Introduction

Oogenesis is a complex process in which germline stem cells (GSCs) develop into mature female gametes or oocytes. This process involves multiple layers of regulations and numerous genes, some of which remain to be discovered and characterized. Drosophila melanogaster, a genetically tractable model organism, has long been at the forefront of oogenesis research.

Adult Drosophila females possess a pair of ovaries, each composed of 12-16 ovarioles. At the anterior tip of each ovariole is a structure called the germarium, which houses two to three GSCs (Fig S1). GSCs reside in the most anterior region of the germarium, where they directly contact cap cells and escort cells, forming a niche crucial for GSC self-renewal ^1–3. GSCs undergo asymmetric mitotic division, producing two distinct cells: one GSC and one cystoblast^1,4. The cystoblast, destined to differentiate, undergoes four mitotic divisions with incomplete cytokinesis, resulting in interconnected cysts of 2, 4, 8, and 16 cells ^5,6. Of the 16 cells, one will differentiate into an oocyte, while the remaining 15 develop into polyploid nurse cells, which synthesize proteins and RNAs for deposition into the oocyte⁷.

GSC proliferation is essential for proper oogenesis. GSCs must be tightly regulated to maintain their undifferentiated state while continuing to divide and differentiate. Failure in GSC self-renewal or inappropriate differentiation leads to stem cell loss, disrupting oocyte production^8,9. Conversely, uncontrolled GSC division or defective differentiation results in an overabundance of GSC-like cells (a tumorous phenotype) and fewer differentiated cells, ultimately inhibiting oocyte production^10,11. Thus, a precise balance between GSC self-renewal and cyst differentiation within the germarium is essential for proper oogenesis and is tightly regulated.

Bone Morphogenetic Protein (BMP) signaling plays a critical role in regulating GSC self-renewal and differentiation during Drosophila oogenesis. In the stem cell niche, GSCs are physically anchored to cap cells via adherens and gap junctions^12,13. Cap cells secrete BMP ligands, such as Decapentaplegic (Dpp) and Glass bottom boat (Gbb), which are recognized by transducing receptors like Thickvein (Tkv) and Saxophone (Sax) on the GSCs^14,15 (Fig S1). Upon activation, these receptors phosphorylate a transcription factor Mad. The phosphorylated Mad (pMad) translocates into the nucleus, where it represses the transcription of bag-of-marbles (bam), a key differentiation factor ^16,17.

Bam, a ubiquitin-associated protein, is essential for cystoblast differentiation. Dpp/BMP signaling in the niche represses bam expression, preventing GSCs from differentiating and enabling self-renew. However, as the daughter cells (cytoblasts) exit the niche, they derepress bam and initiate the differentiation program (Fig S1)⁷. Tight regulation of Dpp/BMP signaling and bam expression ensures that only cells within the niche adopt GSC fate, while adjacent daughter cells differentiate into cystoblasts. Loss of bam leads to blocked differentiation and the accumulation of GSC-like cells, whereas ectopic bam expression forces GSCs differentiation, depleting the stem cell pool ^18–20.

One of the proteins that interacts with Bam is Ovarian Tumor (Otu). Otu forms a deubiquitinase complex with Bam, which deubiquitinates Cyclin A (CycA), stabilizing CycA and promoting GSC differentiation²¹. Otu is crucial for oogenesis and female fertility^22,23. Mutations in the otu gene cause various ovarian defects, including tumorous growths, germ cell loss, and abnormalities in oocyte determination and nurse cell dumping, indicating Otu’s importance in multiple stages of oogenesis^22–25. However, the precise molecular mechanisms by which Otu functions in these processes remain largely unknown.

In this study, we aimed to identify a novel gene required for oogenesis. Through a search for uncharacterized genes expressed exclusively in Drosophila ovaries, we discovered CG14545, a gene encoding a 114 amino acids (aa) long protein without any known functional domain (Fig 1A). CG14545 is expressed specifically in female germline cells, including GSCs. Our genetic analysis revealed that CG14545 is required for oogenesis and female fertility and plays an important intrinsic role in regulating GSCs renewal and differentiation. Additionally, CG14545 is important for the piRNA pathway. We identified Otu as a binding partner of the CG14545 protein, suggesting that both proteins may participate in the same molecular pathways to regulate GSC fate and differentiation. We named the CG14545 gene sakura, meaning “cherry blossom” in Japanese, symbolizing birth and renewal.

Fig 1.

Results

Sakura is exclusively expressed in the ovaries

Based on high-throughput expression data in Flybase, sakura mRNA is highly and exclusively expressed in adult ovaries (Fig 1B). To investigate Sakura protein expression, we generated a polyclonal anti-Sakura antibody against a recombinant full-length Sakura protein. Using this antibody, we examine Sakura protein expression at different stages of Drosophila development and across tissues. We found that Sakura protein is specifically expressed in the ovary, particularly in stage 1-11 oocytes (Fig 1C).

sakura mutant flies

To explore the biological and molecular functions of Sakura in vivo, we generated a sakura mutant allele (sakura^null) with a one nucleotide (nt) insertion replacing 33 nts within the Sakura coding region using the CRISPR/Cas9 genome editing system (Fig 1A)²⁶. This mutation causes a frameshift and a premature stop codon, resulting in a 33-amino acid (aa) N-terminal fragment of Sakura followed by a 22-aa segment produced by frameshifted translation (Fig 1A). This truncated fragment is unlikely to be functional, and we were unable to detect stable protein expression, as shown below. Thus, we consider this allele a null mutation. sakura^null/null flies were viable, revealing that Sakura is not essential for survival.

To validate the sakura^null strain, we performed Western blots using ovary lysates from sakura mutant flies using the anti-Sakura antibody. As expected, full-length Sakura protein was detected in the ovaries of wild-type (sakura^+/+) and heterozygous controls (sakura^null/+), but not in those of homozygous mutants (sakura^null/null) (Fig 1D). No smaller protein corresponding to the truncated Sakura fragment was detected in either sakura^null/+ or sakura^null/null ovaries thus, suggesting that the fragment is unstable or not expressed.

Sakura is cytoplasmic and expressed specifically in germ cells

To determine the localization of Sakura protein within the ovaries, we generated transgenic flies carrying a sakura coding sequence fused with EGFP at the C-terminus (sakura-EGFP) under the control of the sakura promoter. Western blot analysis of ovary lysates showed that Sakura-EGFP is expressed at levels comparable to endogenous Sakura (Fig 1D). Confocal imaging revealed that Sakura-EGFP is specifically expressed in the cytoplasm of germ cells, including GSCs, cyst cells, nurse cells, and developing oocytes (Fig 1E and 1F). Sakura-EGFP was not expressed in somatic cells such as terminal filament, cap cells, escort cells, or follicle cells. (Fig 1E). In the egg chamber, Sakura-EGFP was detected in the cytoplasm of nurse cells and was enriched in developing oocytes (Fig 1F).

sakura is essential for female fertility

Given its ovary-specific expression at both mRNA and protein levels (Fig 1B and 1C), we hypothesized that Sakura plays a critical role in ovarian function. Fertility assays revealed that sakura^null/null females laid no eggs when mated with wild-type males, while control female flies (sakura^+/+ and sakura^null/+) laid numerous eggs (Fig 2A). Dissection of sakura^null/nullovaries revealed that they were rudimentary compared to normal ovaries in controls (Fig 2C). This suggests that the inability of sakura^null/null female to lay eggs is due to their underdeveloped ovaries. In contrast, sakura^null/null males were fertile and showed no significant differences compared to controls (sakura^+/+ and sakura^null/+) (Fig S2).

Fig 2.

To confirm that the observed female sterility was caused by the loss of sakura, we generated sakura rescue flies expressing Sakura-EGFP in the sakura^null/null background (sakura^null/null; sakura-EGFP). Western blotting confirmed that these flies expressed Sakura-EGFP, but not endogenous Sakura (Fig 1D). The sakura-EGFP transgene fully rescued both female fertility and ovary morphology in the sakura^null/null background (Fig 2). sakura-EGFP rescue females showed no significant fertility differences compared to controls (Fig 2A and 2B), and their ovaries were normal in shape and size (Fig 2C). We concluded that Sakura is essential for female fertility and normal ovary morphology but dispensable for male fertility, consistent with the observation that Sakura is predominantly expressed in ovarian germline cells (Fig 1).

sakura^null/null ovaries are germless and tumorous

Given the rudimentary appearance of sakura^null/null ovaries, we questioned whether they contained germ cells. We used flies carrying vasa-EGFP knocked-in reporter in the sakura^null/null background, as Vasa is a known germ cell marker⁷. We observed that some sakura^null/null ovarioles were devoid of germ cells ("germless"), while others retained germ cells (Fig 3A), suggesting that loss of sakura leads to germ cell depletion. Additionally, sakura^null/null ovaries exhibited significantly higher levels of cleaved Caspase-3 staining, indicative of apoptosis, compared to controls (sakura^null/+) (Fig 3B).

Fig 3.

Oogenesis begins in the germarium, which contains 2-3 GSCs identified by round, unbranched spectrosomes that contact cap cells⁷ (Fig S1). In contrast, cysts have branched fusomes. Immunostaining with hu-li tai shao (HTS) antibody, which marks spectrosomes and fusomes, revealed an excess number of GSC-like cells with round spectrosomes in sakura^null/null ovarioles containing germ cells (Fig 3C, red stars), indicative of a "tumorous" phenotype as previously described^27–29. Additionally, we observed an excess number of cyst cells with branched fusomes that persisted throughout the ovarioles (Fig 3C), suggesting abnormal cyst cell differentiation and division.

Within the same ovary, some sakura^null/null ovarioles exhibited the tumorous phenotype, while others were germless, without any spectrosome or fusome staining (Fig 3C, green stars). Quantification revealed that 35% of sakura^null/null ovarioles were tumorous while the remaining 65% were germless (n=74) (Fig 3D). In contrast, no tumorous or germless phenotypes were observed in control (sakura^+/+ and sakura^null/+) or sakura-EGFP rescue flies (Fig 3D). Furthermore, ∼95% of sakura^null/null ovarioles with germ cells contained excess GSC-like cells (>5) (Fig 3E and Fig S3). The mean number of GSCs or GSC-like cells for sakura^+/+, sakura^null/+, and sakura-EGFP rescue was 2.2 ± 0.6, 2.2 ± 0.6, and 2.1 ± 0.7, respectively while that for sakura^null/null was 26.5 ± 17.5 (p-value <0.05). These results suggest that Sakura is essential for regulating GSC division and differentiation. Overall, we conclude that Sakura is required for germ cell survival, division, and differentiation.

Loss of sakura results in reduced germline piRNA

Piwi-interacting RNAs (piRNAs) are produced in germline cells and somatic follicle cells and transcriptionally and post-transcriptionally silence transposon expression through sequence-complementarity. Disruption of the piRNA pathway can result in oogenesis arrest, germ cell loss, rudimentary ovaries, and sterility^9,30. Loss of piRNAs in germ cells leads to the upregulation (desilencing) of transposons, increased DNA damage, and apoptotic cell death³¹. We investigated whether the loss of sakura would result in loss of piRNA and transposon upregulation. We performed high-throughput sequencing (small RNA-seq and polyA+ RNA-seq) on ovary RNA samples from sakura^+/+, sakura^null/+, sakura^null/null, and sakura-EGFP rescue (sakura^null/null with sakura-EGFP) flies.

piRNA levels in sakura^null/null ovaries were reduced compared to controls (sakura^+/+), while sakura^null/+ and sakura-EGFP rescue showed no such reduction (Fig 4A). Consistent with this, transposon RNA levels in sakura^null/null ovaries were upregulated compared to sakura^+/+, while sakura^null/+ and sakura-EGFP rescue showed no such upregulation (Fig 4B). Given that Sakura is specifically expressed in germline cells, but not in somatic cells (Fig 1E and 1F), we hypothesized that loss of sakura interrupts the piRNA pathway in germ cells, leading to the desilencing of germline transposons. A well-characterized germline transposon in Drosophila is Burdock^30,32. We found that Burdock piRNA levels were significantly lower in sakura^null/nullcompared to controls and rescue flies (Fig 4C).

Fig 4.

We employed the Burdock sensor, a transposon reporter tool to monitor germline piRNA activity³⁰. The reporter expresses nuclear GFP and β-gal under the control of the nanos promoter for germline expression, with a target sequence for Burdock piRNAs in the 3′ UTR³⁰ (Fig 4D). Using a sakura RNAi line driven by UAS-Dcr2 and NGT-Gal4 to knock down sakura specifically in the germline, we observed that GFP and β-gal expression were highly elevated in sakura knockdown germlines compared to control RNAi (y^RNAi) knockdowns (Fig 4D). This confirmed that loss of sakura leads to a loss of piRNA-mediated transposon silencing in the germline. We conclude that Sakura is essential for proper piRNA levels and piRNA-based transposon silencing in the germline.

sakura is important for oogenesis in germline cells beyond GSCs and germline cysts

Both sakura^null/null and sakura germline RNAi knockdown driven by UAS-Dcr-2 and NGT-Gal4, which initiates RNAi in germline from GSCs, resulted in rudimentary ovaries (Fig 2C and 5A). These ovaries lacked later-stage germline cells and egg chambers, making it difficult to assess the role of Sakura beyond the germarium. However, because Sakura-EGFP expression is not limited to GSCs and germline cysts (Fig 1E and 1F), we speculated that Sakura might function in later-stage germline cells as well.

Fig 5.

To explore this possibility, we used TOsk-Gal4 (a combination of osk-Gal4 and αTub67C-Gal4) to drive sakura RNAi knockdown in germline cells after the germline cyst stage (from germarium region 2b onward. Fig S1A), sparing GSCs and germline cysts³³. Unlike NGT-Gal4, TOsk-Gal4-driven sakura RNAi knockdown did not drastically affect ovary morphology (Fig 5A), allowing us to study the effects of sakura loss in egg chambers. We confirmed sakura RNAi (sakura RNAi #1 and #2) knockdown efficiency by Western blot, showing effective depletion of Sakura in ovaries from both NGT-Gal4 and TOsk-Gal4-driven RNAi lines (Fig 5B).

Interestingly, TOsk-Gal4-driven sakura RNAi knockdown severely reduced the numbers of eggs laid (Fig 5C) and stage 14 oocytes in ovaries compared to control RNAi (y^RNAi) (Fig 5D), suggesting that sakura is important for oogenesis beyond the germline cyst stage. Additionally, we observed mislocalized and dispersed Oo18 RNA-binding protein (Orb. A marker for oocyte identity) in TOsk-Gal4-driven sakura RNAi knockdowns (Fig 5E). Proper Orb localization is critical for oocyte specification and maintenance³⁴, and its mislocalization likely underlies the reduced ability to produce stage 14 oocytes and eggs sakura knockdown flies (Fig 5C and 5D). We conclude that Sakura is crucial for proper Orb localization and in germline cells beyond GSCs and germline cysts. In subsequent sakura RNAi experiments, we used the sakura RNAi #2 line.

Sakura is required intrinsically for GSC establishment, maintenance, and differentiation

To explore whether sakura functions autonomously in the germline, we conducted mosaic analysis using the FLP-FRT system³⁵. First, we assessed its role in GSC establishment by marking sakura^null clones in primordial germ cells (PGCs)) before the early third instar larval stage and tracking their development into adult GSCs, as described in a previous study³⁶. In control adult flies,12.3% (20/163) of GSCs were marked, while only 1.9% (4/213) were marked in sakura^nullmutants, demonstrating that sakura is autonomously important for GSC establishment.

Next, we assessed GSC maintenance by generating marked sakura^nullGSCs in adult flies using the FLP-FRT system and measuring their loss rate over time ^14,36. The percentage of marked control GSC clones and marked sakura^null GSC clones 4 days after clone induction was 27.8% and 22.1%, respectively (Fig 6A). These numbers were considered the initial percentage. The percentage of marked control GSCs decreased to 25.8% and 16.8% on days 7 and 14, respectively (Fig 6A), resulting in a loss rate of 39.4% over the 10-day testing period (from day 4 to day 14). In contrast, the percentage of marked sakura^null GSC clones decreased to 16.3% and 5.4% on days 7 and 14, respectively. Thus, the loss rate of sakura^null GSCs is 75.6% over the 10-day period, which is higher than controls. We conclude that sakura is intrinsically required for GSC maintenance.

Fig 6.

Germline depletion of sakura caused germless and tumorous phenotypes (Fig 3). To determine whether sakura is intrinsically required in the germ line cells for GSC differentiation, we quantified the number of marked (GFP-negative) and unmarked (GFP-positive) GSC-like cells (germline cells with a round spectrosome) in the germaria carrying marked GSC clones at 4, 7, and 14 days after clone induction. We found that the number of marked sakura^nullGSC-like cells was higher than that of marked control GSC-like cells at all time points (Fig 6C). The number of marked sakura^nullGSC-like cells increased over the 10-day period, while the number of marked control cells did not (Fig 6C). In contrast, the number of unmarked GSC-like cells in germaria containing sakura^null GSC clones did not differ significantly from those in germaria containing control GSC clones, and the number did not increase (Fig S4). These results indicate that sakura is intrinsically important in germline cells, including GSCs, to regulate their proper division and differentiation. Taken together, we conclude that sakura is required intrinsically for GSC establishment, maintenance, and differentiation.

Loss of sakura inhibits Dpp/BMP signaling

Dpp/BMP signaling is a key pathway regulating GSC self-renewal and differentiation (Fig S1)^7,37. We hypothesized that the germless and tumorous phenotypes observed in sakura loss-of-function ovaries could be due to Dpp/BMP signaling misregulation. To test this, we knocked down sakura in the germline driven by UAS-Dcr2 and NGT-Gal4 in the presence of bam-GFP reporter³⁸. In controls (y^RNAi), Bam-GFP expression was restricted to 8-cell cysts and disappeared in 16-cell cysts and onward (Fig 7A)³⁸. However, in sakura knockdowns, Bam-GFP expression persisted throughout the ovariole, with excess Bam-GFP positive cells (Fig 7A). This suggests that Bam expression is not repressed by Dpp/BMP signaling, leading to GSC loss in sakura knockdowns.

Fig 7.

To further confirm this finding, we used the FLP-FRT technique to induce sakura^null clones in the germaria and stained them with anti-Bam antibody²⁰. All sakura^null clones expressed Bam throughout the germarium, whereas in control clones, Bam was not expressed in the GSCs and was restricted to cyst cells (Fig 7B). In GSCs, pMad translocates into the nucleus to repress Bam expression ^16,17(Fig S1). We stained ovaries with anti-pMad antibodies and found that pMad intensity was significantly reduced in sakura^null GSC clones compared to control GSCs (Fig 7C and D). This indicates that the dysregulated Bam expression observed in sakura^nullwas due to reduced pMad levels in GSCs.

Previous studies have shown that ectopic expression of a stable form of CyA leads to germ cell loss³⁹. This germ cell loss phenotype is also observed upon ectopic bam expression in GSCs^14,40,41. It was reported that Bam associates with Ovarian tumor (Otu) to promote deubiquitination and stabilization of CycA²¹. Since we observed derepressed bam expression in sakura^null cells, we investigated whether CycA levels were affected. Staining ovaries with anti-CycA antibodies revealed higher CycA intensity in sakura^null clones compared to neighboring wild-type cells in the same germarium and control clones in control germarium (Fig 7E). sakura^null clones exhibited significantly higher mean CycA intensity compared to control clones (Fig 7F). The elevated CycA levels in sakura^nullclones likely results from Bam misexpression and Bam-mediated stabilization.

Finally, we tested whether we could rescue the germless phenotype caused by sakura loss-of-function by knocking down either bam or cycA in the germline. Double RNAi knockdown of sakura and bam in the germline using UAS-Dcr-2 and NGT-Gal4 partially rescued the germless phenotype (Fig 8). In contrast, knocking down sakura and cycA together failed to rescue the germless phenotype (Fig 8). This suggests that the germless phenotype of sakura is partly due to bam misexpression, not cycA.

Fig 8.

Sakura binds Otu

To explore the molecular function of Sakura in oogenesis, we aimed to identify Sakura-interacting proteins. We conducted a co-immunoprecipitation experiment using anti-GFP magnetic beads on the ovary lysates of flies expressing Sakura-EGFP, followed by mass spectrometry to identify co-immunoprecipitated proteins. Ovary lysates from flies without the sakura-EGFP transgene (w1118) served as negative controls. Mass spectrometric analysis revealed several proteins with peptide signals present in all three biological replicates of the Sakura-EGFP samples without any signals in any of the three biological replicates of the negative control (Table 1). Among these, the most promising candidate, exhibiting the highest unique peptide signal in each of three replicates of the Sakura-EGFP samples was Otu.

Table 1.

To confirm this finding, we generated a polyclonal anti-Otu antibody. Using Western blot analysis, we detected endogenous Otu protein in the Sakura-EGFP immunoprecipitants from ovaries, but not in the negative control (Fig 9A), validating the mass spectrometry results. Next, using wild-type (without Sakura-EGFP) ovary lysates, we immunoprecipitated endogenous Sakura with anti-Sakura antibody and detected endogenous Otu in the immunoprecipitant by Western blot (Fig 9B), demonstrating that endogenous Sakura interacts with endogenous Otu. Furthermore, we found that ectopically expressed Flag-tagged Otu co-immunoprecipitated with ectopically expressed HA-tagged Sakura in S2 cells (Fig 9C).

Fig 9.

The otu gene encodes two annotated protein isoforms, a predominant 98 kDa isoform and a less abundant 104 kDa isoform resulting from alternative splicing⁴². The 98kDa isoform lacks Tudor domain encoded in an alternatively spliced 126-bp exon⁴³. It has been reported that the 104 kDa Otu isoform can perform all functions, while the 98 kDa Otu isoform functions in later stages of oogenesis, such as nurse cell regression and oocyte maturation⁴⁴. The major Otu isoform observed in the inputs and immunoprecipitates from ovaries in Figure 9 corresponds to the 98 kDa isoform.

Next, we sought to identify which regions of Sakura and Otu are important for their interaction. We predicted the structure of the Sakura-Otu protein complex using AlphaFold⁴⁵. AlphaFold suggests that the N-terminal region of Sakura (1-49aa) and the N-terminal region of Otu (1-405aa) are highly likely to form interactions, while the C-terminal region of Sakura (100-144aa) does not interact with Otu (Fig 10A-C). The Tudor domain of Otu is not directly involved in this interaction. To experimentally validate the predictions made by AlphaFold, we generated a series of truncated Otu fragments, including Otu(ΔTudor) (1-336, 389-853aa, corresponding to the endogenous 98 kDa isoform), Otu(N1) (1-405aa), Otu(N2) (1-339aa), and Otu(C) (405-853aa) (Fig 10B). We then performed co-immunoprecipitation assays to test which fragments could interact with full-length Sakura. All fragments except Otu(C) were associated with full-length Sakura protein in S2 cells (Fig 10D), demonstrating that the C-terminal low-complexity region of Otu is dispensable for the interaction. This assay also showed that the Tudor domain is not critical and the N-terminal fragment (1-339aa) of Otu is sufficient for interaction with Sakura. These findings align well with the predictions made by AlphaFold.

Fig 10.

Subsequently, we created several truncated Sakura fragments and conducted co-immunoprecipitation assays in S2 cells to determine which Sakura fragments can interact with full-length Otu. The fragments generated included Sakura(NM) (1-100aa), Sakura(NC) (1-49, 100-144aa), Sakura(MC) (49-144aa), Sakura(N) (1-49aa), Sakura(M) (49-100aa), and Sakura(C) (100-144aa) (Fig 10C). We found that Sakura(NM), Sakura(NC), Sakura(N), and Sakura(M) were associated with Otu (Fig 10E). Compared to Sakura(NM) and Sakura(N), the interaction of Sakura(M) with Otu is relatively weaker, as indicated by a fainter band in the Western blot (Fig 10E). Sakura-NC exhibited a weak interaction with Otu, while Sakura(MC) and Sakura(C) showed no interaction (Fig 10E). These assays demonstrated that the N-terminal region (1-49aa) of Sakura is sufficient for interaction with Otu. Moreover, the results suggest that the C-terminal region of Sakura (100-144aa) does not interact with Otu and that adjoining it with the N-terminal (1-49aa) or the middle (M) region (49-100aa) weakens the interaction with Otu. These findings support the AlphaFold predictions.

Finally, to test whether the interaction between Sakura and Otu can be achieved solely through their N-terminal regions, we performed co-immunoprecipitation assay with Sakura(N) and Otu(N2). We discovered that they interact each other (Fig 10F). A reciprocal co-immunoprecipitation assay also confirmed their interaction (Fig S5), revealing that the interaction between Sakura and Otu can be established with just the 1-49aa of Sakura and the 1-339aa of Otu. Therefore, we conclude that Sakura interacts with Otu in vivo, and the N-terminal regions of both proteins are sufficient for their interaction.

Loss of otu phenocopies loss of sakura

Having established the interaction between Sakura and Otu, we were interested in whether they function together during oogenesis. We speculated that for them to interact and function together, Sakura and Otu must exhibit similar expression and localization patterns in ovaries. We generated two transgenic flies carrying full-length otu-EGFP or otu(ΔTudor)-EGFP transgenes; both constructs are C-terminally fused with EGFP and under the control of the otu promoter. We found that both Otu-EGFP and Otu(ΔTudor)-EGFP display similar expression and localization patterns to Sakura-EGFP (Fig 1E, 1F and Fig 11A). Like Sakura-EGFP, Otu-EGFP and Otu(ΔTudor)-EGFP are specifically expressed in germ cells, but not in follicle cells, and are localized to the cytoplasm, being enriched in the developing oocytes (Fig 11A). We did not observe any difference in localization patterns between Otu-EGFP and Otu(ΔTudor)-EGFP; suggesting that the Tudor domain is not crucial for Otu protein localization. Notably, the localization pattern exhibited by the transgenic Otu-EGFP generated in this study is consistent with a previous report ²³.

Fig 11.

Previous studies have shown that mutations in otu lead to severe defects in germ cell division and differentiation, resulting in a tumorous phenotype^22–24,42. To study the role of otu specifically in the germline, we performed RNAi knockdown of otu in the germline driven by nos-Gal4-VP16. Interestingly, we found that germline depletion of otu results in rudimentary ovaries, similar to the loss of sakura (Fig 2D, Fig 5A, and Fig 11B). Furthermore, similar to sakura RNAi (Fig 5), TOsk-Gal4-driven otu RNAi knockdown did not affect ovary morphology (Fig 11B), but it did result in a reduced number of eggs laid and stage 14 oocytes compared with control RNAi (w^RNAi) (Figure 11C and 11D).

When sakura and otu were independently RNAi-knockdowned in the germline driven by UAS-Dcr2 and NGT-Gal4, the percentages of germless ovaries for otu-RNAi and sakura-RNAi were similar (∼70%) (Fig 8). However, when otu and sakura were knocked down together, the percentage of germless ovaries increased significantly to 79% (Fig 8), suggesting that the double knockdown of otu and sakura enhances the germless phenotype in the ovaries.

Next, we used the Burdock sensor to test if loss of otu leads to loss of piRNA-mediated transposon silencing in the germline. Similar to sakura (Fig 4D), germline depletion of otu resulted in elevated GFP and β-Gal produced by the Burdock sensor (Fig 11E), indicating loss of piRNA-mediated transposon silencing. To determine if germ cell loss caused by otu-RNAi is a result of apoptosis, we stained the rudimentary ovaries of otu-RNAi for cleaved Caspase-3. Similar to sakura^null/null (Fig 3B), cleaved Caspase-3 signals were higher in otu-RNAi ovaries than in controls, suggesting that the observed germ cell loss is due to apoptosis (Fig 11F).

Loss of sakura leads to misexpression of bam, likely due to low levels of pMad in the germ cells (Fig 7A-C). We were curious whether a similar occurrence could be observed in otu-RNAi. In the otu germline RNAi knockdown ovaries, we found that Bam-GFP expression is not restricted to the germarium but persists throughout the ovariole, resulting in an excess number of Bam-GFP-positive cells (Fig 12A). We also observed lower pMad levels in otu^14/14 mutant GSCs (Fig 12B). These data suggest that, similar to sakura, loss of otu results in bam derepression, likely due to low levels of pMad in the germ cells.

Fig 12.

A previous study has shown that Otu possesses deubiquitinase activity²¹. We asked whether Sakura regulates the Otu’s deubiquitinase activity. To test this, we performed an in vitro deubiquitination assay using Ub-Rhodamine 110 as a model substrate. Consistent with previous studies²¹, we detected a deubiquitinase activity in Otu that was ectopically expressed and purified from S2 cells (Fig S6). The presence of Sakura, purified from E.coli, did not affect Otu’s deubiquitinase activity under these assay conditions. Additionally, Sakura itself did not exhibit deubiquitinase activity.

Taken together, these findings show that otu and sakura phenocopy each other in ovaries, specifically in regulating germ cell maintenance and differentiation, raising the possibility that they function together. Similar to sakura, loss of otu inhibits Dpp/BMP signaling and results in the de-repression of bam, leading to aberrant germline differentiation.

Discussion

In this study, we identified Sakura, encoded by a previously uncharacterized gene CG14545, as an essential factor for oogenesis and female fertility. We demonstrated that Sakura is specifically expressed in germline cells in the ovary, including GSCs, is localized to the cytoplasm, and is enriched in the developing oocytes. sakura homozygous null mutant flies are viable but completely female-sterile and male-fertile. Loss of sakura, either through null mutation or germline RNAi including in GSCs, results in rudimentary ovaries exhibiting germless and tumorous phenotypes.

The tumorous phenotype associated with the loss of sakura is characterized by an excess of GSC-like cells, which feature round spectrosomes, as well as cyst cells with branched fusomes (Fig 3C). The increased number of GSC-like cells in loss of sakura suggests dysregulation of GSC self-renewal and differentiation. Meanwhile, the presence of excess cyst cells with branched fusomes indicates abnormal differentiation and division of cysts. Differentiation of cystoblasts typically involves four mitotic divisions with incomplete cytokinesis, leading to 16 cyst cells interconnected by branched fusomes^7,46. The degree of fusome branching serves as a marker for the stages of cyst cell division, with increased branching from 2-cell to 16-cell cysts^47–49. Notably, fusomes begin to degenerate and disappear after 16-cell cysts as the germline cyst enters the meiotic zone or region 2 of the germarium^50,51. The persistence of cyst cells with branched fusomes in sakura^null/null suggests that sakura is crucial for the proper division and differentiation of cysts, and we speculate that it is required for germline cysts to enter meiotic division.

Mosaic analysis of sakura^null indicates that sakura is intrinsically required for establishing GSCs in the ovaries. A significantly reduced number of sakura^null marked GSCs were observed at the adult stage when mutant clones were induced in PGC stages, suggesting that many sakura mutant PGCs die before developing into GSCs. Furthermore, after inducing sakura^null clones in adult ovaries, the number of marked sakura^null GSCs decreased at a higher rate than controls, indicating a rapid loss of sakura^null GSCs (Fig 6A). Thus, sakura is also intrinsically necessary for maintaining GSCs. Additionally, germaria containing sakura^null GSCs became increasingly tumorous over time, with a growing number of sakura^null GSC-like cells (Fig 6B and 6C). This suggests that sakura^null GSCs undergo uncontrolled division, becoming more tumorous over time, which eventually triggers an intrinsic cell death mechanism, as evidenced by elevated levels of cleaved Caspase-3 in sakura^null ovaries (Fig 3B).

Dpp/BMP signaling governs GSC self-renewal and cystoblast differentiation by repressing bam in GSCs and de-repressing it in daughter cystoblasts (Fig S1). This process is mediated by the transcription factor Mad, which, when phosphorylated (pMad), translocates into the nucleus to repress bam transcription^7,46,52. Our findings indicate that Bam expression is not limited to 8-cell cysts but continues throughout the germarium in ovaries lacking sakura function (Fig 7A and 7B). The persistent expression of Bam is likely due to low levels of pMad in GSCs (Fig 7C, and 7D). Given the low pMad levels observed upon loss of sakura, we believe that the misregulation of Bam expression occurs primarily at the transcriptional level, although it has also been reported that Bam expression can be regulated post-transcriptionally⁵³.

Transposons are mobile genetic elements that, if not silenced, can generate DNA damage and genomic instability⁵⁴. piRNAs derived from transposons and other repeats can target and silence transposon RNAs to preserve genome integrity in germ cells³¹. Loss of piRNAs leads to transposon derepression, resulting in increased DNA damage, which subsequently triggers cell death^55,56. Genetic damage in germ cells can cause developmental defects and diseases that may be inherited by the next generation. Thus, elimination of defective germ cells is crucial for maintaining germline integrity of a species^57,58. We found that loss of sakura results in reduced piRNA levels and loss of piRNA-mediated transposon silencing in the germline (Fig 4). The observed apoptosis, indicated by elevated cleaved Caspase-3 levels in sakura mutant ovaries (Fig 3B), suggest that desilencing of transposon due to reduced piRNA levels likely results in increased DNA damage, triggering cell death. We speculate that the germless phenotype in sakura mutants may partly arise from an apoptotic germline elimination program activated to maintain germline integrity. Loss of piwi, a key component of the piRNA pathway, similarly causes GSC-like tumors, as well as reduced bam expression and elevated pMad expression²⁹. While downregulation of bam and upregulation of pMad in piwi mutant is opposite to upregulation of bam and downregulation of pMad in sakura mutant, defects in the piRNA pathway might be a shared underlying cause for the dysregulation of bam and pMad levels in both cases, leading to tumorous phenotype.

In addition to its expression in GSCs and cysts in the germarium, Sakura is also expressed in germline cells in later-stage egg chambers (Fig 1E-F) and is required at this stage for proper oogenesis. When Sakura is depleted from region 2b of the germarium onward, while leaving GSCs and early cyst cells intact, female flies lay significantly fewer eggs and have fewer stage 14 oocytes (Fig 5). The failure to produce stage 14 oocytes likely results from the Orb mislocalization in developing oocytes (Fig 5E). Orb is an RNA-binding protein crucial for egg chamber formation, oocyte specification, and polarity establishment during oogenesis by regulating mRNA localization and translation, such as that of oskar, which is vital for establishing oocyte anterior-posterior polarity^34,59. Interestingly, piRNA pathway mutants exhibit Orb mislocalization in developing oocytes⁶⁰. Therefore, reduced piRNA levels observed (Fig 4) may underlie the Orb mislocalization in loss of sakura (Fig 5E).

Sakura does not possess any known protein domains, complicating investigations into its molecular function. To infer its function, we identified interacting proteins through immunoprecipitation followed by mass spectrometry. We identified Otu, a protein known to be crucial for oogenesis, as a protein partner of Sakura. Mutations in otu gene lead to a range of ovarian phenotypes, including germ cell loss, tumorous egg chambers filled with undifferentiated germ cells, abnormalities in nurse cell chromosome structure, and defects in oocyte determination^22,23. We showed that germline depletion of Otu via RNAi phenocopies loss of sakura. Similar to Sakura, loss of otu inhibits Dpp/BMP signaling, resulting in low pMad levels in GSCs and Bam de-repression (Fig 12A-B). Additionally, loss of otu also results in the loss of piRNA-mediated silencing, paralleling the effects of sakura loss. Furthermore, germline knockdown of otu yielded a similar ratio of germless ovaries as seen in sakura-RNAi, and simultaneous knockdown of both otu and sakura exacerbated the germless ovary phenotype (Fig 8). These observations raise the possibility that Sakura and Otu function together to regulate germ cell maintenance in the ovaries and are involved in Dpp/BMP signaling to balance stem cell renewal and differentiation.

Otu possesses deubiquitinase activity, catalyzed by its N-terminal Otu domain^21,61. In germ cells, Otu interacts with Bam, forming a deubiquitinase complex that deubiquitinates and thereby stabilizes CycA, promoting GSC differentiation²¹. The predicted structure of Sakura and Otu complex suggests that the Mid domain of Sakura directly contacts the Otu domain of Otu (Fig 10A). We found that Sakura lacks a deubiquitinase activity and it does not directly affect Otu’s deubiquitinase rate in our in vitro assays using Ub-Rhodamine 110 as a substrate (Fig S6). This does not preclude the possibility that Sakura may still influence Otu’s deubiquitinase activity, potentially guiding Otu to its substrate and determining substrate specificity. Our in vitro assays may not have detected such specificity changes. Otu is also an RNA-binding protein, and its deubiquitinase activity can be enhanced by bound RNAs⁷. Future studies should explore whether Sakura regulates Otu’s RNA-binding or other RNA-related functions. Notably, while Otu is expressed in various tissues, including testes and gut^42,61, Sakura is exclusively expressed in ovaries, particularly in germline cells such as GSCs, Therefore, Sakura may enhance Otu’s functions specific to female germline cells. We believe that identifying Otu’s deubiquitinase substrates and its RNA partners, along with investigating Sakura’s role in these processes, will deepen our understanding of Sakura’s and Otu’s molecular functions in oogenesis. One speculative hypothesis is that Sakura may regulate Otu’s deubiquitinase activity concerning piRNA pathway proteins crucial for proper piRNA production and function.

In summary, this study identifies and characterizes the previously unknown gene sakura, which is specifically expressed in female germ cells and is essential for oogenesis and female fertility. Together with its protein partner Otu, Sakura likely regulates germline cell fate, maintenance, and differentiation

Materials and methods

Fly strains

We generated the sakura^null strain by introducing indels within the Sakura coding region using the CRISPR/Cas9 genome editing system, as we previously reported^26,62–65. The transgenic sakura-EGFP, otu-EGFP, and otu(Δ-Tudor)-EGFP strains were created following previously published methods^65,66. A fragment containing Sakura cDNA flanked by ∼1 kbp upstream and ∼1 kbp downstream genomic sequences was cloned. Fragments containing Otu and Otu(Δ-Tudor) cDNAs, flanked by ∼2.5 kbp upstream and ∼1 kbp downstream genomic sequences, were also cloned. The EGFP gene was fused in-frame to the C-terminus of the Sakura and Otu coding sequences within these fragments, which were then inserted into a pattB plasmid vector. The sakura-EGFP plasmid was integrated into the 51C1 site within the fly genome using the BDSC:24482 fly strain and the PhiC31 system, while the otu-EGFP and otu(Δ-Tudor)-EGFP plasmids were integrated into the 25C6 site of the genome using the attP40 fly strain and the PhiC31 system.

The sakura RNAi #1 (VDRC: v39727) and #2 (VDRC: v103660), y-RNAi (v106068), TOsk-Gal4 (v 314033), Burdock sensor [UAS-Dcr2; NGT-Gal4; nosGal4-VP16, nos>NLS_GFP_lacZ_vas-3’UTR_burdock-target] (v 313217) strains were from Vienna Drosophila Resource Center. w-RNAi (BDSC: 35573), otu-RNAi (BDSC: 34065), bam-RNAi (BDSC: 33631), cycA-RNAi (BDSC: 29313), UAS-Dcr-2; NGT-Gal4 (BDSC: 25751), FRT82B/TM6C, Sb (BDSC: 86313) and otu¹⁴ (BDSC: 6025) were obtained from the Bloomington Stock Center. The bam-GFP reporter (DGRC: 118177) and vasa-EGFP knocked-in fly (DGRC: 118616) were from Kyoto Drosophila Stock Center.

Fertility assay

The female fertility assay was performed as previously described^64,65,67,68. Briefly, five test virgin females of each test genotype were mated with three wild-type (OregonR) males in a cage with a 6-cm grape juice agar plate supplemented with wet yeast paste. The agar plates were replaced daily, and the number of eggs laid on the third plate (on day 3) was recorded. After incubation at 25°C for an additional day, the number of hatched eggs was counted. At least three cages per genotype were tested.

For the male fertility assay, a single test male was mated with five wild-type (OregonR) virgin females in each vial, following previously published methods^64,65,67. After 3 days, the females were transferred to a new vial (vial 1). Every two days, they were transferred to a new vial until a total of four vials were obtained. After 2 days in the fourth vial, the females were removed, and the total number of progenies emerging from these four vials was counted. At least five males per genotype were tested.

Sakura and Otu antibodies

We expressed a recombinant full-length Sakura protein as an N-terminal 6xHis-tagged protein in E. coli using a modified pET vector and purified it using Ni-sepharose (GE Healthcare) and HiTrapQ HP (GE Healthcare) columns, as we previously performed for other proteins^64,65. Recombinant Otu fragments (145-405aa and 406-853aa) were similarly expressed as N-terminally 6xHis-MBP-fusion proteins in E. coli and purified using Ni-sepharose. These purified proteins were used as antigens to generate polyclonal anti-Sakura and anti-Otu sera in rabbits (Pocono Rabbit Farm & Laboratory, Inc.). The rabbit polyclonal anti-Sakura antibodies were affinity purified using His-MBP-Sakura recombinant protein and Affigel-15 (Bio-rad), following the manufacturer’s instructions. Rabbit anti-Otu sera were first pre-cleared with purified His-MBP protein bound to Affigel-15, and the unbound fraction was affinity purified using purified His-MBP-Otu fragments (145-405aa and 406-853aa) and Affigel-15.

Germline clonal analysis

We generated sakura^null mutant clones using FLP/FRT-mediated recombination⁶⁹. To create sakura^nullGSC clones, 3-day-old female flies with the genotype hs-flp; +; FRT82B, ubi-GFP/FRT82B, sakura^null were heat-shocked at 37°C for 1 hour, twice daily, with an 8-hour interval between heat shocks. Female flies with the genotype hs-flp; +; FRT82B, ubi-GFP/FRT82B were used as controls. Flies were dissected and stained 4, 7, and 14 days after clone induction. To induce primordial germ cell (PGC) clones, early third-instar larvae were heat-shocked similarly, allowed to develop, and dissected and analyzed at 3 to 4 days post-eclosion.

Immunostaining

Stereomicroscope images of dissected ovaries were taken using Leica M125 stereomicrocsope⁶⁴. Ovaries from 2- to 5-day-old, yeast-fed females were hand-dissected in 1X PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, pH 7.4) at room temperature. The dissected ovaries were fixed in a fixative buffer (4% formaldehyde, 15 mM PIPES (pH 7.0), 80 mM KCl, 20 mM NaCl2, 2 mM EDTA, and 0.5 mM EGTA), incubated for 30 minutes at room temperature with gentle rocking. After fixation, the ovaries were rinsed three times with PBX (0.1% Triton X-100 in 1X PBS) and then incubated in a blocking buffer (2% donkey serum in 3% BSA [w/v], 0.02% NAN3 [w/v] in PBX) for 1 hour at room temperature. Then, the ovaries were incubated with primary antibodies diluted in the blocking buffer overnight at 4°C. The following day, the ovaries were rinsed three times with PBX and incubated with Alexa Fluor-conjugated secondary antibodies for 2 hours. The ovaries were rinsed three times with PBX and then mounted in VECTASHIELD® PLUS antifade mounting medium with DAPI (H-2000, Vector lab). Confocal images were acquired on a Zeiss LSM700 confocal microscope at the Johns Hopkins University School of Medicine Microscope Facility.

The primary antibodies used for immunostaining were mouse anti-HTS (1B1) (DSHB, AB_528070, 1/100), mouse anti-Bam (DSHB, AB_10570327, 1/20), mouse anti-CycA (DSHB, AB_528188, 1/100), rat anti-Vasa (DSHB, AB_760351, 1/100), rabbit anti-pMad (Phospho-SMAD1/5 (Ser463/465) mAb #9516, dilution: 1/200, Cell Signaling), and rabbit anti-cleaved caspase-3 (Cleaved Caspase-3 (Asp175) #9661, dilution: 1/200, Cell Signaling). Secondary antibodies used were Alexa Fluor 488 Donkey anti-Mouse Igg (ThermoFisher A21202, 1/100), Alexa Fluor 594 Donkey anti-Rat Igg (ThermoFisher A21209, 1/100), Alexa Fluor 594 Donkey anti-Mouse Igg (ThermoFisher A21203, 1/100), and Alexa Fluor 594 Donkey anti-Rabbit Igg ThermoFisher A21207, 1/100).

Western blot

Lysates of hand-dissected ovaries and tissues were prepared by homogenizing in RIPA buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1% [v/v] IGEPAL CA-630, 0.1% [w/v] sodium dodecyl sulfate (SDS), 0.5% [w/v] sodium deoxycholate, 1 mM ethylenediaminetetraacetic acid (EDTA), 5 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride (PMSF))^63–65,67. The homogenates were centrifuged at 21,000g at 4°C for 10 min, and the protein concentration of the supernatant was determined using the BCA protein assay kit (Pierce) as needed. Fifteen µg of total protein was loaded per lane for Western blot. The sources and dilutions of the primary antibodies were as below. Rabbit anti-Sakura (1/10000, generated in this study), rabbit anti-Otu (1/10000, generated in this study), rabbit anti-alpha-Tubulin [EP1332Y] (1/10000, Abcam, ab52866), mouse anti-alpha-Tubulin [12G10] (1/10000, DSHB, AB_1157911), mouse anti-FLAG (1/10000, Sigma, F1804), mouse anti-HA (1/10000, Sigma, H3663), and mouse anti-GFP [GF28R] (1/3000, Invitrogen, 14-6674-82). IRDye 800CW goat anti-mouse IgG, IRDye 800CW goat anti-rabbit IgG, IRDye 680RD goat anti-mouse, and IgG IRDye 680RD goat anti-rabbit were used as secondary antibodies. The membranes were scanned using the Li-Cor Odyssey CLx Imaging System.

Mass spectrometry

Immunoprecipitation of Sakura-EGFP protein was performed using the GFP-Trap Magnetic Agarose Kit (Proteintech, gtmak-20) on dissected ovaries from flies harboring sakura-EGFP transgene, with w1118 flies as controls. Ovaries were homogenized in 200 µL ice-cold lysis buffer (10 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5 mM EDTA, 0.05% [v/v] IGEPAL CA-630) containing 1× protease inhibitor cocktail (100× protease inhibitor cocktail contains 120 mg/ml 1 mM 4-(2-aminoethyl) benzene sulfonyl fluoride hydrochloride (AEBSF), 1 mg/ml aprotinin, 7 mg/ml bestatin, 1.8 mg/ml E-64, and 2.4 mg/ml leupeptin)⁷⁰. After homogenization, the tubes were placed on ice for 30 minutes, and the homogenates were extensively pipetted every 10 minutes. The lysates were then centrifuged at 17,000x g for 10 minutes at 4°C. The supernatants were transferred to pre-chilled tubes, and 300 µL dilution buffer (10 mM Tris/Cl pH 7.5, 150 mM NaCl, 0.5 mM EDTA) supplemented with 1x protease inhibitor cocktail were added. The diluted lysates were then added to the GFP-trap magnetic beads in 1.5 mL tubes and rotated for 1 hour at 4°C. After separating the beads with a magnetic tube rack, the beads were washed three times with 500 µL wash buffer (10 mM Tris/Cl pH 7.5, 150 mM NaCl, 0.05 % [v/v] IGEPAL CA-630). Proteins were eluted with 40 µL acidic elution buffer (200 mM glycine pH 2.5) followed by immediate neutralization with 5 uL neutralization buffer (1 M Tris pH 10.4).

As a quality control before mass spectrometry, ∼5 µL of the samples were mixed with an equal volume of 2× SDS-PAGE loading buffer (80 mM Tris-HCl [pH 6.8], 2% [w/v] SDS, 10% [v/v] glycerol, 0.0006% [w/v] bromophenol blue, 2% [v/v] 2-mercaptoethanol), heated at 95 °C for 3 min, and run on 4–20% Mini-PROTEAN® TGX™ Precast Protein Gels (Biorad, #4561094). Silver staining was then performed by using Pierce™ Silver Stain Kit (ThermoFisher, 24612) to assess the quality of the immunoprecipitated protein samples. Mass spectrometry was conducted at the Mass Spectrometry Core at the Department of Biological Chemistry, Johns Hopkins School of Medicine, as previously described⁶⁵

Co-immunoprecipitation

For co-immunoprecipitation of endogenous Sakura protein, wild-type ovaries (w1118) were homogenized in RIPA buffer, centrifuged at 21,000g at 4°C for 10 min, and the clear supernatant protein lysates were used for immunoprecipitation. Four µg of rabbit anti-Sakura and normal rabbit IgG (Cell Signaling, #2729) were incubated with 50 µL of Dynabeads Protein G (ThermoFisher, 10004D) for 20 min at room temperature. The beads were washed once with PBST (1x PBS with 0.1% Tween-20). The ovary lysate supernatant was then incubated with the washed beads at room temperature for 30 min, followed by three washes with PBST. The proteins were eluted with 2x SDS-PAGE loading buffer (80 mM Tris-HCl [pH 6.8], 2% [w/v] SDS, 10% [v/v] glycerol, 0.0006% [w/v] bromophenol blue, 2% [v/v] 2-mercaptoethanol) and heated at 70°C for 10 min. After bead separation using a magnetic tube rack, the eluted proteins in 2x SDS-PAGE loading buffer were heated again at 95°C for 3 min.

Transient protein expression in S2 cells was performed using the pAc5.1/V5-HisB plasmid vector (Invitrogen). A total of 1 µg plasmids were transfected using the Effectene transfection reagent (Qiagen, 301425). Three days after transfection, cells were harvested and lysed with ice-cold lysis buffer supplemented with 1× protease inhibitor cocktail. The cell lysates were then centrifuged at 17,000x g for 10 min at 4°C, and the clear supernatants were collected for immunoprecipitation. For anti-HA immunoprecipitation, supernatants were incubated with 25 µL (0.25 mg) of Pierce™ Anti-HA Magnetic Beads (ThermoFisher, 88837) at room temperature for 30 min. The beads were then washed three times with TBST (1x TBS [0.05 M Tris/HCl and 0.15 M NaCl, pH 7.6] with 0.05% Tween-20). For anti-FLAG immunoprecipitation, 2 µg of mouse anti-FLAG (1/10000, Sigma, F1804) was incubated with 50 µL of Dynabeads Protein G (ThermoFisher, 10004D) for 10 min at room temperature. The beads were washed once with PBST. The S2 cell lysate supernatant was incubated with the washed beads at room temperature for 15 min. The beads were washed three times with PBST. In both anti-HA and anti-FLAG immunoprecipitations, proteins were eluted with 2× SDS-PAGE loading buffer. For anti-HA, the beads in 2x SDS-PAGE loading buffer were heated at 95°C for 7 min. For anti-FLAG, the beads were heated at 70°C for 10 min, then were separated using a magnetic tube rack. The eluted proteins in 2x SDS-PAGE loading buffer were heated again at 95°C for 3 min.

Small RNA and mRNA sequencing

Poly-A+ mRNA purification was performed as previously described⁷¹. Small RNA libraries and poly-A+ mRNA libraries were prepared, sequenced on Hiseq2500 (Illumina), and analyzed, as previously reported^72,73. SRA accession number for these datasets is PRJNA1156618.

In vitro deubiquitination assay

C-terminally 6xHis-tagged recombinant Sakura protein was expressed in E. coli using a modified pET vector and was purified using Ni-sepharose. N-terminally 3xHA-HRV3Csite-3xFLAG-tagged Otu protein and a negative control, N-terminally 3xHA-HRV3Csite-3xFLAG-tagged firefly luciferase protein, were expressed in S2 cells using pAc5.1/V5-HisB vector. The proteins were immunoprecipitated using Pierce™ Anti-HA Magnetic Beads (ThermoFisher, 88837) as described previously, with a more stringent washing step to effectively remove interacting proteins. The beads were washed six times with a high salt wash buffer (TBST with 800 mM NaCl). The proteins were eluted by cleaving off 3xHA tag with 25 nM GST-HRV3C protease in a cleavage buffer (25mM Tris-HCl pH 7.4, 150 mM NaCl, 5% glycerol, 2mM EDTA), incubated at 4°C with rotation for 6 hours. An equal volume of 100% glycerol was added to the eluted protein.

Purified proteins, including Otu, luciferase, and Sakura, were mixed with Ub-Rhodamine 110 (Ubiquitin-Proteasome Biotechnologies, M3020) in a total volume of 30 µL in reaction buffer (20 mM Tris-HCl, pH 7.5, 200 mM NaCl, 5 mM MgCl₂, 2 mM DTT). The mixture was added to a black 384-well low-volume plate. Fluorescence measurements were taken using SpectraMax i3x Multi-Mode Microplate Reader at 37°C, with excitation and emission wavelengths set at 485/20 and 530/20 nm respectively. The fluorescence intensity for each condition was averaged from triplicates and plotted as a function of time.

Acknowledgements

We thank Bloomington Drosophila Stock Center, Vienna Drosophila Resource Center, and Kyoto Drosophila Stock Center for fly strain stocks. We thank the Johns Hopkins University School of Medicine Microscope Facility for use of the Zeiss LSM700, supported by NIH grant S10OD016374 awarded to Dr. Scot C. Kuo. We thank Ms. Lauren DeVine and Dr. Bob Cole at the Johns Hopkins University School of Medicine Mass Spectrometry and Proteomics Core Facility for the mass-spec analysis. This work was supported by the grants from the National Institutes of Health [R35GM145352] and Johns Hopkins University Catalyst Award to RF.

Additional information

Author Contributions

Methodology and Investigation, A.A., L.Z., and R.F.; Conceptualization and Writing, A.A. and R.F.; Funding Acquisition and Supervision, R.F.

Competing interests

The authors have no conflicts of interest to declare.

Supplementary figures

Fig S1.

Fig S2.

Fig S3.

Fig S4.

Fig S5.

Fig S6.