Integration of In-Silico and Experimental Methods in Chemical Biology for Peptide and Biocatalyst Discovery

Abstract

Genome mining methodologies were used to advance three high impact projects. In project 1, homology-based genome mining was used to discover a new thermostable poly(ethylene terephthalate) (PET)-degrading enzyme. Rational engineering aided by structural modeling was then used to improve enzyme activity. In project 2, a custom genome mining script was used to discover 1094 examples of lasso pep- tides with putative anti-cyanobacterial activity. These lasso peptides are examples of RiPPs—ribosomally synthesized and post-translationally modified peptides—and contain a unique lariat-like knot. In the 1094 anti-cyanobacterial examples, the lasso peptides likely bind a critical metal needed by cyanobacterial metabolism, after which the metal-lasso complex is imported by the lasso-producing organism and linearized via an isopeptidase. The presence of this isopeptidase gene was used to find the anti-cyanobacterial lasso peptides. Project 3 also involved RiPPs; typically, RiPPs use one class defining enzyme to install a modification onto a ribosomally synthesized peptide precursor. To expand the range of chemistries observed in nature, large-scale genome mining and bioinformatic analysis was used to identify the first examples of hybrid RiPPs, which putatively employ more than one class-defining enzyme to create unique structures. The first examples of hybrid RiPPs were heterologously expressed and purified in E. coli. Nomenclature for these new molecules was created after further bioinformatic exploration of several dozen other putative hybrid RiPPs. The successful computational and experimental workflows used to identify and syn- thesize the first hybrid RiPPs can be applied to find more hybrid RiPPs with unique structures and applications in natural product-based pharmaceuticals.