Zhiqiang Yan*, Yuqing Li, Ying Cao, Xuetao Tao, Jin Wang and Yongsheng Jiang Pages 1 - 11 ( 11 )
Evolution has optimized proteins to balance stability and function by reducing unfavorable energy states, leading to regions of flexibility and frustration on protein surfaces. These locally frustrated regions correspond to functionally important areas, such as active sites and regions for ligand binding and conformational plasticity. Typical strategies of structure-based drug discovery primarily concentrate on enhancing the binding affinity during compound screening and target identification. However, this often overlooks the binding specificity, which is critical for distinguishing specific binding partners from competing ones and avoiding off-target effects. According to the energy landscape theory, optimization of the intrinsic binding specificity involves globally minimizing the frustrations existing in the biomolecular interactions. Recent studies have demonstrated that identifying local frustrations provides a promising approach for screening more specific compounds binding with targets, and quantifying binding specificity complements typical strategies that focus on binding affinity only. This review explores the principles and strategies of computationally quantifying the binding specificity and local frustrations and discusses their applications in structure-based drug discovery. Moreover, given the advancements of artificial intelligence in protein science, this review aims to motivate the integration of AI and available approaches in quantifying the binding specificity and local frustration. We expect that an AI-powered prediction model will accelerate the drug discovery process and improve the success rate of hit compounds.
Energy landscape, biomolecular recognition, binding specificity, local frustration, drug screening, binding site, cryptic site.