The MC2 project is based on the findings of the previous PCV program. We briefly describe hereafter both the progress we made regarding methodological aspects of the work and the results we have obtained, which, together with recent results by other groups working in this field, lead to new questions to be addressed in MC2.
Based on the study of the biogenesis of one of the enzymes of interest [here] we concluded that these molybdoenzymes cannot be produced in large amounts by the cell without also over- expressing the proteins that are part of the biological maturation machineries. Doing so (either by using molecular biology procedures or simply by optimizing the growth conditions) increased not only the purification yields of the two nitrate reductases we study (NarGHI and NapAB) but also also their specific activities. We also established a biochemical procedure for the purification of Ralstonia sp. 22 arsenite oxidase (AroAB), in order to characterize its kinetics properties and its interaction with the physiological redox partners [here]. Improving the biochemical procedures regarding all three enzymes was a prerequisite for mechanistic studies based on biophysical techniques, some of which are very demanding in terms of quantity and/or purity of the biological material. This made it possible to approach all aspects of the mechanism of these enzymes: (i) the interaction with the quinones in the membrane, (ii) the electron transfer chain and (iii) the binding of substrate at the active site.
(i) Numerous membrane bound molybdoenzymes exchange electrons with the quinol/quinone pool in the membrane. Using a combination of site-directed mutagenesis, EPR and ENDOR spectroscopic techniques, we detected and located a semiquinone highly stabilized in the membrane-anchored subunit of membrane bound nitrate reductase (NarGHI, rightmost on fig 3) and concluded that quinol oxidation therein occurs at the distal heme, on the periplasmic side of the membrane [here]. The use of advanced EPR techniques such as X- and S-band (~ 3 GHz) 14N- and 15N-ESEEM and HYSCORE measurements on both the wild-type enzyme and the enzyme uniformly labelled with 15N nuclei allowed us to precisely characterize the interaction between the semiquinone radicals issued from the natural substrates (ubiquinone and menaquinone) and the protein environment [here,here].
(ii) In complex molybdoenzymes, electrons are transferred between the redox partner and the Moco by a chain of redox centers (fig. 3). When these electron relays are FeS clusters, which have ill-defined UV-vis signatures (as is the case of the enzymes we study), this intramolecular process is particularly difficult to study. We have characterized and compared the Rieske-type [2Fe2S] clusters of two arsenite oxidases (Ralstonia sp. 22 and Rhizobium NT-26) using EPR spectroscopy and reevaluated published data on the enzymes from Alcaligenes and Chroroflexus. We corrected published values of reduction potentias and showed that the Rieske-type centers of arsenite oxidases are distinct from their homologues from Rieske/cytb complexes [here,here]. Next to the Moco in NarG is a [4Fe4S] cluster whose very unusual coordination (it is bound to the protein by three cysteines and one histidine, instead of four cysteines) gives it peculiar electronic and spectral characteristics. We could unambiguously assign the g ~ 5 EPR signal of NarGHI to this cluster and we determined its redox properties. We had to develop a new quantitation method because conventional analyses of g ~ 5 signals can lead to large errors [here].
(iii) Regarding the chemistry that occurs at the Moco during turnover, by studying periplasmic nitrate reductase (NapAB) using site-directed mutagenesis, redox potentiometry, EPR spectroscopy and kinetics, we identified two conserved aminoacids that are involved in substrate binding, one of which acts as a filter in the substrate tunnel; it prevents direct reduction of the Mo atom by small reducing molecules, and may contribute to substrate specificity [here]. We developed new methods for precisely determining Michaelis and inhibition constants and how they depend on potential using protein film voltammetry (PFV), which we applied in particular to NapAB [here]. We discussed the unusual kinetic properties of this enzyme by comparing the results of electrochemical and spectrophotometric assays of several mutants [here]. We used PFV to demonstrate that substrate binding in NapAB affects the reduction potential of the active site even at the lowest substrate concentrations. This lowers the driving force required to reduce the active site under turnover conditions and favors intramolecular electron transfer from the proximal [4Fe4S]+ cluster towards the active site Mo(V) [here].
Whereas we succeeded in tackling various aspects of the mechanism of the three enzymes of interest, we found that the inability to detect the Moco or to generate trustworthy catalytic intermediates in which a substrate moiety is bound to the Mo is a major obstacle in mechanistic studies:
– Indeed, in arsenite oxidase, the Mo(V) EPR signature of the Moco has never been detected. This may be a consequence of the thermodynamic instability of the Mo(V) state. However, unpublished results in Team 1 suggest that the Mo(V) may actually be stable in a large potential window, but it would escape detection by EPR because it strongly interacts with the oxidized, paramagnetic [3Fe4S]+ cluster.
– Studying periplasmic nitrate reductase with a combination of PFV and spectroscopy, we have shown that the main Mo(V) EPR signature of the active site in the “as purified” enzyme, the so called Mo(V) “high-g signal,” is actually that of a dead-end species which irreversibly becomes catalytically competent only after full reduction [here]. Moreover, we have revealed the existence of several other forms of the enzyme which stand apart from the reaction pathway, and most importantly, we have demonstrated that these inactive forms accumulate under conditions that were always presumed to favour the formation of catalytic intermediates. We concluded that these inactive forms have been mistaken for catalytic intermediates .
Until very recently, it was admitted that the oxidized form of the Mo in the enzymes of the DMSO reductase family includes an oxo ligand (fig 1 and [Dias 1999]), which originates from the oxidized substrate and is released as a water molecule (fig 2). Crystallographers recently challenged this dogma when they proposed that the Mo in oxidized (“as prepared”) periplasmic nitrate reductase is coordinated by the four thiolates of two pterins, a cysteine and a sulfido ligand, rather than the expected oxo [Najmudin 2008]. This result is supported mainly by anomalous diffraction data. Based on the reinterpretation of former data [Raajimakers 2002], the same crystallographers also propose that the non proteic ligand of the Mo in formate dehydrogenase is a sulfido [Raajimakers 2006, Roamo 2009]. However, we note that the initial interpretation of high-resolution X-ray diffraction data of periplasmic DMSO and TMAO reductases had to be revisited when it became clear that the as-prepared enzymes were heterogeneous, and actually consisted of superpositions of structures [Zhang 2008]. In the case of NapAB, heterogeneity around the Moco has long been reflected by the number of Mo(V) EPR signals [here], but this has never been embraced in mechanistic studies relying on crystallographic data.