Biomimetic assembly and activation of [FeFe]-hydrogenases

Abstract

Hydrogenases are the most active molecular catalysts for hydrogen production and uptake, and could therefore facilitate the development of new types of fuel cell. In [FeFe]-hydrogenases, catalysis takes place at a unique di-iron centre (the [2Fe] subsite), which contains a bridging dithiolate ligand, three CO ligands and two CN(-) ligands. Through a complex multienzymatic biosynthetic process, this [2Fe] subsite is first assembled on a maturation enzyme, HydF, and then delivered to the apo-hydrogenase for activation. Synthetic chemistry has been used to prepare remarkably similar mimics of that subsite, but it has failed to reproduce the natural enzymatic activities thus far. Here we show that three synthetic mimics (containing different bridging dithiolate ligands) can be loaded onto bacterial Thermotoga maritima HydF and then transferred to apo-HydA1, one of the hydrogenases of Chlamydomonas reinhardtii algae. Full activation of HydA1 was achieved only when using the HydF hybrid protein containing the mimic with an azadithiolate bridge, confirming the presence of this ligand in the active site of native [FeFe]-hydrogenases. This is an example of controlled metalloenzyme activation using the combination of a specific protein scaffold and active-site synthetic analogues. This simple methodology provides both new mechanistic and structural insight into hydrogenase maturation and a unique tool for producing recombinant wild-type and variant [FeFe]-hydrogenases, with no requirement for the complete maturation machinery.

Figure 1. Structures of the diiron clusters discussed in the study

Left: the synthetic mimics 1^-, 2, and 3; Middle: proposed structure for the x-HydF (x = 1-3) hybrid proteins; Right: the H-cluster (active site) of [FeFe]-hydrogenase. The protein ribbon and the [4Fe-4S] clusters (shown as balls and sticks with Fe shown as white spheres) are shown only schematically.

Figure 2. Normalized Fourier-transform infrared (FTIR) spectra recorded in liquid solution at 15°C

Panel A: complexes 1-3; Panel B: CaHydF (from ref 19) and x-HydF (x = 1-3) hybrid species; Panel C: HydA1 after treatment of apo-HydA1with 1-HydF (1-HydA1), 2-HydF (2-HydA1) and 3-HydF (3-HydA1). Peak positions in the spectrum of 2-HydA1 are color coded to indicate the contributions from H_ox (red), H_red (violet), Hsred (green) and H_ox-CO (blue) (Figure S6 for a complete dataset). Panel D: HydA1 from C. rheinhardtii expressed in C. acetobutylicum (CrHydA1) (ref 18). Color code as in panel C. Samples of complexes 1-3 and x-HydF (x = 1-3) were prepared in HEPES buffer (20 mM, 100 mM KCl) pH 7.5. Samples of x-HydA1 have been prepared in 10 mM Tris-HCl buffer pH 8 containing sodium dithionite (2mM).

Figure 3. Continuous wave and pulsed EPR spectra on 1-HydF

a: X-band EPR spectra recorded at 10K for dithionite-reduced 1-HydF (black line) and HydF (dashed line) in 50 mM Tris-HCl buffer, 150 mM NaCl, 5 mM DT, pH 8. Microwave power = 100 μW, mod. amp. = 1 mT, mwfreq. = 9.39 GHz. The shoulder observed at g = 1.90 in the 1-HydF spectrum, corresponding to a few percent of the total signal intensity, is assigned to a small fraction of HydF lacking 1. b: X-band 2D pulsed ESEEM spectroscopy (CF-NF) of 1-HydF labelled with ¹³CN– (left) and unlabelled 1-HydF (right). The horizontal ridge seen at (ν₂ = 2·ν_13C = 7.7 MHz) is attributed to a hyperfine interaction between a ¹³C nucleus and the paramagnetic [4Fe-4S] cluster. Its extension yields the magnitude of the coupling (Δν₁ = 4.1 MHz = a_13C). This feature is absent from the unlabelled 1-HydF spectrum.

Figure 4. Specific hydrogenase activity of reconstituted HydA1

Activity (μmol H₂ ·min⁻¹ ·mg HydA1⁻¹ with standard deviation) of HydA1 was measured in the presence of methyl viologen (10 mM) and sodium dithionite (100 mM) after in vitro maturation of apo-HydA1 for 30 minutes at 37°C with 10 equivalents of x-HydF (x = 1-3), HydF or CaHydF. The value for the latter was obtained after 60 min reaction and was taken from reference 23.