Reproduction in sexually dimorphic animals relies on successful gamete production, executed by the germline and aided by somatic support cells. Somatic sex identity in Drosophila is instructed by sex-specific isoforms of the DMRT1 ortholog Doublesex (Dsx). Female-specific expression of Sex-lethal (Sxl) causes alternative splicing of transformer (tra) to the female isoform traF. In turn, TraF alternatively splices dsx to the female isoform dsxF. Loss of the transcriptional repressor Chinmo in male somatic stem cells (CySCs) of the testis causes them to “feminize”, resembling female somatic stem cells in the ovary. This somatic sex transformation causes a collapse of germline differentiation and male infertility. We demonstrate this feminization occurs by transcriptional and post-transcriptional regulation of traF. We find that chinmo-deficient CySCs upregulate tra mRNA as well as transcripts encoding tra-splice factors Virilizer (Vir) and Female lethal (2)d (Fl(2)d). traF splicing in chinmo-deficient CySCs leads to the production of DsxF at the expense of the male isoform DsxM, and both TraF and DsxF are required for CySC sex transformation. Surprisingly, CySC feminization upon loss of chinmo does not require Sxl but does require Vir and Fl(2)d. Consistent with this, we show that both Vir and Fl(2)d are required for tra alternative splicing in the female somatic gonad. Our work reveals the need for transcriptional regulation of tra in adult male stem cells and highlights a previously unobserved Sxl-independent mechanism of traF production in vivo. In sum, transcriptional control of the sex determination hierarchy by Chinmo is critical for sex maintenance in sexually dimorphic tissues and is vital in the preservation of fertility.
Sexually dimorphic adult tissues, like ovaries and testes, require continuous sex-specific instruction for proper function. Establishment of female somatic sex identity in Drosophila is controlled by an alternative splicing cascade wherein Sex-lethal (Sxl) produces the female-specific protein TransformerF (TraF). By contrast, males lack Sxl and undergo default splicing, preventing TraF production. Since TraF expression in males causes sex transformation and impairs tissue function, males must have evolved robust protection against feminization. Here, we investigate the role of a single factor, Chinmo, in protecting male sex identity in the testis: loss of Chinmo in male somatic stem cells causes them to acquire female identity. We demonstrate that this feminization occurs through the induction of TraF and its downstream targets. Surprisingly, Sxl is not induced in these sex transformed cells. Instead, two other alternative splice factors, Virilizer and Female lethal (2)d, are enriched in chinmo-mutant somatic cells and are required for their feminization. Our work demonstrates that transcriptional repression of female-biased alternative splice factors prevents sex transformation in the somatic gonad and that traF production can occur independently of Sxl. Given the importance of sex maintenance in tissue homeostasis, such protective mechanisms may exist in other tissues.
Citation: Grmai L, Hudry B, Miguel-Aliaga I, Bach EA (2018) Chinmo prevents transformer alternative splicing to maintain male sex identity. PLoS Genet 14(2): e1007203. https://doi.org/10.1371/journal.pgen.1007203
Editor: Artyom Kopp, University of California Davis, UNITED STATES
Received: July 18, 2017; Accepted: January 16, 2018; Published: February 1, 2018
Copyright: © 2018 Grmai et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was funded by the National Institute of General Medical Sciences https://www.nigms.nih.gov/Pages/default.aspx (R01 GM085075) (EAB); the New York State Stem Cell Science (NYSTEM) https://stemcell.ny.gov/ (C028132) (EAB); the European Research Council Starting Grant https://erc.europa.eu/funding/starting-grants (310411) (IM-A); Medical Research Council intramural funding https://www.mrc.ac.uk/about/institutes-units-centres/ (IM-A); European Molecular Biology Organization Young Investigator Programme http://www.embo.org/funding-awards/young-investigators (IM-A); United Negro College Fund/Merck Science Initiative http://www.hbcuinnovation.org/my-product/uncf-merck-science-initiative/ (LG); and European Molecular Biology Organization Long-term Fellowship http://www.embo.org/funding-awards/fellowships/long-term-fellowships (BH). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Sexual dimorphism, or the differences between male and female individuals in a species, is observed in many organisms, including insects, reptiles, and mammals. Sex-specific tissue development is essential for proper gonadogenesis, and sexual dimorphism has also been observed in other tissues such as brain, adipose tissue, and intestine [1–4]. While extensive literature has dissected the mechanism of sex determination in early development, recent studies have demonstrated that maintenance of sex identity is also essential for adult tissue homeostasis [5–7]. It is therefore critical to determine the signals that both specify and maintain sex identity.
Differential gene expression via alternative splicing establishes the sex-specific differences observed in the fruit fly Drosophila melanogaster. In flies, the sex of an organism is determined by its number of X chromosomes [8–10]. In XX flies, a positive autoregulatory mechanism activates and maintains expression of the RNA-recognition motif (RRM) containing protein Sex-lethal (Sxl) . In female somatic cells, Sxl binds directly to a polyuridine (poly(U)) tract upstream of exon 2 in transformer (tra) pre-mRNA [12, 13]. This results in the skipping of exon 2, which contains an early stop codon, and synthesis of full-length Tra (TraF) in females. In XY flies, which lack Sxl, tra mRNA incorporates exon 2, resulting in premature translational termination and a presumptive small peptide with no known function . Several other factors have been shown to act in concert with Sxl in sex-specific alternative splicing, such as Virilizer (Vir), Female lethal (2)d (Fl(2)d), and Spenito (Nito). All three proteins have an RRM and are required for sex-specific and non-sex-specific functions in Drosophila [14–19].
One of the best characterized targets of the RNA-binding protein TraF is doublesex (dsx), which can yield one of two functional isoforms . In XX flies, TraF is required for the alternative splicing of dsx and fruitless (fru) pre-mRNAs, generating female-specific DsxF and preventing Fru synthesis [21, 22]. In XY flies, which lack TraF, dsx and fru pre-mRNA undergo default splicing and generate male-specific DsxM and FruM. The DsxF and DsxM transcription factors regulate the majority of known sex-specific differences in gene expression and external appearance in Drosophila, often by direct transcriptional regulation of critical sex-specific genes [20, 23, 24]. DsxF and DsxM have identical DNA binding sites and bind regulatory sites in many common target genes, and it is generally believed that Dsx isoform association with sex-specific co-factors determines whether the target gene is activated or repressed [20, 25–28].
Loss of sex identity in sexually dimorphic tissues has profound effects on organ development and function [1–4, 29–31]. In the gonad, sex identity is specified autonomously in both the germline and the soma; somatic gonadal cells additionally send essential non-autonomous cues to instruct germline sex identity [29, 31–34]. Proper gonadogenesis is impeded when the sex identity of the germline does not match that of the soma, and such a mismatch frequently causes sterility [31, 32]. Despite the importance of maintaining sex identity for tissue development and homeostasis, regulation of canonical sex determinants at the transcriptional level has remained relatively unexplored.
In Drosophila gonads, germline stem cells (GSCs) divide to produce daughters that ultimately differentiate into sperm and oocytes, respectively. Proper gametogenesis proceeds through the ensheathment of GSC daughters by somatic support cells that exhibit sex-specific differences. In the testis, a niche of quiescent somatic cells termed the hub supports GSCs and somatic cyst stem cells (CySCs), which produce somatic support cells (Fig 1A, left). GSCs divide with oriented mitosis, and daughter cells that are displaced from the niche differentiate through 4 rounds of transit-amplifying mitotic divisions. CySCs are the only mitotically active somatic cells in wild type testes, and they divide to produce post-mitotic cyst cells. Two cyst cells ensheath a single GSC daughter and remain associated with the germ cell cluster throughout its transit-amplifying divisions. During somatic differentiation, cyst cells grow dramatically to accommodate the enlarging spermatogonia [35–38].
(A) Schematic of the adult Drosophila testis (left) and ovary (right). In the testis, the niche (green) supports two populations of stem cells, germline stem cells (GSCs, dark pink) and somatic cyst stem cells (CySCs, dark blue). The GSC divides to produce differentiating daughter cells (light pink) that undergo four transit-amplifying divisions. The CySC divides to produce cyst daughter cells (light blue) that exit the cell cycle and ensheath the differentiating GSC daughter. Cyst cells continue to ensheath the associated spermatogonial cyst during transit-amplifying divisions. In the ovary, the niche (green) supports GSCs (dark pink), which divide to give rise to differentiating daughters (light pink) that undergo 4 mitotic divisions. The developing germline cyst is ensheathed by an epithelial layer of follicle cells (light blue). Follicle cells are proliferative descendants of follicle stem cells (FSCs, dark blue) located in the anterior part of the ovary. (B) Chinmo (green) is present in CySCs (B’, green arrowhead), GSCs (B’, magenta arrowhead), and niche cells (B’, cyan arrowhead) in a wild type testis. (C) Chinmo is not expressed in follicle cells of a wild type ovary. Vasa (red) marks the germline and Zfh1 (blue) marks somatic cells in the testis and ovary. Scale bars = 10 μm.
In the ovary, GSCs also divide to produce differentiating daughter cells that undergo 4 mitotic divisions to give rise to 16-cell interconnected germ cysts (Fig 1A, right). The developing germ cyst is surrounded by a layer of somatic follicle cells, which are produced by follicle stem cells (FSCs). CySCs and FSCs require similar self-renewal signals, and both male and female somatic gonadal cells exhibit similar cellular behaviors [39–50]. However, their differentiating offspring exhibit distinct behaviors and markers: cyst cells are quiescent as they differentiate, while follicle cells continue to cycle. Additionally, follicle cells form an epithelium to ensheath the germline, while cyst cells grow in volume and express tight junction proteins to encapsulate spermatogonia [35, 37, 38, 51–53].
Sex-specific anatomical differences are achieved by differential expression of transcription factors . In particular, the transcription factor Chinmo is expressed in male but not female somatic gonadal cells [29, 54, 55]. Chinmo contains a Broad, Tramtrack, and Bric-à-brac/Poxvirus and Zinc finger (BTB/POZ) domain and two C2H2-Zinc fingers (ZFs). Many BTB-ZF proteins in Drosophila and mammals have characterized roles as transcriptional repressors [56–58]. However, while clonal loss of chinmo from imaginal tissue leads to ectopic gene expression in a cell-autonomous manner , no direct targets of Chinmo have been identified. Congruent with its dimorphic expression in the somatic gonad, chinmo has no apparent requirement in follicle cells but is essential for CySC niche occupancy [54, 55]. Chinmo is also required for the maintenance of male sex identity in CySCs, as loss of chinmo from all CySCs causes them to lose male sex identity, express markers of ovarian follicle cells and adopt an epithelial-like organization . These data have led to a model in which single CySC clones lacking chinmo are outcompeted by wild type CySC neighbors, but chinmo depletion in all CySCs removes this competitive environment and leads to sex transformation [29, 54]. We have also observed that chinmo-mutant CySC clones that lack the JAK/STAT and EGFR pathway inhibitor Socs36E can form aggregates, suggesting that CySCs lacking chinmo can feminize so long as they are given a chance to proliferate . This sex transformation was reportedly due in part to a transcriptional loss at the dsx locus, leading to a loss of DsxM; however, sustained expression of UAS-dsxM could not prevent the acquisition of female sex identity in chinmo-mutant CySCs, indicating that the molecular mechanism by which these cells feminize is still unclear .
Our work supports an alternate model whereby male sex identity is maintained not by preventing transcriptional loss of dsxM, but by preventing alternative splicing of dsx pre-mRNA into dsxF. Since TraF is responsible for dsx alternative splicing in canonical sex determination, we investigated a possible role for TraF in CySC feminization upon chinmo loss. Here, we report that Chinmo maintains male sexual identity by preventing the expression of the female sex determinant traF through a two-step mechanism. We first show that Chinmo represses both expression and alternative splicing of tra pre-mRNA. Next, we demonstrate that feminization of chinmo-mutant CySCs does not require Sxl. We instead find that RNA binding proteins Vir and Fl(2)d, which are necessary to alternatively splice traF in the adult ovary, are important for the feminization of chinmo-mutant CySCs. Thus, we uncover a novel mode of sex maintenance involving previously unreported regulation of tra transcription and a Sxl-independent mechanism of traF splicing in the somatic gonad.
Chinmo is expressed dimorphically in the somatic gonad and is required for male identity in CySCs
We found dimorphic expression of Chinmo in Drosophila gonads. While Chinmo protein was expressed in all cell types of the adult testis stem cell niche (Fig 1B and 1B’, arrowheads; S1I Fig), it was not detectable in somatic cells of the adult ovary (Fig 1C; S1J Fig). We next confirmed that loss of Chinmo expression in CySCs leads to the acquisition of female identity. When chinmo was depleted in the CySC lineage by RNAi using the somatic driver tj-gal4 (tj>chinmoRNAi; S1K Fig), expression of the male sex determinant DsxM was lost (Fig 2A–2C), and the follicle cell marker Castor (Cas), normally absent from the testis, was ectopically expressed (Fig 2D–2F). In wild type testes, Fasciclin 3 (Fas3) was expressed in niche cells but not in CySCs (Fig 2G). However, in tj>chinmoRNAi testes, we observed Fas3-expressing somatic aggregates resembling epithelial follicle cells that eventually organized at the periphery (Fig 2H and 2I). A marker of late-stage follicle cell maturation, Slow border cells (Slbo), was absent from wild type CySCs (S1A and S1B Fig) but was ectopically expressed in tj>chinmoRNAi testes (S1C Fig). Finally, transcripts of the DsxF target Yp1 were upregulated in chinmo-deficient testes (S1D Fig; . This sex transformation phenotype is due to loss of chinmo in the CySC lineage and not the niche, as depletion of chinmo specifically in niche cells produced no overt phenotype (S1E and S1F Fig; ).
(A) In a wild type testis, DsxM (green) is present in niche cells (A’, outlined with dotted green line), CySCs (A’, arrowheads) and cyst cells (A’, arrow). (B) DsxM is not expressed in a wild type ovary (B’, arrows indicate early follicle cells). (C) DsxM is lost in the CySC lineage in a tj>chinmoRNAi testis (C’, arrows indicate early somatic cells). (D) Castor (Cas, green) is absent from a wild type testis (D’, arrows indicate early somatic cells). (E) In a wild type ovary, Cas (green) is expressed in early follicle cells (E’, arrowhead) and stalk cells (E’, arrow). (F) Cas (green) is ectopically expressed in feminizing somatic cells in the testis (F’, arrows) upon loss of Chinmo. (G) Fasciclin-3 (Fas3, green) is normally restricted to niche cells in a wild type testis. (H) In a wild type ovary, Fas3 (green) is high in early follicle cells (H’, arrowhead) and is lower in mature follicle cells (H’, arrow). (I) Fas3 (green) is ectopically expressed upon loss of Chinmo in CySC lineage. (J) Relative fertility of tj>chinmoRNAi males (blue bars) is decreased at 9, 16 or 23 days (d) post eclosion compared with control tj>+ males (white bars). tj>chinmoRNAi males are completely sterile by 23 days post eclosion (* denotes p<0.05; *** denotes p<0.001; **** denotes p<0.0001 as determined by single-factor ANOVA). Error bars represent SEM. In A-I, Vasa (red) marks the germline and Tj (blue) marks cyst cells. Scale bars = 20 μm. Time point in A-I is 7 days post-eclosion.
Because CySCs serve a critical role in maintaining GSCs, as well as producing somatic support cells, the stem cell niche in tj>chinmoRNAi testes frequently becomes agametic even at relatively early time points after depletion (S1G and S1H Fig; [29, 60]). Based upon these observations, we hypothesized that tj>chinmoRNAi males would become sterile. To test this, we mated successively tj>chinmoRNAi males to OregonR virgin females and scored the number of progeny. Upon each of two mating rounds, tj>chinmoRNAi males exhibited a significant reduction in fertility (25% and 55% compared to control males, p<0.05 and p<0.001, respectively) (Fig 2J). By the third successive mating, tj>chinmoRNAi males were completely sterile whereas control males were not (p<0.0001). Taken together, our results align with previous work showing that Chinmo is required in adult CySCs to preserve male sex identity . Additionally, we demonstrate that CySC male identity is essential for fertility.
CySC feminization upon loss of chinmo is dependent on the female sex determinant dsxF
We next sought to determine the mechanism by which CySCs undergo feminization upon loss of chinmo. According to a previous report, dsxM mis-expression in chinmo-deficient CySCs (c587>chinmoRNAi; >dsxM) delays feminization , suggesting that dsxM transcription was reduced in chinmo-mutant CySCs. However, at the time point when all c587>chinmoRNAi testes contained Fas3-positive aggregates, nearly all c587>chinmoRNAi; >dsxM testes were also feminized , indicating a delay but not an abrogation of the phenotype. Additionally, depletion of all dsx transcripts using an RNAi transgene (dsxKK111266) targeting the common region of dsxM and dsxF did not recapitulate the defect seen upon loss of chinmo . These data indicate that the loss of DsxM alone cannot fully account for the phenotype of chinmo-mutant CySCs. We reasoned that the loss of DsxM protein observed in chinmo-mutant CySCs could result from alternative splicing of the dsx pre-mRNA into dsxF rather than from a transcriptional decrease at the dsx locus (Fig 3A). If this were true, we would expect to find in chinmo-deficient CySCs: 1) active transcription of the dsx locus; 2) expression of alternatively-spliced dsxF transcripts; and 3) expression of DsxF protein. To assess dsx transcription levels, we surveyed 4 independently-generated dsx transcriptional reporters: two Gal4 knock-in reporters in the dsx locus (dsx-gal4;  and dsx-gal4Δ2; ), one MiMIC allele at the dsx locus (dsxMI03050-GFSTF.1; ) and one Janelia transgene containing a 2.5 kb dsx regulatory element (GMR40A05-gal4; . We selected dsx-gal4 for further use because it was the only line that was robustly expressed in both adult testes and adult ovaries and therefore accurately reflected dsx transcription (Fig 3B and 3C). By contrast, the other 3 lines displayed male-biased or very low expression in gonads (S2A–S2F Fig). We then assessed dsx-gal4 activity as a proxy for transcription of the dsx locus upon chinmo depletion. We used a genetic approach to remove chinmo from the CySC lineage by analyzing testes homozygous for the chinmoST allele  in the dsx-gal4 background. While chinmoST/CyO testes express normal levels of Chinmo, chinmoST/chinmoST males lack Chinmo in the CySC lineage . As expected, GFP was expressed in somatic cells in control chinmoST/CyO; dsx-gal4/UAS-GFP testes and ovaries (Fig 3B and 3C). Importantly, GFP was also expressed in chinmoST/chinmoST; dsx-gal4/UAS-GFP mutant testes (Fig 3D), demonstrating that dsx is still transcribed in chinmo-deficient cyst cells. We also visualized dsx transcript abundance in tj>chinmoRNAi testes using semi-quantitative RT-PCR. Primers that recognize both dsx mRNA isoforms (dsxCOMMON, or dsxC) reveal that dsx is still present in tj>chinmoRNAi testes (Fig 3E). We confirmed that dsxF is produced specifically by chinmo-deficient somatic cells by performing RT-PCR on FACS-sorted CySCs and early cyst cells. As expected, a dsxF-specific band was observed in RNA extracts from wild type ovaries (Fig 3F, left lane). We also observed dsxF in FACS-purified chinmo-deficient cyst cells (Fig 3F, right lane). As expected, dsxF was absent from FACS-purified wild type cyst cells (Fig 3F, middle lane).
(A) Schematic of dsx pre-mRNA splicing. dsxF and dsxM share the first three coding exons and differ at their C-termini. Exon 4 contains a non-canonical splice acceptor, and thus under default splicing conditions, exon 3 is adjoined with exon 5 to yield the male-specific dsxM isoform (blue). In XX somatic cells, TraF is expressed and binds tandem TraF-binding sites (white dashed line) in exon 4. The TraF complex then recruits the spliceosome, leading to synthesis of the female-specific dsxF isoform (pink). Pink and blue stars indicate female and male stop codons, respectively. (B) A knock-in transcriptional reporter for dsx (dsx-gal4) activates UAS-GFP expression (green) in the CySC lineage of a wild type testis. (C) dsx-gal4 activates GFP expression in both escort cells (C’, arrow) and follicle cells (C’, arrowheads) of wild type ovary. (D) dsx-gal4 activates GFP expression (D’, arrowheads) in the somatic lineage of a chinmoST/chinmoST testis. Note that the feminized soma in this testis is organized at the periphery, adjacent to the muscle sheath. (E) Semi-quantitative RT-PCR on homogenized wild type testes (left lane), wild type ovaries (middle lane), and tj>chinmoRNAi testes (right lane). As determined by primers that recognize both dsx mRNA isoforms (dsxCOMMON, or dsxC), dsx transcripts are present in both wild type testes and wild type ovaries (left and middle lanes). dsx mRNA is also expressed in tj>chinmoRNAi testes (right lane). α-tub (tub) was used as a loading control. Flies were aged 9–20 days prior to dissection. (F) Semi-quantitative RT-PCR on RNA extracts from FACS-purified cyst cells from control tj>+ or tj>chinmoRNAi testes. Whole adult ovaries from yw females (labeled “WT ovary”) were homogenized as a positive control for dsxF mRNA detection. dsxF is detected in wild type ovaries (left lane) and in tj>chinmoRNAi cyst cells (right lane) but not in tj>+ cyst cells (middle lane). β-tubulin (tub) was used as a loading control. (G,H) Dsx (green), as detected by the DsxC antibody that recognizes both DsxF and DsxM, is present in somatic cells of a wild type testis (G) and a tj>chinmoRNAi testis (H). Tj (magenta) marks cyst cells. In B-D, Vasa (red) marks germ cells and Tj (blue) marks cyst cells. Scale bars = 20 μm.
We next visualized Dsx protein in control tj>+ and tj>chinmoRNAi testes using an antibody that detects both isoforms of Dsx (anti-DsxC; ). We observed that Dsx protein is still synthesized in somatic cells lacking chinmo (Fig 3G and 3H); because DsxM is lost from tj>chinmoRNAi testes (Fig 2C), we conclude that the Dsx protein present in chinmo-deficient somatic cells is DsxF. We confirmed that anti-DsxC detects DsxF by staining tj>dsxF ovaries (S3 Fig). These results suggest that DsxM loss in chinmo-deficient CySCs is not due to transcriptional loss of dsxM, but rather alternative splicing that generates the female isoform DsxF.
We next tested whether DsxF production is causal to feminization of CySCs lacking chinmo. We took a genetic approach and blocked dsxF splicing by using mutant alleles dsxD/dsx1. dsxD cannot be alternatively spliced into dsxF but produces normal levels of dsxM, and dsx1 is a null allele. dsxD/dsx1 flies only produce DsxM. XX dsxD/dsx1 animals develop male abdominal pigmentation, genitalia, and sex combs due to a masculinized soma (S4A–S4D Fig; ). We introduced dsxD/dsx1 into males homozygous for the chinmoST allele . As expected, control chinmoST/CyO; dsxD/dsx1 sibling testes appeared normal (Fig 4A; Fig 4D, second bar; S1 Table). By contrast, 100% of chinmoST/chinmoST; TM2/TM6B testes at 7 days post-eclosion contained Fas3-positive somatic aggregates outside the hub (Fig 4B; Fig 4D, first bar; S1 Table). Strikingly, only 57% of chinmoST/chinmoST; dsxD/dsx1 testes contained Fas3-positive aggregates (Fig 4C; Fig 4D, purple bar; S1 Table), a significant reduction compared to chinmoST/chinmoST flies (p<0.001). Taken together, our results reveal that (1) acquisition of female identity in chinmo-mutant CySCs occurs by dsx alternative splicing that generates DsxF and (2) dsxF production is required in part for CySC feminization upon loss of chinmo.
(A-C) Representative confocal images of testes from chinmoST/chinmoST; dsxD/dsx1 flies. In a testis from a chinmoST/CyO; dsxD/dsx1 sibling male, only niche cells express Fas3 (A’). By contrast, in a chinmoST/chinmoST testis most of the somatic lineage is positive for Fas3 (B’). In a testis from a “rescued” chinmoST/chinmoST; dsxD/dsx1 male, Fas3 is again restricted to niche cells (C’). Results are quantified in Fig 4D and S1 Table. Vasa (red) marks germ cells and Tj (blue) marks cyst cells. Scale bars = 20 μm.(D) Quantification of Fas3-positive aggregates as a readout of feminization. 100% of chinmoST/chinmoST; TM2/TM6B testes contain Fas3-positive aggregates (dark blue bar) at 7 days post-eclosion. By contrast, none of the testes from chinmoST/CyO; dsxD/dsx1 sibling males contain aggregates (second bar). There is a significant reduction in the percentage of testes from chinmoST/chinmoST; dsxD/dsx1 (purple bar) or chinmoST/chinmoST; tra1/Df(3L)st-j7 (green bar) containing Fas3-positive aggregates, indicating a partial rescue of the feminization. However, testes from the various sibling controls are still feminized (light blue and yellow bars). Sample sizes are indicated within bars. *** denotes p<0.001 and **** denotes p<0.0001 as determined by two-tailed Student’s t-test (for qRT-PCR, compared with tj>+ controls) or Fisher’s Exact Test (for quantifications, compared with chinmoST/chinmoST). n.s. means not significant. See S1 Table for percentage values.
TraF is required for feminization in CySCs lacking chinmo
Given that DsxF is produced in chinmo-deficient CySCs, these cells must also express a factor that promotes alternative splicing of dsx pre-mRNA. In female somatic cells, this alternative splicing is mediated by TraF . We hypothesized that tj>chinmoRNAi testes express ectopic TraF that produces dsxF. To test this, we performed semi-quantitative (Fig 5A) and quantitative RT-PCR (Fig 5B and 5C) analysis on the tra locus in tj>chinmoRNAi testes. We found that total tra mRNA abundance significantly increased (3.6-fold) in tj>chinmoRNAi testes compared with tj>+ testes (p<0.05) (Fig 5A, blue arrowhead; Fig 5B, compare blue to white bar). Furthermore, traF-specific primers revealed a 6.1-fold enrichment of traF mRNA in tj>chinmoRNAi testes compared to tj>+ controls (p<0.001) (Fig 5A, red arrowhead; Fig 5C, compare blue to white bar). These data demonstrate that the ectopic tra in chinmo-deficient CySCs is indeed spliced into the female traF isoform.
(A) Semi-quantitative RT-PCR of total tra and traF mRNA in control tj>+ testes (first lane), control tj>+ ovaries (second lane) and tj>chinmoRNAi testes (third lane). All three samples contain a band for total tra (first row, blue arrowhead). However, a traF band is detected in ovaries, as expected, and in tj>chinmoRNAi testes (second row, red arrowhead) but not in control tj>+ testes. rpl15 was used as a loading control (third row). Flies were aged 9–20 days prior to dissection. (B-C) qRT-PCR of total tra (B) or traF (C) in control tj>+ testes (white bars), control tj>+ ovaries (gray bars) and tj>chinmoRNAi testes (blue bars). There is significantly more total tra mRNA (B) and traF (C) in tj>chinmoRNAi testes compared to control tj>+ testes. Values represent the average of three biological replicates. * denotes p<0.05; *** denotes p<0.001 as determined by two-tailed Student’s t-test (compared with tj>+ testes). Error bars represent SEM. Flies were aged 9–20 days prior to dissection. (D-F) GFP caused by alternative splicing of traFΔT2AGFP pre-mRNA is not observed in control tj>traFΔT2AGFP testes (D’). By contrast, GFP indicative of traF splicing is robustly observed in follicle cells of a control tj>traFΔT2AGFP ovary (E’, arrowheads) and in somatic cells of a tj>traFΔT2AGFP; chinmoRNAi testis (F’, arrowheads). Tj (blue) marks somatic cells. Fas3 (red) marks hub cells in wild type testes, follicle cells in wild type ovaries and feminized cyst cells in tj>chinmoRNAi testes. (G-I) In a control tj>+ testis, only niche cells express Fas3 (G’). By contrast, in a tj>chinmoRNAi testis, most of the somatic lineage is positive for Fas3 (H’). In a “rescued” tj>traRNAi; chinmoRNAi testis, Fas3 is again restricted to niche cells (I’). Vasa (red) marks germ cells, Tj (blue) marks cyst cells, and Fas3 (green) marks niche cells and feminizing somatic cells. Results are quantified in Fig 4J and S1 Table. (J) Quantification of CySC feminization in tj>GFP; chinmoRNAi testes and various genotypes. 100% of tj>GFP; chinmoRNAi testes contain Fas3-positive aggregates (dark blue bar). There is a significant reduction in the percentage of feminized testes when tra is concomitantly depleted from tj>chinmoRNAi testes (purple bar). There is also a significant reduction when vir or fl(2)d is depleted from tj>chinmoRNAi testes (green and yellow bars, respectively). However, there is no rescue of male sex identity in tj>chinmoRNAi testes when Sxl, or nito is concomitantly depleted (light blue and red bars, respectively). Sample sizes are indicated within bars. *** denotes p<0.001; **** denotes p<0.0001 as determined by Fisher’s Exact Test (for rescue quantifications, compared with tj>GFP; chinmoRNAi). n.s. means not significant. See S1 Table for percentage values. (K-M) Fas3 is not ectopically expressed in CySCs from control (tjTS>+) testes (K’), testes with somatic mis-expression of traF (tjTS>traF) (L’), or testes with somatic mis-expression of fl(2)d (tjTS>fl(2)d) (M’). (N-P) Cas is not expressed in control (tjTS>+) (N’), tjTS>traF testes (O’), or tjTS>fl(2)d testes (P’). In K-P, Vasa (red) marks germ cells, and Zfh1 or Tj (blue) mark cyst cells. Scale bars = 20 μm.
To confirm these results, we monitored tra alternative splicing in vivo using a transgene that yields GFP expression when tra pre-mRNA is alternatively spliced (UAS-traFΔT2AGFP). In this transgene, the third exon of tra (which is adjoined with exon 1 in the female isoform) is replaced by the coding sequences for self-cleaving T2A peptide and GFP (S5 Fig). As expected, we detected little to no GFP expression in wild type (male) cyst cells (Fig 5D), while wild type (female) follicle cells expressed high levels (Fig 5E’, arrowheads). Notably, we also observed high levels of GFP in the soma of tj>chinmoRNAi testes (Fig 5F’, arrowheads), demonstrating that tra pre-mRNA is alternatively spliced to traF in these feminized somatic cells. We conclude that Chinmo normally represses tra transcription and alternative splicing in the male somatic gonad.
To determine if ectopic Chinmo is sufficient to repress tra transcription, we mis-expressed it in adult ovarian follicle cells using tj-gal4. To evade lethality caused by Chinmo mis-expression , we used a temperature-sensitive gal80 allele (tj-gal4, tub-gal80TS or tjTS) and reared flies at the permissive temperature (18°C). Adult F1 females were then shifted to the restrictive temperature (29°C) for 5 days before ovaries were homogenized. We observed a 2.2-fold decrease in total tra mRNA abundance (p<0.001) and a 1.8-fold decrease in traF abundance (p<0.001) in tjTS>chinmo ovaries compared with tjTS>+ ovaries (S6A Fig). dsxF was also decreased 5.8-fold (p<0.001) in tjTS>chinmo ovaries compared with tjTS>+ ovaries, presumably as a result of reduced TraF (S6A Fig). Taken together, our results demonstrate that Chinmo is both necessary and sufficient to prevent somatic expression of the female sex determinants traF and dsxF.
These findings suggest that sex transformation in chinmo-deficient cyst cells is due to ectopic TraF. To test this, we concomitantly depleted both tra and chinmo in the somatic lineage of the testis. Whereas 98% of tj>chinmoRNAi testes contained Fas3-positive aggregates outside of the niche, only 48% of tj>traRNAi; chinmoRNAi testes had such aggregates, indicating a significant block in feminization (p<0.0001) (Fig 5I; Fig 5J, purple bar; S1 Table). In these rescued tj>traRNAi; chinmoRNAi testes, CySCs no longer expressed Fas3, and the germline appeared normal (Fig 5I). We also performed epistatic experiments with tra mutant alleles, similar to the dsxD/dsx1 experiment. XX tra1/Df(3L)st-j7 animals develop male somatic structures due to loss of TraF (S4E–S4H Fig; ). Whereas 100% of chinmoST/chinmoST testes were feminized as assessed by Fas3-positive aggregates, only 61% of chinmoST/chinmoST; tra1/Df(3L)st-j7 testes were feminized (p<0.001) (Fig 4D, green bar; S1 Table). The phenotype was not sensitive to tra dose as chinmoST/chinmoST; tra/+ testes were still feminized (Fig 4D, yellow bars; S1 Table). These results demonstrate that Chinmo prevents both tra transcription and alternative splicing in CySCs and that feminization of male somatic cells in the absence of chinmo is due to ectopic traF.
Ectopic TraF impairs somatic differentiation but is not sufficient for CySC feminization
Global expression of TraF in XY flies during development causes female somatic differentiation . To test whether TraF expression alone is sufficient to cause male-to-female sex transformation in adult CySCs, we over-expressed traF cDNA in tj-gal4 expressing cells and used gal80TS to restrict expression to only adult CySCs (tjTS). While we observed accumulation of somatic aggregates in tjTS>traF testes, they did not express Fas3 or Cas, in contrast to those in tj>chinmoRNAi testes (compare Fig 5K and 5L to Fig 2I for Fas3 and compare Fig 5N and 5O to Fig 2F for Cas). These data suggest that traF-misexpressing cyst cells have not fully acquired a follicle-like fate. However, we found on average 121.0±8.8 somatic cells expressing Zinc finger homeodomain 1 (Zfh1), which marks CySCs and their earliest differentiating daughters , in tjTS>traF testes compared with 40.1±1.6 cells in control tjTS>+ testes (p<0.0001) (S7A, S7B and S7I Fig). Upon somatic traF mis-expression, we also observed accumulation of somatic cells expressing Tj, which marks a broader population of CySCs and early cyst cells  (S7C and S7D Fig). tjTS>traF testes contained 158.7±14.5 Tj-positive cells compared with 80.3±3.9 cells in tjTS>+ testes (p<0.001) (S7J Fig). We interpret the accumulation of Zfh1-positive, Tj-positive cells in tjTS>traF testes as a delay in somatic differentiation. Because cyst cells must exit the cell cycle in order to support the developing male germline, there are no somatic cells located away from the niche in wild type testes that are positive for 5-ethynyl-2’-deoxyuridine (EdU), an S-phase marker. We previously showed that when somatic differentiation is delayed, EdU-positive cyst cells are observed several cell diameters away from the niche . Consistent with our prior results, in control tjTS>+ testes, only Tj-positive cells near the hub incorporated EdU (S7E Fig, arrowheads; n = 0/20 testes with EdU-positive cyst cells located away from the niche). By contrast, in tjTS>traF testes we detected EdU-positive somatic cells located many cell diameters away from the niche, suggesting that these cells had delayed differentiation (S7F Fig, arrows; n = 20/26 testes with EdU-positive cyst cells located away from the niche).
Consistent with a defect in somatic differentiation, cyst cells mis-expressing traF were impaired in their ability to support the germline. In tjTS>traF testes, early germ cells accumulated (identified by dot- and dumbbell-shaped α-spectrin-positive fusomes) at the expense of more differentiated spermatogonia, as fewer germ cysts with branched fusomes were observed (S7G and S7H Fig). tjTS>traF testes also contained significantly fewer EdU-positive, 4- and 8-cell spermatogonial cysts than tjTS>+ testes (S7E, S7F and S7K Fig). These results demonstrate that ectopic TraF in CySCs is deleterious to their differentiation, but alone cannot drive CySCs to assume a follicle-like fate. Taken together with our previous finding that tra is downstream of chinmo in CySC feminization, we conclude while TraF induction is important for CySC feminization upon loss of chinmo, it is not sufficient.
Sxl is not required for feminization of chinmo-deficient CySCs
Our finding that chinmo-deficient CySCs produce traF (Fig 5A and 5C) reveals that they possess machinery to splice tra pre-mRNA into the female isoform. We considered the possibility that wild type CySCs might be competent to alternatively splice tra. However, somatic mis-expression of UAS-traFΔT2AGFP in wild type somatic cells did not lead to tra alternative splicing, since GFP was absent from the somatic lineage (Fig 5D). Thus, wild type CySCs are intrinsically unable to generate traF mRNA, precluding this model.
It follows, then, that one or more factors are ectopically expressed upon loss of chinmo that alternatively splice tra pre-mRNA into traF. Since Sxl is required for traF production in wild type females (Fig 6A; ), we investigated whether Sxl is ectopically expressed in chinmo-mutant CySCs. As expected, Sxl protein was absent from wild type testes but was detectable in wild type ovaries (Fig 6B and 6C). Importantly, we did not observe Sxl in chinmo-mutant testes (Fig 6D). These results were validated by assessing Sxl mRNA isoform abundance in adult gonads. Semi-quantitative RT-PCR demonstrated that control tj>+ testes express male-specific SxlM (Fig 6E, left lane), which contains an early stop codon and encodes no functional protein, while control tj>+ ovaries express female-specific SxlF (Fig 6E, middle lane), which encodes functional Sxl. tj>chinmoRNAi testes still express SxlM (Fig 6E, right lane), consistent with the absence of Sxl protein in these testes (Fig 6D). We also tested whether mis-expression of chinmo in female follicle cells could prevent Sxl alternative splicing; however, both tjTS>+ and tjTS>chinmo ovaries expressed only the female-specific SxlF isoform (S6B Fig).
(A) Schematic of tra pre-mRNA splicing. The poly(U) tract upstream of exon 2 is bound by the RRM domain of Sxl in females, causing skipping of exon 2. In wild type males, exons 1–4 comprise tra mRNA and translation terminates at the early stop codon in exon 2 (red star). Pink dashed lines indicate female-specific alternative splicing and blue dashed lines indicate non-sex-specific default splicing. (B-D) Sxl is not expressed in a control tj>+ testis (B’) but is expressed in follicle cells (C’, arrowheads) and in an early germ cell (C’, arrow) of a control tj>+ ovary. Sxl protein is not detected in a tj>chinmoRNAi testis (D’). (E) Semi-quantitative RT-PCR on Sxl in homogenized control tj>+ testes (left lane), control tj>+ ovaries (middle lane), and tj>chinmoRNAi testes (right lane). Control tj>+ testes express SxlM transcripts (blue arrowhead), while control tj>+ ovaries express SxlF transcripts (red arrowhead). SxlM is still present and SxlF is undetectable in tj>chinmoRNAi testes (right lane). SxlJYR primers were used to differentiate between SxlM and SxlF mRNA isoforms in this experiment. α-tubulin (tub) was used as a loading control. (F) Quantification of CySC feminization in Sxl; chinmoST backgrounds. Sample sizes are indicated within bars. **** denotes p<0.0001 as determined by Fisher’s Exact Test (compared to FM7/Y; chinmoST/chinmoST). See S1 Table for percentage values. (G-K) Representative images for Sxl; chinmoST epistasis experiments. Genetic loss of Sxl by 3 different alleles–Sxl f1 (I), Sxl f2 (J), or Sxl f18 (K)–does not prevent feminization of chinmoST/chinmoST cyst cells (G’), defined by the accumulation of Fas3-positive aggregates. Control FM7/Y; chinmoST/CyO cyst cells do not feminize (H’). Scale bars = 20 μm.
Furthermore, unlike depletion of tra, depletion of Sxl in feminizing, chinmo-deficient somatic cells did not suppress Fas3 expression or the epithelial organization of somatic cells (Fig 5J, light blue bar; S1 Table). [As expected, somatic depletion of Sxl in an otherwise wild type background produced no testis phenotype (S1 Table). We confirmed that the UAS-Sxl-RNAi line was effective at knockdown because somatic depletion of Sxl in females led to only a rudimentary ovary with 100% penetrance, n = 23.] Consistent with this, none of three distinct mutant alleles of Sxl prevented feminization in chinmoST/chinmoST testes (Fig 6F–6K; S1 Table). Taken together, these data support a model where the ectopic tra pre-mRNA in chinmo-mutant CySCs is alternatively spliced into traF via a non-canonical, Sxl-independent mechanism.
vir and fl(2)d are upregulated in chinmo-deficient CySCs and are required for sex transformation
We next examined a potential role for other candidates with known roles in female-specific alternative splicing of tra. We found that vir, fl(2)d, and nito transcripts were 1.5-fold (p<0.05), 3.4-fold (p<0.001), and 5.7-fold (p<0.0001) higher in adult ovaries compared with adult testes, respectively, suggesting sex-biased expression in adult gonads (Fig 7A and 7B). This observation is consistent with ModENCODE RNA-seq data demonstrating that vir, fl(2)d, and nito transcripts are present at very low levels in wild type testes . However, levels of all three transcripts significantly increased (2.3-fold, 1.7-fold, and 1.8-fold for vir, fl(2)d, and nito, respectively) in tj>chinmoRNAi testes compared with control tj>+ testes (Fig 7A and 7B; p<0.05 for vir and nito, p<0.01 for fl(2)d). While depleting vir or fl(2)d had no effect on testis development or spermatogenesis (S8A–S8C Fig), we found that depletion of vir in the female somatic gonad caused severe defects in ovary development. tj>virRNAi females develop some female reproductive structures and contain an oviduct, but lack ovaries (S8D and S8E Fig). Both tj>virRNAi and tj>fl(2)dRNAi females failed to lay fertilized eggs. To test whether vir or fl(2)d are necessary for traF splicing in adult ovaries, we depleted vir or fl(2)d in the female somatic gonad using tjTS, rearing flies at the permissive temperature to prevent vir or fl(2)d knockdown during development. After eclosion, adult females were then reared at the restrictive temperature to allow for vir and fl(2)d depletion. While wild type follicle cells express GFP produced by UAS-traFΔT2AGFP (Fig 7D), GFP is dramatically reduced in the follicle cells of tjTS>virRNAi and tjTS>fl(2)dRNAi ovaries (Fig 7C, 7E and 7F). These results demonstrate that vir and fl(2)d are both female-biased in the adult gonad and are required for traF alternative splicing in follicle cells.
(A) Semi-quantitative RT-PCR of vir, fl(2)d, and nito in homogenized control tj>+ testes (left lane), control tj>+ ovaries (middle lane) and tj>chinmoRNAi testes (right lane). vir, fl(2)d, and nito are expressed at higher levels in ovaries (middle lane) than in tj>+ testes (left lane). vir, fl(2)d, and nito are expressed at higher levels in tj>chinmoRNAi testes (right lane) than in control tj>+ testes (left lane). α-tubulin (tub) was used as a loading control. Flies were aged 9–20 days prior to dissection. (B) qRT-PCR analysis of vir, fl(2)d, and nito in homogenized control tj>+ testes (white bars), control tj>+ ovaries (gray bars) and tj>chinmoRNAi testes (blue bars). vir, fl(2)d, and nito are expressed at significantly higher levels in ovaries and tj>chinmoRNAi testes compared to control testes. * denotes p<0.05; ** denotes p<0.01; *** denotes p<0.001; **** denotes p<0.0001 as determined by two-tailed Student’s t-test. Error bars represent SEM. Flies were aged 9–20 days prior to dissection. (C) Quantification of GFP levels (synthesized from UAS-traFΔT2AGFP transgene) in tj>+, tj>virRNAi, and tj>fl(2)dRNAi ovaries represented in D-F. Errors represent SEM. **** denotes p<0.0001 as determined by Student’s t-test. (D-F) Representative images of control tj>+ (D’), tj>virRNAi (E’), and tj>fl(2)dRNAi (F’) ovaries in a UAS-traFΔT2AGFP background. GFP expressed from UAS-traFΔT2AGFP is detectable in escort cells (D’, arrow) and follicle cells (D’, arrowhead). GFP levels are dramatically reduced upon vir or fl(2)d depletion (E’ and F’, respectively) compared with wild type follicle cells (D’). Tj (blue) marks escort/follicle cells and Fas3 (red) marks follicle cells. (G-I) Depletion of nito (I’) does not reduce Fas3-positive (green) aggregates (a readout for feminization) in tj>chinmoRNAi testes. By contrast, depletion of vir (G’) or fl(2)d (H’) in tj>chinmoRNAi reduces the percentage of testes with these aggregates. Results are quantified in Fig 5J and S1 Table. Vasa (red) marks germ cells and Tj (blue) marks somatic cells. Scale bars = 20 μm.
To test whether vir or fl(2)d are required for sex transformation upon somatic loss of chinmo, we depleted each factor concomitantly with chinmo and monitored the frequency of CySC feminization. Depletion of vir or fl(2)d in tj>chinmoRNAi testes significantly reduced the percentage of feminized testes (p<0.001 and p<0.0001, respectively) (Fig 7G and 7H; Fig 5J, green and yellow bars, respectively; S1 Table). In contrast, depletion of nito did not prevent feminization (Fig 7I; Fig 5J, red bar; S1 Table). As expected, somatic depletion of vir, fl(2)d or nito in an otherwise wild type testis had no effect (S1 Table). We also tested the sufficiency of fl(2)d for CySC sex transformation. We found that mis-expression of fl(2)d in the adult CySC lineage (tjTS>fl(2)d) did not cause Fas3-positive aggregates to accumulate (Fig 5M). Furthermore, the follicle cell marker Cas was not induced in tjTS>fl(2)d testes (Fig 5P). Due to the lack of a UAS-vir transgenic Drosophila line, we were unable to test the sufficiency of vir for CySC feminization. Based on these findings, we conclude that Vir and Fl(2)d are epistatic to chinmo and are required, but not sufficient, for feminization of chinmo-mutant CySCs. Taken together with our previous results, this strongly implicates Vir and Fl(2)d in alternative splicing of the ectopic tra pre-mRNA observed in sex-transformed CySCs.
Chinmo prevents female sex identity in adult male CySCs
Here, we show that that one single factor, Chinmo, preserves the male identity of adult CySCs in the Drosophila testis by regulating the levels of canonical sex determinants. We demonstrate that CySCs lacking chinmo lose DsxM expression not by transcriptional loss but rather by alternative splicing of dsx pre-mRNA into dsxF. These chinmo-mutant CySCs ectopically express TraF and DsxF, and both factors are required for their feminization. Furthermore, our results demonstrate that tra alternative splicing in cyst cells lacking chinmo is achieved independently of Sxl. Instead, our work strongly suggests that traF production in the absence of chinmo is mediated by splicing factors Vir and Fl(2)d. We propose that male sex identity in CySCs is maintained by a two-step mechanism whereby traF is negatively regulated at both transcriptional and post-transcriptional levels by Chinmo (Fig 8). In this model, loss of chinmo from male somatic stem cells first leads to transcriptional upregulation of tra pre-mRNA as well as of vir and fl(2)d. Then the tra pre-mRNA in these cells is spliced into traF by the ectopic Vir and Fl(2)d proteins. The ectopic TraF in chinmo-deficient CySCs then splices the dsx pre-mRNA into dsxF, resulting in loss of DsxM and gain of DsxF, and finally induction of target genes usually restricted to follicle cells in the ovary.
Left: In XX animals, the production of Sxl leads to alternative splicing of tra pre-mRNA into traF. TraF protein then alternatively splices dsx pre-mRNA into dsxF. The DsxF protein promotes female-specific transcriptional changes. Right: In XY animals, Sxl is not produced. Neither tra nor dsx pre-mRNA are alternatively spliced, resulting in the production of DsxM protein, which ensures male-specific transcription of target genes. In addition to the absence of Sxl in XY cells, adult somatic stem cells of the Drosophila testis have an extra level of insurance of male sex identity. Chinmo, which is expressed only in male but not female somatic gonadal cells, represses expression of tra, vir and fl(2)d in CySCs. This safeguards male identity by reducing the availability of tra pre-mRNA and of factors (i.e., Vir and Fl(2)d)) that can splice it into traF. Thus, in addition to the canonical sex determination pathway that establishes male and female programs from early development, adult male, sexually dimorphic cells protect their sexual identity by transcriptional repression of tra and its splice factors.
Chinmo regulates levels of tra, vir and fl(2)d
Chinmo has motifs associated with transcriptional repression and its loss clonally is associated with ectopic transcription . One interpretation of our data is that Chinmo directly represses tra, vir, and fl(2)d in male somatic gonadal cells. As the binding site and potential co-factors of Chinmo are not known, future work will be needed to determine whether Chinmo directly regulates expression of these genes. We also note that ~50% of chinmo-mutant testes still feminize in the genetic absence of tra or dsxF. These latter data indicate that Chinmo regulates male sex identity through another, presumably parallel, mechanism that does not involve canonical sex determinants. However, this tra/dsx-independent mode of sex maintenance downstream of Chinmo is not characterized and will require the identification of direct Chinmo target genes.
Regulation of chinmo expression in adult CySCs
We previously showed that JAK/STAT signaling promotes chinmo in several cell types, including CySCs . Since JAK/STAT signaling is itself sex-biased and restricted to the embryonic male gonad, we presume that activated Stat92E establishes chinmo in male somatic gonadal precursors, perhaps as early as they are specified in the embryo [33, 71]. Because loss of Stat92E from CySCs does not result in an apparent sex transformation phenotype [29, 40, 60], we favor the interpretation that Stat92E induces expression of chinmo in CySCs but that other sexually biased factors maintain it. One potential candidate is DsxM, which is expressed specifically in early somatic gonads and at the same time when Stat92E activation is occurring in these cells . In fact, multiple DsxM ChIP-seq peaks were identified in the chinmo locus, suggesting potential regulation of chinmo by DsxM . Taken together with our findings, this suggests a potential autoregulatory feedback loop whereby DsxM preserves its own expression in adult CySCs by maintaining Chinmo expression, which in turn prevents traF and dsxF production.
Non-canonical mechanisms of sex-specific cell fate and tissue homeostasis
Recent studies on tissue-specific sex maintenance demonstrate that while the Sxl/Tra/Dsx hierarchy is an obligate and linear circuit during embryonic development, at later stages it is more modular than previously appreciated. For example, Sxl can regulate female-biased genes in a tra-independent manner [73, 74]. Additionally, Sxl and TraF regulate body size and gut plasticity independently of the only known TraF targets, dsx and fru [3, 4]. We find that negative regulation of the TraF-DsxF arm of this cascade is required to preserve male sexual identity in CySCs but unexpectedly is independent of Sxl. Because depletion of Vir or Fl(2)d significantly blocks sex transformation and both are required for tra alternative splicing in the ovary, our work reveals they can alternatively splice tra pre-mRNA even in the absence of Sxl. To the best of our knowledge, this is the first demonstration of Sxl-independent, Tra-dependent feminization. These results raise the broader question of whether other male somatic cells have to safeguard against this novel mechanism. Because recent work has determined that sex maintenance is important in systemic functions regulated by adipose tissue and intestinal stem cells [3, 4], it will be important to determine whether Chinmo represses traF in these settings. Finally, since the transcriptional output of the sex determination pathway is conserved from Drosophila (Dsx) to mammals (DMRT1), it is possible that transcriptional regulation of sex determinants plays a similar role in adult tissue homeostasis and fertility in higher organisms.
Materials and methods
Fly stocks and husbandry
The following fly stocks were used and are described in FlyBase: OregonR; yw; tj-gal4; tub-gal80TS; dsx-gal4; dsx-galΔ2; GMR40A05-gal4; dsxMI03050-GFSTF.1; dsx1; dsxD; chinmoST; tra1; Df(3L)st-j7; UAS-GFPnls; UAS-dcr2; UAS-chinmoRNAi (HMS00036); UAS-traRNAi (HMS02830); UAS-traF; UAS-3xHAfl(2)d; UAS-dsxF; UAS-5’UTR-chinmo-3’UTR; UAS-SxlRNAi (HMS00609); Sxlf1; Sxlf2; Sxlf18; UAS-virRNAi (HMC03908); UAS-fl(2)dRNAi (HMC03833); UAS-nitoRNAi (HMS00166).
For RNAi-mediated depletion of chinmo, Sxl, tra, vir, fl(2)d, and nito, flies were reared at an ambient temperature (21°C). Adult males were collected twice a week and aged at 29°C to increase Gal4 activity. For temporal control of gene expression, tj-gal4, tub-gal80TS virgins were crossed to UAS-traF or UAS-3xHAfl(2)d males and progeny were reared at the permissive temperature (18°C) to prevent traF or fl(2)d mis-expression during embryonic, larval, and pupal development. Adult males of the correct genotype were collected twice a week and shifted to the restrictive temperature (29°C) to inactivate Gal80.
Generation of UAS-traFΔT2AGFP
In this transgene, most of the third exon of tra is replaced by the coding sequences for self-cleaving T2A peptide and GFP. Specifically, the coding sequences of T2A and GFP were cloned in frame immediately downstream of the 26th nucleotide (nt) of tra exon 3 and immediately upstream of the last 18 nt of this exon. PCR was performed with Q5 high-fidelity polymerase from New England Biolabs (M0491S). The PCR product was digested with EcoRI and XhoI before cloning into the pUASTattb vector . The construct was verified by sequencing, and a transgenic line was established through ΦC-31 integrase mediated transformation (Bestgene, attP site VK05, BDSC#9725).
The following primary antibodies were used: rat anti-Chinmo (1:1000; gift of N. Sokol, Indiana University, IN, USA), goat anti-Vasa (1:50, dC-13, Santa Cruz), rabbit anti-Vasa (1:1500; gift of R. Lehmann, Skirball Institute/NYU School of Medicine, NY, USA), guinea pig anti-Tj (1:5000; gift of D. Godt, University of Toronto, ON, Canada), rabbit anti-Zfh1 (1:5000; gift of R. Lehmann), mouse anti-Fasciclin-3 (1:50; Developmental Studies Hybridoma Bank (DSHB)), mouse anti-Eya (1:20; DSHB), rat anti-DsxM (1:200; gift of B. Oliver, National Institutes of Health, MD, USA), rat anti-DsxC (1:50; gift of M. Arbeitman, Florida State University, FL, USA), rabbit anti-Castor (1:50; gift of W. Odenwald, National Institutes of Health, MD, USA), mouse anti-α-spectrin (1:20, DSHB), mouse anti-SxlM18 (1:5; DSHB), rabbit anti-GFP (1:500; Invitrogen). Secondary antisera used were all raised in donkey (Jackson ImmunoResearch).
Testes and ovaries were dissected in 1x PBS and fixed in 4% paraformaldehyde in 1x PBS for 30 minutes at room temperature (RT). Fixed tissue was washed twice at RT in 0.5% PBST (1x PBS with 0.5% Triton X-100) and blocked in PBTB (1x PBS, 0.2% Triton X-100, 1% BSA) for 1 hour at RT or overnight at 4°C. Primary antibodies were incubated overnight at 4°C and washed off twice at RT in PBTB. Secondary antibodies were incubated for 2 hours at RT in the dark and washed off twice in 0.2% PBST (1x PBS with 0.2% Triton X-100). Tissue was mounted in Vectashield Medium (Vector Laboratories) prior to confocal analysis, and confocal images were captured using a Zeiss LSM 510 confocal microscope, 63x objective.
DIC microscopy on whole ovaries
DIC images of adult female reproductive structures (at 5x) were obtained using a Zeiss Axioplan microscope with a Retiga Evi (QImaging) digital camera and QCapture Pro 6.0 software.
Immunofluorescence using anti-DsxC
Testes and ovaries were dissected in 1x PBS and fixed in 20% EM-grade paraformaldehyde (Electron Microscopy Sciences) in 1x PBS for 20 minutes at RT. Fixed tissue was washed 3 times for 15 minutes each in TNT (0.1M Tris-HCl, 0.3M NaCl, 0.05% Tween-20) and blocked using Image-iT FX Signal Enhancer (ThermoFisher) for 30 minutes at RT, then washed 3 times for 15 minutes each in TNT. Primary anti-DsxC was incubated overnight at 4°C. After anti-DsxC incubation, tissue was blocked in PBTB for 1 hour and then treated with anti-Vasa and anti-Tj. Primary antibodies were washed twice for 15 minutes each in PBTB, then secondary antibodies were incubated overnight at 4°C in PBTB. Finally, the DsxC signal was amplified by TSA (see below) and testes were mounted in Vectashield prior to analysis.
Tyramide signal amplification (TSA)
TSA (Perkin Elmer) was performed to amplify DsxM and DsxC signals. HRP anti-rat (Jackson ImmunoResearch) was used as a secondary antibody and the tertiary Cy3-conjugated tyramide reaction was performed per the manufacturer’s instructions.
CySC purification by fluorescence-activated cell sorting (FACS)
To purify CySCs and early cyst cells, the somatic cell lineage was labeled using tj-gal4 to drive UAS-GFPnls expression. Testes were dissociated in trypsin/collagenase for 15 minutes and the cell suspension was passed through 70μm filters (Falcon). GFP-expressing somatic cells were purified from the resulting filtrate by FACS using a Sony SY3200 highly automated parallel sorting (HAPS) cell sorter into TRIzol LS (ThermoFisher), and RNA was extracted according to the manufacturer’s instructions. Post-sort purity of samples was confirmed by immunocytochemistry and the absence of Vasa-positive germ cells.
5-ethynyl-2’-deoxyuridine (EdU)-labeling of adult testes
EdU-labeling of testes was performed using the Click-iT EdU Alexa Fluor 647 Imaging Kit (ThermoFisher). Testes were dissected in S2 cell culture medium (Life Technologies) then incubated in 10 μM EdU for 30 minutes. Testes were then fixed, washed, and stained as described above. The cycloaddition reaction was performed per the manufacturer’s instructions. Testes were mounted in Vectashield prior to confocal analysis.
Quantitative and semi-quantitative RT-PCR
To detect mRNA levels of canonical sex determinants by PCR, whole ovaries (n = 5–10) or whole testes (n = 55–200) were isolated and homogenized into TRIzol (ThermoFisher). RNA was extracted and DNase-treated (Ambion) per the manufacturer’s instructions. Reverse transcription was performed using Maxima reverse transcriptase (ThermoFisher) according to the manufacturer’s instructions and 1–2 μg of RNA as template. qRT-PCR was performed using SYBR Green PCR Master Mix (ThermoFisher) and a Biorad CFX96 Real-Time PCR Machine. Semi-quantitative RT-PCR was performed on a Biorad iCycler. Because the proportion of somatic cells is significantly increased in tj>chinmoRNAi testes compared to tj>+ controls, the qRT-PCR values were normalized first to tubulin and second to zfh1, an early somatic marker.
total tra: fwd-GAGCCCGCATCGGTATAATC; rev-GACGTGGTAGCCTTTGGTATC
traF: fwd-AACCCAGCATCGAGATTCC; rev-CGAACCTCGTCTGCAAAGTA
dsxC: fwd-GAAAGAACGGCGCCAAT; rev-GGCGTCTGCGTCCTTTAATA
dsxM: fwd-GAGCTGATGCCACTCATGTAT; rev-CTGGGCTACAGTGCGATTTA
dsxF: fwd-GAATGAGTACTCCCGTCAACAT; rev-GGGCAAAGTAGTATTCGTTACTCTA
rpl15: fwd-AGGATGCACTTATGGCAAGC; rev-GCGCAATCCAATACGAGTTC
α-tub84b: fwd-CAACCAGATGGTCAAGTGCG; rev-ACGTCCTTGGGCACAACATC
β-tub56d: fwd-CTCAGTGCTCGATGTTGTCC; rev-GCCAAGGGAGTGTGTGAGTT
SxlJYR: fwd-ACACAAGAAAGTTGAACAGAGG; rev-CATTCCGGATGGCAGAGAATGG
SxlEM: fwd-CGCTGCGAGTCCATTTCC; rev-GTGGTTATCCCCCATATGGC
vir: fwd-CATGAGGAAGTGACGGACATC; rev-GGAAAGTCTGCCTGGACTCG
fl(2)d: fwd-GGCCAACAAGGAGCAAGAA; rev-CGCTCGAACAGGAGATTGAC
nito: fwd-GGTGTACAAGTCCACAACCAGA; rev-CGACGGTGATCCAAAGGAA
The fertility of adult males was assayed by mating individual males with two wild type (OregonR) virgin females (between 5–10 days old) for 48 hours at 25°C. After a 2-day mating period, males were recovered and preserved for subsequent matings using fresh virgin OregonR females. Fertility was scored by counting the number of F1 offspring produced by each individual cross and reported as the average number of F1 offspring for each genotype.
Statistical parameters for each experiment are reported in the figure legends. Data were analyzed using Microsoft Excel and are reported to be statistically significant when p<0.05 by the appropriate statistical test. For qRT-PCR data, significance was determined by two-tailed Student’s t-test. For fertility assays and cyst cell quantifications, significance was determined using single-factor ANOVA. For rescue of CySC feminization (Fas3-positive aggregates), significance was determined using Fisher’s Exact Test.
S1 Table. Quantification of testes with Fas3-positive somatic aggregates (referred to as “feminized”).
Data are presented as the percentage of testes with Fas-3-positive aggregates in testes of the indicated genotypes from the total number of testes examined.
S1 Fig. Loss of chinmo in the CySC lineage causes sex transformation and loss of the germline.
(A-C) A transcriptional reporter for slow border cells (slbo-GFP) is not expressed in a wild type testis (A’, arrowhead). slbo-GFP (green) is expressed in mature follicle cells (B’, arrowhead). slbo-GFP is ectopically expressed in the CySC lineage upon loss of chinmo (C’, arrowhead). Time point is 11 days post-eclosion. Tj (blue) marks cyst cells. Vasa (red) marks the germline.
(D) Semi-quantitative RT-PCR on Yp1 using RNA extracts from homogenized control tj>+ testes (left lane), control tj>+ ovaries (middle lane), and tj>chinmoRNAi testes (right lane). Yp1 is expressed in control tj>+ ovaries (middle lane), but not in control tj>+ testes (left lane). In tj>chinmoRNAi testes (right lane), Yp1 is expressed. α-tubulin (tub) was used as a loading control. Timepoint is 9–14 days post-eclosion.
(E-F) Loss of chinmo in male niche cells using upd-gal4, gal80TS (updTS) causes no overt defects in testis development or spermatogenesis. Time point is 8 days post-eclosion. TOPRO (blue) marks DNA. Fas3 (green) marks niche cells.
(G-H) Representative images of agametic tj>chinmoRNAi testes at 7 days post-eclosion. Fas3-positive somatic aggregates (green) fill the apex of the testis, which is devoid of Vasa-positive (red) germ cells. Zfh1 (blue) marks somatic cells.
(I-K) Expression of Chinmo in adult gonads. Chinmo is expressed in the CySC lineage of the adult testis (I’, arrowheads) but is absent from follicle cells in the adult ovary (J’, arrowheads). Upon chinmo depletion in the testis (tj>chinmoRNAi), Chinmo protein is lost from feminizing cyst cells (K’, arrowheads). The remaining Chinmo protein observed in K’ represents Chinmo expression in the male germline. Vasa (red) marks germ cells and Tj (blue) marks somatic cells.
Scale bars = 20 μm.
S2 Fig. Three different dsx transcriptional reporters show variable expression in adult gonads.
(A-B) Expression of dsx-gal4Δ2 in adult gonads. In the testis, dsx-gal4Δ2 is expressed in the entire CySC lineage (A). In the ovary, dsx-gal4Δ2 is expressed in escort cells, but not follicle cells (B).
(C-D) Expression of GMR40A05-gal4 in adult gonads. In the testis, GMR40A05-gal4 is expressed in the entire CySC lineage (C). In the ovary, GMR40A05-gal4 is expressed in escort cells, but not follicle cells (D).
(E-F) Expression of dsxMI03050-GFSTF.1 in adult gonads. In the testis, dsxMI03050-GFSTF.1 is expressed weakly in the CySC lineage (E) and is undetectable in adult ovaries (F).
Fas3 (red) marks testicular niche cells and ovarian follicle cells. Tj (blue) marks somatic cells in both gonads. Time point for all adults is 5 days post eclosion. Scale bars = 20 μm.
S3 Fig. DsxC antibody detects DsxF protein.
Immunostaining of tj>dsxF ovaries reveals that DsxF protein is detectable by DsxC antibody (magenta). Tj (green) marks somatic cells. Scale bars = 20 μm.
S4 Fig. Blocking DsxF or TraF production genetically in females causes masculinization of the soma.
(A-D) Blocking dsxF production using the dsxD/dsx1 heteroallelic combination masculinizes the soma of XX animals. Chromosomal sex of flies was determined based on inheritance of X-linked traits (eye color, w; cuticle color, y). Genotype for A (XX animal) is yw/y+w+; chinmoST/chinmoST; dsx1/TM2; for B (XX animal) is yw/y+w+; chinmoST/chinmoST; dsxD/dsx1; for C (XY animal) is yw/Y; chinmoST/chinmoST; dsx1/TM2; for D (XY animal) is yw/Y; chinmoST/chinmoST; dsxD/dsx1.
(E-H) Blocking traF production using tra1/Df(3L)st-j7, Ki1 masculinizes the soma of XX animals. Chromosomal sex of flies was determined based on inheritance of X-linked traits (eye color, w). Genotype for E (XX animal) is w/w+; chinmoST/chinmoST; tra1/TM6B, Tb; for F (XX animal) is w/w+; chinmoST/chinmoST; tra1/Df(3L)st-j7, Ki1; for G (XY animal) is w/Y; chinmoST/chinmoST; tra1/TM6B, Tb; for H (XY animal) is w/Y; chinmoST/chinmoST; tra1/Df(3L)st-j7, Ki1.
S5 Fig. Diagram of tra pre-mRNA and UAS-traFΔT2AGFP.
In the transgene, most of the third exon of tra is replaced with self-cleaving T2A peptide and GFP, followed by a poly-adenylation signal (pA). Black shaded regions indicate exons. Red star indicates early stop codon in exon 2. Pink dashed lines indicate female-specific alternative splicing, and blue dashed lines indicate non-sex-specific default splicing.
S6 Fig. Chinmo mis-expression in ovaries leads to reduced traF and dsxF levels.
(A) qRT-PCR analysis of homogenized ovaries demonstrates that mis-expression of chinmo in follicle cells leads to decreased levels of total tra, traF, and dsxF. Lower transcript levels were not due to a change in the relative abundance of somatic cells, as zfh1 levels were unaffected in tjTS>chinmo ovaries. The values were normalized to tubulin. Data are presented as the mean of three biological replicates. *** denotes p<0.001 as determined by two-tailed Student’s t-test. Error bars represent SEM.
(B) Semi-quantitative RT-PCR on RNA extracts from 5 male or 5 female larvae (first two lanes), tjTS>+ adult ovaries (third lane), and tjTS>chinmo adult ovaries (last lane). RNA from male larvae express SxlM (first lane), while RNA from female larvae express SxlF (second lane). Both tjTS>+ (third lane) and tjTS>chinmo (last lane) ovaries express SxlF exclusively. SxlEM primers were used to differentiate between SxlM and SxlF mRNA isoforms in this experiment. α-tubulin (tub) was used as a loading control.
S7 Fig. TraF is necessary but not sufficient for CySC feminization.
(A-B) Zfh1 (blue) expression in tjTS>+ (A) versus tjTS>traF (B) testes. A and B represent single Z slices; A’ and B’ show maximal Z-projections (Z-max) of Zfh1-expressing cells in the entire confocal stack. Fas3 (green) marks the niche.
(C-D) Tj (blue) expression in tjTS>+ (C) versus tjTS>traF (D) testes. C and D represent single Z slices; C’ and D’ show Z-max projections of Tj-expressing cells.
(E-F) EdU (blue)-labeled tjTS>+ (E) and tjTS>traF (F) testes. EdU-positive spermatogonial cysts are outlined. Tj (green) marks cyst cells. Arrowheads (E’) point to EdU-positive CySCs. Arrows (F’) point to EdU-positive differentiating cyst cells away from the niche. Asterisk marks the niche.
(G-H) Visualization of germ cell stages in tjTS>+ (G) and tjTS>traF (H) testes. α-spectrin (green) marks fusomes, which are dot- and dumbbell-shaped in early germ cells (G’, arrowheads) and become branched in later differentiating spermatogonia (G”, arrows). Note that the niche is not in the plane in G’. Tj (blue) marks cyst cells. Arrowheads in H’ indicate spermatogonia away from the niche that have dot and dumbbell shape fusomes in tjTS>traF testes. Asterisk marks the niche.
(I-J) Quantification of Zfh1-expressing (I) and Tj-expressing (J) cells in tjTS>+ (gray bars) versus tjTS>traF (green bars) testes. tjTS>traF testes contain significantly more Zfh1-expressing and Tj-expressing somatic cells than tjTS>+ testes, as determined by single-factor ANOVA.
(K) Quantification of EdU-positive germ cells upon somatic traF mis-expression. tjTS>traF testes contain significantly fewer EdU-positive 4-cell and 8-cell spermatogonia than tjTS>+ testes.
For quantifications, * denotes p<0.05; ** denotes p<0.01; *** denotes p<0.001; **** denotes p<0.0001 as determined by single-factor ANOVA.
Quantification data are presented as mean ± SEM.
Vasa (red) marks the germline in A-H. Scale bars = 20 μm.
S8 Fig. Loss of vir and fl(2)d causes defects in the ovary but not the testis.
(A-C) tj>virRNAi (B) and tj>fl(2)dRNAi (C) testes resemble control tj>+ (A) testes, showing no overt defects in testis development or spermatogenesis. Vasa (red) marks the germline, Tj (blue) marks somatic cells, and Fas3 (green) marks niche cells. Scale bars = 20 μm.
(D-E) Reproductive structures in adult tj>+ (D) and tj>virRNAi (E) females. Ovaries (D, brackets) and accessory structures like spermathecae (SP) (D, arrows) can be observed in tj>+ females. Ovaries, but not somatic accessory structures like SP and oviduct, fail to develop in females lacking vir in the somatic gonad (E, arrows).
We acknowledge the Bloomington Drosophila Stock Center, TRiP and DSHB for stocks and antibodies. We thank E. Rideout for providing dsx-gal4, tra1 and Df(3L)st-j7 stocks, M. Van Doren for dsx-gal4Δ2 and D. Montell for slbo-GFP. We also thank N. Sokol, D. Godt, R. Lehmann, B. Oliver, M. Arbeitman, and W. Odenwald for generously sharing antibodies. We thank M. Fuller for kindly providing the FACS protocol. Additionally, we are incredibly grateful to J. Treisman and H.D. Ryoo for insightful conversations, and M. Amoyel and T. Hurd for manuscript feedback.
- 1. Lee G, Hall JC, Park JH. Doublesex gene expression in the central nervous system of Drosophila melanogaster. Journal of neurogenetics. 2002;16(4):229–48. pmid:12745633.
- 2. Rideout EJ, Dornan AJ, Neville MC, Eadie S, Goodwin SF. Control of sexual differentiation and behavior by the doublesex gene in Drosophila melanogaster. Nature neuroscience. 2010;13(4):458–66. pmid:20305646; PubMed Central PMCID: PMC3092424.
- 3. Rideout EJ, Narsaiya MS, Grewal SS. The Sex Determination Gene transformer Regulates Male-Female Differences in Drosophila Body Size. PLoS genetics. 2015;11(12):e1005683. pmid:26710087; PubMed Central PMCID: PMC4692505.
- 4. Hudry B, Khadayate S, Miguel-Aliaga I. The sexual identity of adult intestinal stem cells controls organ size and plasticity. Nature. 2016;530(7590):344–8. pmid:26887495; PubMed Central PMCID: PMC4800002.
- 5. Mauvais-Jarvis F. Sex differences in metabolic homeostasis, diabetes, and obesity. Biology of sex differences. 2015;6:14. pmid:26339468; PubMed Central PMCID: PMC4559072.
- 6. Casper AL, Van Doren M. The establishment of sexual identity in the Drosophila germline. Development. 2009;136(22):3821–30. pmid:19855024; PubMed Central PMCID: PMC2766343.
- 7. Sacher J, Neumann J, Okon-Singer H, Gotowiec S, Villringer A. Sexual dimorphism in the human brain: evidence from neuroimaging. Magnetic resonance imaging. 2013;31(3):366–75. pmid:22921939.
- 8. Camara N, Whitworth C, Van Doren M. The creation of sexual dimorphism in the Drosophila soma. Current topics in developmental biology. 2008;83:65–107. pmid:19118664.
- 9. Erickson JW, Quintero JJ. Indirect effects of ploidy suggest X chromosome dose, not the X:A ratio, signals sex in Drosophila. PLoS biology. 2007;5(12):e332. pmid:18162044; PubMed Central PMCID: PMCPMC2222971.
- 10. Salz HK, Erickson JW. Sex determination in Drosophila: The view from the top. Fly. 2010;4(1):60–70. pmid:20160499; PubMed Central PMCID: PMC2855772.
- 11. Bell LR, Horabin JI, Schedl P, Cline TW. Positive autoregulation of sex-lethal by alternative splicing maintains the female determined state in Drosophila. Cell. 1991;65(2):229–39. pmid:2015624.
- 12. Inoue K, Hoshijima K, Sakamoto H, Shimura Y. Binding of the Drosophila sex-lethal gene product to the alternative splice site of transformer primary transcript. Nature. 1990;344(6265):461–3. pmid:1690860.
- 13. Boggs RT, Gregor P, Idriss S, Belote JM, McKeown M. Regulation of sexual differentiation in D. melanogaster via alternative splicing of RNA from the transformer gene. Cell. 1987;50(5):739–47. pmid:2441872.
- 14. Hilfiker A, Amrein H, Dubendorfer A, Schneiter R, Nothiger R. The gene virilizer is required for female-specific splicing controlled by Sxl, the master gene for sexual development in Drosophila. Development. 1995;121(12):4017–26. pmid:8575302.
- 15. Hilfiker A, Nothiger R. The temperature-sensitive mutation vir ts(virilizer) identifies a new gene involved in sex determination of Drosophila. Roux's archives of developmental biology: the official organ of the EDBO. 1991;200(5):240–8. pmid:28305793.
- 16. Granadino B, Campuzano S, Sanchez L. The Drosophila melanogaster fl(2)d gene is needed for the female-specific splicing of Sex-lethal RNA. The EMBO journal. 1990;9(8):2597–602. pmid:1695150; PubMed Central PMCID: PMC552292.
- 17. Granadino B, Penalva LO, Sanchez L. The gene fl(2)d is needed for the sex-specific splicing of transformer pre-mRNA but not for double-sex pre-mRNA in Drosophila melanogaster. Molecular & general genetics: MGG. 1996;253(1–2):26–31. pmid:9003283.
- 18. Yan D, Perrimon N. spenito is required for sex determination in Drosophila melanogaster. Proceedings of the National Academy of Sciences of the United States of America. 2015;112(37):11606–11. pmid:26324914; PubMed Central PMCID: PMC4577178.
- 19. Granadino B, San Juan A, Santamaria P, Sanchez L. Evidence of a dual function in fl(2)d, a gene needed for Sex-lethal expression in Drosophila melanogaster. Genetics. 1992;130(3):597–612. pmid:1551580; PubMed Central PMCID: PMC1204876.
- 20. Burtis KC, Coschigano KT, Baker BS, Wensink PC. The doublesex proteins of Drosophila melanogaster bind directly to a sex-specific yolk protein gene enhancer. The EMBO journal. 1991;10(9):2577–82. pmid:1907913; PubMed Central PMCID: PMC452955.
- 21. Inoue K, Hoshijima K, Higuchi I, Sakamoto H, Shimura Y. Binding of the Drosophila transformer and transformer-2 proteins to the regulatory elements of doublesex primary transcript for sex-specific RNA processing. Proceedings of the National Academy of Sciences of the United States of America. 1992;89(17):8092–6. pmid:1518835; PubMed Central PMCID: PMC49862.
- 22. Lynch KW, Maniatis T. Assembly of specific SR protein complexes on distinct regulatory elements of the Drosophila doublesex splicing enhancer. Genes & development. 1996;10(16):2089–101. pmid:8769651.
- 23. Williams TM, Selegue JE, Werner T, Gompel N, Kopp A, Carroll SB. The regulation and evolution of a genetic switch controlling sexually dimorphic traits in Drosophila. Cell. 2008;134(4):610–23. pmid:18724934; PubMed Central PMCID: PMC2597198.
- 24. Shirangi TR, Dufour HD, Williams TM, Carroll SB. Rapid evolution of sex pheromone-producing enzyme expression in Drosophila. PLoS biology. 2009;7(8):e1000168. pmid:19652700; PubMed Central PMCID: PMC2711336.
- 25. Coschigano KT, Wensink PC. Sex-specific transcriptional regulation by the male and female doublesex proteins of Drosophila. Genes & development. 1993;7(1):42–54. pmid:8422987.
- 26. Clough E, Jimenez E, Kim YA, Whitworth C, Neville MC, Hempel LU, et al. Sex- and tissue-specific functions of Drosophila doublesex transcription factor target genes. Developmental cell. 2014;31(6):761–73. pmid:25535918; PubMed Central PMCID: PMC4275658.
- 27. Erdman SE, Burtis KC. The Drosophila doublesex proteins share a novel zinc finger related DNA binding domain. The EMBO journal. 1993;12(2):527–35. pmid:8440242; PubMed Central PMCID: PMCPMC413235.
- 28. Arbeitman MN, New FN, Fear JM, Howard TS, Dalton JE, Graze RM. Sex Differences in Drosophila Somatic Gene Expression: Variation and Regulation by doublesex. G3. 2016;6(7):1799–808. pmid:27172187; PubMed Central PMCID: PMC4938635.
- 29. Ma Q, Wawersik M, Matunis EL. The Jak-STAT target Chinmo prevents sex transformation of adult stem cells in the Drosophila testis niche. Developmental cell. 2014;31(4):474–86. pmid:25453558; PubMed Central PMCID: PMC4254588.
- 30. Shapiro-Kulnane L, Smolko AE, Salz HK. Maintenance of Drosophila germline stem cell sexual identity in oogenesis and tumorigenesis. Development. 2015;142(6):1073–82. pmid:25758221; PubMed Central PMCID: PMC4360176.
- 31. Yang SY, Baxter EM, Van Doren M. Phf7 controls male sex determination in the Drosophila germline. Developmental cell. 2012;22(5):1041–51. pmid:22595675; PubMed Central PMCID: PMC3635078.
- 32. Waterbury JA, Horabin JI, Bopp D, Schedl P. Sex determination in the Drosophila germline is dictated by the sexual identity of the surrounding soma. Genetics. 2000;155(4):1741–56. pmid:10924471; PubMed Central PMCID: PMC1461178.
- 33. Wawersik M, Milutinovich A, Casper AL, Matunis E, Williams B, Van Doren M. Somatic control of germline sexual development is mediated by the JAK/STAT pathway. Nature. 2005;436(7050):563–7. pmid:16049490; PubMed Central PMCID: PMC1421378.
- 34. Murray SM, Yang SY, Van Doren M. Germ cell sex determination: a collaboration between soma and germline. Current opinion in cell biology. 2010;22(6):722–9. pmid:21030233; PubMed Central PMCID: PMC2996260.
- 35. Fairchild MJ, Smendziuk CM, Tanentzapf G. A somatic permeability barrier around the germline is essential for Drosophila spermatogenesis. Development. 2015;142(2):268–81. pmid:25503408.
- 36. Amoyel M, Hillion KH, Margolis SR, Bach EA. Somatic stem cell differentiation is regulated by PI3K/Tor signaling in response to local cues. Development. 2016;143(21):3914–25. pmid:27633989; PubMed Central PMCID: PMC5117146.
- 37. Hardy RW, Tokuyasu KT, Lindsley DL, Garavito M. The germinal proliferation center in the testis of Drosophila melanogaster. Journal of ultrastructure research. 1979;69(2):180–90. pmid:114676.
- 38. Zoller R, Schulz C. The Drosophila cyst stem cell lineage: Partners behind the scenes? Spermatogenesis. 2012;2(3):145–57. pmid:23087834; PubMed Central PMCID: PMC3469438.
- 39. Gilboa L, Lehmann R. Soma-germline interactions coordinate homeostasis and growth in the Drosophila gonad. Nature. 2006;443(7107):97–100. pmid:16936717.
- 40. Amoyel M, Sanny J, Burel M, Bach EA. Hedgehog is required for CySC self-renewal but does not contribute to the GSC niche in the Drosophila testis. Development. 2013;140(1):56–65. pmid:23175633; PubMed Central PMCID: PMC3513992.
- 41. Kiger AA, Jones DL, Schulz C, Rogers MB, Fuller MT. Stem cell self-renewal specified by JAK-STAT activation in response to a support cell cue. Science. 2001;294(5551):2542–5. pmid:11752574.
- 42. Tran J, Brenner TJ, DiNardo S. Somatic control over the germline stem cell lineage during Drosophila spermatogenesis. Nature. 2000;407(6805):754–7. pmid:11048723.
- 43. Margolis J, Spradling A. Identification and behavior of epithelial stem cells in the Drosophila ovary. Development. 1995;121(11):3797–807. pmid:8582289.
- 44. Zhang Y, Kalderon D. Hedgehog acts as a somatic stem cell factor in the Drosophila ovary. Nature. 2001;410(6828):599–604. pmid:11279500.
- 45. Sahai-Hernandez P, Nystul TG. A dynamic population of stromal cells contributes to the follicle stem cell niche in the Drosophila ovary. Development. 2013;140(22):4490–8. pmid:24131631; PubMed Central PMCID: PMC3817939.
- 46. Nystul T, Spradling A. An epithelial niche in the Drosophila ovary undergoes long-range stem cell replacement. Cell stem cell. 2007;1(3):277–85. pmid:18371362.
- 47. Kiger AA, White-Cooper H, Fuller MT. Somatic support cells restrict germline stem cell self-renewal and promote differentiation. Nature. 2000;407(6805):750–4. pmid:11048722.
- 48. Tulina N, Matunis E. Control of stem cell self-renewal in Drosophila spermatogenesis by JAK-STAT signaling. Science. 2001;294(5551):2546–9. pmid:11752575.
- 49. Michel M, Kupinski AP, Raabe I, Bokel C. Hh signalling is essential for somatic stem cell maintenance in the Drosophila testis niche. Development. 2012;139(15):2663–9. pmid:22745310.
- 50. Cook MS, Cazin C, Amoyel M, Yamamoto S, Bach E, Nystul T. Neutral Competition for Drosophila Follicle and Cyst Stem Cell Niches Requires Vesicle Trafficking Genes. Genetics. 2017;206(3):1417–28. pmid:28512187.
- 51. Castanieto A, Johnston MJ, Nystul TG. EGFR signaling promotes self-renewal through the establishment of cell polarity in Drosophila follicle stem cells. eLife. 2014;3. pmid:25437306; PubMed Central PMCID: PMC4298699.
- 52. Wu X, Tanwar PS, Raftery LA. Drosophila follicle cells: morphogenesis in an eggshell. Seminars in cell & developmental biology. 2008;19(3):271–82. pmid:18304845; PubMed Central PMCID: PMC2430523.
- 53. Horne-Badovinac S, Bilder D. Mass transit: epithelial morphogenesis in the Drosophila egg chamber. Developmental dynamics: an official publication of the American Association of Anatomists. 2005;232(3):559–74. pmid:15704134.
- 54. Flaherty MS, Salis P, Evans CJ, Ekas LA, Marouf A, Zavadil J, et al. chinmo is a functional effector of the JAK/STAT pathway that regulates eye development, tumor formation, and stem cell self-renewal in Drosophila. Developmental cell. 2010;18(4):556–68. pmid:20412771; PubMed Central PMCID: PMC2859208.
- 55. Ma Q, de Cuevas M, Matunis EL. Chinmo is sufficient to induce male fate in somatic cells of the adult Drosophila ovary. Development. 2016;143(5):754–63. pmid:26811385; PubMed Central PMCID: PMC4813340.
- 56. Siggs OM, Beutler B. The BTB-ZF transcription factors. Cell cycle. 2012;11(18):3358–69. pmid:22894929; PubMed Central PMCID: PMC3466544.
- 57. Ito H, Fujitani K, Usui K, Shimizu-Nishikawa K, Tanaka S, Yamamoto D. Sexual orientation in Drosophila is altered by the satori mutation in the sex-determination gene fruitless that encodes a zinc finger protein with a BTB domain. Proceedings of the National Academy of Sciences of the United States of America. 1996;93(18):9687–92. pmid:8790392; PubMed Central PMCID: PMC38490.
- 58. Xiong WC, Montell C. tramtrack is a transcriptional repressor required for cell fate determination in the Drosophila eye. Genes & development. 1993;7(6):1085–96. pmid:8504931.
- 59. Amoyel M, Anderson J, Suisse A, Glasner J, Bach EA. Socs36E Controls Niche Competition by Repressing MAPK Signaling in the Drosophila Testis. PLoS genetics. 2016;12(1):e1005815. pmid:26807580; PubMed Central PMCID: PMC4726490.
- 60. Leatherman JL, Dinardo S. Germline self-renewal requires cyst stem cells and stat regulates niche adhesion in Drosophila testes. Nature cell biology. 2010;12(8):806–11. pmid:20622868; PubMed Central PMCID: PMC2917891.
- 61. Pan Y, Robinett CC, Baker BS. Turning males on: activation of male courtship behavior in Drosophila melanogaster. PloS one. 2011;6(6):e21144. Epub 2011/07/07. pmid:21731661; PubMed Central PMCID: PMCPMC3120818.
- 62. Nagarkar-Jaiswal S, DeLuca SZ, Lee PT, Lin WW, Pan H, Zuo Z, et al. A genetic toolkit for tagging intronic MiMIC containing genes. eLife. 2015;4. Epub 2015/06/24. pmid:26102525; PubMed Central PMCID: PMCPMC4499919.
- 63. Pfeiffer BD, Jenett A, Hammonds AS, Ngo TT, Misra S, Murphy C, et al. Tools for neuroanatomy and neurogenetics in Drosophila. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(28):9715–20. Epub 2008/07/16. pmid:18621688; PubMed Central PMCID: PMCPMC2447866.
- 64. Sanders LE, Arbeitman MN. Doublesex establishes sexual dimorphism in the Drosophila central nervous system in an isoform-dependent manner by directing cell number. Developmental biology. 2008;320(2):378–90. pmid:18599032; PubMed Central PMCID: PMCPMC2631280.
- 65. Duncan IW, Kaufman TC. Cytogenic analysis of chromosome 3 in Drosophila melanogaster: mapping of the proximal portion of the right arm. Genetics. 1975;80(4):733–52. pmid:811500; PubMed Central PMCID: PMC1213372.
- 66. Le Bras S, Van Doren M. Development of the male germline stem cell niche in Drosophila. Developmental biology. 2006;294(1):92–103. pmid:16566915.
- 67. McKeown M, Belote JM, Boggs RT. Ectopic expression of the female transformer gene product leads to female differentiation of chromosomally male Drosophila. Cell. 1988;53(6):887–95. pmid:2454747.
- 68. Leatherman JL, Dinardo S. Zfh-1 controls somatic stem cell self-renewal in the Drosophila testis and nonautonomously influences germline stem cell self-renewal. Cell stem cell. 2008;3(1):44–54. pmid:18593558; PubMed Central PMCID: PMC2601693.
- 69. Li MA, Alls JD, Avancini RM, Koo K, Godt D. The large Maf factor Traffic Jam controls gonad morphogenesis in Drosophila. Nature cell biology. 2003;5(11):994–1000. pmid:14578908.
- 70. Celniker SE, Dillon LA, Gerstein MB, Gunsalus KC, Henikoff S, Karpen GH, et al. Unlocking the secrets of the genome. Nature. 2009;459(7249):927–30. pmid:19536255; PubMed Central PMCID: PMC2843545.
- 71. Sinden D, Badgett M, Fry J, Jones T, Palmen R, Sheng X, et al. Jak-STAT regulation of cyst stem cell development in the Drosophila testis. Developmental biology. 2012;372(1):5–16. pmid:23010510; PubMed Central PMCID: PMCPMC3481170.
- 72. Hempel LU, Oliver B. Sex-specific DoublesexM expression in subsets of Drosophila somatic gonad cells. BMC developmental biology. 2007;7:113. pmid:17935627; PubMed Central PMCID: PMC2148063.
- 73. Evans DS, Cline TW. Drosophila switch gene Sex-lethal can bypass its switch-gene target transformer to regulate aspects of female behavior. Proceedings of the National Academy of Sciences of the United States of America. 2013;110(47):E4474–81. pmid:24191002; PubMed Central PMCID: PMC3839706.
- 74. Sun X, Yang H, Sturgill D, Oliver B, Rabinow L, Samson ML. Sxl-Dependent, tra/tra2-Independent Alternative Splicing of the Drosophila melanogaster X-Linked Gene found in neurons. G3. 2015;5(12):2865–74. pmid:26511498; PubMed Central PMCID: PMCPMC4683657.
- 75. Bischof J, Maeda RK, Hediger M, Karch F, Basler K. An optimized transgenesis system for Drosophila using germ-line-specific phiC31 integrases. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(9):3312–7. pmid:17360644; PubMed Central PMCID: PMCPMC1805588.