Synthesis and Characterization of Cobalt(III), Nickel(II) and Copper(II) Mononuclear Complexes with the Ligand 1,3-bis[(2-aminoethyl)amino]-2-propanol and Their Catalase-Like Activity

In this work, we present the synthesis and characterization of two new mononuclear complexes with the ligand 1,3-bis[(2-aminoethyl)amino]-2-propanol (HL), [Co(L)(H2O)](ClO4)2 (1), [Ni(HL)](ClO4)2 (2), as well as the known complex [Cu(HL)](ClO4)2 (3) for comparison. Their abilities to catalyze the dismutation of H2O2 and the oxidation of cyclohexane were investigated. The complexes were characterized by X-ray diffraction, elemental analysis, electronic and infrared spectroscopy, cyclic voltammetry, electrospray ionization mass spectrometry (ESI-MS) and conductivity measurements. The X-ray structures showed that the nickel (2) and copper (3) complexes are tetracoordinated, with the metal ion bound to the nitrogen atoms of the ligand. On the other hand, the cobalt complex (1) is hexacoordinated, possessing additional bonds to the alkoxo group of the ligand and to a water molecule. Neither of the complexes was able to catalyze the oxidation of cyclohexane, but all of them exhibited catalase-like activity, following Michaelis-Menten kinetics, which suggest resemblance with the catalase natural enzymes. The catalytic activity followed the order: [Ni(HL)](ClO4)2 (2) > [Cu(HL)](ClO4)2 (3) > [Co(L)(H2O)](ClO4)2 (1). As far as we know, this is the first description of a nickel complex presenting a significant catalase-like activity.


Introduction
The coordination between a metal ion and peroxide plays an important role in many biological systems [1]. Metalloenzymes such as methane monooxygenase (MMO) and catalase are two examples of dinuclear proteins known or believed to share a peroxide adduct during their catalytic cycle [2]. MMO is responsible for oxidizing methane into methanol at mild conditions, as well as others hydrocarbons and halocarbons [3]. The oxidation starts when a dioxygen molecule is activated in a two-electron oxidation process. On the other hand, catalase is responsible for the biological defense against hydrogen peroxide by its conversion to water and dioxygen (Eq 1). The dismutation process involves a two-electron transfer from the hydrogen peroxide to a diiron-peroxide adduct [4].
Two forms of active sites are found in MMO, a binuclear iron center in the soluble form of the enzyme [2] and mononuclear/binuclear copper centers in the protein-bound form [3]. Similarly, there are mostly two kinds of catalase enzymes, the most abundant possess an heme-type structures that contain an iron(III)-protoporphyrin IX prosthetic group in the active site, and can be found in almost all aerobic organisms. The second is a recently discovered class of manganese catalase that possesses a binuclear manganese center and occurs only in bacteria [5]. Lately, functional models have shown potential biomedical application as therapeutic agents against oxidative stress. Despite the natural enzymes possess iron or manganese in their active sites, complexes of other metals have being investigated as catalase models, such as copper [6][7][8][9][10][11][12][13][14][15][16], cobalt [15,17] and ruthenium [18].
In this work, we present the syntheses and characterization, including the X-ray crystal structure, of two new complexes with the ligand 1,  (2). The known complex [Cu(HL)](ClO 4 ) 2 (3) [19,20] was also studied for comparison. The catalase-like and MMO towards the oxidation of cyclohexane with H 2 O 2 , were also investigated. As far as we know, this is the first study revealing a significant catalase-like activity of a nickel complex.
Infrared spectra were collected on a FTIR Nicolet Magna-IR 760 spectrophotometer (KBr or CsI pellets or film between NaCl window).
UV-Vis spectra were recorded on a Varian Cary 1E spectrophotometer in water solution. Infrared spectra were collected on a FTIR Nicolet Magna-IR 760 spectrophotometer (KBr or CsI pellets or film between NaCl window). Conductivity measurements were carried out with solutions containing 1.0×10 −3 mol dm -3 of the complexes, using a BioCristal NT CVM conductivimeter, employing a conductivity cell CA150.
Electrospray ionization mass spectrometry (ESI-MS) measurements were performed on a high resolution ESI-TOF (micrOTOF, Bruker Daltonics, Bremen, Germany) mass spectrometer. The compound to be tested was dissolved in methanol prior to analysis. Full scans were acquired under the following conditions: capillary, 5.5 kV, capillary exit, 100 V. The spectra were obtained in positive-ion mode.
Cyclic voltammetry experiments were carried out in water using a BAS Epsilon potentiostat/galvanostat and a three-electrode system, consisting of a glassy carbon disk as the working electrode, a platinum wire as the auxiliary electrode and a Ag/AgCl system as the reference electrode. A 0.1 mol dm -3 solution of lithium perchlorate was used as supporting electrolyte and K 3 [Fe(CN) 6 ] (E 1/2 = 0,254 V versus Ag/AgCl; ΔE = 347 mV) was used as internal standard.  [22]. Gas chromatography analyses were conducted on a HP5890 gas chromatograph with a HPDB5 column (30 m × 0.25mm × 0.25 μm) connected to a FID detector, using H 2 (140 kPa) as carrier gas. The analysis conditions for the cyclohexane oxidation reactions were: initial temperature of 50°C, heating ramp of 1.5°C /min to 56°C, then heating ramp of 10°C /min to the final temperature of 127°C. The injector and detector temperature were 200°C C and 250°C, respectively. Products were identified by their mass spectra and the retention times were compared with those of authentic samples. Quantification was made through calibration plots for the detector response of the authentic samples.
The catalase-like activities were followed by measuring the volume of O 2 produced by H 2 O 2 disproportionation reactions. The total reaction volume was kept constant during all experiments at 5.0 mL. The reactions were performed at 25°C, using the assistance of a water bath and a thermostat. TRIS buffer was used as solvent. The buffer pH was adjusted to 7.2 with HCl. The reactor was a kitassato flask (25 cm 3 ) magnetically stirred and closed with a rubber septum. The kitassato was connected to an inverted graduate burette filled with water. Hydrogen peroxide solution (commercial 30% aqueous solution) was injected through the septum with a syringe and the dioxygen production was measured in the burette at appropriate times. The experimental data were plotted in a curve describing the amount of dioxygen evolved versus time.
Caution! The perchlorate salts used in this study are potentially explosive and should be handled with care!

X-ray diffraction experiments
The X-ray data for [Co(L)(H 2 O)](ClO 4 ) 2 (1) and Ni(HL)ClO 4 (2) were collected from a Bruker KAPPA CCD diffractometer [23], using a selected single crystal at 295 K and MoKα monochromatic-graphite radiation. The cell parameters for the complexes were obtained using the PHI-CHI and DIRAX programs [24,25]. The average data were reduced using the EvalCCD program and the absorption correction was performed with the SADABS programs [26,27]. The structure was solved by direct methods via SHELXS97 and refined via SHELXL97 by a fullmatrix least-squares treatment with anisotropic temperature parameters for all non H atoms [28].
For  (2), suggests the presence of racemic twin structures. In this way the crystalline refinement show an inversion twin, with orientation matrices assigned to the twin component, (-100, 0-10, 00-1) twin law. The finally refined ratio of the twin components shows a Flack parameter being 0.18(2): 0.82 (2).
The crystal data are listed in Table 1 and in Supporting Information.

Reactivity studies
Catalase-like activity and kinetics of H 2 O 2 dismutation. In order to study the kinetics of H 2 O 2 dismutation, two sets of experiments were carried out in TRIS buffer solution at pH 7. In a first set of experiments, the initial concentration of H 2 O 2 was kept constant (2.19 mol dm -3 for complex (1) and (3); 1.50 mol dm -3 for complex (2)) while the initial concentration of the catalyst was varied in the ranges indicated in Table 2. In a second set of experiments, the initial concentration of H 2 O 2 was varied while the initial concentration of catalyst was kept constant and equal to 3.0 × 10 −3 mol dm -3 . Table 2 summarizes the H 2 O 2 and catalyst concentrations values used in these experiments.
The dioxygen evolution was measured volumetrically, the reactions were carried out at 25°C, using the assistance of a water bath and a thermostat. The total volume of the reaction solution was 5.0 mL. The reactor was a kitassato flask (25 cm 3 ) magnetically stirred and closed with a rubber septum. The kitassato was connected to an inverted graduated burette filled with water. Hydrogen peroxide solution (commercial 30% aqueous solution) was injected through the septum with a syringe and the dioxygen production was measured in the burette at appropriate times. The experimental data were plotted in a curve describing the amount of dioxygen evolved versus time. Observed initial rates were expressed as mol(O 2 ) s -1 and calculated from the maximum slope of the curves describing the O 2 evolution versus time. All experiments were done, at least, in triplicate and the reported values are average values.

Cyclohexane oxidation
Cyclohexane oxidation tests followed published procedure [29]. The reactions were carried out for 24 hours in CH 3 CN or H 2 O as solvent, at room temperature, under argon atmosphere, using H 2 O 2 as oxidant and the complexes as catalysts. The catalyst:substrate:oxidant proportion used was 1:1000:1000, and the catalyst concentration was 7.0×10 −4 mol dm -3 . The reactions were quenched by the addition of an aqueous 0.4 M Na 2 SO 4 solution, followed by extraction with 10 mL of diethyl ether. The organic layer was dried with anhydrous Na 2 SO 4 and analyzed by gas chromatograph.

Results and Discussion Syntheses
The ligand 1,3-bis[(2-aminoethyl)amino]-2-propanol was synthesized from epichlorohydrin and ethylenediamine, as described in literature [21]. The ligand has four aliphatics nitrogen atoms and one oxygen atom that can coordinate to the metallic center. The complexes were  [19,20], unit cell parameters of obtained single crystals were in accordance with the published data. The three complexes are mononuclear, cationic, with two perchlorate as counter-ions. where the cobalt(III) ion adopts a distorted octahedral geometry by the coordination to the ligand and one water molecule. The ligand is coordinated by the four nitrogen atoms, which occupy the equatorial plane, as well as by the deprotonated alkoxo oxygen, which occupies the axial position trans to the water molecule. The unit cell (Fig 3) contains three crystallographic independent complex cations and six

Cyclic Voltammetry
The electrochemical behavior of the complexes was investigated in water (Figs 7-9). The voltammograms of complexes [Co(L)(H 2 O)](ClO 4 ) 2 and [Ni(HL)](ClO 4 ) 2 showed only one process and the redox potential values are presented in Table 5. For complex [Co(L)(H 2 O)] (ClO 4 ) 2 , the cathodic wave was observed at -0.641 V vs. Ag/AgCl (-0.432 V vs. NHE) and the anodic wave was observed at 0.837 V vs. Ag/AgCl (1.046 V vs. NHE). This can be considered an irreversible process, which is a typical electrochemical behavior for cobalt complexes due to spin shifts from low spin Co(III) to high spin Co(II). The reduction wave can be assigned to the one-electron reduction process from Co(III) ! Co(II) and the anodic wave can be assigned to the reverse process [39]. The voltammogram of [Ni(HL)](ClO 4 ) 2 presented only one process  6 ], this process can be considered as a quasireversible system since i pa /i pc (1.28) is different from one and ΔE changes with the scan rates. This process may be assigned to the oxidation of Ni(II) to Ni(III) and the reverse scan to the reduction of Ni(III) to Ni(II) [40].
The electrochemical behavior of complex [Cu(HL)](ClO 4 ) 2 was different from the other complexes, when scanning over -0.2 V, the voltammogram presented two cathodic and two anodic waves. But when the scan was reversed prior to the second reduction, only one process could be observed, as showed in the inlet of Fig 9. This process presented the cathodic peak at -0.032 V vs. Ag/AgCl (0.177 V vs. NHE-I) and the anodic peak at 0.049 V vs. Ag/AgCl (0.258 V vs. NHE-I´), which can be attributed to the redox couple Cu(II)-Cu(I). Scanning over -0.2 V, another cathodic peak (II) was observed. This peak shifted with the scan rate and can be tentatively attributed to the reduction of Cu(I) to metal copper. The anodic wave (II´) at 0.104 V vs. Ag/AgCl (0.313 V vs. NHE) might be assigned as the redissolution of the metal copper formed previously [41,42]. The electrochemical equilibrium between processes I and Iḿ ay be very slow, since at fast scan rates the anodic peak I´disappeared. Alternatively, process II´may involve so much current that occulted process I´. The three mononuclear complexes were tested as catalysts for the dismutation of H 2 O 2 and for the oxidation of cyclohexane, using H 2 O 2 as oxidant. Neither of the complexes presented satisfactory yields in the formation of cyclohexanol and cyclohexanone (S1 Table), but the three complexes exhibited catalase-like activity under the studied conditions. The oxidations yields and the kinetic parameters are summarized in Table 6. Control experiments with Co(ClO 4 ) 2 , Ni(ClO 4 ) 2 and Cu(ClO 4 ) 2 at 3.0×10 −3 M and H 2 O 2 3.0 M showed no hydrogen peroxide decomposition.
The H 2 O 2 decomposition was followed measuring the dioxygen evolution during the reaction. The initial rates method was applied in order to determine the kinetic parameters. The initial rate values were calculated as the maximum slope of the curves of mol(O 2 ) versus time. The plots of initial rates versus hydrogen peroxide concentration showed saturation kinetics in relation to the substrate, indicating a Michaelis-Menten catalytic behavior (Eq 2), similar to the natural catalase enzymes. Besides that, the solvent used in the experiments was water, which is closer to the medium of the enzymatic reactions and consequently provides more accurate data for comparison. The saturation graphics are presented in Fig 10.  To calculate the Michaelis-Menten parameters (Table 6), the more reliable nonlinear least square fit was used instead of a linear fit, following the recommendation of different authors [43,44]. [Ni(HL)](ClO 4 ) 2 (2) presented the higher affinity to the substrate (1/K M ), followed closely by complex [Co(L)(H 2 O)](ClO 4 ) 2 (1), but [Cu(HL)](ClO 4 ) 2 (3) presented a very low affinity, with K M value roughly five-times higher in relation to the other complexes. Complex (2) also presented the best catalytic efficiency (k cat /K M ) between the three complexes, however, it was still very low if compared with the natural enzymes [45].
Actually, there are few catalase-like activity studies in water because they are usually hampered by the low solubility of most of the complexes or even by the low activity in aqueous solution. In this sense, the results described herein are promising when compared with the previously reported because it shows the activity in a medium similar to that where the natural enzymes work, i.e., aqueous solution and physiological pH. Additionally, as far as we know, complex (2) was the first nickel complex reported presenting a significant catalase-like activity. Siegel at al. reported that the complex [Ni(en) 3 ] 2+ , has a rate constant, k, equal to 3.5×10 −11 dm 3 mol -1 s -1 , which is lower than the catalase activity of Ni 2+ (aq) (k = 1.77×10 −9 dm 3 mol -1 s -1 ) [47]. These authors have also studied Ni(II) complexes with 2,2'-bipyridyl, 2-picolylamine, 4-aminomethylimidazole, and histamine. The experimental data for the first Synthetic Models of Catalase complex (see Figure 5 in the Siegel at al. article [47]) show that the catalase activity follows [Ni 2+ (aq)] and the other complex do not shown catalase activity at all. Then, the relatively high activity promoted by complex 2 should encourage additional studies on nickel synthetic models for catalase.

Conclusions
Three mononuclear complexes were synthesized, characterized and their catalase-like activity were investigated. The kinetic studies were conducted in a very similar condition to the biological enzymes, i.e., aqueous solution and neutral pH. All complexes obeyed Michaelis-Menten kinetics and the following order of activity was found:  (Table A), atomic coordinates, equivalent isotropic displacement parameters (Table B), bond lengths, angles (Table C), anisotropic displacement parameters (Table D), and selected intermolecular interactions parameters (Table E) (Table A), atomic coordinates, equivalent isotropic displacement parameters (Table B), bond lengths, angles (Table C), anisotropic displacement parameters (Table D), and selected intermolecular interactions parameters (Table E)