Disproportionation of H2O2 Mediated by Diiron-Peroxo Complexes as Catalase Mimics

Heme iron and nonheme dimanganese catalases protect biological systems against oxidative damage caused by hydrogen peroxide. Rubrerythrins are ferritine-like nonheme diiron proteins, which are structurally and mechanistically distinct from the heme-type catalase but similar to a dimanganese KatB enzyme. In order to gain more insight into the mechanism of this curious enzyme reaction, non-heme structural and functional models were carried out by the use of mononuclear [FeII(L1–4)(solvent)3](ClO4)2 (1–4) (L1 = 1,3-bis(2-pyridyl-imino)isoindoline, L2 = 1,3-bis(4′-methyl-2-pyridyl-imino)isoindoline, L3 = 1,3-bis(4′-Chloro-2-pyridyl-imino)isoindoline, L4 = 1,3-bis(5′-chloro-2-pyridyl-imino)isoindoline) complexes as catalysts, where the possible reactive intermediates, diiron-perroxo [FeIII2(μ-O)(μ-1,2-O2)(L1-L4)2(Solv)2]2+ (5–8) complexes are known and well-characterized. All the complexes displayed catalase-like activity, which provided clear evidence for the formation of diiron-peroxo species during the catalytic cycle. We also found that the fine-tuning of iron redox states is a critical issue, both the formation rate and the reactivity of the diiron-peroxo species showed linear correlation with the FeIII/FeII redox potentials. Their stability and reactivity towards H2O2 was also investigated and based on kinetic and mechanistic studies a plausible mechanism, including a rate-determining hydrogen atom transfer between the H2O2 and diiron-peroxo species, was proposed. The present results provide one of the first examples of a nonheme diiron-peroxo complex, which shows a catalase-like reaction.


Introduction
Nonheme manganese catalases (MnCAT), as alternatives to the heme-containing catalases, contain a binuclear active sites and deplete hydrogen peroxide in cells through a ping-pong mechanism interconventing between Mn 2 (II,II) and Mn 2 (III,III) states in a two-electron catalytic cycle [1,2]. Manganese catalase enzymes such as Lactobacillus plantarum (LPC) [3,4], Thermus thermophilus (TTC) [5,6], Thermoleophilium album (TAC) [7], and Pyrobaculum calidifontis VA1 (PCC) [8] have been isolated and characterized from both bacteria and archae. Despite their large numbers and importance, few catalase enzyme structures are known and they can also be sub-divided into two groups, each with a different active site.
Among them both LPC and TTC have a dimanganese catalytic core with Glu 3 His 2 ligands coupled by two single-atom solvent bridges (Scheme 1A) [4,9]. In contrast to the above structure, the cyanobacterial manganese catalase (KatB) from Anabaena shows a markedly different active site with canonical Glu 4 His 2 coordination geometry and two terminal water molecules (Scheme 1B). However, it closely resembles the active site of the ferritin-like ruberythrin (RBR), which is a nonheme diiron protein [10]. The Mn-Mn distances (~3. 8 Å) in KatB are similar to that was observed for the iron-containing ribonucleotide reductase R2 (3.6-3.7 Å), but much shorter distances have been found for LPC and TTC (3.0-3.1 Å). The H 2 O 2 binds terminally to one of the Mn center in LPC/TTC, while in KatB and RBR, the symmetrical active sites and the relatively long M-M distances favor its µ-1,2 binding mode [11]. These enzymes, coupled with superoxide dimutates (SOD), are responsible for the full detoxification of O 2 •through the formation of H 2 O 2 and its redox disproportionation into water and dioxygen. H 2 O 2 can also be classified as a ROS particle, since it readily reacts with reduced transition metals and results in highly reactive HO • radical via Fenton-like reactions. Many rationally designed synthetic biomimetics are being developed against oxidative stress to treat various diseases, including Alzheimer's, cancer, aging, neurodegenerative, inflammatory and heart diseases [12][13][14][15][16][17][18][19][20][21]. The identification and study of the structure of the above enzymes can greatly contribute to the design of new catalase mimetics for potential therapeutic use. The structural and functional investigations on low molecular-weight copper, manganese and iron containing model compounds that are focused mainly on the effect of the used metals and their ligand environment (donor sites) on redox potential and through it on reactivity. Based on the literature data, various Schiff base, aminopyridine and azamacrocyclic derivatives have proven to be the most popular and practical ligands [12][13][14][15][16][17][18][19][20][21]. Considering the reactivity of manganese catalases (KatB) toward H 2 O 2 , including a ping-pong mechanism interconventing between Mn 2 (II,II) and Mn 2 (III,III) states in a two-electron catalytic cycle, a similar interaction with H 2 O 2 has not been reported for diiron(III) peroxo intermediates which may be proposed in the catalytic cycles of RBR. Progress in recent years has made a number of diiron(III)-peroxo complexes available that can be generated by stoichiometric amount of H 2 O 2 [22][23][24][25][26][27]. Therefore, the opportunity has arisen to investigate both electrophilic and nucleophilic reactivity towards various substrates, such as phenols, N,N-dimethylanilines, and various aldehydes, respectively [25,26]. The differences in reactivity observed in the investigated systems, as well as the stability of the peroxo intermediates can in principle be explained by the disproportionation of H 2 O 2 as a competing process. Clarification of the relationship between catalase-like activity and catalytic oxidation is an important factor in the design and development of high efficiency catalytic systems. In addition, the investigation of the metal-based disproportionation of H 2 O 2 may help us to understand the nature of the catalytically active species present both, in the enzymatic and biomimetic systems. We have reported that the [Fe II (L 1 )(solvent) 3 ](ClO 4 ) 2 (1) is an efficient catalyst for the oxidation of thioanisoles and benzyl alcohols by the use of H 2 O 2 as cooxidant, where a metastabile green species (5) with a Fe III (µ-O)(µ-1,2-O 2 )Fe III core was observed and characterized by UV-Vis, EPR, rRaman, and X-ray absorption spectroscopic measurements [24]. The reactivity of 5 was also published both in nucleophilic (deformylation of aldehydes) and electrophilic (oxidation of phenols) reactions [27].

Results and Discussions
In view of the above results, further studies were planned to elucidate the mechanism of both the diiron(III)-peroxo formation and its decay mediated by H2O2. Complex 5 from its precursor complex 1 can be generated even at room temperature by the use of excess of H2O2 in CH3CN. Its formation and decay process was followed as an increase and decrease in absorbance at 690 nm (ε = 1500 M −1 cm −1 per Fe ion and assuming 100% conversion), which can be assigned as a charge transfer between Fe III and the O2 2− ligand ( Figure 1a) [24]. Similar spectral feature can be observed for the complexes 6-8: 691 nm (ε = 1404 M −1 cm −1 ), 692 nm (ε = 1460 M −1 cm −1 ), and 693 nm (ε = 1308 M −1 cm −1 ), respectively. It is worth noting that the green species (5), after decomposition, can be regenerated 2-3 times by adding additional H2O2 without significant decrease on the yields (Table 1, Figure 1b and Inset in Figure 2a). Similar behavior and yields were observed for 6, but much lower values were obtained for 7 and 8. These results are consistent with the stability and/or the reactivity of the intermediates at high concentrations of added water, providing further information for the possible role of the binding water during the formation of the reactive oxidant.

Results and Discussions
In view of the above results, further studies were planned to elucidate the mechanism of both the diiron(III)-peroxo formation and its decay mediated by H 2 O 2 . Complex 5 from its precursor complex 1 can be generated even at room temperature by the use of excess of H 2 O 2 in CH 3 CN. Its formation and decay process was followed as an increase and decrease in absorbance at 690 nm (ε = 1500 M −1 cm −1 per Fe ion and assuming 100% conversion), which can be assigned as a charge transfer between Fe III and the O 2 2− ligand ( Figure 1a) [24]. Similar spectral feature can be observed for the complexes 6-8: 691 nm (ε = 1404 M −1 cm −1 ), 692 nm (ε = 1460 M −1 cm −1 ), and 693 nm (ε = 1308 M −1 cm −1 ), respectively. It is worth noting that the green species (5), after decomposition, can be regenerated 2-3 times by adding additional H 2 O 2 without significant decrease on the yields (Table 1, Figure 1b and inset in Figure 2a). Similar behavior and yields were observed for 6, but much lower values were obtained for 7 and 8. These results are consistent with the stability and/or the reactivity of the intermediates at high concentrations of added water, providing further information for the possible role of the binding water during the formation of the reactive oxidant.
Preliminary kinetic studies were performed at 20 • C to establish the role of the observed diiron(III)-peroxo intermediates in the catalytic disproportionation reaction of H 2 O 2 . The formation and decomposition of 5 in the presence of H 2 O 2 was monitored by following the rise and fall of its visible chromophore (Tables 2 and 3), as well as the appearance of the dioxygen formed by gas volumetric methods (Figure 2b). Dioxygen formation started after a short lag phase during which the peroxo-adduct (5) accumulated. The lag phase was followed by a linear dioxygen formation period during which time 5 persisted in a pseudo-steady-state. This profile is similar to that found for the water-assisted decay of Fe III Table 4). The results clearly indicate that diiron-peroxo intermediates play a key role in the formation of the reactive species responsible for the H 2 O 2 oxidation.  Preliminary kinetic studies were performed at 20 ℃ to establish the role of the observed diiron(III)-peroxo intermediates in the catalytic disproportionation reaction of H2O2. The formation and decomposition of 5 in the presence of H2O2 was monitored by following the rise and fall of its visible chromophore (Tables 2 and 3), as well as the appearance of the dioxygen formed by gas volumetric methods (Figure 2b). Dioxygen formation started after a short lag phase during which the peroxo-adduct (5) accumulated. The lag phase was followed by a linear dioxygen formation period during which time 5 persisted in a pseudo-steady-state. This profile is similar to that found for the water-assisted decay of Fe III (H2O)(OOH) intermediate into the reactive Fe V (O)(OH) species [28]. Upon depletion of H2O2, 5 decomposed rapidly and the dioxygen formation ceased.  Table 4). The results clearly indicate that diiron-peroxo intermedi-

Kinetic Studies on the Formation of Diiron-Peroxo Complexes (5-8)
Detailed kinetic studies on the reaction of 1 (2 mM) with H 2 O 2 (24-98 mM) were carried out in CH 3 CN at 15 • C, and the formation of the green species 5 was followed by UV-vis spectroscopy as an increase in absorbance at 690 nm under pseudo-first-order conditions (excess of H 2 O 2 ) ( Table 2)

Kinetic Studies on the Formation of Diiron-Peroxo Complexes (5-8)
Detailed kinetic studies on the reaction of 1 (2 mM) with H2O2 (24-98 mM) were carried out in CH3CN at 15 ℃, and the formation of the green species 5 was followed by UV-vis spectroscopy as an increase in absorbance at 690 nm under pseudo-first-order conditions (excess of H2O2) ( Table 2) 23,26]. It is worth noting that adding large volumes of water (~0.5 M) resulted in nearly twice the rate of the diiron-peroxo formation. Furthermore, a kinetic isotope effect (KIE) of 2.99 and 3.07 were observed for the formation of 6 and 7, when the experiments were carried out in the presence of added H2O or D2O. This value may represent the result of much more multiple effects compared to the elemental step, but clearly indicate the key role of the water during the diiron-peroxo complex formation.    In the next step, we investigated the effect of the the ligand modification by introducing electron withdrawing (Cl) or electron releasing (CH 3 ) substituents on the pyridine arms in the fourth and fifth positions with emphasis on the redox potential and Lewis acidity of the precursor complex. Since the redox potential of the substituted complexes can be measured under the same conditions, it can be used as excellent reactivity descriptor. On the cyclic voltammogram of the Fe(II)-isoindoline complexes (1-4) irreversible reduction and oxidation peaks can be recognized within the range of −440 to +1160 mV (vs. Fc/Fc + ). The huge peak separations (∆E = E pa − E pc , Table 4) indicates the irreversibility of the redox processes and suggests chemical reactions coupled to the electrode reaction. The results, obtained at the scan rate of 100 mV/s, show that oxidation potential is mainly influenced by the electronic properties of the ligands, since the anodic peak potentials (E pa ) of the complexes are increasing in the order of electron releasing ability of the substituent on the pyridine moiety (Figure 4). The E pa spans a 128 mV range from 655 mV (1) to 783 mV (4), which is not too large but informative, in terms of the effect of the substituents. However, the fact that the anodic peak of complex 1 can be found at lower potentials than it could be expected according to electronic effects, suggest that steric effects are also influencing the oxidation reactions. One can assume that the steric effect of the methylated and chlorinated ligands can cause an anodic shift in the oxidation peak potential of the complexes. In the next step, we investigated the effect of the the ligand modification by introducing electron withdrawing (Cl) or electron releasing (CH3) substituents on the pyridine arms in the fourth and fifth positions with emphasis on the redox potential and Lewis acidity of the precursor complex. Since the redox potential of the substituted complexes can be measured under the same conditions, it can be used as excellent reactivity descriptor. On the cyclic voltammogram of the Fe(II)-isoindoline complexes (1-4) irreversible reduction and oxidation peaks can be recognized within the range of −440 to +1160 mV (vs. Fc/Fc + ). The huge peak separations (ΔE = Epa − Epc, Table 4) indicates the irreversibility of the redox processes and suggests chemical reactions coupled to the electrode reaction. The results, obtained at the scan rate of 100 mV/s, show that oxidation potential is mainly influenced by the electronic properties of the ligands, since the anodic peak potentials (Epa) of the complexes are increasing in the order of electron releasing ability of the substituent on the pyridine moiety ( Figure 4). The Epa spans a 128 mV range from 655 mV (1) to 783 mV (4), which is not too large but informative, in terms of the effect of the substituents. However, the fact that the anodic peak of complex 1 can be found at lower potentials than it could be expected according to electronic effects, suggest that steric effects are also influencing the oxidation reactions. One can assume that the steric effect of the methylated and chlorinated ligands can cause an anodic shift in the oxidation peak potential of the complexes.  The four iron complexes, 1-4 have been compared to investigate the effect of the various aryl substituents. Figure 3a shows the results of the activity for the precursor complexes (1-4) during the formation of their appropriate diiron-peroxo intermediates (5)(6)(7)(8). This demonstrates only small differences exist, based on the first-order rate constants obtained under the same conditions. It was found that complex 4 with 5-chloropyridyl side chains is the most reactive complex with k obs of 3.66 M −1 s −1 , whilst complex 1 with unsubstituted side chains was proved to be the least active with k obs of 0.93 M −1 s −1 at 15 • C. Based on the CV data, we have got a clear evidence that the formation rate of the diiron-peroxo species (k 2 ) increases almost linearly with the redox potential (E pa for Fe III /Fe II ) of the precursor complex (Figure 3b). In summary, the higher the redox potentials of Fe III /Fe II redox couple the higher is the activity of the precursor complex towards to H 2 O 2 . Complexes with 4-and 5-chloro-pyridine side chains (3 and 4), whose metal sites are electron-deficient are faster in their reaction with H 2 O 2 than electron-rich derivatives 1 and 2 with more Lewis basic side chains.  The four iron complexes, 1-4 have been compared to investigate the effect of the various aryl substituents. Figure 3a shows the results of the activity for the precursor complexes (1-4) during the formation of their appropriate diiron-peroxo intermediates (5)(6)(7)(8). This demonstrates only small differences exist, based on the first-order rate constants obtained under the same conditions. It was found that complex 4 with 5-chloropyridyl side chains is the most reactive complex with kobs of 3.66 M −1 s −1 , whilst complex 1 with unsubstituted side chains was proved to be the least active with kobs of 0.93 M −1 s −1 at 15 °C. Based on the CV data, we have got a clear evidence that the formation rate of the diiron-peroxo species (k2) increases almost linearly with the redox potential (Epa for Fe III /Fe II ) of the precursor complex (Figure 3b). In summary, the higher the redox potentials of Fe III /Fe II redox couple the higher is the activity of the precursor complex towards to H2O2. Complexes with 4-and 5-chloro-pyridine side chains (3 and 4), whose metal sites are electron-deficient are faster in their reaction with H2O2 than electron-rich derivatives 1 and 2 with more Lewis basic side chains.

Catalase-Like Activity of Peroxo-Diiron Complexes
The ability of the in situ generated diiron-peroxo species (5) to mediate the disproportionation reaction of H2O2 into O2 and H2O was tested in CH3CN by UV-vis spectroscopy as a decrease in absorbance at 690 nm under pseudo-first-order conditions (excess of H2O2) ( Table 3). The initial rate of decay of diiron-peroxo species was measured as a function of the complex and substrate concentrations in CH3CN. At constant [H2O2]0 = 300 mM, the disproportionation reaction shows first-order kinetic behavior on [5], and the first-order rate constant k1 = 4.968 × 10 −3 s −1 was obtained at 20 ℃ from the slope of the plot of Vi versus [5]0 (Figure 5a). At constant [5]

Catalase-Like Activity of Peroxo-Diiron Complexes
The ability of the in situ generated diiron-peroxo species (5) to mediate the disproportionation reaction of H 2 O 2 into O 2 and H 2 O was tested in CH 3 CN by UV-vis spectroscopy as a decrease in absorbance at 690 nm under pseudo-first-order conditions (excess of H 2 O 2 ) ( Table 3). The initial rate of decay of diiron-peroxo species was measured as a function of the complex and substrate concentrations in CH 3 CN. At constant [H 2 O 2 ] 0 = 300 mM, the disproportionation reaction shows first-order kinetic behavior on [5], and the first-order rate constant k 1 = 4.968 × 10 −3 s −1 was obtained at 20 • C from the slope of the plot of V i versus [5] [29]. A kinetic isotope effect (KIE) of 1.73 and 1.66 were observed for the decay of 6 and 7, when the experiments were carried out in the presence of added H 2 O or D 2 O. These results indicate the key role of the water during the diiron-peroxo mediated disproportionation reaction.
In the next step, we investigated the effect of the ligand modification by introducing electron donating (-CH 3 ) and electron withdrawing (-Cl) phenyl-ring substituents. The substituted ligand-containing diiron-peroxo intermediates 5, 6, 7 and 8 show an increasing rate in the listed order (Figure 6a). It was found that complex 8 with 5-chloropyridine side chains is the most efficient oxidant with the fastest rate k 2 = 23.8 × 10 −2 M −1 s −1 , while complexes 5 and 6 with unsubstituted and 4-methylpyridine side chains are the less efficient oxidants with k 2 = 1.52 × 10 −2 M −1 s −1 and k 2 = 3.05 × 10 −2 M −1 s −1 , respectively. These results give clear evidence that substituents affect the redox potential and the catalase activity of the complexes. Table 4 and Figure 6b show that the redox potentials (E pa ) of the complexes and their activity increase with increasing electron-withdrawing ability of the substituent.

Materials and Methods
The ligands 1,3-bis(2 -pyridylimino)isoindolines (L 1 = indH, L 2 = 4Me-indH, L 3 = 4Cl-indH, L 4 = 5Cl-indH,) and their complexes (1-4) were prepared according to published procedures [31]. All manipulations were performed under a pure argon atmosphere using standard Schlenk-type inert-gas techniques. Solvents used for the reactions were purified by literature methods and stored under argon. The starting materials for the ligand are commercially available and they were purchased from Sigma Aldrich. The UV-visible spectra were recorded on an Agilent 8453 diode-array spectrophotometer using quartz cells. IR spectra were recorded using a Thermo Nicolet Avatar 330 FT-IR instrument (Thermo Nicolet Corporation, Madison, WI, USA), Samples were prepared in the form of KBr pellets. Microanalyses elemental analysis was done by the Microanalytical Service of the University of Pannonia.

Cyclic Voltammetry
Cyclic voltammetry experiments were performed on a Radiometer Analytical PGZ-301 potentiostat with a conventional three-electrode configuration, consisting of a glassy carbon working electrode (ID = 3 mm), a platinum wire auxiliary electrode, and Ag/AgCl reference electrode. Potentials are referenced to the Fc/Fc + redox couple. Ferrocene was added as an internal standard at the end of the experiments. Procedure for the cyclic voltammetry experiment: 0.05 mmol of 1-4 was dissolved in 10 mL dry acetonitrile, containing 349.1 mg (1 mmol) of tetrabutylammonium-perchlorate, which served as supporting electrolyte. The solution was bubbled with argon to remove dissolved gas residuals and to ensure inert atmosphere during measurements. The working electrode was wet polished on 0.5 µm alumina slurry or emery paper grade 500, after each measurement. Cyclic voltammograms were recorded with a scan rate of 100 mV s −1 .

Diiron-Peroxo-Mediated Disproportionaton of H 2 O 2
Reactions were carried out by mixing the substrate (H 2 O 2 ) with the in situ generated complex 5-8 (1 mM) in 1.5 mL acetonitrile in different temperature (0, 5, 10 15 • C) and the reactions were followed with UV-Vis spectroscopy at 680 nm, the cuvette length was 1 cm. To ensure the reliability of the measurements, 2-3 parallel measurements were performed.
Catalytic reactions were carried out at 20 • C in a 30 cm 3 reactor containing stirring bar under air. In a typical experiment the complex was dissolved dissolved in 20 cm 3 CH 3 CN, and the flask was closed with a rubber septum. H 2 O 2 was injected by syringe through the septum. The reactor was connected to a graduated burette filled with oil, and the evolved dioxygen was measured volumetrically at time intervals of 2 s.

Conclusions
In summary, the results of the reaction kinetics and CV measurements are shown in Table 4. Based on these data the following reaction mechanism, including the diiron-peroxo complex formation, and its reaction with H 2 O 2 , was proposed (Scheme 3). We obtained clear evidence for the role of the peroxo-intermediate in diiron catalase mimics. We also found that the higher the redox potentials of the Fe III /Fe II redox couple, the higher the catalase-like activity, and the complexes with electron-deficient metal sites are significantly more reactive. It can be explained by their electrophilic character. The results obtained may contribute to the elucidation of the mechanism of both dimanganase and diiron-containing catalase enzymes.