Protein structure-aided prediction of the phylogenetic ancestry of genes. The case of IGF system

Abstract In eutherians, the different elements of the insulin/insulin-like growth factor (IGF) system (ligands, receptors, binding proteins, proteases, inhibitors of proteases…) interact with each other. In the present work, we were interested in the question of when these elements appeared in the course of evolution and whether they appeared at the same evolutive node or one after the other.For this purpose, we have considered phylogenetic relationships extracted from two versions of Ensembl (releases 80 and 113). Moreover, we considered remote relationships detected by comparison of experimental or predicted 3D structures with 3D structures predicted at proteome scale, after validations by independent and reciprocal sequence similarity searches when needed.We showed that insulin-like/IGF peptides and their receptors appeared in non-vertebrate species, as well as the protein encoded by the non-vertebrate ecdysone-inducible gene L2 (Impl2), whereas IGF-binding proteins (IGFBPs) have appeared in vertebrate ancestor. Through structure and sesquence similarity searches, homologs of IGFBP-1/2/5, GH and GHR were detected in more distant species than thought to date (in Lamprey for − 1/2, Amphioxus for − 5, Lamprey for GH/GHR). We also showed that the pregnancy-associated plasma protein A (PAPP-A) and PAPP-A2 metzincin proteases, which degrade several IGFBPs, might have appeared in non-vertebrates and that their two inhibitors, proMBP and STC, probably appeared before both proteases. Overall, these results show that ancestors of insulin-like peptides appeared before the divergence of vertebrate and non-vertebrate species, as well as their receptors, and that the regulation strategies of this ligand might be different between both clades.