Metabolomic Characterization of Escherichia coli-Based Cell-Free Expression Systems for Process Optimization
An important question about cell-free expression (CFE) reactions is why transcription and translation eventually degrade and stop. Commonly-suggested possible causes include exhaustion of transcription or translation substrates, exhaustion of energy sources, cofactor imbalances, buildup of toxic products, or degradation of transcriptional or translational machinery. Something in common to most of these hypotheses is that they involve small-molecule biochemical intermediates - molecules which, if they were still in a cell, we would call metabolites. To this end, we suspected that metabolomics (the systems-scale measurement and analysis of small biochemical intermediates) could be useful in characterizing when, why, and how transcription and translation eventually stop in CFE reactions. Here, we describe our efforts to use metabolomics to relate lysate preparation and performance to metabolic activity. We show that, as expected, lysate processing affects the metabolite makeup of lysates. Nonetheless, we identify some surprising trends, including multiple metabolites that are present in greater amounts in dialyzed lysate than in undialyzed lysate. We also show that, as one might expect, lysate metabolite levels change over the course of a CFE reaction. However, what was not expected was that these changes strictly due to the time of the CFE reaction dwarf any metabolite changes attributable to transcriptional and translational activity, and they occur regardless of whether a target molecule is produced. This "endogenous metabolism" of the CFE reactions is an important, poorly-characterized phenomenon that may contribute substantially to the temporal degradation of CFE. We then used our insights to devise new small molecule supplements that improve CFE protein production. These approaches and insights could be valuable in the future as more groups move towards CFE at scale, for which clear markers of lysate activity and supplementation approaches that compensate for potential variability across batches and experimental protocols would be extremely useful.