SNPC-1.3 is a sex-specific transcription factor that drives male piRNA expression in C. elegans

Piwi-interacting RNAs (piRNAs) play essential roles in silencing repetitive elements to promote fertility in metazoans. Studies in worms, flies, and mammals reveal that piRNAs are expressed in a sex-specific manner. However, the mechanisms underlying this sex-specific regulation are unknown. Here we identify SNPC-1.3, a male germline-enriched variant of a conserved subunit of the small nuclear RNA-activating protein complex, as a male-specific piRNA transcription factor in Caenorhabditis elegans. SNPC-1.3 colocalizes with the core piRNA transcription factor, SNPC-4, in nuclear foci of the male germline. Binding of SNPC-1.3 at male piRNA loci drives spermatogenic piRNA transcription and requires SNPC-4. Loss of snpc-1.3 leads to depletion of male piRNAs and defects in male-dependent fertility. Furthermore, TRA-1, a master regulator of sex determination, binds to the snpc-1.3 promoter and represses its expression during oogenesis. Loss of TRA-1 targeting causes ectopic expression of snpc-1.3 and male piRNAs during oogenesis. Thus, sexually dimorphic regulation of snpc-1.3 expression coordinates male and female piRNA expression during germline development.


Introduction
Piwi-interacting RNAs (piRNAs), a distinct class of small noncoding RNAs, function to preserve germline integrity (Batista et al., 2008;Carmell et al., 2007;Cox et al., 1998;Deng and Lin, 2002;Kuramochi-Miyagawa et al., 2008;Lin and Spradling, 1997;Wang and Reinke, 2008). In Drosophila, mutation of any of the three Piwi genes (piwi, aub, ago3) results in rampant activation of transposons in the germline and severe defects in fertility (Brennecke et al., 2007;Harris and Macdonald, 2001;Lin and Spradling, 1997;Vagin et al., 2006). In Mus musculus, mutation of the Piwi protein MIWI leads to the misregulation of genes involved in germ cell development, defective gametogenesis, and sterility (Deng and Lin, 2002;Zhang et al., 2015b). Caenorhabditis elegans piRNAs can be inherited across multiple generations and trigger the transgenerational silencing of foreign elements such as transgenes. Disruption of this inheritance results in eventual germline collapse and sterility, known as the germline mortal phenotype (Ashe et al., 2012;Buckley et al., 2012;Shirayama et al., 2012). Taken together, piRNAs are essential to preserve germline integrity and ensure the reproductive capacity in metazoans.
piRNA biogenesis is strikingly diverse across organisms and tissue types. In the Drosophila germline, piRNA clusters are found within pericentromeric or telomeric heterochromatin enriched for H3K9me3 histone modifications. The HP1 homolog Rhino binds to H3K9me3 within most of these piRNA clusters and recruits Moonshiner, a paralog of the basal transcription factor TFIIA, which, in turn, recruits RNA polymerase II (Pol II) to enable transcription within heterochromatin (Andersen et al., 2017;Chen et al., 2016;Klattenhoff et al., 2009;Mohn et al., 2014;Pane et al., 2011). Two waves of piRNA expression occur in mouse testes: pre-pachytene piRNAs are expressed in early spermatogenesis and silence transposons, whereas pachytene piRNAs are expressed in the later stages of meiosis and have unknown functions. While the mechanisms of prepachytene piRNA transcription remain elusive, pachytene piRNAs require the transcription factor A-MYB, along with RNA Pol II (Li et al., 2013).
The vast majority of the~15,000 piRNAs in C. elegans are encoded within two large megabase genomic clusters on chromosome IV (Das et al., 2008;Ruby et al., 2006). Each piRNA locus encodes a discrete transcriptional unit that is individually transcribed as a short precursor by Pol II Cecere et al., 2012;Billi et al., 2013). Processing of precursors yields mature piR-NAs that are typically 21 nucleotides (nt) in length and strongly enriched for a 5 0 uracil (referred to as 21U-RNAs). Transcription of these piRNAs requires a conserved eight nt core motif (NNGTTTCA) within their promoters (Billi et al., 2013;Cecere et al., 2012;Ruby et al., 2006). piRNAs enriched during spermatogenesis are associated with a cytosine at the 5 0 most position of the core motif (CNGTTTCA); mutation of cytosine to adenine at this position results in ectopic expression of normally male-enriched piRNAs during oogenesis. In contrast, genomic loci expressing piRNAs enriched in the female germline show no discernable nucleotide bias at the 5 0 position (Billi et al., 2013). While differences in cis-regulatory sequences contribute to the sexually dimorphic nature of piRNA expression, sex-specific piRNA transcription factors that drive distinct subsets of piRNAs in the male and female germlines remain to be identified.
Here, we demonstrate that SNPC-1.3, an ortholog of human SNAPC1, is required specifically for male piRNA expression. Furthermore, TRA-1, a master regulator of sex determination, transcriptionally represses snpc-1.3 during oogenesis to restrict its expression to the male germline. Taken together, our study reports the first example of a sex-specific piRNA transcription factor that drives the expression of male-specific piRNAs.

Results
SNPC-4 is a component of the core piRNA transcription complex that drives all piRNA expression SNPC-4-specific foci are present in both male and female germ cell nuclei (Kasper et al., 2014), but the role of SNPC-4 in the male germline is not well understood. We hypothesized that SNPC-4 is required for piRNA biogenesis in both the male and female germline. To test this, we conditionally depleted the SNPC-4 protein using the auxin-inducible degradation system (Zhang et al., 2015a; Figure 1-figure supplement 1A). We added an auxin-inducible degron (AID) to the C-terminus of SNPC-4 using CRISPR/Cas9 genome engineering, and crossed this strain into worms expressing TIR1 under the germline promoter, sun-1. TIR1 is a plant-specific F-box protein that mediates the rapid degradation of C. elegans proteins tagged with an AID in the presence of the phytohormone auxin. Thus, addition of auxin to the snpc-4::aid; Psun-1::TIR1 strain is expected to degrade SNPC-4::AID, whereas strains with snpc-4::aid alone serve as a negative control; under these conditions, we examined a panel of spermatogenesis-and oogenesis-enriched piRNAs (Billi et al., 2013) during spermatogenesis and oogenesis. Unless otherwise stated, spermatogenesis and oogenesis stages will correspond to time points taken at 48 hr and 72 hr, respectively, post-L1 hatching at 20˚C. Worms depleted of SNPC-4 showed decreased expression of both spermatogenesis-and oogenesis-enriched piRNAs during spermatogenesis and oogenesis time points, respectively ( Figure 1A), confirming that SNPC-4 is a core piRNA transcription factor required for all piRNA expression.

SNPC-1.3 is enriched in the male germline
To determine whether SNPC-1.3 expression is restricted to the germline, we first examined snpc-1.3 mRNA levels during early spermatogenesis (36 hr post-L1 hatching) in worms fed glp-1 RNAi, which abrogates germline development. The mRNA of snpc-4, which is highly expressed in the germline (Kasper et al., 2014), was used as a control. Knockdown of glp-1 mRNA markedly reduced both snpc-1.3 and snpc-4 mRNAs. SNPC-1.3::3xFlag protein expression was also reduced in a glp-4 temperature-sensitive mutant, which fails to develop fully expanded germlines at 25˚C (Beanan and Strome, 1992), suggesting that SNPC-1.3 is predominantly expressed in the germline ( Figure 2A).

SNPC-1.3 is required for transcription of male piRNAs
Given the prominent interaction between SNPC-1.3 and SNPC-4 in the male germline ( Figure 1D), we hypothesized that SNPC-1.3 might be required for piRNA expression during spermatogenesis. To test this hypothesis, we generated a snpc-1.3 null allele by introducing mutations that result in a premature stop codon located eight amino acids away from the start codon at the snpc-1.3 locus. We examined spermatogenesis in hermaphrodites and him-8(-) males and examined oogenesis in adult hermaphrodites and fem-1(-) females. As a control, we analyzed the loss-of-function mutant of the C. elegans Piwi protein, prg-1(-), which almost completely lacked male and female piRNAs ( Figure 3A), as expected. Levels of male piRNAs were dramatically reduced in snpc-1.3(-) hermaphrodites during spermatogenesis and in him-8(-); snpc-1.3(-) males, whereas female piRNAs were largely unaltered in snpc-1.3(-) adult hermaphrodites and in fem-1(-); snpc-1.3(-) females ( Figure 3A,B). Unexpectedly, female piRNAs were also moderately upregulated by at least twofold in snpc-1.3(-) mutants undergoing spermatogenesis and in him-8(-); snpc-1.3(-) males. These findings suggest that, in addition to activating male piRNAs, SNPC-1.3 suppresses the expression of female piRNAs in the male germline, possibly by preferentially recruiting core factors such as SNPC-4 to male piRNA loci. As SNPC-4 is known to activate transcription of snRNAs as well as piRNAs (Kasper et al., 2014), we asked whether SNPC-1.3 is also required for transcribing snRNAs. To test this, we measured U1 snRNA levels in hermaphrodite adults after RNAi-mediated knockdown of snpc-1.3. In contrast to the reduction of U1 observed in snpc-4 RNAi, U1 levels were not significantly altered when snpc-1.3 was depleted (Figure 3-figure supplement 2A), suggesting that, unlike SNPC-4, SNPC-1.3 is likely specific to the transcription of male piR-NAs and does not play a role in snRNA transcription.
We next analyzed the genomic loci of male piRNAs and snpc-1.3-dependent piRNAs. As expected, the intersection of these two piRNA subsets displayed strong enrichment for the eight nt core motif and the 5 0 -most position of this core motif was enriched for cytosine (CNGTTTCA. Figure 3E; Figure 3-figure supplement 2E). In contrast, the core motif found upstream of female piRNAs upregulated upon loss of snpc-1.3 displayed a much weaker bias for the 5 0 cytosine ( Figure 3F). These observations validate our previous findings that male and female core motifs are distinct (Billi et al., 2013). Taken together, these data indicate that SNPC-1.3 is required for male piRNA expression.

SNPC-1.3 binds male piRNA loci in a SNPC-4-dependent manner
Given that SNPC-1.3 interacts with SNPC-4 and is required for expression of male piRNAs, we hypothesized that SNPC-1.3 might bind male piRNA loci in association with SNPC-4. To test this, we performed ChIP-qPCR to investigate SNPC-1.3 occupancy at regions of high piRNA density within the two large piRNA clusters on chromosome IV; an intergenic region lacking piRNAs served as a control. To determine whether SNPC-1.3 binding was dependent on SNPC-4, we again used the auxin-inducible degradation system to deplete SNPC-4 in the snpc-1.3::3xflag strain for 4 hr prior to our spermatogenesis time point. In the presence of SNPC-4 expression, SNPC-1.3 was enriched at both piRNA clusters, albeit to a lesser degree at the small cluster, and this enrichment was lost upon SNPC-4 depletion ( Figure 4A, Figure 4-figure supplement 1A). These data indicate that SNPC-1.3 binds piRNA loci during spermatogenesis in a SNPC-4-dependent manner in vivo.
To examine the genome-wide binding profile of SNPC-1.3 and its dependency on SNPC-4, we performed ChIP-seq of N2, snpc-1.3::3xflag, and snpc-1.3::3xflag; snpc-4::aid; Psun-1::TIR1 worms (Left) snpc-1.3 mRNA expression is dramatically enriched in him-8(-) males over fem-1(-) females during spermatogenesis, whereas snpc-4 mRNA expression shows no specific enrichment. eft-2 was used for normalization. Error bars: ± SD of two technical replicates. (Right) Western blot and quantification of SNPC-1.3::3xFlag in him-8(-) and fem-1(-). Error bars: ± SD of two biological replicates. g-Tubulin was used as the loading control. (C) SNPC-1.3 colocalizes with SNPC-4 in the male germline. Dissected adult male (top) and hermaphrodite (bottom) germlines stained for DNA, SNPC-4::3xFlag (magenta) and SNPC-1.3::Ollas (green) in a N2 background. Yellow insets: transition zone. Blue insets: pachytene. Representative image of three biological replicates is shown (male, n = 21, 18, 15 and hermaphrodite, n = 18, 10, 10). Scale bar, 25 mm. The online version of this article includes the following source data for figure 2: Source data 1. Source data for Figure 2A. Source data 2. Source data for Figure 2A. Source data 3. Source data for Figure 2B. Source data 4. Source data for Figure 2B   To determine whether SNPC-1.3 preferentially binds male piRNA loci, we characterized the SNPC-1.3 signal around individual 5 0 nucleotides of mature piRNAs. Again, we classified piRNAs as male, female, or not significantly enriched (NE) in either sex, based on our small RNA-seq analysis in wild-type hermaphrodites during spermatogenesis and oogenesis ( Figure 3C). SNPC-1.3 binding at male piRNA loci was most enriched just upstream of the piRNA 5 0 nucleotide, which overlaps the conserved core motif ( Figure  TRA-1 represses snpc-1.3 and male piRNA expression during oogenesis As male piRNA expression and SNPC-1.3 protein expression are largely restricted to the male germline, we asked how snpc-1.3 mRNA expression is regulated across development. C. elegans hermaphrodites produce sperm during the L4 stage and transition to producing oocytes as adults. To understand the mRNA expression profile of snpc-1.3 relative to snpc-4 and other developmentally regulated genes, we performed qRT-PCR across hermaphrodite development. snpc-4 mRNA is expressed at low levels during spermatogenesis, but dramatically increases during oogenesis ( Figures 5A and 1D; Figure 1-figure supplement 1H). These data suggest that low levels of SNPC-4 are sufficient for activating male piRNA biogenesis during spermatogenesis. Consistent with SNPC-1.3 protein expression ( Figure 1D), snpc-1.3 mRNA levels peak in L3 to early L4 stages, Source data 1. Source data for Figure 3A. Source data 2. Source data for Figure 3B.   during spermatogenesis ( Figure 5A). Given that snpc-1.3 expression across development is regulated at the mRNA level, we examined the sequences upstream of the snpc-1.3 coding region to identify potential cis-regulatory motifs. Less than 200 bp upstream of the snpc-1.3 start codon, we identified three consensus binding sites for TRA-1 ( Figure 5B), a transcription factor that controls the transition from spermatogenesis to oogenesis (Berkseth et al., 2013;Clarke and Berg, 1998;Zarkower and Hodgkin, 1993).
Given that snpc-1.3 is robustly de-repressed during oogenesis in TRA-1 binding site mutants, we hypothesized that male piRNAs would also be ectopically upregulated during oogenesis. To test this, we performed small RNA-seq and compared piRNA levels in wild-type and snpc-1.3 (2xtbs) worms during oogenesis (Supplementary file 3). Using a 1.2-fold threshold and FDR of 0.05, we observed 1,370 piRNAs in snpc-1.3 (2xtbs) mutants that were upregulated compared to wild type ( Figure 5D). The majority of these upregulated piRNAs overlap with the male piRNAs that we identified in wild-type hermaphrodites ( Figure 5D). We also confirmed this result by Taqman qPCR analysis, which showed that male piRNAs were significantly upregulated in snpc-1.3 (2xtbs) and snpc-1.3 (3xtbs) mutants compared to wild type during oogenesis ( Figure 5E). Taken together, these data suggest that TRA-1 directly binds to tbs sites in the snpc-1.3 promoter to repress its transcription and consequently, male piRNA expression during oogenesis.
Our data showed that female piRNAs are inappropriately upregulated during spermatogenesis upon loss of snpc-1.3 ( Figure 3A). Consistent with this result, female piRNAs show reduced expression during oogenesis upon upregulation of SNPC-1.3 expression in snpc-1.3 (2xtbs) and snpc-1.3 (3xtbs) mutants compared to wild type ( Figure 5E). We posit that SNPC-1.3 plays a direct role in Figure 4 continued is enriched at piRNA clusters on chromosome IV. SNPC-1.3-bound regions are enriched within piRNA clusters compared to regions outside of the piRNA clusters on chromosome IV (****p 0.0001, Wilcoxon rank sum test). The number of bins analyzed is listed in parentheses. (D) SNPC-1.3 enrichment at piRNA clusters is dependent on SNPC-4. SNPC-1.3-bound regions within piRNA clusters are depleted compared to regions outside of the piRNA clusters on chromosome IV upon loss of SNPC-4 (****p 0.0001, Wilcoxon rank sum test). The number of bins analyzed is listed in parentheses. (E) Distribution of SNPC-1.3 reads (mean density ± standard error) around the 5 0 nucleotide of mature piRNAs at the piRNA clusters. To resolve SNPC-1.3 binding between male and female piRNAs despite the high density of piRNAs, we selected 1 kb bins with all male (100), female (19), or non-enriched (279) piRNAs. Heat maps represent ChIP signal in 1 kb bins around the 5 0 nucleotide of all 100 mature male piRNAs, ranked according to SNPC-1.3 signal. (F) Examples of SNPC-1.3 binding at two regions containing two male piRNA loci. Regions are anchored on the 5 0 nucleotide of each mature male piRNA and show mean read density ± standard error. The online version of this article includes the following source data and figure supplement(s) for figure 4: Source data 1. Source data for Figure 4A.
To investigate the cause of snpc-1.3-dependent loss of male fertility, we examined spermiogenesis and sperm morphology in snpc-1.3(-) males. After meiotic differentiation in the male germline, male spermatids are induced by ejaculation and undergo spermiogenesis, a process that converts immature spermatids to motile sperm with a functioning pseudopod. Spermiogenesis can be induced in vitro by isolating spermatids directly from males and treating them with pronase (Shakes and Ward, 1989). Males lacking prg-1 still generate differentiated spermatids, but rarely produce normal pseudopodia upon activation ( Figure 6C,D; Wang and Reinke, 2008). Similar to prg-1(-) mutants, snpc-1.3(-) spermatids were rarely able to form normal pseudopodia. In contrast, snpc-1.3 (3xtbs) sperm formed normal pseudopodia at a frequency similar to wild type ( Figure 6C, D). In addition, many of the snpc-1.3(-) spermatids resembled sperm undergoing intermediate stages of spermiogenesis. Spermiogenesis, in vivo, starts off with spherical spermatids that enter into an intermediate stage characterized by the growth of spiky protrusions. This stage is then followed by fusion of the spiky protrusions into a motile pseudopod ( Figure 6E). To understand the dynamics of snpc-1.3(-) sperm progression through spermiogenesis, we treated spermatids with pronase and observed each activated spermatid over time. Wild-type spermatids spent an average of 6.2 min ± 4.5 min in the intermediate state before polarization and pseudopod development. In contrast, snpc-1.3(-) spermatids occupied the intermediate state for a significantly shorter period of time (2.9 min ± 3.7 min, p<0.05; Student's t-test) before forming pseudopods. By tracking each individual spermatid across spermiogenesis, we found most snpc-1.3(-) spermatids were unable to sustain Figure 5 continued specific mutations shown in red were made in one, two, or three of the TRA-1 binding sites (gray denotes the mutated motifs). (Bottom) TRA-1 binding is reduced in TRA-1 binding site mutants assayed by TRA-1 ChIP-seq. (C) TRA-1 represses snpc-1.3 mRNA expression during oogenesis. (Left) snpc-1.3 mRNA expression is drastically upregulated upon RNAi-mediated knockdown of tra-1 and (middle) in strains bearing mutations in two (2xtbs) or three (3xtbs) TRA-1 binding sites. Error bars indicate ± SD from two technical replicates. (Right) Western blot of SNCP-1.3::3xFlag expression driven under the wild-type and 2xtbs mutant promoter during spermatogenesis (spe.) (top) and oogenesis (oog.) (bottom). H3 was used as the loading control. (D) A subset of male piRNAs are ectopically expressed during oogenesis in snpc-1.3 (2xtbs) mutants. (Top) Volcano plot showing differential piRNA expression between snpc-1.3 (2xtbs) mutants versus wild type during oogenesis (oog.). piRNAs are colored by enrichment scores from Billi et al., 2013. (Bottom) Overlap of male piRNAs defined in Figure 3C with upregulated piRNAs in snpc-1.3 (2xtbs) mutants. (E) Mutations at two (2xtbs) or three (3xtbs) TRA-1 binding sites enhance male piRNA expression (top) but attenuate female piRNA expression (bottom) during oogenesis. Error bars indicate ± SD from two technical replicates. The online version of this article includes the following source data and figure supplement(s) for figure 5: Source data 1. Source data for Figure 5A. Source data 2. Source data for Figure 5C,E.  Figure 6 continued on next page pseudopod growth. While wild-type spermatids exhibited pseudopod growth and motility for an average of 24 min ± 10.35 min, snpc-1.3(-) spermatids sustained growth for a significantly shorter period of time (7.3 min ± 5.7 min, p<0.05; Student's t-test) before becoming immotile ( Figure 6E). These results indicate that spnc-1.3(-) males have defective spermatogenesis processes and exhibit similar fertility defects as prg-1(-) mutants.

Discussion
Our data indicate that C. elegans SNPC-1.3, a human SNAPC1 ortholog, functions as a male piRNA transcription factor. SNPC-1.3 interacts with SNPC-4 in foci in male germ cell nuclei ( Figure 2) and, by preferentially binding male piRNA promoters (Figure 4), is critical for their expression (Figure 3). SNPC-1.3 expression, reflecting the developmental profile of male piRNAs ( Figure 5-figure supplement 1A), is highest during spermatogenesis. We demonstrate that the snpc-1.3 locus itself is regulated by the sex determination regulator, TRA-1 ( Figure 5). During spermatogenesis, tra-1 expression is low, and snpc-1.3 and other male-promoting genes are licensed for expression. In contrast, tra-1 expression is upregulated during oogenesis and TRA-1 binds the snpc-1.3 promoter to repress its transcription, leading to the expression of female over male piRNAs (Figure 7). We propose that SNPC-1.3, via its interaction with SNPC-4, can direct the specificity of the core piRNA complex preferentially to male piRNA loci.
How is the expression of male and female piRNAs coordinated?
Given its role as a putative male piRNA transcription factor, we expected that deletion of snpc-1.3 would result in loss of male piRNAs with no consequences to the expression of female piRNAs. However, loss of snpc-1.3 also results in increased female piRNA expression during spermatogenesis (Figure 3), whereas ectopic overexpression of snpc-1.3 during oogenesis leads to decreased female piRNA levels ( Figure 5). Taken together, our findings suggest that transcription of male and female piRNAs is not completely separable from each other and that the balance in expression of the two piRNA subclasses may be dictated by the allocation of shared core transcription factors such as SNPC-4. Similar to multiple gene classes activated by general transcription factors (Levine et al., 2014), we speculate that male and female promoters compete for access to a limited pool of the core transcription complex, which includes SNPC-4, PRDE-1, TOFU-4, and TOFU-5 ( Figure 1). Therefore, we propose a model in which the expression and binding of SNPC-1.3 to core piRNA factors serves to 'sequester' the core complex away from female promoters. Mechanistically, we posit that the core piRNA transcription complex is specified to female promoters, and that only upon association with SNPC-1.3 is the core machinery directed to male promoters. We predict that when SNPC-1.3 is absent, more SNPC-4 and other previously identified cofactors are available to transcribe female piRNAs. Conversely, overexpression of SNPC-1.3 leads to the disproportionate recruitment of the core machinery to male promoters, leading to the indirect downregulation of female piRNAs. By controlling male piRNA expression, SNPC-1.3 is crucial for maintaining the balance between male and female piRNA levels across development. Source data 1. Source data for Figure 6A. Source data 2. Source data for Figure 6B. Source data 3. Source data for Figure 6E. While the default specification of the core complex to female promoters presents perhaps the most parsimonious explanation underlying male and female piRNA expression, we cannot exclude the possibility that an additional female-specific trans-acting factor may direct the core piRNA complex to female promoters. If true, we speculate that the developmental expression of such a factor (low during spermatogenesis and high during oogenesis), coupled with the developmental expression of SNPC-1.3, would coordinate the differential expression of male and female piRNAs. During spermatogenesis, SNPC-1.3 is more highly expressed such that the core machinery would primarily be directed to male promoters. In contrast, during oogenesis, SNPC-1.3 expression is low, concomitant with elevated expression of a female factor to license transcription of female piRNAs. This model, where both factors are present during both spermatogenesis and oogenesis, but in different  Figure 7. Model illustrating the dynamics of male and female piRNA transcription across C. elegans sexual development. In wild-type worms, male and female piRNA expression peaks during spermatogenesis and oogenesis, respectively. (Top) In snpc-1.3(-) mutants, male piRNA expression is abrogated, and female piRNA expression is moderately enhanced across sexual development relative to wild type. In TRA-1 binding site mutants, snpc-1.3 expression is de-repressed causing ectopic upregulation of male piRNAs and moderate repression of female piRNA expression during oogenesis relative to wild type. (Bottom) During spermatogenesis, SNPC-1.3 interacts with SNPC-4 at male promoters to drive male piRNA transcription. During oogenesis, TRA-1 represses the transcription of snpc-1.3 which results in the suppression of male piRNA transcription, thus leading to enhanced transcription of female piRNAs. ratios, would also be consistent with our piRNA expression analysis in snpc-1.3 loss-of-function and overexpression mutants.

The piRNA pathway co-opts snRNA biogenesis machinery
Our work adds to a growing body of evidence that snRNA machinery has been hijacked at multiple stages in C. elegans piRNA biogenesis, including transcription (Kasper et al., 2014;Weng et al., 2019) and termination (Beltran et al., 2019). Investigating potential parallels between snRNA and piRNA biogenesis may provide useful clues into the role of SNPC-1.3 in the piRNA complex.
The minimal snRNA SNAP complex consists of a 1:1:1 heterotrimer of the subunits SNAPC4, SNAPC1, and SNAPC3 in humans and SNAP190, SNAP43, and SNAP50 in flies Hung and Stumph, 2011;Li et al., 2004;Ma and Hernandez, 2002;Mittal et al., 1999; Figure 1). In vitro studies have shown that the trimer must assemble before the complex is able to bind DNA. Similarly, our data show SNPC-1.3 requires SNPC-4 to bind at the piRNA clusters (Figure 4), although we cannot formally rule out that loss of SNPC-4 only affects the stability of SNPC-1.3, rather than directly recruiting SNPC-1.3 to piRNA promoters. We speculate the piRNA complex is assembled in a similar fashion to the snRNA complex. Based on this model, we expect that SNPC-4 binding at male piRNA loci is abolished in a snpc-1.3 mutant. However, conclusive evidence that SNPC-4 binding at male piRNA promoters requires SNPC-1.3 is still lacking. Due to the highly clustered nature of C. elegans piRNAs, we anticipate that detecting differences in SNPC-4 binding between male and female piRNAs in snpc-1.3(-) mutants may not be possible with traditional ChIPseq methods, and may require application of higher resolution techniques.
Given that piRNAs have co-opted trans-acting factors from snRNA biogenesis (Kasper et al., 2014), it would not be surprising if piRNAs also co-evolved cis-regulatory elements for transcription factor binding from snRNA loci. Recently, Beltran et al., 2019 identified similarity between the 3 0 end of PSEs of snRNA promoters and the eight nt piRNA core motif in nematodes. In addition, Pol II and Pol III transcription from snRNA promoters share a common PSE, but are distinguished by the presence of other unique motifs (Hung and Stumph, 2011). Correspondingly, the canonical Type I and less abundant Type II piRNAs can be discriminated by the presence or absence of the eight nt core motif, respectively. Factors such as TOFU-4 and TOFU-5 function in both Type I and II piRNA expression, whereas PRDE-1 is only required for Type I piRNAs (Kasper et al., 2014;Weng et al., 2019). Altogether, these observations highlight the importance of cis-regulatory elements in specifying the expression of snRNAs and piRNA classes. In addition to enrichment of cytosine at the 5 0 position in the male core motif (Billi et al., 2013), we hypothesize that as-yet unidentified motifs may further discriminate male from female piRNA promoters. While we observed SNPC-1.3 binding to be enriched upstream of male piRNA loci (Figure 4), we cannot definitively conclude that SNPC-1.3 binds to the male-specific core motif, given the limitations of conventional ChIP-seq in resolving the SNPC-1.3 footprint. Identifying the factors that specifically bind the eight nt core motif and other potential cis-regulatory elements important for sex-specific piRNA expression will require further investigation.
What are the functions of male piRNAs in C. elegans?
Our data suggest that SNPC-1.3 is essential for proper spermiogenesis (Figure 6). We hypothesize the global loss of male piRNAs in a snpc-1.3(-) mutant is responsible for the higher incidence of spermiogenesis arrest and subsequent loss in fertility, although it is possible that SNPC-1.3 may have other or additional effects on male fertility. Characterization of prg-1(-) mutants during spermiogenesis agree with our findings that loss of piRNAs in the male germline leads to acute defects directly responsible for fertility (Wang and Reinke, 2008). Since the initial discovery of piRNA function in the targeting and silencing of transposons in Drosophila (Vagin et al., 2006;Brennecke et al., 2007), analyses in other systems have revealed that piRNAs have acquired neofunctions at later points along the evolutionary time scale (Ozata et al., 2019).
While it is estimated that as much as 45% of the human genome encodes for transposable elements (Lander et al., 2001), only 12% of C. elegans genome encodes such elements. Furthermore, nearly all of these regions are inactive in C. elegans (Bessereau, 2006). In contrast to Drosophila piRNAs that target and silence transposons with perfect complementarity (Brennecke et al., 2007), C. elegans piRNAs are thought to bind a broad range of endogenously expressed transcripts by partial complementarity (Ashe et al., 2012;Shen et al., 2018;Zhang et al., 2018). Together, these findings suggest that worm piRNAs function in capacities distinct from canonical transposon silencing. While a recent methodology used crosslinking, ligation, and sequencing of piRNA:target hybrids (CLASH) to determine that female piRNAs engage with almost every germline transcript (Shen et al., 2018), how female piRNAs select their targets has yet to be examined. Like piRNAs characterized in the female germline, male piRNAs may be interfacing with a broad range of targets to regulate gene expression for proper spermatogenesis. Loss of prg-1 in males causes the downregulation of a subset of spermatogenesis-specific genes (Wang and Reinke, 2008), suggesting male piRNAs serve a protective function for spermatogenic processes. The characterization of the in vivo landscape of male piRNA target selection using CLASH may provide insights into piRNA function during spermatogenesis.
Why are male piRNAs restricted from the female germline?
Sperm and oocytes pass epigenetic information such as noncoding RNAs to the next generation (Hammoud et al., 2014;Brykczynska et al., 2010;Tabuchi et al., 2018;Kaneshiro et al., 2019). Recent studies show maternal piRNAs trigger the production of endo-siRNAs, called 22G-RNAs for their 5 0 bias for guanine and 22 nt length, to transmit an epigenetic memory of foreign versus endogenous elements to the next generation (Ashe et al., 2012;Buckley et al., 2012;Shirayama et al., 2012). We predict that misexpression of male piRNAs during oogenesis may perturb the native pool of female piRNAs necessary for appropriate recognition of self versus non-self elements. This may explain the decrease in fertility we observed in multiple TRA-1 binding site mutant hermaphrodites ( Figure 6-figure supplement 1B). As snpc-1.3 (3xtbs) sperm do not seem to exhibit significant morphological defects (Figure 6), the fertility defects in the snpc-1.3 (3xtbs) mutants could be due to problems arising in oogenesis. However, based on our sequencing data in snpc-1.3 (2xtbs) mutants, we cannot distinguish whether fertility defects during oogenesis are due to upregulation of male piRNAs, downregulation of female piRNAs, a combination of the two, or misexpression of downstream endo-siRNAs triggered by piRNAs. Further study of snpc-1.3 gain-offunction mutants in oogenesis will enhance our understanding of the physiological consequences of expressing male piRNAs in the female germline.

The intersection between sex determination and sex specification of piRNA expression
We speculate that gene duplication of the snpc-1 family of genes occurred early during nematode evolution and allowed for the acquisition of new functions by snpc-1 paralogs, specifically, from snRNA to piRNA biogenesis. At least two SNPC-1 paralogs are present within the distantly related nematode species, Plectus sambesii. Furthermore, we predict that co-opting SNPC-1 paralogs for piRNA biogenesis may have occurred in parallel with the evolution of the nematode sex determination pathway. TRA-1 is a sex determination factor that acts to repress male-promoting gene expression in female germ cells to promote female germ cell fate. While Drosophila sex determination utilizes different factors than C. elegans, further investigation into the conservation of TRA-1 shows that it is a common feature in at least the nematode lineage (Pires-daSilva and Sommer, 2004). Additionally, just as we have shown that TRA-1 represses snpc-1.3 in C. elegans (Figure 5), TRA-1 binding motifs GGG(A/T)GG are present in the putative upstream promoter regions of snpc-1.3 homologs identified in C. briggsae, C. brenneri, and C. nigoni (Figure 6-figure supplement 1C). Taken together, these analyses point to a conserved link between sex determination and piRNA biogenesis pathways among nematodes.
In summary, our work reveals that SNPC-1.3 is specified to the male germline and is essential for male piRNA expression. We have identified SNPC-1.3 as a major target of TRA-1 repression in the female germline. Future studies will likely uncover additional factors required to coordinate the proper balance of sex-specific piRNAs required for proper germline development and animal fertility.

Contact for reagent and resource sharing
More details about resources and reagents can be found in the Key Resources Table found in Materials and methods. Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, John K. Kim (jnkim@jhu.edu).

Experimental model and subject details
C. elegans strains were maintained at 20˚C according to standard procedures (Brenner, 1974), unless otherwise stated. Bristol N2 was used as the wild-type strain. Except for RNAi and ChIP experiments, worms were fed E. coli strain OP50. Worms used for ChIP were fed E. coli strain HB101.

Continued on next page
Generations of strains CRISPR/Cas9-generated strains were created as described in Paix et al., 2015 and are listed in Supplementary file 4. crRNA and repair template sequences of CRISPR-generated strains are listed in Supplementary file 4. After initial phenotyping of snpc-1.3a::3xflag and snpc-1.3b::3xflag (Figure 1-figure supplement 1CÀG), snpc-1.3a::3xflag was used for all subsequent experiments (and is referred to as snpc-1.3::3xflag).

RNAi assays
Bacterial RNAi clones were grown from the Ahringer RNAi library (Kamath and Ahringer, 2003). Synchronized L1 worms were plated on HT115 bacteria expressing dsRNA targeting the gene interest or L4440 empty vector as a negative control as previously described (Timmons and Fire, 1998).
All RNAi experiments were performed at 20˚C unless otherwise stated.

RNA extraction, library preparation, and sequencing
After hypochlorite preparation and hatching in M9 buffer, snpc-4::aid::ollas and snpc-4::aid::ollas; Psun-1::TIR1 worms were transferred from NGM plates to plates containing 250 mM auxin 20 hr before collecting L4 and gravid worms, 48 hr and 72 hr after plating L1 worms at 20˚C, respectively. Worms were collected in TriReagent (ThermoFisher Scientific) and subjected to three freeze-thaw cycles. Following addition of 1-bromo-3-chloropropane, the aqueous phase was then precipitated with isopropanol at À80˚C for 2 hr. To pellet RNA, samples were spun at 21,000 Â g for 30 min at 4C . After three washes in 75% ethanol, the pellet was resuspended in water.
RNA concentration and quality were measured using a TapeStation (Agilent Technologies). We size-selected small RNAs of 16-30 nt in length from 5 mg total RNA on 17% denaturing polyacrylamide gels. Small RNAs were treated with 5 0 polyphosphatase (Illumina) to reduce 5 0 triphosphate groups to monophosphates to enable 5 0 adapter ligation. Small RNA-sequencing libraries were prepared using the NEBNext Multiplex Small RNA Library Prep Set for Illumina (NEB). Small RNA amplicons were size-selected on 10% polyacrylamide gels and quantified using qRT-PCR. Samples for each developmental time point were pooled into a single flow cell and single-end, 75 nt reads were generated on a NextSeq 500 (Illumina). An average of 42.01 million reads (range 33.05-50.39 million) was obtained for each library.

Quantitative RT-PCR
Taqman cDNA synthesis was performed as previously described (Weiser et al., 2017). Briefly, for quantification of piRNA levels, TaqMan small RNA probes were designed and synthesized by Applied Biosystems. All piRNA species assessed by qPCR were normalized to U18 small nucleolar RNA. Fifty nanograms of total RNA was used for cDNA synthesis. cDNA was synthesized by Multiscribe Reverse Transcriptase (Applied Biosystems) using the Eppendorf Mastercycler Pro S6325 (Eppendorf). Detection of small RNAs was performed using the TaqMan Universal PCR Master Mix and No AmpErase UNG (Applied Biosystems). For quantification of mRNA levels, cDNA was made using 500 ng of total RNA using Multiscribe Reverse Transcriptase (Applied Biosystems). For quantification of snRNA levels, cDNA was made using 250 ng of total RNA using SuperScript III Reverse Transcriptase (ThermoFisher). Assays for mRNA and snRNA levels were performed with Absolute Blue SYBR Green (ThermoFisher) and normalized to eft-2 using CFX63 Real Time System Thermocyclers (Bio-Rad). All qPCR primers used are listed in Supplementary file 4.

Covalent crosslinking of Dynabeads
Protein G Dynabeads (ThermoFisher Scientific, 1003D) were coupled to monoclonal mouse anti-FLAG antibody M2 (Sigma-Aldrich, F1804). After three washes in 1Â PBST (0.1% Tween), Dynabeads were resuspended with 1Â PBST with antibody, for a final concentration of 50 mg antibody per 100 mL beads. The antibody-bead mixture was nutated for 1 hr at room temperature. After three washes in 1Â PBST and two washes in 0.2 M sodium borate pH 9.0, beads were nutated in 22 mM DMP (Sigma-Aldrich, D8388) in 0.2 M sodium borate for 30 min at room temperature. Following two washes in ethanolamine buffer (0.2 M ethanolamine, 0.2 M NaCl pH 8.5), beads were nutated for 1 hr at room temperature in the same buffer. Beads were placed into the same volume of ethanolamine buffer as the starting bead volume for storage at 4˚C until use.
Unless otherwise stated, all samples for mass spectrometry, co-IP, and western blotting used in this study were subjected to the following procedure. After three washes in M9 and one wash in water, worms were frozen and ground using the Retsch MM400 ball mill homogenizer for two rounds of 1 min at 30 Hz. Frozen worm powder was resuspended in 1Â lysis buffer used previously (Moissiard et al., 2014, 50 mM Tris-HCl pH 8.0, 150 mM NaCl, 5 mM MgCl 2 , 1 mM EGTA, 0.1% NP-40, 10% glycerol) and protease inhibitor cocktail (Roche). After Bradford assay (Thermo-Fisher Scientific), lysates were normalized using lysis buffer and protease inhibitor. Benzonase (Sigma-Aldrich, E1014) was added to a final concentration of 1 mL/mL of lysate and nutated for 10 min at 4˚C. After centrifugation for 10 min at 4,000 x g, 1 mL of supernatant was added to 50 mL of crosslinked Dynabeads and nutated for 15 min at 4˚C. Samples were then washed three times in 1Â lysis buffer with protease inhibitors before 1 hr nutation in 50 mL of 2 mg/mL FLAG peptide (Sigma-Aldrich, F4799) diluted in 1Â lysis buffer. Complete eluate, as well as 5% of crude lysate (after addition of benzonase), input, pellet, and post-IP samples, were added to 2Â Novex Tris-glycine sodium dodecyl sulfate sample buffer (ThermoFisher Scientific, LC2676) to 1Â. Samples were then subjected to western blotting as described below.
For measuring SNPC-1.3 expression levels in various backgrounds, input (for normalization) and immunoprecipitation samples were run on 10% Novex WedgeWell Tris-glycine precast gels (Ther-moFisher). Following transfer, the membrane was dried for 20 min at room temperature. The blot was then recharged in 100% methanol for 1 min, followed by a water rinse and a wash in TBS for 2 min. Blocking was performed in LI-COR Odyssey Blocking Buffer (TBS). Primary antibodies were 1:1,000 mouse anti-Flag (Sigma-Aldrich F1804) and 1:5,000 rabbit anti-gamma-Tubulin (Sigma-Aldrich T1450) in LI-COR Odyssey Blocking Buffer with 0.1% Tween. Washes were performed in TBST (TBS + 0.1% Tween). LI-COR IRDye 800CW goat anti-mouse IgG and 680RD goat anti-rabbit IgG were used at 1:15,000 in Odyssey Blocking Buffer with 0.1% Tween and 0.01% SDS. After three washes in TBST, the membranes were incubated in TBS before imaging in the LI-COR Odyssey Fc.

Mass spectrometry and analysis
Proteins were precipitated with 23% TCA and washed with acetone. Protein pellets solubilized in 8 M urea, 100 mM Tris pH 8.5, and reduced with 5 mM Tris (2-carboxyethyl)phosphine hydrochloride (Sigma-Aldrich, St. Louis, MO, product C4706) and alkylated with 55 mM 2-chloroacetamide (Fluka Analytical, product 22790). Proteins were digested for 18 hr at 37˚C in 2 M urea 100 mM Tris pH 8.5, 1 mM CaCl 2 with 2 mg trypsin (Promega, Madison, WI, product V5111). Single-phase analysis (in replicate) was performed using a Dionex 3000 pump and a Thermo LTQ Orbitrap Velos using an inhouse built electrospray stage (Wolters et al., 2001). Protein and peptide identification and protein quantitation were done with Integrated Proteomics Pipeline, IP2 (Integrated Proteomics Applications, Inc, San Diego, CA; http://www.integratedproteomics.com/). Tandem mass spectra were extracted from raw files using RawConverter (He et al., 2015) with monoisotopic peak option and were searched against protein database release WS260 from WormBase, with FLAG-tagged SNPC-4, common contaminants and reversed sequences added, using ProLuCID (Peng et al., 2003;Xu et al., 2006). The search space included all fully tryptic and half-tryptic peptide candidates with a fixed modification of 57.02146 on C. Peptide candidates were filtered using DTASelect (Tabb et al., 2002).
Using custom R scripts, average enrichment between SNPC-4::3xFlag and no-tag control immunoprecipitation experiments were calculated. For each experiment, enrichment was normalized by dividing the peptide count for each protein by the total peptide count. Adjusted p-values were calculated by applying the Bonferroni method using DESeq2 (Love et al., 2014). Although SNPC-3.1 and SNPC-3.2 are reported to have the same amino acid sequence, we have picked up differential peptide coverage in the fem-1(-) mutant for these two proteins and represented them as two different data points.

Immunofluorescence microscopy
Adult gonads were dissected into egg buffer (25 mM HEPES pH 7.4, 118 mM NaCl, 48 mM KCl, 2 mM EDTA, 0.5 mM EGTA) with 30 mM sodium azide and 0.1% Tween-20, and fixed for 10 s in 1% formaldehyde in egg buffer followed by 1 min in 100% methanol at À20˚C. All washing and staining was completed in suspension. Germlines were blocked in normal goat serum or 1Â Roche blocking buffer in PBST (PBS + 0.2% Tween) for 30 min at room temperature. Primary mouse anti-Flag (Sigma F1804) and rat anti-Ollas (NBP1-06713SS) antibodies were used at 1:200 in blocking agent in PBST. AlexaFluor 555 goat anti-mouse and AlexaFluor 488 goat anti-rat secondary antibodies (Thermo-Fisher) were used at 1:400 in blocking agent in PBST. Germlines were stained with 0.5 mg/mL DAPI and then mounted in Vectashield with DAPI (Vector Laboratories H-1200). Images were acquired at 63x on a Zeiss LSM700 confocal microscope. Publication images were acquired at 100Â on a GE DeltaVision microscope. Image processing was performed using SoftWoRx to collect 3D image stacks, deconvolve (enhanced ratio, 20 cycles), and compile into a maximum intensity projection. Composite images were stitched and colored in Fiji using the Stitching plugin (Preibisch et al., 2009).

Chromatin immunoprecipitation, library prep, and sequencing
Worms were grown in liquid culture as previously described (Zanin et al., 2011). 250 mM auxin was added to snpc-1.3::3xflag; snpc-4::aid::ollas; Psun-1::TIR1 worms 4 hr before collection at 48 hr post-L1 at 20˚C. After washing, the gut was cleared for 15 min by nutation in M9, followed by three washes in M9. Worms were live-crosslinked in 2.6% formaldehyde in water for 30 min at room temperature with nutation. Crosslinking was quenched with a final concentration of 125 mM glycine for 5 min with nutation. After three washes with water, worms were flash-frozen in liquid nitrogen. Frozen worm pellets were ground into powder using the Retsch MM40 ball mill homogenizer for 2 rounds of 1 min at 30 Hz. Frozen worm powder was resuspended in 1Â RIPA buffer (1Â PBS, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) for 10 min at 4˚C. Crosslinked chromatin was sonicated using a Diagenode Bioruptor Pico for three 3 min cycles, 30 s on/off. We nutated 10 mg of chromatin overnight at 4˚C with 2 mg of Flag antibody (Sigma-Aldrich, F1804) and then for 1.5 hr with 50 mL mouse IgG Dynabeads (Invitrogen). Input amount was 10% of IP. Chromatin was de-crosslinked and extracted as described previously (Weiser et al., 2017). Individual input and IP samples of each genotype were processed for both sequencing and quantitative PCR.
Libraries were prepared and multiplexed using the Ovation Ultralow Library Systems v2 (NuGEN Technologies) according to the manufacturer's protocol. The Illumina HiSeq 4000 platform was used to generate 50 bp single-end reads for SNPC-1.3 ChIP-seq libraries. The NovaSeq 6000 platform was used to generate 50 bp paired-end reads for TRA-1 ChIP-seq libraries.

Quantitative PCR of ChIP samples
ChIP DNA was eluted in 18 mL of 1Â TE pH 8.0 and 2 mL of 20 mg/mL RNase A (Invitrogen, Thermo-Fisher Scientific). For a final reaction volume of 25 mL, each reaction consisted of final 1Â Absolute Blue SYBR Green (ThermoFisher Scientific), 35 nM each of forward and reverse primer, and 2 mL ChIP eluate. Reactions were performed in technical duplicates in a Bio-Rad CF96 Real Time PCR thermal cycler.

Hermaphrodite fertility assays
Gravid worms (previously maintained at 20˚C) were subjected to hypochlorite treatment, and their progeny were plated onto NGM at 25˚C (P0). At the L2 or L3 stage, worms were singled onto individual plates and their progeny (F1) counted.

Mating assays
To test male-dependent rescue of fem-1(hc17) fertility, 10-12 hermaphrodites of each strain were grown at 20˚C and embryos were isolated by allowing egg lay for 2 hr before removal. Embryos were shifted to 25˚C, and upon reaching the L4 stage (24 hr), ten him-8(e1489) L4 males were transferred and mated with two fem-1(hc17) females. Brood size was quantified by counting when a majority of progeny had at least reached the young adult stage (about 3 days after transfer). To test the fertility of the hermaphrodites upon mating, 10-12 hermaphrodites of each strain were grown at 20˚C and embryos were isolated after egg lay for 2 hr before removal. Embryos were shifted to 25˚C and 1ten col-19(GFP+) L4-staged males (24 hr) were then transferred with a single hermaphrodite (36 hr), and the number of live cross progeny was counted after reaching adulthood. Brood size was quantified by counting when the majority of progeny had at least reached the young adult stage (about 3 days after transfer).

Sperm activation assay and imaging
To perform sperm activation assays, spermatids were dissected from adult males that were shifted to 25˚C during the embryo stage, and isolated prior to sexual maturity (about 48 hr post-L1). Dissection was performed directly on glass slides in sperm medium (50 mM HEPES pH 7.8, 50 mM NaCl, 25 mM KCl, 5 mM CaCl 2 , and 1 mM MgSO 4 ) supplemented with 20 mg/mL pronase E (Millipore Sigma). For the characterization of sperm morphology, sperm were imaged 30 min after the addition of pronase E. Individual sperm were manually categorized into two types: spermatids with normal pseudopods or spermatids with irregular or no pseudopods (Shakes and Ward, 1989). For Figure 6E,Z, stacks were imaged in 10 s intervals for 30 min and a representative in-focus stack was chosen at every 3 min interval. To characterize sperm activation dynamics, sperm were individually followed across 10 s intervals for 30 min, and the different stages of sperm activation were designated into four categories based on these morphological changes: (1) undifferentiated spermatid, (2) spiky intermediate characterized by the presence of spike growth, (3) growing or motile pseudopod by the presence of a pseudopod, and (4) immobile sperm when little movement was observed either in the sperm body or pseudopod for longer than 30 s. Statistical significance was assessed using Student's t-test.

Quantitative and statistical analysis
Unless otherwise stated, all quantitative analyses are shown as mean with standard deviation represented as error bars. For qRT-PCR, fertility and mating assays, and western blot, at least two independent experiments were performed; one representative biological replicate is shown.
Mapped reads were assigned to genomic features using featureCounts from Subread 1.6.3 (Liao et al., 2014), taking into account overlapping and multi-mapping reads (-O -M). Raw counts were normalized within DESeq2 1.26.0 (Love et al., 2014), and principal component analysis (PCA) was performed using the regularized log transform of normalized counts within DESeq2 ( Figure 5figure supplement 1C). In addition, we distributed mapped reads by size and 5 0 nucleotide identity to verify the presence of small RNA species such as 22G-RNAs and 21U-RNAs (Figure 3-figure supplement 2B, Figure 5-figure supplement 1B).
To identify differentially expressed genes, DESeq2 was applied to piRNAs on chromosome IV. In this study (method 1), we define significant and differentially expressed genes as having an absolute value of log 2 (fold-change) ! 0.26 and FDR of 0.05 (Benjamini-Hochberg). The log 2 (fold-change) threshold and significance level were selected based on benchmarking the differential expression results against the Taqman piRNA expression assays. At the chosen cutoffs, differential expression analysis captures the Taqman assays results for the three male piRNAs (21UR-1258, 21UR-3142, and 21UR-3766) and three female piRNAs (21UR-1848, 21UR-2502, and21UR-4817). Contrasts between mutant and wild type were designed without independent filtering.
For motif discovery, nucleotide sequences were extracted from the reference genome with 60 nt upstream of each piRNA and submitted to the MEME suite 5.1.1 (Bailey et al., 2009). Results from MEME were used to generate the sequence logo plot with the median position of the C-nucleotide of the identified motif, number of piRNAs, and the associated E-value.
A second, independent small RNA-seq analysis workflow (described in Figure 3-figure supplement 1) was implemented to validate our results. Results produced from this analysis are provided in Figure 3-figure supplement 2. We parsed -small RNA sequences that were 16-30 nt from adapters. Reads with >3 nt falling below a quality score of Q30 were discarded. Reads were mapped to the C. elegans WS230 (Stein et al., 2001) reference genome using CASHX v. 2.3 (Fahlgren et al., 2009) allowing for 0 mismatches. Custom Perl, Awk, and R scripts were used to count features and to generate PCA and size distribution plots. Multi-mapping reads were assigned proportionally to each possible locus. Differential expression analysis was done using DESeq2 v. 1.18.1 (Love et al., 2014). A reporting threshold was set at an absolute value of log 2 (fold-change) ! 0.26 and a Benjamini-Hochberg-corrected p 0.20.
To identify and visualize binding sites and peaks for SNPC-1.3 ChIP-seq, filtered SNPC-1.3 ChIPseq reads were extended to 200 bp to account for the average length of ChIP fragments. We then partitioned the genome into consecutive, non-overlapping 1 kb bins and calculated read coverage, normalized by sequencing depth of each library, based on the total read count in each bin. Bins with read coverages in the IP sample that fell below the median read coverage of piRNA-depleted bins on chromosome IV in the relevant input control were excluded from further analysis. Bins containing only male, female, and non-enriched piRNAs (as defined by small RNA-seq analysis) were then extracted to generate binding profiles and heatmaps. For this, the bamCompare tool in deepTools 3.3.1 (Ramírez et al., 2016) was used to calculate the ratio between read coverage of each ChIP sample and input control (-scaleFactorsMethod None -normalizeUsing CPM -operation ratio -binSize 50 -ignoreForNormalization MtDNA -extendReads 200). The ENCODE ce11 blacklist (Amemiya et al., 2019) was also supplied (https://github.com/Boyle-Lab/Blacklist/). The bamCompare output was then used in deepTools computeMatrix to calculate scores for plotting profiles and heatmaps with deepTools plotProfile and plotHeatmap.
TRA-1 ChIP-seq peaks were called by callpeak within MACS 2.1.2 (Zhang et al., 2008) (-p-value 0.05) with filtered TRA-1 ChIP-seq reads and relevant input controls. TRA-1 signal tracks were generated by calculating fold enrichment from read count-normalized genome-wide pileup and lambda track outputs by callpeak (bdgcmp in MACS2). The ENCODE ce 11 blacklist (Amemiya et al., 2019) was supplied in this analysis (https://github.com/Boyle-Lab/Blacklist/). The bamCompare tool in deepTools 3.3.1 (Ramírez et al., 2016) was used to quantify read coverage of each ChIP sample and input control.

Data and software availability
The mass spectrometry, small RNA-seq, and ChIP-seq data have been deposited in NCBI under GEO accession number: GSE152831. Processed data and scripts used for analysis are available at https://github.com/starostikm/SNPC-1.3; Choi, 2021; copy archived at swh:1:rev: b23f652341d999150edf5ae9c8de72e9192b2843.