Insulation of a synthetic hydrogen metabolism circuit in bacteria

Abstract

Background: The engineering of metabolism holds tremendous promise for the production of desirable metabolites, particularly alternative fuels and other highly reduced molecules. Engineering approaches must redirect the transfer of chemical reducing equivalents, preventing these electrons from being lost to general cellular metabolism. This is especially the case for high energy electrons stored in iron-sulfur clusters within proteins, which are readily transferred when two such clusters are brought in close proximity. Iron sulfur proteins therefore require mechanisms to ensure interaction between proper partners, analogous to many signal transduction proteins. While there has been progress in the isolation of engineered metabolic pathways in recent years, the design of insulated electron metabolism circuits in vivo has not been pursued.

Results: Here we show that a synthetic hydrogen-producing electron transfer circuit in Escherichia coli can be insulated from existing cellular metabolism via multiple approaches, in many cases improving the function of the pathway. Our circuit is composed of heterologously expressed [Fe-Fe]-hydrogenase, ferredoxin, and pyruvate-ferredoxin oxidoreductase (PFOR), allowing the production of hydrogen gas to be coupled to the breakdown of glucose. We show that this synthetic pathway can be insulated through the deletion of competing reactions, rational engineering of protein interaction surfaces, direct protein fusion of interacting partners, and co-localization of pathway components on heterologous protein scaffolds.

Conclusions: Through the construction and characterization of a synthetic metabolic circuit in vivo, we demonstrate a novel system that allows for predictable engineering of an insulated electron transfer pathway. The development of this system demonstrates working principles for the optimization of engineered pathways for alternative energy production, as well as for understanding how electron transfer between proteins is controlled.

Figure 1

Overview of synthetic pathway design and insulation strategies A.) Natural and synthetic pyruvate metabolism to acetyl-CoA in E. coli through the pyruvate dehydrogenase complex (PDH), pyruvate formate lyase (PFL), and the heterologous PFOR-ferredoxin (Fd)-hydrogenase synthetic pathway. Native enzymes are indicated in black, heterologous enzymes in blue.~~~B.) Insulation strategies for synthetic electron transfer pathways; deletion of competing reactions, optimization of binding surfaces, direct protein-protein fusion, and localization to a synthetic protein scaffold. We present the maximum fold increase in hydrogen production due to each method, calculated by comparing normalized values of hydrogen production by otherwise identical synthetic pathways with and without the insulation strategy (see Results).

Figure 2

Characterization of synthetic hydrogen production pathway A.) Western blot of Strep-II tagged hydrogenase expression. B.) In vitro hydrogen production from E. coli strains expressing various hydrogenases, measured by the methyl viologen in vitro assay [18]. C.a. = C. acetobutylicum, C.s. = C. saccharobutylicum, C.r. = C. reinhardtii, S.o. = Shewanella oneidensis. C.) Glucose-dependence of hydrogen production. Here and below, in vivo and in vitro hydrogen production values are in units of μmol hydrogen/ml of E. coli culture, normalized to an OD₆₀₀ of 0.15 unless otherwise stated. Assays were performed in triplicate, with error bars indicating standard deviation. D.) In vivo hydrogen production from E. coli strains expressing all combinations of the four hydrogenases vs. three ferredoxins from C. acetobutylicum, Spinacia olearcea (Sp.o), and Zea mays (Zm). E.) In vivo hydrogen production from the C. acetobutylicum hydrogenase paired with combinations of three ferredoxins and three PFOR genes.

Figure 3

Insulation of hydrogenase pathway through deletion of competing reactions Relative hydrogen production of different knockout strains compared to parent strain (ΔhycE, ΔhyaB, ΔhybC) expressing hydrogenase alone (dark blue bar), hydrogenase and ferredoxin only (yellow bars) or the full PFOR-ferredoxin-hydrogenase pathway (green bars).

Figure 4

Insulation of the hydrogenase pathway through ferredoxin binding surface mutagenesis A.) Homology model of C. reinhardtii hydrogenase with mutated residues highlighted in cyan and spinach ferredoxin X-ray structure (1A70 [54]) with negatively charged residues highlighted in yellow. Iron-sulfur clusters and the hydrogenase catalytic cluster are highlighted in orange. B.) Relative in vitro and in vivo hydrogen production for wild type and mutated C. reinhardtii hydrogenase. Mutants D126K and E5K, which enhance charge-complementarity at the putative interaction surface, show a specific enhancement of in vivo activity relative to activity changes seen in the ferredoxin-independent in vitro assay.

Figure 5

Increase in hydrogen production by hydrogenase-ferredoxin fusion A.) Schematic model of the protein fusion, here showing the C. acetobutylicum hydrogenase fused to the spinach ferredoxin N-terminus (N-termini highlighted in blue, C-termini in green, and iron-sulfur clusters in orange). B.) Hydrogenase-ferredoxin fusion proteins are highly expressed and are the predicted size for the chimera, as indicated by western blotting with an anti-ferredoxin antibody. C.) Linker-length dependent behavior of fusion with spinach ferredoxin to hydrogenase C-terminus in vivo or in vitro. D.) Linker-length dependent behavior of fusion with Clostridium ferredoxin at the hydrogenase N- or C-terminus in vivo.

Figure 6

Effect of artificial scaffolding configuration on hydrogen production from the synthetic circuit. A.) Positional effects of ferredoxin targeting to artificial scaffold on hydrogen production. B.) Circuit efficiency is dependent upon length of flexible linker connecting ferredoxin (FD) to scaffolding. C.) Modulation of ferredoxin to hydrogenase ratio on scaffold affects hydrogen production, with decreasing yield observed at higher ferredoxin:hydrogenase ratios. D.) Direct fusion of ferredoxins to one another yields diminishing hydrogen production with increased numbers of fused ferredoxin proteins.