The Acetylcholine Receptor:
A Model For Allosteric
Membrane Proteins
J.-P. Changeux
lnstitut Pasteur, Departement des Biotechnologies, 25, rue du Docteur Roux, 75724 Paris Cedex 15, France
Introduction
The cells of the nervous system, or neurons,
possess a property unique in the organism in their
ability to form multiple finely branched extensions
(called axons and dendrites) by which they
assemble into networks of extreme complexity.
Electron microscopy reveals that at the level of their
contacts (or synapses), the membranes of these cells
juxtapose each other while separated by a distance
of tens of nanometers. Chemical substances, or
neurotransmitters, serve as relays between the electrical signals propagated along the axon of one neuron and those produced by the next cell. The
propagation of the action potential down the axon
stimulates the release of neurotransmitter at a high
concentration (around 1 mM) for a very brief period
of time (around 1 ms) [ 11. This ‘chemical impulse’
passes across the synaptic cleft in a fraction of a
millisecond, and triggers an electrical signal on the
postsynaptic cell as a result of the neurotransmitter
binding to a ‘receptive substance’ or receptor, the
former being postulated by the British pharmacologist John Newport Langley as early as 1905 [Z].
Among neurotransmitter receptors, the open ion
channel is selective either for cations (Na+, K + ,
CaL+),producing an excitation, or for anions (Cl-),
producing inhibition. In all cases, the transduction
of the chemical signal into the electrical signal
involves the opening of an ion channel (resulting in
activation) which occurs with a time course of 1 ps
to several milliseconds. When the neurotransmitter
is applied on to the postsynaptic membrane in a
prolonged manner, and at a concentration which
can be lower than that which elicits activation, the
Abbreviations used: TDF, phenyltrimethyl ammonium
benzene diazonium difluoroborate; DDF, similar to TDF.
but with a dimethyl ammonium group in place of the
quaternary ammonium group; GABA, y-aminobutyric
acid.
Thudichum Medal Lecture
Delivered at Imperial College, London,
on 23 December 1993
JEAN-PIERRE CHANGEUX
amplitude of the response slowly decreases. A
desensitization of the response takes place within a
period of several hundred milliseconds to several
minutes.
In 1964, in my Ph.D. thesis, I mentioned that
the processes involved in neurotransmitter recognition and signal transduction at the synapse might
plausibly be found, in the future, to be relevant to
allosteric mechanisms. I therefore decided to
embark on a research programme based on an
explicit theoretical model, and took advantage of a
well defined experimental system. The theoretical
model [3] was that neurotransmitter receptors were
I995
Biochemical Society Transactions
I96
specialized regulatory proteins, or allosteric proteins. Such molecules, typified by haemoglobin and
diverse regulatory enzymes, are characterized by
the presence of several topologically distinct binding sites, and by a particular structural flexibility
that mediates coupling between these sites [4-91
(Figure 1). These proteins were assumed to be
composed of several subunits organized in the form
of a symmetrical 'microcrystal'. They were further
postulated to spontaneously undergo a reversible
confonnational transition between at least two discrete states which preserve their properties of
symmetry, while having different affinities for substrates and/or regulatory ligands, as well as different biological activities. In the particular case of the
channel-linked neurotransmitter receptor, this is the
state of opening of an ion channel.
The experimental system which led to the first
identification of a neurotransmitter receptor, the
acetylcholine receptor, as well as providing support
for the suggested theoretical model, was the electric
organ of the electric fish (Electrophow and
Torpedo). In 1937, David Nachmansohn [lo] had
recognized the advantages offered by this system:
an exceptional richness of homogeneous cholinergic synapses, as well as the possibility of isolating
a single cell, or electroplaque, from the electric
organ and of studying its pharmacological properties, which later proved to be quite similar to those
of the human neuromuscular junction. A brief postdoctoral stay in the laboratory of David Nachmansohn at Columbia University in New York
permitted me to verify, with the help of Tom
Podleski, that receptor agonists (acetylcholine,
phenyltrimethyl ammonium, decamethonium) and
competitive antagonists (curare, flaxedil) behaved in
accordance with the predictions of the two-state
allosteric model [ 111. At the same time, I demonstrated that an affinity label previously used by
Leon Wofsky and Wolf Singer for identifying the
active site of specific antibodies, phenyltrimethyl
ammonium benzene diazonium difluoroborate
(TDF), behaved as an irreversible competitor of
acetylcholine with the electroplaque, and could thus
potentially be used for the same purpose with the
acetylcholine receptor [ 121.
Electrochemical transduction in vitro
The analysis of the basic mechanisms of electrochemical transduction had until this time occurred
through the recording of the electrical activity of
cells. I quickly realized that the identification of a
receptor, and the analysis of the molecular mechanisms involved in signal transduction, would require
Volume 23
Figure I
Schematic representation of models proposed to
account for the regulatory properties of L-threonine
deaminase and other regulatory enzymes
( a ) Indirect or allosteric interactions between topographically
distinct sites is suggested as an alternative to the classic model
of competitive interaction between ligands binding to overlapping sites [4]. ( b ) Plausible model of allosteric transition
interpreted in terms of quaternary structure which accounts
for both homotropic interactions (between identical sites) and
heterotropic interactions (between different sites). The
transition is thought to take place between two discrete and
symmetrical states possessing distinct binding properties and
biological activities [7-91.
(4
1
Overlapping
Non-overlapping
.
0,F'"'
S
4t
S
Monomer
perforce the establishment of an in vitro 'chemical'
system using which structure/function relationships could finally be assessed in a direct
manner. This step was advanced by the demonstration by Michiki Kasai and myself [13] that membrane fragments of the electric organ of the eel,
when purified as a homogenate, spontaneously
formed closed vesicles, or microsacs, which
responded to acetylcholine with an increase in
permeability to Na+ and K + in vitro (Figure 2). In
agreement with the suggested allosteric mechanism,
Thudichum Medal Lecture
Figure 2
Demonstration of the ionic response in vitro of membrane fragments
purified from the electric organ of Electrophorus electricus
(0) Description of
the filtration method used for measuring the efflux of 22NaCin a
suspension of membrane fragments forming closed vesicles (microsacs). ( b ) When
microsacs rich in acetylcholinesterase and apparently derived from the innervated face of
the electroplaque are used (i) the nicotinic agonist carbamylcholine increases the efflux of
22Na+,and this effect is blocked by d-tubocurarine. With microsacs not enriched for
acetylcholinesterase and apparently derived from the non-innervatedface of the electroplaque (ii), no response is measured [13,14]. Symbols: 0 , control; 0, I O p M dtubocurarine; X , IOOpM carbamylcholine; A , IOOpM carbamylcholine+ IOpM
d-tubocurarine.
I97
w r i h 31 withcold buffrr
o f the dry mdhpora
Sucrose gradient
centrifugation
incubation
tcholinergic
effector
"1
filtration
on millipore
[
BrlO'M "NaCl
(75 m C/ng)
10% NcU
5110'lH Sucrose
- 2 - 5 y dproimna/ml
Cnniglrt I t C C
L l r l O ' M KCl
2110"M GC12
ia 10% Nc phosphate
pn zo
0
60
22'C
I
20
40
0
20
40
60
Time (min)
this regulation was observed in a test-tube assay
[13]. This activation process took place in the
absence of any energy contribution other than the
change in acetylcholine concentration. Starting from
these microsacs, the identification of the receptor
molecule in a functional state became foreseeable as
a realistic goal.
Identification of the receptor protein
A second, no less decisive, step in this direction was
the use of a label that was highly selective for the
active site of the receptor. This was a-
bungarotoxin, a polypeptide of 74 amino acids,
initially characterized and purified by the Taiwanese pharmacologist Chen Yuan Lee [16] from the
venom of the snake Bungarus multicinctus. This
toxin has a paralysing effect, similar to that of
curare, but acts at extremely low concentrations and
in a slowly reversible manner. a-Bungarotoxin
inhibits the electrophysiologicalresponse of isolated
electric eel electroplaque, blocks the increase in ion
permeability caused by acetylcholine in excitable
microsacs, and displaces already bound, radioactively labelled decamethonium in a microsac
1995
Biochemical Society Transactions
I98
preparation solubilized by the detergent deoxycholate [IS] (Figure 3). The protein which binds abungarotoxin
and
the
nicotinic
agonist
decamethonium in a mutually exclusive manner has
a very large size, and could be separated from
acetylcholinesterase.
The receptor protein was hereafter purified at
the Institut Pasteur and in other laboratories, and
was then observed for the first time in the electron
microscope in 1973 by Jean Cartaud, using my
preparations [20-221. Arthur Karlin and Michael
Raftery [23] independently showed that the receptor had a molecular mass of the order of 300000
Da, was composed of four types of subunit assembled into a heteropentamer (a,/?yS), and that
the two a subunits each contributed a binding site
for acetylcholine or a-bungarotoxin.
A biochemist cannot be satisfied on discovering a new type of molecule without knowing if the
purified protein retained all its functions that were
related to electrochemical transduction. In particular, did it function as an ion channel? The strategy
adopted by Gkrald Hazelbauer and myself [24] was
to re-incorporate a purified fraction of the receptor
protein into artificial lipid vesicles and use them,
like the microsacs of the electric organ, to study ion
fluxes as a method of assessing their transduction
function. This experiment succeeded first with a
purified fraction, and then with the purified protein
[25,26]. Finally, the electrical activity of individual
ion channels was recorded by Mauricio Montal and
his collaborators in 1980 [27] with the receptor protein integrated into lipid bilayers. The oligomer
(a2!lyS) thus contained the binding site for acetylcholine and the ion channel, as well as all the structural elements engaged in activation and
desensitization. The critical physiological step of
transducing a chemical signal into an electrical
signal at the synapse was thus 'reduced' to the
properties of a single molecular species.
Figure 3
Identification of the acetylcholine receptor with the aid of a-bungarotoxin
(a-Bgt), a toxin from the venom of Bungarus rnulticintus
(a) Quasi-irreversible inhibition by a-Bgt of the electrical response to carbamylcholine
(Carb) of the isolated electroplaques from Uectrophorus electricus. R, rinsing with buffer.
( b ) The ionic response of excitable microsacs is blocked in vitro by a-Bgt. 0 , Control; X I
I O p M carbamylcholine; 0,IOpM d-tubocurarine.(c) a-Bgt displaces the binding of the
radioactively labelled nicotinic agonist decamethonium measured by equilibrium dialysis
with a membrane extract of Electrophorus electricus solubilized by the detergent sodium
deoxycholate (after [IS]).
c _ _
201 . . .
0 10 20 30
10 min
.
Time (min)
h
.-
(c)
Y
301
0'
0"
Volume 23
.
'
10
50
[ a-Bungarotoxin] (pgjml)
.
.
0 10 20 30
I00
Thudichum Medal Lecture
The contribution of biotechnology
At the end of the 197Os, progress in protein
microsequencing and in genetic engineering provided access to the partial and then complete primary amino acid sequence of the receptor. The
sequence of the 20 amino acids of the N-terminal
domain of the a subunit of Torpedo marmorata was
established by Anne Devillers-Thiiry, Donny
Strosberg and myself in 1979 [28]. Then, in the following year, the partial N-terminal sequence of the
four chains of the receptor from Torpedo californka
was determined by Michael Raftery and his collaborators [29], revealing a 3550% sequence identity
among the subunits. This sequence homology was
interpreted on the basis of an evolutionary relationship among the genes encoding each of the four
chains of the receptor. These sequence data also
suggested a ‘pseudo-symmetrical’ organization of
the receptor oligomer ( a 2 p y d ) around a five-fold
axis of rotation perpendicular to the plane of the
membrane.
Rased on these biochemical data, the cloning
of the cDNAs encoding the chains of the receptor
was reported by four laboratories, including mine
[30]. Then, the complete sequence of the four
chains of the receptor of Torpedo californka was
established by Shosaku Numa and his collaborators
[31], and that of the a chain of Torpedo marmorata
by my group at the Institut Pasteur [32]. An analysis of neighbouring hydrophilic and hydrophobic
amino acids along the length of each of the four
chains revealed a remarkable compartmentalization
into functional domains (see [32]). On this basis, a
transmembrane organization of each polypeptide
subunit was suggested: the large, hydrophilic
N-terminal domain would face the synaptic cleft,
the small hydrophilic C-terminal domain would be
in contact with the cytoplasm, the four hydrophobic
segments of about 20 amino acids each would cross
the membrane, and the C-terminal tail would sneak
into the synaptic cleft.
The same scheme applies to the sequences of
the four chains of the human muscle receptor and
to those of the neuronal nicotinic receptors, of
which there are known to be seven types of a
subunits and three types of non-a subunits
(reviewed in [33]). Among these, the a7 subunit
possesses the remarkable property of being able to
associate with itself and form functional homooligomers when expressed in Xenopus oocytes
[34,35]. This particularly simple system will permit
progress in the analysis of the functional properties
of the receptor. Rut the list is not yet finished. It has
recently extended to include receptors for other
neurotransmitters: the glycine and y-aminobutyric
acid (GABA) receptors contain an ion channel
selective for chloride ions, and play a role as
inhibitory receptors. All these neurotransmitter
receptors comprise a superfamily of ligand-gated
ion channels to which belong, as distant cousins, the
subfamily of glutamate receptors [33]. The work
initially developed for the nicotinic receptor of the
electric eel has thus opened the way for a targeted
molecular pharmacology of neurotransmitter-gated
ion channels involved in fast cerebral
communication.
Structure of the active site
The identification of the amino acids comprising
the active site has benefited from the use of affinity
labels which, like T D F [ 121, serve as ‘self-sticking
labels’ by binding selectively to the active site in a
covalent manner. DDF, a compound very similar to
TDF, but with a dimethyl ammonium group in
place of the quaternary ammonium group [361, acts
as a reversible competitive antagonist in the dark.
Furthermore, it can be activated by energy transfer,
i.e. by light emitted by aromatic amino acids present
in the binding site when they are irradiated by ultraviolet light. This results in a considerable increase
in the specificity of labelling. Work towards identifying the amino acids labelled by DDF, which
involved the participation of Michael Dennis,
Jkr6me Giraudat and Jean-IJuc Galzi from my laboratory, provided evidence for the presence of three
loops (loop A, Tyr-93; loop B, Trp-149; loop C,
Tyr- 190, Cys- 192, Cys- 193, Tyr- 198) comprising,
as expected, a large hydrophilic N-terminal domain
[37,38]. The amino acids labelled in loop C indicated cysteines 192 and 193, initially identified by
Arthur Karlin and his collaborators [39] in the
reduced receptor by using an affinity label reacting
selectively with cysteines. With the exception of
these two cysteines, all the amino acids labelled
were aromatic amino acids (Figure 4). Moreover,
several groups (reviewed in [41]), using different
affinity labels, have confirmed the contribution to
the active site of Tyr-93 (acetylcholine mustard),
Tyr- 190 (lophotoxin, nicotine, d-tubocurarine) and
Tyr- 198 (nicotine). Furthermore, the amino acids
labelled by DDF in the Torpedo receptor are conserved in the nicotinic receptors of the central nervous system. Finally, work performed in
collaboration with Daniel Bertrand of the Centre
Mkdical Universitaire of Geneva, has shown that a
substitution by phenylalanine of the amino acids
homologous to Tyr-93, Trp-149 and Tyr-190 of the
I995
I99
Biochemical Society Transactions
Figure 4
Model of several loops of the binding site for nicotinic ligands
The sphere represents a molecule of DDF in all possible orientations The amino acids
labelled by DDF, lophotoxin, MBTA, acetylcholine mustard and nicotine are indicated
The symbols in bold refer to mutated amino acids The residues labelled by dtubocurarine (dTC) in the y and 6 subunits are Trp-55 and Trp-57 respectively (see the
text) The amino acid sequences of the hydrophilic N-terminal domain of several
peripheral and central nicotinic receptors show that residues labelled by DDF in Torpedo
are highly conserved (modified from [40])
200
2
..
or
6
W
A lophotoxm
MBTA
*
m
Ach mustard
+ nicotine
mutated
LOOP c
Loop B
Loop A
f
a1 T o r p e d o . .
WLPDLVLY".
85
95
GIWTXDG.
147
*.
. VYYTCCPD-TPXLD
153
188
200
...
IWPDLVLY". .
------------
GTWTXDG. . VTYSCCPD-TPXLD
GTWTXDG. . VNYSCCLD-TPXLD
0 2 r a t . . . ...
INIPDIVLY". .
IWPDIVLY". .
ImPDIVLYNN. .
IWPDIVLY". .
GSWTXDK. . KKYDCC-AEI-XPD
GSWSXDK. . IKYNCC-EEI-XQD
GSWTXDK. . RKYECC-AEI-XPD
GSWTXDK. . RKYECC-AEI-XPD
WQPDIVLM". .
LNLPDWLENN. .
WIPDIVLQNN..
IWLPDWLYNN..
KSYTXDT. .
RSQTXNA. .
TALNXDA. .
RSWTXDR. .
a1 human.
a1 cobra . . . .
a3 r a t . . . . . .
a4 r a t . . . . . .
a
7 chick.. . .
p
Torpedo..
Torpedo. .
Torpedo..
p 2 Rat.. ....
a 7 subunit expressed in Xenopus oocytes [42]
results in a decrease in the apparent affinity of this
receptor for acetylcholine and nicotine. These
amino acids thus contribute to the recognition of
the neurotransmitter. They form some kind of
aromatic basket' in which negative charges are
delocalized to accommodate the positively charged
quaternary ammonium group of acetylcholine.
Recently, Nigel Unwin [43] has presented high
resolution electron microscopic images which
reveal that, in the synaptic portion of the a subunit,
three segments are folded into an a helix, which
may correspond to the three loops of the active site.
Volume 23
.
- - - - - RSDDPSXED
NWQLTK-DDTDFQE
-YPDKFPNGTNXQD
RRNEN-PDDSTYVD
Identification and structure of the ion
channel
Affinity labels opened the way towards identifying
the structure of the ion channel. For several
decades, pharmacologists and physiologists had
noted that a rather heterogeneous group of
pharmacological agents, which include local anaesthetics and the frog toxin histrionicotoxin, blocked
the permeability response of the acetylcholine
receptor in a manner different from that of curare.
These agents blocked the permeability response
without significantly inhibiting the binding of acetylcholine. The idea was put forward that these non-
Thudichum Medal Lecture
competitive inhibitors enter into the ion channel
and block the passage of ions in a steric fashion
[44]. One of these inhibitors, chlorpromazine, was
used by Thierry Heidmann, Robert Oswald and
myself for labelling the ion channel. It was used for
at least three reasons: (1) Chlorpromazine, among
its other effects, brings about a reduction in the
mean open time of the channel in response to
acetylcholine in cultured mouse myotubes [45,46],
and inhibits ion fluxes in excitable microsacs in
vztro. (2) Chlorpromazine binds to a single highaffinity site in each receptor oligomer [ a,Byd]; furthermore, when chlorpromazine is bound to this
unique site, UV irradiation results in a labelling of
all the subunits, suggesting an arrangement of this
site in the axis of symmetry of the receptor oligomer [47]. ( 3 ) A rapid mixing of receptor-rich
excitable microsacs with acetylcholine, followed by
brief IJV irradiation, results in a 100-1000-fold
increase in the rate of covalent attachment of chlorpromazine to four of the subunits [48,49]; these
rates of incorporation diminish during exposure to
a fixed concentration of acetylcholine with kinetics
consistent with those of rapid desensitization of ion
flux measured with the same membrane fragments
PI.
In other words, chlorpromazine binds freely
and reacts covalently with its high-affinity site under
the precise conditions where acetylcholine causes
the opening of the ion channel.
Following a particularly difficult fractionation
of the hydrophobic peptide labelled with
chlorpromazine, Jkr6me Giraudat, Michael Dennis,
Thierry Heidmann, Jui-Yoa Chang and I arrived at
the conclusion that chlorproma%ine bound to the
serine at position 262 of the 6 subunit within the
putative segment M2 [5 13. Equivalent serines were
found to be equally labelled in other subunits, as
well as the corresponding leucines at positions 257
and 260 in the B and y subunits respectively, and
threonine-253 in the y subunit [52-541 (Figure 5).
Following the initial work begun in my laboratory,
Ferdinand Hucho and his group identified by the
same method the same serine-262 in the 6 subunit,
using another non-competitive blocker, triphenylmethylphosphonium [55,56]. These results
were in accordance with models already mentioned
in which (1) the binding site of chlorpromazine is
located in the axis of pseudo-symmetry of the molecule [47], and (2) this site is bordered by a segment
common to each subunit, organized in the form of
a-helices (see [XI).
Figure 5
Model of the high-affinity site for non-competitive blockers within the ion channel and a comparison of the
amino acid sequences of the M2 segment of several channel-linked receptors
The numbers on the left refer to the sequence of amino acids of the y-subunit of Torpedo. Based on [41] and [54].
Cleft
aOuter ring 271
AChR
21.
a
Torp.Cal.
0
Torp.C.31.
T o r p Cal.
TOrp.ca1.
rat
Ha3
Valine ring
a4
a2
264
Leucine ring
260
Serine ring
257
rat
Chick
(13
chick
a4
a7
Chick
Chick
6:
y
L
:
mouse
hRD
Gly R
ai
(12, 02.
P
Threonine ring
ar
ii
Intermediate ring 250
246
, .OMPAR
VGLGITIVLTMTTQSSGW
A
. .DAAPAR VALGITTVLTMTTQSSGSR A
. .DASAAR VPLGIFSVLSLAS~CTTLI~
0.
GABAA R
253
al, 0 2 . a5
a3
Inner ring
I.,
. . DSG-EK
. .DAG-EK
, QAGGQK
. .ESG-EK
. .DCG-EY.
. .ECG-EK
. . DCG-EK
. .DCG-EK
.ECG-EK
. .DSG-EK
, .DCG-EK
. ..DAG-EK
KAGGQK
. .EAG-EK
I
/E---D
0
..--D
\i
P
. .ESVPAR TVFGVTTVLTMrr LS 1 SAi1 14
, . ESVPAR
TVFGVTTVLTMTTLSISAR
N..
, .ESVPAR
TVFGVTTVLTMTTK
I sm I4 .
. . ESVPAR TVFGITTVLTMlrLSI $All I/
. .DASAAR VALGITTVLTw1”r ISTI1 LR E
. . DASAAR VALG1TTVLTMT.T I NTII 1.11 I: .
. .DAVPAR TSLGITTVLTMTTLST
I A!: K
. .MVPI\R VSLGITTVLTl~rI.I.MVs(\ll
D
1
’
Cytoplasm
I995
20 I
Biochemical Society Transactions
202
This work performed in my laboratory with
chlorpromazine [51] showed for the first time the
contribution of the M2 segment to the walls of the
ion channel. Mutagenesis experiments carried out
by the groups of Shosaku Numa and Bert Sakmann,
and by Henry A. Lester and Norman Davidson,
confirmed the notion that the M2 segment [51-561
contributes to the inner surface of the ion channel
[57-59].
The analysis of the correspondence between
the selectivity properties of the ion channel and the
three-dimensional structure of the receptor molecule is but the beginning. A collaboration with the
group of Daniel Bertrand has pursued this theme.
The ion channels of the neuronal nicotinic receptors, particularly a7, are relatively more permeable
to Ca2+ compared with the muscle receptor. Mutations at two levels of M2, at the ‘intermediate’ring of
negatively charged residues on the cytoplasmic side
[57,58] and at the double ring of leucines on the
synaptic side, selectively abolish Ca2 permeability
without interfering with Na+ or K + permeability
[60]. The receptors for GARA and glycine possess
an ion channel that is selective for anions. Exchange
of the two rings of critical amino acids and insertion
of a novel ring at the N-terminus of M2 suffices to
convert the ion channel of the a 7 receptor from
cationic to anionic [61]. As in the case of immunoglobulins, the diverse members of the superfamily
of ion-channel-linked receptors possess a similar
molecular architecture. However, they differ in the
nature of amino acids situated in critical positions of
the protein, which are sufficient to determine the
specificities for both neurotransmitter recognition
and ion transport.
Figure 6
Minimal four-state model for the allosteric transitions
of the acetylcholine receptor
ACh, acetylcholine; CB, competitive inhibitor; NCB, noncompetitive blocker; P, phosphate bound in a covalent manner;
R, resting state; A, active state; I, intermediate or rapidly
desensitized state: D, equilibrium or slowly desensitized state
(modified from [63]).
R
A
CB 0
+
Slow allosteric transitions and the
regulation of synaptic efficacy
A brief ‘chemical impulse’ of neurotransmitter
assures the rapid transmission of a signal from one
neuron to another across the synaptic cleft. Neurotransmitter application in a continuous manner, and
at a concentration much lower than that which activates the ion channel, results in a slow, yet reversible, decrease in the amplitude of the permeability
response, known as desensitization. In 1957, Sir
Bernard Katz and Stephen Thesleff [62] proposed,
on the basis of electrophysiological data, a strict
theoretical scheme for activation and desensitization, according to which stabilization at equilibrium
of the receptors produces a ‘refractory’ state with a
high affinity for acetylcholine (Figure 6). The opening of the ion channel would occur through a tran-
Volume 23
sient state of low affinity. Biochemical experiments
have demonstrated the general validity of this
scheme, but with several significant differences.
Soon after identification of the receptor, direct
equilibrium-binding measurements with receptorrich microsacs from Torpedo revealed an affinity for
acetylcholine of the order of 10XM, about
1000- 10 000-fold higher than the concentration
which results in the opening of the ion channel with
the same membranes [64-671. The high affinity
observed at equilibrium could not have corresponded to the active site. Rapid mixing experiments, first carried out ‘by hand by Michel Weber
[68], and then later with the aid of stopped-flow
Thudichum Medal Lecture
apparatus by Hans Grunhagen [69,70] and then
Theirry Heidmann [7 1-73] in my laboratory,
revealed without ambiguity that, at rest, around 80%
of the receptor molecules are in a state of low affinity. Placed in the presence of a fixed concentration
of acetylcholine, these receptors slowly interconvert
towards a high-affinity state at equilibrium. The
system of excitable microsacs from Torpedo thus
gives access to the measurement, in parallel, and in
the same experimental system, of both ion permeability and the binding of agonists which activates
these permeabilities [SO]. The data demonstrated
for the first time, in a direct manner, that at least two
states of high affinity (Kdsaround
to lo-“)
result in a closed ion channel and thus correspond
to desensitized states of the receptor [SO].
The receptor, when purified and reconstituted
into an activatable form, maintains its capacity to
desensitize (see [24-261). This is thus an intrinsic
feature of the receptor which is also found in the
neuronal receptors. A single molecule thus
possesses two classes of regulatory properties: elementary properties for transduction of a chemical
signal into an electrical signal, and higher-order
properties which regulate the efficiency of this
transduction process. The systematic analysis, by
site-directed mutagenesis, of the role of amino acids
labelled by chlorpromazine, singled out a few of
them, in particular leucines from segment M2, that
are specifically involved in this regulation. Mutations into threonine (or serine) of the equatorial
leucine ring present in the M2 segment of the a 7
subunit (1,237T) results in several simultaneous
modifications of the electrophysiological response:
the appearance of a new open channel type, a loss
of desensitization, and an increase in the apparent
affinity of the permeability response to nicotinic
agonists. In accordance with the four-state model of
allosteric transitions, these results are interpreted on
the basis of a permeabilization of the desensitized
state. Mutation of the leucine ring unblocks the ion
channel from its desensitized state. In agreement
with this conclusion, competitive antagonists such
as dihydro-/3-erythroidine or curare, which we
know favour the desensitized state 169.701. stabilize
in the mutant an ‘open desensitized’ state and thus
behave as agonists [74,75].
Neurotransmitter receptors, such as those for
GABA, glycine and glutamate, share several structural properties with the nicotinic receptor [33].
They undergo
but in a manner
which varies with the subunit composition. It seems
plausible, though it has not been demonstrated, that
the ‘desensitization mutants’ of the type mentioned
L
J,
for the nicotinic receptor offer one plausible model
of this diversity [74,75].
Finally, physiological signals other than
neurotransmitters regulate the rate and amplitude of
desensitization of the muscle nicotinic receptor:
membrane potential, Ca2 and phosphorylation
makes possible a regulation of desensitization by
neuronal activity (reviewed in [76]).
All the elements are there for proposing that
desensitization, and in a general sense all structural
transitions of the receptor other than the activation
reaction, play an essential role in the regulation of
synaptic transmission at the postsynaptic level. The
transmembrane organization of the receptor molecule confers on it the capacity to recognize signals
coming from the exterior as from the interior of the
cell. It allows for the integration of multiple signals
and the reading of their coincidence in time. The
allosteric properties of the channel-linked receptor
offer plausible biochemical mechanisms for the
learning law postulated by the Canadian psychologist Donald Hebb [77,78].The relationship between
these elementary properties and learning at the level
of behaviour, however, remains to be established.
In conclusion, the data and interpretation of
work begun with the nicotinic receptor from the
electric eel electric organ can be expanded to
diverse members of the superfamily of ligand-gated
ion channels present in the central nervous system.
A targeted pharmacology for each particular type of
receptor which takes into account conformational
flexibility becomes henceforth plausible, with the
hope that treatments for a select number of neurological and psychiatric disorders will be possible.
+
I thank all my collaborators who, since 1967, have participated in this work, as well as the Association Francaise
Contre Les Myopathies, the Centre National de la
Recherche Scientifique, the College de France, the Institute National de la Santk et de la Recherche Mkdicale, the
Direction des Recherches et Etudes Techniques, the
CEE and Human Frontiers for support.
1 Katz, B. (1966) Nerve, Muscle and Synapse, McGraw
Hill, New York
2 IAangb-J. N. (1905)J- Physiol. (London) 33, 374-413
3 Changeux, J. P. (1969) in Symmetry and Functions in
Biological Systems at the Macromolecular Level.
(Engstrom, A. and Stranberg, B., eds.), Nobel Symposium No. 11, pp. 235-256, Wiley Inter Science Division, New York
4 Changeux, J. P. (1961) Cold Spring Harbor Symp.
Quant. Biol. 26,313-318
5 Monod, J., Changeux, J. P. and Jacob, F. (1963) J. Mol.
Biol. 6,306-328
I995
203
Biochemical Society Transactions
204
6 Monod. J., Wyman, J. and Changeux. J. P. (1965) J.
Mol. Biol. 12.88-1 18
7 Changeux, J. P. (196.5) Bull. SOC.Chim. Biol. 47,
115-139
8 Changeux. J. P. (1965) Bull. SOC.Chim. Biol. 47,
267-280
9 Changeux. J. 1’. (1965) Bull. SOC.Chim. Biol. 47,
28 1-300
10 Nachmansohn, D. (1959) Chemical and Molecular
Basis of Nerve Activity, Academic Press, New York
11 Changeux. J. P. and Podleski, T. (1968) Proc. Natl.
Acad. Sci. U.S.A. 59,944-950
12 Changeux, J. P., Podleski, T. and Wofsy, L. (1967)
Proc. Natl. Acad. Sci. U.S.A. 58,2063-2070
13 Kasai, M. and Changeux. J. P. (1971) J. Membr. Biol.
6, 1-80
14 Kasai, M. and Changeux, J. P. (1970) C. R. Acad. Sci.
Paris 270D. 1400- 1403
15 Changeux, J. P., Kasai, M. and Lee, C. Y. ( 1970) Proc.
Natl. Acad. Sci. U.S.A. 67, 1241-1247
16 Lee, C. Y. and Chang, C. C. (1966) Mem. Inst.
Butantan Sao Paulo 33,555-572
17 Reference deleted
18 Reference deleted
19 Reference deleted
20 Cartaud, J , Benedetti, L., Cohen, J. B., Meunier, J. C.
and Changeux,J. P. ( 1973) FEBS Lett. 33,109- 113
21 Cartaud, J., Benedetti, L.. Sobel, A. and Changeux.
J. P. (1978)J. Cell Sci. 29,313-337
22 Bon, F., Lebrun, E., Gomel, J., Van Rappenbusch, R.,
Cartaud, J., Popot, J. L. and Changeux. J. P. (1984) J.
Mol. Biol. 176,205-237
23 Karlin, A. ( 1991) The Harvey Lectures 85,7 1- 107
24 Hazelbauer, G. and Changeux, J. P. (1974) Proc. Natl.
Acad. Sci. U.S.A. 71, 1479-1483
25 Changeux,J. P., Heidmann, T., Popot, J. L. and Sobel.
A. (1979) FEBS Lett. 105,181-187
26 Popot, J. L., Cartaud, J. and Changeux, J. P. (1981)
Eur. J. Biochem. 118,203-214
27 Nelson, N., Anholt, R., Lindstrom, J. and Montal, M.
(1980) Proc. Natl. Acad. Sci. U.SA. 77,3057-3061
28 Devillers-ThiCry, A., Changeux, J. P., Paroutaud, P.
and Strosberg, A. D. (1979) FEBS Lett. 104,99-105
29 Raftery, M. A., Hunkapiller, M., Strader, C. D. and
Hood, L. E. (1980) Science 208,1454-1457
30 Giraudat, J., Dennis, M., Heidmann, T., Chang, J. Y.
and Changeux, J. P. (1986) Proc. Natl. Acad. Sci.
USA. 83,2719-2723
31 Noda, M., Takahashi, H., Tanabe, T., Toyosato, M.,
Furutani, Y., Hirose, T., Asai. M., Inayama, S., Miyata.
T. and Numa, S. (1982) Nature (London) 299,
793-797
32 Devillers-ThiOry, A., Giraudat. J., Bentaboulet, M. and
Changeux, J. P. (1983) Proc. Natl. Acad. Sci. U.S.A.
80,2067-2071
33 Cockcroft, V. B., Osguthorpe, D. J., Barnard, E. A.,
Friday, A. E. and Lunt, G. G. (1992) Mol. Neurobiol.
4, 129-169
Volume 23
34 Couturier, S., Erkman, I,.. Valera, S., Kungger, D..
Bertrand, S., Houlter, J., Ballivet, M. and Bertrand. D.
(1990)J. Biol. Chem. 265,17560- 17567
3.5 Schoepfer, R., Conroy, W. G., Whiting, P.,Gore, M.
and Lindstrom, J. (1990) Neuron 5.35-48
36 Langenbuch-Cachat,J., Bon, C., Goeldner, M., Hirth,
C. and Changeux, J. P. (1988) Biochemistry 27,
2337-2345
37 Dennis, M., Giraudat. J.. Kotzyba-Hibert, F.,
Goeldner. M., Hirth, C., Chang. J. Y.. Lazure, C.,
ChrCtien, M. and Changeux, J. P. (1988) Biochemistry
27,2346-2357
38 Galzi, J. I,., Revah, F.. Black, D., Goeldner, M., Hirth,
C. and Changeux. J. P. (1990) J. Biol. Chem. 265,
10430- 10437
39 Kao, P. N., Dwork, A. J.* Kaldany, R. R. J., Silver,
M. L., Widemann, J., Stein, J. and Karlin, A. (1 984) J.
Biol. Chem. 259,11662- 11665
40 Galzi, J. L., Revah, F., Bessis, A. and Changeux, J. P.
(1991) Annu. Rev. Pharmacol. Toxicol. 31,37-72
41 Changeux, J. P., Galzi, J. I,., Devillers-Thiiry, A. and
Bertrand, D. (1992) Q. Rev. Biophys. 25,395-432
42 Galzi, J. L., Bertrand, D., Devillers-Thiery, A., Revah,
F., Bertrand, S. and Changeux, J. P. (1991) FEBS Lett.
294, 198-202
43 Unwin, N. (1993)J. Mol. Biol. 229, 1101-1 124
44 Neher, E. and Steinbach, J. H. (1978) J. Physiol.
(London)277,153-176
45 Changeux, J. P., Pinset, C. and Ribera, A. B. (1986) J.
Physiol. (London) 378,497-5 13
46 Benoit, P. and Changeux, J. P. (1993) Neurosci. Lett.
160,8 1-84
47 Heidmann, T., Oswald, R. E. and Changeux, J. P.
(1983) Biochemistry 22,3112-3127
48 Heidmann, T. and Changeux, J. P. (1984) Proc. Natl.
Acad. Sci. U.S.A. 81,1897-1901
49 Heidmann, T. and Changeux, J. P. (1986) Biochemistry 25.6 109-6 113
50 Heidmann, T.. Bernhardt, J., Neumann, E. and
Changeux,J. P. (1983) Biochemistry 22,5452-5459
51 Giraudat, J.. Dennis, M., Heidmann, T., Chang, J. Y.
and Changeux, J. P. (1986) Proc. Natl. Acad. Sci.
U.S.A. 83,27 19-2723
52 Giraudat, J., Dennis, M., Heidmann, T., Haumont,
P. Y..Lederer, F. and Changeux, J. P. (1987) Biochemistry 26,24 10-24 18
53 Giraudat, J., Galzi, J. L., Revah. F.. Changeux, J. P.,
Haumont, P. Y. and Lederer, F. (1989) FEBS Lett.
253,190-198
54 Revah, F., Galzi, J. L., Giraudat. J.. Haumont, P. Y.,
Lederer, F. and Changeux, J. P. (1990) Proc. Natl.
Acad. Sci. U.S.A. 87,4675-4679
55 Oberthur, W., Muhn, P., Baumann, H., Lottspeich, F.,
Wittman-Liebold, B. and Hucho, F. (1986) EMBO J.
5,1815-1819
56 Hucho, F., Oberthur, W. and Lottspeich, F. (1986)
FEBS Lett. 205,137- 142
57 Imoto, K., Methfessel. C., Sakmann, B., Mishina, M.,
Thudichum Medal Lecture
58
59
60
61
62
63
64
65
66
67
68
Mori, Y.. Konno. T., Fukuda, K., Kurasaki, M., Bujo,
H., Fujita, Y. and Numa, S. (1 986) Nature (London)
324,670-674
Imoto, K., Busch, C.. Sakmann, B., Mishina, M.,
Konno, T., Nakai, J., Bujo, H., Mori, Y., Fukuda, K.
and Numan, S. (1988) Nature (London) 335,645-648
Leonard, R. J., Labarca, C. G., Charnet, P., Davidson,
N. and Lester, H. A. ( 1988) Science 242,1578- 1 58 1
Bertrand, D., Galzi, J. I,., Devillers-Thiiry, A.,
Bertrand, S. and Changeux, J. P. (1993) Proc. Natl.
Acad. Sci. U.S.A. 90,6971-6975
Galzi, J. L., Devillers-Thiery, A., Hussy, N., Bertrand,
S.. Changeux, J. P. and Bertrand, D. (1992) Nature
(London) 359,500-505
Katz, B. and Thesleff, S. (1957) J. Physiol. (London)
138,63-80
Changeux, J. P., Devillers-Thidry, A. and Chemouilli,
P. (1984) Science 225,1335-1345
Weber. M. and Changeux, J. P. (1974) Mol. PharmaC O ~ .10, 1-14
Weber, M. and Changeux, J. P. ( 1 974) Mol. Pharmacol. 10, 15-34
Weber, M. and Changeux, J. P. (1974) Mol. PharmaC O ~ .10, 35-40
Cohen, J. B., Weber, M. and Changeux, J. P. (1974)
Mol. Pharmacol. 10,904-932
Weber. M.. David-Pfeuty, M. T. and Changeux, J. P.
(1975) Proc. Natl. Acad. Sci. U.S.A. 712,3443-3447
69 Griinhagen, H. H. and Changeux, J. P. (1976) J. Mol.
Biol. 106,497-5 16
70 Griinhagen, H. H., Iwatsubo, M. and Changeux, J. P.
(1977) Eur. J. Biochem. 80,225-242
71 Heidmann. T. and Changeux, J. P. (1979) Eur. J.
Biochem. 94,255-279
72 Heidmann, T. and Changeux, J. P. (1979) Eur. J.
Biochem. 94,28 1-296
73 Heidmann, T. and Changeux, J. P. (1980) Biochem.
Biophys. Res. Commun. 97,889-896
74 Bertrand, D., Devillers-Thikry, A., Revah, F., Galzi,
J. L., Hussy, N., Mulle, C., Bertrand, S., Ballivet, M.
and Changeux, J. P. (1992) Proc. Natl. Acad. Sci.
USA. 89,126 1 - 1265
75 Devillers-Thiiry, A., Galzi, J. L., Bertrand, S.,
Changeux, J. P. and Bertrand, D. (1 992) NeuroReport
3,1001-1004
76 Lina, C. and Changeux, J. P. (1993) Trends Neurosci.
16,181-186
77 Heidmann, T. and Changeux, J. P. (1 982) C. R. Acad.
Sci. Paris Sirie 111,295,665-670
78 Changeux, J. P. and Heidmann, T. (1981) in Synaptic
Function (Edelman, G., Gall, W. E. and Cowan,
W. M., eds.). pp. 549-604, John Wiley, New York
Received 28 November 1994
I995
205