Quantitative and evolutionary global analysis of enzyme reaction mechanisms

QP601.N2;Enzymes;Enzymes--Classification;Enzymes--Evolution

Nath, Neetika ; Mitchell, John B. O. ; Mitchell, John B. O.

University:University of St Andrews

Department:Chemistry (School of)

关键词: QP601.N2;    Enzymes;    Enzymes--Classification;    Enzymes--Evolution;   

Others  :  https://research-repository.st-andrews.ac.uk/bitstream/handle/10023/6899/NeetikaNathPhDThesis.pdf?sequence=3&isAllowed=y

来源: DR-NTU

【 摘 要 】

The most widely used classification system describing enzyme-catalysed reactionsis the Enzyme Commission (EC) number. Understanding enzymefunction is important for both fundamental scientific and pharmaceuticalreasons. The EC classification is essentially unrelated to the reaction mechanism.In this work we address two important questions related to enzymefunction diversity. First, to investigate the relationship between the reactionmechanisms as described in the MACiE (Mechanism, Annotation,and Classification in Enzymes) database and the main top-level class of theEC classification. Second, how well these enzymes biocatalysis are adaptedin nature.In this thesis, we have retrieved 335 enzyme reactions from the MACiEdatabase. We consider two ways of encoding the reaction mechanism indescriptors, and three approaches that encode only the overall chemicalreaction.To proceed through my work, we first develop a basic model to clusterthe enzymatic reactions. Global study of enzyme reaction mechanismmay provide important insights for better understanding of the diversity ofchemical reactions of enzymes. Clustering analysis in such research is verycommon practice. Clustering algorithms suffer from various issues, such asrequiring determination of the input parameters and stopping criteria, andvery often a need to specify the number of clusters in advance.Using several well known metrics, we tried to optimize the clusteringoutputs for each of the algorithms, with equivocal results that suggested theexistence of between two and over a hundred clusters. This motivated us todesign and implement our algorithm, PFClust (Parameter-Free Clustering),where no prior information is required to determine the number of cluster. The analysis highlights the structure of the enzyme overall and mechanisticreaction. This suggests that mechanistic similarity can influence approachesfor function prediction and automatic annotation of newly discovered proteinand gene sequences.We then develop and evaluate the method for enzyme function predictionusing machine learning methods. Our results suggest that pairs of similarenzyme reactions tend to proceed by different mechanisms. The machinelearning method needs only chemoinformatics descriptors as an input andis applicable for regression analysis.The last phase of this work is to test the evolution of chemical mechanismsmapped onto ancestral enzymes. This domain occurrence and abundancein modern proteins has showed that the/architecture is probablythe oldest fold design. These observations have important implications forthe origins of biochemistry and for exploring structure-function relationships.Over half of the known mechanisms are introduced before architecturaldiversification over the evolutionary time. The other halves of the mechanismsare invented gradually over the evolutionary timeline just after organismaldiversification. Moreover, many common mechanisms includes fundamentalbuilding blocks of enzyme chemistry were found to be associatedwith the ancestral fold.