Naphthalene diimide scaffolds with dual reversible and covalent interaction properties towards G-quadruplex

Article abstract of DOI:10.1016/j.biochi.2011.06.015

Selective recognition and alkylation of G-quadruplex oligonucleotides has been achieved by substituted naphathalene diimides (NDIs) conjugated to engineered phenol moieties by alkyl-amido spacers with tunable length and conformational mobility. FRET-melting assays, circular dichroism titrations and gel electrophoresis analysis have been carried out to evaluate both reversible stabilization and alkylation of the G-quadruplex. The NDIs conjugated to a quinone methide precursor (NDI-QMP) and a phenol moiety by the shortest alkyl-amido spacer exhibited a planar and fairly rigid geometry (modelled by DFT computation). They were the best irreversible and reversible G-quadruplex binders, respectively. The above NDI-QMP was able to alkylate the telomeric G-quadruplex DNA in the nanomolar range and resulted 100-1000 times more selective on G-quadruplex versus single- and double-stranded oligonucleotides. This compound was also the most cytotoxic against a lung carcinoma cell line.

Full text of DOI:10.1016/j.biochi.2011.06.015

Biochimie 93 (2011) 1328e1340
Contents lists available at ScienceDirect
Biochimie
journal homepage: www.elsevier.com/locate/biochi
Research paper
Naphthalene diimide scaffolds with dual reversible and covalent interaction
properties towards G-quadruplex
Matteo Nadai b, Filippo Doria a, Marco Di Antonio c, Giovanna Sattin b, Luca Germani a,
,
a,
**
Claudia Percivalle a, Manlio Palumbo c, Sara N. Richter b , Mauro Freccero
*
a Dipartimento di Chimica, Università di Pavia, V.le Taramelli 10, 27100 Pavia, Italy
b Dipartimento di Istologia, Microbiologia e Biotecnologie Mediche, via Gabelli 63, 35121 Padova, Italy
c Dipartimento di Scienze Farmaceutiche, Università di Padova, Via Marzolo 5, 35131 Padova, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Selective recognition and alkylation of G-quadruplex oligonucleotides has been achieved by substituted
naphathalene diimides (NDIs) conjugated to engineered phenol moieties by alkyl-amido spacers with
tunable length and conformational mobility. FRET-melting assays, circular dichroism titrations and gel
electrophoresis analysis have been carried out to evaluate both reversible stabilization and alkylation of
the G-quadruplex. The NDIs conjugated to a quinone methide precursor (NDI-QMP) and a phenol moiety
by the shortest alkyl-amido spacer exhibited a planar and fairly rigid geometry (modelled by DFT
computation). They were the best irreversible and reversible G-quadruplex binders, respectively. The
above NDI-QMP was able to alkylate the telomeric G-quadruplex DNA in the nanomolar range and
resulted 100e1000 times more selective on G-quadruplex versus single- and double-stranded oligo-
nucleotides. This compound was also the most cytotoxic against a lung carcinoma cell line.
Ó 2011 Elsevier Masson SAS. All rights reserved.
Received 1 March 2011
Accepted 14 June 2011
Available online 16 June 2011
Keywords:
DNA
G-quadruplex
Recognition
Alkylation
Selectivity
Quinone methide
1. Introduction
translation of the related mRNA [7e13]. Moreover, it has been
demonstrated that addition of efficient G-4 ligands in-cellulo affects
Guanine rich oligonucleotides are capable of folding into
supramolecular structures called G-quadruplex (G-4) [1]. In more
detail, a planar tetrameric arrangement of guanines (G-tetrad),
the binding affinity of some proteins, such as SP-1 and Shelterin,
towards specific region of the genome [14e16]. Displacement of the
above proteins activates DNA damage responses and eventually
triggers biological pathways which lead to cell apoptosis or
senescence.
Indeed, it has been shown that stabilization of G-4 folded
structures by small molecules or selective antibodies is respon-
sible for oncogenes down regulation in-vitro (i.e Bcl-2, c-myc) and
telomerase inhibition [17e24]. Since telomerase induces telomere
elongation and hence is required for immortalization of the
tumour cells, the above strategy could be exploited for the
development of new selective antitumour drugs [25]. More
recently, also G-4 RNA has been shown to behave as translational
switch in-vitro, broadening the interest for such structures as tools
for unravelling biological processes regulation at the molecular
level [26].
stabilized by hydrogen bonding, can self-aggregate by
p-stacking
interactions and monovalent cation coordination (i.e Naþ and Kþ),
generating a G-4 [2]. Due to the molecular complexity of the above,
a large number of topologically different G-4 structures can be
formed depending on strand stoichiometry, strand orientation and
guanosine relative conformation [3e5]. In some cases different
monovalent cations could also lead the same oligonucleotide
sequence to fold into different G-4 topologies [6]. G-4 topologies
have been described as parallel, antiparallel or mixed-type,
depending on the direction of the G-4 forming strands. Beside
these supramolecular aspects, the growing interest on G-4 is
justified by their potential implication in biological processes
crucial for genomic stability, such as transcription of oncogenes and
Because of all these reasons, a large number of G-4 selective
ligands has been reported to date [27,28]. The common features
shared among a large number of these compounds are: the
potential for a flat conformation, the presence of an electron poor
aromatic core and cationic or protonable moieties. These charac-
teristics can be modulated in order to improve the specific
* Corresponding author. Tel.: þ390498272346; fax: þ390498272355.
** Corresponding author. Tel.: þ390382987768; fax: þ390382987323.
E-mail addresses: sara.richter@unipd.it (S.N. Richter), mauro.freccero@unipv.it
(M. Freccero).
0300-9084/$ e see front matter Ó 2011 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.biochi.2011.06.015

M. Nadai et al. / Biochimie 93 (2011) 1328e1340
1329
recognition (i) of quadruplex over duplex-DNA and (ii) between
2. Materials and methods
different G-4 sequences and topologies. Particularly promising
from this point of view are the synthetic molecules Braco-19 [29],
the related acridines [30], pyridine-2,6-dicarboxylic acid bis-qui-
nolin-2-ylamides (PDCs) [16] and quinolinium analogues [31,32],
as well as the natural product Telomestatin [33] and synthetic
related structures [34]. A series of naphthalene diimide derivatives
(NDIs) have also been described. In more detail, disubstituted NDIs
showed low selectivity for G-4 [35], while tri- and tetra-
substituted NDIs exhibited high G-4 affinity [36,37], with note-
worthy selectivity for the parallel topology [38,39]. All the
described agents act as reversible G-4 ligands. Our group has
recently developed NDI derivatives coupled to ortho-quinone
methides (o-QMs), which are potent and transient reactive elec-
trophilies [40e42], paving the route for the generation of new G-4
covalent ligands [43]. o-QM have to be generated by activation of
suitable quinone methide precursors (QMPs), under biocompatible
conditions, such as: mild thermal digestion (40 ꢀC) [40,44e46],
UVeVis irradiation [47e51] and mild mono-electronic reduction
[44]. The conjugated NDI-QMPs in Scheme 1 may act as hybrid
ligand-alkylating structures selectively targeting G-4 folding
oligonucleotides [52]. Their analogues, lacking the reactive moiety
CH2NMe3þ, behave as reversible ligands (NDI-AminoPhenol, in
Scheme 1).
All of the NDI-QMPs developed so far, present a quaternary
ammonium moiety (NMe3þ) as good leaving group for mild QM
generation (Scheme 1). Such a key structural feature also repre-
sents an important limiting factor in terms of cells permeability and
biological applications, hampering their potential pharmacological
applications. We herein describe a second generation NDI-QMP
hybrids in which a neutral o-hydroxy benzyl alcohol moiety func-
tion as QMP (Scheme 2).
Due to the alkylation selectivity towards G-4 folded oligonu-
cleotide showed by NDIs 1e6 in the biophysical assays herein
reported, we believe that such a class of ligands, could be further
exploited for unravelling new G-4 related biological processes as
well as a genome wide screening of new relevant G-4 structures.
Their G-4 selectivity is prompted by the good reversible binding
properties.
2.1. Chemical synthesis
The amines 2-amino-N-(4-hydroxyphenyl)acetamide (11) and
3-amino-N-(4-hydroxyphenyl)propanamide (12) have been
synthesized according to an original synthetic protocol, described
in the Supporting information. The compound 13 has been
synthesized by a three-step synthetic protocol starting from
2-(bromomethyl)-4-nitrophenol (see Supporting information for
details).
2-chloro-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)acet-
amide (14): The isopropylidenacetal (13) 500 mg (2.78 mmol) was
dissolved in 15 ml of dry THF followed by the addition of 1 ml of
TEA at 0 ꢀC. Chloroacetyl chloride (0.4 ml, 5.0 mmol) was added
dropwise under vigorous stirring, keeping the temperature at 0 ꢀC.
The resulting solution was stirred at 0 ꢀC for 2 h. Reaction mixture
was quenched by adding 10 ml of ethyl acetate. The resulting
suspension was washed twice with a NaHCO3 solution. The organic
layers were collected, dried on Na2SO4 and the solvent was evap-
orated under vacuum, affording the 2-chloro-N-(4-hydroxyphenyl)
acetamide as a brown solid, which was purified by flash chroma-
tography (cycloesane:ethyl acetate 8:2). Yield 93%. 1H NMR
(300 MHz, CDCl3, 25 ꢀC, TMS):
7.05 (dd, 1H, J ¼ 2.4 Hz, J ¼ 8.7 Hz), 6.8 (d, 1H, J ¼ 8.7 Hz), 4.8 (s, 2H),
4.10 (s, 2H),1.55 (s, 6H). 13C NMR (300 MHz, CDCl3):
129.4, 120.6, 119.7, 117.3, 117.0, 99.5, 60.7, 40.6, 24.5. Anal. Calcd. for
C12H14NO3Cl: C, 56.37; H, 5.52; Cl, 13.87; N, 5.48; O, 18.77. Found: C,
56.51; H, 5.47; N, 5.41.
3-chloro-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)
propanamide (15): NaHCO3 (250 mg) was added to a CH3CN
solution (15 ml) of 3-chloropropionyl chloride (0.34 ml, 3.6 mmol),
keeping the temperature at 0 ꢀC. 2-(hydroxymethyl)-4-amino iso-
propylidenacetal (13) (500 mg 1.8 mmol) in 15 ml of CH3CN was
added dropwise to the suspension and stirred for 2 h. 20 ml of
aqueous NaHCO3 were added to the mixture and the CH3CN was
removed under reduced pressure. CHCl3 extraction (3 ꢁ 50 ml) of
the resulting aqueous solution afforded the product as a brown oil,
which has been used in the next step without further purification.
78% Yield.
d
¼ 8.2 (s, 1H), 7.25 (d, 1H, J ¼ 2.4 Hz),
d
¼ 163.7,148.5,
Here we briefly describe the synthesis and the bioassays of
a small library of the tri-substituted NDI ligands 1e6 and 1ae4a, as
G-4 selective hybrid ligand/alkylating compounds and reversible
ligands, respectively (Scheme 2).
4-chloro-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)buta-
namide (16): 4-Chlorobutyryl chloride (0.47 ml, 4.18 mmol) dis-
solved in 15 ml of DMA has been treated with 260 mg NaHCO3 at r.t.
Y
Y
O
N
O
O
N
O
O
O
N
N
O
O
NDI-QMPs
OH
HO
(CH2)n
(CH2)n
O
O
N
N
O
O
H
N
H
N
NMe3
OH
Me3N
n=0,1-3
NMe3
NDI-QMP
X
HNMe3
O
X
O
HNMe3
O
NDI-QMs
HO
N
O
N
(CH2)n
O
(CH2)n
X=H,Br
NHMe2
Y
NDI-AminoMannich
+
Y
Y=NHMe2
NMe3
Me3N
O
2 HNu
O
+
NDI-QM
O
N
N
O
H
N
O
N
O
N
O
(CH2)n
NDI-adducts
HO
Nu
(CH2)n
OH
Nu
OH
NDI-AminoPhenol
X
O
O
NHMe2
Scheme 1. NDI derivatives conjugated to ortho-quinone methide (QM) precursors (QMPs) recently investigated.

1330
M. Nadai et al. / Biochimie 93 (2011) 1328e1340
Reversible binders
Ligand-alkylating Hybrids
O
O
NH
Cl
O
NH
Cl
N
H
OH
OH
Reactive QM-species
N
N
O
OH
O
O
N
N
O
O
H
N
O
H
N
O
N
H
n
N
H
n
X
X
O
O
Cl
Cl
NH
NH
n= 1, X= H 1a
Br 2a
n= 2, X= H 3a
Br 4a
n= 1, X= H
Br
1
n= 2, X= H
3
n= 3, X= H
5
2
Br 4
Br 6
Scheme 2. Tri-substituted NDIs 1e6 (as hydrochlorides)exploited as G-4 ligand-alkylating hybrid compounds. The dotted lines highlight the presence of intra-molecular H-bonding
which could restrain the conformational mobility of the structures.
A solution of 2-(hydroxymethyl)-4-amino isopropylidenacetal (13)
(500 mg 1.8 mmol) in 5 ml of DMA was added dropwise to the
resulting suspension stirring for 75 min. 20 ml of aqueous NaHCO3
were added to the mixture and after 1 h product 16 was formed as
white solid. 65% Yield. 1H NMR (300 MHz, CDCl3, 25 ꢀC, TMS):
148.4, 129.3, 120.4, 119.7, 117.4, 116.9, 99.6, 60.8, 52.8, 24.6. Anal.
Calcd. for C12H16N2O3: C, 61.00; H, 6.83; N, 11.86; O, 20.32. Found: C,
61.09; H, 6.88; N, 11.88.
3-amino-N-(2,2-dimethyl-4H-benzo[d] [1,3]dioxin-6-yl)prop-
anamide (9): Yield 71%, yellow oil. 1H NMR (300 MHz, CD3OD):
d
¼ 7.39 (bs, 1H), 7.14 (bs, 1H), 7.07 (dd, 1H, J ¼ 8.7, 1.9 Hz), 6.78
d
¼ 7.32 (d, 1H, J ¼ 2.3 Hz), 7.25 (dd, 1H, J ¼ 8.7, 2.5 Hz), 6.74 (d, 1H,
J ¼ 8.7 Hz), 3.32 (s, 2H), 3.03 (t, 2H, J ¼ 6.5 Hz), 2.57 (t, 2H, J ¼ 6.5 Hz),
1.51 (s, 6H). 13C NMR (MeOD):
(d, 1H, J ¼ 8.7 Hz), 4.84 (s, 1H), 3.68 (t, 2H, J ¼ 6 Hz), 2.55 (t, 2H,
J ¼ 7 Hz), 2.21 (q, 2H, J ¼ 6.5), 1.54 (s, 6H). 13C NMR (300 MHz,
d
¼ 172.5, 149.5, 132.9, 122.1, 121.2,
CDCl3):
d
¼ 169.5, 148.0, 130.3, 120.2, 119.7, 117.2, 116.8, 60.8, 44.3,
118.4, 118.3, 101.0, 62.1, 45.0, 39.1, 31.2, 25.2. Anal. Calcd. for
C13H18N2O3: C, 62.38; H, 7.25; N, 11.19; O, 19.18. Found: C, 62.44; H,
7.21; N, 11.23.
33.8, 44.3, 33.8, 27.8, 24.6. Anal. Calcd. for C14H18ClNO3: C, 59.26; H,
6.39; Cl, 12.49; N, 4.94; O, 16.92. Found: C, 58.98; H, 6.41; N, 4.96.
The azides 17e19 have been prepared following the general
procedure of azidation, described in the Supporting information.
2-azido-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)acet-
amide (17): 90% yield; white powder. 1H NMR (300 MHz, CDCl3,
4-amino-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)buta-
namide (10): Yield 92%, yellow oil. 1H NMR(300 MHz, CDCl3, 25 ꢀC,
TMS):
(d, 1H, J ¼ 8.6 Hz), 4.76 (s, 2H), 2.79 (bs, 2H), 2.4 (bs, 2H), 1.83 (bs,
2H), 1.5 (s, 6H). Anal. Calcd. for C14H20N2O3: C, 63.62; H, 7.63; N,
10.60; O, 18.16. Found: C, 63.70; H, 7.41; N, 10.56.
2,6-dibromonaphthalene bisanhydride was synthesized accord-
ing to standard published procedures [53].
Procedure for the synthesis of 7 and 7a. To a stirred suspen-
sion of 2,6-dibromo dianhydride (600 mg, 0.001 mol) in acetic acid
(15 ml) N,N-dimethylethylamine (1.6 ml) was added. After stirring
for 30 min at 130 ꢀC, the reaction mixture was cooled to r.t. and
quenched into ice. The crude orange solid was filtered and purified
by column chromatography (CHCl3:MeOH 9:1), yielding 7 (45%), 7a
(26%).
d
¼ 8.89 (s, 1H), 7.37 (bs, 1H), 7.10 (d, 1H, J ¼ 8.6 Hz), 6.71
25 ꢀC, TMS):
J ¼ 2.4 Hz, J ¼ 8.7 Hz), 6.75 (d, 1H, J ¼ 8.7 Hz), 4.9 (s, 2H), 4.10 (s, 2H),
1.50 (s, 6H) 13C NMR (300 MHz, CDCl3):
120.4, 119.7, 117.4, 116.9, 99.6, 60.8, 52.8, 24.6. Anal. Calcd. for
C12H14N4O3: C, 54.96; H, 5.38; N, 7.82; O, 21.36. Found: C, 55.02; H,
5.32; N, 7.88.
d
¼ 7.9 (s, 1H), 7.30 (d, 1H, J ¼ 2.4 Hz), 7.10 (dd, 1H,
d
¼ 164.2, 148.4, 129.3,
3-azido-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)prop-
anamide (18): 52% yield; brown oil. 1H NMR (300 MHz, CDCl3,
25 ꢀC, TMS):
d
¼ 7.40 (s, 1H), 7.07 (d, 1H, J ¼ 8.6 Hz), 6.79 (d, 1H,
J ¼ 8.6 Hz), 4.85 (s, 2H), 3.73 (t, 2H, J ¼ 6.2 Hz), 2.59 (t, 2H,
J ¼ 6.2 Hz), 1.55 (s, 6H). Anal. Calcd. for C13H16N4O3: C, 56.51; H,
5.84; N, 20.28; O, 17.37. Found: C, 56.49; H, 5.80; N, 20.32.
4-azido-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)buta-
namide (19): 75% yield; white powder. 1H NMR(300 MHz, CDCl3,
N,N0-Di-(N,N-dimethylethyl)-2,6-dibromonaphthalene-
1,4,5,8-tetra-carboxylic acid bisimide (7). Yellow-orange solid. 1H
NMR (200 MHz, CDCl3, 25 ꢀC, TMS):
d
¼ 9.0 (s, 2H), 4.35 (m, 4H),
25 ꢀC, TMS):
d
¼ 7.39 (bs, 1H), 7.13 (bs, 1H), 7.07 (dd, 1H, J ¼ 6.6,
2.72 (m, 4H), 2.37 (s, 12H). Anal. Calcd. for C22H22Br2N4O4: C, 46.66;
H, 3.92; Br, 28.22; N, 9.89; O, 11.30. Found: C, 46.69; H, 3.89; Br,
28.18; N, 9.92.
2 Hz), 6.79 (d, 1H, J ¼ 8.7 Hz), 4.84 (s, 2H), 3.44 (t, 2H, J ¼ 6.4 Hz),
2.46 (t, 2H, J ¼ 7.1 Hz), 2.03 (q, 2H, J ¼ 6.7 Hz), 1.55 (s, 6H). 13C NMR
(300 MHz, CDCl3):
d
¼ 169.7, 148.0, 130.3, 120.2, 119.7, 117.2, 116.8,
N,N0-Di-(N,N-dimethylethyl)-2-bromonaphthalene-1,4,5,8-
99.5, 60.8, 50.6, 33.8, 29.6, 24.6. Anal. Calcd. for C14H18N4O3: C,
57.92; H, 6.25; N, 19.30; O, 16.53. Found: C, 57.97; H, 6.19; N, 19.32.
The amines 8e10 have been synthesized by reduction of the
azido derivatives (17e19), according to the reduction protocol
described in the Supporting information.
tetra-carboxylic acid bisimide (7a). Yellow solid. 1H NMR
(200 MHz, CDCl3, 25 ꢀC, TMS):
d
¼ 8.94 (s, 1H), 8.84e8.76 (m, 2H),
4.37 (m, 4H), 2.69 (bs, 4H), 2.36 (s, 12H). Anal. Calcd. for
C22H23BrN4O4: C, 54.22; H, 4.76; Br, 16.40; N, 11.50; O, 13.13. Found:
C, 54.26; H, 4.71; Br, 16.41; N, 11.55.
2-amino-N-(2,2-dimethyl-4H-benzo[d][1,3]dioxin-6-yl)acet-
Ligands 1e6 and 1ae4a have been prepared according to the
general method described as following.
amide (8): Yield 80%, yellow oil. 1H NMR (300 MHz, CDCl3, 25 ꢀC,
TMS):
6.80 (dd, 1H J ¼ 2.4 Hz, J ¼ 8.7 Hz), 4.81 (s, 2H), 4.15 (s, 2H), 3.75
(broad s, 2H), 1.52 (s, 6H). 13C NMR (300 MHz, CDCl3):
¼ 164.2,
d
¼ 8.25 (s, 1H), 7.35 (d, 1H J ¼ 2.50 Hz), 7.10 (d,1H J ¼ 8.7 Hz),
Nucleophilic aromatic substitution reaction. The NDIs
mixture resulting from the previous step (7, 7a; 250 mg) was solved
into 20 ml of DMF containing the amine (0.6 mmol). The mixture
d

M. Nadai et al. / Biochimie 93 (2011) 1328e1340
1331
was stirred at 35 ꢀC for 24 h under argon. The resulting red solution
was poured in water (100 ml) and crashing out was observed. The
collected solid was washed with water and purified by preparative
HPLC (C-18 reverse phase column, CH3CN:H2O 0.1% TFA, as eluent).
Addition of HCl 1 M solution to each chromatographic portion and
solvent evaporation under vacuum afforded the adducts as
hydrochlorides.
4.54 (m, 4H), 4.01 (m, 2H), 3.54 (m, 4H), 3.47 (m, 2H), 3.05 (s, 6H),
2.98 (s, 6H). 13C NMR (CD3OD):
153.1, 138.9, 131.8, 130.0, 128.7, 124.6, 124.4, 123.8, 122.9, 122.0,
121.3, 116.6, 100.8, 57.5, 57.2, 44.4, 40.8, 37.6, 37.1, 37.0. Anal. Calcd.
for C31H35BrCl2N6O6: C, 50.42; H, 4.78; Br, 10.82; Cl, 9.60; N, 11.38;
O, 13.00. Found: C, 50.43; H, 4.75; Br, 10.79; Cl, 9.63; N, 11.41.
d
¼ 167.2, 163.8, 163.7, 163.0, 155.9,
Ligand 5: Yield 36%. 1H NMR(300 MHz, CD3OD):
d
¼ 8.41 (d, 1H,
Ligand 1: Yield 46%. 1H NMR (300 MHz, DMSO):
d
¼ 10.32
J ¼ 7.7 Hz), 8.15 (d, 1H, J ¼ 7.8 Hz), 8.11 (s, 1H), 7.37 (d, 1H, J ¼ 2 Hz),
7.17 (dd, 1H, J ¼ 8.5, 2.2 Hz), 6.61 (d, 1H, J ¼ 8.6 Hz), 4.55 (bs, 4H),
4.47 (bs, 2H), 3.74 (bs, 2H), 3.56 (bs, 2H), 3.49 (bs, 2H), 3.03 (s, 12H),
2.57 (t, 2H, J ¼ 6.1 Hz), 2.21 (t, 2H, J ¼ 6.2 Hz). 13C NMR (CD3OD):
(s, 1H), 10.22 (s, 1H), 9.71 (bs, 1H), 9.35 (bs, 1H), 8.57 (d, 1H,
J ¼ 7.7 Hz), 8.29 (d,1H, J ¼ 7.7 Hz), 8.04 (s,1H), 7.57 (bs,1H), 7.34 (dd,
1H, J ¼ 1.6, 8.4 Hz), 6.73 (d, 1H, J ¼ 8.4 Hz), 4.57 (s, 2H), 4.41 (m, 4H),
3.48 (m, 4H), 2.94 (s, 6H), 2.92 (s, 6H). 13C NMR (CD3OD):
d
¼ 165.4,
d
¼ 173.2, 167.3, 165.1, 164.8, 164.7, 153.9, 153.0, 132.3, 132.0, 130.9,
163.4, 163.0, 151.4, 150.2, 130.7, 130.2, 128.9, 127.7, 126.1, 124.0,
123.0, 120.4, 119.1, 118.8, 118.7, 118.5, 114.4, 99.4, 58.0, 54.6, 45.8,
42.7, 33.0. Anal. Calcd. for C31H36Cl2N6O7: C, 55.11; H, 5.37; Cl,10.50;
N, 12.44; O, 16.58. Found: C, 55.14; H, 5.31; Cl, 10.54; N, 12.48.
129.2, 129.1,127.4,125.6, 124.4,122.0,121.8,121.4,120.7,115.9,100.7,
61.0, 57.6, 57.3, 44.4, 44.4, 44.1, 43.9, 37.2, 36.7, 35.1, 26.6. Anal.
Calcd. for C33H40Cl2N6O7:C, 56.33; H, 5.73; Cl, 10.08; N, 11.94; O,
15.92. Found: C, 56.35; H, 5.69; Cl, 10.05; N, 11.96.
Ligand 1a: Yield 41%. 1H NMR (300 MHz, CD3OD):
d
¼ 10.09
Ligand 6: Yield 22%, 1H NMR(300 MHz, CD3OD):
d
¼ 8.59 (s, 1H),
(s, 1H), 8.44 (d, 1H, J ¼ 7.8 Hz), 8.19 (d, 1H, J ¼ 7.04 Hz), 7.86 (s, 1H),
7.42 (d, 2H, J ¼ 7.8 Hz), 6.79 (d, 2H, J ¼ 7.8 Hz), 4.53 (bs, 4H), 4.36
(s, 2H), 3.57 (bs, 4H), 3.03 (s, 12H). Anal. Calcd. for C30H34Cl2N6O6: C,
55.82; H, 5.31; Cl, 10.98; N, 13.02; O, 14.87. Found: C, 55.86; H, 5.28;
Cl, 11.02; N, 13.07.
8.26 (s, 1H), 7.30 (bs, 1H), 7.15 (dd, 1H, J ¼ 8.5, 2.3 Hz), 6.57 (d, 1H,
J ¼ 8.5 Hz), 4.55 (bs, 4H), 4.47 (bs, 2H), 3.78 (t, 2H, J ¼ 5.9 Hz), 3.56
(bs, 2H), 3.49 (bs, 2H), 3.04 (s, 12H), 2.56 (t, 2H, J ¼ 6.2 Hz), 2.22
(t, 2H, J ¼ 6.2 Hz). 13C NMR (CD3OD):
d
¼ 173.2, 167.2, 162.0, 163.8,
163.2, 153.5, 153.0, 138.9, 132.0, 130.2, 129.2, 128.7, 124.7, 124.5,
123.0, 122.2, 121.7, 121.6, 121.2, 115.8, 101.0, 60.9, 57.5, 57.3, 50.1,
49.9, 49.6, 49.3, 49.0, 48.7, 48.4, 44.5, 44.0, 37.6, 36.9, 35.2, 26.6.
Anal. Calcd. for C33H39BrCl2N6O7:C, 50.65; H, 5.02; Br, 10.21; Cl,
9.06; N, 10.74; O, 14.31. Found: C, 50.64; H, 5.06; Br, 10.18; Cl, 9.01;
N, 10.78.
Ligand 2: Yield 30%. 1H NMR (300 MHz, CD3OD):
d
¼ 8.65 (s,1H),
8.11 (s, 1H), 7.57 (d, 1H, J ¼ 2.5 Hz), 7.35 (dd, 1H, J ¼ 2.5, 8.6 Hz), 6.77
(d, 1H, J ¼ 8.6 Hz), 4.66 (s, 2H), 4.58 (bs, 4H), 4.47 (s, 2H), 3.57 (bs,
4H), 3.05 (s, 12H). 13C NMR (CD3OD):
d
¼ 171.4, 168.2, 164.0, 162.2,
161.8, 153.3, 153.0, 138.9, 134.4, 129.8, 128.9, 127.9, 125.5, 123.1;
122.7, 122.5, 121.9, 116.1, 101.3, 60.9, 57.7, 57.3, 47.2, 44.5, 38.6, 37.0.
Anal. Calcd. for C31H35BrCl2N6O7: C, 49.35; H, 4.68; Br, 10.59; Cl,
9.40; N, 11.14; O, 14.84. Found: C, 49.39; H, 4.65; Br, 10.58; Cl, 9.37;
N, 11.16.
2.2. Computation
A preliminary conformational search was carried out at the
semi-empiric level using PM3 level of theory. The search was per-
formed with the Spartan stochastic Monte Carlo search algorithm
that is implemented in the software package. This search resulted
in more than 20 minima. All the conformational structures result-
ing from this search provided initial geometries for the electronic
structure calculations. Geometries of all conformers were opti-
mized at the density functional theory (DFT) level with the
Gaussian 03, Revision C.02 software package [54]. These calcula-
tions were performed using gradient corrected DFT with the Becke-
Lee-Young-Parr composite exchange correlation functional (B3LYP)
[55] as implemented in the Gaussian suite of programs for elec-
tronic structure calculations. Single-point energy calculations in
aqueous solvent on the optimized gas phase geometries were
performed for all the conformers by PCM solvation model, using
UAHF atomic radii [56]. For the most stable conformers (NDI-
Mod3a and NDI-Mod3b), geometry optimization was performed
Ligand 2a: Yield 34%. 1H NMR (300 MHz, CD3OD):
d
¼ 10.09
(s, 1H), 8.56 (s, 1H), 8.13 (s, 1H), 7.42 (d, 2H, J ¼ 7.8 Hz), 6.77 (d, 2H,
J ¼ 7.8 Hz), 4.54 (bs, 4H), 4.43 (s, 2H), 3.55 (bs, 4H), 3.02 (s,12H). Anal.
Calcd. for C30H33BrCl2N6O6: C, 49.74; H, 4.59; Br, 11.03; Cl, 9.79; N,
11.60; O, 13.25. Found: C, 49.77; H, 4.61; Br, 11.01; Cl, 9.81; N, 11.62.
Ligand 3: Yield 39%. 1H NMR (300 MHz, CD3OD):
J ¼ 7.8 Hz), 8.03 (m, 2H), 7.54 (d, 1H, J ¼ 2.5 Hz), 7.26 (dd, 1H, J ¼ 2.5,
8.6 Hz), 6.72 (d, 1H, J ¼ 8.6 Hz), 4.62 (s, 2H), 4.50 (m, 4H), 3.94 (bs,
2H), 3.55 (m, 4H), 3.05 (s, 6H), 2.99 (s, 6H), 2.88 (m, 2H). 13C NMR
d
¼ 8.29 (d, 1H;
(CD3OD):
d
¼ 171.6, 167.2, 164.9, 164.7, 164.5, 153.4, 153.3, 132.4,
131.8, 130.7, 129.4, 128.9, 127.1, 125.6, 124.2, 122.5, 122.4, 121.2,
120.4, 116.2, 100.5, 60.9, 57.5, 57.1, 44.5, 44.4, 40.9, 37.2, 36.9. Anal.
Calcd. for C32H38Cl2N6O7: C, 55.74; H, 5.55; Cl, 10.28; N, 12.19; O,
16.24. Found: C, 55.71; H, 5.52; Cl, 10.32; N, 12.14.
Ligand 3a: Yield 52%. 1H NMR (300 MHz, CD3OD):
d
¼ 9.91
(s, 1H), 8.34 (d, 1H, J ¼ 7.8 Hz), 8.07 (d, 1H, J ¼ 7.8 Hz), 8.06 (s, 1H),
7.38 (d, 2H, J ¼ 8.8 Hz), 6.75 (d, 2H, J ¼ 8.8 Hz), 4.52 (m, 4H), 3.96
(t, 2H, J ¼ 5.7 Hz), 3.56 (m, 4H), 3.49 (t, 2H, J ¼ 5.7 Hz), 3.04 (s, 12H).
using a more extended basis set including a double-z basis set
augmented by diffuse s functions on C, N, and O and polarized
d functions on C, N, and O and p function on H [6-31 þ G(d,p)]. Full
geometry optimization in gas phase at PCM-B3LYP/6-31G(d) level
was performed for the conformers NDI-Mod3a and NDI-Mod3b.
Finally, single-point energies on the B3LYP/6-31 þ G(d,p) gas phase
geometries were computed to evaluate solvation effect at PCM-
B3LYP/6-31 þ G(d,p) level of theory.
13C NMR (CD3OD):
d
¼ 169.8, 165.4, 163.2, 162.9, 154.0, 151.6, 130.6,
130.0, 128.9, 127.2, 125.4, 123.8, 122.4, 121.9, 119.4, 118.7, 114.7, 98.7,
55.7, 55.4, 42.6, 42.5, 39.0, 35.3, 35.0. Anal. Calcd. for
C31H36Cl2N6O6: C, 56.45; H, 5.50; Cl, 10.75; N, 12.74; O, 14.55.
Found: C, 56.48; H, 5.47; Cl, 10.72; N, 12.71.
Ligand 4: Yield 27%. 1H NMR (300 MHz, CD3OD):
d
¼ 8.42 (s,1H),
8.18 (s, 1H), 7.54 (d, 1H, J ¼ 2.3 Hz), 7.28 (dd, 1H, J ¼ 2.3, 8.6 Hz), 6.73
(d, 1H, J ¼ 8.6 Hz), 4.76 (s, 2H), 4.52 (m, 4H), 3.97 (bs, 2H), 3.51
(m, 4H), 3.05 (s, 6H), 2.97 (s, 6H), 2.91 (m, 2H). 13C NMR (CD3OD):
2.3. FRET-melting assay
d
¼ 171.6, 167.0, 163.7, 163.6, 162.9, 153.3, 153.0, 138.8, 131.8, 129.8,
All oligonucleotides were purchased from SigmaeAldrich. Afteran
initial dilution at 1 mM in purified water, further dilutions were
carried outintherelevantbuffer. FRETassaywasperformedwithF21T
(50-d(FAM-G3[T2AG3]3-Tamra-30) with FAM: 6-carboxyfluorescein and
Tamra: 6-carboxy-tetramethylrhodamine). Fluorescence melting
curves were determined with a LightCycler II (Roche) real-time PCR
129.5, 128.5, 124.2, 122.6; 122.4, 122.0, 121.3, 116.1, 100.7, 60.9, 57.4,
57.1, 44.4, 40.9, 37.7, 37.1, 37.0. Anal. Calcd. for C32H37BrCl2N6O7: C,
50.01; H, 4.85; Br, 10.40; Cl, 9.23; N, 10.94; O, 14.57. Found: C, 50.05;
H, 4.88; Br, 10.39; Cl, 9.18; N, 10.97.
Ligand 4a: Yield 25%. 1H NMR (300 MHz, CD3OD):
d
¼ 8.53
(s, 1H), 8.26 (s, 1H), 7.38 (d, 2H, J ¼ 8.9 Hz), 6.75 (d, 2H, J ¼ 8.9 Hz),
machine, using a total reaction volume of 20 mL, with 0.25 mM of

1332
M. Nadai et al. / Biochimie 93 (2011) 1328e1340
tagged oligonucleotide in a buffer containing 10 mM lithium caco-
dylate pH 7.4 and 50 mM KCl. After a first equilibration step at 30 ꢀC
during 2 min, a stepwise increase of 1 ꢀC every minute for 65 cycles to
reach 95 ꢀC was performed and measurements were made after each
cycle with excitation at 470 nm and detection at 530 nm. The melting
of the G-quadruplex was monitored alone or in the presence of
various concentrations of compounds and/or of G4 competitor
27NHEG (50-d(TGGGGAGGGTGGGGAGGGTGGGGAAGG)-30), double-
stranded competitor ds26 (50-d(CAATCGGATCGAATTCGATCCGAT
TG)-30) and single-stranded 4GGG scrambled (50-d(GGATGT-
GAGTGTGAGTGTGAGG)-30). Final analysis of the data was carried out
using Excel and Sigma Plot software. EmissionofFAMwasnormalized
between 0 and 1, and T1/2 was defined as the temperature for which
Following this, dissociation from the surface was monitored for
300 s in running buffer. Regeneration was done using 10 mM
glycine. Analysis of the binding sensorgrams was carried out using
1:1 binding sites model using BIA evaluation software 2.0.3. The
experiments were carried out in triplicates and the standard error
was calculated.
2.6. Alkylation
For the oligonucleotide alkylation experiment, the gel-purified
and desalted 4GGG, or 4GGG scrambled (50-d(GGATGTGAGTGT-
GAGTGTGAGG)-30) oligonucleotides were 50-end-labelled with
[g-
32P]ATP by T4 polynucleotide kinase and purified by MicroSpin
the normalized emission is 0.5.
experiments ꢂ standard deviation.
D
T1/2 values are mean of 2e3
G-25 columns (Amersham Biosciences, Europe). After purification,
the oligonucleotides were re-suspended in G4 buffer (lithium
cacodylate 10 mM, pH 7.4, KCl 50 mM), heat-denatured and folded.
Before heat-denaturation 4GGG scrambled DNA was alternatively
incubated with its complementary strand to obtain the ds oligo-
nucleotide. Compounds 1, 2, 3, 4, 5 and 6 reactions with the labelled
oligonucleotide (5 pmol/sample) were performed at 40 ꢀC in G4
buffer. These conditions were selected to maintain the stability of
the target oligonucleotide structures and to obtain thermal acti-
vation of the compounds. After 24 h, samples were ethanol-
precipitated, re-suspended in formamide gel-loading buffer, and
heated at 95 ꢀC for 4 min. Reaction products were analyzed on 20%
denaturing polyacrylamide gels and visualized by phosphorimag-
ing analysis (Molecular Dynamics, Amersham Biosciences). Quan-
tification was performed by ImageQuant software (Molecular
Dynamics).
The alkylation band was excised from the gel and the product
eluted in mQ water, purified by MicroSpin G-25 columns, and
digested with 0.5 units of Exonuclease I (Fermentas) in supplied
buffer (67 mM glycineeKOH, 6.7 mM MgCl2, 1 mM DTT) for 30 min
at 37 ꢀC. Digestion was stopped by ethanol-precipitation and
samples were run on 20% denaturing polyacrylamide gels and
visualized by phosphorimaging analysis (Molecular Dynamics,
Amersham Biosciences).
2.4. CD analysis
CD experiments were performed on a Jasco J 810 spec-
tropolarimeter equipped with a NESLAB temperature controller
and interfaced to a PC100. A quartz cuvette with 5 mm path length
was used for spectra recorded from 230 to 320 nm at 2 nm band-
width, 0.1 nm step size, and 4 s time per point. The reported
spectrum of each sample represents the average of 2 scans.
Observed ellipticities were converted to mean residue ellipticity
(q
) ¼ deg ꢁ cm2 ꢁ dmolꢃ1 (Molar Ellipticity). The oligomer 4GGG
(50-d(AGGG[TTAGGG]3)-30) was diluted from stock to the final
concentration (4 mM) in lithium cacodylate buffer (10 mM, pH 7.4)
with 50 mM KCl and then annealed by heating at 95 ꢀC for 5 min,
gradually cooled to room temperature, and measured after 24 h.
Compounds at 16
mM final concentration were added after 4GGG
annealing. CD titrations were performed at a fixed 4GGG concen-
tration (usually 3e4 mM) with various concentrations (0e16 mol
equiv) of ligands at 20 ꢀC. All samples were allowed to equilibrate
overnight. A buffer baseline was collected in the same cuvette and
subtracted from the sample spectra.
Analysis of the binding data was carried out using Excel and
Sigma Plot software.
For the thermal unfolding experiments, CD spectra were recor-
2.7. Cytotoxicity
ded from 20 ꢀC to 95 ꢀC, with temperature increase of 5 ꢀC, and
processed as above. Tm and
D
H values were calculated according to
Telomerase-positive human cells were: lung carcinoma cells
A549 (ATCC number CCL-185) and human colorectal adenocar-
cinoma cells HT-29 (ATCC number HTB-38). Telomerase-negative
human cells were human foreskin fibroblast HFF-1 (ATCC
number SCRC-1041). Cell lines were obtained from American
Type Culture Collection (Rockville, MD). Cells were maintained as
a monolayer in the logarithmic growth phase at 37 ꢀC in a 5% CO2
humidified atmosphere, using Dulbecco’s modified Eagle’s
medium supplemented with 10% foetal calf serum and penicillin
the van’t Hoff equation, applied for a two state transition from
a folded to unfolded state, assuming that the heat capacity of the
folded and unfolded states are equal [57].
2.5. Surface plasmon resonance study (SPR)
SPR measurements were performed with BIAcore T100 (GE
Healthcare) system using streptavidin-coated sensor chips (Series S
Sensor chip SA, GE Healthcare). The 50- biotinylated sequence
(4GGG DNA) were heated to 95 ꢀC and annealed by slow cooling to
form quadruplex in filtered and degassed 10 mM HEPES buffer with
200 mM KCl with 0.005% surfactant Tween 20, at pH 7.4. Flow cell 2
was used as sample flow cell for the immobilization of the bio-
tinylated oligo. Flow cells 1 was left blank as control to account for
any signal generated owing to bulk solvent effect or any other effect
not specific to the DNA interaction, which was subtracted from the
signal obtained in flow cell 2. All experiments were performed at
(100 units/mL)/streptomycin (100
mg/mL) (A549 and HT-29).
Cytotoxic effects on tumour cell growth were determined by
MTT assay. NDIs were dissolved and diluted into working
concentrations with DMSO. Cells (1.75 ꢁ 104 cells/well) were
plated onto 96-microwell plates to a final volume of 100
m
L and
allowed an overnight period for attachment. At day 1, 1
mL of each
dilution of tested compounds was added per well to get a 1% final
concentration of drug solvent per well. Control cells (without any
compound but with 1% drug solvent) were treated in the exact
same conditions. Cell survival was evaluated by an MTT assay:
25 ꢀC using running buffer (0.22
mm filtered and degassed 10 mM
HEPES with 200 mM KCl and 0.005% surfactant Tween 20) at pH 7.4.
Oligonucleotide immobilized surface was exposed to the running
10
m
L of freshly dissolved solution of MTT (5 mg/mL in PBS) were
L of solu-
added to each well, and after 4 h of incubation, 100
m
buffer for at least 2 h at a flow rate of 30
m
l minꢃ1 for attaining
bilization solution [10% sodium dodecyl sulphate (SDS) and
0.01 M HCl] were added. After overnight incubation at 37 ꢀC,
absorbance was read at 540 nm. Data were expressed as mean
values of three individual experiments conducted in triplicate.
baseline stability. 1 and 1a analyte solutions at different concen-
trations (1 ꢁ 10ꢃ9 M to 1 ꢁ 10ꢃ6 M) were prepared in the running
buffer and were injected (at 30
m
l minꢃ1 for 120 s) in series.

M. Nadai et al. / Biochimie 93 (2011) 1328e1340
1333
The percentage of cell survival was calculated as follows: cell
survival 100, where
blank denotes the medium without cells.
¼
(AwellꢃAblank)/(AcontrolꢃAblank) ꢁ
O
O
O
O
O
O
O
O
i,ii,iii
iv
v
3. Results and discussion
O
NH
n
O
NH
n
O
NH
n
NH2
13
3.1. Structural features of the ligand-alkylating hybrid compounds
Cl
H2N
N3
8
17
18
19
n = 1 14
The ligands (1e6) in Scheme 2 exhibit common structural
features, such as two e(CH2)2NMe2 solubilizing side chains teth-
ered to both the imides and an additional phenol moiety, which has
been linked to the NDI aromatic core by modular alkyl-amido
spacers. The first structural diversity is the length of the spacer
and its conformational mobility, improved by adding methylene
units (from n ¼ 1 to n ¼ 3, Scheme 2). The second key structural
difference is the orto-CH2OH group on the phenol moiety. Its
presence confers alkylating properties to the NDIs 1e6, through
a mild QM generation. Therefore, on the basis of the Y substituent
(Scheme 3), it has been possible to define two distinctive subclasses
among the NDIs listed: (i) potential reversible G-4 binders (1ae4a;
Y]H, Scheme 3) and (ii) hybrid ligand-alkylating NDIs (1e6, Y]
CH2OH). The key step of the synthetic protocol is the mild nucle-
ophilic aromatic substitution (SNAr) on both the NDIs 7 and 7a with
the amines 8e12 (step c, Scheme 3).
9
2
15
10
3
16
Scheme4. Synthesis of the QMP units to be tethered to the NDI core. (i) Chloroacetyl
chloride THF, TEA, 2 h, 0 ꢀC; (ii) 3-Chloropropionyl chloride,CH3CN, NaHCO3, 2 h, r.t.;
(iii) 4-Chlorobutyryl chloride, DMA, NaHCO3, 2 h, r.t.; (iv) NaN3, DMF, 10 h, 80 ꢀC; (v)
NaBH4, NaOH, Pd/C 10%, THF dry, 30 min, r.t.
the resulting azides (17e19) to afford the amines 8e10 in good
yields, which have been used in the final SNAr step. The amines 11
and 12 have been prepared following an identical synthetic
protocol starting from 4-aminophenol (see Supporting
information).
3.3. Conjugation of the QMPs to the NDI core
3.2. Synthesis of the QMPs units
Recently, microwaved assisted functionalization of both 2,6-
dichloro- and 2,6-dibromo-NDIs (7) by aromatic nucleophilic
substitution (SNAr) has been achieved with low yields (11e19%),
using commercial amines as solvent and high temperature
(T ¼ 150 ꢀC) [36]. Such a protocol could not be exploited for teth-
ering the amine 8e10 to the NDI core. In fact, under the harsh
conditions required by the published protocol, the QMP moieties of
the NDIs underwent fast degradation through QM generation,
resulting in a sloppy mixture of oligomers. In addition, the above
conditions are not only inefficient, but also unnecessary. In fact, the
synthesis of the final QMP-NDI conjugated structures (1e6,
Scheme 3) were carried out starting from a mixture of 2-bromo-
and 2,6-dibromo-NDI (7a and 7, respectively) with the amines
8e12, in a stoichiometric ratio, by a mild and efficient SNAr. The
NDIs 7 and 7a have been synthesized in mixture according to an
optimized two-steps synthesis, starting from the commercially
available 1,4,5,8-naphthalene-tetra-carboxylic dianhydride. The
The QMP intermediates 8e10 and the phenol unreactive
analogues 11, 12 (Scheme 3) to be tethered to the NDI scaffold have
been prepared according a general synthetic procedure depicted in
Scheme 4 (Supporting information). The synthesis started from the
2-(hydroxymethyl)-4-aminophenylpropyliden acetal (13) prepared
according to standard procedures. The first step was a nucleophilic
acyl substitution on the chloroacyl chloride by the aniline 13. Three
different acyl chlorides (chloroacetyl chloride, 3-chloropropionyl
chloride, 4-chlorobutyryl chloride) have been used to modulate
the length of the tethering chain. The reaction with chloroacetyl
chloride (i, Scheme 4) was almost quantitative, unlike those using
3-chloropropionyl (ii) and 4-chlorobutyryl chloride (iii), where
reaction yields have been affected by the competing elimination
reaction. The final steps of this synthetic route were: (iv) nucleo-
philic substitution with sodium azide in DMF and (v) reduction of
O
O
O
H2N
N
(CH2)n
H
Ligand-alkylating Hybrids
HO
n=1
2
8
n= 1, Y= CH2OH X= Br
2
1
4
3
6
5
e
9
H
Y
3
N
10
n= 2, Y= CH2OH X= Br
e
H
n= 3, Y= CH2OH X= Br
H
NH
d
e
8-10
O
H
O
N
N
O
O
O
O
O
O
O
O
O
O
O
O
O
(CH2)n
c
1
Br
Br
b
N
a
2
5
4
O
O
3
6
X
11, 12
Br
Cl
N
Cl
NH
c
N
O
O
O
NH
O
Reversible binders
X
N
OH
n= 1, Y= H X= Br 2a
O
e
X=Br
X=H
H
1a
7
H2N
n= 2, Y= H X= Br
4a
7a
N
H
(CH2)n
e
H
3a
n=1
2
11
12
Scheme 3. Synthesis of the tri-substituted NDIs 1e6.Reagent and conditions: (a) dibromoisocyanuric acid (DBI), H2SO4, reflux, 12 h, yield 93%; (b) dimethylethylamine, acetic acid,
130 ꢀC, 30 min. (c) Amine 8e12, DMF, 35 ꢀC, 24 h d) Deprotection, THF:H2O 1:1, 0.5% HCl e) Reductive debromination, Na2S2O4, in water, 2 h.

1334
M. Nadai et al. / Biochimie 93 (2011) 1328e1340
first step (a, Scheme 3) is a selective 2,6-dibromination of the
anhydride with dibromocyanuric acid (DBI) in concentrated H2SO4,
followed by (b) an imidation procedure using N,N-
These calculations were performed with the 6-31G(d) basis, in the
gas phase. NDI-Mod1 exists as a single conformer (NDI-Mod1a,
Fig. 1) characterized by two strong intra-molecular H-bonding,
which freeze the structures of the spacer coplanar to the NDI
aromatic core. On the contrary, 5 different minima, with very
similar energy in gas phase (NDI-Mod2aee, Fig. 1), have been
located for the NDI with a longer spacer (NDI-Mod2). Solvent effect
computed by single-point calculation on the gas phase geometries
at B3LYP/6-31G(d) by PCM solvation model, did not dramatically
affect their relative energies (Fig. 1, data in parenthesis). Among
them, only NDI-Mod2c exhibits a structure with the spacer planar
to the NDI core.
The conformational model NDI-Mod1 was further implemented
to the NDI-Mod3 (Scheme 5) in order to investigate the effect of the
phenol moiety on the molecular shape of the tri-substituted NDIs 1,
2, 1a and 2a at higher level of theory [B3LYP/6-31 þ G(d,p)]. Two
different conformations have been located NDI-Mod3a and NDI-
Mod3b (Fig. 2). The former, which is the most stable both in gas
phase and in water solution, is completely planar. The relative
stability of the planar conformer NDI-Mod3a is remarkable in
aqueous solution where it lays more than 5 kcal/mol below the
non-planar NDI-Mod3b (Table 1).
dimethylethylendiamine, yielding
7 (45%), and 7a (26%) in
mixture. 7a was the result of a competing reductive debromination
process. The key step for the conjugation of the QMP to the NDI
aromatic core is the SNAr reaction (c, Scheme 3), which has been
performed on the 7e7a mixture resulting from the imidization
step, without any further purification. The SNAr selectively gener-
ated the asymmetrically disubstituted NDIs 2, 2a, 4, 4a and 6
(bearing both a bromine and an alkylamino moiety on the NDI
aromatic ring), together with the monosubstituted NDIs 1, 1a, 3, 3a
and 5 (Scheme 3). The resulting substituted NDIs protected as
isopropyledene acetals were used without further purification in
the last deprotection step (d, Scheme 3). The final NDIs 1e6 have
been isolated and purified by preparative HPLC (using acetoni-
trile:water with CF3COOH 0.1%). The trifluoacetate anion has been
exchanged with chloride anion for a better spectroscopic charac-
terization. In order to maximise the efficiency of the synthesis
yielding the NDIs 1, 1a, 3, 3a and 5, the conversion of the Br-
substituted NDIs 2, 2a, 4, 4a and 6 have been achieved by reduc-
tive debromination using Na2S2O4 (e, Scheme 3), according to
a published synthetic protocol [53].
These data suggest that the NDIs 1, 2, 1a and 2a, with the
shortest alkyl-amido spacer (n ¼ 1), exhibit planar and fairly rigid
shape, unlike the NDIs 3e6 and the NDI-AminoPhenol (Scheme 1),
which benefit of a much higher degree of freedom, due to a longer
3.4. Modelling the conformational mobility of the alkyl
amido side chains
(n
¼
2, 3) and conformational more flexible spacers. Such
a conformational difference should be taken into account to ratio-
nalise any striking difference in the biophysical properties between
these NDIs.
The choice of the alkyl-amido spacers connecting the NDI core
to the QMP has been suggested by the following aspects: (i) the
replacement of the quaternary ammonium as good leaving group
with an uncharged one, such as the hydroxyl group, requires an
electron rich phenol aromatic ring (by the amido group) to generate
the alkylating QM under very mild conditions [45]; (ii) the alkyl-
amido spacers may reduce the conformational mobility of the
reactive moiety through an H-bonding network (dotted lines in
Scheme 2), extending the flat shape of the NDI core. In addition, the
length of the spacer should also affect the lipophilicity of the
resulting NDIs. In order to evaluate the conformational mobility, we
compared two simplified tri-substituted NDI models NDI-Mod1,
and NDI-Mod2 (Scheme 5) exhibiting single and double methylene
spacers, respectively, by DFT calculation at B3LYP/6-31G(d) level of
theory. A more refined model NDI-Mod3 has further been inves-
tigated at B3LYP/6-31 þ G(d,p) level of theory in both gas phase and
water bulk, using polarizable continuum models (PCMs).
The conformational potential energy surfaces (PES) of both NDI-
Mod1 and NDI-Mod2 have been preliminary explored by a PM3
semi-empirical method using the Monte Carlo stochastic approach.
This search yielded a single minimum for the NDI-Mod1 and more
several minima for the flexible NDI-Mod2. The structures from this
search provided initial geometries for the electronic structure
calculations. Geometries of all conformers were optimized at the
density functional theory (DFT) level with the Gaussian 03 package.
3.5. Non-alkylating NDI derivatives selectively bind human
telomeric G-quadruplex folded DNA
The NDIs conjugated to the phenol moiety by alkyl-amido
spacers with increasing chain length were tested (n ¼ 1, 2 and 3 for
compounds 1, 3 and 5, respectively, Scheme 2). The presence of a Br
atom on the NDI core centrosymmetric to the side chain was also
evaluated on compounds 2, 4 and 6 (Scheme 2). These compounds
are all NDI-QMP conjugates and could hence be triggered by mild
thermal activation [43]. In addition, compounds that lack the ortho-
hydroxymethyl moiety on the side chain phenol were included as
reversible non-alkylating ligands (NDIs 1a, 2a, 3a and 4a,
Scheme 3), in order to be able to discriminate the effect of revers-
ible and covalent binding towards the nucleic acids.
The ability of the non-alkylating NDI scaffolds to selectively bind
the G-quadruplex (G-4) conformation of the F21T telomere-like
DNA oligonucleotide was tested by FRET-melting analysis. The
alkylating NDIs were not included because digestion at tempera-
ture higher than 40 ꢀC would have triggered QM generation.
Results are reported as the increase in the melting temperature
(D mM
T1/2) of F21T in Kþ solution caused by the treatment with 1.0
O
O
N
O
O
O
N
O
OH
O
N
O
O
O
H
N
H
N
H
N
NH2
NH2
N
H
O
N
O
N
O
O
N
O
NDI-Mod1
NDI-Mod2
NDI-Mod3
Scheme 5. Simplified tri-substituted NDI models (NDI-Mod1-3) used to evaluate the conformational mobility of the tethering alkyl-amido spacers.

M. Nadai et al. / Biochimie 93 (2011) 1328e1340
1335
Fig. 1. Computed geometries in the gas phase [at B3LYP/6-31G(d) level] for the conformers NDI-Mod1a and NDI-Mod2aee. For the latter, relative energies (in kcal/mol) in the gas
phase and in aqueous solvent (in parenthesis, by PCM single-point calculation) are reported.
NDIs (Table 2). Derivative 1a had the most important effect on
telomeric G-4 stabilization, with a
T1/2 value of 20.9 ꢀC, which is
a major positive peak at 290 nm and a shoulder at 265 nm, which
indicated a mixed antiparallel/parallel G-4 form [59]. A minor
positive peak at 250 nm is also present, probably indicating a small
amount of unfolded DNA [60]. Upon incubation with 1a, the posi-
tive peak a 290 nm increased and those at 265 and 250 nm
decreased, indicating a binding event that induced stabilization of
the antiparallel G-4 conformation over the parallel one (Fig. 3A).
The stability of 4GGG DNA in the presence of the compound was
analyzed by CD thermal unfolding and compared to that of 4GGG
DNA alone (Fig. 3B and Fig. 1S Supplementary Information). The CD
spectra of 1a incubated with the folded 4GGG DNA were recorded
at increasing temperature (range 20e95 ꢀC). At 20 ꢀC the spectrum
showed a positive peak at 290 nm, which remained stable up to
50 ꢀC. At 55 ꢀC the peak at 290 nm started to decrease and a positive
peak at 265 nm appeared. This latter reached its maximum at 65 ꢀC,
after which temperature it began to decrease. At 95 ꢀC no evident
peak was present indicating the completeness of DNA unfolding.
Tm, separately calculated for the denaturation at 290 nm and
265 nm by the van’t Hoff equation, was 64 ꢀC and 80 ꢀC, respec-
tively (Fig. 3B). 4GGG in the absence of the NDI showed one major
CD peak at 290 nm, which remained stable up to 50 ꢀC and then
decreased rapidly up to 75 ꢀC, temperature at which no more peaks
were present. The shoulder at 265 nm slightly decreased over
temperature. In these conditions 4GGG exhibited a Tm value of 61 ꢀC
(Fig. 3B). This experiment shows that 1a stabilizes the mixed par-
alleleantiparallel topology of G-4 telomeric DNA at temperature
below 50 ꢀC. However, ligand stabilization of the above confor-
D
lower than
DT1/2 values reported for other NDI compounds
(10.5 ꢀC ꢄ
D
T1/2 ꢄ 35.2 ꢀC at 0.5
mM) [36,39]. In general, compounds
with n ¼ 1 (1a and 2a) were better stabilizers than those with n ¼ 2
(3a and 4a) and H-derivatives (1a and 3a) displayed improved
stabilization over the Br-derivatives (2a and 4a) (Table 2). In addi-
tion 1a performed better than NDI-amino-Mannich (Scheme 1),
a tri-substituted alkylamino NDI recently investigated by us [52],
indicating that the introduction of an alkyl-amido spacer in the NDI
side chain improves G-4 recognition.
Next, G-4 selectivity for 1a and 2a compounds was assessed by
FRET-based competition assay, where the ability of the NDI to
retain G-4 stabilizing affinity was challenged by non-fluorescent
duplex (ds26), single-stranded (scrambled G-4), or G-4 forming
DNA (27NHEG) [58]. The latter was the only DNA capable of
effectively competing with the labelled sequence: at 27NHEG:F21T
molar ratio of 1,
DT1/2 decreased by 65% and 68% for 1a and 2a,
respectively, at molar ratio of 10 the drop was almost complete
(95% and 94% for 1a and 2a, respectively). In contrast, dsDNA was
not able to compete so efficiently: at a molar ratio of 1,
decrease was 15% and 7% for 1a and 2a, respectively; at molar ratio
of 10, T1/2 decreased by 60% and 55% for 1a and 2a, respectively
DT1/2
D
(Table 2). Hence, at molar ratio of 10, G-4 DNA competed 7e8 times
better than dsDNA. Finally, ssDNA did not compete at all, even at
molar ratio of 10. These results show that the present NDI deriva-
tives, can be considered a fairly promising and selective G-4
binding ligands.
The best performing compound, 1a, was incubated with 4GGG
DNA, a non-labelled oligonucleotide exhibiting four repeats of the
human telomeric DNA sequence, and analyzed by CD to check the
preferred stabilized G-4 conformation. To ensure equilibrium,
samples were left to equilibrate 24 h before CD spectra measure-
ment. In these conditions the 4GGG oligonucleotide presented
mation is minor (
prevalent G-4 topology is parallel-like and this conformation is
effectively stabilized by the interaction with the NDI (
Tm 19 ꢀC).
This latter
Tm value is in line with the FRET-melting data (20 ꢀC).
The apparent H difference between the unfolding of the G-4 DNA
DTm 3
ꢀC). At temperature above 50 ꢀC, the
D
D
D
alone or in the presence of 1a was w7 kcal/mol, indicating
a remarkable stabilization of the G-4 conformation upon drug
binding.
Table 1
Relative energies (in kcal/mol) of the two conformers NDI-Mod3a and NDI-Mod3b
computed at DFT level of theory.
Model
Relative energy in gas
phase
Relative free energy in
aqueous solvent
A
B
C
D
NDI-Mod3a
NDI-Mod3b
0.00
þ1.96
0.00
þ1.43
0.00
þ5.57
0.00
þ5.31
A: Geometries optimized in gas phase at B3LYP/6-31G(d) level of theory.
B: Geometries optimized in gas phase at B3LYP/6-31 þ G(d,p) level of theory.
C: Geometries optimized in aqueous solution at PCM-B3LYP/6-31G(d) level of
theory, using UHF radii. D: Computed in gas aqueous solution by single-point
calculation at PCM-B3LYP/6-31 þ G(d,p) level of theory, on gas phase optimized
structures, using UAHF radii.
Fig. 2. Geometries of the conformations for the model NDI-Mod3, located at B3LYP/6-
31G þ G(d,p) level of theory.

1336
M. Nadai et al. / Biochimie 93 (2011) 1328e1340
Table 2
D
T1/2 of reversible binding NDI compounds measured by FRET of the labelled oligonucleotide F21T in Kþ solution in the absence or presence of competitor DNAs.
NDI
D
T1/2/ꢀC
1.0
m
M
1:1 27NHEG:F21T
10:1 27NHEG:F21T
1:1 dsDNA:F21T
10:1 dsDNA:F21T
1a
2a
3a
4a
20.9 ꢂ 0.3
15.9 ꢂ 0.6
14.0 ꢂ 0.4
11.9 ꢂ 0.4
18.2 ꢂ 0.3
7.0 ꢂ 0.5 (65%)
5.0 ꢂ 0.3 (68%)
1.0 ꢂ 0.1 (95%)
1.0 ꢂ 0.2 (94%)
17.0 ꢂ 0.3 (15%)
14.0 ꢂ 0.8 (65%)
8.0 ꢂ 0.5 (60%)
7.0 ꢂ 0.3 (55%)
NDI-AminoPhenol
The stoichiometry of 4GGG/1a complex was analyzed by incu-
versus molar ratio plot for the binding of 1a to 4GGG DNA, obtained
from the CD titration data, indicates a biphasic behaviour: at ratios
below 1 extrapolation of the linear portion (dashed-line) to a ratio
of 2 suggests the initial onset of a ligand:DNA 2:1 complex, while
for ratios between 1 and 3 further interaction is evidenced. Above
a molar ratio of 3, no substantial change in ellipticity occurs. These
data are consistent with the formation of two major types of
complex (NDI:DNA 2:1 and 3:1), both characterized by remarkably
high affinity constants.
bating 4GGG DNA with increasing molar ratios of NDI and analyzing
samples by CD spectroscopy. As shown in Fig. 3C, 1a reached
binding saturation at 3:1 drug:DNA molar ratio. The ellipticity
3.6. Alkylating NDI derivatives selectively alkylate G-quadruplex
folded DNA
Alkylating NDI-QMPs present an additional hydroxymethyl
moiety on the phenol group of the side chain. As such, they are QMP
and they can be activated by mild heating (40 ꢀC), compatible with
physiological treatment.
To check that the introduction of the new chemical function did
not impair efficient reversible ligand interaction with telomeric
DNA, the ability of 1 to retain G-4 stabilization was assessed by CD
analysis. As shown in Fig. 4, incubation of 1 with 4GGG DNA at 20 ꢀC
induced an increase in the peak at 290 nm and a slight decrease in
the peak at 265 nm, similarly to its non-alkylating analogue, 1a.
However, when 1 and 1a were incubated with the telomeric
oligonucleotide at 40 ꢀC for 24 h (alkylating conditions for 1), both
CD spectra presented a much more pronounced shoulder at 265 nm
and a more intense negative peak at 240 nm (Fig. 4), indicating that
a) alkylation does not impair/modify G-4 conformation and b)
a different G-4 conformation is prevalently stabilized by both
compounds at temperatures close to physiological. To note that
4GGG did not show CD spectral variation between 20ꢀ and 40 ꢀC.
Additionally, reversible binding properties towards 4GGG DNA
was evaluated for compound 1 and 1a by surface plasmon reso-
nance (SPR) at 25 ꢀC. Kd values for compounds 1 and 1a were
6e+5
1H 20°C
1
1H 40°C
1
1a
1Hc 40°C
4e+5
G-4 40°C
2e+5
0
-2e+5
240
260
Wavelength (nm)
280
300
320
Fig. 3. CD analysis of 4GGG DNA in the presence of 1a. A) 4GGG was incubated with
increasing molar ratio of NDI (0, 0.25, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 5.0, 6.0) and
spectra were recorded at 20 ꢀC. B) 4GGG was incubated either alone or in the presence
of 1a (4:1 NDI:DNA) and spectra were recorded at increasing temperatures (20e95 ꢀC).
Molar ellipticity values at 290 nm and 265 nm were plotted against temperature. C)
Molar ellipticity values at 290 nm were plotted against different NDI:DNA molar ratios.
Fig. 4. CD analysis of 4GGG DNA in the presence of 1. 4GGG either alone or in the
presence of 1 NDI (4:1 NDI:DNA) was incubated for 24 h at 20 ꢀC or 40 ꢀC before
spectra measurement. 1a was used as a non-alkylating control to check the alkylation
effect of 1 on telomeric DNA when treated at 40 ꢀC.

M. Nadai et al. / Biochimie 93 (2011) 1328e1340
1337
(8.9 ꢂ 2.6) ꢁ 10ꢃ8 M and (1.45 ꢂ 0.29) ꢁ 10ꢃ7 M, respectively,
indicating that the introduction of the QMP into the NDI template
did not impair target recognition properties (Fig. 2S, Supporting
information).
alkylation adducts obtained at 10
16.8% and 12.8% for 1 and 2, respectively, over total DNA amount
(Fig. 5C).
In particular, both compounds induced a faster migrating
alkylation band even at the lowest tested NDI concentration (i.e.
30 nM, lanes 3, Fig. 5A and B). At higher drug amounts (starting
m
M NDI concentration reached
The ability of the alkylating NDIs to attack the 4GGG folded DNA
was next assessed by denaturing gel electrophoresis. Increasing
NDI amounts (30 nMe10
the 32P-labelled 4GGG oligonucleotide. Controls for the highest NDI
concentration incubated with DNA at
ꢀC were included.
m
M) were incubated at 40 ꢀC for 24 h with
from 1.2 mM, lanes 5, Fig. 5A and B), a second slower migrating
alkylation band appeared and increased over NDI concentration.
Contrary, control reversible NDIs did not show any alkylation band.
Based on this behaviour and on previous results [43,47e49], it is
plausible that the faster and slower migrating alkylation bands
correspond to 1:1 and 2:1 NDI:DNA adducts, respectively. We have
shown above that 3 NDI molecules can reversibly bind the G-4
telomeric structure; here we find that 1 or 2 ligand molecules can
covalently bind the G-4 DNA. Therefore, we suggest that only 2 NDI
4
Compounds 1a and 2a were used as non-alkylating control mole-
cules. As shown in Fig. 5A and B, alkylation was assessed by the
formation of two bands migrating slower than the full length
oligonucleotide.
Both in the H- and Br-alkylating NDI series, n ¼ 1 compounds
(1 and 2) were the best alkylating agents, where the sum of the two
Fig. 5. Alkylation effects of 1 NDI on 4GGG DNA. A) Increasing amounts of compounds 1, 3 and 5 along with 1a as non-alkylating control and B) 2, 4 and 6 along with 2a as non-
alkylating control (30 nMe10
M) were incubated with the 32P-labelled 4GGG DNA at 40 ꢀC for 24 h. Reaction samples were run on a 20% denaturing polyacrylamide gel and
m
visualized by phosphorimaging. 1alk and 2alk stand for first (faster migrating) and second (slower migrating) alkylation adduct, respectively. Control lanes are referred as C and are
4GGG DNA incubated at 4 ꢀC (lane 1) or 40 ꢀC (lane 2) without the compounds. C) Quantification of alkylation band amounts obtained at 10
mM for all compounds. Quantifications
were performed by three different operators and reported are mean values with standard deviations.

1338
M. Nadai et al. / Biochimie 93 (2011) 1328e1340
molecules in the reversible complex with G-4 are appropriately
positioned to mediate the successive covalent attack (i.e. loops),
while the third NDI molecule is bound in such a location where the
QM arm cannot reach a close suitable nucleophilic site (i.e. groove).
Between 1 and 2, the former was more efficient in inducing
alkylation (Fig. 5C). Compounds 3, 4, 5 and 6 were also able to
alkylate G-4 DNA at low drug concentration; however alkylation
rate did not increase over concentration. It is possible that precip-
itation occurred at higher drug amounts.
Selectivity of 1 was assessed by incubating increasing NDI
amounts (12 nMe50 mM) with 4GGG, ss-scrambled DNA (a single-
stranded scrambled 4GGG DNA whose sequence base composition,
but not order, corresponded to that of 4GGG oligo), and ds-
scrambled DNA (ssDNA and its complementary sequence). As
shown in Fig. 6, 1 was able to alkylate G-4 DNA even at 12 nM
(lane 3), while with ss-scrambled DNA no clear alkylation band was
readily appreciable; however a smeared band was detectable from
780 nM (lane 6, ss). Finally, alkylation towards ds-scrambled DNA
was very modest and appreciable starting from 12.5 mM (lane 8, ds).
Hence 1 was 100e1000 times more selective for G-4 DNA versus
different conformations adopted by single-stranded and double-
stranded oligonucleotides, respectively. Compound
similar results (data not shown).
2 yielded
Finally, the site of alkylation was investigated by enzymatic
digestion. The slowest migrating band in Fig. 5, alk2, was gel-
purified and subjected to digestion with exonuclease I (exoI), which
catalyzes the removal of nucleotides from single-stranded-DNA in
the 30 to 50 direction. As shown in Fig. 7, the purified adduct was
slightly unstable: in fact, from alk2 a 10e20% of alk1 (see Fig. 5) and
non-alkylated oligonucleotides re-formed. However, this behaviour
is not unexpected for alkylation adducts generated from electron
rich QMs. In fact, only the covalent adducts at the exocyclic NH2
groups of C, G and A moieties are thermodynamically stable, and
alkylation adducts at different sites should revert to free QM and
DNA. Given this initial consideration, the exoI enzyme cut the
control non-alkylated 4GGG all throughout the sequence and pre-
sented a major slow-down region at the level of the first loop (A7,
starting from the 50-end, indicated by the dashed-line arrow in
Fig. 7). In contrast, the alk2 purified sample presented two major
slow-down regions at the level of the second and third G repeats
(G9eG10 and G15eG16, indicated by the continuous-line arrows in
Fig. 7), which can be accounted for NDI putative alkylation sites.
Alk1 and Alk2 characterization by MS is still under investigation.
Fig. 7. ExoI digestion of the slower migrating adduct of compound 1 with 4GGG DNA.
NDI 1/4GGG adduct was gel-purified, digested with ExoI, run on a 20% denaturing
polyacrylamide gel and visualized by phosphorimaging. The non-alkylated 4GGG DNA
was gel-purified and treated in the same condition as a control. Continuous-line
arrows indicate the major slow-down sites (putative alkylation sites) in the purified
adduct. The dashed-line arrow indicate the major slow-down site in the control
non-alkylated G-4 DNA. The upper-right image shows a shorter acquisition of the non-
digested DNA signals.
Alkylating NDIs are More Cytototoxic Against Telomerase-
positive Cell Lines. The effect of the NDI derivatives against two
telomerase-positive human carcinoma cell lines (lung and colon)
was investigated. As shown in Table 3, both alkylating and non-
alkylating NDIs 1, 1a, 2 and 2a were cytotoxic in the low micro-
molar range. In general, all NDIs were more active against the
colon than the lung cell line, and 1-derived compounds were more
active than 2-derived NDIs, which parallel the alkylation/binding
Fig. 6. Selectivity of 1 NDI towards DNA conformations. Increasing amounts (12 nMe50
m
M) of 1 NDI was incubated at 40 ꢀC for 24 h with the G-4 conformed 4GGG DNA, and with
4GGG scrambled single- and double-stranded oligonucleotides whose base composition but not sequence corresponded to that of 4GGG DNA. Samples were run on a 20%
denaturing polyacrylamide gel and visualized by phosphorimaging. 1alk and 2alk stand for first (faster migrating) and second (slower migrating) alkylation adduct, respectively.

M. Nadai et al. / Biochimie 93 (2011) 1328e1340
1339
Table 3
Appendix. Supporting information
Cytotoxicity of NDIs 1 and 2 towards telomerase-positive carcinoma cell lines (Lung,
A549, and Colon, HT-29) and telomerase-negative Human Foreskin Fibroblast (HHF)
at 48 h.
Supporting information related to this article can be found
online at doi:10.1016/j.biochi.2011.06.015.
NDIs
EC50
(
m
M) HT-29
EC50
(m
M) A549
EC50 (mM) HFF
Telomeraseþ
Telomeraseþ
Telomeraseꢃ
References
1
1a
2
4.5 ꢂ 1.9
1.4 ꢂ 0.5
4.2 ꢂ 0.6
9.3 ꢂ 2.2
5.5 ꢂ 1.3
18 ꢂ 3
18 ꢂ 3
40 ꢂ 5
40 ꢂ 5
4.5 ꢂ 0.6
[1] J.T. Davis, G-quartets 40 years later: from 50-GMP to molecular biology and
supramolecular chemistry, Angew. Chem. Int. Ed. Engl. 43 (2004) 668e698.
[2] M. Gellert, M.N. Lipsett, D.R. Davies, Helix formation by guanylic acid, Proc.
Natl. Acad. Sci. 48 (1962) 2013e2018.
20.6 ꢂ 3.2
27.5 ꢂ 5.0
2a
[3] D.E. Gilbert, J. Feigon, Multistranded DNA structures, Curr. Opin. Struct. Biol. 9
(1999) 305e314.
[4] J. Suhnel, Beyond nucleic acid base pairs: from triads to heptads, Biopolymers
61 (2001) 32e51.
efficiency against G-4 DNA, which is higher for compounds 1/1a.
Interestingly, NDI cytotoxicity against telomerase-negative human
foreskin fibroblasts (HFF) was 2e10 times lower, indicating
a moderate effect on telomeres. Disappointingly, no striking
different effect was observed between alkylating and non-
alkylating compounds, which could partially be ascribed to the
low efficiency of the alkylation event. This aspect suggests the
direction for the development of future hybrid ligand-alkylating
agents, which should exhibit higher alkylating efficiency.
[5] T. Simonsson, G-quadruplex DNA structures-variations on
a theme, Biol.
Chem. 382 (2001) 621e628.
[6] S. Burge, G.N. Parkinson, P. Hazel, A.K. Todd, S. Neidle, Quadruplex DNA:
sequence, topology and structure, Nucleic Acids Res. 34 (2006) 5402e5415.
[7] P.H. Wanrooij, J.P. Uhler, T. Simonsson, M. Falkenberg, C.M. Gustafsson, G-
quadruplex structures in RNA stimulate mitochondrial transcription termi-
nation and primer formation, Proc. Natl. Acad. Sci. USA 107 (2010)
16072e16077.
[8] T. Agarwal, S. Roy, T.K. Chakraborty, S. Maiti, Furan based cyclic homo-
oligopeptides bind G-quadruplex selectively and repress c-MYC transcrip-
tion, Bioorg. Med. Chem. Lett. 20 (2010) 4346e4349.
[9] X.D. Wang, T.M. Ou, Y.J. Lu, Z. Li, Z. Xu, C. Xi, J.H. Tan, S.L. Huang, L.K. An, D. Li,
Turning off transcription of the bcl-2 gene by stabilizing the bcl-2 promoter
quadruplex with quindoline derivatives, J. Med. Chem. 53 (2010) 4390e4398.
[10] L.E. Xodo, M. Paramasivam, A. Membrino, S. Cogoi, Protein hnRNPA1 binds to
a critical G-rich element of KRAS and unwinds G-quadruplex structures:
implications in transcription, Nucleic Acids Symp. Ser. (Oxf.) 52 (2008)
159e160.
[11] C. Arnal-Herault, A. Banu, M. Barboiu, M. Michau, A. van der Lee, Amplification
and transcription of the dynamic supramolecular chirality of the guanine
quadruplex, Angew. Chem. Int. Ed. Engl. 46 (2007) 4268e4272.
[12] S. Cogoi, L.E. Xodo, G-quadruplex formation within the promoter of the KRAS
proto-oncogene and its effect on transcription, Nucleic Acids Res. 34 (2006)
2536e2549.
4. Conclusions
A new small library of tri-substituted NDI derivatives has been
prepared and tested towards G-4 folded DNA. These molecules
have been chemically engineered to embed an alkylating quinone
methide precursor (QMP), whose reactivity has been triggered by
mild thermal digestion (T ¼ 40 ꢀC). Substituents on the alkylating
moiety play a dynamical role influencing both the ligand confor-
mational flexibility and the reactivity of the alkylating agent
precursor.
The separate analysis of the reversible and irreversible binding
properties showed that these NDIs are good ligands of G-4 telo-
meric DNA (1ae4a), with a promising duplex vs quadruplex
selectivity. It is worthy to underline that the introduction of the
amido functional group in the ligand structure strongly modifies
the interaction properties of these compounds towards the G-4
folded DNA and its ligand:DNA molecular ratio. In addition, the
semi-rigid amido spacer in NDIs 1a and 2a appears to be a crucial
feature improving G-4 reversible binding properties, by conferring
planarity to the whole ligand structure.
Addition of alkylating moiety to NDIs 1e6, does not impair their
good recognition properties towards the G-4 folded DNA. Alkyl-
ation by NDIs 1 and 2 results particularly efficient against G-4 fol-
ded DNA structures, whilst negligible alkylation has been observed
towards double-stranded-DNA. The tested NDIs were found active
against two telomerase-positive cancer cell lines; in general, NDIs
showed less short-term cytotoxicity than classic anticancer drugs
(i.e. doxorubicin, paclitaxel), suggesting that their activity may be
mediated by a more specific telomerase/telomere-oriented mech-
anism of action. Therefore, these new compounds represent
promising selective telomeric G-4 alkylating agents endowed with
anticancer activity.
[13] A. Siddiqui-Jain, C.L. Grand, D.J. Bearss, L.H. Hurley, Direct evidence for
a G-quadruplex in a promoter region and its targeting with a small molecule
to repress c-MYC transcription, Proc. Natl. Acad. Sci. USA 99 (2002)
11593e11598.
[14] S.L. Palumbo, S.W. Ebbinghaus, L.H. Hurley, Formation of a unique end-to-end
stacked pair of G-quadruplexes in the hTERT core promoter with implications
for inhibition of telomerase by G-quadruplex-interactive ligands, J. Am. Chem.
Soc. 131 (2009) 10878e10891.
[15] A.K. Todd, S. Neidle, The relationship of potential G-quadruplex sequences in
cis-upstream regions of the human genome to SP1-binding elements, Nucleic
Acids Res. 36 (2008) 2700e2704.
[16] R. Rodriguez, S. Muller, J.A. Yeoman, C. Trentesaux, J.F. Riou,
S. Balasubramanian, A novel small molecule that alters shelterin integrity and
triggers a DNA-damage response at telomeres, J. Am. Chem. Soc. 130 (2008)
15758e15759.
[17] A. Rangan, O.Y. Fedoroff, L.H. Hurley, Induction of duplex to G-quadruplex
transition in the c-myc promoter region by a small molecule, J. Biol. Chem.
276 (2001) 4640e4646.
[18] K.I. McLuckie, Z.A. Waller, D.A. Sanders, D. Alves, R. Rodriguez, J. Dash,
G.J. McKenzie, A.R. Venkitaraman, S. Balasubramanian, G-Quadruplex-binding
benzo[a]phenoxazines down-regulate c-KIT expression in human gastric
carcinoma cells, J. Am. Chem. Soc. 133 (2011) 2658e2663.
[19] S.T. Hsu, P. Varnai, A. Bugaut, A.P. Reszka, S. Neidle, S. Balasubramanian,
A G-rich sequence within the c-kit oncogene promoter forms a parallel
G-quadruplex having asymmetric G-tetrad dynamics, J. Am. Chem. Soc. 131
(2009) 13399e13409.
[20] H. Fernando, A.P. Reszka, J. Huppert, S. Ladame, S. Rankin, A.R. Venkitaraman,
S. Neidle, S. Balasubramanian,
A conserved quadruplex motif located in
a transcription activation site of the human c-kit oncogene, Biochemistry 45
(2006) 7854e7860.
Due to the deep structural and environmental modifications
induced by alkylation on G-4 structures, we strongly believe that
designing chemically dynamic G-4 ligands, as the above described
NDI derivatives, could represent a useful tool for unravelling G-4
sensitive tumorigenic pathways.
[21] S. Rankin, A.P. Reszka, J. Huppert, M. Zloh, G.N. Parkinson, A.K. Todd,
S. Ladame, S. Balasubramanian, S. Neidle, Putative DNA quadruplex formation
within the human c-kit oncogene, J. Am. Chem. Soc. 127 (2005) 10584e10589.
[22] H. Li, Y. Liu, S. Lin, G. Yuan, Spectroscopy probing of the formation, recogni-
tion, and conversion of a G-quadruplex in the promoter region of the bcl-2
oncogene, Chem. Eur. J. 15 (2009) 2445e2452.
[23] J. Dai, T.S. Dexheimer, D. Chen, M. Carver, A. Ambrus, R.A. Jones, D. Yang, An
intramolecular G-quadruplex structure with mixed parallel/antiparallel
G-strands formed in the human BCL-2 promoter region in solution, J. Am.
Chem. Soc. 128 (2006) 1096e1098.
[24] C. Douarre, D. Gomez, H. Morjani, J.M. Zahm, M.F. O’Donohue, L. Eddabra,
P. Mailliet, J.F. Riou, C. Trentesaux, Overexpression of Bcl-2 is associated with
apoptotic resistance to the G-quadruplex ligand 12459 but is not sufficient to
Acknowledgements
This work was supported by MIUR, Grant FIRB-Ideas RBI-
D082ATK, Grants AIRC 5826 and 8861 (Associazione Italiana per la
Ricerca sul Cancro) and by Pavia University.

1340
M. Nadai et al. / Biochimie 93 (2011) 1328e1340
confer resistance to long-term senescence, Nucleic Acids Res. 33 (2005)
2192e2203.
formation and stability of their nucleophilic adducts, J. Am. Chem. Soc. 128
(2006) 11940e11947.
[25] J.W. Shay, W.E. Wright, Telomerase therapeutics for cancer: challenges and
new directions, Nat. Rev. Drug Discov. 5 (2006) 577e584.
[46] M. Freccero, R. Gandolfi, M. Sarzi-Amade, Selectivity of purine alkylation by
a quinone methide. Kinetic or thermodynamic control? J. Org. Chem. 68
(2003) 6411e6423.
[47] D. Verga, M. Nadai, F. Doria, C. Percivalle, M. Di Antonio, M. Palumbo,
S.N. Richter, M. Freccero, Photogeneration and reactivity of naphthoquinone
methides as purine selective DNA alkylating agents, J. Am. Chem. Soc. 132
(2010) 14625e14637.
[26] R. Shahid, A. Bugaut, S. Balasubramanian, The BCL-2 50 untranslated region
contains an RNA G-quadruplex-forming motif that modulates protein
expression, Biochemistry 49 (2010) 8300e8306.
[27] A. Arola, R. Vilar, Stabilisation of G-quadruplex DNA by small molecules, Curr.
Top. Med. Chem. 8 (2008) 1405e1415.
[28] D. Monchaud, M.P. Teulade-Fichou, A hitchhiker’s guide to G-quadruplex
ligands, Org. Biomol. Chem. 6 (2008) 627e636.
[29] A.M. Burger, F. Dai, C.M. Schultes, A.P. Reszka, M.J. Moore, J.A. Double,
S. Neidle, The G-quadruplex-interactive molecule BRACO-19 inhibits tumor
growth, consistent with telomere targeting and interference with telomerase
function, Cancer Res. 65 (2005) 1489e1496.
[30] S. Sparapani, S.M. Haider, F. Doria, M. Gunaratnam, S. Neidle, Rational design
of acridine-based ligands with selectivity for human telomeric quadruplexes,
J. Am. Chem. Soc. 132 (2010) 12263e12272.
[48] D. Verga, S.N. Richter, M. Palumbo, R. Gandolfi, M. Freccero, Bipyridyl ligands
as photoactivatable mono- and bis-alkylating agents capable of DNA cross-
linking, Org. Biomol. Chem. 5 (2007) 233e235.
[49] F. Doria, S.N. Richter, M. Nadai, S. Colloredo-Mels, M. Mella, M. Palumbo,
M. Freccero, BINOL-amino acid conjugates as triggerable carriers of DNA-
targeted potent photocytotoxic agents, J. Med. Chem. 50 (2007) 6570e6579.
[50] S. Colloredo-Mels, F. Doria, D. Verga, M. Freccero, Photogenerated quinone
methides as useful intermediates in the synthesis of chiral BINOL ligands,
J. Org. Chem. 71 (2006) 3889e3895.
[31] A. De Cian, E. Delemos, J.L. Mergny, M.P. Teulade-Fichou, D. Monchaud, Highly
efficient G-quadruplex recognition by bisquinolinium compounds, J. Am.
Chem. Soc. 129 (2007) 1856e1857.
[51] (a) S.N. Richter, S. Maggi, S.C. Mels, M. Palumbo, M. Freccero, Binol quinone
methides as bisalkylating and DNA cross-linking agents, J. Am. Chem. Soc. 126
(2004) 13973e13979;
[32] P. Yang, A. De Cian, M.P. Teulade-Fichou, J.L. Mergny, D. Monchaud, Engi-
neering bisquinolinium/thiazole orange conjugates for fluorescent sensing of
G-quadruplex DNA, Angew. Chem. Int. Ed. Engl. 48 (2009) 2188e2191.
[33] M.Y. Kim, H. Vankayalapati, K. Shin-Ya, K. Wierzba, L.H. Hurley, Telomestatin,
a potent telomerase inhibitor that interacts quite specifically with the human
telomeric intramolecular g-quadruplex, J. Am. Chem. Soc. 124 (2002) 2098e2099.
[34] M. Tera, H. Ishizuka, M. Takagi, M. Suganuma, K. Shin-ya, K. Nagasawa,
Macrocyclic hexaoxazoles as sequence- and mode-selective G-quadruplex
binders, Angew. Chem. Int. Ed. Engl. 47 (2008) 5557e5560.
[35] C. Sissi, L. Lucatello, A. Paul Krapcho, D.J. Maloney, M.B. Boxer, M.V. Camarasa,
G. Pezzoni, E. Menta, M. Palumbo, Tri-, tetra- and heptacyclic perylene
analogues as new potential antineoplastic agents based on DNA telomerase
inhibition, Bioorg. Med. Chem. 15 (2007) 555e562.
[36] F. Cuenca, O. Greciano, M. Gunaratnam, S. Haider, D. Munnur, R. Nanjunda,
W.D. Wilson, S. Neidle, Tri- and tetra-substituted naphthalene diimides as
potent G-quadruplex ligands, Bioorg. Med. Chem. Lett. 18 (2008) 1668e1673.
[37] M. Gunaratnam, S. Swank, S.M. Haider, K. Galesa, A.P. Reszka, M. Beltran,
F. Cuenca, J.A. Fletcher, S. Neidle, Targeting human gastrointestinal stromal
tumor cells with a quadruplex-binding small molecule, J. Med. Chem. 52
(2009) 3774e3783.
[38] G. Collie, A.P. Reszka, S.M. Haider, V. Gabelica, G.N. Parkinson, S. Neidle,
Selectivity in small molecule binding to human telomeric RNA and DNA
quadruplexes, Chem. Commun. 48 (2009) 7482e7484.
(b) C. Percivalle, A. La Rosa, D. Verga, F. Doria, M. Mella, M. Palumbo, M. Di
Antonio, M. Freccero, Quinone methide generation via photoinduced electron
transfer, J. Org. Chem. 76 (2011) 3096e3106.
[52] F. Doria, M. Nadai, M. Di Antonio, R. Stocco, L. Germani, C. Percivalle, M.
Petenzi, D. Verga, C. Sissi, N. Zaffaroni, M. Folini, G. Cimino-Reale, M. Palumbo,
S.N. Richter, M. Freccero, Hybrid ligand/alkylating agents targeting telomeric
G-quadruplex structures, Chem. Biol., submitted for publication.
[53] F. Doria, M. Di Antonio, M. Benotti, D. Verga, M. Freccero, Substituted
heterocyclic naphthalene diimides with unexpected acidity. Synthesis, prop-
erties, and reactivity, J. Org. Chem. 74 (2009) 8616e8625.
[54] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, J.A. Montgomery Jr., T. Vreven, K.N. Kudin, J.C. Burant,
J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi,
G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian, J.B. Cross, C. Adamo,
J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi,
C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador,
J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain,
O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz,
Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng,
A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen,
M.W. Wong, C. Gonzalez, J.A. Pople, Wallingford CT, Gaussian 03, Revision
C.02. Gaussian, Inc., 2004.
[39] S.M. Hampel, A. Sidibe, M. Gunaratnam, J.F. Riou, S. Neidle, Tetrasubstituted
naphthalene diimide ligands with selectivity for telomeric G-quadruplexes
and cancer cells, Bioorg. Med. Chem. Lett. 20 (2010) 6459e6463.
[40] E. Modica, R. Zanaletti, M. Freccero, M. Mella, Alkylation of amino acids and
glutathione in water by o-quinone methide. Reactivity and selectivity, J. Org.
Chem. 66 (2001) 41e52.
[55] J.E. Del Bene, W.B. Person, K. Szczepaniak, Properties of hydrogen-bonded
complexes obtained from the B3LYP functional with 6-31G(d,p) and 6-
31þG(d,p) basis-sets: comparison with MP2/6-31þG(d,p) results and exper-
imental-data. J. Phys. Chem. 99 (1995) 10705e10707.
[41] C. Di Valentin, M. Freccero, R. Zanaletti, M. Sarzi-Amade, o-Quinone methide
as alkylating agent of nitrogen, oxygen, and sulfur nucleophiles. The role of H-
bonding and solvent effects on the reactivity through a DFT computational
study, J. Am. Chem. Soc. 123 (2001) 8366e8377.
[42] M. Freccero, C. Di Valentin, M. Sarzi-Amade, Modeling H-bonding and solvent
effects in the alkylation of pyrimidine bases by a prototype quinone methide:
a DFT study, J. Am. Chem. Soc. 125 (2003) 3544e3553.
[43] M. Di Antonio, F. Doria, S.N. Richter, C. Bertipaglia, M. Mella, C. Sissi,
M. Palumbo, M. Freccero, Quinone methides tethered to naphthalene diimides
as selective G-quadruplex alkylating agents, J. Am. Chem. Soc. 131 (2009)
13132e13141.
[44] M. Di Antonio, F. Doria, M. Mella, D. Merli, A. Profumo, M. Freccero, Novel
naphthalene diimides as activatable precursors of bisalkylating agents, by
reduction and base catalysis, J. Org. Chem. 72 (2007) 8354e8360.
[45] E.E. Weinert, R. Dondi, S. Colloredo-Melz, K.N. Frankenfield, C.H. Mitchell,
M. Freccero, S.E. Rokita, Substituents on quinone methides strongly modulate
[56] V. Barone, M. Cossi, J. Tomasi, A new definition of cavities for the computation
of solvation free energies by the polarizable continuum model, J. Chem. Phys.
107 (1997) 3210e3221.
[57] N.J. Greenfield, Using circular dichroism collected as a function of temperature
to determine the thermodynamics of protein unfolding and binding interac-
tions, Nat. Protoc. 1 (2006) 2527e2535.
[58] A. De Cian, L. Guittat, M. Kaiser, B. Saccà, S. Amrane, A. Bourdoncle, P. Alberti,
M.P. Teulade-Fichou, L. Lacroix, J.L. Mergny, Fluorescence-based melting
assays for studying quadruplex ligands, Methods 42 (2007) 183e195.
[59] A. Ambrus, D. Chen, J. Dai, T. Bialis, R.A. Jones, D. Yang, Human telomeric
sequence forms a hybrid-type intramolecular G-quadruplex structure with
mixed parallel/antiparallel strands in potassium solution, Nucleic Acids Res.
34 (2006) 2723e2735.
[60] R.D. Gray, L. Petraccone, J.O. Trent, J.B. Chaires, Characterization of a Kþ
-
induced conformational switch in a human telomeric DNA oligonucleotide
using 2-aminopurine fluorescence, Biochemistry 49 (2010) 179e194.