基因模块中转录调控蛋白识别特异性的预测和改造

在人工基因线路和模块中，蛋白质是最主要的功能执行者。基因线路最终能实现哪些功能取决于我们预测、改造和操控蛋白质分子间特异性识别的能力。我们将以TetR家族转录调控蛋白为对象，建立模型，预测其helix-turn-helix类型DNA结合结构域与操纵序列的特异性识别；对预测序列进行实验评估，并基于此数据集建立机器学习模型，进一步推广到其他天然或人工设计蛋白与DNA的相互作用预测；我们还将以该家族转录调控蛋白为对象，建立方法根据氨基酸序列及配体结合口袋模型对蛋白质与小分子的特异性识别进行预测、评估；采用定向进化等手段，改造蛋白质配体特异性。通过研究这一体系，我们将发展并检验若干原创的、能推广到其他蛋白质家族的方法,挖掘基因组中未被发现的小分子/蛋白/DNA相互作用，或设计、改造已有的特异性相互作用赋予新的“连线”功能，提升合成生物学基因线路和基因模块的应用能力。

In synthetic gene circuits or modules, proteins are the main molecules encoded to carry out functions. A key factor that determines what can be achieved with synthetic gene circuits is our ability to predict, design and manipulate specific molecular recognition of proteins. We plan to use the TetR family of transcription regulatory proteins as model systems. We aim to build models that can predict operator DNA sequences from the amino acid sequences of the helix-turn-helix type DNA binding domains, and to redesign amino acid sequences to recognize orthogonal operators using these models. We will evaluate the predicted sequence experimentally and develop machine-learning models based on this data set to further predict the interaction between other natural or artificially designed proteins and DNA. We also plan to consider the TetR family proteins as examples for the prediction of ligand based on amino acid sequences and comparatively constructed 3-dimensional models of ligand binding proteins. Protein design and directed evolution method will be developed to reengineer ligand specificity, such as the inducer specificity of transcriptional factors. Through the study of these systems, we will develop and test a number of original methods which can be extended to other protein families, design or modify existing specific interactions to give new "wiring" functions and enhance the design capabilities of gene circuits or modules in synthetic biology.