利用X-ray晶體結構最佳化所取得的參數來計算蛋白質內原子間的動態關聯性

Abstract

蛋白質的動態特性可以提供研究人員有用的資訊去了解遠距離的突變位點對於酵素活性的影響，以及蛋白質複合體的協同調控機制。從蛋白質晶體繞射實驗中取得的溫度系數(B-factors) 以及蛋白質核磁共振實驗中得到的有序度參數(order parameters)可提供便捷的方式去評估蛋白質中的原子擾動(atomic fluctuations 或 atomic auto-correlated motions)。然而，蛋白質中原子間動態相互關聯性(atomic cross-correlated motions)並沒有辦法直接由實驗中取得，一般會尋求諸如：動態模擬(molecular dynamics)、簡正模分析法(normal mode analysis)或彈性網絡模型(elastic network models)等複雜的計算方法。在本論文中，我們提出了可以直接利用蛋白質晶體結構最佳化時的資料來合理計算蛋白質中原子間動態相互關聯性的方法。那就是利用最佳化溫度系數時所得到的TLS參數來計算原子間動態的相互關聯性的分析模型。利用一組由100個非同源蛋白結構所集合而成的資料去測試模型，發現用我們的模型去計算所得到的原子動態相關性圖譜(correlation maps)與用機械性模型(mechanical models)所得到的圖譜，平均相關性系數達0.75。總體而論，我們發展了一套可藉由蛋白質結構最佳化時所取得的資料直接計算原子間動態相互關聯性的方法，因為此公式為解析型式，因此可提供快速及高通量的計算。除此之外，由於此方法無需複雜的計算，所以可提供非計算背景的生物研究人員快速可靠的方式，去取得蛋白質原子間動態關聯性來協助進行蛋白質功能的研究。

The dynamic characteristics of proteins can provide researchers with useful information for understanding the long-distance effects of mutated sites on enzyme activity and the co-regulation mechanisms of protein complexes. X-ray B-factors and NMR order parameters can be obtained experimentally, to serve as a convenient way for assessing atomic fluctuations (or atomic auto-correlated motions) in proteins. However, the atomic cross-correlated motions in proteins cannot be obtained directly from experiments. Instead, the following complicated calculation methods are generally used: molecular dynamics, normal mode analysis, or atomic network models. In this thesis, we show that atomic cross-correlations can be reliably obtained directly from protein structure based on X-ray refinement data. We have derived an analytic form of atomic correlated motions in terms of the original translation/libration/screw (TLS) parameters used in the refinement of B-factors for X-ray structures. We used a dataset comprising 100 non-homologous protein structures to test this equation. Our analysis indicated that the correlation maps calculated based on this equation are well correlated with those based on mechanical models (the correlation coefficients is 0.75) for the dataset. In summary, we have developed an approach to compute atomic cross-correlations directly from protein structures by using X-ray refinement data. Because this formula is an analytic form, it can be calculated rapidly and with high throughput to compute motional correlations. In addition, avoiding sophisticated computational operations; it provides a quick, reliable way, especially for biological researchers without mathematical backgrounds to obtain dynamics information directly from protein structure relevant to its function.