Modelling of melanocortin receptors and their ligands

Sammanfattning: Five subtypes of melanocortin receptors (MC1-MC5R) mediate a wide range of physiological actions, including control of skin tanning, inflammation and food intake. The natural ligands for these receptors are the melanocortins, which are linear peptides 11 to 39 amino acids long. The aim of this study was to investigate the interactions between melanocortin receptors and melanocortins using computational modeling tools.Two three-dimensional (3D) models of the MC, receptor were created using homology modeling. The first model was based on the 3D structure of bacteriorhodopsin, while the second was based on the two-dimensional map of rhodopsin. Cyclic analogues of the melanocortins containing the core sequence motif HFRW, common to all melanocortins, could be docked into the receptor models. According to these models, the ligand binding pocket was located between the receptors' first, second, third, sixth and seventh transmembrane regions. The rhodopsin based model indicated that the His residue of the melanocortin's core sequence showed only minor interactions with the receptor. Synthesis and receptor binding of cyclic peptide analogues lacking this His residue gave experimental support for the modeling results. Support for the 30 models had also been obtained in binding studies using chimeric MC1MC3 receptors. However, the visual analysis of this binding data was subjective. A novel objective approach was developed for the analysis of data of chimeric proteins interacting with peptides. The new approach, which is based on the use of multivariate analysis tools, such as PLS, gave a very exact picture for the interactions of melanocortin receptors with their ligands. Moreover, the results of the 3D modeling and PLS modeling agreed well. The novel method seems applicable for the analysis of almost any interactions between several target and ligand molecules.