Macrocyclic DNA-mismatch-binding ligands: Structural determinants of selectivity

Article abstract of DOI:10.1002/chem.200901989

A collection of 15 homodi- meric and 5 heterodimeric macrocyclic bisintercalators was prepared by one- or two-step condensation of aromatic dialdehydes with aliphatic diamines; notably, the heterodimeric scaffolds were synthesized for the first time. The binding of these macrocycles to DNA duplexes containing a mispaired thy- mine residue (TX), as well as to the fully paired control (TA), was investi- gated by thermal denaturation and flu- orescent-intercalator-displacement experiments. The bisnaphthalene derivatives, in particular, the 2, 7-disubstituted ones, have the highest selectivity for the TX mismatches, as these macrocy- cles show no apparent binding to the fully paired DNA. By contrast, other macrocyclic ligands, as well as seven conventional DNA binders, show lesser or no selectivity for the mismatch sites. The study demonstrates that the topology of the ligands plays a crucial role in determining the mismatch-binding affinity and selectivity of the macrocy- clic bisintercalators. 2010 Wiley-VCH Verlag GmbH & Co. KGaA.

Full text of DOI:10.1002/chem.200901989

DOI: 10.1002/chem.200901989
Macrocyclic DNA-Mismatch-Binding Ligands:
Structural Determinants of Selectivity
[a]
Anton Granzhan, Eric Largy, Nicolas Saettel, and Marie-Paule Teulade-Fichou*
Abstract: A collection of 15 homodi-
meric and 5 heterodimeric macrocyclic
bisintercalators was prepared by one-
or two-step condensation of aromatic
dialdehydes with aliphatic diamines;
notably, the heterodimeric scaffolds
were synthesized for the first time. The
binding of these macrocycles to DNA
duplexes containing a mispaired thy-
mine residue (TX), as well as to the
fully paired control (TA), was investi-
gated by thermal denaturation and flu-
orescent-intercalator-displacement ex-
periments. The bisnaphthalene deriva-
tives, in particular, the 2,7-disubstituted
ones, have the highest selectivity for
the TX mismatches, as these macrocy-
cles show no apparent binding to the
fully paired DNA. By contrast, other
macrocyclic ligands, as well as seven
conventional DNA binders, show lesser
or no selectivity for the mismatch sites.
The study demonstrates that the topol-
ogy of the ligands plays a crucial role
in determining the mismatch-binding
affinity and selectivity of the macrocy-
clic bisintercalators.
Keywords:
intercalations
·
bisintercalators · DNA mismatches ·
DNA recognition · macrocycles
Introduction
many cancers, including hereditary nonpolyposis colon can-
[3]
cers, are associated with deficient mismatch repair; in turn,
the inhibition of repair is currently investigated as a novel
Mismatched base pairs in DNA represent defects in the
double-helical DNA structure that are permanently formed
in living cells due to statistical errors in the course of DNA
[4]
approach to potentiate the effect of anticancer drugs.
Thus, molecules that specifically bind to mismatched base
pairs in DNA may interfere with mismatch-recognizing ma-
chineries and provide a basis for novel chemotherapeutic
9
10
replication (1 per 10 –10 base pairs per cell division). They
may also be generated by incorporation of chemically dam-
aged nucleotides or by incorporation of normal nucleotides
[5]
agents. Indeed, it was shown that mismatch-selective rho-
dium complexes preferentially inhibit proliferation of mis-
match-repair-deficient cell lines, compared to those lines
[1]
opposite damaged bases in the DNA template. The rate of
generation of DNA mismatches can be greatly increased by
exogenous factors, such as genotoxic chemicals or UV radia-
tion. If not repaired, mismatched base pairs lead to an en-
hanced frequency of mutations (50–1000-fold increase in or-
ganisms with a genetically inactivated mismatch-repair
system), which leads to an accumulation of mutations and is
harmful for the organism. Therefore, each organism possess-
es an ensemble of systems responsible for the maintenance
of genomic integrity. In particular, the mismatch-repair
[6]
which are MMR proficient. Therefore, the search for other
types of small molecules that selectively recognize mis-
matched base pairs in DNA is currently a challenging task.
Currently, only a limited number of compounds that selec-
tively bind to DNA mismatches and not to the fully paired
DNA are known. The most well-studied examples are repre-
[2]
[6,7]
sented by the aforementioned metalloinsertors.
The bulk-
iness of these cationic metal complexes hinders their associ-
ation with fully matched DNA and provides them with se-
lectivity towards mismatched base pairs, particularly cyto-
(
MMR) system, consisting of many proteins, recognizes and
repairs the mismatched base pairs in the DNA. Notably,
[8]
sine-containing mismatches, and other pairing defects, such
[9]
as abasic sites and single-base bulges, by taking advantage
of the local thermodynamic destabilization of such defects
relative to the Watson–Crick base pairs. On the other hand,
further increase of the steric bulk of metalloinsertors by in-
troduction of p-extended aromatic ligands decreases the se-
lectivity towards mismatches because the resulting com-
plexes also bind to the well-matched sites, which poses a
[
a] Dr. A. Granzhan, E. Largy, Dr. N. Saettel, Dr. M.-P. Teulade-Fichou
UMR176 CNRS, Institut Curie, Centre de Recherche
Centre Universitaire, 91405 Orsay (France)
Fax : (+33)169-07-53-81
E-mail: marie-paule.teulade-fichou@curie.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200901989.
878
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 878 – 889

FULL PAPER
[
10]
[20]
possible limitation of this class. Another important class
of DNA-mismatch binders is represented by the homo- and
heterodimeric derivatives of naphthyridine, which selectively
bind to GG, AG, or CC mismatches by intercalation of the
naphthyridine units into the base stack and formation of
complementary hydrogen bonds with the guanine, adenine,
base-pairing defects, like abasic sites and thymine-contain-
[21]
ing mismatches, through a putative threading bisintercala-
tion mode. Moreover, insertion of BisA at TX-mismatch
sites induces a displacement (“flipping”) of the thymine into
an extrahelical position. This represented the first example
of a small molecule flipping the mismatched base out of the
DNA base stack, in a manner similar to that of DNA meth-
[
11–13]
or cytosine residues, respectively.
were also used to detect mismatched base pairs in heterodu-
plex DNA by analytical methods, such as surface plasmon
These compounds
[22]
yltransferases and DNA glycosylases. Subsequently, sever-
al studies have shown that other mismatch binders, such as
[
12–14]
[15]
[23]
resonance
or affinity chromatography.
This class of
naphthyridine dimers and bulky rhodium-based metalloin-
[8b,24]
mismatch binders is very interesting due to the high fidelity
of the hydrogen-bonding-mediated recognition of both mis-
matched bases; however, studies of the medicinal activity of
sertors,
are also able to displace bases into extrahelical
positions. In a more recent study, we have shown that a bis-
naphthalene macrocyclic compound, 2,6-BisNP, which is
structurally related to BisA, binds to TX mismatches with
[16]
these compounds remain preliminary.
A third class of
[25]
well-studied mismatch binders is the imidazole-containing
polyamides, which selectively bind to TG-mismatched base
pairs by recognition of the hydrogen-bonding pattern of the
bases in the minor groove. The structural and thermodynam-
ic parameters of this binding were thoroughly determined.
This class of agents is important due to the high biological
significance of the TG mismatch, which also represents a
particularly hard target for recognition by small molecules
because it is only slightly less stable than the Watson–Crick
even higher affinity and selectivity than the latter. Con-
versely, we observed that an analogous macrocycle contain-
ing anthracene residues, 9,10-BisAN, is much less selective
towards TX-mismatched duplexes and also binds to the fully
paired duplexes. However, this occurs through different
binding modes, as shown by the drastic differences in the
fluorescence properties of the bound compound, which
allows a “naked-eye” discrimination between the mis-
[17]
[18]
[26]
matched and fully paired duplexes. Subsequently, it was
found that a ruthenium-based DNA metallointercalator also
binds to well-matched and mismatched base pairs by differ-
ent binding modes, with an impact on its luminescence prop-
[
19]
base pairs.
Over the past few years, we have shown that a macrocy-
clic bisacridine compound (BisA, Scheme 1) recognizes
[27]
erties.
These studies led us to conclude that both the aromatic
units and the connectivity between them are structural de-
terminants for the selective recognition of mismatched sites.
Consequently, with the aim of establishing comprehensive
structure–binding relationships and further optimization of
the mismatch-binding ligands in terms of affinity and selec-
tivity towards TX mismatches, we prepared an extended
series of macrocyclic compounds, containing two identical
(
homodimers) or different (heterodimers) aromatic units,
connected by linking chains of the same length. Herein, we
report the synthesis of these compounds, including hetero-
dimers that represent rare examples in the chemistry of
macrocycles, as well as a study of their interaction with
double-stranded oligonucleotides containing one TX-mis-
matched site. The DNA-binding affinities and selectivities
towards mismatches were determined by means of thermal-
denaturation experiments and, for the most representative
macrocycles, fluorescent-intercalator-displacement (FID) ti-
tration experiments were performed to confirm the binding
characteristics.
To firmly establish that the macrocyclic scaffold is neces-
sary for selective mismatch binding and to estimate the con-
tribution of nonspecific binding, we included in our study a
number of well-known “classical” DNA binders, such as
compounds that bind by the prevailing intercalative mode:
[
28–30]
ethidium bromide (EB, Scheme 2),
proflavine hemisul-
[31]
[32,33]
fate (PF), and thiazole orange tosylate (TO),
as well
as representative DNA-minor-groove binders, namely the
Scheme 1. Mismatch-binding macrocycles identified in previous studies.
Presumably, all three compounds are protonated at the four benzylamino
groups at pH 6–7.
bisbenzimidazole derivative Hoechst 33258 trihydrochloride
[
21,25,26]
[29a,34]
(H33258)
and 4’,6-diamidino-2-phenylindole dihydro-
Chem. Eur. J. 2010, 16, 878 – 889
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
879

M.-P. Teulade-Fichou et al.
precipitated directly from the
mixture
(average
purity
>
1
90%, as determined by
H NMR spectroscopy) and
were thus used without further
manipulation. In some cases,
these intermediate tetraimines
were barely soluble or under-
went
decomposition
in
common NMR solvents, which
prevented their complete char-
acterization. The reduction of
the tetraimine intermediates
with NaBH , followed by con-
4
version of the macrocyclic
amines into hydrochloride salts
and purification by recrystalli-
zation or chromatography (see
the Supporting Information),
gave the corresponding poly-
ammonium
macrocycles
Scheme 2. Nonmacrocyclic DNA binders used in this study. Counterions and protonation sites are omitted for
clarity.
(
Table 1), the structure and
[35]
chloride (DAPI). Our interest in these compounds was, in
part, stimulated by the recent reports on H33258 binding to
[
36,37]
bulged and mismatched RNA,
plexes containing TG-mismatched base pairs.
binding of DAPI to a DNA duplex containing a TT mis-
as well as to DNA du-
[38]
Moreover,
[39]
match was demonstrated by NMR spectroscopy.
At the
same time, it was claimed that EB does not bind to short du-
[38]
plexes containing TG mismatches. We also included the
tetracationic porphyrin meso-tetrakis(4-N-methylpyridi-
nium)porphyrin tetratosylate (TMPyP4), which binds to
duplex DNA either by intercalation between GC base pairs
[40]
II
or the grooves of the AT-rich sequences and whose Cu
complex is able to flip the nucleic acid bases out of well-
[41]
matched DNA. Additionally, we included a naturally oc-
curring tetraamine spermine (SPM), which binds to double-
stranded DNA without sequence or base selectivity through
electrostatic interactions, mainly in the major groove, as
well as the bisnaphthalene derivative NP2, which represents
an open-chain analogue of 2,6-BisNP.
Scheme 3. Synthesis of homodimeric macrocycles. For assignment of the
aromatic units, see Table 1. Reagents and conditions: a) MeCN, RT, 5–
[42]
7
days; b) NaBH
4 2 2
, CH Cl /MeOH, RT, 3 h; c) HCl, MeOH or EtOH, 20–
90% over 3 steps.
1
13
Results
purity of which were confirmed by H and C NMR spec-
troscopy, LC–MS, and elemental analysis data. It should be
noted that in several cases the NMR spectra of the macrocy-
clic polyamines displayed broad signals, which did not allow
proper signal assignment, due to conformational restrictions
imposed by the macrocyclic scaffold; in such cases, single
peaks in the HPLC chromatograms and correct MS and ele-
mental analysis data were used as a proof of identity and
purity for the samples.
Synthesis of macrocyclic ligands and reference compounds
Synthesis of homodimeric macrocycles: These macrocycles
were obtained by a straightforward [2+2]-type cycloconden-
sation of aromatic dialdehydes 1a–k with aliphatic diamines
[43]
2
a–c (Scheme 3, Table 1). In a similar manner to the pre-
viously reported synthesis of macrocycles 2,7-BisA-N, 2,6-
BisNP, and 9,10-BisAN (Scheme 1),
[26,44]
the cyclocondensa-
tion was performed in acetonitrile under moderate to high
dilution conditions. In this way, the tetraimine intermediates
Synthesis of heterodimeric macrocycles: Our general ap-
proach to the synthesis of heterodimers containing two dif-
880
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 878 – 889

Macrocyclic DNA-Mismatch-Binding Ligands
FULL PAPER
[48]
Table 1. Homodimeric macrocycles prepared according to Scheme 1.
initial Boc protection of the amino group,
the hydroxy
group of derivative 4 was converted into a primary amino
group by a Mitsunobu reaction with phthalimide, followed
by deprotection of the phthalimide group with hydrazine.
This allowed us to obtain intermediate 5 in 81% yield from
amino alcohol 3. The reaction of 5 with selected aromatic di-
aldehydes (OHCÀAÀCHO; Scheme 4, Table 2) gave the cor-
Macrocycle
Aldehyde
X
O
2
2
2
,7-BisNP
1a
,7-BisNP-S
,7-BisNP-N
1a
1a
S
responding tetraamine derivatives 6a–c, which were isolat-
ed, characterized, and handled as hydrochloride salts. For
the macrocyclization reaction, the salts were quantitatively
converted into free bases by treatment with ion-exchange
resin and were then treated with the second dialdehyde frag-
ment (OHCÀBÀCHO), to give the amino-imino heterodim-
NH
1
,5-BisNP
1b
O
ers contaminated by higher-order oligomers and products of
polymerization, which rendered isolation and purification of
these intermediates impossible. The crude intermediates
3
4
,3’-BisBP
,4’-BisBPy
,8-BisAN
1c
1d
O
O
were reduced with NaBH , and the desired macrocycles
4
were isolated by flash chromatography, either directly or
after conversion into lipophilic tetra-N-Boc-protected deriv-
[49]
atives, which after purification and deprotection with HCl
gave the macrocycles as pure hydrochloride salts. The struc-
1
2
4
2
1e
1 f
1g
1h
O
O
O
O
1
ture and purity of the products were confirmed by H and
[
a]
,7-BisA
,5-BisA
13
C NMR spectroscopy, LC–MS, and elemental analysis
data.
Synthesis of NP2: The nonmacrocyclic analogue of macrocy-
cle 2,6-BisNP, the tetraamine derivative NP2, was synthe-
sized by reductive amination with ethylamine of diformyl
derivative 7, which, in turn, was prepared in 8 steps from 6-
,8-BisPZ
,9-BisPN
bromo-2-naphthoic
acid
as
described
elsewhere
[
b]
2
4
1i
1j
O
[49]
(
Scheme 5).
DNA-mismatch-binding affinities and selectivities
,7-BisPN
O
Thermal denaturation experiments: Thermal denaturation of
DNA is a rapid and straightforward method for determina-
tion of the stabilization effect of ligands towards a given
DNA structure. The extent of the ligand-induced stabiliza-
tion provides a semiquantitative evaluation of the ligand af-
[
45]
[
a] Synthesis was described previously.
[b] Although the synthesis of
[
46]
this macrocycle was recently described, we obtained a material with a
composition (C36 ·6HCl·8H O) that was slightly different from
that given in the literature (C36 ·5HCl·3H O).
H
40
N
8
O
2
2
H
40
N
8
O
2
2
[50]
finity towards duplex DNA. Base mismatches significantly
reduce the thermodynamic stability of DNA duplexes, and
the extent of this destabilization is dependent on the duplex
length. As a consequence, the stabilization effect of mis-
match-binding ligands becomes more pronounced under
conditions of low duplex stability (low ionic strength, short
duplexes). On the other hand, longer duplexes have sharper
melting profiles, which allow more precise determination of
melting temperatures, and offer more potential binding sites
(intercalation between base pairs, groove and surface bind-
ing) to enable better determination of mismatch selectivi-
ty. Hence, we chose to use 17-mer duplexes, which repre-
sent a good compromise, with a general sequence TX (5’-
CCAG TTC GTA GTA ACCC-3’/5’-GGGT TAC TXC
GAA CTGG-3’), containing either a matched (TA) or a
mispaired (TG, TC, TT) thymine residue in the center of
the middle triplet. The reference oligonucleotide (TA) was
used in previous studies and corresponds to a biological se-
ferent aromatic residues A and B relies on fragment-to-frag-
ment assembly from two aromatic aldehydes (Scheme 4,
Table 2). Although a number of unsymmetrical macrocycles
[43b]
have been described,
the heterodimeric-macrocycles of
the type presented in Table 2 have yet to be reported. For
the sake of clarity, macrocycles containing two identical
units but linked at different positions of each aromatic ring
were also included in the heterodimer category (Table 2).
For the synthesis of the heterodimeric macrocycles pre-
sented in Table 2, a common intermediate, monoprotected
diamine 5, was essential. Although the preparation of this
compound by monoprotection of 2,2’-oxydiethylamine had
[25]
[47]
been described, the yield of the monoprotected diamine
[25]
was only about 50% in our hands.
Taking into account
the limited availability of 2,2’-oxydiethylamine, we devel-
oped an alternative preparation for amine 5 by starting from
inexpensive 2-(2-aminoethoxy)ethanol (3; Scheme 4). After
Chem. Eur. J. 2010, 16, 878 – 889
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
881

M.-P. Teulade-Fichou et al.
Scheme 4. Synthesis of heterodimeric macrocycles. Reagents and conditions: a) Boc
2
O, CH
OH, MeOH, reflux, 6 h, 89% over 2 steps; d) benzene, 18 h, reflux; e) NaBH , CH
0–90% over 3 steps; g) Amberlite IRA-420 (OH ), MeOH; h) MeCN, RT, 5–7 days; i) NaBH
0% over 4 steps. Boc: tert-butoxycarbonyl; DIAD: diisopropylazodicarboxylate; THF: tetrahydrofuran.
2
Cl
2
, RT, 18 h, 91%; b) phthalimide, PPh
Cl /MeOH, RT, 3 h; f) HCl, EtOH or MeOH, 40–658C,
, CH Cl /MeOH, RT, 3 h; j) HCl, MeOH or EtOH, 20–
3
, DIAD, THF, RT,
1
8
6
8 h; c) N
2
H
5
4
2
2
À
4
2
2
Table 2. Heterodimeric macrocycles prepared according to Scheme 2.
Linear
Macrocycle
intermediate
1
2
1
,5/2,6-BisNP
6a
,7-NP/9,10-AN 6b
,8/9,10-BisAN
6b
2
2
,7-NP/2,7-A
6c
6d
Scheme 5. Synthesis of the reference compound NP2.
,7-NP/1,1’-FC
equivalent of the ligand or not. Thus, ligands that occupy a
single site with high affinity should show little or almost no
increase in the DTm value upon an increase in the ligand
concentration, whereas ligands with lower affinity and/or
several binding sites should show a larger increase in the
quence recognized by the DNA methyltransferase enzyme
M.TaqI. Under the conditions employed (pH 6, [Na ]=
+
20 mm), the fully matched duplex TA has a melting temper-
DT value.
m
ature of 46.58C, whereas the mismatch-containing duplexes
TG, TT, and TC denaturate at significantly lower tempera-
tures (42.0, 38.4 and 36.68C, respectively). The ligand-in-
Nonmacrocyclic (“classical”) DNA binders: The results of
the thermal denaturation of oligonucleotides TX in the pres-
ence of “classical” DNA binders are presented in Figure 1.
The intercalators EB, TMPyP4, and PF strongly bind to
fully matched and mismatch-containing DNA duplexes, as
indicated by the large DTm values (5–78C at q=1), without
significant selectivity toward one of the four duplexes.
Moreover, the large increase in the DTm values upon an in-
crease in the ligand concentration indicates that saturation
of the potential site(s) is not reached at q=1. The absence
of specific stabilization of a particular duplex by ethidium
duced changes in the melting temperatures (DT ) at ligand-
m
to-duplex ratios (q) of 1 and 2 are listed in Table S1 in the
Supporting Information and are represented as bar graphs
in Figures 1–4 to facilitate direct comparison between the
various ligands. The value of the ligand-induced increase in
the melting temperature of the duplex serves as a measure
of the ligand binding affinity, whereas the difference in the
values observed with one and two equivalents of ligand per
duplex indicates whether saturation is reached with one
882
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 878 – 889

Macrocyclic DNA-Mismatch-Binding Ligands
FULL PAPER
Figure 2. Results of thermal-denaturation experiments for the naphtha-
lene derivatives. For the assignment of datasets and experimental details,
see Figure 1.
Figure 1. Results of thermal-denaturation experiments with duplexes TX
+
(
c=6 mm in cacodylate buffer, pH 6.0, [Na ]=20 mm) and nonmacrocy-
clic DNA binders at ligand-to-duplex ratios of q=1 and q=2. Estimated
error in T
determinations is Æ0.68C.
m
bromide (EB), as observed by thermal denaturation experi-
ments, is in agreement with the results of a recent independ-
ent study, which showed an absence of mismatch selectivity
[27]
of EB. Thiazole orange (TO) also induces thermal stabili-
zation of all duplexes but to a smaller extent than the other
intercalators, although it has been reported to have a higher
[51]
affinity for duplex DNA than EB.
The groove binders DAPI and H33258 induce higher sta-
bilization of all duplexes, without discrimination between
matched and mismatched. The stabilization is particularly
pronounced at a ligand per duplex ratio of q=2, and larger
DTm values are observed than those induced by intercala-
tors. The absence of selective binding of these groove bind-
ers to mismatched duplexes, although expected, is in disa-
greement with studies reporting on binding of DAPI and
[38,39]
H33258 to mismatch-containing DNA duplexes.
Finally,
Scheme 6. Schematic representation of the topology of the BisNP family
the polyamine SPM also stabilizes all four duplexes but to a
smaller extent than all of these minor-groove binders and in-
tercalators, with the exception of TO.
of macrocycles.
pounds (DT (TT)ꢀ9–118C); this result is indicative of simi-
m
Naphthalene derivatives: We previously identified the bis-
naphthalene macrocycle 2,6-BisNP as a promising mis-
lar binding affinities. On the other hand, the degree of selec-
tivity, as represented by the difference in the thermal stabili-
zation of well-matched (TA) and mismatch-containing du-
plexes, is strongly influenced by the structure of the ligands.
[25]
match-selective agent. In the current work, we prepared
and investigated an extended series of homo- and heterodi-
meric naphthalene derivatives, to establish the relationship
between the structure of the macrocycle (substitution pat-
tern of the naphthalene units, heteroatoms in the linking
chains) and the mismatch-binding characteristics. The results
of the duplex-DNA stabilization induced by the novel naph-
thalene derivatives (Figure 2, Scheme 6) show that all of the
bisnaphthalene macrocycles (2,6- and 1,5-BisNP; 2,7-BisNP,
Thus, no stabilization of the TA duplex (DT (TA)ꢁ0) is ob-
m
served with macrocycles that have the 2,7-connectivity of
the naphthalene units (2,7-BisNP, -S, and -N). Moreover, in
these cases, the binding to the mismatch-containing duplexes
is readily saturated at a 1:1 ligand-to-DNA ratio. Among
the three derivatives, the 2,7-BisNP-S shows a lower effect
(DT (TT)ꢀDT (TC)ꢀ6–78C). Thus, the presence of a
m
m
-S, and -N; 1,5/2,6-BisNP), the naphthalene–acridine hetero-
sulfur atom in the linker does not afford any significant ad-
vantage; on the contrary, it has a penalty in terms of affinity.
Notably, all of the 2,7-substituted derivatives are more se-
lective than the prototypal macrocycle with the 2,6 connec-
tivity, that is, 2,6-BisNP, which binds poorly but still signifi-
dimer 2,7-NP/2,7-A, and even the nonmacrocyclic bisnaph-
thalene NP2 preferentially bind to the mismatch-containing
duplexes, as indicated by the larger DTm values with these
than with the matched duplex (DT (TX)>DT (TA)). Inter-
m
m
estingly, almost the same DTm values are observed for the
less stable TT-mismatched duplex with most of the com-
cantly to the TA duplex (DT (TA)=2–38C). Finally, the
heterodimer 2,7-NP/2,7-A also shows selective stabilization
m
Chem. Eur. J. 2010, 16, 878 – 889
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
883

M.-P. Teulade-Fichou et al.
of the mismatch-containing duplexes with a slight preference
for the TC duplex, which may be related to the TC-mis-
match selectivity of the bisacridine macrocycle BisA
(
Scheme 1).
Interestingly, macrocycle 1,5-BisNP, although structurally
very close to the 2,7 and 2,6 isomers, is much less selective
for the mismatched duplexes, as indicated by the significant
stabilization of the matched duplex, particularly pronounced
with two equivalents of ligand (DT (TA)ꢀ78C at q=2).
m
This provides evidence that this macrocycle, unlike the 2,7
and 2,6 isomers, exhibits a strong tendency to bind to fully
paired DNA. It is noteworthy that the heterodimer 1,5/2,6-
BisNP exhibits a behavior that is intermediate between
those of the “parent” homodimers 2,6-BisNP and 1,5-BisNP
(
indicated by thin lines in Figure 2), both in terms of affinity
and selectivity. This observation may establish the “rule-of-
thumb” that the DNA-binding properties of heterodimeric
macrocycles may be estimated by knowing the properties of
the homodimeric parents.
Figure 3. Results of thermal-denaturation experiments for the anthracene
derivatives. For the assignment of datasets and experimental details, see
Figure 1.
A similar decrease of the mismatch selectivity is observed
in the case of the nonmacrocyclic bisnaphthalene derivative
NP2, because this compound induces pronounced stabiliza-
tion of the matched duplex (DT (TA)ꢀ68C at q=2). Com-
m
parison of this to the values obtained with its macrocyclic
analogue 2,6-BisNP allows us to confirm that the macrocy-
clic framework is indispensable for prevention of binding to
the well-matched duplex. Finally, the heterodimer 2,7-NP/
1,1’-FC, which combines a naphthalene unit with a bulky fer-
rocene moiety, displays weak binding to all four duplexes.
The weak DNA-binding affinity of this compound, which
has the same overall tetracationic charge as the other mac-
rocycles but a different shape, in combination with the ab-
sence of a preference for the mismatched structures, leads
us to the conclude that, in the case of bisintercalator macro-
cycles, both aromatic units actively participate in the inter-
calation process and do not merely introduce steric bulk
that prevents interaction with the well-matched duplexes.
Scheme 7. Schematic representation of the topology of the BisAN family
of macrocycles.
es, particularly TC. At the same time, the greatly increased
DT values observed with two equivalents of this compound
m
indicate a significant contribution of nonspecific binding. Fi-
nally, the hybrid anthracene–naphthalene macrocycle 2,7-
NP/9,10-AN shows preferential binding to the TT- and TC-
mismatched duplexes with moderate binding to the TA con-
trol (DT (TT)ꢀDT (TC)ꢀ128C, DT (TA)=48C at q=2).
Anthracene derivatives: Previously, we showed that the bi-
santhracene macrocycle 9,10-BisAN gives different fluores-
cence responses when bound to TX-mismatched DNA and
to well-matched DNA due to both binding selectivity and
m
m
m
The stabilization effect is almost saturated at a ligand-to-
duplex ratio of 1:1 and, altogether, the profile is reminiscent
of that of the bisnaphthalene series.
[26]
different binding modes. Thus, we decided to include this
compound in the present study, to serve as a reference for
the bisanthracene analogues with different topology (1,8-
BisAN and 1,8/9,10-BisAN). The data of the thermal dena-
turation experiments (Figure 3) show that, in contrast to the
moderate but still significant preference of macrocycle 9,10-
BisAN for mismatched duplexes, the isomer 1,8-BisAN
binds strongly the four duplexes without a clear discrimina-
In summary, the anthracene units enhance the interaction
with all forms of DNA, most likely due to their stronger p-
stacking ability, as compared to the results with naphthalene
rings. Additionally, the 1,8-substituted pattern imposes a
particular conformation (Scheme 7) that may be more prone
for insertion into the well-matched duplex, possibly by a
groove-binding mode; this again provides evidence for the
importance of the ligand topology. Finally, the introduction
of a naphthalene unit drives the behavior back to that of the
BisNP series (lower affinity but better mismatch selectivity).
tion (DT =17–228C at q=2). This stabilization profile,
m
along with the very broad, multiphasic melting curves, is
very similar to the ones observed with the groove binders
DAPI and H33258, which are not mismatch-selective either
(
Figure 1). Interestingly, the heterodimer 1,8/9,10-BisAN
displays a smaller stabilization of the fully matched duplex
while retaining a strong affinity for the mismatched duplex-
Other macrocyclic compounds: After having established the
preliminary relationships between the structure and DNA-
884
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 878 – 889

Macrocyclic DNA-Mismatch-Binding Ligands
FULL PAPER
binding properties in the naphthalene and anthracene series
of macrocycles, we turned our attention to other homodi-
meric macrocycles including various aromatic units, such as
acridine, phenazine, phenanthroline, biphenyl, and bipyri-
dine. The bisacridine macrocycle 2,7-BisA-N (BisA) was the
first mismatch-selective macrocycle reported, so we antici-
pated that screening of other heteroaromatic derivatives
could result in interesting new ligands. The results (Figure 4)
duplex structures, presumably by electrostatic-driven surface
binding or accommodation in grooves.
Fluorescent-intercalator displacement (FID) titrations: Taken
together, the thermal denaturation data indicate that the
best candidates for TX mismatch recognition are the bis-
naphthalene homodimers with the 2,7 connectivity, 2,7-
BisNP and 2,7-BisNP-N, because these compounds did not
stabilize the TA duplex at all, under our conditions. To con-
firm the selectivity of these optimized ligands, we investigat-
ed the binding to both TT and TA duplexes by using the
FID method. The FID assay is a versatile method, which is
used to investigate binding of ligands to various DNA struc-
[21]
[51,52]
tures and may provide apparent binding constants.
In
this experiment, the DNA is fluorescently stained with a
dye, typically ethidium bromide or thiazole orange, which is
almost nonfluorescent in the absence of DNA but becomes
fluorescent upon binding to DNA. Binding of ligands to the
DNA–intercalator complex leads to partial displacement of
the bound fluorophore, which is accompanied by a decrease
in the fluorescence intensity. The fluorescence decrease may
serve for evaluation of affinity in a series towards a given
DNA structure. As seen in Figure 1, EB binds equally well
to the mismatched and matched duplexes, which is an im-
portant criterion for choosing this intercalator as the fluores-
cent probe. Moreover, the fluorescence intensity of DNA-
bound EB is not affected by the presence of mismatched
Figure 4. Results of thermal-denaturation experiments for the other mac-
rocyclic compounds. For the assignment of datasets and experimental de-
tails, see Figure 1.
confirm the preferential binding of the macrocyclic bisacri-
dines 2,7-BisA-N and 2,7-BisA to the mismatch-containing
duplexes, with very similar behavior and, as previously ob-
[25,27]
base pairs in the DNA.
In our experiments, the TT and TA duplexes were stained
with EB and the displacement of the probe was calculated
from the changes in fluorescence intensity upon addition of
the ligands (see the Supporting Information). In the case of
the duplex with a TT mismatch, addition of 2,7-BisNP leads
to the rapid displacement of the EB probe, which reaches a
plateau upon addition of an increasing concentration of the
ligand (Figure 5A). The shape of the curve clearly indicates
a binding equilibrium that reaches a limiting value at 1–
2 equivalents of ligand, a result that is fully consistent with
the Tm data. On the other hand, very poor displacement of
the probe from the fully paired TA control was observed. A
totally different behavior was observed with the ligand 3,3’-
BisBP (Figure 5B), which readily displaced the probe from
both duplexes, irrespective of the presence of a mismatch.
Moreover, the same phenomena were observed when the
experiments were repeated at 10-fold higher concentrations
of DNA and EB, respectively (see the Supporting Informa-
tion, Figure S1). Thus, the results of the FID titrations fully
confirm the trend in selectivity drawn from the thermal de-
naturation experiments.
[25]
served, with a lower affinity than their bisnaphthalene an-
alogues. However, modification of the topology in the acri-
dine series (in the case of 4,5-BisA) results in increased
binding to control duplex TA. The replacement of the acri-
dine by other aromatic systems (2,8-BisPZ, 2,9- and 4,7-
BisPN, 4,4’-BisBPy) does not lead to significant improve-
ment, because binding to the TA control is clearly more pro-
nounced. Taken together, these results indicate that, as in
the naphthalene and anthracene series, the substitution pat-
tern of the aromatic residues, which determines the shape of
the macrocycle, plays a crucial role in preventing binding to
well-matched DNA and, thereby, determines the mismatch
selectivity. On the other hand, the electronic structure of the
aromatic units may play a role in the mismatch-binding
strength, as illustrated by the significantly higher affinity of
2,8-BisPZ than that of 2,7-BisA. It should be also noted that
the thermal denaturation studies reveal a slight preferential
stabilization of TC duplexes by the bisacridine and bisphe-
nazine macrocycles, whereas most bisnaphthalene deriva-
tives clearly prefer the TT counterpart (Figure 2).
The thermal denaturation studies show even less mis-
match selectivity of the structurally related derivatives 4,4’-
BisBPy and 3,3’-BisBP, because these macrocycles, contain-
ing nonplanar flexible bipyridine and biphenyl units, strong-
Discussion
ly stabilize the matched duplex TA (DT (TA)ꢀ4–68C at
We conclude that the BisNP compounds and, in particular,
the 2,7-substituted derivatives, exhibit the best selectivity for
the TX-mismatched duplexes over the fully paired duplex
control. In addition, the BisNP series exhibits a strong pref-
m
q=1). Especially in the case of 3,3’-BisBP, the large further
increase in the DTm value at q=2 and the broad, biphasic
melting curves are typical of nonspecific binding to the four
Chem. Eur. J. 2010, 16, 878 – 889
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
885

M.-P. Teulade-Fichou et al.
tions with DNA bases are involved but might not be pre-
dominant in the interaction.
Finally and very importantly, the connectivity between the
two linkers and the two aromatic units appears to be a key
factor for both affinity and selectivity for the mismatched
site. This underscores the crucial role played by the confor-
mation of the ligand. It should be noted that the macrocy-
cles presented herein are flexible molecules that, in princi-
ple, may adopt various conformations upon binding. Howev-
er, it has been shown that the BisA and BisNP derivatives
adopt a semiclosed conformation once bound to DNA or in
complexes with planar aromatic substrates; these data were
obtained in solution (NMR spectroscopy) and in the solid
[20,45]
state (X-ray crystallography), respectively.
In addition,
during the course of our study, an X-ray crystal structure of
the bisphenanthroline compound (called 2,9-BisPN herein)
has been solved and shows that this molecule also adopts a
[46]
semiclosed conformation in the solid state. If these facts
are taken together, it is reasonable to assume that all of the
cyclic bisintercalators studied herein adopt a similar folded
conformation in the free state in aqueous solution due to
the constraints imposed by the two linkers and hydrophobic
forces. On the other hand, the topology of the ligand and its
internal molecular flexibility may vary greatly depending on
the position of attachment of the two linkers on the aromat-
ic platforms (as represented in Schemes 6 and 7), which, in
turn, influences the ability of the ligand to interact with the
DNA residues (bases and phosphates). This subtle influence
is clearly illustrated by the BisNP series, because a change
in the connectivity from the 2,6 to the 2,7 or 1,5 positions in-
creases or decreases selectivity, respectively. To further sup-
port the existence of a folded conformation and to have in-
sight into the flexibility of these macrocyclic systems, we
performed molecular modeling studies for these three iso-
mers. Initial stable conformations of 1,5-, 2,6-, and 2,7-
Figure 5. Displacement of EB upon titration of ligands 2,7-BisNP (A)
and 3,3’-BisBP (B) to solutions of duplexes TA (*) and TT (*; c=
+
0
.1 mm) and EB (c=0.4 mm in cacodylate buffer, pH 6.0, [Na ]=110 mm).
erence for TT and TC mismatches as compared to the TG
mismatch; this result is consistent with the relative thermo-
dynamic stability of the three mismatches, as shown previ-
ously. Although thermal stabilization measurements provide
an indirect evaluation of affinity, an examination of the data
enables us to conclude that the stabilization induced by the
2,7-BisNP homodimers is mainly due to a single binding
event occurring at the mismatch site present in the target
duplex.
[44]
BisNP protonated at the nitrogen atoms in the gas phase
The replacement of the bicyclic naphthalene ring by tricy-
clic aromatic rings with a larger p surface (acridine, anthra-
cene, phenazine, phenanthroline) does not give any tangible
improvement in the recognition properties; in most cases,
the corresponding homodimers do not show a significant in-
crease in the stabilization of the mismatched duplexes and
globally they exhibit a lower selectivity, as reflected by bind-
ing to the control duplex. The anthracene derivatives were
an exception to this trend and showed higher DTm values
were generated (Figure 6A and Figures S2–S4 in the Sup-
[53]
porting Information). All geometries showed a stair-like
structure with aromatic rings sliding away as far as possible.
However, in molecular dynamics (MD) simulations with a
[54,55]
water model (see the Supporting Information),
the
three molecules showed a wide range of molecular motion
À1
over a magnitude of 30–40 kcalmol . The lowest-energy
conformations span from a semiopen structure in the case of
2,7-BisNP (Figure 6B) to a semiclosed conformation for 2,6-
or 1,5-BisNP, respectively (see the Supporting Information,
Figures S2–S4). Notably, in the course of the simulation, 1,5-
BisNP spends more time in its semiopen form than 2,7-
BisNP, whereas 2,6-BisNP shows a significant degree of dis-
(
16–208C for the BisAN derivatives versus 9–118C for the
BisNP series for the TT duplex), but, clearly, these values
result from several contributions, in particular, nonspecific
[26]
binding at diverse sites. The performance of the hetero-
dimers possessing two different aromatic units (naphtha-
lene–acridine, naphthalene–anthracene) demonstrates that
inclusion of the 2,7-NP unit leads to better selectivity. Fur-
thermore, the introduction of flexible bicyclic units, which
are less rigid and less prone to p-stacking interactions, total-
ly abolishes the preferential mismatch binding; this is partic-
ularly obvious in the case of the biphenyl derivative (3,3’-
BisBP). These observations suggest that p-stacking interac-
[55]
tortion. In no case was the completely closed conforma-
tion involving intramolecular stacking of the two aromatic
units observed.
It can be hypothesized that, once bound to mismatched
DNA, these compounds form a catenated complex, as seen
[20b]
previously for BisA in an apurinic site,
to take advantage
of the enhanced breathing motions of mispaired bases. The
strength of the interaction and the stability of the resulting
886
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 878 – 889

Macrocyclic DNA-Mismatch-Binding Ligands
FULL PAPER
+
Figure 6. Geometries of 2,7-BisNP+4H . A) Molecular mechanics opti-
mization in the gas phase. B) Front and side views of the lowest energy
conformation during a molecular dynamics simulation in a neutralized
À
(
4Cl ) water box.
Figure 7. Structure of 2,7-BisNP docked into an 11-mer duplex with a TT
mismatch, d(CGCACT*CACGC):d(GCGTGTGTGCG). The structure
complex are highly dependent on the changes in entropy
and enthalpy due to complexation. Thus, it may be assumed
that the ligands, the conformation of which in the free state
is similar to that when bound with mismatched DNA, have
higher binding affinity due to smaller conformational entro-
py. Indeed, in the course of the MD simulation, a number of
semiclosed structures were encountered within a reasonable
of the oligonucleotide was derived from the NMR spectroscopy data of
[20b]
an abasic duplex
by attachment of a thymine residue (T*); see the
Supporting Information for details.
for groove or surface binding. Studies are currently under-
way to experimentally determine the structure of the com-
plexes between TX-mismatched duplexes and the best iden-
tified candidate 2,7-BisNP.
À1
range of energy (10–20 kcalmol ) compared with the
lowest energy structures. To further support this hypothesis,
the three BisNP isomers in several semiclosed forms were
[56]
docked into an 11-mer duplex containing a TT mismatch.
The docking calculations gave very similar solutions for the
three ligands (see the Supporting Information, Figure S5), in
excellent agreement with the thermal denaturation data. An
example of a docking solution with 2,7-BisNP is represented
in Figure 7.
Conclusion
By using the well-known [2+2] cyclocondensation for the
[43]
synthesis of homodimeric macrocycles and our own versa-
tile two-step approach to the heterodimeric analogues, we
prepared a collection of 20 bisintercalator-type macrocycles.
These compounds display a large molecular diversity, and
the heterodimeric scaffolds have been synthesized for the
first time. This diversity allowed the fine-tuning of the bind-
ing behavior of the ligands with regard to thymine-mis-
match-containing DNA duplexes and the establishment of
comprehensive structure–property guides. The dramatic dif-
ferences between the various series of macrocycles empha-
size the strong influence of the size and topology of the
macrocycle on mismatch recognition and suggest that a
subtle equilibrium exists between the structural features in
which selectivity originates. Finally, the novel compounds
are highly water soluble and thus have great potential for
use as modulators of mismatch-repair pathways or as molec-
ular diagnostic tools.
In summary, on the grounds of the present data and on
previous results indicating that binding of 2,7-BisA at a TT
mismatch induces extrahelical displacement of one thy-
[21]
mine, we can assume that a specific interaction is estab-
lished between macrocycles of the BisNP type and TT- or
TC-mismatched sites in DNA. Although the molecular mod-
eling studies do not address the mismatch versus fully
paired duplex selectivity, it can be hypothesized that the
short distance between the two aromatic units is involved in
the preference for the mispaired site, because a cyclic bisin-
tercalator with long polyamide linkers has recently been
shown to thread into intact DNA through opening of four
[57]
Watson–Crick base pairs.
For this reason, it is unlikely
that 1,5-BisNP could intercalate between Watson–Crick
base pairs and, thus, its higher propensity to bind fully
paired DNA is attributable to a particular topology suitable
Chem. Eur. J. 2010, 16, 878 – 889
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
887

M.-P. Teulade-Fichou et al.
Acknowledgements
[22] a) R. J. Roberts, X. D. Cheng, Annu. Rev. Biochem. 1998, 67, 181–
1
98; b) S. Klimasauskas, S. Kumar, R. J. Roberts, X. D. Cheng, Cell
We sincerely thank Dr. Jason Martin (Institut Curie) for proofreading
the manuscript and Professor Heiko Ihmels (Universitꢁt Siegen, Germa-
ny) for a donation of 2,7-dimethylnaphthalene, and we gratefully ac-
knowledge the generous financial support from the Centre National de la
Recherche Scientifique.
1994, 76, 357–369.
[23] a) T. Peng, K. Nakatani, Angew. Chem. 2005, 117, 7446–7449;
Angew. Chem. Int. Ed. 2005, 44, 7280–7283; b) K. Nakatani, S. Ha-
gihara, Y. Goto, A. Kobori, M. Hagihara, G. Hayashi, M. Kyo, M.
Nomura, M. Mishima, C. Kojima, Nat. Chem. Biol. 2005, 1, 39–43;
c) Y. Oka, T. Peng, F. Takei, K. Nakatani, Nucleic Acids Symp. Ser.
2
008, 52, 435–436.
24] V. C. Pierre, J. T. Kaiser, J. K. Barton, Proc. Natl. Acad. Sci. USA
007, 104, 429–434.
25] M. Bahr, V. Gabelica, A. Granzhan, M.-P. Teulade-Fichou, E. Wein-
hold, Nucleic Acids Res. 2008, 36, 5000–5012.
[
[
[
[
2
[
[
[
1] R. R. Iyer, A. Pluciennik, V. Burdett, P. L. Modrich, Chem. Rev.
006, 106, 302–323.
2] L. C. Young, J. B. Hays, V. A. Tron, S. E. Andrew, J. Invest. Derma-
tol. 2003, 121, 435–440.
3] a) H. T. Lynch, T. Smyrk, J. F. Lynch, Int. J. Cancer 1996, 69, 38–43;
b) S. Zhang, R. Lloyd, G. Bowden, B. W. Glickman, J. G. de Boer,
Environ. Mol. Mutagen. 2002, 40, 243–250.
2
26] A. Granzhan, M.-P. Teulade-Fichou, Chem. Eur. J. 2009, 15, 1314–
1
318.
27] M. H. Lim, H. Song, E. D. Olmon, E. E. Dervan, J. K. Barton,
Inorg. Chem. 2009, 48, 5392–5397.
[
[
28] J.-B. LePecq, C. Paoletti, J. Mol. Biol. 1967, 27, 87–106.
29] a) D. Suh, J. B. Chaires, Bioorg. Med. Chem. 1995, 3, 723–728;
b) E. C. Long, J. K. Barton, Acc. Chem. Res. 1990, 23, 271–273.
30] H. W. Zimmermann, Angew. Chem. 1986, 98, 115–131; Angew.
Chem Int. Ed. Engl. 1986, 25, 115–130.
[
[
[
4] Y. Pommier, DNA Repair Modulators as Anticancer Agents, ASC,
New York, 2008, pp. 128–133.
5] a) G. Damia, M. DꢂIncalci, Eur. J. Cancer 2007, 43, 1791–1801;
b) L. H. Hurley, Nat. Rev. Cancer 2002, 2, 188–200.
6] a) J. R. Hart, O. Glebov, R. J. Ernst, I. R. Kirsch, J. K. Barton, Proc.
Natl. Acad. Sci. USA 2006, 103, 15359–15363; b) C. A. Puckett, J. K.
Barton, J. Am. Chem. Soc. 2007, 129, 46–47; c) C. A. Puckett, J. K.
Barton, Biochemistry 2008, 47, 11711–11716; d) R. J. Ernst, H.
Song, J. K. Barton, J. Am. Chem. Soc. 2009, 131, 2359–2366.
7] a) B. A. Jackson, J. K. Barton, J. Am. Chem. Soc. 1997, 119, 12986–
[
[
[
31] a) L. S. Lerman, J. Mol. Biol. 1961, 3, 18–30; b) M. Aslanoglu, Anal.
Sci. 2006, 22, 439–443.
32] a) L. G. Lee, C.-H. Chen, L. A. Chiu, Cytometry 1986, 7, 508–517;
b) J. A. Bordelon, K. J. Feierabend, S. A. Siddiqui, L. L. Wright, J. T.
Petty, J. Phys. Chem. B 2002, 106, 4838–4843.
[
[
33] B. A. Armitage, Top. Curr. Chem. 2005, 253, 55–76.
34] a) J. Bontemps, C. Houssier, E. Fredericq, Nucleic Acids Res. 1975,
[
[
[
12987; b) B. A. Jackson, V. Y. Alekseyev, J. K. Barton, Biochemistry
1999, 38, 4655–4662; c) B. A. Jackson, J. K. Barton, Biochemistry
2000, 39, 6176–6182.
2
, 971–984; b) F. G. Loontiens, P. Regenfuss, A. Zechel, L. Dumorti-
er, R. M. Clegg, Biochemistry 1990, 29, 9029–9039.
[
35] a) W. D. Wilson, F. A. Tanious, H. J. Barton, R. L. Jones, K. Fox,
R. L. Wydra, L. Strekowski, Biochemistry 1990, 29, 8452–8461;
b) K. Jansen, B. Nordꢃn, M. Kubista, J. Am. Chem. Soc. 1993, 115,
8] a) H. Junicke, J. R. Hart, J. Kisko, O. Glebov, I. R. Kirsch, J. K.
Barton, Proc. Natl. Acad. Sci. USA 2003, 100, 3737–3742; b) C. Cor-
dier, V. C. Pierre, J. K. Barton, J. Am. Chem. Soc. 2007, 129, 12287–
1
0527–10530.
1
2295.
9] a) B. M. Zeglis, J. A. Boland, J. K. Barton, J. Am. Chem. Soc. 2008,
30, 7530–7531; b) B. M. Zeglis, J. A. Boland, J. K. Barton, Bio-
chemistry 2009, 48, 839–849.
[
36] L. Dassonneville, F. Hamy, P. Colson, C. Houssier, C. Bailly, Nucleic
Acids Res. 1997, 25, 4487–4492.
1
[
[
37] J. Cho, R. R. Rando, Nucleic Acids Res. 2000, 28, 2158–2163.
38] K. Yamashita, S. Sata, H. Takamiya, M. Takagi, S. Takenaka, Chem.
Lett. 2001, 680–681.
[
[
10] B. M. Zeglis, J. K. Barton, Inorg. Chem. 2008, 47, 6452–6487.
11] a) K. Nakatani, S. Sando, H. Kumasawa, J. Kikuchi, I. Saito, J. Am.
Chem. Soc. 2001, 123, 12650–12657; b) K. Nakatani, S. Sando, I.
Saito, Bioorg. Med. Chem. 2001, 9, 2381–2385; c) T. Peng, T.
Murase, Y. Goto, A. Kobori, K. Nakatani, Bioorg. Med. Chem. Lett.
[
[
39] E. Trotta, M. Paci, Nucleic Acids Res. 1998, 26, 4706–4713.
40] a) R. F. Pasternack, E. J. Gibbs, J. J. Villafranca, Biochemistry 1983,
2
3
2, 2406–2414; b) A. B. Guliaev, N. B. Leontis, Biochemistry 1999,
8, 15425–15437; c) B. Jin, H. M. Lee, J. Ko, C. Kim, S. K. Kim, J.
2
005, 15, 259–262.
Am. Chem. Soc. 2005, 127, 2417–2424.
[
12] S. Hagihara, H. Kumasawa, Y. Goto, G. Hayashi, A. Kobori, I.
Saito, K. Nakatani, Nucleic Acids Res. 2004, 32, 278–286.
13] A. Kobori, S. Horie, H. Suda, I. Saito, K. Nakatani, J. Am. Chem.
Soc. 2004, 126, 557–562.
14] K. Nakatani, S. Sando, I. Saito, Nat. Biotechnol. 2001, 19, 51–55.
15] Y. Goto, H. Suda, A. Kobori, K. Nakatani, Anal. Bioanal. Chem.
[
[
[
41] L. A. Lipscomb, F. X. Zhou, S. R. Presnall, R. J. Woo, M. E. Peek,
R. R. Plaskon, L. D. Williams, Biochemistry 1996, 35, 2818–2823.
42] A. A. Ouameur, H.-A. Tajmir-Riahi, J. Biol. Chem. 2004, 279,
[
4
2041–42054.
[
[
43] a) J. Jazwinski, J.-M. Lehn, R. Mꢃric, J.-P. Vigneron, M. Cesario, J.
Guilhem, C. Pascard, Tetrahedron Lett. 1987, 28, 3489–3492;
b) N. E. Borisova, M. D. Reshetova, Y. A. Ustynyuk, Chem. Rev.
2007, 107, 46–79.
2
007, 388, 1165–1173.
[
16] K. Nakatani, S. Horie, Y. Goto, A. Kobori, S. Hagihara, Bioorg.
Med. Chem. 2006, 14, 5384–5388.
[
17] a) X. L. Yang, R. B. Hubbard IV, M. Lee, Z. F. Tao, H. Sugiyama,
A. H. Wang, Nucleic Acids Res. 1999, 27, 4183–4190; b) E. R. Lacy,
K. K. Cox, W. D. Wilson, M. Lee, Nucleic Acids Res. 2002, 30, 1834–
[44] a) M.-P. Teulade-Fichou, J.-P. Vigneron, J.-M. Lehn, Supramol.
Chem. 1995, 5, 139–147; b) T. Paris, J.-P. Vigneron, J.-M. Lehn, M.
Cesario, J. Guilheim, C. Pascard, J. Inclusion Phenom. Macrocyclic
Chem. 1999, 33, 191–202.
[45] P. Cudic, J.-P. Vigneron, J.-M. Lehn, M. Cesario, T. Prangꢃ, Eur. J.
Org. Chem. 1999, 2479–2484.
[46] C. Cruz, R. Delgado, M. G. B. Drew, V. Fꢃlix, J. Org. Chem. 2007,
72, 4023–4034.
[47] A. Dahan, T. Ashkenazi, V. Kuznetsov, S. Makievski, E. Drug, L.
Fadeev, M. Bramson, S. Schokoroy, E. Rozenshine-Kemelmakher,
M. Gozin, J. Org. Chem. 2007, 72, 2289–2296.
1
841; c) E. R. Lacy, B. Nguyen, M. Le, K. K. Cox, C. OꢂHare, J. A.
Hartley, M. Lee, W. D. Wilson, Nucleic Acids Res. 2004, 32, 2000–
007.
18] L. A. Frederico, T. A. Kunkel, B. R. Shaw, Biochemistry 1993, 32,
523–6530.
2
[
[
[
6
19] N. Peyret, P. A. Seneviratne, H. T. Allawi, J. SantaLucia, Jr., Bio-
chemistry 1999, 38, 3468–3477.
20] a) N. Berthet, J. Michon, J. Lhomme, M.-P. Teulade-Fichou, J.-P. Vi-
gneron, J.-M. Lehn, Chem. Eur. J. 1999, 5, 3625–3630; b) M. Jour-
dan, J. Garcia, J. Lhomme, M.-P. Teulade-Fichou, J.-P. Vigneron,
J.-M. Lehn, Biochemistry 1999, 38, 14205–14213.
[48] Y.-S. Kim, K. M. Kim, R. Song, M. J. Jun, Y. Sohn, J. Inorg. Bio-
chem. 2001, 87, 157–163.
[49] A. Granzhan, M.-P. Teulade-Fichou, Tetrahedron 2009, 65, 1349–
1360.
[
21] A. David, N. Bleimling, C. Beuck, J.-M. Lehn, E. Weinhold, M.-P.
Teulade-Fichou, ChemBioChem 2003, 4, 1326–1331.
[50] J.-L. Mergny, L. Lacroix, Oligonucleotides 2003, 13, 515–537.
888
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 878 – 889

Macrocyclic DNA-Mismatch-Binding Ligands
FULL PAPER
[
[
51] a) D. L. Boger, B. E. Fink, S. R. Brunette, W. C. Tse, M. P. Hedrick,
J. Am. Chem. Soc. 2001, 123, 5878–5891; b) W. C. Tse, D. L. Boger,
Acc. Chem. Res. 2004, 37, 61–69.
[55] The video files showing the course of MD simulations are provided
as web-only Supporting Information.
[56] GOLD 4.1: a) G. Jones, P. Willett, R. C. Glen, J. Mol. Biol. 1995,
245, 43–53; b) G. Jones, P. Willett, R. C. Glen, A. R. Leach, R.
Taylor, J. Mol. Biol. 1997, 267, 727–748.
52] a) A. J. Geall, I. S. Blagbrough, J. Pharm. Biomed. Anal. 2000, 22,
8
49–859; b) D. Monchaud, C. Allain, H. Bertrand, N. Smargiasso, F.
Rosu, V. Gabelica, A. De Cian, J.-L. Mergny, M.-P. Teulade-Fichou,
Biochimie 2008, 90, 1207–1223.
[57] Y. Chu, D. W. Hoffman, B. L. Iverson, J. Am. Chem. Soc. 2009, 131,
3499–3508.
[
[
53] Marvin 5.2.4, ChemAxon (http://www.chemaxon.com), Budapest,
2
009.
54] Desmond Molecular Dynamics System, Version 2.2.6.2.3, D. E. Shaw
Received: July 17, 2009
Research, New York, 2009.
Published online: November 24, 2009
Chem. Eur. J. 2010, 16, 878 – 889
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
889