Templated photochemical synthesis of a uracil vs. thymine receptor

Article abstract of DOI:10.1039/b103116n

The first templated photochemical synthesis of a receptor capable of differentiating between thymine and uracil is described.

Full text of DOI:10.1039/b103116n

View Article Online / Journal Homepage / Table of Contents for this issue
Templated photochemical synthesis of a uracil vs. thymine receptor
Dario M. Bassani,* Xavier Sallenave, Vincent Darcos and Jean-Pierre Desvergne
Laboratoire de Chimie Organique et Organométallique, CNRS UMR 5802, Université Bordeaux 1,
33505. Talence, France. E-mail: d.bassani@lcoo.u-bordeaux.fr
Received (in Cambridge, UK) 5th April 2001, Accepted 15th June 2001
First published as an Advance Article on the web 20th July 2001
The first templated photochemical synthesis of a receptor
capable of differentiating between thymine and uracil is
described.
Supramolecular self-assembly can be used to build structures of
controlled size, shape, and functionality.1 Of particular interest
is the application of such processes to construct molecular
receptors by moulding small fragments containing molecular
recognition elements around a given substrate.2 However, a
means by which the reversibly-bound adduct can be captured to
allow the isolation and identification of the receptor (lock-in) is
invariably required. We have recently described a system in
which the photoinduced dimerisation of cinnamates appended
Fig. 1 Various hydrogen-bonded assemblies may be formed between 2 and
with molecular recognition units was affected by the presence
of a template molecule.3 A salient feature appeared to be the
possibility to consolidate supramolecular assemblies via light-
triggered photoreactions through a process analogous to the
topochemical control achieved in the solid.4 It therefore seemed
interesting to further explore the use of light to trigger the
covalent capture of reversibly-formed supramolecular re-
ceptors. The binding properties of this new class of hydrogen-
bonding receptors, incorporating a diphenylcyclobutane scaf-
fold, were investigated. In particular, their ability to
discriminate between analogues of the naturally-occurring
pyrimidine bases thymine and uracil makes them potential
candidates in the development of RNA vs. DNA targeting
agents.
Compound 2 was synthesised by Heck coupling between
styrene and 1(Scheme 1), prepared by sequential substitution of
cyanuric chloride† according to literature procedures.5 The
binding of 2 to the template molecule, 5,5-dihexylbarbituric
acid (3), was investigated6 by 1H NMR spectroscopy. In
principle, one may expect the formation of five distinct (two 1+1
and three 2+1) complexes (Fig. 1), only one of which places the
stilbene chromophores in close proximity. Because the inter-
conversion between the hydrogen-bonded complexes is fast on
the NMR timescale, the observed binding isotherm only allows
determination of the overall 1+1 and 2+1 association constants.
Binding of 2 to 3 in CDCl3 is accompanied by a 2 ppm
downfield shift of the N–H resonance of 3, from which
association constants K1 = 1200 60 M21 and K2 = 250 12
M21 can be extracted. A good fit to the experimental data was
3, only one of which places two stilbene chromophores in close proximity,
favouring the formation of syn photodimers.
obtained only for a model comprised of sequential 1+1 and 2+1
complex formation.
Irradiation of dilute (1022 M) CH2Cl2 solutions of 2 results in
rapid E, Z photoisomerisation of the stilbene chromophore
leading to a photostationary equilibrium mixture enriched in the
Z isomer (Z+E = 5.5), known to be unreactive towards
cyclodimerisation.‡ Upon prolonged irradiation, slow grow-in
of five additional products is observed by HPLC (Fig. 2).
Isolation of four photoproducts by preparative HPLC allowed
their characterisation by 1H NMR and mass spectroscopy,
which identified them as [2 + 2] cycloadducts. The structures of
the isolated photoproducts 5a–5d were assigned on the basis of
their spectral properties by comparison with known stilbene
dimers described in the literature.7 The major cycloadducts are
the head-to-head and head-to-tail dimers (5a and 5d), account-
ing for roughly two thirds of the total dimer formation.
Under identical irradiation conditions, the presence of 0.5 eq.
of 3 was found to enhance the formation of three of the
photoproducts (5b, 5c, and 5d), while inhibiting the formation
of 5a. This difference in reactivity is attributed to the formation
of a ternary supramolecular complex in which the stilbene
moieties are held in a face-to-face geometry that is favourable to
dimerisation. The preference for photodimers in which the
Fig. 2 Portion of the HPLC chromatograms showing the formation of
photodimers in the presence (a) and absence (b) of 3 upon irradiation, and
the proposed structures of dimers 5a–5d.
Scheme 1 (a) Styrene, Et3N, Pd2(dba)3 (5 mol%), acetonitrile, 85 °C, 24 h
(71%).
1446
Chem. Commun., 2001, 1446–1447
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b103116n

View Article Online
Table 1 Quantum yieldsa of photodimers 3 103
Dimer
2 Alone
0.5 eq. 3
0.5 eq. 4
5a
5b
5c
5d
0.7
0.1
0.1
0.6
0.5
1.6
0.6
2.4
0.7
0.1
0.1
0.6
a In degassed dichloromethane, [2] = 0.01 M.
Fig. 3 Energy minimised (PM3) structures of complexes formed between 5d
triazine units are oriented syn is in agreement with the
involvement of 3 as a template during the dimerisation reaction.
This is further supported by the lack of activity of 4, in which the
hydrogen bonding sites are blocked by methylation. The
reduced yield of 5a is consistent with the inability of 3 to
promote structures that are not suitable receptors for barbiturate
derivatives. Quantum yields for the formation of dimers 5a–5d
in the presence and absence of 3 or 4 are given in Table 1. From
the measured association constants, and assuming a statistical
distribution of 1+1 and 1+2 complexes, one can estimate the
quantum yield for the formation of dimer 5d within the
supramolecular assembly to be 0.1, approximately a 170-fold
increase with respect to solution. A rationalization of the
catalysis and product distribution for an analogous cinnamate
derivative has already been described,3 and will therefore not be
discussed further.
and uracil (left) and thymine (right).
M21) reflect a more favourable statistical weighting for binding
the first thymine molecule, and a modest anticoopertive effect
towards binding of the second thymine, presumably due to
steric interactions. To adequately compare the binding of 7 or 8
to 5d, one must take into consideration that whereas in the case
of thymine four distinct 1+1 complexes may be formed, only
one complex can be formed between 5d and uracil. Thus, the
microscopic binding constant of thymine is actually only one
half that of uracil (500 vs. 960 M21). An upper limit of 50 M21
was estimated for the association constant between 5d and
9-ethyladenine, suggesting the formation of a rather labile
complex. This is consistent with the binding of adenine in a
fashion similar to that of thymine, but involving only two
hydrogen bonds.
The methyl group in thymine has been recently recognized to
play an important role in the recognition and suppression of
DNA sequences by certain bacteria,11 and receptors capable of
mimicking this form of recognition would be of interest. The
ability to differentiate between uracil and thymine may be
further enhanced by preventing rotation of the binding site in
5d, and this may open new possibilities for the selective
recognition of RNA vs. DNA fragments.
Molecular modelling of 5d indicates that it has a tweezer-like
geometry, with both the aminotriazine groups oriented in the
same direction. Concomitant binding of a substrate to both
triazines is therefore restricted both by the hydrogen-bonding
pattern of triazine and by the size and shape of the cavity.
Titration of 5d with Barbital (5,5-diethylbarbituric acid, 6) in
CDCl3 was monitored using NMR spectroscopy by observing
the barbiturate N–H protons, which underwent a large down-
field shift ( > 4 ppm) upon complexation. The association
constant was found to be 2400 M21, much higher than the
binding of 2 to 3, for which the microscopic binding constant is
We are indebted to Professor M. J. Hynes for making his
program EQNMR available to us, and to the Conseil Régional
Aquitaine and the MRT for supporting this work.
calculated to be 300 M21 8
. The increase in the association
constant is attributed to the binding of 6 within the cleft formed
by the triazine groups in 5d, resulting in the formation of
multiple hydrogen bonds. In this respect, 5d is an example of
substrate-induced receptor synthesis. The magnitude of the
binding constant between 6 and 5d is very similar to that of an
analogous recently synthesised Barbital receptor containing a
ferrocene unit,9 though smaller than those previously reported
by Hamilton and co-workers.10 Thus, the preparation of
receptors via light-induced capture of supramolecular assem-
blies can lead to functional receptors, of similar binding affinity
as those obtained by conventional synthetic methodologies.
The rigid structure of 5d is the basis for the observed
selectivity in the binding of uracil vs. thymine or adenine. NMR
titration (CDCl3) of 5d by 5-(4-tert-butylbenzyl)uracil (7)
results in a binding isotherm indicating the formation of a 1+1
complex with a binding constant of 960 120 M21.§ Molecular
modelling using semi-empirical PM3 calculations (Fig. 3)
reveals that uracil can bind within the diaminotriazine cleft with
formation of 4 hydrogen bonds. In contrast, the presence of the
methyl substituent in thymine is expected to prevent it from
entering the binding cleft, and should therefore result in a lower
binding affinity. This is indeed the case, and the binding of
5-(4-tert-butylbenzyl)thymine (8) to dimer 5d can only be fitted
to a model comprised of sequential 1+1 and 1+2 complex
formation. This is rationalised by the binding of a thymine
molecule to 5d following partial rotation about the C–C bond
connecting the diaminotriazine unit to the cyclobutane scaffold.
Notes and references
1 J.-M. Lehn, Supramolecular Chemistry-Concepts and Perspectives,
VHC, 1995.
2 (a) V. Berl, I. Huc, J.-M. Lehn, A. DeCian and J. Fischer, Eur. J. Org.
Chem., 1999, 3089; (b) V. Berl, M. J. Krische, I. Huc, J.-M. Lehn and
M. Schmutz, Chem. Eur. J., 2000, 6, 1939; (c) I. Huc, M. J. Krische, D.
Funeriu and J.-M. Lehn, Eur. J. Inorg. Chem., 1999, 1415; (d) J. C.
Adrian and C. G. Wilcox, J. Am. Chem. Soc., 1992, 114, 1398; (e) T.
Hayashi, T. Asai, H. Hokazono and H. Ogoshi, J. Am. Chem. Soc., 1993,
115, 12210.
3 D. M. Bassani, V. Darcos, S. Mahony and J.-P. Desvergne, J. Am. Chem.
Soc., 2000, 122, 8795.
4 (a) Y. Ito, Synthesis, 1998, 1, and references therein; (b) D. G. Whitten,
L. Chen, H. C. Geiger, J. Perlstein and X. Song, J. Phys. Chem. B, 1998,
102, 10098.
5 J. A. Zerkowski, J. C. MacDonald, C. T. Seto, D. A. Wierda and G. M.
Whitesides, J. Am. Chem. Soc., 1994, 116, 2382.
6 M. J. Hynes, J. Chem. Soc., Dalton Trans., 1993, 311.
7 H. Schechter, W. J. Link and G. V. D. Tiers, J. Am. Chem. Soc., 1963,
85, 1601; H. Ulrich, D. V. Rao, F. A. Stuber and A. A. R. Sayigh, J. Org.
Chem., 1970, 35, 1121.
8 K. A. Connors, Binding Constants, Wiley Interscience, London,
1987.
9 S. R. Collinson, T. Gelbrich, M. B. Hursthouse and J. H. R. Tucker,
Chem. Commun., 2001, 555.
10 S. K. Chang, D. Van Engen, E. Fan and D. A. Hamilton, J. Am. Chem.
Soc., 1991, 113, 7640.
11 C. S. Chen, A. White, J. Love, J. R. Murphy and D. Ringe, Biochemistry,
2000, 39, 10397.
The binding constants (K1 = 1980 65 M21, K2 = 150
5
Chem. Commun., 2001, 1446–1447
1447