学术报告
题目：Systematic analysis of large enzyme families: modular structure and functionally relevant positions
报告人：Juergen Pleiss
Institute of Technical Biochemistry, University of Stuttgart, Germany
时间：2014年4月10日（周四），10:00-11:00 AM
地点：生命科学学院208会议室
Most of the biocatalytically interesting enzyme families are highly diverse with pairwise sequence identities below 15% and consist of a large number of proteins: 40000 proteins in the current version of the Thiamine-diphosphate dependent Enzyme Engineering Database (www.TEED.uni-stuttgart.de), 16000 in the Cytochrome P450 Engineering Database (www.CYPED.uni-stuttgart.de), and 4000 in the Laccase and Multicopper Oxidase Engineering Database (www.LccED.uni-stuttgart.de). A systematic analysis based on sequence alignment allows for measuring the distance between sequences, identifying equivalent positions in different enzymes, and developing optimal biocatalyst for a desired reaction. However, standard methods of multiple sequence alignment are not feasible in a situation of rapidly increasing number of highly diverse sequences.
Previously we have implemented a strategy to assign standard position numbers to each position in a family of homologous proteins (Vogel 2012). Using this method to generate robust, structure-based alignments, the family of thiamine diphosphate-dependent decarboxylases was systematically analyzed by determining the amino acid variation at equivalent positions. This analysis revealed a mechanistically divergent family, the glyoxylate carboligases (Vogel 2012), predicted stereoselectivity determining positions by generalizing the S-pocket concept (Knoll 2006) to other decarboxylases (Westphal 2013, Gocke 2008; Rother 2011), and identified hotspot positions that resulted in improved substrate specificity.
Systematic sequence analysis was also applied to cytochrome P450 monooxygenases to construct a focused, highly enriched CYP102A1 mutant library by combining five hydrophobic amino acids in two positions which was screened with four terpene substrates (Seifert 2009). 11 of 25 variants demonstrated either a strong shift or improved regioselectivity toward at least one substrate. Using an iterative approach combining sequence analysis and modeling, two further positions were identified that resulted in a triple mutant that converted limonene to perillyl alcohol with a selectivity of 97 %, in comparison to the wild type enzyme where perillyl alcohol was not observed (Seifert 2011).
The thiamine diphosphate-dependent enzymes (Vogel 2012) as well as the multicopper oxidases (Sirim 2011) are multidomain proteins and are active as oligomers. In both protein families, the catalytically active cofactors are bound at domain interfaces. During evolution, Nature created a large variety of domain arrangements with different orders and numbers of domains, and with binding of the cofactor inside one monomer or between different monomers. By applying a systematic and comprehensive analysis of sequences and structures, the modular structure of these enzymes can be represented by simple schemes, despite the apparently high level of structural complexity.
联系人：秦咏梅（电话：62758885）
欢迎各位老师和同学积极参加！