Function of Piwi, a nuclear Piwi/ , is independent of its slicer activity

Nicole Darricarrère, Na Liu, Toshiaki Watanabe, and Haifan Lin1

Yale Center and Department of Cell Biology, Yale University School of Medicine, New Haven, CT 06509

Edited by James A. Birchler, University of Missouri, Columbia, MO, and approved December 11, 2012 (received for review August 1, 2012) The Piwi protein subfamily is essential for Piwi-interacting RNA classical epigenetic phenotypes in Drosophila (15, 17, 18). These (piRNA) biogenesis, transposon silencing, and germ-line develop- observations led us to propose that piRNA molecules can guide ment, all of which have been proposed to require Piwi Piwi to specific genomic loci by sequence complementarity, activity, as validated for two cytoplasmic Piwi in mice. where Piwi recruits epigenetic modifiers (17–19). In this epige- However, recent evidence has led to questioning of the generality netic model, the slicer activity of Piwi seems superfluous. Indeed, of this mechanism for the Piwi members that reside in the nucleus. the involvement of the slicer activity of Piwi has been questioned Drosophila offers a distinct opportunity to study the function of recently based on experiments performed in an ovarian somatic nuclear Piwi proteins because, among three Drosophila Piwi pro- cell line (20). However, this analysis was not definitive. It was teins—called Piwi, Aubergine, and Argonaute 3—Piwi is the only based on a single cultured cell type, relied on RNAi that only member of this subfamily that is localized in the nucleus and incompletely reduced endogenous levels of Piwi and tested only expressed in both germ-line and somatic cells in the gonad, where one transposon element. Thus, these results could not predict it is responsible for piRNA biogenesis and regulatory functions es- the consequence of abolishing the slicer function in an organism. sential for fertility. In this study, we demonstrate beyond doubt that Therefore, we sought to test the role of the endonuclease activity the slicer activity of Piwi is not required for any known functions in of Piwi in all known biological processes that involve Piwi fun- vivo. We show that, in transgenic flies with the DDX catalytic triad ction, such as piRNA biogenesis, transposon silencing, germ-line of PIWI mutated, neither primary nor secondary piRNA biogenesis is development, and gametogenesis. detectably affected, transposons remain repressed, and fertility is normal. Our observations demonstrate that the mechanism of Piwi Results BIOCHEMISTRY is independent of its in vitro endonuclease activity. Instead, it is Slicer Mutations Do Not Affect the Stability or Localization of Piwi consistent with the alternative mode of Piwi function as a molecule in Vivo. We replaced the two critical aspartate residues, D614 and involved in the piRNA-directed guidance of epigenetic factors D685, that form the catalytic triad with alanine, as reported in to chromatin. vitro (16) (Fig. 1A). We used FlyC31 site-specific recombination technology to introduce the slicer-deficient Myc–Piwi transgene small RNA | RNase H fold | transposable element | sterility (“DUO,” which means “two” in Latin, herein refers to two-point mutations) or a wild-type Myc–Piwi control transgene (WT) in- he Piwi protein family plays a fundamental role in sexual to a well-characterized genomic site called attP2 to generate – Treproduction. Not only are Piwi proteins highly enriched in transgenic Drosophila strains (21). The WT Myc Piwi has the gonads of all animals studied to date, but their reproductive been shown to be fully functional compared with the endogenous phenotypes have been revealed in diverse model organisms, such piwi gene (22). The insertion of the DUO and WT transgenes as Caenorhabditis elegans, Drosophila, zebrafish, and mice (1). precisely at the same site in the eliminates potentially fl Piwi proteins are paralogs of Argonaute proteins, sharing the same different in uences of the surrounding genomic context toward domain architecture and the property of binding small noncoding the expression of the inserted transgenes. Additionally, we gen- – erated another DUO Myc–Piwi transgenic line by conventional (2 6). Argonaute proteins bind to small interfering RNAs – (siRNAs) or (miRNAs) and are effectors of RNA P-element transformation, with a WT Myc Piwi transgenic line (G38) previously generated the same way as controls (15). The interference (RNAi) and miRNA pathways; conversely, Piwi pro- – teins bind to Piwi-interacting RNAs (piRNAs) and are involved in expression levels of WT and DUO Myc Piwi proteins in ovaries are very similar among all transgenic lines (Fig. 1B). Likewise, the piRNA pathway in both the germ line and soma during ga- fi – metogenesis (1, 7, 8). the localization pattern of slicer-de cient Myc Piwi (green) both — in the germ-line (red for Vasa, a germ-line marker) and in so- In Drosophila, there are three Piwi proteins Piwi, Aubergine – (Aub), and Argonaute 3 (Ago3) (9)—all of which have been matic cells was indistinguishable from that of WT Myc Piwi (Fig. 1 C–J). Together, these results confirm that these two mutations shown to possess small RNA-dependent endonuclease (slicer) have no effect on the stability or localization of Piwi in vivo. activity by in vitro assays (10, 11), catalyzed by an aspartate cat- alytic triad (DDX) that is positioned within the core of the RNase None of the Essential Functions of Piwi in Somatic and Germ-Line H-fold (12). This activity has been proposed to be essential for the fi Cells Requires its Slicer Activity. To assess the biological role of ampli cation of piRNAs through a mechanism called the ping- the slicer activity of Piwi, we crossed the DUO and WT transgenic pong cycle that occurs in the cytoplasm and uses cRNA targets, such as transposon transcripts, as precursors (13, 14). In addition, the slicer activity has been proposed play a role in silencing target Author contributions: N.D. and H.L. designed research; N.D., N.L., and T.W. performed transposons through cleaving their transcripts (13). research; N.D. and H.L. contributed new reagents/analytic tools; N.D., N.L., T.W., and H.L. Among the three Piwi proteins in Drosophila, only Piwi is analyzed data; and N.D. and H.L. wrote the paper. expressed in both germ-line and somatic cells in the gonad and is The authors declare no conflict of interest. involved in both germ-line and somatic piRNA pathways (15, This article is a PNAS Direct Submission. 16). Piwi is also distinct from the other members of the subfamily Data deposition: The data reported in this paper have been deposited in the Gene in terms of subcellular localization. We showed previously that Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession no. GSE34032). Piwi is a nuclear protein that associates with an epigenetic factor, 1To whom correspondence should be addressed. E-mail: [email protected]. Protein 1a, and binds to piRNA-complementary This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. sequences in the genome and that Piwi deficiencies lead to 1073/pnas.1213283110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1213283110 PNAS Early Edition | 1of6 Downloaded by guest on September 26, 2021 myc-tagged Piwi ABno myc control wildtype slicer mutant Myc-tag W[1118] G38 WT WTDUO DUO D614A D685A N-terminal PAZ MID PIWI myc

β-Actin 0 0 CEDF0 075 µm µm µm µm 75 75 50 0 0 0 GH0 I J µm µm µm µm 75 75 Myc 75 50 VASA DAPI

W[1118] +/+, WT/WT +/+, DUO/DUO +/+, D37/D37

Fig. 1. Slicer-deficient Piwi shows a normal level of expression and pattern of localization. (A) Schematic representation of the domain architecture of PIWI, highlighting the two critical aspartate residues essential for endonuclease activity, which were replaced by alanine in the slicer-deficient transgene. (B) Western blot showing similar levels of WT and DUO Piwi proteins in ovaries. The wild-type nontransgenic line (W1118) is a negative control. Myc–Piwi transgenic line G38 was described (15). Protein levels of two independent site-specific transgenics are shown for wildtype Myc–Piwi at the attP2 site (WT) and + Myc–Piwi slicer mutant (DUO). All transgenes are in the W1118;piwi background. (C–J) Immunofluorescence micrographs of ovarioles from transgenic flies. Myc (green) was costained with Vasa (red) and DAPI (blue). C and G show no Myc staining in Myc-negative W1118 ovaries. D and H show the normal lo- calization of WT Myc–Piwi, whereas E and I show the normal localization of DUO Myc–PIWI (D614A, D685A). F and J show the localization of D37 Myc–PIWI (D614A, D685A).

lines into the homozygous piwi1 mutant background as outlined 3A) and normalized to protein loading controls as quantified by in Fig. S1. To our surprise, ovaries from the resulting piwi1/piwi1; Western blotting (Fig. 3 B and C). By comparing the pull-down DUO/DUO genotype are morphologically indistinguishable from efficiency of DUO and WT Myc–Piwi in a wild-type background, piwi1/piwi1;WT/WT and wild-type flies (Fig. 2 A–F), as indicated we confirmed that the slicer point mutations do not impair piRNA by normal 1B1 staining (green) that outlines somatic cells and binding in vivo. Furthermore, the amount of piRNAs relative to highlights spectrosomes and fusomes in the germ line as well as Piwi protein does not differ between slicer-deficient and wild-type Vasa staining (red) that labels germ-line cells. We then quantified strains in the piwi1 background. the fertility of the slicer-deficient flies by counting the number of To further examine whether the slicer mutant Piwi is defective eggs laid per fly per day. These flies laid as many eggs as the wild- in piRNA biogenesis, we deep-sequenced four piRNA libraries type control (Fig. 2G). Furthermore, the progeny of the slicer- prepared from either total small RNAs or Piwi coimmuno- deficient flies were as viable as those of the wild-type control (Fig. precipitated samples from piwi1/piwi1;WT/WT and piwi1/piwi1; 2H). In fact, we have been keeping piwi1/piwi1;DUO/DUO flies as DUO/DUO genotypes, which we refer to as Total-WT, Total- a healthy stock since November 2009. Thus, the piwi1/piwi1;DUO/ DUO, IP–WT and IP–DUO, respectively. DUO and WT ovaries DUO flies without slicer activity have no detectable defect or show similar distribution of both primary and ping-pong–derived reproductive disadvantage. These results indicate that all essen- secondary piRNA sequences in both Total and IP libraries (Fig. tial functions of Piwi in both somatic and germ-line cells are in- 3D and Fig. S2). Approximately 13% of sequences from Total dependent of its slicer activity. RNA libraries are miRNAs. In contrast, both IP-WT and -DUO libraries only retained ∼1% of miRNA sequences, which could be Slicer Activity of Piwi Is Not Required for piRNA Biogenesis. To di- Piwi-associated miRNAs (23). Regardless of whether the ∼1% of rectly test the involvement of the slicer activity of Piwi in piRNA miRNAs reflect nonspecific background, this finding shows the biogenesis, we purified Piwi–piRNA complexes from piwi1/piwi1; similar effectiveness of both WT and DUO Piwi proteins in DUO/DUO and piwi1/piwi1;WT/WT ovaries by immunoprecipi- binding to piRNAs. In both IP–WT and IP–DUO libraries, ∼40% tation using a high-affinity and high-specificity Myc antibody. of piRNAs are comprised of transposon-derived piRNAs. After Piwi-associated piRNAs were visualized by RNA-PAGE (Fig. bioinformatic subtraction of spurious cellular RNA contaminants

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1213283110 Darricarrère et al. Downloaded by guest on September 26, 2021 A Germline stem cells (GSC) B 050µm C 050µm G 80 slicer mutants Early Fusome GSC 60 n.s. Somatic NC Follicle Cells (FC) 40 FC Developing 20 Oocyte 1B1 Nurse Cells 0 VASA eggs laidper fly per day WT DUO D37 Egg development (NC) DAPI rescue of piwi null Late Developing Oocyte W[1118] piwi1/piwi1

µm µm µm DE075 GSC 075 F075H 80% slicer mutants GSC n.s. NC GSC Fig. 2. Slicer-deficient piwi mutants display normal 60% germ-line stem cell division, oogenesis, and viability. FC FC fl – NC (A) Schematic diagram of a normal y ovariole. (B F) Immunofluorescence images of ovarioles dissected NC 40% from W[1118] strain (B), piwi1/piwi1 mutant (C), piwi1/ FC piwi1;WT/WT (D), piwi1/piwi1;DUO/DUO (E), and piwi1/ 1 Viable eggs eggs Viable 20% piwi ;D37/D37 (F). Vasa staining (red) highlights germ- line cells. 1B1 staining (green) outlines somatic cells, spectrosomes, and fusomes. DAPI reveals nuclear 0% morphology. (G and H) Average fertility (G)andvia- WT DUO D37 bility (H)oftransgenicrescuefly strains are graphed 1 with SE indicated. piwi mutant flies fail to lay any egg BIOCHEMISTRY 1 1 1 1 1 1 rescue of piwi null piwi /piwi ,WT/WT piwi /piwi , DUO/DUO piwi /piwi , D37/D37 due to ovarian defects as shown in C.

and miRNAs, we observed in both total small RNA libraries two piRNA corresponding transposons HOBO, LOOPER1, POGON1, peaks typical of piRNA profiles at 21 and 25 nt, corresponding to and POGO (Fig. 3K, at positions 4, 7, 8, and 10 of the horizon- the sizes of endo-siRNAs and piRNAs, respectively (Fig. 3D). tal axis) are decreased to ∼50% of the original levels. However, Furthermore, the IP–WT and IP–DUO piRNAs showed almost this decrease is not seen in PIWI-associated piRNAs corre- identical peaks at 26 nt—the expected peak size of Piwi-associated sponding to these transposons (Fig. 3J), and its impact on piRNAs. These results indicate that the slicer mutation has no fertility and viability is undetectable and therefore incon- visible effect on the biogenesis of either the PIWI-associated or sequential. To further verify our conclusion, we compared the total piRNA populations. abundance of piRNAs associated with 140 transposons in DUO and WT flies to those in Yb and papi mutants by the same bio- Slicer Activity of Piwi Is Not Required for piRNA Amplification via the informatics analysis (Fig. S3). The Yb mutation caused significant Ping-Pong Mechanism. To further resolve whether the slicer activity decrease of piwi-bound piRNAs, whereas the papi mutation did not of Piwi contribute to the ping-pong mechanism, we analyzed the show any detectable effect. We observed that the piRNA abun- sequence composition of the piRNA pools from the dance in the DUO mutant is very similar to WT and less variable four libraries. The WT and DUO libraries show a striking re- than that of the papi mutant, let alone the Yb mutant. Together, semblance of nucleotide composition at all corresponding posi- our results indicate that the slicer activity of Piwi is not involved in tions. In particular, the first nucleotide exhibits strong uracil bias the biogenesis of primary or secondary piRNAs in either somatic (Fig. 3 E–H). In addition, both total small-RNA libraries dis- or germ-line cells. played a similar and weak adenine bias in the 10th nucleotide (Fig. 3 G and H), as expected for ping-pong–generated piRNAs. Slicer Activity of Piwi Is Not Required for Transposon Silencing. Expectedly, such a bias does not exist in either IP–WT or IP– Lastly, we assessed whether the slicer activity of Piwi is in- DUO libraries (Fig. 3 E and F), because Piwi-associated piRNAs volved in repression of transposons. To this end, we compared are primary piRNAs. Importantly, in either total or IP libraries, the levels of several Drosophila transposon transcripts (invader1, the WT and DUO ovaries showed indistinguishable sequence idefix, mdg1, blood, gtwin, HetA, diver, and 1360) between our composition. Therefore, the slicer activity of Piwi is not required slicer-deficient flies and wild-type flies by quantitative RT-PCR for the ping-pong mechanism. (qRT-PCR) assay. None of the transposons tested were signifi- The above conclusion is based on the fact that DUO mutation cantly up-regulated in piwi1/piwi1,DUO/DUO or piwi1/piwi1,D37/ showed no effect on piRNAs at the population level, which does D37 ovaries (Fig. 4 A–C), consistent with the normal develop- not rule out the possibility that DUO mutations may affect piRNA ment and fertility of these transgenic flies. These results dem- biogenesis at specific loci. To test this possibility, we examined onstrate that the function of Piwi in repressing transposons is not whether the slicer activity of Piwi is involved in the biogenesis of mediated by its endonuclease activity. 17 known piRNA clusters, which comprise 12 somatic-enriched, two germ-line–enriched, and three ubiquitous clusters (16, 24, Discussion 25). As shown in Fig. 3I, no significant change was observed for We have shown that a catalytically null mutant of Piwi (D614A, any of these clusters in the slicer mutant. Finally, we compared D685A) fully rescues all of the known defects of piwi1 mutant— the level of piRNAs corresponding to specific transposon families the most severe mutant allele of Piwi. These slicer-deficient flies and found that no piRNA from any specific transposon was af- have normal ovary morphology, fertility, and viability, and they fected by the slicer mutation (Fig. 3 J and K), except that total retain the ability of generating piRNA molecules and repressing

Darricarrère et al. PNAS Early Edition | 3of6 Downloaded by guest on September 26, 2021 A C IP: anti-Myc Ratio RNA by PAGE Protein by Western RNA/Protein 12345 lane 2/lane 3 0.88 0.97 0.91 lane 4/lane 5 1.85 1.87 0.99 2S rRNA piRNAs D 2,500,0002.5 ) WT_IPIP-WT 6 2,000,0002.0000 MutantMutIPIP-DUOant DUO IP WT_TotalWT_WTTotal-WTotaT TlWota B 1,500,0001.5005 IP: anti-Myc Total-DUOTotal-DU 1,000,0001.00000 123 45 500,000of Reads (x10 0.5005 Piwi Normalized Number 0 18 20 22 24 26 28 30 32 size (base pairs) E F IP-WT IP-DUO 100% 100%

80% G 80% G 60% U 60% U 40% C 40% C 20% 20% A A

piRNA 0% piRNA 0% 1 6 11 16 21 1 6 11 16 21 Position Position G Total-WT H Total-DUO 100% 100% G G 80% 80%

60% U 60% U 40% C 40% C 20% 20% A A

piRNA 0% piRNA 0% 1 6 11 16 21 1 6 11 16 21 Position Position I 2 Total library 1.5 IP library

1

Ratio 0.5

0 Slicer Mutant / Widltype cluster01 cluster06 cluster11 cluster18 Cluster 02 Cluster 06 Cluster 10 Cluster 12 Cluster 18 Cluster 16 Cluster 15 Cluster 01 Cluster 05 Cluster 08 Cluster 11 Cluster 13 Cluster 17 Cluster 14 OSS2

Germline Ubiquitous Somatic J K IP Library Total Library 2 2

1 1

Ratio DUO/ WT 0 0 1 10 20 30 40 50 60 70 80 1 10 20 30 40 50 60 70 piRNA Transposon Family (Reference Number)piRNA Transposon Family (Reference Number)

Fig. 3. Duo mutant has no detectable defects in piRNA biogenesis. (A–C) Total levels of Piwi-associated piRNAs from +/+;WT/WT (lane 2), +/+;DUO/DUO (lane 3), piwi1/piwi1;WT/WT (lane 4), and piwi1/piwi1;DUO/DUO (lane 5) ovaries, resolved by denaturing PAGE (A). (B) Immunoprecipitation efficiency by Western blotting. (A and B) Lane 1 corresponds to W[1118], no-myc negative control. (C) Quantification by densitometry and normalization of piRNA/Piwi protein. (D) Size profile of ovarian piRNAs as revealed by deep sequencing. (E–H) Sequence profiles of piRNAs (∼23–28 nt) from IP libraries of piwi1/piwi1;WT/WT (E)and piwi1/piwi1;DUO/DUO (F) flies and from Total small RNA libraries of piwi1/piwi1;WT/WT (G) and piwi1/piwi1;DUO/DUO (H) flies. (I) Relative abundance of piRNAs derived from 17 known piRNA clusters from fly ovarian Piwi co-IP and total small RNA libraries, calculated as fold change of slicer mutant (DUO) over wildtype (WT). Only unique piRNAs are counted in this calculation, and the abundance is scaled to sequencing depth. The 17 clusters are grouped into three categories: germ-line–enriched (red), somatic-enriched (blue), and a ubiquitous class, comprising piRNAs similarly abundant in germ line and soma (green). Dark colors represent the Piwi co-IP libraries, and light colors represent the Total small RNA libraries. (J and K) Relative abundance of unique piRNAs derived from each transposon class in IP (J) and Total (K) small RNA libraries, shown as fold change of slicer mutant over wildtype. The abundance is scaled to the sequencing depth.

transposon elements. These results indicate that Piwi does not protein that is enriched in the nucleus (26), does not require the require slicer activity in vivo, despite the conservation of the slicer catalytic triad for its essential functions in , catalytic triad and in vitro observations (10). Our results are piRNA biogenesis, and transposon suppression in the consistent with the recent observation that Miwi2, a mouse Piwi testis (27). These congruent findings across species underscore

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1213283110 Darricarrère et al. Downloaded by guest on September 26, 2021 A B C W[1118] 8 250 20 piwi1/cyO 6 200 15 piwi1/piwi1 150 4 10 1 1 100 piwi /piwi , WT/WT 2 50 5 piwi1/piwi1, DUO/DUO 0 0 0 piwi1/piwi1, D37/D37

Expression Fold Change Invader1 Idefix mdg1 Expression Fold Change blood Gtwin Expression Fold Change HetA Diver 1360

Fig. 4. Piwi’s catalytic triad is not necessary to suppress transposons. The expression level of various transposons (invader1, idefix, mdg1, blood, gtwin, HetA, diver, and 1360) in ovaries was measured by qRT-PCR. After normalization of transposon levels per actin, the fold increase for each strain was calculated as a ratio over the wildtype W[1118] strain. The transposons are presented in three panels (A, B,andC) according to the magnitude of Piwi suppression towards their expression.

the distinct and conserved nuclear function of Piwi proteins that Western Blotting, Immunofluorescence, and Antibodies. Western blotting was is independent of their slicer activity. There are multiple lines of performed by using the standard protocol with the following primary anti- evidence supporting the notion that transposon silencing is reg- bodies: mouse anti-Myc (Cell Signaling; 9B11) at a dilution of 1:1,000, mouse ulated by the Piwi–piRNA complex in the nucleus (8, 28). The anti-Piwi (P3G11 and P4D2 mixed; Haru Siomi laboratory, Keio University, Tokyo, Japan.) at 1:5,000 dilution, and mouse anti-β-actin (Sigma Aldrich; association of Piwi with chromatin and with epigenetic factors A1978) at 1:2,000. offers an attractive model for transposon repression via an epige- fl fi Ovarian immuno uorescence protocol has been described (30). Antibodies netic mechanism guided by piRNA sequence-speci c recognition used for ovary immunofluorescence were mouse anti-Myc (Cell Signaling; of the target transposon sequences (17, 18). There is increasing 9B11, 1:1,000 dilution); rabbit anti-Vasa (Paul Lasko laboratory, McGill Uni- evidence suggesting that small RNAs hold answers to many un- versity, Montreal, Canada) at 1:10,000 dilution; mouse anti-1B1 (Developmental solved questions in genome regulation, and this study suggests that Studies Hybridoma Bank; 1:100 dilution); anti-mouse Alexa 488 (Invitrogen), careful in vivo validation of the proposed or in vitro-derived 1:500 dilution; Anti-Rabbit Alexa 456 (Invitrogen) 1:500 dilution. Images were functions of the main players may hold the key to unraveling these acquired with a Leica TCS SP5 Spectral Confocal microscope using sequential mechanisms. scanning mode. Materials and Methods Fertility and Viability Tests. To assess fertility, virgin female flies were col-

lected and paired with two or three males from the same genotype in in- BIOCHEMISTRY DNA Constructs and Site-Directed Mutagenesis. Site-directed mutagenesis was dividual vials. Flies were transferred to clean vials every 24 h, and the number performed by using the Stratagene Quick Change II kit on the pPMB1-6 P- of eggs laid was counted. Egg laying per fly per day was averaged across days element transformation construct that contains myc-tagged piwi genomic 3–9 for several flies of each genotype. Sample size, N, corresponds to the sequences (22), using the following primers: number of vials analyzed. To reduce variability, we assayed fertility for each Forward D614A: GATGACAATTGGCTTTGCCATTGCGAAGAGCACAC; genotype in parallel using the same batch of yeast–molasses medium. Both Reverse D614A: GTGTGCTCTTCGCAATGGCAAAGCCAATTGTCATC; fertility and viability assays were carried out at 25 °C. Viability was scored by Forward D685A: CGAATCGTATTTTATCGAGCCGGTGTGAGCTCCGGC; counting the number of adult flies enclosed per vial normalized per egg Reverse D685A: GCCGGAGCTCACACCGGCTCGATAAAATACGATTCG. counted in that vial.

Mutagenesis was achieved by subcloning a 5.6-kb HindIII–HindIII fragment Transposon Expression Analysis. RNA was isolated from pooled ovaries dis- of the piwi gene into pBluescript, followed by site-directed mutagenesis and sected from each genotype by using the RNeasy kit (Qiagen) with on-column subcloning a 2.3-kb NheI–NheI subfragment containing the relevant point DNase digestion. cDNA was prepared by using the High-capacity cDNA Reverse mutations back into pPMB1-6. For site-directed transgenesis, the wildtype Kit from Applied Biosystems. Real-time PCR was performed with and mutant (D614A,D685A) Myc–Piwi genomic rescue fragments were 2× ready SYBR green Mastermix (BioRad) by using the primer pairs (Table 1). cloned by high-fidelity PCR amplification (Phusion; NEB) from pPMB1-6 into For all genotypes except piwi1/piwi1, biological triplicates of at least 10 the XbaI and SpeI sites on pCa4B2G (GenBank accession no. EU420017.1) ovaries per sample were averaged. For piwi1/piwi1,manymoreovaries (21). All constructs were confirmed by sequencing in full. were pooled to produce one datapoint given their small size and rudi- For convenience we provide the FASTA sequence and positional maps for mentary nature. pPMB1-6, pCa4B2G–Myc–Piwi in Datasets S1 and S2. PIWI–piRNA Co-IP. A total of 100 pairs of ovaries were dissected in ice-cold PBS Drosophila Transgenic Strains and Husbandry. All and homogenized with a pestle in Ovary Lysis Buffer [20 mM Hepes, pH 7.5,

strains were reared between 18 and 25 °C on yeast/molasses agar medium. 100 mM KCl, 5 mM MgCl2, 0.1% SDS, 0.1% Deoxycholate.Na2, 1% (vol/vol) The Piwi mutant strain used, piwi1, has been described (22, 29). Triton X-100, 5% (vol/vol) glycerol supplemented with 1 mM fresh DTT, 1× Transgenic flies were generated by the following two independent complete mini EDTA-free Proteinase inhibitor mixture (Roche), and 0.5 U/μL methods. RNase OUT (Invitrogen)]. Approximately 200 μL of soluble fraction was re- P-element transformation of the slicer mutant. PMB1-6 mutant (D614A, D685A) covered by centrifugation, adjusted to 400 μL, and precleared with 100 μLof construct was coinjected with transposase into white-eyed W[1118] stocks. 50% bead slurry (GE Protein A beads washed and resuspended in ovary lysis Positive red-eyed transformants were balanced with a double balancer strain buffer). Precleared lysate was recovered by centrifugation and incubated (Sco/CyO; Sb/TM6b) and made homozygous. We mapped the integration site with mouse anti-Myc antibody (3 μL per sample) for 12 h at 4 °C after which of transformant fly line D37 to an intergenic region on chromosome 3 (3R:13) 50 μL of 50% Protein A bead slurry was incubated for 1 h. Beads were washed and confirmed the point mutations by sequencing a PCR product from twice with ovary lysis buffer for 5 min and twice with radioimmunoprecipitation transgenic fly genomic DNA. assay buffer with inhibitors. To control for protein immunoprecipitation effi- Site-specific transgenesis by FlyC31 technology. WT (wildtype Myc–Piwi) and ciency, 10% of beads were saved by boiling on SDS-loading buffer and Western DUO (catalytically deficient Myc–Piwi) transgenic fly lines were generated by blotting. The rest was used to recover PIWI-associated RNAs by TRIzol LS ex- injection of pCa4B2G–Myc–Piwi wild-type or point mutant (D614A, D685A) traction and small RNA precipitation. To visualize total piRNA levels on a gel, − constructs into a transgenic line containing white , X-linked integrase, and 10% of the total isolated small RNAs were subjected to phosphonucleotide ki- an attP2 site at 68A4 by Duke University Model System Genomics. Red-eyed nase reaction with P-32 γ-ATP and electrophoresed on a 15% polyacrylamide–6 transformant males were crossed with w[1118] virgin females before bal- M urea denaturing gel. The gel was covered in plastic-wrap and exposed to ancing with Sco/CyO; Sb/TM6b to remove the integrase transgene from the a phosphoimager screen for quantification. background. Site-specific integration of the transgenic constructs at the attP2 site was confirmed by PCR of junctional sequences. The D614A, D685A Small RNA Profiling. Ovaries from piwi1/piwi1,WT/WT and piwi1/piwi1,DUO/ point mutations were also confirmed by sequencing genomic PCR from DUO DUO genotypes were isolated to prepare four piRNA libraries from either transgenic line, which were found absent in the WT transgenic line. total small RNAs (Total-WT and Total-DUO) or from Piwi co-IP samples (IP– All transgenic strains used in this study contain a single copy transgene. WT and IP–DUO). Small RNAs were gel purified, and libraries were prepared

Darricarrère et al. PNAS Early Edition | 5of6 Downloaded by guest on September 26, 2021 Table 1. Primer pairs for real-time PCR analysis of transposon 32-nt sequences were kept. In detail, a linker sequence is searched against expression the small RNA sequence pools. If the linker sequence is detected to match to a sequence, the portion that is beyond the matching region is removed. Transposon Primer Otherwise, the linker is trimmed 1 nt off and subjected to a similar search. < Blood During the search process, mismatch is not allowed for any match that is 9 nt. Otherwise, one mismatch is allowed for a search. The average clipping Forward TGCCACAGTACCTGATTTCG rate is 0.81. Reverse GATTCGCCTTTTACGTTTGC Mapping. For downstream analysis, only sequences perfectly mapped to the Gtwin D. melanogaster Release 5 genome (excluding Uextra) were retained. Li- TTCGCACAAGCGATGATAAG Forward braries were normalized to sequencing depth to allow for cross-analysis. Reverse GATTGTTGTACGGCGACCTT Annotation. To implement the origin annotation of piRNAs represented in our HetA libraries, we used the miRBase Release 16 for miRNA-derived small RNAs, Forward CGCGCGGAACCCATCTTCAGA repeat masker annotation for transposon-derived RNAs, and ensembl gene Reverse CGCCGCAGTCGTTTGGTGAGT annotation for gene-derived RNAs. Up to 2 mismatches were allowed with Diver the insertion/deletion included to account for incomplete annotation of Forward GGCACCACATAGACACATCG noncoding RNAs. fi Reverse GTGGTTTGCATAGCCAGGAT Cluster analysis. We selected 17 piRNA clusters from different scienti c groups, which covers clusters that were expressed preferentially in germ-line, somatic 1360 cells, together with the clusters that showed no preference: 15 piRNA clusters Forward GGAGCTCTGCGTATAGCCAACTT reported by Hannon and coworkers (24) to produce the greatest number of Reverse ACCTAAACCGCCGAGTCCTG piRNAs; 1 piRNA cluster (Traffic Jam) from Siomi and coworkers (16); and 1 Invader1 piRNA cluster (OSS2) located in Chr2R: 936,722–943,706 from Lau, Lai, and GTACCGTTTTTGAGCCCGTA Forward coworkers (25). For this analysis, we only considered uniquely mapped small Reverse AACTACGTTGCCCATTCTGG RNAs, defined as mapping perfectly to only one position in the D. mela- Idefix nogaster Release 5 genome. Forward TGAAGAAAAGAAGGGCGAGA Transposon analysis. We selected ∼80 transposons to show their piRNA ratio Reverse TTCTGCTGTTGATGCTTTGG between DUO and WT for both IP and Total libraries. The criteria for mdg1 choosing transposons were set to include only transposons for which at least Forward AACAGAAACGCCAGCAACAGC 20 piRNA sequences were present in WT. Reverse CGTTCCCATGTCCGTTGTGAT The sequencing data analyzed in this paper have been deposited to the Gene Expression Omnibus database (accession no. GSE34032).

according to the Illumina prep kit. After cloning, 16- to 29-nt small RNAs ACKNOWLEDGMENTS. We thank members of the H.L. laboratory—in partic- — were isolated for Solexa sequencing. ular Drs. John Saxe, Vamsi Gangaraju, Travis Thomson, and Hongyin Qi for input toward this project. pCa4B2G was provided by Dr. Norbert Perrimon, > and the P3G11 and P4D2 antibodies were provided by Dr. Mikiko Siomi. All Bioinformatic Analysis. Clipping. In total, 75 million small RNAs were transgenic injections were performed by Jamie Roebuck (Duke University ∼ obtained from Solexa sequencing. Total-WT library yielded 10 million small Model System Genomics). This work was supported by National Institutes of RNAs, whereas each of the other three libraries yielded >20 million small Health Grant DP1 CA174418 (to H.L.). N.D. was supported by a Fonds de RNAs. Linkers were clipped off by using our homemade program, and 18- to Recherche en Santé Québec Master’s Training Grant.

1. Juliano C, Wang J, Lin H (2011) Uniting germline and stem cells: The function of Piwi 16. Saito K, et al. (2009) A regulatory circuit for piwi by the large Maf gene trafficjamin proteins and the piRNA pathway in diverse organisms. Annu Rev Genet 45:447–469. Drosophila. Nature 461(7268):1296–1299. 2. Grivna ST, Beyret E, Wang Z, Lin H (2006) A novel class of small RNAs in mouse 17. Brower-Toland B, et al. (2007) Drosophila PIWI associates with chromatin and inter- spermatogenic cells. Dev 20(13):1709–1714. acts directly with HP1a. Genes Dev 21(18):2300–2311. 3. Watanabe T, et al. (2006) Identification and characterization of two novel classes of 18. Yin H, Lin H (2007) An epigenetic activation role of Piwi and a Piwi-associated piRNA small RNAs in the mouse germline: -derived siRNAs in oocytes and in Drosophila melanogaster. Nature 450(7167):304–308. germline small RNAs in testes. Genes Dev 20(13):1732–1743. 19. Lin H, Yin H (2008) A novel epigenetic mechanism in Drosophila somatic cells medi- 4. Girard A, Sachidanandam R, Hannon GJ, Carmell MA (2006) A germline-specific class ated by Piwi and piRNAs. Cold Spring Harb Symp Quant Biol 73:273–281. of small RNAs binds mammalian Piwi proteins. Nature 442(7099):199–202. 20. Saito K, et al. (2010) Roles for the Yb body components Armitage and Yb in primary 5. Aravin A, et al. (2006) A novel class of small RNAs bind to MILI protein in mouse testes. piRNA biogenesis in Drosophila. Genes Dev 24(22):2493–2498. Nature 442(7099):203–207. 21. Markstein M, Pitsouli C, Villalta C, Celniker SE, Perrimon N (2008) Exploiting position 6. Lau NC, et al. (2006) Characterization of the piRNA complex from testes. Science effects and the gypsy retrovirus insulator to engineer precisely expressed transgenes. 313(5785):363–367. Nat Genet 40(4):476–483. 7. Siomi MC, Sato K, Pezic D, Aravin AA (2011) PIWI-interacting small RNAs: The van- 22. Cox DN, et al. (1998) A novel class of evolutionarily conserved genes defined by piwi guard of genome defence. Nat Rev Mol Cell Biol 12(4):246–258. are essential for stem cell self-renewal. Genes Dev 12(23):3715–3727. 8. Klenov MS, et al. (2011) Separation of stem cell maintenance and transposon silencing 23. Megosh HB, Cox DN, Campbell C, Lin H (2006) The role of PIWI and the miRNA ma- functions of Piwi protein. Proc Natl Acad Sci USA 108(46):18760–18765. chinery in Drosophila germline determination. Curr Biol 16(19):1884–1894. 9. Senti KA, Brennecke J (2010) The piRNA pathway: A fly’s perspective on the guardian 24. Malone CD, et al. (2009) Specialized piRNA pathways act in germline and somatic of the genome. Trends Genet 26(12):499–509. tissues of the Drosophila ovary. Cell 137(3):522–535. 10. Saito K, et al. (2006) Specific association of Piwi with rasiRNAs derived from retro- 25. Lau NC, et al. (2009) Abundant primary piRNAs, endo-siRNAs, and microRNAs in transposon and heterochromatic regions in the Drosophila genome. Genes Dev a Drosophila ovary cell line. Genome Res 19(10):1776–1785. 20(16):2214–2222. 26. Shoji M, et al. (2009) The TDRD9-MIWI2 complex is essential for piRNA-mediated 11. Gunawardane LS, et al. (2007) A slicer-mediated mechanism for repeat-associated retrotransposon silencing in the mouse male germline. Dev Cell 17(6):775–787. siRNA 5′ end formation in Drosophila. Science 315(5818):1587–1590. 27. De Fazio S, et al. (2011) The endonuclease activity of Mili fuels piRNA amplification 12. Parker JS, Roe SM, Barford D (2004) Crystal structure of a PIWI protein suggests that silences LINE1 elements. Nature 480(7376):259–263. mechanisms for siRNA recognition and slicer activity. EMBO J 23(24):4727–4737. 28. Ishizu H, Nagao A, Siomi H (2011) Gatekeepers for Piwi-piRNA complexes to enter the 13. Thomson T, Lin H (2009) The biogenesis and function of PIWI proteins and piRNAs: nucleus. Curr Opin Genet Dev 21(4):484–490. Progress and prospect. Annu Rev Cell Dev Biol 25:355–376. 29. Lin H, Spradling AC (1997) A novel group of pumilio mutations affects the asymmetric 14. Saito K, Siomi MC (2010) Small RNA-mediated quiescence of transposable elements in division of germline stem cells in the Drosophila ovary. Development 124(12): animals. Dev Cell 19(5):687–697. 2463–2476. 15. Cox DN, Chao A, Lin H (2000) piwi encodes a nucleoplasmic factor whose activity 30. Lin H, Yue L, Spradling AC (1994) The Drosophila fusome, a germline-specificor- modulates the number and division rate of germline stem cells. Development 127(3): ganelle, contains membrane skeletal proteins and functions in cyst formation. 503–514. Development 120(4):947–956.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1213283110 Darricarrère et al. Downloaded by guest on September 26, 2021