Autocrine regulation of ecdysone synthesis by β3-octopamine receptor in the prothoracic gland is essential for Drosophila metamorphosis

Octβ3R Is Required for Ecdysone Biosynthesis to Execute the Larval–Prepupal Transition.

We previously reported that the GPCR Octβ3R is expressed in multiple larval tissues, including the PG (17). To determine whether Octβ3R is involved in ecdysone biosynthesis and metamorphosis, we used RNAi to knock down Octβ3R function specifically in the PG, using the Gal4-upstream activation sequence (UAS) system. Two different UAS-Octβ3R^RNAi constructs targeting distinct regions of the Octβ3R mRNA (Octβ3R^RNAi-1 and Octβ3R^RNAi-2) (Fig. S1) were used to knock down Octβ3R in the PG with the help of a phantom (phm)-22-Gal4 driver (7). Strikingly, larvae expressing Octβ3R^RNAi in the PG never developed into adult flies, and 96% of phm>Octβ3R^RNAi-1 animals (Fig. 1 A and D) and 34% of phm>Octβ3R^RNAi-2 animals (Fig. 1 B and E) arrested at the larval stage. When UAS-dicer2 was introduced into phm>Octβ3R^RNAi-2 larvae (phm>Octβ3R^RNAi-2+dicer2) to increase RNAi activity, all of these animals arrested at the larval stage (Fig. 1 C and F). Using in situ hybridization, we confirmed a significant reduction in the Octβ3R mRNA levels in the PG of knockdown animals (Fig. 1 H and L) relative to control larvae (Fig. 1 G and K). These data suggest that Octβ3R expression in the PG is essential for executing the larval–prepupal transition in metamorphosis.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

PG-specific Octβ3R knockdown causes arrest at the larval–prepupal transition due to reduced levels of ecdysone. (A–F) Developmental profiles for control [UAS-Octβ3R^RNAi-1 (A), UAS-Octβ3R^RNAi-2 (B), and phm > dicer2 (C)] and PG-specific Octβ3R knockdown animals [phm > Octβ3R^RNAi-1 (D), phm > Octβ3R^RNAi-2 (E), and phm > Octβ3R^RNAi-2+dicer2 (F)]. Color bars represent the proportions of larvae (blue), prepupae/pupae (red), adults (green), dead larvae (dark gray), and dead prepupae/pupae (light gray), expressed as percentages, at representative stages [hours after hatching (hAH)]. n, the number of animals examined. (G–N) Section in situ hybridization of the PG (G, H, K, and L) and the brain (I, J, M, and N) of control [UAS-Octβ3R^RNAi-1 (G and I), UAS-Octβ3R^RNAi-2 (K and M)] and PG-specific Octβ3R knockdown larvae [phm > Octβ3R^RNAi-1 (H and J), and phm > Octβ3R^RNAi-2 (L and N)] at 72 hAH, using an antisense RNA probe for Octβ3R. PGs and brains are outlined by dashed lines. (Scale bars, 50 μm.) Octβ3R signal was reduced in the PG of Octβ3R knockdown larvae (H and L), whereas it was unaffected in the brain (J and N). (O and P) Percentages of larvae that developed to prepupae/pupae at the representative stages. PG-specific Octβ3R knockdown larvae [phm > Octβ3R^RNAi-1 (O) and phm > Octβ3R^RNAi-2+dicer2 (P)] were cultured in media supplemented with 1 mg/mL 20E (red) or without 20E (blue) from 48 hAH. n, the number of animals examined. Significance was calculated using the χ² test (**P < 0.01; ***P < 0.001).

Because a similar defect in the larval–prepupal transition occurs in ecdysone-deficient larvae (12), we hypothesized that the Octβ3R knockdown phenotype was due to lack of ecdysone production. Consistent with this idea, the 20E titer was much lower in phm>Octβ3R^RNAi-1 larvae than in control larvae just before the larval–prepupal transition (90 hAH) (Fig. S2). Moreover, administration of 20E by feeding rescued the defect in the larval–prepupal transition caused by Octβ3R knockdown. When phm>Octβ3R^RNAi-1 and phm>Octβ3R^RNAi-2+dicer2 larvae were cultured on media containing 20E (1 mg/mL) from 48 hAH onward, approximately half of them developed to the prepupal stage, compared with only 2–3% of larvae not fed 20E (Fig. 1 O and P). Thus, PG-specific loss of Octβ3R activity causes an arrest in the larval–prepupal transition due to lack of ecdysone.

Octβ3R Function Is Required for Proper Expression of Ecdysone Biosynthetic Genes.

Ecdysone is synthesized in the PG from dietary cholesterol through the action of seven ecdysone biosynthetic genes (neverland, spookier, shroud, Cyp6t3, phantom, disembodied, and shadow) (5, 18⇓–20). We therefore performed quantitative RT-PCR (qPCR) to investigate whether loss of Octβ3R function affects expression of these genes in the PG. In control larvae, expression of these genes increased dramatically between 72 and 96 hAH, when the larval–prepupal transition occurs (Fig. 2 A–G and Fig. S3 A–G). By contrast, in phm>Octβ3R^RNAi-1 and phm>Octβ3R^RNAi-2+dicer2 larvae, the expression of all of these genes was significantly reduced relative to control larvae at 96 hAH (Fig. 2 A–G and Fig. S3 A–G). We confirmed the reduced expression of ecdysone biosynthetic genes in the PG by in situ hybridization (Fig. 2 H′–N′ and Fig. S3 H′–N′). Furthermore, immunostaining revealed that Neverland, Shroud, Phantom, Disembodied, and Shadow protein levels were reduced in the PG of phm>Octβ3R^RNAi-1 and phm>Octβ3R^RNAi-2+dicer2 larvae (Fig. 3 O′–S′ and Fig. S3 O′–S′). Taken together, these data show that Octβ3R function is required in the PG for proper expression of ecdysone biosynthetic genes.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Expression of ecdysone biosynthetic genes is significantly reduced by Octβ3R knockdown. (A–G) Whole-body expression of ecdysone biosynthetic genes in control [UAS-Octβ3R^RNAi-1 (red)] and PG-specific Octβ3R knockdown larvae [phm > Octβ3R^RNAi-1 (blue)] was measured using qPCR. Relative expression levels in gene expression were calculated at the representative stages, and the average values of triplicate data sets are shown with SEs. Significance was calculated using Student t test (*P < 0.05; **P < 0.01; ***P < 0.001). (H–N and H′–N′) Expression of ecdysone biosynthetic genes in the PG of control [UAS-Octβ3R^RNAi-1 (H–N)] and PG-specific Octβ3R knockdown larvae [phm > Octβ3R^RNAi-1 (H′–N′)] at 96 hAH. Whole-mount in situ hybridization was performed using antisense probes for the neverland (H and H′), spookier (I and I′), shroud (J and J′), Cyp6t3 (K and K′), phantom (L and L′), disembodied (M and M′), and shadow transcripts (N and N′). PGs are outlined by dashed lines. (Scale bars, 50 μm.) (O–S and O′–S′) Expression of ecdysone biosynthetic enzymes in the PG of control [UAS-Octβ3R^RNAi-1 (O–S)] and PG-specific Octβ3R knockdown larvae [phm > Octβ3R^RNAi-1 (O′–S′)] at 96 hAH. Immunostaining was performed using antibodies against Neverland (O and O′), Shroud (P and P′), Phantom (Q and Q′), Disembodied (R and R′), and Shadow proteins (S and S′). PGs are outlined by dashed lines. (Scale bars, 50 μm.)

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

PG-specific Tdc2 knockdown causes defects in the larval–prepupal transition due to reduced levels of ecdysone. (A–F) Expression of the Tdc1, Tdc2, and Tbh genes in the PG at 72 hAH. Whole-mount in situ hybridization with antisense (A–C) and sense RNA probes (D–F) for Tdc1 (A and D), Tdc2 (B and E), and Tbh transcripts (C and F) were performed. The stage of the larvae (Oregon-R) from which the PG was obtained and the concentration (400 ng/mL) of the probes used for hybridization were the same for the antisense and sense experiments. PGs are outlined by dashed lines. (Scale bars, 50 μm.) (G–L) The distribution of tyramine and octopamine in the PG of control [phm > dicer2 (G and J)] and PG-specific Tdc2 knockdown larvae [phm > Tdc2^RNAi-1+dicer2 (H and K) and phm > Tdc2^RNAi-2+dicer2 (I and L)] at 72 hAH. Immunostaining was performed with antibodies against tyramine (G–I) and octopamine (J–L). PGs are outlined by dashed lines. (Scale bars, 50 μm.) (M–O) Developmental profiles for control [phm > dicer2 (M)] and PG-specific Tdc2 knockdown animals [phm > Tdc2^RNAi-1+dicer2 (N) and phm > Tdc2^RNAi-2+dicer2 (O)]. Color bars are described in the legend for Fig. 1 A–F. n, the number of animals examined. (P) Expression of Tdc2 in the RG of control (phm > dicer2) and PG-specific Tdc2 knockdown larvae (phm > Tdc2^RNAi-1+dicer2 and phm > Tdc2^RNAi-2+dicer2) were measured at 72 hAH using qPCR. Relative expression level in gene expression was calculated, and the average values of triplicate data sets and SEs are shown. Significance was calculated using the Student t test (*P < 0.05). (Q and R) Percentages of larvae that developed into prepupae/pupae at representative stages. PG-specific Tdc2 knockdown larvae [phm > Tdc2^RNAi-1+dicer2 (Q) and phm > Tdc2^RNAi-2+dicer2 (R)] were cultured in media supplemented with 1 mg/mL 20E (red) or without 20E (blue) from 48 hAH onward. n, the number of animals examined. Significance was calculated using the χ² test (*P < 0.05; ***P < 0.001).

Tyramine Produced in the PG Regulates Ecdysone Biosynthesis.

Octβ3R is thought to be activated by octopamine and tyramine binding (21). Octopamine is synthesized from tyramine by tyramine β-hydroxylase (Tbh) (22), and tyramine is synthesized from tyrosine by tyrosine decarboxylase (Tdc) (22). In Drosophila, two Tdc genes (Tdc1 and Tdc2) and one Tbh gene have been identified (23, 24), and all of them are expressed in the larval CNS (23, 24). We found that Tdc1, Tdc2, and Tbh were also expressed in the PG (Fig. 3 A–F). Furthermore, we detected octopamine and tyramine in the PG by immunostaining (Fig. 3 G and J and Fig. S4 A and C). Thus, octopamine and/or tyramine synthesized in the PG may activate Octβ3R in an autocrine manner to induce ecdysone production.

To test this, we generated PG-specific knockdowns of Tdc1, Tdc2, and Tbh. To knock down Tdc2, two constructs targeting distinct regions of the Tdc2 transcript (Tdc2^RNAi-1 and Tdc2^RNAi-2) (Fig. S1) were expressed along with dicer2 in the PG under the control of the phm-22-Gal4 driver (phm > Tdc2^RNAi-1+dicer2 and phm > Tdc2^RNAi-2+dicer2). All phm > Tdc2^RNAi-1+dicer2 larvae arrested at the larval stage (Fig. 3N), and phm > Tdc2^RNAi-2+dicer2 larvae were significantly delayed at the larval–prepupal transition, relative to control animals (Fig. 3 M and O). Tdc2 mRNA level was reduced in the ring gland (RG) containing the PG in both sets of knockdown animals, as demonstrated by qPCR (Fig. 3P). Moreover, octopamine and tyramine production in the PG was impaired by Tdc2 knockdown (Fig. 3 G–L). By contrast, Tdc1 knockdown (phm > Tdc1^RNAi+dicer2) caused only a subtle delay in the larval–prepupal transition (Fig. S4 J and G) and had no detectable effect on octopamine or tyramine production (Fig. S4 B and D). These results suggest that Tdc2 is the predominant Tdc regulating octopamine and tyramine biosynthesis in the PG and the larval–prepupal transition. Contrary to our findings, a null mutation in Tdc2 does not affect metamorphosis, and these mutant flies are viable (23). Thus, PG-specific knockdown causes a stronger phenotype than complete loss of Tdc2 activity in whole animals. A similar situation has been reported in regulation of metamorphosis by Activin signaling (12). These phenomena can be explained by a model in which some compensatory changes in other mutant tissues rescue the PG-specific knockdown phenotype in null-mutant animals.

PG-specific Tdc2 knockdown caused a reduction in larval 20E concentration (Fig. S2B). Therefore, we next examined whether feeding 20E to Tdc2 knockdown larvae would rescue the larval–prepupal transition defect. To this end, we cultured phm > Tdc2^RNAi-1+dicer2 and phm > Tdc2^RNAi-2+dicer2 larvae in media with or without 20E (1 mg/mL) from 48 hAH onward. Approximately 40% of the 20E-fed phm > Tdc2^RNAi-1+dicer2 larvae developed to the prepupal stage, whereas none of those larvae grown on control media progressed beyond the larval stage (Fig. 3Q). Furthermore, the delay in the larval–prepupal transition in phm > Tdc2^RNAi-2+dicer2 larvae was rescued by 20E feeding (Fig. 3R). These results indicate that the defect in the larval–prepupal transition in Tdc2 knockdown animals results from a lack of 20E production. Thus, octopamine/tyramine synthesized in the PG appears to activate Octβ3R in an autocrine manner to execute the larval–prepupal transition by regulating ecdysone production.

To determine which Octβ3R ligand is responsible for this autocrine signaling, we knocked down Tbh in the PG to prevent conversion of tyramine into octopamine. To knock down Tbh, two constructs targeting distinct regions of the Tbh transcript (Tbh^RNAi-1 and Tbh^RNAi-2) (Fig. S1) were expressed along with dicer2 under the control of phm-22-Gal4 (phm > Tbh^RNAi-1+dicer2 and phm > Tbh^RNAi-2+dicer2) (Fig. S4 K and L). Although the Tbh knockdown caused a reduction in octopamine production in the PG (Fig. S4 E and F), these larvae did not exhibit any obvious defects in the larval–prepupal transition or subsequent metamorphosis (Fig. S4 H and I). These data suggest that tyramine, rather than octopamine, is the Octβ3R ligand that activates ecdysone production in the PG.

Octβ3R Is Required for Activation of Ilps and PTTH Signaling.

Because ecdysone biosynthesis in the PG is under the control of Ilps and PTTH signaling (2⇓–4), we next examined whether Octβ3R function is required to activate these signaling pathways. To detect Ilps signaling activity, we used a pleckstrin-homology domain fused to GFP (PH-GFP), which is recruited to the plasma membrane when insulin signaling is activated (25). In the PG cells of control larvae, PH-GFP was only weakly localized to the plasma membrane at 48 hAH, whereas its membrane localization became increasingly evident at 60, 84, and 90 hAH (Fig. 4 A–D). By contrast, in PG cells of phm>Octβ3R^RNAi-1 larvae, the tight localization of PH-GFP to the plasma membrane was no longer detectable (Fig. 4 A′–D′), indicating that activation of Ilps signaling had been disrupted. Moreover, overexpression of a constitutively active form of the Ilps receptor InR (InR^CA) was able to rescue the larval arrest in phm>Octβ3R^RNAi-1 animals (Fig. 4I). Next, we performed immunostaining of the diphosphorylated form of ERK (dpERK), a downstream signaling component of the PTTH pathway. We found that dpERK expression was very weak at 48 hAH, but was activated in the PG of control larvae at 60, 84, and 90 hAH (Fig. 4 E–H); by contrast, this activation was reduced in the PG of phm>Octβ3R^RNAi-1 larvae (Fig. 4 E′–H′). Expression of a constitutively active form of another downstream PTTH signaling component, Ras (Ras^V12), rescued the larval–prepupal transition defect in phm>Octβ3R^RNAi-1 animals (Fig. 4I). These results show that Octβ3R function is required to activate Ilps and PTTH signaling in the PG and that these signaling pathways execute the larval–prepupal transition. Although activation of both the Ilps and PTTH signaling pathways requires Activin/TGFβ signaling in the PG (12), expression of a constitutively active form of the Activin/TGFβ receptor Baboon (Babo^CA) failed to rescue the larval–prepupal transition defect in phm>Octβ3R^RNAi-1 animals (Fig. 4I). This observation suggests that Octβ3R acts downstream or independent of Activin/TGFβ signaling to regulate Ilps and PTTH signaling in the PG.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

The Ilps and PTTH signaling pathways are disrupted by Octβ3R knockdown in the PG. (A–H and A′–H′) Expression of PH-GFP (A–D and A′–D′) and dpERK (E–H and E′–H′) in the PG of control [phm-22-Gal4+PH-GFP (A–D), phm-22-Gal4 (E–H)] and PG-specific Octβ3R knockdown [phm > Octβ3R^RNAi-1+PH-GFP (A′–D′), phm > Octβ3R^RNAi-1 (E′–H′)] larvae at 48, 60, 84, and 90 hAH. Immunostaining was performed using antibodies against GFP (A–D and A′–D′) and dpERK (E–H and E′–H′). PGs are outlined by dashed lines. (Scale bars, 50 μm.) (I) Percentages of larvae that developed into prepupae/pupae at representative stages. PG-specific Octβ3R knockdown larvae expressing GFP [phm > Octβ3R^RNAi-1+GFP (red)], InR^CA [phm > Octβ3R^RNAi-1+InR^CA (blue)], Ras^V12 [phm > Octβ3R^RNAi-1+Ras^V12 (green)], and Babo^CA [phm > Octβ3R^RNAi-1+Babo^CA (yellow)] were cultured from 48 hAH, and the number of prepupae/pupae was counted at each stage. n, the number of animals examined. Significance was calculated using the χ² test (*P < 0.05; **P < 0.01; ***P < 0.001).

Octβ3R Signaling Is Required at Around 60 hAH in PG for Execution of the Larval–Prepupal Transition.

The observations described above demonstrate that phm>Octβ3R^RNAi affects Ilps and PTTH signaling in the PG as early as 60 hAH, raising the question of when Octβ3R function is required in the PG for execution of the larval–prepupal transition. To address this issue, we used the Gal80^ts and Gal4/UAS system (26), which restricts expression of Octβ3R dsRNA in the PG at 18 °C, but allows its expression at 28 °C. The results of temperature upshift and downshift experiments revealed that the larval–prepupal transition was impaired only when Octβ3R dsRNA was expressed in the PG at around 60 hAH (Fig. S5). Notably, 60 hAH is the critical period during which larvae attain CW under nutrient-rich conditions (2, 13, 14). As noted above, when larvae are starved before attainment of CW, they are unable to transit into the prepupal stage (2, 13, 14). By contrast, starved larvae can successfully transit to prepupal/pupal stage without developmental delay once they have attained CW by growing beyond the critical period (∼56 hAH) under nutrient-rich conditions in standard Drosophila medium (13, 14). Thus, we hypothesized that Octβ3R signaling acts downstream of the body-size checkpoint, or attainment of CW, to allow the larval–prepupal transition.

Several lines of evidence support our hypothesis. First, Octβ3R function is required for activation of Ilps and PTTH signaling detected in the PG at 60 hAH. By contrast, at 48 hAH, before the attainment of CW, neither signaling pathway is active in the PG (Fig. 4). Second, Ilps and PTTH signaling was not activated in the PG when the larvae were starved from 48 hAH onward (early starvation), whereas these signaling pathways were active when the larvae were starved after 60 hAH (late starvation) (Fig. S6). Finally, a ligand for Octβ3R, tyramine, was detectable in the PG at 60 hAH, but decreases after this stage under a nutrient-rich condition (Fig. S7 A and B). This decrease in tyramine was abrogated by early starvation but not by late starvation (Fig. S7 C–E). Assuming that this decrease in tyramine in the PG is due to its secretion from PG cells, it is reasonable to propose that attainment of CW causes tyramine secretion from the PG at around 60 hAH, which in turn activates Octβ3R to regulate the Ilps and PTTH pathways, leading to the larval–prepupal transition (Fig. S8).

This study demonstrates, for the first time to our knowledge, that monoaminergic regulation plays a pivotal role in ecdysone biosynthesis to induce metamorphosis and that Octβ3R acts as an upstream regulator essential for the Ilps and PTTH signaling. In addition, our data indicate that Octβ3R ligands are produced in the PG to stimulate ecdysone biosynthesis in an autocrine manner. Autocrine signaling has been proposed to mediate the community effect, in which identical neighboring cells are coordinated in their stimulation and maintenance of cell type-specific gene expression and their differentiation, as observed in muscle development of amphibian embryos (27, 28). Thus, we propose that monoaminergic autocrine signaling among PG cells acts to increase their responsiveness to Ilps and PTTH, thereby allowing coordinated expression of ecdysone biosynthetic genes within a time window following exposure to neuropeptides.

Our findings raise the larger question of whether monoamine acts as part of an evolutionarily conserved mechanism of steroid hormone production. In vertebrates, there is limited evidence of monoaminergic regulation of steroid hormone biosynthesis. For example, in cultured adrenal glands, catecholamine stimulates the biosynthesis of the steroid hormone cortisol in a paracrine manner to elicit a stress reaction (29). Another example is the Leydig cells of the mammalian testes, in which the steroid hormone testosterone is produced mainly in response to pituitary gonadotropin. However, catecholamine signaling through β-adrenergic receptors, the orthologs of Octβ3R, also promotes the production of testosterone from cultured fetal Leydig cells (30, 31), which may be the site of catecholamine synthesis in the fetal and mature human testes (32). Thus, monoamines may play a conserved role in modulating and/or stimulating steroid hormone production during physiological and developmental transitions.