Development and application of computational tools for the design of ligand bindingin proteins

近年来，计算蛋白质设计已经取得了长足的进步，成功地用于设计新的折叠、结合相互作用和在现有蛋白质支架上的碱性催化剂。然而，由于重大成就不得不受到质疑和撤回，道路一直崎岖不平。这导致了实地的重新校准，有必要将重点重新放在我们认为已经解决的设计的某些方面。我们希望将我们的努力集中在小分子配体相互作用方面，并系统地研究如何改进结合口袋的设计及其特异性。该项目还对催化功能的建立有影响，因为过渡态结合对于特定酶的设计极其重要。因此，我们希望建立我们自己的计算设计管道，并将其用于允许稳健实验验证的模型系统中的增量设计。对于管道，我们不仅自己开发工具，而且还利用现有的经过良好测试的计算方法和算法。它将是一个模块化的设置，具有定义的接口，允许我们为设计过程中的各个步骤交换和测试不同的方法，并可以稍后扩展到其他设计问题。为了进行实验验证，我们使用了稳健的生物物理和结构表征技术。此外，我们希望使用实验库设计和定向进化相结合的方法来调查确定的计算设计是否是可能的最佳解决方案。这一系统的分析将被用作一个反馈回路，以加强计算方法。通过这种方法，我们希望实现一种可靠的设计策略，并对理解蛋白质与配体的相互作用做出重大贡献。

Computational protein design has come a long way in recent years having been used successfully to design new folds, binding interactions and basic catalysts onto existing protein scaffolds. However, the path has been rocky since major achievements had to be questioned and retracted. This has led to a recalibration in the field and it is necessary to refocus on certain aspects of design that we believed to have been solved. We want to concentrate our efforts on the aspect of small molecule ligand interactions and systematically investigate how the design of a binding pocket and its specificity can be improved. The project also has implications for the establishment of catalytic function, since transition state binding is extremely important for the design of specific enzymes. Therefore we want to build our own computational design pipeline and use it for the incremental design in model systems that allow robust experimental validation. For the pipeline we are not only developing tools ourselves, but are also taking advantage of existing well-tested computational methods and algorithms. It will be a modular setup with defined interfaces that allows us to exchange and test different approaches for the various steps in the design process and that can later be extended to other design problems. For experimental validation we use robust biophysical and structural characterization techniques. Further we want to employ experimental library design combined with directed evolution to survey whether the computational design identified is the best solution possible. This systematic analysis will be used as a feedback loop to enhance the computational approach. By this method we hope to implement a reliable design strategy and to contribute significantly to the understanding of protein-ligand interactions.