Two Subtypes of Ecdysis-triggering Hormone Receptor inDrosophila melanogaster *

Insect ecdysis is a hormonally programmed physiological sequence that enables insects to escape their old cuticle at the end of each developmental stage. The immediate events leading to ecdysis, which are initiated upon release of ecdysis-triggering hormones (ETH) into the bloodstream, include respiratory inflation and sequential stereotypic behaviors that facilitate shedding of the cuticle. Here we report that the Drosophila geneCG5911 encodes two functionally distinct subtypes of G protein-coupled receptors through alternative splicing (CG5911a and CG5911b) that respond preferentially to ecdysis-triggering hormones of flies and moths. These subtypes show differences in ligand sensitivity and specificity, suggesting that they may play separate roles in ETH signaling. At significantly higher concentrations (>100-fold), certain insect and vertebrate peptides also activate these receptors, providing evidence that CG5911 is evolutionarily related to the thyrotropin-releasing hormone and neuromedin U receptors. The ETH signaling system in insects is a vital system that provides opportunities for the construction of models for the molecular basis of stereotypic animal behavior as well as a target for the design of more sophisticated insect-selective pest control strategies.

Insect ecdysis is a hormonally programmed physiological sequence that enables insects to escape their old cuticle at the end of each developmental stage. The immediate events leading to ecdysis, which are initiated upon release of ecdysis-triggering hormones (ETH) into the bloodstream, include respiratory inflation and sequential stereotypic behaviors that facilitate shedding of the cuticle. Here we report that the Drosophila gene CG5911 encodes two functionally distinct subtypes of G protein-coupled receptors through alternative splicing (CG5911a and CG5911b) that respond preferentially to ecdysis-triggering hormones of flies and moths. These subtypes show differences in ligand sensitivity and specificity, suggesting that they may play separate roles in ETH signaling. At significantly higher concentrations (>100-fold), certain insect and vertebrate peptides also activate these receptors, providing evidence that CG5911 is evolutionarily related to the thyrotropin-releasing hormone and neuromedin U receptors. The ETH signaling system in insects is a vital system that provides opportunities for the construction of models for the molecular basis of stereotypic animal behavior as well as a target for the design of more sophisticated insect-selective pest control strategies.
A defining feature of growth and metamorphosis in insects is the periodic shedding of cuticle known as ecdysis (1). This complex process, by which the insect escapes its outer shell and sheds the cuticular lining of the foregut, hindgut, and respiratory system, is mediated by a peptide hormone-signaling cascade that results in a sequence of precisely timed physiological and behavioral events. Initiation of the ecdysis sequence coincides with the appearance of ecdysis-triggering hormones (ETHs) 1 in the bloodstream, which act directly on the central nervous system to elicit patterned motor output characteristic of pre-ecdysis and ecdysis behaviors (2,3). In the moth Manduca sexta, two peptides called MasPETH and MasETH act in concert to trigger successive phases of the ecdysis behavioral sequence. Remarkably, the temporal features of these motor patterns recorded from the isolated nervous system closely reflect the behavior observed in whole animals. Related peptides in Drosophila melanogaster known as ETH1 and ETH2 trigger ecdysis, although their respective roles in the recruitment of different parts of the ecdysis sequence are less clear. As yet the molecular and cellular targets of ETHs have not been described. Additional signaling molecules operating downstream of ETH within the central nervous system include the eclosion hormone and the crustacean cardioactive peptide (2, 4 -7).
Identification of ETH receptors would be of great help in defining the cellular elements involved in ecdysis behaviors. With this objective, we took advantage of the Drosophila genome project to investigate G protein-coupled receptors (GPCRs) that are likely to respond to ETH (8,9). The search for ETH receptors was narrowed to a small number of candidates on the basis of an assumption of ligand-receptor co-evolution, in which receptors for peptides having a C-terminal amino acid motif consisting of PRX amide would have high sequence homology. An exhaustive analysis of the Drosophila genome yielded two groups of GPCRs, the neuromedin U (NMU) group and the vasopressin group (10). Functional analysis revealed that the NMU group likely arose through ligand-receptor coevolution, because it responds to several groups of peptides with a C-terminal PRX amide motif characteristic of the ETHs. In contrast, the vasopressin group responds to unrelated peptides, and thus the PRX amide in this group appears to have arisen through convergent evolution.
The responsiveness of CG8795 to only relatively high concentrations of ETH1 suggested that it is unlikely to be the physiological ETH receptor. We therefore expanded the search profile for ETH receptors to include CG5911, which encodes a receptor related to the NMU group. As predicted (8), CG5911 gene products occur as two alternative splice variants termed CG5911a and CG5911b. CG5911b responds to ETH1 and the related peptides MasETH and MasPETH from moths at subnanomolar concentrations while showing relative insensitivity to other peptides sharing C-terminal PRX amide motifs. A pattern of cross-reactivity in a second set of peptides provides further support for the hypothesis that CG5911 encodes the authentic ETH receptor.

EXPERIMENTAL PROCEDURES
Cloning of CG5911-cDNA was synthesized using Superscript (Invitrogen) with mRNA isolated by Dynabeads (Dynal) from whole flies (ϳ50 individuals of Canton S) in both larval and adult stages by priming at poly(A) sites. PCR reactions conducted in 20-l volumes contained 20 mM Tris-HCl, pH 8.4, 50 mM KCl, 2 mM MgCl 2 , 0.2 mM dNTP, 0.5 M of each primer, 0.5 unit of Taq polymerase (Invitrogen), and 0.5 unit of Pfu polymerase (Stratagene). The reaction mixture was denatured initially for 5 min at 94°C and then subjected to 35 cycles of 94°C for 1 min, 53°C for 1 min, and 72°C for 1 min and 30 s. The RT-PCR products were cloned in pGEM-T-EZ vector (Promega) and sequenced. Start and stop codons of the open reading frames (ORF) were confirmed by the primers in the RT-PCR designed on the outside of the in-frame stop codons of the ORF. The primers shown in Fig. 1 are Fa (5Ј-AAGGACGATGGCGATGCTGA) paired with Ra (5Ј-GCTTACCCCT-TACTGCCAGT) and Fb (5Ј-CCCGAGGATGTTGTCCCTGT) paired with Rb (5Ј-GATGGGATTCAGATCTTGCT). The CG5911 ORFs were inserted into the pXOON vector (17), which contains an enhanced green fluorescent protein marker. Phylogenetic analysis was done in PAUP version 4.0b8a for generation of a distance tree with 1000 bootstrapping samplings.
Cell Line Expression-Transient expression of CG5911 gene products was accomplished in the CHO-WTA11 cell line exhibiting stable expression of the luminescent reporter aequorin (18) and the G protein G␣ 16 (19). Cells were grown in complete Ham's F12 medium at 37°C in 5% CO 2 . Transfection with pXOON-CG5911a and pXOON-CG5911b was performed using FuGene6 (Roche Molecular Biochemicals) according to the manufacturer's protocol at a DNA to FuGene6 ratio of 3:1.
Transfected cells were observed under an epifluorescence microscope 1 day later to determine transfection efficiency as measured by the number of cells expressing enhanced green fluorescent protein. Before functional assays, cell suspensions were incubated in coelanterazine h (Molecular Probes) according to previously defined protocols (20).
Functional Analysis-Luminescence assays were performed in opaque 96-well microplates (Corning) using a Beckman model LD400 Luminometer. After the addition of cells to a well, luminescence was monitored using an 0.5-s sampling time for 20 s. Each 96-well microplate contained multiple wells for positive controls (100 and 300 M ETH1) and negative controls (buffer only). Luminescence at each ligand concentration was integrated during the 20-s response interval and normalized to the largest positive control response in each plate after the subtraction of background values obtained from negative controls. Luminescence measured in multiple replica wells (2-4 wells) for one concentration of ligand was averaged for the analysis. Data collected from at least three replica plates were used for analysis using the Origin analysis program (OriginLab Corp., Northampton, MA). The sources of peptides were described previously (10), and additional vertebrate peptides (NMU-8, TRH, and ghrelin) were purchased from the American Peptide Company. MasETHacid was synthesized at the University of California, Riverside, using standard Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry and automated solid state techniques. A list of peptides used in this study is provided in Table I.

Molecular Cloning of CG5911a and CG5911b-The CG5911
gene has homologous mutually alternative exons 4a and 4b encoding the protein sequence from the middle of transmembrane segment 4 to the C terminus (Figs. 1 and 2). The clones used for the expression contain the coding sequence from nucleotide position Ϫ24 upstream of the Met site to ϩ368 nucleotides after the 3Ј stop codon for CG5911a and from Ϫ490 to ϩ8 for CG5911b (see GenBank™ accession numbers AY220741, AY220742, and Celera Genomics sequence GenBank TM accession number AE003734) (Fig. 2). The ORF is supported by the in-frame stop codon in the 5Ј-untranslated region (Fig. 1).
Initial BLASTP searches and phylogenetic analyses identified the translation of CG5911 (GenBank TM accession number AAF55872) as a Drosophila GPCR closely related to the monophyletic group of NMU receptors responding to PRX amide peptides (10). Putative transcripts generating CG5911a and -b are grouped with thyrotropin-releasing hormone (TRH) receptor in a distance tree with a bootstrapping value of 83.8% (Fig. 3).
The prediction of gene structure revealed the presence of homologous mutually alternative exons 4a and 4b from transmembrane segment 4 to the C terminus encoding two splice variants of the receptor (Figs. 1 and 2). Stop codons occurring at C-terminal ORFs for CG5911a and CG5911b were identified by RT-PCR using primers located after the stop codon. RT-PCR with the primer covering a stop codon at the 5Ј end of the ORF for CG5911b revealed a transcript supporting the ORF prediction.
CG5911a and CG5911b Are Preferentially Sensitive to ETHs-Ecdysis-triggering hormones of both Drosophila and FIG. 1. Genomic structure of CG5911 encoding putative ETH receptors. The receptor has two alternative transcripts with mutually alternative exons 4a and 4b. Transcripts were experimentally confirmed by RT-PCR with primers (arrows) located at the boundary of start and stop codons in the predicted open reading frames. The 5Ј and 3Ј ends of the transcript were not determined and are depicted as slanted ends of exon 1 (5Ј) and exon 4 (3Ј).  Manduca elicited robust luminescence responses in the CHO-WTA11 cell line stably expressing G␣ 16 and aequorin transfected with pXOON-CG5911a and pXOON-CG5911b (Fig. 4). The time course of cellular responses was concentration-dependent, with delayed slow responses beginning at 0.1 nM and rapid peak responses occurring at 1 M for CG5911b (Fig. 4A). At the lowest effective concentrations (0.1 nM), the luminescence response was delayed more than 10 s. Integration of each response shown in Fig. 4A during a 20-s sampling period is shown in Fig. 4B.
Both CG5911a and CG5911b showed preferential sensitivity to ETHs, with CG5911b showing a significantly higher appar-ent affinity for all peptides tested (Fig. 4, C and D). Drosophila ETH1 was the most active ligand for both receptors. For CG5911b, responses were evident at 0.25 nM and reached peak values at 10 nM with an EC 50 value of 0.9 nM. ETH2 and Manduca PETH were only slightly less active, with EC 50 values of ϳ2 nM. Interestingly, Manduca ETH was approximately 5 times less active on CG5911b than PETH.
CG5911a was markedly less sensitive to all peptides tested compared with CG5911b. The most active peptide again was ETH1, with an EC 50 of 414 nM, making it ϳ400-fold less active on this receptor than on CG5911b. ETH2, MasPETH, and MasETH had EC 50 values greater than 2 M.
Cross-reactivity of CG5911 to Other Peptide Ligands-CG5911b showed significant sensitivity to a range of ETHrelated peptides having the signature PRXa C-terminal amino acid motif (Fig. 5). These peptides include SCP B , HUG␥, CAP2b-1, CAP2b-2, PBAN, and MasETH acid. The most active of these peptides were SCP B , MasETH acid, HUG␥, and CAP2b-1; the EC 50 values of these peptides were in the range of 300 nM to 3 M. Furthermore, the vertebrate peptides NMU and TRH caused low but significant receptor activation when applied at 10 M as did the molluscan peptide myomodulin (p Ͻ 0.001 in Student's t test; Fig. 5B). No activity was observed for a number of other peptide ligands, including some with the C-terminal PRXa motif. These ligands include Drm-PK-2, CAP2b-3, and Hez PMP (Fig. 5B). DISCUSSION Ecdysis-triggering hormones are vital regulatory signals that govern the stereotypic physiological sequence leading to cuticle shedding in insects. In previous work we identified ETHs in two moth species, Manduca sexta and Bombyx mori, and in the fruit fly Drosophila melanogaster (2,15,16). Using genetic tools in Drosophila to excise the eth gene, lethal ecdysis deficits were demonstrated at the end of the first larval instar (21). Lethality was reversed by the injection of ETH peptides (21). The ETH signaling system therefore has been demonstrated to be both necessary and sufficient for insects to survive the earliest ecdysis. To expand our understanding of the molecular basis for ETH action, we conducted a search for its likely receptor in Drosophila.
In this paper we report on the cloning and functional expression of the Drosophila gene CG5911, predicted by Hewes and Taghert (8) to encode splice variants called CG5911a and CG5911b. Our results show that both splice variants of CG5911 are highly selective for ETH1 and related peptides from both fly and moth, with CG5911b responding to subnanomolar con- FIG. 3. Phylogenetic tree for the G protein-coupled receptors in the clade including the neuromedin U receptor. The phylogenetic tree was generated in PAUP (version 4.0b8a) for the neighborjoining distance method, with numbers at the nodes representing the bootstrapping values in 1000 replications (see Fig. 2 legend for a description of the taxa). The sequence for CG5911b was used only for exon 4b (Fig. 1), which is a mutually exclusive alternative exon. centrations of Drosophila ETH1. For each receptor subtype the highest sensitivity was observed to Drosophila ETH1 and ETH2, but comparable sensitivity also was observed for Mas-PETH and MasETH. Other insect and vertebrate peptides, particularly those having C-terminal PRX amide motifs, also activated CG5911b, but only at ϳ100 -1000-fold higher concentrations. The selectivity and high sensitivity of CG5911 to ETHs alone strongly supports the hypothesis that this gene encodes the authentic ETH receptor.
The gene structure of CG5911 with the results of RT-PCR revealed two homologous exons presenting the mutually alternative splice variants CG5911a and CG5911b. Phylogenetic analysis suggests that divergent exons 4a and 4b are the product of duplication. The portion of the receptor protein predicted by exon 4 extends from the middle of transmembrane segment 4 (T4) to the C terminus and therefore includes two extracellular loops (T4-T5 and T6-T7), one intracellular loop (T5-T6), and the C-terminal intracellular tail. Clearly, alternative splicing of exons 4a and 4b to create these receptor isoforms could affect both their ligand and G protein-coupling specificities.
Functional differences in CG5911a and CG5911b are suggested by several observations. CG5911b is activated by subnanomolar concentrations of ETH1 (EC 50 ϭ 0.9 nM), whereas the sensitivity of CG5911a is more than 400-fold lower (EC 50 ϭ 410 nM). The lower sensitivity of CG5911a to all ligands may be a consequence of poor coupling efficiency to the G ␣16 used in our assay system. Indeed, the smaller size of peak luminescence responses recorded for CG5911a as compared with CG5911b (data not shown) supports this idea. Further investigation of CG5911a sensitivity using different G proteins will be necessary to resolve this question. Nevertheless, differences in G protein coupling and downstream signal transduction steps could provide a basis for functionally distinct roles for these two receptor subtypes in ETH signaling.
In addition to these differences, the ligand sensitivity for the two receptor subtypes is significantly different. The relative potency of ligands on CG5911a (ETH1 Ͼ MasETH Ͼ ETH2 Ͼ MasPETH; see Fig. 4) reflects the same order of biological activity observed for the induction of ecdysis behavior (16). In the case of CG5911b, we observed a somewhat different order of potency (ETH1 Ͼ ETH2 Ͼ MasPETH Ͼ MasETH). The reversal in receptor sensitivity to MasPETH and MasETH suggests that these receptor subtypes discriminate between MasPETH and MasETH and raises the intriguing possibility that CG5911 orthologs in the Manduca system encode receptors that are ligand-specific. Indeed, preliminary findings involving the Manduca ortholog of CG5911b confirm the preference for Mas-PETH. 2 The possible ligand specificity of the two receptor subtypes may account at least in part for divergent activities of these two ligands in the induction of natural pre-ecdysis behavior (2,3). Further study of Manduca orthologous isoforms of CG5911 is under way to test this hypothesis.
Responses of CG5911b to a second tier of ligands that are active only at Ͼ100-fold-higher concentrations reveal further insights into its properties and the likely physiological role of CG5911 as the ETH receptor. All ETHs are C-terminally amidated, and this post-translational modification is essential for biological activity in the silkworm, B. mori (15). Although the injection of synthetic Bombyx ETH elicits the entire ecdysis behavioral sequence in silkworm larvae, the free acid form proved to be largely inactive (15). In the present study, the free acid form of MasETH (MasETHacid, Fig. 5) was found to be more than 100 times less active than native MasETH. CG5911b therefore discriminates between these subtly differ-ent forms of the peptide just as we observed previously in behavioral assays (2).
We tested a variety of both invertebrate and vertebrate peptides on CG5911b, and most were inactive. Nevertheless, some interesting patterns of activity are apparent among the few peptides that showed some biological activity. Of the invertebrate peptides tested, the peptide HUG␥, encoded by the ortholog of the moth PBAN gene as well as PBAN itself, were active at several hundred-fold higher concentrations than authentic ETHs. The low affinity response of CG5911b to HUG␥ and its high affinity response to ETH1 and ETH2 complement our previous results, which showed that the related GPCR encoded by CG8795 is highly sensitive to HUG␥ but considerably less sensitive to ETHs (10). These results confirm our phylogenetic analyses, which indicate the close evolutionary relationship between CG8795 and CG5911. The question remains as to whether the cross-reactivities of these two ligands (ETH and HUG␥) against CG8795 and CG5911b reflect genuine biological phenomena. Evidence in support of this comes from Roos and colleagues (22), who showed that ectopic expression of the hugin gene produced lethal ecdysis defects.
The mammalian peptides neuromedin and TRH, the receptors of which are related phylogenetically to CG5911, also activate CG5911b, albeit at relatively high concentrations. The significant bioactivity of these peptides, with the complete absence of responses to neurotensin and ghrelin (assayed at 10 M), strongly support our phylogenetic analysis, which indicates that CG5911 evolved from a GPCR ancestral to the NMU/ TRH receptors.
Myomodulin (PMSMLRLamide), a neuropeptide first identified in molluscs, also was found to activate CG5911b. We were unable to find myomodulin-like peptides in the Drosophila genomic data base, but a similar sequence occurs in the Caenorhabditis elegans data base (Y45F10A.5). 3 Myomodulin could play an analogous role to ETH in nematodes, which with the arthropods have been newly classified in the Ecdysozoa (23). Indeed, antibodies raised against myomodulin cross-react with ETHs in endocrine Inka cells (24), indicating structural similarities between these peptides.
During preparation of this manuscript, independent evidence appeared to support the hypothesis that CG5911 functions as the ETH receptor (25). These findings confirm that two functional subtypes of CG5911 are produced by alternative splicing and that ETH1 is a more potent agonist against both receptors. These authors (25) also demonstrated the presence of an additional 5Ј-untranslated exon. However, our results demonstrate a higher sensitivity of CG5911 to both ligands and a considerably greater divergence in receptor sensitivity between the two subtypes. These findings could indicate that CG5911a and CG5911b interact with different G proteins and may activate different signal transduction pathways. Our pharmacological and phylogenetic analyses also indicate that ETH and the pyrokinin/CAP2b groups of insect neuropeptides constitute a set of ligands that have evolved with CG5911 and the neuromedin U group of vertebrate receptors and their respective ligands.
In summary, we have demonstrated that alternative splice variants of CG5911 encode two functional receptor subtypes of ETH receptor in Drosophila. Although further work is necessary to confirm their physiological roles in natural ETH signaling, the pronounced ligand sensitivity and specificity of CG5911 gene products along with the phylogenetic relationships of CG5911 to both Drosophila and mammalian receptors for PRX amide peptides provide strong evidence for their roles as the physiological ETH receptors.