G protein-coupled receptors (GPCRs) can form homodimers, heterodimers, or higher-order molecular complexes (oligomers). The reports on the change of functions through the oligomerization have been accumulated. Inhibition of GPCR oligomerization without affecting the protomer's overall structure would clarify the oligomer-specific functions although inhibition experiments are costly and require accurate information about the interface location. Unfortunately, the number of experimentally determined interfaces is limited. The precise prediction of the oligomerization interfaces is, therefore, useful for inhibition experiments to examine the oligomer-specific functions, which would accelerate investigations of the GPCR signaling. However, interface prediction for GPCR oligomerization is difficult because different GPCR subtypes belonging to the same subfamily often use different structural regions as their interfaces. We previously developed a high-performance method to predict the interfaces for GPCR oligomerization, by identifying the conserved surfaces with the sequence and structure information. Then, the structural characteristic of a GPCR structure is regarded to be a thick-tube like conformation that is approximately perpendicular to the membrane plane. Our method had successfully predicted all of the interfaces available on that day. We had launched a web server for our interface prediction of GPCRs (GRIP). We have improved the previous version of GRIP server and enhanced its usability. First, we discarded the approximation of the GPCR structure as the thick-tube-like conformation. This improvement increased the number of structures for the prediction. Second, the FUGUE-based template recommendation service was introduced to facilitate the choice of an appropriate structure for the prediction. The new prediction server is available at http://grip.b.dendai.ac.jp/∼grip/.
Keywords: G protein-coupled receptors (GPCRs); amino acid sequence; interface; prediction; protein–protein interaction; structure.