化学
二硫键
共价键
肽
立体化学
环肽
氢键
分子动力学
肽键
胰蛋白酶
扭转
蛋白质结构
拟肽
分子
蛋白质折叠
非共价相互作用
螺旋(腹足类)
肽构象
连接器
组合化学
折叠(DSP实现)
二面角
测试表
构象集合
计算化学
作者
Myongin Oh,Changin Oh,Tama Ghosh,Eshra Tabassum
标识
DOI:10.1021/acs.jpcb.6c01967
摘要
Backbone cyclization and disulfide bonding are key global covalent constraints in many bioactive peptides, yet their individual and combined effects on peptide conformational ensembles remain incompletely understood. Here, we combine all-atom molecular dynamics (MD) simulations, geometric analysis, and interpretable machine learning to examine these effects in sunflower trypsin inhibitor-1 (SFTI-1) as a minimal model system. Comparison of native SFTI-1 with a cyclic analogue lacking the disulfide bridge, an acyclic analogue lacking backbone cyclization, and a linear analogue lacking both shows that these constraints do not merely rigidify the peptide uniformly, but instead selectively reorganize its conformational ensemble across multiple structural levels. Backbone cyclization favors overall shape isotropy and determines the dominant global twist preference and the multimodal character of the twist distribution. The disulfide bridge plays a central role in maintaining the canonical Arg2-Phe12 and Thr4-Ile10 backbone hydrogen bonds. Coupling between local and global constraints is site-specific: the cis preference of the Ile7-Pro8 peptide bond is observed when either global constraint is present. Finally, by integrating unbiased MD with the Covalent protocol for interpretable collective variable discovery for enhanced sampling, we suggest that prefolding of the linear analogue facilitates compaction but is insufficient to generate the thiol-thiol proximity and orientation required for disulfide formation.
科研通智能强力驱动
Strongly Powered by AbleSci AI