Methods
Sequence sources. The FASTA sequence files for human and coral proteins used in this study were acquired from the UniProt database website at www.uniprot.org.
Sequence analysis. In-depth sequence analysis of coral was performed as (Ye et al., 2017). The 3D reconstruction was made to provide detailed, domain wise information of the coral insulin receptor. Full coral receptor model (left, right, top, and bottom view) was represented as surface representation with color coded description of individual domain.
Hhblits alignment and selection . Protein sequences were imported to the online hhblits coral protein remote homology search tool atwww.hhblits.cs.tufts.edubased on (Remmert et al., 2011). The sequences were provided as a FASTA sequence file format and the results of the remote homology detection were received via email. Coral protein sequence pdam_00006633-RA showed homology with human insulin protein and pdam_00013976-RA protein sequence with human insulin receptor.
Structural mapping of conservation. The ConSurf2016 server atwww.consurf.tau.ac.il was used to identify sequence conservation and plot it onto the structure of insulin and insulin receptors. ConSurf is a bioinformatics application for estimating the evolutionary conservation of amino acid position in a protein molecule between homologous sequences based on the phylogenetic relations.
Insulin signaling information: To uncover the coral insulin signaling we use the literature information from KEGG: Kyoto Encyclopedia of Genes and Genomes atwww.genome.jp/kegg. KEGG database consists of high-level functions of molecular information from large-scale datasets of organisms generated by high-throughput sequencing techniques.
Swiss modeling (Coral IR model construction): The coral insulin receptor homolog model was constructed using Swiss Model atwww.swissmodel.expasy.org. Swiss-Model is an integrated web-based service dedicated to homology modelling of proteins. We used the target-template alignment function of swiss model and provided the full structure of the human insulin receptor (6pxv) as a template to model pdam_00013976-RA. From the same structure we used insulin structure to model pdam_00006633-RA as coral insulin. The model was downloaded and analyzed using PyMOL software (Version-2.3.4, Schrodinger, LLC).
ClusPro protein-protein docking: Protein-protein docking studies were performed using Clus-Pro 2.0 protein-protein docking tool at www.cluspro.bu.edu. We performed the docking of coral insulin protein models with coral IR model (dimer). The docked poses were downloaded and analyzed using PyMOL software. We have analyzed the insulin binding site-1 and site-2 in detail and provided the information of interacting residues between coral insulin and coral IR. The information was extracted using PyMOL.
Autodock Vina (protein-ligand docking): Molecular docking of selected ligands with coral Insulin receptor EC and TK domain was performed using Autodock Vina (Trott and Olson, 2009). We constructed a grid box covering all desired residues using Autodock vina/Autodock tools 1.5.6 atvina.scripps.edu(Trott and Olson, 2009). The gridbox parameters for full protein include: center_x= -11.318, center_y= -15.45, center_z= 18.207, size_x=126, size_y=126, and size_z=126. We also used the full TK structure for identification of putative allosteric ligands. We used the latest inhibitor-bound structure (5HHW) for standardizing our docking experiments. Results are expressed as binding affinity for the top pose with polar contacts for the same.
Coral Insulin Receptor Kinase domain inhibitor binding site analysis: The most recent structures of human kinase domains with their inhibitors were analyzed and essential residues forming polar contacts with inhibitors were identified as the inhibitor pocket (see details in text). The identified residues were also mapped in the coral TK domain model and represented in the coral TK domain using PyMOL.