The development of generalizable methods for rapid optimization of metabolic pathways in yeast using CRISPR/Cas9

Metabolic engineering of the yeast Saccharomyces cerevisiae by altering the abundance of native genes or introducing heterologous genes has led to strains optimized for bio-renewable production of a diverse suite of molecules. However, the commercialization of engineered strains is bottlenecked by the slow rate of genomic modification required to alter gene expression. While the RNA-guided nuclease Cas9 (dCas9) has been repurposed as a plasmid-based transcriptional regulator that can be targeted to desired genes via co-expression with a single guide RNA (sgRNA), its application has typically been limited to binary on/off regulation of individual genes within a cell. This dissertation details work to broaden the scope of dCas9 regulation. First, design rules were developed to created graded gene expression by altering the sgRNA position within target gene promoters. “Stepping” dCas9 within target promoters allowed Systematic Testing of Enzyme Perturbation Sensitivities (STEPS) for discovery of rate limiting enzyme bottlenecks for glycerol and 3-dehydroshikimate production at each step in the strain engineering process, leading to a 5.7 and 7.8-fold increase in titer (respectively). Next, design rules were developed to allow simultaneous up and downregulation of multiple genes from a single plasmid by targeting the dCas9-VPR activator to the promoter region for overexpression and the ORF to block transcription. dCas9-VPR regulation was then streamlined to allow expression of multiple sgRNAs from compact tRNA-sgRNA-tRNA operons that can be synthesized via Ligation Extension of sgRNA Operons (LEGO). LEGO enabled assembly of a 5 sgRNA array to simultaneously downregulate NADH sinks within the cell while overexpressing the NADH-requiring 2,3- butanediol (BDO) pathway, increasing BDO titers 2-fold without any genomic modification. Lastly, a pooled plasmid library was synthesized to target graded expression to all 969 metabolic genes via a single transformation. This graded expression library allowed discovery of gene targets for glucose/xylose co-fermentation and benzyisoquinoline alkaloid production that are optimal at intermediate expression levels, and therefore would have been missed using traditional knockout screens. This work as a whole takes a significant step towards a future where all desired gene expression perturbations at all desired genes can be investigated via simple plasmid transformation, thus accelerating the rate of metabolic engineering