Dokument

Summary:
Aldose Reductase (AR) is the first enzyme of the 'sorbitol pathway'. It is an NADPH dependent enzyme and catalyzes the reduction of various aldehydes to the corresponding alcohols. Its binding pocket shows intrinsic flexibility which is mainly mediated by a small loop region. A specificity pocket can be opened or closed via this loop stretch. In the first part of this work the relevance of considering protein flexibility in structure-based drug design was highlighted. Docking experiments were carried out for two inhibitors which were originally designed to mimic a certain binding mode. By carrying out separate docking simulations for each known major binding pocket conformation, it was shown that the intended binding mode was not predicted to be the most favorable. This hypothesis could later be verified by X-ray crystallography. However, the predictions were by no means perfect. One of the compounds induced a new conformation to the binding pocket. Thus, no appropriate crystal structure was available as template for the docking experiments. For the second molecule the importance of considering water during docking was emphasized. In the second part of this thesis a new method to simplify the tedious process of docking to multiple targets was evaluated in the context of protein flexibility: in-situ cross-docking. With this method, instead of performing sequential docking experiments of multiple ligands into multiple protein structures, several protein conformations can be addressed at once. In the next part of this thesis the flexibility of the AR binding pocket was examined in detail. It was shown that with respect to the binding pocket, flexibility is limited to only a handful amino acids close to the specificity pocket. It was elucidated how the enzyme performs its 'induced-fit' binding mechanism. To further explore the conformational space available to the AR binding pocket, multiple MD simulations were carried out. Good overall agreement between the results from the MD simulations and the crystal structure analysis was found. Residues which exhibited elevated levels of flexibilities in the MD simulations showed in most cases also differences between the single crystal structures. However, a few residues showed unexpected behavior in the MD simulations: Phe 122, Trp 219, and Tyr 309. The behavior of these residues were examined in great detail. In a further project, the generated MD trajectories were used to energetically analyze the process of 'induced-fit' adaptation. The method MM-PBSA was chosen for this purpose. Considering the enormous amount of computational power required to perform the necessary calculations and the remarkable time needed to analyze the results, MM-PBSA did not turn out to be a cost-effective method to predict binding free energies for the dataset of AR inhibitors used in this study. In the final section of this thesis a study was presented where in the first part aspects of flexibility of the C-terminal loop of AR were examined. In a combined study using MD simulations and multiple crystal structures, it was shown that there are clear differences between individual crystal structures of the same protein-ligand complex in this region of the enzyme. A nice agreement between the observations made in the MD and multiple crystal structures derived from different experimental crystallization conditions was found. In the second part of this section the unexpected occurrence of multiple ligands in and close to the binding pocket of AR was described. In summary, this study has dealt with many aspects of protein flexibility using AR as a test case. AR has proves to be a valuable test system to investigate protein flexibility with different methods.

Das Dokument ist im Internet frei zugänglich - Hinweise zu den Nutzungsrechten