The broad objective of this research is to establish the foundation for a novel fully quantum mechanical forcefield for simulations of biocatalysis that can be seamlessly integrated with other multi-scale modeling tools and applied to complex biological problems not accessible by other methods. The design of these new computational tools will greatly extend the scope of biocatalysis applications that can be reliably addressed. The impact of this work with be to create a paradigm shift away from conventional mixed quantum mechanical/molecular mechanical (QM/MM) models toward a united fully quantum mechanical approach for molecular simulations of reactive processes in complex environments. The core methods will be based on a new quantum mechanical model for biocatalysis (Biocat-QM) that combines the advantages of existing semiempirical models and extends their capabilities to accurately model reaction barriers, and charge-dependent many-body exchange, polarization and dispersion effects. The Biocat-QM will form the base of a QM/MM model that contains a new form of the QM/MM interaction where non-bonded terms automatically adjust in response to changes in charge state and hybridization. Ultimately, the Biocat-QM will be made into a novel fully quantum mechanical forcefield for simulations of biocatalysis, based on a new linear-scaling quantum method that utilizes a density-overlap repulsion model to circumvent the need for large local basis projections, and that takes advantage of a recently developed adaptive fast-multipole algorithm for efficient calculation of electrostatic interactions for generalized charge distributions. The new tools for simulations of biocatalysis developed in this proposal are designed to surmount the difficulties presented by specific driving applications: the study of the molecular mechanisms of ribozyme catalysis. The methods will be applied to two ribozyme systems that exhibit large-scale conformational changes and divalent metal ion binding coupled with catalysis, and for which very recent structural data has become available through collaborator Prof. William Scott: the full length hammerhead ribozyme and the L1 ligase ribozyme/riboswitch. These systems present unique challenges for which there currently exists no sufficiently reliable biocatalysis simulation model. The computational tools developed in this proposal will be implemented as publicly available modular software, optimized and ported to several high-performance computing platforms, and integrated with the molecular simulation packages AMBER and CHARMM.
Showing the most recent 10 out of 11 publications
