channel. Recently, X-ray crystallography studies of the KcsA potassium channel have confirmed that Ba2+ binds at a single location just below the selectivity filter, around T141 in Kir2.1 (Jiang & MacKinnon, 2000).

For Kir2.2, data obtained at the single channel level supported observations obtained at the macroscopic level. As expected for a channel blocker, application of extracellular Ba2+ led to a decrease in the mean open time. A similar effect in Kirl.lb channels has been reported previously, although Chepilko et al. (1995) were able to carry out more extensive characterization of Ba2+ block in Kirl.lb at the single channel level, including analysis of closed times and open probability, because they obtained patches containing only one channel opening. They found that not only did Ba2+ decrease the mean open time but also increased the relative number of ‘long closures’. Together, these effects resulted in a decreased open probability of the channel. They proposed that Ba2+ reduced the open probability by introducing a new open, blocked state. However, this putative blocked state displayed similar time constants to the long closed states also seen under control conditions, and therefore could not be distinguished from such ‘intrinsic’ closed states. Subsequently, Choe et al. (1998) have proposed that the long-lived closures seen under control conditions were in fact a result of block by contaminating levels of divalent cations.

Analysis of the steady-state block of Kjr2.2 and Kir2.1 by Ba2+ confirmed that Kir2.2 exhibits a higher affinity for Ba2+ than Kir2.1. Kjr2.2 also exhibited a higher unitary conductance than Kir2.1 when recorded under similar conditions (see also Takahashi et al., 1995). The P- regions of Kir2.2 and Kir2.1 are virtually identical with the exception that leucine 148 in Kir2.2 is replaced by phenylalanine in Kir2.1. In this study, it was questioned whether this single difference could account for the differences in Ba2+ block and unitary conductance exhibited by these two channels.

Mutation L148F in Kir2.2 resulted in an increase in the time constant for block by Ba2+ (Tbi0Ck) and a reduction in the association rate constant, giving values that closely resembled those

169 seen in Kir2.1, without affecting the affinity for Ba2+. However, the reverse mutation F147L in Kir2.1 had no detectable effect on either Tbiock or the association rate constant, yet increased channel affinity for Ba2+ towards that found in Kir2.2.

The effect of mutation L148F in Kir2.2 can be accounted for if the binding of Ba2+ is unaffected but the rate of movement between the external solution and its binding site is slowed by the increment of an energy barrier to ion movement. In contrast, the mutation F147L in Kir2.1 appears to increase channel affinity for Ba2+ by reducing the off rate. The dissociation rate constants (&_b) could not be accurately predicted from analysis of the kinetics of Ba2+ blockage. However, the association rate constant for Ba2+ block in Kir2.1 F147L was similar to that in Kir2.1 wild-type suggesting that F147L increases the affinity for Ba2+ by reducing the rate at which Ba2+ dissociates from the channel. This would be consistent with a localized change at the binding site.

Using the blocking rate constant at OmV, £+b(0), and the Eyring rate theory (Hille, 2001), the energy barrier for the binding of Ba 2+ at OmV can be estimated, as previously described by Alagem et al. (2001), and used to illustrate the effect of mutations L148F in Kir2.2 and F147L in Kjr2.1. Figure 4.14 shows energy profiles for Ba2+ passing through Kir2.2/Kir2.1 wild-type and mutant channels. The blocking rate k+b is related to the difference between the Gibbs free energy of the state in which the ion enters the channel (which is set to be 0 at 1M concentration), and the barrier height, AG+b, and can be described by the following equation:

K i = vexp[~^r) (Equation 4.12)

S. J where v is the oscillation frequency, and R and T have their usual meanings. In a previous study of Ba2+ blockade of Kir2.1, a value of 109 s'1 was assigned to v owing to its prior use in describing the Mg2+ block of Ca2+ and NMDA channels (see Alagem et al., 2001). The same value has been used herein.

170 The depth of the energy well, A G° was calculated for OmV using:

A G° = RT In K d (Equation 4.13) where K& is the equilibrium dissociation constant for Ba2+ binding at OmV obtained from analysis of steady-state fractional currents. The barrier-well energy profile clearly illustrates how mutation L148F in Kir2.2 raises the energy barrier for Ba2+ entry, whilst the mutation F147L in Kir2.1 has no effect on the energy barrier but appears to stabilize the binding site, since the AG° value is more negative in this mutant channel. Since Ba2+ binds at a site internal to the selectivity filter, around the position of T141 in Kir2.1 (Jiang & MacKinnon, 2000), it is unlikely that F147L contributes directly to the binding site. Thus, the effect of substituting leucine for phenylalanine at position 147 in Kir2.1 may be mediated allosterically, by inducing a conformational change that interferes with the binding of Ba . 2+

The F147L mutation in Kir2.1 also reduced the voltage dependence of the steady-state Ba2+ block compared with Kir2.1 wild-type channels. There are two possible explanations for this. Firstly, mutation of phenylalanine to leucine affects the channel structure in such a way that the Ba2+ ion experiences a different electric field. However, this seems unlikely because the voltage dependence of the association and dissociation rate constants were similar for Kir2.1 wild-type and Kir2.1 F147L. Alternatively, this mutation may allow Ba2+ permeation. Indeed, a slight monotonic increase in k_b with increasing hyperpolarization can be seen for Kjr2.1 F147L, suggesting that Ba2+ can permeate the Kir2.1 F147L channel at the voltages tested (figure 4.1 IB).

The different effects of mutating these residues on Ba2+ blockage suggest the involvement of other residues aside from L148 in Kir2.2 and F147 in Kir2.1 in regulating channel block by Ba2+. Analysis of the single channel current amplitude of Kir2.2/Kir2.1 wild-type and mutants also showed that this residue does not affect unitary conductance. Thus, it appears that other residues aside from F147/L148 also determine the single channel conductance in Kjr2.2 and Kir2.1. The lack of effect of mutating leucine to phenylalanine in Kir2.2, and of the reverse mutation in Kir2.1, on single channel conductance suggests that this residue may, like the analogous residue in KcsA (Doyle et al., 1998b), orient its side chain away from the pore (figure 4.15). However, this finding contrasts with results from scanning cysteine mutagenesis studies, which have shown that the side chain of F147 in Kjr2.1 is accessible to

172 Figure 4.15

Schematic representation of the tetrameric structure of K ir2 . 2 viewed from the extracellular side. Model generated by comparative modelling using the crystal structure of KcsA (Doyle et a l 1998b) as a template and a sequence alignment based on mutational analysis of Kjr2.1 (Minor et a l 1999). The side chains of LI 48 are coloured yellow and can be seen oriented away from the pore. The pink sphere represents a K+ ion. Restraints were applied with MODELLER (Sali & Blundell, 1993) to represent (i) four-fold symmetry and (ii) the salt bridge between glutamate 139 and arginine 149 in adjacent sub-units (Yang et a l 1997).

173 block by extracellular Ag+, suggesting that the side chain of F147 is exposed to the pore lumen (Dart et al., 1998b; Kubo et al., 1998). There are, however, several drawbacks with the cysteine scanning mutagenesis approach to identifying pore-lining residues. For example, it is possible that any mutation may uncover a hidden reactivity at other residues. Furthermore, non-cysteine side chains may contribute to the coordination of Ag+ ions (see Holmgren et al., 1998). Due to the limitations of this technique it is only possible to conclude tentatively that residues that induce Ag+ sensitivity, when substituted by cysteine, are likely to interact directly with Ag+ and to face the pore.

Previous studies have shown that residues in several regions affect block of Kn channels by Ba2+. These include residues at the outer mouth of the pore in the Ml-P-region linker (see for example Navaratnam et al., 1995; Alagem et al., 2001; Murata et al., 2002), in the P-region (see for example Zhou et al., 1996; Sabirov et al., 1997; Lancaster et al., 2000; Alagem et al., 2001) and in the M2 transmembrane domain (see Thompson et al., 2000a). Residues in the Ml-P-region and P-region-M2 extracellular linkers have also been shown to determine unitary conductance in Kjr channels (Navaratnam et al., 1995; Repunte et al., 1999; Choe et al., 2000), though residues in the P-region, N- and C-termini and the M2 transmembrane domains also affect unitary conductance (see for example Taglialatela et al., 1994; Yang et al., 1995a; Choe et al., 2000; Kubo & Murata, 2001).

What region of the channel is likely to determine the differences in single channel conductance and Ba2+ blockage in Kir2.2 and Kjr2.1? Comparison of the amino acid sequences of Kir2.2 and Kir2.1 reveals several differences in the Ml-P-region extracellular linker. Thus, in Kjr2.2 the start of the pore helix is glycine-phenylalanine-methionine (G128- F129-M130), whilst in Kir2.1 it is serine-phenylalanine-threonine (S127-F128-T129). The Ml-P-region extracellular linker in Kir2.2 also contains leucine, glutamine and histidine at positions 124, 125 and 127, respectively, which are replaced by serine, glutamate and asparagine in Kir2.1. The differences in the side-chain polarity of the residues in this region might be expected to influence ion conduction, though these residues are far from the channel entrance. However, Repunte et al. (1999) have suggested that residues in this region determine unitary conductance in Kir6.1 and Kir6.2 by modifying the conformation of the permeation pathway via interaction with an arginine, which is highly conserved in Kjr channels and is crucial in stabilizing the ion conduction pore (see Yang et al., 1997). This may also be the case in Kjr2.2 and Kjr2.1.

174 Further support that residues in this region affect the ion conduction pore is provided by the difference in unitary conductance exhibited by chick and mouse homologues of Kir2.1. The chick homologue (cIRKl), which contains a non-charged glutamine at position 125 (Q125) in the Ml-P-region extracellular linker, displays a reduced single channel conductance and a reduced sensitivity to block by Ba2+ compared with the mouse homologue (IRK1), which contains a negatively charged glutamate (E125) at the corresponding position (Navaratnam et al., 1995). Substitution of a glutamate for the glutamine in cIRKl increased both the single channel conductance and sensitivity to Ba2+. Others have confirmed that the reverse mutation, E125Q, reduces channel affinity for Ba2+ and unitary conductance (Alagem et al., 2001; Murata et al., 2002). Given the above findings, it is perhaps surprising that Kir2.2, which displays the higher unitary conductance and a higher affinity for Ba2+ contains a non­ charged glutamine at this position, whilst Kir2.1, which has the lower unitary conductance and 2 i is less sensitive to block by Ba , contains a negatively charged glutamate. However, it is likely that other residues in this region, which clearly differ as highlighted above, are also involved in determining the permeation properties of the ion conduction pore in Kir2.2 and Kir2.1.

Further differences in the amino acid sequences of Kir2.2 and Kir2.1 are also apparent in the M2 transmembrane domains. Thompson et al. (2000a) have found that a serine in M2 regulates the ability of Ba2+ to bind. Thus, the mutation S165L in Kir2.1 lowered channel >y, affinity for Ba . This mutation has also been shown to reduce the affinity for channel blockage by Cs+ and abolishes Rb+ blockage. It has been suggested that serine 165 forms either a binding site for these ions or provides support for such sites (see Thompson et al., 2000b; see also Fujiwara & Kubo, 2002).

Residues in the M2 transmembrane domain have also been found to affect unitary conductance. In Kir4.1 and Kjrl . 1, the residues analogous to D172 in Kir2.1, have been shown to regulate unitary conductance (Xu et al., 2000b). In Ku4.1, which displays a single channel conductance of 24pS, this residue is a glutamate (E158). In contrast, in Kirl.l, which contains an asparagine (N171) at the analogous position, the single channel conductance is 38pS under similar recording conditions. Mutation of glutamate to asparagine (E158N) in Kh4.1 increased the single channel conductance to 35pS, whilst the reverse mutation in Kirl.l reduced the single channel conductance to 27pS (Xu et al., 2000b).

175 Although the serine and aspartate are conserved in Kjr2.2 and Kir2.1, these channels differ at a neighbouring position. Thus, in Kir2.2 the residue at position 164 is an alanine, whilst the analogous residue in Kjr2. lisa phenylalanine. The residues at the analogous positions in Kir2.3 and Kir2.4, which display a lower unitary conductance (~15pS under similar recording conditions) and a lower affinity for Ba2+ (see for example Morishige et al., 1994; Topert et al., 1998; Liu et al., 2001b), are also different. Thus, in Kir2.3 this residue is a valine, whilst in Kir2.4 it is a leucine. As suggested for S165, this residue may also support the binding site for Ba2+ (see Thompson et al., 2000b) and may determine the differences in unitary conductance exhibited by the members of the Kir2.0 subfamily.

Thus, residues in both the Ml-P-region extracellular linker and the M2 transmembrane domains may be important in regulating unitary conductance and Ba2+ blockage in Kir2.2 and Kir2.1. However, without further experimentation it is impossible to conclude which of these residues might account for these differences.

In conclusion, the residue at position 148 in Kir2.2, or at position 147 in Kir2.1, does not account for the difference in unitary conductance exhibited by these two channels. However, it does account for part, but part only, of the difference in channel block by Ba2+. Thus, in Kir2.2 and Kir2.1, residues in regions other than the P-region must determine the differences in unitary conductance and Ba2+ blockage. This is consistent with previous studies in which residues in the Ml-P-region and P-region-M2 extracellular linkers have been shown to be important in determining the properties of the ion conduction pore (see for example Navaratnam et al., 1995; Repunte et al., 1999).

176 Chapter 5

Mutation of Isoleucine 143 affects ionic selectivity and permeation in Kir2.1.

5.1. Introduction

Although Kv channels differ from Kjr channels in their overall primary structure, both contain a ‘P- region’ in which the well-conserved signature sequence resides. As discussed in section 1.6, this region is important for K+ selectivity.

X-ray crystallography studies have shown that in the bacterial potassium channel KcsA, the selectivity filter is formed by backbone carbonyl oxygen atoms from five (T75, V76, G77, Y78 and G79, see table 5.1) amino acids including the two, conserved glycine residues of the GYG motif (Doyle et al., 1998b). With a two-transmembrane-helix topology, KcsA is similar in its overall structure to Kjr channels, yet shares little homology at the amino acid level. In

177 contrast, the amino acid sequence that forms the pore is well conserved between Kv channels and KcsA. Despite the differences in the amino acid sequence of the pore between KcsA and Kir channels, a recent study has shown that the pore of KcsA may be substituted for that of Kir2.1, the resulting chimera retaining the functional hallmarks of an inward rectifier. Likewise, the pore of KcsA could be substituted for that of Shaker such that the resulting chimera retained the functional hallmarks of a voltage-gated potassium channel (Lu et al., 2001b). These results suggest that the ion conduction pore is conserved among potassium channels.

Previous studies with the peptide toxin blockers Lq2 and agitoxin2 have also suggested that the three dimensional structures of potassium channel pores are similar despite the lack of conservation in the P-region sequences. Thus, Lq2 has been shown to inhibit a variety of potassium channels including some members of the Kv, Kca and Kir channel families (see for example Lu & MacKinnon, 1997), whilst agitoxin2 inhibits both eukaryotic Kv channels and the prokaryotic potassium channel KcsA (MacKinnon et al., 1998). Renisio et al. (1999) have suggested that the structure of the external end of the pore serves as a frame to hold the conserved signature sequence, allowing it to form the selectivity filter.

In this chapter, I present results from experiments in which I have investigated the role of an isoleucine at position 143 of Kir2.1 (see Table 5.1) in ionic selectivity. This residue, which resides in the potassium channel signature sequence, corresponds to one of the five residues (V76) implicated in stabilizing the backbone of oxygen rings in the selectivity filter of KcsA (Doyle et al., 1998b). Mutagenesis studies in Kv2.1 have shown the analogous residue (V374) to be an important determinant of K+ and Rb+ selectivity (Taglialatela et al., 1993). If the molecular basis of selectivity is conserved among all potassium channels, the effects of mutating 1143 in Kir2.1 on ionic selectivity should be comparable to the previously reported effects of mutating V374 in Kv2.1 (Taglialatela et al., 1993).

Ionic selectivity was determined by measuring permeability ratios for Rb+ and K+. Whilst Rb+ permeates Kv channels it blocks Kjr channels including the native Kir channels of frog skeletal muscle and starfish eggs (see for example Adrian, 1964; Hagiwara & Takahashi, 1974). This block occurs in a steeply voltage-dependent manner (Standen & Stanfield, 1980), increasing initially with hyperpolarization of the membrane then falling with further hyperpolarization. Although Rb+blocks Kjr channels it can still be substituted as a test ion

178 1 2 3 4 5 6 7 8

KcsA 64 R A LWWS V E TATT V GYG DL Y ....82

Kir2.1 131... AAFLF s I E T Q T T I G YG FR C ...149

ShA 431 ... D AFWWAVV TMTT V GY G D M T ...449

Kv2.1 362 ... A SFWWAT I TMT T V GYG DIY ...380

Kv3.1 389 ... I GFWWAV V T MT T L GYG DMY ...407

Table 5.1.

Alignment of the P-region sequences o f Kir2. /, KcsA and several Kv channels. Residues of the signature sequence are numbered and are marked in bold. Iso leucine 143 in Kir2.1 and its analogous residues in the other channels are highlighted in red. The single amino acid code is given in table 0.1. when measuring the relative permeability of the channel since block of a channel does not alter its reversal potential (Hille, 2001).

5.2. Methods

Whole cell currents were recorded with 140mM K+ in the patch pipette and 70mM K+ or Rb+ as the sole external permeant cationic species. Permeability ratios (P ri/E k) for each cell were calculated using a modified Goldman-Hodgkin-Katz equation (see Hille, 2001):

E -E -RT]nPM AF rev.R b ^ p ^ (Equation 5.l) which can be rearranged to give

(Equation 5.2)

where AFrev is the shift in reversal potential, Frev,Rb and Frev, k are the reversal potentials for Rb+ and K+, respectively, and F, R and T have their usual meaning. The reversal potential was measured by fitting a fifth order polynomial function of the form:

f ( x ) = a + bx + cx2 + dx3 + ex4 + f x 5 (Equation 5.3)

to a region, approximately 15-25 mV either side of where the current intersected the abscissa, of the I-V relationship.

5.3. Results

5.3.1. Ionic selectivity of K ^ .l wild-type

Figure 5.1 A shows representative whole-cell recordings from CHO cells expressing Kir2.1 bathed in equimolar K+ and Rb+ containing solutions. As expected, Kir2.1 exhibited substantial inward currents in 70mM K+. As shown previously, Rb+ blocked Kir2.1 currents (see for example Abrams et al., 1996; Reuveny et al., 1996).

The corresponding mean, steady-state current-voltage ( I-V) relationships are shown in figure 5.IB, and in figure 5.1C are expanded for clarity. When 70mM K+ is replaced with equimolar

180 Figure 5.1.

Ionic selectivity of Kir2.1. A, Representative current recordings of Kjr2.1 wild-type with 70mM [K+]0 or 70mM [Rb+]0. Currents were elicited in response to 50ms voltage steps from a holding potential of -17mV, to test potentials ranging from +63mV to -132mV, in 5mV decrements. Every fifth record is shown for clarity. B, normalized, steady-state current- voltage relationships. The symbols represent the mean ± SEM, where larger than the symbol, for n = 14 cells. Closed circles, 70mM [K+]0; open triangles, 70mM [Rb+]0. C, expansion of current-voltage relationships. Symbols as in B. A. 70K 70Rb+

InA

10ms

B . Vm (mV)

-150 -100 -50

- -0.5

ss

- - 1.0

c. - lOt):

-40 -30 -20

ss

- -0.03

181 Rb+, Erev shifts -17.6 ±1.3 mV (n = 14; mean ± SEM), giving P ri/P k = 0.51 ± 0.02. Although this permeability ratio differs somewhat from previously reported values (0.68, Abrams et a l, 1996; Reuveny et al., 1996), similar values have also been reported (0.49, Choe et a l, 2000; 0.57, Thompson et a l, 2000b).

5.3.2. Ionic selectivity of 1143 mutants

Next, the contribution of isoleucine 143 (1143) to ionic selectivity was investigated by mutation. Figures 5.2A, B and C illustrate representative current traces for mutants in which isoleucine 143 was replaced by valine, leucine or threonine, respectively. The corresponding I-V relationships, which are expanded for clarity, are shown below. The values for the shift in reversal potential in I143V, I143L and I143T following substitution of 70mM Rb+0 for 70mM K+0 are given in Table 5.2. Substitution of isoleucine by leucine (I143L) or valine (1143V) raised the relative Rb+ permeabilities, although the effect was less dramatic for I143V ( P r i/ P k

= 0.62 ± 0.03, n = 9) than for I143L ( P r i / P k = 0.82 ± 0.01, n = 12). In contrast, substitution by the polar residue threonine (I143T) enhanced K+ selectivity ( P r i / P k = 0.44 ± 0.01, n = 11).

5.3.3.1143C and I143S mutants are non-functional

On mutation of isoleucine 143 to cysteine, the mutant channels failed to yield currents (n = 13, not shown). However, I143C mutant channels could be rescued on bathing cells in an external solution containing lOmM dithiothreitol (DTT; figure 5.3A). When 70mM K+ was replaced with equimolar Rb+, Erev shifted -18.9 ± 0.07 mV (n = 9; mean ± SEM; see for example figure 5.3B) giving P r i / P k = 0.52 ± 0.07, similar to that for Kir2.1 wild-type.

The rescue of I143C mutant channels with the reducing agent DTT suggests that a disulphide bond may have formed between the thiol groups of the cysteine introduced at position 143 and another cysteine in either the same or adjacent subunits, or between two of the newly introduced cysteines in adjacent subunits. In the absence of any reducing agent this could render the channel non-conducting. This has been observed in Kv2.1 channels following the substitution of cysteine for isoleucine 379 (Zhang et a l, 1996), which is located near the outer end of the narrow ion conduction pore (the analogous residue to 1379 in Kir2.1 is R148). The Kv2.1 mutant I379C displayed varying levels of current when expressed in Xenopus oocytes, which could be increased several fold after application of ImM DTT. Furthermore, exposure to 0.1% H20 2 significantly reduced currents through Kv2.1 1379C mutant channels.

182 Figure 5.2.

Ionic selectivity of Kir2.11143V, I143L and I143T., B Aand C show representative current recordings with 70mM [K+]0 or 70mM [Rb+]0 from Kir2.1 1143V, I143L and I143T, respectively. Currents were elicited in response to 50ms voltage steps from a holding potential of-17mV, to test potentials ranging from +63mV to -132mV, in 5mV decrements. Every fifth record is shown for clarity. D, E and F show normalized, steady-state current- voltage relationships for Kir2.1 I143V, I143L and I143T, respectively. The symbols represent the mean ± SEM, where larger than the symbol, for n = 9 - 12 cells. Closed circles, 70mM [K+]0; open triangles, 70mM [Rb+]c. A. 70K 70RbH B. 70K 70Rb+ C. 70K 70Rb+ —«■

2nA 2nA 2nA ___1 I 10mss 10ms 10ms

D. 0.03 0.03

(mV)

e s f ^ p | I | U~] I I P-0(j) I | I | I | I | -40-30-20/-10 0 10 20 30 40 50 60 -40:3$-20/-10 0 10 20 30 40 50 60 -40 -30-20/-10 0 10 20 30 40 50 60

ss

- -0.03 - -0.03 - -0.03 Mutagenesis studies and western blot analysis suggested that the cysteine introduced at position 379 could form an inter-subunit disulphide bridge with either a native cysteine at position 394 in the S6 transmembrane domain or with another cysteine 379 in an adjacent subunit (Zhang et al., 1996).

In this study, it was questioned whether a disulphide bond may have formed between the thiol group of the introduced cysteine at position 143 and another cysteine. If the formation of a disulphide bond between C143 and another cysteine underlies the loss of function in I143C mutant channels, mutation of 1143 to a residue with a similar size to cysteine, but without the sulphur moiety, might be expected to give functional channels. To test this hypothesis, isoleucine 143 was mutated to serine. Surprisingly, substitution of serine for isoleucine 143 also generated mutant channels, which failed to yield currents under control conditions (n = 13, data not shown), but as for I143C mutant channels could be rescued by bathing cells in external DTT (Figure 5.3C). These results suggest that the loss of function in 1143 mutant channels is related to the size of the residue substituted at position 143 (see section 5.3.7).

When 70mM K + was replaced with equimolar Rb+, Erev shifted -19.1 ± 3.3 mV (n = 9; mean

± SEM; see for example figure 5.3D), giving P r i / P k = 0.50 ± 0.06, again similar to that for

K ir2 .1 wild-type.

5.3.4. Correlation between side chain properties and Rb+/K+ selectivity

In Kv2.1, a correlation between the hydrophilicity of the side chain of the residue at position 374 and permeability has been shown (Taglialatela et al., 1993). Hydrophilicity is an index of solvent accessibility and is calculated as the equilibrium constant for partition of amino acid side chains between vapour and aqueous phases (Wolfenden et a l, 1979). Whilst substitution of residues with hydrophobic side chains such as isoleucine and valine favoured Rb+, substitution of residues with hydrophilic side-chains such as serine and threonine favoured K+. Thus, the interaction between pore water and the side chains introduced at position 374 in Kv2.1 channels were proposed to determine K+/Rb+ selectivity

A similar correlation between the hydrophilicity of side-chains at position 143 and K+/Rb+ selectivity was sought in Kir channels. Figure 5.4 shows a plot of side-chain hydrophilicity versus Rb+/K+permeability ratios. Data for Kv2.1, taken from Taglialatela et al. (1993), is included in the graph for comparison. As for Kv2.1, linear fits to all the data with the exception of leucine also show a statistically significant (P < 0.05) correlation between side-

184 Figure 5.3.

Ionic selectivity o f Kir2 .1 1143S and I143C. A and C show representative current recordings with 70mM [K+]0 or 70mM [Rb+]0 in the presence of lOmM extracellular DTT from Kir2.1 I143S and I143C, respectively. Currents were elicited in response to 50ms voltage steps from a holding potential of -17mV, to test potentials ranging from +63mV to -132mV, in 5mV decrements. Every fifth record is shown for clarity. B and D show representative membrane currents recorded in response to voltage ramps ranging from -60 to +20mV from Kjr2.1 I143S and I143C, respectively. Blue plots, 70mM [K+]0; Red plots, 70mM [Rb+]0. c + + 70Rb+ 70K 70Rb

500pA 10ms

i- 100 100

-40 -30 0

4 (p A )

-100 L -100 0 .9

0.5

0 .4

0.3

0.2

•2 -1 0 1 2 3 4 5 6 7 8 9 10

Hydrophilicity (kcal/M)

Figure 5.4.

Correlation between P ri/P k and hydrophilicity. Plot of P r JP k against hydrophilicty of residues at position 143 in Kir2 .1 . Data is compared with previous findings for mutants of the corresponding residue, valine 3 7 4 , in K v2.1 (data taken from

Taglialatela et al., 1994). The symbols represent the mean ± SEM, where larger than the symbol, for n = 9 - 14 cells. Red circles, Kjr2 .1 ; Blue squares, K v2 .1 . For each channel, the straight line represents the linear regression fit to all the data with the exception of leucine. The curved line maps the 9 5 % confidence limits.

Hydrophilicity values were calculated as described by Wolfenden et a l, 1979.

186 chain hydrophilicity and permeability in Kir2.1. One notable difference between Kv and Kjr channels is the permeability ratio obtained following substitution of the non-polar residue leucine. Whilst in Kv channels substitution of a leucine gives a Rb+/K+ permeability ratio more like that expected from substitution of a polar residue, reflecting an enhancement in selectivity for K+, in Kir2.1 the Rb+/K+ permeability ratio reflects an enhancement in selectivity for Rb+. Nonetheless, the general correlation between side chain hydrophilicity of the residue at position 5 of the signature sequence (V374 in Kv2.1 and 1143 in Kir2.1) and selectivity for K+ and Rb+ is conserved between Kv2.1 and Kir2.1.

5.3.5. Mutation of 1143 does not alter Rb+ blockage

As mentioned in section 5.1, Rb+ blocks Kir channels. This property depends on residues located in the M2 transmembrane domains, which lie beyond the selectivity filter (Thompson et al., 2000b). In this study it was questioned whether mutations of 1143 alter Rb+ blockage, since Rb+ must traverse the selectivity filter to reach its blocking site. In all 1143 mutants, as for the wild-type channel, the current carried by Rb+ (/Rb) was smaller than that carried by K+

(7 k ). For example, in Kir2.1 wild-type Ir\JIk = 0.058 ± 0.006 at -132mV. As shown previously for mutations in GYG (So et al., 2001), Rb+ blockage was little altered by mutations of 1143 (see Table 5.2).

5.3.6. Mutations of 1143 have little effect on channel macroscopic kinetics.

As described in chapter 1, gating in Kir channels involves the displacement of polyamine molecules from the channel pore by K+. Since permeant K+ activates Kir channels, it was questioned whether mutations of 1143 affect such channel gating. Macroscopic current kinetics were investigated by measuring the relationships between voltage and both K+ chord conductance and rates of activation under hyperpolarization. For each mutant channel, the chord conductance, gx, was calculated from the I-V relationships using:

g K = -~Y ' (Equation 5.4) V ~ E k where gK is the chord conductance to K+ and V - EK is the driving force on K+. gK was normalized to its maximum value and the normalized conductance, gx', was plotted against voltage.

187 Table 5.2. Properties of Isoleucine mutants

H.I.+ Shift in Erev Pri/P k Iri/Ik V! ki v 2 k2 Dependence Tact at -52mV of ’tact On Vm (kcal (mV) (-132mV) (mV) (mV) (mV) (mV) (ms) /M) (e-fold)

1143 0.24 -17.6±1.3(14) 0.5110.02(14) 0.05810.006(14) -25.211.5(6) 7.210.6(6) 2.011.0(6) 10.112.3(6) 22.813.4(8) 0.5410.07(8)

I143L 0.11 -4.9510.4(12)*** 0.8210.01(12)*** 0.07510.01(12) -20.411.0(9) 6.310.7(9) -8.713.5(9) 17.513.0(9) 25.013.1(10) 0.4510.04(10)

I143V 0.4 -12.411.1(9) 0.6210.03(9) 0.03510.003(9) -26.415.8(8) 11.711.8(8) 4.1313.9(8) 9.411.4(8) 36.014.2(8) 0.5310.04(8)

I143C 3.63 -18.910.1(9) 0.5210.07(9) 0.09710.02(9) -22.811.0(9) 6.810.8(9) 1.013.5(9) 16.013.6(9) 20.512.0(9) 1.2610.16(9)* I143T 7.27 -21.110.6(11) 0.4410.01(11) 0.1110.02(11) -19.611.2(5) 10.111.8(5) -3.018.8(5) 10.313.4(5) 36.413.2(10) 0.7810.06(10) I143S 7.45 -19.113.3(9) 0.5010.06(9) 0.07810.02(8) -28.912.0(5) 6.911.4(5) -10.814.2(5) 18.815.3(5) 25.711.9(9) 1.7310.23(9)** tH.I. = Hydrophilicity Index. Average values are given as means ± SEM with the number of experiments in parentheses. Results were compared using a Kruskal Wallis non-parametric ANOVA test followed by a Dunn’s multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001 The relationship was best fitted with a double Boltzmann expression of the form:

(i - b ) + (Equation 5.5) (v-v2) 1 + exp V - V i 1 + exp k\ k2 where b was allowed to vary between 0 and 1. V\ and V2 (mV) give the voltages at which the relative conductance gK = 0.5 and k] and &2 (mV) are the factors effecting the steepness of the relationship. The values found for the half-activation voltages and steepness factors are given in Table 5.2, from which it can be seen that mutation of 1143 to valine, leucine or threonine has little effect on the relationship between chord conductance and membrane potential. The relationship between chord conductance and membrane potential was also similar for I143C and I143S mutants following rescue of channel function with lOmM DTT.

Hyperpolarization from Ek of wild-type and mutant channels produced inward currents that increased with time (see figures 5.1, 5.2 and 5.3), a process that also became more rapid with increasing hyperpolarization. At -52mV, the time constant for activation (Tact) was 0.54 ±

0.07 ms (n = 8) in Kir2.1 wild-type and the rate of activation increased e-fold for a 22.8 ± 3.4 mV hyperpolarization. Little change in the voltage-dependence of activation was seen in any of the mutants (Table 5.2), though the time constant for activation at -52mV was significantly slower (P < 0.05) for currents through Kir2.11143C (xact at -52m V = 1.26 ± 0.16 ms, n = 9) and Kir2.11143S channels (xact at -52mV = 1.73 ± 0.23 ms, n = 9) in the presence of lOmM extracellular DTT. Since the time constant for activation is also slower in wild-type channels with lOmM DTT in the extracellular solution (e.g. in Kjr2.2 xact= 0.31 ± 0.02 ms in 70mM

K+0, n = 7; xact = 0.60 ± 0.1 ms in 70mM K+0 + lOmM DTT, n = 4; P < 0.05), the increase in

Tact in I143C and I143S was not investigated further.

5.3,7. Hypotheses for the mechanism underlying loss of function in I143C and I143S In section 5.3.3, it was shown that on mutating isoleucine 143 to cysteine, channel function was lost. However, by bathing cells in extracellular solution containing the reducing agent dithiothreitol, channel function could be restored. It was initially hypothesized that the introduced cysteine might form a disulphide bond with either a native cysteine or with another C143 in an adjacent subunit, which in the absence of a reducing agent renders the channel non-conducting. However, mutation of isoleucine to serine also generated non-functional

189 channels, which could be rescued with lOmM extracellular DTT. These results suggest that the loss of function is related to the size of the residue at position 143.

To explain this loss of function, it was hypothesized that mutation of isoleucine to smaller residues such as cysteine or serine alters the conformation of the pore, which, in turn, prevents K+ permeation. These changes in conformation might be compared with those associated with C-type inactivation in Kv channels, where the selectivity filter region of the pore is thought to undergo a constriction that pinches off the permeation pathway (see section 1.2.3a). How might a conformational change in the pore structure lead to a loss of conduction in I143C and I143S mutant channels that can be restored following exposure to extracellular DTT?

One possibility is that a conformational change in the pore allows a native cysteine to form a disulphide bond with another residue in an adjacent subunit, which locks the channel in a non­ conducting state. Kir2.1 possesses several extracellular cysteine residues located at positions 122, 149 and 154. The cysteine at position 149 is located sufficiently close to the ion conduction pathway that the binding of only one Ag+ ion to one of four potential binding sites is sufficient to block channel current (Dart et al., 1998a). Previous studies in Shaker have shown that residues in this region move closer together during C-type inactivation. For example, cysteine substitution experiments have shown that a cysteine substituted for a threonine at position 449, which is analogous to C149 in Kir2.1, is more sensitive to block by cadmium in the C-type inactivated state, suggesting that the spatial distribution of each of the four residues at position 449 changes during C-type inactivation (Yellen et al., 1994). Furthermore, substitution of cysteine for a methionine (position 448, equivalent to R148) located at the outer mouth of the Shaker channel pore resulted in the formation of disulphide bonds between subunits in the C-type inactivated state but not in the closed state (Liu et al., 1996). Thus, the formation of an inter-subunit disulphide bond between the cysteines at position 149 represents a plausible explanation for the loss of function in I143C and I143S mutant channels.

An alternative possibility is that a disulphide bond already exists in the wild-type channel, which is required for normal function or for normal folding. Indeed, the two cysteines located at positions 122 and 154, which are absolutely conserved throughout the inward rectifier potassium channel family, form an intra-subunit disulphide bond that is essential for correct channel assembly (Leyland et al., 1999; Bannister et al., 1999; Cho et al., 2000). Mutation of

190 these cysteines has been shown to abolish functional expression, even though mutant channels were expressed in the membrane. However, once assembled the disulphide bond is not essential for channel function since application of an extracellular reducing agent had no effect on normal channel function (Leyland et al., 1999; Bannister et al., 1999; Cho et al., 2000). It has been suggested that the disulphide bond helps to constrain and stabilize the P- loop and selectivity filter (Cho et al., 2000). Thus, an equally plausible explanation for the loss of function is that substitution of cysteine or serine for isoleucine alters the structure of the P-region so that the intra-subunit disulphide bond between C122 and C154, which is required for correct assembly, now disrupts channel function. In either case, the disulphide bond would require breaking to allow K+ permeance. These hypotheses were tested as described below.

5.3.8. Mutation of 1143 to Alanine gives functional channels with a reduced conductance

First, to test whether substitution of other small residues at position 143 in Kir2.1 result in a loss of channel function, isoleucine was mutated to alanine. Although mutation of isoleucine 143 to alanine was not fatal (figure 5.5A), currents were considerably smaller compared with those measured from other mutant channels under similar recording conditions. 1143A currents were very unstable, even in the presence of lOmM DTT in the extracellular solution. However, a substantially bigger current was seen in one cell following exposure to lOmM DTT (figure 5.5B).

5.3.9.1143C C149S and I143S C149S double mutants do not form functional channels

Next, the effects of mutating C149 to serine, in the presence of either a cysteine or serine at position 143, were tested. Expression of the double mutants I143C C149S and I143S C149S in CHO cells resulted in no detectable currents (n = 14 and 11, respectively; figure 5.6A, B). The loss of function in I143C C149S and I143S C149S mutant channels is not consistent with the hypothesis that an inter-subunit disulphide bond between the cysteines at position 149 in adjacent subunits underlies the loss of conduction in I143C and I143S mutant channels, since mutating C149 to serine would disrupt any such bond and would be expected to restore channel function. However, these results do not rule out the alternative hypothesis that the intra-subunit disulphide bond between C122 and C154 disrupts channel function in I143C and I143S mutant channels.

To test this hypothesis further, cells expressing the double mutant channels I143C C149S and I143S C149S were bathed in extracellular solution containing lOmM DTT. However,

191 Figure 5.5.

Expression of Kir2.1 I 143A gives functional channels. A and B, representative current recordings with 70mM [K+]0 in the absence (A) and presence ( B ) of lOmM extracellular DTT. Currents were recorded in response to 50ms voltage steps from a holding potential of -17mV, to test potentials ranging from +63mV to -132mV, in 5 mV decrements. Every fifth record is shown for clarity. C and D, steady-state current-voltage relationships in the absence (C) and presence (.D) of lOmM DTT. In C, the symbols represent the mean ± SEM for n = 8 cells. A B. + DTT

InA 10ms

InA

10ms

D.

-150 -100 -50 -150 -100 -50 •

-2 - -2 -

-4 - -4 -

-6 -

-7 - -8 J -8 J Figure 5.6.

Expression of Kir2.1 I143C CM S, I143S C149S and Kir2.1 C149S. A, B and C, Representative current recordings obtained with 70mM [K+]0 from Kir2.1 I143C C149S, I143S C149S and Kir2.1 C149S respectively. Currents were elicited in response to 50ms voltage steps from a holding potential of -17mV to test potentials ranging from +63mV to -132mV, in 5mV decrements. Every fifth record is shown for clarity. D, normalized, steady-

state, current-voltage relationship for Kir2.1 C149S. The symbols represent the mean ± SEM, where larger than the symbol, for n = 10 cells. £61

0 * 1 - - i

ss

001-

(A«i)" a

SU IQ l

y u g

. m*

tm

D

suioi

V<*OQZ currents could not be detected (n = 10, not shown). This might suggest that the conserved cysteines at position 149 are essential for expression. However, mutation of C149 to serine alone generates functional channels (n =10; figure 5.6C, D). In general, cells expressing a strong green fluorescence signal express large currents when transfected with cDNA encoding wild-type or mutant channels that are functional. Thus, it is suggested that in the double mutants I143C C149S and I143S C149S, the structure of the P-region has been altered so that even disruption of the intra-disulphide bond between C l22 and C l54 does not permit K+ permeance.

5.3.10. Evidence for an intra-subunit disulphide bond

Ideally, to determine whether the intra-subunit disulphide bond between C l22 and C l54 underlies the loss of function in I143C and I143S, site-directed mutagenesis would be used to disrupt this bond. However, this approach cannot be used because this bond is essential for correct channel assembly (see Leyland et al., 1999; Bannister et al., 1999; Cho et a l, 2000).

However, support for this hypothesis can be provided by confirming the presence of an intra­ subunit bond in these mutant channels. Thus, the expression of Kir2.1 wild-type and I143C/I143S mutants was investigated at the level of the protein, under reducing and non­ reducing conditions. The initial approach involved the use of western blotting (see section 2.5.1). However, although bands corresponding to Kir channels were detected in samples from rat cortex and cerebellum, no bands were detected in samples from CHO cells expressing Ku2.1 wild-type or mutant channels (not shown), suggesting that levels of channel expression were too low to detect using western blotting. The formation of disulphide bonds in vitro is either a cotranslational or immediate post-translational event, and can be detected using a reticulocyte lysate system (Scheel & Jacoby, 1982), which is used to detect early events. Thus, an in vitro reticulocyte lysate transcription-translation system was used in this study.

First, to determine whether Kjr channels were expressed in the in vitro translation system, a preliminary experiment was performed in which in vitro translation was carried out in the absence of microsomal membrane. For wild-type and mutant channels, the reticulocyte lysate system yielded proteins with a molecular weight around 48kDa, as expected for a single subunit, when run under non-reducing or reducing conditions on SDS polyacrylamide gels thus showing that Kir proteins are indeed expressed using this system (see Figure 5.7). The smaller bands seen in figure 5.7, which are approximately 42kDa in size, could represent

194 -p-mercaptoethanol +p-mercaptoethanol

r r I143C I143S I143C WT » « C 1143 S I143C WT C149S

48.5 ---- -Monomer

m m . m m m

36.5.

Figure 5.7.

In vitro translation o f K i r2 . 1 wild-type and mutant channels using the reticulocyte system. Autoradiograph of SDS polyacrylamide gels for wild-type (WT), Kir2.1 I143S, Kir2.1 I143C and Kjr2.1 I143S C149S. In vitro translation was carried out in the absence of microsomal membranes and in the absence or presence of p-mercaptoethanol.

195 breakdown products where translation of the protein has temporarily stopped or started at a later codon sequence than the start codon, giving a smaller product. For example, in a previous study of Kir2.0 proteins a band of 43kDa was detected for Kir2.2, although the predicted size of the protein is 48kDa. This difference in the molecular weight could be explained by translation beginning at the third in-frame methionine codon in the Kjr2.2 mRNA molecule (Stonehouse et al., 1999).

Since Kjr proteins are expressed in the in vitro translation system, the next step was to determine whether an intra-subunit disulphide bond is present when the channels are incorporated into membranes. In vitro translation was therefore carried out in the presence of microsomal membranes. The protein expression of Kjr2.1 wild-type in the presence of various concentrations of oxidised glutathione (GSSG) was investigated first (see figure 5.8) to confirm the presence of the intra-subunit disulphide bond, as previously reported (see Bannister et al., 1999; Cho et al., 2000).

Under optimal conditions GSSG would oxidize any thiols, catalyzing the formation of any disulphide bonds (see for example Scheel & Jacoby, 1982). Kir2.1 migrated as a monomer, even in the presence in the reaction mixture of ImM GSSG. In fact, higher concentrations of GSSG (ImM) appeared to inhibit protein synthesis seen as a decreased intensity of the bands in lanes 5 and 8 of figure 5.8. This effect was counterbalanced in the presence of the reducing agent p-mercaptoethanol (lane 8, figure 5.8), which would have converted some of the oxidized glutathione (GSSG) to its reduced form (GSH). Bands corresponding to oligomers were not detected. However, wild-type channels migrated faster under non-reducing conditions (- p-mercaptoethanol) than under reducing conditions (+ p-mercaptoethanol) suggesting the presence of an intra-subunit disulphide bond. In the presence of intact disulphide bonds, these proteins appear as more compact units and migrate further into the polyacrylamide gel than their reduced counterparts. This is likely to be the disulphide bond between the conserved extracellular cysteines at position 122 and 154 (Leyland et al, 1999; Bannister et al., 1999; Cho et al., 2000), which is essential for channel assembly. The difference between the migrations of Kir2.1 wild-type when expressed in microsomal membranes compared with expression in the absence of membranes may be an indication of incorrect glycosylation of the protein. It could be possible to confirm this by pretreatment of protein samples with A-glycopeptidase and O-glycopeptidase before blotting.

196 -p-mercaptoethanol +p-mercaptoethanol

______A______.A. ______

TD -M 0 0.5 1 0 0.5 1 [GSSG] mM

kDa 116 97.4 Tandem Dimer

66

Monomer

45

Figure 5.8.

In vitro translation of Kir2.1 wild-type in the presence o f membranes. Autoradiograph of SDS polyacrylamide gel for Kir2.1 wild-type in the presence of microsomal membranes. In vitro translation was carried out in the absence (lanes 3-5) or presence (lanes 4-6) of p- mercaptoethanol and with 0, 0.5 or ImM GSSG. Two control lanes were run for comparison. Lane 1 shows the migration of a tandem linked dimer (TD) and lane 2 shows migration of Kjr2.1 in the absence of microsomal membranes (-M).

197 -P-mercaptoethanol +p-mercaptoethanol

0 0.5 1 [GSSG] mM

Tandem Dimer

Monomer

Figure 5.9.

In vitro translation of Kir2.l I143S in the presence of membranes. Autoradiograph of SDS polyacrylamide gel for Kir2.1 I143S in the presence of microsomal membranes. In vitro translation was carried out in the absence (lanes 3-5) and presence (lanes 6-8) of p- mercaptoethanol and with 0, 0.5 and ImM GSSG. Two control lanes were run for comparison. Lane 1 shows the migration of a tandem linked dimer (TD) and lane 2 shows the migration of Kjj2.11143S in the absence of microsomal membranes (-M).

198 To verify that the system was capable of expressing dimers, a tandem linked dimer was translated and analyzed by gel electrophoresis. The system yielded a protein (lane 1, figure 5.8) with an approximate molecular weight of 100 kDa, close to that predicted for a dimer of Kir2.1 (96kDa). However, subunits were covalently linked by 10 glutamine residues (Dart et al., 1998a) rather than by disulphide bridges. This is reflected in the slightly higher molecular weight than predicted for dimers of K n channels.

Next, the expression of Kir2.1 I143S was investigated under the same conditions as those described above for Kir2.1 wild-type. Figure 5.9 shows these results. As for wild-type channels, in the presence of microsomal membranes I143S migrated as a monomer, even in the presence in the reaction mixture of ImM GSSG. Similarly, higher concentrations (ImM) of GSSG seemed to inhibit protein synthesis in the absence of P-mercaptoethanol (lane 5, figure 5.9). Furthermore, bands corresponding to oligomers were not detected, though a slight shift in the distance migrated is apparent between I143S channels run under non­ reducing conditions (- p-mercaptoethanol), and under reducing conditions (+ P- mercaptoethanol), again suggesting the presence of an intra-subunit disulphide bond as seen in wild-type channels.

5.4. Discussion

In this chapter, the effects of mutating an isoleucine at position 143 in Kir2.1 on ionic selectivity and macroscopic gating were investigated. In Kjr2.1, isoleucine 143 lies at position five of the signature sequence, a stretch of eight amino acids (TXXTXGY/FG; see table 5.1), which are highly conserved among all potassium channels.

Previous studies in Kv channels have reported various effects of mutating the corresponding residue on ionic selectivity. Thus, in Shaker, wild-type selectivity was retained when the valine at position 5 of the signature sequence was mutated to cysteine, leucine or threonine, whilst mutation to alanine, glycine, asparagine or glutamine rendered the channel non selective (Heginbotham et al., 1994). In contrast, in Kv2.1, functional expression was not obtained on mutation of the valine at position 5 of the signature sequence to alanine, glycine or glutamine. Neither was functional expression obtained on mutation of this valine to charged amino acids such as glutamate. However, mutation to leucine, isoleucine, threonine, cysteine and serine generated functional channels in which there appeared to be a correlation between Rb+/K+ permeability and the side chain hydrophilicity of the substituted amino acid at this position (Taglialatela et al., 1993).

199 In this study, mutation of the isoleucine at position 5 of the signature sequence to valine, leucine and threonine was tolerated, whilst mutation to cysteine and serine resulted in a loss of function. However, channel function could be rescued with lOmM extracellular DTT (see below). As for Kv2.1 channels, a similar correlation between side chain hydrophilicity and Rb+/K+ permeability was also apparent. Substitution of hydrophobic residues, such as valine, favoured Rb+ entry, whilst substitution by more hydrophilic residues, such as threonine, enhanced K+ selectivity. However, one difference between Kir2.1 and Kv2.1 was notable. In Kv2.1 substitution of a leucine for valine enhanced K+ permeability, whilst in this study substitution of leucine for isoleucine in Kir2.1 enhanced Rb+ permeability. Nonetheless, the general correlation between side chain hydrophilicity of the residue at position 5 of the signature sequence (V374 in Kv2.1 and 1143 in Kir2.1) and selectivity for K+ and Rb+ is conserved between Kv2.1 and Kir2.1.

The correlation between P r\JPk and side-chain hydrophilicity of the substituted residue at position 5 of the signature sequence suggests that in both Kv2.1 and Kir2.1, the residue at this position projects its side chain into the aqueous lumen of the pore (see Taglialatela et al., 1993; see also De Biasi et al., 1993). In support of this, cysteine-scanning mutagenesis studies have shown that when mutated to cysteine, the side chains of V374 in Kv2.1 and isoleucine 143 in Kir2.1 are accessible to thiol-labeling reagents such as MTSET or Ag+, which also suggests that the side chains of these amino acids face the pore (Pascual et al., 1995; Dart et al., 1998b). Lti & Miller (1995) have reported similar findings for the analogous residue in Shaker. Mutagenesis studies of the P-region of Kirl.l and block by toxins (Jin & Lu, 1998; Lu & MacKinnon, 1997) provide further evidence that the residue at position 5 of the signature sequence is pore lining. Mutation of isoleucine 142 in Kirl.l lowers the affinity of Kirl.l for the toxins tertiapin and Lq2 by eight- and two-fold, respectively.

The results of Dart et al. (1998b) also indicated that several other residues in the P-region orient their side chains towards the pore lumen, including a phenylalanine residue at position 147 and the tyrosine of the GYG motif. Similarly, in Shaker, three of the P-region’s four aromatic side chains were suggested to point towards the pore lumen (Lii & Miller, 1995). These results provided support for the hypothesis that selectivity for K+ is conferred by

cation- 7i interactions at the face of aromatic residues (Kumpf & Dougherty, 1993).

200 However, in light of the crystal structure of KcsA this hypothesis now seems unlikely. The atomic structure of KcsA shows that the selectivity filter is formed by backbone carbonyl oxygen atoms from five of the (T75, V76, G77, Y78 and G79, see table 5.1) amino acids in the signature sequence (Doyle et al., 1998b), indicating that selectivity is conferred through the interaction of K+ with oxygens as first proposed by Bezanilla & Armstrong (1972) and by Hille (1973; see section 1.6.2). However, this interaction requires the side chains of these residues to face away from the pore.

In contrast to the findings of cysteine scanning mutagenesis, comparative modelling of Kir2.1, based on the structure of KcsA, suggests that 1143 points its side chain away from the pore as shown in Figure 5.10. If isoleucine 143 in Kir2.1 points away from the pore, then it is likely that its hydrophobic side chain interacts with another hydrophobic residue in close proximity. Insertion of a hydrophilic residue would disrupt such hydrophobic interactions, which in turn may alter the packing of amino acid residues in the walls of the pore. This might, in turn, be propagated to the carbonyl oxygen backbone, leading to changes in ionic selectivity. But, why does substitution of the hydrophobic residue leucine, but not valine, significantly alter channel selectivity? The side-chain of leucine differs from those of isoleucine and valine in that it branches at the gamma carbon, whilst the latter amino acid side chains have beta branches. With a leucine at position five of the signature sequence, the backbone might be expected to be more flexible, altering the effective field of the carbonyl oxygens with respect to solvating the permeating ions and resulting in the enhanced Rb+ permeability (see Kirsch et a l, 1992).

It would be interesting to see whether substitution of more hydrophilic residues for isoleucine 143, such as glycine, asparagine or glutamine, result in non-functional channels. This might give us some further indication as to whether the side chains of this residue face the pore lumen, since one would expect pore-facing residues to be more tolerant of hydrophilic residues. In Shaker channels substitution of asparagine or glutamine for valine at this position gave functional channels, though non selective (Heginbotham et al., 1994). In contrast, in Kv2.1, substitution of charged or strongly hydrophilic residues at this position resulted in channels with immeasurable currents (Taglialatela et al., 1993).

The results presented in this chapter also suggest that changes in the signature sequence can alter the conformation of the pore. This region fails to conduct when the isoleucine at position 5 of the signature sequence is replaced by cysteine or serine, and conduction is

201 Figure 5.10.

Schematic representation o f the tetrameric structure o f Kir2A viewed from the extracellular side. Model generated by comparative modelling (MODELLER; Sali & Blundell, 1993) using the crystal structure of KcsA (Doyle et al., 1998b) as a template and a sequence alignment based on mutational analysis of Kjr2.1 (Minor et al., 1999). The side chains of 1143 are coloured white and can be seen oriented away from the pore. The carbonyl oxygens of the glycine (G144) and tyrosine (Y145) of the K+ signature sequence are coloured pink whilst the pink sphere represents a K+ ion. Restraints were applied with MODELLER to represent (i) four-fold symmetry (ii) the salt bridge between glutamate 138 and arginine 148 in adjacent sub-units (Yang et al., 1997), and (iii) the interactions within Kjr2.1 identified by mutagenesis (Minor et al., 1999).

202 reduced when an alanine is substituted at this position. However, the pore is open when the reducing agent DTT is present, which is likely to disrupt the intra-subunit disulphide bond between C l22 and C l54. This hypothesis cannot be tested by mutagenesis, because channels lacking this bond fail to assemble correctly (see Leyland et al., 1999; Bannister et al., 1999; Cho et al., 2000). However, the presence of an intra-subunit bond in I143S was confirmed by in vitro transcription studies using the reticulocyte system. Furthermore, these studies did not indicate the presence of any dimers, ruling out the possibility of an inter-subunit disulphide bond between native cysteines.

The conformational changes in the pore of Kir2.1 following mutation of isoleucine to cysteine or serine appear to bear some resemblance to the slow C-type inactivation process in KV1 channels, which involves conformational changes detectable at the outer region of the pore. In KV1 channels, the stability of the C-type inactivated state can be altered either by changing the residues surrounding the selectivity filter, or other nearby residues (see for example Lopez-Bameo et al., 1993; see also section 1.2.3a). Similarly, the stability of the ‘non­ conducting’ state in the Kir2.1 mutants I143C and I143S can also be altered by changing residues close to the P-region, in this case by disrupting the intra-subunit disulphide bond at the extracellular mouth of the pore.

The results from several other studies support the idea that the P-region can undergo conformation changes, which alter channel gating and permeation. Much of this evidence has been provided by analysis of microscopic gating in wild-type and mutant channels (see Choe et al., 1999; So et al., 2001; Lu et al., 2001a; Proks et al., 2001; Schwalbe et al., 2002).

At negative voltages, Kir channels display spontaneous single-channel gating activity. The gating kinetics of Kirl.lb can be well described by one open state and two closed states (Chepilko et al., 1995; Choe et al., 1998). The closed states are either short (~lms) or long (~40ms) in duration, and both exhibit a biphasic voltage dependence. The long closures can be abolished by EDTA, which suggests that they are due to block by divalent cations. In contrast, the rate of entering the short closed state varies with K+ concentration and is proportional to current amplitude, which suggests that they may result from block by permeating K+ (Choe et al., 1998). Choe et al. (1998) have proposed the existence of a variable energy well in which the pore normally has a shallow energy well for K+ ions, permitting rapid conduction, but can be converted to a deep well, which traps K+ and blocks ion flow long enough to be detected as a channel closure. Since the rate of closure was

203 proportional to the current, Choe et al. (1998) suggested that the change in the channel is triggered when an ion passes a certain part of the pore. Thus, the short closures are thought to reflect an inactivation process, which may be associated with changes in the conformation of the P-region.

These gating properties are known to differ between Kir2.1 and Kirl.lb (see for example Choe et al., 1999). Compared to Kirl.lb , Kir2.1 channels have much longer mean open and closed times, more resolvable closed states and display a stronger voltage dependence. Using chimeras of Kir2.1 and Kirl.lb , Choe et al. (1999) have shown that the extracellular segment between the M l and M2 transmembrane domains, which includes the P-region, is an important determinant of channel kinetics.

Recent mutagenesis studies have shown that changes in the signature sequence of Kjr2.1, either from single point mutations (So et al., 2001) or backbone mutations (where the amide carbonyl of the two conserved glycines is mutated to an ester carbonyl; see Lu et al., 2001a), have a dramatic effect on single channel gating. So et al. (2001) found that channel open times were affected by mutation of the tyrosine in the GYG motif, for example, the mean open time was reduced from 102ms at -120mV in wild-type to 6ms in a wild-type-I145V dimer. Lu et al. (2001b) found that backbone mutations produced distinct subconductance levels, which were attributed to local conformational changes as well as altered ion-ion and ion-carbonyl oxygen interactions. Mutations in the selectivity filter have also been shown to disrupt the fast gating of Kir6.2/SUR1 channels (Proks et al., 2001).

Schwalbe et al. (2002) have reported the effects of several mutations in the P-region, including the mutation I143N/Y145T, on single channel gating in Kir2.1. In this study, the single channel behaviour of Kir2.1 wild-type was well described by three conductance states, corresponding to the main open state and two subconductance states. Whilst the single channel behaviour of Kir2.1 wild-type channels was dominated by two states, a closed and main open state, mutations in the P-region, including the I143N/Y145T mutation, gave rise to Kjr2.1 channels that favoured various intermediate subconductances.

Finally, further evidence that the P-region can undergo conformational changes has been provided by the resolution of the crystal structure of KcsA in high and low [K+]0 (Zhou et al., 2001b). The 'conduction conformation', in which the backbone carbonyls point towards each other, is stabilized by the presence of K+ ions in the selectivity filter. But, when the protein is

204 partially depleted of K+, the selectivity filter adopts an alternative conformation in which the backbone carbonyls twist away from the centre of the channel such that the carbonyl of valine 76, which corresponds to 1143 in Kir2.1, hydrogen bonds with a water molecule outside the pore.

In summary, the results in this chapter show that mutation of isoleucine 143 affects ionic selectivity in Kir2.1. Furthermore, permeation is dependent on the residue at position 143, consistent with the involvement of the selectivity filter in the gating of inward rectifier potassium channels.

205 Chapter 6

6.0. Summary & Future Perspectives

The P-region of potassium channels, as for other voltage-gated channels, is typically regarded as the region primarily responsible for ionic selectivity and permeation characteristics. Recent evidence suggests that this region may also participate in the gating of some potassium channels including the inward rectifiers (see for example Lu et al., 2001a).

In this study, I have used the patch clamp technique to determine the effects of mutating residues in the P-region on the permeation properties of two members of the strong inward rectifier subfamily Kjr2.1 and Kir2.2. The results presented in this study support the notion that the P-region plays an important role in determining the permeation characteristics and ionic selectivity of strong inward rectifiers. However, they also highlight that, as in Kv channels, other regions are involved in determining the properties of the ion conduction pore. Finally, the results presented here also provide further support that this region is involved in

206 channel gating. These studies have raised several interesting points worthy of further investigation:

i) Residues in the P-region do not determine the difference in unitary conductance exhibited by Kjr2.2 and Kir2.1. Repunte et al. (1999) have shown that residues in the Ml-P-region and P-region-M2 extracellular linkers determine the difference in unitary conductance between Kir6.1 and Kir6.2. Is this also the case for members of the Kir2.0 subfamily? ii) The characteristics of Ba2+ blockage in Kir2.2 and Kir2.1 are not fully interchangeable following mutation of L148F in Kir2.2 and the reverse mutation in Kir2.1. Which other residues determine Ba2+ blockage in Kir2.0 subfamily channels? iii) In BKCa and Shaker channels, the kinetics of Ba2+ binding have been shown to be greatly affected by the occupancy of three discreet K+ binding sites known as the iock-in’, ‘enhancement’ and ‘deep high affinity’ sites (see Neyton & Miller, 1988; Harris et al., 1998). These sites impede the exit of the Ba2+ ion to the extracellular side at low concentrations, enhance the exit of Ba2+ in the inward direction in the presence of high K+ concentrations and exhibit high affinity deep binding inside the pore, respectively. It would be interesting to examine the unblocking rate of Ba2+ in low [K+]0 in these channels to determine whether the residues (L148 in Kir2.2 or F147 in Kir2.1) studied here contribute to the 'lock-in' site. iv) Mutation of isoleucine 143 to small residues such as cysteine and serine resulted in a loss of K+ permeation, which was proposed to be a consequence of a conformational change in the pore similar to that seen in KV1 channels during C- type inactivation. Starkus et al. (1997) have shown that Shaker channels are capable of conducting Na+ and Li+ in the C-type inactivated state provided that internal K+ is completely removed. It would be interesting to determine whether I143C and I143S mutant channels conduct Na+ or Li+ under similar conditions. v) Previous studies have shown that mutation of residues in the selectivity filter affect single channel gating behaviour (see for example So et al., 2001; Lu et al., 2001a). Mutating isoleucine to cysteine or serine clearly resulted in a loss of conduction, but what are the effects of mutating isoleucine to valine, leucine and threonine? Do these mutations cause more subtle conformational changes of the pore? It

207 would be interesting to examine the effects of these three mutations on single channel kinetics? vi) The Rb+ currents carried by I143L mutants appear to display inactivation at very negative membrane potentials. A similar phenomenon has been observed in Kir2.1-wild-type and mutant channels with N fV - as the permeant ion and has been attributed to conformational changes of the pore, arising from the permeation and interaction of N H / with the pore region (see for example Chang & Shieh, 2002). Does the inactivation of Rb+ currents in I143L represent a similar phenomenon? It would be interesting to investigate this further through analysis of macroscopic currents and single channel kinetics.

208 Chapter 7

Bibliography

Abrams, C. J. (2000). Studies of the molecular basis of selectivity and gating in the inward rectifier potassium channel Kir2.1. PhD Thesis.

Abrams, C. J., Davies, N. W., Shelton, P. A. and Stanfield, P. R. (1996). The role of a single aspartate residue in ionic selectivity and block of a murine inward rectifier K+ channel Kir2.1. Journal of Physiology. 493.3,643-649.

Abraham, M. R., Jahangir, A., Alekseev, A. E. and Terzic, A. (1999). Channelopathies of inwardly rectifying potassium channels. FASEB Journal. 13, 1901-1910.

Ackermann, M. J. and Clapham, D. E. (1997). Mechanisms of disease: Ion channels - - Basic science and clinical disease. New England Journal of Medicine. 336, 1575-1586.

Adelman, J. P., Shen, K. Z., Kavanaugh, M. P., Warren, R. A., Wu, Y. N., Lagrutta, A., Bond, C. T. and North, R. A. (1992). Calcium-activated potassium channels expressed from cloned complementary DNAs. Neuron. 9,209-216.

Adrian, R. H. (1964). The rubidium and potassium permeability of frog muscle membrane. Journal of Physiology. 175, 134-159.

209 Adrian, R. H. (1969). Rectification in muscle membrane. Progress in Biophysics and Molecular Biology. 19, 341-369.

Adrian, R. H., Chandler, W. K. and Hodgkin, A. L. (1970a). Voltage clamp experiments in striated muscle fibres. Journal of Physiology. 208, 607-644.

Adrian, R. H., Chandler, W. K. and Hodgkin, A. L. (1970b). Slow changes in potassium permeability in skeletal muscle. Journal of Physiology. 208, 645-668.

Adrian, R. H. and Freygang, W. H. (1962a). The potassium and chloride conductance of frog muscle membrane. Journal of Physiology. 163, 61-103.

Adrian, R. H. and Freygang, W. H. (1962b). Potassium conductance of frog muscle membrane under controlled voltage. Journal of Physiology. 163, 104-114.

Aggarwal, S. K. and MacKinnon, R. (1996). Contribution of the S4 segment to gating charge in the Shaker K+ channel. Neuron. 16, 1169-1177.

Agnew, W. S., Levinson, S. R., Brabson, J. S. and Raftery, M. A. (1978). Purification of the tetrodotoxin-binding component associated with the voltage-sensitive sodium channel from Electrophorus electricus electroplax membranes. Proceedings o f the National Academy of Sciences USA. 75, 2626-2610.

Aguilar-Bryan, L. A., Bryan, J. and Nakazaki, M. (2001). Of mice and men: KATP channels and insulin secretion. Recent progress in Hormone Research. 56, 47-68.

Aguilar-Bryan, L., Clement, J. P., Gonzalez, G., Kunjilwar, K., Babenko, A. and Bryan, J. (1998). Toward understanding the assembly and structure of KAtp channels. Physiological Reviews. 78, 227-248.

Aguilar-Bryan, L., Nichols, C. G., Wechsler, S. W., Clement IV, J. P., Boyd HI, A. E., Gonzalez, G., Herrera-Sosa, H., Nguy, K., Bryan, J. and Nelson, D. A. (1995). Cloning of the (3 cell high-affinity sulfonylurea receptor: a regulator of insulin secretion. Science. 268, 423-426.

Ai, T., Fujuwara, Y., Tsuji, K., Otani, H., Nakano, S., Kubo, Y. and Horie, M. (2002). Novel KCNJ2 mutation in familial periodic paralysis with ventricular dysrhythmia. Circulation. 105, 2592-2594.

Aidley, D. J. and Stanfield, P. R. (1996). Ion Channels. Molecules in Action. Cambridge University Press, UK.

Alagem, N., Dvir, M. and Reuveny, E. (2001). Mechanism of Ba2+ block of a mouse inwardly rectifying K+ channel: differential contribution by two discrete residues. Journal o f Physiology. 534.2, 381-393.

Aleksandrov, A., Velimirovic, B. and Clapham, D. E. (1996). Inward rectification of the IRK1 K+ channel reconstituted in lipid bilayers. Biophysical Journal. 70, 2680-2687.

Ali, S., Wei, Y., Lerea, K. M., Becker, L., Rubin, C. S. and Wang, W. (2001). PKA- induced stimulation of ROMK1 channel activity is governed by both tethering and non­

210 tethering domains of an A kinase anchor protein. Cellular Physiology and Biochemistry. 11, 135-142.

Aimers, W. (1972a). Potassium conductance changes in skeletal muscle and the potassium concentration in the transverse tubules. Journal of Physiology. 225, 33-56.

Aimers, W. (1972b). The decline of potassium permeability during extreme hyperpolarization in frog skeletal muscle. Journal of Physiology. 225, 57-83.

Ammala, C., Moorhouse, A., Gribble, F., Ashfield, R., Proks, P., Smith, P. A., Sakura, H., Coles, B., Ashcroft, S. J. H. and Ashcroft, F. M. (1996). Promiscuous coupling between the sulphonylurea receptor and inwardly rectifying potassium channels. Nature. 379, 545- 548.

Andelfinger, G., Tapper, A. R., Welch, R. C., Vanoye, C. G., George, A. L. and Benson, D. W. (2002). KCNJ2 mutation results in Andersen syndrome with sex-specific cardiac and skeletal muscle phenotypes. American Journal of Human Genetics. 71, 663-668.

Appelbaum, E. R. and Shatzman, A. R. (1999). Prokaryotic in vivo expression systems. In Protein expression: a practical approach. Ed. Higgins, S.J. and Hames, B.D. ppl69-200. IRL Press, Oxford.

Arcangeli, A., Wanke, E., Olivotto, M., Camagni, S. and Ferroni, A. (1987). Three types of ion channels are present on the plasma membrane of friend erythroleukemia cells. Biochemical and Biophysical Research Communications. 146, 1450-1457.

Armstrong, C. M. (1966). Time course of TEA+-induced anomalous rectification in squid giant axons. Journal of General Physiology. 50,491-503.

Armstrong, C. M. (1969). Inactivation of the potassium conductance and related phenomena by quaternary ammonium ion injection in squid axons. Journal of General Physiology. 54, 553-575.

Armstrong, C. M. (1971). Interaction of tetraethylammonium ion derivatives with the potassium channels of giant axons. Journal of General Physiology. 58, 413-437.

Armstrong, C. M. (1975). Potassium pores of nerve and muscle membranes. In Membranes: A series of advances, ed. G. Eisenman, 3, 325-358. New York: Dekker, 538pp.

Armstrong, C. M. and Bezanilla, F. (1977). Inactivation of the sodium channel n . Gating current experiments. Journal of General Physiology. 70, 567-590.

Armstrong, C. M. and Binstock, L. (1965). Anomalous rectification in the squid giant axon injected with tetraethylammonium chloride. Journal of General Physiology. 48, 859- 872.

Armstrong, C. M. and Hille, B. (1998). Voltage-gated ion channels and electrical excitability. Neuron. 20,371-380.

Armstrong, C. M. and Taylor, S. R. (1980). Interaction of barium ions with potassium channels in squid giant axons. Biophysical Journal. 30, 473-488.

211 Artigas, P. and Gadsby, D. C. (2002). Ion channel-like properties of the Na+/K+ pump. Annals of the New York Academy of Sciences. 976, 31-40.

Ashcroft, F. M and Gribble, F. M. (1998). Correlating structure and function in ATP- sensitive K+ channels. Trends in Neurosciences. 21, 288-294.

Ashen, M. D., O ’Rouke, B., Kluge, K. A., Johns, D. C. and Tomaselli, G. F. (1995). Inward rectifier K+ channel from human heart and brain: cloning and stable expression in a human cell line. American Journal of Physiology. 268, H506-511.

Ashford, M. L. J., Bond, C. T., Blair, T. A. and Adelman, J. P. (1994). Cloning and functional expression of a rat heart KA t p channel. Nature. 370, 456-459.

Ashmole, I, Goodwin, P. A. and Stanfield, P. R. (2001). TASK-5, a novel member of the tandem pore K+ channel family. Pfliigers Archives. 442, 828-833.

Atkinson, N. S., Robertson, G. A. and Ganetzky, B. (1991). A component of calcium-activated potassium channels encoded by the Drosophila slo locus. Science. 253, 551-555.

Babenko, A. P., Aguilar-Bryan, L. and Bryan, J. (1998). A view of SUR/Kjr6.X KA t p channels. Annual Review of Physiology. 60, 667-687.

Babila, T., Moscucci, A., Wang, H., Weaver, F. E. and Koren, G. (1994). Assembly of mammalian voltage-gated potassium channels: evidence for an important role of the first transmembrane segment. Neuron. 12,615-626.

Bannister, J. P. A., Young, B. A., Sivaprasadarao, A. and Wray, D. (1999). Conserved extracellular cysteine residues in the inwardly rectifying potassium channel Kir2.3 are required for function but not expression in the membrane. FEBS Letters. 458, 393-399.

Barchi, R. L. (1982). Biochemical studies of the excitable membrane sodium channel. International Review of Neurobiology. 23,69-101.

Bargmann, C. I. (1998). Neurobiology of the Caenorhabditis elegans genome. Science. 282, 2028-2033.

Barry, P. H. and Lynch, L. W. (1991). Liquid junction potentials and small cell effects in patch-clamp analysis. Journal of Membrane Biology. 121, 101-107.

Basso, C., Labarca, P., Stefani, E., Alvarez, O. and Latorre, R. (1998). Pore accessibility during C-type inactivation in Shaker K+ channels. FEBS Letters. 429, 375-380.

Baukrowitz, T and Fakler, B. (2000). KA t p channels gated by intracellular nucleotides and phospholipids. European Journal of Biochemistry. 267, 5842-5848.

Baukrowitz, T., Schulte, U., Oliver, D., Herlitze, S, Krauter, T., Tucker, S. J., Ruppersberg, J. P. and Fakler, B. (1998). PIP2 and PIP as determinants for ATP inhibition of KATP channels. Science. 282,1141-1144.

Baukrowitz, T. and Yellen, G. (1996). Use dependent blockers and exit rate of the last ion from the multi-ion. Science. 271, 653-656.

212 Baumann, A., Krah-Jentgens, I., Muller, R., Miiller-Holtkamp, F., Seidel, R., Kecskemethy, N., Casal, J., Ferrus, A. and Pongs, O. (1987). Molecular organization of the maternal effect region of the Shaker complex of Drosophila : characterization of an Ia channel transcript with homology to vertebrate Na+ channel. EMBO Journal. 6, 3419-3429.

Begenesich, T. and Stevens, C. F. (1975). How many conductance states do potassium channels have? Biophysical Journal. 15, 843-846.

Bemheim, L. and Bader, C. R. (2002). Human myoblast differentiation: Ca2+ channels are activated by K+ channels. News in Physiological Sciences. 17, 22-26.

Bettahi, I., Marker, C. L., Roman, M. I. and Wickman, K. (2002). Contribution of the Kir3.1 subunit to the muscarinic-gated atrial potassium channel I kacii - Journal of Biological Chemistry. 277, 48282-48288.

Bezanilla, F. (2000). The voltage sensor in voltage-dependent ion channels. Physiological Reviews. 80, 555-592.

Bezanilla, F. and Armstrong, C. M. (1972). Negative conductance caused by entry of sodium and cesium ions into the potassium channels of squid axons. Journal of General Physiology. 60, 588-608.

Bezanilla, F. and Armstrong, C. M. (1977). Inactivation of the sodium channel: I. Sodium current experiments. Journal of General Physiology. 70, 549-566.

Bianchi, L., Roy, M. L., Taglialatela, M., Lundgren, D. W., Brown, A. M. and Ficker, E. (1996), Regulation by spermine of native inward rectifier K+ channels in RBL-1 cells. Journal of Biological Chemistry. 271, 6114-6121.

Biermans, G., Vereecke, J. and Carmeliet, E. (1987). The mechanism of the inactivation of the inward-rectifying K current during hyperpolarizing steps in guinea-pig ventricular myocytes. Pfliigers Archives. 410,604-613.

Bimboim, H. C. and Doly, J. (1979). A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Research. 7, 1513-1523.

Bissonnette, J. M. (2002). The role of calcium-activated potassium channels in respiratory control. Respiratory Physiology & Neurobiology. 131, 145-153.

Bixby, K. A., Nanao, M. H., Shen, N. V., Kreusch, A., Bellamy, H., Pfaffinger, P. J. and Choe, S. (1999). Zn2+-binding and molecular determinants of tetramerization in voltage- gated K+ channels. Nature Structural Biology. 6, 38-41.

Boim, M. A., Ho, K., Shuck, M. E., Bienkowski, M. J., Block, J. H., Slightom, J. L., Yang, Y., Brenner, B. M. and Hebert, S. C. (1995). ROMK inwardly rectifying ATP- sensitive K+ channel n. Cloning and distribution of alternative forms. American Journal of Physiology. 268, FI 132-F1140.

Bond, C. T., Ammala, C., Ashfield, R., Blair, T. A., Gribble, F., Khan, R. N., Lee, K., Proks, P., Rowe, I. C. M., Sakura, H., Ashford, M. J., Adelman, J. P. and Ashcroft, F. M. (1995). Cloning and functional expression of the cDNA encoding an inwardly-rectifying

213 potassium channel expressed in pancreatic P-cells and in the brain. FEBS Letters. 367, 61- 66.

Bond, C. T., Pessia, M., Xia, X-M., Lagrutta, A., Kavanaugh, M. P. and Adelman, J. P. (1994). Cloning and expression of a family of inward rectifier potassium channels. Receptors and Channels. 2,183-191.

Bradley, K. K., Jaggar, J. H., Bonev, A. D., Heppner, T. J., Flynn, E. R. M., Nelson, M. T. and Horowitz, B. (1999). Kir2.1 encodes the inward rectifier potassium channel in rat arterial smooth muscle cells. Journal of Physiology. 515.3,639-651.

Brahmajothi, M. V., Morales, M. J., Liu, S., Rasmusson, R. L., Campbell, D. L. and Strauss, H. C. (1996). In situ hybridization reveals extensive diversity of K+ channel mRNA in isolated ferret cardiac myocytes. Circulation Research. 78, 1083-1089.

Bredt, D. S., Wang, T-L., Cohen, N. A., Guggino, W. B. and Snyder, S. H. (1995). Cloning and expression of two brain-specific inwardly rectifying potassium channels. Proceedings of the National Academy of Sciences USA. 92, 6753-6757.

Breitwieser, G. E. and Szabo, G. (1985). Mechanism of muscarinic receptor-induced K+ channel activation as revealed by hydrolysis-resistant GTP analogues. Journal of General Physiology. 91, 469-493.

Brenner, R., Perez, G. J., Bonev, A. D., Eckman, D. M., Kosek, J. C., Wiler, S. W., Patterson, A. J., Nelson, M. T. and Aldrich, R. W. (2000). Vasoregulation by the pi subunit of the calcium-activated potassium channel. Nature. 407, 870-876.

Brochiero, E., Wallendorf, B., Gagnon, D., Laprade, R. and Lapointe, J-Y. (2002). Cloning of rabbit Kir6.1, SUR2A, and SUR2B: possible candidates for a renal KA t p channel. American Journal of Physiology. 282, F289-F300.

Brown, A. M. (1994). The pore of voltage-dependent potassium channels. Renal Physiology and Biochemistry. 17, 178-181.

Brown, D. A. and Adams, P. R. (1980). Muscarinic suppression of a novel voltage- sensitive K+-current in a vertebrate neurone. Nature. 283, 673-676.

Brown, A. M., Drewe, J. A., Hartmann, H. A., Taglialatela, M., De Biasi, M., Soman, K. and Kirsch, G. E. (1994). The potassium pore and its regulation. Annals o f the New York Academy of Sciences. 707, 74-80.

Butler, M. (1991). The characteristics and growth of cultured cells. In Mammalian cell biotechnology: a practical approach. Ed. Butler, M. ppl-25. IRL Press, Oxford.

Butler, M. (1996). Animal cell culture and technology - the basics. Ed. Beynon, R.J. and Howe, C. IRL Press, Oxford.

Butler, A., Tsunoda, S., McCobb, D. P., Wei, A. and Salkoff, L. (1993). mSlo, a complex mouse gene encoding ‘Maxi’ calcium-activated potassium channels. Science. 261, 221-224.

214 Butler, A., Wei, A. G., Baker, K. and Salkoff, L. (1989). A family of putative potassium channel genes in Drosophila. Science. 243, 943-947.

Carmeliet, E. (1961). Chloride ions and the membrane potential of Purkinje fibres. Journal of Physiology. 156, 375-388.

Catterall, W. A. (1988). Structure and function of voltage-sensitive ion channels. Science. 242,50-61.

Cha, A. and Bezanilla, F. (1997). Characterizing voltage-dependent conformational changes in the Shaker K+ channel with fluorescence. Neuron. 19, 1127-1140.

Chandy, K. G., Fantino, E., Wittekindt, O., Kalman, K., Tong, L-L., Ho, T-H., Gutman, G. A., Crocq, M-A., Ganguli, R., Nimgaonkar, V., Morris-Rosendahl, D. J. and Gargus, J. J. (1998). Isolation of a novel potassium channel gene hSKCa3 containing a polymorphic CAG repeat: a candidate for schizophrenia and bipolar disorder? Molecular Psychiatry. 3, 32-37.

Chang, H-K. and Shieh, R. C. (2002). Conformational changes in Kir2.1 channels during NH 4 +-induced inactivation. Journal of Biological Chemistry. 278, 908-918.

Cheng, S. H., Gregory, R. J., Marshall, J., Paul, S., Souza, D. W., White, G. A., O’Riordan, C. R. and Smith, A. E. (1990). Defective intracellular transport and processing of CFTR is the molecular basis of most cystic fibrosis. Cell. 63, 827-834.

Chepilko, S., Zhou, H., Sackin, H. and Palmer, L. G. (1995). Permeation and gating properties of a cloned renal K+ channel. American Journal of Physiology. 268, C389-C401.

Cho, H. C., Tsushima, R. G., Nguyen, T-T. T., Guy, H. R. and Backx, P. H. (2000). Two critical cysteine residues implicated in disulfide bond formation and proper folding of Kir2.1. Biochemistry. 39,4649-4657.

Choe, S. (2002). Potassium channel structures. Natures Reviews Neuroscience. 3, 115-121.

Choe, H., Palmer, L. G. and Sackin, H. (1999). Structural determinants of gating in inward-rectifier K+ channels. Biophysical Journal. 76, 1988-2003.

Choe, H., Sackin, H. and Palmer, L. G. (1998). Permeation and gating of an inwardly rectifying potassium channel. Evidence for a variable energy well. Journal of General Physiology. 112, 433-446.

Choe, H., Zhou, H., Palmer, L. G. and Sackin, H. (1997). A conserved cytoplasmic region of ROMK modulates pH sensitivity, conductance, and gating. American Journal of Physiology. 273, F516-F529.

Choe , H., Sackin, H. And Palmer, L. G. (2000). Permeation properties of inward- rectifier potassium channels and their molecular determinants. Journal of General Physiology. 115, 391-404.

Choi, K. L., Aldrich, R. W. and Yellen, G. (1991). Tetraethyl ammonium blockade distinguishes two inactivation mechanisms in voltage-gated K+ channels. Proceedings o f the National Academy of Sciences USA. 88, 5092-5095.

215 Chopra, B., Manolopoulos, J., Stanfield, P. R., Norman, R. I., Stonehouse, A. H., Giblett, S., Evans, R. J. and Grubb, B. D. (1999). The distribution of the inwardly rectifying potassium channel, Kir2.3, in a population of rat dorsal root ganglion neurones. Journal of Physiology. 518.P, 167P.

Christie, M. J., Adelman, J. P., Douglass, J. and North, R. A. (1989). Expression of a cloned rat brain potassium channel in Xenopus oocytes. Science. 244, 221-224.

Christie, M. J., North, R. A., Osbourne, P. B., Douglass, J. and Adelman, J. P. (1990). Heteropolymeric potassium channels expressed in Xenopus oocytes from cloned subunits. Neuron. 4,405-411.

Christophersen, P. (1991). Ca2+-activated K+ channel from human erythrocyte membranes: single channel rectification and selectivity. Journal of Membrane Biology. 119, 75-83.

Chu, G., Hayakawa, H. and Berg, P. (1987). Electroporation for the efficient transfection of mammalian cells with DNA. Nucleic Acids Research. 15, 1311-1326.

Chutkow, W. A., Makielski, J. C., nelson, D. J., Burant, C. F. and Fan, Z. (1999). Alternative splicing of SUR2 exon 17 regulates nucleotide sensitivity of the ATP-sensitive potassium channel. Journal of Biological Chemistry. 274, 13656-13665.

Chutkow, W. A., Simon, M. C., Le Beau, M. M. and Burant, C. F. (1996). Cloning, tissue expression, and chromosomal localization of SUR2, the putative drug-binding subunit of cardiac, skeletal muscle, and vascular K a t p channels. Diabetes. 45, 1439-1445.

Ciani, S., Krasne, S., Miyazaki, S. and Hagiwara, S. (1978). A model for anomalous rectification: electrochemical-potential-dependent gating of membrane channels. Journal of Membrane Biology. 44,103-134.

Cleeman, L. and Morad, M. (1979). Potassium currents in frog ventricular muscle: evidence from voltage clamp currents and extracellular K accumulation. Journal of Physiology. 286, 113-143.

Clemens, M. J. and Pruijn, G. J. (1999). Protein synthesis in eukaryotic cell-free systems. In Protein expression: a practical approach. Ed. Higgins, S.J. and Hames, B.D. ppl29-168. IRL Press, Oxford.

Clement, J. P., Kunjilwar, K., Gonzalez, G., Schwanstecher, M., Panten, U., Aguilar- Bryan, L. and Bryan, J. ( 1 9 9 7 ) . Association and stoichiometry of K a t p channel subunits. Neuron. 18, 8 2 7 - 8 3 8 .

Codina, J., Yatani, A., Grenet, D., Brown, A. M. and Bimbaumer, L. (1987). The a- subunit of the GTP binding protein Gk opens atrial potassium channels. Science. 236, 442- 445.

Coetzee, W. A., Amarillo, Y., Chiu, J., Chow, A., Lau, D., McCormack, T., Moreno, H., Nadal, M. S., Ozaita, A., Pountney, D., Saganich, M., Vega-Saenz de Miera, E and Rudy, B. (1999). Molecular diversity of K+ channels. Annals of the New York Academy of Sciences. 868, 233-285.

216 Cohen, N. A., Brenman, J. E., Snyder, S. H. and Bredt, D. S. (1996a). Binding of the inward rectifier K+ channel Kir2.3 to PSD-95 is regulated by protein kinase A phosphorylation. Neuron. 17, 759-767.

Cohen, N. A., Sha, Q., Makhina, E. N., Lopatin, A. N., Linder, M. E., Snyder, S. H. and Nichols, C. G. (1996b). Inhibition of an inward rectifier potassium channel (Kir2.3) by G-protein py subunits. Journal of Biological Chemistry. 271, 32301-32305.

Cole, K. S. and Curtis, H. J. (1938). Electric impedance of Nitella during activity. Journal of General Physiology. 22, 37-64.

Cole, K. S. and Curtis, H. J. (1939). Electrical impedance of the squid giant axon during activity. Journal of General Physiology. 22,649-670.

Collins, A., German, M. S., Jan, Y. N., Jan, L. Y. and Zhao, B. (1996). A strongly inwardly rectifying K+ channel that is sensitive to ATP. Journal of Neuroscience. 16, 1-9.

Colquhoun, D. and Sigworth, F. L. (1983). Fitting and statistical analysis of single channel records. In Single Channel Recording. Ed. B. Sakmann and E. Neher. New York, Plenum Press.

Colyer, J. (1999). Monitoring protein expression. In Protein expression: a practical approach. Ed. Higgins, S.J. and Hames, B.D. pp224-265. IRL Press, Oxford.

Connor, J. A. and Stevens, C. F. (1971). Voltage clamp studies of a transient outward current in gastropod neural somata. Journal of Physiology. 213, 21-30.

Connors, N. C. and Kofuji, P. (2002). Dystrophin Dp71 is critical for the clustered localization of potassium channels in retinal glial cells. Journal of Neuroscience. 22, 4321- 4327.

Conti, F., Hille, B., Neumcke, B., Nonner, W. and Stamped, R. (1976). Conductance of the sodium channel in myelinated nerve fibres with modified sodium inactivation. Journal of Physiology. 262, 729-742.

Conti-Tronconi, B. M. and Raftery, M. A. (1982). The nicotinic cholinergic receptor: correlation of molecular structure with functional properties. Annual Review of Biochemistry. 51,491-530.

Cook, D. L. and Hales, C. N. (1984). Intracellular ATP directly blocks K+ channels in pancreatic P-cells. Nature. 311,271-273.

Corey, S., Krapivinsky, G., Krapivinsky, L. and Clapham, D. E. (1998). Number and stoichiometry of subunits in the native atrial G-protein-gated K+ channel, I r a q i - Journal of Biological Chemistry. 273, 5271-5278.

Coulter, K. L., Perier, F., Radeke, C. M. and Vandenberg, C. A. (1995). Identification and molecular localization of a pH-sensing domain for the inward rectifier potassium channel HIR. Neuron. 15,1157-1168.

Covarrubias, M. and Steinbach, J. H. (1990). Excision of membrane patches reduces the mean open time of nicotinic acetylcholine receptors. Pflugers Archives. 416, 385-392.

217 Covarrubias, M., Wei, A. A. and Salkoff, L. (1991). Shaker, Shal, Shab and Shaw express independent K+ current systems. Neuron. 7, 763-773.

Cui, J., Mandel, G., DiFrancesco, D., Kline, R. P., Pennefather, P., Haspel, H. C. and Cohen, I. S. (1992). Functional expression of an inwardly rectifying potassium current in Xenopus oocytes. (Abstr.) Biophysical Journal. 61, A377.

Curran, M. E. (1998). Potassium ion channels and human disease: phenotypes to drug targets? Current Opinion in Biotechnology. 9, 565-572.

Dart, C. and Leyland, M. L. (2001). Targeting of an A kinase-anchoring protein, AKAP79, to an inwardly rectifying potassium channel, Kir2.1. Journal of Biological Chemistry. 23, 20499-20505.

Dart, C., Leyland, M. L., Barret-Jolley, R., Shelton, P. A., Spencer, P. J., Conley, E. C., Sutcliffe, M. J. and Stanfield, P. R. (1998a). The dependence of Ag+ block of a potassium channel, murine Kir2.1, on a cysteine residue in the selectivity filter. Journal of Physiology. 511.1, 15-24.

Dart, C., Leyland, M. L., Spencer, P. J., Stanfield, P. R. and Sutcliffe, M. J. (1998b). The selectivity filter of a potassium channel, murine Kir2.1, investigated using scanning cysteine mutagenesis. Journal of Physiology. 511.1, 25-32.

Dascal, N., Doupnik, C. A., Ivanina, T., Bausch, S., Wang, W., Lin, C., Garvey, J., Chavkin, C., Lester, H. A. and Davidson, N. (1995). Inhibition of function in Xenopus oocytes of the inwardly rectifying G-protein-activated atrial K channel (GIRK1) by overexpression of a membrane-attached form of the C-terminal tail. Proceedings o f the National Academy of Sciences USA. 92, 6758-6762.

Dascal, N., Schreibmayer, W., Lim, N. F., Wang, W., Chavkin, C., DiMagno, L., Labarca, C., Kieffer, B. L., Gaveriaux-Ruff, C., Trollinger, D., Lester, H. A. and Davidson, N. (1993). Atrial G protein-activated K+ channel: Expression cloning and molecular properties. Proceedings of the National Academy of Sciences USA. 90, 10235-10239.

Davies, N. W. (1993). A suite of programs for acquisition and analysis of voltage- and patch-clamp data developed using the AxoBASIC library. Journal of Physiology. 459, 11 IP.

Davies, N. W., Standen, N. B. and Stanfield, P. R. (1992). The effect of intracellular pH on ATP-dependent potassium channels of frog skeletal muscle. Journal of Physiology. 445, 549-568.

De Biasi, M., Drewe, J. A., Kirsch, G. E. and Brown, A. M. (1992). Histidine substitution identifies a surface position and confers Cs+ selectivity on a K+ pore. Biophysical Journal. 65, 1235-1242.

Derst, C., Hirsch, J. R., Preisig-Miiller, R., Wischmeyer, E., Karschin, A., Doring, F., Thomzig, A., Veh, R. A., Schlatter, E., Kummer, W. and Daut, J. (2001b). Cellular localization of the potassium channel Kir7.1 in guinea pig and human kidney. Kidney International. 59, 2197-2205.

218 Derst, C., Karschin, C., Wischmeyer, E., Hirsch, J. R., Preisig-Mliller, R., Rajan, S., Engel, H., Grzeschik, K-H., Daut, J. and Karschin, A. (2001a). Genetic and functional linkage of Kir5.1 and Kir2.1 channel subunits. FEBS Letters. 491, 305-311.

Derst, C., Wischmeyer, E., Preisig-Muller, R., Spauschus, A., Konrad, M., Hensen, P., Jeck, N., Seyberth, H. W., Daut, J. and Karschin, A. (1998). A hyperprostaglandin E syndrome mutation in K irl.l (renal outer medullary potassium) channels reveals a crucial residue for channel function in Karl.3 channels. Journal of Biological Chemistry. 273, 23884-23891.

Diaz, F., Wallner, M., Stefani, E., Toro, L. and Latorre, R. (1996). Interaction of internal Ba2+ with a cloned Ca2+-dependent K+ (h slo) channel from smooth muscle. Journal of General Physiology. 107, 399-407.

DiFrancesco, D., Ferroni, A. and Visentin, S. (1984). Barium-induced blockade of the inward rectifier in calf Purkinje fibres. Pfliigers Archives. 402, 446-453.

Dissman, E., Wischmeyer, E., Spauschus, A., Pfeil, D. V., Karschin, C. and Karschin, A. (1996).Functional expression and cellular mRNA localization of a G-protein activated K+ inward rectifier isolated from rat brain. Biochemical and Biophysical Research Communications. 223, 474-479.

Doi, T., Fakler, B., Schultz, J. H., Brandle, U., Weidemann, S., Zenner, H. P., Lang, F. and Ruppersberg, J. P. (1996). Extracellular K+ and intracellular pH allosterically regulate renal Kirl.l channels. Journal of Biological Chemistry. 271, 17261-17266.

Doring, F., Derst, C., Wischmeyer, E., Karschin, C., Schneggenburger, R., Daut, J. and Karschin, A. (1998). The epithelial inward rectifier channel Kir7.1 displays unusual K+ permeation properties. Journal of Neuroscience. 18, 8625-8636.

Doring, F., Wischmeyer, E., Ktihnlein, R. P., Jackie, H. and Karschin, A. (2002). Inwardly rectifying K+ channel in Drosophila: a crucial role of cellular milieu factors for Kir channel function. Journal of Biological Chemistry. 277, 22554-22561.

Doupnik, C. A., Davidson, N. and Lester, H. A. (1995). The inward rectifier potassium channel family. Current Opinion in Neurobiology. 5, 268-277.

Doyle, A., Griffiths, J. B. and Newell, D. G. (1998a). Haemocytometer cell counts and viability studies. In Cell and tissue culture: laboratory procedures. Module 4B:1. John Wiley and Sons, Ltd.

Doyle, D. A., Morais Cabral, J., Pfuetzner, R. A., Kuo, A., Gulbis, J. M., Cohen, S. L., Chait, B. T. and MacKinnon, R. (1998b). The structure of the potassium channel: Molecular basis of K+conduction and selectivity. Science. 280,69-77.

Duprat, F., Lesage, F., Guillemare, E., Fink, M., Hugnot, J. P., Bigay, J., Lazdunski, M., Romey, G. and Barhanin, J. (1995). Heterologous multimeric assembly is essential for K+channel activity of neuronal and cardiac G-protein-activated inward rectifiers. Biochemical and Biophysical Research Communications. 212, 657-663.

Durrell, S. R. and Guy, H. R. (2001). A family of putative Kir potassium channels in prokaryotes. BMC Evolutionary Biology. 1, 1-9.

219 Eaton, D. C. and Brodwick, M. S. (1980). Effects of barium on the potassium conductance of squid axon. Journal of General Physiology. 75, 727-750.

Edwards, F. R., Hirst, G. D. and Silverberg, G. D. (1988). Inward rectification in rat cerebral arterioles: involvement of potassium ions in autoregulation. Journal of Physiology. 404, 455-466.

Ehlert, F., Bierbaum, P., and Schorr, J. (1993). Importance of DNA quality for transfection efficiency. BioTechniques. 14, 546.

England, S. K., Uebele, V. N., Kodali, J et al. (1995). A novel K+ channel beta- subunit (hKv beta 1.3) is produced via alternative mRNA splicing. Journal of Biological Chemistry. 270, 531-534.

Fakler, B., Bond, C. T., Adelman, J. P. and Ruppersberg, J. P. (1996a). Heterooligomeric assembly of inward-rectifier K+ channels from subunits of different subfamilies: Kir2.1(IRKl) and Kir4.1 (BIR10). Pflilgers Archives. 433,77-83.

Fakler, B., Brandle, U., Bond, Ch., Glowatzki, E., Konig, C., Adelman, J. P., Zenner, H-P. and Ruppersberg, J. R. (1994a). A structural determinant of differential sensitivity of cloned inward rectifier K+channels to intracellular spermine. FEBS Letters. 356, 199-203.

Fakler, B., Brandle, U., Glowatzki, E., Weidemann, S., Zenner, H. P. and Ruppersberg, J. P. (1995). Strong voltage-dependent inward rectification of inward rectifier K+channels is caused by intracellular spermine. Cell. 80,149-154.

Fakler, B., Brandle, U., Glowatzki, E., Zenner, H. P. and Ruppersberg, J. P. (1994b). Kir2.1 inward rectifier K+ channels are regulated independently by protein kinases and ATP hydrolysis. Neuron. 13, 1413-1420.

Fakler, B. and Ruppersburg, J. P. (1996). Functional and molecular diversity classifies the family of inward-rectifier K+ channels. Cell Physiology and Biochemistry. 6, 195-209.

Fakler, B., Schultz, J.H., Yang, J., Schulte, U., Brandle, U., Zenner, H. P., Jan, L. Y. and Ruppersberg, J. P. (1996b). Identification of a titratable lysine residue that determines sensitivity of kidney potassium channels (ROMK) to intracellular pH. EMBO Journal. 15, 4093-4099.

Falk, T., Meyerhof, W., Corrette, B. J., Schafer, J., Bauer, C. K., Schwarz, J. R. and Richter, D. (1995). Cloning, functional expression and mRNA distribution of an inwardly rectifying potassium channel protein. FEBS Letters. 367,127-131.

Fanger, C. M., Ghanshani, S., Logsdon, N. J., Rauer, H., Kalman, K., Zhou, J., Beckingham, K., Chandy, K. G., Cahalan, M. D. and Aiyar, J. (1999). Calmodulin mediates calcium-dependent activation of the intermediate conductance Kca channel, IKCal. Journal of Biological Chemistry. 274, 5746-5754.

Fedida, D. and Hesketh, J. C. (2001). Gating of voltage-dependent potassium channels. Progress in Biophysics and Molecular Biology. 75, 165-199.

Feigner, P. L., Gadek, T. R., Holm, M., Roman, R., Chan, H. W., Wenze, M., Northrop, J. P., Ringold, G. M. and Danielsen, M. (1987). Lipofection: a highly efficient,

220 lipid-mediated DNA-transfection procedure. Proceedings of the National Academy of Sciences o f the USA. 84, 7413-7417.

Ficker, E., Taglialatela, M., Wible, B. A., Henley, C. M. and Brown, A. M. (1994). Spermine and spermidine as gating molecules for inward rectifier K+ channels. Science. 266, 1068-1072.

Fink, M., Duprat, F., Heurteaux, C., Lesage, F., Romey, G., Barhanin, J. and Lazdunski, M. (1996). Dominant negative chimeras provide evidence for homo and heteromultimeric assembly of inward rectifier K+ channel proteins via their N-terminal end. FEBS Letters. 378, 64-68.

Finkel, A. S. (1991). Progress in instrumentation technology for recording from single channels and small cells. In Molecular Neurobiology - a practical approach. Ed. Chad, J. and Wheal, H., pg3-25. IRL press, Oxford.

Fischer-Lougheed, J., Liu, J-H., Espinos, E., Mordasini, D., Bader, C. R., Belin, D. and Bemheim, L. (2001). Human myoblast fusion requires expression of functional inward rectifier Kir2.1 channels. Journal of Cell Biology. 153, 677-685.

Flagg, T. P., Tate, M., Merot, J. and Welling, P. A. (1999). A mutation linked with Bartter’s syndrome locks Kiri, la (ROMK1) channels in a closed state. Journal of General Physiology. 114, 685-700.

Forsyth, S. E., Hoger, A. and Hoger, J. H. (1997). Molecular cloning and expression of a bovine endothelial inward rectifier potassium channel. FEBS Letters. 409, 277-282.

Frazier, C. J., George, E. G. and Jones, S. W. (2000). Apparent change in ion selectivity caused by changes in intracellular K+ during whole-cell recording. Biophysical Journal. 78, 1872-1880.

Fujiwara, Y. and Kubo, Y. (2002). Serl65 in the second transmembrane region of the Kir2.1 channel determines its susceptibility to blockade by intracellular Mg2+. Journal o f General Physiology. 120, 677-692.

Fukushima, Y. (1982). Blocking kinetics of the anomalous potassium rectifier of tunicate egg studied by single channel recording. Journal of Physiology. 331, 311-331.

Ganetzky, B., Robertson, G. A., Wilson, G. F., Trudeau, M. C. and Titus, S. A. (1999). The Eag family of K+ channels in Drosophila and mammals. ANNALS o f the New York Academy of Sciences. 868, 356-369.

Ganetzky, B. and Wu, C. -F. (1983). Neurogenetic analysis of potassium currents in Drosophila: Synergistic effects on neuromuscular transmission in double mutants. Journal of Neuro genetics. 1, 17-28.

Gardos, G. (1958). The function of calcium in the potassium permeability of human erythrocytes. Biochimica et Biophysica Acta. 30, 653-654.

Gay, L. A. and Stanfield, P. R. (1977). Cs+ causes a voltage-dependent block of inward K currents in resting skeletal muscle fibres. Nature. 267, 169-170.

221 Gelman, M. S. and Kopito, R. R. (2002). Rescuing protein conformation: prospects for pharmacological therapy in cystic fibrosis. Journal of Clinical Investigations. 110, 1591- 1597.

Ghanshani, S., Coleman, M., Gustavsson, P., Chia-Ling Wu, A., Gargus, J. J., Gutman, G. A., Dahl, N., Mohrenweiser, H. and Chandy, K. G. (1998). Human calcium- activated potassium channel gene KCNN4 maps to chromosome 19ql3.2 in the region deleted in Diamond-Blackfan anemia. Genomics. 51,160-161.

Gibb, A. J. (1995). Patch-clamp recording. In Ion channels - a practical approach. Ed. Ashley, R. H., pgl-42. IRL press, Oxford.

Glowatzki, E., Fakler, G., Brandle, U., Rexhausen, U., Zenner, H. P., Ruppersberg, J. P. and Fakler, B. (1995). Subunit-dependent assembly of inward-rectifier K+ channels. Proceedings of the Royal Society of London Series B. 261, 251-261.

Goldstein, S. A. N. and Colatsky, T. J. (1996). Ion channels: Too complex for rational drug design? Neuron. 16, 913-919.

Goldstein, S. A. N., Price, L. A., Rosenthal, D. N. and Pausch, M. H. (1996). ORK1, a potassium-selective leak channel with two pore domains cloned from Drosophila melanogaster by expression in Saccharomyces cerevisiae. Proceedings of the National Academy of Sciences USA. 93, 13256-13261.

Goldstein, S. A. N., Wang, K.-W., Ilan, N. and Pausch, M. H. (1997). Sequence and function of the two P domain potassium channels: implications of an emerging superfamily. Journal of Molecular Medicine. 76, 13-20.

Gosset, P., Ghezala, G. A., Korn, B., Yaspo, M. L., Poutska, A., Lehrach, H, Sinet, P. M. and Creau, N. (1997). A new inward rectifier potassium channel gene (KCNJ15) localized on chromosome 21 in the Down syndrome chromosome region 1 (DCR1). Genomics. 44,237-41.

Greger, R., Schreiber, R., Mall, M., Wissner, A., Hopf, A., Briel, M., Bleich, M., Warth, R. and Kunzelmann, K. (2001). Cystic fibrosis and CFTR. Pfliigers Archives. 443, S3-S7.

Gribble, G. M., Ashfield, R., Ammala, C. and Ashcroft, F. M. (1997). Properties of cloned ATP-sensitive K+ currents expressed in Xenopus oocytes. Journal of Physiology. 498, 87-98.

Grissmer, S. and Cahalan, M. (1989). TEA prevents inactivation while blocking open K+ channels in human T lymphocytes. Biophysical Journal. 55, 203-206.

Gros, L., Trapp, S., Dabrowski, M., Ashcroft, F. M., Bataille, D. and Blache, P. (2002). Characterization of two novel forms of the rat sulphonylurea receptor SUR1A2 and SUR1BA31. British Journal of Pharmacology. 137,98-106.

Gulbis, J. M., Mann, S. and MacKinnon, R. (1999). Structure of a voltage-dependent K+ channel P subunit. Cell. 97, 943-952.

222 Guo, D. and Lu, Z. (2000a). Mechanism of IRK channel block by intracellular polyamines. Journal of General Physiology. 115, 799-813.

Guo, D. and Lu, Z. (2000b). Mechanism of cGMP-gated channel block by intracellular polyamines. Journal of General Physiology. 115, 783-797.

Guo, D. and Lu, Z. (2000c). Pore block versus intrinsic gating in the mechanism of inward rectification is strongly rectifying IRK1 channels. Journal of General Physiology. 116, 561-568.

Guo, D. and Lu, Z. (2002). IKR1 inward rectifier K+ channels exhibit no intrinsic rectification. Journal of General Physiology. 120,539-551.

Hagiwara, S. (1983). Membrane potential-dependent ion channels in cell membrane. Phylogenetic and developmental approaches. Distinguished lecture series of the society of general physiologists volume 3. Raven Press, New York.

Hagiwara, S. and Jaffe, L.A. (1979). Electrical properties of egg cell membrane. Annual Review of Biophysics and Bioengineering. 8, 385-416.

Hagiwara, S., Miyazaki, S., Krasne, S. and Ciani, S. (1977). Anomalous permeabilities of the egg cell membrane of a starfish in K+-T1+ mixtures. Journal of General Physiology. 70, 269-281.

Hagiwara, S., Miyazaki, S., Moody, W. and Patlak, J. (1978). Blocking effects of barium and hydrogen ions on the potassium current during anomalous rectification in the starfish egg. Journal of Physiology. 279,167-185.

Hagiwara, S., Miyazaki, S. and Rosenthal, N. P. (1976). Potassium current and the effect of cesium on this current during anomalous rectification of the egg cell membrane of a starfish. Journal of General Physiology. 67, 621-638.

Hagiwara, S. and Takahashi, K. (1974). The anomalous rectification and cation selectivity of the membrane of a starfish egg cell. Journal of Membrane Biology. 18, 61-80.

Hagiwara, S. and Yoshii, M. (1979). Effects of internal potassium and sodium on the anomalous rectification of the starfish egg as examined by internal perfusion. Journal of Physiology. 292,251-265.

Hall, A. E., Hutter, O. F. and Noble, D. (1963). Current-voltage relations of Purkinje fibres in sodium-deficient solutions. Journal of Physiology. 166, 225-240.

Hames, B.D. (1990). One-dimensional polyacrylamide gel electrophoresis. In Gel electrophoresis of proteins: a practical approach. Ed. Hames, B.D.and Rickwood, D. ppl- 147. IRL Press, Oxford.

Hamill, O. P., Marty, A., Neher, E., Sakmann, B. and Sigworth, F. J. (1981). Improved patch-clamp techniques for high-resolution current recording from cells and cell- free membrane patches. Pflugers Archives. 391,85-100.

Hanlon, M. R. and Wallace, B. A. (2002). Structure and function of voltage- dependent ion channel regulatory p subunits. Biochemistry. 41, 2886-2894.

223 Harris, R. E., Larsson, H. P. and Isacoff, E. Y. (1998). A permeant ion binding site located between two gates of the Shaker K+ channel. Biophysical Journal. 74, 1808-1820.

Hartmann, H. A., Kirsch, G. E., Drewe, J. A., Taglialatela, M., Joho, R. H. and Brown, A. M. (1991). Exchange of conduction pathways between two related K+ channels. Science. 251, 942-944.

Hayflick, L. and Moorhead, P. S. (1961). The serial cultivation of human diploid cell strains. Experimental Cell Research. 25, 585-.

Hedin, K. E., Lim, N. F. and Clapham, D. E. (1996). Cloning of a Xenopus laevis inwardly rectifying K+ channel subunit that permits GIRK1 expression of Iraqi currents in oocytes. Neuron. 16, 423-429.

Heginbotham, L., Abramson, T. and MacKinnon, R. (1992). A functional connection between the pores of distantly related ion channels as revealed by mutant K+ channels. Science. 258, 1152-1155.

Heginbotham, L., LeMasurier, M., Kolmakova-Partensky, L. and Miller, C. (1999). Single Streptomyces lividens K+ channels. Functional asymmetries and sidedness of proton activation. Journal of General Physiology. 114, 551-559.

Heginbotham, L., Lu, Z., Abramson, T. and MacKinnon, R. (1994). Mutations in the K+channel signature sequence. Biophysical Journal. 66,1061-1067.

Henry, P., Pearson, W. L. and Nichols, C. G. (1996). Protein kinase C inhibition of cloned inward rectifier (HRK1/Kir2.3) K+ channels expressed in Xenopus oocytes. Journal of Physiology. 495.3,681-688.

Hess, E. J. (1996). Identification of the weaver mouse mutation: the end of the beginning. Neuron. 16, 1073-1076.

Hestrin, S. (1981). The interaction of potassium with the activation of anomalous rectification in frog muscle membrane. Journal of Physiology. 317, 497-508.

Hibino, H., Horio, Y., Inanobe, A., Doi, K., Ito, M., Yamada, M., Gotow, T., Uchiyama, Y., Kawamura, M., Kubo, T. and Kurachi, Y. (1997). An ATP-dependent inwardly rectifying potassium channel, Kab-2 (Kir4.1), in cochlear stria vascularis of inner ear: its specific subcellular localization and correlation with the formation of endocochlear potential. Journal of Neuroscience. 17,4711-4721.

Hill, C. E., Briggs, M. M., Liu, J. and Magtanong, L. (2002). Cloning, expression, and localization of a rat hepatocyte inwardly rectifying potassium channel. American Journal of Physiology. 282, G233-G240.

Hille, B. (1973). Potassium channels in myelinated nerve. Selective permeability to small cations. Journal of General Physiology. 61, 669-686.

Hille, B. (2001). Ionic channels of excitable membranes. 3rded. Sinauer Associates, Inc., Sunderland, MA., USA.

Hille, B., Armstrong, C. M. and MacKinnon, R. (1999). Ion channels: From idea to reality. Nature Medicine. 5,1105-1109.

224 Hille, B. and Schwarz, W. (1978). Potassium channels as multi-ion single-file pores. Journal of General Physiology. 72, 409-442.

Ho, I. H. and Murrell-Lagnado, R. D. (1999). Molecular mechanism for sodium- dependent activation of G protein-gated K+ channels. Journal of Physiology. 520, 645-651.

Ho, K., Nichols, C. G., Lederer, W. J., Lytton, J., Vassilev, P. M., Kanazirska, M. V. and Hebert, S. C. (1993). Cloning and expression of an inwardly rectifying ATP-regulated potassium channel. Nature. 362,31-38.

Hodgkin, A. L. (1937a). Evidence for electrical transmission in nerve. Part I. Journal of Physiology. 90,183-210.

Hodgkin, A. L. (1937b). Evidence for electrical transmission in nerve. Part II. Journal of Physiology. 90, 211-232.

Hodgkin, A. L. and Horowicz, P. (1959). The influence of potassium and chloride ions on the membrane potential of single muscle fibres. Journal of Physiology. 148, 127- 160.

Hodgkin, A. L. and Huxley, A. F. (1939). Action potentials recorded from inside a nerve fibre. Nature. 144, 710-711.

Hodgkin, A. L. and Huxley, A. F. (1952a). Currents carried by sodium and potassium ions through the membrane of the giant axon of Loligo. Journal of Physiology. 116, 449- 472.

Hodgkin, A. L. and Huxley, A. F. (1952b). The components of membrane conductance in the giant axon of Loligo. Journal of Physiology. 116, 473-496.

Hodgkin, A. L. and Huxley, A. F. (1952c). The dual effect of membrane potential on sodium conductance in the giant axon of Loligo. Journal of Physiology. 116, 497-506.

Hodgkin, A. L. and Huxley, A. F. (1952d). A quantitative description of membrane current and its application to conduction and excitation in nerve. Journal of Physiology. 117, 500-544.

Hodgkin, A. L., Huxley, A. F. and Katz, B. (1949). Ionic currents underlying activity of the giant axon of the squid. Archives des Sciences Physiologiques. 3, 129-150.

Hodgkin, A.L. and Katz, B. (1949). The effect of sodium ions on the electrical activity of the giant axon of the squid. Journal of Physiology. 108, 37-77.

Hodgkin, A. L. and Keynes, R. D. (1955). The potassium permeability of a giant nerve fibre. Journal of Physiology. 128, 61-88.

Holmgren, M., Shin, K. S. and Yellen, G. (1998). The activation gate of a voltage- gated K+ channel can be trapped in the open state by an intrasubunit metal bridge. Neuron. 21,617-621.

225 Hone, M., Irisawa, H. and Noma, A. (1987). Voltage-dependent magnesium block of adenosine-triphosphate-sensitive potassium channel in guinea-pig ventricular cells. Journal of Physiology. 387,251-272.

Horio, Y., Hibino, H., Inanobe, A., Yamada, M., Ishii, M., Tada, Y., Satoh, E., Hata, Y., Takai, Y. and Kurachi, Y. (1997). Clustering and enhanced activity of an inwardly rectifying potassium channel, Kir4.1, by an anchoring protein, PSD-95/SAP90. Journal of Biological Chemistry. 272, 12885-12888.

Horio, Y., Morishige, K-I., Takahashi, N. and Kurachi, Y. (1996). Differential distribution of classical inwardly rectifying potassium channel mRNAs in the brain: comparison of IRK2 with IRK1 and IRK3. FEBS Letters. 379, 239-243.

Horowicz, P., Gage, P. W. and Eisenberg, R. S. (1968). The role of the electrochemical gradient in determining potassium fluxes in frog striated muscle. Journal of General Physiology. 51, 193S-203S.

Horowitz, B. (1997). Molecular cloning, characterization and assignment of smooth muscle K+ channels. Journal of Physiology. 504P, 10S.

Hoshi, T., Zagotta, W. N. and Aldrich, R. W. (1990). Biophysical and molecular mechanisms of Shaker potassium channel inactivation. Science. 250, 533-538.

Hoshi, T., Zagotta, W. N. and Aldrich, R. W. (1991). Two types of inactivation in Shaker K+ channels: Effects of alterations in the carboxy-terminal region. Neuron. 7, 547- 556.

Houpio, H., Shyng, S. L., Otonkoski, T. and Nichols, C. G. (2002). K a t p channels and insulin secretion disorders. American Journal of Physiology. 283, E207-E216.

Huang, C. L., Feng, S. and Hilgemann, D. W. (1998). Direct activation of inward rectifier potassium channels by PIP 2 and its stabilization by G(3y. Nature. 391, 803-806.

Huang, C. L., Slesinger, P. A., Casey, P. J., Jan, Y. N. and Jan, L. Y. (1995). Evidence that direct binding of GPy to the GIRK1 G protein-gated inwardly rectifying K+ channel is important for channel activation. Neuron. 15, 1133-1143.

Hughes, B. A., Kumar, G., Yuan, Y., Swaminathan, A., Yan, D., Sharma, A., Plumley, L., Yang-Feng, T. L. and Swaroop, A. (2000). Cloning and functional expression of human retinal Kir2.4, a pH-sensitive inwardly rectifying K+ channel. American Journal of Physiology. 279, C771-C784.

Hugnot, J. P., Pedeutour, F., Le Calvez, C., Grosgeorge, J., Passage, E., Fontes, M. and Lazdunski, M. (1997). The human inward rectifying K+ channel Kir2.2 (KCNJ12) gene: gene structure, assignment to chromosome 17pl 1.1, and identification of a simple tandem repeat polymorphism. Genomics. 39,113-116.

Hutter, O. F. and Noble, D. (1960). Rectifying properties of cardiac muscle. Nature. 188,495.

226 Ibarra, J., Morley, G. E. and Delmar, M. (1991). Dynamics of the inward rectifier K+ current during the action potential of guinea pig ventricular myocytes. Biophysical Journal. 60, 1534-1539.

Imoto, Y., Ehara, T. and Matsuura, H. (1987). Voltage- and time-dependent block of Iki underlying Ba2+-induced ventricular automaticity. American Journal of Physiology. 252, H325-333.

Inagaki, N., Gonoi, T., Clement IV, J. P., Namba, N., Inazawa, J., Gonzalez, G., Aguilar-Bryan, L., Seino, S. and Bryan, J. (1995c). Reconstitution of I k a t p - An inward rectifier subunit plus the sulfonylurea receptor. Science. 270, 1166-1170.

Inagaki, N., Gonoi, T., Clement IV, J. P., Wang, C-Z., Aguilar-Bryan, L., Bryan, J. and Seino, S. (1996). A family of sulfonylurea receptors determines the pharmacological properties of ATP-sensitive K+ channels. Neuron. 16, 1011-1017.

Inagaki, N., Gonoi, T. and Seino, S. (1997). Subunit stoichiometry of the pancreatic beta-cell ATP-sensitive K+ channel. FEBS Letters. 409, 232-236.

Inagaki, N., Inazawa, J. and Seino, S. (1995b). cDNA sequence, gene structure, and chromosomal localization of the human ATP-sensitive potassium channel, uK a t p - i , gene (KCNJ8). Genomics. 30, 102-104.

Inagaki, N., Tsuura, Y., Namba, N., Masuda, K., Gonoi, T., Horie, M., Seino, Y., Mizuta, M. and Seino, S. (1995a). Cloning and functional characterization of a novel ATP- sensitive potassium channel ubiquitously expressed in rat tissues, including pancreatic islets, pituitary, skeletal muscle, and heart. Journal of Biological Chemistry. 270, 5691-5694.

Inanobe, A., Horio, Y., Fujita, A., Tanemoto, M., Hibino, H., Inageda, K. and Kurachi, Y. (1999b). Molecular cloning and characterization of a novel splicing variant of the Kir3.2 subunit predominantly expressed in mouse testis. Journal of Physiology. 521.1, 19-30.

Inanobe, A., Ito, H., Ito, M., Hosya, Y. and Kurachi, Y. (1995). Immunological and physical characterization of the brain G protein-gated muscarinic potassium channel. Biochemical and Biophysical Research Communications. 217, 1238-1244.

Inanobe, A., Yoshimoto, Y., Horio, Y., Morishige, K-I., Hibino, H., Matsumoto, S., Tokunaga, Y., Maeda, T., Hata, Y., Takai, Y. and Kurachi, Y. (1999a). Characterization of G-protein-gated K+ channels composed of Kir3.2 subunits in dopaminergic neurons of the substantia nigra. Journal of Neuroscience. 19,1006-1017.

Innis, R. B. and Aghajanian, G. K. (1987). Pertussis toxin blocks autoreceptor- mediated inhibition of dopaminergic neurons in rat substantia nigra. Brain Research. 411, 139-143.

Isacoff, E. Y., Jan, Y. N., and Jan, L. Y. (1990). Evidence for the formation of heteromultimeric potassium channels in Xenopus oocytes. Nature. 345, 530-534.

Isacoff, E. Y., Jan, Y. N. and Jan, L. Y. (1991). Putative receptor for the cytoplasmic inactivation gate in the Shaker K+ channel. Nature. 353, 86-90.

227 Ishihara, K. and Hiraoka, M. (1994). Gating mechanism of the cloned inward rectifier potassium channel from mouse heart. Journal of Membrane Biology. 142, 55-64.

Ishihara, K., Hiraoka, M. and Ochi, R. (1996). The tetravalent organic cation spermine causes the gating of the IRK1 channel expressed in murine fibroblast cells. Journal of Physiology. 491, 367-381.

Ishihara, K., Mitsuiye, T., Noma, A. and Takano, M. (1989). The Mg2+ block and intrinsic gating underlying inward rectification of the K+ current in guinea-pig cardiac myocytes. Journal of Physiology. 419, 297-320.

Ishii, T. M., Silvia, C., Hirschberg, B., Bond, C. T., Adelman, J. P. and Maylie, J. (1997). A human intermediate conductance calcium-activated potassium channel. Proceedings of the National Academy of Sciences USA. 94, 11651-11656.

Ishii, K., Yamagishi, T. and Taira, N. (1994). Cloning and functional expression of a cardiac inward rectifier K+ channel. FEBS Letters. 338, 107-111.

Isomoto, S., Kondo, C., Takahashi, N., Matsumoto, S., Yamada, M., Takumi, T., Horio, Y. and Kurachi, Y. (1996a). A novel ubiquitously distributed isoform of GIRK2 (GIRK2B) enhances GIRK1 expression of the G-protein-gated K+ current in Xenopus oocytes. Biochemical and Biophysical Research Communications. 218, 286-291.

Isomoto, S., Kondo, C., Yamada, M., Matsumoto, S., Higashiguchi, O., Horio, Y., Matsuzawa, Y. and Kurachi, Y. (1996b). A novel sulfonylurea receptor forms with BIR (Kir6.2) a smooth muscle type ATP-sensitive K+ channel. Journal of Biological Chemistry. 271, 24321-24324.

Isomoto, S., Kondo, C. and Kurachi, Y. (1997). Inwardly rectifying potassium channels: their molecular heterogeneity and function. Japanese Journal of Physiology. 47, 11-39.

Ito, M., Inanobe, A., Horio, Y., Hibeno, H., Isomoto, S., Ito, H., Mori, K., Tonosaki, A., Tomoike, H. and Kurachi, Y. (1996). Immunolocalization of an inwardly rectifying K+ channel, KA b - 2 (Kir4.1), in the basolateral membrane of renal distal tubular epithelia. FEBS Letters. 388, 11-15.

Iverson, L. and Rudy, B. (1990). The role of the divergent amino and carboxyl domains on the inactivation properties of potassium channels derived from the Shaker gene of Drosophila. Journal of Neuroscience. 10, 2903-2916.

Jan, L. Y. and Jan, Y. N. (1994). Potassium channels and their evolving gates. Nature. 371, 119-122.

Jan, L. Y and Jan, Y. N. (1997). Cloned potassium channels from eukaryotes and prokaryotes. Annual Review of Neuroscience. 20,91-123.

Jan, L. Y. and Jan, Y. N. (2000). Heartfelt crosstalk: desensitization of the GIRK current. Nature Cell Biology. 2, E l 65-El 67.

Jan, Y. N., Jan, L. Y. and Dennis, M. J. (1977). Two mutations of synaptic transmission in Drosophila. Proceedings of the Royal Society of London Series B. 198, 87- 108.

228 Jelacic, T. M., Kennedy, M. E., Wickman, K. and Clapham, D. E. (2000). Functional and biochemical evidence for G-protein-gated inwardly rectifying K+ (GIRK) channels composed of GIRK2 and GIRK3. Journal of Biological Chemistry. 275, 36211-36216.

Jentsch, T. J. (2000). Neuronal KCNQ potassium channels: physiology and role in disease. Nature Reviews Neuroscience. 1,21-30.

Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B. T. and MacKinnon, R. (2002a). Crystal structure and mechanism of a calcium-gated potassium channel. Nature. 417, 515- 522.

Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B. T. and MacKinnon, R. (2002b). The open pore conformation of potassium channels. Nature. 417, 523-526.

Jiang, Y and MacKinnon, R. (2000). The barium site in a potassium channel by X- ray crystallography. Journal of General Physiology. 115, 269-272.

Jin, W. and Lu, Z. (1998). A novel high-affinity inhibitor for inward-rectifier K+ channels. Biochemistry. 37, 13291-13299.

Joiner, W. J., Tang, M. D., Wang, L-Y., Dworetzky, S. I., Boissard, C. G., Gan, L., Gribkoff, V. K. and Kaczmarek, L. K. (1998). Formation of intermediate-conductance calcium-activated potassium channels by interaction of Slack and Slo subunits. Nature Neuroscience. 1, 462-469.

Joiner, W. J., Wang, L-Y., Tang, M. D. and Kaczmarek, L. K. (1997). hSK4, a member of a novel subfamily of calcium-activated potassium channels. Proceedings o f the National Academy of Sciences USA. 94, 11013-11018.

Kaczorowski, G. J. and Garcia, M. L. (1999). Pharmacology of voltage-gated and calcium-activated potassium channels. Current Opinion in Chemical Biology. 3, 448-458.

Kamb, A., Iverson, L. E. and Tanouye, M. A. (1987). Molecular characterization of Shaker, a Drosophila gene that encodes a potassium channel. Cell. 50, 405-413.

Kaplan, W. D. and Trout, W. E. (1969). The behavior of four neurological mutants of Drosophila. Genetics. 61, 399-409.

Karschin, C., DiBmann, E., Sttihmer, W. and Karschin, A. (1996). IRK(l-3) and GIRK(l-4) inwardly rectifying K+ channel mRNAs are differentially expressed in the adult rat brain. Journal of Neuroscience. 16, 3559-3570.

Karschin, C. and Karschin, A. (1997). Ontogeny of gene expression of Kir channel subunits in the rat. Molecular and Cellular Neuroscience. 10, 131-148.

Karschin, C., Schreibmayer, W., Dascal, N., Lester, H., Davidson, H and Karschin, A. (1994). Distribution and localization of a G protein-coupled inwardly rectifying K+ channel in the rat. FEBS Letters. 348,139-144.

Karschin, C., Wischmeyer, E., Preisig-Muller, R., Rajan, S., Derst, C., Grzeschik, K- H., Daut, J. and Karschin, A. (2001). Expression pattern in brain of TASK-1, TASK-3, and a

229 tandem pore domain K+ channel subunit, TASK-5, associated with the central auditory nervous system. Molecular and Cellular Neuroscience. 18,632-648.

Katz, B. (1949). Les constantes electriques de la membrane du muscle. Archives des Sciences Physiologiques. 3, 285-299.

Kennedy, M. E., Nemec, J., Corey, S., Wickman, K. and Clapham, D. E. (1999). GIRK4 confers appropriate processing and cell surface localization to G-protein-gated potassium channels. Journal of Biological Chemistry. 274, 2571-2582.

Keren-Raifman, T., Ivanina, T., Bismuth, Y. and Dascal, N. (2000). Expression cloning of KCRF, a potassium channel regulatory factor. Biochemical and Biophysical Research Communications. 274, 852-858.

Ketchum, K. A., Joiner, W. J., Sellers, A. J., Kaczmarek, L. K. and Goldstein, S. A. N. (1995). A new family of outwardly rectifying potassium channel proteins with two pore domains in tandem. Nature. 376, 690-695.

Keynes, R. D. (1951). The ionic movements during nervous activity. Journal of Physiology. 114, 119-150.

Khanna, R., Chang, M. C., Joiner, W. J., Kaczmarek, L. K. and Schlichter, L. C. (1999). hSK4/kIKl, a calmodulin-binding Kca channel in human T-lymphocytes. Roles in proliferation and volume regulation. Journal of Biological Chemistry. 274, 14838-14849.

Kim, D. and Gnatenco, C. (2001). TASK-5, a new member of the tandem-pore K+ channel family. Biochemical and Biophysical Research Communications. 284, 923-930.

Kim, S. J., Kerst, G., Schreiber, R., Pavenstadt, H., Greger, R., Hug, M. G. and Bleich, M. (2000). Inwardly rectifying K+ channels in the basolateral membrane of rat pancreatic acini. Pfliigers Archives. 441,331-340.

Kirsch, G. E., Drewe, J. A., Taglialatela, M., Joho, R. H., DeBiasi, M., Hartmann, H. A. and Brown, A. M. (1992). A single nonpolar residue in the deep pore of related K+ channels acts as a K+:Rb+ conductance switch. Biophysical Journal. 62, 136-144.

Kiss, L. and Korn, S. J. (1998). Modulation of C-type inactivation by K+ at the potassium channel selectivity filter. Biophysical Journal. 74, 1849-1849.

Klein, H., Gameau, L., Coady, M., Lemay, G., Lapointe, J. Y. and Sauve, R. (1999). Molecular characterization of an inwardly rectifying K+ channel from HeLa cells. Journal of Membrane Biology. 167, 43-52.

Knot, H. J., Zimmermann, P. A. and Nelson, M. T. (1996). Extracellular K+-induced hyperpolarizations and dilations of rat coronary and cerebral arteries involved inward rectifier K+ channels. Journal of Physiology. 492, 419-430.

Kobayashi, T., Ikeda, K., Ichikawa, T., Abe, S., Togashi, S and Kumanishi, T. (1995). Molecular cloning of a mouse G-protein-activated K+ channel (mGIRKl) and distinct distributions of three GIRK (GIRK1, 2 and 3) mRNAs in mouse brain. Biochemical and Biophysical Research Communications. 208, 1166-1173.

230 Kofuji, P., Davidson, N. and Lester, H. A. (1995). Evidence that neuronal G-protein- gated inwardly rectifying K+ channels are activated by Gfy subunits and function as heteromultimers. Proceedings of the National Academy of Sciences USA. 92, 6542-6546.

Kofuji, P., Hofer, M., Millen, K. J., Millonig, J. H., Davidson, N., Lester, H. A. and Hatten, M. E. (1996). Functional analysis of the weaver mutant GIRK2 K+ channel and rescue of weaver granule cells. Neuron. 16, 941-952.

Kohler, M., Hirschberg, B., Bond, C. T., Kinzie, J. M., Marrion, N. V., Maylie, J. and Adelman, J. P. (1996). Small-conductance, calcium-activated potassium channels from mammalian brain. Science. 273, 1709-1714.

Kondo, C., Isomoto, S., Matsumoto, S., Yamada, M., Horio, Y., Yamashita, S., Takemura-Kameda, K., Matsuzawa, Y. and Kurachi, Y. (1996). Cloning and functional expression of a novel isoform of ROMK inwardly rectifying ATP-dependent K+ channel, ROMK6 (Kiri.If). FEBS Letters. 399, 122-126.

Kondo, C., Repunte, V. P., Satoh, E., Yamada, M., Horio, Y., Matsuzawa, Y., Pott, L. and Kurachi, Y. (1998). Chimeras of Kir6.1 and Kir6.2 reveal structural elements involved in spontaneous opening and unitary conductance of the ATP-sensitive K+ channels. Receptors and Channels. 6, 129-140.

Konstas, A-A., Korbmacher, C. and Tucker, S. J. (2002). Identification of domains which control the heteromeric assembly of Kir5.1/Kir4.0 potassium channels. American Journal of Physiology. 284, C910-C917.

Kovoor, A. and Lester, H. A. (2002). Gi Irks GIRKs. Neuron. 33, 6-8.

Korovkina, V. P. and England, S. K. (2002). Molecular diversity of vascular potassium channel isoforms. Clinical and Experimental Pharmacology and Physiology. 29, 317-323.

Koyama, H., Morishige, K-I., Takahashi, N., Zanelli, J. S., Fass, D. N. & Kurachi, Y. (1994). Molecular cloning, functional expression and localization of a novel inward rectifier potassium channel in the rat brain. FEBS Letters. 341, 303-307.

Krapivinsky, G., Gordon, E. A., Wickman, K., Velimirovic, B., Krapivinsky, L. and Clapham, D. E. (1995a). The G-protein-gated atrial K+ channel iKAChis a heteromultimer of two inwardly rectifying K+-channel proteins. Nature. 374,135-141.

Krapivinsky, G., Kennedy, M. E., Nemec, J., Medina, I., Krapivinsky, L. and Clapham, D. E. (1998b). G(3 binding to GIRK4 subunit is critical for G protein-gated K+ channel activation. Journal of Biological Chemistry. 273, 16946-16952.

Krapivinsky, G., Krapivinsky, L., Wickman, K. and Clapham, D. E. (1995b). GPy

binds directly to the G protein-gated K+ channel, I r a q i - Journal of Biological Chemistry. 270, 29059-29062.

Krapivinsky, G., Medina, I., Eng, L., Krapivinsky, L., Yang, Y. and Clapham, D.E. (1998a). A novel inward rectifier K+ channel with unique pore properties. Neuron. 20, 995- 1005.

231 Kreusch, A., Pfaffinger, P. J., Stevens, C. F. and Choe, S. (1998). Crystal structure of the tetramerization domain of the Shaker potassium channel. Nature. 392, 945-948.

Kubo, Y., Baldwin, T. J., Jan, Y. N. and Jan, L. Y. (1993a). Primary structure and functional expression of a mouse inward rectifier potassium channel. Nature, 362, 127-133.

Kubo, Y., Miyashita, T. and Kubokawa, K. (1996). A weakly inward rectifying potassium channel of the salmon brain. Journal of Biological Chemistry. 271, 15729-15735.

Kubo, Y. and Murata, Y. (2001). Control of rectification and permeation by two distinct sites after the second transmembrane region in Kir2.1 K+ channel. Journal of Physiology. 531.3, 645-660.

Kubo, Y., Reuveny, E., Slesinger, P. A., Jan, Y. N. and Jan, L. Y. (1993b). Primary structure and functional expression of a rat G-protein-coupled muscarinic potassium channel. Nature. 364, 802-806.

Kubo, Y., Yoshimichi, M. and Heinemann, S. H. (1998). Probing pore topology and conformational changes of Kir2.1 potassium channels by cysteine scanning mutagenesis. FEBS Letters. 435, 69-73.

Kukuljan, M., Labarca, P. and Latorre, R. (1995). Molecular determinants of ion conduction and inactivation in K+ channels. American Journal of Physiology. 268, C535- C556.

Kumpf, R. A. and Dougherty, D. A. (1993). A mechanism for ion selectivity in potassium channels: Computational studies of cation-7i interactions. Science. 261, 1708- 1710.

Kunkel, T. A. (1985). Rapid and efficient site-specific mutagenesis without phenotype selection. Proceedings of the National Academy of Sciences USA. 82, 488-492.

Kunkel, M. T. and Peralta, E. G. (1995). Identification of domains conferring G protein regulation on inward rectifier potassium channels. Cell. 83, 443-449.

Kurachi, Y., Ito, H., Sugimoto, T., Katada, T. and Ui, M. (1989). Activation of atrial muscarinic K+ channels by low concentrations of (3y subunits of rat brain G protein. Pfliigers Archives. 413, 325-327.

Kurschner, C., Mermelstein, P. G., Holden, W. T. and Surmeier, D. J. (1998). CIPP, a novel multivalent PDZ domain protein, selectively interacts with Kir4.0 family members, NMDA receptor subunits, neurexins, and neuroligins. Molecular and Cellular Neuroscience. 11, 161-172.

Kuryshev, Y. A., Haak, L., Childs, G. V. and Ritchie, A. K. (1997). Corticotrophin releasing hormone inhibits an inwardly rectifying potassium current in rat corticotropes. Journal of Physiology. 502.2, 265-279.

Kusaka, S., Inanobe, A., Fujita, A., Makino, Y., Tanemoto, M., Matsushita, K., Tano, Y. and Kurachi, Y. (2001). Functional Kir7.1 channels localized at the root of apical processes in rat retinal pigment epithelium. Journal of Physiology. 531.1, 27-36.

232 Lacey, M. G., Mercuri, N. B. and North, R. A. (1988). On the potassium conductance increase activated by GABAB and dopamine D2 receptors in rat substantia nigra neurones. Journal of Physiology. 401, 437-453.

Lancaster, M. K., Dibb, K. M., Quinn, C. C., Leach, R., Lee, J. K., Findlay, J. B. C. and Boyett, M. R. (2000). Residues and mechanisms for slow activation and Ba + block of the cardiac muscarinic K+ channel, Kir3.1/Kir3.4. Journal of Biological Chemistry. 275, 35831-35839.

Lee, W. S. and Hebert, S. C. (1995). ROMK inwardly rectifying ATP-sensitive K+ channel: I. expression in rat distal nephron segments. American Journal of Physiology. 268, F1124-F1131.

Lee, J-K., John, S. A. and Weiss, J. N. (1999). Novel gating mechanism of polyamine block in the strong inward rectifier K channel Kir2.1. Journal of General Physiology. 113, 555-563.

Leech, C. A. and Stanfield, P. R. (1981). Inward rectification in frog skeletal muscle fibres and its dependence on membrane potential and external potassium. Journal of Physiology. 319, 295-309.

Leonoudakis, D., Mailliard, W. S., Wingerd, K. L., Clegg, D. O. and Vandenberg, C. A. (2001). Inward rectifier potassium channel Kir2.2 is associated with Synapse-Associated Protein SAP97. Journal of Cell Science. 114,987-998.

Lesage, F., Duprat, F., Fink, M., Guillemare, E., Coppola, T., Lazdunski, M. and Hugnot, J. P. (1994). Cloning provides evidence for a family of inward rectifier and G- protein coupled K+ channels in the brain. FEBS Letters. 353, 37-42.

Lesage, F., Guillemare, E., Fink, M., Duprat, F., Heurteaux, C., Fosset, M., Romey, G., Barhanin, J. and Lazdunski, M. (1995). Molecular properties of neuronal G-protein- activated inwardly rectifying K+ channels. Journal of Biological Chemistry. 270, 28660- 28667.

Lesage, F., Guillemare, E., Fink, M., Duprat, F., Lazdunski, M., Romey, G. and Barhanin, J. (1996a). A pH-sensitive yeast outward rectifier K+ channel with two pore domains and novel gating properties. Journal of Biological Chemistry. 271, 4183-4187.

Lesage, F., Guillemare, E., Fink, M., Duprat, F., Lazdunski, M., Romey, G. and Barhanin, J. (1996b). TWIK-1, a ubiquitous human weakly inward rectifying K+ channel with a novel structure. EMBO Journal. 15, 1004-1011.

Lesage, F. and Lazdunski, M. (2000). Molecular and functional properties of two- pore-domain potassium channels. American Journal of Physiology. 279, F793-F801.

Lesage, F., Reyes, R., Fink, M., Duprat, F., Guillemare, E. and Lazdunski, M. (1996c). Dimerization of TWIK-1 K+ channel subunits via a disulfide bridge. EMBO Journal. 15, 6400-6407.

Lewis, D. L., Ikeda, S. R., Aryee, D. and Joho, R. H. (1991). Expression of an inwardly rectifying K+ channel from rat basophilic leukemia cell mRNA in Xenopus oocytes. FEBS Letters. 290, 17-21.

233 Leyland, M. L., Dart, C., Spencer, P. J., Sutcliffe, M. J. and Stanfield, P. R. (1999). The possible role of a disulphide bond in forming functional Kir2.1 potassium channels. Pfliigers Archives . 438, 778-781.

Li, M., Jan, Y. N. and Jan, L. Y. (1992). Specification of subunit assembly by the hydrophilic amino-terminal domain of the Shaker potassium channel. Science. 257, 1225- 1230.

Li, M., Unwin, N., Stauffer, K. A., Jan, Y. N., and Jan, L. Y. (1994). Images of purified Shaker potassium channels. Current Biology. 4, 110-115.

Liao, Y. J., Jan, Y. N., and Jan, L. Y. (1996). Heteromultimerization of G-protein- gated inwardly rectifying K+ channel proteins GIRK1 and GIRK2 and their altered expression in weaver brain. Journal of Neuroscience. 16,7137-7150.

Liman, E. R., Tytgat, J. and Hess, P. (1992). Subunit stoichiometry of a mammalian K+ channel determined by construction of multimeric cDNAs. Neuron. 9, 861-871.

Liou, H. H., Zhou, S. S. and Huang, C. L. (1999). Regulation of ROMK1 channel by protein kinase A via a phosphatidylinositol 4, 5-bisphosphate-dependent mechanism. Proceedings of the National Academy of Sciences USA. 96, 5820-5825.

Liu, J. H., Bijlenga, P., Fischer-Lougheed, J., Occhiodoro, T., Kaelin, A., Bader, C. R. and Bemheim, L. (1998). Role of an inward rectifier K+ current and of hyperpolarization in human myoblast fusion. Journal of Physiology. 510.2, 467-476.

Liu, Z., Cashion, L. M. and Twu, J. J. (1997). A systematic comparison of relative promoter/enhancer activities in mammalian cell lines. Analytical Biochemistry. 246, 150- 152.

Liu, G. X., Derst, C., Schlichthorl, G., Heinen, S., Seebohm, G., Briiggemann, A., Kummer, W., Veh, R. W., Daut, J. and Preisig-Miiller, R. (2001a). Comparison of cloned Kir2 channels with native inward rectifier K+ channels from guinea-pig cardiomyocytes. Journal of Physiology. 532.1, 115-126.

Liu, Y., Jurman, M. E. and Yellen, G. (1996). Dynamic rearrangement of the outer mouth of a K+channel during gating. Neuron. 16, 859-867.

Liu, Y., Liu, D., Heath, L., Meyers, D. M., Krafte, D. S., Wagoner, P. K., Silvia, C. P., Yu, W. and Curran, M. E. (2001b). Direct activation of an inwardly rectifying potassium channel by arachidonic acid. Molecular Pharmacology. 59,1061-1068.

Loffler, K and Hunter, M. (1997). Cation permeation and blockade of ROMK1, a cloned renal potassium channel. Pfliigers Archives. 434,151-158.

Logothetis, D. E., Kurachi, Y., Galper, J., Neer, E. J. and Clapham, D. E. (1987). The (3y subunits of GTP-binding proteins activate the muscarinic K+ channel in heart. Nature. 325, 321-326.

Logsdon, N. J., Kang, J., Togo, J. A., Christian, E. P., and Aiyar, J. (1997). A novel gene, hKCa4, encodes the calcium-activated potassium channel in human T lymphocytes. Journal of Biological Chemistry. 272, 32723-32726.

234 Lopatin, A. N., Makhina, E. N. and Nichols, C. G. (1994). Potassium channel block by cytoplasmic polyamines as the mechanism of inward rectification. Nature. 372, 366-369.

Lopatin, A. N., Makhina, E. N. and Nichols, C. G. (1995). The mechanism of inward rectification of potassium channels: ‘Long-pore plugging’ by cytoplasmic polyamines. Journal of General Physiology. 106, 923-955.

Lopatin, A. N., and Nichols, C. G. (1996). [K+] dependence of polyamine-induced rectification in inward rectifier potassium channels (IRK1, Kir2.1). Journal of General Physiology. 108, 105-113.

Lopatin, A. N. and Nichols, C. G. (2001). Inward rectifiers in the heart: an update on IK1. Journal of Molecular and Cellular Cardiology. 33, 625-638.

Lopez-Bameo, J., Hoshi, T., Heinemann, S. H. and Aldrich, R. W. (1993). Effects of external cations and mutations in the pore region on C-type inactivation of Shaker potassium channels. Receptors and Channels. 1, 66-71.

Lopez, G. A., Jan, Y. N. and Jan, L. Y. (1994). Evidence that the S6 segment of the Shaker voltage-gated K+ channel comprises part of the pore. Nature. 367, 179-182.

Lorenz, J. N., Baird, N. R., Judd, L. M., Noonan, W. T., Andringa, A., Doetschman, T., Manning, P. A., Liu, L. H., Miller, M. L. and Shull, G. E. (2002). Impaired renal NaCl absorption in mice lacking the ROMK potassium channel, a model for type II Bartter’s syndrome. Journal of Biological Chemistry. 277, 37871-37880.

Lourdel, S., Paulais, M., Cluzeaud, F., Bens, M., Tanemoto, M., Kurachi, Y., Vandewalle, A. and Teulon, J. (2002). An inward rectifier K+ channel at the basolateral membrane of the mouse distal convoluted tubule: similarities with Kir4-Kir5.1 heteromeric channels. Journal of Physiology. 583.2, 391-404.

Lovelock, J. E. and Bishop, M. W. H. (1959). Prevention of freezing damage to living cells by dimethyl sulphoxide. Nature. 183, 1394-1395.

Lu, Z., Klem, A. M. and Ramu, Y. (2001b). Ion conduction pore is conserved among potassium channels. Nature. 413,809-813.

Lu, Z. and MacKinnon, R. (1994). Electrostatic tuning of Mg2+ affinity in an inward rectifier K+ channel. Nature. 371, 243-249.

Lu, Z. and MacKinnon, R. (1995). Probing a potassium channel pore with an engineered protonatable site. Biochemistry. 34,13133-13138.

Lu, Z. and MacKinnon, R. (1997). Purification, characterization and synthesis of an inward-rectifier K+ channel inhibitor from scorpion venom. Biochemistry. 36, 6936-6940.

Lii, Q. and Miller, C. (1995). Silver as a probe of pore-forming residues in a potassium channel. Science. 268, 304-307.

Lu, T., Ting, A. Y., Mainland, J., Jan, L. J., Schultz, P. G. and Yang, J. (2001a). Probing ion permeation and gating in a K+ channel with backbone mutations in the selectivity filter. Nature Neuroscience. 4, 239-246.

235 Lu, M., Wang, T., Yan, Q., Yang, X., Dong, K., Knepper, M. A., Wang, W., Giebisch, G., Shull, G. E. and Hebert, S. C. (2002). Absence of small conductance K+ channel (SK) activity in apical membranes of thick ascending limb and cortical collecting duct in ROMK (Bartter’s) knockout mice. Journal of Biological Chemistry. 277, 37881-37887.

Lu, T., Zhu, Y-G. and Yang, J. (1999b). Cytoplasmic amino and carboxyl domains form a wide intracellular vestibule in an inwardly rectifying potassium channel. Proceedings o f the National Academy o f Sciences USA. 96, 9926-9931.

MacDonald, C. (1991). Genetic engineering of animal cells. In Mammalian cell biotechnology: A practical approach. Ed. Butler, M. pp57-83. IRL Press, Oxford.

MacKinnon, R. (1991). Determination of the subunit stoichiometry of a voltage- activated potassium channel. Nature. 350, 232-235.

MacKinnon, R., Cohen, S. L., Kuo, A., Lee, A. and Chait, B. T. (1998). Structural conservation in prokaryotic and eukaryotic potassium channels. Science. 280, 106-109.

MacKinnon, R. and Miller, C. (1989). Mutant potassium channels with altered binding of charybdotoxin, a pore-blocking peptide inhibitor. Science. 245, 1382-1385.

MacKinnon, R. and Yellen, G. (1990). Mutations affecting TEA blockage and ion permeation in voltage-activated K+ channels. Science. 250, 276-279.

Makhina, E. N., Kelly, A. J., Lopatin, A. N., Mercer, R. W. and Nichols, C. G. (1994). Cloning and expression of a novel human brain inward rectifier potassium channel. Journal of Biological Chemistry. 269, 20468-20474.

Marrion, N. V. (1997). Control of M-current. Annual Review of Physiology. 59, 483-504.

Marshall, J., Molloy, R., Moss, G. W. J., Howe, J. R. and Hughes, T. E. (1995). The Jellyfish green fluorescent protein: a new tool for studying ion channel expression and function. Neuron. 14,211-215.

Martens, J. R., Kwak, Y-G. and Tamkun, M. M. (1999). Modulation of Kv channel ot/|3 subunit interactions. Trends in Cardiovascular Medicine. 9, 253-258.

Marty, A. (1981). Ca-dependent K channels with large unitary conductance in chromaffin cell membranes. Nature. 291, 497-500.

Mathie, A., Wooltorton, J. R. A. and Watkins, C. S. (1998). Voltage-activated potassium channels in mammalian neurons and their block by novel pharmacological agents. General Pharmacology. 30, 13-24.

Matsuda, H. (1991). Magnesium gating of the inwardly rectifying K+ channel. Annual Review of Physiology. 53, 289-298.

Matsuda, H., Matsuura, H. and Noma, A. (1989). Triple-barrel structure of inwardly rectifying K+ channels revealed by Cs+ and Rb+ block in guinea-pig heart cells. Journal of Physiology. 413, 139-157.

236 Matsuda, H., Oishi, K. and Omori, K. (2003). Voltage-dependent gating and block by internal spermine of the murine inwardly rectifying K+ channel, Kir2.1. Journal o f Physiology. 548.2,361-371.

Matsuda, H., Saigusa, A. and Irisawa, H. (1987). Ohmic conductance through the inwardly rectifying K channel and blocking by internal Mg2+. Nature. 325, 156-159.

McKillen, H. C. (1994). A study of the ATP-dependent potassium channel of rat skeletal muscle. PhD thesis.

McKinney, L. C. and Gallin, E. K. (1988). Inwardly rectifying whole-cell and single­ channel K currents in the murine macrophage cell line J774.1. Journal of Membrane Biology. 103,41-53.

McNicholas, C. M., Guggino, W. B., Schwiebert, E. M., Hebert, S. C., Giebisch, G. and Egan, M. E. (1996). Sensitivity of a renal K channel (ROMK2) to the inhibitory sulfonylurea compound glibenclamide is enhanced by coexpression with the ATP-binding cassette transporter cystic fibrosis transmembrane regulator. Proceedings of the National Academy of Sciences USA. 93, 8083-8088.

McNicholas, C. M., MacGregor, G. G., Islas, L. D., Yang, Y., Hebert, S. C. and Giebisch, G. (1998). pH-dependent modulation of the cloned renal K+ channel, ROMK. American Journal of Physiology. 215, F972-F981.

McNicholas, C. M., Wang, W., Ho, K., Hebert, S. C. and Giebisch, G. (1994). Regulation of ROMK1 K+ channel activity involves phosphorylation processes. Proceedings of the National Academy of Sciences USA. 91, 8077-8081.

Meech, R. W. and Strumwasser, F. (1970). Intracellular calcium injection activates potassium conductance in Aplysia nerve cells. Federation Proceedings. 29, 834.

Meera, P., Wallner, M., Jiang, Z. and Toro, L. (1996). A calcium switch for the functional coupling between a ( hslo ) and p subunits (Kv,Ca|J) of MaxiK channels. FEBS Letters. 385, 127-128.

Melynk, P., Zhang, L., Shrier, A. and Nattel, S. (2002). Differential distribution of Kir2.1 and Kir2.3 subunits in canine atrium and ventricle. American Journal of Physiology. 283, H1123-H1133.

Miller, C. (2000). An overview of the potassium channel family. Genome Biology. 1, 0004.1-0004.5.

Minor, D. L., Masseling, S. J., Jan, Y. N. and Jan, L. Y. (1999). Transmembrane structure of an inwardly rectifying potassium channel. Cell. 96, 879-891.

Miyazaki, S-I., Ohmori, H. and Sasaki, S. (1975). Potassium rectifications of the starfish oocyte membrane and their changes during oocyte maturation. Journal of Physiology. 246, 55-78.

Miyazaki, S-I., Takahashi, K. and Tsuda,, K. (1974a). Electrical excitability in the egg cell membrane of the tunicate. Journal of Physiology. 238, 37-54.

237 Miyazaki, S-I., Takahashi, K., Tsuda, K. and Yoshii, M. (1974b). Analysis of non- linearity observed in the current-voltage relation of the tunicate embryo. Journal of Physiology. 238, 55-77.

Molina, A., Castellano, A. G. and Lopez-Bameo, J. (1997). Pore mutations in Shaker K+ channels distinguish between the sites of tetraethylammonium blockade and C-type inactivation. Journal of Physiology. 499.2, 361-367.

Morais-Cabral, J. H., Zhou, Y. and MacKinnon, R. (2001). Energetic optimization of ion conduction rate by the K+ selectivity filter. Nature. 414, 37-42.

Morishige, K. I., Takahashi, N., Findlay, I., Koyama, H., Zanelli, J. S., Peterson, C., Jenkins, N. A., Copeland, N. G., Mori, N. and Kurachi, Y. (1993). Molecular cloning, functional expression and localization of an inward rectifier potassium channel in the mouse brain. FEBS Letters. 336, 375-380.

Morishige, K-I., Takahashi, N., Jahangir, A., Yamada, M., Koyama, H., Zanelli, J. S. and Kurachi, Y. (1994). Molecular cloning and functional expression of a novel brain- specific inward rectifier potassium channel. FEBS Letters. 346, 251-256.

Mouri, T., Kittaka, N., Horio, Y., Copeland, N. G., Gilbert, D. J., Jenkins, N. A. and Kurachi, Y. (1998). Assignment of mouse inwardly rectifying potassium channel Kcnjl6 to the distal region of mouse . Genomics. 54, 181-182.

Murata, Y., Fujiwara, Y. and Kubo, Y. (2002). Identification of a site involved in the block by extracellular Mg2+ and Ba2+ as well as permeation of K+ in the Kir2.1 K+ channel. Journal of Physiology. 544.3, 665-677.

Murrell-Lagnado, R. U. and Aldrich, R. W. (1993a). Interactions of the amino terminal domains of the shaker channels with a pore blocking site studied with synthetic peptides. Journal of General Physiology. 102, 949-975.

Murrell-Lagnado, R. U. and Aldrich, R. W. (1993b). Energetics of Shaker K+ channels blocking by inactivation peptides. Journal of General Physiology. 102, 977-1003.

Nagashima, M., Ishii, K., Tohse, N., Taira, N. and Yabu, H. (1996). Unitary current through the inward rectifier K+ channel cloned from rabbit heart-comparison with the native K+ channel. Journal of Molecular Cell Cardiology. 28, 957-965.

Nakajima, S., Iwasaki, S. and Obata, K. (1962). Delayed rectification and anomalous rectification in frog’s skeletal muscle membrane. Journal of General Physiology. 46, 97- 115.

Nakamura, T. Y., Artman, M., Rudy, B. and Coetzee, W. A. (1998). Inhibition of rat ventricular /ki with antisense oligonucleotides targeted to Kir2.1 mRNA. American Journal of Physiology. 274, H892-900.

Nakamura, N., Suzuki, Y., Sakuta, H., Ookata, K., Kawahara, K. and Hirose, S. (1999). Inwardly rectifying K+ channel Kir7.1 is highly expressed in thyroid follicular cells, intestinal epithelial cells and choroid plexus epithelial cells: implication for a functional coupling with Na+, K+-ATPase. Biochemical Journal. 342, 329-336.

238 Namba, N., Inagaki, N., Gonoi, T., Seino, Y. and Seino, S. (1996). Kir2.2v: a possible negative regulator of the inwardly rectifying K+ channel Kir2.2. FEBS Letters. 386, 211-214.

Navaratnam, D. S., Escobar, L., Covarrubias, M. and Oberholtzer, J. C. (1995). Permeation properties and differential expression across the auditory receptor epithelium of an inward rectifier K+ channel cloned from the chick inner ear. Journal of Biological Chemistry. 270, 19238-19245.

Navarro, B., Kennedy, M. E., Velimirovic, B., Bhat, D., Peterson, A. S. and Clapham, D. E. (1996). Nonselective and Gpy-insensitive weaver K+ channels. Science, 272,1950- 1953.

Neher, E. (1971). Two fast transient current components during voltage clamp on snail neurons. Journal of General Physiology. 58,36-53.

Neher, E. and Sakmann, B. (1976). Single-channel currents recorded from membrane of denervated frog muscle fibres. Nature. 260, 799-802.

Nehring, R. B., Wischmeyer, E., Doring, F., Veh, R. W., Sheng, M. and Karschin, A. (2000). Neuronal inwardly rectifying K+ channels differentially couple to PDZ proteins of the PSD-95/SAP90 family. Journal of Neuroscience. 20, 156-162.

Nelson, C. S., Marino, J. L. and Allen, C. N. (1997). Cloning and characterization of Kir3.1 (GIRK1) C-terminal alternative splice variants. Brain Research: Molecular Brain Research. 46, 185-196.

Newman, E. A. (1993). Inward-rectifying potassium channels in retinal glial (Muller) cells. Journal of Neuroscience. 13, 3333-3345.

Neyton, J. and Miller, C. (1988). Discrete Ba2+ block as a probe of ion occupancy and pore structure in the high-conductance Ca2+-activated K+ channel. Journal of General Physiology. 92, 569-586.

Nichols, C. G. and Lopatin, A. N. (1997). Inward rectifier potassium channels. Annual Review of Physiology. 59, 171-191.

Nichols, C. G., Shyng, S.-L., Nestorowicz, B. G., Clement IV, J. P., Gonzalez, G., Aguilar-Bryan, L., Permutt, M. A. and Bryan, J. (1996). as an intracellular regulator of insulin secretion. Science. 272, 1785-1787.

Nishida, M. and MacKinnon, R. (2002). Structural basis of inward rectification: cytoplasmic pore of the G protein-gated inward rectifier GERK1 at 1.8A resolution. Cell. Ill, 957-965.

Noble, D. (1965). Electrical properties of cardiac muscle attributable to inward going (anomalous) rectification. Journal of Cell and Comparative Physiology. 66, 127-136.

Noda, M., Ikeda, T., Kayano, T., Suzuki, H., Takeshima, H., Kurasaki, M., Takahashi, H. and Numa, S. (1986). Existence of distinct sodium channel messenger RNAs in rat brain. Nature. 320, 188-192.

239 Noda, M., Shimizu, S., Tanabe, T., Takai, T., Kayano, T., Ikeda, T., Takahashi, H., Nakayama, H., Kanaoka, Y., Minamino, N., Kangawa, K., Matsuo, H., Raftery, M.A., Hirose, T., Inayama, S., Hayashida, H., Miyata, T. and Numa, S. (1984). Primary structure of Electrophorus electricus sodium channel deduced from cDNA sequence. Nature. 312, 121- 127.

Noda, M., Takahashi, H., Tanabe, T., Toyosato, M., Furutani, Y., Hirose, T., Asai, M., Inayama, S., Miyata, T. and Numa, S. (1982). Primary structure of a-subunit precursor of Torpedo califomica acetylcholine receptor deduced from cDNA sequence. Nature. 28, 793- 797.

Noma, A. (1983). ATP-regulated K+ channels in cardiac muscle. Nature. 305, 147- 148.

North, R. A. (1989). Drug receptors and the inhibition of nerve cells. British Journal of Pharmacology. 98, 13-28.

Ogden, D. (1994). Microelectrode electronics. In Microelectrode techniques - the Plymouth workshop handbook. 2nd Ed., pg53-78. The company of biologists limited, Cambridge.

Ogden, D. and Stanfield, P. R. (1994). Patch clamp techniques for single channel and whole-cell recording. In Microelectrode techniques - the Plymouth workshop handbook. 2nd Ed., pg53-78. The company of biologists limited, Cambridge.

Ogielska, E. M., Zagotta, W. N., Hoshi, T., Heinemann, S. H., Haab, J. and Aldrich, R. W. (1995). Cooperative subunit interactions in C-type inactivation of K channels. Biophysical Journal. 69, 2449-2457.

Ohira, M., Seki, N., Nagase, T., Suzuki, E., Nomura, N., Ohara, O., Hattori, M., Sakaki, Y., Eki, T., Murakami, Y., Saito, T., Ichikawa, H. and Ohki, M. (1997). Gene identification in 1.6Mb region of the Down syndrome region on chromosome 21. Genome Research. 7, 47-58.

Ohmori, H. (1978). Inactivation kinetics and steady-state current noise in the anomalous rectifier of tunicate egg cell membranes. Journal of Physiology. 281, 77-99.

Ohmori, H. (1980). Dual effects of K ions upon the inactivation of the anomalous rectifier of the tunicate egg cell membrane. Journal of Membrane Biology. 53, 143-156.

Oliva, C., Cohen, I. S. and Pennefather, P. (1990). The mechanism of rectification of i’k i in canine purkinje myocytes. Journal of General Physiology. 96, 299-318.

Omori, K., Oishi, K. and Matsuda, H. (1997). Inwardly rectifying potassium channels expressed by gene transfection into the green monkey kidney cell line COS-1. Journal of Physiology. 499.2, 369-378.

Orio, P., Rojas, P., Ferreira, G. and Latorre, R. (2002). New disguises for an old channel: MaxiK channel (3-subunits. News in Physiological Sciences. 17, 156-161.

240 Otter-Nilsson, M. and Nilsson, T. (1999). Protein expression in mammalian cells. In Protein expression: a practical approach. Ed. Higgins, S. J. and Hames, B. D. ppl.27. IRL Press, Oxford.

Pallotta, B. S., Magleby, K. L. and Barrett, J. N. (1981). Single channel recordings of Ca2+-activated K+ currents in rat muscle cell culture. Nature. 293,471-474.

Panyi, G., Sheng, Z. F., Tu, L. W. and Deutsch, C. (1995). C-type inactivation of a voltage gated K+ channel occurs by a cooperative mechanism. Biophysical Journal. 69, 896- 903.

Papazian, D. M. (1999). Potassium channels: some assembly required. Neuron. 23, 7-10.

Papazian, D. M., Schwarz, T. L., Tempel, B. L., Jan, Y. N. and Jan, L. Y. (1987). Cloning of genomic and complementary DNA from Shaker, a putative potassium channel gene from Drosophila. Science. 237,749-753.

Papazian, D. M., Timpe, L. C., Jan, Y. N. and Jan, L. Y. (1991). Alteration of voltage-dependence of Shaker potassium channel by mutations in the S4 sequence. Nature. 349, 305-310.

Partiseti, M., Collura, V., Agnel, M., Culouscou, J. M. and Graham, D. (1998). Cloning and characterization of a novel human inwardly rectifying potassium channel predominantly expressed in small intestine. FEBS Letters. 434, 171-176.

Pascual, J. M., Shieh, C-C., Kirsch, G. E. and Brown, A. M. (1995). K+ pore structure revealed by reported cysteines at inner and outer surfaces. Neuron. 14, 1055-1063.

Patel, A. J. and Honore, E. (2001). Properties and modulation of mammalian 2P domain K+ channels. TRENDS in Neuroscience. 24, 339-346.

Patil, N., Cox, D. R., Bhat, D., Faham, M., Myers, R. M. and Peterson, A. S. (1995). A potassium channel mutation in weaver mice implicates membrane excitability in granule cell differentiation. Nature Genetics. 11,126-129.

Pearson, W. L., Dourado, M., Schreiber, M., Salkoff, L. and Nichols, C. G. (1999). Expression of a functional Kir4 family inward rectifier K+ channel from a gene cloned from mouse liver. Journal of Physiology. 514.3, 639-653.

Pennefather, P., Oliva, C. and Mulrine, N. (1992). Origin of the potassium and voltage dependence of the cardiac inwardly rectifying K-current (I k i). Biophysical Journal. 61, 448-462.

Perier, F., Coulter, K. L., Radeke, C. M. and Vandenberg, C. A. (1992). Expression of an inwardly rectifying potassium channel in Xenopus oocytes. Journal of Neurochemistry. 59, 1971-1974.

Perier, F., Radeke, C. M. and Vandenberg, C. A. (1994). Primary structure and characterization of a small-conductance inwardly rectifying potassium channel from human hippocampus. Proceedings of the National Academy of Sciences USA. 91, 6240-6244.

241 Perillan, P. R., Li, X., Potts, E. A., Chen, M., Bredt, D. S. and Simard, J. M. (2000). Inward rectifier K+ channel Kir2.3 (IRK3) in reactive astrocytes from adult rat brain. Glia. 31, 181-192.

Perozo, E., Cortes, D. M. and Cuello, L. G. (1999). Structural rearrangements underlying K+-channel activation gating. Science. 285, 73-78.

Pessia, M., Imbrici, P., D ’Adamo, M. C., Salvatore, L. and Tucker, S. J. (2001). Differential pH sensitivity of Kir4.1 and Kir4.2 potassium channels and their modulation by heteropolymerisation with Kir5.1. Journal of Physiology. 532,359-367.

Pessia, M., Tucker, S. J., Lee, K., Bond, C. T. and Adelman, J. P. (1996). Subunit positional effects revealed by novel heteromeric inwardly rectifying K+ channels. EMBO Journal. 12, 2980-2987.

Peters, M., Jeck, N., Reinalter, S., Leonhardt, A., Tonshoff, B., Klaus, G., Konrad, M. and Seyberth, H. W. (2002). Clinical presentation of genetically defined patients with hypokalemic salt-losing tubulopathies. Amercian Journal of Medicine. 112, 183-190.

Pfaffinger, P. J., Martin, J. M., Hunter, D. D., Nathanson, N. M. and Hille, B. (1985). GTP-binding proteins couple cardiac muscarinic receptors to a K channel. Nature. 317, 536- 538.

Picones, A., Keung, E. and Timpe, L. C. (2001). Unitary conductance variation in Kir2.1 and in cardiac inward rectifier potassium channels. Biophysical Journal. 81, 2035- 2049.

Pineda, J. and Aghajanian, G. K. (1997). Carbon dioxide regulates the tonic activity of locus coeruleus neurons by modulating a proton- and polyamine-sensitive inward rectifier potassium current. Neuroscience. 77, 73-743.

Plaster, N. M., Tawil, R., Tristani-Firouzi, M., Canun, S., Bendahhou, S., Tsunoda, A., Donaldson, M. R., Iannaccone, S. T., Brunt, E., Barohn, R., Clark, J., Deymeer, F., George Jr, A. L., Fish, F. A., Hahn, A., Nitu, A., Ozdemir, C., Serdaroglu, P., Subramony, S. H., Wolfe, G., Fu, Y. H. and Ptacek, L. J. (2001). Mutations in Kir2.1 cause the developmental and episodic electrical phenotypes of Andersen’s syndrome. Cell. 105, 511-519.

Pluger, S., Faulhaber, J., Furstenau, M., Lohn, M., Waldschiitz, R., Gollasch, M., Haller, H., Luft, F. C., Ehmke, H. and Pongs, O. (2000). Mice with disrupted BK channel (31 subunit gene feature abnormal Ca2+ spark/STOC coupling and elevated blood pressure. Circulation Research. 87, e53-E60.

Pond, A. L. and Nerbonne, J. M. (2001). ERG proteins and functional cardiac Ikt channels in rat, mouse, and human heart. Trends in Cardiovascular Medicine. 11, 286-294.

Pongs, O., Kecskemethy, N., Muller, R., Krah-Jentgens, I., Baumann, A., Kiltz, H. H., Canal, I., Llamazares, S. and Ferrus, A. (1988). Shaker encodes a family of putative potassium channel proteins in the nervous system of Drosophila. EMBO Journal. 7, 1087- 1096.

242 Pongs, O., Leicher, T., Berger, M., Roeper, J., Bahring, R., Wray, D., Giese, K. P., Silva, A. J. and Storm, J. F. (1999). Functional and molecular aspects of voltage-gated K+ channel (3 subunits. Annals of the New York Academy of Sciences. 868, 344-355.

Potter, H., Weir, L. and Leder, P. (1984). Enhancer-dependent expression of human k immunoglobulin genes introduced into mouse pre-B lymphocytes by electroporation. Proceedings o f the National Academy of Sciences of the USA. 81, 7161-7165.

Pountney, D. J., Gulkarov, I., Vega-Saenz de Miera, E., Holmes, D., Saganich, M., Rudy, B., Artman, M. and Coetzee, W. A. (1999). Identification and cloning of TWIK- originated similarity sequence (TOSS): a novel human 2-pore K+ channel principal subunit. FEBS Letters. 450, 191-196.

Preisig-Miiller, R., Schlichthorl, G., Goerge, T., Heinen, S., Briiggemann, A., Rajan, S., Derst, C., Veh, R. W. and Daut, J. (2002). Heteromerization of Kir2.x potassium channels contributes to the phenotype of Andersen’s syndrome. Proceedings of the National Academy of Sciences USA. 99, 7774-7779.

Proks, P., Capener, C. E., Jones, P. and Ashcroft, F. M. (2001). Mutations within the P-loop of Kir6.2 modulate the intraburst kinetics of the ATP-sensitive potassium channel. Journal of General Physiology. 118, 341-353.

Proks, P., Gribble, F. M., Adhikari, R., Tucker, S. J. and Ashcroft, F. M. (1999).

Involvement of the N-terminus of Kir6.2 in the inhibition of the K a t p channel by ATP. Journal of Physiology. 514.1, 19-25.

Qu, Z., Zhu, G., Yang, Z., Cui, N., Li, Y., Chanchevalap, S., Sulaiman, S., Haynie, H. and Jiang, C. (1999). Identification of a critical motif responsible for gating of Kir2.3 channel by intracellular protons. Journal of Biological Chemistry. 274, 13783-13789.

Quayle, J. M., Dart, C. and Standen, N. B. (1996). The properties and distribution of inward rectifier potassium currents in pig coronary arterial smooth muscle. Journal of Physiology. 494,715-726.

Quirk, J. C. and Reinhart, P. H. (2001). Identification of a novel tetramerization domain in large conductance Kca channels. Neuron. 32, 13-23.

Raab-Graham, K. F., Radeke, C. M. and Vandenberg, C. A. (1994). Molecular cloning and expression of a human heart inward rectifier potassium channel. Neuroreport. 5, 2501-2505.

Raap, M., Biedermann, B., Braun, P., Milenkovic, I., Skatchkov, S. N., Bringmann, A. and Reichenbach, A. (2002). Diversity of Kir channel subunit mRNA expressed by retinal glial cells of the guinea-pig. Neuroreport. 13, 1037-1040.

Rae, J. L. and Shepard, A. R. (1998). Inwardly rectifying potassium channels in lens epithelium are from the IRK1 (Kir2.1) family. Experimental Eye Research. 66, 347-359.

Rakic, P. and Sidman, R. L. (1973). Sequence of developmental abnormalities leading to granule cell deficit in cerebellar cortex of weaver mutant mice. Journal of Comparative Neurology. 152, 103-132.

243 Reid, J. D., Lukas, W., Shafaatian, R., Bertl, A., Scheurmann-Kettner, C., Guy, H. R. and North, R. A. (1996). The S.cereviasiae outwardly-rectifying potassium channel (DUK1) identifies a new family of channels with duplicated pore domains. Receptors and Channels. 4, 51-62.

Reimann, F and Ashcroft, F. M. (1999). Inwardly rectifying potassium channels. Current Opinion in Cell Biology. 11, 503-508.

Renisio, J-G., Lu, Z., Blanc, E., Jin, W., Lewis, J. H., Bomet, O. and Darbon, H. (1999). Solution structure of potassium channel-inhibiting scorpion toxin Lq2. PROTEINS: Structure, function and genetics. 34, 417-426.

Repunte, V. P., Nakamura, H., Fujita, A., Horio, Y., Findlay, I., Pott, L. and Kurachi, Y. (1999). Extracellular links in Kir subunits control the unitary conductance of SUR/Kir6.0 channels. EMBO Journal. 18,3317-3324.

Rettig, J., Heinemann, S. H., Wunder, F., Lorra, C., Parcej, D. N., Dolly, J. O. and Pongs, O. (1994). Inactivation properties of voltage-gated K+ channels altered by presence of p-subunit. Nature. 369, 289-294.

Reuveny, E., Jan, Y. N. and Jan, L. Y. (1996). Contributions of a negatively charged residue in the hydrophobic domain of the IRK1 inwardly rectifying K+ channel to K+- selective permeation. Biophysical Journal. 70,754-761.

Reuveny, E., Slesinger, P. A., Inglese, J., Morales, J. M., Iniguez-Lluhi, J. A., Lefkowitz, R. J., Bourne, H. R., Jan, Y. N. and Jan, L. Y. (1994). Activation of the cloned muscarinic potassium channel by G protein py subunits. Nature. 370, 143-146.

Robbins, J. (2001). KCNQ potassium channels: physiology, pathophysiology, and pharmacology. Pharmacology & Therapeutics. 90, 1-19.

Robertson, B. (1997). The real life of voltage-gated K+ channels: more than model behaviour. Trends in Pharmacological Sciences. 18, 474-483.

Robertson, G. A., Warmke, J. W. and Ganetzky, B. (1996). Potassium currents expressed from Drosophila and mouse eag cDNAs in Xenopus oocytes. Neuropharmacology. 35, 841-850.

Romey, G. and Lazdunski, M. (1984). The coexistence in rat muscle cells of two distinct classes of Ca2+ -dependent K+ channels with different pharmacological properties and different physiological functions. Biochemical and Biophysical Research Communications. 118, 669-674.

Rossi, P., De Filippi, G., Armano, S., Taglietti, V. and D’Angelo, E. (1998). The weaver mutation causes a loss of inward rectifier current regulation in premigratory granule cells of the mouse cerebellum. Journal of Neuroscience. 18, 3537-3547.

Ruknudin, A., Schulze, D. H., Sullivan, S. K., Lederer, W. J. and Welling, P. A. (1998). Novel subunit composition of a renal epithelial KATP channel. Journal of Biological Chemistry. 273, 14165-14171.

244 Ruppersburg, J. P., Schroter, K. H., Sakmann, B., Stocker, M., Sewing, S. and Pongs, O. (1990). Heteromultimeric channels formed by rat brain potassium-channel proteins. Nature. 345, 535-537.

Sabirov, R. Z., Tominaga, T., Miwa, A., Okada, Y. and Oiki, S. (1997). A conserved arginine residue in the pore region of an inward rectifier K channel (IRK1) as an external barrier for cationic blockers. Journal of General Physiology. 110, 665-677.

Sadja, R., Smadja, K., Alagem, N. and Reuveny, E. (2001). Coupling Gpy-dependent activation to channel opening via pore elements in inwardly rectifying potassium channels. Neuron. 29, 669-680.

Sah, P. (1996). Ca2+-activated K+currents in neurones: types, physiological roles and modulation. Trends in Neuroscience. 19,150-154.

Sah, P. and Faber, E. S. L. (2002). Channels underlying neuronal calcium-activated potassium currents. Progress in Neurobiology. 66, 345-353.

Sakmann, B. and Trube, G. (1984a). Conductance properties of single inwardly rectifying potassium channels in ventricular cells from guinea-pig heart. Journal of Physiology. 347, 641-657.

Sakmann, B. and Trube, G. (1984b). Voltage-dependent inactivation of inward- rectifying single-channel currents in the guinea-pig heart cell membrane. Journal of Physiology. 347, 659-683.

Sakura, H., Ammala, C., Smith, P. A., Gribble, F. M. and Ashcroft, F. M. (1995). Cloning and functional expression of the cDNA encoding a novel ATP-sensitive potassium channel subunit expressed in pancreatic p-cells, brain, heart and skeletal muscle. FEBS Letters. 377, 338-344.

Sali, A. and Blundell, T. L. (1993). Comparative protein modelling by satisfaction of spatial restraints. Journal of Molecular Biology. 234,779-815.

Salinas, M., Reyes, R., Lesage, F., Fosset, M., Heurteaux, C., Romey, G. and Lazdunski, M. (1999). Cloning of a new mouse two-P domain channel subunit and a human homologue with a unique pore structure. Journal of Biological Chemistry. 274, 11751- 11760.

Salkoff, L., Baker, K., Butler, A., Covarrubias, M., Pak, M. D. and Wei, A. (1992). An essential ‘set’ of K+ channels conserved in flies, mice and humans. Trends in Neuroscience. 15, 161-166.

Salkoff, L. and Wyman, R. (1981). Genetic modification of potassium channels in Drosophila Shaker mutants. Nature. 293, 228-230.

Salvatore, L., D’Adamo, M. C., Polishchuk, R., Salmona, M. and Pessia, M. (1999). Localization and age-dependent expression of the inward rectifier K+ channel subunit Kir5.1 in a mammalian reproductive system. FEBS Letters. 449, 146-152.

Sanguinetti, M. C. (2000). Maximal functions of minimal K+channel subunits. Trends in Pharmacological Sciences. 21, 199-201.

245 Sanguinetti, M. C and Spector, P. S. (1997). Potassium Channelopathies. Neuropharmacology. 36, 755-762.

Sato, S., Ward, C. L., Krouse, M. E., Wine, J. J. and Kopito, R. R. (1996). Glycerol reverses the misfolding phenotype of the most common cystic fibrosis mutation. Journal of Biological Chemistry. 271, 635-638.

Scheele, G. and Jacoby, R. (1982). Conformational changes associated with proteolytic processing of presecretory proteins allow glutathione-catalyzed formation of native disulfide bonds. Journal of Biological Chemistry. 257, 12277-12282.

Schlief, T., Schonherr, R. and Heinemann, S. H. (1996). Modification of C-type inactivating Shaker potassium channels by chloramine-T. 431, 483-493.

Schram, G., Melnyk, P., Pourder, M., Wang, Z. and Nattel, S. (2002). Kir2.4 and Kir2.1 K+ channel subunits co-assemble: a potential new contributor to inward rectifier current heterogeneity. Journal of Physiology. 544.2, 337-349.

Schulte, U. and Fakler, B. (2000). Gating of inward rectifier K+ channels by intracellular pH. European Journal of Biochemistry. 267, 5837-5841.

Schulte, U., Hahn, H., Konrad, M., Jeck, N., Derst, C., Wild, K., Weidemann, S., Ruppersberg, J. P., Fakler, B. and Ludwig, J. (1999). pH gating of ROMK (Kirl.l) channels: control by an Arg-Lys-Arg triad disrupted in antenatal . Proceedings o f the National Academy of Sciences USA. 96, 15298-15303.

Schulteis, C. T., Nagaya, N. and Papazian, D. M. (1998). Subunit folding and assembly steps are interspersed during Shaker potassium channel biogenesis. Journal of Biological Chemistry. 273, 26210-26217.

Schulteis, C. T., Nagaya, N. and Papazian, D. (1996). Intersubunit interaction between amino- and carboxy-terminal cysteine residues in tetrameric Shaker K+ channels. Biochemistry. 35, 12133-12140.

Schwalbe, R. A., Wingo, C. S. and Xia, S. L. (2002). Mutations in the putative pore- forming segment favor short-lived wild-type Kir2.1 pore conformations. Biochemistry. 41, 12457-12466.

Schwarz, J. R. and Bauer, C. K. (1999). Theether-a-go-go- related gene K+ current: functions of a strange inward rectifier. News in Physiological Sciences. 14, 135-142.

Schwarz, T. L., Tempel, B. L., Papazian, D. M., Jan, Y. N. and Jan, L. Y. (1988). Multiple potassium-channel components are produced by alternative splicing at the Shaker locus in Drosophila. Nature. 331, 137-142.

Selyanko, A. A., Hadley, J. K., Wood, I. C., Abogadie, F. C., Jentsch, T. J. and Brown, D. A. (2000). Inhibition of KCNQ1-4 potassium channels expressed in mammalian cells via M l muscarinic acetylcholine receptors. Journal of Physiology. 522, 349-355.

Sha, Q., Romano, C., Lopatin, A. N. and Nichols, C. G. (1996). Spermidine release from xenopus oocytes. Electrodiffusion through a . Journal of Biological Chemistry. 271, 3392-3397/

246 Shelton, P. A., Davies, N. W., Antoniou, M., Grosveld, F., Needham, M., Hollis, M., Brammar, W. J. and Conley, E. C. (1993). Regulated expression of K+ channel genes in electrically silent mammalian cells by linkage to P-globin gene-activation elements. Receptors and Channels. 1, 25-37.

Shen, N. V., Chen, X., Boyer, M. M. and Pfaffinger, P. J. (1993). Deletion analysis of K+channel assembly. Neuron. 11,67-76.

Shieh, R-C. (2000). Mechanisms for the time-dependent decay of inward currents through cloned Kir2.1 channels expressed in Xenopus oocytes. Journal of Physiology. 526.2, 241-252.

Shieh, R-C., Chang, J-C. and Arreola, J. (1998). Interaction of Ba2+ with the pores of the cloned inward rectifier channels Kir2.1 expressed in Xenopus oocytes. Biophysical Journal. 75, 2313-2322.

Shieh, C- C., Coghlan, M., Sullivan, J. P. and Gopalakrishnan, M. (2000). Potassium channels: Molecular defects, diseases, and therapeutic opportunities. Pharmacological Reviews. 52, 557-593.

Shieh, R-C., John, S. A., Lee, J-K. and Weiss, J. N. (1996). Inward rectification of the IRK1 channel expressed in Xenopus oocytes: effects of intracellular pH reveal an intrinsic gating mechanism. Journal of Physiology. 494.2, 363-376.

Shieh, R-C. and Lee, Y-L. (2001). Ammonium ions induce inactivation of Kir2.1 potassium channels expressed in Xenopus oocytes. Journal of Physiology. 535.2, 359-370.

Shimoni, Y., Clark, R. B. and Giles, W. R. (1992). Role of an inwardly rectifying potassium current in rabbit ventricular action potential. Journal of Physiology. 448, 709-727.

Shimura, M., Yuan, Y., Chang, J. T., Zhang, S., Campochiaro, P. A., Zack, D. J. and Hughes, B. A. (2001). Expression and permeation properties of the K+ channel Kir7.1 in the retinal pigment epithelium. Journal of Physiology. 531.2, 329-346.

Shin, K. S., Park, J. Y., Kwon, H., Chung, C. H. and Kang, M. S. (1997). A possible role of inwardly rectifying K+ channels in chick myoblast differentiation. American Journal of Physiology. 272, C894-C900.

Shuck, M. E., Bock, J. H., Benjamin, C. W., Tsai, T-D., Lee, K. S., Slightom, J. L. and Bienkowski, M. J. (1994). Cloning and characterization of multiple forms of the human kidney ROM-K potassium channel. Journal of Biological Chemistry. 269, 24261-24270.

Shuck, M. E., Piser, T. M., Bock, J. H., Slightom, J. L., Lee, K. S. and Bienkowski, M. J. (1997). Cloning and characterization of two K+ inward rectifier (Kir) 1.1 potassium channel homologs from human kidney (Kiri.2 and Kiri.3). Journal of Biological Chemistry. 272, 586-593.

Shyng, S. L. and Nichols, C. G. (1997). Octameric stoichiometry of the KA t p channel complex. Journal of General Physiology. 110, 655-664.

247 Shyng, S. L., Sha, Q., Ferrigni, T., Lopatin, A. N. and Nichols, C. G. (1996). Depletion of intracellular polyamines relieves inward rectification of potassium channels. Proceedings of the National Academy of Sciences USA. 93, 12014-12019.

Signorini, S., Liao, Y. J., Duncan, S. A., Jan, L. Y. and Stoffel, M. (1997). Normal cerebellar development but susceptibility to seizures in mice lacking G protein-coupled, inwardly rectifying K+ channel GIRK2. Proceedings of the National Academy of Sciences USA. 94,923-927.

Silver, M. R. and DeCoursey, T. E. (1990). Intrinsic gating of inward rectifier in bovine pulmonary artery endothelial cells in the presence or absence of internal Mg2+. Journal of General Physiology. 96, 109-133.

Silverman, S. K., Kofuji, P., Dougherty, D. A., Davidson, N. and Lester, H. A. (1996b). A regenerative link in the ionic fluxes through the weaver potassium channel underlies the pathophysiology of the mutation. Proceedings of the National Academy of Sciences USA. 93, 15429-15434.

Silverman, S. K., Lester, H. A. and Dougherty, D. A. (1996a). Subunit stoichiometry of a heteromultimeric G protein-coupled inward-rectifier K+ channel. Journal of Biological Chemistry. 271, 30524-30528.

Simon, D. B., Bindra, R. S., Mansfield, T. A., Nelson-Williams, C., Mendonca, E., Stone, R., Schurman, S., Nayir, A., Alpay, H., Bakkaloglu, A., Rodriguez-Soriano, J., Morales, J. M., Sanjad, S. A., Taylor, C. M., Pilz, D., Brem, A., Trachtman, H., Griswold, W., Richard, G. A., John, E. and Lifton, R. P. (1997). Mutations in the chloride channel gene, CLCNKB, cause Bartter’s syndrome type HI. Nature Genetics. 17, 171-178.

Simon, D. B., Karet, F. E., Hamdan, J. M., DiPietro, A., Sanjad, S. A. and Lifton, R. P. (1996a). Bartter’s syndrome, hypokalaemic alkalosis with hypercalciuria, is caused by mutations in the Na-K-2C1 cotransporter NKCC2. Nature Genetics. 13, 183-188.

Simon, D. B., Karet, F. E., Rodriguez-Soriano, J., Hamdan, J. H., DiPietro, A., Trachtman, H., Sanjad, S. A. and Lifton, R. P. (1996b). Genetic heterogeneity of Bartter’s syndrome revealed by mutations in the K+ channel, ROMK. Nature Genetics. 14, 152-156.

Skryma, R., Prevarskaya, N., Vacher, P. and Dufy, B. (1994). Voltage-dependent ionic conductances in Chinese hamster ovary cells. American Journal of Physiology. 267, C544-C553.

Slesinger, P. A., Jan, Y. N. and Jan, L. Y. (1993). The S4-S5 loop contributes to the ion-selective pore of potassium channels. Neuron. 11, 739-749.

Slesinger, P. A., Patil, N., Liao, Y. J., Jan, Y. N., Jan, L. Y. and Cox, D. R. (1996). Functional effects of the mouse weaver mutation on G-protein-gated inwardly rectifying K+ channels. Neuron. 16,321-331.

Slesinger, P. A., Reuveny, E., Jan, Y. N. and Jan, L. Y. (1995). Identification of structural elements involved in G protein gating of the GIRK1 potassium channel. Neuron. 15, 1145-1156.

248 Smeyne, R. J. and Goldowitz, D. (1989). Development and death of external granular layer cells in the weaver mouse cerebellum: a quantitative study. Journal of Neuroscience. 9, 1608-1620.

So, I., Ashmole, I., Davies, N. W., Sutcliffe, M. J. and Stanfield, P. R. (2001). The K+ channel signature sequence of murine Kir2.1: mutations that affect microscopic gating but not ionic selectivity. Journal of Physiology. 531.1, 37-50.

Sobey, C. G. and Faraci, F. M. (2000). Knockout blow for channel identity crisis. Vasodilation to potassium is mediated via Kir2.1. Circulation Research. 87, 83-84.

Soejima, M. and Noma, A. (1984). Mode of regulation of the ACh-sensitive K- channel by the muscarinic receptor in rabbit atrial cells. Pfliigers Archives. 400, 424-431.

Song, W-J. (2002). Genes responsible for native depolarization-activated K+ currents in neurons. Neuroscience Research. 42,7-14.

Songyang, Z., Fanning, A. S., Fu, C., Xu, J., Marfatia, S. M., Chishti, A. H., Crompton, A., Chan, A. C., Anderson, J. M. and Cantley, L. C. (1997). Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science. 275, 73-77.

Soom, M., Schronherr, R., Kubo, Y., Kirsch, C., Klinger, R. and Heinemann, S. H. (2001). Multiple PIP 2 binding sites in Kir2.1 inwardly rectifying potassium channels. FEBS Letters. 490, 49-53.

Spauschus, A., Lentes, K-U., Wischmeyer, E., DiBmann, E., Karschin, C. and Karschin, A. (1996). A G-protein-activated inwardly rectifying K+ channel (GIRK4) from human hippocampus associates with other GIRK channels. Journal of Neuroscience. 16, 930-938.

Spruce, A. E., Standen, N. B. and Stanfield, P. R. (1985). Voltage-dependent ATP- sensitive potassium channels of skeletal muscle membrane. Nature. 16, 763-738.

Standen, N. B., Davies, N. W. and Stanfield, P. R. (1997). Patch clamping methods and analysis of ion channels. In Principles of Medical Biology. Membranes and Cell Signaling 7B, 355-375..

Standen, N. B., Quayle, J. M., Davies, N. W., Brayden, J. E., Huang, Y. and Nelson, M. T. (1989). Hyperpolarizing vasodilators activate ATP-sensitive K+ channels in arterial smooth muscle. Science. 245,177-180.

Standen, N. B. and Quayle, J. M. (1998). K+ channel modulation in arterial smooth muscle. Acta Physiologica Scandinvica. 164, 549-557.

Standen, N. B. and Stanfield, P. R. (1978a). A potential- and time-dependent blockade of inward rectification in frog skeletal muscle fibres by barium and strontium ions. Journal of Physiology. 280, 169-191.

Standen, N. B. and Stanfield, P. R. (1978b). Inward rectification in skeletal muscle: a blocking particle model. Pfliigers Archives. 378,173-176.

249 Standen, N. B. and Stanfield, P. R. (1979). Potassium depletion and sodium block of potassium currents under hyperpolarization in frog sartorius muscle. Journal of Physiology. 294, 497-520.

Standen, N. B. and Stanfield, P. R. (1980). Rubidium block and rubidium permeability of the inward rectifier of frog skeletal muscle fibres. Journal of Physiology. 304, 415-435.

Standen, N. B. and Stanfield, P. R. (1992). Patch clamp methods for single channel and whole cell recording. In Monitoring neuronal activity - a practical approach. Ed Stamford, J.A., pg59-83. IRL press, Oxford.

Stansfeld, C., Ludwig, J., Roeper, J., Weseloh, R., Brown, D. and Pongs, O. (1997). A physiological role for ether-a-go-go K+ channels? Trends in Neuroscience. 20, 13-14.

Stanfield, P. R. (1970). The differential effects of tetraethylammonium and zinc ions on the resting conductance of frog skeletal muscle. Journal of Physiology. 209, 231-256.

Stanfield, P. R., Davies, N. W., Shelton, P. A., Khan, I. A., Brammar, W. J., Standen, N. B. and Conley, E.C. (1994a). The intrinsic gating of inward rectifier K+ channels expressed from the murine IRK1 gene depends on voltage, K+ and Mg2+. Journal of Physiology. 475.1, 1-7.

Stanfield, P. R., Davies, N. W., Shelton, P. A., Sutcliffe, M. J., Khan, I. A., Brammar, W. J. and Conley, E. C. (1994b). A single aspartate residue is involved in both intrinsic gating and blockage by Mg2+ of the inward rectifier, IRK1. Journal of Physiology. 478.1, 1- 6.

Stanfield, P. R., Nakajima, S. and Nakajima, Y. (2002). Constitutively active and G- protein coupled inward rectifier K+ channel s:Kir2.0 and Kir3.0. Reviews of Physiology, Biochemistry and Pharmacology. 145, 50-179.

Starace, D. M. and Bezanilla, F. (2001). Histidine scanning mutagenesis of basic residues of the S4 segment of the Shaker K+ channel. Journal of General Physiology. 117, 469-490.

Starkus, J. G., Kuschel, L., Rayner, M. D. and Heinemann, S. H. (1997). Ion conduction through C-type inactivated Shaker channels. Journal of General Physiology. 110,539-550.

Stockklausner, C., Ludwig, J. L., Ruppersberg, J. P. and Klocker, N. (2001). A sequence motif responsible for ER export and surface expression of Kir2.0 inward rectifier K+ channels. FEBS Letters. 493,129-133.

Stonehouse, A. H., Pringle, J. H., Norman, R. I., Stanfield, P. R., Conley, E. C. and Brammar, W. J. (1999). Characterisation of Kir2.0 proteins in the rat cerebellum and hippocampus by polyclonal antibodies. Histochemistry and Cell Biology. 112, 457-465.

Suh, B. C. and Hille, B. (2002). Recovery from muscarinic modulation of M current channels requires phosphatidylinositol 4, 5-bisphosphate synthesis. Neuron. 35, 507-520.

250 Surmeier, D. J., Mermelstein, P. G. and Goldowitz, D. (1996). The weaver mutation of GIRK2 results in a loss of inwardly rectifying K+ current in cerebellar granule cells. Proceedings of the National Academy of Sciences USA. 93, 11191-11195.

Taglialatela, M., Drewe, J. A., Kirsch, G. E., De Biasi, M., Hartmann, H. A. and Brown, A. M. (1993). Regulation of K+/Rb+ selectivity and internal TEA blockade by mutations at a single site in K+ pores. Pfliigers Archives. 423, 104-112.

Taglialatela, M., Ficker, E., Wible, B. A. and Brown, A. M. (1995). C-terminus determinants for Mg2+ and polyamine block of the inward rectifier K+ channel IRK1. EMBO Journal , 14, 5532-5541.

Taglialatela, M., Wible, B. A., Caporaso, R. and Brown, A. M. (1994). Specification of pore properties by the carboxyl terminus of inwardly rectifying K+ channels. Science. 264, 845-847.

Takahashi, K., Miyazaki, S-I. and Kidokoro, Y. (1971). Development of excitability in embryonic muscle cell membranes in certain tunicates. Science. 171, 415-418.

Takahashi, N., Morishige, K-I., Jahangir, A., Yamada, M., Findlay, I., Koyama, H. and Kurachi, Y. (1994). Molecular cloning and functional expression of cDNA encoding a second class of inward rectifier potassium channels in the mouse brain. Journal of Biological Chemistry. 269, 23274-23279.

Takahashi, N., Morishige, K-I., Yamada, M., Takumi, T. and Kurachi, Y. (1995). Molecular and functional heterogeneity of inward rectifier potassium channels in brain and heart. Heart and Vessels. 9,8-11.

Takano, M., Ishii, T. and Xie, L. H. (1996). Cloning and functional expression of the rat brain KIR6.2 channel. Japanese Journal of Physiology. 46, 491-495.

Takao, K., Yoshii, M., Kanda, A., Kokubun, S. and Nukada, T. (1994). A region of the muscarinic-gated atrial K+ channel critical for activation by protein subunits. Neuron. 13, 747-755.

Takumi, T., Ishii, T., Horio, Y., Morishige, K-I., Takahashi, N., Yamada, M., Yamashita, T., Kiyama, H., Sohmiya, K., Nakanishi, S. and Kurachi, Y. (1995). A novel ATP-dependent inward rectifier potassium channel expressed predominantly in glial cells. Journal of Biological Chemistry. 270, 16339-16346.

Tanabe, T., Takeshima, H., Mikami, A., Flockerzi, V., Takahashi, H., Kangawa, K , Kojima, M., Matsuo, H., Hirose, T. and Numa, S. (1987). Primary structure of the receptor for blockers from skeletal muscle. Nature. 328, 313-318.

Tanemoto, M., Fujita, A., Higashi, K. and Kurachi, Y. (2002). PSD-95 mediates formation of a functional homomeric Kir5.1 channel in the brain. Neuron. 34, 387-397.

Tanemoto, M., Kittaka, N., Inanobe, A. and Kurachi, Y. (2000). In vivo formation of a proton-sensitive K+ channel by heteromeric subunit assembly of Kir5.1 with Kir4.1. Journal of Physiology. 525.3, 587-592.

251 Tang, W., Qin, C. L. and Yang, X. C. (1995). Cloning, localization, and functional expression of a human brain inward rectifier potassium channel (hIRKI). Receptors and Channels. 3, 175-183.

Tang, W. and Yang, X-C. (1994). Cloning a novel human brain inward rectifier potassium channel and its functional expression in Xenopus oocytes. FEBS Letters. 348, 239-243.

Tanouye, M. A., Ferrus, A. and Fujita, S. C. (1981). Abnormal action potentials associated with the Shaker complex locus of Drosophila. Proceedings of the National Academy of Sciences USA. 78, 6548-6552.

Tanouye, M. A., Kamb, C. A. and Iverson, L. E. (1986). Genetics and molecular biology of ionic channels in Drosophila. Annual Review of Neuroscience. 9, 255-276.

Tare, M., Prestwich, S. A., Gordienko, D. V., Parveen, S., Carver, J. E., Robinson, C., and Bolton, T. B. (1998). Inwardly rectifying whole cell potassium current in human blood eosinophils. Journal of Physiology. 506,303-318.

Taylor, J. W., Ott, J. and Eckstein, F. (1985). The rapid generation of oligonucleotide-directed mutations high frequency using phosporothioate-modified DNA. Nucleic Acids Research. 13, 8765-8785.

Tempel, B. L., Jan, Y. N. and Jan, L. Y. (1988). Cloning of a probable potassium channel gene from mouse brain. Nature. 332, 837-839.

Tempel, B. L., Papazian, D. M., Schwarz, T. L., Jan, Y. N. and Jan, L. Y. (1987). Sequence of a probable potassium channel component encoded at Shaker locus of Drosophila. Science. 237, 770-775.

Thomas, P. M., Cote, G. J., Wohllk, N., Haddad, B., Matthew, P. M., Rabl, W., Aguilar-Bryan, L., Gagel, R. F. and Bryan, J. (1995). Mutations in the sulfonylurea receptor gene in familial persistent hyperinsulinemic hypoglycemia of infancy. Science. 268, 426- 429.

Thompson, G. A., Leyland, M. L., Ashmole, I. and Stanfield, P. R. (2000a). Residues in H5 and M2 of murine Kir2.1 regulate Ba2+ block. Journal of Physiology. 526P.

Thompson, G. A., Leyland, M. L., Ashmole, I., Sutcliffe, M. J. and Stanfield, P. R. (2000b). Residues beyond the selectivity filter of the K+ channel Kir2.1 regulate permeation and block by external Rb+ and Cs+. Journal of Physiology. 526.2, 231 -240.

Thompson, G. A., Passmore, G. M., Spencer, P. J., Davies, N. W. and Stanfield, P. R. (1998). Blockage of inward rectifier potassium channels (murine Kir2.1) by Ba2+ is influenced by an aspartate residue (D172). Journal of Physiology. 511P, 146P.

Timpe, L.C., Schwarz, T. L., Tempel, B. L., Papazian, D. M., Jan, Y. N. and Jan, L. Y. (1988). Expression of functional potassium channels from Shaker cDNA in Xenopus oocytes. Nature. 331, 143-145.

Tinker, A., Jan, Y. N. and Jan, L. Y. (1996). Regions responsible for the assembly of inwardly rectifying potassium channels. Cell. 87, 857-868.

252 Tokuyama, Y., Fan, Z., Furuta, FL, Makielski, J. C., Polonsky, K. S., Bell, G. I. and Yano, H. (1996). Rat inwardly rectifying potassium channel Kir6.2: cloning electrophysiological characterization, and decreased expression in pancreatic islets of male Zucker diabetic fatty rats. Biochemical and Biophysical Research Communications. 220, 532-538.

Toneguzzo, F., Hayday, A. C. and Keating A. (1986). Electric field-mediated DNA transfer: transient and stable gene expression in human and mouse lymphoid cells. Molecular and Cellular Biology. 6,716-718.

Toney, M. D., Hohenester, E., Cowan, S. W. and Jansonius, J. N. (1993). Dialkylglycine decarboxylase structure: bifunctional active site and alkali metal sites. Science. 261, 756-759.

Topert, C., Doring, F., Wischmeyer, E., Karschin, C., Brockhaus, J., Ballanyi, K., Derst, C. and Karschin, A. (1998). Kir2.4: A novel K+ inward rectifier channel associated with motoneurons of cranial nerve nuclei. Journal of Neuroscience. 18, 4096-4105.

Toro, L., Wallner, M., Meera, P. and Tanaka, Y. (1998). Maxi-Kca, a unique member of the voltage-gated K channel superfamily. News in Physiological Sciences. 13, 112-117.

Towbin, H., Staehelin, T. and Gordon, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proceedings of the National Academy of Sciences USA. 76, 4350-4354.

Trautmann, A. and Siegelbaum, S. A. (1983). The influence of membrane isolation on single acetylcholine-channel current in rat myotubes. In Single Channel Recording. Ed. Sakmann, B. and Neher, E. pp473-480. Plenum Press, New York.

Tsai, T. D., Shuck, M. E., Thompson, D. P., Bienkowski, M. J. and Lee, K. S. (1995). Intracellular FT1' inhibits a cloned rat kidney outer medulla K+ channel expressed in Xenopus oocytes. American Journal of Physiology. 268. Cell Physiology. 37, C1173-C1178.

Tsaur, M. L., Menzel, S., Lai, F. P., Espinosa, R., Concannon, P., Spielman, R. S., Hanis, C. L., Cox, N. J., Le Beau, M. M., German, M. S., Jan, L. Y., Bell, G. I. and Stoffel, M. ( 1 9 9 5 ) . Isolation of a cDNA clone encoding a KA t p channel-like protein expressed in insulin-secreting cells, localization of the human gene to chromosome band 21q22.1, and linkage studies with NIDDM. Diabetes. 44, 5 9 2 - 5 9 6 .

Tu, L. and Deutsch, C. (1999). Evidence for dimerization of dimers in K+ channel assembly. Biophysical Journal. 76, 2004-2017.

Tu, L., Santarelli, V., Sheng, Z. F., Skach, W., Pain, D. and Deutsch, C. (1996). Voltage-gated K+ channels contain multiple intersubunit association sites. Journal of Biological Chemistry. 271, 18904-18911.

Tucker, S. J., Bond, C. T., Herson, P., Pessia, M. and Adelman, J. P. (1996). Inhibitory interactions between two inward rectifier K+ channel subunits mediated by the transmembrane domains. Journal of Biological Chemistry. 271, 5866-5870.

Tucker, S. J., Gribble, F. M., Zhao, C., Trapp, S. and Ashcroft, F. M. (1997). Truncation of Kir6.2 produces ATP-sensitive K+ channels in the absence of the sulphonylurea receptor. Nature. 387, 179-183.

253 Tucker, S. J., Imbrici, P., Salvatorre, L., D’Adamo, M. C. and Pessia, M. (2000). pH dependence of the inwardly rectifying potassium channel, Kir5.1, and localization in renal tubular epithelia. Journal of Biological Chemistry. 215, 16404-16407.

Tygat, J., Buyse, G., Eggermont, J., Droogmans, G., Nilius, B. and Daenens, P. (1996). Do voltage-gated K vl.l and inward rectifier Kir2.1 potassium channels form heteromultimers? FEBS Letters. 390, 280-284.

Urban, B. W. and Hladky, S. B. (1979). Ion transport in the simplest single file pore. Biochemica et Biophysica Acta. 554, 410-429.

Vandenberg, C. A. (1987). Inward rectification of a potassium channel in cardiac ventricular cells depends on internal magnesium ions. Proceedings of the National Academy o f Sciences USA. 84, 2560-2564.

Vandenberg, J. I., Walker, B. D. and Campbell, T. J. (2001). HERG K+ channels: friend and foe. Trends in Pharmacological Sciences. 22, 240-246.

Vandlen, R. L., Wu, W. C., Eisenach, J. C. and Raftery, M. A. (1979). Studies of the composition of purified Torpedo califomica acetylcholine receptor and of its subunits. Biochemistry. 18, 1845-1854.

Velimirovic, B. M., Gordon, E. A., Lim, N. F., Navarro, B. and Clapham, D. E. (1996). The K+ channel inward rectifier subunits form a channel similar to neuronal G protein-gated K+channel. FEBS Letters. 379,31-37.

Vergara, C. and Latorre, R. (1983). Kinetics of Ca2+-activated K+ channels from rabbit muscle incorporated into planar bilayers. Evidence for a Ca and2 +Ba blockade.2+ Journal of General Physiology. 82, 543-568.

Vergara, C., Latorre, R., Marrion, N. V. and Adelman, J. P. (1998). Calcium- activated potassium channels. Current Opinion in Neurobiology. 8, 321-329.

Wallner, M., Meera, P. and Toro, L. (1996). Determinant for (3-subunit regulation in high-conductance voltage-activated and Ca2+-sensitive K+ channels: An additional transmembrane region at the N terminus. Proceedings of the National Academy of Sciences USA. 93, 14922-14927.

Wang, W. and Giebisch, G. (1991). Dual effect of adenosine triphosphate on the apical small conductance K+ channels of rat cortical collecting duct. Journal of General Physiology. 98,35-61.

Wang, W-H., Lin, D-H. and Sterling, H. (2002). Regulation of ROMK channels by protein tyrosine kinase and tyrosine phosphatase. Trends in Cardiovascular Medicine. 12, 138-142.

Wang, H. S., Pan, Z., Shi, W., Brown, B. S., Wymore, R. S., Cohen, I. S., Dixon, J. E. and McKinnon, D. (1998). KCNQ2 and KCNQ3 potassium channel subunits: molecular correlates of the M-channel. Science. 282,1890-1893.

Wang, W., Sackin, H. and Giebisch, G. (1992). Renal potassium channels and their regulation. Annual Review of Physiology. 54,81-96.

254 Wang, W., Schwab, A. and Giebisch, G. (1990). Regulation of small-conductance K+ channel in apical membrane of rat cortical collecting tubule. American Journal of Physiology. 259, F494-F502.

Warmke, J. W., Drysdale, R and Ganetzky, B. (1991). A distinct potassium channel polypeptide encoded by the Drosophila eag locus. Science. 252, 1560-1562.

Warmke, J. W. and Ganetzky, B. (1994). A family of potassium channel genes related to eag in Drosophila and mammals. Proceedings of the National Academy of Sciences USA. 91,3438-3442.

Wei, A., Covarrubias, M., Butler, A., Baker, K., Pak, M. and Salkoff, L. (1990). K+ current diversity is produced by an extended gene family conserved in Drosophila and mouse. Science. 248, 599-603.

Wei, A., Jegla, T. and Salkoff, L. (1996). Eight potassium channel families revealed by the C.elegans genome project. Neuropharmacology. 35, 805-829.

Weidman, S. (1955). Rectifier properties of Purkinje fibres. American Journal of Physiology. 183,671-.

Welling, P. A. (1997). Primary structure and functional expression of a cortical collecting duct Kjr channel. American Journal of Physiology. 273, F825-F836.

Wible, B. A., De Biasi, M., Majumder, K., Taglialatela, M. and Brown, A. M. (1995). Cloning and functional expression of an inwardly rectifying K+ channel from human atrium. Circulation Research. 76, 343-350.

Wible, B. A., Taglialatela, M., Ficker, E. and Brown, A. M. (1994). Gating of inwardly rectifying K+ channels localized to a single negatively charged residue. Nature. 371, 246-249.

Wickenden, A. D. (2002). K+channels as therapeutic drug targets. Pharmacology & Therapeutics. 94, 157-182.

Wickenden, A. D., Zou, A., Wagoner, P. K. and Jegla, T. (2001). Characterization of KCNQ5/Q3 potassium channels expressed in mammalian cells. British Journal of Pharmacology. 132, 381-384.

Wickman, K. D., Iniguez-Lluhi, J. A., Davenport, P. A., Taussig, R., Krapivinsky, G. B., Linder, M. E., Gilman, A. G. and Clapham, D. E. (1994). Recombinant G-protein py- subunits activate the muscarinic-gated atrial potassium conductance. Nature. 368, 255-257.

Wickman, K., Karschin, C., Karschin, A., Picciotto, M. R. and Clapham, D. E. (2000). Brain localization and behavioural impact of the G-protein-gated K+ channel subunit GIRK4. Journal of Neuroscience. 20, 5608-5615.

Wischmeyer, E., Doring, F. and Karschin, A. (1998). Acute suppression of inwardly rectifying Kir2.1 channels by direct tyrosine kinase phosphorylation. Journal of Biological Chemistry. 273, 34063-34068.

255 Wischmeyer, E., Doring, F. and Karschin, A. (2000). Stable cation coordination at a single outer pore residue defines permeation properties in Kir channels. FEBS Letters. 466, 115-120.

Wischmeyer, E. and Karschin, A. (1996). Receptor stimulation causes slow inhibition of IRK1 inwardly rectifying K+ channels by direct protein kinase A-mediated phosphorylation. Proceedings of the National Academy of Sciences USA. 93, 5819-5823.

Wischmeyer, E., Lentes, K. U. and Karschin, A. (1995). Physiological and molecular characterization of an IRK-type inward rectifier K+ channel in a tumour mast cell line. Pflugers Archives. 429,809-819.

Wolfenden, R. V., Cullis, P. M. and Southgate, C. C. F. (1979). Water, protein folding and the genetic code. Science. 206, 575-577.

Wood, L. S., Tsai, T. D., Lee, K. S. and Vogeli, G. (1995). Cloning and functional expression of a human gene, hIRKI, encoding the heart inward rectifier K+ channel. Gene. 163,313-317.

Woodhull, A. (1973). Ionic blockage of sodium channels in nerve. Journal o f General Physiology. 61, 687-708.

Wu, C. F. and Haugland, F. N. (1985). Voltage clamp analysis of membrane currents in larval muscle fibers of Drosophila: alteration of potassium currents in Shaker mutants. Journal of Neuroscience. 5, 2626-2640.

Xia, X.-M., Fakler, B., Rivard, A., Wayman, G., Johnson-Pais, T., Keen, J. E., Ishii, T., Hirschberg, B., Bond, C. T., Lutsenko, S., Maylie, J. and Adelman, J. P. (1998). Mechanism of calcium gating in small-conductance calcium-activated potassium channels. Nature. 395, 503-507.

Xu, H., Cui, N., Yang, Z., Qu, Z. and Jiang, C. (2000a). Modulation of Kir4.1 and Kir5.1 by hypercapnia and intracellular acidosis. Journal of Physiology. 524.3, 725-735.

Xu, H., Yang, Z., Cui, N., Chanchevalap, S., Valesky, W. W. and Jiang, C. (2000b). A single residue contributes to the difference between Kir4.1 and K irl.l channels in pH sensitivity, rectification and single channel conductance. Journal of Physiology. 528.2, 267- 277.

Xu, Z-C., Yang, Y. and Hebert, S. C. (1996). Phosphorylation of the ATP-sensitive, inwardly rectifying K+ channel, ROMK, by cyclic AMP-dependent protein kinase. Journal of Biological Chemistry. 271, 9313-9319.

Xu, J., Yu, W., Jan, J. N., Jan, L. Y. and Li, M. (1995). Assembly of voltage-gated potassium channels. Conserved hydrophilic motifs determine subfamily specific interactions between a-subunits. Journal of Biological Chemistry. 270, 24761-24768.

Yamada, M., Inanobe, A. and Kurachi, Y. (1998). G protein regulation of potassium ion channels. Pharmacological Reviews. 50, 723-757.

Yamada, M., Isomoto, S., Matsumoto, S., Kondo, C., Shindo, T., Horio, Y. and Kurachi, Y. (1997). Sulphonylurea receptor 2B and Kir6.1 form a sulphonlyurea-sensitive but ATP-insensitive K+channel. Journal of Physiology. 499,715-720.

256 Yamada, M. and Kurachi, Y. (1995). Spermine gates inward-rectifying muscarinic but not ATP-sensitive K+ channels in rabbit atrial myocytes. Journal of Biological Chemistry. 270, 9289-9294.

Yang, Z. and Jiang, C. (1999). Opposite effects of pH on open-state probability and single channel conductance of Kir4.1 channels. Journal of Physiology. 520, 921-927.

Yang, J., Jan, Y. N. and Jan, L. Y. (1995a). Control of rectification and permeation by residues in two distinct domains in an inward rectifier K+ channel. Neuron. 14, 1047- 1054.

Yang, J., Jan, Y. N. and Jan, L. Y. (1995b). Determination of the subunit stoichiometry of an inwardly rectifying potassium channel. Neuron. 15, 1441-1447.

Yang, D., Sun, F., Thomas, L. L., Offord, J., MacCallum, D. K., Dawson, D. C., Hughes, B. A. and Ernst, S. A. (2000b). Molecular cloning and expression of an inwardly rectifying K+ channel from bovine corneal endothelial cells. Investigative Ophthalmology and Visual Science. 41, 2936-2944.

Yang, Z., Xu, H., Cui, N., Qu, Z., Chanchevalap, S., Shen, W. and Jiang, C. (2000a). Biophysical and molecular mechanisms underlying the modulation of heteromeric Kir4.1- Kir5.1 channels by CO2 and pH. Journal of General Physiology. 116, 33-45.

Yang, J., Yu, M., Jan, Y. N. and Jan, L. Y. (1997). Stabilization of ion selectivity filter by pore loop ion pairs in an inwardly rectifying potassium channel. Proceedings o f the National Academy of Sciences USA. 94, 1568-1572.

Yano, H., Philipson, L. H., Kugler, J. L., Tokuyama, Y., Davis, E. M., Le Beau, M. M., Nelson, D. J., Bell, G. I. and Takeda, J. (1994). Alternative splicing of human inwardly rectifying K+channel ROMK1 mRNA. Molecular Pharmacology. 45,854-860.

Yantani, A., Hamm, H., Codina, J., Mazzoni, M. R., Bimbaumer, L. and Brown, A. M. (1988a). A monoclonal antibody to the a subunit of Gk blocks muscarinic activation of atrial K+channels. Science. 241,828-831.

Yantani, A., Mattera, R., Codina, J., Graf, R., Okabe, K., Padrell, E., Iyengar, R., Brown, A. M. and Bimbaumer, L. (1988b). The G protein-gated atrial K+ channel is stimulated by three distinct Gj alpha subunits. Nature. 15, 680-682.

Yellen, G. (1998). The moving parts of voltage-gated ion channels. Quarterly Reviews of Biophysics. 31, 239-295.

Yellen, G., Jurman, M. E., Abramson, T. and MacKinnon, R. (1991). Mutations affecting internal TEA blockade identify the probable pore-forming region of a K+ channel. Science. 251, 939-942.

Yellen, G., Sodickson, D., Chen, T-Y. and Jurman, M. E. (1994). An engineered cysteine in the external mouth of a K+ channel allows inactivation to be modulated by metal binding. Biophysical Journal. 66, 1068-1075.

Yi, B. A., Lin, Y-F., Jan, Y. N. and Jan, L. Y. (2001). Yeast screen for constitutively active mutant G protein-activated potassium channels. Neuron. 29, 657-667.

257 Yifrach, O. and MacKinnon, R. (2002). Energetics of pore opening in a voltage- gated K+channel. Cell. 111,231-239.

Yool, A. J. and Schwarz, T. L. (1991). Alteration of ionic selectivity of a K+ channel by mutation of the H5 region. Nature. 349, 700-704.

Yoshii, K. and Kurihara, K. (1989). Inward rectifier produced by Xenopus oocytes injected with mRNA extracted from carp factory epithelium. Synapse. 3, 234-238.

Zagotta, W. N., Hoshi, T. and Aldrich, R. W. (1990). Restoration of inactivation in mutants of Shaker potassium channels by a peptide derived from ShB. Science. 250, 568-571.

Zaritsky, J. J., Eckman, D. M., Wellman, G. C., Nelson, M. T. and Schwarz, T. L. (2000). Targeted disruption of Kir2.1 and Kir2.2 genes reveals the essential role of the inwardly rectifying K+ current in K+-mediated vasodilation. Circulation Research. 87, 160- 166.

Zaritsky, J. J., Redell, J. B., Tempel, B. L. and Schwarz, T. L. (2001). The consequences of disrupting cardiac inwardly rectifying K+ current (I k i ) as revealed by the targeted deletion of the murine Kir2.1 and Kir2.2 genes. Journal of Physiology. 533.3, 697- 710.

Zerangue, N., Schwappach, B., Jan, Y. N. and Jan, L. Y. (1999). A new ER

trafficking signal regulates the subunit stoichiometry of plasma membrane K a t p channels. Neuron. 22, 537-548.

Zhang, H., He, C., Yan, X., Mirshahi, T. and Logothetis, D. E. (1999). Activation of inwardly rectifying K+ channels by distinct PtdIns(4,5)P2 interactions. Nature Cell Biology. 1, 183-188.

Zhang, H-J., Liu, Y., Ziihlke, R. D. and Joho, R. H. (1996). Oxidation of an engineered pore cysteine locks a voltage-gated K+ channel in a nonconducting state. Biophysical Journal. 71, 3083-3090.

Zhou, H., Chepilko, S., Schtitt, W., Choe, H., Palmer, L.G. and Sackin, H. (1996). Mutations in the pore region of ROMK enhance Ba2+ block. American Journal of Physiology. 271, C1949-C1956.

Zhou, Y., Morais-Cabral, J. H., Kaufman, A. and MacKinnon, R. (2001b). Chemistry of ion coordination and hydration revealed by a K+ channel-Fab complex at 2.0A resolution. Nature. 414, 43-48.

Zhou, M., Morais-Cabral, J. H., Mann, S. and MacKinnon, R. (2001a). Potassium channel receptor site for the inactivation gate and quaternary inhibitors. Nature. 411, 657- 661.

Zhou, H., Tate, S. S. and Palmer, L. G. (1994). Primary structure and functional properties of an epithelial K channel. American Journal of Physiology. 266. Cell Physiology. 35, C809-C824.

258 Zhou, X. L., Vaillant, B., Loukin, S. H., Kung, C. and Saimi, Y. (1995). YKC1 encodes the depolarization-activated K+ channel in the plasma membrane of yeast. FEBS Letters. 373, 170-176.

Zhu, G., Chanchevalap, S. and Jiang, N. C. C. (1999a). Effects of intra- and extracellular acidifications on single channel Kir2.3 currents. Journal of Physiology. 516.3, 699-710.

Zhu, G., Qu, Z., Cui, N. and Jiang, C. (1999b). Suppression of Kir2.3 activity by protein kinase C phosphorylation of the channel protein at threonine 53. Journal o f Biological Chemistry. 274, 11643-11646.

Zhu, L., Wu, X., Wu, M. B., Chan, K. W., Logothetis, D. E. and Thornhill, W. B. (2001). Cloning and characterization of G protein-gated inward rectifier K+ channel (GIRK1) isoforms from heart and brain. Journal of Molecular Neuroscience. 16, 21-32.

259