Experimentally derived quantitative model of protein evolution

点击翻译按钮获取中文摘要

Phylogenetic analysis of sequences is an essential computational tool in biology. It is central to understanding the evolution of protein structure and function, speciation, the relationship among genes and genealogies, and the dynamics of the spread of pathogens. In biochemistry, it is the starting point for understanding the evolution of structure-function relationship in proteins. At the core of any phylogenetics method is a quantitative model of molecular evolution that describes the rate of mutation and the stringency of selection. However, a major challenge is that the widely used quantitative models of coding sequence evolution assume that all sites in a gene (or protein) are independent and equivalent. Both assumptions are inconsistent with protein biophysics. Additionally, this transition matrix is applied to all proteins, ignoring the fact that protein folds affect the stringency of selection in the substitution matrix.The long-term objective of the research program is to address these challenges by developing novel methods in molecular phylogenetics. A major innovative component is that instead of assuming a substitution matrix or estimating them from sequence alignments, we will directly measure the amino acid propensity for each site in the protein and embed it in a phylogenetics model. In the near term, we will focus our methods development on dihydropteroate synthase (DHPS), a core metabolic enzyme in the folate synthesis pathway. DHPS is an excellent model because it is an important target of sulfonamide antibiotics and follows a TIM-barrel fold, the most prevalent type of protein fold in a proteome. First, we will determine the fitness landscape of DHSP by performing high-throughput measurement of the per site amino acid preference. Second, we will develop a phylogenetic inference model based on the experimentally derived substitution matrix. Third, we will reconstruct the evolution of DHPS structure-function relationship using the new phylogenetic inference model. While we focus on DHPS on the short-term, the long-term view of our research program is to generalize this approach to other protein folds. Indeed, this approach could lead to a proteome-wide experimentally derived model of evolution. This multi-disciplinary research program will be used to train students and other highly qualified personnel at the interface of biophysics, biochemistry, molecular evolution, and bioinformatics. This training broad training in both theory and experiment could also provide valuable training for students for their future careers in both academia and industry.