Derepression of Y-linked multicopy protamine-like genes interferes with sperm nuclear compaction in D. melanogaster

Significance Protamines are small, highly positively charged proteins that are required for packaging DNA to produce mature sperm with highly condensed nuclei capable of fertilization. Even small changes in the dosage of protamines in humans is associated with infertility. Our work reveals the presence of dominant-negative protamine genes on the Y chromosome of Drosophila melanogaster and shows that the precise expression of functional protamines and repression of dominant-negative protamines is a critical process to ensure male fertility.

Across species, sperm maturation involves the dramatic reconfiguration of chromatin into highly compact nuclei that enhance hydrodynamic ability and ensure paternal genomic integrity. This process is mediated by the replacement of histones by sperm nuclear basic proteins, also referred to as protamines. In humans, a carefully balanced dosage between two known protamine genes is required for optimal fertility. However, it remains unknown how their proper balance is regulated and how defects in balance may lead to compromised fertility. Here, we show that a nucleolar protein, modulo, a homolog of nucleolin, mediates the histone-to-protamine transition during Drosophila spermatogenesis. We find that modulo mutants display nuclear compaction defects during late spermatogenesis due to decreased expression of autosomal protamine genes (including Mst77F) and derepression of Y-linked multicopy Mst77F homologs (Mst77Y), leading to the mutant's known sterility. Overexpression of Mst77Y in a wild-type background is sufficient to cause nuclear compaction defects, similar to modulo mutant, indicating that Mst77Y is a dominant-negative variant interfering with the process of histone-to-protamine transition. Interestingly, ectopic overexpression of Mst77Y caused decompaction of X-bearing spermatids nuclei more frequently than Y-bearing spermatid nuclei, although this did not greatly affect the sex ratio of offspring. We further show that modulo regulates these protamine genes at the step of transcript polyadenylation. We conclude that the regulation of protamines mediated by modulo, ensuring the expression of functional ones while repressing dominant-negative ones, is critical for male fertility.

protamine | spermatogenesis | Drosophila
In many species, spermatids undergo the process of nuclear compaction, an essential process to produce sperm that are capable of fertilization (1)(2)(3). Nuclear compaction is critical for the sperm's hydrodynamic performance and protecting the paternal genome against mutagens (4)(5)(6). Nuclear compaction involves the dramatic chromatin reorganization mediated by the histone-to-protamine transition (1-5, 7, 8). Sperm nuclear basic proteins, also referred to as protamines, are small, positively charged proteins that replace histone-based nucleosomes to achieve the extreme degree of DNA compaction often seen in sperm (2). As such, these protamines are required for fertility across many different species (4).
Although protamines are essential for fertility, they are rapidly evolving across species (4,9,10), where the primary sequence, the number, and the functionality of protamine genes are not well conserved. For example, human and mouse protamine genes, PRM1 and PRM2, are required for fertility (4,6,7), while PRM2 has become nonfunctional in bulls and boars (4,11). Closely related Drosophila species have independently evolved many different protamine-like genes (10): Drosophila melanogaster has Mst35Ba and Mst35Bb (also known as ProtA and ProtB), which are the most similar to mammalian PRM1 and PRM2 (3,12,13), as well as Mst77F, Prtl99C, and Y-linked multicopy Mst77Y, with evidence that several more uncharacterized genes may also be involved (14). In contrast, in Drosophila simulans, there is just one orthologous copy of the ProtA/B gene (Prot) as well as one ortholog each for Mst77F (GD12157) and Prtl99c (GD21472). D. simulans lacks Mst77Y (10,14), but have evolved their own clade-specific genes that contain large regions of protamine sequences (Dox family genes), which are not present in D. melanogaster (15,16). Surprisingly, while ProtA and ProtB are most similar to their mammalian counterparts, they are not required for fertility in D. melanogaster (12); instead, more divergent genes Mst77F and Prtl99C are essential (17)(18)(19). The potential function of the D. melanogaster-specific multicopy locus of Mst77F homologs (the Mst77Y genes) is unknown (20,21).
Interestingly, it has been observed that mammals appear to feature their own species-specific ratios of protamine dosage (2,11,22,23), and in humans, even small alterations in the ratio of PRM1 and PRM2 are associated with infertility (2, 23-26), suggesting that a specific balance of protamines is important for sperm DNA packaging. However, it remains unknown why carefully balanced protamine expression is important and how it is achieved to support fertility.
While studying D. melanogaster modulo mutants, we discovered that modulo transheterozygotic mutant causes misregulation of protamine genes. modulo mutant spermatids display decreased nuclear incorporation of protamine-like protein Mst77F and increased incorporation of its Y-linked homolog, Mst77Y, which is barely incorporated in the wild type, leading to a DNA compaction defect that explains the reported sterility of modulo mutant. Our data indicate that Mst77Y likely acts as a dominant-negative form of Mst77F, interfering with the process of histone-to-protamine transition. Interestingly, Mst77Y has disproportionate effects on spermatids carrying an X chromosome, leading to biased decompaction of X-bearing spermatid nuclei, although it does not lead to large effects on the sex ratio of offspring. We further find that modulo is involved in safeguarding polyadenylation of Mst77F transcript over that of the Y-linked Mst77Y. Our study reveals a mechanism of protamine gene expression mediated by modulo, balancing the correct ratio of protamine gene expression to ensure male fertility.

modulo Mutant Is Defective in Sperm Nuclear Compaction.
Modulo is the Drosophila homolog of Nucleolin, a nucleolar protein implicated in RNA processing (27,28). Although modulo-mutant males have been known to be sterile (27,29), the cytological defects that lead to their sterility have not been characterized. We find that the modulo transheterozygote mutant (mod L8 /mod 07570 ) exhibits defects in nuclear morphology transformation during late spermiogenesis. In wild-type males, postmeiotic spermatid nuclei undergo well-documented morphological changes (1), from round spermatid stage, to "leaf " stage, to "canoe" stage, resulting in highly condensed "needle-" stage nuclei, which is accompanied by the histone-to-protamine transition (Fig. 1A). Although modulomutant germ cells proceeded through spermatogenesis normally, including early nuclear compaction ( Fig. 1 B and C), the modulo mutant exhibited striking "decompaction" of the nuclei after reaching the canoe stage, coinciding with the individualization of spermatids ( Fig. 1 D and E). Immunofluorescence (IF) staining using anti-dsDNA, which has been previously used to assess the compaction status of spermatid nuclei (30), revealed that defective spermatid nuclei of modulo mutant are indeed decompacted (Fig. 1 F and G). Decompacting nuclei are initially negative via Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL), a method used to identify DNA breaks that occur during apoptosis ( Fig. 1 H and I), then become TUNEL positive (Fig. 1J), suggesting that decompaction is not the result of cell death, but may rather be a cause. Overall, 100% of the modulo-mutant testes exhibited a nuclear decompaction phenotype (Fig. 1K), and it appeared that all nuclei eventually become decompacted and die, filling the distal end of the testis with cellular debris (SI Appendix, Fig. S1 A and B). Because nuclear decompaction in the modulo mutant occurs at stages when sperm chromatin is known to undergo reorganization through the histone-to-protamine transition, we explored whether the modulo mutant is defective in this process. Histone-toprotamine transition occurs step wise: 1) histone modification and removal, 2) transition protein incorporation then removal, and 3) protamine incorporation (1). IF staining revealed that modulo-mutant spermatids undergo proper histone removal and transient transition protein incorporation (SI Appendix, Fig. S2 A-F), but fail to properly accumulate ProtA/B and Mst77F (Fig. 2 A  and B). Moreover, using a specific antibody (SI Appendix, Fig. S3 A and B), we found that Mst77Y, Y-linked multicopy homologs of Mst77F (20, 21) (SI Appendix, Fig. S4A), strongly accumulated in modulo-mutant spermatid nuclei, whereas it was barely detectable in control (Fig. 2 C-G), suggesting that Mst77Y is aberrantly expressed in the modulo mutant. As the deletion of Mst77F and ProtA/B does not cause nuclear decompaction as severe as that of the modulo mutant (18), we infer that the incorporation of Mst77Y (in addition to the depletion of Mst77F and ProtA/B) causes the observed catastrophic nuclear decompaction seen in the modulomutant spermatids.

Ectopic Expression of Mst77Y Alone Is Sufficient to Cause Nuclear
Decompaction. The Mst77Y genes have several interesting features. First, the gene locus contains 18 copies of Mst77F homolog (SI Appendix, Fig. S4 A and B), which originated from a single event of Mst77F translocation to the Y chromosome, followed by gene amplification (20,21). Second, many of the Mst77Y genes have mutations, which have resulted in changes in the position and number of critical arginine, lysine, and cysteine residues believed to be important for protamine function (4,31). Other mutations have resulted in premature truncations (SI Appendix, Fig. S4B) (21). Note that anti-Mst77Y antibody was generated by using multiple peptides from Mst77Y that are distinct from Mst77F to increase specificity. The antibodies were also designed to be able to identify all copies of Mst77Y, which feature similar mutations and were tested to be able to identify both full-length Mst77Y (Mst77Y-12) and normally truncated Mst77Y (Mst77Y-3) (SI Appendix, Figs. S4 B and S5 A-C). Because Mst77Y's mutations likely alter Mst77F's normal function, we hypothesized that Mst77Y genes may function as a dominant-negative form of Mst77F. Accordingly, Mst77Y's aberrantly high expression in the modulo mutant may interfere with the process of normal histoneto-protamine transition.
To test the possibility that the expression of Mst77Y causes the nuclear decompaction phenotype, we generated transgenic lines that express Mst77Y under a male germline-specific tubulin promoter (β2-tubulin promoter) (32)(33)(34). From the 18 copies of Mst77Y homologs present on the Y chromosome (20, 21) we generated two lines expressing either Mst77Y-12 (a full-length version) or Mst77Y-3 (a truncated version due to premature stop codon) (SI Appendix, Figs. S4B and S6), as the transcripts of these two genes have been previously detected by qRT-PCR (21). Strikingly, expression of either Mst77Y-3 or Mst77Y-12 recapitulated a nuclear decompaction phenotype similar to that seen in modulo mutant ( Fig. 3 A-D): 45.7% and 43.2% of testes examined exhibited nuclear decompaction upon expression of Mst77Y-3 or Mst77Y-12, respectively (Fig. 3E), suggesting that high Mst77Y expression is sufficient to cause nuclear decompaction in a subset of spermatids. Notably, in contrast to the eventual decompaction of all spermatids seen in the modulo mutant, Mst77Y overexpression alone does not cause sterility. We speculate that this might be due to the added effect of the decreased incorporation of Mst77F and ProtA/B, in addition to high Mst77Y incorporation, seen in the modulo mutant.
Given that Barckmann et al. utilized the same promoter to overexpress autosomal Mst77F and did not observe such nuclear compaction defects during spermiogenesis (32) as we observed with Mst77Y overexpression, we infer that Mst77Y may act as a dominant-negative form, perhaps interfering with the function of Mst77F (Discussion). This notion is further supported by the fact that a truncated version (Mst77Y-3) also causes the decompaction phenotype. Indeed, spermatid cysts of transgenic males expressing Mst77Y-3 exhibited uneven Mst77F staining, suggesting that some nuclei fail to accomplish proper Mst77F incorporation (SI Appendix, Fig. S5 D and E). It is important to note that the nuclear decompaction was most prominently observed when males were raised in 25 °C after their parents were raised at 18 °C (Methods). Interestingly, using DNA Fluorescence in situ hybridization (FISH) to distinguish X-vs. Y-bearing spermatids, we found that overexpression of Mst77Y results in biased demise of X-bearing spermatids, where 61.8% of decompacting nuclei were X-bearing, compared to only 38.2% being Y-bearing (Fig. 3 F and G). It is important to note that this bias is not due to differential efficiency of hybridization of X chromosome vs. Y chromosome DNA FISH probes: Leaf to canoe stage spermatids of control males (SI Appendix, Fig. S7 A and B), as well as leaf to canoe stage spermatids of modulo-mutant males (before they exhibit decompaction defects), exhibited ~50:50 ratio of X:Y signal (SI Appendix, Fig. S7 C and D), further suggesting that decompaction is biased toward X-bearing spermatids. However, a fertility assay revealed only a minor increase in the male progeny compared to sex chromosome-matched controls (51.8% vs. 47.8%, P = 0.0005) (SI Appendix, Fig. S8A). Likewise, only a small degree of sex ratio distortion was observed in modulo heterozygous mutant, compared to sex chromosomematched control (SI Appendix, Fig. S8B) (Discussion).
Together, these results suggest that Mst77Y acts as a dominant-negative form of Mst77F, interfering with the incorporation of normal protamines and leading to spermatid nuclear decompaction.

Modulo Promotes Polyadenylation of Autosomal Mst77F
Transcript. How does modulo regulate the expression of Mst77F and Mst77Y? Modulo protein is expressed in the nucleolus of spermatogonia and spermatocytes, but is down-regulated prior to the meiotic divisions ( Fig. 4 A and B), days earlier than the stages in which its mutant phenotype manifests. Protamine genes are known to be transcribed many days prior to the sperm nuclear compaction process in both flies and mice (3,32,35). Interestingly, we found that Mst77F transcripts colocalize with Modulo in the spermatocyte nucleolus, while Mst77Y transcripts localize adjacent to the nucleolus (Fig. 4C). These results prompted us to examine whether Mst77F and/or Mst77Y transcripts may be deregulated in modulo mutant. Indeed, we found that Mst77F messenger RNA (mRNA) is dramatically reduced in modulo mutant, whereas Mst77Y mRNA was increased approximately threefold using qRT-PCR of polyA-selected RNA (Fig. 4D). However, when total RNA was used for qRT-PCR or total RNA sequencing, we found that both Mst77F and Mst77Y transcripts were increased in modulo mutant ( Fig. 4D and SI Appendix, Fig. S9A). RNA FISH, which visualizes total RNA, also confirmed the increase of both    Fig. S10). Collectively, these results suggest that Modulo specifically regulates transcripts of Mst77F and Mst77Y at the step of polyadenylation. Given that Modulo protein and Mst77F transcript colocalize in the nucleolus, we speculate that Mst77F is directly regulated by Modulo, whereas increased mRNA level of Mst77Y may be an indirect consequence of reduced functional Mst77F mRNA. Interestingly, RNA FISH using poly(T) probes revealed that poly(A) signal encircles the nucleolus in wild-type spermatocytes, whereas markedly less poly(A) was detected around the nucleolus in the modulo mutant ( Fig. 4 E and F), further suggesting that modulo may function to facilitate polyadenylation of transcripts localized to the nucleolus.
Our findings are consistent with the known importance of polyadenylation to sperm-specific transcripts, such as protamines, which must be translationally repressed for long periods (36)(37)(38)(39). Taken together, these results suggest that modulo plays an essential role in sperm nuclear compaction by facilitating maturation of canonical Mst77F transcript over that of Y-linked Mst77Y (Fig. 4G).

Discussion
The present study reveals a regulatory mechanism mediated by a nucleolar protein Modulo that balances the expression of protamine subtypes in D. melanogaster. This finding may represent a similar theme to what is seen in the fragile balance of PRM1 and PRM2 in mammalian fertility (2,7,24,25 Table S1. multicopy Mst77F homologs, our study suggests that they have the ability to act as dominant-negative protamines and thus must be carefully regulated/repressed. The present study also confirmed that Mst77Y genes are expressed as proteins as suggested previously by the finding that several of the copies contain complete open-reading frames (21) and is also consistent with small RNA sequencing revealing that the Mst77Y locus is not a source of small RNAs (40). We showed that overexpression of Mst77Y dominantly interferes with Mst77F incorporation, leading to decompaction of sperm nuclei and their demise. Mst77Y genes feature differences from their autosomal homolog that further support the idea that they are dominant-negative versions of Mst77F and interfere with sperm chromatin compaction. Mst77Y-12, which retains the full ORF of Mst77F (SI Appendix, Fig. S4), exhibits 87% overall sequence homology to autosomal Mst77F. At the domain/motif level, the MST-HMG-box domain, suggested to be important for DNA binding (14), maintains 100% homology, while the coiled-coil motif and C-terminal domain maintain only ~79.5% and ~85% homology, respectively (SI Appendix, Fig. S6B). It has been shown that the N-terminal domain of Mst77F, which contains the coiled-coil motif, interacts with the C-terminal domain to induce multimerization to mediate DNA compaction (41). The changes to Mst77Y at important regions may thus influence the multimerization of protamines and the formation of proper sperm chromatin structure, by interfering with the ability of the canonical version to multimerize. The notion that Mst77Y behaves as a dominant-negative version of Mst77F is further supported by the fact that overexpression of Mst77Y-3, a truncated version which does not contain the C-terminal domain (SI Appendix, Fig. S6B), is still sufficient to cause defects in nuclear compaction (Fig. 3C).
What is the potential "function" of dominant-negative protamines? We propose a few nonmutually exclusive possibilities. First, dominant-negative protamines may participate in meiotic drive, as suggested by recent works in D. simulans (15,16) as well as D. melanogaster (10). Indeed, our data suggest that Mst77Y has the ability to disproportionally affect X-bearing spermatids. While this did not result in a large sex ratio distortion in offspring (SI Appendix, Fig. S8), this ability to harm a subset of developing spermatids during postmeiotic development may indicate the possibility that these protamine variants could be exploited by a meiotic drive system to unleash its own selfish purpose. Intriguingly, studies on the Winters sex-ratio meiotic drive system in D. simulans revealed that the driver, Dox, contains a large portion of the Protamine gene (15,16). While it has not been confirmed whether this protamine-like region is essential for sex ratio distortion, the derepression of Dox does seem to cause nuclear defects during spermiogenesis (42). We propose that a drive system that would be able to localize a dominant-negative protamine such as Mst77Y to a subset of spermatids containing one homolog over another could be quite successful at achieving bias. Alternatively, the dominant-negative version of a protamine may be utilized when spermatogenesis needs to be aborted (similar to the concept of "programmed cell death"), for example under stress conditions. In such a case, dominant-negative protamines (such as Mst77Y) can be up-regulated to lead to abortive spermatogenesis. In such a scenario, a dominant-negative protamine may have a beneficial function for the organism. Yet another possibility that may contribute toward the rapid divergence of protamines is that protamine genes evolve to optimally package the genome, which may be greatly influenced by the composition of the most abundant sequences in a given genome, i.e., repetitive DNA such as satellite DNA. As these repetitive sequences are known to rapidly diverge across species (43), protamine genes may have to adapt to accommodate diverged repetitive DNA sequences, leading to rapid divergence and/or emergence of multiple protamine genes to optimally package different repetitive DNA with distinct structure/sequence. In such a scenario,  Table S1.
fine-tuning the expression of different protamine genes may be critical. Additionally, if any protamine genes have evolved to optimally package certain satellite DNA, conversion of such protamine into a dominant-negative version can immediately target the chromosome that harbors the given satellite DNA, leading to meiotic drive that selectively harms the specific chromosome. This possibility is intriguing as the Segregation Distorter (SD) meiotic drive system in D. melanogaster is known to target Responder satellite DNA repeats (44)(45)(46) and exhibits sperm nuclei decompaction similar to what is observed in this study (30). The possibility that dominantnegative protamines are involved in the decompaction of spermatid nuclei in the SD drive system remains to be studied. Taken together, our study identified a mechanism by which various protamine variants are coordinately regulated at the posttranscriptional level, possibly to achieve balanced expression of multiple protamine variants. A similar mechanism may be at play to fine-tune the expression levels of protamine variants in human and mouse, disruption of which is associated with compromised fertility.
The Mst77Y transgenic flies were generated by phiC31 site-directed integration into the Drosophila genome. For UAS-Mst77Y-12, β2-tubulin-Mst77Y-3, and β2-tubulin-Mst77Y-12 transgenic lines, the Mst77Y overexpression sequences in D. melanogaster were synthesized by gene synthesis service from Thermo Fisher Scientific (GeneArt Gene Synthesis) and were cloned into pattB vector to insert into specific integration site on second chromosome (attP40) (SI Appendix, Fig. S5C and Table S2). All injection and selection of flies containing integrated transgene were performed by BestGene Inc. Because UAS-Mst77Y-12 transgene was injected to the same host fly strain as β2-tubulin-Mst77Y-3, and β2-tubulin-Mst77Y-12 transgenic lines, we used this (without gal4 driver) as a "background-matched control." Modulo-gfp strain was constructed using CRISPR-mediated knock-in of a Green fluorescent protein (gfp)-tag at the C terminus of Modulo (Beijing Fungene Biotechnology Co.) (SI Appendix, Table S3). Sex Ratio Assay. Individual 1-d-old males raised for at least one generation at 18 °C were crossed with 3× 1-to 3-d-old virgin y w females at 25 °C. After 1 d, males were removed. This was done to maximize the proportion of males exhibiting decompaction phenotype described in Fig. 3. Females were left to produce embryos for a total of 5 d before cleared. Following the onset of eclosion, sex of offspring was scored for 10 consecutive days.
Fluorescently labeled probes were added to the hybridization buffer to a final concentration of 100 nM. Poly(T) probes for recognizing Poly(A) sequence were from Integrated DNA Technologies. Probes against Mst77F and Mst77Y were designed using the Stellaris1 RNA FISH Probe Designer (Biosearch Technologies, Inc.) available online at www.biosearchtech.com/stellarisdesigner. Each set of custom Stellaris1 RNA FISH probes was labeled with Quasar 670 or Quasar 570 (SI Appendix, Table S4).
Images were acquired using an upright Leica TCS SP8 confocal microscope with a 63× oil immersion objective lens (NA = 1.4) and processed using Adobe Photoshop and ImageJ software. DNA Fluorescence In Situ Hybridization. Testes from 1-to 3-d-old flies were rapidly dissected in 4% formaldehyde and 1mM Ethylenediaminetetraacetic acid (EDTA) in 1X PBS and then nutated for 30 min. Then, the testes were washed three times in 1X PBS containing 0.1% Triton-X (PBST) +1 mM EDTA for 30 min each. The testes were briefly rinsed with 1X PBST and then incubated at 37 °C for 10 min with 2 mg/mL Rnase A in PBST. Following Rnase treatment, samples were washed once in 1X PBST + 1 mM EDTA for 10 min. The samples were then briefly rinsed with 2X SSC + 1 mM EDTA + 0.1% Tween-20, and then washed three times in 2X SSC + 0.1% Tween-20 + formamide (20% for first wash, 40% for second, 50% for third) for 15 min each. The samples were then washed with 2X SSC + 0.1% Tween-20 + 50% formamide for 30 min. The samples were then incubated for 5 min at 95 °C with fluorescently labeled probes in hybridization buffer (2X SSC, 10% dextran sulfate, 50% formamide, 1 mM EDTA) and then transferred to 37 °C overnight. Following hybridization, the samples were washed three times in 2X SSC + 1 mM EDTA + 0.1% Tween-20 for 20 min each and then mounted in VECTASHIELD with DAPI (Vector Labs).
qRT-PCR. Total RNA was purified from D. melanogaster adult testes (75 pairs/sample) by Direct-zol RNA Miniprep (Zymo Research), with DNase treatment according to manufacturer's protocol. One microgram total RNA was reverse transcribed following priming with either random hexamers or polyT using SuperScript III® Reverse Transcriptase (Invitrogen) followed by qPCR using Power SYBR Green reagent (Applied Biosystems). Primers for qPCR were designed to amplify mRNA or intron-containing transcript as indicated. Relative expression levels were normalized to Rp49 and control siblings. All reactions were done in technical triplicates with at least two biological replicates. Graphical representation was inclusive of all replicates and P-values were calculated using a t test performed on untransformed average ddct values. Primers used are described in SI Appendix, Fig. S11 A and B. Total RNA-Seq. Total RNA was purified from D. melanogaster adult testes by Direct-zol RNA Miniprep (Zymo Research), with Dnase treatment. Quality of the indexed libraries was confirmed using an Agilent Fragment Analyzer and qPCR. Sequencing libraries were prepared with the KAPA RNA HyperPrep Kit with RiboErase. Samples were sequenced on a HiSeq 2500, producing 100 × 100 nt paired-end reads. The read pairs were mapped to the canonical chromosomes of the D. melanogaster genome (assembly BDGP6/dm6) using STAR 2.7.1a (48); default parameters, except "-alignIntronMax 25000," indexed with all FlyBase genes (FB2020_06 Dmel Release 6.37) and the option "-sjdbOverhang 100." Gene counts were obtained using featureCounts (49); v 2.0.1, with "-M -fraction -p -s 2." After summing gene counts for technical replicates, differential expression was assayed using DESeq2 v1. 26.0 (50), with lfcShrink(type="ashr")). RNA coverage across genes at nucleotide resolution was quantified with "bedtools coverage" (51) and scaled by the total number of reads mapped to genes. Statistics and Reproducibility. Data are presented as mean ± SD unless otherwise indicated. The sample number (n) indicates the number of testes, nuclei, or male flies in each experiment as specified in the figure legends. We utilized two-sided Student's t test to compare paired or independent samples, as applicable and is specified in the figure legends. We calculated probability using exact binomial distribution with parameters specified in Fig. 3G legend. No statistical methods were used to predetermine sample sizes. The experimenters were not blinded to the experimental conditions, and no randomization was performed.
All the statistical details of the experiments are provided in the main text and legends. P-values are listed either in figure, figure legends, or SI Appendix, Table S1. Data, Materials, and Software Availability. Sequencing data is available at National Center for Biotechnology Information Gene Expression Omnibus under accession GSE214456 (52). All other data are included in the manuscript and/ or SI Appendix.