Tài liệu Báo cáo khoa học: Activation of the Torpedo nicotinic acetylcholine receptor The contribution of residues aArg55 and cGlu93 ppt

Activation of the Torpedo nicotinic acetylcholine receptor
The contribution of residues aArg55 and cGlu93
Ankur Kapur, Martin Davies, William F. Dryden and Susan M.J. Dunn
Department of Pharmacology, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Canada
The muscle-type nicotinic acetylcholine receptor
(nAChR) is the prototype of the Cys-loop ligand-gated
ion channel (LGIC) super-family that includes the
neuronal nicotinic, c-aminobutyric acid (GABA) type
A, 5-hydrotryptamine type 3 (5-HT
3
) and glycine
receptors. This is largely a consequence of the abun-
dance of this receptor in Torpedo electric organ, which
facilitated its early purification and characterization.
The Torpedo nAChR is a pentameric transmembrane
protein complex in which four structurally related
subunits (a, b, c, d) in a stoichiometry of 2 : 1 : 1 : 1
assemble to form a central cation-selective ion channel
[1,2]. The a and b subunits of the Torpedo receptor
referred to in this report correspond to the a1 and b1
subunits in the nomenclature recommended by the
International Union of Pharmacology [3]. Radioligand
binding studies have demonstrated that, under equilib-
rium conditions, the nAChR carries two high affinity
Keywords
acetylcholine; loop D; mutagenesis; nicotinic
receptor; oocytes
Correspondence
S.M.J. Dunn, Department of Pharmacology,
University of Alberta, Edmonton, Alberta,
T6G 2H7 Canada

Fax: +780 4924325
Tel: +780 4923414
E-mail:
(Received 22 October 2005, revised 13
December 2005, accepted 23 December
2005)
doi:10.1111/j.1742-4658.2006.05121.x
The Torpedo nicotinic acetylcholine receptor is a heteropentamer (a
2
bcd)in
which structurally homologous subunits assemble to form a central ion
pore. Viewed from the synaptic cleft, the likely arrangement of these sub-
units is a–c–a–d–b lying in an anticlockwise orientation. High affinity bind-
ing sites for agonists and competitive antagonists have been localized to
the a–c and a–d subunit interfaces. We investigated the involvement of
amino acids lying at an adjacent interface (c–a) in receptor properties.
Recombinant Torpedo receptors, expressed in Xenopus oocytes, were used
to investigate the consequences of mutating aArg55 and cGlu93, residues
that are conserved in most species of the peripheral nicotinic receptors.
Based on homology modeling, these residues are predicted to lie in close
proximity to one another and it has been suggested that they may form a
salt bridge in the receptor’s three-dimensional structure (Sine et al. 2002 J
Biol Chem 277, 29 210–29 223). Although substitution of aR55 by phenyl-
alanine or tryptophan resulted in approximately a six-fold increase in the
EC
50
value for acetylcholine activation, the charge reversal mutation
(aR55E) had no significant effect. In contrast, the replacement of cE93 by
an arginine conferred an eight-fold increase in the potency for acetyl-
choline-induced receptor activation. In the receptor carrying the double

mutations, aR55E-cE93R or aR55F-cE93R, the potency for acetylcholine
activation was partially restored to that of the wild-type. The results sug-
gest that, although individually these residues influence receptor activation,
direct interactions between them are unlikely to play a major role in the
stabilization of different conformational states of the receptor.
Abbreviations
5-HT
3A
receptor, serotonin type 3 A receptor; a-BgTx, alpha-bungarotoxin; ACh, acetylcholine; AChBP, acetylcholine binding protein;
dTC, d-tubocurarine; GABA, c-aminobutyric acid; LGIC, ligand-gated ion channel; nAChR, nicotinic acetylcholine receptor; PTMA,
phenyltrimethylammonium; WT, wild-type.
960 FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS
binding sites for agonists and competitive antagonists
[4,5]. It is now generally agreed that these sites lie at
the interfaces between the a–c and the a–d subunits
[6]. Labeling and mutational studies have identified
several key amino acids lying in discrete noncontigu-
ous ‘loops’ of the a-subunits (designated as loops
A–C, the ‘primary component’), together with amino
acids in the neighboring c and d subunits (lying in
loops D–F, the ‘secondary component’) that partici-
pate in forming these binding pockets [7–9].
Although none of the ligand-gated ion channel fam-
ily to which the nAChR belongs has been crystallized,
the published structure of a related protein [10], the
acetylcholine binding protein (AChBP), lends credence
to current ideas of high affinity binding site location.
The AChBP, which is secreted by the glial cells of the
snail, Lymnaea stagnalis , is a truncated homologue
of the extracellular amino terminal domains of the

nAChR[see 10]. Inspection of its structure has rein-
forced earlier predictions that the residues involved in
forming the binding sites occur at subunit–subunit
interfaces and that the stretches of amino acids that
have been implicated in binding are arranged in loop-
like structures.
The structural homology of all subunits in the LGIC
family suggests that each of the five subunit–subunit
interfaces contributes to ligand binding and ⁄ or the
conformational changes that are involved in the trans-
duction mechanism(s) that link agonist binding to
channel opening. This is particularly true of the homo-
pentameric receptors, e.g. the a7 neuronal nAChR and
the 5-HT
3A
homomeric receptor. In these receptors,
there are five identical interfaces that presumably play
equivalent roles in ligand recognition and receptor
function. In the heteromeric receptors, the roles of all
five subunit interfaces are less clear. Due to structural
homology, all subunits carry all putative binding loops
(A–F), suggesting that each interface has the potential
to form a binding site, albeit with a distinct affinity
arising from the nonequivalence of the intersubunit
contacts within the pentamer. Alternatively, these
homologous loops at each interface may contribute to
receptor assembly and ⁄ or play a role in the conforma-
tional changes that result in channel activation or
receptor desensitization.
In the case of the Torpedo nAChR, the importance

of the a-subunit in ligand binding has long been recog-
nized [11-15], but the involvement of non a-subunit
residues has become clear only more recently. The first
direct evidence for the contribution of the c and d sub-
units to ligand recognition came from photoaffinity
labeling studies using [
3
H]nicotine and [
3
H]d-tubocura-
rine (dTC), which identified residues cW55 and the
homologous dW57 (lying in what is now referred to as
the loop D domain) as specific sites of ligand incorpor-
ation [5,15-17]. In the present study, we have investi-
gated the effects of mutations of the equivalent residue
(aR55) lying in loop D of the a-subunit, i.e. at the
opposite side of the subunit from residues (in loops A–
C) that have previously been implicated in agonist
binding (Fig. 1). Within the LGIC family, this residue
in the peripheral nAChR is unique; whereas almost all
subunits in the family have an aromatic residue at this
position, a positively charged arginine residue is con-
served in all peripheral a-subunits (see Fig. 1). Previ-
ous comparative modeling studies have revealed that
E93 of the c-subunit (lying in putative binding loop A)
may lie in close proximity to aR55, leading to the pro-
posal that an ionic interaction between these two resi-
dues may stabilize receptor conformation [18,19]. This
53
A

B
54 55 56 57 58 59 60
T.Ca_nAChR
α 1
NV
R
LRQQ W
H_nAChR
α1
NV
R
LKQQ W
T.Ca_nAChR
γ
NVWI E I Q W
T.Ca_nAChR
δ
NVWMDH A W
Rat_GABA
A
β2
TMYFQQ A W
Rat_5-HT
3
AYIWYRQF W
AChBP
VFWQQT T W
A
C
h

A
C
h
A
B
C
D
E
F
A
B
C
D
D
D
E
F
-
+
+
-
+-
α
α
β
δ
γ
Fig. 1. Loop D of the LGIC family. (A) Amino acid sequence align-
ments of residues lying in loop D of the a1, c and d subunits from
Torpedo californica (T. Ca) nAChR, human (H) a1 nAChR subunit,

b2 of rat GABA
A
receptor, rat 5-HT
3A
subunit and AChBP. Number-
ing shown is for the Torpedo nAChR a1 subunit. The positively
charged R55 residue is unique to the peripheral nAChR a1 subunit
since other members of the LGIC family have an aromatic amino
acid in this position. (B) Schematic representation of the subunit
arrangement of the Torpedo nAChR showing the ‘six binding loop’
model of high affinity ligand binding sites. Also represented is loop
D of the a-subunit (not previously implicated in ligand binding),
which lies at the b-a and c–a subunit interfaces.
A. Kapur et al. Role of a–c subunit interface in nAChR function
FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS 961
residue is also conserved in peripheral nAChR c (and
e) subunits. We therefore also investigated the effects
of a charge reversal mutation of this residue (cE93R)
both alone and in combination with the aR55 muta-
tions. Our results demonstrate that mutations of these
residues, which lie at an interface (c–a) that has not
previously been implicated in receptor function, can
have significant effects on ligand binding and ⁄ or chan-
nel gating. However, we conclude that a direct inter-
action between aR55 and cE93 is unlikely to make a
major contribution to nAChR properties.
Results
Functional effects of aR55 mutations
The functional responses of wild-type (WT) or mutant
receptors expressed in Xenopus oocytes were studied

using two-electrode voltage clamp techniques. Figure 2
shows the concentration-effect curves for ACh-medi-
ated responses. The WT nAChR receptor has an EC
50
value for ACh-induced activation of $24 lm with an
estimated Hill coefficient of 1.6. The substitution of
aArg55 with glutamic acid (aR55E) or lysine (aR55K)
resulted in a statistically insignificant shift in the EC
50
values for ACh activation to 29 and 47 lm, respect-
ively, and had no significant effect on the cooperativity
of receptor activation. In contrast, the aR55F and
aR55W mutations caused a five- to six-fold shift in the
EC
50
for ACh activation to 112 and 151 lm, respect-
ively. In addition, the Hill coefficients for Ach-induced
activation for these mutant receptors were significantly
reduced in comparison with the WT nAChR (Table 1).
The effects of phenyltrimethylammonium (PTMA)
on activation of WT and mutant receptors were also
investigated. PTMA is a poor partial agonist of the
WT nAChR and it elicits a maximum current of only
1.5 ± 0.1% of the ACh response (data not shown).
The WT receptor was activated by PTMA with an
EC
50
of 57 lm and a Hill coefficient of 2.1 ± 0.2
(Fig. 3A, Table 2). In contrast, PTMA failed to acti-
vate the aR55F and aR55W mutant receptors, even at

concentrations up to 10 mm. Instead, PTMA acted as
a competitive antagonist of these mutant receptors (see
Fig. 3A). Co-application of PTMA and ACh to the
aR55F (Fig. 3A) and aR55W (data not shown) recep-
tors resulted in a concentration-dependent inhibition
of ACh-evoked currents with apparent K
I
values of
103 and 88 lm, respectively. Thus PTMA-induced
channel activation (in WT nAChR) and inhibition (in
mutant receptors) occurs over a similar concentration
range suggesting that, although the mutations affected
the apparent efficacy of this ligand, they had little
effect on its affinity.
Effects of d-tubocurarine on aR55 mutant
nAChRs
We further examined the ability of the competitive ant-
agonist, dTC to inhibit ACh-evoked currents in WT
and mutant receptors (Fig. 3B). For WT nAChR,
Fig. 2. ACh activation of wild-type (WT) and aR55 mutant recep-
tors. Concentration-effect curves obtained from oocytes expressing
WT (n), aR55E (h), aR55K (n), aR55W (e)andaR55F (s) nAChR.
Data are normalized to I
max
for each individual point. The data rep-
resent the mean ±
SEM from at least three oocytes. The data
obtained from curve-fitting are summarized in Table 1.
Table 1. Concentration-effect data for ACh activation of wild-type
(WT) and mutant receptors expressed in Xenopus oocytes. Data

represent the mean ±
SEM. Values for log EC
50
and Hill coefficient
(n
H
) were determined from concentration-effect curves using GRAPH-
PAD PRISM
software. Log EC
50
and Hill coefficients from individual
curves were averaged to generate final mean estimates. The val-
ues in parentheses are the number of oocytes used for each recep-
tor type. Statistical analysis was performed by comparing the log
EC
50
and n
H
of the mutant receptors to the WT nAChR (
a
p<0.001,
b
p<0.05) using one-way analysis of variance (ANOVA) followed by
Dunnett’s post-test to determine the level of significance.
c
p<0.001
compared with the aR55F receptor.
Receptor
Log EC
50

± SEM
(M)
EC
50
(lM)n
H
± SEM
EC
50
mutant ⁄
EC
50
WT
WT ) 4.61 ± 0.04 (9) 24.3 1.6 ± 0.1 1
aR55E ) 4.54 ± 0.12 (3) 28.6 1.2 ± 0.1 1.2
aR55K ) 4.33 ± 0.08 (3) 47.2 1.2 ± 0.1 1.9
aR55F ) 3.95 ± 0.06 (4)
a
112 0.8 ± 0.02
a
4.6
aR55W ) 3.82 ± 0.18 (4)
a
151 0.8 ± 0.1
a
6.2
cE93R ) 5.52 ± 0.10 (3)
a
3.03 1.4 ± 0.02 0.12
cE93R-aR55E ) 4.94 ± 0.07 (7)

b,c
11.5 1.1 ± 0.07 0.47
cE93R-aR55F ) 4.89 ± 0.20 (3)
c
12.9 1.3 ± 0.1 0.53
Role of a–c subunit interface in nAChR function A. Kapur et al.
962 FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS
preperfusion with dTC produced a concentration
dependent inhibition of ACh-evoked currents charac-
terized by an apparent K
I
of $42 nm. dTC also inhib-
ited Ach-evoked currents in the receptors carrying the
aR55F and aR55W mutations with apparent K
I
values
of $52 and 34 nm, respectively (see Table 2). These
results suggest that mutations at position 55 of the
a-subunit do not affect either the binding affinity for
dTC or its ability to competitively inhibit ACh-evoked
currents. In these experiments, although dTC alone did
not elicit detectable whole cell currents, we observed
that low concentrations of dTC (1–3 nm) potentiated
ACh- evoked currents (by up to 25%) in both WT and
mutant receptors (Fig. 3B).
Expression levels and maximum amplitude
of WT and mutant nAChR
Fig. 4 compares the density of binding sites for
125
I-

labelled a-BgTx for the WT and mutant receptors with
the maximum ACh-evoked current. Injection of 50 ng
of WT subunit cRNAs resulted in a robust expression
of
125
I-labelled a-BgTx binding sites (approximately
3.1 fmolÆoocyte
)1
). All of the aR55 mutations were
well tolerated and, after their coexpression with WT
b-, c- and d- subunits, their expression levels (in terms
of
125
I-labelled a-BgTx binding sites) were in the same
range as the WT nAChR. Although the expression
levels of the receptors carrying the R55K and R55E
mutations were statistically higher than that of the
WT, the normalized currents (nAÆfmol
)1
) were com-
parable. In contrast, after normalization to the density
of toxin binding sites, the peak currents mediated by
A
B
Fig. 3. The effects of PTMA and dTC on WT and aR55 mutant
receptors. (A) Data show concentration-effect curves for the activa-
tion of WT nAChR(n) and inhibition of the ACh-induced response of
aR55F (s) mutant receptors (Table 2, see text for details). (B) Inhi-
bition of ACh-evoked currents by dTC acting on the WT (n)and
R55F (s) mutant receptors (see Table 2). For each receptor, the

ACh concentration used to induce responses was equivalent to its
EC
50
value for activation of that subtype. Similar data show a lack
of significant effect of dTC on the aR55W and R55E mutant recep-
tors (data not shown).
Table 2. Effects of PTMA and dTC on WT and mutant receptors.
Data were analyzed as described in the legend to Fig. 3. K
I
values
were determined as described in Experimental procedures. Each
experiment was repeated in 3–4 oocytes for each receptor sub-
type. No significant differences were observed between the WT
and mutant receptors.
Receptor
PTMA dTC
log IC
50
± SEM K
I
(lM) log IC
50
± SEM K
I
(nM)
WT ) 4.24 ± 0.06
a
57.0
a
) 7.07 ± 0.11 42.5

aR55F ) 3.68 ± 0.04 103 ) 7.16 ± 0.10 52.0
aR55W ) 3.79 ± 0.18 88 ) 7.01 ± 0.07 34.0
cE93R – – ) 6.77 ± 0.2 54.7
a
Data obtained from PTMA-induced channel activation (Hill coeffi-
cient ¼ 2.1 ± 0.2). All other data are from the effects of the ligand
on ACh-induced currents.
WT
α
R55K
α
R55E
α
R55F
α
R55W
0
2
4
6
8
10
I
max
fmol
0
2
4
6
8

10
I
amx
(
µA)
Surface nAChR (fmol)
Fig. 4. Surface nAChR expression of WT and mutant receptors in
Xenopus oocytes. Maximum ACh-evoked currents (I
max
) were
determined using concentrations determined from concentration-
effect curves (as shown in Fig. 2). Surface receptor levels were
determined in the same oocytes by measuring
125
I-labelled a-BgTx
binding as described in Experimental procedures. The data repre-
sent the mean ±
SEM of 3–11 determinations from individual
oocytes and are presented in Table 3.
A. Kapur et al. Role of a–c subunit interface in nAChR function
FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS 963
the aR55F and aR55W mutants were reduced by
approximately three- to five-fold, respectively, com-
pared with the WT receptor (Table 3; see Discussion).
Overall, these results suggest that mutation of aR55,
which is predicted to lie at the a–c and a–b interfaces,
does not play a major role in receptor assembly or sur-
face expression.
Influence of aR55F and aR55W mutant receptors
on the binding of acetylcholine

The binding properties of ACh were investigated in
intact Xenopus oocytes expressing WT and mutant
nAChR. The affinity of the mutant receptors for ACh
was characterized by its inhibition of the initial rate
of
125
I-labelled a-BgTx binding to Torpedo nAChR
expressed on the surface of oocytes (Fig. 5). ACh
inhibited the initial rate of
125
I-labelled a-BgTx binding
to the WT nAChR in a concentration-dependent man-
ner with an IC
50
of 544 nm (n
H
¼ 0.8). The IC
50
(n
H
)
of the aR55F and aR55W mutants was estimated to
be 454 nm (1.0) and 313 nm (0.9), respectively, and did
not differ from that of the WT nAChR. Thus, despite
the reduction in the potency of ACh in mediating
functional responses in the mutant receptors, the equi-
librium (high affinity) binding of ACh appears to be
unaltered.
Effects of cE93R mutation on the sensitivity of
agonist and antagonist

As noted in the Introduction, it has been suggested
that the proximity of aR55 and c93E may be con-
ducive to an ion-pairing interaction (see Fig. 6).
Surprisingly, the cE89R mutation resulted in an
approximately eight-fold increase in the apparent
potency of ACh-induced activation. As shown in
Fig. 7A (see Table 1), the EC
50
for this mutant recep-
tor was reduced to 3 lm (Hill coefficient of 1.4) from
the value of 24 lm measured in the WT. In contrast,
the apparent affinity for the competitive antagonist,
dTC (as determined by inhibition of ACh-evoked cur-
rents in oocytes), for the cE93R mutant receptor was
unaltered as compared with the oocytes expressing WT
nAChR (K
I
$55 and 42 nm, respectively, see Fig. 7B,
Table 2). This figure also illustrates that, in contrast to
the WT receptor (see above), the potentiating effects
of low concentrations of dTC were abolished by the
cE93R mutation.
Effects of double mutations of aR55 and cE93
In order to investigate whether the aR55F and cE93R
mutations have an additive effect, we studied receptors
carrying the double mutations, aR55E-cE93R and
aR55F-cE93R. These double mutant receptors had
EC
50
values for ACh-induced activation of $11.5 and

12.9 lm (Fig. 7A, Table 1), which approach those of
the WT receptors. The Hill coefficients for these dou-
ble mutants were not significantly different from the
WT receptor.
Discussion
The conformational changes that result in activation
of the nAChR channel are poorly understood, but are
thought to involve an agonist-induced rotation of the
Table 3. Surface expression and normalized ACh-evoked maximum
currents in WT and mutant receptors expressed in Xenopus
oocytes. All oocytes were injected with 50 ng of total cRNA enco-
ding WT or mutant subunit nAChR. The value in parentheses is the
number of oocytes used for each receptor type.
Surface
binding
(fmolÆoocyte
)1
± SEM)
I
max
(nA ± SEM)
Normalized
peak
current
(nAÆfmol
)1
)
%peak
current
(mutant I

max
⁄
WT I
max
)
WT 3.1 ± 0.8 (11) 3316 ± 537 1081 100
aR55E 9.2 ± 0.7 (3) 8680 ± 501 934.4 86.4
aR55K 8.0 ± 1.5 (5) 3652 ± 654 455.1 42.1
aR55F 3.8 ± 0.7 (6) 1445 ± 340 384.6 35.6
aR55W 5.9 ± 1.3 (7) 1154 ± 210 197.0 18.2
-10
-9
-8
-7
-6
-5
-4
0
20
40
60
80
100
120
log [ACh] (M)
[%
521
]I
α
-B xTgBiidngn

Fig. 5. ACh binding to Torpedo nAChR expressed on the surface of
intact Xenopus oocytes. ACh inhibited the initial rate of
125
I-labelled
a-BgTx binding in a concentration dependent manner in WT (n)and
aR55F receptors (s). The data represent the mean ±
SEM of 2–3
determinations performed in duplicate giving log IC
50
± SEM values
of )6.26 ± 0.14 and )6.34 ± 0.06, respectively (IC
50
s of 544 and
454 n
M, respectively). Similar experiments with the R55W mutant
gave a log IC
50
value of )6.50 ± 0.20 (IC
50
of 313 nM). There were
no significant differences between any of these receptor subtypes.
Role of a–c subunit interface in nAChR function A. Kapur et al.
964 FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS
subunits that is eventually communicated to the ion
channel pore [20,21]. It is widely accepted that the
nAChR carries two high affinity binding sites located
at the interfaces between a–c and a–d subunits [7], but
the involvement of other interfaces in receptor function
is relatively unexplored. In the present study, we have
therefore investigated the role of specific residues lying

at an adjacent interface, i.e. residues aR55 (loop D)
and cE93 (loop A) which, in the muscle counterpart,
have been proposed to interact and to possibly play a
critical functional role in receptor properties [19].
Amino acid sequence alignments of loop D (see
Fig. 1A) reveal that the peripheral nAChR a-subunits
carry a unique amino acid at position 55, i.e. an argin-
ine residue rather that an aromatic amino acid that is
conserved in most other subunits of the Cys-loop
LGIC family. There is considerable evidence to suggest
that, in a number of subunits, the residue in the equiv-
alent position plays an important role(s) in modulating
agonist ⁄ antagonist sensitivity. Mutations of cW55 and
dW57 in the Torpedo nAChR have been shown to
affect the affinity for dTC and ACh [17,22] while the
W54 of the neuronal nicotinic a7 receptor has been
shown to contribute to the binding of agonists [23]. In
the GABA
A
receptor, the F64L mutation of the a1
subunit had a dramatic effect on GABA sensitivity
[24] and mutations of the F77 residue of the c2 subunit
significantly affected ligand affinity for the benzodi-
azepine binding site [25]. In addition, the GABA
A
receptor b2Y62 residue has been shown to be an
important determinant of high affinity agonist binding
[26]. In the 5-HT
3A
receptor, the homologous W89

residue has been reported to contribute to both dTC
and granisetron binding [27]. Thus, the conserved aro-
matic residue in this position of most LGIC subunits
appears to play an important role and the unusual
occurrence of a positively charged amino acid in the
peripheral nAChR a-subunit first prompted the present
investigation.
The present results demonstrate that R55, which is
conserved in the peripheral a-subunits, is not essential
for subunit assembly, as mutations in this position did
not have a detrimental effect on the expression of
125
I-
labelled a-BgTx binding sites on the oocyte surface.
Not surprisingly, the conservative substitution, aR55K
had no significant effect on the concentration depend-
ence of ACh-induced activation. More surprisingly, the
charge reversal mutation, aR55E, also had no signifi-
cant effect on receptor activation properties. These
results are strong evidence that a positively charged
residue in this position of the peripheral nAChR
a-subunit is neither obligatory for agonist recognition
nor does it play a major role in the transduction mech-
anism that couples agonist binding to channel activa-
tion.
The aR55F and aR55W mutations resulted in a
modest but significant decrease in the sensitivity to
ACh (by approximately five- and six-fold, respectively).
In these mutant receptors, PTMA failed to induce
a measurable response. However, since PTMA is such

a poor agonist on the WT nAChR, the lack of a
response could be attributable to either a greatly
reduced sensitivity of the receptor towards the ligand
or to a further reduction of conductance to undetecta-
ble levels. The lack of any response to PTMA was
exploited to differentiate between these possibilities
[28], and the results reveal that the effects of the muta-
tions are on PTMA efficacy rather than affinity. The
apparent K
I
for PTMA-inhibition of ACh-responses
mediated by the aR55F and aR55W mutant receptors
γ
γ
α
N
O
N
N
N
O
N
O
O
5.7 Å
γE93 αR55
DINNELVI
Human ε
DVNNELVI
Human γ

DVNNELVV
Torpedo γ
DNNNQLVI
Torpedo δ
DANNYLVL
Torpedo β
DANNYLVL
Torpedo α
9796959493929190Loop A
Fig. 6. Representation of the extracellular
domains of the c–a subunits based on the
crystal structure of AChBP. The positions of
cE93 and aR55 at the subunit interface are
indicated and modeling suggests that these
residues are located approximately 6 A
˚
apart. Also shown are amino acid sequence
alignments of residues in loop A of repre-
sentative nAChR subunits.
A. Kapur et al. Role of a–c subunit interface in nAChR function
FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS 965
were similar to its EC
50
value for activation of the WT
receptor (Fig. 3A, Table 2) suggestive of an unaltered
affinity for its binding site(s).
A reduced sensitivity of the aR55F and aR55W
mutant receptors to ACh-activation was accompanied
by a reduced maximum current response. When the
measured peak currents were normalized to cell-surface

expression (nAÆfmol
)1
) the current responses were sub-
stantially lower than displayed by the WT receptor.
This may reflect a reduction in single channel conduct-
ance of the mutant receptors, a decreased efficacy of
ACh-mediated currents or the possibility that some of
the expressed receptors are nonfunctional. Distinction
between these possibilities requires further analysis at
the single channel level. We also observed a significant
reduction in the Hill slope of the activation curves in
the mutant receptors. While the interpretation of chan-
ges in Hill coefficients is controversial, the simplest
explanation is that these mutations reduce the level of
cooperativity between different agonist binding sites
[29].
Our present findings are consistent with previous
reports that mutations of the homologous residue
(W54) in the a7 nAChR W54 resulted in a reduction of
ACh potency without a disruption of a-BgTx binding
[23]. The present results point to a role of aR55 in the
transduction mechanism rather than in direct agonist
binding. However, there is some evidence in the litera-
ture that this region may also contribute to binding site
formation. A synthetic peptide equivalent to a55–74 of
Torpedo nAChR was shown to be able to bind a-BgTx
but this binding was inhibited by an R55G substitution
in the synthetic peptide [30]. However, since a synthetic
peptide is unlikely to have a similar conformation as
the equivalent domain in the native receptor, it is diffi-

cult to correlate the two sets of results.
The apparent affinity of the competitive antagonist,
dTC, was measured by its ability to inhibit ACh-
induced currents. The potency for dTC-induced inhi-
bition in WT and mutant receptors was similar
suggesting that the mutations had not altered dTC
affinity. When the binding of ACh was measured by
its ability to inhibit the initial rate of
125
I-labelled
a-BgTx binding to individual oocytes, its apparent
affinity was also unaltered from that of the WT recep-
tor. As these experiments are designed to measure high
affinity binding sites that exist under equilibrium con-
ditions, the results suggest that, under these circum-
stances, agonist binding has been unaltered. Thus the
major effect of the aR55F and aR55W mutations
appears to be on the potency for ACh-induced func-
tional responses, i.e. on the transition(s) between rest-
ing and activated states of the receptor.
In the case of the cE93R mutation, the ability of
low concentrations of dTC to potentiate ACh-induced
currents was apparently lost. Steinbach and Chen [31]
previously reported that dTC can act as a weak agon-
ist of the fetal muscle nAChR and suggested that, at
low dTC concentrations, the simultaneous binding of
one agonist molecule and one dTC molecule might eli-
cit channel opening. The above data on the WT recep-
tor are consistent with such a mechanism. The most
parsimonious explanation of the loss of the potentiat-

ing effect in the mutant receptor is that the mutation
results in the loss of the ability of dTC to act as such
a ‘coagonist’.
-8 -7 -6 -5 -4 -3 -2
0
20
40
60
80
100
120
A
B
log [ACh] (M)
%M ixamumr esnopse
-9 -8 -7 -6 -5
0
20
40
60
80
100
120
log [dTC] (M)
I
CTd +AChEC50
I/
hCAEC05
Fig. 7. Effects of the cE93R mutation. (A) Concentration-effect
curves for ACh activation of WT (n), cE93R (,) and the double

mutants, cE93R-aR55E (.)andcE93R-aR55F (e) nAChR. The data
represent the mean ±
SEM from at least three oocytes and are nor-
malized to the I
max
for each oocyte. Data are summarized in
Table 1. (B) Concentration dependent inhibition of ACh evoked cur-
rents by dTC in oocytes expressing the WT (n) and cE93R (,)
mutant receptors. The ACh concentration used in the experiments
corresponded to their EC
50
concentrations determined for each
receptor. Each curve was generated from at least three oocytes.
The apparent K
I
of dTC on oocytes expressing the cE93R receptors
(55 n
M) is not significantly different from the WT nAChR (42 nM).
Role of a–c subunit interface in nAChR function A. Kapur et al.
966 FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS
Although the c–a interface has not previously been
implicated in ligand binding or receptor function,
structural models of the adult human nAChR based
on AChBP have suggested putative interactions at this
interface i.e. a salt bridge between eE93 (a loop A resi-
due) and aR55 (a loop D residue [18,19]). Our results
obtained with the cE93R substitution clearly demon-
strate that this is a ‘gain-of-function’ mutation that
results in an approximately eight-fold decrease in the
EC

50
for ACh activation. One possible explanation is
that this mutation facilitates the rotational movements
at intersubunit contact points that have been suggested
to occur during channel activation [20,21,32]. The
receptor carrying the double charge-reversal mutation
(cE93R-aR55E) was activated by ACh with an EC
50
that approached that of the WT receptor, although the
EC
50
values (Table 1) remained statistically different.
Taken together, these results suggest that, in the WT
receptor, an interaction between aR55 and cE93 is
unlikely to stabilize either the resting conformation (as
the mutation aR55E had little effect on activation) or
the activated state (as mutation cE93R increased ACh
potency). However, these residues lying at the c–a
interface do appear to play a role in receptor activa-
tion and ⁄ or the signal transduction mechanism.
In summary, we have identified a residue, R55 in
loop D of the extracellular ligand binding domain of
a-subunit that modulates ACh sensitivity and that lies
at some distance from the ‘classical’ high affinity bind-
ing sites for ACh. This residue has not previously been
implicated in nAChR function. However, our data
complement earlier work to suggest that loop D resi-
dues occurring in nonbinding domains may play
important roles in receptor function[see 26]. In addi-
tion, we show that E93 of the Torpedo nAChR c-sub-

unit has a significant effect on agonist-induced
activation, as substitution by the positively charged
arginine increased ACh potency by approximately
eight-fold. Although our data do not support a critical
role for a direct interaction between aR55 and cE93,
they demonstrate that residues lying at interfaces adja-
cent to those that have been implicated in agonist
binding influence receptor function.
Experimental procedures
Materials
ACh, a-BgTx and dTC were obtained from Sigma-RBA
(Natick, MA, USA).
125
I-labelled a-BgTx (2000 CiÆmmol
)1
)
was from Amersham Life Science (Arlington Heights, IL,
USA). Restriction enzymes and cRNA transcript prepar-
ation materials were purchased from Invitrogen (Burling-
ton, ON, Canada), Promega (Madison, WI, USA) or from
New England Biolabs (NEB, Pickering, ON, Canada). Pfu
Turbo DNA polymerase for mutagenesis experiments was
from Stratagene (La Jolla, CA, USA). All other chemicals
were obtained from Sigma or other standard sources. The
a-, b- (in the SP64 plasmid) and d-subunit (in the SP65
plasmid) cDNA clones of the Torpedo nAChR were gener-
ous gifts from H. A. Lester (California Institute of
Technology, Pasadena, CA, USA). The c-subunit cDNA
(in the SP64-based plasmid, pMXT) was a gift from J. B.
Cohen (Harvard Medical School, Boston, MA, USA).

In vitro transcription and site-directed
mutagenesis
The plasmid cDNAs were linearized by digestion with
either EcoRI (for the a-subunit), FspI (for the b-subunit)
or XbaI (for the c- and d-subunits). In vitro cRNA tran-
scription was performed using the methods described by
Goldin and Sumikawa [33]. Briefly, the linearized cDNA
templates (5 lg) were transcribed in vitro using SP6 RNA
polymerase (Promega) in the presence of ribonucleotide
triphosphate (NTP mix, Invitrogen) and RNA capping
analogue (NEB). The RNA transcripts were extracted
using 25 : 24 : 1 (v ⁄ v) phenol–chloroform–isoamyl alcohol.
Finally, the RNA pellets were resuspended in diethylpyro-
carbonate-treated water at a concentration of 1 lgÆlL
)1
.
The a-subunit mutants (R55F, R55W, R55K and R55E)
were constructed using Stratagene’s QuikChange site-direc-
ted mutagenesis protocol. Synthetic oligonucleotide muta-
genic primers were typically 23–34 base pairs long (with
10–15 base pairs lying on either side of the mismatch
region). A similar approach was undertaken to engineer the
cE93R mutation. Restriction endonuclease digestion and
DNA sequencing subsequently verified the presence of the
mutation.
Expression in Xenopus oocytes and
electrophysiology
Isolated, follicle-free oocytes were microinjected with 50 ng
of total subunit cRNAs in a ratio of 2a :1b :1c :1d.
Oocytes were maintained in ND96 buffer (96 mm NaCl,

2mm KCl, 1.8 mm CaCl
2
,1mm MgCl
2
,5mm Hepes,
pH 7.6) supplemented with 50 lgÆ mL
)1
gentamicin at 14 °C
for at least 48 h prior to recording. Currents elicited by
bath application of agonist were measured with a Gene-
Clamp 500 amplifier (Axon Instruments, Foster City, CA,
USA) using standard two-electrode voltage clamp tech-
niques at a holding potential of )60 mV. Electrodes were
filled with 3 m KCl and those with resistances of 0.5–3.0
MW were used. The recording chamber was perfused con-
tinuously (at a flow rate of $5mLÆmin
)1
) with low Ca
2+
ND96 buffer, in which the CaCl
2
concentration was
A. Kapur et al. Role of a–c subunit interface in nAChR function
FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS 967
reduced to 0.1 mm in order to slow the rate of receptor
desensitization [34]. Atropine (1 lm) was included in the
perfusion buffer to block endogenous muscarinic receptors
present in the oocytes [35]. Agonist-evoked responses were
measured by applying the drug via the perfusion system for
15–20 s followed by a 15-min wash-out period to ensure

full recovery from desensitization. For measuring the effects
of antagonists, oocytes were preperfused with various con-
centrations of antagonist in low Ca
2+
ND96 buffer for
2 min, before initiating the response by application of solu-
tion containing ACh (at a concentration eliciting 50% of
the maximum response, EC
50
) and including the same con-
centration of antagonist as used in the preperfusion.
Binding of
125
I-labelled a-BgTx to intact oocytes
Binding assays were performed on individual oocytes that
had previously been used in the electrophysiological experi-
ments. To measure the density of nAChR binding sites
expressed on the oocyte surface (fmol), oocytes were incu-
bated with 5 nm
125
I-labelled a-BgTx in a final volume of
100 lL of low Ca
2+
ND96 buffer containing 5 mgÆmL
)1
bovine serum albumin for 2 h at room temperature [36,37].
Excess unbound toxin was removed by washing the oocytes
three times with 1 mL of ice-cold low Ca
2+
ND96 buffer.

This procedure was performed by manually transferring the
oocyte from one solution to another using a broad micropi-
pette tip to pick up the oocyte with minimal transfer of the
original solution. Non-specific binding was estimated by
carrying out parallel studies of
125
I-labelled a-BgTx binding
to uninjected oocytes. Non-specific binding determined in
oocytes expressing WT or mutant receptor in the presence
of excess cold ACh was comparable to that estimated using
uninjected oocytes (data not shown). Bound
125
I-labelled
a-BgTx was measured by c-counting (Gamma8000, Beck-
man). Using these data, the maximum currents (I
max
) meas-
ured for WT and mutant receptors were normalized to the
concentration of binding sites in terms of nAÆfmol
)1
. For
competition curves, oocytes were incubated for 40 min with
various concentrations of ACh in a 96-well plate prior to
the addition of 2.5 nm
125
I-labelled a-BgTx. After 40 min,
125
I-labelled a-BgTx binding was stopped by the addition
of 1 lm unlabeled a-BgTx. In the absence of the competing
ligand, ACh,

125
I-labelled a-BgTx binding was $30% of the
available a-BgTx-binding sites (data not shown). Non-speci-
fic binding was determined in the presence of 100 mm ACh.
Data and statistical analysis
Competition and concentration-effect curves for both elec-
trophysiological and radioligand binding experiments were
analyzed by nonlinear regression techniques using graph-
pad prism 3.0 software (GraphPad, San Diego, CA, USA).
Data from individual oocytes were normalized to the I
max
value obtained for that oocyte.
For receptor activation, concentration-effect curves for
agonist activation were analyzed using the following equa-
tion:
I ¼ I
Ã
max
½L
n
=ðEC
50
þ½LÞ
n
where I is the measured agonist-evoked current, [L] is the
agonist concentration, EC
50
is the agonist concentration
that evokes half the maximal current (I
max

) and n is the
Hill coefficient. In each experiment, the current (I) is nor-
malized to the I
max
and the normalized data are presented
as percentage response.
The IC
50
was determined from competition–inhibition
curves by fitting to the following equation:
f ¼ 100=½1 þð½X=½IC
50
Þ
n

where f is the fractional (%) response remaining in the
presence of inhibitor at concentration [X], IC
50
is the inhib-
itor concentration that reduced the amplitude of ACh-
evoked current by 50% and n is the Hill coefficient. ACh
inhibition of initial rate of
125
I-labelled a-BgTx binding was
also fit by the above equation.
The K
I
(apparent) value was calculated using the Cheng-
Prusoff equation [38]:
K

I
¼ IC
50
=½1 þ½L=EC
50

where [L] is the ACh concentration used in the experiment
and EC
50
is the ACh concentration that evokes half the
maximal current.
Statistical analysis was performed using one-way analysis
of variance (anova, />anova.html) followed by Dunnett’s post-test to determine
the level of significance.
Acknowledgements
This work was supported by the Canadian Institutes
of Health Research. We are especially grateful to
Isabelle Paulsen for expert technical assistance.
References
1 Raftery MA, Hunkapiller MW, Strader CD & Hood
LE (1980) Acetylcholine receptor: complex of homolo-
gous subunits. Science 208, 1454–1456.
2 Dunn SMJ (1993) Structure and function of the nico-
tinic acetylcholine receptor. Adv Struc Biol 2, 225–244.
3 Lukas RJ, Changeux JP, Le Nove
`
re N, Albuquerque
EX, Balfour DJ, Berg DK, Bertrand D, Chiappinelli
VA, Clarke PB, Collins AC et al. (1999) International
Union of Pharmacology. XX. Current status of the

nomenclature for nicotinic acetylcholine receptors and
their subunits. Pharmacol Rev 51, 397–401.
4 Blount P & Merlie JP (1989) Molecular basis of the two
nonequivalent ligand binding sites of the muscle nicoti-
nic acetylcholine receptor. Neuron 3, 349–357.
Role of a–c subunit interface in nAChR function A. Kapur et al.
968 FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS
5 Pederson SE & Cohen JB (1990) d-Tubocurarine bind-
ing sites are located at a–c and a–d subunit interfaces
of the nicotinic acetylcholine receptor. Proc Natl Acad
Sci USA 87, 2785–2789.
6 Grutter T & Changeux JP (2001) Nicotinic receptors in
wonderland. Trends Biochem Sci 26, 459–463.
7 Corringer PJ, Le-Nove
`
re N & Changeux JP (2000)
Nicotinic receptors at the amino acid level. Annu Rev
Pharmacol Toxicol 40, 431–458.
8 Arias HR (2000) Localization of agonist and competit-
ive antagonist binding sites on nicotinic acetylcholine
receptor. Neurochem Int 36, 595–645.
9 Karlin A (2002) Emerging structure of the nicotinic
acetylcholine receptors. Nat Neurosci 3, 102–114.
10 Brejc K, Dijk WJV, Klaassen RV, Schuurmans M, Oost
JVD, Smit AB & Sixma TK (2001) Crystal structure of
an ACh-binding protein reveals the ligand-binding
domain of nicotinic receptors. Nature 411, 269–276.
11 Kao PN, Dwork AJ, Kaldany RR, Silver ML, Wid-
eman J, Stein S & Karlin A (1984) Identification of the
alpha subunit half-cysteine specifically labeled by an

affinity reagent for the acetylcholine receptor binding
site. J Biol Chem 259, 11662–11665.
12 Dennis M, Giraudat J, Kotzyba-Hibert F, Goeldner M,
Hirth C, Chang JY, Lazure C, Chretien M & Changeux
JP (1988) Amino acids of the Torpedo marmorata acetyl-
choline receptor alpha subunit labeled by a photoaffinity
ligand for the acetylcholine binding site. Biochem 27,
2346–2357.
13 Galzi JL, Revah F, Black D, Goeldner M, Hirth C &
Changeux JP (1990) Identification of a novel amino acid
alpha-tyrosine 93 within the cholinergic ligands-binding
sites of the acetylcholine receptor by photoaffinity label-
ling: additional evidence for a three-loop model of the
cholinergic ligands-binding sites. J Biol Chem 265,
10430–10437.
14 Cohen JB, Sharp SD & Liu WS (1991) Structure of the
agonist-binding site of the nicotinic acetylcholine recep-
tor. J Biol Chem 266, 23354–23364.
15 Chiara DC & Cohen JB (1997) Identification of amino
acids contributing to high and low affinity d-tubocurar-
ine sites in the Torpedo nicotinic acetylcholine receptor.
J Biol Chem 272, 32940–32950.
16 Chiara DC, Middleton RE & Cohen JB (1998) Identifica-
tion of tryptophan 55 as the primary site of [
3
H]nicotine
photoincorporation in the c-subunit of the Torpedo nico-
tinic acetylcholine receptor. FEBS Lett 423, 223–226.
17 Xie Y & Cohen JB (2001) Contributions of Torpedo
nicotinic acetylcholine receptor cTrp-55 and dTrp-57 to

agonist and competitive antagonist function. J Biol
Chem 276, 2417–2426.
18 Sine SM (2002) The nicotinic receptor ligand binding
domain. J Neurobiol 53, 431–446.
19 Sine SM, Wang HL & Bren N (2002) Lysine scanning
mutagenesis delineates structural model of the nicotinic
receptor ligand binding domain. J Biol Chem 277,
29210–29223.
20 Miyazawa A, Fujiyoshi Y & Unwin N (2003) Structure
and gating mechanism of the acetylcholine receptor
pore. Nature 423, 949–955.
21 Lee WY & Sine SM (2005) Principal pathway coupling
agonist binding to channel gating in nicotinic receptors.
Nature 438, 243–247.
22 O’leary ME, Filatov GN & White MM (1994) Charac-
terization of d-tubocurarine binding site of Torpedo
acetylcholine receptor. Am J Physiol 266, C648–C653.
23 Corringer PJ, Galzi JL, Eisele JL, Bertrand S, Changeux
JP & Bertrand D (1995) Identification of a new compo-
nent of the agonist binding site of the nicotinic alpha 7
homooligomeric receptor. J Biol Chem 270 , 11749–
11752.
24 Sigel E, Baur R, Kellenberger S & Malherbe P (1992)
Point mutations affecting antagonist affinity and agonist
dependent gating of the GABA
A
receptor channels.
EMBO J 11, 2017–2023.
25 Buhr A, Baur R & Sigel E (1997) Subtle changes in resi-
due 77 of the gamma subunit of alpha1beta2gamma2

GABAA receptors drastically alter the affinity for
ligands of the benzodiazepine binding site. J Biol Chem
272, 11799–11804.
26 Newell JG, Davies M, Bateson AN & Dunn SMJ
(2000) Tyrosine 62 of the c-aminobutyric acid type A
receptor b2 subunit is an important determinant of
high affinity agonist binding. J Biol Chem 275, 14198–
14204.
27 Yan D, Schulte MK, Bloom KE & White MM (1999)
Structural features of the ligand-binding domain of the
serotonin 5HT3 receptor. J Biol Chem 274, 5537–5541.
28 O’leary ME & White MM (1992) Mutational analysis
of ligand-induced activation of the Torpedo acetylcho-
line receptor. J Biol Chem 267, 8360–8365.
29 Colquhoun D (1998) Binding, gating, affinity and effic-
acy: the interpretation of structure–activity relationships
for agonists and of the effects of mutating receptors.
Br J Pharmacol 125, 924–947.
30 Wahlsten JL, Lindstrom JM & Conti-Tronconi BM
(1993) Amino acid residues within the sequence region
a55–74 of Torpedo nicotinic acetylcholine receptor
interacting with antibodies to the main immunogenic
region and with snake a-neurotoxins. J Rec Res 13,
989–1008.
31 Steinbach JH & Chen Q (1995) Antagonist and partial
agonist actions of d-tubocurarine at mammalian muscle
acetylcholine receptors. J Neurosci 15, 230–240.
32 Unwin N (2005) Refined structure of the nicotinic acetyl-
choline receptor at 4A
˚

resolution. J Mol Biol 346,
967–989.
33 Goldin AL & Sumikawa K (1992) Preparation of RNA
for injection into Xenopus oocytes. Methods Enzmol
207, 279–297.
A. Kapur et al. Role of a–c subunit interface in nAChR function
FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS 969
34 Miledi R (1980) Intracellular calcium and the desensiti-
zation of acetylcholine receptors. Proc R Soc Lond B
209, 447–452.
35 Barnard EA, Miledi R & Sumikawa K (1982) Transla-
tion of exogenous messenger RNA coding for nicotinic
acetylcholine receptors produces functional receptors in
Xenopus oocytes. Proc R Soc Lond B 215, 241–246.
36 Sullivan DA & Cohen JB (2000) Mapping the agonist
binding site of the nicotinic acetylcholine receptor:
orientation requirements for activation by covalent ago-
nist. J Biol Chem 275, 12651–12660.
37 Tamamizu S, Guzman GR, Santiago J, Rojas LV,
Mcnamee MG & Lasalde-Dominicci JA (2000) Func-
tional effects of periodic tryptophan substitutions in the
a M4 transmembrane domain of the Torpedo californica
nicotinic acetylcholine receptor. Biochem 39, 4666–4673.
38 Cheng Y & Prusoff WH (1973) Relationship between
the inhibition constant (K1) and the concentration of
inhibitor which causes 50 per cent inhibition (I50) of
an enzymatic reaction. Biochem Pharmacol 22, 3099–
3108.
Role of a–c subunit interface in nAChR function A. Kapur et al.
970 FEBS Journal 273 (2006) 960–970 ª 2006 The Authors Journal compilation ª 2006 FEBS