Evidence for a piwi-Dependent RNA Silencing of the gypsy Endogenous Retrovirus by the Drosophila melanogaster flamenco Gene

Corresponding author: Alain Pélisson, 141 rue de la Cardonille, 34396 Montpellier Cedex 5, France., pelisson{at}igh.cnrs.fr (E-mail)

Communicating editor: M. J. SIMMONS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In Drosophila melanogaster, the endogenous retrovirus gypsy is repressed by the functional alleles (restrictive) of an as-yet-uncloned heterochromatic gene called flamenco. Using gypsy-lacZ transcriptional fusions, we show here that this repression takes place not only in the follicle cells of restrictive ovaries, as was previously observed, but also in restrictive larval female gonads. Analyses of the role of gypsy cis-regulatory sequences in the control of gypsy expression are also presented. They rule out the hypothesis that gypsy would contain a single binding region for a putative Flamenco repressor. Indeed, the ovarian expression of a chimeric yp3-lacZ construct was shown to become sensitive to the Flamenco regulation when any of three different 5'-UTR gypsy sequences (ranging from 59 to 647 nucleotides) was incorporated into the heterologous yp3-lacZ transcript. The piwi mutation, which is known to affect RNA-mediated homology-dependent transgene silencing, was also shown to impede the repression of gypsy in restrictive female gonads. Finally, a RNA-silencing model is also supported by the finding in ovaries of short RNAs (25–27 nucleotides long) homologous to sequences from within the gypsy 5'-UTR.

RECENT genome sequence analyses have shown the quantitative importance of retroelements. For instance, genomic parasites, which replicate by retrotranscription and insertion of a cDNA elsewhere in the genome, make up >40% of the human genome (LANDER et al. 2001 ). The tight control of their mobilization is an important characteristic of their regulation to prevent high mutation rates due to novel insertions. Indeed, their transposition is usually a rare event. To study such regulations, fly geneticists have been paying special attention to some strains in which one or several elements display unusually high transposition rates. For instance, in the 2b3 (PASYUKOVA et al. 1997 ) and U (DESSET et al. 2003 ) unstable stocks, the copia and both the Zam and Idefix retroelements are mobilized, respectively. In two other strains, MG (MEVEL-NINIO et al. 1989 ; PRUD'HOMME et al. 1995 ) and MS (KIM et al. 1990 ), frequent transpositions of the gypsy element result in high levels of insertional mutability.

Gypsy was the first retrovirus described in invertebrates (KIM et al. 1994B ; SONG et al. 1994 ). As shown in Fig 1A, the genomic structure of this retroelement is remarkably similar to the proviral form of vertebrate retroviruses (MARLOR et al. 1986 ). Gypsy proviruses contain two long terminal repeats (LTRs). Their transcription is initiated in the 5'-LTR by a TATA-less promoter, which contains a downstream promoter element (DPE) located 30 bp after the transcription start (ARKHIPOVA and ILYIN 1992 ). Transcription is terminated in the 3'-LTR, giving rise to a full-length RNA. By analogy to vertebrate retroviruses, this RNA is thought to be used both as a messenger RNA, to translate, respectively, the Gag and Pol structural and enzymatic peptides, and as a genomic RNA, packaged into virus-like particles. The env messenger RNA that encodes the Env protein is a splicing derivative of the full-length gypsy transcript from which the gag and pol genes are excised (PELISSON et al. 1994 ). The expression of gypsy is restricted to the salivary gland precursors and fat body of the embryo, imaginal discs and fat body of larvae, and fat body and ovaries of adult females (SMITH and CORCES 1995 ). Both retrovirus-like organization and expression are typical of the functional gypsy proviruses that have inserted recently into the euchromatic DNA of some strains. By contrast, the pericentromeric heterochromatic regions of all strains contain very old gypsy insertions, which have accumulated a lot of deleterious mutations and rearrangements (LAMBERTSSON et al. 1989 ; VAURY et al. 1989 ).

View larger version (63K): In this window In a new window Download PPT slide   Figure 1. Transgenes containing the first 328 nucleotides of gypsy are sensitive to the flam-dependent regulation. (A) Structure of the gypsy proviral DNA and schematic of the gypsy promoter. Coding regions of three genes, gag, pol, and env, are shown by open boxes. Shaded boxes represent the 482-bp long terminal repeats (LTR). Gypsy has a TATA-less promoter, which contains a DPE located 30 bp after the transcription start (ARKHIPOVA and ILYIN 1992 ). The broken arrow indicates the starting point and the orientation of transcription. The Su(Hw)-binding region is shaded in yellow and is denoted "Su(Hw) B. R." (B) Structure and regulation of the p#12 and p#15 constructs. p#12 is a transcriptional fusion of the gypsy promoter and leader with the prokaryotic lacZ reporter gene. The p#15 deletion derivative includes nucleotides 329–1072. In the inset table, the total numbers of transgenes that were made homozygous in both the w flamOR(P) permissive and the w flamRev(R) restrictive backgrounds are given in parentheses in addition to the number of regulated transgenes (i.e., those disclosing a markedly lower expression in the restrictive than in the permissive genotype). (C) flam-dependent follicle cell expression of two homozygous transgenes, #12a and #15(2), representative of the data obtained with the p#12 and p#15 constructs. The blue ß-gal staining is restricted to the permissive follicular epithelium that surrounds the grayish yolk-containing oocyte. (D) flam-dependent female gonadal expression of the #12a homozygous transgene. Third instar female larvae contain acorn-like small gonads. In the permissive gonad, lacZ is expressed nonuniformly throughout the "acorn-cup," including the equatorial region known to contain both precursors of the germinal and somatic cells of the ovary. The histochemical reaction was stopped after 4 hr at 37°.

Genetic analysis of gypsy mobility has shown that it is controlled by an as-yet-uncloned locus of the Drosophila genome. This locus, flamenco (flam), is located in the heterochromatin of the X chromosome (PRUD'HOMME et al. 1995 ; ROBERT et al. 2001 ). Two classes of alleles, restrictive and permissive, have been described. Restrictive alleles, denoted flam^(R), are dominant and repress transposition of functional gypsy proviruses. Strains containing flam^(R) alleles are therefore stable, as far as gypsy mobilization is concerned. Conversely, permissive alleles, denoted flam^(P), allow high levels of gypsy transposition, as illustrated by the two unstable strains MG and MS, which are both flam^(P) homozygous and contain a large number of functional proviruses in their genome. Most of the permissive stocks are stable because they are devoid of functional gypsy proviruses (KIM et al. 1994A ; PELISSON et al. 1997 ). A mutation from the restrictive toward the permissive allelic form was obtained by a P-element insertion (ROBERT et al. 2001 ). Moreover, deficiencies of the locus were shown to behave as permissive alleles (PRUD'HOMME et al. 1995 ). Hence, it is inferred that at least one of the functions of the putative Flamenco protein consists of downregulating gypsy mobilization. This is achieved by preventing the accumulation of first gypsy RNAs and then of virus-like particles into the somatic follicle cells of the restrictive ovaries (PELISSON et al. 1994 ; LECHER et al. 1997 ). However, it is not known whether the reduction of the steady-state level of gypsy RNAs in restrictive females occurs at the transcriptional or post-transcriptional level.

The purpose of the present work was to better understand the mechanisms of the flam-dependent regulation of gypsy. We started from a gypsy-lacZ construct containing the gypsy promoter and its 5'-untranslated region (UTR) fused to ß-galactosidase. This transcriptional fusion can recapitulate the repression of the endogenous proviruses in the restrictive follicle cells (PELISSON et al. 1994 ). Assuming the hypothesis that flam would encode a typical repressor, this putative Flamenco protein would bind to a cis-regulatory sequence that we should be able to characterize by a deletion mapping study of this construct. Alternatively, by analogy to the trans-silencing phenomenon controlling the P element (RONSSERAY et al. 2003 ), the hypothetical Flamenco protein might be involved in some cosuppression-like mechanism that would target any construct containing homology to a putative gypsy trigger element. In this case, the cis-regulatory sequence analysis of the gypsy-lacZ construct should shed some light on the sequence content of this putative silencer gypsy element. Results from this analysis indicated that all the tested fragments originating from the gypsy 5'-UTR could be targeted by the Flamenco repression. The only apparent requisite was that the target had to be included in the tested transcript. This prompted us to further investigate a possible RNA-silencing model (i) by looking for the presence in restrictive ovaries of small interfering-like RNAs (DYKXHOORN et al. 2003 ), which, by virtue of their homology to the targets, should be able to guide the silencing complex toward the gypsy-containing transcripts, and (ii) by testing the impact on gypsy silencing of piwi, a gene known to affect RNA-mediated homology-dependent transgene silencing (PAL-BHADRA et al. 2002 ).

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila strains:
If not otherwise indicated, genetic materials and fly stocks are described in FlyBase (http://flybase.bio.indiana.edu/). Flies were grown at 25° on standard Drosophila medium (GANS et al. 1975 ). The piwi² mutation, kindly provided by U. Bhadra, is a single P-ry¹¹ insertion (COX et al. 1998 ). It was balanced with the CyO; Act5c:GFP(w⁺) chromosome constructed by D. Ferrandon. The permissive wOR(P) and the restrictive wRev(R) strains are both devoid of functional gypsy proviruses and were described previously (MEJLUMIAN et al. 2002 ). By contrast, all the following stocks contain active gypsy proviruses. The A151(P) X chromosome contains a permissive flam allele. It is maintained in the A151(P)/FM7(R) balanced stock where gypsy mobilization is prevented by the FM7 dominant restrictive balancer, as described previously (PELISSON et al. 1994 ). The A151(R) X chromosome was isolated as a restrictive derivative of the A151(P) chromosomal clone. The A151(R)/FM7(R) restrictive and the A151(P)/FM7(R) permissive balanced stocks were isogenized thereafter. The MG1 restrictive strain was described previously (PRUD'HOMME et al. 1995 ).

Plasmid constructions and P-element-mediated germline transformations:
Several different P-element-containing plasmids were made, containing either the gypsy or the yp3 promoter.

Constructs containing the gypsy promoter: The pGypLTR1 construct was kindly provided by V. G. Corces. This pUC18-based plasmid contains a HpaI-XbaI insert corresponding to nucleotides 5–435 in the gypsy sequence (accession no. M12927) and therefore exhibits the following simplified restriction map: EcoRI-KpnI-SmaI/HpaI-BglII-XbaI. It was used to extract the 5–430 gypsy fragment as a EcoRI-KpnI-SmaI/HpaI-BglII fragment, which was then inserted into the corresponding restriction sites of the EcoRI-BglII-BamHI-PstI transformation plasmid pW6 (KLEMENZ et al. 1987 ) to produce the p#6 construct. The KpnI site of p#6 was cut with Acc65I and filled with Klenow to obtain p#6K. The BamHI-KpnI-AUGlacZ-PstI fragment of pCaSpeR-AUG-ßgal (THUMMEL et al. 1988 ) was inserted into p#6K at the corresponding restriction sites, i.e., downstream of the gypsy LTR, to generate p#8. A 5'-UTR gypsy fragment (419–1072) was amplified from the plasmid containing the gypsy insertion into the f¹ mutant (MARLOR et al. 1986 ), with primers containing BglII and KpnI restriction sites (respectively, 5'-GAAGGAAGATCTCTAGACCTACT-3' and 5'-ggggtaccTGGTTGGCACACCACAAA-3', with non-gypsy sequences in lowercase and restriction sites underlined). After digestion with BglII and KpnI, the PCR product was ligated with the BglII-KpnI-digested p#8 plasmid to obtain p#12. The p#12 construct was obtained in two steps: The 59-bp deletion was first generated in the gypsy sequence of pGypLTR1 with the Stratagene (La Jolla, CA) QuikChange site-directed mutagenesis kit (primer: 5'-CGAAATAAACCACAGCCCACAAGGCTAGTGATAATAACTAAGG-3'); the 435-bp EcoRI-BglII wild-type fragment of p#12 was then replaced by the corresponding 376-bp deleted fragment. The p#15 construct was obtained by replacing the XhoI-KpnI fragment of p#8 (which contains the 199–430 gypsy sequence) with the 199–328 gypsy sequence; the latter was provided by the XhoI-KpnI-digested PCR fragment that had been generated with the following primers on the p#12 template: 5'-ggaattCTCGAGGGTAAACTTAG-3' and 5'-ggggtaccatCGATAGCGATTTGATTGT-3'.

Constructs containing the yp3 promoter: The yp3fc plasmid was kindly provided by M. Bownes. This pBluescript-based plasmid contains in its EcoRI site the –285 to +43 fragment of the yp3 gene, sufficient to drive expression in the follicle cells (RONALDSON and BOWNES 1995 ). The Stratagene QuikChange site-directed mutagenesis kit was used to introduce a StuI restriction site at coordinate –42 using the oligonucleotide 5'-GCAGTGCGCTATCAGGCCTCGGAGCTATATAAG-3'. The pES2 parental construct was obtained by subcloning the 383-bp BamHI-EcoRI-StuI-EcoRI-HindIII-XhoI-KpnI mutated sequence upstream of AUGlacZ into the corresponding sites of the pCaSpeR-AUG-ßgal transformation vector (THUMMEL et al. 1988 ). The 270–328 gypsy sequence was obtained by annealing the two corresponding oligonucleotides. The pES3 and pES4 plasmids were constructed by blunt-end cloning this 59-bp fragment at the StuI restriction site of pES2 in the opposite and direct orientations, respectively. This small fragment was also ligated to the Klenow-blunted XhoI site of pES2 in the sense orientation to produce pES5. The Klenow-blunted XhoI site of pES2 was also used to insert the 329–426 and the 426–1072 gypsy fragments in the sense orientation, generating plasmids pES11 and pES9, respectively. Both these fragments originated from p#12 by either PCR or gel purification of a Klenow-blunted BglII-Acc65I fragment, respectively.

P-element-mediated transformations were performed as described (RUBIN and SPRADLING 1982 ). The constructs were coinjected with the pUChs2-3 helper (gift of D. C. Rio) into embryos of the wOR(P) permissive stock. Flies carrying the insertion were identified by rescue of the white phenotype. Inserts in transgenic flies were made homozygous in this w flam^OR(P) permissive background. They were also backcrossed twice to wRev(R) females and then made homozygous in this restrictive background, too. Each pair of transgenes was analyzed (XbaI-restricted genomic DNA probed with the SV40 trailer) to check that the same transgene was indeed present in both backgrounds.

Histochemical analysis of ß-galactosidase:
Ovaries were fixed in 2% formaldehyde, 0.2% (vol/vol) glutaraldehyde/PBS for 5 min at room temperature, washed twice with PBS, rinsed once in staining buffer (1 mM MgCl₂/4 mM K₃Fe(CN)₆/4 mM K₄Fe(CN)₆/1% Triton X-100), and stained for 0.5–4 hr at 37° in staining buffer containing 0.27% X-gal (5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside). For the sake of comparison, each pair of permissive and restrictive genotypes containing the same homozygous transgene was treated simultaneously. The reaction was stopped in PBS and the ovaries were mounted in 90% (vol/vol) glycerol in PBS and viewed in a Leica DMRB microscope.

Wandering female larvae were inverted after cutting off the head. They were treated as above except that the fixative concentration was reduced twofold, the X-gal concentration doubled, and the reaction stopped after 4 hr at 37°. After staining, females gonads were dissected out of the fat body in a drop of 90% glycerol and viewed directly in the dissecting microscope without mounting. Ovary and gonad photomicrographs were taken with a Coolpix990 digital camera (Nikon) and panels were constructed using Adobe Photoshop 5.5.

Sequence alignments:
In an attempt to find common motifs that would explain why gypsy sequences 270–328, 329–426, and 426–1072 share the same regulatory potential, the sequences were compared using the following algorithms: Align (http://www.infobiogen.fr); Multalin (http://probes.toulouse-inra.fr/multalin); Alignment (http://www.genebee.msu.su/cgi-bin/nph-malign.pl); MAP Multiple Sequence Alignment (http://searchlauncher.bcm.tmc.edu/cgi-bin/multi-align.pl); EDTALN (http://www.infobiogen.fr/services/analyseq); Clustalw (http://www.infobiogen.fr/services/analyseq) and MEME (http://meme.sdsc.edu/meme/website/meme.html). Although the latter analysis did disclose a common motif of 10 nucleotides, this similarity turned out not to be significant when compared to the results of the shuffled run.

Detection of small RNA species:
The detection of small RNA was performed essentially according to HAMILTON and BAULCOMBE 1999 with the following modifications: RNA isolated from 25 females using the Trizol reagent was enriched in low-molecular-weight species by precipitation in 5% polyethylene glycol 8000/0.5 M NaCl, loaded in each lane of a 15% denaturating acrylamide:bisacrylamide (19:1) gel, transferred onto Hybond NX membrane (Amersham, Buckinghamshire, UK), and fixed by ultraviolet crosslinking. Before transfer, the gel was imaged after staining in 2 µg/ml ethidium bromide. Hybridization with a short sense gypsy riboprobe (coordinates 310–433 in the gypsy sequence) was performed overnight at 40°; the membrane was washed twice for 30 min at 50° in 2x SSPE/0.5% SDS and exposed for 2 days on a phosphorimager screen (Molecular Dynamics, Sunnyvale, CA). The blots were stripped for 1 min at 90° in 0.2% SDS/10 mM Tris pH 7.5 and probed with an end-labeled oligonucleotide (5'-ACTCGTCAAAATGGCTGTGATA-3') complementary to the mir-13b-1 microRNA.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Larval female gonads also display a flam-dependent gypsy expression:
The p#12 construct contains the promoter and the 5'-UTR of gypsy, transcriptionally fused to the Escherichia coli lacZ reporter (Fig 1B). It is very similar to the previously described pgypCaSpeR plasmid (SMITH and CORCES 1995 ), which is known to have a flam-dependent expression (PELISSON et al. 1994 ). Three transgenic insertions, #12a, -b, and -c, were obtained from this construct and were checked for display of the expected flam-dependent expression in the ovary: lacZ staining was strong in follicle cells of the permissive, but not the restrictive, egg chambers at stages 8–10 (Fig 1B and Fig C, left). Moreover, the female gonads of third instar larvae were also found to respond to the Flamenco regulation (Fig 1D): lacZ staining was observed only in the permissive, not the restrictive, gonads.

The repression of gypsy by the putative Flamenco protein does not require the Su(Hw)-binding region:
Upon binding to the insulator located in the 3' half of the gypsy 5'-UTR, the Su(Hw) protein has been shown to activate the ovarian ß-galactosidase expression of the pgypCaSpeR gypsy-lacZ transgenes (SMITH and CORCES 1995 ). This experiment was done in the Df(1)w, y w^67c23 strain, which has since been shown to be a permissive flam mutant (data not shown). We therefore wondered whether the enhancer activity of the Su(Hw) protein would also have been observed in a restrictive genotype. In other words, could the repressive function of flam be mediated by some competition with the enhancer located in the Su(Hw)-binding region? This hypothesis was clearly ruled out by the observation that the p#15 deleted construct is sensitive to the Flamenco repression in ovaries, even though it is missing the Su(Hw)-binding region (Fig 1B and Fig C, right). As expected, in the permissive background, the #15 transgenes were consistently less expressed than the #12 transgenes. However, since the p#15 deletion also uncovers upstream sequences, this effect might be due to more than the mere absence of the Su(Hw)-binding region. The expression of the #15 transgenes in permissive female gonads was too low for their regulation to be tested in the restrictive background (data not shown).

The restrictive repression does not require the gypsy promoter:
The p#15 deleted construct is a good target for the restrictive function despite the fact that it contains only 328 bp of gypsy sequence, corresponding mostly to the promoter (coordinates 1–270) followed by 60 bp of nonpromoter transcribed sequences (Fig 1A and Fig B). To test whether the gypsy promoter is dispensable for this regulation, we chose to swap it for an alternative promoter from the yp3 gene. The enhancer located immediately upstream of this promoter is known to drive expression in the follicular epithelium of stage 10 egg chambers (RONALDSON and BOWNES 1995 ), where gypsy itself is expressed. As a first control, we checked that the pES2 corresponding reporter construct was not repressed in a restrictive background (Fig 2B and Fig C). Then the nonpromoter gypsy sequence present in p#15 (270–328) was introduced downstream of this yp3 ovarian promoter in the sense orientation. The resulting pES5 construct was repressed by the restrictive flam genotype (Fig 2B and Fig C), indicating that this small 59-bp fragment is sufficient to make a non-gypsy-driven transcription unit sensitive to this regulation.

View larger version (63K): In this window In a new window Download PPT slide   Figure 2. Any fragment from the gypsy 5'-UTR, when inserted into the yp3-lacZ heterologous transcription unit, can be a target of the flam-dependent regulation. (A) Coordinates in the gypsy sequence of each of the three fragments tested. These three fragments scan the whole nonpromoter gypsy sequence present in the p#12 construct. (B) Structure and regulation of the constructs. p#12 differs from p#12 only by the deletion of nucleotides 270–328. pES2 is a transcriptional fusion of the yp3 minimal ovarian promoter (cross-hatched fragment containing the ovarian enhancer shown by an open box) with the prokaryotic lacZ reporter gene. pES2 contains two unique restriction sites (StuI and XhoI) that were used to insert one of the three different gypsy fragments upstream or downstream of the promoter, giving rise to the five other pES constructs. In the inset table the numbers of the corresponding transgenes that were sensitive to the Flamenco repression are given (as in Fig 1). (C) Representative examples of the expression of four of the seven constructs in adult ovaries. For each construct, a pair of stocks homozygous for a given transgene in both the permissive and restrictive genotypes is presented.

When fused to a heterologous transcript, any fragment from the gypsy 5'-UTR appears to be able to target the repression:
To determine whether this 59-bp cis-regulatory sequence is the only possible target of the flam-dependent regulation, we specifically deleted it from the original p#12 gypsy construct. As shown in Fig 2B and Fig C, this deletion did not prevent the #12 transgene from being repressed in the restrictive background, suggesting some level of redundancy in the gypsy control mechanism.

As a first step to dissect the target sequence(s) left in the p#12 construct, the nonpromoter sequence present in this construct (329–1072) was split into two fragments (329–426 and 426–1072), which were both tested in the same way as the 59-bp fragment. Both resulting transgenes (ES11 and ES9) displayed typical regulation (Fig 2B), indicating the existence of at least three different target sequences in the 5'-UTR of gypsy. Several unsuccessful attempts were made to align these three sequences (see MATERIALS AND METHODS), making very unlikely the hypothesis that they would share redundant binding sites for a repressing transcription factor putatively encoded by flam. The only feature common to the gypsy sequences present in pES5, pES9, and pES11 might be that each of them would be homologous to some hypothetical gypsy element(s) that would trigger a homology-dependent silencing-like mechanism.

This putative trigger would target RNA rather than DNA, since the 59-bp sequence does not behave as a target when inserted upstream of the same yp3 promoter (see the ES3 and ES4 transgenes in Fig 2B and Fig C).

The repression of gypsy is impaired in piwi mutant larval female gonads:
The Drosophila argonaute-like piwi gene was reported to be involved in the Adh transgene cosuppression (PAL-BHADRA et al. 2002 ). Hence, we addressed the question of whether piwi would also affect the apparently similar phenomenon of gypsy silencing. Following the mating scheme described in Fig 3, the piwi² mutant allele was introduced into a restrictive background homozygous for the #12a gypsy-lacZ transgene. The stock thus obtained generates both homozygous and heterozygous piwi females, which can be distinguished, at any developmental stage, by the green fluorescent protein (GFP) fluorescence carried by the latter on their CyGFP balancer chromosome. The restrictive genotype of this stock was confirmed by the absence of any lacZ expression in the heterozygous adult ovaries (data not shown). The morphology of homozygous piwi² ovaries is too severely affected for any stage 10 to be recognized and the lacZ expression to be monitored in the corresponding follicle cells. We took advantage of the larval gonadal gypsy regulation described above to study the impact of the piwi² mutation. As expected, no staining was observed in the heterozygous third instar larval female gonads, confirming the restrictive genotype of the stock (Fig 3). By contrast, homozygous piwi² gonads disclosed lacZ expression in spite of their restrictive genotype. The staining pattern appeared to correspond to the apical-most part of the pattern observed for gypsy expression in a flam permissive genotype (see Fig 1D). This derepression suggests that the mechanism of gypsy silencing involves the Piwi function.

View larger version (38K): In this window In a new window Download PPT slide   Figure 3. Gypsy derepression in restrictive piwi mutant larval female gonads. (A) Mating scheme to produce a wRev(R) restrictive balanced stock containing both a piwi2 mutation and the #12a regulated transgene. Three markers were used to keep track of the various chromosomes: GFP stands for the Act5c:GFP(w+) construct, which can be viewed by epifluorescence at any postembryonic developmental stage; the brown-eyed phenotype of this transgene is not epistatic over the red-eyed color of the #12a transgene; and the TM3 balancer was followed by the Sb marker. (B) Derepression of the #12a transgene in wRev(R) restrictive gonads homozygous for the piwi2 mutation. The histochemical reaction was stopped after 4 hr at 37°. The lacZ staining is less widespread than that in the corresponding permissive piwi+ gonads (see Fig 1D). It is mostly located in the central region of the gonad. Gonads were photographed and imaged exactly like those in Fig 1.

Evidence for small gypsy RNAs:
Collectively, these data (involvement of Piwi and redundancy of the gypsy targets and their apparent need to be transcribed) provide circumstantial evidence for a homology-dependent RNA-silencing mechanism triggered by unknown gypsy sequence(s). Such a hypothesis assumes the existence of antisense small gypsy RNAs, which should mediate the silencing information between the hypothetical trigger and the targets. This prediction was tested by Northern blot hybridization with a riboprobe from within the gypsy 5'-UTR (coordinates 310–433). Small RNAs (25–27 nt) were easily detected in the restrictive ovaries and whole females (Fig 4). The signal disclosed by both types of permissive females tested here, A151(P) and wOR(P), was about twofold fainter than that of their isogenic restrictive counterparts, A151(R) and wRev(R), respectively. However, in view of the large interstrain variations, it would be premature to infer that the flam genotype affects the abundance of the small gypsy RNAs.

View larger version (55K): In this window In a new window Download PPT slide   Figure 4. Detection of small RNAs homologous to sequences from within the gypsy 5'-UTR. Low-molecular-weight-enriched RNAs extracted from either ovaries (A) or whole females (B) were hybridized with a sense gypsy riboprobe (coordinates 310–433 in the gypsy sequence; accession no. M12927). The names of the strains studied in this experiment are indicated above the corresponding lanes; A151(R)/A151(R) and A151(P)/A151(P) homozygous females were isolated from the A151(R)/FM7(R) and A151(P)/FM7(R) balanced stocks, respectively (see MATERIALS AND METHODS for a description of these strains). After stripping, the membranes were probed with an end-labeled oligonucleotide complementary to the abundant mir-13b-1 microRNA. The sizes of most gypsy small RNAs were estimated to range between 25 and 27 by extrapolation of this ladder of bands step by step down to the position of this 22-nt microRNA. Thirty nucleotides corresponds to the approximate position of the 2S rRNA as visualized by ethidium bromide staining before transfer.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Tissue specificity of the flam-dependent control of gypsy:
In adult permissive females, gypsy-lacZ transgenes are expressed in both the fat body and the ovarian follicle cells, but the repression is observed only in restrictive ovaries, not in restrictive carcasses (PELISSON et al. 1994 ). We report here another tissue-specific restriction of the flam-dependent control of gypsy, namely of the third instar larval female gonad. Other larval tissues, such as the imaginal discs and larval fat body (SMITH and CORCES 1995 ), do exhibit some expression of the gypsy-lacZ transgenes but, unlike the female gonad, they are not sensitive to the flam genotype (data not shown). Gypsy is expressed in only the basal half of the gonad. This flam-dependent expression pattern includes the central region, which is known to contain both precursors of the germinal and somatic cells of the ovary (GODT and LASKI 1995 ). Since we have not yet identified this new gypsy-expressing cell type, it remains to be seen whether or not this flam-dependent control of gypsy is also restricted to the somatic cells, as it is in the adult ovary.

Circumstantial evidence for a cosuppression-like mechanism involved in the flam-dependent control of gypsy:
The experiments presented in this article were initiated to test the hypothesis that the control of gypsy would operate via a classical transcriptional repressor encoded by flam. This simple hypothesis was systematically ruled out by each of the five following lines of circumstantial evidence:

1. The gypsy promoter sequences are not required for this regulation to operate. The yp3 promoter belongs to a different class from that of the gypsy promoter (presence of a TATA box and absence of a DPE) and does not display a flam-dependent expression. Nevertheless, the yp3-lacZ heterologous transcription unit could be made sensitive to the Flamenco repression by incorporation of nonpromoter gypsy sequences.

2. Three nonhomologous fragments of the gypsy 5'-UTR were shown to be good targets of the repression. Since these fragments do not share any common sequence, the three corresponding hypothetical Flamenco binding sites, scattered over several hundreds of nucleotides, should be very degenerate, although functionally redundant (each sequence is indeed able, by itself, to mediate a strong level of repression).

3. Unlike the usual silencers, these regulatory sequences apparently need to be located downstream of the heterologous promoter, suggesting that they have to be transcribed and targeted as RNA.

4. Small RNAs (25–27 nt) were detected in ovaries by hybridization to a 120-nt probe from the gypsy 5'-UTR. This class of RNAs is considered a hallmark of RNA-mediated silencing phenomena in eukaryotic organisms (HANNON 2002 ). It is worth noting that RNA silencing is considered an adaptative defense mechanism developed by eukaryotic genomes to control genetic parasites like viruses and transposable elements (WATERHOUSE et al. 2001 ).

5. Finally, the involvement in this regulation of the argonaute-like piwi gene, which was already reported to affect Adh cosuppression (PAL-BHADRA et al. 2002 ), is additional circumstantial evidence for a RNA-mediated cosuppression mechanism. Although of unknown biochemical function, several members of the Argonaute gene family are indeed among the most conserved universal components of the RNA-silencing process (CARMELL et al. 2002 ). If this effect of the pleiotropic piwi mutation is not indirect, it might provide the first hint for a role of an Argonaute-like gene in the control of transposable elements.

About the putative trigger(s) of this hypothetical cosuppression phenomenon:
In the aforementioned Drosophila cosuppression model, multicopy Adh transgenes can trigger the repression of the endogenous Adh gene together with their own repression. In the case of the gypsy regulation, we do not know what the trigger(s) would look like, except that it (they) should at least contain some homology to each of the three fragments identified as regulatory targets in the gypsy 5'-UTR. Since the repression can be observed even in those restrictive strains which, like wRev(R), are devoid of active gypsy proviruses, trigger(s) must be searched for among the defective elements shared by all strains and located mainly in the pericentromeric heterochromatin (LAMBERTSSON et al. 1989 ; VAURY et al. 1989 ). In Drosophila, cosuppression is involved in the repression of several transposable elements and repeated sequences, including the P element (RONSSERAY et al. 2003 ), the Stellate repeats (ARAVIN et al. 2001 ), and the I element (CHABOISSIER et al. 1998 ; JENSEN et al. 1999 ; MALINSKY et al. 2000 ; ROBIN et al. 2003 ). The fact that heterochromatic gypsy elements might be responsible for the silencing of the euchromatic ones recalls that, in the case of P and Stellate, the triggers are located in the telomeric and the pericentromeric heterochromatin, respectively. In the case of the I element, it is worth noting that pericentromeric ancestral I-like sequences were also suggested to relay I-element cosuppression (BUCHETON et al. 2002 ; JENSEN et al. 2002 ).

Transcriptional vs. post-transcriptional silencing:
Transposable elements are known to be repressed at both transcriptional (OKAMOTO and HIROCHIKA 2001 ) and post-transcriptional levels (WU-SCHARF et al. 2000 ). The piwi Drosophila gene is required not only for the post-transcriptional but also for some aspects of the transcriptional Adh cosuppression phenomenon (PAL-BHADRA et al. 2002 ). Moreover, recent data indicate that RNA silencing may also result in transcriptional inactivation mediated at the level of chromatin (VOLPE et al. 2002 ). Thus, the effect of the piwi² mutation cannot unambiguously relate to the level at which the gypsy silencing acts and neither can the apparent need for transcription of the gypsy target, nor even the presence of short gypsy RNAs. The differential sensitivity of the spliced and unspliced gypsy RNAs to the restrictive repression (PELISSON et al. 1994 ) would support the post-transcriptional hypothesis, but this question awaits direct testing by experiments such as run-off transcription of ovarian nuclei.

Roles of flam and piwi in this silencing:
The present results suggest that RNA silencing might be responsible for the repression of gypsy in the ovaries that contain functional flam and piwi alleles. Assuming this hypothesis, it is not clear whether flam is involved in the accumulation of the short gypsy RNAs. It is difficult to define a role for flam in this process until it has been clearly identified at the molecular level. It is nevertheless striking that inactivation of this restrictive function was caused by a P-element insertion (ROBERT et al. 2001 ) just 1.7 kb upstream of a gene, DIP1, which was subsequently shown to bind double-stranded RNA (DESOUSA et al. 2003 ), a typical trigger of RNA silencing. DIP1 discloses a ubiquitous expression pattern, including the ovarian tissues in which the Flamenco repression was shown to operate.

PIWI belongs to a highly conserved protein family that is involved in a variety of RNA-silencing phenomena in a diverse set of organisms (CARMELL et al. 2002 ). PIWI itself affects the Adh transgene cosuppression process (PAL-BHADRA et al. 2002 ). To know whether both genes are involved in the same regulatory pathway(s), we are planning to monitor the level of gypsy expression in the flam piwi double mutant. It would also be interesting to test whether flam affects Adh cosuppression, as well as other RNA silencing phenomena. Finally, when the flam product(s) can be detected, it will be worth looking at the possible effect of piwi on the expression of flam itself.

	ACKNOWLEDGMENTS

We thank Maryvonne Mevel-Ninio, Isabelle Busseau, Nicolas Gilbert, and Marc Greener for critically reading the manuscript. The Northern experiments were performed following a detailed protocol kindly provided by Utpal Bhadra and with the help of Alexei Aravin and Natalia Naumova. E.S. was the recipient of a fellowship from the French government (Ministère de l'Enseignement Supérieur et de la Recherche). This work was supported by grants from the Centre National pour la Recherche Scientifique, the Association pour la Recherche sur le Cancer, and the Fondation pour la Recherche Médicale Française.

Manuscript received July 24, 2003; Accepted for publication December 10, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ARAVIN, A. A., N. M. NAUMOVA, A. V. TULIN, V. V. VAGIN, and Y. M. ROZOVSKY et al., 2001  Double-stranded RNA-mediated silencing of genomic tandem repeats and transposable elements in the D. melanogaster germline. Curr. Biol. 11:1017-1027.[CrossRef][Medline]

ARKHIPOVA, I. R. and Y. V. ILYIN, 1992  Control of transcription of Drosophila retrotransposons. BioEssays 14:161-168.[CrossRef][Medline]

BUCHETON, A., I. BUSSEAU and D. TENINGES, 2002 I elements in Drosophila melanogaster, pp. 796–812 in Mobile DNA II, edited by N. L. CRAIG, R. CRAIGIE, M. GELLERT and A. M. LAMBOWITZ. American Society for Microbiology, Washington, DC.

CARMELL, M. A., Z. XUAN, M. Q. ZHANG, and G. J. HANNON, 2002  The Argonaute family: tentacles that reach into RNAi, developmental control, stem cell maintenance, and tumorigenesis. Genes Dev. 16:2733-2742.[Free Full Text]

CHABOISSIER, M. C., A. BUCHETON, and D. J. FINNEGAN, 1998  Copy number control of a transposable element, the I factor, a LINE-like element in Drosophila. Proc. Natl. Acad. Sci. USA 95:11781-11785.[Abstract/Free Full Text]

COX, D. N., A. CHAO, J. BAKER, L. CHANG, and D. QIAO et al., 1998  A novel class of evolutionarily conserved genes defined by piwi are essential for stem cell self-renewal. Genes Dev. 12:3715-3727.[Abstract/Free Full Text]

DESOUSA, D., M. MUKHOPADHYAY, P. PELKA, X. ZHAO, and B. K. DEY et al., 2003  A novel dsRNA-binding protein, DIP1 (Disco interacting protein 1), contributes to cell fate decisions during Drosophila development. J. Biol. Chem. 278:38040-38050.[Abstract/Free Full Text]

DESSET, S., C. MEIGNIN, B. DASTUGUE, and C. VAURY, 2003  COM, a heterochromatic locus governing the control of independent endogenous retroviruses from Drosophila melanogaster. Genetics 164:501-509.[Abstract/Free Full Text]

DYKXHOORN, D. M., C. D. NOVINA, and P. A. SHARP, 2003  Killing the messenger: short RNAs that silence gene expression. Nat. Rev. Mol. Cell Biol. 4:457-467.[CrossRef][Medline]

GANS, M., C. AUDIT, and M. MASSON, 1975  Isolation and characterization of sex-linked female-sterile mutants in Drosophila melanogaster.. Genetics 81:683-704.[Abstract/Free Full Text]

GODT, D. and F. A. LASKI, 1995  Mechanisms of cell rearrangement and cell recruitment in Drosophila ovary morphogenesis and the requirement of bric a brac. Development 121:173-187.[Abstract]

HAMILTON, A. J. and D. C. BAULCOMBE, 1999  A species of small antisense RNA in post-transcriptional gene silencing in plants. Science 286:950-952.[Abstract/Free Full Text]

HANNON, G. J., 2002  RNA interference. Nature 418:244-251.[CrossRef][Medline]

JENSEN, S., M. P. GASSAMA, and T. HEIDMANN, 1999  Taming of transposable elements by homology-dependent gene silencing. Nat. Genet. 21:209-212.[CrossRef][Medline]

JENSEN, S., M. P. GASSAMA, X. DRAMARD, and T. HEIDMANN, 2002  Regulation of I-transposon activity in Drosophila: evidence for cosuppression of nonhomologous transgenes and possible role of ancestral I-related pericentromeric elements. Genetics 162:1197-1209.[Abstract/Free Full Text]

KIM, A. I., E. S. BELYAEVA, and M. M. ASLANIAN, 1990  Autonomous transposition of gypsy mobile elements and genetic instability in Drosophila melanogaster. Mol. Gen. Genet. 224:303-308.[CrossRef][Medline]

KIM, A. I., N. V. LYUBOMIRSKAYA, E. S. BELYAEVA, N. G. SHOSTACK, and Y. V. ILYIN, 1994a  The introduction of a transpositionally active copy of retrotransposon gypsy into the stable strain of Drosophila melanogaster causes genetic instability. Mol. Gen. Genet. 242:472-477.[Medline]

KIM, A. I., C. TERZIAN, P. SANTAMARIA, A. PÉLISSON, and N. PRUD'HOMME et al., 1994b  Retroviruses in invertebrates: the gypsy retrotransposon is apparently an infectious retrovirus of Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 91:1285-1289.[Abstract/Free Full Text]

KLEMENZ, R., U. WEBER, and W. J. GEHRING, 1987  The white gene as a marker in a new P-element vector for gene transfer in Drosophila.. Nucleic Acids Res. 15:3947-3959.[Abstract/Free Full Text]

LAMBERTSSON, A., S. ANDERSSON, and T. JOHANSSON, 1989  Cloning and characterization of variable-sized gypsy mobile elements in Drosophila melanogaster. Plasmid 22:22-31.[CrossRef][Medline]

LANDER, E. S., L. M. LINTON, B. BIRREN, C. NUSBAUM, and M. C. ZODY et al., 2001  Initial sequencing and analysis of the human genome. Nature 409:860-921.[CrossRef][Medline]

LECHER, P., A. BUCHETON, and A. PÉLISSON, 1997  Expression of the Drosophila retrovirus gypsy as ultrastructurally detectable particles in the ovaries of flies carrying a permissive flamenco allele. J. Gen. Virol. 78:2379-2388.[Abstract]

MALINSKY, S., A. BUCHETON, and I. BUSSEAU, 2000  New insights on homology-dependent silencing of I factor activity by transgenes containing ORF1 in Drosophila melanogaster. Genetics 156:1147-1155.[Abstract/Free Full Text]

MARLOR, R. L., S. M. PARKHURST, and V. G. CORCES, 1986  The Drosophila melanogaster gypsy transposable element encodes putative gene products homologous to retroviral proteins. Mol. Cell. Biol. 6:1129-1134.[Abstract/Free Full Text]

MEJLUMIAN, L., A. PELISSON, A. BUCHETON, and C. TERZIAN, 2002  Comparative and functional studies of Drosophila species invasion by the gypsy endogenous retrovirus. Genetics 160:201-209.[Abstract/Free Full Text]

MEVEL-NINIO, M., M. C. MARIOL, and M. GANS, 1989  Mobilization of the gypsy and copia retrotransposons in Drosophila melanogaster induces reversion of the ovoD dominant female-sterile mutations: molecular analysis of revertant alleles. EMBO J. 8:1549-1558.[Medline]

OKAMOTO, H. and H. HIROCHIKA, 2001  Silencing of transposable elements in plants. Trends Plant Sci. 6:527-534.[CrossRef][Medline]

PAL-BHADRA, M., U. BHADRA, and J. A. BIRCHLER, 2002  RNAi related mechanisms affect both transcriptional and post-transcriptional transgene silencing in Drosophila. Mol. Cell 9:315-327.[CrossRef][Medline]

PASYUKOVA, E., S. NUZHDIN, W. LI, and A. J. FLAVELL, 1997  Germ line transposition of the copia retrotransposon in Drosophila melanogaster is restricted to males by tissue-specific control of copia RNA levels. Mol. Gen. Genet. 255:115-124.[CrossRef][Medline]

PÉLISSON, A., S. U. SONG, N. PRUD'HOMME, P. A. SMITH, and A. BUCHETON et al., 1994  Gypsy transposition correlates with the production of a retroviral envelope-like protein under the tissue-specific control of the Drosophila flamenco gene. EMBO J. 13:4401-4411.[Medline]

PÉLISSON, A., L. TEYSSET, F. CHALVET, A. KIM, and N. PRUD'HOMME et al., 1997  About the origin of retroviruses and the co-evolution of the gypsy retrovirus with the Drosophila flamenco host gene. Genetica 100:29-37.[CrossRef][Medline]

PRUD'HOMME, N., M. GANS, M. MASSON, C. TERZIAN, and A. BUCHETON, 1995  Flamenco, a gene controlling the gypsy retrovirus of Drosophila melanogaster.. Genetics 139:697-711.[Abstract]

ROBERT, V., N. PRUD'HOMME, A. KIM, A. BUCHETON, and A. PELISSON, 2001  Characterization of the flamenco region of the Drosophila melanogaster genome. Genetics 158:701-713.[Abstract/Free Full Text]

ROBIN, S., S. CHAMBEYRON, A. BUCHETON, and I. BUSSEAU, 2003  Gene silencing triggered by non-LTR retrotransposons in the female germline of Drosophila melanogaster. Genetics 164:521-531.[Abstract/Free Full Text]

RONALDSON, E. and M. BOWNES, 1995  Two independent cis-acting elements regulate the sex- and tissue-specific expression of yp3 in Drosophila melanogaster. Genet. Res. 66:9-17.[Medline]

RONSSERAY, S., T. JOSSE, A. BOIVIN, and D. ANXOLABEHERE, 2003  Telomeric transgenes and trans-silencing in Drosophila. Genetica 117:327-335.[CrossRef][Medline]

RUBIN, G. M. and A. C. SPRADLING, 1982  Genetic transformation of Drosophila with transposable element vectors. Science 218:348-353.[Abstract/Free Full Text]

SMITH, P. A. and V. G. CORCES, 1995  The suppressor of hairy-wing protein regulates the tissue-specific expression of the Drosophila gypsy retrotransposon. Genetics 139:215-228.[Abstract]

SONG, S. U., T. GERASIMOVA, M. KURKULOS, J. D. BOEKE, and V. G. CORCES, 1994  An env-like protein encoded by a Drosophila retroelement: evidence that gypsy is an infectious retrovirus. Genes Dev. 8:2046-2057.[Abstract/Free Full Text]

THUMMEL, C. S., A. M. BOULET, and H. D. LIPSHITZ, 1988  Vectors for Drosophila P-element-mediated transformation and tissue culture transfection. Gene 74:445-456.[CrossRef][Medline]

VAURY, C., A. BUCHETON, and A. PELISSON, 1989  The beta heterochromatic sequences flanking the I elements are themselves defective transposable elements. Chromosoma 98:215-224.[CrossRef][Medline]

VOLPE, T. A., C. KIDNER, I. M. HALL, G. TENG, and S. I. GREWAL et al., 2002  Regulation of heterochromatic silencing and histone H3 lysine-9 methylation by RNAi. Science 297:1833-1837.[Abstract/Free Full Text]

WATERHOUSE, P. M., M. B. WANG, and T. LOUGH, 2001  Gene silencing as an adaptive defence against viruses. Nature 411:834-842.[CrossRef][Medline]

WU-SCHARF, D., B. JEONG, C. ZHANG, and H. CERUTTI, 2000  Transgene and transposon silencing in Chlamydomonas reinhardtii by a DEAH-box RNA helicase. Science 290:1159-1162.[Abstract/Free Full Text]

This article has been cited by other articles:

				H. Ohta, M. Fujiwara, Y. Ohshima, and T. Ishihara ADBP-1 Regulates an ADAR RNA-Editing Enzyme to Antagonize RNA-Interference-Mediated Gene Silencing in Caenorhabditis elegans Genetics, October 1, 2008; 180(2): 785 - 796. [Abstract] [Full Text] [PDF]

				S. Kuramochi-Miyagawa, T. Watanabe, K. Gotoh, Y. Totoki, A. Toyoda, M. Ikawa, N. Asada, K. Kojima, Y. Yamaguchi, T. W. Ijiri, et al. DNA methylation of retrotransposon genes is regulated by Piwi family members MILI and MIWI2 in murine fetal testes Genes & Dev., April 1, 2008; 22(7): 908 - 917. [Abstract] [Full Text] [PDF]

				C. Klattenhoff and W. Theurkauf Biogenesis and germline functions of piRNAs Development, January 1, 2008; 135(1): 3 - 9. [Abstract] [Full Text] [PDF]

				K. M. Nishida, K. Saito, T. Mori, Y. Kawamura, T. Nagami-Okada, S. Inagaki, H. Siomi, and M. C. Siomi Gene silencing mechanisms mediated by Aubergine piRNA complexes in Drosophila male gonad RNA, November 1, 2007; 13(11): 1911 - 1922. [Abstract] [Full Text] [PDF]

				B. Brower-Toland, S. D. Findley, L. Jiang, L. Liu, H. Yin, M. Dus, P. Zhou, S. C.R. Elgin, and H. Lin Drosophila PIWI associates with chromatin and interacts directly with HP1a Genes & Dev., September 15, 2007; 21(18): 2300 - 2311. [Abstract] [Full Text] [PDF]

				Y. Kirino and Z. Mourelatos The mouse homolog of HEN1 is a potential methylase for Piwi-interacting RNAs RNA, September 1, 2007; 13(9): 1397 - 1401. [Abstract] [Full Text] [PDF]

				J. V. Hartig, Y. Tomari, and K. Forstemann piRNAs--the ancient hunters of genome invaders Genes & Dev., July 15, 2007; 21(14): 1707 - 1713. [Abstract] [Full Text] [PDF]

				M. Mevel-Ninio, A. Pelisson, J. Kinder, A. R. Campos, and A. Bucheton The flamenco Locus Controls the gypsy and ZAM Retroviruses and Is Required for Drosophila Oogenesis Genetics, April 1, 2007; 175(4): 1615 - 1624. [Abstract] [Full Text] [PDF]

				A. Pelisson, E. Sarot, G. Payen-Groschene, and A. Bucheton A Novel Repeat-Associated Small Interfering RNA-Mediated Silencing Pathway Downregulates Complementary Sense gypsy Transcripts in Somatic Cells of the Drosophila Ovary J. Virol., February 15, 2007; 81(4): 1951 - 1960. [Abstract] [Full Text] [PDF]

				A. Pelisson, G. Payen-Groschene, C. Terzian, and A. Bucheton Restrictive flamenco Alleles Are Maintained in Drosophila melanogaster Population Cages, Despite the Absence of Their Endogenous gypsy Retroviral Targets Mol. Biol. Evol., February 1, 2007; 24(2): 498 - 504. [Abstract] [Full Text] [PDF]