New building blocks for the assembly of sequence selective molecular zippers.

Article abstract of DOI:10.1039/b304041k

Synthetic H-bonded molecular zippers contain no sequence information that can be used to engineer the selective binding interactions characteristic of biopolymers; reversing the sense of the amide bonds in the two binding partners generates a new orthogonal recognition motif and the mutually complementary binding partners form complexes an order of magnitude more stable than the corresponding mismatch complexes.

Full text of DOI:10.1039/b304041k

New building blocks for the assembly of sequence selective molecular
zippers
Christopher A. Hunter,*a Philip S. Jones,b Pascale M. N. Tigera and Salvador Tomasa
a Centre for Chemical Biology, Krebs Institute for Biomolecular Science, Department of Chemistry,
University of Sheffield, Sheffield, UK S3 7HF
b Roche Discovery Welwyn, 40 Broadwater Road, Welwyn Garden City, Herts, UK AL7 3AY
Received (in Cambridge, UK) 11th April 2003, Accepted 13th May 2003
First published as an Advance Article on the web 16th June 2003
Synthetic H-bonded molecular zippers contain no sequence
information that can be used to engineer the selective
binding interactions characteristic of biopolymers; reversing
the sense of the amide bonds in the two binding partners
generates a new orthogonal recognition motif and the
mutually complementary binding partners form complexes
an order of magnitude more stable than the corresponding
mismatch complexes.
The bisaniline–isophthalic acid recognition motif used in the
zippers shows excellent complementarity which is not dissi-
pated in long chain lengths. Thus the best chance for success in
designing the new recognition elements required to encode
sequence information is to make a minimal structural change,
i.e. reverse the sense of the amide bonds. The approach is
illustrated in Fig. 2. The geometry of interaction is identical to
the zipper complexes reported previously (Fig. 1), and the
expectation is that the stability will be similar. However, mixing
the recognition motifs in Figs. 1 and 2 should lead to a
geometric mismatch in which the hydrogen bonding groups will
not be optimally positioned or aligned. This should generate
selective mutually exclusive binding interactions between
complementary partners. To test the validity of this approach,
we have studied the properties of simple model compounds
containing the two different recognition motifs and investigated
all possible pairwise binding interactions (Fig. 3).
Oligomeric materials composed of a linear array of recognition
sites possess some unique properties, as typified by the nucleic
acids. DNA stores genetic information as a sequence of
covalently-linked nucleotides, and its ability to reproduce this
information is based on the self-assembly of two com-
plementary linear strands into a double-stranded complex. H-
Bonding interactions between the bases allow one strand to
‘read’ the sequence of another strand, and this property is
responsible for the high specificity observed in the self-
assembly and self-replication of double-stranded nucleic acids.1
This property has been exploited in molecular and macro-
molecular construction for the formation of complex topologi-
cal objects and the encoded organisation of nanoparticles.2 The
assembly of linear proteins into multi-stranded complexes
through encoded recognition sites is the basis of the behaviour
of biological fibres such as muscle.3 Although there are
examples of synthetic systems which self-assemble into double-
stranded complexes,4–8 at present, there are no examples in
which sequence recognition information is encoded, and so
these systems do not exhibit any of the more interesting and
potentially exploitable properties of their biological counter-
parts.
We have developed a synthetic system where hydrogen-
bonding and aromatic interactions direct the assembly of
double-stranded complexes of amide oligomers, ‘molecular
zippers’ (Fig. 1).9,10 The association constant increases by an
order of magnitude for each increment in the length of the
oligomer, indicating that the interactions along the zipper
complexes are highly cooperative.11 However, these systems
rely entirely on length for their recognition properties and the
selectivity between complementary and non-complementary
arrangements is poor. Here, we describe a new strategy for
introducing sequence information into the zipper structures.
Compounds 1 and 2 have been described elsewhere.12
Compound 3 was prepared from bisaniline 5 (Scheme 1).
Reaction with NaNO2 and HCl gave the diazonium salt which
was subsequently treated with KI to give the diiodide, 6.
n
Treatment of 6 with BuLi followed by solid CO2 gave the
dicarboxylic acid 7 in 44% overall yield from 5. The diacid was
converted to the corresponding diacid chloride with oxalyl
chloride and coupled with 4-tert-butyl aniline to give the target
compound 3.† Compound 4 was prepared in a similar fashion
by converting 2,6-diisopropyl aniline to the corresponding
benzoic acid 9 and then coupling with 1,3-diaminobenzene
(Scheme 1).†
The selectivity of the binding interactions between the four
compounds was assessed by 1H NMR titrations, and the results
are summarised in Table 1. The two complementary pairs of
binding partners, 1·2 and 3·4, form complexes that are nearly an
order of magnitude more stable than the mismatched pairs, 1·4
and 3·2. The dimerisation constants and association constants
for the complexes between the isophthalic acid and bisaniline
pairs, 1·3 and 2·4, are all too low to be measured ( < 1 M21). In
these systems, the only possible mode of interaction is via a
single hydrogen bond. Thus the larger association constants
observed for the mismatched complexes, 1·4 and 3·2, indicate
that there is some specific interaction taking place in these
systems, i.e. there are two intermolecular hydrogen bonds in
these complexes.
The complexation-induced changes in chemical shift indicate
that all four complexes, 1·2, 3·4, 1·4 and 3·2, have similar
Fig. 1 Molecular zipper complexes formed between bisaniline–isophthalic
Fig. 2 Reversing the sense of the amide bonds generates a new
acid oligomers in chloroform.
complementary recognition motif.
1642
CHEM. COMMUN., 2003, 1642–1643
This journal is © The Royal Society of Chemistry 2003

View Article Online
Fig. 3 The 1·2 complex represents the parent recognition motif that forms the basis for the zipper complexes shown in Fig. 1. In the 3·4 complex, the
orientation of all of the amide bonds has been reversed to generate the new recognition motif shown in Fig. 2. The 3·2 and 1·4 complexes represent the
mismatched pairs: when one hydrogen bond is optimally positioned as shown, the amide groups on the other side of the complex are too far apart and
incorrectly oriented for optimal hydrogen bonding. The 1H NMR proton labelling used in Table 2 is shown for compounds 2 and 4.
Scheme 1
structures (Table 2). The downfield shifts of the amide protons,
a, show that the amides are involved in H-bonding interactions.
The upfield shifts of protons b and c show that the meta
substituted aromatic rings are located in similar orientations in
the diarylcyclohexane binding pockets. Thus the subtle differ-
ences in geometry associated with reversing the orientation of
the amide bonds are not enough to completely disrupt complex
formation. However, they are sufficient to significantly alter the
relative stabilites of the complexes. In these model complexes,
the ends of the molecules are free to distort to maximise the
intermolecular interactions, but in longer oligomers, the
geometrical constraints imposed by neighbouring interactions
in the chain are likely to amplify the differences in stability
between the matched and mismatched pairs.
This system therefore looks like a promising candidate for the
construction of mutually complementary oligomers that carry
sequence specific binding information. We are currently
developing the synthesis of such oligomers using unsymme-
trical amino acid building blocks to connect the two recognition
motifs described here. These compounds will open the door to
the development of new synthetic molecular information
systems.
Table 1 Association constants (Ka/M21) in CDCl3 at 295 Ka
Compound
1
2
3
4
This work was supported by the Roche (P. M. N. T.) and the
European Union (S. T.).
b
b
1
2
3
4
35 ± 2
5 ± 1
b
b
35 ± 2
6 ± 1
b
b
6 ± 1
31 ± 3
b
b
5 ± 1
31 ± 3
Notes and references
† All new compounds gave satisfactory spectroscopic data.
a Average values for at least two separate experiments. Titration data for
4–6 signals were used to determine the association constants in each
experiment. Errors are quoted as twice the standard error from weighted
mean (weighting based on the observed change in chemical shifts). b No
binding interactions were detected which means that these complexes have
1 J. H. G. Vanroode and L. E. Orgel, J. Mol. Biol., 1980, 144, 579.
2 C. Mao, W. Sun, Z. Shen and N. C. Seeman, Nature, 1999, 397, 144.
3 H. E. Huxley, Science, 1969, 164, 1356.
4 U. Koert, M. M. Harding and J.-M. Lehn, Nature, 1990, 346, 339.
5 R. Krämer, J.-M. Lehn, A. D. Cain and J. Fischer, Angew. Chem., Int.
Ed. Engl., 1993, 32, 703.
association constants less than 1 M21
.
1
Table 2 Complexation-induced changes in H NMR chemical shift (Dd,
ppm) in CDCl3 at 295 Ka
6 J. Sanchez Quesada, C. Seel, P. Prados, J. deMendoza, I. Dalcol and E.
Giralt, J. Am. Chem. Soc., 1996, 118, 277.
Proton
7 H. L. Anderson, Inorg. Chem., 1994, 33, 972.
8 P. N. Taylor and H. L. Anderson, J. Am. Chem. Soc., 1999, 121,
11538.
Complex
a
b
c
9 A. P. Bisson, F. J. Carver, C. A. Hunter and J. P. Waltho, J. Am. Chem.
Soc., 1994, 116, 10292.
10 A. P. Bisson, F. J. Carver, D. S. Eggleston, R. C. Haltiwanger, C. A.
Hunter, D. L. Livingstone, J. F. McCabe, C. Rotger and A. E. Rowan,
J. Am. Chem. Soc., 2000, 122, 8856.
11 A. P. Bisson and C. A. Hunter, Chem. Commun., 1996, 1723.
12 A. P. Bisson, C. A. Hunter, J. C. Morales and K. Young, Chem. Eur. J.,
1998, 4, 845.
1·2
1·4
3·2
3·4
1.2
1.2
1.2
1.2
21.6
21.3
21.3
21.5
20.4
20.5
20.5
20.4
a Errors are of the order of 20%. See Fig. 2 for the proton labelling
scheme.
CHEM. COMMUN., 2003, 1642–1643
1643