Nature 476, 355–359 (2011)

Metabolic engineering of bacterial cells to generate commodity chemicals and biofuels typically entails either the rewiring of existing pathways to direct metabolic flux into desired endpoints or the identification of enzymes from across a host of bacterial species that can function at a given point within a proposed pathway. Dellomonaco et al. take a different approach: they envisioned that, by driving the known pathway for lipid degradation backwards, they could turn the normal products of this process—acetyl-CoA and propionyl-CoA—into substrates for a variety of short- and long-chain molecules. To accomplish this, the authors first introduced known mutations into two regulatory genes to cause constitutive expression of β-oxidation genes. It was also critical to inactivate two regulatory mechanisms that repress the same enzymes. Finally, the authors overexpressed several termination enzymes, specific for different molecules, to release the enzymatically bound or otherwise modified compounds. By feeding the cells glucose and in some cases propionate, the authors successfully obtained the expected products, including butanol and C16 and C18 fatty acids as well as several functionalized carboxylic acids. The high titers, yields and rates observed in the accumulation of these compounds were attributed to the direct usage of acetyl-CoA rather than to an activated version of the molecule. These results provide an intriguing example of the engineering of an endogenous pathway for a new use.