Complex Evolution of Insect Insulin Receptors and Homologous Decoy Receptors, and Functional Significance of Their Multiplicity

Abstract

Evidence accumulates that the functional plasticity of insulin and insulin-like growth factor signaling in insects could spring, among others, from the multiplicity of insulin receptors (InRs). Their multiple variants may be implemented in the control of insect polyphenism, such as wing or caste polyphenism. Here, we present a comprehensive phylogenetic analysis of insect InR sequences in 118 species from 23 orders and investigate the role of three InRs identified in the linden bug, Pyrrhocoris apterus, in wing polymorphism control. We identified two gene clusters (Clusters I and II) resulting from an ancestral duplication in a late ancestor of winged insects, which remained conserved in most lineages, only in some of them being subject to further duplications or losses. One remarkable yet neglected feature of InR evolution is the loss of the tyrosine kinase catalytic domain, giving rise to decoys of InR in both clusters. Within the Cluster I, we confirmed the presence of the secreted decoy of insulin receptor in all studied Muscomorpha. More importantly, we described a new tyrosine kinase-less gene (DR2) in the Cluster II, conserved in apical Holometabola for ∼300 My. We differentially silenced the three P. apterus InRs and confirmed their participation in wing polymorphism control. We observed a pattern of Cluster I and Cluster II InRs impact on wing development, which differed from that postulated in planthoppers, suggesting an independent establishment of insulin/insulin-like growth factor signaling control over wing development, leading to idiosyncrasies in the co-option of multiple InRs in polyphenism control in different taxa.

Keywords: decoy of insulin receptor; gene structure; insects; insulin receptor; insulin signaling; wing polyphenism.

Fig. 1.

Phylogenetic tree of insect insulin receptors and decoy of insulin receptors identified in 98 species from 23 orders. The numbers in condensed branches indicate the numbers of species studied. Orange arrow marks the Cluster II duplication within Polyneoptera. Red arrow (DR2) indicates the loss of the tyrosine kinase domain in advanced Holometabola, giving rise to the decoy of insulin receptor gene DR2 (red). Green arrow marks the Cluster I duplication in Gerromorpha, blue arrow (SDR) marks the loss of the tyrosine kinase domain in Muscomorpha, giving rise to the secreted decoy of insulin receptor gene SDR (blue). The topology and branching supports were inferred using RAxML maximum likelihood algorithm with WAG + Γ model (−ln = 285839.374747). The bootstrap values calculated from 500 replicates are shown for branches represented in >50% of trees. Full version of the tree, together with the taxon coverage and nomenclature used in previous studies, is given in the supplementary figures 2 and 3 (Supplementary Material online).

Fig. 2.

Gene structures and protein domains of InRs and decoys of insect insulin receptors. (A) Intron and exon structures and homologous intron–exon boundaries (gray vertical lines) in Cluster I and Cluster II InR, DR2, and SDR genes in selected insect representatives, including the taxa, in which important duplications were detected (stick insects, termites, semiaquatic bugs, and linden bug) and compared with InR gene structure in the sister group of insects, the crustacean Daphnia pulex. (B) Protein domains recognized in the protein sequences reconstructed from the transcripts of Cluster I and Cluster II InRs, DR2, and SDR, and compared with EGFR protein.

Fig. 3.

Relative expression of insulin receptor genes and decoy of insulin receptor genes in different tissues of termite workers and 10-day-old male and female neotenic reproductive (A), adult male and female fruit flies (B), and adult male and female linden bugs (C). The heatmaps represent mean expression values from two to four replicates per phenotype and tissue, relative to the control gene rp49 and shown on a log scale. Detailed comparisons are given in the supplementary figures 12, 14, and 15 (see Supplementary Material online).

Fig. 4.

Interaction of insulin signaling pathway, stress, and wing polymorphism in Pyrrhocoris apterus studied by means of RNAi experiments. One-day-old fourth instar P. apterus larvae of brachypterous or macropterous strain were subjected to the treatment, raised to adults, and the proportion of obtained short- and long-winged adult phenotypes was scored. (A) Brachypterous strain larvae injected with InR1a, InR1b, InR2, Pilp1 or Pilp2/3 dsRNA, or negative control egfp dsRNA. (B) Macropterous strain larvae subjected to stress stimuli, that is, handling (intact), CO₂ anesthesia and immobilization (CO₂ + glue), buffer injection (Ringer), injection of control dsRNA (trp5), and negative control dsRNA (egfp). (C) Macropterous strain larvae injected with InR1a, InR1b, InR2, Pilp1 or Pilp2/3 dsRNA, or negative control egfp dsRNA. (D‒O) Photographs of dominant phenotypes observed after different treatments for each of the two strains, that is, wild-type adult males (D and J), individuals injected with InR1a dsRNA (E and K), InR1b dsRNA (F and L), InR2 dsRNA (G and M), Pilp1 dsRNA (H and N), and Pilp2/3 dsRNA (I and O). Numbers above each column indicate numbers of evaluated adults in each treatment. Letters inside the columns (e‒o) refer to the photographs of representative phenotypes (E‒O). For each set of experiments (A, B, and C), the obtained proportions of long- versus short-winged adult phenotypes were compared with the control treatment (first column) using an equivalent of Dunnett’s test adjusted for proportion data (Zar 1999). *, **, and *** denote significant differences at P < 0.05, P < 0.01, and P < 0.001, respectively. Scale bars for all photographs indicate 1 mm. Insets show separate views on forewings of the given phenotype.

Fig. 5.

Evolution of insect insulin receptor genes and decoy of insulin receptor genes. The tree represents InR gene duplication events and losses of the tyrosine kinase domain leading to decoy of insulin receptors, mapped on the simplified phylogenetic tree of insects according to Misof et al. (2014). The tree does not display the losses of complete InR genes, nor the duplications detected at the level of individual species, and it only shows selected taxa and lineages important for the understanding of InR evolution.

Fig. 6.

Scheme summarizing the effects of RNAi-mediated gene silencing of insulin receptor genes and inferred roles of individual paralogs in the control of wing polyphenism in the linden bug Pyrrhocoris apterus, compared with the observations by Xu et al. (2015) in the planthopper Nilaparvata lugens.