The sexual identity of adult intestinal stem cells controls organ size and plasticity

SUMMARY

Sex differences in physiology and disease susceptibility are commonly attributed to developmental and/or hormonal factors, but there is increasing realisation that cell-intrinsic mechanisms play important and persistent roles^1,2. Here we use the Drosophila melanogaster intestine to investigate the nature and significance of cellular sex in an adult somatic organ in vivo. We find that the adult intestinal epithelium is a cellular mosaic of different sex differentiation pathways, and displays extensive sex differences in expression of genes with roles in growth and metabolism. Cell-specific reversals of the sexual identity of adult intestinal stem cells uncover its key roles in controlling organ size, its reproductive plasticity and its response to genetically induced tumours. Unlike previous examples of sexually dimorphic somatic stem cell activity, the sex differences in intestinal stem cell behaviour arise from intrinsic mechanisms, which control cell cycle duration and involve a new doublesex- and fruitless-independent branch of the sex differentiation pathway downstream of transformer. Together, our findings indicate that the plasticity of an adult somatic organ is reversibly controlled by its sexual identity, imparted by a new mechanism that may be active in more tissues than previously recognised.

Sex differences in intestinal physiology³ prompted us to investigate possible molecular underpinnings. RNA-seq transcriptional profiling of virgin adult midguts (GEO accession number GSE74775) indicated significant sexual dimorphism in gene expression and/or splicing, with 8.3% of all genes and 5.6% of all isoforms expressed in the midgut displaying sex differences in expression (Extended Data Fig. 1a-f and GutSexRNAseq.xls). Sex-biased expression or splicing was confirmed for a subset of genes by real time qRT-PCR (Extended Data Fig. 1h, i). Genes with sex differences in expression cluster into distinct biological processes; genes assigned to cell division-related processes are more abundantly expressed in females, whereas genes coding for proteins that function in carbohydrate metabolism and redox processes are preferentially expressed in males (Extended Data Fig. 1g).

The above results suggested active expression of sex determinants in the adult intestine. In Drosophila, sex chromosome sensing leads to sexually dimorphic expression of Sex lethal (Sxl): the master regulator of both sexual development and dosage compensation (DC, an epigenetic process by which transcription of the single male X chromosome is upregulated twofold⁴). Sxl controls the sex-specific splicing of its downstream target gene transformer (tra)⁵, leading to functional Tra protein expression only in females, and to the sex-specific splicing of two Tra direct targets – the transcription factors doublesex (dsx) and fruitless (fru) – that sculpt sexually dimorphic anatomical features, reproductive systems and behaviour^6-8 (Extended Data Fig. 2b). As may be expected from their roles in DC⁴, Sxl and the sexually dimorphic DC complex (DCC) member Male-specific lethal 2 (Msl-2) are ubiquitously expressed in the intestinal epithelium of females and males, respectively, in both adult intestinal progenitors (intestinal stem cells (ISCs) and postmitotic enteroblasts (EBs)) and their differentiated progeny (enterocytes (ECs) and enteroendocrine cells (EECs)) (Extended Data Fig. 2a-d). Consistent with our RNA-seq analysis, a newly generated tra-LacZ reporter is also expressed in all four epithelial cell types of the adult intestine (Extended Data Fig. 2a, e) - suggesting that, downstream of Sxl, the sex differentiation pathway remains active in this adult epithelium. However, the two targets of Tra-mediated sex-specific splicing are either absent from all midgut epithelial cells (Fru^M), or are expressed in only a subset of epithelial cells (Dsx^F/M, expressed in ECs but absent from ISCs, EBs and EECs) (Extended Data Fig. 2a, f, g). Hence, the expression of sex determination genes is maintained in the adult midgut, but displays cell type specificity; while adult-born enterocytes express all members of the canonical sex determination pathway, their siblings (the EECs) and both types of adult intestinal progenitors (ISCs and EBs) express the early (Sxl, Tra), but not the late (Dsx^F/M, Fru^M), effectors of the pathway.

Both the enrichment analysis and the presence of Sxl/tra in adult ISCs pointed to sexually dimorphic ISC proliferation. Female flies exhibit a rapid proliferative response to dextran sodium sulphate (DSS)-induced damage of the intestinal epithelium (Fig. 1a and ⁹). This response was less pronounced in male midguts (Fig. 1a), or in female (but not male) midguts following adult-restricted Sxl downregulation in intestinal progenitors (Fig. 1a and Extended Data Fig. 3a, c). Conversely, ectopic expression of Sxl in adult intestinal progenitors enhanced proliferation in male (but not female) midguts (Fig. 1a and Extended Data Fig. 3a). Additional cell type- and adult-specific Sxl downregulation experiments indicated that Sxl acts in ISCs, and not in other cells, to control sexually dimorphic proliferation rather than differentiation (Fig. 1e and Extended Data Figs. 3d and 4a). Mechanistically, females do not have a significantly higher density of ISCs than male flies (Fig. 1b) or a higher ratio of symmetric vs asymmetric divisions (Fig. 1c), suggesting that the proliferative capacity of female ISCs is intrinsically enhanced by their expression of Sxl. Consistent with this idea, a higher percentage of their adult progenitors are found in G2/S phase at the expense of G1 during in homeostatic conditions (Fig. 1d), suggestive of shorter cell cycles, and adult-specific downregulation of Sxl in intestinal progenitors abrogated the sexual dimorphism in G2/S to G1 ratio, without affecting the number of ISCs or their division mode (Figs. 1b-d). Clonal analyses further confirmed the intrinsic nature of the sexual dimorphism in proliferation, its Sxl control and adult reversibility both during regeneration and homeostasis (Fig. 1g and Extended Data Fig. 3b, e and f).

Figure 1. Sxl controls intrinsic sex differences in adult ISC proliferation independently of dosage compensation.

a, Number of mitoses (phospho-histone H3 (pH3)-positive cells, top graph) and percentage of cells positive for intestinal progenitor markers (escargot (esg)-positive/total area, bottom graph) in controls and flies with adult-restricted downregulation or mis-expression of Sxl in ISCs/EBs (achieved by esg-Gal4, tub-Gal80^TS-driven Sxl RNAi or UAS-Sxl, respectively). Flies were exposed to control (sucrose, S) or damage-inducing (DSS) diets. Representative images are shown to the right (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). b, Stem cell number (esg⁺, Supressor of Hairless (Su(H))⁻ cells) in the posterior midgut following 20 days of adult ISC/EB-specific Sxl downregulation. c, Quantifications of symmetric (red) vs asymmetric (green) ISC divisions based on cortical Partner of Numb (Pon)-GFP distribution in metaphase and telophase reveal no differences between the sexes or upon adult ISC/EB-specific Sxl downregulation. Mitoses with lack of clear Pon-GFP signal are displayed in yellow. d, Percentage of progenitors in G1, S or G2 as revealed by ISC/EB-driven expression of the cell cycle indicator Fly-FUCCI. e, Number of mitoses in DSS-treated flies with adult-specific downregulation of Sxl in ISCs (esg-Gal4, Su(H)-Gal80 driver), EBs (Su(H)-Gal4), ECs (midgut expression 1 (mex1-Gal4) or EECs (prospero (pros)^V1-Gal4). f, pH3 quantifications following adult-specific msl-2 downregulation or mis-expression in ISCs/EBs. g, MARCM clone size quantifications (graph) and representative images (labelled in green with GFP) reveal that clones expressing Sxl-RNAi are smaller than control clones in females, but not in males. n denotes the number of midguts (a, e, f), ISCs/EBs (b, d), mitoses (c) or clones (g) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes. In this and all subsequent figures, error bars correspond to standard error of the mean (SEM).

To investigate whether the reported Sxl effects result from deregulated DC, we first confirmed that DC can be functionally inactivated in adults by observing loss of histone H4 lysine 16 acetylation of the X chromosome upon adult-specific downregulation of msl-2 in male intestinal progenitors (Extended Data Fig. 3g). We then investigated whether ectopic msl-2 expression accounted for the reduced proliferation resulting from Sxl downregulation by co-downregulating both genes in adult intestinal progenitors. This did not reinstate female proliferation (Fig. 1f). The converse experiment - mis-expression of msl-2 in adult female intestinal progenitors using a newly generated transgene coding for HA-tagged Msl-2 - did not reduce their proliferation (Fig. 1f) despite efficient Msl-2 protein expression and function (Extended Data Fig. 3h and data not shown). Hence, DC does not account for the ISC sex differences.

This focused our attention on the sex differentiation pathway and its main effector tra. Like Sxl, tra downregulation reduced DSS-induced proliferation in females to levels comparable to those seen in male midguts, but did not affect proliferation in male midguts (Fig. 2a and Extended Data Fig. 5a). Conversely, tra mis-expression – either ubiquitous (Extended Data Fig. 5c) or confined to adult intestinal progenitors (Fig. 2a and Extended Data Fig. 5a) - increased the proliferative response of ISCs to DSS in adult males, but not in females. Clonal and tra mutant rescue experiments further confirmed the adult, cell-intrinsic requirement for tra in regulating sexually dimorphic proliferation, both during regeneration (Extended Data Fig. 5b, d) and in normal homeostasis (Fig. 2c and Extended Data Fig. 5e, f). Strikingly, reintroduction of a tra transgene specifically in adult intestinal progenitors fully rescued the reduced proliferation resulting from Sxl downregulation (Fig. 2b). Together with the Sxl experiments, these results show that the sex of the midgut is actively specified in adult flies. Unlike other adult somatic cell types^6,10,11, adult ISCs have a plastic sexual identity manifested by an intrinsic, Sxl/tra-controlled and DC-independent sexual dimorphism in both their basal and regenerative proliferation. These findings extend previous work in both mouse and Drosophila^12-16 by showing that sexual identity not only needs to be actively maintained, but can also be reversed bidirectionally in the adult cells of a nongonadal organ.

Figure 2. tra, but not tra2, controls intrinsic sex differences in adult ISC proliferation.

a, Mitoses (top graph) and intestinal progenitor area (bottom graph) in flies exposed to control (sucrose, S) or damage-inducing (DSS) diets, in both controls and flies with adult ISC/EB-restricted tra downregulation or mis-expression. Representative images are shown below the graphs (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). b, Comparable quantifications of the Sxl downregulation phenotypes in females, and their rescue by re-expression of tra^F. c, Clone size quantifications (in arbitrary units of GFP fluorescence, see Methods) reveal that tra null mutant MARCM clones are smaller than control clones only in females. d, No significant differences in ISC proliferation in the midguts of DSS-treated males or females lacking tra2 (tra2^B/Df(2R)trix), or lacking tra2 specifically in adults (achieved by shifting flies with the thermosensitive allele tra2^ts1 from 18°C to 29°C in the adult stage), vs controls. n denotes the number of midguts (a, e, f), ISCs/EBs (b, d), mitoses (c) or clones (g) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.

Unexpectedly (and in contrast to tra), mutation of the canonical Tra binding partner Transformer 2 (Tra2)^5,7 failed to affect the sex differences in ISC proliferation (Fig. 2d), despite interfering with the sex-specific expression of dsx^F and Yolk protein 1 (Yp1) transcripts as anticipated (Extended Data Fig. 5g). Together with our finding that ISCs do not express Dsx or Fru^M (Extended Data Fig. 2a, f, g), this result suggested that a non-canonical Tra2-, Dsx-, Fru^M-independent sex determination pathway drives sexually dimorphic proliferation in adult intestinal stem cells. We conducted a series of rescue, gain- and loss-of-function experiments using dsx^F transgenes, dsx/fru mutants, RNAi transgenes and mutants in which dsx/fru splicing was biased towards either male or female isoforms (Extended Data Fig. 6a-g). These ruled out both a direct action of dsx or fru^M in intestinal progenitors as well as indirect contributions to ISC proliferation from other dsx/fru^M-expressing cells such as the neighbouring ECs, thus confirming that the sex of adult somatic stem cells in the intestine is specified by a novel branch of the sex determination pathway.

We then combined new RNA-seq and genetic experiments to identify tra target genes in adult ISCs, defined as genes the expression or splicing of which was different between control and tra mutant females, but returned to levels comparable to those of controls upon adult ISC-specific re-introduction of tra (see Methods, GEO accession number GSE74775 and GutSexRNAseq.xls for details). This yielded 72 genes with tra-regulated expression (34) or splicing (38) (Fig. 3a, b and Extended Data Fig. 7a-c). A genetic screen (Extended Data Fig. 7d, see Methods for details) identified three of them as negative or positive regulators of proliferation accounting for the sexual dimorphism in ISC proliferation. Indeed, downregulation of Imaginal disc growth factor 1 (Idgf1) or Serpin 88Eb (Spn88Eb), normally activated by tra in adult intestinal progenitors, reduced proliferation only in females. By contrast, over-expression of reduced ocelli (rdo), normally repressed by tra in the same cells, reduced proliferation only in females (Fig. 3c and Extended Data Fig. 7e). These three targets are regulated at the expression – rather than splicing – level: consistent with novel, tra2- and, possibly, splicing-independent roles of tra. Once produced, they may be secreted or targeted to the cell surface^17-19, suggesting that sex differences may result from modulation of how ISCs interact with their local environment. All three proteins also appear to be expressed in other tissues and/or during development, and belong to protein families also represented in mammals: leucine-rich repeat (rdo), inhibitory serpins (Spn88Eb) and secreted glycoproteins (Idgf1)^17,19,20. It will therefore be interesting to explore whether they account for how organisms attain their sexually dimorphic body size during development.

Figure 3. tra targets in adult intestinal progenitors.

Heat maps of the genes with tra-regulated expression (a) or splicing (b) in ISCs/EBs, displaying their normalized abundance in females, tra null mutant females and tra null mutant females with feminized ISCs/EBs (esg^TS>tra^F) (see Methods for details). c, Mitoses (pH3) quantifications of adult ISC/EB-confined manipulations of rdo, Idgf1 and Spn88Eb expression. Representative images for each genotype are shown below the graphs (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). n = 10 midguts per genotype/condition. Results combined from at least two independent experiments. See GutSexRNAseq.xls file for a full list of names and quality scores, and Supplementary Information for full genotypes.

Are the sex differences in ISCs physiologically significant? Wild-type female midguts are longer than male midguts at days 3, 5 and 20 of adult life (Fig. 4a and data not shown). Adult-specific and cell-autonomous masculinisation of ISCs (achieved by tra or Sxl downregulation) dramatically shrank the female (but not the male) midgut towards a male-like size (Fig. 4a and Extended Data Fig. 8c). Cell type- and region-specific quantifications indicated that a reduction in ISC proliferation and, consequently, the number of differentiated progeny accounted for the observed shrinkage (Extended Data Fig. 8a-d). Female flies with masculinised adult ISCs also failed to undergo the recently described midgut resizing triggered by mating²¹ (Fig. 4b and Extended Data Fig. 8e) and had slightly reduced fecundity (Extended Data Fig. 8f), indicating that sex differences endow females with enhanced stem cell plasticity to optimise reproduction. Females are, however, more prone to genetically induced tumours; indeed, adult-specific interference with Notch (N) or Apc-ras signalling, previously shown to lead to tissue overgrowth reminiscent of gastrointestinal tumours in females^22,23, resulted in hyperplasia of female (but not male) midguts (Fig. 4c, d and Extended Data Fig. 9a-f). Hyperplasia resulted from sex differences in proliferation rather than sex-specific differentiation defects (Extended Data Fig. 4b), and could be prevented by simultaneously masculinising female adult ISCs by downregulating or mutating Sxl/tra, but not dsx or fru^M (Fig. 4c, d and Extended Data Fig. 9a, c-g). Increased susceptibility of female flies to genetically induced tumours was also observed after mating (data not shown). Thus, the intrinsic sexual identity of adult intestinal stem cells plays key roles in adult life, both in maintaining organ size and in modulating its plasticity.

Figure 4. Physiological significance of the sex differences in intestinal progenitors.

a, Midgut length quantifications and representative images of phenotypes resulting from adult-specific masculinization of ISC/EB-specific masculinization of ISCs (achieved by esg^TS-driven tra downregulation initiated after the phase of midgut post-eclosion growth, see Methods for details). b, The number of mitoses (pH3-positive cells) is higher in control female flies 3 days after mating. The postmating increase is abrogated upon adult ISC/EB-specific tra downregulation. An EB marker (Su(H)LacZ, in red in image panels) reveals that the EB expansion seen in females after mating is reduced upon adult ISC/EB-specific tra downregulation. See also Extended Data Fig. 8 for quantifications. c, pH3 quantifications inside MARCM clones of control flies, Su(H) mutants and Su(H) mutants in which Sxl has been downregulated inside the clone. Su(H) mutation only leads to increased pH3 counts in females, and this increase is Sxl dependent. d, Hyperplasia (quantified by the number of pH3-positive cells) resulting from adult ISC/EB-driven N downregulation and its modulation by tra in female and male midguts. Confocal images show intestinal progenitor coverage of representative midgut portions for each genotype (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). n denotes the number of midguts (a, b, c, d) that were analysed for each genotype. Virgin flies were used in all experiments unless otherwise indicated. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.

Previous observations had pointed to additional branches of the sex determination pathway^24,25. Our discovery of this new branch has important implications for organs such as the nervous system, where sexual identity was thought to be confined to Fru^M and/or Dsx-expressing neurons²⁶, and raises the possibility that every cell has a sexual identity that actively regulates its plasticity and physiology. While early sex specification in Drosophila differs from that of mammals, there is increasing evidence that it converges on common effectors - such as the Dsx/Dmrt family of transcription factors and their targets¹². Hence, the recently reported sex differences in both intestinal gene expression and microbiota in the mammalian intestine^27,28 could at least partly result from similar, so far unexplored, intrinsic sex differences. Similarly, the possible contribution of the intrinsic sexual identity of adult somatic stem cells to both organ plasticity and to the well documented sex differences in susceptibility to many types of cancer^29,30 deserves further investigation.

METHODS

Fly husbandry

Fly stocks were reared on a standard cornmeal/agar diet (6.65% cornmeal, 7.15% dextrose, 5% yeast, 0.66% agar supplemented with 2.2% nipagin and 3.4 mL/L propionic acid). All experimental flies were kept in incubators at 25°C, 65% humidity and on a 12 hr light/dark cycle, except for those containing tub-Gal80^TS transgenes, which were set up at 18°C (restrictive temperature) and transferred to 29°C (permissive temperature) at the time when Gal4 induction was required. This depended on the specific experimental requirements but typically, for loss-of-function (RNAi) or gain-of-function (UAS) experiments, flies were raised and aged as adults for 3 days at 18°C, were then shifted to 29°C to induce transgene expression, and adult midguts were dissected after 10-20 days (as indicated in each figure panel). Flies were transferred to fresh vials every 3 days and fly density was kept to a maximum of 15 flies per vial. Virgin flies were used for all experiments unless indicated otherwise.

For mutant ISC clonal analyses (MARCM clones), 3-day-old adults (raised and aged at 25°C) were heat-shocked for 1 hr at 37°C to induce clones, and were then kept at 25°C (or 29°C for MARCM RNAi clones) until dissection (10 days, 15 days or 4 weeks thereafter as indicated in each figure panel). Flies were transferred to fresh vials every 3 days.

For damage-induced regeneration assays, virgin flies were collected over 72 hrs at 18°C, and were then shifted to 29°C for 7 days on standard media. Flies were then transferred in an empty vial containing a piece of 3.75 cm × 2.5cm paper. 500mL of 5% sucrose solution (control) or 5% sucrose + 3 dextran sulfate sodium (DSS) solution was used to wet the paper, used as feeding substrate. Flies were transferred to new vial with fresh feeding paper every day for 3 days prior dissection.

For fecundity experiments, females were raised and aged as virgins for 3 days at 18°C, and were then shifted to 29°C to induce transgene expression for 10 days. Females were then mated overnight to OregonR males (10 males, 10 females per vial). Males were then removed, and single female flies were transferred individual vials every 24h for a 3-day period. Eggs were counted from the vacated vials.

Fly stocks

UAS transgenes

UAS-Sxl^RNAi (VDRC: GD 3131), UAS-Sxl (generated by ³¹), UAS-pon.GFP (BDSC:42741, generated by ³²), UAS-GFP.E2f1.^1-230, UAS-mRFP1.NLS.CycB.^1-266 (BDSC: 55121, generated by ³³), UAS-GFP (BDSC: 35786), UAS-Dicer2 (VDRC: 60007), UAS-tra^RNAi (BDSC: 28512, TRiP.JF03132), UAS-tra^F (BDSC: 4590, generated by ³⁴), UAS-msl-2^RNAi (VDRC: GD 29356), UAS-Sxl^{RNAi 2} (VDRC: KK 109221), UAS-msl-2^{RNAi 2} (BDSC:35390, TRiP.GL00309), UAS-msl-2^HA (this study), UAS-dsx^RNAi (BDSC: 26716, TRiP.JF02256), UAS-dsx^F (BDSC: 44223, generated by ³⁵), UAS-Arr1 (generated by ³⁶), UAS-rdo (generated by ¹⁷), UAS-GstE4 (generated by ³⁷), UAS-Notch^RNAi (VDRC: KK 100002), UAS-Sxl^{RNAi 3} (BDSC: 34393, TRiP.HMS00609), UAS-tra^{RNAi 2} (VDRC: GD 2560), UAS-tra^{RNAi 3} (NIG: 16724-R2), UAS-fru^RNAi (BDSC: 31593, TRiP.JF01182), UAS-Notch^{RNAi 2} (VDRC: GD 27229), UAS-Hairless (generated by ³⁸), UAS-N^intra (generated by ³⁹), UAS-sspitz (gift from J. Treisman, generated by ⁴⁰), UAS-Ras^V12; FRT82B line (generated by ⁴¹).

Mutants

tra¹, FRT2A, fru^P1-Gal4 chromosome (generated by ⁴²), FRT82b, dsx¹ chromosome (generated by ⁴³), dsx^Δ, Df(3R)^Exel6179 chromosome (generated by ⁴⁴), Df(3R)dsx¹¹ (BDSC: 1865, generated by ⁴⁵), In(3R)dsx²³ (BDSC: 1849, generated by ⁴⁶), FRT82b, dsx¹, fru^Δtra chromosome (gift from B. Baker, generated by ⁴⁷), dsx¹, fru^P1-LexA chromosome (gift from B. Prud’homme, generated by ⁴⁸), tra^KO (this study), Df(3L)^st-j7 (BDSC:5416, generated by ⁴⁹), tra¹ (BDSC: 675, generated by ⁵⁰), tra2^B (BDSC: 25137, generated by ⁵¹), Df(2R)^trix (BDSC:1896, generated by ⁵²), tra2^ts1 (BDSC: 2413, generated by ⁵²), dsx^D (BDSC: 840, generated by ⁵³, fru^F and fru^M (gift from G. Jefferis, generated by ⁴⁷, fru^4-40 (gift from G. Jefferis, generated by ⁵⁴), FRT82b, DI^RevF10 chromosome (gift from S. Bray, generated by ⁵⁵), Su(H)^Δ47, FRT40A chromosome (gift from S. Bray, generated by ⁵⁶), FRT82B Apc2^N175K, Apc^Q8 chromosome and hs-flp;UAS-Ras^V12; FRT82B Apc2^N175K, Apc^Q8 line (both generated by ⁴¹).

Reporters and Gal4 drivers

Su(H)^GBE-LacZ (generated by ⁵⁷), esg-GFP^P01986 (gift from L. Jones, Flytrap), esg-LacZ^k00606 (BDSC : 10359), dsx-Gal4 (generated by ⁵⁸), fru^P1-Gal4 (gift from G. Jefferis, generated by ⁵⁹), esg-Gal4^NP7397, UAS-GFP, Tub-Gal80^TS chromosome (gift from J. de Navascués), Su(H)^GBE-Gal80 (generated by ⁶⁰), Su(H)^GBE-Gal4 (generated by ⁶¹), mex1-Gal4 (generated by ⁶²), pros^V1-Gal4 (generated by ⁶³), Mvl-Gal4^NP2375(Kyoto: 104178), dsx^Δ2-Gal4 (generated by ⁶⁴), vm-Gal4 (BDSC: 48547, GMR13B09-GAL4), btl-Gal4 (generated by ⁶⁵), nSyb-Gal4 (BDSC: 51941), elav-Gal4 (BDSC: 8765), nSyb-Gal80 (gift from J. Simpson), stripe-Gal4 (BDSC: 26663).

MARCM stocks

FRT40A: w, hs-flp,Tub-Gal4, UAS-GFP; tub-Gal80, FRT40A (gift from J. de Navascués), FRT2A: y, w, hs-flp; Tub-Gal4,UAS-GFP; Tub-Gal80, FRT2A (gift from I. Salecker), FRT2A (BDSC: 1997), FRT82b: w, hs-flp; UAS-mCD8GFP, Tub-Gal4; FRT82b, Tub-Gal80 (gift from M. Vidal), FRT82b (BDSC:7369).

Generation of tra^KO and tra^LacZ transgenic lines

To generate the transgenic reporter of tra promoter activity tra^LacZ, the upstream intergenic region between spd-2 and tra was cloned using the following primer pair: 5’-AAAATCTAGAAACTAATAAAGTATATGAG-3’ and 5’-AAAAGGTACCCGGAAAATGCTGGAAATTAAATGATGC-3’. PCR was performed with Q5 high-fidelity polymerase from New England Biolabs (M0491S). The PCR product was digested with XbaI and KpnI prior to cloning into the pH-Pelican attb vector (⁶⁶, gift from C. Thummel). The construct was sequence-verified and a transgenic line was established through ΦC-31 integrase mediated transformation (Bestgene, attP site: attP40).

The new amorphic allele of tra, tra^KO, was generated using the accelerated homologous recombination method recently developed by Baena-Lopez⁶⁷. The 5’ (4302nt) and 3’ (2928nt) homology arms were produced from a tra BAC (BAC-PAC resources: CH322-140P01) with the following primer pairs respectively: 5’-AAAAGCGGCCGCCATTCTACTCTTGAATTGGCTAGC-3’/5’-AAAAGTACCATGATGCACTTTCCTCAGTGTGA-3’ and 5’-AAAAGGCGCGCCAAGAGAATACCATGG-3’/5’-AAAAGGCGCGCCATTGTCGACACAATCAAACTG-3’. PCRs were performed with Q5 high-fidelity polymerase from New England Biolabs (M0491S). The PCR products were digested with NotI/KpnI or AscI respectively prior to cloning into the pTV^Cherry vector to generate pTV^Cherry[tra]. The vector was sequence-verified and inserted into random genomic locations by P-element-mediated transformation (Bestgene). Transformants (not necessarily mapped or homozygosed) were crossed to hs-FLP, hs-SceI flies (DBSC: 25679) and the resulting larvae were heat-shocked at 48, 72, 96 and 120 hrs after egg laying for 1 hr at 37°C. Approximately 200 adult females with mottled eyes (indicating the presence of pTV^Cherry[tra] and the transgenes carrying hs-FLP and hs-SceI) were crossed in pools of 15 to ubiquitin-Gal4[3xP3-GFP] males and the progeny was screened for the presence of red-eyed flies. The ubiquitin-Gal4[3xP3-GFP] transgene was subsequently removed by selecting against the presence of GFP in the ocelli. The generated deletion removed 342 nucleotides starting 13 nucleotides upstream of the transcription start site.

Generation of the UAS-msl-2^HA line

To overexpress msl-2, a transgenic UAS line was generated from msl-2 cDNA adding an HA tag at the C-terminal (BDGP Gold cDNAs, clone ID: GH22488) with the following primer pair: 5’-AAA AAGATCTATGGCTCAGACGGCATACTTG and 5’-AAAATCTAGATTAAGCGTAATCTGGAACATCGTATGGGTACAAGTCATCCGAGCCCGAC-3’. PCR was performed with Q5 high-fidelity polymerase from New England Biolabs (M0491S). The PCR product was digested with BglII and XbaI prior to cloning into the pUASTattb vector⁶⁸. The construct was sequence-verified and a transgenic line was established through ΦC-31 integrase mediated transformation (Bestgene, attP site ZH-86Fb, DBSC: 24749).

Immunohistochemistry

Intact guts were fixed at room temperature for 20 min in PBS, 3.7% formaldehyde. All subsequent incubations were done in PBS, 4% horse serum, 0.2% Triton X-100 at 4°C following standard protocols.

The following primary antibodies were used: mouse anti-Sxl (M114, DSHB Hydridoma) 1/50, mouse anti-Sxl (M18, DSHB Hydridoma) 1/50, goat anti-Msl-2 (dC-20, sc-32459, Santa Cruz Biotechnology) 1/50, chicken anti-beta galactosidase (ab9361, Abcam) 1/200, rabbit anti-phospho-Histone H3 Ser10 (9701L, Cell Signalling Technology) 1/500, rabbit anti-fru^M (Male-2, generated by ⁶⁹) 1/500, mouse anti-Dsx^DBD (DSHB Hybridoma) 1/100, mouse anti-Prospero (MR1A, DSHB Hybridoma) 1/50, goat anti-ac-Histone H4 Lys16 (sc-8662, Santa Cruz Biotechnology) 1/ 500, rat anti-HA (11867423001, Roche) 1/500, mouse anti-GFP (11814460001, Roche) 1/1000, mouse anti-Pdm1 (kind gift of Steve Cohen, generated by ⁷⁰) 1/20. Fluorescent secondary antibodies (FITC-, Cy3- and Cy5-conjugated) were obtained from Jackson Immunoresearch. Vectashield with DAPI (Vector Labs) was used to stain DNA.

Cell, clone and midgut length quantifications

Mitotic indices were quantified by counting phospho-Histone H3-positive cells in >10 midguts per genotype, time point, and/or condition (e.g. male/female), and are displayed as means ± standard error of the mean (SEM). For posterior midgut cell counts, a midgut portion immediately anterior to the junction with the tubules and hindgut was imaged at 20× magnification. Cells were counted using ImageJ in areas of identical size across all genotypes to control for size differences. Threshold was adjusted for the GFP channel (ImageJ function: Image>Adjust>Threshold) to subtract background, and the percentage of area above the threshold were considered (ImageJ function: analyse particles). Data was collected from at least 10 midguts per genotype and/or condition, and is displayed as mean of % of GFP-positive area ± SEM.

MARCM clones were quantified as number of cells per clone (when the GFP reporter was expressed in the nucleus) or by the size of the GFP area/clone (in arbitrary units, when the reporter was membrane-tagged GFP).

To measure midgut length, guts were dissected and were then straightened on polyLysine-coated slides. After imaging, a line was draw between the most anterior point of the proventriculus and the midgut-hindgut junction using ImageJ. The number of pixels contained in the line was used as a proxy for midgut length. For display purposes, the value obtained for a control female midgut selected at random was used as reference to normalise all the other values, which are shown as percentage of control female length value.

Functional RNAi screen of tra targets

To test the functional significance of tra targets, only those with female/tra null mutant female fold differences in transcript abundance >±2 that were also under tra control in adult intestinal progenitors were selected for functional validation. The contribution of these tra targets to sex differences in adult ISC proliferation was investigated using publicly available RNAi and UAS lines (see below for details of the RNAi lines, and Methods section ‘Fly stocks’ for details of the UAS lines) expressed from esg-Gal4^ts. Results obtained with 11 out of the 16 lines belonging to the VDRC KK collection were not considered and are not shown in Extended Data Fig. 7 because their expression in adult ISCS from esg^TS-Gal4 resulted in a recurrent ISC differentiation phenotype in both males and females. We attribute this effect to the previously reported dominant Gal4-dependent toxicity issue with this VDRC KK collection⁷¹).

Gene	Transformant ID	Library
Idgf1	12416	GD
Snp88Eb	28425	GD
CG17470	100414	KK
CG8008	4159	GD
Klp54D	100140	KK
CG4500	34852	GD
CG4500	106260	KK
Arr1	22196	GD
Arr1	109860	KK
Arr1		UAS
rdo	107213	KK
rdo		UAS
Spn47C	25534	GD
Spn47C	105933	KK
Cyp12d1-p	49269	GD
Cyp12d1-p	109256	KK
TwdlE	24867	GD
TwdlE	107483	KK
Cyp12d1-d	109248	KK
Cyp12d1-d	50507	GD
GstE2	32945	GD
GstE4	100986	KK
GstE4	20472	GD
GstE4		UAS
lectin-28C	45634	GD
lectin-28C	104290	KK
AANATL2	44677	GD
AANATL2	102802	KK
CG5348	1698	GD
CG5348	106565	KK
CG15128	100238	KK
CG15128	19302	GD
CG15236	101045	KK
Ir10a	45403	GD
Ir10a	100181	KK

Statistics and data presentation

All statistical analyses were carried out in the R environment⁷². Comparisons between two genotypes/conditions were analyzed with the Mann-Whitney-Wilcoxon rank sum test (R function wilcox.test). All graphs were generated using Adobe Illustrator. All confocal and bright field images belonging to the same experiment and displayed together in our figure were acquired using the exact same settings. For visualization purposes, level and channel adjustments were applied using ImageJ to the confocal images shown in the figure panels (the same correction was applied to all images belonging to the same experiment), but all quantitative analyses were carried out on unadjusted raw images or maximum projections. In all figures, n denotes the number of midguts, ISCs/EBs, mitoses or clones that were analysed for each genotype. Values are presented as average ± standard error of the mean (SEM), p-values from Mann-Whitney-Wilcoxon test (non-significant (ns): p>0.05; *: 0.05>p>0.01; **: 0.01>p>0.001; ***: p>0.001). Lines and asterisks highlighting significant comparisons across sexes are displayed in red, whereas those highlighting significant comparisons within same-sex datasets are displayed in black.

RNA-seq experiments

RNA extraction

RNA from 30 pooled dissected midguts was extracted using Trizol (Invitrogen), and 3 such samples were used for each sex and genotype. RNA-seq libraries were prepared from 500ng of total RNA using the Illumina Truseq mRNA stranded library prep kit (Illumina Inc. San Diego, USA) according to the manufacturer’s protocol. Library quality was checked on a Bioanalyser HS DNA chip and concentrations were estimated by Qubit measurement. Libraries were pooled in equimolar quantities and sequenced on a Hiseq2500 using paired end 100bp reads. At least 30 million reads passing filter were achieved per sample. After demultiplexing, raw RNASeq reads were aligned with Tophat splice junction mapper⁷³, version 2.0.11 against Ensembl Drosophila genome reference sequence assembly (dm3) and transcript annotations.

For differential gene expression analysis, gene-based read counts were then obtained using HTSeq count module (version 0.5.3p9)⁷⁴. Differential expression analysis was performed on the counts data using DESeq2 Bioconductor package⁷⁵. The analysis was run with the default parameters. DESeq2 package uses negative binomial model to model read counts and then performs statistical tests for differential expression of genes. Raw p values were then adjusted for multiple testing with the Benjamini-Hochberg procedure. GO enrichment analysis was performed using FlyMine v40.1.b. In Fig. 3a, b and Extended Data Fig. 1g, each column corresponds to one of three different replicates (30 midguts each) for each sex, and transcript abundance for each gene was normalized to a scale of −2 (white) to +2 (gray).

Isoform expression analysis

Reads were first aligned to the transcript sets using Bowtie software (version 1.0.0)⁷⁶. Transcript annotations for cDNA and non-coding RNAs were obtained from Ensembl (release 75). Isoform abundances were then calculated using mmseq R package (version 1.0.8)⁷⁷. Differential expression analysis was performed on these abundances using the DESeq2 Bioconductor package (version 1.42.1)⁷⁸. Isoforms not expressed in any of the samples were not included in the differential expression analysis. Raw p values were adjusted for multiple testing with the Benjamini-Hochberg procedure. Initial isoform analysis identified differential expression for 1379 isoforms between male and female midguts (p<0.05 cutoff), which, following subtraction of those corresponding to single-isoform genes, yielded 704 isoforms (see main text for subsequent analysis).

RNA-seq data displays

For the volcano plots in Extended Data Fig. 1, log 2 fold change values were plotted against log 10 of adjusted p values. Genes and isoforms significantly upregulated in males are coloured blue, while those significantly upregulated in females are coloured red. Selected genes and isoforms are further highlighted as empty circles. For the scatter charts with quadrants in Extended Data Fig. 7, log 2 fold change values (between control and tra mutant females) were plotted against log 2 fold change values (between tra mutant females and tra mutant females and female with adult-specific, progenitor-specific tra expression). Genes and isoforms significantly repressed by tra are coloured blue, whereas those significantly activated by tra are coloured red. Selected genes and isoforms are further highlighted as empty circles. Volcano plots and scatter charts were generated using Adobe Illustrator. To generate the heat maps, a matrix containing the relative values of gene/isoform expression was built. Transcript abundance for each gene/isoform was normalized to a scale ranging from −2 to +2. A hierarchical clustering algorithm (with Euclidian distance and average linking) was applied to the matrix using the MeV software suite⁷⁹. Area-proportional Venn diagrams were generated using BioVenn (http://www.cmbi.ru.nl/cdd/biovenn/).

Real time qRT-PCRs

RNAs were extracted from 30 dissected midguts using Trizol (Invitrogen). RNAs were cleaned using RNAeasy mini Kit (QIAGEN), and cDNAs were synthesized using the iScript cDNA synthesis kit (Bio-Rad) from 300ng of total RNAs. Quantitative PCRs were performed by mixing cDNA samples (5ng) with iTaq^™ Universal SYBR^® Green Supermix (Bio-Rad, #172-5124) and the relevant primers in a 96-well plate. Expression values were normalized to eIF4G. For each gene/isoform, at least 3 independent biological replicates were used, and 2 technical replicates were performed. See Supplementary Information for a list of all primers used.

Extended Data

Extended Data Fig. 1. Sexually dimorphic transcription and splicing in the adult midgut.

a, Number and percentage of genes with sexually dimorphic gene expression, as revealed by RNA-seq transcriptional profiling of virgin male and female dissected midguts (p<0.05 cutoff). b, Volcano plot displaying all genes with detectable midgut expression. Female/male ratio of gene expression is shown on the X axis (in log2 scale) and significance is displayed on the Y axis as the negative logarithm (log10 scale) of the adjusted p-value. Genes with significantly upregulated (p<0.05 cutoff) expression in males and females are coloured in blue and red, respectively. Other genes are displayed in black. Genes with known sex-specific transcription are displayed as red (female-enriched) or blue (male-enriched) open circles. c and d, Comparable analyses for sex-biased isoforms belonging to genes with multiple transcripts. We identity 714 sex-biased isoforms belonging to a total of 603 genes. Isoforms resulting from known sex-specific alternative splicing are displayed as in panel b. e, Female/male ratios of overall transcript abundance (left graph) and abundance of sex-biased isoforms (right graph) for the members of the Drosophila sex determination pathway as revealed by RNA-seq analysis of the adult midgut. We note a sexual dimorphism in dsx transcript levels. f, Venn diagram illustrating the overlap between the genes showing sex-biased expression (overall transcript abundance, light grey, 1305 genes) and sex-biased alternative splicing (sex-biased isoforms, dark grey, 603 genes) in the adult midgut. Known members of the sex determination pathway are displayed as examples. g, Heat maps displaying genes with sexually dimorphic expression clustered by enrichment in specific biological processes, as revealed by Gene Ontology enrichment analysis. Genes with sexually dimorphic expression belonging to the top 4 enriched biological processes are shown. h and i, Real-time qRT-PCR data for a subset of genes for which RNA-seq transcriptional profiling experiments revealed sexually dimorphic expression (h) or isoforms (i). RNA was obtained from midguts from virgin male and female samples (same genotypes as for the RNA-seq experiments). For each gene/isoform, expression abundance was arbitrarily set up at 100% for the sex with the highest expression level, and percentage of that expression is displayed for the other sex. See Methods for details, GutSexRNAseq.xls for a full list of names and quality scores, and Supplementary Information for full genotypes.

Extended Data Fig. 2. Cell type-specific expression of sex determinants in the adult intestinal epithelium of virgin flies.

a, In the adult Drosophila midgut, resident stem cells (ISCs) and their postmitotic daughter cells (EBs) maintain the adult intestinal epithelium during normal homeostasis and regenerate it after injury by giving rise to two types of differentiated progeny: ECs and EECs^80,81. The posterior midgut area used to visualise and quantify phenotypes is boxed. The following Gal4 drivers were used to label and/or genetically manipulate these four different cell types present in the adult intestinal epithelium: midgut expression 1 (mex1) for ECs, prospero (pros^v1) for EECs, escargot (esg) to target both ISCs and EBs, and Supressor or Hairless (Su(H)) for EBs alone. b, The canonical sex determination pathway in somatic cells of Drosophila melanogaster, consisting of a cascade of sex-specific alternative splicing events culminating in the production of sex-specific transcription factors encoded by Dsx and Fru^M. In females, Sex lethal (Sxl) is activated and regulates the splicing of transformer (tra) pre-mRNA, resulting in the production of Tra^F. Tra^F regulates the female-specific splicing of dsx pre-mRNA (dsx^F) and fru transcript coming from the P1 promoter (fru^P1, giving rise to fru^M). In males, Sxl is not expressed and no functional Tra is produced, resulting in default splicing of dsx and fru pre-mRNAs, leading to Fru^M and Dsx^M proteins, respectively. The resulting male- and female-specific Dsx and Fru isoforms confer sexual identity to the cells in which they are produced. In addition, in females, Sxl represses dosage compensation by inhibiting Msl-2 expression. The tables summarize the cell-specific expression profiles of the sex determinants in adult midguts of virgin males and females, shown in the panels below. c, Sxl protein (in red) is expressed only in female midguts. Co-staining with ISC/EB reporters indicates that Sxl is found esg-positive progenitors (ISCs: GFP-positive and LacZ-negative cells, and EBs: GFP-positive and LacZ-positive cells). It is also expressed in female polyploid ECs (GFP- (in green) and LacZ- (in blue) negative cells). Co-staining with pros^V1 reporter indicates that it is also expressed in EECs. d, Msl-2 protein is found in the same cell types only in males (staining is confined to the X chromosome, consistent with the signal observed in non-intestinal tissues⁸². e, A new reporter of tra promoter activity (tra^LacZ, see Methods for details) is broadly expressed in the epithelium of both male and female midguts, including ISCs and EBs (as revealed by co-staining with esg-Gal4-driven GFP) and ECs (GFP-negative cells with large nuclei). Co-staining with pros^V1 reporter indicates that it is also expressed in EECs. f, A dsx-Gal4 reporter (visualised with a GFP reporter that has been false-coloured in red for consistency with the other panels) is active in male and female polyploid ECs (LacZ-negative cells), and is inactive in esg-positive progenitors (LacZ-positive cells, in green). Dsx protein (visualised using a Dsx^DBD-specific antibody in red) is expressed strongly only in males bot not females, indicating that the sexual dimorphism in dsx transcript levels found in our RNA-seq analyses (Extended Data Fig. 1e) is further enhanced at the protein level. Co-staining of the same antibody with the EC marker mex1-Gal4 confirms expression in ECs. Cytoplasmic dsx-Gal4-driven expression of an mTomato reporter is apparent in ECs (visualised as large cells by co-staining with the membrane-enriched marker Armadillo), but is absent from EECs (as revealed by the gaps in mTomato expression in cells that are labelled with Pros). g, The fru^P1-Gal4 reporter (which labels the only sexually dimorphic fru transcript that gives rise to Fru^M protein) is inactive in both male and female midguts, as revealed by lack of GFP signal (false-coloured in red for consistency with other panels). Consistent with the lack of fru^P1-Gal4 expression, a Fru^M-specific antibody (in red) is not expressed in the male midgut. An independent dsx reporter (dsx^Δ2-Gal4) is expressed in polyploid ECs, consistent with the data displayed in f. See Supplementary Information for full genotypes.

Extended Data Fig. 3. Sxl controls intrinsic sex differences in adult ISC proliferation independently of dosage compensation.

a, Immunocytochemistry using a Sxl-specific antibody indicates that adult-restricted downregulation of Sxl in intestinal progenitors (ISCs/EBs) - achieved by Gal80^TS-controlled expression of a Sxl RNAi transgene – efficiently downregulates Sxl expression in progenitors, but not large polyploid ECs. Conversely, efficient ectopic Sxl protein expression is obtained by expression of a UAS-Sxl transgene in adult ISCs/EBs of male flies. In all panels Sxl antibody is in red; DNA: DAPI, in blue; ISC/EB marker: GFP, in green. b, Quantifications of the number of cells inside control MARCM⁸³ clones, or MARCM clones expressing RNAi transgenes directed against Sxl following DSS treatment. 5 days after clone induction by heat shock, female clones are larger than male clones only when Sxl is present, confirming the cell autonomy of Sxl action. c, Additional controls for Fig. 1a, and confirmation of phenotypes using an independent RNAi transgene. This second RNAi transgene against Sxl (different from the one used in Fig. 1a) reduces the number of pH3-positive cells in DSS-treated female midguts when expressed from esg^TS in adults ISCs/EBs, confirming an adult progenitor-specific requirement for Sxl in promoting damage-induced cell divisions in female flies. d, Adult-specific downregulation of Sxl in adult visceral muscle (using the vm driver), trachea (btl-Gal4 and DSRF-Gal4), neurons (nSyb-Gal4, Elav-Gal4), or fat body (stripe-Gal4) does not reduce DSS-induced ISC proliferation in females. By contrast, Sxl downregulation using an ISC/EB driver with suppressed neuronal expression (esg-Gal4 combined with nSyb-Gal80) effectively reduces DSS-induced ISC proliferation in females. Together, these results indicate that Sxl does not control sexually dimorphic DSS-induced ISCs proliferation from non-ISC cells. e, MARCM clones expressing a third RNAi transgene against Sxl (distinct from those used in Fig. 1 and above) are smaller than control clones in females, whereas their size is comparable to that of wild-type or Sxl-RNAi clones in males. This confirms that, during normal homeostasis, female ISCs divide more often than male ISCs because of the cell-autonomous action of Sxl. The graph shows quantifications of the number of cells within each clone 15 days after clone induction by heat shock, and the confocal images show representative clones (labelled in green with GFP) for each genotype. f, Clonal analyses of homeostatic proliferation using the inducible esg flip-out system - which labels progenitors and their progeny generated within a defined temporal window⁸⁴ - in midguts of control males, control females and females in which Sxl downregulation has been confined to adult progenitors. 15 days after induction, the size (assessed as the percentage of GFP-positive area) of control female clones was significantly larger than that of male clones, but both became comparable upon adult-specific Sxl downregulation using Sxl RNAi transgenes. The graph shows area quantifications for each sex/genotype, and the confocal images show representative clones for each genotype. g, Immunohistochemical detection of histone H4 lysine 16 (H4Lys16) acetylation (in red) indicates that adult-specific downregulation of msl-2 in male intestinal progenitors (ISCs/EBs marker: GFP, in green) results in loss of H4Lys16 acetylation of the X chromosome. h, Efficient Msl-2 mis-expression in adult female intestinal progenitors (ISCs/EBs marker: GFP, in green) is confirmed by immunocytochemistry using an HA-specific antibody (in red). n denotes the number of guts (c, d, f), or clones (b, e) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.

Extended Data Fig. 4. Sexually dimorphic proliferation does not result from sex differences in differentiation.

a, Markers for all four intestinal cell types are still apparent following adult-specific downregulation of Sxl in the intestinal progenitors of females - achieved by Gal80^TS-controlled expression of a Sxl RNAi transgene. Indeed, expression of esg-Gal4 (ISC/EBs), Su(H)-LacZ (EBs), Pdm1 (ECs) and Pros (EECs) can be readily detected, suggesting that Sxl downregulation in females (which results in reduced ISC proliferation) does not have a major effect on differentiation. Sxl staining confirms efficient downregulation in ISCs/EBs, but not neighbouring cells. b, The same markers reveal that the differentiation defect resulting from N downregulation, previously reported in females, is also apparent in males (note loss of Su(H)-LacZ following N downregulation in both males and females), suggesting that sex differences in differentiation do not contribute to the sex differences in susceptibility to N-induced tumours. Co-expression of a mitogen (secreted Spitz, sSpi) abrogates the sex differences in tumour susceptibility by efficiently triggering hyperplasia also in males, as revealed by an expanded progenitor (GFP-positive) area in both males and females. The identity of these tumours in males is also comparable to that previously show for N tumours in females (consisting of high Pros-positive EEC-like cells and low Pros-positive neoplastic ISC-like cells⁸⁵. This further suggests that the sex differences in N-induced tumour susceptibility do not arise from sexually dimorphic differentiation effects, but result from sex differences in ISC proliferation. See Supplementary Information for full genotypes.

Extended Data Fig. 5. tra, but not tra2, controls intrinsic sex differences in adult ISC proliferation.

a, Additional controls for Fig. 2a, and confirmation of phenotypes using independent RNAi transgenes. Two additional RNAi transgenes against tra reduce the number of pH3-positive cells in DSS-treated female – but not male - midguts when expressed from esg^TS in adults ISCs/EBs, confirming an adult progenitor-specific requirement for tra in promoting damage-induced cell divisions in female flies. b, tra¹ MARCM mutant clones are smaller than control clones in females, whereas their size is comparable to that of wild-type or tra¹ clones in males. This confirms that, during normal homeostasis, female ISCs divide more often than male ISCs because of the cell-autonomous action of tra. The graph shows quantifications of clone size (in arbitrary units of GFP fluorescence, as described in Methods) 15 days after clone induction by heat shock, and the confocal images show representative clones (labelled in green with GFP) for each genotype. c, Ubiquitous, adult-restricted tra^F expression from tubP^TS in males increases the number of pH3-positive cells following DSS treatment to levels comparable to those of female flies. d, Re-introduction of this tra^F transgene specifically in adult ISCs/EBs rescues the reduced, male-like intestinal proliferation (as assessed by the number of pH3-positive cells) of tra null mutant females entirely lacking the tra gene from all their tissues (tra^KO/tra¹) to levels comparable to those of control females. Expression of this transgene in control heterozygous female flies (tra^KO/+ esg^TS>traF) does not significantly increase their proliferation (in fact, it reduces it slightly relative to tra^KO/+ esg^TS> controls, likely as a consequence of its over-expression). e, Clonal analyses of homeostatic proliferation using the inducible esg flip-out system - which labels progenitors and their progeny generated within a defined temporal window⁸⁴ - in midguts of control females and females in which tra downregulation has been confined to adult progenitors. 15 days after induction, the size (assessed as the percentage of GFP-positive area) of control clones is significantly larger than that of tra-RNAi clones. The graph shows area quantifications for each genotype, and the confocal images show representative clones for each genotype. f, Consistent with the tra mutant clonal analysis in Fig. 2c, quantifications of clone size (number of cells per clone) reveal that MARCM clones in which tra has been downregulated are smaller than control clones only in females. Their size is comparable to that of wild-type or tra-downregulated mutant clones in males. The confocal images show representative clones (labelled in green with GFP) for each genotype in females. g, qRT-PCR quantifications of relative abundance of tra^F, dsx^F, dsx^M and Yp1 transcripts in adult-specific tra2 mutants (tra2^B/ts1 grown at permissive temperatures, then switched to the restrictive temperature 4 days after eclosion and transcriptionally profiled following 10 additional days at the restrictive temperature) and controls (tra2^B/+). In tra2 mutant females, dsx^F is lost, dsx^M is upregulated to levels comparable to those of control males and Yp1 (a Dsx^F target) is lost (to levels also comparable to those of males). tra mutants (tra^{Df(3L)st-j7/KO}) were also used as a positive control. n denotes the number of guts (a, c, d, e), or clones (b, f) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.

Extended Data Fig. 6. dsx- and fru^M-independent control of sexually dimorphic proliferation in adult intestinal stem cells.

a, Adult-restricted downregulation of dsx (achieved by co-expression of a dsx-RNAi transgene and Dicer-2 (Dcr-2) in ISCs/EBs) has no effect on the compensatory ISC proliferation observed upon DSS treatment in neither males nor females. dsx^F expression in the same conditions does not increase ISC proliferation in either males or females. b, dsx^F expression does not rescue the reduced proliferation resulting from tra downregulation in females. Representative images for each genotype are shown in both a and b (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). c, The size of dsx null mutant (dsx¹) MARCM clones (quantified in arbitrary units of GFP fluorescence as described in Methods) is comparable to that of controls in both sexes 15 days after clone induction by heat shock. Confocal images show representative clones (labelled in green with GFP) for each genotype. d, Quantifications of clone size in control and dsx null mutant (dsx¹) MARCM clones in the midguts of DSS-treated males and females. 5 days after clone induction by heat shock, there are no significant differences in clone size (quantified in arbitrary units of GFP fluorescence as described in Methods) between control and mutants clones in either males or females. Confocal images show representative clones (labelled in green with GFP) for each genotype. e, An RNAi transgene against fru does not reduce the number of pH3-positive cells in DSS-treated midguts when expressed from esg^TS in the adult ISCs/EBs of either males or females. Confocal images show that number of intestinal progenitors (esg-positive cells in green) is also unaffected by this manipulation. f, Quantifications of the number of pH3-positive cells upon DSS treatment indicates that the sexual dimorphism in ISC proliferation is unaffected in females with forced fru^M expression (fru^M/fru^4-40) or in males with forced fru^F expression (fru^F/fru^4-40). g, ISC proliferation is unaffected in the migduts of DSS-treated males and females entirely lacking dsx (dsx^Δ/dsx¹), producing only Dsx^F (dsx^Δ/dsx¹¹) or Dsx^M (dsx^Δ/dsx^D). ISC proliferation is also unaffected in dsx, fru^M double null mutant males and females (dsx^Δ,Df(3R)^Exel6179/dsx¹, fru^P1.LexA), and in dsx null mutants in which fru^M is ectopically produced in females (dsx^Δ,Df(3R)^Exel6179/dsx¹, fru^Δtra). n denotes the number of guts (a, b, e, f, g), or clones (c, d) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.

Extended Data Fig. 7. tra targets in adult ISCs.

a, Scatter plot of all 1346 genes with tra-dependent expression in the adult fly midgut. For each gene, control female/tra null mutant female (Df(3L)^st-j7/tra^KO) fold differences in transcript abundance (X axis, log2 scale) are plotted against tra mutant female with feminized ISCs (adult-restricted rescue of tra^F in ISCs/EBs)/tra mutant female fold differences (Y axis, log2 scale). Genes with tra-sensitive expression and significantly repressed by tra^F in ISC/EBs (p<0.05 cutoff) are therefore found in the left-bottom quadrant and are displayed in blue, whereas those significantly activated by tra^F are found in the top-right quadrant and are displayed in red. Genes with tra-dependent transcription, independent of the action of tra^F in intestinal precursors are displayed in black. b, Comparable analysis of tra-dependent alternative splicing. c, tra expression in adult ISCs affects splicing of 38 transcripts by at least 5 different mechanisms. The outcome of each of the alternative splicing mechanisms is shown in yellow for a representative gene. d, Adult-restricted downregulation (RNAi lines) or mis-expression (UAS lines) of tra targets in adult ISCs/EBs by means of esg-Gal4, tubGal80^TS. Genes normally repressed in female progenitors in a tra^F-dependent manner were downregulated in males (top row) and/or misexpressed in females (bottom row). Genes upregulated in female progenitors in a tra^F-dependent manner were downregulated in females (bottom row). Adult-restricted downregulation of Idgf1 and Spn88Eb reduces the number of mitoses (pH3-positive cells) in DSS-treated females. Conversely, rdo mis-expression inhibits DSS-induced ISC proliferation in females. Adult-restricted downregulation of other tra^F targets in the same conditions does not affect ISC proliferation in either males or females. e, Male controls for Fig. 3c. In contrast to their effects on females, adult ISC/EB-restricted mis-expression of rdo or downregulation of Idgf1 and Spn88Eb does not reduce the percentage of midgut area covered by esg-positive cells in DSS-treated males (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). n denotes the number of guts (d) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.

Extended Data Fig. 8. Effects of the sexual identity of adult ISCs on midgut size and reproductive plasticity.

a, The number of cells in the R3a-b and R4a midgut regions, defined by expression of Cut and Mvl^NP2375 respectively (as described in ⁸⁶), is higher in females, and can be significantly reduced in females to numbers comparable to those found in males after 20 (but not 3) days of adult-specific downregulation of tra in intestinal progenitors (achieved by esg^TS-driven tra downregulation initiated after the phase of midgut post-eclosion growth, see Methods for details). No effects are apparent following downregulation in males. Representative images of these midgut regions (labelled in red with Cut or in green with Mvl^NP2375,esg^TS–driven GFP) are shown to the right for each genotype. b, Adult ISC/EB-specific tra downregulation does not affect the number of ISCs (esg-positive, Su(H)-negative cells) in either males or females, but strongly reduces EB (esg-positive, Su(H)-positive cells) production in females. c and d, Quantifications as in a and b for midguts with adult-specific downregulation of Sxl in intestinal progenitors. c, Reduced number of cells in the R4a midgut region (top graph) and total midgut length (bottom graph) in female flies following 20 days of adult-specific and cell-autonomous masculinization of their intestinal progenitors (achieved by-down regulation of Sxl over 20 days with esg-Gal4). The same manipulation has no discernible effects in males. d, The same genetic manipulation does not affect the number of ISCs (esg-positive, Su(H)-negative cells) in either males or females, but strongly reduces EB (esg-positive, Su(H)-positive cells) production in females. e, The number of EBs (esg-and Su(H)-positive cells, bottom graph), but not ISCs (esg-positive only cells, top graph) is higher in control female flies 3 days after mating. Adult ISC/EB-specific tra downregulation abrogates the postmating increase in EBs in females without affecting EB number in males, or ISC number in males or females. f, Adult-specific Sxl downregulation in intestinal progenitors leads to a small, but significant, reduction in egg production. An unrelated manipulation that also reduces ISC proliferation by inducing differentiation of ISCs (esg^TS>Notch^intra, images to the right of the graph and ⁸⁷) also results in reduced egg production, whereas downregulation of dsx (which does not control sex differences in progenitor proliferation) has no such effect. It should, however, be noted that esg-Gal4 is expressed in a subset of cells in the ovary niche²¹. Hence, the possibility that these cells contribute to the observed phenotype cannot entirely be ruled out. Images to the right show loss of intestinal progenitor cell makers esg-Gal4 and Su(H)-LacZ following expression of N^intra in adult intestinal progenitors, indicative of loss of progenitor identity. n denotes the number of guts (a, c), ISCs/EBs (b, d, e), or female flies (f) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.

Extended Data Fig. 9. Effects of the sexual identity of adult ISCs on the susceptibility to genetically induced tumours.

a, The number of mitoses in Apc-ras mutant midguts is larger than that of control clones in both sexes, but it is higher and dependent on Sxl in females. b, The size of Delta (Dl, the N ligand) null mutant (Dl^RevF10) MARCM clones is larger than that of control clones in both sexes, but female mutant clones are larger than male mutant clones. The graph shows quantifications of the number of cells within each clone 15 days after clone induction by heat shock, and the confocal images show representative clones (labelled in green with GFP) for each genotype. c, In tra “female” mutant flies (tra^KO and tra^KO/Df(3L)^st-j7) reduced N signalling in intestinal progenitors fails to induce the hyperplasia (quantified by the number of pH3 cells) normally observed in control females. d, Following 15 days of adult-specific downregulation of Notch (N) in intestinal progenitors, hyperplasia (quantified by the number of pH3-positive cells and also shown in representative images) is observed in female, but not male midguts. Adult-specific and cell-autonomous reversal of ISC/EB female identity – achieved by esg^TS-driven downregulation of Sxl – fully prevents the hyperplasia induced by Notch downregulation in females, but has no effects on males. Confocal images show intestinal progenitor coverage of representative midgut portions for each genotype (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). e, pH3 quantifications show a comparable effect for an independent RNAi transgene against Sxl. f, Adult-specific downregulation of Notch (N) signalling by ectopic expression of the downstream N signalling antagonist Hairless (H)³⁸ leads to hyperplasia (quantified by the number of pH3-positive cells and also shown in representative images) in female, but not male midguts. Adult-specific and cell-autonomous reversal of ISC/EB female identity – achieved by esg^TS-driven downregulation of Sxl – fully prevents the hyperplasia induced by H overexpression in females, but has no effects on males. Confocal images show intestinal progenitor coverage of representative midgut portions for each genotype (DNA: DAPI, in blue; ISC/EB marker: GFP, in green). g, The number of pH3-positive cells 15 days after N downregulation in adult intestinal progenitors of double null mutant flies lacking dsx and fru^M (dsx^Δ,Df(3R)^Exel6179/dsx¹, fru^P1.LexA) is comparable to that of controls in both males and female flies. Like control flies, it is significantly higher in female flies. Virgin flies were used in all these experiments. n denotes the number of guts (a, c, d, e, f, g), or clones (b) that were analysed for each genotype. Results combined from at least two independent experiments. See Supplementary Information for full genotypes.