Nicotinic pharmacophore: The pyridine N of nicotine and carbonyl of acetylcholine hydrogen bond across a subunit interface to a backbone NH

Article abstract of DOI:10.1073/pnas.1007140107

Pharmacophore models for nicotinic agonists have been proposed for four decades. Central to these models is the presence of a cationic nitrogen and a hydrogen bond acceptor. It is now well-established that the cationic center makes an important cation-π interaction to a conserved tryptophan, but the donor to the proposed hydrogen bond acceptor has been more challenging to identify. A structure of nicotine bound to the acetylcholine binding protein predicted that the binding partner of the pharmacophore's second component was a water molecule, which also hydrogen bonds to the backbone of the complementary subunit of the receptors. Here we use unnatural amino acid mutagenesis coupled with agonist analogs to examine whether such a hydrogen bond is functionally significant in the α4β2 neuronal nAChR, the receptor most associated with nicotine addiction. We find evidence for the hydrogen bond with the agonists nicotine, acetylcholine, carbamylcholine, and epibatidine. These data represent a completed nicotinic pharmacophore and offer insight into the design of new therapeutic agents that selectively target these receptors.

Full text of DOI:10.1073/pnas.1007140107

Nicotinic pharmacophore: The pyridine N of nicotine
and carbonyl of acetylcholine hydrogen bond across
a subunit interface to a backbone NH
Angela P. Bluma, Henry A. Lesterb, and Dennis A. Doughertya,1
aDivision of Chemistry and Chemical Engineering; and bDivision of Biology, California Institute of Technology, Pasadena, CA 91125
This contribution is part of the special series of Inaugural Articles by members of the National Academy of Sciences elected in 2009.
Contributed by Dennis A. Dougherty, June 2, 2010 (sent for review April 1, 2010)
H
H
N+
Pharmacophore models for nicotinic agonists have been proposed
for four decades. Central to these models is the presence of a
cationic nitrogen and a hydrogen bond acceptor. It is now well-
established that the cationic center makes an important cation-π
interaction to a conserved tryptophan, but the donor to the pro-
posed hydrogen bond acceptor has been more challenging to iden-
tify. A structure of nicotine bound to the acetylcholine binding
protein predicted that the binding partner of the pharmacophore’s
second component was a water molecule, which also hydrogen
bonds to the backbone of the complementary subunit of the recep-
tors. Here we use unnatural amino acid mutagenesis coupled with
agonist analogs to examine whether such a hydrogen bond is func-
tionally significant in the α4β2 neuronal nAChR, the receptor most
associated with nicotine addiction. We find evidence for the hydro-
gen bond with the agonists nicotine, acetylcholine, carbamylcho-
line, and epibatidine. These data represent a completed nicotinic
pharmacophore and offer insight into the design of new therapeu-
tic agents that selectively target these receptors.
A
Cl
N
N+
N+
H
H
N
S-MPP
Epibatidine
S-Nicotine
O
O
N+
N+
N+
O
H2N
O
HO
Acetylcholine
(ACh)
Carbamylcholine
(CCh)
Choline
(Ch)
O
i
O
O
i
O
H
N
H
N
H
N
H
N
B
N
H
O
i-1
O
i+1
i-1
O
i+1
O
O
H
H
H
H
Fig. 1. Key structures considered in the present work. (A) Structures of
agonists used. Hydrogen bond acceptor moieties are red and cationic nitro-
gens are blue. (B) Backbone amide to ester mutation strategy for perturbing
a hydrogen bond.
he nicotinic acetylcholine receptor (nAChR) is a pentameric,
T
ligand-gated ion channel activated by the neurotransmitter
ligand-binding domain, which contains the cation-π binding site,
and three segments from the β2 subunit (D, E, and F) form the
complementary face. A major advance in the study of nAChRs
was the discovery of the water-soluble acetylcholine binding pro-
teins (AChBP) (17–22). AChBP serves as a structural template
for the extracellular, N-terminal, ligand-binding domain of the
nAChRs, sharing 20–24% sequence identity with the ligand-bind-
ing domain of the much larger ion channel proteins. Several
AChBP structures with ligands bound have been published, in-
cluding structures of AChBP in complex with the ACh analog
carbamylcholine (CCh) and with nicotine (18) and the nicotine
analog epibatidine (21). Drugs that target the nAChR, such as
nicotine and epibatidine, typically contain a protonatable amine
rather than the quaternary ammonium seen in ACh. Along with
the cation-π interaction, the crystallography indicated a hydrogen
bond between the NþH and the backbone carbonyl of the tryp-
tophan that also forms the cation-π interaction, and functional
studies on intact receptors confirmed the hydrogen bonding
interaction (14, 23).
acetylcholine (ACh), and also by nicotine and structurally related
agonists (1–3). Nicotinic receptors mediate fast synaptic transmis-
sion at the neuromuscular junction of the peripheral nervous
system. In addition, a family of paralogous nAChRs termed
the neuronal receptors function in the central nervous system
and certain autonomic ganglia, and the addictive and cognitive
properties of nicotine are associated with these neuronal recep-
tors (4, 5). Neuronal receptors comprised of α4 and β2 subunits
are most strongly associated with nicotine addiction (6–9). They
are upregulated during chronic nicotine exposure and are impli-
cated in various disorders, including Alzheimer’s disease and
schizophrenia, and in protection against Parkinson disease. Inter-
est in the development of molecules that selectively target α4β2
receptors has been growing, highlighted by the development of
the smoking cessation drug, varenicline (6).
Many have undertaken the task of dissecting nicotinic agonists
into a core pharmacophore, since the first publication on the
topic in 1970 (10). While the details are debated, two aspects
are clear. Nicotinic agonists contain a cationic nitrogen and a
hydrogen bond acceptor (Fig. 1A) (11, 12). In 1990, we proposed
that binding of the cationic nitrogen of acetylcholine would be
mediated through a cation-π interaction with an aromatic residue
of the nAChRs (13). We subsequently validated this model with
the identification of a cation-π interaction to a conserved trypto-
phan residue for both acetylcholine and nicotine (14, 15). In fact,
the cation-π interaction has been shown to be a general contri-
butor to agonist affinity across the entire family of Cys-loop
(pentameric) neurotransmitter-gated ion channels (16).
Concerning the second component of the pharmacophore, the
hydrogen bond acceptor, the AChBP structure produced intri-
guing results. With nicotine bound, the pyridine nitrogen makes
a hydrogen bond to a water molecule that is positioned by hydro-
gen bonds to the main chains of two residues, the CO of N107 and
the NH of L119, both in the complementary subunit (α4β2
Author contributions: A.P.B. and D.A.D. designed research; A.P.B. performed research;
A.P.B., H.A.L., and D.A.D. analyzed data; and A.P.B., H.A.L., and D.A.D. wrote the paper.
The authors declare no conflict of interest.
In the nAChRs, ligand binding occurs at the interface between
adjacent principal (α4 in α4β2) and complementary (β2) subunits.
Three segments from the α4 subunit (historically referred to as
the A, B, and C “loops”) form the principal face of the
See Commentary on page 13195.
1To whom correspondence should be addressed. E-mail: dadougherty@caltech.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/
doi:10.1073/pnas.1007140107/-/DCSupplemental.
13206–13211 ∣ PNAS ∣ July 27, 2010 ∣ vol. 107 ∣ no. 30
www.pnas.org/cgi/doi/10.1073/pnas.1007140107

numbering; residues are in the β2 sequence FYSNAVVSYDG-
SIFWLPPA) (Fig. 2) (18). In other structures, including those
with CCh or epibatidine bound, the overall binding site structure
is preserved, although the key water molecule is not always
evident, especially in lower-resolution structures.
A key question, then, is the extent to which predictions based
on the AChBPs, which evolved to bind a target molecule, relate to
the nAChRs, which evolved to undergo a global structural change
(to gate) on binding ACh. Here we describe an approach to probe
with high precision a specific structural interaction in a complex
receptor protein. Using unnatural amino acid mutagenesis and
agonist analogs, we find that nicotine, acetylcholine, epibatidine,
and carbamylcholine make the same hydrogen bond involving
the backbone NH of β2L119 of the α4β2 receptor, supporting
a common pharmacophore for acetylcholine and nicotine.
ditive, this would provide compelling evidence for the proposed
interaction.
The metric used to evaluate receptors is EC50, the effective
concentration of agonist required to achieve half-maximal re-
sponse. This is a functional measure that can be influenced by
changes to drug binding and/or efficacy of activation of the recep-
tor. Previously we have shown that subtle mutations to TrpB of
the binding site primarily, if not exclusively, affect agonist binding
(14), but we cannot assume the same for Leu119. Because the
goal here is to map the pharmacophore for a collection of ago-
nists, we are interested in factors that influence receptor activa-
tion. We consider EC50 to be an appropriately useful guide for
understanding agonism and designing new agonists, but more
detailed studies of the mutations considered here would be
valuable.
Optimization of Nonsense Suppression Experiments. The α4β2 recep-
tor is a pentamer with two possible stoichiometries, ðα4Þ ðβ2Þ
and ðα4Þ3ðβ2Þ termed A2B3 and A3B2, respectively. Our s2tudie3s
have focused2on the A2B3 receptor, which shows the higher
sensitivity to nicotine and is thought to be upregulated during
chronic nicotine exposure. Subunit stoichiometry can be mana-
ged by controlling mRNA injection ratios. Exclusive expression
of A2B3 can be verified by monitoring I-V relationships of
agonist-induced currents, as described previously (14).
This study represents the first report of unnatural amino acid
mutagenesis in the β2 subunit of α4β2. Since nonsense suppres-
sion often produces low protein yields of the subunit where the
suppression occurs, it was critical to ensure that a receptor with
excess β2 subunit, i.e., the A2B3 stoichiometry, was exclusively
produced in nonsense suppression experiments. To that end,
mRNA ratios substantially favoring the β2 subunit were explored.
We found that an injected mRNA ratio of 1∶20 of α4∶β2 (with β2
containing the nonsense suppression site) gave I-V relationships
indicative of A2B3 (14), while still providing enough current to
conduct meaningful dose-response experiments. The α4 subunit
also contained a known mutation in the M2 transmembrane helix
(L9’A), which improves receptor expression and lowers whole-
cell EC50 values, but does not influence the binding trends of
the receptor (29).
One challenge in incorporating a hydroxy acid at β2L119 was
to limit the amount of current observed from oocytes injected
with full length tRNA that was not synthetically appended to
an amino or α-hydroxy acid. Such current would indicate that
the suppressor tRNA was aminoacylated by an endogenous
aminoacyl-tRNA synthetase and delivered a natural amino acid
at the mutation site. We observed significant background currents
attributable to such infidelity when using the suppressor tRNA
THG73, which has been the workhorse of our unnatural amino
acid mutagenesis experiments (30). Employing the recently
developed opal suppressor tRNA TQOpS’ (31, 32) significantly
reduced this background current at β2L119. Aminoacylation
from TQOpS’ was assessed for each agonist by injection of
unacylated TQOpS’, and full dose-response relations were gen-
erated for agonists displaying >20 nA of current. Suppression ex-
periments typically produced ≥1 μA of current and yielded Hill
and EC50 values that were markedly different from unacylated
TQOpS’ control experiments, and so the small background
currents are not expected to distort the reported EC50 values.
With these conditions, characterization of mutant receptors
was straightforward (Fig. 3).
Results
Strategy. A well-established strategy for probing potential back-
bone hydrogen bonds is to replace the residue that contributes
the hydrogen bond donor with its α-hydroxy analog (Fig. 1B)
(24–28). This mutation converts a backbone amide to a backbone
ester, a subtle change that impacts backbone hydrogen bonding in
two ways. The backbone NH that can donate a hydrogen bond is
removed, and the carbonyl oxygen, by virtue of being part of an
ester rather than an amide, is a weaker hydrogen bond acceptor.
In the present context, simply seeing a change in receptor func-
tion in response to appropriate backbone ester substitutions
would not prove the presence of the proposed interaction. Back-
bone mutation is certainly subtle, but when installed in an impor-
tant region of the receptor it could affect function in a number of
ways. As such, we sought a way to provide a direct connection
between any consequences of backbone mutation and the pro-
posed hydrogen bond. To do this, we considered the molecule
S-N-methyl-2-phenylpyrrolidine (S-MPP, Fig. 1A). In this struc-
ture a phenyl ring replaces the pyridyl group of nicotine, obliter-
ating the possibility of forming the proposed hydrogen bond.
This would allow a “double mutant cycle” analysis that links
the backbone NH to the pyridine N. If the mutant cycle analysis
shows that the effects of the two changes—the backbone muta-
tion and the modification of the drug—are substantially nonad-
Amide to Ester Backbone Mutation at β2L119 Impacts Receptor Func-
tion. To probe the hydrogen bond suggested by the AChBP
structures, β2L119 was replaced with its α-hydroxy analog (leu-
cine, α-hydroxy; Lah). Meaningful increases in EC50 for the back-
bone amide to ester mutation were seen for the conventional
agonists nicotine, ACh, CCh, and epibatidine, suggesting a signif-
Fig. 2. Key interactions seen in the crystal structure of nicotine bound to
AChBP (PDB ID code 1UW6). Residue numbering is for the α4β2 receptor, with
AChBP homologs in brackets.
Blum et al.
PNAS
∣
July 27, 2010
∣
vol. 107
∣
no. 30
∣
13207

Fig. 3. Representative current waveforms and dose-response relations for S-nicotine and S-MPP. Agonist-induced current waveforms for (A) S-Nic on wild-type
α4β2. (B) S-Nic on α4β2L119Lah. (C) S-MPP on wild-type α4β2. (D) S-MPP on α4β2L119Lah. Concentrations are in μM. (E) Dose-response relations for S-Nic and S-
MPP on wild-type α4β2 or α4β2L119Lah.
icant functional role for the backbone NH (Table 1 and Fig. 3). In
contrast, no shift was seen for the very weak agonist choline.
As noted above, we considered S-MPP as a potentially
informative structure for probing the pyridine hydrogen bond.
As such, we adapted existing synthetic protocols (33) to prepare
N-methyl-2-phenylpyrroline (MPP). Recrystallization of the
dibenzoyl tartrate salt (at the phenylpyrrolidine stage) gave the
S enantiomer.
As expected, S-MPP is a much poorer agonist than nicotine,
showing a ∼120-fold higher EC50 with the wild-type receptor.
For nicotine, the S enantiomer is the higher affinity enantiomer
and the one traditionally used in studies of nicotinic receptors.
We find that S-MPP has a twofold lower EC50 than racemic
MPP, indicating that the higher affinity enantiomer is being used.
Incorporation of a backbone ester at β2L119 leads to a remark-
able change in relative agonist potencies. Instead of the increase
in EC50 seen with nicotine, S-MPP actually shows a decrease in
EC50; S-MPP is a more potent agonist when the backbone ester
is present than when the natural backbone amide is present. In
fact, when the backbone ester is present, nicotine and S-MPP
display comparable potency.
The AChBP structure also predicts that a second residue in the
complementary subunit positions the water molecule in proximity
to the pyridine N of nicotine. The backbone carbonyl of β2N107 is
expected to make a hydrogen bond to the water molecule in con-
junction with the first hydrogen bond made by β2L119 (Fig. 2). As
noted above, an established strategy for attenuating the hydrogen
bonding ability of a backbone carbonyl is to mutate the (i þ 1)
residue to its α-hydroxy acid (Fig. 1B). However, nonsense sup-
pression experiments at the β2A108 site gave inconsistent results
that suggested we could not reliably control the stoichiometry of
the mutant receptor. As such, we have been unable to probe this
interaction.
Mutant Cycle Analyses Indicate Strong Receptor-Agonist Interactions
at β2L119. As noted above, a mutant cycle analysis (Fig. 4) is the
standard way to determine whether pairs of mutations are
independent or are coupled. EC50-based mutant cycle analyses
have been performed by our lab and others to investigate multiple
interactions in Cys-loop receptors and related structures (28, 34–
36). For several different agonist pairs, coupling coefficients (Ω)
and coupling energies (ΔΔG) were calculated (Table 2).
Mutant cycle analysis for the S-nicotine/S-MPP pair and the
β2L119/β2L119Lah pair predicts a substantial coupling energy
of 2.6 kcal∕mol. This is a relatively large energy for a putative
hydrogen bond, and it provides strong evidence for a hydrogen
bonding interaction between the pyridine N of nicotine and
the backbone NH of β2L119.
We also considered double mutant cycle analyses for the ago-
nists ACh and CCh using choline as the reference compound, as it
lacks the key hydrogen bond acceptor. This is a much less subtle
probe than the S-nicotine/S-MPP pair, but it still could produce
relevant results. Indeed, we find that for both the ACh∕Ch and
CCh∕Ch pairs, smaller, but still meaningful, coupling energies are
seen (Table 2).
Table 1. EC50 values, Hill coefficients, and relative efficacies*
Agonist
S-Nic
Mutation
EC50, nM
nH
WT
Leu
Lah
WT
Leu
Lah
WT
Leu
Lah
WT
Leu
Lah
WT
Leu
Lah
WT
Leu
Lah
120
120
5
3
1.3 0.05
1.5 0.05
1.3 0.04
1.7 0.08
1.5 0.11
1.5 0.05
1.3 0.07
1.3 0.08
1.2 0.04
1.3 0.02
1.2 0.03
1.2 0.04
1.6 0.06
1.2 0.09
1.4 0.05
1.4 0.07
1.5 0.15
1.3 0.03
800 30
11,000 400
14,000 900
1,100 40
360 20
440 20
3,000 100
7,200 80
7,900 200
29,000 800
140,000 4,000
140,000 20,000
150,000 5,000
0.79 0.04
0.58 0.05
2.9 0.06
S-MPP
ACh
CCh
Ch
Discussion
The nicotinic receptor has produced one of the longest-known,
best-studied pharmacophores. The original study of Beers and
Reich (10) proposed that two points, a cationic nitrogen and a
hydrogen bond acceptor, were required for successful interaction
with biological receptors. Later discussion debated the optimal
distance between the two points (deemed the internitrogen dis-
tance), and more recent models have alluded to pharmacophore
binding partners within the biological receptors. Despite 40 years
of interest in the nicotinic pharmacophore, the binding partners
of the essential two point pharmacophore have only recently been
identified. Pioneering mutagenesis and affinity labeling studies of
Epi
*All studies showed current values at þ70 mV, normalized to −110 mV,
≤0.08, confirming the A2B3 stoichiometry. Errors are standard error of
the mean. Epi is epibatidine. Mutations identified as “Leu” represent
recovery of the wild-type receptor by nonsense suppression.
13208
∣
www.pnas.org/cgi/doi/10.1073/pnas.1007140107
Blum et al.

gen bond partner(s) to agonists such as ACh and nicotine before
the structure of AChBP with nicotine bound. Nevertheless,
AChBP is not a nAChR. AChBP evolved to bind ligands, not
to gate an ion channel in response to ACh binding. As such, tests
of predictions from AChBP structures in real receptors are always
essential.
Here we employ a unique strategy to test the water-mediated
hydrogen bonding model of Fig. 2 in the neuronal, α4β2 nAChR.
The α4β2 receptor shows high affinity for nicotine, and it is
generally accepted to be the dominant receptor subtype that con-
tributes to nicotine addiction. Our studies of α4β2 are made pos-
sible by recent advances (14) that allow us to express significant
quantities of α4β2 in Xenopus oocytes, to control subunit stoichio-
metry, and to efficiently incorporate unnatural amino acids into
the receptor. Recently, we have shown that the cation-π interac-
tion and the hydrogen bond to TrpB are strong in the ðα4Þ2ðβ2Þ3
receptor (14).
N+
N+
CH3
CH3
H
H
N
H
O
H
O
O
O
H
H
Asn 107
Asn 107
H
H
HN
HN
N
N
Leu 119
Leu 119
O
O
120
14000
Ω = (120*1100)/(14000*800) = 0.012
G = –RTln(Ω) = 2.6 kcal/mol
To probe the second hydrogen bond suggested by AChBP, we
mutated β2L119 to its α-hydroxy analog. This removes the critical
NH, and, indeed, the agonists nicotine, ACh, CCh, and epibati-
dine, all show 5- to 7-fold increases in EC50 in response to the
mutation. While consistent with the hydrogen bonding model,
these observations certainly do not prove it. It could be that
the backbone mutation is simply generically disruptive to recep-
tor function.
N+
N+
CH3
CH3
H
H
N
H
O
H
O
O
O
H
H
Asn 107
Asn 107
HN
HN
O
O
Lah 119
Lah 119
O
O
To make an explicit connection between the pyridine N of ni-
cotine and the backbone NH of β2L119, we combined backbone
mutagenesis with a modification of the agonist, removing the
pyridine N to create S-MPP. Of course, S-MPP would never be
the target of a medicinal chemistry study; it can be anticipated
to be a terrible drug at the nAChR. Here it is used as a chemical
probe, to evaluate a key binding interaction of the potent drug
nicotine.
Studies with S-MPP produced remarkable results. As expected,
it is a very poor agonist at the wild-type receptor. However, com-
pletely opposite to what is seen with nicotine, ACh, CCh, or
epibatidine, introduction of the backbone ester at β2L119 lowers
EC50 for S-MPP. In fact, S-MPP and nicotine are comparably
potent at the mutant receptor. Clearly the backbone mutation
has had dramatically different effects on the two agonists. The
effect can be quantified by a mutant cycle analysis, which reveals
a coupling energy of 2.6 kcal∕mol between the backbone muta-
tion and the agonist “mutation.” This is a quite substantial energy,
especially when one considers that these chemical changes—both
in the protein and in the ligand—are more structurally subtle than
those typically employed in mutant cycle analysis studies using
conventional mutagenesis.
The results with S-MPP provide strong support for the nicotine
binding model based on the AChBP structure. As noted above,
however, AChBP structures with CCh or epibatidine bound do
not include the key water molecule, although other components
of the hydrogen bonding network are comparably positioned. We
find that ACh, CCh, and epibatidine all respond to the backbone
ester mutation in a way that is comparable to that seen for nico-
tine. In addition, choline, a weak agonist that lacks the hydrogen
bond acceptor of ACh and CCh, is not influenced by the back-
bone mutation. We thus conclude that all the drugs studied here
make a hydrogen bonding interaction with the backbone NH of
β2L119; the nicotinic pharmacophore has thus been completed
by interactions with the complementary subunit. Note that these
studies do not establish that the interaction between the hydrogen
bond acceptor component of the agonists and the backbone NH
of β2L119 is mediated by a water molecule; a direct interaction
would be just as compatible with our data. At present, we feel the
water-mediated interaction is the most reasonable interpretation,
but further experiments to address this point would be valuable.
We have now used chemical-scale investigations of functional
receptors to establish a three-point interaction between nicotine
800
1100
EC50 Values (nM)
Fig. 4. Double mutant cycle analysis for S-Nic and S-MPP on wild-type α4β2
and α4β2L119Lah.
the receptor from Torpedo rays identified a number of aromatic
amino acids near the binding site (1, 3). Early unnatural amino
acid mutagenesis studies showed that one of these aromatics,
now termed TrpB, makes a cation-π interaction with ACh in
the muscle-type nAChR (15), and more recent studies established
a comparable interaction to both ACh and nicotine in the α4β2
receptor (14).
The search for the presumed hydrogen bond donor to the acet-
yl group of ACh and the pyridine N of nicotine was much more
challenging. A breakthrough came with the discovery of the
AChBPs, and in 2004 a structure of nicotine bound to AChBP
was reported (18). As shown in Fig. 2, that AChBP structure
confirmed the cation-π interaction to TrpB. It also implicated
a hydrogen bond between the pyrrolidine NþH and the backbone
carbonyl of TrpB, an interaction that was subsequently confirmed
by unnatural amino acid mutagenesis (14, 23).
Importantly, the AChBP structure also suggested the binding
partner for the second element of the pharmacophore. In AChBP,
the pyridine N of nicotine makes a water-mediated hydrogen
bond to a backbone NH and to a backbone carbonyl (Fig. 2). This
elegant arrangement emphasizes the interfacial nature of the
agonist binding site, as the pyridine N interacts with residues that
are on the complementary subunit, while TrpB, which makes the
cation-π interaction and the hydrogen bond to the pyrrolidine
NþH, lies in the principal subunit. The value of AChBP in guiding
nAChR research is undeniably large, especially in the present
context. It would have been very challenging to guess the hydro-
Table 2. Coupling parameters (Ω) and ΔΔG values for mutant cycle
analyses
Agonist
Ω
ΔΔG, kcal mol−1
S-Nic∕S-MPP
ACh∕Ch
0.012
0.16
0.29
2.6
1.1
0.73
CCh∕Ch
Blum et al.
PNAS
∣
July 27, 2010
∣
vol. 107
∣
no. 30
∣
13209

the ligand-binding trends of the receptor (29). Stage V–VI Xenopus laevis
oocytes were injected with mRNA in a 1∶1 or 1∶20 ratio of α4L90 A∶β2 for
wild-type experiments or suppression with α-hydroxy acids, respectively.
Hydroxy or amino acids were appended to the dinucleotide dCA and enzy-
matically ligated to the truncated 74-nucleotide TQOpS’ tRNA as previously
described (30). Each cell was injected with 75 nL of a 1∶1 mixture of mRNA
(20–25 ng of total mRNA): tRNA (20–30 ng), with oocytes injected with Leu
ligated to TQOpS’ receiving an additional 75 nL after 24 h of incubation at
18 °C. Wild-type recovery experiments (injection of tRNA appended to the
natural amino acid) were performed to evaluate the fidelity of the unnatural
suppression experiments. Additional controls, mRNA only and 74-mer
TQOpS’ ligated to dCA (TQOpS’-dCA), were also examined. While small
currents (typically less than 200 nA) were seen for TQOpS’-dCA control experi-
ments, EC50 and Hill values were substantially different from suppression
values.
and the α4β2 neuronal nAChR, the receptor most strongly asso-
ciated with nicotine addiction. A cation-π interaction to TrpB
has been established by progressive fluorination of the key tryp-
tophan. Backbone mutagenesis has been used to establish two
key hydrogen bonds: the pyrrolidine NþH hydrogen bonds to
the backbone carbonyl of TrpB and the pyridine N of nicotine
hydrogen bonds to the backbone NH of β2L119. Studies of these
two hydrogen bonds were inspired by the AChBP structures, em-
phasizing the substantial impact of AChBP on nAChR research.
At the same time, AChBP is not a neurotransmitter-gated ion
channel; it evolved to serve a different function than a nAChR.
As such, we should anticipate some differences between the two
structures. Indeed, two features of the nicotine-AChBP structure
have been shown to be not functionally significant in studies of
nAChRs. The AChBP structure clearly shows a cation-π interac-
tion between the CH3 of nicotine and a tyrosine at the agonist
binding site termed TyrC2 (Fig. S1) (18). This methyl group—
which carries a charge comparable to a CH3 attached to the Nþ
of ACh—points directly at the center of the aromatic ring of
TyrC2 and essentially makes van der Waals contact with the ring,
unquestionably a cation-π interaction. However, we find no ex-
perimental support for this cation-π interaction in either the mus-
cle-type or the α4β2 nAChR. In each system, inserting 4-CN-Phe
at TyrC2 gives essentially wild-type receptor function (14, 15). A
CN group is very strongly deactivating in a cation-π interaction,
and so this result is in conflict with the AChBP structure. Note
that in a different Cys-loop receptor, the residue at position C2
does make a functionally significant cation-π interaction to the
natural agonist serotonin (37).
In addition, all AChBP structures—the nicotine, CCh, and epi-
batidine bound structures considered here as well as the “apo”
structure—contain a strong hydrogen bond between the indole
NH of TrpB and the backbone carbonyl of the residue that cor-
responds to β2L119 (Fig. S1). N • • • O distances range from 2.7
to 3.0 Å. However, earlier studies of the muscle-type receptor
found no evidence for an important interaction of this kind. In
particular, TrpB can be substituted by unnatural amino acids
in which the indole ring is replaced by a naphthalene or an N-
methylindole with very little impact on EC50 (15). All of these
analogs lack the critical hydrogen bond-donating NH of the
Trp indole ring.
In summary, we have used a combination of unnatural amino
acid mutagenesis and chemical synthesis to provide strong evi-
dence for a functionally important hydrogen bond between the
pyridine N of nicotine and the backbone NH of β2L119 in the
nicotine-sensitive α4β2 receptor. A similar interaction contri-
butes to the binding of ACh, CCh, and epibatidine. We have
now used unnatural amino acid mutagenesis to establish three
strong contact points between this critical receptor and nicotine:
the cation-π interaction to the side chain of TrpB, the hydrogen
bond between the pyrrolidine NþH and the backbone carbonyl of
TrpB, and the hydrogen bond between the pyridine N and the
backbone NH of β2L119. There is much interest in the pharma-
ceutical industry in developing subtype-selective agonists of
neuronal nAChRs, and it seems likely that the complementary
subunit will play the dominant role in discriminating among
subtypes. As such, these studies of a key binding interaction invol-
ving the complementary binding site suggest a general strategy
for developing insights that could lead to subtype-specific pharma-
ceuticals.
Electrophysiology Protocols. Electrophysiology experiments were performed
24–48 h after injection using the OpusXpress 6000A instrument (Axon Instru-
ments) in two-electrode voltage clamp mode at a holding potential of
−60 mV. The running buffer was Ca2þ-free ND96 solution (96 mM NaCl,
2 mM KCl, 1 mM MgCl2, and 5 mM Hepes, pH 7.5). During typical recordings,
agonists were applied for 15 s followed by a 116-s wash with the running
buffer. For recordings with epibatidine, the first eight drug concentrations
were applied for 90 s with a 116-s wash with running buffer, while the re-
maining concentrations were applied for 15 s with a 116-s wash. Dose-
response data were obtained for ≥8 agonist concentrations on ≥6 cells.
All EC50 and Hill coefficient values were obtained by fitting dose-response
relations to the Hill equation and are reported as averages ꢀ standard error
of the fit. A detailed error analysis of nonsense suppression experiments
reveals data are reproducible to ꢀ50% in EC50 (38, 39). Voltage jump experi-
ments were conducted to verify stoichiometry as described previously (14).
Double mutant cycle analyses were performed with EC50 values to calcu-
Leu;ligand
late coupling coefficients (Ω) using the equation Ω ¼ ðEC50
•
Lah;ligand analog
Lah;ligand
EC50
Þ∕ðEC50Leu;ligand analog • EC50
Þ, where Leu, ligand and
Leu, ligand analog represent the EC50 of the wild-type receptor with either
ligand and Lah, ligand; and Lah, ligand analog represent the EC50 of the
ester mutation with either ligand. Coupling energies ΔΔGint were calculated
from the equation ΔΔGint ¼ −RTlnΩ.
Synthesis of N-Methyl-2-Phenylpyrrolidine Hydrochloride. Racemic 2-phenyl-
pyrrolidine (5.0 g, 34 mmol), prepared according to a published protocol
(33), was mixed with dibenzoyl-L-taratric acid (6.1 g, 17 mmol) in a 100-mL
round-bottom flask equipped with a reflux condenser. To this was added
35% ethanol in ethylacetate (30 mL). The solution was heated to boiling
for 10 min and then cooled to room temperature overnight. The white
crystals were collected, rinsed with cold ethylacetate, and then submitted
to five sequential recrystallizations. The yield was (10%, 2.2 g). Spectral data
are 1H NMR (CDCl3, 300 MHz) δ 8.20 (4H, m), 7.61–7.32 (16H, m), 5.92 (2H, s),
5.03 (4H, b), 4.54 (2H, dd, J ¼ 9.1, 6.7 Hz), 3.38 (4H, m), 2.27–2.00 (8H, m); 13
C
NMR (CDCl3, 75 MHz) δ 172.58, 166.45, 134.91, 132.73, 130.54, 129.70, 128.83,
128.76, 128.03, 127.37, 75.60, 62.74, 44.80, 30.42, 23.37. High resolution mass
spectrometry (HRMS) (FABþ) m∕z calculated for C10H14N [Mþ]: 148.1126,
found 148.1081. To obtain enantioenriched 2-phenylpyrrolidine, the product
was vigorously stirred in a 1∶1 mixture of 2 M NaOH: CH2Cl2. The organic
layer was then extracted with additional CH2Cl2 (3×), washed with brine,
dried over Na2SO4, and concentrated to yield enantioenriched 2-phenylpyr-
rolidine as a yellow oil (yield: 95%). NMR spectra are consistent with pre-
viously reported data. HRMS (FABþ) m∕z calculated for C10H14N [M þ H]:
148.1126, found 148.1134. To establish enantiomeric excess, the product
was converted to ethyl 2-phenylpyrrolidine-1-carboxylate via a previously
described procedure (40), and this material was evaluated by analytical chiral
HPLC analysis using a Chiralcel OD-H column (4.6 mm × 25 cm) from Daicel
Chemical Industries, Ltd., with 2% isopropyl alcohol in hexanes, giving an en-
antiomeric excess of 96%. 1H NMR of ethyl 2-phenylpyrrolidine-1-carboxylate
gave (CH3OD, 300 MHz) δ; 7.32–7.15 (5H, m), 4.92 (1H, m), 4.08 (IH, m), 3.92
(1H, m), 3.59 (2H, q, J ¼ 7.7 Hz), 2.34 (1H, m), 1.95–1.86 (4H, m), 1.26 (1H, t,
J ¼ 7.0 Hz), 0.94 (1H, t, J ¼ 7 Hz); 13C NMR of ethyl 2-phenylpyrrolidine-1-
Materials and Methods
carboxylate (CDCl3, 75 MHz)
δ 155.40, 144.32, 128.22, 126.59, 125.44,
60.85, 47.34, 47.03, 35.71, 23.58, 14.79. HRMS of ethyl 2-phenylpyrrolidine-
1-carboxylate (FABþ) m∕z calculated for C13H18O2N [M þ H]: 220.1338, found
220.1336.
Enantioenriched 2-phenylpyrrolidine from above, (0.13 g, 0.86 mmol) was
added to a two-neck, 25-mL round-bottom flask equipped with a reflux
condenser. To this was added 4 mL of formic acid and 2 mL of 37 wt%
formaldehyde (in H2O). The mixture was stirred and heated to reflux at
Molecular Biology Protocols. Rat α4 and β2 cDNA in the pAMV vector was
linearized with the restriction enzyme Not 1. mRNA was prepared by in vitro
transcription using the mMessage Machine T7 kit (Ambion). Unnatural
mutations were introduced by the standard Stratagene QuickChange proto-
col, using a TGA mutation at the site of interest. The α4 subunit contained
a known mutation in the M2 transmembrane helix (L9’A) that improves
receptor expression and lowers whole-cell EC50 values, but does not influence
13210
∣
www.pnas.org/cgi/doi/10.1073/pnas.1007140107
Blum et al.

80 °C for 3 h. The solution was cooled to room temperature and made basic
(pH 12) by the addition of 2 M NaOH. The organics were extracted with
CH2Cl2, washed with brine, dried over Na2SO4, and concentrated. The result-
ing yellow oil was placed into a 25-mL round-bottom flask and dissolved in
5 mL of cold ether. HCl (g) was generated and passed into the solution by slow
addition of HCl (aq, 12 M) into H2SO4 (aq). The resulting white crystals were
J ¼ 4.7 Hz), 2.29 (4H, m); 13C NMR (CDCl3, 75 MHz)
δ 131.99, 129.89,
129.31, 128.76, 73.05, 44.43, 37.67, 31.94, 20.95; HRMS (FABþ) m∕z calculated
for C11H16N [Mþ]: 162.1283, found 162.1325.
ACKNOWLEDGMENTS. We thank Ariele P. Hanek and Sean M. A. Kedrowski for
helpful discussions. This work was supported by the National Institutes of
Health (NS 34407; NS 11756) and the California Tobacco-Related Disease
Research Program of the University of California, Grant 16RT-0160.
collected by filtration and dried. The yield was 83%, 140 mg, and the spectral
24
data were ½αꢁ
¼ −110° (c ¼ 1, CHCl3); 1H NMR (CDCl3, 300 MHz) δ 7.66
D
(2H, m), 7.32 (3H, m), 4.14 (1H, m), 3.95 (1H, b), 3.05 (2H, m), 2.60 (3H, d,
1. Corringer PJ, Le Novere N, Changeux JP (2000) Nicotinic receptors at the amino acid
level. Annu Rev Pharmacol Toxicol 40:431–458.
2. Grutter T, Changeux JP (2001) Nicotinic receptors in wonderland. Trends Biochem Sci
26:459–463.
3. Karlin A (2002) Emerging structure of the nicotinic acetylcholine receptors. Nat Rev
Neurosci 3:102–114.
4. Gotti C, Zoli M, Clementi F (2006) Brain nicotinic acetylcholine receptors: Native
subtypes and their relevance. Trends Pharmacol Sci 27:482–491.
5. Romanelli MN, et al. (2007) Central nicotinic receptors: Structure, function, ligands,
and therapeutic potential. ChemMedChem 2:746–767.
6. Coe JW, et al. (2005) Varenicline: An alpha 4 beta 2 nicotinic receptor partial agonist
for smoking cessation. J Med Chem 48:3474–3477.
7. Mansvelder HD, Keath JR, McGehee DS (2002) Synaptic mechanisms underlie nicotine-
induced excitability of brain reward areas. Neuron 33:905.
8. Nashmi R, et al. (2007) Chronic nicotine cell specifically upregulates functional alpha
4* nicotinic receptors: Basis for both tolerance in midbrain and enhanced long-term
potentiation in perforant path. J Neurosci 27:8202–8218.
9. Tapper AR, et al. (2004) Nicotine activation of alpha4* receptors: Sufficient for reward,
tolerance, and sensitization. Science 306:1029–1032.
10. Beers WH, Reich E (1970) Structure and activity of acetylcholine. Nature 228:917–922.
11. Glennon RA, Dukat M (2000) Central nicotinic receptor ligands and pharmacophores.
Pharm Acta Helv 74:103–114.
12. Glennon RA, Dukat M, Liao L (2004) Musings on alpha4beta2 nicotinic acetylcholine
(nACh) receptor pharmacophore models. Curr Top Med Chem 4:631–644.
13. Dougherty DA, Stauffer DA (1990) Acetylcholine binding by a synthetic receptor.
Implications for biological recognition. Science 250:1558–1560.
14. Xiu X, Puskar NL, Shanata JAP, Lester HA, Dougherty DA (2009) Nicotine binding to
brain receptors requires a strong cation-π interaction. Nature 458:534–537.
15. Zhong W, et al. (1998) From ab initio quantum mechanics to molecular neurobiology:
23. Cashin AL, Petersson EJ, Lester HA, Dougherty DA (2005) Using physical chemistry to
differentiate nicotinic from cholinergic agonists at the nicotinic acetylcholine receptor.
J Am Chem Soc 127:350–356.
24. Koh JT, Cornish VW, Schultz PG (1997) An experimental approach to evaluating the
role of backbone interactions in proteins using unnatural amino acid mutagenesis.
Biochemistry 36:11314–11322.
25. England PM, Zhang Y, Dougherty DA, Lester HA (1999) Backbone mutations in trans-
membrane domains of a ligand-gated ion channel: Implications for the mechanism
of gating. Cell 96:89–98.
26. Deechongkit S, et al. (2004) Context-dependent contributions of backbone hydrogen
bonding to beta-sheet folding energetics. Nature 430:101–105.
27. Deechongkit S, Dawson PE, Kelly JW (2004) Toward assessing the position-dependent
contributions of backbone hydrogen bonding to beta-sheet folding thermodynamics
employing amide-to-ester perturbations. J Am Chem Soc 126:16762–16771.
28. Gleitsman KR, Kedrowski SMA, Lester HA, Dougherty DA (2008) An intersubunit
hydrogen bond in the nicotinic acetylcholine receptor that contributes to channel
gating. J Biol Chem 283:35638–35643.
29. Fonck C, et al. ( 7, 2005) Novel seizure phenotype and sleep disruptions in knock-in
mice with hypersensitive alpha 4* nicotinic receptors. J Neurosci 25:11396–11411.
30. Nowak MW, et al. (1998) In vivo incorporation of unnatural amino acids into ion
channels in a Xenopus oocyte expression system. Methods Enzymol 293:504–529.
31. Rodriguez EA, Lester HA, Dougherty DA (2007) Improved amber and opal suppressor
tRNAs for incorporation of unnatural amino acids in vivo. Part 1: Minimizing misacyla-
tion. RNA 13:1703–1714.
32. Rodriguez EA, Lester HA, Dougherty DA (2007) Improved amber and opal suppressor
tRNAs for incorporation of unnatural amino acids in vivo. Part 2: Evaluating suppres-
sion efficiency. RNA 13:1715–1722.
33. Dunsmore CJ, Carr R, Fleming T, Turner NJ (2006)
A chemo-enzymatic route to
A
cation-π binding site in the nicotinic receptor. Proc Natl Acad Sci USA
enantiomerically pure cyclic tertiary amines. J Am Chem Soc 128:2224–2225.
34. Kash TL, Jenkins A, Kelley JC, Trudell JR, Harrison NL (2003) Coupling of agonist
binding to channel gating in the GABA(A) receptor. Nature 421:272–275.
35. Price KL, Millen KS, Lummis SC (2007) Transducing agonist binding to channel
gating involves different interactions in 5-HT3 and GABAC receptors. J Biol Chem
282:25623–25630.
36. Venkatachalan SP, Czajkowski C (2008) A conserved salt bridge critical for GABAA
receptor function and loop C dynamics. Proc Natl Acad Sci USA 105:13604–13609.
37. Mu TW, Lester HA, Dougherty DA (2003) Different binding orientations for the same
agonist at homologous receptors: A lock and key or a simple wedge? J Am Chem Soc
125:6850–6851.
38. Torrice MM (2009) Chemical-Scale Studies of the Nicotinic and Muscarinic Acetylcho-
line Receptors (California Institute of Technology, Pasadena, CA).
39. Torrice MM, Bower KS, Lester HA, Dougherty DA (2009) Probing the role of the
cation-pi interaction in the binding sites of GPCRs using unnatural amino acids. Proc
Natl Acad Sci USA 106:11919–11924.
95:12088–12093.
16. Dougherty DA (2008) Cys-loop neuroreceptors: Structure to the rescue? Chem Rev
108:1642–1653.
17. Brejc K, et al. (2001) Crystal structure of an ACh-binding protein reveals the ligand-
binding domain of nicotinic receptors. Nature 411:269–276.
18. Celie PH, et al. (2004) Nicotine and carbamylcholine binding to nicotinic acetylcholine
receptors as studied in AChBP crystal structures. Neuron 41:907–914.
19. Rucktooa P, Smit AB, Sixma TK (2009) Insight in nAChR subtype selectivity from AChBP
crystal structures. Biochem Pharmacol 78:777–787.
20. Hansen SB, et al. (2006) Structural characterization of agonist and antagonist-
bound acetylcholine-binding protein from Aplysia californica.
J Mol Neurosci
30:101–102.
21. Hansen SB, et al. (2005) Structures of Aplysia AChBP complexes with nicotinic agonists
and antagonists reveal distinctive binding interfaces and conformations. EMBO J
24:3635–3646.
22. Taylor P, et al. (2007) Structure-guided drug design: Conferring selectivity among
neuronal nicotinic receptor and acetylcholine-binding protein subtypes. Biochem
Pharmacol 74:1164–1171.
40. Felpin F-X, et al. (2001) Efficient enantiomeric synthesis of pyrrolidine and piperidine
alkaloids from tobacco. J Org Chem 66:6305–6312.
Blum et al.
PNAS
∣
July 27, 2010
∣
vol. 107
∣
no. 30
∣
13211