Independent Metrics for Protein Backbone and Side-Chain Flexibility: Time Scales and Effects of Ligand Binding

Abstract

Conformational dynamics are central for understanding biomolecular structure and function, since biological macromolecules are inherently flexible at room temperature and in solution. Computational methods are nowadays capable of providing valuable information on the conformational ensembles of biomolecules. However, analysis tools and intuitive metrics that capture dynamic information from in silico generated structural ensembles are limited. In standard work-flows, flexibility in a conformational ensemble is represented through residue-wise root-mean-square fluctuations or B-factors following a global alignment. Consequently, these approaches relying on global alignments discard valuable information on local dynamics. Results inherently depend on global flexibility, residue size, and connectivity. In this study we present a novel approach for capturing positional fluctuations based on multiple local alignments instead of one single global alignment. The method captures local dynamics within a structural ensemble independent of residue type by splitting individual local and global degrees of freedom of protein backbone and side-chains. Dependence on residue type and size in the side-chains is removed via normalization with the B-factors of the isolated residue. As a test case, we demonstrate its application to a molecular dynamics simulation of bovine pancreatic trypsin inhibitor (BPTI) on the millisecond time scale. This allows for illustrating different time scales of backbone and side-chain flexibility. Additionally, we demonstrate the effects of ligand binding on side-chain flexibility of three serine proteases. We expect our new methodology for quantifying local flexibility to be helpful in unraveling local changes in biomolecular dynamics.

Large-scale, quantitative protein assays on a high-throughput DNA sequencing chip

“…3G). We investigated this possibility by examining the relationship between the observed enzymatic activity and amino acid backbone flexibility [34] of amino acid substitutions at position 153. We observed a strong relationship (R 2 =0.43; p-value=1.6x10 -3 ; Fig.…”

Molecular dynamics simulation of local structural models of PrPScreveals how codon 129 polymorphism affects propagation of PrPSc

“…Besides L125-L125 interactions, other hydrophobic interactions in the hydrophobic core of the β-arch were also significantly affected, e.g., V121-M129 ( Fig 1E ). The heat map of local B-factors of the G127V amyloid ( Fig 1D, local B ), which is a metric for local flexibility of backbone [32], showed more bright spots in the region of the U-shaped β-arch(120-133) than that of WT, further confirming the structural disorder. Interestingly, the hydrophobic interaction network outside the U-shaped β-arch was similar to that of WT ( Suppl Fig S2B, G127V ).…”

Structure‐Based Target‐Specific Screening Leads to Small‐Molecule CaMKII Inhibitors

“…Per‐residue flexibility : Per‐residue flexibility was calculated from the set of 1000 snapshots for the apo‐CaMKIIδ protein and three docked poses of compound 1 : the ( R )‐ 1 pose from the initial AutoDock Vina‐docked virtual screen and the ( S ) and ( R) configuration poses of 1 from the follow up binding mode analysis using Glide. Per‐residue flexibility is adapted using the scheme from Fuchs et al . Per‐residue B ‐factor is calculated using the atomicfluc function in cpptraj .…”

“…Ad istance matrix between the vectors was calculated using Euclidean distance and used to hierarchically cluster the kinases using average linkage in R. [90] Per-residue flexibility:P er-residue flexibility was calculated from the set of 1000 snapshots for the apo-CaMKIId protein and three docked poses of compound 1:t he (R)-1 pose from the initial Auto-Dock Vina-docked virtual screen and the (S)a nd (R) configuration poses of 1 from the follow up binding mode analysis using Glide. Per-residue flexibility is adapted using the scheme from Fuchs et al [91] Per-residue B-factor is calculated using the atomicfluc function in cpptraj. The per-residue B-factors of the apo-structure are used to normalize the B-factors of the bound complexes.…”