Predictive modelling of ligand binding to flexible proteins

现在人们认识到，人类蛋白质组的很大一部分是由极其灵活的蛋白质组成的，这些蛋白质的结构不能用独特的折叠来表征。这些固有的无序蛋白(IDPs)在细胞中发挥关键作用，并与癌症、神经退行性疾病和糖尿病等一系列疾病有关。这些疾病将在老龄化的英国人口中变得越来越普遍。因此，开发能够与国内流离失所者结合并调节其功能的小的类药物分子是非常可取的。然而，使用实验技术很难确定境内流离失所者所采用的构象的范围和性质。因此，国内流离失所者通常被制药业认为是“无法下药的”。重要的是开发新的技术和技术，以应对国内流离失所者造成的国家和全球健康挑战。计算机模拟有可能提供IdP/小分子相互作用的详细结构模型，以指导合理的基于结构的药物设计工作。然而，标准的模拟技术不能完成这项任务。他们对分子间相互作用的描述过于近似。探索境内流离失所者复杂的能源状况太耗时了。因此，有必要开发新的模拟方法来处理IDPs的高度灵活性。我们建议开发新的分子模拟算法，使快速计算IDPs与类药物分子的络合物的结构、热力学和动力学性质成为可能。我们的研究围绕三个目标展开：1.我们将开发一种力场优化方法，该方法迭代偏向分子动力学模拟和能量重新加权，以最大限度地减少预测IDP/小分子络合物的分子可观测值(例如，核磁共振化学位移)的系统误差。2.我们将开发一种模拟方法，将“即时”计算工作引导到对分子构象的探索上，这些分子构象对预测IDP/小分子络合物动力学的总体不确定性贡献最大。我们将进行模拟研究，以阐明小分子与选定的IDPs结合的机制。目前，这种相互作用挑战了我们对分子识别的理解。例如，测试系统将包括在几种癌症的发展过程中发挥关键作用的国内流离失所者、c-myc和p53。因此，这项研究的主要目标是开发新的计算方法，使临床前药物发现工作能够通过基于结构的方法来瞄准内在无序的蛋白质。此外，在这项研究中开发的算法和软件将具有广泛的适用性，这项研究还将使其在广泛的软凝聚物质研究领域得到应用。

It is now recognised that a large fraction of the human proteome is made of extremely flexible proteins whose structure cannot be characterized by a unique fold. These intrinsically disordered proteins (IDPs) play key roles in cells and have been implicated in a striking range of diseases such as cancers, neurodegenerative disorders and diabetes. These diseases will become increasingly prevalent in an aging UK population. It is therefore highly desirable to develop small drug-like molecules that could bind to IDPs and modulate their function. Yet it is very difficult to determine the range and nature of conformations adopted by an IDP using experimental techniques. IDPs are therefore generally considered "undruggable" by the pharmaceutical industry. It is important to develop new techniques and technologies to address the national and global health challenges caused by IDPs. Computer simulations have the potential to provide detailed structural models of IDP/small molecule interactions to guide rational structure-based drug design efforts. However standard simulation techniques are unable to perform this task. Their description of intermolecular interactions is too approximate. The exploration of the complicated energy landscape of IDPs is too time consuming. It is therefore essential to develop novel simulation methodologies that can handle the high flexibility of IDPs.We propose the development of new molecular simulation algorithms that will enable the rapid computation of the structural, thermodynamic and kinetic properties of IDPs in complex with drug-like molecules. Our research is structured around three objectives:1. We will develop a force-field optimisation method that iterates biased molecular dynamics simulations and energy reweighting to minimize systematic errors in the prediction of molecular observables for IDP/small molecule complexes (e.g. NMR chemical shifts). 2. We will develop a simulation method to steer "on-the-fly" computational efforts towards the exploration of molecular conformations that contribute the most to overall uncertainties in predicting the dynamics of IDP/small molecule complexes.3. We will perform simulation studies to elucidate the mechanisms of small molecule binding to selected IDPs. Such interactions currently challenge our understanding of molecular recognition. Test systems will include for instance the IDPs c-myc and p53 that play key roles in the progression of several cancers. The primary goal of this research is therefore to develop new computational methodologies that will enable preclinical drug discovery efforts to target intrinsically disordered proteins with structure-based approaches. In addition the algorithms and software developed during this research will be widely applicable and this research will also enable applications in a broad range of soft-condensed matter research areas.