A computational pipeline for predicting distal hotspots in an artificial enzyme

Targeting distal mutations holds promising implications for enzyme engineering. Here, we present an open-source computational workflow designed to explore the functional impact of distal sites, demonstrated on an artificial enzyme built on the widely used Lactococcal multidrug resistance regulator (LmrR) scaffold. By integrating residue network analysis, allosteric pathway mapping and a machine learning-based functional site modeling, we prioritized five distal positions predicted to influence the protein dynamical behavior. After exploring a list of only 20 single point mutations, we identified one variant (Y27H) that significantly enhanced both activity (20 % over the parent) and thermostability (12.5 °C over the parent). Further exploration of double mutants revealed even greater enhancements, with up to a 50 % activity increase and 22.7 °C gain in thermostability. This study highlights the potential of distal mutations in enzyme design and provides a computational strategy for guiding such efforts together with a comprehensive protocol to make it accessible to other users.