A Framework Based on Membrane Metrics to Improve the Performance of Biocatalysts

The cell membrane plays a central role in the fitness of all cells, including microbial cell factories, which are subjected to various stimuli from the fermentation broth. These stimuli often inhibit cell growth, decreasing productivity. The Jarboe Lab and others have extensively shown that engineering strategies targeting the cell membrane can enhance tolerance and bio-production. To investigate the effect of multi-stressors in the fermentation broth and prioritize membrane engineering strategies according to specific conditions, we developed a framework for identifying informative membrane features for improving cell growth and productivity in multi-stressor conditions. We modulated the membrane lipid composition via the expression of des, fabA, and fabB with three different strength promoters. We focused on the membrane lipid composition because of its fast adjustment depending on the environmental condition. Rigorous characterization of membrane composition and properties across several individual stressor conditions allowed us to correlate membrane metrics, such as the newly defined ratio of linear and non-linear membrane fatty acids (L/nL ratio), to outcomes given specific input or stimulus. We modeled 62 hypothetical scenarios with multi-stressor conditions and identified relevant membrane metrics predicted to impact cell growth. The L/nL ratio and the membrane hydrophobicity were the membrane metrics predicted to impact cell growth with the highest frequency. We selected the scenario of 42 °C, 2.5 mM furfural, and 2% v/v ethanol in minimal media with a modest connection between a membrane metric and cell growth for experimental validation. We found that the membrane hydrophobicity significantly correlated with cell growth and productivity (ethanol) in this multi-stressor condition experimentally. This work demonstrated that membrane properties can predict the performance of biocatalysts in single and multiple inhibitory conditions and possibly as an engineering target. In this manner, membrane properties can be used as screening or selection metrics for library- or evolution-based strain engineering.