Designed surface residue substitutions in [NiFe] hydrogenase that improve electron transfer characteristics

Abstract

Photobiological hydrogen production is an attractive, carbon-neutral means to convert solar energy to hydrogen. We build on previous research improving the Alteromonas macleodii "Deep Ecotype" [NiFe] hydrogenase, and report progress towards creating an artificial electron transfer pathway to supply the hydrogenase with electrons necessary for hydrogen production. Ferredoxin is the first soluble electron transfer mediator to receive high-energy electrons from photosystem I, and bears an electron with sufficient potential to efficiently reduce protons. Thus, we engineered a hydrogenase-ferredoxin fusion that also contained several other modifications. In addition to the C-terminal ferredoxin fusion, we truncated the C-terminus of the hydrogenase small subunit, identified as the available terminus closer to the electron transfer region. We also neutralized an anionic patch surrounding the interface Fe-S cluster to improve transfer kinetics with the negatively charged ferredoxin. Initial screening showed the enzyme tolerated both truncation and charge neutralization on the small subunit ferredoxin-binding face. While the enzyme activity was relatively unchanged using the substrate methyl viologen, we observed a marked improvement from both the ferredoxin fusion and surface modification using only dithionite as an electron donor. Combining ferredoxin fusion and surface charge modification showed progressively improved activity in an in vitro assay with purified enzyme.

Figure 1

Electrostatic models of S. elongatus ferredoxin PetF (A); Spinacia oleracea PFOR (PDB: 1FNB) (B); Clostridial [FeFe] hydrogenase (PDB: 1FEH) (C); and a structural model of the A. macleodii hydrogenase small subunit (D). Negatively charged residues found near the docking site are colored red in the structural model. Electrostatic models of A. macleodii hydrogenase variants G1 (E) and G2 (F) in the same orientation as (D). Circled areas highlight the region near the distal Fe-S cluster. All models were obtained from the PDB where codes are given, or generated by the threading modeler Phyre. Charges were modelled using the default vacuum electrostatic package in PyMOL. In all electrostatic models, red is negatively charged and blue is positively charged.

Figure 2

Methyl viologen-mediated (A) and methyl viologen-free (dithionite only) (B) in vitro hydrogen production assay from extracts of E. coli over-expressing the WT hydrogenase, G1 hydrogenase, and progressive substitutions to G2 hydrogenase (see Table 1 for sequence identities). Activities are plotted on a log scale over a 10-fold and 100-fold ranges, respectively to compare fold improvements.

Figure 3

Methyl viologen-mediated (A) and methyl viologen-free (dithionite only) (B) in vitro hydrogen production assay from extracts of E. coli over-expressing the G1 hydrogenase and Δ15 and Δ22 truncations of the C-terminal tail.

Figure 4

Methyl viologen-mediated in vitro hydrogen production assay (A) from tandem immobilized metal affinity chromatography (IMAC)/streptactin-purified G1 and G2 hydrogenases, and their respective ferredoxin fusions. In vitro H₂ production of the same hydrogenases in the methyl viologen-free assay (dithionite only) (B) and expressed as the ratio of methyl viologen-free activity to activity in assays containing methyl viologen (C), plotted on a log scale over an 100-fold range to compare fold-changes. Sypro-ruby stained gel (D) and Anti-HynL western blot (E) of the same constructs from the same protein preparations presented in (A,B).