Induction of Chk2 signaling in the germarium is sufficient to cause oogenesis arrest in Drosophila

The conserved RNA helicase Vasa is required for germ cell development in many organisms. It is established that in Drosophila loss of piRNA pathway components, including Vasa, causes Chk2-dependent oogenesis arrest, however the stage at which Chk2-signaling is triggered was unknown. We found that absence of Vasa during the germarial stages arrests oogenesis due to Chk2 activation. Importantly, once induced in the germarium, Chk2-mediated arrest of oogenesis cannot be overcome by restoration of Vasa to the arrested egg-chambers. We conclude that Vasa activity specifically in the germarium is essential for germ cell lineage development.

Development of the Drosophila female gonad begins during the third larval instar with the formation of 16-25 somatic niches that will give rise to the future germaria (Panchal et al., 2017). Each germarium hosts germ line stem cells (GSCs) that produce the germ cell lineage (Wieschaus and Szabad, 1979). In adult females, germ cell development begins with the division of a GSC into a self-renewing stem cell and a differentiating daughter cell, the cystoblast (CB). The CB undergoes four rounds of mitosis with incomplete cytokinsesis, such that a stage 1 egg-chamber is ultimately composed of an oocyte and 15 nurse cells, surrounded by a layer of follicular epithelial cells (reviewed in Gilboa and Lehmann, 2004). A newly formed egg-chamber buds off from the germarium and joins a linear array of developing egg-chambers to form an ovariole. Each Drosophila ovary consists of 16-25 ovarioles, corresponding to the number of germaria formed in the third instar larva.
Drosophila oogenesis has been intensively studied and many genes found to regulate development of the germ cell lineage. Among the germline proteins essential for oogenesis is the conserved RNA helicase Vasa (Vas). Vas is expressed throughout oogenesis and localizes to the posterior pole of the oocyte and early embryo. In situations leading to absence of Vas from the oocyte posterior pole, germline and posterior patterning determinants fail to localize, germ (or pole) plasm does not form, and the resulting embryos lack posterior structures and primordial germ cells (Hay et al., 1990;Ashburner, 1988, 1990). In contrast to late oogenesis and embryos, little is known about the role of Vas during early oogenic stages. In early oogenesis vas has been implicated in the translational control of mei-p26 and in regulation of GSC mitotic chromosome condensation (Liu et al., 2009;Pek and Kai, 2011). As complete absence of vas triggers oogenesis arrest induced by checkpoint kinase 2 (Chk2) (Durdevic et al., 2018;Lasko and Ashburner, 1990), the question remains at which stage Vas activity is critically required for oogenesis to complete.
To address the importance of Vas activity in early germ cell lineage development we took a genetic approach. We found that targeted exclusion of Vas from the germarium arrests oogenesis and reduces the number of egg-chamber-producing ovarioles. Furthermore, in the absence of Vas in the germarium, Chk2-signaling induces germ cell developmental arrest. Importantly, once induced in the germarium, Chk2-mediated arrest of oogenesis cannot be overcome by restoration of Vas to the arrested egg-chambers. Our data show that the activity of Vas RNA helicase early in oogenesis prevents activation of Chk2-signaling, ensuring sustained development of the germline component of the ovary.

Generation of transgenic flies and expression of the transgenes
The vasa transgene carrying the K295N substitution (vas GNT  All transgenes are expressed using the GAL4/UAS-system. Gal4-drivers were under control of two promoters with distinct expression patterns: vas-Gal4 is expressed throughout oogenesis and matTub-Gal4 is excluded from the germarium (Supplementary Figure S1B).

Fecundity and hatching assays
Virgin females of w 1118 control and vas PD/D1 and vas D1/D1 genetic backgrounds with or without expressed transgenes were mated with w 1118 males for 24h at 25°C. The crosses were then transferred to apple-juice agar plates, which were used to collect eggs at 24h intervals over 3 or 3-20 days. The number of laid eggs on each plate was counted and the plates were kept at 25°C for 24h and the number of hatched larvae was also counted (Supplementary Table S1 and S2). Experiments were performed in five independent replicates.

Ovarian morphology and quantification of ovariole number
Ovaries were dissected from 3-day to 20-day old flies in PBS. To assess ovarian morphology, ovaries were directly imaged on Olympus SZX16 stereomicroscope. The length of the ovaries was measured using Fiji (Supplementary Table S3, S4 and S5). For determination of eggchamber-producing ovariole number females were frozen and held at -20°C prior dissection.
Ovaries were manually dissected under magnification in a drop of PBS. The ovarioles were gently separated from each other using wolfram needles. The ovariole number of each female was defined as a summary of the number of egg-chamber-containing ovarioles in the right and left ovary (Supplementary Table S6 and S7).
Fixed ovaries were incubated with primary antibodies Vasa (rat; 1:500; (Tomancak et al., 1998)) and subsequently with secondary antibodies Alexa 647 conjugated donkey anti-rat IgG (1:1000; Jackson ImmunoResearch). Ovaries from flies expressing fusion proteins were fixed in 2% PFA and 0.01% Triton X-100 for 15 min at RT. Fixed ovaries were mounted on glass slides for examination of GFP fluorescence for the fusion proteins and Alexa 647 fluorescence for wild-type Vas using a Zeiss LSM 780 confocal microscope. Nuclei were visualised with DAPI.

Cuticle preparation
To examine larval cuticles, eggs were allowed to develop fully for 24h at 25°C, dechorionated in bleach, and then transferred to a microscope slide bearing a drop of Hoyer's medium mixed 1:1 with lactic acid. Cuticle preparations were heated at 65° overnight before examination using Zeiss Axiophot microscope. Number of counted larvae with or without abdomen is represented in Supplementary Table S8.
The samples were observed using a Zeiss LSM 780 or Leica SP8 confocal microscope. Oocytes and embryos with Aub and Ago3 positive pole plasm were counted in three independent replicates (Supplementary Table S9 and S10).
Quantification of relative protein expression levels was performed using ImageJ. A frame was placed around the most prominent band on the image and used as a reference to measure the mean gray value of all other protein bands, as well as the background. Next, the inverted value of the pixel density was calculated for all measurements by deducting the measured value from the maximal pixel value. The net value of target proteins and the loading control was calculated by deducting the inverted background from the inverted protein value.
The ratio of the net value of the target protein and the corresponding loading control represents the relative expression level of the target protein. Fold-change was calculated as the ratio of the relative expression level of the target protein in the wild-type control over that of a specific sample.
Proteomic experiments were performed as described in (Casabona et al., 2013). In brief, three samples were stacked in the SDS polyacrylamide gel and after Coomassie staining each lane was cut into 3 blocks, which were processed separately. After digestion with trypsin (Promega, sequencing grade), the resulting peptides were analyzed by LC-MS/MS (LTQ-Orbitrap Velos pro, Thermo Fisher Scientific). Peptides and proteins from each MS run were identified using Scaffold software and results for each lane were displayed. The selection criteria for the displayed proteins were: a minimum of 2 peptides per protein should be identified and the peptide Mascot score should be at least 20. The experiment was performed in two biological replicates and specific interaction partners were determined by statistical analysis of control and positive samples using extracted spectral counts. A protein was considered as a high confidence binding partner if its enrichment was equal to or above 2 and the p-value was ≤ 0.05 in positive IPs compared to controls. All proteins that showed enrichment equal to or above 2 but had a higher p-value were considered low confidence hits.
P-values were computed using the web-based Quantitative Proteomics p-value Calculator (QPPC) (Chen et al., 2014) that applies a distribution-free permutation method based on simulation of the log(ratio). A pseudocount of 1 was used in all samples for proteins with no spectral counts. Un-weighted spectrum counts for both replicates and the results of the statistical analysis are provided in Supplementary Table S11.

Fluorescent in situ RNA hybridization
FISH experiments were performed as described in (Gaspar et al., 2017). In brief, ovaries were dissected in PBS and immediately fixed in 2% PFA, 0.05 % Triton X-100 in PBS for 20 min at RT. After washing in PBT (PBS + 0.1% Triton X-100) samples were treated with 2 µg/mL proteinase K in PBT for 5 min and then were subjected to 95°C in PBS + 0.05% SDS for 5 min. Samples were pre-hybridized in 200 µL hybridization buffer (300 mM NaCl, 30 mM sodium citrate pH 7.0, 15 % ethylene carbonate, 1 mM EDTA, 50 µg/mL heparin, 100 µg/mL salmon sperm DNA, 1% Triton X-100) for 10 min at 42°C. Fluorescently labeled oligonucleotides (12.5-25 nM) were pre-warmed in hybridization buffer and added to the samples. Hybridization was allowed to proceed for 2 h at 42°C. Samples were washed 3 times for 10 min at 42°C in pre-warmed buffers (1x hybridization buffer, then 1x hybridization buffer:PBT 1:1 mixture, and then 1x PBT). The final washing step was performed in prewarmed PBT at RT for 10 min. The samples were mounted in 80% 2,2-thiodiethanol in PBS and analyzed on a Leica SP8 confocal microscope.

Labeling of DNA oligonucleotides for fluorescent in situ RNA hybridization
Labeling of the oligonucleotides was performed as described in (Gaspar et al., 2017). Briefly, non-overlapping arrays of 18-22 nt long DNA oligonucleotides complementary to mnk (Supplementary Table S12) were selected using the smFISHprobe_finder.R script (Gaspar et al., 2017). An equimolar mixture of oligos for a given RNA was fluorescently labelled with Alexa 565-or Alexa 633-labeled ddUTP using terminal deoxynucleotidyl transferase. After ethanol precipitation and washing with 80% ethanol, fluorescently labeled oligonucleotides were reconstituted with nuclease-free water.

RNA extraction and quantitative PCR analysis
Total RNA was extracted from ovaries of 3-day old flies using Trizol reagent (Thermofisher).
For first-strand cDNA synthesis, RNA was reverse transcribed using the QuantiTect Reverse  (Livak and Schmittgen, 2001) and normalized to rp49 mRNA levels and normalized to respective RNA levels from w 1118 flies. Sequences of primers used for qPCR reaction are presented in Supplementary Table S12.

Localization of PIWI proteins is affected by Vasa's helicase activity
Localization of Vas in the egg-chamber is independent of RNA-binding and helicase activity (Dehghani and Lasko, 2016;Liang et al., 1994). We analyzed whether the E400Q  showed wild-type localization (Figure 2A-B). This was true for the localization of Aub and Ago3 (Supplementary Figure S2A-B and S3A). These findings indicate that an open helicase conformation of the Vas is required for its correct localization, as well as for the localization of Aub and Ago3, whereas helicase activity per se is not.
In oocytes and embryos, GFP-Vas WT showed a wild-type localization at the posterior pole ( Figure 2A-B), whereas GFP-Vas DQAD was not detected. And, although we could detect GFP-Vas GNT at the posterior pole of the oocyte and the protein was transmitted to the embryo, it was not detected at the posterior pole ( Figure 2B and Supplementary Figure S1E

Vasa associated proteins in the Drosophila ovary
The dynamic association of DEAD-box RNA helicases with multiprotein complexes (Linder and Jankowsky, 2011) renders challenging the biochemical detection of their interaction partners. The E400Q mutation, which locks Vas-containing protein complexes, is an ideal biochemical tool for identifying Vasa's interaction partners in vivo (Xiol et al., 2014). We protein, which interacts with Vas at the posterior pole (Breitwieser et al., 1996;Jeske et al., 2017;Wang et al., 2015) was not detected in the co-IPs, whether in the case of GFP-Vas WT or of the "locked" GFP-Vas DQAD ( Figure 2B upper panels). Among the Vas interactors we identified were Aub, Piwi, Fragile X Mental Retardation1 (FMR1) and eIF4A ( Figure 3C), which have been shown to be in complex with Vas also in early embryos (Megosh et al., 2006;Thomson et al., 2008). Ago3 was not among the interactors, in agreement with previous findings that Bombyx Vas directly associates with Siwi (B. mori Aub homolog) but not with Ago3 (Nishida et al., 2015).

Vasa activity in the germarium is essential for oogenesis
Absence of Vas in Drosophila females causes oogenesis arrest Ashburner, 1988, 1990). To determine at which stage of oogenesis Vas is required, we used either the vas-Gal4 or the matTub-Gal4 driver to express GFP-Vas WT at distinct stages of oogenesis: the vas promoter is active throughout oogenesis, whereas the matTub promoter is inactive in the germarium, but active during the subsequent stages (Supplementary Figure S1B). matTub-Gal4 driven expression of GFP-Vas WT in vas D1/D1 females fully rescued oogenesis in 3-day-old flies, but as these progressed in age, oogenesis arrested ( Figure 4A and Supplementary Figure S5A). In contrast, vas-Gal4 driven expression of GFP-Vas WT in the same vas D1/D1 background restored oogenesis independently of the age of the flies ( Figure 4A and Supplementary Figure S5A). Of note, expression of helicase inactive GFP-Vas DQAD and GFP-Vas GNT proteins did not rescue oogenesis, whichever Gal4-driver was used. In addition, analysis of egg-chamber development showed that ovarian atrophy takes place between oogenesis stage 6 and 8 and is a result of pyknosis (Supplementary Figure 5B).
To test whether absence of Vas in the germarium interferes with germ cell development, we determined the number of egg-chamber-producing ovarioles per female.
Strikingly, in the case of matTub-Gal4>GFP-Vas WT expressing vas D1/D1 and vas D1/D1 ; flies, the number of ovarioles decreased with the age of the females, whereas it did not, in the case of vas-Gal4>GFP-Vas WT expressing vas D1/D1 ; flies or of wild-type flies ( Figure 4B). Furthermore, egg-laying analysis showed that the number of eggs produced by vas D1/D1 ; matTub-Gal4>GFP-Vas WT females decreased with the age of the females and eventually stopped altogether ( Figure 4C left diagram). However, the hatching rate of eggs produced by vas D1/D1 females as a result of matTub-Gal4 or of vas-Gal4 GFP-Vas WT driven expression of GFP-Vas WT did not differ significantly ( Figure 4C right diagram). Taken together, these results indicate that progression and completion of oogenesis depends on the activity of Vas in the germarium.

Chk2 signaling in the germarium induces oogenesis arrest in vas mutant Drosophila
We recently showed that mnk (Chk2) and vas interact genetically, and that depletion of Chk2 signaling in loss of function vas D1/D1 flies rescues oogenesis, but that the embryos die due to severe DNA damage (Durdevic et al., 2018). Moreover, previous studies determined that Vas is phosphorylated in a Chk2-dependent manner (Abdu et al., 2002;Klattenhoff et al., 2007).
Furthermore, whereas vas-Gal4 driven silencing of mnk in vas D1/D1 females restored oogenesis, matTub-Gal4 driven knockdown of mnk did not ( Figure 5A). This indicates Chk2mediated signaling activity in the germarium determines the fate of developing egg-chambers.
Although the efficiency of vas-Gal4 RNAi-driven downregulation of mnk decreased over time, finally resulting in ovarian atrophy, we observed a more severe age-dependent decrease in the number of egg-chamber-producing ovarioles when mnk knockdown was driven by matTub-Gal4 ( Figure 5B). These results suggest that in vas mutants Chk2-signaling specifically in the germarium induces arrest of germ cell development.

Discussion
We have demonstrated that development of the Drosophila female germline depends on Vas activity in early oogenesis. Our data indicate that progression and completion of oogenesis require helicase active Vas. However, as our fusion proteins show low expression levels, we cannot rule out that when expressed at higher levels Vas might support oogenesis independently of helicase activity (Dehghani and Lasko, 2015). Using stage-specific promoters, we manipulated the expression of Vas and determined that activity of Vas in the germarium is crucial for germ cell lineage development. Our conclusion that oogenesis depends on an early helicase activity of Vas is consistent with the finding that Vas directly interacts with meiotic P26 (mei-P26) mRNA and activates its translation (Liu et al., 2009). Mei-P26 itself has been found to cooperate with proteins such as Bag of marbles and Sex lethal to promote both GSC self-renewal and germline differentiation (Li et al., 2012;Li et al., 2013).
In addition, we identified Vas interaction partners Lingerer, Rasputin, FMR1 and Caprin, which have been shown to cooperate in restricting tissue growth in a non-germline tissue, the Drosophila eye (Baumgartner et al., 2013). Interestingly, these proteins were found to interact in Drosophila ovaries as well (Costa et al., 2013;Costa et al., 2005) suggesting formation of a complex that could act in conjunction with Vas to control growth of Drosophila germline tissue. FMR1 was previously shown to interact with Vas in embryos and to be important for PGC formation (Deshpande et al., 2006;Megosh et al., 2006). In Drosophila ovaries, FMR1 was proposed to participate in the regulation of germline proliferation and GSC maintenance (Epstein et al., 2009;Yang et al., 2007). In addition, Vas interaction with numerous other factors implicated in promoting the self-renewal of GSC such as Rm62 (Ma et al., 2017), Bel (Kotov et al., 2016), and Nop60B (Kaufmann et al., 2003) indicates an intricate network of Vas-associated processes involved in sustaining the germ cell lineage. Further studies will be required to determine how these proteins collaborate to regulate early germ cell development.
In Drosophila, Chk2-signaling triggered by DNA damage, replication stress or nuclear lamina dysfunction induces GSC loss (Barton et al., 2018;Ma et al., 2016;Molla-Herman et al., 2015). Our previous study showed that removal of Chk2 in vas mutant flies fully restores oogenesis, while progeny embryos succumb to transposon up-regulation and DNA damage (Durdevic et al., 2018). Here we went further and genetically determined that, in vas mutants,

Data availability
The authors declare that all data supporting the findings of this study are available within the manuscript and its supplementary files.

Author contributions
Z.D. and A.E. conceived and designed the experiments. Z.D carried out the experiments and analysed the data. Z.D. and A.E. wrote the manuscript.

Conflict of interest
The authors state that there is no conflict of interest.