Biochemistry 50: 5858- 5869 (2011)

Molecular features that enable certain [NiFe] hydrogenases to catalyze hydrogen (H2) conversion in the presence of dioxygen (O2) were investigated in the present study. By means of X-ray absorption spectroscopy (XAS) we compared the [NiFe] active site and FeS clusters in the O2-tolerant membrane-bound hydrogenase (MBH) of Ralstonia eutropha and the O2sensitive periplasmic hydrogenase (PH) of Desulfovibrio gigas. Fe-XAS indicated an unusual complement of iron-sulfur centers in the MBH, likely based on an unusual structure of the FeS cluster proximal to the active site. This cluster is a [4Fe4S] cubane in PH. For MBH it comprises less ~2.7 Å Fe-Fe distances and additional longer vectors of 3.4 Å, consistent with an Fe-trimer plus a more isolated Fe ion. Ni-XAS indicated a similar architecture of the [NiFe] site in MBH and PH, featuring Ni-coordination by four thiolates of conserved cysteines, i.e. in the fully reduced state (Ni-SR). For oxidized states, short Ni-μO bonds due to Ni-Fe bridging oxygen species were detected in the Ni-B state of the MBH and in Ni-A of the PH. Furthermore, a bridging sulfenate (CysSO) is suggested for an inactive state (Niia-S) of the MBH. We propose that the O2-tolerance of the MBH is mainly based on a dedicated electron donation from a modified proximal FeS cluster to the active site, which may favour formation of the rapidly re-activated Ni-B state instead of slowly re-activated Ni-A. Thereby, catalytic activity of the MBH is facilitated in the presence of both, H2 and O2. Biochemistry 50: 58585869 (2011) Hydrogenases (H2ases) are enzymes that catalyze the cleavage and production of molecular hydrogen (H2) at high turnover rates (1-2). They are of interest in the worldwide effort to substitute fossil fuels by renewable energy sources, such as H2 (3-5). A limitation for biotechnological application of most H2ases is their sensitivity towards dioxygen (O2) (6). However, a few H2ases are catalytically active in the presence of O2 (7-11). The elucidation of the molecular basis of their O2-tolerance may open new strategies to improve the catalytic features of O2-sensitive H2ases by genetic engineering (12-15) and could be useful for the design of novel biomimetic synthetic catalysts (16-18). The facultative chemolithoautotrophic bacterium Ralstonia eutropha H16 (Re) has the ability to grow on H2, CO2, and O2, and harbours three H2ases, all of which are active in the presence of O2. These are the regulatory H2ase (RH), which acts as an H2 sensor (19-21), the soluble NAD + -reducing H2ase (SH) (22-24), and the membrane-bound H2ase (MBH) (25-26). All enzymes belong to the [NiFe] class and their active sites contain one Ni and one Fe atom (1, 27-29) as opposed to the [FeFe] H2ases possessing a homo-metallic center (1, 27, 30). The MBH, which is in the focus of the present study, is bound to the periplasmic side of the cytoplasmic membrane via a membrane-integral b-type cytochrome (31). The enzyme shows high H2-oxidation activity (26) and feeds the derived electrons into the respiratory chain (32). The MBH sustains growth of R. eutropha with H2 even at atmospheric levels of O2 (26, 33-35). Electrochemical experiments have shown that H2-oxidizing activity at ambient O2 pressure (~230 mbar) can be up to 60 % of the activity determined in the absence of O2; nearly full activity is recovered at high redox potential when O2 is replaced by H2 (33-36). Furthermore, the MBH is insensitive to inhibition by CO (36). These features are unique and not found for the standard type of [NiFe] H2ases, e.g., from Desulfovibrio and Allochromatium species, in which H2-oxidizing activity is completely inhibited by traces of O2 and low CO levels. Reactivation of O2-inhibited standard H2ases occurs extremely slowly and only at low potentials (6, 37-38). Comentario [MSOffice1]: ? Biochemistry 50: 58585869 (2011) The high O2-tolerance of the MBH is surprising because its amino acid sequence (26, 29) suggests a high structural similarity with standard [NiFe] H2ases, e.g. the periplasmic H2ase (PH) of Desulfovibrio gigas (Fig. 1). The MBH consists of a large subunit (HoxG) comprising the [NiFe] site and a small subunit (HoxK), which harbours binding sites for three FeS clusters (Fig. 1) (39-40). According to the crystal structures of several [NiFe]hydrogenases, generally four conserved cysteines in the large subunit are involved in coordinating the active site Ni and Fe atoms (27, 41). A further common feature of [NiFe] H2ases is the ligation of the Fe by two cyanides (CN ) and one CO as observed by FTIR spectroscopy (36, 42-43). Apart from these similarities, biochemical, electrochemical, and spectroscopic results (26, 36, 43-45) suggest that O2 tolerance of the MBH is related to the structural arrangement of its metal cofactors, rather than to the restricted access of O2 and CO to the [NiFe] site due to a narrow gas-channel as proposed for H2-sensing H2ases (12, 26, 46). Notably, compared to standard H2ases, biosynthesis of active MBH requires a significantly larger set of maturation proteins (47-49). Spectroscopic studies on the MBH using EPR and FTIR techniques uncovered several features differing from those of standard enzymes (43-46). The so-called Ni-A state, corresponding to oxidized, inactive unready enzyme, was not detectable in wild-type MBH. For standard H2ases residing in the Ni-A state, a (hydro)peroxo species was suggested to occupy the bridging position between Ni and Fe, whereas in the Ni-B state a hydroxide is assigned as bridging ligand (50). Remarkably, in addition to the absence of Ni-A, a complex EPR spectrum indicates modification of the FeS cluster in the proximal position to the [NiFe] site in the MBH (43-46). The structural basis of these modifications needs to be elucidated. In the present study, the metal centers of the MBH were characterized by an elementspecific technique, namely X-ray absorption spectroscopy (XAS) (51-52), by which the coordination environment and redox changes of the protein-bound metal atoms were determined. The spectroscopic features of the MBH are compared to those of the PH standard Biochemistry 50: 58585869 (2011) H2ase, for which crystal structures are available (27, 53). Our results suggest a different structure of the proximal FeS cluster, but a rather similar, although more disordered [NiFe] active site in MBH compared to PH. MATERIALS AND METHODS Enzyme purification and sample preparation. Three different MBH protein preparations were used for this study. The most active MBH (mean activity of 14013 U/mg) was isolated from R. eutropha strain HF649, which was grown on fructose-glycerol mineral medium under oxygen-limited conditions (43-44, 54). The membrane fraction was prepared under an argon atmosphere and oxidized by the addition of 50 mM K3[Fe(CN)6] prior to aerobic solubilisation with Triton X-114. Purification via Strep-Tactin affinity chromatography was carried out in K-PO4 buffer, pH 7.0. The protein samples were concentrated using Amicon Ultra-15 (PL-30) and Amicon Microcon (YM-30) filtration devices (Millipore) in buffer containing 40 mM K-PO4, pH 5.5, 150 mM NaCl, 20 % glycerol. MBH mutant protein (HoxK Cys19/120Gly) was prepared as previously described (54) and concentrated as shown above (pH 5.5, ~86 U/mg). Less active MBH protein (mean activity of 604 U/mg) was purified as described previously (26, 43) and concentrated under aerobic conditions, and the purified protein was finally stored in 50 mM K-PO4 buffer at pH 8.0. All preparations lacked the membrane-integral cytochrome-b that serves as primary electron acceptor. Protein concentrations were determined by the Bradford method (55) with bovine serum albumin as a standard. Purity of samples was examined by SDS-PAGE. H2ase activity was determined photometrically (26, 40) in 50 mM K-PO4 buffer at pH 5.5 using the artificial electron acceptor methylene blue. For treatments in the presence of H2, concentrated protein solutions (~30 μL) were enclosed in gas-tight reaction vessels, the atmosphere was exchanged by repeated degassing (23), and solutions were incubated under a gentle stream of moistured gas for 10-45 min at ~20 °C. The final MBH protein concentrations in the samples Biochemistry 50: 58585869 (2011) were 0.5-0.8 mM. PH protein from D. gigas was prepared and concentrated under aerobic conditions and protein concentrations were determined as described before (56). Activation of PH was performed upon incubating Eppendorf reaction vessels containing concentrated protein (~30 μl) in repeatedly degassed buffer at 35 °C under pure H2 for 90-180 min. Reduction of PH was achieved by incubation with Na-dithionite (10 mM) for 20 min. Final PH concentrations were 0.7-1.0 mM. Immediately after the treatments, ~20 μl of protein samples were transferred under argon atmosphere to Kapton-covered sample holders for XAS and EPR measurements, rapidly frozen, and stored in liquid nitrogen until use. Metal content quantification. The metal content in H2ase proteins was determined by total-reflection X-ray fluorescence detection (TXRF) (57) on a PicoFox instrument (Bruker). Ni and Fe concentrations were determined using the built-in spectrometer functions relative to a gallium standard (Sigma), which had been added prior to the measurements to the protein solutions. FTIR spectroscopy. Infrared spectroscopy was carried out on a Bruker TENSOR 27 spectrometer as described before (43) with aliquots of the protein samples used for XAS. X-ray absorption spectroscopy (XAS). XAS measurements were performed at beamline D2 of the EMBL outstation (at HASYLAB, DESY, Hamburg, Germany) and at beamline KMC-1 of BESSY (Helmholtz-Zentrum Berlin). K fluorescence-detected XAS spectra at the Ni and Fe K-edges were collected for samples held in liquid-helium cryostats at 20 K with energy-resolving 13-element Ge detectors (Canberra) (58-59). Fe spectra were recorded for a monochromator scan range of 7000-8200 eV. For Ni, in an extended-range approach (60), the scan range was 8200-9200 eV. XAS spectra were averaged (5-12 scans) after energy calibration of each scan by the Bragg-reflection method (61) or using the spectrum of a Ni metal foil as an energy standard, normalized, and EXAFS oscillations were extracted (52). The energy scale of EXAFS spectra was converted to a wavevector (k) scale using E0 values of 8333 eV (Ni) and 7112 eV (Fe) E0 refined to ~7120 eV (Fe) and ~8336 eV (Ni) in the Biochemistry 50: 58585869 (2011) simulations. Unfiltered k 3 -weighted spectra were used for least-squares curve-fitting and Fourier-transform (FT) calculation with the in-house program SimX (52). In EXAFS simulations, phase-functions from FEFF7 (62) and amplitude reduction factors (S0 2 ) of 0.9 (Ni) and 0.85 (Fe) were used. XANES simulations. XANES calculations were done as described in (23, 63) using the code FEFF8.2 (62) with the full-multiple-scattering (FMS) and the self-consistent-field (SCF) options activated. Atomic coordinates of FEFF input files were generated using the crystal structure of the [NiFe] site of D. gigas H2ase (41) as a template and Ni-ligand distances from the EXAFS analysis. Calculated spectra were shifted by 1 eV to lower energies and smoothed over data points within 3 eV for better comparison with the experimental data. Density functional theory (DFT) calculations. DFT calculations were performed using the program ORCA (64-66). Geometry optimizations of [NiFe] site models involved the BP86 exchange-correlation functional (67-68) and a double-zeta basis set (69) with polarization functions from the TURBOMOLE library (ftp://ftp.chemie.uni-karlsruhe.de/pub/basen) and a dielectric constant of ε = 4 in a COSMO solvation model (70, 71).