Engineering Cross-Feeding Co-Cultures As a Platform for High-Throughput Screening of Microbial Strain Libraries for Enhanced Biomolecule Production

A variety of methods have been developed to explore vast phenotype landscapes in a controlled manner. The applicability of these methods to the development of bacterial production strains is, however, limited by the requirement of a screening method with high enough throughput to investigate all the variants generated. Products that require chromatography-based screens typically reach a throughput of only ~10² variants per day, substantially lower than those with colorimetric or growth-based assays. Here, we present a new method for screening Escherichia coli strains for improved production of amino acids and/or their intermediates. Amino acid intermediates can be diverted towards production of a variety of other useful compounds, including higher alcohol biofuels.

This method uses a cross-feeding auxotroph system to convert production levels of amino acid or amino acid intermediates into co-culture growth characteristics that are detectable in a high-throughput context. As proofs of concept we examine two target molecule cases: tryptophan and 2-ketoisovalerate (a precursor of valine, leucine, and the drop-in biofuel isobutanol). We demonstrate that suitable pairs can be selected so that improvements in production by the production strain lead to increased co-culture growth rate and changes in the final composition of the population. We explore implementation of this co-culture screening method on petri dishes and in microdroplets, demonstrating its ability to identify rare over-producers from a library. We investigate the dynamic ranges of our model systems and ways in which they may be expanded.