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 GSC and oogenesis research (Kirilly and Xie 2007).

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 1A). 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 (Lin and Spradling 1993; de Cuevas and Matunis 2011; Losick et al. 2011). GSCs undergo asymmetric mitotic division, producing two distinct cells: one GSC and one cystoblast. The cystoblast, destined to differentiate, undergoes four mitotic divisions with incomplete cytokinesis, resulting in interconnected cysts of 2, 4, 8, and 16 cells. 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 oocyte.

Fig 1.

GSCs must be tightly regulated to maintain their undifferentiated state while dividing and differentiating to produce one GSC and one cystoblast. Failure in GSC self-renewal or inappropriate differentiation leads to stem cell loss, disrupting oocyte production (Lin 1997; Cox et al. 1998). 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 (Gateff et al. 1996; Ohlstein et al. 2000). 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 (Song et al. 2002; Gilboa et al. 2003). 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 (Fig S1) (Xie and Spradling 1998; Casanueva and Ferguson 2004). 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 (Kai and Spradling 2003; Song et al. 2004).

Bam, a ubiquitin-associated protein, is essential for cystoblast differentiation (McKearin and Spradling 1990; McKearin and Ohlstein 1995). Dpp/BMP signaling in the niche represses bam expression, preventing GSCs from differentiating and enabling self-renew (Song et al. 2004). However, as the daughter cells (cytoblasts) exit the niche, they derepress bam and initiate the differentiation program (Fig S1). 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 (tumorous phenotype), whereas ectopic bam expression forces GSCs differentiation, depleting the stem cell pool (McKearin and Spradling 1990; McKearin and Ohlstein 1995; Ohlstein and McKearin 1997).

One of the proteins that interacts with Bam is Ovarian Tumor (Otu) (Ji et al. 2017). Otu forms a deubiquitinase complex with Bam, which deubiquitinates Cyclin A (CycA), stabilizing CycA and promoting GSC differentiation (Ji et al. 2017). Otu also binds RNAs (Ji et al. 2019). Otu is crucial for oogenesis and female fertility. Mutations in the otu gene cause various ovarian defects, including tumorous growths, germ cell loss, abnormalities in oocyte determination and nurse cell dumping, and defects in the sexual identity determination of germ cells, indicating Otu’s importance in multiple stages of oogenesis (Smith and King 1966; Gans et al. 1975; Gollin and King 1981; King and Riley 1982; Storto and King 1988; Pauli et al. 1993; Rodesch et al. 1997; Glenn and Searles 2001). 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 1B). We found that 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 and for the female-specific mRNA splicing pattern of sex-lethal (sxl), a master regulator of sex determination. 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.

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 S2). 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 ovaries, particularly in stage 1-11 egg chambers (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 1B). 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 1B). This truncated fragment is unlikely to be functional, and we were unable to detect stable protein expression, as shown below (Fig 1D). 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).

Oogenesis begins in the germarium, which contains 2-3 GSCs identified by round, unbranched spectrosomes that contact cap cells (Fig 1A) (Kirilly and Xie 2007). In contrast, cysts exhibit branched fusomes. Immunostaining with hu-li tai shao (HTS) antibody marks both spectrosomes and fusomes (Fig 1E), while Vasa is a known germ cell marker (Fig 1F). 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 1C and S2), 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 females (sakura^+/+ and sakura^null/+) laid numerous eggs (Fig 2A and 2B). Dissection of sakura^null/null ovaries 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 S3).

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 marker. We observed that some sakura^null/null ovarioles were devoid of germ cells (“germless”, cyan stars), while others retained germ cells (Fig 3A and 3B). Immunostaining with HTS antibody revealed an excess number of GSC-like cells with round spectrosomes in sakura^null/null ovarioles containing germ cells (Fig 3C, orange stars), indicative of a ”tumorous” phenotype as previously described for mutants of other genes including bam, otu, and sxl (Smith and King 1966; Gans et al. 1975; Gollin and King 1981; King and Riley 1982; McKearin and Spradling 1990; Bopp et al. 1993; Eliazer et al. 2011; Jin et al. 2013; Yang et al. 2019). Additionally, we observed an excess number of cyst cells with branched fusomes that persisted throughout the ovarioles, suggesting abnormal cyst cell differentiation and division. Thus, loss of sakura results in both germ cell depletion and overgrowth.

Fig 3.

Within the same ovary, some sakura^null/null ovarioles exhibited the tumorous phenotype, while others were germless, without any spectrosome or fusome staining (Fig 3A-3C, cyan stars). Quantification revealed that 35% of sakura^null/null ovarioles from 2-to 5-day-old flies were tumorous, while the remaining 65% were germless (n=74) (Fig 3D). In contrast, all ovarioles in control (sakura^+/+ and sakura^null/+) and sakura-EGFP rescue flies were normal with no tumorous or germless phenotypes observed (Fig 3D). Approximately 95% of sakura^null/null ovarioles containing germ cells had an excess of GSC-like cells (>5) (Fig 3E and Fig S4). The mean number of GSCs or GSC-like cells in 2- to 5-day-old 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).

We investigated the age-dependent effects on the tumorous and germless phenotypes by examining the ovarioles from 0-1 day old, 7-day old, and 14-day old flies. Ovarioles in control and sakura-EGFP rescue flies remained normal throughout this time course (Fig S5A). In the very young, 0-1 day old sakura^null/null, 78% of ovarioles were tumorous and 22% were germless. Overtime, the ratio of germless ovarioles increased while the ratio of tumorous ovarioles decreased. By 14 days old, only 5% of ovarioles were tumorous, with the remaining 95% being germless. The number of GSC-like cells in sakura^null/null ovarioles was already high in 0-1 day old flies, and it decreased over time (Fig S5B). Tumorous ovarioles, but not germless ones, exhibited significantly higher levels of cleaved Caspase-3 staining, indicative of apoptosis, compared to controls (Fig 3F). These results suggest that tumorous ovarioles undergo apoptosis, and that these ovarioles eventually become germless.

Together, these results suggest that Sakura is essential for regulating the survival, proliferation, and differentiation of germline cells, including GSCs.

Loss of sakura results in reduced germline piRNA

In controls, Vasa-EGFP is enriched in the perinuclear structure known as the nuage within nurse cells (Fig 3A and 3B), which is essential for piwi-interacting RNA (piRNA) biogenesis. In sakura^null/nulltumorous ovarioles, Vasa-EGFP retains its perinuclear localization, indicating that sakura is not required for the proper localization of Vasa to the nuage.

piRNAs are produced in germline cells and somatic follicle cells and transcriptionally and post-transcriptionally silence transposon expression through sequence-complementarity (Huang et al. 2017; Yamashiro and Siomi 2018). Disruption of the piRNA pathway can result in oogenesis arrest, germ cell loss, rudimentary ovaries, and sterility. 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 (Donertas et al. 2013; Handler et al. 2013). We found that Burdock piRNA levels were significantly lower in sakura^null/null compared to controls and rescue flies (Fig 4C).

Fig 4.

We employed the Burdock sensor, a transposon reporter tool to monitor germline piRNA activity (Handler et al. 2013). 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′ UTR (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, suggesting that Sakura is essential for proper piRNA levels and piRNA-based transposon silencing in the germline.

Sex-specific sxl mRNA alternative splicing is dysregulated in sakura^null/null ovaries

Sxl is a master regulator of sex determination in Drosophila (Penalva and Sanchez 2003; Salz and Erickson 2010; Grmai et al. 2022). Sex-specific alternative splicing of sxl transcripts produces distinct mRNA isoforms: the female-specific and male-specific mRNA isoforms in respective sex and only the female-specific mRNA isoform encodes the functional Sxl protein, whereas the male-specific isoform does not. In the female germline, loss of Sxl function or disruption of female-specific sxl splicing leads to developmental defects, including germline tumors and sterility. Notably, several mutants with tumorous ovariole phenotypes—including otu mutants—exhibit aberrant sxl mRNA splicing, resulting in the expression of the male-specific isoform in ovaries and defects in germ cell sexual identity (Bopp et al. 1993; Pauli et al. 1993).

We examined sxl splicing in sakura mutants. As expected, ovaries from control and sakura-EGFP rescue flies expressed exclusively the female-specific sxl mRNA isoform and testes from control flies expressed only the male-specific isoform (Fig S6). In contrast, sakura^null/nullovaries exhibited the male-specific isoform and a reduced level of the female-specific isoform. These results indicate that female-specific sxl alternative splicing is disrupted in the absence of sakura.

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 1A), sparing GSCs and germline cysts (ElMaghraby et al. 2022). 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 ovaries beginning around stage 6 of oogenesis (Fig 5E). In controls, Orb was enriched and properly localized to the posterior end of stage ∼6-8 egg chambers (Fig 5E, yellow arrows). In sakura^RNAi ovaries, although Orb localization appeared normal at stage ∼6 (yellow arrow), mislocalization was evident by stage ∼8 (magenta arrow), and signs of cytoskeletal disorganization were apparent in later-stage egg chambers (white arrowheads). Phalloidin staining to visualize F-actin further supported cytoskeletal defects in sakura^RNAi later-stage egg chambers (Fig S7).

These defects likely contribute to the impaired production of stage 14 oocytes and eggs in sakura^RNAi flies (Fig 5C and 5D). We conclude that Sakura plays an essential role in oogenesis beyond the early germline stages. For all subsequent sakura RNAi experiments, we used the sakura RNAi #2 line.

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

To investigate whether sakura functions autonomously in the germline, we performed mosaic analysis using the FLP-FRT system with heat shock promoter-driven FLP (hs-flp) (Xu and Rubin 1993; Rubin and Huynh 2015). First, to assesse its role in GSC establishment, we induced sakura^null clones in primordial germ cells (PGCs) by applying heat shock before the early third instar larval stage and tracked their development into adult GSCs, following a previously established protocol (Yang et al. 2007). In control adult flies (FRT82B),12.3% (20/163) of GSCs were marked, wheras only 1.9% (4/213) were marked in sakura^nullmutants (FRT82B, sakura^null) (p-value = 4.8 * 10^-11, chi-square test), indicating that sakura is autonomously required for GSC establishment.

Next, we assessed GSC maintenance using the same system by generating sakura^null GSC clones in adult flies and measuring clone loss over time, as described previously (Xie and Spradling 1998; Yang et al. 2007). Four days after clone induction, 27.8% of GSCs were marked in controls (FRT82B) and 22.1% in sakura^null (FRT82B, sakura^null), which we defined as initial levels (Fig 6A). In controls, the percentage of marked GSCs declined to 25.8% on day 7 and and 16.8% on 14, resulting in a 39.4% loss rate over 10 days (from day 4 to day 14). In contrast, the proportion of marked sakura^null GSC clones dropped more sharply—to 16.3% on day 7 and and 5.4% on day 14—yielding a higher loss rate of 75.6% over 10 days. These results demonstrate that sakura is intrinsically important for GSC maintenance.

Fig 6.

We further tested whether sakura is intrinsically required 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 containing marked GSC clones at days 4, 7, and 14 post-induction (Fig 6B). We found that marked sakura^null GSC-like cells were significantly more numerous than 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, unmarked GSC-like cells in germaria containing sakura^nullGSC clones did not differ significantly in number compared with those in germaria containing marked control GSC clones, and the number did not increase over time (Fig S8).

These results demonstrate that sakura is intrinsically important in germline cells, including GSCs, to regulate proper division and differentiation. In the absence of intrinsic sakura, germline cells become tumorous and undergo uncontrolled proliferation. Taken together, we conclude that sakura is required intrinsically for GSC establishment, maintenance, and differentiation.

Loss of sakura inhibits Dpp/BMP signaling

The Dpp/BMP signaling pathway plays a central role in regulating bam expression and GSC self-renewal and differentiation (Fig S1) (Kirilly and Xie 2007; Hayashi et al. 2020). 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 reporter (Chen and McKearin 2003b). In controls (y^RNAi), Bam-GFP expression was restricted to 8-cell cysts and disappeared in 16-cell cysts and onward (Fig 7A). In contrast, sakura knockdowns ovaries showed persistent Bam-GFP expression throughout the germarium including the GSC niche region (Fig 7A).

Fig 7.

To further confirm this finding, we used the FLP-FRT system to induce sakura^null clones in germaria and performed immunostaining with anti-Bam antibody. In control clones (FRT82B), Bam expression was observed exclusively in 8-cell cysts. However, in all sakura^null clones (FRT82B, sakura^null), Bam was aberrantly expressed throughout the germarium, including in GSCs (Fig 7B). These results suggest that in the absence of sakura, Bam expression is no longer repressed by Dpp/BMP signaling in the GSC niche, leading to GSC loss, and is no longer shut off after the 16-cell cyst stage.

In GSCs, pMad translocates into the nucleus to repress Bam expression (Fig S1) (Kai and Spradling 2003; Song et al. 2004). 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^null was due to reduced pMad levels in GSCs.

Previous studies have shown that ectopic expression of a stable form of CycA leads to germ cell loss (Chen et al. 2009). This germ cell loss phenotype is also observed upon ectopic bam expression in GSCs (Xie and Spradling 1998; Chen and McKearin 2003a; Xia et al. 2010). It was reported that Bam associates with Ovarian tumor (Otu) to promote deubiquitination and stabilization of CycA (Ji et al. 2017). Since we observed derepressed bam expression in sakura^nullcells, 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 (FRT82B, sakura^null) and control clones in control germarium (FRT82B) (Fig 7E). sakura^null clones exhibited significantly higher mean CycA intensity compared to control clones (Fig 7F). The elevated CycA levels in sakura^null clones likely results from Bam misexpression and Bam-mediated stabilization.

To determine whether bam misexpression occurs in egg chambers in sakura knockdown conditions, we examined bam expression in TOsk-Gal4-driven sakura RNAi flies. Bam was appropriately restricted to 8-cell cysts in germaria and not misexpressed in egg chambers compared to yRNAi controls (Fig. S9, cyan arrowheads). Thus, sakura is specifically required in GSCs and cyst cells to regulate bam expression.

We also examined DE-Cadherin (DE-Cad), which plays a role in oocyte positioning within the egg chamber (Godt and Tepass 1998). DE-Cad level and localization appeared normal in both control and sakura RNAi ovaries through approximately stage 8. However, later-stage egg chambers in sakura RNAi flies displayed structural disorganization (Fig. S9, white arrowheads), consistent with defects observed in Fig. 5E and Fig. S7.

Attempts to rescue sakura loss-of-function ovariole phenotypes

To determine whether the phenotypes caused by sakura loss-of-function could be rescued, we first performed genetic interaction experiments by knocking down either bam or cycA in the germline. As expected, germline-specific bam RNAi driven by UAS-Dcr2 and NGT-Gal4 resulted in 100% tumorous ovarioles, while cycA RNAi produced 100% normal ovarioles in 2–5-day-old flies (Fig. S10A). Double knockdown of sakura and bam reduced the proportion of germless ovarioles and increased the proportion of tumorous ovarioles compared to sakura RNAi alone or the double knockdown of sakura and control w RNAi. However, no normal ovarioles were observed (Fig. S10A). sakura and cycA double knockdown had no effect on the phenotype distribution, suggesting that the germless phenotype in sakura mutants is partially attributable to ectopic bam expression but not to cycA misregulation.

The number of GSC-like cells was increased in sakura and bam double knockdowns compared to sakura RNAi or sakura and w double RNAi, whereas sakura and cycA knockdown did not alter GSC-like cell numbers (Fig. S10B),, further supporting a genetic interaction between sakura and bam.

Next, we tested whether the sakura loss-of-function phenotypes could be rescued by overexpressing components of the Dpp/BMP signaling pathway. Germline expression of UASp-Mad-GFP driven by nos-Gal4-VP16 in both sakura^null/+ controls and sakura^null/null did not alter the proportions of ovariole phenotypes (Fig S11A) or the number of GSC-like cells (Fig S11B). Thus, transgenic overexpression of Mad did not rescue the sakura mutant phenotypes.

We then tested whether overexpression of a constitutively active form of the Dpp receptor Thickveins (Tkv.Q253D) (Casanueva and Ferguson 2004) could rescue the phenotypes. Expression of UASp-tkv.Q253D in the germline using a NGT-Gal4 in a control (y^RNAi) background resulted in 100% tumorous ovarioles, as expected (Fig. S12A). When co-expressed with sakura RNAi, Tkv.Q253D increased the proportion of tumorous ovarioles and decreased the proportion of germless ovarioles relative to sakura^RNAi alone but did not restore normal ovarioles. The number of GSC-like cells was also increased by Tkv.Q253D expression in sakura^RNAi (Fig S12B), indicating partial phenotypic modulation via BMP pathway activation.

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 and validated it by Western blots using ovary lysates from the hypomorphic otu¹⁴ mutant (Fig S13A) (King et al. 1986; Storto and King 1987; Steinhauer and Kalfayan 1992a; Pauli et al. 1993). Using this antibody, we detected endogenous Otu protein in the Sakura-EGFP immunoprecipitants from ovaries of 3–7-day-old flies, but not in the negative control (Fig 8A), validating the mass spectrometry results. Next, using wild-type (without Sakura-EGFP) ovary lysates from 3–7-day old-flies, we immunoprecipitated endogenous Sakura with anti-Sakura antibody and detected endogenous Otu in the immunoprecipitant by Western blot (Fig 8B), 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 8C).

Fig 8.

The otu gene encodes two annotated protein isoforms, a predominant 98 kDa isoform and a less abundant 104 kDa isoform generated by alternative splicing (Steinhauer and Kalfayan 1992b). The 98 kDa isoform lacks the Tudor domain, which is encoded by an alternatively spliced 126-nucleotide exon (Fig 9B. Otu(ι1Tudor)) (Van Buskirk and Schupbach 2002). Previous studies have shown that the 104 kDa Otu isoform is more abundant in predifferentiated germline cells and is sufficient to carry all known Otu functions, whereas the 98 kDa Otu isoform becomes prominent during later stages of oogenesis and is implicated in nurse cell regression and oocyte maturation (Sass et al. 1995). Consistent with these findings, we observed that ovaries from very young (2–5 hour old) flies—containing only germaria and previtellogenic egg chambers—predominantly expressed the 104 kDa isoform (Fig S13B). In contrast, ovaries from 3–7 day old flies—containing all 14 stages of oogenesis—predominantly expressed the 98 kDa isoform. Importantly, both the 104 kDa and 98 kDa isoforms co-immunoprecipitated with endogenous Sakura using anti-Sakura antibody, demonstrating that Sakura interacts with both isoforms of Otu and that the Tudor domain is not required for this interaction (Fig. S13B).

Fig 9.

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 AlphaFold (Jumper et al. 2021; Yang et al. 2023). AlphaFold suggested 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 9A-C). The Tudor domain of Otu is not directly involved in this interaction, consistent with our findings (Fig S13B). 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 9B). 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 9D), 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.

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 9C). We found that Sakura(NM), Sakura(NC), Sakura(N), and Sakura(M) were associated with Otu (Fig 9E). 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 9E). Sakura-NC exhibited a weak interaction with Otu, while Sakura(MC) and Sakura(C) showed no interaction (Fig 9E). 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 Sakura(N) and Otu(N2) physically interact (Fig 9F). A reciprocal co-immunoprecipitation assay also confirmed their interaction (Fig S14), 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 10A). 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 10A). 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 (Glenn and Searles 2001).

Fig 10.

Previous studies have shown that mutations in otu lead to defects in germ cell division and differentiation, resulting in phenotypes including tumorous ovarioles (King and Riley 1982; King et al. 1986; Storto and King 1988; Steinhauer and Kalfayan 1992b). To study the role of otu specifically in the germline, we performed RNAi knockdown of otu in the germline driven by nos-Gal4-VP16 (Fig 10B and 10C). Interestingly, we found that germline depletion of otu results in rudimentary ovaries, similar to the loss of sakura (Fig 2C, Fig 5A, and Fig 10B). Furthermore, similar to sakura RNAi (Fig 5), TOsk-Gal4-driven otu RNAi knockdown did not affect ovary morphology (Fig 10B), but it resulted in a reduced number of eggs laid and stage 14 oocytes compared with control RNAi (w^RNAi) (Figure 10D and 10E).

Given that sakura loss of function causes tumorous and germless ovarioles, apoptosis, piRNA pathway defects, bam misexpression, and reduced pMad levels (Fig 3, 4 and 7A-D), we asked whether similar defects occur upon otu depletion. Germline otu RNAi knockdown in 2-5 days old flies using UAS-Dcr2 and NGT-Gal4 lead to both germless (∼70%) and tumorous (∼30%) ovarioles, closely resembling the phenotype of sakura RNAi (Fig S10A). Similar germless and tumorous phenotypes were also observed in nos-Gal4-VP16 > otu^RNAi flies (Fig 11A). In these otu^RNAiovarioles, Bam-GFP expression was no longer restricted to 8-cell cysts, instead persisting throughout the germarium and egg chambers (Fig 11A). Additionally, otu^14/14 mutant GSCs showed reduced pMad levels (Fig 11B), and cleaved Caspase-3 staining revealed elevated apoptosis in otu RNAi ovaries (Fig 11C). Germline depletion of otu resulted in elevated expression of GFP and β-Gal produced by the Burdock sensor, indicating loss of piRNA-mediated transposon silencing (Fig 11D).

Fig 11.

Together, these data demonstrate that otu loss-of-function closely phenocopies sakura loss-of-function, including defects in germline maintenance and differentiation, bam misexpression, reduced pMad signaling, and compromised piRNA pathway activity. These findings, along with physical interaction between Sakura and Otu (Figs 8, 9, S13, and S14), suggest that they function together in regulating germline maintenance and differentiation.

To further explore their genetic interaction, we compared the ovariole phenotypes caused by single and double knockdowns of sakura and otu in the germline driven by UAS-Dcr2 and NGT-Gal4. Individual knockdown of either gene resulted in ∼70% germless and ∼30% tumorous ovarioles in 2-5 days old flies (Fig S10A). Simultaneous knockdown of both genes increased the propotion of germless ovarioles to 86%, with a corresponding decrease in tumorous phenotypes to 14%, suggesting a synergistic enhancement of the germless phenotype. The number of GSC-like cells were not affected by the double knockdown of sakura and otu compared with their respective single knockdown (Fig S10B).

Sakura and Otu proteins do not depend on each other for their abundance

We investigated whether Sakura and Otu proteins depend on each other for their abundance. We performed Western blot analysis on ovaries with germline-specific knockdown of sakura or otu, driven by TOsk-Gal4. Otu protein levels were not reduced in sakura^RNAi ovaries compared to control y^RNAi (Fig S15A), and conversely, Sakura levels were not decreased in otu^RNAi ovaries compared with control w^RNAi ovaries (Fig S15B). These results indicate that Sakura and Otu do not require each other for their protein expression or stability.

To further test this conclusion and to examine whether Otu enrichment to the posterior of egg chambers is affected by loss of sakura, we generated marked sakura^null germline clones using the FLP-FRT system. Clones were marked by the absence of RFP, and experiments were performed in the flies expressing the otu-EGFP or otu(ΔTudor)-EGFP transgene in an otherwise wild-type (otu^+/+) background. Otu-EGFP fluorescence signal was comparable between sakura^nullclones (FRT82B, sakura^null) and control clones (FRT82B) (Fig 12A). Similarly, Otu(ΔTudor)-EGFP levels were unaffected in sakura^null clones compared to the control clones (Fig 12B), confirming that Sakura is not required for Otu protein expression or stability.

Fig 12.

Both Otu-EGFP and Otu(ΔTudor)-EGFP exhibited clear enrichment at the posterior of egg chambers in the control clones (FRT82B) —consistent with localization to the developing oocyte (Fig 12A and 12B). This posterior enrichment was absent in sakura^nullclones, demonstrating that Sakura is crucial for this process.

Sakura does not affect Otu’s deubiquitinase activity in vitro

A previous study has shown that Otu possesses deubiquitinase activity (Ji et al. 2017). 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 the previous study (Ji et al. 2017), we detected a deubiquitinase activity in Otu that was ectopically expressed and purified from S2 cells (Fig S16). 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.

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. 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 (de Cuevas and Spradling 1998). 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. The persistence of cyst cells with branched fusomes in sakura^null/nullsuggests 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 the establishment of GSCs in the ovary. When mutant clones were induced in PGC stage, significantly fewer sakura^null marked GSCs were observed in adult ovaries, suggesting that many sakura^null PGCs fail to survive or differentiate into GSCs. Furthermore, induction of sakura^null clones in adult ovaries led to a more rapid decline in marked sakura^null GSCs compared to controls, indicating that sakura is also intrinsically required for GSC maintenance (Fig 6A). Over time, germaria containing sakura^null GSC clones became increasingly tumorous, with an expanding population of sakura^null GSC-like cells (Fig 6B and 6C). In contrast, the number of GSC-like cells across entire and all ovarioles regardless of whether GSCs in the niche remain, declined over time in sakura^null ovaries (Fig S5B). These results suggest that sakura^null GSCs initially undergo aberrant proliferation, leading to tumor formation, but are ultimately sakura^nullGSCs and GSC-like cells are lost through a cell death mechanism, as evidenced by the elevated levels of cleaved Caspase-3 in sakura^null ovaries (Fig 3F).

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 (Kirilly and Xie 2007; Kahney et al. 2019; Hinnant et al. 2020). 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 misexpression of Bam in GSCs likely results from reduced levels of pMad in the GSCs (Fig 7C, and 7D). Given the low pMad levels observed upon loss of sakura, we propose that bam misregulation in GSCs occurs primarily at the transcriptional level, although Bam is also known to be subject to post-transcriptional regulation (Pek et al. 2009). The continued expression of Bam throughout the germarium, well beyond the GSC niche, suggests a failure to terminate bam expression at the 16-cell cyst stage —a developmental transition when bam is normally turned off. Thus, sakura mutants likely exhibit two distinct molecular defects contributing bam overexpression: (1) impaired Dpp/BMP signaling within the GSC niche, and (2) a failure to downregulate bam at 16-cell stage. Given the limited space within the GSC niche, the latter defect may be the predominant contributor to the observed bam overexpression. The molecular mechanism that silences bam at the 16-cell cyst stage remains poorly understood, and how sakura loss perturbs this regulation is unclear. The sakura mutant may therefore serve as a valuable model to investigate the mechanism that normally suppresses bam expression at the 16-cell stage.

Transposons are mobile genetic elements that, if not silenced, can generate DNA damage and genomic instability (Levin and Moran 2011). piRNAs derived from transposons and other repeats can target and silence transposon RNAs to preserve genome integrity in germ cells (Siomi et al. 2011). Loss of piRNAs leads to transposon derepression, resulting in increased DNA damage, which subsequently triggers cell death (Kang et al. 2018; Moon et al. 2018). 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 (Chu et al. 2014; Ota and Kobayashi 2020). 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 3F), 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.

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 —leaving GSCs and early cyst cells intact—females have significantly fewer stage 14 oocytes and lay significantly fewer eggs (Fig 5). The failure to produce stage 14 mature oocytes likely stems from cytoskeletal disorganization that disrupt oocyte development, as evidenced by mislocalization of Orb (Fig 5E, S7, and S12). Notably, piRNA pathway mutants also exhibit similar Orb mislocalization (Ohtani et al. 2013). Therefore, the reduced piRNA levels (Fig 4) may contribute to the oocyte development defects, including Orb mislocalization, in sakura mutants. (Fig 5E).

Sakura does not possess any known protein domains. To infer its function, 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, defects in germlne sexual identity determination, abnormalities in nurse cell chromosome structure, and defects in oocyte determination (King and Riley 1982; Storto and King 1988; Pauli et al. 1993; Glenn and Searles 2001). 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 11A-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 S10). 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 (Ji et al. 2017; Ji et al. 2019). In germ cells, Otu interacts with Bam, forming a deubiquitinase complex that deubiquitinates and thereby stabilizes CycA, promoting GSC differentiation. The predicted structure of Sakura and Otu complex suggests that the Mid domain of Sakura directly contacts the Otu domain of Otu (Fig 9A). 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 S16). 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 an RNA-binding protein whose deubiquitinase activity is enhanced by RNA binding (Ji et al. 2019). Bam, along with other proteins such as Bgcn, Mei-P26, and Sxl, binds —a key stem cell maintenance factor (Wang and Lin 2004)— and represses its translation in GSCs (Wang and Lin 2004; Li et al. 2009; Chau et al. 2012; Li et al. 2013). Bam and Otu form a protein complex (Ji et al. 2017). Future studies should explore whether Sakura modulates Otu’s RNA-binding properties and its interaction with other proteins. Although Sakura is exclusively expressed in ovaries, particularly in germline cells including GSCs (Fig 1), Otu is broadly expressed in various tissues, including testes and gut (Steinhauer and Kalfayan 1992b). This restricted expression pattern suggests that Sakura may serve as a female germline-specific cofactor that enhances or modifies Otu’s molecular functions. Identifying Otu’s RNA targets and deubiquitinase substrates beyond CycA, as well as determining how Sakura binding influences these activities, will be key to elucidating their roles in oogenesis. One intriguing possibility is that Sakura regulates Otu’s activity toward piRNA pathway components, thereby contributing proper piRNA biogenesis and function. Additionally, the Sakura-Otu complex may directly regulate essential post-transcriptional processes such as sxl alternative splicing and translational control of other oogenic RNAs.

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 (Zhu et al. 2018a; Zhu et al. 2018b; Zhu et al. 2019b; Zhu and Fukunaga 2021). The transgenic sakura-EGFP, otu-EGFP, and otu(ΔTudor)-EGFP strains were created following previously published methods (Fukunaga et al. 2012; Kandasamy and Fukunaga 2016; Zhu and Fukunaga 2021). A fragment containing Sakura cDNA flanked by ∼1 kbp upstream and ∼1 kbp downstream genomic sequences was cloned.

Fragments containing Otu and Otu(ΔTudor) 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(ΔTudor)-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_GF’_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), otu¹⁴ (BDSC: 6025), UASp-Mad-GFP (BDSC: 604592), and UASp-tkv.Q253D (BDSC: 604934) 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 (Zhu et al. 2018b; Liao et al. 2019; Zhu et al. 2019a; Zhu and Fukunaga 2021). 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 (Zhu et al. 2018b; Zhu et al. 2019a; Zhu and Fukunaga 2021). 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 (Fukunaga and Doudna 2009). 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.

Immunostaining

Stereomicroscope images of dissected ovaries were taken using Leica M125 stereomicrocsope. 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, dilution: 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), rat anti-DE Cadherin (DCAD2) (DSHB, AB_528120, 1/20), rabbit anti-pMad (Cell Signaling, Phospho-SMAD1/5 (Ser463/465) mAb #9516, 1/200,), and rabbit anti-cleaved caspase-3 (Cell Signaling, Cleaved Caspase-3 (Asp175) #9661, 1/200). 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). Rhodamine phalloidin (ThermoFisher, R415, 1/100) was used to stain F-Actin.

Germline clonal analysis

We generated sakura^null mutant clones using FLP/FRT-mediated recombination (Rubin and Huynh 2015). To induce sakura^null GSC clones, 3-day-old female flies of the genotype hs-flp/w; +; 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 of the genotype hs-flp/w; +; FRT82B, ubi-GFP/FRT82B were used as controls. Ovaries were dissected and stained 4, 7, and 14 days after clone induction. To induce primordial germ cell (PGC) clones, early third-instar larvae were subjected to the same heat shock regime, then allowed to develop to adulthood. Ovaries were dissected and analyzed at 3 to 4 days after eclosion.

To induce sakura^null mutant clones in the presence of otu-EGFP or otu(ΔTudor)-EGFP transgenes, 3-day-old female flies with the genotype hs-flp/w; otu-EGFP/+; FRT82B, ubi-RFP/FRT82B, sakura^null and hs-flp/w; otu(i1Tudor)-EGFP /+; FRT82B, ubi-RFP/FRT82B, sakura^null were heat-shocked under the same conditions. Control genotypes were hs-flp/w; otu-EGFP/+; FRT82B, ubi-RFP/FRT82B and hs-flp/w; otu(i1Tudor)-EGFP/+; FRT82B, ubi-RFP/FRT82B were used as controls. Flies were dissected 3-4 days after clone induction.

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)) (Kandasamy et al. 2017; Zhu et al. 2018b). 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). 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 described (Zhu and Fukunaga 2021).

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 described (Fukunaga et al. 2012). Small RNA libraries and poly-A+ mRNA libraries were prepared, sequenced on Hiseq2500 (Illumina), and analyzed, as previously reported (Fukunaga et al. 2014; Kandasamy et al. 2017; Liao et al. 2018; Zhu et al. 2018a; Zhu et al. 2018b; Liao et al. 2019; Zhu et al. 2019a; Zhu et al. 2019b). SRA accession number for these datasets is PRJNA1156618.

RT-PCR

Total RNAs from testes and ovaries were prepared using miRVana (Thermo Fisher Scientific). RNAs were treated with Turbo DNase (Thermo Fisher Scientific) to remove potential genomic DNA contamination. A total of 1 μg of RNA was reverse-transcribed into cDNA using SuperScript^TM VILO^TM MasterMix (Thermo Fisher Scientific). To assess sxl alternative splicing, PCR was performed using GoTaq Green Master Mix (Promega) with the primers sxl-F (CTCACCTTCGATCGAGGGTGTA) and sxl-R (GATGGCAGAGAATGGGAC).

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.

Supplementary figures

Fig S1.

Fig S2.

Fig S3.

Fig S4.

Fig S5.

Fig S6.

Fig S7.

Fig S8.

Fig S9.

Fig S10.

Fig S11.

Fig S12.

Fig S13.

Fig S14.

Fig S15.

Fig S16.

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 R03AI178064] 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.