ABSTRACT
The Netrin receptor Frazzled/Dcc (Fra in Drosophila) functions in diverse tissue contexts to regulate cell migration, axon guidance and cell survival. Fra signals in response to Netrin to regulate the cytoskeleton and also acts independently of Netrin to directly regulate transcription during axon guidance in Drosophila. In other contexts, Dcc acts as a tumor suppressor by directly promoting apoptosis. In this study, we report that Fra is required in the Drosophila female germline for the progression of egg chambers through mid-oogenesis. Loss of Fra in the germline, but not the somatic cells of the ovary, results in the degeneration of egg chambers. Although a failure in nutrient sensing and disruptions in egg chamber polarity can result in degeneration at mid-oogenesis, these factors do not appear to be affected in fra germline mutants. However, similar to the degeneration that occurs in those contexts, the cell death effector Dcp-1 is activated in fra germline mutants. The function of Fra in the female germline is independent of Netrin and requires the transcriptional activation domain of Fra. In contrast to the role of Dcc in promoting cell death, our observations reveal a role for Fra in regulating germline survival by inhibiting apoptosis.
INTRODUCTION
Netrin and its receptor Deleted in colorectal cancer (Dcc, Frazzled in Drosophila) play crucial roles in the development and maintenance of multiple tissue types, including the Drosophila heart and gut, as well as the vertebrate pancreas, lung, mammary glands, vascular system and musculature (Lai Wing Sun et al., 2011; Macabenta et al., 2013; Pert et al., 2015). In the developing nervous systems of invertebrates and vertebrates, Netrin signals through its receptors Frazzled (Fra)/Dcc to promote attractive axon guidance (Boyer and Gupton, 2018). This activity requires receptor interactions with intracellular effector proteins that remodel the growth cone cytoskeleton to steer the navigating axon (Zang et al., 2021).
In Drosophila commissural neurons, Fra also acts independently of Netrin to regulate gene expression (Neuhaus-Follini and Bashaw, 2015; Yang et al., 2009). In this context, Fra is proteolytically processed to release its intracellular domain (ICD), which can translocate into the nucleus and activate transcription of commissureless (comm) (Neuhaus-Follini and Bashaw, 2015). It is unknown whether this mode of signaling is conserved or functions outside of the nervous system. In human embryonic kidney cells, vertebrate orthologs of Fra, Neogenin and Dcc, activate transcription of a luciferase reporter gene, and the Neogenin ICD can bind upstream of open reading frames and regulate their expression (Goldschneider et al., 2008; Taniguchi et al., 2003). This suggests that the ability of Fra to activate transcription is conserved across species. However, it is unclear how the transcriptional activity of Fra is regulated. Nor is it known how Fra interacts with transcriptional machinery and what other target genes it may regulate.
A second non-canonical function of Dcc is to act as a tumor suppressor to promote cell death. In the absence of Netrin, expressing Dcc in human embryonic kidney cells, prostate and colon carcinoma cells, and neuroblastomas lead to cleavage of the Dcc ICD by caspase 3, activating caspase-mediated cell death (Chen et al., 1999; Forcet et al., 2001; Mehlen et al., 1998). These studies have led to the ‘dependence receptor’ hypothesis, which posits that Dcc depends on the presence of Netrin to prevent cell death. Dcc also promotes cell death when Netrin expression is limited in the mouse spinal cord and enteric nervous system, and the chick neural tube (Castets et al., 2012; Furne et al., 2008). Furthermore, netrin conditional knockout in rodent brains leads to Dcc-mediated dopaminergic neuron death (Jasmin et al., 2021). Although Dcc may function as a dependence receptor in human tumor cells and in some vertebrate neurons, whether this function is conserved in other species and other tissue contexts remains to be determined.
Netrin-Fra signaling has been predominantly studied in the developing nervous system; however, this signaling pathway plays diverse and essential roles in many tissues (Lai Wing Sun et al., 2011; Macabenta et al., 2013; Pert et al., 2015). Netrin and its receptors may also play a role in reproduction. In Drosophila, netrinAB mutant females have decreased fertility (Newquist et al., 2013), although it is unclear whether this reflects tissue-intrinsic or neuronal requirements. Although the nervous system profoundly influences organismal physiology, including reproduction (Drummond-Barbosa, 2019), Netrin also affects cell migration and adhesion by acting on its receptors in other tissues. For example, Netrin (Unc-6) is required for C. elegans reproductive system innervation (Asakura et al., 2007), and Unc-6 secreted from neurons also shapes the developing reproductive system (Ziel et al., 2009). In other cases, Netrin and its receptors have clear tissue-intrinsic roles. For example, NetrinA is expressed in the Drosophila germarium and is required in escort cells for germline stem cell maintenance (Tu et al., 2020). Intriguingly, Netrin and Dcc are expressed in porcine and human adult female reproductive tissues, respectively (Basini et al., 2011; Enomoto et al., 1995; Maeda et al., 2008; Saegusa et al., 2000). Although there are some hints that Netrin may be important for blood vessel development in porcine reproductive tissues, the importance of Netrin and Dcc to reproductive tissue development and function remains largely unknown (Basini et al., 2011; Enomoto et al., 1995). Furthermore, the mechanism of Dcc signaling in these tissues has yet to be explored.
To further investigate the diverse signaling mechanisms of Fra/Dcc, we sought to define a novel tissue context that would allow us to directly observe changes in cell morphology and survival; therefore, we turned to the Drosophila ovary. The Drosophila ovary is an excellent system with which to address links between cell morphology and survival, as the process of oogenesis requires coordination of multiple morphogenetic events as egg chambers grow and differentiate. Furthermore, germline survival depends on the suppression of apoptosis (Peterson et al., 2003), allowing us to test whether Fra regulates this process. Drosophila ovaries consist of ovarioles, which are strings of developing egg chambers. Oocyte development begins in the germarium. At the anterior tip of the germarium, germline stem cells divide to give rise to daughter cystoblasts, which divide four times with incomplete cytokinesis to create 16-cell cysts containing one oocyte and 15 nurse cells (Spradling, 1993). Nurse cells endoreplicate, producing mRNA and proteins that are eventually transferred to the oocyte and are necessary for its growth (Spradling, 1993). At the midpoint of the germarium, somatic follicle cells encapsulate the cyst in a single layer as it buds off of the germarium to form an egg chamber (Kirilly and Xie, 2007) (Fig. 1A). In the vitellarium, egg chambers progress through 14 stages of growth that are characterized by well-established morphological criteria (King, 1970). At mid-oogenesis, also known as vitellogenesis, the oocyte grows dramatically as it takes up yolk, and follicle cells migrate to surround the growing oocyte. Shortly thereafter, nurse cells dump their contents into the oocyte, follicle cells create the vitelline membrane of the egg and the mature egg is ovulated (McLaughlin and Bratu, 2015).
Fra localizes to the cell membrane of both follicle and germ cells in the Drosophila ovary. (A) Schematic of an ovariole with the germarium at the anterior and multiple egg chamber stages, each encapsulated by a single layer of follicle cells. Mid-oogenesis starts at stage eight, when the oocyte begins to take up yolk. (B-D) Single-channel images of a fra-MiMIC ovariole stained for (B) GFP (GFP-Fra, green) and (C) Fra (magenta), along with the merged image (D). (B′-D′) Detailed views of egg chamber indicated in B-D, respectively. Arrows indicate Fra on nurse cell membranes; arrowheads indicate Fra enrichment on the apical side of follicle cell membranes. Asterisks mark Fra localized to a ring canal. (E-G) Single channel images of an ovariole with fra3 clones, where the GFP+ cells are wild type and GFP− cells are mutant for fra, stained for (E) GFP (green) and (F) Fra (magenta), along with the merged image (G). Scale bars: 20 μm.
Fra localizes to the cell membrane of both follicle and germ cells in the Drosophila ovary. (A) Schematic of an ovariole with the germarium at the anterior and multiple egg chamber stages, each encapsulated by a single layer of follicle cells. Mid-oogenesis starts at stage eight, when the oocyte begins to take up yolk. (B-D) Single-channel images of a fra-MiMIC ovariole stained for (B) GFP (GFP-Fra, green) and (C) Fra (magenta), along with the merged image (D). (B′-D′) Detailed views of egg chamber indicated in B-D, respectively. Arrows indicate Fra on nurse cell membranes; arrowheads indicate Fra enrichment on the apical side of follicle cell membranes. Asterisks mark Fra localized to a ring canal. (E-G) Single channel images of an ovariole with fra3 clones, where the GFP+ cells are wild type and GFP− cells are mutant for fra, stained for (E) GFP (green) and (F) Fra (magenta), along with the merged image (G). Scale bars: 20 μm.
Oogenesis is an energy-intensive process, and it stands to reason that such an investment should be reserved for the production of high-quality eggs (Laws and Drummond-Barbosa, 2017). Poor nutrient conditions can trigger programmed cell death both in the germarium and during mid-oogenesis (Drummond-Barbosa and Spradling, 2001). This checkpoint activation leads cell death effector caspase Dcp-1 cleavage and activation, then egg chamber degeneration (Peterson et al., 2003). Similarly, egg chamber abnormalities, such as disrupted polarity (Beachum et al., 2021; Tanentzapf et al., 2000) or follicle cell death (Chao and Nagoshi, 1999), can trigger the mid-oogenesis checkpoint. Little is known about the mechanism whereby these developmental events trigger the checkpoint at mid-oogenesis.
Here, we find that, although Fra is expressed in both germline and somatic cells in the Drosophila ovary, it is required specifically in the germline for progression through mid-oogenesis. The starvation response in fra mutant germline cysts is unaffected, indicating that fra is unlikely to regulate the ovarian response to diet. Furthermore, both germline and follicle cell polarity appear to be intact in egg chambers with fra mutant germlines. Nevertheless, ovarioles containing these mutant egg chambers express activated Dcp-1 and initiate apoptosis. Thus, in contrast to vertebrate systems where Dcc promotes apoptosis in some contexts, our results indicate that Fra can play the opposite role to promote germline survival by negatively regulating apoptosis. Global netrin mutants have morphologically normal ovaries, suggesting that Fra acts independently of Netrin in this context. Intriguingly, the transcriptional activation domain of Fra is required for egg chambers to progress through mid-oogenesis, providing in vivo evidence that Fra may act as a transcription factor outside of the nervous system. Together, this work reveals a crucial Netrin-independent role for Fra in allowing progression through mid-oogenesis by preventing apoptosis, and establishes the ovary as a system for investigating Fra signaling.
RESULTS
Fra is expressed in the ovarian germline and the soma
To determine whether and where Fra is expressed in the ovary, we took advantage of the fra-MiMIC allele (Nagarkar-Jaiswal et al., 2015), which produces a GFP-tagged Fra from its endogenous locus. GFP-Fra is expressed throughout the ovariole, with higher expression in egg chambers that have bud from the germarium (Fig. 1B,B′). In the vitellarium, GFP-Fra is present on the membrane of somatic follicle cells, where it is enriched at the apical domain (Fig. 1B′, arrowhead). We also detect GFP-Fra on both nurse cell (Fig. 1B′, arrow) and oocyte membranes in the germline. Additionally, GFP-Fra is present on F-actin-enriched ring canals (Fig. 1B′, asterisk), the intracellular bridges between syncytial germ cells. A similar expression pattern is seen with a c-terminal Fra antibody (Fig. 1C,C′) (Kolodziej et al., 1996). To test the specificity of this antibody in the ovary, we generated genetic mosaic females and compared Fra expression in homozygous null clones with neighboring cells (Fig. 1E-G). As expected, GFP-negative cells do not contain Fra (Fig. 1F).
Fra is required for oogenesis
We generated fra mosaic flies using the Flp-dominant female sterile technique to determine whether there is an ovary-intrinsic role for Fra (Chou and Perrimon, 1996). Ovo is a transcription factor involved in female germline differentiation, and the ovoD allele produces a dominant-negative protein that causes germline degeneration early in oogenesis (Vazquez-Pianzola et al., 2011). We used a heat-shock-inducible flippase to induce recombination at FRT sites on chromosome 2R, where one chromosome carried the ovoD allele and the other carried either a wild-type or mutant fra allele. As germline cells carrying ovoD die early in oogenesis, we were able to compare control ovarioles with germlines that are nearly completely mutant for fra. We generated fra mutant germlines using three different alleles: fra3 and fra4 (null alleles), and fra6 (a hypomorphic allele) (Kolodziej et al., 1996; Yang et al., 2009). In control ovarioles, egg chambers bud from the germarium, grow progressively larger and rarely degenerate (McLaughlin and Bratu, 2015) (Fig. 2A). In fra germline mutants, egg chambers appear morphologically normal prior to mid-oogenesis; however, a striking number of ovarioles contain degenerating egg chambers at the onset of mid-oogenesis (Fig. 2B-D). This degeneration is easily recognized by the presence of pyknotic nurse cell nuclei (Fig. 2B′), and is accompanied by the apparent enlargement of some follicle cells (Fig. 2B″), suggesting germline engulfment (Etchegaray et al., 2012). Consistent with differences in Fra function in these alleles, only 42% of fra6 ovarioles contain degenerating egg chambers, whereas fra3 and fra4 ovarioles exhibit 60.22% and 85.71% degeneration, respectively (Fig. 2C,D). As both fra3 and fra4 are protein null alleles, the increased degeneration seen in fra4 is most likely due to a linked background mutation. These observations suggest that Fra is required in the germline for egg chambers to progress through mid-oogenesis.
Fra is required for germline survival in the ovary. (A) Wild-type ovariole from ovoD control flies stained for 1B1 (magenta) to mark cell membranes and DRAQ5 (gray) to mark nuclei. (A′) DRAQ5 channel of boxed region in A. Arrow indicates a healthy nurse cell nucleus. (B) Ovariole with a fra3 mutant germline. (B′) DRAQ5 channel of boxed region in B. The arrowhead indicates a pyknotic nurse cell nucleus. (B″) 1B1 channel of boxed region in B illustrates morphological changes to follicle cell membrane. (C) Percentage of ovarioles containing a degenerating egg chamber in ovarioles with a wild-type germline versus a fra3 mutant germline. n=122 and 36 ovarioles from one trial. (D) Percentage of ovarioles containing a degenerating egg chamber in ovarioles with wild-type and mutant germlines. n=26, 50, 93 and 21 ovarioles from one trial. Statistical significance determined by Fisher's exact test and P-value adjusted using Bonferroni-Dunn for multiple comparisons, ***P<0.0003, ****P<0.0001. Scale bars: 20 μm.
Fra is required for germline survival in the ovary. (A) Wild-type ovariole from ovoD control flies stained for 1B1 (magenta) to mark cell membranes and DRAQ5 (gray) to mark nuclei. (A′) DRAQ5 channel of boxed region in A. Arrow indicates a healthy nurse cell nucleus. (B) Ovariole with a fra3 mutant germline. (B′) DRAQ5 channel of boxed region in B. The arrowhead indicates a pyknotic nurse cell nucleus. (B″) 1B1 channel of boxed region in B illustrates morphological changes to follicle cell membrane. (C) Percentage of ovarioles containing a degenerating egg chamber in ovarioles with a wild-type germline versus a fra3 mutant germline. n=122 and 36 ovarioles from one trial. (D) Percentage of ovarioles containing a degenerating egg chamber in ovarioles with wild-type and mutant germlines. n=26, 50, 93 and 21 ovarioles from one trial. Statistical significance determined by Fisher's exact test and P-value adjusted using Bonferroni-Dunn for multiple comparisons, ***P<0.0003, ****P<0.0001. Scale bars: 20 μm.
Fra is cell-autonomously required in the germline for egg chamber survival
Although ovoD clones generate germlines that are almost entirely mutant for fra, this approach also creates undetectable follicle cell clones, albeit less frequently. As Fra is expressed in both the soma and the germline, we investigated where Fra is required for egg chambers to progress through mid-oogenesis. Because all three fra alleles lead to degeneration with the ovoD system, we selected one allele, fra3, to continue our analysis. To determine whether Fra is required for oogenesis in the germline or follicle cells, or both, we generated negatively marked homozygous fra clones, which we identified by the absence of GFP (Fig. 3A,C). We identified ovarioles containing GFP-negative clones (fra mutants) in either follicle cells or the germline, and determined whether these ovarioles also contained degenerating egg chambers (Fig. 3A,C). We compared the rate of degeneration with control mosaic ovarioles, where all cells are wild type at the fra locus. As cell death leads to membrane perforation and cytoplasmic GFP leakage, we could not definitively determine the GFP status of degenerating egg chambers. Therefore, we restricted our analysis to ovarioles that had germline or follicle cell clones in non-degenerating egg chambers. Consistent with our results using the ovoD system, ovarioles with at least one GFP-negative fra mutant germline cyst contain more degenerating egg chambers than ovarioles with control cysts (Fig. 3A,B). Furthermore, egg chamber degeneration in fra mutant mosaic germlines primarily occurs at mid-oogenesis (Table 1). To evaluate the contribution of follicle cells to this phenotype, we quantified degeneration in ovarioles with large follicle cell clones (>50% of each egg chamber). Similar to control ovarioles, fra3 mosaic ovarioles with large follicle cell clones rarely contain degenerating egg chambers (Fig. 3C,D), suggesting that Fra is dispensable in follicle cells for progression through mid-oogenesis. To confirm that the degeneration in germline fra mutants is due to the loss of Fra, we used the germline-specific driver nanos-GAL4 to express a full-length Fra transgene in fra mutant mosaic flies (Fig. 3E). As expected, germline expression of the full-length Fra receptor rescues the fra mutant degeneration phenotype (Fig. 3F). Thus, Fra is required specifically in the germline to promote egg chamber progression through mid-oogenesis.
Fra is required in the germline for egg chambers to progress through mid-oogenesis. (A) Ovariole stained for GFP (green) and 1B1 (cell membranes; magenta) with fra3 germline clones (GFP−, white asterisks). A degenerating egg chamber is at the posterior end of the ovariole (white arrow). (A′) GFP channel from A. (B) Percentage of ovarioles containing a degenerating egg chamber out of all ovarioles with at least one GFP− germline cyst. n=82 and 63 ovarioles from one trial. (C) Ovariole stained for GFP (green) and 1B1 (cell membranes; magenta) with fra3 mutant follicle cells (GFP−, white arrowhead) and few wild-type follicle cells (GFP+, yellow arrowhead). (C′) GFP channel from C. (D) Percentage of ovarioles containing a degenerating egg chamber in ovarioles with large GFP− follicle cell clones (more than 50% GFP− follicle cells in all egg chambers). n=44 and 15 ovarioles from two independent trials. (E) Ovariole with fra3 germline clones (GFP−, white asterisks) expressing full-length Fra tagged with HA (magenta) in the germline. DRAQ5 marks nuclei (gray). (E′) HA channel from E. (F) Percentage of ovarioles containing a degenerating egg chamber. This graph also appears in Fig. 7C with additional genotypes that were tested simultaneously using the same controls. n=208, 260 and 116 ovarioles from at least three independent trials. Statistical significance determined by Fisher's exact test and P-values were adjusted using Bonferroni-Dunn for multiple comparisons, *P<0.05, ***P=0.0003, ****P<0.0001. Error bars represent s.d. Scale bars: 20 μm.
Fra is required in the germline for egg chambers to progress through mid-oogenesis. (A) Ovariole stained for GFP (green) and 1B1 (cell membranes; magenta) with fra3 germline clones (GFP−, white asterisks). A degenerating egg chamber is at the posterior end of the ovariole (white arrow). (A′) GFP channel from A. (B) Percentage of ovarioles containing a degenerating egg chamber out of all ovarioles with at least one GFP− germline cyst. n=82 and 63 ovarioles from one trial. (C) Ovariole stained for GFP (green) and 1B1 (cell membranes; magenta) with fra3 mutant follicle cells (GFP−, white arrowhead) and few wild-type follicle cells (GFP+, yellow arrowhead). (C′) GFP channel from C. (D) Percentage of ovarioles containing a degenerating egg chamber in ovarioles with large GFP− follicle cell clones (more than 50% GFP− follicle cells in all egg chambers). n=44 and 15 ovarioles from two independent trials. (E) Ovariole with fra3 germline clones (GFP−, white asterisks) expressing full-length Fra tagged with HA (magenta) in the germline. DRAQ5 marks nuclei (gray). (E′) HA channel from E. (F) Percentage of ovarioles containing a degenerating egg chamber. This graph also appears in Fig. 7C with additional genotypes that were tested simultaneously using the same controls. n=208, 260 and 116 ovarioles from at least three independent trials. Statistical significance determined by Fisher's exact test and P-values were adjusted using Bonferroni-Dunn for multiple comparisons, *P<0.05, ***P=0.0003, ****P<0.0001. Error bars represent s.d. Scale bars: 20 μm.
Germline fra is not required for nutrient sensing and does not appear to impact polarity
In fra germline clones, degeneration occurs at mid-oogenesis. Although a low level of egg chamber degeneration occurs stochastically at this checkpoint, flies subjected to specific stressors, including starvation and disruptions to egg chamber polarity, experience higher levels of degeneration (Beachum et al., 2021; Drummond-Barbosa and Spradling, 2001; Tanentzapf et al., 2000). Given their morphological similarities, we reasoned that Fra could impinge on the ovarian response to diet. Alongside egg chamber degeneration at mid-oogenesis, starved flies exhibit a dramatic shift in the localization of the Insulin-responsive transcription factor Forkhead Box O (FoxO). Under well-fed conditions, Drosophila insulin-like peptides signal through the Insulin receptor, leading to FoxO phosphorylation and sequestration in the cytoplasm (Manning and Toker, 2017; Nässel et al., 2015). When insulin signaling is low, unphosphorylated FoxO is transported into the nucleus to activate target genes (Nässel et al., 2015). Insulin signaling is required by the germline for egg chamber progression through mid-oogenesis (Drummond-Barbosa and Spradling, 2001; LaFever and Drummond-Barbosa, 2005), and although FoxO is not an effector of insulin signaling in this context (LaFever et al., 2010), its re-localization is insulin dependent. We tested whether the absence of fra could shift FoxO from the cytoplasm into the nucleus. However, we find that FoxO localization is unchanged in GFP-negative fra mutant germline cysts (Fig. 4A-C), indicating that insulin signaling is not compromised in these cells. To explore whether Fra controls a different aspect of the ovarian response to diet, we tested whether starvation could further increase degeneration in ovarioles with fra mutant germline cysts. When flies are starved or fed a protein-poor diet, egg chamber degeneration at mid-oogenesis increases (Drummond-Barbosa and Spradling, 2001; Shimada et al., 2011). If fra mutant germline cysts degenerate due to a failure to sense the nutrient environment, then starving these flies would not dramatically increase egg chamber degeneration. However, upon starvation, flies with fra mutant germlines have a drastic increase in degeneration, closely mirroring the response of flies with wild-type germlines (Fig. 4D). Taken together, these results indicate that germline Fra is unlikely to be involved in nutrient sensing during oogenesis.
Loss of fra does not affect the ovarian response to diet or the polarity markers Orb, Armadillo and Discs large. (A) Ovariole from a fed fly with negatively marked fra3 germline clones (GFP−, white asterisks) stained for GFP (green) and (A′) Foxo (gray). (B) Ovariole from a starved fly with negatively marked fra3 germline clones (GFP−, white asterisks) stained for GFP (green) and (B′) Foxo (gray). (C) Percentage of egg chambers with negatively marked germline clones that had Foxo localized to the nurse cell nuclei. n=47, 18 and 19 egg chambers. (D) Percentage of control and fra3 germline mutant ovarioles containing a degenerating egg chamber. Flies were either fed a regular diet or starved for 6 h before dissection. Ovarioles with fra3 germlines still respond to diet. n=234, 137, 229 and 159 ovarioles from two independent trials. (E) Schematic depicting normal localization of Orb, Arm and Dlg in an ovariole. (F,G) Egg chamber with either a negatively marked control germline clone (F,F′) or a negatively marked fra3 germline clone (G,G′) stained for GFP (green), Orb and 1B1 (both gray). Orb is localized to the oocyte in both egg chambers (arrowhead). (H,I) Egg chambers with either negatively marked control germline clones (H,H′) or a negatively marked fra3 germline clone (I,I′) stained for Armadillo (magenta) and GFP (green). (J,J′) Ovariole containing negatively marked fra3 germline clones (white asterisks) stained for Discs large (magenta), GFP (green) and DRAQ5 (gray). Error bars represent s.d. Scale bars: 20 μm.
Loss of fra does not affect the ovarian response to diet or the polarity markers Orb, Armadillo and Discs large. (A) Ovariole from a fed fly with negatively marked fra3 germline clones (GFP−, white asterisks) stained for GFP (green) and (A′) Foxo (gray). (B) Ovariole from a starved fly with negatively marked fra3 germline clones (GFP−, white asterisks) stained for GFP (green) and (B′) Foxo (gray). (C) Percentage of egg chambers with negatively marked germline clones that had Foxo localized to the nurse cell nuclei. n=47, 18 and 19 egg chambers. (D) Percentage of control and fra3 germline mutant ovarioles containing a degenerating egg chamber. Flies were either fed a regular diet or starved for 6 h before dissection. Ovarioles with fra3 germlines still respond to diet. n=234, 137, 229 and 159 ovarioles from two independent trials. (E) Schematic depicting normal localization of Orb, Arm and Dlg in an ovariole. (F,G) Egg chamber with either a negatively marked control germline clone (F,F′) or a negatively marked fra3 germline clone (G,G′) stained for GFP (green), Orb and 1B1 (both gray). Orb is localized to the oocyte in both egg chambers (arrowhead). (H,I) Egg chambers with either negatively marked control germline clones (H,H′) or a negatively marked fra3 germline clone (I,I′) stained for Armadillo (magenta) and GFP (green). (J,J′) Ovariole containing negatively marked fra3 germline clones (white asterisks) stained for Discs large (magenta), GFP (green) and DRAQ5 (gray). Error bars represent s.d. Scale bars: 20 μm.
Polarity of both follicle cells and germline cysts determines the embryonic body plan (Merkle et al., 2020), and disruptions in the polarity of either follicle cells or germ cells can lead to an increase in egg chamber degeneration during oogenesis (Beachum et al., 2021; Tanentzapf et al., 2000). We evaluated fra germline clones for defects in germline and somatic polarity. Shortly after the formation of the 16-cell cyst, Orb accumulates in the oocyte, where its expression is maintained throughout oogenesis (Lantz et al., 1994). In the vitellarium, the oocyte is positioned at the posterior end of the egg chamber (Fig. 4E) (King, 1970). To evaluate germline cyst polarity, we monitored Orb and oocyte localization in GFP-negative cysts. Orb accumulates normally in fra germline cysts, and oocytes in mutant cysts are appropriately oriented at the posterior end of the egg chamber (Fig. 4F,G). Thus, Fra does not appear to control germline polarity preceding the checkpoint at mid-oogenesis.
Although Fra is expressed robustly in follicle cells, it is not intrinsically required in the soma for germline cyst survival (Fig. 3C,D). We tested the possibility that Fra non-autonomously regulates apicobasal and lateral follicle cell polarity. Armadillo (Arm, β-catenin) is localized to the cell membrane of both follicle and germline cells, and is enriched at the follicle cell apical (Fig. 4E,H). Arm localization in follicle cells adjacent to fra germline cysts is indistinguishable from its localization in wild-type ovarioles (Fig. 4I). Similarly, Discs large (Dlg), which localizes to lateral domains of follicle cells (Goode and Perrimon, 1997), is normally distributed in egg chambers with fra mutant germline cysts (Fig. 4J). These results suggest that Fra does not regulate apicobasal or lateral follicle cell polarity non-autonomously. As expected, localization of both Arm and Dlg is unperturbed in fra mutant follicle cells (Fig. S1). Although we cannot exclude the possibility that Fra controls other aspects of egg chamber polarity, the grossly normal morphology of fra germline mutants prior to degeneration suggests that any effects Fra has on polarity are subtle. Overall, the degeneration in fra mutants does not appear to be due to an activation of known triggers of the mid-oogenesis checkpoint.
Fra prevents apoptosis to promote progression through the mid-oogenesis checkpoint
How does fra germline degeneration compare with degeneration induced by poor nutrition and abnormal egg chambers? In nutrient-dependent egg chamber degeneration, follicle cells upregulate Draper and engulf the germline following nurse cell nuclei condensation and fragmentation (Etchegaray et al., 2012). Similarly, the follicle cells in degenerating egg chambers from ovarioles with fra germline cysts upregulate Draper (Fig. S2A,B), indicating that engulfment signaling in the follicle cells is active in degenerating egg chambers. Thus, although known triggers of the mid-oogenesis checkpoint appear unaffected in fra mutants, downstream degeneration appears similar.
In contrast to Dcc, which promotes cell death in the absence of Netrin (Chen et al., 1999; Forcet et al., 2001; Mehlen et al., 1998), loss of fra from the germline results in egg chamber degeneration, suggesting that Fra has a pro-survival function. The cell death effector caspase Dcp-1 is required at mid-oogenesis for germline cell death in response to checkpoint activation (Peterson et al., 2003). We hypothesized that loss of germline Fra results in activated Dcp-1 expression and leads to apoptosis. Although control cysts rarely express activated Dcp-1, we often detect it in late-stage fra germline mutant egg chambers and degenerating egg chambers in ovarioles with fra germline cysts (Fig. 5A,B). Furthermore, germline-specific expression of the baculovirus caspase inhibitor p35 (nanos-Gal4>UASp-p35) rescues the degeneration phenotype in ovarioles with fra germline clones (Fig. 5C,D), creating egg chambers with a persistent germline and missing follicle cells (‘balding’ egg chambers, Fig. 5C,E). Based on Dcp-1 staining and pyknotic nuclei, the follicle cells appear to be dying (not shown). This is consistent with previous reports describing the effect of inhibiting caspases in the germline of starved flies (Mazzalupo and Cooley, 2006). Thus, in contrast to the role of Dcc as a ‘dependence receptor’, Fra has an anti-apoptotic role in the Drosophila female germline.
Fra is required in the germline to prevent apoptosis. (A) Ovariole with fra3 germline clones (GFP−, white asterisks) stained for GFP (green) and activated Dcp-1 (gray). Ovariole contains a degenerating egg chamber with Dcp-1 expression (white arrowhead). (A′) Dcp-1 channel from A. (B) Percentage of ovarioles with a Dcp-1-positive egg chamber. n=22 and 33 ovarioles from a single experiment. (C) Ovariole with fra3 germline cysts (GFP−, white asterisks) and germline-specific p35 to inhibit caspases stained for GFP (green) and DRAQ5 (gray). Arrow indicates follicle cell death. (D,E) Percentage of ovarioles containing germline clones with germline degeneration (D) and balding egg chambers (E) when caspases are inhibited in fra3 germline cysts. For both graphs, n=226, 164 and 192 ovarioles scored across three independent trials. Error bars represent s.d. Statistical significance determined by Fisher's exact test, ****P<0.0001. (F) Schematic depicting Diap1 expression in a wild-type ovariole. (G,G′) Ovariole with fra3 germline cysts (GFP−, white asterisks), wild-type germline cysts (GFP+) stained for GFP (G, green) and Diap1 (G and G′, gray). Diap1 is not prematurely downregulated in fra3 germline clones. Image is representative of 44 ovarioles. Scale bars: 20 μm.
Fra is required in the germline to prevent apoptosis. (A) Ovariole with fra3 germline clones (GFP−, white asterisks) stained for GFP (green) and activated Dcp-1 (gray). Ovariole contains a degenerating egg chamber with Dcp-1 expression (white arrowhead). (A′) Dcp-1 channel from A. (B) Percentage of ovarioles with a Dcp-1-positive egg chamber. n=22 and 33 ovarioles from a single experiment. (C) Ovariole with fra3 germline cysts (GFP−, white asterisks) and germline-specific p35 to inhibit caspases stained for GFP (green) and DRAQ5 (gray). Arrow indicates follicle cell death. (D,E) Percentage of ovarioles containing germline clones with germline degeneration (D) and balding egg chambers (E) when caspases are inhibited in fra3 germline cysts. For both graphs, n=226, 164 and 192 ovarioles scored across three independent trials. Error bars represent s.d. Statistical significance determined by Fisher's exact test, ****P<0.0001. (F) Schematic depicting Diap1 expression in a wild-type ovariole. (G,G′) Ovariole with fra3 germline cysts (GFP−, white asterisks), wild-type germline cysts (GFP+) stained for GFP (G, green) and Diap1 (G and G′, gray). Diap1 is not prematurely downregulated in fra3 germline clones. Image is representative of 44 ovarioles. Scale bars: 20 μm.
Because Fra prevents germline apoptosis, we explored the possibility that it regulates cell death more directly. Dcp-1 is inhibited by Drosophila Inhibitor of apoptotic protein 1 (Diap1) (Hawkins et al., 1999). Diap1 protein and mRNA are detected in egg chambers prior to mid-oogenesis; expression decreases at mid-oogenesis, then increases again after stage eight (Baum et al., 2007; Foley and Cooley, 1998) (Fig. 5F). Germline overexpression of Diap1 suppresses Dcp-1-induced germline cell death at mid-oogenesis (Peterson et al., 2003). Similarly, overexpressing Diap1 in the germline prevents degeneration caused by starvation (Baum et al., 2007; Mazzalupo and Cooley, 2006). We reasoned that if Fra were preventing Dcp-1 activation through its negative regulator Diap1, then fra germline cysts might have prematurely reduced Diap1 levels, causing increased degeneration at mid-oogenesis. We compared Diap1 expression in GFP-negative fra germline cysts to neighboring GFP-positive control germline cysts and detected no differences in Diap1 levels (Fig. 5G,H). Thus, if Fra interacts with cell death machinery, it does not do so by regulating Diap1 levels.
Fra acts independently of Netrin in the ovary to promote germline survival
In the Drosophila nervous system, Fra has Netrin-independent and -dependent signaling mechanisms (Boyer and Gupton, 2018; Neuhaus-Follini and Bashaw, 2015; Yang et al., 2009). As germline Fra is required during oogenesis, we asked whether Netrin is also present in the ovary. In the developing nervous system, the two Drosophila Netrin genes, netrinA and netrinB (NetA and NetB – FlyBase) have overlapping expression domains and can function interchangeably to control axon guidance (Harris et al., 1996; Mitchell et al., 1996). We first evaluated Netrin expression with the netrinA-MiMIC allele (Nagarkar-Jaiswal et al., 2015), which produces NetrinA-GFP from the endogenous locus. Consistent with a recent report (Tu et al., 2020), we detect NetrinA-GFP in a subset of escort cells, which are somatic cells in the germarium (Fig. 6A). We do not detect NetrinA outside the germarium. In contrast, flies expressing NetrinB-Myc from its endogenous locus (Brankatschk and Dickson, 2006) have Myc signal throughout the ovariole (Fig. 6B,B′). Thus, although NetrinA is unlikely to signal through Fra in the vitellarium, the NetrinB expression pattern is consistent with such a role.
Netrin is expressed in the ovariole but is not required for egg chambers to progress through mid-oogenesis. (A) NetrinA-MiMIC germarium stained for GFP (NetrinA-GFP, green) and 1B1/LamC (cell membranes and cap cell nuclear envelopes, magenta). (A′) GFP channel from A. (B) NetrinBMyc ovariole stained for Myc (green) and 1B1 (magenta). (B′) Myc channel from A. (C) Ovariole from a netrinABΔGN escaper stained with 1B1 (cell membranes; magenta) and DRAQ5 (nuclei, green). (D) Percentage of netrinABΔGN ovarioles containing a degenerating egg chamber compared with sibling heterozygotes. n=280 and 193 ovarioles from two independent trials. Error bars represent s.d. Scale bars: 20 μm.
Netrin is expressed in the ovariole but is not required for egg chambers to progress through mid-oogenesis. (A) NetrinA-MiMIC germarium stained for GFP (NetrinA-GFP, green) and 1B1/LamC (cell membranes and cap cell nuclear envelopes, magenta). (A′) GFP channel from A. (B) NetrinBMyc ovariole stained for Myc (green) and 1B1 (magenta). (B′) Myc channel from A. (C) Ovariole from a netrinABΔGN escaper stained with 1B1 (cell membranes; magenta) and DRAQ5 (nuclei, green). (D) Percentage of netrinABΔGN ovarioles containing a degenerating egg chamber compared with sibling heterozygotes. n=280 and 193 ovarioles from two independent trials. Error bars represent s.d. Scale bars: 20 μm.
The transcriptional activation domain of Fra is required for egg chamber progression through mid-oogenesis. (A-C) Ovarioles with fra3 germline clones (GFP−, white asterisks) and (A) HAFraE1354A, (B) HAFraE1354A-VP16 or (C) FraICDMYC driven by nanos-GAL4. Ovarioles stained for GFP (green), HA or MYC (tagged transgenes, magenta), and DRAQ5 (nuclei, white). Arrowheads indicate degenerating egg chambers. (A′-C′) HA (A′,B′) or MYC (C′) from A-C. (D,E) Graphs showing the percentage of ovarioles containing a degenerating egg chamber in ovarioles containing fra3 mutant germline cysts. (D) HAFraE1354A and FraICDMYC are unable to rescue egg chamber degeneration. The first three genotypes of this graph also appear in Fig. 3F, as these were tested simultaneously and use the same controls. n=208, 260, 116, 123 and 80 ovarioles across at least three trials for each genotype. (E) Degeneration in ovarioles containing fra3 germline clones is rescued by HAFraE1354A-VP16. n=326, 184 and 224 ovarioles across three independent trails. Statistical significance determined by Fisher's exact test and P-values were adjusted by Bonferroni-Dunn for multiple comparisons, ***P=0.0003, ****P<0.0001. Error bars represent s.d. Scale bars: 20 μm.
The transcriptional activation domain of Fra is required for egg chamber progression through mid-oogenesis. (A-C) Ovarioles with fra3 germline clones (GFP−, white asterisks) and (A) HAFraE1354A, (B) HAFraE1354A-VP16 or (C) FraICDMYC driven by nanos-GAL4. Ovarioles stained for GFP (green), HA or MYC (tagged transgenes, magenta), and DRAQ5 (nuclei, white). Arrowheads indicate degenerating egg chambers. (A′-C′) HA (A′,B′) or MYC (C′) from A-C. (D,E) Graphs showing the percentage of ovarioles containing a degenerating egg chamber in ovarioles containing fra3 mutant germline cysts. (D) HAFraE1354A and FraICDMYC are unable to rescue egg chamber degeneration. The first three genotypes of this graph also appear in Fig. 3F, as these were tested simultaneously and use the same controls. n=208, 260, 116, 123 and 80 ovarioles across at least three trials for each genotype. (E) Degeneration in ovarioles containing fra3 germline clones is rescued by HAFraE1354A-VP16. n=326, 184 and 224 ovarioles across three independent trails. Statistical significance determined by Fisher's exact test and P-values were adjusted by Bonferroni-Dunn for multiple comparisons, ***P=0.0003, ****P<0.0001. Error bars represent s.d. Scale bars: 20 μm.
To determine whether Fra function in the ovary is dependent on Netrin, we tested whether Netrin is required in the ovary. Female flies homozygous for a small deletion removing both netrinA and netrinB (netrinABΔGN; Brankatschk and Dickson, 2006; Newquist et al., 2013) survive to adulthood at low frequency, and we used these ‘escapers’ to examine the effect of global Netrin removal on oogenesis. A previous study found that netrinABΔGN escaper females lay fewer eggs than control flies; however, no defects were observed in ovary morphology (Newquist et al., 2013). Consistent with these results, ovarioles from netrinABΔGN escapers appear morphologically indistinguishable from control ovarioles (Fig. 6C). Specifically, egg chambers progress through mid-oogenesis normally, and sibling heterozygote controls and netrinABΔGN mutants have similar rates of egg chamber degeneration (Fig. 6D). Thus, Netrin is dispensable for progression of egg chambers through mid-oogenesis, and the role of Fra in this process must be Netrin independent.
The transcriptional activation domain of Fra is required in the germline for egg chamber survival
Fra has a Netrin-independent function in the embryonic nerve cord, where it activates transcription to regulate axon guidance (Neuhaus-Follini and Bashaw, 2015; Yang et al., 2009). To activate transcription, Fra must be proteolytically processed by γ secretase, which releases the Fra ICD and allows it to enter the nucleus (Neuhaus-Follini and Bashaw, 2015). Once there, the Fra ICD activates transcription of comm (Neuhaus-Follini and Bashaw, 2015), the protein product of which downregulates the expression of the repulsive guidance receptor Robo1 (Keleman et al., 2005, 2002).
As Fra functions independently of Netrin in the ovary, we considered the possibility that Fra regulates transcription in this context. Previously, the Fra ICD (UAS-FraICDMyc) and a transgene with a point mutation that inactivates the transcriptional activation domain of Fra (UAS-HAFraE1354A) were used to rescue fra mutant phenotypes in the embryonic nervous system (Neuhaus-Follini and Bashaw, 2015). Although neuronal expression of UAS-FraICDMyc rescued the transcriptional regulation of comm by Fra, UAS-HAFraE1354A did not, demonstrating that the activation domain of Fra is required in that context. To test the possibility of a similar mechanism operating in the germline, we cloned the FraICDMyc and HAFraE1354A constructs into the germline-optimized pUASp vector (Rørth, 1998). Importantly, all transgenes were inserted at the same location, and none changed the rate of degeneration at mid-oogenesis when overexpressed in the germline of wild-type flies (Table 2). We then tested the ability of each Fra variant to rescue degeneration in ovarioles containing fra mutant germline cysts and compared the level of rescue to that of the full-length Fra receptor (Fig. 3). Although the full-length Fra receptor is able to rescue degeneration in ovarioles with fra mutant germline cysts, UASp-FraE1354A fails to rescue this degeneration (Fig. 7A,D). As this E1354A point mutation disrupts the transcriptional activation domain in Fra without affecting the nuclear export signal or Netrin-dependent Fra signaling (Neuhaus-Follini and Bashaw, 2015), we hypothesized that the transcriptional activation domain of Fra is specifically required for its function in the ovary. Indeed, a version of this transgene with an added VP16 activation domain (UASp-FraE1354A-VP16; Neuhaus-Follini and Bashaw, 2015) rescues degeneration in ovarioles with fra germline clones, consistent with the model that the transcriptional activation domain of Fra is necessary for its anti-apoptotic role in the ovary (Fig. 7B,E). Surprisingly, the Fra ICD alone fails to rescue degeneration in fra germline clones (Fig. 7C,D) despite rescuing the transcriptional function of Fra in the nerve cord (Neuhaus-Follini and Bashaw, 2015). One possibility is that the full-length receptor contains interaction domains that are necessary for Fra function in the germline but dispensable in the nervous system. Alternatively, the levels of ICD expression achieved using the nanos-GAL4 element may not be sufficient to rescue the germline phenotype. Nevertheless, the failure of UASp-FraE1354A to rescue degeneration in ovarioles with fra mutant germline clones and the rescue provided by UASp-FraE1354A-VP16 both suggest that the transcriptional activation domain in Fra is required to promote germline survival.
Although Fra most likely activates the transcription of multiple genes, comm is its only known target (Neuhaus-Follini and Bashaw, 2015). To determine whether comm is expressed in the ovary, we conducted reverse transcription polymerase chain reaction (RT-PCR) using two sets of primers directed against comm cDNA on mRNA extracted from both Drosophila ovaries and embryos. Although we detected comm mRNA in the ovary, this method cannot resolve its expression pattern (Fig. S3A). To identify comm-expressing cells, we used small molecule fluorescence in situ hybridization (smFISH) (Little and Gregor, 2018). Unexpectedly, we did not detect endogenous comm mRNA (Fig. S3B). A positive control, where we induce transgenic Comm expression in follicle cells using traffic jam-GAL4, demonstrates that our probe can detect comm mRNA (Fig. S3C), indicating that comm is either not expressed in the ovary or is expressed at levels below our threshold of detection. Indeed, a recently published RNA-seq study detected comm mRNA at low levels in certain follicle cells in the ovary but did not detect germline Comm (Jevitt et al., 2020). Although these observations do not support a germline role for Comm, it remains possible that low-level expression is functionally relevant for oogenesis. To further evaluate potential expression and function of comm in the germline, we used two approaches that have revealed functional connections between fra and comm during axon guidance. First, we tested whether mis-expression of Fra could induce comm transcription. Overexpressing Fra in the germline is unable to induce comm expression (Fig. S3D). Moreover, when we compared fra/+; comm/+ female flies with sibling controls, we observed no increase in egg chamber degeneration. Taken together, our observations indicate that Fra is unlikely to be regulating comm in the ovary. Nevertheless, the clear requirement for the transcriptional activation domain of Fra to promote germline survival suggests that Fra is regulating the transcription of key target genes in the germline.
DISCUSSION
Here, we explore the role of Fra in the ovary and demonstrate that Fra intrinsically promotes germline survival independently of Netrin. Fra localized to the cell membrane of nurse cells, oocytes and follicle cells. Loss of germline, but not follicle cell, fra leads to egg chamber degeneration at mid-oogenesis. Degeneration at mid-oogenesis is often caused by starvation or disruptions in egg chamber polarity (Beachum et al., 2021; Drummond-Barbosa and Spradling, 2001; Tanentzapf et al., 2000), and fra germline degeneration shares the morphological hallmarks of these pathways. However, fra mutants have wild-type starvation-induced degeneration response and FoxO localization, suggesting that Fra is not involved in nutrient sensing. Furthermore, the normal localization of Orb, Armadillo and Discs large in egg chambers with fra mutant germline cysts indicates that loss of fra is unlikely to affect germline or apicobasal/lateral follicle cell polarity. In fra mutant germlines, Dcp-1 expression is elevated, suggesting that the degeneration observed at mid-oogenesis is triggered by the activation of apoptosis. Accordingly, specifically inhibiting caspases with the UASp-p35 transgene results in a robust rescue of the fra mutant degeneration phenotype. Ovarioles from netrinABΔGN mutants appear morphologically normal and do not degenerate at mid-oogenesis, indicating that Fra functions independently of its canonical ligand Netrin at the mid-oogenesis checkpoint. Consistent with a Netrin-independent role for Fra here, rescue experiments demonstrate that the transcriptional activation domain of Fra is required for germline cyst survival at mid-oogenesis. Together, our results demonstrate that Fra is required in the germline, independently of Netrin, to promote egg chamber progression through mid-oogenesis. We have established the ovary as a novel tissue context in which to further investigate the Netrin-independent activity of Fra.
Fra functions independently of known regulators of the mid-oogenesis checkpoint
At mid-oogenesis, both external and internal factors can activate a checkpoint that leads to cell death. As vitellogenesis requires significant energy input, this checkpoint may prevent a costly investment in a low-quality oocyte (Drummond-Barbosa and Spradling, 2001; Mazzalupo and Cooley, 2006; Pritchett et al., 2009). Starvation (Buszczak et al., 2002; Drummond-Barbosa and Spradling, 2001; Terashima and Bownes, 2006) and disruption to egg chamber polarity (Beachum et al., 2021; Tanentzapf et al., 2000) trigger the mid-oogenesis checkpoint. When wild-type flies are starved or fed a protein-poor diet, degeneration at mid-oogenesis (also described as a ‘block to vitellogenesis’) increases dramatically (Drummond-Barbosa and Spradling, 2001). Although nutrient-dependent degeneration is morphologically similar to the fra mutant germline phenotype, FoxO localization indicates that insulin signaling is functioning in these cells. In addition, when flies with fra mutant germlines are starved, egg chamber degeneration at mid-oogenesis increases compared with well-fed counterparts. This further increase in degeneration suggests that fra germline cysts are still competent to respond to dietary signals. Taken together, these observations indicate that Fra is unlikely to be involved in the ovarian response to diet.
Disruptions in egg chamber polarity can also increase degeneration at mid-oogenesis (Beachum et al., 2021; Tanentzapf et al., 2000). However, Orb, Armadillo and Discs large localization in and adjacent to fra mutant clones indicates that neither germline polarity nor apicobasal or lateral follicle cell polarity are controlled by Fra. It remains possible that other aspects of egg chamber polarity are affected in fra germline cysts. Based on the absence of diet-related phenotypes and the normal morphology of egg chambers prior to degeneration, it is unclear why fra mutant germline cysts undergo apoptosis. A better understanding of downstream Fra signaling in the ovary will give insights into the functions of Fra here. Furthermore, Netrin is expressed in the germarium, and appears to regulate germline stem cell maintenance (Tu et al., 2020). It would be interesting to see whether Fra is also required for this process and functions with Netrin in the germarium.
Fra is anti-apoptotic in the ovary
In several contexts, Dcc has been shown to act as a ‘dependence receptor’ (Castets et al., 2012; Furne et al., 2008; Jasmin et al., 2021). Limiting Netrin, either in vitro by its absence in the serum, or in vivo through conditional knockouts, prevents Netrin from interacting with Dcc. This ultimately leads to caspase-mediated cell death in many contexts, including the nervous system and in both human embryonic kidney and cancer cell lines (Forcet et al., 2001; Goldschneider and Mehlen, 2010; Mehlen et al., 1998; Mehlen and Mazelin, 2003). Whether this function is limited to select cells, and whether the homolog of Dcc in other organisms can also act in a similar way, is unknown.
In contrast to the pro-apoptotic role of Dcc in some tissues, Fra has an anti-apoptotic role in the Drosophila ovarian germline. Removing Netrin has no effect on egg chamber degeneration. However, Fra loss causes an increase in egg chamber degeneration and a concomitant increase in ovarioles with Dcp-1-positive egg chambers. Future studies should address whether Fra is a substrate for caspase cleavage and how Fra/Dcc can have both pro- and anti-apoptotic activity. Indeed, it is unclear whether the mechanism through which Fra prevents apoptosis in this context bears any similarity to that in which Dcc engages the caspase signaling pathway to promote cell death in vertebrate systems. Interestingly, although the precise Caspase3 cleavage site in Dcc is not conserved in Fra, Fra ICD cleavage generates multiple fragments that are similar in size to Dcc ICD fragments (Neuhaus-Follini and Bashaw, 2015; Taniguchi et al., 2003). One intriguing possibility is that the Fra ICD may interact directly with Dcp-1 to prevent its activation.
Netrin-independent Fra transcriptional regulation
Netrin is required for fecundity in Drosophila, and global netrinABΔGN mutants lay fewer eggs than wild-type controls (Newquist et al., 2013). We observe no defects in the morphology of netrinABΔGN mutant ovaries, and, in contrast to flies with fra germline clones, we do not observe changes in egg chamber survival. Global removal of Netrin is likely to affect multiple tissues in adult flies, including the nervous system, and reproduction is sensitive to organismal physiology (Laws and Drummond-Barbosa, 2017). Nevertheless, the absence of egg chamber degeneration in global netrinABΔGN mutants indicates that Fra acts independently of Netrin to promote germline survival.
We have previously shown that, in addition to its Netrin-dependent role in axon guidance, Fra signals independently of Netrin in the nerve cord to transcriptionally activate comm (Neuhaus-Follini and Bashaw, 2015; Yang et al., 2009). Consistent with this Netrin-independent mode of signaling, the transcriptional activation domain of Fra is required for egg chamber progression through mid-oogenesis. Unlike the embryonic nervous system, where the Fra ICD partially rescues fra mutant phenotypes (Neuhaus-Follini and Bashaw, 2015), expression of the Fra ICD in the germline fails to rescue the fra mutant oogenesis phenotype. This difference may reflect different requirements for Fra in these two tissue contexts. One possibility is that germline Fra binds a co-activator at the cell membrane, facilitating its transport to the nucleus following γ-secretase cleavage. Alternatively, the failure to rescue may reflect a technical limitation due to insufficient expression levels of the Fra ICD in these experiments. In the nervous system, gain-of-function effects of ICD expression are dose dependent, and multiple copies of the transgene are required to generate robust phenotypes. Transcriptional signaling requires the Fra ICD to translocate to the nucleus (Neuhaus-Follini and Bashaw, 2015). Although we are unable to detect the Fra ICD in the nucleus, this does not eliminate the possibility that it enters the nucleus to regulate transcription. Indeed, in the nerve cord, nuclear Fra ICD is detected only occasionally when UAS-FraICDMyc is overexpressed in neurons, and is detected more often when the nuclear export signal of the Fra ICD is also removed (Neuhaus-Follini and Bashaw, 2015).
Currently, the only known transcriptional target of Fra is comm (Neuhaus-Follini and Bashaw, 2015; Yang et al., 2009), and we do not detect Comm in the germline. Furthermore, germline-specific expression of Fra using nanos-GAL4 does not induce comm mRNA expression, suggesting that comm is not a transcriptional target of Fra in these cells. As the transcriptional activation domain is required for Fra to promote germline survival, this indicates that Fra has other transcriptional targets that are necessary for preventing apoptosis in the germline. Indeed, the Neogenin ICD binds upstream of several genes and regulates their transcription in vitro in human embryonic kidney cells (Goldschneider et al., 2008). Future studies should determine other transcriptional targets of Fra.
Our results establish the ovary as a second in vivo tissue context where Fra regulates transcription. In the nervous system, Fra functions both via cytoskeletal modifications and transcriptional regulation (Neuhaus-Follini and Bashaw, 2015; Yang et al., 2009; Zang et al., 2021). However, teasing apart the different functions of Fra is challenging: both occur in the same cells and depend on the conserved P3 motif within the Fra ICD (Garbe et al., 2007; Neuhaus-Follini and Bashaw, 2015). Our work in the ovary provides a complementary tractable system to specifically study how Fra regulates transcription and to identify the upstream and downstream components involved in this signaling pathway.
MATERIALS AND METHODS
Fly stocks
Fly lines used in this study were: w*; P{FRT(whs)}G13 fra4/CyO, P{lacZ.w+}276 [Bloomington Drosophila Stock Center (BDSC) #8743], w*; P{FRT(whs)}G13 fra3/CyO, P{lacZ.w+}276 (BDSC #8813), w[*]; P{w[+mW.hs]=FRT(w[hs])}G13 P{w[+mC]=ovoD1-18}2R/T(1;2)OR64/CyO (BDSC #4434), y1 w67c23; Mi{PT-GFSTF.1}fraMI06684-GFSTF.1 (BDSC #59835), NetABΔGN/FM7 (provided by Thomas Kidd, University of Nevada, Reno, USA), w[*]; Bac{w[+mW.hs]=GreenEye.nosGAL4}Dmel6 (BDSC #32180), hsFLP1; FRT42D Ubi-GFP/CyO (provided by Elizabeth Ables, East Carolina University, USA), y1 w* Mi{PT-GFSTF.1}NetAMI04563-GFSTF.1/FM7j (BDSC #59409), NetB-MYC (Brankatschk and Dickson, 2006), and FRT42B (FRT G13) fra6, FRT42D fra3, hsFlp1; FRT42D UbiGFP; nos-GAL4 and UASp-p35 (provided by Andreas Bergmann, University of Massachusetts Chan Medical School, USA). Transgenic fly lines generated and used in this study were: UASp-HA-Fra, UASp-HA-FraE1354A, UASp-FraICD-MYC and UASp-HA-FraE1354A-VP16.
Generation of UASp Fra transgenes for germline expression
To generate UASp-FraICD-MYC, FraICD-MYC was amplified from UASt-FraICD-MYC by PCR and subcloned into pUASp-attB (DGRC #1358). To generate UASp-HA-Fra, HA-Fra was amplified from UASt-HA-Fra by PCR and subcloned into UASp-attB. To generate UASp-HA-FraE1354A, the C-terminal end of Fra was cut from the UASp-HA-Fra plasmid using XbaI, and the C-terminal end of HA-FraE1354A (containing the E1354A mutation) was cut from the UASt-HA-FraE1354A plasmid by XbaI and inserted into the cut UASp-HA-Fra plasmid. To generate UASp-HA-FraE1354A-VP16, the construct was cut from UASt-HA-FraE1354A-VP16 using NotI and inserted into UASp-attB. Constructs were verified by sequencing at the Penn Genomics Core. Transgenic flies were generated by phiC31 targeted insertion into the 86F8 site by BestGene.
Immunostaining and imaging
Ovaries were processed as described previously (Laws and Drummond-Barbosa, 2015) with minor modifications. Briefly, ovaries were dissected in ice-cold phosphate-buffered saline (PBS, Roche) and teased apart, then fixed for 13-15 min in 5.3% PFA in PBS (Electron Microscopy Services). The fix was washed off with 0.01% Triton-X100 in PBS (PBT), and ovaries were blocked overnight in PBT with 5% bovine serum albumin (w/v) and 5% normal goat serum. Primary antibodies were diluted in blocking solution and incubated with samples overnight at 4°C with rocking. Antibodies were washed off with PBT, then samples were stained with secondary antibodies and/or stains diluted in block for at least 1 h at room temperature with rocking. After washing, samples were cleared in 90% glycerol with antifade (20 µg/ml propyl gallate) overnight, then mounted onto slides. Ovaries were analyzed on a Nikon Ti-U inverted microscope with a Nikon OFN25 40× objective and imaged on a PerkinElmer spinning disk confocal system with a Hamamatsu C10600-10B CCD camera and Yokogawa CSU-10 scanner head with Volocity imaging software. Images were tiled using a FIJI pairwise stitching macro (Preibisch et al., 2009) and equally and minimally adjusted using FIJI and Adobe Illustrator.
Antibodies and stains
Primary antibodies used were: chick anti-GFP (1:1000, Abcam #13970), mouse anti-1B1 (1:100, DSHB concentrate), mouse anti-LamC (1:20, DSHB supernatant), rabbit anti-Fra (1:100, provided by Michael Murray, University of Melbourne, Australia), mouse anti-MYC (1:250, DSHB #9E10 concentrate), mouse anti-HA (1:250, Biolegend #901502), rabbit anti-dFoxO (1:500, provided by Pierre Leopold, Institut Curie, France), rabbit anti-cleaved Dcp-1 (1:50, Cell Signaling #9578S), mouse anti-Diap1 (1:100, provided by Bruce Hay, California Institute of Technology, USA) and mouse anti-Draper (1:20, DHSB 5D14 supernatant).
Secondary antibodies used (all at 1:200) were: goat anti-chick 488 (Invitrogen #A11039), goat anti-mouse 488 (Invitrogen #A11209), goat anti-mouse CY3 (Jackson Immuno #115-165-003), and Goat anti-rabbit CY3 (Jackson Immuno #111-165-144). DRAQ5 (1:1000 Cell Signaling #40845) was included with secondary antibodies.
Genetic mosaic analysis
ovoD clones were generated by heat shocking late second/early third instar larvae (hsFlp1; FRT42B/ovoD, FRT42B, hsFlp1; fra3, FRT42B/ovoD, FRT42B, hsFlp1; fra4, FRT42B/ovoD, FRT42B, hsFlp1; fra6, FRT42B/ovoD, FRT42B) in vials for 1 h in a 37°C water bath. Female flies 0-3 days old were collected and cultured with healthy males in vials with yeast paste for 2 days prior to dissection.
Negatively marked clones were generated by heat shocking progeny ([hsFlp1; FRT42D UbiGFP/FRT42D*], [hsFlp1; FRT42D UbiGFP/FRT42D*; UASp-HA-Fra/nos-GAL4], [hsFlp1; FRT42D UbiGFP/FRT42D*; UASp-HA-FraE1354A/nos-GAL4], [hsFlp1; FRT42D UbiGFP/FRT42D*; UASp-FraICD-Myc/nosGAL4], [hsFlp1; FRT42D UbiGFP/FRT42D*; UASp-HA-FraE1354A-VP16/nosGAL4] and [hsFlp1; FRT42D UbiGFP/FRT42D fra3, UAS-p35; nosGAL4/+] where * is a wild-type or mutant allele of fra) in vials for 1 h in a 37°C water bath once each day for 3 consecutive days (approximately days 5, 6 and 7 after egg-laying). Female flies 0-3 days old were cultured with healthy males in vials with yeast paste. Flies were fed with fresh yeast paste every 1-2 days for 4 days prior to dissection.
Scoring degeneration or follicle cell death
Ovarioles were scored blind to genotype. Degeneration was scored by the presence or absence of pyknotic nuclei visualized by the nuclear stain DRAQ5. Follicle cell death was scored by the absence of follicle cells surrounding nurse cell nuclei that are not condensed.
Diet experiment
Female flies with fra3 germline clones (see ovoD clone generation in ‘Genetic mosaic analysis’ section) were collected at 0-3 days old and placed on wet yeast paste with healthy males. Half of the flies were placed in a vial with a wet Kimwipe and no food for 6 h prior to dissection.
RT-PCR
Approximately 25 female flies were fed yeast paste for 3 days prior to dissection in RNase-free PBS and put on ice. w1118 fly embryos were collected from apple juice plates after adding 50% bleach for 3 min and washing with distilled water. RNA was extracted from both ovaries and embryos using the Qiagen RNeasy mini kit protocol. Qiagen One-step RT-PCR kit protocol was used for RT-PCR. Two sets of primers were used to detect comm mRNA: set 1 FWD, CTCTCCAAGTCGGTGGTTCT; set 1 REV, TTCATGCCGTAGGCAAAGTG; Set 2 FWD ATCTGTGGATCGGAGTGGTC, REV TTATTCAGCGGCTCCTGCTT.
Acknowledgements
We thank members of the Bashaw lab, Lauren Anllo and anonymous reviewers for their feedback and comments on the manuscript. We thank Thomas Kidd, Elizabeth Ables, Michael Murray, Pierre Leopold, Andreas Bergmann and Bruce Hay for sending us fly lines or antibodies, as well as BDSC and FlyBase. We also thank Kevin Wiles for assistance in interpreting publicly available RNA-seq data.
Footnotes
Author contributions
Conceptualization: S.A.R., K.M.L., G.J.B.; Methodology: S.A.R., K.M.L., G.J.B.; Formal analysis: S.A.R.; Investigation: S.A.R., K.M.L.; Writing - original draft: S.A.R.; Writing - review & editing: S.A.R., K.M.L., G.J.B.; Visualization: S.A.R., K.M.L.; Supervision: K.M.L., G.J.B.; Project administration: G.J.B.; Funding acquisition: G.J.B.
Funding
This research was supported by the National Institutes of Health (T32 HD083185 to S.A.R., K12 GM081259 to K.M.L., and R35 NS097340 and NSF IOS-1853719 to G.J.B.). Deposited in PMC for release after 12 months.
Peer review history
The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.199762.
References
Competing interests
The authors declare no competing or financial interests.