DNA-binding and in vitro cytotoxic activity of platinum(II) complexes of curcumin and caffeine

Article abstract of DOI:10.1016/j.jinorgbio.2019.110749

Three Pt(II) complexes containing the natural ligands curcumin and caffeine, namely [Pt(curc)(PPh₃)₂]Cl (1), [PtCl(curc)(DMSO)] (2) (curc = deprotonated curcumin) and trans-[Pt(caffeine)Cl₂(DMSO)] (3), were synthesized and

Full text of DOI:10.1016/j.jinorgbio.2019.110749

Accepted Manuscript
DNA-binding and in vitro cytotoxic activity of platinum(II)
complexes of curcumin and caffeine
Valentina Censi, Ana B. Caballero, Marta Pérez-Hernández,
Vanessa Soto-Cerrato, Luís Korrodi-Gregório, Ricardo Pérez-
Tomás, Maria Michela Dell'Anna, Piero Mastrorilli, Patrick
Gamez
PII:
S0162-0134(19)30086-8
DOI:
https://doi.org/10.1016/j.jinorgbio.2019.110749
Article Number:
Reference:
110749
JIB 110749
To appear in:
Journal of Inorganic Biochemistry
Received date:
Revised date:
Accepted date:
12 February 2019
3 June 2019
5 June 2019
Please cite this article as: V. Censi, A.B. Caballero, M. Pérez-Hernández, et al., DNA-
binding and in vitro cytotoxic activity of platinum(II) complexes of curcumin and caffeine,
Journal of Inorganic Biochemistry, https://doi.org/10.1016/j.jinorgbio.2019.110749
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.

ACCEPTED MANUSCRIPT
DNA-binding and in vitro cytotoxic activity of platinum(II)
complexes of curcumin and caffeine
d
Valentina Censi a,b, Ana B. Caballero a,c*, Marta Pérez-Hernández , Vanessa Soto-
Cerrato d, Luís Korrodi-Gregório d, Ricardo Pérez-Tomás d, Maria Michela Dell´Anna b,
Piero Mastrorilli b*, Patrick Gamez a,c,e*
a
Department of Inorganic and Organic Chemistry, Inorganic Chemistry Section,
University of Barcelona, Martí i Franquès 1-11, 08028 Barcelona, Spain
b DICATECh, Politecnico di Bari, via Orabona, 4 70125 Bari, Italy
c
Institute of Nanoscience and Nanotechnology (IN2UB), Universitat de Barcelona,
Spain
d
Department of Pathology and Experimental Therapeutics, Faculty of Medicine,
University of Barcelona, Campus Bellvitge, Feixa Llarga s/n, 08907 L'Hospitalet de
Llobregat, Spain
e
Catalan Institution for Research and Advanced Studies (ICREA), Passeig Lluís
Companys 23, 08010 Barcelona, Spain
Keywords: Platinum(II) complex ─ natural ligands ─ caffeine ─ curcumin ─ DNA
interaction ─ cytotoxicity ─ photoactivation
Abstract
Three Pt(II) complexes containing the natural ligands curcumin and caffeine, namely
[Pt(curc)(PPh3)2]Cl (1), [PtCl(curc)(DMSO)] (2) (curc = deprotonated curcumin) and
trans-[Pt(caffeine)Cl2(DMSO)] (3), were synthesized and fully characterized. The data
obtained suggest that, for both 1 and 2, the anion of curcumin is coordinated to the
platinum ion via the oxygen atoms of the-diketonate moiety. Spectroscopic features
reveal that in 2 and 3, a DMSO molecule is S-bonded to the metal centre. For 3, all data
indicate a square-planar geometry formed by a 9-N bonded caffeine, two trans chloride
anions and a DMSO. The three complexes undergo changes in solution upon incubation
for 24 h; 1 and 2 release curcumin while 3 isomerizes from trans to cis configuration.
The DNA-binding and cytotoxic properties of 1−3 were evaluated in vitro. Despite their
structural similarity, curcuminate-containing 1 and 2 exhibit distinct DNA interactions.
While 1 appears to intercalate between nucleobase pairs, inducing the oxidative
degradation of the biomolecule, 2 behaves as a groove binder, by means of electrostatic
forces. Caffeine-containing 3 exhibits a behaviour that is comparable to that of 2.
Complexes 1 and 2 showed moderate to high cytotoxicity and selectivity against several
cancer cell lines, while 3 is inactive. Compounds 1 and 2 can be further activated by
visible-light irradiation.

ACCEPTED MANUSCRIPT
1. Introduction
The International Agency for Research on Cancer (IARC) estimated that 14.1 million
new cancer cases and 8.2 million cancer deaths (which correspond to about 22,000
cancer deaths a day) occurred in 2012.[1] These figures are expected to grow
significantly, so that 21.7 million new cancer cases and 13 million cancer deaths will be
witnessed by 2030.[2] Thus, cancer is a major public health problem worldwide and
investigation is definitively required to develop new efficient drugs.[3-6]
Cisplatin[7] is one of the most used anticancer agents worldwide.[8] However, it
exhibits several drawbacks, e.g. its use is limited to a restricted number of cancers,[9]
acquired or intrinsic cisplatin resistance may be shown by some tumors,[10] and it
causes severe side effects, such as nausea, nephrotoxicity, ototoxicity or
myelosuppression.[11, 12] Therefore, the design of new platinum-based compounds is a
hot topic of current research in the field; actually, a number of platinum drugs have been
developed and marketed, viz. the second-generation platinum chemotherapy agents
carboplatin and nedaplatin, and the third-generation complexes oxaliplatin and
lobaplatin.[13-16]
Naturally occurring ligands, especially those approved by the Food and Drug
Administration (FDA) for diverse applications, represent molecules of choice for the
preparation of metal-based compounds, as they may accelerate the procedure to reach
marketing approval for an active compound. In this context, a number of metal
complexes based on natural ligands have been described in the literature; for example,
coordination compounds from curcumin (1,7-bis-(4-hydroxy-3-methoxyphenyl)-1,6-
heptadiene-3,5-dione) have been reported for medicinal applications.[17, 18] Caffeine-
derived complexes of rhodium(I) or gold(I) have shown promising anticancer
properties.[19]
In fact, curcumin displays a wide therapeutic profile as it has been reported for its
antioxidative, anti-inflammatory, antiviral, antibacterial, antifungal, and anticancer
properties.[20-22] However, curcumin possesses serious disadvantages that hinder its
use as a therapeutic agent: low aqueous solubility, rapid metabolism leading to
degradation, and reduced bioavailability.[23] Caffeine, apart from its known anti-
inflammatory activity, modulates the antitumor activity of several drugs; it is thought to
increase the antitumor effect of cisplatin or DNA-damaging agents by inhibiting DNA
repair.[24-26]
In the search for new platinum-based anticancer therapeutics, DNA is still the first
selected target to test the therapeutic potential of newly synthesized compounds. This is
because of the structural resemblance of Pt(II) complexes with its well-known, DNA-
targeted predecessors cisplatin and oxaliplatin, which are DNA binders producing
unrepairable damages to the biomolecule that lead to cell death.
In the present study, curcumin and caffeine have been selected as natural and bioactive
building-blocks to obtain three new platinum(II) complexes with in vitro
antiproliferative activity against some cancer cell lines. Reports have shown that the
coordination of these bioactive molecules to d[27][26][25]- and f-block metal ions can
greatly improve their aqueous solubility, bioavailability, and antitumor activity.[25, 27,
28] Hence, the complexes [Pt(curc)(PPh3)2]Cl (1) and [PtCl(curc)(DMSO)] (2) (curc =
deprotonated curcumin), in which the anion of curcumin should act as a leaving group
once the complexes enter the cells, and trans-[Pt(caffeine)Cl2(DMSO)] (3) were

ACCEPTED MANUSCRIPT
prepared and fully characterized. As carrier ligands we used triphenylphosphane (PPh3)
and dimethylsulfoxide (DMSO); PPh3 may enhance the diffusion through cell
membranes and DMSO should improve the complex solubility in water.
The changes taking place upon incubation of the complexes in solution are discussed
and related to their DNA-binding features and cytotoxicity. Cytotoxicity assays against
several cancer cell lines and a non-tumorigenic one, reveal the great therapeutic
potential of curcuminate complexes 1 and especially 2, which displayed half-maximal
inhibitory concentrations (IC50) in the range of 6-23 μM. The cytotoxic effect of
[PtCl(curc)(DMSO)] (2) clearly surpassed those of its parent drug cisplatin and free
curcumin. Moreover, the photosensitizing nature of curcumin exerted an even higher
cytoxicity to this complex, reaching IC50 values as low as 2.5 μM upon irradiation with
visible light for 1 h.
2. Experimental
2.1. Materials and instrumentation. cis-PtCl2(PPh3)2 [29] and K[PtCl3(DMSO)] [30]
were synthetized according to literature methods. IR spectra were recorded on a Bruker-
Vector 22 spectrometer. NMR spectra were recorded on a Bruker Advance 400
spectrometer Bruker Advance 400 spectrometer or on a Bruker Avance III 700 MHz
instrument equipped with cryoprobe; chemical shifts are in ppm and coupling constants
13
in Hz; the frequencies are referred to Me4Si for 1H (400 MHz) and for C (100 MHz),
to 85% H3PO4 for 31P (121.5 MHz) and to H2PtCl6 for 195Pt (85.99 MHz) NMR.
Elemental analyses were obtained on a EuroVector CHNS EA3000 elemental analyser
using acetanilide as analytical standard material. Triplicate analysis runs have been
performed to ensure reproducibility. High-resolution mass spectrometry (HR-MS)
analyses were performed using a time-of-flight mass spectrometer equipped with an
electrospray ion source (Bruker micrOTOF). The calculated (exact mass) and the
experimental (accurate) m/z values were compared considering the isotope pattern of the
main ion (that is giving the most intense peak), by using the software Bruker Daltonics
Data Analysis (version 3.3). The melting points were determined with a Büchi Melting
Point B-540 apparatus and are uncorrected. Caffeine was purchased from Alpha Aesar,
potassium tetrachloroplatinate(II) was purchased from ChemPur, curcumin, PPh3 and
the sodium salt of calf thymus DNA (ct-DNA, Type I fibrous) were purchased from
Sigma-Aldrich. Plasmid pBR322 DNA (4361 bp, 0.25 mg mL–1) was obtained from
1
Thermo Scientific. Ethidium bromide (EB) 10 mg mL–
solution and
tris(hydroxymethyl)aminomethane (Tris base) were purchased from Promega. Agarose
(D-1, Low EEO) was purchased from Pronadisa. N-(2-hydroxyethyl)piperazine-N’-(2-
ethanesulfonic acid) (HEPES) and (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-
tetrazolium bromide (MTT) were obtained from Sigma-Aldrich. All reagents were of
molecular biology grade and used without further purification.
2.2. Synthesis of [Pt(curc)(PPh3)2]Cl (1). A solution of KOH (1.46 mmol, 0.0853 g) in
ethanol (4.0 mL) at 40 °C was added dropwise to an orange solution obtained by
dissolving curcumin (1.46 mmol, 0.5408 g) and cis-PtCl2(PPh3)2 (0.725 mmol, 0.5766
g) in 40 mL of CH2Cl2 at room temperature. The resulting dark-red mixture was stirred
at room temperature overnight. Then, the reaction mixture was filtered to remove KCl
and dried under high vacuum. The resulting brown oleum was recrystallized from
CH2Cl2/Et2O and the dark-green solid obtained was filtered off and washed with ethanol
and diethyl ether, and subsequently dried under high vacuum. Yield: 0.695 g, 85%. M.p.

ACCEPTED MANUSCRIPT
= 200.1 °C. Anal. Calcd for C57H49ClO6P2Pt [Mw: 1122.49]: C, 60.99; H, 4.40. Found:
C, 61.17; H, 4.45.
HRMS (ESI, acetonitrile, positive ion mode) m/z: calcd. for C57H49O6P2Pt [M-Cl]+
1086.2643; found. 1086.2651.
1H NMR (400 MHz, DMSO-d6, 298 K) , ppm: 7.53 (m, 6H, Hpara, PPh3), 7.52 (m,
12H, Hmeta, PPh3 ), 7.38 (m, 12H, Hortho, PPh3) , 7.27 (s, 2H, H10), 7.11 (d, 2H, H6, 3JHH
3
3
= 8.0 Hz), 6.79 (d, 2H, H7, JHH = 8.0 Hz ), 6.45 (d, 2H, H3, JHH = 15.0 Hz ), 5.97 (s,
1H, H1), 5.79 (d, 2H, H4, 3JHH = 15.0 Hz ), 3.81 (s, H6, -OCH3). 13C{1H}[31] NMR (100
MHz DMSO-d6, 298 K) , ppm: 175.0 (C2), 148.9 (C8), 148.8 (C9), 141.2 (C4), 134.7
(Cmeta, PPh3), 132.4 (Cpara, PPh3), 129.2 (Cortho, PPh3), 126.8 (Cipso, PPh3), 126.4 (C5),
124.1 (C3), 123.7 (C6), 116.4 (C7), 111.7 (C10), 105.1 (C1), 56.0 (OCH3). 31P{1H} NMR
(121.5 MHz, DMSO-d6, 298 K) , ppm: 8.6 (s, 1JPt,P = 3876 Hz). 195Pt{1H} NMR (85.99
MHz, DMSO-d6, 298 K) , ppm: –3951 (t, 1JPt,P = 3876 Hz).
IR (KBr, cm–1): 3496 (b, m, O-H), 1971-1813 (w, aromatic overtones), 1620 (s,

C=O+C=C), 1601 (s, C=C+C=O), 1511 (s, C=C), 1292 (m, δC-O-C), 1266 (m, C-O), 1158
(m, O-H).
2.3. Synthesis of [PtCl(curc)(DMSO)] (2). 0.0686 g (0.164 mmol) of K[PtCl3(DMSO)]
was dissolved in 5 mL of water. A solution of an equimolar amount of curcumin
(0.0710 g, 0.1604 mmol) and KOH (9.00 mg, 0.1604 mmol) in ethanol (2.5 mL) was
added dropwise to the K[PtCl3(DMSO)] solution, resulting in an immediate darkening
of the resulting mixture. The mixture was left under vigorous stirring at room
temperature for 36 h. Then, it was kept at 277 K overnight, giving rise to the
precipitation of a brown solid, which was filtered off, washed with ethanol and diethyl
ether, and dried under high vacuum. Yield: 0.0786 g, 71%. M.p. = 142.4 °C. Anal. calcd
for C23H25ClO7PtS [Mw: 676.04]: C, 40.88; H, 3.73. Found: C, 40.50; H, 3.47.
HRMS (ESI, acetonitrile, positive ion mode) m/z: calcd. for C23H25ClO7PtSNa [M+Na]+
699.0550; found. 699.0538.
IR (nujol mull, cm–1): 3378 (b, m, O-H), 1623 (s, C=O+C=C), 1590 (s, C=C+C=O), 1511
(s, C=C), 1268 (m, C-O), 1123 (s, S=O), 296 (m, Pt-Cl).
3
1H NMR (700 MHz CDCl3, 298 K) , ppm: 7.70 (d, 1H, H4, JH,H = 15.6 Hz), 7.56 (d,
3
1H, H4’, 3JHH = 15.6 Hz), 7.16 (d, 1H, H6, 3JHH = 8 Hz), 7.14 (d, 1H, H6’, JHH = 8 Hz),
7.04 (broad s, 1H, H10), 7.03 (broad s, 1H, H10’), 6.92 (d, 1H, H7, 3JHH = 8 Hz), 6.91 (d,
3
3
3
1H, H7’, JHH = 8 Hz), 6.51 (d, 1H, H3, JHH = 15.6 Hz), 6.48 (d, 1H, H3’, JHH = 15.6
Hz), 5.83 (s, 1H, H1), 3.94 (s, 3H, -OCH3), 3.93 (s, 3H, -OCH3’), 3.57 (-CH3, 6H,
coordinated DMSO). 13C{1H} NMR (176 MHz CDCl3, 298 K) d, ppm: 176.2 (C2),
176.0 (C2'), 147.7 (C8,8’), 146.6 (C9,9’), 142.0 (C4), 141.1 (C4’), 127.8 (C5), 127.6 (C5’),
122.4 (C6), 122.8 (C6’), 122.1 (C3), 121.8 (C3’), 114.8 (C7), 115.0 (C7’), 110.0 (C10),
109.5 (C10’), 103.9 (C1), 55.9(-OCH3), 44.8 (-CH3, coordinated DMSO). 195Pt{1H}
NMR (86 MHz CDCl3, 298 K) , ppm: –2381 (s).
2.4. Synthesis of trans-[Pt(caffeine)Cl2(DMSO)] (3). A solution containing 0.1276 g
(0.657 mmol) of caffeine in 9 mL of water was added dropwise to a solution obtained
by dissolving 0.2503 g (0.597 mmol) of K[PtCl3(DMSO)] in 5 mL of water. The

ACCEPTED MANUSCRIPT
addition of the first drops of caffeine solution caused a colour change of the platinum-
containing solution, from orange to peach pink, with concomitant precipitation of a
white solid. The mixture was left under vigorous stirring at room temperature for 12 h.
Then, the white solid was removed by filtration from the reaction mixture, washed with
ethyl alcohol and diethyl ether, and dried under high vacuum. Yield: 0.1786 g, 58%.
M.p. = 232.4 °C. The complex is soluble in DMSO and slightly soluble in CHCl3. Anal.
Calcd for C10H16Cl2N4O3PtS [Mw: 538.31 Da]: N, 10.41; C, 22.31; H, 3.00. Found: N,
10.34; C, 22.36; H, 2.79. 1H NMR (400 MHz CDCl3, 298 K) , ppm: 7.94 (q, H1, 8-H,
4
4JHH = 0.7 Hz), 4.40 (s, 3H, H3), 4.12 (d, 3H, H7, JHH = 0.7 Hz), 3.52 (s, 6H, -CH3 of
DMSO group), 3.44 (s, 3H, H1). 13C{1H} NMR (101 MHz CDCl3, 298 K) , ppm: 154.3
(C6), 151.0 (C2), 145.1 (C4), 140.7 (C8), 108.4 (C5), 44.2 (-CH3 of DMSO), 35.1 (C7),
33.4 (C3), 28.5 (C1). 195Pt{1H} NMR (85.99 MHz, CDCl3, 298 K) , ppm: –3009 (s).
IR (nujol mull, cm–1): 1716 (s, C=O), 1668 (s, C=O), 1612 (w, C=N), 1240 (m, C-N),
1145 (s, S=O), 341 (m, Pt-Cl).
HRMS (ESI, methanol, positive ion mode) m/z: calcd. for C10H16Cl2N4O3PtSNa
[M+Na]+ 560.9857, found. 560.9834; calcd. for C10H16Cl2N4O3PtSK [M+K]+ 576.9596,
found 576.9577; calcd. for C20H32Cl4N8O6Pt2S2Na [2M+Na]+ 1098.9790, found
1098.9798.
2.5. UV-Vis measurements. Electronic spectra were recorded on a Varian Cary 100
Scan spectrophotometer at room temperature employing a 1.0 cm path length cuvette.
Stock solutions of complexes 1−3 were prepared in DMSO. For the stability studies,
UV-Vis spectra of 25 µM solutions of 1−3 in PBS (containing 1% DMSO) were
recorded over a period of 24 h. For the DNA-binding assays, the concentration of ct-
DNA was determined from the absorption intensity at 260ꢀnm and the corresponding
molar absorptivity of 6600ꢀM−1ꢀcm−1 (nucleobase concentration). Absorption titration
experiments were carried out by adding increasing amounts of ct-DNA (0−50 µM) to a
25 µM solution of the compound in cacodylate-NaCl buffer (1.0 mM sodium
cacodylate, 20 mM NaCl, pH 7.2). The final concentration of DMSO was 0.25 %. The
spectra were recorded after equilibration for 5 min at 20 ºC. Corrections were made for
the absorbance of ct-DNA itself.
The intrinsic binding constant (Kb) was determined from the titration data using
equation (1):
[퐷푁퐴]
[퐷푁퐴]
1
=
+
퐾 (휀 −휀 )
푏 푓
0
(Eq. 1)
휀 −휀
휀 −휀
푎
푓
0
푓
where εa, ε0, and εf are respectively, the molar extinction coefficients of the free
complexes in solution, fully bound complex with DNA, and complex bound to DNA at
a definite concentration.
2.6. Fluorescence measurements. Emission intensity of the DNA-intercalator ethidium
bromide (EB) was measured using an iHR320 HORIBA JOBIN YVON
spectrofluorometer. The samples contained 15 μM (base pairs) of ct-DNA and 75 μM
EB in 1.0 mM cacodylate – 20 mM NaCl buffer (pH 7.2). In a first instance, EB was
incubated with ct-DNA for 30 minutes, to allow the intercalation of EB between DNA

ACCEPTED MANUSCRIPT
base pairs. Afterwards, increasing amounts of compounds 1−3 (0 - 25 μM) were added
and the final samples were incubated for 24 h. A maximum of 5% DMSO was used to
solubilise the complexes. The fluorescence spectra of all complexes were recorded at
room temperature, at λexc= 514 nm.
For the displacement assays of the minor-groove binder Hoechst 33258, 0.10 mM of the
dye was used and the same conditions as those used for the EB-displacement assays
were applied. The fluorescence spectra were recorded at λexc = 350 nm.
2.7. Agarose gel electrophoresis. Stock solutions of the platinum(II) compounds were
prepared in 1.0 mM cacodylate – 20 mM NaCl buffer (pH = 7.2) containing 2% DMSO.
pBR322 plasmid DNA aliquots (50 µM base pairs) in 1 mM cacodylate – 20 mM NaCl
buffer were incubated with the complexes for 24 h at 37 °C. Next, the reaction samples
were quenched with 4 μL of loading buffer (5 mM xylene cyanol and 30% glycerol) and
electrophoresed on 1% agarose gel in 1x Tris-Borate-EDTA buffer (89 mM Tris pH 7.6,
89 mM boric acid and 2 mM EDTA) for 90 min at 6.5 V cm–1. The electrophoresis was
run in a BioRad horizontal tank connected to a CONSORT EV231 electrophoresis
power supply. Free DNA was used as control. Afterwards, the DNA was stained with
SYBR® Safe overnight and the gel was photographed with a BioRad Gel Doc EZ
Imager.
2.8. Circular dichroism (CD) spectroscopy. Circular dichroism (CD) spectra were
recorded using a JASCO-815 spectropolarimeter, equipped with a xenon-arc lamp of
450 W, and connected to a computer to carry out the data processing. The sample
compartment was air-purged with N2 before use. For the measurements, quartz cuvettes
with an optical path of 5 mm were used. The concentration of ct-DNA was determined
from its absorption intensity at 260 nm, with a molar extinction coefficient of 6600 M–1
cm–1. A solution of 200 µM ct-DNA in 1.0 mM cacodylate – 20 mM NaCl buffer (pH
7.2) was incubated for 24 h with metal complex / DNA ratios of 0 – 1 (except for
compound 1, which precipitated at ratios over 0.2 at 37 °C) before recording the CD
spectra at room temperature.
2.9. Atomic force microscopy (AFM). pBR322 plasmid DNA aliquots (0.2 mg mL–1) in
40 mM HEPES – 10 mM MgCl2 buffer were incubated with the complexes for 24 h at
37 °C. The AFM samples were prepared by casting a 5 μL drop of test solution onto
freshly cleaved muscovite green mica disks as the support. The drop was allowed to
stand for 3 min to favour the adsorbate–substrate interaction. Each DNA laden disk was
rinsed with Milli-Q water and was blown dry with clean compressed argon gas directed
normal to the disk surface. The samples were stored over silica prior to AFM imaging.
Images were recorded with a Bruker AFM Multimode8 with nanoscope V electronics
using a SNL tip and Scan Asyst mode (1hz).
2.10. Cell culture and cell-viability assays. Human melanoma A375, lung A549, breast
MCF-7, ovary SKOV3 colorectal SW620 adenocarcinoma and non-tumorigenic
epithelial breast MCF10A cell lines were obtained from the American Type Culture
Collection (ATCC, Manassas, VA, USA). The cell lines A549, A375, SKOV3 and
SW620 were cultured in DMEM medium with 10% heat-inactivated fetal bovine serum

ACCEPTED MANUSCRIPT
(FBS; Life Technologies, Carlsbad, CA, USA), 100 U mL−1 penicillin, 100 μg mL−1
streptomycin, and 2.0 mM glutamine. The MCF-7 cell line was cultured in DMEM–F12
(HAM) media (1:1) with 10% FBS, 50 μM sodium pyruvate (Sigma-Aldrich Chemical
Co., St. Louis, MO, USA), 10 μg mL−1insulin (Sigma-Aldrich), 100 U mL−1 penicillin,
100 μg mL−1 streptomycin, and 2 mM glutamine. MCF10A cells were cultured in
DMEM–F12 (HAM) media (1:1) with 5% horse serum (Life Technologies), 100 U
mL−1 penicillin, 100 μg mL−1 streptomycin, 2.0 mM glutamine, 10 μg mL−1 insulin 0.5
µg mL-1 hydrocortisone (Sigma-Aldrich), 20 ng mL−1 epidermal growth factor (EGF,
Sigma-Aldrich), 100 ng mL−1 choleric toxin (Calbiochem,San Diego, CA, USA). Both
the media and the added chemicals whenever not stated were obtained from the
Biological Industries, Beit Haemek, Israel. Cells were grown at 37 °C in a 5% CO2
atmosphere.
104 cells were seeded in 96-well plates (105 cells mL−1) and allowed to grow for 24 h.
For single-point experiments, the cells reacted with the investigated compounds at two
different concentrations (10 μM and 50 μM) and incubated for 24 h. For the dose–
response curves, the cells were treated with different concentrations of the compounds
(from stock solutions in DMSO), in the range of 0.16–20 µM for compound 1 and
cisplatin dissolved in PBS or in the range of 0.78-100 µM for compound 2, curcumin
and cisplatin dissolved in DMSO, and incubated for 24 h. Prior to cell treatment, the
compounds were freshly dissolved (referred as Fresh in Table 1) or incubated in the
corresponding culture medium for 24 h at 37 ºC (referred as Pre-incubated). After
dosing the compounds, the cells were incubated 1 h at room temperature in the dark
(Non-irradiated) or under visible light (Irradiated). These steps were followed by 24 h
incubation at 37 ºC. Next, a 10 μM solution of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyl-tetrazolium bromide) was added to each well and the plates were incubated at
37 °C for two additional hours. The medium was removed, and the purple formazan
crystals were dissolved in 100 µL of DMSO. The absorbance was measured at 570 nm
using a multi-well plate reader (Multiskan FC, Thermo Scientific). The cell viability
was calculated according to the relation: viability (%) = [(absorbance of the treated
wells)/(absorbance of the control wells)] × 100.
The IC50 values (corresponding to the compound concentrations that produce 50%
reduction in cell viability) were obtained from the dose–response curves using
GraphPad Prism V5.0 for Windows™ (GraphPad Software, San Diego, CA, USA). All
data are shown as the mean value ± S.D. of three independent experiments for single-
point assays and for the dose–response curves.
3. Results and Discussion
3.1. Synthesis and characterization of compounds 1−3.
Deprotonated curcumin binds almost all metals due to its β-diketonate moiety, which is
known to be an excellent chelating group, owing to the delocalization of the  electrons
in the resulting ring and the consequent stabilization of the overall system.[32] In
addition, the coordination of curcumin to a metal centre protects curcumin itself against
homo-polymerization and photo-degradation, and modifies its solubility in some
solvents, depending on the nature of the other ligands bound to the metal centre. In fact,
the major problems associated to the use of the sole curcumin as a drug arise from its
poor solubility in water under physiological conditions, enhanced by its tendency to
polymerize, and its low stability to sunlight. Thus, the binding of curcumin to a metal
could increase its biological features, as it has been observed for [Pt(curc)(NH3)2]NO3

ACCEPTED MANUSCRIPT
(curc = deprotonated curcumin),[28] which is a cisplatin analogue, showing synergic
anticancer effects between a platinum-based drug and curcumin. Recently,
triarylphosphane-bearing platinum complexes have been revaluated from a cytotoxic
point of view,[33] due to their ability to across the cell membrane, thanks to lipophilic
phosphane groups.[34] Triphenylphosphane platinum complexes carrying the
deprotonated form of biologically active flavonoids showed interesting cytotoxic
activity against selected cell lines.[35] Following our studies on platinum complexes
with naturally occurring biologically active ligands, we decided to synthetize curcumin-
containing Pt complexes bearing triphenylphosphane or dimethyl sulfoxide (DMSO) as
carrier ligands.
Curcumin exhibits a tautomeric equilibrium with two possible structures through
intramolecular hydrogen transfer. The typical predominance of the enol form in solution
facilitates the α-deprotonation under alkaline conditions and promotes the O,O’-
bidentate coordination to a metal centre. In fact, we easily obtained [Pt(curc)(PPh3)2]Cl
(1) by reaction of curcumin with cis-[PtCl2(PPh3)2] in the presence of an equimolar
amount of KOH with respect to curcumin (Scheme 1).
The HR ESI-MS(+) spectrogram of a diluted acetonitrile solution of 1 shows an intense
peak at m/z 1086.2643, whose isotope pattern is perfectly superimposable to that
calculated for the cation [Pt(curc)(PPh3)2]+.
The 31P{1H} NMR spectrum of 1 in DMSO-d6 exhibits a singlet at δ 8.6 ppm flanked by
1
satellites from which a JPtP value of 3876 Hz could be extracted. The 195Pt{1H} NMR
spectrum shows a triplet at δ 4400 ppm. These data are consistent with a Pt(II)
phosphane complex,[36] bearing two magnetically equivalent 31P nuclei mutually in cis
positions,[37] and in trans position with respect to an anionic dioxygen ligand.[33, 35]
1
Accordingly, in the H NMR spectrum, the signal of the intramolecular hydrogen
bonded enolic proton, which falls at 10.05 ppm for the free ligand,[38] is not
observed and the -proton signal of the coordinated diketone moiety (5.97 ppm) is
shifted 0.08 ppm up-field with respect to that of free curcumin ( 6.15 ppm), as it occurs
for similar curcumin-chelated metal complexes.[39, 40] Analogously, the 13C{1H}
NMR spectrum of 1 in DMSO-d6 reveals a 7 ppm up-field shift of the magnetically
13
equivalent -diketonate C carbonyl signal ( 176.0 ppm) with respect to the carbonyl
peak of free curcumin ( 183.6 ppm), as it has been found in the 13C{1H} NMR
spectrum of [Pt(curc)(NH3)2]NO3,[28] whose structure is very similar to that proposed
13
1
for complex 1. In addition, the C and H NMR data evidence only one set of signals
for 1, indicating a symmetric molecule, thus corroborating a chelate curcumin
coordination. The other NMR spectroscopic features were unequivocally assigned by
2D NMR experiments and are reported in the experimental section.
The IR spectrum of 1 shows the two characteristic bands at 1620 cm–1 and 1601 cm–1,
[39, 41] assigned to C=O coupled with C=C and to C=C coupled withC=O, respectively,
which are typical for metal chelate -diketonate systems.[39, 40],[42] No Pt-Cl
stretching peak was detected in the far IR region for complex 1, suggesting that metal
coordination by chloride[43] did not occur, as already observed in similar platinum(II)
complexes.[35]

ACCEPTED MANUSCRIPT
Scheme 1. Synthetic pathways for the preparation of compounds 1−3.
Some Pt(II) and Pd(II) complexes containing-diketonate ligands and DMSO have
been recently synthetized, the platinum ones displaying higher cytotoxicity against
selected cancer cells than the free ligands.[44] Inspired by these findings, we decided to
synthesize a platinum compound bearing anionic curcumin, as leaving group, and
DMSO, as carrier ligand, exploiting the procedure already successfully used for the
synthesis of some ß-diketonate Pt complexes containing DMSO as ligand.[45] Thus, the
reaction of K[PtCl3(DMSO)] with deprotonated curcumin (Scheme 1) gave in high yield
complex [PtCl(curc)(DMSO)] (2). To prevent metal reduction, a slight excess of
curcumin was used with respect to the 1:1 ratio of KOH/ platinum precursor.[45]
The results of the elemental analyses (see Experimental Section) and of the HRMS
measurements (Figure S1) are in accordance with the proposed structure.
The IR spectrum of 2 shows typical bands for chelate curcumin complexes, namely at
1623 cm–1 (C=O + C=C) and 1590 cm–1 (C=C + C=O), along with a strong absorption at
1511 cm– ascribed to C=C. The presence of an intense S=O stretching band at 1123
cm–1 indicates that DMSO is coordinated to platinum through its sulfur atom.[42] The
1
Pt-Cl stretching band appears at 296 cm–1. In the H NMR spectrum, the expected
differentiation between the two halves of the coordinated curcumin was observed. For
instance, the signals of H3 and H3’ were found at  6.51 and  6.48, respectively, while
the signals of H4 and H4’ were found at  7.70 and  7.56, respectively. The 195Pt{1H}
signal was found at  –2381, a value comparable to that observed for the related
complex [PtCl(acac)(DMSO)] (Pt –2399). [32]
As mentioned above, 1 and 2 contain the biologically active ligand (i.e. the
curcuminate) as leaving group. Next, a Pt complex bearing a biologically active
molecule as carrier ligand was designed and prepared, namely trans-
[Pt(caffeine)Cl2(DMSO)] (3). Caffeine is a purine-type alkaloid that is well-known for
its anticancer effects.[46] The caffeine complex 3 was synthesized modifying a reported
procedure used for the synthesis of trans-[PtCl2(DMSO)(py)] (py = pyridine),[30] i.e.
by reaction between K[PtCl3(DMSO)] and a slight excess of caffeine.

ACCEPTED MANUSCRIPT
The HR ESI-MS data achieved for 3 in methanol showed an intense peak at m/z
560.9834, ascribed to the sodium adduct [3+Na]+ along with peaks at m/z 576.9577
([3+K]+) and at m/z 1098.9798 (attributed to the sodium adduct of a dimer of 3).
The IR spectrum showed, beside the Pt-Cl stretching band at 341 cm–1, two bands at
1716 cm–1 and 1668 cm–1, ascribed to C=O stretching vibrations. The C=O values for
free caffeine are 1702 cm–1 and 1662 cm–1, hence indicating that the C=O groups are
not coordinated to the metal.[47] The presence of a strong band at 1145 cm–1 assigned
to the S=O stretching indicates that DMSO is S-coordinated to platinum.[39]
1
The H NMR spectrum of complex 3 in CDCl3 showed the signal attributable to H8 at
7.94 ppm, typical for Pt coordination of caffeine through the nitrogen atom N9.[48] In
1
addition, the H-195Pt HMQC spectrum of 3 in CDCl3 (Figure 1) clearly revealed
couplings between the 195Pt nucleus and: H8 (7.94), N3-CH3 ( 4.40), and the
coordinated DMSO (3.52 ppm), whose CH3 protons at  3.52 ppm are typical for an
S-bonded ligand.[49]
The 195Pt NMR signal in CDCl3 was found at  –3009, which is consistent with trans,
square-planar Pt(II) dichloride complexes containing DMSO and N-bonded
13
heterocycles.[50] The C{1H} NMR spectrum of 3 showed signals for the coordinated
caffeine which were slightly de-shielded with respect to those of free caffeine.[51]
Figure 1. 1H-195Pt HMQC spectrum of 3 (CDCl3, 298 K).
3.2. Stability of complexes 1−3
The behaviour/stability of compounds 1−3 in solution was first evaluated by UV-Vis
spectroscopy, in PBS containing 1% DMSO at 37 ºC. UV-Vis spectra solutions of the
different compounds were thus recorded during 24 h. As observed in Figure S2, the
curcumin-based complexes 1 and 2 display a strong decrease of the absorbance at 440

ACCEPTED MANUSCRIPT
nm, ascribed to the alteration/modification of the original compounds. The caffeine-
containing complex 3 is characterised by an initial, slight decrease of its main
absorption followed by an increase during the first hour; subsequently, the data obtained
up to 24 hours do not show any further changes.
Next, the stability of the new Pt complexes was investigated by NMR, which
corroborated the results obtained by UV-Vis spectroscopy. The modifications
experienced by the complexes in solution after 24 h of incubation are illustrated in
Scheme 2. NMR studies indeed confirmed that complexes 1 and 2 were progressively
1
13
hydrolyzed. In fact, the H and C{1H} spectra of a DMSO-d6 solution of complex 1
left overnight in the NMR tube revealed the presence of free curcumin (half amount
with respect to the coordinated ligand). The hydrolysis is significantly faster in the case
of complex 2, which easily loses curcumin during the first hour in solution. In addition,
1H NOESY-EXSY experiments revealed the occurrence of an intense exchange cross-
peak between the -proton of free curcumin and water (Figure S3). Furthermore, the 1H-
195Pt HMQC spectrum of a sample left 24 h in the NMR tube (containing non-
anhydrous solvent) showed the presence of an extra platinum signal at –3446 ppm,
correlating to coordinated DMSO methyl protons (Figure S4). This signal is tentatively
assigned to a platinum hydroxy species.
Complex 3 did not undergo hydrolysis. Even after a few days in (non-anhydrous)
CDCl3, its NMR spectrum did not show any signals belonging to free caffeine.
However, it slowly isomerizes from the trans to the cis form, leading to an equilibrium
mixture. In fact, the 1H-195Pt HMQC spectrum registered after 24 h (Figure S5) showed
the presence of an extra cross-peak at –3457 ppm for 195Pt (correlating with DMSO
1
methyl protons), consistent with a cis Pt(II) complex.[50] The H and 13C signals of
trans and cis isomers were isochronous.[52]
Scheme 2. Hydrolysis of 1 and 2 and isomerization of 3 during 24 h incubation.
3.3. Photoluminescence properties of curcuminate complexes 1 and 2
Given the emissive nature of curcumin, the photoluminescence spectra of complexes 1
and 2 have been recorded. Figure 2 displays both the absorbance and the emission
spectra of curcumin and complexes 1 and 2 at the same concentration, for comparison.

ACCEPTED MANUSCRIPT
The emission of free curcumin results from its delocalized π-conjugated electronic
system, which is strongly influenced by solvents, tautomerism, and structural
modifications.[53] Coordination to metal ions may quench the fluorescence through
either energy or electron-transfer processes. Namely, square-planar platinum(II)
complexes have been shown to exhibit interesting and intriguing photophysical
behaviours, associated with the strong tendency to form π-π stacking and Pt···Pt
interactions.[54]
Curcumin emission (560 nm) is quenched when forming the complexes with
platinum(II). Such phenomenon has been observed with other platinum(II) complexes,
and is most likely due to the presence of low-lying thermally accessible d–d LF state,
which would result in the quenching of the excited state via non-radiative decay, totally
quenching
or
diminishing
the
luminescence.
[54]
Interestingly,
the
bis(triphenylphosphane) complex 1 does not only show a stronger quenching of the
emission but also a remarkable red shift of the absorbance and emission maxima, with
respect to the free curcumin; i.e. its absorbance maximum changes from 430 to 470 nm,
and its emission shifts from 560 to 660 nm. On the contrary, no shift is observed in the
emission maxima of complex 2.
Figure 2. Absorbance (a) and emission spectra (b) of freshly prepared 30 µM solutions
of curcumin and its platinum(II) complexes 1 and 2 in cacodylate-NaCl buffer pH 7.2
(0.3% DMSO).
3.4. DNA-binding studies
DNA is the classical pharmacological target of metal-based anticancer agents; therefore,
the in vitro study of the interaction of coordination compounds with DNA is of
paramount importance to assess the therapeutic potential of the compounds investigated.
The four main ways (but not the only ones) in which small molecules can bind to
double-stranded DNA are: (i) intercalation between two adjacent base pairs and
perpendicular to the helical axis; (ii) outside-edge binding to the sugar-phosphate
backbone of the helix through electrostatic interactions; (iii) groove binding with
functional groups into either the major or minor groove and (iv) covalent interaction
between DNA and the metal complexes at the nitrogen atoms of nucleobases.[55]
Several characterization techniques have been employed to study the interactions of
platinum(II) complexes 1−3 with DNA, namely UV-Vis spectroscopy, displacement

ACCEPTED MANUSCRIPT
assays of DNA binders, circular dichroism, agarose gel electrophoresis and atomic-force
microscopy. It should be stressed that the studies with DNA were carried out using
incubation periods of 24 h, during which the complexes hydrolysed (compounds 1 and
2) and isomerized (compound 3).
3.4.1. UV-Vis spectroscopy
Using increasing amounts of ct-DNA (applying a short incubation time, viz. 5min), the
absorption maximum of complex [Pt(curc)(PPh3)2]Cl (1) at 440 nm is shifted towards
longer wavelengths, with concomitant decrease of its intensity (Figure 3c). Such
bathochromic and hypochromic effects are generally associated to an intercalative
process into the double-stranded DNA. This reduction of the absorption intensity
combined with a red shift may arise from strong π - π* stacking interactions between the
planar aromatic curcumin ligand and DNA base pairs.[56, 57] To take into
consideration the hydrolysis of compound 1, it was also incubated for 24 h (instead of 5
min) with increasing amounts of ct-DNA; however, the precipitation of the modified
complex did not allow to carry out this study.
In the case of [PtCl(curc)(DMSO)] (2), 5 min-incubation with increasing amounts of
ct-DNA, the absorption maximum at 444 nm initially exhibits a hyperchromic effect
upon the first addition of DNA, followed by a hypochromic effect in the presence of
increasing amounts of the biomolecule, with no wavelength shift (Figure 3a). These
features are consistent with changes of the DNA conformation upon complex binding,
either by electrostatic interactions or groove binding.[58] The binding constant (Kb) was
calculated applying Equation 1, giving a value of 8.8 × 105 M–1. Hydrolysed 2 (after 24
h incubation with ct-DNA) produced a distinct hypochromic effect (Figure 3b), which
may arise from either electrostatic interactions or groove binding. The [DNA]/(εa−εf) vs.
[DNA] plot reveals the occurrence of two different Kb values, which can be due to the
interaction of hydrolysed 2 and free (released) curcumin with the biomolecule.[20]
For [Pt(caffeine)Cl2(DMSO)] (3), no changes of the absorption (at λmax = 272 nm) were
observed upon incubation with ct-DNA for 5 min (data not shown). Upon incubation for
24 h, the main absorbance band increased strongly with the first addition of DNA, and
only experienced slight changes with more DNA added (Figure 3d). The observed
hyperchromicity indicates a decrease in the degree of parallel stacking in the DNA
double helix, which may result from covalent bonding of Pt(II) to purine bases. Similar
hyperchromic effects have been reported for cisplatin interacting with DNA, although
no batochromism is detected in this case.[59] A high Kb value of 3 × 107 M–1 was
obtained for 3.

ACCEPTED MANUSCRIPT
Figure 3. Absorption spectra of 25 µM solutions of compounds 1−3 without DNA and
with increasing amounts of ct-DNA (0–25 µM), in cacodylate buffer (pH = 7.2), at the
indicated incubation times (5 min or 24 h). The arrows show the change upon addition
of DNA. Insets: Linear fitting of the plots of [DNA]/(εa−εf) vs. [DNA] for the titration
of ct-DNA with complexes 1−3 at the specified wavelengths.
3.4.2. Displacement assays of fluorescent DNA-binders
To further investigate the nature of the interactions between the platinum(II) complexes
1−3 and ct-DNA, competitive binding studies using ethidium bromide (EB) bound to ct-
DNA were carried out. EB is a DNA-intercalating agent whose fluorescence emission is
nearly 20-fold stronger when bound to the polynucleotide molecule.[60, 61]
Hence, if a complex is able to displace DNA-bound EB upon interaction with DNA,
then a decrease of fluorescence (due to EB release) will be detected. The extent of this
displacement can provide information regarding the DNA-binding affinity of the
compound considered.[62] It should be stated here that EB displacement does not imply
that the interacting compound as an intercalator like EB; electrostatic interactions or
groove binding may be sufficient to alter significantly the conformation of the DNA
double helix, inducing the release of EB.
Fluorescence spectra of solutions containing constant concentrations (i.e. 30 and 75 µM,
respectively) of ct-DNA and EB (previously incubated) and increasing amounts of
complexes 1−3, viz. in the range 0–25 µM, were recorded after 24 h. In all cases, a clear
decrease in emission intensity is noticed (Figure 4). To assess the respective affinity of
the different complexes for ct-DNA (compared to EB), their quenching efficiency was

ACCEPTED MANUSCRIPT
evaluated determining the Stern–Volmer quenching constant KSV, applying equation
(2):
퐼
0
= 1 + 퐾푆푉[푐표푚푝푙푒푥]
(Eq. 2)
퐼
where I0 and I are the emission intensities in the absence and the presence of the
complex, respectively. Figure 4 shows the I0/I values in function of the concentration,
for 1−3. The slopes of the corresponding lines gave KSV values of 5.0 × 104 M–1, 1.2 ×
104 M–1, and 7.7 × 103 M–1 for 1, 2 and 3, respectively. These values reveal moderate
EB-displacing abilities, complex [Pt(curc)(PPh3)2]Cl (1) inducing a significantly
stronger displacement than complexes 2 and 3. The higher propensity of 1 to act as a
possible intercalator (see Section 3.4.1) corroborates the UV-Vis data, which showed a
shift of the absorbance maximum for this compound.
Figure 4. a) Emission spectra of the DNA–EB complex, upon addition of increasing
amounts of compound 1. The arrow shows the diminution of the emission intensity with
the addition of 1; b) Plots of I0/I vs. complex concentration for the titration of DNA–EB
with complexes 1−3: experimental data points and linear fitting of the data. I0
corresponds to the maximum emission of the DNA–EB complex in sodium cacodylate
buffer (pH = 7.2) and I stands for the maximum emission intensity of the DNA–EB
complex after 24 h-incubation with increasing concentrations of complexes 1, 2 and 3.
[EB] = 75 µM, [DNA] = 30 µM, [complex] = 0–25 µM, λex = 514 nm and λem = 610
nm. Measurements were carried out in duplicate; error bars represent standard deviation
values.
Competitive binding studies were next conducted with the minor-groove binder Hoechst
33258. When this dye is bound to ct-DNA, it fluoresces at 458 nm when excited at 350
nm, while its free form emits at 508 nm when excited at 337 nm. As for EB, its
fluorescence is dramatically increased upon binding to the biomolecule, over 28-fold as
reflected by the corresponding quantum yields.[63] Therefore, displacement of DNA-
bound Hoechst 33258 by a metal complex will result in fluorescence quenching if it
interacts in the minor groove of the double helix.
Increasing amounts of complexes 1−3 were added to ct-DNA (30 µM, in terms of
nucleobases) and Hoechst 33258 (100 µM). In all cases, the emission of the Hoechst-
DNA complex gradually decreased upon complex addition (see Figure 5a for compound
2). The fluorescence decrease was not linear for the bulkier curcumin complexes 1 and
2 at concentrations above 3 µM, as observed in Figure 5b. Such a non-linear quenching

ACCEPTED MANUSCRIPT
phenomenon may be attributed to an energy-transfer process between the dye and the
released curcumin from complexes 1 and 2; it can indeed be noted that the emission of
the dye (λem = 458 nm) and the absorbance of free curcumin (λmax = 430 nm; Figure 2a)
overlap. Hence, the extent of dye displacement and the energy-transfer contribution
could not be discerned for any of these complexes. A weak groove binding of free
curcumin should also not be ignored in this displacement assay. [20]
Therefore, the Stern-Volmer equation (2) could only be applied to 3, which gave a KSV
value of 3.6 × 104 M–1. This KSV value, obtained with the groove-binder Hoechst 33258,
is five times higher than achieved by competitive binding with the intercalator EB. Such
an effect may be caused by the kinking of the DNA double helix through cis-bidentate,
covalent binding, which is suggested as well by the UV-Vis data.
From the relatively low KSV value obtained for 2 in the EB displacement assays, it can
be considered that it is not acting as intercalator either; most likely, it interacts with ct-
DNA electrostatically or via groove binding.
Figure 5. a) Emission spectra of the DNA–Hoechst33258 complex upon addition of
increasing amounts of 2. The arrow shows the diminution of the emission intensity with
the addition of the metal complex; b) Plots of I0/I vs. complex concentration for the
titration of Hoechst-DNA with complexes 1−3: experimental data points and linear
fitting of the data. I0 corresponds to the maximum emission of the Hoechst–DNA
complex in sodium cacodylate buffer (pH = 7.2) and I stands for the maximum emission
intensity of the Hoechst-DNA after 24 h-incubation with increasing concentrations of
complexes 1, 2 and 3. [Hoechst33258] = 100 µM, [DNA] = 30 µM, [Complex] = 0–25
µM, λex = 350 and λem = 458 nm. Measurements were carried out in duplicate; error bars
represent standard deviation values.
3.4.3. Circular dichroism (CD).
The CD spectrum of natural calf thymus DNA consists of a positive band at 275 nm due
to base stacking and a negative band at 245 nm due to the characteristic helicity of the
right-handed B form. The effect of 1−3 on the secondary structure of ct-DNA was
studied with a constant concentration of DNA by increasing the complex concentration
up to one equivalent with respect to the concentration of nucleobases, except for
complex 1, which precipitated at complex:DNA ratios over 0.2 (Figure 6).

ACCEPTED MANUSCRIPT
In the presence of 1, the intensity of the positive peak increased, and the intensity of the
negative peak decreased. This phenomenon was accompanied by a hypsochromic shift
of both peaks. These observations are consistent with an intercalative process.[64]
Figure 6. Circular dichroism spectra of ct-DNA (200 µM) in the presence of increasing
amounts of complexes 1 (a), 2 (b) and 3 (c), in 1.0 mM cacodylate – 20 mM NaCl
buffer (pH = 7.2). The spectra were recorded after 24 h incubation at 37 ºC.
Precipitation was observed for 1 at complex:DNA ratios above 0.2.

ACCEPTED MANUSCRIPT
The incubation of ct-DNA with complexes 2 and 3 gave rise to a decrease of the
intensity of both the positive and negative signals, accompanied by a bathochromic
shift. These changes are indicative of alterations of the secondary structure of B-DNA,
probably as a result of a non-intercalative interaction of the complexes with the double
helix. The effect was remarkably stronger for 3 and is comparable to that observed for
cisplatin, i.e. cis-[PtCl2(NH3)2]; these features suggest a cis-bidentate covalent binding
of 3 to DNA, which is consistent with the UV-Vis measurements and allowed by a
trans-to-cis isomerization upon incubation of the compounds.[65]
Finally, it can be pointed out that, at the complex concentrations used, induced CD
signals in the metal-to-ligand charge transfer region were not detected.
3.4.4. Agarose gel electrophoresis.
The interaction of the Pt(II) complexes with DNA can be visualized by agarose gel
electrophoresis through the changes they produce to the DNA conformation. These
structural changes are reflected by distinct electrophoretic mobilities. pBR322 DNA
usually presents about 90% of supercoiled form (Form I) and 10% of the nicked form
(Form II). The most intense band of the non-treated DNA sample in Figure 7, lane 1
corresponds to form I (supercoiled DNA), which migrates faster than form II (nicked or
relaxed circular DNA); being more compact, form I is less retained in the gel. Figure 7
also shows the effects of increasing concentrations of complexes 1−3 (from 5 to 50 µM)
on the DNA structure, after incubation of 24 h.
A slight decrease of the electrophoretic mobility of DNA is observed with increasing
amounts of 1 (Figure 7, lanes 3−6), which results from the unwinding of the double
helix due to the intercalation of the complex. The bands gradually lose their intensity,
suggesting that the DNA is cleaved into undetectable single-stranded pieces; the
intermediate linear form III is not detected. Since the samples were exposed to daylight
during their preparation (except during the incubation step), the observed degradation
may result from photocleavage, a feature that has been reported for other platinum(II)
complexes containing curcumin, which has a photosensitizing character.[28] The
oxidative photo-induced damage of 1 may be favoured by its intercalation between
nucleobase pairs; this would also explain why such a damage is not observed for the
other curcumin-containing compound 2.
The behaviour of 2 displays is similar to that of the caffeine-containing complex 3. At
low concentrations of these two complexes, the respective bands observed (Figure 7,
lanes 7−8 and 11−12, respectively) suggest an unwinding of the supercoiled form,
leading to the coalescence of form I and II. At higher concentrations, namely 25 and 50
µM, complexes 2 and 3 produce a gradual shrinking of the relaxed circular form and,
more especially, of the supercoiled form (Figure 7, lanes 9−10 and 13−14, respectively);
this effect results in a higher electrophoretic mobility. These data are consistent with a
non-intercalative interaction of the compounds with the biomolecule. It can be pointed
out that the effect is more pronounced for 2.

ACCEPTED MANUSCRIPT
Figure 7. Agarose gel electrophoresis images of pBR322 plasmid DNA (50 µM base
pairs) incubated for 24 h at 37 °C with increasing concentrations of the different
complexes in cacodylate-NaCl buffer pH 7.2. Lanes 1 and 2 contain non-treated DNA
and DNA treated with 10 µM cisplatin. The other lanes correspond to increasing
concentrations, namely 5, 10, 25 and 50 µM, of compounds 1 (lanes 3−6), 2 (lanes
7−10) and 3 (lanes 11−14).
3.4.5. AFM analysis
Atomic force microscopy (AFM) investigations were carried out with pBR322 plasmid
DNA incubated at 37 °C in the dark for 24 h with complexes 2 and 3. The effect of 1
could not be examined with this technique because it co-precipitated with DNA at the
concentrations used for these studies.
The AFM images presented in Figure 8 show the open circular form of plasmid DNA
with some degree of supercoiling (as observed by electrophoresis), and their subsequent
morphological alteration upon incubation with different concentrations of compounds 2
and 3. Figures 8b and 8c clearly illustrate that the incubation of DNA with 2 caused a
marked shrinking (supercoiling) of the biomolecule, generating small aggregates with a
nearly globular shape, especially at higher complex concentrations. A similar but less
pronounced effect could be observed for 3 (Figs. 8d and 8e). The drastic morphological
changes noticed are consistent with the gel electrophoresis results.

ACCEPTED MANUSCRIPT
Figure 8. AFM micrographs of plasmidic pBR322 DNA incubated for 24 h (a) without
complex and in the presence of (b) 10 µM 2, (c) 50 µM 2, (d) 10 µM 3 and (e) 50 µM 3.
[DNA]bp = 15 µM.

ACCEPTED MANUSCRIPT
3.5. In vitro cytotoxic activity
First, single-point cell-viability assays (MTT) were carried out with the free ligands
(curcumin and caffeine) and their complexes 1−3 at two concentrations, namely 10 µM
and 50 µM (except for 1), using five cancer cell lines, namely A549 (lung
adenocarcinoma), A375 (melanoma), MCF-7 (breast adenocarcinoma), SKOV3 (ovary
adenocarcinoma) and SW620 (colorectal adenocarcinoma). The results obtained upon
incubating the freshly prepared solutions of the compounds with the cells during 24 h
are depicted in Figure S6. Compounds [Pt(curc)(PPh3)2]Cl (1) and [PtCl(curc)(DMSO)]
(2) exhibited the best activities, particularly against SW620 and MCF-7 cells. It should
be noted that the extremely low aqueous solubility of the curcumin-based complex 1 did
not allow testing it at 50 μM, where visible aggregates formed in the wells. Neither
caffeine nor its complex trans-[Pt(caffeine)Cl2(DMSO)] (3) showed significant toxicity
against the selected cell lines.
Next, dose-response curves and the corresponding IC50 values were produced for the
cancer cell lines SW620 and MCF-7, and the non-tumorigenic cell line MCF-10A to
assess the potential selectivity of the compounds towards malignant cells (Table 1).
Because of the above-mentioned low aqueous solubility of compound 1, the
corresponding IC50 at %DMSO below cell tolerance could not be obtained.
Since the described compounds herein are inspired by cisplatin, this drug has been
selected as reference for these studies. Cisplatin showed remarkably different
cytotoxicities depending on whether its stock solution was prepared in DMSO or in
PBS. The higher stability of the complex cis-[Pt(DMSO)2(NH3)2], which results from
the replacement of the chlorido ligands of cisplatin by the softer S-ligand DMSO, most
likely is responsible for its inertness and consequent lack of cytotoxicity (IC50 > 100
µM). In fact, cisplatin dissolved in DMSO was nontoxic for all the cell lines tested
(maximum concentration used 100 µM, data not shown), whereas cisplatin dissolved in
PBS reduced the cell viability at lower doses (IC50 ranging from 12 to 28 µM).
The nontoxic nature of DMSO-dissolved cisplatin is in sharp contrast with the high
toxicity of the Pt(II) complexes 1 and 2, which were also (and necessarily) dissolved in
DMSO prior to their addition to the cell media. Curcumin and its complex 2 showed
interesting antiproliferative activities and revealed a clear enhancement of the activity
upon metal complexation; for instance, compound 2 exhibited an IC50 value as low as
6.1 µM against SW620 cells while the IC50 of free curcumin was 15.5 µM. With MCF-7
cells, the toxicity increase was even higher upon metal complexation. To assess the
effect of the hydrolysed form of 2 in combination with the released curcumin, the
compounds were pre-incubated for 24 h in the cell media before IC50 determination. The
resulting toxicity was significantly lower in all cell lines for both curcumin and the
hydrolysis products of 2. This observation may be due to the sequestration/binding of
the hydrolysed platinum(II) compound (see Scheme 2) and the free curcumin by the
serum proteins present in the cell media; curcumin can interact with serum albumins
through the hydroxyl phenolic groups, [66] and the aquated form(s) of platinum(II)
complex 2 may bind covalently with protein thiols. Similarly, the cytotoxic effect of
cisplatin was reduced when it was pre-incubated in the cell media (> 85% cell viability
at 20 µM).
Next, we evaluated the effect of the photosensitizing nature of curcumin and its
complexes on their cytotoxicity against SW620 and MCF-7 cells, and the non-
tumorigenic cells MCF-10A. Curcumin reaches a triplet state upon photoactivation with
visible light (strong absorption centred at 430 nm), and it reacts with molecular oxygen

ACCEPTED MANUSCRIPT
to generate reactive oxygen species (ROS).[28] The irradiation of free curcumin and
compound 2 (both freshly prepared and pre-incubated) with visible light for 1 h
increased their toxicity (Table 1). According to the IC50 values obtained, the
photosensitization was generally higher for free curcumin than for 2. In line with this
observation, the photosensitizing effect was also more significant for the hydrolysed
forms of the compounds. The smaller differences observed with SW620 cells upon
irradiation may be attributed to the higher amounts of the ROS-scavenger cysteine
contained in the culture media used (for this cell line). Expectedly, the irradiation of
cisplatin with visible light did not improve its toxicity.
In all cases (namely with or without pre-incubation and with or without visible-light
irradiation), [PtCl(curc)(DMSO)] (2) displayed similar or higher cytotoxicities than its
parent compound, viz. cisplatin. Furthermore, the freshly prepared solution of 2 showed
better selectivity towards cancer cells than both curcumin and cisplatin.
Table 1. IC50 values obtained with cisplatin (stock solution prepared in PBS), free
curcumin, and complex 2, against MCF-7 (breast adenocarcinoma), MCF-10A (non-
tumorigenic epithelial breast cell line) and SW620 (colorectal adenocarcinoma) cells,
after incubation for 24 h at 37 °C. Freshly prepared or pre-incubated (for 24 h) solutions
of the compounds in the culture medium were evaluated. Once the compounds were
added, the cells were incubated for 1 h at room temperature in the dark or under visible-
light irradiation, and for further 23 h at 37 ºC. The results are means ± SD of four
independent experiments.
IC50 (μM)
MCF7
28.3 ± 2.2
26.2 ± 1.0
>20
SW620
17.2 ± 8.8
23.1 ± 0.5
>20
MCF10A
12.0 ± 0.8
13.1 ± 0.2
>20
Non-irradiated
Irradiated
Freshly prepared
Pre-incubated 24 h
Freshly prepared
Pre-incubated 24 h
Freshly prepared
Pre-incubated 24 h
Cisplatin (PBS)
Curcumin
Non-irradiated
Irradiated
>20
>20
>20
40.5 ± 10.2
5.6 ± 1.4
>100
Non-irradiated
Irradiated
15.5 ± 2.9
10.8 ± 3.4
85.6 ± 4.6
51.1 ± 6.3
6.1 ± 1.8
4.1 ± 1.0
37.4 ± 18.9
5.9 ± 1.7
>100
Non-irradiated
Irradiated
36.4 ± 12.2
11.6 ± 5.0
2.5 ± 0.3
66.1 ± 12.9
19.5 ± 6.8
60.2 ± 22.5
22.9 ± 9.9
4.3 ± 1.9
Non-irradiated
Irradiated
Compound 2
Non-irradiated
Irradiated
83.0 ± 15.6 85.2 ± 23.8
36.5 ± 3.3 31.5 ± 11.9
4. Conclusions
Three new Pt(II) complexes, containing curcuminate, viz. [Pt(curc)(PPh3)2]Cl (1) and
[PtCl(curc)(DMSO)] (2), or caffeine, viz. trans-[Pt(caffeine)Cl2(DMSO)] (3), were
synthesized and characterized. The curcuminate ligands coordinate through its β-
diketonate moiety, while caffeine binds Pt through its N9 atom. A thorough study of
their interaction with DNA through a range of spectroscopic techniques, AFM and gel
electrophoresis, revealed distinct behaviours for the different compounds. Complex 1
tends to intercalate in between nucleobase pairs while for complex 2, which contains a
chloride and DMSO ligands instead of PPh3, the binding seems to occur in the grooves,

ACCEPTED MANUSCRIPT
predominantly driven by electrostatic forces at high compound:DNA molar ratios.
Complex 3 appears to interact by covalent bonding with DNA nucleobases, in a
cisplatin-like fashion. The three complexes underwent important changes during the
first 24 h in solution, and this fact determined their interaction with DNA; this
observation has been nicely illustrated in the case of 3. Speciation studies are required
to determine the nature of the species generated.
In addition, the cytotoxic activity of these Pt(II) compounds has been evaluated in vitro
against five cancer cell lines (A549, A375, MCF-7, SKOV3 and SW620) and one
normal breast cell line (MCF-10A) to assess their selectivity. The curcuminate
complexes 1 and 2 were rather cytotoxic after 24 h incubation with the cells, while the
caffeine complex 3 was not. Remarkably, the curcuminate complex 2 exhibited IC50
values in the low micromolar range. The curcuminate moiety allowed the
photosensitization of this complex, producing an even more efficient system. This
“combined” compound surpassed its parent drugs cisplatin and curcumin, both in
toxicity and selectivity. The pre-incubation of all compounds in the cell media lowered
their efficacy, which is ascribed to their sequestration by the serum proteins. This
observation highlights the importance of including pre-incubation tests in routine cell-
viability studies in order to better assess their potential therapeutic use.
The data achieved in the present study clearly illustrate the high potential of curcumin-
based metal complexes for possible applications in anticancer drug design.
Acknowledgements
The authors gratefully acknowledge financial support from MICINN (project
CTQ2017-88446-R AEI/FEDER, UE) and from Italian MIUR (PRIN project 2010–
2011 n. 2010FPTBSH, NANOMED) and from ISCIII (FIS PI18/00441, FEDER).
A.B.C. acknowledges the European Union’s Horizon 2020 research and innovation
programme for her Marie Skolodowska-Curie grant No. 656820. P.G. acknowledges the
Institució Catalana de Recerca i Estudis Avançats (ICREA). P.M. thanks Dr Hamid
Reza Shahsavari for useful discussions. MPH acknowledges Fundación la Caixa for her
postdoctoral fellowship.

ACCEPTED MANUSCRIPT
References
[1] J. Ferlay, I. Soerjomataram, M. Ervik, R. Dikshit, S. Eser, C. Mathers, M. Rebelo, D.M. Parkin,
D. Forman, F. Bray, vol. 2017, International Agency for Research on Cancer, Lyon, France, 2013.
[2] F. Bray, A. Jemal, N. Grey, J. Ferlay, D. Forman, Lancet Oncol, vol. 13, 2012, pp. 790-801.
[3] R.F. Brissos, L. Korrodi-Gregório, R. Péres-Tomás, O. Roubeau, P. Gamez, Chem. Sq., vol. 2,
2018, 4.
[4] J.-A. Cuello-Garibo, C.C. James, M.A. Siegler, S. Bonnet, Chem. Sq., vol. 1, 2017, 2.
[5] A. Presa, G. Vazquez, L.A. Barrios, O. Roubeau, L. Korrodi-Gregorio, R. Perez-Tomas, P.
Gamez, Inorg. Chem., vol. 57, 2018, pp. 4009-4022.
[6] J.M. Saborit, A. Caubet, R.F. Brissos, L. Korrodi-Gregorio, R. Perez-Tomas, M. Martinez, P.
Gamez, Dalton Trans., vol. 46, 2017, pp. 11214-11222.
[7] B. Rosenberg, L. VanCamp, J.E. Trosko, V.H. Mansour, Nature, vol. 222, 1969, pp. 385-386.
[8] F. Arnesano, G. Natile, Coord Chem Rev, vol. 253, 2009, pp. 2070-2081.
[9] B.A. Chabner, T.G. Roberts, Jr., Nat Rev Cancer, vol. 5, 2005, pp. 65-72.
[10] C.A. Rabik, M.E. Dolan, Cancer Treat Rev, vol. 33, 2007, pp. 9-23.
[11] R.P. Miller, R.K. Tadagavadi, G. Ramesh, W.B. Reeves, Toxins, vol. 2, 2010, pp. 2490-2518.
[12] F.E. de Jongh, R.N. van Veen, S.J. Veltman, R. de Wit, M.E. van der Burg, M.J. van den Bent,
A.S. Planting, W.J. Graveland, G. Stoter, J. Verweij, Br J Cancer, vol. 88, 2003, pp. 1199-1206.
[13] C. Monneret, Ann Pharm Fr, vol. 69, 2011, pp. 286-295.
[14] G. Momekov, D. Momekova, Expert Opin Ther Pat, vol. 16, 2006, pp. 1383-1403.
[15] T. Boulikas, M. Vougiouka, Oncol Rep, vol. 10, 2003, pp. 1663-1682.
[16] T.C. Johnstone, K. Suntharalingam, S.J. Lippard, Chem Rev, vol. 116, 2016, pp. 3436-3486.
[17] S. Wanninger, V. Lorenz, A. Subhan, F.T. Edelmann, Chem Soc Rev, vol. 44, 2015, pp. 4986-
5002.
[18] M. Pröhl, U.S. Schubert, W. Weigand, M. Gottschaldt, Coord Chem Rev, vol. 307, 2016, pp.
32-41.
[19] U. Ndagi, N. Mhlongo, M.E. Soliman, Drug Des Devel Ther, vol. 11, 2017, pp. 599-616.
[20] S. Nafisi, M. Adelzadeh, Z. Norouzi, M.N. Sarbolouki, DNA Cell Biol, vol. 28, 2009, pp. 201-
208.
[21] M. Heger, R.F. van Golen, M. Broekgaarden, M.C. Michel, Pharmacol Rev, vol. 66, 2014,
pp. 222-307.
[22] Y. Sunagawa, Y. Katanasaka, K. Hasegawa, T. Morimoto, PharmaNutrition, vol. 3, 2015, pp.
131-135.
[23] S. Prasad, A.K. Tyagi, B.B. Aggarwal, Cancer Res Treat, vol. 46, 2014, pp. 2-18.
[24] Y. Kawano, M. Nagata, T. Kohno, A. Ichimiya, T. Iwakiri, M. Okumura, K. Arimori, Biol
Pharm Bull, vol. 35, 2012, pp. 400-407.
[25] B. Bertrand, L. Stefan, M. Pirrotta, D. Monchaud, E. Bodio, P. Richard, P. Le Gendre, E.
Warmerdam, M.H. de Jager, G.M. Groothuis, M. Picquet, A. Casini, Inorg Chem, vol. 53, 2014,
pp. 2296-2303.
[26]
[27] A.K. Renfrew, N.S. Bryce, T.W. Hambley, Chem Sci, vol. 4, 2013, pp. 3731.
[28] K. Mitra, S. Gautam, P. Kondaiah, A.R. Chakravarty, Angew Chem Int Ed Engl, vol. 54, 2015,
pp. 13989-13993.
[29] F.R. Hartley, The chemistry of platinum and palladium: with particular reference to
complexes of the elements, Applied Science Publishers (1973).
[30] Y. Yu.N. Kukushkin , E. Vyaz'menskii, L. I. Zorina, Russ J Inorg Chem, vol. 13, 1968, pp.
1573-1576.
[31] K. Nakamoto, M. Tsuboi, G.D. Strahan, Methods Biochem Anal, vol. 51, 2008, pp. 1-366.
[32] K.I. Priyadarsini, Molecules, vol. 19, 2014, pp. 20091-20112.

ACCEPTED MANUSCRIPT
[33] P. Bergamini, V. Ferretti, P. Formaglio, A. Marchi, L. Marvelli, F. Sforza, Polyhedron, vol.
78, 2014, pp. 54-61.
[34] A. Medrano, S.M. Dennis, A. Alvarez-Valdes, J. Perles, T. McGregor Mason, A.G. Quiroga,
Dalton Trans, vol. 44, 2015, pp. 3557-3562.
[35] M.M. Dell'Anna, V. Censi, B. Carrozzini, R. Caliandro, N. Denora, M. Franco, D. Veclani, A.
Melchior, M. Tolazzi, P. Mastrorilli, J Inorg Biochem, vol. 163, 2016, pp. 346-361.
[36] M. Ravera, E. Gabano, M. Sardi, E. Monti, M.B. Gariboldi, D. Osella, Eur J Inorg Chem,
2012, pp. 3441-3448.
[37] D.G. Gusev, Angew Chem Int Ed Engl, vol. 51, 2012, pp. 12930-12931.
[38] R. Benassi, E. Ferrari, S. Lazzari, F. Spagnolo, M. Saladini, J Mol Struct, vol. 892, 2008, pp.
168-176.
[39] M.S. Refat, Spectrochim Acta A, vol. 105, 2013, pp. 326-337.
[40] F. Kühlwein, K. Polborn, W. Beck, Z Anorg Allg Chem, vol. 623, 1997, pp. 1211-1219.
[41] F. Kühlwein, K. Polborn, W. Beck, Z Anorg Allg Chem, vol. 623, 1997, pp. 1931-1944.
[42] K. Nakamoto, Handbook of Vibrational Spectroscopy, Major Reference Works, 2001.
[43] M.M. Dell'Anna, U. Englert, M. Latronico, P.L. Luis, P. Mastrorilli, D.G. Papa, C.F. Nobile, M.
Peruzzini, Inorg Chem, vol. 45, 2006, pp. 6892-6900.
[44] J. do Couto Almeida, I.M. Marzano, F.C.S. de Paula, M. Pivatto, N.P. Lopes, P.C. de Souza,
F.R. Pavan, A.L.B. Formiga, E.C. Pereira-Maia, W. Guerra, J Mol Struct, vol. 1075, 2014, pp. 370-
376.
[45] S.A. De Pascali, P. Papadia, A. Ciccarese, C. Pacifico, F.P. Fanizzi, Eur J Inorg Chem, 2005,
pp. 788-796.
[46] M. Melnik, O. Sprusansky, P. Musil, Adv Biol Chem, vol. 4, 2014, pp. 274-280.
[47] N. Raman, K. Pothiraj, T. Baskaran, J Mol Struct, vol. 1000, 2011, pp. 135-144.
[48] N. Shahabadi, N. Moeini, J Coord Chem, vol. 68, 2015, pp. 2871-2885.
[49] C. Sacht, M.S. Datt, S. Otto, A. Roodt, J Chem Soc, Dalton Trans, 2000, pp. 727-733.
[50] L.G. Marzilli, Y. Hayden, M.D. Reily, Inorg Chem, vol. 25, 1986, pp. 974-978.
[51] J. Sitkowski, L. Stefaniak, L. Nicol, M.L. Martin, G.J. Martin, G.A. Webb, Spectrochim Acta
A, vol. 51, 1995, pp. 839-841.
[52] I. Łakomska, B. Golankiewicz, J. Wietrzyk, M. Pełczyńska, A. Nasulewicz, A. Opolski, J.
Sitkowski, L. Kozerski, E. Szłyk, Inorg Chim Acta, vol. 358, 2005, pp. 1911-1917.
[53] R. Ghosh, J.A. Mondal, D.K. Palit, J Phys Chem B, vol. 114, 2010, pp. 12129-12143.
[54] V.W. Yam, K.M. Wong, Chem Commun, vol. 47, 2011, pp. 11579-11592.
[55] M.J. Hannon, Pure Appl Chem, vol. 79, 2007, pp. 2243-2261.
[56] J.M. Kelly, A.B. Tossi, D.J. McConnell, C. OhUigin, Nucleic Acids Res, vol. 13, 1985, pp.
6017-6034.
[57] S.U. Rehman, T. Sarwar, M.A. Husain, H.M. Ishqi, M. Tabish, Arch Biochem Biophys, vol.
576, 2015, pp. 49-60.
[58] M. Sirajuddin, S. Ali, A. Badshah, J Photochem Photobiol B, vol. 124, 2013, pp. 1-19.
[59] Y. Zou, L. Biao, F. Xu, R. Liu, Z. Liu, Y. Fu, Scanning, vol. 38, 2016, pp. 880-888.
[60] J.B. LePecq, C. Paoletti, J Mol Biol, vol. 27, 1967, pp. 87-106.
[61] F.J. Meyer-Almes, D. Porschke, Biochemistry, vol. 32, 1993, pp. 4246-4253.
[62] R.F. Brissos, E. Torrents, F.M. dos Santos Mello, W. Carvalho Pires, P. Silveira-Lacerda Ede,
A.B. Caballero, A. Caubet, C. Massera, O. Roubeau, S.J. Teat, P. Gamez, Metallomics, vol. 6,
2014, pp. 1853-1868.
[63] G. Cosa, K.S. Focsaneanu, J.R. McLean, J.P. McNamee, J.C. Scaiano, Photochem Photobiol,
vol. 73, 2001, pp. 585-599.
[64] N. Shahabadi, N.H. Moghadam, Spectrochim Acta A, vol. 99, 2012, pp. 18-22.
[65] J.-P. Macquet, J.-L. Butour, Eur J Biochem, vol. 83, 1978, pp. 375-387.
[66] F. Mohammadi, A.K. Bordbar, A. Divsalar, K. Mohammadi, A.A. Saboury, Protein J, vol. 28,
2009, pp. 189-196.

ACCEPTED MANUSCRIPT
Graphical Abstract
Three Pt(II) compounds have been prepared that contain the natural ligands curcumin
(complexes 1 and 2) or caffeine (complex 3). DNA-intercalating 1 and groove-binding 2
exhibit cytotoxic properties, which can be enhanced through visible-light irradiation.

ACCEPTED MANUSCRIPT
Highlights



Platinum complexes from naturally occurring ligands
Selectivity towards cancer cells
Distinct DNA-binding modes