Amide bond direction modulates G-quadruplex recognition and telomerase inhibition by 2,6 and 2,7 bis-substituted anthracenedione derivatives

Article abstract of DOI:10.1016/j.bmc.2007.09.040

G-quadruplex structures of DNA represent a potentially useful target for anticancer drugs. Stabilisation of this arrangement at the ends of chromosomes may inhibit the action of telomerase, an enzyme involved in immortalization of cancer cells. Appropriately substituted amido anthracenediones are effective G-quadruplex stabilizers, but no information is available as yet on the possible modulation of G-quadruplex recognition and telomerase inhibition produced by the direction of the amide bond. To understand the basis of amido anthracenedione selectivity, we have synthesized a number of derivatives bearing the -CO-NH- or -NH-CO- group linked to the planar anthraquinone (AQ) moiety at 2,6 and 2,7 positions. The various isomers were tested in terms of telomerase inhibition, determined by the TRAP assay, G-quadruplex stabilisation measured by the increase in melting temperature of the appropriately folded oligonucleotide using FRET, and conformational and G4 binding properties examined by molecular modelling techniques. In all cases, enzymatic inhibition and G-quadruplex stabilization were directly related, which strongly supports the proposed molecular mechanism of telomerase interference. Interestingly, the AQ-NH-CO- arrangement performs invariantly better than the AQ-CO-NH- arrangement, showing a clear preference among isomeric derivatives. Theoretical calculations suggest that the former amide arrangement is co-planar with the aromatic system, whereas the latter is tilted by about 30° when considering the most stable conformation. A more extended planar surface would allow more efficient stacking interactions with the quadruplex structure, hence more effective telomerase inhibition.

Full text of DOI:10.1016/j.bmc.2007.09.040

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 354–361
Amide bond direction modulates G-quadruplex recognition
and telomerase inhibition by 2,6 and 2,7 bis-substituted
anthracenedione derivatives
Giuseppe Zagotto,a Claudia Sissi,a Stefano Moro,a Diego Dal Ben,d Gary N. Parkinson,c
Keith R. Fox,b Stephen Neidlec and Manlio Palumboa,*
aDepartment of Pharmaceutical Sciences, University of Padova, Via Marzolo, 5, 35131 Padova, Italy
bSchool of Biological Sciences, University of Southampton, Bassett Crescent East, Southampton SO16 7PX, UK
cCRUK Biomolecular Structure Group, The School of Pharmacy, University of London, London WC1N 1AX, UK
dDepartment of Chemical Sciences, University of Camerino, Via S.Agostino, 1, 62032 Camerino(MC), Italy
Received 2 July 2007; revised 11 September 2007; accepted 19 September 2007
Available online 25 September 2007
Abstract—G-quadruplex structures of DNA represent a potentially useful target for anticancer drugs. Stabilisation of this arrange-
ment at the ends of chromosomes may inhibit the action of telomerase, an enzyme involved in immortalization of cancer cells.
Appropriately substituted amido anthracenediones are effective G-quadruplex stabilizers, but no information is available as yet
on the possible modulation of G-quadruplex recognition and telomerase inhibition produced by the direction of the amide bond.
To understand the basis of amido anthracenedione selectivity, we have synthesized a number of derivatives bearing the –CO–
NH– or –NH–CO– group linked to the planar anthraquinone (AQ) moiety at 2,6 and 2,7 positions. The various isomers were tested
in terms of telomerase inhibition, determined by the TRAP assay, G-quadruplex stabilisation measured by the increase in melting
temperature of the appropriately folded oligonucleotide using FRET, and conformational and G4 binding properties examined by
molecular modelling techniques. In all cases, enzymatic inhibition and G-quadruplex stabilization were directly related, which
strongly supports the proposed molecular mechanism of telomerase interference. Interestingly, the AQ–NH–CO– arrangement per-
forms invariantly better than the AQ–CO–NH– arrangement, showing a clear preference among isomeric derivatives. Theoretical
calculations suggest that the former amide arrangement is co-planar with the aromatic system, whereas the latter is tilted by about
30ꢀ when considering the most stable conformation. A more extended planar surface would allow more efficient stacking interac-
tions with the quadruplex structure, hence more effective telomerase inhibition.
ꢁ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
cells, however, have the facility to maintain telomere
length constant as they express a nucleoprotein, telome-
rase, consisting of an RNA subunit that provides the nu-
cleic acid template and a protein catalytic subunit, which
is able to add telomeric sequences at the 30 terminal of
chromosomes.3
Telomeres are DNA sequences located at the terminal
ends of all chromosomes.1 They consist of tandem re-
peats of guanine rich sequences (50-TTAGGG in hu-
mans) which cap the ends of chromosomes, protecting
them from deleterious processes such as fusion, recom-
bination or loss of genetic information during replica-
tion steps. Indeed, due to the so called ‘end-replication
problem’ telomeres shorten after each doubling in nor-
mal cells. When the number of tandem repeats falls be-
low a critical value (Hayflick number), the cell enters
senescence and eventually, if it can undergo additional
cell division cycles, into apoptosis.2 Undifferentiated
The interest in telomerase as an anticancer drug target
stems from the observation that telomerase is activated
in ca 85% of human cancers, but it is silent, or expressed
at very low levels in normal somatic cells.4 Telomerase
prevents cells from entering senescence and ensures that
they remain immortal. Hence, inhibiting this enzyme
would allow telomere shortening and finally cell death
to occur. At present, among different strategies actively
pursued to inhibit telomerase activity in cancer cells, the
development of G-quadruplex stabilizers has emerged as
a highly promising approach.5 This is based on the fact
Keywords: Anthracenedione; G-quadruplex; Telomerase; Cancer.
*
Corresponding author. Tel.: +39 049 8275699; fax +39 049
8275366; e-mail: manlio.palumbo@unipd.it
0968-0896/$ - see front matter ꢁ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.09.040

G. Zagotto et al. / Bioorg. Med. Chem. 16 (2008) 354–361
355
that single-stranded G-rich telomeric repeats, in the
presence of monovalent cations, are able to assemble
into quadruplex DNA structures stabilized by guanine
tetrads. Since telomerase reverse transcriptase activity
depends on the 30 single-stranded telomeric DNA end
acting as a primer, small molecules that bind and stabi-
lize the folded quadruplex form of the primer can inhibit
enzyme action. Efficient inhibitors are generally charac-
terized by the presence of an extended planar system
which stacks on the guanine tetrad, together with two
or three positively-charged side-chains located in the
DNA grooves, which additionally stabilize the resulting
complex through charge interactions. To date, a number
of families of compounds have been identified and their
interactions with G-quadruplexes have been extensively
studied.6 The cellular effects produced by several such
compounds have now been studied in detail, and indi-
cate that classic telomerase inhibition does occur, with
subsequent telomere length attrition. However more ra-
pid senescence and apoptosis effects occur, which have
been ascribed to quadruplex formation promoting disso-
ciation of single-strand binding proteins such as hPOT1
and uncapping of telomerase from the 30 end of the telo-
mere.7 Exposure of the 30 end as a quadruplex-ligand
complex then acts as a DNA damage response signal.
Among the quadruplex ligands studied, disubstituted
amido-anthraquinones represent an example of how
the pattern of disubstitution modulates DNA duplex
vs. quadruplex selectivity. In fact, the presence of 1,4
substituents shifts the binding preference to the duplex
structure, whereas 2,6 or 2,7 congeners prefer G-quad-
ruplex arrangements.8
NH– group was labelled as the amide-form and the
AQ–NH–CO– as the ‘inverse’ amide-form.
2. Synthetic procedures
Reaction of 1,4-benzoquinone with 2 equivalents of iso-
prene at 130 ꢀC under pressure gave the mixed 2,6 and
2,7- Diels-Alder adducts in 48% yield. The 2,6 and 2,7
isomers could be separated after oxidation of the above
adducts with oxygen to the anthraquinones. In fact, the
2,6-isomer is completely insoluble in ethanol and can be
filtered off.
Both 2,6 and 2,7-dimethylanthraquinones can be oxi-
dized to the di-acid in 77% yield by refluxing in glacial
acetic acid with chromium trioxide.
The acid chlorides of the 2,6 and 2,7-diacid anthraqui-
nones were obtained using thionyl chloride in THF and
these were converted, in the same pot, to the 2,6 and
2,7-Boc-ethylenediamido-
anthraquinones
using
mono-protected Boc-ethylenediamine in THF. The Boc
protecting group was finally removed using 90%
Trifluoroacetic acid (TFA) in water. The synthesis of
anthracenedione-peptide 2,6-AQ–CO–NH conjugates
(ethylenediamide anthraquinones or ED-AQ) is shown
in Scheme1. The same procedure was applied to obtain
the 2,7-AQ–CO–NH derivatives (numbered with a
prime: 10, 20, 30, etc.).
The 2,6-diaminoanthraquinone is commercially avail-
able, however the 2,7-diaminoanthraquinone is not.
According to previously reported protocols, the 2,7-
diaminoanthraquinone was obtained by Zinin reduction
of 2,7-dinitroanthraquinone using sodium sulfide nono-
hydrate, prepared, in turn, by the oxidative nitration of
anthrone8d (see Scheme 1).
One of the issues remaining still unexplored is an assess-
ment of the role, if any, played by the amide bond direc-
tion in amido-anthraquinones, that would favour G-
quadruplex recognition. In this connection, we prepared
and examined anthraquinone derivatives (AQ) charac-
terised by two identical side chains at positions 2 and
6 or 2 and 7, linked to the aromatic ring by an amide
bond as reported in Figure 1. The choice of the side
chain groups rests on the effective anti-telomerase activ-
ity they exhibited in previous studies.8 The AQ–CO–
The acid chloride of Fmoc-bAla-OH was generated by
reaction of the acid with thionyl chloride. The Fmoc-
bAla-Cl was then coupled with the 2,6- or 2,7-diam-
inoanthraquinone in refluxing THF for 7 h to give the
O
R3
R1
R2
O
Compound
R1
R2
R3
+
+
+
4
4’
5
-CO-NH-(CH2)2-NH3
-CO-NH-(CH2)2-NH3
-CO-NH-(CH2)2-NH3
H
H
+
-CO-NH-(CH2)2-NH3
-CO-NH-(CH2)2-pyrrolidine -CO-NH-(CH2)2-pyrrolidine
-NH-CO -(CH2)2-pyrrolidine -NH-CO-(CH2)2-pyrrolidine
H
5’
10
10’
H
+
+
-NH-CO-(CH2)2-NH3
-NH-CO-(CH2)2-NH3
-NH-CO-(CH2)2-NH3
H
H
+
+
-NH-CO-(CH2)2-NH3
Figure 1. Chemical structures of tested compounds.

356
G. Zagotto et al. / Bioorg. Med. Chem. 16 (2008) 354–361
O
O
O
O
O
OH
Cl
a
HO
Cl
O
O
O
1 [2,6 isomer], 1'
2 [2,6 isomer], 2'
b
O
O
NHBoc
N
H
H
N
BocHN
3 [2,6 isomer], 3'
O
O
O
c
O
NH+3
N
H
-
CF3CO2
H
N
H3N+
4 [2,6 isomer], 4'
-
CF3CO2
O
O
Scheme 1. Reagents and conditions. (a) SOCl2, THF, reflux, 6 h; (b) Boc-ethylenediamine, NEt3, THF, reflux, 3 h; (c) TFA, water, rt, 1 h.
2,6-Fmoc-ethylenediamine reverse amide anthraquinon-
es as shown in Scheme 2. The 2,6-bis-[N-(3-Fmoc-ami-
the 2,6-bis-(3-amino)-propionamide anthraquinones.
Also, active ester couplings between the 2,6-diam-
inoanthraquinone and Fmoc-bAla-OH using EDCI/
HOBt and HBTU/HOBt did not work either and only
starting material could be recovered. This is due to the
no)-propionamide]
anthraquinones
were
then
deprotected in DMF and the free amines were treated
with 90% trifluoroacetic acid (TFA) in water to give
O
O
O
O
NO2
NH2
O2N
H2N
a
b
O
6
7
8 [2,6-isomer], 8'
O
H
N
H
N
c
O
NHFmoc
O
O
FmocNH
FmocNH
O
O
9 [2,6-isomer], 9'
Cl
d
H
N
NH3+
H
H3N+
N
CF3COO-
CF3COO-
O
O
10 [2,6-isomer], 10'
O
Scheme 2. Reagents and conditions. (a) HNO3, 0–5ꢀC, 2 h—AcOH, reflux, 2 h; (b) Na2S.9H2O, NaOH, EtOH, reflux, 6 h; (c) Fmoc-b-Ala-Cl, THF,
7 h; (d) piperidine, DMF, rt 2 h—TFA/water.

G. Zagotto et al. / Bioorg. Med. Chem. 16 (2008) 354–361
357
low solubility of the 2,6-diaminoanthraquinone. The
synthesis of anthracenedione-peptide 2,6-AQ–NH–CO
conjugates (2,6-Fmoc-b-Ala-AQ) and of the 2,7-AQ–
NH–CO derivatives (numbered with a prime: 10, 20, 30,
etc.) is shown in Scheme 2.
rescein) at the 50-end and a quencher (MethylRed) at the
30-end.11 When this sequence folds into an intramolecu-
lar G-quadruplex, fluorescein and MethylRed are in
close proximity and the fluorescence is quenched,
whereas, when DNA denaturation occurs, they move
apart and a large increase in the fluorescence signal is
observed. The mid point of this temperature transition
(Tm) is related to G-quadruplex stability. Ligands that
bind effectively to this structure produce an increase in
the melting temperature, which provides a reliable indi-
cation of the G-quadruplex affinity for different
compounds.
2,6-Bis[3-(1-pyrrolidino)]propionamido]anthracene-9,l0-
dione 5 was prepared as described8e but reacting
2-bromopropionylbromide with the 2,6-diaminoanthra-
cene-9,10-dione while the analogous 2,6-bis-[3-(1-pyrr-
olidino)propionamido]anthracene-9,10-dione 50 was
obtained reacting the cloride 2 and 2-bromoethylam-
mine and then substititing the bromine in the
2,6-bis-(3-bromopropionamido)anthracene-9,10-dione
with pyrrolidine.
The results obtained using the test AQs at two concen-
trations in 50 mM potassium phosphate buffer, pH
7.4, are reported in Table 1. A representative example
is shown in Figure 2.
3. Telomerase inhibition
All derivatives stabilise the G-quadruplex structure,
which is indicative of effective binding. Interestingly,
derivatives characterised by the ‘inverse’ amide form
are more efficient stabilisers as they extensively increase
thermal stability at drug concentration as low as 1 lM,
which is close to the IC50 values for telomerase inhibi-
tion. Further increase in AQ concentration induced a
more pronounced shift in the oligonucleotide melting
temperature. These experimental data further confirm
that amido-anthraquinones containing the AQ–NHCO–
linkage binds more efficiently to G-quadruplex struc-
tures than those containing the AQ–CO–NH– linkage.
All purified anthraquinones were assayed for telomerase
inhibition by a modified telomere repeat amplification
protocol (TRAP) assay,9 using a JR8 cell extract. At
the drug concentrations used, no TAQ-mediated DNA
amplification was observed.10
Results are summarized in Table 1.
All test derivatives showed telomerase inhibition activity
in the micromolar range, although with variable
efficiency.
When comparing the differences in potency between
derivatives carrying the two different amide linkages, it
clearly emerged that compounds presenting the inverse
amide link are consistently more efficient in interfering
with the enzymatic activity.
Melting experiments were also performed with
(GGGTTA)4 annealed to its complementary sequence
to start form a double helical arrangement. In these
experiments two transitions were observed, one corre-
sponding to the change from duplex to quadruplex,
the other referring to quadruplex melting. Hence, differ-
ential stabilization of G-quadruplex vs. double helix by
The activity ratio is close to 4 for 2,6 derivatives but in-
creases up to 12 for 2,7 derivatives.
1.0
0.8
4. G-quadruplex stabilisation
To confirm that telomerase inhibition rests on G-quad-
ruplex recognition, ligand interaction with telomeric G-
quadruplexes was investigated using a synthetic oligonu-
cleotide formed by four repeats of the human telomeric
sequence, (GGGTTA)4, containing a fluorophore (fluo-
increasing drug concentration
0.6
0.4
0.2
0.0
Table 1. Telomerase inhibition EC50 and thermal stabilization of
human telomeric G-quadruplex by test amido-anthraquinones
Compound
EC50 (lM)
Ratioa
DTm (ꢀC)
1 lM
10 lM
30
45
60
75
90
4
10
5
5.2 0.4
1.2 0.3
10.4 0.4
2.5 0.4
14.5 0.4
1.2 0.3
4.3
4.2
2.0 0.2
9.1 0.3
1.8 0.3
5.9 0.3
3.3 0.2
11.3 0.3
10.0 0.3
18.1 0.3
10.8 0.3
19.2 0.3
12.9 0.3
27.3 0.3
Temperature (°C)
50
Figure 2. Fluorescence melting curves of end-labelled (GGGTTA)4
recorded in the absence (solid line) and in the presence of increasing
concentrations of compound 100 (dashed lines) in 50 mM potassium
phosphate buffer, pH 7.4. Ligand concentrations were 0.5; 1.0; 2.0; 5.0
and 10.0 lM., respectively. Curves are normalized in terms of relative
fluorescence changes.
40
100
12.1
a Ratio of the IC50 values obtained for the compound containing the
linker in the amide form and in the ‘inverse’ amide form,
respectively.

358
G. Zagotto et al. / Bioorg. Med. Chem. 16 (2008) 354–361
the test drugs could be monitored. Indeed, the first tran-
sition was poorly affected by drug presence (DTm max
4 ꢀC at 10 lM) and essentially independent upon the
nature of the side-chains. On the contrary, the second
transition was effectively stabilised by the test anthra-
cenediones, the inverse amide form producing remark-
ably higher quadruplex stabilisation. While confirming
the results previously described, these data show modest
double helix stabilisation by the test drugs and clear-cut
preference for the G-quadruplex arrangement.
5. Conformational and modelling studies
To explain the molecular basis for preferential recogni-
tion of G-quadruplexes by amido-anthraquinones hav-
ing the ‘inverse’ amide bond,
a
systematic
conformational analysis was performed to describe
and compare the conformational space generated by
the rotation of both AQ–NH–CO– and AQ–CO–NH–
dihedrals.
Figure 4. Molecular model of compound 10 bound in a low-energy
position to
a 45-nucleotide human quadruplex dimer molecule
constructed from two of the units determined crystallographically.15
As shown in Figure 3, the inverse amide is characterized
by a deep potential energy minimum corresponding to a
perfect co-planar arrangement (/ = 0ꢀ) of the AQ–NH–
CO– (inverse amide) dihedral. On the contrary, the AQ–
CO–NH– dihedral reaches its energy minimum in corre-
spondence of a torsional angle of 30ꢀ. Hence, in this case
a tilt conformation seems to represent the most stable
arrangement. It is noteworthy the non-symmetrical
shape of the curve in the left panel, which is consistent
with loss of planarity of the amide nitrogen substituents.
Considering that the likely molecular mechanism of ac-
tion of these AQ-based telomerase inhibitors rests upon
G-quadruplex stabilisation, it can be inferred that the
derivative showing the largest planar surface can better
recognize, mainly through p–p interactions, the quadru-
plex target structure. Qualitative molecular modelling
(Fig. 4) using a two-quadruplex model for human telo-
meric DNA indicates that the reverse amide substitution
results in improved interactions compared to the normal
amide molecules; more detailed computations of interac-
tion energies will be undertaken in the future.
In conclusion, the data presented here show that the
direction of the amide bond connecting the side chains
to the anthraquinone moiety plays a critical role in
modulating telomerase inhibition. According to the
proposed mode of action of this class of compounds,
the observed differences are linked to the extension of
the drug’s planar surface which confirms the impor-
tance of stacking interactions onto the G-quartet
arrangement for effective telomerase inhibition. A large
planar structure of the G-quadruplex binder plays an
additional role in enhancing selectivity. In fact, besides
Figure 3. Conformational analysis of both amide- and ‘inverse’ amide form of the newly synthesized anthracenedione derivatives.

G. Zagotto et al. / Bioorg. Med. Chem. 16 (2008) 354–361
359
increasing the energy of interaction with the nucleic
acid target arranged as a G-quaduplex, it renders
intercalation into a double helix less likely. Accord-
ingly, future design of G-quadruplex-directed amido-
anthraquinone derivatives for effective telomerase
inhibition should preferentially consider the ‘inverse’
direction of the amide bond.
1H NMR (DMSO-d6): d 13.75 (bs, 2H), 8.78 (s, 2H),
8.46 (d, J = 7.9 Hz, 2H), 8.37 (d, J = 7.9 Hz, 2H); 13C
NMR (b, DMSO-d6): d 182.47, 182.17, 166.70, 136.64,
136.38, 135.42, 133.98, 134.14, 128.32.
6.3. 2,6-and 2,7-dicarboxyl-anthraquinones chloride (2
and 20)
2,6- (or 2,7-dicarboxyl-anthraquinone) 1 or 10 (0.70 g,
2.38 mmol) was suspended in dry THF (150 mL). To
the suspension was added freshly distilled thionyl chlo-
ride (1.73 mL, 23.81 mmol) and the reaction mixture re-
fluxed for 7 h. The solvent was removed in vacuo and
the product was washed with chloroform (3 · 60 mL)
to remove any residual thionyl chloride. The 2,6- and
2,7- dicarboxyl-anthraquinone chloride 2 and 20 were
used without any further purification or characterization
directly in the next reaction (quantitative yield
assumed).
6. Experimental
Chemicals, reagents and solvents, were purchased from
Aldrich or Fluka and used without further purification.
Aminoacids were from Novabiochem. 1H and 13C NMR
spectra were recorded on a Bruker AMX300 spectrom-
eter in deuterated methyl sulfoxide (DMSO-d6) from
Cambridge Lab. and referenced to TMS (d scale). Chro-
matography and flash chromatography were performed
using silica gel Merck 60 (70-230 mesh) and Merck 60
(0.040-0.063 mm), respectively. TLC plates were also
from Merck: Merck 60 F-254, 0.25 mm precoated
plates. Melting points were determined in capillary tubes
and are uncorrected using a Gallenkamp apparatus. Ele-
mental analysis (C, H, N) were performed by the Micro-
analytical Laboratory of the Department of
Pharmaceutical Sciences of the University of Padova
on a within 0.40% of the calculated values and were per-
formed on a Carlo Erba 1016 Elemental Analyser. Exact
masses (HRMS) were obtained from a Electrospray
mass spectra were obtained using a MarinerTM API-
TOF (Perseptive Biosystems Inc.- Framingham MA
01701, USA).
6.4. 2,6- and 2,7-Boc protected ethylenediamide anthra-
quinones (3 and 30)
In a 250 mL round bottomed flask flushed with nitrogen
were introduced the 2,6- (or 2,7-dicarboxyl-anthraqui-
none chloride) (5.07 mmol) and dry tetrahydrofuran
(150 mL). To the stirring solution was added triehylamine
(11.1 mol) and then the N-(tert-butoxycarbonyl)ethylene-
diamine (9.9 mmol) in dry tetrahydrofuran (5 mL) and
the reaction mixture was warmed to reflux for 3 h. The
reaction mixture was allowed to cool down to 0 ꢀC and
left for 1 h. The resulting suspension was transferred in
to two 50 mL centrifuge tubes and spun at 4500 rpm for
5 min at +4 ꢀC.
Within the experimental part AQ stays for anthraqui-
none, ED for 1,2-ethylenediamine and b-Ala for b-
Alanine.
The supernatant was removed and the solid washed with
water (5 · 20 mL). The solid product was dried in vacuo
for 4 h at 50 ꢀC to give the 2,6-Boc protected ethylene-
diamideanthracene-9,10-dione (3.9 mmol, 68%) or 2,7-
Boc protected ethylenediamideanthracene-9,10-dione
(2.66 mmol, 52%).
Both the 2,6 and 2,7-dimethylanthracene-9,10-diones
were obtained using a previously-reported procedure.12
6.1. 2,6-Dicarboxyanthracene-9,10-dione (1)
2,6-Dimethyanthracene-9,10-dione (4.04 g, 17.10 mmol)
was dissolved in glacial acetic acid (220 mL) and CrO3
(26.44 g, 264.00 mmol) was added. The mixture was
heated to reflux for 24 h (the colour changed form or-
ange to green). The cooled mixture formed a precipitate
that was collected, washed several times with water until
the supernatant was colourless, and then with acetone.
Product 1 was obtained as slightly yellow powder
(3.92 g, yield 77%).
1H NMR (DMSO-d6): d 8.95 (bs,2H), 8.66 (s, 2H), 8.32
(d, J = 7.7 Hz, 2H), 8.29 (d, J = 7.7 Hz, 2H), 6.95 (bs,
2H), 3.33 (m, 4H), 3.16 (m, 4H), 1.37 (s, 18H).
6.5. 2,6- and 2,7- ethylenediamide anthraquinones TFA
salt (4 and 40)
In a 250 mL round bottomed flask were introduced the
2,6- (or 2,7-) Boc protected ethylenediamideanthracene-
9,10-dione (2.0 mmol) and 90% trifluoroacetic acid
water (50 mL). The reaction mixture was stirred 1 h at
room temperature and then cooled to 0 ꢀC. To the reac-
tion mixture was added cold diethyl ether (120 mL) and
resulting solid was transferred in to two 50 mL centri-
fuge tubes. The solid was spun at 4500 rpm for 5 min
at +4 ꢀC. The supernatant was removed and discarded
and the solid product in each of the centrifuge tubes
was washed with diethyl ether (5 · 30 mL) to remove
any excess trifluoroacetic acid. The solid product was
dried in vacuo for 1 h to give 2,6-ethylenediamidean-
thracene-9,10-dione TFA salt as a pale yellow solid
1H NMR (DMSO-d6): d 13.77 (bs, 2H), 8.81 (s, 2H),
8.48 (d, J = 7.9 Hz, 2H), 8.40 (d, J = 7.9 Hz, 2H); 13C
NMR (DMSO-d6): d 182.47, 166.50, 139.45, 137.03,
136.47, 135.44, 134.14, 128.33.
6.2. 2,7-Dicarboxyanthracene-9,10-dione (10)
2,7-Dimethyanthracene-9,10-dione (4.04 g, 17.10 mmol)
was oxidized as decribed before for the 2,6-isomer.
Product 10 was obtained as slightly yellow powder
(3.88 g, yield 76%).

360
G. Zagotto et al. / Bioorg. Med. Chem. 16 (2008) 354–361
(0.002 mol, 100%) or the 2,7-ethylenediamideanthra-
cene-9,10-dione TFA salt as
(0.0019 mol, 97%).
2.77 (t, J = 6.5 Hz, 4H). 13C NMR (DMSO-d6): d
181.6, 169.7, 144.6, 134.6, 128.8, 128.4, 123.8, 116.2,
34.95, 33.75. HRMS, m/z: 191.0865, (C20H20N4O4
M + 2H requires 191.0815).
a
yellow solid
1H NMR (DMSO-d6): d 9.13 (t, J = 5.3 Hz, 2 H), 8.69
(d, J = 1.5 Hz, 2H), 8.38 (dd, J = 8.2 and 1.5 Hz, 2H),
8.33 (d, J = 8.2 Hz, 2H), 8.02 (bs, 6H), 3.62-3.55 (m,
4H), 3.13–3.01 (m, 4H); 13C NMR (DMSO-d6): d 4
181.78, 165.26, 139.02, 134.72, 133.07, 133.03, 127.09
125.68, 38.41, 37.30.
6.7.3. Taq polymerase assay. Compounds were assayed
against Taq polymerase using pBR322 (2.5 ng) as a
DNA template and appropriate primer sequences
(0.5 lM) to amplify the 906–1064 sequence of plasmid
by PCR. The reaction was carried out in a Perkin-Elmer
thermocycler performing 25 cycles of: 30 s at 94 ꢀC, 30 s
at 65 ꢀC and 30 s at 72 ꢀC. The reaction products were
resolved on a 2% agarose gel in TBE 1X (89 mM TRIS
base, 89 mM boric acid, 2 mM Na2EDTA) and stained
with ethidium bromide.
13C NMR (DMSO-d6): d 40 183.45, 183.39, 166.94,
140.65, 136.43, 134.72, 134.68, 128.70, 127.32, 40.05,
38.92.
HRMS, m/z: 4 191.0853, 40 191.0905, (C20H20N4O4
M + 2H requires 191.0815).
6.7.4. Telomerase activity assay. An aliquot of 5 · 106
JR8 cells in exponential phase of growth was pelleted
and lysed for 30 min on ice using 100 ll of 0.5%
CHAPS, 1 mM EGTA, 25% 2-mercaptoethanol, 1.74%
PMSF and 10% w/v glycerol. The lysate was centrifuged
at 13000 rpm for 30 min at 4 ꢀC and the supernatant col-
lected, stored at ꢀ80 ꢀC, and used as the telomerase
source.
6.6. 2,6- or 2,7-bis-[N-(3-Fmoc-amino)-propiona-
mide]anthracene-9,10-dione – General procedure (2,6-
Fmoc-b-Ala-AQ (9 or 90)
Commercial
2,6-diamino-anthraquinone
(0.40 g,
1.68 mmol) was suspended in dry THF (60 mL) and
Fmoc-b-Ala-Cl (1.36 g, 4.12 mmol) was added. The
mixture was heated to reflux for 7 h and a colour change
from dark red to orange was observed. The product was
collected by centrifugation from the cooled solution and
washed with 1 M HCl (3 · 30 mL) and acetone
(3 · 30 mL). Product 9 (90) was obtained as a red orange
powder (1.32 g, 95%).
Telomerase activity was assayed using a modified telo-
mere repeat amplification protocol (TRAP) assay.8
Briefly, a proper primer (TS) was 50-labeled with
[c-32P]ATP and T4 polynucleotide kinase. After enzyme
inactivation (85 ꢀC for 5 min), a 50 ll TRAP reaction
mix (50 lMdNTPs, 0.2 lg labelled TS, 0.1 lg return pri-
mer ACX, 500 ng proteic extract, 2 U Taq polymerase)
was prepared in the presence/absence of increasing drug
concentration. An internal control template (0.01 mol
TSNT) with its return primer (1 ng NT) was added to
the reaction mixture. Then, telomerase elongation step
was performed (30 min at 30 ꢀC) followed by a PCR
amplification step (30 cycles of: 30 s at 37 ꢀC and 30 s
at 58 ꢀC). The reaction products were loaded onto a
10% polyacrylamide gel (19:1) in TBE. Gels were trans-
ferred to Whatman 3MM paper, dried under vacuum at
80 ꢀC and read using a phosphorimager apparatus
(Amersham).
1H NMR (DMSO-d6): d 10.59 (s, 2H), 8.50 (d,
J = 1.8 Hz, 2H), 8.15 (d, J = 8.3 Hz, 2H), 8.06 (dd,
J = 8.3 and 1.8 Hz, 2H), 7.86 (dd, J = 7.4 and 0.9 Hz,
4H), 7.68 (dd, J = 7.4 and 0.9 Hz, 4H), 7.46 (t,
J = 5.4 Hz, 2H), 7.38 (dt, J = 7.4 and 0.9 Hz, 4H), 7.29
(dt, J = 7.4 and 0.9 Hz, 4H), 4.27 (d, J = 6.4 Hz, 4H),
4.21 (t, J = 6.5 Hz, 2H), 3.33 (t, J = 6.3 Hz, 4H), 2.59
(t, J = 6.3 Hz, 4H).
6.7. 2,6- or 2,7-bis-(3-amino-propionamide)-anthracene-
9,10-dione di-trifluoroacetate—General procedure (2,6-
b-Ala-AQ TFA) (10 or 100)
Values were expressed as percent of telomerase inhibi-
tion relative to control (no drug) lanes.
6.7.1. First step. 2,6-Fmoc-b-Ala-AQ
9
(0.84 g,
1.00 mmol) was dissolved in DMF (54 mL) and piperi-
dine (6 mL). The solution was stirred 2 h at room tem-
perature and then poured in Et2O (250 mL). The free
base was collected by centrifugation of the resulting sus-
pension, washed with Et2O (3 · 50 mL) then with water
(3 · 50 mL) and dried.
6.7.5. Fluorescence melting studies. Fluorescence melting
experiments were performed in a Roche Light Cycler,
using excitation at 488 nm and recording fluorescence
emission at 520 nm. Oligonucleotides labelled with fluo-
rescein (FAM) at the 50-end of the oligonucleotide and
Methyl Red (MeRed) at 30-end were synthesized and
HPLC purified by OSWEL Research Products Ltd
(Southampton, UK).
6.7.2. Second step. The free base was dissolved in TFA/
water 9/1 (20 mL), stirred 2 h at rt and then poured in
Et2O (150 mL). The product was collected by centrifu-
gation, washed with Et2O (3 · 50 mL) then water
(3 · 50 mL) and dried. 10 was obtained as red/orange
solid (0.56 g, overall yield 92%).
Melting experiments were performed in a total volume
of 20 lL containing 0.25 lM quadruplex forming la-
belled oligonucleotide and variable concentrations of
test derivatives in LiP buffer (10 mM LiOH; 50 mM
KCl pH 7.4 with H3PO4). Alternatively, the quadru-
plex-forming sequence was annealed with its comple-
mentary sequence and then submitted to the melting
1H NMR (DMSO-d6): d 10.76 (s, 2H), 8.53 (d,
J = 1.8 Hz, 2H), 8.18 (d, J = 8.3 Hz, 2H), 8.01 (dd,
J = 8.3 and 1.8 Hz, 2H), 7.75 (bs, 6H), 3.12 (m, 4H),

G. Zagotto et al. / Bioorg. Med. Chem. 16 (2008) 354–361
361
Drug Discov 2002, 1, 383; (d) Kelland, L. R. Eur. J. Cancer
2005, 41, 971.
procedure in the presence/absence of added drugs under
the above conditions.
6. See for example: (a) 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;
(b) Kim, M. Y.; Vankayalapati, H.; Shin-Ya, K.; Wierzba,
K.; Hurley, L. H. J. Amer. Chem. Soc. 2002, 124, 2098; (c)
Leonetti, C.; Amodei, S.; D’Angelo, C.; Rizzo, A.;
Benassi, B.; Antonelli, A.; Elli, R.; Stevens, M. F. G.;
D’Incalci, M.; Zupi, G.; Biroccio, A. Mol. Pharmacol.
2004, 66, 1138; (d) 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; (e) Barbieri,
C. M.; Srinivasan, A. R.; Rzuczek, S. G.; Rice, J. E.;
LaVoie, E. J.; Pilch, D. S. Nucleic Acids Res. 2007, 35,
3272.
7. (a) Incles, C. M.; Schultes, C. M.; Kempski, H.; Koehler,
H.; Kelland, L. R.; Neidle, S. Mol. Cancer Ther. 2004, 3,
1201; (b) Pennarun, G.; Granotier, C.; Gauthier, L. R.;
Gomez, D.; Hoffschir, F.; Mandine, E.; Riou, J.-F.;
Mergny, J.-L.; Mailliet, P.; Boussin, F. D. Oncogene
2005, 24, 2917; (c) Gomez, D.; O’Donohue, M.-F.;
Mixtures were denatured by heating to 95 ꢀC at a rate of
0.1 ꢀC minꢀ1 and Tm values were determined from the
first derivatives of the melting profiles using the Roche
LightCycler software.
6.7.6. Computational methods for analyzing conformers.
All computational modeling studies were carried out on
a 10 AMD64-CPU linux cluster running under open-
Mosix architecture.
All described compounds were built using the ‘Builder’
module of Molecular Operation Environment (MOE,
version 2005.06). They were minimized using the
MMFF94 atom force field,13 implemented in the MOE
modeling package, until the rms value using the Trun-
ꢀ1
˚
A .
cated Newton method (TN) was <0.001 kcal molꢀ1
Partial charges have been calculated using the point elec-
trostatic potential (ESP) charge approach implemented
in the MOPAC (v.7) program as part of the MOE suite.
To model the effects of solvent more directly, a set of
electrostatic interaction corrections were used. MOE
suite implemented a modified version of GB/SA contact
function described by Still and co-authors.14 Conforma-
tional analyses were performed using the ‘Dihedral En-
ergy Plot’ module implemented in the MOE suite.
´
Wenner, T.; Douarre, C.; Macadre, J.; Koebel, P.;
Giraud-Panis, M.-J.; Kaplan, H.; Kolkes, A.; Shin-ya,
K.; Riou, J.-F. Cancer Res. 2006, 66, 6908; (d) Gunarat-
nam, M.; Greciano, O.; Martins, C.; Reszka, A. P.;
´
Schultes, C. M.; Macadre, J.; Riou, J.-F.; Neidle, S.
Biochem. Pharmacol. 2007, 74, 679.
8. (a) Collier, D. A.; Neidle, S. J. Med. Chem. 1988, 31, 874;
(b) Lown, J. W.; Morgan, A. R.; Yen, S. F.; Wang, Y. H.;
Wilson, W. D. Biochemistry 1985, 24, 4028; (c) Perry, P. J.;
Gowan, S. M.; Reszka, A. P.; Polucci, P.; Jenkins, T. C.;
Kelland, L. R.; Neidle, S. J. Med. Chem. 1998, 41, 3253;
(d) Perry, P. J.; Reszka, A. P.; Wood, A. A.; Read, M. A.;
Gowan, S. M.; Dosanjh, H. S.; Trent, J. O.; Jenkins, T. C.;
Kelland, L. R.; Neidle, S. J. Med. Chem. 1998, 41, 4873;
(e) Agbandje, M.; Jenkins, T. C.; McKenna, R.; Reszka,
A. P.; Neidle, S. J. Med. Chem. 1992, 35, 1418.
9. Kim, N. W.; Piatyszek, M. A.; Prowse, K. R.; Harley,
C. B.; West, M. D.; Ho, P. L.; Coviello, G. M.;
Wright, W. E.; Weinrich, S. L.; Shay, J. W. Science
1994, 266, 2011.
10. Sissi, C.; Lucatello, L.; Krapcho, A. P.; Maloney, D. J.;
Boxer, M. B.; Camarasa, M. V.; Pezzoni, G.; Menta, E.;
Palumbo, M. Bioorg. Med. Chem. 2007, 15, 555.
Acknowledgments
This work was carried out with the financial support
from the University of Padova, Italy, the Italian Minis-
try for University and Research (MIUR), Rome, Italy,
EU (Grant HPRN-CT-2000-00009) and Cancer Re-
search UK. The scientific and technical partnership of
Chemical Computing Group is gratefully acknowledged
(Padua).
References and notes
11. (a) Risitano, A.; Fox, K. R. Biochemistry 2003, 42, 6507;
(b) Darby, R. A. J.; Sollogoub, M.; McKeen, C.; Brown,
L.; Risitano, A.; Brown, M.; Brown, T.; Fox, K. R.
Nucleic Acids Res. 2002, 30, e39; (c) De Cian, A.; Guittat,
1. Greider, C. W. Cell 2001, 97, 419.
2. Blackburn, E. H. Cell 2001, 106, 61.
`
L.; Kaiser, M.; Sacca, B.; Amrane, S.; Bourdoncle, A.;
3. (a) Greider, C. W.; Blackburn, E. H. Cell 1987, 51, 887; (b)
Autexier, C.; Lue, N. F. Ann. Rev. Biochem. 2006, 75, 493.
4. (a) 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.; Weinrich, R.; Shay, J. W. Science 1994, 266, 2011;
(b) Counter, C. M.; Hirte, H. W.; Bacchetti, S.; Harley, C.
M. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 2900; (c) Shay,
J. W.; Bacchetti, S. Eur. J. Cancer 1997, 33, 787.
Alberti, P.; Teulade-Fichou, M.-P.; Lacroix, L.; Mergny,
J.-L. Methods 2007, 42, 183.
12. (a) Staab, H. A.; Sauer, M. Liebigs Ann. Chem. 1984, 742;
(b) Broene, R. D.; Diedirich, F. Tetrahedron Lett. 1991,
32, 5227.
13. Halgren, T. J. Comput. Chem. 1996, 17, 490.
14. Qiu, D.; Shenkin, S.; Hollinger, F. P.; Still, W. C. J. Phys.
Chem. 1997, 101, 3005.
15. Parkinson, G. N.; Lee, M. P.; Neidle, S. Nature 2002, 417,
87.
5. (a) Oganesian, L.; Bryan, T. M. Bioessays 2007, 29, 155;
(b) Shay, J. W.; Wright, W. E. Nature Rev. Drug Discov.
2006, 5, 577; (c) Neidle, S.; Parkinson, G. N. Nature Rev.