Organoselenium functionalized oxacycles as ligands in some trans-palladium(II) complexes: biological evaluation and interaction with small biomolecules

Abstract In this article, the biological evaluation and interaction of three selenium-based trans-palladium complexes (Pd–Se1, Pd–Se2, and Pd–Se3) with small biomolecules were performed. The UV–Vis kinetic studies on substitution reaction of synthesized complexes with selected biologically important N- and S-bonding ligands (L-Cys, GSH, L-Met, 5′-GMP, L-His) has emphasized greater affinity of sulphur-based ligands for complex-binding, while the general order of the reactivity was L-Cys > GSH > L-Met > 5′-GMP > L-His. The complex reactivity toward small biomolecules is influenced by ligand flexibility, i.e. the steric hindrances of their corresponding ligands near the coordination site (Se–Pd) during the substitution process. The in vitro cytotoxicity of the investigated complexes against colorectal carcinoma HCT-116 and healthy fibroblast MRC-5 cells exhibited moderate prooxidative and significant cytotoxic character against HCT-116 cells, while such effect on MRC-5 was much weaker. Antioxidant activity (DPPH assay) indicated that all three complexes and their ligands have significant antioxidant activity, with Pd–Se2 being significantly stronger than its ligand L2 or the other complexes. Molecular docking simulations on tyrosinase (Tyr) singled out Pd–Se1 as the most compatible to bind to the cavity of Tyr. At all concentrations, the tested compounds demonstrated genotoxic effects (comet assay) compared to negative control.


Introduction
Selenium is an important micronutrient for bacteria, mammals, and birds [1].Organoselenium compounds have attracted significant attention as biologically active compounds due to their low toxicity and antioxidant properties [2][3][4].Selenium forms stable bonds with most elements, including C, N, O, and some metal elements.Organoseleno moieties form part of the active site of different selenoenzymes, such as GPx and thioredoxin reductase [5], protecting tissues and cell membranes against oxidative stress [6].This has prompted efforts toward the design and synthesis of different organoselenium compounds which can mimic the same activities as those expressed by natural enzymes.
Organoselenium compounds are regarded as promising pharmaceutical agents, such as antioxidants, enzyme inhibitors, anti-tumor, and anti-infection agents [7].Moreover, the introduction of selenium into biologically active organic scaffolds improves the physical and biochemical properties of the original compounds.This applies to transition metal complexes that incorporate organoselenium moieties as part of their spectator ligands.Because of this, the proper choice of ligands is critical to attain the desired properties such as the drug's pharmaco-kinetics and physiological distribution, physical interactions, chemical reactivity, and bio-activity.There has been tremendous development in the design of anticancer platinum [8,9] and palladium [10,11] complexes following the discovery of cisplatin (cis-Pt(NH 3 ) 2 Cl 2 ), the prototype drug in cancer chemotherapy.Although excellent results have been achieved in the treatment of some cancer cell lines with platinum complexes, chemotherapy patients endure increased nephrotoxicity, neurotoxicity, hair loss, drug resistance, or ototoxicity.Therefore, we design structurally related palladium complexes as alternatives that could overcome these problems.The kinetics and stability of palladium complexes are significantly different from those of platinum complexes, and the selection of the ligands (spectator or leaving group) is a crucial step for maintaining the structural integrity necessary for reaching molecular targets.The functional advantage of incorporating organoselenium moieties in the ligand framework of potential anticancer metallodrugs is motivated by the neutri-pharmaceutical significance of elemental Se.This has caused an increased interest for their coordination to transition metal ions, even though organosulphur ligands have been exploited more than related selenium ligands [12][13][14][15][16][17].
Studies have demonstrated that transition metal complexes bearing organoselenium compounds as ligands in their structure could be promising therapeutic targets.For example, some of the metal complexes with seleno-semicarbazones have been investigated as anticancer drugs [18], while organoselenium-based zinc complexes have shown excellent potential as antidiabetic [12], as well as antimicrobial and antioxidant agents [19].Tin complexes prepared from organoselenium compounds containing pyrazole or phenylthiazole groups have antiproliferative activity toward the mouse colon carcinoma C26 cell line, better than 5-fluorouracil [13].Inspired by the role of palladium and platinum complexes in chemotherapy and the interesting biological features that could be obtained by their combination with organoseleniumbased pharmacophores, we have synthesized and conducted several biological screening tests of the various palladium(II) and platinum(II) complexes bearing trans-oriented selenium-functionalized cyclic ethers as ligands [20][21][22][23].
These compounds have interesting pharmacological activities (such as antimicrobial, antioxidative, and cytotoxic) [20][21][22][23].Some of our platinum and palladium complexes have shown significant cytotoxic activity against human colorectal cancer cell line HCT-116, as well as moderate to strong binding affinity for calf thymus DNA and bovine serum [20,22,23].Moreover, these studies have shown a distinct connection between the ligands' structure (the presence of five-or six-membered cyclic ether) or its structural rigidity and exhibited biomedical potential, thus calling for further assays on the ligands' structure/activity relation.
In previous study [23] we have reported the synthesis, characterization, and biological screening of three trans-palladium(II) complexes, Pd-Se1, Pd-Se2, and Pd-Se3 (Figure 1), bearing organoselenium ligands.The research was motivated by previous studies where it was shown that palladium complexes with trans-geometry had higher cytotoxicity against cancer cells and lower cytotoxicity in healthy cells than their cisisomers [10,24].In order to synthesize high purity trans-palladium complexes where the cis/trans isomerism to form the more stable cis-isomer is prevented, bulkier ligands were used [10].Since it is known that lower kidney toxicity of palladium complexes in contrast to platinum is due to the inability of proteins in the kidney tubules to replace the tightly bound ligands of Pd(II) with thiol group, we designed palladium complexes which have selenium as a donor atom due to its preference toward palladium as a soft base.Also, complexes with higher lability tend to decompose before entering the cell and reaching the cellular target [25].
Knowing that both electronic properties and bulkiness of the ligand have a noticeable influence on the activity of the palladium(II) complexes, monodentate organoselenium ligands which differ structurally from each other in their heterocyclic moieties had a tetrahydrofuran ring (Pd-Se2), tetrahydropyran ring (Pd-Se3), as well as lactone ring (Pd-Se1).Also, tetrahydrofuran, tetrahydropyran, and lactone rings are entities present in many natural products with remarkable bioactivities [26], and therefore it was interesting to compare how the presence of different organoselenium-enriched saturated rings in the palladium-complex structure could affect the pharmacological profile.
To get more insight into the mode of interactions governing the metal-protein binding process, we measured rate constants for chloride substitution reactions from Pd-Se1-2 by small biologically important biomolecules (L-Cys, GSH, L-Met, L-His, and 5 0 -GMP).
Taking into consideration the prominent antimicrobial activity of these compounds against 17 strains of bacteria and fungi, we evaluated main redox status parameters of these compounds.As previous results on cell viability and cytotoxicity of synthesized ligands and complexes on human epithelial colorectal cancer cell line HCT-116 have shown cytotoxic effects [23], we decided to repeat the experiments and to investigate their effect on healthy MRC-5 cells and to establish if selectivity between cancer and healthy cell lines exists; previous studies with complexes of similar structure showed higher cytotoxicity toward healthy rather than cancer cell lines [22].Since the investigated complexes showed a strong binding affinity with serum albumin and DNA, we proceeded to investigate their genotoxicity.

Compounds
Pd-Se1, Pd-Se2, and Pd-Se3 were synthesized according to previously reported procedures [20,23].NMR spectra were recorded on a 200 MHz Varian Gemini-2000.Chemical shifts (d) are reported in ppm and referenced to tetramethylsilane.Elemental analyses (C, H, O) were performed by combustion and gas chromatographic analysis with an Elementar Vario MICRO elemental analyzer.Mass spectrometry was measured on a Waters Quadrupole-ToF Synapt 2 G using electrospray ionization.Corresponding data can be found in the Supporting Material.
L-Methionine (L-Met), L-histidine (L-His), guanosine-5 0 -monophosphate sodium salt (5 0 -GMP), L-cysteine (L-Cys), glutathione (GSH), and PdCl 2 were obtained from Acros Organics or Sigma Aldrich and were used without purification.Hepes buffer (N-2hydroxyethylpiperazine-N0-2-ethanesulfonic acid) was obtained from Sigma Aldrich.All other chemicals were of the highest purity commercially available and used without purification.Ultra-pure water was used in all experiments.Nucleophile stock solutions were prepared shortly before use.

Kinetic measurements
Kinetic measurements were conducted on an Applied Photophysics SX.18 MV stoppedflow instrument coupled to an online data acquisition system.The temperature was controlled throughout all kinetic experiments to ±0.1 K.The interaction of Pd-Se1 and Pd-Se2 with small biomolecules was followed as a substitution reaction on the Pd(II) center by purposely selected nucleophiles (L-Met, L-His, 5 0 -GMP, L-Cys, and GSH).The course of the reaction was monitored by the change in absorbance at suitable wavelengths as a function of time, where working wavelengths were determined by recording the spectral changes of the reaction mixture from 200 to 400 nm.The reaction was conducted under pseudo-first-order conditions and initiated by mixing equal volumes of complex and ligand solutions directly in the stopped-flow instrument.The reaction was followed for at least eight half-lives, and rate constants, k obsd , were calculated as an average value from three to five independent kinetic runs.

Cell culturing
In vitro evaluation of Pd-Se1-3 included using cell line model systems, human colorectal carcinoma HCT-116 and human healthy fibroblasts MRC-5, both purchased from ECACC (Lot Nr. 16I034 for HCT-116 and 18A018 for MRC-5).Cells were maintained in standard conditions (5% CO 2 at physiological temperature of 37 � C) in Dulbecco's Modified Eagle Medium (DMEM; Sigma, D5796) supplemented with 10% fetal bovine serum (Sigma, F4135-500ML) and 1% penicillin/streptomycin (Sigma, P4333-100ML) in culture flasks.All experiments were conducted with the same passage of the cells and started at cells confluence of about 90%.
Evaluation of the cell viability was conducted by the MTT assay according to standardized assay (Laboratory for Bioengineering protocol CB-005) in the concentration range of 0.1-500 mM for Pd-Se1-3.Cytotoxic effect was evaluated 24 and 72 h after treatment.More details are available in Nikolic et al. [27].

Redox status evaluation
Glutathione is a tripeptide that serves as a metabolic redox buffer with significant capacity (as our results confirm).Lewis acids, including medicinal drugs that enter a cell, interact with GSH.In most cases, the drugs induce formation of reactive oxygen/nitrogen species (ROS/RNS).Toxic ones often induce an increase in the content of free radical species in biological systems, which damage cellular compartments and key molecular pathways.In our study, we assess the ability of Pd-Se1-3 to scavenge nitric oxide (NO) and superoxide species, which are known to be induced by toxic substances and can cause damage to cellular compartments and molecular pathways.To evaluate this oxidative potential, we conducted incubation experiments using nitroblue tetrazolium (NBT) and Griess assays, which are commonly employed methods for measuring the content of superoxide anion radical and nitrite (an indicator of NO), respectively.In addition, we also investigated the impact of Pd-Se1-3 on the levels of GSH (glutathione), an important antioxidant that plays a crucial role in maintaining cellular redox balance.By examining the changes in GSH content, we aimed to determine whether the substances could modulate the cellular antioxidant defense system.Therefore, the experiments involved incubating the substances with appropriate reagents and substrates required for the NBT, Griess, and GSH assays.These assays allowed us to quantify the scavenging potential toward NO and superoxide species, as well as their influence on the cellular antioxidant capacity mediated by GSH.By employing these techniques, we were able to comprehensively evaluate the effects of the tested substances on the scavenging of ROS and RNS, as well as their impact on the cellular antioxidant defense system mediated by GSH.The protocols for superoxide anion radical determination by NBT assays (Laboratory for Bioengineering protocol CB-006), nitrites by Griess assay (CB-007), and GSH (CB-008) were conducted on both the HCT-116 and MRC-5 cell lines.The GSH assay relies on the oxidation of reduced glutathione using DTNB, resulting in formation of 5 0 -thio-2-nitrobenzoic acid (a yellow product) whose concentration was measured spectrophotometrically at 405 nm.For this assay, a total of 5 � 10 4 cells were initially distributed in 96-well plates and subjected to the same treatment protocol employed in the MTT, NBT, and Griess assays.Initially, the cells were disrupted using 2.5% ice-cold sulfosalicylic acid, and subsequently, 50 mL of the resulting supernatant was combined with 100 mL of a reaction mixture.This reaction mixture consisted of 100 mM phosphate-EDTA buffer with a pH of 7.4, as well as 1 mM DTNB.Following this, the absorbance was measured.The control samples consisted of untreated cells.All biochemical in vitro reactions were performed in 96-well plates following the colored reactions in concentration range from 0.1 to 500 mM for Pd-Se1-3.More details are provided in Petrovi� c et al. [28].Due to its simplicity, cytotoxicity is usually the first screening assay to be performed.However, cellular viability is a consequence of biochemical reactions and the cellular response to xenobiotic influences.The tested substance interacts first with the cell membrane and later with the cell redox equilibrium maintenance system.

Antioxidant activity-DPPH radical scavenging capacity assay
The ability of these ligands and complexes to scavenge 2,2-diphenyl-1-picrylhydrazyl (DPPH) free radicals was assessed using the method described by Takao et al. [29].
The methanolic DPPH solution (2 mL, 20 lg mL À 1 ) was added to the sample solutions (Pd-Se1-3 or free ligand) in methanol (2 mL) at various concentrations (31.25-500 lg mL À 1 ).After 30 min in darkness at room temperature, the absorbance was read on a spectrophotometer at 517 nm.Methanol was used as a reagent control and ascorbic acid as a positive control.The experiment was performed in triplicate and their means were used for analysis.
The inhibition percentage was calculated using the following equation: where Ac is the absorbance of the reagent control and As is the absorbance of the samples (ligands or complexes).
The EC 50 value (an effective concentration of the complex or free ligand at which 50% of DPPH radicals were scavenged) was obtained from the graph of scavenging activity (%) versus the concentration of the complexes or ligands.A low EC 50 value indicates strong ability of the complex to act as a DPPH scavenger.
The antioxidant activity, expressed as the antioxidant activity index (AAI), was calculated using the following equation: and was classified as follows: if AAI < 0.5 has poor antioxidant activity; AAI > 0.5-1 has moderate antioxidant activity; AAI > 1-2 has strong antioxidant activity; and AAI > 2 has very strong antioxidant activity [30].

Single-cell gel electrophoresis (comet assay)
The standard comet assay was performed as described by Singh et al. [31] with some modifications.Peripheral venous blood was obtained from three healthy donors aged 41, 38, and 37 years, and lymphocytes isolated using Histopague-1077.Lymphocyte suspension was treated with the trans-palladium(II) complexes at four different concentrations (1, 10, 20, and 40 lg mL À 1 ) for half an hour at 37 � C.After the incubation period, cell viability was determined using a Trypan blue assay.Treated cells were suspended in 100 lL of 1% low melting point agarose and spread onto the slide per two drops of 90 lL and kept at 4 � C for 5 min.Afterwards the slides were transferred into ice-cold prepared lysis solution for 2 h (2.5 M NaCl, 100 mM EDTA, 10 mM Tris, 1% triton X-100, 10% DMSO, pH 10) in the dark at 4 � C. The slides were then kept in an electrophoresis buffer (10 M NaOH, 200 mM EDTA, pH >13) at 4 � C for 30 min.
Electrophoresis was run at 25 V and 300 mA for 30 min at 4 � C in an ice bath, after which the slides were submerged in a neutralization buffer (0.4 M Tris-HCl, pH 7.5) three times for 5 min and stained with 50 lL ethidium bromide for 10 min.Negative control (without treatment) and positive controls with only hydrogen peroxide (the final concentration of H 2 O 2 was 10 mg mL À 1 ) were also simultaneously evaluated.Three hundred randomly selected cells from each slide (50 cells from each of the two replicate slides) were viewed under a Nikon E50i microscope at 400� magnification.The DNA damage was quantified by visual classification of cells into five comet classes: 0-no damage, class 1-low damage, class 2-medium damage, class 3-high damage, and class 4-total destruction.The genetic damage index (GDI) was calculated using the formula [32]:

Statistical analyses
The data were expressed as mean ± standard error (SE) as in vitro activity from three individual experiments, performed in duplicate for each tested dose.Statistical significance was determined using the Student's t-test or the one-way analysis of variance (ANOVA) test for multiple comparisons.The correlation between variables was done using the SPSS (Chicago, IL) statistical software package (SPSS for Windows, version17, 2008).The IC 50 values were calculated from the dose curves by a computer program (CalcuSyn).IC 50 values represent the half maximal inhibitory concentration of tested complexes that inhibits cell growth by 50%.Because investigated complexes greatly influence cell viability, all redox parameters were calculated in relation to the number of surviving cells.ANOVA with Tukey's post hoc test was used to compare differences among groups in the comet assay.The relationship between the tested concentrations of complexes and ligands and GDI values were determined by Pearson correlation coefficient.A difference at p < 0.05 was considered significant.

Molecular docking simulations
Geometric optimizations of Pd-Se1, Pd-Se2, and Pd-Se3 were performed using M062X hybrid meta function [33] in combination with 6-311þþg(d,p) basis set.Calculations were performed using Gaussian 09 software program [34].All geometries were optimized without symmetry constraints.Vibrational frequency calculations were performed to confirm the minima on the potential surface at the same theory level.These structures were docked into a rigid three-dimensional (3-D) structural representation of tyrosinase (Tyr; PDB: 6QXD).Structure of host molecule was obtained from the Protein Data Bank (PDB; http://www.rcsb.org)[35].All water molecules, ligands, and heteroatoms were removed.Grid resolution of the binding side was 0.3 Å.The parameters of docking procedure were: maximum number of iterations 1500, population size 50, energy threshold 100.00, and maximum number of steps 300.A maximum population of 100 and maximum number of iterations of 10,000 were used for each run.The MolDock SE search algorithm was used with the number of runs set to 100 and the number of generated poses was five.For docking simulations Molegro Virtual Docker (MVD, version 2013.6.0.1) was used and the estimation of investigated compounds and host interactions were described by the MVD-related scoring functions: MolDock, Docking, Rerank, and Hbond [36].Molegro scores were evaluated in a relative fashion.Docked poses with DNA fragments were visualized using CHIMERA (http://www.cgl.ucsf.edu/chimera/)molecular graphics program.

Kinetics and mechanism of interaction with small biomolecules
To gain better insight and more accurately predict the transformations the complexes could undergo through the interaction with biomolecules, detailed kinetic and mechanistic investigations were conducted by the use of stopped-flow technique.The rate of chloride substitution from Pd-Se1, Pd-Se2, and Pd-Se3 with small biomolecules L-Met, L-Cys, L-His, GSH (glutathione), and 5 0 -GMP (guanosine monophosphate) were measured under the physiological conditions, i.e. reactions were performed in 25 mM Hepes buffer (pH � 7.2) at 310 K.All reactions were carried out in a 30 mM NaCl solution to suppress the solvolytic pathway because it is well known that substitution reactions in square-planar complexes can occur via one of two competing pathways: (a) solvolytic, whose kinetics are independent of the type and concentration of entering nucleophile, or (b) direct displacement of the leaving group by nucleophile.In Scheme 1, the proposed mechanism of substitution is presented, where k 1 represents the rate constant of the direct and k À 1 of the reverse reaction.The reactions were studied under pseudo-first-order conditions where the concentrations of the ligands were at least 10 times higher than the concentration of the complex.
The pseudo-first-order rate constants (k obsd ) were determined with Eq. (1) by fitting each of the kinetic runs to a single-exponential function: where A 0 , A t , and A 1 represent the absorbance of the reaction mixture initially, at time t, and at the end of the reaction, respectively.
Scheme 1. Proposed mechanism of the substitution reactions of studied complexes.
The values of k obsd show linear dependence on nucleophile concentration, as shown in Figures 2 and 3. Equation (2) describes k obsd as a function of total ligand concentration and can be used to determine second-order rate constants (k 1 ) by linear regression of the k obsd values versus the nucleophile concentration, where k 1 is the second-order rate constant for the forward reaction and can be determined from the slope of the line, and k À 1 is the rate constant of the reverse reaction and represents the intercept.From Figures 2 and 3, it can be seen that all lines start from the origin of the graph; therefore the values of k -1 are insignificantly small.The obtained values of the rate constant k 1 are represented in Table 1.
Considering the fact that during the experiments on determination of working wavelengths, no additional processes were observed, and that all reactions were completed within 500 ms, we presume that substitution of chloride ions went simultaneously or that the substitution of first Cl À was too fast to be observed by this method and that only substitution of second Cl À was detected.Because S-donor ligands have greater affinity for binding Pd(II), their rate constants k 1 are from 10 2 to 10 3 , higher than for N-donor ligands.The following order of nucleophilic ligand reactivity was observed: L-Cys > GSH > L-Met � 5 0 -GMP > L-His.As per the order of their nucleophilicity and steric properties, thiols (L-Cys and GSH) have shown higher affinity toward Pd(II) complexes than thioether (L-Met).The greater values of k 1 for the 5 0 -GMP compared to the L-His can be attributed to the better nucleophilic properties of purine base than of the imidazole ring.
From Figure 3 and Table 1, the following order of reactivity can be observed: Pd-Se2 > Pd-Se3 > PdSe1, probably the result of greater flexibility (lower steric factors) of their corresponding ligands near the coordination site (Se-Pd) during the substitution process.The steric hindrance is most pronounced in the case of Pd-Se3 making its substitution reactions slower than that of Pd-Se2 as observed in the experimental trend.The rigidity of c-lactone moiety makes Pd-Se1 the least reactive.The presumption that steric factors of the organoselenium ligands have the greatest influence on the order of reactivity of examined complexes can be validated by comparing the obtained k 1 values for the reaction of thiol-nucleophiles.The bulky GSH ligand's rate constants are, on average, two times smaller than rate constants of less bulky L-Cys, although the nucleophilicity of both thiols is the same.This observation is the most pronounced in the substitution reaction of the most rigid complex Pd-Se1 (2.75 times smaller rate constant for GSH vs. L-Cys reaction).When comparing Pd-Se2 with corresponding trans-Pd(II) complex bearing unsubstituted THF ring (see Ref. [21]) the values of rate constant significantly differ only in the case with the slowest nucleophile L-His (k 1 ¼ (6.6 ± 0.1) � 10 3 M À 1 s À 1 ), while in all other cases is within the limits of error.
We used the Eyring equation to calculate activation parameters (enthalpy of activation (DH 6 ¼ ) and entropy of activation (DS 6 ¼ )) for further information on the mechanism of substitution reaction.The reaction of Pd-Se1 and Pd-Se2 with L-Cys were followed at three different temperatures (288, 298, and 310 K).On the base of the negative values of the DS 6 ¼ , it could be concluded that the substitution of halide ions by nucleophile follows the associative mechanism (Table S1 and Figure S4, Supporting Information).

In vitro evaluation
We have focused on in vitro cytotoxicity testing and determination of the main redox status parameters.The induced formation of key ROS-superoxide anion radical-and Table 1.The rate constants for the substitution reactions of Pd-Se1, Pd-Se2, and Pd-Se3 with L-Met, L-His, L-Cys, GSH, and 5 0 -GMP and at pH 7.2 in 25 mM Hepes buffer and presence of 30 mM NaCl at 310 K.
RNS-nitrites-were confirmed, as well as the key factor in maintaining redox balance, glutathione (GSH).As previous experiments have shown strong interactions of Pd-Se1, Pd-Se2, and Pd-Se3 (Figure 1) with bovine serum albumin and DNA [23], a genotoxic in vitro assay was conducted to test genetic damage that could be caused by interaction, directly or indirectly, with the tested compounds [31].

Cytotoxic activity
Results for cytotoxic activities of Pd-Se1-3 are presented in Figure 4 (IC 50 values) and in Figure 5 (viability curves).All three complexes decrease cell viability in a time and dose dependent manner.The cytotoxic effect is more significant 72 h from treatment than after 24 h.Also, the selectivity indices (SI) of Pd-Se1 and Pd-Se2 toward HCT-116 cancer cells over the healthy cells have been estimated.Results indicate very weak toxicity on healthy MRC-5 cells with the IC 50 values close to 400 mM.On the other hand, Pd-Se1 and Pd-Se2 induce significant cell toxicity on HCT-116 in the prolonged treatment of 72 h, with IC 50 values of 218.8 and 81.9 mM, respectively.The SI of the complexes for the cancer cell line CT-116 over the healthy cell line MCR5 were calculated by dividing the IC 50 values of two cell lines for the same treatment time point, i.e.IC 50 MRC-5 /IC 50 HCT-116 [37].The SI value for Pd-Se1 was 1.82, while that for Pd-Se2 was 4.99 after 72 h of treatment.The SI values indicate very good selectivity of the two trans-Pd(II) complexes toward cancer cells without affecting healthy cells.

Redox status evaluation
The first cellular response to xenobiotic intrusion or attack is to receive or donate an electron through the scavenging roles of the GSH/GSSG buffer pair.In the cell, the GSH/GSSG buffer ratio is significantly shifted toward the GSH content due to the more pronounced need to accept foreign electrons.The results shown in Figure 6 indicate a strong cellular response to the oxidative induction by Pd-Se1-3.GSH content does not change significantly over time and with increasing concentration (except at the highest dose).However, with Pd-Se1 and especially Pd-Se2, GSH capacity was not sufficient to prevent an increase in superoxide anion radical (Figure 7) and nitrites (Figure 8).The presented results indicate a slight increase in ROS/RNS in cell after their incubation with increasing concentration of Pd-Se1-2, especially for the 72 h of treatment.The origins of the free electrons, after cell treatment with cytotoxic complexes, is not surprising given the associated external Fenton reactions [38], which further allow for cascading radical reactions.Moreover, the possible reaction of the induced superoxide anion radicals and NO forms highly reactive peroxynitrites.

Antioxidant activity-DPPH radical scavenging activity
DPPH is a stable free radical often used for detection of radical scavenging (antioxidant) activity of chemical substances.DPPH radical neutralization capacity is expressed as EC 50 .Pd-Se1, Pd-Se2, and Pd-Se3 exhibited significant antioxidant activities, with that of Pd-Se2 being the best.According to AAI index, the highest scavenging activity was from Pd-Se2 and was half that of ascorbic acid, the positive control (Table 2).The antioxidant activities of L1, L2, and L3 were also examined.They show antioxidant activity that is not dose dependent, so the EC 50 could not be calculated.The dose at which they show approximately 50% of their activity ranged from 62.50 to 250 lg mL À 1 .While L1 and L3 showed similar antioxidant activity as their complexes, this was not the case for L2 where its complex had significantly stronger antioxidant activity.
The study by Tetteh et al. [39] showed that the antioxidant activity of the ligand after palladium complexation increases significantly, which is in agreement with our results.Tetteh et al. reported that coordination of Pd(II) to the ligand transfers electron density onto the metal center and as consequence, ligand hydrogens are easily lost in the presence of a free radical.The electron density drift is shown in 1 H-NMR downfield shifts of the proton signals upon complexation to palladium(II).This phenomenon, that coordination of metal ion to ligand facilitates the release of hydrogen to reduce DPPH radicals, is also reported by other authors [40,41].
Compounds that contain both selenium and palladium have multifunctional effects on cells.While selenium-containing compounds usually exhibit antioxidant properties, some compounds containing both palladium(II) and selenium are cytotoxic and prooxidative [42].In that study, the free ligand (without palladium(II)) also showed the same properties but only to a lesser extent [42].In contrast to these compounds, complexes Pd-Se1-3 are poorer cytotoxic agents, but they have good antioxidant activity.The basis of the discrepancy in cytoxicity can be found in different test methods and the existing possibility of different action of ligands in these complexes.

Molecular docking simulations
Antioxidant potentials of Pd-Se1, Pd-Se2, and Pd-Se3 were tested using molecular docking simulations on Tyr, which is a widely distributed copper-containing enzyme that catalyses two enzymatic reactions: the hydroxylation of monophenols to diphenols and the oxidation of diphenols to quinones [43].The key stable interactions predicted from docking the complexes onto Tyr are illustrated on Figure 9 and in Table 3.
According to the stability scores of the tested compounds, Pd-Se1 was predicted to form the most compatible noncovalent adduct within the cavity of Tyr.This can be understood by examining the structures of the complexes.L3 can be classified as the most sterically hindered to fit into the Tyr cavity, which is reflected in its lowest values of the docking scoring functions.

Single-cell gel electrophoresis (comet) assay
Ligands (L1, L2, and L3) and complexes (Pd-Se1, Pd-Se2, and Pd-Se3) at all tested concentrations demonstrated genotoxic effects on cultured human lymphocytes compared to the negative control.However, there is a significant statistical difference in  the concentration effect of ligands and their respective complexes (p < 0.05).Genotoxic effects that are dose dependent were observed for Pd-Se1 and L1, while this was not the case with Pd-Se2, L2, Pd-Se3, and L3 for concentrations above 1 lg mL À 1 .Pd-Se2, L2, Pd-Se3, and L3 have no statistically significant difference in genotoxicity at all tested concentrations.Pd-Se3 has the greatest genotoxic effect, and L1 has the least genotoxic effect (Table 4).

Conclusion
Given that interactions between metal ions and biomolecules have far-reaching biological consequences, the rates of chloride substitution from Pd-Se1-3 by biologically relevant small biomolecules (L-Cys, GSH, L-Met, 5 0 -GMP, and L-His) were examined.The results gave additional insight into the physiological action of Pd-Se1, Pd-Se2, and Pd-Se3.The reactivity was: L-Cys > GSH > L-Met > 5 0 -GMP > L-His which emphasized greater affinity of the complexes for the sulphur-based than the nitrogen-based ligands as per the soft acid-base theory as well as the order of nucleophilicity of the ligands.Pd-Se1 and Pd-Se2 had moderate prooxidative and strong cytotoxic effect on colon cancer HTC-116 cells, with prominent selectivity toward healthy fibroblasts MRC-5 cells.Pd-Se3 did not show any of these effects on either cell line.Pd-Se2 exerted very significant selectivity toward colon cancer cells.The three complexes showed significant antioxidant activity, with the best complex being Pd-Se2; its activity was significantly stronger than L2.These experimental results are confirmed by trends in the molecular docking simulations (stability scores) of the Tyr-Pd-Se1-3 adducts.Genotoxicity testing as a preliminary step in safety assessment for newly synthesized compounds showed that Pd-Se3 has the greatest genotoxic effect.

Figure 2 .
Figure 2. Pseudo-first-order rate constants k obsd as a function of nucleophile concentration for substitution reactions of Pd-Se2 with L-Met, L-His, L-Cys, GSH, and 5 0 -GMP at pH 7.2 and 310 K in 25 mM Hepes buffer and 30 mM NaCl.

Figure 3 .
Figure 3. Pseudo-first-order rate constants k obsd as a function of nucleophile concentration for chloride substitution reactions of Pd-Se1, Pd-Se2, and Pd-Se3 with L-Cys at pH 7.2 and 310 K in 25 mM Hepes buffer and 30 mM NaCl.

Figure 5 .
Figure 5. Cell viability after 24 and 72 h of exposure, expressed in percentages of viable cells.

Figure 9 .
Figure 9. Best poses with Tyr for Pd-Se1, Pd-Se2, and Pd-Se3 according to H bond scoring values: (A) molecular docking results illustrated regarding the Tyr backbone; (B) complex embedded inside the active site of Tyr in the electrostatic view; and (C) binding sites of the investigated compounds on Tyr and selected amino acid residues represented by stick models.Hydrogen bonds are shown in blue dotted lines.

Table 2 .
Radical neutralization capacity of Pd-Se1, Pd-Se2, and Pd-Se3 and positive control expressed as EC 50 and AAI index.

Table 3 .
Top-score values for

Pd-Se1, Pd-Se2, and Pd-Se3 with Tyr.
Best complex pose according to MolDock, Docking, and Rerank scoring functions.b Best complex pose according to Hbond scoring function. a

Table 4 .
Genotoxic effects of trans-palladium(II) complexes with organoselenium compounds as ligands on cultured human lymphocytes using comet assay.Values are significant different (statistically) in comparison to negative control cells (ANOVA, p < 0.05); values of cell classes are expressed as mean of three parallel measurements.