A process for microbial hydrocarbon synthesis: Overproduction of fatty acids in Escherichia coli and catalytic conversion to alkanes

Abstract

The development of renewable alternatives to diesel and jet fuels is highly desirable for the heavy transportation sector, and would offer benefits over the production and use of short-chain alcohols for personal transportation. Here, we report the development of a metabolically engineered strain of Escherichia coli that overproduces medium-chain length fatty acids via three basic modifications: elimination of beta-oxidation, overexpression of the four subunits of acetyl-CoA carboxylase, and expression of a plant acyl-acyl carrier protein (ACP) thioesterase from Umbellularia californica (BTE). The expression level of BTE was optimized by comparing fatty acid production from strains harboring BTE on plasmids with four different copy numbers. Expression of BTE from low copy number plasmids resulted in the highest fatty acid production. Up to a seven-fold increase in total fatty acid production was observed in engineered strains over a negative control strain (lacking beta-oxidation), with a composition dominated by C(12) and C(14) saturated and unsaturated fatty acids. Next, a strategy for producing undecane via a combination of biotechnology and heterogeneous catalysis is demonstrated. Fatty acids were extracted from a culture of an overproducing strain into an alkane phase and fed to a Pd/C plug flow reactor, where the extracted fatty acids were decarboxylated into saturated alkanes. The result is an enriched alkane stream that can be recycled for continuous extractions. Complete conversion of C(12) fatty acids extracted from culture to alkanes has been demonstrated yielding a concentration of 0.44 g L(-1) (culture volume) undecane.

Figure 1

Fatty acid biosynthesis pathway in Escherichia coli and metabolic engineering strategy for overproduction of free fatty acids. Acetyl-CoA is converted to malonyl-CoA by the four subunits of acetyl-CoA carboxylase (AccABCD). Malonyl-CoA is converted to malonyl-ACP (acyl–acyl carrier protein) by malonyl-CoA:ACP transacylase (FabD), and to the acetoacetyl-ACP, the first β-ketoacyl-ACP in the fatty acid elongation cycle, by multiple pathways catalyzed by FabD, β-ketoacyl-ACP synthase III (FabH), and β-ketoacyl-ACP synthase I (FabB). The ketoacyl-ACP is reduced twice and dehydrated once to yield an acyl-ACP in the elongation cycle by β-ketoacyl-ACP reductase (FabG), enoyl-ACP reductase (FabI), and β-hydroxyacyl-ACP dehydratase (FabZ). The acyl-ACP is then condensed with malonyl-ACP by FabB or β-ketoacyl-ACP synthase II (FabF). Glycerol-3-phosphate acyltransferase (PlsB) and 1-acylglycerol-3-phosphate acyltransferase (PlsC) utilize C₁₆ to C₁₈ acyl-ACPs as substrates for phospholipid biosynthesis. Acyl-ACP thioesterases hydrolyze acyl-ACPs to yield free fatty acids. Free fatty acids can be degraded, ultimately to acetyl-CoA, by enzymes of β-oxidation. This process is initiated by acyl-CoA synthetase (FadD), which converts free fatty acids to acyl-CoAs. Bolded arrows represent reactions enhanced by overexpression (ACC) or heterologous expression (BTE) of indicated enzymes. X represents a disruption of the indicated pathway by gene deletion.

Figure 2

Time course of growth of Escherichia coli RL08 cultures harboring plasmids expressing BTE (filled markers) or nonfunctional BTE-H204A (open markers) monitored by optical density (OD₆₀₀). Cells were grown at 37°C in shake flasks containing Luria–Bertani medium supplemented with 0.4% (v/v) glycerol and appropriate antibiotics for each vector. Cells were induced at an OD₆₀₀ of 0.2 by the addition of a final concentration of 0.2% (w/v) L-arabinose.

Figure 3

Total fatty acid titers (μg mL⁻¹ culture medium, open bars) and C₁₂ chain length fatty acid titers (saturated and estimated unsaturated, filled bars) extracted from cultures shown in Figure 2 at approximately 23 h postinoculation. Error bars represent standard deviations about the mean of two or three replicate samples for either C₁₂ fatty acids (lower bars) or total fatty acids including C₁₂ (upper bars).

Figure 4

Growth and fatty acid production of strain RL08 harboring combinations of plasmids pBAD33 or pBAD33-ACC, and pBAD35-BTE-H204A or pBAD35-BTE. Cells were grown at 37°C in shake flasks containing Luria–Bertani medium supplemented with 0.4% (v/v) glycerol, 50 μg mL⁻¹ of ampicillin, and 34 μg mL⁻¹ of chloramphenicol. A: OD₆₀₀ as a function of time from inoculation; (B) total (filled bars) and C₁₂ chain length (saturated and estimated unsaturated, open bars) of fatty acid titers (μg mL⁻¹ culture medium) for selected times during cell growth as indicated (6, 10, 18, 29, and 34 h). Error bars represent standard deviations about the mean for three replicate samples for either C₁₂ fatty acids (upper bars) or total fatty acids including C₁₂ (lower bars).

Figure 5

Chromatograms of collected decane layers from a 34-h fatty acid overproducing culture (RL08/pBAD33-ACC/pBAD35-BTE, second from top trace) and a negative control culture (RL08/pBAD33/pBAD35-BTE-H204A, second from bottom trace) following decarboxylation at 300°C in a plug flow reactor containing 1% (w/w) Pd/C catalyst in the presence of hydrogen. A standard containing undecane and dodecane in a decane solvent (top trace) and a blank decane sample (bottom trace) are shown for comparison.