From Cyclen to 12-Crown-4 Copper(II) Complexes: Exchange of Donor Atoms Improves DNA Cleavage Activity

Article abstract of DOI:10.1002/ejic.201500596

Macrocyclic Cu^II complexes with [N_XO_Y] donor sets of different N/O ratios were synthesised resulting in a series ranging from cyclen (X = 4, Y = 0: 1) to 12-crown-4 (X = 0, Y = 4: 6) complexes. In order to elucidate the structure of the complexes UV/Vis spectroscopy and X-ray crystallography were applied, focusing especially on the literature-unknown compounds with regioisomeric [N₂O₂] and [NONO] donor sets (3, 4). The complexes were subjected to DNA cleavage experiments under reducing conditions and were also tested in the absence of a reducing agent. Although 3 and 6 were the most active DNA cleavers in the presence of reducing ascorbate, both of them and 4 also cleaved DNA (not hydrolytically) in its absence. This brings up questions regarding the cleavage mechanism. The present study is an expansion of our previously reported finding that heterosubstitution in macrocyclic ligands leads to changes in oxidative DNA cleavage activity of Cu^II complexes.

Full text of DOI:10.1002/ejic.201500596

FULL PAPER
DOI:10.1002/ejic.201500596
From Cyclen to 12-Crown-4 Copper(II) Complexes:
Exchange of Donor Atoms Improves DNA Cleavage
Activity
Jan Hormann,[a] Margarethe van der Meer,[a] Biprajit Sarkar,[a] and
Nora Kulak*[a]
Keywords: DNA cleavage / Copper / Macrocyclic ligands / Cyclen / Cyclic voltammetry
Macrocyclic CuII complexes with [NXOY] donor sets of dif-
ferent N/O ratios were synthesised resulting in a series rang-
ing from cyclen (X = 4, Y = 0: 1) to 12-crown-4 (X = 0, Y = 4:
6) complexes. In order to elucidate the structure of the com-
plexes UV/Vis spectroscopy and X-ray crystallography were
applied, focusing especially on the literature-unknown com-
pounds with regioisomeric [N2O2] and [NONO] donor sets (3,
absence of a reducing agent. Although 3 and 6 were the most
active DNA cleavers in the presence of reducing ascorbate,
both of them and 4 also cleaved DNA (not hydrolytically) in
its absence. This brings up questions regarding the cleavage
mechanism. The present study is an expansion of our pre-
viously reported finding that heterosubstitution in macro-
cyclic ligands leads to changes in oxidative DNA cleavage
4). The complexes were subjected to DNA cleavage experi- activity of CuII complexes.
ments under reducing conditions and were also tested in the
DNA in a hydrolytic (like enzymes) or oxidative manner.[5,6]
Introduction
While hydrolytic nucleases cleave phosphor–oxygen bonds
of the phosphate backbone, oxidative DNA cleavers gener-
ate reactive oxygen species (ROS) that degrade DNA by
attacking the sugar and base moieties. Due to the irrevers-
ibility of the oxidative cleavage these nucleases are promis-
ing candidates for future chemotherapeutics that could
overcome the problem of Pt anticancer drug resistance. The
topic of hydrolytic and oxidative artificial nucleases has
been reviewed several times in the past few years.[6–11]
A commonly used ligand in bioinorganic and medicinal
chemistry is 1,4,7,10-tetraazacyclododecane, the so called
cyclen or [12]aneN4 ligand. It coordinates to transition
metal atoms and rare earth metal atoms and can readily
be substituted due to its secondary amine functionalities.
“Artificial nucleases” based on the cyclen ligand and cop-
per(II) are extensively used for the design of oxidative DNA
cleavers. On the other hand complexes of redox-inert metals
as zinc(II) and cobalt(III) are employed to hydrolytically
cleave DNA.[6,7,12–15] Recent examples of such (cyclen)CuII-
based systems involve complexes that are further derivatised
with DNA-affine moieties such as DNA intercalators, pept-
ide nucleic acids or positively charged groups.[15–20] Other
approaches are centered on the synthesis of multinuclear
systems.[21,22]
According to the world health organisation (WHO) can-
cer is among the leading causes of death worldwide. In
2012, 14.1 million new cases of cancer (excluding non-mela-
noma skin cancer) were reported, and cancer is estimated
to have caused the death of 8.2 million people.[1] The treat-
ment of cancer remains a key challenge. The effectiveness
of even advanced chemotherapeutics is restricted by dose-
limiting side effects. This and the increasing resistance of
some cancer types against chemotherapeutics like cisplatin
call for the development of novel drugs.[2] Among the candi-
dates of such new drugs are metallonucleases, which are
able to cause oxidative stress and promote oxidative cleav-
age of DNA. This may harm cells due to programmed cell
death, e.g. apoptosis.[3]
Under physiological conditions the half-life of DNA is
estimated to be between ten and one hundred billion
years.[4] This stability relies mainly on the repulsion of pos-
sible nucleophiles by its negatively charged phosphate back-
bone. This is likely the reason as to why DNA evolved into
the carrier of the genetic code. Nevertheless, there are sev-
eral enzymes that can cleave DNA. Restriction enzymes
have inspired chemists over the last decades to develop so
called “artificial nucleases” that are also able to cleave
In contrast to the above-mentioned cyclen-based sys-
tems, whose cutting activity is increased by conjugating
DNA-affine moieties to the cyclen ligand, we achieved an
increase in oxidative cleavage activity by simply exchanging
one of the four donor atoms with either a sulfur or an oxy-
[a] Institut für Chemie und Biochemie, Freie Universität Berlin,
Fabeckstr. 34/36, 14195 Berlin, Germany
E-mail: nora.kulak@fu-berlin.de
www.bcp.fu-berlin.de/en/chemie/kulak
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201500596.
Eur. J. Inorg. Chem. 2015, 4722–4730
4722
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 1. Complexes 1–6, synthesised by gradually exchanging cyclen donor atoms by oxygen atoms.
gen atom. We found that the reactivity strongly depends on copper(II) nitrate as a metal source. The synthesis of com-
the heteroatom inserted into the macrocycle. An increase of plex 5, however, was not successful despite variation of sol-
oxidative DNA cutting activity was observed going from vent, metal source and ligand/metal stoichiometry. In the
nitrogen through sulfur to oxygen.[23]
literature we found only one example for complexes with
Herein, we report the influence of exchanging several L5: In 1995, Bourson and co-workers described coumarin-
nitrogen donor atoms of cyclen with oxygen atoms on the substituted [12]aneNO3 complexes of the MxLy type. They
solid-phase structure, nuclease activity and electrochemical also observed that the synthesis of complexes with M being
properties of its CuII complexes. Based on (cyclen)CuII (1), copper and x = y = 1 was not successful.[30] While the cop-
we synthesised the (oxacyclen)CuII complex (2), the two re- per complex 6[31] is already known in the literature, to the
gioisomeric (dioxacyclen)CuII complexes Cu([12]aneN2O2) best of our knowledge we are the first to describe copper
(3) and Cu([12]aneNONO) (4) and the all oxygen-substi- nitrato complexes 3 and 4.
tuted (12-crown-4)CuII complex Cu([12]aneO4) (6) (Fig-
ure 1). The [12]aneNO3 analogue 5 could not be isolated.
Complexes 1–4 and 6 were characterised by elemental
analysis, IR spectroscopy, mass spectrometry and UV/Vis
spectroscopy. Absorption spectra are presented in Figure 2,
the maximum absorption wavelengths and the correspond-
ing extinction coefficients are shown in Table 1. X-ray crys-
tal structures of the so far unknown nitrato complexes 3
and 4 were also obtained (Figure 3, Table 2).
Results and Discussion
(a) Synthesis, Characterisation and Structure
Cu([12]aneN4) (1) shows a maximum absorption band at
ca. 600 nm, which is consistent with the literature value.[32]
As extinction coefficients are relatively small, these absorp-
tions are probably due to the Laporte-forbidden d-d transi-
tions. Going from complex 1 to complex 6 a bathochromic
shift is observed in the maximum absorption wavelengths.
The reason for this is the exchange of the N atoms by the
more electronegative O atoms. With increasing electronega-
The synthesis of the macrocyclic ligands L1, L2, L4 and
L5 was performed as previously reported.[23–25]
L3 was synthesised by starting from tosylated ethylenedi-
amine and tosylated tetraethylene glycol by cyclisation in
DMF under high-dilution conditions using NaH as a base.
The deprotection step was conducted in hydrogen bromide
saturated acetic acid solution. Phenol was added to the re-
action mixture to improve the ligand solubility and to pre-
vent the ligand from being attacked by bromide ions
(Scheme 1).[28,29] Complexes 1–4 and 6 were readily access-
ible by reaction of the respective ligand in methanol with
Figure 2. Absorption spectra of complexes 1–4 and 6 (5 mm) in
50 mm Tris-HCl buffer (pH 7.4).
Table 1. Maximum absorption wavelength λmax with corresponding
extinction coefficients ε of complexes 1–4 and 6.
λmax [nm]
ε [Lmol–1 cm–1]
1
2
3
4
6
601
705
620
638
670
235.9
120.5
57.5
39.7
23.0
Scheme 1. Synthesis of complex 3. (a) TsCl, pyridine, 0 °C Ǟ room
temp., 16 h;[26] (b) TsCl, KOH, DCM, 0 °C, 3 h;[27] (c) NaH, DMF,
80 °C, 3 h; (d) phenol, 33% HBr in acetic acid, 80 °C, 3 d; (e)
Cu(NO3)2·3H2O, MeOH, 50 min reflux.
Eur. J. Inorg. Chem. 2015, 4722–4730
4723
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 2. Crystallographic and structure refinement data.
3
4
Name
Cu([12]aneN2O2)
C8H18CuN4O8
361.80
Cu([12]aneNONO)
C8H18CuN4O8
361.80
Empirical formula
Mr [gmol–1
]
Appearance
Crystal system
Space group
a [Å]
b [Å]
c [Å]
dark blue crystals
monoclinic
P21/c
7.665(3)
13.662(5)
13.392(4)
102.110(7)
1371.2(8)
4
light blue prisms
orthorhombic
Pca21
15.637(4)
11.872(3)
14.502(4)
90
2692.2(11)
8
1.785
β [°]
V [Å3]
Z
ρcalcd. [gcm–3
T [K]
]
1.753
133(2)
100(2)
θ range [°]
F(000)
2.15–26.41
747.8
1.7–30.5
1495.6
No. of parameters
μ [mm–1
No. of reflections
198
1.640
16307
395
1.671
17286
]
No. of total reflections 2813 (Rint = 0.0406) 7792 (Rint = 0.0414)
R1/wR2 (all data)
R1/wR2 [I Ͼ 2σ(I)]
Goodness of fit
0.0510/0.0823
0.0339/0.0748
1.110
0.0807/0.1317
0.0545/0.1192
0.987
tivity of donor atoms the energy of d-d transitions decreases
(i.e. longer wavelength). The rupture of this order by
Cu([12]aneN3O) (2), might be due to changes in the coordi-
nation environment (Table 3 and Figure 4).
Complexes 1–4 and 6 can be classified into three main
coordination geometries (Figure 4).
+
Cu([12]aneN4) (1) belongs to type I: Cu([12]aneN4)NO3
cations with square-pyramidal coordination environment
–
and NO3 counterions.[34] Cu([12]aneN3O) (2) and Cu([12]-
Figure 3. ORTEP[33] diagrams of the molecular structure of 3 (top)
and 4 (bottom). Thermal ellipsoids are drawn at a probability level
of 50%. The counterion of 4 and H atoms are omitted for clarity.
aneNONO) (4) can be classified as type II, which also com-
prises NO3 anions and CuLNO3 cations, but in contrast
–
+
Table 3. Selected bond lengths [Å] and angles [°] of complexes 1–4 and 6, X = N or O.
c.e.[a]
1
2
3
4
6
type I
type II
type III
type II
type III
Cu(1)–O(14)
Cu(1)–O(15)
Cu(1)–O(18)
Cu(1)–X(1)
Cu(1)–X(4)
Cu(1)–X(7)
Cu(1)–X(10)
C(2)–C(3)
2.183(4)
–
–
2.022(5)
2.023(5)
2.019(5)
2.001(6)
1.484(1)
1.251(5)
–
2.0312(1)
2.470
–
2.2310(1)
2.0224(2)
1.9837(2)
2.0267(1)
1.518(2)
1.2955(2)
–
2.026(2)
2.0316(1)
2.447
–
2.2470(1) O(1)
1.9875(2) O(4)
1.9893(1) O(7)
1.9915(2) O(10)
1.505(2)
1.235(3)
–
57.01
98.62(6) O(4)
162.07(5) O(7)
95.21(6) O(10)
1.943(4)
–
1.9842(2)
2.478(2)
2.766(2)
2.045(3)
2.009(2)
1.495(4)
1.293(3)
90.78(8)
–
128.96(7)
156.68(9)
89.65(9)
63.30(6)
114.62(9)
85.86(7)
117.09(9)
85.35(10)
1.965(5)
2.246(4)
2.040(4)
2.387(5)
2.125(4)
1.49(1)
1.297(6)
92.0(2)
–
N(1)
N(4)
N(7)
N(10)
O(1)
N(4)
N(7)
N(10)
O(1)
O(4)
N(7)
N(10)
O(1)
N(4)
O(7)
N(10)
O(14)–N(13)
O(14)–Cu(1)–O(18)
O(14)–Cu(1)–O(15)
X(4)–Cu(1)–O(14)
X(7)–Cu(1)–O(14)
X(10)–Cu(1)-(O14)
X(1)–Cu(1)–X(4)
X(1)–Cu(1)–X(7)
X(1)–Cu(1)–O(14)
X(4)–Cu(1)–X(10)
X(7)–Cu(1)–X(10)
–
56.72
N(4)
N(7)
N(10)
N(1), N(4)
N(1), N(7)
N(1)
N(4), N(10)
N(7), N(10)
105.9(2)
100.6(2)
104.9(2)
84.7(2)
151.6(2)
107.8(2)
149.1(2)
87.1(2)
N(4)
N(7)
N(10)
O(1), N(4)
O(1), N(7)
O(1)
N(4), N(10)
N(7), N(10)
95.04(5)
152.88(5)
99.78(6)
82.29(5)
110.88(5)
96.15(4)
159.40(6)
86.54(6)
O(4)
N(7)
N(10)
O(1), O(4)
O(1), N(7)
O(1)
O(4), N(10)
N(7), N(10)
N(4)
O(7)
N(10)
162.1(2)
125.0(2)
88.6(2)
O(1), N(4) 81.39
O(1), O(7) 106.07(5) O(1), O(7) 136.9(2)
O(1) 91.84(5) O(1) 87.1(2)
N(4), N(10) 158.24(6) O(4), O(10) 90.2(2)
O(7), N(10) 85.50(6) O(7), O(10) 75.4(2)
O(1), O(4) 75.1(2)
Reference
Murray–Rust[34]
Kulak[23]
this work
this work
Song[31]
[a] c.e.: coordination environment, cf. Figure 4.
Eur. J. Inorg. Chem. 2015, 4722–4730
4724
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
cleavage activity of Cu([12]aneO4) complex 6 is comparable
to that of complex 3, the regioisomer of 3, Cu([12]-
aneNONO) complex 4, cleaves in the range of complex 2.
Although the classification is based on solid-state struc-
tures, it is striking that the most active complexes, 3 and 6,
both belong to the type III coordination class (Figure 4),
whereas mediocre activity is shown by type II complexes 2
and 4. The lowest activity is exhibited by complex 1, a
type I complex.
Figure 4. Representation of possible solid-state coordination envi-
ronments.
to type I shows an anisodentate η2 chelation by the nitrato
ligand. Cu([12]aneN2O2) (3) and Cu([12]aneO4) (6)
(type III) are neutral complexes made up of the cyclen de-
rivative and two nitrato ligands which in turn coordinate in
η1 fashion.
Complex 2 and complex 4 employ a square-pyramidal
coordination sphere with additional anisodentate coordina-
tion by an O atom of the nitrato ligand. The bottom of the
pyramid is made up of three heteroatoms of the macrocyclic
ligand and one nitrato O atom. The top of the pyramid is
built by the O(1) atom of the macrocyclic ligands.
According to Song et al. the CuII ion of complex 6 is
coordinated in a distorted octahedral fashion, resulting in
a distortion of the ligand’s C4 symmetry and an elongation
of the Cu–O(1) and Cu–O(7) bonds [2.246(4) Å, 2.387(5) Å;
Table 3].[31] Similar observations can be made for complex
3: here the Cu–O(1) and Cu–O(4) bonds are elongated
[2.478(2) Å, 2.766(2) Å], leaving the CuII ion in a distorted
square-pyramidal coordination sphere rather than in an oc-
tahedral one. In contrast, all other Cu–N/O bonds are ca.
2 Å long, and the N(7)–Cu(1)–N(10), N(10)–Cu(1)–O(14),
O(14)–Cu(1)–O(18) and O(18)–Cu(1)–N(10) angles are al-
most 90°.
Figure 5. Cleavage activities of complexes 1–4 and 6 (0.04 mm) on
pBR322 (0.025 μgμL–1) in Tris-HCl buffer (50 mm, pH 7.4) and
ascorbic acid (0.32 mm) at 37 °C for 2 h. Illustrated is the average
of four experiments, the standard deviations are shown as error
bars.
In Figure 6 the dependence between cleavage activity and
complex concentration is presented. As all complexes but
Cu([12]aneN4) (1) readily convert plasmid DNA into linear
form III DNA, the linear form was employed to study the
concentration dependence of complexes 2–4 and 6 and their
cleavage activity. For a detailed comparison of the concen-
tration dependence by using form II and form III DNA in-
cluding complex 1, please refer to the Supporting Infor-
(b) DNA Cleavage in the Presence of Ascorbate
Oxidative cleavage of pBR322 supercoiled plasmid DNA
by complexes 1–4 and 6 was studied under approximate
physiological conditions at 37 °C in Tris-HCl buffer
(50 mm) with a pH value of 7.4. Ascorbic acid was chosen
as a reducing agent in a 0.32 mm concentration, since its
intracellular concentration is also in this range.[35] Agarose
gel electrophoresis (1% agarose) was used to monitor the
conversion of supercoiled (I) pBR322 into its nicked (II)
and linear (III) form. Cleavage activity of ligands without
any metal atom could be excluded (cf. Supporting Infor-
mation S-2,6). Experiments were carried out at least three
times in order to generate reliable data with standard devia-
tion.
Figure 5 shows a comparison of DNA cleavage activities
between complexes 1–4 and 6. As reported before, the ex-
change of only one donor atom (N Ǟ O) results in triplica-
tion of DNA cleavage, thus Cu([12]aneN3O) (2) is more
active than Cu([12]aneN4) (1) measured by the amount of
form III DNA.[23] Exchange of another N atom leads to an
even more effective cleaver: the Cu([12]aneN2O2) complex 3
cleaves supercoiled DNA completely into form II and even
Figure 6. Effect of different concentrations of complexes 2–4 and 6
on the pBR322 (0.025 μgμL–1) cleavage activity in Tris-HCl buffer
(50 mm, pH 7.4) and ascorbic acid (0.32 mm) at 37 °C for 2 h. Illus-
trated is the average of three experiments, the standard deviations
are shown as error bars. See the Supporting Information for a com-
generates about 15% of linear form III DNA. Whereas the plete overview of gels (cf. S-2,1–S-2,5).
Eur. J. Inorg. Chem. 2015, 4722–4730
4725
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
mation (cf. S-2). For complexes 2 and 4 a linear dependence (c) DNA Cleavage without External Reducing Agent
between concentration and cleavage activity is found, while
DNA cleavage was also conducted without the addition
of external reducing agent since some bis(aminoethyl)- and
bis(guanidinoethyl)-substituted [12]aneNONO copper com-
plexes have been shown to hydrolytically cleave DNA. How-
ever, in the reported case the guanidinium groups were tak-
ing an active part in the cleavage mechanism.[39] The time-
dependent cleavage activities of complexes 1–4 and 6 are
shown in Figure 8 (background cleavage is subtracted).
complexes 3 and 6 show a plateau forming within the error
limits above a concentration of 80 μm due to full consump-
tion of the reducing agent.
To elucidate the mechanism of DNA cleavage, incubation
was conducted under the addition of scavengers for reactive
oxygen species (ROS) like hydroxyl radicals (DMSO),[36]
singlet oxygen (NaN3),[37] hydrogen peroxide (catalase,
CAT)[38] and superoxide (superoxide dismutase, SOD)[37]
and compared to the cleavage without any scavenger (w/o).
As shown exemplarily for complex 3 in Figure 7 catalase
shows a strong inhibition effect on DNA cleavage for all
five complexes. Complexes 1, 2 and 4 are also inhibited by
DMSO indicating that hydroxyl radicals are involved in the
DNA cleavage. In the presence of SOD, the cleavage activity
of complexes 3 and 4 increases. As SOD converts superox-
ide anion radicals into hydrogen peroxide species, the in-
crease might be due to concomitance of superoxide anion
radicals that do not participate in the cleavage of DNA.
Figure 8. Time-dependent cleavage of complexes 1, 2, 3, 4 and 6
(0.4 mm) on pBR322 (0.025 μgμL–1) in Tris-HCl buffer (50 mm, pH
7.4) at 37 °C. Illustrated is the average of three experiments, the
standard deviations are shown as error bars. For an overview of
representative agarose gels refer to the Supporting Information (cf.
S-5).
While complexes 1 and 2 show no significant cleavage
activity, complexes 3, 4 and 6 can cleave pBR322 plasmid
DNA into form II even without external reducing agent.
After approximately 3 d, a plateau is formed at an amount
of 30–50% of form II DNA. This might indicate the mecha-
nism of DNA cleavage to be of oxidative rather than hydro-
lytic nature. As hydrolytic DNA cleavers are not consumed
in the cleaving process, they act catalytically and should not
lead to plateau formation. Presumably, the ligand itself acts
as reducing agent and is consumed during the course of the
reaction. As a result there is no catalytic CuII/CuI cycle.
This hypothesis was verified by using the active ester
bis(4-nitrophenyl) phosphate (BNPP) as a representative
substrate for hydrolytic cleavage of phosphodiester bonds
in an UV/Vis experiment: none of the complexes 3, 4 and
6 were able to cleave BNPP within an incubation time of
2 d (cf. Supporting Information S-6). As a positive control,
the enzyme phosphodiesterase was added to the incubation
Figure 7. Cleavage of pBR322 plasmid DNA (0.025 μg μL–1) in
Tris-HCl buffer (50 mm, pH 7.4), PBS (1.25X) and ascorbic acid
(0.32 mm) by complex 3 (0.02 mm) in the presence of ROS scaven-
gers at 37 °C for 2 h. See the Supporting Information for a com-
plete overview of agarose gels (cf. S-3,1–S-3,4).
These results prove the mechanism of DNA cleavage to
be oxidative involving hydroxyl radicals for complexes 1, 2
and 4, peroxide species for all complexes, and to some ex-
tent singlet oxygen as ROS for complexes 1 and 2 (summary
cf. Supporting Information S-3,5).
Furthermore, the effect of the pH value on the cleavage mixture leading to the release of the cleavage product nitro-
of pBR322 plasmid DNA was investigated; pH-dependent phenolate with its characteristic absorption at 400 nm.
studies showed, that the pH does not have an influence on
A hydrolytic cleavage mechanism could further be ex-
the cleavage activity of complexes 1, 2 and 4, whereas com- cluded by a religation assay using T4 ligase: form II DNA
plexes 3 and 6 showed a decrease in activity when going was collected from the agarose gel and incubated with ligase
from pH 7 to pH 9 (cf. Supporting Information S-4). Stud- (cf. Supporting information S-7 for complex 3). However,
ies carried out at pH 7.4 (Figures 5 and 6) thus fall into the after incubation, no form I band was detected, rendering
pH range with the highest activity of 3 and 6.
the hydrolytic cleavage mechanism unlikely.
Eur. J. Inorg. Chem. 2015, 4722–4730
4726
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
(d) Electrochemical Studies
Electrochemical studies were performed to gain further
insight into the redox reaction involved in the cleavage
mechanism (Figure 9).
Figure 10. Cyclic voltammogram of complex 3 in 0.1 m KCl solu-
tion at 50 mV/s and several scans.
However, differences in the electron-transfer process have
to be considered: while the electrochemical reduction is due
to an outer-sphere process, the reduction of oxygen by the
CuI species that leads to ROS is due to an inner-sphere
electron-transfer process. This means that activity in DNA
cleavage does not necessarily have to correlate with re-
duction potentials.[41] Nevertheless, a slight correlation be-
tween the number of O atoms, cleavage activities and redox
potentials between complexes 1–4 and 6 cannot be denied.
Figure 9. Cyclic voltammograms for complexes 3, 4 and 6 in 0.1 m
KCl solution at 100 mV/s and room temperature.
As we have shown for complexes 1 and 2 before,[23] com-
plexes 3, 4 and 6 also display a one-electron reduction wave
in their cyclic voltammograms. This process is due to the
reduction of CuII to CuI, being electrochemically irrevers-
ible for all complexes. A comparison of the reduction po-
tentials is presented in Table 4.
(e) Cytotoxicity Studies
In order to investigate the applicability of the compounds
as potential chemotherapeutic agents, cytotoxicity tests
using the MTT test were carried out (Supporting Infor-
mation S-8). Within 48 h even at 100 μm concentrations the
copper complexes did no show any effect on the viability of
A549 cells. Thus, DNA cleavage activity does not correlate
with cytotoxic behaviour. This might be due to low cell in-
ternalization or deactivation of the complexes by the com-
ponents of the cell culture medium. The latter case can be
excluded, though, due to the high stability of the macro-
cyclic complexes investigated (e.g. logβ = 23.4 for 1 and
15.9 for 2).[42,47]
Table 4. Reduction and reoxidation potentials of complexes 1–4
and 6.
Ered [V]
Eox [V]
1[23]
2[23]
3
–1.00
–0.74
–0.58
–0.54
–0.53
–0.50
–0.38
–0.36
–0.32
–0.32
4
6
Conclusions
A correlation between the reduction potentials and the
number of O atoms can be observed: as the number of O
atoms increases, there is a shift in the reduction potentials
towards less negative values.
In this paper we have investigated the oxidative cleavage of DNA
by macrocyclic copper complexes based on 1,4,7,10-tetraazacyclo-
dodecane. By stepwise exchange of the donor atoms (N Ǟ O) we
obtained five different complexes comprising also a regioisomeric
pair (NONO, N2O2).
By a closer inspection of the reoxidation waves it can be
concluded that the affinity for CuI of the cyclen analogue
As already shown before, insertion of one O atom leads to an in-
ligands decreases with an increase of the number of O crease in cleavage activity when compared to the tetraaza parent
compound.[23] We could now identify, however, the species, which
outreach oxacyclen: the [N2O2] derivative and 12-crown-4 as the
all-oxygen substituted form. At first surprisingly, the regioisomer
of the former (NONO) showed much lower activity. This might be
due to structural differences of the [NONO] and [N2O2] complexes
resulting in different reactivity. As expected according to the HSAB
principle, a CuI species generated during reduction has a lower af-
finity to an [NXOY] ligand with increasing O content. This results
in a general trend in which a higher O content leads to higher
cleavage activity due to an efficient ROS generation by reaction of
CuI with O2.
atoms, leading to the release of free CuI ions. To further
prove this hypothesis the cyclic voltammogram of complex
3 was scanned for several times leading to a growth of the
wave, which in turn can be attributed to the redox processes
of free copper (* in Figure 10).[40] This indicates that DNA
cleavage might also be accomplished by free copper ions.
This might also be the reason for an increase in cleavage
activities when going from complex 1 to 6, even if the re-
duction potentials do not correlate with the cleavage activi-
ties.
Eur. J. Inorg. Chem. 2015, 4722–4730
4727
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Whereas Xu and co-workers have shown that copper complexes of
1,7-dioxa-4,10-diazacyclododecane bearing two bis(guanidino-
ethyl) groups are able to cleave DNA in a hydrolytic fashion,[39] we
could not observe this behaviour for our NXOY donor ligands. This
was to a certain extent expected, since the hydrolytic pathway was
taken mostly due to the interaction of the positively charged guani-
dinium groups with the DNA backbone. A cleavage experiment
carried out in the absence of reducing agent did, nevertheless, result
in cleaved DNA in the case of the macrocyclic complexes with two
and four oxygen atoms. The reduction potentials of these deriva-
rial was dissolved in water (50 mL) and washed two times with
dichloromethane. The pH value of the solution was adjusted to
14, and the product was extracted five times with dichloromethane
(50 mL). The combined organic layers were dried with sodium sulf-
ate, and the solvent was removed in vacuo to yield L3 as a colour-
less oil (0.25 g, 1.4 mmol, 23%). HRMS (ESI): calcd. for [M –
NO3 ] 175.1441; found 175.1456. H NMR (CDCl3, 400 MHz): δ
= 3.71–3.67 (m, 4 H, OCH2CH2N), 3.64 (d, J = 1 Hz, 4 H, OCH2),
3.06 (d, J = 1 Hz, 4 H, NCH2), 3.02–2.95 (m, 4 H, OCH2CH2N)
ppm. 13C NMR (CDCl3, 100 MHz): δ = 69.56 (OCH2CH2N), 68.98
– +
1
tives (Ered = –0.58 to –0.53 V) suggest that DNA is oxidised directly (OCH2), 46.48 (NCH2), 45.51 (OCH2CH2N) ppm.
by the CuII complexes. Even the nucleobase easiest to oxidise, gua-
Synthesis of [Cu([12]aneN2O2)(NO3)2] (3): A solution of Cu(NO3)2·
nine (Eox = 0.8 V under physiological conditions), might hardly be
3H2O (282 mg, 1.2 mmol) in methanol (5 mL) was added to a boil-
oxidised by cyclen and oxacyclen complexes (Ered = 1.0 and
ing solution of 1,4-dioxa-7,10-diazacyclododecane (204 mg,
–0.74 V, respectively).[43]
1.2 mmol) in 5 mL methanol. The reaction mixture was heated to
reflux for 50 min, and the solvent removed in vacuo. To yield the
desired product 3, the leftover solid was recrystallised from aceto-
Our results show that CuII complexes of [N2O2] and [O4] macro-
cyclic ligands are the most efficient oxidative DNA cleavers within
a series of [NXOY] cyclen derivatives. Cleavage activity was even nitrile. Upon cooling, azure crystals of 3 (37 mg, 0.1 mmol, 9%)
– +
observed in the absence of a reducing agent.
formed. HRMS (ESI): calcd. for [M – NO3 ] 299.0542; found
299.0537. C8H18CuN4O8 (361.80): C 26.56, H 5.010, N 15.49;
found C 27.05, H 5.183, N 15.67. UV/Vis (50 mm, pH 7.4, Tris-
One should not disregard, however, the presence of Tris buffer in
this study as a potential competitor ligand to the cyclen derivatives.
At a high excess, which is usually the case when it is used as pH
buffer (here 50 mm), Tris binds CuII with an apparent affinity of
2ϫ106 m–1.[44] The stability constants for the macrocyclic ligands
are expected to be much higher, for which we found evidence in the
literature (e.g. logβ = 23.4 for 1 and 15.9 for 2).[42,47]
HCl): λmax (ε) = 620 nm (57.5 Lmol–1 cm–1). IR (solid): ν = 3215,
˜
3181 (w, ν NH), 2921 (w, ν CH2), 1749 (vw, δ NH), 1479 (vw, δ
CH2), 1281 (s, ν NO3 ), 1085 (m), 1012 (s), 858 (m) cm–1.
–
Synthesis of [Cu([12]aneNONO)NO3]NO3 (4):
A solution of
Cu(NO3)2·3H2O (207 mg, 0.9 mmol) in methanol (10 mL) was
added to a boiling solution of 1,7-dioxa-4,10-diazacyclododecane
(150 mg, 0.9 mmol) in 10 mL methanol. The reaction mixture was
heated to reflux for 10 min, the volume was reduced to 3 mL, and
product 5 precipitated upon cooling. Light-blue prisms were col-
lected by filtration and dried in vacuo to yield 4 (152 mg, 0.4 mmol,
Since cytotoxicity experiments showed insignificant effects of the
CuII complexes 1–4 and 6 on the viability of cancer cells, strategies
for improving DNA targeting and cellular internalization of the
compounds have still to be developed.
– +
49%). HRMS (ESI): calcd. for [M – NO3 ] 299.0542; found
299.0537. C8H18CuN4O8 (361.80): calcd. C 26.56), H 5.010), N
Experimental Section
15.49; found C 26.58, H 5.136, N 15.39. UV/Vis (50 mm pH 7.4
Tris-HCl): λmax (ε) = 638 nm (39.7 Lmol–1 cm–1). IR (solid): ν =
General: All chemical reagents were purchased from Sigma–Aldrich
if not stated otherwise. Only HPLC-grade solvents were used. UV/
Vis absorption spectra were recorded with a Varian Cary 100 spec-
trophotometer by using precision cells made of quartz (1 cm) at
25 °C. Electrospray mass spectra were obtained in positive-ion
mode by using an Agilent 6210 ESI-TOF mass spectrometer. Ele-
mental analyses were performed by using a Vario EL elemental
analyser.
˜
3164 (w, ν NH), 2945, 2894 (w, ν CH2), 1748 (vw, δ NH), 1499,
1448 (vw, δ CH2), 1364, 1326 (s, ν NO3), 1056 (m), 1015 (s), 853
(m) cm–1.
Synthesis of [Cu([12]aneO4)(NO3)2] (6): A solution of Cu(NO3)2·
3H2O (343 mg, 1.4 mmol) in methanol (10 mL) was added to a
boiling solution of 12-crown-4 (250 mg, 1.4 mmol) in 10 mmmL
methanol. The reaction mixture was heated to reflux for 10 min,
the volume was reduced to 3 mL, and product 6 precipitated upon
cooling. Light-blue crystals were collected by filtration and dried
in vacuo to yield 6 (345 mg, 1.0 mmol, 67%). HRMS (ESI): calcd.
Synthesis: Tosylated triethylene glycol, tosylated ethylenediamine
and 1,7-dioxa-4,10-diazacyclododecane (L4) were synthesised as
published earlier.[24,26,27] Complexes 1 and 2 were synthesised ac-
cording to the method published by our group.[23]
– +
for [M – NO3 ] 301.0223; found 301.0232. C8H16CuN2O10
(363.77): C 26.41), H 4.430, N 7.700; found C 26.27, H 4.610, N
Synthesis of 1,4-Dioxa-7,10-diazacyclododecane, [12]aneN2O2 (L3):
Under argon a solution of sodium hydride (1.96 g, 40.0 mmol,
60%) in absolute DMF (40 mL) was added to a solution of tosyl-
ated ethylenediamine (8.70 g, 23.6 mmol) in absolute DMF
(125 mL). The white suspension was heated to 100 °C as a solution
of tosylated tetraethylene glycol (10.02 g, 23.8 mmol) in absolute
DMF (125 mL) was added through a syringe pump over a period
of 4 h. Afterwards, the solution was stirred at the same temperature
for 3 h, then the reaction mixture was reduced to 50% of its volume
in vacuo. The solution was poured into ice-cold water (250 mL),
the precipitate was collected by filtration and recrystallised from
toluene. The white solid and phenol were dissolved in hydrogen
bromide (33% solution in acetic acid, 150 mL) and stirred at 80 °C
for 3 d. After the solvent had been removed in vacuo, the remaining
oil was washed three times with ethanol, which was then also re-
moved in vacuo. In order to remove leftover acid, the crude mate-
7.792. UV/Vis (50 mm, pH 7.4, Tris-HCl): λmax (ε) = 670 nm
(23.0 Lmol–1 cm–1). IR (solid): ν = 2969 (w, ν CH ), 1479 (vw, δ
˜
2
–
CH2), 1282 (s, ν NO3 ), 1127 (m), 1071 (s), 1009 (m), 929 (m), 896
(m), 807 (m) cm–1.
DNA Cleavage: DNA cleavage activity of complexes 1–4 and 6
towards pBR322 plasmid DNA (Carl Roth GmbH) was monitored
by gel electrophoresis. All experiments were carried out in triplicate
(except for the ROS quenching studies). In a typical experiment,
plasmid DNA (0.025 μg mL–1) in Tris-HCl buffer (50 mm, pH 7.4,
Fisher Scientific) was mixed with different concentrations of com-
plexes 1–4 and 6 in the presence and absence of ascorbic acid
(0.32 mm, Acros). Deionised water (Millipore system) was added
up to a total reaction volume of 8 μL before the samples were incu-
bated for a given time. After incubation, samples were analysed
directly or stored at –196 °C (liquid nitrogen) before usage in gel
Eur. J. Inorg. Chem. 2015, 4722–4730
4728
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
electrophoresis. For analysis, 1.5 μL of loading buffer (containing
3.7 mm bromophenol blue, 1.2 m saccharose in deionised water) was
added to the incubation solution and loaded onto an agarose
(SeaKem LE, Lonza) gel [1% in 0.5X Tris-borate-EDTA (TBE)
buffer, Fisher Scientific] containing ethidium bromide (0.2 μgmL–1,
Fisher Scientific). Electrophoresis was carried out at 40 V for 2 h
with an electrophoresis unit (Carl Roth; power supply: consort
EV243) in 0.5X TBE buffer. Bands were visualised by UV light and
photographed by using a gel documentation system (GelDoc, Bio-
Rad). The quantity of different DNA forms was estimated by using
Image Lab software. To do so, a value of 1 was assigned to the
supercoiled control DNA incubated under the same conditions as
in the other experiments but without the addition of metal com-
plexes (last lane of the agarose gels). All other values were ex-
pressed relatively to that. Taking into account that the supercoiled
form I of plasmid DNA has a smaller affinity towards ethidium
bromide, its intensity was afterwards multiplied with a correction
factor of 1.22.[45] Small DNA fragments were added together
within form III DNA. To make sure that DNA bands were cor-
rectly assigned to forms I, II and III, a DNA ladder was estab-
lished: 250 μL of pBR322 plasmid DNA (0.025 μgmL–1) were lin-
earised by using EcoRI nuclease, purified by agarose gel electro-
phoresis and then extracted by using a GenEluteTM extraction kit.
The extract was diluted to 250 μL and mixed with another solution
of 250 μL of pBR322 plasmid DNA (0.025 μgmL–1). 2 μL of this
aliquot were loaded into the first pocket of every agarose gel. Ex-
periments in the presence of ROS scavengers were conducted as
described above by using either 200 mm DMSO as hydroxyl radical
scavenger, 10 mm NaN3 as singlet oxygen scavenger, 2.5 mgmL–1
catalase (bovine liver, 2000–5000 unitsmL–1, Sigma Aldrich) as
hydrogen peroxide scavenger or 313 unitsmL–1 superoxide dismut-
ase (SOD, bovine liver, 2000–6000 unitsmL–1, Sigma Aldrich) as
superoxide anion scavenger. Addition of PBS to all samples was
necessary, because catalase and SOD were preincubated at 37 °C
in 10X PBS for 30 min resulting in a 1.25X PBS concentration in
the incubation mixture.
Studienstiftung is acknowledged for a fellowship for Jan Hormann.
We thank the Kulak group for proofreading of the article.
[1] D. Forman, J. Ferlay, B. W. Stewart, C. P. Wild, in: World Can-
cer Report 2014 (Eds.: B. W. Stewart, C. P. Wild), World Health
Organization, Lyon, France, 2014, pp. 77–180.
[2] Y. Jung, S. J. Lippard, Chem. Rev. 2007, 107, 1387–1407.
[3] C.-P. Tan, Y.-Y. Lu, L.-N. Ji, Z.-W. Mao, Metallomics 2014, 6,
978–995.
[4] F. H. Westheimer, Science 1987, 235, 1173–1178.
[5] H. W. Boyer, Annu. Rev. Microbiol. 1971, 25, 153–176.
[6] F. Mancin, P. Scrimin, P. Tecilla, U. Tonellato, Chem. Commun.
2005, 2540–2548.
[7] F. Mancin, P. Scrimin, P. Tecilla, Chem. Commun. 2012, 48,
5545–5559.
[8] J. A. Cowan, Curr. Opin. Chem. Biol. 2001, 5, 634–642.
[9] C. Santini, M. Pellei, V. Gandin, M. Porchia, F. Tisato, C. Mar-
zano, Chem. Rev. 2014, 114, 815–862.
[10] U. Jungwirth, C. R. Kowol, B. K. Keppler, C. G. Hartinger, W.
Berger, P. Heffeter, Antioxid. Redox Signaling 2011, 15, 1085–
1127.
[11] Q. Jiang, N. Xiao, P. Shi, Y. Zhu, Z. Guo, Coord. Chem. Rev.
2007, 251, 1951–1972.
[12] M. Subat, K. Woinaroschy, C. Gerstl, B. Sarkar, W. Kaim, B.
König, Inorg. Chem. 2008, 47, 4661–4668.
[13] B. Gruber, E. Kataev, J. Aschenbrenner, S. Stadlbauer, B.
König, J. Am. Chem. Soc. 2011, 133, 20704–20707.
[14] J. Li, Y. Yue, J. Zhang, Q.-S. Lu, K. Li, Y. Huang, Z.-W. Zhang,
H.-H. Lin, N. Wang, X.-Q. Yu, Trans. Met. Chem. 2008, 33,
759–765.
[15] Q. Liu, J. Zhang, M.-Q. Wang, D.-W. Zhang, Q.-S. Lu, Y. Hu-
ang, H.-H. Lin, X.-Q. Yu, Eur. J. Inorg. Chem. 2010, 45, 5302–
5308.
[16] Q.-S. Lu, Y. Huang, J. Li, Z.-W. Zhang, H.-H. Lin, X.-Q. Yu,
Chem. Biodiversity 2009, 6, 1273–1282.
[17] J. A. Krauser, A. L. Joshi, I. O. Kady, J. Inorg. Biochem. 2010,
104, 877–884.
[18] Y. Zhang, Y. Huang, J. Zhang, D.-W. Zhang, J.-L. Liu, Q. Liu,
H.-H. Lin, X.-Q. Yu, Sci. China Chem. 2011, 54, 129–136.
[19] R. Hettich, H.-J. Schneider, J. Am. Chem. Soc. 1997, 119, 5638–
5647.
[20] Q.-L. Li, J. Huang, Q. Wang, N. Jiang, C.-Q. Xia, H.-H. Lin,
J. Wu, X.-Q. Yu, Bioorg. Med. Chem. 2006, 14, 4151–4157.
[21] A. Bencini, E. Berni, A. Bianchi, C. Giorgi, B. Valtancoli, D.
Kumar Chand, H.-J. Schneider, Dalton Trans. 2003, 793–800.
[22] Y.-G. Fang, J. Zhang, S.-Y. Chen, N. Jiang, H.-H. Lin, Y.
Zhang, X.-Q. Yu, Bioorg. Med. Chem. 2007, 15, 696–701.
[23] J. Hormann, C. Perera, N. Deibel, D. Lentz, B. Sarkar, N. Ku-
lak, Dalton Trans. 2013, 42, 4357–4360.
X-ray Crystallography: X-ray diffraction data were collected with a
Bruker-AXS SMART CCD system. The structures were solved by
direct methods and refined by full-matrix least-squares methods
(SHELX-97). CCDC-1049602 (for 3) and CCDC-1049603 (for 4)
contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
Electrochemical Experiments: Cyclic voltammetry was carried out
in 0.1 m KCl solutions (Millipore water) by using a three-electrode
configuration (glassy carbon working electrode, Pt counter elec-
trode, Ag wire as pseudo reference) and a PAR VersaSTAT 4 po-
tentiostat. The ferrocene/ferrocenium (Fc/Fc+) couple served as in-
ternal reference. Experimental conditions were adapted from Lima
et al.[46]
[24] L. Börjesson, C. J. Welch, Acta Chem. Scand. 1991, 45, 621–
626.
[25] F. Bottino, M. Di Gracia, P. Finocchiaro, F. R. Fronczek, A.
Mamo, S. Pappalardo, J. Org. Chem. 1988, 53, 3521–3529.
[26] Y. Fukudome, H. Naito, T. Hata, H. Urabe, J. Am. Chem. Soc.
2008, 130, 1820–1821.
[27] F. Friscourt, C. J. Fahrni, G.-J. Boons, J. Am. Chem. Soc. 2012,
134, 18809–18815.
[28] H. R. Snyder, R. E. Heckert, J. Am. Chem. Soc. 1952, 74, 2006–
2009.
[29] D. I. Weisblat, B. J. Magerlein, D. R. Myers, J. Am. Chem. Soc.
1953, 75, 3630–3632.
Supporting Information (see footnote on the first page of this arti-
cle): Crystallographic data, DNA cleavage and quench experiments,
BNPP and religation assay, cytotoxicity studies.
[30] F. Z. Badaoui, J. Bourson, Anal. Chim. Acta 1995, 302, 341–
354.
Acknowledgments
[31] R. D. Rogers, Y. Song, J. Coord. Chem. 1995, 34, 149–157.
[32] M. C. Styka, R. C. Smierciak, E. L. Blinn, R. E. DeSimone,
J. V. Passariello, Inorg. Chem. 1978, 17, 82–86.
[33] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565–565.
[34] R. Clay, P. Murray-Rust, J. Murray-Rust, Acta Crystallogr.,
Sect. B 1979, 35, 1894–1895.
We thank Dr. Kai Licha from mivenion GmbH for donations of
cyclen. We thank Prof. Dr. Lentz for the measurement of crystal
structures, Dr. Naina Deibel for cyclovoltammetric measurements,
and Dr. Stefanie Wedepohl for cytotoxicity studies. The Deutsche
Eur. J. Inorg. Chem. 2015, 4722–4730
4729
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[35]
[43] H. Xie, D. Yang, A. Heller, Z. Gao, Biophys. J. 2007, 92, L70–
L72.
[44] J. Nagaj, K. Stokowa-Sołtys, E. Kurowska, T. Fra˛czyk, M.
Jezowska-Bojczuk, W. Bal, Inorg. Chem. 2013, 52, 13927–
13933.
M. Kashiba-Iwatsuki, M. Yamaguchi, M. Inoue, FEBS Lett.
1996, 389, 149–152.
E. L. Hegg, J. N. Burstyn, Inorg. Chem. 1996, 35, 7474–7481.
A. Sreedhara, J. D. Freed, J. A. Cowan, J. Am. Chem. Soc.
2000, 122, 8814–8824.
Y. Jin, J. A. Cowan, J. Am. Chem. Soc. 2005, 127, 8408–8415.
[36]
[37]
[45] R. P. Hertzberg, P. B. Dervan, Biochemistry 1984, 23, 3934–
[38]
3945.
[39] X. Sheng, X.-M. Lu, Y.-T. Chen, G.-Y. Lu, J.-J. Zhang, Y. Shao,
F. Liu, Q. Xu, Chem. Eur. J. 2007, 13, 9703–9712.
[40] S.-H. Chiou, J. Biochem. 1983, 94, 1259–1267.
[41] J. Schnödt, M. Sieger, B. Sarkar, J. Fiedler, J. S. Manzur, C.-Y.
Su, W. Kaim, Z. Anorg. Allg. Chem. 2011, 637, 930–934.
[42] R. D. Hancock, M. S. Shaikjee, S. M. Dobson, J. C. A. Boey-
ens, Inorg. Chim. Acta 1988, 154, 229–238.
[46] L. M. P. Lima, D. Esteban-Gómez, R. Delgado, C. Platas-Igle-
sias, R. Tripier, Inorg. Chem. 2012, 51, 6916–6927.
[47] Thermodynamic and kinetic data for macrocycle interaction
with cations and anions: R. Izatt, M. K. Pawlak, J. S. Brad-
shaw, R. L. Bruening, Chem. Rev. 1991, 91, 1721–2085.
Received: June 2, 2015
Published Online: August 3, 2015
Eur. J. Inorg. Chem. 2015, 4722–4730
4730
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim