The Insect Neuropeptide PTTH Activates Receptor Tyrosine Kinase Torso to Initiate Metamorphosis

See allHide authors and affiliations

Science  04 Dec 2009:
Vol. 326, Issue 5958, pp. 1403-1405
DOI: 10.1126/science.1176450


Holometabolous insects undergo complete metamorphosis to become sexually mature adults. Metamorphosis is initiated by brain-derived prothoracicotropic hormone (PTTH), which stimulates the production of the molting hormone ecdysone via an incompletely defined signaling pathway. Here we demonstrate that Torso, a receptor tyrosine kinase that regulates embryonic terminal cell fate in Drosophila, is the PTTH receptor. Trunk, the embryonic Torso ligand, is related to PTTH, and ectopic expression of PTTH in the embryo partially rescues trunk mutants. In larvae, torso is expressed specifically in the prothoracic gland (PG), and its loss phenocopies the removal of PTTH. The activation of Torso by PTTH stimulates extracellular signal–regulated kinase (ERK) phosphorylation, and the loss of ERK in the PG phenocopies the loss of PTTH and Torso. We conclude that PTTH initiates metamorphosis by activation of the Torso/ERK pathway.

Many organisms undergo distinct temporal transitions in morphology as a part of their normal life process. In humans, for example, passage through puberty is accompanied by changes in body mass and the acquisition of sexual maturity. Likewise, in all holometabolous insects, metamorphosis transforms the immature larva into a completely new body form that is capable of reproductive activity. In both cases, neuropeptide signaling in response to environmental and nutritional cues triggers the transition process (15). In insects, the process is initiated by the neuropeptide known as prothoracicotropic hormone (PTTH) (3, 6, 7). PTTH signals to the prothoracic gland (PG), the primary insect endocrine organ, which triggers the production and release of ecdysone, the precursor of the active steroid molting hormone 20-hydroxyecdysone (20E) (8). The increased level of 20E provides a systemic signal that ends the larval growth period and initiates metamorphosis.

PTTH has been proposed to be structurally similar to certain mammalian growth factors (9) that are ligands for receptor tyrosine kinases (RTKs). Previous studies have also indicated that PTTH signaling results in the phosphorylation of cellular signaling molecules that are linked to the mitogen-activated protein kinase (MAPK) pathway in the PG (1013). In light of the potential involvement of MAPK pathway components in PTTH signaling, we examined the expression of all Drosophila RTKs in the PG to determine whether any showed a tissue-specific expression profile that was consistent with a possible role as a PTTH receptor. We found that after early embryogenesis, the RTK encoded by torso is expressed specifically in the PG (Fig. 1, A and B).

Fig. 1

Torso is expressed in the PG, and its embryonic ligand Trk is structurally related to PTTH. (A) RNA in situ hybridization of torso in a stage-17 Drosophila embryo (dorsal view with anterior to the left). (B) RNA in situ hybridization of torso in the PG and brain (Br) from a wandering L3 larva of Drosophila. (C) Sequence similarity between Drosophila PTTH and Trk. Processing of the precursors is believed to release the mature C-terminal peptides that form cysteine knot-type structures. Red vertical lines indicate cysteines and numbers refer to amino acid residues.

The gene torso belongs to the so-called terminal group of genes that are required for the correct patterning of anterior and posterior structures during early embryogenesis (1416). The presumed ligand for Torso during terminal patterning is Trunk (Trk), which contains a cysteine knot-type motif in the C-terminal region similar to the motif in PTTH (3, 15). Also like PTTH, Trk is thought to be proteolytically processed from a precursor molecule to generate an active C-terminal fragment that is comparable in length to that of PTTH (17). Alignment of the protein sequences of Trk and PTTH reveals that they share some conserved structures in the C-terminal region that compose the mature peptide, including all of the six cysteines that are important for intramonomeric bonds of the PTTH homodimeric molecule (7) (Fig. 1C). Previously, it was noted that Trk is related to Spatzle (17), but a phylogenetic analysis of different insect cysteine knot-type proteins shows that Trk and PTTH form a separate cluster and that PTTH is the closest paralog of Trk (fig. S1). These results raise the possibility that Trk and PTTH share a conserved three-dimensional structure enabling both to activate Torso despite the modest conservation of primary sequence. We did not detect expression of trk in the wandering third-instar larval (L3) stage using real-time polymerase chain reaction (PCR) (no product after 30 PCR cycles) or by in situ hybridization to the brain-PG complex (fig. S2), supporting the idea that PTTH, and not Trk, is a ligand for Torso in post-blastoderm stages.

To investigate possible post-embryonic roles of Torso in Drosophila development, we used RNA interference (RNAi) to knock down torso specifically in the PG. The PG-specific phantom (phm)-Gal4 line (phm>) was used to drive expression of RNAi constructs under control of upstream activator sequences (UASs) in the PG. As shown in Fig. 2A, this expression of a torso RNAi construct produced a phenotype that was almost identical to the one created by the loss of PTTH-expressing neurons (3). Reduction of torso expression in the PG of phm>torso-RNAi larvae delays the onset of pupariation by 5.8 days as compared with the phm> + control animals, similar to the 5.4-day delay of pupariation in animals lacking PTTH. As with the loss of the PTTH-producing neurons, torso silencing in the PG also leads to excessive growth during the prolonged L3 stage, resulting in increased pupal size (Fig. 2B). To test the specificity of the RNAi, we confirmed that torso mRNA levels are reduced in phm>torso-RNAi larvae and that the PG cells are morphologically normal, although slightly smaller (fig. S3).

Fig. 2

Disruption of Torso signaling phenocopies loss of PTTH-producing neurons. The ptth-Gal4 line (ptth>) expresses Gal4 specifically in the PTTH-producing neurons (3), and the phm> expresses Gal4 specifically in the PG (20). (A) Compared with the control (phm>+), reducing Torso (phm>torso-RNAi), Ras (phm>ras85D-RNAi), Raf (phm>Draf-RNAi), and ERK (phm>ERK-RNAi) signaling in the PG prolongs the duration of L3, similar to the removal of PTTH [ptth>grim: expression of the pro-apoptotic gene grim under control of the ptth promoter genetically ablates the PTTH-producing neurons (3)] in Drosophila. Expression of the constitutively active rasV12 rescues the developmental delay of torso silencing in the PG (phm>rasV12, torso-RNAi). Feeding with 20E rescues the delay of both torso and ERK silencing in the PG. Error bars represent SD. A model of PTTH signaling (right) is proposed based on the Torso pathway components in the embryo (15). Multiple arrowheads indicate intermediate steps in the pathway. (B) Effect on pupal size of torso silencing in the PG and 20E feeding. (C) Effect on pupal size of ERK silencing in the PG and 20E feeding. (D) RasV12 expression in the PG rescues the overgrowth phenotype of torso silencing. E, embryogenesis; L1, first larval instar; L2, second larval instar.

Because torso is a maternal-effect gene, homozygous mutants derived from heterozygous parents are viable. Therefore, we examined the developmental profile and adult size of animals homozygous and transheterozygous for three different torso mutations. Larvae with mutations in torso exhibited substantial developmental delays, although not as long as those seen by RNAi knockdown, in the time to pupariation as compared with heterozygous controls (fig. S2), and the mutants produced larger adults (table S1). The difference in time delay may result from residual maternally loaded torso mRNA. In contrast, trk mutants developed on a normal time scale (fig. S2), and adults were similar in size to heterozygous control adults (table S1), demonstrating that the phenotype of torso mutants is independent of early embryonic signaling.

In animals lacking PTTH-producing neurons, it is the low level of the active molting hormone 20E that causes the developmental delay and tissue overgrowth (3). To investigate whether the torso loss-of-function phenotype is also caused by low 20E levels, we fed 20E to phm>torso-RNAi larvae. Similar to what was found when the PTTH-producing neurons were removed, feeding these larvae with 20E completely rescued the developmental delay and overgrowth (Fig. 2, A and B). Taken together, these results demonstrate that reducing Torso signaling in the PG alone phenocopies the loss of PTTH, which is consistent with the notion that Torso mediates PTTH signaling in the PG. If this is the case, we expect that the constitutively active torsoRL3 allele (14) might produce precocious pupation, as would overexpression of PTTH. Consistent with this conjecture, we found, using the daughterless (da)-Gal4 driver (da>), that ubiquitous overexpression of PTTH advances the onset of pupariation by 11.5 hours as compared with (da> +) balancer controls (P < 0.001) and produces smaller adults, as shown previously (3). At 25°C, torsoRL3 is activated, and heterozygous torsoRL3/+ animals pupariate 9.2 hours before controls (P < 0.001) and form smaller adult males (table S1).

To establish whether PTTH can activate Torso in vivo, we reasoned that if PTTH is a ligand for Torso, then ectopic expression of PTTH in the embryo might elicit partial rescue of trk mutants. To examine this, we used the maternal nanos (nos)-Gal4 line (nos>) to drive ubiquitous early embryonic expression of a UAS-PTTH-hemagglutinin (HA)–tagged transgene in trk mutant embryos. In the blastoderm-stage embryo, activation of Torso by Trk induces expression of the downstream target gene tailless (tll) in the anterior and posterior regions (15, 17). The inability to activate this target gene in trk (Fig. 3) or torso (15) mutants leads to the loss of structures posterior to the seventh abdominal segment. Early embryonic expression of PTTH was observed in 13% (n = 70 embryos) of blastoderm-stage embryos derived from trk1/trk1; nos>PTTH females (Fig. 3). Ectopic expression of PTTH in these embryos was sufficient to activate tll in the posterior part of the embryos (11% of embryos showing posterior tll expression; n = 56). Although PTTH expression did not fully restore wild-type tll expression, the partial rescue elicited by PTTH was sufficient to restore posterior structures, such as the Filzkörper, in several trk mutant embryos. These results provide genetic evidence that PTTH functions as a ligand for Torso in vivo.

Fig. 3

PTTH is a ligand for Torso in vivo. PTTH partially rescues the terminal phenotype of Drosophila trk1 mutants. RNA in situ hybridization shows that in wild-type (wt) embryos, trk is ubiquitously expressed. The nos> allows the uniform expression of PTTH in the embryo, which is confirmed by immunolocalization of HA-tagged PTTH in embryos from trk1/trk1;nos>PTTH females. In trk1 mutants, tll expression is restricted to the anterior domain because of the lack of Trk-dependent activation of Torso. Ectopic expression of PTTH in trk1/trk1;nos>PTTH embryos partially restores posterior expression of tll. Cuticle structures posterior to the seventh abdominal denticle belt are missing in trk1 mutants, but expression of PTTH in these embryos can rescue the morphology of denticle belt structures (yellow arrowheads) and induce formation of posterior structures such as the Filzkörper (white arrowheads). Shown are lateral views of embryos with anterior to the left. DAPI, 4′,6′-diamidino-2-phenylindole.

In the embryo, Torso signaling is transduced through the canonical MAPK pathway that includes the Drosophila homologs of Ras (Ras85D), Raf (Draf), MAPK kinase (MEK), and extracellular signal–regulated kinase (ERK) (15). If Torso is indeed the PTTH receptor, we expect that disrupting MAPK signaling in the PG would result in a phenotype similar to that resulting from loss of the PTTH-producing neurons and Torso signaling. So far, the role of the MAPK pathway in transduction of the PTTH signal has been determined only by in vitro studies of lepidopteran PG (1013). In Drosophila, the expression of dominant negative forms of Ras and Raf is known to delay development (18). To further examine the importance of the MAPK pathway in mediating PTTH/Torso signaling, we used RNAi to reduce the expression of several core components of this pathway, including Ras, Raf, and ERK, in the PG. Loss of either Ras, Raf, or ERK delayed pupariation by 4.3, 2.7, and 6.1 days, respectively (Fig. 2A). ERK silencing in the PG delays pupariation as severely as the reduction of Torso signaling or the complete loss of the PTTH-producing neurons does. The increase in size of phm>ERK-RNAi pupae and adults (Fig. 2C and table S1) was also similar to the increase caused by the loss of PTTH or loss of Torso. The developmental delay, as well as the size increase caused by ERK silencing, were negated by 20E feeding (Fig. 2, A and C). The less-severe phenotypes produced by the loss of Raf and Ras may result from less-efficient knockdown or, in the case of Ras, may reflect partial redundancy with Rap1 (19). Consistent with Ras being downstream of torso, we also found that expression of constitutively active Ras in the PG completely rescued the torso-RNAi–induced delay and overgrowth phenotype (Fig. 2, A and D). Taken together, these results indicate that, as during embryonic terminal patterning, Torso regulation of ecdysone production in the PG is primarily mediated by the MAPK pathway, resulting in the activation of ERK.

To test directly whether stimulation of Torso by PTTH could lead to ERK phosphorylation, we sought to develop a cell culture–based signaling assay. Because we have not been able to produce active Drosophila PTTH in tissue culture, we turned to the silkworm Bombyx mori (7) and cloned a full-length Bombyx torso cDNA (fig. S4). As in Drosophila, the Bombyx torso ortholog is expressed predominantly in the PG of the final (fifth)–instar larvae (figs. S4 and S5). Stimulation of Drosophila S2 cells transfected with Bombyx torso and Drosophila ERK with 10−9 M PTTH led to robust phosphorylation of ERK (Fig. 4). PTTH stimulation of ERK phosphorylation was not detected in control S2 cells, either incubated in the absence of PTTH or those stimulated with PTTH but not expressing Bombyx torso. Bombyx PTTH did not stimulate activation of ERK through Drosophila Torso or through the insulin receptor (fig. S6), demonstrating that ERK stimulation by Bombyx PTTH is specific to Bombyx Torso. These results demonstrate that Torso is a functional PTTH receptor that is able to mediate PTTH signaling through the activation of the ERK pathway.

Fig. 4

Characterization of Bombyx Torso as a functional PTTH receptor in vitro. Quantified levels of phosphorylated ERK (top) in the Bombyx PG and Drosophila S2 cell extract determined by immunoblotting (bottom) with an antibody that is specific to the phosphorylated form of ERK (p-ERK) are shown. Bombyx PGs from day 4 fifth-instar larvae were incubated for 15 min in medium (control) or medium containing 1 nM Bombyx PTTH. Drosophila S2 cells overexpressing ERK-Myc and Bombyx torso were incubated in medium (control) or medium containing 1 nM Bombyx PTTH for 15 and 25 min. As a control to show that Bombyx torso is responsible for the activation by PTTH that induces ERK phosphorylation, S2 cells transfected with ERK-Myc, but not with torso, were subjected to the same treatment. Phospho-ERK levels were normalized against total ERK (ERK-Myc). As a loading control, blots were also probed with an antibody to tubulin. Fold changes are relative to control. Error bars represent SD. *P < 0.05, **P < 0.01.

Our observations define another role for the terminal system, which is the initiation of metamorphosis at the end of larval growth. Therefore, insects apparently use the same core system for two developmentally distinct processes: the establishment of terminal cell fate in the embryo and the termination of larval growth at the correct time to ensure an appropriate final adult body size. Our identification of the PTTH receptor will facilitate further characterization of the system that determines body size in insects. It will be of interest to ascertain just how similar this system is in overall design to the hypothalamus-pituitary-gonadal axis, which controls the timing of puberty in mammals (5).

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S6

Table S1


References and Notes

  1. K.F.R. was supported by a postdoctoral fellowship from the Danish Natural Science Research Council (grant no. 272-07-0340). M.B.O. is an investigator with the Howard Hughes Medical Institute. We are grateful to our collaborators J. T. Warren and R. Rybczynski and to reviewers for helpful comments. We also thank H. Kataoka for Bombyx PTTH. The GenBank accession number for Bombyx Torso is GQ477743.

Stay Connected to Science

Navigate This Article