Gene silencing mechanisms mediated by Aubergine–piRNA complexes in Drosophila male gonad

Expression of Aub in germline cells

To biochemically elucidate the functional involvement of Aub in the presumptive RNA silencing process in fly gonads, we produced monoclonal antibodies against Aub. Western blotting on ovary and S2 cell lysates using the anti-Aub antibody revealed that Aub is strongly expressed in ovaries, but little in S2 cells (Fig. 1A). AGO1 and AGO2 expressed in S2 cells (Miyoshi et al. 2005) were not cross-reacted by anti-Aub (Fig. 1A). Western blotting on Piwi and AGO3 recombinant proteins produced in E. coli confirmed that the anti-Aub cross-reacts to neither of those (Supplemental Fig. 1A), indicating that the antibody is specific to Aub. In early embryos, Aub is strongly expressed, but its protein levels gradually diminish throughout development (Supplemental Fig. 1B).

To examine the expression patterns of endogenous Aub in ovaries and embryos, immunofluorescent staining was carried out. In the germarium, Aub was found in the cytoplasm of germline stem cells (GSC) and germline cyst cells (GCC) (Fig. 1B). There are strong accumulations of Aub at the cytoplasmic side of perinuclear regions, called the nuage, a germline-specific body conserved in species (Eddy 1975). In egg chambers at stages 6 and 10A, Aub was clearly detected in the cytoplasm of both nurse cells and oocytes (Fig. 1C,D). Nuage staining was observed only in nurse cells, not in oocytes at stage 6 (Fig. 1C), which is consistent with the notion that nuage is known to be lost from oocytes (Mahowald and Strassheim 1970). Clear, sharp accumulation of Aub was observed at the posterior of oocytes at stage 10A (Fig. 1D). Aub signals were under the detection limit in somatic cells such as terminal filament cells, cap cells, and follicle cells (Fig. 1B,C), where strong Piwi expression has been observed (Saito et al. 2006). These staining patterns rather resembled those previously reported with GFP-tagged Aub (GFP-Aub) overexpressed in fly ovaries (Harris and Macdonald 2001) and with polyclonal antibodies against Aub (Brennecke et al. 2007). In embryos, endogenous Aub was clearly accumulated in the cytoplasm of pole cells, as expected (Fig. 1E–G).

Testis staining with anti-Aub revealed Aub in GSC, gonialblast, and spermatogonia cells (Fig. 1H,K). However, Aub expression was undetected in somatic cells surrounding gonialblast and spermatogonia cells, and in hub (Fig. 1H,K), where the strong expression of Piwi was observed (Fig. 1I,L; Cox et al. 2000; Saito et al. 2006). Hub is a tiny cluster of post-mitotic somatic cells localized at the apical tip of the testis and functions in the maintenance of GSC identity and influences its behavior (Ohlstein et al. 2004; Snee and Macdonald 2004). Double-staining of testes with anti-Aub and anti-Piwi confirmed that Aub signals were only detected in germ cells, whereas Piwi signals were restricted to somatic cells in testes (Fig. 1J,M). Namely, expression of these two proteins was mutually exclusive in respect to both cellular localization and cell type.

Comparison of piRNAs associated with Aub and Piwi in fly ovaries

We previously identified small RNAs associated with Aub immunoprecipitated from ovary lysate with the anti-Aub antibody (Gunawardane et al. 2007). Silver staining patterns of protein and RNA components in the immunoprecipitate are shown in Supplemental Figure 2. Western blotting analyses on the immunoprecipitate using anti-AGO3 and anti-Piwi antibodies confirmed that the Aub complex does not contain either protein (Supplemental Fig. 2C). To gain insight into the piRNAs associated with Aub and Piwi in ovaries, we reexamined the compositions of those clones, which are listed in Supplemental Tables 1, 2, and 3. Northern blotting analyses demonstrated that piRNAs associated with Aub seemed also to be associated with Piwi (Fig. 2A–C). A size difference was clearly observed between piRNAs with Piwi and Aub (Fig. 2A–C). This agrees with previous data showing that ³²P-end-labeled piRNAs associated with Piwi in fly ovaries contained molecules migrating slightly slower than those associated with Aub on denaturing acrylamide gels (Brennecke et al. 2007; Gunawardane et al. 2007). piRNAs associated with Aub and Piwi most likely contain an equal amount of monophosphate at the 5′-ends (data not shown) and are methylated at their 3′-ends (Saito et al. 2007). Thus, piRNAs with Aub might simply be shorter than those associated with Piwi.

We previously reported that miRNAs were not present in Piwi complex immunopurified from ovaries and that siRNAs pre-programmed in ovary lysate were not loaded on Piwi (Saito et al. 2006). We speculated that neither siRNAs nor miRNAs were loaded onto Aub, as was the case for Piwi. To examine this possibility, small RNAs isolated from Aub complex in ovaries were probed with an oligo that specifically recognizes miR-310, one of the miRNAs expressed in fly ovaries. A positive band was detected in just the AGO1 lane (Fig. 2D). Another oligo specifically recognizing miR-186 showed the same result (data not shown). This strongly suggested that Aub does not function together with miRNAs. Indeed, only a trace of miRNAs was obtained (nine of a total 925 Drosophila small RNAs; 0.97%) as Aub-associating small RNAs (Supplemental Table 1). We then assessed if the siRNA duplex of 21 nt in length, a triggering molecule of RNAi mediated by AGO2, was loaded onto Aub as well as AGO2 upon addition to ovary lysate. luc guide siRNA was efficiently loaded onto AGO2 but not onto Aub (Fig. 2E). These data suggested that the 21 nt siRNA loading machinery composed with Dicer-2 and R2D2 (Liu et al. 2003) was unable to load the small RNA onto both Piwi (Saito et al. 2006) and Aub (this study) proteins.

Su(Ste) piRNAs are the most abundant associated with Aub in testes

Immunoprecipitation of Aub from testes was also performed using the anti-Aub antibody. As in ovaries, small RNAs 24–30 nt long were observed (Supplemental Fig. 3). A small RNA library was constructed, and 458 clones were sequenced (Table 1). The list of piRNAs obtained is shown in Supplemental Table 4. Although piRNAs associated with Aub in testes also contain those derived from LINE and LTR-type transposons (Table 1) and show a strong preference for U at their 5′-ends (∼87%) (Supplemental Fig. 4), the most abundant class of piRNAs identified was made up of those derived from Su(Ste) antisense transcripts (195 of 425 piRNAs; ∼46%) (Fig. 3A; Table 1; Supplemental Table 4). Consistent with these findings, previously genetic studies have shown that both aub and Su(Ste) mutations cause an accumulation of Stellate mRNA in testes (Bozzetti et al. 1995, Aravin et al. 2004; Vagin et al. 2006). Su(Ste) piRNAs in a sense orientation were found only once in this screening. This bias was apparently not a consequence of any loading bias to Aub because it has been reported that even in total RNAs, Su(Ste) piRNAs in sense are hardly detected in testes (Vagin et al. 2006). piRNAs generated from Stellate sense transcript were also obtained in this experiment (two of 425) (Fig. 3A; Table 1; Supplemental Table 4). Notably, Su(Ste) piRNAs seemed not evenly produced from the precursor molecules (Fig. 3A). Particularly, one Su(Ste) piRNA, termed Su(Ste)-4, was so abundant that it appeared 120 times [including closely related ones such as Su(Ste)-4a shown in Supplemental Table 4] in all the clones sequenced. The association of Su(Ste)-4 with Aub in testes was further confirmed by Northern blotting analysis using a DNA oligo specific for the small RNA (Fig. 3B).

A second large class of piRNAs associated with Aub in testes was made up of those derived from two tiny regions (A, 274 nt; and B, 173 nt) on chromosome X (Fig. 4A). The A and B regions are both located between retrotransposons (Fig. 4A). Interestingly, one of these piRNAs, termed AT-chX-1, whose sequence appeared 103 times in this screening (∼24% of the total piRNAs), showed strong complementarity to vas mRNA (Fig. 4B), a germline-specific transcript involved in oogenesis (Lasko and Ashburner 1988; Styhler et al. 1998). Northern blotting using an oligo specific for AT-chX-1 verified its association with Aub in testes (Fig. 4C). AT-chX-2, another piRNA found in this subgroup, appeared only once in this experiment but also showed high complementarity to vas mRNA (Fig. 4B). Western blotting revealed that the protein levels of VAS in aub ⁻/aub ⁻ mutant fly testes were increased by about twofold compared with those of wild-type (wt) and aub ⁻/+ (Fig. 4D; Supplemental Fig. 5A). Mislocalization or ectopic expression of VAS was not observed in aub mutant testes (data not shown). We also found that VAS expression was not increased in piwi mutant testes (Supplemental Fig. 5B). Although it has been shown that the phosphorylation of VAS in ovaries is altered in mutants defective in piRNA production and/or function (Klattenhoff et al. 2007), we found no change in the electrophoretic mobility of VAS in aub mutant and wt testes (Fig. 4D; Supplemental Fig. 5A; data not shown), suggesting that the observed twofold changes in VAS abundance do not reflect an effect on its phosphorylation. In contrast, in ovaries, where little AT-chX-1 and -2 was found in the Aub complex (three of 689), we did not observe such an alteration of VAS levels (Fig. 4D) as reported (Wilson et al. 1996). Taken together, it is suggested that vas expression in testes might possibly be down-regulated by the Aub complex containing AT-chX-1 and -2 through RNA silencing mechanisms (see below).

Piwi and piRNA association in testes

We also carried out a parallel examination of small RNAs residing within the Piwi complex in testes (Table 1; Supplemental Fig. 3; Supplemental Table 5). We detected only a few Su(Ste) piRNAs (four of 246 piRNAs) and no AT-chX-1 and -2 as Piwi-associating small RNAs (Table 1; Supplemental Table 5), which suggested that Piwi hardly contributes to the repression of Stellate and vas gene expression. It is, however, worth noting that the loss of Piwi in testes increases the production of Su(Ste) piRNAs (Vagin et al. 2006). The majority of the small RNAs associated with Piwi in testes were piRNAs derived from transposable elements and other repetitive sequences in the genome, as observed in ovaries (Saito et al. 2006). Interestingly, pieces of protein-coding sense mRNAs were conspicuous in this screening (80 of 347 small RNAs present in the Piwi complex in testes) (Table 1; Supplemental Table 5) as has been reported in the small RNA profiling study done for MIWI, a Piwi homolog expressed in mouse (Girard et al. 2006). However, the significance of the association of Piwi/MIWI in testes with small RNAs originating from sense mRNAs remains unclear.

Immunoprecipitated Aub–piRNA complexes from ovaries and testes show Slicer activity

Previously, we demonstrated that Drosophila AGO1 and AGO2 exhibit Slicer activity in vitro (Miyoshi et al. 2005). Moreover, we recently showed that recombinant Piwi, Aub, and AGO3 produced in E. coli also exhibit Slicer activity (Saito et al. 2006; Gunawardane et al. 2007). We then assessed if the Aub–piRNA complexes immunopurified from ovaries and testes could, indeed, show the target RNA cleavage activity in vitro. A target RNA containing a sequence perfectly complementary to #4 roo piRNA, one of the piRNAs residing in the Aub complex in ovaries (Fig. 2C), was produced and used in the assays. As shown in Figure 5A, the #4 roo target RNA was efficiently cleaved with Aub complex immunopurified from ovaries, indicating that endogenous Aub residing in fly gonads also has Slicer activity. Similar experiments were carried out using Aub complex immunopurified from testes. An RNA target containing a sequence completely matching Su(Ste)-4 piRNA [Su(Ste)-4 target] or part of the Stellate sense mRNA (Stellate target), which should be recognized by the Aub–Su(Ste)-4 complex, was efficiently cleaved, as expected (Fig. 5B,C). These results are consistent with data demonstrating that mutations in aub cause an increase in both Stellate mRNA and sense Su(Ste) RNA (Schmidt et al. 1999; Aravin et al. 2001; Vagin et al. 2006). Under the same conditions, Aub complex obtained from ovaries, where Su(Ste) transcripts are presumably unexpressed, did not cleave the Su(Ste)-4 target (Fig. 5B). In fact, Su(Ste)-4 piRNA did not appear at all in the experiment identifying small RNAs in Aub complex in ovaries (Supplemental Table 1). Although the activity was relatively weak, a target RNA containing a sequence of vas mRNA showing complementarity to AT-chX-1 was also cleaved by Aub complex immunopurified from testes (Fig. 5D). These results imply that vas is a candidate for in vivo target genes for Aub–piRNA complexes in testes.

Drosophila strains

The y w strain was used as a wild type. The strains bearing aub mutations, aub^HN2 and aub^QC42, were provided by P. Zamore (University of Massachusetts). Females of aub^HN2 and males of aub^QC42 were crossed to yield aub heterozygote flies, aub^QC42/aub^HN2.

Western blot analysis

Two hundred amino acids at the N terminus of Aub fused with GST were used as the antigen to immunize mice. Anti-Aub monoclonal antibodies were produced essentially as described previously (Ishizuka et al. 2002). Western blotting was performed as described previously (Miyoshi et al. 2005). Ten micrograms of proteins from each sample were loaded on gels (Figs. 1A, 4D). Anti-tubulin was from the Developmental Studies Hybridoma Bank (1:1000 dilution). Anti-VAS was provided by A. Nakamura (Kobe-RIKEN CDB) and used at 1:2500 dilution.

Immunofluorescence

Testes and ovaries were dissected manually from adult flies in 1× PBS. Embryos were collected and dechorionated. Immunostaining was performed following standard procedures. Anti-Aub was purified from culture supernatants of hybridoma cells using Thiophilic-Superflow Resion (BD Biosciences) and directly labeled using a HiLyte Fluor 555 Labeling Kit-NH₂ (Dojindo Molecular Technologies). Culture supernatants of anti-Piwi hybridoma cells were used without dilution. Alexa 488-conjugated anti-mouse IgG (Molecular Probe) was used as the secondary antibody to detect the anti-Piwi antibody. All images were collected using a Zeiss LSM510 laser scanning microscope.

Small RNA cloning and sequence analysis

Cloning of small RNAs associated with Aub in ovaries and testes and Piwi in testes was carried out essentially as described (Saito et al. 2006) with minor modifications. Immunoprecipitation was performed essentially as described previously (Miyoshi et al. 2005). Immunoprecipitation buffer contained 30 mM HEPES-KOH (pH 7.3), 150 mM KOAc, 2 mM MgOAc, 5 mM DTT, 0.1% NP-40, 2 μg/mL Pepstatin, 2 μg/mL Leupeptin, and 0.5% Aprotinin. About 300 ovaries or about 1000 testes were used per immunoprecipitation. After immunoprecipitation, total RNAs were isolated from the immunopurified complexes with phenol:chloroform and precipitated with ethanol. RNAs were dephosphorylated with CIP (NEB) and labeled with [γ-³²P]ATP with T4 polynucleotide kinase (TaKaRa) for visualization. For cloning of small RNAs isolated from Aub complex in ovaries, we used adapters and primers (i.e., RT and PCR primers) as described by Saito et al. (2006). Adapters (MI-5′ Linker and MI-3′ Linker) and primers used for small RNAs associated with Aub and Piwi in testes were included in the DynaExpress miRNA Cloning Kit (BioDynamics Laboratory). The sequences of each oligo were as follows:

• MI-3′ Linker: 5′-pCTGTAACTCGGGTCAATddC-3′ (DNA);

• MI-5′ Linker: 5′-AUCGUCUCGGGAUGAAA-3′(RNA);

• 3′ RT primer: 5′-ATTGACCCGAGTTACAG-3′ (DNA);

• 5′ Primer: 5′-ATCGTCTCGGGATGAAA-3′ (DNA).

First-strand cDNA synthesis was performed with Stratascript RT (Stratagene) or Reverse Transcriptase (BioDynamics Laboratory). KOD plus (TOYOBO) or Ex Taq polymerase (TaKaRa) was used as the polymerase. PCR products were cloned into the EcoRV site of the pBS SK+ vector (for small RNAs associated with Aub in ovaries) or pTAC-1 (for small RNAs associated with Aub or Piwi in testes) (BioDynamics Laboratory) and sequenced. RNA sequences between 5′ and 3′ adapters were analyzed by undertaking searches in annotated genomic databases (UCSC: http://genome.ucsc.edu/cgi-bin/hgBlat; NCBI: http://www.ncbi.nlm.nih.gov/blast/; and FlyBase: http://flybase.bio.indiana.edu/blast/) to determine whether the cloned RNAs map to the genome. We only used the best matches up to a maximum of two differences (mismatch, insertion, or deletion) for each small RNA sequence. For repeat annotation, we used results from the RepeatMasker program (UCSC BLAT), NCBI BLAST, or FlyBase BLAST.

Northern blot analysis

Total RNA either from fly ovaries/testes or immunoprecipitated complexes was isolated using ISOGEN (Nippon Gene). Five micrograms of total RNA were used in Figures 2, C and D, 3B, and 4C. Northern blotting was performed as reported previously (Saito et al. 2006). Probes used for roo piRNA and miR-310 were as described previously (Saito et al. 2006). Probes used for detecting #4 roo piRNA, Su(Ste)-4 and AT-chX-1 were as follows:

• For #4 roo piRNA: 5′-TCGACTCAGTGGCACAATAAAT-3′;

• for Su(Ste)-4: 5′-TCGGGCTTGTTCTACGACGATGAGA-3′;

• for AT-chX-1: 5′-GCCCGAGCCGTCTAACGATGAAACA-3′.

DNA fragments for detecting piRNAs that originated from 17.6 (accession number: XD1472) and GATE (accession number: AJ010298) retrotransposons were obtained by PCR on the Drosophila melanogaster genomic DNA and cloned into a pBS SK+ vector. The primers used are as follows:

• For 17.6: 5′-ATTCTTGTAAACAAATCTTA-3′ and 5′-CAGCCTCTCACAAATTCAAT-3′;

• for GATE: 5′-TCGAGCAGGGGCACTTCTCAGTC-3′ and 5′-ACTATGGCATGCGCGTCGCCA-3′.

PCR was again performed using primers for the T7 and T3 promoter sequences on each construct, and the PCR products were used as templates for in vitro transcription using MAXIscript T7 and T3 kits (Ambion) in the presence of ³²P-UTP. RNAs transcribed were extracted with phenol:chloroform, precipitated with ethanol, and partially hydrolyzed as described previously (Saito et al. 2006).

siRNA loading assay

Three hundred ovaries were homogenized in a hypotonic buffer (30 mM HEPES-KOH at pH 7.4, 2 mM MgOAc, 5 mM DTT, and 1 mg/mL Pefablock SC) to prepare ovary lysate. luc siRNA duplex (Miyoshi et al. 2005) was incubated in the lysate for 1 h at 26°C. Of the luc siRNA duplex, only guide siRNA was labeled with [γ-³²P]ATP for visualization. Aub and AGO2 complexes were then immunopurified from the mixture using the specific antibodies. Sodium chloride was added to the lysates to 800 mM just before immunoprecipitation was started. Reaction mixtures were rocked for 1 h at 4°C, and the beads were washed extensively with a washing buffer (30 mM HEPES-KOH at pH 7.4, 800 mM NaCl, 2 mM MgOAc, 5 mM DTT, 1 mg/mL Pefablock SC). RNAs were isolated from the immunoprecipitates, separated on a denaturing polyacrylamide gel, and visualized on BAS-2500 (Fuji).

In vitro target RNA cleavage assay

Target RNA cleavage assays using immunoprecipitates were performed in a reaction buffer containing 30 mM HEPES-KOH (pH 7.4), 100 mM KOAc, 2 mM MgOAc, 5 mM DTT, 10 mM creatine phosphate, 0.5 mM ATP, 30 μg/mL creatine kinase (Roche), 0.1 U/mL RNasin (Promega), and 0.5 μg of yeast RNA (Ambion). After reaction for 3 h at 26°C, RNAs were isolated from the whole reaction mixtures, run on gels, and visualized on the BAS 2500. To make a target harboring a sequence completely complementary to miR-310, a short DNA fragment produced by annealing a set of oligo DNAs (5′-AGCTTAAAGGCCGGGAAGTGTGCAATAA-3′ and 5′-AGCTTTATTGCACACTTCCCGGCCTTTA-3′) was inserted in a pBS SKII+ vector digested with HindIII (pBS310). To make a target harboring a sequence completely complementary to #4 roo piRNA (#4 roo piRNA target), a short DNA fragment produced by annealing a set of oligo DNAs (5′-GATCCATAGTCGACTCAGTGGCACAATAAATAAAGGC-3′ and 5′-GGCCGCCTTTATTTATTGTGCCACTGAGTCGACTATG-3′) was inserted in a pBS310 vector digested with BamHI and NotI. To make a target harboring a sequence completely complementary to Su(Ste)-4 piRNA [Su(Ste)-4 target], a short DNA fragment produced by annealing a set of oligo DNAs (5′- GATCTCGGGCTTGTTCTACGACGATGAGA-3′ and 5′- GGCCTCTCATCGTCGTAGAACAAGCCCGA-3′) was inserted in a pBS310 vector digested with BamHI and NotI. To make a target harboring a partial sequence of Stellate sense mRNA (Stellate target), a short DNA fragment produced by annealing a set of oligo DNAs (5′-AATTCTCTGGCTTGTTGTACGGCGATGAAAG-3′ and 5′-GATCCTTTCATCGCCGTACAACAAGCCAGAG-3′) was inserted into a pBS SK+ vector digested with BamHI and EcoRI. To make a target harboring a partial sequence of vas mRNA (vas target; as a target for AT-chX-1), a short DNA fragment produced by annealing a set of oligo DNAs (5′-TTCCTCCCGAGCCGTCCAACGATGCAATTGAGA-3′ and 5′-TCTCAATTGCATCGTTGGACGGCTCGGGAGGAA-3′) was inserted in a pBS SKII+ vector that was digested with BamHI and NotI and blunted. PCR was again performed using primers for the T7 and T3 promoter sequences, and the PCR products were then used as templates for in vitro transcription using a MEGAscript T7 kit (Ambion). The resultant RNAs were radiolabeled at the 5′-G cap by guanylyltransferase (Ambion) as previously described (Okamura et al. 2004) or with [γ-³²P]ATP using T4 polynucleotide kinase and gel-purified. vasa target (Fig. 5D) was labeled with pCp at the 3′-end and gel-purified before use.

Accession numbers

The accession numbers are: Aub-associated piRNAs in ovary, AB296387-AB297014; Aub-associated piRNAs in testis, AB297015-AB297190; and Piwi-associated piRNAs in testis, AB297191-AB297373.