Proteins depend on defined molecular plasticity for their functionality. How to comprehensively capture dynamics correctly is of ubiquitous biological importance. Approaches commonly used to probe protein dynamics include model-free elucidation of site-specific motion by NMR relaxation, molecular dynamics (MD)-based approaches, and capturing the substates within a dynamic ensemble by recent eNOE-based multiple-structure approaches. Even though MD is sometimes combined with ensemble-averaged NMR restraints, these approaches have largely been developed and used individually. Owing to the different underlying concepts and practical requirements, it has remained unclear how they compare, and how they cross-validate and complement each other. Here, we extract and compare the differential information contents of MD simulations, NMR relaxation measurements, and eNOE-based multi-state structures for the SH3 domain of chicken α-spectrin. The data show that a validated, consistent, and detailed picture is feasible both for timescales and actual conformational states sampled in the dynamic ensemble. This includes the biologically important side-chain plasticity, for which experimentally cross-validated assessment is a significant challenge.

蛋白质的功能取决于定义的分子可塑性。如何全面正确地捕捉动力学具有普遍的生物学重要性。通常用于探测蛋白质动力学的方法包括通过 NMR 弛豫对位点特定运动的无模型阐明、基于分子动力学 (MD) 的方法，以及通过最近的基于 eNOE 的多结构方法捕获动态集合内的亚状态。尽管 MD 有时与整体平均 NMR 限制相结合，但这些方法主要是单独开发和使用的。由于不同的基本概念和实际要求，目前尚不清楚它们如何比较，以及它们如何交叉验证和互补。在这里，我们提取并比较 MD 模拟的差异信息内容，核磁共振弛豫测量，以及鸡 α-血影蛋白 SH3 域的基于 eNOE 的多态结构。数据表明，经过验证、一致和详细的图片对于动态集合中采样的时间尺度和实际构象状态都是可行的。这包括生物学上重要的侧链可塑性，对此，实验交叉验证评估是一个重大挑战。