Linge Jens P, Habeck Michael, Rieping Wolfgang, Nilges Michael
Unité de Bio-Informatique Structurale, Institut Pasteur, 25-28 rue du docteur Roux, F-75015 Paris, France.
J Magn Reson. 2004 Apr;167(2):334-42. doi: 10.1016/j.jmr.2004.01.010.
Indirect magnetization transfer increases the observed nuclear Overhauser enhancement (NOE) between two protons in many cases, leading to an underestimation of target distances. Wider distance bounds are necessary to account for this error. However, this leads to a loss of information and may reduce the quality of the structures generated from the inter-proton distances. Although several methods for spin diffusion correction have been published, they are often not employed to derive distance restraints. This prompted us to write a user-friendly and CPU-efficient method to correct for spin diffusion that is fully integrated in our program ambiguous restraints for iterative assignment (ARIA). ARIA thus allows automated iterative NOE assignment and structure calculation with spin diffusion corrected distances. The method relies on numerical integration of the coupled differential equations which govern relaxation by matrix squaring and sparse matrix techniques. We derive a correction factor for the distance restraints from calculated NOE volumes and inter-proton distances. To evaluate the impact of our spin diffusion correction, we tested the new calibration process extensively with data from the Pleckstrin homology (PH) domain of Mus musculus beta-spectrin. By comparing structures refined with and without spin diffusion correction, we show that spin diffusion corrected distance restraints give rise to structures of higher quality (notably fewer NOE violations and a more regular Ramachandran map). Furthermore, spin diffusion correction permits the use of tighter error bounds which improves the distinction between signal and noise in an automated NOE assignment scheme.
在许多情况下，间接磁化转移会增加两个质子之间观测到的核欧沃豪斯效应（NOE），从而导致对目标距离的低估。需要更宽的距离界限来考虑这种误差。然而，这会导致信息丢失，并可能降低由质子间距离生成的结构的质量。尽管已经发表了几种自旋扩散校正方法，但它们通常未被用于推导距离约束。这促使我们编写一种用户友好且计算效率高的方法来校正自旋扩散，该方法已完全集成到我们的程序“迭代分配的模糊约束”（ARIA）中。因此，ARIA允许使用经自旋扩散校正的距离进行自动迭代NOE分配和结构计算。该方法依赖于通过矩阵平方和稀疏矩阵技术对控制弛豫的耦合微分方程进行数值积分。我们从计算得到的NOE体积和质子间距离中推导距离约束的校正因子。为了评估我们的自旋扩散校正的影响，我们用小家鼠β - 血影蛋白的普列克底物蛋白同源（PH）结构域的数据广泛测试了新的校准过程。通过比较经自旋扩散校正和未经校正的结构精修结果，我们表明经自旋扩散校正的距离约束能产生更高质量的结构（特别是更少的NOE违反和更规则的拉马钱德兰图）。此外，自旋扩散校正允许使用更严格的误差界限，这在自动NOE分配方案中改善了信号与噪声之间的区分。
