Barrett, S. E. | Search bioRxiv

bioRxiv2026

De novo Folding Mechanisms of Lasso Peptides

Lasso peptides adopt a distinctive [1]rotaxane conformation, yet the principles governing the folding of this kinetically trapped structure have remained elusive. Here, we integrated extensive molecular dynamics simulations and deep learning to elucidate the de novo folding mechanism of 20 lasso peptides lacking secondary post-translational modifications. We constructed Multi-Ensemble Markov Models for each lasso peptide and uncovered a universal uphill folding landscape with spontaneous folding probabilities consistently below 0.8%. Loop stability strongly correlated with folding propensity, and targeted experiments further validated that enhancing loop {beta}-hairpin formation promotes folding of microcin J25, the well-studied lasso peptide extensively characterized as an in vitro model. Additionally, the substantial entropy cost opposed lasso peptide folding. Simulations mimicking enzymatic spatial confinement reduced this penalty and stabilize folding. Leveraging Variational AutoEncoder-based pathway clustering, we resolved distinct pathway channels and representative folding pathways. Together, these findings establish representative folding models and fundamental thermodynamic and kinetic principles for rational engineering of lasso peptides.

biophysics↗

bioRxiv2024

LassoESM: A tailored language model for enhanced lasso peptide property prediction

Ribosomally synthesized and post-translationally modified peptides (RiPPs) comprise a structurally and functionally diverse group of natural products. Lasso peptides represent one of about 50 known molecular classes of RiPPs, which display a characteristic [1] rotaxane conformation formed by a lasso cyclase. This unique, threaded conformation endows lasso peptides with diverse biological activities and remarkable thermal and proteolytic stability. The prediction of lasso peptide properties, such as substrate compatibility with a particular lasso cyclase or desired biological activity, remains challenging due to limited experimental data and the intricate nature of the substrate fitness landscapes. Protein language models (PLMs) have demonstrated impressive performance in predicting protein structure and function. However, general-purpose PLMs perform poorly in lasso peptide-related predictive tasks. Therefore, there is a need to provide effective representations for lasso peptides to enable enhanced property prediction. In this study, we developed a lasso peptide-specific language model (LassoESM) by leveraging advances in pre-trained PLMs to aid the prediction of lasso peptide related properties and experimentally validate the model predictions. We demonstrate that LassoESM embeddings can accurately predict substrate compatibility for a lasso cyclase of interest when experimental data for model training was scarce. Using a deep learning framework incorporating cross-attention between lasso cyclase and substrate peptide embeddings, we identify non-cognate pairs of lasso cyclases and substrate peptides with predicted compatibility. We further show that LassoESM embeddings improve the prediction of RNA polymerase inhibitory activity, which represents a biological activity of several known lasso peptides. We anticipate that LassoESM and future iterations thereof will be instrumental for the rational design of lasso peptides with desired properties.

biophysics↗