Novel activities and sequence-function relationships of a synthetic P450 family

对自然产生的蛋白质序列的比较分析被广泛用于获得关于蛋白质序列与其结构和功能之间关系的信息。虽然与进化相关的序列分析往往受到有限和有偏见的信息的影响，但合成蛋白质家族提供了新的机会来探索结构和生物功能的决定因素，而不受自然选择的限制。最近在Arnold教授的团队中通过重组构建了一个由6,561个P450血红素蛋白组成的独特的合成家族，并被证明同时具有高度突变和正确折叠[1]。这个P450嵌合体家族代表了一种前所未有的资源，可以探索通过重组以及序列-功能关系获得的一系列新活动。我建议在一系列化合物上测试这些嵌合酶的活性，包括那些不被任何亲本蛋白质接受的化合物，以确定嵌合体所展示的底物分布。我将使用Logistic回归、k-均值聚类和两者的新组合来分析数据，以便分别建立给定底物或一组底物上嵌合体活性的预测模型。我的工作旨在了解P450蛋白质的序列-功能关系，并将允许识别能够羟化所需化合物的蛋白质序列；这些预测将得到实验验证。

Comparative analyses of naturally occurring protein sequences are widely used to obtain information about the relationship between a protein’s sequence and its structure and function. Whereas analyses of sequences that are related by evolution often suffer from limited and biased information available, synthetic protein families offer new opportunities to explore the determinants of structure and biological function free from the constraints of natural selection. A unique synthetic family comprising 6,561 P450 heme proteins was recently constructed in the group of Prof. Arnold by recombination and shown to be simultaneously highly mutated and properly folded [1]. This family of P450 chimeras represents an unprecedented resource to explore the range of new activities that are accessible by recombination as well as sequence-function relationships. I propose to test the activities of these chimeric enzymes on an array of compounds, including ones that are not accepted by any of the parental proteins, in order to determine the substrate profiles exhibited by the chimeras. I will analyze the data using logistic regression, k-means clustering and a novel combination of both in order to build predictive models of chimera activity on a given substrate or a cluster of substrates, respectively. My work aims at understanding the sequence-function relationship of P450 proteins and shall allow to identify protein sequences capable of hydroxylating a desired compound; these predictions will be experimentally verified.