Substrate interactions guide cyclase engineering and lasso peptide diversification

Lasso peptides are a diverse class of naturally occurring, highly stable molecules kinetically trapped in a distinctive rotaxane conformation. How the ATP-dependent lasso cyclase constrains a relatively unstructured substrate peptide into a low entropy product has remained a mystery owing to poor enzyme stability and activity in vitro. Here, we combine substrate tolerance data with structural predictions, bioinformatic analysis, molecular dynamics simulations, and mutational scanning to construct a model for the three-dimensional orientation of the substrate peptide in the lasso cyclase active site. Predicted peptide-cyclase molecular contacts are validated by rationally engineering multiple, phylogenetically diverse lasso cyclases to accept substrates rejected by the wild-type enzymes. Finally, we demonstrate the utility of lasso cyclase engineering by robustly producing previously inaccessible variants that tightly bind to integrin αvβ8, which is a primary activator of transforming growth factor β, and thus an important anticancer target.

Methods
MD simulations were performed using OpenMM 7.6.0 on distributed computing platform Folding@home by Song Yin [[e-mail]] and Prof. Diwakar Shukla [[e-mail]]. Resulting analysis code was created via python based scripts and softwares (Numpy 1.19.2, MDTraj 1.9.9, etc.). For more information about specific data and code, please refer to the paper and Github Repository: https://github.com/ShuklaGroup/lasso_rotaxane/. 

This Dryad dataset has been made available in compliance with Data Availability guidelines by Nature Chemical Biology. For inquiries concerning further data availability, please contact Prof. Diwakar Shukla [[e-mail]] and Prof. Douglas A. Mitchell [[e-mail]].