Extracellular spreading of Wingless is required for Drosophila oogenesis

Summary

Recent studies have investigated whether the Wnt family of extracellular ligands can signal at long range, spreading from their source and acting as morphogens, or whether they signal only in a juxtacrine manner to neighboring cells. The original evidence for long-range Wnt signaling arose from studies of Wg, a Drosophila Wnt protein, which patterns the wing disc over several cell diameters from a central source of Wg ligand. However, the requirement of long-range Wg for patterning was called into question when it was reported that replacing the secreted protein Wg with a membrane-tethered version, NRT-Wg, results in flies with normally patterned wings. We and others previously reported that Wg spreads in the ovary about 50 μm or 5 cell diameters, from the cap cells to the follicle stem cells (FSCs) and that Wg stimulates FSC proliferation. We used the NRT-wg flies to analyze the consequence of tethering Wg to the cap cells. NRT-wg homozygous flies are sickly, but we found that hemizygous NRT-wg/null flies, carrying only one copy of tethered Wingless, were significantly healthier. Despite their overall improved health, these hemizygous flies displayed dramatic reductions in fertility and in FSC proliferation. Further, FSC proliferation was nearly undetectable when the wg locus was converted to NRT-wg only in adults, and the resulting germarium phenotype was consistent with a previously reported wg loss-of-function phenotype. We conclude that Wg protein spreads from its source cells in the germarium to promote FSC proliferation.

Keywords:

DAPI staining – Eggs – Homozygosity – Ovaries – Wnt signaling cascade – Juxtacrine signaling – Bird eggs

Introduction

Wnt signaling is an important and conserved mechanism of cellular communication that controls proliferation and differentiation in many cell types and organisms [1]. The first Wnt ligand identified in any animal was encoded by the wingless (wg) gene in Drosophila melanogaster [2], and subsequent studies of the Wg protein have provided important paradigms for its function in all animals. Beginning over 25 years ago, accumulated studies have established that Wg meets the criteria for a morphogen: it spreads in a gradient from its source exposing cells to different concentrations of the protein, and it triggers different transcriptional responses at different concentrations [3–9]. This morphogen model is based largely on studies in the wing disc, an immature larval structure (anlage) that gives rise to the adult wing. More recently, however, the morphogen function of Wg was significantly challenged by the finding that flies are viable and patterned normally when the gene encoding the secreted Wg ligand was replaced with a membrane-tethered version of Wg, NRT-Wg, capable of signaling but not spreading from its source [10]. Although these flies were reported to be less healthy than their control siblings, their wings were patterned normally, indicating that Wg extracellular spreading is not important for wing pattern. This highly influential study was interpreted to mean that the observed spreading of Wg protein is not important for its signaling, as Wg signaling occurs only in a juxtacrine manner between adjacent cells rather than spreading in a diffusive manner from source cells to target cells [11].

Several groups have re-examined the question of extracellular spreading of Wg using the NRT-wg allele. Two studies have found that Wg protein signals at a distance–requiring extracellular spreading—from gut tissues. In embryos, NRT-Wg cannot replace Wg emanating from the midgut needed to pattern the renal tubules [12], and in pupae Wg spreading is important for normal development of the adult gut [13]. These results, combined with the direct visualization of Wg protein as far away as 11 cells from its source in the wing disc [14], support the model that Wg is not limited to juxtracrine signaling but rather spreads extracellularly to signal at a distance from its source.

We have previously reported that Wg spreads from source cells in the Drosophila germarium, the most anterior region of the ovary [15]. Extracellular Wg protein forms a visible gradient in the germarium, emanating from its source in the cap cells at the anterior tip of the germarium and spreading posteriorly over a distance of about 50 μm or 5 cell diameters to reach the follicle stem cells (FSCs) [16,17]; this extracellular spreading of Wg is facilitated by the glypican Dally-like protein (Dlp), in a mechanism similar to that reported for the wing disc [15,18,19]. In the germarium, Wg acts as a proliferative signal for the FSCs, inducing them to increase their proliferation rate [15,17]. The progeny of FSCs, the follicle cells, encapsulate cysts of germline cells to form immature Drosophila eggs, called egg chambers. Increasing the level of Wg protein that reaches the FSCs, either by overexpressing wg in cap cells or by promoting its extracellular spread, results in increased follicle cell numbers; in contrast, decreasing the level of Wg protein that reaches the FSCs decreases the number of follicle cells, resulting in encapsulation defects [15,17]. Because wg in situ hybridization signal and transcriptional reporters localize to cap cells, and because the visible Wg protein gradient is most concentrated at the cap cells, we concluded that Wg spreads in the germarium [15–17,20]. However, it has also been proposed that Wg signals in a juxtacrine manner to the FSCs, with the ligand emanating from neighboring escort cells, raising questions about the role of Dlp in this process [21,22]

The existence of the NRT-wg allele affords an opportunity to test the model that Wg spreads from the cap cells to the FSCs. If Wg signaling is extracellular over a distance, then NRT-Wg would not be able to substitute for wild-type Wg in the germarium; in contrast, if Wg signaling is juxtacrine in nature, emanating either from neighboring cells or presented via cytonemes from distant cells, then NRT-Wg would be able to substitute for wild-type Wg. In this study, we analyze the consequences of eliminating Wg spreading by tethering Wg in the germarium. In the course of these studies, we determined that NRT-Wg has an inherent dose-dependent toxicity, which we minimized by analyzing NRT-wg/null animals. We find that tethered Wg accumulates around the cap cells and cannot substitute for wild-type Wg in egg development. Further, FSC proliferation is reduced to undetectable levels when Wg is tethered, supporting the idea that wild-type Wg spreads in the germarium and has long-range function. Thus, Wg signals at a distance from the source cells in the germarium, a result that adds to the evidence that Wg signals at a distance and not only in a juxtacrine manner.

Results and discussion

Tethering Wg to the plasma membrane generates a dose-dependent toxicity

Previously it was reported that flies expressing membrane-tethered NRT-wg from the endogenous wg locus–flies which lack all wild-type Wg protein–have reduced fitness [10]. We quantified their survival to adulthood by crossing heterozygous balanced NRT-wg flies and counting the progeny classes (details of the NRT-wg and other genotypes used in this study are shown in Fig 1A). Full survival would be indicated by 33% homozygous NRT-wg flies because the CyO balancer is homozygous lethal; however, we found that only 8% of the progeny were homozygous NRT-wg (Fig 1B). The NRT-wg flies were generated by inserting a cDNA encoding NRT-wg into the wg locus, which was previously mutated to delete wg coding sequence. The appropriate controls for these NRT-wg flies have a wild-type wg cDNA inserted into the same deleted wg locus, which results in a homozygous viable, completely healthy line (called “control” from here on; see Fig 1A for schematic) [10]. In comparison to the homozygous NRT-wg flies, control flies survived to 5 days, as 32% of the progeny were homozygous control, remarkably close to the expected 33% (Fig 1B). If NRT-wg simply reduced wg function (e.g., by reducing Wg spreading) we would expect survival to decrease further when the dose of NRT-wg was halved, in trans to a null allele of wg. We crossed NRT-wg flies to wg null flies (wg^CX4) and counted NRT-wg/null flies and were surprised to discover that survival ratios were significantly improved, with NRT-wg/null flies representing 18% of the progeny (Fig 1B). These results indicate that NRT-wg is not simply a loss-of-function of Wg spreading.

The increased lethality of homozygous NRT-wg flies compared to NRT-wg/null flies can be explained either by a second mutation on the chromosome that contributes to the sickly phenotype when homozygous, or by NRT-wg itself having a weak neomorphic toxicity. To distinguish these possibilities, we utilized another allele, also engineered by the Vincent lab, which can be converted from wild-type wg to NRT-wg by FLP-mediated recombination (Fig 1A), without changing any other locus on the chromosome. We converted this allele to NRT-wg in the male germline with TubP-FLP; this newly flipped NRT-wg allele and the initial unflipped wg allele had identical chromosomal backgrounds. When we tested viability, we found that homozygotes of the unflipped wg line were completely viable whereas homozygotes of the newly flipped NRT-wg allele were barely viable (6% of the population, Fig 1B), mirroring our earlier results despite the identical background. As a final test of possible background elements contributing to lethality, we crossed the newly flipped NRT-wg line to the unflipped wg line, testing the viability of offspring heterozygous for NRT-wg yet homozygous for potential background elements. These progeny were viable (Fig 1C). Thus, despite extensive testing, we found no evidence to support the hypothesis that NRT-wg chromosomes carry a background mutation contributing to lethality, and we conclude NRT-Wg itself has a dose-dependent toxicity. Such toxicity could be explained by increased juxtacrine Wg signaling in neighbors of wg-expressing cells, since Wg that should spread to distant cells is tethered at the cell surface.

After eclosion, homozygous adult flies were difficult to work with because they were short-lived: even when fed well and housed without crowding, 50% of the homozygous NRT-wg female flies died by ~5 days after eclosion (Fig 1D). Like the eclosion phenotype, the lifespan phenotype was partially suppressed in NRT-wg/null flies: 50% lethality was observed by ~12 days after eclosion (Fig 1D). Thus, the dose-dependent toxicity of NRT-Wg was apparent in lifespan as well as in viability to adulthood.

Tethered Wg specifically reduces female fertility

To determine whether Wingless tethering affected egg formation in the germarium, we examined the fertility of NRT-wg flies, mating them to wild-type males and counting the number of eggs laid per day. On days 4 and 5 after eclosion, control flies laid on average 61 and 78 eggs, whereas NRT-wg homozygous flies laid on average 5 eggs on each of these days, a reduction in fertility of 92–94% (Fig 1E). The sharp reduction in fertility indicates that Wg spreading is required for normal egg development. Importantly, with respect to fertility, hemizygous NRT-wg/null flies were not significantly different from homozygotes, with fertility reduced by 85–89% (Fig 1E), indicating that the NRT-wg fertility phenotype is not a result of the unexpected toxicity of homozygous NRT-wg and instead can be attributed to the lack of Wg spreading. These results are consistent with the model that Wg spreading is required for normal egg production.

Upon dissection, the ovaries of NRT-wg homozygous flies were observed to be much smaller than control ovaries with a reduction in the stages of vitellogenesis. In contrast, ovaries of NRT-wg/null hemizygotes were normally sized with representation of all stages of egg chamber development (Fig 2A–2D), suggesting that defects in later-stage vitellogenic eggs are caused by NRT-Wg toxicity, like the viability and lifespan phenotypes. Consistent with the relatively normal appearance of the ovaries, the architecture of the germarium of NRT-wg/null females appeared similar to control females, with egg chambers budded at the posterior of the germarium surrounded by a full complement of follicle cells (Fig 2E). The normal germarium architecture was surprising, because previous studies by us and others have found that conditional loss of wg in adult females results in defects in the germarium, with too many germline cysts and not enough follicle cells to encapsulate them [15,17]. This mismatch of germline cysts to follicle cells can result in the accumulation of germline cysts in an enlarged germarium and/or encapsulation defects in which two germline cysts are squeezed together in one follicle cell covering [15,17].

In wild-type germaria, the Wg protein forms a gradient that extends from the anterior to the FSCs, identified by lineage tracing (Fig 2F) [15]. To validate the NRT-wg allele was tethered as expected, we examined Wg localization in NRT-wg homozygotes and in controls (with the same copy number of the wg gene). In contrast to the gradient of Wg protein found in the controls (Fig 2G), in NRT-wg germaria the Wg protein was sequestered at the cap cells, the cells that produce the Wg protein (Fig 2H). The change in Wg localization provided confirmation that the NRT transmembrane domain functioned in the germarium to tether Wg to the plasma membrane, but it did not explain why germarium architecture appeared normal when egg laying was so infrequent in NRT-wg/null females.

fz3-RFP is a complex reporter of signaling by many Wnt ligands

Previously we reported fz3-RFP was a faithful reporter of Wnt signaling activity in the germarium [15]. The specificity of the fz3-RFP reporter for Wg has not been established, as four Wnts are expressed in the germarium (Wnt2, Wnt4, Wnt6, and wg; [16,20,23,24]). To determine if fz3-RFP responded to multiple Wnt ligands, we analyzed how fz3-RFP responded to the loss of each Wnt ligand (S1 Fig). We measured RFP fluorescence after knocking down each Wnt with a Gal4 driver in the cell type that generates it. bab1^Agal4-5 was used to knock down wg or Wnt6 in cap cells (S1A–S1D Fig), and C587-Gal4 was used to knock down Wnt2 or Wnt4 in escort cells (S1E–S1J Fig). These experiments showed that fz3-RFP reports the signaling activities of at least three Wnt ligands–Wg, Wnt2, and Wnt4: when wg was knocked down in cap cells with bab1^Agal4-5, RFP fluorescence decreased by 35% compared to controls; when Wnt2 was knocked down in escort cells with C587-Gal4 driving either of two RNAi lines, RFP fluorescence decreased by 51 or 45%; when Wnt4 was knocked down in escort cells, RFP fluorescence decreased by 69%. In contrast RFP levels were not affected by knocking down wg in escort cells, confirming that wg is not expressed in escort cells at appreciable levels (S1J Fig). To analyze fz3-RFP in NRT-wg germaria, we needed to recombine these two loci on the 2^nd chromosome. fz3-RFP is unmapped and identifying a recombinant required molecular screening of hundreds of progeny, suggesting that it is tightly linked to the wg locus. Puzzlingly, when germaria were analyzed from the fz3-RFP, NRT-wg recombinant in trans to a wg null allele (fz3-RFP, NRT-wg/null), RFP levels were unexpectedly increased throughout the anterior germarium (S2 Fig). Although fz3-RFP reports the activity of three Wnt ligands, these results suggest there may be a relay of Wnt signaling from the cap cells to the escort cells, and proper levels depend on Wg spreading. We conclude that fz3-RFP is a complex reporter for many types of Wnt signaling, and that the Wnt signaling pattern in the germarium indicated by this reporter is the sum of signaling from several ligands.

Wg spreading is required for FSC proliferation

Previously we found that FSC proliferation is sensitive to the levels of Wg spreading from the cap cells, with more proliferation occurring when the range of Wg spreading is visibly increased. Because the range of Wg spreading is sharply limited by NRT-Wg, we hypothesized that FSC proliferation would be reduced, a possible explanation for the lack of fertility in NRT-wg females. In our earlier study, FSC proliferation was measured by two assays–indirectly by the frequency of FSC mitotic clone induction (see Fig 2F), and directly by visualizing mitotic cells in the region of FSCs by phospho-histone H3 (pH3) staining–and the results of these two assays were in close agreement [15]. For this study, we were unable to use lineage tracing to evaluate FSC proliferation in the NRT-wg/null germaria we were analyzing, as the allele we used contains an FRT at the wg locus, precluding the use of FRTs for mitotic recombination. Fortunately, we were able to use the alternative approach of direct visualization of mitotic cells with anti-pH3 (Fig 3A and 3B), and we analyzed their frequency in the germarium. Currently it is thought that there are three layers or rings of FSCs, extending anteriorly from the 2a/2b boundary, with proliferation occurring mostly in the posterior ring at the boundary of bright Fas3 staining [25–27]. This understanding of FSCs is at odds with an older model, which held there were only two FSCs [28,29], but even in this model, FSC proliferation is at the border between regions 2a and 2b, at the boundary of bright Fas3 antibody staining [29–32]. Thus, we counted pH3 stained nuclei at this border. For comparison, we also measured mitosis in other regions of the germarium: the germline cells found anterior to the 2a/2b boundary; region 2b where mitotic cells are prefollicle cells, the immediate daughters of the FSCs; region 3 where mitotic cells are follicle cells; and stage 6, an ovarian tissue outside the germarium. We analyzed three sets of NRT-wg genotypes. As described previously, we analyzed germaria from homozygous NRT-wg females (Fig 3C) and from hemizygous NRT-wg/null females (Fig 3D), as toxicity is reduced in the latter. We also analyzed germaria from females carrying a conditional allele, in which wg was converted to NRT-wg only after adults eclosed, so as to avoid any contribution of NRT-wg during pupal development of the ovary; the final adult genotypes compared were control-wg (unflipped)/ null versus NRT-wg (flipped)/null (Fig 3E). Examining all three data sets, it is striking that pH3 staining was never observed in the FSC-containing region in any of the 156 NRT-wg germaria, compared to 33 mitotic pH3-positive cells observed in 131 control ovaries. Although we counted only the pH3-positive cells, DAPI staining showed the total number of cells to be similar, consistent with our observations that the germarium is not disrupted in these mutants (Fig 2D and 2E). In the other regions, some reductions in mitosis were observed at varying levels in different NRT-wg backgrounds, with the reduction in mitosis most pronounced in the homozygous NRT-wg, but they were not consistent across genotypes. Together, these results indicate that in NRT-wg germaria, the FSC proliferation rate is so low as to be undetectable, and such a low rate would explain the ten-fold reduction in egg-laying. These results are consistent with the model that Wg spreads from a distance to signal the FSCs, a process previously shown to be mediated by the glypican Dlp [15].

Several mechanisms compensate for the reduction in FSC proliferation

Given the absence of measurable FSC proliferation, we were puzzled by the architecture of the NRT-wg germaria, since germline proliferation was not as dramatically reduced as follicle stem cell proliferation (Fig 3C–3E). Such a mismatch in proliferation rates would result in either the accumulation of too many germline cells in the germarium, encapsulation defects with multiple germline cysts within one follicle, and/or death of germline cells. We noted that some NRT-wg/null germaria displayed encapsulation defects (15/19 germaria), although this defect was not sufficiently penetrant to compensate on its own for the drastic reduction in FSC proliferation. To assay germline death, we performed TUNEL-staining to identify apoptotic cells in germaria (Fig 4A–4C). In NRT-wg/null germaria, the rate of germline death was slightly but significantly elevated over controls, which combined with the encapsulation defects may compensate for the lack in FSC proliferation. Remarkably, in NRT-wg homozygotes nearly every germarium contained a dying germline cyst. These two mechanisms together, apoptosis of unencapsulated cysts and encapsulating multiple cysts into one follicle, likely compensate for the continued proliferation of germline stem cells while the follicle stem cells do not proliferate.

It has been documented that homozygous NRT-wg animals have defects in gut development during pupal metamorphosis, specifically having defects in muscle patterning, cell fate specification, organ folding, and the relative size of various regions [13]. It is reasonable to consider that NRT-wg malformed guts may result in reduced nutritional uptake in adults, and interestingly, several of the ovarian phenotypes we observed in NRT-wg homozygotes—small ovaries, a reduction in follicle cell proliferation, and apoptosis of germline cysts—are observed in wild-type flies with poor nutrition [32,33]. This nutrition interpretation is less likely to explain the reduced fertility of NRT-wg/null hemizygotes because ovary size was normal and overall follicle cell proliferation was not significantly different from controls except for the region of follicle stem cells and their immediate daughters, but nevertheless it is impossible to rule out developmental causes when Wg was tethered throughout development. To separate the role of Wg spreading in development from oogenesis, we generated conditional mutants in which Wg was tethered only after eclosion, using {FRT wg FRT NRT-wg} which converted wg to NRT-wg in adults via intrachromosomal recombination (Fig 4D and 4E). To convert this allele, we first expressed the FLP recombinase ubiquitously with Act5C-Gal4 restricted by Gal80^ts to be expressed only after eclosion, but few adults were recovered. Fortunately, expressing FLP recombinase in the wg expression pattern with wg-Gal4, restricted by Gal80^ts to adults, resulted in NRT-wg/null viable adults. Although these germaria did not generally display encapsulation defects (only 1/32 had a visible encapsulation defect), they did display a significant increase in 16-cell germline cysts (Fig 4F–4H), a previously reported phenotype for the adult-onset loss of wg [17]. These results are consistent with the model that extracellular spreading of Wg is required in the germarium for normal FSC proliferation, and without sufficient FSC proliferation, the germarium backs up with too many germline cysts.

Extracellular spreading of Wg may be part of a complex relay

In these NRT-wg experiments, and in previously reported wg loss-of-function germarium experiments, the wg gene product was manipulated throughout the animal [17]. Thus, these experiments cannot completely exclude the possibility that wg function and Wg spreading are required in another adult tissue to promote FSC proliferation. A conclusive experiment about whether Wg spreads from the cap cells to the FSCs would be to convert wg to NRT-wg only in the cap cells (converting wg to a null allele as a control). Unfortunately, despite substantial efforts we have not been unable to identify a cap-cell specific Gal4 driver that expresses in all of the cap cells, the source of most or all Wg for the germarium, and the escaping cap cells appear to provide sufficient Wg protein to maintain signaling. Nevertheless, because NRT-Wg is capable of signaling to neighboring cells [13], these data clearly demonstrate that extracellular Wg spreading is required for oogenesis and follicle stem cell proliferation. Because Wingless spreading is not required in the wing for patterning but is required for egg production, our findings suggest that different tissues may have different requirements for diffusible versus juxtacrine wingless signaling.

On its own, our data are consistent with a simple model that Wg spreads from the cap cells to the FSCs where it directly signals FSC proliferation. We developed this model from our previous study, where we found that the glypican Dlp plays a central role in spreading Wg to the FSCs, and those data are consistent with our results here. This simple model is challenged, however, by studies of FSCs that cannot transduce Wnt signals. In two papers, the Kalderon lab has found that when FSC clones decrease Wnt signaling by overexpressing dnTCF [25] or when they lose Wnt signaling by mitotic recombination of an arrow mutant [27], FSCs have a small decrease in proliferation measured by EdU incorporation, a decrease insufficient to explain our data that proliferation rates are below the threshold of detection when Wg is tethered. Interestingly, the Kalderon group found that increasing activity of the Wnt signaling pathway dramatically reduces FSC proliferation [25]. One model to reconcile these conflicting data is that Wg indirectly signals FSC proliferation, through an intermediate cell type that cannot be reached with NRT-Wg. Our fz3-RFP Wnt reporter data suggest that total Wnt signaling is unexpectedly increased when Wg is tethered, hinting at a complex relay of Wnt signaling in the germarium. Thus the loss of FSC proliferation in NRT-wg germaria may be indirectly caused by the increase in Wnt signaling. Together, these data suggest that Wg spreading is part of a complex system of Wnt regulation required for FSC proliferation.

Methods

Fly stocks

Flies were cultured on cornmeal-molasses media at 25°C unless otherwise noted. Females were fed on media with wet yeast paste in the presence of w¹¹¹⁸ males, and tossed to new vials every 1–2 days until dissection. The following stocks were a generous gift from Dr. Jean-Paul Vincent: wg{KO; FRT NRT-wg FRT QF}/CyO. wg{KO; FRT wg FRT QF}/CyO. wg{KO; Gal4/CyO}. wg{KO; FRT wg FRT NRT-wg}/CyO. wg{KO; Gal4]. The NRT-wg flies used in this study were wg{KO; FRT NRT-wg FRT QF}, which express NRT-wg and not QF in their unflipped state; the corresponding control flies were wg{KO; FRT wg FRT QF}, which express wg and not QF in their unflipped state. We did not use the non-flippable wg{KO; NRT-wg} line because of reported problems with its control, wg{KO; wg}. In its unflipped state, the wg{KO; FRT NRT-wg FRT QF} chromosome is equivalent to wg{KO; NRT-wg}. The wg^CX4 null allele was obtained from the Bloomington Drosophila Stock Center (BDSC).

For adult-onset tethering of Wg in which wg is replaced with NRT-wg, the cross wg{KO; Gal4}/CyO; Tub-Gal80^ts/TM6 x wg{KO; FRT wg FRT NRT-wg}/CyO; UAS-FLP/TM6 was performed at 18°C. Immediately after progeny eclosed as adults, females were collected of genotype wg{KO; Gal4}/wg{KO; FRT wg FRT NRT-wg}; UAS-FLP/Tub-Gal80^ts and moved to 29°C for 5 days (Fig 3E) or 10 days (Fig 4) before dissection. UAS-FLP, Tub-Gal80^ts, and the alternative driver Act5C-Gal4 were obtained from the BDSC.

FSC mitotic clones (Fig 2F) were induced and stained for lacZ as previously described [15]

To assess fz3-RFP expression in control and Wnt knockdown conditions, fz3-RFP; bab1^Agal4-5, Tub-Gal80^ts or C587-Gal4; fz3-RFP; Tub-Gal80^ts females were crossed to w¹¹¹⁸ (control), UAS-wg^RNAi (NIG-FLY 4889 R-4), UAS-Wnt2^RNAi (BL 29441, BL 36722), UAS-Wnt4^RNAi (BL 29442) or UAS-Wnt6^RNAi (VDRC 104020) males at 18°C. F1 females of the appropriate genotypes were collected within one day of eclosing, crossed with w¹¹¹⁸ males, transferred to fresh vials supplemented with yeast paste and incubated at 29°C for Gal80^ts inactivation and Gal4/UAS-RNAi-mediated knockdown of Wnts for 7 days. During the incubation at 29°C, flies were transferred to fresh vials supplemented with yeast paste every other day.

To assess fz3-RFP expression in germaria in control and NRT-wg heterozygous flies, fz3-RFP/CyO or fz3-RFP, wg{KO; FRT NRT-wg FRT QF, Pax-Cherry}/CyO flies were analyzed. To assess fz3-RFP expression in germaria of NRT-wg homozygotes, fz3-RFP, wg{KO; FRT NRT-wg FRT QF, Pax-Cherry}/wg{KO; FRT NRT-wg FRT QF, Pax-Cherry} were analyzed. To assess fz3-RFP in NRT-wg/null flies, fz3-RFP, wg{KO; FRT NRT-wg FRT QF, Pax-Cherry}/wg^CX4 were analyzed. All crosses were performed at 25°C. Females of the appropriate genotypes were aged at 25°C for 5–7 days.

Genotypes in each figure panel

Fig 1B:

1^st column: wg{FRT wg FRT QF}/wg{FRT wg FRT QF}

2^nd column: wg{FRT NRT-wg FRT QF}/wg{FRT NRT-wg FRT QF}

3^rd column: wg{FRT NRT-wg FRT QF}/wg^CX4

4^th column: wg{KO; FRT wg FRT NRT-wg}/ wg{KO; FRT wg FRT NRT-wg}

5^th column: wg{KO; FRT NRT-wg}/wg{KO; FRT NRT-wg} (germline NRT-wg line generated by flipping out FRT wg FRT cassette in the germline of wg{KO; FRT wg FRT NRT-wg}/ wg{KO; FRT wg FRT NRT-wg} with tub-FLP)

Fig 1C: numerator of fraction: wg{KO; FRT wg FRT NRT-wg}/wg{KO; FRT NRT-wg}

Denominator is sum of two genotypes: wg{KO; FRT NRT-wg}/ wg{KO; FRT wg FRT NRT-wg} + wg{KO; FRT wg FRT NRT-wg}/CyO

Fig 1D:

Both controls: wg{FRT wg FRT QF}/wg{FRT wg FRT QF}

NRT-wg/NRT-wg: wg{FRT NRT-wg FRT QF}/wg{FRT NRT-wg FRT QF}

NRT-wg/null: wg{FRT NRT-wg FRT QF}/wg^CX4

Fig 1E:

Control: wg{FRT wg FRT QF}/wg{FRT wg FRT QF}

NRT-wg/NRT-wg: wg{FRT NRT-wg FRT QF}/wg{FRT NRT-wg FRT QF}

NRT-wg/null: wg{FRT NRT-wg FRT QF}/wg^CX4

Fig 2:

A left, B, and G: wg{FRT wg FRT QF}/wg{FRT wg FRT QF}

A middle, C, and H: wg{FRT NRT-wg FRT QF}/wg{FRT NRT-wg FRT QF}

A right, D, and E: wg{FRT NRT-wg FRT QF}/wg^CX4

F: hsFLP/+; X15-33/X15-29

Fig 3:

A, C control, D control: wg{FRT wg FRT QF}/wg{FRT wg FRT QF}

B, C: wg{FRT NRT-wg FRT QF}/wg{FRT NRT-wg FRT QF}

D: wg{FRT NRT-wg FRT QF}/wg^CX4

E: Control: wg{KO; FRT wg FRT NRT-wg}/wg{KO; Gal4}

Flipped as adults: wg{KO; FRT wg FRT NRT-wg}/wg{KO; Gal4}; UAS-

FLP/Tub-Gal80^ts

Fig 4:

A: wg{FRT wg FRT QF}/wg{FRT wg FRT QF}

B: wg{FRT NRT-wg FRT QF}/wg{FRT NRT-wg FRT QF}

C: wg{FRT wg FRT QF}/wg{FRT wg FRT QF}

wg{FRT NRT-wg FRT QF}/wg{FRT NRT-wg FRT QF}

wg{FRT NRT-wg FRT QF}/wg^CX4

D: wg{KO}/wg{KO; FRT wg FRT NRT-wg}; Actin5C-Gal4, Tub-Gal80^ts/UAS-FLP

E: wg{KO; Gal4}/wg{KO; FRT wg FRT NRT-wg}; Tub-Gal80^ts/UAS-FLP

F: wg{KO; Gal4}/+; Tub-Gal80^ts /UAS-FLP

G: wg{KO; Gal4}/wg{KO; FRT wg FRT NRT-wg}; Tub-Gal80^ts /UAS-FLP

H: control: +/wg{KO; Gal4}; UAS-FLP/Tub-Gal80^ts

Flipped as adults: wg{KO; Gal4}/wg{KO; FRT wg FRT NRT-wg}; UAS-FLP/Tub-Gal80^ts

S1 Fig:

A: w^-/w^-; fz3-RFP/+; bab1^Agal4-5, Tub-Gal80^ts/+

B: w^-/w^-; fz3-RFP/UAS-wg^RNAi; bab1^Agal4-5, Tub-Gal80^ts/+

C: w^-/y^-,w^-; fz3-RFP/UAS-Wnt6^RNAi; bab1^Agal4-5, Tub-Gal80^ts/+

D: Genotypes A-C are shown on the graph

E: C587-Gal4/w^-; fz3-RFP/+; Tub-Gal80^ts/+

F: C587-Gal4/w^-; fz3-RFP/ UAS-wg^RNAi; Tub-Gal80^ts/+

G: C587-Gal4/y^-, v^-; fz3-RFP/ UAS-Wnt2^RNAi; Tub-Gal80^ts/+ (Line #1)

H: C587-Gal4/ y^-, sc^-, v^-; fz3-RFP/ UAS-Wnt2^RNAi; Tub-Gal80^ts/+ (Line # 2)

I: C587-Gal4/ y^-, v^-; fz3-RFP/+; Tub-Gal80^ts/ UAS-Wnt4^RNAi

J: Genotypes E-I are shown on the graph

S2 Fig:

A: w^-/w^-; fz3-RFP/CyO; +/+

B: w^-/w^-; fz3-RFP, wg{KO; FRT NRT-wg FRT QF, Pax-Cherry}/CyO

C: w^-/w^-; fz3-RFP, wg{KO; FRT NRT-wg FRT QF, Pax-Cherry}/wg{KO; FRT NRT-wg FRT QF, Pax-Cherry}

D: w^-/w^-; fz3-RFP, wg{KO; FRT NRT-wg FRT QF, Pax-Cherry}/wg^CX4

E: Genotypes A-D are shown on the graph

Tissue fixation and staining

Ovaries were stained as previously described [15]. Briefly, ovaries were dissected in Schneider’s Drosophila media, fixed in 4% paraformaldehyde (Ted Pella) in PBS and washed with PBST (PBS + 0.1% Triton X-100) before being blocked with 5% normal goat serum in PBST and incubated with primary antibodies at 4°C overnight. To visualize DNA, ovaries were incubated in 1 μg/ml DAPI (Sigma) in PBST for 10 min. Extracellular Wg staining was performed according a published protocol [6] and was also described in [15]. The following primary antibodies were from Developmental Studies Hybridoma Bank (DSHB): mouse anti-Fas3 (7G10, 1:8), mouse anti-Hts (1B1, 1:5), mouse anti-LamC (LC28.26, 1:20), mouse anti-Wg (4D4, 1:3) and rat anti-Vasa (1:10). Rabbit anti-phospho-Histone H3 (Millipore, 1:1,000) was used to label mitotic cells. Secondary antibodies used were goat anti-rabbit IgG and goat anti-rat IgG conjugated to Alexa Fluor 488 (Molecular Probes), Cy3-conjugated goat anti-mouse IgG1 or IgG2a. Stained samples were mounted in Vectashield (Vector Laboratories).

For Fz3-RFP visualization, ovaries were dissected in incomplete Schneider’s media, fixed in 4% paraformaldehyde (Ted Pella) in 1XPBS at RT, washed with 1X PBST, and mounted in Vectashield mounting media containing DAPI (Vector Laboratories).

For TUNEL staining, ovaries were dissected and fixed as described above, washed thoroughly in PBS, and permeabilized with PBS containing 0.1% Triton X-100 and 0.1% sodium citrate. 100 μl of TUNEL reaction mixture (In Situ Cell Death Detection Kit TMR Red, Roche) was added per 5 pairs of ovaries, and samples were incubated at 37°C in the dark for 1 hr. Ovaries were washed thoroughly in PBST, blocked and co-stained with primary antibodies overnight as described above.

Microscopy and imaging

Fluorescent images were taken by a Zeiss Axioimager M2 equipped with an Apotome system and an AxioCam MRm camera (Zeiss). Images were acquired using 40X/1.3 oil EC Plan-NeoFluor or 63X/1.4 oil Plan-Apochromat objective lens at room temperature. The Zeiss Axiovision 4.8 software was used for data acquisition, and projections of Z-stacks were compiled using the Orthoview functions. Images were exported as 16-bit TIFF files and processed with Adobe Photoshop CS4 or Affinity Photo. Brightfield images of ovaries were acquired with a Zeiss Axiocam MRc camera mounted to a Zeiss Lumar V12 stereomicroscope, using a Neolumar S 1.5x objective, X-Cite 120Q light source and Axiovision 4.8 software.

Quantification and statistics

Viability (eclosion) assays were performed at 25°C. Parents were allowed to lay eggs in bottles for 2–3 days. Adults emerged ~10 days later, and all progeny were counted until they stopped eclosing (never longer than 8 days). For the crosses in Fig 1B, 8 bottles (replicates) were counted for control and NRT-wg homozygotes; 4 bottles for NRT-wg/null; and 6 bottles each for the germline flipped and non-flipped genotypes. For the cross in Fig 1C, 6 bottles were counted. For longevity assays, about 20 adult females were collected upon eclosion and cultured with w¹¹¹⁸ males in a vial. Dead females were counted every 2 days when flies were tossed to new vials.

For the fertility assay, female flies were fattened on cornmeal-molasses media with males and fresh yeast for 2–4 days, then 10–12 females and half as many males were put in an egg-collecting cage fed daily with a grape juice plate with yeast paste. Each 24 hours the plate was changed, eggs laid on the plate were counted, and the number of eggs/female/day was determined by dividing by the number of eggs by the number of females present at the start of the collection period. Dead females were removed daily and the number of surviving females was recorded for the next day’s egg collection.

For quantification of fz3-RFP intensities, Z-stacks spanning the entire germarium were acquired at 20X for each genotype, sum Z projection of each germarium was generated in ImageJ, and average fz3-RFP intensity was obtained by outlining the fz3-RFP expression domain in the germarium. All germaria were imaged at same settings for comparisons shown in individual graphs.

To determine the number of mitotic cells in regions of the ovary, ovarioles were dissected and stained with anti-phospho-Histone H3 (pH3), Fas3, and DAPI. Mitotic cells were recognized by anti-pH3 staining. DAPI and Fas3 were used to identify the region of the ovary. Cells at the 2a/2b border, proximal to the region of high Fas3 staining, were considered as FSCs. Sixteen-cell cysts were identified by their typical structure of fusomes labeled with anti-Hts staining and vasa germline staining.

For main figures, student’s t-test (two-tailed, two-sample equal variance) was used for statistical analysis and a p value of <0.05 was considered significant. For dot plots in S1D and S1J and S2E Figs, comparisons of means across indicated genotypes was done using ordinary one-way ANOVA, and significant differences between genotypes was determined by Tukey’s test. Mean and standard error of mean are represented on the dot plots.

Supporting information

S1 Fig [b]
reports signaling activity of Wg, Wnt2, and Wnt4 in the germarium.

S2 Fig [a]
expression is higher germaria.

S1 Table [xlsx]