Neutral and cationic manganese(II)–diclofenac complexes: structure and biological evaluation

The interaction of MnCl2 with the non-steroidal anti-inflammatory drug sodium diclofenac in the presence of 2,2′-bipyridine and pyridine resulted in the formation of cationic and neutral mononuclear complexes [Mn(diclofenac)(2,2′-bipyridine)(H2O)2] (diclofenac) (1) and [Mn(diclofenac)2(pyridine)2(H2O)2] (2), respectively. The structure of 1 was characterized by X-ray crystallography. In a preliminary attempt to evaluate the biological properties and possible application, the interaction of the complexes with calf-thymus DNA and human or bovine serum albumins was monitored. Additionally, the ability of the compounds to scavenge radicals such as 1,1-diphenyl-picrylhydrazyl, 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid), and hydroxyl radicals was evaluated; the complexes were more potent scavengers than free sodium diclofenac.

Manganese is among the most significant biometals mainly because of its existence in the active center of many enzymes of diverse functionalities [12,13]. Additionally, the manganese-containing compounds SC-52608 and Teslascan are used in medicine as an anticancer agent and a MRI contrast agent, respectively [14]. In the context of bioinorganic chemistry, many manganese complexes with diverse ligands have been tested for potential activity and have shown noteworthy in vitro antimicrobial [15,16], antiproliferative [17][18][19], and antifungal [20] activities.
DNA stock solution was prepared by dilution of CT DNA to buffer (containing 15 mM trisodium citrate and 150 mM NaCl at pH 7.0) followed by exhaustive stirring for three days, and kept at 4°C for no longer than a week. The stock solution of CT DNA gave a ratio of UV absorbance at 260 and 280 nm (A 260 /A 280 ) of 1.87, indicating that the DNA was sufficiently free of protein contamination [27]. The DNA concentration was determined by UV absorbance at 260 nm after 1 : 20 dilution using ε = 6600 M −1 cm −1 [28].
Infrared (IR) spectra (400-4000 cm −1 ) were recorded on a Nicolet FT-IR 6700 spectrometer with samples prepared as KBr disk. UV-visible (UV-vis) spectra were recorded as nujol mulls and in solution at concentrations of 10 −5 -10 −3 M on a Hitachi U-2001 dual beam spectrophotometer. Room temperature magnetic measurements were carried out by the Faraday method using mercury tetrathiocyanatocobaltate(II) as a calibrant. C, H and N elemental analyses were performed on a Perkin-Elmer 240B elemental analyzer. Molar conductivity measurements were carried out with a Crison Basic 30 conductometer. Fluorescence spectra were recorded in solution on a Hitachi F-7000 fluorescence spectrophotometer. Viscosity experiments were carried out using an ALPHA L Fungilab rotational viscometer equipped with an 18 mL LCP spindle and the measurements were performed at 100 rpm.

X-ray structure determination
Crystals of 1 were taken from the mother liquor and mounted at room temperature on a Bruker Kappa APEX2 diffractometer equipped with a triumph monochromator using Mo Kα radiation. Unit cell dimensions were determined and refined by using the angular settings of at least 100 high intensity reflections (>10 σ(I)) in the range 15°< 2θ < 40°. Intensity data were recorded using φ and ω scans. Crystal presented no decay during the data collection. The frames collected were integrated with the Bruker SAINT software package [29] using a narrow-frame algorithm. Data were corrected for absorption using the numerical method (SADABS) based on crystal dimensions [30]. The structure was solved using the SUPERFLIP package [31], incorporated in Crystals. Data refinement (full-matrix leastsquares methods on F 2 ) and all subsequent calculations were carried out using the Crystals version 14.40b program package [32].
All non-hydrogen atoms were refined anisotropically. Hydrogens were located by difference maps at their expected positions and refined using soft constraints. By the end of the refinement, they were positioned geometrically using riding constraints to bonded atoms. Crystal data as well as details of data collection and structure refinement for the compounds are given in table 1. Illustrations were drawn by CAMERON [33]. Further details on the crystallographic studies as well as atomic displacement parameters are given as Supporting Information in the form of cif files.

Antioxidant biological assay
Each experiment in the in vitro assays was performed at least in triplicate and the standard deviation of absorbance was less than 10% of the mean. Ethanol was used as control solution. The absorbance of the final solution at 517 nm was recorded at room temperature after 20 and 60 min in order to examine the time dependence of the radical scavenging activity [34]. The radical scavenging activity of the compounds was expressed as the percentage reduction of the absorbance values of the initial DPPH solution (RA%). NDGA and BHT were used as reference compounds.

2.4.2.
Competition of the tested compounds with DMSO for hydroxyl radicals. The hydroxyl radicals which were generated by the Fe 3+ /ascorbic acid system were detected according to Nash [35], by the determination of formaldehyde produced from the oxidation of DMSO. The reaction mixture contained EDTA (0.1 mM), Fe 3+ (167 μM), DMSO (33 mM) in phosphate buffer (50 mM, pH 7.4), ascorbic acid (10 mM) and the tested compounds (0.1 mM). After 30 min of incubation at 37°C, the reaction was stopped by the addition of CCl 3 COOH (17% w/v) and the absorbance at λ = 412 nm was measured. Trolox was used as an appropriate reference standard. The competition of the compounds with DMSO for Å OH, generated by the Fe 3+ /ascorbic acid system, was expressed as percent inhibition of formaldehyde production and was used for the evaluation of hydroxyl radical scavenging activity ( Å OH%).

2.4.3.
Assay of radical cation scavenging activity. ABTS was dissolved in water to a 2 mM concentration. The cationic radical ABTS Åþ was produced by reacting ABTS stock solution with 0.17 mM potassium persulfate and the mixture was allowed to stand in the dark at room temperature for 12-16 h before use [34]. The oxidation of ABTS is incomplete, because stoichiometric reaction ratio of ABTS and potassium persulfate is 1 : 0.5. The oxidation of ABTS commenced immediately, but the absorbance became maximal and stable after more than 6 h of reaction. The radical was stable in this form for more than two days when stored in the dark at room temperature. The ABTS Åþ solution was diluted with ethanol to an absorbance of 0.70 at 734 nm. After addition of 10 μL of the compounds or reference standards (0.1 mM) in DMSO, the absorbance reading was taken exactly 1 min after initial mixing [36]. The radical scavenging activity of the complexes was expressed as the percentage inhibition of the absorbance of the initial ABTS solution (ABTS%). Trolox was used as an appropriate standard.

Soybean LOX inhibition study in vitro.
The soybean LOX inhibition was evaluated in vitro as reported in the literature [36]. The compounds were dissolved in ethanol and incubated at room temperature with sodium linoleate (0.1 mM) and 0.2 mL of enzyme solution (1/9 × 10 −4 w/v in saline). The conversion of sodium linoleate to 13-hydroperoxylinoleic acid at 234 nm was recorded and compared with the appropriate standard inhibitor caffeic acid.

DNA-binding studies
The interaction of 1-4 with CT DNA was studied by UV spectroscopy in order to investigate the possible binding mode to CT DNA and to calculate the binding constants to CT DNA (K b ). In the UV titration experiments, the binding constant of the complexes with DNA, K b (in M −1 ), was determined by the Wolfe-Shimer equation (equation (1) where [DNA] is the concentration of DNA in base pairs, ε A = A obsd /[compound], ε f = the extinction coefficient for the free compound and ε b = the extinction coefficient for the compound in the fully bound form [37]. Control experiments with DMSO were performed and no changes in the spectra of CT DNA were observed. The interactions of 1-4 with CT DNA were also investigated by monitoring the changes observed in the cyclic voltammogram of a 0.40 mM 1 : 2 DMSO : buffer solution of complex upon addition of CT DNA at diverse r values. The buffer was also used as the supporting electrolyte and the cyclic voltammograms were recorded at ν = 100 mV s −1 .
The viscosity of DNA ([DNA] = 0.1 mM) in buffer solution (150 mM NaCl and 15 mM trisodium citrate at pH 7.0) was measured in the presence of increasing amounts of the compounds (up to the value of r = 0.22). All measurements were performed at room temperature. The obtained data are presented as (η/η 0 ) 1/3 versus r, where η is the viscosity of DNA in the presence of the compound and η 0 is the viscosity of DNA alone in buffer solution.
Competitive studies of the compounds with EB were investigated by fluorescence spectroscopy in order to examine whether they can displace EB from its CT DNA-EB complex. The CT DNA-EB complex was prepared by pre-treating 20 μM EB and 26 μM CT DNA in buffer (150 mM NaCl and 15 mM trisodium citrate at pH 7.0). The possible intercalating effect of the compounds was evaluated by adding a certain amount of the compound into a solution of the DNA-EB complex. The influence of the addition of each compound to the DNA-EB complex solution was obtained by recording the variation of fluorescence emission spectra with excitation wavelength at 540 nm. Complexes 1-4 show no fluorescence at room temperature in solution or in the presence of DNA under the same experimental conditions and the quenching may be attributed to the displacement of EB from its EB-DNA complex. The values of the Stern-Volmer constant (K SV , in M −1 ) were calculated according to the linear Stern-Volmer equation (equation (2)) [38]: where I o and I are the emission intensities in the absence and the presence of the quencher, respectively, and [Q] is the concentration of the quencher (i.e. complexes). K SV was obtained from the Stern-Volmer plots by the slope of the diagram I o I versus [Q].

Albumin binding studies
The albumin binding studies were performed by tryptophan fluorescence quenching experiments using BSA (3 μM) or HSA (3 μM) in buffer (containing 15 mM trisodium citrate and 150 mM NaCl at pH 7.0). The fluorescence spectra were recorded from 300 to 500 nm at an excitation wavelength of 295 nm [39]. The quenching of the tryptophan-emission intensity of BSA or HSA was monitored at 343 nm or 351 nm, respectively, using 1 and 2 as quenchers at increasing concentration. The fluorescence spectra of the complexes were also recorded with an excitation at 295 nm and an emission maximum appeared at 365 nm, as previously reported for diclofenac compounds [4,9,10]; thus, the serum albumin (SA) fluorescence emission spectra were corrected by subtracting the spectra of the compounds. The influence of the inner-filter effect [40] on the measurements was evaluated according to equation (3): where I corr = the corrected intensity, I meas = the measured intensity, c = the concentration of the quencher, d = the cuvette (1 cm), ε(λ exc ), and ε(λ em ) = the ε of the quencher at the excitation and the emission wavelength, respectively, as calculated from the UV-vis spectra of the complexes [40]. The Stern-Volmer and Scatchard equations (equations (4)- (6)) [41] and graphs were used in order to calculate the (i) dynamic quenching constant K SV (in M −1 ) and the approximate quenching constant k q of the SAs (in M −1 s −1 ) and (ii) the SA-binding constant K (in M −1 ) and the number of binding sites per albumin n, respectively. The Stern-Volmer equation is [41]: where : The Scatchard equation is [40]: K is calculated from the slope and n is given by the ratio of y intercept to the slope in plots  . The complexes were characterized by elemental analysis, IR and UV-vis spectroscopy and room temperature magnetic measurements. The solid state structure of 1 was determined by X-ray crystallography.

Spectroscopic study of the complexes
In the IR spectrum of Nadicl, two bands attributed to the antisymmetric, ν asym (CO 2 ), and the symmetric, ν sym (CO 2 ), stretching vibrations were observed at 1575(s) and 1399(s) cm −1 [4], respectively. In the IR spectra of the complexes, the ν asym (CO 2 ) vibration was located at 1577-1578(vs) cm −1 , while the ν sym (CO 2 ) vibration was observed at 1422(s) and 1396 (s) cm −1 for 1 and 1373(s) cm −1 for 2 (figure S1). Furthermore, the parameter Δν(CO 2 ) [42] was determined for both complexes in order to suggest the coordination mode of the carboxylato group of diclofenac ligands. For 1, two Δν(CO 2 ) values (156 and 182 cm −1 ) were obtained; the former (156 cm −1 ) is indicative of bidentate chelating binding and the latter (182 cm −1 ) may be attributed to the ionic form of diclofenac found as counter-ion in 1 [9]. For 2, the Δν(CO 2 ) has a value of 204 cm −1 which is rather indicative of monodentate binding of the carboxylato group of diclofenac. UV spectra of the complexes were recorded as nujol mull, in DMSO solution and in the presence of solvents and the buffer solution (150 mM NaCl and 15 mM trisodium citrate at pH 7.0) used in the biological evaluation. In all cases, the complexes exhibit the same UV spectral features which in combination to the literature [10] and the molar conductivity measurements may suggest that the compounds keep their integrity in solution [21,22]. group P-1. There are two crystallographically independent molecules in an asymmetric unit, denoted as 1A and 1B, respectively, which present slight differences in bond distances and angles (table 2). The two diclofenac found in the molecule have different roles; one is a ligand bound to manganese in a bidentate chelating mode (a symmetric mode is found in   (17) Mn (51) [11], as described below. According to the IR spectra, the diclofenac ligands are deprotonated in monodentate binding mode coordinated to manganese via a carboxylato oxygen. On the basis of the magnetic data at room temperature, the complex is mononuclear exhibiting an octahedral geometry around Mn(II). The octahedron is formed by two oxygens of the two diclofenac ligands, two nitrogens from two pyridine ligands, and two oxygens from two aqua ligands.

Interaction with CT DNA
The potential biological (antiproliferative and/or anti-inflammatory) activities of the NSAIDs and their complexes have been closely examined in regard to their affinity to DNA [45,46]. Within this context and as continuation of our recent studies [11,[21][22][23][24][25][26], the interaction of 1-4 with CT DNA was investigated by UV spectroscopy, cyclic voltammetry, and DNA viscosity measurements, while their ability to displace the classic intercalator EB was monitored by fluorescence emission spectroscopy.
In  suggest safely a possible mode interaction; therefore, more experiments including cyclic voltammetry and viscosity measurements were conducted to come to a safe conclusion concerning the interaction mode [47].
The cyclic voltammograms of the complexes were recorded in a 1/2 DMSO/buffer solution (0.33 mM) in the absence and presence of CT DNA (representatively shown for 3 in figure S3). The observed decrease of the current intensity may suggest the existence of equilibrium between free and DNA-bound complex as evidence of the complex-DNA interaction [52]. The cathodic (E pc ) and anodic (E pa ) potentials of the redox couple Mn(II)/Mn (I) for the complexes along with the corresponding shifts in the presence of CT DNA are cited in table 4. When CT DNA is added in the solution of the complex, both the cathodic and the anodic potentials exhibit a positive shift (ΔE pc/a = (+5) − (+65) mV) from which we may conclude intercalation as the most possible mode of interaction between the complexes and CT DNA bases [11,[21][22][23][24][25][26]; a conclusion which clarifies preliminary UV spectroscopic finding and is in accordance to viscosity experiments.
Monitoring of the viscosity changes of a DNA solution in the presence of increasing amounts of the complexes may provide significant information in regard to the DNA-binding mode of the complexes; the DNA-viscosity is sensitive to DNA-length changes, since the DNA-viscosity and DNA-length are related via the equation L/L 0 = (η/η 0 ) 1/3 , with η/η 0 Notes: (i) "+" denotes hyperchromism, "−" denotes hyprochromism.
(ii) "+" denotes red-shift, "−" denotes blue-shift. denoting the relative solution viscosity and L/L 0 the DNA-length [53]. In the present study, the viscosity measurements were carried out on a CT DNA solution (0.1 mM) in the presence of increasing amounts of 1-4 (up to the value of r = 0.22, figure 4). The addition of increasing amounts of the complexes to the DNA solution resulted in a moderate to significant (in the case of 4) increase of the relative DNA-viscosity; this result may be explained via the insertion of the complexes between the DNA base pairs resulting in an increase in the separation distance of DNA-base pairs at intercalation sites and, thus, an increase in the DNA-length [21][22][23][24][25][26]54]. Therefore, the DNA-viscosity increase may be considered evidence of intercalation between the complexes and DNA; a conclusion which is in agreement with cyclic voltammetry data. The ability of the complexes to displace the typical DNA-intercalator EB from its EB-DNA complex was monitored by fluorescence emission spectroscopy [39,54]. The intercalating system EB-DNA was prepared by pre-treatment of EB and CT DNA ( (2)) [38], indicating that EB is replaced by 1-4 in the EB-DNA complex [54]. The obtained K SV values (table 5) are higher than that of free Nadicl and show the tight binding of the complexes to DNA with 3 bearing the highest K SV value (= 4.98(±0.30) × 10 6 M −1 ) among the compounds. The K SV of 1-4 are of the same magnitude as those of a series of metal complexes with NSAIDs as ligands [9][10][11][21][22][23][24][25][26].

Interaction with SAs
The transfer of ions and drugs to cells and tissues through the bloodstream is the main role of the most abundant serum protein, i.e. SA. Within this context and as a preliminary step for potential applications, it is important to investigate the ability of potential drug candidates to bind to SA [55]. The interactions of 1 and 2 with HSA and BSA were monitored by the quenching of the fluorescence emission (due to tryptophan residues) at 351 nm or 343 nm, respectively, upon excitation at 295 nm in the presence of increasing amounts of the complexes. The fluorescence quenching of SA solutions in the presence of the complexes was moderate to significant (quenching up to 73.7% of the initial fluorescence, figure 6). Such quenching is usually assigned to possible changes in secondary structure of SA due to the binding of the compounds to SA [39]. The values of the quenching constant (k q ) for the interaction of the complexes with the albumins were calculated from the Stern-Volmer quenching equation (equations (4) and (5)) and the corresponding Stern-Volmer plots (figures S5 and S6) and are cited in table 6. The k q values are 3.01 × 10 12 − 2.04 × 10 14 M −1 s −1 , significantly higher than the values found for other quenchers interacting with biopolymers (2.0 × 10 10 M −1 s −1 ); therefore, a static quenching mechanism may be indirectly suggested [56]. The k q values are within the range found for a series of complexes previously reported by our lab [9][10][11][21][22][23][24][25][26].    [10]. Similarly, the values of the SA-binding constant (K) of the complexes were calculated from the Scatchard equation (equation (6)) and the corresponding Scatchard plots (figures S7 and S8) and are given in table 6. The K values for the complexes are relatively high and of the same magnitude with a series of NSAIDs-complexes previously reported [9][10][11][21][22][23][24][25][26].  Compound potential transportation. These values are also lower than the association constant of one of the strongest known non-covalent interactions, i.e. avidin with diverse ligands with K ≈ 10 15 M −1 . Thus, the complexes are not too tightly bound to the SAs and they may get released upon arrival at the targets [57].

Antioxidant capacity of the complexes
Manganese superoxide dismutase is a well-known manganese antioxidant and has the role to protect the organism from damaging oxygen-containing radicals by catalyzing the disproportionation of superoxide radical to oxygen and hydrogen peroxide [12,58]. As a continuation of our recent studies concerning the antioxidant activity of metal-NSAID complexes [9][10][11][21][22][23][24][25][26], the scavenging ability of 1 and 2 towards DPPH, ABTS, and hydroxyl radicals and the in vitro soybean LOX inhibition were investigated in regard to well-known antioxidant agents (e.g. BHT, trolox, NDGA, and caffeic acid) used as reference compounds.
Compounds that scavenge DPPH radicals may provide protection against inflammation and rheumatoid arthritis, leading to potentially effective drugs [59]. The DPPH scavenging activities of the complexes are not time-dependent as concluded by measurements performed after 20 and 60 min (table 7). Complexes 1 and 2 exhibit low to moderate ability to scavenge DPPH radicals compared to reference compounds NDGA and BHT and are more effective DPPH scavengers than free Nadicl. The scavenging ability of the compounds against the cationic ABTS radical (ABTS þÅ ) is a marker of the total antioxidant activity [60]. Complex 2 shows significantly higher ABTS radical scavenging activity (ΑΒΤS% = 85(±1)%) than free Nadicl and of the same magnitude to the reference compound trolox (table 7). Hydroxyl radical scavengers may serve as protectors via the activation of the synthesis of the prostaglandins since hydroxyl radicals and other reactive oxygen species are generated during the inflammatory process [59]. Complexes 1 and 2 are more potent hydroxyl radical scavengers than free Nadicl and equally active as the reference compound trolox (table 7). Among the Mn-diclofenac complexes, 4 is the most potent radical scavenger.
LOXs are important enzymes in several allergic and inflammatory diseases and are involved in the transformation of arachidonic acid to leukotrienes. Thus, LOX inhibitors may be excellent antioxidants or free radical scavengers [60] and the in vitro inhibition of the compounds against soybean LOX is as a marker of the antioxidant activity. All Mn(II)diclofenac complexes exhibit noteworthy inhibitory activity against soybean LOX (table 8)   Table 7. % DPPH scavenging ability (RA%, for 0.1 mM), % superoxide radical scavenging activity (ABTS%, for 0.1 mM), and competition % with DMSO for hydroxyl radical ( Å OH%, for 0.1 mM)) for Nadicl and 1-4. NDGA, BHT and trolox are the reference compounds used in the studies. compared to the reference compound caffeic acid (IC 50 = 600 μM), with 2 being the best LOX inhibitor (IC 50 = 24(±1) μM). Mn(II)-diclofenac complexes exhibit higher activity than free sodium diclofenac. The activity of the complexes may be considered selective, especially against hydroxyl and ABTS radicals since they show low to moderate activity against DPPH and high activity against hydroxyl and ABTS radicals [60][61][62][63][64].

Conclusion
The synthesis and characterization of two mononuclear manganese(II) complexes with the NSAID sodium diclofenac in the presence of 2,2′-bipyridine or pyridine has been presented. Spectroscopic, hydrodynamic, and electrochemical techniques were employed to examine the binding of 1 and 2 as well as of previously reported Mn(II)-diclofenac complexes [Mn(dicl) 2 (bipyam)] (3) and [Mn 3 (dicl) 6 (phen) 2 (MeOH)] (4) to CT DNA. The DNA-binding constants of the complexes were calculated by UV spectroscopic titrations with 4 having the highest K b value among the compounds. Measurement of the DNA viscosity and cyclic voltammetry experiments indicated intercalation as the possible interaction mode with CT DNA, a conclusion which was further verified by the EB-displacement ability of the complexes.
Interaction of the complexes with SAs revealed an enhanced quenching ability of the SA fluorescence and higher binding affinity in comparison to free sodium diclofenac providing relatively high binding constants (2.70 × 10 4 − 3.14 × 10 6 M −1 ) and indicating their ability to bind to SA, get transferred by them and be released upon arrival at their targets. Therefore, the studies with SA provide useful information concerning potential application of the complexes and may be expanded to more serum proteins.
The in vitro antioxidant activity of the compounds was studied revealing their ability to scavenge moderately DPPH radicals and significantly ABTS and hydroxyl radicals as well as the high inhibitory activity on soybean LOX; in most cases, the complexes were more active than free Nadicl. According to the existing results from the in vitro biological evaluations of complexes 1-4 might be considered interesting for use as potential manganese metallodrugs. The biological activity of the complexes could be further evaluated by investigating the anti-inflammatory activity and antiproliferative activity in diverse cell lines.

Supplementary material
CCDC 1408663 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB21EZ, UK; Fax: (+44) 1223-336-033).