Transposon silencing in the Drosophila female germline ensures genome stability in progeny embryos

The piRNA pathway functions in transposon control in the germ line of metazoans. The conserved RNA helicase Vasa is an essential piRNA pathway component, but has additional important developmental functions. Here we address the importance of Vasa-dependent transposon control in the Drosophila female germline and early embryos. We find that transient loss of vasa expression during early oogenesis leads to transposon up-regulation in supporting nurse cells of the fly egg-chamber. We show that elevated transposon levels have dramatic consequences, as de-repressed transposons accumulate in the oocyte where they cause DNA damage. We find that suppression of Chk2-mediated DNA damage signaling in vasa mutant females restores oogenesis and egg production. Damaged DNA and up-regulated transposons are transmitted from the mother to the embryos, which sustain severe nuclear defects and arrest development. Our findings reveal that the Vasa-dependent protection against selfish genetic elements in the nuage of nurse cell is essential to prevent DNA damage-induced arrest of embryonic development.


Introduction
Transposons and other selfish genetic elements are found in all eukaryotes and comprise a large fraction of their genomes. Although transposons are thought to be beneficial in driving evolution (Levin and Moran, 2011), their mobilization in the germline can compromise genome integrity with deleterious consequences: insertional mutagenesis reduces the fitness of the progeny and loss of germ cell integrity causes sterility. Therefore, it is of great importance for sexually reproducing organisms to firmly control transposon activity in their germ cells. Metazoans have evolved a germline specific mechanism that, by relying on the activity of PIWI family proteins and their associated Piwi-interacting RNAs (piRNAs), suppresses mobile elements.
Drosophila harbors three PIWI proteins: Piwi, Aubergine (Aub) and Argonaute 3 (Ago3) which, guided by piRNAs, silence transposons at the transcriptional and posttranscriptional levels (reviewed in Guzzardo et al., 2013). Besides PIWI proteins, other factors such as Tudor-domain proteins and RNA helicases are involved in piRNA biogenesis and transposon silencing. Mutations in the majority of piRNA pathway genes in Drosophila females cause transposon up-regulation that leads to an arrest of oogenesis.
This effect can be rescued by suppression of the DNA damage checkpoint proteins of the ATR/Chk2 pathway (Chen et al., 2007;Klattenhoff et al., 2007;Pane et al., 2007). In contrast, inhibition of DNA damage signaling cannot restore embryonic development (Chen et al., 2007;Klattenhoff et al., 2007;Pane et al., 2007). Recent studies suggest that PIWI proteins might have additional roles during early embryogenesis independent of DNA damage signaling (Khurana et al., 2010;Mani et al., 2014). However, functions of the piRNA pathway during early embryonic development remain poorly understood.
One of the essential piRNA pathway factors with an important role in development is the highly conserved RNA helicase Vasa. First identified in Drosophila as a maternal-effect gene (Hay et al., 1988;Lasko and Ashburner, 1990;Schüpbach and Wieschaus, 1986), vasa (vas) was subsequently shown to function in various cellular and developmental processes (reviewed in Lasko, 2013). In the Drosophila female germline Vasa accumulates in two different cytoplasmic electron-dense structures: the pole (or germ) plasm at the posterior pole of the oocyte, and the nuage, the perinuclear region of nurse cells. In the pole plasm, Vasa interacts with the pole plasm inducer Oskar (Osk) (Breitwieser et al., 1996;Jeske et al., 2015) and ensures accumulation of different proteins and mRNAs that determine primordial germ cell (PGC) formation and embryonic patterning (Hay et al., 1988;Lasko and Ashburner, 1990). In the nuage, Vasa is required for the assembly of the nuage itself (Liang et al., 1994;Malone et al., 2009) and facilitates the transfer of transposon RNA intermediates from Aub to Ago3, driving the piRNA amplification cycle and piRNA-mediated transposon silencing (Nishida et al., 2015;Xiol et al., 2014). As Vasa's involvement in many cellular processes renders it difficult to analyze its functions in each process individually, it remains unknown whether Vasa's functions in development and in the piRNA pathway are linked or independent.
In this study, we address the role of Vasa in transposon control in Drosophila development. We find that failure to suppress transposons in the nuage of nurse cells causes DNA double-strand breaks (DSBs), severe nuclear defects, and lethality of progeny embryos. Even transient interruption of Vasa expression in early oogenesis de-represses transposons and impairs embryo viability. Depletion of the Drosophila Chk2 ortholog maternal nuclear kinase (mnk) restores oogenesis in vas mutants, but does not suppress defects in transposon silencing or DSB-induced nuclear damage and embryonic lethality.
We show that up-regulated transposons invade the maternal genome, inducing DNA DSBs that, together with transposon RNAs and proteins, are maternally transmitted and consequently cause embryogenesis arrest. Our study thus demonstrates that Vasa function in the nuage of Drosophila nurse cells is essential to maintain genome integrity in both the oocyte and progeny embryos, ensuring normal embryonic development.

Vasa dependent transposon control is not essential for oogenesis
Vasa is required for piRNA biogenesis and transposon silencing in Drosophila, as in vas mutants piRNAs are absent and transposons are up-regulated (Czech et al., 2013;Handler et al., 2013;Malone et al., 2009;Vagin et al., 2004;Zhang et al., 2012). To investigate the importance of transposon control in Drosophila development, we expressed wild-type GFP- We first assessed the ability of GFP-Vas WT fusion protein to promote transposon silencing in the female germline, and examined the effect of GFP-Vas WT on the level of expression of several transposons in vas mutant ovaries. We chose the long terminal repeat (LTR) retrotransposons burdock and blood, and the non-LTR retrotransposon HeT-A, which were previously reported to be up-regulated upon Vasa depletion (Czech et al., 2013;Vagin et al., 2004). The LTR retrotransposon gypsy, which belongs to the so-called somatic group of transposons and is not affected by vasa depletion, served as a negative control (Czech et al., 2013). Loss-of-function vas D1/D1 ovaries contained elevated levels of burdock, blood and HeT-A RNA ( Figure 1A). Remarkably, silencing of transposons by GFP-Vas WT in vas D1/D1 flies depended on which Gal4 driver was used (Supplementary Figure S1B-C): when driven by nos-Gal4, GFP-Vas WT had no effect on transposon levels, whereas when driven by vas-Gal4 it led to the re-silencing of transposons ( Figure 1A). This differential effect presumably reflects the stages of oogenesis at which the nos and vas promoters are active, and suggests that lack of Vasa between stages 2 and 6 of oogenesis

Loss of Vasa during early oogenesis affects viability of progeny embryos
Concentration of Vasa protein at the posterior pole of the embryo is essential for PGC and abdomen formation during embryogenesis (Hay et al., 1988;Lasko and Ashburner, 1990;Schüpbach and Wieschaus, 1986). We analyzed number of PGC positive embryos and the hatching rate of eggs produced by vas D1/D1 flies expressing GFP-Vas WT either under control of the nos or the vas promoter (vas D1/D1 ; nos-Gal4>GFP-Vas WT and vas D1/D1 ; vas-Gal4>GFP-Vas WT embryos). PGC formation was restored in approximately 50% of vas D1/D1 ; nos-Gal4>GFP-Vas WT and vas D1/D1 ; vas-Gal4>GFP-Vas WT embryos ( Figure 1C). However, DAPI staining revealed nuclear damage in some vas D1/D1 ; nos-Gal4>GFP-Vas WT embryos (see below), which we excluded from the quantification.
Expression of GFP-Vas WT also partially rescued the hatching of eggs produced by vas D1/D1 flies ( Figure 1D). However, expression of GFP-Vas WT led to a significantly lower hatching rate in vas D1/D1 ; nos-Gal4>GFP-Vas WT than in vas D1/D1 ; vas-Gal4>GFP-Vas WT flies ( Figure 1D). Expression of GFP-Vas WT in heterozygous loss-of-function vas D1/Q7 females led to a low hatching rate similar to vas D1/D1 (Supplementary Figure S2A vas-Gal4>GFP-Vas WT embryos was higher than that of vas D1/D1 ; nos-Gal4>GFP-Vas WT embryos, suggests that transient loss of vas expression during early oogenesis impairs viability of progeny embryos ( Figure 1D). One of the up-regulated transposons in vas mutants is HeT-A, whose RNA and protein expression is strongly de-repressed in piRNA pathway mutant ovaries (Aravin et al., 2001;Lopez-Panades et al., 2015;Vagin et al., 2006;Zhang et al., 2014). Analysis of HeT-A/Gag protein expression in 0-1 hour old embryos showed that levels of HeT-A/Gag were much higher in vas D1/D1 ; nos-Gal4>GFP-Vas WT than in vas D1/D1 ; vas-Gal4>GFP-Vas WT embryos ( Figure 2E). Additionally, we stained embryos with antibodies against HeT-A/Gag protein and observed that in cellularized wild-type embryos, HeT-A localized at distinct perinuclear foci ( Figure 3A panel a and Supplementary Figure S3B panel a), as previously described for HeT-A/Gag-HA-FLAG fusion protein (Olovnikov et al., 2016). In vas D1/D1 ; nos-Gal4>GFP-Vas WT embryos displaying nuclear damage HeT-A protein accumulated in large foci throughout the embryo ( Figure 3A

Chk2 mutation restores oogenesis but not transposon silencing and embryogenesis in vas mutants
To test genetically whether DNA damage signaling contributes to the oogenesis arrest of vas loss-of-function mutants (Hay et al., 1988;Lasko and Ashburner, 1990;Schüpbach and Wieschaus, 1986 and Figure Figure S2C). Taken together, these findings demonstrate that the oogenesis arrest of loss-of-function vas mutants results from activation of the Chk2mediated DNA damage-signaling checkpoint.
Although removal of mnk allowed oogenesis progression, it did not reduce transposon levels in vas D1/D1 , mnk P6/P6 ovaries and the eggs laid failed to hatch ( Figure 3B and Supplementary Figure S4B). Further analysis revealed that vas D1/D1 , mnk P6/P6 early embryos contained elevated levels of maternally transmitted transposon RNAs ( Figure 3C). This was also the case of ago3 single mutant embryos, which displayed nuclear damage (Mani et al., 2014;Supplementary Figure S4D) similar to that of vas D1/D1 ; nos-Gal4>GFP-Vas WT embryos ( severe nuclear defects and embryogenesis arrest. We conclude that tight regulation of transposons throughout oogenesis is essential to maintain genome integrity in the oocyte and in early syncytial embryo, hence for normal embryonic development.

Discussion
Our study shows that a transient loss of vas expression during early oogenesis leads to upregulation of transposon levels and compromised viability of progeny embryos. The observed embryonic lethality is due to DNA DSBs and nuclear damage that arise as a consequence of the elevated levels of transposon mRNAs and proteins, which are transmitted from the mother to the progeny. We thus demonstrate that transposon silencing in the nurse cells is essential to prevent maternal transmission of transposons and DNA damage, protecting the progeny from harmful transposon-mediated mutagenic effects.
Our finding that suppression of Chk2-mediated DNA damage signaling in loss-offunction vas mutant flies restores oogenesis and egg production demonstrates that Chk2 is epistatic to vas. However, hatching is severely impaired, due to the DNA damage sustained by the embryos. The defects displayed by vas, mnk double mutant embryos resembled those of PIWI (piwi, aub and ago3) single and mnk; PIWI double mutant embryos (Klattenhoff et al., 2007;Mani et al., 2014). Earlier observation that inactivation of DNA damage signaling does not rescue the development of PIWI mutant embryos led to the assumption that PIWI proteins might have an essential role in early somatic development, independent of cell cycle checkpoint signaling (Mani et al., 2014). By tracing transposon protein and RNA levels and localization from the mother to the early embryos we have shown that, independent of Chk2 signaling, de-repressed transposons are responsible for nuclear damage and embryonic lethality. Our study demonstrates that transposon insertions occur in the maternal genome where they cause DNA DSBs that together with transposon RNAs and proteins are passed on to the progeny embryos. Transposon activity and consequent DNA damage in the early syncytial embryo cause aberrant chromosome segregation, resulting in unequal distribution of the genetic material, nuclear damage and ultimately embryonic lethality. Our study shows that early Drosophila embryos are defenseless against transposons and will succumb to their mobilization if the first line of protection against selfish genetic elements in the nuage of nurse cell fails.
A recent study showed that in p53 mutants transposon RNAs are up-regulated accumulate at the posterior pole of the oocyte, without deleterious effects on oogenesis or embryogenesis (Tiwari et al., 2017). It is possible that the absence of pole plasm in vas mutants (Lehman and Ephrussi, 1994)  Transposon up-regulation in the Drosophila female germline triggers a DNA damage-signaling cascade (Chen et al., 2007, Klattenhoff et al., 2007. In aub mutants, before their oogenesis arrest occurs, Chk2-mediated signaling leads to phosphorylation of Vasa, leading to impaired grk mRNA translation and embryonic axis specification (Klattenhoff et al., 2007). Considering the genetic interaction of vas and mnk (Chk2) and the fact that Vasa is phosphorylated in Chk2-dependent manner (Abdu et al., 2002, Klattenhoff et al., 2007 it is tempting to speculate that phosphorylation of Vasa might stimulate piRNA biogenesis, reinforcing transposon silencing and thus minimizing transposon-induced DNA damage ( Figure 5). The arrest of embryonic development as a first, and arrest of oogenesis as an ultimate response to DNA damage thus prevents the spreading of detrimental transposon-induced mutations to the next generation.

Generation of mnk, vas double mutant flies
To generate mnk, vas double mutants, +, +, mnk P6 Supplementary Table S1 and sequences of primers used for RT-PCR reaction are shown in Supplementary Table   S7.

Fecundity and hatching assays
Virgin females of all genetic backgrounds tested were mated with w 1118 males for 24h at 25°C. The crosses were then transferred to apple-juice agar plates, and eggs collected in 24h intervals over 3-4 days. The number of eggs laid on each plate was counted; the plates were kept at 25°C for 2 days, then the number of hatched larvae counted. Experiments were performed in three independent replicates, and w 1118 females were used as a control.

Ovarian morphology and Vasa localization analysis
Ovaries of 3-7-days old flies were dissected in PBS. Ovarian morphology was evaluated under a Olympus SZX16 stereo microscope. Vasa localization was assessed in ovaries of 3-7-days old flies expressing the GFP-Vasa proteins after fixation in 2% PFA and 0.01% Triton X-100 for 15 min at RT. Fixed ovaries were mounted on glass slides and GFP fluorescence examined under a Zeiss LSM 780 confocal microscope. Vasa localization in wild-type and vas mutant ovaries and progeny embryos was analyzed by antibody staining (see below). Nuclei were visualised with NucBlue Fixed Cell Stain (Thermofisher).

Fluorescent in situ RNA hybridization
All FISH experiments were performed as described in Gaspar et al., 2018. 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. Embryos (1-3h) were collected and dechorionated in 50% bleach, fixed for 25 min at RT in the heptane/2% PFA interface and devitellinized by vigorous shaking after adding 1V of methanol. 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. Proteinase K treatment was omitted when samples were subsequently to be immunohistochemically stained (see bellow). Samples were prehybridized 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 pre-warmed 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.
For simultaneous FISH and immunohistochemical staining, ovaries and embryos were fixed as described above. Samples were simultaneously incubated with fluorescently labeled oligonucleotides (12.5-25 nM) complementary to HeT-A RNA and primary antibodies against γH2Av (rabbit; 1:5000; Rockland) or HeT-A/Gag (rabbit 1:100; gift of Elena Casacuberta) overnight at 28°C in PBT. Samples were washed 2 times for 20 min at 28°C in PBT and subsequently incubated with secondary Alexa 488 conjugated goat antirabbit antibodies (1:1000; Invitrogen). The samples were mounted in 80% 2,2-thiodiethanol in PBS and analyzed on a Leica SP8 confocal microscope.  Table S7) 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.
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.

RNA extraction and quantitative PCR analysis
Total RNA was extracted from ovaries of 3-7-day old flies or 0-1h old embryos using Trizol reagent (Thermofisher). For first-strand cDNA synthesis, RNA was reverse-transcribed using a QuantiTect Reverse Transcription Kit (QIAGEN). Quantitative PCR was performed on a StepOne Real-Time PCR System (Thermofisher) using SYBR Green PCR Master Mix (Thermofisher). Relative RNA levels were calculated by the 2 -ΔΔCT method (Livak and Schmittgen, 2001) and normalized to rp49 mRNA levels for ovaries, and 18S rRNA for embryos. Fold-enrichments were calculated by comparison with the respective RNA levels in w 1118 flies. Sequences of primers used for qPCR reaction are shown in Supplementary   Table S7.

Data availability
The authors declare that all data supporting the findings of this study are available within the manuscript and its supplementary files.  Table S2). P-value was determined by Student's t-test. vas-Gal4>GFP-Vas WT flies. Tubulin was used as a loading control. Table shows quantification of HeT-A/Gag protein levels relative to wild-type. HeT-A/Gag signal was normalized to Tubulin signal in individual experiments and was set to 1 in wild-type.