An orientation-selected ENDOR and HYSCORE study of the Ni-C active state of Desulfovibrio vulgaris Miyazaki F hydrogenase

Abstract

Electron nuclear double resonance (ENDOR) and hyperfine sublevel correlation spectroscopy (HYSCORE) are applied to study the active site of catalytic [NiFe]-hydrogenase from Desulfovibrio vulgaris Miyazaki F in the reduced Ni-C state. These techniques offer a powerful tool for detecting nearby magnetic nuclei, including a metal-bound substrate hydrogen, and for mapping the spin density distribution of the unpaired electron at the active site. The observed hyperfine couplings are assigned via comparison with structural data from X-ray crystallography and knowledge of the complete g-tensor in the Ni-C state (Foerster et al. (2003) J Am Chem Soc 125:83–93). This is found to be in good agreement with density functional theory calculations. The two most strongly coupled protons (a_iso=13.7, 11.8 MHz) are assigned to the β-CH₂ protons of the nickel-coordinating cysteine 549, and a third proton (a_iso=8.9 MHz) is assigned to a β-CH₂ proton of cysteine 546. Using D₂O exchange experiments, the presence of a hydride in the bridging position between the nickel and iron—recently been detected for a regulatory hydrogenase (Brecht et al. (2003) J Am Chem Soc 125:13075–13083)—is experimentally confirmed for the first time for catalytic hydrogenases. The hydride exhibits a small isotropic hyperfine coupling constant (a_iso=−3.5 MHz) since it is bound to Ni in a direction perpendicular to the z-axis of the Ni ${\left( {3d_{{z^{2} }} } \right)}$ orbital. Nitrogen signals that belong to the nitrogen N_ε of His-88 have been identified. This residue forms a hydrogen bond with the spin-carrying Ni-coordinated sulfur of Cys-549. Comparison with other hydrogenases reveals that the active site is essentially the same in all proteins, including a regulatory hydrogenase.

Appendix

Choosing signs for the direction cosines from the ENDOR spectra for D. gigas hydrogenase

When analyzing orientation-selected ENDOR spectra, information is extracted in the form of principal values of the hyperfine tensor and direction cosines. However, the absolute sign of the direction cosines cannot be determined from analysis of the experimental data.

Müller et al. [26] used their choice of signs for the direction cosines to reconstruct the orientation of the g-tensor for Ni-C in D. gigas hydrogenase. We propose a different sign choice for the following reasons. (1) By inspecting the field-frequency plot (see Fig. 3, top), protons a and b are found to be the two most strongly coupled protons and display a similar field-dependence, whereas the curve for proton c looks different. This is a strong indication that protons a and b (or H-A and H-B for D. gigas hydrogenase) belong to the same cysteine residue (and likewise protons H-C and H-D in [26]). (2) The assignment of Müller disagrees with the spin-density distribution calculated by two independent groups [16, 24, 25], which indicates that the sulfur of Cys-549 (Cys-533 for D. gigas) carries significant spin-density, so the protons of this residue should have the largest hyperfine coupling constants. Calculated proton hyperfine coupling constants from [25] are given in Table 6 for comparison. The calculated spin density distributions show that the principal z-axis of the g-tensor is directed from nickel towards the vacant coordination position, which is also expected based on ligand field theory. Lastly, the direction cosines of the x axis of proton H-B suggest that the minimum hyperfine coupling of 10.329 MHz (see Table 5) is attained for a molecule oriented such that the effective g value is [(2.194×0.283)² + 2.146×0.631)² + (2.011×0.723)²]^1/2=2.08. Inspection of the field-frequency plot in [26], however, reveals that proton H-B attains its minimum hyperfine coupling at g ≈2.15. The direction cosines of proton H-B seem to be in disagreement with the corresponding curve in the field-frequency plot of [26]. When the data for D. gigas hydrogenase [26] are used with the sign convention and assignment as chosen in the present work, a similar picture is obtained for the Ni-C state in D. gigas hydrogenase. Probably, the g-tensor for the Ni-C state of D. gigas hydrogenase is similar to the one determined by single crystal EPR for D. vulgaris hydrogenase [23], which is also compatible with theoretical studies [16, 24, 25].