The maturase HydF enables [FeFe] hydrogenase assembly via transient, cofactor-dependent interactions

Abstract

[FeFe] hydrogenases have attracted extensive attention in the field of renewable energy research because of their remarkable efficiency for H₂ gas production. H₂ formation is catalyzed by a biologically unique hexanuclear iron cofactor denoted the H-cluster. The assembly of this cofactor requires a dedicated maturation machinery including HydF, a multidomain [4Fe4S] cluster protein with GTPase activity. HydF is responsible for harboring and delivering a precatalyst to the apo-hydrogenase, but the details of this process are not well understood. Here, we utilize gas-phase electrophoretic macromolecule analysis to show that a HydF dimer forms a transient interaction complex with the hydrogenase and that the formation of this complex depends on the cofactor content on HydF. Moreover, Fourier transform infrared, electron paramagnetic resonance, and UV-visible spectroscopy studies of mutants of HydF show that the isolated iron-sulfur cluster domain retains the capacity for binding the precatalyst in a reversible fashion and is capable of activating apo-hydrogenase in in vitro assays. These results demonstrate the central role of the iron-sulfur cluster domain of HydF in the final stages of H-cluster assembly, i.e. in binding and delivering the precatalyst.

Keywords: Fourier transform IR (FTIR); chaperone; cofactor; hydrogenase; iron-sulfur protein; mass spectrometry (MS); metal ion–protein interaction; metallo-cofactor assembly; metalloenzyme; protein–protein interaction; scaffold.

Figure 1.

Schematic representation of the metal cofactors in HydF and HydA and the maturation of [FeFe] hydrogenase by HydF.

Figure 2.

A, SEC chromatograms of the apo-HydF (top) and apo-HydA1 (bottom). Running conditions were 0.8 ml/min flowrate, 50 mm Tris-HCl, pH 8.0, 150 mm NaCl. B, the corresponding GEMMA spectra recorded for the different fractions of the apo-HydF and apo-HydA1 proteins obtained from SEC. Running conditions were 0.02 µg/µl protein in 20 mm ammonium acetate and 0.005% Tween 20.

Figure 3.

GEMMA spectra recorded on HydF and HydA1 proteins with different cofactor loadings. A, HydA1 (top, holo-HydA1; middle, [4Fe4S]-HydA1; bottom, apo-HydA1); B, HydF (top, holo-HydF; middle, [4Fe4S]-HydF; bottom, apo-HydF). The protein concentration in all samples were 0.02 µg/µl in 20 mm ammonium acetate and 0.005% Tween-20.

Figure 4.

GEMMA spectra recorded on combinations of HydF and HydA1 proteins with different cofactor content. A, apo-HydA1 interaction with different cofactor containing forms of HydF (top, holo-HydF; middle, [4Fe4S]-HydF; bottom, apo-HydF). B, [4Fe4S]-HydA1 interaction with different cofactor containing forms of HydF (top, holo-HydF; middle, [4Fe4S]-HydF; bottom, apo-HydF). The 150-kDa peak representing the HydF₂–HydA1 interaction complex is indicated with an asterisk. The experiments were performed with 0.02 µg/µl each protein (total protein concentration, 0.04 µg/µl) in 20 mm ammonium acetate and 0.005% Tween-20.

Figure 5.

A, a rigid-body protein-protein docking model of the complex between dimeric [4Fe4S]-HydF (red) and [4Fe4S]-HydA1 (blue). Models are based on reported crystal structures (PDB entries 5KH0 and 3LX4). B, crystal structure of the tetrameric form of apo-HydF (one HydF dimer shown in red, analogously to panel A, and one in brown), revealing a blockage of the interaction crevice (PDB entry 3QQ5).

Figure 6.

Spectroscopic characterization of the truncated HydF proteins. A and B, UV-visible spectra of HydFΔD (A) and HydFΔDG (B). Shown are the reconstituted ([4Fe-4S]²⁺) form of the HydF variants (black spectra), the [2Fe]^adt cofactor-loaded holo-forms ([2Fe]^adt-[4Fe-4S]²⁺) (red spectra), and the Na-DT reduced ([4Fe-4S]⁺) forms (dashed spectra). The samples were prepared in a buffer containing 100 mm Tris-HCl and 300 mm KCl with a protein concentration of 100 μm (HydFΔD) or 50 μm (HydFΔDG). Na-DT (0.5 mm) was added to generate the reduced samples. C and D, low-temperature EPR spectra of the reduced forms of HydFΔD (200 μm) (C) and HydFΔDG (200 μm) (D). Shown are both reconstituted (black spectra) and [2Fe]^adt-loaded forms (red spectra), and observed g-values are indicated. E, FTIR spectra of the [2Fe]^adt-loaded proteins, holo-HydFΔD and HydFΔDG. Spectra recorded for [2Fe]^adt and holo-HydF are displayed for comparison, and the peak positions of [2Fe]^adt are indicated with vertical dashed lines. The EPR spectra were recorded at 10 K, 1 mW microwave power, 10-Gauss modulation amplitude, and 100-kHz modulation frequency. The microwave frequency was 9.28 GHz. The FTIR spectra were recorded at room temperature, and the samples were prepared in a solution containing approximately 2.5 mm protein, 100 mm Tris-HCl, pH 8.0, 300 mm KCl.

Figure 7.

Influence of the different domains of HydF on the transfer of the precatalyst to apo-HydA1. A representative titration experiment in which [4Fe-4S]-HydA1 (8 nm) was titrated with 8–120 nm of holo-HydF (gray squares), 8–80 nm holo-HydFΔD (blue triangles), and 8–80 nm holo-HydFΔDG (orange diamonds). The extent of HydA1 activation was determined by calculating the resulting specific activity. 15-Fold molar excess (120 nm) of [2Fe]^adt complex was used as a positive control (green circles). Individual data points are indicated, see also Fig. S7. The maturation reactions were performed in 100 mm K-phosphate buffer (pH 6.8), and H₂ evolution was initiated via addition of dithionite and MV²⁺ after 15 min.