BACKGROUND EPR-based distance measurements between spin labels in proteins have become a valuable tool in structural biology. The direct translation of the experimental distances into structural information is however often impaired by the intrinsic flexibility of the spin labelled side chains. Different algorithms exist that predict the approximate conformation of the spin label either by using pre-computed rotamer libraries of the labelled side chain (rotamer approach) or by simply determining its accessible volume (accessible volume approach). Surprisingly, comparisons with many experimental distances have shown that both approaches deliver the same distance prediction accuracy of about 3 Å. RESULTS Here, instead of comparing predicted and experimental distances, we test the ability of both approaches to predict the actual conformations of spin labels found in a new high-resolution crystal structure of spin labelled azurin (T21R1). Inside the crystal, the label is found in two very different environments which serve as a challenging test for the in silico approaches. CONCLUSIONS Our results illustrate why simple and more sophisticated programs lead to the same prediciton error. Thus, a more precise treatment of the complete environment of the label and also its interactions with the environment will be needed to increase the accuracy of in silico spin labelling algorithms.

中文翻译：

---

来自铜绿假单胞菌的自旋标记 (T21R1) 天青蛋白的高分辨率晶体结构：计算机自旋标记算法的具有挑战性的结构基准。

背景 蛋白质中自旋标记之间基于 EPR 的距离测量已成为结构生物学中的一种有价值的工具。然而，将实验距离直接转换为结构信息通常会受到自旋标记侧链的固有灵活性的影响。存在不同的算法，通过使用标记侧链的预计算旋转异构体库（旋转异构体方法）或通过简单地确定其可访问体积（可访问体积方法）来预测自旋标签的近似构象。令人惊讶的是，与许多实验距离的比较表明，两种方法都提供了大约 3 Å 的相同距离预测精度。结果 在这里，不是比较预测距离和实验距离，我们测试了两种方法预测自旋标记天青蛋白 (T21R1) 的新高分辨率晶体结构中发现的自旋标记实际构象的能力。在晶体内部，标签是在两种截然不同的环境中发现的，这是对计算机方法的一项具有挑战性的测试。结论 我们的结果说明了为什么简单和更复杂的程序会导致相同的预测错误。因此，需要对标签的完整环境及其与环境的相互作用进行更精确的处理，以提高计算机自旋标记算法的准确性。结论 我们的结果说明了为什么简单和更复杂的程序会导致相同的预测错误。因此，需要对标签的完整环境及其与环境的相互作用进行更精确的处理，以提高计算机自旋标记算法的准确性。结论 我们的结果说明了为什么简单和更复杂的程序会导致相同的预测错误。因此，需要对标签的完整环境及其与环境的相互作用进行更精确的处理，以提高计算机自旋标记算法的准确性。