Subunit-specific determinants of function and pharmacology of nicotinic acetylcholine receptors


Material Information

Subunit-specific determinants of function and pharmacology of nicotinic acetylcholine receptors
Physical Description:
Francis, Michael Marvin, 1970-
Publication Date:


Spatial Coverage:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 29450555
oclc - 51005237
System ID:

Table of Contents
    Title Page
        Page i
        Page ii
        Page iii
        Page iv
    Table of Contents
        Page v
        Page vi
        Page vii
        Page viii
    Chapter 1. Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
    Chapter 2. Materials and methods
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
    Chapter 3. Results
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
    Chapter 4. Discussion
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
    Chapter 5. Summary and conclusions
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
    Biographical sketch
        Page 161
        Page 162
        Page 163
        Page 164
Full Text







I look back on the last five years and wonder why it took me so long to accomplish the things which I have accomplished. Upon reflection, I realize how much longer it would have taken without the patience, understanding and guidance of family, friends and instructors. I would like to thank my father, Miles, mother, Sandy, and sister, Kathy, for putting up with five years of having to take a backseat to my pursuit of this degree. It has been a real benefit to me to know that I have a wonderful family to fall back on whenever I have needed to. It has been a great privilege to have the unquestioning support of my family especially in light of the fact that they have no clue as to exactly what I am studying or why it is taking so long.

I would also like to thank the many friends, old and new, near and far, who have helped encourage, distract, intoxicate and otherwise lend a vitality to my life which really keeps graduate school in perspective. In particular, I would like to thank the many friends I have made in Gainesville during the course of graduate school for not allowing me to waste the few free hours I have on worrying about what I could be doing in the lab.

I would like to thank the members of my graduate committee past and present, Dr. Peter Anderson, Dr. Jeff Harrison, Dr. Ben Horenstein, Dr. Mike King, Dr. Bob Lenox, and Dr. Janet Zengel, for their time and effort in helping me to achieve my goals as a scientist and as a person. I would like to thank the current members of the Papke laboratory (Nik, Gill, Anatolii, Julia and Chris) as well as past members (Clare, Hugo, Amy, Rick, Becky and Wayne) for their encouragement and intellectual as well as technical support.

Finally, I would like to express my deep gratitude to my mentor, Dr. Roger Papke, foremost for his guidance throughout my course of study and also for his patience in

understanding the times when I needed a break from my course of study. The free exchange of ideas with him has been indispensable in transforming a psychology major with a vague notion of wanting to study synaptic transmission into a successful neuroscientist.





ABSTRACT ...................................................................... vii


1. INTRODUCTION ............................................................... 1

Nicotinic Receptor Gene Family ................................................ 3
General Features of Muscle-type Receptors.................................... 3
Cloning of the Neuronal nAChRs ............................................. 4
General Features of Neuronal nAChRs........................................ 5
Function of Nicotinic Acetylcholine Receptors.................................... 6
Pharmacology and Electrophysiology in the Study of nAChR .................. 6
Neuromuscular Junction nAChR .............................................12
In vivo Subunit Composition and Distribution of Neuronal nAChRs............ 15
Functional Roles of Neuronal nAChRs....................................... 19
Sensitivity to Noncompetitive Inhibitors........................................ 28
Structure of nAChRs...........................................................33
General Structural Features ..................................................34
N-Terminal Domain and Agonist Binding Site ............................... 37
Transmembrane Domains 1-4 (TM 1-4) and Cytoplasmic Loop................. 39
Relating Structure to Function for Neuronal nAChRs......................... 46

2. MATERIALS AND METHODS................................................ 48

Chemicals and Synthesis ...................................................... 48
Production of Chimeras and Sequencing........................................ 49
Preparation of RNAs and Oocyte Expression ................................... 49
Electrophysiology ..............................................................50
Two-Electrode Voltage Clamp ............................................... 50
Cut-Open Vaseline Gap Voltage Clamp ...................................... 51
Experimental Protocols and Data Analysis.................................... 52

3. RESULTS..................................................................... 56

Structural Determinants of Sensitivity to the TMP Family of NCIs................ 56
Muscle Delta Subunit Effects ................................................ 56
Neuronal Beta Subunit Effects............................................... 73
Mechanism of Inhibition of nAChRs by bis-TMP- 10............................ 87
Inhibition by Bis-TMP- 10 Is Independent of Voltage ......................... 87
Inhibition of nAChRs by QX-314 ........................................... 97


Requirements for Long-Term Inhibition by Bis-TMP- 10..................... 105
Inhibition by an Amphipathic Analogue of Bis-TMP-10 ....................... 116
Inhibition by ATMP- 10 Independent of Activation by Agonist ................ 119
Voltage-Dependence of Inhibition by ATMP-10............................. 121
TMP Protects a1I31(I34TM2)y5 Receptors from Inhibition by ATMP-10......... 126

4. DISCUSSION................................................................127

The Mechanism of Action of Mecamylamnine Is Distinct from that of Bis-TMP-10 .... 128 TM2 Determines Kinetics of Long-Term Inhibition by Bis-TMP-l0.............. 129
Significance of Voltage -Independent Inhibition................................. 131
Inhibition by Bis-TMP-l0 Is Distinct from Inhibition by QX-314 ............... 133
Bis-TMP-10 and QX-314 Do Not Compete for the Same Site................... 134
Significance of Compound Length Requirement for Long-Term Inhibition ....... 137

5. SUMMARY AND CONCLUSIONS........................................... 142

REFERENCES................................................................... 147

BIOGRAPHICAL SKETCH ..................................................... 161


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SUBUNIT-SPECIFIC DETERMINANTS OF FUNCTION AND PHARMACOLOGY OF NICOTINIC ACETYLCHOLINE RECEPTORS


Michael M. Francis

December, 1998

Chairman: Roger L. Papke
Major Department: Neuroscience

Some noncompetitive inhibitors (e.g., ganglionic blockers) exhibit selectivity for the inhibition of neuronal nicotinic acetylcholine receptors (nAChRs). The main goal of the present study is to characterize the mechanism of selective long-term inhibition of neuronal and muscle-neuronal chimeric nAChRs by bis-TMP-10 (bis (2,2,6,6-tetramethyl-4piperidinyl) sebacate or BTMPS), a bifunctional form of the potent ganglionic blocker tetramethylpiperidine. Long-term inhibition of neuronal nAChRs by bis-TMP- 10 has been previously demonstrated to arise, at least in part, from the binding of the bis- compound to neuronal beta subunits. In this study, long-term inhibition is demonstrated to be dependent upon the presence of sequence element(s) within the pore-lining second transmembrane domain (TM2) of either neuronal beta subunits or muscle delta subunits; however, for either class of subunit, because the onset of inhibition does not appear to be strongly voltage-dependent, the inhibitor binding site itself does not appear to be contained within the segment of the channel pore influenced by the membrane electric field. In the case of the neuronal beta subunits, long-term inhibition is also not affected by preapplication of the


open-channel blocker QX-314. Furthermore, we demonstrate a compound length requirement for long-term inhibition which would be consistent with binding to multiple sites contributed by separate subunits (either beta or delta) located on the extracellular portion of the receptor. Our results may imply that bis-TMP-10 interacts with an activation-sensitive element, the availability of which may be regulated by sequence in the TM2 domain. It is interesting to note that the mechanism of inhibition by bis-TMP-10 appears to be distinct from that of other ganglionic blockers such as mecamylamine. Knowledge of the mechanism underlying the basis for inhibition by pure antagonists may be useful for consideration of observations of mixed agonist/antagonist properties of certain experimental therapeutics for nicotinic receptors. Furthermore, if bis-TMP-10 is binding to an activation-sensitive element distinct from TM2, it may be possible to use sensitivity to inhibition by this compound to investigate the structural changes which take place with channel gating.



The idea that osmosis and secretion across a biological membrane could be regulated by pores in the membrane dates back to the work of Ernst Brucke in the mid- 19th century (Brucke, 1843). Only much more recently has the critical role of membrane pores in nervous system function come to be recognized. With the almost simultaneous description of the ionic basis of action potential generation and propagation by Alan Hodgkin and Andrew Huxley (1952) and the observation of quantal changes in electrical potential (called miniature endplate potentials or MEPPs) during voltage recordings at the endplate of the neuromuscular junction by Paul Fatt and Bernard Katz (195 1), the electrochen-fical nature of communication between cells of the nervous system (neurons) began to emerge. However, it was only much later still that changes in membrane permeability to ions could be definitively linked to the opening and closing of membrane pores or ion channels as they would come to be called. Katz and Miledi (1972) were able to demonstrate that the electrical noise in their recordings increased with application of the endogenous activator of muscle fibers, acetylcholine, to the muscle fiber. Via noise analysis, these investigators were able to infer that this noise increase could be correlated with opening of individual ion channels in the post-synaptic membrane. Their analysis even went so far as to predict an estimate for average open time (length of time a channel is conducting before closing) and conductance (capacity of a channel to allow ions to permeate) of individual channels. The subsequent refinement of techniques such as the voltage-clamp (Anderson and Stevens, 1973) together with the development and application of improved techniques such as the patch clamp (Neher and Sakmann, 1976) allowed direct measurement of the opening and closing of ion channels. More recent studies employing these improvements have, in



general, confirmed much of the theory suggested by the work of earlier electrophysiologists. Furthermore, the early studies by investigators such as Hodgkin, Huxley and Katz have provided the framework for consideration and interpretation of research even today.

The process of neuronal communication is now commonly referred to as synaptic

transmission and, in general, consists of the release and subsequent diffusion of a chemical messenger or neurotransmitter across junctions between neurons called synapses. Upon reaching the post-synaptic neuron, neurotransmitter binds to post-synaptic sites (receptors) coupled to closed pores in the membrane. Binding of the neurotransmitter to these receptors triggers a change in the structure of the pore so that ions can flow freely through the ligand-gated ion channel, so called because the binding of a ligand opens the ion channel pore. The resulting redistribution of charged ions causes a change in the electrical potential across the membrane and hence, elicits a response in the post-synaptic cell. This response can be either inhibitory or excitatory depending upon the particular type of receptors present on the post-synaptic cell. If an excitatory response reaches a certain threshold, it will cause an action potential to be generated and propagated by the activity of another class of ion channels, voltage-gated sodium channels (a change in the distribution of charge opens the ion channel pore). When the action potential reaches the presynaptic terminal, calcium enters via the activation of voltage-gated calcium channels. This calcium influx initiates the events involved in the release of packets containing neurotransrriitter (synaptic vesicles) and, for excitatory neurotransmitters, the process repeats itself. Although the efficacy of this process can be modulated on a cellular and even subcellular level, this general form of electrochemical coupling underlies all rapid transmission in the nervous system.

Since the time of these early studies, the ongoing development of biochemical and

molecular biological techniques has provided for the identification of a great diversity of ion channels in the nervous system and permitted the study of homogeneous populations of ion


channel subtypes in isolated systems which allow more detailed analyses of the relationship between ion channel structure and function. The particular subtype of ion channel found at the neuromuscular junction, the nicotinic acetylcholine receptor (nAChR), remains a prototype system for these studies. Additionally, cloning of the neuromuscular junction nAChR has permitted the identification of a great diversity of nAChR subtypes in the central and peripheral nervous system.

Nicotinic Receptor Gene Family

The genes encoding nicotinic acetylcholine receptors are members of a superfamily of genes coding for ligand-gated ion channels which share considerable sequence homology and seem to have roughly equivalent membrane topographies. Other receptors encoded for by members of the gene family include GABA (gamma-amino butyric acid), glycine and 5HT3 (serotonin) receptors. All of the proteins encoded by members of this gene family share a characteristic 13 residue loop between disulfide-linked cysteine residues in the Nterminal domain (Kao and Karlin, 1986). Additionally, although each of these receptor types differs in selectivity for agonist and relative ionic permeabilities, they share certain essential functional properties. Specifically, agonist binds to distinct sites on the receptor complex causing a conformational change in the protein which allows ions (principally sodium, calcium and potassium in the case of nAChR and 5HT3 receptors or chloride in the case of GABA and glycine receptors) to flow down electrical and chemical gradients in relative proportions specific to each receptor type.

General Features of Muscle-type Receptors

Muscle-type nAChRs are the best characterized of the ligand-gated ion channels and as such are a model system for the study of structure-function relationships in the gene family. Muscle-type nAChRs are highly concentrated in the junctional folds at the endplate of the neuromuscular junction where they serve to transmit neuronal impulses to the muscle fiber.


Much of the initial characterization of the structural features of muscle-type nAChR can be attributed to the ready availability of large quantities of a homologue of muscle-type nAChR in the electric organs of the Torpedo ray. From biochemical experiments on Torpedo nAChR, it was possible to deduce a putative membrane topology and structural organization which has proven to be, for the most part, conserved in mammalian muscletype nAChR.

Muscle-type nAChRs are pentameric complexes, consisting of four distinct protein subunits (a1131y8) with protein molecular weights of about 50,000 (alpha 1), 53,700 (beta 1), 56,300 (gamma), and 57,600 (delta) in the ratio of 2: 1:1:1 (Conti-Tronconi et al., 1982). Two molecules of acetylcholine bind to sites believed to be located at the interface of the alpha subunits with the delta and gamma subunits (Blount and Merlie, 1989; Pedersen and Cohen, 1990; Sine and Claudio, 1991) to activate the receptor. It has now become apparent that the function of muscle-type nAChRs is developmentally regulated via the substitution of the epsilon subunit for the gamma subunit in the adult form of the receptor (Mishina et al., 1986).

Cloning of the Neuronal nAChRs

N-terminal sequencing of the muscle nAChR alphal subunit (Raftery et al., 1980) and production of degenerate oligonucleotides allowed for the identification of a cDNA encoding the muscle alpha subunit (Noda et al., 1982) and subsequent cloning of cDNAs encoding the muscle beta, gamma and delta subunits (Noda et al., 1983a, 1983b). Subsequently, a neuronal homologue of alpha 1 was cloned via low stringency hybridization screening of a PC 12 cell cDNA library with an oligonucleotide designed from the muscle alphal sequence (Boulter et al., 1986). At about the same time, another group working independently cloned a second neuronal alphal homologue from a chick brain cDNA library (Nef et al., 1988). Based on the conservation of a set of vicinal cysteines additional to the cysteine pair characteristic to all members of the gene family and shown to


be important for formation of the agonist binding sites (Kao and Karlin, 1986), these neuronal homologues were eventually designated as alpha3 and alpha2 respectively. Subsequent identification of additional neuronal homologues of the muscle alpha subunit clone from both rat and chick proceeded rather quickly after the identification of alpha2 and alpha3 (Boulter et al., 1990; Couturier et al., 1990a, 1990b; Deneris et al., 1988; Goldman et al., 1987; Schoepfer et al., 1988; Seguela et al., 1993; Wada et al., 1988). All neuronal subunits which share the conserved vicinal cysteine residues in the N-terminal domain have been designated as alpha subunits while the first non-alpha subunit was designated beta 2 based on it's ability to substitute for the betal subunit of muscle nAChR. Subsequently, certain other non-alpha subunits have been designated as neuronal beta subunits because of their ability to form functional receptors when expressed in pairwise combination with neuronal alphas.

General Features of Neuronal nAChRs

Eight neuronal nicotinic receptor subunits (alphas2-9) and three beta (betas2-4) subunits have been cloned to date (for review see (Lindstrom, 1996)). The receptors formed from these subunits fall into two major groups. Based on heterologous expression studies of alphas2-4, it has been demonstrated that these alpha subunits require coexpression with either beta2 or beta4 for function and form heteromeric receptors consisting of nonidentical subunits. Data from studies of heterologously expressed receptors indicate that both the alpha and beta subunits influence activation and inhibition characteristics of neuronal nAChRs (Luetje and Patrick, 1991). Based on these data and by analogy with the situation for muscle-type receptors, it is hypothesized that the agonist binding sites of neuronal nAChR lie at the interface of the alpha subunits with the beta subunits. The alpha subunit appears to function as more of a structural or modulatory subunit although functional effects of this subunit have recently been described (Wang et al., 1996). While the alpha5 protein does contain the conserved vicinal cysteine residues characteristic of alpha subunits,


it lacks other conserved residues in the N-terminal domain which are believed to contribute to the ACh binding site (see below). Alpha6 appears to require coexpression with the beta4 subunit in particular for function (Gerzanich et al., 1996). A functional role for the beta3 subunit in heterologous expression systems has not been described thus far although recent evidence suggests that it is incorporated into heteromeric receptors in the brain (Forsayeth and Kobrin, 1996).

The second major category of neuronal nAChR subtypes is defined by subunits which have the ability to function as homomeric receptors (receptors formed entirely from identical subunits) and is comprised of alphas7-9. Alpha8 has only been found in chick thus far where it most likely functions both as a homomer and as a heteromer with alpha7 (Gerzanich et al., 1993). In addition to possessing unique and heterogeneous functional properties (see below), these receptors exhibit unique pharmacological characteristics.

Function of Nicotinic Acetylcholine Receptors

Pharmacology and Electrophysiology in the Study of nAChRs

Before the development of molecular biology, ion channels were generally categorized according to their pharmacology. For example, although the endogenous agonist for the nicotinic acetylcholine receptor is acetylcholine, the receptor can be distinguished pharmacologically on the basis of high affinity binding of the drug nicotine and other agonists (e.g., cytisine). This property distinguishes the nicotinic receptor from another classification of acetylcholine receptors in the nervous system which bind muscarine with relatively high affinity, the muscarinic receptors. Muscarinic acetylcholine receptors are not directly coupled to ion channels but instead serve to, among other things, modulate neuronal excitability via initiation of second messenger cascades through the action of a class of proteins known as GTP-binding proteins or simply G-proteins (Hille, 1992). This class of acetylcholine receptors will not be discussed further here.


Although the identification of new ion channel subtypes currently relies, for the most part, on molecular biological rather than pharmacological techniques, pharmacology still provides a strong tool to identify the distribution and study functional relationships of particular ion channel subtypes. Pharmacological agents are generally described as either agonists or antagonists. Agonists bind to the receptor at a particular site and activate the receptor. For example, the neurotransmitter acetylcholine is an agonist of the nicotinic acetylcholine receptor. In contrast, antagonists act to inhibit channel function via a variety of different mechanisms (discussed below). The relative ability of particular agonists and antagonists to either activate or inhibit channel function also provides a tool to distinguish between different ion channel subtypes. In addition to their role as channel activators and inhibitors, agonists and antagonists, particularly high affinity agonists and antagonists, can be valuable tools in binding experiments examining the tissue distribution of particular ion channel subtypes.

Although a great deal can be learned by studying the binding characteristics of ligands to specific receptor sites, the study of ion channel function requires a technique whereby the functional effects of ligand binding can be observed. The direction of flow of ions is determined by the relative ionic concentrations present on the interior and exterior of the cell. Because the cell membrane acts as a barrier to the free diffusion of ions between the cytoplasm and the extracellular space, ATP-activated ion transport systems can bring about the production of ionic gradients which accumulate with respect to the interior and exterior of the cell. Potassium ions are concentrated intracellularly while sodium ions are concentrated extracellularly. This gradient is actively maintained by the activity of a Na-K pump which exchanges intracellular sodium for extracellular potassium. Because this gradient represents potential energy it is referred to as the membrane potential. Because the membrane potential arises from the separation of charged ions, it necessarily has both an electrical and chemical component. When the membrane permeability for a specific ion increases (by receptor activation, for example) it is possible to describe an electrical


potential which exactly counterbalances the chemical potential arising from the concentration gradient. At this potential, there will be no flow of ions across the membrane. This potential is known as the ionic equilibrium potential. In contrast, the magnitude and direction of net charge flow through a nonspecific cation channel like nAChR is dictated by the voltage difference between the membrane potential (Vm) and the combined and weighted equilibrium potentials for each permeant ion (the reversal potential or Erev), a quantity known as the driving force (Vm-Erev).

Electrophysiology takes advantage of the fact that we can treat the flow of ions across a membrane in a similar manner to the flow of electrons in electricity. That is, the movement of charge arising from the activation of ion channels can be described by Ohm's law:




V=voltage or driving force

G=conductance (the inverse of resistance)

In other words, the amount of current (movement of charge) observed involves both the magnitude of the electrical potential gradient and the ability of current to flow through the conductor (in this case, the ion channels). Many electrophysiology experiments are carried out under conditions of voltage clamp where the membrane potential is held constant. In this case, it is possible to relate the currents observed directly to the conductance of the membrane (or more specifically, the conductance of the ion channels in the membrane). For a system such as a heterologous expression system (e.g., Xenopus oocyte), in which the expressed channels are of a homogeneous nature, a constant number of channels of equivalent conductance are present. Therefore, it is useful to expand Ohm's law to:




N=# of ion channels

po=probability of the channels to be open

and "-=the conductance of a single channel

From the discussion above, it is apparent that the electrical potential gradient is defined by the driving force on the ions (Vm-Erev). Because in an oocyte voltage-clamp experiment, Vm-Erev, N and y are all constant, it is possible to directly relate current to the probability of the channels to be open. For a ligand-gated channel, open probability depends upon agonist concentration. Thus, it is possible to relate directly agonist concentration to receptor response. The ability to reliably quantify the activation and inhibition of specific receptor subtypes in response to the application of specific drugs has allowed a detailed description of the mechanism of action of a variety of different drugs and, furthermore, has permitted the comparison of functional differences across receptor subtypes. For example, although a binding experiment may indicate that a particular agonist has a similar affinity across receptor subtypes, electrophysiological characterization of the effects of this drug may reveal a selectivity for the activation of a particular subtype. In the case of neuronal nAChRs, the agonist cytisine exhibits a high binding affinity for beta2-containing receptors; however, voltage clamp studies of heterologously expressed receptors indicate cytisine to be only a partial agonist with very low efficacy for receptors containing the beta2 subunit and a full agonist for receptors containing the beta4 subunit (Papke and Heinemann, 1994). Similarly, the agonist DMPP is more efficacious on receptors containing the beta2 subunit and less so on receptors containing the beta4 subunit (Luetje and Patrick, 1991). Thus, binding affinity can not be viewed as a reliable indicator of agonist efficacy or potency. By using electrophysiology in combination with molecular


biological approaches it is possible to localize the structural elements specific to individual subunits which underlie each of these characteristics.

The voltage-clamp technique has permitted detailed analysis of the mechanism of action of inhibitors of ion channel function as well. The mechanism of antagonist action is varied and can be quite complex; however, antagonists can be broadly classified as either competitive or noncompetitive. Competitive antagonists bind to the same site as agonist but do not activate the receptor and therefore compete with agonist for occupation of the receptor binding site. A hallmark of this type of antagonist activity is the observation that high concentrations of agonist are able to compete off more moderate concentrations of the competitive antagonist and fully activate a population of channels, albeit at a higher concentration than if the antagonist was not present at all. In pharmacological terms, the presence of the competitive antagonist would shift the concentration-response curve for the agonist to the right indicating an apparent change in potency while no change in efficacy or maximal response would be observed. As is the case for agonists, certain antagonists show selectivity for the inhibition of particular receptor subtypes. For example, in the case of neuronal nAChRs, beta2-containing receptors exhibit more prolonged inhibition by the competitive antagonist neuronal bungarotoxin compared with beta4-containing receptors. Structural elements underlying this difference in sensitivity have been localized to the Nterminal 121 amino acids of the beta subunit (Papke et al., 1993). In addition, high affinity competitive antagonists (e.g., ax-bungarotoxin), like potent agonists, can be extremely valuable as reagents for binding experiments examining the tissue distribution of particular receptor subtypes.

Noncompetitive antagonists act at sites that are distinct from the agonist binding sites and are, in some cases, also use-dependent meaning that they require the presence of agonist in order to inhibit the receptor. The study of the mechanism of action of noncompetitive inhibitors has proven to be quite complex and has provided a great deal of information about ion channel function. However, in general the activity of noncompetitive


inhibitors is characterized by a reduction in the efficacy of agonist for activation of the receptor. Thus, in contrast to the situation for competitive inhibitors in which high concentrations of agonist can alleviate the effects of the inhibitor, the effects of noncompetitive inhibitors remain unaffected by application of high concentrations of agonist. It should be noted that, in the case of use-dependent noncompetitive inhibitors, which require prior activation of the channel, a low agonist concentration can affect the amount of inhibition observed for a population of channels because only channels which have been activated are available for inhibition.

The most well described class of noncompetitive inhibitors are known as open-channel blockers (Neher and Steinbach, 1978). Open-channel blockers are generally small, charged molecules which act by occupying a site within the ion channel pore and physically occluding ion permeation. Because open channel blockers do bind within the membranespanning region of the ion channel, their binding and unbinding can be affected by the membrane electric field, an electric field created by the voltage gradient across the membrane. By analyzing the degree to which inhibition is affected by membrane potential, it is possible to use this feature of open-channel block to assess the depth of an inhibitor binding site within the channel pore. Accordingly, open-channel blockers have been extremely useful in the process of identifying structural components of the receptor which contribute to the pore-lining region. A second more heterogeneous and less well characterized class of noncompetitive inhibitors are known as allosteric inhibitors because they are not believed to physically occlude ion channel permeation but most likely bind to a site outside the pore-lining region and act to stabilize the receptor in a closed configuration (Papke and Oswald, 1989). In general, inhibition by this class of compounds is not statedependent (i.e., does not require prior activation of the channel). Additionally, there is some evidence for a class of inhibitors with characteristics intermediate to those of openchannel blockers and allosteric inhibitors (Papke et al., 1994). For this class of


compounds, inhibition seems to require prior activation of the channel but does not appear to have an appreciable voltage-dependence.

Channel block by agonist constitutes a final type of noncompetitive inhibition of nAChR. It has been well documented in the literature that, at high concentrations, acetylcholine and a number of acetylcholine analogues can have secondary inhibitory effects on muscle-type/Torpedo nAChRs (Arias, 1996; Marshall et al, 199 1; Ogden and Colquhoun, 1985; Sine and Steinbach, 1984; Tonner et aL, 1992). Data from this laboratory suggest that nicotine can have a similar effect on certain subtypes on neuronal nAChR (de Fiebre et aL, 1995). In particular, the 00j2(x5 subtype of neuronal nAChR seems to exhibit enhanced sensitivity to secondary inhibition by both ACh and nicotine (Francis et al, unpublished observations). Moreover, many of the experimental new nicotinic agents being considered for clinical development are in fact only partial agonists for high affinity nicotinic receptors and also share, with nicotine, the ability to function as antagonists. The identification and characterization of subtype-specific structural determinants of pharmacological sensitivity to each of the classes of drugs described above remains of major importance as it could prove critical in developing more effective and more specific pharmacological agents for nAChRs.

Neuromuscular Junction nAChR

As a principal component of the most well studied synapse in neurophysiology, the muscle-type nAChR is the prototype ion channel against which all others are compared. However, in many respects, the neuromuscular junction is one of the most specialized synapses in the body. The major function of muscle nAChR seems to be the faithful transmission of pre-synaptic impulses to the post-synaptic muscle fiber and the organization of the synapse serves to optimize nAChR activation for this role. At the mature neuromuscular junction, the surface of the muscle fiber is characterized by a series of invaginations. Nicotinic AChRs are concentrated at the tips of these folds (at a density


of about 20, 000 binding sites per 4m2) in a location known as the endplate. The presynaptic terminals of the nerve contact the muscle fiber in this endplate region, bringing the release sites for neurotransmitter (the active zone) and the post-synaptic receptors together within a distance of about 500 'A at the synaptic cleft. When an action potential causes release of synaptic vesicles containing ACh, the post-synaptic receptors are rapidly activated initiating a local change in the membrane potential of the muscle fiber. This change in membrane potential activates voltage-gated sodium channels located deep within the junctional folds and a muscle action potential can be generated. A safety factor of about 10 in terms of excess acetylcholine and receptors is built in to this synapse in order to ensure faithful signal transmission. Furthermore, the presence of acetylcholine esterase, an enzyme which cleaves acetylcholine, in the extracellular matrix of the synaptic cleft ensures that ACh is present for only a very short time. Thus, each ACh receptor probably only binds ACh once during a single synaptic event and the termination of the response is mediated by the intrinsic rate of deactivation or closing of the receptor (Katz and Miledi, 1973). This organization permits rapid and efficient signaling of even high frequency impulses between the nerve and the muscle fiber which it contacts. Functional States of Muscle nAChR

Exploration of the role of nAChR in mediating transmission at the nerve-muscle synapse has led to an understanding of the gating transitions associated with ion channel function itself. After studying the response properties of muscle-type nAChR to acetylcholine in the presence of a variety of choline derivatives and the esterase inhibitor prostigmine, del Castillo and Katz (1957) were able to conclude that "the first step in a depolarizing endplate reaction is the formation of an intermediate, inactive compound between drug and receptor" (p. 369). This result eventually led to a kinetic model for the functional transitions involved in the activation of nAChR. Although subsequent models have grown increasingly complex, the most basic kinetic model consists of three general states of the receptor: an unoccupied closed state, an agonist bound closed state and the open state.


Subsequent analyses have demonstrated that, at high agonist concentrations, activation is predominantly associated with the binding of two molecules of acetylcholine and that the prolonged presence of agonist at the receptor can induce a non-conducting state which has a high affinity for ligand, known as a desensitized state. Thus, a reasonably accurate description of most response properties of nAChR can be given by the relation:

R AR--_ A2R A2R*- RD




A2R*=agonist bound open state

D= desensitized state

The arrows represent transitions between discrete states which proceed at characteristic rates. The direct measurement of rates for transitions between states was made possible by the development of the patch clamp technique (Hamill et al., 1981; Neher and Sakmann, 1976). During the analysis of single channel responses to the agonist suberyldicholine, Colquhoun and Sakmann (198 1) observed what appeared to be short, non-conducting states within the current responses of single channels. Analysis of these gaps or fast closures led Colquhoun and Sakmann to conclude that the response to a single agonist binding event consists of a burst of activity. A burst is represented by transitions between the agonist bound closed state and agonist bound open state while longer closed states are predominantly associated with extraburst closures and the dissociation of agonist. Other long closed states which may observed include desensitized states and blocked states.


Physiological Characteristics of Muscle versus Neuronal nAChRs

Although neuronal and muscle nAChR exhibit many similar functional characteristics, there do exist significant differences. In particular, neuronal nAChR subtypes generally have a higher calcium permeability than muscle nAChR (Mulle et aL., 1992; Rogers and Dani, 1995; Vernino et al., 1992, 1994). Regulation of calcium influx via neuronal nAChRs could activate second messenger pathways which play a role in processes as diverse as neuronal survival, transmitter release and synaptic strengthening. For example, calcium influx via nAChRs on neurons of the medial habenula has been shown to be sufficient to activate a calcium-dependent chloride conductance and, in addition, cause a decrease in the response of GABAA receptors (Mulle et a!., 1992). The alpha7 subtype of neuronal nAChR in particular has been demonstrated to have a very high calcium permeability and much interest has focused on potential functional roles for this receptor subtype. In addition to differences in calcium permeability, neuronal and muscle nAChRs differ in their capacity to conduct outward current at positive potentials. Muscle nAChRs exhibit a linear current-voltage relation (equal amount of outward current at positive potentials as inward current at negative potentials) while neuronal nAChRs exhibit an inwardly rectifying current-voltage relationship (greater inward current at negative potentials than outward current at positive potentials). Recent evidence indicates that, similar to voltage-gated potassium channels and glutamate receptors, the inward rectification of neuronal nAChRs is mediated via channel block by intracellular polyanines at positive potentials (Haghighi and Cooper, 1998).

In vivo Subunit Composition and Distribution of Neuronal nAChRs

As is apparent from the number of genes encoding subunits for neuronal nAChRs, there is a much greater diversity of neuronal nAChR in terms of subtype, localization and presumed functional role than is the case for muscle-type nAChR. Each of the various neuronal nAChR subunits appears to have a tissue-specific distribution (Wada et al., 1989)


and there is potential for a great variety of subunit combinations to appear in the nervous system. Functional nicotinic receptors are present both centrally in the brain and spinal cord and peripherally on neurons of the autonomic ganglia. In general, the typical subunit composition of neuronal nAChR from peripheral ganglia appears to be distinct from the most widely expressed subunit combinations centrally. However, evidence from a number of studies supports the notion that a variety of subunit combinations exist both centrally and peripherally.

Peripheral Neuronal nAChRs

It is apparent that receptors containing the alpha3 and beta4 subunits are the predominant nAChR of the peripheral ganglia. However, it remains unclear to what extent these subunits coassemble with other nicotinic subunits and the description of receptor subunit composition on individual neurons remains problematic. In fact, it appears that multiple nAChR subtypes can exist on the same ganglionic neuron (Conroy and Berg, 1995; Poth et al., 1997; Vernallis et al., 1993). Furthermore, in the case of chick ciliary ganglion neurons, it has been shown that the alpha3 subunit nearly always coassembles with the beta4 subunit while the alpha5 subunit always coassociates with the alpha3 and beta4 gene products. About 20% of these synaptic ciliary ganglion nAChR also contain the beta2 subunit (Conroy and Berg, 1995). To add further complexity, an additional subtype of nAChR containing the alpha7 gene product is present but seems to exhibit a non-synaptic or perisynaptic distribution on the neuron (Vernallis et aL, 1993; Wilson Horch and Sargent, 1995). These findings are discussed in more detail below. Central Neuronal nAChRs
It has been well established that the nAChR of the neuromuscular junction can be bound nearly irreversibly by the snake venom toxin a-bungarotoxin (a-BTX) and that this ligand labels putative receptor sites in brain. However, studies employing labeling of central nAChR by agonists such as acetylcholine or nicotine demonstrate a labeling pattern quite distinct from that of a-BTX (Clarke et aL, 1985) and, in addition, early studies using


heterologously expressed heteromeric neuronal nAChR show these receptors to be insensitive to inhibition by a-BTX (Deneris et al., 1988). With the cloning of the alpha7 gene product (Couturier et at., 1990; Schoepfer et al., 1990; Seguela et at., 1993), it became apparent that a distinct population of nAChR with relatively low binding affinity for the traditional nicotinic agonists and very high affinity for a-BTX exist in brain. Therefore, it is useful to divide the description of central neuronal nAChR subtypes into two general populations: those that bind nicotinic agonists with high affinity and are insensitive to inhibition by a-BTX and those nAChR which bind agonist with relatively low affinity and are sensitive to low concentrations of a-BTX. High Affinity Brain nAChRs

Towards describing the high affinity brain nAChR population, Whiting and Lindstrom (1986) demonstrated that an antibody raised to chick neuronal nAChR (mAb 270) also cross-reacts with rat neuronal nAChR and is able to bind greater than 90% of the high affinity nicotine binding sites in detergent extracts of rat brain. It has since been shown that mAb 270 binds what is now designated the beta2 subunit and this antibody has been used to show the distribution of beta2 subunits in rat brain (Swanson et al., 1987). Via analysis of immunoprecipitation of nAChRs labeled by the specific nicotinic agonist cytisine with antibodies specific for the alpha4 and beta2 subunits, Flores et al. (1992) also show that high affinity binding sites are composed a402 receptors and moreover that these sites are upregulated after chronic exposure to nicotine. The highest densities of mAb 270 labeling are in interpeduncular nucleus, the nuclei of the thalamus, superior colliculus, medial habenula. More moderate labeling occurs in the presubiculum, layers I and IIIW of cerebral cortex and in the substantia nigra pars compacta/ventral tegmental areas. This distribution corresponds quite well with the previously described distribution of high affinity nicotinic agonist binding sites (Clarke et al., 1985). It is now apparent that beta2 has the most widespread distribution of the neuronal nicotinic receptor subunits and can coassemble with alpha2, alpha3 or alpha4; however, the majority of this labeling most


likely represents the expression pattern of receptors containing the alpha and beta2 subunits. It may be the case that a significant proportion of these receptors coassemble with the alpha5 subunit in the mature brain (Conroy and Berg, 1998). More recent studies examining the brains of beta2 knockout mice show a complete lack of high affinity nicotine binding sites providing further evidence consistent with the idea that the most prevalent high affinity nAChRs in brain are 132-containing (Picciotto et al., 1995). Interestingly, residual binding of other more subtype-specific ligands can be detected in the brains of beta2 knockout mice (Zoli et al., 1998) suggesting that there exist high affinity nicotinic subtypes in which beta2 does not participate. Electrophysiological studies have also indicated that neurons of the habenulo-interpeduncular system have responses consistent with expression of a variety of nAChR subtypes (Mulle et al., 1991). Low Affinity Brain nAChRs

The second grouping of brain nAChR is termed low affinity because the desensitized state of this receptor does not exhibit a high affinity for agonist. Thus, in binding experiments in which the desensitized state of nAChR predominates, this class of receptor appears to have a very low affinity for agonist compared to the class of neuronal nAChR described above. In contrast, the low affinity brain nAChRs are characterized by a high affinity for the binding of ct-BTX. The binding of a-BTX is high in cerebral cortex (layers I and VI), superior and inferior colliculus, hypothalamus and hippocampus (Clarke et al., 1985). The only regions of significant overlap with high affinity agonist binding are in layer I of cortex and superior colliculus. Until recently, the description of a functional nicotinic receptor underlying brain a-BTX binding was problematic because, although labeling of protein by a-BTX in brain and peripheral ganglia is readily apparent, it is difficult to detect a-BTX sensitive responses in brain and application of a-BTX does not inhibit the most easily detected nAChR responses on ganglionic neurons. However, with the cloning of the alpha7 subunit it has become apparent that the labeling by c-BTX does in fact represent the distribution of functional neuronal nAChRs.


Affinity purification of high affinity a-BTX binding proteins from chick brain allowed N-terminal protein sequencing (Conti-Tronconi et al., 1985). From the this sequence, Schoepfer et al (1990) prepared degenerate oligonucleotides which isolated clones associated with a-BTX binding. Working independently, Couturier et al. (1990) were able to demonstrate that the cDNA encoding a-BTX binding protein from chick brain expresses a functional low affinity nicotinic receptor that is sensitive to block by c-BTX and exhibits rapid desensitization. This finding was later confirmed by Seguela et al. (1993) for the alpha7 clone from rat brain. With the development of rapid agonist application systems, a number of investigators have demonstrated the existence of rapidly desensitizing nicotinic responses both centrally and peripherally. Thus, it is now clear that the a-BTX labeling in rat brain is associated with the alpha7 subtype of nAChR whereas in chick brain the alpha8 subunit also contributes. The related low-affinity nAChR subunit alpha9 also forms homomeric receptors which are blocked by c-BTX but has a very limited distribution (expressed in the cochlea for the most part) and recognizes nicotine only as an antagonist (Elgoyhen et al., 1994).

Functional Roles of Neuronal nAChRs

Interest in the function of neuronal nAChRs stems from the observation that nicotine can increase performance in some measures of memory. Although this effect is well documented, it is most likely the case that memory performance is also linked to the state of arousal of the test subject. Thus, the effects of nicotine may be mediated through nonspecific effects on arousal. However, the additional observation that there is a selective loss of cholinergic neurons during the progression of Alzheimer's disease also implicates nicotinic systems in the process of memory formation. Nicotinic receptors have also been linked to the motor deficits associated with Parkinson's disease, the sensory gating deficits associated with schizophrenia and, of course, to nicotine addiction. Furthermore, a number of recent reports have described the utility of nicotinic agonists as non-opioid


analgesics. With these observations in mind, a number of laboratories have proceeded to characterize the function of neuronal nAChRs (Holladay et al., 1997).

The heterogeneity of neuronal nAChR in terms of subunit composition may encode a corresponding functional diversity which is required for neuronal function (for review see Papke, 1993). It has been demonstrated that the time-course of desensitization and single channel kinetics of heterologously expressed alpha3-containing receptors depends on the particular beta subunit with which the alpha subunit is expressed (Papke and Heinemann, 1991). It is possible that the pattern of activity of individual receptor subtypes could be important for coincidence detection in the brain. However, it must also be noted that neuronal nAChRs expressed in heterologous expression systems do not fully recapitulate the characteristics of native receptors (e.g., in terms of single channel conductance, (Sivilotti et al., 1997)). It is unclear whether the differences between heterologously expressed receptors and native receptors arise because of the potential for complex subunit arrangements in vivo that are not reproduced in heterologous expression systems or possibly because of differences in modulation of receptor function by different cell types (e.g., by phosphorylation).

It is also the case that calcium permeability may be dependent upon subunit

combination. In whole-cell recordings of transfected cells in conjunction with fluorescence imaging of a calcium indicator dye, Ragozzino et al. (1998) demonstrate that a314 receptors have a slightly higher fractional calcium conductance than alpha4-containing receptors. However, the bulk of work in this field indicates that any differences in calcium permeability across the heteromeric subtypes of neuronal nAChRs are most likely minor. By measuring reversal potential shifts with different concentrations of extracellular calcium in the cut-open oocyte system (allowing access to both sides of the oocyte membrane), Costa et al. (1994) demonstrated that heterologously expressed a334 receptors have a pCa/pNa ratio of 1.1 compared to a pCa/pNa ratio of 0.12 for muscle-type nAChR, while other heteromeric subtypes have a permeability ratio in the range of 1.0-1.5. For this class


of receptors, a direct effect of external calcium on open probability has also been described (Amador and Dani, 1995). This direct effect of external calcium may also prove to be of physiological significance. By way of contrast, the alpha7 homomeric nAChR has pCa/pNa ration in the range of 20 (Seguela et al., 1993). Additional studies of the permeability of heterologously expressed alpha7 receptors, in which barium was substituted for calcium to minimize any contaminating influence of calcium-activated chloride currents to the measurements, indicate a pBa/pNa ratio of 17 (Sands et al., 1993). However, it should also be noted that similar measures for a-BTX-sensitive responses in hippocampus indicate a pCa/pCs ratio of 6.1 (Castro and Albuquerque, 1995). In the same study, the pCa/pCs ratio for the highly calcium-permeable NMDA subtype of glutamate receptor was measured to be 10.3. The high calcium permeability of the alpha7 subtype may be prove to be of major importance as this receptor subtype could provide a route for calcium entry at negative potentials for which both voltage-gated calcium channels and NMDA receptors would be inactive.

The function of nAChR of the peripheral ganglia may be similar to that of muscle processing at the level of the ganglion. However, the functional organization of the ganglionic synapse has proven to be quite elaborate. The most well characterized ganglionic synapse is that of the chick ciliary ganglion. Accessory motor neurons of the chick midbrain send processes to this ganglion which terminate in calycal boutons enveloping the post-synaptic cell. The post-synaptic receptors at this synapse are of two major types: those labeled by mab35 (monoclonal antibody specific for the alpha subunit) and those bound irreversibly by a-BTX. The mab35 labeled receptors most likely correspond to different classes of alpha3-containing receptors (with alpha) while the aBTX subtype most likely corresponds the alpha7-containing receptors (discussed in more detail above). It has been clearly demonstrated that the a-BTX-sensitive subtype can raise intracellular calcium and function as rapidly desensitizing nAChRs (Vijayaraghavan et al., 1992). Studies examining the distribution of these two receptor classes in ganglionic


neurons show that alpha7-containing receptors outnumber mab35 labeled receptors and seem to be localized in clusters with a perisynaptic distribution. Although not strictly necessary for transmission through the ganglion, the responses of alpha7-containing receptors seem to contribute a large portion of the synaptic current (Ullian, 1997; Zhang et al., 1996). In consideration of the high calcium permeability of the alpha7 subtype, this observation may imply that there exist functional roles in the ganglion for this class of receptor apart from participating directly in synaptic transmission. In fact, the regulation of intracellular calcium concentration for ganglionic neurons seems to be exceedingly complex. Increases in calcium concentration in the post-synaptic ganglionic neuron can be initiated by activation of nicotinic or muscarinic acetylcholine receptors (Rathouz et aL, 1995). Activation of muscarinic receptors results in oscillatory increases in intracellular calcium that are dependent upon the release of calcium from intracellular stores and coupled to phosphatidylinositol turnover while activation of nicotinic receptors results in sustained calcium increases dependent upon extracellular calcium. Differences in concentration dependence and receptor localization between muscarinic and nicotinic receptors and between classes of nicotinic receptor may allow for selective activation of distinct second messenger cascades by each of these processes.

Additional studies of receptor distribution on mature ganglionic neurons by laser confocal microscopy indicate that only about 10% of the mab35 labeled receptors lie directly across from presynaptic active zones (as labeled by the synaptic vesicle antigen SV2) (Wilson Horch and Sargent, 1995). Thus, according to this study the bulk of both receptor classes on chick ciliary ganglion neurons are located perisynaptically. It may be the case that the nonsynaptic receptors are activated in a long range fashion by the diffusion of ACh out of the synaptic cleft. However, the presence of acetylcholine esterase in the cleft makes this route of activation by acetylcholine seem inefficient for transmission (Zhang et al., 1996). Although the functional significance of perisynaptic receptors (of both classes) for transmission dependent upon acetylcholine remains unclear., it is possible


that activation of perisynaptic alpha7 receptors by choline could influence cellular excitability. To complicate matters further, it has been shown that the post-synaptic neurons produce arachidonic acid in an activity- and calcium-dependent fashion and that arachidonic acid can inhibit alpha7-containing receptors raising the possibility of retrograde effects on the presynaptic cell (Vijayaraghavan et al., 1995). Moreover, the presence of functional presynaptic ct-BTX sensitive nAChRs on the presynaptic terminals has recently been demonstrated (Coggan et aL, 1997). Interestingly, the responses of these presynaptic a-BTX sensitive nAChRs do not appear to be rapidly desensitizing. This result may imply a different subunit composition or different modulation between a-BTXsensitive nAChRs of the pre- and post-synaptic neuron. However, desensitization of alpha7 receptors is strongly dependent upon concentration and it may be the case that solution exchange at the presynaptic terminal was not complete such that the effective concentration reaching the receptors was lower than the bath concentration of agonist in the recording chamber.

Function of Central Neuronal nAChRs

Brain nAChRs have been implicated in the pathology of a number of disease states including Alzheimer's disease, Parkinson's disease and schizophrenia. A number of biochemical studies in synaptosomal and slice preparations have shown that nicotinic agonists can affect the release of neurotransmitters including 5-HT, dopamine (DA) and noradrenaline (NE) implying a potential contribution of presynaptic nAChRs (for review see (Wonnacott, 1997)). However, until recently there was scant electrophysiological evidence for a functional role of brain nAChR. Consistent with the biochemical data, an increasing amount of electrophysiological evidence points to the fact that although somatic responses to nicotinic agonists have been recorded, the major functional contribution of brain nAChRs arises from a presynaptic receptor population (for review see (Role and Berg, 1996)). In general, studies conducted in various synaptosomal and slice preparations are consistent with a contribution of heteromeric nAChRs (e.g., a4132 or a3134) to the


stimulation of neurotransmitter release because the effects are not inhibited by ca-BTX or the competitive antagonist of alpha7 receptors MLA (methllycaconitine). However, it is unclear if an effect of alpha7 receptors could be detected in these studies because of the rapidly desensitizing nature of alpha7 responses.

The medial habenula (MHB) and interpeduncular nucleus (IPN) are connected via the fasciculus retroflexus and each of these regions show high expression of a variety of nicotinic mRNAs and the presence of nicotine binding sites. In addition, electrophysiological characterization of the response of rat MHB neurons indicates the presence of at least two distinct subtypes of nAChR (Connolly et al., 1995). Effects of presynaptic nAChRs on the responses of post-synaptic IPN neurons have been described in separate studies.

Mulle et al. ( 1991) characterized the nAChRs of isolated rat MHB and IPN neurons and showed a differential rank order potency for agonist activation between the two regions. The pharmacological profile of MHB nAChRs is most consistent with the a304 subtype while that of IPN neurons is most consistent with alpha2-containing receptors. In addition, it was demonstrated that the presence of nicotine reduces the amplitude of afferent volleys stimulated in the fasciculus retroflexus. These investigators attribute this reduction to a shunting effect as a result of depolarization of the presynaptic terminal upon activation of presynaptic nAChRs in the presence of nicotine. Consistent with the pharmacological profiles of somatic nAChRs, this effect is insensitive to a-BTX and sensitive to other inhibitors of neuronal nAChRs such as hexamethonium and mecamylamine indicating that it is most likely mediated by heteromeric nAChRs.

Further characterization of presynaptic nAChRs via whole-cell recording of neurons in slices of rat IPN have demonstrated that nicotine increases the frequency of GABAergic and glutamatergic post-synaptic events (Lena et al., 1993). This effect was found to be TTX-sensitive and, based on the fact that the frequency increase seemed to be dependent upon activation of voltage-gated sodium channels, it was hypothesized that presynaptic


nAChRs may be located "preterminally". Furthermore, this effect was insensitive to aBTX but sensitive to traditional inhibitors of neuronal nAChRs such as hexamethonium, mecamylamine and DHI3E (dihydro-beta-erythroidine). GABAergic innervation of IPN neurons is thought to arise from local IPN interneurons rather than via the fasciculus retroflexus so activation of this population of presynaptic nAChRs may provide a method for modulating a local inhibitory circuit.

In addition to the above results, McGehee et aL. (1995) have demonstrated an a-BTX sensitive effect of presynaptic nAChRs in cocultures of chick MHB and IPN neurons. In these cultures, application of nicotine enhances both evoked and spontaneous release (EC50~ 120 nM) in the presence of TTX indicating that nAChRs are most likely located directly on the terminals. The EPSCs were shown to be sensitive to CNQX, a specific inhibitor of the non-NMDA subtype of ionotropic glutamate receptors, indicating that transmission at this synapse is glutamatergic. The half-maximal inhibitory concentration for a-BTX inhibition of the nicotine effect on synaptic transmission was about 70 times higher than reported IC50s for the homomeric alpha7 receptor indicating that the nAChRs mediating the effects in this study may possibly be heteromeric alpha7-containing receptors.

Although the effects presynaptic nAChRs in the habenulo-interpeduncular tract have

been the most well characterized, functional effects via putatively presynaptic nAChRs have been described in rat prefrontal cortex, rat hippocampus, rat dorsal raphe nucleus, mouse thalamus (ventrobasal complex and dorsolateral geniculate nucleus), chick lateral spiriform nucleus (part of the avian basal ganglia) and chick lateral geniculate nucleus. Nicotinic receptors in prefrontal cortex seem to modulate excitatory transmission via non-NMDA glutamate receptors in a neuronal BTX-sensitive but a-BTX-insensitive manner (Vidal and Changeux, 1993). Presynaptic nAChRs of the dorsal raphe nucleus seem to modulate release of both NE and 5HT to metabotropic (G-protein coupled) receptors in an MLAsensitive and MLA-insensitive manner respectively (Li et al., 1998). Release of GABA in


the thalamus was also found to be modulated by presynaptic nAChRs (Lena and Changeux, 1997). In different sensory nuclei, presynaptic effects either required the simultaneous activation of voltage-gated calcium channels or could be mediated directly by calcium influx through nAChRs. Furthermore, these effects were absent in P32 knockout mice implicating presynaptic a4132 receptors. Presynaptic nAChRs in chick ventrolateral geniculate nucleus (LGN) modulate the release of both GABA and glutamate in an a-BTX sensitive and a-BTX insensitive manner respectively; however, the a-BTX sensitive effects on glutamate release in chick LGN were also found to be MLA insensitive indicating an nAChR of previously undescribed pharmacology (Guo et al., 1998). Nicotinic facilitation of GABA release in chick lateral spiriform nucleus is DH1OE sensitive but sensitivity to ax-BTX was not tested (McMahon et al., 1994). Consistent with the involvement of nicotinic systems in a variety of cognitive disorders, these studies demonstrate the potential involvement of multiple subtypes of presynaptic nicotinic receptors in the activity of a variety of other neurotransmitter systems ranging from the major excitatory and inhibitory ligands for central ionotropic receptors to activators of metabotropic neurotransmitter systems. Although functional roles for presynaptic receptors are likely specific to each neurotransmitter system in which they are expressed, it is almost certainly the case that the presence of presynaptic nAChRs increases the spatial and temporal range of inputs which may result in neurotransmitter release and thereby increases the receptive field of the post-synaptic neuron. Moreover, the presence of presynaptic nAChRs may allow for release independent of a requirement for depolarization to a potential which would activate voltage-gated calcium channels. Both of these effects may provide a mechanism whereby the probability of release and thus efficiency of a particular synapse cam be modulated.

In view of the importance of hippocampus for memory consolidation and the

demonstration of multiple forms of synaptic plasticity within hippocampus, elucidation of the function of neuronal nAChRs within this circuit is particularly exciting. Whole-cell


recordings of pyramidal neurons in hippocampal slices indicate that application of nicotine increases the frequency of EPSCs at mossy fiber-CA3 synapses (Gray et al., 1996). Consistent with the studies in MHB-IPN cocultures, this effect was found to be a-BTX and MLA-sensitive indicating that this effect is mediated via alpha7-containing receptors. Moreover, calcium imaging indicates that application of nicotine in the mossy fiber region induces similar amounts of calcium influx as invasion of an action potential to the terminal region. Therefore, calcium entry directly through nAChRs without the requirement of a contribution of voltage-activated calcium channels may be sufficient to facilitate the release of glutamate at this synapse.

In addition to the presynaptic role for nAChRs in hippocampus, it may be the case that nAChRs mediate synaptic transmission in the hippocampus directly. There is cholinergic innervation of the hippocampus via the septum and high levels of a-BTX binding are present in hippocampus. Moreover, robust a-BTX sensitive responses to nicotinic agonists are present on cultured hippocampal neurons (Alkondon and Albuquerque, 1993; Alkondon et al., 1994; Castro and Albuquerque, 1995; Zorumski et al., 1992). In addition, several recent studies report the existence of rapidly desensitizing, a-BTXsensitive responses on interneurons but not pyramidal cells of rat hippocampus (Frazier et al., 1998; Jones and Yakel, 1997).

The demonstration of the existence of functional presynaptic nAChRs raises a question as to the proximity of release sites for acetylcholine to presynaptic terminals. In principle, activation of presynaptic receptors could arise via at least three mechanisms: direct axoaxonic cholinergic synapses, diffusion of synaptically released ACh to nicotinic autoreceptors or synaptic spillover of acetylcholine between adjacent synapses. As there seem to be few, if any, purely nicotinic responses in brain, it is most likely that the first of these options will prove to underlie activation of presynaptic brain nAChRs. The demonstration of modulation of evoked neurotransmitter release via stimulation of intact


nicotinic axo-axonic terminals should provide definitive evidence for the functional significance of presynaptic nAChRs.

However, in the absence of such direct evidence, it is interesting to speculate on

alternative roles for nicotinic receptors in the brain. The observation that choline is a fully efficacious agonist for alpha7 receptors provides a challenge to the notion that the major function of this receptor subtype is to participate in fast synaptic transmission (Papke et al., 1996). As choline would be predicted to be fairly ubiquitous in brain, activation by choline may provide an alternate, non-synaptic route by which neuronal calcium influx may be regulated. Nonetheless, in light of recent progress in the description of the function of nicotinic receptors in brain, it should be noted that both alpha7 knockout mice and beta2 knockout mice survive to adulthood without any severe anatomical or behavioral abnormalities. This result may imply that a high degree of functional redundancy exists in cholinergic brain systems such that knockout of a single receptor population is readily compensated for (Orr-Urtreger et al., 1997; Picciotto et al., 1995).

Sensitivity to Noncompetitive Inhibitors In addition to differences in distribution and function, neuronal and muscle nAChRs differ in their sensitivities to certain classes of noncompetitive inhibitors. These compounds are known as ganglionic blockers and show selectivity for the inhibition of neuronal nAChRs. The ganglion blocking activity of compounds such as mecamylamine, hexamethonium, chlorisondamine, TMP (2,2,6,6 tetramethylpiperidine) and PMP (1,2,2,6,6 pentamethylpiperidine or pempidine), has been well documented in the literature (Lee et al., 1958; Spinks and Young, 1958).

The effort to develop pharmaceutical agents specific for neuronal nAChRs dates back to the middle of this century. The observation that synaptic transmission through the autonomic ganglia was mediated by the chemical messenger acetylcholine led scientists to hypothesize that inhibition of ganglionic synapses might be a route by which disorders


related to autonomic nervous function could be regulated. Since that time, a heterogeneous group of compounds with varying selectivities for ganglionic nicotinic receptors has been developed and characterized. However, because of the wide range of functions which are affected by inhibition of the entire ganglia, clinical applications for these compounds were never pursued. In fact, compounds specific for post-ganglionic noradrenergic receptors have proven to have a better clinical utility both in terms of safety and effectiveness. Although of little clinical utility, these compounds remain of considerable scientific importance in part because an understanding of the mechanism of action of pure antagonists may allow for a more useful consideration of the observation of mixed agonist/antagonist effects of nicotine and other experimental nicotinic agents.

Although ganglionic blockers have been used extensively as selective blockers of neuronal nAChRs, there is a relative paucity of experimental evidence regarding the mechanism of selectivity of the various compounds. Each of the compounds seems to act in a noncompetitive manner and varying degrees of voltage-dependence for inhibition of neuronal nAChRs has been reported. Examination of inhibition of rat submandibular ganglion nicotinic receptors by hexamethonium indicate an open channel block mechanism. Additionally, recovery from inhibition seems to be dependent upon subsequent application of ACh consistent with a model in which hexamethonium becomes trapped in the pore (Gurney and Rang, 1984). Consistent with this result, studies of heterologously expressed human and rat a402 neuronal nAChRs indicate that inhibition by hexamethonium is usedependent and profoundly voltage-dependent (Bertrand et al., 1990; Buisson and Bertrand, 1998). Moreover, analysis of voltage-jump relaxations indicate that inhibition of neuronal a402 receptors by hexamethonium is voltage-dependent and relatively long-lived while inhibition of muscle nAChRs by hexamethonium occurs independent of voltage (Charnet et al., 1992).

The ganglionic blocker chlorisondamine (Plummer er al., 1955) also appears to act via a use-dependent mechanism for both cultured ganglionic neuronal nAChRs (Amador and


Dani, 1995) and striatal synaptosome preparations (El-Bizri and Clarke, 1994). Similar to the inhibition of neuronal nAChR by hexamethonium, chlorisondamine inhibition of nAChR at the frog neuromuscular junction also seems to exhibit a dependence on subsequent application of ACh for recovery consistent with trapping of the inhibitor in the ion channel pore (Neely and Lingle, 1986).

In contrast to the general agreement in the literature about the mechanism of action of chlorisondamine and hexamethonium, there is less consensus about the mechanism of action of the ganglionic blockers TMP (Spinks and Young, 1958) and mecamylamine (Stone et al., 1956). In particular, although mecamylamine is arguably the most widely used inhibitor of nicotinic receptors, accounts of the mechanism of action of mecamylamine vary greatly. Based on the fact that mecamylamine does not effectively compete with ACh or nicotine for nicotinic binding sites, it is surmised that the action of mecamylamine is noncompetitive. However, Ascher et al. (1979) conclude that the effects of mecamylamine are largely competitive based on the dual observations that inhibition decreases at high agonist concentrations and that inhibition appears for the most part voltage-independent. Similarly, Bertrand et al. (1990) report a relatively long-lived block of 4132 receptors by mecamylamine which appears voltage-independent; however, these authors also report that inhibition requires the co-application of agonist indicating a noncompetitive effect. Varanda et al. (1985) report a strictly noncompetitive inhibition at neuromuscular junction nAChR. In addition, the range of reported IC50s for the effects of mecamylamine vary from about 40 nM (Ascher et al., 1979; Fieber and Adams, 1991) to 1 pM (Connolly et al., 1992). There is also some evidence that differences in sensitivity are associated with the particular beta subunit expressed. Specifically, Cachelin and Rust (1995) report an increased sensitivity of beta4-containing receptors relative to beta2-containing receptors. Studies examining mecamylamine inhibition of heterologously expressed ca3134 receptors in this laboratory indicate that inhibition is strongly voltage-dependent (Webster et al., unpublished observations).


The ganglionic blockers TMP and PMP were originally developed as more potent and less toxic alternatives to mecamylamine for the treatment of hypertension (Spinks and Young, 1958). However, as more effective adrenergic blockers were developed soon after, extensive characterization of this class of blockers did not occur. The observation that an analogue of TMP, bis-TMP-10 or BTMPS, used as a light and radiation stabilizer in medical plastics, functions as an extremely potent and selective use-dependent inhibitor of neuronal nAChRs and a less potent inhibitor of voltage-gated calcium channels has renewed interest in the TMP family of ganglionic blockers (Glossmann et al., 1993; Papke et aL, 1994). Bis-TMP- 10 is a member of the bis-TMP-n series of compounds which share a common structure consisting of a symmetrical diester of tetramethyl piperidinol rings linked by an aliphatic diacid chain containing n carbons (Figure 1-1). Neuronal nAChRs exhibit prolonged inhibition after co-application of BTMPS with ACh while muscle nAChRs recover completely within five minutes. Both subtypes exhibit only very short-term inhibition in response to co-application of the monofunctional inhibitor TMP with ACh indicating that the presence of two piperidinol rings may be critical for the conversion from short-term inhibition to long-term inhibition. The inhibition by bis-TMP10 exhibits an IC50 of about 200 nM for the open state of heterologously expressed a304 receptors. Substitution of a neuronal beta subunit for the muscle betal subunit increases the time course of recovery from inhibition consistent with a role for the beta subunit in determining sensitivity to long-term inhibition (Papke et al., 1994).

Insight into a possible basis for the selectivity of ganglionic blockers may be gained

from consideration of a previously described mechanism for inhibition of muscle nAChRs. Neher and Steinbach (1978) were able to demonstrate that the effects of application of the charged local anesthetic derivatives of lidocaine, QX-222 and QX-314, on the responses of muscle nAChR to suberyldicholine were consistent with a sequential channel block scheme. When these inhibitors are coapplied with ACh, the square pulses of current

HN 0
N 0
2,2,6,6-Tetramethylpiperidine 2,2,6,6-Tetramethyl-4-piperidinyI decanoate TMP ATMP-1 0

0 NH

4 0 0 0 0 -r (CHA-2
0 VHN 0

B is(2,2,6,6-tetramethyl-4-pi pe rid inyl) decanedioate

Bis-TMP-10 Bis-TMP-n

Figure 1-1. The TMP series of monofunctional and bi-functional inhibitors. Bis-TMP-n
compounds with linkers of 4, 6, 8, 10 and 12 carbons are included in the study.


normally observed for single channel openings are divided into smaller bursts as a result of blocking events. The time-course of these bursts are as long or longer than the mean channel open time for responses in the presence of agonist alone. This result implies that, for these inhibitors, both the binding and unbinding of inhibitor is specific for the open state of the channel. Furthermore, inhibition by these compounds is voltage-dependent providing further evidence that they act at a site within the ion channel pore. It should be noted that there may also be lower affinity sites for the binding of these inhibitors, as the utility of the sequential channel block scheme for describing inhibition by QX-222 is not maintained at high inhibitor concentrations (>40 jiM) (Neher, 1983).

As inhibition by hexamethonium and chlorisondamnine is consistent with open-channel block, it follows that sensitivity to inhibition by these compounds may be determined by sequence within the pore-lining region of the channel itself. However, the mechanism of inhibition of mecamylamnine and TMP-related compounds has been described less completely. It may be the case these compounds also function as open channel blockers. Alternatively, it may be the case the specificity of these ganglionic blockers arises via sequence elements located outside the pore-lining domains and these compounds function as either allosteric inhibitors or state-dependent, voltage-independent inhibitors. Consideration of these possibilities requires a more detailed discussion of the structure of nAChR.

Structure of nAChRs

Ligand-gated ion channels (LGICs) can be thought of as modular proteins consisting of three major components each with an associated high affinity site for the binding of ligands: 1) the receptor portion which contains the agonist binding sites and is located extracellularly, 2) the ion channel portion which contains a high affinity site for the binding of noncompetitive inhibitors and spans the cell membrane and 3) the intracellular domain which is essential for functional regulation by second messengers and for cytoskeletal


interactions. Muscle-type nAChRs are the most well characterized of the LGLCs. As most information about the structure of neuronal nAChR is based on the extensive studies of the muscle-type nAChR, structure of muscle nAChR will be considered first and comparisons to neuronal nAChR will follow.

General Structural Features

Imaging of the nAChR shows the five subunits of nAChR organized symmetrically around a central pore (Unwin, 1993). The receptor consists of a vestibule about 30 A in width which narrows as it spans the membrane to a diameter of about 8- 10 A and subsequently widens again at its intracellular extent. The length of the receptor is about 120 A with the larger portion situated extracellularly while the intracellular portion only extends about 15 A into the cytoplasm (Figure 1-2A). The ordering of the subunits around the pore remains a matter of some controversy. While it has been shown conclusively that the alpha subunits are adjacent to the delta and gamma subunits to form the agonist binding sites of muscle-type nAChRs (Blount and Merlie, 1989; Pedersen and Cohen, 1990; Sine, 1993; Sine and Claudio, 199 1), the orientation of the beta subunit with respect to the alphagamma and alpha-delta pairs is less well characterized. On the basis of electron microscopy of biotinylated neurotoxin binding to the alpha subunits and the study of induced I3-p and native 8-5 dimerization of Torpedo receptors, Karlin (1983, 1995) and Machold (1995) proposed that the muscle beta subunit lies adjacent to the delta subunit while the receptor models of Unwin (1995) place the beta subunit between the two alpha subunits. However, the latter proposed subunit arrangement poses a symmetry problem with regards to the polarity of the alpha subunit interfaces with gamma and delta subunits. It seems likely that the sidedness of the alpha subunit interaction with the gammna or delta subunit would remain consistent around the receptor as it would be presumed that very specific intersubunit contacts are required for the formation of a functional agonist binding site.

Figure 1-2. Cartoons depicting receptor structure (A) and putative membrane topology
(B) for nAChRs. In each case, the encircled area denotes the putative location of the sequence elements which are the focus of this work (TM2: second transmembrane domain; ECL: extracellular loop).
The receptor dimensions are taken from Unwin, 1993, The nicotinic acetylcholine receptor at 9 A resolution. J. Mol. Biol., 229(4): p. 1101-24. The cartoons are the work of Dr. Roger L. Papke.


A 80

60 a

40 A

00000000 0000000.0 +





Each subunit of the pentameric muscle nAChR shares a characteristic configuration (for review see (Karlin and Akabas, 1995)). Hydrophobicity analysis suggests a membrane topology consisting a large hydrophilic N-terminal putatively extracellular sequence followed by four hydrophobic, putative transmembrane domains, with a large cytoplasmic domain between transmembrane domains three and four (Figure 1-2B) (Claudio et al., 1983; Devillers-Thiery et al., 1982; Noda et al., 1983). The results of a number of studies are consistent with this proposed topology. The second transmembrane domain of each subunit seems to contribute to the lining of the ion channel pore (for discussion, see below). It has been presumed that each of the four putative transmembrane domains of each subunit would have the form of an a-helix. However, somewhat surprisingly, imaging of two-dimensional crystalline arrays of Torpedo nAChR at a resolution of 9 A has demonstrated that only the second transmembrane of each subunit is helical while the surrounding transmembrane domains seem to form beta structures (Unwin, 1995). N-Terminal Domain and Agonist Binding Site

General structural requirements of a ligand binding site can be inferred from the structure of the particular ligand. Relevant features of the ACh molecule include a positively charged ammonium moiety and the carbonyl oxygen of the acetyl group separated by a distance of about 5.9 A. It can be readily appreciated that the ACh binding site might include both a negative subsite and a polar subsite with a hydrogen donor function separated by an appropriate distance. Most evidence is consistent with this idea although the fact that tetramethylammonium (TMA) alone can function as a very low efficacy agonist for muscle nAChR and a much higher efficacy agonist for neuronal nAChRs (Papke et al., 1996) indicates that the hydrogen bond acceptor group is not strictly required. The agonist binding site of muscle nAChR is comprised of elements contributed by each of the alpha subunits with either the gamma or delta subunits such that two distinct sites are present on each receptor. Photoaffinity labeling of Torpedo nAChR by the competitive antagonist derivative d-tubocurarine (d-TC) yields covalent incorporation into


each of the alpha, gamma and delta subunits (Pedersen and Cohen, 1990) indicating that each of the gamma and delta subunits contributes to a single binding site with a partner alpha subunit. However, activation and binding properties of nAChR show a Hill coefficient in the range of 2 indicating cooperativity or nonequivalence between acetylcholine binding sites. The molecular basis for this nonequivalence is associated with the specific subunit contacts of a particular alpha subunit within the pentamer (Blount and Merlie, 1989; Sine and Claudio, 1991). Interestingly however, while the a-y interface provides the higher affinity site for d-TC, the a4 interface provides the higher affinity site for the agonist carbamylcholine and is presumed to be the higher affinity site for the binding of acetylcholine (Sine and Claudio, 199 1).

Three distinct portions of the alpha 1 linear sequence contribute residues that are labeled by derivatives of agonists and competitive antagonists (for review see (Changeux et aL, 1992)). It is therefore hypothesized that these regions of alpha subunit linear sequence must come together in the mature receptors to contribute to one subsite of an ACh binding site. These clusters of contributing residues have been referred to as "loops" by some investigators. Specifically, the vicinal cysteine residues Cys-192 and Cys-193 with aromatic residues Trp-86, Tyr-93, Trp-149, Tyr-190 and Tyr-198 are believed to contribute to one subsite of an ACh binding site. It is interesting to note that all of these residues are conserved across muscle and neuronal alphas (with the exception of alpha5) and across species (with the exception of human alpha7). It is hypothesized that negatively charged and aromatic residues from the delta or gamma subunit (e.g., the homologous residues gamma Trp-55 or delta Trp-57) contribute to the second subsite. On a more gross level, images of Torpedo nAChR indicate depressions in the alpha subunits surrounded by three rods which are presumed to represent a-helices at the approximate level of the ACh binding site (Unwin, 1995).

The activation properties of neuronal nAChR by acetylcholine also exhibit a Hill

coefficient of about 2 and both alpha and beta subunits contribute to activation by agonists


and sensitivity to competitive antagonists (Hussy et al., 1994; Luetje and Patrick, 1991; Papke et al., 1993; Papke and Heinemann, 1994). Therefore, it is presumed that, like the agonist binding sites of muscle nAChRs, the agonist binding sites of neuronal nAChRs lie at the subunit interfaces. Recent studies of the structural requirements for agonist binding have taken advantage of the fact that the alpha7 subunit forms a homomeric receptor. Chimeric exchanges between the homomeric serotonin receptor (5HT3) and the homomeric

7 receptor demonstrate that the necessary requirements for agonist binding are contained within the N-terminal 194 amino acids of the alpha7 subunit (Eisele et al., 1993). Additionally, within this stretch of amino acids there are three consensus sites for glycosylation. Disruption of these sites by mutagenesis does not affect receptor homooligomerization or protein surface expression but does affect the expression of functional a-BTX binding sites (Chen et al., 1998). Thus, it seems to be the case that receptor glycosylation is also an important determinant for the formation of a functional agonist or binding site.

Transmembrane Domains 1-4 (TM1-4) and Cytoplasmic Loop

From hydrophobicity analysis, it is possible to assess which portions of amino acid sequence could potentially contribute to transmembrane domains. However, transmembrane domains which function to line the ion channel also contain polar residues, the side chains of which are able to interact with permeant ions. Because charged ions in solution are partially hydrated, most theories of pore function require the side chains of amino acids forming the pore lining to interact with or possibly even substitute for the waters of hydration surrounding the permeant ion (Hille, 1992). A full understanding of the process of ionic selectivity and permeability will required detailed description of the structural elements contributing the pore-lining domain.

Much of the initial work in describing the three-dimensional arrangement of the putative transmembrane domains relied on using photoreactive derivatives of noncompetitive


inhibitors as photoaffinity labels for the Torpedo receptor and subsequent peptide mapping and sequencing. One of the first pieces of evidence that each subunit contributes to a common pore-lining domain comes from the observation that, in the presence of agonist, the inhibitor chlorpromazine labels residues located on each of the four nonidentical subunits of the nAChR and that this labeling can be reduced by application of the noncompetitive inhibitor phencyclidine (Oswald and Changeux, 1981). Purification and trypsin cleavage of the delta subunit followed by HPLC fractionation of the peptides allowed partial peptide sequencing and delta subunit Ser-262 was found to incorporate label (Giraudat et al., 1986). Subsequently, homologous residues in the beta subunit (Ser-254 and Leu-257 (Giraudat et al., 1987)), alpha subunit (Ser-248 (Giraudat et al., 1989)) and gamma subunit (Thr-253, Ser-257, and Leu-260 (Revah et al., 1990)) were also identified as incorporating label. All of these residues lie on homologous portions of each subunit in a hydrophobic region, the putative second transmembrane domain. Many of the same residues are labeled by the noncompetitive antagonist triphenylmethylphosphonium (TPMP+) providing further evidence the homologous regions of each subunit contribute to a single binding site (Hucho, 1986; Oberthur et al., 1986). In contrast to these results, photolabeling by other NCI derivatives has been incorporated into TM 1 and the extracellular loop region between TM2 and TM3. Specifically, an alkylating derivative of the desensitizing noncompetitive antagonist meproadifen shows incorporation of label at alpha subunit Glu-262, a site predicted to lie in the extracellular loop region between TM2 and TM3 (Pedersen et al., 1992) while the photoreactive NCI derivative quinacrine azide incorporates within a hydrophobic region corresponding to the putative TM I region of the Torpedo alpha subunit (Cox et al., 1985; DiPaola et al., 1990).

Additional evidence that the TM2 region of each subunit contributes to the lining of the ion channel pore comes from examination of the effects of site-directed mutagenesis on channel conductance and the binding of open-channel blockers. By altering the charge of particular residues (i.e., mutating negatively charged amino acids to positively charged or


neutral amino acids) located at homologous positions of the various subunits of Torpedo nAChR, Lmoto et al. ( 1988) demonstrated that the rate of ion transport through the channel is regulated by three rings of negatively charged and glutamine residues designated the intracellular, intermediate and extracellular anionic rings. These rings are situated adjacent to the hydrophobic TM2 region as part of the intracellular linker region between TM 1 and TM2, within TM2 itself and as part of the extracellular linker region between TM2 and TM3 respectively. Additionally, mutation of these rings of negative charge can influence sensitivity to reduction of current flow by the presence of either extracellular or intracellular magnesium (Imoto et al., 1988). The sidedness of the magnesium effect was used to confirm the presumed orientation of the receptor with respect to its synaptic and cytoplasmic domains. It is also of interest to note that clusters of positively charged amino acids adjacent to the internal and external rings of negative charge do not affect channel conductance. This observation is consistent with the presumed a-helical structure of the TM2 domain because adjacent residues would be predicted to face away from the pore region by about 1000.

Examination of the effects of mutation of residues in the putative TM2 region on the binding of open-channel blockers has provided a direct evaluation of the contribution of specific amino acids to the lining of the ion channel pore. Based on the voltage-dependence of inhibition and analysis of the opening and closing rates of single ion channels of muscle nAChR (see discussion above), it was concluded that the quaternary lidocaine derivatives QX-222 and QX-3 14 function as open-channel blockers (Neher and Steinbach, 1978). Leonard et al. (1988) were able to verify this conclusion directly by demonstrating that mutation of polar residues at homologous sites within the TM2 regions of the muscle subunits to nonpolar residues not only decreases channel conductance but also decreases the residence time and equilibrium binding affinity of QX-222 for the open state of muscle nAChR. This site was designated the inner polar site and is located six amino acids downstream (in the linear sequence) from the first residue of TM2 as predicted from


hydrophobicity analysis (Leonard et al., 1988). Given the proposed nAChR membrane topology, this site would lie near the midpoint of the pore closer to the inner mouth of the channel. The voltage-dependence of block indicates that the QX-222 binding site experiences about 78% of the membrane electric field consistent with a site deep in the pore. Subsequent studies demonstrated that similar mutation of poiar residues at homologous positions on each subunit of nAChR located four residues downstream (extracellular) to the inner polar site increases the residence time of QX-222 in the pore (Charnet et al., 1990). It was proposed that QX-222 interacts with residues in adjacent helices of TM2 via binding of the charged ammonium moiety to residues at the inner polar site in conjunction with hydrophobic interactions of the aromatic tail with nonpolar residues at a site located more extracellularly in the pore. Based on these studies and the predictions from hydrophobicity analysis, a system of nomenclature for the pore-lining region has been proposed which numbers the consecutive residues of TM2 from F to 20'. According to this nomenclature, the inner polar site corresponds to position 6' while the site of interaction of the aromatic portion of QX-222 would lie at position 10'.

More recent studies employing mutation of single residues of interest to cysteine and subsequent examination of availability for covalent modification by small, charged sulthydryl-selective reagents, a process known as substituted-cysteine accessibility method or SCAM, has provided further insights into accessible residues in both the open and closed states of the receptor (Akabas and Karlin, 1995; Akabas et al.. 1994; Akabas et al., 1992; Zhang and Karlin, 1997). These studies have attempted to examine higher-order structure of the pore-lining domain. Of particular interest is the location of two elements presumed to be critical for ion channel function: the channel gate and selectivity filter. From image reconstruction of pseudochrystalline arrays of Torpedo receptor, Unwin (1995) hypothesized that a kink in the middle of the TM2 region at the level of a pair of highly conserved leucine residues (9' and 10') could function as the channel gate. However, SCAM analysis indicates that alpha subunit residues as deep in the pore as E241


(position -1') are'accessible even in the closed state of the channel and that the pattern of accessibility of residues throughout TM2 is consistent with an interrupted a-helical structure containing a gate at the cytoplasmic end. It may be the case that the interruption of the a-helical segment detected by SCAM analysis is analogous to the kink in TM2 detected by Unwin. Additional SCAM analysis of alpha subunit TM 1 indicates that N-terminal residues of this domain may also contribute directly to extracellular mouth of the ion channel.

Electrophysiology studies of receptors incorporating mutant subunits have highlighted the importance of highly conserved 9' leucine residues in regulating channel gating (Filatov and White, 1995-, Kearney et al., 1996; Labarca et al., 1995). These studies indicate that mutation of this residue to a polar residue decreases the EC50 for acetylcholine most likely by stabilizing the open state of the channel through polar interactions with the hydrated ions of the pore. Similar effects were noted upon mutation of the homologous residue of the homomeric a7 receptor (L247T). However, these effects were interpreted in terms of reduction in channel desensitization such that the mutant channels exhibit a desensitized but conducting state (Revah et al., 199 1).

Because each family of ion channels exhibits a characteristic ionic selectivity, it is also reasonable to hypothesize that particular elements of the pore-lining domain may contribute to a selectivity filter which would function to dictate which ions can permeate the channel either via steric or electrostatic mechanisms. As discussed above, charged residues have been shown to be critical for determining the conductance of the muscle nAChR. Somewhat surprisingly however, the presence of charged residues in the pore-formning domains alone does not seem to be sufficient to determine the selectivity for anions versus cations. By exchanging residues between the cation-selective a7 receptor and the anionselective GABAA receptor, Galzi et al. (1992) demonstrated that insertion of a proline residue in the TM 1 -TM2 linker seems to be critical for converting selectivity from cationic to anionic. While receptors incorporating only this insertion are nonfunctional, pairing of


the proline insertion with mutation of two other pore-lining amino acids to the homologous GABA receptor residues produces an anion-selective receptor. Because mutation of these residues without insertion of the proline residue in the TM1-TJM2 linker was insufficient to confer anion selectivity, it may be the case that insertion of the proline alters the orientation of the transmembrane helices. It seems likely that the pattern of exposure of amino acids in the transmembrane helices together with the character of the exposed amino acids serves to regulate selectivity for anions versus cations.

A closely related question concerns the mechanism by which the permeability to

different ions of the same charge species is regulated once selectivity for anions versus cations has been established. Most studies indicate that a major determinant of ionic selectivity of this nature is discrimination on the basis of hydrated ion size. Mutation of alpha subunit Thr-244 (position 2') has a large effect on the permeability profile of muscle nAChR to large monovalent cations consistent with this region being located near a narrow region of the pore and possibly a component of the selectivity filter (Cohen et at., 1992; Imoto et al., 199 1; Villarroel et al., 199 1). It seems to be the case that permeability to divalent cations can be regulated by multiple sequence elements suggesting a more complex mechanism for determination of this property than selection on the basis of ion size alone. For the highly calcium permeable neuronal 0x receptor, it has been demonstrated that mutation of residues located either at the intracellular (E237, position -FI) or extracellular (L,254 or L255, positions 16' and 17' respectively) mouth of the channel can abolish calcium permeability. Mutations at position -1V had no other detectable effect on receptor function while mutation of residues at positions 16' or 17' seem to effect channel gating as evidenced by higher agonist potency and prolongation of response time-course (Bertrand et al., 1993).

Because the position of the extracellular loop region between TM2 and TM3 in the threedimensional arrangement of the receptor may allow for residues in this domain to play a role in transmitting conformational changes from the agonist binding site to the channel


pore, it has been suggested that this region may contribute to regulation of channel gating. In fact, mutation of an asparagine residue in the extracellular loop of homomeric a7 receptors substantially reduces current responses to agonist while binding of a-BTX remains intact (Campos-Caro et al., 1996). This residue is conserved across both homomer-forming alpha subunits and functional beta subunits of rat and human clones. Mutation of the homologous residue in beta4 and expression of the mutant subunit with alpha3 yields qualitatively similar results suggesting that this region in structural subunits in particular may be important for linking agonist binding to channel activation.

The role of the intracellular domain between transmembrane domains three and four in determining receptor function has also been characterized. This region is the most variable across subunits and contains consensus sites for phosphorylation, a finding which has suggested a role for differential subunit phosphorylation in determining receptor functional characteristics. There is evidence for both serine/threonine phosphorylation and tyrosine phosphorylation of muscle nAChRs. The most well described effect of serine/threonine phosphorylation is an increase in the rate of desensitization with phosphorylation of the gamma and/or delta subunits at a site in the intracellular domain. However, it is unclear if this effect is physiologically relevant because of the rapid action of acetylcholine esterase at the neuromuscular junction. For neuronal nAChRs, the effects of phosphorylation appear to be more heterogeneous. An increase in the response of alpha3-containing receptors of chick ciliary ganglion neurons via a cAMP-dependent mechanism has implied that phosphorylation of the alpha subunit may convert a population of receptors from a "silent" state to a "functionally available" state (Margiotta et al., 1987; Vijayaraghavan et al., 1990). In addition, some recent studies suggest that inhibition of phosphatase activity can increase rate of recovery from desensitization (Eilers et al., 1997; Khiroug et al., 1998). These results would be consistent with a model in which the desensitized state of the receptor is selectively modulated by phosphorylation. However, as these studies involve the prolonged application of nicotine rather than the endogenous agonist acetylcholine, it is


difficult to assess the significance of these findings beyond involvement in pathology related to nicotine addiction.

Relatively few studies have described a specific role for TM3 or TM4 in regulating receptor function although it has been widely speculated that these domains mediate the interactions of the receptor with surrounding lipid. Chimeric exchanges of TM3 between alpha3 and alpha7 subunits and site-directed mutagenesis of residues in TM4 have each been shown to have an effect on channel gating (Campos-Caro et al., 1997; Ortiz-Miranda et al., 1997). However, it is difficult to assign any direct functional implications to these effects.

Relating Structure to Function for Neuronal nAChRs

It has long been noted that nicotine can increase performance on some measures of

memory performance. This observation, in conjunction with the fact that a large number of cholinergic neurons are lost during the progression of Alzheimer's disease, has led some investigators to hypothesize that nicotinic systems may be involved in the processes of learning and attention (for review see Levin, 1992). Additionally, brain nAChRs have been implicated in the symptomnatology of diseases ranging from schizophrenia to nicotine addiction. The heterogeneity of potential neuronal nAChR subtypes and their possible dysfunction in disease states has provided the impetus for development of subtype-specific agonists as candidate therapeutics. However, as is the case for nicotine, many of these drugs have both agonist and antagonist effects. In order to define a profile for agonist and antagonist specificity, it will be necessary to determine the structural components of the neuronal receptor involved in each process.

This study seeks to expand our knowledge of the structural determinants of sensitivity to use-dependent inhibition for neuronal nAChRs. As noted above, bis-TMP- 10 is a bifunctional analogue of the ganglionic blocker TMP and shows selectivity for the long-term inhibition of neuronal nAChRs. The experiments described in this study characterize the


mechanism of action of bis-TMP- 10 and related analogues and, in addition, attempt to localize structural determinants of sensitivity to long-term inhibition. It is the ultimate goal of these studies to use the knowledge gained from the study of the mechanism of action of pure inhibitors to understand the structural basis for the mixed agonism/antagonism observed for certain nicotinic experimental therapeutics.


Chemicals and Synthesis

Fresh acetylcholine (Sigma; St. Louis, MO) stock solutions were made daily in Ringer's solution and diluted. All other drugs were stored at 4oC at a concentration of 100 mM in methanol for a period of no longer than two weeks before use. Bis-TMP- 10, bis(2,2,6,6tetramethyl-4-piperidinyl) sebacate (BTMPS), was obtained from Ciba-Geigy (Hawthorne, NY), and tetramethylpiperidine was obtained from Aldrich Chemical Company (Milwaukee, WI). Bis-TMP-4 (bis (2,2,6,6,-tetramethyl-4-piperidinyl) succinate was synthesized by Ciba-Geigy and obtained from Dr. H. Glossmann (Glossmann, et al., 1993). All other chemicals for electrophysiology were purchased from Sigma Chemical Company (St. Louis, MO) or synthesized. Chemicals used for the synthesis were purchased from Aldrich Chemical Company.

Compounds were synthesized by Drs. Kyung I1 Choi and Benjamin A. Horenstein in the laboratory of Dr. Horenstein. To a mixture of 2,2,6,6-tetramethyl-4-piperidinol (3.0 mmol) and one equivalent of the corresponding ester, unless otherwise specified, in 2 mL of dimethyl formamide was added 250 mg of powdered potassium carbonate. The resulting mixture was heated at 145- 150 oC for 24-72 hrs under a gentle stream of N2. After cooling, the reaction mixture was partitioned between water and methylene chloride. The organic layer was separated, washed with water and brine, dried (anhydrous MgSO4) and evaporated to dryness to give a crude product, which was purified via salt formation (HCl or acetic acid), extraction, or column chromatography. Compounds were characterized by mass spectrophotometry and high resolution nuclear magnetic resonance. All compounds were recrystallized before use in electrophysiology experiments.



Production of Chimeras and Sequencing All chimeric genes were constructed by the method of overlap extension PCR (Horton et al., 1989). The beta subunit chimeras were designed and produced in large part by Wayne Gottlieb and Ricardo Quintana. The genes encoding the beta 1, beta4, gamma and delta subunits were cloned into p-Bluescript SK-. Specific PCR primers were designed to generate mutants exchanging just the bases necessary to code the TM2 or BCL region. Each primer contained 27 bases of the sequence flanking the TM2 or ECL sequence on one side and 24 bases that coded for the TM2 or ECL region to be exchanged. Oligonucleotides were designed to contain a unique silent restriction site in the mutant region for future screening, and synthesized by the University of Florida DNA Synthesis Core. Separate PCR reactions consisting of the appropriate PCR primer with template and either T3 or T7 primer selectively amplified the upstream and downstream portions of the gene of interest with overhanging chimeric sequence. These two products were then put together in a second PCR reaction with T3 and T7 primers. The region of chimeric sequence overlap formed double stranded DNA that primed elongation in both directions, and the full length product was amplified using T3 and T7 primers. The region coding for mutant sequence was then cut out with restriction enzymes and cloned back into the original plasmid, reducing the amount of PCR-generated sequence in the final constructs. Clones were evaluated by both restriction analysis and sequencing through the PCR generated region either by the dideoxy chain termination method (Sanger et al., 1977) using the Sequenase 2.0 kit from United States Biochemical Corporation (Cleveland, Ohio) in the laboratory of Dr. Jeffrey K. Harrison or by automated fluorescence sequencing in the University of Florida DNA Sequencing Core.

Preparation of RNAs and Gocyte Expression

In vitro cRNA transcripts were prepared using the appropriate mMessage mMachine kit from Ambion Inc. (Austin, TX) after linearization and purification of plasmids containing


cloned cDNAs. RNA transcripts were stored at -80 'C as water stocks at a concentration of either 200 ng/gl or 600 ng/gd. Concentrations were determined from measures of percent incorporation of 32p-labeled UTP via scintillation counting.

Ovarian lobes were surgically removed and then cut open to expose the oocytes. The ovarian tissue was then treated with collagenase from Worthington Biochemical Corporation (Freehold, NJ) for about 2 hours at room temperature (in calcium-free Barth's solution: 88 mM NaCl, 10 mM HEPES pH 7.6, 0.33 mM MgS04, 0.1 mg/mi gentamicin sulfate). Subsequently, stage 5 oocytes were isolated and injected with 50 nl each of a mixture of the appropriate subunit cRNAs following harvest. Barth's solution was changed daily under semi-sterile conditions. Recordings were made 2 to 7 days after injection depending on the cRNAs being tested.


Two-Electrode Voltage Clamp

Initial recordings were made on a Warner Instruments (Hamden, CT) OC-725C oocyte amplifier and RC-8 recording chamber interfaced to a Macintosh personal computer, while the majority of experiments employed a Gene Clamp 500 amplifier (Axon Instruments; Foster City, CA) interfaced to a Gateway 2000 (N. Sioux City, SD) P5-75 personal computer. Comparable results were obtained on both sets of equipment. Initial experiments were performed in a configuration such that a 2 ml bolus of drug was applied after loading of a loop at the terminus of the drug delivery system, while subsequent experiments were conducted in a configuration where drug application was electronically controlled and regulated by duration rather than volume, permitting more rapid solution exchange without stoppage of flow through the chamber. Oocytes were placed in a Lexan recording chamber with a total volume of about 0.6 mi and perfused at room temperature by frog Ringers (115 mM NaCl, 2.5 mM KC1, 10 mM HEPES pH 7.3, 1.8 mM CaC12) containing 1 giM atropine to block potential muscarinic responses. A Mariotte flask filled


with Ringers was used to maintain a constant hydrostatic pressure for drug deliveries and washes. Drugs were diluted in perfusion solution and applied from a reservoir for 10 seconds using a 2-way electronic valve. Data were acquired using Axoscope 1.1 software (Axon Instruments; Foster City, CA) at a 20 Hz sample rate and filtered at a rate of 10 Hz using either a CyberAmp 320 external filter (Axon Instruments; Foster City, CA) or the filter in the amplifier. The rate of drug application and perfusion was 6 ml/min in all cases. Current electrodes were filled with a solution containing 250 mM CsCl, 250 mM CsF and 100 mM EGTA and had resistances of 0.5-2 ML. Voltage electrodes were filled with 3 M KC1 and had resistances of 1-3 M92. Oocytes with resting membrane potentials more positive than -30 mV were not used.

Cut-Open Oocyte Vaseline-Gap Voltage Clamp

Experiments were conducted using the modified chamber described by Costa et al.

(1994) and available commercially from Dagan. Measurements were made using a Dagan amplifier interfaced to a Gateway 2000 (N. Sioux City, SD) P5-75 personal computer running pClamp 7 software. For most experiments, data were acquired at a rate of 200 Hz and filtered at a rate of 20 Hz using the filter in the amplifier. Ringers solution was perused to the external face of the oocyte by gravity flow from a Mariotte flask while an internal solution consisting of 100 mM KCI, 10 mM HEPES, and 10 mM EGTA (pH 7.4) was perfused to the internal face of the oocyte by syringe pump (World Precision Instruments) at a rate of 0.01 ml/hr. Agonist was applied via a modified U-tube controlled by an electronic 2-way valve (General Valve Corporation) activated digitally by the computer. An insert in the recording chamber produced a region of laminar flow across the exposed portion of oocyte membrane in the chamber allowing for relatively rapid application of agonist. Internal perfusion pipettes were pulled to a length of about 4.7 cm and broken to a diameter of >100 gtm to allow adequate perfusion and reduce series resistance. Internal


perfusion pipettes were then coated with a mixture a paraflm and mineral oil to prevent leaking of solution between the pipette and the floor of the chamber.

Experimental Protocols and Data Analysis

For the majority of experiments, current responses to drug application were studied

under two-electrode voltage clamp at a holding potential of -50 mV unless otherwise noted. Holding currents immediately prior to agonist application were subtracted from measurements of the peak response to agonist. All drug applications were separated by a wash period for a length of time as noted (five minutes in most cases). At the start of recording, all oocytes received an initial control application of ACh to which subsequent drug applications were normalized in order to control for the level of channel expression in each oocyte. An experimental application of ACh with inhibitor was followed by an application of ACh alone ten minutes after the control ACh application used for normalization. In some receptor subtypes (e.g., O3P4), rundown was observed to stabilize after a second application of ACh. For these subtypes, responses were normalized to the second of two initial control ACh applications. Means and standard errors (SEM) were calculated from the normalized responses of at least 4 oocytes for each experimental concentration.

For all experiments involving use-dependent inhibitors, a concentration of ACh was

selected sufficient to stimulate the receptors to a level representing a reasonably high value Of Ppen at the peak of the response, while minimizing rundown with successive ACh applications. For potent use-dependent inhibitors, this concentration is adequate to achieve maximal inhibition (Papke et al., 1994). Specific concentrations for each receptor subtype are as noted.


For concentration-response relations, data were plotted using Kaleidagraph 3.0.2 (Abelbeck Software. Reading, PA) and curves were generated using the following modified Hill equation (Luetje and Patrick, 199 1)

Response = 1I,,a[agonist]fl
[agonist]n + (EC5O)n

where Imax denotes the maximal response for a particular agonist/subunit combination, and n represents the Hill coefficient. 1ma, n, and the EC50 were all unconstrained for the fitting procedures, and the r values of the fits were all > 0.96 (the average r value = 0.97).

For use-dependent inhibitors, measurements of peak response at the time of coapplication of agonist with inhibitor underestimate steady-state inhibition in our system. Therefore in experiments assessing rate of recovery from use-dependent inhibition, inhibitor alone was pre-applied for a length of time sufficient to achieve maximal concentration in the chamber prior to application of agonist. This pre-application protocol maximized the probability of block upon channel activation as a function of inhibitor concentration and permitted the use of peak current as a more accurate measure of steadystate inhibition for applications of ACh in the presence of the TMP compounds. After normalization to the control response (as described above), total inhibition can be calculated by subtracting the normalized value from 1. In this manner, nearly complete inhibition at the time of co-application of agonist with inhibitor is observed and a recovery rate from this point in time can be estimated. Agonist concentrations were selected in order to minimize rundown with successive ACh applications while still providing a high enough probability of channel activation during the time-course of a response to achieve close to 100% inhibition, and are noted in the figure legend.


Recovery rate data were fitted by the equation

% Inhibition = Io(e-th)

which describes a first order process where Io represents the inhibition at time t=0 and 'r is the time constant for recovery. The r values of the displayed fits were all > 0.96 (the average r value = 0.98).

For experiments assessing voltage-dependence of inhibition, oocytes were initially voltage clamped at a holding potential of -50 mV and a control application of ACh alone was delivered. The holding potential was stepped to +20 mV for 30-60 s prior to coapplication of either ACh with bis-TMP-10 or ACh alone. Thirty to sixty seconds after the peak of the co-application response, voltage was stepped back down to -50 mV and residual inhibition was evaluated with two subsequent applications of ACh alone separated by 5 minutes.

For experiments assessing protection from long-term inhibition by application of a short term inhibitor, a saturating concentration of the short term inhibitor [either QX-314 (lidocaine N-ethyl bromide) or TMP (2, 2, 6, 6-tetramethylpiperidine)] was applied for 3060 s prior to the application of ACh and the long-term inhibitor (bis-TMP- 10). The application of the short term inhibitor continued throughout the time-course of the coapplication of ACh with long-term inhibitor until at least 30 s after the co-application. The concentration of inhibitor and period of application was selected to maximize the potential for protection effects. Recovery from inhibition was evaluated in 3 minute intervals after the application of inhibitor in the case of C l P (134TM2)'8 receptors and 10 minutes after application of inhibitor in the case of a304 receptors.

In experiments using the cut-open oocyte system, data from each oocyte was normalized to an initial 5 s application of ACh in the presence of normal internal solution (100 mM KCI, 10 mM EGTA, 10 mM HEPES, pH 7.4). For experiments evaluating the effects of


intracellularly applied inhibitors, control of perfusion was then switched to a second syringe pump containing inhibitor diluted into the internal solution to a concentration of 2 g.M. Internal solution is pumped (0.01 mllhr) into the perfusion pipette via the base of the pipette holder with an approximate pipette solution exchange time of 10 minutes based on the inclusion of a dye in the solution. Experimental measurements for activation in the presence of intracellularly applied inhibitor were made in 3 minute intervals starting 15 minutes after switching to the second syringe pump.


Structural Determinants of Sensitivity to the TMP Family of Noncompetitive Inhibitors

Muscle Delta Subunit Effects

It has been previously published that a bis- analogue of the ganglionic blocker TMP

produces long-term inhibition of heterologously expressed cc304 receptors while TMP itself only produces short-term inhibition (Papke, 1994). One of the main goals of the present work is to characterize the basis for this selectivity. Although muscle nAChRs show rapid recovery from inhibition by bis-TMP-10 (within five minutes), this compound does produce appreciable inhibition of muscle nAChRs at the time of co-application with ACh while a34 receptors show pronounced inhibition at both time points (Papke et al, 1994); Figure 3-1). This observation demonstrates the presence of a high affinity bis-TMP-10 binding site(s) on neuronal nAChRs and furthermore suggests the presence of relatively low affinity site(s) for binding of bis-TMP- 10 on muscle nAChR. Moreover, because bisTMP-10 is a bi-functional molecule composed of two TMP moieties linked by an aliphatic chain and TMP itself is an effective ganglionic blocker, it may be the case that the time course of recovery from inhibition by bis-TMP-10 is determined by the number of available TMP binding sites per pentameric receptor. In this case, the long-term inhibition of neuronal receptors may be a result of the presence of multiple neuronal beta subunits in the receptor pentamer while the short-term inhibition of muscle receptors may be associated with the presence of only a single TMP binding site. It has been shown previously that expression of either the neuronal beta2 or beta4 subunit with the other muscle subunits prolongs the time course of recovery from inhibition by bis-TMP- 10. This result may


Figure 3-1. Long-term inhibition by bis-TMP-10 is specific for neuronal nAChR subunit combinations. Representative traces are shown in A while mean data are shown in B. A) In each set of traces, the responses on the far left and far right are to control applications of 30 gM ACh alone while the middle trace is the response to coapplication of ACh with 2 giM bis-TMP-10. All responses are separated by five minutes. B) All mean data are expressed relative to the initial control response. Each column represents the mean of at least 4 oocytes. Similar results have been published previously by Papke et al., 1994.


B 1.25T


30 jiM ACh 30 jiM ACh alone
with 2 p.M bis-TMP- 10 five minutes later


imply that a single site for TMP binding is present on the wild-type muscle receptor and substitution of a neuronal beta subunit for the muscle beta subunit provides a second potential TMP binding site. To characterize structural elements of muscle subunits which may contribute to a site for bis-TMP-10 binding, muscle-type receptors lacking either the gamma or delta subunit were expressed and characterized according to their sensitivity to inhibition by bis-TMP-10. aip18 and al Ply receptors have been previously characterized in a number of studies (Charnet et al., 1992; Jackson et aL, 1990; Kullberg et al., 1990; Lo et al., 1990) and show concentration-response relationships typical of nAChRs in our system as well with EC50s for activation of 35 and 25 piM respectively (not shown). Because the agonist binding sites for muscle nAChRs are located at the interface of the alpha subunits with the gamma and delta subunits, it is presumed that the receptors formed by omission of either gamma or delta subunit RNA contain two copies of the included subunit. For example, a1P18 (gamma-less) receptors would have a subunit stoichiometry of 2:1:2.

Contributions of Delta Subunit to Inhibition by Bis-TMP-10

Muscle-type and az1f 1y receptors show nearly complete recovery from inhibition by coapplication of bis-TMP- 10 with ACh within five minutes, while a11318 receptors display prolonged inhibition as measured five minutes after co-application of ACh with bis-TMP10 (Figure 3-2A). The mean response of a1il1i receptors to application of ACh alone five minutes after co-application with bis-TMP-10 is 2403% of the response to the initial control application of 10 pM ACh (n=8) while the corresponding responses of al1P1y (n=4) and wild-type muscle receptors (n=6) are near control levels. Omission of gamma subunit RNA seems to have effects on the time-course of recovery from inhibition by bis-TMP-10 comparable to the effects seen previously after substitution of neuronal beta subunit RNA (beta4 or beta2) for muscle beta subunit RNA (beta 1) (Papke et al., 1994).

Figure 3-2. The blockade of aPS nAChRs by bis-TMP-10 (A) and TMP (B). The cluster of bars on the left represents the mean peak responses (SEM) of oocytes to the coapplication of 10 gM ACh and the inhibitor, while the cluster of bars on the right represents the normalized responses to 10 gM ACh alone after a 5 minute wash period. The response of each oocyte was normalized to an initial response to 10 giM ACh applied 5 minutes prior to the co-application of ACh with bis-TMP-10.


A 1.25- 2gM Bis-TMP-10

2T 0T


0 I
10 tM ACh 10 giM ACh alone
with 2 pM bis-TMP-10 five minutes later

[L] u(py8 (Wild-type muscle) g a 3y (Delta-less muscle)
0 a38 (Gamma-less muscle)






10 gM ACh with 10 gM ACh alone
4 nM TMP five minutes later


Contributions of Delta Subunit to Inhibition by TMP

ali1 receptors also show increased sensitivity to the monofunctional inhibitor TMP (2, 2, 6, 6-tetramethylpiperidine) at the time of co-application as compared to wild-type muscle and aip1ly receptors (Figure 3-2B). The mean response of a118 receptors to coapplication of 10 gM ACh with 4 gM TMP is about 60% of the control response to an initial application of ACh alone (n=4) while the responses of normal muscle-type receptors and all1y receptors are near control levels. As these data are for a single concentration of TMP only, the possibility that the difference in sensitivity observed in these experiments simply reflects a shift in the concentration dependence of the effect cannot be ruled out. In fact, although an effect of TMP on wild-type muscle receptor is not observed at the concentration used in this study, muscle-type nAChRs do exhibit a weak sensitivity to short-term inhibition by this compound. The IC50 for short-term inhibition of muscle receptors by TMP appears to be in the range of 40 pM, a concentration about 200 times higher than the reported IC50 for long-term inhibition of a3p4 receptors by bis-TMP-10 (Kabakov and Papke, unpublished observations). Although ac1018 receptors do exhibit an increased sensitivity to short-term inhibition by TMP compared to wild-type muscle receptors, it is interesting to note that a corresponding sensitivity to long-term inhibition by TMP is not observed. Thus, the effects of TMP are strictly short-term in nature and seem to be equivalent to the effects of bis-TMP- 10 on wild-type muscle or aft receptors. This observation is consistent with the results of a previous study in which neuronal beta subunits were substituted for muscle beta subunits in that, for both cases, although sensitivity to short-term inhibition by TMP increases with incorporation of a TMP sensitive subunit, long-term inhibition only occurs with application of bi-functional TMP compounds such as bis-TMP- 10. The requirement for a bi-functional molecule may imply that a secondary process associated with interactions between the receptor and the aliphatic linker or second TMP moiety of bis-TMP-l0 contributes to the time course of recovery from inhibition.


Contributions of Delta Subunit to Inhibition by Mecamylamine

In order to have a context for consideration of the effects of TMP-related compounds, similar experiments assessing sensitivity to the ganglionic blocker mecamylamine were conducted. Omission of gamma subunit RNA also leads to increased sensitivity to inhibition by mecamylamine (Figure 3-3A). Additional experiments assessing the contribution of beta subunits to inhibition by mecamylamine indicate that, as expected for a ganglionic blocker, 304 neuronal receptors show more pronounced inhibition by mecamylamine than any of the other subunit combinations tested (Figure 3-3B). Substitution of the neuronal beta subunit for the muscle betal subunit seems to increase sensitivity to inhibition by mecamylamine at the time of co-application slightly. However, inhibition five minutes after application of mecamylamine was detected only in the case of a34 (about 31%) and a1l310 (about 38%) receptors (Figure 3-3B). Response Kinetics Imply Mechanism of Inhibition

Responses of both muscle and neuronal nAChRs to the co-application of bis-TMP- 10 with ACh exhibit a decreased to time to peak response compared to the peak of the control response to ACh alone consistent with a dependence on prior activation of the channel for inhibition (Figure 3-4). In contrast, responses to the co-application of mecamylamine with ACh exhibit a clear decrease in time to peak only for a304 receptors. Based on these observations, inhibition by bis-TMP-10 appears to be purely-use dependent and relatively long-lived for all receptor subtypes whereas, for some subtypes of nAChR, inhibition by mecamylamine may occur via multiple mechanisms. The differences in time to peak response for the two inhibitors indicate that mecamylamine may not act by a purely usedependent mechanism on non-neuronal receptors, particularly the a i p18 receptor subtype. Alternatively, it may be the case that the off-rate of mecamylamine is rapid compared to the time-course of drug application so that the decay phase of the macroscopic response is not appreciably affected by the presence of the inhibitor. For open-channel blockers with a very fast off-rate, it would not be expected that time to peak would be affected appreciably

Figure 3-3. The blockade of a1py, a4, and muscle-type (ali3rTS) receptors (A) and muscletype, alo4yS and neuronal (3034) receptors (B) by mecamylamine. The cluster of bars on the left represents the mean peak responses (SEM) of at least 4 oocytes to the coapplication of 10 [tM ACh and the inhibitor, while the cluster of bars on the right represents the peak responses to 10 gM ACh alone after a 5 minute wash period. The response of each oocyte was normalized to its response to an initial application of 10 p.M ACh 5 minutes prior to the co-application of ACh with inhibitor represented in the figure.



10 gM ~h 10gM ki alone


10 jM A~wh 10jM ACh alone
wih10 jiM mec. five minutes later

Figure 3-4. Inhibitor effects on the kinetics of macroscopic currents. Representative waveforms of responses from a038 (A and B), wild type muscle (C and D), aikt4S (E and F) and neuronal (G and H) receptors to a pulse of 30 AM ACh alone (line 1, gray), and a co-application with either 2 4.M bis-TMP-lO (traces on left) or 10 p.M mecamylamine (traces on right) and 30 A.M ACh (line 2, black). For all the traces, the thin black line (3) plots the inhibited current scaled to the same peak value as the control in order to visualize the kinetics of inhibition. The thick bars under the traces represent the period of drug application. Note that due to the prolonged response of c4p8 receptors, the time scale in A and B is expanded.


A Bis-TMP-10: aIP18 injected oocyte, B Mecamylamine: a 3P1~ injected
gamma-less receptor oocyte, gamma-less receptor



0 30 60 90 120 0 30 60 9*0 120
Seconds Seconds
C Bis-TMP-10: a 1pl1y injected D Mecamylamine: al3iy8 injected
oocyte, Muscle-type receptor oocyte, Muscle-type receptor

3 2

1 31

0 15 30 45 60 0 15 30 45 o60
Seconds Seconds

E Bis-TMP-10: alr4ya injected oocyte, F Mecamylamine: al14yB injected
neuronal/muscle hybrid receptor oocyte, neuronal/muscle hybrid receptor


0 15 30 45 o60 0 15 io30 45 60
Seconds Seconds
G Bis-TMP-10: a34 injected H Mecamylamine: a304 injected
oocyte, neuronal receptor oocyte, neuronal receptor
2 2

3F 3

0 15 30 45 60 0 15 30 45 60
Seconds Seconds


whereas the falling phase of the response may in fact be prolonged by the presence of the inhibitor. However, the observation of long-term inhibition after application of bis-TMP10 with ACh to aPl receptors makes it seem unlikely that a rapid off-rate of the inhibitor underlies the change in response waveform. In some cases, where inhibition is clearly use-dependent, a secondary peak is observed in the waveform of the co-application response (Figure 3-4; A and H). Since this secondary peak corresponds exactly with the removal of agonist and occurs only in conditions where inhibitor is present and acting in a clearly use-dependent manner, it is likely that this peak represents relief from a short-time course, low affinity inhibition. Because this inhibition relaxes within the time course of the agonist response, it most likely represents an entirely different form of inhibition (or at least inhibition with different kinetics) from the relatively long-term inhibition discussed above.

It is interesting to note that even though muscle receptors show rapid recovery from

inhibition by bis-TMP-10, the waveform of the response to co-application of bis-TMP-10 with ACh looks very similar across all receptor subtypes tested. This similarity may indicate that the mechanism underlying the initial phase of inhibition by bis-TMP- 10 is similar across all receptor subtypes while the rate of recovery from inhibition may be related to some secondary process associated with an interaction between the inhibitor and structural elements specific to individual subunits. Furthermore, the affinity of this secondary interaction may be dependent upon the number of TMP-sensitive subunits present within a particular receptor subtype. Because inhibition by bis-TMP-10 shows the most pronounced selectivity for neuronal nAChRs and also has the longest time-course, subsequent investigations focused on inhibition by this compound. Time-course of Recovery of ap38 Receptors after Inhibition by Bis-TMP-10

In order to examine the rate of recovery from inhibition by bis-TMP-10 for a10 18

receptors, ACh was applied at time points beyond five minutes after co-application of ACh with bis-TMP- 10 and residual inhibition was evaluated (Figure 3-5). Because in this system, peak currents at the time of co-application of bis-TMP- 10 with ACh underestimate

Figure 3-5. Sequence elements on the delta subunit determine the time course of recovery from inhbition by bis-TMP- 10. A) Representative responses of #%, aft, and c4y(5'-6'8) receptors. For each receptor subtype, the trace at the far left and series of 3 traces at the far right are responses to ACh alone while the middle trace is the response to the coapplication of 30 pM ACh with 2 tM bis-TMP-10. All responses are separated by 5 minutes. B) Recovery from inhibition by bis-TMP-10 as a function of time. The data points at t=0 minutes represent a measure of total inhibition observed after coapplication of ACh with 2 p.M bis-TMP-10 (see Methods). Data points at t=5, t=10 or t=15 minutes represent mean values of residual inhibition measured after application of ACh alone. Since for ailly8 and ay receptors, nearly full recovery was effectively achieved after only 5 minutes of wash, the data at the 10 and 15 minute time points for these receptor subtypes are omitted for clarity. All drug applications are separated by five minute wash periods. The concentration of ACh in each experiment is either 10 p.M (alpiy8) or 30 p.M (alplfy, al183I and a131y(5-6'8)). All data points represent the mean responses of at least 4 oocytes and are fit with an equation describing exponential recovery (see Methods).



'/ 20 nAl 60s

apy -r Y3 A

aPY(5'-6'5) y
150 nAl


e aP8 cu

-M apY8
--B -(XPY

5 0 5 10 15 20
Time (mins.)


total inhibition at this time point, even with application of a saturating concentration of bisTMP-10, for these experiments, 2 M bis-TMP-10 alone was applied continuously for 1520 seconds prior to application of ACh in the continued presence of 2 gM bis-TMP-10. Using this protocol, nearly total inhibition at the time of co-application of 2 VM bis-TMP10 with ACh was observed and the time-course of recovery from this time point could be measured by fitting the data with a single exponential. Consistent with the observations five minutes after application of inhibitor, a11315 receptors exhibit prolonged inhibition with a time constant of recovery of about 25 minutes while aOlly receptors recover rapidly from inhibition with a time constant of about 2.5 minutes. Residues in the Delta Subunit TM2 Region Regulate Time-Course of
Recovery from Inhibition by Bis-TMP-10
Based on previous descriptions of residues in the pore-forming domains which contribute to the binding of other use-dependent inhibitors, it was hypothesized that sequence within the TM2 domain of delta subunits may regulate sensitivity to long-term inhibition by bis-TMP-10. In particular, a pair of adjacent residues at positions 5' and 6' were found to differ between the gamma and delta subunits (underlined sequence below).

intracellular MEMBRANE SPANNING II extracellular
1' 20'

In the gamma subunit sequence, a threonine residue is at position 5' while an asparagine residue occupies position 6'. For the delta subunit, the residues at these positions are isoleucine and serine respectively. As position 6' has previously been found to be important for regulating the binding kinetics of local anesthetics such as QX-222 and QX314 (Leonard et al., 1988), it is likely that amino acids at this position face the ion channel pore. Therefore, it seems reasonable to speculate that, if bis-TMP- 10 acts as an openchannel blocker, residues at this position may regulate the time-course of inhibition of by


this drug also. Additionally, it may be the case that the size of the adjacent residue also influences exposure of the amino acid. Towards testing these hypotheses, a pair of mutant cDNAs were constructed which exchange the bases coding for these two amino acids between the gamma and delta subunits. In order to assess the effects of these exchanges on the time-course of inhibition by bis-TMP- 10 without the influence of a wild-type gamma or delta, these mutant subunits were then expressed individually with wild-type muscle alpha and beta subunits only. Receptors including the delta(5'.6' gamma) double mutant either do not express sufficiently or are nonfunctional in this configuration. However, receptors including the gamma mutant subunit (z1P31Y(5'..6'8) receptors) give robust responses to application of ACh within 5-6 days after injection and exhibit concentration-response relationships typical of nAChRs. However, this class of mutant receptors a higher EC50 for activation compared to c41y or c435 receptors in the range of 135 WM (not shown). Also, a1fP1Y(5'_6') receptors show about 70% inhibition five minutes after application of 30 M ACh with 2 jiM bis-TMP- 10 (Figure 3 -5A) and recover from inhibition with a time constant of recovery of about 10 minutes (Figure 3-5B). Receptors including the wild-type alphalI, betalI and delta subunits with the gamma subunit double mutant also exhibit prolonged inhibition while receptors including wild-type alphal, betal and gamma subunits with the delta subunit double mutant recover from inhibition within five minutes (data not shown). Interestingly, receptors in which three different wild-type subunits are included with either the gamma or delta mutant subunit exhibit expression levels similar to wild type muscle receptors suggesting that all four subunit types are in fact incorporated into the pentameric receptor. These results are consistent with the hypothesis that sequence within the intracellular portion of the delta subunit TM2 region (specifically the 5' and 6' residues) contributes to the determination of the time-course of recovery from inhibition by bis-TMP10. Moreover, the fact that long-term inhibition requires the presence of at least two sensitive subunits (either wild-type delta or ganl a(5'-6delta) mutant) implies that prolonged inhibition requires contributions from sequence elements on separate subunits.


Neuronal Beta Subunit Effects

Previously published reports have indicated that substitution of a neuronal beta subunit for the muscle beta subunit confers sensitivity to long-term inhibition on the resulting a1I1PN-tS receptor (13N=neuronal beta subunit, either beta 2 or beta4). This result implies that sequence elements important for long-term binding are contained, at least in part, within the beta subunit sequence. Results from other laboratories have implicated either sequence elements in TM2 or the extracellular loop region (ECL) between TM2 and TM3 of muscle subunits in regulating the binding of particular noncompetitive inhibitors (see Introduction). Thus, these regions are likely candidates for regulation of the binding of bis-TMP-lO as well. In order to characterize the mechanism for the selective long-term inhibition of neuronal nAChRs, two pairs of chimeric beta subunits were created which exchange eight amino acids of either the TM2 or ECL region between muscle (131) and neuronal beta subunits (134 in this case). Although beta2 and beta4 are identical within this region of TM2. the beta4 subunit was chosen for these exchanges because this subunit is more prevalent in receptors of the autonomic ganglia and TMP has been demonstrated to be effective as a ganglionic blocking agent. Receptors resulting from coexpression of the beta4 subunit with alpha3 are likely to represent a reasonable approximation of one class of peripheral nAChRs.

Chimeric DNAs exchanging sequence coding for eight amino acids of the TM2 (in underline below) or ECL regions (in double underline below) between a neuronal beta subunit (134) and the muscle beta subunit (131) were constructed by overlap extension PCR (Horton et al, 1989). The TM2 chimeric region extends from position 4' to position I 1F including the position homologous to the inner polar site of Leonard et al (1988) at which the charged amino group of the local anesthetic QX-222 has been hypothesized to bind in muscle-type nAChRs (position 6'). The exchanged ECL region begins at position 18' incorporating the last three residues of the putative TM2 region and extends through the first five residues of the loop between TM2 and TM3.


intracellular MEMBRANE SPANNING II extracellular
4' 11' extracellular

These chimeric beta subunits were then expressed with the other muscle subunits (al,y,8) to produce alIl(34TM2)y8, aII34(P31TM2)yS, a l1(034ECL)y8 or atl4(IPiECL)y8 receptors. Note that only four of the eight amino acids in the underlined region (at positions 4', 6',7' and 10') differ between the two subunits in the TM2 region while five of the eight amino acids exchanged differ between the two subunits in the ECL region. The effects of these exchanges were initially evaluated in the muscle receptor because only a single beta subunit is included per receptor, while neuronal nAChRs include multiple beta subunits. As noted above, the presence of a single neuronal beta subunit in combination with the other muscle subunits has been previously shown to be sufficient for achieving long-term inhibition after co-application of bis-TMP-10 with ACh (Papke et al., 1994).

Coinjection of chimeric or wild type beta subunit RNA with RNA coding for the other muscle subunits provides for the expression of functional at1317y, aI4Y76, al l(f4TM2)y6 and a134(P31TM2)y8 receptors with activation profiles typical of nAChRs. In order to interpret data comparing the magnitude of use-dependent inhibition across receptor subtypes, it is necessary to first define a relationship between the experimental concentration of agonist applied and the EC50 for each receptor subtype. Although the EC50s for each of the receptor subtypes differ somewhat, the Hill coefficients for all of the receptor subtypes are in the range of 1-2, typical of nAChRs (not shown). While txpl31y6, a1134y8 and a1P4(31TM2)y6 receptors show comparable EC50s in the range of 3 to 8 RM, a131(04TM2)y8 receptors require about a five fold higher concentration of ACh (30 PM) for 50% activation. Based on the concentration-response studies, the concentration of ACh to be used in specific experiments was determined. This concentration ranges between 5 and


30 pM depending on receptor type or experimental design and is noted in the figure legends.

Dependence of Long-Term Inhibition by Bis-TMP-10 on Sequence in the TM2 Region

The time-course of recovery from inhibition by bis-TMP- 10 was examined for a number of different subunit combinations (Figures 3-6, 3-7, 3-8). While normal muscle-type receptors consistently recover from inhibition within five minutes after co-application of 30 ptM ACh and 2 p.M inhibitor, a1P1(fMTM2)y8 chimeric receptors show more prolonged inhibition (Figure 3-6). In order to demonstrate a reciprocal dependence of this effect, the time-course of recovery from inhibition of a134y8 receptors after co-application of 2 W.M bis-TMP-10 and 30 M ACh was compared with that of aII4( I TM2)y receptors (Figure 3-7). While a1i14 y receptors remain about 73% inhibited after five minutes, cclaI4(P1TM2)y8 receptors recover to near control levels. Additionally, coexpression of the neuronal a3 subunit with the chimeric t34(j31TM2) subunit produces receptors which recover completely from inhibition within five minutes, while normal a304 receptors remain about 93% inhibited (Figure 3-8). It should also be noted that continuous application of 2 PM bis-TMP-10 to either a I134TM2)y8 or a3 f4 receptors for up to one minute in duration without co-application of agonist does not produce any significant inhibitory effects providing further evidence of the use-dependence of this form if inhibition (data not shown).
The rapid recovery of c3134(131TM2) receptors from inhibition by bis-TMP-10

demonstrates that the neuronal c3 subunit is insensitive to long-term inhibition after application of bis-TMP- 10. The sensitivities of the neuronal a2 and a4 subunits to inhibition by bis-TMP- 10 were also evaluated. While attempts to get expression of 4I4(IITM2) receptors were unsuccessful, ox204(I31TM2) receptors recover completely within five minutes from the inhibition elicited with co-application of 30 pM ACh and 2 p.M bisTMP-10 (data not shown). Since the a4 and c2 subunits are identical within the TM2

Figure 3-6. Substitution of eight amino acids of neuronal beta subunit tm2 region into the muscle beta subunit confers sensitivity to long-term inhbiition by bis-TMP-10. Traces representative of the mean data are shown on the left. The trace marked "a" is the response to a coapplication of ACh with inhibitor while the trace marked "b" is the response to ACh alone five minutes later and is used as a measure of recovery from inhibition. To generate mean data, the coapplication (a) and recovery (b) responses are normalized to the trace marked control. Mean data are displayed in the bar graphs on the right. In each graph, the pair of bars on the left (a) represent the normalized response to a coapplication of ACh with 2 iiM bis-TMP- 10 and correspond to the trace marked "a" at left. The pair of bars on the right (b) represent the normalized response to an application of ACh alone five minutes after the coapplication response and correspond to the trace marked "b" at left. Each column represents the mean response SEM of at least 4 oocytes. All responses are separated by 5 minute wash intervals.

A B xijpiy8

control a b X OfI1(pTM2)'y8

125 nAj>

control a b r .5

ax I3 1 (I34TM2)y

60OnA L 0mp
30 s a b
co-application recovery

A B al 4y8
~ l P4(PlTM2)yi control a b 1 --- --

x1 47 20 nA

control a b 0.55

ulP4(PITM 2)y86

60 nAL
30s 0
a b
co-application recovery

Figure 3-7. Substitution of eight amino acids of the muscle beta subunit TM2 sequence into the neuronal beta4 subunit reverses sensitivity of alp4y8 receptors to long term-inhibition by bis-TMP-10. The organization of the figure is identical to that of Figure 3-6A and a description appears in the legend to that figure.

A B M a34
control a b I a3P4(131TM2)

500 nA


control a b

ITM2) 500 nA L 30 s

a b
co-application recovery

Figure 3-8. Substitution of eight amino acids of the muscle beta subunit TM2 sequence reverses sensitivity of neuronal a3134 receptors to long term-inhibition by bis-TMP-10. The organization of the figure is identical to that of Figures 3-6 and Figure 3-7. A description of the layout appears in the legend to Figure 3-6.


domains, it is hypothesized that these alpha subunits are not sensitive to long-term inhibition by bis-TMP- 10; however, it is interesting to note that 7 receptors do exhibit long-term inhibition after application of bis-TMP- 10.

For ctl31(jMTM2)-y8 receptors, examination of the relative amount of inhibition remaining five minutes after co-application of varying concentrations of bis-TMP- 10 with ACh indicates that long-term inhibition of this subtype has an IC50 in the range of 30 nM (Figure 3-9). Previous reports indicate that inhibition of c334 receptors by bis-TMP-10 has an IC50 of about 200 nM. It should be noted that, for use-dependent inhibitors, the observed IC50 is dependent upon open probability (i.e., agonist concentration). Thus, some variability in IC50 across receptor subtype may be attributable to differing levels of open probability across experiments.

In order to examine the rate of recovery from inhibition for individual receptor subtypes, ACh was applied at time points beyond five minutes after co-application of ACh with a saturating concentration of bis-TMP-10 (2 IpM) and residual inhibition was evaluated (Figure 3-10). For these experiments, 2 gM bis-TMP- 10 alone was applied continuously for 15-20 seconds prior to application of ACh in the continued presence of the same concentration of bis-TMP- 10. The ACh concentrations used were either 10 gM for the a 1 yl, al 04y6, 1[31(034TM2)y8, a104(31TM2)y8 subunit combinations or 100 jiM ACh for ax334 receptors. It is possible to estimate a time constant of recovery from inhibition by fitting these data with a single exponential. Muscle-type receptors show the most rapid recovery from inhibition with a time constant of recovery ('tr) of about 3 minutes, while chimeric aIj31([34TM2)y receptors exhibit the most prolonged inhibition ('tr=81 minutes). As expected, 304 receptors also show prolonged inhibition with a time constant of recovery of about 70 minutes. The rates of recovery of x104(P1TM2)-rS (Tr=6 minutes) and a304(3 I TM2) (tr=3 minutes) receptors are most comparable to that of muscle-type receptors, while the recovery rate of aI 4y8 receptors falls intermediate (Tr=l 6 minutes).


1.2 (1i1 (134TM2)'y8i

S 0.8

2 0.6

g 0.4


0.0001 0.001 0.01 0.1 1 10 100 1000

Concentration (gM)

Figure 3-9. Concentration dependence of residual inhibition of aIPI(P4TM2y 8 receptors. Each data point represents the mean peak response of at least 4 oocytes to 30 gpM ACh alone five minutes after coapplication of 30 p.M ACh with the indicated concentration of bis-TMP-10. Data are expressed relative to the response
to an initial control application of ACh alone.

Figure 3-10. Recovery from inhibition by bis-TMP-10 as a function of time. The data points at t=0 minutes represent a measure of total inhibition observed after coapplication of ACh with 2 pM bis-TMP-10 (see Methods). Data points at t=5, t=10 or t=15 minutes represent mean values of residual inhibition measured after application of ACh alone. Since for ali[ly8 receptors, full recovery was effectively achieved after only 10 minutes of wash, the data at the 15 minute time point for this receptor subtype is omitted for clarity. All drug applications are separated by five minute wash periods. The concentration of ACh in each experiment is either 10 p.M (alily8, a1x I(p4TM2)ya, al04y and a14(p1TM2)y8) or 100 pLM (034 and a30I4(I ITM2)). All data points represent the mean responses of at least 4 oocytes and are fit with an equation describing exponential recovery (see Methods). In some cases (e.g., ailiy8), the recovery process appears to contain two components. However, the normalized responses at later time points reflect the minimal contribution of response rundown over time.




0 2 4 6 8 10 12 14 16
Time (minutes)


Effects of Exchange of the ECL Region on Time-Course of Inhibition by

The bis-TMP- 10 sensitivity of receptors resulting from the exchange of the extracellular loop region was also evaluated (Figure 3-11). Coinjection of the p3(4iECL) subunit with either the other muscle subunits (a l,y,8) or the neuronal alpha3 subunit provides for the expression of functional nAChRs. Substitution of the 131 ECL sequence does not reverse the long-term inhibition normally observed with co-application of bis-TMP-10 and ACh to either the a 104y6 or a334 subunit combinations. Chimeric a 1134(131ECLy8 receptors have a time constant of recovery of about 12 minutes (Figure 3-1 1A). It should be noted however that for a 134(1 ECL)yS receptors, response rundown over time is observed so for the purposes of calculating a recovery rate, responses at each time point are normalized to the amount of rundown observed in response to consecutive applications of ACh alone in control oocytes. Similar rundown is observed for a304(PIECL) receptors. However, for this receptor subtype, normalization for rundown can not correct for the observed biexponential recovery from inhibition. Thus, data measuring recovery for this receptor subtype are plotted together with measurements for rundown and compared to similar measures for a304 receptors (Figure 3-1 1B). Even taking rundown into account, only slight effects of substitution of the extracellular loop region are observed. The reciprocal chimeric beta subunit 13 I(4ECL) was also tested for bis-TMP-10 sensitivity. Coinjection of the 1 3(ECL) subunit with the other muscle subunits (a,y,8) provides for the expression of functional nAChRs. As would also be predicted from the above results, chimeric aol(1MECL)y8 receptors show little residual inhibition five minutes after application of bisTMP-10 (,rr=6.5 minutes). Thus, the capacity to regulate the time course of inhibition by bis-TMP-10 appears to be limited to sequence within the N-terminal half of TM2.

Figure 3-11. Exchange of extracellular loop region between muscle and neuronal beta subunits does not reverse the timecourse of inhibition by bis-TMP-10 after coexpression with either (A) the other muscle subunits (cxi,y,8) or (B) the neuronal alpha3 subunit. For each panel, mean residual inhibition is plotted as a function of time. For both receptor subtypes, significant decreases in the responses to consecutive applications of ACh alone were observed. In the case of al34(PIECL)y receptors, it was possible to normalize for this effect by taking into account the average amount of response rundown observed over time for a control population of oocytes. However, in the case of a3p34(P I ECL) receptors, recovery from inhibition could not be fit by a single exponential even after normalizing for rundown in this manner. Therefore, data for ot3134(P3IECL) receptors is plotted with data for responses to consecutive applications of ACh alone and with similar data for a304 receptors for comparison. For the chimeric receptor subtypes, each data point represents the mean of at least 4 oocytes. For 3P4, each data point is an average of the responses of 2 oocytes. Additional data for a304 receptors is included in Figure 3-8.

00.1 0.1 11
> > 00

0 (XIM(PIECL)-18 0 a3 P4
(XIDI(MECLYYB 0 oc3p4(pIECL)/ACh only
1[3 a3p4/ACh only
0.011 0.011 . . Ell I
-5 0 5 10 15 20 -5 0 5 10 15 20
Time (mins.) Time (n-fins.)


Mechanism of Inhibition of nAChRs by Bis-TMP-10

Although the above results are consistent with a role for sequence in the muscle delta and neuronal beta subunit TM2 regions in determining the time-course of recovery from inhibition by bis-TMP- 10, from these data alone, it is not clear whether these effects are mediated via a direct contribution of the respective TM2 regions to a binding site for bisTMP- 10. It may be the case that sequence elements in TM2 participate in the exposure of a bis-TMP-10 binding site distinct from the site of sequence exchanges in TM2.

Inhibition by Bis-TMP-10 Is Independent of Voltage

Since the residence time of the previously characterized open-channel blockers QX222 and QX-314 has been shown to be dependent on membrane voltage (Leonard et al., 1988; Neher and Steinbach, 1978) and block by a variety of bis-ammonium compounds has also been shown to be voltage-dependent (Ascher et al., 1979; Bertrand et al., 1990; Kurenny et al., 1994; Zhorov et al., 1991), we hypothesized that inhibition by bis-TMP-10 should also show voltage-dependence if bis-TMP- 10 is predominantly charged and inhibition occurs via binding to the chimeric TM2 region directly.

The voltage-dependence of bis-TMP-10 inhibition of a1318 receptors was assessed by measuring the response to ACh alone at a holding potential of -50 mV five minutes after coapplication of ACh with bis-TMP- 10 at a holding potential of either -50 mV or +20 mV. This response was then compared to an initial response to ACh alone at a holding potential of -50 mV. Five minutes after co-application of 2 pM bis-TMP-10 with 30 AIm ACh at a holding potential +20 mV, the inhibition by bis-TMP- 10 is not significantly different from that observed after application of bis-TMP- 10 with ACh at -50 mV (not shown). Because it may be the case that any effects of membrane potential on inhibition can not be detected at high concentrations of inhibitor, this experiment was repeated at a bis-TMP- 10 concentration of 500 nM and holding potentials of -80 and +20 mV (Figure 3-12).

Figure 3-12. Residual inhibition of of a38 receptors by bis-TMP-10 does not show a measurable voltage-dependence. A) For both sets of traces, the responses at the far left and far right are to application of 30 pM ACh alone at a holding potential of -50 mV while the middle response is to application of 30 glM ACh with 500 nM bis-TMP-10 at a holding potential of either -80 mV (top) or +20 mV (bottom). The voltage steps begin about 30 s prior to the start of recording and are maintained for the duration of the middle traces. Each response is separated by five minutes. B) Mean data for peak response five minutes after coapplication of ACh with bis-TMP-10 are normalized to the initial control application of ACh alone. The mean data correspond to the traces at far right in A.


-80 mV
ACh with bis-TMP-10

100 nA +20 mV 30
ACh with bis-TMP-10

T =5 minutes

o 0.75C


o 0.250
-80 mV +20 mV

Holding potential


Similar results were observed. Thus, membrane potential does not appear to affect the long-term inhibition of a 10 18 receptors by bis-TMP- 10.

Interestingly, a1I p11 receptors exhibit more inward rectification of current than do either muscle ora pl~3y receptors (Figure 3-13). Because holding potential does not have any measurable effect on long-term inhibition by bis-TMP- 10, it also does not appear that the process underlying current rectification influences the ability of bis-TMP- 10 to bind to the channel.

Since alp I1q34TM2)y8 receptors resemble neuronal nAChRs in their time-course of recovery from inhibition but maintain the linear current-voltage relationship typical of muscle-type nAChRs (Figure 3-14, panel B), it is possible to examine the effects of voltage on sensitivity to inhibition independent of any potential effects of voltage on channel gating. A voltage step to +20 mV for the duration of the co-application of 30 gM ACh with

2 gM bis-TMP-l0 does not increase the relative magnitude of inhibition of x1i3(4rM2y receptors from that observed with a steady holding potential of -5O mV (Figures 3-14 and 3-16). Thus, for a1131(j4TM2)y6 receptors also, under these conditions at least, the binding of bis-TMP- 10 to its activation-sensitive site appears to be independent of membrane voltage. It should be noted that this experiment examines only the voltage-dependence of the onset of inhibition. Therefore, it is possible that some measure of voltage-dependence would be observed at a lower concentration of bis-TMP- 10. However, the fact that no affect on the onset of inhibition was observed even at a positive holding potential coupled with the fact that no voltage-dependence of inhibition is observed for c4S receptors at a lower concentration of bis-TMP- 10 (500 nM) makes this possibility seem somewhat less likely. It may also be possible that the process underlying recovery exhibits a voltagedependence not detected in these experiments.

The same protocol was used to examine the dependence of inhibition on membrane voltage for the 00~4 receptor subtype (Figure 3-15). However, since neuronal receptors show pronounced inward rectification (Figure 3-15, panel B), a lack of inhibition at


A 1.5 B 8

cUaIly8 c3.4
current current
(pA) (nA/100)

-50 voltage 50 -50 volta e 50
x t(mV (MV

-1.5 -8

C 1.5 D 6.5
(X1P17 t 1P18
current current
(cA)t (nA/100)

-5 0 Volta e 50 -5=0 Volta 50
voltaevolte (mVT (mV

-1.5 -6.5

Figure 3-13. Representative current-voltage relations for (A) muscle-type, (B) neuronal a34, (C) cap3y and (D) aPl8 nAChRs. The holding potential was ramped from -50 mV to +50 mV in the plateau phase of the response to a prolonged application of ACh. Measurements were made in Ringers solution with barium substitued for calcium.

Figure 3-14. Long-term inhibition of alPl(3l4tTM2)y8 receptors by bis-TMP-10 is independent of voltage. A) An initial control application of 30 pM ACh (far left) is followed by either successive applications of ACh alone (upper trace) or by a single coapplication of ACh with 2 gxM bis-TMP-10 followed by subsequent application of ACh alone (lower trace). The timing of the voltage step from -50 mV to +20 mV is represented by the gray line and begins about 30 s prior to and ends about 30 s after the peak of the middle response in each case. Five minute wash periods separate each response. Mean data are shown in Figure 3-16. B) A representative current-voltage relationship during the plateau phase of the response to extended application of 30 M ACh alone is shown. The mean reversal potential for aill1(p4tTM2)y5 receptors is -4.02.1 mV (n=3). Holding currents during ramps in the absence of agonist were point to point subtracted.