Tri- and tetra-substituted naphthalene diimides as potent G-quadruplex ligands

Article abstract of DOI:10.1016/j.bmcl.2008.01.050

A series of tri- and tetra-substituted naphthalene diimides have been designed and synthesized. Several compounds show exceptional affinity for telomeric G-quadruplex DNA in classical and competition FRET assays and SPR studies. They inhibit telomerase in the TRAP assay, and show potent senescence-based short-term anti-proliferative effects on MCF7 and A549 cancer cell lines, and localize in the nucleus and particularly the nucleolus of MCF7 cells.

Full text of DOI:10.1016/j.bmcl.2008.01.050

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1668–1673
Tri- and tetra-substituted naphthalene diimides as potent
G-quadruplex ligands
Francisco Cuenca,a Olga Greciano,a Mekala Gunaratnam,a Shozeb Haider,a
Deeksha Munnur,a Rupesh Nanjunda,b W. David Wilsonb and Stephen Neidlea,*
aCRUK Biomolecular Structure Group, The School of Pharmacy, University of London, 29/39 Brunswick Square,
London WC1N 1AX, UK
bDepartment of Chemistry, Georgia State University, Atlanta, GA 30303, USA
Received 20 December 2007; revised 10 January 2008; accepted 14 January 2008
Available online 18 January 2008
Abstract—A series of tri- and tetra-substituted naphthalene diimides have been designed and synthesized. Several compounds show
exceptional affinity for telomeric G-quadruplex DNA in classical and competition FRET assays and SPR studies. They inhibit tel-
omerase in the TRAP assay, and show potent senescence-based short-term anti-proliferative effects on MCF7 and A549 cancer cell
lines, and localize in the nucleus and particularly the nucleolus of MCF7 cells.
ꢀ 2008 Elsevier Ltd. All rights reserved.
Telomeres are highly specialized DNA-protein struc-
tures at the ends of eukaryotic chromosomes1 whose
integrity is required for the cell to avoid end-to-end fu-
sions and recombination.2 Telomere length progres-
sively shortens in somatic cells with successive rounds
of cell division, leading eventually to senescence and
apoptosis. In contrast telomere length is maintained in
cancer cells and is a major factor in immortalization
and tumorigenesis.3 The reverse transcriptase enzyme
telomerase is over-expressed in ca 80–85% of cancer cell
types4 where its ability to catalyze the synthesis of telo-
meric DNA repeats is responsible for telomere homo-
statis. Inhibition of telomerase induces cell senescence
and cell death in cancer cells and it is a validated target
for cancer therapeutics.5
complexes at the 30-single-strand telomeric DNA over-
hang displaces telomere-associated proteins such as
hPOT1 and telomerase.8 This unmasking of the 30-over-
hang invokes a rapid DNA damage response, which
rapidly leads to selective cell death and anti-tumour
activity in vivo.9
A number of naphthalene imide and diimide (ND)
monomer and dimer derivatives bind to duplex
DNA.10 Several show in vivo anti-cancer activity and
two (Amonafide and Elinafide) have been evaluated in
anti-cancer clinical trials in humans,11 although they
have not progressed due to toxicity combined with lim-
ited activity. A series of disubstituted NDs have been
screened as G4 ligands but showed only low affinity.12
We hypothesized that the planarity of the ND moiety
would nevertheless be an interesting starting point for
the development of more complex ligands since recent
chemical developments13 enabled us to envisage a syn-
thetic route to tetra-substituted analogues, which we ar-
gued, would possess high affinity for G4s such as the
parallel telomeric G4 structure which possesses four tar-
getable grooves.14 This supposition was supported by
molecular modelling studies15 (Fig. 1), and more re-
cently, by crystallographic analyses on several G4-ND
complexes.16 Furthermore, the synthetic route allows
the introduction to the ND core of up to three different
side chains, which in principle can be exploited to
produce ligand diversity that may discriminate between
One approach to telomerase inhibition involves seques-
tering its substrate, single-stranded telomeric DNA, by
inducing it to form G-quadruplex (G4) structures.6
Induction of a G-quadruplex conformation can be
achieved by small-molecule ligands; a large number have
been reported,7 although none as yet have reached the
clinic. It has been shown that the induction of G4-ligand
Keywords: G-quadruplex; Telomere; Telomerase inhibition.
*
Corresponding author. Tel.: +44 207 753 5969; fax: +44 207 753
5970; e-mail: bmcl@pharmacy.ac.uk
0960-894X/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.01.050

F. Cuenca et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1668–1673
1669
reason compound 2 was used for the synthesis of com-
pounds 23 and 24 where two different side chains were
used. For this two-step synthesis compound 2 was trea-
ted with the first amine in acetic acid and subsequently
with the second amine for the substitution at the chlo-
rines. The disubstituted compounds 25 and 26 previ-
ously reported,10a were synthesized using commercially
available naphthalene dianhydride and the correspon-
dent amine.
The initial assessment of the DNA stabilization ability
of the compounds was done using a FRET melting as-
say.19 We used two different DNA sequences, a human
telomeric G4 and a self-complementary duplex DNA
hairpin. We also ran FRET competition experiments20
in which affinity for the G4 sequence is evaluated in
the presence of increasing concentrations of duplex
DNA. Together, these experiments enabled us to assess
the selectivity of the ligands for G4 versus duplex DNA
(Table 1 and Fig. 2).
Figure 1. Molecular model of 11 bound to the 50 face of the human
intramolecular G4 structure (van der Waals surface coloured grey with
the 50 G-tetrad highlighted in yellow), after 5 ns of molecular dynamics
simulation. The MM-PBSA-calculated binding energy is
ꢀ60.6 kcal moleꢀ1, compared to ꢀ39.4 kcal moleꢀ1 for the acridine
ligand BRACO19.8a,9a
DTm values are given at ligand concentrations of 0.5 lM
since a high level of G4 DNA stabilization was observed
for a number of the ligands. For example, the DTm of
35.2 ꢁC for compound 7 contrasts with the behaviour
of the tri-substituted acridine BRACO19,8a,9a which
produced an equivalent DTm value at a concentration
of 2.5 lM. Overall, the tetra-substituted ligands (4-
ND) show superior G4 stabilization compared to the
tri-substituted ones (3-ND) and these in turn are supe-
rior to the di-substituted counterparts (2-ND). Com-
pounds with four side chains have higher stabilizing
ability regardless of the overall number of cationic
charges as the effect is observed for compounds 15/16,
17/18 and 21/22, which are uncharged or weakly cationic
at pH 7.4 (the pH at which the FRET experiments were
performed). Compounds 3, 7 and 11 and 4, 8 and 12
have the highest DTm within the 4-ND and 3-ND series,
respectively. They all have side chains of the same length
(3 carbons) and similar end groups (tertiary amines, pro-
tonated at pH 7.4). Their counterparts with shorter side
chains, 5, 9 and 13 and 6, 10 and 14, respectively, all give
lower DTm values. Side chains of at least three carbon
atoms are necessary to deliver the end groups to the
grooves of the G4, as the molecular modelling suggests
(Fig. 1). Substitution of the tertiary amines for morpho-
lino groups in compounds 15, 16, 17 and 18 reduces the
DTm, as observed previously in other series of G4
ligands.21 However, these morpholino and also the
hydroxyl analogues 21 and 22 all show a high level of
G4 stabilization. With the notable exception of
telomestatin, to our knowledge, very few other neutral
ligands show a similar degree of G4 affinity.7 The
2-substituted NDs examined here (compounds 25 and
26), show poor G4 affinity, in accord with earlier
observations.12
different types of G-quadruplexes. This feature will be
reported elsewhere.
We report here on a library of tri- and tetra-substituted
ND analogues accessible from commercially available
amines. The differences in the amines have provided
diversity in the library, enabling exploration of the ef-
fects of differing end groups, side-chain lengths, and to
some extent, non-equivalent side chains.
The synthesis of the final compounds (Table 1) was con-
ducted in one or two steps from the key intermediates
2,6-dibromo-1,4,5,8-naphthalene tetra-carboxylic acid
dianhydride (1) or 2,6-dichloro-1,4,5,8-naphthalene tet-
ra-carboxylic acid dianhydride (2) (Scheme 1). The tet-
ra-substituted analogues containing four equal side
chains could be synthesized from either 1 or 2 using neat
amine as a solvent and heating at 150 ꢁC for 10 min in a
microwave. The more economical dibromo compound 1
was preferentially used.
It was found that the tri-substituted (and occasionally
di-substituted) analogues were regularly obtained as
by-products in the reactions using 1. We hypothesized
that the occurrence of these sub-products was due to
radical debromination of the starting materials or inter-
mediates caused by traces of DBI used in a previous syn-
thetic step.13 Although initially unwanted, this side
reaction proved to be useful as the tri-substituted ana-
logues together with the targeted compounds were con-
sistently isolated from the crude reaction product using
HPLC.17
Most NDs reported here showed at least some interac-
tion with duplex DNA. Compounds 21 and 22 however,
slightly destabilized the duplex sequence. With the
exception of the pair of compounds 5/6, the tri-substi-
tuted NDs are more effective binders to the duplex
DNA than the tetra-substituted analogues, probably be-
It has been reported that the substitution reaction at the
anhydride groups of compound 1 can be successfully
decoupled from the reactions at the bromines.18 In our
hands, however, this reaction did not show the expected
selectivity and complex mixtures were obtained. For this

1670
F. Cuenca et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1668–1673
Table 1. ND compounds synthesized and the results of FRET melting studies, where DTm values are estimated as the mid-points of the melting
transition for Q (human G4 quadruplex) and d (a duplex DNA sequence)
Compound
Corea
Side chain
End groupc
FRET-Qd (ꢁC)
DTm (0.5 lM)
FRET-de (ꢁC)
DTm (0.5 lM)
nb
3
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
4-ND
3-ND
(2 + 2)-ND
3
DMA
DMA
DMA
DMA
DEA
DEA
DEA
DEA
Pyrr
33.2
26.5
28
4.5
12.5
7.5
4.2
4.2
11.5
4
4
3
5
2
6
2
17.2
35.2
26.2
30
7
3
8
3
9
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2
21
5
3
34.5
28.2
29.7
21.2
20.5
10.2
13.7
6.2
2
3
Pyrr
Pyrr
10.7
3.5
6.5
3.7
3.2
3
2
2
Pyrr
Mor
3
3
Mor
Mor
2
2
Mor
Pip
3
2
27.5
17.5
14.2
10.5
27.7
2.2
4.7
ꢀ0.5
ꢀ0.5
0.2
2
Pip
OH
5
5
OH
R1: 3
R2: 2
R1: 3
R2: 3
3
R1: DMA
R2: Pip
R1: DMA
R2: OH
DMA
DMA
24
(2 + 2)-ND
23.7
3.5
25
26
2-ND
2-ND
5.2
3.2
2.7
1
2
Values are the means of two determinations.
a Nomenclature used in the table corresponds to:
R1
R
N
R
N
R:
O
O
O
O
O
O
O
O
N
O
O
H
N
H
(CH2)n
2 N
R
R
R2
N
end group
H
N
R
O
N
R
O
N
R1
4-ND (R1=R2)
(2+2)-ND
3-ND
2-ND
Side Chain
b Number of carbons in the side chain linker.
c DMA, dimethylamine; DEA, diethylamine; Pyrr, pyrrolidine; Mor, morpholine; Pip, piperidine; OH, hydroxyl.
d See supplementary data for materials and methods.
R
N
R
N
cause of steric reasons. Compound 23, with mixed-
length side chains, has the highest G4/duplex selectivity
of all while being a strong G4 binder. The reason for this
selectivity is not understood and further investigation is
underway.
O
O
O
O
O
O
O
O
O
O
O
H
N
H
N
a)
Br
Cl
R
R
R
N
H
Br
Cl
N
R
N
R
O
O
O
1
O
O
The 4-NDs showed the best selectivity for G4 versus du-
plex DNA in the competition experiments, followed by
the 3-NDs and then the 2-NDs (Fig. 2). Compounds
15, 16 and 18, and particularly the hydroxyl compounds
21 and 22, stood out as the most G4 selective com-
pounds within the series.
R1
R1
O
N
O
O
O
O
N
O
H
N
b)
c)
Cl
R2
R2
N
H
Cl
O
N
O
O
O
2
O
N
O
R1
R1
Biosensor-SPR binding studies were conducted on a rep-
resentative set of NDs with an immobilized human G4
model sequence as previously described.22,23 A table of
binding (Ka) and dissociation rate constants (kd) is in-
Scheme 1. Synthesis of tri- and tetra-substituted naphthalene diimide
analogues. Reactions and conditions: (a) Amine RNH2, 150 ꢁC,
10 min, MW; (b) Amine R1NH2, acetic acid, 120 ꢁC, 10 min, MW;
(c) Amine R2NH2, 150 ꢁC, 10 min, MW.

F. Cuenca et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1668–1673
1671
120
100
80
60
40
20
0
7. 7 has a single very strong binding site with Ka of
3.1 · 107 Mꢀ1 and a weaker secondary site with a K ca
50· weaker. The binding of 25 is almost 10· lower with
Ka of 4.5 · 106 Mꢀ1 and a still weaker second binding
site. In agreement with the Tm studies, compound 3
has a Ka that is similar to but slightly weaker than that
for 7 while binding of 5 is somewhat weaker than for 3.
These results show that the 4-NDs are among the stron-
gest G4 binding compounds found to date.23 In agree-
ment with its Tm, 5, with a two-methylene linker, is
the weakest binding 4-ND, with the highest kd while 7,
with three methylenes, has the highest Ka and lowest kd.
A group of ligands representative of the diversity in the
library were also evaluated as telomerase inhibitors
using a modified TRAP assay24 (Table 2). Values of
EC50 between 12 and 28 lM were obtained with the
exception of compound 17 (>50 lM). No simple corre-
lation between the FRET and TRAP data is evident
but compound 17 is the poorest ligand in the FRET as-
say in the group. Two established G4 ligands, BRAC-
O198a,9a and TMPyP425, were also evaluated in the
TRAP assay and the EC50 values are consistent with a
recent report using a direct telomerase assay.26 The
ND compounds have potencies of the same order as
BRACO19 or TMPyP4. Note that these values are not
comparable with most previously published TRAP as-
say results, which have not taken into account the need
to remove ligand from the second (PCR) step of the as-
say. These results are more comparable with recent di-
rect telomerase assay data.26
1:1
1:10
1:100
1:300
Experiment
-20
Average values 4-NDs
Compound 15
Average values 3-NDs
Compound 16
Average values 2-NDs
Compound 18
Compound 21
Compound 22
Figure 2. Competition FRET results. Y-axis represents the G4
stabilization ability of the compounds when in presence of a duplex
DNA competitor (in the ratios depicted in X-axis, G4:duplex)
normalized using 100% for the experiment without duplex competitor.
See Supplementary Data for a full description of methods and results.
cluded in the Supplementary Data. The sensorgrams for
all compounds reach a steady-state plateau between 100
and 200 s after initiation of compound injection. The
steady-state RU values were determined by averaging
in the steady-state region and were plotted versus the
free compound concentration in the flow solution for
determination of the Ka values. Sensorgrams for binding
of 4-ND 7 and 2-ND 25 are compared in Figure 3 and a
sensorgram plus a binding plot for compound 8 are
shown in the Supplementary Data. Although both com-
pounds in Figure 3 bind to the G4, significant differences
can be seen in the sensorgrams, especially in the dissoci-
ation rate, which is much slower for 7 than for 25. Be-
cause of surface absorption of the compounds in the
initial period of injection, it is not possible to quantita-
tively determine their association kinetics constants. For
the dissociation reaction, however, it is clear that the 2-
ND dissociates in the first few sec of buffer flow while
the apparent half-life for dissociation of the 4-ND is
approximately 35 s. The 3-NDs have kd rates that are
more similar to the 4-NDs than the 2-NDs. An excep-
tion is compound 22, which has considerably weaker
binding and faster dissociation rates from the G4 than
The ability of the compounds to produce short-term cell
growth inhibition against two cancer cell lines (MCF7
and A549, both of which express telomerase), and a nor-
mal human fibroblast line (WI38) was assessed using the
sulforhodamine B (SRB) assay. The compounds as a
class show potent cell growth arrest, especially in the
cancer cell lines (Table 3), although there is wide varia-
tion in behaviour, with the most potent compounds such
as 4, 5, 12–14 having IC50 values of 5–10 nM. The
MCF7 cell line is consistently the most sensitive. In gen-
eral, potency correlates with high G4 affinity. This is
especially evident for the morpholino compounds 16–
18, which are relatively non-toxic in the three cell lines
(although the cell uptake of these compounds may be
low—see below).
Table 2. TRAP telomerase inhibition resultsa
Compound
TRAP EC50 (lM)
3
25
17
20
15
10
5
4
7
21.3
27.9
21
8
11
13
17
12.3
>50b
7.9
BRACO19
TMPyP4
0
16.4
0
100 200 300 400 500
0
100 200 300 400 500 600
a Values are the averages of two independent experiments, apart from
13.
b No inhibition observed at concentrations of up to 50 lM.
Time (seconds)
Time (seconds)
Figure 3. SPR sensorgrams for compounds 7 and 25.

1672
F. Cuenca et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1668–1673
Table 3. SRB and senescence data for the ND compounds
Compound
IC50 (nM)a
MCF7 A549
Senescence
(% of cells)
after 1 wk treatment
WI38
3
104.7
5.4
9.2
28.7
13.3
15.3
25.1
10.7
48
292.3
42.8
40.1
137.5
86.7
83.7
971.9
466.8
167.2
62.6
63.4
63.9
965
35.0 3 (at 70 nM)
4
n.a.b
n.a.
5
6
18.5
18
n.a.
n.a.
Figure 4. Cell uptake detection using fluorescence confocal micros-
copy. (1) Transmission/fluorescence composite image of 11 localized in
the nucleus of MCF7 cells after 30 min exposure at 0.5 lM; (2) non-
specific and low uptake of 17 after 30 min exposure at 50 lM; (3)
Image of 3 localized in the nucleolus after 30 min at 0.5 lM.
7
8
17.9
219
35 3 (at 15 nM)
n.a.
9
273.9
144
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
164.4
81.3
10.5
10.2
11
n.a.
45 6 (at 75 nM)
115.3
96
n.a.
n.a.
n.a.
n.a.
n.a.
BRACO19, and significantly greater than the effects
produced by TMPyP4 and telomestatin (data not
shown). Whether this is a significant factor in the po-
tency shown by a number of these NDs in inhibiting
cancer cell growth, remains to be demonstrated. The
flexible synthetic route described here may yield new li-
gands with improved G4/duplex selectivity, which will
help to further define their mechanism of action. The
crystal structures of G4-ND complexes16 may also facil-
itate future structure-based design studies.
13.6
18.3
1190
1320
437
800
1220
6700
1610
295.7
128.9
8300
9200
26.9
287.7
135.2
118.9
>10000 >10000 51 3 (at 3.5 lM)
2750
204.5
145.6
8500
1460
438.9
n.a.
n.a.
n.a.
>10000 >10000 n.a.
>10000 >10000 n.a.
57.1
57.7
n.a.
n.a.
n.a.
n.a.
1700
113.8
99.1
10000
210.9
171.1
Acknowledgments
a Standard deviations omitted for clarity.
b n.a., not available.
We are grateful to CRUK, EU FP6 (Molecular Cancer
Medicine) and the Heptagon Fund for support, and
Drs. J.-L. Mergny and A. de Cian for preliminary Tm
experiments and much useful discussion.
The observation of cellular senescence as a result of G4
ligand action at telomeres is well-documented.8,27 The
senescence phenotype was investigated in MCF7 cells
treated for one week with sub-cytotoxic concentrations
for a few compounds.28 All those evaluated showed a
significant increase in the percentage of senescence cells
after treatment (Table 3), surprisingly even for those
with higher IC50 values. We conclude that the senes-
cence effects are consistent with telomerase inhibition
and that the ability of NDs to selectively kill cancer cells
is at least in part a consequence of telomere targeting
and telomerase uncapping effects.
Supplementary data
Supplementary data for this article can be found online
at, doi:10.1016/j.bmcl.2008.01.050.
References and notes
1. McEachern, M. J.; Krauskopf, A.; Blackburn, E. H. Ann.
Rev. Genetics 2000, 34, 331.
2. Smogorzewska, A.; De Lange, T. Annu. Rev. Biochem.
2004, 73, 177.
3. Hahn, W. C.; Counter, C. M.; Lundberg, A. S.; Beijers-
bergen, R. L.; Brooks, M. W.; Weinberg, R. A. Nature
1999, 400, 464.
4. Kim, N. W.; Piatyszek, M. A.; Prowse, K. R.; Harley, C.
B.; West, M. D.; Ho, P. L. C.; Coviello, G. M.; Wright, W.
E.; Shay, J. W. Science 1994, 266, 2011.
5. (a) Hahn, W. C.; Stewart, S. A.; Brooks, M. W.; York, S.
G.; Eaton, E.; Kurachi, A.; Beijersbergen, R. L.; Knoll, J.
H. M.; Meyerson, M.; Weinberg, R. A. Nat. Med. 1999, 5,
1164; (b) Shay, J. W.; Wright, W. E. Nat. Rev. Drug
Discov. 2006, 5, 577.
6. (a) Zahler, A. M.; Williamson, J. R.; Cech, T. R.; Prescott,
D. M. Nature 1991, 350, 718; (b) Sun, D.; Thompson, B.;
Cathers, B. E.; Salazar, M.; Kerwin, S. M.; Trent, J. O.;
Jenkins, T. C.; Neidle, S.; Hurley, L. H. J. Med. Chem.
1997, 40, 2113; (c) Oganesian, L.; Bryan, T. M. Bioessays
2007, 25, 155.
7. Comprehensively reviewed in: De Cian, A.; Lacroix, L.;
Douarre, C.; Temime-Smaali, N.; Trentesaux, C.; Riou,
J.-F.; Mergny, J.-L. Biochimie. 2007, in press.
The ND-3 and ND-4 compounds reported here show
profound fluorescence. This facilitated localization of a
selected group of compounds within MCF7 cells using
confocal microscopy. Compounds 3, 8 and 11 localize
exclusively in the nucleus following 30 min exposure at
a concentration of 0.5 lM (Figure 4-1). For 17, less in-
tense and more widespread distributed fluorescence
was observed upon cell exposure even up to a concentra-
tion of 50 lM (Figure 4-2). This may partially account
for the lower toxicity of 17 and the other morpholino
analogues 15, 16 and 18. The compounds that concen-
trated in the nucleus showed preference for the nucleolus
(Figure 4-3). It may be significant for their action that
telomerase trafficking is associated with the nucleolus.29
We have shown here that a number of tri- and tetra-
substituted NDs are exceptional G4 binding agents, with
DTm values for the human intramolecular G4 that are at
least twofold greater than that of the acridine compound

F. Cuenca et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1668–1673
1673
8. (a) Gunaratnam, M.; Greciano, O.; Martins, C.; Reszka,
A. P.; Schultes, C. M.; Morjani, H.; Riou, J.-F.; Neidle, S.
Biochem. Pharmacol. 2007, 74, 679; (b) Brassart, B.;
Gomez, D.; De Cian, A.; Paterski, R.; Montagnac, A.;
Qui, K. H.; Temime-Smaali, N.; Trentesaux, C.; Mergny,
J.-L.; Gueritte, F.; Riou, J.-F. Mol. Pharmacol. 2007, 72,
631.
9. (a) Burger, A. M.; Dai, F.; Schultes, C. M.; Reszka, A. P.;
Moore, M. J.; Double, J. A.; Neidle, S. Cancer Res. 2005,
65, 1489; (b) Phatak, P.; Cookson, J. C.; Dai, F.; Smith,
V.; Gartenhaus, R. B.; Stevens, M. F. G.; Burger, A. M.
Brit. J. Cancer 2007, 96, 1223.
10. See for example: (a) Yen, S.-F.; Gabbay, E. J.; Wilson, W.
D. Biochemistry 1982, 21, 2070; (b) Tanious, F. A.; Yen, S.
F.; Wilson, W. D. Biochemistry 1991, 30, 1813; (c) Lee, J.;
Guelev, V.; Sorey, S.; Hoffman, D. W.; Iverson, B. L.
J. Amer. Chem. Soc. 2004, 126, 14036.
11. (a) Brana, M. F.; Ramos, A. Curr. Med. Chem. Anticancer
Agents 2001, 1, 237; (b) Van Quaquebeke, E.; Mahieu, T.;
Dumont, P.; Dewelle, J.; Ribaucour, F.; Simon, G.;
Sauvage, S.; Gaussin, J.-F.; Tuti, J.; El Yazidi, M.; Van
Vynckt, F.; Mijatovic, T.; Lefranc, F.; Darro, F.; Kiss, R.
J. Med. Chem. 2007, 50, 4122.
C), 125.79 (2· C), 149.19 (2· C), 163.05 (2· C@O), 166.12
(2· C@O); HRMS (ES+) calcd C34H52N8O41 [M+H]+
637.4190. Found: 637.4199. Analytical data 4: H NMR
(CDCl3) d: 1.90 (5q, 2H, J = 7.2 Hz), 1.91 (5q, 2H,
J = 7.5 Hz), 1.96 (5q, 2H, J = 6.9 Hz), 2.24 (s, 6H), 2.26
(s, 6H), 2.27 (s, 6H), 2.41–2.47 (m, 6H), 3.67 (m, 2H), 4.23
(m, 4H), 8.27 (s, 1H), 8.32 (d, 1H, J = 7.8 Hz), 8.63 (d, 1H,
J = 7.8 Hz), 10.21 (t, 1H, J = 5.5 Hz); 13C NMR (CDCl3)
d: 26.00 (CH2), 26.08 (CH2), 27.48 (CH2), 38.68 (CH2),
39.25 (CH2), 41.32 (CH2), 45.36 (2· CH3), 45.41 (2· CH3),
45.48 (2· CH3), 56.65 (CH2), 57.22 (CH2), 57.31 (CH2),
99.88 (C), 119.42 (C), 119.97 (CH), 123.56 (C), 124.36
(CH), 126.18 (C), 127.93 (C), 129.57 (C), 131.22 (CH),
152.44 (C), 162.99 (C@O), 163.05 (C@O), 163.39 (C@O),
166.12 (C@O); HRMS (ES+) calcd C29H40N6O4 [M+H]+
537.3189. Found: 537.3217.
18. Jones, B. A.; Facchetti, A.; Marks, T. J.; Wasielewski, M.
R. Chem. Mater. 2007, 19, 2703.
19. The FRET DNA melting assay was performed as described
previously Schultes, C. M.; Guyen, B.; Cuesta, J.; Neidle, S.
Bioorg. Med. Chem. Lett. 2004, 14, 4347, The full protocol is
given in the Supplementary Data. Tagged DNA sequences
were used, 50-FAM-d(GGG[TTAGGG]3)-TAMRA-30 for
the G4 and 50-FAM-dTATAGCTATA-HEG-TATAGC-
TATA-TAMRA-30 (HEG linker: [(-CH2-CH2-O-)6]) for the
duplex experiment. The assay employs 96-well plates and
was processed with a DNA Engine Opticon instrument (MJ
Research).
12. Sissi, C.; Lucatello, L.; Paul, K. A.; Maloney, D. J.; Boxer,
M. B.; Camarasa, M. V.; Pezzoni, G.; Menta, E.;
Palumbo, M. Bioorg. Med. Chem. 2007, 15, 555.
13. Thalacker, C.; Roger, C.; Wurthner, F. J. Org. Chem.
2006, 71, 8098.
14. Parkinson, G. N.; Lee, M. P.; Neidle, S. Nature 2002, 417,
876.
20. Assay performed as described above for the G4 experi-
ment but with addition of varying concentrations of calf
thymus DNA (Sigma-Aldrich, UK). Moorhouse, A. D.;
Santos, A. M.; Gunaratnam, M.; Moore, M.; Neidle, S.;
Moses, J. E. J. Am. Chem. Soc. 2006, 128, 15972.
21. For example: Harrison, R. J.; Gowan, S. M.; Kelland, L.
R.; Neidle, S. Bioorg. Med. Chem. Lett. 1999, 9, 2463.
22. Moore, M. J.; Schultes, C. M.; Cuesta, J.; Cuenca, F.;
Gunaratnam, M.; Tanious, F. A.; Wilson, W. D.; Neidle,
S. J. Med. Chem. 2006, 49, 582.
23. White, E. W.; Tanious, F.; Ismail, M. A.; Reszka, A. P.;
Neidle, S.; Boykin, D. W.; Wilson, W. D. Biophys. Chem.
2007, 126, 140.
24. This TRAP assay includes a ligand removal step between
elongation and PCR to avoid ligand interference effects
(Reed et al., to be published).
15. The crystal structure of the 22-mer intramolecular paral-
lel-stranded human telomeric G4 DNA (PDB id 1KF1)
was used to model interactions with ND ligands. Ligand
structures were randomly positioned over the 50 terminal
G-tetrad. A simulated annealing docking protocol located
and then optimized ligand placements, followed by 5–
10 ns molecular dynamics simulations. The MM-PBSA
method was used to calculate intermolecular interaction
energies.
16. Crystal structures of intra- and intermolecular G4 ND
ligand complexes have been solved in this laboratory, and
will be reported separately. All show G4s with parallel
topology, in accord with the structural model shown here.
17. Synthesis and analytical data for compounds 3 and 4.
Data on other compounds are available as Supplementary
Data. Compound 1 (100 mg, 0.234 mmol) was suspended
in N,N-dimethyl-1,3-propanediamine (0.5 ml) in a micro-
wave reaction vessel, then flushed with nitrogen, sealed
and treated at 150 ꢁC for 10 min in the microwave. The
amine was then evaporated off under high vacuum. The
crude mixture was purified by HPLC to obtain 3 and 4 as
a blue and an orange solid, respectively. Yield 3 (23.1 mg,
0.036 mmol, 15.5%), 4 (12.6 mg, 0.024 mmol, 10.0%).
Analytical data 3: 1H NMR (CDCl3) d: 1.90 (5q, 4H,
J = 7.4 Hz), 1.94 (5q, 4H, J = 7.0 Hz), 2.26 (s, 12H), 2.27
(s, 12H), 2.44 (m, 8H), 3.57 (m, 4H), 4.22 (m, 4H), 8.16 (s,
2H), 9.41 (t, 2H, J = 5.1 Hz); 13C NMR (CDCl3) d: 26.10
(2 · CH2), 27.51 (2 · CH2), 38.71 (2· CH2), 41.25 (2·
CH2), 45.41 (4· CH3), 45.52 (4· CH3), 56.99 (2· CH2),
57.32 (2· CH2), 101.93 (2· C), 118.37 (2· CH), 121.17 (2·
25. Shi, D. F.; Wheelhouse, R. T.; Sun, D.; Hurley, L. H. J.
Med. Chem. 2001, 44, 4509.
26. De Cian, A.; Cristofari, G.; Reichenbach, P.; De Lemos,
E.; Mondchaud, D.; Teulade-Fichou, M. P.; Shin-Ya, K.;
Lacroix, L.; Lingner, J.; Mergny, J.-L. Proc. Natl. Acad.
Sci. U.S.A. 2007, 104, 17347.
27. Riou, J.-F.; Guittat, L.; Mailliet, P.; Laoui, A.; Renou, E.;
´ `
Petitgenet, O.; Megnin-Chanet, F.; Helene, C.; Mergny, J.-
L. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 2672.
28. 1 · 105 cells per well were cultivated in 2 ml of medium
plus compound in 96-well plates (Fisher Scientific, UK).
After 24 h the cells were stained for senescence with a b-
galactosidase kit (Cell Signalling Technology). Senescent
cells, stained blue, were quantified by microscopy.
29. Tomlinson, R. L.; Ziegler, T. D.; Supakorndej, T.; Terns,
R. M.; Terns, M. P. Mol. Cell. Biol. 2006, 17, 955.