Ryabov Yaroslav E, Geraghty Charles, Varshney Amitabh, Fushman David
Department of Chemistry and Biochemistry, Center for Biomolecular Structure and Organization, University of Maryland, 1115 Biomolecular Sciences Building, College Park, Maryland 20742, USA.
J Am Chem Soc. 2006 Dec 6;128(48):15432-44. doi: 10.1021/ja062715t.
We propose a new computational method for predicting rotational diffusion properties of proteins in solution. The method is based on the idea of representing protein surface as an ellipsoid shell. In contrast to other existing approaches this method uses principal component analysis of protein surface coordinates, which results in a substantial increase in the computational efficiency of the method. Direct comparison with the experimental data as well as with the recent computational approach (Garcia de la Torre; et al. J. Magn. Reson. 2000, B147, 138-146), based on representation of protein surface as a set of small spherical friction elements, shows that the method proposed here reproduces experimental data with at least the same level of accuracy and precision as the other approach, while being approximately 500 times faster. Using the new method we investigated the effect of hydration layer and protein surface topography on the rotational diffusion properties of a protein. We found that a hydration layer constructed of approximately one monolayer of water molecules smoothens the protein surface and effectively doubles the overall tumbling time. We also calculated the rotational diffusion tensors for a set of 841 protein structures representing the known protein folds. Our analysis suggests that an anisotropic rotational diffusion model is generally required for NMR relaxation data analysis in single-domain proteins, and that the axially symmetric model could be sufficient for these purposes in approximately half of the proteins.
我们提出了一种预测蛋白质在溶液中旋转扩散性质的新计算方法。该方法基于将蛋白质表面表示为椭球壳的思想。与其他现有方法不同，此方法使用蛋白质表面坐标的主成分分析，这使得该方法的计算效率大幅提高。与实验数据以及最近基于将蛋白质表面表示为一组小球形摩擦元件的计算方法（Garcia de la Torre等人，《磁共振杂志》，2000年，B147，138 - 146）直接比较表明，这里提出的方法能够以至少与另一种方法相同的准确度和精度重现实验数据，同时速度快约500倍。使用新方法，我们研究了水化层和蛋白质表面形貌对蛋白质旋转扩散性质的影响。我们发现，由大约一个水分子单层构成的水化层使蛋白质表面变得平滑，并有效地使整体翻滚时间加倍。我们还计算了代表已知蛋白质折叠的841个蛋白质结构的旋转扩散张量。我们的分析表明，单结构域蛋白质的核磁共振弛豫数据分析通常需要各向异性旋转扩散模型，并且轴对称模型在大约一半的蛋白质中可能足以满足这些目的。
