Department of Chemistry, University of Zurich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland.
Department of Biochemistry, University of Zurich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland.
J Mol Biol. 2015 May 22;427(10):1916-33. doi: 10.1016/j.jmb.2015.02.022. Epub 2015 Mar 24.
The specific recognition of peptide sequences by proteins plays an important role both in biology and in diagnostic applications. Here we characterize the relatively weak binding of the peptide neurotensin (NT) to the previously developed Armadillo repeat protein VG_328 by a multidisciplinary approach based on solution NMR spectroscopy, mutational studies, and molecular dynamics (MD) simulations, totaling 20μs for all MD runs. We describe assignment challenges arising from the repetitive nature of the protein sequence, and we present novel approaches to address them. Partial assignments obtained for VG_328 in combination with chemical shift perturbations allowed us to identify the repeats not involved in binding. Their subsequent elimination resulted in a reduced-size binder with very similar affinity for NT, for which near-complete backbone assignments were achieved. A binding mode suggested by automatic docking and further validated by explicit solvent MD simulations is consistent with paramagnetic relaxation enhancement data collected using spin-labeled NT. Favorable intermolecular interactions are observed in the MD simulations for the residues that were previously shown to contribute to binding in an Ala scan of NT. We further characterized the role of residues within the N-cap for protein stability and peptide binding. Our multidisciplinary approach demonstrates that an initial low-resolution picture for a low-micromolar-peptide binder can be refined through the combination of NMR, protein design, docking, and MD simulations to establish its binding mode, even in the absence of crystallographic data, thereby providing valuable information for further design.
蛋白质对肽序列的特异性识别在生物学和诊断应用中都起着重要作用。在这里，我们通过基于溶液 NMR 光谱学、突变研究和分子动力学 (MD) 模拟的多学科方法，对先前开发的 Armadillo 重复蛋白 VG_328 与肽神经降压素 (NT) 的相对较弱的结合进行了表征，所有 MD 运行的总时间为 20μs。我们描述了由于蛋白质序列的重复性而产生的分配挑战，并提出了新的方法来解决这些问题。VG_328 的部分分配与化学位移扰动相结合，使我们能够识别不参与结合的重复序列。随后消除这些重复序列导致与 NT 具有非常相似亲和力的结合物尺寸减小，对于该结合物，实现了几乎完整的骨架分配。自动对接建议的结合模式，并通过显式溶剂 MD 模拟进一步验证，与使用自旋标记 NT 收集的顺磁弛豫增强数据一致。在 MD 模拟中观察到有利的分子间相互作用，这些相互作用对于先前在 NT 的 Ala 扫描中显示出对结合有贡献的残基。我们进一步表征了 N-帽内残基在蛋白质稳定性和肽结合中的作用。我们的多学科方法表明，即使没有晶体学数据，也可以通过 NMR、蛋白质设计、对接和 MD 模拟的组合，对低分辨率的低微摩尔肽结合物进行初始图像进行细化，以确定其结合模式，从而为进一步设计提供有价值的信息。
