Abstract:
Mixed coarse-graining approach has been recently introduced to the elastic network model to enable the modeling of a protein’s native conformation with regions of high- and low-resolution. In this model, each node of the elastic network may represent a heavy atom (high-resolution), a residue or a group of residues (low-resolution), and close-neighboring nodes are connected by springs. To assign suitable cutoffs and force constants to the nodes in different resolution regions, two alternative procedures which either take the residue-based or atom-based elastic network model parameters as reference are developed. The calculation of parameters with the atom-based approach has proved to be superior due to its straightforward and realistic description of interactions and its applicability to both proteins and their complexes with RNA/DNA. The mixed coarse-graining method is validated by exploring the internal dynamics of the enzyme triosephosphate isomerase with 494 residues. The role of the dimeric enzyme’s collective motions in controlling loop 6 closure and hence the catalytic activity is revealed for the first time. The supramolecular assemblage ribosome in its complex with mRNA, three tRNAs and elongation factor Tu (~11000 amino acids and nucleotides in total) is studied with residue-based and mixed coarse-grained models. As a result, large domain motions with functional importance are clearly observed, a translocation mechanism for mRNA and tRNAs is proposed, and the dynamics of the ribosomal tunnel are investigated. The decoding center with codons and anticodons in ribosome structure is modeled at atomistic level to investigate the local vibrational dynamics, difficult to attain with classical atomistic techniques. The results indicate that retaining the whole structure is critical to describe the collective dynamics of specific components in a large multi-subunit protein. The mixed resolution elastic network models have proven to be a powerful tool to study the dynamics of extremely large supramolecular assemblages at the atomistic scale with high computational efficiency.
