Functional Annotation of Two Orphan G-protein-coupled Receptors, Drostar1 and -2, from Drosophila melanogaster and Their Ligands by Reverse Pharmacology*

By combining a Drosophila genome data base search and reverse transcriptase-PCR-based cDNA isolation, two G-protein-coupled receptors were cloned, which are the closest known invertebrate homologs of the mammalian opioid/somatostatin receptors. However, when functionally expressed in Xenopus oocytes by injection of Drosophila orphan receptor RNAs together with a coexpressed potassium channel, neither receptor was activated by known mammalian agonists. By applying a reverse pharmacological approach, the physiological ligands were isolated from peptide extracts from adult flies and larvae. Edman sequencing and mass spectrometry of the purified ligands revealed two decapentapeptides, which differ only by an N-terminal pyroglutamate/glutamine. The peptides align to a hormone precursor sequence of the Drosophila genome data base and are almost identical to allatostatin C from Manduca sexta. Both receptors were activated by the synthetic peptides irrespective of the N-terminal modification. Site-directed mutagenesis of a residue in transmembrane region 3 and the loop between transmembrane regions 6 and 7 affect ligand binding, as previously described for somatostatin receptors. The two receptor genes each containing three exons and transcribed in opposite directions are separated by 80 kb with no other genes predicted between. Localization of receptor transcripts identifies a role of the new transmitter system in visual information processing as well as endocrine regulation.

Insect development and behavior are largely controlled by hormones and neurotransmitters often identified using a diverse array of bioassays. Besides the biogenic amines and the steroid-like hormones, insect hormones have been frequently classified as neuropeptides, which are widely distributed throughout the invertebrate kingdom (1,2). Despite the large number of neuropeptides, the number of known cognate receptors in insects is still rather limited, with only a few examples in Drosophila that have been cloned based on homology with mammalian G-protein-coupled receptors (GPCRs) 1 (i.e. neuropeptide Y-like and tachykinin-like receptors) (3,4). With the completion of the Drosophila genome project, a more thorough analysis of neuropeptide/receptor relations in insects is now possible. Whereas this genome data base allows the identification of peptide hormones previously isolated from other insect species as part of larger precursors (5), Drosophila GPCR-like sequences have been predicted mostly based on structural analogy of the transmembrane regions to mammalian neuropeptide receptor groups (6).
Structural evidence for the existence of ligands identical or similar to their mammalian neuropeptide counterparts are lacking when searching the Drosophila genome data base. This may indicate that in insects these receptors are activated by an entirely different set of ligands. This view is supported by data reported here on the identification of two novel GPCRs from Drosophila melanogaster, termed Drostar1 and -2, which are structurally related to the mammalian opioid/somatostatin receptor family yet are activated by peptide ligands, pyro-Gluand Gln-allatostatin C, unrelated to any known mammalian agonists.

MATERIALS AND METHODS
Molecular Cloning-Several PCR primers were designed to amplify SST-like receptor cDNAs from Drosophila RNA (7); 5Ј-atggatccatgacagccgattcagagg-3Ј and 5Ј-taaagcttttacaaatctgtctgctgc-3Ј were used to amplify a 1.4-kb cDNA corresponding to the CG7285 gene product of the Drosophila genome data base. 5Ј-gtctgcatcattggactctttg-3Ј and 5Ј-ttttctagattataagtccgtgtggagcacg-3Ј were used to amplify a central fragment of the CG13702 gene product. Based on the putative 3Ј end of the coding region of this gene (predicted by similarity to the CG7285 gene product), we used primers 5Ј-gtctgcatcattggactctttg-3Ј and 5Ј-gtctgcatcattggactctttg-3Ј to amplify the 3Ј-end of this cDNA. The 5Ј-end of this cDNA was obtained by 5Ј-rapid amplification of cDNA ends (7). All PCR fragments were cloned into TOPO vectors (Invitrogen) and sequenced. Site-directed mutagenesis was performed by PCR using mutagenic primers and Vent DNA polymerase.
Functional Expression in Xenopus Oocytes-Wild type and mutant receptor cDNAs were cloned into a modified pGEMHE expression vector, which carries a sequence coding for the signal peptide of the 5-hydroxytryptamine receptor 3 (amino acid sequence MVLWLQLAL-LALLLPTSLAQGEVDI); the pGEMHE vector itself contains 5Ј-and 3Ј-untranslated regions derived from the Xenopus laevis globin gene, * This work was supported by Deutsche Forschungsgemeinschaft Grants SFB545/B7 (to H.-J. K. and D. R.), GRK 255/2 (to D. R.), and SFB 444/A6 (to T. R.) and European Commission Grant QLG3-CT-1999-00908 (to D. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AY017416 and AY046072.
‡ These authors contributed equally to this work. ¶ To whom correspondence should be addressed. E-mail: richter@ uke.uni-hamburg.de. which stabilize exogenous RNA in Xenopus oocytes (8). After linearization of the vector with NheI, cRNA was transcribed with T7 RNA polymerase and coinjected with the mouse GIRK1 cRNA into Xenopus oocytes (7). Agonist-induced potassium currents were recorded in high potassium (hK) solution under whole cell voltage clamp conditions (7). For the testing of fractions obtained by various chromatographic steps, aliquots were lyophilized and dissolved in hK buffer.
Peptide Purification-Crude peptide extracts from adult flies and a mixture of larvae and pupae (both about 50 ml of material) were prepared by homogenizing frozen animals in 200 ml of 0.5 M acetic acid, 10 mM ascorbic acid on ice; the homogenate was incubated in boiling water for 10 min and again incubated on ice for 1 h. Insoluble material was removed by centrifugation (30,000 ϫ g, 30 min) followed by filtration of the supernatant fraction. The extract was then applied to C-18 reverse phase material; after washing with water plus 0.1% trifluoroacetic acid, peptides were eluted with 80% methanol and lyophilized.
A crude purification of this extract was achieved by gel filtration using a Sephadex G-25 column (43 ϫ 1.4 cm; in 0.1 M acetic acid); active fractions from both the adult flies and the larvae/pupae mixture were pooled and further purified using Vydac C-8, C-18, C4, and Nucleosil C-8 (two times) reverse phase HPLC columns.
Sequencing by Edman Degradation and Mass Spectrometric Analysis-Peptide sequences were determined by standard Edman degradation using an automatic pulsed-liquid protein sequencer (476A; Applied Biosystems, Foster City, CA). Mass spectroscopic analysis was performed with a hybrid tandem mass spectrometer (QTOF II; Micromass, Manchester, UK) equipped with a nanoelectrospray ion source. Samples were purified by binding to C18 reverse phase material in a pipette tip (ZipTip; Millipore Corp.) and eluted with 1 l of 50% methanol, 5% formic acid into a gold-coated borosilicate vial for nanoelectrospray measurements.
Synthetic Peptides-Synthetic, linear Gln-allatostatin C (QVRYRQ-CYFNPISCF) was obtained by custom synthesis from Genemed Synthesis, Inc. (South San Francisco, CA). For oxidation, the peptide was treated with K 3 Fe(CN) 6 (9) for reduction with 1 mM dithiothreitol (30 min, 65°C). Conversion of linear Gln-allatostatin C peptide into pyro-Glu-allatostatin C was achieved with glacial acetic acid at 95°C for 60 min, followed by K 3 Fe(CN) 6 oxidation as above. The purity and the identity of active agonists was then determined by reverse phase HPLC as well as by mass spectroscopy.
In Situ Hybridization and Immunocytochemistry-Heads of adult flies were snap frozen in Tissue Tek (Sakura, Tokyo, Japan). Sections (10 m) were dried for 1 h at room temperature and postfixed in 4% paraformaldehyde/phosphate-buffered saline, pH 10. Washing was performed with phosphate-buffered saline, pH 7.4, 0.3% Triton X-100. Prehybridization was done with Ultrahyb (Ambion, Austin, TX) supplemented with salmon sperm and tRNA for 2 h at 60°C. cRNA was in vitro transcribed (Ambion) with digoxigenin UTP. Hybridization was carried out overnight at 60°C in prehybridization buffer followed by immunohistochemical detection with alkaline phosphatase-labeled anti-digoxigenin antibodies (Roche Molecular Biochemicals).

RESULTS
Receptor Identification-When screening the Drosophila genome data base with mammalian somatostatin or opioid receptors as probes, two highly similar sequences partially overlapping with the predicted CG7285 and CG13702 gene products were detected on a contig (accession number AE003520) derived from chromosome 3L.
Using total RNA isolated from adult flies and primers based on conserved regions of the mammalian SSTR2 and on the coding sequences predicted from the Drosophila genome project annotation data base, we isolated and sequenced cDNAs coding for both receptors by reverse transcriptase-PCR. The CG7285 cDNA could be amplified using forward and reverse primers directed at the 5Ј-and 3Ј-ends of the predicted coding sequences; this was not possible for the CG13702 cDNA, since both ends of the coding region were incorrectly predicted. Here primers for the 3Ј-end were designed based on similarity to the CG7285 gene product, whereas the 5Ј-end of the cDNA was determined by 5Ј-rapid amplification of cDNA ends. Open reading frames of 1401 bp (CG7285) and 1467 bp (CG13702) were detected, coding for proteins of 467 and 489 amino acids, respectively ( Fig. 1). The deduced protein sequences exhibit the characteristic pattern of GPCRs (seven TMs; N-linked glycosylation sites at the N terminus). Both receptors share 60% identical amino acid residues; considering only the TMs, this value increases to 76%, strongly suggesting that the two receptors are distinct members of a novel GPCR subfamily. When comparing both sequences with GPCRs present in the Gen-Bank TM data base, mammalian SSTR2 appears as the closest relative (42% amino acid identity in the TMs), followed by the other SSTR subtypes and then the opioid receptors (Fig. 2). Other insect receptors (AlstR1 and -2) appear closer to the mammalian galanin receptor family.
Functional Expression of the Two Receptors and Identification of Their Cognate Ligands-Functional expression in Xenopus oocytes was achieved by coinjection of CG7285 cRNA together with GIRK1 cRNA. Significant signals were recorded when oocytes expressing the novel Drosophila receptor were exposed to a crude peptide extract from Drosophila larvae or adult flies (Fig. 3A). Since weaker signals were also recorded from oocytes expressing the GIRK1 channel alone, the specificity of signals could only be ascertained after Sephadex G-25 gel filtration of the crude peptide extracts (Fig. 3B). Signals could be detected in fractions corresponding to a molecular mass of about 1000 -2000 Da; these signals were specific, since they were not observed with oocytes expressing the GIRK1 channel only.
Active fractions were further purified by reverse phase HPLC (Vydac C-8 column). The activity was eluted as two distinct peaks (Fig. 4A), which were purified separately. After three further reverse phase HPLC steps, the first (slightly more hydrophilic) active ligand was finally isolated in an apparently pure peptide fraction, as judged by UV absorbance of the HPLC elution profile (Fig. 4B). Analyzed by Edman degradation, this peptide revealed the sequence XVRYRQXYFNPIX. Table IA shows that this sequence aligns with part of a hormone precursor sequence derived from the Drosophila genome data base and referred to as Drostatin precursor (11). The resulting decapentapeptide is reminiscent to allatostatin C isolated from the tobacco hornworm, M. sexta (12), except that in Manduca the first amino acid residue is a pyroglutamate, and in position 4 the tyrosine is replaced by phenylalanine (Table  IB). Our ability to sequence the Drosophila agonist of peak 1 (Fig. 4B) by Edman degradation indicated that the pyro-Glu modification was not present in the purified peptide. In agreement with this, mass spectroscopy of the pure peptide yielded a peptide mass of 1920.67, which corresponds to the predicted mass of the unmodified (but Cys-oxidized) peptide starting with a glutamine (Table IB).
The second peptide peak (Fig. 4) was purified accordingly, but due to its rather low abundance when compared with peak 1 it was analyzed by mass spectroscopy; the detected peptide mass of 1903.26 corresponded to the decapentapeptide carrying an N-terminal pyroglutamate, indicating that Drosophila contains two types of allatostatin C carrying N-terminally either a glutamine or a pyroglutamate; they are present in the peptide extract with an estimated ratio (in terms of agonistic activity) of 1:2 (pyro-Glu-allatostatin C/Gln-allatostatin C; see Fig. 4A). Accordingly, the two receptors were termed Drostar1 and -2 (for Drosophila allatostatin C receptors 1 and 2).
To verify our findings that the deduced sequences of the decapentapeptides are indeed that of the agonists of Drostar1 and -2, the linear Gln-allatostatin C peptide was custom-synthesized and tested for ligand activity. The linear peptide as obtained after synthesis yielded a rather low receptor-induced potassium current in the Xenopus system. However, after oxidation of the two cysteine residues (positions 7 and 14), strong agonist-induced currents were obtained, which were specific to oocytes expressing either Drostar1 or -2 and GIRK1 but which were not observed with oocytes expressing either GIRK1 alone or SSTR2/GIRK1 (Fig. 5, A-C). Strong currents were also induced by the peptide after generation of the N-terminal pyro-Glu modification, followed again by peptide oxidation. A comparison of HPLC elution profiles of synthetic, oxidized peptide carrying N-terminally either the pyro-Glu or Gln showed that the pyro-Glu peptide version elutes about 2 min later from a reverse phase Vydac C-8 HPLC column (data not shown), confirming that the second peak of agonistic activity obtained during the purification procedure (Fig. 4B) is the Drosophila pyro-Glu-allatostatin C reminiscent of the M. sexta peptide. The reduced form of the Gln-allatostatin C peptide exhibited a potency that was about 100-fold lower than that of the oxidized form.
To test whether the two Drostars discriminate between Glnand pyro-Glu-allatostatin C peptides, dose-response curves were recorded. Both peptides were almost equally active at both receptors with EC 50 values for GIRK1 activation in the low nanomolar range (Gln-allatostatin C: 9.45 nM at Drostar1, 7.0 nM at Drostar2; pyro-Glu-allatostatin C: 25.4 nM at Dro-  (AlstR1 and -2). All sequences were obtained from GenBank TM . The lengths of the horizontal lines indicate reciprocally the sequence similarities. The bar at the bottom corresponds to a mutation rate of 10 bases per 100 residues. Amino acid alignments were generated using the PileUp software in the GCG package. Evolutionary distances were then calculated using the Distances program (GCG), and the phylogenetic tree was reconstructed using the GrowTree program. D, Drosophila; h, human.

FIG. 3. Functional expression of the novel receptor in Xenopus oocytes.
A, oocytes expressing GIRK1 and the CG7285 receptor (upper trace) or the GIRK1 channel alone (lower trace) were voltage-clamped at Ϫ80 mV; measurements were initiated by switching oocytes to hK medium (downward arrow); after the initial current reached a plateau, peptide extracts (diluted in hK) from adult flies (AF) or larvae/pupae (L/P) were applied for the time indicated (bars); measurements were terminated by perfusion with ND-96 medium (upward arrows). Similar expression profiles were obtained with the CG13702 receptor. B, gel filtration of peptide extracts from L/P or AF on a Sephadex G-25 column. Fractions were collected starting with the void volume of the column determined by the appearance of a small amount of Dextran Blue added to the extracts. Aliquots were lyophilized and tested for activity on oocytes expressing GIRK1 with (squares) or without (triangles) the CG7285 receptor. Bars, active fractions pooled for further purification. star1, 8.7 nM at Drostar2; Fig. 5B), suggesting that the modification at the N terminus of allatostatin C does not contribute significantly to the ligand binding affinity toward its receptors. As known from other peptides starting with a pyroglutamate, the corresponding amino acid in the respective precursor sequence is always a glutamine (Table IA), suggesting that Drostatin is converted into both types of allatostatin C. The reason for the presence of more unprotected versus protected allatostatin C in Drosophila is at present not quite clear. It may reflect that the capacity to produce pyro-Glu-allatostatin C is restricted to a limited number of cells.
Mutational Analysis-The affinity and ligand specificity of mammalian somatostatin receptors are determined by residues in transmembrane region 3 (TM3; D 124 in SSTR3 (13)) and in those portions of TM6 and -7 facing the extracellular side of the plasma membrane (14). These residues in Drostar1 were changed into the corresponding residues of the SSTR2 sequence by site-directed mutagenesis, and recombinant receptors were challenged with pyro-Glu-allatostatin C and somatostatin 14. A point mutation in TM3 reduced the affinity of Drostar1 about 3-fold (Fig. 6), whereas a Drostar1 mutant carrying SSTR2 sequence from the middle of TM6 to the middle of TM7 was not activated by allatostatin C at all concentrations tested. Neither mutant nor wild type receptors were activated by SST14 (data not shown). It should be noted that when constructs were transfected into human embryonic kidney cells, there were no differences in the localization pattern between wild type and mutant receptors, indicating that the transport to the cell surface was not impaired by the mutations (data not shown). Gene Organization and Expression Pattern-Both Drostars are encoded by distinct genes located very close to each other at a distance of about 80 kb with no other genes predicted between them (Fig. 7). The open reading frames of both genes are directed in opposite directions; both receptor mRNAs are transcribed from three exons, including a small exon with 48 bp in size that codes for a part of the sixth transmembrane region (Fig. 1). Exon/intron boundaries, as deduced by comparison with the genomic sequence, are entirely conserved between both receptors. Although receptor-coding transcripts were predicted for both the 138-kb region (CG13702) and the 224-kb region (CG7285) of this part of the Drosophila genome by the GADFLY data base, predictions for both genes proved to be inaccurate, since the small exon was apparently missed in each case by the prediction software used; in the case of Drostar2, both the N-and C-terminal part were also not correctly assigned (see data base entry, accession number AE003520).
In situ hybridization experiments revealed expression of both Drostar genes within the optic lobes of adult flies, an area devoted to the processing of visual information (Fig. 8A). In the lamina region, these cells might represent lamina monopolar cells, the main cellular components of the neuropil (15). In addition, a region in the anterior median part of the brain, the pars intercerebralis, has also few Drostar-expressing cells. Since allatostatin C is also present in the pars intercerebralis (16), the Drostar receptors on these cells might be involved in autoregulatory processes. In addition, a group of cells at the dorsal margin of the optic lobes also displayed allatostatin C-like immunoreactivity (16), which may project to Drostarexpressing cells of the optic lobe. The presence of both the peptide and the receptor in the optic lobes suggests a function in the modulation of visual information processing. Interestingly, the disperse labeling within the neuropil regions indicates Drostar expression in glial cells (Fig. 8A, arrowheads).
In larvae, Drostar expression was detected only in few cells of the nervous system (Fig. 8B). In contrast, immunolabeling of the peptide was found in the brain and corpora allata of the larvae, presumably on axon terminals arising from the pars intercerebralis of the brain (Fig. 8C). The allatostatin C peptide may be released from these terminals into the hemolymph or onto the corpora allata; thus, allatostatin C might function as the main allatostatic activity in Drosophila. In agreement, injection of pyro-Glu allatostatin C into the hemolymph of second instar larvae led to a slightly shortened time until they undergo pupariation (data not shown). DISCUSSION The data presented here identify two novel receptors, Dro-star1 and -2, from the fruit fly and their cognate ligands as  Fig. 3B were purified by reverse phase HPLC (Vydac C-8 column, 250 ϫ 4.6 mm; flow rate 1 ml/min; solvent A, 0.1% trifluoroacetic acid; solvent B, 70% acetonitrile, 0.1% trifluoroacetic acid). Aliquots of fractions were lyophilized and tested for activity on oocytes expressing CG7285 and GIRK1; peak currents obtained are indicated by columns. Note the split of agonistic activity into peaks 1 (P1) and 2 (P2). B, final purification step of peaks 1 and 2 (inset in gray) using a Nucleosil C-8 column (1 ϫ 250 mm; flow rate 0.2 ml/min; solvents as in A). Eluted fractions were tested for activity as in A; bars, single active fraction of each peak. Dotted lines, slope of gradients.

TABLE I
Identification of the native agonists of the Drosophila receptors A, alignment of the partial amino acid sequence of peak 1 (Fig. 4B) obtained by Edman degradation with the Drosophila precursor Drostatin (DAP C 102-121 ) sequence (11). Identical amino acid residues are inverted. The sequence within DAP C flanked by basic amino acids (underlined) predicts that this peptide is proteolytically released from its precursor. B, sequence and mass spectrometric comparison of peptide peak 1 or 2 ( Fig. 4B) with Manduca allatostatin C (pyro-Glu-AIC) (11). A comparison of mass spectroscopic data with the calculated values for Drosophila allatostatin C is given in the two right columns. X, unidentified residues; *, stop codon. pyro-Glu-allatostatin C and Gln-allatostatin C. Different peptides have been described as allatostatins based on their ability to inhibit juvenile hormone synthesis from the retrocerebral corpora allata complex of insects. They can be grouped into the allatostatin subfamilies A, B, and C. Members of subfamily A share the conserved C-terminal peptide motif YXFGL-amide, whereas their N terminus may vary in sequence and length. As reported, the conserved pentapeptide is sufficient to activate the cognate invertebrate AlstR1 and -2 receptors (7,17). Members of the allatostatin B peptide family are characterized by the C-terminal motif XWXXXXXXW-amide. Neither of the two sequence motifs are present in the allatostatin C peptides of M. sexta or Drosophila. Thus, only the allatostatin C peptides activate the two Drostars, whereas no cross-activation of the two receptors occurs when complemented with allatostatin A peptides (Table II); conversely, AlstR1 is not activated by allatostatin C. As in M. sexta, our immunocytochemical data show that in Drosophila allatostatin C is present in the corpora allata presumably on axon terminals arising from the pars intercerebralis of the brain (Fig. 8), pointing to an allatostatic function. Using antibodies against an allatostin A-type peptide allatostatin-like immunoreactivity in Drosophila has been reported earlier mainly in nerve cells but not in fibers terminating on the corpora allata (18), also suggesting that in Drosophila allatostatin C rather than an allatostatin A is involved in the regulation of juvenile hormone synthesis. This assumption is supported by the observation that second instar larvae injected with allatostatin C show a slightly shortened period until they start pupariation. On the other hand, pyro-Glu allatostatin C does not reduce neurotransmitter release at the larval neuromuscular junction, indicating that Drostars are not located presynaptically on glutamatergic terminals. 2 Clearly, the final elucidation of the physiological function of the ligand-Drostar system in the fly remains to be solved in the future. Genetic means may be the alternative approach by using, for instance, the chimeric Drostar receptor (Fig. 6) as a dominant negative construct for generating transgenic flies.
Structurally, the two Drostars are most closely related to the mammalian opioid/somatostatin receptor family; homology between Drostar1 and -2 described here and their mammalian counterparts is remarkably high in those parts of the transmembrane regions facing the cytosol (i.e. that part where the receptors couple to their associated heterotrimeric G-proteins). In accordance with this observation, both Drostars and mammalian opioid and SST receptors can be coupled to the same intracellular pathway in Xenopus oocytes (i.e. activation of GIRK1 currents presumably via G i/o subunits). In contrast, those portions of the TMs facing the extracellular side are less well conserved, in keeping with their role in agonist binding and specificity. As depicted in Table II, the natural agonists of Drostars, pyro-Glu-and Gln-allatostatin C, bear only superficial similarity to somatostatin-14, since they are rather similar in their molecular weights (14 versus 15 amino acid residues) and also in carrying a disulfide loop absent in all other known allatostatin peptides and essential for high affinity to the Drostar receptors. Otherwise, there is no detectable sequence similarity between the allatostatin C peptides and SST-14.
The affinity and ligand specificity of mammalian somatostatin receptors are determined by two sites; one consists of a single aspartate residue in transmembrane region 3 (TM3; Asp 124 in SSTR3) (13), and the other consists of a region between TM6 and -7 facing the extracellular side of the plasma membrane (14). The aspartate residue of the SSTR is part of a narrow and selective pocket that forms an ion pair with the positively charged lysine residue at position 4 or 9 of somatostatin (13). In the somatostatin-like receptors of Drosophila, Drostar1 and 2, this aspartate is replaced by threonine (Thr 140 ). When converting this residue into an aspartate by site-directed mutagenesis, the recombinant Drostar1 receptor showed a 3-fold reduced affinity for pyro-Glu-allatostatin C; no activity was observed with somatostatin 14. A Drostar1 chimeric mutant carrying the second binding pocket of SSTR2 from the middle of TM6 to the middle of TM7 was activated by neither pyro-Glu-allatostatin C nor somatostatin 14 at all concentrations tested. Accordingly, there is no cross-reactivity between the Drostar and the SSTR ligand/receptor systems even in the mutant constructs. This suggests that during evolution of this group of receptors, a common scaffold of a receptor protein coupled to inhibitory G-proteins appears to be maintained, whereas the extracellular parts facing the ligand are rather variable in order to accommodate different peptide ligands. Both SSTRs and Drostars presumably use part of TM3 and the extracellular portions of TM6 and TM7 to accommodate their agonists, as is underlined by our mutagenesis study.
Since the Drostars are the closest relatives of mammalian SST/opioid receptors but are not activated by these peptides (Table II), it may be speculated that neither opioid nor SST-like peptides exist in Drosophila. This view is supported by our inability to identify SST or opioid precursor-like sequences in the Drosophila genome data base (data not shown), although exon-intron boundaries within a predicted peptide may complicate such a search. As a conclusion, whereas many receptors from vertebrates have a structural analog in invertebrates (6), the latter may use an entirely different set of peptides to activate these receptors. For a functional annotation of the many receptor sequences generated in genome projects such as 2 M. Heckmann, personal communication. . Specific hybridization was detected in the optic lobe (ol) and the pars intercerebralis (pi). At higher magnification of optic lobe, numerous labeled laminar monopolar cells are visible (arrow in A) as well as scattered cells in the neuropil, which are presumably glial cells (arrowheads in A). In larvae, Drostar1 signals were detected in few cell bodies in the central nervous system (arrowheads). C, section of a larvae. Immunostaining of cells in brain (arrowheads) and corpora allata (ca) using Manduca pGlu-allatostatin C antibodies. lam, laminar; bars, 50 m. those for D. melanogaster or Caenorhabditis elegans, it may therefore in many cases become necessary to identify the physiological ligands of novel receptors by the reverse pharmacology approach that has been used here.