Characterization of RNA Polymerase II Trigger Loop Mutations using Molecular Dynamics Simulations and Machine Learning

Abstract

Catalysis and fidelity of multisubunit RNA polymerases rely on a highly conserved active site domain called the trigger loop (TL), which achieves roles in transcription through conformational changes and interaction with NTP substrates. The mutations of TL residues cause distinct effects on catalysis including hypo- and hyperactivity and altered fidelity. We applied molecular dynamics simulation (MD) and machine learning (ML) techniques to characterize TL mutations in the Saccharomyces cerevisiae RNA Polymerase II (Pol II) system. We did so to determine relationships between individual mutations and phenotypes and to associate phenotypes with MD simulated structural alterations. Using fitness values of mutants under various stress conditions, we modeled phenotypes along a spectrum of continual values. We found that ML could predict the phenotypes with 0.68 R2 correlation from amino acid sequences alone. It was more difficult to incorporate MD data to improve predictions from machine learning, presumably because MD data is too noisy and possibly incomplete to directly infer functional phenotypes. However, a variational auto-encoder model based on the MD data allowed the clustering of mutants with different phenotypes based on structural details. Overall, we found that lethal mutations tended to increase distances of TL residues to the NTP substrate, while viable loss-of-function (LOF) substitutions tended to confer an increase in distances between TL and bridge helix (BH). In contrast, GOF mutants generally have a disrupting effect on hydrophobic contacts among TL and nearby helices. AUTHOR SUMMARYRNA polymerase II (Pol II) synthesizes RNA with the help of an active site domain called trigger loop (TL). The mutations of TL cause changes in the activity of Pol II that could range from gain-of function (GOF) to loss-of-function (LOF) or lethal. This study provides a systematic characterization of the structural and functional outcomes of the TL mutations using molecular dynamics (MD) simulations and machine learning (ML). We obtained functional phenotypes of mutants by ML using the genetic fitness scores as the input. We revealed that mutant TL sequences could predict the functional outcomes at a relatively high correlation. Then, we performed MD simulations to relate the structural information to the phenotypes. The analysis of the MD data suggested that the lethal mutants had increased distances between the TL and the substrate, while a subset of LOF mutants showed increased distances between TL and another active site domain called bridge helix (BH). On the other hand, GOF mutants had effects on the hydrophobic interactions around the active site. Overall, this study enhances our understanding of the effects of TL mutations to the Pol II function.