In silico optimization of the production of amino-acids in Escherichia coli

The increasing need to replace chemical synthesis of compounds of interest by more environmentally friendly biological processes is driving the research for microbial cell factories. The industrial production of amino and organic acids includes several examples of success stories using microorganisms to convert inexpensive substrates into added value products. Traditionally, the design of such microbes relied on cycles of random mutagenesis followed by phenotypic selection [1], but a deeper knowledge of the microbial physiology allowed a more rational approach to this optimization problem [2,3]. However, this task is not straightforward, since the cell metabolism has proved to be highly complex and hard to predict. One of the approaches to tackle this problem is to use Systems Biology simulation tools to predict the microorganism behavior when subjected to genetic modifications. Using genome scale stoichiometric models, such as the latest iAF1260 for Escherichia coli [4] one can simulate a great diversity of possible metabolic phenotypes under steady state conditions by imposing flux-balance constrains. The use of flux balanced analysis (FBA) allows the determination of flux values through all the reactions in the network under a set of environmental conditions and genetic manipulations, by using an objective function, such as the maximization of growth [5]. In this work, we used genetic algorithms, such as OptGene [6] to search for sets of gene knockouts that result in the overproduction in silico of amino-acids in Escherichia coli. From all the proteinogenic amino-acids, glycine yielded the best results in the optimizations. A careful analysis of the in silico flux distribution in some of the mutants revealed an interesting and non-intuitive mechanism behind glycine accumulation. Furthermore, in these mutants the growth is coupled to the production of glycine, which makes them excellent candidates for in vivo implementation. We are reaching a point where bioinformatics tools are advanced enough to aid in complex tasks, such as the optimization of microbial cell factories. Here we described an effort to optimize in silico the production of amino-acids in Escherichia coli, which resulted in the discovery of a potential set of knock-outs that leads to glycine overproduction. This serves to show the increasing importance of in silico optimizations to aid in the metabolic engineering projects, especially to search for non-intuitive beneficial genome modifications