Quinoline-2-carboxaldehyde thiosemicarbazones and their Cu(II) and Ni(II) complexes as topoisomerase IIa inhibitors

Article abstract of DOI:10.1016/j.jinorgbio.2015.08.008

A series of quinoline-2-carboxaldehyde thiosemicarbazones and their copper(II) and nickel(II) complexes were synthesized and characterized. In all complexes the ligands are in the E configuration with respect to the imino bond and behave as terdentate. Th

Full text of DOI:10.1016/j.jinorgbio.2015.08.008

Journal of Inorganic Biochemistry 152 (2015) 10–19
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Quinoline-2-carboxaldehyde thiosemicarbazones and their Cu(II) and
Ni(II) complexes as topoisomerase IIa inhibitors
Franco Bisceglie a,d, Anastasia Musiari a, Silvana Pinelli b,d, Rossella Alinovi b,d, Ilaria Menozzi c,
Eugenia Polverini c, Pieralberto Tarasconi a,d, Matteo Tavone a, Giorgio Pelosi a,d,
⁎
a
Department of Chemistry, University of Parma, Parco Area delle Scienze 17A, 43124 Parma, Italy
Department of Clinical and Experimental Medicine, University of Parma, Via Gramsci 14, 43126 Parma, Italy
Department of Physics and Earth Sciences, University of Parma, Parco Area delle Scienze, 43124 Parma, Italy
b
c
d
CIRCMSB (Consorzio Interuniversitario di Ricerca in Chimica dei Metalli nei Sistemi Biologici), Parma Unit, University of Parma, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 8 May 2015
Received in revised form 7 July 2015
Accepted 5 August 2015
Available online 8 August 2015
A series of quinoline-2-carboxaldehyde thiosemicarbazones and their copper(II) and nickel(II) complexes were
synthesized and characterized. In all complexes the ligands are in the E configuration with respect to the imino
bond and behave as terdentate. The copper(II) complexes form square planar derivatives with one molecule of
terdentate ligand and chloride ion. A further non-coordinated chloride ion compensates the overall charge.
Nickel(II) ions form instead octahedral complexes with two ligands for each metal ion, independently from the
stoichiometric metal:ligand ratio used in the synthesis. Ligands and complexes were tested for their antiprolifer-
ative properties on histiocytic lymphoma cell line U937. Copper(II) derivatives are systematically more active
than the ligands and the nickel complexes. All copper derivatives result in inhibiting topoisomerase IIa in vitro.
Computational methods were used to propose a model to explain the different extent of inhibition presented
by these compounds. The positive charge of the dissociated form of the copper complexes may play a key role
in their action.
Keywords:
Topoisomerase IIa
Quinoline
Thiosemicarbazone
Copper
Nickel
© 2015 Elsevier Inc. All rights reserved.
1. Introduction
when thiosemicarbazones act as antiproliferating agents, a cell cycle
block in phase G2/M is observed [10–12]. Among putative targets
Thiosemicarbazones are a class of compounds extensively studied
for their biological properties [1–6]. Two fundamental steps in their
development as drugs were made in the Sixties with the discovery of
their activity as antitumor agents [7] and of their antiviral properties;
in particular, these latter were successfully exploited to treat smallpox
[8]. Since then, their anticancer properties have received increasing at-
tention and in recent years one of them, Triapine® (3-aminopyridine-
2-carboxaldehyde thiosemicarbazone), has been developed as an anti-
cancer drug [5]. For some types of tumors, clinical trials for this molecule
have now reached clinical phase II [9].
In contrast with the large quantity of data from preliminary anti-
proliferative assays carried out with these molecules on several cell
lines in vitro found in the literature [4], the amount of information
on specific targets (DNA, enzymes or proteins) is still wanting. In the
majority of cases, the biological activity of thiosemicarbazones is
strongly dependent on the presence of metal ions, since their coordina-
tion compounds are systematically more efficient than the parent
ligands. According to our experience, another common feature is that,
consistent with this fact, a prominent place is certainly occupied by
topoisomerase IIa (TopoIIa). This homodimeric enzyme, in fact, plays
an essential role in all eukaryotic cells and it is particularly active in
cancerous cells because they present a higher growth rate [13]. The
main role of TopoIIa is the regulation of the topological states of DNA
through transient cleavage strand passing and re-ligation of double-
stranded DNA, resulting in decatenation of intertwined DNA molecules
and relaxation of supercoiled DNA. In addition, TopoIIa plays a key role
in DNA replication and is required for condensation and segregation of
chromosomes. The expression of TopoIIa is cell-cycle dependent with
both protein levels and catalytic activity peaking at G2/M. As said
above, this protein is therefore a plausible candidate as a target for
thiosemicarbazones, since its phosphorylation/dephosphorylation
is part of the regulatory checkpoints at the entry and progression of
mitosis. For this reason, molecules able to inhibit TopoIIa are valuable
compounds as anticancer drugs thanks to their capacity to stop prolifer-
ation and to lead cells to apoptosis. Known TopoIIa inhibitors are
extremely varied and act through different mechanisms. The target for
most of these molecules is the DNA–protein complex, rather than the
protein itself or DNA [14]. Some of the compounds interact predomi-
nantly with the protein (quinones), while others with DNA (ellipticine).
DNA-intercalating drugs (etoposide) [15] probably bind to both DNA
⁎
Corresponding author at: Department of Chemistry, University of Parma, Parco Area
delle Scienze 17A, 43124 Parma, Italy.
E-mail address: giorgio.pelosi@unipr.it (G. Pelosi).
http://dx.doi.org/10.1016/j.jinorgbio.2015.08.008
0162-0134/© 2015 Elsevier Inc. All rights reserved.

F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
11
and the enzyme through intercalation of the aromatic part at the
enzyme–DNA interface. Some drugs work by inhibiting the re-ligation
of cleaved DNA, whereas others, such as ellipticine, genistein, and
quinolones, are presumed to accelerate the forward rate of complex
formation.
Exp.: C 57.38%, H 4.22%, N 24.55%. 1H NMR (300 MHz, DMSO-d6)
11.79 ppm (1H, s, C=N–NH), 8.45 ppm (2H, d, J = 8.7 Hz, NH2),
8.35 ppm (2H, m, CH arom.), 8.23 ppm (1H, s, CH=N), 8.00 ppm (2H,
t, J = 8.7 Hz, CH arom.), 7.78 ppm (1H, t, J = 8.4 Hz, CH arom.),
7.62 ppm (1H, t, J = 8.1 Hz, CH arom.). IR: 3395 cm−1, 3262 cm−1
and 3152 cm−1 ν N–H, 3066 cm−1 ν C–H arom., 2986 cm−1 ν C–H
aliph., 1609 cm−1 ν C=N, 1533 cm−1 and 1503 cm−1 ν C=C,
1110 cm−1 and 839 cm−1 ν C=S.
In this paper, we report the synthesis of a series of quinoline-2-
carboxaldehyde thiosemicarbazones variously substituted on the termi-
nal nitrogen (Fig. 1), of the corresponding copper(II) complexes, square
planar of stoichiometry M:L 1:1, and of the nickel(II) derivatives,
prepared with the same stoichiometry, but which spontaneously re-
solve in compounds with an M:L 1:2 ratio, with octahedral geometry.
All compounds have been systematically tested for their capacity of
inhibiting proliferation of a human histiocytic lymphoma, cell line U937,
and their activity has been determined by evaluating their IC50. Sub-
sequently, the inhibition activity on TopoIIa was assayed in vitro in
order to evaluate the capacity of the compounds to inhibit directly the
enzyme. Several docking simulations of the compounds were eventually
performed on TopoIIa in an attempt to understand the inhibition
mechanisms.
2.2.2. Quinoline-2-carboxyaldehyde N4,N4-dimethylthiosemicarbazone
(HL2)
N4,N4-dimethyl-3-thiosemicarbazide (271 mg, 2.16 mmol) was
dissolved in 120 mL of ethanol at room temperature under stirring. An
equimolar amount of quinoline-2-carboxaldehyde (351 mg, 2.16 mmol)
was then added. The solution obtained was clear and deep yellow. The
reaction mixture was then kept under stirring and placed in an ice bath.
After an hour a light yellow precipitate began to form, but the reaction
was left running for 24 h. The precipitate was dried on a Buchner funnel,
rinsed with ethanol, allowed to dry and eventually weighed. The product
was recrystallized from acetonitrile and the crystals so formed were
allowed to solve the structure by X-ray diffraction. Yield: 39%. Elemental
analysis: C13H14N4S Calc: 60.44%, H 5.46% N 21.69% Exp.: C 60.68%, H
5.40%, N 21.84%. 1H NMR (300 MHz, DMSO-d6) 11.36 ppm (1H, s, NH),
8.37 ppm (1H, s, CH=N), 8.03 ppm (4H, m, CH4 arom.), 7.78 ppm (1H,
t, J = 5.7 Hz, CH arom.), 7.62 ppm (1H, t, J = 5.7 Hz, CH arom.),
3.36 ppm (6H, s, N–(CH3)2). IR: 3066 cm−1 ν N–H, 3008 cm−1 ν C–H
2. Experimental
2.1. General experimental conditions
Infrared spectra were recorded in the range 4000–400 cm−1 using a
Nicolet 5PC FTIR spectrophotometer. Elemental analyses (C, H, N) were
carried out using a flash1112 Series CHNS-O Analyzer by CE Instru-
ments. 1H NMR spectra were recorded using a Bruker Avance 300
spectrometer.
arom., 2921 cm−1 ν C–H aliph., 1619 cm−1 ν C=N, 1576 cm−1
1545 cm−1 and 1499 cm−1 ν C=C, 1104 cm−1 and 822 cm−1 ν C=S.
,
2.2.3. Quinoline-2-carboxyaldehyde N4-methylthiosemicarbazone (HL3)
2.2. Syntheses of the ligands
N4-methylthiosemicarbazide (234.2 mg, 2.16 mmol) was dissolved
in 40 mL of ethanol at room temperature and under stirring. An equi-
molar amount of quinoline-2-carboxyaldehyde (355 mg, 2.16 mmol)
was added. The resulting solution was clear and yellow-orange. The
reaction mixture was maintained in an ice bath under stirring. After
about half an hour a white precipitate began to form. The reaction was
left standing for 24 h and the white precipitate was filtered on a Buchner
funnel, washed with ethanol, dried on air and weighed. The product was
recrystallized from acetonitrile and the crystals obtained were suitable
for X-ray diffraction analysis. Yield: 83%. Elem. Anal.: C12H12N4S calc.
C 58.99%, H 4.95%, N 22.93%, exp. C 59.80%, H 4.92%, N 22.82%. 1H
NMR (300 MHz, DMSO-d6) 11.87 ppm (1H, s, C=N–NH), 8.81 ppm
(1H, q, J = 4.2 Hz, S=C–NH), 8.45 ppm (1H, d, J = 8.7 Hz, CH arom.),
8.39 ppm (1H, d, J = 8.7 Hz, CH arom.), 8.23 ppm (1H, s, CH=N),
8.00 ppm (2H, t, J = 7.5 Hz, CH arom.), 7.78 ppm (1H, t, J = 7.8 Hz,
CH arom.), 7.62 ppm (1H, t, J = 7.5 Hz, CH arom.), 3.07 ppm (3H, d,
J = 4.5 Hz, NH–CH3). IR: 3322 cm−1 and 3117 cm−1 ν N–H,
3069 cm−1 ν C–H arom., 2937 cm−1 ν C–H aliph., 1602 cm−1 ν C=N,
1536 cm−1 and 1502 cm−1 ν C=C. 1112 cm−1 and 846 cm−1 ν C=S.
2.2.1. Quinoline-2-carboxyaldehyde thiosemicarbazone (HL1)
Thiosemicarbazide (337.6 mg, 3.7 mmol) was dissolved in 80 mL
of methanol at room temperature under stirring, and the solution
was added with an equimolar amount of quinoline-2-carboxaldehyde
(582.1 mg, 3.7 mmol). A clear orange solution was obtained. This reac-
tion mixture was kept under stirring and placed in an ice bath. The reac-
tion was left standing for 20 h and the white precipitate afforded was
filtered on a Buchner funnel and washed using ethanol. Yield: 77%.
Elemental analysis: C11H10N4S Calc.: C 57.37%, H 4.38%, N 24.33%;
2.2.4. Quinoline-2-carboxyaldehyde N4-phenylthiosemicarbazone (HL4)
N4-phenylthiosemicarbazide (391 mg, 2.3 mmol) was dissolved in
60 mL of ethanol at room temperature and under stirring. An equimolar
amount of quinoline-2-carboxyaldehyde (375.3 mg, 2.3 mmol) was
added. The clear dark-yellow solution was kept in an ice bath under
stirring. After 24 h, a white precipitate was filtered and washed with
ethanol. Yield: 52%. Elem. Anal.: C17H14N4S calc. C 66.64%, H 4.61%, N
18.29%; exp. C 66.82%, H 4.74%, N 18.06%. 1H NMR (300 MHz, DMSO-d6)
12.20 ppm (1H, s, C=N–NH), 10.39 ppm (1H, s, S = C–NHPhe),
8.62 ppm (1H, d, J = 8.4 Hz, CH arom.Quin), 8.40 ppm (1H, d, J = 8.7 Hz,
CH arom.Quin), 8.35 ppm (1H, s, CH=N), 8.02 ppm (2H, t, J = 9 Hz, CH
arom.Quin), 7.79 ppm (1H, t, J = 7.5 Hz, CH arom.Quin), 7.64 ppm (1H, t,
J = 7.5 Hz, CH arom.Quin), 7.57 ppm (2H, d, J = 8.1 Hz, CH arom.Phe),
7.41 ppm (2H, t, J = 7.5 Hz, CH arom.Phe), 7.26 ppm (1H, t, J = 7.5 Hz,
CH arom.Phe). IR: 3313 cm−1 and 3119 cm−1 ν N–H, 3054 cm−1 and
Fig. 1. Scheme of the ligands.

12
F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
3041 cm−1 ν C–H arom., 2945 cm−1 ν C–H aliph., 1598 cm−1 ν C=N,
1537 cm−1 and 1503 cm−1 ν C=C, 1096 cm−1 and 825 cm−1 ν C=S.
2.3.4. Quinoline-2-carboxyaldehyde N4-phenylthiosemicarbazonato
copper(II) chloride ([Cu(L4)Cl])
An equimolar amount of copper(II) chloride dihydrate (41.8 mg,
0.25 mmol) was added to 40 mL of a methanol solution of ligand HL4
(45.1 mg, 0.15 mmol). The reaction mixture became immediately
dark, but, as soon as the precipitate began to form, it turned light
green. The flask was left under reflux and stirring for 4 h and then was
left cooling to room temperature. The red precipitate was washed
with methanol. Yield: 68%. Elem. anal.: C17H13ClCuN4S calc. C 50.49%,
H 3.24%, N 13.85%; exp. C 49.89%, H 3.22%, N 13.67%. IR: 3199 cm−1 ν
N–H, 3104 cm−1 and 3059 cm−1 ν C–H arom., 2951 cm−1 ν C–H
aliph., 1597 cm−1 ν C=N, 1535 cm−1 and 1504 cm−1 ν C=C,
1100 cm−1 and 825 cm−1 ν C=S.
2.2.5. Quinoline-2-carboxyaldehyde N2-methylthiosemicarbazone (L5)
N2-methylthiosemicarbazide (412.5 mg, 3.9 mmol) was dissolved in
60 mL of methanol at room temperature and under stirring. An equi-
molar amount of quinoline-2-carboxyaldehyde (616.5 mg, 3.9 mmol)
and a few drops of glacial acetic acid were added to obtain a clear,
yellow solution. The solution was left standing for 20 h and the white
precipitate that formed was filtered and washed with ethanol. The
product was recrystallized from ethanol and the crystals obtained
were suitable for X-ray diffraction analysis. Yield: 76%. Elem. Anal.:
C12H12N4S calc. C 58.99%, H 4.95%, N 22.93%; exp. C 58.86%, H 4.98%,
N 23.08%. 1H NMR (300 MHz, DMSO-d6) 8.67 ppm (1H, m, CH arom.),
8.60 ppm (2H, d, J = 8.7 Hz, NH2), 8.39 ppm (1H, d, J = 8.7 Hz, CH
arom.), 8.02 ppm (2H, m, CH arom.), 7.96 ppm (1H, s, CH=N),
7.79 ppm (1H, t, J = 7.5 Hz, CH arom.), 7.64 ppm (1H, t, J = 7.5 Hz,
CH arom.), 3.87 ppm (3H, s, N–CH3). IR: 3403 cm−1 and 3230 cm−1 ν
N–H, 3060 cm−1 ν C–H arom., 2890 cm−1 ν C–H aliph., 1595 cm−1 ν
C=N, 1561 cm−1 ν C=C, 1118 cm−1 and 838 cm−1 ν C=S.
2.3.5.
Quinoline-2-carboxyaldehyde
N2-methylthiosemicarbazone
copper(II) chloride ([Cu(L5)Cl2])
Ligand L5 (80.8 mg, 0.33 mmol) was dissolved in 70 mL of ace-
tonitrile at room temperature and under stirring. An equimolar amount
of copper chloride dihydrate (56.4 mg, 0.33 mmol) was added and
the mixture was left under reflux and stirring for 6 h. The light green
precipitate was filtered and washed with methanol. Yield: 88%. Elem.
anal. C12H12Cl2CuN4S: calc. C 38.05%, H 3.19%, N 14.79%; C 38.21%, H
3.12%, N 14.56%. IR: 3304 cm−1 and 3189 cm−1 ν N–H, 2998 cm−1 ν
C–H, 1592 cm−1 ν C=N, 1560 cm−1 ν C=C, 1145 cm−1 and
843 cm−1 ν C=S.
2.3. Syntheses of the copper(II) complexes
2.3.1. Quinoline-2-carboxyaldehyde thiosemicarbazonato copper(II)
chloride dihydrate ([Cu(L1)Cl]·2H2O)
Ligand HL1 (104 mg, 0.45 mmol) was dissolved in 200 mL of warm
acetonitrile and under stirring. An equimolar amount of copper(II)
chloride dihydrate (77 mg, 0.45 mmol) was added. The solution became
immediately dark and a green precipitate separated. The reaction
mixture was left under reflux and stirring for about 4 h. It was then
left cooling at room temperature and filtered on a Buchner funnel. The
light green precipitate was washed with acetonitrile and dried. Yield:
58%. Elem. Anal.: C11H13ClCuN4O2S calc. C 36.16%, H 3.86%, N
15.33%; exp. C 35.78%, H 2.83%, N 14.50%. IR: 3267 cm−1 ν N–H,
3057 cm−1 ν C–H arom., 2970 cm−1 ν C–H aliph., 1573 cm−1 ν C=N,
1508 cm−1 ν C=C, 1141 cm−1 and 831 cm−1 ν C=S.
2.4. Syntheses of the nickel(II) complexes
2.4.1. (Quinoline-2-carboxyaldehyde thiosemicarbazonato) (quinoline-2-
carboxyaldehyde thiosemicarbazone) nickel(II) chloride methanol solvate
([Ni(L1)(HL1)]Cl·CH3OH)
To 70 mL of a warm methanol solution of ligand HL1 (88.2 mg,
0.38 mmol) were added 3 drops of NaOH 1 N and an equimolar amount
of nickel chloride hexahydrate (91 mg, 0.38 mmol). The solution im-
mediately turned dark red. The reaction mixture was left under stirring
for about 5 h and then cooled to room temperature. The precipitate was
a brick-red microcrystalline powder. The product was recrystallized
from methanol and crystals apt for an XRD analysis were obtained.
Yield: 66%. Elem. Anal. C23H23ClN8NiOS2 Calc. C 47.16%, H 3.96%, N
19.13%; exp. C 47.46%, H 3.70%, N 19.27%. 1H NMR (300 MHz, DMSO-d6):
11.79 ppm (1H, s, C=N–NH), 8.45 ppm (2H, d, J = 8.7 Hz, NH2),
8.35 ppm (2H, m, CH arom.), 8.23 ppm (1H, s, CH=N), 8.00 ppm (2H, t,
J = 8.7 Hz, CH arom.), 7.78 ppm (1H, t, J = 8.4 Hz, CH arom.), 7.62 ppm
(1H, t, J = 8.1 Hz, CH arom.). IR: 3329 cm−1, 3242 cm−1, 3134 cm−1 ν
C–H arom.; 2938 cm−1 ν C–H aliph.; 1575 cm−1 ν C=N; 1539 cm−1
and 1505 cm−1 ν C=C; 1116 cm−1 and 817 cm−1 ν C=S.
2.3.2. Quinoline-2-carboxyaldehyde N4,N4-dimethylthiosemicarbazonato
copper(II) chloride ([Cu(L2)Cl])
To a warm solution obtained by dissolving ligand HL2 (62.7 mg,
0.23 mmol) in 20 mL of ethanol, an equimolar amount of copper(II)
chloride dihydrate (40.2 mg, 0.23 mmol) was added, and the solution
immediately turned dark and a precipitate appeared. The reaction
mixture was kept under reflux and under stirring for 2 h. It was then
left to cool down and filtered. The green product was washed with
ethanol and allowed to dry. From the ethanol solution crystals suitable
for X-ray structure determination were afforded. Yield: 46%. Elem.
Anal.: C13H13ClCuN4S calc. C 43.78%, H 3.68%, N 15.72%; exp. C
43.41%, H 3.43%, N 15.35%. IR: 3011 cm−1 ν C–H arom., 2924 cm−1 ν
C–H aliph., 1596 cm−1 ν C=N, 1576 cm−1 and 1510 cm−1 ν C=C,
1119 cm−1 and 821 cm−1 ν C=S.
2.4.2. Bis(quinoline-2-carboxyaldehyde N4,N4-dimethylthiosemicarbazone)
nickel(II) dichloride ([Ni(HL2)2]Cl2·2H2O)
Nickel chloride hexahydrate (41.4 mg, 0.17 mmol) was added to
35 mL of a warm ethanol solution of ligand HL2 (44.5 mg, 0.17 mmol).
The solution turned red-orange. The mixture was kept under reflux
and stirring for 5 h and was then left to cool down to room temper-
ature. The product was a red-orange powder. Yield: 56%. Elem. Anal,
C26H28Cl2N8NiS2 calc. C 48.32%, H 4.38%, N 17.33%; exp. C 48.24%, H
4.86%, N 17.46%. IR: 3080–2800 cm−1 ν C–H, 1606 cm−1 ν C=N,
1570 cm−1 ν C=C, 1118 cm−1 and 798 cm−1 ν C=S.
2.3.3. Quinoline-2-carboxyaldehyde N4-methylthiosemicarbazonato
copper(II) chloride dihydrate ([Cu(L3)Cl]·2H2O)
Ligand HL3 (45.6 mg, 0.19 mmol) was dissolved in 100 mL of warm
acetonitrile and kept under stirring. An equimolar amount of copper(II)
chloride dihydrate (31.8 mg, 0.19 mmol) was added and the reaction
mixture was kept under stirring for 1 h and a half. The mixture
was cooled at room temperature and filtered. The green product was
washed with acetonitrile. The compound was crystallized using a mix-
ture of acetonitrile/methanol (2:1) and crystals suitable for an X-ray
analysis were obtained. Yield: 39%. Elem. Anal.: C12H15ClCuN4O2S,
calc. C 37.99%, H 3.98%, N 14.77%; exp. C 38.21%, H 3.84%, N 14.64%. IR:
3137 cm−1 ν N–H, 3072 cm−1 ν C–H arom., 2952 cm−1 ν C–H aliph.,
1582 cm−1 ν C=N, 1530 cm−1 and 1508 cm−1 ν C=C, 1124 cm−1
and 835 cm−1 ν C=S.
2.4.3. Bis(quinoline-2-carboxyaldehyde N4-methylthiosemicarbazonato)
nickel(II) ([Ni(L3)2])
Ligand HL3 (65.2 mg, 0.26 mmol) was dissolved in 50 mL of warm
ethanol under stirring. An equimolar quantity of nickel chloride hexahy-
drate (59.8 mg, 0.26 mmol) was then added and the solution imme-
diately turned dark red. The reaction mixture was kept under reflux
and stirring for 2 h and a half. The solution was left cooling down to
room temperature and poured into a crystallizer. The product was

F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
13
recrystallized in a mixture of methanol/ethanol (1:1) and crystals
2.6.2 . Cell culture
suitable for X-ray diffraction studies were obtained. Yield: 51%. Elem.
Anal. C24H22N8NiS2 Calc. C 52.84%, H 4.06%, N 20.54%, exp. C 52.64%,
H 3.84%, N 20.36%. IR: 3174 cm−1 ν N–H; 3038 cm−1 ν C–H arom.;
2961 cm−1 ν C–H aliph.; 1574 cm−1 ν C = N; 1531 cm−1 and
1505 cm−1 ν C=C; 1123 cm−1 and 834 cm−1 ν C=S.
U937, a monoblastoid line, was obtained from American Type
Culture Collection (ATCC, Rockville, MD, USA). Cells were cultured in
RPMI 1640 medium supplemented with 10% (v/v) fetal bovine serum,
100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine
and maintained at 37 °C in a humidified atmosphere with 5% CO2.
2.6.3. Cell treatment
2.4.4. Bis(quinoline-2-carboxyaldehyde N4-phenylthiosemicarbazonato)
nickel(II) ([Ni(L4)2])
Ligand and complex solutions were prepared from a concentrated
10 mg/mL stock solution in DMSO. U937 cells at a mid-log phase were
seeded at a density of 1.5 × 105 cells/mL in RPMI 1640 medium in plates
or flasks. After 24 h the cells in the exponential phase of growth were
treated with DMSO 0.1% (control) and all the compounds for different
periods.
Ligand HL4 (58.7 mg, 0.19 mmol) was dissolved in 80 mL of warm
methanol. To this mixture, an equimolar amount of nickel chloride
hexahydrate (45.5 mg, 0.19 mmol) was added and the solution became
dark red. The mixture was kept 5 h under stirring and then left cooling
to room temperature. The product obtained by slow evaporation of the
solvent was an orange-brown microcrystalline powder. Yield: 48%.
Elem. Anal. C34H26N8NiS2 Calc. C 61.00%, H 3.91%, N 16.74%; exp. C
59.66%, H 3.13%, N 16.55%. IR: 3403–3270 cm−1 ν N–H, 3057 cm−1 ν
C–H arom., 1596 cm−1 ν C=N, 1541 cm−1 and 1507 cm−1 ν C=C
1096 cm−1 and 825 cm−1 ν C=S.
2.6.4. Proliferation assay
Cell proliferation was measured using the Cell-Titer 96® AQueous
One Solution Cell Proliferation Assay (MTS) kit. U937 cells were seeded
into 96-well plates in triplicate. One hour prior to the end of each treat-
ment MTS reagent (20 μL) was added to each well. At the end of expo-
sure time, the absorbance of the formazan product was measured
at 490 nm by a microwell plate reader (Multiskan Ascent, Thermo
Labsystems, Helsinki, Finland). Untreated cells were used for the cali-
bration curve. Medium containing MTS and different concentrations of
complexes, but without cells, was included to detect possible inter-
action between the MTS and the test compounds. Each experiment
was repeated on three separate occasions.
2.4.5. Bis(quinoline-2-carboxyaldehyde N2-methylthiosemicarbazone)
nickel(II) chloride tetrahemihydrate ([Ni(L5)2]Cl2·4.5H2O)
Ligand L5 (88.8 mg, 0.36 mmol) was dissolved in 70 mL of methanol
at room temperature and an equimolar amount of nickel chloride hexa-
hydrate (86.4 mg, 0.36 mmol) was added under stirring. The solution
immediately turned dark red. After 20 h under stirring the mixture
was left cooling down to room temperature and then poured into a crys-
tallizer. The slow evaporation afforded a microcrystalline precipitate
and the yield was 75%. The product was recrystallized from methanol
and crystals suitable for an XRD structure determination were obtained.
Yield: 75%. Elem. Anal. C24H33Cl2N8NiO4.5S2 Calc. C 41.22%, H 4.76%, N
16.02%; exp. C 40.34%, H 4.82%, N 15.63%. 1H NMR (300 MHz, DMSO-d6):
8.67 ppm (1H, m, CH arom.); 8.60 ppm (2H, d, J = 8.7 Hz, NH2);
8.39 ppm (1H, d, J = 8.7 Hz, CH arom.); 8.02 ppm (2H, m, CH arom.);
7.96 ppm (1H, s, CH=N); 7.79 ppm (1H, t, J = 7.5 Hz, CH arom.);
7.64 ppm (1H, t, J = 7.5 Hz, CH arom.); 3.87 ppm (3H, s, N–CH3). IR:
Cell viability was monitored also by counting viable cells in a hemo-
cytometer (trypan blue exclusion); after incubation with copper com-
plexes, cells collected from eight wells were pooled and placed in
trypan blue for 5 min.
2.6.5. TopoIIa inhibition assay
Inhibition assays on TopoIIa were carried out using a kit by TopoGen
(Port Orange, FL, USA). The standard procedure was followed [20]: 14 μL
di H2O, 4 μL of buffer solution, 1 μL supercoiled DNA and 1 μL of eukary-
otic topoisomerase IIα (~2 A.U. as described by the producer) are mixed
in a microcentrifuge tube (1.7 mL). In the experiments which require
the presence of our compounds we added 2 μL of a stock solution of
the compound to be tested dissolved in DMSO and the initial volume
of water was corrected to reach a final volume of 20 μL. The reaction
mixture was gently shaken, centrifuged for a short time and incubated
at 37° for 30 min. After 30 min 2 μL of a 10% SDS solution was added
to quench the reaction, 2 μL of proteinase K to digest the remaining
proteins and the reaction mixture was incubated at 37° for 15 min.
After this, 2.5 μL of loading dye (10×) was added to the reaction mixture
and the tube was subsequently gently shaken and centrifuged. The
sample was then loaded on a 1% agarose gel and an electrophoresis
was run for 5 h at 5 V/cm. The resulting gel was then dyed in a TAE
solution containing 5 μg/L of ethydium bromide. The resulting gels
were visualized with a BioRad Fluor-S TM MultiImager.
3400–3200 cm−1 ν N–H, 3180–2950 cm−1 ν C–H, 1585 cm−1
C=N, 1555 cm−1 ν C=C, 1140 cm−1 and 829 cm−1 ν C=S.
ν
2.5. Crystal structure determination
The X-ray structures of ligands HL2, HL3 and HL5 were collected on a
Siemens AED diffractometer with Cu-Kα radiation (λ = 1.54178 Å) and
those of compounds [Cu(L2)Cl], [Cu(L3)Cl], [Ni(L1)(HL1)]Cl·CH3OH,
[Ni(L3)2] and [Ni(L5)2]Cl2·4.5H2O were collected on a Bruker SMART
APEX2 with Mo-Kα radiation (λ = 0.71069 Å). The structures were
solved through direct methods with SIR97 [16] and refined with
SHELXL97 [17] implemented in WinGX [18]. The drawing were obtained
using the ORTEPIII program [19]. The relevant crystal data are reported in
the Supplementary material file as Table 1S. CCDC 1062429–1062436
contain the supplementary crystallographic data. These data can be
obtained free of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.
html, or from the Cambridge Crystallographic Data Centre, 12 Union
Road, Cambridge CB2 1EZ, UK; fax: (+44) 1223-336-033, or e-mail:
deposit@ccdc.cam.ac.uk.
2.7. Docking simulations
The crystal structures of both the ATPasic and the cleavage domains
of human TopoIIa were retrieved from the Protein Data Bank (PDB) [21]
and used after deletion of all water molecules and other ligands, but
keeping the DNA chains. The structure of the ATP binding domain was
used in the dimeric form (PDB ID: 1zxm [22]), while the structure of
cleavage domain is available only in monomeric form (PDB ID: 4fm9
[23]).
2.6. Biological experiments
2.6.1. Materials
For ligands HL2, HL3, and HL5 and complexes [Cu(L2)]+ and
[Cu(L3)]+ the crystal structures determined from our structural studies
were used. The structure of ANP (phosphoaminophosphonic acid-
adenylate ester), used as a control, was extracted from the ATP binding
domain crystal structure. ANP is an ATP molecule in which the oxygen
For cell cultures, RPMI 1640 medium, fetal bovine serum (FBS),
penicillin/streptomycin, L-glutamine, and phosphate buffered saline
(PBS) were purchased from Euroclone (Milan, Italy) and flasks/plates
from Costar, Corning (Amsterdam, The Netherlands).

14
F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
between Pβ and Pγ is substituted by a nitrogen to avoid the hydrolysis
of the molecule and to make possible the crystallization of the complex,
The structures of the other compounds were built by means of
Hyperchem8 [24] and then refined by minimization with the Avogadro
v. 1.1.0 [25] software. Both the E and the Z configurations of the ligands
were built and used in the simulations.
more favorable E configuration. When the terminal nitrogen is completely
substituted and becomes a tertiary amine, as in compound HL2, it is the
hydrazinic hydrogen that forms a hydrogen bond with the quinoline ni-
trogen, forcing the C=N bond to a Z configuration (Fig. 3). This trend is
corroborated also by the structure of the quinoline-2-
carboxaldehyde thiosemicarbazone reported in the literature [35].
As far as the packing of the above mentioned ligands are concerned,
it is interesting to observe how each one presents a peculiar hydrogen
bond that characterizes the packing. Ligand HL2 shows an intra-
molecular H-bond between the hydrazinic NH and the quinoline nitro-
gen, in ligand HL3 the H-bond involves the same atoms but belonging to
two different adjacent molecules, while L5 uses the NH of the terminal
amino group to form the H-bond with the quinoline nitrogen of a
nearby molecule. Weak van der Waals interactions between adjacent
molecules contribute to the packing and only molecules HL3 and L5
show intermolecular stacking interactions between the aromatic
system of the quinoline and the thiosemicarbazide moiety.
The introduction of a coordinating metal ion affects dramatically the
configuration of these ligands, since the energy implied in coordination
prevails on that of the hydrogen bonds found in the ligand molecules
and complexation imposes an E configuration on the ligands by forming
three coordinative bonds, i.e. with the quinoline nitrogen, the imino ni-
trogen and the sulfur.
In the case of the copper derivatives the syntheses afforded com-
plexes of stoichiometry M:L 1:1. The ligand chelates the metal with its
three donor atoms which are sterically more favored and the fourth
position is occupied by a chloride ion that completes the square planar
geometry.
In complexes [Cu(L1)Cl], [Cu(L2)Cl], [Cu(L3)Cl] and [Cu(L4)Cl],
the deprotonation causes the negative charge to delocalize partly on
the nitrogen and partly on the sulfur, as can be inferred by the bond
lengths observed in the two structures characterized through XRD.
The distances C1–N2 become shorter passing from 1.359 Å and 1.356 Å
in the ligands to 1.342 Å and 1.324 Å in the complexes, and contempo-
rarily the C1–S1 bond changes from 1.687 Å and 1.677 Å in the ligands
to 1.750 Å and 1.746 Å in the complexes. The crystal structures of com-
pounds [Cu(L2)Cl] and [Cu(L3)Cl] are reported in Fig. 4. In the structures,
the two copper derivatives are characterized by a distorted square
planar coordination geometry where the chlorine atom is markedly
out of the plane (the Cu–Cl bond forms angles of 21.05° and 28.91°
with the coordination plane, in [CuL2Cl] and [CuL3Cl] respectively).
Moreover, the complex molecules present, pairwise, copper–sulfur
interactions of 2.82 and 2.86 Å, respectively in which the sulfur occupies
the apical position of a square pyramid centered on the metal ion.
Neither stacking interactions nor hydrogen bonds are present in the
packing. In [Cu(L5)Cl2], where the ligand cannot undergo deprotonation,
the compound contains a further chloride ion to compensate the missing
negative charge.
Differently from copper, solutions of nickel(II) chloride added of
each of the ligands have afforded complexes with stoichiometry M:L
1:2. The complexes present a nickel ion in an octahedral environment
with two terdentate ligands perpendicular to each other. The molecular
structures of compounds [Ni(L1)(HL1)]Cl, [Ni(L3)2] and [Ni(L5)2]Cl2
are reported in Fig. 5. The nickel derivatives present the most diverse
packing systems. The crystals of [Ni(L1)(HL1)]Cl·CH3OH contain a
chloride ion and a methanol. The chloride plays a relevant role in deter-
mining the packing through a complex H-bond system. It is involved in
The docking simulations were performed with the Autodock4 soft-
ware package, with the aid of the AutoDockTools (ADT) interface [26].
The Kollmann partial charges [27] were added to the receptors by
means of ADT. The partial charges were added to the compounds by
means of the UCSF Chimera package [28] using the Antechamber mod-
ule [29] and the Amber ff12SB force field. After the deletion of chlorine,
the copper atom was replaced with iron, because Autodock4 does not
contains the docking potential parameters for copper; we exploited
the similar ionic radius presented by the two ions in the oxidation
state of +2. The charge on the deprotonated ligand was −1
(delocalized partly on the nitrogen and partly on the sulfur, as sug-
gested by the experimental results), therefore giving a net charge of
+1 to the complex. The only exception in the series was complex
[Cu(L5)]2+, in which the presence of a methyl group bound to the
hydrazinic nitrogen prevents the deprotonation of the ligand and
confers a net charge of +2 to the whole complex.
After an initial check, all torsions of both the receptors and the com-
pounds were kept fixed. The grid maps centered on the ATP binding
pocket had a box size of 72 × 66 × 64 points, spaced 0.375 Å and the
one at the cleavage site had a box size of 62 × 68 × 72 points, with the
same spacing. For the docking calculations, the Lamarckian Genetic Al-
gorithm [30] was used, performing 100 runs with 150 individuals in
the population, 27,000 generations and 2.5 × 106 energy evaluations.
A cluster analysis was performed and the ligand–receptor complexes
with the lowest binding energy and/or in the more populated cluster
were chosen for the analysis.
Structural analysis of the binding modes was made using ADT, VMD
[31] and Swiss-PdbViewer [32] software and the web tool for macro-
molecular visualization FirstGlance in Jmol (http://bioinformatics.org/
firstglance/fgij/index.htm).
3. Results and discussion
3.1. Synthesis and characterization
In a paper published by Huang et al. in 2010 [33], among a variety
of compounds assayed in search for TopoIIa inhibitors, quinoline-2-
carboxaldehyde N4-N4dimethylthiosemicarbazone was observed to
possess the highest activity, with IC50 between 0.02 and 0.08 μM,
depending on the tumor cell line. In this study we have synthesized
derivatives modified on the thiosemicarbazide nitrogens and their coor-
dination compounds with copper(II) and nickel(II) in order to explore
the influence of the substituents, of the metal ions and of the molecular
shape on the overall biological behavior. More specifically, we have
tested the capacity of these molecules to inhibit one of the putative
targets of thiosemicarbazones in cells, TopoIIa, and to propose a model
that could explain their different activities.
All five ligands were synthesized and characterized according to the
procedures reported in the experimental section and for three of them
the XRD structure has also been determined. All these molecules are
planar due to an extended system of conjugated double bonds and to
a thione-thiol tautomerism of the thiosemicarbazone fragment (see
Fig. 2). The configuration of the molecules with respect to the imino
double bond in solution easily undergoes E and Z interconversion [34],
but in the solid state only one of the two conformations is observed.
The determining factor seems to be, besides possible packing interac-
tions, the pattern of intramolecular hydrogen bonds that the molecule
can form. For instance, if at least one hydrogen atom is present on the ter-
minal nitrogen, an intramolecular hydrogen bond with the imino nitro-
gen is usually present. This pattern is usually associated with a sterically
Fig. 2. Thiosemicarbazide thione/thiol tautomerism equilibrium.

F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
15
Fig. 3. ORTEP drawings of compound HL2, HL3 and L5 with 50% probability ellipsoids. Intramolecular hydrogen bonds are represented with dashed lines.
three hydrogen bonds: two with the terminal NH of two different
complexes and a third with the methanol OH. One of the terminal NH
also binds the methanol oxygen. In addition, the aromatic rings form
loose dimers through parallel and perpendicular interactions of the
pi-systems. Compound [Ni(L3)2] contains neither counterions nor
solvent and its structure is characterized by helices formed by roto-
translation of the asymmetric unit, where the molecules are joined
together by hydrogen bonds between the terminal NH, which acts as a
donor, and the non-coordinated hydrazinic nitrogen. Another peculiar
feature presented by these molecules is the short contact between the
terminal methyl and the pi-system of the quinoline moiety of an adja-
cent molecule. The packing of [Ni(L5)2]Cl2·4.5H2O is the most complex
one, as can be predicted by the presence of two chloride ions and by the
large number of water molecules. The structure is characterized by an
extensive system of hydrogen bonds connecting the water molecules,
the chloride ions and the terminal NH2 of the ligand. The pi–pi interac-
tions are limited to a single weak contact of two aromatic systems at
3.599 and 3.593 Å. Figs. 1S to 6S in the Supplementary material show
the described interactions.
the main variables that can influence the biological activity of these
compounds. It is worth noting here that equivalent amounts of
copper(II) and nickel(II) salts do not have an influence on cell prolifer-
ation [36] and copper(II) on topoisomerase IIa inhibition [20].
Cell viability was checked after a 24 h exposure to increasing con-
centrations of ligands and complexes (range 0.1–100 μM) (Table 1).
From these IC50 values it is possible to do some considerations. If we
compare each ligand with the corresponding metal complexes, we can
observe that the copper derivatives are systematically more active
than both the ligands and the analogous nickel complexes. The IC50
values for the copper(II) derivatives are in the range 0.48 μM (compound
[Cu(L2)Cl]) to 16.2 μM (compound [Cu(L5)Cl2]). In this latter case, it
is noteworthy that the parent ligand (L5), and the analogous Ni(II)
complex ([Ni(L5)2]Cl2), do not show any activity up to a 100 μM concen-
tration suggesting that the disappearance of the thione/thiol tautomeric
equilibrium plays a relevant role. Also the substituents seem to play a
relevant role in the potency of the compounds. The ligands and com-
plexes with the doubly methylated terminal nitrogen are the most effec-
tive of the series (ligand HL2: IC50 = 0.82 μM, Cu(II) complex [Cu(L2)Cl]:
IC50 = 0.48 μM and Ni(II) complex [Ni(HL2)2]Cl2: IC50 = 8.69 μM). The
copper complexes with ligands possessing a partial substitution on the
3.2. Biology
very same nitrogen either with a single methyl ([Cu(L3)Cl]: IC50
=
All the compounds reported, ligands and complexes, were tested for
their capacity to inhibit proliferation in cell line U937 and assayed
on TopoIIa. The differences in the substitution on the ligands and the
nature of the metals, together with the shape of their complexes, are
2.25 μM) or with a phenyl ([Cu(L4)Cl]: IC50 = 1.8 μM) do not present
a significant difference in their IC50 values. Compound L5 does not
show any activity up to a concentration of 100 μM. The related copper
complex [Cu(L5)Cl2] is the least active among the analogous tested
Fig. 4. ORTEP plots of two copper(II) derivatives, compound [Cu(L2)Cl] and [Cu(L3)Cl], with 50% probability ellipsoids.

16
F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
Fig. 5. ORTEP plots of three nickel(II) derivatives, compound [Ni(L1)(HL1)Cl], [Ni(L3)2] and [Ni(L5)2Cl2], with 50% probability ellipsoids.
compounds, with an IC50 of 16.2 μM; it is therefore apparent that the
methyl on the N2 position seems to reduce the efficacy of this com-
pounds. In addition, its higher positive charge could interfere with its
ability to cross the cell membrane.
Molecular docking simulations were performed on the compounds
to analyze their interactions with both the ATPasic domain, led by
its structural similarity with ATP, and at the cleavage domain, led by a
putative analogy with etoposide-like inhibitors.
As far as the TopoIIa tests are concerned, the results are quite inter-
esting. In fact, we find an overall correlation between proliferation inhi-
bition and TopoIIa inhibition: the copper derivatives are outstandingly
more active than all the other compounds (Fig. 7S).
The simulations were performed on the ligands (both in the E and in
the Z conformation) and on the copper complexes, after a substitution of
the copper atom with iron (see Experimental section).
3.3.1. ATPasic domain
3.3. Docking simulations
At first, we performed a control simulation with the ANP molecule,
that was present in the binding site of the crystal structure of TopoIIa
(PDB ID: 1ZXM, see Experimental section). Our control simulation re-
trieved for ANP the same identical position that it has in the crystal,
checking the reliability of the docking software and parameters and per-
mitting the estimation of a reference binding energy for the endogenous
ligand (Eb = −13.8 kcal/mol) that could be used for comparison.
The docking simulations of all the ligands in both E and Z conforma-
tions show a binding mode inside the ATP pocket, in the same place of
ANP/ATP. Nevertheless, their binding energies were too high to compete
with the endogenous ligand (ranging from −4.13 to −3.01 kcal/mol).
On the contrary, the complexes with metal do not find any binding
site at the ATP place, but they all bind externally, in the same place at
the edge of the pocket (Fig. 6). In fact, they find a very favorable site
for Cu2+ in the loop of the pocket lid, that contains four consecutive
acidic residues (152–155) creating a strong negative potential region.
The metal complexes [Cu(L1–L4)]+ link the loop of the lid with the
The structure of TopoIIa can be divided in three distinct domains.
Two of them host two functionally different active sites: the C-terminal
domain with the ATP-binding site and the central domain with the
catalytic DNA cleavage-religation site [14]. The ATP-binding domain
possesses a lid whose opening mechanism permits not only the binding
of ATP and its hydrolysis, but also the dimerization of this N-terminal
region [22,37]. Interestingly, the dimer presents a domain swapping
involving a little helix, named “strap” helix, that positioning over the
lid that closes the ATP-binding site, stabilizes the closed conformation
[38].
The DNA cleavage-religation domain, which is already in the dimeric
form, binds DNA, cuts it and re-ligates it, after the passage of a second
DNA strand. It requires for its function two Mg2+ ions that, binding to
a cluster of acidic residues (Asp541, Asp543, Glu461) and to a phos-
phate, are involved in the cutting and anchoring of DNA [37,39].
Table 1
Proliferation inhibition concentrations, IC50, (the compound concentrations used in the test were 0.01, 0.05, 0.1, 0.5, 1, 2, 5, 10, 20 50 and 100 μM) and TopoIIa inhibition assay outcome
(the compound concentration used for the topoisomerase IIa inhibition test following the Topogen® kit protocol is 50 μM).
Ligands
IC50 (μM)
TopoIIa inhibition assay
Cu(II) derivatives
IC50 (μM)
TopoIIa inhibition assay
Ni(II) derivatives
IC50 (μM)
TopoIIa inhibition assay
HL1
HL2
HL3
HL4
L5
55
0.82
N100
12.05
N100
−
−
−
−
−
[Cu(L1)Cl]
[Cu(L2)Cl]
[Cu(L3)Cl]
[Cu(L4)Cl]
[Cu(L5)Cl2]
4.1
0.48
2.25
1.8
+
+
+
+
+
[Ni(L1)(HL1)Cl]
[Ni(HL2)2Cl2]
[Ni(L3)2]
[Ni(L4)2]
[Ni(L5)2Cl2]
42
−
−
−
−
−
8.69
N100
N100
N100
16.2

F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
17
modes, with higher binding energies than the metal complexes
(ranging from −5.18 to −3.09). All metal complexes, instead, adopt
two equally favorable binding modes, both involving the interaction of
the metal with the oxygen atoms of the last phosphate at the edge of
the cut. We have to pinpoint that, since the available crystal structure
of TopoIIa was monomeric, we used the superimposition with the
dimeric structure of topoisomerase IIb (PDB ID: 3QX3 [15]) to check
the hindrance of the other protomer containing the other DNA fragment
at the opposite edge of the cut.
In the first binding mode the molecules are oriented almost parallel
to the DNA bases, in an etoposide-like manner, with the rings in contact
with the sugar of the first base and the tail in contact with the pyrimi-
dine ring of the second base (Fig. 7a). In particular, the interaction of
the second base with the methyl groups of complexes [Cu(L2)]+ and
[Cu(L3)]+ are favored if compared with complex [Cu(L1)]+; while
complex [Cu(L4)]+, that has the phenyl ring in stacking with the base
ring, shows the lower, i.e. more favorable, binding energy (Table 2).
Complex [Cu(L5)]2+ is the only one for which this binding mode is
not present.
The second binding mode involves the other oxygen of the last
phosphate, but the conformations are at the external edge of the DNA
strand, almost perpendicular to the basis plane (Fig. 7b). As in the
ATPasic domain, complex [Cu(L5)]2+ inverts its orientation and has a
lower binding energy than the other complexes. In this case, it is com-
plex [Cu(L1)]+ that does not assume this position.
Fig. 6. Best docked positions of metal complexes in the ATPasic domain. Complex
[Cu(L1)]+ = blue, [Cu(L2)]+ = red, [Cu(L3)]+ = gray, [Cu(L4)]+ = yellow, [Cu(L5)]2+
=
green. The metal is in a spacefill representation, in the same color as the ligand. The ATP
pocket (orange) with the lid loop (cyan) and the strap helix of the opposite chain are in
opaque cartoon representation. All the rest of the dimeric molecule (gray) is in transparent
cartoon. The main interacting residues (in ball and stick) are labeled. The crystal position of
ANP is shown in purple.
The energies of the two binding modes are comparable. As it happens
in the ATPasic domain, the complexes [Cu(L4)]+ and [Cu(L5)]2+ show
the lowest binding energies. Probably, the higher net charge of the first
compound and the hydrophobic phenyl ring of the latter, give a strong
contribution to the binding.
In neither of the two binding modes, the metal is in the place of
Mg2+ (i.e. coordinated by the acidic cluster) preferring instead the
DNA phosphate, probably because of the more favorable hydrophobic
interactions that can be formed. This fact suggests a second putative
inhibition mechanism that prevents the re-ligation of DNA, with the
formation of cleavage intermediates resulting in being toxic for cells.
strap helix, by means of their hydrophobic rings. Among them, complex
[Cu(L4)]+ has the most favorable binding energy, due to its phenyl ring
in the R2 position, forming additional hydrophobic interactions with the
carbons of the Arg98 side chain. Complex [Cu(L5)]2+, that has the
methyl group in R1 position, inverts its orientation in a way that makes
the CH3 interact with Ile33 on the strap helix while the rest of the tail
of the molecule can make H-bonds with Arg32 and Gln35 (Fig. 6).
This position and the particular charge distribution of this complex
(see Experimental section) produce a net gain in binding energy of com-
plex 10 in respect to all the other metal complexes (Table 2).
It is worthwhile to note that, even if ATP has a 2+ ion (Mg2+) that
could also bind in the site of the metal complex, it has nevertheless a
highly negative net charge, due to the phosphate groups, that fits well
in its very basic binding pocket; the positive charge of the pocket is
instead not favorable to accommodate our metal complexes, that have
a net charge of +1 or +2.
Observing the binding mode of the metal complexes in the ATPasic
domain, we can hypothesize a first inhibition mechanism that, linking
the strap helix to the lid, could prevent the release of the monomeric
form of this domain and the restart of the enzyme cycle.
4. Conclusions
A series of quinoline-2-carboxaldehyde thiosemicarbazones, differ-
ing in the substituents of the thiosemicarbazide moiety, were synthe-
sized and characterized. In all complexes the ligands are in the E
configuration with respect to the imino bond and behave as terdentate.
Copper(II) complexes present a metal:ligand ratio 1:1 and form square
planar derivatives with one molecule of ligand and chloride ions to
compensate the overall charge. Independently from the stoichiometric
amounts used, nickel(II) ions form, instead, octahedral complexes with
two ligands for each metal ion. Ligands and complexes were tested
for their antiproliferative properties on histiocytic lymphoma cell line
U937. As already observed by Zeglis et al. [20] metal complexes are
more active than the corresponding ligands, and copper compounds
are the most active. All the compounds were tested for their specific
activity on TopoIIa and among them only copper derivatives inhibit
the enzyme in vitro. Computational methods have been used to propose
a model that explains the different extent of inhibition presented by
these compounds. Both Huang and co-workers [33] and Zeglis and co-
workers [20] propose a catalytic inhibition mechanism, that is to say,
an interaction of these compounds with the enzyme through the ATP
pocket, but the first author tests only the ligands as probes, while the
second one assumes the same mechanism for the metal complexes
without exploring other possible target sites in the enzyme. A feature
of these copper complexes is certainly the possibility to dissociate the
chloride ion and give rise to positively charged metal ions as soon as
they enter the cytoplasm, a mechanism already proposed for cisplatin
[40]. We propose that the positive charge of the dissociated form of
the copper complexes may play a key role in their action. The charge
3.3.2. Cleavage domain
The docking simulations on the cleavage domain show again for the
ligands, in both E and Z conformations, spread and aspecific binding
Table 2
Binding energies (kcal/mol) of metal complexes in the ATPasic and cleavage domains. For
the latter, the energies of the two preferential binding mode are reported, i.e. parallel (//)
and perpendicular (⊥) to the DNA base plane.
Cu(II)
derivatives
ATPasic domain Cleavage domain
Binding energy Binding energy
Cleavage domain
Binding energy
(kcal/mol)
(kcal/mol)
(kcal/mol)
First binding mode (//) Second binding mode (⊥)
[Cu(L1)]+ −6.6
[Cu(L2)]+ −6.7
[Cu(L3)]+ −6.7
[Cu(L4)]+ −8.3
[Cu(L5)]2+ −8.2
−6.1
−6.5
−6.5
−7.0
–
–
−6.4
−6.0
−7.1
−8.4

18
F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
Fig. 7. Best docked positions of metal complexes in the cleavage domain. (a) First binding mode, parallel to the bases rings; (b) second binding mode, perpendicular to the bases rings.
Complex [Cu(L1)Cl] = blue, [Cu(L2)]+ = red, [Cu(L3)]+ = gray, [Cu(L4)]+ = yellow, [Cu(L5)]2+ = green. The metal atom is in a space-fill representation, in the same color as the ligand.
In black is the putative position of the Mg2+ ion, coordinated by the acidic cluster (red sticks). The protein is in transparent cartoon, colored by secondary structure (beta strands = yellow;
alpha-helices = purple; 310-helices = blue; turns = cyan). The DNA fragment (gray) is in opaque sticks representation, with the phosphate group colored by element.
drives the metal complex to find its way in the enzyme structure and
binds more favorably in negatively charged sites.
Acknowledgments
In our calculations we observe that the ligands, in both their E and Z
forms, do enter in the ATP pocket and bind to the site, but with much
lower binding energies that the endogenous ligand. What we observe
with the complexes is something different: the metal derivatives tend
to get stuck between the lid and the strap helix, binding to an acidic
cluster. The access into the ATP pocket is made even more difficult by
the presence of basic residues which are ideal for attracting a negatively
charged ATP molecule. Compound [Cu(L5)Cl2] is the least active of the
copper compounds as far as the IC50 is concerned but a good inhibitor
of TopoIIa both in vitro and in silico. This peculiar behavior can be easily
explained taking into account that the methylation on the N2 nitrogen
prevents the deprotonation of the ligand and confers to the complex,
differently from the other copper compounds which are neutral,
[Cu(L)Cl], a cationic nature, [Cu(L5)Cl]+, and this could certainly
account for a possible hindrance to the access into the cell by passive
diffusion.
Thanks are due to the Centro Interdipartimentale Misure “G. Casnati”
of the University of Parma for instrumental facilities.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jinorgbio.2015.08.008.
References
[1] Y. Yu, D.S. Kalinowski, Z. Kovacevic, A.R. Siafakas, P.J. Jansson, C. Stefani, D.B. Lovejoy,
P.C. Sharpe, P.V. Bernhardt, D.R. Richardson, J. Med. Chem. 52 (2009) 5271–5294.
[2] D.X. West, A.E. Liberta, S.B. Padhye, R.C. Chikate, P.B. Sonawane, A.S. Kumbhar, R.G.
Yerande, Coord. Chem. Rev. 123 (1993) 49–71.
[3] F. Tisato, C. Marzano, M. Porchia, M. Pellei, C. Santini, Med. Res. Rev. 30 (2010)
708–749.
[4] G. Pelosi, Open Crystallogr. J. 3 (2010) 16–28.
We docked our molecules also in the cleavage site, finding again for
the ligands a spread and aspecific binding, with higher energies than the
complexes and comparable with those at the ATP site. Considering the
features of the binding, the ATP domain seems a more reliable target
for the ligands, in agreement with what was found by Huang and co-
workers [33]. The copper compounds, instead, strongly interact with
the DNA phosphate group at the cut edge, likely preventing the re-
ligation to the opposite strand. Again, compound [Cu(L5)]2+ has the
most favorable energy, probably due to its cationic feature that, in this
case, favors the binding at the DNA negatively charged group. In both
domains, compound [Cu(L1)]+ has the worst energy, confirming the
relevant role of substitutions in enhancing the binding affinity.
Comprehension of the inhibition mechanisms, based on the struc-
tural characterization, and the relevance of the Cu role, could lead to a
new approach for the synthesis of differently substituted
thiosemicarbazones and for the study of their biomedical application
as anticancer drugs.
[5] D.S. Kalinowski, P. Quach, D.R. Richardson, Future Med. Chem. 1 (2009) 1143–1151.
[6] H. Beraldo, D. Gambino, Mini-Rev. Med. Chem. 4 (2004) 31–39.
[7] A.C. Sartorelli, B.A. Booth, Cancer Res. 27 (1967) 1614–1619.
[8] G.A. Kune, Br. Med. J. 2 (1964) 621.
[9] J.F. Zeidner, J.E. Karp, A.L. Blackford, B.D. Smith, I. Gojo, S.D. Gore, M.J. Levis, H.E.
Carraway, J.M. Greer, S.P. Ivy, K.W. Pratz, M.A. McDevitt, Haematologica 99 (2014)
672–678.
[10] F. Bisceglie, S. Pinelli, R. Alinovi, M. Goldoni, A. Mutti, A. Camerini, L. Piola, P.
Tarasconi, G. Pelosi, J. Inorg. Biochem. 140 (2014) 111–125.
[11] A. Buschini, S. Pinelli, C. Pellacani, F. Giordani, M.B. Ferrari, F. Bisceglie, M. Giannetto,
G. Pelosi, P. Tarasconi, J. Inorg. Biochem. 103 (2009) 666–677.
[12] M.B. Ferrari, F. Bisceglie, G. Pelosi, P. Tarasconi, R. Albertini, P.P. Dall'Aglio, S. Pinelli,
A. Bergamo, G. Sava, J. Inorg. Biochem. 98 (2004) 301–312.
[13] A.K. Larsen, A. Skladanowski, K. Bojanowski, Prog. Cell Cycle Res. 2 (1996) 229–239.
[14] A.K. McClendon, N. Osheroff, Mutat. Res. Fundam. Mol. Mech. Mutagen. 623 (2007)
83–97.
[15] C.C. Wu, T.K. Li, L. Farh, L.Y. Lin, T.S. Lin, Y.J. Yu, T.J. Yen, C.W. Chiang, N.L. Chan, Science
333 (2011) 459–462.
[16] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C. Giacovazzo, A. Guagliardi,
A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr. 32 (1999) 115–119.
[17] G. Sheldrick, SHELXL 97 A Program for Structure Refinement, University of
Göttingen, Göttingen, 1997.
[18] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837–838.
[19] L.J. Farrugia, J. Appl. Crystallogr. 45 (2012) 849–854.
Abbreviations
[20] B.M. Zeglis, V. Divilov, J.S. Lewis, J. Med. Chem. 54 (2011) 2391–2398.
[21] H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H. Weissig, I.N.
Shindyalov, P.E. Bourne, Nucleic Acids Res. 28 (2000) 235–242.
[22] H. Wei, A.J. Ruthenburg, S.K. Bechis, G.L. Verdine, J. Biol. Chem. 280 (2005) 37041–37047.
[23] T.J. Wendorff, B.H. Schmidt, P. Heslop, C.A. Austin, J.M. Berger, J. Mol. Biol. 424
(2012) 109–124.
RPMI
FBS
Roswell Park Memorial Institute medium;
fetal bovine serum;
PBS
PHosphate buffered saline;
[24] HyperChem (Molecular Modeling System) Hypercube, Inc, 1115 NW, 4th Street,
Gainesville, FL 32601, USA, 2007 http://www.hyperchem.com/.
[25] M.D. Hanwell, K. Lutz, Abstr. Pap. Am. Chem. Soc. 244 (2012).
[26] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson,
J. Comput. Chem. 16 (2009) 2785–2791.
MTS
3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-
2-(4-sulfophenyl)-2H-tetrazolium
Sodium Dodecyl Sulfate
SDS
ANP
phosphoaminophosphonic acid-adenylate ester

F. Bisceglie et al. / Journal of Inorganic Biochemistry 152 (2015) 10–19
19
[27] S.J. Weiner, P.A. Kollman, D.A. Case, U.C. Singh, C. Ghio, G. Alagona, S.J. Profeta, P.
Weiner, J. Am. Chem. Soc. 106 (1984) 765–784.
[28] E.F. Pettersen, T.D. Goddard, C.C. Huang, G.S. Couch, D.M. Greenblatt, E.C. Meng, T.E.
[35] P.N. Bourosh, M.D. Revenko, M. Gdaniec, E.F. Stratulat, Y.A. Simonov, J. Struct. Chem.
50 (2009) 510–513.
[36] F. Bisceglie, R. Alinovi, S. Pinelli, M. Goldoni, A. Buschini, S. Franzoni, A. Mutti, P.
Tarasconi, G. Pelosi, Metallomics 5 (2013) 1510–1518.
[37] J.E. Deweese, N. Osheroff, Metallomics 2 (2010) 450–459.
[38] S. Classen, S. Olland, J.M. Berger, Proc. Natl. Acad. Sci. U. S. A. 100 (2003)
10629–10634.
Ferrin, J. Comput. Chem. 25 (2004) 1605–1612.
[29] J. Wang, W. Wang, P.A. Kollman, D.A. Case, J. Mol. Graph. Model. 25 (2006) 247–260.
[30] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew, A.J. Olson,
J. Comput. Chem. 19 (1998) 1639–1662.
[31] W. Humphrey, A. Dalke, K. Schulten, J. Mol. Graph. Model. 14 (1996) 33–38.
[32] N. Guex, M.C. Peitsch, Electrophoresis 18 (1997) 2714–2723.
[33] H. Huang, Q. Chen, X. Ku, L. Meng, L. Lin, X. Wang, C. Zhu, Y. Wang, Z. Chen, M. Li, H.
Jiang, K. Chen, J. Ding, H. Liu, J. Med. Chem. 53 (2010) 3048–3064.
[34] A.T. Ota, M.L.A. Temperini, E.P.G. Areas, M. Loos, THEOCHEM J. Mol. Struct. 451
(1998) 269–275.
[39] C.C. Chang, Y.R. Wang, S.F. Chen, C.C. Wu, N.L. Chan, Curr. Opin. Struct. Biol. 23
(2013) 125–133.
[40] N.D. Eljack, H.Y. Ma, J. Drucker, C. Shen, T.W. Hambley, E.J. New, T. Friedrich, R.J.
Clarke, Metallomics 6 (2014) 2126–2133.