Hydrogen-bond interactions of nicotine and acetylcholine salts: A combined crystallographic, spectroscopic, thermodynamic and theoretical study

Article abstract of DOI:10.1002/chem.200600808

The hydrogen-bond (HB) interactions of the monocharged active forms of nicotine and acetylcholine (ACh) have been compared theoretically by using density functional theory (DFT) calculations and experimentally on the basis of crystallographic observations and the measurement of equilibrium constants in solution. The 2,4,6-trinitrophenolate (picrate) counterion was used to determine the experimental HB basicity of the cations despite its potential multisite HB acceptor properties. The preferred HB interaction site of the ammonium picrate salts was determined from a survey of crystallographic data found in the Cambridge Structural Database (CSD) and is supported by theoretical calculations. Two distinct classes of ammonium groups were characterised depending on the absence (quaternary ammonium) or presence (tertiary, secondary and primary ammoniums) of an N⁺ H...O hydrogen bond linking the two ions. The crystal structure of nicotinium picrate was determined and compared with that of ACh. This analysis revealed the peculiar behaviour of the ammonium moiety of nicotinic acetylcholine receptor (nAChR) ligands towards the picrate anion. Dedicated methods have been developed to separate the individual contributions of the anion and cation accepting sites to the overall HB basicity of the ion pairs measured in solution. The HB basicities of the picrate anions associated with the two different ammonium classes were determined in dichloromethane solution by using several model ion pairs with non-basic ammonium cations. The experimental and theoretical studies performed on the nicotine and ACh cations consistently show the significant HB ability of the acceptor site of nAChR agonists in their charged form. Both the greater HB basicity of the pyridinic nitrogen over the carbonyl oxygen and the greater HB acidity of the N⁺H unit relative to N⁺CH could contribute to the higher affinity for nAChRs of nicotine-like ligands relative to ACh-like ligands.

Full text of DOI:10.1002/chem.200600808

FULL PAPER
DOI: 10.1002/chem.200600808
Hydrogen-Bond Interactions of Nicotine and Acetylcholine Salts: A
Combined Crystallographic, Spectroscopic, Thermodynamic and Theoretical
Study
[a]
[b]
Virginie Arnaud,[a] MichelBertheol t,
Jean-Yves Le Questel*[a]
MichelEvain, JØrôme Graton,[a] and
Abstract: The hydrogen-bond (HB) in-
teractions ofthe monocharged active
forms of nicotine and acetylcholine
(ACh) have been compared theoreti-
cally by using density functional theory
(DFT) calculations and experimentally
on the basis ofcrystallographic obser-
vations and the measurement ofequi-
librium constants in solution. The 2,4,6-
trinitrophenolate (picrate) counterion
was used to determine the experimen-
tal HB basicity ofthe cations despite
its potential multisite HB acceptor
properties. The preferred HB interac-
tion site ofthe ammonium picrate salts
was determined from a survey of crys-
tallographic data found in the Cam-
bridge Structural Database (CSD) and
is supported by theoretical calculations.
Two distinct classes ofammonium
groups were characterised depending
on the absence (quaternary ammoni-
um) or presence (tertiary, secondary
anion and cation accepting sites to the
overall HB basicity ofthe ion pairs
measured in solution. The HB basici-
ties ofthe picrate anions associated
with the two different ammonium
classes were determined in dichlorome-
thane solution by using several model
ion pairs with non-basic ammonium
cations. The experimental and theoreti-
cal studies performed on the nicotine
and ACh cations consistently show the
significant HB ability of the acceptor
site ofnAChR agonists in their charged
form. Both the greater HB basicity of
the pyridinic nitrogen over the carbon-
yl oxygen and the greater HB acidity
ofthe N +H unit relative to N+CH
could contribute to the higher affinity
for nAChRs of nicotine-like ligands
relative to ACh-like ligands.
+
and primary ammoniums) ofan N
H···O hydrogen bond linking the two
ions. The crystal structure ofnicotini-
um picrate was determined and com-
pared with that ofACh. This analysis
revealed the peculiar behaviour ofthe
ammonium moiety ofnicotinic acetyl-
choline receptor (nAChR) ligands to-
wards the picrate anion. Dedicated
methods have been developed to sepa-
rate the individual contributions ofthe
Keywords: density functional calcu-
lations · hydrogen bonds · IR spec-
troscopy · molecular recognition ·
nAChR agonists · X-ray diffraction
Introduction
Nicotinic acetylcholine receptors (nAChRs) are ligand-gated
ion channels that mediate neurotransmission at the neuro-
muscular junction, both in the peripheral nervous system
and at numerous sites in the central nervous system (CNS).
During the last decade, nAChRs have appeared as valid tar-
gets for treatments against a wide variety of diseases such as
cognitive and attention deficits, Parkinsonꢀs and Alzheimerꢀs
diseases and anxiety and pain management.[1,2] The function-
al importance ofCNS nAChRs and their involvement in nu-
merous pathologies have led to a great number ofinvestiga-
tions aimed at their structural characterisation.[3,4] However,
as for most membrane proteins, structural information
about nAChRs is sparse because they are difficult to purify.
The best experimental structure derived for nAChRs con-
[a] Dr. V. Arnaud, Prof. Dr. M. Berthelot, Dr. J. Graton,
Prof. Dr. J.-Y. Le Questel
Laboratoire de Spectrochimie et ModØlisation
EA 1149, FR CNRS 2465, UniversitØ de Nantes
Nantes Atlantique UniversitØs
FacultØ des Sciences et des Techniques de Nantes
2, rue de la Houssini›re, BP 92208, 44322 Nantes Cedex 3 (France)
Fax : (+33)251-125-567
E-mail: Jean-Yves.Le-Questel@univ-nantes.fr
[b] Prof. Dr. M. Evain
Laboratoire de Chimie du Solide
(Institut Jean Rouxel, CNRS UMR 6502)
UniversitØ de Nantes, Nantes Atlantique UniversitØs
FacultØ des Sciences et des Techniques de Nantes
2, rue de la Houssini›re, BP 92208, 44322 Nantes Cedex 3 (France)
Chem. Eur. J. 2007, 13, 1499 – 1510
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1499

[5]
sists of4.0 Š data obtained rfom cryoelectron microscopy,
In preliminary work[16] we have shown that the picrate
anion (Picꢀ) is the most convenient counterion because it
combines several interesting and/or essential properties: 1)
the salts are readily synthesised by direct neutralisation of
the amine with picric acid or by precipitation ofsilver
iodide between silver picrate and the ammonium iodide,[17]
2) the picrates ofternary and quaternary ammonium ions
are generally soluble enough in low polarity solvents to
enable quantitative physico-chemical measurements, 3) the
asymmetrical geometry and the large size ofthe anion dras-
a resolution that does not allow structure-based design ap-
proaches.
Recently, the crystal structure ofan acetylcholine binding
protein (AChBP) was established as the model for the ex-
tracellular domain ofnAChRs, which contains the ligand-
binding domain (LBD).[6] Furthermore, the resolution ofits
complexes with nicotine (1) and carbamylcholine,[7] both ag-
onists ofACh ( 2), have validated the pharmacophore
tically reduce the association ofthe ion pair into polymeric
[18,19]
species in solvents oflow dielectric constants
and 4) the
competing hydrogen-bond accepting strength ofthe anion is
very weak, ofthe same order ofmagnitude as that ofthe
perchlorate ion.[20] This work supplements our preliminary
investigation ofthe HB properties ofcharged nicotine in
two ways. First, we investigated the HB interactions ofam-
monium–picrate systems from a survey of crystallographic
data found in the Cambridge Structural Database (CSD).
The crystal structure ofnicotinium picrate was determined
allowing a direct comparison with the known structure of
ACh picrate. Then, we carried out experimental measure-
ments and theoretical (B3LYP/6-31+G** level) estimations
ofthe HB basicity ofthe sp nitrogen atom of different mo-
noprotonated nicotinoids and ofthe carbonyl oxygen atom
ofACh and its models.
Altogether, the different analyses carried out in the solid
state, in solution and in the gas phase provide a coherent set
of data allowing, for the first time, a detailed comparison of
the two charged ligands and revealing important informa-
tion about the hydrogen-bonding interactions ofboth phar-
macophore elements ofACh and nicotine.
[8]
models ofBeers and Reich
that have been the subject of
extensive study since the 1970s.[9–11] The two key elements of
the nAChR pharmacophore models, a quaternised nitrogen
atom and a hydrogen-bond (HB) acceptor site, are directly
involved in non-covalent interactions with residues ofthe re-
ceptor. However, a detailed analysis ofthese interactions
and oftheir relative importance in the docking process into
the nAChRs requires high-resolution X-ray data. Dougherty
and co-workers[12] recently showed how a physical chemistry
approach can provide high-precision insights into nAChR–
agonist interactions. In the same vein, in a comprehensive
investigation based on a wide variety ofcomputational
methods, Huang et al.[13] identified the dominant factors dif-
ferentiating the binding strength of agonists in several
forms. Among these factors, hydrogen bonding was recog-
nised as playing a key role, both at the HB acceptor site of
the agonists and at the ammonium site. Despite this impor-
tance, investigations devoted to the study ofnAChR HB
properties are scarce. The neutral form of nicotine, nornico-
2
Results and Discussion
CSD searches
tine, and models ofthese compounds have been the subject
[14]
ofdetailed experimental and theoretical studies,
whereas
Preferred site of interaction of the picrate anion: The picrate
anion contains several non-equivalent organic functionalities
that are able to interact with a positively charged entity: the
phenolate and the three nitro groups. This situation is com-
plicated even more by the fact that the nitro groups them-
selves present several interaction motifs (Scheme 1), for ex-
the agonists interact with nAChRs in their charged form. A
recent computational investigation ofprotonated nicotine
has thrown light on the inlfuence ofthis charge, and ofits
partial or full neutralisation, on the HB accepting properties
ofthe pyridine sp nitrogen atom.[15] Moreover, on the basis
2
[7]
ofcrystal structures and calorimetric studies, Celie et al.
have shown that the greater affinity of nicotine relative to
carbamylcholine is partly due to the presence ofan HB in-
teraction between the protein and nicotine. We have recent-
ly confirmed by using experimental and theoretical data that
the sp2 nitrogen atom is indeed a significant HB acceptor in
monoprotonated nicotine,[16] but this is, to the best ofour
knowledge, the only experimental study devoted to the anal-
ysis ofthe HB properties ofACh agonists in their active
form. This quantification is clearly important for a deeper
characterisation ofthe hydrogen-bonding interactions occur-
ring in their docking into nAChRs and a rationalisation of
their respective affinities.
Scheme 1. Directionality ofthe HB contacts with the nitro group.
ample, with regard to hydrogen bonding.[21] Allen et al. have
shown in their study ofthe HB properties ofnitro oxygen
atoms[21] that ofthe three possibilities presented in
1500
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1499 – 1510

Hydrogen-Bond Interactions ofNicotine and Acetylcholine Salts
FULL PAPER
Scheme 1, the two bidentate
motifs I and II are clearly pre-
ferred over the monodentate
motifIII. On this basis, ifve di-
rections ofapproach ofthe pos-
itively charged group, corre-
sponding to three types ofasso-
ciation, A, B and C, have been
considered (Scheme 2).
In addition to these potential
interactions of the different
oxygen atoms with charged
groups and/or HB donors, p–p
interactions through the picrate
interactions involve the phenolate group. Contacts observed
at the ortho- and para-nitro groups can thus only be defined
as secondary interactions, generally present when the am-
monium group is already engaged with the phenolate
moiety ofa ifrst picrate anion. Another interesting observa-
tion that can be made from the data in Table 1 is the pecu-
liar behaviour ofthe quaternary ammonium–picrate sub-
structure, which presents a unique interaction motifA with
a very low number ofcontacts and a lack ofany short con-
tact. These features reflect the greater steric crowding
around the ammonium nitrogen atom, preventing the ap-
proach ofthe picrate oxygen atoms.
Scheme 2. Directionality ofthe
HB contacts with the picrate
anion.
Geometry of ammonium–picrate interactions: Geometrical
analyses ofthe N +···Oꢀ interactions were carried out by
using the parameters indicated in Figure 1, which represents
aromatic ring also contribute to molecular recognition pro-
cesses.[22] For all these reasons, picric acid appears a versatile
agent particularly well suited to probing the competition be-
tween various intermolecular forces in crystal engineering.[23]
To determine the favoured motif of HB association (A, B or
C), we have listed in Table 1 their distribution among the
Table 1. Distribution ofthe ammonium–picrate contacts observed in the
CSD according to the ammonium and association types.
+
+
+
+
NH4
NH3Z+
NH2Z2
NHZ3
NZ4
[a]
NT
14 (5)
144 (43) 77 (28)
60 (41) 8 (7)
[b]
motifA
NA (NA/NT [%])[c]
8 (57)
75 (52) 43 (56)
48 (83) 8 (100)
short contacts [%][d] 100
99
93
100
0
[b]
motifB
NB (NB/NT [%])[c]
4 (29)
22 (15)
68
7 (9)
57
1 (2)
0
0 (0)
–
Figure 1. Numbering and geometrical parameters ofthe picrate anions.
short contacts [%][d] 100
[b]
motifC
an example ofthe interaction motisf. The distances between
the positive nitrogen atom and the phenolic or nitro oxygen
atoms are noted as d1 and d2, respectively. The a1 and a2 va-
lence angles indicate the interaction directionality, whilst the
dihedral angles f1 and f2 describe, respectively, the position
ofthe ammonium nitrogen atom and the rotation ofthe ad-
jacent NO2 group relative to the plane ofthe picrate anion
ring.
NC (NC/NT [%])[c]
2 (14)
47 (33) 27 (35)
33 15
11 (15) 0 (0)
short contacts [%][d] 100
9
–
[a] Total number ofinteractions with the number ofrefcodes given in pa-
rentheses. [b] The A, B and C notation is defined in Scheme 2. [c] The
number ofcontacts shorter than 3.6 Š with the corresponding percentage
in the ammonium series is given in parentheses. [d] Percentage ofcon-
tacts shorter than the sum ofthe van der Waals radii ( V=3.10 Š) among
the Ni contacts.
Because the preferred picrate site of interaction corre-
sponds without ambiguity to the A motif, we only analysed
the geometric parameters ofthe ammonium–picrate con-
tacts in this configuration. The distances d and the angles a
are presented in Table 2 as a function of ammonium type.
five ammonium series, in addition to the total number of in-
teractions. An intermolecular contact was defined as in the
IsoStar library,[24] that is, as any contact shorter than V+
0.5 Š, where V is the sum ofthe Bondi van der Waals radii.
In the present case, this criterion corresponds to a distance
[25]
of3.6 Š.
Table 2. Geometries[a] observed in the CSD for ammonium–picrate con-
tacts in motifA.
Table 1 shows that the highest number ofcontacts is sys-
tematically observed for mode A, which corresponds to in-
Ammonium
d1 [Š]
d2 [Š]
a1 [8]
a2 [8]
teractions between the ammonium and the phenolate group.
+
+
NH4
2.88(2)
2.86(2)
2.82(2)
2.76(3)
3.49(2)
3.00(5)
3.03(2)
3.12(3)
3.18(3)
4.32(5)
125(5)
131(2)
135(2)
146(1)
166(4)
131(5)
132(2)
140(2)
135(2)
135(3)
Then, with the exception ofthe NH
series, mode C ap-
4
NH3Z+
NH2Z2
pears to be preferred significantly over mode B. Moreover,
the limitation ofthe distance criterion to V=3.10 Š drasti-
cally reduces the percentage ofshort contacts in motisf B
and C as nitrogen alkylation increases, while motifA re-
mains unchanged. These observations confirm that the main
+
+
NHZ3
+
NZ4
[a] The geometric parameters are defined in Figure 1. The standard
errors referring to the last significant digit are indicated in parentheses.
Chem. Eur. J. 2007, 13, 1499 – 1510
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1501

J.-Y. Le Questel et al.
The f1 and f2 dihedral histograms ofthe primary ammoni-
um contacts NH3Z+, which present the greatest number of
interactions, are shown in Figure 2.
an equatorial position, that is to say, anti with respect to the
pyridine ring. The C1-C2-C6-C7 dihedral angle of ꢀ126.88
corresponds to a roughly perpendicular orientation ofthe
pyridine and pyrrolidine rings. These features are in agree-
ment with previous investigations ofneutral and charged
forms of nicotine and appear general whatever the physical
state.[26–30] Conversely, the envelope conformation of the pyr-
rolidine ring has been shown to depend on the state ofthe
molecule.[30,31] In this work, the five-membered ring enve-
lope is characterised by an out-of-plane position of the C6
carbon atom, whereas in its neutral form the N2 nitrogen
atom occupies this position. The geometry ofthe nicotinium
cation corresponds to the A conformer, the most stable in
the gas phase.[29,31] There has been much debate about the
optimal internitrogen N1···N2 distance in the nicotinic phar-
macophore for maximal binding affinity.[10] The N1···N2 dis-
tance of4.70 Š observed in nicotinium picrate is ofthe
same order ofmagnitude as the value obtained for the nico-
tine hydroiodide salt by Barlow and Johnson[32] and is close
to the 4.4ꢁ0.1 Š distance observed for nicotine in the
AChBP binding site.[7]
Figure 2. Histograms ofthe f1 and f2 dihedral angles for primary ammo-
nium–picrate contacts.
It appears from the data in the last column of Table 2 that
the a2 valence angle does not differ significantly between
the different types of ammonium groups. The average value
ofaround 135 8 indicates a cooperative participation ofa
nitro oxygen lone pair in the overall interaction. Conversely,
the a1 valence angle does vary significantly, from 1258 for
NH4+–picrate contacts to 1668 for quaternary ammonium
ones. This evolution reflects the progressive crowding of the
nitrogen atom, which induces more linear interactions with
the phenolate oxygen in the case ofquaternary ammonium
picrates.
Comparison of ACh and nicotinium picrates: The HB inter-
actions between the ammonium groups ofthe nicotinium
and ACh cations (refcode: GEBMEF)[33] and the picrate
anion are displayed in Figure 3a and 3b, respectively. Note
As seen in Figure 2, the two dihedral angles f1 and f2
also behave differently. On the one hand, the values of f1,
with a predilection for values of ꢁ608, reveal that the am-
monium nitrogen atom is not in the plane ofthe picrate
anion. On the other hand, the very large range of f2 values
indicates that the ortho-nitro group freely twists its plane to
better accommodate the cation position.
The variations in the distances d reported in Table 2 are
also very informative. Direct comparison of d1 and d2 shows
that the bonding motifs are always asymmetric, the ammoni-
um nitrogen being closer to the phenolate oxygen atom for
all types ofcation. The distances d1 and d2 discriminate
clearly between two very distinct classes ofammonium pic-
rates. The first class, with mean values d1 =2.83(5) and d2 =
3.08(8) Š, includes all types ofammonium cation bearing at
Figure 3. Crystal structures ofa) nicotinium and b) ACh picrate ion pairs.
The short contact network observed in the asymmetric unit is shown by
the dotted lines. The contacts specific to the ion pairs are shown as bold
dots.
least one N+ H bond. A closer inspection ofthe structures
ꢀ
shows that the interactions correspond to strong (N+)H···Oꢀ
hydrogen bonds. The separations d1 and d2 between the two
+
ions are distinctly larger for quaternary NZ4 ammoniums
that both crystalline structures reveal p–p stacking interac-
tions between two Picꢀ anions in GEBMEF and between
the pyridyl ring and the Picꢀ anion in the nicotinium picrate.
We only examine here the HB interactions between the am-
monium moieties and the picrate anion. The main hydrogen
bond in the nicotinium picrate ion pair involves the ammo-
(d1 =3.49(2), d2 =4.32(5) Š), which constitute the second
class.
Crystalstructures
Conformation of the nicotinium cation in the picrate salt:
The geometrical parameters generally used to describe nico-
tine (1) derivatives are 1) the syn or anti position ofthe N2-
methyl substituent relative to the pyridine ring, 2) the rela-
tive position ofthe two rings and 3) the conformation ofthe
pyrrolidine ring. In the crystal structure ofnicotinium pic-
rate, the methyl group carried by the N2 nitrogen atom is in
nium NH+ and the phenolate Oꢀ groups (dH+
=1.863 Š).
···Oꢀ
The d1 and d2 values of2.767 and 3.146 Š, respectively, are
ofthe same order ofmagnitude as the corresponding mean
+
values for the NHZ3 ammonium cations (2.76(3) and
3.18(3) Š) presented in Table 2. This is also true for the di-
rectionalities as the a1 and a2 angles of150.3 and 126.5 8, re-
spectively, compare reasonably well with the corresponding
1502
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1499 – 1510

Hydrogen-Bond Interactions ofNicotine and Acetylcholine Salts
FULL PAPER
146(1) and 135(2)8 values given in Table 2. The ammonium
nitrogen atom is located in the plane ofthe picrate anion
ring, as shown by the f1 value (2.68). In addition, a signifi-
cant twist ofthe nitro group is revealed by the f2 value
(ꢀ67.38). This rotation allows additional short contacts with
the hydrogen atoms ofthe methyl and methylene groups to
form (Figure 3a). With its phenolate oxygen atom and its
three nitro groups, the picrate anion contains many HB ac-
ceptor sites able to provide secondary interactions with the
hydrogen atoms ofthe cation. Indeed, not less than 8 ofthe
15 hydrogen atoms ofthe nicotinium cation are engaged in
10 contacts that are less than the sum ofthe van der Waals
[25,34]
radii ofthe hydrogen and oxygen atoms (2.65 Š).
The
Figure 4. Electrostatic potential map ofthe naked picrate anion at the
shortest one (1.863 Š) corresponds as expected to the (N+
)H···Oꢀ interaction, this ammonium hydrogen atom also
being engaged in a secondary contact with the oxygen atom
ofan ortho-nitro group. The eight remaining contacts can be
divided into two groups involving six hydrogen atoms ofthe
methylpyrrolidinium group on the one side and a pyridyl hy-
drogen atom on the other. In comparison, 10 ofthe 16 hy-
drogen atoms ofthe ACh picrate are involved in 11 short
contacts with either different picrate moieties or the carbon-
yl group ofanother ACh molecule. These hydrogen atoms
belong both to CH groups directly linked to the ammonium
nitrogen atom and also to the acetyl group (Figure 3b). The
mean distance ofthese 11 (C)H···O contacts (2.413 Š) is sig-
nificantly shorter than the corresponding one for the nicoti-
nium cation (six (C)H···O contacts, 2.551 Š). This suggests
that the positive charge on the nitrogen atom is spread con-
siderably more over the adjacent CH bonds in ACh than in
the nicotinium cation, in which the charge is more localised
on the NH group. Finally, note that such CH···O interactions
are observed in the crystal structures ofnicotine and carba-
mylcholine bound to AChBP[7] and have recently been re-
ported by Dougherty[12] and Huang[13] and their co-workers
in their molecular modelling studies devoted to their dock-
ing into nAChR models. These contacts involve nicotine and
carbamylcholine CH groups and the oxygen atoms ofpro-
tein residues (the carbonyl group oftryptophan and the phe-
nolic oxygen atom oftyrosine).
molecular surface (values shown are Vs,min [kJmolꢀ1]).
line picrate environments (Table 1). Furthermore, the loca-
tion ofthe minima clearly supports the afvoured bidentate
nature ofthe interaction with the nitro groups (motisf I and
II ofScheme 1).
The most positive electrostatic potentials ofthe naked
ACh cation surface appear approximately in the direction of
the centres of the four triangular faces formed by the
carbon atoms ofthe CH 2N+
(CH3)3 distorted tetrahedron.
N
This preference of a basic centre to bind to hydrogen atoms
belonging to different carbon atoms rather than to the same
one has already been rationalised through theoretical calcu-
lations.[37] The electrostatic potential maximum is located on
the face formed by the methylene and two methyl groups,
which is opposite the acetoxy substituent. As shown in
Figure 3, this is indeed the face selected by the negative
oxygen atom ofthe picrate anion in the solid state. The
+
NH+ proton ofthe naked NicH cation surface exhibits the
maximum electrostatic potential. The Vs,max =593 kJmolꢀ1
value is about 90 kJmolꢀ1 higher than the maximum induced
by three cooperative protons ofthe CH N+
A
2
ACh.
Electrostatic potentials of the ion pairs: The ion-pair electro-
static potentials presented in Figure 5 for different ammoni-
um picrates are also very useful. First, the electrostatic po-
tential minima around the two nitro groups in the ortho po-
sition are now always different owing to the asymmetry of
the ion pair. Secondly, the electrostatic potential minima of
the nitro oxygen atoms in the para position are always more
negative (indicating more basic atoms) by 15–30 kJmolꢀ1
than the nitro oxygen atoms in the ortho positions. The
para-nitro group appears therefore as the major site of asso-
ciation ofa hydrogen-bond donor in the ion pairs. Thirdly,
Theoretical calculations
Electrostatic potentials of the naked anion and cations: Anal-
ysis ofthe electrostatic potentials on the picrate ion surafce
(Vs,min) is particularly well suited to highlighting the behav-
iour ofits various organic ufnctionalities. Sawada et al.,
through Vmin calculations (HF/6-31G*),[35] and Troxler et al.,
from an analysis of partial atomic charges fitted on electro-
static potentials,[36] found that the most electron-rich site of
the anion is the phenolate oxygen. These results indicate
that the internal charge transfer toward the strong electron-
withdrawing nitro substituents is not sufficient to completely
deplete its negative charge in favour of the nitro oxygen
atoms. Our calculations at the B3LYP/6-31+G** level of
theory, presented in Figure 4, confirm the pre-eminence of
site A in agreement with the preference observed in crystal-
+
as observed in Table 2 for the NZ4 fragment, the geometri-
cal parameters ofthe corresponding tetramethylammonium
picrate (NMe4+Picꢀ) model differ significantly from those of
all the other substituted cations. The longer charge separa-
tion in NMe4+Picꢀ reduces the withdrawing effect of the
positive charge on the electrons ofthe phenolate oxygen
atom. Thus, the internal charge transfer of this negative
oxygen atom toward the nitro groups is considerably en-
Chem. Eur. J. 2007, 13, 1499 – 1510
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1503

J.-Y. Le Questel et al.
+
Figure 5. Geometries [Š] and local electrostatic potential minima [kJmolꢀ1] of alkylammonium picrates (from left to right: MeNH3+, Me2NH2
Me3NH+ and NMe4).
,
hanced giving much more basic nitro HB accepting sites. Fi-
nally, the distances d computed for the ammonium species
bearing at least one hydrogen atom correspond to strong
HBs, as revealed by the (N+)H···Oꢀ distances (from 1.41 to
1.53 Š). These theoretical tendencies are in good accord
with the former solid-state observations. They also follow
the order of measured nAChR–agonist affinities because the
ligands bearing tertiary (nicotine, cytosine) or secondary
ammonium moieties (epibatidine) generally have greater af-
finities than ACh-like ligands carrying quaternary ammoni-
ums.[7,12]
sponds to the one in which the phenolate oxygen atom inter-
acts with the NH+ ammonium group. Note that not less
than three different orientations of the picrate anion with
respect to the nicotinium cation are found within a range of
3.6 kJmolꢀ1, corresponding to 97% ofthe ion-pair popula-
tion in the NicH+Picꢀ system. The AChPicꢀ PES is in good
agreement with the previous electrostatic potential analysis.
Indeed, the CH2N+
(CH3)3 tetrahedron face showing the
R
greatest Vs,max value is involved in the three most stable
structures found. These geometries correspond to 86% of
the ion-pair population in the AChPicꢀ complexes and thus
the complexes relative to the three other faces are consid-
ered negligible.
Finally, the tertiary ammonium cation gives an interaction
energy 50 kJmolꢀ1 stronger than the quaternary one (DEel =
ꢀ364 and ꢀ313 kJmolꢀ1, respectively). In the same vein, op-
timisations ofthe NicH +Clꢀ (DEel =ꢀ462 kJmolꢀ1) and
AChClꢀ complexes (DEel =ꢀ388 kJmolꢀ1) indicate the
stronger HB donor ability ofthe tertiary ammonium cation
and the stronger HB acceptor ability ofthe chloride relative
to the picrate anion.
Exploration of the NicH+Picꢀ and AChPicꢀ potential energy
surfaces (PES): In order to scan thoroughly the potential
energy surfaces (PESs) of the cation–anion interactions in
NicH+Picꢀ and AChPicꢀ ion pairs, an intermolecular model
potential was used to optimise the interaction geometry be-
tween the two ions, starting with 400 initial structures (see
the Experimental Section for details). Two stable conform-
ers, denoted A and B,[30] in which the pyridine and pyrroli-
dine rings are roughly perpendicular to one another, have
previously been found for the naked NicH+ cation. As the
energy difference between these two rotamers is insignifi-
cant, two PESs were investigated: the interaction ofeach of
the two conformers with the picrate anion. The most signifi-
cant minima for the three ion pairs were then fully opti-
mised at the B3LYP/6-31+G** level oftheory.
In agreement with the trends revealed by crystallographic
analysis, the phenolate oxygen atom is found again to be, in
vacuo, the predominant interaction site ofthe picrate anion
(motifA). An interaction with the para-nitro group (motif
C) only occurs in NicH+Picꢀ and is too unfavourable to be
significant (about 65 and 80 kJmolꢀ1 above the most stable
A motif, respectively, for the A and B isomers), whereas the
interaction geometries ofmotifB do not lead to any mini-
mum. In this system, the optimised ion pairs have very dis-
tinct energies depending on the cationic isomer. Interactions
with rotamer B are significantly destabilised (by at least
8.6 kJmolꢀ1) in comparison with the global minimum and so
they have not been further considered in this study. In
accord with the crystal data, the most stable ion pair corre-
Calibration of HF hydrogen-bond association with pyridine
and ester families: Following the preliminary analysis ofLa-
marche and Platts,[38] we have recently shown[39] that the ex-
2
perimental pKHB HB basicity scale ofpyridine Nsp and
amide-ketone CO bases can be precisely predicted from the
variation in the electronic energies DEel for their association
with hydrogen fluoride provided that the data for the two
families are considered separately. In this study, we used
Equation (1) to determine pKHB values for the association
2
ofpFP with pyridine sp nitrogen sites ofnicotine-like cat-
ions. We also calibrated a new family corresponding to the
association with oxygen esters in order to determine the ba-
sicity ofthe carbonyl site ofACh and its models with the
best possible precision. These experimental and theoretical
data are presented in Table 3 and lead to Equation (2). The
validity ofthis latter equation can be tested with 2-dimethyl-
aminoethyl acetate, a bifunctional Nsp3 +CO ester base. Ex-
perimentally, the only accessible constant is Kt =
63 dm3 molꢀ1 and this constant can be separated into the two
1504
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1499 – 1510

Hydrogen-Bond Interactions ofNicotine and Acetylcholine Salts
Table 3. Experimental and theoretical energies ofHB association with esters.
FULL PAPER
of0.74 between the protonated
and the neutral species corre-
Compound
Formula
Experimental donor: pFP Theoretical donor: HF
pKHB
ꢀDEel [kJmolꢀ1
]
sponds to
a
decrease of
methyl 1,1,1-trichloroacetate
2,2,2-trichloroethyl acetate
ethyl formate
2-chloroethyl acetate
methyl benzoate
CCl3CO2Me
MeCO2CH2CCl3
HCO2Et
MeCO2CH2CH2Cl
C6H5CO2Me
MeCO2Me
Me2C=CHCO2Me
MeCO2CH2CH2NMe2
MeCO2Et
0.11[a]
0.47[b]
0.66[a]
0.81[b]
0.89[a]
1.00[a]
32.3
40.8
41.7
43.2
44.2
45.9
45.2
46.7
47.0
47.1
61.1
5 kJmolꢀ1 in Gibbs free energy.
In the same way, the carbonyl
ꢀ
group basicity ofAChPic
is
significantly less than that of 2-
dimethylaminoethyl acetate
methyl acetate
methyl 3,3-dimethylacrylate
2-dimethylaminoethyl acetate
ethyl acetate
1.02[b]
(dpKHB =0.57). Finally, the ter-
tiary ammonium NicH+ exhib-
its a greater maximum electro-
static potential and higher bind-
ing energies with both picrate
and chloride than the quaterna-
[c]
1.07[a]
1.11[b]
2.11[b]
propyl acetate
MeCO2CH2CH2CH3
ethyl trans-dimethylaminoacrylate Me2NCH=CHCO2Et
[a] See ref. [40]. [b] This work. [c] Bifunctional (tertiary amine+carbonyl ester) compound: the experimental
constant is Kt =63 dm3 molꢀ1
.
Table 4. Theoretical estimations ofthe HB accepting strength ofACh
and nicotine cations.
individual contributions by using the frequency shifts
method[41,42] as follows. In addition to its free OH band at
n˜ =3644 cmꢀ1, the association ofmethanol with this base in
CCl4 presents two large OH absorptions at 3571 and
3276 cmꢀ1 corresponding to associations with the carbonyl
and tertiary amine groups, respectively. The correlation be-
tween Dn˜(OH) and pKHB proposed by Besseau et al.[40] for
the ester family allows an experimental value to be calculat-
ed: pKHB =0.94. On the other hand, the variation ofthe
electronic energy ofthe most stable complex (Table 3)
yields pKHB =1.05 by using Equation (2). Thus, both indirect
experimental and calculated pKHB values agree fairly well
with the value expected for an acetate carbonyl group sub-
stituted in the g position by a weak electron-withdrawing
group.[43]
[a]
Compound
Vs,min
ꢀDEel
pKHB
Vs,max
[kJmolꢀ1
]
[kJmolꢀ1
]
[kJmolꢀ1
]
C=O site
N+ site
ACh
+114.1
ꢀ88.6
+502.6
ACh Picꢀ
38.7
46.7
0.48[b]
1.05[b]
ACh Clꢀ
ꢀ105.8
2-dimethylaminoethyl ꢀ152.0
acetate
Nsp2 site
55.2
N+ site
+593.2
NicH+(A)[c]
+93.5
ꢀ113.5
ꢀ132.0
ꢀ163.3
NicH+ Picꢀ (A)[c]
NicH+ Clꢀ (A)[c]
nicotine (A)[c]
1.37[e]
2.11[d]
62.2
[a] Hydrogen-bond binding energies for the complexation with HF.
[b] Calculated from Equation (2). [c] Rotamer A (see text). [d] Calculat-
ed from Equation (1).
pKHBðpyridinic sp2
¼ ꢀ0:1057 DEelꢀ4:464
nitrogenÞ
ð1Þ
ð2Þ
r ¼ 0:9995, n ¼ 11, s ¼ 0:03
ry ammonium ACh. In summary, the theoretical calculations
show that both charged ligands keep a significant HB basici-
ty and that protonated nicotine appears not only slightly
more basic, but also much more acidic than ACh.
pKHBðC¼O estersÞ ¼ ꢀ0:0719 DEelꢀ2:302
r ¼ 0:991, n ¼ 10, s ¼ 0:07
Hydrogen-bonding ability of ACh and NicH+: The results of
the theoretical calculations on the most stable ionic forms of
the two ligands are summarised in Table 4. Comparison of
the minimum electrostatic potentials ofthe carbonyl group
and the pyridine nitrogen atom indicates that 1) the strong
electron-withdrawing charge ofthe positive nitrogen atom
totally depletes the HB basicities ofthe acceptor sites in
both naked cations (compare ACh with 2-dimethylami-
noethyl acetate and NicH+ with nicotine), 2) these basic
sites recover proportionally greater HB affinities when their
positive charges are neutralised by more basic anions (Clꢀ >
Picꢀ) and 3) the Vs,min values for the two cations in the pic-
rate salts are still very different from those of their neutral
reference bases indicating a substantial loss of basicity be-
tween the neutral models and their picrate salts. The elec-
tronic energy values DEel confirm the notable reduction in
sp2 nitrogen HB basicity ofnicotine resulting rfom protona-
tion ofthe pyrrolidine nitrogen atom. The p KHB difference
Experimentaldetermination of the basicities
HB basicity of the anion: The experimental data for the pic-
rate derivatives ofnicotine, ACh and their model amines
are shown in Table 5. The frequency shifts Dn˜OH(NO2) of
R
the 4-fluorophenol (pFP) hydroxy vibration upon associa-
tion with the nitro groups clearly characterise two distinct
series ofammonium picrates, in line with the trends re-
vealed by crystal and theoretical data. On the one hand, the
very large Dn˜OH shifts of the three quaternary ammonium
picrates 2, 10 and 11 indicate the very important HB basicity
[44,45]
oftheir nitro oxygen atoms.
The thermodynamic quan-
tification of this basicity is given by the total equilibrium
constant Kt ofthe tetrabutylammonium ( 10) and trimethyl-
dodecylammonium (11) picrates in which the only potential
basic sites are located on the para-nitro groups ofthe pic-
rate anions (see above). Unexpectedly, the significant reduc-
Chem. Eur. J. 2007, 13, 1499 – 1510
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1505

J.-Y. Le Questel et al.
Table 5. Association of 4-fluorophenol with ammonium picrates. Equilibrium constants and frequency shifts in dichloromethane.
Kt[c]
KC+
KPicꢀ
n[f]
[d]
[e]
Compound
no.
Picrate (Picꢀ)[a]
Dn˜OH
U
[dm3 molꢀ1
]
[dm3 molꢀ1
]
[dm3 molꢀ1
]
[b]
3
4
5
6
1
7
8
pyrrolidinium
methyldodecylammonium
N-methylpyrrolidinium
112
120
119
115
108
112
108
5.3
2.7
3.9
–
–
–
–
9.7
8.7
5.3
2.7
3.9
3.1
2
4
2
4
3
4
1
N,N-dimethyldodecylammonium
3.1
nicotinium (NicH+)
14.5
12.4
10.4
ACHTREUNG
N,N-dimethylaminomethylpyridinium (Me2AMP)
N-ethyl-N-methylaminomethylpyridinium (Et-
MeAMP)
ACHTREUNG
AHCTREUNG
9
noracetylcholine (norACh)
tetrabutylammonium
trimethyldodecylammonium
acetylcholine (ACh)
111
369
364
359
2.8
2
2
4
9
10
11
2
22.9[g]
23.6
–
–
22.9
23.6
22.7
[a] The acronyms ofthe compounds used throughout the text are indicated in parentheses. [b] Frequency shitf ofpFP upon association with the nitro
group. [c] Equilibrium constant determined from the OH absorption of pFP: Kt =KC+ +KPicꢀ. [d] Experimental individual association constants measured
at the pyridinic site ofthe cation. [e] Experimental or calculated (in parentheses) equilibrium association constants measured ofr the picrate ani on.
[f] Number of independent measurements. [g] Kuntz (see ref. [20]) found Kt =20 dm3 molꢀ1 for the phenol–tetrabutylammonium picrate complex in the
same solvent.
tion in the effective size of the trimethylammonium head of
11 by comparison with the tributylammonium head of 10
does not induce any great change in nitro basicity. In the
same vein, on passing from the alkylated model 11 to ACh
(2), the presence ofan electron-attracting acetoxy substitu-
ent in 2 does not influence the frequency shift Dn˜OH due to
the association with the nitro group. On the other hand, the
substitution ofthe first alkyl group ofa quaternary ammoni-
um cation by a hydrogen atom leads to a very large decrease
in the OH frequency shift. The lower basicity of the nitro
groups ofthe tertiary ammonium picrates 1 and 5–9 can be
attributed to the lower tendency ofthe phenolic negative
charge to delocalise in the picrate ring as it is held back by
the hydrogen bond between the two ions.[46,47] Substitution
ofa second alkyl group by a hydrogen atom in the pyrrolidi-
nium (3) and methyldodecylammonium (4) cations causes
little change in basicity, as previously observed by Wit-
schonke and Kraus.[48] One can expect a strengthening of
the cation···anion hydrogen bond when an electron-with-
drawing pyridinyl substituent is added to the molecule lead-
ing to a further decrease in the anion basicity. However,
direct comparison ofthe association constants for the associ-
ation ofpFP on the nitro groups of N-methylpyrrolidinium
(5) and N,N-dimethyldodecyl-
HB basicity of the nicotine and ACh cations: The missing
equilibrium constant corresponding to the association of
pFP with the accepting pyridinic nitrogen atom ofthe Et-
MeAMP cation (8) can then be obtained by simple subtrac-
tion: KC+ =KtꢀKPicꢀ =6.6 dm3 molꢀ1. The situation is differ-
ent for the HB associations of pFP on the carbonyl groups
ofthe ACh ( 2) and norACh (9) cations which require specif-
ic evaluations. Independent estimations ofthese basicities
can be obtained from their carbonyl absorptions because
the C=O stretching frequency is known to be a very sensi-
tive probe ofmolecular interactions ofcarbonyl groups. The
limiting condition is that mass and coupling effects are mini-
mised by using compounds with similar structures.[49,50] We
therefore completed our database on the hydrogen-bond ba-
sicity pKHB (in CCl4) ofesters [40] by studying two methyl ace-
tates substituted by electron-attracting atoms on the ether
moiety in order to mimic more closely the substituent effect
ofthe positive nitrogen atom in ACh and norACh. The
pKHB data reported in Table 6 can be converted into logK
values in dichloromethane by using Equation (3), the linear
free energy relationship (LFER) relevant to the family of
carbonyl compounds studied herein. Then, the good linear
relationship [Equation (4)] obtained between the thermody-
ammonium (6) picrates on the
Table 6. Equilibrium constants and carbonyl frequencies for the association of pFP with some esters substitut-
ed on the ether group.
one hand and nicotinium (1),
Me2AMP (7) and EtMeAMP
(8) picrates on the other, does
not show any influence of the
substituent. Thus, we can safely
predict that the mean value of
KPicꢀ =3.8 dm3 molꢀ1 for com-
pounds 1 and 5–7 also holds for
EtMeAMP (8). This conclusion
is supported by the similar fre-
quency shifts observed for these
compounds.
Compound
n˜C
N
pKHB
U
U
K(CH2Cl2)
N
=
]
MeCOOCH2CCl3
MeCOOCH2CH2Cl
MeCOOCH3
1759.8
1742.7
1739.5
1733.3
1750.4
0.47
0.81
ꢀ0.15[a]
0.14[a]
0.31[a]
0.37[a]
0.03[c]
0.70
1.39
2.04
2.36
1.08
1.00[b]
1.07[b]
–
MeCOOCH2CH3
MeCOOCH2CH2NMe2H+
(NorACh Picꢀ; 9)
MeCOOCH2CH2NMe3 (ACh Picꢀ;
1753.5
–
ꢀ0.03[c]
0.94
+
2)
[a] Calculated from the pKHB value by using Equation (3). [b] See ref. [40]. [c] Calculated from n˜(C
by using
=
O)
Equation (4).
1506
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1499 – 1510

Hydrogen-Bond Interactions ofNicotine and Acetylcholine Salts
FULL PAPER
namic parameter logK and the carbonyl frequency enables
the equilibrium constants for the association of the carbonyl
site with the two cations 2 and 9 to be calculated (Table 6).
As predicted by the theoretical calculations, the sp2 nitrogen
atom ofnicotine keeps a significant basicity in the monopro-
tonated form since the electron-withdrawing effect of the
ammonium charge is largely neutralised by the interaction
with the picrate anion. Experimentally, the equilibrium con-
stant for the association of pFP with the sp2 nitrogen atom
ofnicotine picrate is seven times greater than the associa-
tion constant for the interaction with the carbonyl group of
ACh picrate. The corresponding ratio (corresponding to
5 kJmolꢀ1 in free energy units) measured for the cations
does not differ significantly from that found in the corre-
log KðCH2Cl2Þ ¼ 0:878 pKHBꢀ0:56
ð3Þ
r ¼ 0:997, n ¼ 14, s ¼ 0:05
~
log KðCH2Cl2Þ ¼ ꢀ0:0205 nC¼O þ 35:8
r ¼ 0:984, n ¼ 4, s ¼ 0:05
ð4Þ
sponding neutral forms (K
(nicotine)=108;[51] K(2-dimethyl-
A
The comparison ofthe anion–cation interactions ofACh
and norACh picrates through their carbonyl and nitro
probes is totally coherent with previous observations. Thus,
the presence ofan (N +)H···Oꢀ hydrogen bond in norACh
Picꢀ definitely results in a neutralisation of the two opposite
charges leading concomitantly 1) to an attenuation ofthe
electron-withdrawing inductive effect of the ammonium sub-
stituent on the carbonyl group (the carbonyl group is more
basic, Table 6) and 2) to a large reduction in the electron-
donating resonance effect of the phenolic oxygen negative
charge on the nitro group (the nitro group is less basic,
Table 5).
aminoethyl acetate)=9.9) and seems to be insensitive to the
nature ofthe counterion, as indicated by our Vs,min calcula-
tions on the hydrochloride salts (Table 4).
Comparison with docking energies: In the last few years, the
X-ray crystal structure ofAChBP has been used to build,
through homology modelling, the LBD ofnAChR subtypes.
From these three-dimensional atomic models, molecular
docking ofACh, nicotine and epibatidine into nAChR bind-
ing sites has been achieved. Through their docking studies
ofnicotine and ACh into homomeric chick a7 nAChR, Le
Nov›re et al.[52] estimated a greater binding free energy (of
about 6 kJmolꢀ1) for nicotine. Similarly, Schapira et al.[53]
docked both ACh and nicotine into the a4b2, a3b2 and a7
atomic models. They found that the inclusion of a water
molecule in the binding pocket provides additional stability
to the complexes by forming small arrays of hydrogen bonds
between the accepting sp2 nitrogen atom ofnicotine or the
carbonyl oxygen ofACh and residues ofthe surrounding
protein. In these studies, nicotine is again favoured by about
3 kJmolꢀ1 compared with ACh. Celie et al.[7] confirmed the
presence ofa bridging water molecule in high-resolution
crystallographic structures ofnicotine–AChBP complexes.
Furthermore, their thermodynamic measurements ofthe
ligand dissociation by isothermal titration calorimetry
(DG=ꢀ34.6 and ꢀ41.8 kJmolꢀ1 for ACh and nicotine, re-
spectively) correctly match the docking results. Ofall the in-
teractions, the numerous HBs created between the ligands
and the protein may play a crucial role. Those involving the
accepting centres ofthe ligand have long been recognised as
essential features of the nicotinic pharmacophore,[8,10] but
they have never been quantified: first, because the HB
donor was not known and secondly, because the acceptor
atoms belong to charged molecules for which there are no
reference data in the literature. Quantitative LFERs be-
tween numerous kinds ofdonor have been set up by Abra-
ham and co-workers[54,55] so that, provided the pKHB value of
an accepting group is known, its equilibrium association
constant with water can be calculated. In this analysis, equi-
librium constants of K=2.1 and 0.7 dm3 molꢀ1 were obtained
for the HOH···Nsp2 (nicotine) and HOH···O=C (ACh) com-
plexes, respectively. The energy difference of 2.7 kJmolꢀ1
between the two associations corresponds fairly well to the
differentials measured by calorimetry or calculated with the
docking models. One must bear in mind that the specific hy-
Table 7. Hydrogen-bond basicity ofthe anionic and cationic sites ofthe
ligand picrates.
Compound no.
Picrate (Picꢀ)
K
N
pKHB
K
G
pKHB
1
7
8
Nic
Me2AMP
EtMeAMP
27.5
24.4
18.1
1.44
1.39
1.26
40.7
30.7
31.6
1.61
1.49
1.60
C=O cation site
NO2 anion site
13.1
213.0
9
2
norACh
ACh
4.9
4.2
0.69
0.62
1.12[a]
2.33[a]
+
[a] Obtained from the difference KtꢀKC in CH2Cl2 and converted to the
corresponding value in CCl4 by using Equation (6).
Finally, we have summarised in Table 7 the individual as-
sociation constants ofthe pFP on the pyridyl and carbonyl
sites of the different ligands. These data were obtained from
the results given in Tables 5 and 6 and were converted from
dichloromethane to CCl4 using the LFERs described by
Equations (5), (6) and (7), which correspond, respectively,
to the different families of the pyridinic sp2 nitrogen, the
nitro oxygen (among other oxygen bases) and the carbonyl
oxygen bases.
Nsp2 :
pKHB ¼ 1:094 log KðCH2Cl2Þ þ 0:36
r ¼ 0:9997, n ¼ 7, s ¼ 0:02
ð5Þ
ð6Þ
ð7Þ
OðNO2Þ : pKHB ¼ 1:092 log KðCH2Cl2Þ þ 0:87
r ¼ 0:997, n ¼ 15, s ¼ 0:09
OðC¼OÞ : pKHB ¼ 1:132 log KðCH2Cl2Þ þ 0:65
r ¼ 0:997, n ¼ 14, s ¼ 0:06
Chem. Eur. J. 2007, 13, 1499 – 1510
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1507

J.-Y. Le Questel et al.
The model potential, dubbed MP2fit/DMA, has been fully described pre-
viously.[15,61–63] It consists ofan atom–atom 6-exp potential to depict the
repulsion and dispersion terms augmented with an accurate distributed
multipole analysis (DMA)[64,65] model to depict the dominant electrostat-
ic interaction term. The GDMA[66] program was used to compute the
DMA models from the Gaussian MP2/6-311G**//B3LYP/6-31+G** mo-
nomer wave functions and the model potential of the ion pairs was calcu-
lated with the ORIENT program.[67]
drogen-bond energy between a water molecule and the ac-
cepting centre ofthe ligand is only one component ofthe
total interaction energy expressed in the experimental DG
measured by calorimetry. We have also seen that the dra-
matic differences in the HB donating properties of the two
cationic heads play a crucial role in their interactions with
the picrate anion. Owing to the number and diversity ofthe
hydrogen bonds involved in the docking process, this result
appears to be a first quantitative step in the understanding
ofmultiple ligand interactions with nAChRs.
Electrostatic potentials around all the optimised ion pairs were mapped
on the 0.001 electronbohrꢀ3 contour ofthe electronic density.
2
The HB ability ofthe basic sites (the pyridine sp nitrogen or the carbon-
yl oxygen atom in the reference compounds and in the salts) was calcu-
lated in vacuo using hydrogen fluoride (HF) as the HB donor probe
[Equation (8)]). HF has already been used with success to predict HB en-
ergetic parameters both for charged[16] and neutral[39] forms of ligands
through LFERs between the experimental Gibbs energy ofpFP complex-
ation, DGA298 (DG2A98/kJmolꢀ1 =ꢀ5.71pKHB), measured in CCl4, and the
calculated density functional (B3LYP/6-31+G**) electronic energy var-
iation DEel ofHF complexation. In this work, the binding energy ofHF
to a basic site was calculated through Equation (9) according to the su-
permolecule approach.
ExperimentalSection
X-ray crystallography: A crystal ofnicotinium picrate was mounted on
the tip ofa Lindemann capillary using solvent-rfee glue. The intensity
was measured at 150 K on a Bruker-Nonius Kappa CCD diffractometer
using graphite-monochromatised MoK-L2,3 radiation (l=0.71073 Š) and
an Oxford cryostream cooler to obtain the low temperature. Data were
corrected for Lorentzian polarisation effects and absorption corrections
were applied using a Gaussian integration. Friedel pairs were merged.
The structure was solved by Sir2004 direct methods[56] and subsequent
calculations were carried out using the Jana2000 program package.[57] All
non-hydrogen atoms were refined anisotropically and hydrogen atoms
were introduced with geometrical constraints and riding atomic displace-
ment parameters.
Base þ HF Ð Base Á Á Á HF
ð8Þ
ð9Þ
DEel ¼ EelðBase Á Á Á HFÞꢀEelðHFÞꢀEelðBaseÞ
The basis set superposition error (BSSE) was not taken into account be-
cause it is expected to be quasiconstant within the two homogeneous
families of pyridines and carbonyl model compounds that are considered
separately in this study. All stationary points (molecular bases, ion pairs
and their HB complexes) were confirmed as true minima by vibrational
frequency calculations.
CCDC 604836 contains the supplementary crystallographic data for this
paper. These data can be obtained free of charge from the Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Crystaldata for nicotinium picrate : C16H17N5O7; Mr =391.3; orthorhom-
bic; P212121; a=7.2166(4), b=12.4243(6), c=19.5563(16) Š; V=
Synthesis and purification: 2,2,2-Trichloroethyl acetate and 2-choroethyl
acetate were synthesised by standard methods from the corresponding al-
cohol and acetic anhydride in the presence ofpyridine. The compounds
were distilled, dried over molecular sieves and their purity checked by
gas-phase chromatography. The picrates were generally synthesised by
direct reaction ofthe amine with picric acid in methanol/water or etha-
nol/water. Quaternary ammonium picrates were obtained by precipita-
1753.44(19) Š3; Z=4; 1calcd =1.4819 gcmꢀ3; m=0.119 mmꢀ1; F
(000)=816;
A
yellow plate; 0.5ꢂ0.4ꢂ0.13 mm3; 2qmax =69.988; T=150 K; 21072 reflec-
tions, 4277 unique (99% completeness); Rint =0.075; 253 parameters;
GOF=2.26; wR2 =0.1212; R=0.0449 for 2604 reflections with I>2s(I).
Cambridge structuraldatabase searches : Crystallographic data were re-
trieved from the 5.27 release (November 2005, 355064 entries) of the
CSD.[58] The ConQuest program[59] was used to search for bonded sub-
structures and intermolecular non-bonded contacts. The structures were
visualised with MERCURY[59] and the geometrical data analysis was per-
formed with VISTA.
[17]
tion ofsilver iodide between silver picrate and the ammonium iodide.
The crystals were recrystallised in ethanol/water and carefully dried
under vacuum. Their purity was checked by thin-layer chromatography
and infrared spectroscopy.
FTIR spectroscopy: The experimental determination ofthe strength ofa
remote hydrogen-bond accepting site in an organic cation, such as the
pyridinic nitrogen ofthe nicotinium ion ( 1) or the carbonyl oxygen of
ACh 2, is drastically affected by the nature of the counteranion Aꢀ.
Among the 384 entries containing the picrate fragment, we found 157
crystal structures with an ammonium group. We then arranged the am-
monium substructures according to the degree ofnitrogen substitution.
At this stage, we looked for N+···Oꢀ contacts as a function of the picrate
site ofassociation and the degree ofammonium nitrogen substitution. In
the first step of this work, an intermolecular contact was defined as in
the IsoStar library,[24] that is, as any contact shorter than V+0.5 Š, where
The association equilibrium constants of a reference donor 4-fluorophe-
nol (pFP) on the C+Picꢀ salts and on some neutral carbonyl bases were
measured by FTIR spectroscopy using the model system selected to
[68, 69]
define the general HB basicity scale pKHB
[Equations (10) and (11)].
V is the sum ofthe Bondi van der Waals radii. In the present case, this
[25]
criterion corresponds to a distance of3.6 Š.
In a second step this crite-
4-FC6H4OH þ CþPicꢀ Ð 4-FC6H4OH Á Á Á CþPicꢀ
ð10Þ
ð11Þ
rion was reduced to V (3.1 Š) in order to investigate the shortest con-
tacts.
½4-FC6H4OH Á Á Á CþPicꢀꢃ
Kcð25 ꢂCÞ ¼
½CþPicꢀꢃ½4-FC6H4OHꢃ
Theoretical calculations: All calculations were performed at the B3LYP/
[60]
6-31+G** level oftheory using the Gaussian 03
package supported on
The concentrations [4-FC6H4OH···C+Picꢀ], [C+Picꢀ] and [4-FC6H4OH]
in Equation (11) were calculated from the equilibrium infrared absorp-
tion intensities and from the weighted initial amounts of the donor and
acceptor. For the picrate salts, which are sparingly soluble in carbon tet-
rachloride, all measurements were carried out in dichloromethane and
the experimental data converted to the reference pKHB scale (in CCl4)
using calibrated family-dependent LFERs.[16] Owing to the very low
transparency ofdichloromethane in the OH region and to the low solu-
bility ofsome salts, the precision ofthe equilibrium constants, carried out
in 10, 2 and 1 mm quartz or fluorine cells, is estimated to be about 15–
the CCIPL, CINES and IDRIS supercomputers.
The starting geometries for the ion pairs must take into account not only
the conformational flexibility of the cation, but also the relative positions
of the anion and cation. Starting from the fully optimised monomer struc-
tures (B3LYP/6-31+G**), the PES ofthe ion-pair interaction can be
thoroughly scanned in the rigid-body approximation using an intermolec-
ular model potential based on an accurate distributed multipole analysis
(DMA) for the dominant electrostatic term. The interaction geometry
was optimised starting from 400 points randomly generated at both short
and long distances. This approach led to a limited number ofcomplexes
which were then fully optimised at the B3LYP/6-31+G** level oftheory.
20%. The frequency shifts are accurate to 5–10 cmꢀ1
.
1508
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1499 – 1510

Hydrogen-Bond Interactions ofNicotine and Acetylcholine Salts
FULL PAPER
The decrease in the OH absorption ofthe donor induced by the introduc-
tion ofthe salt into solution is a consequence ofthe association ofpFP
with all the available accepting sites. Therefore, the calculated equilibri-
[19] K. Sawada, E. Takahashi, T. Horie, K. Satoh, Monatsh. Chem. 2001,
132, 1439–1450.
[20] I. D. Kuntz, Jr., R. P. Taylor, J. Phys. Chem. 1970, 74, 4573–4577.
[21] F. H. Allen, C. A. Baalham, J. P. M. Lommerse, P. R. Raithby, E.
Sparr, Acta Crystallogr., Sect. B 1997, 53, 1017–1024.
[22] J. Harrowfield, J. Chem. Soc., Dalton Trans. 1996, 3165–3171.
[23] A. Szumna, J. Jurczak, Z. Urbanczyk-Lipkowska, J. Mol. Struct.
2000, 526, 165–175.
[24] I. J. Bruno, J. C. Cole, J. P. M. Lommerse, R. S. Rowland, R. Taylor,
M. L. Verdonk, J. Comput. Aided Mol. Des. 1997, 11, 525–537.
[25] A. Bondi, J. Phys. Chem. 1964, 68, 441–451.
[26] M. Evain, F. X. Felpin, C. Laurence, J. Lebreton, J. Y. Le Questel, Z.
Kristallogr. 2003, 218, 753–760.
[39]
um constant Kt is the sum ofthe individual 1:1 equilibrium constants
for the C+ cation and Picꢀ anion [Equations (12)–(14)].
4-FC6H4OH þ CþPicꢀ Ð 4-FC6H4OH Á Á Á CþPicꢀ
4-FC6H4OH þ CþPicꢀ Ð 4-FC6H4OH Á Á Á PicꢀCþ
KCþ
ð12Þ
ð13Þ
KPicꢀ
½4-FC6H4OH Á Á Á CþPicꢀꢃ þ ½4-FC6H4OH Á Á Á PicꢀCþꢃ
Kt ¼
¼ KCþ þ KPicꢀ
ð14Þ
½4-FC6H4OHꢃ½CþPicꢀꢃ
[27] T. P. Pitner, W. B. Edwards III, R. L. Bassfield, J. F. Whidby, J. Am.
Chem. Soc. 1978, 100, 246–251.
[28] J. F. Whidby, J. I. Seeman, J. Org. Chem. 1976, 41, 1585–1590.
[29] T. Takeshima, R. Fukumoto, T. Egawa, S. Konaka, J. Phys. Chem. A
2002, 106, 8734–8740.
[30] D. E. Elmore, D. A. Dougherty, J. Org. Chem. 2000, 65, 742–747.
[31] J. Graton, M. Berthelot, J.-F. Gal, S. Girard, C. Laurence, J. Lebre-
ton, J.-Y. Le Questel, P.-C. Maria, P. Naus, J. Am. Chem. Soc. 2002,
124, 10552–10562.
[32] R. B. Barlow, O. Johnson, Br. J. Pharmacol. 1989, 98, 799–808.
[33] K. Frydenvang, L. Gronborg, B. Jensen, Acta Crystallogr. Sect. B
1988, 44, 841–845.
In a few favourable systems in which dibasic species show intense charac-
teristic absorptions shifted by association, the two individual equilibrium
constants can be obtained from additional infrared experiments.[39] For
instance, the individual association constant for the Nsp2 site ofthe nico-
tinium cation 1 has been obtained separately from the pyridine n˜1 vibra-
[16]
tion at 1026 cmꢀ1
.
In most cases, however, the only direct quantitative
information is the total equilibrium constant. Various indirect methods
for their separation into individual association constants have been used
that involve 1) experimental Dn˜(OH) frequency shifts,[41,42] 2) substituent
effects,[51,70] 3) molecular fragment models[71] or 4) theoretical descrip-
tors.[42,72,73]
[34] R. S. Rowland, R. Taylor, J. Phys. Chem. 1996, 100, 7384–7391.
[35] K. Sawada, F. Chigira, K. Satoh, T. Komatsuzaki, J. Chem. Soc.,
Faraday Trans. 1997, 93, 1903–1908.
[36] L. Troxler, J. M. Harrowfield, G. Wipff, J. Phys. Chem. A 1998, 102,
6821–6830.
Acknowledgements
The authors gratefully acknowledge AurØlien Planchat for his active par-
ticipation in the experimental work, Prof. Anthony J. Stone for providing
the Orient program and the Centre de Calcul Intensifdes Pays de la
Loire (CCIPL), the Centre Informatique National de lꢀEnseignement Su-
pØrieur (CINES) and the Institut Des Ressources en Informatique Scien-
tifique (IDRIS) for computer time.
[37] C. E. Cannizzaro, K. N. Houk, J. Am. Chem. Soc. 2002, 124, 7163–
7169.
[38] O. Lamarche, J. A. Platts, Chem. Eur. J. 2002, 8, 457–466.
[39] V. Arnaud, J.-Y. Le Questel, M. MathØ-Allainmat, J. Lebreton, M.
Berthelot, J. Phys. Chem. A 2004, 108, 10740–10748.
[40] F. Besseau, C. Laurence, M. Berthelot, J. Chem. Soc., Perkin Trans.
2 1994, 485–489.
[41] M. Berthelot, C. Laurence, M. Safar, F. Besseau, J. Chem. Soc.
Perkin Trans. 2 1998, 283–290.
[42] E. Marquis, J. Graton, M. Berthelot, A. Planchat, C. Laurence, Can.
J. Chem. 2004, 82, 1413–1422.
[43] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165–195.
[44] C. Laurence, M. Berthelot, M. Lucon, D. G. Morris, J. Chem. Soc.,
Perkin Trans. 2 1994, 491–493.
[1] J. A. Dani, M. De Biasi, Y. Liang, J. Peterson, L. Zhang, T. Zhang,
F. M. Zhou, Bioorg. Med. Chem. Lett. 2004, 14, 1837–1839.
[2] P. Newhouse, A. Singh, A. Potter, Curr. Top. Med. Chem. 2004, 4,
267–282.
[3] R. C. Hogg, M. Raggenbass, D. Bertrand, Rev. Physiol. Biochem.
Pharmacol. 2003, 147, 1–46.
[4] T. Grutter, J. P. Changeux, Trends Biochem. Sci. 2001, 26, 459–463.
[5] A. Miyazawa, Y. Fujiyoshi, N. Unwin, Nature 2003, 423, 949–955.
[6] K. Brejc, W. J. van Dijk, R. V. Klaassen, M. Schuurmans, J. van der
Oost, A. B. Smit, L. K. Sixma, Nature 2001, 411, 269–276.
[7] P. H. N. Celie, S. E. Van Rossum-Fikkert, W. J. Van Dijk, K. Brejc,
A. B. Smit, T. K. Sixma, Neuron 2004, 41, 907–914.
[8] W. H. Beers, E. Reich, Nature 1970, 228, 917–922.
[9] R. A. Glennon, M. Dukat, Pharm. Acta Helv. 2000, 74, 103–114.
[10] R. A. Glennon, M. Dukat, L. Liao, Curr. Top. Med. Chem. 2004, 4,
631–644.
[11] J. E. Tonder, P. H. Olesen, J. B. Hansen, M. Begtrup, I. Pettersson, J.
Comput. Aided Mol. Des. 2001, 15, 247–258.
[12] A. L. Cashin, E. J. Petersson, H. A. Lester, D. A. Dougherty, J. Am.
Chem. Soc. 2005, 127, 350–356.
[13] X. Huang, F. Zheng, P. A. Crooks, L. P. Dwoskin, C.-G. Zhan, J.
Am. Chem. Soc. 2005, 127, 14401–14414.
[14] J. Graton, M. Berthelot, J. F. Gal, C. Laurence, J. Lebreton, J. Y. Le
Questel, P. C. Maria, R. Robins, J. Org. Chem. 2003, 68, 8208–8221.
[15] J. Graton, T. Van Mourik, S. L. Price, J. Am. Chem. Soc. 2003, 125,
5988–5997.
[45] A. Chardin, C. Laurence, M. Berthelot, D. G. Morris, Bull. Soc.
Chim. Fr. 1996, 133, 389–393.
[46] S. C. Mohr, W. D. Wilk, G. M. Barrow, J. Am. Chem. Soc. 1965, 87,
3048–3052.
[47] H. W. Aitken, W. R. Gilkerson, J. Am. Chem. Soc. 1973, 95, 8551–
8559.
[48] C. R. Witschonke, C. A. Kraus, J. Am. Chem. Soc. 1947, 69, 2472–
2481.
[49] L. J. Bellamy, The Infrared Spectra of Complex Molecules, Vol. 2,
Chapman and Hall, London, 1980.
[50] T. Gramstad, W. J. Fuglevik, Acta Chem. Scand. 1962, 16, 1369–
1377.
[51] J. Graton, M. Berthelot, J.-F. Gal, C. Laurence, J. Lebreton, J.-Y. Le
Questel, P.-C. Maria, R. Robins, J. Org. Chem. 2003, 68, 8208–8221.
[52] N. Le Nov›re, T. Grutter, J.-P. Changeux, Proc. Natl. Acad. Sci.
U.S.A. 2002, 99, 3210–3215.
[53] M. Schapira, R. Abagyan, M. Totrov, BMC Struct. Biol. 2002, 2, 1–
8.
[54] M. H. Abraham, P. L. Grellier, D. V. Prior, R. W. Taft, J. J. Morris,
P. J. Taylor, C. Laurence, M. Berthelot, R. M. Doherty, M. J. Kamlet,
J.-L. M. Abboud, K. Sraidi, G. GuihØneuf, J. Am. Chem. Soc. 1988,
110, 8534–8536.
[55] M. H. Abraham, P. L. Grellier, D. V. Prior, P. P. Duce, J. J. Morris,
P. J. Taylor, J. Chem. Soc., Perkin Trans. 2 1989, 699–711.
[16] V. Arnaud, M. Berthelot, J.-Y. Le Questel, J. Phys. Chem. A 2005,
109, 3767–3770.
[17] S. Roelens, R. Torriti, J. Am. Chem. Soc. 1998, 120, 12443–12452.
[18] D. T. Copenhafer, C. A. Kraus, J. Am. Chem. Soc. 1951, 73, 4557–
4561.
Chem. Eur. J. 2007, 13, 1499 – 1510
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1509

J.-Y. Le Questel et al.
[56] M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G. L. Cascara-
no, L. De Caro, C. Giacovazzo, G. Polidori, R. Spagna, J. Appl.
Crystallogr. 2005, 38, 381–388.
[57] JANA2000, Structure Determination Software Programs, V. Petri-
cek, M. Dusek, Praha (Czech Republic), 2000.
[58] F. H. Allen, R. Taylor, Chem. Soc. Rev. 2004, 33, 463–475.
[59] I. J. Bruno, J. C. Cole, P. R. Edgington, M. Kessler, C. F. Macrae, P.
McCabe, J. Pearson, R. Taylor, Acta Crystallogr. Sect. B 2002, 58,
389–397.
[60] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomer-
y, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyen-
gar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
[61] T. van Mourik, S. L. Price, D. C. Clary, Faraday Discuss. 2001, 118,
95–108.
[62] T. van Mourik, Phys. Chem. Chem. Phys. 2001, 3, 2886–2892.
[63] T. van Mourik, Phys. Chem. Chem. Phys. 2004, 6, 2827–2837.
[64] A. J. Stone, Chem. Phys. Lett. 1981, 83, 233–239.
[65] A. J. Stone, M. Alderton, Mol. Phys. 1985, 56, 1047–1064.
[66] GDMA: a program for performing distributed multipole analysis of
wave functions calculated using the Gaussian program system, ver-
sion 1.0, A. J. Stone, University ofCambridge, Cambridge (UK),
1999.
[67] ORIENT, version 4.4, A. J. Stone, A. Dullweber, O. Engkvist, E.
Fraschini, M. P. Hodges, A. W. Meredith, P. L. A. Popelier, D. J.
Wales, University ofCambridge, Cambridge (UK), 2001.
[68] R. W. Taft, D. Gurka, L. Joris, P. von R. Schleyer, J. W. Rakshys, J.
Am. Chem. Soc. 1969, 91, 4801–4808.
[69] C. Laurence, M. Berthelot, Perspect. Drug Discovery Des. 2000, 18,
39–60.
[70] F. Besseau, M. Lucon, C. Laurence, M. Berthelot, J. Chem. Soc.,
Perkin Trans. 2 1998, 101–108.
[71] J.-Y. Le Questel, G. Boquet, M. Berthelot, C. Laurence, J. Phys.
Chem. B 2000, 104, 11816–11823.
[72] J.-Y. Le Questel, M. Berthelot, C. Laurence, J. Phys. Org. Chem.
2000, 13, 347–358.
[73] C. Ouvrard, M. Lucon, J. Graton, M. Berthelot, C. Laurence, J.
Phys. Org. Chem. 2004, 17, 56–64.
Received: June 9, 2006
Published online: November 14, 2006
1510
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 1499 – 1510