Pointed is expressed in the somatic cells of the female larval gonad and ovarian GSC niche.

pnt spans ≈56 kb and encodes four different transcriptional variants. The transcription start sites (TSS) of pnt^long and pnt^short, are separated by ≈50 kb of mostly intronic sequence. The TSS of pnt^intermediate falls within this interval (Fig 1A). Considering the essential function of the EGFR pathway for proper female gonadogenesis and that pnt is a downstream target of the pathway [10,20], we studied a potential role for pointed in regulating GSC proliferation and/or niche formation throughout development. We used droplet digital PCR (ddPCR) to measure expression levels of the long, intermediate and short isoforms in embryos, female larval gonads and during early oogenesis in the adult. All three isoforms were found in embryos, consistent with the known roles of pnt in embryonic central nervous system development [8,21]. In contrast, female late third instar larval (LL3; 5 days after egg laying) gonads only expressed pnt short and long, while in ovary tips (germaria plus few, early egg chambers) we could only detect consistent levels of pnt long (Fig 1E).

To analyze further pointed expression in LL3 female gonads and throughout adult oogenesis, we utilized a number of lines expressing the Gal4 trans-activator under the control of different pnt enhancers and monitored their expression using a UASp-Tau::GFP reporter line to distinguish clearly cell shapes. In all ten lines analyzed, we detected pointed expression in most populations of somatic cells of the larval gonad including subsets of TFCs (often those cells in contact with the CpCs), CpCs and SwC. In the adult ovary, pnt-Gal4-driven Tau::GFP was consistently observed in TFCs, especially in TFCs near the CpC rosette, CpCs and in anterior ECs in the germarium (Figs 1F, 1G and S1). In addition, and as described before [10,22,23]Click or tap here to enter text., we could also detect expression of pnt in follicle cells at the posterior of the egg chamber from stage 6, as shown for the Gal4 line NP5370, inserted within the pnt locus (S1 Fig). Thus, pnt expression seems to be restricted to the somatic component of the female gonad and adult ovary.

pointed restricts PGC numbers in the female gonad and ensures cyst encapsulation in the germarium.

pnt null alleles are embryonic lethal [11]. Thus, studying pnt function during gonad formation requires either genetic tools to manipulate gene function in groups of cells or the use of hypomorphic mutations that give rise to viable adults. To evaluate the role of pnt in early oogenesis, first we used the UAS/Gal4 system to induce RNAi-mediated gene silencing. We examined the effect of lowering pnt expression in the larval gonad and/or the adult germarium by expressing different RNAi lines (HMS01452, KK100473 and JF02227; S2A Fig) under the control of c587-Gal4 (c587 > pnt^RNAi), which is expressed in the somatic cells of the gonad [24] and germarium [25] (S2B, S2E and S2F Fig). We observed a significant increase in PGC numbers for any of the RNAi lines compared to controls in female larval gonads (Fig 2A–C; similar results were obtained for the traffic jam-Gal4 line, also expressed in the somatic cells of the gonad; S2B and S2D Fig). Growing larvae at increasing temperatures (22, 25 and 29^oC) produced a gradation in the PGC overproliferation phenotype. Furthermore, scoring PGCs in mitosis with the phospho-histone H3 marker confirmed higher PGC proliferation in RNAi conditions (Fig 2C and 2D). These results support previous reports showing that decreasing EGFR signalling in the somatic cells of the female larval gonad gave rise to excess PGCs [20,26]. Our findings thus suggest that pnt acts downstream of the EGFR pathway in the intermingled cells of the female gonad to control proliferation and survival of PGCs. Moreover, the expression of the JF02227 RNAi construct with the c587-Gal4 driver in adult females gave rise to germaria larger than controls containing more germline cells, in agreement with a previous report [27]. In addition, we also observed egg-chamber fusions (Fig 2E and 2F), a novel phenotype suggesting that pnt is involved in somatic cell fate specification, proliferation and/or migration in the germarium. Since the RNAi lines used in these experiments are designed to target regions common to all the pnt isoforms and since the c587- and traffic jam-Gal4 lines are expressed in somatic cells (S2A, S2D and S2E Fig), we conclude from these results that reduction of pnt function in the somatic component of female gonads and ovaries controls germline proliferation and gametogenesis.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. pointed restricts PGC numbers in the gonad and ensures proper cyst encapsulation in the germarium.

(A, A’) Control LL3 gonad stained with anti-Vasa (green) to show PGCs and with a DNA dye (blue) to mark chromatin. (B, B’) LL3 gonad from a c587 > pnt RNAi^JF02227 female larva showing the excess of PGCs found in pnt loss-of-function gonads. (C) Quantification of the number of Vasa+ cells in control and c587 > pnt RNAi^HMS01452 at different temperatures in LL3 gonads. The arithmetic mean and the SEM are shown. (D) Quantification of the number of germ cells in mitosis using the PH3 marker. (E, E’) Ovariole from a control female stained to visualize F-actin (red), the germline (anti-Vasa; green) and with a DNA dye (blue). Developing egg chambers are formed in the germarium and each contains a 16-cell germline cyst surrounded by a monolayered follicular epithelium. The oocyte is always found at the posterior of the cyst. (F, F’) Ovariole from a c587 > pnt RNAi^JF02227 female stained as in (D). White arrowheads point to fused egg chambers containing more than one oocyte and 15 nurse cells. (E,” F”) High magnifications of control and c587 > pnt RNAi^JF02227 germaria. *** = p < 0.0005; **** = p < 0.0001. Scale bar: 20 μm. Related to S2 Fig.

https://doi.org/10.1371/journal.pgen.1011809.g002

pnt^aga, a novel mutation that induces germline cell death.

We have isolated and characterized a spontaneous mutation in the pnt locus that we named pnt agametic (pnt^aga). Contrary to the above results, adult pnt^aga homozygous males and females were viable but displayed a highly penetrant agametic phenotype that rendered them sterile (Figs 3, S3A and S3B). This phenotype was not a consequence of abnormal embryonic gonad development, since gonads from control and pnt^aga embryos were indistinguishable (S4 Fig). In contrast, the phenotypic characterization of ML3 gonads from female pnt^aga larvae (4–5 days after egg laying) already showed significant differences with controls. Thus, compared to controls (average 98.22 ± 5 n (number of gonads analyzed) = 9), the mutant PGC pool had an aberrant and disorganized morphology and the number of mutant PGCs at ML3 was considerably reduced (average 55.5 ± 4.35 n = 10; Fig 3A, 3B and 3I). To analyze the consequences for gonad maturation of the effects observed in larval tissues, we stained female control and pnt^aga white pupal gonads (∼24 hours later in development) to find that mutant gonads contained greatly reduced PGC numbers compared to controls (controls: average 206.87 ± 8.7 n = 8; pnt^aga: average 34 ± 3.8 n = 13). In addition, the few remaining mutant germ cells appeared aberrant in shape and larger in size (Fig 3C, 3D and 3I). Finally, homozygous mutant adult ovaries were considerably smaller than controls and were completely deprived of germ cells including GSCs, cystoblats and cysts. This phenotype was fully penetrant, as 100% of adult ovarioles were negative for the germline marker Vasa (Fig 3E-H and 3I). We also used an anti-Engrailed antibody to establish that mutant niches, while depleted of germline cells, seemed to form properly and to contain a normal pool of TFCs and CpCs (Fig 3G and 3H). This further confirmed that the reduced ovaries of agametic females maintain some characteristics of normal ovarian niches [28]

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. pnt^aga regulates germline survival.

(A) Control and (B) pnt^aga ML3 gonads stained with anti-Hts (red; to visualise cell outlines and germline spectrosomes and fusomes), anti-Vasa (green; to label the germline) and with a DNA dye (blue; to mark chromatin). (C) Control and (D) pnt^aga white pupa gonads stained as in (A). Note the progressive decline in germline cells in the mutant condition. (E, F) Darkfield images of control and pnt^aga ovaries. (G, H) Control and pnt^aga germaria stained with anti-Engrailed (red; to label TFCs and CpCs), anti-Vasa (green) and DE-cad (white). Note that pnt^aga ovaries and germaria are devoid of germline cells. (I) Quantification of the number of Vasa+ cells in control and pnt^aga embryos, ML3 gonads and white pupa gonads. Quantification of the percentage of ovarioles with germline in control and mutant adult ovaries. Numbers in columns denote sample size (n). The arithmetic mean and the SEM are shown for each of the genotypes. (J) Control and (K) pnt^aga LL3 gonads stained with anti-Caspase 3 (red; to visualise apoptotic cells), anti-Vasa (green) and with a DNA dye (blue). (L) Quantification of the number of Casp3 + cells in somatic or germline cells in control and pnt^aga gonads. *** = p < 0.0005; **** = p < 0.0001. Scale bars: 20 μm unless otherwise noted. Related to S3 and S4 Figs.

https://doi.org/10.1371/journal.pgen.1011809.g003

To test whether the agametic phenotype observed in mutant ovaries is due to PGC death concomitant with niche formation, we stained ML3 gonads with anti-Caspase 3 or DCP1, broadly used markers for the execution of apoptosis. We observed that, compared to controls, Casp3 and DCP1 were up-regulated in the middle region of the mutant gonad, particularly in PGC nuclei (Fig 3J, 3K, S3C and S3D). Altogether, our observations demonstrate that homozygous pnt^aga germ cells die by apoptosis during larval and pupal development, resulting in the final, highly penetrant agametic phenotype. To rule out the possibility that apoptosis is also upregulated in other somatic cells, including the ICs and that PGCs die as a consequence of this, we studied in detail higher magnification views of mutant gonads and found that Caspase3 signal is generally restricted to PGCs (controls: 2.33 ± 0.56 (n = 6) apoptotic PGCs and 6.7 ± 1.17 (n = 6) apoptotic somatic cells; pnt^aga: 15.13 ± 1.76 (n = 8) apoptotic PGCs and 5.75 ± 1.5 (n = 8) apoptotic somatic cells; Fig 3L). In addition, the use of the IC-specific marker Traffic-jam did not coincide with DCP1-positive cells. From the above results, we conclude that the initial steps of ovary morphogenesis take place correctly and that, as development progresses, pnt^aga germ cells are lost from the mutant gonads. Hence, pnt^aga uncovers a specific function for pointed in gonad development.

pnt^aga affects pointed transcription

Using deficiency mapping, genetic tests and genomic sequencing, we mapped the molecular lesion responsible for the pnt^aga phenotype to a two-nucleotide change (AC to CT; FlyBase coordinates: 23,295,576–7) in an intronic region upstream of the TSS of pnt^short (Figs 4A, S5 and S6). Knowing that the molecular lesion of pnt^aga does not alter the coding sequence of any of the three Pnt proteins, it is likely that pnt^aga causes PGC degeneration by altering transcript levels of the pnt locus in the somatic cells of the female gonad. To test this, we performed ddPCR to assess the relative mRNA levels of the different pnt isoforms in mutant LL3 larval gonads, and compared them to controls. We observed a significant reduction in the expression levels of both the long and short isoforms in pnt^aga (controls: long isoform 22.29 ± 10.8 copies/ng of cDNA; short 11.46 ± 7.5. pnt^aga: long 8.70 ± 0.6; short 3.60 ± 2.1. Two biological and two technical replicates per genotype; Fig 4B), demonstrating that pnt^aga represents a partial loss-of-function mutation. In fact, utilizing the expression of the diphosphorylated form of the extracellular signal-regulated kinase (dp-ERK) as a read-out of the EGFR pathway, we could confirm that the pnt^aga condition showed reduced dp-ERK levels (Fig 4C). Thus, pnt^aga impacts pnt transcription, leading to altered transcript levels of the different isoforms. Similarly, pnt^aga mutant embryos showed a decrease in long, intermediate and short isoforms (controls: long isoform 27.72 ± 2.7 copies/ng of cDNA; intermediate 13.89 ± 1.4; short 28.17 ± 8.5. pnt^aga: long 13.84 ± 3.7; intermediate 5.71 ± 0.8; short 13.47 ± 3.55. Two biological and two technical replicates per genotype; S4C Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. pnt^aga affects pointed locus transcription.

(A) Mapping of the pnt^aga mutation. Numbers refer to genomic coordinates according to FlyBase (release FB2024_02). (B) Quantification of the long, intermediate and short pnt mRNA levels in control and pnt^aga gonads using droplet-digital PCR. Measurements correspond to three biological replicates and to one or two technical replicates. (C) Quantification of dp-ERK fluorescence intensity in control and c587 > pnt RNAi^HMS01452 IC. (D) Control, (E) pnt^aga and (F) c587 > pnt^intermediate; pnt^aga ovarioles stained to visualise F-actin (red), the germline (anti-Vasa; green) and with a DNA dye (blue). (G) Quantification of germline-containing ovaries in different genotypes. Ectopic expression in the soma of any of the pnt isoforms rescues the agametic phenotype. (H) Quantification of the long, intermediate and short pnt mRNA levels in control and pnt^aga/Df, UASt-pnt^short pnt^aga/Df, control and c587 > pnt^intermediate; pnt^aga/Df gonads using droplet-digital PCR. Measurements correspond to two or three biological replicates and to one or two technical replicates. (I) Quantification of germline-containing ovaries in pnt^aga and in c587 > pnt RNAi^KK100473; pnt^aga females. Reduction of pnt function in a pnt^aga background rescues partially the agametic phenotype. Scale bars: 20 or 50 μm, as indicated. Related to S4 and S5 Figs.

https://doi.org/10.1371/journal.pgen.1011809.g004

To determine whether an increase in pnt levels could rescue the pnt^aga phenotype, we generated flies containing UAS-pnt constructs in a pnt^aga background. We tested the three isoforms, long and short (UASt-pnt^long, UASt-pnt^short; [21]) and intermediate (UASp-pnt^intermediate; this work). In all instances, flies carrying any of the constructs were viable and fertile. Interestingly, pnt^aga females carrying two copies of any of the UAS-pnt constructs showed a significant proportion of ovaries containing germline (Fig 4D–G). Since UAS constructs can display some basal activity in absence of a Gal4 driver [29], these phenotypic rescues confirmed the loss-of-function nature of the pnt^aga allele. Moreover, molecular analyses of pnt isoform levels upon combination with rescue constructs demonstrated a positive feed-back loop regulating pnt^long transcription in experimental conditions over-expressing pnt^short or pnt^intermediate (Fig 4H). Next, we determined that pnt was required in somatic cells of the gonad, as the somatic c587-Gal4 driver, but not the germline-specific nanos-Gal4 (S2G and S2H Fig), was able to rescue the agametic phenotype when overexpressing UASp-pnt^intermediate (Fig 4G).

As stated above, the pnt^aga phenotype is opposite to that of pnt^RNAi, the former giving rise to gonads and germaria devoid of germ cells and the latter causing excess germ cell proliferation. Considering that pnt^aga LL3 gonads express reduced levels of long and short pnt and that pnt^RNAi knockdown reduces transcript levels of all isoforms (S2A Fig), it is possible that germline progression depends on the levels and/or combination of the different pnt isoforms expressed. In our working model, a strong reduction in pnt levels in somatic cells of the gonad (c587 > pnt^RNAi) impairs EGFR signaling and thus results in germ cell overproliferation, while a novel combination of long and short pnt induces germline disappearance by cell death (pnt^aga). If this hypothesis were right, the pnt^aga phenotype should be suppressed by the removal of all pnt function. To test this, we generated flies that carried the pnt^RNAi construct in a pnt^aga background (c587 > pnt^RNAi + pnt^aga) and scored the germline phenotype in adult females. We found that the expression of pnt^RNAi rescued partially the agametic phenotype typical of pnt^aga, as experimental c587 > pnt^RNAi + pnt^aga ovaries contained developing egg chambers in 38% of the cases (n = 34), whereas control pnt^aga showed complete absence of germline (n = 79; Fig 4I). From these results we interpret that germline development relies on the presence of the proper levels and combination of Pnt isoforms — mainly the long and short ones — in the somatic cells of the larval gonad. Thus, the RNAi-induced decrease in pnt-mediated signaling causes excess germ cell proliferation, most likely since this would mimic the abrogation of EGFR pathway activity in intermingled cells in the gonad [20]. Conversely, a new isoform combination of long and short pnt, as seen in the pnt^aga condition, induces the opposite effect, germline apoptosis.

A transcriptomic analysis of pnt^aga gonads identifies cell-type specific changes in cell adhesion, ion binding and membrane activity

Considering the effect of pnt^aga on the various pointed isoforms, we aimed to determine how the altered levels of pointed transcripts influenced the global transcriptomics of mutant gonads. Using Affymetrix microarrays (GeneChip Drosophila Genome 2.0; > 13,500 genes arrayed), we performed a differential gene expression (DGE) analysis of control vs mutant LL3 gonads. We identified 253 genes whose expression was changed in pnt^aga gonads, 118 were upregulated and 135 down-regulated (Fig 5A and S1 Data; fold change<0.5 or >2; false discovery rate<5%). Gene Ontology (GO) enrichment analyses of the differentially expressed genes yielded a number of terms associated to cell adhesion or ion binding (up-regulated genes) or to membrane activity, including membrane transport, and response to mechanical stimulus (down-regulated genes), among others (Fig 5B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Differential GE in pnt^aga gonads.

(A) Changes in gene expression (GE) — represented as log2 fold change of pnt^aga/control versus the -log10 false discovery rate (FDR) — identified 118 genes upregulated and 135 genes downregulated in pnt^aga compared to controls (linear fold change <0.5 or >2; FDR < 5%). (B) GO terms identified (p-value<0.01) in the analysis of either up- or down-regulated genes. Numbers in parenthesis correspond to the genes identified in each term. (C) Diagram to indicate the number of genes showing differential GE in pnt^aga gonads common to the list of pnt targets in the eye disc [31] and the genes bound by Pnt in embryos [13]. (D) Diagram representing the number of differentially expressed genes in pnt^aga gonads and in pnt^null embryos, and of direct targets identified in the eye disc. Data related to pnt^aga correspond to three biological replicates per genotype. Related to S1 Data.

https://doi.org/10.1371/journal.pgen.1011809.g005

The ETS family of transcriptions factors, comprising Pnt, are versatile in their functions and interact with other transcription factors and cofactors to allow combinatorial control of gene expression. As previously stated, the output of some Ras/MAPK signalling pathways depends upon the biochemical equilibrium between the partners Pnt and Yan, switching from an activation state of high Pnt to a contrasting inactivation state of high Yan. It is important to acknowledge that this bistable model depends on the specific cellular milieu and the intricacies of transcriptional regulation. Hence, a significant proportion of pnt targets most likely are tissue specific [30,31]. With this in mind, we compared the list of up- and down-regulated genes in pnt^aga gonads with known pnt targets in the eye disc (157 genes) and with pnt-bound genes (2584) in the embryo [13,31]. We found 8 genes in common with the eye disc targets and 44 also bound by Pnt in the embryo, with only 3 genes common to the three datasets (Fig 5C and S1 Data). While the identification of Pnt direct targets in the gonad requires further experimentation, from our results we conclude that the outcome of diminishing pnt transcript levels seems to be largely cell-type specific. Next, considering that the lower levels of long and short pnt observed in pnt^aga gonads indicated a partial loss of function situation, we compared the differentially expressed genes in pnt^aga with those identified in pnt null mutant embryos [13] and identified 7 common genes out of the ∼9,000 genes spotted on the array [13]. Interestingly, 3 of the latter (hibris, CG1773 and 3-Hydroxymethyl-3-methylglutaryl-CoA lyase) were up-regulated in pnt^aga, while they were down-regulated in pnt null embryos, emphasizing the idea that pnt-mediated regulation of gene expression is context-specific. This view is further supported by the limited number (15) of direct pnt targets in the eye disc that were also differentially expressed in pnt null embryos and by the fact that only 1 gene was common to the three datasets (Fig 5D and S1 Data). Finally, we performed two statistical tests to calculate the probability of the overlap by chance between the different sets of genes, an analytical test (hypergeometric distribution) and an empirical test (permutation test with 10,000 permutations). These analyses dictated that the only significant overlap (p-value<0.05) in all of the pairwise comparisons was between the identified Pnt targets in the eye disc and the Pnt-bound genes in the embryo (p-value (analytical)=9.99x10^-5; (empirical)=4.83x10^-12) and between the Pnt eye targets and the differentially expressed genes in pnt^null embryos (p-value (analytical)=1.16x10^-4; (empirical)=3.00x10^-4). These results provide strong support for the context-dependent outcome of pnt transcriptional regulation.

Fly stocks and RNAi protocol

We used FlyBase (release FB2024_02) to find information on genes and their functions [41]. Flies were grown at 25°C on standard medium unless otherwise stated. The lines used include:

TM3, twist-Gal4, UAS-nlsGFP [42]

UASp-tau::mGFP6 [43]

pnt^aga (this work)

pnt^Δ88 [11] (Bloomington Drosophila Stock Centre (BDSC) #861)

P{w[+mC]=lacW} pnt¹²⁷⁷ [11](BDSC #837)

P{w[+mW.hs]=GawB} pnt^NP5370 (Kyoto Stock Centre #104978)

tj-Gal4 [44]

c587-Gal4 [45]

nos-Gal4 [46] (BDSC #4442)

UASp-pnt^intermediate (this work)

UASt-pnt.P1 (pnt^short) [21] (BDSC #869)

UASt-pnt.P2 (pnt^long) [21] (BDSC #399)

FRT82B ubi-nls::GFP (BDSC #32655)

Df(3R)Exel6280, P{w[+mC]=XP-U}Exel6280 (BDSC #7746)

Df(3R)Exel9012, P{w[+mC]=XP-U}Exel9012 (BDSC #7990)

Df(3R)ED6105, P{w[+mW.Scer\FRT.hs3]=3’.RS5 + 3.3’}ED6105 (Kyoto Stock Center #150336)

P{TRiP.JF02227}attP2 (UAS-pnt RNAi^JF02227) (BDSC #31936)

UAS-pnt RNAi^KK100473 (Vienna Drosophila Resource Center #105390)

y v;; P{TRiP.HMS01452}attP2 (UAS-pnt RNAi ^HMS01452) (BDSC #35038)

The following enhancer-Gal4 constructs from the Janelia Farm collection were used (BDSC stock numbers in parenthesis): P{GMR44C09-GAL4}attP2 (#41263), P{GMR43E07-GAL4}attP2 (#45304), P{GMR44B07-GAL4}attP2 (#45717), P{GMR46C10-GAL4}attP2 (#46271), P{GMR44C01-GAL4}attP2 (#48145), P{GMR45D11-GAL4}attP2 (#49563), P{GMR45F08-GAL4}attP2 (#49565), P{GMR45B10-GAL4}attP2 (#50223), P{GMR45E10-GAL4}attP2 (#50233).

Generation of the UASp-pnt^intermediate construct

To express pnt^intermediate in the germline, we first synthesized the pnt-RD cDNA from 2-day old control (y w) ovaries. mRNA was isolated and purified using the QuickPrep Micro mRNA Purification Kit (GE healthcare) from approx. 100 ovary pairs. 1–2 μg of mRNA were used to synthesized cDNA with 0.5 μg of oligo(dT) (Sigma Genosys) and the Superscript II RNase H Transcriptase (Invitrogen LifeTechnology) in a 20 μl final volume. 2 μl of total cDNA were used in a PCR reaction to amplify the pnt-RD cDNA.

The following primers were used:

Forward primer

5’ GCTAGGATCCATGACCAATGAGTGGATCGAT 3’

Reverse primer

5’ GAATGCGGCCGCTTTGCGGTTGGTTGAGTACA 3’.

The pnt-RD cDNA generated by reverse transcription was then cloned as a BamHI/Not I fragment into pUASp [47]. The resulting construct was verified by restriction digests and sequencing. Transgenic lines were generated by standard procedures.

Mapping of genomic deficiencies

Genomic DNA from single flies was prepared following the protocol by Georg Dietzl (Barry Dickson’s Laboratory, IMP, Vienna). In short, one fly of the desired genotype was place in a 0.5 ml tube containing 5 µl of squishing buffer and mashed for 5–10 seconds with a pipette tip. After the addition of another 45 µl of squishing buffer, the mix was incubated at room temperature for 30 minutes. The Proteinase K was inactivated by heating the mix at 95^oC for 1–2 minutes. Typically, 2 µl of supernatant were used per 20–50 µl PCR reaction.

Squishing buffer: 10 mM Tris-Cl pH 8.2, 1mM EDTA, 25 mM NaCl, 200 µg/ml Proteinase K (enzyme diluted fresh from a frozen stock).

The Exel6280 and Exel9012 deficiencies were generated by FLP/FRT deletions [48] and their breakpoints were originally mapped with reasonable precision and reported in FlyBase (flybase.org). They both harbour hybrid elements from the XP5’ (plus and minus) and the WH5’ P-elements, which we used to map the breakpoints in the pnt locus accurately. Internal primers corresponding to the XP elements were paired with genomic primers designed to bind to flanking sequences according to the published breakpoints. The PCR reaction would only amplify a DNA product from the chromosome bearing the deficiency, as internal primers are only complementary to the P-element sequences. To sequence the 3’ genomic flanking region of Df(3R)ED6105 we used a similar strategy, as this deficiency was generated by FLP/FRT-mediated recombination of two RS elements [49] In all cases, genomic DNA was prepared from Deficiency/Balancer flies, diluted 1:25 and used as template in PCR reactions. The primers used were the following (S6 Fig):

Df(3R)Exel6280, distal breakpoint:

XP3’ + Flanking genomic region primers:

XP1 forward 5’GCCACCACTTCAAGAACTCTGTAGC 3’

G1 reverse 5’CGAGTTGGCGTTGTTAATGA 3’

Df(3R) Exel9012, distal breakpoint:

XP + Flaking genomic region:

52B forward 5’ TTTACTCCAGTCACAGCTTTG 3’

G5 reverse 5’ GCACAGTGGTTGTCATGGTC 3’

Df(3R) ED6105, distal breakpoint:

Pry4 (maps to the RS3r 3’ terminal repeat end) forward 5’ CAATCATATCGCTGTCTCACTCA 3’

Genomic reverse 5’ GTTTGTGAGAGGCGGCAGCCA 3’

Mapping of pnt^aga

Genomic DNA from control y w and pnt^aga homozygous flies was prepared as described, diluted 1:25 and used as template in PCR reactions. Genomic primers were designed based on the reference sequence in flybase.org. The following primers were used (S6 Fig):

Forward 5’ AGAACGCGTTGTATGAGGGACCCA 3’

Reverse 5’ GTTTGTGAGAGGCGGCAGCCA 3’

Sample preparation, microarray hybridization and data analysis

RNA from 2 larval gonads (approximately 2000–3000 cells) per sample (Oregon-R controls and pnt^aga homozygous; three biological replicas per genotype) was isolated using magnetic beads. cDNA synthesis, library preparation and amplification (Pico Profiling) are described elsewhere [50]. In short, after reverse-transcription each cDNA sample was added to an amplification mix that was subdivided into five equivalent parts for PCR amplification (26 cycles). Amplified cDNA collections were purified with the PureLink PCR Purification Kit (ThermoFisher Scientific), resuspended in 40μl and their concentrations measured using a Nanodrop 1000 spectrophotometer. The six cDNA collections were hybridized to GeneChip Drosophila Genome 2.0 Arrays from Affymetrix. Raw data generated with Affymetrix’s AGCC software were used as input for the DTT software. For the resulting datasets, RMA (Robust Microarray Average) normalization and linear modelling by limma (Linear Models for Microarray Analysis) were performed. The p-value for the False Discovery Rate (FDR) was ≤ 0.05 and fold change was set at <0.5 or >2. To test the overlap of different gene sets in Fig 5, we utilized the hypergeometric probability distribution and a permutation test with 10,000 permutations. We assumed a number of detectable D. melanogaster genes of 9,000, presuming that all of the ∼9,000 genes present in the embryo microarray were also in the gonad microarray and that all were putative targets for the ChIP-seq analyses.

Homozygous pnt^aga larvae were selected based on their loss of the Tubby marker on the TM6B balancer.

Detection and quantification of mRNA levels by ddPCR

We used droplet digital PCR (ddPCR) to quantify mRNA levels of pnt long, intermediate and short isoforms from embryos, LL3 gonads and ovary tips (obtained after severing and collecting the germaria and few egg chambers from the whole ovaries). For embryos, RNA was isolated from ~30 embryos per genotype (y w controls and pnt^aga homozygous; three biological replicas each; homozygous pnt^aga embryos were selected based on their loss of the TM3, twist-Gal4, UAS-GFP chromosome) with RNAeasy Micro (Qiagen 74004) and QIAshredder (Qiagen 79654) columns. To synthesize cDNA, 10 μl of mRNA were incubated for 5 minutes at 65°C with 1 μl (0.5 μg) of Anchored Oligo(dT)₂₃ (Sigma O4387) + 1 μl of 10mM dNTP mix and then placed on ice for 1 minute. 4 μl of 5x First Strand buffer (Y02321) + 1 μl RNaseOUT Recombinant Ribonuclease Inhibitor (40 units/ μl) (100000840) + 2 μl 0.1 M DTT (Y00147) from Invitrogen were added and incubated for 1 minute at 42°C. Next, 1 μl of Super Script II RT from Invitrogen (18064–014) was added and incubated for 50 minutes at 42°C, 15 minutes at 70°C and then placed on ice for 1 minute. Finally, the mix was incubated for 20 minutes at 37°C with 1 μl of RNase H from Invitrogen (18021–014). For LL3 gonads, see above. Finally, to determine pnt expression in ovary tips, we dissected 30 adult ovaries per genotype and manually severed off the tips (which included the germarium and few young egg chambers per ovariole) using tungsten needles. RNA was isolated as in the case of embryos.

Primers used were the following (5’-3’): pnt-RB (long) isoform, forward (exon 2): TTCTGTCCAGCCTAGTTGAG, reverse (exon 5): AGTACCTCGTTGACCTTGCG; pnt-RD (intermediate) isoform, forward (exon 4): AACCACCAATCATAGCAAGC, reverse (exon 5): AGTACCTCGTTGACCTTGCG; and pnt-RC/E (short) isoform, forward (exon 8): TGCGAATGCCTACTACACGG, reverse (exon 9): TACCGCTGCCATTGACAGTC. Ribosomal protein L32 (RpL32) and β-Tubulin (β-Tub) were used for normalization. RpL32, forward: ATGACCATCCGCCCAGCATAC, reverse: GCTTAGCATATCGATCCGACTGG; β-Tub, forward: GCAGTTCACCGCTATGTTCA, reverse: CGGACACCAGATCGTTCAT.

For the ddPCR, each reaction contained 10 μl of Master Mix ddPCR EVAGreen (Bio-Rad), 150 nM of each primer and 2,5 ng of cDNA template (except in the case of the housekeeping control, where we used 0,5 ng). Samples were prepared in duplicate with 10% additional volume. Droplet generation, PCR amplification and droplet analysis were done using the QX200 AutoDG ddPCR system (Bio-Rad). PCR conditions were: 95°C for 5 minutes (1x), 95°C for 1 minute and 65.2°C for 2 minutes (44x). For all steps a ramp rate of 2°C/s was used. Data were analysed with the Quantasoft software 1.7.4.0917 (Bio-Rad).

Developmental staging of larvae and pupae

At 25^oC, the end of embryogenesis is 24 hours after egg laying (AEL). The first and second larval instars are 24 hours each. Third larval instar lasts for 48 hours and puparium formation starts 120 hours AEL. In this study, ML3 (mid-third instar larvae) refers to larvae that crawled out of the food but that have not initiated pupariation. At this stage, most of TFs are still forming and cap cells are not ready visible. LL3 (late-third instar larvae) is used to identify larvae about to initiate pupariation. Their gonads present a number of already-formed TFs and some cap cells can be recognised. The white pupa stage marks the initiation of pupariation and it is characterised by pale and clear pupae. At this stage, the GSC niches with their respective TFs and cap cell rosettes are well established [19].

Immunohistochemistry

Adult flies were yeasted for 2 days before dissection in cold Ringer’s. Ovaries and testes were then fixed for 10 minutes in 6% formaldehyde in fix buffer (600μl of heptane were added per 100μl of fix). After fixing, samples were permeabilized in 1% Triton in PBS for 2 hours and then blocked for 1 hour with PAT before overnight incubation with primary antibodies in PAT. The following morning primary antibodies were rinsed with PAT and the samples washed three times with PBT for 30 minutes. The incubation with secondary antibodies was done in PBT during 2–4 hours. Ovaries and testes were then washed three times with PTW, 10 minutes each. Mounting of testes and individual ovarioles was done in Vectashield medium (Vector Laboratories; Cat# H1000; RRID: AB_2336789).

To stain larval gonads, gonads embedded in the fat body were fixed in 5% formaldehyde in Ringer’s for 20 minutes then washed for 5, 10 and 45 minutes with 1% PBT. The fat body was then blocked in 0.3% PBTB for 1 hour with gentle agitation. After blocking, the tissue was incubated overnight with the desired primary antibody diluted in 0.3% PBTB at 4^oC with gentle agitation. The next day the fat body was washed thrice (30 minutes each) with 0.3% PBTB. After blocking with 0.3% PBTB supplemented with 5% fetal bovine serum for 1 hour, the fat body was incubated for 2 hours with the secondary antibodies in blocking solution, followed by three washes of 30 min each with 0.3% PBT. Gonads were dissected from the fat body and mounted in Vectashield.

Embryos collected overnight on yeasted agar plates were dechorionated with bleach for 2 minutes and fixed for 20 minutes in 2 ml of 4% formaldehyde in PBS with an equal volume of heptane. After fixing, the aqueous phase was removed and 2 ml of methanol were added to remove the vitelline membrane. Devitellinized embryos were rinsed in methanol thrice and stored at -20^oC. For antibody staining, embryos were re-hydrated in PBS:methanol for 5 minutes and then washed in PBS, in PBS:PAT and finally rinsed in PAT twice. Next, they were blocked for 1–2 hours in PAT at 4^oC. Primary antibody incubation was done overnight at 4^oC in PAT, followed by a blocking step for several hours in PBT supplemented with 4% goat serum prior to secondary antibody incubation for 2–4 hours in PBT. Finally, embryos were washed thrice with PBT, 10 minutes each, and mounted in Vectashield.

Primary antibodies were used at the following concentrations: mouse anti-Hts (1B1, Developmental Studies Hybridoma Bank (DSHB) Cat# 1b1, RRID:AB_528070), 1:100; rabbit anti-Vasa 1:2000 (a gift from R. Lehmann); rat anti-DE-Cadherin (DSHB Cat# DCAD2, RRID:AB_528120) 1:100; mouse monoclonal anti-Engrailed (DSHB Cat# 4D9; RRID: AB_528224) 1:10; rat anti-Bab2 (a gift from F. Laski) 1:4000; rabbit anti-cleaved-Caspase3 (from Cell Signaling Technology) 1:50; goat anti-GFP FITC-conjugated (Abcam Cat# ab6662, RRID:AB_305635) 1:500; rabbit anti-dp-ERK (Cell Signaling) 1:20; Rabbit anti-Caspase DCP1 (Cell Signaling) 1:100. Secondary antibodies Alexa-Fluor 488, Cy2, Cy3 and Cy5 (Jackson Immuno Research Laboratories, Inc.; final concentrations of 1:100) and conjugated anti-GFP-488 nanobody (alpaca anti-GFP-Booster_Atto488 from ChromoTek, Cat# gba488–100, RRID:AB_2631386; final concentration 1:200) were incubated for four hours. To stain DNA, ovaries were incubated for 10 minutes with Hoechst (Sigma B2883 10 mg/ml in H₂O). To visualise F-actin, fixed ovaries were incubated in PBT + Rhodamine-labelled Phalloidin (Biotium Cat# BT-00027; final concentration 1:200) for 20 minutes.

Ringer’s: 128 mM NaCl, 2 mM KCl, 1.8mM CaCl2, 4 mM MgCl2, 35.5 mM Sucrose, 5 mM Hepes pH 6.9

Fix buffer: 16.7 mM KPO4 pH 6.8, 75 mM KCl, 25 mM NaCl and 3.3 mM MgCl2

PAT: PBS, 1% BSA, 0.1% Triton, 0.05% azide

PBT: PBS, 0.1 BSA, 0.1% tween20

PTW: PBS, 0.1% tween20

1% PBT: 1% Triton x-100 in PBS

0.3% PBTB: 0.3% Triton X-100 and 1% BSA in PBS

Imaging, processing and quantification of larval, pupal and adult samples

Images were acquired with Leica’s TCS-SP5 or Stellaris confocal microscopes, analysed utilizing ImageJ, and processed with Adobe Photoshop and Adobe Illustrator. Z stacks of fixed samples were taken at 0.7 μm intervals using 40x/1.3 and 63x/1.4 NA oil immersion objectives.

Quantification of the number of Vasa+ cells was done as described in [51]. Quantification of the number of Casp3 + cells was done manually using the multi-point tool in ImageJ.

Statistical analysis

Experiments involving gonads or ovaries were performed with at least three biological replicates. Samples were collected from at least 5 different larvae, white pupae or adult females grown in equivalent environmental conditions. For each of the quantifications, the arithmetic mean and the standard error of the mean (SEM) of the different experimental settings are shown. Sample sizes correspond to the number of gonads or adult ovaries analysed. Statistically significant differences between control and experimental samples were calculated with a Student’s t-test. * = p < 0.05; ** = p < 0.005; *** = p < 0.0005; **** = p < 0.0001. Only significant differences are indicated in the graphs. To test the overlap of different gene sets in Fig 5, we utilized the hypergeometric probability distribution and a permutation test with 10,000 permutations. We assumed a number of detectable D. melanogaster genes of 9,000, presuming that all of the 9,000 genes present in the embryo microarray are also in the gonad microarray and that all are putative targets for the ChIP-seq analyses.

The quantifications of mRNA concentrations using digital PCR were performed with a minimum of two biological replicates.

Experimental genotypes

Fig 1.

• (F, G) pnt enhancer 45E10>tau::GFP: y w/w; P{GMR45E10-GAL4}attP2/ + ; UASp-tau::mGFP6/+

Fig 2.

• (A, E) control: w, c587-Gal4/y w;; UASp-tau::mGFP6/+
• (B, F) c587 > pnt RNAi: w, c587-Gal4/y v;; P{TRiP.JF02227}attP2/+
• (C, D) control: c587-Gal4 > w RNAi: w, c587-Gal4/y v;; P{TRiP.JF02387}attP2/+
• c587 > pnt RNAi (HMS01452): w, c587-Gal4/y v;; P{TRiP.HMS01452}attP2/+

Fig 3.

• (A-L) control: w;; FRT-82B cu sr pnt^aga/TM3, twist-Gal4, UAS-nlsGFP
• pnt^aga: w;; FRT-82B cu sr pnt^aga

Fig 4.

• (B, D) control: w;; FRT-82B cu sr pnt^aga/TM3, twist-Gal4, UAS-nlsGFP
• pnt^aga: w;; FRT-82B cu sr pnt^aga
• (C) control: c587-Gal4 > w RNAi: w, c587-Gal4/y v;; P{TRiP.JF02387}attP2/+
• c587 > pnt RNAi (HMS01452): w, c587-Gal4/y v;; P{TRiP.HMS01452}attP2/+
• pnt^aga/Df: w;; FRT-82B cu sr pnt^aga/ Df(3R)ED6105
• (E, F, G) UASp-pnt^intermediate; pnt^aga: UASp-pnt^intermediate/ + ; FRT-82B cu sr pnt^aga
• nos > pnt^intermediate; pnt^aga: y w/w; nos-Gal4/UASp-pnt^intermediate/ + ; FRT-82B cu sr pnt^aga
• c587 > pnt^intermediate; pnt^aga: w, c587-Gal4/w; UASp-pnt^intermediate; FRT-82B cu sr pnt^aga
• UASt-pnt^short pnt^aga: UASt-pnt^shortcu sr pnt^aga
• UASt-pnt^long pnt^aga: UASt-pnt^longcu sr pnt^aga
• (H) control: w, c587-Gal4/+; CyO:GFP/+
• pnt^aga/Df: w;; FRT-82B cu sr pnt^aga/ Df(3R)ED6105
• UASt-pnt^short pnt^aga/Df: UASt-pnt^short, cu sr pnt^aga/ Df(3R)ED6105
• c587 > pnt^intermediate; pnt^aga/ Df(3R)ED6105: w, c587-Gal4/ + ; UASp-pnt^intermediate/ + ; FRT-82B cu sr pnt^aga/ Df(3R)ED6105
• (I) c587 > pnt RNAi; pnt^aga: w, c587-Gal4/w; UAS-pnt RNAi^KK100473/ + ; FRT-82B cu sr pnt^aga

Fig 5.

• control: Oregon-R
• pnt^aga: w;; FRT-82B cu sr pnt^aga

S1 Fig.

• (A–C) NP5370 > tau::GFP: y w;; P{w[+mW.hs]=GawB} pnt^NP5370/UASp-tau::mGFP6.
• (D–L) pnt enhancer>tau::GFP: y w/w; P{GMR enhancer-GAL4}attP2/ + ; UASp-tau::mGFP6/+
• The nine enhancers used are listed in the Fly stocks section above.

S2 Fig.

• (B,C) control: w; + /CyO.
• c587 > pnt RNAi (JF02227): w/y v; tj-Gal4/ + ; P{TRiP.JF02227}attP2/ + w, c587-Gal4/y v;; P{TRiP. JF02227}attP2/+
• c587 > pnt RNAi (KK100473): w; UAS-pnt RNAi^KK100473/tj-Gal4
• tj > pnt RNAi (JF02227): w/y v; tj-Gal4/ + ; P{TRiP.JF02227}attP2/+
• tj > pnt RNAi (KK100473): w; UAS-pnt RNAi^KK100473/tj-Gal4
• (D) tj > tau::GFP: w/y w; tj-Gal4/ + ; UASp-tau::mGFP6/+
• (E, F) c587 > tau::GFP: w, c587-Gal4/y w;; UASp-tau::mGFP6/+
• (G, H) nos > tau::GFP: w/y w; nos-Gal4/ + ; UASp-tau::mGFP6/+

S3 and S4 Figs.

• (A–D) control: y w
• pnt^aga: w;; FRT-82B cu sr pnt^aga

S5 Fig.

• (B, C) control: w.
• pnt^aga: w;; FRT-82B cu sr pnt^aga
• (D) pnt^Δ88/Df(3R)6280: y w f/w;; FRT-82B pnt^Δ88/Df(3R)Exel6280, P{w[+mC]=XP-U}Exel6280
• pnt^Δ88/Df(3R)9012: y w f/w;; FRT-82B pnt^·88/Df(3R)Exel9012, P{w[+mC]=XP-U}Exel9012
• pnt^Δ88/Df(3R)ED6105: y w f/w;; FRT-82B pnt^Δ88/Df(3R)ED6105, P{w[+mW.Scer\FRT.hs3]=3’.RS5 + 3.3’}ED6105
• pnt^aga/ pnt^Δ88: w/y w f;; FRT-82B cu sr pnt^aga/FRT-82B pnt^Δ88
• pnt^aga/pnt¹²⁷⁷: w;; FRT-82B cu sr pnt^aga/P{w[+mC]=lacW} pnt¹²⁷⁷
• pnt^aga/Df(3R)9012: w;; FRT-82B cu sr pnt^aga/Df(3R)Exel9012, P{w[+mC]=XP-U}Exel9012
• pnt^aga/Df(3R)6280: w;; FRT-82B cu sr pnt^aga/Df(3R)Exel6280, P{w[+mC]=XP-U}Exel6280
• pnt^aga/Df(3R)ED6105: w;; FRT-82B cu sr pnt^aga/Df(3R)ED6105, P{w[+mW.Scer\FRT.hs3]=3’.RS5 + 3.3’}ED6105