TGF-β signaling in insects regulates metamorphosis via juvenile hormone biosynthesis

Abstract

Although butterflies undergo a dramatic morphological transformation from larva to adult via a pupal stage (holometamorphosis), crickets undergo a metamorphosis from nymph to adult without formation of a pupa (hemimetamorphosis). Despite these differences, both processes are regulated by common mechanisms that involve 20-hydroxyecdysone (20E) and juvenile hormone (JH). JH regulates many aspects of insect physiology, such as development, reproduction, diapause, and metamorphosis. Consequently, strict regulation of JH levels is crucial throughout an insect's life cycle. However, it remains unclear how JH synthesis is regulated. Here, we report that in the corpora allata of the cricket, Gryllus bimaculatus, Myoglianin (Gb'Myo), a homolog of Drosophila Myoglianin/vertebrate GDF8/11, is involved in the down-regulation of JH production by suppressing the expression of a gene encoding JH acid O-methyltransferase, Gb'jhamt In contrast, JH production is up-regulated by Decapentaplegic (Gb'Dpp) and Glass-bottom boat/60A (Gb'Gbb) signaling that occurs as part of the transcriptional activation of Gb'jhamt Gb'Myo defines the nature of each developmental transition by regulating JH titer and the interactions between JH and 20E. When Gb'myo expression is suppressed, the activation of Gb'jhamt expression and secretion of 20E induce molting, thereby leading to the next instar before the last nymphal instar. Conversely, high Gb'myo expression induces metamorphosis during the last nymphal instar through the cessation of JH synthesis. Gb'myo also regulates final insect size. Because Myo/GDF8/11 and Dpp/bone morphogenetic protein (BMP)2/4-Gbb/BMP5-8 are conserved in both invertebrates and vertebrates, the present findings provide common regulatory mechanisms for endocrine control of animal development.

Keywords: GDF8/11; Gryllus bimaculatus; RNA interference; juvenile hormone; metamorphosis.

Fig. 1.

Phenotypes observed after depletion of Gb’mad and Gb’tkv was achieved with RNAi in the nymph stage of G. bimaculatus. (A and B) The effects of RNAi targeting Gb’mad or Gb’tkv in nymphs on day 1 of the third instar. In each box, the control nymph is on the left, and the RNAi-treated nymph is on the right. The instar and adult stages for each box are indicated at the bottom. The RNAi-treated nymphs remained small but underwent precocious adult metamorphosis at the seventh instar. (C and D) Body length (C) and weight (D) of male () and female () adults that developed following injections of RNAi targeting DsRed2 (as a control) or Gb’mad. The data presented are the mean ± SD. *P < 0.05 according to Student’s t test. (E) The wing pads (indicated by red asterisks) of the sixth-instar Gb’mad RNAi nymphs exhibited abnormal growth and displayed an extended side. (F) The morphology of the ovipositor (indicated by arrows) in the Gb’mad RNAi sixth-instar nymphs was smaller than that of the control nymphs (Fig. 2O and Fig. S4J). (G) Precocious adults were produced following the injection of RNAi targeting Gb’mad. The wings of these adults were significantly smaller than those of controls and were wrinkled. (H) The ovipositors of the adults produced following the injection of RNAi targeting Gb’mad were cleaved at the tip and became abnormally short. (Scale bars: 10 mm in A and B; 2 mm in E, G, and H; 1 mm in F.)

Fig. S1.

Wing pad and ovipositor development following the injection of RNAi targeting DsRed2 (control) (A–C), Gb’medea (D–F), Gb’gbb (G–I), and Gb’dpp + Gb’gbb (J–L) into sixth-instar nymphs and adult crickets as indicated. Wing pads are marked with asterisks. The wing pads of the Gb’medea- (D), Gb’gbb- (G), and Gb’dpp + Gb’gbb- (J) targeted sixth-instar nymphs exhibited abnormal growth and displayed an extended side, whereas the wings of the precocious adults (E, H, and K) were significantly reduced and wrinkled. The ovipositors of the precocious adults (F, I, and L) were cleaved at the tip and became short. (M) The male () and female () adults that developed following injections of RNAi targeting DsRed2 (as a control), Gb’dpp, or Gb’dpp + Gb’dpp-like1 + Gb’dpp-like2. (Scale bars: 2 mm in A–L; 10 mm in M.)

Fig. 2.

Phenotypes observed after depletion of Gb’myo was achieved with RNAi in G. bimaculatus. (A and B) RNAi targeting DsRed2 (control) or Gb’myo were injected into third-instar nymphs on day 1. Morphological variations during supernumerary molts (3′–3′′–fourth–4′–4′′) and during metamorphosis were subsequently observed in A and B, respectively. In A, the control nymph is on the left and the RNAi-treated nymph is on the right in each box. The instar and adult stages for each box are indicated at the bottom (male: ; female: ). (C–E) Lateral views of third- (C), fourth- (D), and fifth- (E) instar nymphs injected with RNAi targeting DsRed2 on day 1 of the third instar. The red lines indicate the contours of the wing pads (indicated by asterisks). T1–3; thorax 1–3. (F) Dorsal view of the wing pads (indicated by asterisks) in a representative sixth-instar nymph injected with RNAi targeting DsRed2 on day 1 of the third instar. (G–J) Lateral views of supernumerary 3′- (G), 3′′- (H), 4′- (I), and 4′′- (J) instar nymphs injected with RNAi targeting Gb’myo on day 1 of the third instar. (K) Dorsal view of a representative supernumerary 4′′-instar nymph injected with RNAi targeting Gb’myo on day 1 of the third instar. (L–O) Ventral views of third- (L), fourth- (M), fifth- (N), and sixth- (O) instar nymphs injected with RNAi targeting DsRed2 on day 1 of the third instar. Morphological alterations in the ovipositors (indicated by arrows) at the abdomen 8 (Abd8; indicated by arrowheads) were observed. (P–S) Ventral views of supernumerary 3′- (P), 3′′- (Q), 4′- (R), and 4′′- (S) instar nymphs injected with RNAi targeting Gb’myo on day 1 of the third instar. (T and U) Body length (T) and weight (U) of nymphs and adults treated with RNAi targeting DsRed2 (black) or Gb’myo (red). Weeks postinjection (w) are indicated on the x axis. The data presented are the mean ± SD. (Scale bars: 10 mm in A and B; 0.5 mm in C and L–O; 2 mm in F and K.)

Fig. S2.

(A) The amino acid sequence of Gb’Myo. The predicted RXXR processing site is underlined, and conserved cysteine residues are marked with asterisks. (B) The Gb’Myo TGF-β family domain amino acid sequence was aligned with orthologs in Drosophila melanogaster, Tribolium castaneum, and Bombyx mori. Residues outlined in red are common between two or more of the sequences. (C) Pairwise comparisons were made among G. bimaculatus TGF-β family members. The numbers represent the percent amino acid identity from the processing site (underlined in A) to the C terminus. Gb’Myo exhibited the highest identity with Gb’Activineβ (42%).

Fig. S3.

Phenotypes of G. bimaculatus nymphs after depleting Gb’smox and Gb’babo with RNAi and following treatment with a JH analog (JHA). (A and B) RNAi targeting Gb’smox or Gb’babo were injected into nymphs on day 1 of the third instar. Control crickets (labeled in black) are shown on the left of each box; RNAi-treated crickets (labeled in red) are shown on the right. The RNAi-treated nymphs underwent progression series of third–3′–3′′–fourth–4′–4′′–fifth or third–3′–fourth–4′–4′′–fifth (instead of third–fourth–fifth) and subsequently developed into large-sized adults compared with the nymphs that were administered DsRed2 RNAi (control). (C) Following the injection of methoprene, a JH analog, third-instar nymphs underwent a progression series of third–3′–fourth–4′–fifth and subsequently developed into large-sized adults compared with the nymphs that were injected with ethanol (control). Control crickets (labeled in black) are shown on the left of each box. Crickets treated with a JH analog (labeled in blue) are shown on the right of each box. (D and E) Body length (D) and weight (E) of the nymphs and adults treated with ethanol (control; black), methoprene (JHA; blue), or RNAi targeting Gb’smox (red) are shown. Weeks postinjection are indicated on the x axis. The data presented are the mean ± SD. (Scale bar: 10 mm.)

Fig. S4.

RNAi-mediated knockdown of Gb’myo in G. bimaculatus nymphs. (A) After RNAi targeting Gb’myo was injected into fourth-instar nymphs on day 1 (the crickets shown on the far right in each box and labeled in red), a fourth–4′–4′′–4′′′–fifth progression (instead of fourth–fifth) resulted in large-sized adults compared with nymphs injected with RNAi targeting DsRed2 (control; crickets shown on the left in each box and labeled with black text). (B) The phenotypes of fourth-instar nymphs injected with RNAi targeting Gb’myo. These phenotypes resemble those of the third-instar nymphs injected with the same RNAi. (C) The fifth-instar nymphs that received a second dose of RNAi targeting Gb’myo initiated supernumerary molts such as fifth–5′–5′′ or fifth–5′–sixth (instead of fifth–sixth–seventh–eighth–adult) and then developmentally arrested and died. (D) Phenotypes of the fifth- (Left) and sixth- (Center) instar control nymphs administered RNAi targeting DsRed2 on day 1 of the fifth instar and a 5′ nymph (Right) that received RNAi targeting Gb’myo on day 1 of the fifth instar. The nymphs that received RNAi targeting Gb’myo died after supernumerary molts of fifth–5′–5′′ or fifth–5′–sixth (instead of fifth–sixth–seventh–eighth–adult). (E–H) The effect of RNAi targeting DsRed2 (control) or Gb’myo on the wing pad (E and G) and ovipositor (arrows in F and H) that developed in fifth-instar nymphs. In the nymphs treated with RNAi targeting Gb’myo, the wing pad morphology in the resultant 5′-instar nymphs was abnormal (G), and the ovipositor was smaller than in the sixth-instar control nymphs (H). T1–3; thorax 1–3. (I–L) The effect of RNAi targeting DsRed2 (control) or Gb’myo on the wing pad (asterisks in I and K) and ovipositors (arrows in J and L) of sixth-instar nymphs. In the nymphs treated with RNAi targeting Gb’myo, wing pad morphology in the resultant 6′-instar nymphs was abnormal (K), and the ovipositor was smaller than in the control nymphs (L). (M) A qPCR analysis revealed transcript levels of Gb’myo, Gb’jhamt, and Gb’CYP15A1 on day 1 of 5′-instar nymphs injected with RNAi targeting Gb’myo. The transcript levels at day 5 of the fifth-instar nymphs injected with RNAi targeting DsRed2 (control) were set to 1. Asterisks represent significant differences between the control and the Gb’myo RNAi-treated nymphs (Student’s t test; *P < 0.05; **P < 0.005). The data presented are the mean ± SD. (N) A comparison of the different life stages and supernumerary molts induced by RNAi targeting Gb’myo at various stages of development. (Scale bars: 10 mm in A–D; 2 mm in E and I; 1 mm in F and J.)

Fig. 3.

Expression profiles of Gb’myo, Gb’dpp, Gb’jhamt, and Gb’CYP15A1 transcripts in G. bimaculatus during development and the effect of RNAi targeting Gb’myo and Gb’mad on the hemolymph titer of JH. (A) Developmental changes in JH III titer in the hemolymph of male (dotted line) and female (solid line) nymphs that were collected from the fourth to the eighth instars. (B) JH III titer measurements in the hemolymph of nymphs treated with RNAi targeting Gb’myo (red) or Gb’mad (blue) in the third instar. Asterisks represent significant differences between control and RNAi nymphs: *P < 0.05 according to Student’s t test. (C–F) Temporal expression of Gb’myo (C), Gb’dpp (D), Gb’jhamt (E), and Gb’CYP15A1 (F) as detected in qPCR analyses of nymph heads. Relative fold changes in the mRNA levels were plotted, and the average expression level in the heads on day 1 of the third instar (D1 third) was set to 1. The mRNA levels were also normalized to Gb’β-actin mRNA levels. Developmental stages were defined as days (D) after molting. Nymphs were unsexed during the third to fifth instars and were sexed during the sixth to eighth instars and the adult (ad) stage (male data: dotted lines; female data: solid lines). The data presented are the mean ± SD. (G–M) Expression levels of Gb’myo (G), Gb’baboon (H), Gb’dpp (I), Gb’tkv (J), Gb’jhamt (K), and Gb’CYP15A1 (L) in the corpus allatum–corpus cardiacum (CA–CC) complex on day 3 of the seventh instar were examined by whole-mount in situ hybridization. A control experiment using the Gb’myo sense probe is shown in M. (N and O) Expression levels of Gb’myo, Gb’dpp, Gb’jhamt, and Gb’CYP15A1 as detected in qPCR analyses of RNA samples collected from the CA (N) and CC (O) of seventh-instar nymphs. The expression level of Gb’jhamt was set to 1. The data presented are the mean ± SD.

Fig. S5.

(A) Relative levels of Gb’myo expression in the head, thorax 1, thorax 2 and 3, leg, and abdomen of third-, sixth-, and eighth-instar G. bimaculatus nymphs after normalization to Gb’β-actin mRNA. The average expression level in the heads of third-instar nymphs was set to 1. The data are presented as the mean ± SD. (B) The effect of Gb’smox depletion by RNAi on the transcript levels of Gb’smox, Gb’brk, and Gb’jhamt at day 5 of 3′ and fourth instars. The transcript levels of Gb’smox, Gb’brk, or Gb’jhamt on day 5 (D5 fourth) or day 1 (D1 fourth) of the control fourth-instar nymphs, respectively, were set to 1. The data presented are the mean ± SD. (C) A qPCR analysis revealed the transcript levels of Gb’medea, Gb’brk, and Gb’jhamt on day 1 following the injection of RNAi targeting Gb’medea into third-instar nymphs. The transcript levels at day 1 for the fourth instar (D1 fourth) nymphs that were injected with RNAi targeting DsRed2 (control) were set to 1. The data presented are the mean ± SD. (D) A qPCR analysis revealed the transcript levels of Gb’gbb, Gb’brk, and Gb’jhamt on day 1 following the injection of RNAi targeting Gb’gbb into third-instar nymphs. The transcript levels at day 1 for the fourth-instar (D1 fourth) nymphs injected with RNAi targeting DsRed2 (control) were set to 1. The data presented are the mean ± SD. (E) Relative transcript levels of Gb’dad detected in the heads of 3′- and fourth-instar nymphs on day 5 or fourth- and sixth-instar nymphs on day 1 following treatment with RNAi targeting Gb’smox (maroon) or Gb’mad (blue). The transcript levels on day 5 (D5 fourth for Gb’smox RNAi) or day 1 (D1 fourth for Gb’mad RNAi) of the fourth-instar nymphs that received RNAi targeting DsRed2 (control) were set to 1. The data presented are the mean ± SD. Asterisks in B–E represent significant differences between control and RNAi nymphs (Student’s t test; *P < 0.05; **P < 0.005; n.s., not significant). (F) Relative levels of Gb’jhamt, Gb’CYP15A1, Gb’myo, and Gb’dpp detected in brain extracts collected from seventh-instar nymphs. The expression level of Gb’jhamt was set to 1. The data presented are the mean ± SD.

Fig. 4.

The effects of RNAi-mediated depletion of Gb’myo and Gb’mad on the expression of Gb’jhamt, Gb’CYP15A1, and Gb’brk. (A–C) RNAi targeting DsRed2 control or Gb’myo were injected on day 1 of the third instar. Transcript levels of Gb’myo (A), Gb’jhamt (B), and Gb’CYP15A1 (C) were subsequently determined on days 1 and 5 in the heads of the supernumerary third-, 3′-, fourth-, and 4′-instar nymphs. The transcript levels determined on day 1 of the third-instar control nymphs (D1 third) for A–C were set to 1. The data presented are the mean ± SD. (D) Transcript levels of Gb’mad, Gb’jhamt, and Gb’CYP15A1 also were determined on day 1 of the fourth and sixth instars following the injection of RNAi targeting Gb’mad. The transcript levels of these genes in control nymphs on day 1 of the fourth instar (D1 fourth) were set to 1. The data presented are the mean ± SD. (E) Gb’brk mRNA levels in the heads of fourth- (3′) and sixth- (fourth-) instar nymphs on days 1 and 5 after the injection of RNAi targeting Gb’myo (red) and on day 1 for the fourth- and sixth-instar nymphs that received RNAi targeting Gb’mad (blue). The transcript levels of both sets of control nymphs on day 1 of the fourth instar were set to 1. The data presented are the mean ± SD. (F) After RNAi-mediated depletion of Gb’brk (green) or Gb’mad + Gb’brk (yellow) in the third instar, transcript levels of Gb’brk or Gb’jhamt were measured on days 1 and 5 of the fourth and sixth instars as indicated. The transcript levels measured on day 5 (D5 fourth) or day 1 (D1 fourth) of the control fourth-instar nymphs, respectively, were set to 1. The data presented are the mean ± SD. Asterisks in A, B, and D–F represent significant differences between the control and RNAi nymphs. n.s., not significant; *P < 0.05; **P < 0.005; ***P < 0.001 according to Student’s t test.

Fig. S6.

Rescue of the Gb’myo-RNAi phenotype by dual targeting of Gb’mad and Gb’jhamt with RNAi in the nymphal stage of G. bimaculatus. (A) The effects of RNAi targeting Gb’jhamt in nymphs on day 1 of the third instar. (Left) Adult control males and males treated with Gb’jhamt-targeted RNAi (). (Right) Control females and females treated with Gb’jhamt-targeted RNAi (). The RNAi-treated nymphs underwent precocious adult metamorphosis at the seventh instar. (B and C) The supernumerary molts caused by targeting of Gb’myo by RNAi (the middle cricket in the first three boxes of B and C) were inhibited by simultaneously injecting RNAi targeting Gb’mad and RNAi targeting Gb’jhamt (the crickets at the far right in all boxes in B and C), and adults of approximately normal body size eventually formed (far right boxes in B and C). As a control, RNAi targeting DsRed2 was injected into nymphs; the resulting crickets are shown at the left of each box in B and C. (D and E) Body length (D) and weight (E) of the nymphs and adults treated with RNAi targeting Gb’myo and RNAi targeting Gb’jhamt were compared with nymphs and adults treated with RNAi targeting DsRed2. Weeks postinjection are indicated on the x axis. The data presented are the mean ± SD. (F) A qPCR analysis of the transcript levels of Gb’myo, Gb’jhamt, and Gb’CYP15A1 detected on day 5 for the fifth instar (D5 fifth) and day 1 for the sixth instar (D1 sixth) after third-instar nymphs were injected with RNAi targeting Gb’myo and RNAi targeting Gb’jhamt. The transcript levels for the fifth-instar nymphs injected with DsRed (control) on day 5 (D5 fifth) were set to 1. The asterisks represent significant differences between the control nymphs and nymphs treated with RNAi targeting Gb’myo + Gb’jhamt (Student’s t test; *P < 0.05; **P < 0.005). The data presented are the mean ± SD. (Scale bars: 10 mm.)

Fig. 5.

Regulation of Gb’jhamt expression. (A–E) Schematic diagrams of Gb’brk and Gb’jhamt transcriptional regulation based on the results obtained from experiments targeting Gb’mad (A and E), Gb’brk (B), Gb’mad + Gb’brk (C), and Gb’smox (D) genes by RNAi. Gray denotes gene depletion and transcriptional regulatory effects by RNAi. Red arrows indicate the down- and up-regulation of target gene expression. (F) A diagram depicting the function of Dpp/Gbb (blue) and Myo (pink) signaling pathways in the regulation of jhamt expression and JH action. P indicates the phosphorylation of Mad and Smox.