Dinuclear Zn Complex: Phenoxyl Radical Formation Driven by Superoxide Coordination

The dinuclear zinc complex Zn2L2 (HL=2-{[[di(2-pyridyl) methyl](methyl)amino]methyl}phenol) has been synthesized and isolated as colorless crystals. The interaction of the compound with superoxide in anhydrous organic solvents has been evaluated by CV, stopped-flow Uv-vis, EPR and ESI-MS suggesting the binding and the activation of the coordinated superoxide, thanks to the Lewis acidity of the Zn(II) centers. The results obtained in this study highlight the formation of a Zn2L2 O2 * intermediate and a metastable phenoxyl-radical driven by the coordinated superoxide.


Introduction
Molecular dioxygen O 2 is a fundamental component for aerobic life. However, during the biological processes, O 2 is converted into reactive oxygen species (ROS) including the superoxide radical anion (O 2 * À ), hydrogen peroxide (H 2 O 2 ), and the hydroxyl radical (HO * ). Once generated, the superoxide radical might react with redox active metal centers starting a cascade of reactions that lead to the formation of HO * , a highly reactive specie responsible for the unselective oxidation of proteins, DNA and other biomolecules. These phenomena are responsible for the so-called "oxidative stress", a condition strongly related to several pathologies like Alzheimer and cancer. [1,2] The natural defence against ROS is provided by the combined action of antioxidant enzymes. A cascade detoxification is initiated by superoxide dismutases (SODs) via O 2 * À dismutation to O 2 and H 2 O 2 and completed by catalases and glutathione peroxidase, to convert H 2 O 2 to O 2 . [3] In particular the Cu,Zn-SOD is an enzyme based on the Cu(II) active center that efficiently dismutate O 2 * À cycling between the Cu II/I redox states. [4] The zinc ion instead appears to play a role in the overall folding stability and facilitates a broader pH independence. However, the first example of a Zn-based complex, bearing a redox active ligand able to catalytically dismutate superoxide has been recently reported. [5] Another well-established function of the Zn(II) center in the Cu,ZnÀ SOD, is to act as Lewis acid. In fact, upon binding of the superoxide to form a ZnÀ O 2 * À adduct, both the processes of reduction and oxidation of the superoxide become accelerated. [6] Herein the reactivity of the Zn 2 L 2 complex, a structural analogue of the highly reactive SOD mimic Cu 2 L 2 , with O 2 * À has been investigated in order to evaluate the interaction and the contribution of the Zn center in the process of superoxide activation. [7]

Results and Discussion
The ligand HL (2-{[[di(2-pyridyl)methyl](methyl)amino]methyl} phenol) has been synthesized according to the literature procedures. [8] This ligand has been recently reported in the literature to form mono and dimeric complexes with first row transition metals (Mn, Cu, Fe) exhibiting excellent properties in the processes of H 2 O 2 and O 2 * À dismutation and proton reduction. [7,9] The tetradentate N 3 O donor-set provided by the ligand can bind Zn(II) with a bis-pyridyl, tertiary amine and phenolate terminals. The Zn 2 L 2 complex has been obtained by reacting the ligand HL with zinc perchlorate and triethylamine in methanol (Scheme 1). [7,9] Colourless crystals suitable for X-ray analysis have been isolated upon addition of acetonitrile and cooling to 4°C. The X-ray analysis of Zn 2 L 2 shows a C i point-group symmetry for the complex, with a distorted octahedral geometry for the metal centers, analogously to the structure previously reported for the Cu 2 (II)L 2 complexes (Figure 1). [7] Each Zn(II) atom is coordinated in a facial configuration by the three N atoms of the tetradentate ligand, while the phenolate ligands act as a bridge between the two metal centers. The metal coordination sphere is completed by two MeOH molecules (Figure 1). From the crystallographic structures, minor differences were observed in the M 2 O 2 core between the Zn and Cu complexes: by way of comparison the ZnÀ Zn and CuÀ Cu distances were found to be 3.129 Å and 3.007 Å, respectively, whereas the ZnÀ OÀ Zn and CuÀ OÀ Cu angles differ only slightly, being 99.9°and 100.2°, respectively. The key difference between the Cu 2 L 2 and Zn 2 L 2 structures is related to the lack of the Jahn-Teller distortion in Zn 2 L 2 , resulting in a relatively strong coordination of the MeOH molecules ( Figure 1; for the structural parameters see the Supplementary Information). However, in solution the coordinated solvent can be easily exchanged, allowing the access of a substrate to the zinc sites. The FT-ATR spectrum of Zn 2 L 2 confirms the ligand coordination, as the pyridines and phenol absorption bands are shifted towards higher frequencies upon metal binding and found at 1607-1599 cm À 1 (Figures S2 and  S3). [10] The retention of the dinuclear structure in solution has been confirmed by ultra-high resolution cold-spray ionization mass spectrometry (UHR-CSI-MS). The MS peaks were detected for the expected molecular ion and for the formate adducts (formic acid was present in the eluent) with m/z = 368.0738 [Zn 2 L 2 ] 2 + and 781.1424 [Zn 2 L 2 + HCO 2 ] + ( Figure S11). The UV-Vis spectrum was collected for Zn 2 L 2 (100 μM) in CH 3 CN ( Figure S4). The absorption bands between 260-290 nm are ascribable to the π-π* transitions of pyridines and phenolate ( Figure S4).
Concerning the electrochemical characterization, even if Zn 2 L 2 contains the non-redox active metal centers, its redox properties has been addressed by cyclic voltammetry (CV) experiments, in order to evaluate the ligand-based processes. Under reductive scan in acetonitrile, as expected, the complex did not show any redox features ( Figure S8). In the oxidative scan instead, two irreversible waves have been detected at E 1 a =340 and E 2 a =633 mV (vs Fc/Fc + ) and attributed to the oxidation of the coordinated phenolates to phenoxyl radicals ( Figure S9 and S10). [11] These relatively high redox potentials should hinder the reaction of the complex with mild oxidants like O 2 * À via an outer-sphere electron transfer (ET). The SOD activity of Zn 2 L 2 has been tested in water media upon KO 2 addition by stopped-flow. However, the complex does not exhibit any effect on the rate of superoxide dismutation. As known in literature, in aqueous media, the key step in the superoxide dismutation process is represented by the protonation of the superoxide. [12] On the other hand, superoxide is quite stable in organic aprotic solvents. For this reason, we used dry dimethyl sulfoxide (DMSO) and acetonitrile to investigate the interaction between Zn 2 L 2 and O 2 * À . First of all, we monitored the reaction between Zn 2 L 2 (25 μM) and an excess of KO 2 (4 mM) by stopped-flow UV-vis in DMSO at room temperature. Immediately after mixing the appearance of a new band at 435 nm has been detected ( Figure 2). [11] This spectral feature, attributable to the formation of a phenoxyl radical on the organic ligand moiety, quickly decays after its formation ( Figure 2). As reported above, the redox potential of the Zn 2 L 2 /Zn 2 L 2 + redox couple precludes an oxidation process occurring via an outer sphere ET suggesting instead an inner sphere mechanism involving the superoxide coordination to the Zn center followed by the oxidation of the chelating ligand. The same experiment performed on the free ligand indeed did not show the formation of any band compatible with the formation a phenoxyl radical ( Figure S5). The broad band observed for the reaction of the free ligand at ca. 340 nm, like previously reported in literature for various phenols, is associated to a radical based degradation including the coupling of the phenolic moiety and the addition/ condensation of the superoxide on it. [13] For longer rection time  (ca. 10 min), after the decay of the phenoxyl band, a similar trend can also be observed for Zn 2 L 2 ( Figure S6). The UV-vis band, associated to the phenoxyl radical (425 nm, ɛ = 7.5 · 10 3 ), has been also obtained upon addition of 1 eq of ceric ammonium nitrate (CAN) to an acetonitrile solution of the complex ( Figure S7). [11] The radical generated in this way exhibits a longer lifetime (t 1/2 = 25 min) in comparison with the one generated in presence of an excess of KO 2 . The shorter lifetime of the latter is expected due to the additional reactions in which the phenoxyl radical may be involved in the presence of a superoxide excess. [13] However, the high reactivity of the phenoxyl radical does not allow its detection by MS, even when it has been directly generated with CAN.
Once formed, the phenoxyl radical in fact can be reduced again to phenolate by an additional equivalent of superoxide or couple with another phenoxyl or superoxide radical generating the rise of the broad band observed below 400 nm ( Figure S6). This evidence proofs the crucial role played by the Zn center in promoting the formation of a phenoxyl radical and also in increasing its lifetime. [14] To further support the effective interaction and coordination of the O 2 * À to the Zn 2 L 2 complex, the Zn 2 L 2 /O 2 * À solution has been analyzed by CV, EPR and ESI-MS. The first hint of the effective coordination of superoxide to the Zn-core of the complex has been provided by CV experiments in presence of KO 2 . The performed reductive scan shows the presence of two characteristic waves associated to the formation of two intermediates in solution ( Figure 3). The two irreversible cathodic waves peaked at À 1.49 V and À 2.10 V (vs Fc/Fc + ) arise respectively from the reduction of O 2 to O 2 * À and O 2 * À to O 2 2À (Figure 3). [6] It is important to underline that the O 2 /O 2 * À reduction is shifted to a higher potential (less negative) in presence of the Zn complex ( Figure 3) in comparison with the reduction of the free O 2 , peaked at À 1.54 V (vs Fc/Fc + ). In addition, in presence of Zn 2 L 2 , in the reverse anodic scan a reoxidation of O 2 * À to O 2 was not observed. However, a new wave appeared at lower potential (À 2.10 V), indicative of the further reduction of the coordinated O 2 * À to O 2 2À , a feature lacking in absence of Zn 2 L 2 . As previously reported, this speaks in favor of an electrostatic interaction between the Zn(II)-complex and the superoxide anion. [6] The Zn center in fact acts as a Lewis acid, tuning the redox properties of the coordinated substrate and in our case allows the electrochemical reduction of superoxide to peroxide and the chemical oxidation of the phenolate to phenoxyl radical. The phenol oxidation in particular appears very interesting considering that this process usually requires very strong chemical oxidants like CAN. [15] A further confirmation of the effective coordination of the O 2 * À to the Zn-center comes from CW X-band EPR measurements. Adding 5 eq of KO 2 (2.5 per Zn center) at room temperature to a Zn 2 L 2 solution (1 mM) in DMSO, and immediately freezing the solution in liquid nitrogen, the formation of an intense signal characteristic for a Zn-superoxo intermediate has been detected. [6] The EPR spectrum measured at 95 K shows the formation of the desired Zn(II)À O 2 * À complex overlapped with the spectrum of the free KO 2 , added in excess ( Figure 4a, blue trace). Collecting the EPR spectrum at higher temperature (240 K, Figure 4a green trace) only the anisotropic signal associated to the coordinated superoxide Zn(II)À O 2 * À has been detected. This allowed us to simulate the EPR spectrum and to determine the g-values for the coordinated superoxide g // = 2.06 and g ? = 2.01 (Figure 4b), which are significantly smaller than those measured for the metal-free O 2 * À (g // = 2.12 and g ? = 2.00, Figure S13), due to the binding to the Zn(II) ion. [17] By way of comparison, 1 eq of CAN has been added to the complex solution in acetonitrile to promote the phenoxyl radical formation. The obtained EPR spectrum of the frozen solution ( Figure S14), shows an isotropic signal (S = 1/2) g iso = 2.00, associated to the phenoxyl radical. [11] The marked differences between the EPR spectra obtained with KO 2 and CAN further confirm the formation of two distinct species and exclude a significant contribution of phenoxyl radical in the spectrum of the Zn(II)-O 2 * À complex under these experimental conditions. The deviation observed for the g // value from the free spin value (g e = 2.0023) is caused by the spin-orbit interaction as given in Eq. 1 when ΔE @ λ. Where λ is the spinorbit coupling constant (0.014 eV) and ΔE is the energy splitting between the 2pπ x * and the 2pπ y * levels of the superoxide due to the metal binding. [17,18] (1) Accordingly, the g // value obtained can be used to determine the ΔE value for the superoxide in the ZnÀ O 2  Thus, the EPR results follow the CV data confirming a more favorable reduction of the coordinated superoxide thanks to the Lewis acidity of the Zn center ( Figure 3).
Finally, the CSI-MS analysis has been performed on a Zn 2 L 2 solution in acetonitrile, upon addition of an excess of solid KO 2 (see Experimental Section), in order to determine the nuclearity of the Zn-O 2 * À complex. The experiment highlights the formation of a dinuclear Zn-superoxo complex Zn 2 L 2 -O 2 * À with m/z = 768.1368 ( Figure S12). The retainment of the dinuclear structure in presence of the coordinated superoxide is remarkable and constitute, to the best of our knowledge, the first example of dinuclear-Zn-superoxide complex reported up to date.

Conclusions
In summary, a new dinuclear zinc complex Zn 2 L 2 has been synthesized and fully characterized. The reaction of the complex with superoxide has been addressed by UV-vis, CV, EPR and CSI-MS revealing the formation of a Zn 2 L 2 -O 2 * À intermediate. The coordinated superoxide is activated by the Lewis acidity of the Zn center and its reduction comes out to be more favorable in comparison with the free superoxide, as evidenced by CV and EPR analysis. In addition, the inner-sphere ET occurring between the ligand and the coordinated O 2 * À allows the oxidation of the phenolic moiety of the ligand to phenoxyl radical, a process that cannot occur via an outer-sphere ET due to the high redox potential of the phenolate/phenoxyl radical couple. These results pave the way for the reactivity study of the Zn 2 L 2 À O 2 * À intermediate obtained, in order to determine its electrophilic properties and evaluate its activity in the oxidation of external substrates like phosphines and other standard organic molecules.

Experimental Section
Materials and Methods: Commercially available reagents were used directly, unless otherwise noted. The ligand precursors 2-{[[di(2-pyridyl)methyl](methyl)amino]methyl}phenol (HL) was synthesized according to the literature procedures. [8] UV-Vis spectra were recorded with an Analytik Jena AG Specord 200 UV-VIS-NIR spectrometer. FT-IR spectra were recorded Shimadzu ATR MIRacle 10 as solid samples from a diamond crystal. Cyclic voltammetry measurements were performed using an Autolab instrument with a PGSTAT 101 potentiostat. A three-electrode arrangement was employed consisting of a glassy carbon disk working electrode (A = 0.07 cm 2 ) (Metrohm), a platinum counter electrode (Metrohm) and a silver wire (Metrohm) as reference electrode. Potentials were referred to a Fc/Fc + redox couple. Prior to each experiment, the electrode was polished with 1 μm alumina, rinsed with deionised water and wiped with a paper tissue. The CSI-MS measurements were performed on a UHR-TOF Bruker Daltonik maXis plus, an ESI-quadrupole time-of-flight (qToF) mass spectrometer capable of a resolution of at least 60.000 (FWHM), which was coupled to a Bruker Daltonik Cryospray unit. Detection was in positive ion mode, the source voltage was 3.5 kV and the flow rate was 240 μL/hour. The temperature of the spray gas (N 2 ) was held at À 40°C and the temperature of the dry gas for solvent removal was kept at À 35°C. The mass spectrometer was calibrated prior to every experiment via direct infusion of Agilent ESI-TOF low concentration tuning mixture, which provided an m/z range of singly charged peaks up to 2700 Da in both ion modes. Applied solvents were not extra dry and could provide a source of protons. Processing of the measured data was done with Bruker DataAnalysis. EPR spectra were recorded on a JEOL continuous wave spectrometer JES-FA200, equipped with an X-band Gunn diode oscillator bridge, a cylindrical mode cavity, and an N 2 cryostat. The samples were measured in solution, under nitrogen atmosphere, in quartz glass EPR tubes at 95 and 240 K, respectively. The spectra shown were measured using the following parameters: microwave frequency ν = 8.959 GHz, modulation width 0.1 mT, microwave power 1.0 mW, modulation frequency 100 kHz, time constant 0.1 s. Analysis and simulation of the data was carried using the software "eview" and "esim", written by E. Bill (MPI for Chemical Energy Conversion, Mülheim an der Ruhr). [16] Caution ! Although the salts and the complexes reported do not appear to be mechanically sensitive, perchlorates should be treated with due caution.

X-ray Diffraction:
A colourless crystal of the composition [Zn 2 L 2 (MeOH) 2 ](ClO 4 ) 2 was embedded in inert perfluoropolyalkylether (viscosity 1800 cSt; ABCR GmbH) and mounted using a Hampton Research CryoLoop. The crystal was then flash cooled to 100.0(1) K in a nitrogen gas stream and kept at this temperature during the experiment. The crystal structure was measured on a SuperNova diffractometer with Atlas S2 detector using a CuKα microfocus source. The measured data was processed with the CrysAlisPro (v40.53) software package. [19] Using Olex2, [20] the structure was solved with the ShelXT [21] structure solution program using Intrinsic Phasing and refined with the ShelXL [22] refinement package using Least Squares Minimization. All non-hydrogen atoms were refined anisotropically. Most hydrogen atoms were placed in ideal positions and refined as riding atoms with relative isotropic displacement parameters. The position of H2, which is connected to O2 of the methanol ligand, was observed from difference Fourier maps and refined. The crystal structure data has been deposited with the Cambridge Crystallographic Data Centre. CCDC 2036662 contains the supplementary crystallographic data for complex [Zn 2 L 2 (MeOH) 2 ](ClO 4 ) 2 . This data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Crystallographic and refinement data are summarized in Tables S1-S3 (see Supporting Information).

UV-vis measures with CAN:
The reaction between Zn 2 L 2 (100 μM) and 1 eq of CAN has been monitored in dry acetonitrile (every 60 s for 2 h) by using an Analytik Jena AG Specord 200 UV-VIS-NIR spectrometer.
CSI-MS Experiments with KO 2 : The Zn 2 L 2 -O 2 * À complex have been obtained by adding solid KO 2 to a 0.1 mM solution of Zn 2 L 2 in acetonitrile previously cooled down to À 40°C.