Synthesis, characterization, HSA/DNA interactions and antitumor activity of new [Ru(η6-p-cymene)Cl2(L)] complexes

Jou r n a l o f I n o r g an i c B i o che m i s tr y 2 1 3 (2 020 )1 11256

Contents lists available at ScienceDirect

Journal of Inorganic Biochemistry

journal homepage: www.elsevier.com/locate/jinorgbio

Synthesis, characterization, HSA/DNA interactions and antitumor activity of

new [Ru(η⁶-p-cymene)Cl₂(L)] complexes

Maja B. Đukić^a, Marija S. Jeremić^a, Ignjat P. Filipović^a, Olivera R. Klisurić^b, Vesna V. Kojić^c,

⁎

Dimitar S. Jakimov^c, Ratomir M. Jelić^d, Valentina Onnis^e, Zoran D. Matović^a,

^aUniversity of Kragujevac, Faculty of Science, Department of Chemistry, Radoja Domanovića 12, 34000 Kragujevac, Serbia

^bUniversity of Novi Sad, Faculty of Sciences, Department of Physics, Trg Dositeja Obradovića 4, 21000 Novi Sad, Serbia

^cOncology Institute of Vojvodina, Faculty of Medicine, University of Novi Sad, Put Doktora Goldmana 4, 21204 Sremska Kamenica, Serbia

^dUniversity of Kragujevac, Faculty of Medical Sciences, Department of Pharmacy, Svetozara Markovića 69, 34000 Kragujevac, Serbia

^eDepartment of Life and Environmental Sciences, Unit of Pharmaceutical, Pharmacological and Nutraceutical Sciences, University of Cagliari, University Campus, S.P. n° 8,

Km 0.700, I-09042 Monserrato (CA), Italy

A R T I C L E I N F O

A B S T R A C T

Keywords:

Three new ruthenium(II) complexes were synthesized from diﬀerent substituted isothiazole ligands 5-(methy-

Ruthenium(II) complexes

HSA interactions

DNA interactions

Molecular docking

Antiproliferation

lamino)-3-pyrrolidine-1-ylisothiazole-4-carbonitrile

(1),

5-(methylamino)-3-(4-methylpiperazine-1-yl)iso-

thiazole-4-carbonitrile (2) and 5-(methylamino)-3-morpholine-4-ylisothiazole-4-carbonitrile (3): [Ru(η⁶-p-

cymene)Cl₂(L1)]·H₂O (4), [Ru(η⁶-p-cymene)Cl₂(L2)] (5) and [Ru(η⁶-p-cymene)Cl₂(L3)] (6). All complexes were

characterized by IR, UV–Vis, NMR spectroscopy, and elemental analysis. The molecular structures of all ligands

and complexes 4 and 6 were determined by an X-ray. The results of the interactions of CT-DNA (calf thymus

deoxyribonucleic acid) and HSA (human serum albumin) with ruthenium (II) complexes reveal that complex 4

binds well to CT-DNA and HSA. Kinetic and thermodynamic parameters for the reaction between complex and

HSA conﬁrmed the associative mode of interaction. The results of Quantum mechanics (QM) modelling and

docking experiments toward DNA dodecamer and HSA support the strongest binding of the complex 4 to DNA

major groove, as well as its binding to IIa domain of HSA with the lowest ΔG energy, which agrees with the

solution studies. The modiﬁed GOLD docking results are indicative for Ru(p-cymene)LCl··(HSA··GLU292) binding

and GOLD/MOPAC(QM) docking/modelling of DNA/Ligand (Ru(II)-N(7)dG7) covalent binding. The cytotoxic

activity of compounds was evaluated by MTT (3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bro-

mide) assay. Neither of the tested compounds shows activity against a healthy MRC-5 cell line while the MCF-7

cell line is the most sensitive to all. Compounds 3, 4 and 5 were about two times more active than cisplatin, while

the antiproliferative activity of 6 was almost the same as with cisplatin. Flow cytometry analysis showed the

apoptotic death of the cells with a cell cycle arrest in the subG1 phase.

1. Introduction

speciﬁc activity toward metastatic cancer cells. The reader may ﬁnd

more relevant intro about the anticancer activity of half-sandwich ru-

A signiﬁcant contribution to the ﬁght against malignancy was

achieved using transition metal complexes. Probably the most sig-

niﬁcant among them is cisplatin, which has been used with more or less

success for more than half a century [1]. Although eﬃcient against the

number of cancers, cisplatin also causes severe side-eﬀects, which leads

to e.g. peripheral neuropathy, hair loss, and myelotoxicity in patients

[2]. Furthermore, for decades, chemists have been persistently working

on the synthesis of new potentially bio-active metal complexes, chan-

ging either the structure of the ligand or the nature of the metal ion (Ru,

Os, Fe, Cu …). Ruthenium is very often seen in complexes with a

thenium(II) complexes in many reviews and papers in which the

synthesis and the mechanisms of its action were reported [3–6]. Also,

Ru(III) complexes such as NAMI-A ((ImH)[trans-RuCl₄(dmso-S)(Im)],

Im = imidazole) and KP1019 ((IndH)[trans-RuCl₄(Ind)₂], Ind = in-

dazole) have undergone phase I of a clinical trial [7] (Fig. 1). Never-

theless, half-sandwich piano stool Ru(II) complexes have attracted

greater interest in recent years for their antimetastatic and anticancer

properties.

RAPTA-C

([Ru(η⁶-p-cymene)(PTA)]Cl₂],

where

PTA = 1,3,5-triaza-7-phosphaadamantane) [8] is anti-metastatic and

anti-angiogenic agent able to form adducts with histone proteins [9].

^⁎Corresponding author.

E-mail address: zmatovic@kg.ac.rs (Z.D. Matović).

https://doi.org/10.1016/j.jinorgbio.2020.111256

Received 23 July 2020; Received in revised form 8 September 2020; Accepted 10 September 2020

A v a i l a b l e o n l i n e 1 5 S ep t e m b e r 2 020

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Fig. 1. Ruthenium(III) and ruthenium(II) arene complexes currently in clinical trials.

Encouraged by these results, many research groups have based their

research on the synthesis of ruthenium(II) complexes, which will show

greater activity against malignant cells and less toxicity to healthy cells.

Thus, Romerosa et al. synthesized a large number of ruthenium com-

plexes with diﬀerent ligands: [RuCp(PPh₃)₂(HdmoPTA)](OSO₂CF₃)₂

(HdmoPTA = 3,7-H-3,7-dimethyl-1,3,7-triaza-5 phosphabicyclo[3.3.1]

biologically active substances, isothiazoles can serve as ligands for

transition metal complexes [31,35–38].

Recently, we published the results of our research on syntheses and

biology of [Ru(η⁶-p-cymene)(R-imidazole)Xn]^2-ncomplexes [39]. Fol-

lowing our work on heterocycles and half-sandwich ruthenium(II)

complexes, we describe here the synthesis, characterization and crystal

structures of some isothiazole ligands and corresponding ruthenium(II)

complexes. Also, the results of their cytotoxicity against four cancer cell

lines and one non-cancer cell in vitro have been reported. Additionally,

we examined the in vitro interaction of the complexes with calf thymus

deoxyribonucleic acid (CT-DNA), as well as the in vitro aﬃnity of the

complexes for human serum albumin by UV–Vis spectroscopy and

ﬂuorescence emission spectroscopy as well as the in vitro interaction of

the complexes with CT-DNA. Furthermore, quantum mechanics mod-

elling and docking simulations toward DNA dodecamer and HSA have

been done. AutoDock and Vina have been used with new parameters for

ruthenium(II) as well as modiﬁed GOLD software in order to establish

binding energies of the Ligand—HSA(DNA) system.

nonane)

[10],

[RuCp(PPh₃)₂-μ-dmoPTA-1κP:2κ²N,N′-CoCl₂]

(OTf)·0.25H₂O (OTf = triﬂuoromethanesulfonate, i.e. CF₃SO₃) [11],

bimetallic RueZn complex [RuCp(PPh₃)₂-μ-dmoPTA-1κP:2κ²N,N′-

ZnCl₂](CF₃SO₃) [12] etc…. Such complexes are generally insoluble or

sparingly soluble in water. Similar antitumor active ruthenium(II)

complexes were summarized in review articles [13]. On the other hand,

data on the antitumor activity of metal complexes with isothiazoles are

very scarce. The mechanism of action of these ruthenium(III) com-

pounds involves at least their interaction with serum proteins such as

albumin and transferrin that endows them with tumor seeking prop-

erties [14]; second, ruthenium(III) complexes appear to be activated

through intracellular reduction to allow generation of toxic ruthenium

(II) species [15,16]; and third, at the tumor site reactions (binding) with

proteins are preferred to DNA binding, which contrasts with the be-

havior of platinum(II) complexes such as cisplatin [17–19]. Reaction

rates between Ru-p-cymene and human serum albumin (HSA) co-

ordinating ligands are strongly aﬀected by the type of the donor atoms.

Complex formation with (O,O) or (O,O,O) donor ligands is a fast pro-

cess at physiological pH, while (N) and (N, N, N) donor ligands chelate

the Ru-p-cymene rather slowly (hours to days) [20–23]. Therefore

metalation of important serum proteins by ruthenium(II) appears to be

probable and should be caused by the nature of the ruthenium drugs

[24].

2. Experimental

2.1. Materials and physical measurements

[Ru-(η⁶-p-cymene)Cl₂]₂, ethyl acetate, toluene, dichloromethane,

methanol, NaOH, HCl, NaCl, Tris-HCl, highly polymerized CT-DNA and

ethidium bromide were purchased from Sigma-Aldrich, used as re-

ceived.

The stock DNA solution was prepared by diluting CT-DNA with a

buﬀer solution (10 mM tris(hydroxymethyl)–aminomethane (Tris-HCl)

and 150 mM NaCl to pH 7.4), after which it was stirred at 4 °C for

3 days, and kept at 4 °C for no longer than a week. The concentration of

CT-DNA in stock solution was determined by UV absorption at 260 nm

using a molar absorption coeﬃcient (ε₂₆₀= 6600 M⁻¹cm⁻¹) [40]. The

purity of the DNA was checked by monitoring the ratio of the absor-

bance at 260 nm to that at 280 nm. The solution gave a ratio of > 1.8 at

Multidonor heterocyclic ligands demonstrate diﬀerent and diverse

capabilities for the coordination of various metal ions. The electron-rich

polyfunctional derivatives of thiazole and isothiazole belong to this

type of ligand, and their bioactive coordinating compounds have at-

tracted signiﬁcant attention in medicine and pharmacy [25–27]. Basi-

cally, the thiazole heterocyclic ring can be found in two isomeric forms,

1,3-thiazole (known as thiazole) and 1,2-thiazole (known as iso-

thiazole). They showed a wide range of useful properties which led

researchers to study the synthesis and chemical transformations of its

derivatives [28–30]. Some of them have shown eﬀectiveness in the

treatment of Alzheimer's disease, where their eﬀects are based on in-

hibition of serine protease, on anti-inﬂammation and anticonvulsion

[31]. Also, these types of compounds show great antiviral activity and

inﬂammatory properties [32–34]. Besides their role in the synthesis of

A

₂₆₀/A₂₈₀, which indicates that DNA was adequately free from protein

[41]. Due to the limited solubility of the ligands and metal complexes in

the water, stock solutions were prepared by dissolving the compound in

DMSO. By diluting the stock solution with buﬀer (10 mM Tris-HCl and

150 mM NaCl at pH 7.4), working solutions were obtained in which the

ﬁnal concentration of DMSO was at most 8%. Doubly distilled water

was used for the preparation of all solutions.

UV–Vis spectra were recorded on

a double beam UV–Vis

2

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

spectrophotometer model Cary 300 (Agilent Technologies, Santa Clara,

USA) with 1.0 cm quartz cells. Emission measurements were carried out

using an RF-1501 PC spectroﬂuorometer (Shimadzu, Japan). The ex-

citation wavelength was ﬁxed, and the emission range was adjusted

before measurements, with the excitation and emission slit widths set at

10 nm. All the measurements were conducted in buﬀer containing Tris-

HCl (10 mM) and NaCl (150 mM) and adjusted to pH 7.4 with hydro-

chloric acid, at 25 °C.

method described elsewhere [44] with little modiﬁcation. A mixture of

[Ru-(η⁶-p-cymene)Cl₂]₂(0.0998 g; 0.163 mmol) and ligand 2 (0.2373 g;

1 mmol; 6.1 equiv.) in dichloromethane (15 mL) was reﬂuxed for 3 h

(39 °C), cooled and ﬁltered. The solution was evaporated to a volume of

~5 mL at room temperature. After a few days, the precipitate was ﬁl-

tered, yielding a green powdery residue. This complex was washed with

diethyl ether and air-dried (0.192 g, 35.52%). Melting point = 137 °C.

Anal. Cal. for C₂₀H₂₉Cl₂N₅RuS (Mw = 543.52) C: 44.20, H: 5.38, N:

12.89; found: C: 44.09, H: 5.33, N: 12.65. ¹H NMR (200 MHz, CDCl₃):

δ(ppm) 1.36 (d, 6H, CH₃), 2.21 (s, 3H, 4-CH₃), 2.33 (s, 3H, 1^V-CH₃),

2.51 (t, 3′-H, 5′-H), 2.81 (sept, 1H, 1-CH(CH₃)₂), 2.99 (s, 3H, 1″-CH₃),

3.58 (t, 2′-H, 6′-H), 5.40 (d, 2H, 2-H, 6-H), 5.62 (d, 2H, 3-H, 5-H) (Fig.

S5, ESI). ¹³C{¹H} NMR (50 MHz, CDCl₃): δ(ppm) 19.15 (-CH₃), 22.34

(1-CH(CH₃)₂), 31.36 (C-1″), 33.22 (1-CH(CH₃)₂), 46.10 (Ce1^V), 47.50

(C-2′, C-6′), 54.61 (C-3′, C-5′), 87.49 (C-1‴), 97.94 (C-1^IV), 115.35 (C-2,

C-6), 117.20 (C-3, C-5), 135.50 (C-4), 136.38 (C-1), 159.65 (C-5‴),

164.34 (C-2‴) (Fig. S6, ESI). IR (KBr, pellet): ν (cm⁻¹) 3266 (ν = CH),

2961, 2935 (νCH), 2239, 2200 (νC ≡ N), 1565 (νC=C) (Fig. S7, ESI).

UV–Vis (CH₃OH, C = 2 × 10⁻³M): λ_max/nm(ε/L mol⁻¹cm⁻¹): 390

Elemental microanalyses for C, H, N were performed at the

Microanalytical laboratory, Faculty of Chemistry, University of

Belgrade, Serbia. IR spectra in the 400–4000 cm⁻¹region were re-

corded on a Perkin-Elmer FT-IR spectrophotometer Spectrum One,

using the KBr pellets technique. ¹H and ¹³C{¹H} NMR spectra were

recorded on a Varian Gemini 2000 spectrometer (200 MHz). Chemical

shifts are expressed as δ values (ppm) relative to Tetramethylsilane

(TMS) as an internal standard. Electronic absorption spectra were re-

corded on a double beam UV–Vis spectrophotometer model Cary 300

(Agilent Technologies, Santa Clara, USA) with 1.0 cm quartz cells.

Kinetic measurements were performed on UV–Vis Perkin-Elmer

Lambda 35 double beam spectrophotometer with water thermostated

1.0 cm path-length quartz cuvettes (3.0 mL). Melting points were

measured by the Stuart melting device with accuracy 1 °C. Molar

conductivities were measured at room temperature on a digital con-

ductivity-meter Crison Multimeter MM 41. The concentration of the

solutions of complexes 4–6 in DMSO used for conductivity measure-

ments was 1 × 10⁻³M.

(sh) (406) (Fig. S4, ESI). Λ_M(DMSO): 0.01 Ω⁻¹cm²mol⁻¹

2.2.4. Synthesis of [Ru(η⁶-p-cymene)Cl₂(L3)] (6)

.

The [Ru(η⁶-p-cymene)Cl₂(L3)] was synthesized by following the

method described elsewhere [44]. A mixture of [Ru-(η⁶-p-cymene)Cl₂]₂

(0.0998 g; 0.163 mmol) and ligand 3 (0.2242 g; 1 mmol; 6.1 equiv.) in

toluene (15 mL) was reﬂuxed for 3 h (110 °C) and cooled. The orange

crystals were obtained from the toluene (0.255 g, 48.06%). Melting

point = 216 °C. Anal. Cal. for C₁₉H₂₆Cl₂N₄ORuS (Mw = 530.47) C:

43.02, H: 4.94, N: 10.56; found: C: 43.31, H: 4.92, N: 10.46. ¹H NMR

(200 MHz, CDCl₃): δ(ppm) 1.36 (d, 6H, CH₃), 2.16 (s, 3H, 1″-CH₃), 2.32

(s, 3H, 4-CH₃), 2.92 (sept, 1H, 1-CH(CH₃)₂), 3.50 (t, 2′-H, 6′-H), 3.78 (t,

5′-H, 3′-H), 5.41 (d, 2H, 2-H, 6-H), 5.62 (d, 2H, 3-H, 5-H) (Fig. S8, ESI).

¹³C{¹H} NMR (50 MHz, CDCl₃): δ(ppm) 19.09 (-CH₃), 22.42 (1-CH

(CH₃)₂), 31.38 (C-1″), 33.23 (1-CH(CH₃)₂), 48.00 (C-2′, C-6′), 66.42 (C-

3′, C-5′), 80.59 (C-1‴), 83.45 (C-1^IV), 96.75 (C-2, C-6), 98.97 (C-3, C-5),

101.26 (C-4), 102.84 (C-1), 128.16 (C-5‴), 128.97 (C-2‴) (Fig. S9, ESI).

IR (KBr, pellet): ν (cm⁻¹) 3432 (ν = CH), 2969, 2923 (νCH), 2228

(νC ≡ N), 1567 (νC=C) (Fig. S10, ESI). UV–Vis (CH₃OH,

C = 2 × 10⁻³M): λ_max/nm(ε/L mol⁻¹cm⁻¹): 337 (1118) and 430

2.2. Synthesis of compounds

2.2.1. Synthesis of ligands (1, 2, 3)

The ligands 5-(methylamino)-3-pyrrolidine-1-ylisothiazole-4‑carbo-

nitrile (1), 5-(methylamino)-3-(4-methylpiperazine-1-yl)isothiazole-

4‑carbonitrile (2) and 5-(methylamino)-3-morpholine-4-ylisothiazole-

4‑carbonitrile (3) were prepared by the previously reported procedures

[42,43]. Herein, we have provided the synthesis of ligand 1 as an ex-

ample: the ice-cooled solution of 3-amino-2-(methylthiocarbamoyl)-3-

pyrrolidino-2-propenenitrile (1.05 g, 5 mmol) in anhydrous CHCl₃

(50 mL) was mixed with Br₂(0.8 g, 5 mmol) and the resulting mixture

was stirred for 1 h. The mixture was washed with a solution of NaOH

and water. The ligand 1 was obtained by recrystallization from cyclo-

hexane.

(654) (Fig. S4, ESI). Λ_M(DMSO): 0.01 Ω⁻¹cm²mol⁻¹

.

2.3. Crystal structure determination

2.2.2. Synthesis of [Ru(η⁶-p-cymene)Cl₂(L1)]·H₂O (4)

The [Ru(η⁶-p-cymene)Cl₂(L1)]·H₂O was synthesized by following

the method described elsewhere [44]. A mixture of [Ru-(η⁶-p-cyme-

ne)Cl₂]₂(0.0998 g; 0.163 mmol) and ligand 1 (0.2082 g; 1 mmol; 6.1

equiv.) in toluene (15 mL) was reﬂuxed for 3 h (110 °C), cooled and

ﬁltered. The solution was evaporated to a volume of ~5 mL at room

temperature. Crystals suitable for X-ray crystal analysis were obtained

by the addition of ethyl acetate (0.228 g, 42.84%). Melting

point = 174 °C. Anal. Cal. for C₁₉H₂₈Cl₂N₄ORuS (Mw = 532.48) C:

42.86, H: 5.30, N: 10.52; found: C: 43.03, H: 5.51, N: 10.78. ¹H NMR

(200 MHz, CDCl₃): δ(ppm) 1.35 (d, 6H, CH₃), 1.82 (s, 3H, 4-CH₃), 1.95

(qui, 3′-H, 4′-H), 2.31 (s, 3H, 1″-CH₃), 2.92 (sept, 1H, 1-CH(CH₃)₂),

3.58 (t, 5′-H, 2′-H), 5.38 (d, 2H, 2-H, 6-H), 5.60 (d, 2H, 3-H, 5-H) (Fig.

S1, ESI). ¹³C{¹H} NMR (50 MHz, CDCl₃): δ(ppm) 19.12 (-CH₃), 22.42

(1-CH(CH₃)₂), 25.56 (C-3′, C-4′), 31.38 (C-1″), 33.14 (1-CH(CH₃)₂),

48.25 (C-5′, C-2′), 80.59 (C-1‴), 81.50 (C-1^IV), 83.35 (C-2‴), 96.82 (C-

2, C-6), 98.86 (C-3, C-5), 101.26 (C-4), 102.85 (C-1) (Fig. S2, ESI). IR

(KBr, pellet): ν (cm⁻¹) 3407 (ν = CH), 2959, 2925 (νCH), 2235

(νC ≡ N), 1580, 1524 (νC=C) (Fig. S3, ESI). UV–Vis (CH₃OH,

C = 2 × 10⁻³M): λ_max/nm(ε/L mol⁻¹cm⁻¹): 342 (sh) (1000), 428

Single crystals of complexes 4, 6 and ligands 1, 2 and 3 were

mounted on a glass ﬁber and crystallographic data were collected using

the Rigaku (Oxford Diﬀraction) Gemini S diﬀractometer with a CCD

area detector (λ_MoKα= 0.71073 Å, monochromator: graphite) at 293 K.

CrysAlisPro and CrysAlis RED software packages [45] were used for

data collection and data integration. Analysis of the integrated data did

not reveal any decay. Collected data were corrected for absorption ef-

fects by using a Numerical absorption correction based on Gaussian

integration over a multifaceted crystal model [46] for compound 4 and

analytical numeric absorption correction applying a multifaceted

crystal model [47] for compound 6, while collected data for ligands 1, 2

and 3 were corrected for absorption eﬀects by using a Multi-scan ab-

sorption correction [48]. Structure solution and reﬁnement were car-

ried out with the programs SHELXT and SHELXL-2014/6 respectively

[49]. All calculations were performed using PLATON [50] while

MERCURY [51] was employed for molecular graphics, all implemented

in the WINGX [52] system of programs. Full-matrix least-squares re-

2

ﬁnement was carried out by minimizing (F_o– F_c²). All nonhydrogen

atoms were reﬁned anisotropically. Hydrogen atoms in all crystal

structures attached to carbon atoms in methyl, methylene and methine

groups were placed in geometrically idealized positions and reﬁned as

riding on their parent atoms where C–H(CH₃) = 0.96 Å with U_iso

(H) = 1.5 U_eq(C), C–H(CH₂) = 0.97 Å with U_iso(H) = 1.2 U_eq(C) and

(564) (Fig. S4, ESI). Λ_M(DMSO): 0.01 Ω⁻¹cm²mol⁻¹

2.2.3. Synthesis of [Ru(η⁶-p-cymene)Cl₂(L2)] (5)

.

The [Ru(η⁶-p-cymene)Cl₂(L2)] was synthesized by following the

3

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

4

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

C–H(CH) = 0.98 Å with U_iso(H) = 1.2 U_eq(C). Positions of all other

hydrogen atoms were taken from ∆F map and reﬁned as riding, while

their U_isowere reﬁned. Atoms C18 form the terminal methyl group form

η⁶-p-cymene in complex 6 showed large displacement parameters. The

electron-density diﬀerence map revealed alternate sites for this C atom,

therefore, an eﬀort to model the disorder was made. Atom C18 was

found to be disordered over two sites A and B with site occupancy

factors of 0.60 (5) and 0.40 (5), respectively. After the reﬁnement, atom

C18B was still found to be disordered, but the electron-density diﬀer-

ence map did not reveal alternate sites for C18B, and non-Fourier peak

in the region exceeds 0.3 e A⁻³. Crystal data and experimental details

of the structures determination for ligands 1, 2 and 3 and corre-

sponding complexes 4 and 6 are listed in Table 1.

2.5. Computational chemistry

2.5.1. Quantum mechanics and molecular docking

Geometry optimizations were carried out with Gaussian 09 D.01

program [55], using hybrid functional of Truhlar and Zhao (M06) [56]

and in each case, the SDDall basis set was used [57]. Starting geome-

tries were taken either from experimental X-ray structures or were

modeled, starting with conformational search from which the best hits

were taken in order to pre-optimize them using the semi-empirical

methods (MOPAC2016 [58]/PM6-D3H4 [59] Hamiltonian). The sys-

tems were treated within restricted formalism. All calculations were

performed with toluene (ligands) or H₂O (complexes) as solvent by use

of the Integral Equation Formalism Polarizable Continuum Model

(IEFPCM) as implemented in Gaussian 09. Geometry optimizations

were conducted without symmetry constraints. The minimum energy

was achieved in all cases which were conﬁrmed by the frequency cal-

culations for each optimized structure. The best conformations along

with atom charges have been taken for molecular docking simulations.

2.4. Solution studies

2.4.1. HSA-binding experiments

The protein binding study was performed by tryptophan ﬂuores-

cence quenching experiments using HSA (2.0 × 10⁻⁶M) in a buﬀer

(10 mM Tris–HCl and 150 mM NaCl at pH 7.4). Fluorescence spectra

were recorded from 300 to 450 nm at an excitation wavelength of

295 nm. The ﬂuorescence spectra of the ligands and complexes in buﬀer

solutions were recorded under the same experimental conditions. The

ﬂuorescence quenching is described by the Stern–Volmer relation (Eq.

(3)) (see ESI), similarly as described below for CT-DNA binding studies.

The values of K and n were obtained from the intercept and slope of

the plots of log (F₀− F)/F versus log [Q] using Eq. (1) (see ESI).

Kinetics and mechanism of the interaction between complex 4 and

HSA were studied by conventional UV–Vis spectrophotometry re-

cording the change of the absorbance (in the range between 300 and

800 nm) with the time (Fig. S11, ESI). The reaction was initiated by

mixing the solutions in quartz cuvette and followed at three diﬀerent

temperatures (288, 298 and 308 K) several hours. According to the

observed kinetic traces, second-order rate constants and thermo-

dynamic parameters (ΔH^≠and ΔS^≠) were calculated using Microsoft

Excel 2007 and Origin Pro 8.

2.5.1.1. Docking. Molecular docking of ligands and corresponding

ruthenium(II) complexes was simulated to either three-dimensional X-

ray structure of human serum albumin (HSA), PDB (Protein Data Bank)

code 1HK1 [60] and 1O9X [61] or to the double-stranded (ds)

dodecamer

sequence

of

5′-d(CCTCTG*GTCTCC)-3′*5′-d(

GGAGACCAGAGG)-3′ the three-dimensional X-ray structures of

Lippard's DNA duplex (PDB code 1AIO) [62]. Docking processes were

carried out using AutoDock 4.2.6 [63 ] and AutoDock Vina [64]

software equipped with the graphical user interface (GUI) Auto-

DockTools (ADT 1.5.6rc3) [63]. Adapted GOLD [65] suite of

programs (Hermes and Gold), 5.6.2 release [66] were used for

docking as well.

Details of Autodock 4.2.6 docking procedures are described else-

where [39,67]. Since ruthenium is not included in the AutoDock

parameter ﬁle, extended parameter ﬁle including Ru(II) parameters was

used [68]. Docking assessment against AutoDock has been validated

across two X-ray structures: PDB codes 1HK1 and 1O9X as described in

our recent paper [39].

All GOLD calculations were performed with the GoldScore (GS)

scoring function. All modiﬁcations to the GOLD force-ﬁeld were made

according to Sciortino et al. [66]. The GS parameter ﬁle was modiﬁed

to include parameters of atom types not included in GOLD's database,

such as metals and possible coordinating atom groups and, in parti-

cular, sp²oxygens (e.g., those of carboxylate group), sp³oxygen of

water, and sp³negatively charged oxygens of deprotonated serine,

threonine, and tyrosine residues. The metal atom type (M.Ru) was built

with the values reported in the literature, the keto (O.pl3), the water

oxygen atoms (O.H₂O) and the sp³negatively charged oxygens were

ﬁxed considering the geometry of the electron pairs derived from the

valence shell electron pair repulsion (VSEPR) theory. Genetic algorithm

(GA) parameters were set with a number of GA runs equal to 50 and a

minimum of 100,000 operations. The rest of the parameters, including

pressure, a number of islands, niche size, crossover, mutation, and

migration were left to default. Docking assessment against GOLD in the

case of protein has been already validated by Sciortino et al. [66]. Since

modiﬁed GOLD forceﬁeld has not been validated against DNA by the

authors [66] we carried out a small portion of this process. For this

purpose, we took the DNA part of nucleosome structure (PDB code

4KGC [9]) and one of the ruthenium(II) complex RAED-C coordinated

to DNA. The results are more or less promising (Fig. S12, ESI) with a

best docked conformation close to X-ray one.

2.4.2. DNA-binding studies

Interactions of ligands 1–3 and complexes 4–6 with CT-DNA have

been studied with UV spectroscopy in order to investigate the possible

binding modes to CT-DNA and to calculate the binding constants to CT-

DNA, K_b. The binding constants, K_b, of the compounds with CT-DNA

have been determined using the UV spectra of the compound recorded

for a constant concentration (8.0 × 10⁻⁵M) in the absence or presence

of CT-DNA for diverse r (0 to 2.2) values. From the absorption data, the

binding constant K_bwas determined from a plot of [DNA]/(ε_a- ε_f) vs.

[DNA] using Eq. (2) (see ESI).

The competitive binding of each compound against EB (3,8-

Diamino-5-ethyl-6-phenyl-phenanthridinium bromide) has been in-

vestigated with ﬂuorescence spectroscopy in order to establish whether

the compound can displace EB from its CT-DNA-EB complex. The CT-

DNA-EB complex was prepared by adding 1.8 × 10⁻⁴M CT-DNA and

1.2 × 10⁻⁴M EB in a buﬀer (10 mM Tris-HCl and 150 mM NaCl at

pH 7.4). Recording of ﬂuorescence spectra was carried out using RF-

1501 PC spectroﬂuorometer by keeping the concentration of CT-DNA-

EB complex constant while varying the compound concentration from 0

to 2.0 × 10⁻⁴M. The ﬂuorescence emission spectra were measured in

the wavelength range of 550–700 nm with an excitation wavelength at

520 nm.

HSA molecule was cleaned up from the water and co-crystallized

ligands; hydrogens were added and Gasteiger charges were computed.

All the ligands used for the docking process were taken from either their

X-ray structures or mother macromolecules and optimized by QM using

the procedure given above. Finally, the “.mol2” 3D ligand ﬁles were

extracted from Gaussian outputs along with corresponding charges that

The Stern-Volmer constant K_SVis used to evaluate the quenching

eﬃciency for each complex according to Eq. (3) [53,54] (see ESI). K_SVis

the Stern-Volmer constant and can be obtained by the slope of the

diagram F₀/F versus [Q].

5

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Scheme 1. Synthetic route for isothiazole ligands 1, 2 and 3 (A) and corresponding ruthenium(II) [Ru(η⁶-p-cymene)Cl₂(L)] complexes 4, 5 and 6 (B).

have been used for the docking purpose. The receptor-ligand interac-

tions were calculated and visualized by Discovery Studio Visualizer

v17.2.0.16349 [69]. For the DNA docking of the standard [Ru(Ar)(L)

(H₂O)_n]²⁺(n = 0, 1, 2) complexes the grid box of 60 × 60 × 60 points

(default spacing is 0.375 Å) has been used with DG6(N7) atoms taken as

its center.

compounds was expressed as the IC₅₀value.

2.6.4. Flow cytometry

2.6.4.1. Cell cycle analysis. HeLa cells were treated with tested

compounds for 48 h at their IC₅₀concentrations. After treatment,

HeLa cells were washed in cold phosphate buﬀer (PBS) and incubated

for 30 min in 70% ethanol on ice, centrifuged and incubated with

500 mL RNAse A (100 units mL⁻¹) and 500 mL propidium iodide

(400 μg mL⁻¹) for 30 min at 37 °C. The cell cycle was analyzed by

FACS Calibur E440 (Becton Dickinson) ﬂow cytometer and the Cell

Quest software. Results were presented as the percentage of cell cycle

phases.

The best GOLD hit (DNA-Ligand) was further optimized by MOPAC-

MOZYME PM6 [58] method. This includes preoptimization of hydrogen

atoms only with MOPAC's keywords NOOPT and OPT-H following by

the full optimization of the whole system using MOZYME. Covalently

bounded sphere around ruthenium(II) ion has been further optimized

by QM using the above given details.

2.6. Biological tests

2.6.4.2. Annexin-V-FLUOS assay. Apoptosis of HeLa cells was evaluated

with an Annexin V-FITC detection kit. Treated cells from each sample

were collected (800 rpm 5 min⁻¹, Megafuge 1.0R, Heraeus, Thermo

Fisher Scientiﬁc) and the pellet was re-suspended in 1 mL of PBS

pH 7.2. HeLa cells were washed twice with cold PBS and then re-

suspended in the binding buﬀer to reach the concentration of 1 × 10⁶

cells mL⁻¹. The cell suspension (100 mL) was transferred to 5 mL

culture tubes and mixed with Annexin-V (5 mL) and propidium iodide

(5 mL). The cells were gently vortexed and incubated for 15 min at

25 °C. After incubation, 400 mL of binding buﬀer was added to each

tube and suspension was analyzed after 1 h on FACS Calibur E440

(Becton Dickinson) ﬂow cytometer. Results were presented as a percent

of Annexin-V positive gated cells. The percentage of speciﬁc apoptosis

was calculated according to Bender et al. [71].

2.6.1. Cell lines

The cell lines used in the study were: MCF-7 (ATCC HTB 22 - human

breast adenocarcinoma ER+), HeLa (ATCC CCL 2 - cervix epithelial

carcinoma), HT-29 (ATCC HTB 38 - human colon adenocarcinoma), A-

549 (ATCC CCL185 - lung carcinoma) and MRC-5 (ATCC CCL 171 -

normal fetal lung ﬁbroblasts). The cells were grown in Dulbecco's

modiﬁed Eagle's medium (DMEM) with 4.5% of glucose, supplemented

with 10% of fetal calf serum (FTS, Sigma) and antibiotics and anti-

mycotics solution (Sigma). The cell lines were cultured in ﬂasks (Costar,

25cm²) at 37 °C in the atmosphere of 100% humidity and 5% of CO₂

(Heraeus). Exponentially growing viable cells were used throughout the

assays.

2.6.2. Tested substances

Cisplatin and tested compounds were used at concentrations from

10⁻⁴M to 10⁻⁸M in order to deﬁne IC₅₀concentration for a particular

3. Results and discussion

time point. The substances were added in a volume of 10 μL well⁻¹

.

The 3,5 diaminoisothiazole derivatives 1–3 were prepared in ex-

cellent yields by following a literature protocol [42,43]. These ligands

were used for synthesis with the [Ru-(η⁶-p-cymene)Cl₂]₂in the presence

of toluene (complexes 4 and 6) or dichloromethane (complex 5) under

reﬂux. All synthesized complexes of the general formula [Ru(η⁶-p-

cymene)Cl₂(L)] were obtained in good yields (Scheme 1).

2.6.3. MTT

(3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium

bromide) assay

Antiproliferative activity was investigated by tetrazolium colori-

metric MTT assay [70], as previously described [39]. The eﬀect of the

6

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

obvious that the ligands are coordinated to the Ru(II)-ion via the -C ≡ N

group. The complexes also displayed ν = _CH,ν_CH, and ν_C=Cabsorptions

in the region of 3266–3432 cm⁻¹

1524–1580 cm⁻¹, respectively.

,

2923–2969 cm⁻¹and

UV–Vis spectra for all complexes were recorded from 200 to 800 nm

in methanol. Absorption bands that were observed in the region

337–342 nm region can be attributed to n-π^⁎and π -π^⁎ligand-centered

transitions, respectively. Furthermore, the low-intensity bands in the

visible region 390–430 nm could be assigned to the metal to ligand (dπ-

π^⁎) charge transfer (MLCT) transition from the ﬁlled 4d orbital to the

empty π^⁎orbital, similar to the MLCT observed in other reported Ru(II)-

arene complexes [72].

The NMR spectra were recorded in CDCl₃to conﬁrm the bonding of

the ligand to the ruthenium(II) ion. In the ¹H NMR spectra from each

complex, a sharp singlet appeared at around 2.19–2.99 ppm that has

been assigned to the CH₃group. The complexes 5 and 6 show two

triplets each originating from protons belonging to 4-methylpiperazine

(δ = 2.51 ppm and δ = 3.58 ppm) and morpholine (δ = 3.50 ppm and

δ = 3.78 ppm). Complex 4 is the only one to show quintet at δ = 1.95,

which is derived from pyrrolidine protons. In complex 5, an additional

methyl signal is observed at 2.33 ppm, which is due to the methyl group

in the 4-methylpiperazine. All the other signals belong to ruthenium-p-

cymene.

3.3. Crystal structures

The molecular structures of 1, 2 and 3 are presented in Fig. S15 (see

ESI) while the selected bond lengths and bond angles are presented in

Table S1 (see ESI). Both, 1 and 2, structures crystallize in centrosym-

metric P2₁/n while 3 crystallizes in the triclinic crystal system and P¯1

space group, and all asymmetric parts of the unit cells in the crystal

structures contains neutral molecules. If we validate the conformation

of ligand structures through the values of dihedral angles (Table S1,

ESI) we can notice that the isothiazole-4‑carbonitrile part in all three

compounds (1, 2, 3) can be considered as almost perfectly planar. On

the other hand, pyrrolidine, methylpiperazine and morpholine part in

1, 2 and 3, respectively, gives rise to conformational diﬀerences since

the methylpiperazine in 2 and morpholine in 3 are adopting the same

chair conformation where the puckering parameters are:

QT = 0.556(2) Å, θ₂= 177.3(2)° and QT = 0.552(2) Å, θ₂= 174.3(2)°,

respectively, while the ﬁve-membered pyrrolidine in 1 has almost

planar conformation with puckering parameters q₂= 0.118(5) Å,

φ₂= − 84(2)°. The planarity and conformation diﬀerences of 1, 2 and

3 compounds are visually depicted in Fig. S15 (see ESI) which displays

an overlay of three structures. The structures of all three ligands are

stabilized by two hydrogen bonds between the cyano group of one

molecule and the amino group of another molecule, as follows: 2.327

(1), 2.232 (2), 2.242 (3).

Fig. 2. MERCURY [51] drawings of the molecular structure of complexes 4 and

6 with labeled non-H atoms.

3.1. Stability

The detailed syntheses of [Ru(η⁶-p-cymene)LCl₂] are presented in

Scheme 1. Before use in the biological experiments, the purities of the

investigated compounds were determined by UV–Vis and NMR methods

(Figs. S13 and S14, ESI). Because of the probable metalation of S atoms

in DMSO with Ru metal, we ﬁrst tested the chemical stability of N-

coordinated [Ru(η⁶-p-cymene)LCl₂] in 1% DMSO (MTT assay) through

the above methods. As shown in Fig. S13 (ESI), the shape of peaks and

their relative position remained unchanged during all tested time points

(24 h). Taken together, these data indicated that the structures of [Ru

(η⁶-p-cymene)LCl₂] are stable in DMSO solution at least for 48 h.

A perspective view of the molecular structures of complexes 4 and 6

with the adopted atom-numbering scheme is shown in Fig. 2. Selected

metal-ligand bond lengths, bond angles, and torsion angles are listed in

Table 2. Complex 4 crystallizes in the triclinic crystal system and P¯1

space group where each asymmetric unit consists of two moieties: one

neutral Ru-p-cymene-complex and one water molecule. In the crystal

structure of 4 the Ru(II) ion is coordinated by two Cl⁻ions and nitrogen

donor atom of 1 with the Ru1-N1 bond distance of 2.064 (3) Å, while

Ru1—Ar_centroidbond distance (π bond) has the value of 1.661 Å. The

crystal packing of complex 4 is arranged by hydrogen bonds involving

H₂O molecules and nitrogen and chlorine atoms (Table S2, ESI).

Complex 6 crystallizes in the monoclinic crystal system and P2₁/n

space group where each asymmetric unit consists of the one neutral

form of Ru-p-cymene-complex. In the crystal structure of 6, the Ru(II)

ion is coordinated by two Cl⁻ions and nitrogen donor atom of ligand 3

with a Ru1-N1 bond distance of 2.059 (3) Å. Ru1—Ar_centroidbond

distance (π bond) in 6 has a value of 1.662 Å. The crystal packings of all

ligands 1, 2 and 3 are arranged in forms of dimers that are formed by N-

3.2. Spectroscopic analysis

The FT-IR, ¹H NMR, ¹³C{¹H} NMR, and UV–Vis were used to

characterize the synthesized compounds (Figs. S1-S10, ESI). The spec-

tral data obtained were in good agreement with the proposed struc-

tures.

Perhaps the most important analyses use FT-IR results. The IR

spectra of the complexes displayed a strong band in the region of

2200–2239 cm⁻¹which is characteristic of the -C ≡ N functional

group. This band is located at a higher frequency relative to the same

band in the spectra of the uncoordinated ligands (Fig. S3, ESI). It is

7

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Table 2

The K_SVand quenching constants (k_q) of the interactions of the

compounds with the albumins were calculated (Table 3) from the

corresponding Stern-Volmer plot (see inset in Figs. S16–S21, ESI) and

Stern-Volmer quenching equation (Eq. (3)) (see ESI), where the ﬂuor-

escence lifetime of tryptophan in HSA was taken as τ₀= 10⁻⁸s.

As seen in Table 3, the quenching constants (> 10¹¹M⁻¹s⁻¹) are

higher than diverse kinds of quenchers for biopolymers ﬂuorescence

(10¹⁰M⁻¹s⁻¹), suggesting that the interaction of the ligands and

complexes with the albumins takes place via a static quenching me-

chanism, which indicates the formation of a new conjugate between

each complex and HSA [67].Using the equation (Eq. (3)) (see ESI), the

values of K (association binding constant) and n (number of binding

sites per albumin) for the ligands 1–3 and complexes 4–6 were obtained

from the intercept and slope of the plots of log (F₀− F)/F versus log [Q]

(Figs. S22 and S23, ESI).

Selected geometric parameters for complexes 4 and 6.

4

6

Bond length [Å]

Ru1—N1

Ru1—Cl1

Ru1—Cl2

Ru1—C10

Ru1—C11

Ru1—C12

Ru1—C13

Ru1—C14

Ru1—C15

2.064 (3)

2.412 (1)

2.423 (2)

2.158 (4)

2.183 (4)

2.210 (4)

2.160 (4)

2.154 (4)

2.190 (4)

2.059 (3)

2.395 (1)

2.422 (1)

2.154 (4)

2.166 (4)

2.218 (3)

2.175 (3)

2.147 (4)

2.194 (3)

Bond angles [°]

Cl1—Ru1—Cl2

N1—Ru1—Cl1

N1—Ru1—Cl2

N1—Ru1—C10

N1—Ru1—C11

N1—Ru1—C12

N1—Ru1—C13

N1—Ru1—C14

N1—Ru1—C15

88.36 (4)

86.61 (8)

84.14 (8)

92.37 (1)

94.92 (1)

121.28 (1)

159.73 (2)

153.3 (2)

115.71 (2)

86.91 (6)

83.40 (9)

81.97 (8)

91.91 (1)

110.92 (1)

145.91 (1)

168.58 (1)

131.17 (1)

100.03 (1)

The values of the binding constant, K and those of n are given in

Table 3. The calculated value of n is around one for all of the com-

pounds, indicating the existence of just a single binding site in HSA for

all of the compounds. From the values of K, it is inferred that complex 4

interacts with HSA more strongly than the rest of the compounds. The

HSA–binding constants of complexes

4

(1.04 × 10⁶M⁻¹),

5

(2.80 × 10⁵M⁻¹) and 6 (4.60 × 10⁵M⁻¹) showed that there is a

strong binding force between complex and HSA, which implies that

HSA can transfer investigated complexes toward potential biotargets.

As shown in Table 3, ligands have lower binding aﬃnity for albumin in

relation to their complexes.

Torsion angles [°]

Ru1—C15—C16—C17

Ru1—C15—C10—C11

Ru1—C11—C12—C13

Ru1—C13—C12—C19

Ru1—C13—C14—C15

−68.4 (6)

−54.1 (3)

−53.1 (3)

−123.6 (4)

−53.9 (4)

−177.6 (3)

−54.5 (3)

−52.1 (3)

−123.9 (4)

−55.4 (3)

UV–Vis absorption measurement is a very simple but eﬀective

method that is used to investigate structural changes and to explore

complex formation. The changes (hypochromism/hyperchromism and/

or red shift/blue shift) observed in the UV spectra during titration may

provide evidence of the existing interaction mode of compounds and

HSA. Collisional (dynamic) quenching only aﬀects the excited states of

the ﬂuorophores and, thus, no changes in the absorption spectra are

expected. In contrast, the groundstate complex formation will fre-

quently result in perturbation of the absorption spectrum of the ﬂuor-

ophore [52]. In the present study, the change in the UV–Vis absorption

spectra of the HSA-compound system (Fig. 4 and Figs. S24–S28, ESI)

was measured under simulated physiological conditions.

H···N hydrogen bonds; the crystal structure of 6 is stabilized by in-

tramolecular hydrogen bond involving NH group of ligand 3 and Cl⁻

ion (Table S2, ESI).

3.4. Interaction of the ligands and complexes with HSA

The most important role of the serum albumins is the transportation

of many biologically active compounds (drugs, natural products, metal

ions, metal complexes, etc.) in the blood. The investigation of binding

interactions between the biologically potent compounds and HSA can

be important in exploring their potential biological activity and appli-

cation. The degree of binding of the compound to albumin determines

the availability of the compound in the organism. It is well known that

only the unbound fraction of the compound (drug) in plasma is free to

cross the cell membrane and to produce pharmacological eﬀects [73].

In order to investigate the structural changes in HSA caused by the

addition of ligand or complex and determine the binding constant (K)

and the number of binding sites (n) for the complex formed between

ligand or complex and HSA, absorption and ﬂuorescence spectra were

measured.

HSA has a weak absorption peak at about 280 nm because of the

cumulative absorption of three aromatic amino acid residues (Trp, Tyr,

and Phe). The absorption intensity at 280 nm increased progressively

(Fig. 4 and Figs. S24–S28, ESI) with the addition of ligands 1–3 or

complexes 4–6, suggesting that the complex was formed between li-

gand or complex and HSA and that the polypeptide chain of HSA suc-

cessively unfolded upon the addition of this complex. The maximum

peak position of HSA-compounds was shifted slightly toward the lower

wavelength region. The change in λ_maxindicates the change in polarity

around the tryptophan residue and the change in the peptide strand of

HSA molecules and hence the change in hydrophobicity [74]. This re-

sult reconﬁrms that the probable ﬂuorescence quenching mechanism of

HSA by compounds is a static quenching process [75].

The study of ligand binding to a variety of (usually biological)

molecules is often performed using ﬂuorescence quenching metho-

dology. Quenching is any process that decreases the ﬂuorescence in-

tensity (excited–state reactions, molecular rearrangements, energy

transfer, ground-state complex formation, and collisional quenching).

The quenching mechanism is usually classiﬁed as either dynamic or

static quenching. HSA solutions exhibit a strong ﬂuorescence emission

with a peak at about 350 nm, due to the tryptophan residues, when

excited at 295 nm [54]. The ﬂuorescence spectra of HSA with diﬀerent

concentrations of ligands (1–3) and complexes (4–6) were recorded and

are shown in Figs. S16-S21 (ESI).

3.4.1. Kinetics

Metal complexes can bind to HSA by diﬀerent binding modes.

Associative manner includes the replacement of labile aqua (or chorido)

ligand by protein donor to the metal center. This binding mechanism is

usually governed by the thermodynamic stability and kinetic inertness

of the complexes. On the other hand, the dissociative mode is speciﬁc

for labile complexes that during the interaction decompose partially or

completely. Considering the rate constants and activation parameters

(Table 4), studied interaction is very slow, while the negative value of

the entropy of activation indicates an associative mode that includes

strong coordinate (covalent) bonding of the complex through the dis-

placement of the labile aqua (or chlorido) ligand by protein donor

atoms preserving its original entity. IR spectral changes conﬁrmed

lower aﬃnity binding of this complex, and non-dissociated complex is

As shown in Figs. S16–S21 (ESI), adding of complex to the HSA

solution, the ﬂuorescence intensity of HSA decreased gradually with an

increase in complex concentration. This result suggests that complex

can interact with HSA and quench its intrinsic ﬂuorescence. The addi-

tion of ligands (1–3) or their complexes (4–6) to HSA results in ﬂuor-

escence quenching (up to about 15–20% of the initial ﬂuorescence in-

tensity of HSA for ligands and about 30–70% for complexes) (Fig. 3).

8

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Fig. 3. Plot of EB relative ﬂuorescence intensity at λem = 350 nm (%) versus r (r = [complex]/[HSA]) for ligands (A) and their complexes (B) in buﬀer solution

(10 mM Tris–HCl and 150 mM NaCl at pH 7.4).

Table 3

The HSA binding constants and parameters (K_SV, k_q, K, n) for the ligands 1–3 and complexes 4–6.

Compound

K_SV(M⁻¹

)

K_q(M⁻¹s⁻¹

)

R^2a

K (M⁻¹

)

ΔG (kcal)^b

N

R^2a

1

2

3

4

5

6

1.45 × 10⁴

7.17 × 10³

1.20 × 10⁴

1.47 × 10⁵

2.93 × 10⁴

7.71 × 10⁴

1.45 × 10¹²

7.17 × 10¹¹

1.20 × 10¹²

1.47 × 10¹³

2.93 × 10¹²

7.71 × 10¹²

0.985

0.963

0.988

0.974

0.985

4.95 × 10⁴

7.28 × 10³

4.39 × 10⁴

1.04 × 10⁶

2.80 × 10⁵

4.60 × 10⁵

−6.40

−5.27

−6.33

−8.21

−7.43

−7.72

1.12

0.98

1.13

1.17

1.20

1.16

0.987

0.990

0.987

0.984

0.983

0.980

a

b

R²is the correlation coeﬃcient.

ΔG° = -RTlnK.

assumed to bind mainly on the protein. UV–Vis studies showed minimal

spectral changes in the presence of HSA, implying small, or no re-

arrangement in the coordination sphere of Ru(II) ion. Combining the

results of the solution studies applied in this work, we can conclude that

the binding of the ligand- complex to HSA is thermodynamically less

favored.

coordination sphere of Ru-p-cymene is saturated by the donor groups of

the protein only [77]. Metal complexes studied display diﬀerent

binding modes. Our results suggest the following mechanism for al-

bumin binding: as initial step a protein donor atom coordinates

monodentately to the metal center replacing the aqua (or chlorido)

leaving the group in the metal complex (slow associative binding), that

is optionally followed by coordination of additional protein donor atom

(s) and consecutive release of the original ligand (dissociative binding).

The binding mechanism is governed by the thermodynamic stability

and kinetic inertness of the complexes.

X-ray single crystal structures reported for protein–Ru-p-cymene

complex adducts often show coordination of amino acids His or Glu to

the metal center [76–82], and coordination of Lys, Arg, Asp and Cys

side chains occur as well [77,82,83]. Coordination of one or two of

these side chains is typical, and the loss of p-cymene ligand can be

detected in some cases as well [76,77,83]. Apo-ferritin coordinated

[(NHis, NHis, OGlu)] is the only reported structure, where the

9

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Fig. 4. Absorption spectra of HSA (2 × 10⁻⁶M), with various

amounts of the 4 complex (0–1.6 × 10⁻⁵M) at room tem-

perature. Inset: UV–Vis absorption spectra of HSA in the ab-

sence and presence of complex 4: A, the absorption spectrum

of HSA only; B, the absorption spectrum of complex 4 only; C,

diﬀerence between the absorption spectrum of HSA-complex

4 and complex 4; D, the absorption spectrum of HSA-complex

4.

In the presence of CT-DNA, the absorption band of ligand 1 and

complex 4 at 264 nm exhibited hyperchromism (as in all other in-

vestigated compounds). This increasing absorbance indicates that there

are interactions between the complex and the base pairs of DNA. The

extent of the hyperchromism in the charge transfer band is generally

consistent with the strength of interaction [90]. As DNA double helix

possesses many hydrogen bonding sites which are accessible both in the

minor and in the major grooves, N and O atoms on ligands likely form

hydrogen bonds with DNA, which may contribute to the hyperchro-

mism observed in absorption spectra. The increasing absorbance in-

dicates that there are groove binding modes [91]. The observed hy-

perchromicity for ligands and complexes after addition of CT-DNA

suggested their electrostatic interaction with DNA. Also, ligands con-

taining aromatic moiety in the complex may bind to the base pairs of

DNA in intercalative mode [92]. The intrinsic binding constants K_bof

ligands and complexes were calculated according to the classic equation

(Eq. (1)) and are given in Table 5. The calculated K_bvalues for ligands

(1–3) and complexes (4–6), were obtained from the plot of [DNA]/(ε_a-

ε_f) vs. [DNA] (see Figs. S33 and S34, ESI) (Table 5), suggest a moderate

binding of the compounds to CT-DNA.

Table 4

Second-order rate constants and activation parameters for the reaction between

complex 4 and HSA.

T (K)

k₁(M⁻¹s⁻¹

)

ΔH^≠(kJ mol⁻¹

)

ΔS^≠(JK⁻¹mol⁻¹

)

288

298

308

(6.80

(1.86

(3.07

0.02) × 10⁻⁴

0.03) × 10⁻⁴

53

2

−154

7

3.5. Interaction of the ligands and complexes with CT-DNA

Transition metal complexes can bind to DNA via both covalent and/

or noncovalent (intercalation, electrostatic or groove binding) interac-

tions [84,85]. Absorption spectroscopy is one of the most useful

methods for studying the binding of compounds to DNA [86]. The in-

teraction of the metal complexes with the base pairs of DNA is usually

followed by a hypochromic shift with a small red/blue shift [87]. On

the other hand, the hyperchromic shift might be ascribed to external

contact (electrostatic interactions) or to the capacity of the complex to

uncoil the helix structure of DNA [88,89]. The absorption spectra of the

ligand 1 and complex 4 in the absence and presence of CT-DNA are

shown in Fig. 5 (others are given in Figs. S29-S32, ESI).

The K_bvalues of complexes 5 and 6 are lower than their ligands 2

and 3 (Table 5) suggesting that complexes have lower binding aﬃnity

Fig. 5. UV–Vis absorption spectra of (a) ligand 1 and (b) complex 4 in the absence and presence of CT-DNA: A, the absorption spectrum of ligand 1 or complex 4 only;

B, the absorption spectrum of DNA only; C, the absorption spectrum of DNA-compound; D, the diﬀerence between the absorption spectrum of DNA-compound

complex and compound. Inset: absorption spectra of the compound in the presence of

a

range of DNA concentrations.

[Ligand] = [Complex] = 8.0 × 10⁻⁵mol dm⁻³, [DNA] = 0–1.7 × 10⁻⁴mol dm⁻³

.

10

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Table 5

The DNA binding constants (K_b), calculated from UV spectra and the Stern–Volmer constants (K_sv), and Quenching constant (K_q) calculated from ﬂuorometric

spectra.

Compounds

In the absence of EB

In the presence of EB

K

_b(M⁻¹⁾

ΔG (kcal)^b

R^2a

K_SV(M⁻¹

)

K_q(M⁻¹s⁻¹

)

R^2a

1

2

3

4

5

6

2.63 × 10⁴

3.67 × 10⁴

9.17 × 10⁴

4.97 × 10⁴

1.85 × 10⁴

1.04 × 10⁴

−6.03

−6.22

−6.77

−6.41

−5.82

−5.48

0.991

0.915

0.976

0.965

0.987

0.975

5.39 × 10³

2.08 × 10³

1.80 × 10³

1.34 × 10³

1.50 × 10³

2.32 × 10³

5.39 × 10¹¹

2.08 × 10¹¹

1.80 × 10¹¹

1.34 × 10¹¹

1.50 × 10¹¹

2.32 × 10¹¹

0.941

0.938

0.996

0.967

0.913

0.992

a

b

R²is the correlation coeﬃcient.

ΔG° = -RTlnK.

Fig. 6. Plot of EB relative ﬂuorescence intensity at λ_em= 613 nm (%) versus r (r = [complex]/[DNA]) for ligands (A) and their complexes (B) in buﬀer solution

(10 mM Tris-HCl and 150 mM NaCl at pH 7.4).

to CT-DNA. On the other hand, complex 4 exhibits higher K_bvalue than

its ligand. The K_bvalues of all compounds are lower than the classical

out with ﬂuorescence measurements. The emission spectra of EB bound

to CT-DNA in the absence and presence of ligands or complexes were

recorded for [DNA] = 1.8 × 10⁻⁴M, [EB] = 1.2 × 10⁻⁴M EB and in-

creasing amounts of each compound. The emission spectra of EB bound

to CT-DNA in the presence of a compound 1–6 derived for diverse r

values are shown in Figs. S35–40 (see ESI). The intensity of the emis-

sion band of the CT-DNA-EB complex at 613 nm decreased with

intercalator EB binding aﬃnity for CT-DNA, (K_b= 1.23 × 10⁵M⁻¹

)

[93] and these values are in agreement with those of well–established

groove binding rather than classical intercalation [94].

In order to examine the ability of the compounds to displace EB

from the EB-DNA complex, competitive EB binding studies were carried

11

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Table 6

3.6. Computational chemistry

X-Ray vs. DFT bond length for complexes 4 and 6.^a

The structure of isothiazole ligands and corresponding Ru(II) com-

plexes have been solved by the X-ray analysis (an exception is complex

5 with piperazine substituent). This fact allowed us to compare their

experimental structural properties with the results of density functional

theory (DFT) calculations.

X-ray

DFT

Complex 4

Ru1—N1

Ru1—C10

Ru1—C11

Ru1—C12

Ru1—C13

Ru1—C14

Ru1—C15

Ru1—Cl1

Ru1—Cl2

S1—N2

Complex 6

Ru1—N1

Ru1—C10

Ru1—C11

Ru1—C12

Ru1—C13

Ru1—C14

Ru1—C15

Ru1—Cl1

Ru1—Cl2

S1—N2

2.064 (3)

2.158 (4)

2.183 (4)

2.210 (4)

2.160 (4)

2.154 (4)

2.190 (4)

2.412 (1)

2.423 (2)

1.670 (3)

Ru1—N25

Ru1—C5

Ru1—C21

Ru1—C6

Ru1—C42

Ru1—C29

Ru1—C43

Ru1—Cl2

Ru1—Cl4

S3—N9

2.0252 (1)

2.2246 (1)

2.1932 (1)

2.2283 (1)

2.2618 (1)

2.2324 (1)

2.2163 (1)

2.4872 (1)

2.5142 (1)

1.8018 (1)

For this purpose, we carried out the geometry optimization of all

ligands and their ruthenium(II) complexes starting from X-ray struc-

tures except for the complex 5 which has been modeled, preoptimized

and then fully relaxed by the Gaussian. At ﬁrst glance, the optimized

structures are in good agreement with the experimental ones (Table 6).

This means that the bonds, angles, and torsions ﬁt well with X-ray

ﬁndings except for slightly longer bonds (particularly in case of NeS

bonds: 1.670 Å/DFT vs. 1.8018 Å/X-ray (complex 4), 1.663 Å/DFT vs.

1.797 Å/X-ray (complex 6)). DFT optimized structure of the unique

unsolved complex 5 is given in Fig. 7, showing structural consistency

with all the crystallographic structures.

2.059 (3)

2.154 (4)

2.166 (4)

2.218 (3)

2.175 (3)

2.147 (4)

2.194 (3)

2.395 (1)

2.422 (1)

1.663 (4)

Ru52—N32

Ru52—C1

Ru52—C2

Ru52—C3

Ru52—C4

Ru52—C5

Ru52—C6

Ru52—Cl53

Ru52—Cl54

S1—N2

2.0221 (1)

2.2186 (1)

2.2321 (1)

2.2242 (1)

2.2587 (1)

2.1966 (1)

2.2270 (1)

2.5115 (1)

2.4867 (1)

1.7970 (1)

3.6.1. Docking

The study of metal–protein systems is of fundamental importance in

biology, pharmacy, and medicine [96]. Covalent (coordinative)

bonding (metalation) of Ru(II) piano stool complexes with proteins has

been described by many authors [24]. Often these systems cannot be

investigated through X-ray diﬀraction analysis, and other methods are

necessary to characterize the structure and active site. The common

spectroscopic and analytical techniques, such as NMR, EPR, ESEEM,

ENDOR, ESI-MS, CD, and UV–Vis spectroscopy, often do not provide

information on the region of the protein where the metal species are

bound or on the amino acid residues involved in the coordination.

Concerning the systems with covalent interactions, some authors re-

cently tried to simulate the binding of small inorganic molecules to

proteins, [97,98] but until now there were very few examples in which

the docking methods have been systematically applied to the simulation

of the covalent bonds between a protein and a metal species which, as

mentioned above, represent most of the situations [66]. The method of

Sciortino et al. [66] was thoroughly validated and results were always

successful for 39 “pdb” structures, the pose with the highest aﬃnity was

the one suggested by the X-ray analysis; moreover, the crystallographic

structure is reproduced with a success rate of 100% with an RMSD <

2.5 Å and of 90% with an RMSD < 1.5 Å [66].

a

[Å].

We decided to check the interactions between our ligand and

complexes against the two most abundant macromolecules HSA and

DNA.

Fig. 7. DFT optimized structure of the complex 5 (hydrogen atoms were

omitted having better clarity of picture).

3.6.2. HSA-L Covalent approach

increasing concentration of the compounds (Figs. S35-S40, ESI).

The addition of the ligands 1–3 and complexes 4–6 at diverse r

values (Fig. 6) results in a decrease of the ﬂuorescence intensity. This

decrease of EB ﬂuorescence (up to 30% of the initial CT-DNA-EB

ﬂuorescence intensity) suggests that they displace EB from the CT-DNA-

EB complex so they can interact with CT-DNA probably by the inter-

calative mode [95].

HSA is an abundant plasma protein that binds a wide variety of

hydrophobic ligands including fatty acids, bilirubin, and thyroxine. The

GOLD suite was used for the study of HSA-Ligand docking. According to

reported works on similar systems, we followed the next suggestions:

docking on surface-exposed His-128, His-247, His-510 and Met-298 are

denoted as coordination sites of Ru(arene) complexes [99].

Here we investigated the new GOLD docking method in the pre-

diction of ruthenium(II)–HSA structures without using any geometrical

constraints or energy restraints. In particular, our systems with ruthe-

nium(II) metal-containing ligands were examined following the above

ﬁndings and new coordination scoring parameters were generated.

Following the results of kinetic studies that involve two HSA binding

steps (associative slow and dissociative fast) we choose, at the ﬁrst

glance, the active species which were assumed on the basis of quantum

mechanical thermodynamic parameters for the reactions given in

Scheme 2.

The quenching parameters of ligands and complexes were calcu-

lated according to the Stern–Volmer equation (Eq. (3)) (see ESI) and are

given in Table 5. The K_SVvalue was obtained from the slope in the plot

of F₀/F versus [compound] (see inset in Figs. S38, S41 and S42, ESI).

The complex 4 exhibits the lowest K_SVvalue (Table 5) and among all

the compounds studied have the lowest ability to displace EB from its

CT-DNA-EB complex. On the other hand, ligand 1 has the highest K_SV

value and the greatest ability to displace EB from its CT-DNA-EB

complex.

As can be seen from Table 5, the values of K_qwere greater than

10¹⁰M⁻¹s⁻¹, indicating that the quenching mechanism, as a result of

the formation of the CT-DNA-EB-compound complex, is a static

quenching process.

According to the results, we may assume that reactions 1. and 2. are

very probable in protein surroundings and have been taken in the Gold

docking process. Every leaving ligand has been replaced by hydrogen

simulating bonding d orbital of ruthenium(II) (hydrogen acceptor). The

12

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Scheme 2. Ru(II)-complexes hydrolytic reaction as well as their reactions with amino acids and nucleic bases involved in binding with HSA and DNA.

Table 7

GOLD Docking results for reactive species of complex 4 within HSA molecule.

^⁎t1, t2, t3 and t4 denotes tyrosine bounded within 1HK1 molecule.

docking has been carried out on every active site of HSA and surface

exposed cysteine and histidine. Our results are summarized in Table 7

and Fig. 8. It is more than clear that active species [Ru(p-cymene)LCl

(H₂O)]⁺and [Ru(p-cymene)Cl₂(H₂O)] are less active toward any his-

tidine or cysteine but glutamine-292 being the common binding place

in an associative mechanism of protein binding.

The results are indicative for HSA GLU-292 binding. Gold scores of

[Ru(p-cymene)LClH], [Ru(p-cymene)Cl₂H] and [Ru(p-cymene)LH₂] are

the best scores of all reactive species. However, the last one appears to

be a product of a dissociative step. Therefore our choice falls on the [Ru

(p-cymene)LClH] regarding to higher Gold score 49.8461 and a much

higher VdW score 50.1322 than [Ru(p-cymene)Cl₂H] species. More

13

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Fig. 8. (a) HSA with docked reactive species of complex 4, and orientations of docked conformations in domain IIA cavity for (b) [Ru(p-cymene)LClH] (c) [Ru(p-

cymene)Cl₂H] (d) complex 4 (e) [Ru(p-cymene)LH₂] and (f) [Ru(p-cymene)H₂].

Fig. 9. Potential covalently binding residues inside cavity at the domain IIa of HSA with docked structures of reactive species of complex 4 (a) [Ru(p-cymene)Cl₂L]

(b) [Ru(p-cymene)Cl₂H] and (c) [Ru(p-cymene)LClH] (d) GOLD model as the Ru-H-O(GLU292) hydrogen bond interactionin case of (c).

important is the associative mechanism involving [Ru(p-cymene)LCl₂]

and HSA-GLU-292 account on chloride abstraction which is enough

energy demanding to slow the binding process. Docked positions in

relation to the residues that could potentially enter into covalent in-

teractions inside the cavity of HSA at the IIa domain for complex 4 and

some of the reactive species can be seen in Fig. 9.

fast reaction 7., appears to be very likely where neutral [Ru(p-cymene)

LClGLU] complex is formed according to high negative ΔG_reactionvalue.

3.6.3. DNA

The above theory is indicative in terms of the formation of the

protein-ligand complex. However, solution studies showed minimal

spectral changes in the presence of HSA, implying small, or no re-

arrangement in the coordination sphere of Ru(II) ion. Therefore we can

conclude that fast strong binding of the ligand/complex to HSA is

thermodynamically less favored. This may lead us to conclusion that

HSA actually has a transporter role at least within the timeframe of 48 h

(according to kinetics experiments). One may expect for this prodrug to

The results of docking studies that involve ﬁrst associative HSA

binding step has been studied by the quantum mechanics as well

(Scheme 2) and, one may see, given thermodynamic parameters for the

exchange reactions, results are in accordance with our conclusion. This

means that taking into account the ﬁrst slow associative binding step of

[Ru(p-cymene)LClH₂O]⁺to HSA and GOLD results (Fig. 9d) the second

14

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Table 8

Docking results of [Ru(η⁶-p-cymene)X₂(L)] (X = Cl or H₂O) and diﬀerent li-

gands toward DNA protein.

Compound

DNA

AutoDock

ΔG^a

Vina

K_i^b

Aﬃnity^a

1

2

3

4

5

6

−3.90

−4.11

−4.27

−7.00

−6.75

−6.40

1.38^c

−5.30

−5.60

−5.70

−7.10

−7.50

−6.90

974.64

745.49

7.36

Fig. 10. [A] MOPAC-MOZYME PM6 optimized DNA/(aquo complex of 4)

structure of the best GOLD docking hit; [B] Overlay of MOPAC-MOZYME PM6

(green atoms) and QM (g09) optimized ruthenium(II) coordination sphere with

RMSD = 1.051 Å.

11.30

20.32

a

b

c

kcal mol⁻¹

.

μM.

Table 10

mM.

The IC₅₀values of tested compounds and cisplatin.^a

Samples

Cell lines

Table 9

GOLD Docking results for complex 4 and reactive species with the DNA.

MRC-5

A549

90.25

MCF-7

˃100

HeLa

˃100

HT-29

Molecule

GoldScore

H-bond

VdW

Closest

1

˃100

1.48

˃100

[Ru(p-cymene)L(H₂O)H]

46.89

7.80

41.55

dG7

2

˃100

25.64

45.3

3

89.07

7.04

6.34

˃100

42.48

4

32.46

18.15

28.14

6.75

˃100

5

7.98

31.45

6

13.59

11.6

5.94

51.73

cis Pt

1.77

15.9

a

[μM].

[Ru(p-cymene)(H₂O)₂H]

[Ru(p-cymene)H₂]

37.50

38.98

30.88

30.04

10.00

9.80

2.23

0.00

29.66

29.21

28.98

30.12

dG7

groove binding inside the DNA has been found as the best binding site

in the case of all ligands and corresponding complexes. The results of

the best hits are given in Table 8.

Evidently, the data given in Table 8, supports the hardest binding to

DNA in the case of complex 4, which agrees with the solution study.

The general positions of the ligands found inside the DNA major groove

along with principal interactions are given in Fig. S43 (see ESI).

Natively Gold has been built up for protein-ligand interactions.

Nevertheless, here we have used GOLD with the same modiﬁed para-

meter ﬁle [66] to dock ligands to the DNA dodecamer (as given in

Experimental). For this experiment we have used several possible active

species for which it might be reasonable to expected to form by hy-

drolytic processes inside of the cell (Scheme 2). The results of these

docking experiments (GoldScore ﬁtness, H-bond and VdW indexes) are

given in Table 9. One may see that [Ru(p-cymene)L(H₂O)H] species has

a highest GoldScore (46.89) and VdW index (41.55) with pretty high H-

Bond index (7.80). This implies that this species is very likely when it

comes to the covalent binding of ruthenium(II) complexes to DNA.

Naturally, caution must be taken into account here because full para-

meterization of GOLD in the case of docking to the DNA has not been

done.

[Ru(p-cymene)LH₂]

[Ru(p-cymene)(H₂O)₃]

To check the best docked conformation of [Ru(p-cymene)L(H₂O)

H]²⁺species we have used MOPAC [58] PM6-MOZYME method to

optimize whole DNA/Ligand complex. The ﬁnal solution is given in

Fig. 10 which demonstrate the covalent bonding of Ru(II) and N(7) of

dG7 nucleotide. This complex, that comprehends just guanine from

dG7, was further QM optimized giving structure very similar to the one

obtained by MOZYME (Fig. 10).

enter cancer tissue and release a drug into the cell. Therefore it became

of interest to investigate interactions of ligands with crucial cytoplas-

matic macromolecule such as DNA.

3.7. Biological tests

Firstly two common docking software packages (parametrized

Autodock and Vina) were used to reveal main interactions of DNA

against all the prepared ligands and ruthenium analogues. Therefore,

we have done a docking routine in the case of all the compounds using

the procedure given in Experimental. Commonly established major

Based on the IC₅₀value (Table 10), it is seen that the most sensitive

cell lines to the eﬀects of the tested substances were MCF-7, then A549

and HeLa, while the least sensitive were HT-29 cells. The MCF-7 cell

line is the most sensitive to all of the tested compounds (except 1). In

this case 3, 4 and 5 were about two times more active than cisplatin,

15

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Fig. 11. Cell cycle analysis. (A) Histograms and (B) graph presenting cell cycle distribution in treated HeLa cells through the diﬀerent cell cycle phases after 48 h

exposure to the equitoxic doses (IC₅₀concentrations) of the tested complex 6 and cisplatin (cis Pt), along with untreated control sample.

while the antiproliferative activity of 6 was almost the same as with

cisplatin. The tested compounds did not inhibit the growth of the HT-29

cell line, while the reference compound cisplatin exhibited a strong

antiproliferative activity. Regarding the healthy MRC-5 cell line, the

tested compounds were not toxic, only for the reference compound was

shown to have signiﬁcant activity.

4. Conclusion

Three new [Ru(η⁶-p-cymene)Cl₂(L)] complexes with diﬀerent sub-

stituted izothiazole ligands: [Ru(η⁶-p-cymene)Cl₂(L1)]·H₂O (4), [Ru(η⁶-

p-cymene)Cl₂(L2)] (5) and [Ru(η⁶-p-cymene)Cl₂(L3)] (6) were isolated

and characterized using experimental and computational techniques.

Crystal structures of three ligands and two complexes (4 and 6) have

been veriﬁed by X-ray diﬀraction analysis. The interactions of CT-DNA

and HSA with ligands and new ruthenium(II) complexes have been

studied through absorption and ﬂuorescence measurements. The high

value of the binding constants, K_b, and the Stern–Volmer quenching

constant, K_SV, is the result of good binding of complex 4 for CT-DNA

and HSA. Docking experiments toward DNA dodecamer and HSA have

been done. Kinetic studies on HSA-Ligand interaction suggest an asso-

ciative manner that includes the replacement of labile aqua (or

chlorido) ligand by protein donor to the metal center. The revised

The cell cycle of HeLa cells was examined using ﬂow cytometry to

determine the antiproliferative eﬀect of synthesized substances.

Complex 6 and cisplatin, after a 48-h treatment, reduced the percentage

of cells in the G2/M phase by 40% and 10% relative to control, re-

spectively. Complex 6 and the reference compound reduced the per-

centage of cells in the G0/G1 phase. Only cisplatin increases the per-

centage of cells in the S phase, compared to the control sample. After

exposure of HeLa cells to complexes for 48 h, it was found that complex

6 and cisplatin increase the percentage of apoptotic cells (subG1 phase,

Fig. 11). Flow cytometric analysis of Annexin-V-FLUOS stained cells

showed that both complex 6 induces apoptosis in HeLa cells (Fig. 12).

After 48 h treatment, the majority of cells were apoptotic (early apop-

totic 7.23% and late apoptotic 8.8%), while only a small percent of cells

was necrotic.

GOLD

docking

results

are

indicative

for

Ru(p-cymene)

LCl··(HSA··GLU292) binding. Gold docking scores are in accordance

with the associative mechanism involving [Ru(p-cymene)LClH] and

HSA-GLU-292 that rely on chloride abstraction which is high enough to

slow the binding process. The same EES was used to dock ligands to the

DNA dodecamer giving major groove binding of ligands as the principal

ones. MOPAC-MOZYME optimization of the whole system of the best

docked complex and QM relaxation of the initial ruthenium(II) co-

ordination sphere are indicative for Ru(II)-N(7)dG7 covalent binding.

The results of biological tests showed that the ligands and new Ru(II)

complexes in comparison to the reference compound cisplatin, have

more desirable cytotoxic activity and unlike cisplatin, they are selective

Also, the apoptotic response, shown as a percentage of speciﬁc

apoptosis (Fig. 12), shows that cisplatin multiplies the percentage of

HeLa cells that are positive for Annexin-V compared to the tested

complex. The percentage of speciﬁc apoptosis of complex 6 is 13.36%.

These results are in correlation with the results of the cell cycle analysis.

The percentage of speciﬁc necrosis after 48 h for complex 6 is twice as

high as in the reference compound.

16

M.B. Đukić, et al.

JournalofInorganicBiochemistry213(2020)111256

Fig. 12. Flow cytometric analysis of Annexin-V-FLUOS staining. Dot plots presenting the percentage of viable (lower left quadrant), early apoptotic (lower right

quadrant), late apoptotic (upper right quadrant), and necrotic cells (upper left quadrant). Columns showing the percentage of speciﬁc apoptosis and speciﬁc necrosis

of HeLa cells induced by cisplatin (cis Pt) and complex 6 after 48 h treatment. The percentage of speciﬁc apoptosis was calculated according to reference [71].

against cancer and healthy cell lines. For all the tested compounds re-

sults show that the MCF-7 cell line is the most sensitive. Flow cytometry

analysis showed the apoptotic death of the cells with a cell cycle arrest

in the subG1 phase.

Appendix A. Supplementary data

Crystallographic data are deposited in the Cambridge

Crystallographic Data Centre as supplementary material number CCDC

1939949 - 1939953. Copies of this information may be obtained free of

charge from the Director, CCDC, 12 Union Road, Cambridge, CB21FZ,

UK (email: deposit@ccdc.cam.ac.uk or http://www.ccdc.cam.ac.uk).

Supplementary data to this article can be found online at https://doi.

org/10.1016/j.jinorgbio.2020.111256.