Heterobinuclear copper(II)‑platinum(II) complexes with oxindolimine ligands: Interactions with DNA, and inhibition of kinase and alkaline phosphatase proteins

Journal of Inorganic Biochemistry 203 (2020) 110863

Contents lists available at ScienceDirect

Journal of Inorganic Biochemistry journal homepage: www.elsevier.com/locate/jinorgbio

Heterobinuclear copper(II)‑platinum(II) complexes with oxindolimine ligands: Interactions with DNA, and inhibition of kinase and alkaline phosphatase proteins

T

Esther Escribano Arandaa, Juliana Silva da Luzb, Carla Columbano Oliveirab, ⁎ Philippe A. Divina Petersenc, Helena M. Petrillic, Ana M. da Costa Ferreiraa, a b c

Departamento de Química Fundamental, Instituto de Química, Universidade de São Paulo, São Paulo 05508-000, SP, Brazil Departamento de Bioquímica, Instituto de Química, Universidade de São Paulo, São Paulo 05508-000, SP, Brazil Departamento de Física dos Materiais e Mecânica, Instituto de Física, Universidade de São Paulo, 05508-090 São Paulo, SP, Brazil

ARTICLE INFO

ABSTRACT

Keywords: Heterobinuclear metal complexes Copper Platinum DNA binding Nuclease activity Kinase and alkaline phosphatase inhibition

Two mononuclear copper(II) compounds, [Cu(isad)(H2O)Cl]Cl 1 and [Cu(isah)(H2O)Cl]Cl 2, and its corresponding heterobinuclear species containing also platinum(II), [CuCl(isad)Pt(NH3)Cl2] 3 and [CuCl(isah)Pt (NH3)Cl2] 4 (where isad and isah are oxindolimine ligands, (E)-3-(2-(3-aminopropylamino)ethylimino)indolin-2one, and (E)-3-(3-amino-2-hydroxypropylimino)indolin-2-one, respectively), have been previously synthesized and characterized by different spectroscopic techniques in our laboratory. Cytotoxicity assays performed with B16F10 murine cancer cells, and MES-SA human uterine sarcoma cells, showed IC50 values lower or in the same order of cisplatin. Herein, in order to better elucidate their probable modes of action, possible interaction and damage to DNA, as well as their effect on the activity of crucial proteins were verified. Both mononuclear complexes and the binuclear compound 4 displayed a significant cleavage activity toward plasmid DNA, while compound 3 tends to protect DNA from oxidative damage, avoiding degradation. Complementary experiments indicated a significant inhibition activity toward cyclin-dependent kinase (CDK1/cyclinB) activity in the phosphorylation of histone H1, and only moderate inhibition concerning alkaline phosphatase. Results also revealed that the reactivity is reliant on the ligand structure and on the nature of the metal present, in a synergistic effect. Simulation studies complemented and supported our results, indicating different bindings of the binuclear compounds to DNA. Therefore, the verified cytotoxicity of these complexes comprises multiple modes of action, including modification of DNA conformation, scission of DNA strands by reactive oxygen species, and inhibition of selected proteins that are crucial to the cellular cycle.

1. Introduction Platinum compounds (cisplatin, carboplatin and oxaliplatin) are the only anticancer pharmaceuticals so far approved by the Food and Drugs Administration (FDA) in USA, despite the wide range of metallodrugs under investigation [1,2]. They have been intensively investigated, and applied in cancer treatment for several decades [3–5], although their use provokes various side effects, and often leads to drug resistance [6,7] Usually, DNA was identified as a main target for these Pt compounds [8], since they tightly bind to nucleic acids via covalent interactions, and cause reversible or irreversible structural modifications that can cause disruption of the transcription and/or replication, inducing cancer cells death. However, further studies demonstrated that

there are many others vital biomolecules capable of interacting with platinum compounds and that could possibly also contribute to its anticancer activity [9,10]. Because of its high toxicity and induced tumor resistance, different non‑platinum metal complexes have been developed as alternative anticancer metallodrugs, exhibiting varied modes of action [11,12]. Among these metals, copper is a promisor candidate, being an essential redox-active trace element, whose homeostasis is strongly controlled [13,14]. Also, it forms very stable complexes, and many of them have been already investigated as potential pharmacological agents with usually low toxicity [15,16]. In addition, compared to other essential metal ions, the affinity of copper ions to DNA appears to be remarkable [17], and in many cancer patients, concentration of copper in serum and tumor cells is higher than in healthy people [18].

⁎ Corresponding author at: Departamento de Química Fundamental, Instituto de Química, Universidade de São Paulo, Av. Prof. Lineu Prestes, 748 – 05508-000 São Paulo, SP, Brazil. E-mail address: [email protected] (A.M. da Costa Ferreira).

https://doi.org/10.1016/j.jinorgbio.2019.110863 Received 1 May 2019; Received in revised form 12 September 2019; Accepted 14 September 2019 Available online 28 October 2019 0162-0134/ © 2019 Elsevier Inc. All rights reserved.

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 1. Structures of the oxindolimine metal complexes under comparative studies with cisplatin.

ascorbate followed by re-oxidation with dioxygen, or in the presence of hydrogen peroxide [20,21]. In other studies, dysfunction of mitochondria is reported as the cause for apoptosis [22]. In our previous studies, some oxindolimine copper(II) complexes, as [Cu(isaepy)2]2+ and [Cu (isapn)]2+, where the ligands are isatin-derivatives (isaepy = (E)-3-(2(pyridin-2-yl)ethylimino)indolin-2-one, and isapn = (E)-3-(3-((E)-2oxoindolin-3-ylideneamino)propylimino)indolin-2-one, respectively, a dual role both on DNA and mitochondria was observed [23,24]. This type of ligand is able to carry the active metal across cellular membranes, inserting it into the cell, and triggering oxidative stress inside the cell. Behaving as delocalized lipophilic cation-like molecules, they target preferentially the organelles mitochondria and nuclei, in a process modulated by the imine ligand. Further, these copper(II) complexes were also able to target kinases [25] and topoisomerases, as already reported [26]. On the other hand, cisplatin-type complexes bind preferentially to guanosine bases of DNA, especially to GG or AG sites [27], and this preference could be used to drive a nuclease moiety to specific DNA sites. Some platinum‑copper complexes (named 3-Clip-Phen) have been reported exhibiting higher DNA cleavage efficiency than similar copper-only complexes [28,29]. In the present work, it is investigated the DNA binding ability of some oxindolimine‑copper(II) complexes in comparison to the corresponding bimetallic ones containing also platinum, and how this property can influence the nuclease activity of these complexes, as well as its ability of inhibiting particular proteins. Protein phosphorylation is a signaling event that plays a crucial role in biological functions and controls nearly every cellular process. It consists on the transfer of a phosphoryl group from a nucleoside triphosphate (usually ATP) to the hydroxyl group of an amino acid residue

Fig. 2. Curves of fluorescence decay in DNA-EtBr solution observed after addition of different concentrations of metal complexes: [CuL1] (1), [CuL2] (2), [CuL1Pt] (3), [CuL2Pt] (4) or cisplatin (5), after 24 h incubation at 37 °C.

Copper complexes and platinum drugs exhibit a quite different biodistribution and intracellular accumulation, and it seems a good strategy to have both metals in the same compound to verify its preferential binding sites, and eventually detect if they can overcome the tumor resistance to conventional platinum drugs [19]. Frequently, the cytotoxicity of copper complexes is credited to the generation of reactive oxygen species (ROS), induced by reduction with glutathione or 2

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 3. Cleavage patterns of the agarose gel electrophoresis and the corresponding cleavage extent (%) for pBluescript II plasmid DNA (240 ng) by compound [CuL1] 1 (1.5 mM) and H2O2 (150 μM) in buffer solution (50 mM Tris-HCl/50 mM NaCl, pH 7.4) at 37 °C, after 30 or 60 min incubation. Lane 1, MW; lane 2, DNA control; lane 3, DNA + H2O2 (150 μM) 30 min; lane 4, DNA + H2O2 (150 μM) 60 min; lane 5–7, DNA + H2O2 (150 μM) + 1 at 0.5, 2, and 6 μM, at 30 min; and lane 8–10, DNA + H2O2 (150 μM) + 1 at 0.5, 2, and 6 μM, at 60 min.

in a protein substrate, catalyzed by kinases [30]. These processes depend on the highly regulated and opposing actions of kinases (PKs) and phosphatases (PPs), through changes in the phosphorylation of key proteins. Histone phosphorylation, along with methylation, ubiquitination, and acetylation, also regulates access to DNA through chromatin reorganization. Phosphatases, on the other hand, are enzymes that remove a phosphate group from its substrate by hydrolyzing phosphoric acid monoesters into a phosphate ion, and a molecule with a free hydroxyl group, acting in opposition to kinases [31]. Being central protagonists in cells, these proteins are promising therapeutic targets in cancer and other diseases. Many inhibitors of kinases, especially cyclindependent proteins (CDKs), have been developed [32] since Gleevec was clinically approved for the treatment of chronic myelogenous leukemia [33]. In this work, some possible mechanisms of action of two mononuclear oxindolimine‑copper(II) compounds (1 and 2) and corresponding heterobinuclear Cu(II)-Pt(II) compounds (3 and 4) with the same ligands (Fig. 1), against vital biomolecules (DNA and selected proteins) were investigated, in order to provide a better understanding of their cytotoxicity versus tumor cells. Compounds 1 and 2 differ in their ligands, while compounds 3 and 4, differ additionally in the

length of the linker between metals, which could be determinant in driving them to the place of action and in the type of interaction established. 2. Experimental 2.1. Materials Most of the reagents, as hydrogen peroxide, ethidium bromide (EtBr), CT-DNA, N-(2-hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid) (HEPES), Tris-acetate, MgCl2, EDTA and PEI cellulose for TLC, as well as solvents (ethanol, DMSO, DMF) were purchased from different sources, Sigma-Aldrich (St. Louis, MI, USA), Merck (Darmstadt, Germany), or Amresco (Solon, OH, USA). CDK1 protein, cyclin B and human recombinant Histone H1 were purchased from BioLabs (Ipswich, MA, USA), and alkaline phosphatase protein (EF0651) was acquired from Thermo Fisher Scientific (Waltham, MA, USA). Deionized water from a Milli-Q (Millipore, Bedford, MA, USA) was used throughout for the preparation of all solutions. Metal complexes 1, 2, 3 and 4 (shown in Fig. 1) were previously synthesized and characterized by several methods as described 3

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 4. Cleavage patterns of the agarose gel electrophoresis and the corresponding cleavage extent (%) for pBluescript II plasmid DNA (240 ng) by compound [CuL2] 2 (1.5 mM) and H2O2 (150 μM), in buffer solution (50 mM Tris-HCl/50 mM NaCl, pH 7.4) at 37 °C, after 30 or 60 min incubation. Lane 1, MW; lane 2, DNA control; lane 3, DNA + H2O2 (150 μM) at 30 min.; lane 4, DNA + H2O2 (150 μM) at 60 min.; lane 5–7, DNA + H2O2 (150 μM) + 2 at 0.5, 2 and 6 μM, at 30 min; and lane 8–10, DNA + H2O2 (150 μM) + 2 at 0.5, 2 and 6 μM, at 60 min.

elsewhere [34]. Complexes 1 and 2 contain only copper(II), coordinated by an oxindolimine ligand, isad = (E)-3-(2-(3-aminopropylamino)ethylimino)indolin-2-one, or isah = (E)-3-(3-amino-2-hydroxypropylimino)indolin-2-one respectively, both in the keto form. Complexes 3 and 4 contain additionally a platinum(II) ion coordinated to the terminal amine group of the corresponding oxindolimine ligands in the enol form, (E)-3-(2-(3-aminopropylamino)ethylimino)-3H-indol2-ol (isad) or (E)-3-(3-amino-2-hydroxypropylimino)-3H-indol-2-ol (isah). They are designated as [CuL1], [CuL2], [CuL1Pt] and [CuL2Pt] respectively, along the text. All of them are stable in aqueous or DMSO/ buffer solutions for more than 24 h.

2.3. DNA damage monitored by gel electrophoresis The plasmid pBluescript II (Stratagene, San Diego, USA) was purified using Qiagen plasmid purification kit (Qiagen, Hilden, Germany). Reaction mixtures (20 μL total volume) containing 240 ng of supercoiled DNA (Form I) in 50 mM phosphate buffer (pH 7.4), in the presence or absence of different concentrations of H2O2 (150 μM), and varying concentrations of compounds 1–4 (0.5 to 30 μM), were incubated at 37 °C, for different periods of time. After incubation, a quench buffer solution (5 μL) was added, and the final solution was subjected to electrophoresis on an 1% agarose gel, in 1× TAE buffer (40 mM Tris-acetate, 1 mM EDTA) at 100 V, for 1 h. Analogous experiments were carried out in the presence of distamycin, a known minor group DNA binder.

2.2. Fluorescence measurements of DNA interactions DNA interaction assays were performed by fluorescence measurements, in the presence of different concentrations of each complex under study. A 50 μM solution of Calf-Thymus DNA (CT-DNA) in a TE (10 mM Tris pH 8.0, 1 mM EDTA) buffer solution was prepared. To samples of 3.0 mL of this DNA solution, 30 μL of a 5 mM solution of ethidium bromide (EtBr) were added. The resulting mixtures were incubated at 37 °C for 24 h. Next, different amounts of the complex solutions (containing no more than 2% DMSO) were added to 3 mL aliquots of DNA-EtBr solution, giving final concentrations of 10, 20, 30, 40 and 50 μM in metal complex, respectively. The emission spectra were recorded in a SPEX-fluorolog 2 spectrofluorimeter, in the range 530–670 nm, with excitation at 502 nm. All experiments were made in triplicate.

2.4. Computational studies Molecular docking was used to investigate the possible sites at which [Cu(isad)Pt(NH3)]3+ 3 and [Cu(isah)Pt(NH3)]3+ 4 complexes interact with DNA supercoiled. The AutoDock 4.0 [35] was used to perform the docking and the graphical user interface AutoDockTools (ADT 1.5.6) was employed to setup the DNA and build the gridbox. In the case of DNA, only the polarized hydrogens were considered and Gasteiger charge distributions were calculated. The 3D structure of DNA was obtained from the Protein Data Bank (entry 1ZYG). [Cu(isad) Pt(NH3)]3+ 3 and [Cu(isah)Pt(NH3)]3+ 4 complexes optimized geometrical structures were obtained using the Gaussian09 code [36] with Hartree-Fock HF and CEP-31G basis set [37]. The choice of this 4

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 5. Cleavage patterns of the agarose gel electrophoresis and the corresponding cleavage extent (%) for pBluescript II plasmid DNA (240 ng) by [CuL1Pt], compound 3 (1.5 mM) and H2O2 (150 μM), in buffer solution (50 mM Tris-HCl/50 mM NaCl, pH 7.4) at 37 °C, after 30 or 60 min incubation. Lane 1, MW; lane 2, DNA control; lane 3, DNA + H2O2 (150 μM) 30 min.; lane 4, DNA + H2O2 (150 μM) 60 min.; lane 5–7, DNA + H2O2 (150 μM) + 3 at 0.5, 2 and 6 μM, at 30 min; and lane 8–10, DNA + H2O2 (150 μM) + 3 at 0.5, 2 and 6 μM, at 60 min.

computational level was based on reference [35]. Complexes partial charge calculations were performed using Charges from Electrostatic Potentials and a Grid-based method (CHELPG) implemented in the Gaussian09 package [38]. All hydrogens were included for the complexes and ADT parameters of no bonding hydrogens were considered for nonpolar hydrogen atoms. Since the AutoDock does not provide the parameters for metal ions we set: the charges obtained with the CHELPG calculations for both metal ions, the initial parameters with van der Waals radius 2.14 Å [39] for copper(II) and 2.15 Å for platinum (II), van der Waals well depth 0.005 kcal/mol for copper(II) and 0.080 kcal/mol for platinum(II). The complexes and DNA were maintained frozen with no torsional energy loss during the molecular docking simulations. The ADT program was used to generate the docking input file and the grid box with size set to 100 × 126 × 100 grid points in x, y, and z directions, with 0.364 Å grid spacing. The grid box was centered in the center of the DNA structure (obtained with AutoDockTools). The Lamarckian algorithm was used to calculate the most favorable free energy of [Cu(isad)Pt(NH3)]3+ 3 and [Cu(isah)Pt (NH3)]3+ 4 docking positions. The initial population of 250 individuals, a maximum number of 2.5 × 106 energy evaluations, a maximum number of 27,000 generations and 9 best results of the most favorable binding free energy were chosen for the docking results of [Cu(isad)Pt (NH3)]3+ 3 complex structure, and 2 best results for [Cu(isah)Pt (NH3)]3+ 4.

5 min. Samples were then subjected to sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). The gel was revealed, after an exposure time of 30 min, in a Typhoom TRIO apparatus (Variable Mode Imager, from GE Healthcare). The software used to treat the obtained image was Image avant TL. 2.6. Inhibition of alkaline phosphatase protein The test was performed with high concentrations of the metal compounds (1 mM stock solution, 50 μM in each well) to ensure there is free compound able to interact with the phosphatase. The compounds were allowed to react with the kinase for 45 min, inhibiting histone phosphorylation. Then, 1 μL of alkaline phosphatase (EF0651, Thermo Fisher Scientific, Waltham, USA), was added to the mixture and the reaction was allowed to proceed for 15 min. To follow and evaluate the dephosphorylating, a control with 32P-ATP was also performed. Then, the reaction was stopped by denaturing the proteins at 95 °C for 5 min. After that, samples run in a cellulose TLC plate for about 2 h, in 0.8 M LiCl +0.5 M acetic acid solution. The TLC plate was revealed, after an exposure time of 30 min, in the apparatus Typhoom TRIO (Variable Mode Imager) GE Healthcare, by using the software Image avant TL to treat the images. The images obtained allowed the detection of the free phosphates released in the reaction. A vertical electrophoresis was also performed on SDS-PAGE gel with the highest concentration of the compounds, to be able to display the 32P-Histone. Tris-glycine buffer was used to run the gel, and the system was left for circa 2 h at 150 V.

2.5. Inhibition of CDK1 kinase/cyclin B protein The study of cyclin-dependent kinase protein inhibition (CDK1/cyclin B) was performed by agarose gel electrophoresis. For this study, 0.1 and 1 mM stock compound concentrations and a commercial 10 X CDK1 kinase buffer, 32P-ATP and histone H1 were used. In each sample, 18 μL of the mix, 1 μL of the solution of the compounds and 1 μL of water were added to each Eppendorf tube. The samples were then incubated for 45 min at 37 °C. After this time, the protein was denatured at 95 °C for

3. Results and discussion A set of mononuclear copper(II) complexes and bimetallic Cu(II)-Pt (II) complexes with Schiff base ligands have been prepared and characterized in our laboratory, as reported in previous work [34]. The ligands are derivatives of isatin, an endogeneous oxindole, formed as a metabolite of amino acids. These complexes can induce apoptosis and 5

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 6. Cleavage patterns of the agarose gel electrophoresis and the corresponding cleavage extent (%) for pBluescript II plasmid DNA (240 ng) by [CuL2Pt], compound 4 (1.5 mM) and H2O2 (150 μM), in buffer solution (50 mM Tris-HCl/50 mM NaCl, pH 7.4) at 37 °C, for 30 and 60 min incubation. Lane 1, MW; lane 2, DNA control; lane 3, DNA + H2O2 (150 μM) 30 min; lane 4, DNA + H2O2 (150 μM) 60 min; lane 5–7, DNA + H2O2 (150 μM) + 4 at 5, 15 and 30 μM, at 30 min; and lane 8–10, DNA + H2O2 (150 μM) + 4 at 5, 15 and 30 μM, 60 min.

already showed promising values of IC50 against B16F10 murine melanoma cancer cells, and MES-SA human uterine sarcoma cells. In order to provide complementary information about their possible anticancer activity, experiments by fluorescence, and gel electrophoresis were performed, after incubation with CT-DNA or pBS plasmid DNA. Experiments with kinase and phosphatase proteins were also carried out to better elucidate their modes of action.

DNA. The increase of compound concentration is responsible for a gradual decrease observed in fluorescence intensity. This could be attributed to the DNA distortion once platinum is bound covalently to the bases, after chlorides dissociation by aquation. Such distortion modifies the secondary structure of DNA and force the release of EtBr. In the case of bimetallic compounds, 3 and 4, a decrease in EtBrDNA fluorescence intensity was verified after increases of the compound concentration. This effect was much stronger in the case of compound 3, which was very similar to that observed for cisplatin. Therefore, the coordination of these compounds to the double helix seems to occur via the Pt center preferentially, and is responsible for its distortion (See Section 3.3. Molecular Docking analysis)

3.1. Fluorescence measurements Copper compounds, [CuL1] 1 and [CuL2] 2, do not cause a remarkable change in the relative fluorescence intensity of the DNA-EtBr system, as seen in Fig. 2. Even increasing the concentration of compounds, no appreciable decrease was observed. This fact could be due to possible hydrophobic interactions or hydrogen bonding occurring between compounds and the double helix. On the contrary, cisplatin is able to displace efficiently EtBr from

3.2. Gel electrophoresis studies Changes in DNA topology, monitored by unwinding of closed-circular plasmid DNA in electrophoretic gel-mobility shift assays, are

6

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 7. Changes in patterns of the agarose gel electrophoresis of pBluescript II plasmid DNA (240 ng) with addition of [cisplatin], compound 5, in buffer solution (50 mM Tris-HCl/50 mM NaCl, pH 7.4) at 37 °C, after 30 or 60 min incubation. Lane 1, DNA control; lane 2–3, DNA + cisplatin (0.5 μM); lane 4–5, DNA + cisplatin (2.0 μM); lane 6–7, DNA + cisplatin (6.0 μM).

Fig. 8. 2D and 3D structures for (A) [Cu(isad)Pt(NH3)]3+ 3 and (B) [Cu(isah)Pt(NH3)]3+ 4 complexes theoretically obtained (see text). Atom colors representation: Cooper (orange), Oxygen (red), Carbon (grey), Nitrogen (blue), Hydrogen (white) and Platinum (Dark blue). The green lines exhibit some bond distances. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

7

Journal of Inorganic Biochemistry 203 (2020) 110863

100 – –

– – 100

3.2.1. Studies of the DNA damage caused by the compounds For compound 1, it was observed an evident effect at higher concentration and longer incubation time, with a strong increase in the intensity of the linear DNA (Form III). Thus, at 6 μM and 60 min incubation, a significant amount of DNA form III was verified (see Fig. 3). Compound 2 was even more active compared to compound 1 at the concentrations and incubation times tested. DNA form I gradually disappears as the compound 2 concentration and the incubation time increase. As shown in Fig. 4, compound 2 can convert DNA form I into form II (85.7%) and form III (6.5%) at 0.5 μM, after 30 min of incubation (lane 4). It can completely cleave DNA form I into form II (72.2%) and form III DNA (25.0%) after 60 min of incubation (lane 7). At 60 min, and 6 μM of compound 2 no form I is observed, increasing the proportion of form II and form III (linear). The binuclear compound 3 was the most active in the series of tested compounds against B16F10 murine melanoma cells, showing the lowest IC50 value [34]. As shown in Fig. 5, the weak band corresponding to the DNA form III that appears with H2O2 (lanes 2 and 3) disappears completely at increasing concentration of the compound 3, while the intensity of the band corresponding to DNA form I increases. The effect is more obvious at longer incubation time. In contrast to compounds 1 and 2, compound 3 stabilizes the supercoiled form of plasmid DNA (form I), as can be seen in Fig. 5. Also, form III which appears in the presence of H2O2 is barely formed in the presence of 0.5 μM compound 3 concentration (14.3%), and not visible at 6 μM, both at 30 and 60 min. These results also confirm those of fluorescence decay. Interestingly, this remarkable relative higher levels of supercoiled DNA (Form I) is very similar to results obtained with cisplatin. For compound 4, the effect is the opposite. As shown in Fig. 6, DNA does not agglomerate, but suffers significant cleavage after treatment with this compound. It was reported as the most active in the series versus MES-SA sarcoma human cells [34]. Linear DNA formation becomes evident at 5 μM after 30 min incubation (29.8%), increasing slightly at increasing compound concentration. At higher concentration, 30 μM, and after 60 min incubation, the highest value of linear DNA is obtained (31.3%). Furthermore, the greatest amount of DNA form II is obtained with 30 μM, both at 30 min (40.4%) or 60 min (41.0%) of incubation. The different reactivity between compounds 3 and 4 toward DNA is probably due to their structural differences. Compound 3 has a longer linker between the two metal centers that confers better mobility to the system. This flexible structure could facilitate the winding around the DNA and would lead to an agglomeration of the biomolecule, similar to what was observed with cisplatin (see Fig. 7). Moreover, in compound 4 the linker containing an amine group and a hydroxyl group that bind the Pt and Cu centers, respectively, form a shorter and more rigid bimetallic structure. This structural feature of complex 4 influences the type of interaction it has with the biomolecule.

100 – – Phosphate backbone Nucleobases (DNA matrix) Other regions

52.8 3.0 44.2

[Cu(isah)Pt(NH3)]3+ Cu interactions (%) [Cu(isah)Pt(NH3)]3+ Pt interactions (%) [Cu(isad)Pt(NH3)]3+ Pt interactions (%)

[Cu(isad)Pt(NH3)]3+ Cu interactions (%)

commonly used to study the binding or disturbance that small-molecules cause to DNA [40]. The rate of migration of closed circular DNA in an agarose gel depends on its topological state. The supercoiled form (form I) shows a higher mobility than the completely uncoiled relaxed form (form II), or the cleaved linear form (form III). In these studies, DNA cleavage activity of synthesized compounds was monitored using pBS plasmid DNA in the presence of H2O2 by agarose gel electrophoresis in TAE buffer.

Active sites at DNA

Table 1 Percentage of binding preferences of platinum and copper for the most populated docking poses (250 total poses) obtained for [Cu(isad)Pt(NH3)]3+ 3 and [Cu(isah)Pt(NH3)]3+ 4 complexes in active sites of DNA.

E.E. Aranda, et al.

3.2.2. DNA damage in the presence of distamycin In order to verify mechanistic aspects of the DNA oxidative cleavage verified in the presence of the studied complexes, the effect of the minor groove binder distamycin on this cleavage was investigated. Distamycin is known to bind side-by-side to the minor groove of duplex DNA, although it can also interact with quadruplex DNA arrangements [41], and the structural modifications verified in both cases are analogous 8

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 9. Most populated clusters representative structures obtained by molecular docking simulations: A) [Cu(isad)Pt(NH3)]3+ 3, with Pt binding to the DNA phosphate backbone, and Cu coordinating to the backbone; B) [Cu(isad)Pt(NH3)]3+ 3, with Pt binding to the DNA phosphate backbone, and Cu coordinating to the cytosine basis; (C) [Cu(isah)Pt(NH3)]3+ 4,with only Pt binding to the DNA phosphate backbone.

[42]. The assays were performed by binding distamycin to plasmidial DNA before the treatment with the studied complexes and analyzing its possible inhibition effect on DNA cleavage. No significant effect of distamycin was observed in the verified DNA cleavage in the presence of the metal complexes (data not shown). This fact reveals that the compounds under study do not show high affinity for DNA minor groove, thus corroborating the information obtained through DNA interaction studies with ethidium bromide.

the frequency of each docking poses, instead of the energy values provided by the scoring function, are discussed. The results of docking simulations for both complexes, also identified the binding preferences of platinum and copper related to DNA sites. In Table 1 the binding preferences of Pt and Cu in [Cu(isad)Pt(NH3)]3+ 3 and [Cu(isah)Pt (NH3)]3+ 4 obtained for the most populated docking poses in phosphate backbone and nucleobases of the DNA matrix sites are presented. In both binuclear complexes, the platinum center interacts with the phosphate backbone site in all analyzed poses; no interaction with the nucleobases is observed. Besides that, no interaction of copper with DNA is observed in the case of complex [Cu(isah)Pt(NH3)]3+ 4. On the other hand, for complex [Cu(isad)Pt(NH3)]3+ 3, copper interactions with the DNA phosphate backbone are observed in 52.8% of poses, and with the DNA nucleobases in 3% of the poses. These interactions between the complexes and active regions of DNA are in the vast majority close contacts. For these close contacts and hydrogen bond interactions, 1.0 Å of van der Waals radius was considered. The most frequent poses configurations of the docking for both complexes are shown in Fig. 9.

3.3. Molecular docking analysis The [Cu(isad)Pt(NH3)]3+ 3 and [Cu(isah)Pt(NH3)]3+ 4 geometries of minimum energy obtained in the calculations are presented in Fig. 8. The bond distances obtained for [Cu(isad)Pt(NH3)]3+ 3 are: 2.178 Å for NePt, 2.188 Å and 2.144 Å for NeCu and 2.804 Å for OeCu. For [Cu (isah)Pt(NH3)]3+ 4, the bond distances are: 2.187 Å and 2.167 Å for NePt and 2.012 Å for OeCu. Molecular docking results performed on DNA identified two main regions for complex 3-DNA and complex 4-DNA interactions: the phosphate backbone and nucleobases in direct binding interactions. The energy analysis obtained by free energy scoring function of the complexes are similar, varying from −6.45 kcal/mol to −6.31 kcal/ mol for [Cu(isad)Pt(NH3)]3+ 3 clusters and from −7.86 kcal/mol to −7.80 kcal/mol for [Cu(isah)Pt(NH3)]3+ 4 clusters. Since the variation of the energy clusters for both complexes are very small, the analyses of

3.4. Inhibition of CDK1/cyclin B kinase protein The purpose of this experiment was to verify if the compounds under study could act as inhibitors of CDK1 protein, activated by cyclin B, changing the conformation of the active site, and thereby preventing phosphorylation of histone H1. The assay was performed at two

9

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

numbers), confirming that histone H1 was being dephosphorylated by the phosphatase. Interestingly, however, alkaline phosphatase was not inhibited by the copper compounds, as was CDK1. To further analyze the possible effects of the compounds on alkaline phosphatase, aliquots from the same reactions were subjected to thin layer chromatography (TLC) to separate 32P-ATP from released 32Pi. The results show that upon addition of alkaline phosphatase to the reactions (even numbers), 32P-ATP is completely hydrolyzed to ADP and 32Pi (the only radioactive product, and, therefore the only detected in this assay), confirming that the copper compounds do not inhibit this phosphatase. As a control of position for the TLC experiment, 32P-ATP was also loaded on the membrane. Quantification of the signals obtained by TLC (see Fig. 12) show that in the presence of the compounds, there is more free 32P-ATP due to their inhibition of CDK1, which phosphorylated H1 less efficiently. However, 32Pi levels are not much affected, showing that the compounds do not strongly inhibit alkaline phosphatase. 4. Conclusions Based on the results, addition of copper(II) compounds, [Cu(isad) (H2O)Cl]Cl 1 and [Cu(isah)(H2O)Cl]Cl 2, do not cause a noteworthy change in the relative fluorescence intensity of the DNA-EtBr system, showing that intercalation in DNA structure is not a predominant event. In the case of bimetallic compounds, [Cu(isad)Pt(NH3)]3+ 3 and [Cu (isah)Pt(NH3)]3+ 4, a more significant decrease in DNA-EtBr fluorescence intensity was observed. So, the binding of these compounds to the double helix occurs via the platinum center and is responsible for the DNA structure distortion. Studies of the DNA damage monitored by agarose gel electrophoresis reveal that compound 2 is the most active in the series, showing more oxidative cleavage capacity in the presence of H2O2. This is not expected by considering that the reduction potential of copper in all complexes (1 to 4) are very similar, around −0.86 V.[34] On the other hand, compound 3 with the ligand containing the longest linker, is the only one that does not show nuclease activity, being unable of degrading DNA Form I. In this case, the complex is undoubtedly bound to DNA by the platinum moiety. The binding of the studied complexes to DNA was tested in relation to the minor groove binder distamycin. The results showed that these compounds do not show affinity for DNA minor grooves higher than distamycin, thus corroborating the information obtained through DNA interaction studies with ethidium bromide. In the literature, the influence of bridging ligands in determining DNA-binding and cross-linking abilities has been verified for two binuclear platinum(II) complexes, containing different linkers: methylenedianiline (1) or diamino-p-xylene (2) [43]. Formation of 1,3-intrastrand cross-links was observed with compound (1), while for (2) predominant 1,4-intrastrand cross-links were formed. Results showed that the linker properties play a critical role in controlling both capabilities and in modulating the cytotoxicity of such compounds. In our studies, different linkers between the copper and the platinum centers lead to different reactivity toward DNA. The predominant damage to DNA occurred via oxidative cleavage, caused by ROS generated at the copper center, in compounds 1, 2 and 4. However, only for complex [CuL1Pt] 3 a remarkable binding to DNA was observed in DNA-ethidium bromide assays, in a similar action to cisplatin (see Fig. 2). Both copper-only complexes 1 and 2 exhibited better reactivity, in the nuclease action as well as in the inhibition of the proteins, in comparison to the binuclear complexes containing also platinum. The cytotoxicity of these compounds toward B16F10 murine melanomas and human sarcoma uterine cancer cells (standard MES-SA and resistant MES-SA/Dox5) has been reported previously, and indicated

Fig. 10. Gel electrophoresis assay that shows the ability of compounds to inhibit kinase protein. Lane 1: mix + H2O, control; lane 2: mix + DMSO; lane 3–4: mix + [CuL1] 1, 5 or 50 μM; lane 5–6: [CuL2] 2, 5 or 50 μM; lane 7–8: [CuL1Pt] 3; lane 9–10: [CuL2Pt] 4; mix = buffer, kinase, cyclin-CDK1, 32PATP and Histone H1.

different concentrations of compounds (5 μM and 50 μM), and it was also considered if the proportion of DMSO (up to 1%) used to dissolve the compounds could affect the experiment. As shown in Fig. 10, DMSO does not significantly affect the histone phosphorylation (control). Looking at the quantification results, resumed in Table S1 (Supplementary data), compounds 1 and 2 show similar inhibitory effects. It is also observed that mononuclear copper (II) compounds inhibit more efficiently CDK1, decreasing significantly the extent of histone phosphorylation, than the corresponding binuclear species containing also platinum. 3.5. Inhibition of alkaline phosphatase protein With the aim of discerning if the tested compounds exhibit selectivity for kinases, an assay with the alkaline phosphatase enzyme was also performed. In this assay, CDK1 was incubated with histone H1 and 32P-ATP, in the presence or absence of the compounds for 45 min to allow H1 phosphorylation by CDK1. Alkaline phosphatase was then added to one half the tubes that were incubated for additional 15 min for the dephosphorylation reaction. Samples were then subjected to SDS-PAGE for the analysis of H1 phosphorylation and subsequent dephosphorylation. As can be seen in Fig. 11, copper compounds inhibited H1 phosphorylation by CDK1 (odd numbers), as also shown in Fig. 10. Addition of alkaline phosphatase led to a decrease in the levels of 32PeH1 (even

10

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 11. A) Polyacrylamide gel electrophoresis. Lane 1: mix + H2O, lane 2: mix + H2O + phosphatase, odd lanes 3, 5, 7, 9: mix +5 μM compounds (45 min incubation, at 37 °C), respectively; pairs lanes 4, 6, 8, 10: mix +50 μM compounds (45 min incubation, at 37 °C) + phosphatase (15 min incubation, at 37 °C), respectively. mix = buffer, kinase, cyclin-CDK1, 32P-ATP and Histone H1; B) Action of protein phosphatase (15 min incubation) in compounds-kinase system (45 min of incubation) monitored by cellulose TLC. Lane 1: mix + H2O; lane 2: mix + H2O + phosphatase; odd lanes 3, 5, 7 and 9: Mix +5 μM compounds 1 to 4 (after 45 min incubation, at 37 °C), respectively; pairs lanes 4, 6, 8 and 10: mix +50 μM compounds 1 to 4 (after 45 min incubation, at 37 °C) + phosphatase (15 min incubation, at 37 °C), respectively; lane 11: 32P-ATP. mix = buffer, kinase, cyclinCDK1, 32P-ATP and Histone H1.

similar IC50 values, in the range 0.5 to 2 μM against the melanomas, and 2 to 20 μM against the sarcoma cells, after 24 h incubation at 37 °C [34]. The bimetallic compounds 3 and 4 were more reactive than the corresponding mononuclear, and exhibited IC50 values of the same order or even lower than those observed with cisplatin. Therefore, a synergistic effect between the two metal centers was detected. However, the copper-only complexes 1 and 2 exhibited similar toxicity against both types of sarcoma cells, standard and resistant. On the contrary, for cisplatin the IC50 values were higher for the resistant cells [34]. Simulation studies complemented and supported our experimental data. The platinum center in both bimetallic complexes interacts with the phosphate backbone site in all analyzed poses, and no interaction with the nucleobases is observed. For complex [Cu(isad)Pt(NH3)]3+ 3, with the longest linker, copper interactions with the DNA phosphate backbone are observed in 52.8% of poses, and with the DNA nucleobases in 3% of the poses. In contrast, no interaction of copper with DNA is observed in the case of complex [Cu(isah)Pt(NH3)]3+ 4. All the compounds (1, 2, 3 and 4) were able to inhibit the activity of the CDK1/cyclin B kinase protein. Compound 2 is the most active, followed by 1, 4 and finally, 3. Further, they do not strongly inhibit the alkaline phosphatase protein. Finally, the effect of all these complexes is more pronounced toward kinases than against phosphatases, influencing more significantly the phosphorylation steps. In conclusion, although their modes of action are different, both

copper-only and copper‑platinum compounds showed high activity in DNA damage and inhibition of kinase and phosphatase proteins. These properties explain their high reactivity toward tumor cells. Our results indicate that oxindolimine metal complexes are promising candidates to therapeutic anticancer agents. Abbreviations CDK1/cyclin B cyclin dependent kinase, activated by cyclin B CT-DNA calf thymus DNA EPR electron paramagnetic resonance ESI electrospray ionization isad (E)-3-(2-(3-aminopropylamino)ethylimino)indolin-2-one (keto form), or (E)-3-(2-(3-aminopropylamino)ethylimino)3H-indol-2-ol (enol form), ligand prepared from isatin and N(2-aminoethyl)-1,3-propanediamine isah (E)-3-(3-amino-2-hydroxypropylimino)indolin-2-one (keto form), or (E)-3-(3-amino-2-hydroxypropylimino)-3H-indol-2ol (enol form), ligand prepared from isatin and 1,3-diamino2-propanol MALDI matrix-assisted laser desorption/ionization MTT 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide NMR nuclear magnetic resonance

11

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

Fig. 12. A) Graphic representation of the results obtained from testing compounds with kinase and phosphatase proteins, considering the area of the first column, orange, as 100% of 32P-histone, and the area of the second column, green, as 100% of free phosphate after the action of the phosphatase; B) Graphic representation of the results obtained from testing compounds with kinase and phosphatase proteins, considering the area of the first column, blue, as 100% of 32P-histone, and the area of the second column, red, as 100% of free phosphate after the action of the phosphatase. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

ROS reactive oxygen species SDS-PAGE sodium dodecyl sulphate-polyacrylamide gel electrophoresis Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jinorgbio.2019.110863.

analysis and interpretation of data; AMDCF did the conceptualization, design of complexes, participated in analysis and interpretation of data; CCO and AMDCF provided financial resources. All authors wrote and approved the final version of the manuscript being submitted.

Acknowledgements

Declaration of competing interest

Financial support was provided by Brazilian agencies: São Paulo State Research Foundation (FAPESP, grants 2011/50318-1, and 2012/ 17671-2), CEPID Redoxoma Project (FAPESP, grant 2013/07937-8), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, grant 304776/2014-9), and ProReitoria de Pesquisa da Universidade de São Paulo (PRPUSP) Grant 2011.1.9352.1.8.

The University of São Paulo, former employer of AMDCF and CCO, has filed patent applications relating to the metal complexes under study at AMDCF laboratory (BR 10 2015 001045-1). References [1] L. Kelland, Nat. Rev. Cancer 7 (2007) 573–584. [2] K.M. Deo, D.L. Ang, B. McGhie, A. Rajamanickam, A. Dhiman, A. Khoury, J. Holland, A. Bjelosevic, B. Pages, C. Gordon, J.R. Aldrich-Wright, Coord. Chem. Rev. 375 (2018) 148–163. [3] N.P. Farrell, Curr. Top. Med. Chem. 11 (2011) 2623–2631. [4] S. Komeda, T. Moulaei, K.K. Woods, M. Chikuma, N.P. Farrell, L.D. Williams, J. Am. Chem. Soc. 128 (2006) 16092–16103. [5] X. Wang, X. Wang, Z. Guo, Acc. Chem. Res. 48 (2015) 2622–2631. [6] P. Heffeter, U. Jungwirth, M. Jakupec, C. Hartinger, M. Galanski, L. Elbling,

Authors' statement We would like to make an statement about the author's main contribution: EEA synthesized all the compounds, and characterized them by different techniques; JSL and EEA were responsible for biological data acquisition and analysis; CCO developed methodology, formal

12

Journal of Inorganic Biochemistry 203 (2020) 110863

E.E. Aranda, et al.

[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]

M. Micksche, B. Keppler, W. Berger, Drug Resist. Updates 11 (2008) 1–16. S. van Zutphen, J. Reedijk, Coord. Chem. Rev. 249 (2005) 2845–2853. H.-K. Liu, P.J. Sadler, Acc. Chem. Res. 44 (2011) 349–359. A. Casini, J. Reedijk, Chem. Sci. 3 (2012) 3135–3144. F. Arnesano, G. Natile, Coord. Chem. Rev. 253 (2009) 2070–2081. E. Alessio, Eur. J. Inorg. Chem. 2017 (2017) 1549–1560. C.S. Allardyce, P.J. Dyson, Dalton Trans. 45 (2016) 3201–3209. B.-E. Kim, T. Nevitt, D.J. Thiele, Nature Chem. Biol. 4 (2008) 176–185. T.D. Rae, P.J. Schmidt, R.A. Pufahl, V.C. Culotta, T.V. O’Halloran, Science 284 (1999) 805–808. C. Santini, M. Pellei, V. Gandin, M. Porchia, F. Tisato, C. Marzano, Chem. Rev. 114 (2014) 815–862. D. Denoyer, S. Masaldan, S. La Fontaine, M.A. Cater, Metallomics 7 (2015) 1459–1476. B. Deka, T. Sarkar, S. Banerjee, A. Kumar, S. Mukherjee, S. Deka, K.K. Saikia, A. Hussain, Dalton Trans. 46 (2017) 396–409. A. Gupte, R.J. Mumper, Cancer Treat. Rev. 35 (2009) 32–46. Z.D. Liang, Y. Long, W.-B. Tsai, et al., Mol. Cancer Ther. 11 (2012) 2483–2494. F. Tisato, C. Marzano, M. Porchia, M. Pellei, C. Santini, Med. Res. Rev. 30 (2010) 708–749. V.C. de Silveira, J.S. Luz, C. Columbano, I. Graziani, M.R. Ciriolo, J. Inorg. Biochem. 102 (2008) 1090–1103. K. Banerjee, S. Basu, S. Das, A. Sinha, M.K. Biswas, S.K. Choudhur, Free Rad. Res. 50 (2016) 426–447. G. Filomeni, G. Cerchiaro, A.M. Da Costa Ferreira, A. De Martino, J.Z. Pedersen, G. Rotilio, M.R. Ciriolo, J. Biol.Chem. 282 (2007) 12010–12021. G. Filomeni, S. Piccirillo, I. Graziani, S. Cardaci, A.M. Da Costa Ferreira, G. Rotilio, M.R. Ciriolo, Carcinogenesis 30 (2009) 1115–1124. R.B. Miguel, P.A.D. Petersen, F.A. Gonzales-Zubiate, C.C. Oliveira, N. Kumar, R.R. do Nascimento, H.M. Petrilli, A.M. Da Costa Ferreira, J. Biol. Inorg. Chem. 20 (2015) 1205–1217. S. Castelli, M.B. Gonçalves, P. Katkar, G.C. Stuchi, R.A.A. Couto, H.M. Petrilli, A.M. Da Costa Ferreira, J. Inorg. Biochem. 186 (2018) 85–94.

[27] Y. Jung, S.J. Lippard, Chem. Rev. 107 (2007) 1387–1407. [28] P. de Hoog, C. Boldron, P. Gamez, K. Sliedregt-Bol, I. Roland, M. Pitié, R. Kiss, B. Meunier, J. Reedijk, J. Med. Chem. 50 (2007) 3148–3152. [29] S. Ozalp-Yaman, P. de Hoog, G. Amadei, M. Pitié, P. Gamez, J. Dewelle, T. Mijatovic, B. Meunier, R. Kiss, J. Reedijk, Chem. Eur. J. 14 (2008) 3418–3426. [30] G. Manning, D.B. Whyte, R. Martinez, T. Hunter, S. Sudarsanam, Science 298 (2002) 1912–1934. [31] J.B. Vincent, M.W. Crowder, B.A. Averill, Trends Biochem. Sci. 17 (1992) 105–110. [32] A. Uzma, A.K. Witkiewicz, N.C. Turner, E.S. Knudsen, Na. Rev. Drug Discov. 14 (2015) 130–146. [33] M.H. Cohen, G. Williams, J.R. Johnson, J. Duan, J. Gobburu, A. Rahman, K. Benson, J. Leighton, S.K. Kim, R. Wood, et al., Clin. Cancer Res. 8 (2002) 935–942. [34] E.E. Aranda, T.A. Matias, K. Araki, A.P. Vieira, E.A. de Mattos, P. Colepicolo, C.P. Luz, F.L.N. Marques, A.M. Da Costa Ferreira, J. Inorg. Biochem. 165 (2016) 108–118. [35] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson, J. Comput. Chem. 30 (2009) 2785–2791. [36] D.J.F.M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Peterson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Had, Gaussian 09, (n.d.). [37] K. Sayin, D. Karakas, Spectrochim Acta Part A: Mol. Biol. Spec. 188 (2018) 537–546. [38] C.M. Breneman, K.B. Wiberg, J. Comput. Chem. 11 (1990) 361–373. [39] S.S. Batsanov, Inorg. Materials 37 (2001) 871–885. [40] A. Kellett, Z. Molphy, C. Slator, V. McKee, N.P. Farrel, Chem. Soc. Rev. 48 (2019) 971–988. [41] X. Chen, B. Ramakrishnan, S.T. Rao, M. Sundaralingam, Nat. Struct. Biol. 1 (1994) 169–175. [42] L. Martino, A. Virno, B. Pagano, A. Virgilio, S. DiMicco, A. Galeone, C. Giancola, G. Bifulco, L. Mayol, A. Randazzo, J. Am. Chem. Soc. 129 (2007) 16048–16056. [43] J. Zhu, M. Lin, D. Fan, Z. Wu, Y. Chen, J. Zhang, Y. Lu, Z. Gueo, Dalton Trans (2009) 10889–10895.

13