National Mathematics Initiative, Indian Institute of Science, Bangalore 560012, India.
Department of Biochemistry, Indian Institute of Science, Bangalore 560012, India.
Structure. 2018 Mar 6;26(3):499-512.e2. doi: 10.1016/j.str.2018.02.001.
Protein-ligand interactions form the basis of most cellular events. Identifying ligand binding pockets in proteins will greatly facilitate rationalizing and predicting protein function. Ligand binding sites are unknown for many proteins of known three-dimensional (3D) structure, creating a gap in our understanding of protein structure-function relationships. To bridge this gap, we detect pockets in proteins of known 3D structures, using computational techniques. This augmented pocketome (PocketDB) consists of 249,096 pockets, which is about seven times larger than what is currently known. We deduce possible ligand associations for about 46% of the newly identified pockets. The augmented pocketome, when subjected to clustering based on similarities among pockets, yielded 2,161 site types, which are associated with 1,037 ligand types, together providing fold-site-type-ligand-type associations. The PocketDB resource facilitates a structure-based function annotation, delineation of the structural basis of ligand recognition, and provides functional clues for domains of unknown functions, allosteric proteins, and druggable pockets.
蛋白质-配体相互作用是大多数细胞事件的基础。鉴定蛋白质中的配体结合口袋将极大地促进对蛋白质功能的合理化和预测。许多具有已知三维 (3D) 结构的蛋白质的配体结合位点是未知的，这在我们对蛋白质结构-功能关系的理解中造成了一个空白。为了弥补这一空白，我们使用计算技术在具有已知 3D 结构的蛋白质中检测口袋。这个扩充的口袋组 (PocketDB) 包含 249,096 个口袋，大约是目前已知数量的七倍。我们推断出大约 46%的新识别口袋可能存在配体关联。扩充的口袋组基于口袋之间的相似性进行聚类后，产生了 2,161 种位点类型，这些位点类型与 1,037 种配体类型相关联，共同提供了折叠-位点-类型-配体类型的关联。PocketDB 资源有助于基于结构的功能注释、配体识别的结构基础的描绘，并为未知功能域、别构蛋白和可成药口袋提供功能线索。
