A hydrophilic three side-chained triazatruxene as a new strong and selective G-quadruplex ligand

Article abstract of DOI:10.1039/b904723a

A new hydrosoluble triazatruxene derivative (AZATRUX) is reported to selectively bind to G-quadruplex DNA, as derived by ESI-MS measurements and competition experiments. The Royal Society of Chemistry 2009.

Full text of DOI:10.1039/b904723a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
A hydrophilic three side-chained triazatruxene as a new strong and selective
G-quadruplex ligand†
Luca Ginnari-Satriani, Valentina Casagrande, Armandodoriano Bianco, Giancarlo Ortaggi and
Marco Franceschin*
Received 9th March 2009, Accepted 1st May 2009
First published as an Advance Article on the web 13th May 2009
DOI: 10.1039/b904723a
A new hydrosoluble triazatruxene derivative (AZATRUX) is
reported to selectively bind to G-quadruplex DNA, as derived
by ESI-MS measurements and competition experiments.
studies suggest a ligand/G-quadruplex general model in which
ligands interact mainly through terminal p–p stacking at the
end of the quadruplex (threading intercalation), without internal
intercalation, and by electrostatic interactions of the side chains in
the grooves.6 Classical G-quadruplex ligands, such as substituted
acridines, triazines, porphyrins, as well as perylene,7 coronene8 and
berberine9 derivatives, present an aromatic core equipped with
hydrophilic side chains.10
In recent years, triazatruxene (Fig. 2) derivatives have been of
great interest in supramolecular chemistry and in particular in
organic electronics, with regard to their potential applications
as light-emitting diodes,11 battery and capacitors.12 For their
intrinsic photophysical and redox properties and their p-stacking
capability,13 these molecules can also behave as electroactive
discotic liquid-crystals.14 More recently, a triaza-fullerene was
synthesized, using the triazatruxene core as a precursor.15 Never-
theless, so far, triazatruxene derivatives have not been reported for
important pharmaceutical applications and we thought that this
kind of compound could represent a useful basis to develop new
G-quadruplex ligands, provided suitable hydrophilic side chains
were added to the triazatruxene moiety.
Introduction
G-quadruplexes are a family of nucleic acid secondary struc-
tures stabilized by G-tetrads, coplanar quartets of guanines
held together by a cyclic arrangement of eight unconventional
hydrogen bonds (Hoogsteen bonds, Fig. 1).1 These structures are
stabilized by the presence of monovalent cations and the stack-
ing interactions between overlapping G-tetrads: G-quadruplexes
can be formed by G-rich strands of DNA as the telomeric
sequences.2 Telomeric single-strand DNA is the substrate of
telomerase, an enzyme necessary for telomeric replication, which
is over-expressed in most cancer cells and participates in tumor
genesis.3 The formation of a telomeric G-quadruplex blocks
telomerase activity and offers an original strategy for new anti-
cancer agents.4 During the last few years, several families of
compounds have been identified which specifically bind to the
telomeric quadruplex. These derivatives, called “G-quadruplex
DNA ligands”, are able to block telomeric replication in cancer
cells and to cause replicative senescence and/or apoptosis after
a few cell cycles.5 Crystallographic and molecular modelling
Fig. 2 Chemical structure of the triazatruxene core, showing its complete
numbering.
Results and discussion
The triazatruxene core consists of a C3 symmetric cyclotrimer
of indoles, which presents a wide aromatic surface with three
useful points for the attachment of side chains, namely the
three indolic NH at 5, 10 and 15 positions (Fig. 2). Previously,
the synthesis of this core has been reported by reaction of
indole with bromine,16 leading to a mixture of brominated
symmetric indole trimers, and subsequent dehalogenation.17 In
our experience, this way resulted in uncontrolled reactions with
poor yields. So, we have considered another synthetic strategy,
starting from N-substituted-2-indolones, which are reacted with
Fig. 1 Schematic representation of a G-tetrad, showing Hoogsteen
hydrogen bonds.
Dipartimento di Chimica, SAPIENZA Universita` di Roma, Piazzale A.
Moro 5, 00185, Roma, Italy. E-mail: marco.franceschin@uniroma1.it;
Fax: +39-06-4991-3841; Tel: +39-06-4991-3341
† Electronic supplementary information (ESI) available: Detailed synthetic
procedures, molecular modelling simulations protocol, full characteriza-
tion of the synthesized compounds, 1H-NMR spectrum of AZATRUX as
a hydrochloride in D2O, full data of ESI-MS competition experiments and
in vitro antitumor activities. See DOI: 10.1039/b904723a
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2513–2516 | 2513
©

View Article Online
phosphoryl chloride to give the corresponding tris-N-susbstituted-
triazatruxenes.18 Since the preparation of N-substituted precursors
with hydrophilic chains can strongly limit the variability of the
molecular features of possible substituents,19 we have found more
convenient the symmetric cyclotrimerization of 2-indolone 1 in
◦
POCl3 at 100 C,20 leading to the unsubstituted azatruxene core
2 in 48% yield (Scheme 1). Attachment of the basic side chains
consists of two steps: the N-alkylation in basic conditions with
an excess of diiodobutane to give the corresponding tris-N-
substituted derivative and subsequent substitution of the iodide
at the end of the chain with an excess of piperidine to get the
target compound 3.‡ It is worth noting that the proposed pathway
represents a versatile synthetic approach for the preparation of
many compounds of this series with different side chains in terms
of length and basicity, which are two critical parameters for
modulating G-quadruplex ligands properties.7
Fig. 3 Complex of AZATRUX (stick model with yellow transparent
surface) with the human monomeric G-quadruplex DNA25 (red surface),
obtained by simulated annealing:8 (a) top view, (b) lateral view.
Scheme 1 Synthetic scheme for the preparation of AZATRUX.
Fig. 4 UV-vis absorption spectra of AZATRUX in MES-KCl aqueous
buffer at different temperatures.
The new compound AZATRUX shows optimal molecular
features for the interaction with the G-quadruplex, as confirmed
by molecular modelling simulations:8 the triazatruxene core stacks
upon the terminal G-tetrad (Fig. 3a), while the hydrophilic side
chains fit the DNA grooves (Fig. 3b). Since drug self-aggregation
has been shown to be related to a lower telomerase inhibition
and weaker interactions with the G-quadruplex with respect to
unstacked molecules,21 the self-aggregation in aqueous solution of
G-quadruplex ligands must be carefully considered in the model of
their interaction with G-quadruplex DNA.7,8 For this reason, we
have studied the self association of AZATRUX in MES-KCl buffer
by UV-vis absorption spectroscopy (Fig. 4). The poor variability
of the spectrum appearance when increasing temperature suggests
that no significant self-association occurs, as confirmed by the
clearly resolved aromatic region of the 1H-NMR spectrum in D2O
(Fig. S1, ESI†).22
The efficiency of AZATRUX in binding the G-quadruplex
and its selectivity with respect to duplex DNA have been stud-
ied by electrospray ionization mass spectrometry (ESI-MS),23
according to the protocol described in a previous paper.24
For this study, we have chosen two oligonucleotides that can
form different G-quadruplex structures: TG4T (5¢-TGGGGT-
3¢) which gives a tetrameric G-quadruplex and F21T (5¢-
GGGTTAGGGTTAGGGTTAGGG-3¢) which is composed by
human telomeric repeats and is able to fold in a monomeric
G-quadruplex structure, characterized by X-ray crystallogra-
phy in K+ solution.25 The formation of stable complexes be-
tween azatrux and the G-quadruplex forming oligonucleotides
is clearly demonstrated by the presence in the mass spectra (at
1:1 drug/quadruplex ratio) of intense peaks corresponding to
2514 | Org. Biomol. Chem., 2009, 7, 2513–2516
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
with this kind of model.24 In order to achieve a more reliable study
of selectivity, we have performed competition experiments with
the same technique, carried out in the simultaneous presence of G-
quadruplex forming oligonucleotides and fragments of a double-
stranded genomic DNA in 1:1 and 1:5 ratios, calculated on the
basis of the phosphate group concentrations. Provided that calf
thymus DNA (CT) cannot be detected in the used experimental
conditions due to its high molecular weight, it is possible to
report the percentage of bound quadruplex DNA28 in the presence
of different duplex concentrations (Fig. 6): any decrease of this
percentage must be caused by the drug binding to genomic duplex
DNA.24
Table 1 K1 and K2 values on a logarithmic scale for the complexes
between AZATRUX and the indicated oligonucleotides as derived by ESI-
MS experiments24
a
a
Oligo
Log K1
Log K2
(TG4T)4
F21T
DK66
6.3 0.2
5.5 0.1
4.1 0.1
5.3 0.2
5.0 0.1
n.d.
a Standard deviations are reported over at least three independent experi-
ments.
drug-quadruplex complexes (Fig. 5 and Fig. S2, ESI†). The
evaluation of the binding constants obtained by the collected
data (Table 1)26 demonstrates that the studied molecule is a good
G-quadruplex ligand able to form both 1:1 and 2:1 drug–DNA
complexes. This agrees with the general model of a G-quadruplex–
ligand interaction characterized by the presence of two binding
sites on the external tetrad surfaces of the G-quadruplex structures
(Fig. 3), even though other explanations are possible for the 2:1
stoichiometry, since the external tetrads of the quadruplex are not
identical. The values of K1 are above 105 for F21T and above 106
for (TG4T)4, while the order of magnitude for K2 is 105 with both
the oligonucleotides. Comparing K1 and K2 we can assume that
no cooperativity is involved in the binding mechanism since K2
values are always lower than those for K1.24
Fig. 6 Competition experiments on F21T and (TG4T)4 oligos. Histogram
diagrams reporting the percentage of bound quadruplex DNA24 (%bound
G-q) for samples containing a fixed amount of both drug and G-quadru-
plex DNA (5 mM, 1:1 ratio) and different amounts of calf thymus DNA
(CT), at the indicated duplex/quadruplex ratios (in phosphate ions). All
values should be considered with an error estimated on at least three
independent experiments of about 5%.
Fig. 5 ESI mass spectra of DK66 in the presence of AZATRUX at 1:4
molar ratio (top) and F21T in the presence of AZATRUX at 1:1 molar
ratio (bottom). Arrows indicate the peaks corresponding to the oligos
alone and to the relative 1:1 or 2:1 drug/DNA complexes, in different
charge states.
The analysis of the measures performed on both oligonu-
cleotides shows that the percentage value of quadruplex bound
at a 1:1 duplex/quadruplex ratio has a poor decrease: from 75
to 64 with TG4T and from 46 to 39 with F21T, which is more
than 80% left with respect to the value in the absence of CT
(Tables S1 andS2, ESI†). This result confirms the strong preference
of this molecule for the quadruplex structure of DNA. At a 5:1
duplex/quadruplex ratio, AZATRUX is still able to bind 44%
of the quadruplex DNA formed in the sample by TG4T, which
is about 2/3 of the quadruplex bound in the absence of duplex
DNA, while in the case of F21T the percentage of quadruplex
bound is remarkably reduced, probably due to a weaker affinity
for this oligonucleotide with respect to TG4T, even alone (Fig. S3,
ESI†).
As an independent confirmation of this quadruplex/duplex
selectivity, we have performed titrations of the human telomeric
F21T oligonucleotide and the biologically relevant CT DNA with
AZATRUX by UV spectroscopy (Fig. S4, ESI†). We noticed that
the binding curve slope for quadruplex DNA is steeper than for
duplex DNA and thus saturation is reached at a lower ratio in
the first case. This different behaviour of AZATRUX towards the
two DNA structures confirms a stronger interaction for telomeric
In order to evaluate the selectivity of AZATRUX for quadruplex
over duplex DNA we have studied preliminarily its affinity for a self
complementary dodecamer: DK66 (5¢-CGCGAATTCGCG-3¢),
one of the simplest duplex models widely used in the literature.27
In this case, the spectra acquired at 1:1 ratios do not show any trace
of the 1:1 or 2:1 complex peaks, suggesting a weaker interaction
with respect to that revealed for G-quadruplexes. In order to detect
an appreciable peak of the 1:1 drug/DNA complex, this molecule
must be present in the sample at 3:1 and 4:1 ratios, but even at
this concentration, there is no evidence of the peak relative to
the 2:1 complex (Fig. 5). The order of magnitude for K1 with
DK66 is 104, about 25 times lower than that of K1 with F21T
and more than 150 times lower than K1 with TG4T (Table 1).
These values suggest a good selectivity for quadruplex structures
with respect to duplex DNA. Nevertheless, we have previously
shown that a short duplex oligonucleotide is a really simplified
model and physiologically nonrelevant interactions are possible
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2513–2516 | 2515
©

View Article Online
Table 2 Growth percent of the indicated cell lines after 48 hours of
incubation with 10 mM AZATRUX solution, with respect to no-drug
control, as reported by a preliminary study at the National Cancer Institute
(NIH, Bethesda, MD, USA)29
are due to Manuel Volpe for help in the final revision of the
manuscript, to Lorenzo Velluti for graphical aid and to Anita
Scipioni for helpful discussion.
Cell line
Tumour
Growth percent (%) a
Notes and references
RPMI-8226
K-562
A549/ATCC
HOP-62
HCT-15
HCT-116
U251
leukemia
leukemia
-49.4
-81.8
-49.7
-91.4
-68.4
-97.1
-68.4
-98.7
-56.5
-96.6
-31.4
-98.8
-54.5
-95.2
-93.2
-95.2
-48.0
-93.0
‡ Details on materials, instrumentation and synthetic procedures have been
provided in the Electronic Supplementary Information†, together with the
full characterization of all the intermediate compounds.
lung cancer
lung cancer
colon cancer
colon cancer
CNS cancer
CNS cancer
melanoma
1 J. L. Huppert, Chem. Soc. Rev., 2008, 37, 1375.
2 S. Neidle and G. N. Parkinson, Curr. Opin. Struct. Biol., 2003, 13,
275.
3 J. L. Mergny, J. F. Riou, P. Mailliet, M. P. Teulade-Fichou and E.
Gilson, Nucleic Acids Res., 2002, 30, 839.
4 S. Neidle and G. Parkinson, Nat. Rev. Drug Discov., 2002, 1, 383.
5 C. M. Incles, C. M. Schultes and S. Neidle, Curr. Opin. Investig. Drugs,
2003, 4, 675.
6 S. M. Haider, G. N. Parkinson and S. Neidle, J. Mol. Biol., 2003, 326,
117.
7 L. Rossetti, M. Franceschin, S. Schirripa, A. Bianco, G. Ortaggi and
M. Savino, Bioorg. Med. Chem. Lett., 2005, 15, 413; M. Franceschin,
E. Pascucci, A. Alvino, D. D’Ambrosio, A. Bianco, G. Ortaggi and M.
Savino, Bioorg. Med. Chem. Lett., 2007, 17, 2515; M. Franceschin, C.
M. Lombardo, E. Pascucci, D. D’Ambrosio, E. Micheli, A. Bianco, G.
Ortaggi and M. Savino, Bioorg. Med. Chem., 2008, 16, 2292.
8 M. Franceschin, A. Alvino, V. Casagrande, C. Mauriello, E. Pascucci,
M. Savino, G. Ortaggi and A. Bianco, Bioorg. Med. Chem., 2007, 15,
1848.
SF-539
UACC-257
SK-MEL-5
IGROV1
SK-OV-3
RXF 393
A498
PC-3
DU-145
HS 578T
MDA-MB-468
melanoma
ovarian cancer
ovarian cancer
renal cancer
renal cancer
prostate cancer
prostate cancer
breast cancer
breast cancer
a Only the lowest and the highest values for each tumour type are reported,
the complete data over 60 cell lines have been included in the ESI†.
9 M. Franceschin, L. Rossetti, A. D’Ambrosio, S. Schirripa, A. Bianco,
G. Ortaggi, M. Savino, C. Schultes and S. Neidle, Bioorg. Med. Chem.
Lett., 2006, 16, 1707.
G-quadruplex with respect to genomic duplex DNA, also in these
conditions (50 mM KCl solution) more similar to physiological
conditions than those used in ESI-MS experiments.
10 M. Franceschin, Eur. J. Org. Chem., 2009, 2225.
A preliminary study at the National Cancer Institute (USA) on
the effects of AZATRUX on the growth of 60 different tumour
cell lines29 suggests a good potential anticancer activity in vitro
(Table 2). Growth percents after 48 hours of incubation with 10 mM
AZATRUX solution range from -31 to -99%, with respect to no-
drug control, suggesting a potentially selective growth inhibition
for different tumour types. Following these results, NCI has
decided to perform further multi-dose studies.
11 W. Y. Lai, R. Zhu, Q. L. Fan, L. T. Hou, Y. Cao and W. Huang,
Macromolecules, 2006, 39, 3707; W. Y. Lai, Q. Y. He, R. Zhu, Q. Q.
Chen and W. Huang, Adv. Funct. Mater., 2008, 18, 265.
12 M. Talarico, R. Termine, E. M. Garcia-Frutos, A. Omenat, J. L.
Serrano, B. Gomez-Lor and A. Golemme, Chem. Mater., 2008, 20,
6589.
13 E. M. Garcia-Frutos and B. Gomez-Lor, J. Am. Chem. Soc., 2008, 130,
9173.
14 B. Gomez-Lor, G. Hennrich, B. Alonso, A. Monge, E. Gutierrez-Puebla
and A. M. Echavarren, Angew. Chem., Int. Ed., 2006, 45, 4491.
15 G. Otero, G. Biddau, C. Sanchez-Sanchez, R. Caillard, M. F. Lopez,
C. Rogero, F. J. Palomares, N. Cabello, M. A. Basanta, J. Ortega, J.
Mendez, A. M. Echavarren, R. Perez, B. Gomez-Lor and J. A. Martin-
Gago, Nature, 2008, 454, 865.
Conclusion
From the reported data, the triazatruxene derivative AZATRUX
results in a potent and selective G-quadruplex ligand, which surely
deserves further investigation on its biological activity. In fact, due
to its high water solubility and DNA binding properties, this new
compound is very promising in terms of biological and possible
pharmacological effects, not only at telomeres but also with respect
to regulation of several oncogenes.10 The results of a preliminary
study on the effects of this molecule on the growth of several
different tumour cell lines suggest an interesting and possibly
selective inhibitory action, although obviously at the moment it
is not possible to correlate this finding to quadruplex binding.
Moreover, the versatility of the synthetic pathway has been applied
in our lab to synthesize other compounds of this series, with side
chains of different length and basicity, which will soon be reported
in a complete comparative study.
16 N. Robertson, S. Parsons, E. J. MacLean, R. A. Coxall and A. R.
Mount, J. Mater. Chem., 2000, 10, 2043.
17 B. Gomez-Lor and A. M. Echavarren, Org. Lett., 2004, 6, 2993.
18 H. Hiyoshi, H. Kumagai, H. Ooi, T. Sonoda and S. Mataka, Heterocy-
cles, 2007, 72, 231; M. A. Eissenstat, M. R. Bell, T. E. D’Ambra, E. J.
Alexander, S. J. Daum, J. H. Ackerman, M. D. Gruett, V. Kumar and
K. G. Estep, J. Med. Chem., 1995, 38, 3094.
19 D. Evans and I. M. Lockhart, J. Chem. Soc., 1965, 4806.
20 K. K. Kim and J. G. Jang, U.S. Pat,. Appl. Publ. 2006, US2006063037.
21 C. Sissi, L. Lucatello, A. Paul Krapcho, D. J. Maloney, M. B. Boxer,
M. V. Camarasa, G. Pezzoni, E. Menta and M. Palumbo, Bioorg. Med.
Chem., 2007, 15, 555.
22 A. Alvino, M. Franceschin, C. Cefaro, S. Borioni, G. Ortaggi and A.
Bianco, Tetrahedron, 2007, 63, 7858.
23 F. Rosu, E. De Pauw and V. Gabelica, Biochimie, 2008, 90, 1074.
24 V. Casagrande, A. Alvino, A. Bianco, G. Ortaggi and M. Franceschin,
J. Mass Spectrom., 2009, 44, 530.
25 G. N. Parkinson, M. P. Lee and S. Neidle, Nature, 2002, 417, 876.
26 F. Rosu, V. Gabelica, C. Houssier and E. De Pauw, Nucleic Acids Res.,
2002, 30, e82.
27 F. Rosu, S. Pirotte, E. De Pauw and V. Gabelica, Int. J. Mass Spectrom.,
2006, 253, 156.
Acknowledgements
28 C. L. Mazzitelli, J. S. Brodbelt, J. T. Kern, M. Rodriguez and S. M.
Kerwin, J. Am. Soc. Mass Spectrom., 2006, 17, 593.
29 R. H. Shoemaker, Nat. Rev. Cancer, 2006, 6, 813.
Financial support for this work came from MIUR (Italian
Ministry of University) and Sapienza Universita` di Roma. Thanks
2516 | Org. Biomol. Chem., 2009, 7, 2513–2516
This journal is
The Royal Society of Chemistry 2009
©