Directed evolution has been used extensive to engineer enzymatic activity for the synthesis of fuels, chemicals, and bioactive compounds. The process of directed evolution consists of an iterative process of creating mutant libraries and choosing desired phenotypes through screening or selection until the enzymatic activity reaches the desired goal. Screening is accomplished through linking the desired property to some visual output using colorimetric or fluorometric assays, while selection links the desired property to survival of host cell. Between the two, selection is more desirable because it offers far higher throughput capabilities with lesser effort. Furthermore, in screening, every mutant in the library needs to be examined individually for the desired property, whereas using a selection enables elimination of the majority of undesirable candidates. However, selection is harder to implement since it requires identification of a unique selection mechanism for every given directed evolution undertaking. That is, given an enzymatic reaction, the challenge is to automatically identify a consumption pathway from the desired product to a metabolite native to the host cell. In our prior work, presented at International Workshop on Bio Design Automation, June 2014, we developed an algorithm, ASF (Automated Selection Finder), for constructing a selection pathway utilizing metabolites and reactions catalogued in a database (e,g., KEGG). Here, we improve on our earlier work by providing a systemic way to engineer the host cell to maximize the yield through the selection pathway.
Given a selection pathway, we describe in this work a computational method for coupling the pathway with cell survival. Our method aims to maximize the yield of the selection pathway. To reach this goal, our method restricts carbon uptakes, making the selection pathway the only possible carbon source within the host. The added pathway thus becomes essential for producing cellular biomass. Additionally, our method identifies possible knockout targets to improve the yield through the pathway. To assess the effectiveness of our approach, we applied our method to construct selection pathways for several beneficial metabolites including xylitol, aniline, methanol, and D-ribulose-1,5-bisphosphate. We used Escherichia coli as the host organism. For all compounds except aniline, high-consumption pathways were identified. Further, for xylitol and D-ribulose-1,5-bisphosphate, our method identified high-consumption pathways that were experimentally validated and confirmed in the literature.