基于序列和结构特征的人类蛋白酶底物裂解的生物信息学研究

The ability of proteases to catalytically hydrolyse protein substrates with the removal of damaged or undesirable products is fundamental for all forms of life. The importance of protease-controlled proteolysis is apparent in numerous human pathological conditions related to alterations in proteases, including cancer, arthritis, inflammation, degenerative and cardiovascular diseases. Proteases act as processing enzymes that carry out either highly or moderately selective cleavage of the scissile bond after the specific cleavage site in their substrate. The key bottleneck to understanding the physiological role of a protease is to identify its natural substrate(s). Knowledge of the substrate specificity of a protease can dramatically improve our ability to predict target protein substrates and develop specific inhibitors; however, this information can at present only be derived from expensive and time-consuming experimental approaches. However, a major problem in the field is accurate prediction of the target specificity of proteases is currently not possible. In this project, we aim to develop bioinformatic approaches to make testable predictions in regards to the substrate specificity and biological targets of several cysteine and serine proteases, by focusing on the human cysteine and serine protease families, including human calpain, caspase-3, caspase-6, granzyme B and thrombin. We will characterize the sequence and structural determinants of substrate cleavage sites. We will ascertain whether structural information of protease/inhibitor/substrate complexes can help guide our ability to predict the preference of a protease for binding a particular amino acid at each of its subsites. We will use the developed approaches to perform in silico screening in human proteomes and experimentally verify the predicted novel substrates to shed light on functional regulation of the proteases. This study will provide novel basic understanding of the substrate specificity of proteases and provide novel insights into the effective screening of novel substrates and rational design of inhibitors.

由蛋白酶所调控的底物水解对于所有生命形式具有重要的基础意义。失控的蛋白酶底物水解则跟许多人类疾病密切相关，包括癌症、炎症以及心血管疾病等。蛋白酶对于所作用的反应底物有着严格的选择性，仅能特异性水解底物中特定位置的肽键。蛋白酶功能研究的核心瓶颈是如何准确识别所有的天然底物。掌握蛋白酶底物特异性知识能够极大提高我们预测酶底物以及设计抑制剂的能力，然而，目前这一信息获取只能依赖于实验手段。因而，本研究领域的一个核心问题是对底物特异性的准确预测能力相当有限。本课题以具有重要生物功能的几种人类半胱氨酸和丝氨酸蛋白酶为研究对象，旨在针对底物裂解位点临近序列和结构信息进行创新性、系统性的智能计算和生物信息分析及深入研究，找出决定底物特异性的重要序列和结构特征以及规律性信息，开发高效准确的蛋白酶底物裂解位点机器学习模型，为深入了解蛋白酶生化反应规律、底物识别和抑制剂设计提供重要信息与理论指导。

在自然基金青年基金资助下，我们获取了以下的研究成果：（i）对胱天蛋白酶（Caspases，又称半胱天冬酶）和颗粒酶B（Granzyme B）的底物裂解特异性开展了一系列深入的研究。（ii）研究了几种主流的大肠杆菌重组蛋白可溶性的预测工具，从一下几个方面进行了全名综合的比较和评价，包括：预测性能、可用性、功用、工具开发和确证方法。（iii）构建了四种物种（包括酵母、人、老鼠和拟南芥）特异性的蛋白质泛素化数据集，并全面系统的比较综述了目前存在的各类蛋白质泛素化位点的预测方法的优劣性，它们各自的特点以及用户在使用这些方法时应注意的问题。（v）研发了一种基于氨基酸指数(AAindex)的有效生物信息算法，来预测区分蛋白质结构中存在的平行二聚体和三聚体卷曲螺旋(coiled coils)，这一工具称为RFCoil。（vi）对蛋白质结晶过程的重要理化性质与蛋白质克隆、表达、纯化和结晶过程多个实验步骤的相关性进行了深入研究。开发出新的序列分析工具，称为PredPPCrys (http://www.structbioinfor.org/PredPPCrys/)，用以精确预测目标蛋白质结晶多步骤实验过程成功的倾向性。通过该基金的资助，我们共发表了17篇SCI论文。