PROPERTIES of DRUG BLOCKADE of a LARGE CONDUCTANCE CALCIUM-ACTIVATED POTASSIUM CHANNEL in CULTURED RAT HIPPOCAMPAL NEURONS by XUEPING WANG M.D

PROPERTIES OF DRUG BLOCKADE OF A LARGE CONDUCTANCE CALCIUM-ACTIVATED POTASSIUM CHANNEL IN CULTURED RAT HIPPOCAMPAL NEURONS by XUEPING WANG M.D. Beijing Medical University, 1984 M.Sc. Beijing Medical University, 1989

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF

THE REQUIREMENTS FOR THE DEGREE OF MASTER'S OF SCIENCE

in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF PHARMACOLOGY & THERAPEUTICS, FACULTY OF MEDICINE

We accept this thesis as conforming

to the required standard

THE UNIVERSITY OF BRITISH COLUMBIA

AUGUST 1991

©XUEPING, WANG, 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.

(Signature)

Department of Pharmacology and Therapeutics Vancouver, Canada

Date July 4th, 1992

DE-6 (2/88) ii

ABSTRACT The objective of these experiments was to investigate the actions of a number of novel putative class III antiarrhythmic agents on the large conductance calcium-activated potassium channel (BKca) in cultured hippocampal neurons. The experiments were carried out in three steps. In the first set of experiments BKca channel, isolated from cultured rat hippocampal CA1 neuronal somatic membrane, was first identified and its physiological and pharmacological properties were characterized at a single channel level. This BKca channel had an average single channel conductance of 80 pS with physiological transmembrane K+ (140 mM inside and 5 mM outside). It was very selective to K+over Na+, a ratio of 100 to 7 was determined from this experiment. Calcium, at the internal side of the membrane, was necessary to activate the BKcaa channel. With low internal calcium concentration, depolarization could promote the rate of channel openings. An e-fold increase in p0 was found with 1 n M internal calcium and 15 mV depolarization. With 200 nM internal calcium, p0 was virtually independent of voltage. 0.1 mM external TEA showed fast blockade of this channel. Internal TEA and 4-AP (internal or external) showed no effect on BKca. In the second set of experiments, the actions of putative class III drugs on BKca was studied. The drugs were RP-62719, UK-68798, tedisamil (KC-8857) and risotilide (WY-48986) at concentrations 0.1 ~ 10uM. All these agents, applied both to the inside or to the outside of the patch membrane, resulted in the opening of the BKca channel to exhibit rapid flicking from open to nonconducting levels. This effect was dose-dependent and for KC-8857 and UK-68798 was evident at concentrations of 0.1 uM. The blocking rate constants were determined from a simple open channel blockade scheme and were not dependent on voltages. Single channel conductance and ionic selectivity were not affected by the drugs. The potencies for channel block of the drugs acting either externally or internally iii were in the order UK-68798>tedisamil>RP-62719> risotilide with UK-68798 reducing the mean open time of BKca by one-half at a concentration near 0.4/iM. In the final set of experiments, the thermodynamics associated with RP- 62719 block BKca channel was studied in order to better understand the molecular mechanisms of channel block. The Q10 associated with the channel mean open time was found to be 2.2 with 5 /iM RP-62719 at the inner surface of the patch membrane. The blocking and unblocking rate constants were determined using the simple open channel block scheme. Thermodynamic analysis, using transition rate theory, showed that the blocking rate constant was associated with a large increase in entropy. The relatively high temperature dependence for channel blockade was not consistent with a rate-limiting process established by simple diffusion of the agent to a channel blocking site. Channel block may involve conformational changes in the channel protein as a consequence of hydrophobic interactions between drug and channel sites. iv

TABLE OF CONTENTS

CHAPTER PAGE n ABSTRACT ii IV LIST OF TABLES vii V LIST OF FIGURES viii VI LIST OF LEGENDS x VH ACKNOWLEDGEMENT xii 1 INTRODUCTION 1

1.1 BACKGROUNDS OF SINGLE KCa CHANNEL STUDIES 1

1.1.1 KCa Classifications 1

1.1.2 BKCa Channels 2 1.1.2.1 BKca Channel Distributions and Function 2 1.1.2.2 Ca2+ Sensitivity and Selectivity of Ca2 + -Binding Sites 3

1.1.2.3 BKCa Channel Voltage Dependence 5 1.1.2.4 BKca Channel Selectivity and Conductance 5

1.1.2.5 BKCa Channel Gating Kinetics 6

1.1.2.6 BKCa Channel Pharmacology 7

2+ 2+ 1.1.2.6.1 Na + , Ca and Ba Blockade of BKCa 7

1.1.2.6.2 TEA Blockade of BKCa 8

1.1.2.6.3 Toxin Blockade of BKCa 8

1.1.3 SKCa Channels 10

1.1.3.1 SKfja Channels Distributions and Function 10 1.1.3.2 SKca Conductance and Calcium Dependence 11 TABLE OF CONTENTS (CONT'S)

CHAPTER PAGE

1.1.3.3 SKCa Channel Pharmacology 11 1.1.4 Ca2+-activated K+ Currents in Hippocampal Neurons 12 1.2 SEQUENTIAL OPEN CHANNEL BLOCKADE MODEL 12

1.3 PATCH CLAMP TECHNIQUE 13 1.4 SYNTHETIC CLASS III ANTIARRHYTHMIC DRUGS 14 1.5 THERMODYNAMIC ANALYSIS 15 2 METHODS 19 2.1 CELL CULTURE PREPARATION 19 2.2 ELECTROPHYSIOLOGY 20 2.2.1 Glass and Pipettes 20 2.2.2 Mechanical Setup 21 2.2.3 Single Channel Patch-Clamp Configurations 21 2.2.4 Solutions Used 22 2.2.5 Recordings 25

2.2.6 Analysis Procedures 25

2.3 TEMPERATURE STUDIES 30 2.3.1 Equations for Thermodynamic Studies 31

3 RESULTS 33

3.1 IDENTIFICATION AND PROPERTIES OF SINGLE Krja CHANNELS 33 IN CULTURED RATCAl HIPPOCAMPAL NEURONS 3.1.1 Identification: Selectivity and Channel Conductance 33

3.1.2 Calcium Dependence 38

3.1.3 Voltage Dependence 39 vi

TABLE OF CONTENTS (CONT'S)

CHAPTER PAGE

3.1.4 Channel Kinetics 46

3.1.5 Pharmacology of BKCa Channel 49

3.2 ACTIONS OF ANTIARRHYTHMIC AGENTS ON BKCa 52 CHANNELS: SINGLE CHANNEL ANALYSIS 3.2.1 Inside-out Experiments 55 3.2.2 Outside-out Patch Experiments 63 3.3 THERMODYNAMIC STUDIES 69

3.3.1 Temperature Dependence of KCa Channel 72 3.3.2 Temperature Dependence of Drug Actions 79 4 DISCUSSION 91

4.1 BKCa CHANNEL PROPERTIES 91 4.2 OPEN CHANNEL BLOCKADE AS THE MECHANISM OF 92

CLASS III ANTIARRHYTHMICS BLOCK BKCa CHANNELS 4.3 MOLECULAR MECHANISM OF RP-62719 BLOCKADE 95

OF BKCa CHANNELS 5 REFERENCES 101 vii LIST OF TABLES

TABLE PAGE

Chemical Components of Solutions Used for Cell- 23 Attached and Excised Patch-Clamp Single Channel Studies

Normalized Mean Open Time, Onward (Blocking) 68 Rate Constants for Drugs

Transition State and Associated State 90 Thermodynamic Parameters viii

LIST OF FIGURES

FIGURE PAGE

1 Chemical Structures of Drugs Used for Experiments 24 2 Typical Records of Unitary Currents Passing 34 Through BKca in the Somatic Membrane of the Cultured Rat Hippocampal Neurons.

3 Current-Voltage Plots: KCa Channel Conductance 36 and Ion Selectivity

4 Calcium Dependency of BKCa Channel Behaviour 40

5 Calcium Dependency of BKCa Channel Kinetics 42-44

6 Voltage Dependence of BKCa 47 7 Kinetic Analysis of m 50

8 Effect of TEA on BKCa 53

9 Effect of RP-62719 on BKCa 56

10 Effect of RP-62719 on BKCa Channel Kinetics and 58 Current Amplitude 11 TO'-Drug Concentration Relationships 61

12 Voltage Dependence of RP-62719 Effect on BKCa 64 13 Effect of UK-68798, Tedisamil and Risotilide on 66 BKCa Channels

14 Effect of Drugs on BKCa at V = 0 mV; Outside-Out 70 Experiments

15 Temperature Dependence of BKCa Single-Channel 73 Currents

16 Temperature Dependence of BKCa Channel Mean 75 Open Time ix

LIST OF FIGURES (Cont's)

FIGURE PAGE

17 Effect of Temperature on BKrja Channel Kinetics 77 and Single-Channel Current Amplitudes 18 Temperature Dependence of RP-62719 Effects on 80 BKca Channel Behaviour 19 Temperature Dependence of RP-62719 Effects on 82 BKrja Channel Kinetics 20 Temperature Dependence of RP-62719 Effects on 84 BKfja Channel Mean Open Time (T0)

21 Arrhenius Plot for k2 87 X LIST OF LEGENDS

FIGURE PAGE

2 Typical Records of Unitary Currents Passing 35 Through Kfja in the Somatic Membrane of the Cultured Rat Hippocampal Neurons

3 Current-Voltage Plots: KQ3 Channel Conductance 37 and Ion Selectivity

4 Calcium Dependency of BKCa Channel Behaviour 41

5 Calcium Dependency of BKfja Channel Kinetics 45

6 Voltage Dependence of BKca 48

7 Kinetic Analysis of BKCa 51

8 Effect of TEA on BKCa 54

9 Effect of RP-62719 on BKCa 57

10 Effect of RP-62719 on BKCa Channel Kinetics and 59 Single-Channel Current Amplitude

11 To'- Drug Concentration Relationships 62

12 Voltage Dependence of RP-62719 Effect on BKCa 65

13 Effect of UK-68798, Tedisamil and Risotilide on 67 BKfja Channel Behaviour

14 Effect of Drugs on BKCa at V = 0 mV; outside-out 71 Experiments

15 Temperature Dependence of BKfja Single-Channel 74 Currents in Control Solution

16 Temperature Dependence of BKCa Channel Mean 76 Open Time (To)

17 Effect of Temperature on BKca Channel Kinetics 78 and Single-channel Current Amplitude

18 Temperature Dependence of RP-62719 Effect on 81 BKfja Channel Behaviour xi LIST OF LEGENDS (Cont's)

FIGURE PAGE

19 Temperature Dependence of RP-62719 Effect on 83 BKCa Channel Kinetics 20 Temperature Dependence of RP-62719 Effect on 85 BKCa Channel Mean Open Time (T0)

21 Arrhenius Plot for k2 88 xii

ACKNOWLEDGMENTS:

I would like to express my sincere appreciation to Dr. James G. McLamon for giving me the opportunity to work in his laboratory and for his continuing advice and guidance throughout my graduate studies.

Thanks must also be extended to the faculty members of the department for their commitment to teaching, which enriched my knowledge in Pharmacology. I must also express my gratitude to the members of my examining committee for their valuable advice. Special thanks to Dr. Morley C. Sutter for his encouragement and for teaching me how to handle literature efficiently.

"Thank-you" also goes to Dr. Peter B. Reiner for letting me tap his knowledge in handling scientific figures and to Mr. Andrew Laycock for helping me making projective slides.

Last, but not least, I would like to thank my dear husband Dr. Jiren Tan. Mere words can not express my emotion towards him. Without his love, support and belief in me, I would never have made it as far as I have. Wang, X.P. l

1. INTRODUCTION

1.1 BACKGROUNDS OF SINGLE KCa CHANNEL STUDIES

1.1.1 Kfa Classifications:

Calcium-activated potassium currents (Kca) were first described in erythrocytes (Gardos, 1958), and later identified in molluscan nerve cells (Meech and Strumwasser, 1970; Meech and Standen, 1975; Heyer and Lux, 1976; Gorman and Thomas, 1980) on the basis of the effects induced by calcium channel blockade or by alteration of the intracellular calcium concentration. Similar calcium-dependent currents have then been found in almost every cell type examined: the nerve cells (Moolenaar and Spector, 1978; Adams, Constanti, Brown and Clark, 1982), endocrine and exocrine tissues (Marty, 1981; Wong, Lecar and Adler, 1982; Maruyama and Petersen, 1982; Petersen and Maruyama 1984), macrophages (Gallin, 1984), rat skeletal muscle (Pallotta, Magleby and Barrett 1981; Barrett,

Magleby and Pallotta, 1982) and smooth muscle cells (Singer and Walsh, 1987). Kca channels may provide a link between cell metabolism and membrane potential (Latorre, Oberhauser, Labarca and Alvarez, 1989) and in some cells are demonstrated to contribute to spike generation (MacDermott and Weight, 1982; Brown et al. 1983; Grega and MacDonald, 1987), transient hyperpolarizations (Brown et al., 1983; Blatz and Magleby, 1984 & 1986; Lancaster and Nicoll, 1987), the resting potential (Bourque, 1988), fluid secretion (Wong and Lecar, 1982; Petersen and Maruyama, 1984), control of cell volume (Christensen, 1987) and pacemaker activity (Bourque, 1988).

With the advent of the patch-clamp technique, single KQQ channel conductance has been determined. Based on conductance differences there are at least two types of Kca channels: the large-conductance Kca channel (BKca or maxi-Kcg channel)

+ and the low conductance K channel (SKca). The discovery of two invertebrate toxins, apamin and charybdotoxin, which are able to block Kca channels with Wang, X.P. 2

considerable potency provided a pharmacological method to distinguish the different

types of Kca channels. However the tentative classification of Kca channels may require future revision as the structures and the amino acid sequences of the channels become known. 1.1.2 B^ Channels:

1.1.2.1 BKra Channel Distributions and Function

BKca channels, with single channel conductance generally in excess of 120 pS in symmetrical K+ concentration of -140 mM across the membrane, were the first to be observed both in the cell membranes of chromaffin cells (Marty, 1981), in a pituitary cell line (Wong, 1982), in the planar lipid bilayers incorporated with this channel from rat brain (Krueger, 1982) and in transverse tubule membranes isolated from rabbit skeletal muscle (Pallotta, 1981; Latorre, 1982). They are widely distributed in a very large range of excitable and inexcitable cells. Patch-clamp studies have shown that in neurons activation of BKCa underlies lc [fast phase of action potential repolarization (Adams, 1982)]; in exocrine gland cells (in pancreas acinar cells), they mediate the hyperpolarization effect of secretagogues (acetylcholine or cholecystokinin) (Petersen, 1987, 1985); in the basolateral plasma membrane of

+ salivary gland acini, the activity of BKca explains the K loss induced by nerve stimulation or by direct stimulation with secretagogues (Petersen, 1984); in lacrimal glands, carbamoylcholine (cholinergic) induces activation of BKrja channels by increasing the [Ca2+]j, allowing the extrusion of K+ during secretion of tears (Trantman and Marty, 1984). In pancreatic B cells it was thought that insulin release promoted by glucose and quinine was due to a blockade of the BKCa (Atwater, 1979). However, since the finding that glucose inhibits ATP sensitive K+ channels, the role of BKca is under re-interpretation (Petersen, 1984). BKca probably provides a feedback control of Ca2+ uptake (Findlay and Petersen, 1985) and hence, they may play a key role in generating the burst pattern of electrical activity which leads to a Wang, X.P. 3

pulsatile release of insulin. BKca channels also provide a passive transport pathway from the principal cell to the lumen in rabbit renal collecting duct anlagen (Gitter,

1987). Similarly, BKca channels have been identified as the electrodiffusive pathway for K+ ions in the epithelium that secrets cerebrospinal fluid into the ventricles of the brain from the blood (Christens, 1987). 1.1.2.2 Ga2+ Sensitivity and Selectivity of Ca2+-Bindina Sites: 2+ 2+ The gating of BKca channel is Ca dependent. The binding of several Ca ions are required to open fully the channel.The Hill coefficients that best describe the 2+ probability of the open state (p0) vs intracellular calcium concentration ([Ca ]j) curves are in the range 2-4 . This result is usually interpreted as a need for two to four Ca2+ to bind for maximal channel activation (Moczydlowski and Latorre, 1983), However, in the presence of millimolar concentrations of intracellular Mg2+, the Hill coefficients can increase up to 6 (Golowasch, Kirkwood and Miller 1986). Mg2+ itself does not activate the channel in the absence of Ca2+. It was suggested that Mg2+ may unmask Ca2+-binding sites which are not operational in the absence of Mg2+ (Golowasch et al 1986). The steepening of the relationship between [Ca2+]j and channel open state probability (P0) would make the cell more sensitive to small changes in [Ca2+]j. 2+ Maximal sensitivity to [Ca ]j of BKca also varies widely from one cell type to another. In mammalian salivary glands (Petersen, 1984), maximum activation is 2+ 8 7 achieved in the [Ca ]j range of 10" - 10" M; the BKca channel can be activated by membrane depolarization in the virtual absence of [Ca2+]j. Similar [Ca2+]j levels are required for activation of the BKca channels in smooth muscle (Benham, et al., 1986), pancreatic acinar cells (Maruyama, et al., 1983), chromaffin cells (Marty, 1981) and clonal anterior pituitary glands (Wong, 1982). In contrast, with cultured rat myotubes (Barrett et al., 1982) and T-tubules (Latorre et al., 1982), [Ca 2+] j of 10 ^ ~ 104 M) is needed to obtain maximum activation of the BKca channel. Wang, X.P. 4

The surface charge of lipid surrounding the BKQ,, channel may be involved in controlling Ca2+ sensitivity (minimum [Ca2+]j required to activate the channel). Moczydlowski and Latorre (1985) found that negatively charged lipids increase the 2+ 2+ apparent Ca sensitivity. Interestingly, the Ca sensitivity of BKca channels in spinal neurons was observed to be dependent on development (Blair et al., 1985). In young neurons, Ca2+ sensitivity is poor or absent, whereas in mature cells raising the [Ca2+]j increases the probability of opening. This increase in calcium sensitivity is related to the disappearance of the Ca2+-dependent plateau of the action potential in these cells.

BKca channels can also be activated by other divalent ions. In a planar bilayer 2+ system, the effectiveness of activating the BKca by divalent cations other than Ca was in the order Cd2+ > Sr2+ > Mn2+ > Fe2+ > Co2+, Cadmium was about 100-fold less effective than Ca2+. Some other divalent cations such as Mg2+, Ni2+, Ba2+, Cu2+, Zn2+, Hg2+, and Sn2+ were ineffective alone (Oberhauser and Latorre, 1988). Cation size has been suggested to be important for the agonistic effect; only cations with radii >0.072nM (Co2+) or less than 0.113nM (Sr2+) were able to activate the channel. In the presence of [Ca2+]j some divalent cations potentiate Ca2+ activation of 2+ 2+ 2+ 2+ 2+ BKca. These include Cd , Co , Mn , Ni , Mg , which were able to increase the apparent affinity of the channel for Ca2+ and to increase the Hill coefficient in a concentration-dependent fashion. These divalent cations were only effective when added to the cytoplasmic side of the channel, suggesting the existence of an internal modulatory site that controls channel activity. A similar result has been found in rat pancreatic islet beta cells (Joseph, 1989). Mg2+ alone from 1 mM to 50 mM, can not 2+ activate Kca channels incorporated into lipid bilayers, but in the presence of Ca , Mg2+ can increase the channel open state probability (Latorre, 1988). It has also Wang, X.P. 5

been found that millimolar concentrations of [Mg2+]j allowed the maintenance of membrane stability with lower than 1 /iM free [Ca2+]j (Rudy, 1988).

1.1.2.3 BKca Channel Voltage Dependence

The gating of BKca channels is voltage dependent. The voltage dependency of

BKca channels varies with cell types. The probability of open channel state (P0) increases e-fold per 11-15 mV in muscle (Barret, 1982; Methfessel), chromaffin cells (Marty, 1983) and in skeletal muscle channels incorporated into bilayers (Latorre,

1982; Moezydlowski and Latorre, 1983). (P0) increases e-fold per 9 mV in the pituitary cell line (Wong, 1986), and per 10-30 mV in medullary thick ascending limb cells (Guggino, 1987). These results suggest that in various tissues a common gating mechanism is involved with channel activation. The voltage-dependent gating may

account for the long quiescent periods of the BKCa kinetic behavior (Latorre, 1991). 1.1.2.4 BKca Channel Selectivity and Conductance BKca channels are highly selective for K+ among other monovalent ions. The permeability sequence, as determined from bionic potentials, for myotubes (Blatz + + + + + + andMagleby, 1984) isTI (1.1)> K (1)>Rb (0.67)>NH4 (0.11)>Cs ~Na ~Lt .A similar sequence was found with pancreatic islet beta cells (Joseph, 1989). The permeability ratio between Na+ and K+ is less than 0.01 (Blatz, 1984; Cecci, 1986; Singer, 1987; Yellen, 1984; Joseph, 1989). Although, Tl+ is more permeable that K+ judging from the channel reversal potential, the current-voltage relationship in Tl+ shows that the channel conducts Tl+ significantly less well than it conducts K+ (100 pS for Tl+, 200 pS for K+). Hence, the relative cation permeability, as calculated from reversal-potential measurements, may be a poor estimator of relative conductance through the channel (Joseph 1989). Intracellular Na+ ions which hardly permeate, + can in fact block BKca (Yellen, 1984; Joseph 1989). This effect of Na can be + reversed by raising [K ]0 and this observation suggests interactions between ions in Wang, X.P. 6

the channel (Yelien 1984; Eisenman 1986; Cecchi 1987; Joseph, 1989) i.e., BKca channels are not single-ion channels but rather are multionic pores.

The single-channel conductances (G) of BKca channels are characteristically very high. For channels exposed to symmetrical 140 mM K+, G's range from 100 pS to 300 pS. G was found to saturate at 325-350 pS with ratio of K+ concentration + + + ([K ]j/[K ]0) -400 (mM) (Joseph, 1989) and at 640 pS with ratio of [K ] -600 mM (Blatz, 1984). Under a physiological K+ gradient, the conductance is lower, -100 pS (Barrett 1982).

The molecular mechanisms underlying BKca channel selectivity have been studied. Mackinnon and Miller (1988) have shown direct evidence that negative charges are located in the BKca channel conduction system. Trimethyloxonium (TMO) was used, which methylates carboxylic groups. TMO decreases channel conductance and shifts the channel conductance vs [KCI] curve to the right. This implies the presence of a carboxylic group located at the channel entrance that is altering the local electrostatic potential.

The BKca channel possess both a large conductance and a high cation selectivity. This situation is somewhat surprising since a high selectivity implies strong interactions of the permeate ion with the selectivity filter (Latorre, 1983). The complexity is due in part to the multiionic nature of the channel and fixed negative charge in the channel. Ion repulsion in a multiply-occupied pore at high permeate ion concentrations increases the rate of K+ exit, and channel conductance becomes higher than in a single-ion pore. (Latorre 1986,1989 and Cecchi 1986).

1.1.2.5 BKca Channel Gating Kinetics Markov models have been frequently used in describing channel gating. In essence Markov models assume that the rate constants for leaving any given kinetic state are independent of previous channel activity, i.e. rate constants are time- invariant. This model has been proven to be valid for the BKca channel cultured rat Wang, X.P. 7 skeletal muscle by McManus and Magleby (1989). They examined the channel open and closed-time distributions conditional on adjacent interval durations, and found r0 and r c (respective time constants of the open and closed exponential components describing the distributions) to be independent of adjacent interval duration and hence previous channel activity. Since r0 and rc represent rate constants for transitions between the kinetic states, this finding suggests that the rate constants are also independent of previous channel activity. Thus the BKca channel kinetics are consistent with Markov gating.

1.1.2.6 BKCa Channel Pharmacology

2 + 2 + 1.1.2.6.1 Ba and Ca blockade of BKCa

2+ 2+ Non-permeant inorganic cations such as Ca and Ba block BKca channel from the inside. Ba2+ is the most potent ion with blocking actions at 10"7M (Benham et al. 1985; Gitter et al. 1987; Guggino et al. 1987; Hunter et al 1984; Vergara et al. 1983 and 1984). The blockade periods in the presence of Ba2+ can last for several seconds and are characterized by prolonged non-conducting periods in single- channel patch-clamp recordings. This block is also voltage dependent, indicating binding to a site within the channel. Channel closure can trap Ba2+ in the channel with little or no dissociation from the blocking site until after the channel opens (Miller, 1987; Miller and Latorre, 1987). When the channel is in the closed configuration, the site is inaccessible to Ba2+ from both the internal and external sides, suggesting that Ba2+ binds to the open channel.

2+ In the absence of Ba , the kinetic pattern of the BKca channel also showed long-lived non-conducting states. This feature has been observed in many preparations, such as in clonal anterior pituitary cells (Wong, 1982) muscle cells (Latorre, 1983) and pancreatic islet cells (Findlay, 1985). The slow kinetic process has been interpreted in terms of a Ca2+ blockade, i.e., a voltage-dependent binding of Ca2+ to the conduction system, which blocks the K+ flux (Vergara and Latorre, Wang, X.P. 8

1983). However, Neyon and Miller (1988) have proposed that most of the supposed "Ca2+ blockade" could be due to Ba2+ contamination of the solutions. Thus "Ca2+ blockade" can only be seen at [Ca2+] > 10"4 M in muscle, in which the contaminating Ba2+ is in the VM range. Latorre (1991) have studied the

characteristics of the long closure kinetics in Kca channels incorporated into planar lipid bilayer membrane. In this experiment, in which Ba2+ was buffered to ~3 nM, the slow kinetic processes were not due to a Ba2+ blockade one important functional

2+ role for calcium blockade of BKCa channels at high [Ca ]j is that the BKca channels provide a feedback control of Ca2+ uptake (Findlay, 1985).

1.1.2.6.2 TEA blockade of BKCa

TEA can block BKca channels with both external and internal application. The potency of TEA block was found to be different for different cell types. For example,

in a skeletal muscle preparation, TEA was more effective from the outside (KD = 0.29

mM) than the inside (KD = 45 mM) (Latorre, 1986). In rat anterior pituitary cells, TEA

was more effective from the inside (KD = 0.08 mM) than the outside (KD = 52.2 mM) of the membrane (Wong, 1982, 1986). The main kinetic feature of TEA block was an apparent reduction in single-channel conductance. This result suggested large values for the on- and off-rate constants for channel blockade. In essence the blocking events were so rapid that the unitary transitions were not resolvable in patch-clamp recordings.

1.1,2.6.3.Toxin blockade of KCa

Some scorpion toxins are known to block BKca channels. A toxin obtained from the venom of the scorpion Leiurus quinquestriatus (Miller, 1985; Smith, 1986)

charybdotoxin (CTX) blocks BKCa from the outside with high affinity (KD = 10 nM). Miller (1988) demonstrated that TEA and CTX were mutually exclusive in their binding to the channel "mouth". Using site-directed mutants of the shaker gene, Mackinnon and Miller have shown that glutamate 422, a negatively charged residue placed near Wang, X.P. 9

or in the externally facing mouth of this channel (which has similar properties as Kca channels in other types of tissue) is responsible for the electrostatic focusing of the positively charged CTX peptide toward its blocking site, i.e., CTX binds to the external negative charge of the channel, affecting a physical occlusion of the conduction pathway. This blocking action can be destabilized at the binding site by K+ ions moving through the channel from the opposite side (Mackinnon and Miller, 1988).

It should be noted that CTX may be non-specific in action. For example small conductance Ca2+-activated K+ channels (SK^a) can be blocked by CTX including

SKCa in molluscan nerve (Hermann and Erxleben, 1987) mammalian brain (Farley and Rudy, 1988) and red blood cells (Castle and Strong, 1986). Furthermore not all

BKca channels are sensitive to CTX. Reinhart found recently (1989) that one type of

BKca channel from rat brain is not blocked by charybdotoxin.

BKca channels are also blocked by noxiustoxin, which is also a voltage-

dependent channel blocker, with an affinity of KD = -0.5 A*M (Valdivia, 1988; Slaughter, 1991). Another scorpion toxin, iberiotoxin (IbTX), which is a peptide toxin purified from the venom of the scorpion Buthus famulus and with a high degree of homology with CTX, has been shown to block the BK^g channel incorporated into planar lipid bilayers. It was also found that 1mM [TEA]0 increased the mean blocking time about 4-fold suggesting that IbTX block is due to channel occlusion at a site near the TEA external binding site. Thus, the mode of action of the IbTX is similar to that of CTX, but IbTX needs to overcome a larger energy barrier in order to leave the channel external mouth (Sebastian and Latorre, 1991).

Not all toxins act exclusively from the external site. Dendrotoxin-I (DTX-1), a 60- residue peptide belonging to the dendrotoxin family of Mamba snake neurotoxins and potent inhibitor of various types of voltage-gated K+ currents, was found to alter Wang, X.P. 10

the conductance when applied internally to BKca channels in rat skeletal muscle incorporated into planar bilayers. The binding of DTX-I modified conduction of K+ ions through the pore without affecting the Ca2+ dependence or voltage dependence of gating suggesting a unique internal binding (Kathryn and Moczydlowski 1990).

1.1.3 SKCa Channels

1.1.3.1 SKCa Channels: Distributions and Function

The SKCa channel, conductance range 6-20 pS in symmetrical 140mM KCI solution, can be potently and selectively blocked by apamin (Banks, 1979; Maas, 1980). This channel has been found in many cell types such as in the longitudinal smooth muscle of the intestine, in hepatocytes (Banks, 1979) and in neuroblastoma cells (Huguest, 1982). It is widely distributed and is involved in many important functions. For example, in intestinal longitudinal smooth muscle, its activation mediates the inhibitory action of a!-adrenoceptors, and of the receptors for neurotensin and ATP (P2-subtype). In hepotocytes, it underlies adrenaline hyperkalemia (Coats, 1983, 1985). In rat brain, apamin-binding sites, which are considered as SKCa channel proteins, distributed widely (Mourre, 1986) and an endogenous apamin-like peptide (Lazdunski, 1988) has been found, suggesting that

SKca may play an important role in CNS.

Opening of SKCa channels underlies the prolonged hyperpolarization that follows action potentials in many neurons and myotubes which possess repetitive firing properties. For example, mammalian myotubes in culture contract spontaneously by generating of spontaneous action potentials. During the hyperpolarization phase of this spontaneous action potential, afterhyperpolarization

(lAHP) works as a pace-maker which enable the sodium channel that has been activated and then inactivated during the first action potential to be reactivated then allowing the generation of a new action potential. The lAHP is blocked by apamin concentrations in the 1-10 nM range. The SKca channel is the basis for lAHP in these Wang, X.P. 11

myotubes (Hugues et al., 1982a,b; Romey et al., 1984; Lazdunski et al, 1984 and

1989). The expression of SKca channel in these cells is subject to the regulation of innervation. Innervating rat myotubes by co-culturing them with spinal cord neurons will result in the disappearance of IAHP- This is important functionally, because it ensures the control of skeletal muscle by nerves during development, i.e., the skeletal muscles no longer contract spontaneously after innervation.

1.1.3.2 SKca Conductance and Calcium dependence

The SKca single channel conductance and calcium dependence has been studied in many preparations. Blatz and Magleby (1986) found in cultured rat skeletal 2+ muscle, the conductance of SKCa channel to be 10-14 pS and the Ca -dependence of the SKCa channel to be at EC50 = 200-500 nM. Similar results were found by Capiod and Ogder (1988,1989) in puinea-pig hepatocytes with 6 pS conductance in + + 2+ asymmetrical K and 20 pS in symmetrical K solutions; the EC50 for Ca to activate half-maximally the SKCa channel was 600 nM.

1.1.3.3 SKca Channel Pharmacology Apamin, derived from bee venom, consists of 18 amino-acids, is by far the most active and selective SKca blocker available. Its active site is made of 2 arginines located at positions 13 and 14 (Habermann, 1984). Apamin can selectively and potently block SKCa channels in many cell types with a KD ~ 20 pM and the neuromuscular blocking agent tubocurarine can inhibit apamin binding to hepatocytes, implying a common site of action. This is not surprising since apamin is a positively changed molecule, and tubocurarine is a bis-charged molecule. Tubocurarine and related neuromuscular blockers are less selective and less potent than apamin (Smart, 1987). Another toxin derived from Leiurus quinquestriatus hebrasus can also block the SKCa channel (Abia, 1986; Castle and Strong, 1986) and is distinct from both apamin and charybdotoxin. Cecchi et al. (1988) have purified Leiurotoxin I and sequenced its Wang, X.P. 12

amino acid sequence, which showed little homologue with apamin. Nevertheless, it is able to inhibit apamin binding to rat brain synaptical membranes. SKca is not blocked by TEA (Romey, 1984) or CTX in skeletal muscle. In contrast, Van Rentergehm (1988) has found with an aortic smooth muscle cell line

having a repetitive activity that the IAHP which is responsible for this repetitive activity is blocked by CTX but not by apamin. 1.1.4 Ca2+-Activated Macroscopic K+ Currents in Hippocampal Neurons Hippocampal cells possess several currents gated by intracellular Ca2+, including two K+ currents and a CI" current. These may be activated by the rise in intracellular Ca2+ produced by Ca2+ entry through voltage-gated Ca2+ channels or through those activated by glutamate (NMDA channel) (Nicoll and Alger, 1981; Kudo and Ogura, 1986) or by the release of intracellular Ca2+ resulting from activation of muscarinic receptors (Kudo, 1988), or 5-HT receptors (Menahem 1990). Two distinct Ca2+-activated K+ currents have been identified in hippocampal neurons: First, a rapidly activated current contributes to action potential repolarization and the fast AHP following individual action potentials, (labled lc). In addition, a slowly developing current (IAHP) underlies the slow AHP, which occurs after a burst of action potentials and contributes substantially to the spike-frequency accommodation in these cells during a prolonged depolarizing current pulse. Ic is typically a large, non- inactivating outward current which is blocked by low concentrations of external TEA and charybdotoxin but not by apamin. The current is very voltage-dependent even at

2+ constant Ca concentrations. In contrast to lc, IAHP is a small apamin sensitive current with little voltage-dependence, higher Ca2+-sensitivity and is not blocked by TEA.

1.2 SEQUENTIAL OPEN CHANNEL BLOCKADE MODEL Wang, X.P. 13

The drugs used in the present study exhibit open channel block of BKca in CA-j neurons and some considerations of open channel block models are discussed. Adams (1976) was the first to propose a sequential reaction scheme to fit his voltage-clamp data for amylobarbitone, thiopentone, methohexitone, and methyprylone on end-plate channels. In this model the channel has open, closed and blocked states with the open state the drug target. Neher et al. (1978) examined the effects of QX-222 and QX-314 on the currents through individual end-plate channels in the extrajunctional regions of denervated frog skeletal muscle fibers. The following results were obtained: (I) After initial opening, the channel switched several times between open and blocked before finally closing. (II) The l-V relation was unchanged by the drugs since the channel must open prior to block. (Ill) Rates for channel closing, blocking and unblocking were voltage sensitive. Channel blocking had the reverse voltage dependence of channel closing; channel blocking and unblocking had approximately equal and opposite voltage dependences. (IV) The apparent blocking rate increased linearly with drug concentration, while the unblocking rate was concentration independent. (V) The total time (on the average) that a channel spent in the open state after initial opening is invariant (i.e. the sum of all the short open periods in a burst was equal to the mean open time of channels before adding drug.) These findings suggested that local anaesthetics transiently occlude open channels.

1.3 PATCH CLAMP TECHNIQUE

In the patch clamp, a micropipette is initially brought into contact with a cell- attached configuration. Recording of unitary currents can be done with cell-attached or from excised inside-out or outside-out membrane patches; macroscopic currents with whole-cell configurations. For the patch clamp to work satisfactorily, since unitary currents are in the picoampere range, attention to a number of details Wang, X.P. 14

concerning the preparation is required to ensure an adequate signal-to-noise ratio. The following is a brief list of pertinent considerations. 1. A high-resistance seal is essential to ensure that the leakage currents (and their noise components) are small compared to the desired transmembrane currents.

2. The pipettes should be carefully fire-polished and clean (in general, pipettes may be used only once). 3. Cells should be clean and free of connective tissue, adherent cells, and basement membrane. Good seals are most readily obtained on cultured cells. 4. The application of gentle suction will increase the resistance of the seal to greater than 1010 (ohms). This is the desired giga-ohm seal (gigaseal), permitting the measurement of currents from membrane areas of the order of 1 nxw2 or less. Typical patch current morphology is shown in Fig.1. It reflects the opening and closing of a single channel with a constant conductance. The single channel behavior forms a link between the microscopic membrane property and the macroscopic behavior which is the statistical summation of a large number of single ion channel

=, N events lmacro singie*Po* > where N is the number of active channels; p0 is the channel open state probability (percentage of time spent in the open state). This link permits the single-channel measurements to be used as a tool for elucidating quantitative macroscopic membrane properties.

1.4 SYNTHETIC CLASS III ANTIARRHYTHMIC DRUGS

Drugs which can prolong the action potential duration via specific potassium channel block without interfering with intraventricular conduction velocity are defined as class III antiarrhythmics. The chemical requirements for selective potassium channel block potentially provide the insights of the basic molecular pharmacology of channel activation and gating. At this time, these requirements are not clear. The basic structure of some class III agents is similar to that for local anaesthetic and Wang, X.P. 15

class I antiarrhythmics (i.e., an aromatic moiety joined to a terminal amine by a short alkyl chain containing an electronegative functional group such as an amide, ether or alcohol), other putative class III agents however have very different chemical structures (eg. tedisamil).

The structure-activity relationships for class III antiarrhythmics are presently under intensive investigation. It seems that the substitution of an electron-withdrawing for an electron-donating group of the class I agent tends to minimize the drug action on Na+ channels and increase drug activity on potassium channels. This kind of substitution produces a large electronic change in the molecule, but only slightly alters lipid solubility and molecular weight (Follmer, 1988). It has been reported that if the N-acetyl group of sematilide is replaced with the methylsulfonamide group to get sotalol, the class I activity of sematilide will change to a class III profile. Another

example is that whereas the 4-NH2 benzene sulfonamide (WY-47804) is a class I

agent, the 4-N02 substituent (WY-47,792) is a class III drug (Buzby.et al. 1987). WY- 47,792 related compounds did not show clear dependence of class III activity on molecular size or lipid solubility.

1.5 THERMODYNAMIC ANALYSIS

Determination of whether a specific ligand-receptor interaction at equilibrium is enthalpy- and/or entropy-driven can be achieved by thermodynamic analysis (Weiland et al 1980). Since its early application to enzyme radioligand binding data in the 1970's (Laidler et al, 1979) thermodynamic analysis has provided the potential for an insight into the molecular events underlying drug-receptor interactions not obtainable by other techniques. Embodied in thermodynamics are the laws governing the interconvertibility of heat and work and, hence, it is a particularly apt framework for the analysis of the transduction of information from ligand to biological tissue during the initiation of a drug effect (Raffa, et al, 1989). Wang, X.P. 16

Quantitative measurement of the driving forces involved in the drug-receptor interaction is implicated in the thermodynamic analysis of pharmacological data involved in the drug-receptor interaction in place of less precise terms such as "affinity". In addition, the cautious interpretation of thermodynamic analysis can give clues to the underlying mechanisms of the drug-receptor interaction that is beyond the resolving power of other parameters, such as the dissociation constant (Miklavc, 1990). An enthalpy-driven process is usually associated with the formation of new bonds, e.g. hydrogen bonds and Van der Waals's interactions, in the drug-receptor- membrane array. An entropy-driven process is usually characterized by the displacement of ordered water molecules coupled to the formation of new hydrophobic interactions (Weiland, 1979). The temperature dependency of the dissociation constant is a property of the drug - receptor interaction that appears to be as characteristic and as informative as the dissociation constant itself (Raffa et al, 1989). Two types of information are generally obtained from thermodynamic analysis. Firstly the equilibrium thermodynamic parameters AG°, AH° and AS° can be determined, which yield the overall driving forces for the binding reaction. The standard free energy change (AG°) is calculated from

AG° = -RTInKa or AG° = RTInK (1)

where

1 1 R = The gas constant (1.99 cal Mor K" = 8.31 JMOMK"1)

T = The absolute temperature (K=° C+273)

Ka= The equilibrium association constant

1 K = The equilibrium dissociation constant (K=Ka" ). For the simplest case of a reversible bimolecular drug-receptor interaction, the association constant is the ratio of the concentration of the complex [AR] to the Wang, X.P. 17

product of drug and receptor concentrations [A][R]. i.e. Ka = [AR]/[A][R]. Ka is given by the slope of the Scatchard plot. The relationship between change in free energy (^G°), change in enthalpy (the heat content of a substance per unit mass), and change in entropy (a measure of unusable energy, or disorder in a system) is given by

AG°= AH°-TAS° (2)

where AG° = Standard free energy change of the reaction, KCal Mol'1 or KJ Mol"1 (1 KCal = 4.184 KJ). AH° = Enthalpy change of the reaction (KCal Mol"1 or KJ Mol"1) AS° = Entropy change of the reaction (KCal Mol"1K"1 or KJ Mol"1K"1) Combination of equations (1) and (2) yields the van't Hoff equation,

lnKa = - AH°/RT + AS°/R (3)

1 AH° is calculated from the van't Hoff plot (InKa vs T" ). The slope is - AH°/R, - DS ° is then obtained from equation (2). Equilibrium is attained in a drug-receptor reaction at the point at which the free energy of the system is a minimum for the given set of initial conditions. The change in free energy (AG°) accompanying a state change in a system represents the maximum capacity for performing useful work. Most drug-receptor interactions are allowed to come to equilibrium at which point AG=0, the system is no longer capable of doing work. At conditions other than the standard state (298 K,1 atm, pH = 0.0 [H+] = 1.0M) for a reversible, bimolecular drug-receptor interaction.

AG = AG° - RTInKa (5) The free energy of a molecule is inherent in its structure. The AG° is the difference in state values of the sum of free energies of the products and the sum of free energies of the reactants (each at standard state). Wang, X.P. 18

AG°, the standard free energy change, is a constant for any given chemical reaction at a given temperature. AG, the actual free energy change, varies with the concentrations of the reactants and products. AG is a indicator of spontaneous

reaction, but it is not an indicator of the reaction rate. If AG<0 the drug-receptor reaction will occur spontaneously in the direction indicated from given concentrations

of reactions and products. If AG>0 then a spontaneous reaction in the given system is not possible. The second type of information obtained from thermodynamic studies involves a determination of the activation thermodynamic parameters, AGt, AS* and AHt, as adapted from transition state theory (Cornish-Bowden, 1976). Arrhenius plots Ink! vs T "1 or Ink.-, vs T "1 can be used to obtain the energy of activation Ea. Where kj and k. 1 represent on- and off-rate constants for drug respectively. The enthalpy A Ht can then be determined using. AHt = Ea-RT (6) The free energy of activation AGt were then found from: AGt = RTInA(rate constant) - RTIn (kT/h) where k is the Boltzmann's constant and h is Planck's constant, from AG t andAH t . AG+ can be easily found according to AGt = AHt -TASt (Hitzemann, 1988)

A large enthalpy of activation indicates that a large amount of stretching, squeezing or even breaking of chemical bonds is necessary for the formation of the transition state (Cornish-Bowden 1976). The entropy of activation ASt gives a measure of the inherent probability of the transition state (Laidler 1979). A positive A St could result from the breaking of bonds within the receptor, causing receptor unfolding and greater freedom of movement. (Hitzemann, 1988). Wang, X.P. 19

2. METHODS 2.1 CELL CULTURE PREPARATION

The procedures used to maintain the CAi hippocampal neurons from fetal rat were a modification of the method described by Banker and Cowan in 1977 as used in Dr. Baimbridge's laboratory. Briefly 18 day embryonic-age rat fetuses were utilized. Following dissection the hippocampi were subject to enzymatic (trypsinization) and mechanical treatment to dissociate the cell. The cell count was adjusted with Dulbecco's Modified Eagle's Medium (DMEM) and laminin-coated coverslips, treated with poly-D-lysine to decrease the rate of proliferation of non-neuronal cells, were plated with a low cell density of approximated 105 cells/cm2. The coverslips were

then incubated, with the growth side downwards, in DMEM and 5% C02 at 37° C. After 2-5 days the cells were treated with 5-fluoro-deoxy-uridine(5-FDU) to prevent glial cells from multiplication. The cells were maintained in culture solution with the growth medium half-changed every third day. Little reaggregation occurs and the number of neuronal cells present were over 98% as determined by immunocytochemistry. The survival cells were those which had completed DNA synthesis about 48 hours before dissociation, and which were in the process of migration to the cortical plate. The morphology of most neurons in their long-term culture bears a striking resemblance to the growing pyramidal neurons seen in vivo in the field CA1 of the developing hippocampus. They have a characteristically slender apical dendrite which does not bifurcate, but gives off a few short side branches, especially near its end. The electrophysiological studies were conducted utilizing cells from days 5-10 after isolation.

The culturing of cells with the growing surface facing the bottom of the dish was found to be essential for the long-term survival of the neurons. It would appear that the normal content of oxygen in air is toxic to neurons since the cultured neurons can Wang, X. P. 20

survive "face-up" but only if they are incubated in an atmosphere of less than 10% oxygen.

2.2 ELECTROPHYSIOLOGY

2.2.1 Glass and Pipettes: Patch-clamp electrodes were made from Corning #7052 microtubes composed of borosilicate glass, this type of glass has less energy loss when voltage is supplied and hence less noise. Pipettes were fabricated with a vertical pipette puller (PP-83, Narishige) using the standard two-pull method (Hamill et al., 1981). A fixed pulling length and fixed settings for the two stages were maintained to obtain pipettes of similar properties. The resultant pipettes have steep tapers at the tip and outside diameters of about 1 ~2 jum. Polishing of the glass wall at the pipette tip was done on a homemade microphorge connected to a D.C. voltage supply and consisting of a U- shaped platinum filament. Under a light microscope (10 x 20 magnification),the electrode tip was brought, for a few seconds, to within 5 ^m of the heated platinum wire coated with soft glass.

Fire-polished pipettes were then filled with pipette solution and bubbles were eliminated by gentle taps to the pipette. The pipette was always filled only far enough from the tip so that the end of the internal reference wire was immersed. Any solution that reached near the back of the pipette was drawn to keep it from getting into the holder. This is important since the holder may become a significant noise source if fluid gets into it. The resistances of the electrodes were generally between 4-8 M n. In most cases, when the holder with a filled pipette had been inserted into the headstage connector, and the pipette tip was positioned just above the bathing solution, the RMS current noise was about 0.1-0.2 pA, which meant a low noise setup had been achieved. Wang, X. P. 21

2.2.2 Mechanical Setup: The patch pipette was mounted on a pipette holder which was connected by a silicone rubber tubing through which pressure was applied. Then the pipette holder was connected to an amplifier headstage which was mounted on a Newport motor drive micromanipulator. This, in turn, was mounted onto another manipulator for coarse movements (Newport). The pipette holder was always kept clean and dry by a jet of oxygen to maintain a low noise level. 2.2.3 Single Channel Patch-Clamp Configurations: The neurons were viewed at x 300 magnification using phase contrast microscopy (Nikon). Patch electrodes were manually positioned against the cell surface with a micromanipulator (Newport). The high resistance seals of different configurations were formed according to the requirement of the experiment. The cell attached mode was the initial mode for all recording configurations. In this mode, the micropipette was brought close to the cell membrane with a small positive pressure applied to the pipette holder to keep debris in the solution away from the pipette. After gentle contact between the pipette and the cell a small negative pressure was applied: the procedures usually resulted in the formation of a giga-ohm seal (resistances in excess of 5 G n). This cell-attached mode is also called on-cell patch clamp configuration. Following this, if a microelectrode was introduced into the cell by transient voltage pulses or by additional suction to rupture the cell membrane, a whole cell recording mode would be established. If after the establishment of the cell-attached configuration, the pipette was quickly withdrawn, then a patch of membrane was isolated in the inside-out mode. In this case the patch was formed with the inside surface adjacent to the bath solution. If instead the pipette was pulled away after the whole cell configuration was established then the membrane would reform in an outside-out arrangement. That is, in this case, Wang, X.P. 22 the extracellular side of the membrane was facing the bathing solution. Further details of the patch clamp configurations can be found in Hamill (1981). 2.2.4 Solutions used: In these experiments, single-channel recordings were taken from excised patches with inside-out and outside-out patch-clamp modes. The solutions used for the different configurations are shown in Table 1. With the inside-out configuration the pipette contained solution B. The bath solution used prior to and during patch excision was solution A as shown in the same table. After patch excision the bath solution was then exchanged to solution C or C With the outside-out configuration the pipette solution contained solution D and the bath solution, prior to and during patch excision, was solution E. Free calcium concentrations generally were 200 ^M in the bath and 1 mM in the pipette solutions for inside-out recordings. In some experiments, internal Ca2+ was decreased to 3 - 1000 nM using Ca/EGTA buffers. HEPES buffer was used instead of phosphate buffer to avoid crystals precipitating at the tip of the pipette.

Putative Class III antiarrhythmic agents of different concentrations were added to the high K+ bath solution (Solution C) for perfusion of the excised patches. These drugs were tedisamil (KC-8857, Kali-Chemie; Hannover, Germany), UK-68798 (Pfizer; Kent, England), risotilide (WY-48986, Wyeth-Ayerst; Princeton, U.S.A.) and RP-62719 (Rhone-Poulenc; Vitry, France).The chemical structures of these compounds are shown in Fig.1. The following compounds were also used in some experiments; tetraethylammonium (TEA at 0.1 - 5 mM), 4-aminopyridine (4-AP at 1 - 10 mM) and ethylenglycol-bis(2 aminoethyl)-tetraacetic acid (EGTA at 1 - 2 mM). All pipette solutions were filtered through 0.22 nm filter and all experimental solutions were made using deionized water. Wang,X.P 23

TABLE 1

Chemical Components of Solutions Used for Cell-Attached and Excised Patch- Clamp Single Channel Studies:

PRE-EXCISION INSIDE-OUT OUTSIDE-OUT

in mM BATH PIPETTE BATH PIPETTE BATH

A B c C D E

NaCI 140.0 140.0 5.0 70.0 5.0 140.0

KCI 5.0 5.0 140.0 70.0 140.0 5.0

CaCl2 0.2 1.0 0.2 0.2 0.2 1.0

MgCl2 1.0 1.0 1.0 1.0 1.0 1.0

TTX — 0.001 — — — 0.001

HEPES 10.0 10.0 10.0 10.0 10.0 10.0

PH 7.3 7.3 7.3 7.3 7.3 7.3 Wang,X.P 24 Figure 1 Chemical Structures of a: UK-68798; b: risotilide c: tedisamil.

CH3S02NH-/~~\

^-NHS02CH3

CH3S02NH-^\ CH|CH3I2

^S/^NH/CH,CH3'2

O

c Wang, X. P. 25

2.2.5 Recordings: The currents passing through single channels were recorded using standard patch clamp techniques (Hamill et al. 1981). The headstage amplifier (current to voltage converter with 50 G a feedback resistance) was connected to the patch clamp amplifier (Axopatch, Axon Instruments) which was used in the voltage-clamp mode to amplify current and to control voltage across the patch of membrane. The standard convention was used, in which the bath served as reference ground i.e., the external side of the channel was defined as zero potential. Thus a negative potential applied to the pipette would result in a depolarization of the patch. A downward deflection from the baseline was produced when positive ions cross the membrane and enter the pipette. This was an outward cationic current recorded from the cell- attached or inside-out mode. The same current would be recorded as an upward deflection in outside-out configurations since the membrane was reversed relative to the situation of inside-out and cell-attached modes. The bath was connected through a 0.15 M KCI agar bridge to the ground. The output of the amplifier was low pass filtered (4-pole with Bessel characteristics which induces less than 1% overshoot) with a -3 dB corner frequency of 2 or 5 kHz. Data were acquired using an IBM compatible computer programmed with pCLAMP software (version 5.0 or 5.5.1, Axon Instruments Inc. Burlingame, CA). Membrane currents were digitized at sampling rates of 5 kHz (low-pass filter set at 2 kHz) or 20 kHz (low-pass filter set at 5 kHz). The latter setting was used for recording rapid channel flickering events between open and non-conducting states. Holding potential usually was set at 0 mV and step depolarizations or hyperpolarizations were applied to study the voltage dependence of the channel. 2.2.6 Analysis Procedures: The single channel data were used to characterize the gating mechanism of Kca channel, which primarily means to define 1) the number of states or Wang, X. P. 26 conformations the channel can enter; 2) the time spent in each state as a function of voltage or ligand concentration; 3) the rates at which transitions between the various states occur and the underlying reaction mechanism and; 4) also the single channel current amplitude and transmembrane voltage relationship (conductance). In order to get the desired information, digitized single channel current records were analyzed to measure the amplitudes and durations of the single channel events, as well as the length of time between events. The distributions of the current amplitudes and the open and closed durations were determined by analysis of 5-200 second consecutive recordings. It is important to determine the baseline associated with the current. If there was only one conductance channel in the patch, the baseline was adjusted so that current averaged 0 pA when channels were closed. Sometimes a smaller conductance channel was activated with the large conductance channel in the patch, which made the baseline relatively noisy. In this case the baseline was kept so that current averaged 0 pA when both channels were closed. It should be pointed out that in the presence of the small conductance channel the amplitudes of the large conductance channels tended to be underestimated due to the slight baseline drifting despite the effort to keep it stable. Openings were detected using a half-amplitude threshold, and current amplitude data were derived from open channel currents, fitted with Gaussian distributions. The Gaussian fitting terms used to fit amplitude data were of the form: N= EWn exp [-(A-n)2/(2(j2)]/[(2n)^2a] dt. where N = the number of counts in the given bin contributed by the term, i = the total detected levels. Wn = the weight of the n-th term. A = the center value of bin width Wang, X.P. 27

n = the mean or the peak value a = the standard deviation •K = the trigonometric constant (?r =3.1415)

Measurement of channel kinetics requires determination of mean open and closed times. It seems the switchings of a channel from a closed to open, or vice- versa, is always signalled by a discrete current jump. The measurement of the time the channel spent in the closed or open state is unambiguous and easy to do: one simply measures the duration between two successive current jumps. But in reality, the analysis is less simple. When a channel opens, a small current of the order of a picoampere flows through it. Because of the electrical background noise the single channel current can be resolved only if the signal is subjected to low-pass filtering that excludes the high frequency component of the noise. The side effect of low-pass filtering is the distortion of the original square-wave signal typical of single channel current. The distortion, due to the attenuation by the filter, delays the response following the input current step due to a single channel transition. In the case of a brief current pulse, the output may fail to reach full amplitude. The amplitude reached is a function both of pulse length and filter setting. Because of the large number of single channel events needed for statistical accuracy, the pCLAMP software provides a way to process data automatically. In this case a threshold level is defined as an estimate of one half of the single channel amplitude and, every crossing of the threshold is interpreted as an opening or a closing of the channel. The time between two consecutive threshold crossings represents state lifetimes. This method can clearly miss brief events which, after low-pass filtering, do not reach threshold. The major consequence of missing events is an overestimate of state lifetimes. In some cases, appreciable errors in estimating underlying kinetic mechanisms can be Wang, X.P. 28

introduced. In order to prevent this type of error in our data processing, the analysis was done manually to ensure missing brief events were less than 1% the total. Once the open times of an appropriate number of events have been measured, the rate constant of the open state can be derived from the mean open time as k_i = 1/(mean open time). The mean open time is not calculated by algebraically summing of all the open lifetime values and divided by the number of events considered. Instead histograms of event counts versus lifetime are used to calculate the mean open time, since this representation of data also provides information on the reaction mechanism. In this type of plot, the vertical scale gives the number of events whose lifetimes fall in the corresponding time interval (bin) into which the abscissa has been divided. Open or closed times were counted into the same bin widths respectively and the first low event bin was eliminated from the fit so that missed openings due to the limited bandwidth of the recordings did not distort the analysis. In most cases a minimum of 20-200 sec consecutive current records, containing only a single level of openings, were used to define the distributions. The likelihood exponential functions used to fit the data were of the form N = Wn/r exp (-T/r)dt (mode = 1-4) where

T was the center value of a given bin of width dt;

T was the time constant; the number of modes meant the number of exponential terms needed to be summed in order to fit the curve and, the other terms were the same as above (Sigworth, 1987). A non-linear least-squared curve-fitting routine was used, assuming a Markov process to fit the open or closed-time histograms. The data were weighted by the x2 value (to measure the badness of fit). The fit proceeded until convergence was reached. Convergence was defined when successive improvements in parameter values were less than 2.5 x 10"7, or if all of the Wang, X.P. 29 parameters converged when parameter vectors went to zero (Schreiner, 1985). The fitting statistics and the values of the fitting parameters were given by the computer.

The most important ones were the sum of squares (SS) given by SS=E(XO-X)2/N, where XQ was the observed value and X was the expected value; x2 was given by

2 2 1 2 2 * =SS/sX and R was given by (SS/SXO ) / X100. x and R were kept as low as possible, the goodness of fit was normally around 2 and * 2/degree of freedom was around 1. The number of exponentials which best fitted the data were determined by comparison of the statistical results of the fitting. Since most of the data were recorded at 200 fis intervals (sampling filtered at 5 kHz) and processed at a resolution of 500 A*S (low-pass filtered at 2 kHz), brief exponential components with time constants of 200-300 ^s as observed in many of the closed-time distributions might be artificially enhanced and should be identified as "apparent" until greater experimental resolution confirms their accuracy. This problem is an unavoidable result of limited resolution and occurs in all single channel studies. (Roux B, 1985; Blatz, 1986; Robert A Maue et al., 1989;).

The probability of an individual channel being in its open state (p0) was also determined. If only one active channel was present in a membrane patch, the open- state probability (P0) was simply obtained as the time spent in the open-current level divided by the total time of the record. If two active channels were present in a patch, an all-points histogram constructed from 20-200 sec consecutive current recordings was then used to calculate p0. The integral areas of the amplitude histogram peaks were calculated. When the patch contained two active channels, amplitude histograms had a maximum of three peaks, corresponding to: (1) both channels closed (peak centered at 0; relative area Sc); (2) either channel open (relative area 3 tn Si,o)»' ( )- b° channels open (S2l0). Assuming the two channels to be independent 2 in their gating, the following relations hold: Sc = (1 - p0) , S1>0 = 2p0(1 - p0), S2,0 = Wang, X.P. 30

2 p0 where p0 = (A1i0 + 2A2,0)/2. If a membrane patch contained more than two

channels, this type of analysis were not carried out in terms of p0 calculations. At high Ca2+ concentrations (200/iM), long silent periods occurs. These long closure are thought to be due to a slow blockade of the channel by divalent ions

(Vergara and Lattorre 1983, Miller et al. 1987). Therefore P0 was calculated excluding channel closures lasting longer than 200 msec (see result for details) to ensure that measurements were not related to the slow blocking of the channels. For studies of drugs blocking Kca channel, the Ca2+ concentration bathing the intracellular face of patch membrane was 200 nM. At this concentration the Kca channel generally had a

p0 in excess of 80% and the mean open times were in the range of 10-20 msec which were useful targets for drug actions.

The values given for single channel conductance were slope conductances derived from linear regression analysis of the average single channel current amplitude recorded at different membrane potentials.

2.3 TEMPERATURE STUDIES:

The bath temperature of the recording chamber was varied from 24° C to 14°C using a proportional temperature control system. This system, which utilized thermistors and Peltier devices, was similar in design to that described in Chabaly etal. (1985). Data were initially recorded at T = 24° C and the set-point was then changed to T = 14° C. Data were then recorded at intervals of T = 2°C as the bath temperature was cooled to the low temperature set-point. Since the time required to change the bath temperature by 10° C was 4 minutes, then the time period for recording at an intermediate temperature (example 20 ± 0.2° C) was about 10 s. In most experiments, after data were obtained at the steady state temperature of T = 14° C, additional single channel currents were recorded at the intermediate temperatures by alternating the temperature control system between T = 14° C and a Wang, X.P. 31

given intermediate temperature. No significant differences in the amplitudes or mean open times associated with the unitary currents were observed at a given temperature between the data recorded during the initial cooling run and that obtained in a step to that temperature from T = 14° C (warming up run). Furthermore, no systematic changes in the properties of the unitary currents were found with prolonged exposures of the patches at either of the steady state temperatures of 24° C and 14° C. In a few experiments data were also recorded at T = 19° C in order to illustrate the changes in current amplitudes and channel open times at a temperature intermediate between the two extremes of set-point temperatures. A digital thermometer was used for calibration purposes to ensure accuracy in the readings from the temperature controller. As a measure for the temperature

dependence of channel kinetics, Q10 values have been determined from the ratios of mean open times or rate constants at 24° C to those determined at 14° C. All errors are stated as ± standard deviation. 2.3.1 Equations for Thermodynamic Studies: A sequential open channel block model (see results) was used to determine the

rate constants for channel block of KCa by drugs. An analysis of the thermodynamic parameters for the transition state associated with the onward (blocking) and the off (unblocking) rate constants followed the methods outlined in Minneman et al. (1980) and Weiland and Molinoff (1981). A plot of In (rate constant) vs. T"1 (Arrhenius plot)

was first used to determine the transition state activation energy Ea. The AHt and free

energy of activation A Gt were then found from AHt = Ea-RT and AHt = RTIn(rate constant) - RTIn(kT/h), respectively, k is Boltzmann's constant and h is Planck's

constant. The entropy of activation ASt was then found from AG* = AHt . TAS*. Values for the associated state thermodynamic parameters were also obtained by subtraction of the values of the parameters for the channel unblocking transition state from the magnitudes of the parameters for the channel blocking transition state Wang, X.P. 32

(Minneman et al., 1980). The parameters for the associated state are equilibrium values if the channel block and channel unblock processes are governed by the same rate-limiting reaction. Wang, X. P. 33

3. RESULTS

3.1 IDENTIFICATION AND PROPERTIES OF SINGLE BKCa CHANNELS IN CULTURED RAT CA1 HIPPOCAMPAL NEURONS 3.1.1 .Identification: Selectivity and Channel Conductance:

Typical records of unitary currents passing through the BKca channel are shown in Fig. 1. In these experiments inside-out patches, with 140 mM K+ in the bath solution and 5 mM K+ in the pipette, were isolated. The amplitude of single-channel current is plotted vs patch potential in Fig. 3, where data has been accumulated from 8 patches (including the patch activity illustrated in FIG. 2). The extrapolated zero- current potential was —76 mV in asymmetrical 5 mM K+ outside and 140 mM tf solution facing the inside of the membrane (Fig. 3 solid line). The observed single channel currents could arise from an efflux of K+ from the intracellular side of the membrane and/or an influx of CI". Na+ could not be a major charge carrier as the flow of Na+ down its electrochemical gradient would produce single channel currents with a polarity opposite to that observed. There were no fluctuations observed above base line at any voltages and this eliminated the possibility of Na+ channel openings. To determine the channel ion selectivity, the concentrations of ions on the intracellular sides (bath) of excised membrane patches were changed from the high K+ bath solution (140 mM) to one containing 70 mM KCI and 70 mM NaCI, the total CI" concentration remained unchanged (Fig. 3 dotted line). This caused a 19 mV shift in the zero-current potential in a depolarizing direction which was consistent with the predictions of the Nernst relation with channel selectivity for K+ (17 mV). If the current were carried by CI", the reversal potential should not have changed since the CI" concentration gradient was not changed by this manipulation. Thus in the conditions of these experiments the primary charge carrier through this channel was K+. WANG, X.P. 34

Figure 2

o

5 pA

a: 100 ms b: 10 ms Wang, X.P. 35

Figure 2. Typical Records of Unitary Currents Passing Through Kca in the Somatic Membrane of the Cultured Rat Hippocampal Neurons.

A: Currents (downward deflections from the baseline) recorded from an inside-out patch with 140 mM K+ in the bath solution and 5 mM K+ in the patch solution. The record length was 1.2 sec. B: The same data were expanded 10 times to show a detailed channel kinetic behavior. The closed and open states of the channel were designed C and O respectively. WANG, X.P. 36

Figure 3 Wang, X. P. 37

Figure 3. Current-Voltage Plots: KCa Channel Conductance and Ion Selectivity

Single-channel fluctuations were examined in excised inside-out patch with 5 mM K+ outside and 140 mM K+ inside (closed circles, n = 20) or 70 mM K+ inside (closed triangles, n = 6). The data were fitted by linear regressions and the error bars were S.E.M. values. The slope conductance obtained from the solid line was 80± 21 pS and the extrapolated zero-current (reversal) potential was -76 mV; whereas the slope conductance for the dotted line was 60 ± 5 pS and the extrapolated zero- current potential was -57 mV. Wang, X.P. 38

The l-V relation was linear over the voltage ranges examined (t 40 mV). It was

not possible to measure current at Vm less than - 40 mV because the currents were too small to be distinguished from the baseline noise. If the discrepancy between the calculated and measured shifts in the reversal potential with changes in transmembrane ion gradients was indicative of an imperfect selectivity of the channel for K+ over Na+, then the estimated permeability ratio calculated using the Goldman-Hodgkin-Katz relationship (Goldman, 1943; Hodgkin & Katz, 1949) was about 100:7. The channel conductance under different transmembrane ionic conditions were different. With 140 mM internal (5 mM external) K+, the channel conductance differed from patch to patch within a normal distribution and with a central tendency of -80 ± 21 pS (n = 20). While with 70 mM internal (5 mM external) K+, the averaged single channel conductance was 60+ 5 pS (n = 5). Occasionally, activation of a second channel was seen in some patches. It is likely that this activity was from SKca channels in the neurons; recent work has documented some of the properties of such channels in CA1 neurons (Lancaster et al. 1991). Because of the low probability of these channel activities, these channels were not studied further. 3.1.2 Calcium Dependence: Calcium at the inner surface site of the membrane was required for the activation of this channel. In the experiments discussed above the internal calcium concentration [Ca2+],- was 200 »M. Upon perfusion of inside-out patches with a bath (control) solution containing 140 mM K+ and 200 A*M Ca2+ (Table 1, solution C), activation of a large conductance channel was apparent (Fig. 4a); In order to determine Ca2+ dependency of channel activity, the [Ca2+]j was lowed to nano­ molar level by addition of 1.99 mM EGTA to the 140 mM K+ solution (free Ca2+ concentration was 3 nM, calculated by Maxchelater program, assuming that the dissociation constant for Ca2+/EGTA is 10*7 M, Caldwell 1970, Fibiosco 1979). This Wang, X. P. 39 procedure resulted in abrupt cessation of channel openings (Fig. 4b). Re-perfusion of control solution (C) re-established channel activity (Fig. 4c). Thus Ca2+ activation of ® these channels was reversible by reducing free Ca2+ concentration at the inner surface of the membrane to less than 10 nM. 2+ The sensitivity of KCa to other concentrations of Ca were also studied. When a solution containing 10/iM was applied to an inside-out patch, double level openings were recorded indicating that two active channels were present in the patch (Fig. 4d, arrow 2); when a low concentration of calcium solution was introduced to the bath, the channel activity ceased (Fig. 3d, arrow 3, The "break" in 2 indicates a time gap of 1 sec when no channel openings were recorded); re-perfusion with higher calcium solution resulted in the re-appearance of channel activity (Fig. 4d, arrow 4). In other experiments (n = 3) the [Ca2+]j was near 1 ^M and the pattern of channel activity was clearly changed compared with that at 10/iM [Ca2+]j. At high [Ca2+]j, the channel exhibited typical long bursting openings (Fig. 5a) whereas with 1 2+ nM [Ca ]j KCa showed very short openings interspersed with long closures (Fig. 5b). A detailed comparison of channel kinetic analysis at different [Ca2+]j is shown in Fig. 5c-e. With [Ca2+],- ~ 1 ^M, the channel open state probability DQ was decreased 2+ by about 90% compared to that measured at [Ca ]j = 10pM. The ratio between r0 with [Ca2+]j = 10/iM and that with [Ca2+]j = 1 pM was ~9. At 1 /iM [Ca2+ J the fast time constants (rcf) of closed time distributions within the burst openings had a value similar to that at 10/iM [Ca2+]j and the slow time constants of closed time distribution 2+ (rcs) were increased. Thus, the decrease in p0 with 1 /*M [Ca ]j was due to lengthening of long closed times and also shortening of mean open times. 3.1.3 Voltage Dependence: The voltage dependency of channel mean open time was studied by changing the patch command potential (Vp) from 0 to ± 60 mV. Wang, X.P 40

Figure 4 a.

5 pA 50 ms

5 pA 100 ms Wang, X. P. 41

Figure 4. Calcium Dependency of BKca:

a-c: In this experiment, free Ca2+ concentrations were obtained by addition of EGTA to the bath solution containing 200^M total Ca2+. a: Single channel openings with [Ca2+]j = 200^M. b: Addition of 1.99 mM EGTA to the bath solution with 200^M Ca2+ (free Ca2+ was buffered to less than 3 nM), the channel activity abruptly disappeared. c: Channel openings after return to the 200 ^M Ca2+ solution.

d: In this experiment, EGTA was kept constant whereas free Ca2+ concentrations were adjusted by changing total [Ca2+]. Arrow 1 indicates inside-out patch bathed in 3 nM free Ca2+ solution. No channel activity was observed. From arrow 2 onwards solution containing lO^M Ca2+ was introduced to the same patch, after -100 msec delay (which represented perfusion time), two channel openings were obtained. Return to 3 nM Ca2+ solution (arrow 3) resulted in the disappearance of channel openings. Re-application of 10/iM Ca2+ solution (arrow 4) lead to the re­ opening of BKCa. Wang, X.P. 42

in

<0 Figure 5c

2S00 2+ [Ca ], - 10 MM 5000 [Ca2+] i = 10 pM r0 = 14.4 ± 0.1 msec rcf = 0.4 ± 0.2 msec n = 8633 r0, = 1.62 ± 0.4 msec

N

0 J T 20 40 60 80 1 2 3 Open Time (ms) Closed Time (ms)

1500 [Ca2*], = 1 IJM 800 [Ca2+]i = 1 MM T0 = 0.94 ± 0.2 msec rct = 0.3 ± 0.2 msec

n = 2664 rc, = 26.8 ± 0.2 msec

0) Z3 (Q I X LRAOJIJ 0J ^ —i o 60 80 100 Open Time (ms) Closed Time (ms) CO Figure 5d

80n * p < 0.05

2 • fCa ^ = 10 MM 40 A 2+ i [Ca ] | = 1 MM 15 A ms

5H m oJ—L_£2 'A T r c, TC< Wang, X.P. 45

Figure 5. Calcium Dependency of BKCa Channel Kinetics 2+ 5a: single-channel behaviour of BKca with 10/xM internal [Ca ]; openings are downward. 5b: the same patch with 1 ^M internal [Ca2+]. Low pass filter was set at 2 kHz and the sampling rate was 5 kHz, 200 msec for sweep. 5c: quantitative analysis of single channel kinetic for the same cell. With [Ca2+li = 10^Mand[Ca2+]j = 1^M. 5d: comparison of channel kinetics at 10 ^M [Ca2+]j with that at 10/*M [Ca2+],-. Wang, X. P. 46

The voltage dependency of channel mean open time was studied by changing the patch command potential (Vp) from 0 to ± 60 mV. As shown in Fig. 6a, the mean open times were not strongly influenced during these potential changes with 200 ^M [Ca2+]j in the bath solution. At a patch command potential (Vp) of 0 mV, the channel mean open time was 14.7 ± 1.0 msec (n = 11); and at - 20 mV the mean open time was 15.3± 1.1 msec (n = 11). In three of these patches, the mean open times were measured over a wider range of patch potentials; the values of open time at - 60 mV were not significantly different from those determined at V = 0 mV. Therefore the mean open times were not strongly dependent on potential with [Ca2+],- = 200 ^M. However, the rate of channel opening increased with patch depolarization. Since

open state probability p0 is the product of mean open time and open rate then p0

increased with depolarization (Fig. 6b). The relationship between p0 and the membrane potential can be fit reasonably well by Boltzman's equation:

1 Po= {1 + exp[-K(V-V1/2)]}- where V-j /2 is tne membrane potential at which the open state probability is 0.5. V is the membrane potential and K is the logarithmic potential sensitivity which is given by

the slope of ln(p0/(1-p0) versus membrane potential. At 1 ^M [Ca]j the open state probability changed e-fold with a membrane potential change between 10-15 mV. Increasing the Ca2+ concentration to 200 ^M altered the voltage sensitivity. As

shown in the Fig.6, with high internal calcium, p0 is virtually not dependent on voltage.

3.1.4 Channel Kinetics:

A kinetic analysis of the gating transitions is shown in Fig. 7. At Vp = 0 mV and [Ca2+]j = 200 /iM, the channel showed bursting behavior consisting of groups of long openings with some brief openings. Within the burst, the channel occasionally had rapid (few milliseconds) transitions into and out of the closed state (flickering), Wang, X.P 47

Figure 6

A 25 r

ms

0 L l_ 0 20 40 60 B V (mV)

Po= 0.99

e I

P0= 0.01 -s L

-40 -20 0 20 40 60 V (mV) Wang, X.P. 48

Figure 6. Voltage Dependence of BKca:

A: Channel mean open time (r0)-voltage plot. The data were fitted with a horizontal linear line, showing r 0 was not sensitive to voltage, (n = 7)

B: Effect of voltage on the open state probability (p0) of BKca- With 1 f*M 2+ [Ca ]j, at applied voltages between 0 and 60 mV, the p0 changed e-fold per 15 mV (data points solid triangles and fitted by broken line) whereas with 200/iM [Ca2+]j; the p0 was virtually independent of voltage (data points were designed as solid circles and fitted by continuous line).

The p0 was calculated using recording periods from 20 to 60 seconds of recording time at each voltage. Wang, X. P. 49

mainly spending most of its time in the open state. The average single channel open state probability (p0) was near -80%. The distribution of the open duration of events of this channel was well fitted by a single exponential (Fig. 7a) although on a few occasions there appeared to be an excess of brief openings. Because brief openings were concentrated in the first few bins of the histogram and were a relatively small percentage of the total, the uncertainty in fitting this component with a second exponential was large. Furthermore, the fit of the entire distribution was not substantially improved using the sum of two exponential fits. The distribution of closed intervals recorded during periods of high activity were described by the sum of two exponential probability density functions (Fig. 7b). The time constants of these components differed by about 10 fold (0.6 msec for the fast component and 4.2 msec for the slower component; n = 12) and they appeared to be independent of voltage. It should be mentioned that these closed time constants mainly represent nonconducting states within the burst (fast component) and short closures between bursts (slow component). Occasional, very long channel closures (>2 sec) were also observed in channel records. It should be noted that considerable variation was seen in channel kinetics from patch to patch. Thus from kinetic analysis the large conductance calcium- activated potassium channel appeared to have at least one open state and two closed states.

3.1.5 Pharmacology of the Kga channel: The channel activity was strongly affected with TEA included in solutions bathing outside-out patches (n = 4). Fig. 8 shows an example of the effects of TEA applied to the extracellular face of the membrane. TEA (at 100A*M) reduced unitary current amplitude. These effects could be explained by TEA blocking the open channel with a time course that was faster than the frequency response of the recording system. TEA (at 1 mM) virtually abolished the channel Wang, X.P. 50 Figure 7

120 r0 = 14.7 ± 0.2 ms Control

N

20 *0 60 80 Open Time (ms)

300 TCf = 0.58 ± 0.01 ms

TC, = 4.2 ± 0.2 ms Control

N

Closed Time (ms)

M = 5.4 ± 0.1 pA

N

i (PA) Wang, X. P. 51

Figure 7: Kinetic Analysis of BKCa

A: Open time histogram constructed from 1470 consecutive samples of openings. The data were well fitted by a single exponential decay function with a time 2+ constant of 14.7± 0.2 msec. [Ca ],- = 200MM. B: Closed time histogram for the same patch. The sum of two exponential functions was needed to fit the data. C: Amplitude histograms were generated at V = 0 mV, channel openings longer than 0.2 msec were placed into bins spaced 0.1 pA apart. The data were fitted with a Gaussian distribution with mean/i = 5.4 ± 0.1 pA. Current signals were filtered at 2 kHz and digitized at 5 kHz. Durations of open and closed times were placed into 2 msec and 0.4 msec bin widths respectively. Wang, X.P. 52 activity. The action of TEA was fast and reversible and channel activity was restored with TEA washout. TEA, at concentrations up to 5 mM, did not block this channel when applied to the inside of the membrane. 4-AP (up to 10 mM) did not show any effect on this channel with application to either inside or outside of the membrane. The actions of TEA from outside (and not from inside) and the lack of effect of 4-AP are consistent with the known pharmacology of large conductance Kca channels.

3.2 ACTIONS OF ANTIARRHYTHMIC AGENTS ON KCa CHANNELS: SINGLE CHANNEL ANALYSIS: Considerable efforts are presently being directed towards the development of drugs that possess actions on specific ion channels in cardiac tissue. For example, drugs that would act selectively on sodium, calcium, or potassium channels in cardiac cells could have potential therapeutic use as antianginal or antiarrhythmic agents. In addition to establishing specificity and mechanisms of actions for the drugs on ion channels in cardiac cell membrane, it is also important to determine whether or not such agents act, at low concentrations, on other tissue. A previous study had shown that a putative Class III antiarrhythmic agent KC-8851 blocked KQ3 in hippocampal neurons (McLarnon, 1990) at concentrations similar to those at which a chemically related agent, KC-8857 (tedisamil), increased the inactivation rate of the transient outward potassium current in rat ventricular myocytes. (Dukes et al, 1989). Of interest was the observation that tedisamil, at a dose twice that producing maximum antiarrhythmic actions in rat, also caused adverse effects on respiration that could be a consequence of drug activity in the central nervous system (Beatch, Walker et al. 1991). In this work, the actions of four newly synthesized cardiac drugs, RP-62719, UK-68798, KC-8857 (tedisamil) and WY-48986 (risotilide), Wang, X.P. 53

CL E LO O o. CM1

O

< LU h-

00 _J O

1 BMM o LL o Wang, X.P. 54

Figure 8: Effect of TEA on BKCa Records were obtained from an isolated outside-out pipette with 140 mM KCI in the patch and 5 mM KCI in the bath solutions. Patch potential was held at 0 mV. Current fluctuations (upward from the baseline designed C) of two BKca channels in control solution (control arrow); after 0.1 mM TEA was added to the external side of the membrane (TEA 0.1 mM arrow) the channel unitary current apparently decreased in amplitude due to fast block; addition of more TEA (TEA 1mM, arrow) resulted in further reduction of the single channel current amplitude; re-application of control solution to remove TEA reversed this effect. Abbreviations as in text. Wang, X.P. 55

have been studied on Kca in CA1 hippocampal neurons. All agents have potential Class III antiarrhythmic utility. 3.2.1 Inside-out Experiments:

The actions on BKCa of the drugs RP-62719, UK-68798, tedisamil and risotilide were studied by adding the agents to the 140 mM K+ solution (C) bathing the inside of the cell. The addition of submicro- to micromolar quantities of the individual drugs to the patch showing Ca2+-activated K+ channel activity caused an increase in transitions from the open to closed state. The frequency of flickers increased with increasing drug concentration. An example of RP-62719 block is given in Fig. 9. Openings in the absence of the drug are shown in Fig. 9a. The addition of 0.5/*M RP- 62719 (Fig. 9b) resulted in an increase in the frequency of transitions from the open level to a non-conducting state with no change in the amplitude of the currents. Increasing the concentration of RP-62719 (to 5 ^M) caused a further increase in flickering between open and non-conducting states as shown in Fig. 9c. In order to resolve channel kinetics the sampling rate was increased to 20 kHz and the low-pass filter was raised to 5 kHz, (Fig. 9c'). Drug actions could be readily washed out (not shown).

As FIG. 10a illustrates, open time distributions for drug were fitted reasonably well with single exponential probability density functions (smooth curves). (The values of the exponential time constants are indicated in each panel). The distributions for closed times with RP-62719 (Figure 10b) were best fitted with two components, reflecting the bursting activity of the channel. The fast component of the closed distributions was due to gaps within bursts, and the longer component represented the closed times between bursts. RP-62719 had no significant effect on either of the two components in the distributions of closed Wang, X.P. 56

Figure 9

---V\^^ o

b.

Jwt5^^ o c.

5 pA > 10 ms

'• \J

5 pA 2.5 ms Wang, X.P. 57

Figure 9: Effect of RP-62719 on BKCa:

Single-channel currents with V = 0 mV and inside-out patch, a, openings in control bath solution; b, openings with RP-62719 = 0.5 ^M. c, openings with RP- 62719 = 5/xM. In a-c the sampling frequency was 5 kHz and the low-pass filter was set at 2 kHz; c', openings with RP-62719 (5A*M), with sampling frequency increased to 20 kHz and low-pass filter at 5 kHz. Wang, X.P. 58 Figure 10 A. 200 T = 3.4 ± 0.2 ms RP-62719 = 5 yM n = 834

N

—i— —i— —i 10 20 30 40 50 Open Time (ms) B. 500 T Tcf = 0.8 ± 0.01 ms

rCf = 2.4 ± 0.4 ms n = 796 RP-62719 = 5 nM

N

i 1 1 i 1 1 1 1 i 012345678 Closed Time (ms)

C. ji = 5.5 ± 0.1 pA 120 T

N

0-> r 5 10 i (pA) Wang, X.P. 59

Figure 10. Effect of RP-62719 on BKCa Channel Kinetics and Amplitude.

Distributions with RP-62719 (5/zM). In A for open time with the mean open time

T0 of 3.4 ± 0.2 msec and in B for closed time [two components with values of 0.8 ± 0.01 msec for the fast component (rcf) and 2.4 ± 0.4 msec for the slow component 0"cs)]- (P) Amplitude distribution with a mean of 5.5 ± 0.1 pA. Refer to Figure 7 for distributions with control solution. Wang, X.P. 60

times. These results are representative of collected data from seven patches. The holding potential of all patches was 0 mV.

Typical distribution for the amplitudes of the open state is shown in FIG. 10c. The shape of amplitude distributions were not changed by the drug, which indicates that there were no other channel or channel subconductance state induced. A quantitative comparison was done in which the amplitude analysis was performed both in control and in the presence of RP-62719. The ratio of the amplitude between the control and in the presence of drug with similar conditions was ~ 1.

These results are consistent with open channel blockade of KQQ. i.e. within a particular flickering burst. The kinetic behaviour of the channel can be described in terms of a closed state (C), an open state (0) and a blocked state (B), connected in sequence as follows:

k, k2 C o B

k-i k_2 where k-| and k^ represent the respective onward- and off-rate constants from the closed state to the open state. The blocking (onward) rate constant k2 is a second

1 1 order rate constant for association of the blocker (in M" S" ). In this model k_2 is the (first order) rate constant for the dissociation of the blocker. In this scheme the distribution of open channel lifetimes in the presence of drug would be a single exponential whose rate of decay («') should be given by the sum of the rate constants

1 leading away from the open state i.e. <*' = r0" = k^ + k2 [D]. The rate constant (a') for channel block was ploted as a function of drug concentration in Fig 11. The value of a' increased linearly with RP-62719 concentration. The slope of the fitted line corresponds to the onward (blocking) rate constant k2. The Y intercept of the fitted line gives k^, the closing rate with no drug present. The values of k2 were not Wang, X.P. 61 Figure 11

a

9UK68798 _* 800 Tedisamil •RP62719 .ARisotilide 1.0 Inside-out Mode 1 UK68798 _* Tedisamil .. * RP62719 *Riso«ide hinside-ou t mode ---* 600 0.0

* ^"' In a' (lO-'toc-'J ^ ,*•«* 200

*•-' -2.0 H ..-••#' A'

-3.0

10 -2.5 [Drug] frM) 0.0 23 In [Drug] (JJM)

C

800 "j OUK68798 _*Tedisamil ...jftflPB2719 .AHisotilide 1.0 I Outslde-ouOlltfil/in—suit ModIJ^-Je_ UK68798 _* Tedisamil ...• RP62719 A 1, * RisotNlde 600 o.o H Outside-out Mode aW'H In a' MO-W-')

200 H • -2.0 '/ A A-'

-3.0 >••••" * 6 10 -2.5 -, [Drug] ( M) 0.0 M 23 In [Drug] (JJM) Wang, X.P. 62

Figure 11. r 0' - Drug Concentration Relationships Data shown in a and b were obtained from inside-out experiments. Data were plotted on linear and natural logarithmic scale respectively. The rate constants with drugs (<*') increased linearly with increasing drug concentrations within the concentration range studied (0.1 ~ 10i*M). For RP-62719 each point was averaged from 6 patches, the standard deviations were within the range of each point. Refer to Table 2 for detailed sample sizes of UK-68798, tedisamil and risotilide. Data shown in c and d were obtained from outside-out experiments. Wang, X.P. 63

significantly changed when Vp = -20 mV or Vp = -40 mV thus the blocking actions of the drugs were not dependent on patch potential (Fig 12).

The distribution of shut times (blocked and closed times) were a sum of the two exponential functions. The slow component (which should attribute to the closed state) was not significantly changed by the drug. The fast component is a measure of

k_2, which is essentially constant with changing RP-62719 concentration. Such a result is consistent with an open channel block model.

Very similar behavior was found with other putative class III drugs, including UK-68798, tedisamil and risotilide. However the potency for channel block varied considerably with different drugs. The actions of UK-68798 (at 1 /*M), tedisamil (at 2 nM) and risotilide (at 5/xM) are shown in FIG. 13. UK-68798 was the most potent drug in terms of shortening single channel mean open times as shown in FIG. 11. The

results of the drug actions on BKca are summarized in Fig. 11 and Table 2. For all drugs the rate constant a' showed a linear dependence on drug concentration. This point is important in establishing validity of open channel block for all drugs over the concentrations studied (Fig. 11). The approximate values for the drug concentrations required to halve the mean open time were as follows: UK-68798 (0.4 nM), tedisamil

(1PM), RP-62719 (3.5/xM) and risotilide (7.5PM) (Table 2). 3.2.2 Outside-out Patch Experiments: The external actions of the four drugs were also studied by adding the agents to the solution bathing outside-out patches. In these experiments the pipette solution contained 140 mM K+ and 200 nM Ca2+. As shown in FIG. 14a, with the patch potential at 0 mV, relatively long openings in control solution were obtained however after the application of the agents the openings showed rapid flickering transitions to a non-conducting state (FIG. 14b,c,d) in a manner similar to that found with inside- Wang, X.P. 64

Q.E uo o C\J

CM > E o CM O Wang, X.P. 65

Figure 12: Voltage dependence of RP-62719 effect on BKca

Effect of RP-62719 (5 MM) on BKca, at different patch potentials; inside-out patch. Solid lines indicate closed levels whereas dotted lines represent open levels. The low-pass filter was set at 2 kHz and the sampling rate was 5 kHz. Wang, X.P. 66 Figure 13

d

wWlAiW 5 pA 10 ms] Wang, X.P. 67

Figure 13: Effect of UK-68798, tedisamil and risotilide on BKca Single-channel currents with other putative class III drugs. Data obtained from an inside-out membrane patch with V = 0 mV. a, in control solution; b, with UK-68798 (1 /*M) added to control bath solution; c, return to control solution; d, Addition of tedisamil (2/iM) to control solution; e, risotilide (5/iM) was added to control solution after recontrol (not shown) solution to washout tedisamil effect. The low-pass filter was set at 2 kHz and the sampling rate was at 5 kHz. Wang, X.P 68

TABLE 2

Normalized Mean Open Time, Onward (blocking) Rate Constants for Drugs

The normalized mean open times (TN = TO'/TO), at V = 0 mV, were determined by dividing the open times in the presence of the drugs (To') by the mean open times (To) in the absence of the drug (control) solution. For inside-out mode. To = 14.7 ± 1.0 msec, n = 11, thus k _f = 68.0 sec"1; for outside-out mode, To = 17.7 ± 1.8 msec, n = 9, therefore _1 k .1 = 56.5 sec . The rate constant (a') is given by a ' = k ., + k 2 [DJ. See Fig. 11 for linear regression functions describing each drug action.

Drug [Drug] n a ' k TN 2 ^M sec"1 107 IvHsec"1

INSIDE-OUT MODE

UK68798 0.1 3 0.71 95.8 0.5 4 0.40 170.0 29.5 1.0 2 0.19 358.0

Tedisamil 1.0 5 0.48 141.7 5.0 4 0.16 425.2 5.2 10.0 3 0.11 618.4

RP-62719 0.5 5 0.64 106.3 6.0 5.0 6 0.18 377.9

Risotilide 5.0 3 0.62 109.7 1.3 10.0 2 0.35 194.4

OUTSIDE-OUT MODE

UK68798 0.2 3 0.52 130.8 0.5 5 0.34 200.0 26.4 1.0 3 0.20 340.1

Tedisamil 2.0 4 0.24 283.4 10.8

RP-62719 0.5 2 0.75 55.0 1.0 2 0.57 72.0 4.0 2.0 2 0.36 114.0 4.0 2 0.23 174.0

Risotilide 5.0 3 0.50 136.1 10.0 3 0.35 194.4 1.3 Wang, X.P. 69

out patches. Analysis of the data from outside-out patches was done using the

simple open channel block scheme and values for the blocking rate constant k2 are included in Table 2. The potency for channel block, as defined from the magnitudes

of k2 for external application of the drugs was in the order UK- 68798>tedisamil>RP62719>risotilide; that is the same as determined for inside-out patches. The values of k2 were not significantly affected when the patch potential was changed to 20 mV or 40 mV. The approximate drug concentrations resulting in mean open times being halved were: UK-68798 (0.2 ^M), risotilide (5/iM), and RP-62719 (3.5 fi M). Although only one concentration of tedisamil was used, a value close to that found with inside-out patches (1 A«M) is suggested from the data.

3.3 THERMODYNAMIC STUDIES

A number of drugs depress macroscopic currents through voltage-and chemically-gated ion channels in a manner qualitatively consistent with channel blockade. At present, however, very little is known concerning the molecular mechanisms underlying channel block. It is often assumed that blockade involves simple diffusion of the blocker to a site and that the blocking action represents occlusion of the channel. However, there are little data available to support this argument. One useful approach in the determination of molecular mechanisms involved in channel block would be to examine the temperature dependence of the process. For example, the temperature dependence of channel mean open time could be used to determine the effects of temperature on specific rate constants, assuming a particular channel blocking model. A simple model for channel block is a sequential open channel scheme whereby an activated channel can undergo transitions to the usual closed state or to an additional blocked state. Measurement of the temperature dependence of the kinetic rate constants in the blocking interaction could then be used to better characterize the molecular nature of the SUJ 09 o Vd g a. X c CO -W^UiAWU/ irafcist^^

J ll, 1, I Ji *''•*' . . J i k . ., , I

'Ti-jr > •••» -|T M^^mmmmmmwim."" r'T"-T»m*— -p.1|pnT P^.„. ~.M_,,rv „, inr 0-

o-"^^^

frl 9jn6jj Wang, X. P. 71

Figure 14: Effect of Drugs on BKfja at V = 0 mV; Outside-out Experiments. a. With control bath solution containing 5 mM KCI and 140 mM KCI patch pipette solution. The mean open time (rQ) was 17.7 ±1.8 msec, n = 10. b. With UK-68798 (0.1 /zM) addedto the control bath solution facing outside of the patch membrane. c. With tedisamil (1 n M) added to control bath solution. d. With RP-62719 (5 MM) added to control. e. With risotilide (5 n M) added to control bath solution. All data were from the same patch, with re-control solution applied between each drug application (not shown) to ensure the recovery of channel kinetics from previous drug action. Low-pass filter was set at 2 kHz; each trace represented 400 msec of consecutive sweep. Wang, X.P. 72

interactions between the drug and the ion channel site. For example, if the mean open time and the onward (blocking) rate constant were relatively temperature independent, then the results would be in reasonable accord with a diffusion-limited process. If, on the other hand, the kinetics of channel block were strongly temperature-dependent, then a non-diffusional process would be suggested. In this case, information regarding molecular mechanisms may be obtained by measurement of the temperature dependence of channel blocking rate constants in order to determine the thermodynamic parameters, such as enthalpy and entropy, which are associated with the blocking reaction.

3.3.1 Temperature Dependence of Kga Channel: Typical channel events, recorded from an inside-out patch with V = 0 mV, are shown in Fig. 15 (trace a) with the bath temperature held at 24° C. When the temperature was decreased to 19°C (Fig. 15, trace b) and 14° C (Fig. 15, trace c), the amplitudes of the unitary currents were progressively diminished; however, the durations of open times remained relatively constant. In Fig. 16, a plot of the channel mean open time dependence on temperature is shown for 6 patches, including the patch with the channel openings illustrated in Fig. 15. The mean open times were not

significantly changed over a temperature range from 24° C to 14°C. A Q10 value of 1.2 ± 0.2 was determined for the temperature dependence of mean open time. At all temperatures the open time distributions were well-fit with single exponential functions. Typical distributions for channel open times are shown in FIG. 17A with T = 24° C and in Fig. 17B with T = 14° C. In most instances the distributions for channel closed times required fitting with a two-component exponential function. Typical distributions for channel closed times are illustrated in Fig. 17C with T = 24° C and in Fig. 17D with T= 14°C. Wang, X.P. 73 o o

Q- E LO O C\J

10

Figure 15: Temperature Dependence of BKca Single-Channel Currents

Data recorded from an inside-out membrane patch bathed in control 140 mM KCI solution at V = 0 mV. The temperatures of the bath solution were: The top trace: 24 °C The middle trace: 19 ° C The bottom trace: 14 °C The low-pass filter was set at 2 kHz and the sampling frequency was 5 kHz. Wang, X.P. 75

Figure 16

25 -,

20 H

r0(ms)i5 J

1

5 -•

—r— —r— ~~r~ —1 14 16 18 20 22 24 T (°C) Wang, X.P. 76

Figure 16: Temperature Dependence of BKfja Channel Mean Open Time

Plot of the dependence of mean open time (r0) on temperature. Each point was the average value of 6 patches. The data were fitted using linear regression analysis. The Q10 associated with the temperature dependence of mean open time was 1.2. Figure 17 Wang, X.P. 77

B

T = 24"C 150

T0 = 15.2 + 0.2 ms T0 = 17.6 ± 0.2 ms Control Control n = 1951 n = 747

N

o-> —I— 20 40 60 20 40 60 Open Time [ms] Open Time (ms)

C D 400 T = 24°C 150 T = 14°C rCf = 0.79 + 0.03 ms TCf = 0.53 ± 0.03 ms rc, = 3.86 ± 2.09 ms rCs = 5.34 ± 0.49 ms Control Control

N N

J o 0J ••H^- '*•• jw^Mwi^i^rh. 16 Closed Time (ms) Closed Time (ms) E F

150 -i T = 14°C 400 T = 24*C /J = 3.71 ± 0.1 pA \i = 5.2 ± 0.1 pA

N N

-? 10 12 10 i (pA) i (pA) Wang, X.P. 78

Figure 17: Effect of Temperature on BKQQ Channel Kinetics and Single-channel Current Amplitude.

Data were obtained from an inside-out membrane patch, n represented the total number of events included in the curve fit. For data obtained at 24° C, 1951 events were included in channel open, closed times and amplitude distributions; for 14°C, 747 channel events were used. Other abbreviations were as indicated before. Wang, X.P. 79

As shown in Fig. 15, the amplitudes of the single channel currents were diminished with decreasing temperature. Typical amplitude distributions of 24° C and 14° C are shown in Fig 17E and 17F. In six patches, the ratios of the channel height at

14° C to that at 24° C were determined. The Q10 value associated with the temperature dependence of current height, over the temperature range from 24° C to 14° C, was 1.4 ± 0.1 (n = 6). Thus both the mean open time and the amplitude of

single channel currents for Kca were not strongly dependent on temperature. 3.3.2 Temperature Dependence of Drug Actions: The dependence of drug actions on temperature was investigated over the temperature range from 24° C to 14°C. In Fig. 18, unitary currents, in the presence of RP-62719 (5 pM), are shown for temperatures of 24° C, 19° C and 14° C, respectively. The amplitudes of the currents and the channel open times were sensitive to temperature with decreasing T acting to diminish amplitude and to prolong open time. Typical distributions for channel open and closed times with RP-62719 are shown for temperatures of 24° C and 14° C in Fig. 19. The open time distributions, (FIG. 19a,b), were well-fit with single exponential functions at both temperatures however decreasing T by 10°C caused a considerable prolongation of mean open time. Distributions for closed times at both temperatures (FIG. 19c,d), over the range from 0 to 15 msec, were fit with two exponential component functions. As shown in Fig 19, the distributions for closed times were not significantly dependent on temperature in this patch. The average values of mean open times from 7 patches have been plotted as a function of temperature in Fig. 20. The Q10, as determined from linear regression fitting to the data, was 2.2 ± 0.3. Thus, the channel blocking action of RP-62719 exhibited a considerable dependence on temperature, whereas in the absence of the Wang, X.F> 80 o o

< 00 CL E If) o CM

GO

a> Wang, X.P. 81

Figure 18: Temperature Dependence of RP-62719 Effect on BKCa.

Single-channel currents with RP-62719 (5/^M) in the bath solution facing the inside of the patch membrane. The bath temperatures were: top trace (24 °C); middle trace (19 °C); bottom trace (14 °C). The patch potential was held at 0 mV throughout the experiment. Wang, X.P. 82

Figure 19

350 T = 24°C T = 14°C r = 2.65 ± 0.16 ms 0 r0 = 5.82 ± 02 ms RP-62719 = 5 uM RP-62719 = 5 jiM

N N

Open Time (ms) Open time (ms)

350 T = 24'C T = 14°C 7"c, = 0.9 ± 0.04 ms Tc, = 1.1 ± 0.01 ms To. = 2.5 ± 0.38 ms RP-62719 = 5 uM To, = 2.6 ± 12 ms RP-62719 = 5 JJM N N

4 « 12 16 8 12 Closed Time (ms) 16 Closed Time (ms) Wang, X. P. 83

Figure 19: Quantitative Analysis of Temperature Dependence of RP-62719 Effect on

BKca Channel Kinetics

RP-62719 (5 A*M) was included in the bath solution bathing an inside-out membrane patch, V = 0 mV. Low-pass filter was set at 2 kHz and the sampling rate was at 5 kHz. Wang, X. P. 84

Figure 20

8 n

6 J J" r'fms)

24

o -J - r ~T~ 16 14 18 20 22 24 T (°C) Wang, X.P. 85

Figure 20. Temperature Dependence of RP-62719 Effect on BKQa Channel Mean Open Time (r0')

Plot of mean open time dependence on temperature with RP-62719 (at 5/xM) added to the bath solution. The values of mean open time were averaged from 7 patches. The fit to the data was determined by linear regression analysis. A Q 10 of 2.2 was found for the temperature dependence of mean open time (r 0'). Wang, X.P. 86

drug, channel open times were not strongly sensitive to temperature (Q10 value of 1.2 ± 0.2). The mean open time dependence on temperature was determined at a single potential (V = 0 mV). The blocking rate constant k2 was determined at different temperatures using the data from the graphs of mean open time versus temperature for control (Fig. 16) and with RP-62719 (Fig. 20). The magnitudes of k-i, as a function of T, were first found by inverting the mean open times in Fig. 16. Substitution for the values of k^ and the values of mean open time from Fig. 20 (in the presence of the drug) into the channel block equation allowed solution for k2. The results obtained at the extremes

7 1 1 of the temperature range were: k2 (24° C) = 5.9 ± 0.2 x 10 M" s" and k2 (14° C) = 2.2 ± 0.2 x107 M"1 s"1. Using these values for the channel blocking rate constants, at the temperatures of 24° C and 14° C, yields a Q10 of 2.7 for the blocking rate constant. Although the temperature dependence was determined for a single concentration of RP-62719 (at 5/iM), a value for the onward rate constant was also found with RP-62719 at 0.5 ^M (T = 24° C). At the lower concentration of the drug,

7 1 1 the k2 was 5.5 ± 0.3 x 10 M" s" (n = 3). This value for k2 was not significantly different compared to the magnitude of k2 found with RP-62719 at 5 /*M with T = 24° C.

The transition state thermodynamic parameters associated with k2 were found using the procedures described in METHODS (Minneman et al., 1980). An Arrhenius

1 plot (In k2 vs. T " ) was first constructed and is shown in Fig. 21. The transition state activation energy Ea was found from the product of the slope of the linear relation in Fig. 21 and the universal gas constant R. This value of 17.3 kcal/mol, was then used (see METHODS) to determine the enthalpy of activation, AHt and the free energy of activation, AG* (Table 3). From these results the entropy of activation, AS* was then calculated to be +33 entropy units (e.u). Wang, X.P. 87

Figure 21

18 n

ln(k2)

16.5

336 338 340 342 344 346 348

TVK-'MCT5 Wang, X.P. 88

Figure 21. Arrhenius Plot for k2

The Arrhenius equation expresses the dependence of the rate constant k2 on Ea RT temperature T (° K) as follows k2 = A*e" / or lnk2 = InA - Ea/RT where A was a constant for the reaction and R was the universal gas constant. Linear regression analysis was used to fit the data. The slope of the fit was -Ea/R. Wang, X.P. 89

An estimate for the temperature dependence for the channel unblocking rate

constant k_2 was also done from examination of the single channel records. Since

Kca exhibits bursting behavior in the absence of drug it was necessary to differentiate between channel transitions to the non-conducting state during bursts (blocking

events) and longer channel closing events between bursts. The values of k_2 were: at

1 1 T = 24° C (k_2 = 550 s" ); at 14° C (k_2 = 377 s" ). These values were not dependent on the magnitude of the burst duration for durations in excess of about 2 msec. The

ratio of the k_2 values yields a Q10 of 1.5. An analysis, using the transition rate theory, was also applied to the channel

unblocking rate constant k_2. The parameters for free energy of activation, enthalpy and entropy are included in Table 3. A value for the activation energy of 6.7 kcal/mol was considerably lower than the corresponding activation energy determined for the channel block rate constant. The lower activation energy for k_2 suggests that the channel unblocking process has a lower temperature sensitivity than the channel blocking step. The negative entropy associated with the unblocking rate constant (Table 3) indicates an increase in the order of the system. Thus, the entropy calculated for the channel unblocking rate constant is opposite in sign to that found for the entropy change determined for k2. The thermodynamic parameters for the associated state of the system can be determined from the transition state parameters (see METHODS). The data are presented in Table 3 and show positive values for both enthalpy and entropy. The overall free energy change in the blocking interaction, determined by subtracting the free energy value associated with k_2 from the corresponding value with k2, was -6.8 kcal/mol. A negative value for the associated state free energy would be expected for a reaction which decreases the total energy of the system. The driving force for the blocking interaction is the large increase in the entropy. Wang, X.P 90

TABLE 3

Transition State and Associated State Thermodynamic Parameters

TRANSITION STATE PARAMETERS T ,t RATE CONSTANTS Ea AH' AG. t AS* (kcal/mol) (kcal/mol) (kcal/mol) (e.u.)

1 k2 (M-V ) 5.9 x 107 17.3 16.7 6.9 + 33.0

1 k 2 (S" ' 550 6.7 6.1 13.7 -25.6

ASSOCIATED STATE PARAMETERS 10.6 -6.8 + 58.6

All data areforT = 24°C The values for associated state parameters were determined from the differences between the transition state parameters for channel blocking and unblocking. The rate constants are values with V = 0 mV at T = 24 °C. Wang, X. P. 91

4. DISCUSSION This work has concerned study of the single channel properties of a large conductance calcium-activated potassium channel in cultured rat CA1 hippocampal neurons and the changes in properties caused by class III antiarrhythmic drugs. The channel under study can be considered as a ligand- and voltage- activated ion channel macromolecule that contains both the agonist (calcium) receptor site and the voltage sensitive molecular machinery to generate a response. The channel serves an important function in the regulation of neuronal excitability.

4.1. BKca CHANNEL PROPERTIES: With physiological concentrations of K+ (140 mM intracellular KCI, 5 mM extracellular KCI) the single channel conductance was 80 ± 20 pS, which was close to

values found for other large conductance Kca channels. For example, values of 65- 140 pS Kca were found using a rat brain synaptosomal membrane preparation

(Joseph et al, 1987) and Levitan et al (1991) has described a family of Kca channels from rat brain including conductances of 135 ± 10 pS and 76 ± 6 pS. Somewhat higher values have also been found with 150 mM K+ across patches: 180 pS (Marty 1981); 200 pS (Benham et al. 1986); 210 pS (Wong et al. 1982) ; 220 pS (Franciolini 1988).

The activity of the channel was a function of both membrane potential and [Ca2+]j. At lower [Ca2+]j (near 10 /*M), channel activity can be modulated by varying the membrane potential whereas at high [Ca2+]j (200 ^M) the channel activity was independent of voltage. The channel was preferentially active at depolarized voltages due to increased rate of opening. The percentage of time the channel spends in the open state increases with a voltage dependence of 30 mV per e-fold change of p0 (FIG.6). This voltage sensitivity was similar to that reported for the

+ macroscopic BKCa currents in Aplysia neurons (25 mV per e-fold change in K Wang, X.P. 92

conductance: Gorman and Thomas 1980), or for single BKca channel of clonal anterior pituitary cells and smooth muscle cells (28 and 30 mV respectively: Wong et al. and Benham et al.).

The BKca channel described above is likely the single channel basis for the macroscopic current Ic in hippocampal neurons. Ic was blocked by TEA (at 1 mM) applied from the bath solution bathing the cytoplasmic surface of the cells and was not altered by 4-AP (at 5 mM). Depolarization potentials evoked Ic presumably by increasing the frequency of single channel openings as noted in the present study. Ic is involved with both the repolarization phase and the fast after-hyperpolarization phase of action potentials thus serves an important role in neuronal excitability.

4.2 OPEN CHANNEL BLOCKADE AS THE MECHANISM OF CLASS III

ANTIARRHYTHMICS BLOCK OF THE BKCa CHANNEL: The data was consistent with open channel blockade as the mechanism of action for class III antiarrhythmic drugs including RP-62719 UK-68798, tedisamil, and risotilide. Similar results were found for the effects of KC-8851 (analog of tedisamil) in previous work (McLarnon 1990). The amplitudes of the single-channel currents were not significantly altered by any of the drugs. The channel conductance and selectivity were not changed by the drugs within the tested dose ranges.

Channel openings in the presence of the agents showed increased numbers of transitions from the open level to a nonconducting state. The frequency of the transitions was increased with the dose of the drugs. A simple open-channel sequential scheme was used to determine the onward (blocking) rate constants (k2) for the drugs, and the potency of the agents to diminish channel mean open time, for both external and internal patch applications, was in the order UK- 68798 > tedisamil > RP-62719 > risotilide. Wang, X.P. 93

The mean open times of the Kca channel were not significantly changed with the drug applications to inside-out patches at different transmembrane potentials (V = - 20 mV,-40 mV) or to outside-out patches (with V = 20 mV, 40mV). These results would suggest that the neutral species of the drugs were the active forms, and that hydrophobic membrane sites were involved in the drug actions. A hydrophobic binding site is also consistent with the result that UK-68798 has a considerably higher potency for channel block than does risotilide. The two agents have similar structures; however, UK-68798 would have a higher degree of lipid solubility than would risotilide.Tedisamil (a dihydrochloride derivative, not a sulfonamide compound) and RP-62719 (a benzopyran compound) exhibited intermediate potency between those of UK-68798 and risotilide. In this case potency is defined as the concentration of drug which halved the mean open time of Kca-

The similarities in the K2 values for the drug actions on inside-out or outside-out patches would suggest that the blocking sites might be the same. For example, if the binding site was reached from the internal side of the membrane or from the lipid, then external applications of the drugs could block the channel subsequent to diffusion of the agents through the lipid region.

A detailed analysis for the off-rate constant (k_2 in the open channel block scheme) was not possible. In the absence of the drugs, the single-channel records showed bursting behavior of channel activity, and two-component exponential functions were generally required to fit distributions of closed times. In most cases, the magnitudes of the two components were similar, and on the basis of the single- channel records it was difficult to distinguish between shut and blocked states.

In essence, bursts of channel-blocking events were often not resolvable from sojourns to the shut level (such as occurred in the absence of the drug) in the records. Estimates for the dissociation constant, KD, could be obtained from the single-channel records, however, if one assumed that in the presence of the drug the Wang, X.P. 94 majority of rapid transitions from the open state represent blocking episodes. Thus, for example, in the records for UK-68798 (Fig. 12) the off-rate constant was estimated to be near 1000/sec, which, coupled with the blocking rate constant from Table 2, would yield a value for KD near 5 n M. All of the drugs studied here are generally classified as class III antiarrhythmic agents, by virtue of actions to prolong action potentials in cardiac tissue with little or no effect to slow conduction velocity; such actions are consistent with block of K channels (Colatsky,1989). This work includes studies of UK-68798 and risotilide (Follmer, 1989) in ventricular muscle and of the effects of tedisamil on rat myocytes (Dukes, 1989) and mammalian cardiac and glial cells (Dukes, 1990). Some studies have been carried out to determine the sites and molecular basis for the effects of the drugs in cardiac tissue, in preliminary experiments (using whole-cell patch clamp), UK-68798 at 2MM decreased the amplitudes of outward tail currents in guinea pig ventricular myocytes (Gwilt et al. 1989). These results were suggested to be due to actions of the drug on l^. It should be noted that in pharmacological and toxicological studies, at a wide range of doses, no obvious effects of UK-68798 on normal neuronal function have been observed. Block of lk was also suggested from preliminary voltage-clamp experiments on the actions of risotilide on cat ventricular myocytes (Follmer and Colatsky, 1990). Recent studies using other sulfonamide agents have been described (Sanguinetti et al. 1990). Both d-sotalol and E-4031

(IC50 is 397 nM) were found to block a rapidly activating component of the l^ in guinea pig ventricular myocytes. It would seem reasonable to assume that the greater potency for the blocking actions of E-4031, relative to sotalol, could be attributable to increased lipid solubility of the former compound.

Tedisamil has been studied extensively. Whole-cell patch-clamp experiments have shown that tedisamil at 1-20/^M significantly hastened the decay phase of lt0 in rat ventricular myocytes (Dukes, 1989). The actions on lt0 were independent of Wang, X.P. 95 membrane potential. In addition, the drug was equally efficacious when applied from inside or outside the cells. Interestingly, lt0 in cardiac cells has been suggested to depend, at least partly, on calcium concentration (Siegelbaum, SA 1980, Coraboent E 1982). More recently, tedisamil has been applied to guinea pig ventricular myocytes and mouse astrocytes (Dukes 1990). With the cardiac cells, tedisamil at 3 or 10 >iM decreased lk with a KD of 2.5 /uM. With the glial cells, tedisamil at 10 /*M

+ decreased both transient (la) and delayed (lk) K currents. These results were interpreted as voltage-independent channel block to tedisamil acting at a site accessible from the cytoplasm. A comparison of the data on tedisamil block of lt0 in cardiac cells and the present data on the block of IKca in hippocampal neurons shows a significant degree of commonalty. This includes similar concentrations of drug applications that were effective in channel blockade, effects from the drug applied to the outside or the inside often cell membrane, and drug actions that were independent of potential.

4.3 MOLECULAR MECHANISM OF RP-62719 BLOCKADE OF BKca CHANNELS: A sequential open channel block scheme (Adams, 1977, Neher and Steinbach,

1978) was used to determine the blocking (k2) and unblocking (k_2) rate constants for the interaction of RP-62719 with the Kca channel. In the open channel block model the rate constant k2 represents the binding of the drug to a channel site. This scheme, which is perhaps most relevant to charged ion block of ion channels from outside, assumes that the only pathway from the blocked state is via the open state.

Such a simple model is only an approximation however, since RP-62719, with a benzopyran structure, would be expected to exhibit a considerable degree of lipid solubility.

In the presence of RP-62719, the mean open time of the channel was significantly prolonged with diminished bath temperature (Q10 of 2.2). In the absence Wang, X.P. 96

of drug the kinetics of Kca were essentially independent of temperature; in this case a

Q10 value of 1.2 was found for the mean open time. The temperature dependence of the onward (blocking) rate constant k2 was determined from measurements of the mean open times (with RP-62719 at 5/«M) and the channel closing rate constants

(with no drug) for bath temperatures in the range 14° C to 24° C. The rate constant k2 was strongly dependent on temperature with a Q10 value of 2.7. The magnitude of k2 (24° C) was about 6 x 107 M"1 s'1 which is respectively, about three times and double, the values for the k2 found with octanol and procaine block of the end-plate channel (McLamon et al, 1984). Values of k2 in the range from (1-25) x 107 M'V1 were found at room temperature with other antiarrhythmic drugs: UK-68798, Tedisamil and risotilide. These values for the blocking rate constant are at least ten times the magnitudes of the blocking rate constants found for QX222 block of inactivating Na+ channels (Starmer et al, 1986).

The relatively high Q10 values associated with the channel mean open time and with k2 would not be consistent with simple diffusion of the blocking moiety to a channel site. If diffusion were rate limiting in the channel blocking step, then the process would be expected to be characterized by a low temperature sensitivity with a Q10 value in the range of 1-1.4. For example, a Q10 of 1.4 has been determined for the temperature dependence of the single channel conductance for the cation selective gramicidin channel in lipid bilayer membranes (Hladky and Haydon, 1972; see also Hille, 1984). It is possible, however, that the movement of a particular drug in an aqueous channel or through lipid may not be characteristic of simple diffusion and a higher temperature sensitivity could reflect the drug pathway to a channel site.

In order to clarify the high temperature sensitivity for the channel block rate constants, an analysis for the thermodynamic parameters of enthalpy, free energy and entropy was carried out using the methods of transition rate theory (Minneman etal., 1980; Hitzemann, 1988; Raffa and Porreca, 1989). The results showed that in Wang, X.P. 97

the formation of the transition state (activated complex of drug and channel site) the blocking rate constant is associated with large positive entropy and enthalpy changes. Since A G+ = AHt-TASt, the blocking process is favored (decrease in free energy) by the change in entropy and hindered by the change in enthalpy. These results are similar to those found with block of end-plate channels with n-alkanols and local anaesthetics (McLarnon and Quastel, 1984) and with binding of ligands to adrenergic receptors of turkey membrane (Minneman et al., 1980). In both cases the large increases in entropy were suggested to be consistent with increased disorder in the systems through the formation of hydrophobic bonds (Kauzmann, 1959) between the interacting species. Within the series of n-alkanols, the potency for block of end- plate channels was correlated with the degree of hydrophobicity of the agent (McLarnon and Quastel, 1984, McLarnon etal., 1986). The magnitudes of the activation energies and entropy values for both block of the end-plate channel by alcohols and for block of Kca by RP-62719 were very similar, strongly suggesting that hydrophobic binding could also account for the latter actions. Indeed, a significant hydrophobic component in the channel block of Kca by several other drugs has been documented here. The large positive value for enthalpy would suggest that the contribution of non-hydrophobic interactions, e.g., in the formation of hydrogen bonds, to be relatively unimportant in channel blockade.

The large values of Ea (in excess of 10 kcal/mol) and entropy determined for k2 are considerably higher than would be expected if the kinetics of channel block were determined by simple diffusion of the drug in an aqueous medium. The results suggest that channel block could be a consequence of the drug interaction with a hydrophobic portion of the channel protein. Such an interaction could then account for the substantial increase in entropy through a decrease in the structured water of hydration at the site and possible destabilization of the protein segment. Wang, X.P. 98

The Q10 associated with k_2 was 1.5; Thus, the channel unblocking step exhibited a considerably lower dependence on temperature compared with the blocking step. Another difference between the blocking and unblocking steps was the calculation of a negative entropy for the unblocking rate constant. A negative

entropy signifies increasing order which is opposite to that found with k2. The channel unblocking step could also indicate changes in conformation of the channel protein or perhaps in the structured water near the site of interaction. In this case the changes are opposite to those associated with the blocking step. It seems

reasonable to assume, that the magnitude of k_2 is strongly dependent on the degree

of lipid solubility of the drug. The k_2 for RP-62719 block of KCa is very close to that found with block of the end-plate channel by the lipid soluble n-alkanol, octanol, at the mammalian neuromuscular junction (McLarnon and Quastel, 1984). The associated state thermodynamic parameters can be found from the transition rate values (Minneman et al., 1980). For example, the associated state free

energyAG° is AG* (k2) -AG* (k_2) = 6.8 kcal/mol. A negative free energy is to be expected for the spontaneous blocking reaction.

The associated state entropy AS° is ASt(k2) - ASt(k_2) = + 58.6 c.u.. Thus, the blocking reaction appears to be consistent with a rate-limiting, entropy-driven, step through the formation of hydrophobic bonding between the drug and channel site. It should be noted that the associated state parameters are equilibrium values if channel block and channel unblock are established from the same rate-limiting process. If this was the case (as suggested from the different signs of the entropy terms for the rate constants), then the blocking and unblocking steps have a large difference in temperature dependence. Although the interpretation of the present data is necessarily speculative, the results suggest that channel block could be a complex process. In this case it is interesting that the thermodynamic parameters for block of KCa by RP-62719 are Wang, X.P. 99

close to those determined for n-alkanol block of the end-plate channel (McLarnon and Quastel, 1984). This point may reflect the general commonalty in channel

structures. The Kca channel, like other ion channels, is an aqueous pore; it is likely that the channel also includes regions of hydrophobic interfacial pockets. In terms of energy considerations, a non-polar region of a blocking molecule would preferentially interact with hydrophobic segments of the channel protein. The large positive

magnitudes of entropy associated with the channel block rate constant k2 could then be due to increased disorder in the system involving changes in structured water or changes in protein conformation. It should be noted that alterations in the temperature of the bath solution could also have effects on other aspects of the system. For example, changes in the buffering capacity of the solution or on the degree of ionization of the drug, could occur. It would be expected, however, on the basis that the temperature variation in this study was over a range of only 10°C, that such changes would be relatively small. Methods, other than changing ambient temperature, have been utilized in order to determine the biophysical properties of drug interactions with ion channel proteins. Patch clamp measurements have been combined with alterations in solution viscosity to study the interaction of the neurotoxin charybdotoxin with a calcium-dependent K+ channel in planar lipid bilayers (Miller, 1990). The results were in accord with rate- limiting diffusion controlling binding of the agent to a site in the channel mouth. Measurements of conformational changes in membrane proteins, induced by a number of compounds, have recently been reported (McCarthy and Stroud, 1989a,b). The kinetics of tritium-hydrogen exchange were used to demonstrate that binding of competitive inhibitors to the acetylcholine receptor (AChR), such as d- tubocurarine, caused significant conformational changes in the receptor complex (McCarthy and Stroud, 1989a). In addition, a probe specific for hydrophobic Wang, X.P. 100 segments associated with ion channel proteins, was used to show conformation changes in the subunits of AChR with the agonist carbamylcholine (McCarthy and Stroud, 1989b). It is apparent that measurements with a diversity of techniques will be necessary in order to characterize the complex interactions between ion channel proteins and modulators of the properties of the proteins. Wang, X.P. 101

5. REFERENCES Adams PR. Voltage jump analysis of procaine action at frog endplates. Journal of Physiology 268:291-318 1977.

Adams PR. Drug blockade of open end-plate channels. Journal of Physiology; 1976 Sep; 260(3); P 531-52

Adams PR. Constanti A. Brown DA. Clark RB. Intracellular Ca2+ activates a fast

voltage-sensitive K+ current in vertebrate sympathetic neurones. Nature.

296(5859):746-9, 1982 Apr 22.

Arena JP. Walsh KB. Kass RS. Measurement, block, and modulation of potassium channel currents in the heart. Progress In Clinical & Biological Research. 334:43-63, 1990.

Armstrong CM. Hille B. The inner quaternary ammonium ion receptor in

potassium channels of the node of Ranvier. Journal of General

Physiology. 59:388-400 1972.

Atwater I. Dawson CM. Ribalet B. Rojas E. Potassium permeability activated by

intracellular calcium ion concentration in the pancreatic beta-cell. Journal of

Physiology. 288:575-88, 1979 Mar.

Banker GA. Cowan WM. Rat hippocampal neurons in dispersed cell culture. Brain Research. 126:397-425 1977. Wang, X.P. 102

Banks BE. Brown C. Burgess GM. Burnstock G. Claret M. Cocks TM. Jenkinson DH. Apamin blocks certain neurotransmitter-induced increases in potassium permeability. Nature. 282(5737):415-7, 1979 Nov 22.

Banks BE. Dempsey CE. Pearce FL. Vernon CA. Wholley TE. New methods of isolating been venom peptides. Analytical Biochemistry. 116(1):48-52, 1981 Sep 1.

Banks BE. Garman AJ. Habermann E. Structure-activity studies on apamin and

mast cell degranulating peptide(MCDP)-401 [proceedings] Journal of

Physiology. 284:160P-161P, 1978 Nov.

Barrett JN. Magleby KL. Pallotta BS. Properties of single calcium-activated potassium channels in cultured rat muscle. Journal of Physiology. 331:211-30, 1982.

Beatch GN. Courtice ID. Salari H. A comparative study of the antiarrhythmic

properties of eicosanoid inhibitors, free radical scavengers and potassium

channel blockers on reperfusion induced arrhythmias in the rat.

Proceedings of the Western Pharmacology Society. 32:285-9, 1989.

Beatch GN. MacLeod BA. Abraham S. Walker MJ. The in vivo electrophysiological actions of the new potassium channel blockers, tedisamil and UK-68,798. Proceedings of the Western Pharmacology Society. 33:5- 8, 1990. Wang, X.P. 103

Benham CD. Bolton TB. Lang RJ. Takewaki T. The mechanisms of action of Ba2+ and TEA on single Ca2+-activated K+ channels in arterial and intestinal smooth muscle cell membrane. Pflugers Arch-European Journal of

Pharmacology. 403:120-27

Benham CD. Bolton TB. Lang RJ. Takewaki T. Calcium-activated potassium

channels in single smooth muscle cells of rabbit jejunum and guinea-pig

mesenteric artery. Journal of Physiology. 371:45-67, 1986 Feb.

Blatz AL. Magleby KL. Ion conductance and selectivity of single calcium-activated potassium channels in cultured rat muscle. Journal of General Physiology; 84(1); P 1-23, 1984 Jul;

Blatz AL. Magleby KL. Single apamin-blocked Ca-activated K+ channels of small conductance incultured rat skeletal muscle. Nature. 323(6090):718-20, 1986 Oct 23-29.

Blatz AL. Magleby KL. Correcting single channel data for missed events.

Biophysical Journal. 49(5):967-80, 1986 May.

Blatz AL. Magleby KL. Single apamin-blocked Ca-activated K+ channels of small conductance incultured rat skeletal muscle. Nature. 323(6090):718-20, 1986 Oct 23-29.

Bolton TB. Lang RJ. Takewaki T. Benham CD. Patch and whole-cell voltage-

clamp studies on single smooth muscle cells. Journal of Cardiovascular

Pharmacology. 8 Suppl 8:S20-4, 1986. Wang, X.P. 104

Bourque CW. Transient calcium-dependent potassium current in magnocellular neurosecretory cells of the rat supraoptic nucleus. Journal of Physiology. 397:331-47, 1988 Mar.

Brown DA. Constanti A. Adams PR. Calcium-activated potassium current in vertebrate sympathetic neurons. Cell-Calcium; 4(5-6); P 407-20, 1984 Jul;

Brown DA. Griffith WH. Calcium-activated outward current in voltage-clamped hippocampal neurones of the guinea-pig. Journal of Physiology; 337; P 287- 301; 1983.

Brown DA. G:ahwiler BH. Griffith WH. Halliwell JV. Membrane currents in hippocampal neurons. Progress In Brain Research. 83:141-60, 1990.

Burgess GM. Claret M. Jenkinson DH. Effects of quinine and apamin on the calcium-dependent potassium permeability of mammalian hepatocytes and red cells. Journal of Physiology. 317:67-90, 1981 Aug.

Capiod T. Ogden DC. Properties of membrane ion conductances evoked by hormonal stimulation of guinea-pig and rabbit isolated hepatocytes. Proceedings of the Royal Society of London - Series B: Biological Sciences. 236(1283):187-201, 1989 Mar 22. Wang, X.P. 105

Capiod T. Ogden DC. The properties of calcium-activated potassium ion channels in guinea-pig isolated hepatocytes. Journal of Physiology. 409:285-95, 1989 Feb.

Castle NA. Haylett DG. Jenkinson DH. Toxins in the characterization of potassium channels. Trends in Neurosciences. 12(2):59-65, 1989 Feb.

Cecchi X. Wolff D. Alvarez O. Latorre R. Mechanisms of Cs+ blockade in a Ca2+-activated K+ channel from smooth muscle. Biophysical Journal. 52(5):707-16, 1987 Nov.

Chabala LD. Sheridan RE. Hodge DC. Power JN. Walsh M. P. A microscope stage temperature controller for the study of whole-cell or single channel currents. Pflugers Arch-European Journal of Pharmacology. 404:374-377 1985.

Chicchi GG. Gimenez-Gallego G. Ber E. Garcia ML. Winquist R. Cascieri MA. Purification and characterization of a unique, potent inhibitor of apamin binding from Leiurus quinquestriatus hebraeus venom. Journal of Biological

Chemistry. 263(21):10192-7, 1988 Jul 25.

Christensen C, Zeuthen T, Maxi K+ channels in leaky epithelia are regulated by

internal Ca2+, pH and membrane potential. Pflugers Arch-European Journal

of Pharmacology. 408:249-59 1987 Wang, X.P. 106

Coats RA. Effects of apamin on alpha-adrenoceptor-mediated changes in plasmapotassium in guinea-pigs. British Journal of Pharmacology. 80(3):573-80, 1983 Nov.

Coats RA. The effects of adrenoceptor agonists and antagonists on plasma potassiumconcentration in anaesthetized guinea-pigs, rabbits and rats. British Journal of Pharmacology. 86(4).-827-36, 1985 Dec.

Colatsky TJ. Follmer CH. K+ channel blockers and activators in cardiac arrhythmias. Cardiovascular Drug Review. 7:199-209 1989.

Colatsky TJ. Follmer CH. Starmer CF. Channel specificity in antiarrhythmic drug action. Mechanism of potassium channel block and its role in suppressing and aggravating cardiac arrhythmias. Circulation. 82(6):2235-42, 1990 Dec.

Colquhoun D. Sigworth FJ. Fitting and statistical analysis of single-channel records. In Single-Channel Recording. B.Sakmann and E. Neher, editors. New York, Plenum Press. 191-263 1983.

Coraboeuf E. Deroubaix E. Escande D. Coulombe A. Comparative effects of three class I antiarrhythmic drugs on plateau and pacemaker currents of sheep cardiac Purkinje fibres. Cardiovascular Research. 22(6):375-84, 1988 Jun.

Cornish-Bowden A. Principles of enzyme kinetics, Butterworths 1976. Wang, X.P. 107

Cross PE. Arrowsmith JE. Thomas GN. Gwilt M. Burges RA. Higgins AJ. Selective class III antiarrhythmic agents. 1 Bis(arylalkyl)amines. Journal of Medicinal Chemistry. 33(4): 1151-5, 1990 Apr.

Cukierman S. Yellen G. Miller C. The K+ channel of sarcoplasmic reticulum. A new look at Cs+ block. Biophysical Journal. 48(3).-477-84, 1985 Sep.

Dimpfel W. Pierau FK. Haider SG. Electrophysiological studies on the effects of the neurotoxin apamin on cultured neurons. Advances In Cytopharmacology. 3:395-9, 1979.

Dukes ID. Cleemann L Morad M. Tedisamil blocks the transient and delayed rectifier K+ currents in mammalian cardiac and glial cells. Journal of Pharmacology & Experimental Therapeutics. 254(2):560-9, 1990 Aug.

Eisenberg H. Thermodynamics and the structure of biological macromolecules. European Journal of Biochemistry. 187(1):7-22, 1990 Jan 12.

Eisenman G. Latorre R. Miller C. Multi-ion conduction and selectivity in the high- conductance Ca2+-activated K+ channel from skeletal muscle. Biophysical Journal. 50(6): 1025-34, 1986 Dec.

Farley J. Rudy B. Multiple types of voltage-dependent Ca2+-activated K+ channels of large conductance in rat brain synaptosomal membranes. Biophysical Journal; 53(6); P 919-34; 1988 Jun. Wang, X.P. ios

Findlay IM. Dunne MJ. Petersen OH. High conductance K+ channel in pancreatic islet cells can be activated and inactivated by internal calcium. Journal of Membrane Biology 83:169-75 1985.

Fisher HF. A unifying model of the thermodynamics of formation of dehydrogenase-ligand complexes. Advances in Enzymology & Related Areas of Molecular Biology. 61:1-46, 1988.

Franciolini F. Calcium and voltage dependence of single Ca2+-activated K+ channels from cultured hippocampal neurons of rat. Biochim-Biophys-Acta; 943(3); P 419-27. 1988 Sep 1.

French RJ. Shoukimas JJ. Blockage of squid axon potassium conductance by internal Tetra-n-alkylammonium ions of various sizes. Biophysical Journal. 34:271-291. 1981.

Gallin EK. Calcium- and voltage-activated potassium channels in human

macrophages. Biophysical Journal. 46(6):821-5, 1984 Dec.

Gardos G. The function of calcium in the potassium permeability of human erythrocytes. Biochim Biophys. Acta 30:653-54. 1958.

Gitter AH. Beyenbach KW. Christine CW. Gross P. Minuth WW. Fr:omter E.

High-conductance K+ channel in apical membranes of principal cells cultured

from rabbit renal cortical collecting duct aniagen. Pflugers Archiv - European

Journal of Physiology. 408(3):282-90, 1987 Mar. Wang, X.P. 109

Goldstein SA. Miller C. Site-specific mutations in a minimal voltage-dependent K+ channel alter ion selectivity and open-channel block. Neuron. 7(3):403-8, 1991 Sep.

Golowasch J. Kirkwood A. Miller C. Allosteric effects of Mg2+ on the gating of Ca2+-activated K+ channels from mammalian skeletal muscle. Journal of Experimental Biology. 124:5-13, 1986 Sep.

Gorman AL. Thomas MV. Intracellular calcium accumulation during depolarization in a molluscan neurone. Journal of Physiology. 308:259-85, 1980a.

Gorman AL. Thomas MV. Potassium conductance and internal calcium accumulation in a molluscan neurone. Journal of Physiology. 308:287- 313, 1980b.

Grega DS. Macdonald RL. Activators of adenylate cyclase and cyclic AMP prolong calcium-dependent action potentials of mouse sensory neurons in culture by reducing a voltage-dependent potassium conductance. Journal of Neuroscience. 7(3):700-7, 1987 Mar.

Grover GJ. Sleph PG. Dzwonczyk S. Pharmacologic profile of cromakalim in the treatment of myocardial ischemia in isolated rat hearts and anesthetized dogs. Journal of Cardiovascular Pharmacology. 16(6):853-64, 1990 Dec.

GugginoSE. GugginoWB. Green N. SacktorB. Ca2+-activated K+channels in cultured medullary thick ascending limb cells. American Journal of Physiology. 252:C 128-37; 1987. Wang, X.P. no

Gwilt M. Arrowsmith JE. Blackburn KJ. Burges RA. Cross PE. Dalrympie HW. Higgins AJ. UK-68,798: a novel, potent and highly selective class III antiarrhythmic agent which blocks potassium channels in cardiac cells. Journal of Pharmacology & Experimental Therapeutics. 256(1).'318-24, 1991 Jan.

Habermann E. Apamin. Pharmacology & Therapeutics. 25(2).-255-70, 1984.

Habermann E. Neurotoxicity of apamin and MCD peptide upon central application. Naunyn-Schmiedebergs Archives of Pharmacology. 300(2):189-91, 1977 Nov 10.

Habermann E. Fischer K. Apamin, a centrally acting neurotoxic peptide: binding and actions. Advances In Cytopharmacology. 3:387-94, 1979.

Habermann E. Fischer K. Bee venom neurotoxin (apamin): iodine labeling and characterization of binding sites. European Journal of Biochemistry. 94(2):355-64, 1979 Mar.

Haworth RA. Goknur AB. Berkoff HA. Inhibition of ATP-sensitive potassium channels of adult rat heart cells by antiarrhythmic drugs. Circulation

Research. 65 (4): 1157-60, 1989 Oct.

Haylett DG. Jenkinson DH. Calcium-Activated Potassium channels. Chapter 3 in

Potassium Channels Edited by Nigel S. Cook, 1990 Wang, X.P. 111

Heyer CB. Lux HD. Control of the delayed outward potassium currents in bursting pace-maker neurones of the snail, Helix pomatia. Journal of Physiology. 262(2):349-82, 1976 Nov.

Hille B Potassium Channel in Myelinated Nerve: selectivity and permeability to small cations. Journal of General Physiology. 61:669-686, 1973

Hille B. Ionic channels of excitable membranes. Published by Sinauer Associates Inc., Sunderland, Mass. 1984.

Hille B. Local anesthetics: Hydrophilic and hydrophobic pathways for the drug-

receptor reaction. Journal of General Physiology. 69:497-515. 1977.

Hitzemann R. Thermodynamic aspects of drug-receptor interactions. Trends in Pharmacological Sciences. 9:408-411. 1988.

Hladky SB. Haydon DA. Ion transfer across lipid membranes in the presence of

gramicidin A. Biochim. Biophys. Acta 274:294-312. 1972.

Hondeghem LM. Snyders DJ. Class III antiarrhythmic agents have a lot of potential but a long way to go. Reduced effectiveness and dangers of reverse use dependence. Circulation. 81(2):686-90, 1990 Feb.

Howard PG. Abraham S. Courtice ID. Walker MJ. Further studies into the ECG

effects of tedisamil (KC 8857) in rats. Proceedings of the Western Pharmacology Society. 32:183, 1989. Wang, X.P. 112

Hugues M. Duval D. Schmid H. Kitabgi P. Lazdunski M. Vincent JP. Specific binding and pharmacological interactions of apamin, theneurotoxin from bee venom, with guinea pig colon. Life Sciences. 31 (5):437-43, 1982a Aug 2.

Hugues M. Romey G. Duval D. Vincent JP. Lazdunski M. Apamin as a selective blocker of the calcium-dependent potassium channelin neuroblastoma cells: voltage-clamp and biochemical characterization of the toxin receptor. Proceedings of the National Academy of Sciences of the United States of America. 79 (4): 1308-12, 1982b Feb.

Hugues M. Schmid H. Lazdunski M. Identification of a protein component of the Ca2+-dependent K+ channel byaffinity labelling with apamin. Biochemical & Biophysical Research Communications. 107(4): 1577-82, 1982 Aug 31.

Hugues M. Schmid H. Romey G. Duval D. Frelin C. Lazdunski M. The Ca2+- dependent slow K+ conductance in cultured rat muscle cells:characterization with apamin. Embo Journal. 1(9):1039-42, 1982.

Hunter M. Lopes AG. Boulpaep EL. Giebish GH. Single chennal recordings of single Ca2+-activited K+ channels in apical membrane of coetical collecting tubules. Proceedings of the National Acadamy of Sciences of the United States of America 81:4237-39 1984

Ikemoto Y. Ono K. Yoshida A. Akaike N. Delayed activation of large- conductance Ca2+-activated K+ channels in hippocampal neurons of the rat. Biophysical Journal; 56(1); P 207-12; 1989 Jul. Wang, X.P. 113

Jacobs ER. DeCoursey TE. Mechanisms of potassium channel block in rat alveolar epithelial cells. Journal of Pharmacolology and Experimental Therapeutics. 255:459-472 1990.

Kauzmann W. Some factors in the interpretation of protein denaturation. Advances in Protein Chemistry. 14:1-63 1959.

Kawai T. Watanabe M. Blockade of Ca2+-activated K+ conductance by apamin in rat sympathetic neurones. British Journal of Pharmacology. 87(1):225-32, 1986 Jan.

KurachiY. NakajimaT. Ito H. SugimotoT. AN-132, a new class I anti-arrhythmic agent, depresses the acetylcholine-induced K+ current in atrial myocytes. European Journal of Pharmacology. 165 (2-3) :319-22, 1989 Jun 20.

Laidler KJ. Peterman BF. Temperature effects in enzyme kinetics. Methods- Enzymol; 63; P 234-57; 1979.

Lancaster B. Nicoll RA. Properties of two calcium-activated hyperpolarizations in

rat hippocampal neurones. Journal of Physiology. 389:187-203, 1987 Aug.

Latorre R. Miller C. Conduction and selectivity in potassium channels. Journal of Membrane Biology. 71 (1 -2): 11 -30, 1983.

Latorre R. Oberhauser A. Labarca P. Alvarez O. Varieties of calcium-activated

potassium channels. Annual Review of Physiology. 51:385-99, 1989. Wang, X.P. 114

Lazdunski M. Fosset M. Hughes M. Mourre C. Romey G. Schmid-Antomarchi H. The apamin-sensitive Ca2+-dependent K+ channel molecular properties, differentiation and endogenous ligands in mammalian brain. Biochemical Society Symposia. 50:31-42, 1985.

Lotorre R. Cecchi X. Calcium and long-lived closed states in a Ca2+ - activated K+ channel from skeletal muscle. Biophysical Journal 59:(2 pt 2) 213a 1991 Feb.

Maas AJ. The effects of apamin on responses evoked by field stimulation inguinea- pig taenia caeci. European Journal of Pharmacology. 73(1):1-9, 1981 Jul 17.

Maas AJ. Den Hertog A. Ras R. Van den Akker J. The action of apamin on guinea-pig taenia caeci. European Journal of Pharmacology. 67(2-3) :265- 74, 1980 Oct 17.

MacDermott AB. Weight FF. Action potential repolarization may involve a

transient, Ca2+-sensitive outward current in a vertebrate neurone. Nature.

300(5888): 185-8, 1982 Nov 11.

MacKinnon R. Miller C. Functional modification of a Ca2+-activated K+ channel by trimethyloxonium. Biochemistry. 28(20) :8087-92, 1989 Oct 3 Wang, X.P. us

MacKinnon R. Miller C. Mechanism of charybdotoxin block of the high- conductance, Ca2+-activated K+ channel. Journal of General Physiology. 91(3):335-49, 1988 Mar.

MacKinnon R. Miller C. Mutant potassium channels with altered binding of charybdotoxin, a pore-blocking peptide inhibitor. Science. 245(4924):1382-5, 1989 Sep 22.

MacKinnon R. Smith C. Miller C. Charybdotoxin block of single Ca2+-activated K+ channels. Effects of channel gating, voltage, and ionic strength. Journal of General Physiology. 91 (3):317-33, 1988 Mar.

MacKinnon R. Latorre R. Miller C. Role of surface electrostatics in the operation of

a high-conductance Ca2+-activated K+ channel. Biochemistry. 28(20):8092-9, 1989 Oct 3.

Magleby KL Pallotta BS. Burst kinetics of single calcium-activated potassium channels in cultured rat muscle. Journal of Physiology; 344; P 605-23 1983 Nov.

Magleby KL. Pallotta BS. Calcium dependence of open and shut interval

distributions from calcium-activated potassium channels in cultured rat

muscle. Journal of Physiology; 344; P 585-604; 1983 Nov.

Marty A. Ca2+-dependent K+ channels with large unitary conductance in

chromaffin cell membranes. Nature. 291(5815):497-500, 1981Jun11. Wang, X.P. 116

Maruyama Y. Petersen OH. Cholecystokinin activation of single-channel currents is mediated by internal messenger in pancreatic acinar cells. Nature. 300(5887).-61-3, 1982 Nov 4.

Massefski W Jr. Redfield AG. Hare DR. Miller C. Molecular structure of charybdotoxin, a pore-directed inhibitor of potassium ion channels [retracted by Massefski W Jr, Redfield AG, Hare DR, Miller C. In: Science 1991 May 3;252(5006):631] Science. 249(4968) :521-4, 1990 Aug 3.

McCarthy MP. Stroud RM. Changes in conformation upon agonist binding and nonequivalent labeling of the membrane-spanning regions of the nicotinic acetylcholine receptor subunits. Journal of Biological Chemistry. 264: 10911-10916 1989b.

McCarthy MP. Stroud RM. Conformational states of the nicotinic acetylcholine receptor from Torpedo californica induced by the binding of agonists, antagonists and local anesthetics: Equilibrium measurements using tritium- hydrogen exchange. Biochemistry 28:40-48 1989a.

McLarnon JG. Block by a putative antiarrhythmic agent of a calcium-dependent potassium channel in cultured hippocampal neurons. Neuroscience Letters. 112(2-3):210-5, 1990 May 4.

McLarnon JG. Block by a putative antiarrhythmic agent of a calcium-dependent

potassium channel in cultured hippocampal neurons. Neuroscience Letters.

112:210-215 1990. Wang, X.P. 117

McLarnon JG. Pennefather P. Quastel DMJ. Mechanism of nicotinic channel blockade by anesthetics. In Molecular and Cellular Mechanisms of Anesthetics. S. H. Roth and K. W. Miller, editors, Plenum Publishing, New York. 155-164 1986.

McLarnon JG. Quastel DM. Postsynaptic effects of magnesium and calcium at the mouse neuromuscularjunction. Journal of Neurosciences; 3(8); P 1626- 33; 1983 Aug.

McLarnon JG. Quastel DM. Thermodynamic parameters of end-plate channel blockade. Journal of Neurosciences; 4(4); P 939-44; 1984 Apr.

Meech RW. Calcium dependent potassium activation in nervous tissues. Annual Review of Biophysics and Bioengineering 7:1-18 1978

Meech RW. Intracellular calcium injection causes increased potassium

conductance in Aplsia nerve cells. Comp. Biochem. Physiol. 42A:493

1972

Meech RW. Strumwasser F. Intracellular calcium injection activates potassium

conductance in Aplsia nerve cells. Fed. Proc. 29:834a 1970

Methfessel C. Boheim G. The gating of single calcium-dependent potassium channels is described by an activation/blockade mechanism. Biophys- Struct-Mech; 9(1); P 35-60; 1982. Wang, X.P. us

Miklavc A. Kocjan D. Mavri J. Koller J. Had:zi D. On the fundamental difference in the thermodynamics of agonist and antagonist interactions with beta- adrenergic receptors and the mechanism of entropy-driven binding. Biochemical Pharmacology. 40(4):663-9, 1990 Aug 15.

Miller C. Competition for block of a Ca2+-activated K2+ channel by charybdotoxin and tetraethylammonium. Neuron. 1(10):1003-6, 1988 Dec.

Miller C. Diffusion-controlled binding of a peptide neurotoxin to its K+ channel receptor. Biochemistry. 29(22):5320-5, 1990 Jun 5.

Miller C. Diffusion-controlled binding of a peptide neurotoxin to its K+ channel receptor. Biochemistry 29:5320-5325 1990.

Miller C. Genetic manipulation of ion channels: a new approach to structure and mechanism. Neuron. 2(3): 1195-205, 1989 Mar.

Miller C. Trapping single ions inside single ion channels. Biophysical Journal. 52(1): 123-6, 1987 Jul.

Miller C. Latorre R. Reisin I. Coupling of voltage-dependent gating and Ba2+ block in the high-conductance, Ca2+-activated K+ channel. Journal of General Physiology. 90(3):427-49, 1987 Sep.

Miller C. Moczydlowski E. Latorre R. Phillips M. Charybdotoxin, a protein

inhibitor of single Ca2+-activated K+ channels from mammalian skeletal

muscle. Nature. 313(6000):316-8, 1985 Jan 24-30. Wang, X.P. 119

Miller C. Stahl N. Barrol M. A thermodynamic analysis of monovalent cation permeation through a K+-selective ion channel. Neuron. 1(2):159-64, 1988 Apr

Minneman KP. Weiland GA. Molinoff PB. A comparison of the beta-adrenergic receptor of the turkey erythrocyte with mammalian betal and beta2 receptors. Molecular Pharmacology. 17:1-7 1980.

Moczydlowski E. Alvarez O. Vergara C. Latorre R. Effect of phospholipid surface charge on the conductance and gating of a Ca2+-activated K+ channel in planar lipid bilayers. Journal of Membrane Biology. 83(3):273- 82, 1985.

Moczydlowski E. Latorre R. Gating kinetics of Ca2+-activated K+ channels from rat muscle incorporated into planar lipid bilayers. Evidence for two voltage- dependent Ca2+ binding reactions. Journal of General Physiology. 82(4):511-42, 1983 Oct.

Molinoff PB. Wolfe BB. Weiland GA Quantitative analysis of drug-receptor

interactions: II. Determination of the properties of receptor subtypes. Life

Sciences; 29(5); P 427-43; 1981 Aug 3.

Moolenaar WH. Spector I. Ionic currents in cultured mouse neuroblastoma cells

under voltage-clamp conditions. Journal of Physiology. 278:265-86,

1978 May. Wang, X.P. 120

Mourre C. Hugues M. Lazdunski M. Quantitative autoradiographic mapping in rat brain of the receptor of apamin, a polypeptide toxin specific for one class of Ca2+-dependent K+ channels. Brain Research. 382(2).-239-49, 1986 Sep 24.

Mourre C. Schmid-Antomarchi H. Hugues M. Lazdunski M. Autoradiographic localization of apamin-sensitive Ca2+-dependent K+ channels in rat brain. European Journal of Pharmacology. 100(1): 135-6, 1984 Apr 13.

Neher E. The charge carried by single-channel currents of rat-cultured muscle cells in the presence of local anaesthetics. Journal of Physiology 339:663-678. 1983.

Neher E. Steinbach JH. Local anaesthetics transiently block currents through single acetylcholine-receptor channels. Journal of Physiology 277:153-176 1978.

Neyton J. Miller C. 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(5):569-86, 1988 Nov.

Oberhauser A. Alvarez O. Latorre R. Activation by divalent cations of a Ca2+-

activated K+ channel from skeletal muscle membrane. Journal of General

Physiology. 92(1):67-86, 1988 Jul. Wang, X.P. 121

Pallotta BS N-bromoacetamide removes a calcium-dependent component of channel opening from calcium-activated potassium channels in rat skeletal

muscle. Journal of General Physiology; 86(5); P 601 -11; 1985 Nov.

Pallotta BS. Magleby KL Barrett JN. Single channel recordings of Ca2+- activated K+ currents in rat muscle cell culture. Nature. 293(5832):471-4, 1981 Oct 8.

Pearson GT. Singh J. Petersen OH. The effects of L-alanine and acetylcholine on membrane potential, 45Ca2+ and 8^Rb+ efflux and amylase secretion in the isolated mouse pancreas. Quarterly Journal of Experimental Physiology. 69(3):531-40, 1984 Jul.

Petersen OH Calcium-activated potassium channels and fluid secretion by

exocrine glands. American Journal of Physiololgy; 251(1 Pt 1); P G1-13;

1986 Jul.

Petersen OH. Maruyama Y Calcium-activated potassium channels and their role in secretion. Nature; 1984 Feb 23-29; 307(5953); P 693-6

Privalov PL. Thermodynamic problems of protein structure. Annual Review of

Biophysics & Biophysical Chemistry. 18:47-69, 1989.

Raffa RB. Porreca F. Thermodynamic analysis of the drug-receptor interaction.

Life Sciences. 44(4):245-58, 1989. Wang, X.P. 122

Reichardt B. Konzen G. Hauswirth O. Pirmenol, a new antiarrhythmic drug with potassium- and sodium-channel blocking activity; a voltage-clamp study in rabbit Purkinje fibres. Naunyn-Schmiedebergs Archives of Pharmacology. 341(5):462-71, 1990 May.

Reinhart PH; Chung S; Levitan IB A family of calcium-dependent potassium channels from rat brain. Neuron; 2(1); P 1031-41; 1989 Jan.

Renaud JF. Desnuelle C. Schmid-Antomarchi H. Hugues M. Serratrice G. Lazdunski M. Expression of apamin receptor in muscles of patients with myotonicmuscular dystrophy. Nature. 319(6055).-678-80, 1986 Feb 20- 26.

Rials SJ. Friehling TD. Marinchak RA. Kowey PR. Potassium channels in cardiac arrhythmias: focus on antiarrhythmic drug action. Progress In Clinical & Biological Research. 334:111-21, 1990.

Romey G. Lazdunski M. Lipid-soluble toxins thought to be specific for Na+ channels block Ca2+ channels in neuronal cells. Nature. 297(5861):79-8, 1982 May 6.

Roux B. Sauv:e R. A general solution to the time interval omission problem applied

tosingle channel analysis. Biophysical Journal. 48(1):149-58, 1985 Jul.

Rudy B Diversity and ubiquity of potassium channels. Neuroscience; 25(3); P 729-49; 1988 Jun. Wang, X.P. 123

Sakmann B. Neher E. Patch clamp techniques for studying ionic channels in excitable membranes. Annual Review of Physiology. 46:455-72, 1984.

Sakmann B. Neher E. editors Single-Channel Recording. New York, Plenum Press. 1983

Sandberg BE. Solid phase synthesis of 13-lysine-apamin, 14-lysine-apamin and thecorresponding guanidinated derivatives. International Journal of Peptide & Protein Research. 13(3).-327-33, 1979 Mar.

Sanguinetti MC. Jurkiewicz NK. Two components of cardiac delayed rectifier K+ current. Differential sensitivity to block by class III antiarrhythmic agents. Journal of General Physiology. 96(1):195-215, 1990 Jul.

Schmid-Antomarchi H. Hugues M. Lazdunski M. Properties of the apamin- sensitive Ca2+-activated K+ channel in PC12 pheochromocytoma cells which hyper-produce the apamin receptor. Journal of Biological Chemistry. 261(19):8633-7, 1986 Jul 5.

Schmid-Antomarchi H. Hugues M. Norman R. Ellory C. Borsotto M. Lazdunski M. Molecular properties of the apamin-binding component of the Ca2+- dependent K+ channel. Radiation-inactivation, affinity labelling and solubilization. European Journal of Biochemistry. 142(1):1-6, 1984 Jul 2. Wang, X.P. 124

Schmid-Antomarchi H. Renaud JF. Romey G. Hugues M. Schmid A. Lazdunski M. The all-or-none role of innervation in expression of apamin receptor and of apamin-sensitive Ca2+-activated K+ channel in mammalian skeletal muscle. Proceedings of the National Academy of Sciences of the United States of America. 82(7):2188-91, 1985 Apr.

Schreiner W. Kramer M. Barkakati N. Nonlinear least-squares fitting. PC Tech J. 170-181, 1985 May.

Segal M, Potassium currents activated in hippocampal neurons by serotonin are mediated by a change in intracellular calcium concentrations European Journal of Pharmacology. 181; P 299-301 1990.

Sigworth FJ, An example of analysis. In Single-Channel Recording. B.Sakmann and E. Neher, editors. New York, Plenum Press. 301-321. 1983.

Singer JJ; Walsh JV Jr Characterization of calcium-activated potassium channels

in single smooth muscle cells using the patch-clamp technique. Pflugers

Archiv-European Journal of Physiology. 408(2); P 98-111 1987 Feb.

Singh J. Petersen OH. Adrenergic nervous control of cAMP-mediated amylase

secretion in the rat pancreas. American Journal of Physiology. 246(5 Pt

1):G563-73, 1984 May.

Smith C. Phillips M. Miller C. Purification of charybdotoxin, a specific inhibitor of

the high-conductance Ca2+-activated K+ channel. Journal of Biological

Chemistry. 261(31):14607-13, 1986 Nov 5. Wang, X.P. 125

Stanley EF; Ehrenstein G A model for exocytosis based on the opening of calcium-activated potassium channels in vesicles. Life Sciences; 37(21); P 1985-95; 1985 Nov 25.

Starmer CF. Yeh JZ. Tanguy J. A quantitative description of QX222 blockade of sodium channels in squid axons. Biophysical Journal. 49:913-920 1986.

Storm JF. Potassium currents in hippocampal pyramidal cells. Progress In Brain Research. 83:161-87, 1990.

Strong PN. Evans WH. Receptor-mediated endocytosis of apamin by liver cells. European Journal of Biochemistry. 163(2):267-73, 1987 Mar 2.

Testa B. Jenner P. Kilpatrick GJ. Tayar N. Van de Waterbeemd H. Marsden CD. Do thermodynamic studies provide information on both the binding to and the activation of dopaminergic and other receptors?. Biochemical Pharmacology. 36(23):4041-6, 1987 Dec 1

Triggle DJ. Drugs acting on ion channels and membranes. Chaper 14 p1027-1099, in Comprehensive Biomedical Chemistry the fourth edition 1990

Vaidivia HH. Smith JS. Martin BM. Coronado R. Possani LD. Charybdotoxin and noxiustoxin, two homologous peptide inhibitors of the Ca2+-activated K+ channel. Febs Letters. 226(2).-280-4, 1988 Jan 4. Wang, X.P. 126

Vergara C. Latorre R. Kinetics of Ca2+-activated K+ channels from rabbit muscle incorporated into planar bilayers. Evidence for a Ca2+ and Ba2+ blockade. Journal of General Physiology. 82(4):543-68, 1983 Oct.

Vergara C. Moczydlowski E. Latorre R. Conduction, blockade, and gating in a Ca2+ -activated K+ channel incorporated into planar bilayers. Biophysical Journal 45:73-76 1984.

Villarroel A. Alvarez O. Oberhauser A. Latorre R. Probing a Ca2+-activated K+ channel with quaternary ammonium ions. Pflugers Archiv - European Journal of Physiology. 413 (2): 118-26, 1988 Dec.

Weiland GA. Minneman KP. Molinoff PB Thermodynamics of agonist and antagonist interactions with mammalian beta-adrenergic receptors. Molecular Pharmacology; 18(3); P 341-7; 1980 Nov.

Weiland GA. Molinoff PB Quantitative analysis of drug-receptor interactions: I. Determination of kinetic and equilibrium properties. Life Sciences; 29(4); P

313-30; 1981 Jul 27.

Wolff D. Cecchi X. Spalvins A. Canessa M. Charybdotoxin blocks with high affinity the Ca2+-activated K2+ channel of Hb A and Hb S red cells: individual differences in the number of channels. Journal of Membrane Biology. 106(3):243-52, 1988 Dec. Wang, X.P. 127

Wong BS Quinidine blockade of calcium-activated potassium channels in dissociated gastric smooth muscle cells. Pflugers Archiv-European Journal of Physiology. 414(4); P 416-22; 1989 Aug.

Wong BS. Adler M. Tetraethylammonium blockade of calcium-activated potassium channels in clonal anterior pituitary cells. Pflugers Archiv- European Journal of Physiology. 407(3); P 279-84; 1986 Sep.

Wong BS. Lecar H. Adler M. Single calcium-dependent potassium channels in clonal anterior pituitary cells. Biophysical Journal. 39(3):313-7, 1982 Sep.

Yanagisawa T. Hashimoto H. Taira N. The negative inotropic effect of nicorandil is independent of cyclic GMP changes: a comparison with pinacidil and cromakalim in canine atrial muscle. British Journal of Pharmacology. 95(2):393-8, 1988 Oct.

Yellen G. Ionic permeation and blockade in Ca2+-activated K+ channels of bovine chromaffin cells. Journal of General Physiology. 84(2): 157-86, 1984 Aug.

Yellen G. Permeation in potassium channels: implications for channel structure. Annual Review of Biophysics & Biophysical Chemistry. 16:227-46, 1987.

Yellen G. Relief of Na+ block of Ca2+-activated K+ channels by external cations. Journal of General Physiology. 84(2): 187-99, 1984 Aug. Wang, X.P. 128

Zhang L. Krnjevi:c K. Apamin depresses selectively the after-hyperpolarization of cat spinal motoneurons. Neuroscience Letters. 74(1):58-62, 1987 Feb 10.