Advances in Engineering of Bacterial Methyl Ketone Synthesis for Biofuels

We have engineered Escherichia coli to overproduce saturated and monounsaturated
aliphatic methyl ketones in the C₁₁ to C₁₇ (diesel)
range; this group of methyl ketones includes 2-undecanone and 2-tridecanone,
which have favorable cetane numbers and are also of
importance to the flavor and fragrance industry.  We have made specific improvements that
resulted in more than 10,000-fold enhancement in methyl ketone titer relative
to that of a fatty acid-overproducing E.
coli strain, including the following: (a) overproduction of beta-ketoacyl-coenzyme A (CoA) thioesters achieved by
modification of the beta-oxidation pathway (specifically, overexpression
of a heterologous acyl-CoA oxidase and native FadB,
and chromosomal deletion of fadA) and (b) overexpression of a native thioesterase
(FadM). The first generation of engineered E. coli (Goh et al. 2012) produced ~380 mg/L of methyl ketones in rich medium.  We
have subsequently made additional genetic modifications that included balancing
overexpression of fadR
and fadD to
increase fatty acid flux into the pathway, consolidation of the pathway from
two plasmids into one, codon optimization, and knocking out key acetate
production pathways (Goh et al. 2014). These modifications have led
to a methyl ketone titer of 1.4 g/L with 1%
glucose in shake flask experiments, which represents 40% of the maximum
theoretical yield, and also attained titers of 3.4 g/L after ~45 h of fed-batch
glucose fermentation (the best values reported to date).  We have also conducted in vitro assays with purified pathway
enzymes, which confirmed that a decarboxylase is not required to convert beta-keto acids
into methyl ketones and that FadM is promiscuous and
hydrolyzes not only beta-ketoacyl-CoAs but also other
CoA-thioester pathway intermediates. These in vitro results have provided insight
on how to fine-tune expression of pathway genes for further optimization of
methyl ketone production.  As part of efforts to further improve methyl
ketone production, metabolic modeling was used to identify gene deletions that
could improve flux through the methyl ketone pathway.  One of the specified knockouts, ΔscgE,
which was annotated as a homolog of Rpe
(ribulose-5-phosphate epimerase) in the pentose
phosphate pathway, improved methyl ketone production by >50% relative to the
base strain (EGS1710). ¹³C-glucose experiments have recently been
performed on the knockout strain along with the DH1 wild-type
and base strains to obtain more comprehensive metabolic flux profiles that will
be leveraged to enable additional improvements in methyl ketone production.