COM, a Heterochromatic Locus Governing the Control of Independent Endogenous Retroviruses From Drosophila melanogaster

Corresponding author: Chantal Vaury, Faculté de Médecine, BP38, 28 place Henri-Dunant, 63001 Clermont-Ferrand Cedex, France., chantal.vaury{at}inserm.u-clermont1.fr (E-mail)

Communicating editor: M. VEUILLE

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ZAM and Idefix are two endogenous retroviruses whose expression is tightly controlled in Drosophila melanogaster. However, a line exists in which this control has been perturbed, resulting in a high mobilization rate for both retroviruses. This line is called the U (unstable) line as opposed to the other S (stable) lines. In the process of analyzing this control and tracing the genetic determinant involved, we found that ZAM and Idefix expression responded to two types of controls: one restricting their expression to specific somatic cells in the ovaries and the other silencing their expression in S lines but permitting it in U lines. While studying this second control in the U or S backgrounds, we found that the heterochromatic locus 20A2-3 on the X chromosome, previously implicated in the regulation of a third retroelement, gypsy, also controlled both ZAM and Idefix. We report here that genetic determinants necessary for endogenous retrovirus silencing occur at the 20A2-3 locus, which we call COM, for centre organisateur de mobilisation. We propose that if this point of control becomes mutated during the life of the fly, it may trigger processes reactivating dormant endogenous retroviruses and thus bring about sudden bursts of mobilization.

TRANSPOSABLE elements are a substantial component of the eukaryotic genome. Current estimates predict that retroelements that replicate by transcription of an RNA intermediate, subsequent reverse transcription, and insertion of a new copy elsewhere in the genome make up at least 40% of a mammalian genome (SMIT 1999 ). The tight control of their mobilization is thus an important feature of their regulation to prevent high mutation rates due to novel insertions. This implies that their transposition is normally a rare event. However, current evidence indicates that these elements are not always inert components of the genome (WHITELAW and MARTIN 2001 ), and Drosophila offers several examples of genetic instability associated with massive mobilization of mobile elements. For instance, elements such as I, P, and hobo are three transposable elements that recently invaded the genome of Drosophila melanogaster. They have only a limited transpositional activity within this genome, but when they are introduced by crossing into Drosophila genomes devoid of such elements, a high-frequency transposition of the incoming transposon is observed. Several studies report that multiple transposable elements may be reactivated and exhibit simultaneous mobilization in the genome of their host in which they previously had a limited transposition rate (LIM et al. 1983 ; PASYUKOVA and NUZHDIN 1993 ). A system of hybrid dysgenesis in which at least four unrelated transposable elements are mobilized following dysgenic crosses has been described for D. virilis (PETROV et al. 1995 ). The occurrence of simultaneous mobilization of more than one type of element suggests that they may share a common pathway of regulation in their host.

We previously reported the identification of a strain, Rev, that had suffered a recent massive amplification of two retroelements, ZAM and Idefix. We found that all D. melanogaster strains contain either no copies or very few copies of ZAM and Idefix outside the chromocenter. In contrast, Rev and its derivatives have 20–30 copies of these elements dispersed on the chromosomal arms (DESSET et al. 1999 ). In the present study, we investigated the control exerted by the host and governing ZAM and Idefix invasion of Rev genome and not the other strains of D. melanogaster. Previous experiments indicated that this regulation was exerted on their RNA products. ZAM transcripts are detected only in the posterior follicle cells of the ovaries of Rev lines, which will be referred to as unstable lines or U lines, and not in the ovaries of other D. melanogaster lines, referred to as stable lines or S lines. Similarly, Idefix transcripts are detected in the germarium of U-line but not of S-line ovaries (TCHERESSIZ et al. 2002 ).

In this study, we show that the control of ZAM and Idefix mobilization is conferred by a genetic determinant located at 20A2-3 in the pericentromeric region of the X chromosome. With the aid of transgenic lines bearing the long terminal repeat (LTR) of ZAM or Idefix fused to LacZ, we show that the genetic determinant in U lines is dominant or semidominant over the determinant in S lines. Similarly, gypsy, another endogenous D. melanogaster retrovirus that is stable in most Drosophila strains, but is able to transpose at high frequency in a few others, was shown by PRUD'HOMME et al. 1995 to be controlled by a host gene called flamenco (flam), located at 20A1-3. Our findings, together with those for gypsy, identify the 20A locus as a chromosomal point of regulation for the mobilization of several retroelements. We propose to name this chromosomal point controlling mobilization of several retroelements COM, for centre organisateur de mobilisation.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly stocks and crosses:
The S lines w^IR6 and w¹¹¹⁸ and the U line Rev are from the collection of the Institut National de la Sante et de la Recherche Medicale U384. The ZAM-LacZ and Idefix-LacZ transgenic lines were obtained by injection of ZAM-LacZ and Idefix-LacZ transformation vectors into w¹¹¹⁸ flies. All stocks were maintained at 20°. Expression of transgenes in a genetic context allowing ZAM and Idefix mobilization was analyzed in the progeny of crosses of ZAM-LacZ and Idefix-LacZ transgenic lines with the U line Rev (DESSET et al. 1999 ). Flies used for analysis of expression were raised and kept at 25°.

Stock carrying phenotypic markers on the stable X chromosome, yellow2 (y² w^a ct⁶ lz^bg v¹ f¹/FM7c), was provided by Pedro Santamaria. Stocks with chromosome balancers Bal [Muller5/Muller5; CyO; TM3/C(2;3) Ap^-] and FM6/FM7 were kindly provided by Stephane Ronsseray and Dario Coen, respectively. These lines are stable for ZAM and Idefix mobilization. Stocks carrying deficiencies of the proximal part of the X chromosome [Df(1) in the text] were provided by the Drosophila stock center at Bloomington.

Transgenic constructs:
The LTR region of ZAM was amplified by Taq polymerase on genomic DNA using the pair of primers, whs and zaas (see below). The ZAM-LTR fragment extending from nucleotide 1 to 499 was inserted into the EcoRI site of the P7 vector (gift from A. Pélisson) carrying the Escherichia coli lacZ gene. The 499-bp fragment contains the ZAM promoter and regulatory sequences present within the LTR. This construct was called ZAM-LacZ.

Sequences of the primers referred to are as follows: whs (5'-GGA ATT CCC AAC GGA TGT TTT GAT ACG-3') and zaas (5'-TCT GAA TTC CAT TCC AGT TTT CCG GCT-3'). Each contains an additional EcoRI restriction site at its 5' end. The construct was injected into S-line w¹¹¹⁸ flies, and four transgenic lines were established. Transgenic lines bearing Idefix-LTR fused to LacZ have been previously reported in TCHERESSIZ et al. 2002 . Expression of transgenes was assayed by a histochemical method for ß-galactosidase.

Histochemical staining for ß-galactosidase:
Ovaries were dissected in 1x phosphate-buffered saline (PBS), fixed in 0.5% glutaraldehyde in PBS for 5–10 min at 4°, and rinsed twice in 1x PBS and once in Fe/NaP buffer [0.003 M Na₂HPO₄, 0.072 M NaH₂PO₄, 0.003 M K₃Fe(CN)₆, 0.003 M K₄Fe(CN)₆, 0.15 M NaCl, 0.001 M MgCl₂]. Staining was performed in Fe/NaP buffer with X-Gal (0.2 mg/ml) final concentration at 37°. All samples were stained simultaneously and for the same length of time (2 hr). Stained tissues were washed four times in 1x PBS and mounted in 1:1 PBS/glycerol and examined under an Axiophot microscope (Zeiss) using Nomarski optics.

In situ hybridization on polytene chromosomes:
Larval salivary glands were dissected in saline solution (0.7% NaCl) and squashed in 45% acetic acid. Idefix DNA probes (BH clone described in DESSET et al. 1999 ) were labeled by nick translation with DIG-UTP (Boehringer Mannheim, Indianapolis) and detected by rhodamine-conjugated antibody (Boehringer Mannheim). Fluorescent in situ hybridization (FISH) and CCD camera analyses were carried out as described in detail elsewhere (GATTI et al. 1994 ). Preparations of salivary gland chromosomes were stained with 4',6-diamidino-2-phenylindole. Images were merged and analyzed using the Adobe Photoshop 2.5 program.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The LTR of ZAM responds to the differential regulation exerted by stable and unstable lines:
To investigate the mechanism driving ZAM expression, we constructed a tool that permits an assessment of its restricted activity in unstable lines. A PCR-amplified ZAM fragment extending from nucleotides 1 to 499 was fused to the E. coli lacZ gene in a P-element transformation vector. This construct (ZAM-LacZ) contains the full-length LTR of ZAM and thus allows the test of the potential response of its promoter to the genetic control exerted by S and U lines. This construct was injected into an S line, w¹¹¹⁸, in which ZAM is not expressed and not mobilized (data not shown). Four independent transgenic lines were established.

We designed crosses to test the influence of the genetic background on ZAM expression: (i) Transgenic flies homozygous for ZAM-LacZ were crossed to S flies from the w^IR6 line and (ii) transgenic flies were crossed to U flies from the Rev line. ZAM-driven expression of LacZ was studied in the F₁ progeny of these two series of crosses. F₁ progeny displayed two main characteristics. First, half of its genome derived from the transgenic S line and half was transmitted by either the S w^IR6 line or the U Rev line used for crossing. F₁ individuals were thus either homozygous S/S or heterozygous S/U for the genetic determinant restricting ZAM mobilization. Second, whichever cross is considered, F₁ progeny were heterozygous for the transgene ZAM-LacZ, which was transmitted only by the transformed line (see genotype on Fig 1).

View larger version (48K): In this window In a new window Download PPT slide   Figure 1. Mating schemes used to test the expression pattern of the E. coli LacZ gene fused to ZAM LTR as a function of genetic determinants U and S. (A) Mating scheme and ovary picture of tested females from a cross of w1118 transgenic females (noted as S*) with w1118 nontransgenic males (noted as S). The S/S* ovaries display no LacZ staining. (B) Mating scheme and a typical ovariole picture of females from the cross of w1118 transgenic females (noted S*) with Rev nontransgenic males (noted U). The U/S* ovaries display a typical lacZ pattern in the follicle cells, which is restricted to stage 10 of adult oogenesis. In both series of tests, stained ovarioles are from females heterozygous for the transgene ZAM-LacZ.

From previous experiments, we knew that ZAM transcripts are detected in the ovaries of the Rev unstable lines and specifically occur in a small set of follicle cells located at the posterior pole of each Rev follicle. We therefore followed the expression of the transgene present in S/S and S/U flies in this specific tissue.

When S/S flies were stained for ß-galactosidase expression, no enzyme activity was detected in the F₁ female follicles (Fig 1). All the assays were performed on the four independent transformants obtained from the injections. The effects were the same for all of them, so only results from one transgene located on chromosome 2 are shown here and hereafter. In the other series of dissected females with the S/U genotype, a typical staining of ß-galactosidase activity was detected in the ovaries. This pattern is identical to the pattern of transcription of ZAM in the Rev lines: It is detected in the posterior follicle cells of the follicles (LEBLANC et al. 2000 ). However, it is restricted to later stages of ovary development, starting from stage 9 of oogenesis.

These results show that the LTR of ZAM drives ZAM transcription in the specific lineage of somatic cells located at the posterior part of the oocyte and responds to the control exerted by the genotype of the flies. Furthermore, ZAM transcripts are detected within U/S heterozygotes, indicating that the genetic determinant controlling ZAM expression in U lines is dominant or semidominant over the genetic determinant of S lines.

ZAM is controlled by a genetic determinant located on the X chromosome:
To map the genetic determinant involved in ZAM mobilization, we first sought the chromosome involved in this control. Through a series of crosses, we successively changed chromosomes of the Rev line by their homologs from the stable w^IR6 line. With the help of balancer chromosomes, we established a line bearing an X chromosome from Rev (X^U) and chromosomes 2 (II^S) and 3 (III^S) from w^IR6. The genotype of this line will be referred to here as [X^U; II^S; III^S]. The reciprocal line [X^S; II^U; III^U] was also established with an X chromosome from w^IR6 and autosomes 2 and 3 from Rev (Fig 2).

View larger version (25K): In this window In a new window Download PPT slide   Figure 2. Identification of the chromosome bearing the determinant controlling ZAM expression. The sexual chromosome, chromosome 2, and chromosome 3 are noted X, II, and III, respectively. An exponent indicates which line the chromosome is from. For example, XS is an X chromosome from an S line; XU is an X chromosome from a U line. Crosses performed to establish lines [XU; IIS; IIIS] and [XS; IIU; IIIU] are presented. ZAM-LacZ transgene (not indicated on the scheme) is present as a single copy in the genome of the dissected females.

To test which chromosome was necessary for ZAM activity, females from these two series of lines were then crossed with males from the stable transgenic w¹¹¹⁸ line bearing the ZAM-LacZ construct. We found that the observed ß-galactosidase expression was dependent on the presence of the X chromosome supplied by the U line. Only flies from crossing between the [X^U; II^S; III^S] line and the transgenic ZAM-LacZ line displayed LacZ staining in the follicle cells located at the posterior part of the follicle. Flies from crosses involving the [X^S; II^U; III^U] line displayed no staining (Fig 2).These results indicate that the genetic determinant responsible for ZAM regulation is located on the X chromosome and that chromosomes 2 and 3 present in Rev do not play a role.

A series of crosses was assayed to characterize genetic properties of this determinant. Three types of lines differing by their genotype were established. First, one line, called [U-ZAM-LacZ], with the genotype [X^U/X^U; ZAM-LacZ/ZAM-LacZ; III^S/III^S] displayed two copies of X^U and two copies of the transgene. Second, individuals with the genotype [X^U/X^U; ZAM-LacZ/II^S; III^S/III^S] displayed two copies of the unstable X chromosome and only one copy of the transgene. Third, flies with the genotype [X^U/X^S; ZAM-LacZ/ZAM-LacZ; III^S/III^S] displayed one copy of the unstable X chromosome and two copies of the ZAM-LacZ transgene.

LacZ staining in these three categories of flies revealed several features of the ZAM regulation process. lacZ expression is dependent on the X^U copy number, LacZ staining being much stronger in [X^U/X^U] than in [X^U/X^S] flies (Fig 3). lacZ expression was found throughout the ovarioles from the very first stages of oogenesis to the later stages in [X^U/X^U] flies, whereas it was detected only at later stages 9–10 in [X^U/X^S] flies (Fig 3B and Fig C). By contrast, the degree of LacZ expression in ovaries was independent of the ZAM-LacZ copy number. Flies [X^U/X^U; ZAM-LacZ/ZAM-LacZ] or [X^U/X^S; ZAM-LacZ/ZAM-LacZ] displayed the same intensity of LacZ staining as flies [X^U/X^U; ZAM-LacZ/+] or [X^U/X^S; ZAM-LacZ/+], respectively. These results indicate that the genetic determinant allowing ZAM expression in the unstable lines acts in a dose-dependent manner. In contrast, the LacZ staining is independent of the heterozygous or homozygous status of ZAM-LacZ.

View larger version (87K): In this window In a new window Download PPT slide   Figure 3. Expression pattern of the E. coli LacZ gene fused to ZAM LTR as a function of X chromosomes from a U line. (A) LacZ staining in ovaries of [XS/XS; IIS*/IIS*; IIIS/IIIS] flies; no staining is detected. (B) LacZ staining in the ovaries of [XU/XS; IIS*/IIS*Z; IIIS/IIIS] flies. A typical lacZ pattern is observed in the follicle cells and restricted to stage 10 of adult oogenesis. (C) LacZ staining in the ovaries of [XU/XU; IIS*/IIS*; IIIS/IIIS] flies. LacZ staining is observed in the follicle cells all along the ovariole. Transgene ZAM-LacZ is homozygous in all the lines tested. It is located on chromosome 2 and indicated by an asterisk.

ZAM is regulated by a genetic determinant located at position 20A on the X chromosome:
To map the genetic determinant controlling ZAM more precisely, recombined X chromosomes were generated through meiotic recombination assays performed in a set of crosses between two stocks of flies: U flies (Rev) with an X^U chromosome and S flies with an X^S chromosome bearing the phenotypic markers [y w ct lz v f] dispersed along the whole length of the chromosome (see MATERIALS AND METHODS). These markers help to identify which parts of the chimeric chromosomes obtained from the meiotic recombination have been brought by X^U, which does not display the markers, or by X^S, which does. Some 53 lines with a recombined X chromosome were established and tested for their control of ZAM-LacZ expression (Fig 4).

View larger version (23K): In this window In a new window Download PPT slide   Figure 4. Localization of the genetic determinants present on the X chromosome and involved in ZAM regulation. (A) Details of crosses performed to establish lines with chimeric chromosomes displaying regions from a U and an S genetic context. (B) Localization of the genetic determinant deduced from LacZ assay experiments carried out with lines bearing the chimeric chromosomes obtained in A. The X chromosome is represented with the phenotypical markers and their respective positions. The centromere is represented by a disc.

We found that all the 29 lines that retained an X^U portion of the chromosome located between the f-marker and the centromere were still able to activate ZAM-LacZ expression in the follicles. Conversely, when this portion of the X chromosome was transmitted by the X^S multimarked chromosome, 20 out of 24 lines tested were found to be unable to activate ZAM-LacZ. This finding indicates that ZAM-LacZ expression is clearly associated with the presence of a fragment located on the X^U chromosome between marker f at the cytogenetic position 15A and the centromere located at 20F. No such correlation was ever found for any of the other phenotypic markers y, w, ct, lz, or v, the presence or absence of which within the recombinant line indifferently allows or restricts ZAM-LacZ expression. Even so, 4 lines out of the 24 that display an f-marker transmitted from an X^S chromosome were nevertheless able to activate ZAM-LacZ. This certainly indicates that some recombination events occurred within the 5 cM separating the f-marker and the centromere. These lines therefore may contain portions of the X^U chromosome in the region extending proximally from the f-marker that have not been detected in our screen.

As reported above, LacZ staining depends on the dose of the X^U chromosome. It is strong in homozygous U/U flies, lower in heterozygous U/S flies, and absent in S/S flies. Taking advantage of this property, we performed a further series of crosses to define more precisely the position of the determinant between locus 15A and 20F. We reasoned as follows: If a deficiency Df(1) covers the locus involved in ZAM regulation, then Df(1)/X^U flies might display a LacZ staining similar to homozygous X^U/X^U flies. In contrast, if the genetic determinant is outside the deficiency, then Df(1)/X^U individuals should behave as heterozygous flies and give a weaker LacZ staining, similar to X^U/X^S flies.

This was indeed the case, since two categories of flies differing from the lacZ staining profile were observed in these crosses. Either LacZ was observed in each follicle from early to late stages of oogenesis, as expected in a line with an X^U/X^U genotype, or LacZ staining was observed only from stage 9, as expected with an X^U/X^S genotype. Deficiencies used in this experiment and the corresponding LacZ profile are presented in Fig 5. Maximum LacZ staining was observed when deficiencies affected the X chromosome locus present at cytological position 20A2-3. These results indicate that the genetic determinant responsible for ZAM regulation is present at position 20A2-3 on the X chromosome.

View larger version (19K): In this window In a new window Download PPT slide   Figure 5. X chromosomal deficiencies used for cytogenetic mapping of genetic determinant controlling ZAM. The full chromosomal region is presented at the top. The line below indicates the deficiencies tested. Dissected females display an XU/Df(1) genotype and one copy of the ZAM-LacZ or Idefix-LacZ transgene as indicated in columns on the right. A plus indicates LacZ staining as intense as in Fig 3C. A minus sign represents LacZ staining as weak as in Fig 3B. Data from PRUD'HOMME et al. 1995 concerning gypsy control are reported in the third column. Deficiencies are indicated as plus for flam permissive and as minus for flam restrictive according to the ovoD1 inactivation assay previously described by PRUD'HOMME et al. 1995 . ND, no data.

Region 20A2-3 is also responsible for Idefix regulation:
We previously reported that another retroelement, Idefix, is mobilized in the unstable Rev lines. We thus wondered whether locus 20A2-3 was also responsible for Idefix regulation. Initial evidence that Idefix was regulated by a genetic determinant present on the X chromosome came from in situ experiments performed on polytene chromosomes. FISH experiments performed on the [X^U; II^S; III^S] line revealed that numerous copies of Idefix were distributed on all the chromosome arms of this line. Since no copy was present on chromosomes 2 and 3 when the [X^U; II^S; III^S] line was established, this result indicates that the X^U chromosome is also competent for Idefix mobilization (Fig 6).

View larger version (81K): In this window In a new window Download PPT slide   Figure 6. FISH mapping of Idefix insertions on polytene chromosomes from the salivary glands of line [XU; IIS; IIIS]. Signals are dispersed on all the chromosomes.

To go further into the control of Idefix in different genetic backgrounds, we used transgenic flies bearing the LTR of Idefix fused to LacZ (Idefix-LacZ). These transgenic lines established for an earlier study indicated that LacZ expression from Idefix-LacZ displays the same pattern of expression as endogenous Idefix when observed in a U context. LacZ expression is detected early in the developing follicles of females from the Rev stocks in a structure called the germarium (TCHERESSIZ et al. 2002 ). This blue staining is not observed in the germarium of S flies.

Tests conducted with X chromosome deficiency stocks confirmed that the genetic determinants regulating the expression of both Idefix and ZAM were located in the 20A2-3 region (Fig 5). As previously noted for ZAM, the locus responsible for Idefix regulation is included in a region located between a series of deficiencies, including Df(1)B12, Df(1)mal6, and Df(1)16-3-22, that extend distally, and another, including deficiencies Df(1)R21, Df(1)16-2-13, Df(1)R44, Df(1)GA22, and Df(1)JC12, that extend proximally. Once we had mapped the genetic determinant to this locus, we compared our results to those obtained by PRUD'HOMME et al. 1995 for flamenco, the gene controlling gypsy. Although localized to 20A1-3, flamenco appears to be located outside the region controlling ZAM and Idefix, as clearly shown by crosses involving Df(1)R21, Df(1)16-2-13, and Df(1)DCB1-35c. These deficiencies are located outside the region controlling ZAM and Idefix while, according to PRUD'HOMME et al. 1995 , they include flam determinant able to revert the ovoD1 mutation.

Taken together, these results indicate that locus 20A2-3 is involved in the regulation of the two tested retroelements, ZAM and Idefix. Although close to the flamenco gene, the genetic determinant allowing ZAM and Idefix mobilization is independent from that of gypsy. Additionally, since a deficiency affecting the locus behaves as an activator, the wild-type locus must act as a repressor of ZAM and Idefix mobilization, this repression having been lost in the Rev stocks.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this work, we identify a ß-heterochromatic locus that controls the regulation of multiple transposable elements within the genome of D. melanogaster. Using transgenic lines carrying a LacZ reporter gene for two D. melanogaster retroelements, ZAM and Idefix, we found that the cis elements essential for their proper control are included within their respective LTRs. The LacZ expression patterns indicate that ZAM and Idefix display two main characteristics. First, they are restricted to a very specific subset of somatic ovarian cells. Second, they are expressed in the U lines in which ZAM and Idefix are mobilized and absent in the S lines in which ZAM and Idefix are silenced.

Our analysis of the LacZ expression patterns of ZAM and Idefix enabled us to map and identify the locus responsible for their differential regulation in the U or S backgrounds. We have found that both ZAM and Idefix are controlled by a single locus located at position 20A2-3 in the ß-heterochromatin at the distal tip of the X chromosome. This part of the chromosome is composed of repeated sequences mainly due to the accumulation of defective transposable elements.

In a series of crosses involving recombined chromosomes and deficiencies affecting the X chromosome, we show that this locus acts as a silencer of ZAM and Idefix expression and prevents these elements from being mobilized in S strains. A null allele, such as the deficiency covering the locus, confers a lower repression than that of the allele present in S lines and thus a greater expression of a LacZ reporter gene controlled by ZAM and Idefix promoters. This demonstrates that both elements are downregulated in S lines and have undergone a process of reactivation in U lines. One additional characteristic of this silencing is that the strength of ZAM- or Idefix-driven LacZ staining is dependent on the chromosomal makeup of the 20A2-3 locus. Hence LacZ expression is lowest in flies bearing both X chromosomes from stable lines (S/S), higher in heterozygous (S/U) flies, and highest in U/U or U/def (20A2-3) flies.

Few data concerning the genetic determinants that initiate a burst of transposition within a genome, such as that observed for ZAM and Idefix in the U lines, are available. This is primarily because it is often difficult to trace the history of these genomes and thus to accurately identify the molecular event at the source of such losses of control. This is not an issue in the case of ZAM and Idefix, since both w^IR6, the original parental line in which they are silenced, and its derived Rev lines in which they have been reactivated after a P-element mutagenesis (DESSET et al. 1999 ) are available and were used for comparative studies.

The control of gypsy, another LTR retrotransposon from D. melanogaster, has also been analyzed in detail (PRUD'HOMME et al. 1995 ; ROBERT et al. 2001 ) and appears to display strikingly similar features to those reported here for ZAM and Idefix: (i) The distribution of gypsy in the D. melanogaster genome resembles that of ZAM and Idefix, with all D. melanogaster strains carrying very few elements and a few strains carrying additional copies on chromosome arms (BUCHETON 1995 ); (ii) a high copy number of gypsy is correlated with high rates of mobilization; and (iii) the stability of gypsy mobilization is controlled by the flamenco gene at 20A1-3.

Although the locus involved in gypsy mobilization has been identified, the exact nature of this involvement is still unknown. In a recent study, ROBERT et al. 2001 isolated 100 kb of genomic DNA flanking a P-element-induced mutation of flamenco. Deficiency mapping indicated that the sequences required for normal flamenco function were located 130 kb proximal to the insertion. However, this region falls in a gap in the Celera genome sequence data, where the shotgun sequence could not be assembled because of the presence of long stretches of repetitive DNA. Nevertheless, the distal part of the cloned DNA did contain several unique sequences. ROBERT et al. 2001 have proposed that Dip1, the closest open reading frame to the P-element insertion point, is the gypsy regulator since it putatively encodes a nuclear protein containing two double-stranded RNA (dsRNA)-binding domains. However, transgenes containing dip1 were unable to rescue flamenco mutant flies. Thus, the identity of the missing flamenco sequence remains unclear.

At present, the identity of the determinant governing the control of ZAM and Idefix is also unknown. Genetic analysis of the U line indicates that several lines bearing deficiencies, such as Df(1)16-2-13 or Df(1)DCB1-35c, and affecting dip1 function do not activate ZAM and Idefix expression, but are permissive for gypsy mobilization (Fig 5). Thus, dip1 is unlikely to be involved in ZAM and Idefix regulation. Our study also suggests that gypsy regulation is genetically separable from ZAM and Idefix regulation. Indeed, ZAM-LacZ and Idefix-LacZ transgenes are activated in the U lines, whereas a gypsy-LacZ transgene is not (data not shown). In contrast, neither ZAM-LacZ nor Idefix-LacZ staining is observed in w^IR6, a line permissive for gypsy-LacZ expression.

What, then, is the mechanism involved in ZAM and Idefix regulation? Our results point to the heterochromatic region 20A2-3, where numerous deleted copies of transposable elements are accumulated, as the key region involved in the regulation of these retroelements. This locus is responsible for their repression, a deficiency behaving as an activator of their expression. Previous evidence has indicated that retrotransposons can be silenced by RNA-mediated cosuppression, a process in which mRNA degradation is triggered by homologous RNA (ARAVIN et al. 2001 ). Additionally, in Caenorhabditis elegans, mutants defective in dsRNA interference display activation of their transposons (TABARA et al. 1999 ; KETTING and PLASTERK 2000 ) and in Drosophila, the I element is silenced by cosuppression, a process that may have evolved to suppress transposable elements (CHABOISSIER et al. 1998 ; JENSEN et al. 1999 ).

Since the 20A2-3 locus is composed mainly of vestiges of transposable elements, aberrant transcripts initiated from these vestiges and homologous to ZAM, Idefix, and possibly other elements can potentially be synthesized and then become a basis for their repression by a mechanism involving cosuppression. This hypothesis would explain why ZAM and Idefix controls seem to be independent from that of gypsy, although they are initiated from a genetic determinant located at the same locus. While we failed to detect any ZAM or Idefix sequences at this locus by in situ hybridization on polytene chromosomes, it is possible that fragments of these elements were present but undetected by our experiments and absent from the preliminary sequence of this region available in the data banks. However, it is also possible that other retrotransposons present in the 20A region may display domains that are highly conserved between retroelements (such as the reverse transcriptase domain), which can potentially recognize and regulate ZAM, Idefix, and/or numerous other retroelements. Recently, VOLPE et al. 2002 proposed a model in wild-type fission yeast, whereby dsRNAs of centromeric heterochromatin repeats produce siRNAs, which in turn trigger gene silencing and repress their own transcription. They suggest that these dsRNA could potentially silence other loci with homologous DNA sequences. If such a mechanism is indeed involved in the regulation of ZAM and Idefix, it should also account for the dose-dependent response we observe. Although dsRNA is a robust inhibitor of gene activity, it has been shown that varying levels of dsRNA cause a dose-dependent response of RNA-mediated interference (KENNERDELL and CARTHEW 1998 ; SCHWARZ et al. 2002 ). Furthermore, the effectiveness of RNA interference may also be changed by the quality of the interfering RNA duplex, depending on the presence or absence of mismatches between the two strands (PARRISH et al. 2000 ). If the targeting signal is not absolutely identical to the target sequences of ZAM and Idefix, then heteroduplexes with mismatches between the two interfering strands could trigger an RNAi response, but with reduced effectiveness. The strength of the response may then depend on the dosage of the targeting signal.

In summary, although RNA interference is an attractive hypothesis, it remains to be demonstrated that this cellular defense system is indeed the mechanism accounting for ZAM and Idefix silencing. Regardless of the mechanism involved, however, our results clearly indicate that a control system for retroelements is present in this particular region of the Drosophila genome and prevents high rates of transposition. If altered, this system could easily trigger the reactivation of dormant retrotransposons and lead to sudden bursts of mobilization that affect an individual and thereby, potentially, a species. Since our data indicate that diverse retroelements are controlled by determinants located in a single locus at 20A2-3, we propose calling it COM and note its specificity for a family as a superscript. Hence, COM^U represents the locus that permits ZAM and Idefix mobilization, while COM^flam permits gypsy mobilization.

	ACKNOWLEDGMENTS

We thank Pradeep Das and members of the group for critical review of the manuscript and P. Dimitri for in situ hybridizations on polytene chromosomes. This work was supported by grants from INSERM (U384) and CNRS (GDR 2157) and by a project grant from Association pour la Recherche contre le Cancer and MAE (Galilée project) to C.V. C.M. received a grant from the Ministère de l'Enseignement Supérieur et de la Recherche and the Fondation pour la Recherche Médicale.

Manuscript received October 3, 2002; Accepted for publication February 13, 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.[Medline]

BUCHETON, A., 1995  The relationship between the flamenco gene and gypsy in Drosophila: how to tame a retrovirus. Trends Genet. 11:349-353.[Medline]

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]

DESSET, S., C. CONTE, P. DIMITRI, V. CALCO, and B. DASTUGUE et al., 1999  Mobilization of two retroelements, ZAM and Idefix, in a novel unstable line of Drosophila melanogaster. Mol. Biol. Evol. 16:54-66.[Abstract]

GATTI, M., S. BONACCORSI, and S. PIMPINELLI, 1994  Looking at Drosophila mitotic chromosomes. Methods Cell Biol. 44:371-391.[Medline]

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

KENNERDELL, J. R. and R. W. CARTHEW, 1998  Use of dsRNA-mediated genetic interference to demonstrate that frizzled and frizzled 2 act in the wingless pathway. Cell 95:1017-1026.[Medline]

KETTING, R. F. and R. H. PLASTERK, 2000  A genetic link between co-suppression and RNA interference in C. elegans. Nature 404:296-298.[Medline]

LEBLANC, P., S. DESSET, F. GIORGI, A. R. TADDEI, and A. M. FAUSTO et al., 2000  Life cycle of an endogenous retrovirus, ZAM, in Drosophila melanogaster. J. Virol. 74:10658-10669.[Abstract/Free Full Text]

LIM, J. K., M. J. SIMMONS, J. D. RAYMOND, N. M. COX, and R. F. DOLL et al., 1983  Homologue destabilization by a putative transposable element in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 80:6624-6627.[Abstract/Free Full Text]

PARRISH, S., J. FLEENOR, S. XU, C. MELLO, and A. FIRE, 2000  Functional anatomy of a dsRNA trigger: differential requirement for the two trigger strands in RNA interference. Mol. Cell 6:1077-1087.[Medline]

PASYUKOVA, E. G. and S. V. NUZHDIN, 1993  Doc and copia instability in an isogenic Drosophila melanogaster stock. Mol. Gen. Genet. 240:302-306.[Medline]

PETROV, D. A., J. L. SCHUTZMAN, D. L. HARTL, and E. R. LOZOVSKAYA, 1995  Diverse transposable elements are mobilized in hybrid dysgenesis in Drosophila virilis. Proc. Natl. Acad. Sci. USA 92:8050-8054.[Abstract/Free Full Text]

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]

SCHWARZ, D. S., G. HUTVAGNER, B. HALEY, and P. D. ZAMORE, 2002  Evidence that siRNAs function as guides, not primers, in the Drosophila and human RNAi pathways. Mol. Cell 10:537-548.[Medline]

SMIT, A. F., 1999  Interspersed repeats and other mementos of transposable elements in mammalian genomes. Curr. Opin. Genet. Dev. 9:657-663.[Medline]

TABARA, H., M. SARKISSIAN, W. G. KELLY, J. FLEENOR, and A. GRISHOK et al., 1999  The rde-1 gene, RNA interference, and transposon silencing in C. elegans. Cell 99:123-132.[Medline]

TCHERESSIZ, S., V. CALCO, F. ARNAUD, L. ARTHAUD, and B. DASTUGUE et al., 2002  Expression of the Idefix retrotransposon in early follicle cells in the germarium of Drosophila melanogaster is determined by its LTR sequences and a specific genomic context. Mol. Genet. Genomics 267:133-141.[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]

WHITELAW, E. and D. I. MARTIN, 2001  Retrotransposons as epigenetic mediators of phenotypic variation in mammals. Nat. Genet. 27:361-365.[Medline]

This article has been cited by other articles:

				A. A. Aravin, G. J. Hannon, and J. Brennecke The Piwi-piRNA Pathway Provides an Adaptive Defense in the Transposon Arms Race Science, November 2, 2007; 318(5851): 761 - 764. [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, 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]

				P. Lopez-Sanchez, J. C. Costas, and H. F. Naveira Paleogenomic Record of the Extinction of Human Endogenous Retrovirus ERV9 J. Virol., June 1, 2005; 79(11): 6997 - 7004. [Abstract] [Full Text] [PDF]

				X. Zhang and T. H. Eickbush Characterization of Active R2 Retrotransposition in the rDNA Locus of Drosophila simulans Genetics, May 1, 2005; 170(1): 195 - 205. [Abstract] [Full Text] [PDF]

				A. I. Kalmykova, M. S. Klenov, and V. A. Gvozdev Argonaute protein PIWI controls mobilization of retrotransposons in the Drosophila male germline Nucleic Acids Res., April 7, 2005; 33(6): 2052 - 2059. [Abstract] [Full Text] [PDF]

				C. Meignin, B. Dastugue, and C. Vaury Intercellular communication between germ line and somatic line is utilized to control the transcription of ZAM, an endogenous retrovirus from Drosophila melanogaster Nucleic Acids Res., July 19, 2004; 32(13): 3799 - 3806. [Abstract] [Full Text] [PDF]

				E. Sarot, G. Payen-Groschene, A. Bucheton, and A. Pelisson Evidence for a piwi-Dependent RNA Silencing of the gypsy Endogenous Retrovirus by the Drosophila melanogaster flamenco Gene Genetics, March 1, 2004; 166(3): 1313 - 1321. [Abstract] [Full Text] [PDF]

				C. Meignin, J.-L. Bailly, F. Arnaud, B. Dastugue, and C. Vaury The 5' Untranslated Region and Gag product of Idefix, a Long Terminal Repeat-Retrotransposon from Drosophila melanogaster, Act Together To Initiate a Switch between Translated and Untranslated States of the Genomic mRNA Mol. Cell. Biol., November 15, 2003; 23(22): 8246 - 8254. [Abstract] [Full Text] [PDF]