Radu Silaghi-Dumitrescu

 

Activation of molecular oxygen and of peroxides by hemoproteins and by non-heme iron systems

 

 

 

Proteins such as cytochrome P450 (P450), horseradish peroxidase (HRP), hemoglobin and catalase are all known to bind and/or reduce (activate) dioxygen at their heme active sites, in the ferrous form. A brief overview of these processes is given below. While the thiolate-ligated ferrous heme of P450 binds dioxygen and promotes O-O bond cleavage, the ferrous histidine-ligated heme of hemoglobin only binds dioxygen in a reversible manner. Thiolate- as well as histidine-ligated active sites are also known to form ferric-hydroperoxo complexes. In HRP the hydroperoxo complex decays via proton-assisted heterolytic cleavage of the O-O bond to yield water and a [Fe=O]3+ unit (known as Compound I), where the iron is considered to be in the formal oxidation state +4 and a further oxidizing equivalent is proposed to reside on the porphyrin. Two active-site “distal” residues, a histidine and an arginine, are crucial in promoting O-O bond cleavage in the HRP ferric-hydroperoxo complex. The hemoglobin active site lacks the distal arginine and is much less efficient than HRP at activating peroxide. The heme-thiolate active site of P450 does not contain the histidine-arginine catalytic pair of HRP, yet still cleaves the O-O bond of its ferric-hydroperoxo complex. This situation is generally rationalized in terms of the thiolate ligand being able to push electron density (“the push effect”) more efficiently onto the iron-peroxo moiety, thereby facilitating O-O bond cleavage even in the absence of an efficient proton source such as the HRP histidine/arginine pair.

 

Physiologically-relevant reactions of active site hemes with O2 and/or H2O2. “S” denotes an organic substrate molecule. “X” may be a protein-derived cysteinate or histidine ligand.

 

Our interest has been in defining the properties and reactivity of the transient adducts depicted in the Figure above (nature of ferrous-dioxygen bonding, nature of ferric-peroxo interaction, electronic structure and protonation state of ferryl, i.e. Compounds I and II). Traditional manners of cleaving the O-O bond in metal-hydroperoxo complexes, and especially in Fe(III)-OOH, have been either proton-unassisted, or proton-assisted involving an Fe(III)-O-OH2 intermediate/transition state. However, an alternative mechanism has been proposed by Li et al, based on computations and involving an Fe(III)-O(H)-OH intermediate that undergoes O-O bond cleavage directly, avoiding the Fe(III)-O-OH2 species. The Figure below illustrates two examples of particularly facile O-O bond activation in Fe(III)-OH-OH adducts computed by us.

Upper panel: potential energy surface, following O-O bond elongation in a Fe(III)-OH-OH model of bleomycin (structure shown in inset, left lower corner). Geometries were optimized with the HO-OH distance constrained to values indicated in the plots, starting from the equilibrium value of 1.51 Å. Energy differences are plotted for each model, with the energy of the equilibrium structure (far left side of the plot) taken as reference for each model. At an O-O distance of 2.2 Å, both protons originating from the HOOH ligand were found on the leaving oxygen atom, yielding the structure depicted in the inset, upper right corner. Further optimization without any geometry constraints led to a final O-O distance of ~2.7Å (rightmost data point in the plot). Lower panel: potential energy surface, following meso-hydroxylation reactions in a Fe(III)-H2O2; X = imidazole, i.e. a possible mechanism for the reaction performed at the active site of heme oxygenase. Geometries were optimized with the O---C distances (shown as dotted lines in the reaction scheme) constrained to values indicated in the plots, starting from the equilibrium geometry with C---O ~3.8 Å, and shortening this distance by amounts indicated in the plot. The leftmost data point (solid triangle, S=1/2 model) was obtained by geometry optimization after lifting all geometry constraints. Energy differences are plotted for each model, with the energy of the equilibrium structure (far right side of the plot) taken as reference. No O-O bond cleavage was found to occur for the equivalent S=3/2 Fe(III)-OH-OH model under the conditions used here (instead, at the left-most point of the plot, O---C = 1.8 Å, the Fe-O bond was irrevocably broken)