G Protein-Coupled Receptors (GPCRs) are one of the most important pharmaceutical targets. They function not only as monomers but also as dimers or higher-order molecular complexes . Recently, GPCR oligomerization and its functional meaning have been extensively investigated. These studies revealed the wide variety of biochemical functions of the oligomers in cells, the combinations of subtypes required for complex formation, and the biological functions of the oligomers. Nevertheless, biochemical and biological functions of various combinations of hetero oligomers have not been clarified. One of the strategies to elucidate such functions is the modification the interfaces for GPCR oligomerization, by introducing a point mutation at the interface. Then, positions of the interfaces for GPCR oligomerization is required . However, few experimental studies have identified the interfaces. Therefore, it is important to predict the interfaces for protein-protein interactions between GPCRs. The prediction of the interfaces for GPCR oligomerization based on residue conservation is difficult, because the various subtypes of GPCRs often use different regions of the three-dimensional structure as the interfaces, even when the subtypes belong to the same subfamily. Therefore, we developed a method to predict the interfaces for oligomerization, based on the spatial distribution of conserved residues for each subtype . We launched a web service to predict the interfaces for GPCR oligomerization on this server. To our knowledge, it is the only web service that predicts the interfaces for GPCR oligomerization. It will be useful to investigate molecular functions of GPCR oligomers and molecular network via GPCRs. GRIP website is free and open to all users, and there is no login requirement.
Usage (Fig. 1)
- Input an AMINO ACID sequence of a GPCR as a query in the sequence input form ((A) in Fig. 1). The sequence should be submitted in FASTA format.
- Select a pair of a GPCR abbreviation name and its PDB ID from the select box ((B) in Fig. 1) that is used for the prediction. When a user is not sure which structure should be adopted for the prediction, click the link "structure recommendation" where simple calculation is performed to recommend several candidate structures.
- Write a PDB chain ID of the adopted GPCR structure in the chain input form ((C) in Fig. 1). The default setting is chain "A".
- Write an E-value used as a threshold in the E-value input form ((D) in Fig. 1) to set the higher bound of the E-value when collecting homologous sequences of a query sequence by BLAST. The default setting is 1e-50.
- Write a percent sequence identity threshold in the percent sequence identity threshold form ((E) in Fig. 1) to set the lower bound of the percent sequence identity when adopting close homologues from the set of sequences obtained by BLAST. The default setting is 50%.
- Select a significance level from the select box ((F) in Fig. 1) for binomial test to identify a cluster of conserved residues at the molecular surface along transmembrane helices. The default setting is 0.01. The alternative threshold is 0.05.
- Press the Submit button ((G) in Fig. 1) to execute the GRIP program.
Figure 1. Input page of GRIP server.
Clicking Figure 1 enlarges the image.
- Input an AMINO ACID sequence of a target GPCR as a query in the sequence input form ((A) in Fig. 2). The sequence should be submitted in FASTA format.
- Press the Recommend button ((B) in Fig. 2) to execute the structure recommendation program.
Figure 2. Input page of Structure Recommendation.
Clicking Figure 2 enlarges the image.
Three query candidates with parameter settings are provided in Table 1.
Table 1. Three query candidates and parameters for GRIP execution.
|No.||Accession Number||Protein Name||Adopted Structure: PDB ID||E-value Threshold||Percent Sequence Identity Threshold||Significance Level|
|1||NP_000530.1||rhodopsin [Homo sapiens]||Rhodopsin: 1L9H||1e-50||50%||0.05|
|2||NP_002554.1||P2Y purinoceptor 1 [Homo sapiens]||P2Y12R: 4NTJ||1e-50||50%||0.05|
|3||NP_057658.2||D(2) dopamine receptor isoform short [Homo sapiens]||D3R: 3PBJ||1e-50||50%||0.05|
The two types of input data are processed by the following six steps . At first, the MSA is aligned together with the sequence of the structure, for which the three-dimensional structure is available at atomic resolution. Second, the conservation score is calculated at each alignment site of the MSA by the method of Valdar and Thornton . Then, the sequence of the structure is ignored, to calculate the conservation score. Third, the calculated conservation scores are assigned to the corresponding residues of the structure. Fourth, the buried residues and the residues at the extracellular and the intracellular loops of the structure, and the corresponding residues in the target sequence are ignored because the interfaces for the oligomerization are assumed to reside at the lipid-facing surface of the transmembrane helices. Fifth, the remaining residues are projected on a plane that is assumed to be parallel to the membrane plane. In order to define the plane for the projection, we defined two planes at both extracellular and intracellular sides of transmembrane helices. Each plane was defined by fitting a plane to the three-dimensional coordinates of the extracellular or intracellular terminus of seven transmembrane helices by least-square method. The plane for the projection is defined by averaging a pair of normal vectors of both planes. As a result, a ring-like distribution of the projected residues is generated on the plane. Finally, we adopt a sector within the ring-like distribution as a region corresponding to an interface, when the sector includes a significantly large number of conserved residues, and an ad hoc objective function to evaluate the accumulation of conserved residues yields the maximum score. If the GPCR of interest could have more than one interface, then the interface residues predicted by the above procedure are removed from the projected coordinate data on the plane, and the sixth step is performed again to the remaining residues to predict the other interface. Further details are described in reference .
- Nemoto W, Toh H. Prediction of interfaces for oligomerizations of G-protein coupled receptors. Proteins. 2005;58:644-60. PubMed PMID: 15593372.
- Valdar WS. Scoring residue conservation. Proteins. 2002;48:227-41. Review. PubMed PMID: 12112692.
Figure 3. Flowchart of the data processing in GRIP server.
Structure selection by FUGUE should be performed when a user is not sure which structure is appropriate for a structure of each query sequence.Clicking Figure 3 enlarges the image.
Akira Saito1, Seiji Satoh2, Atsushi Okamoto2, Yoichi Murakami3,4, Kenji Mizuguchi3, Hiroyuki Toh5, Wataru Nemoto1,6
- Division of Life Science and Engineering, School of Science and Engineering, Tokyo Denki University (TDU), JAPAN.
- Asubio Pharma Co., Ltd, JAPAN.
- Laboratory of Bioinformatics, National Institutes of Biomedical Innovation, Health and Nutrition (NIBIOHN), JAPAN.
- Department of Informatics, Tokyo University of Information Sciences, JAPAN.
- Department of Biomedical Chemistry, School of Science and Technology, Kwansei Gakuin University, JAPAN.
- Division of Life Science and Engineering, Department of Life Science and Engineering, Tokyo Denki University (TDU), JAPAN.
Computational Biology Laboratory