PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Mol Cell. Author manuscript; available in PMC May 18, 2012.
Published in final edited form as:
PMCID: PMC3236501
NIHMSID: NIHMS340674
Heterotypic piRNA Ping-Pong Requires Qin, a Protein with Both E3 ligase and Tudor Domains
Zhao Zhang,1 Jia Xu,2,3 Birgit S. Koppetsch,4 Jie Wang,2 Cindy Tipping,1 Shengmei Ma,1 Zhiping Weng,2* William E. Theurkauf,4* and Phillip D. Zamore1*
1Biochemistry and Molecular Pharmacology and Howard Hughes Medical Institute, University of Massachusetts Medical School, Worcester, Massachusetts 01605, USA
2Program in Bioinformatics and Integrative Biology, University of Massachusetts Medical School, Worcester, Massachusetts 01605, USA
3Bioinformatics Core Facility, University of Massachusetts Medical School, Worcester, Massachusetts 01605, USA
4Program in Molecular Medicine, University of Massachusetts Medical School, Worcester, Massachusetts 01605, USA
* Correspondence: zhiping.weng/at/umassmed.edu (ZW), william.theurkauf/at/umassmed.edu (WET), phillip.zamore/at/umassmed.edu (PDZ)
piRNAs guide PIWI proteins to silence transposons in animal germ cells. Reciprocal cycles of piRNA-directed RNA cleavage—catalyzed by the PIWI proteins Aubergine (Aub) and Argonaute3 (Ago3) in Drosophila melanogaster—expand the population of antisense piRNAs in response to transposon expression, a process called the Ping-Pong cycle. Heterotypic Ping-Pong between Aub and Ago3 ensures that antisense piRNAs predominate. We show that qin, a piRNA pathway gene whose protein product contains both E3 ligase and Tudor domains, co-localizes with Aub and Ago3 in nuage, a perinuclear structure implicated in transposon silencing. In qin mutants, less Ago3 binds Aub, futile Aub:Aub homotypic Ping-Pong prevails, antisense piRNAs decrease, many families of mobile genetic elements are reactivated, and DNA damage accumulates in nurse cells and oocytes. We propose that Qin enforces heterotypic Ping-Pong between Aub and Ago3, ensuring that transposons are silenced and maintaining the integrity of the germline genome.
In animals, PIWI-interacting RNAs (piRNAs) silence transposons and other repetitive elements (Klattenhoff and Theurkauf, 2008; Ghildiyal and Zamore, 2009; Siomi et al., 2011). For flies, mammals, and other bilateral animals, the piRNA pathway protects the germline genome from DNA damage and mutation, ensuring that genetic information passes faithfully from generation to generation.
piRNAs are typically 23–30 nucleotides (nt) long and bind to members of the PIWI clade of Argonaute proteins, a family that includes the proteins that mediate RNA interference and microRNA-directed gene regulation. The Drosophila PIWI clade comprises three proteins: Argonaute3 (Ago3), Aubergine (Aub) and Piwi. In the germline nurse cells, which support development of the oocyte, Ago3 and Aub reside in a perinuclear, cytoplasmic structure called “nuage”; Piwi resides in the nuclei of ovary germ cells and the somatic follicle cells that surround the germ cells (Wilson et al., 1996; Cox et al., 1998; Cox et al., 2000; Saito et al., 2006; Brennecke et al., 2007; Gunawardane et al., 2007).
In fly germ cells, primary piRNAs are thought to derive from discrete genomic loci, “piRNA clusters,” that contain complex arrays of transposon sequences (Brennecke et al., 2007). Reciprocal cycles of Aub- and Ago3-mediated RNA cleavage are thought to increase piRNA abundance by producing secondary piRNAs. Secondary piRNAs are disproportionately antisense to germline-expressed transposons (Aravin et al., 2006; Girard et al., 2006; Grivna et al., 2006; Lau et al., 2006; Vagin et al., 2006; Brennecke et al., 2007; Houwing et al., 2007). Amplification of antisense piRNAs requires both Aub and Ago3, and piRNA production in the germline collapses in ago3 mutants (Brennecke et al., 2007; Li et al., 2009; Malone et al., 2009). The “Ping-Pong” model for piRNA amplification (Brennecke et al., 2007; Gunawardane et al., 2007) proposes that Aub, guided by an antisense primary piRNA, binds to and cleaves a complementary transposon mRNA, simultaneously destroying the transposon transcript and generating a new, sense piRNA that is loaded into Ago3. The model envisions that sense secondary piRNA direct Ago3 to cleave an antisense sequence in the original piRNA precursor transcript to generate another Aub-bound antisense piRNA (Brennecke et al., 2007; Li et al., 2009; Malone et al., 2009). Heterotypic Ping-Pong between Aub and Ago3 produces more antisense piRNAs than sense, but the molecular mechanisms that bias the Ping-Pong cycle towards antisense remain unknown. Strikingly, the absence of Ago3 results in a futile, homotypic Ping-Pong cycle that generates more sense than antisense piRNAs, both of which bind Aub. Homotypic Aub:Aub Ping-Pong fails to silence all transposons, and ago3 mutant females are infertile (Li et al., 2009).
The fly genome encodes 23 predicted Tudor d omain proteins, seven of which have been shown to act in the piRNA pathway (Lim and Kai, 2007; Patil and Kai, 2010; Siomi et al., 2010; Liu et al., 2011; Zamparini et al., 2011). The binding of Tudor-domain proteins to di-methylarginine-modified PIWI proteins is conserved from flies to mammals, but its molecular function remains unknown (Kirino and Mourelatos, 2007; Kirino et al., 2009; Nishida et al., 2009; Vagin et al., 2009; Wang et al., 2009; Siomi et al., 2010). Here, we identify a Drosophila Tudor-domain protein, Qin, which is required for piRNA production in the germline. The qin gene encodes an unusual protein with an amino-terminal E3 ligase domain and five carboxy-terminal Tudor domains. In qin mutants, futile homotypic Aub:Aub Ping-Pong replaces heterotypic Ping-Pong between Aub and Ago3, activating transposon expression, and triggering DNA damage in the nurse cells and the oocyte. Thus, qin mutants phenocopy ago3 mutants, although the abundance of Ago3 is greater than that in ago3 heterozygotes. Qin localizes to nuage and appears to be required for the physical interaction of Aub with Ago3.
qin Encodes a Protein with Five Tudor and One E3 Ligase Domain
FlyBase (Drysdale et al., 2005) predicts that the putative transcription unit CG14303 encodes a protein with five Tudor domains (Figures 1A and S1). Disruption of CG14303 by insertion of a piggyBac transposon ({RB}PBacCG14303e03728; Thibault et al., 2004) in the largest predicted exon caused female sterility and produced embryos with fused or missing dorsal appendages (Table S1), phenotypes associated with a failure of the piRNA pathway (Schupbach and Wieschaus, 1989; Schupbach and Wieschaus, 1991; Wilson et al., 1996; Theurkauf et al., 2006; Chen et al., 2007; Klattenhoff et al., 2007). Because CG14303 mutant females produce offspring that fail to develop into adults, we named the gene qin, after the ancient Chinese dynasty ( An external file that holds a picture, illustration, etc.
Object name is nihms340674ig1.jpg) that ended after just two generations.
Figure 1
Figure 1
The qin Gene and Qin protein
PBac{RB}CG14303e03728 (qin1) homozygotes produced few eggs, all of which display dorsal appendage defects (Table S1). Similarly, only 4.3% of embryos from qin1/Df(3R)Excel6180 females had wild-type appendages (Table S1 and Figures 1A and S2), suggesting that mutation of qin causes the observed phenotypes. The deficiency Df(3R)Excel6180, henceforth Df, removes qin as well as other genes (Figure S2; Parks et al., 2004). Although qin1 is predicted to encode a truncated protein, qin1 behaved like a genetic null allele: the dorsal appendage phenotype, egg hatching, and eggs laid per day per female were as severe or worse in qin1 homozygotes compared to qin1/ Df (Table S1).
To further verify that these defects resulted from mutation of qin, we attempted to rescue qin1/ Df by creating a transgenic fly expressing genomic fragment CH322-81J04 (henceforth, gf-1; Venken et al., 2009), which encompasses CG14303 (Figure 1A). The gf-1 transgene failed to rescue the defects in qin1/ Df ovaries, eggs, and embryos (Table S1 and Figure 2A). Moreover, the transgene carrying gf-1 generated a transcript that was shorter than wild -type qin mRNA (Figure 1B). To understand why gf-1 failed to rescue, we mapped the 52 end of the qin mRNA by rapid amplification of cDNA ends (52 RACE). The 52 end of the qin mRNA mapped 51,682 bp upstream of CG14303, in the putative gene CG14306, which is predicted to encode a protein with a RING domain and two B-Box domains, the hallmarks of a Ubiquitin or SUMO E3 ligase (Figures 1A, 1C, and S1).
Figure 2
Figure 2
Qin Silences Transposons in Drosophila Ovaries
In control (w1118) ovaries, a Northern hybridization probe for CG14306 detected the same size mRNA as a probe for CG14303 (Figure 1D). Probes for CG14306 and CG14303 both detected a longer RNA in qin1 whose size is consistent with transcription of three exons in CG14306 and five exons plus part of a sixth in CG14303 fused to PBac{RB}CG14303e03728, the piggyBac insertion that creates the qin1 allele. Neither hybridization probe detected the qin mRNA in ovaries from females bearing both the deficiency and PBac{RB}e01936 (qin2), a piggyBac insertion 52 to the first predicted exon of CG14306 (Figures 1D and S3A). Consistent with qin encompassing both CG14306 and CG14303, qin2 failed to complement qin1 (Figure S3B). Moreover, a cDNA comprising the predicted exons of CG14306 and CG14303 (UASp-(FLAG)3-(Myc)6-Qin, henceforth FM-Qin) and driven from a UASp promoter by nanos-Gal4-VP16, partially rescued the female sterility and dorsal appendage (Table S1) and transposon silencing defects associated with loss of Qin (Figures 2A and 2B). (The incomplete rescue m ay reflect the poor expression of nanos-Gal4-VP16 in stage 2–6 egg chambers (Figure S3A) (Ni et al., 2011) or an effect of the amino-terminal FLAG-Myc tag on Qin function or stability.)
As a final test of the idea that the qin locus comprises both CG14303 and CG14306, we constructed a transgenic fly bearing a 97,532 bp genomic fragment (gf-2, CH321-88L14) (Venken et al., 2009) encompassing all of the qin exons, as well as 6,602 bp 52 to exon 1 and 32,343 bp 32 to exon 10 (Figures 1A and S3B). To facilitate detection of the predicted Qin protein encoded by gf-2, we inserted the EGFP coding sequence before the qin stop codon of the genomic fragment. Gf-2 includes part or all of the qin promoter, because transgenic flies bearing gf-2 expressed the Qin::EGFP fusion in the germline throughout oogenesis (Figure S3C). A single copy of gf-2 partially rescued the dorsal appendage defects, low egg hatch rate, and the low number of eggs laid per day per fly observed for w; Sp/ +; qin1/ Df females (Table S1). We conclude that the qin locus comprises the three exons of CG14306 and the seven exons of CG14303 joined by the removal of a 51,682 nt intron and that the Qin protein contains an amino-terminal E3 ligase domain and five carboxy-terminal Tudor domains (Figures 1A and S1).
Qin is Required to Protect the Germline Genome and Silence Transposons
Defects in the piRNA pathway cause double-stranded breaks in the germline genome, and γH2Av, a phosphorylated histone variant, accumulates at the sites of DNA damage (Theurkauf et al., 2006; Chen et al., 2007; Klattenhoff et al., 2007). Homozygous qin1, qin1/Df, and qin2/Df mutant ovaries showed many nuclear γH2Av foci (Figures 3A and S4), unlike control ovaries, which displayed only the expected small number of γH2Av foci likely to be associated with normal endoreduplication (Figure 3A) or meiotic recombination (Figure S4). The increased number and intensity of γH2Av foci in qin mutants suggest that Qin is required for transposon silencing.
Figure 3
Figure 3
Qin Colocalizes with Aub and Ago3 in Nuage
We used whole-genome tiling microarrays to survey the expression of transposons and mRNAs in qin mutant ovaries. In qin mutants, 56 of 93 transposon families showed a 1.5- to 3-fold increase in steady-state transcript abundance (Figure 2C). The expression of another 13 transposon families increased more than 3-fold (False Discovery Rate, FDR, < 0.1). In contrast, the expression of none of the 19,987 protein coding genes detected by the microarray, including 613 annotated heterochromatic genes (Smith et al., 2007), increased or decreased >3-fold at FDR < 0.1 (Figure S5). We used quantitative RT-PCR to measure the expression of those elements that increased more than 3-fold with an FDR < 0.1 in the microarray data. Of the 13 such transposon families, 11 showed a statistically significant increase in RNA expression when measured by qRT-PCR (Table S2). For example, expression of the transposon HeT-A increased 72-fold (p-value = 0.004), Springer increased 55-fold (p-value = 0.002), and Burdock increased 41-fold (p-value = 0.006).
The transposon families whose expression increased in qin mutants correspond to a subset of those transposons whose silencing requires Ago3 (Figure 2D; Pearson correlation, r = 0.77, p-value < 10−16). In contrast, the transposons hyper-expressed in qin mutants and those activated in aub mutants were more weakly correlated (Figure 2E; r = 0.58, p-value = 1.7 × 10−9). A single copy transgene expressing full length qin cDNA in the germline via a UASp promoter driven by a nanos-Gal4-VP16 transgene restored transposon silencing (Figures 2A and 2B). ZAM, a transposon silenced by Piwi in the somatic follicle cells (Brennecke et al., 2007; Mevel-Ninio et al., 2007; Desset et al., 2008; Malone et al., 2009), was unaffected in qin1/ qin1 and qin1/ Df ovaries (Figure 2B). We also did not detect a change in the expression of blood, a transposon thought to be silenced in both the germline and the follicle cells (Li et al., 2009; Malone et al., 2009).
In testes, piRNAs silence the multi-copy gene Stellate, a gene whose hyper-expression causes Stellate protein crystals to form during spermatogenesis, reducing male fertility (Hardy, 1980; Balakireva et al., 1992; Aravin et al., 2001; Vagin et al., 2006). Anti-Stellate piRNAs derive from transcripts from the many copies of Suppressor of Stellate (Su(Ste)) (Hardy et al., 1984; Livak, 1984; Livak, 1990; Balakireva et al., 1992). Accumulation of Su(Ste)-derived piRNAs requires Aub (Aravin et al., 2004) and Ago3 (Li et al., 2009), but not Piwi (Vagin et al., 2006). Stellate silencing also requires Qin. Stellate crystals accumulated in the testes of qin1/ Df mutant males but not their heterozygous siblings (Figure 3B). Expression of FM-Qin cDNA in testes restored Stellate silencing (Figure 3B).
Qin Co-localizes with Ago3 and Aub in Nuage
To define the subcellular distribution of Qin, we examined the localization of the Qin::EGFP fusion encoded by the 97,532 bp genomic fragment gf-2. We detected EGFP fluorescence in the germline throughout oogenesis, but not in the somatic follicle cells (Figure S3C). Together with the finding that a FM-Qin cDNA expressed in the germline rescued the dorsal appendage defects and sterility of qin mutant females and rescued Stellate silencing in males, these data suggest that Qin acts mainly or solely in the germline.
The Qin::EGFP fusion protein produced by the gf-2 transgene and the epitope tagged FM-Qin produced by the full-length qin cDNA transgene expressed from a UASp promoter driven by nanos-Gal4-VP16 were both predominately cytoplasmic and localized in perinuclear foci typical for nuage (Figures S3A and S3C). Moreover, in all nurse cells examined (n = 30, two measurements for each nurse cell), quantitative fluorescence microscopy showed that Qin::EGFP co-localized with Aub and Ago3 (Figure 3C and 3D), which have been shown previously to localize to nuage (Harris and Macdonald, 2001; Snee and Macdonald, 2004; Brennecke et al., 2007; Lim and Kai, 2007). However, mutation of qin did not appear to affect the assembly or stability of nuage, as Aub and Ago3 all retained their perinuclear, punctate distribution in both qin1/Df and qin2/ Df ovaries (Figure S6). Neither a truncated Qin protein comprising the Qin E3 ligase-like domain alone nor a truncated Qin protein comprising the five Tudor domains but missing the E3 ligase domain localized to nuage (Figure S3A) or rescued the phenotypic defects observed in qin mutant ovaries (data not shown).
Compared with qin heterozygotes, the abundance of Ago3, Aub, and Piwi was not significantly changed in qin mutants (Figures 4A and S7). We did observe a small decrease in Ago3 abundance in the absence of Qin, but this 27% reduction is unlikely to explain the defects observed in qin mutants for two reasons. First, over-expression of Ago3 in the germline of qin mutant ovaries failed to rescue female sterility (hatch rate = 1.1%, n = 1,031). Second, the abundance of Ago3 was lower in ago3 heterozygotes than in qin1/Df ovaries, yet ago3/TM6B females are fertile, lay eggs with wild -type dorsal appendages, and silence their transposons (Li et al., 2009).
Figure 4
Figure 4
Mutation of qin Decreases the Antisense Character of piRNA Populations
Compared with the control strain (w1118), the abundance of Vasa increased in both qin heterozygous and mutant ovaries (Figures 4A and S7), although its localization to nuage was not detectably altered (Figure S6). Both qin heterozygous and mutant ovaries contained an additional Vasa isoform, likely phosphorylated Vasa, which has been associated with loss of transposon silencing and DNA damage in the germline (Ghabrial and Schupbach, 1999; Abdu et al., 2002; Chen et al., 2007; Klattenhoff et al., 2007).
The Ratio of Sense to Antisense piRNAs Increases in qin Mutants
qin mutants show little change in the total abundance of piRNAs: the percent of genome-mapping, small RNA sequence reads corresponding to transposon-derived, 23–29 nt small RNAs was essentially indistinguishable among control (cn1; ry506; 360,387/ 949,504 = 38%), qin/ TM6B heterozygous (5,558,981/ 14,619,724 = 38%), and qin1/Df mutant (6,064,405/ 16,096,174 = 38%) ovaries (Figure 4B and Table S3A and S3B). However, the effects of mutation of qin on the structure of piRNA populations can be readily detected by analyzing the fraction of piRNAs bearing the sense orientation (sense fraction = sense/ [sense + antisense]) for each of the 93 transposon families for which we sequenced = 100 parts per million (ppm) piRNA reads in both control and qin/ TM6B ovaries. Mutation of qin increased the median sense fraction for the 93 transposon families fr om 0.24 for the control ovaries to 0.33 for qin heterozygotes and 0.41 for qin1/Df mutants. We note that an increase in sense fraction implies a decrease in the proportion—but not necessarily the abundance—of antisense piRNAs, those piRNAs believed to direct transposon silencing.
One potential explanation for an increase in sense piRNAs in qin mutant ovaries is that the increased abundance of transposon transcripts directly feeds the production of sense piRNAs. In this view, the observed increase in the fraction of sense piRNAs would be a consequence, rather than the cause, of the loss of transposon silencing in qin mutant ovaries. To test this idea, we compared the change in transposon transcript abundance to the change in sense piRNA abundance for the 93 transposon families we analyzed (Figure S8A). Transposon expression and sense piRNA abundance were uncorrelated, even when we restricted our analysis to only those 11 transposon families whose expression increased significantly. We conclude that increased transposon expression cannot explain the increased sense piRNA abundance in qin mutant ovaries.
Among the 93 transposon families analyzed, 27 lost more than half their antisense piRNAs in qin1/ Df ovaries (Figures 4C; analyses of each transposon family are available at http://www.umassmed.edu/zamore/datasets.aspx?linkidentifier=id&itemid=66736). The median abundance of antisense piRNAs for these 27 transposon families in qin1/ Df ovaries was 32% of their median in the control, whereas the median sense piRNA abundance in qin1/ Df ovaries was 67% of the control median. Consequently, the piRNA populations from these 27 transposon families became less antisense and more sense biased. That is, loss of antisense piRNAs and not gain of sense piRNAs underlies the increase in sense fraction for these transposon families. Of the 27 transposon families, eight were among the 11 families that showed significantly elevated steady-state mRNA expression in qin1/Df compared with control ovaries (Figure 2C). Overall, antisense piRNAs from the 11 derepressed transposon families decreased more than those from the other 82 transposon families (p-value = 0.018, Wilcoxon rank-sum test), suggesting that loss of antisense piRNAs caused the transposon desilencing. This correlation was particularly striking for HeT-A, Burdock, TAHRE, and I element, whose expression increased ~40-, 38-, 21-, and 14-fold in qin mutants and whose antisense piRNA pools declined to only 14%, 18%, 35%, and 21% of control levels (Table S2A).
Of the remaining 66 transposon families, antisense piRNA abundance either increased or decreased by less than a factor of two for 59 families, and more than doubled for 7 families. While the median antisense piRNA abundance for these 66 families was unchanged from control, median sense piRNA abundance increased 2.3-fold (p-value = 5.7 × 10−10, two-tailed, paired t-test; Figure 4C). Even for these transposons, the normal antisense bias of piRNAs was reduced in qin mutant ovaries. Notably, copia expression increased 23-fold and Transpac expression increased 14-fold in qin mutant ovaries, yet copia antisense piRNA abundance increased 3-fold and Transpac antisense piRNA abundance was unchanged, compared to cn1; ry506 control ovaries (Table S2A). Thus, our data suggest that the ratio of sense to antisense piRNAs, not sim ply the abundance of antisense piRNAs, determines the efficacy of piRNAs in silencing transposons. The data also suggest that Qin acts to maintain the antisense bias of transposon piRNA populations.
Indeed, 74 of the 93 transposon families we analyzed had a greater piRNA sense fraction (qin mutants – control > 0.05) in the ovaries of the qin mutants than in the control (Figure 4D). The proportion of sense piRNAs declined in qin mutants compared with control (control – qin heterozygotes > 0.05) for only 6 transposon families (accord2, diver2, hopper, hopper2, gypsy, and gyspsy12). Among these, four families (accord2, diver2, hopper, hopper2) correspond to transposons whose sense piRNAs have been previously shown to be loaded into Aub, causing antisense piRNAs to accumulate in Ago3 (Brennecke et al., 2007; Li et al., 2009; Malone et al., 2009). It is not known if the Ago3-bound antisense piRNAs act to repress these elements. A fifth transposon family, gypsy, is thought to be active in the somatic follicle cells surrounding the oocyte where it is silenced by a mechanism that requires Piwi but not Aub or Ago3 (Li et al., 2009; Malone et al., 2009). We do not currently understand the mechanism that causes the sixth transposon family, gypsy12, to gain antisense piRNAs in qin1/ Df ovaries.
The overall increase in the fraction of piRNAs corresponding to sense sequences was reflected in the sense fraction of piRNAs bound to Aub in qin mutant ovaries: among the 93 transposon families analyzed, the median sense fraction for the piRNAs co-immunoprecipitated with Aub increased from 0.31 in qin heterozygotes to 0.42 in qin mutants (Figure 4E); the median sense fraction of Aub-bound piRNAs in wild-type Oregon R ovaries is 0.16 (Brennecke et al., 2007). In contrast, the sense fraction of piRN As bound to Piwi was essentially unaltered in qin mutant ovaries (Figure S8B). Perhaps the excess sense piRNA bound to Aub in qin mutant ovaries sequesters factors needed for antisense piRNAs to direct transposon silencing.
Qin is Required to Maintain Aub:Ago3 Ping-Pong
To understand why in qin mutants the proportion of sense piRNAs increased for so many transposon families, we examined the frequency of Ping-Pong piRNA pairs among the piRNAs bound to Aub and Ago3. The first 10 nt of Ping-Pong piRNA pairs are complementary, evidence of piRNA amplification by reciprocal cycles of cleavage by PIWI proteins, typically Aub and Ago3. Our preliminary analysis, in which we used only piRNA sequence reads unique to Ago3, Aub, or Piwi immunoprecipitates detected no significant Ping-Pong between Piwi and Ago3, Piwi and Aub, or Piwi and itself (z-scores = 2.8, i.e., p-value = 0.05). Because so many Aub-bound species are also bound to Piwi, but Piwi participates so little in Ping-Pong, we analyzed the piRNAs bound to Ago3 or Aub, excluding only species bound to both Ago3 and Aub. In contrast to our previously published Ping-Pong analyses, we computed a z-score for the occurrence of Ping-Pong piRNA pairs using a method that is uninfluenced by sequencing depth (JX and ZW, manuscript in preparation).
The Ping-Pong model for piRNA amplification suggests a simple explanation for the effect of mutation of qin: in qin mutants homotypic Ping-Pong between Aub and itself replaces heterotypic Ping-Pong between Aub and Ago3. Homotypic Ping-Pong, which occurs at low levels in control ovaries, is predicted to generate equal amounts of sense and antisense piRNAs. Thus, homotypic Ping-Pong is predicted to diminish the antisense bias of the piRNA population (Brennecke et al., 2007; Gunawardane et al., 2007; Malone et al., 2009). Aub:Aub homotypic Ping-Pong dominates in ago3 mutants, which not only disrupt piRNA amplification and produce fewer piRNAs overall, but also show increased piRNA sense fraction (Li et al., 2009).
To test if inappropriate Aub:Aub Ping-Pong replaced productive Aub:Ago3 Ping-Pong in qin mutant ovaries, we calculated Ping-Pong z-scores for Aub:Aub, Aub:Ago3, and Ago3:Ago3 pairs, using only those piRNAs that could be assigned uniquely to Ago3 or Aub. In qin heterozygotes, Ping-Pong between Aub and Ago3 predominated (Figures 5A), with far more Aub:Ago3 Ping-Pong pairs (z-score = 25) detected than Aub:Aub (z-score = 13) or Ago3:Ago3 (z-score = 3.9). In qin mutants, far fewer Aub-bound piRNAs showed the characteristic 10 nt 52 complementarity to Ago3-bound piRNAs (Aub:Ago3 z-score = 12), and Aub:Aub Ping-Pong emerged as the main pairing (Aub:Aub z-score = 26); Ago3:Ago3 Ping-Pong was lost (z-score = 1.8; Figures 5A and S8C). HeT-A, Springer, Burdock, I element, Transpac, and TAHRE, six of the seven transposons that had the greatest increase in transposon expression, all switched from Aub:Ago3 Ping-Pong to Aub:Aub Ping-Pong in qin mutants (Table S2B).
Figure 5
Figure 5
Futile Aub:Aub Homotypic Ping-Pong Prevails in qin Mutants
Intriguingly, a single transposon family, the non-long-terminal-repeat retroelement Doc, was hyper-silenced in both qin and ago3 mutant ovaries. Doc expression decreased 2.6-fold in ago3 (Li et al., 2009) and 3.7-fold in qin, even though seemingly inappropriate Aub:Aub Ping-Pong increased (the Aub:Aub z-score for Doc increased from 7.3 to 21), normally productive Aub:Ago3 Ping-Pong decreased (the z-score decreased from 12 to 4.5), the proportion of piRNAs antisense to Doc declined, the abundance of antisense piRNAs for Doc was unchanged, and the abundance of sense piRNAs more than doubled. We do not currently understand why qin and ago3 mutants show enhanced Doc silencing.
We envision that in wild -type ovaries, piRNAs shared between Aub and Ago3 correspond to secondary piRNAs generated by Aub:Aub and by Aub:Ago3 Ping-Pong, respectively. In qin mutants, Aub:Aub Ping-Pong occurs more often and Aub:Ago3 Ping-Pong occurs less frequently. Therefore, mutation of qin should decrease the number of secondary piRNAs common to Aub and Ago3. Consistent with this idea, the fraction of piRNAs shared between Ago3 and Aub declined from 37% in qin/ TM6B to 13% in qin1/Df ovaries (p-value < 2.2 × 10−16, Fisher’s exact test).
When Aub participates in heterotypic Ping-Pong, Aub-bound sense piRNAs often begin with uridine (U) but rarely contain an adenosine (A) at position 10. Primary piRNAs derived from cluster transcripts are believed to begin with U, whereas a position 10 A is the signature of a secondary piRNA produced by the Ping-Pong mechanism. (Aub-bound secondary piRNAs generated by cleavage of longer RNAs by Ago3-bound secondary piRNAs also typically begin with U, but cannot be distinguished from primary piRNAs on that basis.) We computed the nucleotide composition for all Aub- and Ago3-bound piRNAs (Figure 5B and S9A) and for the subset of piRNAs that had a Ping-Pong partner piRNA (Figure 5C). Aub-bound sense piRNAs were more likely to begin with U in qin heterozygous ovaries than in qin mutant ovaries (Figure 5B, 5C and S9A). Conversely, Aub-bound sense piRNAs were more likely to have an A at position 10 in qin mutants than in heterozygotes, consistent with the emergence in the mutant ovaries of inappropriate Aub:Aub Ping-Pong that produces Aub-bound, sense secondary piRNAs. The Aub-bound putative secondary piRNAs did not, however, favor an initial U (Figure 5C). These data suggest that the strong tendency for Aub-bound primary piRNAs to begin with U does not reflect an intrinsic property of Aub, but rather derives from a step in piRNA production before Aub loading, such as fragmentation of primary piRNA transcripts.
Aub:Ago3 Ping-Pong continues in qin mutants, but it no longer comprises the majority of Ping-Pong interactions. Ping-Pong between Aub and Ago3 declined both for Aub-bound antisense piRNAs that Ping-Pong with Ago3-bound sense piRNAs (z-score = 20.2 in qin heterozygotes; 12.5 in qin1/ Df) and for Aub-bound sense piRNAs that Ping-Pong with Ago3-bound antisense piRNAs (z-score = 16.9 qin heterozygotes; 11.4 in qin1/ Df). Despite the replacement in qin mutants of Aub:Ago3 Ping-Pong with Aub:Aub Ping-Pong as the main mechanism for piRNA amplification, Ago3-bound piRNAs that participate in Ping-Pong in qin1/ Df ovaries retained their 10A signature (Figure 5C). That is, those Ago3 piRNAs—all of which are expected to be secondary piRNAs—that are still made in qin1/Df ovaries appear to be mainly generated by Aub-catalyzed RNA cleavage directed by piRNAs bearing a U at position 1, rather than by the Aub-bound secondary piRNAs produced by Aub:Aub Ping-Pong that bear an A at position 10.
To understand why, we set out to quantify Ago3-bound tertiary piRNAs generated by Aub:Aub:Ago3 Ping-Pong. We readily detected Aub-bound piRNAs that bear an A at position 10 and Ping-Pong with Ago3-bound piRNAs. These Ago3-bound piRNAs would be expected to begin with U, but might also fortuitously contain an A at position 10, making it difficult to determine whether they represent the product of an Aub-bound primary piRNA (i.e., U at position 1) or an Aub-bound secondary piRNA generated by Aub:Aub Ping-Pong (i.e., an A at position 10). To avoid this ambiguity, we restricted our analysis to Ago3-bound piRNAs that begin with U but bear a C, G, or U but not an A at position 10. Such “1U, non-10A” piRNAs comprised ~4.9% of all Ago3-bound piRNAs possessing Ping-Pong partners in qin heterozygotes, but encompassed ~8.2% of all Ago3-bound piRNAs with Ping-Pong partners in qin1/ Df mutant ovaries. We conclude that the increase in Aub-bound secondary piRNAs generated by inappropriate Aub:Aub Ping-Pong in qin mutants leads to a corresponding increase in Ago3-bound tertiary piRNAs generated by Aub:Aub:Ago3 Ping-Pong. These tertiary piRNAs do not generate a 1U signature in the Ago3 piRNA sequence logos, likely because most Ago3-bound piRNAs are still produced by Aub-bound primary piRNAs even in qin1/ Df mutant ovaries.
Association of Ago3 with Aub Requires Qin
Ago3 co-immunoprecipitates with Aub from ovary lysate, suggesting that at least a fraction of Ago3 is bound to or present in a common complex with Aub (Nishida et al., 2009). We immunoprecipitated Aub from ovary lysate prepared from flies bearing a transgene expressing FLAG-Myc–tagged Ago3 (FM-Ago3), then measured the amount of co-immunoprecipitated FM-Ago3 by Western blotting using anti-FLAG antibody (Figure 6A). The association of Ago3 with Aub was not bridged by RNA: pre-incubation of the lysate with RNase A did not affect the co-association of FM-Ago3 with Aub (Figure 6B); control experiments demonstrated that the RNase treatment reduced miR-317 levels to background and reduced 2S rRNA, a highly structured RNA component of the ribosome, by 15-fold (Figure 6C). Qin was required for the association of FM-Ago3 with Aub. Compared to qin heterozygotes, > 6-fold less FM-Ago3 co-immunoprecipitated with Aub in lysate from qin1/ Df ovaries (mean ± standard deviation of co-immunoprecipitated FM-Ago3 in qin mutant ovaries was 16 ± 12% of the heterozygous control, n = 3; Figures 6D, S9B, S9C, and S9D).
Figure 6
Figure 6
The Association of Ago3 with Aub Requires Qin but not RNA
Compared with the 11 Drosophila piRNA pathway mutants previously subjected to high throughput sequencing analysis, qin is unique (Figure 7). Like qin, mutations in piwi or zucchini increase overall Ping-Pong z-score and decrease antisense piRNA abundance, but neither piwi nor zucchini mutants increase the abundance of sense piRNAs. Moreover, Piwi and Zucchini appear to function mainly in the somatic follicle cells, where piRNAs are not amplified by the Ping-Pong cycle, so the increase in Ping-Pong observed in piwi and zucchini mutants largely reflects the loss of somatic piRNAs, rather than a direct effect on germline piRNAs.
Figure 7
Figure 7
Among piRNA Pathway Mutants, qin Has a Unique Effect on the Structure of the Ovarian piRNA Population
Mutation of ago3, aub, armitage, vasa, krimper, vret, rhino, spindle-E, or squash disrupts the Ping-Pong mechanism, with mutations in ago3, aub, vasa, krimper, vret, rhino, and spindle-E eliminating Ping-Pong piRNA amplification for most or all transposon families. We note that krimper, a Tudor-domain protein, localizes to nuage like Qin, but plays a very different role in piRNA biogenesis from Qin: Ping-Pong amplification collapses in krimper mutants, whereas mutation of qin dramatically increases piRNA Ping-Pong by triggering non-productive Aub:Aub Ping-Pong. While the overall abundance of piRNAs is preserved in qin mutant ovaries, piRNA antisense bias declines, largely because of an increase in the abundance of sense piRNAs.
The replacement of heterotypic Aub:Ago3 Ping-Pong with homotypic Aub:Aub Ping-Pong in qin mutants suggests either that Qin acts to suppress homotypic Ping-Pong or that Qin promotes heterotypic Aub:Ago3 Ping-Pong. Our data cannot distinguish between these two models. Mutation of qin disrupts the interaction of Ago3 with Aub in ovary lysate. If Qin acts to suppress Aub self-association, the decrease in Ago3 bound to Aub might reflect the redirection of Aub from Ago3 to Aub itself. Alternatively, if Qin acts—directly or indirectly—to promote the binding of Aub to Ago3, a decrease in Qin function would lead to futile homotypic Ping-Pong by default, especially if the intracellular concentration of Aub were greater than that of Ago3.
In contrast to the germline piRNAs of flies, which require Aub:Ago3 Ping-Pong to silence transposons during gamete development, piRNA production in the somatic follicle cells of the ovaries of Drosophilidae and of piRNAs targeting non-repetitive sequences during the pachytene phase of mammalian spermatogenesis require just a single PIWI protein. These piRNA pathways proceed without Ping-Pong amplification (Aravin et al., 2007; Lau et al., 2009; Li et al., 2009; Malone et al., 2009; Robine et al., 2009; Saito et al., 2009). We do not know if Qin in flies or Qin-like proteins in other species play a role in the production of this second type of piRNA. It is intriguing that the mouse protein Tdrd4 (also called RNF17) resembles Qin. Like Qin, Tdrd4 contains amino-terminal E3 ligase motifs and five carboxy-terminal Tudor domains. While Tdrd4 has yet to be implicated in piRNA biogenesis or function, it is required for normal mouse spermatogenesis and localizes to nuage (Pan et al., 2005).
Heterotypic Ping-Pong between pairs of PIWI proteins drives the amplification of antisense piRNAs, the species believed to silence transposon expression during gamete formation in insects and likely many other invertebrates and during the prenatal development of the testes in mammals (Brennecke et al., 2007; Gunawardane et al., 2007; Aravin et al., 2008; Li et al., 2009; Malone et al., 2009). Our data, however, suggest that the proposal that antisense piRNA abundance alone determines the extent of repression of transposon families is too simple to explain how piRNAs silence transposons. We find that neither normal amounts of the PIWI proteins Aub and Ago3 nor the organization of these proteins into nuage suffices to promote productive amplification of silencing-competent antisense piRNAs. Nor are near wild -type levels of antisense piRNAs sufficient to ensure adequate silencing of active transposons during Drosophila oogenesis. Instead, effective transposon silencing requires that Aub partner with Ago3, rather than itself, generating more Aub-bound antisense piRNAs than sense. We have shown here that this heterotypic partnership requires Qin. Key challenges for the future will be to determine how Qin promotes heterotypic Ping-Pong or represses homotypic Ping-Pong and what role its E3 ligase and Tudor domains play in this process.
General Methods
Female fertility tests, tiling arrays, immunoprecipitation, immunoblotting, and quantitative RT-PCR were as described (Li et al., 2009). Small RNA data for total small RNAs in cn1; ry506 ovaries (Klattenhoff et al., 2009) was previously deposited in the NCBI trace archives (accession number SRP002060); tiling microarray data (Fig. 2, C–D) for total RNA in w1, aubHN2/QC42 and ago3t2/t3 ovaries (Klattenhoff et al., 2009; Li et al., 2009) was previously deposited in the NCBI gene expression omnibus (accession number GSE14370). Small RNA data for Figure 7 was previously deposited in the NCBI trace archives SRP000458 (Li et al., 2009), GEO GSE15186 (Malone et al., 2009), and GSE30088 (Zamparini et al., 2011). Aub- and Ago3-associated small RNA data from Oregon R was published previously (Brennecke et al., 2007). Unless otherwise specified, p-values were calculated from at least three independent biological replicates using a two-tailed, two-sample unequal variance t-test (Excel, Microsoft).
Drosophila stocks
All flies were raised at 25°C. PBac{RB}CG14303e03728 (qin1) and Df flies were from the Bloomington Stock Center at Indiana University; PBac{RB}e01936 (qin2) was from the Harvard Medical School stock center. P{w+mc, UASp-FM-Ago3-C2} (Chengjian Li and PDZ, unpublished) was used to express Ago3.
Bioinformatic analyses
Analysis was as described (Li et al., 2009) except for the computation of Ping-Pong z-scores. For two piRNAs that were sufficiently complementary to each other at a particular 52-to-52 distance, a score was defined as the product of their abundances. The Ping-Pong z-score was then the difference of the score at the 52-to-52 distance of 10 nt and the mean scores of background distances, divided by the standard deviation of the scores of background distances, defined as distances of 0–9 and 11–20 nt. Two piRNAs were sufficiently complementary to each other when the nucleotides 2–10 of the first piRNA were perfectly paired with the second piRNA and there was at most one mismatch among positions 1 and 11–16 of the first piRNA. Genomic sequence adjacent to the second piRNA was used to determine complementarity when the 52-to-52 distance was less than 16 nt.
Accession Numbers
Sequence data generated in this study are available via the NCBI trace archives (http://www.ncbi.nlm.nih.gov/Traces/) with accession number SRP007101. Microarray data generated in this study are available via the NCBI gene expression omnibus (http://www.ncbi.nlm.nih.gov/geo/) as GSE30061.
Highlights
  • Qin silences transposons and ensures genome stability in the Drosophila germline.
  • Qin maintains antisense piRNA populations.
  • Qin comprises an amino-terminal E3 ligase motif and five carboxy-terminal Tudor domains.
  • In qin mutants, futile Aub:Aub Ping-Pong replaces productive Aub:Ago3 Ping-Pong.
Supplementary Material
01
Acknowledgments
We thank A. Boucher and G. Farley for technical assistance, J. Brennecke, G. Hannon, A. Hyman, and M. Siomi for reagents, and members of our laboratories for advice and critical comments on the manuscript. This work was supported in part by National Institutes of Health grant HD049116 to W.E.T., Z.W., and P.D.Z. and GM62862 and GM65236 to P.D.Z.
Footnotes
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
  • Abdu U, Brodsky M, Schupbach T. Activation of a meiotic checkpoint during Drosophila oogenesis regulates the translation of Gurken through Chk2/ Mnk. Curr Biol. 2002;12:1645–1651. [PubMed]
  • Aravin A, et al. A novel class of small RNAs bind to MILI protein in mouse testes. Nature. 2006;442:203–207. [PubMed]
  • Aravin AA, Klenov MS, Vagin VV, Bantignies F, Cavalli G, Gvozdev VA. Dissection of a natural RNA silencing process in the Drosophila melanogaster germ line. Mol Cell Biol. 2004;24:6742–6750. [PMC free article] [PubMed]
  • Aravin AA, Sachidanandam R, Bourc’his D, Schaefer C, Pezic D, Toth KF, Bestor T, Hannon GJ. A piRNA pathway primed by individual transposons is linked to de novo DNA methylation in mice. Mol Cell. 2008;31:785–799. [PMC free article] [PubMed]
  • Aravin AA, Sachidanandam R, Girard A, Fejes-Toth K, Hannon GJ. Developmentally regulated piRNA clusters implicate MILI in transposon control. Science. 2007;316:744–747. [PubMed]
  • Aravin AA, Naumova NM, Tulin AV, Vagin VV, Rozovsky YM, Gvozdev VA. Double-stranded RNA-mediated silencing of genomic tandem repeats and transposable elements in the D. melanogaster germline. Curr Biol. 2001;11:1017–1027. [PubMed]
  • Balakireva MD, Shevelyov Y, Nurminsky DI, Livak KJ, Gvozdev VA. Structural organization and diversification of Y-linked sequences comprising Su(Ste) genes in Drosophila melanogaster. Nucleic Acids Res. 1992;20:3731–3736. [PMC free article] [PubMed]
  • Brennecke J, Aravin AA, Stark A, Dus M, Kellis M, Sachidanandam R, Hannon GJ. Discrete small RNA-generating loci as master regulators of transposon activity in Drosophila. Cell. 2007;128:1089–1103. [PubMed]
  • Brennecke J, Malone CD, Aravin AA, Sachidanandam R, Stark A, Hannon GJ. An epigenetic role for maternally inherited piRNAs in transposon silencing. Science. 2008;322:1387–1392. [PMC free article] [PubMed]
  • Chen Y, Pane A, Schupbach T. cutoff and aubergine mutations result in retrotransposon upregulation and checkpoint activation in Drosophila. Curr Biol. 2007;17:637–642. [PMC free article] [PubMed]
  • Cox DN, Chao A, Baker J, Chang L, Qiao D, Lin H. A novel class of evolutionarily conserved genes defined by piwi are essential for stem cell self-renewal. Genes Dev. 1998;12:3715–3727. [PubMed]
  • Cox DN, Chao A, Lin H. piwi encodes a nucleoplasmic factor whose activity modulates the number and division rate of germline stem cells. Development. 2000;127:503–514. [PubMed]
  • Desset S, Buchon N, Meignin C, Coiffet M, Vaury C. In Drosophila melanogaster the COM Locus Directs the Somatic Silencing of Two Retrotransposons through both Piwi-Dependent and -Independent Pathways. PLoS ONE. 2008;3:e1526. [PMC free article] [PubMed]
  • Drysdale RA, et al. FlyBase: genes and gene models. Nucleic Acids Res. 2005;33:D390–5. Database Issue. [PMC free article] [PubMed]
  • Ghabrial A, Schupbach T. Activation of a meiotic checkpoint regulates translation of Gurken during Drosophila oogenesis. Nat Cell Biol. 1999;1:354–357. [PubMed]
  • Ghildiyal M, Zamore PD. Small silencing RNAs: an expanding universe. Nat Rev Genet. 2009;10:94–108. [PMC free article] [PubMed]
  • Girard A, Sachidanandam R, Hannon GJ, Carmell MA. A germline-specific class of small RNAs binds mammalian Piwi proteins. Nature. 2006;442:199–202. [PubMed]
  • Grivna ST, Beyret E, Wang Z, Lin H. A novel class of small RNAs in mouse spermatogenic cells. Genes Dev. 2006;20:1709–1714. [PubMed]
  • Gunawardane LS, Saito K, Nishida KM, Miyoshi K, Kawamura Y, Nagami T, Siomi H, Siomi MC. A Slicer-Mediated Mechanism for Repeat-Associated siRNA 5′ End Formation in Drosophila. Science. 2007;315:1587–1590. [PubMed]
  • Hardy RW, Lindsley DL, Livak KJ, Lewis B, Siversten AL, Joslyn GL, Edwards J, Bonaccorsi S. Cytogenetic Analysis of a Segment of the Y Chromosome of Drosophila melanogaster. Genetics. 1984;107:591–610. [PubMed]
  • Hardy RW. Crystal aggregates in the primary spermatocytes of XO males in D. melanogaster. Drosophila Inform Serv. 1980;55:54–55.
  • Harris AN, Macdonald PM. aubergine encodes a Drosophila polar granule component required for pole cell formation and related to eIF2C. Development. 2001;128:2823–2832. [PubMed]
  • Houwing S, et al. A role for Piwi and piRNAs in germ cell maintenance and transposon silencing in Zebrafish. Cell. 2007;129:69–82. [PubMed]
  • Kirino Y, Kim N, de Planell-Saguer M, Khandros E, Chiorean S, Klein PS, Rigoutsos I, Jongens TA, Mourelatos Z. Arginine methylation of Piwi proteins catalysed by dPRMT5 is required for Ago3 and Aub stability. Nat Cell Biol. 2009;11:652–658. [PMC free article] [PubMed]
  • Kirino Y, Mourelatos Z. Mouse Piwi-interacting RNAs are 2′-O-methylated at their 3′ termini. Nat Struct Mol Biol. 2007;14:347–348. [PubMed]
  • Klattenhoff C, Bratu DP, McGinnis-Schultz N, Koppetsch BS, Cook HA, Theurkauf WE. Drosophila rasiRNA pathway mutations disrupt embryonic axis specification through activation of an ATR/ Chk2 DNA damage response. Dev Cell. 2007;12:45–55. [PubMed]
  • Klattenhoff C, Theurkauf W. Biogenesis and germline functions of piRNAs. Development. 2008;135:3–9. [PubMed]
  • Klattenhoff C, et al. The Drosophila HP1 homolog Rhino is required for transposon silencing and piRNA production by dual-strand clusters. Cell. 2009;138:1137–1149. [PMC free article] [PubMed]
  • Lau NC, Robine N, Martin R, Chung WJ, Niki Y, Berezikov E, Lai EC. Abundant primary piRNAs, endo-siRNAs, and microRNAs in a Drosophila ovary cell line. Genome Res. 2009;19:1776–1785. [PubMed]
  • Lau NC, Seto AG, Kim J, Kuramochi-Miyagawa S, Nakano T, Bartel DP, Kingston RE. Characterization of the piRNA complex from rat testes. Science. 2006;313:363–367. [PubMed]
  • Li C, et al. Collapse of Germline piRNAs in the Absence of Argonaute3 Reveals Somatic piRNAs in Flies. Cell. 2009;137:509–521. [PMC free article] [PubMed]
  • Lim AK, Kai T. Unique germ-line organelle, nuage, functions to repress selfish genetic elements in Drosophila melanogaster. Proc Natl Acad Sci U S A. 2007;104:6714–6719. [PubMed]
  • Liu L, Qi H, Wang J, Lin H. PAPI, a novel TUDOR-domain protein, complexes with AGO3, ME31B and TRAL in the nuage to silence transposition. Development. 2011;138:1863–1873. [PubMed]
  • Livak KJ. Organization and mapping of a sequence on the Drosophila melanogaster X and Y chromosomes that is transcribed during spermatogenesis. Genetics. 1984;107:611–634. [PubMed]
  • Livak KJ. Detailed structure of the Drosophila melanogaster Stellate genes and their transcripts. Genetics. 1990;124:303–316. [PubMed]
  • Malone CD, Brennecke J, Dus M, Stark A, McCombie WR, Sachidanandam R, Hannon GJ. Specialized piRNA Pathways Act in Germline and Somatic Tissues of the Drosophila Ovary. Cell. 2009;137:522–535. [PMC free article] [PubMed]
  • Mevel-Ninio M, Pelisson A, Kinder J, Campos AR, Bucheton A. The flamenco locus controls the gypsy and ZAM retroviruses and is required for Drosophila oogenesis. Genetics. 2007;175:1615–1624. [PubMed]
  • Ni JQ, et al. A genome-scale shRNA resource for transgenic RNAi in Drosophila. Nat Methods. 2011;8:405–407. [PubMed]
  • Nishida KM, et al. Functional involvement of Tudor and dPRMT5 in the piRNA processing pathway in Drosophila germlines. EMBO J. 2009;28:3820–3831. [PubMed]
  • Pan J, Goodheart M, Chuma S, Nakatsuji N, Page DC, Wang PJ. RNF17, a component of the mammalian germ cell nuage, is essential for spermiogenesis. Development. 2005;132:4029–4039. [PMC free article] [PubMed]
  • Parks AL, et al. Systematic generation of high–resolution deletion coverage of the Drosophila melanogaster genome. Nat Genet. 2004;36:288–292. [PubMed]
  • Patil VS, Kai T. Repression of Retroelements in Drosophila Germline via piRNA Pathway by the Tudor Domain Protein Tejas. Curr Biol. 2010;20:724–730. [PubMed]
  • Poser I, et al. BAC TransgeneOmics: a high-throughput method for exploration of protein function in mammals. Nat Methods. 2008;5:409–415. [PMC free article] [PubMed]
  • Robine N, Lau NC, Balla S, Jin Z, Okamura K, Kuramochi-Miyagawa S, Blower MD, Lai EC. A broadly conserved pathway generates 32 UTR-directed primary piRNAs. Curr Biol. 2009;19:2066–2076. [PMC free article] [PubMed]
  • Saito K, Inagaki S, Mituyama T, Kawamura Y, Ono Y, Sakota E, Kotani H, Asai K, Siomi H, Siomi MC. A regulatory circuit for piwi by the large Maf gene traffic jam in Drosophila. Nature. 2009;461:1296–1299. [PubMed]
  • Saito K, Nishida KM, Mori T, Kawamura Y, Miyoshi K, Nagami T, Siomi H, Siomi MC. Specific association of Piwi with rasiRNAs derived from retrotransposon and heterochromatic regions in the Drosophila genome. Genes Dev. 2006;20:2214–2222. [PubMed]
  • Schneider TD, Stephens RM. Sequence logos: a new way to display consensus sequences. Nucleic Acids Res. 1990;18:6097–6100. [PMC free article] [PubMed]
  • Schupbach T, Wieschaus E. Female sterile mutations on the second chromosome of Drosophila melanogaster. I. Maternal effect mutations. Genetics. 1989;121:101–117. [PubMed]
  • Schupbach T, Wieschaus E. Female sterile mutations on the second chromosome of Drosophila melanogaster. II. Mutations blocking oogenesis or altering egg morphology. Genetics. 1991;129:1119–1136. [PubMed]
  • Siomi MC, Mannen T, Siomi H. How does the royal family of Tudor rule the PIWI-interacting RNA pathway? Genes Dev. 2010;24:636–646. [PubMed]
  • Siomi MC, Sato K, Pezic D, Aravin AA. PIWI-interacting small RNAs: the vanguard of genome defence. Nat Rev Mol Cell Biol. 2011;12:246–258. [PubMed]
  • Smith CD, Shu S, Mungall CJ, Karpen GH. The Release 5.1 annotation of Drosophila melanogaster heterochromatin. Science. 2007;316:1586–1591. [PMC free article] [PubMed]
  • Snee MJ, Macdonald PM. Live imaging of nuage and polar granules: evidence against a precursor-product relationship and a novel role for Oskar in stabilization of polar granule components. J Cell Sci. 2004;117:2109–2120. [PubMed]
  • Theurkauf WE, Klattenhoff C, Bratu DP, McGinnis-Schultz N, Koppetsch BS, Cook HA. rasiRNAs, DNA damage, and embryonic axis specification. Cold Spring Harb Symp Quant Biol. 2006;71:171–180. [PubMed]
  • Thibault ST, et al. A complementary transposon tool kit for Drosophila melanogaster using P and piggyBac. Nat Genet. 2004;36:283–287. [PubMed]
  • Vagin VV, Sigova A, Li C, Seitz H, Gvozdev V, Zamore PD. A distinct small RNA pathway silences selfish genetic elements in the germline. Science. 2006;313:320–324. [PubMed]
  • Vagin VV, Wohlschlegel J, Qu J, Jonsson Z, Huang X, Chuma S, Girard A, Sachidanandam R, Hannon GJ, Aravin AA. Proteomic analysis of murine Piwi proteins reveals a role for arginine methylation in specifying interaction with Tudor family members. Genes Dev. 2009;23:1749–1762. [PubMed]
  • Venken KJ, Carlson JW, Schulze KL, Pan H, He Y, Spokony R, Wan KH, Koriabine M, De Jong PJ, White KP, Bellen HJ, Hoskins RA. Versatile P[acman] BAC libraries for transgenesis studies in Drosophila melanogaster. Nat Methods. 2009;6:431–434. [PMC free article] [PubMed]
  • Wang J, Saxe JP, Tanaka T, Chuma S, Lin H. Mili Interacts with Tudor Domain-Containing Protein 1 in Regulating Spermatogenesis. Curr Biol. 2009;19:640–644. [PMC free article] [PubMed]
  • Warming S, Costantino N, Court DL, Jenkins NA, Copeland NG. Simple and highly efficient BAC recombineering using galK selection. Nucleic Acids Res. 2005;33:e36. [PMC free article] [PubMed]
  • Wilson JE, Connell JE, Macdonald PM. aubergine enhances oskar translation in the Drosophila ovary. Development. 1996;122:1631–1639. [PubMed]
  • Zamparini AL, Davis MY, Malone CD, Vieira E, Zavadil J, Sachidanandam R, Hannon GJ, Lehmann R. Vreteno, a gonad-specific protein, is essential for germline development and primary piRNA biogenesis in Drosophila. Development. 2011;138:4039–50. [PMC free article] [PubMed]