Computer Simulations of Enzymes

Project Summary Our first research theme is the accurate and efficient description of bio-molecular interactions, which is at the core of biological modeling, particularly for the understanding of protein structures and the reactions catalyzed by enzymes. Achieving optimal results requires an energy function that balances accuracy with computational efficiency, in both quantum mechanical (QM) methods and the molecular mechanical (MM) force fields. The second theme is to study and reveal the complex reaction mechanisms of key enzymes. Enzymes are biological catalysts that drive biochemical reactions essential for life. Studying reaction mechanisms in enzymes is crucial for obtaining insights into fundamental biological processes, developing inhibitor and drug, aiding in the development of diagnostic tests and prognostic tools for diseases, optimizing biomedical technology, and inferring evolutionary relationships. This proposed MIRA research project focuses on three directions. (A) Developing QM methods based on density functional theory (DFT) to achieve accuracy and efficiency needed for biological systems. DFT is the most widely used QM for biological molecules. Despite its success, DFT can still suffer from large pervasive systematic errors. We will develop the localized orbitals scaling corrections to eliminate these errors. (B) Constructing protein and biomolecular force fields based on multiscale QM and MM and machine learning (ML). Although MM force fields have been successfully applied to many problems, they still frequently fail to capture the correct conformational and energetic features of diverse biomolecules. We will build on the residue-based systematic molecular fragmentation (rSMF) approach developed in the PI lab to develop a general and scalable ML force field for all biomolecules based on QM/MM and rSMF, providing a unified model for not only proteins, but also protein-ligands, RNA-ligands, and other complexes. (C) Investigating reaction mechanisms of 5’-deoxyadenosyl radical formation in radical SAM enzymes. Radical S-adenosylmethionine (SAM) enzymes form one of the largest enzyme superfamilies with more than 700,000 unique sequences. They are abundant in Nature, involving in many medically important pathways. We will characterize the mechanism of SAM cleavage using MoaA radical SAM enzyme as a model system, in collaboration with experimental efforts. The plan research, harnessing the latest QM, QM/MM and ML developments, will advance simulation methods for biological systems and provide valuable insights into enzyme functions.