ERK regulates mitochondrial membrane potential in fission deficient Drosophila follicle cells during differentiation

Author’s Accepted Manuscript ERK regulates mitochondrial membrane potential in fission deficient Drosophila follicle cells during differentiation Darshika Tomer, Rohan Chippalkatti, Kasturi Mitra, Richa Rikhy www.elsevier.com/locate/developmentalbiology

PII: DOI: Reference:

S0012-1606(17)30580-8 https://doi.org/10.1016/j.ydbio.2017.11.009 YDBIO7629

To appear in: Developmental Biology Received date: 27 August 2017 Revised date: 4 November 2017 Accepted date: 15 November 2017 Cite this article as: Darshika Tomer, Rohan Chippalkatti, Kasturi Mitra and Richa Rikhy, ERK regulates mitochondrial membrane potential in fission deficient Drosophila follicle cells during differentiation, Developmental Biology, https://doi.org/10.1016/j.ydbio.2017.11.009 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ERK regulates mitochondrial membrane potential in fission deficient Drosophila follicle cells during differentiation Darshika Tomera, Rohan Chippalkattia1, Kasturi Mitrab and Richa Rikhy*a

aBiology,

Indian Institute of Science Education and Research, Homi Bhabha Road,

Pashan, Pune, 411008, India

Phone: +91-20-25908065 bDepartment

of Genetics, University of Alabama at Birmingham, Birmingham, AL

35294, USA * To whom correspondence is addressed: [email protected] Abbreviations PFC, Posterior Follicle Cells; EGFR, Epidermal Growth Factor Receptor; NICD, Notch Intra-Cellular Domain; ROS, Reactive Oxygen Species; ETC, Electron Transport Chain; Ψm, Mitochondrial Membrane Potential Abstract Mitochondrial morphology regulatory proteins interact with signaling pathways involved 1

Present address: Institute of Cell Biology, University of Bern, Baltzerstrasse 4/6, 3012 Bern, Switzerland

1

in differentiation. In Drosophila oogenesis, EGFR signaling regulates mitochondrial fragmentation in posterior follicle cells (PFCs). EGFR driven oocyte patterning and Notch signaling mediated differentiation are abrogated when PFCs are deficient for the mitochondrial fission protein Drp1. It is not known whether fused mitochondrial morphology in drp1 mutant PFCs exerts its effects on these signaling pathways through change in mitochondrial electron transport chain (ETC) activity. In this study we show that aggregated mitochondria in drp1 mutant PFCs have increased mitochondrial membrane potential. We perform experiments to assess the signaling pathway regulating mitochondrial membrane potential and how this impacts follicle cell differentiation. We find that drp1 mutant PFCs show increase in phosphorylated ERK (dpERK) formed downstream of EGFR signaling. ERK regulates high mitochondrial membrane potential in drp1 mutant PFCs. PFCs depleted of ERK and drp1 are able to undergo Notch mediated differentiation. Notably mitochondrial membrane potential decrease via ETC inhibition activates Notch signaling at an earlier stage in wild type and suppresses the Notch signaling defect in drp1 mutant PFCs. Thus this study shows that the EGFR pathway maintains mitochondrial morphology and mitochondrial membrane potential in follicle cells for its functioning and decrease in mitochondrial membrane potential is needed for Notch mediated differentiation. Key words: mitochondria, signaling, Drosophila, ERK, Notch

Introduction 2

The mitochondrial fusion/fission machinery has a key role in embryogenesis and stem cell differentiation (Kasahara and Scorrano, 2014). Mitochondrial fusion proteins play a critical function in mouse embryogenesis and neuronal differentiation (Chen et al., 2003; Williams et al., 2010). Mitochondrial fission protein, Drp1 is essential for mouse and C. elegans embryogenesis and Drosophila follicle cell differentiation (Ishihara et al., 2009; Labrousse et al., 1999; Mitra et al., 2012). Mitochondrial functions in energy metabolism, calcium homeostasis, cell cycle and signaling are modulated by changes in its morphology (Nunnari and Suomalainen, 2012). For example, periodic changes in mitochondrial morphology, mitochondrial membrane potential and ATP generation occur during the cell cycle (Mitra et al., 2009; Schieke et al., 2008). Hence mitochondrial morphology may modulate differentiation during an organism’s development via its effect on key mitochondrial functions. Recent literature links mitochondrial morphology to conserved signaling pathways such as Notch and epidermal growth factor (EGF) involved in cellular patterning and differentiation. The Notch pathway is activated by ligand binding to the Notch receptor followed by transmembrane cleavage to generate an intracellular domain (NICD) that is transported to the nucleus (Struhl and Adachi, 1998). NICD increases mitochondrial fusion protein Mitofusin2 (Mfn2) by the Akt pathway and inhibits mitochondrial fragmentation thereby protecting cancer cells from apoptosis (Perumalsamy et al., 2010). Notch signaling is increased in mouse embryonic cardiomyocytes mutant for mitochondrial fusion proteins Opa and Mfn (Kasahara et al., 3

2013) and decreased in oogenesis in Drosophila drp1 mutant follicle cells containing fused mitochondria (Mitra et al. 2012). Elevated cell calcium and increased protease activity to generate NICD are possible mechanisms by which fragmented mitochondrial morphology affect Notch signaling in cardiomyocytes (Kasahara et al. 2013). EGF binds the EGFR and activates Ras-MEK-ERK through a series of phosphorylation events eventually leading to regulation of cytoplasmic and/or nuclear substrates and gene expression (Hughes, 1995). Src mediated activation of EGFR leads to its mitochondrial translocation and binding to cytochrome oxidase subunit II enhancing tumorigenesis (Boerner et al., 2004). Mitochondrial EGFR supports its fusion by Prohibitin2 and Opa1 activation (Bollu et al., 2014). Activated Ras along with dysfunctional mitochondria leads to metastasis in Drosophila eye epithelia (Ohsawa et al., 2012) and directly induces mitochondrial fragmentation causing a metabolic switch to glycolysis in cancer cells (Chiaradonna et al., 2006). ERK2 inactivates Mfn1 and activates Drp1 by phosphorylation in mammalian cancer cells (Kashatus et al., 2015; Pyakurel et al., 2015). Taken together the EGFR-Ras-ERK pathway increases glycolysis and regulates mitochondrial fragmentation for appropriate activation in cancer cells. Mitochondrial morphology and mitochondrial electron transport chain (ETC) are both likely to affect signaling pathways and the mechanism of interaction is unknown. Fused mitochondrial morphology in drp1 mutant ovarian follicle cells (FCs) results in loss of Notch mediated FC differentiation (Mitra et al., 2012). EGFR signaling regulates mitochondrial fragmentation in FCs (Mitra et al., 2012). Fused mitochondrial 4

morphology is correlated with high mitochondrial membrane potential, high calcium sequestration and increased ATP generation (Chen et al., 2003; Lodi et al., 2004; Thayer and Miller, 1990; Zanna et al., 2008). Fragmented mitochondrial morphology is associated with lowered ETC activity and decreased ATP generation (Ashrafian et al., 2010; Jheng et al., 2012). It is not known whether the EGFR pathway alters mitochondrial ETC and mitochondrial membrane potential on fused mitochondria in drp1 mutant FCs. Such an alteration may have downstream effects on Notch signaling. Hence we developed assays for measuring mitochondrial membrane potential in drp1 mutant FCs and tested the effect of its loss in fused mitochondria on EGFR and Notch signaling. The Drosophila ovary consists of chains of successively mature chambers housing the oocyte. FCs surround the germ cells and divide until stage 6 and then differentiate to form endocycling cells (Klusza and Deng, 2011; Nystul and Spradling, 2010) (Figure 1A). Endocycling is induced by Hindsight (Hnt) expression activated by the Notch pathway (Sun and Deng, 2007). Gurken from the oocyte activates EGFR in stage 8 posterior follicle cells (PFCs) covering the oocyte and this results in relatively fragmented mitochondrial morphology (Mitra et al., 2012). The EGFR pathway manifests via Ras-ERK and controls oocyte nucleus movement to the dorso-anterior position thus patterning the body axis of the embryo (Roth and Lynch, 2009) (Figure 1A). In the present analysis of effects of fused mitochondria in drp1 mutant FCs, we identify that ERK regulates increased mitochondrial membrane potential in drp1 mutant FCs. ERK depletion or mitochondrial ETC disruption shows Notch mediated 5

differentiation in drp1 mutant FCs.

Materials and Methods Drosophila genetics: All Drosophila crosses were performed in standard cornmeal agar medium at 25oC. The drp1KG03815 (drp1KG) (BL13510), drp1i (BL51483), erki (BL36058), rasi (BL31469), pdswi (BL29592), collagen-Gal4 (BL7011) lines were obtained from the Bloomington (BL) Stock Center. The stock hsflp; Gal80FRT40A/CyO; tub-Gal4,UAS CD8-GFP/TM6 was obtained from Nicole Grieder. The stock FRT40A/Cyo was obtained from Mary Lilly. The UASP-mito-GFP flies were obtained from Rachel Cox (Cox and Spradling, 2003). The drp1KGFRT40A/CyO, drp1KG FRT40A/CyO; erki/TM6 , drp1KGFRT40A/CyO; rasi/TM6 and drp1KGFRT40A/CyO; pdswi/TM6 stocks were generated using standard genetic crosses. Generation of follicle cell clones Homozygous drp1KG mutant clones were generated by the FLP-FRT mediated sitespecific recombination using MARCM system (Golic and Lindquist, 1989). Flies containing drp1KGFRT40A were crossed with hs-flp; Gal80 FRT40A/CyO; tub-Gal4, UAS CD8-GFP/TM6. 1-3 day adult females carrying the genotype hsflp/+; drp1KGFRT40A/Gal80 FRT40A; tub-Gal4, UAS CD8-GFP/+ were heat pulsed at 37.5 oC for 60 min in a water bath to generate FC clones. After the heat shock, flies were transferred to fresh vials supplemented with yeast granules. Clones were marked with plasma membrane associated CD8-GFP. The staging of the egg chambers was done 6

following Spradling 1993. For epistasis experiments, mosaic females were obtained by crossing either control FRT40A/CyO or those carrying RNAi with FRT40A or RNAi with drp1KG03815FRT40A/CyO to hs-flp; Gal80 FRT 40A/CyO; tub Gal4, UAS CD8-GFP/TM6. The protocol and time of heat shock followed by dissection was maintained identical across genotypes allowing quantification across genotypes. Immunostaining of follicle cells Ovaries dissected in Schneider’s medium were fixed with 4%PFA in PBS (or 4% PFA/PBS with 0.3% Triton X for dpERK immunostaining), permeabilized in 0.3% TritonX-100 in PBS (0.3% PBST), blocked in 2%BSA in 0.3%PBST, stained with primary antibody overnight at 4 0C, washed in 0.3%PBST, stained with secondary antibody for 1 hr in 0.3% PBST at RT, washed and mounted in Slow-fade Gold (Molecular Probes). 515 animals were dissected for each experiment at 10d after heat shock. All the ovarioles (30-35 per fly) containing clones at the stage 7-9 were imaged and experiments were repeated 2 or more times as indicated by “N” value in the figure legends. The primary antibodies used were: mouse anti-Hnt 1:10 (DSHB), mouse anti-Cut 1:10 (DSHB), mouse anti-CycB 1:10 (DSHB), rabbit anti-CycE 1:500 (Santa Cruz), mouse anti-NICD 1:100 (DSHB), rabbit anti-Ras 1:200 (Cell Signaling), mouse anti-dpERK 1:200 (Cell Signaling), rabbit anti-cleaved caspase-3 1:200 (Cell Signaling). Fluorescently coupled Streptavidin 633 (Molecular Probes) (1:1000) was used to mark mitochondria in FCs (Chowdhary et. al., 2017) and fluorescently coupled secondary antibodies (Molecular Probes) were used at dilution 1:1000. 7

Larval hemocyte isolation, immunostaining and imaging Third instar larvae expressing mito-GFP with collagen-Gal4 were dissected as described by Goyal et al., 2007. For mitochondrial fluorescence area estimation, isolated hemocytes were incubated on poly-lysine coated 18mm coverslips for 30min. Cells were fixed using 4% PFA, permeabilized by 0.1% PBST (15min) and mounted in Slow-fade Gold (Molecular Probes). Imaging was done on 63X Plan-Apochromat, NA 1.4 objective of Zeiss LSM 710/780. Experiments were repeated 3 times and a total of 90 cells were imaged for analysis. Mitochondrial membrane potential CMXRos assay CMXRos (Molecular Probes) (100nM) in Schneider’s medium was added to live ovaries for 30min and washed thrice in Schneider’s medium for 5min each. Fixation was done in 4%PFA in PBS for 15min. The experiments for control and mutant samples used the same CMXRos dye aliquot, were done at the same time and imaged on the same day. Uncoupler FCCP (10μM) (Sigma) was added after 3 washes for 30 min followed by one 5min wash and fixation followed by staining. Live hemocytes expressing mito-GFP from control and mutant were incubated with 50nM CMXRos for 10min and were imaged on the same day. For drug treatment, cells were incubated with 25μM U0126 (Sigma) for 60min or with 10μMFCCP for 30min, prior to incubation with CMXRos. To quantify mitochondria specific CMXRos signal, a thresholded mask was created based on the mito-GFP signal in ImageJ and was applied to the CMXRos channel. CMXRos fluorescence was compared using non-parametric two-tailed Kruskal Wallis, Mann 8

Whitney and Dunn’s test for statistical analysis to allow comparison of different sample sizes and non-gaussian distribution. Image acquisition and phenotypic estimation in FCs The ovaries were imaged using a Plan apochromat 40X 1.3/1.4 NA objective on Zeiss LSM 710/780. The images were grabbed with an averaging of 4 at 512 x 512 pixels. The laser power was kept similar between samples and the gain varied between 800850 for antibody staining and 600-650 for CMXRos experiments. The range indicator mode was used for acquisition of each image so that the intensity did not reach 255 on an 8-bit scale. An ROI of 5-30 cells across an optical plane was chosen in each clone for estimation of phenotypes and “n” refers to the numbers of clones in different ovarioles observed for 2 or more experiments as represented by “N” value in the figure legends. Percentage values in the figures with aggregated mitochondria, loss of Ras and dpERK immunostaining, mitochondrial morphology, loss of Hnt and oocyte nucleus positioning at the dorsal anterior position are representative of percentage of independent clones showing the defective phenotype. Cytoplasm to nuclear ratio of dpERK was measured by creating a thresholded mask for DNA and cytoplasmic signal and intensity was measured using ImageJ. Quantification of oocyte nuclear position The oocyte nucleus moves to the dorsal anterior position (away from the anteroposterior axis) at stage 7 in response to EGFR signaling. The oocyte nucleus remained 9

at the posterior location when PFCs were mutant for drp1 or EGFR signaling. If the oocyte nucleus was positioned between 0-10 degrees along the antero-posterior axis drawn in the center of the egg chamber it was considered mislocalized. Oocyte nucleus mislocalization is represented as the percentage of chambers with posterior FC clone and lack of oocyte nucleus at the dorsal anterior position out of total chambers imaged. Image analysis for estimation of fluorescence in FCs Quantification of fluorescent intensity for Ras, dpERK, NICD, Cyclin B, Cyclin E, CMXRos signal was performed using ImageJ. The optical slice was manually chosen depending on clone visibility and the GFP boundary in the MARCM system. A region of interest in one independent clone containing cell numbers between 5-30 across an optical plane was marked and average fluorescence intensity for GFP positive (clonal) and GFP negative (background) region was computed. A ratio of these intensities was obtained for each clone thus normalizing for variations in accessibility to dyes, drugs, antibodies or imaging conditions. For mitochondrial membrane potential quantification, CMXRos fluorescence intensity was also normalized to mitochondrial fluorescence intensity in the same cell in separate quantifications. The mitochondrial fluorescence intensity was estimated by measuring fluorescence intensity of Streptavidin staining in clones as a ratio to the background cells. The non-parametric two-tailed Kruskal Wallis test, Dunn’s test and Mann-Whitney test were used to compare non-gaussian distribution of fluorescence ratios between different genotypes. Images for CMXRos and dpERK are represented in the form of pseudocolor which is a 16 color LUT from Image 10

J, where red represent the highest pixels and blue represents the lowest pixels. Mitochondrial morphology quantification in hemocytes Mitochondria were selected from 3D immunostained images by thresholding and were quantified using 3D-object counter plugin in ImageJ. The cross-sectional area for individual mitochondria was measured and was compared between the wild type and drp1i cells using two-tailed Mann-Whitney test on GraphPad Prism 5.01 software. The average cross sectional area per cell was obtained from 90 cells in 3 experiments. High resolution imaging of mitochondria For visualization of mitochondrial morphology, immunostained FCs or hemocytes were imaged using a Plan-Apochromat 63X Oil, 1.4 NA objective of Zeiss LSM 800 microscope with AiryScan module (Engelmann and Weisshart, 2014). Images were deconvolved and surface rendered using AutoQuant X3 and BitPlane Imaris version 8.0 respectively. NRE-GFP and Hnt onset quantification in oogenesis The length in the antero-posterior axis of the earliest chamber in an ovariole expressing NRE-GFP and Hnt immunostaining was measured using ImageJ. These lengths were quantified and the percentage frequency is shown in bins of 30 μm size. The values were plotted using Microsoft Excel. Results 11

Fused mitochondria in drp1 mutant PFCs and hemocytes have increased mitochondrial membrane potential The transposon tagged mutation drp1KG03815 (drp1KG) was used to deplete FCs of Drp1 using mitotic recombination with the MARCM system. The drp1KG mutant is a null allele and its mitochondrial fusion phenotype is reversed by a genomic duplication or a transgene expressing drp1 (Mitra et al., 2012; Rikhy et al., 2007). GFP positive FCs homozygous for the drp1KG mutant and control FRT40A were induced under the same conditions (materials and methods). We first assessed the mitochondrial morphology in drp1KG FCs. drp1KG FCs showed an aggregated mitochondrial mass in smaller cells as compared to background heterozygous control cells and FRT40A control cells (Figure 1B-C, S1A-B). This is similar to aggregated mitochondrial morphology shown previously and consists of fused mitochondria as assessed by micro irradiation induced loss in mitochondrial membrane potential (Mitra et al., 2012). In order to confirm that the smaller drp1KG mutant PFCs were not dying by apoptosis, we stained them with Caspase antibody. drp1 mutant PFCs did not show Caspase staining (Table S1). This observation was also consistent with the known necessity of Drp1 in mediating mitochondrial fragmentation during apoptosis (Frank et al., 2001; Goyal et al., 2007). Hence the drp1 mutant cells were small, had aggregated mitochondria and did not contain apoptotic markers. Progressive cell cycle stages have different mitochondrial morphology. G1-S cells exhibit fused mitochondria and have higher mitochondrial membrane potential (Mitra et. al., 2009). Cells lacking mitochondrial fusion proteins Mfn1/2, show 12

mitochondrial fragmentation and loss of mitochondrial membrane potential (Chen et al., 2003). To check if fused mitochondria in drp1KG mutant PFCs had an altered mitochondrial potential we used the Mitotracker Red CMXRos uptake assay in the presence or absence of an ETC uncoupler FCCP. CMXRos is a fluorescent dye which concentrates in mitochondria proportional to their potential difference (Pendergrass et al., 2004). CMXRos forms thiol conjugates with peptides in the matrix thereby making it resistant to depletion on fixation with paraformaldehyde (Cottet-Rousselle et al., 2011; Susin et al., 1999). Mitochondrial potential can be abrogated by treatment of oxidative phosphorylation uncoupling drug FCCP (Carbonyl cyanide-4(trifluoromethoxy)phenylhydrazone) (Nicholls and Budd, 2000). FCCP is an ionophore and mediates passage of hydrogen ions across the mitochondrial membrane thereby dissipating the mitochondrial membrane potential. CMXRos was used to measure the mitochondrial membrane potential by estimating the fluorescence in GFP positive clone cells as a ratio to the background heterozygous cells. The fluorescence ratio was close to 1 when control FRT40A clone cells were compared to the corresponding background cells (Figure 1D,G). The fluorescence ratio of more than 1 showed greater accumulation of the potentiometric dye CMXRos and an increase in mitochondrial membrane potential in clone PFCs as compared to background control cells. drp1KG clone PFCs showed increased CMXRos fluorescence as compared to background control cells and the ratio was more than 1 (Figure 1E,G, Table S2). CMXRos fluorescence in living and fixed ovarioles containing

13

drp1KG PFCs did not show a significant difference (Figure S1C-D). Fixation allowed costaining and imaging with relevant markers along with mitochondrial potential. FCCP was used to disrupt the mitochondrial ETC and confirm that the increase in CMXRos fluorescence was specific to increase in ETC activity in drp1 mutant PFCs. FCCP disrupted the mitochondrial membrane potential in control and drp1 mutant (Figure 1D,E,G) cells and quantification of CMXRos fluorescence confirmed a drop in mitochondrial potential to control levels (Figure 1D,E,G, Table S2). In order to assess if the change in CMXRos was not due to altered mitochondrial mass, we measured mitochondrial fluorescence in different genotypes and found that this was not significantly different from each other (Figure S1E-F). We also normalized the CMXRos fluorescence to mitochondrial fluorescence and expressed this as a ratio between mutant GFP and non-GFP cells. We found that CMXRos fluorescence was significantly higher in drp1KG mutant PFCs as compared to background control cells and FRT40A clones. Hence CMXRos intensity change was not due to an alteration in mitochondrial intensity. Thus, drp1 mutant FCs had a high mitochondrial membrane potential. In order to further test if ETC activity plays a role in eliciting the increased mitochondrial potential in drp1KG PFCs, we combined RNAi against ETC Complex I protein, Pdsw (pdswi) with drp1KG and measured CMXRos uptake as compared to background control cells. pdswi expressing PFCs showed a decrease in the CMXRos fluorescence ratio. The CMXRos fluorescence in drp1KG; pdswi combination was comparable to background control cells (Figure 1F-G). Hence the increased 14

mitochondrial potential defect in drp1KG mutant cells occurred by ETC activity and complex I protein Pdsw function. To test if mitochondrial morphology and membrane potential difference was affected in other cell types we depleted drp1 in hemocytes using an RNAi (drp1i) with collagen-Gal4. Hemocytes are differentiated cells of the immune system. Primary hemocyte cultures can be generated from larval hemolymph for studying mitochondrial morphology and activity in response to signaling pathways in well-separated single cells (Goyal et al., 2007). Mitochondria were predominantly tubular in organisation in wild type hemocytes. Mito-GFP revealed an aggregated and optically unresolvable mitochondrial area in drp1i hemocytes (Figure S2A-B). Thus, loss of drp1 led to mitochondrial aggregation in both FCs and hemocytes. In hemocytes, the CMXRos uptake imaging for controls and mutants was done in living cells at the same time. CMXRos fluorescence was increased in drp1i expressing hemocytes as compared to control cells and treatment with FCCP decreased the mitochondrial potential in drp1i hemocytes (Figure 1H-I). In both FCs and hemocytes mitochondria remained clustered on FCCP treatment. In summary drp1 mutant PFCs and hemocytes had aggregated mitochondria with increased membrane potential. The increased mitochondrial membrane potential can be depleted on ETC disruption either pharmacologically by acute FCCP treatment or chronically by genetic depletion of ETC complex I protein Pdsw. drp1 mutant PFCs have elevated cytoplasmic ERK 15

EGFR signaling via Ras and ERK at stage 7-8 regulates oocyte nucleus migration to the dorso-anterior position (Figure 1A) thus patterning its dorso-ventral and antero-posterior axis. The EGFR pathway activation results in increased activated ERK immunostaining (diphospho, dpERK) in stage 8 PFCs (Dammai and Hsu, 2003). The EGFR pathway regulates mitochondrial morphology in PFCs (Mitra et al., 2012) and could also be responsible for increased mitochondrial membrane potential in drp1KG PFCs. Hence, we first tested the status of EGFR signaling components Ras and dpERK in drp1KG PFCs. Ras and dpERK immunostaining were significantly increased in stage 8 drp1KG PFCs as compared to background control cells while they were similar to nonGFP cells in FRT40A control (Figure 2A-D). Ras increase was seen on the plasma membrane and this has been correlated previously with increased active GTP bound Ras (Arozarena et al., 2000). This is also in accordance with increased dpERK (Figure 2C-D). The activated ERK however was both cytoplasmic and nuclear in localization in drp1 mutant PFCs at stage 8 in contrast to primarily nuclear in control FRT40A cells (Figure 2C,E). The predominantly cytoplasmic dpERK as compared to background control cells was also seen at earlier stages 2-5 (Figure 2F-G). Thus EGFR was activated and led to increased Ras and dpERK in drp1 mutant FCs and PFCs. However dpERK was enriched in the cytoplasm in PFCs and this might interact with mitochondria to affect the mitochondrial membrane potential.

ERK depletion in drp1 mutant PFCs and hemocytes suppressed the mitochondrial membrane potential defect 16

EGFR loss of function causes mitochondrial aggregation in PFCs (Mitra et al. 2012) and ERK activates Drp1 by phosphorylation to cause fragmentation in cancer cells (Kashatus et al. 2015). In order to check if drp1KG mitochondrial membrane potential and morphology defects were regulated by Ras/ERK accumulation, we depleted them using RNAi against Ras (rasi) and ERK (erki). As expected, Ras and pERK antibody staining were depleted in PFCs expressing rasi and erki respectively (Figure S3A-B). rasi lowered but did not completely deplete Ras (Figure S3A) or dpERK (Table S1-2) as compared to erki. rasi was weaker in depletion of Ras mRNA as indicated by a smaller change in Ras immunostaining. erki on the other hand efficiently depleted dpERK (Figure S3B). We therefore analysed PFCs containing the drp1KG; erki combination for effect on mitochondrial membrane potential and found that the mitochondrial membrane potential was decreased as compared to drp1KG alone (Figure 3A-B, S4A-B). EGFR-DN expression also lowered the membrane potential in drp1KG PFCs (Table S1-2). Moreover ERK depletion alone lowered the mitochondrial membrane potential as compared to FRT40A control cells (Figure 3A-B, S4A-B). The rasi depletion did not lower the mitochondrial membrane potential like erki in drp1 mutant PFCs and this was probably due to the RNAi being less efficient in reducing Ras (Table S1-2). We also depleted activated ERK in control and drp1i hemocytes by adding an inhibitor to MEK (an upstream component of ERK) (Newton et al., 2000) called U0126 (1,4-diamino-2,3-dicyano- 1,4-bis[2-aminophenylthio] butadiene). CMXRos fluorescence was lowered in U0126 treated, drp1i and control cells (Figure 3C-F). We did not find a 17

significant change in mitochondrial organization in this acute treatment (Figure 3C-F). We next assessed if the change in mitochondrial membrane potential in drp1 mutant PFCs correlated with a change in mitochondrial morphology. We observed mitochondrial morphology using confocal microscopy and super-resolution microscopy using the Airyscan detector. EGFR depletion resulted in aggregated mitochondrial morphology in FCs (Mitra et al 2012). However erki and rasi did not give an appreciable change in mitochondrial organization as compared to neighboring control cells (Figure 3G, S4). We further visualized mitochondrial morphology in drp1KG PFCs additionally depleted of Ras and ERK. The aggregated mitochondrial phenotype in drp1KG PFCs was not present in the drp1KG; rasi and the drp1KG; erki combination and mitochondria were present on all sides of the nucleus (Figure 3G, Figure S4C-G, Table S2). The cell size was also similar to background control cells. Hence mitochondrial arrangement and morphology were similar to control in cells depleted of both drp1 and ras/erk (Figure 3G, S4). Mitochondrial morphology is maintained via a balance between fusion and fission. It is likely that in the absence of fission protein Drp1, the dispersed mitochondrial morphology may occur in drp1KG; erki by inactivation of mitochondrial fusion proteins. Thus, regulation of mitochondrial fragmentation in PFCs by the EGFR-ERK pathway is not entirely dependent on Drp1 and other proteins such as those involved in mitochondrial fusion could be regulated by the pathway. In summary, the EGFR pathway inhibition by EGFR-DN, MEK and ERK diminished the elevated membrane potential in drp1 mutant hemocytes and PFCs. ERK depletion alone also reduced the mitochondrial membrane potential in both cell types. It 18

was therefore possible that cytoplasmic ERK accumulation in drp1 mutant PFCs regulated the increase in mitochondrial membrane potential.

Mitochondrial membrane potential depletion does not impact the EGFR driven oocyte nucleus patterning Since ERK depletion showed a drop in mitochondrial membrane potential in ovarioles with drp1 mutant PFCs, we investigated if mitochondrial ETC activity depletion also had an effect on EGFR signaling. ETC complex I mutant pdsw allowed an analysis of impact of ETC activity reduction on EGFR signaling. We first checked the mitochondrial architecture in pdswi and drp1KG; pdswi PFCs and found that even though the mitochondrial membrane potential was lowered in these cells (Figure 1F-G), the mitochondrial aggregation phenotype remained unchanged (Figure 4A). As mentioned earlier, EGFR signaling regulates dorso-anterior movement of the oocyte nucleus. The oocyte nucleus failed to migrate to the dorso-anterior position by stage 8 in ovarioles with drp1 mutant PFCs (Figure 4B, Table S2) showing a loss of EGFR pathway mediated oocyte patterning even with elevated Ras and dpERK levels (Figure 2). Loss of oocyte nucleus migration was also seen in erki and was unchanged in the drp1KG; erki combination (Figure 4B, Table S2). A similar phenotype of dpERK accumulation and loss of oocyte patterning is also seen in polarity mutants of scrib and dlg (Li et al., 2009). Thus, both increase and decrease in ERK levels failed to bring about oocyte axis determination and this argued for the requirement of activated ERK at appropriate levels along with nuclear localisation downstream of EGFR for oocyte patterning. 19

We next assessed EGFR signaling induced oocyte nucleus migration on ETC depletion in pdswi and drp1KG; pdswi PFCs. pdswi expressing PFCs deficient for the ETC did not show a defect on oocyte nucleus migration thus confirming that mitochondrial ETC did not interact or feedback onto the EGFR driven oocyte patterning. The drp1 mutant oocyte nucleus migration defect remained unchanged in drp1KG; pdswi PFCs (Figure 4C). This was consistent with the lack of change in the increased dpERK accumulation in drp1KG; pdswi combination (Figure 4D-E) as compared to drp1KG (Figure 2C-D). dpERK levels in pdswi expressing PFCs were similar to background control cells (Figure 4D-E). Also an acute FCCP treatment to disrupt mitochondrial potential on drp1KG or wild type ovarioles at stage 8 did not alter dpERK immunostaining (Table S1-S2). Mitochondrial potential decrease in pdswi did not remove dpERK accumulation and alter the oocyte patterning defect in drp1 mutant PFCs. Thus, higher mitochondrial membrane potential in drp1 mutant PFCs was not responsible for dpERK accumulation and oocyte nucleus mislocalization.

ERK reduction in drp1 mutant PFCs suppresses the defect in Notch mediated differentiation We tested whether ERK dependent increase in mitochondrial membrane potential could affect Notch signaling in drp1 mutant PFCs and if mitochondrial potential reduction could independently affect Notch signaling. First we assessed if the Notch signaling defect in drp1 mutant PFCs was affected by elevation in ERK/Ras levels. As previously reported, Notch induced transcription factor Hnt was absent in drp1 mutant 20

PFCs (Figure 5A) (Mitra et al., 2012). Notch signaling driven loss of transcription factor Cut was also aberrant in drp1 mutant PFCs and Cut remained in these cells (Table S12). FRT40A control cells showed the same Hnt and Cut staining as the neighboring cells (Figure S5A-B). Plasma membrane accumulation of NICD can be monitored by a specific antibody and quantified as a ratio to the background control FCs. The ratio of the plasma membrane NICD fluorescence in drp1KG mutant PFCs to the background PFCs was elevated (Figure 5A,E). FRT40A cells displayed comparable NICD immunostaining in GFP positive and background control cells (Figure S5A). The drp1KG; erki combination did not show the NICD accumulation and Hnt depletion seen in drp1KG PFCs (Figure 5B-C, Table S2). Thus, the Notch pathway mediated differentiation defects were not seen in drp1 mutant PFCs additionally depleted of erk. The drp1KG; rasi combination also did not show these Notch defects but Hnt expression was seen at a lower percentage than the drp1KG; erki combination (Figure 5B-C, Table S2). The Notch pathway regulates the mitosis to endocycle switch by downregulating Cyclin levels through proteins such as Dacapo, APC and String (Schaeffer et al., 1998; Shcherbata et al., 2004). Since drp1KG FCs had defects in the Notch pathway, we checked the status of Cyclins in these cells. Cyclin B and E are downregulated in stage 7 endocycling FCs. drp1KG FCs at stage 7-9 clones did not enter the endocycle and had increased Cyclin B and E antibody staining consistent with loss of Notch activation in these cells. FRT40A control clones had comparable levels of Cyclin B and E immunostaining in GFP positive and neighbouring cells (Figure S5B). In the drp1KG; erki 21

combination, Notch activation was seen while Cyclin B and E staining decreased and was similar to background control cells (Figure 5D-G, Table S2). This decrease in Cyclin B and E was not significant for the drp1KG; rasi combination as compared to drp1KG alone, even though there was a decrease in averages (Figure S6A-D). We attribute this difference to rasi being less efficient in removing Ras mRNA. In summary elevated Ras/ERK in drp1 mutant PFCs was responsible for loss of Notch mediated differentiation.

Mitochondrial membrane potential depletion suppresses the Notch signaling defect in drp1 mutant PFCs and enhances Notch signaling in wild type ovarioles ERK depletion in drp1 mutant PFCs reduced the mitochondrial membrane potential and did not show the Notch signaling defect. We therefore tested if mitochondrial membrane potential depletion by ETC mutant pdsw in drp1 mutant PFCs could alleviate the Notch signaling defect. The drp1KG; pdswi combination showed appearance of Hnt in greater number of clones as compared to drp1KG alone (Figure 6A). Hnt immunostaining was present in the drp1KG; pdswi combination at a weaker intensity than the neighbouring control cells and similar to stage 6 nuclei when Notch signaling is beginning to be activated. NICD accumulation seen in drp1KG PFCs was also absent in the drp1KG; pdswi combination (Figure 6B,D, Table S2). The Notch pathway is activated at stage 6 and Hnt expression correlates with increase in nuclear size due to activation of endocycle. Hence we decided to quantify nuclear size as a 22

readout of endocycle activation by Notch signaling. We found that drp1KG nuclei were approximately half the size of neighbouring control cells and FRT40A GFP positive cells (Figure S7A-B). Nuclear size in drp1KG; erki and drp1KG; rasi was comparable to the control cells, thus showing that Notch signaling was present and endocycling was appropriately activated (Figure S7A-B). Interestingly stage 8 drp1KG; pdswi nuclei were comparable in area to stage 6 FRT40A control nuclei (Figure S7A-B). This showed that even though Notch signaling was activated in the drp1KG; pdswi combination, it was delayed as compared to background control cells. This may be due to delay in Hnt appearance on pdsw depletion only after mitotic clone generation in the MARCM strategy. Thus, pdsw depletion in drp1KG PFCs could suppress the drp1KG phenotype of loss of Notch mediated differentiation. Since pdswi lowered the membrane potential (Figure 1F-G) but did not change the mitochondrial organization in drp1KG FCs (Figure 4A), it was possible that mitochondrial potential depletion specifically brought about Notch mediated differentiation. We therefore analyzed Notch phenotypes on acute treatment of FCCP for 30 min in drp1KG PFCs and did not find a significant change in Hnt appearance and NICD accumulation (Figure 6C-D). FCCP treatment however resulted in Hnt appearance at an earlier stage in wild type ovarioles (Figure S7C). A distribution of chamber sizes showed the onset of Hnt in stages with smaller sizes corresponding to stage 5 chambers in FCCP treated ovarioles as compared to stage 6-7 in controls (Figure S7D). We further found that the Notch response element construct NRE-GFP 23

(Jia et al., 2015; Jouandin et al., 2014) showed an early appearance at Stage 5 instead of Stage 6 in wild type along with Hnt when acutely treated with FCCP for 30 min (Figure 6E-F). NICD depletion was also seen from the plasma membrane on FCCP treatment (Figure 6G). This suggested that mitochondrial potential decrease in wild type chambers in the presence of Drp1 could activate Notch signaling. These data together show that mitochondrial membrane potential depletion is required for Notch pathway activation. Acute reduction of mitochondrial membrane potential activated Notch signaling at an earlier stage in wild type ovarioles. Genetic depletion of ETC component pdsw in drp1 mutant PFCs also showed activation of Notch signaling even though it was slower than background control cells. In summary, increased Ras-ERK pathway in drp1KG PFCs with aggregated mitochondria increased mitochondrial potential thus resulting in loss of Notch mediated differentiation. Discussion Developmental signaling pathways integrate cues at the cellular level by coordinating activities of various subcellular organelles to determine cell fate. Here we show that coregulation of two well characterized developmental pathways, EGFR and Notch, can occur through mitochondrial morphology and mitochondrial membrane potential during FC differentiation and oocyte patterning in oogenesis. EGFR and Notch pathways work antagonistically in regulation of Drosophila eye cell fate (Rohrbaugh et al., 2002) and together in cell fate determination of photoreceptor R7 (Tomlinson and Struhl, 2001). In follicle stem cells EGFR phosphorylates Groucho which inhibits Notch 24

driven differentiation of prefollicle cells (Johnston et al., 2016). Our studies show that EGFR signaling regulates mitochondrial fragmentation (Mitra et al.,2012) and mitochondrial membrane potential thereby affecting Notch activity. drp1 mutant PFCs with aggregated mitochondria and elevated mitochondrial potential result in loss of both EGFR and Notch signaling leading to aberrant oocyte patterning and PFC differentiation (Figure 7). The oocyte patterning is lost despite increased Ras/ERK in the drp1 mutant cells possibly due to cytoplasmic retention of dpERK. Depleting drp1 mutant PFCs of ETC activity with pdsw RNAi suppresses the Notch loss of function phenotype with no effect on aberrant EGFR signaling and oocyte mispatterning. While activated ERK enters the nucleus to promote gene expression, it can be enriched in the cytoplasm or organelles such as mitochondria (Horbinski and Chu, 2005). Activated ERK is seen in mitochondria during brain development (Alonso et al., 2004). Fused mitochondria in the drp1 mutant cells may act as docking platform for activated ERK and retain it outside the nucleus. dpERK shows a predominant nuclear staining in wild type as opposed to cytoplasmic in drp1KG PFCs and FCs. Absence of nuclear ERK is a possible way for failure of signaling and oocyte nucleus localization to dorso-anterior position. Cytoplasmic ERK can dimerize and activate different cytoplasmic substrates (Casar et al., 2008; Ebisuya, 2005) and result in modification of the mitochondrial activity (Monick et al., 2008). Activated ERK in the cytoplasm may interact with the mitochondrial ETC proteins on fused mitochondria and increase the membrane potential. Our data for MEK-ERK depletion shows a decrease in 25

mitochondrial potential in hemocytes and FCs, also suggesting that the pathway could control mitochondrial activity directly. In order to understand the mechanism by which mitochondrial membrane potential regulates Notch signaling, we discuss here the possible interactions between Notch and mitochondrial ETC outputs of ROS and ATP. drp1 mutant PFCs also had increased ROS and pAMPK, the sensor activated during ATP depletion. ROS remained elevated in the drp1KG; erki combination showing that alleviation of Notch signaling in these FCs was not due to ROS. ROS and pAMPK were also elevated in pdsw mutant PFCs and the elevation persisted in the drp1KG; pdswi mutant PFCs (data not shown). Hence, ETC reduction activates Notch signaling via a pathway independent of ROS or ATP concentration in PFCs. The impact of membrane potential increase on Notch deactivation can also be through calcium or activation of PARL family of proteases resulting in NICD cleavage. Membrane potential difference across the mitochondrial inner membrane allows calcium influx (Gunter and Pfeiffer, 1990). Lowering of mitochondrial membrane potential using an uncoupler releases stored calcium (Thayer and Miller, 1990). Released calcium can be involved in Notch activation as shown in cardiomyocyte differentiation (Kasahara et. al, 2012). NICD localizes to mitochondria to regulate mitochondrial genes (Xu et al., 2015) and could be cleaved into an activated form by mitochondrial intermediate peptidase (Lee et al., 2011). Hence, future studies could test if mitochondria can directly or indirectly activate the Notch pathway via calcium. 26

drp1 mutant FCs were present in multiple layers. EGFR signaling through Ras/MEK/ERK is important for basal polarity in follicle stem cells and its reduction is essential for induction of apical polarity in prefollicle cells (Castanieto et al., 2014). Thus it is likely that cytoplasmic increase of dpERK in drp1 depleted FCs affects apical polarity and results in multilayering. The apical polarity protein aPKC also plays a role in microtubule activation and movement of the oocyte nucleus (Tian and Deng, 2008). A study of polarity pathways in these cells will be important to understand the mechanism of its reversal in drp1 mutant cells additionally depleted of Ras/ERK. Mitochondrial morphology in most cells exists as a balance between tubular, fragmented and aggregated. Stem cells undergo a distinct change in both inner and outer mitochondrial membrane shape as they differentiate (Chung et al., 2007; Facucho-Oliveira et al., 2009; Mandal et al., 2011). EGFR regulated fragmented mitochondrial morphology in PFCs (Mitra et al., 2012) is also in agreement with higher Ras/ERK driven mitochondrial fission activity in mammalian cancer cells (Kashatus et al., 2015). Loss of mitochondrial fusion in drp1 mutant PFCs has been previously shown to partially suppress their Notch signaling defect (Mitra et al., 2012). In our study, ERK depletion alone did not change mitochondrial morphology appreciably. While higher resolution live imaging or electron microscopic analyses will discern mitochondrial morphology more completely, the aggregated mitochondrial structure is resolved by codepletion of Drp1 and Ras/ERK. We reason that this rescue of mitochondrial organization could occur by multiple pathways. ERK could directly activate 27

mitochondrial fusion in these cells but this is not in agreement with mammalian literature, which shows reduced mitofusin Mfn1 activity by ERK induced phosphorylation. The residue phosphorylated in mitofusin Mfn1 by ERK2 is not present in mitofusin Mfn2 or its Drosophila orthologue Marf (Pyakurel et al., 2015). However, the stress induced p38 MAPK2 mediated Mfn2 phosphorylation residue (Leboucher et al., 2012) is conserved in Marf and could result in its inactivation and degradation. The ERK2 induced Drp1 phosphorylation residue is conserved from Drosophila to mammals suggesting that it is a key regulator of mitochondrial morphology downstream of ERK (Kashatus et al., 2015). Taken together ERK depletion in drp1 mutant PFCs may inactivate mitochondrial fusion by an unidentified mechanism and Drosophila ERK may regulate both Marf and Drp1 activity by phosphorylation such that there are more fragmented mitochondria. A second possibility is that a different p38K based stress induced mechanism results in reversal of fusion in drp1 and erk mutant cells. Future studies on the activation of stress pathways in drp1 mutant cells will reveal the mechanism by which they affect cell fate in combination with Ras/ERK. These interactions between mitochondrial architecture, potential and Ras/ERK point to their role in regulating the threshold of signaling in the EGF pathway. This study establishes mitochondrial membrane potential as key interactor of EGF signaling alongside the already known players: ROS and morphology proteins. Along with a role in proliferation alluded to by number of studies, the mitochondria-ERK interaction is essential for cell fate determination. In addition to control at transcriptional level, ERK 28

can modulate Notch via the mitochondrial membrane potential in the drp1 mutant. In conclusion, our study warrants an analysis of mitochondrial membrane potential in tumours with a dysregulated EGFR-Ras-ERK pathway.

Acknowledgements: We thank L.S. Shashidhara, Girish Ratnaparkhi, Anuradha Ratnaparkhi, Girish Deshpande and RR lab members for critical comments on the data and the manuscript. We thank the Drosophila fly and Microscopy facility for help with media, stock maintenance and microscopy. We thank Zeiss, India for imaging with Airyscan on the LSM800 demonstration system. DT thanks UGC for the graduate fellowship. RR thanks IISER, Pune, DBT and DST for funding. Author contributions: DT and RR designed the experiments in the FCs. DT performed the experiments and did the analysis in the FCs. RC, DT and RR designed the experiments in hemocytes. RC and DT performed and analysed the experiments in hemocytes. DT, RC, KM and RR analyzed the data and wrote the manuscript. Figure 1: Aggregated mitochondria in drp1 mutant PFCs have increased mitochondrial membrane potential. (A) Drosophila ovariole structure. Drosophila ovariole consists of chambers containing germline and follicle cells in successive ages. Germline and follicle stem cells are present in stage 1. Follicle cells in stage 2-6 29

enclosing the oocyte and the nurse cells undergo mitosis; Notch pathway is activated and cells enter endocycle at stage 7. Stage 8 has three differentiated populations of follicle cells, follicle cells covering the oocyte are known as posterior follicle cells. EGFR pathway is activated by the ligand Gurken in posterior follicle cells at stage 7 and is responsible for antero-dorsal movement of the oocyte nucleus (A). (B-C) Mitochondria are aggregated in drp1KG PFCs. drp1KG homozygous mutant cells (CD8-GFP; yellow boundary) are generated by mitotic recombination in a heterozygous background. Mitochondria labelled by fluorescent Streptavidin (grey) are aggregated in GFP positive drp1KG PFCs (97%, percentage of chambers with fused mitochondria; n=47, N=4) (B). Mitochondria are dispersed in GFP positive FRT40A clones and morphology is similar in GFP positive to GFP negative cells (n=40, N=3) (C) . (D-G) drp1KG PFCs have increased and pdswi PFCs have decreased mitochondrial membrane potential. CMXRos intensity (pseudocolored bar, first panel) ratio in control FRT40A GFP to non-GFP cells is close to 1 with and without FCCP treatment (n= 13,11, N= 3,3) (D). drp1KG FCs (n=22, N=5) show higher CMXRos fluorescence as compared to background control cells (ratio higher than 1) and this is not seen in the presence of FCCP (n=30, N=4) (E). CMXRos fluorescence ratio decreases in drp1KG; pdswi as compared to background control cells (n=21,N=4) (F). CMXRos intensity ratio quantification shows a significant decrease on FCCP addition and in pdswi (n=30,8, N=4,4; FRT40A (n=13, N=3), ***,P < 0.001, *,P < 0.05,two-tailed Kruskal Wallis and Dunn’s test) (G). 30

(H-I) drp1i hemocytes have increased mitochondrial membrane potential. drp1i hemocytes expressing mito-GFP have increased CMXRos fluorescence as compared to control cells (H). FCCP treatment causes reduction of CMXRos intensity in control and drp1i expressing cells (I) (n=90,90,89,97 cells, N=3 experiments each; ***,P < 0.001, two-tailed Kruskal Wallis and Dunn’s test). The pseudocolor scale is the 16 color LUT from Image J where highest intensity pixel is red and lowest is blue. Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Each data point in the box plot is an average from 5-30 cells in independent ovarioles. Scale Bar: 10μm FCs and 5μm hemocytes. st= stage, n=PFC clones in independent ovarioles/n= number of individual hemocytes, N=Experimental replicates. Figure 2: drp1 mutant PFCs have increased Ras and dpERK. (A-B) drp1KG PFCs have increased Ras. Ras (intensity pseudocolored scale) (n=16, N=4) antibody staining shows significantly more red pixels in drp1KG PFCs implying increased intensity in comparison to neighbouring control cells (CD8GFP; green). FRT40A control (n=13) does not show any difference as compared to neighbouring control cells (A). Quantification of relative fluorescence is shown in (B) (n=13,16, N=4 ,**,P <0.01, two tailed Mann-Whitney test). (C-G) drp1KG FCs have increased cytoplasmic dpERK. dpERK (intensity pseudocolored scale, DNA; blue, CD8GFP; green) immunostaining is 31

higher in drp1KG mutant cells as compared to neighbouring control cells (C). Quantification for relative fluorescence intensity shown in (D) (n=22,23, N=5,***,P < 0.001). FRT40A cells showed a high nuclear dpERK signal (grey in the merged image, DNA magenta in merged image) while in drp1KG mutant cells dpERK is mainly cytoplasmic. Quantification for cytoplasmic to nuclear ratio (C/N) is shown in (E) (n=9,8, N=5,*,P < 0.05). In stages 3-5 drp1KG mutant FCs show an increased cytoplasmic to nuclear dpERK abundance as compared to FRT40A control cells (F). Quantification in (G) (n=10, 10, N=3,**,P < 0.01). Statistical test used was two-tailed Mann-Whitney test. The pseudocolor scale is the 16 color LUT from Image J where highest intensity pixel is red and lowest is blue. Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Each data point in the box plot is an average from 5-30 cells in independent ovarioles. Scale Bar: 10μm. n=FC clones in independent ovarioles, N=Experimental replicates. Figure 3: MEK-ERK depletion reduces mitochondrial membrane potential and suppresses the increased mitochondrial potential defect in drp1 PFCs and hemocytes. (A-B) erki causes reduction in mitochondrial membrane potential in drp1KG PFCs. drp1KG; erki and erki PFCs show reduced CMXRos fluorescence (intensity pseudocolored scale) as compared to drp1KG (A). Quantitative analysis of CMXRos fluorescence ratio shows a significant reduction in drp1KG; erki as compared to drp1KG 32

alone. erki shows decrease in CMXRos fluorescence compared to FRT40A control (B). (n=11,8,10,12, N=5,3,3,3, **,P < 0.01, *,P < 0.05, two-tailed Kruskal Wallis and Dunn’s Test). (C-F) MEK inhibition by U0126 reduces the mitochondrial membrane potential in hemocytes. Inhibition of MEK by U0126 in wild type (C,E) as well as drpi hemocytes (D,F) shows decreased CMXRos fluorescence (intensity pseudocolored scale) (n=90,89,91,90, N=3 experiments, ***,P < 0.001). Statistical significance is calculated with two-tailed Mann-Whitney test. (G) Mitochondrial aggregation in drp1KG PFCs is suppressed by Ras/ERK depletion. rasi and erki expression with drp1KG causes loss of mitochondrial aggregation (red; percentage aggregation depicted in each panel, n=47,62,62,24,21, N=3 for all) and appearance of dispersed mitochondrial punctae. Percentage mitochondrial aggregation phenotype is represented. Yellow line: clone boundary, white dashed line: cell boundary. The pseudocolor scale is the 16 color LUT from Image J where highest intensity pixel is red and lowest is blue. Data is presented as

box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Each data point in the box plot is an average from 5-30 cells in independent ovarioles. ns =not significant, Scale Bar: 10μm FCs and 5μm hemocytes. n=PFC clones in independent ovarioles/n= number of individual hemocytes, N=Experimental replicates. Figure 4: Mitochondrial membrane potential decrease by pdswi does not affect EGFR driven oocyte nucleus localization in the dorso-anterior position. (A) drp1KG; pdswi mitochondrial morphology is similar to drp1KG. drp1KG; pdswi PFCs show 33

aggregated mitochondria (87.5%, percentage of chambers with fused mitochondria; n=47, N=4). pdswi mitochondria are similar to neighbouring control cells (n=20, N=3)(A). (B) Oocyte nucleus mislocalization to a central position instead of dorso-anterior is observed in drp1KG (41% mislocalization, n=47, N=3) and drp1KG; erki (39% mislocalization, n=62, N=3) and erki (31.5% mislocalization, n=24, N=3) FRT 40A (0% mislocalization, n=40,N=3). (C-E) Oocyte nucleus mislocalization to central position in drp1KG chambers is not suppressed in drp1KG; pdswi (40.4% mislocalization, n=47, N=3). (Oocyte nucleus; Orange dash line) (C). Increased dpERK in drp1KG is not reversed in drp1KG; pdswi as antibody fluorescence ratio is more than 1 while dpERK in pdswi is similar to neighboring control cells (n=18, N=3, n=13, N=2) Quantification in (E). The pseudocolor scale is the 16 color LUT from Image J where highest intensity pixel is red and lowest is blue. Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Each data point in the box plot is an average from 5-30 cells in independent ovarioles. ns=not significant, Scale Bar: 10μm. n=PFC clones in independent ovarioles, N=Experimental replicates. Figure 5: Increased Ras/ERK leads to loss of Notch mediated differentiation in drp1KG PFCs. (A) drp1KG PFCs show loss of Notch mediated differentiation. Notch induced Hnt (red, first panel) is absent (81% show Hnt loss, n= 26, N=3) in drp1KG PFCs and, NICD (red, second panel) accumulates on the plasma membrane (quantification in C, n=15,n=3) in comparison to neighbouring cells. Percentage phenotype is shown in 34

the figure. (B-C) Notch mediated signaling defect is not seen in drp1KG PFCs depleted of rasi and erki. Hnt absence in drp1 PFCs is recovered in drp1KG; rasi (% Hnt loss is 52%, % Hnt loss is reduced as compared to drp1 (A), n=31, N=4) and drp1KG; erki (24% Hnt loss n=29, N=3) (B) and not seen in rasi (0% Hnt loss n=22, N=2) and erki (0% Hnt loss, n=14, N=3) (yellow line: clone boundary). NICD fluorescence ratio quantification shows that the increased staining in drp1KG PFCs when additionally depleted of ERK (C) (n=15,21,15,10,10, N=3,3,6,2,2, ***,P < 0.001,**,P < 0.01,two-tailed Kruskal Wallis and Dunn’s test). Percentage defect is shown in each panel. (D-G) Increased Cyclin E and B is not seen in drp1KG cells depleted of ERK. Increased Cyclin E and B immunostaining in stage 8 drp1KG PFCs is not seen in stage 8 erki or in stage 8 drp1KG, erki (D, F). Quantification of Cyclin E fluorescence intensity ratio of the mutant clone to the background shows a significant decrease in drp1KG; erki (E) (n=7,20,7,7, N= 2,2,4,6, **,*,P < 0.01, P < 0.05, two-tailed Kruskal Wallis and Dunn’s test). Quantification of Cyclin B fluorescence intensity ratio of the clone to the background shows a significant decrease in stage 8 drp1KG; erki (G) (n=9,14,12,11, N=2,3,3,4, ***,P < 0.001, **,P< 0.01, *,P<0.05, two-tailed Kruskal Wallis and Dunn’s test). Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Each data point in the box plot is an average from 5-30 cells in independent ovarioles. ns=not significant, Scale Bar: 10μm. n=PFC clones from independent ovarioles, N=Experimental replicates.

35

Figure 6: Mitochondrial membrane potential depletion enhances Notch signaling in drp1 mutant and wild type FCs. (A-D) drp1KG; pdswi shows Notch signaling. drp1KG; pdswi reduces the Hnt loss (% Hnt loss is 38%, n=29, N=4) (A) and NICD accumulation (B) as compared to drp1KG. FCCP treatment of drp1KG mutant clones does not show change in Hnt loss (86% is Hnt loss, n=7, N=7) (C). Quantitative analysis of NICD plasma membrane accumulation as a ratio to the background control cells shows decreased intensity for pdswi and drp1KG; pdswi as compared to drp1KG (D) (n=15,24,14,7, N=3,4,3,3, *** P < 0.001, two-tailed Kruskal Wallis and Dunn’s test). (EG) FCCP treated wild type ovarioles show Notch activation at an earlier stage. NREGFP containing ovarioles show NRE-GFP (green) activation in stage 6-7 and on FCCP treatment NRE-GFP appears at stage 5 (96% chambers show early onset, n=26, N=2). CMXRos staining is decreased in fluorescence on FCCP treatment (E). Quantitative analysis of chamber size of Hnt onset shows a shift to the left to a lower size as compared to untreated ovarioles (Control; n= 30,N=2) (F). NICD plasma membrane staining in wild type PFCs on FCCP treatment is lowered (n=7, N=3) (G). The pseudocolor scale is the 16 color LUT from Image J where the highest intensity pixel is red and lowest is blue. Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Each data point in the box plot is an average from 5-30 cells in independent ovarioles. ns =not significant, Scale Bar: 10μm. n=PFC clones in independent ovarioles, N=Experimental replicates. 36

Figure 7: Schematic showing an interaction between mitochondrial morphology and potential in drp1 depleted PFCs and the EGFR and Notch signaling pathways. The EGFR-Ras-ERK pathway controls mitochondrial membrane potential in PFCs and maintains active Notch signaling (left panel). Fused mitochondria and increased mitochondrial membrane potential in drp1 depleted PFCs lead to faulty dpERK accumulation in the cytoplasm and increased Cyclin B/E accumulation and loss of differentiation due to inactive Notch pathway (right panel). References Alonso, M., Melani, M., Converso, D., Jaitovich, A., Paz, C., Carreras, M.C., Medina, J.H., Poderoso, J.J., 2004. Mitochondrial extracellular signal-regulated kinases 1/2 (ERK1/2) are modulated during brain development. J. Neurochem. 89, 248–256. doi:10.1111/j.1471-4159.2004.02323.x Arozarena, I., Aaronson, D.S., Matallanas, D., Sanz, V., Ajenjo, N., Tenbaum, S.P., Teramoto, H., Ighishi, T., Zabala, J.C., Silvio Gutkind, J., Crespo, P., 2000. The Rho family GTPase Cdc42 regulates the activation of Ras/MAP kinase by the exchange factor Ras-GRF. J. Biol. Chem. 275, 26441–26448. doi:10.1074/jbc.M002992200 Ashrafian, H., Docherty, L., Leo, V., Towlson, C., Neilan, M., Lygate, C.A., Hough, T., Townsend, S., Williams, D., Wells, S., Glyn-jones, S., Land, J., Barbaric, I., Lalanne, Z., Denny, P., Bhattacharya, S., Griffin, J.L., Hargreaves, I., Fernandezfuentes, N., Cheeseman, M., Watkins, H., Dear, T.N., 2010. A Mutation in the 37

Mitochondrial Fission Gene Dnm1l Leads to Cardiomyopathy 6, 1–18. doi:10.1371/journal.pgen.1001000 Boerner, J.L., Demory, M.L., Silva, C., Parsons, S.J., 2004. Phosphorylation of Y845 on the Epidermal Growth Factor Receptor Mediates Binding to the Mitochondrial Protein Cytochrome c Oxidase Subunit II. Mol. Cell. Biol. 24, 7059–7071. doi:10.1128/MCB.24.16.7059-7071.2004 Bollu, L.R., Ren, J., Blessing, A.M., Katreddy, R.R., Gao, G., Xu, L., Wang, J., Su, F., Weihua, Z., 2014. Involvement of de novo synthesized palmitate and mitochondrial EGFR in EGF induced mitochondrial fusion of cancer cells. Cell Cycle 13, 2415–30. doi:10.4161/cc.29338 Casar, B., Pinto, A., Crespo, P., 2008. Essential Role of ERK Dimers in the Activation of Cytoplasmic but Not Nuclear Substrates by ERK-Scaffold Complexes. Mol. Cell 31, 708–721. doi:10.1016/j.molcel.2008.07.024 Castanieto, A., Johnston, M.J., Nystul, T.G., 2014. EGFR signaling promotes selfrenewal through the establishment of cell polarity in Drosophila follicle stem cells. Elife 3, 1–18. doi:10.7554/eLife.04437 Cereghetti, G.M., Stangherlin, a, Martins de Brito, O., Chang, C.R., Blackstone, C., Bernardi, P., Scorrano, L., 2008. Dephosphorylation by calcineurin regulates translocation of Drp1 to mitochondria. Proc. Natl. Acad. Sci. U. S. A. 105, 15803–8. doi:10.1073/pnas.0808249105 38

Chen, H., Detmer, S. a., Ewald, A.J., Griffin, E.E., Fraser, S.E., Chan, D.C., 2003. Mitofusins Mfn1 and Mfn2 coordinately regulate mitochondrial fusion and are essential for embryonic development. J. Cell Biol. 160, 189–200. doi:10.1083/jcb.200211046 Chiaradonna, F., Gaglio, D., Vanoni, M., Alberghina, L., 2006. Expression of transforming K-Ras oncogene affects mitochondrial function and morphology in mouse fibroblasts. Biochim. Biophys. Acta 1757, 1338–56. doi:10.1016/j.bbabio.2006.08.001 Chowdhary, S., Tomer, D., Dubal, D., Sambre, D., Rikhy, R., 2017. Analysis of mitochondrial organization and function in the Drosophila blastoderm embryo. Sci. Rep. 1–17. doi:10.1038/s41598-017-05679-1 Chung, S., Dzeja, P.P., Faustino, R.S., Perez-terzic, C., Behfar, A., Terzic, A., 2007. Mitochondrial oxidative metabolism is required for the cardiac differentiation of stem cells. Nat Clin Pr. Cardiovasc Med. 4, 1–12. doi:10.1038/ncpcardio0766.Mitochondrial Cottet-Rousselle, C., Ronot, X., Leverve, X., Mayol, J.F., 2011. Cytometric assessment of mitochondria using fluorescent probes. Cytom. Part A 79 A, 405–425. doi:10.1002/cyto.a.21061 Cox, R.T., Spradling, A.C., 2003. A Balbiani body and the fusome mediate mitochondrial inheritance during Drosophila oogenesis. Development 130, 1579– 39

1590. doi:10.1242/dev.00365 Dammai, V., Hsu, T., 2003. EGF-dependent and independent activation of MAP kinase during Drosophila oogenesis. Anat. Rec. A. Discov. Mol. Cell. Evol. Biol. 272, 377– 82. doi:10.1002/ar.a.10037 Ebisuya, M., 2005. The duration, magnitude and compartmentalization of ERK MAP kinase activity: mechanisms for providing signaling specificity. J. Cell Sci. 118, 2997–3002. doi:10.1242/jcs.02505 Engelmann, R., Weisshart, K., 2014. Airyscanning. G.I.T Imaging Microsc. 20–21. Facucho-Oliveira, J.M., Facucho-Oliveira, J.M., St John, J.C., St John, J.C., 2009. The relationship between pluripotency and mitochondrial DNA proliferation during early embryo development and embryonic stem cell differentiation. Stem cell Rev. reports. doi:10.1007/s12015-009-9058-0 Frank, S., Gaume, B., Bergmann-Leitner, E.S., Leitner, W.W., Robert, E.G., Catez, F., Smith, C.L., Youle, R.J., 2001. The Role of Dynamin-Related Protein 1, a Mediator of Mitochondrial Fission, in Apoptosis. Dev. Cell 1, 515–525. doi:10.1016/S15345807(01)00055-7 Golic, K.G., Lindquist, S., 1989. The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome. Cell 59, 499–509. doi:10.1016/00928674(89)90033-0 40

Goyal, G., Fell, B., Sarin, A., Youle, R.J., Sriram, V., 2007. Role of mitochondrial remodeling in programmed cell death in Drosophila melanogaster. Dev. Cell. Gunter, T.E., Pfeiffer, D.R., 1990. Mechanisms by which mitochondria transport calcium. Am. J. Physiol. 258, C755–C786. Horbinski, C., Chu, C.T., 2005. Kinase signaling cascades in the mitochondrion: a matter of life or death. Free Radic. Biol. Med. 38, 2–11. doi:10.1016/j.freeradbiomed.2004.09.030 Ishihara, N., Nomura, M., Jofuku, A., Kato, H., Suzuki, S.O., Masuda, K., Otera, H., Nakanishi, Y., Nonaka, I., Goto, Y.-I., Taguchi, N., Morinaga, H., Maeda, M., Takayanagi, R., Yokota, S., Mihara, K., 2009. Mitochondrial fission factor Drp1 is essential for embryonic development and synapse formation in mice. Nat. Cell Biol. 11, 958–966. doi:10.1038/ncb1907 Jheng, H.-F., Tsai, P.-J., Guo, S.-M., Kuo, L.-H., Chang, C.-S., Su, I.-J., Chang, C.-R., Tsai, Y.-S., 2012. Mitochondrial Fission Contributes to Mitochondrial Dysfunction and Insulin Resistance in Skeletal Muscle. Mol. Cell. Biol. 32, 309–319. doi:10.1128/MCB.05603-11 Jia, D., Soylemez, M., Calvin, G., Bornmann, R., Bryant, J., Hanna, C., Huang, Y.-C., Deng, W.-M., 2015. A large-scale in vivo RNAi screen to identify genes involved in Notch-mediated follicle cell differentiation and cell cycle switches. Sci. Rep. 5, 12328. doi:10.1038/srep12328 41

Johnston, M.J., Bar-Cohen, S., Paroush, Z., Nystul, T.G., 2016. Phosphorylated Groucho delays differentiation in the follicle stem cell lineage by providing a molecular memory of EGFR signaling in the niche. Development dev.143263. doi:10.1242/dev.143263 Jouandin, P., Ghiglione, C., Noselli, S., 2014. Starvation induces FoxO-dependent mitotic-to-endocycle switch pausing during Drosophila oogenesis. Development 3013–3021. doi:10.1242/dev.108399 Kasahara, A., Scorrano, L., 2014. Mitochondria: from cell death executioners to regulators of cell differentiation. Trends Cell Biol. 24, 761–770. doi:10.1016/j.tcb.2014.08.005 Kasahara Atsuko, Sara Cipolat, Yun Chen, G.W.D.I.L.S., 2013. Mitochondrial Fusion Directs Cardiomyocyte Differentiation via Calcineurin and Notch Signaling. Science (80-. ). 342, 734–738. Kashatus, J.A., Nascimento, A., Myers, L.J., Sher, A., Byrne, F.L., Hoehn, K.L., Counter, C.M., Kashatus, D.F., 2015. Erk2 Phosphorylation of Drp1 Promotes Mitochondrial Fission and MAPK-Driven Tumor Growth. Mol. Cell 57, 537–551. doi:10.1016/j.molcel.2015.01.002 Klusza, S., Deng, W.-M., 2011. At the crossroads of differentiation and proliferation: Precise control of cell-cycle changes by multiple signaling pathways in Drosophila follicle cells. Bioessays 33, 124–134. doi:10.1002/bies.201000089.At 42

Labrousse, a M., Zappaterra, M.D., Rube, D. a, van der Bliek, a M., 1999. C. elegans dynamin-related protein DRP-1 controls severing of the mitochondrial outer membrane. Mol. Cell 4, 815–826. doi:10.1016/S1097-2765(00)80391-3 Leboucher, G.P., Tsai, Y.C., Yang, M., Shaw, K.C., Zhou, M., Veenstra, T.D., Glickman, M.H., Weissman, A.M., 2012. Stress-Induced Phosphorylation and Proteasomal Degradation of Mitofusin 2 Facilitates Mitochondrial Fragmentation and Apoptosis. Mol. Cell 47, 547–557. doi:10.1016/j.molcel.2012.05.041 Lee, S.F., Srinivasan, B., Sephton, C.F., Dries, D.R., Wang, B., Yu, C., Wang, Y., Dewey, C.M., Shah, S., Jiang, J., Yu, G., 2011. Γ-Secretase-Regulated Proteolysis of the Notch Receptor By Mitochondrial Intermediate Peptidase. J. Biol. Chem. 286, 27447–27453. doi:10.1074/jbc.M111.243154 Li, Q., Shen, L., Xin, T., Xiang, W., Chen, W., Gao, Y., Zhu, M., Yu, L., Li, M., 2009. Role of Scrib and Dlg in anterior-posterior patterning of the follicular epithelium during Drosophila oogenesis. BMC Dev. Biol. 9, 60. doi:10.1186/1471-213X-9-60 Lodi, R., Tonon, C., Clementi, V., Malucelli, E., 2004. Deficit of In Vivo Mitochondrial ATP Production in OPA1-Related Dominant Optic Atrophy 719–723. doi:10.1002/ana.20278 Mandal, S., Lindgren, A.G., Srivastava, A.S., Clark, A.T., Banerjee, U., Angeles, C.L., 2011. Mitochondrial function controls proliferation and early differentiation potential of embryonic stem cells. Stem Cells 29, 486–495. 43

doi:10.1002/stem.590.Mitochondrial Mitra, K., Rikhy, R., Lilly, M., Lippincott-Schwartz, J., 2012. DRP1-dependent mitochondrial fission initiates follicle cell differentiation during Drosophila oogenesis. J. Cell Biol. 197, 487–97. doi:10.1083/jcb.201110058 Mitra, K., Wunder, C., Roysam, B., Lin, G., Lippincott-Schwartz, J., 2009. A hyperfused mitochondrial state achieved at G1-S regulates cyclin E buildup and entry into S phase. Proc. Natl. Acad. Sci. U. S. A. 106, 11960–5. doi:10.1073/pnas.0904875106 Monick, M.M., Powers, L.S., Barrett, C.W., Hinde, S., Ashare, A., Groskreutz, D.J., Nyunoya, T., Coleman, M., Spitz, D.R., Hunninghake, G.W., 2008. Constitutive ERK MAPK activity regulates macrophage ATP production and mitochondrial integrity. J. Immunol. 180, 7485–7496. doi:180/11/7485 [pii] Newton, R., Cambridge, L., Hart, L.A., Stevens, D.A., Lindsay, M.A., Barnes, P.J., 2000. The MAP kinase inhibitors , PD098059 , UO126 and SB203580 , inhibit IL-1 b dependent PGE 2 release via mechanistically distinct processes 1353–1361. Nunnari, J., Suomalainen, A., 2012. Mitochondria: in sickness and in health. Cell 148, 1145–59. doi:10.1016/j.cell.2012.02.035 Nystul, T., Spradling, A., 2010. Regulation of epithelial stem cell replacement and follicle formation in the Drosophila ovary. Genetics 184, 503–515. doi:10.1534/genetics.109.109538 44

Ohsawa, S., Sato, Y., Enomoto, M., Nakamura, M., Betsumiya, A., Igaki, T., 2012. Mitochondrial defect drives non-autonomous tumour progression through Hippo signaling in Drosophila. Nature 490, 547–551. doi:10.1038/nature11452 Pendergrass, W., Wolf, N., Poot, M., 2004. Efficacy of MitoTracker Green? and CMXrosamine to measure changes in mitochondrial membrane potentials in living cells and tissues. Cytometry 61A, 162–169. doi:10.1002/cyto.a.20033 Perumalsamy, L.R., Nagala, M., Sarin, A., 2010. Notch-activated signaling cascade interacts with mitochondrial remodeling proteins to regulate cell survival. Proc. Natl. Acad. Sci. U. S. A. 107, 6882–6887. doi:10.1073/pnas.0910060107 Pyakurel, A., Savoia, C., Hess, D., Scorrano, L., 2015. Extracellular Regulated Kinase Phosphorylates Mitofusin 1 to Control Mitochondrial Morphology and Apoptosis. Mol. Cell 58, 244–254. doi:10.1016/j.molcel.2015.02.021 Rikhy, R., Kamat, S., Ramagiri, S., Sriram, V., Krishnan, K.S., 2007. Mutations in dynamin-related protein result in gross changes in mitochondrial morphology and affect synaptic vesicle recycling at the Drosophila neuromuscular junction. Genes, Brain Behav. 6, 42–53. doi:10.1111/j.1601-183X.2006.00218.x Rohrbaugh, M., Ramos, E., Nguyen, D., Price, M., Wen, Y., Lai, Z.C., 2002. Notch activation of yan expression is antagonized by RTK/pointed signaling in the Drosophila eye. Curr. Biol. 12, 576–581. doi:10.1016/S0960-9822(02)00743-1

45

Roth, S., Lynch, J. a, 2009. Symmetry breaking during Drosophila oogenesis. Cold Spring Harb. Perspect. Biol. 1, a001891. doi:10.1101/cshperspect.a001891 Schaeffer, V., Althauser, C., Shcherbata, H.R., Wu-Min Deng, Ruohola-Baker, and H., 1998. Notch-Dependent Fizzy-Related/Hec1/Cdh1 Expression Is Required for the Mitotic-to-Endocycle Transition in Drosophila Follicle Cells. Development 125, 2457–2467. doi:10.1016/j Schieke, S.M., McCoy, J.P., Finkel, T., 2008. Coordination of mitochondrial bioenergetics with G(1) phase cell cycle progression. Cell Cycle 7, 1782–1787. Shcherbata, H.R., Althauser, C., Findley, S.D., Ruohola-Baker, H., 2004. The mitotic-toendocycle switch in Drosophila follicle cells is executed by Notch-dependent regulation of G1/S, G2/M and M/G1 cell-cycle transitions. Development 131, 3169– 81. doi:10.1242/dev.01172 Struhl, G., Adachi, A., 1998. Nuclear access and action of Notch in vivo. Cell 93, 649– 660. doi:10.1016/S0092-8674(00)81193-9 Sun, J., Deng, W.-M., 2007. Hindsight mediates the role of notch in suppressing hedgehog signaling and cell proliferation. Dev. Cell 12, 431–42. doi:10.1016/j.devcel.2007.02.003 Susin, S. a, Lorenzo, H.K., Zamzami, N., Marzo, I., Snow, B.E., Brothers, G.M., Mangion, J., Jacotot, E., Costantini, P., Loeffler, M., Larochette, N., Goodlett, D.R., 46

Aebersold, R., Siderovski, D.P., Penninger, J.M., Kroemer, G., 1999. Molecular characterization of mitochondrial apoptosis-inducing factor. Nature 397, 441–446. doi:10.1038/17135 Thayer, S.A., Miller, R.J., 1990. Regulation of the intracellular free calcium concentration in single rat dorsal root ganglion neurones in vitro. J. Physiol. 425, 85–115. Tian, A.-G., Deng, W.-M., 2008. Lgl and its phosphorylation by aPKC regulate oocyte polarity formation in Drosophila. Development 135, 463–71. doi:10.1242/dev.016253 Tomlinson, A., Struhl, G., 2001. Delta/Notch and Boss/Sevenless signals act combinatorially to specify the Drosophila R7 photoreceptor. Mol. Cell 7, 487–495. doi:10.1016/S1097-2765(01)00196-4 Williams, P. a., Morgan, J.E., Votruba, M., 2010. Opa1 deficiency in a mouse model of dominant optic atrophy leads to retinal ganglion cell dendropathy. Brain 133, 2942– 2951. doi:10.1093/brain/awq218 Xu, J., Chi, F., Guo, T., Punj, V., Lee, W.N.P., French, S.W., Tsukamoto, H., 2015. NOTCH reprograms mitochondrial metabolism for proinflammatory macrophage activation 125, 1579–1590. doi:10.1172/JCI76468DS1 Zanna, C., Ghelli, A., Porcelli, A.M., Karbowski, M., Youle, R.J., Schimpf, S., Wissinger, 47

B., Pinti, M., Cossarizza, A., Vidoni, S., Valentino, M.L., Rugolo, M., Carelli, V., 2008. OPA1 mutations associated with dominant optic atrophy impair oxidative phosphorylation and mitochondrial fusion. Brain 131, 352–67. doi:10.1093/brain/awm335 Figure S1: Controls for mitochondrial morphology and membrane potential (A-B) Surface view of aggregated mitochondria in drp1KG mutant follicle cells. Mitochondria are tightly aggregated in drp1 mutant follicle cells (CD8-GFP;green) as compared to background control cells (A). FRT 40A clone cells have dispersed mitochondria and their organization is similar to background control cells (B). (C-D) Ratio of CMXRos fluorescence intensity in drp1KG GFP positive to control cells is similar before and after fixation of FCs (n= 4). (E-F) CMXRos intensity is increased and mitochondrial intensity does not change in drp1KG. CMXRos fluorescence intensity normalized to mitochondrial fluorescence intensity in GFP positive cells and GFP negative cells shows that drp1 mutant cells have increased CMXRos and increased mitochondrial membrane potential. CMXRos/Mito ratio is lowered in drp1KG; pdswi combination. FCCP treatment lowers the CMXRos fluorescence in drp1KG mutant follicle cells (E, n= 22,21,8,30,13,11, N= 5,4,4,4,3,3, ***, P<0.001, **, P<0.01, *, P<0.05). Mitochondrial fluorescence intensity ratio per cell is unchanged between GFP and non-GFP cells in all the genotypes tested (n=17,16,18,15,12,16, N= 3 for all) (F).

48

The pseudocolor scale is the 16 color LUT used from Image J where highest intensity pixel is red and lowest is blue. Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). ns= non significant. Scale Bar: 10μm FCs n=PFC clones in independent ovarioles, N= Experimental replicates. Percentage mitochondrial aggregate phenotype is represented in the figure.

Figure S2: Drp1 depletion leads to mitochondrial aggregation in hemocytes (A-B) Mitochondria are aggregated in hemocytes expressing drp1i: Collagen Gal4 driven mito-GFP shows aggregated mitochondria in drp1i hemocytes (yellow: cell outline) (A) with a significantly larger optically unresolvable fluorescence area in drp1i as compared to wild-type (B) (n= 75, 90, N=3 experiments ,***, P< 0.001, Two-tailed MannWhitney). The mitochondrial fluorescence density per cell did not change between control and drp1i expressing cells (data not shown). Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Scale Bar: 5μm. n= number of individual hemocytes, N=Experimental replicates.

49

Figure S3: Ras and ERK RNAi lower Ras and ERK in FCs (A-B) ras and erk RNAi mutants show depletion of corresponding antibody levels: rasi and erki cells show lowering of Ras (94% show Ras depletion, n= 16, N=2) (A) and pERK (100% show pERK depletion, n= 15, N=2) (B) antibody staining (intensity pseudocolored scale) as compared to neighbouring control cells. Zoomed images of the clonal region are shown in bottom panels. Scale Bar: 10μm. The pseudocolor scale is the 16 color LUT used from Image J where highest intensity pixel is red and lowest is blue. n= PFC clones in independent ovarioles, N=Experimental replicates.

Figure S4: Confocal images of rescue of mitochondrial organization in drp1KG by additional depletion of erk and ras (A-B) Increased CMXRos in drp1KG mutant is not seen in drp1KG;erki. drp1 mutant cells have increased CMXRos which is suppressed by erk reduction in drp1 mutant cells. Graph depicts lowering of CMXRos intensity normalized to mitochondrial intensity (A, n= 11,8,10,12, N= 5,3,3,3, ***, P<0.001, **, P<0.01, *, P<0.05). Mitochondrial intensity is similar in drp1, drp1;erki and erki clones (B, n=12,16,10,12, N=3 for all). (C-G) Mitochondria are aggregated in drp1KG PFCs. Expressing rasi and erki in the drp1KG background changes the mitochondrial distribution to that of neighbouring control cells (n= 47, 62, 62, 24, 21, N=3 for all). Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations 50

outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). ns= non significant. Scale Bar: 10μm FCs n=PFC clones in independent ovarioles, N= Experimental replicates. Percentage mitochondrial aggregate phenotype is represented in the figure.

Figure S5: FRT40A clone cells are indistinguishable for Cyclin and Notch immunostaining from background cells (A-B) Immunostaining is similar to the background for FRT40A clones. Cyclin B (n=15) and Cyclin E (n=10) (A) remain same and Notch pathway components; Hnt (n=10), NICD (n=10) and Cut (n=3) (B) also are comparable. Scale Bar: 10μm n=PFC clones in independent ovarioles

Figure S6: Ras downregulation does not reverse the Cyclin B/E accumulation in drp1KG PFCs (A-D) Increased Cyclin E (A, B (n= 19,12,11, N=4,3,3) and B (C, D (n= 13,10,12, N=3,3,2) immunostaining in drp1KG (Fig 5D, F) is not quantitatively reversed (B, D) in drp1KG; rasi PFCs. Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). Statistical significance tested with two-tailed Kruskal Wallis and Dunn’s Test,

51

ns =not significant, Scale Bar: 10μm. n=PFC clones in independent ovarioles, N=Experimental replicates.

Figure S7: Hnt appears early in FCCP treated wild-type ovarioles (A-B) Small nuclear size in drp1KG mutant is reversed with the pdswi, erki and rasi combination. drp1KG mutant follicle cell nuclei (red boundary) at stage 7-9 are significantly smaller as compared to neighbouring control cell (green boundary) as well as FRT40A control cells (A) (n=50,50 nuclei, 5 clones, ***,P<0.001,two-tailed MannWhitney). Small size is rescued partially in drp1KG; pdswi (n=42,46 (pdswi) nuclei, 5 clones, ***,P < 0.001). drp1KG; erki and drp1KG; rasi show a complete rescue of nuclear size (n=52,50,52 (erki), 47 (rasi) nuclei, 5 clones, ***,P < 0.001). drp1KG; pdswi is comparable to Stage 6 FCs suggesting a delay in Hnt activation in the same. Representative images are shown in (A) and quantification in (B). Statistical significance tested with two-tailed Mann-Whitney test. (C-D) FCCP treated wild-type ovarioles show Notch induced Hnt appearance at an earlier stage. Wild-type ovarioles show Hnt (green) in stage 6-7 and on FCCP treatment Hnt appears at stage 5 (100% show early Hnt appearance, n=39, N=2). CMXRos (intensity pseudocolored scale) staining has decreased fluorescence on FCCP treatment (C). Quantitative analysis of chamber size of Hnt onset shows a shift to the left as compared to untreated ovarioles (control; n=10, N=2)(D).

52

The pseudocolor scale is the 16 color LUT used from Image J where highest intensity pixel is red and lowest is blue. Data is presented as box plots where horizontal bar represents mean, box limits 25th and 75th percentiles, whiskers 10th and 90th percentiles and dots are observations outside 10th and 90th percentiles. Numbers within the box represent number of data points (n). ns =not significant, Scale Bar: 10μm. n=number of clonal nuclei in independent ovarioles (A-B) and individual ovarioles (C-D), N=Experimental replicates.The pseudocolor scale is a rainbow where highest intensity pixel is red and lowest is blue.

Supplementary Table Legends Table S1: Additional quantification for mitochondrial membrane potential, dpERK, Caspase and Cut levels The table shows the EGFR-DN rescue of the mitochondrial membrane potential in the drp1KG background. drp1KG; rasi however does not rescue the same. drp1KG treated with FCCP does not show a significant difference in dpERK immunostaining from the neighboring control cells. dpERK in rasi clones is not significantly different from the neighboring control cells. drp1KG does not show a difference in activated caspase immunostaining from neighboring control cells.

53

The transcription factor Cut is present in drp1KG (percent clones containing Cut at stage 7-8 is 100) PFCs and this defect is rescued in drp1KG; rasi (percent clones containing Cut at stage 7-8 is reduced to 22) and drp1KG; erki (percent clones containing Cut at stage 7-8 is 23). erki and rasi do not show any defect in Cut appearance (percent clones containing Cut at stage 7-8 is 0). Lowering of membrane potential alone via pdswi and FCCP also does not rescue drp1KG Cut appearance in PFCs (percent clones containing Cut at stage 7-8 is 100%). The genotypes and corresponding phenotypes along with quantification and sample sizes are shown. SD= Standard deviation, n= number of ovarioles, N= Experimental replicates.

Table S2: Summary of phenotypes and epistasis analysis The table summarizes phenotypes in FCs with different genetic combinations in four classes: mitochondrial morphology and activity, multilayering, Ras/ERK signaling and Notch signaling. The control phenotypes are shown in the right hand side columns. drp1KG FCs had aggregated mitochondrial morphology, elevated mitochondrial potential, multilayering, accumulation of Ras/dpERK, smaller nuclear size and loss of Notch signaling. ERK depletion in the drp1KG background rescued all phenotypes except oocyte localization. pdsw depletion with drp1KG rescued Notch signaling with no effect on dpERK levels and mitochondrial morphology. FCCP treatment reduced the 54

mitochondrial potential to background levels in drp1 mutant FCs and induced Notch signaling at an earlier stage in wild-type FCs. The results are based on statistically significant change in the phenotypes across the combinations as shown in the main and supplementary figures.

Highlights ● Aggregated mitochondria in drp1 mutant PFCs have higher membrane potential Ψm ● ERK regulates Ψm in wild type and drp1 mutant cells ● Ψm decrease does not impact EGFR signaling driven oocyte patterning ● Ψm decrease activates Notch signaling in wild type and drp1 mutant cells

55

56

57

58

59

60

61

62