Comparative chemical-genomic profiling of plant-based hydrolysate toxins enables predictive assessment of responses to complex mixtures

Many toxins and stressors found in hydrolysates inhibit microbial metabolism and product formation, requiring mitigation strategies including strain engineering. To identify mechanisms of toxicity and targets for genetic engineering, we used a chemical genomics approach with a library of S. cerevisiae deletion mutants cultured anaerobically in dozens of individual compounds found in different types of hydrolysates, to explore shared and divergent gene requirements across inhibitors. Relationships in chemical-genomic profiles identified classes of toxins that provoked similar cellular responses, spanning inhibitor relationships that were not expected from chemical classification. Our results also revealed widespread antagonistic effects across inhibitors, such that the same gene deletions are beneficial for surviving some toxins but detrimental for others. As a proof of principle, we used the gene-deletion responses to single inhibitors to successfully predict strains whose fitness was improved in complex inhibitor mixes found in synthetic hydrolysates. We discuss the implications for strain engineering and the potential of this rich dataset for identifying engineering targets. Overall design: We conducted a chemical genomics experiment using a barcode-sequencing approach with a gene-deletion library of S. cerevisiae under anaerobic conditions, and quantified knockout strains growth, in triplicates, in cultures without (control) and with each of 45 inhibitory chemicals individually, including solvents used in pre-treatment, toxins generated during hydrolysis of plant material, and biofuel products that are stressful to cells at high levels.