Aub, Vasa, and Armi concentration in phase-separated nuage is dispensable for piRNA biogenesis and transposon silencing

Abstract

The piRNA biogenesis machinery localizes to phase-separated nuage granules, but nuage function is not well understood. We therefore assayed nuage composition, piRNA expression, and transposon silencing in Drosophila mutants that disrupt piRNA precursor production and nuclear export, ping-pong amplification, and phased piRNA biogenesis. These mutations destabilize the genome and activate Chk2 signaling, and chk2/mnk double mutants were therefore analyzed in parallel. Aub and Vasa are required for ping-pong amplification, and Armi promotes phased piRNA processing. We show that Chk2 activation releases Aub and Vasa from nuage and that piRNA precursors are required for nuage localization of the ping-pong and phased biogenesis machinery. However, this analysis also indicates that Vasa, Aub, and Armi concentration in nuage is dispensable for piRNA production and transposon silencing, indicating that dispersed cytoplasmic proteins can drive these processes. We speculate that nuage sequesters silencing effectors, which are released by Chk2 in response to transposon mobilization.

INTRODUCTION

Twenty-four to thirty-one nucleotides long PIWI interacting RNAs (piRNAs) have a conserved function in silencing transposons in the germline (Vagin et al. 2006; Aravin et al. 2007; Brennecke et al. 2007; Siomi et al. 2011). This genome defense system is best understood in Drosophila melanogaster, where the most abundant germline piRNAs originate from pericentromeric and subtelomeric “clusters” composed of nested transposon fragments, but piRNAs are also derived from a subset of isolated transposon insertions (Bergman et al. 2006; Brennecke et al. 2007; Mohn et al. 2014). piRNA clusters are bound by the Rhino–Deadlock–Cutoff complex (RDC) (Klattenhoff et al. 2009; Mohn et al. 2014; Zhang et al. 2014; Chen et al. 2016), which promotes transcription from both genomic strands and suppresses splicing and polyadenylation, leading to production of piRNA precursor transcripts (Mohn et al. 2014; Zhang et al. 2014; Andersen et al. 2017). UAP56 and the THO complex are components of the transcript and export complex (TREX), which binds piRNA precursors and facilitates nuclear export by the noncanonical Nxf3–Nxt1 mediated Crm1 export pathway (Zhang et al. 2012, 2018; Hur et al. 2016; ElMaghraby et al. 2019; Kneuss et al. 2019; Mendel and Pillai 2019).

Following nuclear export, piRNAs are processed by proteins that localize to perinuclear “nuage” granules, mitochondria, or both. “Nuage” was first identified as germline-specific electron-dense material associated with the nuclear periphery and mitochondria (André and Rouiller 1956) and subsequently found in germ cells of evolutionarily diverse organisms, including sea urchins, zebrafish, frogs, nematodes, flies, mice, rats, and humans (Eddy 1975). Little was known about the function and composition of nuage until Liang et al. (1994) found that Drosophila Vasa, a germline-specific DEAD-box RNA binding protein, localizes to nuage granules. The PIWI clade Argonautes Aub and Ago3, which bind piRNAs, were subsequently found to colocalize with Vasa at perinuclear nuage granules, and all three proteins are required for processing of piRNA precursors by the ping-pong amplification cycle (Brennecke et al. 2007; Lim and Kai 2007; Malone et al. 2009; Webster et al. 2015). Ago3 catalyzed cleavage also generates precursors for phased piRNA biogenesis, which produces piRNAs through processive processing of precursors (Wang et al. 2015). The products of phased biogenesis are loaded into Piwi, which localizes to the nucleus and guides transcriptional silencing (Cox et al. 2000; Kalmykova et al. 2005; Klenov et al. 2011; Wang et al. 2015). Phased biogenesis requires the helicase Armi and endonuclease Zuc and Armi localizes to nuage and mitochondria and Zuc localizes to the outer mitochondrial membrane (Han et al. 2015; Mohn et al. 2015; Ge et al. 2019). These findings support a model in which nuage is the site of both ping-pong amplification and production of the precursors for phased biogenesis.

Prominent perinuclear nuage granules containing Aub and Vasa also localize across the nuclear envelope from piRNA clusters marked by the RDC, and Aub and Vasa are displaced from nuage in uap56 mutants, which disrupt piRNA precursor export from the nucleus (Zhang et al. 2012). Mutations that disrupt piRNA precursor production and ping-pong amplification also block Aub and Vasa localization (Klattenhoff et al. 2009; Pane et al. 2011; Zhang et al. 2012). Based on these findings, we previously proposed that clusters are linked to perinuclear nuage by the export machinery, forming a piRNA production and processing compartment that spans the nuclear envelope (Zhang et al. 2012).

Loss of Aub and Vasa localization to nuage in uap56 mutants supports a role for precursor export in nuage assembly (Zhang et al. 2012), but piRNA pathway mutations lead to genome instability and checkpoint kinase 2 (Chk2) DNA damage signaling, which could trigger secondary defects in nuage organization. We therefore compared piRNA production, transposon silencing, and nuage organization in uap56 single mutants and uap56 double mutants with mnk, which encodes the Drosophila homolog of Chk2. Aub and Vasa localize to nuage in mnk; uap56 double mutants, but piRNA biogenesis and transposon silencing are not restored. We extended this analysis by systematically evaluating localization of the piRNA machinery, transposon silencing, and piRNA biogenesis in single and mnk double mutants in genes that drive piRNA precursor expression and export, ping-pong amplification, phased piRNA biogenesis, and transcriptional silencing. These studies indicate that Chk2 activation displaces the ping-pong amplification machinery from nuage and defines a nuage assembly pathway, but also show that concentration of Vasa, Aub, and Armi in nuage is not required for piRNA biogenesis or transposon silencing. Nuage function is reconsidered in light of these findings.

RESULTS

Chk2-dependent modification of perinuclear nuage

To determine if loss of Aub and Vasa from nuage in uap56 mutants is secondary to DNA damage signaling, we localized nuage components in uap56^sz15/28 double mutant with a null allele of mnk, which encodes the Drosophila Chk2 homolog. For this analysis, we assayed the ping-pong factors Vasa, Aub, and Ago3, which localize to nuage, and the helicase Armi, which promotes phased piRNA biogenesis and associates with nuage and mitochondria. To quantify localization, we used laser scanning confocal imaging to optically section labeled egg chambers and computationally defined nuage as foci adjacent to the nuclear envelope with signal 4 SD above average signal for the image (see Materials and Methods). We confirmed specificity of the Vasa, Aub, and Ago3 antibodies by staining corresponding mutant ovaries, which showed no perinuclear localization (Supplemental Fig. S1A–C).

In control egg chambers, Aub and Ago3 colocalize with Vasa in prominent perinuclear foci (filled arrows, Fig. 1A–F), and Aub and Ago3 show essentially identical distributions (Supplemental Fig. S1G–K). Confirming these qualitative observations, image quantification shows that nuage signal for Aub and Ago3 broadly correlates with Vasa signal (Fig. 1G–J). Armi also colocalizes with Vasa in perinuclear nuage (Supplemental Fig. S1D–F; Ge et al. 2019), indicating that Vasa, Aub, Ago3, and Armi are components of that same prominent nuage foci. In uap56^sz15/28 mutants, in contrast, direct inspection and quantification of fluorescent signal shows that Aub is dispersed in the cytoplasm (Fig. 2E,K), and Vasa and Ago3 localization to perinuclear granules is reduced (Fig. 2B,K; Supplemental Fig. S2B,E,K). In striking contrast, nuage localization of Aub, Ago3, and Vasa is restored in mnk, uap56^sz15/28 (Fig. 2C,F,L; Supplemental Fig. S2C,F,L).

FIGURE 1.

Colocalization of the ping-pong machinery to prominent nuage granules. (A–C) Confocal images of w¹ nurse cell nucleus double labeled for Vasa (green) and Aub (red) show colocalization of the two nuage markers (filled arrow heads) along with instances where there is Vasa localized without Aub (open arrow heads). Scale bar, 2 μm. (D–F) Confocal images of w¹ nurse cell nucleus double labeled for Vasa (green) and Ago3 (red) show similar trends as above. (G) Density scatterplot comparing average Vasa signal versus average Aub signal where each point represents a nuage granule in w¹. Only granules proximate to nuclear pore signal were used for this analysis. The Pearson correlation coefficient was calculated for each comparison. (H) Violin plot showing distribution of Aub/Vasa signal ratios, where 0 means equal amount of both signals, positive values indicate a bias toward more Aub signal, and negative values indicate a bias toward more Vasa signal. (I,J) Same image quantification analysis as G and H, respectively, except using Ago3 instead of Aub as the second nuage marker.

FIGURE 2.

Chk2 regulates perinuclear nuage composition in uap56 piRNA mutant. (A–I) Immunofluorescence for Vasa (green) and Aub (red), in heterozygous controls, uap56^sz15/28, and mnk, uap56^sz15/28 double mutants. Scale bar, 2 μm. (J–L) Scatterplot comparing average Vasa signal versus average Aub signal where each point represents a perinuclear nuage granule for heterozygous controls, uap56^sz15/28, and mnk, uap56^sz15/28 double mutants, respectively. (M) Violin plot showing distribution of Aub/Vasa signal ratios. (N) Western blot images probed for anti-Vasa (red) and α-tubulin (green). The lanes are as follows: Ladder (L), w¹ (WT control, 1), uap56^sz15/28 (2), and mnk, uap56^sz15/28 (3), and vas^RG/PH (negative control, 4). The black and white arrowheads point to 75 and 50 kDa sizes, respectively. (O) Quantification of western blots looking for Vasa protein levels relative to α-tubulin (n = 4). t-test was performed and degree of significant differences for the indicated comparisons is above the bars, and not significant is marked as (nsec).

Loss of Aub, Ago3, and Vasa from perinuclear foci in uap56 mutants is therefore dependent on Chk2, presumably through activation of the kinase. This could displace these proteins from nuage granules, disrupt nuage structure, or decrease steady-state protein expression. All three gene transcripts are expressed at comparable levels in uap56^sz15/28, w¹, and mnk, indicating that the Chk2 activation does not block transcription or destabilize cognate mRNAs (data not shown). In addition, quantitative western blots show that Vasa protein is expressed at similar levels in w¹, uap56^sz15/28 and mnk, uap56^sz15/28 double mutants (Fig. 2N,O). The vasa^RG allele carries single nucleotide deletion that reduced coding transcript expression by five- to sevenfold, and Vasa protein is also reduced by approximately sevenfold, confirming specificity of the western blots (Fig. 2N,O, vasa^RG/PH). We were unable to directly assay Aub and Ago3 levels, but uap56 mutations do not reduce total piRNAs or the ping-pong signature for piRNAs mapping to a significant fractions of transposon families (see below), indicating that the protein function is retained. Armi is required for phased piRNA biogenesis and localizes to nuage and mitochondria. Armi fails to accumulate in nuage in uap56^sz15/28 and mnk, uap56^sz15/28 double mutants (Supplemental Fig. S2M). Chk2 activation thus appears to disrupt Vasa, Aub, and Ago3 localization to perinuclear granules, while UAP56, presumably through the TREX-dependent export of piRNA precursors, appears to have a direct role in Armi localization to nuage.

To determine if loss of Aub, Ago3, Vasa, and Armi accumulation in perinuclear granules reflects loss of nuage structure, we used thin section transmission electron microscopy to directly assay for electron-dense perinuclear structures in uap56^sz15/28 and mnk, uap56^sz15/28 mutants. As shown in Figure 3A–C, electron-dense nuage is present in both genetic backgrounds. Chk2 activation thus reduces Vasa, Aub, Ago3, and Armi association with nuage, but does not block nuage assembly.

FIGURE 3.

Electron-dense nuage granules detected in various piRNA mutants. (A) Electron microscopy (EM) images of Stages 3–5 nurse cells for WT (w¹) control along with various piRNA mutants show electron-dense nuage structures designated by the black arrowheads. (N) Nucleus, (C) cytoplasm, and (M) mitochondria. (B,C) There are no detectable differences in nuage structure between uap56^sz15/28 and mnk^-/-, uap56^sz15/28 double mutants. (D,E) Flies carrying the vas^RG/PH allelic mutation produce minimal Vasa early in oogenesis, but Vasa is undetectable in the stages imaged. Despite this, intact nuage structures were still observed via EM. vas^PH homozygous flies contains a complete deletion of the vas open reading frame still show intact electron-dense nuage structure via EM.

Electron-dense nuage granules detected in various piRNA mutants

Liang et al. (1994) did not detect electron-dense nuage in ovaries hemizygous for two hypomorphic vas mutant alleles that block Vasa immunolocalization to perinuclear granules (Liang et al. 1994). However, we found that nuage persists in uap56^sz15/28 mutants, which also displace Vasa from nuage. We therefore determined the impact of genetically eliminating Vasa protein on nuage assembly by performing thin section transmission electron microscopy (EM) on ovaries homozygous for a deletion of the vas locus (vas^PH/PH). To eliminate potential defects associated with Chk2 activation, the deletion was combined with mnk null allele. We also analyzed nuage formation in vas^RG/PH mutant ovaries, and the corresponding mnk double mutants. The vas^RG allele disrupts Aub localization of perinuclear granules (Supplemental Fig. S3B) and significantly reduces Vasa protein expression (Fig. 2N). Electron-dense perinuclear granules were present in vas^PH/PH null and vas^RG mutant ovaries (Fig. 3D–F). Ago3 and Armi also localize to perinuclear granules in vas^RG/PH mutants, supporting the presence of nuage (Supplemental Fig. S3C,D). Since Aub does not localize to perinuclear granule in vas^RG/PH or mnk, vas^RG/PH mutants (Supplemental Fig. S3B), Vasa is required to recruit Aub to nuage, but is dispensable for nuage assembly.

To determine if piRNA precursors, ping-pong amplification, and phased piRNA biogenesis are required for nuage assembly, we performed thin section transmission electron microscopy on rhi^KG/02, thoc7^d/Df, aub^HN/QC, and armi^1/72.1 mutant ovaries (Supplemental Fig. S4). As observed in uap56 and vas mutants, electron-dense perinuclear nuage was present in all of these genotypes. In addition, Ago3 localizes to perinuclear granules in thoc7^d/Df, aub^HN/QC, and mnk; armi^1/72.1 mutants (Supplemental Fig. S3C), supporting the presence of nuage structure. In contrast, nuage localization of Aub, Ago3, Vas, and Armi is disrupted in rhi^KG/02 and mnk; rhi^KG/02 mutants (Supplemental Fig. S3). Rhino is required for piRNA cluster transcription, implying that piRNA precursors are required to recruit the downstream ping-pong and phased biogenesis machinery to nuage.

piRNA production and transposon silencing

Localization of piRNA pathway proteins to perinuclear granules is generally thought to have an important role in biogenesis and transposon silencing (Lim and Kai 2007; Batista et al. 2008; Aravin et al. 2009), and uap56^sz15/28 allelic combination displaces Aub, Vasa, and Armi from nuage (Zhang et al. 2012). To determine if the defects in piRNA biogenesis and transposon silencing in uap56^sz15/28 mutants are suppressed by mnk, which restores Aub and Vasa localization, we sequenced short and long RNAs from mnk controls, uap56^sz15/28, and mnk, uap56^sz15/28 double mutant ovaries and measured piRNA abundance and transposon expression. Total piRNA expression is reduced in uap56^sz15/28 single mutants and mnk, uap56^sz15/28 double mutants (Fig. 4A), consistent with a direct role for UAP56 in piRNA biogenesis. Antisense piRNAs targeting transposons are approximately twofold more abundant, and transposon expression for some families is approximately twofold lower in mnk, uap56^sz15/28 double mutants relative to uap56^sz15/28 single mutants (Fig. 4A,D). The ping-pong cycle produces piRNAs from opposite strands that overlap by 10 nt. Despite the twofold change in piRNA abundance, ping-pong overlap bias is similar in mnk, uap56^sz15/28 double mutants, uap56^sz15/28 single mutants, and mnk controls (Fig. 4B). Additionally, long piRNA precursor expression from clusters is also similar in mnk, uap56^sz15/28 single mutants and mnk, uap56^sz15/28 double mutants (Fig. 4E). Within the context of uap56 mutants, loss of Aub and Vasa from nuage thus has a relatively modest impact on piRNA ping-pong amplification.

FIGURE 4.

Chk2 activation shows subtle effects on piRNA production and transposon silencing. (A) Scatterplot comparing antisense transposon mapping piRNA abundance (log₁₀) in uap56^sz15/28 and mnk, uap56^sz16/28 double mutants relative to mnk controls. Each point represents a transposon family. Red and blue indicate the twofold increase and decrease, respectively. (B) Scatterplot comparing ping-pong Z-scores (log₂) detected for each transposon family with the same setup as above. (C) Armi is required for phased piRNA biogenesis and is lost from nuage in uap56^sz15/28 and mnk, uap56^sz15/28 double mutants. However, phasing is comparable to controls in both genotypes. Phasing was measured by determining the shortest distance between the 3′ end of one piRNA and the 5′ end of the subsequent piRNA downstream. Zero nucleotide distance represents the start of a piRNA abutted directly downstream from the end of another piRNA with no nucleotide separating the two. The primary 0 nt and secondary 28 nt peak in mnk, uap56^sz15/28 and mnk, uap56^sz16/28 double mutants are comparable to controls, indicating that phasing is intact when Armi is displaced from nuage. (D) Scatterplots show pairwise comparisons of transposon expression levels (log₁₀ rpkm) between mnk control, uap56^sz15/28, and mnk, uap56^sz15/28 double mutants. Each point represents a transposon family. Red and blue indicate the twofold increase and decrease, respectively. A subset of transposons are significantly overexpressed in uap56 and mnk doubles relative to controls. However, recovery of nuage localization in the double mutants is not associated with a consistent increase in silencing. (E) Scatterplots with same pairwise comparisons as above showing long RNA expression from piRNA clusters. piRNA cluster expression is less sensitive to activation of Chk2.

The phased biogenesis pathway produces piRNAs that are loaded into Piwi and guide transcriptional silencing (Han et al. 2015; Mohn et al. 2015). Armi is proposed to shuttle precursor from nuage to mitochondria during phased biogenesis (Ge et al. 2019). Since Armi does not localize to nuage in uap56^sz15/28 or mnk, uap56^sz15/28 mutant ovaries, it is possible that phased piRNA biogenesis is disrupted. To determine if phasing is compromised, we quantified the distance between the 3′ ends of one piRNA and the 5′ end of the downstream piRNA. A primary peak at 0 nt and secondary peak at 28 nt are hallmarks of phased processing (Han et al. 2015). These peaks are essentially identical in mnk controls, uap56^sz15/28, and mnk, uap56^sz15/28 double mutants (Fig. 4C). Phased piRNA biogenesis thus does not require Armi concentration in nuage. However, Armi could transiently associate with perinuclear granules in uap56^sz15/28 mutants.

Spatial juxtaposition of perinuclear nuage and nuclear clusters

Nuage granules are biased toward regions of the nuclear envelope opposite piRNA clusters (Zhang et al. 2012). uap56 mutations disrupt TREX-mediated cluster transcript nuclear export and Aub and Vasa localization to nuage, leading us to propose that piRNA precursor export organizes perinuclear nuage (Zhang et al. 2012). However, Aub and Vasa localization to nuage is restored in mnk, uap56^sz15/28 double mutants. To determine if these granules are associated with clusters, we quantified the spatial relationship between clusters and perinuclear granules in control, uap56^sz15/28, and mnk, uap56^sz15/28 double mutants. For this analysis, we triple labeled egg chambers for Rhino, the nuclear envelope, and the nuage-associated PIWI proteins Aub or Ago3. Triple labeling for Rhino, nuclear pore, and Armi was also performed as a negative control, since Armi does not localize in the uap56^sz15/28 or mnk, uap56^sz15/28. Labeled egg chambers were optically sectioned by laser scanning confocal microscopy (Fig. 5A–I), and clusters and nuage were defined as foci with signal 3 SD above average for the field. Clusters at the nuclear periphery were defined by proximity to the nuclear envelope, and the fraction of clusters at the periphery with nuage directly across the nuclear envelope, termed percent juxtaposed, was calculated (see Materials and Methods).

FIGURE 5.

UAP56 is required for Armi nuage localization and organizing nuage adjacent to nuclear piRNA machinery. (A–C) Confocal immunofluorescence images of WT control (w¹) nurse cell nucleus stages 6–8 triple labeled for Rhino (green), nuclear pore (blue), and a nuage marker (red), Aub, Ago3, or Armi. White arrowheads show instances where peripheral Rhino is juxtaposed to cytoplasmic nuage marker. (D–F) uap56^sz15/28 stained the same as WT show no Aub and Armi localization and little to no detectable juxtaposition events between nuage marker and Rhino. (G–I) mnk, uap56^sz15/28 stained the same as WT show partial restoration of juxtaposing for Aub and Ago3; however, localization of Armi remains perturbed. (J–L) Images from the staining experiments above were quantified for percent of peripheral rhino juxtaposed to Aub, Ago3, or Armi (see Materials and Methods). Box plot shows the percent juxtaposed distribution for w¹, uap56^sz15/28, and mnk, uap56^sz15/28, where each point represents a nucleus. P-value shown whenever significantly different.

In w¹ control egg chambers, an average of 52% and 58% of clusters at the nuclear periphery are juxtaposed to perinuclear nuage marked by Aub or Ago3, respectively (Fig. 5A,B,J,K). In uap56^sz15/28 single mutants, very little Aub is detected in nuage, and juxtaposition collapses to 2% (Fig. 5D,J). In contrast, Ago3 localization to nuage is reduced but not eliminated, and juxtaposition drops to 31%, suggesting that some “linkage” persists (Fig. 5E,K). In mnk, uap56^sz15/28 double mutants, Aub juxtaposition increases to 20% (Fig. 5G,J), and Ago3 juxtaposition increases to 49%, which is not significantly different from controls (Fig. 5K,H). Armi does not localize in either uap56^sz15/28 nor mnk, uap56^sz15/28 double mutants, and juxtaposition with Rhino ranges from 0% to 3% in both genotypes (Fig. 5F,I,L). These observations indicate that UAP56, presumably through precursor export, enhances association of perinuclear nuage with clusters, but also suggest that a UAP56-independent mechanism also contributes to wild-type juxtaposition.

Chk2 modification of piRNA pathway mutations

To determine the impact of Chk2 signaling on additional piRNA pathway mutants, we compared pathway organization, piRNA production, and transposon silencing in single and mnk double mutants that disrupt piRNA precursor transcription (rhi), precursor nuclear export (thoc7), ping-pong amplification (aub, ago3, and vas), and phased biogenesis (armi). To allow direct comparisons with published phenotypes, we analyzed well-characterized allelic combinations of these genes (see Materials and Methods). For each single and double mutant combination, we immunolocalized Aub, Ago3, Vasa, and Armi localization and assayed transposon silencing and piRNA production by sequencing short and long RNAs.

Eliminating Chk2 signaling reveals hierarchy of protein recruitment to nuage

The HP1 homolog Rhino anchors that RDC, which promotes piRNA cluster transcription (Klattenhoff et al. 2009; Mohn et al. 2014; Zhang et al. 2014). Mutations in rhi displace Vasa, Aub, Ago3, and Armi from perinuclear nuage and produce a profound reduction in germline piRNA precursors and mature piRNAs (Klattenhoff et al. 2009; Mohn et al. 2014). The piRNA production and protein localization defects are not suppressed by mnk (Supplemental Figs. S3, S5A), implying that RDC-dependent cluster transcripts are required to recruit Vasa, Aub, Ago3, and Armi to nuage.

The TREX functions downstream from Rhino in precursor transcript export (Hur et al. 2016; Zhang et al. 2018; ElMaghraby et al. 2019; Kneuss et al. 2019; Mendel and Pillai 2019). Intriguingly, mnk, uap56^sz15/28 and thoc7^d/Df mutations disrupt the TREX but do not block Ago3 localization to nuage (Supplemental Figs. S2F, S3C) and show significantly less severe defects in germline piRNA production than rhi mutants (Fig. 4A; Supplemental Fig. S5A,B). This suggests that the piRNA precursors are exported and processed into piRNAs independent of TREX. In addition, uap56 and thoc7 mutations displace Vasa from nuage, but localization is restored in mnk; uap56^sz15/28 and mnk; thoc7^d/Df double mutants (Fig. 2A–C; Supplemental Fig. S3A). In contrast, Armi localization to nuage is disrupted in both uap56, thoc7, and rhi single and mnk double mutants (Supplemental Figs. S2F,M, S3C,D). RDC-dependent transcripts thus appear to be essential to Ago3, Aub, Vasa, and Armi localization to nuage, but nuclear export of these transcripts appears to be mediated by TREX-dependent and TREX-independent mechanisms, with TREX-dependent transport required for Armi localization.

Extending this analysis to mutations that disrupt ping-pong amplification revealed a series of interdependencies that determine nuage composition. Mutations in both ago3 and vasa displace Aub from nuage, and Aub localization is not restored in mnk double mutants (Supplemental Fig. S3B). In contrast, Ago3 and Vasa localize to nuage in aub^HN/QC or mnk, aub^HN/QC mutants (Supplemental Fig. S3A,C). Aub localization to nuage thus requires both Ago3 and Vasa. In contrast, Vasa localizes to nuage in ago3^t2/t3 and mnk; ago3^t2/t3 mutations (Supplemental Fig. S3A), and Ago3 localizes to nuage in vasa^RG/PH or mnk, vas^RG/PH mutants (Supplemental Fig. S3C). Ago3 and Vasa thus localize to nuage through independent mechanisms.

Uncoupling protein localization and nuage function

To determine the impact of Chk2 signaling on piRNA biogenesis and transposon silencing, we assayed small and long RNA expression in single mutants and the corresponding mnk double mutants. For each mutant pair, we quantified antisense transposon mapping piRNAs, ping-pong Z-score, transposon, and piRNA cluster expression, as described for uap56 mutants (Supplemental Figs. S5, S6). All of the mnk double mutations analyzed show significant defects in piRNA expression and transposon silencing relative to mnk (Supplemental Figs. S5A,B, S6A) and w¹ (Supplemental Fig. S7) control strains, supporting primary functions for the genes in the piRNA pathway. Most of the mnk double mutants also show comparable levels of antisense transposon mapping piRNAs, ping-pong Z-score, and phased piRNA production relative to the corresponding single mutants (Supplemental Fig. S5), with thoc7 and uap56 mutants being the only exceptions. Thoc7 functions with UAP56 in the TRE, and mnk; thoc7 double mutants show a modest increase in antisense piRNA production relative to the single mutants (Fig. 4A,B; Supplemental Fig. S5A,B), and Vasa localization to nuage is restored in the double mutants (Fig. 2; Supplemental Fig. S3A). Within the context of both TREX mutants, Vasa localization to nuage is linked to a modest increase in piRNA biogenesis. This suggests that there may be other drivers of piRNA production in addition to Vasa location to nuage.

Armi is proposed to shuttle precursors from nuage to mitochondria during phased piRNA biogenesis (Ge et al. 2019). However, Armi localization to nuage is disrupted in rhi, uap56, and thoc7 single and mnk double mutants (Supplemental Figs. S2M, S3D), and phased piRNA biogenesis is comparable to wild type in all of these single and double mutants (Fig. 4C; Supplemental Fig. S5C). Armi accumulation at nuage thus appears to be dispensable for phased piRNA production.

A separation of function allele of vas

Vasa localizes to the posterior pole and nuage and is required for embryonic patterning and ping-pong piRNA amplification (Schupbach and Wieschaus 1986, 1991; Malone et al. 2009; Xiol et al. 2014). The vas^RG allele was identified in a screen for patterning mutations and we find that piRNA production, ping-ping amplification, and transposon silencing are also disrupted in vas^RG hemizygous mutant ovaries (Supplemental Figs. S5A,B, S6A). The vas^RG allele does not carry a stop or missense point mutation (Liang et al. 1994), but our RNA sequencing revealed a single nucleotide deletion near the C terminus and the resulting frameshift that truncates the protein before domain VI, which is conserved in all DEAD box proteins (see Materials and Methods). This domain is essential to RNA unwinding by eIF-4A, the founding member of the DEAD box family (Pause and Sonenberg 1992; Pause et al. 1993). The vas^RG allele thus appears to be catalytically dead and disrupts both known functions of Vasa.

In contrast, vas^PD appears to selectively disrupt Vasa function in posterior patterning (Schupbach and Wieschaus 1986, 1991). Both homozygous and hemizygous vas^PD combinations produce identical, severe axis specification and fertility defects (Schupbach and Wieschaus 1986, 1991). However, our small and long RNA sequencing shows that transposon mapping antisense piRNA levels, ping-pong bias, and transposon silencing are similar between wild type, w¹ controls and vas^PD/PH mutant ovaries (Fig. 6A–C). Nonetheless, we also find that Aub is dispersed in the nurse cell cytoplasm and does not accumulate in perinuclear granules (Fig. 6D). In addition, we also find that vas mRNA expression is reduced by ∼100-fold (Fig. 6E) and that Vasa protein is not detectable by immunofluorescence labeling and confocal imaging (Fig. 6F). These findings indicate that transposon silencing and ping-pong amplification in adult ovaries do not require Aub localization to nuage or significant Vasa expression.

FIGURE 6.

vas^PD/PH allele shows uncoupling of Aub localization to nuage from piRNA biogenesis, ping-pong, and transposon silencing. (A) Scatterplot shows antisense transposon mapping piRNA abundance (log₁₀) of vas^PD/PH or mnk, vas^PD/PH double mutants compared to w¹ control. Each point represents a transposon family. Red and blue indicate the twofold increase and decrease, respectively. piRNA production is comparable between vas^PD/PH and WT controls and unaffected by activation of Chk2. (B) Scatterplot showing ping-pong Z-scores (log₂) detected for each transposon family with the same comparisons as above. Transposon ping-pong Z-scores are scattered around the diagonal suggesting ping-pong is comparable between vasa^PD/PH mutants and w¹ controls. (C) Long RNA sequencing was performed on vas^PD/PH and compared to w¹ controls. vas^PD/PH show no global changes in transposon and piRNA cluster. (D) OreR control and mnk, vas^PD/PH fly ovaries were stained for Aub, nuclear pore, and Rhino. Rhino localization remains intact in mnk, vas^PD/PH; however, Aub localization to nuage is disrupted. Scale bar, 6 μm. (E) Long RNA sequencing was performed on vas^PD/PH and compared to w¹ controls. Gene expression shows no global changes, while vas mRNA expression is decreased by ∼100-fold. (F) w¹ controls, vas^RG/PH, and vas^PD/PH fly ovaries were stained for Vasa. Images show germarium (toward the left) and increasingly older egg changers (toward the right). w¹ control flies show Vasa localization in germarium and in all stages of the egg chambers. vas^RG/PH allelic combination is a protein null mutant and shows little to no Vasa protein at any stages, while low levels of Vasa are detected only in early germarium of vas^PD/PH mutants. Scale bar, 20 μm.

The vas^PD allele does not alter protein-coding capacity (Liang et al. 1994). The extremely low levels of wild-type Vasa expressed in adult vas^PD ovaries could be sufficient to catalytically drive the ping-pong cycle. Alternatively, vas^PD could support wild-type protein expression in pre-adult stages and initiate the ping-pong cycle, which is subsequently propagated by Aub and Ago3. The available data cannot distinguish between these alternatives, but indicate that concentration of Aub and Vasa in nuage is not required for piRNA biogenesis or transposon silencing in adult ovaries.

DISCUSSION

DNA damage control of nuage organization

Forward genetic screens have defined the Drosophila piRNA biogenesis machinery and identified the targets for these small silencing RNAs. However, piRNA mutations activate transposons, leading to genome instability and DNA damage signaling, which has the potential to mask primary phenotypes. By genetically suppressing DNA damage signaling through Chk2, the genetic, cytological, and molecular studies reported here provide new insight into the primary functions of piRNA pathway genes and the role of subcellular localization in piRNA biogenesis. We show that DNA damage signaling displaces the ping-pong piRNA biogenesis machinery from nuage granules but does not block nuage assembly. Surprisingly, these studies also indicate that nuage localization of central components of the ping-pong and phased biogenesis machinery is not required for piRNA production or transposon silencing. Table 1 summarizes the phenotypes we observed in this study.

TABLE 1.

Summary of piRNA single mutant and mnk double mutant phenotypes

		Nuage localization of				piRNA			Long RNA
	Genotype	Vasa	Aub	Ago3	Armi	antisense TE	Z-score	Phasing	TE
RDC	rhi KG/02	−	O	−	−	− −	− −	WT	++
	mnk, rhi KG/02	−	O	−	−	− −	− −	WT	++
TREX export machinery	uap56 sz16/28	−	−	−	O	−	−	WT	++
	mnk, uap56 sz15/28	WT	WT	WT	O	−	−	WT	+
	thoc7 d/Df	−	O	WT	−	−	WT	WT	+
	mnk; thoc7 d/Df	WT	O		−	−	WT	WT	+
Ping-pong machinery	vas RG/PH	NA	O	WT	WT	−	−	WT	++
	mnk, vas RG/PH	NA	O	WT	WT	−	−	WT	++
	aub HN/QC	WT	NA	WT	WT	− −	− −	WT	+
	mnk, aub HN/QC	WT	NA	WT	WT	− −	− −	WT	+
	ago3 t2/t3	WT	O	NA	−	− −	− −	WT	++
	mnk; ago3 t2/t3	WT	−	NA	−	− −	− −	WT	++
Transcriptional silencing	armi 1/72.1	WT		WT	NA	− −	− −	WT	++
	mnk; armi 1/72.1	WT		WT	NA	− −	− −	WT	++
	mnk, piwi 02/ΔNLS	Ovaries were unrecoverable

Perinuclear nuage localization of Vasa, Aub, Ago3, and Armi proteins was evaluated in various piRNA mutants and designated as more localized than wild type (+), wild-type-like localization (WT), less localized than wild type (−), not localized at all (O), and not applicable (NA). Long and small RNA sequencing were performed on various piRNA mutants, and changes in global levels relative to w¹ controls were evaluated. Up relative to controls is designated by +, down relative to controls is designated by −, and comparable levels are designated by WT.

Nuclear control of perinuclear nuage composition

Drosophila germline piRNA biogenesis is initiated by RDC-mediated cluster transcription, which produces precursors that are exported from the nucleus (Klattenhoff et al. 2009; Mohn et al. 2014; Zhang et al. 2014). These precursors undergo ping-pong piRNA amplification mediated by Aub, Ago3, and Vasa, which localize to perinuclear nuage, as well as phased piRNA processing that involved Armi, found both in nuage and mitochondria, and the mitochondrial nuclease Zuc (Brennecke et al. 2007; Lim and Kai 2007; Han et al. 2015; Mohn et al. 2015). Rhino anchors the RDC at clusters, and rhi mutants lead to collapse of germline piRNA production and loss of Aub, Ago3, Vas, and Armi from nuage. These defects are not suppressed by a null allele of mnk, which encodes the DNA damage–activated kinase Chk2 (Supplemental Figs. S3, S5A). Transcripts produced through Rhino-driven transcription thus appear to recruit the cytoplasmic biogenesis machinery to nuage. Mutations in uap56 and thoc7, which encode TREX complex components, disrupt cluster transcript export from the nucleus (Hur et al. 2016; Zhang et al. 2018). However, these mutations produce less severe defects in piRNA production than rhi mutants (compare Fig. 4A and Supplemental Fig. S5A), and both displace Vasa from nuage, but localization is restored in mnk double mutants (Fig. 2; Supplemental Fig. S3). In contrast, uap56 and thoc7 mutations displace Armi from nuage, and localization is not restored in mnk double mutants. These findings imply that piRNA precursors are exported by TREX-dependent and TREX-independent mechanisms and that the TREX-dependent pathway is required for Armi localization to nuage (Fig. 7).

FIGURE 7.

Model of Chk2 modulating nuage composition and interdependency between piRNA pathway proteins and cluster transcripts. Rhino bound to dual-stranded piRNA clusters initiates transcription of long piRNA precursors from both genomic strands. We propose that these transcripts can be exported out of the nucleus in a TREX-dependent and TREX-independent manner. piRNA precursors in the cytoplasm can facilitate Ago3 and Vasa localization to nuage granules, and their localization is independent of each other. However, Aub localization to nuage is dependent on Ago3 and Vasa. Lastly, Armi localization to nuage requires intact TREX complex and Ago3. Upon genomic instability, activation of DNA damage signaling though Chk2 affects Vasa and Aub localization to nuage. We propose that Chk2 signaling may target Vasa, and Aub displacement may be a downstream effect loss of Vasa at nuage.

Mutations in armi disrupt phased piRNA biogenesis, and Armi protein localizes to nuage and colocalizes with the mitochondrial nuclease Zuc (Han et al. 2015; Mohn et al. 2015; Rogers et al. 2017; Ge et al. 2019). These observations suggest that Armi binds precursors in nuage, where they are generated by Ago3 cleavage, and shuttles these transcripts to mitochondria, where they are processed by Zuc (Rogers et al. 2017; Ge et al. 2019). However, phased piRNA biogenesis is not disrupted in uap56 or thoc7 mutants, or the corresponding mnk double mutants, which disrupt Armi accumulation at nuage (Fig. 4C; Supplemental Fig. S5C). Most proteins shuttle between subcellular organelles and the cytoplasm, and cytological “localization” reveals the rate-limiting step of a dynamic cycle. In wild-type ovaries, exit from nuage could be the slow step in a cycle in which Armi shuttles between nuage and mitochondria. In contrast, Armi association with nuage could be rate limiting in rhi and TREX mutations, leading to reduced accumulation in perinuclear granules. However, our findings indicate that steady-state accumulation of Armi in nuage is not required for phased biogenesis.

TREX function in spatial juxtaposition of nuage and clusters

In wild-type ovaries, prominent nuage granules localize to regions of the nuclear periphery opposite piRNA clusters, and uap56 and thoc7 mutations disrupt localization of nuage components, leading us to propose that TREX-mediated export of cluster transcripts promotes nuage assembly (Zhang et al. 2012). However, nuage localization of Aub, Ago3, and Vasa is restored in uap56, mnk double mutants, indicating that loss of localization is secondary to Chk2 activation. In the mnk, uap56^sz15/28 double mutants, nuage localization to regions opposite clusters is significantly reduced, but not eliminated. TREX-dependent precursor export from the nucleus thus contributes to localization of perinuclear nuage but may be partially redundant with TREX-independent export.

Nuage assembly and function

Drosophila Vasa was the first molecularly characterized nuage component, and two hypomorphic point mutations that disrupt localization were also reported to block nuage formation as assayed by thin section electron microscopy (Liang et al. 1994). However, we observe electron-dense nuage in ovaries homozygous for a chromosomal deletion that completely removes the vas locus (vas^PH) and in an allelic combination that blocks Vasa and Aub localization in adult ovaries (vas^RG/PH; Fig. 3; Supplemental Fig. S3C,D). In addition, Armi and Ago3 accumulate in perinuclear granules in vas^RG/PH mutant ovaries, and the level of accumulation is comparable to controls (Supplemental Fig. S5). Assembly of perinuclear nuage thus does not require Vasa, but does not explain the discrepancy with previously published work (Liang et al. 1994). Electron-dense perinuclear nuage is also present in aub, ago3, armi, rhi, uap56, and thoc7 mutants (Fig. 3; Supplemental Figs. S3, S4), indicating that nuage assembly is also independent of piRNA precursor production, export, and processing. Several proteins with unstructured and Tudor domains localize to nuage (Suyama and Kai 2025), and redundant interactions mediated by these proteins could drive nuage assembly, with no single component playing an essential role.

While proteins required for nuage assembly remain to be identified, our analysis of piRNA mutations and mnk double defines a genetic pathway controlling nuage composition (Fig. 7). In this pathway, Rhino-dependent clusters are transcribed and produce piRNA precursors. Once precursors are exported into the cytoplasm, they can recruit Armi, Aub, Ago3, and Vas to nuage. Precursor export appears to be comprised of TREX-dependent and TREX-independent processes, with TREX-dependent export essential to Armi localization (Fig. 7). Nuage composition is also regulated in response to genome instability, which triggers Chk2-dependent loss of Vasa and Aub localization to nuage.

Nuage composition thus appears to be carefully controlled, but we observe remarkably subtle changes in piRNA expression and transposon silencing when nuage localization of the biogenesis machinery is disrupted. Armi promotes phased piRNA processing and localizes to nuage and mitochondria, but uap56 and thoc7 displace Armi from nuage and do not disrupt phased biogenesis. In addition, Vasa and Aub are required for ping-pong piRNA amplification and localize to nuage. However, in vas^PD mutant ovaries, Aub fails to localize to nuage, and Vasa protein is not detectable, yet ping-pong amplification and antisense piRNA production remain comparable to wild type (Fig. 6C). Phased biogenesis and ping-pong amplification thus do not require Armi, Aub, or Vasa accumulation to nuage.

In striking contrast to the vas^PD allele, null vas mutations produce profound defects in ping-pong amplification, piRNA expression, and transposon silencing (Malone et al. 2009). Biochemical studies indicate that Vasa promotes release of cleaved products from PIWI proteins during the ping-pong cycle (Xiol et al. 2014). Previous studies suggest that the vas^PD produces low levels of Vasa protein during very early stages of oogenesis (Lasko and Ashburner 1990; Johnstone and Lasko 2004). Low levels of wild-type protein could be sufficient to catalytically drive the ping-pong cycle. However, we could not detect protein in adult vas^PD ovaries but speculate that this allele supports expression earlier in development. This could initiate the ping-pong amplification cycle, which is then propagated by reciprocal Aub and Ago3 cleavage.

Localization of the piRNA biogenesis machinery to perinuclear granules is conserved from nematodes to placental mammals, consistent with positive selection driven by enhanced reproductive fitness. However, our data indicate that piRNA production in Drosophila females does not require concentration of key biogenesis factors in nuage. We therefore speculate that nuage localization regulates piRNA pathway activity. For example, nuage could sequester transposon silencing activity. Within this framework, Chk2 activation in response to transposon mobilization and genome instability would enhance silencing by increasing the active cytoplasmic pool of Aub, which directly mediates posttranscriptional silencing.

MATERIALS AND METHODS

Fly strains and husbandry

All fly crosses were maintained at 25°C on cornmeal medium except for crosses generating thoc7 and mnk; thoc7 mutants, which were raised between 18°C and 20°C. mnk; thoc7^d/Df double mutants were not recoverable with crosses maintained at 25°C. Females were collected at 1–2 days old and fed on yeast for 2 days. Ovaries were collected from 2–4 day old females unless otherwise stated. Standard genetic procedures were used to generate double mutant combinations and to generate mutant trans-heterozygous.

Long RNA sequencing of vas^RG allele combined with vas deletion revealed a single adenosine deletion within a string of five consecutive adenosines in all transcripts mapping to the vas gene. The string of five adenosines corresponds to dm6/genomic location 2L:15073772–15073776 and is found in the seventh exon of vas gene, which is present in all three reference transcripts.

Immunofluorescence

Generally, 10–20 females were dissected for each staining condition. The protocol was previously described in McKim et al. (2009) and Zhang et al. (2012). Briefly, 2–4 day old female ovaries were dissected in Robb's media, fixed in 4% formaldehyde, washed, incubated in primary overnight, washed, incubated in fluorophore conjugated secondary and DAPI for DNA staining overnight, washed, and mounted on a slide with mounting medium. Visual inspection of ovaries was performed, and 10–25 z-sections were acquired from minimally six nuclei from distance egg chambers for image analysis. Representative images were chosen for the figures presented.

Image analysis

Automated quantification of perinuclear nuage localization

z-stacks were taken of nurse cell nuclei stained for two different nuage markers and nuclear pore. Images were analyzed in ImageJ, where signal was defined as 4 SD above the mean for each marker. Signals for the two nuage markers were merged and used to define nuage particles that were larger than 0.013–0.026 μm² in size. Only particles within 0.25–0.36 μm proximity to the nuclear membrane signal were considered perinuclear granules. Signal intensity of each nuage marker was then measured for every granule identified previously. See extended methods in Supplemental Material for annotated scripts used.

Automated quantification of nuclear and cytoplasmic spatial juxtaposition

To measure the percent of peripheral Rhino juxtaposed with the cytoplasmic nuage in an unbiased manner, we developed an automated image analysis pipeline using ImageJ. z-stacks were taken of nuclei stained for Rhino, nuclear membrane, and nuage marker. Signal was defined as 3 SD above the mean signal for each marker. The nuclear membrane signal was used to identify any Rhino and nuage proximal to the nuclear membrane. Total number of Rhino at the periphery was counted. To count how many peripheral Rhino was adjacent to nuage marker, we looked for overlap between nuage and Rhino signal. Normally, nuage and Rhino signals do not colocalize; therefore, we expanded the nuage marker signal equally in all directions and measured overlap with each exaptation. A range of —zero to six pixel expansions were tested. All expansions showed similar trends to each other and to the manual method of measuring spatial juxtaposition described in Zhang et al. (2012). Expansion 3 was chosen moving forward. See extended methods in Supplemental Material for annotated scripts used.

Electron microscopy

Ovaries were collected from 2–4 day old females and fixed overnight in fixative buffer containing 2% glutaraldehyde and 100 mM sodium cacodylate. Dehydration, embedding, sectioning, and staining were done at the UMass Chan Medical School EM core facility. Visual inspection of ovary sections was performed, and images were taken from sections of four to eight distinct nurse cells. Representative images were chosen for the figures presented.

Western blotting

Ovaries were collected from 2–4 day old females as previously specified and flash frozen in liquid nitrogen. Total protein lysate was prepared by adding lysis buffer (50 mM HEPES KOH, pH 7.5, 150 mM KCl, 0.5% NP-40, 3.2 mM MgCl₂, 5 mM DTT, 1 mM PMSF, and 1× protease inhibitor) to frozen tissue, homogenizing, centrifuging at 4°C for 15 min, and collecting supernatant. Total protein concentration was measured for each lysate using the Pierce BCA Protein Assay kit (23225) by following the manufacturer's protocol. Twenty-five milligrams of total protein was loaded into each well of an 8% acrylamide gel and subsequently transferred to nitrocellulose membrane. Blocking, primary, and secondary antibody incubations and washes were performed following the protocol from Li-Cor. Blots were imaged and quantified using the Odyssey DLx LI-COR system. Rat IgM anti-Vasa (1:500) and mouse anti-α-tubulin B512 (Sigma-Aldrich T5168) (1:2000) were used as primary antibodies, and in our hands, rabbit anti-Ago3 (MAb Technologies, Inc.) does not blot. The secondary antibodies used were Alexa Fluor 680 Conjugate anti rat IgM (Jackson ImmunoResearch 112-625-075, 1:100,000) and anti-mouse IgG 800CW (LI-COR 926-32210, 1:20,000). Four replicates were performed.

Ovary small RNA-seq

Small RNA-seq library construction was described previously (Zhang et al. 2021). Briefly, total RNA was extracted from ovaries collected from 10 to 20 females flies 2–4 days old using the mirVana kit (Ambion). Small RNAs 18–30 nt long were isolated using polyacrylamide gel purification. Sequencing libraries were made by first depleting the 2S rRNA followed by 3′ adaptor ligation, gel purification, 5′ adaptor ligation, gel purification, reverse transcription, and PCR amplification. Libraries were sequenced using Illumina NextSeq. Two or more biological replicates were generated for each genotype.

Ovary long RNA-seq

Strand-specific long RNA-seq library construction was described previously (Zhang et al. 2021). Briefly, total RNA was extracted from 10–20 female flies 2–4 days old using the mirVana kit (Ambion). rRNA was depleted using antisense rRNA oligo hybridization followed by RNase H digestion. Library construction consisted of RNA fragmentation, reverse transcription, dUTP incorporation, end repair, size selection using Ampure XP beads (∼100–500 nt), A-tailing, adaptor ligation, UDG treatment, and PCR amplification. Libraries were sequenced using the Illumina NextSeq. Two or more biological replicates were generated for each genotype.

Bioinformatics analysis

The bioinformatic analysis was performed as described previously (Zhang et al. 2021). The Drosophila reference genome (dm6), rRNA sequences, gene annotations, and hairpin sequences are from FlyBase (version 6.13). Transposon consensus sequences were from Repbase (Bao et al. 2015).

piRNA cluster annotation

We used a similar method as Yu et al. (2019) to annotate piRNA clusters in the dm6 genome using w¹ control small RNA-seq data from Zhang et al. (2021) (see Table 2). We considered 24–32 nt small RNA reads that could map to the dm6 genome, after rRNA, miRNA, tRNA, snRNA, and snoRNA were removed, as piRNAs. piRNAs were then assigned to 20 kb sliding windows (with a 1 kb step), and windows with more than 100 piRNAs per million uniquely mapped piRNAs were further considered as potential piRNA clusters. To remove false positives due to unannotated miRNA, rRNA, tRNA, snRNA, and snoRNA, which mostly produce reads with the same sequences, we also filtered out those 20 kb genomic windows with fewer than 200 distinct reads (called species). We then calculated the first-nucleotide content for each 20 kb window, and those windows with 1U/10A percentage <50% were also discarded. The remaining contiguous 20 kb windows were deemed putative piRNA clusters. Finally, we performed manual curation for putative piRNA clusters using piRNA profile. Determined by the direction of the piRNAs produced, piRNA clusters are classified as uni-strand and dual-strand piRNA clusters.

TABLE 2.

Key resources

Reagent or resource	Source	Identifier
Antibodies
Guinea pig anti-Rhino—(1:500)	Klattenhoff et al. 2009	n/a
Mouse antinuclear pore complex, MAb414 clone—(1:1000)	Covance MAb414 clone	MMS-120R
Rat IgM anti-Vasa—(1:250-IF, 1:500-western)	Developmental Studies Hybridoma bank	ID: AB_760351
Rabbit anti-Ago3—(1:3000)	MAb Technologies, Inc.	Product # Ago3-3
Rabbit anti-Aub—(1:500)	Klattenhoff et al. 2009	n/a
Rabbit anti-Armi-CT (1:500)	Cook et al. 2004	n/a
Mouse anti-α-tubulin B512 (1:2000—western)	Sigma	T6074
Antidigoxigenin POD	Roche	Cat# 11207733910
Chemicals, peptides, and recombinant proteins
SuperScript III	Thermo Fisher Scientific	Cat# 18080-085
RNase OUT	Thermo Fisher Scientific	Cat# 10777-019
TURBO DNase	Thermo Fisher Scientific	Cat# AM2238
RNase H	Thermo Fisher Scientific	Cat# 18021-071
T4 RNA Ligase	Thermo Fisher Scientific	Cat# AM2141
dNTP Set (100 mM)	Thermo Fisher Scientific	Cat# 10297018
dUTP solution (100 mM)	Thermo Fisher Scientific	Cat# R0133
dNTP mix	NEB	Cat# N0447L
DNA polymerase I	NEB	Cat# M0209S
T4 DNA polymerase	NEB	Cat# M0203L
Klenow DNA polymerase	NEB	Cat# M0210S
T4 PNK	NEB	Cat# M0201L
Klenow 3′ to 5′ exo	NEB	Cat# M0212L
UDG	NEB	Cat# M0280S
Phusion Polymerase	NEB	Cat# M0530S
T4 RNA Ligase 2, truncated	NEB	Cat# M0242L
50% PEG8000	NEB	Cat# B1004S
T4 DNA ligase	Enzymatics Inc.	Cat# L6030-HC-L
Hybridase Thermostable RNase H	Biosearch Technologies	Cat# H39500
16% Formaldehyde	Ted Pella Inc.	Cat# 18505
8% Glutaraldehyde solution	Millipore Sigma	Cat# G7526-10ML
Critical commercial assays
mirVana miRNA isolation kit	Thermo Fisher Scientific	Cat# AM1560
Direct-zol RNA MicroPrep	Zymo Research	Cat# R2060
Ribo-Zero Magnetic Kit (Human/Mouse/Rat)	Illumina	Cat# MRZG12324
RNA Clean & Concentrator-5	Zymo Research	Cat# R1015
Agencourt AMPure XP	Beckman Coulter	Cat# A63880
TSA Cyanine 3 System	Akoya Biosciences	Cat# NEL704A001KT
Deposited data
High-throughput sequencing	This study	GEO: GSE236035
Small RNA-seq for uap56sz15/28 Rep 1	Zhang et al. 2021	SRA: SRR10541139
Small RNA-seq for uap56sz15/28 Rep 2	Zhang et al. 2021	SRA: SRR10541140
Small RNA-seq for w1 rep1 for cluster definition	Zhang et al. 2018	SRA: SRR7408119
Small RNA-seq for w1 rep2 for cluster definition	Zhang et al. 2021	SRA: SRR10541189
Experimental models: organisms/strains
D. melanogaster/w 1	William Theurkauf laboratory	n/a
D. melanogaster/Oregon-R	William Theurkauf laboratory	n/a
D. melanogaster/mnk p6	Brodsky et al. 2004	n/a
D. melanogaster/rhi2(rhi02086)	Klattenhoff et al. 2009	Bloomington Drosophila Stock Center—stock # 12226
D. melanogaster/rhiKG (rhiKG00910)	Klattenhoff et al. 2009	Bloomington Drosophila Stock Center—stock # 13161
D. melanogaster/uap56 28	Zhang et al. 2012	n/a
D. melanogaster/uap56 sz15	Zhang et al. 2012	n/a
D. melanogaster/thoc7d (thoc7d05792)	Zhang et al. 2018	Harvard Exelixis Stock Collection—stock # d05792
D. melanogaster/thoc7Df (DF(3L)BSC128)	Zhang et al. 2018	Bloomington Drosophila Stock Center—stock # 9293
D. melanogaster/vas RG53	Schupbach and Wieschaus 1991	n/a
D. melanogaster/vasPH (vasPH165)	Styhler et al. 1998	n/a
D. melanogaster/vasPD (vas1)	Styhler et al. 1998	n/a
D. melanogaster/aub HN	Schupbach and Wieschaus 1991	n/a
D. melanogaster/aub QC	Schupbach and Wieschaus 1991	n/a
D. melanogaster/ago3 t2	Li et al. 2009	Bloomington Drosophila Stock Center—stock #28269
D. melanogaster/ago3 t3	Li et al. 2009	Bloomington Drosophila Stock Center—stock #28270
D. melanogaster/armi 1	Cook et al. 2004	n/a
D. melanogaster/armi 72.1	Cook et al. 2004	n/a
D. melanogaster/piwi 02	Lin and Spradling 1997	Bloomington Drosophila Stock Center—stock #43319
D. melanogaster/piwi NLS	Klenov et al. 2011	n/a
D. melanogaster/w1 ; ; nos:GAL4, UASp:GFP:aub nanos>Gal4, P[UAS-GFP-aub]	Harris and MacDonald 2001
Oligonucleotides
Random primers	Thermo Fisher Scientific	Cat# 48190011
Software and algorithms
RStudio		https://www.rstudio.com/
ImageJ		https://imagej.nih.gov/ij/
Inkscape		https://inkscape.org
Adobe Creative Suite 6		Adobe Systems Inc.
BEDTools	Langmead et al. 2009
piSet		https://github.com/tianxiongbb/piSet

RNA-seq analysis

piSet_rnaseq pipeline from GitHub was used to analyze long RNA sequence reads. Briefly, Bowtie2 (version 2.2.5) with default settings was used to map raw sequencing reads to rRNA sequence (Langmead and Salzberg 2012). The remaining reads were mapped to the Drosophila genome (dm6) and transposon consensus sequences using STAR (version 2.5.2b) and HISAT2 with default parameters (Dobin et al. 2013; Kim et al. 2015). The transcript abundance for each gene, transposon, and cluster (reads per kilobase per million mapped reads [RPKM]) was counted using BEDTools (version 2.27.1) (Quinlan and Hall 2010) and normalized to total number of genome mapping reads after excluding rRNA mapping reads.

Small RNA-seq analysis

piSet_srnaseq was used to analyze small RNA sequencing reads. Briefly, the 3′ end adaptor was removed via cutadapt (version 1.15) (Martin 2011), and the raw small RNA-seq reads were mapped to rRNA, miRNA hairpin, snoRNA, snRNA, and tRNA sequence allowing for no mismatches using Bowtie (version 1.1.0) (Langmead et al. 2009). The remaining reads were mapped to the Drosophila genome (dm6) and transposon consensus sequences. Small RNA abundance was normalized to reads mapping to Flamenco. For ping-pong analysis, the 5′ go 5′ overlaps between piRNAs mapping to opposite genomic strands were calculated, and Z-score for 10 nt overlap was calculated by using 1–9 and 11–30 nt overlaps as background (Li et al. 2009). For phasing analysis, the shortest distance between the 3′ end of piRNAs and the 5′ end of the piRNA downstream was measured. Z-score for 0 nt between piRNAs was calculated using 1–19 nt as background (Han et al. 2015).

SUPPLEMENTAL MATERIAL

Supplemental material is available for this article.

MEET THE FIRST AUTHORS

Samantha Ho.

Nicholas P. Rice.

Meet the First Author(s) is an editorial feature within RNA, in which the first author(s) of research-based papers in each issue have the opportunity to introduce themselves and their work to readers of RNA and the RNA research community. Samantha Ho and Nicholas P. Rice are co-first authors of this paper, “Aub, Vasa, and Armi concentration in phase-separated nuage is dispensable for piRNA biogenesis and transposon silencing.” Samantha Ho and Nicholas P. Rice completed their graduate studies in Dr. William Theurkauf's laboratory at UMass Chan Medical School, where their research focused on understanding how stress influences piRNA biogenesis and transposon silencing. After earning their PhDs, both continued pursuing research in RNA biology—Samantha as a postdoctoral researcher in Dr. Wenwen Fang's lab studying double-stranded RNA sensing, and Nicholas as a scientist at ADViRNA exploring siRNA biology and therapeutic applications.

What are the major results described in your paper, and how do they impact this branch of the field?

piRNAs guide transposon silencing, and much of the piRNA biogenesis machinery localizes to phase-separated perinuclear granules known as “nuage.” Although it is believed that nuage plays a role in piRNA biogenesis, the assembly and functional significance of nuage remain poorly understood. To address this gap, we systematically analyzed nuage ultrastructure (by EM), pathway organization (by IF), transposon silencing, and piRNA expression in Drosophila mutants that disrupt nuclear, perinuclear, or mitochondrial components of the piRNA pathway. Notably, mutations that impair the piRNA pathway induce genome instability and activate checkpoint kinase 2 (Chk2)–mediated DNA damage signaling, potentially masking primary phenotypes. This led us to investigate whether Chk2 regulates nuage organization. Indeed, we found that Chk2 activation displaces key piRNA proteins from nuage. Importantly, our findings also demonstrate that the accumulation of essential piRNA biogenesis factors in nuage is not required for piRNA production or transposon silencing, challenging prevailing view of nuage function.

What led you to study RNA or this aspect of RNA science?

SH: When I started graduate school at UMass, I was struck by how passionate the small RNA community was and how they were constantly pushing the field forward. Their excitement was contagious and made me want to be part of their discoveries. Ever since then, I've been amazed not only by how much we've learned about RNA, but also by how much remains unknown.

NPR: Before UMASS, I had done research on microRNAs and knew only little about piRNAs. Bill was a big reason I ended up staying in the RNA field. His ideas and enthusiasm encouraged me to pursue questions that I found interesting and opened my eyes to the wide field of RNA.

What are some of the landmark moments that provoked your interest in science or your development as a scientist?

NPR: My undergraduate lab mentors from the Wangh lab at Brandeis showed me that science was not just about answering questions; it is also about understanding what we didn't know, which can lead to very fun and new directions in your research.

If you were able to give one piece of advice to your younger self, what would that be?

SH: Don't be afraid to ask questions.

Are there specific individuals or groups who have influenced your philosophy or approach to science?

SH: I truly feel that every member of the Theurkauf lab and the broader RNA community at UMass has shaped this project and the way I think about science. Some of the most important ideas came from casual hallway conversations, lab meetings, and discussions in the piRNA group and RNA journal club. Those interactions made this work possible and have been a big part of my growth as a scientist.

NPR: Both Larry Wangh and Bill Theurkauf showed me that questioning our assumptions and independently testing an idea or model often leads to a new direction and that pursuing those new directions is a great way to learn fundamentally new things.

How did you decide to work together as co-first authors?

SH: When Nick and I joined Bill's lab as graduate students, we each started on our own projects, connected only by a shared interest in the piRNA pathway. But as we regularly talked through our ideas and results, we began to see that our data were leading us toward the same underlying question. It was exciting to watch our separate projects naturally grow into a shared story.

NPR: As Sam wrote, our projects began independently, but as we generated data, the experiments and planning became increasingly interconnected. Long before we began writing the manuscript together, our work had naturally evolved into a joint effort.