[FeFe]-hydrogenases, HydAs, are unique biocatalysts for proton reduction to H. However, they suffer from a number of drawbacks for biotechnological applications: size, number and diversity of metal cofactors, oxygen sensitivity. Here we show that HydA from Megasphaera elsdenii (MeHydA) displays significant resistance to O. Furthermore, we produced a shorter version of this enzyme (MeH-HydA), lacking the N-terminal domain harboring the accessory FeS clusters. As shown by detailed spectroscopic and biochemical characterization, MeH-HydA displays the following interesting properties. First, a functional active site can be assembled in MeH-HydA in vitro, providing the enzyme with excellent hydrogenase activity. Second, the resistance of MeHydA to O is conserved in MeH-HydA. Third, MeH-HydA is more biased toward proton reduction than MeHydA, as the result of the truncation changing the rate limiting steps in catalysis. This work shows that it is possible to engineer HydA to generate an active hydrogenase that combines the resistance of the most resistant HydAs and the simplicity of algal HydAs, containing only the H-cluster.

Many enzymes that produce or transform small molecules such as O, H, and CO embed inorganic cofactors based on transition metals. Their active site, where the chemical reaction occurs, is buried in and protected by the protein matrix, and connected to the solvent in several ways: chains of redox cofactors mediate long-range electron transfer; static or dynamic tunnels guide the substrate, product and inhibitors; amino acids and water molecules transfer protons. The catalytic mechanism of these enzymes is therefore delocalized over the protein and involves many different steps, some of which determine the response of the enzyme under conditions of stress (extreme redox conditions, presence of inhibitors, light), the catalytic rates in the two directions of the reaction and their ratio (the "catalytic bias"). Understanding all the steps in the catalytic cycle, including those that occur on sites of the protein that are remote from the active site, requires a combination of biochemical, structural, spectroscopic, theoretical, and kinetic methods. Here we argue that kinetics should be used to the fullest extent, by extracting quantitative information from the comparison of data and kinetic models and by exploring the combination of experimental kinetics and theoretical chemistry. In studies of these catalytic mechanisms, direct electrochemistry, the technique which we use and contribute to develop, has become unescapable. It simply consists in monitoring the changes in activity of an enzyme that is wired to an electrode by recording an electric current. We have described kinetic models that can be used to make sense of these data and to learn about various aspects of the mechanism that are difficult to probe using more conventional methods: long-range electron transfer, diffusion along gas channels, redox-driven (in)activations, active site chemistry and photoreactivity under conditions of turnover. In this Account, we highlight a few results that illustrate our approach. We describe how electrochemistry can be used to monitor substrate and inhibitor diffusion along the gas channels of hydrogenases and we discuss how the kinetics of intramolecular diffusion relates to global properties such as resistance to oxygen and catalytic bias. The kinetics and/or thermodynamics of intramolecular electron transfer may also affect the catalytic bias, the catalytic potentials on either side of the equilibrium potential, and the overpotentials for catalysis (defined as the difference between the catalytic potentials and the open circuit potential). This is understood by modeling the shape of the steady-state catalytic response of the enzyme. Other determinants of the catalytic rate, such as domain motions, have been probed by examining the transient catalytic response recorded at fast scan rates. Last, we show that combining electrochemical investigations and MD, DFT, and TD-DFT calculations is an original way of probing the reactivity of the H-cluster of hydrogenase, in particular its reactions with CO, O, and light. This approac...

International audienc

We report on upconverting luminescent nanoparticles (UCLNPs) that are spectrally tuned such that their emission matches the absorption bands of the two most important species associated with enzymatic redox reactions. The core-shell UCLNPs consist of a β-NaYF4 core doped with Yb(3+)/Tm(3+) ions and a shell of pure β-NaYF4. Upon 980 nm excitation, they display emission bands peaking at 360 and 475 nm, which is a perfect match to the absorption bands of the enzyme cosubstrate NADH and the coenzyme FAD, respectively. By exploiting these spectral overlaps, we have designed fluorescent detection schemes for NADH and FAD that are based on the modulation of the emission intensities of UCLNPs by FAD and NADH via an inner filter effect.

Along with aggregation of the amyloid-β (Aβ) peptide and subsequent deposit of amyloid plaques, oxidative stress is an important feature in Alzheimer's disease. Cu bound to Aβ is able to produce reactive oxygen species (ROS) by the successive reductions of molecular dioxygen, and the ROS produced contribute to oxidative stress. In vitro, ascorbate consumption parallels ROS production, where ascorbate is the reductant that fuels the reactions. Because the affinity of Cu for Aβ is moderate compared to other biomolecules, the rate of ascorbate consumption is a combination of two contributions. The first one is due to peptide-unbound Cu and the second one to peptide-bound Cu complexes. In the present Article, we aim to determine the amounts of the second contribution in the global ascorbate consumption process. It is defined as the intrinsic rate of ascorbate oxidation, which mathematically corresponds to the rate at an infinite peptide to Cu ratio, i.e., without any contribution from peptide-unbound Cu. We show that, for the wild-type Cu(Aβ) complex, this value equals 10% of the value obtained for peptide-unbound Cu and that this value is strongly dependent on peptide alterations. By examination of the dependence of the intrinsic rate of ascorbate oxidation, followed by UV-vis spectroscopy, for several altered peptides, we determine some of the key residues that influence ROS production.

FeFe hydrogenases catalyze H oxidation and production using an "H-cluster", where two Fe ions are bound by an aza-dithiolate (adt) ligand. Various hypotheses have been proposed (by us and others) to explain that the enzyme reversibly inactivates under oxidizing, anaerobic conditions: intramolecular binding of the N atom of adt, formation of the so-called "Hox/inact" state or nonproductive binding of H to isomers of the H-cluster. Here, we show that none of the above explains the new finding that the anaerobic, oxidative, H-dependent reversible inactivation is strictly dependent on the presence of Cl or Br. We provide experimental evidence that chloride uncompetitively inhibits the enzyme: it reversibly binds to catalytic intermediates of H oxidation (but not to the resting "Hox" state), after which oxidation locks the active site into a stable, saturated, inactive form, the structure of which is proposed here based on DFT calculations. The halides interact with the amine group of the H-cluster but do not directly bind to iron. It should be possible to stabilize the inhibited state in amounts compatible with spectroscopic investigations to explore further this unexpected reactivity of the H-cluster of hydrogenase.

Hydrogenases are metalloenzymes that catalyze the redox conversion between H2 and protons. The so-called [NiFeSe] hydrogenases are highly active for both H2 production and oxidation, but like all hydrogenases, they are inhibited by O2. In the [NiFeSe] enzyme from Desulfovibrio vulgaris Hildenborough this inhibition results from the oxidation of an active site cysteine ligand. We designed mutations that constrict a hydrophilic channel which connects the protein surface to this active site cysteine. Two of the variants show markedly increased tolerance to O2 inactivation, while they retain high catalytic activities in both directions of the reaction, and structural studies confirm that these mutations prevent the oxidation of the cysteine. Our results indicate that the diffusion of O2 or ROS to the active site can occur through a hydrophilic water channel, in contrast to the widely held assumption that only hydrophobic channels are involved in active site inactivation. This provides an original strategy for optimizing the enzyme by protein engineering.

Redox metalloenzymes are omnipresent in living organisms where they catalyze key cellular reactions with great efficiency. These enzymes can often be reversibly placed into inactive states following changes in redox conditions. This is a hindrance for their use in biotechnological devices, and also a complication for their study via a structure/function approach, because structural data alone usually is not enough to discriminate between active and inactive states. However, these inactive states can also inform on the chemistry of the enzyme's active sites and on their catalytic cycles. A technique that has proved particularly valuable in the last decades for studying these processes is protein film voltammetry (PFV), in which an enzyme is immobilized on an electrode in a configuration where direct electron transfer is possible. In this article, we review the studies of redox (in)activation processes using PFV, present the theory for a number of cases (reversible inactivations, irreversible activations), and give guidelines to obtain and interpret suitable kinetic data.[a] Dr. M. del Barrio, Dr. V. Fourmond