A palladium(II) complex with the Schiff base 4-chloro-2-(N-ethyliminomethyl)-phenol: Synthesis, structural characterization, and in vitro and in silico biological activity studies

Article abstract of DOI:10.1016/j.jinorgbio.2019.110792

The synthesis and characterization of the Pd(II) complex of the formula [Pd(L)₂] 1 with the Schiff base 4-chloro-2-(N-ethyliminomethyl)-phenol (HL) as derived in situ via the condensation reaction of 5-chloro-salicylaldehyde and ethylamine was

Full text of DOI:10.1016/j.jinorgbio.2019.110792

Accepted Manuscript
A palladium(II) complex with the Schiff base 4-chloro-2-(N-
ethyliminomethyl)-phenol: Synthesis, structural characterization,
and in vitro and in silico biological activity studies
Ariadni Zianna, George D. Geromichalos, Anna Pekou, Antonios
G. Hatzidimitriou, Evdoxia Coutouli-Argyropoulou, Maria Lalia-
Kantouri, Anastasia A. Pantazaki, George Psomas
PII:
S0162-0134(19)30210-7
DOI:
https://doi.org/10.1016/j.jinorgbio.2019.110792
Article Number:
Reference:
110792
JIB 110792
To appear in:
Journal of Inorganic Biochemistry
Received date:
Revised date:
Accepted date:
6 April 2019
16 July 2019
21 July 2019
Please cite this article as: A. Zianna, G.D. Geromichalos, A. Pekou, et al., A palladium(II)
complex with the Schiff base 4-chloro-2-(N-ethyliminomethyl)-phenol: Synthesis,
structural characterization, and in vitro and in silico biological activity studies, Journal of
Inorganic Biochemistry, https://doi.org/10.1016/j.jinorgbio.2019.110792
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
A palladium(II) complex with the Schiff base 4-chloro-2-(N-ethyliminomethyl)-
phenol: Synthesis, structural characterization, and in vitro and in silico
biological activity studies
Ariadni Zianna,1,* George D. Geromichalos,1,* Anna Pekou,2 Antonios G. Hatzidimitriou,1
Evdoxia Coutouli-Argyropoulou,3 Maria Lalia-Kantouri,1 Anastasia A. Pantazaki,2,* George
Psomas,1, *
1
Laboratory of Inorganic Chemistry, Department of Chemistry, Aristotle University of
Thessaloniki, Thessaloniki GR-54124, GREECE.
2
Laboratory of Biochemistry, Department of Chemistry, Aristotle University of Thessaloniki,
Thessaloniki GR-54124, GREECE.
3 Laboratory of Organic Chemistry, Department of Chemistry, Aristotle University of Thessaloniki,
Thessaloniki GR-54124, GREECE.
Abstract
The synthesis and characterization of the Pd(II) complex of the formula [Pd(L)2] 1 with the
Schiff base 4-chloro-2-(N-ethyliminomethyl)-phenol (HL) as derived in situ via the condensation
reaction of 5-chloro-salicylaldehyde and ethylamine was undertaken. The structure of 1 was
verified by single-crystal X-ray crystallography. The ability of 1 to interact with calf-thymus (CT)
DNA was studied by UV-vis and viscosity experiments, and its ability to displace ethidium bromide
(EB) from the DNA-EB conjugate was revealed by fluorescence spectroscopy. It was found
intercalation as the most possible mode of interaction with CT DNA. Additionally, DNA
electrophoretic mobility experiments showed that 1 interacts with the plasmid pBluescript SK(+)
(pDNA) as proved by the formation of unusual mobility DNA bands and degradation of relaxed
pDNA at concentration of 5 mM. The interaction of 1 with human (HSA) and bovine serum
albumin (BSA) was monitored revealing its reversible binding to albumins. The complex showed
noteworthy antimicrobial activity against one (Bacillus subtilis) of the five tested bacteria. In order
to explain the described in vitro activity of the compound, we adopted molecular docking studies on
the crystal structure of HSA, BSA, CT DNA and DNA-gyrase. Furthermore, in silico predictive
tools have been employed to study the properties of the complex. The in silico studies are adopted
on a multitude of proteins involved in cancer growth, as well as prediction of drug-induced changes
* Corresponding authors e-mails:
ariadnezianna@gmail.com (A. Zianna);
natasa@chem.auth.gr (A.A. Pantazaki);
gerom@chem.auth.gr (G.D. Geromichalos);
gepsomas@chem.auth.gr (G. Psomas)
1

ACCEPTED MANUSCRIPT
of gene expression profile, protein- and mRNA-based prediction results, prediction of sites of
metabolism, cytotoxicity for cancer cell lines, etc.
Keywords: Palladium(II) complex; Schiff base; DNA-interaction; interaction with serum albumins,
in silico molecular docking; in silico predictive tools
1. Introduction
The interest concerning the synthesis of palladium complexes has been significantly
increased recently because of their numerous potential applications in areas, such as catalysis,
medicine and material science [1-4]. In the field of Bioinorganic Chemistry, palladium compounds
have gained attention mainly due to their similarity regarding coordination modes with platinum
compounds used as anticancer agents [4], although the Pd complexes seem to hydrolyze more
rapidly [5]. Furthermore, many Pd complexes have been reported to exhibit in vitro anti-viral, anti-
fungal, antimicrobial [6], anti-inflammatory [7], and antioxidant properties [8]. The ligands in these
complexes are mainly N-, S- or P-donors, such as thiosemicarbazones, dithiocarbamates,
thioamides, amines, and especially Schiff bases [6,9].
As known, Schiff bases are organic compounds containing an imine group (-RC=N-). They
are usually prepared by the condensation reaction of a primary amine and a carbonyl compound
(either aldehyde or ketone) [10,11]. They are at least bidentate N,O-donor ligands when coordinated
to diverse metal ions. The Schiff bases and their metal complexes have shown a variety of activities
[12], including catalytic [13] and biological (antibacterial [14], antibiofilm [15], antifungal [16],
antimalarial [17], antioxidant [18], anticancer [19], antiviral [20]), properties (luminescence [21],
fluorescence [22] and non-linear optical [23]) and potential applications (as in sensors [24] and
photovoltaic materials [25]).
Considering the palladium(II) complexes with Schiff bases as ligands, a plethora of
compounds has been reported; in many of them, the Schiff base ligand is a derivative of
salicylaldehyde and is a bidentate N,O-donor [26-29]. The Pd(II)-Schiff base complexes have been
examined for their catalytic activity [29-31] or have shown in vitro biological activity
(antimicrobial [26,32], antifungal [27] and cytotoxic [28,33,34]).
Serum albumins (SAs) are the most abundant plasma proteins in mammals, playing a key-
role in drug transportation. SAs are multifunctional proteins with extraordinary ligand-binding
capacity, making them ideal transporter molecules for a diverse range of metabolites, drugs,
nutrients, metals and other molecules. Due to their ligand-binding properties, SAs have wide
clinical, pharmaceutical, and biochemical applications [35,36], since a wide variety of drugs are
delivered to their targeting organs/tissues by binding with human serum albumin (HSA). Such
2

ACCEPTED MANUSCRIPT
studies are of great importance for pharmaceutical industry, not only because of the impact as an
efficient drug-delivery system, but also for the ability of the albumin to modify the pharmacokinetic
profile of drugs [37,38]. DNA is recognized as one of the common biological targets of anticancer
drugs, since one of the main mechanisms of action of the anticancer drugs is the damage of DNA
[39]. Bacterial DNA-gyrase is a topoisomerase type II enzyme that has attracted attention since its
discovery in 1976, when it was first isolated from Escherichia coli [40]. DNA-gyrase catalyzes
changes in DNA-topology by breaking and rejoining double stranded DNA, introducing negative
supercoils of the closed-circular DNA in front of the replication fork [41]. As this function is
essential for DNA-replication and transcription, DNA-gyrase is really a suitable target for
antibacterial agents, since its blocking induces bacterial cell death.
As a continuation of our research concerning palladium(II) metal complexes with
substituted salicylaldehydes (X-saloH, X = 3-OCH3,5-NO2,, 5-Cl or 5-Br) derivatives [42], we have
isolated and characterized a palladium(II) complex with a Schiff base ligand (4-chloro-2-(N-
ethyliminomethyl)-phenol, HL) derived from the condensation reaction of 5-chloro-salicylaldehyde
and ethylamine. The resulting complex has the formula [Pd(4-chloro-2-(ethyliminomethyl)-
phenolato-κ2N,O)2] ([Pd(L)2], 1) and its structure was verified by physicochemical and
spectroscopic techniques (IR, UV-vis) as well as single-crystal X-ray crystallography. The calf-
thymus (CT) DNA-binding mode of the complex was investigated by (i) UV-vis spectroscopy and
the DNA-binding constants of the complex (Kb) were determined, (ii) DNA-viscosity measurements
and (iii) competitive DNA-binding studies with the classic intercalator ethidium bromide (EB)
investigated by fluorescence emission spectroscopy. Moreover, in order to visualize a possible
interaction and the influence of complex 1 on the structure of plasmid pBluescript SK(+) (pDNA),
electrophoretic mobility experiments were performed. The affinity of the complex for HSA and
bovine serum albumin (BSA) was studied by fluorescence emission spectroscopy and the
corresponding binding constants were determined. The antimicrobial activity of complex 1 was
evaluated against five bacterial species: Escherichia coli XL1 (E. coli), Xanthomonas campestris
ATCC 33013 (X. campestris), Staphylococcus aureus ATCC 29213 (S. aureus), Bacillus subtilis
ATCC 6633 (B. subtilis) and Bacillus cereus ATCC 11778 (B. cereus). Furthermore, in silico
studies adopting various procedures with the employment of molecular docking simulations on the
crystal structure of CT DNA, DNA-gyrase and the target albumins HSA and BSA, were employed
with the aim to explore the ability of complex 1 to bind to these macromolecules, contributing thus
in the understanding of the role to act as antibacterial agent and suggesting a mode of action of
compound 1. Additionally, a variety of computational tools were employed to predict the complete
biological activity profile of the synthesized complex 1. Predictive tools include general
pharmacological potential (prediction of activity spectra for substances (PASS)), prediction of
3

ACCEPTED MANUSCRIPT
target proteins affected by the compound, prediction of drug-induced changes of gene expression
profile, protein and mRNA based prediction results, prediction substrate/metabolite specificity
(SMP) and sites of metabolism (SOMP), prediction of cytotoxicity for tumor and non-tumor cell
lines, rodent organ-specific carcinogenicity prediction (ROSC-Pred) and calculation of absorption,
distribution, metabolism, excretion, and toxicity (ADMET) parameters, pharmacokinetic properties,
druglike nature and medicinal chemistry friendliness of the compound. It is among the few cases
that so many extensive in vitro and in silico studies of the biological activity of a Pd(II) complex
with Schiff base ligands have been employed and reported.
2. Experimental
2.1. Materials - Instrumentation - Physical measurements
All chemical reagents used in the present study, i.e. 5-Cl-saloH, Pd(CH3COOH)2,
ethylamine (EtNH2), trisodium citrate, NaCl, CT DNA, EB, BSA and HSA were obtained as
reagent grade from Sigma-Aldrich Co. and used as received. All solvents were obtained from
Chemlab and were used without further purification.
The CT DNA stock solution was prepared by dilution of CT DNA to buffer (containing 150
mM NaCl and 15 mM trisodium citrate at pH 7.0) followed by exhaustive stirring at 4°C for 3 days,
and kept at 4°C for no longer than ten days. The DNA concentration per nucleotide was determined
by the UV absorbance at 260 nm after 1:20 dilution using ε = 6600 M-1cm-1 [43]. The stock solution
of CT DNA gave a ratio of UV absorbance at 260 and 280 nm (A260/A280) of 1.88, indicating that
the DNA was sufficiently free of protein contamination [44].
Infrared (IR) spectra (400-4000 cm-1) were recorded on a Nicolet FT-IR 6700 spectrometer
with samples prepared as KBr pellets. UV-visible (UV-vis) spectra of the compounds as nujol mulls
or in DMSO solution (in the range 1×10-5 - 5×10-3 M) were recorded on a Hitachi U-2001 dual
1
beam spectrophotometer. H-NMR spectra were recorded at 300 MHz on a Bruker AVANCE III
300 spectrometer using DMSO-d6 as solvent. Mass spectra were determined on an LC-MS 2010 EV
Instrument (Shimadzu) under electrospray ionization (ESI) conditions. C, H and N elemental
analyses were performed on a PerkinElmer 240B elemental microanalyzer. Molecular conductivity
measurements of 1 mM DMSO solution of the complexes were carried out with a Crison Basic 30
conductometer. Fluorescence spectra were recorded in solution on a Hitachi F-7000 fluorescence
spectrophotometer. Viscosity experiments were carried out using an ALPHA L Fungilab rotational
viscometer equipped with an 18 mL LCP spindle and the measurements were performed at 100
rpm.
2.2. Synthesis of the complex
4

ACCEPTED MANUSCRIPT
The Schiff base HL was prepared in situ by the addition of ethylamine (1 mmol, 45 mg) into
a CH3CN solution (15 mL) of 5-Cl-saloH (1 mmol, 156 mg) which was stirred at room temperature
for 30 min, A CH3CN solution (15 mL) of Pd(CH3COOH)2 (0.5 mmol, 112 mg) was added to this
solution and the resulting solution was stirred for additional 30 min and then turned into yellow.
Orange-yellow crystals of complex [Pd(L)2], 1 suitable for single-crystal X-ray structure
determination, (yield 42%, 198 mg) were formed after a few days, Anal. Calc. C18H18Cl2N2O2Pd
(MW = 471.6): C: 45.83, H: 3.85, N: 5.93 %; found: C: 45.90, H: 3.85, N: 5.91 %. ESI(+) MS in
MeCN (m/z): found: 473, calcd. for (1+H)+: 472.6; found: 186, calcd for HL+: 184; found: 156,
calcd for 5-Cl-saloH: 156, IR spectrum (KBr), selected peaks (v, in cm-1): 1662 (strong) v(C=N),
1308 (strong-to-medium) v(C-O→Pd), 507 (medium) v(Pd-O), 460 (medium) v(Pd-N). UV-vis
spectra: as nujol mull, λ/nm: 395, 450(shoulder (sh)); in DMSO, λ/nm (ε/M-1cm-1): 270 (18900),
397 (4900), 445(sh) (1500). 1H NMR spectrum in DMSO-d6 (δ/ppm) (the numbering of H atoms is
shown in Scheme 1): 8.06 (2H, s, H7), 7.43 (2H, d, J = 2.8 Hz, H6), 7.23 (2H, dd, J = 9.0, 2.8 Hz,
H4), 6.77 (1H, d, J = 9.0 Hz, H3), 3.83 (4H, q, J = 6.9 Hz, -CH2-), 1.25 (6H, t, J = 6.9 Hz, -CH3). ΛM
(1 mM DMSO) = 7 mho∙cm2∙mol-1.
2.3. X-ray crystal structure determination
Single-crystals of complex 1 suitable for X-ray crystallography structure analysis were
obtained by slow evaporation of the mother liquid at room temperature. A crystal of 1 was mounted
at room temperature on a Bruker Kappa APEX2 diffractometer equipped with a triumph
monochromator using Mo Kα radiation. Unit cell dimensions were determined and refined by using
the angular settings of at least 191 high intensity reflections (>10 σ(I)) in the range 11 < 2θ ω< 36o.
Intensity data were recorded using  and ω scans. Crystal presented no decay during the data
collection. The frames collected were integrated with the Bruker SAINT Software package [45],
using a narrow-frame algorithm. Data were corrected for absorption using the numerical method
(SADABS) based on crystal dimensions [46]. The structures was solved using the SUPERFLIP
package [47], incorporated in Crystals. Data refinement (full-matrix least-squares methods on F2)
and all subsequent calculations were carried out using the Crystals version 14.61_build_6236
program package [48]. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were
located by difference maps at their expected positions and refined using soft constraints. By the end
of the refinement, they were positioned geometrically using riding constraints to bonded atoms.
Illustration was drawn by CAMERON [49]. Crystallographic data for complex 1 are presented in
Table S1.
2.4 In vitro biological activity studies
5

ACCEPTED MANUSCRIPT
In order to study in vitro the biological activity of complex 1 (i.e., antimicrobial activity and
interaction with DNA or albumins), the complex is initially dissolved in DMSO (1 mM). Mixing of
such solutions with the aqueous buffer DNA solutions used never exceeded 5% DMSO (v/v) in the
final solution, which was needed due to low aqueous solubility of most compounds.
The interaction of the complex with CT DNA was investigated by UV-vis spectroscopy and
viscosity measurements and via the evaluation of its EB-displacing ability which was studied by
fluorescence emission spectroscopy. The interaction of the complex with pDNA was monitored by
agarose gel electrophoresis experiments. The albumin (BSA or HSA) binding studies were
performed by tryptophan fluorescence quenching experiments. The antimicrobial activity of the
complex was evaluated by determining the half-minimal inhibitory concentration (IC50) values
toward two Gram-(−) (E. coli and X. campestris) and three Gram-(+) (S. aureus, B. cereus and B.
subtilis) bacterial species. Detailed procedures regarding the in vitro study of the biological activity
of the complex are given in the Supporting Information (section S1).
2.5 In silico computational methods
A series of in silico calculations were employed in order to predict the biological activity of
the complex. These studies include: (i) molecular modeling and docking calculations on the crystal
structure of CT DNA dodecamer d(CpGpCpGpApApTpTpCpGpCpG), HSA, BSA and DNA-
gyrase, (ii) plasma protein binding (PPB) for target prediction in ChEMB database, (iii) PASS, (iv)
prediction of cytotoxicity for tumor and non-tumor cell lines via Cell Line Cytotoxicity Predictor
(CLC-Pred), (v) drug-induced gene expression profiles prediction (DIGEP-Pred), (vi) SMP, (vii)
SOMP, (viii) ROSC-Pred, (ix) prediction of toxicity of chemicals (ProTox-II) and oral rat toxicity
prediction, (x) SwissTargetPrediction and (xi) ADMET (SwissADME prediction and ADMET
structure-activity relationship server (admetSAR) prediction). Details concerning the in silico
biological properties of the complex are given in the Supporting Information (section S2).
3. Results and discussion
3.1. Synthesis
The Schiff base HL was prepared in situ via the reaction of 5-Cl-saloH with EtNH2 in
CH3CN. The addition of the Pd(II) salt in this solution in a 1:2 Pd:HL ratio resulted in the formation
of complex 1 (Scheme 1).
6

ACCEPTED MANUSCRIPT
Scheme 1. Synthetic route for complex 1. The numbering of H atoms in L-1 is also provided.
The obtained palladium(II) complex 1 is neutral (the molar conductivity in DMSO solution
was found 7 mho∙cm2∙mol-1) [50] and possesses an 1:2 Pd:L composition, as it has been indicated
from elemental analyses. The ESI(+) mass spectrum of the complex (Fig. S1) was recorded in
acetonitrile; the obtained molecular ion was found at m/z 473 which is satisfactorily close to the
calculated value of 472.6. The complex can be well-dissolved in DMSO and DMF, but it is less
soluble or insoluble in common organic solvents and water.
Evidence of the formation and the coordination mode of the ligand in complex 1 was also
1
derived from the interpretation of the IR, H NMR and UV-vis spectra, as well as by single-crystal
X-ray crystallography analysis.
3.2 Crystal structure of complex 1
The molecular structure of complex [Pd(L)2], 1 was verified by X-Ray crystallography and
is presented in Fig. 1.
Fig. 1. Molecular structures of complex 1 with the displacement ellipsoids shown at the 50%
probability level.
Complex 1 is centrosymmetric with the four-coordinate palladium atom Pd(1) on the
inversion center with a square planar geometry. The deprotonated L ligands are coordinated to the
palladium ion in a chelate bidentate mode through the carbonyl oxygen O(1) and the azomethine
nitrogen N(1) (O(1)-Pd(1)-N(1) = 92.05(14)°). The Pd(1)-O(1) bond distance (Pd(1)-O(1) =
7

ACCEPTED MANUSCRIPT
1.968(3) Å) is slightly shorter than the Pd(1)-N(1) bond distance (Pd(1)-N(1) = 1.996(4) Å). In
contrast, in complex [bis-(N-ethylsalicylaldimine)palladium] found in the literature, the Pd-O1
distance (1.94 Å) is much longer than the Pd-N1 distance (1.86 Å) [51].
A thorough search of the Cambridge Crystallographic Data Centre concerning the reported
crystal structures of metal complexes with 2-(ethyliminomethyl)-phenol and its derivatives as
ligands has revealed the following:
(a) regarding the nature of the metal, crystal structures of a variety of transition metal ions including
Ti(IV) [52], Mn(IV) [53,54], Co(II) [55,56], Co(III) [57], Ni(II) [58-62], Cu(II) [63-70], Zn(II)
[71,72], Zr(IV) [73,74], Ru(III) [75], Pd(II) [51], Au(III) [76,77] as well as Th(IV) [78] have been
reported.
(b) regarding the substitution of the aromic ring, metal complexes with alkyl (tBu- [74], methyl-
[62,71,75]), alkoxy (methoxy- [54]) and (di)halogen (chloro- [58,68], dichloro- [57,61], bromo-
[61] and dibromo- [55]) groups have been also structurally characterized.
(c) regarding the presence of halogen group(s) as substitute(s) in the aromatic ring of 2-
(ethyliminomethyl)-phenolate ligands, the crystal structures of Co(II) (for 3,5-dibromo-) [55],
Co(III) (for 3,5-dichloro-) [57], Ni(II) (for 5-chloro-, 5-bromo- and 3,5-dichloro-) [58,60,61], and
Cu(II) (for 5-chloro-) complexes exist; i.e. there are two more complexes, a mononuclear Ni(II)
[58] and a dinuclear Cu(II) [68] one, bearing 5-chloro-2-(ethyliminomethyl)-phenolate ligands as
complex 1.
(d) the structure of one more palladium complex with 2-(ethyliminomethyl)-phenolate ligands [51]
has been reported.
(e) three discrete coordination modes of these ligands have been observed (I-III as shown in fig. 2);
i.e. the monodentate mode where the coordination to the metal takes place via the phenolate oxygen
atom (mode I), the bidentate fashion when the ligand is bound to metal via the phenolate oxygen
and the imine nitrogen leading to the formation of a six-membered chelate (mode II) and the
tridentate mode upon coordination of the ligand via the imine N to one metal and via the phenolate
O to two metal ions forming a monoatomic O bridge (mode III). Mode III has been observed in the
case of dinuclear Cu(II) complexes [66-70], while mode II is the often and has been observed in all
other cases [51-65,71-78], and also in complex 1. Mode I has been observed only in the crystal
structure of a Ti(IV) complex [52] and only in co-existence with mode II.
8

ACCEPTED MANUSCRIPT
Fig. 2. Coordination modes of 2-(ethyliminomethyl)-phenol and its derivatives (R = tBu-, Me-,
MeO-, Cl-, Br-, Cl,Cl-, Br,Br-) as ligands. (I) Monodentate, (II) Bidentate, (III) Tridentate.
3.3 Spectroscopic characterization of the complex
Infrared spectroscopy has confirmed the deprotonation and the coordination of HL. In the IR
spectra of complex 1, no peaks attributed to the stretching and bending vibrations of the phenolic
OH around 3200 cm-1 and 1400 cm-1, respectively, were observed and my suggest the deprotonation
of the phenolic group of HL. In addition, the band originating from the C-O stretching vibration at
1285 cm-1, exhibits in the IR spectra of the complex a shift to 1308 cm-1 implying coordination
through the phenolate oxygen of the ligand [79,80]. The band assigned to the v(C=N) of the ligand
is located at 1662 cm-1, and may denote the coordination of the ligand to the metal through the
nitrogen.
1H-NMR spectroscopy was also used in order to confirm the formation and the
deprotonation of the Schiff base ligand (L), as well as the stability of the complex in solution. The
deprotonation of the phenolic hydrogen may be concluded by the absence of any -OH signal, which
1
was observed in the H-NMR spectra of the free 5-Cl-saloH as single peak at δ = 12 ppm. The
signals expected for the aromatic protons of L were found in the range 8.06-6.77 ppm, whereas the
signal attributable to the ethyl protons of L were observed at 3.83 ppm and 1.25 ppm for the -CH2
and the -CH3 groups, respectively (Fig. S2). The absence of any additional signals which might be
related to dissociated ligands may indicate that complex 1 remains intact in solution keeping its
structure.
In the UV-vis spectra of complex 1, two bands attributable to d-d transitions were observed;
the first one as a shoulder at 445 nm and the second one as distinct broad band at 397 nm.
According to literature, for four-coordinate palladium(II) complexes with a square planar geometry
around palladium, three d-d transition bands are expected in the regions 460-520 nm, 405-420 nm
and 320-380 nm [81,82] which may be attributed to transitions from ground state (1A1g) to the
excited states 1A2g,1B1g and 1E1g, respectively. In the case of complex 1, these bands are rather close
and seem to overlap; the band at 445 nm may be assigned to the first transition while the remaining
two transitions are ascribed to the second high-intensity band at 397 nm.
The UV-vis spectra of the complex were also recorded as nujol mull and in buffer solutions
used for the biological experiments and present similar pattern with those recorded in DMSO
1
solution. The combination of data derived from molar conductivity measurements, ESI-MS, H
NMR and UV-vis spectroscopy may leads us to conclude that the complex retains its structure in
solution within the timeframe used for the biological experiments.
9

ACCEPTED MANUSCRIPT
3.4. Interaction of complex 1 with CT DNA
The interaction of metal complexes with CT DNA usually takes place in a covalent and/or
non-covalent fashion. In covalent interactions, one or more labile ligand of the complex may be
displaced by a nitrogen atom of a DNA-base (i.e. guanine-N7) leading to the formation of novel
covalent bond between the metal and DNA-nitrogen atom. The non-covalent interactions usually
occur in the case of stable metal complexes and include three main modes, i.e. (i) intercalation
through the insertion of the complex in-between DNA-base pairs resulting in the appearance of π-π
stacking interactions, (ii) groove-binding to the major and/or minor groove mainly through van der
Waals interactions and hydrogen bonding and (iii) electrostatic interaction to the external phosphate
groups of the DNA-helix via Coulomb forces [83,84].
UV-vis spectroscopic titrations were employed as a preliminary approach to explore the
existence of any interaction between the compound and CT DNA and to calculate its binding
strength of this interaction. Any noncovalent interaction with DNA may perturb the band of CT
DNA or the intra-ligand bands of the complex during these titrations [84].
So the UV-vis spectra of a DNA solution were recorded in the presence of increasing
concentration of complex 1 (Fig. 3(A)). Addition of 1 into DNA solution resulted in slight
hyperchromism of the DNA band at 258 nm, which may be attributed to the formation of a new
adduct of 1 with double-helical DNA [85]. Additionally, a slight hyperchromism of the intraligand
band at 397 nm was observed in the UV-vis spectra of complex 1 (Fig. 3(B)) upon the addition of
increasing amounts of CT DNA. The changes observed in the UV spectra are not really intense and
the existing spectroscopic features may not reveal a definite conclusion regarding the DNA-binding
mode of 1 [84]; therefore, further DNA-viscosity measurements were conducted in order to shed
light to the DNA-binding mode.
2.0
(B)
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
2.0
1.5
1.0
0.5
0.0
(A)
240
260
280
300
320
300
350
400
450
500
 (nm)
 (nm)
Fig. 3. UV-vis spectra of a solution of: (A) CT DNA ([DNA] = 1.5×10-4 M) in buffer solution (150
mM NaCl and 15 mM trisodium citrate at pH 7.0) in the presence of increasing amounts of complex
1 (r = [complex]/[DNA] = 0-0.5). The arrow shows the changes upon increasing amounts of
complex 1. (B) complex [Pd(L)2] (10-4 M) in DMSO in the presence of increasing amounts of CT
10

ACCEPTED MANUSCRIPT
DNA (r’ = [DNA]/[complex] = 0-0.5). The arrow shows the changes upon increasing amounts of
CT DNA.
The Kb constant of complex 1 was obtained by the plot [DNA]/(εA-εf) versus [DNA] (Fig.
S3) and the Wolfe-Shimer equation (eq. S1) [86]. The Kb constant of complex 1 was found equal to
4.30(±0.74)104 M-1 (Table 1) and has the same magnitude with a series of reported Pd complexes
with derivatives of carboxamides or salicylaldehydes as ligands found in the literature [42,87,88].
Table 1. Binding, Stern-Volmer (KSV) and quenching constants (kq) regarding the interaction of
complex 1 with DNA, HSA and BSA and its competition with EB.
Constant
1
Kb(DNA) (M-1)
KSV(EB) (M-1)
kq(EB) (M-1s-1)
KSV(BSA) (M-1)
kq(BSA) (M-1s-1)
K(BSA) (M-1)
KSV(HSA) (M-1)
kq(HSA) (M-1s-1)
K(HSA) (M-1)
4.30(±0.74)104
5.87(±0.27)104
2.55(±0.12)1012
9.06(±0.40)104
9.06(±0.40)1012
2.43(±0.11)104
3.91(±0.17)104
3.91(±0.17)1012
3.21(±0.14)104
The relative DNA-viscosity (η/η0) is sensitive to changes of the relative DNA-length (L/L0)
upon the interaction of DNA with a compound, because of the equation (L/L0) = (η/η0)1/3 [89]. As
known, an intercalative interaction will result in an increase of the separation distance of base pairs
lying at intercalation sites in order to host the inserting compound, inducing an increase of the
length of the DNA-helix and, subsequently an increase of DNA viscosity will be observed, the
magnitude of which is usually in accordance to the strength of the interaction. In the case of a
partial or non-classic intercalation (i.e. electrostatic interaction or external groove-binding), a bend
or kink in the DNA-helix may occur, leading to a slight shortening of its effective length, and the
relative DNA-viscosity may present a less pronounced decrease or even remain unchanged [89].
These characteristics may be proved helpful in clarifying the interaction with DNA.
The viscosity measurements were carried out on CT DNA solutions (0.1 mM) at room
temperature upon addition of increasing amounts (up to the value of r = 0.35) of 1 (Fig. 4).
Therefore, the increase of relative DNA-viscosity shown in the presence of complex 1 may be
considered evidence of an intercalative interaction of complex 1 in-between the DNA-bases.
11

ACCEPTED MANUSCRIPT
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
r = [1]/[DNA]
Fig. 4. Relative viscosity (η/η0)1/3 of CT DNA (0.1 mM) in buffer solution (150 mM NaCl and 15
mM trisodium citrate at pH 7.0) in the presence of complex 1 at increasing amounts (r = 0-0.36).
A solution containing the EB-DNA conjugate shows an intense fluorescence emission band
at 592 nm (with λexcitation = 540 nm) as a result of the intercalation of the planar EB-phenanthridine
ring in-between adjacent DNA-base pairs; for this reason, EB is considered a typical indicator to
study the intercalation of a compound to DNA. If an intercalating compound which can bind to
DNA equally or stronger than EB is added into this solution, a significant quenching of the EB-
DNA fluorescence emission may be induced. The solution of complex 1 does not show any
fluorescence emission bands at room temperature either alone or in the presence of DNA or EB
under the same experimental conditions (λexcitation = 540 nm). Therefore, the changes in the
fluorescence emission spectra of the EB-DNA solution, when complex 1 is added, may be
informative regarding the EB-displacing ability of the complex [90].
The fluorescence emission spectra of pre-treated EB-DNA ([EB] = 20 µM, [DNA] = 30
µM) were recorded (Fig. 5) in the presence of increasing concentration of complex 1 (up to r =
0.46). The addition of the complex resulted in a moderate-to-significant quenching (up to 62.3% of
the initial EB-DNA fluorescence emission intensity) of the fluorescence emission band of the initial
DNA-EB compound at 592 nm, showing the potency of 1 to displace EB for the EB-DNA
conjugate. According to the Stern-Volmer plots (Fig. S4), the EB-DNA fluorescence emission
quenching is in good agreement (R = 0.99) with the linear Stern-Volmer equation (eq. S2); thus, the
observed quenching may be attributed to the displacement of EB from EB-DNA by each complex
and let us conclude indirectly an intercalative mode of interaction between complex 1 and CT DNA
[91].
12

ACCEPTED MANUSCRIPT
6000
5000
4000
3000
2000
1000
0
100
80
60
40
20
0
0.0
0.1
0.2
0.3
0.4
0.5
r= [complex]/[DNA[
560 580 600 620 640 660 680 700 720
 (nm)
Fig. 5. Fluorescence emission spectra (λexcitation = 540 nm) for EB-DNA ([EB] = 20 μM, [DNA] =
26 μM) in buffer solution in the absence and presence of increasing amounts of complex 1 (up to r
= 0.05). The arrow shows the changes of intensity upon increasing amounts of 1. Inset: Plot of EB
relative fluorescence emission intensity at λemission = 592 nm (I/Io, %) versus r (r =
[complex]/[DNA]) (up to 62.3 % of the initial EB-DNA fluorescence intensity for 1).
The calculated KSV constant of 1 is equal to 5.87(±0.27)104 M-1 and is relatively high
suggesting tight binding to DNA. The fluorescence lifetime of EB-DNA system is τo = 23 ns [92],
the EB-DNA quenching constant of complex 1 (calculated with eq. S3) is 2.55(±0.12)1012 M-1s-1.
The kq constant of 1 is significantly higher than 1010 M-1s-1 [91] and may suggest that the quenching
of the EB-DNA fluorescence takes place via a static quenching mechanism [93].
3.5 Interaction of complex 1 with plasmid DNA
The DNA interaction efficiency of a compound may be estimated after DNA treatment with
it by its effect on the pDNA bands mobility and integrity, as visualized in agarose gel
electrophoresis. A delayed mobility of pDNA (up shift) could be attributed to the binding of the
compound to pDNA, increasing consequently, its molecular weight. Concerning DNA topology, the
supercoiled conformation of pDNA (S) or covalently closed circular pDNA (ccc), due to its high
folding and coiling degree, migrates faster than the relaxed conformation (R) or open circular (oc).
At the side of the pDNA integrity, diminution or disappearance of the pDNA forms' amount
suggests pDNA degradation due to a strong interaction with the compound. At the side of the
pDNA integrity, diminution or disappearance of the pDNA forms’ amount suggests pDNA-
degradation due to a strong interaction with the compound.
13

ACCEPTED MANUSCRIPT
Fig. 6. Agarose (1% w/v) gel electrophoretic pattern of EB-stained mixture of supercoiled (S) and
relaxed (R) plasmid DNA [pBluescript SK(+)], after 30 min and 1 h duration of electrophoresis.
Each sample containing 5 μL (8 μg) of pDNA-incubated with 1 at 37 °C (A1): Lanes C1, C2:
Controls, pDNA with buffer A, pDNA with buffer A and DMSO, respectively. Lanes 1-8: pDNA
treated for 30 min (A1) and 1 h (A1΄) with 5 μM, 10 μM, 20 μM, 40 μM, 60 μM, 100 μM, 200 μM,
500 μM of 1, respectively. (A2) Lanes C1, C2: Controls, pDNA with buffer A, pDNA with buffer A
and DMSO, respectively. Lanes 1-6: pDNA treated with 0.5 mM, 1 mM, 2 mM, 3 mM, 4 mM, 5
mM of 1, respectively, for 1 h.
The effect of the pDNA treatment with complex 1 is depicted for 30 min (Fig. 6(A1)) and 1
h (Fig. 6(A1΄)) at concentrations 5-500 μM and 0.5-5 mM (Fig. 6(A2)). In both cases, a small
delayed electrophoretic mobility (an up-shift) of both pDNA forms (supercoiled and relaxed) was
observed, evidently, due to the pDNA binding with complex 1. It is strange that, at the
concentration of 60 μM, an obvious decrease of the relaxed (R) form occurred and a band is created
which proceeds from the supercoiled pDNA and migrate ahead of all the other forms of the plasmid
in a gel suggesting the formation of single-stranded closed-circles or a more compact form of
supercoiled pDNA, which exhibit faster electrophoretic mobility due to the interaction with
compound 1 (Fig. 6(A1 and A1΄), lane 5). Furthermore, when higher concentrations of the complex
were tested a delay of the electrophoretic mobility (up shift) of both pDNA forms occurs (Fig. 6,
A2) confirming the binding, while at concentration of 5 mM, the relaxed form is completely
degraded as deduced by its complete disappearance (Fig. 6(A2), lane 6).
14

ACCEPTED MANUSCRIPT
A similar strong up-shift of the pDNA bands was also observed with platinum complexes
[94,95] as well as with other metal complexes [96,97] and nanoparticles [98].
3.6. Interaction of complex 1 with serum albumins
Serum albumins are the major soluble protein constituents of the circulatory system. Among
their many physiological functions [99], the most important one is to transport drugs and other
bioactive small molecules and ions through the bloodstream [99,100]. BSA has been the most
extensively albumin studied in the literature, mainly because of its structural homology with HSA
[101]. BSA consists of three homologous domains (I, II, III) and has two tryptophans, Trp-134 and
Trp-212, located in the first sub-domain IB and sub-domain IIA, respectively. HSA is a globular
protein composed of 585 aminoacid residues in three homologous a-helices domains (I-III) bearing
only one tryptophan found at position 214 along the chain and in sub-domain IIA [101,102]. The
solutions of BSA and HSA exhibit an intense fluorescence emission band, due to the tryptophan
residues, at λem,max = 342 nm and 351 nm, respectively, when they are excited at 295 nm [103]. The
solution of complex 1 did not exhibit any appreciable emission band, under the same experimental
conditions. Furthermore, the inner-filter effect was evaluated with eq. S4; it was found too
negligible to affect the measurements [104].
The addition complex 1 to a SA solution induced a low-to-moderate quenching of the
respective SA fluorescence emission band (i.e. at λem = 351 nm for HSA and at λem = 342 nm for
BSA) which is in the case of BSA was more pronounced (Fig. 7). This SA-quenching may be
attributed to changes occurring in tryptophan environment of SA resulting from possible changes in
secondary SA-structure due to binding of compound 1 [105]. The kq constants for compound 1
(Table 1) were calculated from the Stern-Volmer plots (Figs. S5 and S6) and the Stern-Volmer
quenching equation (eq. S2 and S3); the values were found 9.06(±0.40)1012 M-1 for BSA and
3.91(±0.17)1012 M-1 for HSA. The kq constants of the complexes are higher than the value 1010 M-
1s-1 suggesting the existence of static quenching mechanism that may indirectly confirm the
interaction of complex 1 with the albumins [91].
15

ACCEPTED MANUSCRIPT
100
80
60
40
20
0
(A)
4000
3000
2000
1000
0
0
1
2
3
4
5
6
r = [complex]/[HSA]
320 340 360 380 400 420 440 460 480
 (nm)
100
80
60
40
20
0
(B)
5000
4000
3000
2000
1000
0
0
1
2
3
4
5
6
r = [complex]/[BSA]
320 340 360 380 400 420 440 460 480
 (nm)
Fig. 7. (A) Fluorescence emission spectra (λexcitation = 295 nm) for HSA ([HSA] = 3 μM) in buffer
solution (150 mM NaCl and 15 mM trisodium citrate at pH 7.0) in the absence and presence of
increasing amounts of complex 1. The arrow shows the changes of intensity upon increasing
amounts of 1. Inset: Plot of relative fluorescence emission intensity of HSA at λem = 351 nm (I/Io,
%) versus r (r = [complex]/[HSA]) for complex 1 (up to 60.9 % of the initial HSA fluorescence).
(B) Fluorescence emission spectra (λexcitation = 295 nm) for BSA ([BSA] = 3 μM) in buffer solution
(150 mM NaCl and 15 mM trisodium citrate at pH 7.0) in the absence and presence of increasing
amounts of complex 1. The arrow shows the changes of intensity upon increasing amounts of 1.
Inset: Plot of relative fluorescence emission intensity of BSA at λem = 342 nm (I/Io, %) versus r (r =
[complex]/[BSA]) for complex 1 (up to 37.7 % of the initial BSA fluorescence).
The SA-binding constants (K) of the compounds for both SAs were calculated from the
corresponding Scatchard plots (Figs. S7 and S8) and the Scatchard equation (eq. S5); the K values
of 1 were equal to 2.43(±0.11)104 M-1 for BSA and 3.21(±0.14)104 M-1 for HSA. The SA-binding
constants of complex 1 are of the same magnitude with those reported for diverse palladium
complexes with carboxamides or semicarbazones as ligands reported in the literature [42,87,88].
Furthermore, the magnitude of K constants may indicate the tight binding of complex 1 to the SAs
in order to get transported towards the possible biological targets. It should be also noted that, upon
16

ACCEPTED MANUSCRIPT
arrival at these biotargets, the compound may have the potential to get released, since its
noncovalent interaction with the SAs is less tighter than the strongest known noncovalent
(irreversible) binding attributed to the interaction of diverse ligands to the protein avidin with K
value ≈ 1015 M-1 [106,107].
3.7. Evaluation of antibacterial activity of complex 1
The results of the antibacterial efficacy of the Pd(II) complex 1 against the five bacterial
species (E. coli, X. campestris, S. aureus, B. subtilis and B. cereus) are presented in Fig. S9 and the
IC50 and Minimal Bactericidal Concentration (MBC) values are cited in Table 2.
Table 2. Antibacterial activity of complex 1 evaluated by the Half-Minimal Inhibitory
Concentration (IC50, in μg/mL) and Minimal Bactericidal Concentration (MBC, in μg/mL) values
provided by a nonlinear curve fit-growth/sigmoidal-dose response on the experimental optical
density data. Values are mean of two replicates.
1
1
Microbial strains
IC50 (μg/mL) MBC (μg/mL)
Gram negative
E. coli
Not detected
>100
Not detected
160
X. campestris
Gram positive
S. aureus
Not detected
>100
Not detected
130
B. cereus
B. subtilis
94
110
The Pd(II) complex 1 paradoxically at low concentrations seems to favor bacterial growth of
almost all the bacteria except E. coli that it is not at all infected, while shows bactericidal activity
only for B. subtilis from concentration higher than 100 μg/mL (Fig. S9). Although Pd(II) complexes
have shown noteworthy antibacterial activity, regarding other studies [94], in this study the newly
synthesized complex of Pd(II) seems to be a significant source of nutrients for the bacteria at the
low concentrations tested.
3.8. Molecular docking calculations
Molecular docking calculations were employed to evaluate the ability of complex 1 to bind
to macromolecules HSA, BSA, CT DNA and DNA-gyrase, in order to explain the in vitro activity
of this compound. Binding energies for the best docking poses of complex 1 with HSA, BSA, CT
DNA and DNA-gyrase are shown in Table 3. The binding interactions of complex 1 with HSA,
17

ACCEPTED MANUSCRIPT
BSA, CT DNA, and DNA-gyrase are illustrated in Tables S2-S5. The modular structural
organization of HSA and BSA provides a variety of ligand binding sites shown in Fig. S10. The
secondary structures of HSA and BSA shown with the sub-domains color-coded are assigned based
on Sugio et al [108] and Bujacz et al [109], respectively.
Table 3. Global binding energies (in kcal/mol) of complex 1 docked on HSA, BSA, CT DNA and
DNA gyrase targets (Protein Data Bank (PDB) accession numbers: 2BXG, 4OR0, 1BNA and
1KZN, respectively) along with the occupied drug binding sites (BSs) in Ligand-Binding Domains
(LBDs) 1 and 2 (*denote common binding site with co-crystallized drug naproxen (NPS)).
Complex 1
HSA
BSA
CT DNA DNA gyrase
Lowest energy
-42.22
-37.13
-28.94
-28.02
-42.78
-30.53
(most favorable binding pose)
Higher energy
LBD1 - BSV
-37.46
LBD1 - BSIII*
-36.83
(next favorable binding pose)
LBD2 - Between
BS’s I, II, IV
LBD2 - Between
BS’s I, II, IV
The computational approach revealed that complex 1 is anchored in a cavity of HSA protein
in sub-domain IIIB formed by helices IIIB-h10, IIIB-h9 and IIIB-h8 (Ligand-Binding Domain
(LBD) 1, LBD1), in binding site V (Fig. 8, upper part). Complex 1 can also be stabilized, less
favorably, in another binding pocket between binding sites I, II and IV (LBD2) formed by helices
IIIA-h3, IIIA-h4 and IIA-h1 (Fig. 8, lower part, the designation of sub-domains and helices are
shown in Fig. S10). Hydrogen-bond (H-b), polar (P) and hydrophobic (H-ph) contacts of the studied
complex with the target protein HSA are illustrated in Table S2. It is obvious (Table S2) that
complex 1 is predicted to bind to LBD1 of the protein with lower binding energy, compared to
LBD2, reasoning a better stabilization in the protein’s binding pocket, a fact that is in accordance
with the more extended hydrogen bond network. Two H-bond contacts between phenyl oxygen and
imino nitrogen atoms of Schiff base L of complex 1 and Ser579 residue, additionally contribute to
the stabilization of the molecule in protein’s binding pocket. Further stabilization of the docked
molecule inside most energetically favorable LBD1 cavity is achieved with the participation of π-π
interactions observed between aromatic ring of L and Phe507 and His535 residues and the
construction of a hydrophobic cavity, formed by residues Leu575, Val547, Phe502, Val576,
Phe507, and His535. Binding site LBD1 of complex 1 is adjacent to two sites in sub-domain IIIB,
where the thyroid hormone thyroxine also binds [110].
18

ACCEPTED MANUSCRIPT
Fig. 8. Docking poses of complex 1 superimposed with the co-crystallized drug ibuprofen (IBP)
into HSA (PDB ID 2BXG) (chain A). Target protein is illustrated as cartoon with sub-domains
color-coded according to chainbow, with additional depiction of semi-transparent surface colored
also by chainbow. Two individual poses for both complex 1 and IBP in the most common binding
sites: site V and between sites I, II and IV for complex 1, and site III and IV for IBP are depicted.
Complex 1 is represented as CPK spheres colored according to atom type (deep purple or violet C
atoms) and IBP molecules are depicted in stick model and colored according to atom type in hot
pink C atoms. Complex 1 in its lower energy binding pose (LBD1) (deep purple C atoms) is
anchored in binding site V, in the cavity formed by helices IIIB-h10, IIIB-h9 and IIIB-h8 (upper
part), while docking simulation resulting in next higher energy pose placed it in LBD2 (violet C
atoms), in a crevice formed by helices IIIA-h3, IIIA-h4 and IIA-h1, revealed by a rotated view of
the initial structure by 180 degrees around the main vertical axis of heart-shaped protein (the
designation of sub-domains and helices are shown in Fig. S9). Selected sidechain residues in the
binding network are shown as sticks color-coded by chain. Heteroatom color-code: Cl: green, N:
blue, O: red, Pd: petrol. Binding contacts are shown as yellow dotted lines. Molecular docking
simulations of both ligands were performed individually. H atoms are omitted for clarity. The final
structure was ray-traced and illustrated with the aid of PyMol Molecular Graphics System. Binding
interactions are shown in Table S2.
19

ACCEPTED MANUSCRIPT
The identified drug-binding sites of complex 1 in BSA protein, LBD1 and LBD2, are
illustrated in Fig. 9. Docked molecule is positioned in drug-binding site III (most favorable LBD1)
at the same place with the co-crystallized drug naproxen (NPS), in a cavity surrounded by helices
IIA-h2 and IIB-h9. The second most favorable binding site revealed to be LBD2, formed by helices
IIA-h1 and IIIA-h4 (the designation of sub-domains and helices are shown in Fig. S10). The
binding interactions of the docked molecule inside LBD1 and LBD2, involving hydrogen bond,
hydrophobic and polar contacts, are shown in Table S3. An extended network of hydrophobic and
hydrogen bond contacts in LBD1 explains the favorable accommodation of complex 1 inside this
pocket. π-π interactions observed between aromatic ring of Schiff base HL and His145 residue
additionally contribute to the stability of the molecular structure between complex 1 and BSA
protein. As obvious from Table S3, complex 1 and NPS share a number of amino acid residues
taking part in the anchorage of the molecules in LBD1 binding pocket.
20

ACCEPTED MANUSCRIPT
Fig. 9. Docking poses of complex 1 superimposed with the co-crystallized drug NPS into BSA
(PDB entry code 4OR0) (chain A). The best poses for complex 1 and NPS (in two binding pockets)
in the most common binding sites: site III designated as LBD1 between sub-domains IIA, IIB and
IIIA, and between sites I, II and IV (the designation of sub-domains and helices are shown in Fig.
S9) designated as LBD2 between sub-domains IB, IIA and IIIA. Complex 1 is docked in the LBD1
cavity surrounded by helices IIA-h2 and IIB-h9 (most favorable binding with lower binding
energy), and LBD2 cavity surrounded by helices IIA-h1, IIIA-h4 (second most favorable binding
with higher binding energy), while for NPS there are depicted three individual poses in the binding
sites III and IV. Complex 1 is illustrated in sphere representation, while NPS is depicted in stick
model. Both molecules are colored according to atom type: deep purple C atoms for complex 1 in
LBD1 and violet C atoms in LBD2, and hot pink C atoms for NPS. Target protein is illustrated as
cartoon with sub-domains color-coded according to chainbow, with additional depiction of
semitransparent surface colored also by chainbow. A close up view of the binding cavities of
complex 1 labeled as LBD1 and LBD2 is shown in lower part. Selected contacting amino acid
residues of the binding pocket are rendered in stick model and colored according to atom type
(LBD1) or chainbow (LBD2). Heteroatom color-code: Cl: green, N: blue, O: red, Pd: petrol.
Binding contacts are shown as yellow dotted lines. Molecular docking simulations of both ligands
were performed individually. H atoms are omitted for clarity. The final structure was ray-traced and
illustrated with the aid of PyMol Molecular Graphics System. Binding interactions are shown in
Table S3.
We chose to use for docking experiments the DNA-gyrase in complex with bound co-
crystallized drug chlorobiocin (CBN), which include only the B subunit exhibiting the crucial
ATPase activity (A subunit is mainly involved in DNA-breakage and reunion) [111,112]. The
binding of complex 1 in the crystal structure of DNA-gyrase, superimposed with the co-crystallized
inhibitor CBN ligand, is depicted in Fig. 10, where the best-fitted docking pose of complex 1 inside
the ATP-binding site of DNA-gyrase is shown. The compound is stabilized inside the same binding
pocket of the protein, occupied by CBN [109]. CBN is one of the most important members of the
coumarin family inhibitors of DNA-gyrase activity, by competitively binding to the ATP-binding
site. The binding pocket is defined by a cavity of approximately 20 Å inside the B subunit of the
protein which consists of two domains, an N-terminal domain of ~44 kDa and a C-terminal one of
47 kDa. The N-terminal domain includes two sub-domains (24 kDa C-terminal part and 20 kDa N-
terminal part). The ATP-binding site is located in the first sub-domain of the protein, i.e. in the C-
terminal part of the N-terminal domain. Complex 1 is shown to lie at the entry of the catalytic
pocket, partially covering the ATP-binding site. Stabilization of compound 1 may be attributed to
hydrogen-bond, polar and hydrophobic contacts inside the ATP-binding site of DNA-gyrase
protein. The docking predicts the formation of hydrogen bonds between the imino N atom of the
Schiff base ligand and H-bond donors: OH of Thr165 and HND2 of Asn46 (Table S4). Aromatic C
atoms of ligand L form additional hydrophobic contacts with Ile90, Ile78, Pro79 and Asp73
residues. A hydrophobic cavity is additionally constructed with the inclusion of residues Val120,
21

ACCEPTED MANUSCRIPT
Met91, Gly77 and Thr165. Complex 1 is further stabilized inside the protein’s binding pocket with
polar contacts with Asn46, Asp73, Val43, Thr165, and Gln72 residues. Common binding residues
with CBN revealed to be Asn46 (also taking part in hydrogen bond), Ile90, Pro79, Asp73, Gly77.
and Gln72, while the rest residues are sited in adjacent to those of CBN ligand [113,114].
Fig. 10. Docking pose orientation of complex 1 (deep purple C atoms rendered in sphere mode)
superimposed with CBN (hot pink C atoms rendered in stick representation) on DNA-gyrase (PDB
ID: 1kzn). The ligand-binding site of the molecules depicting the extent of the pocket as determined
by the computation process labeling the critical amino acid residues interacting with both molecules
is shown in the upper part. Heteroatom color-code: Cl: green, N: blue, O: red, Pd: petrol. Binding
contacts are shown as yellow dotted lines. Molecular docking simulations of both ligands were
22

ACCEPTED MANUSCRIPT
performed individually. H atoms are omitted for clarity. The final structure was ray-traced and
illustrated with the aid of PyMol Molecular Graphics System (upper part) and UCSF Chimera
visualization system (lower part). Binding interactions are shown in Table S4.
Fig. 11. Binding architecture of complex 1 in the crystal structure of (a) CT DNA (PDB accession
no. 1bna) (b) depicting its stabilization in the binding cavity of minor groove of DNA. (a) DNA
structure is illustrated with smooth ribbons colored in rainbow, stick bonds, and special
representations of the sugars and bases created with the nucleotides extension of UCSF Chimera
version 1.11, while complex 1 is rendered in ball-and-stick mode with deep purple C atoms. (b) In
the ligand-binding site, DNA structure is illustrated as cartoon color-coded according to chain in
brown color with the aid of PyMol Molecular Graphics System, while docked molecule is
represented in sphere model and colored according to atom type (deep purple C atoms). The
docking pose from a view above the axis of the helix to illustrate the extent of insertion of docked
molecule in the interior of double strand DNA is illustrated In lower part of (a). Binding
interactions with nucleotide molecules are shown in Table S4. Nucleotides are rendered in
semitransparent stick model and colored according to atom type. Yellow dotted lines indicate
hydrogen bond, polar and hydrophobic interactions between the docked molecule and the
nucleotides of the binding pocket in the minor groove of DNA. H atoms are omitted from all
molecules for clarity. Nucleotides are numbered according to PyMol software. The final structure
was ray-traced.
Our model for predicted binding poses of complex 1 into CT DNA, suggests intercalation of
the complex with A and B helices of DNA, between purines and pyrimidines of the same strand and
23

ACCEPTED MANUSCRIPT
between strands as well, in minor groove of DNA, due to the development of van der Waals forces,
hydrogen or hydrophobic bonds (Table S5 and Fig. 11). Interaction of Cl atom with N9 guanine
atom of DG22 in B-strand interrupts interstrand H-bond with its pair cytosine. The same is obvious
for contacts of Cl atom with N2 guanine atom of DG4 in A-strand (interrupts H-bond with carbonyl
O of its paired cytosine) and O2 cytosine atom of DC21 in B-strand (interrupts H-bond with NH2 of
its paired guanine). The docking orientation of complex 1 in the crystal structure of CT DNA in the
binding cavities of minor groove of DNA is shown in Fig. 11. Compound 1 seems to achieve its
stabilization inside minor groove of DNA due to a hydrogen bonds formed between phenyl O and
imino N of the Schiff base ligand and OH group of water96 and water78 molecules, respectively.
Ligand-binding site architecture of the docked molecule in CT DNA crystal structure is illustrated
in Fig. 11(b). Additional stabilization of complex 1 inside the minor groove of DNA is achieved by
the formation of π-π interactions between phenyl ring carbon atoms of the Schiff base ligand with
the aromatic rings of nucleotides DA6 and DG22 (Fig. 11 and Table S5). A construction of
hydrophobic pocket is mediated with the inclusion of DA6, DT7 and DG22 nucleotides. Compound
1 is also interacting with the minor groove of DNA with five polar contacts with the participation of
DG22, DC21 and DG4 nucleotides. Due to the bulk size of the complex, it enters the minor groove
by its half size, i.e. one of its HL Schiff bases as is illustrated in Fig. 11(a).
3.9. Polypharmacology Browser for Target Prediction in ChEMB
Furthermore, in silico predictive tools have been employed to identify new targets. The
ability of complex 1 to interact with more than one target proteins was explored with
Polypharmacology Browser. Target identification has become an area of intense research owing to
widespread application ranging from the identification of the biological active molecule, drug re-
purposing, toxicity prediction to the identification of promiscuous drugs or compounds (Poly-
Pharmacology). Interactions of the drug molecule with unintended side targets (off-targets) often
leads to adverse-drug reactions and shown to be one of the most common reason for the failure of
drug molecule in the clinical trial. The Poly-pharmacology profile of complex 1 identifying
probable target proteins is illustrated in Table S6. It is deduced (Table S6) that P = 0 for the
following ChEMBL-Names and IDs of targets retrieved from a variety of fingerprints: 1)
Thromboxane A2 receptor, TBXA2R (CHEMBL2069) in Topological Pharmacophore-Xfp, 2)
Fatty acid binding protein adipocyte, FABP4 (CHEMBL2083) in Topological Pharmacophore-Xfp,
3) Fatty acid binding protein muscle, FABP3 (CHEMBL3344) in Topological Pharmacophore-Xfp,
4) Prostaglandin E synthase 2, PTGES2 (CHEMBL4411) in Topological Pharmacophore-Xfp, 5) G
protein coupled receptor 44, PTGDR2 (CHEMBL5071) in Topological Pharmacophore-Xfp, 6)
Aldose reductase, AKR1B1 (CHEMBL1900) in Topological Pharmacophore-Xfp, 7) Sodium
24

ACCEPTED MANUSCRIPT
channel protein type IX alpha subunit, SCN9A (CHEMBL4296) in Atom-Pair fingerprint-APfp, 8)
Retinoid X receptor alpha, RXRA (CHEMBL2061) in Topological Pharmacophore-Xfp, 9) Aryl
hydrocarbon receptor, AHR (CHEMBL3201) in Extended-Connectivity Fingerprint-ECfp4, 10)
Cyclooxygenase-1, PTGS1 (CHEMBL221) in Topological Pharmacophore-Xfp, 11)
Cyclooxygenase-2, PTGS2 (CHEMBL230) in Topological Pharmacophore-Xfp, and 12)
Cytochrome P450 2A5, CYP2A5 (CHEMBL4085) in Extended-Connectivity Fingerprint-ECfp4.
3.10 PASS Biological Activity Prediction Profile
Prediction of activity spectra is based on PASS technology which can predict over 4,000
kinds of biological activity, including pharmacological effects, mechanisms of action, toxic and
adverse effects, interaction with metabolic enzymes and transporters, influence on gene expression,
etc.
3.10.1 Activity Spectra prediction
The results of PASS prediction are given as a list of biological activities, for which the
difference between probabilities to be active (Pa) and to be inactive (Pi) was calculated. The results
for activity spectra prediction for complex 1 with Pa > 0.7 are reported in Table S7. The output file
represents a list of predictable biological activities. It is interesting to note that the second place in
the ranking is occupied for antineoplastic activity with Pa = 0.851, maybe because of its probable
activity against a number of proteins involved in cancer growth.
3.10.2 CLC-Pred: in silico prediction of cytotoxicity for tumor and non-tumor cell lines
The in silico prediction of cytotoxic activity of complex 1 on a large number of cancer cell
lines is shown in Table S8. It is deduced that for probability level higher than 0.7, only one cancer
cell line, the brain derived oligodendroglioma Hs683, was found to be sensitive to complex 1.
3.10.3 DIGEP-Pred: Prediction of drug-induced changes of gene expression profile
The DIGEP-Pred of drug-induced changes of gene expression profile for complex 1 is
shown in Tables S9-S14. With this in silico tool it is possible to estimate the influence of complex 1
on gene expression, based on mRNA, protein and activity related to human cancer cell lines MCF-7
(breast) and VCaP (prostate) and the combination prediction results, respectively. The output file
represents a list of activities with two probabilities, Pa (probability to be active) and Pi (probability
to be inactive). The more probable changes of gene expression are at the top of the list. Only the
down-regulated and up-regulated target proteins predicted with a Pa > 0.5 are shown. From Table
S9 it is interesting to notice that the two first entries with the highest Pa values (0.937) in the
25

ACCEPTED MANUSCRIPT
mRNA-based training set prediction of 500 up-regulation genes are FABP4 (fatty acid binding
protein 4) with Pa = 0.748 and MYBL1 (MYB proto-oncogene like 1) with Pa = 0.726. The
prediction of FABP4 is in accordance with the previous PPB target prediction in ChEMB (section
3.9), where FABP4 protein is positioned in the 2nd place of probable target proteins.
For mRNA based predictions and Pa > 0.7, two down-regulated genes were identified:
MYBL1 (MYB proto-oncogene like 1) and FABP4 (fatty acid binding protein 4), while for
upregulated genes, REL (REL proto-oncogene, NF-kB subunit) and MGST1 (microsomal
glutathione S-transferase 1), genes found to play a role. For protein based predictions (Table S10)
and Pa > 0.7, it was identified one downregulated gene: VIM (vimentin, for up-regulated genes
there was no result with Pa > 0.7). For MCF7 based predictions and Pa > 0.7, two genes were
identified and only for the upregulation mode of action DDX60 (DExD/H-box helicase 60) and
IFIT2 (interferon induced protein with tetratricopeptide repeats 2). For VCAP cells in 6h, only one
down-regulated gene was found in probability level of Pa > 0.7: HP (haptoglobin). Vimentin is
found to be over-expressed in many tumor types [115]. Down-regulation of its protein could
contribute to anticancer activity of complex 1. Rel/NF-kappaB proto-oncogene transcription factors
are key regulators of immune, inflammatory and acute phase responses and are also implicated in
the control of cell proliferation and apoptosis [116]. This gene encodes a protein that belongs to the
Rel homology domain/immunoglobulin-like fold, plexin, transcription factor (RHD/IPT) family.
Members of this family regulate genes involved in apoptosis, inflammation, the immune response,
and oncogenic processes. This proto-oncogene plays a role in the survival and proliferation of B
lymphocytes. Mutation or amplification of this gene is associated with B-cell lymphomas, including
Hodgkin's lymphoma. Persistent nuclear NF-kappaB activity was also described in several human
cancer cell types, as a result of constitutive activation of upstream signaling kinases or mutations
inactivating inhibitory IkappaB subunits. Experiments implicating NF-kappaB in the control of the
apoptotic response also support a role in oncogenesis and in the resistance of tumor cells to
chemotherapy.
For MCF-7 human breast cancer cells based predictions (Table S11), it was identified as
downregulated the DEK gene which plays a role in steroid hormone receptor signaling in breast
cancer, supporting cell proliferation, invasion and the maintenance of the breast cancer stem cell
population [117]. DxEK oncogene expression is associated with positive hormone receptor status in
primary breast cancers and is up-regulated in vitro following exposure to the hormones estrogen,
progesterone, and androgen [118]. MCF-7 cells are known to be hormone positive (estrogen and
progesterone receptor positive, ER+, PR+) and DEK is a target gene whose expression promotes
estrogen-induced proliferation. Additionally, complex 1 from ADMET predicted profile-
Classifications (Table S19) found to act as an ER and androgen receptor (AR) binder, due to
26

ACCEPTED MANUSCRIPT
estimated probability levels of 0.7039 and 0.6812, respectively. Since complex 1 is able to
downregulate DEK gene in MCF-7 cells, this targeted inhibition of translated DEK protein may
enhance the efficacy of conventional hormone therapies in breast cancer. Furthermore, the also
downregulated gene ALDH1A3, encodes a metabolic target for cancer therapy and is associated
with the development, progression, and prognosis of cancers. [119]. ALDH1A3, as well as other
members of the ALDH1 subfamily, can function in cells as a retinaldehyde dehydrogenase to
generate retinoic acid (RA) from retinal. The enzymatic activity of ALDH1A3 and its product RA,
are necessary for the observed expression of tissue transglutaminase (tTG) in mesenchymal (MES)
glioma stem cells (GSCs). tTG, a dual-function enzyme with GTP-binding and acyltransferase
activities, has been implicated in the survival and chemotherapy resistance of aggressive cancer
cells and cancer stem cells (CSC), including glioma stem cells. Therefore, there is a link between
tTG and ALDH1A3 since its up-regulated expression in cancer stem cell populations. For this
reason, it may be therapeutically targeted for various types of glioma cells. This result of
downregulation of DEK gene by complex 1, can also explain, in part, the results from cancer cell
line prediction of cytotoxic activity of complex 1 (Table S8), revealing Hs 683 oligodendroglioma
cell line to be sensitive to the activity of complex 1.
3.10.4 SMP: Prediction of substrate/metabolite specificity (Pa > Pi)
In Table S15 are illustrated the substrate and metabolite based specificity predictions for
complex 1. The interactions with metabolic enzymes and transporters are shown with their
corresponding probability levels.
3.10.5 SOMP: Prediction of Sites of Metabolism
For the prediction of SOMPs for complex 1, sets for the five isoforms of CYP P450 that
metabolize the majority of xenobiotics have been prepared: 3A4, 2C9, 2C19, 2D6 and 1A2. The
reaction of glucoronidation was also included, which is catalyzed by UGT (Table S16). In Table
S16 the SoLAs are illustrated, in which labeled atom is a SOM with positive ΔP values. The labeled
atoms in SOM (ΔP+) revealed to be 11, 9, 11, 9, 9 and 2 atoms of complex 1 for CYP3A4,
CYP2D6, CYP2C9, CYP2C19, CYP1A2 and UGT, respectively.
3.10.6 ROSC-Pred: Rodent Organ-Specific Carcinogenicity Prediction (Pa>Pi)
No carcinogenicity was predicted for complex 1 and for probability level higher than 0.7 (Pa
> 0.7), except for female mice and specifically for the thyroid gland organ, as is illustrated in Table
S17.
27

ACCEPTED MANUSCRIPT
3.11 ProTox-II - Prediction of toxicity of chemicals
Oral toxicity prediction results
Predicted Lethal Dose value in mg/kg body weight (LD50) and toxicity class for oral toxicity
of complex 1 found to be 540mg/kg (LD50), and 4 (predicted toxicity class). The comparison of
complex 1 with dataset compounds (training set of 38,515 compounds) is illustrated in Figure S11,
leading to the conclusion that complex 1 could be harmful if swallowed (300 < LD50 ≤ 2000).
Toxicity risk assessment parameters and drug related properties were within the acceptable interval
which validate complex 1 as drug. The validation approach used is leave-one-out cross-validation
with balanced accuracy of 0.86. The descriptors that was adopted was fragments and molecular
fingerprints.
3.12 SwissTargetPrediction
In Table S18, are shown the target proteins with higher possibility for binding with complex
1, derived from Swiss Target Prediction. Two of the proteins depicted in Table S18 with higher
possibility to acta s targets for complex 1, are microtubule-associated protein tau (MAPT) and
Cannabinoid receptor 1 (CNR1).
MAPT: Structural and biochemical alterations of the MAPT are associated with degenerative
disorders referred to as tauopathies. We have previously shown that MAPT is present in human
islets of Langerhans, human insulinomas, and pancreatic beta-cell line models, with biophysical
similarities to the pathological MAPT in the brain. Furthermore, MAPT plays a role in pancreatic
endocrine tissue. It was found upregulation of MAPT protein expression in human insulinomas
when compared to human pancreatic islets of Langerhans and an imbalance between MAPT
isoforms in insulinomas tissue [120]. MAPT may play an important role in insulin granule
trafficking and indicate the importance of balanced MAPT phosphorylation and dephosphorylation
for adequate insulin release. Aggregates of abnormally hyperphosphorylated MAPT are the main
constituent of neurofibrillary tangles (NFTs) and neuropil threads (NTs) that are characteristic
neuropathological hallmark lesions of Alzheimer’s disease [121,122]. MAPT gene encodes the
MAPT protein whose transcript undergoes complex, regulated alternative splicing, giving rise to
several mRNA species. MAPT transcripts are differentially expressed in the nervous system,
depending on stage of neuronal maturation and neuron type. MAPT gene mutations have been
associated with several neurodegenerative disorders such as Alzheimer's disease, Pick's disease,
frontotemporal dementia, cortico-basal degeneration and progressive supranuclear palsy.
CNR1: Cannabinoids have been shown to activate an Emdoplasmic Reticulum ER-stress related
pathway that leads to the stimulation of autophagy-mediated cancer cell death. In addition,
cannabinoids inhibit tumor angiogenesis and decrease cancer cell migration [123]. CNR1 receptor
28

ACCEPTED MANUSCRIPT
was found to be upregulated in Hodgkin lymphoma cells [124] and in chemically induced cellular
hepatocarcinoma [125]. CNR1 receptor levels are also increased and correlate with disease severity
in human epithelial ovarian tumors [126] and have been proposed to be a factor of bad prognosis
following surgery in stage IV colorectal cancer [127]. Cannabinoid treatment promotes cancer cell
death, impair tumor angiogenesis and block invasion and metastasis [128]. The molecular
mechanisms that have been proposed to be involved in cannabinoid anticancer actions have been
thoroughly reviewed elsewhere [129,130].
3.13 ADME predictions
For the ADMET predictions, two methods were adopted, SwissADME and admetSAR.
3.13.1 SwissADME predictions
Physicochemical descriptors and predicted ADME parameters, pharmacokinetic properties,
drug-likeness nature and medicinal chemistry friendliness and bioavailability radar, are depicted in
Figure S12. From the predicted descriptors it is deduced that complex 1 bears a good lipophilicity
and human intestinal absorption (high). Furthermore, it is shown to be able to pass the blood brain
barrier (BBB), to inhibit various CYP metabolism enzymes, to act as a P-glycoprotein substrate,
and finally to possess drug likeness properties, since it is not violating none of the adopted rule tests
(Lipinski, Ghose, Veber, Egan and Muegge).
3.13.2 admetSAR predictions
In Table S19 is illustrated a series of ADMET properties of complex 1, leading to the
conclusion that this compound, with high probability, could be absorbed by the intestinal
(probability, 0.9898), enable Caco-2 cells permeability (probability, 0.8674), and pass the BBB
(probability, 0.9807). Additionally, ER- and AR-binding of the compound could be a factor
contributing to its possible anticancer activity. One undesired property of the compound is that it
appears to be eye corrosive and irritant with probability levels 0.9772 and 0.9580, respectively.
Carcinogenicity and Ames mutagenicity, could also play a limiting therapeutic role (probability
levels 0.7857 and 0.7000, respectively), although organ-specific carcinogenicity was not predicted
for probability level higher than 0.7. Nevertheless, Acute Oral Toxicity of 2.104 kg/mol, is a rather
high value, maybe due to limited human oral bioavailability (probability level of 0.5429). Plasma
protein binding (probability of 100 %) is in a good agreement with in vitro and in silico molecular
docking results.
4. Conclusions
29