Production of Petrol Using Engineered E Coli Strains

The production of petrol using engineered E coli strains

Frontiers in Interdisciplinary Bioscience

Introduction

Concerns over the global environment has led to more efforts to use greener fuel options. However, there is still an increasing demand for transport fuels including petrol and so there has been a focus on developing biofuels that are eco-friendly and sustainable, Previous reports have shown that diesel can be produced microbially(Howard et al., 2013) but production of petrol has not yet been established. Petrol consists of a homogeneous mixture of small hydrocarbons with chain lengths of 4-12 carbons including paraffins and cycloalkanes (Altin et al., 2004). Short-chain alcohols have been investigated (Atsumi et al., 2008, Jun Choi et al., 2012) but their properties are significantly inferior to petrol. A promising alternative is the use of hydrocarbons, specifically short chain alkanes (SCAs) owing to their many advantages, including high energy content (Lennen et al., 2010). It is proposed that petrol can be produced by using lignocellulosic hydrolysates as a feedstock. Lignocellulose biomass is composed of two carbohydrate polymers, cellulose and hemicellulose, and the non-carbohydrate phenolic polymers, lignin. Hydrolysis of the polysaccharides produces a series of monosaccharides which are used for fermentation processes. Glucose is particularly interesting as it can be preferentially consumed by Escherichia coli (E coli).  This paper presents a process for the production of petrol, using lignocellulosic hydrolysates as a feedstock with E coli strains engineered metabolically and enzymatically to produce the desired short chain alkanes (SCAs).

Figure 1: Summary of the biological process for the production of petrol

Metabolic and Enzyme engineering

There are numerous engineered E coli strains that can metabolise many sugars from xylose to arabinose from lignocellulosic biomass however all sugars lag behind glucose due to the preferential uptake by E coli. This is due to a phenomenon called Carbon Catabolite Repression (CCR) (Bailey et al., 1986). Since glucose is a key material in the biological production of petrol and is a significant portion of the lignocellulosic hydrolysates, modifications have to be made to disrupt or completely stop the effect of CCR.  Therefore, we will use a metabolically engineered E coli strain which will grow using all monosaccharides found in lignocellulosic hydrolysates exceptglucose. (Sun et al., 2018). Since we do not want E coli to uptake glucose, improvement of xylose consumption will be done using a strain which has xylH naturally mutated. XylH encodes for high affinity xylose transport into the cell. To disrupt utilisation of glucose, several genes will be modified in the phosphotransferase system for carbohydrate transport to prevent glucose entering E coli (Gosset., 2005). The ptsG, manZ and glk genes will be modified sequentially due to their essential contribution for glucose uptake and metabolism in E coli. 

Figure 2: Modification of genes in the phosphotransferase system

The engineered strain which will be used is B0013-2020H as E coli does not grow on a series of different temperatures. It is required that we separate the glucose from the other monosaccharides. Therefore, we propose to set up batch reactors in series and extract the glucose from one reactor and send it through to the adjacent reactor where it will be used by a different E coli strain for the eventual production of short-chain alkanes (SCAs). It is worth pointing out that glucose concentration decreases slightly when cell mass is large due to the cell exhausting all other sugars in increasing its size and mass. 

Figure 3: Utilisation of glucose from metabolically engineered E coli strain B0013-2020H (Adapted from: Consumption of Lignocellulosic hydrolysates by E coli B0013-2020H, Sun et al., 2018)

Figure 4: Modified metabolic pathway for utilisation of glucose (Sun et al., 2018)

The metabolic pathway has been successfully engineered to incorporate the consumption of other monosaccharides in lignocellulosic hydrolysates for growth while restricting E coli’s use of glucose for further processing. The pathway has also been engineered to remove the conversion of glucose to ethanol. To improve the conversion of lignocellulosic biomass to monosaccharides, our process will use an optimized cocktail of enzymes which will include cellulases and hemicellulases. They will increase our yield of monosaccharides which can be used for growth of E coli strains, ultimately increasing short chain alkane production (Gao et al., 2011).

In a recent study, it was demonstrated that glucose can be utilized to produce short-chain alkanes through the metabolic engineering of several different E coli strains (Choi and Lee, 2013). The process is outlined below:

Figure 5: Metabolic engineering of E coli strains for the production of short-chain alkanes (Choi and Lee, 2013)

Initially, glucose enters E. coli strain W3110 which has had the fadR gene, a global transcriptional regulator protein, deleted using the one-step deactivation method. The deletion of this particular gene will enhance the formation of short-chain fatty acids  as it helps prevent fabA and fabB upregulation which are both responsible for the synthesis of unsaturated fatty acids. (Nunn et al., 1983). The conversion of fatty acyl-acyl carrier protein (ACP) to free fatty acids (FFAs) is an important metabolic process. It was found that expression of TesA thioesterase enzyme provided the highest amount of FFAs, however it is known that this enzyme preferentially hydrolyses long chain fatty acids (Steen et al., 2010).Therefore, it is necessary to structurally engineer an enzyme capable of hydrolysing short-chain FFAs. It is not sufficient to simply increase the level of TesA because a specific ratio between each component including fatty acid synthase and TesA is required for optimal production (Shin et al., 2016). Expression of TesA with the mutation L109P has been found to hydrolyse short-chain FFAs. The crystal structure of the mutated enzyme undergoes structural changes in which Leucine 109 (Leu109) mutates to Proline 109 (Pro 109) which abolishes a conformational change called the ‘switch loop movement which is acyl chain length dependent (Lo et al., 2005). The movement of the switch loop influences substrate specificity by stabilizing the Michaelis complex (MC) (Lo et al., 2005). Therefore Tes(L109P) will be used in the process due to its altered substrate specificity. This not only increases yield of FFAs but also allows for controlling the chain length. In our process, we want to maximize the amount of FFAs produced, so the fadE gene will be deleted to block beta-oxidation which created a new strain known as the GAS1 where the fadD gene was also deleted like the W3110 strain. The fadE gene facilitates the conversion of fatty acyl-CoA to 2-trans-enoyl-CoA, an unsaturated compound. This is a highly undesired by-product since the desired products are SCAs, which are saturated compounds. Furthermore, deletion of fadD gene via one-step inactivation is done to prevent conversion of FFAs to fatty acyl-CoA (coenzyme A).  Overall, the synthesis of short-chain FFAs will be optimised by deletion of fadE, fadR and fadD genes (Choi and Lee., 2013). Future work will look at further engineering the tesA enzyme by substitution of Arg64 for Cys64 to improve FFA production further (Shin et al., 2016).

For conversion of FFAs to fatty-acyl CoA’s,  E. coli is able to utilize long-chain FFAs however not able to grow on short-medium sized FFAs. However, with overexpression of the fadD gene, recombinant E. coli (GAS2 strain) can grow on short-chain FFAs (C8-C10 FFAs) (Zhang et al., 2006). So the fadD gene is amplified by plasmid-based overexpression under a tac promoter. pTacFad was constructed to clone to the tac promoter in front of a fadD gene. To further increase the fadD level, it will be amplified with primers fadf and fadr using the genomic DNA of E. coli as a template. Overexpression of the fadD gene can induce the beta-oxidation pathway so in our process we will ensure that the fadE gene is deleted. Following this step is the conversion to fatty aldehydes by using the acr gene encoding a fatty acyl co-A reductase. The gene will be introduced by plasmid-based overexpression under the trc promoter.  To produce short chain alkanes, the CER1 gene encoding the fatty aldehyde decarbonylase enzyme(Aarts et al., 1995) will be introduced and subsequently overexpressed via plasmids under the tac promoter.

Through fed-batch fermentation, the final engineered strain (GAS3) with two promoters pTacCer1FadD and pTrcAcR’TesA(L109P) produced a reasonable yield of final product at 31˚C but the literature states that there was a significant amount of short-chain fatty alcohols produced. It has been discovered previously that overexpression of fatty acyl-coA reductase can convert fatty aldehydes to fatty alcohols (Schirmer et al., 2010).  Short-chain hydrocarbons (SCHCs) production will be enhanced by increasing the activity of the fatty aldehyde decarbonylase enzyme. Enzymes work best at optimal temperatures so operating the fed-batch fermentation process just one degree at 30˚C resulted in the enzyme activity was 1.7 fold higher than at 31˚C through SDS-PAGE analysis (Choi and Lee, 2013). Such an increase in activity for a minor change in temperature would suggest a change in the enzyme structure to better conform with the substrate. Given the specific and sensitive nature of enzymes, the reaction temperature needs to be carefully controlled to obtain optimal enzyme performance. There was no alcohol conversion which suggests that at 30˚C decanal preferentially converts to nonane as opposed to decanol (Choi and Lee, 2013), which is preferential to our process.. This is shown through GC-MS analysis which provides clear evidence for the removal of the nonane signal at 30˚C through the differences in the product metabolite profiles shown in Figure 6.  Our process will keep the production reactor at a constant temperature of 30˚C to ensure maximum production.

Figure 6: GC-MS metabolite profile of fermentation products (Choi and Lee, 2013)

Immobilisation

To retain biomass in the bioreactor, the engineered E coli strains will be in immobilised form. Immobilisation is the confinement of viable cells within a certain area of space and there are several main types including cell entrapment, flocculation and encapsulation (Freeman and Lilly, 1998). Advantages of immobilised cells include their resistance to changing environmental conditions and toxic substances. However, when immobilising cells, particularly at an industrial level, there are several engineering constraints that have to be considered, including a support which allows effective diffusion leading to effective cell growth, stability and control under operation. Failure to adhere to these constraints for a given system can spoil cell growth conditions.

Entrapment is a common immobilisation technique with gels often used due to long operational stabilities of at least 100 h (Freeman and Lilly, 1998). In our process, we will look to utilize the common entrapment material Calcium alginate. Recently, it was shown that immobilised E coli strains utilising pentoses and hexoses in Ca-alginate beads for conversion to ethanol had a successful, long-term operational stability. The conversion of ethanol was high-yielding and operated through repeated batch reactors (Zajkoska et al., 2013). We will look to utilise the same method of operation except our strains will have different purposes. A major concern for immobilised cell systems is mass transfer as it not known if the conditions in the bulk solution and the matrix are the same because the matrix can act as a diffusion barrier. In all mass transfer models, the size of the immobilised cells is the most important parameter. Internal mass transfer can be described by the Damkohler number:   $\mathit{Da}=\frac{k_2{r_0}^2}{D_e{K}_s}$

r₀, the size parameter, is one of the few parameters in the diffusion model of that can be measured with accuracy. There are several ways the Damkohler number can be calculated experimentally so it is not easy to compare literature values but the calculation provides an estimate for mass transfer limitations. When Da is much less than 1, diffusion occurs much faster than the rate of the reaction so it is important in our process that the size of the particles are not so small that diffusion dominates but we want the size to be considerably small as this will limit internal diffusion.  

Our design choice for our process is very deliberate as it takes into consideration that each intermediate product is produced by different E coli strains. There will be a total of seven batch reactors. The initial batch reactor will contain the hydrolysates which will be purified and then used by the first E coli strain (B0013-2020H) to produce glucose at maximum concentration. The glucose will then be up taken by E coli W3110. For each intermediate product produced from different E coli strains, centrifugation will occur to ensure purity and minimise cross-contamination between intermediates. The batch reactors used will be small in size to limit fermentation times. As a process, we feel that this is the best way to produce the maximum of SCAs of high purity.

Figure 7: Process flow diagram for the production of short-chain alkanes (Petrol)

References

• Aarts, M., Keijzer, C., Stiekema, W. and Pereira, A. (1995). Molecular Characterization of the CER1 Gene of Arabidopsis Involved in Epicuticular Wax Biosynthesis and Pollen Fertility. The Plant Cell, 7(12), p.2115.
• Altin,O. & Eser,S. (2004) Carbon deposit formation from thermal stressing of petroleum fuels. Am. Chem. Soc. Div. Fuel Chem. 49, pp. 764–766.
• Atsumi, S., Hanai, T. and Liao, J. (2008). Non-fermentative pathways for synthesis of branched-chain higher alcohols as biofuels. Nature, 451(7174), pp.86-89.
• Bailey, JE. and Ollis, DF. (1986) Biochemical Engineering Fundamentals.2^nd Ed. Singapore: McGraw-Hill. p.321.
• Choi, Y. and Lee, S. (2013). Microbial production of short-chain alkanes. Nature, 502(7472), pp.571-574.
• Freeman, A. and Lilly, M. (1998). Effect of processing parameters on the feasibility and operational stability of immobilized viable microbial cells. Enzyme and Microbial Technology, 23(5), pp.335-345.
• Gao, D., Uppugundla, N., Chundawat, S., Yu, X., Hermanson, S., Gowda, K., Brumm, P., Mead, D., Balan, V. and Dale, B. (2011). Hemicellulases and auxiliary enzymes for improved conversion of lignocellulosic biomass to monosaccharides. Biotechnology for Biofuels, 4(1), p.5.
• Gosset, G., 2005. Improvement of Escherichia coli production strains by modification of the phosphoenolpyruvate:sugar phosphotransferase system. Microbial Cell Factories , 4(14), pp.1–11.
• Howard, T., Middelhaufe, S., Moore, K., Edner, C., Kolak, D., Taylor, G., Parker, D., Lee, R., Smirnoff, N., Aves, S. and Love, J. (2013). Synthesis of customized petroleum-replica fuel molecules by targeted modification of free fatty acid pools in Escherichia coli. Proceedings of the National Academy of Sciences, 110(19), pp.7636-7641.
• Jun Choi, Y., Hwan Park, J., Yong Kim, T. and Yup Lee, S. (2012). Metabolic engineering of Escherichia coli for the production of 1-propanol. Metabolic Engineering, 14(5), pp.477-486.
• Lennen, R., Braden, D., West, R., Dumesic, J. and Pfleger, B. (2010). A process for microbial hydrocarbon synthesis: Overproduction of fatty acids in Escherichia coli and catalytic conversion to alkanes. Biotechnology and Bioengineering, 106(2), pp.193-202.
• Lo, Y., Lin, S., Shaw, J. and Liaw, Y. (2005). Substrate Specificities of Escherichia coli Thioesterase I/Protease I/Lysophospholipase L1 Are Governed by Its Switch Loop Movement. Biochemistry, 44(6), pp.1971-1979.
• Nunn,W.D.,Giffin,K.,Clark,D.& Cronan,J.E.Jr. (1983). Role for fadR in unsaturated fatty acid biosynthesis in Escherichia coli. J.Bacteriol. 154, pp. 554–560.
• Schirmer, A., Rude, M., Li, X., Popova, E. and del Cardayre, S. (2010). Microbial Biosynthesis of Alkanes. Science, 329(5991), pp.559-562.
• Shin, K., Kim, S. and Lee, S. (2016). Improvement of free fatty acid production using a mutant acyl-CoA thioesterase I with high specific activity in Escherichia coli. Biotechnology for Biofuels, 9(1).
• Steen, E., Kang, Y., Bokinsky, G., Hu, Z., Schirmer, A., McClure, A., del Cardayre, S. and Keasling, J. (2010). Microbial production of fatty-acid-derived fuels and chemicals from plant biomass. Nature, 463(7280), pp.559-562.
• Sun, J., Tian, K., Wang, J., Dong, Z., Liu, X., Permaul, K., Singh, S., Prior, B. and Wang, Z. (2018). Improved ethanol productivity from lignocellulosic hydrolysates by Escherichia coli with regulated glucose utilization. Microbial Cell Factories, 17(1), pp.1-8.
• Zajkoska, P., Rebroš, M. and Rosenberg, M. (2013). Biocatalysis with immobilized Escherichia coli. Applied Microbiology and Biotechnology, 97(4), pp.1441-1455.
• Zhang, H., Wang, P. and Qi, Q. (2006). Molecular effect of FadD on the regulation and metabolism of fatty acid in Escherichia coli. FEMS Microbiology Letters, 259(2), pp.249-253.