hide
Free keywords:
-
Abstract:
Iron-sulfur (FeS) clusters are omnipresent in nature, where they are involved in a variety of tasks such as electron transfer, DNA repair, Fe storage, and substrate activation. The enzyme nitrogenase contains one of the largest known biological FeS clusters. Nitrogenase is responsible for the conversion of inert N2 gas to bioavailable ammonia, a reaction that is copied by the industrial Haber-Bosch process. However, despite decades of research, little is known about the molecular mechanism behind enzymatic N2 reduction. The active site of nitrogenase is the iron-molybdenum cofactor (FeMoco, [MoFe7S9C] ) in the MoFe protein, which contains a unique μ6−C4 – center. The complex electronic structure of FeMoco pushes state-of-the-art quantum mechanical methods to their limits, therefore, smaller model compounds are often explored with high-level methods. The present work discusses the electronic structure of FeS clusters, ranging from monomeric and dimeric FeS model compounds up to a QM/MM model for FeMoco in the MoFe protein, and relates the results to experimental findings. The discussion of the results is separated into four chapters, each focusing on different electronic structure aspects of FeS clusters relevant to FeMoco.
Some S ligands in FeMoco can be selectively replaced with Se, which can be used as a probe for the electronic structure with element-specific spectroscopic techniques. However, it is unknown how much the electronic structure of FeMoco is altered by the replacement. The present work quantifies the perturbation of the S→Se substitution in [Fe(XH)4]1 – ,2 – and [Fe2X2]2+,1+ (X = S, Se) model compounds. The analysis of the electronic structure focuses mostly on the multiconfigurational complete active space self-consistent field (CASSCF) wave function. The local electronic structure of the Fe−X bonds is characterized with ab initio ligand field theory (AILFT ) and the angular overlap model (AOM), while the metal-metal interactions are related to spin Hamiltonian parameters. Se-based ligands show a smaller ligand field splitting and have a reduced donor strength compared to the S-based counterparts. For the homo-valent [Fe2X2]2+ clusters, S →Se substitution reduces the antiferromagnetic Heisenberg exchange coupling constant J by about 10 %. In the mixed-valent [Fe2X2]1+ clusters, on the other hand, the coupling strength decreases by about 50 % upon S→Se substitution. The latter trend can be explained by increasing contributions from double exchange and vibronic coupling and is consistent with the mixture of spin states observed experimentally. S→Se substitution may therefore have a noticeable effect on the electronic structure of FeMoco, where antiferromagnetic coupling plays an important role.
Higher-nuclearity FeS clusters, such as FeMoco, often exhibit valence-delocalized Fe2.5+Fe2.5+ pairs as part of the electronic structure. In contrast, nearly all FeS dimers in the [Fe2S2]1+ redox state have been reported to have a valence-localized Fe2+Fe3+ electronic structure. Cys→Ser variants of the [Fe2S2] ferredoxins from Clostridium pasteurianum (Cp) and Aquifex aeolicus (Aae) are the only known examples for a valence-delocalized [Fe2S2]1+ cluster. This work presents density functional theory (DFT) cluster model calculations for the [Fe2S2]1+ ferredoxins from Cp and Aae. The electronic structure in the wild type model is localized, but delocalized in the Cys→Ser variant, consistent with experiment. Furthermore, protonation in the variant model leads to a localized electronic structure, which is consistent with the experimentally observed pH dependence of the delocalization. The results suggest that the terminal ligands are central to valence delocalization.
The complex wave function of FeMoco is most often modeled using broken-symmetry (BS ) DFT. Here, the local spins on each metal center can be aligned in a multitude of ways, but the most stable BS determinants are usually those that maximize the number of antiferromagnetically coupled Fe pairs. Furthermore, the localized orbital analysis yields the distribution of unpaired electrons across the metal centers and is helpful to rationalize structure-reactivity relations. The present work shows that calculating the coupling constants for different Fe pairs explicitly leads to an improved correlation between the energies and antiferromagnetic coupling. A comparison of localization algorithms suggests that Foster-Boys orbitals are the most robust in the context of the localized orbital analysis. Atomic charges, on the other hand, do not correlate with the localized orbitals analysis, but capture differences in the protein environment, such as in quantum mechanics/molecular mechanics (QM/MM) models of Mo nitrogenase compared to V nitrogenase. Furthermore, the atomic charges show that if the QM/MM boundary runs too close to the active site, the metal cluster becomes overpolarized by the MM charges. These findings help to interpret the results of DFT calculations for FeMoco and to choose a suitable QM region for QM/MM models.
The molecule CO is isoelectronic to N2. It binds reversibly to Mo nitrogenase, where it acts as an inhibitor to N2 reduction. V nitrogenase, on the other hand, reduces CO to hydrocarbons in a Fischer-Tropsch-like reaction. CO-inhibited nitrogenase gives rise to a number of experimentally well-characterized electron paramagnetic resonance (EPR) and infrared (IR) species, as well as X-ray diffraction (XRD) structures. However, details of the initial binding event are unknown, such as the redox state of the active site (En) or the role of the protein environment. In this work, the mechanism of CO binding to Mo and V nitrogenase is studied using QM/MM models. The models for the E1 redox state feature a terminal and a bridging CO binding motif, where the calculated frequencies (1922 cm−1 and 1716 cm−1) agree well with the experimentally observed IR bands (1904 cm−1 and 1715 cm−1). Therefore, the QM/MM calculations are consistent with CO binding happening in the E1 redox state. Alternatively, the calculated frequency for a semi-bridging CO in the E2 QM/MM model (1718 cm−1) is also consistent with the latter IR band and the topology is the same as in the CO-bound XRD structure. Analogous models for V nitrogenase do not show significant differences, even though here CO binding has been reported without enzymatic turnover conditions. Furthermore, the careful analysis of the electronic structure reveals that CO coordination induces local spin pairing at the binding site. This, in turn, affects the magnetic interaction between the metal center and leads to an energy reordering of the BS determinants.
