Synthesis of Cu ( II ) complex with schiff bases derived from aryl-S-benzyildithiocarbazate : Antimicrobial activity and in silico biological properties evaluations

1 Grupo de Investigación en Química de Coordinación y Bioinorgánica, Departamento de Química, Facultad de Ciencias, Av. Carrera 30 # 45-03, Universidad Nacional de Colombia – Sede Bogotá, Colombia. 2 Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Minas Gerais, Av. Antônio Carlos, 6627, CEP 31270-901, Belo Horizonte, Minas Gerais, Brazil. 3 Grupo de Investigación en Compuestos de Coordinación y Catálisis, Facultad de Ciencia y Tecnología, Universidad de Ciencias Aplicadas y Ambientales, U. D. C. A., Sede Norte, Calle 72 # 14-20, Bogotá, Colombia.


INTRODUCTION
In the last decades the coordination complex derived from S-benzyldithiocarbazate has been widely studied, mainly due to its activity against viruses, bacteria and fungus and as antitumor, pesticide and others (Tarafder et al., 2002a,b;Zangrando et al., 2015;Islam et al., 2014;Nanjundan et al., 2014).The presence in their structure *Corresponding author.E-mail: aeburgosc@unal.edu.co.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License of two electron-donors facilitates the coordination processes with metal ions involving nitrogen and sulfur atoms.This property makes these compounds even more interesting for the development of further studies involving biological and pharmacological activities (Garoufis et al., 2009).
Schiff bases are molecules used in the generation of supra molecular structures (Ziessel, 2001).This type of compounds are broad chemical and biological interest, mainly because of the possible ways of coordinating to the metal ion, the stability of the complexes which, in general present an increased activity compared to their respective free ligands (Ravoof et al., 2004;Ravoof et al., 2007;Monika et al., 2014).The shiff bases derived from alkyl-S-benzyldithiocarbazates obtained by the condensation of the an aldehyde or a ketone with dithiocarbazate, present different biological properties increasing the interest of researchers for this kind of ligands (How et al., 2008;Tarafder et al., 2001;Islam et al., 2011;Ali et al., 2012).To metallic complexes obtained reacting these ligands with Cd(II), Zn(II), Ni(II), Co(II), Sn(IV) or Cu(II) were attributed in vitro pharmacological activities such as antibacterial, antifungal, and inhibitory effect of the leukemia and ovarian cancer cell growth (Ali, 1997;Tarafder et al., 2002a,b;Islam et al., 2014;Nanjundan et al., 2014;Monika et al., 2014;Esmaielzadeh et al., 2014;Zangrando et al., 2015).And, studies involving schiff bases such as thiosemicarbazones, showed that the complexes often present greater pharmacological activity than the original free ligands.Thus, the study of these compounds becomes more important (Beraldo, 2004;Rigol et al., 2005).
In recent years, gold complexes have been used for the treatment of arthritis, silver complexes for alleviate antibacterial infections and platinum complexes have been used against cancer (Kelland, 2007;Ott, 2009;Chen et al., 2009;Alanne et al., 2013).In these and other cases, the complex has been more active than the original free ligand or may induce a lower cell drug resistance (Alanne et al., 2013).In addition, some side effects produced by ligands may also be reduced due to complexation with metals.However, to acquire greater knowledge about the biological effect of these compounds, it is necessary a long period of researches and the involvement of high financial resources.This fact makes it difficult or unfeasible to conduct more comprehensive studies to prove the true pharmacological potential of these compounds.A good alternative to selecting the first biological targets to be evaluated is the use of tools in which the chemical structure of the synthesized compounds is compared with the active compounds already known.In some situations, the use of these tools allows establishing structural similarities with known active compounds, and provides good starting point to investigate the potential pharmacological activity associated to specific compound.The strategy of PASSonline web tool involves decomposition of chemical structure on topological structure superposition fragmental notation (SSFN) descriptors type.Then, this tool develops a comparison of these descriptors with about 250,000 data of drugs, drug candidates, compounds under registration processes, toxic substances, chemical oncogenes and other biologically active compounds (Lagunin et al., 2000;Filimonov et al., 2014).And, by means of Bayesian statistical methods, the potential pharmacologic and/or toxicological activities are established based on its relationships with similar compounds available on the base data.The final results are established in accordance with quantitative structure-activity relationship (QSAR) models (Lagunin et al., 2000;Lagunin et al., 2010).
The SHApe-FeaTureSimilarity (SHAFTS) is used in ChemMapper tool (Gong et al., 2013a) in which it was validated in relation to base data Directory of Useful Decoys (DUD-E) (Wang et al., 2014;Mysinger et al., 2012).In the ChemMapper analysis, the superposition of the tridimensional chemical structures of compounds, together with identification of potential pharmacophores, is used to estimate the potential biological effects of target compound (Gong et al., 2013b;Lu et al., 2011;Liu et al., 2011).
The aldehyde used in each reaction was firstly dissolved in ethanol (5 ml) and mixed with S-benzyldithiocarbazate also dissolved in ethanol (5 ml).The mixture was kept under reflux at 78°C, with constant magnetic stirring for 2 h.Then, the solution was slowly cooled.The solvent was withdrawn in a Buchi rotavapor ® under reduced pressure.After filtration this solid was washed with hot ethanol (3x 15 ml) and kept under vacuum in desiccators, protected against sun light.Each obtained solid L1 (yellow), L2 (yellow) and L3 (light yellow) was respectively cleaned with hexane (3x 15 ml), dried and kept under vacuum in desiccators over anhydrous calcium chloride, protected against sun light.

Synthesis of Cu(II)-complex C1, C2 and C3
The reactant CuCl2.2H2O was dissolved in CHCl3: MeOH (1:1) (10 ml) and then, it was added the ligand dissolved in a sufficient quantity of CHCl3.The reaction mixture was kept under reflux at 68°C, with constant magnetic stirring for 2 h, carefully protected against sun light.The solvent was withdrawn in a Buchi rotavapor ® under reduced pressure.Thus, the brown solid formed was washed with CHCl3, dried and then kept under vacuum in desiccators, protected against sun light.The following conditions ratios were used in the synthesis of complex C1, C2 and C3 as shown in Table 2

Compound characterization
The chemical characterization of compounds was realized by means of physicochemical techniques.The fourier transform infrared (FTIR) spectra were obtained (1% KBr pellet) on SHIMADZU IR Prestige21 FTIR Spectrometer, with absorption band ranging from 4000 a 400 cm -1 .The ultraviolet spectra (UV) were performed at Thermo Scientific Evolution 300 UV-VIS spectrophotometer (280 to 900 nm), using quartz cuvette with 1 cm optical path.The sample was dissolved in Dimethyl Sulfoxide (DMSO), at concentration of 1x10 -3 mol/L.The nuclear magnetic resonance of hydrogen ( 1 H NMR) spectra of ligand was recorded on Bruker AVANCE DRX-400 spectrometer (400 MHz).The ligand was dissolved in deuterated solvent DMSO-d6 and CDCl3 and TMS used as internal standard (δH = δC = 0).The molar conductance values of the 10 -3 M DMSO solutions of the complexes were measured with a Hanna conductivity meter HI 5321.The electron paramagnetic resonance (EPR) was performed at room temperature on Bruker ESP300 Electron Spin Resonance ESP300-E X-band-Spectrometer, g-values were obtained from g=h/H,  is the Bohr magneton, h the Planck constant,  is the frequency and H is the centers field at which the resonance occurs.The g-value is often a key parameter in identifying paramagnetic centers in a particular symmetry.

Antimicrobial activity
The bacteria E. coli (ATCC 25922) (gram-positive) and S. aureus (ATCC 292139) (gram-negative), and the yeasts C. albicans (ATCC 10231) and S. cerevisiae (ATCC AH22) were used to evaluate the antimicrobial activity of synthesized ligands and its respective Cu(II)-complex.The antimicrobial assays were performed using the disk diffusion method (Bauer and Kirby, 1966;Tamayo et al., 2014).
The solution of microorganism (1.0 ml) with 600 nm optical density (OD600) ranging from 0.1 to 0.2 was added to Petri plates.As culture medium, trypticase soja Agar (TSA) was added for bacteria and potato dextrose Agar (PDA) for the yeasts.Cefalothin (40.0 mg/ml) for bacteria and clotrimazole (1.0 %) for yeasts were used as positive control.Standard solutions (300.0 µg/ml) of ligands and complexes were prepared from stock solutions.Then, 20.0 L of ligand L1, L2 or L3 and of the Cu(II) complex C1, C2 or C3 was added on culture medium.The microorganism susceptibility to each compound was determined by means of the growth inhibition halo (mm) and the minimum inhibitory concentration (MIC) was determined in µg/mL.The assays were performed in triplicate considering the median  standard deviation of the mean (s.e.m.).

PASSonline
To develop studies by means of PASSonline it is necessary to upload sdf file of the structure of compound.The PASSonline tool realizes the decomposition of structure in descriptors and a comparison with descriptors of biologically active substances available in its data base.The ligand L1, L2 and L3, and its (CuII)complex C1, C2 and C3 were compared with more than 250,000 compounds, including drugs, potential drugs, compounds under register processes, toxic substances, oncogenes and other biologically active compounds ( (Lagunin et al., 2000;Lagunin et al., 2010).Simultaneously this tool develops predictions of different biological activities for the structure of organic compound, correlating the probability "to be active" (Pa) with probability "to be inactive" (Pi).Pa represent the probability that the studied compound is belonging to the sub-class of active compounds (resembles the structures of molecules, which are the most typical in a sub-set of "actives" in PASS training set).Pi represent the possibility that the studied compound is belonging to the sub-class of inactive compounds (resembles the structures of molecules, which are the most typical in a sub-set of "inactives" in PASS training set).At the end of analytical process, the results were established by difference (Pa-Pi), and the potential biological activities generated for ligands and complexes were organized in Table 6.

ChemMapper
The ligands L1, L2 and L3 and complexes C1, C2 and C3 also were analyzed using the ChemMapper tool.The predictions of biological activities for these compounds were established by means of the following base data (active compounds): PDB (7072), KEGG (5928), Drug Bank (4.645),ChEMBL (339.624) and BindingDB (364.221).The similarity score suggested by this tool were converted in percent of similarity with the compounds of the data base (1.2 similarity score = 0.6 % similarity).In the present study, the results of the similarity indices obtained through ChemMapper were converted in percent of similarity in the range of 0 to 1.0.The normalized values (score) presented by the tools were not used.The maximum values selected for the similarity score for the same target were considered only when it is > 60 % similarity.The profile of L1, L2 and L3 and C1, C2 and C3 was analyzed both in terms of chemical structure, such as in relation to the type of molecule superposition, correlating the average number of pharmacophores found in each one, as well as the same function in terms of similarity of the spatial conformation.At the end of analytical process applied to the compounds, the potential biological targets generated by this tool were organized in Table 7.

RESULTS AND DISCUSSION
The FTIR spectrum of the L1, L2 and L3 present characteristic absorption band of alkyl-Sbenzyldithiocarbazate as reported by Ali and Tafafder (1997).In the spectrum, was observed strong absorption band at 3177 cm -1 that corresponds to the v(N-H) of NH2 group present on the free ligand (Ali and Tafafder, 1997;Silverstein and Webster, 1997).In aqueous solutions, the ligand L1, L2 and L3 present a thione-thiol tautomerism which is brought due to the presence of thioamide NHC=S functional group (Beckford et al., 2011).The absence in the IR spectra of absorption band at 2600 cm -1 , attributed to sulfhydryl group stretch (S-H), and the presence of band at 940 to 970 cm -1 characteristic of C=S stretch, indicated that these ligands has in thione form, typical of similar compounds in solid state (Beckford et al., 2011;Tamayo et al., 2014).The schiff base derivatives also present strong absorption bands at 1597 cm -1 (L1 and L2) and 1577 cm -1 (L3) (Table 3).These are assigned to the (C-N) modes for the free ligand.In metal complexes, this band is shifted to lower frequencies, as observed in the FTIR spectrum of complex C2, and was associated to the delocalization of bonding electrons that occur in complexation.Higher frequencies observed for C1 and C3 were correlated to metal-nitrogen bond formation, which induced shorter C-N bond lengths.
The ligand acts as a bidentate sulphur-nitrogen chelating agent.The ligands (L1, L2 and L3) also showed absorption bands at 929, 948 and 933 cm -1 , that were attributed to (C=S) (Ali and Tafafder, 1997;Silverstein and Webster, 1997).The (C=S) observed in the spectrum of free ligand is shifted to the higher energy in the complexes (Table 1), thus supporting the thione bonding with metal ions, such as the sulfur atom of the C=S group forming a coordination site (Al-Amin et al., 2014).
This data together with the absence of hydrogen signal around  4.00 ppm in the NMR spectra related to sulfhydryl group confirm the ligands as thione tautomers (Beckford et al., 2011).In the 1 H NMR spectra of ligands was observed a signal at  4.5 ppm (2H) which was attributed to methylene of S-benzyldithiocarbazate and a signal at  7.8 to 8.2 ppm correlated to hydrogen of azomethine imine (Beckford et al., 2011).The presence of these signals is typical of shiff bases and contributes to confirm the formation of ligands.By means of spectroscopic data L1 was identified as 3,4methylenedioxybenzyl-S-benzyldithiocarbazate, L2 as pchloro-benzyl-S-benzyldithiocarbazate, and L3 as 2,6dichloro-benzyl-S-benzyldithiocarbazate: Regarding the complex formation, the following yields were obtained: 32 (C1), 24 (C2) and 20% (C3).By means of spectral data, the complexes were identified as In comparison with the ligand L1, L2 and L3, the IR absorption bands of (C=N) and (C=S) of the respective complex C1, C2 and C3 occurred in longer wavelengths in the spectra (Table 3), indicating coordination of nitrogen and sulfur atoms with ion copper (II).By means of this result was possible to infer the formation of Cu(II) complex with an octahedral structure due to the nature of copper and the stereochemistry employed in the synthesis reaction (Figure 1).
The FTIR and UV-V is absorption bands of ligands L1, L2 and L3 compiled in Table 3, and physical properties of the Cu(II) complexes C1, C2 and C3 as shown in Table 4.The absorption band in the visible region UV-Vis spectra of these complexes was attributed to n→π* transition for the S-benzyldithiocarbazate moiety.Although the n→π* band of dithiocarbazate group also showed a blue shift (339 to 375 nm) in Cu(II) complexes (Tarafder et al., 2002a,b;Monika et al., 2014).The bathochromic shift observed in the UV-Vis spectra of Cu(II) complexes suggesting a structural change or electronic behavior associated to a selected transition.In the Cu(II) complexes, the band that is related to the possible transition 2 T 2g  2 E g, and feature of a octahedral geometry was not observed.
In the electronic spectra of the complexes C1, C2 and C3, the absence of d-d transition, is due to the intrusion of tails of intense charge transfer bands into visible portion of the spectrum which masks the expected d-d bands, the tail of this band covers all d-d transition (Takjoo and Centore, 2013).In ligands containing sulphur atoms, the strong inter-ligands charge transfer transitions (SM) interfere in the observation of  bands which can be correlated to the azomethine group and ring (Takjoo and Centore, 2013).These results are in accordance with published data for similar complexes (Tarafder et al., 2002a,b;Ali et al., 2011).At room temperature were observed Bohr magneton (μB) for Cu(II) complex C1, C2 and C3 effective in the range of 1.7 to 2.1 μB corresponding to one unpaired electron, suggesting an octahedral environment.This behavior is expected for transition metals (3d9), as in this case with the Cu(II) in dilute environment.The molar conductance of these complexes dissolved in DMSO is presented in Table 2.The low molar conductance values show that the metal complexes are non-electrolytes.These results indicate that the complexes do not dissociate in this solvent because the values are very small compared with those expected for 1:1 electrolytes (Ali et al., 2011).In the EPR spectra (Figure 2), obtained at room temperature for complex C1, was observed a signal at 330 mT (3.3 x10 7 Gauss) (West and Liberta, 1993) that confirms the presence of Cu(II) in the structure.
The g values of the complex were obtained by means of simulation and the experimental results were 2.04, 2.05 and 2.13, respectively correspondent to g x and g y y g z .For constant A x , A y and A z were 10, 10 and 90, respectively.Based on the values obtained for g z (2.13) and A z (90), it was determined that complex C1 has an elongated octahedral structure (Geary, 1971;West and Liberta, 1993).Having gx and gy, it is suggested that this complex is a slight rhombic distortion octahedral (Geary, 1971;West and Liberta, 1993;Hamid et al., 2009).These values differ from those previously reported for this type of structure.However, the observed differences is herein attributed due to the published data were obtained at temperatures below 100 K (Hamid et al., 2009;Garribba and Micera, 2006), thus the working temperature affects the behavior of the analyzed molecule.

Antimicrobial activity
Regarding ligand L1, L2 and L3 were not observed activity against bacteria E. coli.Nevertheless, the complexes inhibit the growth of this bacterium, with MIC of C1 ≥ 15 µg/mL, C2 ≥ 150 µg/mL and C3 ≥ 200 µg/mL.The complex C1 induced growth inhibition of S. aureus at ≥ 50 µg/ml and of C. albicans at ≥ 15 µg/mL (Table 5).In the assays with S. cerevisiae , were observed inhibition activity induced by C1 at ≥ 15 µg/mL and C2 ≥ 50 µg/mL (Table 5).These results suggest C1 as compound with higher inhibitory property than C2 and C3.This fact can be attributed to the interaction between the binder whose structure has electron donor atoms (N and S), and also has the dioxolane ring B (Figure 1).
The presence of oxygen atoms completing the five member ring with carbons of benzyl group has being studied indicating that they influence the activity of both ligands as its metal complexes with Ni(II) and Sn(II) (Ali et al., 2008;Muhammad et al., 2010).In this study the higher activity presented by complex C1 probably occurred due to the presence of two oxygen atoms in its structure (Table 5).

In silico activity prediction
By means of ChemMapper and PASSonline tools were  not possible to observe in silico predictions for Cu(II) complex C1, C2 and C3.This fact may be associated to the absence of similar compounds tested and available biological effect information, or to the low similarity with active compounds available in the libraries evaluated mainly with the aid of this tool.Using the PASSonline tool were not observed prediction of adverse effect for ligands L1,L2 and L3 related to abortion inducer,arrhythmogenic,bronchoconstrictor,carcinogenic,cardiotoxic,convulsant,cytotoxic,DNA damaging,depression,embryotoxic,emetic,eye irritation,hallucinogen,hepatotoxic,hypercholesterolemic,hyperglycemic,hypertensive,hyperthermic,hypnotic,hypokalemia,hypothermic,narcotic, nephrotoxic, neurotoxic, QT interval prolongation, rubefacient, sedative, sensitization, skin irritation, spasmogenic, mutagenic, teratogen, thrombocytopoiesis inhibitor, torsades de pointes, toxic for respiratory center, vasodilator and vasopressor.
By means of ChemMapper tool were predicted antiviral and antitumor activities for the targets beta-1 adrenergic receptor (Hjalmarson et al., 2002), beta-2 adrenergic receptor (Irwin et al, 1990), cellular tumor antigen P53 (Cuff and Ruby, 1996), cyclin-A2 (Cribier et al., 2013), 5hydroxytryptamine 1A receptor (Fiorino et al., 2014), cell division protein kinase 2 (Chow et al., 2009) and collagenase 3[MMP-13] (Casini et al., 2002), protooncogene tyrosine-proteinkinase LCK (Hansen et al., 2010) (Dolai et al., 2011), thymidylatesynthase (Balzarini, et al., 1987) (Table 7).The target trypsin beta indicated by ChemMapper tool is associated to bacterial infectious diseases, and the target glutathionereductase is associated to parasitic diseases.Based on the results was possible to deduce that ligand L1, the Schiff base that originate complex C1, present significant similarity with compounds that acts as inhibitors of trypsin-beta (Table 7) which is involved in the growth of bacteria.The chemical structure of ligand L2 and L3 is similar to compounds that induce the inhibition of thymidylate synthase (Table 7).This enzyme is related to the conversion of folic acid to dihydrofolate (DHF) and then to tetrahydrofolate (THF) which is directly involved in the synthesis of bacterial DNA.

Conclusion
Three new ligands L1, L2 and L3 derivatives of S-  ---] Absence of similar ligands in the data base benzyldithiocarbazate and its Cu(II) complexes C1, C2 and C3 were synthesized with a good yield, and its chemical structures were identified by means of its respective IR and 1 H NMR data.Through the spectral data was established that both in solid as in solution, the ligands have thione tautomer form.The structure of complex C1 in solid state was confirmed by EPR, at room temperature.Complex C1 showed the major in vitro antimicrobial activity.The bacterial inhibition effect was observed for all microorganism subjected to assays.The inhibition property was attributed mainly to the metalligand interaction and probably by the presence of two oxygen atoms in the structure of C1, commonly found in active compounds.This study results contribute to future studies of similar complexes involving quantitative structure-activity relationships (QSAR).
The pharmacological effect or potential biological target, established by means of computer (in silico) tools can be further proven through experiments performed in vitro and/or in vivo.PASSonline (Prediction Activity Spectra of Substances) and ChemMapper represent a good examples of web tools that contain large amount of information available in open access databases.
 Indicates no observed activity in the concentration range evaluated.

Table 6 .
In silico test prediction of potential biologic effect of ligand L1, L2 and L3 found using PASSonline * tool.

Table 7 .
In silico test prediction of potential biologic targets of ligand L1, L2 and L3 found using ChemMapper** tool.