蛋白质-DNA复合物中残基界面偏好性分析及在识别界面中的应用

Analysis of Residue Interface Preference in Protein-DNA Complexes and Its Application in Recognition of Binding Interface

摘要: 蛋白质-DNA识别在生物过程中起着重要作用，其结合是由序列特异性识别和结构特征共同影响的。为了研究残基类型和蛋白质二级结构对结合的贡献，本文构建了一个新的非冗余蛋白质-DNA复合物数据库，其中包含1545个结构。经过统计分析发现，残基和二级结构类型对蛋白质与DNA结合有很大贡献，二级结构中π-helix和β-ladder是最偏好界面的类型。对蛋白质二级结构按界面偏好进行分类，构建了60 × 4氨基酸–核苷酸成对界面偏好性。从该偏好性中获得氨基酸界面偏好性，并探讨了将该信息用于预测蛋白质-DNA结合界面的可能性，研究对象为对接基准数据集中的47个复合物体系。结果发现成对界面偏好性信息可以将87.23%的体系的真实界面打分排在所有表面区域的前10%。这说明本文构建的60 × 4氨基酸–核苷酸成对界面偏好性很好地反映了蛋白质-DNA的界面识别，对界面和复合物结构预测具有重要意义。

Abstract: Protein-DNA recognition plays an important role in biological processes, and its binding is influ-enced by sequence specific recognition and structural characteristics. To investigate the contribu-tion of residue types and protein secondary structure elements to binding, a new non-redundant protein-DNA database with 1545 complex structures was constructed. Statistical analysis reveals that protein residue and secondary structure types have significant contributions to its binding with DNA. Among the secondary structures, π-helix and β-ladder have the highest preferences. We classified the protein secondary structures according to their interface preferences, and construct-ed the 60 × 4 amino acid-nucleotide pairwise interface preferences. The amino acid interface pref-erences obtained from the pairwise ones were used to explore the possibility of predicting pro-tein-DNA binding interfaces for 47 complex systems from the docking benchmark dataset. The re-sult shows that the pairwise interface preferences can rank the real interfaces in the top 10% of all surface patches for 87.23% of all cases. These results indicate that the 60 × 4 amino acid-nucleotide pairwise interface preferences constructed by us can well reflect protein-DNA recognition, which is of great significance for interface and complex structure predictions.