#29 · Structural bio & protein design

MD simulations resolve conformational switching in ezrin, σ2R homodimers, and NLRP3

The common thread is simulation-driven resolution of conformational mechanism, with molecular dynamics as the principal tool. The ezrin preprint uses MD to show that PI(4,5)P2 binding to the FERM domain triggers spontaneous CTD dissociation, and that T567 phosphorylation blocks reassociation. The σ2R/TMEM97 study applies adaptive sampling MD with quasi-anharmonic analysis and unsupervised machine learning to reveal asymmetric dynamics between the protomers of the homodimer. The NLRP3 study integrates nudged elastic band calculations, MD, and Markov state model simulations to map ATP-driven activation across five metastable states. What is new is methodological layering: adaptive sampling, kinetic state modeling, and ML-assisted analysis within integrated workflows rather than MD alone. The systems differ, so the shared element is methodological integration rather than a single mechanistic finding.