Surface plasmon resonance biosensor based fragment screening using acetylcholine binding protein identifies ligand efficiency hot spots (le hot spots) by deconstruction of nicotinic acetylcholine receptor α7 ligands

Article abstract of DOI:10.1021/jm100834y

The soluble acetylcholine binding protein (AChBP) is a homologue of the ligand-binding domain of the nicotinic acetylcholine receptors (nAChR). To guide future fragment-screening using surface plasmon resonance (SPR) biosensor technology as a label-free, direct binding, biophysical screening assay, a focused fragment library was generated based on deconstruction of a set of α7 nAChR selective quinuclidine containing ligands with nanomolar affinities. The interaction characteristics of the fragments and the parent compounds with AChBP were evaluated using an SPR biosensor assay. The data obtained from this direct binding assay correlated well with data from the reference radioligand displacement assay. Ligand efficiencies for different (structural) groups of fragments in the library were correlated to binding with distinct regions of the binding pocket, thereby identifying ligand efficiency hot spots (LE hot spots). These hot spots can be used to identity the most promising hit fragments in a large scale fragment library screen.

Full text of DOI:10.1021/jm100834y

7192 J. Med. Chem. 2010, 53, 7192–7201
DOI: 10.1021/jm100834y
Surface Plasmon Resonance Biosensor Based Fragment Screening Using Acetylcholine Binding Protein
Identifies Ligand Efficiency Hot Spots (LE Hot Spots) by Deconstruction of Nicotinic Acetylcholine
Receptor r7 Ligands
§
§
^
^
^
Gerdien E. de Kloe,†,# Kim Retra,‡,# Matthis Geitmann, Per Kallblad, Tariq Nahar, Rene van Elk, August B. Smit,
€
ꢀ
Jacqueline E. van Muijlwijk-Koezen,† Rob Leurs,† Hubertus Irth,‡ U. Helena Danielson, ,§ and Iwan J. P. de Esch*,†
†Leiden/Amsterdam Center for Drug Research (LACDR), Division of Medicinal Chemistry, Faculty of Sciences, VU University Amsterdam,
De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands, ‡Leiden/Amsterdam Center for Drug Research (LACDR), Division of BioMolecular
Analysis, Faculty of Sciences, VU University Amsterdam, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands, §Beactica AB, Box 567,
SE-751 22 Uppsala, Sweden, Department of Biochemistry and Organic Chemistry, Uppsala University, BMC, Box 576, SE-751 23, Uppsala,
Sweden, and ^Department of Molecular and Cellular Neurobiology, Center for Neurogenomics and Cognitive Research, Neuroscience Campus
Amsterdam, VU University Amsterdam, De Boelelaan 1085, 1081 HV, Amsterdam, The Netherlands. # These authors contributed equally.
Received July 5, 2010
The soluble acetylcholine binding protein (AChBP) is a homologue of the ligand-binding domain of the
nicotinic acetylcholine receptors (nAChR). To guide future fragment-screening using surface plasmon
resonance (SPR) biosensor technology as a label-free, direct binding, biophysical screening assay, a
focused fragment library was generated based on deconstruction of a set of R7 nAChR selective
quinuclidine containing ligands with nanomolar affinities. The interaction characteristics of the
fragments and the parent compounds with AChBP were evaluated using an SPR biosensor assay. The
data obtained from this direct binding assay correlated well with data from the reference radioligand
displacement assay. Ligand efficiencies for different (structural) groups of fragments in the library were
correlated to binding with distinct regions of the binding pocket, thereby identifying ligand efficiency
hot spots (LE hot spots). These hot spots can be used to identity the most promising hit fragments in a
large scale fragment library screen.
Introduction
expressed R4β2 and R7 subtypes.7 They are important drug
targets for Alzheimer’s disease, Parkinson’s disease, and
schizophrenia.8,9 The development of subtype selective
ligands is one of the major challenges in the area of nAChRs.
However, many of the biophysical screening technologies that
have been developed and used for interaction studies with
water-soluble proteins are not readily available for the trans-
membrane LGICs. In recent years the soluble acetylcholine
binding protein (AChBP) has been adopted as a structural and
pharmacological homolog of the extracellular domain (ECD)
of nAChRs; the ECD constitutes the ligand-binding domain.10,11
AChBP displays comparable ligand pharmacology to the R7
nAChR in particular.11,12 Next to nAChRs,13 AChBP has
been used as a molecular tool to study structural and ligand-
binding properties of other LGICs, i.e., 5-HT3,14 GABA,15
and glycine16 receptors. AChBP is also fused to the transmem-
brane ion pore of 5-HT3 and glycine receptors to form
chimeric channels that are useful for studying binding sites,
activation mechanisms, and the different states of LGICs.17,18
Several AChBPs have been identified from various snail
species, e.g., Aplysia californica (Ac) and Lymnaea stagnalis
(Ls), that differ slightly in pharmacology.19
Fragment-based drug discovery (FBDDa) has become a
central part in modern medicinal chemistry, and an increasing
number of fragment screening and optimization studies are
being published.1 As the initial focus of FBDD is on low
molecular weight fragments, there is a great need for technol-
ogies that can measure the weak binding of a fragment to a
protein target. Biophysical techniques such as NMR spectroscopy,
X-ray crystallography, mass spectrometry, and surface plasmon
resonance (SPR biosensors) are particularly suited.2-6 These
assays do not require labeling of the ligand or protein, as is the
case for displacement of radiolabeled or fluorescent ligands.
In the search for new compounds interacting with ligand-
gated ion channels (LGIC), we are implementing fragment-
based approaches. The LGICs of our interest are the nicotinic
acetylcholine receptors (nAChRs) of the subgroup of Cys-loop
receptors, also including the GABAA receptors, 5-HT3-
serotonin receptors, and glycine receptors. Many different
subtype nAChRs are present in the CNS, including the highly
*To whom correspondence should be addressed. Phone: þ31-(0)20-
5987841. Fax: þ31-(0)20-5987610. E-mail: i.de.esch@few.vu.nl.
a Abbreviations: FBDD, fragment-based drug discovery; AChBP,
acetylcholine binding protein; Ls, Lymnaea stagnalis; Ac, Aplysia
californica; nAChR, nicotinic acetylcholine receptor; LGIC, ligand-
gated ion channel; CNS, central nervous system; ECD, extracellular
domain; SPR, surface plasmon resonance; RBA, radioligand binding
assay; RU, response unit; LE, ligand efficiency; Tc-IFP, interaction
fingerprint Tanimoto scores.
To develop nAChR directed compounds, we consider SPR
biosensor analysis as suitable for our FBDD studies. It allows
for sensitive and label-free detection of binding fragments.3,20
By immobilization of AChBP on the chip surface, protein
consumption is very low. An additional advantage of direct
binding biosensor-based interaction assays is the potential
r
pubs.acs.org/jmc
Published on Web 09/09/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7193
Figure 1. Generic structures of fragmentized quinuclidine compounds. Group I (green): high affinity R7 and AChBP compounds
2,3-disubstituted quinuclidines. Group II (purple): 3-substituted quinuclidines. Group III (aqua): 2,3-disubstituted quinuclidines with a
truncated substituent in the 3 position. Group IV (orange): smaller fragments containing a basic amine. Group V (red): smaller fragments
without a basic amine.
to identify orthosteric and ECD-bound allosteric compounds.
The latter is of interest not only in the nAChRs field but also
ligand efficiencies (LE), enabling better evaluation in the
fragment hit selection phase, as well as subsequent efficient
fragment growing. In the process we want to validate our SPR
biosensor assay24 for fragment screening with AChBP. The
obtained data are also used as a preliminary evaluation for the
use of AChBP as molecular bait for nAChR ligands.
A set of R7 selective nAChR ligands was used to design
and synthesize groups of fragments by deconstruction. To
evaluate our SPR biosensor assay,24 the SPR biosensor
data for the fragment set were compared with the corre-
sponding data obtained with the well established radioli-
gand binding assay.35 The data were generated with two
different types of AChBP and compared with R7 nAChR
affinity data. The affinity data for the fragments were
used to calculate LEs. The LEs were projected on the
surface of the binding pocket of Ls-AChBP using predicted
binding modes, to identify potential “LE hot spots” in the
pocket.
21
for other LGICs such as GABAA and 5-HT3 receptors.22
Earlier we have developed two biosensor-based assays: an
indirect assay23 (with bungarotoxin as displacement ligand)
and a direct assay.24 This direct assay is used and evaluated in
the current paper.
Ideally, the technology used for fragment screening is able
not only to identify the most promising hit fragment but also
to guide efficient growing of these entities into bigger and
more potent compounds. Here the ability of SPR biosensors
to readily obtain kinetic and thermodynamic binding data can
be a major advantage over traditional radioligand displace-
ment assays.25 An important concept for both fragment
selection and fragment growing is ligand efficiency (LE).26,27
LE indicates the contribution of the atoms in a compound to
the affinity or activity and is defined as the binding energy
divided by the molecular weight or number of heavy atoms of
a ligand. Compounds with the highest LE are thought to have
the highest potential to be optimized into potent leads. During
the optimization process, LE is used to guide the efficient
growing, linking, or merging of fragments into druglike
molecules. In a hallmark paper by Hadjuk, a retrospective
study indicated that LE ideally is maintained throughout the
discovery process.28 However, other studies, as well as an
FBDD literature survey, indicate that LE is not always a
constant factor during the fragment growing process.1,29,30
Deconstruction of druglike molecules is an attractive tool to
probe the binding energy contributions of ligand substruc-
tures to distinct areas and subpockets.31-34
Results and Discussion
Ligand Deconstruction. A fragment set was designed from
high-affinity quinuclidine-containing R7 receptor ligands
(shown in green box in Figure 1, published by Mazurov
et al.36). The similarity of R7 and AChBP ligand pharmacol-
ogies directed our choice for R7 selective ligands, while large
ligands with high affinities were preferable to allow the
synthesis of several groups of fragments with measurable
affinities. These original ligands, with molecular weights of
∼400 Da and nanomolar affinities, were included as druglike
compounds representing the final products of a (fragment)
hit optimization program (group I, Figure 1). Four groups of
fragments were designed (groups II-V, Figure 1) covering
molecular weights from 100 to 400 Da. A color scheme for
We here present a deconstruction study of nAChR ligands.
The primary aim of this deconstruction study is to benchmark
the expectations of upcoming fragment hits with regard to

7194 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
de Kloe et al.
Scheme 1a
a Reagents and conditions: (i) 3-pyridinecarboxaldehyde, KOH, MeOH, reflux; (ii) H2, Pd/C, MeOH, 3.45 bar, room temp; iii) (i-PrO)3Al,
isopropanol, reflux; (iv) substituted isocyanates, triethylamine, DCM, room temp; (v) HCO2NH2, NaCNBH3, ZnCl2, MeOH, room temp;
(vi) carboxylic acid, diphenylphosphoryl chloride, triethylamine, DCM, room temp; (vii) phenylchloroformate, triethylamine, DCM, room temp;
(viii) dimethylaniline, 200 ꢀC; (ix) formaldehyde, formic acid, reflux.
these groups is used consistently throughout the manuscript.
Nicotine and an R7 selective nAChR ligand (26, PNU-
28298737) were used as reference compounds.
direct assay and radioligand displacement ([3H]epibatidine)
assay; affinities for R7-nAChR were determined in a radi-
oligand displacement ([3H]MLA) assay (Table 1). The affi-
nities obtained for Ls-AChBP with the two assays were
highly correlated (R2 = 0.97), as shown in Figure 3A. This
excellent correlation, for a diverse set of compounds with
respect to molecular weight and affinities, between the novel
SPR biosensor direct assay and the established radioligand
displacement assay validates the quality of the SPR bio-
sensor assay for the detection of binding of fragments.
The validity of AChBP as a model for R7-nAChR and the
similarities of the binding sites were also evaluated by
correlation analysis (Figure 3B,C). SPR biosensor affinity
data (for Ls-AChBP in Figure 3B, for Ac-AChBP in
Figure 3C) are correlated with R7-nAChR radioligand affi-
nity data. The correlations between R7 and AChBP are quite
low. Ls-AChBP data correlate better with the R7-nAChR
data than the Ac-AChBP data, particularly for group I
compounds. Therefore, our focus in the remainder of the
article is on Ls-AChBP. In fact, two groups of fragments,
group II (pink squares) and group III (blue triangles), show
selectivity for either AChBP or R7. The group II urea and
carbamates are better accommodated in the R7-nAChR,
while group III 2-pyridine substituted quinuclidines have
better interactions with Ls- and Ac-AChBP. This indicates
significant differences between the binding sides of AChBP
and the R7 receptor.
Chemistry. The synthesis routes are given in Scheme 1.
Compounds in groups I and III were synthesized from the
commercially available quinuclidin-3-one (1) according to a
literature procedure.36 An aldol condensation with 3-pyridi-
naldehyde, followed by a reduction of the formed double
bond and a reduction of the ketone to the corresponding
alcohol with aluminum isopropoxide or to the correspond-
ing amine with ammonium acetate and sodium cyanobor-
ohydride, yielded 3 and 4, respectively. For compounds in
group II, amine 17 and the corresponding alcohol 14 were
commercially available, as well as 3-(3-hydroxy)pyridine (20)
for compounds in group V. Coupling with isocyanates gave
the corresponding urea and carbamates. A peptide coupling
with carboxylic acids gave the corresponding amides. In-
verted carbamate 13 was synthesized by coupling 4 with
phenylchloroformate. Fragments in group V were synthe-
sized starting from 3-(3-pyridyl)-D-alanine (23): 24 was
obtained by decarboxylation at high temperature,38 and
Eschweiler-Clarke methylation resulted in 25.
Validation of SPR Biosensor Interaction Assay. Ls-AChBP
and Ac-AChBP were immobilized to levels around 3000 RU.
Higher amounts were avoided in order to reduce the effects
of limited mass transport of analytes. Examples of sensor-
grams are given for one representative compound per struc-
tural group (Figure 2). These examples show that high
quality sensorgrams could be obtained for all fragment
groups. Affinity values (KD) could be determined for all
compounds but not interaction rate constants due to very
fast kinetics.24 In some cases saturation could not be reached
within the concentration range that could be reliably used.
The affinities of all 20 compounds for Ls-AChBP, along
with nicotine and 26, were determined in the SPR biosensor
Ligand Efficiencies. The ligand efficiencies (LE) for all
compounds were calculated from the experimentally deter-
mined ΔG values, using the (adjusted) formula of Hopkins
et al.:27 LE = (ΔG/MW) ꢀ 1000 (Table 1). Dividing by the
molecular weight instead of the number of heavy atoms is
more accurate in describing the difference between substituents
with large molecular weight differences.26 To monitor LE
during the (retrospective) growth of low affinity fragments

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7195
Figure 2. Examples of sensorgrams for one representative fragment from each group: (A) 8, group I; (B) 17, group II; (C) 3, group III; (D) 25,
group IV; (E) 22, group V.
into druglike compounds, LE is plotted against pKi for
Ls-AChBP (Figure 4A) and R7 nAChR (Figure 4B).
the final ligand than others. To visualize this, we defined
LE “hot parts” of a molecule and translate these parts to “LE
hot spots” in the binding pocket. This is presented in Figure
5A-D, where docking modes of a representative compound
of four fragment groups are presented.
The focused fragment set gives insight into the contribu-
tion of the different parts of the quinuclidine compounds to
binding. It is apparent that LEs differ remarkably for the
different groups (Figure 4). The highest LEs are obtained for
group IV fragments containing a basic nitrogen atom as part
of a pyridine or amine group. Decoration of the quinuclidine
3-position leads to compounds with higher affinity but with
significantly reduced LEs. Actually, for the majority of the
compounds, the LE of the starting fragments (group IV,
orange squares) is about twice as high as the LE of the “final”
optimized compound (group I, green diamonds). Perhaps not
surprisingly, the lowest LEs are found for the compounds in
group V (red circles) that lack a basic nitrogen group. The basic
center forms a cation under physiological conditions, and it is
known that this charged group forms key cation-π inter-
actions when binding to AChBP and nAChRs.39
When the LEs for the R7 nAChR are monitored, the trend
is comparable with that for Ls-AChBP except for group II
(purple stars) compounds (Figure 4B). The affinities of a
majority of the compounds (groups I, III, IV, and V) are
about 10 times higher for R7, but for three of the group II
compounds the affinity is 100 times higher for R7. Urea 17
(group II) even has a LE of 39, which is higher than the group
III compounds (while the group III compounds have lower
LEs for R7 than for Ls). This illustrates that the binding
region that accommodates the substituent in the 3-position
of quinuclidine is more effectively addressed in the R7
nAChR than in AChBP. It is also noted that the dependency
on the basic amine is greater for AChBP than for the
R7 nAChR, as is suggested by the trend in the group V
fragments.
Docking modes for the compound set were generated with
Gold/Chemscore and ranked according to interaction fin-
gerprint Tanimoto scores (Tc-IFP).40 The binding mode of
nicotine in the crystal structure of Ls-AChBP (1UW6) was
used to generate the reference IFP. The use of IFP can easily
overcome the imperfection of scoring functions with respect
to induced fit,41 since emphasis is given to important inter-
actions known from available crystal structures. Compar-
ison with experimental data from a crystal structure makes
the resulting docking poses more reliable than using docking
and scoring functions to select the best pose. Previously,
we have used an induced-fit docking method to study flex-
ibility of the loop regions upon ligand binding and as
such addressed the dynamic process of ligand recognition.24
The present docking studies with rigid protein as described
here represent the end point of that process and the energy
minimum as validated by X-ray analyses in which identical
interactions are consistently present (i.e., pyridine nitrogen
atoms of the ligands interact with a water molecule, etc.).
The LEs of the ligands are projected on the surface. The
docking modes of 25, 3, and 8 (Figure 5A-C) indicate that a
better resemblance of the nicotine binding mode yields a
higher LE. Or in other words, the compound with the highest
proportion of binding to the “hot spot” of the pocket gives
the highest LE. Amine 25 (Figure 5A) makes all three
important interactions that nicotine also makes in the crystal
structure (Figure 5E), namely, (1) cation-π, mainly with
aromatic side chain of Trp143, (2) a hydrogen bond with the
backbone of Trp143, and (3) a hydrogen bond through a
bridging water molecule (shown) to the main chain of
residues Leu102 and Met114 (not shown). Quinuclidine 3
(Figure 5B) extends already in the direction of the cysteines,
and the LE drops slightly. Ligand 8 (Figure 5C) extends
Ligand Efficiency Hot Spots. Thus, from a retrospective
growing consideration, the high LEs of the small fragments
are not maintained in the optimized compounds. Following
a typical fragment growing path (an example is shown in
Figure 4), LE drops from 50 to 42 to 20 while affinity goes up
with 2 log units. Apparently, certain parts of the molecule
contribute much more to the binding energy and LE of
further, whereas its LE drops to only 22 kcal mol-1 Da-1
.
Ligand 22 (Figure 5D) lacks a positive nitrogen and makes
3

7196 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
de Kloe et al.
Table 1. Structures and Affinities of All Compounds Determined by a Direct Binding Biosensor-Based Approach and an Indirect Radioligand
Displacement Assay for Ls-AChBP, Ac-AChBP, and R7 nAChRa
a (1) [3H]MLA displacement. (2) Ligand efficiencies (LE) were calculated using the (adjusted) formula of Hopkins et al.:27 LE = (ΔG/MW) ꢀ 1000. (3)
[3H]Epibatidine displacement. (4) Highest measured concentration: 10-3 M. (5) LE calculated using SPR biosensor determined pKD.
only one interaction in the highest LE part of the pocket (the
pyridine moiety with the conserved water molecule). Thus,
its LE is even lower.
The identification of LE hot spots is relevant, as it allows
for a more considered nomination of most promising frag-
ment hits. For prioritization of the hits from a full fragment

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7197
Figure 3. Comparison of the direct SPR data with radioligand displacement and AChBP with R7 data: (A) RBA pKi versus SPR pKD
Ls-AChBP; (B) SPR pKD data Ls-AChBP versus pKi RBA R7; (C) SPR pKD Ac-AChBP versus pKi RBA R7.
Figure 4. Monitoring ligand efficiency (LE) during (retrospective) fragment growing. Colors represent different groups: (A) Ls-AChBP data;
(B) R7 nAChR data.
Figure 5. Docking modes in Ls-AChBP (1UW6) of (A) 25, (B) 3, (C) 8, and (D) 22 and (E) crystal structure of nicotine in complex with
Ls-AChBP. Ligand efficiencies were calculated from Ls-AChBP affinity data (RBA). The color coding of the surface was obtained by
˚
exchanging B-factors with LEs of ligands projected on protein atoms at 2.5 A distances.
library screen, we intend to use a combination of predicted
binding modes with weighting according to LE hot spots.
For example, if hit fragments lacking a basic nitrogen atom
have at least higher LEs than group V fragments (the

7198 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
de Kloe et al.
compounds lacking a cationic center as described in this
study), the potential to develop these fragments into high
affinity ligands by growing the core fragment toward the
cation-π interaction site is considered as high.
An HCl salt was made of the product by stirring it in a saturated
HCl solution in diethyl ether. A solid was formed and filtered.
The product was recrystallized from acetone/methanol. Yield:
1
1.5 g (5.9 mmol, 51%). H NMR (400 MHz, MeOD) δ (ppm)
8.59 (d, J = 1.81 Hz, 1H), 8.43 (dd, J = 4.89, 1.47 Hz, 1H),
7.98-7.86 (m, 1H), 7.42 (dd, J = 7.64, 4.70 Hz, 1H), 4.07-3.98
(m, 1H), 3.98-3.88 (m, 1H), 3.62-3.44 (m, 2H), 3.44-3.32 (m,
2H), 3.26-3.10 (m, 1H), 3.00 (dd, J = 13.77, 5.27 Hz, 1H),
2.37-2.25 (m, 1H), 2.25-2.17 (m, 1H), 2.07-1.94 (m, 1H),
1.94-1.69 (m, 2H). Absence of NOE signals between 2R and 3S
protons confirmed trans-configuration (see Supporting Infor-
mation for NOESY spectrum).
Conclusion
Deconstruction of known R7 nAChR ligands has led to the
identification of “LE hot spots”. These hot spots are useful for
evaluating the binding efficiency of fragments based on their
LE. Fragments bound to these hot spots should have a very
high LE, for growing will probably only decrease their LE.
Fragments that bind to other areas than the LE hot spot can
potentially be grown toward thesehotspot, thereby increasing
LE and affinity considerably.
We have used an SPR biosensor assay that is capable of
detecting the binding of low molecular weight and low affinity
fragments. SPR data are able to successfully guide fragment
growing toward druglike and highly potent compounds. To
our knowledge, the present SAR study based on SPR bio-
sensor datafor fragments is one of the first2,3,20 in an emerging
field and the data illustrate the use of SPR for fragment
screening and optimization programs.
2-(Pyridin-3-ylmethyl)quinuclidin-3-amine (4). 2-(Pyridin-3-
ylmethyl)quinuclidin-3-one (2, 18 g, 83 mmol) was dissolved in
900 mL of methanol. To this yellow solution, ammonium
acetate (52.5 g, 0.83 mol) was added. The solution was stirred
for 20 min at room temperature. Then NaCNBH3 (19.3 g,
0.31 mol) and ZnCl2 (1.13 g, 8.3 mmol) were added. The reaction
mixture was stirred overnight at room temperature. The next
day, another amount of NaCNBH3 (0.5 equiv, 2.6 g, 41 mmol)
was added. The reaction mixture was stirred for another 4 h,
then concentrated and dissolved again in 1 M HCl solution until
pH ∼6/7 was obtained. The mixture was washed with DCM (2 ꢀ
100 mL), then basified with 2.5 M NaOH to pH > 14. The water
layer was extracted with DCM (3 ꢀ 200 mL). The organic layer
was dried with Na2SO4, filtered, and concentrated. An HCl salt
was made of the product by dissolving it in DCM and adding
2 M HCl solution in diethyl ether. The resulting solid was
filtered and recrystallized from EtOH with 10% isopropanol.
Yield: 4.6 g (14.7 mmol, 18%, 2.6 equiv of HCl). 1H NMR (400
MHz, D2O) δ (ppm) 8.82 (s, 1H), 8.72 (d, J = 5.62 Hz, 1H), 8.56
(d, J = 8.24 Hz, 1H), 8.00 (dd, J = 8.05, 5.75 Hz, 1H), 4.01 (dt,
J = 10.74, 5.41 Hz, 1H), 3.72 (dd, J = 5.00, 2.54 Hz, 1H),
3.66-3.51 (m, 2H), 3.50-3.18 (m, 4H), 2.55-2.44 (m, 1H),
2.22-1.92 (m, 4H). Absence of NOE signals between 2R and 3S
protons confirmed trans-configuration (see Supporting Infor-
mation for NOESY spectrum).
Experimental Section
Synthetic Methods. General Remarks. Chemicals and re-
agents were obtained from commercial suppliers and were used
without further purification. The following compounds were
purchased from Sigma-Aldrich: nicotine, 26, and 17. The fol-
lowing compounds were synthesized according to the liter-
ature:36 3-12, 15, 16, 18, 19. The synthesis of 3 and 4 is described
in detail because our chemical characterization (most notably
through NMR analysis) identified the obtained material as the
trans-isomers of the quinuclidines, whereas Mazurov et al. have
reported the isolation of the cis-isomers in scientific literature.
Very recently, however, the same group published different
findings in the patent literature, where the isolated quinuclidine
amines were indeed identified as the trans-isomers, as validated
by X-ray analysis of the crystallized product.42 Dry DCM and
THF were obtained by distillation from CaCO3. Flash column
chromatography was typically carried out on a Biotage flash
chromatography system, using prepacked Biotage Si columns
with the UV detector operating at 254 nm. Analytical HPLC-
MS analyses were conducted using a Shimadzu LC-8A pre-
parative liquid chromatograph pump system with a Shimadzu
SPD-10AV UV-vis detector with the MS detection performed
with a Shimadzu LCMS-2010 liquid chromatograph mass
spectrometer. Conditions were as follows: an Xbridge (C18)
5 μm column (100 mm ꢀ 4.6 mm) with solvent A (90%
MeCN-10% buffer) and solvent B (90% water-10% buffer),
flow rate of 1.0 mL/min, start 5% A, linear gradient to 100% A
in 10 min, then 22.5 min at 90% A, then 12.5 min at 5% A, total
run time of 45 min. The buffer is a 0.4% (w/v) NH4CO3 solution
in water, adjusted to pH 8.0 with NH4OH. Compound purities
were calculated as the percentage peak area of the analyzed
compound by UV detection at 254 nm. HRMS data were
collected using a Bruker micrOTOF-Q (ESI). Purity and HRMS
data for all compounds are listed in Supporting Information
Table S1.
Phenyl-2-(pyridine-3-ylmethyl)quinuclidin-3-ylcarbamate (13).
The free base of 4 (0.5 g, 1.5 mmol) was dissolved in 5 mL of
dry DCM and triethylamine (0.8 mL, 6.1 mmol). Phenylchloro-
formate (226 μL, 1.8 mmol) was added. The reaction mixture
was stirred for 24 h and quenched with saturated NaHCO3
solution (20 mL). This water layer was extracted with DCM
(2 ꢀ 50 mL). The organic layer was dried with Na2SO4, filtered,
and concentrated. The product was purified over SiO2 (EtOAc
100%, TEA 2%, gradient to MeOH 20%), yielding 102 mg (0.30
1
mmol, 20%) of a white solid. H NMR (400 MHz, CDCl3) δ
(ppm) 8.54-8.36 (m, 2H), 7.54 (t, J = 9.69 Hz, 1H), 7.33 (t, J =
7.60 Hz, 2H), 7.18 (dd, J = 7.63, 4.81 Hz, 2H), 7.08 (dd, J =
14.35, 8.01 Hz, 2H), 5.86-5.67 (m, 1H), 3.49 (d, J =6.73 Hz, 1H),
3.12-2.57 (m, 6H), 1.96 (s, 1H), 1.65 (s, 3H), 1.43 (t, J = 11.62
Hz, 1H). 13C NMR (101 MHz, CDCl3) δ 154.18, 150.98, 150.19,
147.44, 136.63, 135.00, 129.25, 125.25, 123.28, 121.52, 65.60,
55.14, 49.62, 40.84, 28.19, 26.27.
3-(Pyridine-3-yl)propyl phenylcarbamate (21). 3-Pyridinepro-
panol (20, 1 mL, 7.7 mmol) was dissolved in 25 mL of dry THF.
Phenyl isocyanate (0.93 mL, 8.5 mmol) was added. The reaction
mixture was stirred overnight at room temperature and then
concentrated. Saturated NaHCO3 solution was added (50 mL),
and this water layer was extracted with DCM (2 ꢀ 50 mL). The
organic layer was dried with Na2SO4, filtered, and concentrated.
The product was purified over SiO2 (EtOAc 100%, TEA 2%,
gradient to MeOH 20%), yielding 1.58 g (6.2 mmol, 80%) of a
white solid. 1H NMR (400 MHz, CDCl3) δ (ppm) 8.45 (d, J =
5.68 Hz, 2H), 7.48 (d, J = 7.54 Hz, 1H), 7.39 (d, J = 7.55 Hz,
2H), 7.28 (t, J = 7.60 Hz, 2H), 7.23-7.14 (m, 1H), 7.04 (t, J =
7.22 Hz, 1H), 4.18 (t, J = 6.27 Hz, 2H), 2.69 (t, J = 7.69 Hz, 2H),
2.10-1.90 (m, 2H). 13C NMR (101 MHz, CDCl3) δ 153.67,
149.79, 147.50, 138.09, 136.60, 135.90, 129.03, 123.44, 123.39,
118.75, 64.12, 30.26, 29.38.
2-(Pyridin-3-ylmethyl)quinuclidin-3-ol (3). 2-(Pyridin-3-yl-
methyl)quinuclidin-3-one (2, 2.5 g, 11.6 mmol) was dissolved
in 35 mL of isopropanol. Aluminum tert-isopropoxide (7.1 g,
34.7 mmol) was added. The reaction mixture was refluxed for
6 h, concentrated, and dissolved again in water (100 mL) and
dichloromethane (75 mL). The layers were separated; the water
layer was extracted with dichloromethane (2 ꢀ 75 mL). The
organic layer was dried with Na2SO4, filtered, and concentrated.

Article
3-(Pyridine-3-yl)propyl 4-bromophenylcarbamate (22). 3-Pyr-
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7199
by amine coupling according to the manufacturer’s instructions.
An unmodified dextran surface was used as a reference surface.
A phosphate buffer (10 mM phosphate, pH 7.0, 137 mM
NaCl, 3 mM KCl, with addition of 5% DMSO and 0.005% v/v
surfactant P20 (GE Healthcare, Uppsala, Sweden)) was used as
a running buffer at a flow rate of 90 μL/min. All compounds
were dissolved as 10 mM stock solutions in pure DMSO and
diluted in the running buffer. Suitable concentration series were
determined for all compounds separately. Typically, samples
were injected for 60 s, and the dissociation was recorded for
300 s.
Signals were corrected for nonspecific binding to the surface
by subtracting signals from the reference surface from those of
the AChBP surfaces (reference subtraction). In addition, correc-
tions for minor differences between AChBP and reference sur-
face interactions with DMSO were performed by using a series
of solvents standards (solvent correction). Moreover, signals
were corrected for background by subtracting signals from a
blank injection from those of compound injections (blank
subtraction).
The affinity was determined by fitting a Langmuir binding
equation to steady state binding signals at different concentra-
tions. For the estimation of the affinity of compounds not
reaching saturation within the range of concentrations that
could be used for the measurements, the maximum binding
level (Rmax) was fixed to a value determined by a reference
compound of similar molecular weight. Unless otherwise stated,
standard deviations were based on at least four experimental
series.
idinepropanol (20, 1 mL, 7.7 mmol) was dissolved in 25 mL of
dry THF. 4-Bromophenyl isocyanate (1.7 g, 8.5 mmol) was
added. The reaction mixture was stirred overnight at room
temperature and then concentrated. Saturated NaHCO3 solu-
tion was added (50 mL), and this water layer was extracted with
DCM (2 ꢀ 50 mL). The organic layer was dried with Na2SO4,
filtered, and concentrated. The product was recrystallized from
tert-butyl methyl ether/THF, yielding 0.71 g (2.1 mmol, 27%) of
an off-white solid. 1H NMR (500 MHz, DMSO) δ (ppm) 9.82 (s,
1H), 8.46 (d, J = 2.00 Hz, 1H), 8.41 (dd, J = 4.75, 1.58 Hz, 1H),
7.66 (dt, J = 7.80, 1.92 Hz, 1H), 7.50-7.39 (m, 4H), 7.32 (dd,
J = 7.78, 4.76 Hz, 1H), 4.09 (t, J = 6.55 Hz, 2H), 2.75-2.66 (m,
2H), 2.01-1.90 (m, 2H). 13C NMR (126 MHz, DMSO) δ
153.42, 149.57, 147.29, 138.61, 136.55, 135.79, 131.51, 123.48,
120.02, 113.88, 63.54, 29.74, 28.44.
2-(Pyridine-3-yl)ethanamine HCl (24). 3-(3-Pyridyl)-D-ala-
3
nine (23, 600 mg, 3.6 mmol) was stirred in 12 g of diphenylamine.
This mixture was heated to 200 ꢀC. When the formation of CO2
gas was over (after 1 h), to the resulting yellow solution was
added carefully 1 M HCl solution (50 mL). The mixture was
cooled to room temperature and extracted with diethyl ether
(6 ꢀ 50 mL). The organic layer was dried with Na2SO4, filtered,
and concentrated to yield 309 mg (2.5 mmol, 70%) of the crude
product, a yellow oil. An HCl salt was made with 4 M HCl in
dioxane, and the product was recrystallized from ethanol,
yielding 98 mg (0.50 mmol, 14%) of the double HCl salt, a
white solid. 1H NMR (500 MHz, DMSO) δ (ppm) 8.91 (s, 1H),
8.82 (d, J = 5.54 Hz, 1H), 8.55 (d, J = 8.04 Hz, 1H), 8.40 (s, 3H),
8.02 (dd, J = 7.96, 5.71 Hz, 1H), 3.18 (s, 4H). 13C NMR (126
MHz, DMSO) δ 146.36, 142.07, 140.06, 137.47, 126.90, 38.73,
29.55.
Radioligand Displacement Assay. For the radioligand displa-
cement study, AChBP (Ls- or Ac-) was diluted in PBS-Tris
binding buffer (final concentration of 1.4 mM KH2PO4, 4.3 mM
Na2HPO4, 137 mM NaCl, 2.7 mM KCl, 20 mM Trizma base,
4% DMSO, 0.05% Tween 20, pH 7.4) to obtain a quantity of
N,N-Dimethyl-2-(pyridine-3-yl)ethanamine HCl (25).2-(Pyridine-
3
3-yl)ethanamine (24, 100 mg, 0.8 mmol) was refluxed in formalde-
hyde (2.5 mL of a 35 wt % solution in water/methanol) and formic
acid (2.5 mL). After the mixture was cooled to room temperature,
10% NaOH solution (50 mL) was added and extracted with
dichloromethane (3 ꢀ 50 mL). The organic layer was dried with
Na2SO4, filtered, and concentrated. An HCl salt was made with 4 M
HCl in dioxane, and the product was recrystallized from ethanol,
yielding 66 mg (0.30 mmol, 37%) of the double HCl salt, a yellow
solid. 1H NMR (500 MHz, DMSO) δ (ppm) 11.17 (s, 1H), 8.94 (s,
1H), 8.83 (d, J = 5.43 Hz, 1H), 8.54 (d, J = 8.00 Hz, 1H), 8.02 (dd,
J = 7.92, 5.69 Hz, 1H), 3.42 (s, 2H), 3.35-3.24 (m, 2H), 2.78 (d, J =
2.91 Hz, 6H). 13C NMR (126 MHz, DMSO) δ 145.86, 142.17,
140.44, 137.01, 126.84, 55.75, 41.96, 26.48.
1.3 ng per well. AChBP was incubated with 10-4-10-11
M
ligands (stock concentration of 10 mM in DMSO) in the
presence of approximately 1.5 nM [3H]epibatidine (KD=0.875
nM, Perkin-Elmer Life Scince, Inc.) and 0.2 mg of PVT Copper
His-Tag SPA beads (GE Healthcare). For fragments with pKi
values below 4.5, AChBP was incubated with 10-3-10-10
M
solutions. Final well volume was 100 μL, and the incubation
time was 60 min followed by 3 h in the dark. Thereafter, the
label-bead complexes were counted in a micro-β-counter. Binding
assays for the human R7 nAChR were performed in a similar way
as described for AChBP. However, the cells with R7 nAChR were
homogenized immediately for use and no PVT copper His-Tag
SPA beads were added. Asradioligand, 3H-MLA (KD = 1.81nM,
American Radiolabeled Chemicals, Inc.) was used at 2.5 nM. The
binding assay data were analyzed using Prism 4.0 (Graphpad
Software, Inc.).
Protein Expression and Purification. His6-Lymnaea stagnalis-
acetylcholine binding protein (Ls-AChBP) and His6-Aplysia
california-acetylcholine binding protein (Ac-AChBP) were ex-
pressed using the Bac-to-Bac baculovirus expression system
(Invitrogen, Carlsbad, CA) following the manufacturer’s re-
commendations. Secreted protein was trapped on an HIS-Select
cartridge (Sigma, St. Louis, MO). The cartridge was washed
with 24 mL of 10 mM imidazole, left overnight at 4 ꢀC in
250 mM imidazole/Tris at pH 8, and thereafter eluted with 5 mL
of 250 mM imidazole. A Slide-A-Lyzer dialysis cassette (Pierce,
Rockford, IL) was used according to the manufacturer’s re-
commendations to replace the imidazole by standard binding
buffer. The purity of the protein was checked on an SDS gel, and
the protein concentration was determined by Bradford analysis.
Protein aliquots were stored at -80 ꢀC until use. Human neuro-
blastoma cells (SH-SY5Y) expressing human R7 nAChRs were
kindly provided by Christian Fuhrer (University of Zurich,
Switzerland). Cells were washed with PBS, pelleted by centrifu-
ging, and stored at -80 ꢀC until use.
Modeling. Compound Preparation. Three-dimensional struc-
tures were generated using MOE (version 2008.10, Chemical
Computing Group, Montreal, Canada). In order to compare
racemic compounds 3 and 8 with enantiopure R-isomer 17, the
2S,3R-enantiomers were used in the docking experiments. Pro-
tonation was set such that strong acids and bases were charged.
Partial atomic charges were calculated using Gasteiger charges,
and the molecules were energy-minimized in vacuo using the
MMFF94x force field in MOE.
Template Preparation. The Ls-AChBP crystal structure in
˚
complex with nicotine (PDB accession code 1uv6, 2.5 A) and the
Ac-AChBP crystal structure in complex with epibatidine (PDB
˚
accession code 2byq, 3.4 A) were used for the generation of
docking modes. The ligand and water molecules (except two
conserved waters in Ls-AChBP in contact with nicotine) were
removed, and hydrogen atoms were added to the protein
models. Partial atomic charges were calculated, and energy-
minimization was performed using the AMBER99 force field in
MOE. The docking procedure was performed in one of the
binding pockets formed by two adjacent subunits.
SPR Biosensor Assay. SPR biosensor experiments were per-
formed at 25 ꢀC on a BIAcore T100 (GE Healthcare, Uppsala,
Sweden). Ls-AChBP and Ac-AChBP were diluted to 0.03-0.1
mg/mL in 50 mM NaAc (pH 5.5) and immobilized on a CM5
sensor chip (research grade, GE Healthcare, Uppsala, Sweden)

7200 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
de Kloe et al.
In Silico Docking Procedure. Docking studies were performed
using the GOLD docking program (version 2.0)43 and Chem-
Score and GoldScore scoring functions using default settings
unless stated otherwise. For each ligand, 25 docking poses were
calculated, allowing a cluster size of three ligands within a rmsd
Griffey, R. H. SAR by MS: a ligand based technique for drug lead
discovery against structured RNA targets. J. Med. Chem. 2002, 45,
3816–3819.
(5) Nienaber, V. L.; Richardson, P. L.; Klighofer, V.; Bouska, J. J.;
Giranda, V. L.; Greer, J. Discovering novel ligands for macro-
molecules using X-ray crystallographic screening. Nat. Biotechnol.
2000, 18, 1105–1108.
˚
of 1.5 A. Docking poses were chosen on the basis of their fitness
(6) Fesik, S. W.; Shuker, S. B.; Hajduk, P. J.; Meadows, R. P. SAR by
NMR: an NMR-based approach for drug discovery. Protein Eng.
1997, 10, 73–73.
(7) Gotti, C.; Zoli, M.; Clementi, F. Brain nicotinic acetylcholine
receptors: native subtypes and their relevance. Trends Pharmacol.
Sci. 2006, 27, 482–491.
(8) Taly, A.; Corringer, P. J.; Guedin, D.; Lestage, P.; Changeux, J. P.
Nicotinic receptors: allosteric transitions and therapeutic targets in
the nervous system. Nat. Rev. Drug Discovery 2009, 8, 733–750.
(9) Jensen, A. A.; Frolund, B.; Lijefors, T.; Krogsgaard-Larsen, P.
Neuronal nicotinic acetylcholine receptors: structural revelations,
target identifications, and therapeutic inspirations. J. Med. Chem.
2005, 48, 4705–4745.
(10) Brejc, K.; van Dijk, W. J.; Klaassen, R. V.; Schuurmans, M.;
van der Oost, J.; Smit, A. B.; Sixma, T. K. Crystal structure of
an ACh-binding protein reveals the ligand-binding domain of
nicotinic receptors. Nature 2001, 411, 269–276.
(11) Smit, A. B.; Syed, N. I.; Schaap, D.; van Minnen, J.; Klumperman,
J.; Kits, K. S.; Lodder, H.; van der Schors, R. C.; van Elk, R.;
Sorgedrager, B.; Brejc, K.; Sixma, T. K.; Geraerts, W. P. M. A glia-
derived acetylcholine-binding protein that modulates synaptic
transmission. Nature 2001, 411, 261–268.
(12) Hansen, S. B.; Radic, Z.; Talley, T. T.; Molles, B. E.; Deerinck, T.;
Tsigelny, I.; Taylor, P. Tryptophan fluorescence reveals conforma-
tional changes in the acetylcholine binding protein. J. Biol. Chem.
2002, 277, 41299–41302.
score, ΔG score, and a pharmacophoric filter. This pharmaco-
phoric filter consisted of the known interactions of nicotinic
ligands, especially nicotine and epibatidine in their crystal
structures, namely, cation-π interaction, H-bond with Trp143
backbone, H-bond with conserved water. The pose with the
highest amount of these interactions and the highest fitness
score was chosen (see Table S2 in Supporting Information). In
the case of 8 this corresponded to the pose with the highest ΔG;
however, this pose scored lower because of higher clashes.
Interaction Fingerprint Scoring. Nicotine (in the Ls-AChBP
X-ray structure) and epibatidine (in the Ac-AChBP X-ray
structure) were used to generate reference interaction finger-
prints (IFPs) as previously described.40 Eight different interac-
tion types (cation-π, negatively charged, positively charged,
H-bond acceptor, H-bond donor, aromatic face-to-edge, aro-
matic face-to-face, and hydrophobic interactions) were used to
define the IFP. The cavity used for the IFP analysis consisted of
the 15 residues and 2 water molecules: Y89, S142, W143, T144,
Y185, C187, C188, Y192, W53, L102, A103, R104, L112, Y113,
M114, HOH1090, and HOH1045. Standard IFP scoring
parameters40 and a Tanimoto coefficient (Tc-IFP) measuring
IFP similarity with the reference molecule pose were used to
rank the docking poses generated with Gold/Goldscore and
Chemscore. For the four docking poses in Figure 5, IFP, Goldscore,
and Chemscore values are given in the Supporting Information
Table S2.
(13) Grutter, T.; Le Novere, N.; Changeux, J. P. Rational understand-
ing of nicotinic receptors drug binding. Curr. Top. Med. Chem.
2004, 4, 645–650.
(14) Thompson, A. J.; Lummis, S. C. R. 5-HT3 receptors. Curr. Pharm.
Des. 2006, 12, 3615–3630.
LE Projection on Surface Maps. To visualize LEs of the ligands
on the complementary surface of the binding pocket, protein atoms
(15) Ernst, M.; Bruckner, S.; Boresch, S.; Sieghart, W. Comparative
models of GABA(A) receptor extracellular and transmembrane
domains: important insights in pharmacology and function. Mol.
Pharmacol. 2005, 68, 1291–1300.
(16) Chen, Z. L.; Dillon, G. H.; Huang, R. Q. Molecular determinants
of proton modulation of glycine receptors. J. Biol. Chem. 2004, 279,
876–883.
(17) Grutter, T.; de Carvalho, L. P.; Dufresne, V.; Taly, A.; Fischer, M.;
Changeux, J. P. A chimera encoding the fusion of an acetylcholine-
binding protein to an ion channel is stabilized in a state close to the
desensitized form of ligand-gated ion channels. C. R. Biol. 2005,
328, 223–234.
(18) Paulo, J. A.; Hawrot, E. Effect of homologous serotonin receptor
loop substitutions on the heterologous expression in Pichia of a
chimeric acetylcholine-binding protein with alpha-bungarotoxin-
binding activity. Protein Expression Purif. 2009, 67, 76–81.
(19) Hansen, S. B.; Talley, T. T.; Radic, Z.; Taylor, P. Structural and
ligand recognition characteristics of an acetylcholine-binding pro-
tein from Aplysia californica. J. Biol. Chem. 2004, 279, 24197–24202.
(20) Nordstrom, H.; Gossas, T.; Hamalainen, M.; Kallblad, P.;
Nystrom, S.; Wallberg, H.; Danielson, U. H. Identification of
MMP-12 inhibitors by using biosensor-based screening of a frag-
ment library. J. Med. Chem. 2008, 51, 3449–3459.
ꢀ
˚
within 2.5 A distance of a ligand were scored according to the LE of
that ligand. The ligand with the highest LE was projected first (25).
For the second ligand (3) additional contacts were scored accord-
ing to ΔΔG/ΔMW (i.e., group efficiency44). This procedure was
repeated for the third (8) ligand (22). These LE scores were stored
in the PDB structure of Ls-AChBP as B-factors; the surface was
subsequently colored by B-factors.
Acknowledgment. The authors thank Obbe Zuiderveld for
conducting the radioligand displacement assay and Frans de
Kanter for performing the NOESY experiments. K.R. was
financially supported by Drug Discovery Center Amsterdam,
G.E.d.K., A.B.S., and T.N. were financially supported within
the framework of Top Institute Pharma Project No. D2-103.
The research leading to these results has received funding
from the European Union Seventh Framework Programme
under Grant Agreement No. HEALTH-F2-2008-202088
(“NeuroCypres” project). M.G., A.B.S., and I.J.P.d.E were
financially supported by NeuroCypres.
(21) Sieghart, W. GABA(A) receptors as targets for different classes of
drugs. Drugs Future 2006, 31, 685–694.
(22) Thompson, A. J.; Lummis, S. C. R. The 5-HT3 receptor as a
therapeutic target. Expert Opin. Ther. Targets 2007, 11, 527–540.
(23) Retra, K.; Geitmann, M.; Kool, J.; Smit, G.; de Esch, I. J. P.;
Danielson, U. H.; Irth, H. Development of SPR biosensor assays
for primary and secondary screening of AChBP ligands. Anal.
Biochem. [Online early access]. DOI: 10.1016/j.ab.2010.06.021. Pub-
lished Online: June 17, 2010.
Supporting Information Available: Purity data, as determined
by LC-MS, and exact masses, as determined by HRMS, for all
compounds; NMR and NOESY spectra for 3 and 4; and Gold,
Chem, and IFP scores for relevant docking poses. This material
is available free of charge via the Internet at http://pubs.acs.org.
(24) Geitmann, M.; Retra, K.; De Kloe, G. E.; Homan, E.; Smit, A. B.;
Esch, I. J. P.; Danielson, U. H. Interaction kinetic and structural
dynamic analysis of ligand binding to acetylcholine-binding pro-
tein. Biochemistry [Online early access]. DOI: 10.1021/bi1006354.
Published Online: August 12, 2010.
References
(1) de Kloe, G. E.; Bailey, D.; Leurs, R.; de Esch, I. J. P. Transforming
fragments into candidates: small becomes big in medicinal chem-
istry. Drug Discovery Today 2009, 14, 630–646.
(2) Danielson, U. H. Fragment library screening and lead character-
ization using SPR biosensors. Curr. Top. Med. Chem. 2009, 9,
1725–1735.
(3) Navratilova, I.; Hopkins, A. L. Fragment screening by surface
plasmon resonance. ACS Med. Chem. Lett. 2010, 1, 44–48.
(4) Swayze, E. E.; Jefferson, E. A.; Sannes-Lowery, K. A.; Blyn, L. B.;
Risen, L. M.; Arakawa, S.; Osgood, S. A.; Hofstadler, S. A.;
(25) Danielson, U. H. Integrating surface plasmon resonance biosensor-
based interaction kinetic analyses into the lead discovery and optimi-
zation process. Future Med. Chem. 2009, 1, 1399–1414.
(26) Abad-Zapatero, C.; Metz, J. T. Ligand efficiency indices as guide-
posts for drug discovery. Drug Discovery Today 2005, 10, 464–469.
(27) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a useful
metric for lead selection. Drug Discovery Today 2004, 9, 430–431.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7201
(28) Hajduk, P. J. Fragment-based drug design: How big is too big?
J. Med. Chem. 2006, 49, 6972–6976.
(29) Kuntz, I. D.; Chen, K.; Sharp, K. A.; Kollman, P. A. The maximal
affinity of ligands. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 9997–
10002.
(30) Reynolds, C. H.; Tounge, B. A.; Bembenek, S. D. Ligand binding
efficiency: trends, physical basis, and implications. J. Med. Chem.
2008, 51, 2432–2438.
(31) Bissantz, C.; Kuhn, B.; Stahl, M. A medicinal chemist’s guide to
molecular interactions. J. Med. Chem. 2010, 53, 5061–5084.
(32) Hajduk, P. J. Puzzling through fragment-based drug design. Nat.
Chem. Biol. 2006, 2, 658–659.
(33) Barelier, S.; Pons, J.; Marcillat, O.; Lancelin, J.-M.; Krimm, I.
Fragment-based deconstruction of Bcl-xL inhibitors. J. Med.
Chem. 2010, 53, 2577–2588.
(34) Brandt, P.; Geitmann, M.; Danielson, U. H. Deconstruction of
non-nucleoside reverse transcriptase inhibitors of human immu-
nodeficiency virus type-1 for exploration of the optimization land-
scape of fragments. Manuscript submitted.
(35) Ulens, C.; Akdemir, A.; Jongejan, A.; van Elk, R.; Bertrand, S.;
Perrakis, A.; Leurs, R.; Smit, A. B.; Sixma, T. K.; Bertrand, D.;
de Esch, I. J. P. Use of acetylcholine binding protein in the search
for novel alpha 7 nicotinic receptor ligands. In silico docking,
pharmacological screening, and X-ray analysis. J. Med. Chem.
2009, 52, 2372–2383.
(36) Mazurov, A.; Klucik, J.; Miao, L.; Phillips, T. Y.; Seamans, A.;
Schmitt, J. D.; Hauser, T. A.; Johnson, R. T.; Miller, C. 2-(Arylmethyl)-
3-substituted quinuclidines as selective alpha 7 nicotinic receptor
ligands. Bioorg. Med. Chem. Lett. 2005, 15, 2073–2077.
(37) Bodnar, A. L.; Cortes-Burgos, L. A.; Cook, K. K.; Dinh, D. M.;
Groppi, V. E.; Hajos, M.; Higdon, N. R.; Hoffmann, W. E.; Hurst,
R. S.; Myers, J. K.; Rogers, B. N.; Wall, T. M.; Wolfe, M. L.;
Wong, E. Discovery and structure-activity relationship of quinu-
clidine benzamides as agonists of alpha 7 nicotinic acetylcholine
receptors. J. Med. Chem. 2005, 48, 905–908.
(38) Niemann, C.; Hays, J. T. The relation between structure and
histamine-like activity. J. Am. Chem. Soc. 1942, 64, 2288–2289.
(39) Xiu, X. A.; Puskar, N. L.; Shanata, J. A. P.; Lester, H. A.;
Dougherty, D. A. Nicotine binding to brain receptors requires a
strong cation-pi interaction. Nature 2009, 458, 534–537.
(40) Marcou, G.; Rognan, D. Optimizing fragment and scaffold dock-
ing by use of molecular interaction fingerprints. J. Chem. Inf.
Model. 2007, 47, 195–207.
(41) de Graaf, C.; Rognan, D. Selective structure-based virtual screen-
ing for full and partial agonists of the beta 2 adrenergic receptor.
J. Med. Chem. 2008, 51, 6620–6620 (correction to J. Med. Chem.
2008, 51, 4978-4985) .
(42) Bencherif, M. B.; Benson, L.; Dull, G. M.; Fedorov, N.; Gatto,
G. J.; Genus, J.; Jordan, K. G.; Mathew, J.; Mazurov, A. A.; Miao,
L.; Munoz, J. A.; Pfeiffer, I.; Pfeiffer, S.; Phillips, T. Y. (2S,3R)-
N-(2-((3-Pyridinyl)methyl)-1-azabicyclo[2.2.2]oct-3-yl-benzyofur-
an-2-carboxamide, novel salt forms, and methods of use thereof.
U.S. Patent Application 20090048290, 2008.
(43) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible
docking. J. Mol. Biol. 1997, 267, 727–748.
(44) Verdonk, M. L.; Rees, D. C. Group efficiency: a guideline for hits-
to-leads chemistry. ChemMedChem 2008, 3, 1179–1180.