Novel Strategies for Model-Building of G Protein-Coupled Receptors
The G protein-coupled receptors constitute still the most densely populated proteinfamily encompassing numerous disease-relevant drug targets. Consequently, medicinal chemistry is expected to pursue targets from that protein family in that hits need to be generated and subsequently optimized towards...
|Online Access:||PDF Full Text|
No Tags, Be the first to tag this record!
|Summary:||The G protein-coupled receptors constitute still the most densely populated proteinfamily encompassing numerous disease-relevant drug targets. Consequently, medicinal chemistry is expected to pursue targets from that protein family in that hits need to be generated and subsequently optimized towards viable clinical candidates for a variety of therapeutic areas. For the purpose of rationalizing structure-activity relationships within such optimization programs, structural information derived from the ligand's as well as the macromolecule's perspective is essential. While it is relatively straightforward to define pharmacophore hypotheses based on comparative modelling of structurally and biologically characterized low-molecular weight ligands, a deeper understanding of the molecular recognition event underlying, remains challenging, since the principally available amount of experimentally derived structural data on GPCRs is extremely scarse when compared to, e.g., soluble enzymes. In this context, the protein modelling methodologies introduced, developed, optimized, and applied in this thesis provide structural models that are capable of assisting in the development of structural hypotheses on ligand-receptor complexes. As such they provide a valuable structural framework not only for a more detailed insight into ligand-GPCR interaction, but also for guiding the design process towards next-generation compounds which should display enhanced affinity. The model building procedure developed in this thesis systematically follows a hierarchical approach, sequentially generating a 1D topology, followed by a 2D topology that is finally converted into a 3D topology. The determination of a 1D topology is based on a compartmentalization of the linear amino acid sequence of a GPCR of interest into the extracellular, intracellular, and transmembrane sequence stretches. The entire chapter 3 of this study elaborates on the strengths and weaknesses of applying automated prediction tools for the purpose of identifying the transmembrane sequence domains. Based on an once derived 1D topology, a type of in-plane projection structure for the seven transmembrane helices can be derived with the aide of calculated vectorial property moments, yielding the 2D topology. Thorough bioinformatics studies revealed that only a consensus approach based on a conceptual combination of different methods employing a carefully made selection of parameter sets gave reliable results, emphasizing the danger to fully automate a GPCR modelling procedure. Chapter 4 describes a procedure to further expand the 2D topological findings into 3D space, exemplified on the human CCK-B receptor protein. This particular GPCR was chosen as the receptor of interest, since an enormous experimentally derived and structurally relevant data-set was available. Within the computational refinement procedure of constructed GPCR models, major emphasis was laid on the explicit treatment of a non-isotropic solvent environment during molecular mechanics (i.e. energy minimization and molecular dynamics simulations) calculations. The majority of simulations was therefore carried out in a tri-phasic solvent box accounting for a central lipid environment, flanked by two aqueous compartments, mimicking the extracellular and cytoplasmic space. Chapter 5 introduces the reference compound set, comprising low-molecular weight compounds modulating CCK receptors, that was used for validation purposes of the generated models of the receptor protein. Chapter 6 describes how the generated model of the CCK-B receptor was subjected to intensive docking studies employing compound series introduced in chapter 5. It turned out that by applying the DRAGHOME methodology viable structural hypotheses on putative receptor-ligand complexes could be generated. Based on the methodology pursued in this thesis a detailed model of the receptor binding site could be devised that accounts for known structure-activity relationships as well as for results obtained by site-directed mutagenesis studies in a qualitative manner. The overall study presented in this thesis is primarily aimed to deliver a feasibility study on generating model structures of GPCRs by a conceptual combination of tailor-made bioinformatics techniques with the toolbox of protein modelling, exemplified on the human CCK-B receptor. The generated structures should be envisioned as models only, not necessarily providing a detailed image of reality. However, consistent models, when verified and refined against experimental data, deliver an extremely useful structural contextual platform on which different scientific disciplines such as medicinal chemistry, molecular biology, and biophysics can effectively communicate.|