Stability tests on known and misfolded structures with discrete and all atom molecular dynamics simulations

Area of Science:

• Computational Biology
• Biophysics
• Structural Biology

Background:

• Molecular dynamics (MD) simulations are crucial for protein modeling but are often limited by simulation length.
• Discrete molecular dynamics (DMD) offers significantly longer simulation times by employing approximations like coarse-graining and implicit solvent.
• These approximations in DMD can compromise the accuracy of protein structural dynamics and model evaluation.

Purpose of the Study:

• To enhance the accuracy and utility of discrete molecular dynamics (DMD) for protein structure modeling and evaluation.
• To identify optimal simulation parameters for DMD, including a 'physiological' temperature.
• To compare the performance of DMD with conventional all-atom molecular dynamics (MD) in distinguishing correct from incorrect protein structures.

Main Methods:

• Performed comparative simulations using both DMD and all-atom MD on known and misfolded protein structures.
• Investigated the effect of temperature on DMD simulations to find conditions mimicking conventional MD at 310K.
• Utilized two different implicit solvent models in DMD: one based on Miyazawa-Jernigan potentials and another on Kyte-Doolittle hydropathy.
• Simulated monomeric and dimeric α β-barrel structures, both correctly and incorrectly folded.

Main Results:

• DMD simulations showed greater deviations from correct protein structures compared to all-atom MD, particularly in coiled loops.
• Incorrectly folded protein structures were poorly preserved in DMD simulations.
• Despite deviations, both DMD and all-atom MD methods could differentiate between correct and incorrect protein models based on root mean squared deviation (RMSD) magnitudes.
• A new implicit solvent model based on Miyazawa-Jernigan potentials was tested.

Conclusions:

• Discrete molecular dynamics (DMD) can be optimized to improve its accuracy for protein modeling.
• While DMD introduces greater structural deviations than all-atom MD, it remains capable of distinguishing valid from invalid protein structural models.
• Further development of implicit solvent models and parameter optimization can enhance DMD's utility in structural biology research.