Application of replica exchange umbrella sampling to protein structure refinement of nontemplate models

Abstract

We provide an assessment of a computational strategy for protein structure refinement that combines self-guided Langevin dynamics with umbrella-potential biasing replica exchange using the radius of gyration as a coordinate (Rg -ReX). Eight structurally nonredundant proteins and their decoys were examined by sampling conformational space at room temperature using the CHARMM22/GBMV2 force field to generate the ensemble of structures. Two atomic statistical potentials (RWplus and DFIRE) were analyzed for structure identification and compared to the simulation force-field potential. The results show that, while the Rg -ReX simulations were able to sample conformational basins that were more structurally similar to the X-ray crystallographic structures than the starting first-order ranked decoys, the potentials failed to detect these basins from refinement. Of the three potential functions, RWplus yielded the highest accuracy for recognition of structures that refined to an average of nearly 20% increase in native contacts relative to the starting decoys. The overall performance of Rg -ReX is compared to an earlier study of applying temperature-based replica exchange to refine the same decoy sets and highlights the general challenge of achieving consistently the sampling and detection threshold of 70% fraction of native contacts.

Evaluation of Unrestrained Replica-Exchange Simulations Using Dynamic Walkers in Temperature Space for Protein Structure Refinement

“…The average maximum f N sampled from the simulations of all targets is 0.74 for the adaptive T-ReX and 0.72 for the static method. While these sampling excursions seem to approach the downhill refinement regime on the force-field potential energy landscape for both methods [7], [22], the conformational populations of these basins and their detection from the energy functions were disappointedly poor.…”

“…4c and d). The GOAP energy landscape for the static method shows a downhill refinement funnel that likely represents the all-atom resolution of the force field in sampling conformational space [7], [22], whereas the adaptive T-ReX reveals a landscape that favors a non-native collapsed basin. For the adaptive method, GOAP identified a conformer of f N 0.52 (RMSD 5.6 Å), while the static T-ReX method detected f N 0.71 (RMSD 1.7 Å).…”

“…SGLD parameters of the friction constant was set to γ of 1 ps −1 for all heavy atoms, the guiding factor λ set to a value of 1, and the averaging time was set to 1 ps. Selection of these values was taken from our previous studies of the SGLD model [7], [22]. Non-bonded interaction cutoff parameters for electrostatics and vdW terms were set at a radius of 22 Å with a 2-Å potential switching function.…”

“…While there are theoretical virtues and limitations of the adaptive T-ReX method, we investigate in this work the practical performance of the method applied to the structure refinement of a dataset of protein conformational decoys [7], [22]. Given the success of the adaptive T-ReX simulations for modeling the sharp energy differences between the folded and unfolded states of SH3 [18], it is of general interest to determine whether this approach is beneficial in modeling less-cooperative transitions that are thought to govern the structure refinement of decoys taken from either comparative modeling or knowledge-based structure predictions.…”

“…Given the success of the adaptive T-ReX simulations for modeling the sharp energy differences between the folded and unfolded states of SH3 [18], it is of general interest to determine whether this approach is beneficial in modeling less-cooperative transitions that are thought to govern the structure refinement of decoys taken from either comparative modeling or knowledge-based structure predictions. We examine a set of protein targets that offers a challenging benchmark of simulation methods for structure refinement [7], [22]. The targets are 49–92 residues in size and amenable to an analysis by unrestrained all-atom simulation methods.…”