Conformer-independent ureidoimidazole motifs-tools to probe conformational and tautomeric effects on the molecular recognition of triply hydrogen-bonded heterodimers

Article abstract of DOI:10.1002/chem.201102128

Linear arrays of hydrogen bonds are useful for the reversible assembly of "stimuli-responsive" supramolecular materials. There is thus an ongoing requirement for easy-to-synthesise motifs that are capable of presenting hydrogen-bonding functionality in a

Full text of DOI:10.1002/chem.201102128

DOI: 10.1002/chem.201102128
Conformer-Independent Ureidoimidazole Motifs—Tools to Probe
Conformational and Tautomeric Effects on the Molecular Recognition of
Triply Hydrogen-Bonded Heterodimers
Maria L. Pellizzaro,[a] Andrea M. McGhee,[a] Lisa C. Renton,[a] Michael G. Nix,[a]
Julie Fisher,[a] W. Bruce Turnbull,[a] and Andrew J. Wilson*[a, b]
Abstract: Linear arrays of hydrogen
bonds are useful for the reversible as-
sembly of “stimuli-responsive” supra-
molecular materials. There is thus an
ongoing requirement for easy-to-syn-
thesise motifs that are capable of pre-
senting hydrogen-bonding functionality
in a predictable manner, such that
high-affinity and high-fidelity recogni-
tion occurs. The design of linear arrays
is made challenging as a consequence
of their ability to adopt multiple con-
formational and tautomeric configura-
tions; with each additional hydrogen-
bonding heteroatom added to an array,
the available tautomeric and conforma-
tional space increases and it can be dif-
ficult to anticipate where unproductive
conformers/tautomers will arise. This
paper describes a detailed study on the
conformations are postulated, both of
which present a DDA array that is
complementary to appropriate DAA
1
partners. 1D and 2D H NMR spectros-
complementary
ureidoimidazole
copy, isothermal titration calorimetry,
and ab initio structure calculations con-
firm 1 interacts with 2 (Ka ꢀ33000mꢁ1
in CDCl3) in a conformer-independent
fashion driven by enthalpy. Compari-
son of the binding behaviour of 1 with
hexylamidocytosine (4) and amido-
naphthyridine (5) provides insight on
the role that intramolecular hydrogen-
bonding plays in mediating affinity to-
wards DAA partners.
donor–donor–acceptor (DDA) array
(1) and amidoisocytosine donor–ac-
ceptor–acceptor (DAA) array (2). A
specific feature of 1 is that two degen-
erate, intramolecular hydrogen-bonded
Keywords: hydrogen bonds · linear
arrays · molecular recognition · self-
assembly · supramolecular chemis-
try
Introduction
development of motifs with high-fidelity recognition behav-
iour,[24] whilst using the minimum number of hydrogen
bonds, which will potentially simplify synthesis. Much recent
research has focused upon linear arrays employing
four[12,13,25–42] or more[43–47] hydrogen bonds because of the
enhanced strength of association that may be achieved.
However, recent interest in triply hydrogen-bonded sys-
tems[15,21–23,48–52] has illustrated that useful affinity can be
achieved for assembly of dynamic architectures.[53–56] Within
this family of hydrogen-bond receptors two classes emerge:
I) rigid (poly)cyclic aromatic systems with pendant exocyclic
hydrogen-bond donor/acceptor functionality[3,4] and II) sys-
tems in which individual donors (D) and acceptors (A) are
separated from one another.[27,42] Although arrangement,
conformation and tautomeric configuration can be more
easily controlled for the latter, the synthetic simplicity of the
former is appealing. A major issue with the design and syn-
thesis of linear arrays of class I, however, is the potential for
different (unanticipated) conformational and tautomeric
configurations, which can complicate analysis of molecular
interactions or have a deleterious effect on the desired
mode of molecular recognition. For instance the elegantly
designed deazapterin array introduced by Zimmerman as a
tautomer-independent quadruple array[19] potentially adopts
five different tautomers/conformers, of which four have
been experimentally observed. Two of the configurations
The design and synthesis of linear arrays of hydrogen bonds
is an area of continuing investigation in modern supramolec-
ular chemistry.[1–5] Not only do these systems represent ex-
cellent motifs to further our fundamental understanding of
molecular recognition, but they also serve as building blocks
for self-assembly[6–8] and biological probes.[9,10] In addition to
the number of primary interactions,[11–14] the arrange-
ment,[15,16] acidity/basicity of the hydrogen-bond donor/ac-
ceptors,[12,17,18] tautomerism,[19] electronic substituent ef-
fects[20,21] and conformation[22,23] all affect the affinity and
specificity of the interactions made by linear arrays. Careful
consideration of these factors should allow the design and
[a] M. L. Pellizzaro, Dr. A. M. McGhee, L. C. Renton, Dr. M. G. Nix,
Dr. J. Fisher, Dr. W. B. Turnbull, Dr. A. J. Wilson
School of Chemistry
University of Leeds, Leeds, LS2 9 JT (UK)
Fax : (+44)113-343-6565
E-mail: a.j.wilson@leeds.ac.uk
[b] Dr. A. J. Wilson
Astbury Centre for Structural Molecular Biology
University of Leeds, Leeds, LS2 9 JT (UK)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102128.
14508
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 14508 – 14517

FULL PAPER
are productive for homodimerisation through a DDAA
array and one is present in minor amounts as a self-comple-
mentary DADA array, whilst at least one of two plausible
ADDA arrays is adopted in the presence of a complementa-
ry partner which presents a DAAD array—hence greater
fundamental understanding of these factors is desirable
within the broader context of supramolecular design. Our
own studies in this area have focused on linear arrays em-
ploying three hydrogen-bonding groups, specifically the in-
troduction of conformer-independent arrays.[22,23] We previ-
ously introduced the complementary ureidoimidazole DDA
and 5 (Ka ꢀ3000mꢁ1 in CDCl3) reveals a key role for unde-
sired tautomers and steric effects in mediating molecular
recognition within this series. These results are anticipated
to be broadly applicable to the design of supramolecular
heterodimers.
Results
Synthesis: The hydrogen-bonding motifs 1 and 2 were syn-
thesised (Scheme 2) using routes developed previously.[21,22]
A key point to note is that we revised the purification pro-
(1) and amidoisocytosine DAA (2) arrays (Scheme 1).[22]
A
Scheme 1. Structures of molecules 1–5 (i and ii denote different conform-
ers).
specific feature of 1 is that two degenerate intramolecular
hydrogen-bonded conformations 1i and 1ii are postulated,
both of which present a DDA array that is complementary
to appropriate DAA partners (it should be noted that two
tautomeric configurations are possible for each conformer
that position the tBu group in two different positions not
shown). Indeed, this property confers significantly enhanced
dimerisation constants of 1 towards 2 in deuterochloroform
Scheme 2. Synthesis of molecules 1, 2, 4 and 5.
cedure[21] from our earlier work[22] for preparation of 1; that
is, isolating the product by crystallisation from moist aceto-
nitrile/methanol.-?This had a significant effect on its binding
properties because trace impurities (presumably diphenyl
urea) that could not be detected using our analytical meth-
ods were removed. We considered several routes[57–59] for
the synthesis of 4. Despite the availability of methods for se-
lective alklyation on N1 with a methyl group,[57] we were
unable to selectively alkylate cytosine with hexylbromide.
On the other hand protection of the exocyclic nitrogen as an
amidine followed by alkylation,[58] deprotection and acyla-
tion would introduce an additional protection/deprotection
into the sequence. Therefore, 4 was obtained (Scheme 2) by
the reaction of cytosine 9 and benzoyl chloride, which gave
an insoluble intermediate 10. Reaction with 1-bromohexane
gave 4 (in overall yield of 15%), which is soluble in chloro-
form, the solvent of choice for testing the binding affinities.
The low yield for the alkylation of cytosine derivatives on
N1 has precedence[60] and is noted to arise in part due to
competing alkylation on O2 although acylation mitigates
against this. Compound 5 was synthesised following previ-
ously described methodology[61] (Scheme 2) by first forming
relative to ureidopyridine derivatives such as
3
(Scheme 1).[22]
In the current investigation our aim was to further study
compound 1 to establish the extent to which the array is
truly “conformer-independent” and probe its binding prop-
erties towards a series of complementary triply hydrogen-
bonded arrays. Three compounds that all present DAA
arrays were designed for this study to evaluate the role of
intramolecular hydrogen-bonding on affinity; one that con-
tains an intramolecular hydrogen-bond, 2, and two that are
not pre-organised by intramolecular hydrogen-bonds, hexy-
lamidocytosine (4) and amidonaphthyridine (5; Scheme 1).
Each of these compounds was synthesised and subjected to
a series of experimental and in silico structural studies
alongside binding experiments. The results confirm 1 inter-
acts with 2 (Ka ꢀ33000mꢁ1 in CDCl3) in a conformer inde-
pendent fashion, driven by enthalpy. Comparison with the
binding behaviour of 1 with 4 (Ka ꢀ17,000mꢁ1 in CDCl3)
Chem. Eur. J. 2011, 17, 14508 – 14517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14509

A. J. Wilson et al.
2-amino-7-methylnaphthyridine 12 from 2,6-diaminopyri-
dine 11 and 4,4-dimethoxy-2-butanone. 2-Amino-7-methyl-
naphthyridine 12 was then treated with benzoyl chloride to
give the desired 5.
lised by an intramolecular hydrogen-bond. Only weak inter-
molecular contacts are observed in this structure. As a mon-
omeric building block, 4 is incapable of adopting different
tautomeric forms, but free rotation around the amide bond
offers different rotamer possibilities. The 1H–1H NOESY
NMR spectrum of 4 exhibits only cross-peaks consistent
with the predominance of a conformation presenting the de-
sired DAA array (e.g., absence of correlation between NH
and cytosine ring protons; see the Supporting Information).
Structural studies: We first performed structural analyses on
individual components prior to characterisation of the heter-
odimers. For 1 in CDCl3 at room temperature, a single set of
sharp signals for the tert-butyl group and the aromatic pro-
tons was observed along with two very broad NH signals.
This suggests fast interconversion between all of the possible
conformers/tautomers in addition to any intermolecular as-
sociation that is occurring. The absence of well-resolved NH
resonances in the 1H NMR spectrum hampered our attempts
to perform conformational analysis using 1H-1H NOESY.
The crystal structure of 1 revealed only the conformation in-
volving intramolecular urea carbonyl to imidazole NH hy-
drogen-bonding in addition to intermolecular hydrogen-
bonding by a bifurcated interaction between the urea NHs
and the imidazole N through a bridging methanol molecule
(Figure 1a).[62] Recently we have shown that several deriva-
tives of 1 all adopt a similar mode of solid-state organisa-
1
The H–1H NOESY of 5 also suggested the presence of the
desired DAA conformation.
1H–1H NOESY data were then collected for equimolar
mixtures of each of the three complexes. In the case of 1·2,
these clearly illustrate that both conformers of 1 interact
with 2 (Figure 2); for instance diagnostic cross peaks be-
tween Hg and HC and between Hb and HA can be observed.
1
Low-temperature H NMR of 1·2 (40 mm, 400 MHz, CDCl3,
213-333 K) suggested that all four possible conformers/tau-
tomers are present, with one being the dominant structure
as evidenced by the observation of four different tBu peaks
(see the Supporting Information). The NH peaks did not
fully resolve, however, giving a sufficiently complicated
spectrum to prevent a full assignment using 2D correlation
1
tion.[62] The H NMR spectra of 2 also exhibited poorly re-
solved NH resonances. This is likely owing to rapid ex-
change between different conformers/tautomers in solution,
but likewise prevents detailed conformational analysis. We
previously published a crystal structure of 2[22] (Figure 1b),
in which an undesired tautomeric form was observed, stabi-
Figure 1. X-ray structures of 1 and 2. a) Structure packing diagram of 1
ꢁ
shown in stick format; key distances: O7A H4A 2.29 (intramolecular),
ꢁ
ꢁ
ꢁ
ꢁ
N1A H2S 1.89, H6A O1S 2.16, H8A O1S 2.03, O7A H4B 2.29 ꢁ (in-
termolecular). b) Structure packing diagram of 2 shown in stick format;
key distances: O2 H16 1.94, O22 H36 1.99 (intramolecular), H1 O35
Figure 2. 1H-1H NOESY spectrum (10 mm, CDCl3, 500 MHz), for 1·2
showing that both conformations of 1 can interact with 2.
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
2.11, N12 H36 2.54, O15 H21 2.07, H16 N32 2.59 ꢁ (intermolecular).
14510
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 14508 – 14517

Molecular Recognition
FULL PAPER
spectroscopy, precluding a detailed conformational analyses
using 1H-1H NOESY. 1H-1H NOESY experiments on 1·4
and 1·5 only confirmed the presence of one conformation of
1. There are three plausible explanations for this: 1) both 4
and 5 preferentially interact with a single conformer of 1
(urea-NH/urea-NH/imidazole-N 1i or imidazole-NH/urea
NH/carbonyl-O 1ii), 2) 1 possesses steric features that pre-
vent one of the confomers interacting with 4 and 5 and 3)
the geometry of the interaction is such that any nOe effect
is not observed noting that intermolecular nOeꢂs are typical-
ly of low intensity if observed at all.
DOSY data sets were acquired for the individual mole-
cules 1 and 2, and the complex 1·2. The observed diffusion
coefficient enabled the hydrodynamic radius for the com-
plex to be derived through utilisation of the Stokes–Einstein
equation [Eq. (1)], in which r is the radius (in m), k is the
Boltzmann constant, T is the temperature (in K), h is the
viscosity of the solvent (in Pas) and D is the diffusion coeffi-
cient (in m2 sꢁ1 ꢃ10ꢁ10). In small molecule assemblies such as
these, the radius is related to the cube root of the molecular
mass (MW), allowing the mass of the species that is present
at different monomer concentrations to be calculated by
using Equation (2). (Note: The calculated mass will always
be larger than expected because the DOSY technique in-
cludes a layer of solvent around the molecule when calculat-
ing the radius.) The association behaviour of each of the
molecules was established by measuring diffusion coeffi-
cients at two concentration extremes. For 1 diffusion coeffi-
cients were 7.092ꢃ10ꢁ10 and 5.247ꢃ10ꢁ10 m2 sꢁ1 at concentra-
tions of 10 and 100 mm, respectively. By using Equation (2)
it was established that the species present in solution at high
concentration is a dimer of 1. Since the mass of the mono-
mer is 259 Da it was calculated that the mass present in the
100 mm solution was 639 Da, which is consistent with dimer
formation at very high concentrations. Likewise, the diffu-
Figure 3. Typical a) 1H NMR spectra (starting at 5 mm 2, 300 MHz,
CDCl3) and b) titration curve for the titration of 1 into 2.
resonances was observed—visual analysis of the titration
curve revealed a non-ideal fit in this region of the binding
isotherm, suggestive of additional equilibria. Incorporation
of the dimerisation constant for 1 (Kdim =11mꢁ1)[22] into the
model used for fitting improved the quality of the fit, with
little change in the resulting association constant. Thus for
sion coefficients for
2
were 9.575ꢃ10ꢁ10 and 7.058ꢃ
10ꢁ10 m2 sꢁ1 at concentrations of 10 and 100 mm, respectively.
Assuming that 2 is only present as monomer (228 Da) at
10 mm a mass of 569 Da was calculated for the 100 mm
sample, corresponding to a homodimer. These data are con-
sistent with 1 and 2 being monomeric at 10 mm and hence
this concentration was used to study the diffusion properties
of the 1·2 complex, for which a diffusion coefficient of
7.721ꢃ10ꢁ10 m2 sꢁ1 was found that corresponds to a mass of
493 Da, suggesting that a heterodimer is present.
1
all H NMR titrations, we performed fitting using a simple
1:1 model incorporating no additional equilibria; data are
collated in Table 1. In the case of compounds 1 and 2, we
measured an association constant Ka =33400ꢃ8000mꢁ1, by
curve fitting using HypNMR[63] (Figure 3b), which corre-
sponds to a standard free energy change DG=ꢁ25.8ꢃ
1.1 kJmolꢁ1. This value is fourfold higher than that deter-
mined in our original study,[22] and was discussed in our sub-
sequent study on electronic substituent effects[21]—briefly,
we attribute the enhanced affinity to 1) the improved purifi-
cation of 1 and 2) more rigorous drying of the NMR titra-
r ¼ kT=6phD
ð1Þ
ð2Þ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3
MW
D
D
½highꢂ
½lowꢂ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼
3
MW
½highꢂ
½lowꢂ
Table 1. Binding constants determined by 1H NMR titration (300 MHz,
CDCl3). The values determined are the average of three titrations with
the standard deviation given in brackets.
1H NMR titrations: Upon titration of a solution of 1 into 2,
complexation-induced shifts in the aromatic resonances
were observed (Figure 3a); however, the NH resonances
were broadened and not observed throughout the experi-
ment. In the early part of the titration, broadening of proton
K [mꢁ1
]
K [mꢁ1
]
1·2
1·3
33400 (ꢃ8000)
84 (ꢃ10)
1·4
1·5
17289 (ꢃ2205)
2436 (ꢃ368)
Chem. Eur. J. 2011, 17, 14508 – 14517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14511

A. J. Wilson et al.
1
tion solvent (noting that impurities can have a dramatic
effect on the binding affinities determined by H NMR titra-
the H NMR experiments, standard analysis of the data by
subtraction of dilution curves from titration curves gave
data that could not be fitted with standard MicroCal models
available in the Origin software. We therefore resorted to a
global analysis of all titration and 1 dilution data in the
SEDPHAT software programme using a mixed association
model that allows for guest dimerisation coupled to host–
guest binding (see Experimental Section for additional de-
tails).[66] This data analysis provided a rigorous determina-
tion of the binding parameters for the interaction of 1 with
2, giving values of Ka =24700ꢃ3800mꢁ1, DHa =ꢁ41.9ꢃ
1
tion[64,65]). For titration of ureidopyridine (3) into 2, the NH
resonances were observed and larger complexation induced
shifts for these were also observed, but the variation with
concentration was less dramatic and a much smaller associa-
tion constant (Ka =84ꢃ10mꢁ1) was obtained, consistent with
previous observations for related compounds.[45] This pro-
vides further evidence that 1 indeed operates as a confor-
mer-independent array.
The binding affinities of 1·4 and 1·5 were determined
1
using the same H NMR titration method. The interaction
2.1 kJmolꢁ1
and
K
dim =519ꢃ72mꢁ1,
DHdim =ꢁ28.3ꢃ
of 1·4 occurs with a Ka of 17289ꢃ2205mꢁ1, corresponding to
a standard free energy change of DG=ꢁ24.2ꢃ0.3 kJmolꢁ1,
which is comparable to that of 1·2. The interaction of 1·5
was much weaker, with a Ka of 2436ꢃ368mꢁ1, correspond-
ing to a standard free energy change of DG=ꢁ19.3ꢃ
0.4 kJmolꢁ1. We attributed the lower binding affinity ob-
served for interaction of 1·5 to the steric bulk of the methyl
group in the 7-position of the ring in 5 on the basis of our
molecular modelling results, vide infra; however, an unfav-
ourable inductive effect from this methyl group cannot be
excluded, although alkyl groups were not found to exert a
large effect on binding affinities in our earlier work.[21]
1.3 kJmolꢁ1. The association constant corresponds to DGa =
ꢁ24.6 kJmolꢁ1 and DSa =58.9 JKꢁ1 molꢁ1. The analysis re-
veals, as expected, an enthalpy-driven process. Furthermore,
the association constant obtained from ITC is comparable
1
to that determined by H NMR spectroscopy. The dimerisa-
tion constant that was extracted however, is one order of
magnitude different between the ITC and NMR measure-
ments—this may reflect the small signal changes observed
and limited concentration range accessible in the NMR ex-
periment due to solubility limitations.
1
Molecular modelling: Because the 2D H NMR experiments
could not conclusively reveal the preferred conformation of
individual components or provide sufficient information to
rationalise differences in binding affinity across the series,
we turned to molecular modelling. Gaussian 03[67] was em-
ployed to perform density functional theory (DFT) calcula-
tions, which are able to predict the geometry of complexes
that interact using intermolecular hydrogen-bonds. Gas-
phase calculations were run by
Isothermal titration calorimetry (ITC): To obtain further in-
formation on the nature of interaction we attempted ITC
experiments. Titrations between 1 and 2 provided usable en-
dotherms for curve fitting. Figure 4 provides the titration
and dilution curves for the titration of 1 into 2 alongside the
reverse experiment (i.e., 2 into 1). In a similar manner to
using the hybrid B3LYP func-
tional. For reliable treatment
of dispersive interactions and
hydrogen bonding, a basis in-
cluding diffuse and polarisation
functions was required. The
complexes were too large to
allow the use of the required
6-31+ +G** global basis set.
To overcome this limitation
whilst treating the non-cova-
lent interactions, we applied a
mixed basis wherein C atoms
were modelled with 6-31G, N/
O atoms with 6-31G* and H
atoms had 6-31+ +G**. This
allowed
the
heteroatoms,
which are involved in hydrogen
bonding, to be polarised, but
the carbon atoms, which are
not directly involved in the for-
mation of hydrogen bonds, to
be calculated using a much
smaller basis set, reducing the
computational load. The accu-
Figure 4. a) Titration of 1 into a solution of 2 (black line) and dilution of guest 1 into chloroform (grey line).
b) Titration of 2 into a solution of 1 (black line) and dilution of 1 into chloroform (grey line). c) and d) Inte-
grated heat changes for titration (squares) and dilution (circles) experiments with lines of best fit derived from
a global analysis of five titration curves and three dilution curves performed in SEDPHAT using a model that
allows for 1·2 binding coupled to dimerisation of 1.
14512
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 14508 – 14517

Molecular Recognition
FULL PAPER
racy of this custom basis set was tested by comparing it to a
larger one that included polarisation on all of the atoms (C,
6-311G*; N/O, 6-311G*; H, 6-311+ +G**) with the individ-
ual molecules (rather than the full complexes). Minimal rel-
ative energy differences were found between the two basis
sets when used on our hydrogen-bonding units, indicating
that the small basis was not deficient in any critical functions
and should provide a reasonable estimate of the binding en-
ergies. Firstly, the geometries and minimum energies of the
individual molecules were optimised. (Note: tBu of 1 was re-
moved and the hexyl group of 4 reduced to Me in order to
simplify the calculations.) It was found that both conforma-
tions of 1 that present the DDA array, stabilised by an intra-
molecular hydrogen-bond, have similar energies. A differ-
ence of 1.9 kJmolꢁ1 was calculated, with 1ii being slightly
more stable than 1i (Figure 5a). There are several plausible
tautomers/conformers of 2. We excluded pyrimidinol tauto-
mers on the grounds that these are likely to be disfavoured
through loss of an intramolecular hydrogen-bond and fo-
cused only on the desired DAA conformation and the tauto-
mer found in the crystal structure. It was found that the
solid-state structure is more stable (30.8 kJmolꢁ1) than the
tautomer that presents the desired DAA hydrogen-bond
array (Figure 5b). Given that both structures possess an in-
tramolecular hydrogen-bond, the reason for this difference
in energy is not immediately evident. The structure observed
in the solid state might be favoured because this structure
minimises the overall dipole. The energies of four possible
rotamers of 4 and 5 were calculated. It was found that for
both molecules, the lowest energy conformation presents
the desired AAD array (Figure 5b), which is in agreement
with the 1H NMR structural characterisation.
for the global minimum found. (Note: Because all calcula-
tions were carried out in the gas phase, it is unsurprising
that larger binding affinities were obtained when compared
to the experimental results because no solvation energies
were included in the calculations.) The difference in energy
between the dimer and that of the sum of the monomers
gives the calculated association constant. The energies of
each molecule binding to both conformers of 1 were calcu-
lated and it was found that only small differences in binding
affinity were obtained between 1i and 1ii (Figure 6). This
shows that both conformations of 1 interact with 2, 4 and 5
in agreement with the 1H-1H NOESY data (obtained for
1·2) and further emphasising the conformer independent
nature of the 1 DDA array.
A noteworthy feature of the calculations is that the aro-
matic ring of 2 twists partially out of conjugation to accom-
modate binding of the terminal acceptor within the partner
DAA array. This steric feature clearly has a detrimental
effect on binding over the entire series. Despite the rigour
with which the energies were calculated, caution must still
be exercised in comparing these values with experimentally
The structures of the complexes were optimised with
counterpoise correction, in order to overcome the basis set
superposition error, using the custom basis set previously
described. Energies are counterpoise-corrected and given
Figure 6. Energies of the complexes were calculated. The two conforma-
tions of 1 (1i and 1ii) have similar binding affinities with all three com-
plementary molecules, showing that it is a conformer independent array.
Complex 1·2 has the strongest association constant, 1·4 is slightly lower
and 1·5 is much lower. The methyl group in 5 sterically hinders the bind-
ing of both conformations of 1, resulting in a lower association constant
for complex 1·5.
Figure 5. a) Both conformations of 1 have similar energies. There is
1.9 kJmolꢁ1 difference with 1ii being slightly lower in energy. b) The
lowest energy conformers/tautomers of 2, 4 and 5. Compound 2 does not
present the desired AAD array, but 4 and 5 do.
Chem. Eur. J. 2011, 17, 14508 – 14517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14513

A. J. Wilson et al.
determined Kaꢂs, however, several conclusions can be drawn.
The overall trend in Kaꢂs is consistent across experiment and
theory that is, 1·2>1·4>1·5; however, experimentally we
found that 1·2 and 1·4 had very similar binding affinities,
whereas a much larger difference was observed for the cal-
culations. A plausible explanation stems from the structural
and theoretical studies on individual components—com-
pound 2 is likely to be present in an undesired conformation
in solution which is less stable by about 30 kJmolꢁ1 and this
factor is not accounted for in the calculated binding ener-
gies. Compound 5, is calculated to have a much lower asso-
ciation constant towards 1 in good agreement with our ex-
perimental results. The calculated structure reveals that nei-
ther conformation of 1 can bind to 5 without steric hin-
drance from the methyl group on the naphthalene ring and
so formation of three strong hydrogen-bonds is not possible.
Discussion
Compound 1 indeed appears to exhibit conformer-independ-
1
ent binding properties as evidenced by H NMR studies and
molecular modelling. It therefore serves as a useful motif
with which to study interactions with complementary DAA
arrays. In this study, we were particularly interested in the
analysis of any differences between the binding properties
of 2 and 4 given the evident parallels with the self-comple-
mentary ureidoisocytosine (13) series developed by Meijer
and Sijbesma[12] and the self-complementary ureidocytosine
(14) developed by the Hailes group.[36] The main conforma-
tional features of each array are summarised in Scheme 3.
In series 13, a number of different tautomers are possible,
all of which are pre-organised by intramolecular hydrogen-
bonds—the principle of maximum site occupancy dictates
that quadruple hydrogen-bonded homodimers predominate,
with the pyrimidinone form biased in apolar solvents. For
the 14 series, the tautomeric variation is lost, but an intra-
molecular hydrogen-bond can form that leads to an array
that can dimerise through two rather than four hydrogen
bonds. Ultimately, this has a minor effect on the binding af-
finity and series 14 still undergoes strong self-association
(with the quadruple hydrogen-bonded form perhaps biased
through additional CH···O interactions). This last observa-
tion appears to contravene Etterꢂs rules;[68] it seems that
compounds do not necessarily form intramolecular hydro-
gen-bonds first before using the remaining hydrogen-bond-
ing atoms to make intermolecular contacts, but instead form
the maximum number of hydrogen bonds, whether they in-
clude intramolecular contacts or not. Intuitively, in the triple
hydrogen-bond systems described here there is a tendency
to assume that 2 will be the higher affinity partner for 1 in
comparison to 4, because the intramolecular hydrogen-bond
should rigidify/pre-organise the binding conformation. How-
ever, for 2, like 13, intramolecular hydrogen-bonding is pos-
sible in all of the alternative tautomeric states, one of which
is preferred and (in contrast to 13) non-productive for high
affinity dimerisation. In contrast, no alternative tautomeric
Scheme 3. Conformational and tautomeric summary for 2, 4, 13 and 14.
conformations are possible for 4 and the desired conformer
predominates (presumably due to repulsion between the ni-
trogen and carbonyl groups and/or weak CH···O interac-
tion). Both 2 and 4 exhibit similar binding affinities towards
1 and a number of features may be relevant: 1) the intramo-
lecular hydrogen-bond formed for the binding conformation
of 2 results in a) a more rigid binding conformation and b)
stronger hydrogen-bonding due to inductive effects[69] rela-
tive to 4, but this is offset against the cost of adopting the
desired tautomeric configuration and 2) the overall driver in
the system is to minimise the total free energy and this can
be best achieved through formation of three intermolecular
hydrogen-bonds in each case, overriding conformational
preferences. For instance, such behaviour is observed for the
ureidoisocytosine/diamido-1,8-naphthyridine system,[70] in
which the arrangement of donors and acceptors is less fa-
vourable in terms of secondary electrostatics.
Conclusion
We have reported the synthesis of a series of hydrogen-
bonding motifs (2, 4 and 5) that are capable of interacting
with conformer-independent 1 through three hydrogen-
bonds. Firstly, we have shown that both conformations of 1
14514
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 14508 – 14517

Molecular Recognition
FULL PAPER
100 mL) and saturated aqueous sodium chloride (2ꢃ100 mL). The organ-
ics were dried (Na2SO4) and evaporated in vacuo to give the crude prod-
uct which was purified by column chromatography (EtOAc) to give 4
(214 mg, 15%) as a colourless powder; Rf =0.58 (EtOAc). 1H NMR
(500 MHz, [D6]DMSO): d=11.17 (s, 1H; NH), 8.19 (d, J=7.3 Hz, 1H;
ArCH), 8.01 (d, J=8.1 Hz, 2H; 2ArCH), 7.68 (t, J=8.1 Hz, 1H; ArCH),
7.57 (t, J=8.1 Hz, 2H; 2ArCH), 7.32 (d, J=7.3 Hz, 1H; ArCH), 3.85 (t,
J=7.3 Hz, 2H; CH2), 1.66 (m, 2H; CH2), 1.29 (m, 6H; 3CH2), 0.91 ppm
(t, J=7.3 Hz, 3H; CH3); 13C NMR (75 MHz, CDCl3): d=162.0, 148.8,
133.2, 129.1, 127.5, 96.3, 51.1, 31.3, 28.9, 26.2, 22.5, 14.0 ppm; IR (neat):
have similar binding affinities towards a range of DAA
arrays, establishing it as a generally applicable conformer-in-
dependent array. Secondly, our analyses confirm that intra-
molecular hydrogen-bonding, tautomeric configuration and
unfavourable electrostatic interactions play a key role in de-
termining the preferred conformation of such arrays, but
that this does not necessarily correlate with binding affinity
towards complementary partners. Instead steric effects and
maximising the number of hydrogen bonds play a more
dominant role as is clearly exemplified for other examples
of linear arrays employed for supramolecular polymer syn-
thesis.[30,70] Finally, the results suggest that intramolecular hy-
drogen-bonding has a positive effect on binding affinity to-
wards complementary arrays. Our future studies will aim to
test this hypothesis by introducing preorganising hydrogen-
bonding functionality without introducing variable tauto-
meric configurations. In addition, these motifs will be em-
ployed within a variety of self-assembled architectures in-
cluding main chain supramolecular polymers by exploiting
additional lateral interactions to augment hydrogen-bonding
interactions.[71,72]
n˜ =3381 (br), 2954, 1671, 1625 cmꢁ1
; HRMS (ESI): m/z calcd for
C17H21N3O2 +H+: 300.1707 [M+H+]; found: 300.1695; elemental analy-
sis calcd (%) for C17H21N3O2: C 68.2, H 7.1, N 14.0; found: C 67.9, H 7.1,
N 14.2.
2-Amino-7-methylnaphthyridine (12): Was prepared as described previ-
ously,[61] briefly: 2,6-Diaminopyridine (11; 2.99 g, 27.3 mmol) and phos-
phoric acid (35.0 g) were heated to 908C until melted. 4,4-Dimethoxy-2-
butanone (3.75 mL, 28.1 mmol) was then added dropwise over 30 mins
and the reaction mixture was then heated to 1158C for 3 h. After cooling
to room temperature, ammonium hydroxide was added dropwise until
pH>10. The resultant mixture was extracted with chloroform (3ꢃ
100 mL) and the combined organics were washed with saturated aqueous
sodium chloride (2ꢃ100 mL), dried (Na2SO4) and concentrated in vacuo
to give 12 (3.32 g, 82%) as a red-brown solid and used in subsequent re-
action without further purification. 1H NMR (500 MHz, CDCl3): d=7.83
(d, J=8.0 Hz, 2H; 2ArCH), 7.08 (d, J=8.0 Hz, 1H; ArCH), 6.71 (d, J=
8.7 Hz, 1H; ArCH), 4.95 (s, 2H; NH2), 2.17 ppm (s, 3H; CH3).
2-Benzamido-7-methylnaphthyridine (5): Benzoyl chloride (2.00 mL,
17.2 mmol) was added dropwise to a solution of 12 (2.10 g, 14.1 mmol)
and triethylamine (2.96 mL, 21.1 mmol) in CH2Cl2 (50 mL) and stirred at
room temperature for 16 h. The reaction mixture was then diluted with
CH2Cl2 (50 mL) and washed with saturated aqueous ammonium chloride
(3ꢃ50 mL) and saturated aqueous sodium chloride (2ꢃ50 mL). The or-
ganics were dried (Na2SO4) and evaporated in vacuo. The resultant solid
was purified by column chromatography (0:1!1:19 MeOH/EtOAc) and
then crystallised (CH2Cl2/hexane) to give 5 (1.66 g, 47%) as a colourless
powder. Rf =0.68 (1:9 MeOH/CH2Cl2); m.p. 160–1628C; 1H NMR
(500 MHz, CDCl3): d=8.98 (s, 1H; NH), 8.66 (d, J=8.8 Hz, 1H; ArCH),
8.21 (d, J=8.8 Hz, 1H; ArCH), 8.05 (d, J=8.2 Hz, 1H; ArCH), 7.97 (d,
J=7.0 Hz, 2H; ArCH), 7.51–7.61 (m, 3H; 3ArCH), 7.31 (d, J=8.2 Hz,
1H; ArCH), 2.79 ppm ( s, 3H;CH3); 13C NMR (75 MHz, [D6]DMSO):
d=166.7, 162.5, 154.5, 154.3, 139.1, 136.9, 133.9, 132.3, 128.5 (ꢃ 2), 128.3
(ꢃ 2), 121.6, 118.3, 115.0, 25.3 ppm; IR (neat): n˜ =3225, 1677, 1610, 1579,
Experimental Section
General considerations: All reagents were purchased from Aldrich or
Alfa-Aesar and used without further purification unless otherwise stated.
Yields are not optimised. Purification by column chromatography was
carried out using Merck Kieselgel 60 silica gel. Analytical thin-layer chro-
matography (TLC) was conducted using Merck Kieslegel 0.25 mm silica
gel pre-coated aluminium plates with fluorescent indicator active at
UV254. Where anhydrous solvents were required, THF was freshly dis-
tilled from sodium-benzophenone ketyl radical, CH2Cl2 freshly distilled
from calcium hydride and CHCl3 freshly distilled from calcium chloride
under a nitrogen atmosphere. Triethylamine was distilled from calcium
hydride under a nitrogen atmosphere and stored, under nitrogen, over
potassium hydroxide pellets. Anhydrous DMF was obtained “sure-
sealed” from Sigma–Aldrich. All melting points reported were measured
using a Griffin D5 variably temperature apparatus and are uncorrected.
NMR spectra were obtained using Bruker AMD300 or Bruker DMX500
spectrometers operating at 500 or 300 MHz for 1H spectra and 100 or
75 MHz for 13C spectra as stated. NMR solvent was CDCl3 unless other-
wise stated. 1H NMR spectra are referenced to tetramethylsilane (TMS)
and all chemical shifts are displayed in parts per million relative to TMS
with all coupling constants being reported to the nearest 0.1 Hz. IR spec-
tra were obtained on a Perking-Elmer FTIR spectrometer using attenuat-
ed total reflectance (ATR) apparatus. Mass spectroscopy and microanaly-
sis were carried out at the University of Leeds. Compounds 1, 2, 3, 7 and
8 were prepared as described previously by our group.[21,22]
1505, 1437, 1391, 1322, 1272 cmꢁ1
; HRMS (ESI): m/z calcd for
C16H13N3O+H+: 264.1129 [M+H+]; found: 264.1131;
NOESY data acquisition: 1H-1H NOESY data was collected on the indi-
vidual components and their 1:1 mixtures at 10 mmol concentrations in
CDCl3. Spectra were recorded on a Bruker Avance500 instrument oper-
ating at 300 K at a frequency of 500 MHz. A pulse sequence of 8.2 ms
pulse, 2.0 s delay, 16.4 ms pulse, 1.2 s delay was used.
DOSY data acquisition: DOSY NMR measurements were made on a
Varian Inova 500 MHz spectrometer. All experiments were conducted at
208C on CDCl3 solutions, and used a 5 mm ID probe. The bipolar pulse
pair simulated echo (BPPSTE) sequence[73] was employed operating in
ONESHOT mode.[74] Additional parameters: number of different gradi-
ent strengths, 20; gradient stabilisation delay 0.002 s; gradient length
0.002 s; diffusion delay 0.03 s; relaxation delay 2.5 s (following measure-
ment of T1); acquisition time 2 s; kappa (unbalancing factor) 0.2. Data
were processed using a 3 Hz line broadening and exponential multiplica-
tion. Data were zero-filled once. Spectra were phased and baseline-cor-
rected prior to production of the pseudo 2D DOSY plots. For measure-
ment of diffusion co-efficients, a calibration curve was plotted (using the
N-Hexylbenzamidocytosine (4): Triethylamine (1.04 mL, 7.46 mmol) and
benzoyl chloride (0.69 mL, 5.97 mmol) were added to a solution of cyto-
sine
9
(0.55 g, 4.97 mmol) and 4-dimethylaminopyridine (0.61 g,
0.50 mmol) in chloroform (55 mL) and heated to reflux for 16 h. The re-
action mixture was then allowed to cool before being filtered. The resul-
tant solid was washed with chloroform (100 mL) and water (100 mL)
before being dried in vacuo to give benzamidocytosine 10 (1.00 g) as the
crude product which was taken onto the next stage without further purifi-
cation. Potassium carbonate (1.93 g, 13.9 mmol) and 1-bromohexane
(0.71 mL, 5.12 mmol) were added to a suspension of benzamidocytosine
(1.00 g) in DMF (80 mL) and the mixture was heated to 808C for 20 h.
The reaction mixture was then cooled to room temperature before the
solid was filtered. The filtrate was diluted with EtOAc (200 mL) and
washed with 10% aqueous hydrochloric acid (3ꢃ100 mL), water (3ꢃ
Stokes–Einstein relationship) for diffusion coefficient (ꢃ106 cm2 sꢁ1
)
versus the reciprocal cube root of the molecular mass (1/(molecular
mass)1/3). The compounds used for this calibration exercise were 4-amino-
pyrimidine (95.11 Da, 17.281ꢃ10ꢁ6 cm2 sꢁ1), dipyridyl (184.24 Da, 13.573),
4-(2,’,6’,2’’-terpyridinyl)benzyl (323.4 Da, 10.197), benzene-1,3,5-tricar-
boxylic acid tris(6-methyl-2-pyridinyl)amide (480.0 Da, 11.503), hexa-
Chem. Eur. J. 2011, 17, 14508 – 14517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14515

A. J. Wilson et al.
kis(2-pyridlymethly)CTV (CTV=cyclotriveratrylene) (913.11 Da, 8.703),
tris[4-(2-pyridyl)benzyloxy]CTG (CTG= cyclotriguaiacylene) (910.14 Da,
9.037). The straight line determined was y=73.275x+1.1263 (R2 =0.93).
Molecular modelling: Ab initio density-functional calculations were car-
ried out using the Gaussian 03 Package and performed at the level of
theory stated in the text.
1
1H NMR titration experiments: For H NMR titrations anhydrous CDCl3
was purchased from Aldrich and prior to use distilled over CaCl2 under a
stream of nitrogen to remove residual water and other impurities then
used immediately (Note 1, below). For titrations the 1H NMR spectrum
was recorded for a solution of host (2–10 mm) in CDCl3 and the change
in chemical shift of key proton resonances was recorded upon sequential
additions of a solution of guest (20–120 mm) containing host (2–10 mm)
in CDCl3. The data was subsequently analysed using the HypNMR pro-
gram using the appropriate model.[63] HypNMR uses data from multiple
resonances for curve fitting. Representative data for individual resonan-
ces is shown in the supplementary data. Each titration was performed in
triplicate and the mean binding constant given with the standard devia-
tion as the error. Dimerisation constants where applicable were calculat-
Acknowledgements
This work was supported by the Leverhulme Trust (F/00122/AN) and
EPSRC (EP/C007638/1).
[1] S. C. Zimmerman, P. S. Corbin, Struct. Bonding (Berlin) 2000, 96,
63–94.
[2] G. Cooke, V. M. Rotello, Chem. Soc. Rev. 2002, 31, 275–286.
[3] J. L. Sessler, J. Jayawickramarajah, Chem. Commun. 2005, 1939–
1949.
1
ed by taking H NMR spectra of a solution of the chosen compound as it
was diluted from about 20 mm to around 0.1 mm. The change in chemical
shift of key proton resonances was recorded at each dilution point. Data
were analysed in HypNMR (see supplementary data) using the appropri-
ate model (Note 2).
[4] S. Sivakova, S. J. Rowan, Chem. Soc. Rev. 2005, 34, 9–21.
[5] A. J. Wilson, Soft Matter 2007, 3, 409–425.
[6] L. J. Prins, D. N. Reinhoudt, P. Timmerman, Angew. Chem. 2001,
113, 2446–2492; Angew. Chem. Int. Ed. 2001, 40, 2382–2426.
[7] T. F. A. De Greef, M. M. J. Smulders, M. Wolffs, A. P. H. J. Schen-
ning, R. P. Sijbesma, E. W. Meijer, Chem. Rev. 2009, 109, 5687–
5754.
[8] W. Binder, R. Zirbs in Advances in Polymer Science, Vol. 207: Hy-
drogen Bonded Polymers (Ed.: W. Binder), Springer, Berlin/Heidel-
berg, 2007, pp. 1–78.
[9] K. Nakatani, S. Sando, I. Saito, Nat. Biotechnol. 2001, 19, 51–55.
[10] J. F. Arambula, S. R. Ramisetty, A. M. Baranger, S. C. Zimmerman,
Proc. Natl. Acad. Sci. USA 2009, 106, 16068–16073.
[11] H.-J. Schneider, R. K. Juneja, S. Simova, Chem. Ber. 1989, 122,
1211–1213.
Note 1: It has been noted that the presence of HCl can have a significant
effect on association constants—we find in our hands that the use of po-
tassium carbonate during distillation or its addition following distillation
results in weaker association constants as a result of higher water content
1
that is observable in the H NMR spectrum.
Note 2: The self-dimerisation constants were sufficiently small for all
compounds that they were not included in the determination of the asso-
ciation constant for heterodimer formation. In our earlier study we in-
cluded self-dimerisation in our treatment of the data (Note 3).
Note 3: A fourfold difference in the dimerisation constant between our
original report[22] and subsequent papers[21] including this one is report-
ed—this is in part ascribed to the different treatment of the data as out-
lined in Note 2, but also an improved purification procedure for 1 that re-
moves traces of diphenyl urea by-product, not observed by 1H NMR
spectroscopy.
[12] F. H. Beijer, R. P. Sijbesma, H. Kooijman, A. L. Spek, E. W. Meijer,
J. Am. Chem. Soc. 1998, 120, 6761–6769.
[13] R. P. Sijbesma, E. W. Meijer, Chem. Commun. 2003, 5–16.
[14] G. B. W. L. Ligthart, H. Ohkawa, R. P. Sijbesma, E. W. Meijer, J.
Org. Chem. 2006, 71, 375–378.
ITC experiments: Titrations were performed at 258C in a MicroCal VP-
ITC with a cell volume of 1.4109 mL. The concentration of receptor in
the cell was 0.2 mm and the concentration of titrant in the syringe was
2.5 mm. For data fitting, the standard MicroCal Origin software was used
for initial data processing to calculate host and guest concentrations and
to integrate heat changes for each step in the titrations. For reverse titra-
tions (2 in syringe and 1 in the cell), the heat associated with diluting the
host was subtracted from the titration data prior to curve fitting. Dilution
experiments with 1 in the syringe gave non-linear curves that are indica-
tive of significant self-association at the concentrations used for the titra-
tions. Therefore, rather than subtracting the heat of 1 dilution on a point-
by-point basis, several direct and reverse titrations, along with 1 dilution
experiments, were subjected to a global analysis in the SEDPHAT soft-
ware[66] using a mixed association model that allows for 1 dimerisation
coupled to 1·2 binding. The model was constrained to exclude the possi-
bility of 2 dimerisation or the formation of any ternary or quaternary
complexes. First, 2 into 1 titration was fit to a simple 1:1 binding model
to obtain an initial estimate for Ka and DHa for the 1·2 interaction. These
values were fixed while three dilution experiments for 1 were simultane-
ously fit to the mixed association model to provide initial estimates for
Kdim and DHdim. A wide range of initial parameter values and repetitive
switching between simplex and Marquardt–Levenberg fitting methods
were required to avoid local minima at this stage in the fitting process.
Then five titration experiments (three 1 into 2 titrations, and two 2 into 1
titrations) were fit to the global model, while keeping Kdim and DHdim
fixed. Any apparent deviations from 1:1 binding were accommodated by
using the “incompetent fraction” parameter for each experiment to esti-
mate any errors in concentrations of the 1 or 2.[66] Finally, Ka, DHa, Kdim
and DHdim were all allowed to float freely in a global fit of eight datasets.
Parameter errors (assuming Gaussian-distributed errors) were estimated
using Monte Carlo simulations[75] as implemented in the SEDPHAT soft-
ware.[66]
[15] T. J. Murray, S. C. Zimmerman, J. Am. Chem. Soc. 1992, 114, 4010–
4011.
[16] W. L. Jorgensen, J. Pranata, J. Am. Chem. Soc. 1990, 112, 2008–
2010.
[17] L. Brunsveld, B. J. B. Folmer, E. W. Meijer, R. P. Sijbesma, Chem.
Rev. 2001, 101, 4071–4098.
[18] Y. Kyogoku, R. C. Lord, A. Rich, Proc. Natl. Acad. Sci. USA 1967,
57, 250–257.
[19] P. S. Corbin, S. C. Zimmerman, J. Am. Chem. Soc. 1998, 120, 9710–
9711.
[20] V. G. H. Lafitte, A. E. Aliev, H. C. Hailes, K. Bala, P. Golding, J.
Org. Chem. 2005, 70, 2701–2707.
[21] A. Gooch, A. M. McGhee, M. L. Pellizzaro, C. I. Lindsay, A. J.
Wilson, Org. Lett. 2011, 13, 240–243.
[22] A. M. McGhee, C. Kilner, A. J. Wilson, Chem. Commun. 2008, 344–
346.
[23] A. Gooch, A. M. McGhee, L. C. Renton, J. P. Plante, C. I. Lindsay,
A. J. Wilson, Supramol. Chem. 2009, 21, 12–17.
[24] E. M. Todd, J. R. Quinn, T. Park, S. C. Zimmerman, Isr. J. Chem.
2005, 45, 381–389.
[25] F. H. Beijer, H. Kooijman, A. L. Spek, R. P. Sijbesma, E. W. Meijer,
Angew. Chem. 1998, 110, 79–82; Angew. Chem. Int. Ed. 1998, 37,
75–78.
[26] U. Lꢄning, C. Kuhl, Tetrahedron Lett. 1998, 39, 5735–5738.
[27] B. Gong, Y. Yan, H. Zeng, E. Skrzypczak-Jankuun, Y. W. Kim, J.
Zhu, H. Ickes, J. Am. Chem. Soc. 1999, 121, 5607–5608.
[28] U. Lꢄning, C. Kꢄhl, A. Uphoff, Eur. J. Org. Chem. 2002, 4063–4070.
[29] S. Brammer, U. Luning, C. Kuhl, Eur. J. Org. Chem. 2002, 4054–
4062.
[30] X.-Z. Wang, X.-Q. Li, X.-B. Shao, X. Zhao, P. Deng, X.-K. Jiang, Z.-
T. Li, Y.-Q. Chen, Chem. Eur. J. 2003, 9, 2904–2913.
14516
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 14508 – 14517

Molecular Recognition
FULL PAPER
[31] T. Park, S. C. Zimmerman, S. Nakashima, J. Am. Chem. Soc. 2005,
127, 6520–6521.
[58] D. L. Helfer, R. S. Hosmane, N. J. Leonard, J. Org. Chem. 1981, 46,
4803–4804.
[59] T. Krꢄger, C. Bruhn, D. Steinborn, Org. Biomol. Chem. 2004, 2,
2513–2516.
[32] P. K. Baruah, R. Gonnade, U. D. Phalgune, G. J. Sanjayan, J. Org.
Chem. 2005, 70, 6461–6467.
[60] V. G. H. Lafitte, A. E. Aliev, E. Greco, K. Bala, P. Golding, H. C.
Hailes, New J. Chem. 2011, 35, 1522–1527.
[33] P. Prabhakaran, V. G. Puranik, G. J. Sanjayan, J. Org. Chem. 2005,
70, 10067–10072.
[61] E. E. Fenlon, T. J. Murray, M. H. Baloga, S. C. Zimmerman, J. Org.
Chem. 1993, 58, 6625–6628.
[62] A. M. McGhee, J. P. Plante, C. A. Kilner, A. J. Wilson, Supramol.
Chem. 2011, 23, 470–479.
[63] C. Frassineti, S. Ghelli, P. Gans, A. Sabatini, M. S. Moruzzi, A.
Vacca, Anal. Biochem. 1995, 231, 374–382.
[34] H. C. Ong, S. C. Zimmerman, Org. Lett. 2006, 8, 1589–1592.
[35] A. M. Martin, R. S. Butler, I. Ghiviriga, R. E. Giessert, K. A.
Abboud, R. K. Castellano, Chem. Commun. 2006, 4413–4415.
[36] V. r. G. H. Lafitte, A. E. Aliev, P. N. Horton, M. B. Hursthouse, K.
Bala, P. Golding, H. C. Hailes, J. Am. Chem. Soc. 2006, 128, 6544–
6545.
[64] J. L. Cook, C. A. Hunter, C. M. R. Low, A. Perez-Velasco, J. G.
Vinter, Angew. Chem. 2007, 119, 3780–3783; Angew. Chem. Int. Ed.
2007, 46, 3706–3709.
[65] J. L. Cook, C. A. Hunter, C. M. R. Low, A. Perez-Velasco, J. G.
Vinter, Angew. Chem. 2008, 120, 6371–6373; Angew. Chem. Int. Ed.
2008, 47, 6275–6277.
[66] J. C. D. Houtman, P. H. Brown, B. Bowden, H. Yamaguchi, E. Ap-
pella, L. E. Samelson, P. Schuck, Protein Sci. 2007, 16, 30–42.
[67] Gaussian 03, Revision D.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
[37] Y. Yang, H.-J. Yan, C.-F. Chen, L.-J. Wan, Org. Lett. 2007, 9, 4991–
4994.
[38] J. Taubitz, U. Luning, Aust. J. Chem. 2009, 62, 1550–1555.
[39] Y. Hisamatsu, N. Shirai, S.-i. Ikeda, K. Odashima, Org. Lett. 2009,
11, 4342–4345.
[40] Y. Hisamatsu, N. Shirai, S.-i. Ikeda, K. Odashima, Org. Lett. 2010,
12, 1776–1779.
[41] B. A. Blight, C. A. Hunter, D. A. Leigh, H. McNabb, P. I. T. Thom-
son, Nat. Chem. 2011, 3, 246.
[42] P. Zhang, H. Chu, X. Li, W. Feng, P. Deng, L. Yuan, B. Gong, Org.
Lett. 2011, 13, 54–57.
AHCTUNGTREGeNNUN ry, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, 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.
[43] S. K. Chang, A. D. Hamilton, J. Am. Chem. Soc. 1988, 110, 1318–
1319.
[44] S. K. Chang, D. Van Engen, E. Fan, A. D. Hamilton, J. Am. Chem.
Soc. 1991, 113, 7640–7645.
[45] P. S. Corbin, S. C. Zimmerman, P. A. Thiessen, N. A. Hawryluk, T. J.
Murray, J. Am. Chem. Soc. 2001, 123, 10475–10488.
[46] X. Yang, S. Martinovic, R. D. Smith, B. Gong, J. Am. Chem. Soc.
2003, 125, 9932–9933.
[47] H. Zeng, R. S. Miller, R. A. Flowers II, B. Gong, J. Am. Chem. Soc.
2000, 122, 2635–2644.
[48] S. Djurdjevic, D. A. Leigh, H. McNab, S. Parsons, G. Teobaldi, F.
Zerbetto, J. Am. Chem. Soc. 2007, 129, 476–477.
[49] Y. Hisamatsu, Y. Fukumi, N. Shirai, S.-i. Ikeda, K. Odashima, Tetra-
hedron Lett. 2008, 49, 2005–2009.
[50] B. A. Blight, A. Camara-Campos, S. Djurdjevic, M. Kaller, D. A.
Leigh, F. M. McMillan, H. McNab, A. M. Z. Slawin, J. Am. Chem.
Soc. 2009, 131, 14116–14122.
[51] Y. L. Jiang, P. Patel, S. M. Klein, Bioorg. Med. Chem. 2010, 18,
7034–7042.
[52] M. Herder, M. Patzel, L. Grubert, S. Hecht, Chem. Commun. 2011,
47, 460–462.
[68] M. C. Etter, Acc. Chem. Res. 1990, 23, 120–126.
[69] A. P. Bisson, C. A. Hunter, J. C. Morales, K. Young, Chem. Eur. J.
1998, 4, 845–851.
[70] G. B. W. Ligthart, H. Ohkawa, R. P. Sijbesma, E. W. Meijer, J. Am.
Chem. Soc. 2005, 127, 810–811.
[71] H. Kautz, D. J. M. van Beek, R. P. Sijbesma, E. W. Meijer, Macro-
molecules 2006, 39, 4265–4267.
[53] C. Dohno, S.-n. Uno, K. Nakatani, J. Am. Chem. Soc. 2007, 129,
11898–11899.
[72] S. Sivakova, D. A. Bohnsack, M. E. Mackay, P. Suwanmala, S. J.
Rowan, J. Am. Chem. Soc. 2005, 127, 18202–18211.
[54] W. H. Binder, S. Bernstorff, C. Kluger, L. Petraru, M. J. Kunz, Adv.
Mater. 2005, 17, 2824–2828.
[73] M. D. Pelta, H. Barjat, G. A. Morris, A. L. Davis, S. J. Hammond,
Magn. Reson. Chem. 1998, 36, 706–714.
[55] M. N. Higley, J. M. Pollino, E. Hollembeak, M. Weck, Chem. Eur. J.
2005, 11, 2946–2953.
[74] M. D. Pelta, G. A. Morris, M. J. Stchedroff, S. J. Hammond, Magn.
Reson. Chem. 2002, 40, S147–S152.
[56] O. Uzun, A. Sanyal, H. Nakade, R. J. Thibault, V. M. Rotello, J. Am.
Chem. Soc. 2004, 126, 14773–14777.
[75] L. M. Schwartz, Anal. Chem. 1975, 47, 963–964.
[57] E. Greco, A. E. Aliev, V. G. H. Lafitte, K. Bala, D. Duncan, L.
Pilon, P. Golding, H. C. Hailes, New J. Chem. 2010, 34, 2634–2642.
Received: July 11, 2011
Published online: November 23, 2011
Chem. Eur. J. 2011, 17, 14508 – 14517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14517