High throughput screen for cells with high extracellular metabolite consumption--secretion rates using microfluidic droplets

Abstract

Metabolic engineering has contributed significantly to the improvement of strains for the industrial production of various compounds. Traditionally, enzymatic steps closely associated with the product-forming pathway have been engineered to prune side reactions and eliminate kinetic bottlenecks. Other, so-called distal genes may also impact production in a profound way due to (often unknown) kinetic and regulatory effects. Inverse Metabolic Engineering (IME) emerged as an approach to identify such distal genetic factors. IME employs combinatorial methods whereby libraries are constructed harboring random genomic variants of the host or other strains, cells with superior properties are selected, and genetic inserts impacting the superior phenotype are characterized. While many strategies can be deployed in library construction, broad applicability of IME to strain improvement for overproduction of secreted metabolites is severely limited by the availability of high-throughput methods for selecting strains with significantly improved metabolite secretion or uptake rates. As soon as a metabolite is secreted, its association with the cell that secreted it is lost. Thus, each clone must grow in a separate environment to allow for the measurement of clone-specific metabolite concentrations. Furthermore, since many mutant libraries are large (=/>104 unique clones), a high throughput screening platform must be used to culture and measure each unique clone.

(cont.) Traditional methods such as microwell plates for culturing and assaying can be utilized. However, this method is laborious, expensive, and low-throughout. Other selection systems have been developed which are very specialized for the metabolite of interest so it is difficult to apply these systems to other applications. Here, we will describe a flexible high throughput screening platform which utilizes microfluidics to encapsulate cells in monodisperse nanoliter aqueous droplets surrounded by an immiscible fluorinated oil phase. This system contains integrated modules for cell culturing, measurement of an extracellular metabolite with a fluorescent enzymatic assay, and sorting. To demonstrate the functionality and flexibility of this system, high xylose and glucose consuming Saccharomyces cerevisiae strains were enriched from mixtures of known strains. These high consuming strains could be identified even in a cell population of 1 cell/104. Several S. cerevisiae and Escherichia coli libraries were also screened in this system. One of these libraries was a genomic DNA library which was screened for high xylose consumption, an important phenotype for the utilization of lignocellulosic feedstocks. This library was constructed using DNA from an evolutionary engineered strain containing the Piromyces sp. E2 xylose isomerase gene. This library was transformed into the original unevolved strain to identify the beneficial mutations which occurred during the strain evolution.

(cont.) After screening the library and analyzing the highest xylose consuming strain, we determined that multiple copies of the xylose isomerase gene was one of the genomic changes responsible for high xylose consumption. This microfluidic screening system is general and has broad applications in strain selection for the overproduction of fuels, chemicals, and pharmaceuticals.