Dinuclear nickel(II) triple-stranded supramolecular cylinders: Syntheses, characterization and G-quadruplexes binding properties

Article abstract of DOI:10.1016/j.saa.2013.12.102

Three dinuclear nickel triple-stranded supramolecular cylinders [Ni ₂(L1)₃][ClO₄]₄ (1), [Ni ₂(L2)₃][ClO₄]₄ (2) and [Ni ₂(L3)₃][ClO₄]₄ (3) with bis(pyridylimine) Schiff base containing triphenyl groups in the spacers as ligands were synthesized and characterized. The human telomeric G-quadruplexes binding properties of cylinders 1-3 were evaluated by means of UV-Vis spectroscopy, circular dichroism (CD) spectroscopy and fluorescence resonance energy transfer (FRET) melting assay. UV-Vis studies revealed that the supramolecular cylinders 1-3 could bind to G-quadruplex DNA with high binding constants (K_b values ranging from 0.11-2.2 × 10⁶ M^-1). FRET melting studies indicated that the cylinders 1-3 had much stronger stabilizing effect on G-quadruplex DNA (ΔT_m up to 24.5 C) than the traditional cylinder Ni₂L34+ just containing diphenylmethane spacers (ΔT_m = 10.6 C). Meanwhile, cylinders 1-3 were found to have a modest degree of selectivity for the quadruplex DNA versus duplex DNA in competition FRET assays. Moreover, CD spectroscopy revealed that complex 1 could induce G-quadruplex formation in the absence of metal ions solution and convert antiparallel G-quadruplex into hybrid structure in Na ⁺ solution. These results provided a new insight into the development of supramolecular cylinders as potential anticancer drugs targeting G-quadruplex DNA.2014 Elsevier B.V. All rights reserved.

Full text of DOI:10.1016/j.saa.2013.12.102

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and
Biomolecular Spectroscopy
journal homepage: www.elsevier.com/locate/saa
Dinuclear nickel(II) triple-stranded supramolecular cylinders: Syntheses,
characterization and G-quadruplexes binding properties
Xin-Xin Xu a, Jing-Jing Na a, Fei-Fei Bao a, Wen Zhou a, Chun-Yan Pang a, Zaijun Li a, Zhi-Guo Gu a,b,
⇑
a School of Chemical and Material Engineering, Jiangnan University, Wuxi 214122, PR China
b The Key Laboratory of Food Colloids and Biotechnology, Ministry of Education, School of Chemical and Material Engineering, Jiangnan University, Wuxi 214122, PR China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
ꢀ
ꢀ
ꢀ
Three dinuclear nickel triple-stranded
supramolecular cylinders were
synthesized.
The cylinders could bind to G-
quadruplexes with high binding
constants.
1 could convert G-quadruplexes from
antiparallel to hybrid structure in Na+
solution.
a r t i c l e i n f o
a b s t r a c t
Article history:
Three dinuclear nickel triple-stranded supramolecular cylinders [Ni2(L1)3][ClO4]4 (1), [Ni2(L2)3][ClO4]4
(2) and [Ni2(L3)3][ClO4]4 (3) with bis(pyridylimine) Schiff base containing triphenyl groups in the spacers
as ligands were synthesized and characterized. The human telomeric G-quadruplexes binding properties
of cylinders 1–3 were evaluated by means of UV–Vis spectroscopy, circular dichroism (CD) spectroscopy
and fluorescence resonance energy transfer (FRET) melting assay. UV–Vis studies revealed that the supra-
molecular cylinders 1–3 could bind to G-quadruplex DNA with high binding constants (Kb values ranging
from 0.11–2.2 ꢁ 106 Mꢂ1). FRET melting studies indicated that the cylinders 1–3 had much stronger sta-
Received 3 November 2013
Received in revised form 10 December 2013
Accepted 23 December 2013
Available online 7 January 2014
Keywords:
Dinuclear
Supramolecular cylinders
Schiff base
bilizing effect on G-quadruplex DNA (
ing diphenylmethane spacers (
D
Tm up to 24.5 °C) than the traditional cylinder Ni2L4þ just contain-
3
DTm = 10.6 °C). Meanwhile, cylinders 1–3 were found to have a modest
degree of selectivity for the quadruplex DNA versus duplex DNA in competition FRET assays. Moreover,
CD spectroscopy revealed that complex 1 could induce G-quadruplex formation in the absence of metal
ions solution and convert antiparallel G-quadruplex into hybrid structure in Na+ solution. These results
provided a new insight into the development of supramolecular cylinders as potential anticancer drugs
targeting G-quadruplex DNA.
G-quadruplex
DNA binding
Stabilizers
Ó 2014 Elsevier B.V. All rights reserved.
Introduction
the past decades [1]. G-quadruplex DNA sequences have an unu-
sual structure formed from stacks of four in plane guanine bases
G-quadruplex DNA has attracted much interest in several re-
search fields including chemistry, biology and pharmacology over
in a circle through hoogsteen hydrogen-bonding interactions [2].
Alkali metal cations (such as Na+ and K+) can further stabilize sec-
ondary structures of G-quadruplex DNA via engaging in electro-
static interactions with the guanine carbonyl groups [3]. Folding
telomeric DNA into four-stranded G-quadruplexes can inhibit telo-
merase elongation of telomeres by interrupting the interaction be-
tween the enzyme and single-stranded telomeric DNA substrate
⇑
Corresponding author at: School of Chemical and Material Engineering,
Jiangnan University, Wuxi 214122, PR China. Tel.: +86 510 85917090; fax: +86
510 85917763.
E-mail address: zhiguogu@jiangnan.edu.cn (Z.-G. Gu).
1386-1425/$ - see front matter Ó 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.saa.2013.12.102

22
X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
[4]. Telomere quadruplexes are potential tumor-selective drug tar-
gets for cancer chemotherapy and modulation of gene transcrip-
tion [5]. Therefore, G-quadruplexes stabilizing compounds have
become an attractive platform for designing anticancer drugs [6].
Molecules can interact with the G-quadruplexes by face, edge,
loop or groove recognition through stacking, H-bonding, and elec-
trostatic interactions. The weak non-covalent interactions are very
important in controlling the structure and are vital for molecular
recognition of G-quadruplex DNA [7]. The majority of quadruplex
DNA binders reported to date based on organic molecules with
spectroscopy, CD spectroscopy and FRET melting assay were used
to assess the human telomeric G-quadruplex DNA binding proper-
ties of the cylinders 1–3. The experimental results have shown that
the structure characters of the cylinders played an important role
in the stabilizing ability of G-quadruplexes.
Experimental
Materials and methods
large planar
p-aromatic surface could stack on the quadruplex
DNA oligomers or primers, 50-AG3(T2AG3)-30 (HTG22), double
stranded competitor 50-CAATCGGATCGAATTCGATCCGATTG-30
(ds26), 50-FAM-AG3(T2AG3)-TAMRA-30 (F22T) (FAM = 6-carboxy-
fluorescein, TAMRA = 6-carboxytetrame-thylrhodamine), were
synthesized by Sangon Biotechnology (Shanghai, China) and were
used without further purification. Calf thymus DNA (ct-DNA) was
purchased from Sigma–Aldrich (Shanghai, China) Co. Ltd. Concen-
trations of HTG22 and F22T were determined by measuring the
absorbance at 260 nm after melting. Single-strand extinction coef-
ficients were calculated from mononucleotide data using a near-
[8]. Some metal complexes with hetero aromatic multidentate li-
gands have been proved to be excellent stabilizers of quadruplex
DNA [9]. For example, functionalized nickel(II) salphen complexes
with cationic side arms displayed a high degree of G-quadruplexes
stabilization (at best
DTm = 33.8 °C) and considerable selectivity
(50-fold) for G-quadruplex DNA versus duplex DNA [10]. Recently
supramolecular complexes have emerged as a new and very excit-
ing class of quadruplex DNA stabilizers [11]. Qu and co-workers re-
ported that tetracationic chiral bimetallic cylinders [M2L3]4+
(M = FeII, NiII) based on bis(pyridylimine) ligand (Scheme 1a),
whose shapes were comparable with the recognition surface of
zinc finger protein, were shown to interact with quadruplex DNA
and to exhibit good antitumor activities [12,13]. Only P-enantio-
mer of [M2L3]4+ had the ability to convert antiparallel G-quadru-
plex into a hybrid form in the presence of sodium and to
selectively induce G-quadruplex formation under salt-deficient
conditions [12,13]. However, to the best of our knowledge, no
other bimetallic cylinder based on diimine ditopic ligands except
for [M2L3]4+ systems has been reported as G-quadruplex
stabilizers.
In this paper, three new bis(pyridylimine) Schiff base ligands
(Scheme 1b) containing triphenyl groups in the spacers (L1–L3) in-
stead of L have been used to construct novel dinuclear nickel(II)
supramolecular cylinders [Ni2(L1)3][ClO4]4 (1), [Ni2(L2)3][ClO4]4
(2) and [Ni2(L3)3][ClO4]4 (3) and analyze the influence of this sub-
stitution on G-quadruplex DNA binding abilities. As we know, nick-
el is an essential element in many biological processes [14], and
many new nickel(II) complexes have been researched for potential
applications in various fields [15]. Meanwhile, the four positive
charges of the cylinders will significantly contribute to the strength
of the electrostatic interactions of such species to the anionic DNA
[16]. More importantly, the introduction of the additional aromatic
rings in the spacers can increase the size of the cylinders which is
possible to enhance the non-covalent interactions between the
synthetic supramolecular cylinders and G-quadruplexes. UV–Vis
est-neighbor
approximation
[17].
The
formations
of
intramolecular G-quadruplexes were carried out as follows: the
oligonucleotide samples, dissolved in different buffers (buffer A:
10 mM tris(hydroxymethyl)methanamin-HCl (Tris–HCl), pH = 7.4;
buffer B: 100 mM NaCl, 10 mM Tris–HCl, pH = 7.4; buffer C:
100 mM KCl, 10 mM Tris–HCl, pH = 7.4.), were heated to 95 °C for
5 min, gently cooled to room temperature and then incubated at
4 °C overnight. The solution of ct-DNA was prepared in buffer D
(50 mM NaCl, 5 mM Tris–HCl, pH = 7.2), and gave the ratio of UV
absorbance at 260 and 280 nm, A260/A280 = 1.9, indicating that the
ct-DNA was sufficiently free of protein [18]. The ct-DNA concentra-
tion (moles of bases per liter) was determined spectroscopically by
using the molar extinction coefficient:
e -
260 = 6600 cmꢂ1 molꢂ1
dm3. All stock solutions were stored at 4 °C and used after no more
than 4 days. All other reagents and solvents were purchased from
commercial sources and used without further purification. Ultra-
pure water (18.2 MX cm) was used in all experiments. Concentra-
tions of stock solutions of the metal complexes were 1 mM in
acetonitrile. Further dilution was made in the corresponding buffer
according to the required concentrations for all the experiments.
All other reagents employed were of analytical reagent grade or
with highest quality and were purchased from commercial sources
and used without further purification.
1H NMR spectra ware recorded on AVANCE III (400 MHz) instru-
ment at 298 K using standard Bruker software. The spectra were
Scheme 1. (a) Schematic drawing of ligand L and triple-stranded cylinder cations [M2L3]4+ (M = FeII, NiII). (b) Schematic drawing of the ligands L1–L3.

X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
23
internally referenced using the residual protio solvent resonance
relative to tetramethylsilane (d = 0 ppm). The electrospray ioniza-
tion mass spectrometry (ESI-MS) spectra were recorded using an
LCQ fleet APT/SSQ-710 ESI-MS spectrometer (Finnigan MAT). Infra-
red spectra were measured on an ABB Bomem FTLA 2000-104
spectrometer with KBr pellets in the 400–4000 cmꢂ1 region. Ele-
ment analyses were conducted on elementar corporation vario EL
III analyzer. UV–Vis absorbance spectra were collected on Shima-
dzu UV-2101 PC scanning spectrophotometer.
days following the slow diffusion of ether into the MeCN solution.
Yield, (44%). ESI-MS (m/z): 382.33 [Ni2(L1)3]4+ 100%, 542.75 [Ni2(-
L1)3(ClO4)]3+ 30%, 863.25 [Ni2(L1)3(ClO4)2]2+ 23%; IR (KBr cmꢂ1):
V = 3070 (w), 1631 (w), 1593 (s), 1497 (vs), 1477 (w), 1444 (w),
1263 (w), 1223 (w), 1078 (vs), 968 (w), 845 (m), 777 (w), 623
(m); UV–Vis (H2O/MeCN nm): 331 (
e = 15150), 282 (e = 23000),
241 ( = 37500), 205 ( = 77600); Anal. Calcd (%) for Ni2C90N12O22-
e
e
H66Cl4: C 56.10, H 3.45, N 8.72, Found: C 56.21, N 8.67, H 3.51.
Synthesis of [Ni2(L2)3][ClO4]4 (2)
Synthesis
[Ni2(L2)3][ClO4]4 (2) was prepared by the same method as
above for 1 but with ligand L2 (0.09 mmol, 0.047 g) to give a yield
of 40%. ESI-MS (m/z): 421.42 [Ni2(L2)3]4+ 93%, 594.83 [Ni2(L2)3(-
ClO4)]3+ 100%, 941.42 [Ni2(L2)3(ClO4)2]2+ 67%; IR (KBr cmꢂ1):
V = 3057 (w), 2972 (w), 2870 (w), 1624 (w), 1599 (w), 1541 (w),
1504 (w), 1451 (w), 1361 (w), 1117 (vs), 1086 (s), 841 (w), 777
Synthesis of L1
A solution of 2-pyridinecarboxaldehyde (0.512 g, 4.8 mmol) in
acetonitrile (15 mL) was added dropwise to a solution of 1,3-
bis(4-aminophenoxy)benzene (0.702 g, 2.4 mmol) in 15 mL aceto-
nitrile at room temperature. The reaction mixture was heated at
reflux for 3 h. After the reaction was completed, the solvent was re-
moved under reduced pressure to get the crude ligand, which was
washed with ethanol several times and dried in vacuo to get L1 as
yellow powder in sufficient high purity. Yield, 76%. 1H NMR
(400 MHz, CDCl3 d ppm): 8.715 (d, J = 4.4 Hz, 2H, H2), 8.630 (s,
2H, HAC@N), 8.207–8.188 (d, J = 7.6 Hz, 2H, H5), 7.810 (t, 2H, H4),
7.367–7.338 (m, 2H, H3), 7.338–7.264 (m, 5H, Ha,e), 7.083 (d,
J = 12.4 Hz, 4H, Hb), 6.786–6.765 (d, 3H, Hc,d); IR (KBr cmꢂ1):
(w), 629 (m); UV–Vis (H2O/MeCN nm): 329 (
e = 18,700), 245
(e
= 25,500), 205 ( = 81,050); Anal. Calcd (%) for Ni2C108N12O16-
e
H102Cl4: C 62.27, H 4.93, N 8.07, Found: C 62.33, H 4.89, N 8.10.
Synthesis of [Ni2(L3)3][ClO4]4 (3)
[Ni2(L3)3][ClO4]4 (3) was prepared by the same method as
above for 1 but with ligand L3 (0.09 mmol, 0.047 g) to give a yield
of 35%. ESI-MS (m/z): 421.42 [Ni2(L3)3]4+ 100%, 594.92 [Ni2(L3)3(-
ClO4)]3+ 71%, 941.25 [Ni2(L3)3(ClO4)2]2+ 17%; IR (KBr cmꢂ1):
m
= 3051 (w), 2916 (w), 2850 (w), 1631 (m), 1581 (s), 1498 (vs),
m
= 3070 (w), 2968 (w), 2887 (w), 1629 (w), 1597 (w), 1500 (w),
1363 (w), 1261 (w), 1086 (s), 1020 (w), 802 (m), 625 (m); UV–
Vis (H2O/MeCN nm): 338 = 18,750), 247 = 31,250), 205
= 75,000); Anal. Calcd (%) for Ni2C108N12O16H102Cl4: C 62.27, H
1479 (vs), 1347 (w), 1269 (w), 1228 (vs), 1126 (w), 962 (w), 881
(w), 837 (m), 775 (m); Anal. Calcd (%) for C30N4H22O2: C 76.58, H
4.71, N 11.91, Found: C 76.63, H 4.65, N 11.84.
(
e
(e
(e
4.93, N 8.07, Found: C 62.32, H 4.97, N 8.12.
Synthesis of L2
The ligand L2 was prepared from 1,3-bis[1-(4-aminophenyl)-1-
methylethyl]benzene (0.826 g, 2.4 mmol) and 2-pyridinecarboxal-
dehyde (0.512 g, 4.8 mmol) by the method described above for L1,
and was obtained as light yellow oil. Yield, (70%). 1H NMR
(400 MHz, CDCl3 d ppm): 8.702 (d, J = 4.4 Hz, 2H, H2) 8.640 (s,
2H, HAC@N), 8.202 (d, J = 8.0 Hz, 2H, H5), 7.792 (t, 2H, H4), 7.356
(t, 2H, H3), 7.283–7.216 (m, 9H, Ha,b,d), 7.147 (s, 1H, Hc), 7.116–
Molecular mechanics modelling studies
Molecular mechanics calculations on [Ni2(L1)3]4+, [Ni2(L2)3]4+
and [Ni2(L3)3]4+ were carried out with Hyperchem Version 7.52
[19]. The modelling studies were limited to use MM+ with the Po-
lak-Ribiere algorithm of Hyperchem and the energy minimized at
an RMS gradient of 0.01. Molecular dynamics was also used (sim-
ulated heating to 3000 K) to make sure the true energy minima had
been reached. The original rac- and meso-configurations of com-
plexes 1–3 were arranged roughly by eye and guided by the ESI-
MS, UV–Vis and IR spectroscopic data.
7.097 (t, 2H, He), 1.683 (s, 12H, CH3); IR (KBr cmꢂ1):
m = 3057
(w), 2968 (s), 2870 (w), 1628 (w), 1583 (m), 1566 (w), 1500 (vs),
1465 (s), 1435 (m), 1361 (w), 1205 (w), 1178 (w), 993 (w), 883
(w), 833(s), 775 (m); Anal. Calcd (%) for C36N4H34: C 82.72, H
6.56, N 10.72, Found: C 82.77, H 6.49, N 10.68.
DNA-binding studies
Synthesis of L3
The ligand L3 was prepared from 1,4-bis[1-(4-aminophenyl)-1-
methylethyl]benzene (0.826 g, 2.4 mmol) and 2-pyridinecarboxal-
dehyde (0.512 g, 4.8 mmol) by the method described above for L1,
and was obtained as light yellow powder. Yield, (73%). 1H NMR
(400 MHz, CDCl3 d ppm): 8.700 (d, J = 4.8 Hz, 2H, H2), 8.634 (s,
2H, HAC@N), 8.200 (d, J = 8.0 Hz, 2H, H5), 7.800 (t, 2H, H4), 7.352
(t, 2H, H3), 7.306–7.229 (m, 8H, Ha,b), 7.150 (s, 4H, Hc), 1.694 (s,
m = 3055 (w), 2962 (m), 2927 (w), 1630
(m), 1564 (w), 1504 (vs), 1469 (m), 1389 (w), 1207 (w), 1090
(w), 1016 (w), 883 (m), 838 (s), 779 (m); Anal. Calcd (%) for
Absorption spectroscopy
To determine the binding affinity of the complexes 1–3 with
HTG22 and ct-DNA, the complexes (4–20
DNA solutions (HTG22 in buffer B: from 0 to 1.26
DNA in buffer D: from 0 to 200 M). The same volume of DNA solu-
tion was added to each cuvette to eliminate the absorbance of DNA
itself, and the solutions were mixed by repeated inversion. A buffer
solution baseline was subtracted from each data set. After mixing
for 5 min, the absorption spectra were recorded. The changes in
the metal complex concentration due to dilution at the end of each
titration were negligible. The UV–Vis titrations for each sample
were repeated at least three times.
l
M) were titrated with
l
M, and ct-
l
12H, CH3); IR (KBr cmꢂ1):
C36N4H34: C 82.72, H 6.56, N 10.72, Found: C 82.79, H 6.50, N 10.71.
Synthesis of [Ni2(L1)3][ClO4]4 (1)
To a stirred suspension of ligand L1 (0.09 mmol, 0.042 g) in
10 mL acetonitrile, a MeCN solution (10 mL) of Ni(ClO4)2 6H2O
(0.06 mmol, 0.022 g) was added dropwise. The resultant orange
mixture was heated at reflux temperature for 5 h and allowed to
cool. The resulting orange solution was filtered through Celite
and the filtrate was evaporated to dryness under reduced pressure.
The residue was dissolved in 10 mL MeCN from which orange
microcrystalline powder was retrieved by filtration after a few
Circular dichroism spectroscopy
Circular dichroism (CD) titrations were carried out using a
MOS-450/AF-CD spectropolarimeter at room temperature with
fixed concentration constant at 2 lM of HTG22 DNA. By adding
the solutions of metal complex to the 0.5 cm pathlength cell, the
HTG22 DNA:metal complex ratios were 2:1, 1:1, 2:3, 2:4 and 2:6,
respectively. All solutions were mixed thoroughly and allowed to
equilibrate for 5 min before data collection. For each sample, the

24
X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
spectrum was scanned at least three times and accumulated over
the wavelength range. The instrument was flushed continuously
with pure evaporated nitrogen throughout the experiment. The
scan of the buffer alone was subtracted from the average scan for
each sample.
was minimized to 173.13, 181.69 and 183.46 kcal molꢂ1, which
corresponded to the structure of meso-[Ni2(L1)3]4+
,
rac-
[Ni2(L2)3]4+ and rac-[Ni2(L3)3]4+, respectively (Fig. 1). Complex 1
mainly exists in meso-configuration, while complexes 2 and 3
mainly exist in rac-configuration. As listed in Table 1, the sizes of
cylinders 1–3 were found slight larger than the traditional nicke-
4þ
l(II) cylinder Ni2ðLÞ3 (length ꢃ 18 Å, width ꢃ 8 Å) [21]. The cylin-
Fluorescence resonance energy transfer (FRET) studies
der 1 containing 1,3-bisphenoxylbenzene spacers is longer and
narrower than 2 and 3, because the O linked bridging ligand L1
is more rigid than L2 and L3 formed CH(CH3)2 bind and cannot
spiral easily. However, 2 and 3 with rac-configuration have larger
intracavities than 1 with meso-configuration. The structural simu-
lation information of complexes 1–3 was favorable to explain their
different DNA binding features.
The fluorescent-labeled oligonucleotide, F22T, used as the FRET
probes were diluted in buffer B solution and annealed after heating
to 95 °C for 5 min, and then incubated at 4 °C overnight. Fluores-
cence melting curves were determined by using a Bio-Rad iQ5
real-time PCR detection system. Different concentrations of com-
plexes were added into a total reaction volume of 20
lL with
0.2 M of labeled oligonucleotide. Fluorescence readings with exci-
l
tation were taken at intervals of 1 °C over the range 20–90 °C, with
a constant temperature being maintained for 30 s prior to each
reading to ensure a stable value. The melting of the G-quadruplex
was monitored alone or in the presence of various concentrations
of complexes. Various concentrations of competitors (double
stranded self-complementary ds26 DNA) were added to test the
binding selectivity of the compound to the quadruplex structure.
Final analysis of the data was carried out using Origin 7.5 software
(Origin Lab Corp.).
DNA binding experiments
Electronic absorption spectroscopy
UV–Vis spectrum was performed to determine the binding
affinities for complexes 1–3 toward quadruplex (HTG22) and du-
plex DNA (ct-DNA). As the successive addition of HTG22 to the
solution of complexes 1–3, both intraligand absorption band from
200 to 300 nm and metal ? ligand charge transfer (MLCT) at about
330 nm displayed significantly hypochromism (H%), indicating the
strong interaction between the dinuclear triple-stranded cylinders
and G-quadruplex DNA bases (Fig. 2). In general, the extent of hyp-
ochromism parallels the DNA-binding strength of the complex
Results and discussion
Synthesis and characterization
[22]. As addition of 1.26 lM HTG22 to 4 lM complexes 1, 2 and
3, the band at 205 nm exhibited the max hypochromism of
34.7%, 29.9% and 19.4%, respectively. This indicated that the com-
plex 1 bound more tightly to G-quadruplexes than complexes 2
and 3 did. In contrast, as the ratio of [ct-DNA]/[complex] increased
to 10, all three complexes demonstrated large hypochromicity
The dinuclear triple-stranded cylinders 1–3 were prepared by
reactions of Ni(ClO4)2 with the tetradentate ligands L1, L2 and L3,
respectively. The ligands and complexes were characterized by
various spectroscopic and analytical techniques. In the 1H NMR
spectra of the ligands L1–L3 (Fig. S1), HAC@N protons showed sin-
glet at about 8.63, the protons of pyridine appeared at 8.70–7.35,
and the peeks at 7.30–6.70 belonged to the protons of benzene.
In the IR spectra, the characteristic peaks near 1630 cmꢂ1 were as-
signed to the mC@N stretching vibration of the ligands. After coordi-
nating with nickel(II) ions, the characteristic mC@N peaks shifted to
lower wave numbers. Meanwhile, the peeks at 1088 cmꢂ1 and
630 cmꢂ1 revealed the existence of ClOꢂ4 , which confirmed metal
center involvement in bis(pyridylimine) Schiff base units. UV–Vis
spectra observed from 200 to 300 nm for 1–3 were due to in-ligand
(IL) transitions, while the peaks around 330 nm were characteristic
metal ? ligand charge transfer (MLCT) transitions for nickle(II)
complexes. The ESI-MS of complexes 1–3 in acetonitrile-methanol
solution revealed the presence of molecular ion peaks [Ni2(L1)3]4+
,
[Ni2(L2)3]4+ and [Ni2(L3)3]4+ with varying numbers of associated
ClOꢂ4 (Fig. S2). For example, the peaks of complex 1 at m/z 382,
542 and 863 corresponded to [(Ni2(L1)3)(ClO4)n](4ꢂn)+ (n = 0, 1, 2),
4þ
indicating that Ni2ðL1Þ3 is the most stable fragment in solution.
The peak values of the great abundant fragment ions observed in
ESI-MS spectra for complexes 1–3 were in agreement with the cal-
culated values (Table S1). From the above, three new dinuclear
nickel(II) triple-stranded supramolecular cylinders have been syn-
thesized successfully.
Molecular mechanics modelling studies
To further confirm the structures of the Ni(II) Schiff base
complexes 1–3, the molecular simulate modelling studies were
performed. The compounds we used, all have bimetallo triple-
stranded structures, in which two configurations are possible: a
rac isomer (the isomer is chiral and is one of a pair of enantiomers)
and a meso isomer rendering the structure achiral [20]. According
to the lowest energy principle, the energy of the cylinders 1–3
Fig. 1. The molecular models of tetracationic triple-stranded supramolecular
cylinders: (a) [Ni2(L1)3]4+, (b) [Ni2(L2)3]4+, and (c) [Ni2(L3)3]4+
.

X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
25
Table 1
Structural simulation data of complexes 1–3.
Complex
1
Configuration
Energy (kcal molꢂ1
)
Length (Å)
Diameter (Å)
rac-
meso-
177.14
173.13
22.38
22.67
9.42
9.00
2
3
rac-
meso-
181.69
193.28
19.69
23.37
13.43
12.47
rac-
meso-
183.46
190.59
19.26
22.73
14.86
13.41
quadruplex and duplex DNA of 1–3 could be attributed to the struc-
tural differentiation.
(>30%) at 205 nm suggesting 1–3 could also bind to duplex ct-DNA
(Fig. S3).
In order to compare quantitatively the binding strength of 1–3
to G4-DNA and ct-DNA, the intrinsic binding constants Kb with
DNA were obtained using the following equation [23]:
Circular dichroism spectroscopy
Circular dichroism (CD) spectroscopy is uniquely sensitive to
the absorption of the solution conformations of HTG22. Without
any metal cations in HTG22 solution, the DNA molecules not only
exist as a mixture of parallel and antiparallel G-quadruplex confor-
mations, but also are dissociated partially to single-stranded mol-
ecules [25]. With the addition of complex 1 to HTG22 aqueous
solution, the distinct spectral changes that the intensity of the po-
sitive CD band at 256 nm decreased gradually from positive to neg-
ative, and a remarkable increase of the CD band at 290 nm were
observed. (Fig. 3a-1) This induced CD spectrum suggested complex
1 could fold single-stranded human telomeric DNA to form a hy-
brid G-quadruplex structure under salt deficient conditions. How-
½Dꢄ=
D
eap ¼ ½Dꢄ=
D
e
þ 1=ðKb
ꢁ
D
e
Þ;
D
eap ¼ jea
ꢂ
ef j;
D
e
¼ jeb
ꢂ
ef
j
where [D] is the concentration of DNA in base pair, ea
the apparent extinction coefficient (Aobs/[complex]) and the extinc-
,
eb and ef are
tion coefficient for nickle(II) complex in the free and fully bound
form, respectively. Kb is the equilibrium binding constant in Mꢂ1
.
As listed in Table 2, the G4-DNA binding constants of 1–3 were
comparable to that of other compounds reported to be good G-
quadruplex DNA binders (106 Mꢂ1 range) [24]. Moreover,
complexes 1–3 displayed reasonable selectivity (ca. 2 orders of
magnitude) for quadruplex versus duplex DNA. The trend of G4-
DNA binding constant G4-DNAKb was 1 > 2 > 3 which was consistent
with the order of hypochromicity. However, when bound to ct-DNA,
the ct-DNAKb order was 3 > 2 > 1. The different binding affinities to
ever, addition of 1–10 lM 2 or 3 to HTG22 DNA in the absence of
metal ions solution, the positive CD band at 256 nm decreased
gradually but the peak close to 290 nm initially increased and then
slowly declined when the added complex was more than 6 lM
Fig. 2. UV–Vis spectra of complexes: (a) 1, (b) 2 and (c) 3 (4
added. Inset: plot of [D]/4eap versus [D].
lM) in buffer B with HTG22 (0–1.26 lM). The arrows indicate the change upon increasing amount of HTG22

26
X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
Table 2
Absorption spectra (kmax/nm) and DNA-binding data of complexes 1–3.
Complexes
1
kmax(free) (nm)
H (%)
G4-DNAKb (Mꢂ1
)
ct-DNAKb (Mꢂ1
)
331
282
241
205
21.6
25.3
19.6
34.7
2.2 ꢁ 106
2.0 ꢁ 106
1.1 ꢁ 105
4.0 ꢁ 103
2.0 ꢁ 104
2.3 ꢁ 104
2
3
329
245
205
18.4
22.6
29.9
338
247
205
12.6
22.9
19.4
(Fig. 3a-2 and a-3). The latter suggested that complexes 2 and 3
were able to interact with the single-strand human telomeric
DNA and partially induced it to form G-quadruplex structures.
The structures of G-quadruplexes were also investigated in the
Na+ or K+ buffer solution. Previous studies showed that human
telomeric G-quadruplex could form an antiparallel structure in
the presence of sodium or adopt a hybrid structure in the presence
of potassium [26,27]. With addition of complex 1 to HTG22 in Na+
buffer solution, major positive CD band significantly increased and
shifted from 295 to 290 nm, accompanied by a sharp decrease
around 260 and 245 nm with band shift. (Fig. 3b-1). The induced
CD spectra produced by complex 1 was similar to the findings of
[Ni2L3]4+ complex [12]. Moreover, after addition of complex 1 to
HTG22 in K+ buffer solution, no significant change in CD spectra
was found (Fig. 3c-1). These results indicated that complex 1 could
convert the antiparallel G-quadruplex into a hybrid-type G-quad-
ruplex in Na+ solution and stabilize the hybrid G-quadruplex struc-
ture in K+ solution. However, adding complexes 2 or 3 to HTG22 in
Na+ (Fig. 3b-2 and b-3) or K+ (Fig. 3c-2 and c-3) buffer solution, the
CD spectra exhibited a maxima-minima pattern, similar but not
identical to the original spectra without addition of the complex.
These CD phenomena of 2 and 3 were consistent with the reported
g
6-arene ruthenium complexes [28]. Similar conclusions were
drawn that complexes 2 and 3 disturbed DNA structure slightly
and the DNA secondary structure did not change even at high ionic
strength. The above CD studies clearly demonstrated that 1 bound
to G4-DNA more strongly than complexes 2 and 3, and only cylin-
der 1 of the three complexes could lead to the conversion of anti-
parallel G-quadruplexes to hybrid G-quadruplexes. To the best of
our knowledge, a very high level CD signal conversion is rare
among the molecules as G-quadruplexes stabilizers [12].
Thermodynamic stabilization of the telomeric G-quadruplex by
complexes
Fluorescence resonance energy transfer (FRET) was used as a
convenient method to provide information on the folded or
Fig. 3. CD titration spectra of HTG22 (2
lM) induced by the three compounds (from left to right: 1, 2 and 3) in 10 mM Tris–HCl, pH 7.4, at room temperature. (a) In the
absence of metal ions, (b) in 100 mM NaCl, and (c) in 100 mM KCl.

X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
27
unfolded state of the quadruplex by monitoring the 30- to -50 end
distance [29]. The effect on the thermal stabilities of the cylinders
1–3 upon adding G4-DNA was explored by UV–Vis absorbance at
205 nm. The results (Fig. S4) showed that all cylinders 1–3 were
stabilized in the presence of the G4-DNA at high temperature.
We performed the FRET melting assay to investigate thermody-
namic stability of complexes 1–3 to G-quadruplex DNA F22T (se-
quence: FAM-AG3(T2AG3)3-TAMRA, mimicking the human
telomeric repeat). As shown in Fig. 4a–c, all the three complexes
1–3 significantly increased the melting temperature of the F22T
with [complex]/[DNA] ratios = 1:2, 1.25:1, 2.5:1, 3.75:1 and 5:1,
(
D
Tm = 33.2 °C) [10]. This was probably because the planar
arrangement of the salphen rings was able to interact with the
G-quadruplexes through strong stacking.
p-p
A competition FRET experiment of the telomeric sequence F22T
versus non-fluorescent duplex DNA (ds26) was assessed to further
determine the G-quadruplexes selectivity of cylinders 1–3 (Fig. 5).
The competition FRET experiment was processed by maintaining
the concentration of complex (1.0 lM) and F22T (0.2 lM) as con-
stant with varying the ratio of [ds26]:[F22T] from 0 to 50. With
the [ds26]:[F22T] ratio of 10, the thermal stabilization of F22T en-
hanced by 1–3 was slightly affected. Even at the addition of a 50-
while ligands L1–L3 had no effect on G4-DNA with
(Fig. S5). At the ratio of 5, complex 1 induced a high degree of sta-
bilization for G-quadruplex DNA, as demonstrated by an increase
D
Tm ꢅ 0 °C
fold excess of ds26 DNA, the enough high DTm values of F22T for
1–3 were still observed in Fig. 5d. This means that complexes 1–
3 can stabilize the G-quadruplex even with the addition of sub-
stantial amounts of competitive ds26 DNA. The results of FRET
competition assay demonstrated that complexes 1–3 can be con-
sidered as a new class of G-quadruplex DNA binders, and 1 was
found to have stronger preference for binding to the G-quadruplex
over duplex DNA than 2 or 3 did. However, the selectivity (at least
10-fold) of G-quadruplex versus duplex DNA for complexes 1–3
was lower than that of nickel salphen complexes with cyclic amine
side arms (50-fold) [10]. It’s probable that such cylinders with
proper width and length are suitable for the configuration of the
major groove of duplex DNA helix and have a bending effect on du-
plex DNA [30].
in melting temperature (
DTm) of 24.5 °C (Fig. 4d). Compared with
1, complexes 2 and 3 ( Tm = 17.3 °C and 10.8 °C respectively) were
D
less effective G-quadruplex DNA stabilizers, which was consistent
with UV and CD studies. Moreover, the FRET-melting data con-
firmed that 1–3 were better G-quadruplexes binders than the chi-
ral nickel(II) cylinder P-Ni2L4þ
(D
Tm = 10.6 °C) [12]. The changes in
3
size of the cylinders 1–3 might afford larger molecular surfaces
which were favorable for the stacking through the extensive
hydrophobic exterior of the cylinders on the top face of the G-
quadruplexes. Nevertheless, the
DTm values for cylinders 1–3 were
lower than the reported planar nickel(II) salphen complexes
Fig. 4. FRET melting curves for experiments carried out with F22T alone and separately with (a) 1, (b) 2 and (c) 3. F22T concentration was 0.2 lM in Buffer B. (d) Plot of DNA
stabilization temperature versus the concentration of 1 (black), 2 (red) and 3 (green) binding to F22T. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)

28
X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
Fig. 5. Competitive FRET melting curves of F22T (0.2
lm) with 1.0 lm of (a) 1, (b) 2, (c) 3 and duplex competitor ds26 in Buffer B. (d) Competition FRET experiment of
complexes for the G-quadruplex DNA sequence over duplex DNA. Results are the mean values of at least three independent experiments.
Possible G4-DNA binding behavior
likely as a whole to stack on the tail of G-quadruplexes. Such
The sizes of synthetic dinuclear nickel(II) triple-stranded supra-
molecular cylinders 1–3 were calculated by molecular mechanics
modelling studies (Table 1), which were compatible with G-quar-
tet (length ꢃ 14 Å, width ꢃ 11 Å). From the point of view of
structures, it is difficult for cylinders 1–3 to be embedded in a
end-stacking model of compounds onto the G-quartet is in agree-
ment with the previous reported cylindrical quadruplex binders
[12,13]. The precise size and shape of the cylinders are crucial to
determine such binding sites and modes to G-quadruplexes. In
addition, four positive charges schlepped by cylinders 1–3 enhance
the electrostatic interactions between the metal complex fragment
and DNA’s phosphate backbone. Moreover, the large hydrophobic
exterior of 1–3 may also enhance the non-covalent interactions be-
tween the complexes and grooves or loops of quadruplexes. The
details of the binding modes are not clear yet and deserve further
investigation. Fig. 6 is the representative illustration of a proposed
model of complex 1 binding to human telomeric DNA (PDB code
1KF1). It is the largest length (ꢃ22 Å) and suitable width (ꢃ9 Å)
in the three cylinders that lead complex 1 to stacking most
strongly on the top of a terminal G-quartet. However, the larger
width (ꢃ14 Å) of complexes 2 and 3 slightly exceed the recognition
surface of G-quadruplex DNA (width ꢃ 11 Å) and this may reduce
the binding affinity of 2 and 3 to G-quartet. This is consistent with
the conclusion of experiential results that complex 1 has the most
excellent binding affinity to G-quadruplexes in the three
complexes.
p-stacking unit to the G-tetrads. Combining with UV and CD spec-
tra studies, we can speculate that the charged cylinder is most
Conclusion
In summary, three dinuclear nickel(II) triple-stranded supramo-
lecular cylinders 1–3 with different bis(pyridylimine) Schiff base
ligands have been successfully prepared to study their
Fig. 6. Predicted interaction between 1 (shown as a ball model) and the intermo-
lecular G-quadruplex (shown as a stick model).

X.-X. Xu et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 124 (2014) 21–29
29
(c) M. Micco, G.W. Collie, A.G. Dale, S.A. Ohnmacht, I. Pazitna, M. Gunaratnam,
A.P. Reszka, S. Neidle, J. Med. Chem. 56 (2013) 2959–2974.
[9] (a) D.P.N. Gonçalves, R. Rodriguez, S. Balasubramanian, J.K.M. Sanders, Chem.
Commun. 45 (2006) 4685–4687;
(b) S.N. Georgiades, N.H. Abd Karim, K. Suntharalingam, R. Vilar, Angew. Chem.
Int. Ed. 49 (2010) 4020–4034.
[10] (a) J.E. Reed, A.A. Arnal, S. Neidle, R. Vilar, J. Am. Chem. Soc. 128 (2006) 5992–
5993;
(b) A.A. Arnal, J. Benet-Buchholz, S. Neidle, R. Vilar, Inorg. Chem. 47 (2008)
11910–11919.
[11] (a) R. Kieltyka, P. Englebienne, J. Fakhoury, C. Autexier, N. Moitessier, H.F.
Sleiman, J. Am. Chem. Soc. 130 (2008) 10040–10041;
(b) S. Shi, J. Liu, T. Yao, X. Geng, L. Jiang, Q. Yang, L. Cheng, L. Ji, Inorg. Chem. 47
(2008) 2910–2912;
G-quadruplexes binding properties. The three novel nickel(II) cyl-
inders were good G-quadruplex DNA stabilizers and showed differ-
ent features on DNA binding affinity. The suitable size and four
positive charges of the complexes are the major factors determin-
ing the high efficiency of binding G4-DNA. Such supramolecular
cylinders definitely represent a new and exciting strategy on
designing new antitumor agents for targeting quadruplexes DNA.
Other novel supramolecular metallaclusters as G-quadruplexes
inducers and stabilizers are currently being investigated in our
laboratory.
(c) Y.-L. Wu, K.E. Brown, M.R. Wasielewski, J. Am. Chem. Soc. 135 (2013)
13322–13325.
Acknowledgements
[12] H.J. Yu, X.H. Wang, M.L. Fu, J.S. Ren, X.G. Qu, Nucleic Acids Res. 36 (2008)
5695–5703.
[13] C.Q. Zhao, J. Geng, L.Y. Feng, J.S. Ren, X.G. Qu, Chem. Eur. J. 17 (2011) 8209–
8215.
[14] (a) M.J. Maroney, Curr. Opin. Chem. Biol. 3 (1999) 188–199;
(b) P.E. Carrington, P.T. Chivers, F. Al-Mjeni, R.T. Sauer, M.J. Maroney, Nat.
Struct. Biol. 10 (2003) 126–130.
[15] (a) S. Sarkar, A. Datta, A. Mondal, D. Chopra, J. Ribas, K.K. Rajak, S.M. Sairam,
S.K. Pati, J. Phys. Chem. B110 (2006) 12–15;
This work was supported by the National Natural Science Foun-
dation of China (21101078 and 21276105), the Program for New
Century Excellent Talents in University of China (NCET-11-0657),
the Natural Science Foundation of Jiangsu Province (BK2011143),
the Fundamental Research Funds for the Central Universities (JUS-
RP51314B), and the Ministry of Education and the State Adminis-
tration of Foreign Experts Affairs for the 111 Project (B13025).
(b) N.S. John, G.U. Kulkarni, A. Datta, S.K. Pati, F. Komori, G. Kavitha, C.
Narayana, M.K. Sanyal, J. Phys. Chem. C 111 (2007) 1868–1870;
(c) A. Datta, N.S. John, G.U. Kulkarni, S.K. Pati, J. Phys. Chem. A 109 (2005)
11647–11649.
Appendix A. Supplementary material
[16] X.D. Dong, X.Y. Wang, Y.F. He, Z. Yu, M.X. Lin, C.L. Zhang, J. Wang, Y.J. Song, Y.M.
Zhang, Z.P. Liu, Y.Z. Li, Z.J. Guo, Chem. Eur. J. 16 (2010) 14181–14189.
[17] M.E. Reichmann, S.A. Rice, C.A. Thomas, P. Doty, J. Am. Chem. Soc. 76 (1954)
3047–3053.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.saa.2013.12.102.
[18] J. Marmur, P. Doty, J. Mol. Biol. 3 (1961) 585–594.
[19] S. Goetz, P.E. Kruger, Dalton Trans. 35 (2006) 1277–1284.
[20] C. Uerpmann, J. Malina, M. Pascu, G.J. Clarkson, V. Moreno, A. Rodger, A.
Grandas, M.J. Hannon, Chem. Eur. J. 11 (2005) 1750–1756.
[21] (a) L.J. Childs, M.J. Hannon, Supramol. Chem. 16 (2004) 7–22;
(b) M.J. Hannon, I. Meistermann, C.J. Isaac, C. Blomme, J. Aldrich-Wright, A.
Rodger, Chem. Commun. (2001) 1078–1079.
[22] C. Wei, G. Jia, J. Yuan, Z. Feng, C. Li, Biochemistry 45 (2006) 6681–6691.
[23] A. Wolfe, G.H. Shimer Jr., T. Meehan, Biochemistry 26 (1987) 6392–6396.
[24] (a) R. Kieltyka, J. Fakhoury, N. Moitessier, H.F. Sleiman, Chem. Eur. J. 14 (2008)
1145–1154;
(b) K. Suntharalingam, A.J.P. White, R. Vilar, Inorg. Chem. 49 (2010) 8371–
8380.
[25] Y. Xu, Y. Noguchi, H. Sugiyama, Bioorg. Med. Chem. 14 (2006) 5584–5591.
[26] J.S. Ren, X.G. Qu, J.O. Trent, J.B. Chaires, Nucleic Acids Res. 30 (2002) 2307–
2315.
References
[1] (a) J.T. Davis, Angew. Chem. Int. Ed. 43 (2004) 668–698;
(b) H.Z. He, D.S.H. Chan, C.H. Leung, D.L. Ma, Nucleic Acids Res. 41 (2013)
4345–4359.
[2] (a) S. Burge, G.N. Parkinson, P. Hazel, A.K. Todd, S. Neidle, Nucleic Acids Res. 34
(2006) 5402–5415;
(b) O. Doluca, J.M. Withers, V.V. Filichev, Chem. Rev. 113 (2013) 3044–3083.
[3] P. Murat, Y. Singh, E. Defrancq, Chem. Soc. Rev. 40 (2011) 5293–5307.
[4] (a) W.D. Wilson, H. Sugiyama, ACS Chem. Biol. 2 (2007) 589–594;
(b) D.S. Pilch, C.M. Barbieri, S.G. Rzuczek, E.J. Lavoie, J.E. Rice, Biochimie 90
(2008) 1233–1249.
[5] (a) S. Balasubramanian, S. Neidle, Curr. Opin. Chem. Biol. 13 (2009) 345–353;
(b) A. De Cian, L. Lacroix, C. Douarre, N. Temime-Smaali, C. Trentesaux, J.-F.
Riou, J.-L. Mergny, Biochimie 90 (2008) 131–155.
[6] T.M. Ou, Y.J. Lu, J.H. Tan, Z.S. Huang, K.Y. Wong, L.Q. Gu, Chem. Med. Chem. 3
(2008) 690–713.
[7] (a) A.K. Jissy, A. Datta, ChemPhysChem 13 (2012) 4163–4172;
(b) A.K. Jissy, U.P.M. Ashik, A. Datta, J. Phys. Chem. C 115 (2011) 12530–12536.
[8] (a) A. Arola, R. Vilar, Curr. Top. Med. Chem. 8 (2008) 1405–1415;
(b) D. Monchaud, M.-P. Teulade-Fichou, Org. Biomol. Chem. 6 (2008) 627–636;
[27] J. Dai, M. Carver, C. Punchihewa, R.A. Jones, D. Yang, Nucleic Acids Res. 35
(2007) 4927–4940.
[28] D.D. Sun, R. Zhang, F. Yuan, D. Liu, Y.H. Zhou, J. Liu, Dalton Trans. 41 (2012)
1734–1741.
[29] G.W. Collie, G.N. Parkinson, Chem. Soc. Rev. 40 (2011) 5867–5892.
[30] M.J. Hannon, Chem. Soc. Rev. 36 (2007) 280–295.