Rapid molecular evolution across amniotes of the IIS/TOR network

Abstract

The insulin/insulin-like signaling and target of rapamycin (IIS/TOR) network regulates lifespan and reproduction, as well as metabolic diseases, cancer, and aging. Despite its vital role in health, comparative analyses of IIS/TOR have been limited to invertebrates and mammals. We conducted an extensive evolutionary analysis of the IIS/TOR network across 66 amniotes with 18 newly generated transcriptomes from nonavian reptiles and additional available genomes/transcriptomes. We uncovered rapid and extensive molecular evolution between reptiles (including birds) and mammals: (i) the IIS/TOR network, including the critical nodes insulin receptor substrate (IRS) and phosphatidylinositol 3-kinase (PI3K), exhibit divergent evolutionary rates between reptiles and mammals; (ii) compared with a proxy for the rest of the genome, genes of the IIS/TOR extracellular network exhibit exceptionally fast evolutionary rates; and (iii) signatures of positive selection and coevolution of the extracellular network suggest reptile- and mammal-specific interactions between members of the network. In reptiles, positively selected sites cluster on the binding surfaces of insulin-like growth factor 1 (IGF1), IGF1 receptor (IGF1R), and insulin receptor (INSR); whereas in mammals, positively selected sites clustered on the IGF2 binding surface, suggesting that these hormone-receptor binding affinities are targets of positive selection. Further, contrary to reports that IGF2R binds IGF2 only in marsupial and placental mammals, we found positively selected sites clustered on the hormone binding surface of reptile IGF2R that suggest that IGF2R binds to IGF hormones in diverse taxa and may have evolved in reptiles. These data suggest that key IIS/TOR paralogs have sub- or neofunctionalized between mammals and reptiles and that this network may underlie fundamental life history and physiological differences between these amniote sister clades.

Keywords: insulin growth factor; insulin signaling; molecular evolution; rapamycin.

Fig. 1.

Medians of pairwise measures between all reptiles and mammals per gene for K_a, Ks, and K_a/Ks calculated in PAML. Control = 1,417 genes not in the IIS/TOR network, with 66 taxa represented in the alignments. Intracellular = 51 genes. Extracellular = 10 genes (hormones, receptors, IGF binding proteins). Extracellular genes exhibit significantly greater median K_a/Ks and median K_a than control or intracellular genes.

Fig. 2.

Protein structures for reptile and mammal IGF hormones and receptors. Reptile protein structures predicted from snake sequence homology modeled onto human protein structures from the Protein Data Bank. Enlarged positions indicate the amino acid sites predicted to be under positive selection (Table S3). (A) Reptile and mammal IGF hormones with their protein domains color coded. Positively selected sites cluster on the C-domain of reptile IGF1 but are not present in the C-domain of the reptilian IGF2. In contrast, positively selected sites cluster on the C-domain of mammal IGF2. (B) The α chain of reptile IGF1R homodimer with hormone binding domains L1, CR, and L2 labeled. The square is an enlargement with IGF1 orientated in the IGF1R binding pocket to demonstrate the clustering of positively selected sites on the interacting IGF1-IGF1R binding surfaces (36). Labeled sites (IGF1 S37 and N251; human numbering) are known to affect IGF hormone and receptor binding (Table S4). (C) Domain 11 of reptile and mammal IGF2R with IGF2 oriented toward the binding pocket to demonstrate the clustering of positively selected sites on the reptile IGF2R binding surfaces (43, 44). The magenta sites on reptile IGF2 and IGF2R were identified as coevolving amino acids using CAPS (46). Labeled sites IGF2R (1558 and 1609; human numbering) are predicted to regulate IGF2-IGF2R binding (Table S4). Like mammals, some lizards have IGF2R N1558.