A DFT and QM/MM Investigation on Models Related to the [FeFe]-Hydrogenase Active Site

In the present thesis, a theoretical investigation is described regarding hydroge- nases - enzymes that are able to catalyze the reversible oxidation of molecular hydrogen: H2 2H+ + 2e− . Such a very simple reaction could have fundamen- tal importance for the possible future development of a hydrogen-based econ- omy. However, the current approaches for molecular hydrogen oxidation imply the use of very expensive platinum-containing catalysts, while H2 production at industrial level still depends on hydrocarbons. In this framework, hydrogenases represent a model for the development of new-generation catalysts, as they con- tain only inexpensive transition metal cofactors (iron and/or nickel ions) and are able to evolve hydrogen directly from acidic aqueous solutions supplied with a convenient source of electrons. The present work deals with the characterization of a specific class of hydro- genases, termed [FeFe]-hydrogenases. These enzymes contain in their active site a peculiar Fe6 S6 cluster - the so-called H-cluster - which can be ideally subdi- vided in two distinct portions: a classical Fe4 S4 moiety, and a Fe2 S2 subcluster (commonly termed [2Fe]H ) bearing CO and CN− ligands; these subclusters are linked to each other through the sulphur atom of a cysteine residue. The two iron atoms of the binuclear sub-site are termed proximal (Fep ) or distal (Fed ), de- pending on their positions with respect to the Fe4 S4 moiety. Notably one of the carbonyl groups included in the [2Fe]H subsite bridges the Fep and Fed centers, and it moves to a semibridging position when the enzyme is in its completely reduced form. The coordination environment of the iron ions included in the binuclear cluster is completed by a bidentate ligand which has been proposed to correspond either to a di(thiomethyl)amine (DTMA) or to a propanedithiolate (PDT) residue. Direct metal-hydrogen interaction at the binuclear sub-site is required for the enzymatic activity of [FeFe]-hydrogenases; however, there is still some debate about the way in which the interaction takes place, and about the catalytic mechanism leading to H2 splitting/formation. In fact, despite the large number of theoretical and experimental investigations carried out to clarify the catalytic mechanism of [FeFe]-hydrogenases, a direct comparison between the two more plausible routes for dihydrogen evolution/oxidation - i.e. a path involving the formation of metal-bound terminal hydrides, as opposed to a route that implies the presence of a hydride bridging Fep and Fed - was still lacking. Such study has then been carried out in our laboratories, using computational models of the H-cluster binuclear subsite in the context of a Density Functional Theory (DFT) representation; this work is presented in Chapter 2. It turns out that H2 formation can take place according to reaction pathways that imply initial protonation of the Fe(I)-Fe(I) form of [2Fe]H , leading to a formal Fe(II)-Fe(II) hydride species, subsequent monoelectron reduction to an Fe(II)-Fe(I) species, further protonation, and H2 release. A comparison of pathways involving either the initial protonation of Fed or protonation of the Fep -Fed bond shows also that the former pathway is characterized by smaller activation barriers, as well as a downhill free-energy profile, suggesting that it could be the H2 production pathway operative in the enzyme. The next chapter in the present thesis is devoted to the characterization of CO-mediated enzyme inhibition; indeed, the enzyme active site is able to bind exogenous carbon monoxide, and such an interaction impairs the catalytic process of H2 production/oxidation. Experimental and computational studies have converged towards the assignment of a Fe(I)Fe(II) state to the CO-inhibited binuclear sub-cluster, while there is still much debate about the disposition of CO and CN− ligands around Fed in this form. Our analysis is carried out us- ing a hybrid quantum mechanical/molecular mechanical (QM/MM) approach; this means that an all-atom model of the enzyme is used for studying different geometrical configurations of the active site. This allows us to show that the protein environment surrounding the H-cluster plays a crucial role in influenc- ing the mechanism of CO-inhibition; as a result, the CO-inhibited H-cluster is expected to be characterized by a terminal CO ligand trans to the μ-CO group on Fed . A QM/MM approach is also used in order to unravel key issues regarding the activation of the enzyme from its completely oxidized inactive state (Hox inact , an enzyme form in which the [2Fe]H subcluster attains the Fe(II)Fe(II) redox state), and the influence of the protein environment on the structural and cat- alytic properties of the H-cluster (see Chapter 4). Our results show that, in Hox inact , a water molecule is bound to Fed . The computed QM/MM energy values for water binding to the diferrous subsite are in fact over 17 kcal mol−1 ; however, the affinity towards water decreases by one order of magnitude af- ter a one-electron reduction of Hox inact , thus leading to release of coordinated water from the H-cluster. The investigation of a catalytic cycle of the [FeFe]- hydrogenase that implies formation of a terminal hydride ion and a DTMA molecule acting as acid/base catalyst indicates that all steps have reasonable reaction energies, and that the influence of the protein on the thermodynamic profile of H2 production catalysis is not negligible; QM/MM results show that the interactions between the Fe2 S2 subsite and the protein environment could give place to structural rearrangements of the H-cluster functional for catalysis, provided that the bidentate ligand that bridges the iron atoms in the binuclear subsite is actually a DTMA residue. In the last two studies included in the present thesis (Chapter 5 and Chapter 6), DFT investigations are presented regarding the characterization of two syn- thetic model complexes that represent structural and functional model of the [2Fe]H cluster: Fe2 (S2 C3 H6 )(CO)6 and (S2 C3 H6 )[Fe2 (CO)5 P(NC4 H8 )3 ]. Both of them are known to be able to catalyze proton reduction in an electrochemical cell, but the details of the electrocatalytic mechanisms leading to H2 produc- tion needed clarification. As for Fe2 (S2 C3 H6 )(CO)6 (a), it is showed that, in the early stages of the catalytic cycle, a neutral μ-H adduct is formed; mono-electron reduction and subsequent protonation can give rise to a diprotonated neutral species (a-μH-SH), which is characterized by a μ-H group, a protonated sulfur atom and a CO group bridging the two iron centers, in agreement with experi- mental IR data indicating the formation of a long-lived μ-CO species. H2 release from a-μH-SH and its less stable isomer a-H2 is kinetically unfavourable, while the corresponding monoanionic compounds (a-μH-SH− and a-H2 − ) are more reactive in terms of dihydrogen evolution, in agreement with experimental data. As far as (S2 C3 H6 )[Fe2 (CO)5 P(NC4 H8 )3 ] (A) is concerned, experimental results have suggested that the presence of the electron donor P(NC4 H8 )3 ligand in A could favour the formation of a μ-CO species similar to that observed in the enzymatic cluster. However, insight into the structural features of key catalytic intermediates deriving from reduction and protonation of A was still lacking. Thus, in Chapter 6 we present results obtained using Density Functional Theory to evaluate structures, relative stabilities and spectroscopic properties of several species relevant for the electrocatalytic H2 evolving process. The results enable us to unravel the structure of the μ-CO complex ex- perimentally detected after monoelectronic reduction of A. Moreover, we show that the introduction of the large electron-donor ligand P(NC4 H8 )3 in the bio- mimetic complex does not favour the stabilization of terminal-hydride adducts, which are expected to be very reactive in terms of H2 production. The comparison of our findings with previous theoretical and experimental results obtained on similar model complexes suggests that the introduction of an electron donor ligand as good as P(NC4 H8 )3 , but less sterically demanding, could represent a better choice to facilitate the formation of μ-CO complexes more closely resembling the structure of the enzymatic cluster.