Rational design of acridine-based ligands with selectivity for human telomeric quadruplexes

Article abstract of DOI:10.1021/ja1003944

Structure-based modeling methods have been used to design a series of disubstituted triazole-linked acridine compounds with selectivity for human telomeric quadruplex DNAs. A focused library of these compounds was prepared using click chemistry and the selectivity concept was validated against two promoter quadruplexes from the c-kit gene with known molecular structures, as well as with duplex DNA using a FRET-based melting method. Lead compounds were found to have reduced effects on the thermal stability of the c-kit quadruplexes and duplex DNA structures. These effects were further explored with a series of competition experiments, which confirmed that binding to duplex DNA is very low even at high duplex:telomeric quadruplex ratios. Selectivity to the c-kit quadruplexes is more complex, with some evidence of their stabilization at increasing excess over human telomeric quadruplex DNA. Selectivity is a result of the dimensions of the triazole-acridine compounds, and in particular the separation of the two alkyl-amino terminal groups. Both lead compounds also have selective inhibitory effects on the proliferation of cancer cell lines compared to a normal cell line, and one has been shown to inhibit the activity of the telomerase enzyme, which is selectively expressed in tumor cells, where it plays a role in maintaining telomere integrity and cellular immortalization.

Full text of DOI:10.1021/ja1003944

Published on Web 08/18/2010
Rational Design of Acridine-Based Ligands with Selectivity for
Human Telomeric Quadruplexes
Silvia Sparapani, Shozeb M. Haider, Filippo Doria, Mekala Gunaratnam, and
Stephen Neidle*
CR-UK Biomolecular Structure Group, The School of Pharmacy, UniVersity of London,
London WC1N 1AX, U.K.
Received January 15, 2010; E-mail: stephen.neidle@pharmacy.ac.uk
Abstract: Structure-based modeling methods have been used to design a series of disubstituted triazole-
linked acridine compounds with selectivity for human telomeric quadruplex DNAs. A focused library of
these compounds was prepared using click chemistry and the selectivity concept was validated against
two promoter quadruplexes from the c-kit gene with known molecular structures, as well as with duplex
DNA using a FRET-based melting method. Lead compounds were found to have reduced effects on the
thermal stability of the c-kit quadruplexes and duplex DNA structures. These effects were further explored
with a series of competition experiments, which confirmed that binding to duplex DNA is very low even at
high duplex:telomeric quadruplex ratios. Selectivity to the c-kit quadruplexes is more complex, with some
evidence of their stabilization at increasing excess over human telomeric quadruplex DNA. Selectivity is a
result of the dimensions of the triazole-acridine compounds, and in particular the separation of the two
alkyl-amino terminal groups. Both lead compounds also have selective inhibitory effects on the proliferation
of cancer cell lines compared to a normal cell line, and one has been shown to inhibit the activity of the
telomerase enzyme, which is selectively expressed in tumor cells, where it plays a role in maintaining
telomere integrity and cellular immortalization.
end-capping4 and catalytic functions5 of the telomerase enzyme
from maintaining telomere integrity in cancer cells. The
Introduction
Quadruplex nucleic acids can be formed by the folding of
DNA or RNA sequences that have several repeats of short
guanine-containing tracts, into structures comprising a stem of
stacked G-quartets held together by loop arrangements arising
from the sequences occurring between the G-tracts.1 Quadru-
plex-forming sequences occur in particular at the ends of
eukaryotic chromosomes, in promoter regions and/or in both
5′ and 3′ untranslated regions of a number of genes.2 Stabiliza-
tion of quadruplex structures within their genomic environments
by the binding of small molecules can lead to a range of
biological effects. Promoter quadruplex formation and stabiliza-
tion can result in inhibition of the transcription of a targeted
gene,3 and formation of telomeric quadruplexes can inhibit the
antitumor activity of several acridine derivatives in xenograft
models, notably the molecules BRACO-196 and RHPS4,7 has
been ascribed to their binding to telomeric quadruplexes.
However these acridines are also relatively nonselective, with
significant binding affinity to duplex DNA, as well as to a
(4) Gomez, D.; Wenner, T.; Brassart, B.; Douarre, C.; O’Donohue, M. F.;
El, K. V.; Shin-Ya, K.; Morjani, H.; Trentesaux, C.; Riou, J. F. J. Biol.
Chem. 2006, 281, 38721. Phatak, P.; Cookson, J. C.; Dai, F.; Smith,
V.; Gartenhaus, R. B.; Stevens, M. F.; Burger, A. M. Br. J. Cancer
2007, 96, 1223.
(5) (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) Read, M.; Harrison,
R. J.; Romagnoli, B.; Tanious, F. A.; Gowan, S. H.; Reszka, A. P.;
Wilson, W. D.; Kelland, L. R.; Neidle, S. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 4844.
(1) Davis, J. T. Angew. Chem., Int. Ed. 2004, 43, 668. Burge, S.; Parkinson,
G. N.; Hazel, P.; Todd, A. K.; Neidle, S. Nucleic Acids Res. 2006,
34, 5402. Phan, A. T.; Kuryavyi, V.; Luu, K. N.; Patel, D. J. Nucleic
Acids Res. 2007, 35, 6517.
(6) Burger, A. M.; Dai, F.; Schultes, C. M.; Reszka, A. P.; Moore, M. J.;
Double, J. A.; Neidle, S. Cancer Res. 2005, 65, 1489. Gunaratnam,
M.; Greciano, O.; Martins, C.; Reszka, A. P.; Schultes, C. M.; Morjani,
H.; Riou, J.-F.; Neidle, S. Biochem. Pharmacol. 2007, 74, 679.
(7) Salvati, E.; Leonetti, C.; Rizzo, A.; Scarsella, M.; Mottolese, M.; Galati,
R.; Sperduti, I.; Stevens, M. F.; D’Incalci, M.; Blasco, M.; Chiorino,
G.; Bauwens, S.; Horard, B.; Gilson, E.; Stoppacciaro, A.; Zupi, G.;
Biroccio, A. J. Clin. InVest. 2007, 117, 3236.
(2) Todd, A. K.; Johnston, M.; Neidle, S. Nucleic Acids Res. 2005, 33,
2901. Huppert, J. L.; Balasubramanian, S. Nucleic Acids Res. 2005,
33, 2908. Huppert, J. L.; Balasubramanian, S. Nucleic Acids Res. 2007,
35, 406. Balasubramanian, S.; Neidle, S. Curr. Opin. Chem. Biol. 2009,
13, 345.
(3) See for example: (a) Siddiqui-Jain, A.; Grand, C. L.; Bearss, D. J.;
Hurley, L. H. Proc. Natl. Acad. Sci USA 2002, 99, 11593. Cogoi, S.;
Xodo, L. E. Nucleic Acids Res. 2006, 34, 2536. Gunaratnam, M.;
Swank, S.; Haider, S. M.; Galesa, K.; Reszka, A. P.; Beltran, M.;
Cuenca, F.; Fletcher, J. A.; Neidle, S. J. Med. Chem. 2009, 52, 3774.
(b) Kumari, S.; Bugaut, A.; Huppert, J. L.; Balasubramanian, S. Nature
Chem. Biol 2007, 3, 218. Halder, K.; Wieland, M.; Hartig, J. S. Nucleic
Acids Res. 2009, 37, 6811. Balkwill, G. D.; Derecka, K.; Garner, T. P.;
Hodgman, C.; Flint, A. P.; Searle, M. S. Biochemistry 2009, 48, 11487.
(8) For recent surveys see: (a) Monchaud, D.; Teulade-Fichou, M. P. Org.
Biomol. Chem. 2008, 6, 627. Tan, J.-H.; Gu, L.-Q.; Wu, J.-Y. Mini-
reV. Med. Chem. 2008, 8, 1163. Franceschin, M. Eur. J. Org. Chem.
2009, 14, 2225. (b) Laronze-Cochard, M.; Kim, Y.-M.; Brassart, B.;
Riou, J.-F.; Laronze, J.-Y.; Sapi, J. Eur. J. Med. Chem. 2009, 44, 3880.
Debary, J.; Zeghida, W.; Jourdan, N.; Monchard, D.; Dheu-Andries,
M.-L.; Dumy, P.; Teulade-Fichou, M.-P.; Demeunynck, M. Org.
Biomol. Chem. 2009, 7, 5219. Fu, Y.-T.; Keppler, B. R.; Soares, J.;
Jarstfer, M. B. Bioorg. Med. Chem. 2009, 17, 2030.
9
10.1021/ja1003944  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 12263–12272 12263

A R T I C L E S
Sparapani et al.
number of genomic quadruplexes. Many other quadruplex-
binding small molecules have been reported,8a including a
number of acridine-based ligands;8b a particular challenge is
for a given ligand to have high selectivity for any chosen one
over other quadruplex structures, as well as to have low affinity
for duplex DNA, which is the overwhelming nucleic acid
background in the genome. Structural information has revealed
the presence of several distinctive topologies for telomeric qua-
druplexes, dependent on the nature of the cations present, and on
molecular crowding conditions.9,10 Although only a small number
of ligands have been evaluated to date by structural studies, these
at least generally appear to bind preferentially to parallel folded
arrangements of human telomeric G-quadruplexes.11,12
We describe here the rational design, synthesis and evaluation
of a series of novel triazole-acridine conjugates that bind with
high selectivity to human telomeric quadruplex structures, as
judged by interactions with intramolecular human parallel and
(3 + 1) hybrid antiparallel telomeric quadruplex DNAs.9a Their
design is based on a comparative study of the crystallographic
and NMR data on these human telomeric quadruplexes and the
two quadruplexes found in the promoter region of the c-kit
gene,13 and their selectivity is derived from differences in
accessible binding site dimensions between the different
quadruplexes.
postulated that the ring nitrogen atom of the acridine moiety
would be preferentially positioned over the central ion channel
in a quadruplex, and that directly attaching appropriate hetero-
cyclic rings at the 3- and 6-positions would maximize overlap
with guanine bases. Initially the ionization status of this nitrogen
atom was unknown; an initial supposition that it is protonated
at physiological pH was overturned by subsequent experiment
(see below). We also reasoned that triazole rings would be
optimal substituents at the 3- and 6-positions, because (a) they
offer minimal steric resistance to coplanarity with the acridine
group, and (b) their planarity and electron-richness could
enhance possible π-π interactions with G-quartets. A small
library of end-groups5c was designed (each attached to a phenyl
ring, to further enhance possible π-π G-quartet interactions).
Some variation in side-chain length has also been incorporated
in order to explore a range of distances between their two distant
3- and 6-ends.
Preliminary circular dichroism studies on ligand complexes
of the human telomeric quadruplex (see below) indicated that
in solution they probably exist as a mixture of topologies. Thus
modeling was performed on two distinct (parallel and antipar-
allel) topologies in order to approximate the solution situation.
The coordinate files for the human telomeric quadruplex X-ray
structure9a (PDB id 1KF1), an NMR-derived (3 + 1) hybrid
antiparallel human telomeric quadruplex structure9e (hybrid-2,
PDB id 2JPZ), together with the two c-kit quadruplex NMR
Results
structures (c-kit1, PDB id 2KQG;13b c-kit2, PDB id 203M13c
)
Molecular Modeling. Previous crystallographic analyses of
quadruplex-acridine complexes11a,14 have shown a singular
binding geometry for the acridine core, stacked onto the end of
the G-quartet stem of the quadruplex, with some π-π overlap
between two guanine bases and the acridine. This general type
of arrangement was therefore taken as a starting point for
subsequent modeling and structure prediction. It was initially
were used as starting-points for a series of systematic docking
studies followed by short molecular dynamics simulations to
find low-energy positions for each ligand bound to these
quadruplexes. The (3 + 1) hybrid-2 structure was employed
because it represents one of the several forms found by NMR
studies in solution, and by contrast with other hybrid structures
it did not require any structural modification in order to produce
a ligand binding platform. It was assumed, in accord with data
from crystallographic11,14 and NMR12,15 studies on quadruplex-
ligand complexes, that the ligands bind at one or other of the
terminal G-quartets in each structure. The docking thus explored
both 3′ and 5′ faces of each of these three quadruplexes. In
each case detailed ligand docking was focused on compounds
with side-chains each containing a methylene linker with 1-3
carbon atoms, and terminating in a pyrrolidine or a diethylamine
group (compounds 8a,b; 9a,b; 10a,b). In accord with poten-
tiometric studies, the central ring acridine nitrogen atom was
not protonated during the simulations, whereas the terminal side-
chain amine groups were considered to each carry a proton,
and thus a positive charge.
The modeling provides estimates of relative binding energies,
although these cannot reliably be compared between different
quadruplexes. Modeling with the two sets of human telomeric
quadruplex complexes (Table 1), indicated that compounds 9a,b
with two-carbon linkers overall bind consistently best to both
the parallel and the (3 + 1) structure, as judged by the calculated
relative binding energies. Examination of the models themselves
confirms this and shows (Figure 1a,b) that the side-chains are
able to make a large number of attractive close contacts with
the floor and walls of the grooves in these structures, and
significantly that there are few parts of the ligands that are not
on contact with DNA. It does appear that the complexes with
the parallel quadruplex have a slightly greater number of
(9) (a) Parkinson, G. N.; Lee, M. P. H.; Neidle, S. Nature 2002, 417,
876. (b) Phan, A. T.; Modi, Y. S.; Patel, D. J. J. Am. Chem. Soc.
2004, 126, 8710. (c) Ambrus, A.; Chen, D.; Dai, J.; Jones, R. A.;
Yang, D. Biochemistry 2005, 44, 2048. (d) Ambrus, A.; Chen, D.;
Dai, J.; Bialis, T.; Jones, R. A.; Yang, D. Nucleic Acids Res. 2006,
34, 2723. (e) Dai, J.; Carver, M.; Punchihewa, M.; Jones, R. A.; Yang,
D. Nucleic Acids Res. 2007, 35, 4927. (f) Lim, K. W.; Amrane, S.;
Bouaziz, S.; Xu, W.; Mu, Y.; Patel, D. J.; Luu, K. N.; Phan, A. T.
J. Am. Chem. Soc. 2009, 131, 4301. (g) Haider, S. M.; Parkinson,
G. N.; Neidle, S. Biophys. J. 2008, 95, 296. Haider, S. M.; Neidle, S.
Biochem. Soc. Trans. 2009, 7, 583.
(10) Xue, Y.; Kan, Z. Y.; Wang, Q.; Yao, Y.; Liu, J.; Hao, Y. H.; Tan, Z.
J. Am. Chem. Soc. 2007, 129, 11185. Renciuk, D.; Kejnovska´, I.;
Skola´kova´, P.; Bedna´rova´, K.; Motlova´, J.; Vorl´ıckova´, M. Nucleic
Acids Res. 2009, 37, 6625. Zheng, K.-W.; Chen, Z.; Hao, Y.-H.; Tan,
Z. Nucleic Acids Res. 2010, 38, 327. Martino, L.; Pagano, B.;
Fotticchia, I.; Neidle, S.; Giancola, C. J. Phys. Chem. B 2009, 113,
14779.
(11) (a) Parkinson, G. N.; Ghosh, R.; Neidle, S. Biochemistry 2007, 46,
2390. (b) Parkinson, G. N.; Cuenca, F.; Neidle, S. J. Mol. Biol. 2008,
381, 1145. (c) Campbell, N. H.; Parkinson, G. N.; Reszka, A. P.;
Neidle, S. J. Am. Chem. Soc. 2008, 130, 6722.
(12) Phan, A. T.; Kuryavyi, V.; Gaw, H. Y.; Patel, D. J. Nature Chem.
Biol. 2005, 1, 167.
(13) (a) Rankin, S.; Reszka, A. P.; Huppert, J.; Zloh, M.; Parkinson, G. N.;
Todd, A. K.; Ladame, S.; Balasubramanian, S.; Neidle, S. J. Am. Chem.
Soc. 2005, 127, 10584. Phan, A. T.; Kuryavyi, V.; Burge, S.; Neidle,
S.; Patel, D. J. J. Am. Chem. Soc. 2007, 129, 4386. (b) Fernando, H.;
Reszka, A. P.; Huppert, J.; Ladame, S.; Rankin, S.; Venkitaraman,
A. R.; Neidle, S.; Balasubramanian, S. Biochemistry 2006, 45, 7854.
(bb) Phan, A. T.; Kuryavyi, V.; Burge, S.; Neidle, S.; Patel, D. J.
J. Am. Chem. Soc. 2007, 129, 4386. (c) Hsu, S. T.; Varnai, P.; Bugaut,
A.; Reszka, A. P.; Neidle, S.; Balasubramanian, S. J. Am. Chem. Soc.
2009, 131, 13399.
(14) Haider, S. M.; Parkinson, G. N.; Neidle, S. J. Mol. Biol. 2003, 326,
117. Campbell, N. H.; Patel, M.; Tofa, A. B.; Ghosh, R.; Parkinson,
G. N.; Neidle, S. Biochemistry 2009, 48, 1675.
(15) Gavathiotis, E.; Heald, R. A.; Stevens, M. F.; Searle, M. S. J. Mol.
Biol. 2003, 334, 25.
9
12264 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010

Design of Acridines Selective for Human Telomeric Quadruplex
A R T I C L E S
Table 1. Calculated Binding Energies for Selected Compounds to
the Two Telomeric Quadruplex Structures, to the Two c-kit
Promoter Quadruplexes, and to a 45-mer
1,3-dipolar cycloaddition between bisazide 2 and alkyne building
blocks 5a-i through 7a-i. The “click” reaction17 was performed
under typical conditions, using CuSO4 ·5H2O and sodium
ascorbate in a 1:1 mixture of t-BuOH/H2O, with stirring for
72 h and at room temperature. Attempts to increase the rate of
reaction by thermolysis or microwave irradiation were unsuc-
cessful, due to the formation of unidentified degradation
products, as revealed by LC/MS analysis. The reactions
proceeded smoothly and final bis-triazole derivatives that
precipitated in water were initially isolated by filtration and
ultimately purified by semipreparative HPLC.
∆Ebinding (kcal mol-1
)
compound
parallel G4
(3 + 1) G4
c-kit1
c-kit2
45-mer G4
8a
8b
9a
9b
10a
10b
-14.0
-19.2
-19.5
-19.4
-12.3
-12.2
-19.3
-22.7
-25.2
-23.8
-17.8
-16.3
-8.0
-8.2
-8.4
-8.3
-8.1
-8.0
-7.0
-7.0
-6.8
-7.1
-7.0
-7.1
-24.3
-26.0
-29.1
-27.5
-23.4
-23.1
Ligand Protonation Status. The 9-position of many 3,6-
disubstituted acridines such as the diamino compound 1, is
normally protonated at physiological pH. In order to assess the
effect of the direct attachment of the triazole rings to the 3-
and 6-positions, a potentiometric method was used to obtain
pKa values. The titration of NaOH with (arbitrarily chosen)
compound 10a shows two points of inflection (Figure 2a), which
analysis show (Figures 2b,c) correspond to pKa values of 4.03
( 0.03 and 8.54 ( 0.10, respectively. We interpret this in terms
of two distinct protonation categories, the two equivalent
terminal amines having a pKa value of 8.54, whereas the lower
value of 4.03 corresponds to the acridine ring nitrogen atom,
which is therefore not protonated under physiologically relevant
conditions.
Solution Binding Studies. The CD spectra of the human 22-
mer telomeric quadruplex (Figure 3a,b) in the presence of
increasing concentrations of one of the ligands (9b) in K+ and
Na+ solution, show characteristics typical of quadruplex be-
havior, although the two spectra show distinct maxima. The
spectra in Na+ solution show a prominent peak at 295 nm,
consistent with an antiparallel structure, which is retained on
titrating increasing equivalents of ligand. The spectra in K+
solution show a strong shoulder at ca. 260-265 nm and a major
peak at 290 nm, consistent with a mixed population of parallel
and (probably several) antiparallel species, as has been observed
in a number of previous studies with ligands binding to human
telomeric quadruplexes.18,19 It is apparent that, at least under
the solution conditions of a typical CD experiment, that different
ligands can stabilize or induce differing topologies, with some
inducing a parallel fold, such as the oxazine compound
oxazine170,20 several amido phthalocyanines21 and diarylethynyl
amides.22
contacts, suggesting that this may be the preferred form, but
experimental verification will have to await a crystal structure.
The more compact shape of the hybrid structure means that part
of one ligand side-chain is unable to fully contact the quadruplex
surface (Figure 1b). We have also examined the binding of the
molecules to a 45-mer model of human telomeric DNA, with
two consecutive quadruplexes and a single ligand binding site
between them.9g This arrangement may be a more realistic
model for ligand binding to the human telomere than a single
quadruplex. The ranking order of ligand binding (Table 1) for
the 45-mer is the same as those for the other telomeric
quadruplexes.
The modeling also clearly indicates that compound 9a cannot
bind as effectively to the c-kit1 quadruplex (Figure 1c). There
are extensive steric clashes when it is bound at the 5′ end of
the quadruplex, which push the central acridine chromophore
away from the G-quartets while the side-chains of the ligand
cannot interact fully with the backbone atoms. At the 3′ end
the optimal position of the ligand is such that its termini are
unable to make effective contacts with the single-residue loop
of the quadruplex. An analogous situation occurs with the c-kit2
quadruplex (Figure 1c). In this instance the ends of the side-
chains are too long to be able to fit into the loop structure and
protrude outward. Ligands with one- or three-carbon linkers,
series 8a-i and 10a-i (see Scheme 1) have also been modeled,
and similarly, these are also predicted to not bind as effectively
to these two quadruplexes. No low-energy arrangement was
found for duplex DNA binding to compound 9a (or indeed any
others in the series), using either an intercalation site or groove
binding model.
Synthesis of Ligands. The 3,6-diazoacridine 2 was readily
obtained by diazotization of the readily available 3,6-aminoacri-
dine 1, as previously reported.16 The straightforward synthesis
of intermediates 5a-i and 6a-i was accomplished in one step
by reaction of the commercially available 3-ethynylaniline 3
with the appropriate acid chloride and different cyclic and
aliphatic amines (see Scheme 1). The same approach was also
possible with longer side chains (n ) 3), as in the case of
compounds 7a-i. However, from initial attempts it was
observed that yields were affected by the formation of elimina-
tion products. Their isolation from the reaction mixture, followed
by subsequent amination, was necessary to force formation of
the final intermediates 7a-i. In order to circumvent the problem,
the presence of unreacted materials was reduced by a two-step
reaction, which involved acylation of 3 with 4-chlorobutyryl
chloride in the presence of NaHCO3 at 0 °C, prior to aminolysis
of 4 with the required amine under mild thermal conditions, to
afford compounds 7a-i in good yields. The target ligands 8a-i,
9a-i, and 10a-i were synthesized by Cu(I) catalyzed Huisgen
Compounds were assessed for their ability to stabilize the
melting temperature (Tm) of all three quadruplexes and a DNA
duplex, using a fluorescence resonance energy transfer (FRET)
procedure23 (Table 2). This shows that the two-methylene linker
(17) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. Angew.
Chem., Int. Ed. 2002, 41, 2596.
(18) See for example: Shirude, P. S.; Gillies, E. R.; Ladame, S.; Godde,
F.; Shin-ya, K.; Huc.;Balasubramanian, S. J. Am. Chem. Soc. 2007,
129, 11890. Li, G.; Huang, J.; Zhang, M.; Zhou, Y.; Zhang, D.; Wu,
Z.; Wang, S.; Weng, X.; Zhou, X.; Yang, G. Chem. Commun. 2008,
4564. Tera, M.; Ishizuka, H.; Takagi, M.; Suganuma, M.; Shin-ya,
K.; Nagasawa, K. Angew. Chem., Int. Ed. 2008, 47, 1.
(19) Garner, T. P.; Williams, H. E. L.; Gluszyk, K. I.; Roe, S.; Oldham,
N. J.; Stevens, M. F. G.; Moses, J. E.; Searle, M. S. Org. Biomol.
Chem. 2009, 7, 4194.
(20) Chen, M.; Song, G.; Wang, C.; Hu, D.; Ren, J.; Qu, J. Biophys. J.
2009, 97, 2014.
(21) Alzeer, J.; Luedtke, N. W. Biochemistry 2010, 49, 4339.
(22) Dash, J.; Shirude, P. S.; Hsu, S. D.; Balasubramanian, S. J. Am. Chem.
Soc. 2008, 130, 16048.
(23) Schultes, C. M.; Guyen, B.; Cuesta, J.; Neidle, S. Bioorg. Med. Chem.
Lett. 2004, 14, 4347. Gue´din, A.; Lacroix, L.; Mergny, J.-L. Methods
Mol. Biol. 2010, 613, 25.
(16) Kanakarajan, K.; Haider, K.; Czarnik, A. W. Synthesis 1988, 7, 566.
9
J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010 12265

A R T I C L E S
Sparapani et al.
Figure 1. (a) Views of the docked low-energy positions of ligands 9a and 9b on the parallel-stranded human telomeric quadruplex. The upper views show
the position of each ligand in stick representation stacking onto the terminal G-quartet, shown as a solvent-accessible surface. The lower views show the
positively charged side-chains of each ligand accessing the propeller loops and interacting with the negatively charged sugar phosphate backbone. (b) Views
of the docked low-energy positions of ligands 9a and 9b on the antiparallel (3 + 1) hybrid human telomeric quadruplex. The lower views show side-chains
of the ligands accessing the grooves formed by the distinctive (3 + 1) antiparallel topology. (c) Views of the docked low-energy positions of ligand 9a
bound to the c-kit1 and c-kit2 quadruplexes. In the c-kit1 complex, the shape of the ligand ring system positions the side-chains in close vicinity to the
backbone, resulting in steric repulsion. In the c-kit2 complex there are two single-nucleotide loops and one five-nucleotide loop. The conformations adopted
by these loops prevent the side-chains from making favorable interactions with them.
Scheme 1. Synthesis of Triazole-Linked Acridines
series that two compounds, 9a and 9b, have ∆Tm values >12
°C, and one other, 9d, has a ∆Tm > 6 °C with the human
telomeric quadruplex. This suggests that the size of the cationic
group at the terminus of the side chains has a significant effect
on binding, and the pyrrolidine ring in 9a (and the equivalent
in 9b), is an optimal size, in accord with the findings of the
modeling that the groove pocket in both telomeric quadruplexes
modeled here is relatively small and would require distortion
to accommodate larger groups.
Only one compound in the series, the morpholino derivative
9f, has a significant effect on duplex DNA stability; by contrast
all the other compounds have ∆Tm values with duplex that are
essentially 0 °C within the limits of the sensitivity of the FRET
method. G4 selectivity vs duplex DNA has been further assessed
by a series of FRET-based competition assay for the two
approximately equipotent ligands 9a and 9b, which shows that
increasing duplex:quadruplex ratios, result in almost complete
retention of their G4 stabilizing ability (Figure 4a), with <10%
9
12266 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010

Design of Acridines Selective for Human Telomeric Quadruplex
A R T I C L E S
Figure 2. Titration curves for compound 10a with 0.1 M NaOH, (a) showing the two inflection points produced with increasing volume with respect
to pH, and (b,c) showing the straight lines produced by multiple linear regression analysis of the pH vs equilibrium constant plot to give the two pKa
values.
9
J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010 12267

A R T I C L E S
Sparapani et al.
Figure 3. Circular dichroism spectra for compound 9b titrated against the 23-mer human intramolecular telomeric quadruplex (top) in 100 mM Na+ solution
and (bottom) in 100 mM K+ solution.
change in ∆Tm values at the highest, 300-fold excess of duplex
DNA that was employed.
compound BRACO-19, which in common with 9a, has basic
pyrrolidino group at positions 3 and 6, does bind to the human
telomeric quadruplex with high affinity. However this is at the cost
of selectivity. Thus BRACO-19 stabilizes all three quadruplexes
to an approximately equal extent, and has significant interaction
with duplex DNA.
Telomerase Activity and Inhibition of Cell Growth. The
finding that compounds 9a, 9b can bind selectively to human
telomeric quadruplexes suggests that they may have biological
activity as telomerase inhibitors. Their effects on cell prolifera-
tion have been examined using a standard SRB assay. Three
human carcinoma cell lines (A549, lung; A2780, ovarian; Panc1,
pancreatic) were used, together with a normal cell fibroblast
line (WI38). Table 3 shows that both compounds have activity
in the low µM range with at least one of the carcinoma lines;
neither shows significant activity against the normal cell line.
All three carcinoma lines express the telomerase enzyme at
elevated levels, and the inhibitory ability of one compound, 9a,
was examined using a modified TRAP assay,18 which showed
a significant level of inhibition, and a telIC50 of ca. 35 µM.
Interactions with the c-kit1 and c-kit2 quadruplexes are generally
weak, and for compounds 9a and 9b, are on the borderline of
significance, as assessed by their ∆Tm values. Competition assays
with the human telomeric quadruplex vs the c-kit1 and c-kit2
quadruplexes show a more complex picture (Figure 4b-d). The
results with increasing ratios of the c-kit2 quadruplex are surprising
(Figure 4c) since they show ∆Tm values increasing beyond a ratio
of 1:10. ∆Tm values in the absence of ligand, of 5.4 °C at a 1:100
excess and 9.2 °C at a 1:300 excess of c-kit2 sequence indicate
that the c-kit2 quadruplex sequence itself is stabilizing the human
telomeric sequence, possibly by forming an intermolecular qua-
druplex. Accordingly the competition data was normalized by
subtracting the ∆Tm values solely due to stabilization of the human
telomeric quadruplex-c-kit2 interaction, to give Figure 4d, which
shows that these ligands are interacting, now with the c-kit2
quadruplex, as with c-kit1. These results suggest that, contrary to
the suggestion from the low ∆Tm values with the c-kit1 and c-kit2
quadruplex alone, these ligands do not have negligible affinity for
them.
Discussion
Alterations of the side chains to one or three CH2- groups (series
8a-i and 10a-i) as well as changes in their basicity, do not result
in any positive stabilizing effects. The trisubstituted acridine
The structure-based design process described here has resulted
in a series of compounds with high selectivity for telomeric
9
12268 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010

Design of Acridines Selective for Human Telomeric Quadruplex
A R T I C L E S
Table 2. ∆Tm Data (°C) for the Acridine Ligand BRACO-19 and
Compounds Reported Here, at a 1 µM Concentration and in 100
mM KCl Salt, with the Human Telomeric Quadruplex (htel), the
Two c-kit Promoter Quadruplexes, and a Duplex (ds) DNA
Sequence, Using the FRET Methoda
quadruplex, and to design out telomeric quadruplex affinity by
judicious changes in the side-chains.
Experimental Section
DNA Melting Experiments Using Fluorescence Resonance
Energy Transfer (FRET). Dual labeled oligonucleotides (Euro-
gentec Ltd., U.K.) with the following sequences were used (where
FAM is the donor fluorophore 6-carboxyfluorescein and TAMRA
is the acceptor 6-carboxytetramethylrhodamine):
htel G4
c-kit1 G4
c-kit2 G4
ds DNA
BRACO-19
8a
8b
8c
8d
8e
8f
9a
9b
9c
9d
9e
9f
9g
9h
9i
10a
10c
10d
10e
10f
10h
10i
25.9 ( 0.2
4.0 ( 0.5
6.2 ( 0.6
1.5 ( 0.1
3.8 ( 0.8
3.5 ( 0.3
2.5 ( 0.0
15.8 ( 0.7
15.5 ( 0.6
2.8 ( 0.2
6.7 ( 0.4
4.5 ( 0.3
2.6 ( 0.6
3.0 ( 0.5
1.7 ( 0.2
0.4 ( 0.2
3.5 ( 0.5
0.9
20.1 ( 0.3
1.8 ( 0.5
1.5 ( 0.2
1.6 ( 0.2
1.0 ( 0.8
2.0 ( 0.7
0.1 ( 0.1
0.8 ( 0.2
0.7 ( 0.5
2.1 ( 0.6
2.4 ( 0.2
1.3 ( 0.1
0.0 ( 0.6
0.5 ( 0.6
0.7 ( 0.5
0.4 ( 0.3
0.4 ( 0.5
0.6
25.3 ( 0.4
4.3 ( 0.1
3.2 ( 0.8
3.5 ( 0.4
3.0 ( 0.5
3.2 ( 0.5
3.1 ( 0.4
1.4 ( 0.5
2.5 ( 0.0
2.2 ( 0.1
4.6 ( 0.3
3.4 ( 0.1
6.5 ( 1.0
6.3 ( 0.3
2.4 ( 0.9
0.4 ( 0.7
0.2 ( 0.2
3.8
11.2 ( 0.6
0.3 ( 0.1
0.2 ( 0.8
0.1 ( 0.6
0.5 ( 0.3
0.4 ( 0.0
0.4 ( 0.5
0.4 ( 0.2
0.6 ( 0.0
0.0 ( 0.1
0.1 ( 0.2
0.2 ( 0.5
5.5 ( 0.6
0.4 ( 0.5
0.0 ( 0.3
0.1 ( 0.5
0.1 ( 0.3
0.3
G4 DNA, 5′-FAM-GGGTTAGGGTTAGGGTTAGGGTTAGGG-
TAMRA-3′
dsDNA,5′-FAM-TATAGCTATATTTTTTTATAGCTATA-TAM-
RA-3′
c-kit1, 5′-FAM-AGAGGGAGGGCGCTGGGAGGAGGGGCT-
TAMRA-3′
c-kit2, 5′-FAM-CCCGGGCGGGCGCGAGGGAGGGGAGG-
TAMRA-3′
Each oligonucleotide was initially diluted to 100 µM in nuclease-
free water (not DEPC-treated), purchased from Ambion Applied
Biosystems, UK. Stock solution of 20 µM and subsequent dilutions
were obtained in FRET buffer (60 mM KCl, K cacodylate, pH 7.4).
Oligonucleotides at concentrations of 400 µM were annealed by
heating at 85 °C for 10 min and cooled to room temperature.
Original compounds were prepared as 10 mM DMSO stock
solutions and diluted to 1 mM using 1 mM HCl in HPLC grade
water. The rest of the dilutions were performed using FRET buffer.
Next, 50 µL of annealed DNA and 50 µL of compound solution
were distributed across 96-well RT-PCR plates (Bio-Rad) (MJ
Research, Waltham, MA) and processed in a DNA Opticon Engine
(MJ Research). All experimental values were determined in
triplicate. Fluorescence readings were taken at intervals of 0.5 °C
over the range 30-100 °C, with a constant temperature being
maintained for 30 s prior to each reading. The incident radiation
was 450-495 nm and the detection was conducted at 515-545
nm. The raw data were imported into the Origin program (version
7.0, Origin Lab Corp.) and the graphs were smoothed using a 10-
point running average and subsequently normalized. All data were
analyzed with the Origin Pro 7.0 software package (Origin Lab
Corp., Northampton, MA). The change in melting temperature at
a 1 µM ligand concentration (∆Tm(1µM)) was calculated from three
experiments by subtraction of the blank from the averaged 1 µM
ligand melting temperature.
3.1
1.7
0.7
2.3
0.6
0.2
0.8
0.2
0.2
0.9
0.6
2.2
0.1
0.7
1.2
0.1
2.1
2.0
2.3
0.3
a A small number of ligands were not able to be purified to a
sufficient extent to be analyzed by FRET and are not reported here.
quadruplex DNA over duplex DNA, and of a lead compound,
9a, with significant telomerase inhibitory activity (although we
have as yet to definitively demonstrate that this inhibition is
due to induction of a quadruplex structure in the telomeric DNA
primer). It is notable that compounds 9a, 9b are active inhibitors
of cellular proliferation in at least one cancer cell line yet do
not show comparable activity (>10-fold less) against a normal
fibroblast cell line, which does not express telomerase. The
observed selectivity of these compounds for the parallel and a
(3 + 1) hybrid human telomeric quadruplex over the two parallel
topology c-kit quadruplexes is in accord with the modeling
analysis and is due to differences in loop sequence and
geometry, at least on the basis of the modeling presented here.
The competition studies do also suggest that single-point ∆Tm
values are not by themselves reliable indicators of selectivity.
For compounds 9a, 9b selectivity is limited to low telomeric
quadruplex:c-kit quadruplex ratios. In potential therapeutic terms
this may not be a problem since per cell there maybe just one
c-kit1 and one c-kit2 quadruplex on the basis that the c-kit gene
occurs uniquely in the human genome. By contrast, each human
cell contains 46 × 2 single-stranded 3′ telomeric DNA ends,
so telomeric quadruplexes would be at least in 100-fold excess.
An increasing number of other promoter quadruplexes have
been reported,2,3 although detailed structural data is as yet only
available for a very few. Most have parallel folds, with diversity
in the loops and thus asymmetry in the 3′ and 5′ ends. The
same principles of ligand selectivity would apply as determined
here. It is thus realistic to use a given quadruplex geometry as
the basis for the design of particular analogues of the triazole-
acridines or other ligands such as diarylethynyl amides22 or
nonpolycyclic triazoles25 with selectivity for a given promoter
FRET Competition assays were performed in a similar manner.
Compound dilutions were prepared in concentrations four times
the final ones and added as 25 µL aliquots. To G4 DNA (50 µL),
ligand (25 µL, 4 µM), and the appropriate competitor calf thymus
DNA (Sigma-Aldrich, UK) or unlabeled c-kit1 or c-kit2 oligo-
nucleotides (25 µL), synthesized in-house, were added at the
following concentration ratios of 1:1, 1:10, 1:100, 1:300 (G-quartets:
pair of bases when calf thymus DNA was used). The percentage
of retained F21T melting temperatures (∆Tm) upon addition of
increasing concentrations of competitor was averaged from three
experiments, normalized with standard deviation to the ∆Tm(1 µM)
for that ligand in the absence of competitor.
Circular Dichroism (CD) Studies. CD spectra of the 23mer
DNA in sodium and potassium phosphate buffers were measured
on the Applied Photophysics Ltd. (Leatherhead, UK) Chirascan/
Plus spectrometers at King’s College, London. The CD spectra were
measured between 340 and 200 nm at room temperature in a strain-
free rectangular 5 mm cell. The instrument was flushed continuously
with pure evaporated nitrogen throughout the measurements. Spectra
were recorded with a 1 nm step size, a 1.5 s measurement time-
per-point and a spectral bandwidth of 1 nm. CD experiments were
carried out on annealed 23-mer human telomeric DNA sequence
at a concentration of 4.4 µM DNA. Annealing was performed by
heating to 85 °C for 10 min in the appropriate 100 mM sodium or
potassium phosphate buffer solution, pH 7.4, followed by cooling
overnight. Concentration of the ligand 9b was determined using
the recorded value of the absorbance at 260 nm by using the
(24) Reed, J. E.; Gunaratnam, M.; Beltran, M.; Reszka, A. P.; Vilar, R.;
Neidle, S. Anal. Biochem. 2008, 380, 99.
(25) Moorhouse, A. D.; Santos, A. M.; Gunaratnam, M.; Moore, M.; Neidle,
S.; Moses, J. E. J. Am. Chem. Soc. 2006, 128, 15972.
9
J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010 12269

A R T I C L E S
Sparapani et al.
Figure 4. Competition experiments showing the effects of compounds 9a, 9b on the thermal stabilization of the human telomeric quadruplex, as measured
by the FRET method, (a) using increasing ratios of duplex DNA (as base pairs), (b) with increasing ratios of c-kit1 oligonucleotide (as base pairs), (c) with
increasing ratios of c-kit2 oligonucleotide (as base pairs), (d) as in (c) but now normalized by subtracting the ∆Tm values solely due to stabilization of the
human telomeric quadruplex-c-kit2 interaction.
Table 3. Quantitation of the Inhibition of Cell Proliferation by Two
(3 + 1) human telomeric G-quadruplex and NMR structures of the
Selected Compounds, in Three Human Cancer Cell Lines (A549,
A2780 and Panc1) and a Normal Fibroblast Line (WI38), shown as
human c-kit gene (c-kit1 (PDB 2O3M) and c-kit2 (PDB 2KQG))
IC50 values (µM)
intramolecular G-quadruplexes from the promoter sequence of the
were obtained from the Protein Data Bank and used as a starting
compound
A549
A2780
Panc1
WI38
point to examine plausible interactions with the ligands. The
molecular models for the triazole-acridine series of ligands were
constructed, minimized and partial charges calculated semiempiri-
cally using the MOPAC program26 from the Insight suite software
(www.accelrys.com). Prior to docking on the human telomeric
quadruplex crystal structure, the extreme 5′ terminal residue was
removed in order to expose the flat planar quartet to the ligand.
The ligands were minimized and docked on both 5′ and 3′ surfaces
of the four G-quadruplex structures using the AFFINITY docking
program (www.accelrys.com), employing the grid docking method
available with AFFINITY. This approach has been validated in
previous studies on the rational design of quadruplex ligands.27
The automated docking process identifies and ranks positions based
on high interaction energies within the binding site. The complexes
were then subjected to 1000 steps of molecular mechanics
minimization and 200 ps of molecular dynamics simulations in
explicit solvent at 300 K using the DISCOVER3 program in the
Insight II suite software. The data was visualized by means of the
BRACO-19
9a
9b
2.4 ( 1.0
31.4 ( 0.5
>50 ( 2
2.5 ( 0.9
6.6 ( 0.6
4.4 ( 0.8
ND
>50 ( 2
5.3 ( 1.0
10.7 ( 0.6
>50 ( 2
>50 ( 2
Beer-Lambert law. All raw data were and analyzed with OriginPro
7.0 software package (Origin Lab Corp., Northampton, MA) and
spectra were buffer baseline corrected.
Potentiometric Titrations. A 10 mM stock solution in DMSO
of compound 10a was diluted to 5 mM with deionized water and
titrated with a 0.1 M NaOH solution, previously standardized with
potassium hydrogen phthalate (Vep ) 0.66 mL), starting with 5 µL
additions until the first end-point region and then with increments
of 15-50 µL near the second end-point region. Titrations were
carried out using a glass-body pH semi-micro electrode (Hanna
Instruments Ltd.). Data points were interpolated and curves were
smoothed using the Boltzmann equation to determine end-point
values. Dissociation constants were obtained by multiple linear
regression analysis. Data were analyzed with the OriginPro 7.0
software package (Origin Lab Corp., Northampton, MA).
Molecular Modeling. The crystal structures of the 22-mer
parallel-stranded propeller-loop topology (PDB 1KF1; resolution
2.10 Å) human telomeric G-quadruplex, the NMR structure of the
(26) Stewart, J. J. P. J. Comput. Aided Mol. Des. 1990, 4, 100.
(27) 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.
9
12270 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010

Design of Acridines Selective for Human Telomeric Quadruplex
A R T I C L E S
VMD program.28 Electrostatic contributions to the overall binding
energy were calculated using the APBS software, employing an
implicit solvent model.29 The docked complexes were prepared
using the PDB2PQR server (http://agave.wustl.edu/pdb2pqr). Charges,
atom types and radii were assigned to each ligand atom based on
the AMBER force field. The complexes were then subjected to
APBS electrostatic calculations. A dielectric constant of 2, a solvent
dielectric constant of 80 were used, together with grid spacing that
were all optimally chosen such that the grid was always finer than
0.5 Å. The remaining parameters were kept at default values. The
binding free energies were calculated on the basis of the simple
thermodynamics cycle.
in order to remove any traces of ethanol from the purification step.
The purified samples were freeze-dried and then redissolved in
PCR-grade water at room temperature prior to the second ampli-
fication step. Step 3: The purified extended samples were then
subject to PCR amplification. For this, a second PCR master mix
was prepared consisting of ACX reverse primer (1 µM; 5′-GCG
CGG [CTTACC]3 CTA ACC-3′), TS forward primer (0.1 µg; 5′-
AAT CCG TCG AGC AGA GTT-3′), TRAP buffer, BSA (5 µg),
0.5 mM dNTP and 2U of TAQ polymerase (RedHot, ABgene,
Surrey, UK). An aliquot of 10 µL of the master mix was added to
the purified telomerase extended samples and amplified for 35
cycles of 94 °C for 30 s, at 61 °C for 1 min and 72 °C for 1 min.
Samples were separated on a 12% PAGE and visualized with SYBR
green (Aldrich) staining. Gels were quantified using a gel scanner
and gene tool software (Sygene, Cambridge, UK). Intensity data
were obtained by scanning and integrating the total intensity of
each PCR product ladder in the denaturing gels. Drug samples were
normalized against positive control containing protein only. All
samples were corrected for background by subtracting the fluores-
cence reading of negative controls. Data for all ligands was collected
at a range of concentrations in order to obtain dose-response
curves, from which EC50 values (the concentration required for 50%
enzyme inhibition, corresponding to a 50% decrease in total
integrated ladder intensity), were obtained by inspection. Graphs
were fitted to dose-response curves using the Origin 6.0 software
package.
where [complex, quadruplex, ligand] and [complex(s), quadruplex(s),
ligand(s)] represent the nonsolvated and solvated states of the system
respectively. The program APBS calculates binding free energies
using the equation:
∆Gbinding ) -∆3G ) ∆4G - d∆1G - ∆2G
Cellular Proliferation Assay. Short-term growth inhibition was
measured using the Sulphorhodamine B (SRB) assay as described
previously.30 Briefly, cells were seeded (4000 cells/wells) into the
wells of 96-well plates in DMEM and incubated overnight as before
to allow the cells to attach. Subsequently cells were exposed to
freshly made solutions of ligand at increasing concentrations and
incubated for a further 96 h. Following this the cells were fixed
with ice-cold trichloacetic acid (TCA) (10%, w/v) for 30 min and
stained with 0.4% SRB dissolved in 1% acetic acid for 15 min.
All incubations were carried out at room temperature. The IC50
value, the concentration required to inhibit cell growth by 50%,
was determined from the mean absorbance at 540 nm for each
ligand concentration expressed as a percentage of the control
untreated well absorbance.
Synthetic Chemistry. All chemicals were purchased from
Sigma-Aldrich and Fisher Scientific and were reagent grade. They
were used as supplied without further purification. Solvents were
supplied by BDH, Sigma-Aldrich and Fisher Scientific (HPLC
grade). Flash chromatography employed Merck silica gel [Kieselgel
60 (0.040-0.063 mm)]. Analytical TLC was performed using 0.2
mm silica-coated aluminum sheets [60 F254] from Merk. 1H NMR
and 13C NMR spectra were recorded at 295K on a Bruker Avance
400 spectrometer at 400 and 100 MHz respectively using the
specified deuterated solvent purchased from GOSS Scientific or
Sigma-Aldrich. NMR spectra were analyzed in MestReC 4.5.6.0
and reported as observed with coupling constants (J) in hertz (Hz).
Melting points (m.p.) were conducted using a Bibby Stuart Scientific
SMP3 melting point apparatus, High resolution accurate mass
spectra (HRMS) were conducted on a Micromass Q-TOF Ultima
Global tandem mass spectrometer using electrospray ionization
mode and 50% acetonitrile in water and 0.1% formic acid as solvent,
and processed using the MassLab 3.2 software. All ligands were
purified by semipreparative reversed-phase HPLC (Semi-Prep
HPLC) on a Gilson chromatograph with a Gilson 215 liquid handler
and a Gilson 845Z injection module coupled to a Gilson UV/vis
155 detector and YMC C18 5 µm (100 × 20 mm) column or YMC
C18 5 µm (1000 × 30 mm) using 0.1% formic acid in acetonitrile
and 0.1% formic acid in water (flow rate: 10 mL min-1). Analytical
HPLC analyses were carried out on a Gilson chromatograph with
YMC C18 5 µm (100 × 4.6 mm) column and an Agilent 1100
series photodiode array detector (flow rate: 1 mL min-1) to assess
This binding free energy cycle illustrates binding in terms of the
transfer of free energies from a homogeneous dielectric environment
to an inhomogenous dielectric environment with differing internal
and external dielectric constants.29
Telomerase Inhibition Assay. The telomerase repeat amplifica-
tion protocol (TRAP) assay was modified from a two-step to the
three-step TRAP-LIG procedure:24 step 1, initial primer elongation
by telomerase and addition of ligand; step 2, subsequent removal
of the ligand; step 3, PCR amplification of the products of
telomerase elongation. Step 1: This was carried out by preparing a
master mix containing the TS forward primer (0.1 µg; 5′-AAT CCG
TCG AGC AGA GTT-3′), TRAP buffer (20 mM Tris-HCl [pH
8.3], 68 mM KCl, 1.5 mM MgCl2, 1 mM EGTA, 0.05% v/v Tween-
20), bovine serum albumin (0.05 µg), and dNTPs (125 µM each),
protein extract (500 ng/sample) diluted in lysis buffer (10 mM Tris-
HCl, pH 7.5, 1 mM MgCl2, 1 mM EGTA, 0.5% CHAPS, 10%
glycerol, 5 mM ꢀ-mercaptoethanol, 0.1 mM AEBSF). The PCR
master mix was added to tubes containing freshly prepared ligand
at various concentrations and to a negative control containing no
ligand. The initial elongation step was carried out for 10 min at 30
°C, followed by 94 °C for 5 min and a final maintenance of the
mixture at 20 °C. Step 2: To purify the elongated product and to
remove the bound ligands the QIA quick nucleotide purification
kit (Qiagen) was used according to the manufacturer’s instructions.
This kit is especially designed for the purification of both double
and single-stranded oligonucleotides, from 17 bases in length. It
employs a high-salt buffer to bind the negatively charged oligo-
nucleotides to the positively charged spin-tube membrane through
centrifugation, so that all other components, including positively
charged and neutral ligand molecules would be eluted. PCR-grade
water was then used (rather than the manufacturers recommendation
of an ethanol-based buffer) to wash any impurities away before
elution of the DNA using a low-salt concentration solution. The
TRAP assay is sensitive to even trace amounts of ethanol, which
was present as residual in the PCR reaction mixture from the spin
tube purification step. Even when 1% ethanol was present it was
found to severely inhibit the subsequent PCR step. To overcome
this problem it was first necessary to freeze-dry the eluted samples
(28) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graph 1996, 14, 33.
(29) Baker, N. A.; Sept, D.; Joseph, S.; Holst, M.; McCammon, A. J. Proc.
Natl. Acad. Sci. U.S.A. 2001, 98, 10037.
(30) Gowan, S. M.; Harrison, R. J.; Patterson, L.; Valenti, M.; Read, M. A.;
Neidle, S.; Kelland, L. R. Mol. Pharmacol. 2002, 61, 1154.
9
J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010 12271

A R T I C L E S
Sparapani et al.
sample purity. Spectra were processed using the Unipoint 5.11
software package, with compound purity and retention time (tR)
assessed at 254 nm. Compounds were concentrated using freeze
drier Savant A160 SpeedVac Concentrator and Vacuubrand CVC
3000. Infrared measurements were carried out on a Perkin-Elmer
Spectrum 100 FT-IR spectrometer. Compound names were obtained
from CS ChemDraw Ultra 12.0.
L-ascorbate and copper(II) sulfate pentahydrate in H2O were added
in one portion. The heterogeneous mixture was stirred vigorously
at room temperature for 3 days and it was then concentrated under
reduced pressure. The resulting residue was washed with 1:1
mixture of water/methanol and collected by filtration. The crude
product was subjected to preparative HPLC purification, using the
following eluents: A ) H2 O and 0.1% FA, B ) ACN and 0.1%
FA, 0 min (95% A, 5% B), 2 min (95% A, 5% B), 21 min (50%
A, 50% B), 23.50 min (95% A, 5% B). Further details and analytical
data are provided as Supporting Information; data on the two most
active compounds are detailed below.
3,6-Diazidoacridine (2). 3,6-diaminoacridine hydrochloride (0.50
g, 2.40 mmol) was dissolved in TFA (15 mL) and stirred at 0-5
°C for 10 min. NaNO2 (0.60 g, 8.70 mmol) was then added in small
portions and over a period of 5 min, followed by NaN3 (1.60 g,
25.0 mmol), under vigorous stirring and cooling with ice (special
care must be taken during this time to avoid any possible foaming
formation). The reaction mixture was stirred for 75 min, followed
by the addition of two portions of ice water (2 × 20 mL) to induce
the precipitation of the product, which was then collected by
filtration (0.06 g, 85%). mp 155-159 °C black dec [Lit.31 168-169
N,N′-((1,1′-(Acridine-3,6-diyl)bis(1H-1,2,3-triazole-4,1-diyl))bis(3,1-
phenylene))bis(3-(pyrrolidin-1-yl)propanamide) (9a). To a suspen-
sion of N-(3-ethynylphenyl)-3-(pyrrolidin-1-yl)propanamide (0.14
g, 0.58 mmol) and 3,6-diazidoacridine (0.07 g, 0.27 mmol) in 2
mL of tBuOH, a solution of (+)-sodium L-ascorbate (0.026 g, 0.135
mmol) and copper(II) sulfate pentahydrate (0.007 g, 0.027 mmol)
in 2 mL of H2O was added in one portion. The heterogeneous
mixture was treated according to the general procedure to afford a
dark brown solid (0.142 g, 82%). Purity by HPLC analysis, 99%
(tR ) 18.13 min). mp >275 °C. 1H NMR (TFA+D2O, 400 MHz):
δ 13.43 (1H, s), 12.65-12.56 (4H, m), 12.23 (2NH, s), 12.01-12.00
(2H, m), 11.63-11.58 (4H, m), 11.16-11.15 (2H, m), 11.02-10.86
(4H, m), 7.29-7.28 (2H, br s), 7.01-7.00 (2H, br s), 6.48-6.44
(10H, m), 5.61 (6H, br s), 5.20-5.18 (4H, m). HRMS theoretical
C43H43N11O2, mass 746.367 [M+1]+, found 746.370.
1
°C dec.]. H NMR (MeOD, 400 MHz): δ 9.45 (2H, s), 8.31 (2H,
d, J ) 10 Hz), 7.65 (2H, d, J ) 2 Hz), 7.48 (2H, dd, J ) 10.2 Hz).
IR (film) 2114.36, 1678.21, 1165.72.
4-Chloro-N-(3-ethynylphenyl)butanamide (4). To an ice cooled
mixture of 3-ethynylbenzenamine (0.72 g, 0.9 mL, 6.14 mmol) and
sodium bicarbonate (0.95 g, 11.31 mmol) in DMA (6 mL),
4-chlorobutanoyl chloride (1.59 g, 1.25 mL, 11.30 mmol) was added
dropwise and it was vigorously stirred for 75 min. 0.5 M NaHCO3
aq (8 mL) was slowly added dropwise, keeping the yellow mixture
at 0 °C and it was left to stir for 30 min. The white solid was
collected by filtration, washed with H2O and dried overnight, to
yield 0.89 g of product (65%). The product was crystallized from
N,N′-((1,1′-(Acridine-3,6-diyl)bis(1H-1,2,3-triazole-4,1-diyl))bis(3,1-
phenylene))bis(3-(diethylamino)propanamide) (9b). To a suspen-
sion of 3-(diethylamino)-N-(3-ethynylphenyl)propanamide (0.142
g, 0.58 mmol) and 3,6-diazidoacridine (0.07 g, 0.27 mmol) in 2
mL of tBuOH, a solution of (+)-sodium L-ascorbate (0.026 g, 0.135
mmol) and copper(II) sulfate pentahydrate (0.007 g, 0.027 mmol)
in 2 mL of H2O was added in one portion. The heterogeneous
mixture was treated according to the general procedure to afford a
dark brown solid (0.16 g, 78%). Purity by HPLC analysis, 98% (tR
1
hexane/ethyl acetate, 9:1. mp 83-85 °C. H NMR (CDCl3, 400
MHz): δ 7.60 (s, 1H), 7.50 (d, J ) 7.2 Hz, 1H), 7.23 (m, 3H),
3.67 (m, 2H), 3.03 (s, 1H), 2.53 (m, 2H), 2.17 (m, 2H). 13C NMR
(CDCl3, 400 MHz): δ 169.99, 137.67, 129.04, 128.11, 123.24,
122.89, 120.37, 83.04, 77.53, 44.35, 34.10, 27.82. HRMS theoretical
C12H12ClNO mass 222.068 [M+1]+, found 222.067.
General Procedure of Synthesis for Compounds 5a-i,
6a-i, and 7a-i. To an ice cooled solution of 3-ethynylbenzenamine
in THF and TEA, 2-chloroacetyl chloride (for 5a-i), 2-chloropro-
pionyl chloride (for 6a-i), or 4-chloro-N-(3-ethynylphenyl)butana-
mide (for 7a-i) was added dropwise and the mixture was allowed
to warm up to r.t., stirring vigorously for 2 h. The appropriate amine
was then added to the ice-cooled mixture and it was stirred
overnight, for approximately 17 h at r.t. The THF was removed on
a rotary evaporator and the residue was basified (pH ) 8.0) with
a sat. NaHCO3(aq.) (5 M NH4OH for 7a-i) solution and extracted
with portions of dichloromethane. The combined fractions were
dried over MgSO4 and evaporated in Vacuo. The crude product
was purified by flash column chromatography over silica gel, eluting
with dichloromethane, then dichloromethane:methanol (99:1, 97:
3, 95:5 for 5a-i), (99:1, 97:3, 95:5, 93:7 for 6a-i), or (98:2, 95:5,
90:10 for 7a-i) The pure compound was further purified by HPLC
using the following eluents: A ) H2O and 0.1% FA, B ) ACN
and 0.1% FA, 0 min (75% A, 25% B), 3 min (75% A, 25% B),
8-15 min (75% A, 25% B). Further details and analytical data are
provided as Supporting Information.
1
) 16.88 min). mp >275 °C. H NMR (TFA+D2O, 400 MHz): δ
12.34 (1H, s), 11.64 (2H, s), 11.12-11.10 (2H, m), 10.91 (2H, br
s), 10.47 (2H, s), 10.37 (2H, s), 10.26 (2NH, s), 10.04 (2H, br s),
9.88 (2H, br s), 9.77 (2H, br s), 5.82 (4H, s), 5.63 (8H, br s), 5.37
(4H, s), 3.68-3.66 (12H, m). 13C NMR (TFA+D2O, 400 MHz): δ
167.11, 162.54, 146.84, 138.05, 136.45, 132.41, 129.67, 126.61,
122.94, 120.73, 120.54, 119.71, 118.51, 117.38, 115.71, 44.32,
43.53, 24.28, 2.08. HRMS theoretical C43H47N11O2, mass 750.399
[M+1]+, found 750.398.
Acknowledgment. S.S. is a Maplethorpe Fellow of The
University of London. This work was supported by Programme
Grant No. C129/A4489 to S.N. from Cancer Research UK, and by
the FP6 framework grant “Molecular Cancer Medicine” from the
EU.
Supporting Information Available: Analytical and synthetic
data on intermediate compounds 5a-i, 6a-i, 7a-i, 8a-i, 9a-i,
and 10a-i. This material is available free of charge via the
Internet at http://pubs.acs.org.
General Synthetic Procedures for Compounds 8a-i, 9a-i,
and 10a-i. To a suspension of the appropriate building block
(5a-i, 6a-i, 7a-i) and 3,6-diazidoacridine in tBuOH, (+)-sodium
(31) Firth, W. J. Org. Chem. 1982, 47, 3002.
JA1003944
9
12272 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010