|Home | About | Journals | Submit | Contact Us | Français|
In Drosophila melanogaster, fruitless (fru) encodes male-specific transcription factors (FRUM; encoded by fru P1) required for courtship behaviors [reviewed in 1]. However, downstream effectors of FRUM throughout development are largely unknown [2-5]. During metamorphosis the nervous system is remodeled for adult function, the timing of which is coordinated by the steroid hormone 20-hydroxy ecdysone (ecdysone) through the ecdysone receptor, a heterodimer of the nuclear receptors EcR (isoforms are EcR-A, EcR-B1, or EcR-B2) and Ultraspiracle (USP) [reviewed in 6]. Here, we show that genes identified as regulated downstream of FRUM during metamorphosis are significantly overrepresented with genes known to be regulated in response to ecdysone or EcR. FRUM and EcR isoforms are co-expressed in neurons in the CNS during metamorphosis in an isoform-specific manner. Reduction of EcR-A levels in fru P1-expressing neurons of males caused a significant increase in male-male courtship activity and significant reduction in size of two antennal lobe glomeruli. Additional genes were identified that are regulated downstream of EcR-A in fru P1-expressing neurons. Thus, EcR-A is required in fru P1-expressing neurons for wild type male courtship behaviors and the establishment of male-specific neuronal architecture.
Differences in transcript abundance between fru P1 mutant and wild type Drosophila males were examined using microarrays [7, 8] in whole pupae and CNS tissues at the 48-hour after puparium formation (APF) stage (pupal FRUM- and CNS FRUM-regulated sets, Table S1), when FRUM is at highest abundance . Transcripts from two fru P1 genotypes (fru440/P14 and w; fruw12/ChaM5) were compared to two wild-type strains (Canton S and w Berlin, respectively). Each experiment included at least four replicates, and all showed high correlations and similar numbers of expressed genes (Table S2). Transcripts from 236 and 94 genes show significant expression differences in the pupae or CNS, respectively (q < 0.15, moderated t-test, Table S3).
Genes functioning in the ecdysone pathway were significantly overrepresented in the gene sets identified as regulated downstream of fru P1, based on comparisons to previously identified ecdysone-regulated gene sets (Table 1) [10-12]. This includes genes regulated downstream of EcR in pupae (129 genes), genes regulated downstream of ecdysone in larval organ cultures (30 genes), and genes known to function in the ecdysone pathway (8 genes) (1.4, 1.8 and 5.4 fold-enrichment over what is expected at random, respectively, p < 0.0002, Tables Tables1,1, S4 and S5). Additional microarray experiments examining expression when a FRUM transgene with either the A, B, or C DNA binding domain was over-expressed identified 156, 116, and 109 genes, respectively (q < 0.05, fold-difference > 2, Table S6); each of these sets was significantly overrepresented with genes regulated downstream of EcR or ecdysone (Table S7) .
Given this observation, the presence of ecdysone receptor binding sites (EcREs) in regulatory sequence was determined for genes regulated downstream of fru P1. A significant overrepresentation is observed, with 92/317 genes containing an EcRE (1.2 fold-enrichment over what is expected at random, p = 0.014, hypergeometric test). The regulatory region of the fru locus contains three EcREs, which together with the observation that fru was found to be expressed downstream of EcR in pupae  suggests that ecdysone receptor may regulate fru P1, ultimately regulating expression of some downstream target genes.
EcR isoforms have distinct temporal and spatial expression patterns in the CNS (Figure S1) . Examination of co-expression patterns of EcR-A and EcR-B1 with FRUM demonstrated that EcR isoforms are in few FRUM-expressing cells in white pre-pupal (wpp) brains (Figure S2). By 48 hour APF and through the adult stage, no co-localization is observed between EcR-B1 and FRUM in the CNS (Figures (Figures11 and S2). All of the examined FRUM-expressing brain neurons, and most ventral nerve cord (VNC) neurons, co-express EcR-A at 48 hour APF and 0-24 hour adult stages in regions previously described (Figure 1 and S2) [9, 14]. Several cells in the abdominal ganglion express FRUM, but not EcR-A.
EcR function was reduced in fru P1-expressing neurons and male behaviors were assayed by the courtship index (CI) and wing extension index (WEI, Figure 2). EcR function was reduced by expressing EcR RNA interfering (UAS-IR[EcR]) or EcR dominant negative transgenes (UAS-DN[EcR]) via activation by fru P1-GAL4 (Figure 1A) . Males of one experimental genotype showed a small, but significant reduction in male-female WEI (p < 0.05, ANOVA and post-hoc u-tests), but no flies of experimental genotypes had significant differences in CI. Males of several experimental genotypes showed significant and substantial increases in male-male WEI and CI (p < 0.05, ANOVA and post-hoc u-tests). This effect was due to reduction of EcR-A, but not EcR-B1, function. Similar results were obtained using a different fru P1-GAL4 strain (Tables S8-S9) . Given that UAS-IR[EcR-A] and UAS-IR[EcR-B1] transgenes were reported to be similarly efficacious at reducing protein levels, this suggests that EcR-A is the predominant isoform functioning in fru P1-expressing neurons to establish wild type courtship behaviors. Although potential functions of EcR-B isoforms cannot be ruled out, we demonstrate EcR-A is necessary within fru P1-expressing neurons for wild type courtship behavior.
When the UAS-DN[EcR]-3 transgene was expressed to reduce EcR function within fru P1-expressing neurons of males during development, adulthood, or all stages, a significant increase in male-male courtship was observed with transgene expression during development or all stages (p < 0.05, u-tests, Figure 2G, Tables S8-S9). Reduction of EcR function throughout all stages caused higher levels of male-male courtship than reduction only during development. Male-male courtship did not significantly increase when UAS-DN[EcR]-3 expression was restricted to the adult stage. This indicates that the male-male courtship phenotype is primarily due to reduction of EcR function in fru P1-expressing cells during development, and is enhanced by combined reduction during adulthood.
A study that examined males with a temperature-sensitive EcR allele demonstrated that male flies shifted to the non-permissive temperature at adult stages displayed increased male-male courtship . The apparent difference between the results might be explained by an additional adult requirement for EcR outside of fru P1-expressing cells. Also, the temperature-sensitive allele of EcR might have caused developmental phenotypes in fru P1-expressing cells at the permissive temperature, which were enhanced by shifting to the non-permissive temperature at the adult stage.
fru P1-expressing olfactory receptor neurons (fru P1-ORNs) are required for the sexually dimorphic size of antennal lobe glomeruli (DA1, VA1lm, and VL2a) . Because mate discrimination has an olfactory component, EcR isoform co-expression within fru P1-ORNs was examined at 48-hour APF. All fru P1-ORNs co-express EcR-A and EcR-B1, each having similar levels of expression of the respective EcR-isoform, though expression of EcR-A appears higher than EcR-B1 throughout the antennal segment (Figures (Figures33 and S2). All glomeruli innervated by fru P1-ORNs in the adult are innervated by 48 hour APF (Figure 3), suggesting that all third antennal segment fru P1-ORNs innervating antennal lobe glomeruli also express EcR.
Antennal lobe morphology was examined in flies where fru P1-GAL4 drove expression of UAS-EcR transgenes that reduce EcR function, and volumetric analyses were performed (Figure 3, Tables S10-S11). The volumes of two fru P1-ORN innervated glomeruli, DA1 and VA1lm, were significantly smaller when EcR-A, but not EcR-B1, function was reduced (p < 0.05, ANOVA and post-hoc t-tests, Figure 3). Flies expressing the transgene causing the highest levels of male-male courtship (UAS-DN[EcR]-3) had the smallest DA1 and VA1lm volume, which are the two glomeruli with the largest sexual dimorphism . No significant changes in size were seen for the VL2a and VA6 glomeruli, which are also innervated by fru P1-ORNs. This is perhaps because these glomeruli do not show substantial differences in volume between wild type males and females and the largest reduction in volume observed above never results in glomeruli smaller than what is observed in females [16, 18].
To determine if these decreases in glomeruli volume are due to a general requirement of EcR, and not a sex-specific effect, the volume of female glomeruli innervated by fru P1-ORNs was analyzed. Expression of the transgene with the largest effect in males (UAS-DN[EcR]-3) resulted in no reduction in volume (Figure 3), suggesting that in male fru P1-ORNs, EcR-A, which is not sex-specific, may act with FRUM, or some other male-specific factor, to affect glomeruli volume. The reduction of glomeruli volumes may also be due to effects from the fru P1-expressing projection neurons (PNs) innervating these glomeruli; all fru P1-expressing PNs examined express EcR-A (data not shown).
EcR-A levels were reduced in fru P1-expressing neurons in CNS tissues from wpp, 48-hour APF, and 0-24 hour stages, and gene expression was examined using microarrays (EcR-A/fru P1 gene sets, Tables S1-2). Previous studies have shown that two distinct high-titer ecdysone pulses coincide with wpp and 48-hour APF stages, but not the adult stage [reviewed in 6, 19]. Here, the numbers of differentially expressed genes correlate with levels of ecdysone hormone, with a higher number of genes identified during wpp and 48-hour APF stages (176 and 166 genes, respectively), as compared to adults (36 genes, Figure S4, Table S12).
Significant overlap between genes expressed downstream of fru P1 and the wpp and 48-hour APF EcR-A/fru P1 gene sets was observed (p = 0.01 and 0.019, respectively, hypergeometric test, Figure S4). Additionally, there was significant overlap among the wpp and 48-hour APF EcR-A/fru P1 sets and previously identified sets of genes regulated downstream of EcR (Table 1). The observation that more genes are regulated downstream of EcR-A during development, as compared to adult stages, is consistent with the behavioral experiments above that demonstrated male-male courtship phenotypes require a reduction of EcR-A function during development.
broad (br) is directly regulated by the ecdysone receptor [reviewed in 20] and is necessary for CNS development [21-23]. Here, br expression is significantly increased in wpp CNS in males with reduced EcR-A function in fru P1-expressing neurons (Table S12). Co-localization of BR with FRUM is observed in the CNS at three time points (wpp, 48-hour APF, and adult, Figure S4 and data not shown), further confirming that the ecdysone hierarchy functions in fru P1-expressing neurons during development.
This study demonstrates that the ecdysone hierarchy, in concert with FRUM, helps establish the neural circuitry required to prevent male-male courtship. Genes regulated downstream of FRUM are also regulated downstream of ecdysone or the ecdysone receptor, and contain an enrichment of EcREs in their regulatory sequence. EcR-A, but not EcR-B1, is expressed in many FRUM-expressing neurons throughout development, and reduced EcR-A in fru P1-expressing cells results in male-male courtship and reduced glomeruli volumes.
Males with reduced EcR-A in fru P1-expressing neurons display normal courtship behavior towards females, suggesting the neural circuitry is largely unaltered. Nevertheless, fine-scale morphological differences may underlie the male-male courtship phenotype, consistent with the observed reduction in volume of DA1 and VA1lm glomeruli. This difference may cause a defect in processing sensory information, such as the male-specific pheromone cis-vaccenyl acetate detected by ORNs that synapse on DA1 glomeruli . Here, flies of the genotype that showed the strongest male-male courtship phenotype showed the most substantial reduction in glomerulus volume. Identifying the causes of this male-male courtship phenotype, which our results suggest may be due to deficits in antennal lobe glomeruli, but may also be explained by additional sensory or higher-order processing defects, will provide insight into how neural substrates that underlie complex behaviors develop.
Our data suggest that ecdysone, through EcR-A, provides temporal input to the development of the spatially restricted fru P1-expressing neurons, directing the precise timing of sex-specific development. EcR-A appears to primarily function in fru P1-expressing neurons during periods of high ecdysone titers, given that the reduction of EcR-A function results in more genes with transcriptional changes at stages of metamorphosis than the adult stage. Moreover, reduction of EcR during development, but not during adulthood, leads to increased levels of male-male courtship.
Three non-mutually exclusive models are proposed for how EcR and FRUM might coordinate to regulate gene expression (Figure S5). Worth noting is many of the genes identified here are likely to be indirect targets of EcR and FRUM. Genes may be regulated downstream of the EcR and FRUM in parallel pathways (Model A). Alternatively, gene regulation may occur in linear pathways, with EcR regulating FRUM and FRUM regulating gene expression, or vice versa (Models B and C). Evidence for the model in which EcR regulates FRUM is that during development EcR is detected in the CNS before FRUM, fru itself contains putative EcREs and fru is regulated downstream of EcR , whereas there is not similar data supporting FRUM regulating EcR or usp. Further investigation of the identified genes will provide insight into how these two independent genetic-regulatory hierarchies coordinate the large-scale changes that remodel the nervous system during metamorphosis, setting the stage for the performance of adult behaviors.
See supplemental materials.
We thank Garwin Chin, Brandon Ishaque Peter Liu, Jason Portillo and Rianna Wurzburger for their contributions to the experiments. Maria Spletter, and Liqun Luo for assistance with glomeruli images, and Marc Green and Susan Forsburg for assistance in deconvolution. We thank Eric Johnson and his lab (University of Oregon) for generously printing microarrays. We are grateful to all members of the Arbeitman lab for their help. The work was funded by NIH grant 1R01GM073039 awarded to MNA.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.