CASOP: a computational approach for strain optimization aiming at high productivity.

The identification of suitable intervention strategies increasing the productivity of microorganisms is a central issue in metabolic engineering. Here, we introduce a computational framework for strain optimization based on reaction importance measures derived from weighted elementary modes. The objective is to shift the natural flux distribution to synthesis of the desired product with high production rates thereby retaining the ability of the host organism to produce biomass precursors. The stoichiometric approach allows consideration of regulatory/operational constraints and takes product yield and network capacity--the two major determinants of (specific) productivity--explicitly into account. The relative contribution of each reaction to yield and network capacity and thus productivity is estimated by analyzing the spectrum of available conversion routes (elementary modes). A result of our procedure is a reaction ranking suggesting knockout and overexpression candidates. Moreover, we show that the methodology allows for the evaluation of cofactor and co-metabolite requirements in conjunction with product synthesis. We illustrate the proposed method by studying the overproduction of succinate and lactate by Escherichia coli. The metabolic engineering strategies identified in silico resemble existing mutant strains designed for the synthesis of the respective products. Additionally, some non-intuitive intervention strategies are revealed.