An RNAi screen of the kinome in epithelial follicle cells of the Drosophila melanogaster ovary reveals genes required for proper germline death and clearance

Abstract

Programmed cell death and cell corpse clearance are an essential part of organismal health and development. Cell corpses are often cleared away by professional phagocytes such as macrophages. However, in certain tissues, neighboring cells known as nonprofessional phagocytes can also carry out clearance functions. Here, we use the Drosophila melanogaster ovary to identify novel genes required for clearance by nonprofessional phagocytes. In the Drosophila ovary, germline cells can die at multiple time points. As death proceeds, the epithelial follicle cells act as phagocytes to facilitate the clearance of these cells. We performed an unbiased kinase screen to identify novel proteins and pathways involved in cell clearance during two death events. Of 224 genes examined, 18 demonstrated severe phenotypes during developmental death and clearance while 12 demonstrated severe phenotypes during starvation-induced cell death and clearance, representing a number of pathways not previously implicated in phagocytosis. Interestingly, it was found that several genes not only affected the clearance process in the phagocytes, but also non-autonomously affected the process by which germline cells died. This kinase screen has revealed new avenues for further exploration and investigation.

Keywords: Drosophila; cell death; kinase; phagocytosis; phosphoinositide 3-kinase.

Figure 1

The Drosophila ovary is composed of developing egg chambers. (A) Two Drosophila ovaries. (B) Cartoon of Drosophila ovaries, illustrating strings of developing egg chambers called ovarioles. The drawing was adapted from Mahowald and Kambysellis (1980). (C) The central plane of a mid-stage egg chamber from a GR1>Luc^RNAi (control) fly. DNA is labeled with DAPI (cyan). (D) Cartoon of mid-stage egg chamber. Each egg chamber consists of three cell types, nurse cells (NC), oocyte (O), and a surrounding epithelial FC layer. In the cartoon, the nuclei are in cyan and membranes are white. (E) Two ovarioles from a single Drosophila ovary. The inset is the original image from which these ovarioles were taken. (F) Cartoon of an ovariole. Early stages are labeled in yellow. Mid-oogenesis is labeled in green. Late oogenesis is labeled in cyan. Egg chambers complete the 14 stages of oogenesis while progressing from anterior to posterior in the ovariole. Death events can be induced by starvation during mid-stage oogenesis (stages 7–9) and are seen at the completion of development of healthy egg chambers (stages 12–14). The ovariole drawing was adapted from Frydman and Spradling (2001). All scale bars depict 50 μm.

Figure 2

Experimental approach to detect different cell death phenotypes. (A) The bipartite Gal4-UAS system was used to express kinase RNAi constructs in the FCs of Drosophila ovaries. A FC specific promoter drives expression of the Gal4 transcription factor. This construct is expressed in one parental fly. An upstream activating sequence (UAS) is bound to a gene of interest in the other parental fly. When flies are mated, their progeny have Gal4 expressed in the FCs which can then bind the UAS to promote the expression of the desired gene or, in this case, kinase RNAi. (B) Experimental set up for analysis of developmental NC death. Flies with the desired genotype were conditioned with extra yeast paste which was applied to the inner wall of a vial twice over 48 h. Ovaries were then dissected, stained, and imaged. (C) Experimental set up for starvation induced cell death. Flies with the desired genotype were conditioned for 24 h with extra yeast paste. After 24 h, flies were transferred to an apple juice agar vial to starve the flies of protein. Flies were kept on apple juice agar for 16–20 h until their ovaries were dissected, stained, and imaged.

Figure 3

Classification of defective developmental death and clearance using the persisting nurse cell nuclei phenotype. (A) Cartoon of late stage oogenesis morphology adapted from King (1970). As egg chambers progress through the final stages of oogenesis, the volume of the NCs (anterior, left) decreases while that of the oocyte (posterior) increases. The NCs are then depleted and the oocyte development is complete with the growth of DA and a chorion shell. (B–F) Anterior (NC) region of GR1>Luc^RNAi (control) egg chambers during the final stages of oogenesis. DNA is labeled with DAPI (cyan). NC nuclei are indicated by yellow arrows, and small nuclei are FCs. (B) At stage 10, NC chromatin is dispersed and the nuclei are surrounded by cytoplasm. (C) During stage 11, NCs dump their cytoplasmic contents into the oocyte (O). (D and E) By stage 12, dumping is complete and throughout stages 12 and 13, the NC nuclei are cleared. (F) By stage 14, all nuclei are cleared and the oocyte has fully developed DA. (G–O) Representative images of anterior regions of stage 14 egg chambers demonstrating different developmental death and clearance phenotypes. DNA labeled with DAPI (cyan). (G) A stage 14 egg chamber from GR1>drpr^RNAi fly demonstrates a very strong PN phenotype. (H) GR1>Adck^HMS02533 demonstrates no PN phenotype. (I) GR1>Pak^HM05156 demonstrates a mild PN phenotype. (J) GR1>rok^JF03225 demonstrates a moderate PN phenotype. (K) GR1>SNF4Agamma^JF02060 demonstrates a strong PN phenotype. (L) GR1>tefu^HMS02790 demonstrates a very strong phenotype. (M and N) GR1>Wnk^HMJ02087 demonstrates a dumpless phenotype. (M) Several GR1>Wnk^HMJ02087 egg chambers shown at 100X. (N) A 600X image of square in box M. (O) GR1>rl^HMS00173 demonstrates a gnarled dorsal appendage phenotype. (P) GR1>sdt^HMS00953 demonstrates a “bumpy” chorion phenotype (white arrows). All scale bars depict 50 μm. Scale bar in B is representative of all images in this figure unless otherwise indicated.

Figure 4

Engulfment phenotypes during starvation-induced death during mid-oogenesis. (A–D) Cartoon of starvation-induced egg chamber death. DNA is depicted by cyan, germline cytoplasm is depicted in green, and membranes are depicted in magenta. (A) A healthy, mid-oogenesis egg chamber is composed of germline cells that are approximately equal in size that contain large, dispersed, polyploid nuclei surrounded by a layer of FCs. The posterior-most germline cell is the developing oocyte (O). (B) Early phase dying egg chamber where the NC nuclei of the egg chambers become disorganized, condense, and begin to fragment. (C) Mid-phase dying egg chamber where the NC nuclei further degrade and the FCs increase in size and engulf the germline material. (D) Late phase dying egg chamber where all of the germline has been consumed. (E–M) Representative egg chamber images depicting different germline death and clearance phenotypes during mid-oogenesis. DNA labeled with DAPI (cyan), germline cytoplasm is labeled with GFP (green), and membranes are labeled with Dlg (magenta). White arrows identify the egg chamber of interest. Yellow arrows indicate missing FCs. (E) GR1>drpr^RNAi egg chamber shows pronounced engulfment defects with missing FCs and unengulfed germline. (F–I) Healthy, early phase, mid- and late-phase dying egg chambers from a protein-starved GR1>Luc^RNAi control fly. (J) GR1>CG10702^HMS02499 produces a wild-type phenotype. (K) GR1>LIMK1^JF02063 produces a weak clearance defect in that some of the FCs have been eliminated before the dying germline has been cleared. (L) GR1>aPKC^JF01966 demonstrates a strong clearance defect in that most FCs have disappeared before the dying germline has been cleared. (M) GR1>Taf1^HMS00416 demonstrates an “undead” phenotype where no FCs are seen surrounding the living germline. All scale bars depict 50 μm. Scale bar in (F) is representative of all images in this figure.

Figure 5

Knockdowns of classes II and III PI3-kinases result in strong defects during late oogenesis. (A–C) Stage 14 egg chambers from well-fed flies stained with DAPI (blue) to label DNA. Insets—Egg chambers imaged with DIC to identify stage 14 egg chambers using DAs. (A) Sibling control egg chambers did not have PN. (B) GR1>Pi3K68D^JF0119 egg chambers exhibited a strong PN phenotype. (C) GR1>Pi3K59F^HMS00261 egg chambers exhibited a strong PN phenotype. (D) Quantification of PN in sibling control, GR1>Pi3K59F^HMS00261, and GR1>Pi3K68D^JF0119 stage 14 egg chambers. Data presented are mean + SEM. (E–G) Stage 13 egg chambers from well-fed flies stained with DAPI (cyan) to label DNA. DIC was used to visualize the developing DA to identify egg chambers as stage 13. (Eʹ–Gʹ) Egg chambers labeled with LysoTracker (red) to identify acidified compartments. (Eʹʹ–Gʹʹ) Egg chambers co-labeled with DAPI (cyan) and LysoTracker (red). (E–Eʹʹ) Sibling control egg chambers contained a few acidified NCs, with the majority already degraded. (F–Fʹʹ) GR1>Pi3K68D^JF0119 contained several NC nuclei that were not LysoTracker positive. (G–Gʹʹ) GR1>Pi3K59F^HMS00261 contained several nuclei, several of which were not LysoTracker positive. (H) Quantification of LysoTracker staining in stage 13 egg chambers. Data presented are mean + SEM. All scale bars depict 50 μm. Scale bar in (E) is representative of images in (E–Gʹʹ).

Figure 6

Knocking down of class I PI3K results in delayed engulfment during mid-oogenesis. (A–C) Sibling control egg chambers stained with DAPI (cyan) to visualize the nuclei of healthy, early dying, and mid phase dying respectively. (D) GR1>Luc^RNAi egg chamber stained with DAPI (cyan) to visualize nuclei. (Aʹ–Dʹ) Control egg chambers stained with α-Dlg (grey) to label membranes of FCs and NCs. (A–Aʹ) Healthy egg chamber shows a thin FC layer surrounding germline NCs. (B–Bʹ) Egg chamber undergoing early phase of death shows a thin FC layer surrounding condensing and fragmenting NC nuclei. (C–Cʹ) Egg chamber in mid-phase death continues to show condensing NC nuclei while the FC layer increases in size as it begins to engulf germline material. (D–Dʹ) Late phase dying egg chamber shows a FC layer that completely encroaches on the germline region as it engulfs any remaining condensed NC nuclei and other germline material. (E–Hʹ) GR1>Pi3K92E^RNAi (Pi3K92E^JF02770) egg chambers in various phases of health and death. (E–H) GR1>Pi3K92E^RNAi egg chambers stained with DAPI (cyan) to visualize the nuclei of healthy, early dying, mid, and late phase dying respectively. (Eʹ–Hʹ) GR1>Pi3K92E^RNAi egg chambers stained with α-Dlg (grey) to label membranes of FCs and NCs. (E–Eʹ) Healthy egg chambers demonstrating diffuse NC nuclei and a single thin layer of FCs. (F–Fʹ) Early phase dying GR1>Pi3K92E^RNAi egg chamber with condensing nuclei with no change in the FC layer. (G–Gʹ) Mid phase dying GR1>Pi3K92E^RNAi egg chamber with nuclei undergoing further condensation and fragmentation, but the FC layer still has not enlarged to start clearing the germline as it has in the control. (H–Hʹ) Late phase dying GR1>Pi3K92E^RNAi egg chamber where the germline has been cleared by an enlarged FC layer. All scale bars depict 50 μm. Scale bar in A is representative of all images in this figure unless otherwise indicated.

Figure 7

Members of the PI3K/PTEN/AKT pathway are required for egg chamber survival during mid-oogenesis in well-fed flies. (A–C) Low magnification images displaying multiple egg chambers labeled with DAPI (blue). White arrows indicate degenerating egg chambers. (A) GR1>Luc^RNAi egg chambers from well-fed conditioned flies. No degenerating egg chambers are observed. (B) GR1>Pi3K92E^JF02770 egg chambers from well-fed (conditioned) flies demonstrate a moderate level of degeneration during mid-oogenesis. (C) GR1>Pdk1^JF02807 egg chambers from well-fed (conditioned) flies demonstrated an excessive amount of degeneration during mid-oogenesis.