A Drosophila gene that encodes neuropeptides related to molluscan Phe-Met-Arg-Phe-NH2 (FMRFamide) was isolated by screening a genomic library with a fragment of an Aplysia Phe-Met-Arg-Phe-NH2 cDNA and with synthetic oligonucleotides. This gene was used to isolate a cDNA from a Drosophila adult head cDNA library. The cDNA was defined by sequence analysis to encode 13 peptides that have Phe-Met-Arg-Phe-NH2 or related sequences at their carboxyl termini. Other putative neuropeptides, including one that has homology to mammalian corticotropin-releasing factor, are present in the deduced approximately equal to 39-kDa precursor. Southern blot analysis suggested the presence of a single Phe-Met-Arg-Phe-NH2-like gene within the haploid genome. RNA blot analysis indicated the expression of at least two transcripts of approximately equal to 1.7 and approximately equal to 0.7 kilobases. Both transcripts are evident throughout larval, pupal, and adult developmental stages. In situ hybridization was used to localize this neuropeptide gene to band 46C on the right arm of the 2nd chromosome. These data provide the basis for utilizing the advanced genetics and molecular techniques of Drosophila to address complex aspects of neuropeptide expression and function.

We have mapped protein expression of the FMRFamide neuropeptide gene in Drosophila with polyclonal antisera against three small peptides whose sequences were derived from the Drosophila proFMRFamide precursor. One antiserum was affinity-purified and extensively characterized. The enriched antibodies labeled 15-21 bilaterally symmetric pairs of neurons in a pattern that corresponded very closely to the pattern of in situ hybridization that was determined previously (Schneider et al. [1991] J. Comp. Neurol. 304:608-622; O'Brien et al. [1991] J. Comp. Neurol. 304:623-638). The other antisera produced complementary results. These findings suggest that the antisera specifically label cells that express the FMRFamide gene. In larvae we consistently observed strong staining in identified interneurons and neuroendocrine cells, and moderate to weak staining in neurons of unknown function. The adult pattern of expression included both larval neurons whose immunoreactivity persisted through metamorphosis and adult-specific neurons. During metamorphosis, we observed transient staining in a small number of neurons and in specific neuropil regions that included the central body, the protocerebral bridge, and the optic ganglia. Based on these morphological features, we suggest that the FMRFamide-like neuropeptides in Drosophila play a number of functional roles, perhaps affecting both physiological and developmental phenomena. Such roles include general modulation throughout all post-embryonic stages, via the blood, and also more stage- and region-specific modulation within the CNS.

No abstract

In order to identify functionally important regions of a neuropeptide gene in Drosophila melanogaster, we have studied its occurrence in related species and have characterized the structure of a homologous gene in Drosophila virilis. The melanogaster gene encodes a precursor that contains 13 neuropeptides related to the molluscan tetrapeptide FMRFamide (Nambu et al., 1988; Schneider and Taghert, 1988). Using the melanogaster gene as a probe in Southern blot analysis, related sequences were detected in DNA from each of 7 species tested. D. virilis, which is estimated to have diverged from D. melanogaster between 60 and 80 million years ago (Throckmorton, 1975), was chosen for more detailed study. Immunocytochemical staining using an antibody to authentic FMRFamide revealed a similar set of immunoreactive neurons in the CNS of larvae from the 2 Drosophila species. Using a melanogaster gene probe, overlapping clones were isolated from a virilis genomic library; DNA sequence analysis indicated the presence of a homologous gene. Comparisons of the genes and deduced proteins between the 2 species revealed the following points. (1) Both genes are divided into 2 exons: in D. melanogaster the exons are 106 and 1352 bp long; in D. virilis, they are 169 and at least 1232 bp long; in both species, the intron is approximately 2.5 kb long. (2) The sequence of exon I has largely diverged, and in neither species are exon I sequences translated. In this vicinity of the gene, sequence conservation is limited to a 67 bp region that spans the TATA box and the RNA start site. (3) The deduced neuropeptide precursors have very similar sizes (347 vs 339 amino acids) and the presumed signal sequences are perfectly conserved. (4) While the melanogaster precursor contains 13 FMRFamide-related peptides, the virilis precursor contains only 10. (5) The sequences of some but not all of the FMRFamide-like peptides are perfectly conserved. (6) In the rest of the precursor, significant sequence conservation is found only in the N-terminal portion; immediately downstream of the final FMRFamide-like peptide, the protein sequences are highly divergent. (7) 5' to the RNA start sites (1.2 kb of melanogaster DNA and 1.8 kb of virilis DNA), 17 small (9-52 base pairs) regions are evolutionarily conserved (greater than 80% sequence conservation). We discuss neuropeptide biosynthesis, the functions and evolution of FMRFamide-like neuropeptides in insects, and the cell-specific regulation of neuropeptide gene expression in the contexts of these results.

We have studied changes in the pattern of specific neuropeptide gene expression during the metamorphosis of the Drosophila nervous system. Prior to metamorphosis, the Drosophila FMRFamide gene is expressed exclusively within the central nervous system in a stereotyped pattern that comprises roughly 60 neurons (Schneider et al., '91). Using in situ hybridization, we found that the FMRFamide gene was continuously expressed throughout all stages examined: at each of 15 stages of adult development and through at least the first 10 days of adult life. There were no differences between the results observed with 2 exon-specific hybridization probes, thus indicating little if any alternative splicing during postembryonic development. Despite many changes in the positions of individual hybridization signals due to the large-scale reorganization of the nervous system, the continuous pattern of gene expression through adult development permitted many adult signals to be identified as larval signals. We concluded that the adult pattern of FMRFamide gene expression was largely derived from persistent larval neurons. Adult-specific hybridization signals in the brain and ventral ganglion were also detected and these corresponded to many of the approximately 40 adult-specific FMRFamide-immunoreactive neurons. One specific larval signal was lost during adult development and the intensities of other signals fluctuated in reproducible manners. These stereotyped differences in hybridization signal intensity resemble similar observations made in larval stages (Schneider et al., '91) and support the hypothesis that the steady-state levels of FMRFamide transcripts are differentially regulated among the diverse neurons that express the gene.

We have used in situ hybridization techniques to describe the cellular distribution of transcripts from a Drosophila gene that encodes multiple FMRFamide-related neuropeptides. The Drosophila FMRFamide gene consists of two exons and is expressed predominantly as a approximately 1.7 kb RNA throughout postembryonic stages (Nambu et al., '88; Schneider and Taghert, '88, '90). We used exon-specific oligonucleotide probes to assay transcription in both embryonic and larval stages and found a pattern of hybridization signals that was restricted to the central nervous system and, within that tissue, was cell-specific. The pattern included 36 distinct signals distributed throughout both the brain and segmental nerve cord (ventral ganglion). These observations suggest that the cell-specific pattern of FMRFamide-like neuropeptide expression in the Drosophila CNS (White et al., '86; Taghert and Schneider, '90) is due to the restricted expression of specific gene transcripts. The results also indicate that, with few exceptions, all previously identified FMRFamide-immunoreactive neurons in Drosophila larvae express FMRFamide gene transcripts. The 36 hybridization regions of the CNS could be divided into three categories, based on their signal intensities (strong, moderate, and weak). The differences in intensity were reproducible and suggest that steady-state levels of specific neuropeptide RNA differ among individual neurons. The two exon-specific probes produced patterns that were indistinguishable both in pattern and in intensity. This result supports the previous conclusion that the one detectable FMRFamide transcript contains both exons (Schneider and Taghert, '90). A single identifiable signal was detected during embryogenesis (beginning at stage 16), but the mature complement of signals was not fully established until the final larval stages.

G protein-coupled receptors mediate their biological responses through the generation of second messengers, such as cAMP. The down-regulation of their activity (desensitization) is carried out, in part, by the family of G protein-coupled receptor kinases, which phosphorylate activated receptors. The Gprk2 gene in Drosophila melanogaster is a putative member of this family. The GPRK2 protein is expressed most abundantly in the ovaries and in the mushroom bodies, the brain region that is implicated in learning and memory in insects. Many of the genes that are involved in learning in Drosophila are members of a cAMP-signaling pathway and are also expressed in the mushroom bodies. These observations suggest that the Gprk2 gene may be involved in a cAMP-mediated pathway. To investigate this possibility, we tested for a genetic interaction between Gprk2 and dunce (which encodes cAMP-specific phosphodiesterase). A mutant allele of Gprk2, called gprk2(6936), has decreased fertility as a result of reduced levels of egg laying and hatching, and developing egg chambers display defects in the formation of anterior structures. Similarly, many alleles of dunce are sterile, with an ovary phenotype that resembles gprk2(6936). Introduction of a single copy of a hypomorphic or null allele of dunce into the gprk2(6936) background suppressed all of these defects to a significant degree. Suppression was also observed when a single copy of gprk2(6936) was introduced into a dunce background. Like mutants of rutabaga (which encodes a calcium/calmodulin-dependent adenylate cyclase), gprk2(6936) has reduced levels of cAMP. Ovaries from gprk2(6936) females contain about one third of the normal amount of cAMP. In addition, in every mutant combination where fertility is increased, cAMP levels are closer to wild type levels. These results suggest that Gprk2 is functioning in a cAMP-signaling pathway and that the underlying basis of the interaction between Gprk2 and dunce is a normalization of cAMP levels.

No abstract