Highly Tissue Specific Expression of Sphinx Supports Its Male Courtship Related Role in Drosophila melanogaster

Sphinx is a lineage-specific non-coding RNA gene involved in regulating courtship behavior in Drosophila melanogaster. The 5′ flanking region of the gene is conserved across Drosophila species, with the proximal 300 bp being conserved out to D. virilis and a further 600 bp region being conserved amongst the melanogaster subgroup (D. melanogaster, D. simulans, D. sechellia, D. yakuba, and D. erecta). Using a green fluorescence protein transformation system, we demonstrated that a 253 bp region of the highly conserved segment was sufficient to drive sphinx expression in male accessory gland. GFP signals were also observed in brain, wing hairs and leg bristles. An additional ∼800 bp upstream region was able to enhance expression specifically in proboscis, suggesting the existence of enhancer elements. Using anti-GFP staining, we identified putative sphinx expression signal in the brain antennal lobe and inner antennocerebral tract, suggesting that sphinx might be involved in olfactory neuron mediated regulation of male courtship behavior. Whole genome expression profiling of the sphinx knockout mutation identified significant up-regulated gene categories related to accessory gland protein function and odor perception, suggesting sphinx might be a negative regulator of its target genes.


Introduction
The genetic basis of species-specific courtship behavior has been one of the major interests of evolutionary biology. Behavioral differences between Drosophila species have been well-documented, especially the ones that influence mate choices and have important fitness consequences [1]. But little study has been done to reveal whether there is any relationship between these phenotypic differences and lineage specific genes. Our recent study of sphinx has been one of the few studies that directly related novel behavior to newly evolved gene. Sphinx is a lineage specific chimeric gene [2,3] involved in regulating male courtship behavior [4]. The sphinx gene was formed by the insertion of a retroposed sequence of the ATP synthase F-chain gene (CG4692) from chromosome 2 into the 102F region of chromosome 4 (first exon), recruiting sequences upstream to form an intron and a second exon [2]. The sphinx gene appears to be functional because the gene contains indel polymorphisms only in the non-exonic sequences; it has a rate of evolution significantly above neutral expectations, suggesting rapid adaptive evolution. However, although it is derived, in part, from a protein-coding gene, it is most likely a noncoding RNA (ncRNA) because its parental-inherited coding regions are disrupted by several nonsense mutations [2].
We previously showed that knocking-out of this gene led to increased male-male courtship in D. melanogaster, while leaving other aspects of mating behavior unchanged [4]. Comparative studies of courtship behavior in other closely related Drosophila species suggested that this mutant phenotype of male-male courtship was the ancestral condition, since these related species showed much higher levels of male-male courtship than D. melanogaster. The recruitment of sphinx in D. melanogaster therefore, might have increased male-female mating by suppressing malemale courtship behavior [4].
Male courtship in Drosophila is an elaborate ritual involving multiple sensory inputs with olfactory and/or gustatory stimuli being particularly important during mate recognition [5,6]. In flies, different subsets of olfactory receptor neurons (ORNs) of the olfactory appendage, the antenna and the maxillary palps project axons to different functional processing units called glomeruli in the antennal lobe (AL). The AL is the primary olfactory association center in insects where ORNs synapse onto second order neurons called projection neurons (PNs). An essentially complete olfactory map has been constructed by large-scale genetic efforts to label ORNs expressing each of the 62 known OR genes and map their projections to approximately 50 morphologically defined glomeruli in the adult AL [7,8]. The axons of PNs project to the mushroom body (MB) and lateral horn via inner antennocerebral tract (iACT) [9,10]. The organization of gustatory system is more dispersed than olfactory system. The main taste organs are the labial palps at the distal end of the proboscis, and the labral and cibarial sense organ inside the pharynx [11]. Gustatory receptor neurons in these sensilla project axons to the subesophageal ganglion(SOG) of the brain [12].
Here, we utilized a promoter GFP transformation system to dissect promoter region of sphinx and to investigate its expression pattern in relation to its function. We found that 1 kb upstream region of sphinx was able to drive GFP expression in accessory gland, and possibly peripheral and central nervous system, suggesting existence of putative promoter elements within this region. The highly tissue specific expression pattern also supported sphinx's reproductive related role. We carried out microarray analysis of a sphinx mutant to identify possible pathways in which sphinx might be involved. Results from our analysis suggest sphinx might function as a negative regulator in the courtship network.

Promoter region of sphinx
We identified several conserved elements in the 59 regulatory region of sphinx from multiple species syntenic alignment [4]. The very proximal conserved region encompassed ,300 bp upstream of the transcriptional start site and is very conserved out to Drosophila virilis (90% identical). The homology extends further upstream to around 2600 bp position with high sequence conservation within the melanogaster subgroup. An additional distal conserved element sits around 2850 to 21000 bp region ( Figure 1A).
To further investigate the evolutionary processes in the sphinx promoter region, we performed a polymorphism survey among world-wide populations. Consistent with the divergence data, the polymorphism level of the promoter region in D. melanogaster was very low. We sequenced 12 lines from a geographically diverse panel of populations and found only 7 segregating sites among the 1.3 kb region ( Figure 1B). The average nucleotide diversity (p) was estimated to be 0.00163, which was about half the level of diversity seen in the 4 th chromosome CG11091-toy region (0.0028), sequenced in the same lines. Thus the promoter region of sphinx seems to be under selective constraint.

Expression patterns of sphinx
To dissect the promoter region and understand the molecular functions of sphinx, we used Green Fluorescent Protein (GFP) as a reporter [13,14] to investigate the expression pattern of sphinx in various tissues. We generated two P-element derived constructs of Psphinx-GFP that contained GFP tagged 265 bp sequence of the proximal conserved region and 1067 bp (which includes all three conserved regions) of the 59 upstream genomic region of sphinx. By the standard p-element transformation procedure, we obtained 12 transgenic lines with a 265 bp fragment insertion and 6 lines with a 1067 bp insertion on different chromosomes. We examined GFP signals of all the transgenic lines at different developmental stages: embryo, three larvae stages, pupae ( Figure S1), and in a variety of tissues: head, wing, leg, testis, ovary, accessory gland, and brain ( Figure S2, S3). Consistent patterns of GFP expression were observed in the brain, accessory gland, wing and leg across all transgenic lines, with noted differences of expression in proboscis between the short and long promoter-GFP constructs (see below).
Anti GFP staining in the Drosophila brain showed a distinctive GFP signal in a pair of glomeruli of the antennal lobes (Figure 2). At a slightly different confocal plane, we observed signals in the inner antennal glomerular tract ( Figure 2A). To further identify the exact glomeruli in which GFP signal was visualized, we counter stained the brains of Psphinx_GFP lines with neurophile marker nc-82 ( Figure 2D). By comparing to the 3D reconstruction of the antennal lobe [15], we observed that the structure and positioning of GFP stained glomerulie is very similar to glomeruli VA2, which corresponds to Or92a projection in the AL. Odor ligands that activate Or92a and its corresponding glomerulus are carvone and octanal. The neuronal expression pattern of sphinx, however, did not appear to be sexually dimorphic, as there were little observable differences between male and female brains ( Figure 3).
The possible co-expression of sphinx and Or92a in VA2 suggests that sphinx might function in the same pathway as Or92a. Or92a has not previously been identified in courtship behavior, but it is one of the few OR genes that were found to be co-expressed with GR genes, in this case with Gr21a in ab1 of large basiconics of antenna [7]. Our microarray analysis (details see below) suggested that both Or92a and Gr21a genes were significantly up-regulated in the sphinx mutant with p-value of 0.014 and 0.017 respectively. Thus sphinx might play an integrative role between the olfactory and gustatory system.
A strong and consistent signal was observed in male accessory glands that are known to be involved in regulating male reproduction and courtship behavior [16] (Figure 4). There are two types of secretory cells: main cells and secondary cells. Only main-cell secretions are essential for the short-term inhibition of remating [17]. In the accessory gland, it was the main cells but not the secondary cells that show strong GFP expression ( Figure 4C, D). There was, however, no GFP expression in male testis and female ovaries. The endogenous expression of sphinx in male accessory gland was confirmed by antisense RNA in situ hybridization, which clearly showed that sphinx mRNA exists in the accessory gland and anterior ejaculation duct ( Figure 4A). We found very little or no sphinx mRNA in the testes and ovaries. This is consistent with the GFP expression pattern. The lack of GFP signal in ejaculation duct was probably due to fusion of GFP protein with target gene product making it difficult to be secreted from the accessory gland to the anterior ejaculation duct.
Furthermore, we observed GFP expression in chemosensory organs in both male and female adults: the bases of bristles on forelegs ( Figure 5A); and the bristles of the anterior margin of the wings ( Figure 5B) in all the lines. We also observed expression in the labral sense organ in the adult proboscis and the corresponding larval and pupae terminal sensory organ in the lines containing the long promoter insert ( Figure 5C, D, 6A, C) but no expression in the lines containing the shorter construct ( Figure 6B, D); Chemosensory organs are important in Drosophila pheromonal communication, especially in male perception of cuticular hydrocarbons during courtship [18,19].  Figure 6B). There was also no evidence of expression in larvae terminal sensory organ with the 265 bp insertion lines ( Figure 6D). These results implied that the promoter region of sphinx might have different functional units, with upstream 800bp carrying enhancer element that might be important for sphinx's expression in labral sense organ.
We were able to confirm sphinx expression in accessory gland by in situ hybridization and RT-PCR. We do not have independent evidence supporting the expression of the sphinx-promoter driven GFP expression in pheripheral and central nervous system due to technical difficulties. Yet previous behavior experiment and the consistent GFP expression patterns among replicate lines strongly suggest that the observed signals are real.

Whole genome expression profiling of the sphinx mutant
The expression of sphinx in accessory gland indicates sphinx's involvement in the male reproductive system, while the possible expression in chemosensory organs and AL suggests its possible neuronal function, suggesting the behavior of the sphinx mutant might involve expression changes in many other genes. Using Affymetrix microarrays, we compared expression profiles of the sphinx null mutant versus Oregon-R (mutant's genetic background) flies to search for candidate genes that might interact with sphinx. We identified differentially expressed genes as those that showed at least a 2-fold (up or down) difference in expression, with a p value of ,0.001. This resulted in a list of 84 up-regulated (Table S1) and 55 down-regulated genes (Table S2) in the sphinx mutant compared to Oregon-R. Obp99d was significantly up-regulated in the sphinx mutant (P,,0.001), which was probably not by accident as it was previously showed that the transcript abundance of this gene affected mating speed [20]. Obp56h and Elav were significantly down-regulated in the sphinx mutant, which might also potentially interact with sphinx. It was rather surprising that two female spermatogenesis genes (kelch and E2F) were among the down-reulated gene list. It might be that male and female sometime share the use of same genes during gametogenesis, or that sphinx might affect female reproductive behavior in aspects that have not been uncovered.
To further identify potentially important biochemical processes in a statistically rigorous way, we made use of the freely available software package CATMAP (http://bioinfo.thep.lu.se/Catmap)  [21]. This program assigns significance to gene categories based on their relative statistical ranking or representation within the data set. We ran Catmap analysis on the ranked gene list based on the Bayes t statistic of all genes for over-representation of functional categories from a number of biological databases, including GO and Interpro, and several customized databases that contain microarray data and functional classifications from previously published studies. The top lists of over-represented up-regulated gene categories in the sphinx mutant line are accessory gland protein and chemosensory and odorant receptor genes, while down regulated genes show no bias towards gene categories related to male reproduction or courtship (Table 1). One plausible explanation is that sphinx acts as a negative regulator in the biological processes related to accessory gland protein and neuropeptide secretion. Thus the inability of sphinx mutants to discriminate male from female might due to the fact that certain sensory circuits, which are normally only turned on in the presence of females, are being turned on constitutively.

Fly Strains
All fly strains, w1118 and GFP transformant lines, were kept at 25c on standard agar medium.

Sequence comparison
Homologous sequences of the sphinx gene and its upstream region were retrieved by running BLAST against the D.  After digestion by proper restriction enzymes, these fragments were inserted into a GFP expression vector pEGFP-N1. The chimeric pSphinx-GFP fragments were introduced in a P-element vector pCasper4. Microinjection was performed on w1118 (white eye) embryos. Successful transformants (red eyes) were screened from progenies of the crosses between the hatched injected flies to w1118 individuals. Red eye offspring were further inbred and purified for several generations until homozygous lines were established. GFP expression was visually examined using UV illumination with an Olympus BX60 stereomicroscope and fluorescence module.

Anti-GFP staining
Dissection and antibody staining of adult brain whole mounts was performed exactly as described in [15], using the nc82 antibody (kindly provided by Professor Reini Stocker), which was visualized with a 1:100 dilution of goat anti-mouse IgG coupled to CY3 (Jackson ImmunoResearch). Expression of Psphinx-GFP was visualized with a 1:1000 dilution of anti-GFP antibody (Molecular Probes) and a 1:250 dilution of goat anti-mouse secondary antibody coupled to Alexa Fluor 488 (Molecular Probes). Brains were mounted in Vectashield (Vector Labs) using small cover slips as spacers and analyzed with a LeicaSP5 2photon confocal microscope.

Oligonucleotide Microarray Analysis
Changes in transcript abundance were measured using D. melanogaster whole genome oligonucleotide microarrays 2.0 (Affymetrix). Total RNA was extracted from 5 days old adult male flies of oregon R and sphinx mutant sphinx720RW [4] by using Qiagen Rneasy mini kit according to the manufacturer's procedures. We performed three biological replicates of each genotype. All Affymetrix protocols were performed at the University of Chicago Functional Genomics Core Facility. The cRNA probe was generated by using standard Affymetrix protocols (www.affymetrix.com). Fragmented biotinylated probe was then hybridized to D. melanogaster whole genome arrays. Washing, labeling (streptavidin-phycoerythrin), and scanning followed standard procedures at the Core Facility.

Statistical Analysis
To calculate gene expression measures, the data sets were normalized as follows. Raw image files were converted to probe set data (.cel files) in Microarray Suite (MAS 5.0). The 20 probe set data files were normalized together, and expression values were determined, using the Robust Multichip Average method [22], implemented in the Affymetrix package (version 1.4.14) of the free statistical programming language R (www.r-project.org). We calculated a new t test value for all gene changes, using a more stringent two-tailed Student's t test and assuming unequal variance. We selected probes that showed at least a 2-fold (up or down) difference in expression and had a p value of ,0.001. This resulted in a list of 139 differentially expressed genes (84 upregulated, 55 down-regulated).
For Catmap analysis, a ranked gene list based on the Bayes t statistic from the Goldenspike [23] analysis was used as input. The Wilcoxon rank sum was used to generate a score based on the sum of the rankings of all genes with a particular functional annotation, and the significance of that score (the p value) was calculated analytically based on a random gene-rank distribution [21]. Gene categories were considered significantly differentially regulated at FDR (false discovery rate) ,0.01.