Substituent effects on aromatic stacking interactions

Article abstract of DOI:10.1039/b617576g

Synthetic supramolecular zipper complexes have been used to quantify substituent effects on the free energies of aromatic stacking interactions. The conformational properties of the complexes have been characterised using NMR spectroscopy in CDCl₃, and by comparison with the solid state structures of model compounds. The structural similarity of the complexes makes it possible to apply the double mutant cycle method to evaluate the magnitudes of 24 different aromatic stacking interactions. The major trends in the interaction energy can be rationalised using a simple model based on electrostatic interactions between the π-faces of the two aromatic rings. However, electrostatic interactions between the substituents of one ring and the π-face of the other make an additional contribution, due to the slight offset in the stacking geometry. This property makes aromatic stacking interactions particularly sensitive to changes in orientation as well as the nature and location of substituents. This journal is The Royal Society of Chemistry.

Full text of DOI:10.1039/b617576g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Substituent effects on aromatic stacking interactions†
Scott L. Cockroft,a Julie Perkins,a Cristiano Zonta,a Harry Adams,a Sharon E. Spey,a Caroline M. R. Low,b
Jeremy G. Vinter,b Kevin R. Lawson,c Christopher J. Urchc and Christopher A. Hunter*a
Received 1st December 2006, Accepted 17th January 2007
First published as an Advance Article on the web 7th March 2007
DOI: 10.1039/b617576g
Synthetic supramolecular zipper complexes have been used to quantify substituent effects on the free
energies of aromatic stacking interactions. The conformational properties of the complexes have been
characterised using NMR spectroscopy in CDCl3, and by comparison with the solid state structures of
model compounds. The structural similarity of the complexes makes it possible to apply the double
mutant cycle method to evaluate the magnitudes of 24 different aromatic stacking interactions. The
major trends in the interaction energy can be rationalised using a simple model based on electrostatic
interactions between the p-faces of the two aromatic rings. However, electrostatic interactions between
the substituents of one ring and the p-face of the other make an additional contribution, due to the
slight offset in the stacking geometry. This property makes aromatic stacking interactions particularly
sensitive to changes in orientation as well as the nature and location of substituents.
conformationally well-defined system to measure intermolecular
interactions, using hydrogen bonding to force two aromatic rings
Introduction
For decades, researchers from across the chemical sciences
into a stacked geometry.
have used ‘aromatic interactions’ to rationalise their observa-
The double-mutant cycle is a robust thermodynamic tool that
tions.1Aromatic interactions have been exploited in template-
has been employed by numerous researchers to isolate individual
directed synthesis to prepare topologically complex molecules
weak non-covalent interactions from a noisy background of
multiple secondary interactions.41,42 We have previously used
supramolecular zipper complexes in conjunction with the double-
mutant cycle approach to quantify a wide range of aromatic
interactions, including the effects of substituents on edge-to-face
aromatic interactions and cation–p interactions.43–53 The method
was successfully adapted to quantify a number of aromatic
stacking interactions (see Fig. 1).54,55 Since those studies our
compound library has been significantly expanded. Here we
present a thorough conformational analysis of the complexes and
a complete analysis of the entire data set. In the light of the
work presented here and the studies of other investigators, we
and to control the enantioselectivity of reactions in asymmetric
syntheses.2–5 The arrangements of molecules in solids, liquid
crystals and solution are known to be influenced by aromatic
stacking interactions.6–8 In biological systems, aromatic stacking
interactions have been identified as key factors in determining the
structural and molecular recognition properties of nucleic acids,
peptides and proteins.9–11 For example, X-ray crystal structures
have identified stacked aromatic contacts between drug molecules
and the aromatic side-chains of proteins.12–14 These observations
have motivated interest in the prediction of stacking interaction
energies for use in quantitative structure activity relationships.15
Whilst ab initio calculations16–22 and qualitative models23 that
describe aromatic stacking interactions exist, experimental data
are required to test these theories. Aromatic interactions have
been investigated in a range of model systems, which have been
extensively reviewed,24–27 and new studies continue to emerge.
One approach involves the study of intramolecular aromatic
stacking interactions using rotameric or conformationally flexible
molecules.28–37 Other studies have taken a supramolecular host–
guest approach to the assessment of intermolecular aromatic
interactions.1,38–40 In the work presented here, we exploit a
aCentre for Chemical Biology, Department of Chemistry, Krebs Institute
for Biomolecular Science, University of Sheffield, Sheffield, UK S3 7HF.
E-mail: c.hunter@sheffield.ac.uk; Fax: (+44)114 2738673; Tel: (+44)114
2229476
bJames Black Foundation, 68 Half Moon Road, London, UK SE24 9JE
cSyngenta, Jealott’s Hill International Research Centre, Bracknell, Berk-
shire, UK RG42 6EY
† Electronic supplementary information (ESI) available: detailed synthetic
procedures and characterisation data for new compounds; crystallographic
data (CCDC reference numbers 238547 and 615388–615406). See DOI:
10.1039/b617576g
Fig. 1 A chemical double-mutant cycle for measuring the free energy
contribution of the aromatic stacking interaction in complex A.
1062 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
draw new conclusions concerning the main factors influencing the
magnitudes of aromatic stacking interactions.
induced changes in chemical shifts, ROESY experiments and X-
ray crystallography‡ on model compounds. These conformational
insights have been used to establish which of the many possible
double-mutant cycles can be reliably used to systematically survey
the effects of substituents on aromatic stacking interactions.
Design and synthesis
Fig. 1 shows an example of a double-mutant cycle used in
this study. The stacking interaction highlighted in complex A is
measured by chemical mutations that remove it. A single mutation
(e.g. comparing the stabilities of complexes A and B) is not
sufficient, because this has secondary effects, such as changing
the H-bond strength. The double-mutant complex D quantifies
these secondary interactions, and the free energy difference of any
two parallel mutations in Fig. 1 allows the interaction of interest
to be dissected out of the complicated array of weak interactions
present in complex A.
The compounds used in this work were prepared as outlined
in Fig. 2 and 3. Isophthaloyl derivatives 8a–e, 9 and 10 were
synthesised from the appropriate substituted anilines, which are
available commercially or via relatively simple syntheses (Fig. 2;
ESI†). Bisaniline derivatives 15a–c, 15e–g and 20d–f were prepared
according to the routes shown in Fig. 3.
Solid state conformational studies
Isophthaloyl derivatives 8d, 8e, 9 and bisaniline derivatives similar
to those used in the 1H NMR titrations of the present study have
been successfully crystallised. The conformational attributes of
the bisaniline derivatives in the solid state have been discussed
previously.54 The crystal structures of the individual components
of a complex are of limited utility, since they provide little informa-
tion regarding the geometry of the aromatic interactions present in
the supramolecular zipper complexes. However, the X-ray crystal
structures of simple model compounds have been able to provide a
useful indicator of the likely geometry of the edge-to-face aromatic
interactions in zipper complexes in solution (Fig. 4a).43,46,49 The
X-ray crystal structures of the model compounds for the edge-
to-face zipper complexes reveal H-bonded chains with the head-
to-tail packing arrangement shown in Fig. 4a. Based on the
1H NMR titrations were used to measure association con-
stants for 34 different combinations of the isophthaloyl and
bisaniline compound libraries. The structural properties of the
complexes have been analysed using 1H NMR complexation-
‡ CCDC reference numbers 238547 and 615388–615406. For crystallo-
graphic data in CIF or other electronic format see DOI: 10.1039/b617576g
Fig. 2 Synthesis and proton labelling scheme for isophthaloyl derivatives 8a (Y = NMe2), 8b (Y = H), 8c (Y = OMe), 8d (Y = Cl), 8e (Y = NO2), 9 and
10.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1063
©

View Article Online
Fig. 3 Synthesis and proton labelling scheme for the 15x and 20x series of bisaniline derivatives.
success of this approach, new model compounds were synthesised
and crystallised (Fig. 4b and 5). Compounds 21a–24f held the
promise of revealing the geometry of the aromatic stacking
interactions in the ‘full-sized’ zipper complexes (the structures of
compounds 21a, 21b and 23d have been previously deposited in
the CCDC).54,56,57
Each of the model compounds 21a–24f were found to form H-
bonded chains in the solid state, as expected. However, none of
1064 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
polymorphic forms (Fig. 6). Based on the propensity of pentaflu-
orophenyl groups to form stacks with electron-rich aromatic
systems in the solid state,6,58–61 it was anticipated that compound
21f would crystallise in H-bonded chains with head-to-tail stacks
in line with the intended design. However, the molecules in the a-
polymorph were found to form twisted H-bonded chains accom-
modating close phenyl–pentafluorophenyl stacking interactions
between adjacent H-bonded chains (Fig. 6a). In contrast, the b-
polymorph forms a parallel head-to-head arrangement containing
offset phenyl-phenyl, and pentafluorophenyl–pentafluorophenyl
stacks within the H-bonded chain (Fig. 6b).
Whilst the design of the model compounds containing edge-
to-face aromatic interactions was successful, the new stacking
analogues 21a–24f fall short of the mark. Fig. 4b shows that
formation of a stacking interaction on one side of the H-bonded
amide chain would create a cavity between the aromatic rings on
the opposite side of the dimer. The model compounds shown
in Fig. 4a do not suffer from this problem, because edge-to-
face aromatic contacts occur on both sides of the H-bonded
amide chain, with the 2,6-isopropylated aniline rings providing an
additional conformational constraint.46 In contrast, the aromatic
groups in stacking compounds 21a–24f are free to rotate about
the H-bonded amide chain. Head-to-head and twisted amide H-
bonded chains form in preference to the intended head-to-tail
orientation because of the free energy penalty associated with the
loss of dispersion interactions in a poorly packed, cavity-filled
crystal.
Fig. 4 a) Solid state structures of simple model compounds (right)
have previously been used to infer the geometry of edge-to-face aromatic
interactions in zipper complexes in solution (left). b) Complex proposed
for the measurement of aromatic stacking interactions in solution (left)
and the corresponding model compound for X-ray crystallisation studies
(right).
The range of structures obtained for compounds 21a–24f,
and the identification of polymorphic forms of 21f are evidence
that there is not a single well-defined pathway involved in the
crystallisation of this series of compounds. Nevertheless, useful
conformational information can be gained from the X-ray crystal
structures of these molecules. Fig. 7 shows that anthracene,
acridine, 2,6-dimethylphenyl, or 2,6-fluorophenyl aromatic groups
provide sufficient steric influence to twist the aromatic rings out
of the plane of the amide bond, which is an essential feature of
the supramolecular designs used in this investigation. The aniline
rings in compounds 21a–24f are tilted out of the plane of the amide
bond by 70◦ 10◦ (Fig. 7d), and the aromatic rings on the other
side of the amide are oriented at 65◦ 15◦ to the plane of the amide
(Fig. 7b). Furthermore, the presence of intermolecular aromatic
stacks in the crystal structures demonstrate that close stacking of
the terminal aromatic rings in the supramolecular complexes is not
sterically compromised by the 2,6-dimethyl and fluorine aromatic
substituents.
Fig. 5 Model compounds designed to probe the geometry of aromatic
stacking interactions in the solid-state. X group labels are shown in Fig. 3.
these compounds crystallised in the desired head-to-tail conforma-
tion seen for the edge-to-face model compounds (Fig. 4a). Instead,
the stacking compounds crystallised in head-to-head or twisted
H-bonded chains. Both of these packing modes were present in
the structures of compound 21f, which was crystallised in two
Fig. 6 Interaction motifs in the twisted a-polymorph a) and head-to-head b-polymorph b) of 21f.
The Royal Society of Chemistry 2007 Org. Biomol. Chem., 2007, 5, 1062–1080 | 1065
This journal is
©

View Article Online
Fig. 7 Overlays of the structures of model compounds in the solid state. a) View of model compounds 21a–24f orthogonal to the amide group. b) View
of model compounds 21a–24f along the axis indicated in a). c) View of model compounds 21a–24f in the plane of the amide bond. d) View of model
compounds 21a–24f along the axis indicated in a). e) View of N-methylamide model compounds 26a, 26e and 26f orthogonal to the amide group. f) View
of N-methylamide model compounds 26a, 26e and 26f in the plane of the amide bond.
Self-association studies of the isophthaloyl derivatives
8a–10
The solubilities of the isophthaloyl derivatives (Fig. 2) in CDCl3
are in the range 0.2–4 mM. Attempts to determine accurate self-
association constants of the isophthaloyl derivatives using 1H
NMR dilution experiments were hampered by small observed
changes in chemical shift and the limited solubility of these
compounds. The dimerisation constants are less than 5 M−1, and
since the isophthaloyl derivatives were used as the host at mM
1
concentrations in the H NMR titrations, this has a negligible
effect on the binding experiments.48
Self-association studies of the 15x and 20x bisaniline
derivatives
The nitropyrrole moiety in the bisaniline derivatives (Fig. 3 and
Fig. 8 Conformations of nitropyrrole amides. a) The major conformer
of the nitropyrrole moiety in the unbound state. b) 1H NMR experiments
indicate a 10% population of the cis-amide is in slow exchange with a). The
geometry of the cis-conformer explains the large upfield change in chemical
shift of p3. c) In the bound state, the pyrrole NH (p1) and the adjacent
amide NH (n1) signals experience large downfield shifts consistent with
the formation of two hydrogen bonds.
8) is known to promote self-association of these compounds
in CDCl3.54 1H NMR dilution experiments were performed for
each of the bisaniline derivatives, and the data were fit to a
dimerisation model. The limiting dimerisation-induced changes in
chemical shift and dimerisation constants for each of the bisaniline
derivatives are given in Table 1. The large downfield shifts on the
nitropyrrole NH (p1) and the adjacent amide NH (n1) suggest that
both of these protons are acting as H-bond donors (Fig. 8c). Each
of the bisaniline derivatives also have minor conformers, as seen
the free energy effects of dimerisation cancel in the double-mutant
cycle,54 it is important to account for the effects of dimerisation to
obtain accurate Dd values for the bisaniline guest from 1H NMR
titration data (see Experimental section).
1
before in the H NMR spectra of 15b, 15f and 15g in CDCl3.54
The minor conformer (10%) is in slow exchange with the major
conformer on the NMR timescale. The minor conformer differs
in the chemical shifts of p1 and p3: 9.3 and 5.4 ppm respectively,
compared to 9.6 and 7.3 ppm in the major conformer. These
differences in chemical shift are consistent with the cis-amide
shown in Fig. 8b. Although it has previously been shown that
Binding studies using the 15x bisaniline derivatives
The 15x series of bisaniline derivatives were titrated into the isoph-
1
thaloyl derivatives 8a, 8b, 8c, 8e and 10. H NMR titration data
1066 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Table 1 Dimerisation constants Kdim (M−1) and limiting dimerisation-induced changes in 1H chemical shift (ppm) for the bisaniline derivative series 15x
and 30x in CDCl3 at 298 Ka
Signal
Compound
Kdim
p1
p2
n1
n2
b1
b2
b3
b4
a1
a2
a3
a4
a5
m1/m2
15a
15b
15c
15f
15g
20d
20e
20f
9
13
14
11
20
33
30
16
2
2
2
2
2
5
4
6
2.4
2.2
2.5
2.2
2.5
2.3
2.6
2.7
−0.7 2.3 0.7
−0.8 2.0 0.6
−0.7 2.1 1.9
−0.5 2.7 1.2
−0.7 2.0 1.1
−0.3 −0.2
0.0 −0.1 −0.3 −0.3 −0.2
—
—
—
—
—
—
−0.2
−0.1
−0.1
−0.2
−0.4 −0.3 −0.1 −0.2 −0.3 −0.3 −0.3 −0.2 −0.3
−0.3 −0.2 −0.1 −0.1 −0.6 −0.5 −0.4 −0.4
—
—
—
—
—
—
−0.2 −0.1 −0.1 −0.1
−0.3 −0.2 −0.1 −0.2 −0.2
−0.3 −0.2 0.0 −0.1
−0.5 −0.3 −0.1 −0.1 −0.3 −0.2
−0.3 −0.2 −0.2 −0.1
—
—
—
—
—
—
—
—
—
—
−0.7 2.2
−0.7 2.5
−0.6 2.4
—
—
—
—
−0.2 −0.2
—
—
—
—
a See Fig. 3 for proton labelling scheme. p3 shifts could not be accurately determined due to overlap with the residual CHCl3 peak throughout most of
the 1H NMR experiments. There were no significant changes in the chemical shift of the signals for protons on the piperidine ring or solubilising group.
Dilution experiments were repeated at least twice, and Kdim is the weighted mean based on the observed changes in chemical shift for all signals monitored.
The error is twice the standard error.
were fit to a 1 : 1 binding isotherm that allowed for dimerisation
ROESY studies of complexation
of the guest. The bisaniline derivatives contain a large solubilising
1
Valid application of the double-mutant cycle methodology re-
group, intended to facilitate their use as guest molecules in H
quires that the core structure of each complex is conserved. It
NMR titrations. Despite the solubilising group, the nitro derivative
1
is therefore important to check the geometries of the complexes
before attempting to determine (and interpret) interaction free
15e was poorly soluble in CDCl3 (∼3 mM). Analysis of host H
NMR chemical shifts from titration experiments with 15e gave
energies. Intermolecular NOEs from two-dimensional ROESY
less than 15% coverage of the binding isotherm, preventing the
experiments provide limited, but useful information about the
simultaneous determination of reliable 1 : 1 host–guest association
structures of the complexes (Table 2 and Fig. 9). The NOEs b1–d
constants and Dd values. Accordingly, further studies with 15e were
and b4–d are consistently observed indicating that the isophthaloyl
group is docked into the bisaniline pocket in the core of the
complex. Other NOEs confirm that the nitropyrrole group sits over
abandoned. The pentafluorophenyl compound 15f also suffered
from low solubility in CDCl3 (∼10 mM); in the worst case 30% of
compound 10 was bound at the end of the 10 : 15f titration,
the terminal aromatic ring in at least three of the complexes (p2–
although the favourable binding of 8a with 15f allowed 70%
me), and that on the opposite side of the complex the anthracene
coverage of the 8a : 15f binding isotherm. All of the other
group in bisaniline derivative 15b is in close proximity to the
compounds in the 15x series had sufficient solubility in CDCl3
terminal aniline rings of the isophthaloyl derivatives (a1–me and
(∼35–40 mM) to allow the accurate determination of association
a2–me).
constants and Dd values.
Table 2 Intermolecular NOEs observed in two-dimensional ROESY experimentsa
Bisaniline compound
Isophthaloyl compound
15a X = o-Me2-phenyl
15b X = anthracene
15c X = acridine
15f X = F5-phenyl
15g X = Me
8a
b4–d
b1–d
b1–d
b4–d
b1–d
b4–d
b4–d
Y = NMe2
p2–me
a1–me
a2–me
8b
Y = H
b1–d
b4–d
b1–d b
b4–d
p2–me
a1–me
a2–me
b1–d
b4–d
b1–db
b4–d
b1–d b
8c
b1–d
b4–d
p2–me
n.d.c
n.d.
b1–d
b4–d
n.d.
Y = OMe
10
b1–d
b4–d
b1–db
b4–d
b4–d
b1–d b
b4–d
b1–db
b4–d
a1–nh
a See Fig. 2 and 3 for proton labelling schemes. Experiments were carried out on 1 : 1 mixtures of the two components in CDCl3 at the maximum
concentration possible 0.2–5 mM. Intermolecular NOEs of complexes with compounds 8d, 8e and 9 were not detectable because of their low solubilities
in CDCl3. b Previously reported but included for comparison purposes. c Not determined.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1067
©

View Article Online
Not all of the complexes are so well behaved. Attention is
immediately drawn to the s signal of the 8b·15c complex that
experiences a surprisingly large upfield shift. In addition to this
unusual chemical shift behaviour, the titration data did not fit well
to a 1 : 1 complexation model but was well described by a 2 : 1
(guest–host) model which included dimerisation of the bisaniline
guest. Molecular modelling using XED62 provided a plausible
explanation for this anomalous behaviour. The 1 : 1 8b·15c
complex is the ideal geometry to bind a second 15c molecule.
The structure of the XED-minimised 2 : 1 complex is shown in
Fig. 10 and accounts for the strange patterns of chemical shifts
that we observe for this complex. The acridine nitrogen of the
first 15c molecule is capable of accepting two H-bonds from the
nitropyrrole group of the second 15c molecule. The s proton is held
directly over the acridine ring of the second 15c molecule, and this
is the reason for the large upfield shift of s. Similarly, the aa and
y protons form a favourable edge–face interaction with one of
the 15c aniline rings, explaining their large upfield complexation-
induced changes in chemical shift. The other signals are unaffected
relative to the other complexes, since their environment is not
altered by formation of the 2 : 1 complex. The final piece of
support for the assembly presented in Fig. 10 comes from the
binding properties of 15c with the other isophthaloyl derivatives.
Both 8a (Y = NMe2) and 8e (Y = NO2) bind to 15c following a 1 :
1 binding isotherm (rather than 2 : 1), with very similar Dd values
to those obtained with the anthracene bisaniline derivative 15b.
This is because the p-nitro and p-dimethylamino Y-substituents of
compounds 8a and 8e are too large to be accommodated in the
same the position as the y proton in the bisaniline pocket of the
second 15c molecule in the 2 : 1 complex.
Fig. 9 Intermolecular NOEs observed for the complexes of the 15x series
of bisaniline derivatives with compounds 8a–c and 10.
Complexation-induced changes in chemical shift of the
isophthaloyl hosts 8x, 9 and 10
Complexation-induced changes in chemical shift (Dd) can be
1
determined from H NMR titrations and are a particularly rich
source of structural information. Dd Values for the hosts are easily
obtained from the direct extrapolation of 1H NMR binding curves
to the 100% bound state. Table 3 lists the Dd values for all of the
isophthaloyl hosts 8a–8e, 9 and 10 in each titration experiment.
There are consistent patterns in the changes in chemical shift.
The large upfield shifts of the d and t signals indicate that these
protons are docked into the bisaniline aromatic cleft. The large
downfield shifts on the amide nh show that it is involved in
hydrogen bonding. The signals on the terminal aromatic rings (me,
aa) experience moderate upfield shift changes that are consistent
with the shielding effect of the aromatic stack that is formed upon
complexation. The magnitude of these changes is generally larger
in the anthracene (15b) and acridine (15c) complexes, where there
are extended aromatic surfaces with larger ring currents, and lower
in the single mutant (15g) complexes, where the aromatic group is
replaced by a methyl group.
The s proton in the isophthaloyl hexyl mutant 10·15c complex
also experiences an upfield complexation-induced change in
chemical shift. This indicates that the complex may be similar
to the 8b·15c complex discussed above. Accordingly, the acri-
dine complexes will not be used to construct double-mutant
cycles.
Fig. 10 Energy minimised structure showing how the 8b·15c complex (blue H-bonds) is able to bind a second molecule of 15c (red H-bonds). This
structure is consistent with the unusual complexation-induced chemical shift changes of the labelled protons. The DDd values are the additional changes
in chemical shift for this complex compared to the simple 1 : 1 complexes. The solubilising group and non-polar protons of 15c are omitted for clarity.
1068 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Table 3 Limiting complexation-induced changes in 1H chemical shift (Dd in ppm) from NMR titrations in CDCl3 at 298 K for isophthaloyl host series
8a–e, 9 and 10
Signal
Guest
Aromatic group
Host
Y
d
t
nh
aa
me
s
y
Complex A
15a
o-Me2-phenyl
anthracene
acridine
8a
8a
8a
8a
8a
8a
8a
8b
8b
8b
8b
8b
8b
8b
8c
8c
8c
8c
8d
8d
8d
8e
8e
8e
8e
8e
8e
8e
9
NMe2
NMe2
NMe2
NMe2
NMe2
NMe2
NMe2
H
H
H
H
H
H
H
OMe
OMe
OMe
OMe
Cl
−0.34
−0.32
−0.41
−0.51
−0.49
−0.49
−0.57
−0.29
−0.34
−0.73c
−0.48
−0.44
−0.45
−0.52
−0.30
−0.51
−0.46
−0.53
−0.24
−0.38
−0.38
−0.06
−0.27
−0.18
−0.05
−0.28
−0.19
−0.19
−0.33
−0.31
−0.36
−1.07
−1.08
−1.09
−1.23
−1.35
−1.48
−1.47
−0.97
−1.04
−1.16c
−1.20
−1.22
−1.32
−1.34
−0.96
−1.14
−1.33
−1.39
−0.82
−1.14
−1.09
−0.49
−0.83
−0.83
−0.30
so
1.46
1.55
1.54
1.10
1.59
1.66
1.62
1.58
1.70
2.09c
1.31
1.78
1.82
1.74
1.48
1.51
1.67
1.76
1.66
1.81
1.81
1.70
1.67
1.61
1.31
1.80
1.76
1.68
1.76
2.09
1.87
−0.14
−0.37
−0.47
−0.12
−0.17
−0.23
−0.14
−0.19
−0.40
−0.73c
−0.15
−0.19
−0.21
−0.12
−0.18
−0.14
−0.20
−0.14
−0.39
−0.29
−0.32
−0.34
−0.61
−0.42
−0.30
−0.32
−0.31
−0.34
−0.22
so
−0.20
−0.38
−0.46
−0.19
−0.23
−0.24
−0.20
−0.25
−0.42
−0.42c
−0.24
−0.25
−0.27
−0.22
−0.25
−0.20
−0.26
−0.24
−0.29
−0.26
−0.30
−0.26
−0.46
−0.34
−0.24
−0.29
−0.26
−0.30
—
0.21
0.16
0.02
0.03
0.16
0.15
0.13
0.25
0.24
−0.02
−0.06
−0.11
0.01
−0.02
−0.02
0.03
sob
−0.19
−0.75c
−0.08
−0.12
−0.10
−0.04
−0.04
−0.01
−0.02
0.01
—
—
—
—
—
—
—
—
—
15b
15c
15f
F5-phenyl
20d
o-F2-phenyl
p-NO2-o-Me2-phenyl
F5-phenyl
o-Me2-phenyl
anthracene
acridine
20e
20f
15a
15b
15c
−1.12c
0.10
0.23
0.17
0.19
0.21
0.07
0.20
0.18
0.25
0.20
0.23
0.44
0.20
0.17
0.41
0.26
0.40
0.40
0.23
0.24
0.24
15f
F5-phenyl
20d
o-F2-phenyl
p-NO2-o-Me2-phenyl
F5-phenyl
o-Me2-phenyl
F5-phenyl
o-F2-phenyl
F5-phenyl
o-Me2-phenyl
o-F2-phenyl
F5-phenyl
o-Me2-phenyl
anthracene
acridine
20e
20f
15a
15f
20d
20f
15a
20d
Cl
Cl
20f
15a
NO2
NO2
NO2
NO2
NO2
NO2
NO2
—
15b
15c
15f
F5-phenyl
20d
o-F2-phenyl
p-NO2-o-Me2-phenyl
F5-phenyl
o-Me2-phenyl
o-F2-phenyl
F5-phenyl
20d
−0.62
−0.59
−0.74
−0.74
−0.82
20f
—
so
so
so
15a
20d
9
—
—
—
—
20f
9
−0.21
Complex B
15g
—
—
—
—
—
—
8a
8b
8c
8d
8e
9
NMe2
H
OMe
Cl
NO2
—
−0.48
−0.43
−0.43
−0.33
−0.15
−0.44
−1.23
−1.12
−1.02
−0.92
−0.53
−0.89
1.34
1.65
1.45
1.88
1.86
1.95
−0.06
−0.15
−0.12
−0.36
−0.33
so
−0.12
−0.18
−0.18
−0.31
−0.25
—
0.14
0.23
0.18
0.29
0.53
0.23
0.00
−0.09
−0.06
—
—
so
15g
15g
15g
15g
15g
Complex C
15a
o-Me2-phenyl
anthracene
acridine
10
10
10
10
10
10
10
—
—
—
—
—
—
—
−0.38
−0.42
−0.63
−0.49
−0.45
−0.41
−0.45
−1.00
−1.11
−1.06
−0.69
−1.00
−1.02
−0.91
1.07
1.02
0.89
0.58
1.21
1.22
1.22
—
—
—
—
—
—
—
−0.28
−0.41
−0.55
−0.27
−0.26
−0.27
−0.32
0.08
0.03
—
—
—
—
—
—
—
15b
15c
−0.24
15f
F5-phenyl
0.02
20d
o-F2-phenyl
p-NO2-o-Me2-phenyl
F5-phenyl
0.05
0.11
0.09
20e
20f
Complex D
15g
—
10
—
−0.43
−0.89
1.01
—
−0.21
0.06
—
a See Fig. 2 for proton labelling scheme. Titrations were performed at least twice. b so—Not determined due to signal overlap. c Data for the 8b·15c complex
were fit to a 2:1 model (guest:host) including dimerisation of the guest.
Experimental section). The Dd values for all of the bisaniline guests
in the complexes with isophthaloyl derivatives 8a, 8b, 8c and 10, are
given in Table 4. The intermolecular hydrogen bonds in the zipper
Complexation-induced changes in chemical shift of the
15x bisaniline guests
The qualitative interpretations of host Dd values presented above
can provide a lot of information about the structure of a complex,
but this is only half the story, as the complexation-induced
changes in chemical shift of the bisaniline guests have yet to be
discussed. Dd Values for the guest molecules can be determined
by separating the contributions of guest dimerisation and host–
guest complexation to the observed chemical shifts changes (see
complexes are a crucial part of the supramolecular design. They
have a central role in stabilising and controlling the geometry of the
complexes. The bisaniline derivatives in the 15x series contain three
substantial H-bond donors. The Dd values of these NH protons
enable any structural irregularities in the complexes to be easily
identified. The nitropyrrole p1 and n1 signals should have large
positive Dd values from the formation of H-bonds in the complex
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1069
©

View Article Online
and the n2 signal should have zero (or a small) change in chemical
shift, since it should not be involved in H-bonding. This was found
to be the case for all of the 15x complexes, except in two situations.
Firstly, the 10·15c complex, which has already been identified as
having conformational problems from the host Dd values, has a
larger than expected shift for n2, consistent with the structure of
the 2 : 1 complex shown in Fig. 10. The second case affects all
complexes containing 15f. Dd Values of n2 in complexes with 15f
range from +0.5 ppm in the complex with 8a, to +2.2 ppm in
the complex with 8c. Previous observation of this anomalous shift
in the 8b·15f complex was attributed to dimerisation.54 However,
dimerisation cannot be the cause, because it has been taken into
account in determination of the guest Dd values presented in
Table 4.
Comparing the Dd values of the 15f complexes with the others in
Tables 3 and 4 also exposes reduced complexation-induced shifts
for the nitropyrrole p1 and n1 guest signals. The shifts of the isoph-
thaloyl host signals nh and aa are also reduced in these complexes,
but the d and t signals remain relatively unaffected. Clearly, the n2
amide proton adjacent to the pentafluorophenyl group is acting as
a H-bond donor in the 15f complexes. The shift patterns described
above can be explained by the equilibria shown in Fig. 11. When
Y = NMe2 (8a) the equilibria lie to the left (conformers a and
c) as indicated by the smallest n2 Dd of the 15f complexes. In
conformers b and d on the right-hand side of the equilibria in
Fig. 11, the nh amide breaks its H-bond with the carbonyl adjacent
to the pentafluorophenyl ring. The n2 Dd values signify that the
intermolecular stacking interaction that we wish to measure is
not present to a significant extent in the 8x·15f complexes. The
physical basis for this conformational problem seems to lie with
the electron-withdrawing pentafluorophenyl group in 15f, which
makes the amide adjacent to the pentafluorophenyl ring a poor
H-bond acceptor, but a good H-bond donor. This problem should
affect all bisaniline derivatives containing electron-withdrawing
aromatic rings, but there is a simple solution. New bisaniline
derivatives 20d, 20e and 20f were synthesised by the route shown in
Fig. 3 to replace the offending amide proton with a methyl group,
blocking access to conformers b and d in Fig. 11.
spectra of model compound 26f, precursors 18d, 18e, 19d, 19e, and
the final bisaniline guest compounds 20d and 20f all showed the
presenceofboth cis- andtrans-amide conformers in slow exchange.
Integration of the N-methyl H NMR signals indicated a 30%
population of the unwanted cis-conformer in CDCl3 for all of
these fluoroaromatic compounds.
1
Each of the N-methylated bisaniline derivatives 20d, 20e and
20f were very soluble in CDCl3 compared to the non-methylated
equivalents, so it seems that the polar amide NH (n2) was a major
contributor to low solubility as postulated earlier. The improved
solubility of the 20x series of bisaniline derivatives allowed large
fractions of the binding isotherms to be covered (up to 90%)
which improved the accuracy of the binding constant and Dd
value determinations. NMR titration data for 20e were fit to a
1 : 1 model including dimerisation of the guest, as for the 15x
bisaniline derivatives. Limiting dimerisation-induced changes in
chemical shift for the 20x bisaniline derivatives are included in
Table 1. The cis–trans amide ratio of compounds 20d and 20f did
not change during the titrations, because the guest molecules were
present in large excess during most of the experiment, and the
rate of equilibration is on the time-scale of hours. Titration data
obtained using compounds 20d and 20f were therefore fit to a 1 : 1
model including dimerisation of the guest, with the concentration
of guest corrected to account for the proportion of inactive cis-
conformer (as determined by integration of the N-methyl m2/m2
+ signals). The trans N-methyl signals m2 of compounds 20d and
20f showed small positive changes in chemical shift, but the cis N-
methyl m2+ signals did not change during the titration, confirming
that the cis isomers do not bind to the isophthaloyl derivatives
(Table 4). Dd Values for the hosts (8a–e, 9 and 10) and the 20x
guests used in the NMR titrations are included in Tables 3 and 4,
alongside the values obtained for the 15x complexes.
NMR solution structure determination of complexes
So far, we have used Dd values to characterise the structures of
the complexes in a qualitative fashion. However, it is possible
to use Dd values in a quantitative manner to gain a clearer
view of the conformational ensemble of the zipper complexes.
We have developed a computational method for determining
the three-dimensional structures of intermolecular complexes
in solution using Dd values. This approach has been shown
to give high-resolution structural information that agrees well
with the corresponding X-ray crystal structures where they are
available.57,64–67 This method has been used to determine the
solution structures of nine representative complexes containing the
isophthaloyl derivatives 8a–c, and bisaniline derivatives 15a, 15b,
15g, 20d, 20e and 20f. The calculated Dd values of the optimised
structures shown in Fig. 12 are in excellent agreement with the
experimental values (Table 5). While the absolute magnitudes of
the Dd values vary from one complex to another, the patterns are
very similar and the corresponding structures of the complexes are
remarkably consistent. The geometry of the stacked region of the
complex appears to be particularly well defined and is unaffected
by the magnitude of the aromatic stacking interaction (the 8a·15a
complex contains the most repulsive stacking interaction in the
zipper complexes, and the 8a·20f complex the most attractive, see
later).
Conformational properties of the 20x N-methylamide
bisaniline derivatives
The corresponding model compounds for the N-methylamide
bisaniline derivatives 20e and 20f (26e and 26f) were synthesised
by the route shown in Fig. 5. The crystal structure of related
compound 26a has been previously determined and is included for
comparison in the structural overlays in Fig. 7e and f.56 In the solid
state, the aromatic rings in all three model compounds are twisted
90◦ out of the plane of the N-methylamide; the ideal geometry
for the zipper complexes. The steric effects of substituents on
the aromatic rings influence the position of the amide cis–trans
equilibrium.63 Previous studies have shown that the four methyl
groups ortho to the amide in compound 26a ensure that it is found
exclusively in the trans-conformation in CDCl3, just as it is in the
1
solid state (Fig. 7e).56 Similarly, the H NMR spectra of model
compound 26e, the bisaniline precursors 18e and 19e, and the
final bisaniline guest compound 20e only showed the presence
1
of the trans-amide conformer. In contrast, the H and 19F NMR
1070 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1071
©

View Article Online
Fig. 11 Conformational equilibria of 15f complexes. Dd Values indicate that 15f is not only able to complex the isophthaloyl hosts in its intended binding
mode a), but also in conformations b) and d) where rearrangement of the amide H-bonds disrupts the intermolecular stacking interaction.
Fig. 12 Overlay of solution structures calculated from the Dd values in Table 5. The inset shows the geometry of the offset aromatic stacks. The dotted
grey ring shows the position of the aniline rings of the isophthaloyl derivatives, and the yellow dot indicates the position of the m1 methyl carbon in
compound 15g.
The determination of guest Dd values in complexes containing
isophthaloyl derivatives 8d (Y = Cl), 8e (Y = NO2) and 9 was not
possible due to the low solubility of these compounds in CDCl3.
However, the Dd values of the isophthaloyl hosts 8d, 8e and 9
can be used to establish how similar these complexes are to the
NMR solution structures that have been determined. There are
some small differences in the Dd values of the isophthaloyl host
compounds containing the most electron-withdrawing groups.
The largest differences are observed for the complexes of 8e (Y =
NO2) with both the 15x and 20x series of bisaniline guests. The
Dd values of the d and t signals are lower than those seen in
the other complexes. The isophthaloyl derivative 8e is likely to
be a poorer H-bond acceptor and a better H-bond donor than
the other isophthaloyl derivatives. A small population of the
alternative mode of complexation shown in Fig. 11c is consistent
with experimental Dd values. The central isophthaloyl aromatic
ring is moved out of the bisaniline pocket, accounting for the
reduced Dd values of d and t. Further support for the existence
of this conformational equilibrium can be obtained from the Dd
values of the s signal, since this chemical shift is affected by the
orientations of the adjacent amides (Fig. 13). Dd Values of s are
∼0.2 ppm larger in complexes with 8e compared to the other
1072 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
isophthaloyl derivatives. If the +1.3 ppm Dd value calculated for
the b) to c) transition in Fig. 13 is accurate, then ∼15% of the
bound state of the complexes involving 8e can be attributed to
the binding mode shown in Fig. 13c. A second estimate of the
position of the equilibrium can be obtained from the Dd values of
the nh protons in complexes involving 8e compared with 8a. Some
of the shift of the nh protons in the well-behaved 8a complexes
will be caused by polarisation of the isophthaloyl amide when it
accepts two H-bonds from the nitropyrrole unit of the bisaniline
derivatives. The 10% increase in Dd observed for the nh protons in
complexes with 8e versus 8a is consistent with the 15% population
of the alternative binding mode as estimated above (Fig. 11c and
13c).
Fig. 13 The chemical shift of the s proton in isophthaloyl derivatives
8a–10 is influenced by the orientations of the amides. Chemical shift
changes were calculated using HF/6-31G*.
The Dd values of the aa and me signals on the terminal stacked
aromatic rings in complexes with 8e are similar to the other
complexes, which suggests that the geometry of terminal stacking
interaction in the conformational ensemble is not greatly affected
by the alternative binding mode. The geometry of the stacking
interaction in the alternative binding mode of 8e is probably not
identical to the main mode of complexation, because the bisaniline
guest has to twist so that the amide carbonyl oxygens are the
correct separation to accept two H-bonds from 8e. The occurrence
of the additional binding mode of 8e is certainly not ideal, but
Dd values indicate that 8e binds all bisaniline derivatives in a
similar way, including the single-mutant compound 15g (complex
B in Fig. 1). Thus, the free energy differences arising from the
alternative geometry of the minor mode of complexation are
probably cancelled in the double-mutant cycle. The double-mutant
cycle for complexes of 8e measures the population-weighted
average of the two similar (but probably not identical) stacking
interactions shown in Fig. 11a and c.
Summary of conformational studies
In summary, 1H NMR complexation-induced changes in chemical
shift (Dd) have proved to be of exceptional utility for the
characterisation of the supramolecular complexes used in this
study. It has been possible to identify conformational complexities.
Some complexes behaved as originally designed (complexes with
15a, 15b) and others behaved in unexpected ways that prevent
valid double-mutant cycles from being constructed (complexes
with 15c, 15e and 15f). The information obtained from Dd values
has guided the design of compounds that overcame some of the
conformational problems (20d, 20e and 20f). Small conforma-
tional differences in the complexes containing compound 8e mean
that some caution may need to be applied when interpreting the
results where Y = NO2.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1073
©

View Article Online
variations, in addition to the anthracene and pentafluorophenyl
groups. This diverse cross-section of aromatic groups are not
readily described using Hammett substituent constants. Instead,
calculated electrostatic surface potentials (ESPs) have been used
to generate a scale that describes the properties of the aromatic
groups employed in this study.
Fig. 14 shows the variation in the electrostatic potential surfaces
of molecules representative of the aromatic groups used in binding
studies. The values of the ESPs at the centre of the each aromatic
group are included in Table 7. As has been noted by others, an
OH or OMe substituent very slightly reduces the ESP at the ring
centre, in accord with the small and positive Hammett meta-
substituent constants (rm) of OH and OMe.69,70 Since a subset
Double-mutant cycle results and discussion
The 1 : 1 association constants and free energies of complexation
for all complexes used in this study are given in Table 6. Using
the data in Table 6 and the equation from Fig. 1, the magnitude
of the intermolecular stacking interaction was calculated for each
of the aromatic pairs in Table 7. Although many of the values
are the same within experimental error, the interactions range
from +1.5 kJ mol−1 to −3.2 kJ mol−1 and are clearly sensitive
to the nature of the aromatic substituents. Hammett substituent
constants are frequently used in physical organic chemistry to
rationalise electrostatic trends in experimental data.68 In the work
presented here, there is a mixture of ortho and para substituent
Table 6 Association constants (Ka in M−1) and free energies of complexation (DG in kJ mol−1) measured in CDCl3 at 298 Ka
Isophthaloyl host
Bisaniline
guest
15a
15b
106 18
−11.4 0.4
69 10
−10.3 0.4 −10.3 0.4
69 11
70.1
8
149 31
56
9
46
3
−10.4 0.3 −12.2 0.5
−9.8 0.4
−9.3 0.1
151 30
−12.2 0.5
70
7
n.d.b
n.d.
n.d.
n.d.
n.d.
175 47
−12.6 0.7
n.d.
54
6
−10.3 0.2
15c
15f
116 20
−11.6 0.4
2 : 1c
128 37
−11.8 0.7
n.d.
n.d.
2 : 1
513 113
133 22
124 29
n.d.
73 16
−15.2 0.5
−11.9 0.4 −11.7 0.6
−10.4 0.5
15g
20d
206 44
−13.0 0.5
83 15
−10.8 0.5 −11.3 0.5
103 20
69 15
−10.3 0.5 −11.4 −0.6
110 28
66 13
−10.2 0.5 −9.4 0.3
47
5
357 82
−14.3 0.8
248 90 235 69
−13.4 0.9 −13.3 0.7
234 89 330 223
−13.3 0.9 −14.1 1.7
111 53 88 23
−11.5 1.2 −10.9 0.6
20e
20f
386 94
−14.5 0.6
196 52
−12.9 0.6
n.d.
n.d.
73 22
−10.4 0.7
n.d.
43
4
−9.2 0.2
627 206
−15.7 0.8
227 58
−13.2 0.6 −13.1 0.6
212 53
94 18
−11.1 0.5 −11.2 0.7
97 26
68 13
−10.3 0.5 −8.9 0.4
39
7
a Titration experiments were repeated at least twice and Ka and DG are the weighted mean based on the observed chemical shift of the d, t and nh signals.
Quoted errors are twice the standard error. b Not determined. c Binds as a 2 : 1 complex as described in the text.
1074 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Table 7 Aromatic stacking interaction energies (DDG in kJ mol−1) measured in CDCl3 at 298 K. The electrostatic surface potentials (kJ mol−1) of each
group are indicateda
Substituted aniline ring (on isophthaloyl host)
ESP
Aromatic group (on
bisaniline guest)
ESP
−92
+1.5 0.7
+1.1 0.8
+0.4 0.6
+0.8 0.6
+0.9 0.7
−0.1 0.7
+0.3 0.7
−0.8 0.9
−0.8 1.0
−54
n.d.b
n.d.
n.d.
−21
−13
+73
+0.2 1.2
−1.7 0.9
−3.2 1.1
−1.2 1.2
−2.3 0.9
−2.9 0.9
−0.5 1.1
−1.5 1.3
+0.2 1.6
−1.2 1.9
+0.8 1.0
−0.2 1.0
n.d.
n.d.
n.d.
−2.2 0.9
−1.2 0.9
−0.5 0.8
a Electrostatic surface potentials were calculated at the B3LYP/6-31G* level at the centre of each aromatic group. Titration experiments were repeated at
least twice. Quoted errors are twice the standard error. b Not determined.
of the experimental stacking interaction energies reported here
have previously been correlated against rm values,55 it is valuable
to know how rm relates to calculated ESPs. Very good agreement
is observed between rm and the ESP at the ring centres of a series
of para-substituted meta-xylenes (Fig. 15).68 Similar plots using
AM1 level calculations or rp were less good.
difluorophenyl (yellow points) and nitrophenyl (green points)
groups are part way between those of the pentafluorophenyl (blue
points) and dimethylphenyl groups (red points), and accordingly,
their stacking interaction energies are generally found to lie
between those of the pentafluorophenyl and dimethylphenyl
groups.
A plot of the experimental aromatic stacking interaction
energies determined using the zipper complex double-mutant
cycles against the ring centre ESPs of the aromatic groups in the
isophthaloyl derivatives reveals some interesting general trends
(Fig. 16). The stacking interaction of the negative surface of
the dimethylphenyl group (red points in Fig. 16) with the most
electron-rich aromatic group (Y = NMe2) is the most repulsive
interaction encountered in these studies. As the ESP of the partner
ring becomes less negative, the stacking interaction becomes less
repulsive. The pentafluorophenyl group (blue points) has a positive
surface and it interacts most favourably with electron-rich aro-
matics, inverting the interaction trend seen for the dimethylphenyl
group. The groups of Gung and Siegel observed similar effects
in intramolecular systems.32,71 For all of the interaction trends,
as the ESP on the interacting partner approaches zero, then so
does the magnitude of stacking interaction. The ESPs of the
While the major trends in the aromatic stacking interaction
energies can be attributed to electrostatic effects using the very
simplistic model described above, this is not the complete picture.
For example, why are there favourable stacking interactions
between the nitro-substituted aromatic groups and electron-rich
aromatics, when the ESPs of the nitro-substituted rings also have
partial negative charges? This discrepancy arises because the ring-
centre ESP model is too simple to fully describe the interactions
that have been measured. Clearly, when two aromatic rings stack
upon one another the surfaces of the rings are brought into contact
and so are the substituents. In an offset-stacked conformation, the
ring substituents may also come into contact with the surface
of the aromatic ring.21,30,72 The NMR solution structures in
Fig. 12 indicate that the ortho-methyl (or fluorine) substituents
are positioned close to the ortho carbon of the opposing ring. The
methyl groups in o-dimethyl-p-nitrophenyl (Fig. 14h) are rather
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1075
©

View Article Online
Fig. 15 B3LYP/6-31G* electrostatic surface potentials (ESPs) at the
ring centres of para-substituted meta-xylenes correlate well with Hammett
meta-substituent constants that have been previously used in structure
activity relationships describing aromatic stacking interactions.
Fig. 14 B3LYP/6-31G* calculated electrostatic potential surfaces (ESPs)
of molecules representative of the stacking aromatic groups used in
this study. Position 1 in the compound nomenclature used refers to the
position at which the group is connected to the isophthaloyl or bisaniline
derivative in the zipper complex; a) 2,6-dimethyl-4-dimethylaminobenzene,
b) 2,6-dimethylbenzene, c) anthracene, d) 2,6-dimethyl-4-methoxybenzene,
e) 4-chloro-2,6-dimethylbenzene, f) acridine, g) 2,6-difluorobenzene, h)
2,6-dimethyl-4-nitrobenzene, i) pentafluorobenzene. Colours are scaled
from −100 to +100 kJ mol−1 (red to blue), green represents neutral charge.
ESP −21 kJ mol−1). Although the ESP at the ring centre of the
difluorophenyl ring is similar to that of the nitro-substituted ring,
the polar methyl groups (+79 kJ mol−1) have been replaced by
fluorine substituents with a partial negative charge (−42 kJ mol−1).
Thus, the favourable polar CH3–p interaction is replaced by a
small but repulsive F–p interaction. Using this departure from the
dimethylphenyl trend line, an upper limit for the CH3–p interaction
in this complex can be tentatively assigned as −1.0 kJ mol−1.
Interestingly, this value is similar to the intercept of the red line
in Fig. 16. In other words, the contribution from the stacking
interaction is approximately zero when the ESP at the centre of
the aniline ring is zero, but there is an additional contribution from
attractive CH3–p interactions.
polar, with an ESP of +79 kJ mol−1, a value similar to the ring
centre of a pentafluorophenyl group. This explains the favourable
interaction between the nitro-substituted aromatics and electron-
rich rings.
All of the aromatic groups used in this study contain ortho-
methyl or fluorine substituents, and our data contain further
evidence of the importance of substituent-to-ring interactions in
the zipper complexes. As mentioned earlier, there is a clear trend
for the dimethylphenyl stacking interactions (red line in Fig. 16).
The sole significant departure from this trend corresponds to the
stacking interaction with the difluoroaniline group (ring centre
The stacking interactions of the pentafluorophenyl group (blue
points in Fig. 16) appear to be better described using a simple ring-
centre ESP model, and this trend line meets the y-axis close to the
origin. This may be explained by the relative simplicity of the ESP
Fig. 16 Plot of experimental aromatic stacking interaction energies measured in zipper complexes (y-axis) against the B3LYP/6-31G* calculated
electrostatic surface potential at the ring centre of substituted isophthaloyl derivatives (Y = NMe2 to Y = NO2) (x-axis).
1076 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
distribution over the surface of the pentafluorophenyl group. The
centre of the pentafluorophenyl group has partial positive charge,
but the ESPs of the fluorine substituents at the face of the ring
are close to zero (cf. the partial negative charge of the fluorine
substituents in the difluorophenyl group). This means that the
trend in the interaction energies for the pentafluorophenyl group is
not complicated by electrostatic interactions involving the fluorine
substituents.
electrostatic surface potentials (not shown). Although dispersion
interactions are almost certainly important to the overall stability
of these compounds, the conformational changes and inferred
changes in the stacking interactions are in perfect accord with
the results of Siegel, Gung and those presented in the current
study.28,31,32,55 i.e. the negative charge on the p-face of acridine is
reduced by N-oxidation, which decreases electrostatic repulsion
between the rings. Additionally, the skewed arrangement of the
rings in the acridyl di-N,Nꢀ-oxide can be attributed to electrostatic
repulsion between the negatively charged oxygen atoms.
The
o-dimethylphenyl·difluorophenyl
and
the
o-
dimethylphenyl·o-dimethyl-p-nitrophenyl stacking interactions
have been measured twice; once with the dimethylphenyl group
present in the isophthaloyl host (8b·20d) and again with the
dimethylphenyl group in the bisaniline guest (9·15a). While the
stacking interactions that are reported are the same within the
quoted error margins, it may be more than a coincidence that
in both cases, the interaction is 1.5 kJ mol−1 more stable when
the dimethylphenyl ring is contained within the isophthaloyl
host. It is important to note that the geometry of the stacking
interaction in the zipper complexes is not symmetrical, and
that the relative orientation of the rings depends on which
half of the complex the aromatic group is attached to. The
difference in the interaction energies may be an indication of how
sensitive aromatic stacking interactions are to subtle changes
in geometry and suggests that the geometric constraints of the
zipper architecture prevents the aromatic groups from reaching
the minimum energy arrangement. However, without the ability
to fix the geometry of the interaction, it would not have been
possible to examine the effects of substituents which would
otherwise be obscured by structural differences.
Theoretical calculations of aromatic interactions in the gas
phase have indicated that electrostatic interactions provide less of
a contribution to the total interaction energy than van der Waals
interactions (consisting of dispersion, induction and repulsion
terms).16,18–21,59,73–75
The experimental studies of Wilcox and co-workers, and
Nakamura and Houk, have suggested that dispersion forces make
important contributions in edge-to-face aromatic interactions.76–78
However, recent insights have shown that solvent competition
provides a more general explanation for the behaviour of the
torsion balance molecules employed in these studies.79,80
The strained intramolecular systems employed by Siegel and
Wolf for the investigation of aromatic stacking interactions
represent a special case where the solvent is completely excluded
from interaction of interest.28,29,37 The intermolecular approach
taken in the current study is able to provide additional insights
into the contribution of van der Waals interactions to aromatic
stacking interactions in the solution phase. At this point it is
important to note that although the binding experiments were
performed in CDCl3, much of the solvent may be displaced from
the terminal aromatic rings by the mutant alkyl groups in the
reference complexes in the double-mutant cycle (Fig. 1). This
means that the measured aromatic interaction energies effectively
refer to an alkane pseudo-solvent environment rather than CDCl3.
Anthracene has a higher polarisability than simple aromatics,81
yet the stacking interactions of the anthracene group (orange
points in Fig. 16) are practically identical to those obtained with
the smaller dimethylphenyl group (red points) and follow the same
electrostatic trend as the aniline ring Y-substituent is varied. Any
contributions from differences in dispersion interactions in the
anthracene stacking interactions are small enough to be masked
by electrostatic effects and the experimental errors. Dispersion
interactions play a lesser role in solution compared to the gas
phase because of competitive dispersion interactions with the
solvent. When a stacking interaction forms between two aromatic
rings, the solvent molecules coating the interacting surfaces are
displaced. Thus, if the solvent is able to coat the entire surface
of the rings evenly, the favourable free energy of the dispersion
interaction between the stacked aromatic rings will be reduced by
the cost of breaking the dispersion interactions of each aromatic
group with the solvent. Theory suggests that there is little variation
of the dispersion interaction energy per unit surface area for
intermolecular interactions between organic molecules, and so we
should expect electrostatic effects to dominate in solution.80,82
From a simple consideration of relative areas of molecular
contact, dispersion interactions should be expected to contribute
more to aromatic stacking energies compared with aromatic edge-
to-face interactions. Gas phase calculations predict that stacking
free energies of all substituted benzene dimers are more stable than
the unsubstituted stacked benzene dimer. Thus, electron-donating
substituents have been predicted to stabilise aromatic stacking
interactions even when both p-faces are electron-rich.
Experimental
Computational procedures
Mei and Wolf found that the degree of splay between two
acridine groups forced into an intramolecular stack decreased
upon sequential oxidation to the mono-N-oxide and the di-N,Nꢀ-
oxide.37 The change in geometry was attributed to an increase
in the stability of the stacking interaction between the acridine
rings. This observation was cited as experimental support for the
computational predictions of Sherrill and co-workers, since the
authors believed that N-oxidation increased the electron density of
acridine.21 On the contrary, the nitrogen atom in acridine N-oxide
bears a formal positive charge and the electron density of the entire
acridine p-face is decreased, a fact supported by DFT/6-31G*
Ab initio calculations were performed using Spartan, Wavefunc-
tion, Inc., Irvine, CA, USA. Molecular modelling was performed
using XED 6.1.0, Cresset BioMolecular Discovery Ltd., Hertford-
shire, UK.
General NMR procedures
1H, 19F and 13C spectra were recorded on either a Bruker AC250
or a AMX400 spectrometer with residual solvent as an internal
standard. Fluorine chemical shifts were referenced to an external
CFCl3 reference. Two-dimensional ROESY experiments were
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1077
©

View Article Online
recorded on a Bruker AMX400 with 300 ms mixing time and
a 3 s relaxation delay between pulses.
free concentration of guest in this system can be shown to be:
ꢀ
−1 + 1 + 8KdimG([G]0 − [HG])
[G] =
(2)
(3)
4Kdim G
1H NMR dilutions
and the concentration of guest dimer is:
A saturated analyte solution of known concentration (∼mM) was
prepared in CDCl3. Aliquots of this solution were sequentially
added to a small volume (0.25–0.5 ml) of CDCl3 in an NMR
tube, and the 1H NMR chemical shifts for each signal were
recorded. The program NMRDil_Dimer was used to fit data to a
dimerisation isotherm using a non-linear curve fitting procedure.
Dimerisation constants and the chemical shifts of the dimer and
free analyte were determined for each signal monitored. Dilution
experiments were performed in duplicate and the weighted mean
dimerisation constant (based on the observed changes in chemical
shift) was used in the fitting of NMR titration data as outlined
below.
[GG] = KdimG[G]2
where KdimG is the dimerisation constant of G determined from
NMR dilution experiments. The observed chemical shift of the
guest signal has three contributing terms:
[HG]
[G]0
[G]
[GG]
[G]0
dobsG
=
dboundG
+
dfreeG + 2
ddimG
(4)
[G]0
where [G]0 is the total concentration of guest, dboundG, dfreeG and ddimG
are the chemical shifts of a guest signal in the three states [HG],
[G] and [GG] respectively. The concentrations in this equation are
known from eqn (1)–(3), and dfreeG and ddimG were determined in the
NMR dilution experiment. Thus, the equation can be rearranged
to yield the desired limiting complexation-induced chemical shift
of a particular guest signal, dboundG. The difference between this
number and dfreeG gives the limiting complexation-induced change
in chemical shift which is referred to as the guest Dd elsewhere in
this report.
1H NMR titrations
A few ml of host solution of known concentration (0.2–5 mM) was
prepared in CDCl3. A small sample (0.2–0.5 ml) of this solution
was added to an NMR tube, and a 1H NMR spectrum was
recorded. A guest solution of known concentration (7–40 mM)
was then prepared using the remaining host solution. This solution
was saturated with guest to allow as much coverage of the binding
isotherm as possible (generally 50%–90% was achieved). This
procedure also ensured that the concentration of host remained
constant throughout the titration. Aliquots of the saturated guest
solution were added successively to the NMR tube containing the
host solution, and the 1H NMR spectrum was recorded after each
addition. The program NMRTit_HGHHGG was used to fit host
signals to a 1 : 1 binding isotherm allowing for dimerisation of the
guest using a non-linear curve fitting procedure. For titrations
with compounds 20d and 20f the concentration of guest was
corrected to allow for the presence of ∼30% inactive conformer
NMR structure determination
The method used to determine three-dimensional structures from
limiting changes in complexation-induced chemical shift Dd has
been described in detail elsewhere.65 Molecules were built and
minimised using standard bond lengths and angles in XED.62 Ring
current factors used in the calculation of Dd values, relative to
the phenyl group, were: pyrrole 0.55, central anthracene ring 1.27,
outer anthracene rings 0.8. Ring currents for the anthracene group
were based on ring currents calculated by Fowler and Steiner.83
A
genetic algorithm was used to optimise the conformation of the
complex so that the calculated Dd values matched the experimental
values as closely as possible (Table 5). The structure of each
complex was refined using a population of 2000, a replacement
rate of 400 per generation and five sequential steps of 1000
generations. The first step allowed intermolecular translations
1
as determined by relative integrals of the me and me+ H NMR
signals. The mean association constant for each experiment was
evaluated as the weighted mean based on the observed change
in chemical shift for the d, t and nh signals in all complexes (for
proton labelling scheme see Fig. 2 and 3). The error was taken as
twice the standard error.
Limiting complexation-induced chemical shifts of the 15x and
20x guest signals (Table 4) were extrapolated by determining
the relative concentrations of complex [HG], free guest [G],
guest dimer [GG], and the values of the guest chemical shifts
in these three states. Since the concentration of host was constant
throughout the titration (and low enough that dimerisation was
insignificant), the host concentration was in excess of the guest
during the early stages of a titration. Under these conditions, com-
plexation provides a larger contribution to the observed chemical
shifts of the guest than guest dimerisation. The concentration of
complex (and therefore complexed guest) is given by:
of 10 A, rotations of 360◦ and intramolecular torsional
˚
changes of 360◦ for the bonds highlighted in Fig. 17. After
each step, the calculated structure in closest agreement with
experimental Dd values was used to generate a new population
of 2000, and the search space was reduced by a half in all
[HG] = [H]0 (DdobsH/DdboundH
)
(1)
where [H]0 is the total concentration of host, the observed change
in chemical shift of a particular host signal is DdobsH, and DdboundH
is the limiting complexation-induced change in chemical shift of
the host as determined from the NMR titration experiment. The
Fig. 17 The bond torsions that were free to rotate during the NMR
solution structure determination and the intermolecular NOE constraint
are indicated. X groups are shown in Fig. 3.
1078 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
˚
˚
˚
12 P. A. Williams, J. Cosme, A. Ward, H. C. Angove, D. M. Vinkovic and
H. Jhoti, Nature, 2003, 424, 464–468.
13 G. Kryger, I. Silman and J. L. Sussman, Structure (London), 1999, 7,
297–307.
dimensions: intermolecular translations ( 5 A, 2.5 A, 1.25 A),
rotations ( 180◦, 90◦, 45◦, 22.5◦) and torsions ( 180◦, 90◦, 45◦,
22.5◦). The solubilising R group and piperidine protons did not
move in NMR experiments and were therefore excluded from
calculations. van der Waals clashes were penalised at distances of
14 G. Kryger, I. Silman and J. L. Sussman, J. Physiol. (Paris), 1998, 92,
191–194.
15 D. F. V. Lewis, M. N. Jacobs and M. Dickins, Drug Discovery Today,
2004, 9, 530–537.
16 S. Tsuzuki, K. Honda, T. Uchimaru, M. Mikami and K. Tanabe, J. Am.
Chem. Soc., 2002, 124, 104–112.
˚
˚
less than 2 A for intermolecular clashes and 1 A for intramolecular
clashes for non-hydrogen atoms. The NOE constraint illustrated
in Fig. 17 was imposed by applying a penalty, if the inter-
17 S. Perez-Casas, J. Hernandez-Trujillo and M. Costas, J. Phys. Chem. B,
2003, 107, 4167–4174.
˚
proton separation exceeded 5 A. Each geometry optimisation was
run at least five times. Structure calculations yielding RMSDs
between calculated and experimental Dd values less than 0.17 ppm
were accepted. The optimised structures obtained from repeat
calculations of the complexes shown in Fig. 12 were very similar,
but only the structures with the lowest RMSDs are reported in
this work.
18 M. O. Sinnokrot and C. D. Sherrill, J. Am. Chem. Soc., 2004, 126,
7690–7697.
19 M. O. Sinnokrot and C. D. Sherrill, J. Phys. Chem. A, 2003, 107, 8377–
8379.
20 M. O. Sinnokrot and C. D. Sherrill, J. Phys. Chem. A, 2004, 108, 10200–
10207.
21 A. L. Ringer, M. O. Sinnokrot, R. P. Lively and C. D. Sherrill, Chem.–
Eur. J., 2006, 12, 3821–3828.
22 B. W. Gung and J. C. Amicangelo, J. Org. Chem., 2006, 71, 9261–9270.
23 C. A. Hunter and J. K. M. Sanders, J. Am. Chem. Soc., 1990, 112,
5525–5534.
Synthetic procedures
Detailed synthetic procedures and compound characterisation
data are provided in the ESI.†
24 C. A. Hunter, K. R. Lawson, J. Perkins and C. J. Urch, J. Chem. Soc.,
Perkin Trans. 2, 2001, 651–669.
25 M. L. Waters, Curr. Opin. Chem. Biol., 2002, 6, 736–741.
26 E. A. Meyer, R. K. Castellano and F. Diederich, Angew. Chem., Int.
Ed., 2003, 42, 1210–1250.
Conclusions
27 M. L. Waters, Biopolymers, 2004, 76, 435–445.
28 F. Cozzi and J. S. Siegel, Pure Appl. Chem., 1995, 67, 683–689.
29 F. Cozzi, R. Annunziata, M. Benaglia, M. Cinquini, L. Raimondi,
K. K. Baldridge and J. S. Siegel, Org. Biomol. Chem., 2003, 1, 157–162.
30 M. J. Rashkin and M. L. Waters, J. Am. Chem. Soc., 2002, 124, 1860–
1861.
31 B. W. Gung, X. Xue and H. J. Reich, J. Org. Chem., 2005, 70, 3641–3644.
32 B. W. Gung, M. Patel and X. Xue, J. Org. Chem., 2005, 70, 10532–10537.
33 M. S. Cubberley and B. L. Iverson, J. Am. Chem. Soc., 2001, 123,
7560–7563.
Aromatic stacking interactions are sensitive to changes in geome-
try and the degree of overlap. Using H-bonded supramolecular
zipper complexes, it has been possible to lock the geometry
of two aromatic rings in an offset stacked arrangement. This
has enabled the effects of substituents on the interaction free
energy to be quantified. The conformational behaviour of the
complexes in the solid state and in solution has been thoroughly
investigated. Some complexes were found to be incompatible with
the double-mutant cycle methodology and were excluded from
the analysis due to significant conformational changes. Insights
obtained from these conformational studies guided the design
and synthesis of new compounds better suited to the approach.
To a first approximation, the electrostatic properties of the ring
surfaces dominate the trends in the interaction energy. However,
direct electrostatic interactions with the ring substituents also
make important contributions. The interplay of these two factors
could lead to complicated behaviour, for example, quite different
interactions between similar aromatic groups in different contexts.
34 L. F. Newcomb and S. H. Gellman, J. Am. Chem. Soc., 1994, 116,
4993–4994.
35 R. R. Gardner, S. L. McKay and S. H. Gellman, Org. Lett., 2000, 2,
2335–2338.
36 S. L. McKay, B. Haptonstall and S. H. Gellman, J. Am. Chem. Soc.,
2001, 123, 1244–1245.
37 X. Mei and C. Wolf, J. Org. Chem., 2005, 70, 2299–2305.
38 M. Gray, A. J. Goodman, J. B. Carroll, K. Bardon, M. Markey, G.
Cooke and V. M. Rotello, Org. Lett., 2004, 6, 385–388.
39 R. Faraoni, M. Blanzat, S. Kubicek, C. Braun, W. B. Schweizer, V.
Gramlich and F. Diederich, Org. Biomol. Chem., 2004, 2, 1962–1964.
40 T. Yajima, R. Takamido, Y. Shimazaki, A. Odani, Y. Nakabayashi and
O. Yamauchi, Dalton Trans., 2007, 299–307.
41 S. L. Cockroft and C. A. Hunter, Chem. Soc. Rev., 2007,
DOI: 10.1039/b603842p.
42 S. M. Butterfield and J. Rebek, J. Am. Chem. Soc., 2006, 128, 15366–
15367.
References
43 H. Adams, F. J. Carver, C. A. Hunter, J. C. Morales and E. M. Seward,
Angew. Chem., Int. Ed. Engl., 1996, 35, 1542–1544.
44 H. Adams, K. D. M. Harris, G. A. Hembury, C. A. Hunter, D.
Livingstone and J. F. McCabe, Chem. Commun., 1996, 2531–2532.
45 F. J. Carver, C. A. Hunter and E. M. Seward, Chem. Commun., 1998,
775–776.
46 H. Adams, P. L. Bernad, Jr., G. A. Hembury, C. A. Hunter, J. F. McCabe,
D. S. Eggleston, R. C. Haltiwanger, D. J. Livingstone, K. D. M. Harris
and B. M. Kariuki, Chem. Commun., 2001, 1500–1501.
47 H. Adams, J.-L. J. Blanco, G. Chessari, C. A. Hunter, C. M. R. Low,
J. M. Sanderson and J. G. Vinter, Chem.–Eur. J., 2001, 7, 3494–3503.
48 F. J. Carver, C. A. Hunter, P. S. Jones, D. J. Livingstone, J. F. McCabe,
E. M. Seward, P. Tiger and S. E. Spey, Chem.–Eur. J., 2001, 7, 4854–
4862.
49 F. J. Carver, C. A. Hunter, D. J. Livingstone, J. F. McCabe and E. M.
Seward, Chem.–Eur. J., 2002, 8, 2847–2859.
50 C. A. Hunter, C. M. R. Low, C. Rotger, J. G. Vinter and C. Zonta, Proc.
Natl. Acad. Sci. U. S. A., 2002, 99, 4873–4876.
51 C. A. Hunter, C. M. R. Low, C. Rotger, J. G. Vinter and C. Zonta,
Chem. Commun., 2003, 834–835.
1 B. R. Hamilton and D. L. Hammick, J. Chem. Soc., 1938, 1350–1352.
2 D. B. Amabilino and J. F. Stoddart, Chem. Rev., 1995, 95, 2725–2829.
3 H. C. Kolb, P. G. Andersson and K. B. Sharpless, J. Am. Chem. Soc.,
1994, 116, 1278–1291.
4 M. Keller, C. Lehmann and M. Mutter, Tetrahedron, 1999, 55, 413–422.
5 B. M. Trost, D. O’Krongly and J. L. Belletire, J. Am. Chem. Soc., 1980,
102, 7595–7596.
6 M. Weck, A. R. Dunn, K. Matsumoto, G. W. Coates, E. B. Lobkovsky
and R. H. Grubbs, Angew. Chem., Int. Ed., 1999, 38, 2741–2745.
7 G. J. Gabriel, S. Sorey and B. L. Iverson, J. Am. Chem. Soc., 2005, 127,
2637–2640.
8 G. J. Gabriel and B. L. Iverson, J. Am. Chem. Soc., 2002, 124, 15174–
15175.
9 K. M. Guckian, B. A. Schweitzer, R. X. F. Ren, C. J. Sheils, D. C.
Tahmassebi and E. T. Kool, J. Am. Chem. Soc., 2000, 122, 2213–2222.
10 W. Saenger, Principles of Nucleic Acid Structure, Spring-Verlag,New
York, 1988.
11 G. B. McGaughey, M. Gagne and A. K. Rappe, J. Biol. Chem., 1998,
273, 15458–15463.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1062–1080 | 1079
©

View Article Online
52 C. A. Hunter, C. M. R. Low, J. G. Vinter and C. Zonta, J. Am. Chem.
Soc., 2003, 125, 9936–9937.
53 H. Adams, S. L. Cockroft, C. Guardigli, C. A. Hunter, K. R. Lawson,
J. Perkins, S. E. Spey, C. J. Urch and R. Ford, ChemBioChem, 2004, 5,
657–665.
54 H. Adams, C. A. Hunter, K. R. Lawson, J. Perkins, S. E. Spey, C. J.
Urch and J. M. Sanderson, Chem.–Eur. J., 2001, 7, 4863–4877.
55 S. L. Cockroft, C. A. Hunter, K. R. Lawson, J. Perkins and C. J. Urch,
J. Am. Chem. Soc., 2005, 127, 8594–8595.
56 I. Azumaya, K. Yamaguchi, H. Kagechika, S. Saito, A. Itai and K.
Shudo, Yakugaku Zasshi, 1994, 114, 414–430.
57 A. Spitaleri, C. A. Hunter, J. F. McCabe, M. J. Packer and S. L.
Cockroft, CrystEngComm, 2004, 6, 489–493.
58 C. R. Patrick and G. S. Prosser, Nature, 1960, 1021.
59 P. Mignon, S. Loverix, J. Steyaert and P. Geerlings, Nucleic Acids Res.,
2005, 33, 1779–1789.
60 F. Ponzini, R. Zagha, K. Hardcastle and J. S. Siegel, Angew. Chem., Int.
Ed., 2000, 39, 2323–2325.
61 J. C. Collings, K. P. Roscoe, E. G. Robins, A. S. Batsanov, L. M. Stimson,
J. A. K. Howard, S. J. Clark and T. B. Marder, New J. Chem., 2002, 26,
1740–1746.
67 C. A. Hunter, C. M. R. Low, M. J. Packer, S. E. Spey, J. G. Vinter,
M. O. Vysotsky and C. Zonta, Angew. Chem., Int. Ed., 2001, 40, 2678–
2682.
68 C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91, 165–195.
69 S. Mecozzi, A. P. West, Jr. and D. A. Dougherty, Proc. Natl. Acad. Sci.
U. S. A., 1996, 93, 10566–10571.
70 F. Hof, D. M. Scofield, W. B. Schweizer and F. Diederich, Angew.
Chem., Int. Ed., 2004, 43, 5056–5059.
71 F. Cozzi, F. Ponzini, R. Annunziata, M. Cinquini and J. S. Siegel,
Angew. Chem., Int. Ed. Engl., 1995, 34, 1019–1020.
72 B. W. Gung, X. Xue and H. J. Reich, J. Org. Chem., 2005, 70, 7232–7237.
73 E. C. Lee, B. H. Hong, J. Y. Lee, J. C. Kim, D. Kim, Y. Kim, P.
Tarakeshwar and K. S. Kim, J. Am. Chem. Soc., 2005, 127, 4530–4537.
74 P. Mignon, S. Loverix and P. Geerlings, Chem. Phys. Lett., 2005, 401,
40–46.
75 A. Olasz, P. Mignon, F. De Proft, T. Veszpremi and P. Geerlings, Chem.
Phys. Lett., 2005, 407, 504–509.
76 S. Paliwal, S. Geib and C. S. Wilcox, J. Am. Chem. Soc., 1994, 116,
4497–4498.
77 E.-i. Kim, S. Paliwal and C. S. Wilcox, J. Am. Chem. Soc., 1998, 120,
11192–11193.
62 J. G. Vinter, J. Comput.-Aided Mol. Des., 1994, 8, 653–668.
63 A. Itai, Y. Toriumi, N. Tomioka, H. Kagechika, I. Azumaya and K.
Shudo, Tetrahedron Lett., 1989, 30, 6177–6180.
64 C. A. Hunter, M. J. Packer and C. Zonta, Prog. Nucl. Magn. Reson.
Spectrosc., 2005, 47, 27–39.
65 C. A. Hunter and M. J. Packer, Chem.–Eur. J., 1999, 5, 1891–1897.
66 M. Gardner, A. J. Guerin, C. A. Hunter, U. Michelsen and C. Rotger,
New J. Chem., 1999, 23, 309–316.
78 K. Nakamura and K. N. Houk, Org. Lett., 1999, 1, 2049–2051.
79 T. Ren, Y. Jin, K. S. Kim and D. H. Kim, J. Biomol. Struct. Dyn., 1997,
15, 401–405.
80 S. L. Cockroft and C. A. Hunter, Chem. Commun., 2006, 3806–3808.
81 D. R. E. Lide, CRC Handbook of chemistry and physics, 75th edn, CRC
Press, Boca Raton, 1994.
82 C. A. Hunter, Angew. Chem., Int. Ed., 2004, 43, 5310–5324.
83 P. W. Fowler and E. Steiner, Chem. Phys. Lett., 2002, 364, 259–266.
1080 | Org. Biomol. Chem., 2007, 5, 1062–1080
This journal is
The Royal Society of Chemistry 2007
©