Downregulation of homeodomain protein Cut is essential for Drosophila follicle maturation and ovulation

Oogenesis is a process of generating mature follicles that are destined to release functional oocytes for fertilization. Although the anatomical structure of ovaries is hugely different in diverse species, many of the molecular and cellular principles governing oogenesis are conserved across different species (Matova and Cooley, 2001). For example, oocyte differentiation from insects to mammals involves incomplete cytokinesis and the receiving of organelles and cytoplasm from sister germ cells (Lei and Spradling, 2016; Matova and Cooley, 2001). Oocyte development is fulfilled in a follicle unit, which consists of somatic follicle cells that encase the oocyte. Proper proliferation and differentiation of somatic follicle cells are essential for oocyte development and, ultimately, for release of the functional oocytes through ovulation (Khristi et al., 2018; Robker et al., 2018); however, the molecular mechanisms regulating somatic follicle cell differentiation and follicle maturation in the final preovulatory stages are poorly understood.

The somatic follicle cells encapsulating oocytes in Drosophila ovaries provide an exceptional model for study of the progressive regulation of cell proliferation, differentiation and maturation. Each ovary in Drosophila consists of about 16 ovarioles, which include a string of developing follicles (egg chambers) wrapped by a circular muscle sheath (Spradling, 1993). At the anterior tip of the ovariole (also named germarium), germline stem cells proliferate to give rise to a 16-cell germline cyst, which becomes wrapped by a layer of somatic follicle cells derived from follicle stem cells to form a stage-1 egg chamber at the end of the germarium (Margolis and Spradling, 1995). Each egg chamber develops through 14 distinct stages to become a mature follicle (Fig. 1A) (Spradling, 1993). Somatic follicle cells undergo three separate cell cycle phases: from stages 1-6, follicle cells are in the mitotic cycle and produce about 855 follicle cells to cover the growing germ cells; from stages 7-10A, they enter three rounds of endocycles to increase DNA content without cell division; and from stages 10B-13, they undergo gene amplification to increase the copy number of chorion genes in order to produce the eggshell (Calvi et al., 1998; Edgar and Orr-Weaver, 2001; Klusza and Deng, 2011; White et al., 2009). The homeodomain transcription factor Cut is expressed in stage 1-6 follicle cells and is required for follicle cell proliferation (Sun and Deng, 2005). The transition from mitotic cycle to endocycle (M/E transition) is induced by Notch signaling, which activates the zinc-finger transcription factor Hindsight (Hnt) to downregulate Cut and inhibit Hedgehog signaling (Deng et al., 2001; Lopez-Schier and St. Johnston, 2001; Sun and Deng, 2007). In the last decade, multiple factors have been found to regulate Notch signaling and the M/E transition (Domanitskaya and Schüpbach, 2012; Fic et al., 2019; Jia et al., 2015; Jouandin et al., 2014; Lee and Spradling, 2014; Lo et al., 2019; Poulton et al., 2011; Starble and Pokrywka, 2018; Vaccari et al., 2010). At the endocycle/gene amplification (E/A) transition, downregulation of Notch signaling permits the activation of ecdysteroid signaling, which upregulates another zinc-finger transcription factor Tramtrack-69 (Ttk69) at stage 10B (Sun et al., 2008). Regulated by microRNA-7, Ttk69 upregulates Cut expression, suppresses Hnt expression and promotes the chorion gene amplification at stage 10B (Sun et al., 2008; Huang et al., 2013). In addition, ecdysteroid signaling also regulates proper follicle cell migration (Hackney et al., 2007), apical microvilli morphogenesis partly through Ttk69 (Romani et al., 2015) and microRNA-318 expression, which also promotes the E/A switch independently of Ttk69 (Ge et al., 2015).

Several populations of somatic follicle cells are established at stage 10B (Fig. 1A), and they go through complex morphogenetic events during stages 10B-14 to synthesize the sophisticated eggshell capable of protecting the developing embryo (Cavaliere et al., 2008; Duhart et al., 2017; Osterfield et al., 2017; Waring, 2000). In addition to regulating chorion gene amplification, Ttk69 plays essential roles in the formation of dorsal appendages (French et al., 2003; Peters et al., 2013). Upregulated Cut in centripetal follicle cells counteracts C/EBP transcription factor in Slow border cells to regulate centripetal follicle cell migration (Levine et al., 2010). In addition, multiple eggshell products have been identified and temporal gene expression patterns in stage 10B-14 follicles have been characterized (Fakhouri et al., 2006; Niepielko et al., 2014; Pyrowolakis et al., 2017; Tootle et al., 2011). Despite extensive study of eggshell formation, it still remains unknown how somatic follicle cells transition into a final maturation status at stage 14 to be ready for ovulation.

We recently characterized the cellular process of ovulation in Drosophila and identified multiple factors that function in stage-14 follicle cells for ovulation. Like mammals, Drosophila ovulation involves an initial step of follicle wall breakdown through trimming of posterior follicle cells, a process dependent on matrix metalloproteinase 2 (Mmp2) (Deady et al., 2015). The exposed oocyte then ruptures into the oviduct, while the rest of the follicle cells compress and remain in the ovary. This process depends on NADPH oxidase (Nox), which produces the signaling molecule hydrogen peroxide in the presence of superoxide dismutase 3 (SOD3) (Li et al., 2018). Both Mmp2 and Nox are activated by elevated intracellular calcium, which is induced by octopamine (OA) binding to its receptor (octopamine receptor in mushroom body; Oamb) in stage-14 follicle cells (Li et al., 2018; Deady and Sun, 2015).

We also discovered that Hnt is re-upregulated in stage-14 follicle cells and is essential for Mmp2 expression in posterior follicle cells at stage 14B and Oamb expression in all follicle cells at stage 14C, a final maturation stage (Deady et al., 2017). Furthermore, the enzyme Shade (Shd; required for 20-hydroxyedcysone synthesis) is also upregulated in stage-14 follicle cells, whereas two of its receptors (EcR.A and EcR.B1) are downregulated, which leads to specific ecdysteroid signaling via EcR.B2 in stage-14 follicle cells for follicle rupture and ovulation (Knapp and Sun, 2017). Thus, it seems there could be another crucial follicle cell transition from stage 13 to stage 14 that allows follicle cells to mature into a fully competent pre-ovulatory state. The question still remains as to what actually regulates this transition of follicle cells into an ovulatory competent state.

In this study, we characterized the follicle cell transition at stage 13/14. We report that Cut, a homolog of human CCAAT displacement protein (Neufield et al., 1992), is downregulated at the stage 13/14 transition. Its downregulation is crucial for this transition, as extended expression of Cut into stage-14 follicle cells leads to ovulation defects and suppressed upregulation of Hnt, a key regulator in follicle rupture. Furthermore, continuation of Cut expression in stage-14 follicle cells also promotes Ttk69 expression, whose downregulation in stage-14 follicle cells is also crucial for full activation of Oamb signaling. Our findings elucidate a maturation pathway involving downregulation of Cut/Ttk69 and upregulation of Hnt at the stage 13/14 transition that is necessary for follicle cells to transition into stage 14 and thus gain ovulatory competency.

Our recent work demonstrated the crucial role of Hnt upregulation in stage-14 follicle cells for ovulation; however, little is known about the mechanism of Hnt upregulation and follicle maturation. To identify potential transcription factors that regulate the follicle cell transition into the stage-14 preovulatory stage, we characterized the expression of several transcription factors in late oogenesis. Our experiments confirmed the previous finding that Hnt is downregulated in main-body follicle cells after stage 10B and through stages 11-13, and then re-upregulated at stage 14 (Fig. 1B-E) (Deady et al., 2015; Sun et al., 2008). Interestingly, we observed that Cut and Ttk69 showed an inverse expression pattern to Hnt from stage 10B to stage 14. After upregulated at stage 10B, Cut expression exhibited a steady decrease in main-body follicle cells and was undetectable by stage 14, when Hnt is upregulated (Fig. 1F-I). Ttk69 showed a similar expression pattern as Cut and was also undetectable in stage-14 main-body follicle cells (Fig. 1J-M).

The antagonistic relationship between Hnt and Cut at the M/E switch in stage-6/7 follicle cells led us to hypothesize that Hnt upregulation in stage-14 follicle cells downregulates Cut expression. To test this hypothesis, we knocked down hnt (also known as peb) in stage-14 follicle cells with the 44E10-Gal4 driver (Deady and Sun, 2015) and examined Cut expression. Cut expression was not detected in stage-14 follicle cells (data not shown). Because 44E10-Gal4 was not efficient at knocking down hnt in early stage 14 (stage 14A) (Deady et al., 2017), we generated an alternative Gal4 line (CG13083-Gal4), which is expressed in stage-13 follicle cells (Fig. S1A). hnt was completely knocked down in stage-14 follicle cells with CG13083-Gal4; however, Cut was still not detectable in these main-body follicle cells and neither was Ttk69 (Fig. S1B-E). Therefore, upregulation of Hnt is not the cause of downregulation of Cut and Ttk69 in stage-14 follicle cells.

The finding that Hnt does not downregulate Cut and Ttk69 expression in stage-14 follicle cells led us to question whether their downregulation is important for follicle maturation. To address this question, we extended Cut expression in stage-14 follicle cells with the 44E10-Gal4 driver (Fig. 2A,B). Females bearing such genetic manipulation laid significantly fewer eggs than control females (Fig. 2C). In addition, ovaries from these females retained significantly more mature follicles after egg laying (Fig. 2D). No morphological defects were observed in mature follicles. These results suggest that ectopic Cut expression in stage-14 follicle cells leads to an ovulation defect.

Drosophila ovulation involves proteolytic degradation of posterior follicle cells and follicle rupture to release the oocyte into the oviduct (Deady and Sun, 2015; Deady et al., 2015). This follicle rupture process is stimulated by octopaminergic signaling in stage-14 follicle cells and can be recapitulated in our ex vivo culture system (Deady and Sun, 2015; Knapp et al., 2018). To test whether ectopic Cut in stage-14 follicle cells disrupts follicle rupture, we isolated mature follicles and stimulated them with OA in our ex vivo culture system. Consistent with the ovulation defect observed in vivo, follicles overexpressing cut showed less than 3% follicle rupture compared with ∼50% follicle rupture in control follicles (Fig. 2E-G). Together, these results suggest that downregulation of Cut in stage-14 follicle cells is essential for proper follicle rupture and ovulation.

To determine why cut-overexpressing follicles are defective in OA-induced follicle rupture, we first investigated whether these follicles can respond to ionomycin stimulation, which bypasses OA/Oamb to induce calcium influx and follicle rupture (Deady and Sun, 2015). Control follicles demonstrated a robust rupture of 95% in response to ionomycin; however, mature follicles with ectopic Cut showed less than 6% follicle rupture, indicating that these follicles were defective in ovulatory signaling downstream of calcium influx (Fig. 3A-C).

Our recent work showed that OA-induced calcium elevation in stage-14 follicle cells leads to Mmp2 activation and reactive oxygen species (ROS) production, both of which are required for follicle rupture (Deady and Sun, 2015; Li et al., 2018). To determine whether ectopic Cut blocks Mmp2 activation in mature follicle cells, we performed in situ zymography in control and cut-overexpressing follicles to examine OA-induced Mmp2 activation. Consistent with previous findings, approximately 63.8% of control follicles exhibited posterior gelatinase activity when stimulated with OA (Fig. 3D,E) (Deady and Sun, 2015). In contrast, only ∼15.5% of cut-overexpressing follicles showed posterior gelatinase activity (Fig. 3D,F), indicating that cut-overexpressing follicles are defective in OA-induced Mmp2 activation. Next, we examined OA-induced superoxide production in control and cut-overexpressing follicles. To do this, we isolated stage-14 follicles and performed a luminescence assay utilizing the dye L-012 to detect the levels of superoxide. Our results showed that OA is able to induce a robust increase in superoxide production in control follicles, consistent with previous results (Fig. 3G) (Li et al., 2018). However, cut-overexpressing follicles showed minimal, if any, superoxide production after OA stimulation (Fig. 3G). Together, these results demonstrate that ectopic expression of Cut is sufficient to inhibit follicle rupture through disruption of both Mmp2 activation and ROS generation.

To find out how ectopic Cut inhibits Mmp2 activation and ROS production, we first examined the expression of Mmp2 and Nox (the key enzyme in ROS production) in stage-14 follicles. Using an Mmp2::GFP fusion gene, we found that Mmp2 was detected in the posterior follicle cells of controls, but not in cut-overexpressing stage-14 follicles (Fig. 4A,B). To confirm this result, we also analyzed Mmp2 mRNA in stage-14 follicles. To our surprise, cut-overexpressing follicles only showed a minimal decrease in Mmp2 mRNA levels (Fig. 4C). However, closer examination of Mmp2::GFP expression elucidated that ectopic Cut downregulated Mmp2 in posterior follicle cells, but not in anterior follicle cells (Fig. S2A,B). This was further supported by qRT-PCR analysis with either posterior or anterior halves of mature follicles. Mmp2 mRNA was significantly reduced in the posterior halves of cut-overexpressing follicles but not in the anterior halves of those follicles (Fig. S2C,D). Therefore, ectopic Cut prevents the upregulation of Mmp2 mRNA and protein in posterior follicle cells. The results also suggest that regulation of Mmp2 expression utilizes different mechanisms in posterior and anterior follicle cells. Overall, these results indicate that downregulation of Cut is required for the proper upregulation of Mmp2 in posterior follicle cells.

Next, we measured the expression of Nox in control and cut-overexpressing follicles. We did not observe any significant difference in Nox mRNA levels (Fig. 4C); however, we did see a significant decrease in the level of Oamb mRNA in cut-overexpressing follicles compared with controls (Fig. 4C), indicating that ectopic Cut inhibits Oamb expression. To support this conclusion, we analyzed the expression of Oamb-RFP, a reporter with the Oamb enhancer region fused with the RFP gene, in control and cut-overexpressing follicles. Oamb-RFP was uniformly expressed in stage-14 follicle cells of control egg chambers (Fig. 4D). In contrast, Oamb-RFP expression was considerably reduced in cut-overexpressing follicle cells (Fig. 4E). Together, these data suggest that ectopic Cut inhibits Oamb expression in all stage-14 follicle cells and Mmp2 expression in posterior follicle cells.

Because both Oamb and Mmp2 are downstream targets of Hnt (Deady et al., 2017), we therefore hypothesize that ectopic Cut inhibits Hnt expression. Consistent with this hypothesis, we observed that Hnt is severely disrupted in stage-14 follicle cells with ectopic Cut expression (Fig. 4F,G). Together, our data suggest that downregulation of Cut in stage-14 follicle cells is essential for upregulation of Hnt, which promotes Oamb and Mmp2 expression.

To determine whether loss of Hnt in cut-overexpressing follicles is responsible for the loss of Mmp2, Oamb and ovulatory competency, we rescued the ovulation defect of cut-overexpressing females with ectopic Hnt. Antibody staining confirmed that both Cut and Hnt were robustly expressed in stage-14 follicle cells when driven by the 44E10-Gal4 driver (Fig. 5A,B). Females with both ectopic Cut and Hnt expression in stage-14 follicle cells (cut/hnt-overexpressing females) only showed a minimal increase in egg-laying capacity compared with females with ectopic Cut expression alone and were significantly different from control females (Fig. 5C). In addition, cut/hnt-overexpressing females still showed egg retention after a 2 day egg-laying experiment (Fig. 5D); follicles from these females did not respond to OA or ionomycin stimulation (Fig. 5E,F). Furthermore, cut/hnt-overexpressing follicles were still defective in OA-induced ROS production (Fig. 5G).

Interestingly, when examining Mmp2 activation in response to OA stimulation, we observed that significantly more cut/hnt-overexpressing follicles than cut-overexpressing follicles showed posterior gelatinase activity; however, the rate of gelatinase activity in cut/hnt-overexpressing follicles was still significantly lower than in control follicles (Fig. 5H). This result suggests that overexpression of Hnt partially restores the ability of cut-overexpressing follicles in OA-induced Mmp2 activation. Consistent with these findings, we also observed partial restoration of Mmp2 and Oamb expression in cut/hnt-overexpressing follicles. More than 50% of cut/hnt-overexpressing follicles exhibited moderate or high levels of Mmp2::GFP expression in their posterior follicle cells, whereas only ∼27% of cut-overexpressing follicles had moderate or high levels of Mmp2::GFP expression (Fig. 5I-K). In addition, qRT-PCR analysis showed that cut/hnt-overexpressing mature follicles exhibited increased levels of Oamb mRNA compared with cut-overexpressing follicles (Fig. 5L). Together, these results suggest that loss of Hnt in cut-overexpressing follicles is partially responsible for the loss of Oamb and Mmp2 expression, consistent with the role of Hnt in regulating these genes. In addition, these results suggest that Cut regulates additional factors (apart from Hnt) to prevent full expression of Oamb and Mmp2 and to inhibit ROS production, both of which lead to defective follicle rupture and ovulation.

To determine what other factors could be regulated by Cut expression in stage-14 follicle cells, we analyzed the expression of genes that show differential expression at the stage-13/14 transition. Both EcR.A and EcR.B1, two isoforms of the ecdysone receptor, are downregulated in stage-14 follicle cells, whereas the 20-hydroxyedcysone-synthesizing enzyme Shd is robustly upregulated in stage-14 follicle cells (Knapp and Sun, 2017). All of these proteins still exhibited normal expression patterns in cut-overexpressing follicle cells (Fig. S3A-F), indicating it is unlikely that ectopic Cut targets ecdysone signaling to disrupt follicle rupture and ovulation. Because Ttk69 is also downregulated in stage-14 follicle cells (Fig. 1M), we then examined Ttk69 expression in cut-overexpressing follicles. Interestingly, Ttk69 was still detected in the nuclei of stage-14 follicle cells with Cut overexpression but was undetectable in the nuclei of control follicle cells (Fig. 6A,B). These findings suggest that ectopic Cut is sufficient to promote Ttk69 expression in stage-14 follicle cells.

To determine whether the expression of Ttk69 in stage-14 follicle cells leads to an ovulation defect, we directly overexpressed ttk69 (the short isoform encoded by ttk) in stage-14 follicle cells using the 44E10-Gal4 driver. Unfortunately, females with such genetic manipulation were not viable, probably because 44E10-Gal4 drives ttk69 expression in developmental stages. We thus included tubGal80^ts to bypass the developmental defect by culturing the animals at 18°C and shifted the newly eclosed females to the restrictive temperature of 29°C. Females with such genetic manipulations were viable, and Ttk69 was indeed overexpressed in stage-14 follicle cells (Fig. 6C). Consistent with our hypothesis, ttk69-overexpressing females exhibited a severely decreased egg-laying rate and significant retention of mature follicles after an egg-laying experiment (Fig. 6D,E). In addition, mature follicles with Ttk69 overexpression were defective in OA-induced follicle rupture and ionomycin-induced follicle rupture (Fig. 6F,G). Furthermore, these follicles were also defective in Mmp2 activation and ROS production (Fig. 6H,I). All these defects were similar to those in cut-overexpressing follicles.

Next, we investigated whether ectopic Ttk69 affects Oamb, Mmp2 and Hnt expression. Similar to cut-overexpressing follicles, ttk69-overexpressing follicles showed normal expression of Nox, but significant reductions in levels of Oamb and Mmp2 mRNA (Fig. 6J). The reduction in Mmp2 mRNA was also limited to posterior halves of the follicles (Fig. S2C). Unfortunately, owing to the complication of the genetic crosses, we were unable to examine Mmp2::GFP expression directly in ttk69-overexpressing follicle cells. In addition, we found that ectopic Ttk69 did not affect the expression of Hnt in stage-14 follicle cells (Fig. 6K,L), nor did it extend Cut expression in stage-14 follicle cells (Fig. S4). Taken together, these results suggest that ectopic Ttk69 in stage-14 follicle cells is able to disrupt follicle rupture by preventing Oamb and Mmp2 expression independently of Hnt and Cut.

Our study showed that Ttk69 is upregulated in cut-overexpressing follicles and that ectopic Ttk69 mimics cut overexpression independently of Hnt. In addition, we also found that Hnt did not block Ttk69 expression in stage-14 follicle cells, as Ttk69 was still detected in cut/hnt-overexpressing follicle cells (Fig. 7A). Therefore, we hypothesized that Ttk69 is the additional factor (aside from Hnt) involved in regulation of follicle rupture in cut-overexpressing follicles. To test this hypothesis, we aimed to rescue the ovulation defect of cut-overexpressing females by simultaneously overexpressing hnt and knocking down ttk with RNA interference (RNAi; for simplicity, we named these females cut/hnt/ttk^RNAi-overexpressing females). Antibody staining confirmed that Ttk69 was indeed knocked down in stage-14 follicle cells of cut/hnt/ttk^RNAi-overexpressing females (Fig. 7B). To our surprise, cut/hnt/ttk^RNAi-overexpressing females still showed an egg-laying defect and follicle retention (Fig. 7C,D), indicating an ovulation defect. Mature follicles from these females did not rupture in response to OA stimulation (Fig. 7E) and were defective in producing ROS (Fig. 7F). In contrast, mature follicles with cut/hnt/ttk^RNAi overexpression showed normal OA-induced Mmp2 activation (Fig. 7G). Overall, these results indicate that simultaneous upregulation of hnt and downregulation of ttk69 is sufficient to rescue the defective OA/Oamb-Ca²⁺-Mmp2 pathway in cut-overexpressing follicle cells, but not the defective OA/Oamb-Ca²⁺-Nox-ROS pathway (Fig. 7H). Indirectly, these results also suggest that Cut influences additional factors in Nox-ROS production downstream of Ca²⁺ (Fig. 7H). In conclusion, our work suggests that downregulation of Cut in stage-14 follicle cells is essential for follicles to transition into full maturation, a preovulatory stage ready for ovulatory stimuli.

Previous studies have focused on the two follicle cell transitions that occur in early and mid-oogenesis. The first transition occurs at stages 6/7, when follicle cells switch from the mitotic cycle to the endocycle. The second transition happens at stages 10A/10B, as follicle cells complete endoreplication and shift into gene amplification (Klusza and Deng, 2011). Together with our previous data, our results here demonstrate that follicle cells in stage-13 and stage-14 egg chambers differ in expression of multiple proteins, including Cut, Ttk69, Hnt, EcR.A, EcR.B1, Shd, Oamb and Mmp2. Therefore, follicle cells experience a novel third follicle cell transition when they develop into stage-14 egg chambers, and this transition is key to reaching final maturation, a state primed for ovulation and follicle rupture. In comparison to the two previous transitions, this final transition has received little attention and no regulatory mechanism has been identified.

In this study, we have illustrated the crucial role of Cut downregulation at this final transition and characterized the epistatic relation between Cut and the rest of the known factors showing changes in expression at this transition. Ectopic Cut expression at stage 14 influences both Mmp2 activity and ROS production by regulating Ttk69, Hnt and other unknown factors; however, ectopic Cut does not influence ecdysteroid signaling at the stage 13/14 transition. Therefore, other signaling pathways are also involved in upregulation of ecdysteroid signaling in follicle cells during this transition. Alternatively, ecdysteroid signaling might function upstream of Cut to downregulate Cut expression and promote this transition. Thus, the late follicle cell transition is a complex process that ensures proper maturation of follicles for ovulation.

Recent work has started to unveil multiple signaling pathways that are required for Drosophila ovulation and are conserved in mammals. For example, Mmps show spatiotemporal activation in both Drosophila and mammalian mature follicles and are essential for ovulation (Curry and Osteen, 2003; Deady and Sun, 2015). ROS, particularly hydrogen peroxide, play essential roles in ovulation in Drosophila and mice (Li et al., 2018; Shkolnik et al., 2011) and, in its role in ovulation, Hnt can be replaced by human homolog Ras response-element binding protein 1 (Deady et al., 2017; Fan et al., 2009). Elevation of intracellular calcium in mature follicle cells, either by OA/Oamb signaling in Drosophila or luteinizing hormone (LH) signaling in mammals, seems important for ovulation (Breen et al., 2013; Deady and Sun, 2015). Paralleled to LH-induced progesterone production prior to ovulation (Richards and Ascoli, 2018), stage-14 follicle cells upregulate the monooxygenase Shd to produce active steroid hormone for ovulation (Knapp and Sun, 2017). Thus, it stands to reason that molecular mechanisms regulating the stage 13/14 transition in follicle cells at the end of oogenesis could be fairly well conserved.

The transcription factors analyzed in this study (Cut, Hnt and Ttk69) exhibit an intricate epistatic relationship throughout Drosophila oogenesis. In follicle cells during early oogenesis, upregulation of Hnt expression at stage 7 is required for suppression of Cut expression and transition from the mitotic cycle to the endocycle (Sun and Deng, 2007). Loss of Hnt leads to extended Cut expression in stage-7 follicle cells, whereas misexpression of Hnt in earlier follicle cells can drive premature downregulation of Cut. It is interesting that Cut knockdown in mitotic follicle cells also leads to premature upregulation of Hnt (Lo et al., 2019), indicating that Hnt and Cut antagonize each other to regulate the M/E switch. Hnt/Cut and Ttk69 do not regulate each other at the M/E transition (Sun and Deng, 2007); however, Ttk69 acts upstream of Hnt and Cut in the E/A switch during mid-oogenesis. Ecdysteroid-induced upregulation of Ttk69 expression leads to the suppression of Hnt and the induction of Cut expression in stage-10B follicle cells (Sun et al., 2008). After stage-10B, Ttk69 also plays a crucial role in regulating dorsal appendage (DA) morphogenesis, as loss of Ttk69 in these stages causes defects in DA tube expansion (French et al., 2003). Ttk69 is required for follicle cells to take on proper cell fates and shapes during these stages, through regulation of multiple downstream genes such as broad, mirror, paxillin and shibire. Furthermore, expression of Ttk69 after stage-10B is crucial for regulation of the expression of numerous chorion genes and is thus essential for proper eggshell formation (Boyle and Berg, 2009; Peters et al., 2013). In this study, we illustrate once again the opposing nature of Hnt and Cut in follicle cells during late oogenesis. Our work demonstrates that misexpression of Cut in stage-14 follicle cells can inhibit the re-upregulation of Hnt in these follicle cells, whereas Hnt knockdown in stage-14 follicle cells does not extend Cut expression. In addition, we found that misexpression of Cut in stage-14 follicle cells is able to induce expression of Ttk69, independently of Hnt. Our data suggest that Cut acts upstream of Hnt and Ttk69 during the late follicle cell transition. Depending on the cellular environment, the epistatic relationship among Cut, Hnt and Ttk69 can change dramatically in the follicle cell lineage.

More work still needs to be done to understand what mechanism regulates induction of Hnt at stage 14. Induction of Hnt in stage 14 is not solely a result of relief from transcriptional repression by Cut, as experiments knocking down Cut protein in follicle cells from stage 10B onwards are not sufficient to induce premature Hnt expression before stage 14 (Fig. S5). Thus, there must be another factor that is required to activate Hnt expression in stage-14 follicle cells or, alternatively, multiple transcriptional repressors need to be relieved in order for Hnt to be turned on. It will be interesting to investigate further the regulation of the complex signaling network necessary for follicle cells to transition into a stage-14 preovulatory state.

Flies were reared on standard cornmeal and molasses food at 25°C and experiments performed at 29°C, unless noted otherwise. Two stage-14 follicle cell-specific Gal4 drivers (47A04-Gal4 and 44E10-Gal4) from the Janelia Gal4 collection (Pfeiffer et al., 2008) and Vm26Aa-Gal4 (Peters et al., 2013) were used. CG13083-Gal4 and Oamb-RFP were generated in this study. The following transgenic lines were used to knock down or overexpress genes in experiments: UAS-cut (a gift from Dr Wu-Min Deng, Florida State University, Tallahassee, USA), UAS-hnt (Bloomington Drosophila Stock Center, stock #5358), UAS-ttk69 (UAS-ttk69, ttk^1e11/TM3, Sb; a gift from Dr Celeste Berg, University of Washington, Seattle, USA), UAS-ttk^RNAi [Vienna Drosophila Resource Center (VDRC), stock #10855], UAS-cut^RNAi (VDRC, stock #5687) and UAS-hnt^RNAi (VDRC, stock #3788). Isolation and identification of stage-14 follicles for ex vivo assays were performed using 44E10-Gal4 to drive UAS-RFP expression. The protein trap line Mmp2::GFP (Deady et al., 2015) was used for Mmp2 expression. UASp-GFP::act79B; UAS-mCD8::GFP was used to map CG13083-Gal4 expression pattern. Control flies for all experiments were prepared by crossing Gal4 drivers to Oregon-R flies.

To generate the CG13083-Gal4 driver, we amplified the 896 bp promoter region of CG13083, whose mRNA is detected in stage-12 follicle cells (Fakhouri et al., 2006), using Oregon-R genomic DNA and the primer pair 5′-CACCTCATGGAAATCATATGCATCAAC-3′ and 5′-CCCGCTGAGTGTCTCTTTTC-3′. This fragment was inserted into pENTR vector (Invitrogen) and subcloned into pBPGUw vector (Addgene) using LR Clonase™ II enzyme mix (Invitrogen) to generate the final construct CG13083-Gal4. To generate the Oamb-RFP reporter, we amplified the 803 bp promoter region of the Oamb gene, the same fragment used to generate the 47A04-Gal4 driver, using the primer pair 5′-TggggtaccccAACTGGCCAGAACTAACGGTTC-3′ and 5′-TTcgggatcccGCCGGGTTTTGTGAAATTAATAG-3′ (restriction enzyme sites are shown in lowercase). The fragments were then digested with KpnI and BamHI (New England Biolabs) and inserted into pRed H-Pelican vector (DGRC) to generate the Oamb-RFP vector. Both vectors were injected into fly embryos to generate the transgenic CG13083-Gal4 and Oamb-RFP stocks using BestGene through standard procedures.

Egg laying and mature follicle analyses were performed as previously described (Deady and Sun, 2015). Five-day-old females fed with wet yeast for 1 day were placed with Oregon-R males (5 females to 10 males) in one bottle for egg laying on molasses plates over 2 days at 29°C (with removal and replacement of plates every 22 h). Ovaries were dissected out after egg laying and mature follicles within each ovary pair were quantified.

The ex vivo follicle rupture assays were performed as described previously (Knapp et al., 2018). Ovaries from 6-day-old virgin females fed with wet yeast for 3 days were dissected out and stage-14 follicles isolated in Grace's insect medium (Caisson Labs, Smithfield, UT, USA). Isolated follicles were divided into groups of ∼30 follicles and cultured at 29°C for 3 h in culture medium containing 20 μM OA (Sigma-Aldrich) or 5 μM ionomycin (Cayman Chemical, Ann Arbor, MI, USA). Each data point represents the percentage (mean±s.d.) of ruptured follicles per experimental group.

In situ zymography to detect gelatinase activity was performed as described previously (Knapp et al., 2018). Mature follicles were cultured for 3 h with 25 μg/ml of DQ-gelatin conjugated with fluorescein (Invitrogen) and 20 μM OA. The number of mature follicles with posterior fluorescein signal was counted afterwards, and data represented as percentage (mean±s.d.) of mature follicles with posterior fluorescein signal.

For qRT-PCR, 4- to 6-day-old virgin females fed with wet yeast for 3 days were used for isolation of mature follicles according to 44E10-Gal4 driving UAS-RFP expression. For experiments in which quantification was performed in either posterior or anterior halves, mature follicles were first isolated and cut in half with Vannas-Tubingen spring scissors (Fine Science Tools, Item No. 15003-08) in cold Grace's medium. Anterior and posterior halves were separated into different tubes in batches of 30. Total RNAs were extracted using Direct-zolTM RNA MicroPrep Kit (Zymo Research) from 60 stage-14 follicles or follicle halves. cDNA synthesis and real-time PCR amplification with three technical repeats were performed as previously described (Knapp and Sun, 2017; Li et al., 2018). Data from one representative biological replicate are presented as mean±s.d. Two to three biological replicates were performed to ensure reproducibility.

Measurement of superoxide production was performed as previously described (Li et al., 2018). Mature follicles (30) were isolated and placed in each well of a 96-well plate with 250 μl of culture media containing 200 μM of L-012 (Wako Chemicals). Plates were placed in a synergy H1 plate reader (BioTek Instruments) for a 45 min L-012 luminescence reading. After 5 min, 20 μM of OA was added to each well. Three to four wells (technical repeats) were used in each experiment for each genotype, and the mean±s.d. of the technical repeats was calculated. Each experiment was repeated at least twice.

Immunostaining was performed following a standard procedure comprising ovary dissection, fixation in 4% EM-grade paraformaldehyde for 13 min, blocking in PBTG (PBS plus 0.2% Triton, 0.5% BSA and 2% normal goat serum), and primary and secondary antibody staining diluted in PBTG. The following primary antibodies were used: mouse anti-Cut (2B10, 1:15), anti-Hnt (1G9, 1:75), anti-EcR.A (15G1a, 1:15) and anti-EcR.B1 (AD4.4, 1:15) antibodies from the Developmental Study Hybridoma Bank; rabbit anti-Ttk69 (1:100; a gift from Dr Wanzhong Ge, Zhejiang University, China) (Ge et al., 2015), anti-GFP (1:4000; Invitrogen) and anti-Shd (1:250; a gift from Dr Michael O'Connor, University of Minnesota, Minneapolis, USA). Alexa Fluor 488 and Alexa Flour 568 goat secondary antibody (1:1000; Invitrogen) were used as secondary antibodies. Images were acquired using a Leica TCS SP8 confocal microscope or Leica MZ10F fluorescent stereoscope with a sCOMS camera (PCO.Edge) and assembled using Photoshop software (Adobe) and ImageJ.

Statistical tests were performed using Prism 7 (GraphPad, San Diego, CA). Quantification results are presented as mean±s.d. Statistical analysis was conducted using Student's t-test. For comparison of more than two means, one-way ANOVA with post-hoc Fisher's least significant difference test was used. For comparison of distribution, the chi-squared test was used.

We thank Drs Wu-Min Deng, Celeste Berg, Wanzhong Ge and Michael O'Connor for sharing reagents and fly lines; BDSC and VDRC for fly stocks; and the Developmental Studies Hybridoma Bank for antibodies. We thank Dr Kyle Hadden for sharing equipment and Lylah Deady and Timothy King in J.S.’s laboratory for technical support and discussion. The Leica SP8 confocal microscope was supported by a National Institutes of Health Award (S10OD016435) to Akiko Nishiyama.