The interactions between small molecules and diverse enzyme, membrane receptor and channel proteins are associated with important biological processes and diseases. This makes the study of binding motifs between proteins and ligands appealing to scientists. We use multiple computational techniques to unveil the protein-ligand interaction motifs from three perspectives. Firstly, from the perspective of proteins, by comparing the structure differences and common features of different binding sites for the same ligand, 3-dimensional motifs that represent the favorable interactions of the same ligands can be extracted. The goal is for such a motif to represent the shared features for binding a certain ligand in unrelated proteins, while discriminating from other ligands. The 3-dimensional motifs for cholesterol and cholate binding to non-homologous protein sites have been extracted, using SimSite3D alignment and analysis of the conserved interactions between these sites. The 3-dimensional protein motif for cholesterol binding can give about 80% accuracy of true positive sites with a low false positive rate. Furthermore, an online server CholMine was established so that the users can use this approach to predict cholesterol and cholate binding sites in proteins of interest. These motifs can help annotation of protein functions, drug discovery and the design of mutations. Secondly, from the perspective of ligands, interaction motifs can be represented as molecular features important for biological activities of ligands. Searching and summary of shared motifs from pretested series of ligand candidates can provide rational guidance to further drug improvement and screening. Here, we report a series of potential sea lamprey olfactory receptor 1 antagonists discovered from databases we designed of molecules that are similar to the native ligand, 3kPZS. Compounds with overall electrostatic and shape similarity to 3kPZS were assessed by using ROCS software, and their initial important feature matches to 3kPZS were analyzed, to prioritize compounds for biological testing. Then, the molecular features important to biological activities were summarized using SALI analysis and functional group matchprint analysis. By combining theses approaches, 12 compounds were discovered that suppress the detection of 3kPZS by at least 45%, and the most active compounds have entered field testing. Thirdly, dissecting the components of protein-ligand binding energies is also important to define the key determinants of ligand interaction with a protein site. Through analyzing the correlation coefficient of interaction energies between a series of alpha-phenylalanine substitutes and PaPAM and biological activities of these compounds, the dominant factor that determine the activities of the compounds was revealed, which was steric effect between the binding site and these compounds. From the analysis, mutations at the residues of the binding site were suggested to change or improve the catalytic efficiency of the enzyme. Given these three approaches, we envision a more integrated approach in the future that combines the analysis of shared protein-ligand interactions, shared interaction features from active ligands and shared features of protein binding sites to identify even more selective and tight-binding ligands.
Read
