Ruthenium(II)-arene complexes with monodentate aminopyridine ligands: Insights into redox stability and electronic structures and biological activity

Article abstract of DOI:10.1016/j.jorganchem.2018.11.036

The synthesis and spectroscopic characterization of four ruthenium(II) arene complexes with monodentate pyridine derivatives ([(η⁶–p-cymene)RuCl₂L]: L = 2-aminopyridine, 2-methylaminopyridine, 2-benzylaminopyridine, and pyridine) are

Full text of DOI:10.1016/j.jorganchem.2018.11.036

Accepted Manuscript
Ruthenium(II)-arene complexes with monodentate aminopyridine ligands: Insights into
redox stability and electronic structures and biological activity
Flávia Marszaukowski, Ivelise Dimbarre Lao Guimarães, Juliana Paula da Silva, Luis
Henrique da Silveira Lacerda, Sergio Ricardo de Lazaro, Márcio Peres de Araujo,
Patrícia Castellen, Tania Toyomi Tominaga, René T. Boeré, Karen Wohnrath
PII:
S0022-328X(18)30775-7
DOI:
https://doi.org/10.1016/j.jorganchem.2018.11.036
JOM 20656
Reference:
To appear in: Journal of Organometallic Chemistry
Received Date: 4 September 2018
Revised Date: 28 November 2018
Accepted Date: 30 November 2018
Please cite this article as: Flá. Marszaukowski, I.D. Lao Guimarães, J.P. da Silva, L.H. da Silveira
Lacerda, S.R. de Lazaro, Má.Peres. de Araujo, Patrí. Castellen, T.T. Tominaga, René.T. Boeré, K.
Wohnrath, Ruthenium(II)-arene complexes with monodentate aminopyridine ligands: Insights into redox
stability and electronic structures and biological activity, Journal of Organometallic Chemistry (2019),
doi: https://doi.org/10.1016/j.jorganchem.2018.11.036.
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
Ruthenium(II)-arene complexes with monodentate aminopyridine
ligands: insights into redox stability and electronic structures and
biological activity†
Flávia Marszaukowskia, Ivelise Dimbarre Lao Guimarãesa, Juliana Paula da Silvaa, Luis
Henrique da Silveira Lacerdaa, Sergio Ricardo de Lazaroa, Márcio Peres de Araujob,
Patrícia Castellena, Tania Toyomi Tominagac, René T. Boeréd and Karen Wohnratha*
Graphical abstract

ACCEPTED MANUSCRIPT
Ruthenium(II)-arene complexes with monodentate aminopyridine
ligands: insights into redox stability, electronic structures and
biological activity†
Flávia Marszaukowskia, Ivelise Dimbarre Lao Guimarãesa, Juliana Paula da Silvaa, Luis
Henrique da Silveira Lacerdaa, Sergio Ricardo de Lazaroa, Márcio Peres de Araujob,
Patrícia Castellena, Tania Toyomi Tominagac, René T. Boeréd and Karen Wohnratha*
a Departamento de Química, Universidade Estadual de Ponta Grossa, Av. Gal. Carlos
Cavalcanti, 4748, Uvaranas, CEP 84030-900, Ponta Grossa - PR, Brasil.
b Departamento de Química, Universidade Federal do Paraná, Rua Coronel Francisco
Heraclito dos Santos, 100 - Jardim das Américas, CEP 81531-990, Curitiba – PR,
Brasil.
c Departamento de Física, Universidade Estadual do Centro-Oeste, Rua Simeão
Camargo Varela de Sá, 03, Vila Carli, CEP 85040-080, Guarapuava – PR, Brasil.
d Department of Chemistry and Biochemistry, University of Lethbridge, 4401University
Drive, Lethbridge, Alberta, Canada T1K 3M4
† Supplementary information (SI) available: X-ray crystallography data; DFT computed
and crystallographic structure depictions; vibrational spectroscopic data; archival NMR
data; Uv-vis spectroscopic data; solubility and stability of the complexes; additional
cyclic voltammograms and scan-rate studies; biological data and literature survey;
Cartesian coordinates of DFT geometries. CCDC 1822445. For SI and crystallographic
data in CIF or other electronic format see DOI: xxxxxx
*Corresponding author: Karen Wohnrath. karen.woh@gmail.com
Phone: +55 (42) 3220-3731
1

ACCEPTED MANUSCRIPT
Abstract
The synthesis and spectroscopic characterization of four ruthenium(II) arene complexes
with monodentate pyridine derivatives ([(η6–p-cymene)RuCl2L]: L = 2-aminopyridine,
2-methylaminopyridine, 2-benzylaminopyridine, and pyridine) are reported. Full
1
13
characterization was undertaken using H and C NMR spectroscopy, vibrational and
electronic spectroscopies and crystallography (2-methylaminopyridine derivative).
UB3LYP//(6-31+G(d),SPK-DZCD) density functional theory calculations determined
the molecular and electronic structures. Cyclic voltammetry determined a large
electrochemical stability window (>2.2 V) extending well beyond the physiological E°.
Interactions with CT-DNA and BSA, and activity against four cell lines (HeLa,
B16F10, HEp-2 and Vero) were evaluated. The 2-methylaminopyridine shows weak
cytotoxicity (IC50 = 346 µmol·L-1) towards HeLa cells. All the complexes interact with
DNA at relatively high concentrations as determined by UV-vis spectroscopic titration.
Results of circular dichroism spectroscopy, ethidium bromide competition, fluorescence
spectroscopy and DNA viscosity measurements identify electrostatic interactions
between partly hydrolyzed cationic complexes and the phosphate backbone of DNA as
the most likely interaction mode. Slower rates of hydrolysis may be the origin of lower
cytotoxicity for κ1 these complexes.
Keywords
:
Ruthenium(II)-arene; pyridine ligands; DNA interaction; BSA interaction; cytotoxicity;
crystal structure.
2

ACCEPTED MANUSCRIPT
Abbreviations table
B16F10
Murine melanoma cell line
BSA
Bovine serum albumin
CT-DNA
Calf-Thymus deoxyribonucleic acid
DMSO
Dimethyl sulfoxide
EtBr
Ethidium bromide
HeLa
Cervical carcinoma cell line
HEp-2
Laryngeal carcinoma cell line
PBS
Phosphate buffer
RAPTA
Ruthenium arene PTA
TBAP
Tetrabutylammonium perchlorate
Vero
Kidney murine cell line
3

ACCEPTED MANUSCRIPT
1.
Introduction
Interest in utilizing ruthenium-based metal complexes as vital new drugs for the
treatment of cancer, either to replace platinum therapies where resistance has developed
or to expand the therapeutic range of tumor types, remains very strong [1-9]. Whereas
all the platinum-based drugs in clinical use are based on square-planar L2PtCl2 in the
classic cis configuration, the range of structures, ligand-types and geometries of
ruthenium complexes which have demonstrated anti-proliferative action is extremely
diverse. In this regard, a strong warning has recently been given against making false
generalizations (low toxicity because of similarity to iron; slow rates of ligand
exchange; activation is by reduction to Ru(II); specific accumulation in cancerous
tissues; uptake mediated by transferrin) about ruthenium cytotoxic agents [7]. In short,
different types of ruthenium-based metallodrugs are under investigation, with
distinctive behaviors. One class of ruthenium complex that differs significantly in
structure and properties from the platinum-type agents are (η6-arene)ruthenium(II)
tripodal organometallic complexes, the vast majority of which have either one (type A)
or two (type B: E = NR) nitrogen donor ligands (Chart 1). The remaining coordination
sites are almost always occupied by chlorides. The type A complexes are thus neutral
ruthenium complexes with κ1-pyridyl ligands [1]. In the chelated type B complexes, the
second donor group “E” is frequently nitrogen but can also be oxygen or carbon
moieties [5]. There are now numerous derivates of type A where the nitrogen ligand is a
simple pyridine group [10-19].
Chart 1 Structure types and ligands of importance to this work.
In the first step of a larger research program utilizing pyridine-based ligands of
both types A and B, we wished to investigate the cytotoxicity of a selected series of
simple neutral 2-aminopyridine complexes incorporating the ligands 2-aminopyridine
(apy), 2-methylaminopyridine (meapy) and 2-benzylaminopyridine (bzapy) (Chart 1).
2-Aminopyridines are important nitrogen-containing ligands due to a strong nitrogen
4

ACCEPTED MANUSCRIPT
donor in the ring, while the aromatic amino substituent provides additional electronic
and hydrogen-bonding stability. Such complexes have attracted considerable interest
recently because of their applications in pharmaceutical research, for instance
glucokinase activators [6] or selective inhibitors of neuronal nitric oxide synthase [21].
Herein we report on the synthesis, structures and biological activity of four [(η6-p-
cymene)RuCl2L] complexes 1 – 3 where L = apy, meapy and bzapy, respectively. In
addition to cytotoxicity tests in vitro against cancer and normal cell lines (HeLa,
B16F10, HEp-2 and Vero), studies on the interaction of the complexes with DNA and
BSA were undertaken to search for understanding of biological activity. The long-
known complex 4 of the parent ligand py was included for comparison to the
aminopyridines [13,22-25]. Complex 4 did not display any cytotoxic effect as
previously reported in the literature [26].
While this work was in progress, studies of a number of other type-A pyridine
complexes appeared in the literature [24-28]. The biological properties of these
analogues, in particular the substituted 2-phenylaminopyridine derivatives reported by
Richter et al. [27], serve as interesting parallels to our results. By incorporating remote
carboxylate donor groups at the 4-position of the NC6H4X rings, moderate
antiproliferative activity was observed. A comparison of this closely parallel series with
1 to 4 (i.e. with and without these remote donor groups) proved helpful for delineating
their structure-property relationships.
Quantum calculations on ruthenium complexes are relatively rare in the
literature but there has been an increase in interest in using such such methods to
investigate molecular structures [29-30] and catalysis [31] of Ru complexes. Kreitner
and co-authors applied Density Functional Theory (DFT) in a study of excited state
behaviour of cyclometalated bis(tridentate)ruthenium(II) complexes [32], while Das and
co-authors investigated the interaction of aquated ruthenium(III) complexes with DNA
base pairs using computational methods [33]. Herein we employ DFT computational
methods to provide a basic understanding of the electronic structures of the title
complexes to provide insights into their spectroscopic and redox behaviours in vitro and
in vivo. To date, these aspects of ruthenium-arene complexes have been unduly ignored
despite an intensive literature regarding their potential utility, or otherwise, as new
antiproliferative agents.
5

ACCEPTED MANUSCRIPT
2.
Experimental section
2.1. Materials and general methods
RuCl3.3H2O was purchased from Strem and the precursor complex [{(η6 p-
cymene)Ru(µ-Cl)Cl}2] was prepared according to published procedures [34,35].
Ligands apy, meapy, bzapy and py, as well as α-phellandrene, calf thymus DNA (CT-
DNA), bovine serum albumin (BSA) and tetrabutylammonium perchlorate (TBAP)
were purchased from Sigma-Aldrich and used without further purification. The solvents
were rigorously purified by standard procedures [36]. All synthesis manipulations were
carried out under an argon atmosphere using modified Schlenk techniques.
2.2. Physical measurements
FT-IR spectra (4000-550 cm-1) were recorded on a DRS-8000/Shimadzu
IRPrestige-21 spectrometer. Raman spectra (40-4400 cm-1) were obtained with a Bruker
Senterra dispersive Raman microscope. UV-vis spectra (0.1 mmol) were recorded on a
Varian Cary 50 Bio spectrophotometer using quartz cells, in the range 200-900 nm.
Conductivity values were obtained using an Infolab WTW TetraCon® 325 conductivity
bridge in a thermostated bath held at 25.0 °C. Aqueous solution of 1·10-3 mol·L-1 NaCl
was used as the 1:1 electrolyte standard, where the conductivity value for this solution
was 124.7 µS·cm-1. [nBu4N][ClO4] was used similarly used as the standard in CH3CN,
for which the molar conductance is 197.1 µS·cm−1 [37]. X-ray crystallography was
undertaken for complex 2 using Cu Kα radiation at 100 K on a Rigaku-Oxford
Diffraction SuperNova diffractometer equipped with a Pilatus 200K HPAD detector.
Details of the structure solution and refinement are provided in the SI. The structure is
complicated by wholesale disorder of the meapy ligand, presence of CHCl3 solvent and
Z' = 2. CCDC 1822445 contains the supplementary crystallographic data for this paper.
13
1
1D and 2D solution-phase NMR experiments (1H, C, H–1H gCOSY,1H–13C
1
gHSQC, and H–13C gHMBC) were recorded on a Bruker Model DRX, 400 MHz
spectrometer, at probe temperature using, in general, 20 mg samples of complexes
dissolved in CDCl3, containing a trace amount of tretramethylsilane (TMS) that was
6

ACCEPTED MANUSCRIPT
used as an internal reference (0 ppm). The 1H and 13C NMR spectra were acquired with
16 and 2048 scans; spectral widths (sw) of 6393.862 and 25510.203 Hz; relaxation
1
delays (d1) of 1 and 0.2 s; and 90° and 80° pulse lengths, respectively. The H–1H
1
1
gCOSY, H–13C gHSQC and H–13C gHMBC, spectra were acquired with 8, 16, and 8
scans and spectral widths of 4595.588 for F2 and 18852.455 for F1, respectively. The
relaxation
delay
of
1 s, with 256 data points at F1 and 4 K at F2 were the same in all 2D NMR experiments.
Archival spectral data is presented in the SI.
2.3. Computational methodology
DFT calculations in gas phase, zero Kelvin and vacuum were carried out on 1 as
an electronic model for all the complexes and on 2 to verify the geometry due to the
disorder encountered in the X-ray diffraction study. All models were fully optimized
from 10-6 Hartree and 10-8 Hartree for self-consistent field (SCF) based on the Hartree-
Fock formalism and total energy criteria, respectively. Harmonic frequency calculations
were undertaken to confirm that the geometries are at least local minima on the potential
energy surface. Similar calculations were undertaken for 1+ (doublet), 12+ (singlet and
triplet), 1− (doublet) and 12− (singlet and triplet) to help interpret the voltammetry
results. In this study, all calculations were performed on GAUSSIAN09 program using
B3LYP [38] functional with Gaussian-type 6-31+G(d) basis set for C, H, N, P, and F
atoms; whereas the all-electron, relativistically corrected, Sapporo double-zeta (SPK-
DZCD) basis set was applied to describe the Ru atom [39].
2.4. Electrochemistry
Cyclic voltammetric experiments were recorded on a potentiostat/galvanostat
µAutolab (Type III, Metrohm-Eco Chemie) connected to a computer with GPES 4.9
(General Purpose Electrochemical System) software. The measurements were
performed in nitrogen atmosphere at room temperature in dry CH3CN with
tetrabutylammonium perchlorate (TBAP Sigma-Aldrich) in 1.0·10-3 mol·L-1 as a
supporting electrolyte. The electrochemical cell was equipped with a glassy carbon (A =
3 mm2) working electrode, a platinum foil auxiliary electrode and Ag/AgCl as the
reference electrode in a Luggin capillary probe. Voltammograms were performed at a
scan rate of 50 to 2000 V·s-1, with complex concentrations of 1 mM. The
7

ACCEPTED MANUSCRIPT
ferrocenium/ferrocene redox couple was used as an internal reference (E1/2 = 0.46 V vs
Ag/AgCl).
2.5. Synthesis of complexes 1 - 4
Synthesis of [(η6 p-cymene)RuCl2(apy)] (1). Complex 1 was synthesized according to
the literature method [40]. A solution of the precursor (100 mg, 0.16 mmol) with an excess
of 2-aminopyridine (76 mg, 0.82 mmol) in toluene (10 mL) was stirred for 12 h at room
temperature. The orange solid that precipitated was filtered off, washed with diethyl ether
and dried under vacuum. Yield: 82 mg, 63%. Elemental analysis (%) calcd. for
C15H20Cl2N2Ru (400.34 g·mol-1): C, 45.01; H, 5.04; N, 7.00. Found (%): C, 44.92; H, 5.17;
N, 6.70. Its identity was established by agreement of the 1H NMR and IR with the original
13
report. 1H NMR: (see Table 2). C (400 MHz, CDCl3, δ ppm): 18.21 (Ca); 22.27 (Cg);
30.49 (Cf); 81.70 (Cc); 82.97 (Cd); 97.75 (Cb); 103.05 (Ce); 112.21 (Ck); 114.04 (Ci);
138.47 (Cj); 152.24 (Ch); 162.59 (Cl). (FTIR, cm-1): 3372 and 3290 νasN-H (w); 3040
2
3
νasCsp -H (w); 2966 and 2863 νasCsp -H (w); 1594 νasC=N (s); 1611, 1469 and 1438 νasC=C
(s); 1251 νC-H (w); 1061 νasc (w); 753 δC-H (s). UV-vis. (CH3CN, Max nm): 419 (420
mol-1·L·cm-1), 292 (5600 mol-1·L·cm-1), 233 (9200 mol-1·L·cm-1).
Synthesis of [(η6-p-cymene)RuCl2(meapy)] (2). A solution of the precursor (200 mg,
0.32 mmol) with an excess of 2-methylaminopyridine (173 mg, 1.60 mmol) in toluene
(10 mL) was stirred for 4 h at room temperature. The orange solid that precipitated was
filtered off, washed with diethyl ether and dried under vacuum. Yield: 238.7 mg, 88.5%.
An analytical sample was obtained by vapour diffusion of diethyl ether into a
concentrated CHCl3 solution at RT (with some occluded CHCl3). Elemental analysis
(%) calcd. for C16H22Cl2N2Ru·0.17CHCl3 (414.32 g·mol-1): C, 44.68; H, 5.14; N, 6.45.
Found (%): C, 44.25; H, 5.09; N, 6.64. Removal of residual CHCl3 was achieved by
redissolving and precipitating with diethyl ether. Final purity was monitored by very
1
careful NMR measurements showing the absence of C or H containing impurities. H
NMR: (see Table 2). 13C (400 MHz, CDCl3, δ ppm): 18.14 (Ca); 22.50 (Cg); 30.12 (Cf);
30.51 (Cm); 82.07 (Cc); 82.74 (Cd); 97.34 (Cb); 103.47 (Ce); 108.81 (Ck); 113.28 (Ci);
2
138.99 (Cj); 153.62 (Ch); 163.13 (Cl). (FTIR, cm-1): 3259 νasN-H (w); 3051 νasCsp -H (w);
3
2953 and 2870 νasCsp -H (w); 1573 νasC=N (s); 1615, 1473 and 1429 νasC=C (s); 1259 νC-H
(w); 1073 νasc (w); 853 δC-H (s), 749 δC-H (s). UV-vis. (CH3CN, Max nm): 419 (800 mol-
1·L·cm-1), 305 (5400 mol-1·L·cm-1), 242 (22000 mol-1·L·cm-1).
8

ACCEPTED MANUSCRIPT
Synthesis of [(η6 p-cymene)RuCl2(bzapy)] (3). Complex 3 was synthesized according
to the literature method [40]. A solution of the precursor (200 mg, 0.32 mmol) with an
excess of 2-benzylaminopyridine (300 mg, 1.60 mmol) in toluene (10 mL) was stirred for
4 h at room temperature. The orange solid that precipitated was filtered off, washed with
diethyl ether and dried under vacuum. Yield: 237.4 mg (74.3%). Elemental analysis (%)
calcd. for C15H20Cl2N2Ru (400.34 g·mol-1): C, 53.88; H, 5.34; N, 5.71. Found (%): C,
53.62; H, 5.37; N, 6.75. It proved impossible to obtain this complex in pure crystalline
form and crystals suitable for X-ray diffraction could not be obtained. To obtain pure
material, it was repeatedly dissolved and re-precipitated as an amorphous solid with ether,
followed by vacuum drying. Finally, the identity and purity could be confirmed by very
careful 1H NMR analysis. 1H NMR (see Table 2). 13C (400 MHz, CDCl3, δ ppm): 19.49
(Ca); 23.67 (Cg); 31.91 (Cf); 49.12 (Cm); 83.43 (Cc); 84.10 (Cd); 98.91 (Cb); 104.84 (Ce);
109.08 (Ck); 115.06 (Ci); 128.99 (Cp); 130.39 (Co); 138.84 (Cn); 140.38 (Cj); 154.96
(Ch); 163.22 (Cl). (FTIR, cm-1): 3238 νasN-H (w); 3046 νasCsp -H (w); 2953 and 2860
2
3
νasCsp -H (w); 1573 νasC=N (s); 1618, 1475 and 1434 νasC=C (s); 1237 νC-H (w); 1067 νasc
(w); 843 δC-H (s), 763 and 703 δC-H (s). UV-vis. (CH3CN, Max nm): 420 (600 mol-
1·L·cm-1), 305 (4400 mol-1·L·cm-1), 242 (17200 mol-1·L·cm-1).
Synthesis of [(η6 p-cymene)RuCl2(py)] (4). The complex 4 was synthesized according
to the previously reported method [41-42]. A solution of the precursor (100 mg, 0.16
mmol) with an excess of pyridine (30 mg, 0.37 mmol) in toluene (5 mL) was stirred for 4 h
at room temperature. The orange solid that precipitated was filtered off, washed with
diethyl ether and dried under vacuum. Yield: 108.5 mg (80.2%). Elemental analysis (%)
calcd. for C15H20Cl2N2Ru (400.34 g·mol-1): C, 46.76; H, 4.97; N, 3.64. Found (%): C,
1
46.75; H, 4.74; N, 3.94. Its identity and purity were established by agreement of the H
13
NMR data with ref [40]. 1H NMR (see Table 2). C (400 MHz, CDCl3, δ ppm): 19.61
(Ca); 23.71 (Cg); 32.08 (Cf); 83.68 (Cc); 84.26 (Cd); 95.51 (Cb); 104.99 (Ce); 125.94 (Ci);
2
3
138.97 (Cj); 156.36 (Ch). (FTIR, cm-1): 3046 νasCsp -H (w); 2963 and 2852 νasCsp -H (w);
1531 νasC=N (s); 1600, 1468 and 1437 νasC=C (s); 1210 νC-H (w); 1067 νasc (w); 881 δC-H
(s). UV-vis. (CH3CN, Max nm): 408 (720 mol-1·L·cm-1), 244 (28200 mol-1·L·cm-1).
2.6. DNA interaction studies
All measurements with calf-thymus deoxyribonucleic acid (CT-DNA) were
taken in a PBS buffer (NaCl 0.137 mol; KCl 2.68·10-3 mol; KH2PO4 1.47·10-3 mol;
9

ACCEPTED MANUSCRIPT
Na2HPO4 0.016 mol; pH 7.6). The CT-DNA concentration per nucleotide was
determined by absorption spectrophotometric analysis using a molar absorption
coefficient of 6600 mol−1·L·cm−1 at 260 nm [43]. The spectroscopic titrations were
carried out by adding increasing amounts of CT-DNA to a solution of the complex in a
quartz cell and recording the UV–vis spectrum after each addition. The binding
affinities (Kb) were obtained by using the Benesi-Hildebrand equation: [DNA]/(εa-εf) =
[DNA]/(εb-εf) + 1/[Kb(εb-εf)] [44], where εa is the apparent molar absorptivity, which
corresponds to the ratio between an absorption of the measurement and a concentration
of the complex (Aobserved [complex]); εf is molar absorptivity of the free complex
(without addition of DNA); εb is molar absorptivity of the DNA-bound complex; Kb is
the binding constant. Plotting a graph of [DNA]/(εa - εf) versus [DNA] gives the ratio of
the angular and linear coefficients of intrinsic binding (Kb) between the complex and
DNA.
Ru-complex/CT-DNA solution viscosities at different concentration ratios were
measured using a Lovis 2000 M/ME Rolling-Ball Viscometer maintained at 25 °C in a
constant temperature bath. Aqueous solutions of CT-DNA were studied by viscosity
measurements at ambient pressure. The DNA concentration was fixed at 20 µmol·L-1,
and flow time was measured with a digital stopwatch. The mean values of three
measurements were used to evaluate the viscosity η of the samples. Specific viscosity
was plotted as a function of DNA and Ru-complex concentrations. The values for
relative specific viscosity (η/η0)1/3, where η0 and η are the specific viscosity
contributions of DNA in the absence (η0) and in the presence of the complex (η), were
plotted against [complex]/[DNA] [45].
Circular dichroism spectra were measured on a Jasco J-810 spectropolarimeter
equipped with a Peltier temperature control unit held at 25°C (Jasco Corp. Tokyo,
Japan). CT-DNA and Ru-complexes were measured alone or at different mixture
concentrations in PBS (pH 7.6) in a 1 mm path length quartz cell between 220 and 340
nm at a scanning speed of 100 nm·min−1 and by the averaging of 10 scans. The absence
of CD signal for Ru-complexes (200 µmol·L-1) was verified. Modification of the
mixture signal was monitored after addition of Ru-complexes solutions to a fixed
concentration of CT-DNA solution in two ratios [DNA] [Ru-Complex] 1:1 and 1:2.
Interaction of Ru-complexes with DNA was studied by ethidium bromide (EtBr)
competition assays. All measurements were performed on a Varian Cary Eclipse
Fluorescence Spectrophotometer using a 1 cm pathlength cuvette. These competition
10

ACCEPTED MANUSCRIPT
experiments were carried out in PBS (pH 7.6), by keeping the molar ratio of DNA
(nucleotide) to EtBr (5:1) constant and varying the Ru-complex concentrations (0-420
µmol·L-1). The excitation wavelength was 530 nm, and the emission range was set
between 550 and 700 nm for all bromide fluorescence measurements.
2.7. BSA interaction studies
Circular dichroism spectra of bovine serum albumin (BSA) were recorded
using a Jasco J-720 Spectropolarimeter at 25°C The measurements in presence and
absence of the complexes were made in the range of 203–260 nm using a 0.1 cm cell
with ten scans averaged for each CD spectra. The BSA concentration was maintained at
2.5 µmol·L-1, and the molar ratio of complexes to BSA concentration was 1:2, 1:1 and
10:1. The thermal denaturation experiments were performed over 15–95°C, with
intensity measurements taken at 208 and 222 nm, every 5 C°. Melting temperatures
were calculated with sigmoidal fit employing the software Origin (Microcal).
2.8. Cell culture and Cytotoxicity assays
In vitro cytotoxicity assays on cultured human tumor cell lines still represent the
standard method for initial screening of antitumor agents. The complexes were assayed
against human cell lines: cervical carcinoma HeLa (ATCC® CCL-2™), the complexes
1 and 2 were assayed against human laryngeal carcinoma HEp-2 (ATCC CCL-23) and
the complex 1 against the murine melanoma B16F10 (ATCC® CCL-6475™) and Vero
Cell (ATCC CCL-81) derived from the kidney of an African green monkey. The cells
were routinely maintained with Iscove's Modified Dulbecco's medium, supplemented
with 10% fetal bovine serum (FBS), at 37 °C in a humidified 5% CO2 atm. For the
cytotoxicity assay, 5 x 105 cells·mL-1 were seeded in 200 µL of complete medium in 96-
well plates. Stock solutions were prepared by dissolving the complexes in dimethyl
sulfoxide (DMSO) followed by dilution with PBS, and serial dilutions of these stock
solutions were made using culture media. In this way the lowest possible DMSO
concentration was used in these experiments. The cells were exposed to the complex in
different concentrations for a 24 and 48 h period. However, it was necessary to deviate
from the standard MTT test protocol because of the sensitivity of 1 to 4 to ligand
displacement by DMSO (see below). In this modified protocol, the stock solutions were
prepared by dissolving the complexes in PBS buffer. The viability of cultured cells for
these protocols described above was evaluated using MTT (3-(4,5-dimethylthiazol-2-
11

ACCEPTED MANUSCRIPT
yl)-2,5-diphenyltetrazolium bromide) assays [46]. In this method, the MTT conversion
to formazan by metabolically viable cells was monitored by SpectraMax 190 Microplate
Reader (Molecular Devices) at 540 nm. Cell survival rate (%) versus drug concentration
(logarithmic scale) was plotted to determine the IC50 (drug concentration at which 50%
of the cells are viable relative to the control), with its estimated error derived from the
average of 3 trials.
3. Results and discussion
3.1. Synthesis and characterization of Ruthenium(II)-arene complex
The synthesis of the series of [(η6 p-cymene)RuCl2L] complexes (Scheme 1)
was achieved via bridge cleavage of [{(p-cymene)Ru(µ-Cl)Cl}2] with a 5:1 ratio of
ligands L (apy, meapy and bzapy) in toluene at RT for 1 - 3. Complex 4 required a
2.3:1 mole ratio of py in toluene at reflux to achieve complete reaction. The complexes
were prepared as orange solids, stable to light and in air, with yields ranging from 63-
88%. In addition, all the complexes are soluble in water, halogenated solvents and polar
organic solvents such as DMSO, dichloromethane and acetonitrile, but insoluble in
diethyl ether (see Figures S11 - S14). The molar conductance of CH3CN solutions of 1
– 4 were measured (1.00·10−3 M) after mixing and after 24 h to determine if solvolysis
was a factor for the voltammetric and electronic spectroscopic experiments. The results
are convincingly attributable to non-electrolyte solutions with no change after 24 h
within experimental error (Table 1) [36]. In addition, conductivity measurements on
aqueous solutions of 2 were performed (1.00·10−3 M) to confirm the labilization of the
chloride ligand and the results suggest that partial hydrolysis occurs rapidly.
12

ACCEPTED MANUSCRIPT
Scheme 1 Synthesis and structures of complexes 1-4.
13

ACCEPTED MANUSCRIPT
Table 1 Molar conductance data for the series of complexes [(η6 p-cymene)RuCl2L]
Λm H2O (µS·cm-1) b
Complex Λm CH3CN (µS·cm-1)
Λm CH3CN (µS·cm-1)
0 h a
24 h a
35.5
35.1
32.6
36.8
33.4
33.3
31.4
35.6
___
69.8
___
___
1
2
3
4
a CH3CN 1·10-3 mol·L-1. b H2O 1·10-3 mol·L-1.
3.2. Molecular structures by X-ray crystallography and DFT computation
Single-crystal structures on 1 and 4 have previously been reported were
deposited in the Cambridge Structural Database, version 5.38, Nov 2016 (CSD) [47]
under CSD refcodes: JOBCOS [39] for 1 and MIXSOD [41] and MIXSOD01 [13] for
4, respectively. For complex 3, a structure of a closely related complex exists in which
bzapy is coordinated in [(η6-ethylbenzoate)RuCl2L] (3b), CSD refcode: OHICAL [48].
In this work, we report the new crystal structure of the 2-methylaminopyridine complex
2 (see Table S1, Figure S1 in the Supplementary Information (SI) for structure details).
The derived interatomic parameters of these four structures in their crystalline lattices
are compared to those from UB3LYP//6-31+G(d), SPK-DZCD hybrid DFT calculations
(Table S2, Figure S2) on 2 and more extensively on 1 as a representative electronic
model system. The crystal structure of 2 displayed considerable complexity, containing
CHCl3 solvent of crystallization, two independent molecules in the asymmetric unit and
a wholesale disorder of the meapy ligands (only the major components of the disorder
are shown in Figure 1). In each case, there are hydrogen bonds from the ligand N–H to
metal-bound chloride ligands. Note that in the Ru1 molecule for the major (75% refined
occupancy) component, the CH3NH group (N2) is oriented towards the end of the
cymene ligand bearing the isopropyl group (Cl1), whereas in the Ru2 molecule in the
major (63% refined occupancy) component, the equivalent group (N4) is oriented
towards the methyl group side of the cymene ligand (Cl3). The minor components of
each disordered pair have the meapy groups flipped towards the opposite ends of the
respective cymene ligands. The same disorder was observed for crystals grown solvent-
free from diethyl ether and chlorobenzene; apparently the molecular volume of meapy
and para-cymene are extremely similar.
14

ACCEPTED MANUSCRIPT
Fig. 1 Displacement ellipsoids (50% probability) plot of one of two similar independent
molecular structures of 2 as found in the crystal lattice, with the atom numbering
scheme. Only the major components of the meapy disorder models are shown. The H-
bond geometry is indicated by a dashed orange line (see Table 1)
All four molecules adopt very similar molecular structures to that of 2 in their
crystal lattices (see Fig. S1 in the Supporting Information). Key features are the classic
“piano-stool” architecture wherein the η6-arene groups fill three facial sites in the
pseudo-octahedral geometry at ruthenium. In each case, including 4, the coordinated
pyridine ring is oriented approximately parallel to that of the η6-arene albeit angled
downwards by ~30°. Consideration of space filling models (Figure S3) indicates that
this conformation represents a rotational minimum due to constraints between the ortho
hydrogen atoms in pyridine with Cl1,2. This conformational preference is augmented by
N2H2···Cl1 hydrogen bonding in 1 – 3b, which is retained in gas-phase DFT calculated
structures. Notably, the pyridyl ligands are all κ1 coordinated through N1 in these [(η6
p-cymene)RuCl2L] complexes, whereas the κ2-N,N’-chelating geometry is known in (2-
aminopyridine)-dichloro-bis(triphenylphosphine)-ruthenium(II),
CSD
refcode:
IHIWAX (Figure S4) [49]. Thus, despite the 2-NHR substitution in 1–3, all complexes
are confirmed to be of type A (see the Introduction). Interestingly, a recent paper reports
on several [(η6‐p‐cymene)RuCl2(2‐halogenated-5‐aminopyridines)] with unexpected
coordination from the amino group rather than the more basic ring N [50].
3.3. Solid state structural features from FT-IR and Raman spectra
The FT-IR spectra (Figure S5 and Table S3) contain the expected bands for the
pyridine ligands in addition to those of the η6-p-cymene and chloride ligands but have
low diagnostic information save for the νas(N-H) bands in 1–3 (and their absence in 4).
The observed stretching frequencies follow the inter- and intramolecular N···Cl
15

ACCEPTED MANUSCRIPT
hydrogen bonding distances from respective crystal structures of 3.234 (ν1(N-H) = 3390
cm−1) and 3.243 Å (ν2(N-H) = 3372 cm−1) in 1, and with intramolecular distances 3.165 Å
in 2 (ν(N-H) = 3259 cm−1) and for 3 3.147 Å in OHICAL (ν(N-H) = 3238 cm−1). The
correlation of N-H···Cl distances with N-H stretching frequencies is long known from
the literature [51]. The Raman spectra (Figure S6, Table S4) corroborate the FT-IR
results.
3.4. Solution structures as established by NMR
1
13
A full assignment of the solution-phase H and C NMR signals was achieved
on the basis of 1D (NMR) and 2D experiments (Table S5 and Scheme S1, along with
1
full archival spectra in Figures S7-9). These data, especially the H spectra, provide
convincing information about the solution structures and the purities of 1–4, showing
that (i) the p-cymene ring is rotationally fluxional, rendering an effective C2v symmetry
despite coordination to the ruthenium and (ii) the apy, meapy and bzapy ligands retain
similar geometries in solution in CDCl3 as deduced from the solid-state structures (see
above). Thus, whilst the NH2 signal in 1 integrates to 2H, it has a chemical shift (δ =
6.16) intermediate between that of hydrogen-bonded and non-hydrogen-bonded NH. By
contrast, the single NH proton signals of 2 (δ = 7.28) and 3 (δ = 8.10) are deshielded.
This data is consistent with the intermolecular H-bonding strengths 3 > 2 > 1 also
shown by crystallographic d(N···Cl) data and the trend of the νas(N-H) bands in the FTIR
spectra. The other 1H chemical shifts correlate well with the expected substituent effects
from the presence of the 2-amino groups on pyridine in 1 – 3 and its absence in 4. The δ
and the ∆δ values for p-cymene ring protons H2,3 are remarkably invariant for the series
and quite similar to those in the chloride-bridged precursor complex (pseudo-AB
doublets at 5.37 and 5.51 ppm, and 3JHH = 6.04 Hz). The observation of 3J(Hn,Hm) = 5.0
and 5.5 Hz in spectra of 2 and 3 in CHCl3 solution indicates lack of exchange of the NH
signals. This is also consistent with a dominant H-bonded conformation in solution.
This coupling is confirmed by gCOSY experiments (see Figure S9 in the SI).
16

ACCEPTED MANUSCRIPT
3.5. Electronic structure from DFT calculations on 1 as a model system
Fig. 2 Frontier molecular orbitals (FMO) and energy levels from gas phase DFT
calculations on complex 1 (singlet state, left) and 1+ (doublet state, right; for clarity only
the β-spin orbitals in 1+ are shown). The energy scale at right is displaced upwards by
about 4 eV.
The electronic structure of complex 1 was examined in detail using hybrid DFT
calculations at the UB3LYP//6-31+G(d),SPK-DZCD level of theory. The neutral
molecule was geometry optimized both in the gas phase and in an aqueous solvent
model. In addition, the oxidation states −2, −1, +1 and +2 were all optimized (see
Figure S10 in the SI). The need for computation was indicated to assist with assignment
of the electronic absorption spectra and particularly the rather odd voltammetric
behavior of the complexes (see below). All complexes optimized to a reasonable
geometry with the strongest bonding between the p-cymene and metal in the neutral
(18e) state, as expected (p-cymene ring centroid to Ru distance of 1.766 Å). Both
oxidized and reduced forms have weaker bonding and the anions optimize with one
chloride ligand migrating from Ru to sites that only hydrogen-bond to the NH group. In
the 20e dianion, the arene converts to η4–coordination, fully consistent with classical
organometallic bonding models. All charge states display NH···Cl H-bonding, either to
17

ACCEPTED MANUSCRIPT
coordinated or displaced halides (see Fig. S10). The following discussion deals
specifically with neutral 1 in the gas-phase model.
The calculated electronic structure presented in Figure 2 (left) indicates that the
highest occupied orbitals (HOMO, HOMO-1, HOMO-2, corresponding closely to the t2g
set of the pseudo-octahedral geometry) have mixed Ru 3d orbitals and Cl 2p-π*
character along with minor participation from C 2p orbitals of the p-cymene ring. These
three frontier molecular orbitals (FMO) are almost degenerate and lying only slightly
lower in energy are two (accidentally) degenerate levels: HOMO-3, with mostly non-
interacting Cl pz character, and HOMO-4, the essentially unperturbed aminopyridine
filled π3 level. In turn, the two lowest unoccupied orbitals (LUMO and LUMO+1,
corresponding to the eg set) have both Ru 3d and significant Ru-Cl σ* character. Above
these, LUMO+2 is an almost unperturbed pyridine π4 MO [52].
3.6. Electrochemical characterization in solution by voltammetry
The redox behavior of 1 – 4 was investigated by cyclic voltammetry (CV). The
CVs of 1 – 4, recorded at a glassy carbon electrode in 0.1 M TBAP/CH3CN solutions as
the supporting electrolyte (vs. Ag/AgCl), are shown in Figure 3 and the pertinent data
are presented in Table 2 (see also Figures S11–S16 in the SI). Most importantly, and the
goal of the voltammetric study undertaken for this project, is the evidence for a very
I
wide redox stability window (Ewindow = 2.2 V), defined by the difference between Epa
and EpcIV, extending far out in both the anodic and cathodic regions. This provides
direct evidence for the oft-claimed stabilization of the Ru(II) oxidation state by η6-arene
ligands [9]. This stability is only marginally affected, compared to py in 4, by amino-
substitution in 1 – 3. Importantly, this range is much wider than the physiological Eo
range, between ~−1V and +1V vs. SHE, so this class of complex can be expected to
remain as Ru(II) in vivo so long as the arene ligand remains attached [53].
The detailed CV behaviour of 2, representative of the 2-aminpyridine
complexes, was studied first in the anodic region at a scan rate of 100 mV·s-1, whereby
it exhibits two very closely-spaced chemically irreversible processes at around 1.15 and
1.23 V, designated I and II, respectively, followed by process III at Epa 1.37 V, for
which a small return wave can be detected. These processes could in principle
correspond to
a
possible sequence of one-electron oxidations [(η6-p-
cymene)RuCl2(meapy)]0/+/2+/3+, but such closely spaced waves would be quite unusual
for an isolated metal complex due to the effects of charge buildup.
18

ACCEPTED MANUSCRIPT
Fig. 3 Cyclic voltammograms of 1 – 4 at 2.0·10-3 mol·L-1 in TBAP/ACN 0.1 mol·L-1 vs.
Ag/AgCl; obtained at 100 mV·s-1.
Further insight into these oxidation processes is provided by the computed
electronic structure of the 1+ state. After removal of the first electron from a Ru(d)-
Cl(p-π) orbital (the HOMO at left in Figure 2), significant orbital re-organization
occurs. The result is that the HOMO of the doublet state ion derives from the filled
aminopyridine ligand π3 level (which is HOMO-4 of the neutral state.) The unpaired
electron retains Ru(d) character at lower energy (it is the α spin orbital that has a very
similar topology to the β LUMO shown on the diagram; the energy evolution is shown
in dashed red lines on Figure 2). Because the aminopyridine π orbitals are non-
interacting with the metal, they do not experience the same amount of energy-lowering
in the cationic state as the metal orbitals do. This suggests the possibility that second
(and third) electrons in oxidation processes II (and III) come from the ligand and the
metal remains Ru(III) in all these oxidized states. In short, aminopyridine ligands
appear to function as redox-non-innocent ligands in the oxidized cationic states [54].
19

ACCEPTED MANUSCRIPT
Table 2 Potential data for complexes 1 – 4.
I
II
III
III
III
IIIc
IV
V
Complex
Epa
/V
Epa
Epa
Epc
Em
/V
Ipc/Ipa
Epc
/V
Epc
/V
Ewindow
/Vd
∆Ep
III /Vb
/V
/V
/V
1.37
1.15
1.22
1.16
1.41
1.52
1.43
1.47
1.37
1.44
1.36
0.09
0.09
0.09
0.06
0.11
0.20
0.15
0.04
-0.91 -1.22
-0.98 -1.30
-0.92 -1.22
-1.04 -1.32
2.28
2.13
2.14
2.20
1
2
3
4
1.23
1.32
1.42
1.49
1.39
1.33
1.40
1.33
c
a E1/2 = (Epa + Epc)/2; b Measured by difference between Epa and Epc; Evaluated as by ref [55] (to 2 V·s-1); d
IV
Measured by difference between EpaI and Epc
.
In the cathodic region, two irreversible reduction peaks are observed (processes
IV
V
IV and V, Epc at -0.98 and Epc at -1.30 V, respectively). These processes can be
confidently attributed to sequential occupation of the LUMO of the neutral complex
(Figure 2). Although formally this involves RuII/RuI reduction, this orbital also has
significant Ru-Cl σ* character. In this regard, it is interesting to observe that DFT
calculations optimize to geometries where one of the Cl− ions leaves the metal and
attaches remotely to the NH moiety via H-bonding (see Figure S10 in the SI). Ligand
dissociation could thus be responsible for the (chemical) irreversibility of these
processes.
The CVs of all four complexes were also recorded upon scanning from +1.0 to
+1.9 V and from -0.6 to -1.7 V over v = 50 – 2000 mV·s-1 (see Figures S13-S16 in the
SI). In all cases, the linearity of Ip vs. v1/2 plots demonstrates that mass transport of these
compounds to the electrode surface is diffusion-controlled. The voltammetric features
(Ipa/Ipc less than unity and ∆Epeak values about 90 mV) show that oxidation-reduction of
these compounds is chemically and electrochemically almost reversible in fast scans
and irreversible at lower scan rates.
3.7. Electronic absorption spectroscopy
Electronic absorption spectra of complexes 1 – 4 were acquired in the
concentration 10-4 mol·L-1, in different solvents at RT, such as: acetonitrile, water and
PBS buffer (Figure 4). The spectra in all these solvents showed similar broad low-
energy bands with maxima at 389-416 nm (ɛ = 2100-800 L·mol-1·cm-1), which can be
attributed to LUMO←HOMO transitions. TD-DFT calculations, carried out in CH3CN
and water solvent models, indicate that several transitions involving the cluster of
20

ACCEPTED MANUSCRIPT
highest filled orbitals probably contribute to these bands which thus have significant
d←d character (albeit with covalent contributions from Cl and pyridine N σ orbitals).
In addition, there are two sets of higher-energy bands. The first, with maxima
close to 300 nm for 1 – 3, are noticeably absent in the spectra of 4. These bands may
involve excitation from HOMO-4 to the LUMOs or from the higher filled orbitals to
LUMO+2. Probably the 2-aminopyridine π orbitals are involved and are thus either
LMCT or MLCT bands of modest intensity. Very intense bands at ~250 nm could
include lower-lying metal electron excitation to high virtual orbitals; furthermore, p-
cymene and pyridine π*←π transitions. The TD-DFT calculations indicate that most
bands have multiple transition contributions with varying oscillator strengths.
A noticeable feature (see the insets in Figure 4) is that the LUMO←HOMO
bands are blue-shifted by 12–26 nm in the aqueous spectra. There is a strong
expectation that these complexes will undergo hydrolysis either partly or completely in
water (Scheme 2) [18,56-58]. Our conductivity data (Section 3.2) suggest that the
hydrolysis is relatively slow and thus likely to stop at a single halide replacement.
Replacement of one chloride ligand by water is consistent with lowering the highest
occupied FMOs due to weaker π* character, and thus with a blue-shift. Since the
hydrolyzed complexes will tend to be cationic (especially in the enhanced acidic
environment of cancer cells), these processes have important implications for the in
vitro biological test results (see below).
Scheme 2 Some of the equilibria connected to chloride hydrolysis in aqueous phases.
21

ACCEPTED MANUSCRIPT
Fig. 4 UV-vis spectra of the series of complexes [(η6-p-cymene)RuCl2L] in acetonitrile,
water and PBS buffer. In each case, the main window covers 200-600 nm, with an
expanded inset from 300-600 nm. (A) Complex 1. (B) Complex 2. (C) Complex 3. (D)
Complex 4. Tabulated data provided in Table S4.
3.8. DNA interaction studies
Among organometallic ruthenium complexes, a range of compounds exhibit
potent anticancer activity. Many cytotoxic agents were proven to have DNA as the
cellular target. These molecules elicit a range of cellular responses which implies
different mechanisms of action [5,7]. The interaction of drug molecules with DNA can be
categorized using simple limiting models as shown schematically in Figure 5 as non-
covalent (intercalation, groove binding, electrostatic attraction) or covalent
(condensation or hydrogen-bonding with nitrogen bases or condensation with phosphate
of the DNA backbone).
Fig. 5 Cartoons showing common modes of interaction between drug molecules and
double-stranded DNA. Adapted from Rev. Virtual Quim., 2015, 7 (6), 1998-2016, with
the permission of the Brazilian Chemical Society.
22

ACCEPTED MANUSCRIPT
3.8.1. UV–vis spectroscopy
UV–vis absorption measurements have been successfully used to study the
mode and magnitude of interaction of complexes 1 – 4 with CT-DNA. DNA–complex
interactions are evidenced by this technique through changes in absorbance intensity
and position of the absorption band. When complexes 1 – 4 were titrated with CT-DNA,
hyperchromism (i.e. increased intensity) was observed, along with a small red-shift of
2-4ꢀnm in the presence of complexes 2 and 3 (Figure 6). The binding strengths of 1 – 4
were quantified from the values of intrinsic binding constant Kb, determined using the
Benesi-Hildebrand equation (Table 3) [43]. Whereas hypochromism (i.e. decreased
intensity) is indicative of DNA intercalation (due to contraction of the helix and
conformational changes caused by changes in the π-π stacking interactions), [59]
hyperchromism is attributed to electrostatic interaction between complexes and the
negatively charged phosphate backbone at the periphery of the double helix CT-DNA or
to secondary damage of the DNA double helix structure [60-63]. The presence of
complexes 1 to 4 resulted in hyperchromism in DNA absorption spectra suggesting non-
intercalative binding between DNA and the complexes. This is corroborated by the size
of the intrinsic binding constants Kb being in the micromolar range [64].
Fig. 6 Electronic spectra of the complexes in the absence and after successive additions
of CT-DNA. (A) complex 1; (B) complex 2; (C) complex 3; (D) complex 4.
23

ACCEPTED MANUSCRIPT
Table 3 Intrinsic binding constants (Kb) for interaction between CT-DNA and
complexes 1 – 4.
Wavelength (nm)
Kb (M-1)
2.24 x105
1.40 x105
7.34 x104
3.68 x104
Complexes
227
237
243
257
1
2
3
4
3.8.2. Circular dichroism spectroscopy
The circular dichroism (CD) technique is responsive to changes in the chiral
structure of DNA and is used to study variations in DNA conformation upon its
interaction with small molecules [65]. B-form calf thymus DNA exhibits a negative
band at 245 nm caused by helicity and a positive band with maximum at 275 nm
caused by base stacking [66,67].
Distinct modes of interaction can be distinguished by changes in the spectra.
Intercalation clearly enhances the signal intensity of both the base stacking and
helicity bands, while groove binding and electrostatic interactions cause slight
perturbations on positive and negative bands. In addition, changes just in the intensity,
and not shape, of the observed CD results suggest a single binding mode. Changes in
the shape of CD signals may indicate multiple ligand-DNA binding modes, changes in
DNA conformation or ligand–ligand interactions [68].
To verify whether binding of the complex causes any conformational change of
the DNA double helix, CD spectra of CT-DNA were recorded at different complex/CT-
DNA ratios (Figure 7, black lines). The addition of complexes 1, 2 and 4 to CT-DNA
slightly increased the intensity of the positive peak and decreased the intensity of the
negative peak (Figure 7). These changes in the CD spectra in the presence of the
complexes indicate that complexes 1, 2 and 4 interact with CT-DNA and stabilize the
right-handed B-form of CT-DNA structure. These alterations in spectra are common in
groove binding and electrostatic interactions [65].
24

ACCEPTED MANUSCRIPT
Fig. 7 CD spectra for CT-DNA (100 and 200 µmol·L-1) in PBS buffer (pH 7.6), with
addition of 200 µmol·L-1 of the complexes. (A) complex 1; (B) complex 2; (C) complex
3 and (D) complex 4.
3.8.3. Ethidium bromide competition.
To confirm a non-intercalative binding mode, competition experiments with
EtBr were performed. Classical intercalators displace EtBr from DNA bases, thereby
decreasing its fluorescence emission [69]. Addition of complexes 1 – 4 to EtBr-DNA
solutions does not alter emission intensity of EtBr by more than the dilution effect,
which, together with the spectroscopic titration and CD results, indicates that these
complexes are not DNA intercalators.
25

ACCEPTED MANUSCRIPT
Fig. 8 Emission spectra of the EtBr/CT-DNA system, with successive additions of the
complexes: (A) complex 1; (B) complex 2; (C) complex 3 and (D) complex 4.
Confirmed dilution effect after successive additions to PBS buffer.
3.8.4. Viscosity Measurements
Hydrodynamic measurements are considered as unequivocal tests of DNA
binding models in solution, clarifying the interaction mode of a compound with the
nucleic acid. An interaction between the DNA double helix and a small molecule may
cause length changes in DNA and as a result viscosity changes [44]. The values of
relative specific viscosity (η/η0)1/3 were plotted against [DNA]/[complex] (Figure 9). In
this study, it was observed that increasing concentrations of complexes 1 – 4 do not
significantly alter the DNA viscosity. Thus, it is possible to infer that these complexes
are not covalent binders, and neither partial nor classical DNA intercalators.
Fig. 9 Viscosity graph plotting the increase of the concentration of the complexes vs.
the relative viscosity of the CT-DNA (20 µmol·L-1) at 25 °C.
26

ACCEPTED MANUSCRIPT
3.9. Cytotoxicity assays
The cytotoxic effects were examined for the [(η6
‐p-cymene)RuCl2L] series
1
complexes as well as for the ([(η6–p-cymene)RuCl2(DMSO)] obtained in situ, since H
NMR (see Figure S18) tests performed for the complexes in the presence of DMSO
demonstrate that there is a pyridine ligand lability with coordination by the DMSO
molecule. The complexes thus formed by replacing monodentate pyridine ligands with
DMSO were confirmed to also be neutral by conductivity tests (see Table S7).
Coordination of DMSO in ruthenium pyridine complexes has been previously reported
in the literature [18,70] and by comparison of data we can indicate that the DMSO is
coordinated to the metal through sulfur [71,72].
The cytotoxic tests for [(η6
‐p-cymene)RuCl2L] complexes in DMSO were carried
out against HeLa, Hep-2, B16F10 and Vero cells line (Table 4, Figure S19), while tests
carried in phosphate buffer were evaluated only against Hep-2 and B16F10 (Table 4,
Figure S20). Emphasis was placed on testing against the resistive HeLa line. The results
obtained using an MTT assay showed that only 2 with DMSO achieved an IC50 against
the HeLa tumor line after a 24 h with IC50 = 346 µmol·L-1. The absence of toxicity in
aqueous solution suggests that the toxicity is due to the liberated meapy ligand. For the
other complexes of the series with or without DMSO as well as for other tumor lines,
HEp-2 and B16F10, it was not possible to determine IC50 values in the range of
concentrations investigated.
27

ACCEPTED MANUSCRIPT
Table 4 Cytotoxicity results for 1–4 on HeLa, HEp-2, B16F10 and Vero cell lines,
after 24 h incubation performed in DMSO solution or in aqueous buffer.
IC50 (µmol·L-1)
HeLa HEp-2 B16F10 Vero
Complex
1 – DMSO a
>500 >500
346±3 >500
>500 >500
—
—
2 – DMSO a
—
—
—
—
—
—
—
3 – DMSO a
>500
>500
—
—
—
—
4 – DMSO a
—
—
1 b
2 b
3 b
4 b
>630
>600
>600
>650
—
>600
—
—
—
a From DMSO stock solution; displacement of L by DMSO-S is assumed. b Test
performed using an aqueous buffer stock solution.
The low activity of [(η6
‐p-cymene)RuCl2L] series 1 – 4 against cancer cell lines
is similar to results reported on a range of other type A complexes with simple
substituted pyridines [10,12,14,16,18,28,73,74]. By contrast, the substituted
aminopyridines reported by Richter et al. [27], bearing 4-carboxylatophenyl substituents
at the amine, were shown to be more active against 8500C, MCF-7, SW-480 and 518A2
cancer cell lines, although only marginally more than the direct use of the corresponding
substituted amino-pyridines in control tests [75].
3.10. BSA Interactions
An important feature of biologically active compounds is their binding to proteins.
The ability to interact with proteins affects the activity of a molecule in biological systems,
so protein-binding studies are carried out to reveal the potential of new drug molecules.
Bovine serum albumin protein is frequently used in these protein-binding studies because of
its structural homology with human serum albumin (HSA). In order to characterize binding
of complexes 1 – 4 to BSA, we carried out circular dichroism studies. The protein
conformation was not significantly altered in the presence of complexes 1 – 4 (Figure S17 in
the SI). We further investigated the thermal stability of BSA in the presence of complexes 1
– 4. When bound to a protein, small molecules tend to enhance the thermal stability,
resulting in increased melting temperatures (Tm) [76]. Thermal stability curves for BSA-
complexes and BSA alone were plotted from 15 to 95ºC, as shown in Figure 10. The melting
temperature (Tm) of BSA was estimated in 72.1ºC and in the presence of complexes 1 – 4,
temperature varied from 67.2 to 75.4ºC, which indicates no significant increase in BSA
stability. The addition of complexes 1 – 4 did not increase the melting temperature (Tm) of
28