Engineered Respiro-Fermentative Metabolism for the Production of Biofuels and Biochemicals from Fatty Acid-Rich Feedstocks

Although lignocellulosic sugars have been proposed as the primary feedstock for the biological production of renewable fuels and chemicals, the availability of fatty acid (FA)-rich feedstocks and recent progress in the development of oil-accumulating organisms make FAs an attractive alternative. In addition to their abundance, the metabolism of FAs is very efficient and could support product yields significantly higher than those obtained from lignocellulosic sugars. However, FAs are metabolized only under respiratory conditions, a metabolic mode that does not support the synthesis of fermentation products. In the work reported here we engineered several native and heterologous fermentative pathways to function in Escherichia coli under aerobic conditions, thus creating a respiro-fermentative metabolic mode that enables the efficient synthesis of fuels and chemicals from FAs. Representative biofuels (ethanol and butanol) and biochemicals (acetate, acetone, isopropanol, succinate, and propionate) were chosen as target products to illustrate the feasibility of the proposed platform. The yields of ethanol, acetate, and acetone in the engineered strains exceeded those reported in the literature for their production from sugars, and in the cases of ethanol and acetate they also surpassed the maximum theoretical values that can be achieved from lignocellulosic sugars. Butanol was produced at yields and titers that were between 2- and 3-fold higher than those reported for its production from sugars in previously engineered microorganisms. Moreover, our work demonstrates production of propionate, a compound previously thought to be synthesized only by propionibacteria, in E. coli. Finally, the synthesis of isopropanol and succinate was also demonstrated. The work reported here represents the first effort toward engineering microorganisms for the conversion of FAs to the aforementioned products.Concerns about climate change and the depletion and cost of petroleum resources have ignited interest in the establishment of a bio-based industry (5, 49, 61), and the conceptual model of a biorefinery has emerged (27, 28, 45). Given its abundance in nature, the carbohydrate portion of edible crops such as sugarcane, sugar beet, maize (corn), and sorghum is currently used as the primary feedstock in the biological production of fuels and chemicals (12, 49, 52). Although the use of nonedible lignocellulosic sugars has been proposed as an efficient and sustainable avenue to the aforementioned processes, the availability of fatty acid (FA)-rich feedstocks and recent progress in the development of oil-accumulating organisms make FAs an attractive alternative. Edible oil-rich crops such as rapeseed, sunflower, soybean, and palm are currently available and widely used as feedstocks for chemical conversion to biodiesel (6), while oleaginous algae and nonedible FA-rich crops along with industrial by-products are receiving greater attention as longer-term alternatives. These nonedible FA-rich feedstocks are presently generated in large amounts and can be exploited for the biological production of fuels and chemicals (14, 22, 51, 56, 57). Unfortunately, microbial platforms to enable this are at present almost absent.FAs not only are abundant but also offer several advantages when used for fuel and chemical production. For example, their metabolism to the key intermediate metabolite acetyl coenzyme A (acetyl-CoA) is very efficient, as it results in 100% carbon recovery (Fig. ​(Fig.1).1). Since many fuels and chemicals can be derived from acetyl-CoA, high yields can be realized if FAs are used as the carbon source. In contrast, sugar metabolism generates one molecule of carbon dioxide (or formate) per molecule of acetyl-CoA, limiting the yield of products derived from acetyl-CoA (Fig. ​(Fig.1).1). The product yield advantage of FAs over sugars is also supported by the more highly reduced nature of their carbon atoms. Table ​Table11 provides a comparison of maximum theoretical yields, on both weight and carbon bases, for the production of biofuels and biochemicals from FAs and lignocellulosic sugars. Maximum theoretical yields have been calculated from stoichiometry based on the pathways shown in Fig. ​Fig.11 for the utilization of FAs and glucose, the synthesis of products, the tricarboxylic acid (TCA) cycle, and oxidative phosphorylation. The stoichiometric coefficients were obtained by conducting elemental balances on carbon, hydrogen, and oxygen, and an ATP balance was also included in the analysis. As an example, when production of biofuels (e.g., ethanol and butanol) is considered, utilization of FAs (e.g., palmitic acid C_16:0]) as a substrate returns product yields 2.7-fold (wt/wt) or 1.4-fold (C/C) higher than those for sugars (calculations are provided for glucose but are equally valid for other lignocellulosic sugars). Although the current prices of feedstocks on a weight basis are comparable (lower than 20¢/pound), the data reported in Fig. S1a in the supplemental material show that the price per carbon for glucose derived from corn is remarkably higher. Regardless of the basis used for calculations (i.e., weight or carbon basis), when maximum theoretical yields and costs of FA and sugar feedstocks are accounted for, the advantages of using FAs are remarkable (see Fig. S1b in the supplemental material).Open in a separate windowFIG. 1.Pathways engineered in E. coli for the conversion of fatty acids to fuels (red) and chemicals (green). Also shown is the catabolism of fatty acids via the β-oxidation pathway (orange) and of glucose through the Embden-Meyerhof-Parnas pathway (blue). Relevant reactions are represented by the names of the genes coding for the enzymes (E. coli genes unless otherwise specified in parentheses as follows: C. acetobutylicum, ca; C. beijerinckii, cb): aceA, isocitrate lyase; aceB, malate synthase A; adc, acetoacetate decarboxylase (ca); ackA, acetate kinase; adh, secondary alcohol dehydrogenase (cb); adhE, acetaldehyde/alcohol dehydrogenase; adhE2, secondary alcohol dehydrogenase (ca); atoA and atoD, acetyl-CoA:acetoacetyl-CoA transferase; atoB, acetyl-CoA acetyltransferase; bcd, butyryl-CoA dehydrogenase (ca); crt, crotonase (ca); etfAB, electron transfer flavoprotein (ca); fadA, 3-ketoacyl-CoA thiolase; fadB, enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase; fadD, acyl-CoA synthetase; fadE, acyl-CoA dehydrogenase; hbd, β-hydroxybutyryl-CoA dehydrogenase (ca); icd, isocitrate dehydrogenase; pta, phosphate acetyltransferase; sdhABCD, succinate dehydrogenase; scpA, methylmalonyl-CoA mutase; scpB, methylmalonyl-CoA decarboxylase; scpC, propionyl-CoA:succinate CoA transferase; sucA, 2-oxoglutarate dehydrogenase; sucB, dihydrolipoyltranssuccinylase; and sucCD, succinyl-CoA synthetase. Abbreviations: 2H] = NADH = FADH₂ = QH₂ = H₂; P/O, amount of ATP produced per oxygen consumed in the oxidative phosphorylation.

TABLE 1.

Comparison of maximum theoretical yields for the production of biofuels and biochemicals from fatty acids (palmitic acid) and lignocellulosic sugars (glucose)

Pathway stoichiometry for the synthesis of the specified product from glucose (C6H12O6) or palmitic acid (C16H32O2)a	Maximum yield (wt basis/C basis)
Biofuels
Ethanol (C2H6O)
C6H12O6 → 2C2H6O + 2CO2	0.51/0.67
C16H32O2 → 23/3C2H6O + 2/3CO2	1.38/0.96
C16H32O2 + 51/7H2O → 53/7C2H6O + 6/7CO2 + 8/7H]; 8/7H] + 2/7O2 → 4/7H2O	1.36/0.95
Butanol (C4H10O)
C6H12O6 → C4H10O + 2CO2 +H2O	0.41/0.67
C16H32O2 + 7/2H2O → 53/14C4H10O + 6/7CO2 + 8/7H]; 8/7H] + 2/7O2 → 4/7H2O	1.10/0.95
Biochemicals
Acetate (C2H4O2)
C6H12O6 + 2H2O → 3C2H4O2	1.00/1.00
C16H32O2 + 7H2O + 7CO2 → 23/2C2H4O2	2.70/1.44
Acetone (C3H6O)
C6H12O6 → 3/2C3H6O + 3/2CO2 + 3/2H2O	0.48/0.75
C16H32O2 + 5/4H2O + 5/4CO2 → 23/4C3H6O	1.30/1.08
Isopropanol (C3H8O)
C6H12O6 → 4/3C3H8O + 2CO2 + 2/3H2O	0.44/0.67
C16H32O2 + 40/9H2O → 46/9C3H8O + 2/3CO2	1.20/0.96
Succinate (C4H6O4)
C6H12O6 + 6/7CO2 → 12/7C4H6O4 + 6/7H2O	1.12/1.14
C16H32O2 + 152/17CO2 + 86/17H2O → 106/17C4H6O4 + 80/17H]; 80/17H] + 20/17O2 → 40/17H2O	2.87/1.56
Propionate (C3H6O2)
C6H12O6 → 12/7C3H6O2 + 6/7CO2 + 6/7H2O	0.70/0.86
C16H32O2 + 262/83CO2 + 370/83H2O → 530/83C3H6O2 + 216/83H]; 216/83H] + 54/83O2 → 108/83H2O	1.81/1.20

Open in a separate window^aStoichiometry is based on the pathways shown in Fig. ​Fig.11 for the utilization of FAs and glucose, the synthesis of products, the TCA cycle, and oxidative phosphorylation. For the synthesis of biochemicals, CO₂ fixation via the Wood-Ljungdahl pathway (50) (2CO₂ + ATP + 8H] → acetyl-CoA) or the carboxylation of phosphoenolpyruvate (54) (phosphoenolpyruvate + CO₂ → oxaloacetate + ATP) were also considered (not shown in Fig. ​Fig.1).1). The stoichiometric coefficients were obtained by conducting elemental balances on carbon, hydrogen, and oxygen. An ATP balance was also included in the analysis for the reactions shown in italics. All other reactions represent ATP-generating pathways. Every acetyl-CoA oxidized through the TCA cycle generates three NADH, one reduced flavin adenine dinucleotide (FADH₂), and one ATP equivalent. Eleven ATPs can be generated from the oxidation of the NADH and FADH₂ produced in the TCA cycle (two and three ATPs per FADH₂ and NADH, respectively) via coupling between the electron transfer chain and oxidative phosphorylation.Despite the aforementioned advantages, biological conversion of FA-rich feedstocks has been exploited only for the production of polyhydroxyalkanoates (46, 47), with no report to date of organisms engineered for the conversion of FAs to fuels and chemicals (see the text in the supplemental material for more details).Escherichia coli is one of the most amenable organisms to industrial applications and has been engineered for biofuel production (52). The utilization of FAs in E. coli is mediated by enzymes encoded by the fad regulon and the ato operon (11) (Fig. ​(Fig.1).1). Products of the fad regulon mediate the transport, acylation, and β-oxidation of medium-chain (C₇ to C₁₁) and long-chain (C₁₂ to C₁₈) FAs. Two additional enzymes encoded by the atoD-atoA and atoB genes (part of the atoDAEB operon) are also required for the growth of E. coli on short-chain (C₄ to C₆) FAs (25). The expression of the fad regulon and ato operon is controlled by FadR (fadR) and AtoC (atoC), respectively (44).While advantageous, the high degree of reduction of carbon in FAs also poses a metabolic challenge because their average degree of reduction per carbon is higher than in biomass. Therefore, the incorporation of fatty acids into cell mass generates reducing equivalents (Fig. ​(Fig.1)1) and hence requires the presence of an external electron acceptor. That is, the aforementioned pathways are active only in the respiratory metabolism of FAs, which leads to the synthesis of cell mass and carbon dioxide but no other metabolic product. Therefore, fuel and chemical production from FAs requires the engineering of a respiro-fermentative metabolic mode that would support the synthesis of fermentative products during respiratory metabolism of FAs. To this end, we metabolically engineered native and heterologous pathways for the efficient catabolism of FAs and the synthesis of fuels and chemicals in E. coli. Biofuels, commodity chemicals, and polymer building blocks were chosen as model products to illustrate the feasibility of the proposed approach.