ABSTRACT
Mitochondrial morphology dynamics regulate signaling pathways during epithelial cell formation and differentiation. The mitochondrial fission protein Drp1 affects the appropriate activation of EGFR and Notch signaling-driven differentiation of posterior follicle cells in Drosophila oogenesis. The mechanisms by which Drp1 regulates epithelial polarity during differentiation are not known. In this study, we show that Drp1-depleted follicle cells are constricted in early stages and present in multiple layers at later stages with decreased levels of apical polarity protein aPKC. These defects are suppressed by additional depletion of mitochondrial fusion protein Opa1. Opa1 depletion leads to mitochondrial fragmentation and increased reactive oxygen species (ROS) in follicle cells. We find that increasing ROS by depleting the ROS scavengers, mitochondrial SOD2 and catalase also leads to mitochondrial fragmentation. Further, the loss of Opa1, SOD2 and catalase partially restores the defects in epithelial polarity and aPKC, along with EGFR and Notch signaling in Drp1-depleted follicle cells. Our results show a crucial interaction between mitochondrial morphology, ROS generation and epithelial cell polarity formation during the differentiation of follicle epithelial cells in Drosophila oogenesis.
INTRODUCTION
Signaling pathways regulate the onset and remodeling of epithelial cell polarity in stem cell differentiation and embryogenesis. Epithelial cells have a polarized plasma membrane organization with an apical, lateral and basal domain containing different protein complexes. Epithelial cell differentiation accompanies the presence of elaborate and fused mitochondria in mammalian hepatocytes. A crucial role for mitochondrial fusion-derived OXPHOS-dependent energy has been seen in epithelial polarity establishment (Fu et al., 2013). Mammalian MDCK cells need an ATP-rich environment for the integrity of junctional complexes (Tsukamoto and Nigam, 1997, 1999). These studies show a role for mitochondrial morphology and associated metabolites in the form of ATP in regulating polarity formation and maintenance (Madan et al., 2021). However, the mechanisms by which mitochondrial morphology dynamics and function regulate polarity complexes during stem cell differentiation remain to be investigated.
An immature apical domain is present at the stem cell stage, compared with the differentiated stage, in certain epithelial cells in Drosophila and mammalian systems. Drosophila intestinal stem cells lack an apical domain. A mature apical domain is formed on the differentiation of these cells into enteroblasts and enterocytes. The inhibition of mitochondrial fusion protein Opa1 (Deng et al., 2018) and mitochondrial respiratory (ETC) complexes reduces intestinal stem cell differentiation into enterocytes (Zhang et al., 2020). Thus, epithelial cell differentiation is influenced by mitochondrial morphology and function.
The Drosophila follicle stem cell (FSC) is another example of a partially polarized epithelial cell that lacks an apical domain (Castanieto et al., 2014). In this study, we have analyzed the role of regulation of mitochondrial dynamics in apical polarity maintenance in late Drosophila follicle cell (FC) differentiation. FCs encase germ cell cysts during Drosophila oogenesis. The germline stem cells (GSCs) and FSCs are located in the germarium (Fadiga and Nystul, 2019). Each GSC undergoes four incomplete divisions and forms a 16-cell cyst. Each FSC divides and forms pre-follicle cells, which further divide to form FCs (Castanieto et al., 2014; Ulmschneider et al., 2016). FSCs lack an apical domain, whereas mature follicle epithelial cells possess apical, basolateral and basal domains, characterized by a domain-specific set of polarity proteins (Castanieto et al., 2014; Dobens and Raftery, 2000; Wu et al., 2008). The apical domain has the aPKC-PAR3-PAR6-PatJ complex and the basolateral domain has the Dlg-Scribble-Lgl complex. The apical and basolateral domains are separated by the adherens junction complex (Kaplan et al., 2009). FCs undergo mitosis in stages 1-6 and enter the endocycle from stage 6 following activation of the Notch signaling pathway (Klusza and Deng, 2011) (Fig. 1A). The mitotic stage FCs express the homeodomain transcription factor Cut (Sun and Deng, 2005). Transitioning from the mitotic stage to the endocycling stage requires the activation of the Notch signaling pathway and the expression of the transcription factor Hindsight (Hnt; also known as Pebbled) (Jia et al., 2015; Klusza and Deng, 2011; Shcherbata et al., 2004). An active EGFR signaling pathway inhibits the formation of the apical domain in FSCs, whereas newly formed pre-FCs show an apical domain due to the suppression of EGFR signaling (Castanieto et al., 2014). Activation of the EGFR signaling pathway also regulates differentiation of the posterior follicle cells (PFCs) at late stages. The stepwise polarity establishment in the follicle epithelial cells during Drosophila oogenesis is an insightful model for studying the mechanisms that regulate epithelial cell differentiation and polarity establishment (Franz and Riechmann, 2010; Tepass et al., 2001).
Multilayering in Drp1-depleted PFCs is reduced by additional depletion of Opa1. (A) Schematic showing Drosophila ovariole and stages of egg chambers. (B) Representative images of FC arrangement in control FRT40A (Ba), drp1KG (Bb), opa1i (Bc) and drp1KG;opa1i (Bd) FC clones at mitotic stage. (C) Representative images of FC arrangement in control FRT40A (Ca), drp1KG (Cb), opa1i (Cc), and drp1KG;opa1i (Cd) PFC clones. (D) Representative images showing multiple layers of PFC clones in drp1KG (66% clones have three layers or more, n=50) (Da) and drp1KG;opa1i (42% clones have three layers or more, n=33) (Db). (E) The graph shows a comparison of the normalized PFC clone height of control FRT40A (n=27), drp1KG (n=56), opa1miRNA (n=37), drp1KG;opa1miRNA (n=37). Data are mean±s.e.m. Each data point (n) represents a clone in a separate egg chamber. The statistical test performed is one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; **P<0.01; ***P<0.001. mCD8-GFP (green)-expressing FC clones are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Scale bars: 10 μm.
Multilayering in Drp1-depleted PFCs is reduced by additional depletion of Opa1. (A) Schematic showing Drosophila ovariole and stages of egg chambers. (B) Representative images of FC arrangement in control FRT40A (Ba), drp1KG (Bb), opa1i (Bc) and drp1KG;opa1i (Bd) FC clones at mitotic stage. (C) Representative images of FC arrangement in control FRT40A (Ca), drp1KG (Cb), opa1i (Cc), and drp1KG;opa1i (Cd) PFC clones. (D) Representative images showing multiple layers of PFC clones in drp1KG (66% clones have three layers or more, n=50) (Da) and drp1KG;opa1i (42% clones have three layers or more, n=33) (Db). (E) The graph shows a comparison of the normalized PFC clone height of control FRT40A (n=27), drp1KG (n=56), opa1miRNA (n=37), drp1KG;opa1miRNA (n=37). Data are mean±s.e.m. Each data point (n) represents a clone in a separate egg chamber. The statistical test performed is one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; **P<0.01; ***P<0.001. mCD8-GFP (green)-expressing FC clones are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Scale bars: 10 μm.
FCs show increased mitochondrial density during Drosophila oogenesis (Tourmente et al., 1990) and dispersed mitochondria at early stages (Cox and Spradling, 2009). FSCs are lost in mutants with mitochondrial dysfunction and in mutants which exhibit an increase in reactive oxygen species (ROS) (Wang et al., 2012). Blocking mitochondrial fission by depleting Drp1 leads to fused mitochondria, inhibiting the transition to the endocycling stage in Drosophila oogenesis (Mitra et al., 2012; Tomer et al., 2018). The Drp1-depleted PFCs show an accumulation of phosphorylated ERK downstream of EGFR signaling and loss of Notch signaling. They are present in multiple layers. In this study, we assess the mechanism by which mitochondrial morphology regulates epithelial cell polarity maintenance during Drosophila oogenesis. We find that Drp1 deficient FCs show depletion of apical polarity protein aPKC even before the formation of multilayers. This defect is suppressed by the fragmentation of mitochondria by additional depletion of mitochondrial fusion protein Opa1 in Drp1-depleted FCs. An increase in ROS by Opa1 depletion or inhibition of ROS scavenging enzymes leads to mitochondrial fragmentation and the formation of apical polarity in Drp1-depleted FCs. The increase in ROS also leads to suppression of the accumulation of cytoplasmic ERK and loss of Notch-mediated differentiation defect in Drp1-depleted FCs. Our study shows an interaction of mitochondrial dynamics in the regulation of the levels of ROS and follicle epithelial cell formation in Drosophila oogenesis.
RESULTS
Drp1-depleted PFCs are present in multiple layers in the endocycling stage, and this defect is partially suppressed by the depletion of Opa1
To analyze the arrangement and polarity of Drosophila epithelial FCs depleted of mitochondrial fission protein Drp1, we generated MARCM clones (see Materials and Methods) with the transposon tagged null mutant drp1KG03815 (drp1KG) (Mitra et al., 2012; Rikhy et al., 2007; Tomer et al., 2018). Drp1-depleted FCs are marked by the expression of mCD8-GFP. Ovaries containing FC clones depleted of Drp1 were stained for nuclear DNA and mitochondria. As expected, Drp1 depletion led to mitochondrial clustering in FCs (Fig. S1Aa,Ab,Ba,Bb), similar to the previous analysis in neuroblasts, neurons, FCs and spermatocytes (Aldridge et al., 2007; Dubal et al., 2022; Mitra et al., 2012; Tomer et al., 2018). Drp1-depleted FC clones were in a monolayer at the mitotic stage egg chambers (Fig. 1Ba,Bb). Our previous studies have shown that Drp1-depleted FC clones in the main body at the endocycling stage were present in monolayers, whereas PFC clones adjacent to the oocyte encountering the EGFR signaling were present in multiple layers (Mitra et al., 2012; Tomer et al., 2018). As expected, PFC clones surrounding the oocyte at the endocycling stage depleted of Drp1 were present in multiple layers (Fig. 1Ca,Cb). We have previously shown that the cells depleted of Drp1 are in the mitotic stage and do not transition to the endocycling stage. The nuclei in PFCs are smaller than the control nuclei owing to lack of entry into the endocycle (Mitra et al., 2012; Parker et al., 2015; Tomer et al., 2018). The number of cell layers per clone in the drp1KG-depleted PFCs varied from one to six, with the highest frequency of three layers (Fig. S1C).
Mitochondrial clustering in Drp1-depleted cells requires the activity of mitochondrial fusion proteins Opa1 and Marf (Dubal et al., 2022; Sessions et al., 2022). We depleted Opa1 using an shRNA to inhibit mitochondrial clustering in FCs containing the homozygous null drp1KG allele. The MARCM clone strategy was used to generate homozygous drp1KG FCs expressing Opa1 shRNA (opa1i). Mitochondria were fragmented in Opa1-depleted FCs (Fig. S1Ac,Bc). Even though the mitochondria were clustered in the drp1KG;opa1i combination, the cluster appeared to be decreased in compaction compared with drp1KG alone (Fig. S1Ad,Bd). FCs depleted of Opa1 were present in a single layer similar to controls in the mitotic and endocycling stages (Fig. 1Bc,Cc). The drp1KG;opa1i combination contained PFCs in multiple layers; however, the extent of multilayering was reduced compared with drp1KG (Fig. 1D). We quantified the height of the clone and clone area in different genotypes as a readout of the extent of multilayering (see Materials and Methods). Consistent with the decrease in multilayering, we found a significant reduction in the height of the clone (Fig. 1E) and clone area of drp1KG;opa1i PFCs (Fig. S1D) compared with drp1KG alone. We also estimated the extent of multilayering in FCs depleted of outer mitochondrial fusion protein Marf. We found that a similar alleviation of the extent of multilayering was also observed by combining a second independent RNAi against Opa1 (Fig. S2Ab,B,C) and Marf (Fig. S2Ac,C,D) in drp1KG homozygous FCs. In summary, Drp1-depleted PFCs were present in multiple layers at the endocycling stage, and additional depletion of Opa1 and Marf partially alleviated the extent of multilayering.
Drp1-depleted FCs show a decrease of apical polarity protein aPKC
Depletion of polarity proteins aPKC, Baz, Crumbs, Scribble, Dlg and Lgl leads to the distribution of FCs in multiple layers (Baum and Perrimon, 2001; Benton and St Johnston, 2003; Bergstralh et al., 2013; Bilder et al., 2000; Dent et al., 2019; Fletcher et al., 2012; Khoury and Bilder, 2020; Luo et al., 2016; Moreira et al., 2019; Romani et al., 2009; Sun and Deng, 2005; Ventura et al., 2020; Wang et al., 2021; Woolworth et al., 2009). The apical protein aPKC phosphorylates several other polarity proteins including Dlg, Bazooka, Crumbs and Lgl, and coordinates epithelial cell polarization (Betschinger et al., 2003; Golub et al., 2017; Morais-de-Sá et al., 2010; Plant et al., 2003; Sotillos et al., 2004). In addition, aPKC also maintains a single layer of FCs by inhibiting actomyosin contractility (Osswald et al., 2022). We assessed the distribution of apical polarity marker aPKC in drp1KG PFC clones in endocycling and FC clones in mitotic stages. The aPKC levels were reduced or lost from PFCs homozygous for the drp1KG null allele and located adjacent to the oocyte in the endocycling stage (Fig. 2Aa,Ab,B). In addition, when two drp1KG clones were present adjacent to each other with control tissue in the middle, aPKC decreased in these cells in a non-autonomous manner. At earlier mitotic stages, despite being arranged in a monolayer similar to controls, the drp1KG FCs also showed decreased aPKC (Fig. 2Ca,Cb,D). This aPKC decrease was suppressed in drp1KG;opa1i in the endocycling PFCs adjacent to the oocyte (Fig. 2Ad,B) and the mitotic stage FCs (Fig. 2Cd,D). This aPKC depletion was also suppressed in a second independent RNAi against Opa1 (Fig. S3Ad,B,Cd,D).
Drp1-depleted FC clones show reduction or loss of aPKC in endocycling and mitotic stages, and this defect is decreased in drp1KG;opa1i clones. (A) Representative images showing aPKC (red) in control FRT40A (Aa), drp1KG (Ab), opa1i (Ac) and drp1KG;opa1i (Ad) PFC clones. (B) The graph shows the percentage of cells per PFC clone adjacent to the oocyte with reduced aPKC, showing a higher percentage in drp1KG and drp1KG;opa1i. (C) Representative images showing aPKC (red) in control FRT40A (Ca), drp1KG (Cb), opa1i (Cc) and drp1KG;opa1i (Cd) in mitotic stage FCs. (D) The graph shows the percentage of cells per FC clone in mitotic stages with reduced aPKC, showing a higher percentage in drp1KG and drp1KG;opa1i. (E,F) The graphs show the normalized apical length of the FCs from the clone in control FRT40A (n=52), drp1KG (n=80), opa1i (n=24) and drp1KG;opa1i (n=21) at the endocycling stage (E) and in control FRT40A (n=51) and drp1KG (n=26), opa1i (n=20) and drp1KG;opa1i (n=23) at the mitotic stage (F). mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads show defective FCs with reduced aPKC. Asterisks mark FCs with aPKC in double mutants. Data are mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the percentage of cells with reduced aPKC in a single clone of a separate egg chamber. The values for the genotypes control FRT40A, drp1KG, opa1i and drp1KG;opa1i are given as n (the number of clones) and N (the number of independent replicates): n=11,11,22,21 and N=3,5,3,4, respectively, for endocycling stages (B) and n=17,17,21,15 and N=3,5,3,4, respectively, for mitotic stages (D). The statistical tests performed were one-way ANOVA with Tukey's multiple comparisons for analysis of aPKC and two-tailed unpaired Student's t-test for comparing normalized apical length and clone area. ns, non significant; *P<0.05; ***P<0.001. Scale bars: 10 μm.
Drp1-depleted FC clones show reduction or loss of aPKC in endocycling and mitotic stages, and this defect is decreased in drp1KG;opa1i clones. (A) Representative images showing aPKC (red) in control FRT40A (Aa), drp1KG (Ab), opa1i (Ac) and drp1KG;opa1i (Ad) PFC clones. (B) The graph shows the percentage of cells per PFC clone adjacent to the oocyte with reduced aPKC, showing a higher percentage in drp1KG and drp1KG;opa1i. (C) Representative images showing aPKC (red) in control FRT40A (Ca), drp1KG (Cb), opa1i (Cc) and drp1KG;opa1i (Cd) in mitotic stage FCs. (D) The graph shows the percentage of cells per FC clone in mitotic stages with reduced aPKC, showing a higher percentage in drp1KG and drp1KG;opa1i. (E,F) The graphs show the normalized apical length of the FCs from the clone in control FRT40A (n=52), drp1KG (n=80), opa1i (n=24) and drp1KG;opa1i (n=21) at the endocycling stage (E) and in control FRT40A (n=51) and drp1KG (n=26), opa1i (n=20) and drp1KG;opa1i (n=23) at the mitotic stage (F). mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads show defective FCs with reduced aPKC. Asterisks mark FCs with aPKC in double mutants. Data are mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the percentage of cells with reduced aPKC in a single clone of a separate egg chamber. The values for the genotypes control FRT40A, drp1KG, opa1i and drp1KG;opa1i are given as n (the number of clones) and N (the number of independent replicates): n=11,11,22,21 and N=3,5,3,4, respectively, for endocycling stages (B) and n=17,17,21,15 and N=3,5,3,4, respectively, for mitotic stages (D). The statistical tests performed were one-way ANOVA with Tukey's multiple comparisons for analysis of aPKC and two-tailed unpaired Student's t-test for comparing normalized apical length and clone area. ns, non significant; *P<0.05; ***P<0.001. Scale bars: 10 μm.
We assessed the lateral domain marker Dlg for its localization and spread into the apical domain in drp1KG FCs due to the depletion of aPKC (Schmidt and Peifer, 2020). Dlg remained in the lateral domain in PFCs adjacent to oocyte at the endocycling stage, and in FCs depleted of Drp1 in the mitotic stage (Fig. S4Aa,Ab,Ba,Bb). However, when we quantified the apical length from these images, we found that the Drp1-depleted FCs were constricted compared with controls in the endocycling and mitotic stages (Fig. 2E,F). We found that this defect of apical constriction was partially rescued in drp1KG;opa1i (Fig. 2E,F; Fig. S4A,B).
We also assessed the distribution of apical protein PatJ (Fig. S4C-F) and adherens junction protein DE-cadh (Fig. S4G-J) for their distribution. We found that they were depleted in both endocycling and mitotic stage FCs. As aPKC recruitment occurs in pre-FCs during the formation of the apical domain (Castanieto et al., 2014), and aPKC inhibits the contractility of the actomyosin cytoskeleton (Osswald et al., 2022), its depletion likely contributes to constriction in FCs in mitotic stages and occurrence of PFCs in multiple layers on Drp1 depletion.
Increased expression of aPKC has been achieved by expressing aPKC-ΔN in FCs, renal tubes and neuroblasts (Betschinger et al., 2003; Campbell et al., 2010; Drier et al., 2002; Khoury and Bilder, 2020). We expressed aPKC-ΔN in FCs homozygous for the null allele drp1KG and found an increase in cytoplasmic aPKC staining (Fig. 3A,C). We observed a rescue of the distribution of apical aPKC in drp1KG overexpressing aPKC-ΔN in both mitotic and endocycling stages (Fig. 3A-D). We further found that the phenotypes of apical constriction (Fig. 3E,F) and multilayering (Fig. 3G-I) were alleviated in drp1KG;aPKC-ΔN compared with drp1KG alone. Together, these data indicate that the decrease of aPKC leads to multi-layering of FCs in endocycling stages on Drp1 depletion.
Co-expression of aPKC-ΔN suppresses the defects seen in Drp1-depleted FCs. (A,C) Representative images showing aPKC (red) in control FRT40A, drp1KG, aPKC-ΔN and drp1KG;aPKC-ΔN PFC clones at the endocycling stage (Aa-Ad) and at the mitotic stage (Ca-Cd). Arrowheads mark FCs deficient in aPKC. Asterisks mark FCs that show the presence of aPKC in double mutants. (B,D) The graphs show the quantification of the percentage of cells with reduced aPKC in PFC clone adjacent to the oocyte in drp1KG (n=35) and drp1KG;aPKC-ΔN (n=4) (B) and the percentage of cells with reduced aPKC in each FC clone in drp1KG (n=48) and drp1KG;aPKC-ΔN (n=4) at mitotic stages (D). (E,F) The graphs show the normalized apical length of the PFCs from drp1KG (n=80) and drp1KG;aPKC-ΔN (n=23) at the endocycling stage (E) and from drp1KG (n=26) and drp1KG;aPKC-ΔN (n=19) at the mitotic stage (F). (G) Representative images showing multiple layers of PFC clones in drp1KG (66% clones have three layers or more, n=50) and drp1KG;aPKC-ΔN (55% clones have three layers or more, n=11). (H) The graph shows a comparison of the normalized PFC clone height of control FRT40A (n=27), drp1KG (n=56), aPKC-ΔN (n=12) and drp1KG;aPKC-ΔN (n=14). (I) The graph shows a comparison of the PFC clone area of drp1KG (n=38) and drp1KG;aPKC-ΔN (n=14). Data are mean±s.e.m. Each data point (n) represents a clone in a separate egg chamber except for the apical length (E,F) where each data point (n) represents a normalized apical length of the constricted individual cells of the clone from multiple egg chambers. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons for the analysis of normalized height and aPKC and two-tailed unpaired Student's t-test for the analysis of clone area and normalized apical length. ns, non-significant; **P<0.01; ***P<0.001. The control and drp1KG clone height data in 3H is repeated from Fig. 1E and drp1KG apical length data in 3E,F is repeated from Fig. 2E,F, respectively, for comparison. mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Scale bars: 10 μm.
Co-expression of aPKC-ΔN suppresses the defects seen in Drp1-depleted FCs. (A,C) Representative images showing aPKC (red) in control FRT40A, drp1KG, aPKC-ΔN and drp1KG;aPKC-ΔN PFC clones at the endocycling stage (Aa-Ad) and at the mitotic stage (Ca-Cd). Arrowheads mark FCs deficient in aPKC. Asterisks mark FCs that show the presence of aPKC in double mutants. (B,D) The graphs show the quantification of the percentage of cells with reduced aPKC in PFC clone adjacent to the oocyte in drp1KG (n=35) and drp1KG;aPKC-ΔN (n=4) (B) and the percentage of cells with reduced aPKC in each FC clone in drp1KG (n=48) and drp1KG;aPKC-ΔN (n=4) at mitotic stages (D). (E,F) The graphs show the normalized apical length of the PFCs from drp1KG (n=80) and drp1KG;aPKC-ΔN (n=23) at the endocycling stage (E) and from drp1KG (n=26) and drp1KG;aPKC-ΔN (n=19) at the mitotic stage (F). (G) Representative images showing multiple layers of PFC clones in drp1KG (66% clones have three layers or more, n=50) and drp1KG;aPKC-ΔN (55% clones have three layers or more, n=11). (H) The graph shows a comparison of the normalized PFC clone height of control FRT40A (n=27), drp1KG (n=56), aPKC-ΔN (n=12) and drp1KG;aPKC-ΔN (n=14). (I) The graph shows a comparison of the PFC clone area of drp1KG (n=38) and drp1KG;aPKC-ΔN (n=14). Data are mean±s.e.m. Each data point (n) represents a clone in a separate egg chamber except for the apical length (E,F) where each data point (n) represents a normalized apical length of the constricted individual cells of the clone from multiple egg chambers. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons for the analysis of normalized height and aPKC and two-tailed unpaired Student's t-test for the analysis of clone area and normalized apical length. ns, non-significant; **P<0.01; ***P<0.001. The control and drp1KG clone height data in 3H is repeated from Fig. 1E and drp1KG apical length data in 3E,F is repeated from Fig. 2E,F, respectively, for comparison. mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Scale bars: 10 μm.
ROS increase on depletion of Opa1 and ROS scavenging proteins leads to mitochondrial fragmentation
Opa1 depletion partially alleviated the aPKC decrease in Drp1-deficient FCs (Fig. 2). Loss of Opa1 function has been previously shown to increase mitochondrial ROS (mtROS), and overexpression of Opa1 along with the ATP synthase complex oligomerization reduced mtROS (Jang and Javadov, 2020; Quintana-Cabrera et al., 2021; Tang et al., 2009). The mtROS was elevated in the Opa1-depleted eye cells in Drosophila (Yarosh et al., 2008); however, ROS was also found to be decreased on Drp1 depletion in Drosophila embryos (Chowdhary et al., 2020). To detect mitochondrial superoxides, we stained living ovaries containing control or mutant clones with a fluorescent dye mitoSOX. Drp1-depleted FCs showed fluorescence in a cluster on one side (Fig. 4Aa,Ab). Depletion of Opa1 led to an increase in the mitoSOX fluorescence compared with their neighboring control FCs (Fig. 4Ac). ROS levels are regulated by antioxidant enzymes such as superoxide dismutase 1 (SOD1), SOD2, catalase and glutathione peroxidase. Therefore, we tested whether ROS increase was seen when we depleted mitochondrial SOD2 (Celotto et al., 2012; Kirby et al., 2002; Mukherjee et al., 2011) and catalase using shRNA expression. The fluorescence intensity of mitoSOX was higher in sod2i (Fig. 4Ad) and catalasei (Fig. 4Ae) FC clones compared with the neighboring cells.
Mitochondrial ROS and fragmentation is increased on depletion of ROS scavengers. (A) Representative images showing the fluorescence intensity in a rainbow scale (red pixels are higher intensity and blue pixels are lower intensity) of MitoSOX in control FRT40A, drp1KG, opa1i, sod2i and catalasei (Aa-Ae). (B) Mitochondria stained with ATP-β (red) are shown in representative images from control FRT40A (100% intermediate, n=21), drp1KG (92% clustered, n=26), sod2i (82% fragmented, n=28) and catalasei (79% fragmented, n=14) (Ba-Bd). mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with solid (A) or dotted (B) yellow outlines. Asterisks mark clustered mitochondria in surface view. Arrowheads mark the punctate mitochondria in surface view. The nucleus (blue) is stained with Hoechst. Scale bars: 10 μm.
Mitochondrial ROS and fragmentation is increased on depletion of ROS scavengers. (A) Representative images showing the fluorescence intensity in a rainbow scale (red pixels are higher intensity and blue pixels are lower intensity) of MitoSOX in control FRT40A, drp1KG, opa1i, sod2i and catalasei (Aa-Ae). (B) Mitochondria stained with ATP-β (red) are shown in representative images from control FRT40A (100% intermediate, n=21), drp1KG (92% clustered, n=26), sod2i (82% fragmented, n=28) and catalasei (79% fragmented, n=14) (Ba-Bd). mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with solid (A) or dotted (B) yellow outlines. Asterisks mark clustered mitochondria in surface view. Arrowheads mark the punctate mitochondria in surface view. The nucleus (blue) is stained with Hoechst. Scale bars: 10 μm.
SOD2 depletion affects mitochondrial dynamics, leading to fragmentation in delaminating cells during Drosophila dorsal closure (Muliyil and Narasimha, 2014). We therefore assessed mitochondrial morphology in sod2i and catalasei FC clones by immunostaining with an antibody against Complex V. Mitochondria were found to be punctate in appearance in FCs depleted of sod2i or catalasei (Fig. 4Ba-Bd) compared with neighboring control cells. We made combinations to obtain PFCs that are homozygous for the drp1KG null allele with sod2i and catalasei. The mitochondrial morphology in drp1KG;sod2i and in drp1KG;catalasei (Fig. S5Aa-Af) showed mitochondrial clustering, but the compaction was decreased compared with drp1KG (Fig. S5Ab,Bb). Mitochondria were typically distributed to one side in multilayered PFCs depleted of Drp1 (Fig. S5Bb). Mitochondria were present all around the nucleus, consistent with fragmentation in the drp1KG;sod2i (Fig. 5Bd) and in drp1KG;catalasei (Fig. 5Bf) combinations compared with drp1KG (Fig. S5Bb). We also assessed the change in ROS and mitochondrial distribution on depletion of aPKC using two independent RNAis in FCs by crossing with GR-Gal4 (Fig. S6A-D). We found that although there was a decrease in ROS in the FCs (Fig. S6E,F), there was no change in mitochondrial distribution (Fig. S6G). It is likely that the decrease in ROS is not sufficient to cause a change in mitochondrial morphology on aPKC depletion.
Drp1-depleted FC clones show reduction or loss of aPKC in endocycling and mitotic stages, and this is partially recovered on additional depletion of ROS scavengers. (A) Representative images showing aPKC (red) staining in PFC clones from control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei (Aa-Af). (B) The graph shows quantification of the percentage of cells with reduced aPKC in each PFC clone adjacent to the oocyte in control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei. (C) Representative images showing aPKC (red) staining in control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei FC clones at the mitotic stages (Ca-Cf). (D) The graph shows the quantification of the percentage of cells with reduced aPKC in each FC clone in mitotic stages in control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei mCD8-GFP (green)-expressing FC clones. In A,C, clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads mark FCs deficient in aPKC. Asterisks mark FCs that show the presence of aPKC in double mutants. Data are mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the percentage of cells with reduced aPKC in a single clone of a separate egg chamber. The n values of the genotypes control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei are given as n (the number of clones) and N (the number of independent replicates): n=12,13,9,14,13,26 and N=3,5,3,3,3,3, respectively, for endocycling stages and n=17,17,26,11,15,11 and N=3,4,3,3,3,3, respectively, for mitotic stages. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; *P<0.05; ***P<0.001. Scale bars: 10 μm.
Drp1-depleted FC clones show reduction or loss of aPKC in endocycling and mitotic stages, and this is partially recovered on additional depletion of ROS scavengers. (A) Representative images showing aPKC (red) staining in PFC clones from control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei (Aa-Af). (B) The graph shows quantification of the percentage of cells with reduced aPKC in each PFC clone adjacent to the oocyte in control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei. (C) Representative images showing aPKC (red) staining in control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei FC clones at the mitotic stages (Ca-Cf). (D) The graph shows the quantification of the percentage of cells with reduced aPKC in each FC clone in mitotic stages in control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei mCD8-GFP (green)-expressing FC clones. In A,C, clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads mark FCs deficient in aPKC. Asterisks mark FCs that show the presence of aPKC in double mutants. Data are mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the percentage of cells with reduced aPKC in a single clone of a separate egg chamber. The n values of the genotypes control FRT40A, drp1KG, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei are given as n (the number of clones) and N (the number of independent replicates): n=12,13,9,14,13,26 and N=3,5,3,3,3,3, respectively, for endocycling stages and n=17,17,26,11,15,11 and N=3,4,3,3,3,3, respectively, for mitotic stages. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; *P<0.05; ***P<0.001. Scale bars: 10 μm.
In summary, we found that the depletion of Opa1 led to increased ROS and mitochondrial fragmentation, similar to SOD2 and catalase depletion.
ROS increase in Drp1-depleted PFCs shows an alleviation of polarity defects and decrease of aPKC
ROS alters the activity and distribution of protein kinases by oxidation (Corcoran and Cotter, 2013). Protein kinase C (PKC) is activated by mtROS in human peritoneal mesothelial cells (HPMC), and PKC can in turn further enhance the ROS, to continue the activation cycle (Lee et al., 2004). ROS plays a significant role in the regulation of PKC activation. As drp1KG;opa1i FCs showed alleviation of the defect of multilayering and aPKC decrease compared with drp1KG alone, we tested whether depletion of ROS scavengers in Drp1-depleted FCs improved their organization and aPKC levels. The height and area of the clone in drp1KG;sod2i (Fig. S2Ad,B,C) and drp1KG;catalasei (Fig. S2Af,B,C) mutant clones was reduced compared with drp1KG. The levels of aPKC in endocycling PFCs adjacent to the oocyte (Fig. 5A,B) and mitotic stage FC clones (Fig. 5C,D) of drp1KG;sod2i and drp1KG;catalasei increased on the apical membrane compared with drp1KG alone. This rescue of the multilayering defect (Fig. S2Ae,B,C) and aPKC depletion (Fig. S3Ag,B,Cg,D) was also seen when a second independent RNAi for SOD2 was expressed in drp1KG homozygous cells. This increase in aPKC was similar to that observed in the double mutant drp1KG;opa1i clones (Fig. 2A-D). In addition to aPKC levels, we found that the apical length from drp1KG;sod2i and drp1KG;catalasei mutant FCs was less constricted compared with the Drp1-depleted FCs in the endocycling stage (Fig. S7A,B), whereas drp1KG;sod2i follicle cells were less constricted in mitotic stages (Fig. S7C,D). Thus, a partial recovery of multilayering, apical constriction and aPKC levels observed in drp1KG;sod2i and drp1KG;catalasei suggested an important role of ROS in regulating mitochondrial morphology and distribution of apical polarity protein aPKC in follicle epithelial cells.
ERK accumulation in Drp1-depleted FCs is suppressed by additional depletion of Opa1 and ROS scavengers
EGFR signaling leads to the activation of ERK by phosphorylation (González-Reyes and St Johnston, 1998). Previous studies have shown that EGFR signaling-driven accumulation of doubly phosphorylated ERK (dpERK) increases in the cytoplasm of Drp1-depleted PFCs (Tomer et al., 2018). A decrease in EGFR signaling in pre-FCs is essential for recruiting aPKC and forming the apical domain (Castanieto et al., 2014). We found that aPKC is decreased in Drp1-depleted cells (Fig. 2). As ROS increase in Drp1-depleted FCs led to an increase in aPKC at the apical membrane, we assessed whether this was coincident with a decrease in dpERK. For this, we quantified the levels of dpERK on reducing Opa1 and ROS scavengers in Drp1-depleted FCs. We found that there was a reduction in dpERK similar to controls in endocycling stage PFCs adjacent to the oocyte and mitotic stage FCs in the drp1KG;opa1i (Fig. 6Aa,Ad,B,Cd,D), drp1KG;sod2i (Fig. 6Aa,Af,B,Cf,D) and drp1KG;catalasei (Fig. 6Aa,Ah,B,Ch,D) combination compared with drp1KG alone (Fig. 6Ab,B,Cb,D). In summary, we found that an increase in aPKC in Drp1-depleted FCs on additionally depleting Opa1 and ROS scavengers is accompanied by a loss of dpERK in the cytoplasm.
Increase in dpERK in drp1KG clones is rescued on additional depletion of Opa1 and ROS scavengers. (A,C) Representative images showing dpERK (red) in control FRT40A, drp1KG, opa1i, drp1KG;opa1i, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei PFC clones at endocycling stage (Aa-Ah) and FC clones at mitotic stage (Ca-Ch). (B,D) The graphs show the quantification of relative dpERK fluorescence compared with neighboring control cells in control FRT40A, drp1KG, opa1i, drp1KG;opa1i, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei PFC clones at endocycling stage (B) and FC clones at mitotic stages (D). mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. The data of each genotype is presented with its respective mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the relative fluorescence intensity of dpERK from a single clone from a separate egg chamber. The n values for genotypes control FRT40A, drp1KG, opa1i, drp1KG;opa1i, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei are given as n (the number of clones) and N (the number of independent replicates): n=25,24,20,18,16,18,14,55 and N=4,4,3,4,3,3,3,6, respectively, for endocycling stages and n=52,40,31,12,27,30,30,47 and N=5,4,2,4,3,3,2,5, respectively, for mitotic stages. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; *P<0.05; **P<0.01; ***P<0.001. Scale bars: 10 μm.
Increase in dpERK in drp1KG clones is rescued on additional depletion of Opa1 and ROS scavengers. (A,C) Representative images showing dpERK (red) in control FRT40A, drp1KG, opa1i, drp1KG;opa1i, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei PFC clones at endocycling stage (Aa-Ah) and FC clones at mitotic stage (Ca-Ch). (B,D) The graphs show the quantification of relative dpERK fluorescence compared with neighboring control cells in control FRT40A, drp1KG, opa1i, drp1KG;opa1i, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei PFC clones at endocycling stage (B) and FC clones at mitotic stages (D). mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. The data of each genotype is presented with its respective mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the relative fluorescence intensity of dpERK from a single clone from a separate egg chamber. The n values for genotypes control FRT40A, drp1KG, opa1i, drp1KG;opa1i, sod2i, drp1KG;sod2i, catalasei and drp1KG;catalasei are given as n (the number of clones) and N (the number of independent replicates): n=25,24,20,18,16,18,14,55 and N=4,4,3,4,3,3,3,6, respectively, for endocycling stages and n=52,40,31,12,27,30,30,47 and N=5,4,2,4,3,3,2,5, respectively, for mitotic stages. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; *P<0.05; **P<0.01; ***P<0.001. Scale bars: 10 μm.
ERK loss in Drp1-depleted FCs leads to alleviation of the polarity and aPKC defects
As dpERK decreased in Drp1-depleted FCs on the additional expression of RNAi against Opa1 and ROS scavengers, we tested whether ERK depletion would change the apical aPKC accumulation. As expected, ERK RNAi (erki) expression decreased dpERK levels in the FCs of both endocycling and mitotic stages in drp1KG (Fig. S8A-D). There was a decrease in height (Fig. S8E,F) and area of the clone (Fig. S8E,G) in PFCs homozygous for the null allele drp1KG expressing erki compared with drp1KG alone. We observed that, unlike drp1KG, aPKC was present apically in drp1KG;erki in endocycling FCs adjacent to the oocyte (Fig. 7A,B) and mitotic FCs (Fig. 7C,D). This rescue of multilayering (Fig. S2Ag,B,C) and apical aPKC (Fig. S3Aj,B,Cj,D) was also seen in a second independent RNAi against ERK in combination with drp1KG. A decrease in ERK in PFCs, which contained EGFR signaling, led to alleviating polarity defects in Drp1-depleted FCs.
aPKC depletion in drp1KG FC clones recovers on additional depletion of ERK. (A,C) Representative images showing aPKC (red) in control FRT40A, drp1KG, erki and drp1KG;erki PFC clones at endocycling stage (Aa-Ad) and FC clones at mitotic stages (Ca-Cd). (B) The graph shows the quantification of the percentage of cells with reduced aPKC in each PFC clone adjacent to the oocyte in control FRT40A, drp1KG, erki and drp1KG;erki. (D) The graph shows the quantification of the percentage of cells with reduced aPKC in each FC clone in control FRT40A, drp1KG, erki and drp1KG;erki at mitotic stages. mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dashed yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads mark FCs deficient in aPKC. Asterisks mark FCs that show the presence of aPKC in double mutants. Data are mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the percentage of cells with reduced aPKC in a single clone of a separate egg chamber. The n values for the genotypes control FRT40A, drp1KG, erki and drp1KG;erki is given as n (the number of clones) and N (the number of independent replicates): n=9,11,10,19 and N=3,5,3,4, respectively, for endocycling stages and n=23,14,12,22 and N=5,5,3,3, respectively, for mitotic stages. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; ***P<0.001. Scale bars: 10 μm.
aPKC depletion in drp1KG FC clones recovers on additional depletion of ERK. (A,C) Representative images showing aPKC (red) in control FRT40A, drp1KG, erki and drp1KG;erki PFC clones at endocycling stage (Aa-Ad) and FC clones at mitotic stages (Ca-Cd). (B) The graph shows the quantification of the percentage of cells with reduced aPKC in each PFC clone adjacent to the oocyte in control FRT40A, drp1KG, erki and drp1KG;erki. (D) The graph shows the quantification of the percentage of cells with reduced aPKC in each FC clone in control FRT40A, drp1KG, erki and drp1KG;erki at mitotic stages. mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dashed yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads mark FCs deficient in aPKC. Asterisks mark FCs that show the presence of aPKC in double mutants. Data are mean±s.e.m. Each data point (n) in the endocycling and mitotic stages represents the percentage of cells with reduced aPKC in a single clone of a separate egg chamber. The n values for the genotypes control FRT40A, drp1KG, erki and drp1KG;erki is given as n (the number of clones) and N (the number of independent replicates): n=9,11,10,19 and N=3,5,3,4, respectively, for endocycling stages and n=23,14,12,22 and N=5,5,3,3, respectively, for mitotic stages. The statistical test performed was one-way ANOVA with Tukey's multiple comparisons. ns, non-significant; ***P<0.001. Scale bars: 10 μm.
Even though Drp1-depleted FCs showed an increase in dpERK, this increase occurred in the cytoplasm, and there was a loss of oocyte patterning depicted by loss of oocyte nuclear migration to the dorso-anterior location (Mitra et al., 2012; Tomer et al., 2018). We found that the oocyte nucleus remained in the posterior location in drp1KG;opa1i, drp1KG;sod2i and drp1KG;catalasei at a frequency similar to drp1KG (Fig. S9). These data showed that even though the cytoplasmic accumulation of dpERK was reduced on depletion of Opa1 and ROS scavengers in Drp1-depleted FCs, the defect in oocyte nuclear migration and patterning still remained.
The Notch signaling defect in Drp1-deficient FCs is suppressed by the depletion of ROS scavengers
Notch signaling activation promotes the transition of the mitotic stage to the endocycling stage in FCs during oogenesis by the expression of transcription factor Hnt (Kim-Yip and Nystul, 2018; López-Schier and St Johnston, 2001; Ruohola et al., 1991). Mitochondrial fission by Drp1 leads to an opposing interaction between EGFR and Notch signaling. Drp1 depletion leads to loss of Hnt due to altered mitochondrial activity and accumulation of dpERK in PFCs (Mitra et al., 2012; Tomer et al., 2018). We further assessed whether the increase in ROS on depletion of ROS scavenging enzymes in Drp1-deficient FCs could suppress the defect in Notch signaling activation. We found that the depletion of Opa1, SOD2, Catalase and ERK showed activation of Hnt at levels similar to controls. The drp1KG;opa1i, drp1KG;sod2i, drp1KG;catalasei and drp1KG;erk1i combinations showed expression of Hnt unlike drp1KG (Fig. 8). In order to examine the impact of aPKC on Hnt loss in Drp1-depleted PFCs, we stained the drp1KG;aPKC-ΔN combination for assessing mitochondrial morphology and Hnt. We found that mitochondria were similar to controls on aPKC-ΔN overexpression and remained clustered in drp1KG;aPKC-ΔN (Fig. S10A). Whereas apical constriction and multilayering were rescued (Fig. 3), Hnt was still missing in these clones (Fig. S10B). These data showed that the loss of Notch-signaling-mediated differentiation was alleviated by the increase in ROS but not aPKC in FCs depleted for Drp1. aPKC decrease, therefore, was likely to be due to the change in mitochondrial morphology and ROS but not due to the impact of mitochondrial dynamics on Notch signaling.
Hnt depleted in drp1KG PFC clones recovers on additional depletion of Opa1, ROS scavengers and ERK. Representative images showing Hnt (red) in PFCs in the genotypes control FRT40A (100% show Hnt, n=10; a), drp1KG (0%, n=11; b), opa1i (100%, n=10; c), drp1KG;opa1i (94%, n=18; d), sod2i (100%, n=10; e), drp1KG;sod2i (60%, n=20; f), catalasei (100%, n=10; g), drp1KG;catalasei (29%, n=24; h), erki (100%, n=10; i) and drp1KG;erk1i (100%, n=10; j). Numbers denote the percentage of egg chambers with PFC clones that express Hnt. mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads mark PFCs containing Hnt. Scale bars: 10 μm.
Hnt depleted in drp1KG PFC clones recovers on additional depletion of Opa1, ROS scavengers and ERK. Representative images showing Hnt (red) in PFCs in the genotypes control FRT40A (100% show Hnt, n=10; a), drp1KG (0%, n=11; b), opa1i (100%, n=10; c), drp1KG;opa1i (94%, n=18; d), sod2i (100%, n=10; e), drp1KG;sod2i (60%, n=20; f), catalasei (100%, n=10; g), drp1KG;catalasei (29%, n=24; h), erki (100%, n=10; i) and drp1KG;erk1i (100%, n=10; j). Numbers denote the percentage of egg chambers with PFC clones that express Hnt. mCD8-GFP (green)-expressing FC clones of the indicated genotype are marked with dotted yellow outlines. The nucleus (blue) is stained with Hoechst. Arrowheads mark PFCs containing Hnt. Scale bars: 10 μm.
In summary, our results show that increased ROS can alleviate the mitochondrial fusion, aPKC and Notch signaling defect in Drp1-depleted FCs.
DISCUSSION
In this study, we show that mitochondrial fission by Drp1 regulates apical polarity and signaling via mitochondrial morphology and ROS in Drosophila follicle epithelial cell differentiation. We find that appropriate ROS levels regulated by mitochondrial fragmentation are important for the interaction between activated ERK and aPKC in mitotic stage FCs (Fig. 9A,B). ROS also plays a crucial role in the interaction between EGFR and Notch signaling to differentiate PFCs during oogenesis (Fig. 9A,B). We discuss our results in the following three contexts: (1) the role of depletion of aPKC in causing multilayering of PFCs, (2) the function of ROS in regulating the onset of apical polarity and (3) the interaction of mitochondrial dynamics, ROS and activation of signaling pathways in epithelial polarity formation and differentiation.
Interaction between mitochondrial morphology, ROS and signaling in Drosophila FC differentiation. (A,B) The schematic shows the interaction between mitochondrial fission regulated by Drp1, ROS and polarity protein aPKC in mitotic (A) and endocycling (B) FCs of Drosophila oogenesis. Drp1 regulates mitochondrial fission, thereby controlling the levels of ROS and appropriate activation of ERK (dpERK), in turn affecting apical aPKC distribution in FCs in mitotic stages. In endocycling PFCs, Drp1 leads to mitochondrial fission and appropriate ROS and dpERK levels, thereby regulating aPKC accumulation at the apical membrane and induction of Hnt in Notch signaling.
Interaction between mitochondrial morphology, ROS and signaling in Drosophila FC differentiation. (A,B) The schematic shows the interaction between mitochondrial fission regulated by Drp1, ROS and polarity protein aPKC in mitotic (A) and endocycling (B) FCs of Drosophila oogenesis. Drp1 regulates mitochondrial fission, thereby controlling the levels of ROS and appropriate activation of ERK (dpERK), in turn affecting apical aPKC distribution in FCs in mitotic stages. In endocycling PFCs, Drp1 leads to mitochondrial fission and appropriate ROS and dpERK levels, thereby regulating aPKC accumulation at the apical membrane and induction of Hnt in Notch signaling.
The apical polarity protein aPKC decreases from the apical membrane in Drp1-depleted FCs in the mitotic and endocycling stages. Appropriate recruitment of polarity proteins in stem cell proliferation is important for the correct orientation of the cell division machinery. aPKC regulates the mitotic spindle orientation and helps epithelial cell division in the Drosophila wing disc (Guilgur et al., 2012). However, aPKC is not required for correct spindle orientation in FCs (Bergstralh et al., 2013). aPKC depletion by optogenetic inactivation has been recently found to result in the formation of gaps and multilayered epithelia (Osswald et al., 2022). This change in organization occurs due to the increased activation of Myosin II. Loss of Myosin II activation in aPKC-depleted FCs leads to an inhibition of actomyosin contractility (Osswald et al., 2022). It is, therefore, likely that an increase in actomyosin-based contractility occurs due to the decrease of apical aPKC in Drp1-depleted cells, leading to the occurrence of FCs in multiple layers. Future experiments testing the extent of increase in Myosin II activation in Drp1-depleted FCs will outline the mechanism by which multilayering occurs in the FC epithelium. An opposite effect of loss of activation of Myosin II-driven contractility due to the decrease in ROS has been observed in Drp1-depleted Drosophila embryos during cellularization and dorsal closure (Chowdhary et al., 2020; Muliyil and Narasimha, 2014). Therefore, the effects on contractility due to mitochondrial fragmentation and ROS are likely tissue specific and dependent upon the different factors regulating Myosin II activity. Further, an increase in aPKC in Drp1-depleted FCs alleviates the multilayering and polarity defects but not Notch signaling, thereby emphasizing that loss of ROS affects both polarity and signaling. EGFR signaling in turn, regulates polarity via aPKC.
Alterations in mitochondrial ROS levels are known to occur on the change in mitochondrial morphology. An increase in mitochondrial ROS occurs on the depletion of Opa1, and a decrease occurs on the depletion of Drp1 (Celotto et al., 2012; Chowdhary et al., 2020; Quintana-Cabrera et al., 2021; Tang et al., 2009; Yarosh et al., 2008). In addition, an increase in ROS is also likely to impact mitochondrial morphology. For example, an increase in ROS leads to mitochondrial fragmentation in cells undergoing delamination in dorsal closure in Drosophila embryogenesis (Muliyil and Narasimha, 2014). We also found that increased ROS, due to the depletion of ROS scavenger proteins such as SOD2 and Catalase, leads to mitochondrial fragmentation. ROS acts as a secondary messenger, oxidizing proteins and changing their activity and distribution. An increase in ROS removes Opa1 from the mitochondria to the cytosol in mammalian H22 T cells (Sanderson et al., 2015). This Opa1 loss is likely a mechanism by which ROS elevation leads to mitochondrial fragmentation, similar to Opa1 depletion in FCs. Decreased ROS during wound healing in epithelial cells leads to the inhibition of activation of Src kinase and inhibition of actomyosin constriction (Hunter et al., 2018; Muliyil and Narasimha, 2014; Ponte et al., 2020). ROS is known to regulate the activity of protein kinases and phosphatases (Corcoran and Cotter, 2013; Ray et al., 2012). ROS leads to oxidation of key catalytic cysteine residues and alteration of their enzymatic activity, thereby affecting a variety of cellular processes. Therefore, an increase in ROS by the depletion of ROS scavengers in Drp1-depleted FCs is likely to oxidize and stabilize aPKC on the apical domain and correct epithelial polarity defects.
Our results show that ROS increase in Drp1-depleted FCs not only led to the suppression of polarity defects but also decreased activated ERK and increased Notch signaling, promoting the differentiation of PFCs. Multilayering of PFCs produced by depletion of polarity complexes often shows aberrant EGFR and reduced Notch signaling in the formation of PFCs. Depletion of the Dlg-Lgl-Scribble complex in PFCs leads to multilayering along with accumulation of dpERK and loss of Notch signaling (Li et al., 2008, 2009; Tian and Deng, 2008). aPKC mutant clones show multilayering, but their analysis for defects in EGFR and Notch signaling has yet to be reported in detail (Kim et al., 2009; Wodarz et al., 2000). Intestinal epithelial cells in Drosophila proliferate to form multilayers against bacterial infection and show increased EGFR signaling (Buchon et al., 2010). EGFR and its downstream factors regulate mitochondrial fragmentation (Kashatus et al., 2015; Mitra et al., 2012). EGFR signaling is also known to require fragmented mitochondria produced by the activation of Drp1-driven fission for the proliferation of cancer cells (Kashatus et al., 2015). We found that increased ROS by depletion of Opa1 and ROS scavengers prevented the accumulation of dpERK. Decreased EGFR signaling and downstream factors are important for forming apical polarity in pre-FCs (Castanieto et al., 2014). This loss of dpERK is likely to be responsible for stabilizing aPKC in the apical membrane in mitotic- and endocycling-stage FCs depleted of Drp1. It is further likely that aberrant EGFR signaling on the loss of polarity complexes also leads to a defect in mitochondrial morphology. Our data motivate a systematic analysis of the regulation of mitochondrial morphology proteins on the change in epithelial polarity complexes in FCs and other epithelial cells in Drosophila.
Increased EGFR signaling is known to oppose Notch signaling in various contexts, such as the Drosophila wing disc and the FCs (Hasson et al., 2005). The transcription factor Groucho executes the repression between EGFR and Notch signaling in the wing disc. We have previously found that the depolarization of mitochondria promotes the Notch signaling-driven differentiation in FCs (Tomer et al., 2018). Moreover, decreased mitochondrial membrane potential partially reverses the Notch-driven differentiation defect in Drp1-depleted PFCs (Tomer et al., 2018). We found that a decrease in ERK leads to restoring aPKC and Notch signaling in Drp1-depleted PFCs. A link between apical polarity and Notch signaling has been shown in intestinal stem cells. An increase in aPKC in intestinal stem cells increases Notch signaling (Goulas et al., 2012). In this study, we find that mitochondrial fragmentation and an increase in ROS driven by the depletion of Opa1 and ROS scavengers alleviates the Notch-driven differentiation defect in Drp1-depleted PFCs. In addition, an increase in ROS also leads to a rescue of the defect in accumulation of dpERK in Drp1-depleted FCs. In summary, these studies show that mitochondrial activity, dynamics and appropriate ROS are essential for mediating the antagonistic interaction between EGFR and Notch signaling in FCs.
MATERIALS AND METHODS
Drosophila genetics
Drosophila strains were grown and maintained at 25°C on a standard cornmeal agar medium. The Drosophila stocks obtained from the Bloomington Drosophila Stock Center are drp1KG (y[1]; P{y[+mDint2] w[BR. E. BR]=SUPor-P}Drp1[KG03815]/CyO; ry[506], stock number #BL13510 FRT40A); UAS-opa1 RNAi (opa1i) (y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP. HMS00349}attP2, #BL32358); UAS-opa1 miRNA (opa1 miRNA) (w[*]; Bl[1]/CyO; P{w[+mC]=UAS-Opa1.RNAi.CDS}3, #67159); UAS-marf RNAi (marf(MG)i), UAS-erk RNAi (erki) (y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP. HMS00173}attP2, #BL34855); UAS-erk RNAi (erki) (y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP. GL00215}attP2, #BL36058); UAS-sod2 RNAi (sod2i) (y[1] v[1]; P{y[+t7.7] v[+t1.8]=TRiP. JF01989}attP2, # BL25969); UAS-sod2 RNAi (sod2i) (y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP. HMS00783}attP2, #BL32983); UAS-catalase RNAi (catalasei) (y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP. HMS00990}attP2, #BL34020); UAS-aPKC RNAi (aPKCi) (y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP. HMS01320}attP2, #BL34332); UAS-aPKC RNAi (aPKCi) (y[1] sc[*] v[1] sev[21]; P{y[+t7.7] v[+t1.8]=TRiP. GL00007}attP2, #BL35140); UAS-aPKC-ΔN) (w[*]; P{w[+mC]=UAS-aPKC.ΔN}3, #BL51673); GR-Gal4 (w[*]; P{w[+mW.hs]=GawB}GR1, #BL36287). For the MARCM experiments, hs-FLP; Gal80-FRT40A/CyO; tub-Gal4,UAS-CD8-GFP/TM6 and FRT40A/SM6a were obtained from Nicole Grieder and Mary Lilly (National Institutes of Health, MD, USA). Standard genetic crosses were performed to make the following combinations for experiments:
drp1KG FRT40A/CyO; opa1i 32358/TM6 and FRT40A/CyO; opa1i 32358/TM6,
drp1KG FRT40A/CyO; opa1 miRNA 67159/TM6 and FRT40A/CyO; opa1 miRNA 67159/TM6,
drp1KG FRT40A/CyO; marf(MG)i/TM6 and FRT40A/CyO; marf(MG)i/TM6,
drp1KG FRT40A/CyO; erki 34855/TM6 and FRT40A/CyO; erki 34855/TM6,
drp1KG FRT40A/CyO; erki 36058/TM6 and FRT40A/CyO; erki 36058/TM6,
drp1KG FRT40A/CyO; sod2i 25969/TM6 and FRT40A/CyO; sod2i 25969/TM6,
drp1KG FRT40A/CyO; sod2i 32983/TM6 and FRT40A/CyO; sod2i 32983/TM6,
drp1KG FRT40A/CyO; catalasei 34020/TM6 and FRT40A/CyO; catalasei 34020/TM6,
drp1KG FRT40A/CyO; aPKC-ΔN 51673/TM6 and FRT40A/CyO; aPKC-ΔN 51673/TM6.
Induction of the mitotic clones using the MARCM technique
The flies having genotype drp1KG FRT40A/CyO were crossed to hs-FLP; Gal80 FRT40A/CyO; tub-Gal4,UAS-CD8-GFP/TM6. F1 flies having the genotype hs-FLP/+; drp1KG FRT40A/Gal80 FRT40A; tub-Gal4,UAS-CD8-GFP/+ were heat shocked at 37°C for 60 min in a water bath to induce FC clones in the ovaries. The flies were transferred to the vials containing fresh media with yeast granules and were further transferred into fresh media every 3 days for 10 days post heat shock to assess germline clones. To perform RNAi-mediated knockdown for epistatic analysis of flies carrying opa1i, opa1 miRNA, marf(MG)i, catalasei, sod2i (25696), sod2i (32983), erki (34855), erki (36058), aPKC-ΔN in the FRT40A and drp1KG FRT40A shown above were crossed to hs-FLP; Gal80 FRT40A/CyO; tub-Gal4,UAS-CD8-GFP/TM6, and appropriate flies were selected to perform heat shocks and assess the phenotypes in clones after 10 days.
Immunostaining of the Drosophila ovaries
Drosophila ovaries were dissected 10 days post-heat shock in a serum-free Schneider's insect medium, and ovarioles were separated using fine needles. The dissected ovaries were washed in fresh Schneider's insect medium to remove debris released during dissection and fixed using freshly prepared 4% paraformaldehyde in 1× PBS for 20-25 min followed by three washes in 1× PBS containing 0.3% Triton X-100 (T8787, Sigma-Aldrich) (hereafter called 0.3% PBST) for 10 min each. The ovaries were transferred to a blocking solution containing 2% bovine serum albumin (MB083, HIMEDIA) dissolved in 0.3% PBST for 1 h and incubated with primary antibodies for 16-18 h followed by three 10 min washes with 0.3% PBST. The secondary antibodies, dissolved in 0.3% PBST, were added for 2 h. The ovaries were washed thrice in 0.3% PBST for 5 min each. The second wash of 0.3% PBST contained Hoechst 33258 (H3569, Thermo Fisher Scientific) at 1:1000 dilution. The ovarioles were separated and mounted on a glass slide with Slowfade (S36937, Thermo Fisher Scientific). The primary antibodies used were: chicken GFP (A10262, Thermo Fisher Scientific, 1:500), rabbit aPKC (C-20, Santa Cruz Biotechnology, 1:500), mouse aPKC (H-1, Santa Cruz Biotechnology, 1:500), rabbit Patj (Bhat et al., 1999; 1:1000), mouse Dlg [4F3, Developmental Studies Hybridoma Bank (DSHB), 1:10], mouse Hindsight (1G9, DSHB, 1:10), rat DE-cadh (DCAD2, DSHB, 1:5), rabbit dpERK (4370, Cell Signaling Technology, 1:200), mouse ATP-β (ab14730, Abcam, 1:200). DCAD2 antibody staining was carried out as previously described (Nilangekar et al., 2019). The secondary antibodies were obtained from Thermo Fisher Scientific and used at a 1:1000 dilution: anti-chicken 488 (A11039), anti-mouse 568 (A11004), anti-mouse 633 (A21050), anti-mouse 647 (A21235), anti-rabbit 568 (A11011), anti-rabbit 633 (A21070), anti-rabbit 647 (A21245).
MitoSOX staining for mitochondrial ROS
The MitoSOX dye (M36008, Thermo Fisher Scientific) gets oxidized by mitochondrial superoxides and fluoresces (Robinson et al., 2006). The ovaries were dissected and washed in Schneider's medium. The ovaries were immersed in Schneider's medium containing MitoSOX dye with a final concentration of 5 μM for 15 min followed by a 5 min wash with Schneider's medium (Parker et al., 2017). The ovaries were immediately transferred to a glass slide and mounted using Schneider's medium for imaging. The 16-color rainbow scale was used to show the change in the fluorescence relative to the background in clonal experiments. The control clone cells show similar fluorescence to non-clone cells. The fluorescence is relatively increased in clones expressing an RNAi against Opa1 and ROS scavengers. For the experiments where GR-Gal4 was used to drive aPKC RNAi, we imaged egg chambers with the same parameters across different genotypes to compare the mitoSOX intensity.
Identification of stages of development in ovarioles
The stage identification was performed by measuring the surface area of the middle stack of the egg chambers (Jia et al., 2016). Images were quantified with the help of Fiji and ImageJ (Schindelin et al., 2012; Schneider et al., 2012). PFCs were identified by their presence at the posterior side and their placement adjacent to the oocyte in stage 7-10 ovarioles.
Quantification of multilayering, clone area and clone height
The estimation of the number of layers of FCs was performed for PFCs in controls and mutants. The rows of cells per clone were counted in the mCD8-GFP-positive clones to report the phenotype of multilayering. The area of the clone was estimated by using a central section of the PFC clone. We quantified the height of the clones in multilayers as a ratio to the height of the adjacent PFCs in different genotypes in stages 7 to 9. We quantified the clone area of different genotypes using a middle stack of each PFC clone in stages 7 to 9.
Quantification of mitochondrial morphology
To evaluate the mitochondrial morphology, the Streptavidin or ATP-β staining of the mitochondria in the GFP-positive follicular cell clones and GFP-negative neighboring cells were compared visually. Streptavidin or ATP-β staining in the FCs that are more punctate was used to characterize the fragmented mitochondria. The compactness of the mitochondrial staining and a punctate appearance from the ATP-β staining were used to distinguish between differentially clustered mitochondria. Furthermore, the mitochondrial morphology of FCs, which lacked any discernible variations, was classified as intermediate. From the observations, the percentage of egg chambers with the respective mitochondrial morphology was estimated.
Estimation of numbers of cells containing aPKC, Patj, DE-cadh and Hnt
For the analysis of cells containing polarity protein aPKC on the apical membrane, we estimated the numbers of FCs per clone containing reduced or loss of aPKC at the apical membrane in the mCD8-GFP-positive clones. We expressed them as a percentage of the total cells in the clone. To quantitate the appearance of the aPKC on the apical membrane in epistasis experiments, the percentage of PFCs containing apical aPKC was counted in the layer of cells adjacent to the oocyte. A similar analysis was performed for the subapical protein Patj and adherens junction protein DE-cadh between control and drp1KG mutants. To analyze Notch-signaling-mediated differentiation, we counted the percentage of the egg chambers with Hnt to compare genotypes. Hnt was present in patches all over the clones in double mutants, which showed a rescue of its expression.
Estimation of the aPKC intensity
For the experiments where Gal4 was used to drive aPKC knockdown by RNAi, we imaged egg chambers with the same laser power and gain settings on the confocal microscope across different genotypes to compare the aPKC intensity.
Estimation of the apical constriction
The length of the apical membrane was estimated by drawing a line in ImageJ between two adjacent membranes marked by mCD8-GFP. For ovarioles stained with Dlg, the apical length between lateral membranes marked with Dlg and mCD8-GFP was measured for the clonal FCs and Dlg alone in adjacent non-clonal FCs. For ovarioles marked with aPKC, apical length was determined in mCD8-GFP- and aPKC-stained clonal FCs and with aPKC alone in adjacent non-clonal FCs. A ratio was obtained for each clone with the adjacent heterozygous FCs and represented for different genotypes in a graph to estimate the extent of apical constriction.
Quantification of dpERK
For the analysis of relative dpERK fluorescence in different genotypes, we marked a region of interest (ROI) around the clone expressing mCD8-GFP and neighboring FCs. A background ROI was used to subtract the noise from the actual signal. Appropriate thresholding was applied to quantitate the dpERK in cells. The ratio of the thresholded intensity value of dpERK from the clone to their neighboring cell was computed.
n values and statistics
Each data point (n) represents quantification from a distinct clone in different egg chambers. A minimum number of four egg chambers were used for the statistics. The data for each genotype is shown with its respective mean and s.e.m. The statistical test applied for two groups is two-tailed unpaired Student's t-test, and for more than two groups is a one-way analysis of variance (ANOVA) with Tukey's multiple comparison test.
Acknowledgements
We thank Deepa Subramanyam, Mayurika Lahiri, Nagaraj Balasubramanian and R.R. lab members for continuous discussions on the data. Stocks obtained from the Bloomington Drosophila Stock Center (NIH P40OD018537) were used in this study. We have used Microsoft Excel, GraphPad Prism, Fiji and ImageJ for data analysis. We acknowledge the Drosophila and Microscopy facilities of the Indian Institute of Science Education and Research, Pune, India, for stock maintenance, fly food and microscopy.
Footnotes
Author contributions
Conceptualization: B.U., D.T., R.R.; Methodology: B.U., R.K.V., D.T., R.R.; Validation: B.U., R.K.V., R.R.; Formal analysis: B.U., R.K.V., D.T., R.R.; Investigation: B.U., R.K.V., R.R.; Resources: R.R.; Writing - original draft: B.U., R.R.; Writing - review & editing: B.U., D.T., R.R.; Visualization: B.U., R.K.V., D.T., R.R.; Supervision: R.R.; Project administration: B.U., R.R.; Funding acquisition: R.R.
Funding
B.U., R.K.V. and D.T. thank the Council of Scientific and Industrial Research, India, for the graduate fellowship. B.U. thanks the Science and Engineering Research Board, India and Wellcome Trust DBT India Alliance for the project fellowship, and Innoplexus Consulting Services Pvt and Scivic Engineering Pvt for additional financial support. R.R. thanks the Science and Engineering Research Board (CRG/2018/003347) and the Wellcome Trust DBT India Alliance Senior Fellowship (IA/S/22/1/506232) for funding. Open access funding provided by the Indian Institute of Science Education Research Pune. Deposited in PMC for immediate release.
Data availability
All relevant data can be found within the article and its supplementary information.
Peer review history
The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201732.reviewer-comments.pdf
References
Competing interests
The authors declare no competing or financial interests.