engleski

In silico analysis of polyketide synthases

With the development of new high capacity DNA sequencing techniques the use of a computational approach in biology is gaining even greater importance. In this thesis several of the bioinformatic methods have been employed on one class of enzymes – polyketide synthases (PKSs) - which were used to decode information stored in DNA into its more useful form - a chemical compound synthesised by the enzyme. To get the information about the chemical compound synthesized by the enzyme, DNA sequences coding for modular biosynthetic clusters first have to be identified. For that purpose a top-down approach relying on Hidden Markov Model (HMM) profiles, describing all type I PKS domains was used. HMM profiles were chosen due to superior sensitivity coming from capturing information from multiple members of the protein family. Another advantage of HMM profiles is their robustness which was demonstrated in annotation of PKS genes in the genome of Dictyostelium discoideum where profiles were "retrained" in several steps with organism-specific sequences and were able, at the end, to accurately detect all deviations within the sequence, as was the case with introns occurring within domains. When all domains constituting PKS are identified their activity and/or specificity has to be determined. In this thesis several methods were used - comparison of motifs consisting of specificity-determining residues, the statistical parameters of similarity search or predefined rules based on existing knowledge, depending on the type of the domain. After all existing components of the system (all PKS domains) were found and their properties (activity/specificity) determined they were organised into a "functioning system" which is able to predict the chemical entity synthesised by the system. In addition to the information about protein function and specificity/activity, information about the structure of the protein as well as its interactions can also be extracted from the DNA sequence. Structures of polypeptides constituting the daunorubicin PKS were built from their DNA sequences using homology modelling methods. These structures were later used for rigid body docking simulations which revealed interacting partners and revealed some information about the overall structure of the complex.

Polyketide synthases; Hidden Markov Models; Dictyostelium discoideum; in silico studies; daunorubicin

nije evidentirano