Bioinformatics approaches for engineering L-Tyrosine production in Escherichia coli

Aromatic compounds, such as aromatic amino acids, petrochemical aromatics, and aromatic polymers, are very important industrial materials. Among them, L-tyrosine (L-Tyr) has been considered an appealing metabolite to produce thanks to its wide variety of applications in the pharmaceutical and chemical industries. This metabolite is an important precursor of a diverse number of secondary metabolites or natural products, such as phenylpropanoic acids, flavonoids, curcuminoids, stilbenoids, Parkinsons disease drug 3,4-dihydroxy-L-phenylalanine (L-DOPA) and melanin [1][4]. There are several studies where Escherichia coli L-Tyr overproducing strains are used, but there is still a need to increase its yield to make the bioprocess economically feasible [1], [2]. Numerous computational tools have been developed for strain design that identify genetic modification strategies that increase targeted biochemicals production [5]. The construction of a biochemical network map, and the development of mathematical models and simulations that reproduce experimental data and phenotypes under different conditions, allow computer-aided cell design [6]. Stoichiometric models alone cannot quantitatively measure the effect of concentration levels, enzyme saturation and regulation. Kinetic models yield a system of ordinary differential equations that describe the time evolution of metabolite concentrations, enzyme activities and reaction fluxes, therefore have the potential to capture these interdependencies [5], [7]. These types of models may be used to model E. coli to overproduce L-Tyr. The aim of this work was to perform the in silico insertion of L-Tyr pathway in a kinetic model of the central carbon metabolism of E. coli, which will allow to identify the best target genes for the design of the L-Tyr overproducing strain. Three stoichiometric models, iML1515 [8], iJO1366 [9] and iJR904 [10], as well as three kinetic models, the Millard [11], the Oliveira [12] and the Jahan [6], were used to design an accurate model containing all the reactions needed to produce L-Tyr. This new model was then used to test the different knock-in/out strategies to design an E. coli L-Tyr overproducing strain.