UC Berkeley

Pseudomonas putida is an important chassis strain for the valorization of lignin. While we know much about this host and have developed numerous tools to engineer it, we still lack a complete understanding of its accessory genes that may be important for a variety of metabolisms and stress responses. Using a functional genomics method called barcode abundance sequencing (BarSeq), thousands of conditionally essential genes can be assayed in a pooled manner to elucidate their functions. With these assays novel metabolisms and novel gene functions can be identified and used to inform genetic engineering efforts. To validate the practicality of this method, we first assayed the aromatic metabolic pathways of Pseudomonas putida KT2440, and then applied the initial step of hydroxycinnamate catabolism toward the synthesis of a curcuminoid. Following this success, we sought to characterize the other diverse metabolisms of P. putida. Fatty acids and alcohols are highly sought after bioproducts for use as fuels and commodity chemicals; however, P. putida is known to metabolize most of these compounds. We therefore used BarSeq to elucidate the metabolism of 23 of these compounds with applications ranging from surfactants to plastic monomers. The results of these assays could then be used to inform engineering efforts for the production of these molecules and their derivatives. Given the diversity of metabolisms, next we sought to analyze the nitrogen metabolisms of P. putida as these again have applications as commodity chemicals. To this end, 52 different nitrogen-containing compounds were assayed via BarSeq as sole nitrogen sources. These molecules have many overlapping gene requirements in their metabolism, thus requiring a method for effectively analyzing and interpreting the data. The machine learning method known as t-stochastic neighbor embedding (tSNE) was employed to generate an interactive map of clustered genes with the distances between genes determined by the similarity of each gene’s BarSeq data.