Refine
Has Fulltext
- yes (1)
Is part of the Bibliography
- yes (1)
Year of publication
- 2010 (1)
Document Type
- Doctoral Thesis (1) (remove)
Language
- English (1) (remove)
Keywords
- Integrasen (1) (remove)
Institute
In this thesis, computational structure-based design approaches were employed to target the HIV-1 integrase and the macrophage infectivity potentiator (MIP) of Legionella pneumophila. The thesis yields valuable information about the mechanism of action of a known class of integrase inhibitors and a novel approach towards enzyme inhibition, which still is mainly unaddressed in current integrase research. For the MIP enzyme, two small-molecule MIP inhibitors were discovered. The computational studies of HIV-1 integrase have provided valuable information for IN inhibitor design. Docking experiments supported the hypothesis that the well-known diketo acid inhibitors enter the IN active site not as free ligands, but rather as metal complexes. These results help to reveal the mechanism of action of this important class of IN inhibitors.To give an impulse for the development of a novel class of inhibitors, a new strategy towards IN inhibition was introduced: An alternative binding site, the dimerization interface of an IN catalytic core domain monomer, was explored for inhibitor design. The lack of structural data of the free monomer was overcome by extensive MD studies. Snapshots derived from the MD simulation were used as protein input structures in a docking study with the inhibitory peptide YFLLKL to reveal its potential binding mode. The docking procedure showed that the peptidic ligand binds to a dimerization interface conformation which shows a Y-shaped binding site.. The next step was to address this protein conformation with small, non-peptidic molecules. The first strategy towards finding small-molecule interface binders was to create a pharmacophore model with hydrophobic features and shape constraints, aiming to find molecules with a good complementarity to the Y-shaped dimerization interface. Virtual screening yielded a total of 10 compounds, which all displayed good shape complementarity and favorable hydrophobic interactions. Unfortunately, none of the compounds showed a reproducible inhibitory activity in biological assays. Some doubts remain about the validity of the assay results: The use of BSA was critical, since it is not unlikely that BSA “intercepted” the hydrophobic candidate compounds. The first strategy towards finding small-molecule dimerization inhibitors was reconsidered: In the second approach, the satisfaction of hydrogen bonding residues at the dimerization interface, was of major interest. Two pharmacophore models were employed, which retrieved several hundred hit molecules. However, docking of these molecules showed that still many hydrogen bonding groups of the protein remained unaddressed by the ligands. Eventually, after visual inspection, only eight molecules were selected as candidate compounds for further testing (results pending). This small “yield” underlines the difficulties in finding interface binders: The IN dimerization interface is a peculiar target with frequently alternating basic, acidic, and hydrophobic residues. It is not a well-ordered binding site with continuous hydrophobic areas and distinct hydrogen bond donors / acceptors. Other protein-protein interfaces show such well-ordered binding sites. Accordingly, the peculiarity of the IN dimerization interface, in addition to the delicate task of disrupting protein-protein interactions at all, makes the development of IN dimerization inhibitors very challenging. For MIP, the studies revealed two experimentally validated MIP inhibitors, which significantly reduce MIP enzymatic activity. To our knowledge, no small-molecule MIP inhibitor has been reported in the literature so far. A detailed analysis of the available structural data of MIP and a comparison to the human PPIase counterpart, FKBP12, pointed out a conformational diversity among the MIP structures and a crucial difference between the two PPIases, which could be traced to mainly one residue (Tyr109). The detailed comparison of FKBP12 and MIP complex structures made it possible to give an explanation, why a ketoacyl-substituted pipecoline derivative most probably does not bind to MIP, but a sulfone-substituted pipecoline derivative does bind to MIP. Knowledge of Legionella MIP inhibitors could be transferred also to other organisms (e.g. trypanosoms), where homologous MIP proteins are also pathological factors.