Artificial and Epigenetic Regulation of the I Factor, a Nonviral Retrotransposon of Drosophila melanogaster

Corresponding author: Hubert Pinon, Centre de Génétique Moléculaire et Cellulaire, UMR 5534 bât. 741, Université Claude Bernard, 43, Blvd. du 11 Novembre 1918, F-69622 Villeurbanne Cedex, France., hpinon{at}biomserv.univ-lyon1.fr (E-mail)

Communicating editor: M. J. SIMMONS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The I factor (IF) is a LINE-like transposable element from Drosophila melanogaster. IF is silenced in most strains, but under special circumstances its transposition can be induced and correlates with the appearance of a syndrome of female sterility called hybrid dysgenesis. To elucidate the relationship between IF expression and female sterility, different transgenic antisense and/or sense RNAs homologous to the IF ORF1 have been expressed. Increasing the transgene copy number decreases both the expression of an IF-lacZ fusion and the intensity of the female sterile phenotype, demonstrating that IF expression is correlated with sterility. Some transgenes, however, exert their repressive abilities not only through a copy number-dependent zygotic effect, but also through additional maternal and paternal effects that may be induced at the DNA and/or RNA level. Properties of the maternal effect have been detailed: (1) it represses hybrid dysgenesis more efficiently than does the paternal effect; (2) its efficacy increases with both the transgene copy number and the aging of sterile females; (3) it accumulates slowly over generations after the transgene has been established; and (4) it is maintained for at least two generations after transgene removal. Conversely, the paternal effect increases only with female aging. The last two properties of the maternal effect and the genuine existence of a paternal effect argue for the occurrence, in the IF regulation pathway, of a cellular memory transmitted through mitosis, as well as through male and female meiosis, and akin to epigenetic phenomena.

TRANSPOSABLE elements are mobile DNA sequences found in all organisms from bacteria to higher eukaryotes. In eukaryotes, transposable elements represent a substantial fraction of the genome: at least 15% of the Drosophila genome, 35% of the human genome, and more than 50% of some plant genomes (FINNEGAN 1992 ; SAN MIGUEL et al. 1996 ; VOYTAS and NAYLOR 1998 ). Their mobility is associated with deleterious effects such as insertional mutagenesis and chromosomal rearrangements (LIM and SIMMONS 1994 ), but it is also believed to contribute positively to the evolution of the host genome (reviewed in KIDWELL and LISCH 1997 ).

Transpositions are usually rare events and therefore are difficult to study, except during the hybrid dysgenesis process (BREGLIANO and KIDWELL 1983 ). Hybrid dysgenesis is observed in Drosophila melanogaster when flies containing silenced but potentially active transposable elements such as P, the I factor (IF), or hobo are crossed with a strain devoid of autonomous elements (RUBIN et al. 1982 ; FAWCETT et al. 1986 ; BLACKMANN et al. 1987 ). In the germline of the progeny the elements are mobilized at high frequency, and sterility is observed. For that reason hybrid dysgenesis has been proposed as one of the mechanisms that contribute to the isolation of subpopulations and that are involved in speciation (BREGLIANO and KIDWELL 1983 ; KIDWELL and LISCH 1997 ; O'NEILL et al. 1998 ). In the case of the IF, a high level of transposition occurs only when IF-containing males (from inducer, or I strains) are crossed with IF-lacking females (from reactive, or R strains). IF, which is silenced in I strains, is subsequently mobilized at high frequency in the germlines of the female progeny, which exhibit the specific traits of I-R hybrid dysgenesis, the most conspicious one being a more or less severe sterile female (SF) phenotype. These so-called SF females lay a normal number of eggs, but the eggs do not hatch because of early embryonic death (LAVIGE 1986 ). The reciprocal cross (I female x R male) gives rise to RSF females, which show normal fertility and a fivefold lower transposition rate than SF females (PICARD 1978 ).

In ovaries from SF females IF silencing is disrupted and IF comes under the positive control of a quantitative trait, called reactivity, which is provided by the R strain and shows a complex mixture of chromosomal and maternal inheritance (reviewed in BREGLIANO and KIDWELL 1983 ; FINNEGAN 1989 ); reactivity is for the most part maternally inherited from one generation to the next. However, the genetic determinants appear to be chromosomal (PICARD 1978 ; BUCHETON and PICARD 1978 ). The reactivity level of an R female is measured by the percentage of sterility of its SF daughters. Accordingly, it is possible to identify strong and weak reactive strains (BUCHETON et al. 1976 ). Reactivity level can be reversibly modified over several generations by various experimental conditions: female aging, breeding temperature (BUCHETON 1978 , BUCHETON 1979A , BUCHETON 1979B ; reviewed in BREGLIANO and KIDWELL 1983 ; FINNEGAN 1989 ), inhibitors of DNA synthesis, and gamma rays (BREGLIANO et al. 1995 ). In addition, the reactivity level controls the transcriptional activity of the IF (LACHAUME and PINON 1993 ; MCLEAN et al. 1993 ; DE LA ROCHE SAINT ANDRE and BREGLIANO 1998 ), the cellular localization of the IF translation products (SELEME et al. 1999 ), and the percentage of crossing over in the maternal germline (LAURENCON et al. 1997 ).

Reactivity is also believed to exist in I strains although it induces neither sterility nor high rates of IF transposition, which suggests that in I females IF autoregulation interferes with reactivity (LAURENCON et al. 1997 ). Reactivity has been likened to an error-prone recombination-repair process called the variability modulation system (VAMOS), thought to resemble the SOS system in bacteria (BREGLIANO et al. 1995 ; LAURENCON et al. 1997 ). Indeed, the SOS system shares two interesting similarities with reactivity: it is able to regulate the expression of the transposable element Tn5 through the LexA protein (ECHOLS and GOODMAN 1990 ), and it leads to insertional mutagenesis when induced (KUAN and TESSMAN 1991 ). However, whether IF transposition is directly responsible for SF sterility or whether it leads to the overinduction of the VAMOS system has yet to be resolved.

In the generations following a dysgenic cross, the IF invades the genome until it reaches 10–15 copies per haploid genome and is progressively repressed through an autoregulation process (PELISSON and BREGLIANO 1987 ). So far, IF autoregulation has been explained by four nonexclusive mechanisms: (1) The protein encoded by ORF1 of IF contains zinc-finger domains and may be an IF self-repressor. (2) Alternatively, the so-called "titration" model proposed that ovarian-specific activators able to bind and activate the IF promoter are present in a limited amount (UDOMKIT et al. 1996 ). As IF copy number increases, the local concentration of this factor on each copy of the IF decreases. This would decrease the overall level of transcription of the IF under a threshold where transposition does not occur. From this hypothesis, CHABOISSIER et al. 1998 demonstrated that, indeed, the IF promoter induced an increasing repressive effect as its copy number increased. However, deletion of the binding site of the ovarian activator previously described by UDOMKIT et al. 1996 did not abolish this repression. (3) These data were then reinterpreted (CHABOISSIER et al. 1998 ) as indicating a phenomenon analogous to a transcriptional silencing, perhaps due to a change in IF chromatin structure as described in Drosophila (PAL-BHADRA et al. 1997 , PAL-BHADRA et al. 1999 ). (4) Recently, JENSEN et al. 1999A , JENSEN et al. 1999B showed that IF autoregulation could be assigned not to the ORF1 polypeptide but to the production of IF RNAs. These RNAs appeared responsible for a phenomenon similar to the post-transcriptional silencing previously described in plants and fungi (MATZKE and MATZKE 1995 ; COGONI et al. 1996 ; GRANT 1999 ; SELKER 1999 ) and known as cosuppression. Thus, CHABOISSIER et al. 1998 and JENSEN et al. 1999B agree that IF regulation is controlled through epigenetic phenomena showing the characteristics of homology-dependent gene silencing.

IF belongs to the LINE (long interspersed nucleotidic element) family of transposable elements and transposes through reverse transcription of an RNA intermediate (JENSEN and HEIDMANN 1991 ; PELISSON et al. 1991 ). Consequently, reducing the level of this RNA is expected to reduce IF transposition. To test this idea, the RNA intermediate of IF has been targeted by artificial antisense RNAs (asRNAs) previously evaluated in transient expression assays (TATOUT et al. 1998 ). Here the use of these transgenic asRNAs demonstrates that downregulation of IF abolishes the sterile female phenotype. Moreover, we show that some transgenes confer a repressive effect with zygotic but also with particular maternal or paternal components. These inherited effects may be induced at the DNA and/or RNA (sense or antisense) level, provided that these sequences contain specific features of the IF. These results support the idea that IF expression is controlled through epigenetic mechanisms, provide a genetic description of this epigenetic phenomenon, and show that the mechanisms involved are reminiscent of reactivity itself.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

D. melanogaster strains:
Because breeding conditions strongly influence the reactivity level and, consequently, the intensity of the hybrid dysgenesis syndrome, experimental cultures must be grown under carefully controlled procedures. First, strains were reared at 20 ± 1° under uncrowded conditions on the axenic medium of DAVID 1959 . Second, all stocks were maintained with short generations; i.e., each generation was obtained from very young (2-day-old) parents. Standard strains used in this study are the R strain w^K (LUNING 1981 ) and the I strain w¹¹¹⁸ (HAZELRIGG et al. 1984 ). The R strains LH23 and Valence, which carry balancer chromosomes, are from the stocks of the former Laboratoire de Génétique, Université de Clermont-Ferrand; their I derivatives were obtained by chromosome contamination with I factor sequences after crossing with the standard w¹¹¹⁸ strain. Additional information about the genetic markers can be found in LINDSLEY and ZIMM 1992 .

DNA constructs:
Fragments of the IF ORF1 were derived from the pSPI1 clone (CHABOISSIER et al. 1990 ). The IF-lacZ fusion is described in LACHAUME et al. 1992 and the RZI construct in TATOUT et al. 1998 . RZAP was derived from RZI using primers P1 to P6 (see below) in recombinant PCR. P1 and P2 contain a BamHI site (underlined) used for cloning; P1 allows deletion of the first five nucleotides of IF essential for IF promoter activity (MCLEAN et al. 1993 ). P3 and P4 mutate ATG187 to TAG (boldface type), giving rise to a new XbaI site (underlined). P5 and P6 change ATG586 to CGT (boldface type), giving rise to a new SnaBI site (underlined). P1 and P4 yield a 198-bp PCR fragment; P3 and P6 yield a 429-bp fragment; P5 and P2 yield a 546-bp fragment. The 429- and 546-bp fragments were combined in a subsequent PCR using P3 and P2 to generate a 954-bp fragment. Finally, the 198-bp and verified 954-bp fragments were combined together with P1 and P2 to generate a 1124-bp mutated ORF1 sequence corresponding to RZAP. The BB construct contains the BstBI-BstBI fragment from the pSPI1 clone. DNA sequencing was performed using the T7 sequencing kit (Pharmacia, Piscataway, NJ). After appropriate enzymatic treatments, RZI, RZAP, and BB were cloned in antisense orientation in the BamHI-digested pCaSpeR-act5C vector (THUMMEL et al. 1988 ), which contains the promoter and the polyadenylation signals from the actin5C gene. The 8S construct is the same as RZI, but this truncated ORF1 fragment was cloned in the pW8 transformation vector (KLEMENZ et al. 1987 ) digested by BamHI.

• P1 = (+6)CGGGATCCACTTCAACCTCCGAAGAGATAAGTCG(+33)

• P2 = (+1104)CGGGATCCATTAGGTGATGGAGTGTTTGTTGTCC(+1079)

• P3 = (+171)CCCTTAACCAACAATCTAGACAGACCCACC(+200)

• P4 = (+197)GGGTCTGTCTAGATTGTTGGTTAAGGGC (+170)

• P5 = (+579)AAAAATACGTAAACGGCAAAACCCC(+603)

• P6 = (+600)GTTTTGCCGTTTACGTATTTTTTTAACTTCA G(+569)

P-mediated transformation and establishment of transgenic Drosophila stocks:
Except that plasmid pUChs2-3 was the source of transposase, P-mediated germline transformation was performed as usual (RUBIN and SPRADLING 1982 ) in the w^K strain that is M and R in the P-M and I-R systems, respectively. The transformation vectors contain the mini-white gene, which allows the selection of transformants on the basis of their colored-eyed phenotype. Inserts were mapped to chromosomes by segregation analysis against balancer chromosomes M5, CyO, and Dcxf; results were confirmed by in situ hybridization on salivary gland polytene chromosomes (BIEMONT 1986 ). Transgene integrity and copy number were determined by Southern blots probed with I factor or actin5C sequences. For the RZI construct, 10 independent lines bearing one homozygous copy (1RZI lines) were established. Among them, some insertions on the chromosomes X, II, and III were arbitrarily chosen to construct, by standard genetic procedures, three lines with two homozygous insertions (2RZI lines) and two lines with three homozygous insertions (3RZI lines) on the chromosomes II and III, and X, II, and III, respectively. Independent transgenic lines were also obtained for the other constructs: 17 for RZAP (1RZAP lines), 4 for 8S (18S lines) with all insertions on chromosome III, 16 for BB (1BB lines). For the RZAP construct some insertions were also combined to establish three lines with two homozygous insertions on chromosomes II and III (2RZAP lines). For the BB construct, three lines with two (2BB) and two lines with three (3BB) homozygous insertions distributed among the three main chromosomes were established.

Ten lines carrying one homozygous copy of RZI have been studied in parallel. In this report, the term 1RZI refers to the "average line"—whatever the chromosomal position of its transgene—and, by extension, depicts the average effects obtained from these 10 single-copy-carrying lines. 2RZI and 3RZI have the same meaning for the lines carrying two or three insertions, respectively. In the same way, 1RZAP, 2RZAP, 18S, 1BB, 2BB, and 3BB are used to stand for the respective average line and average effect from all the corresponding transgenic lines. For each construct, all lines were subsequently contaminated by functional I factors (LACHAUME and PINON 1993 ) to obtain the same set of transgenic lines but in the inducer state.

RNA probes and standards:
The sense (antisense) RNA probes used in antisense (sense) Northern blot or slot-blot analysis were synthesized from a HindIII- (EcoRI-) digested pSPI1 vector (CHABOISSIER et al. 1990 ), using [³²P]rUTP and SP6 (T7) RNA polymerase (Boehringer Mannheim, Indianapolis). Sense and antisense RNAs used to calibrate RNA slot-blot analysis were obtained in the same way using unlabeled rUTP. Prior to use, sense and antisense RNAs were treated with Dnase I, extracted with phenol-chloroform, and precipitated twice by ethanol.

Northern and slot-blot analysis:
Ovaries from 5-day-old flies were dissected and stored at -80°. Poly(A)⁺ RNAs were prepared using the Quick prep micro mRNA purification kit (Pharmacia). For the 8S transgenic lines, total RNAs were extracted (DE LA ROCHE SAINT ANDRE and BREGLIANO 1998 ) because the pW8 transformation vector does not contain a polyadenylation sequence. Extracted RNAs were transferred to nylon N membranes (Amersham, Buckinghamshire, UK) either after electrophoresis in formaldehyde-agarose gels or with a slot-blot apparatus. Before electrophoresis 5 µg of poly(A)⁺ RNAs (or 10 µg of total RNAs) were denatured for 5 min at 65° in 1x MOPS, 50% formamide, and 7% formaldehyde. For slot-blot analysis, RNAs were denatured 15 min at 68° in 1x SSC, 50% formamide, and 7% formaldehyde. Then, 0.25, 0.5, 1.0, and 1.5 µg of poly(A)⁺ RNAs (or 0.5, 1.0, 1.5, and 2.0 µg of total RNAs) were loaded on a membrane. In parallel, known quantities of in vitro transcribed antisense or sense RNAs, estimated by UV absorption at OD₂₆₀ nm, were used as standard. Membranes were hybridized as described in SAMBROOK et al. 1989 . Slot-blots were quantified using the PhosphorImager system (Bio-Rad). For each transgenic line quantifications were performed from three independent extractions to obtain mean values with standard errors.

ß-Galactosidase quantification:
Ovarian ß-galactosidase activities were measured using an ELISA kit (TEBU, 5'-3' Inc.). After dissection, 10 ovaries were disrupted in 200 µl of PBS containing 1 mM phenylmethylsulfonyl fluoride, frozen in liquid nitrogen and allowed to thaw thrice, and centrifuged 15 min at 15,000 x g. The supernatant was recovered and checked for ß-galactosidase activity as described by the manufacturer. Quantifications were performed from at least three independent extractions to obtain mean values with standard errors.

Measurement of the sterility level and statistical analysis:
Three sets of five to eight SF females from each dysgenic cross were mated with sib males and allowed to lay eggs for 1 wk at 20 ± 1°. Every 24 hr, eggs were recovered and allowed to develop for 36 hr. Percentages of hatched and unhatched eggs were scored to evaluate SF sterility. For a given transgene copy number, mean values and standard deviations were estimated from at least three independent experiments performed on all the corresponding strains, unless otherwise stated.

For each line and for each dysgenic cross, the SF sterility data were recorded daily during the first 7 days of egg laying and combined to obtain the average sterility level. From these data, zygotic and maternal and paternal effect values were then computed as described in the legend of Fig 3. Finally, data were combined over lines, for each transgene and according to copy number, to give the data presented under the terms 1RZI, 1RZAP, 18S, 1BB, 2RZI, 2RZAP, 2BB, 3RZI, and 3BB (see Table 1 for an example).

View larger version (20K): In this window In a new window Download PPT slide   Figure 1. Structure of the I factor and transgenes. Nucleotide positions are given according to FAWCETT et al. 1986 . IF, general organization of the I factor (IF) with its two nonoverlapping open reading frames (not drawn to scale). IF transcription is driven by a weak internal promoter and is initiated at nucleotide +1 (arrow). Functional and potential initiating codons are indicated. 290, 1014, and 1104 indicate, respectively, the BstBI, BstBI, and SspI restriction sites used for transgene construction. Zn-f indicates the zinc-finger domain also found within the gag domain of retroviruses. IF-lacZ is a translational fusion between IF and the ß-galactosidase gene from Escherichia coli previously described in LACHAUME et al. 1992 ; the Zn-f domain has been deleted (). This construct is present at two homozygous loci in the previously described 4I-lacZ strain (TATOUT et al. 1994 ). RZI contains part of the ORF1 of IF (from +1 to +1104) transcribed in antisense orientation from the act5C promoter and in sense orientation from the IF promoter. RZAP is similar to RZI, but contains three mutations: , a deletion of the first five nucleotides essential for IF promoter activity (MCLEAN et al. 1993 ), and X, missense mutations of both ATG187 and ATG586 to TAG and CGT codons. 8S expresses from the IF promoter a truncated sense RNA (+1 to +1104) that can be potentially translated from ATG187 or ATG586. All four insertions have been obtained on the same chromosome, and no line with multiple insertions has been constructed. BB produces an antisense RNA corresponding to the central part of ORF1 (+290 to +1014). n, number of independent reactive lines established for each transgene. At the right of the figure the phenotypic effects recorded and detailed in the course of this study are summarized.

View larger version (17K): In this window In a new window Download PPT slide   Figure 2. Correlations among asRNA production, IF expression, and SF sterility. First, reactive females carrying two homozygous copies of IF-lacZ (TATOUT et al. 1994 ) were crossed with reactive males carrying zero to three homozygous copies of RZI. Reactive female progeny were kept for 5 days, and then their ovaries were dissected. Ovarian asRNA accumulation was evaluated by slot-blot analysis on poly(A)+ RNA extracts and the amount of bacterial ß-galactosidase was estimated using ELISA tests (see MATERIALS AND METHODS). Second, isogenic (but not transgenic) reactive females were crossed with inducer males carrying zero to three homozygous copies of RZI. SF progeny were kept until 5 days old and checked for their sterility. (A) asRNA accumulation [in nanograms of asRNA/µg of poly(A)+ RNA] as a function of RZI copy number. (B) Bacterial ß-galactosidase amount (in nanograms of ß-galactosidase/milligrams of total ovarian protein, open triangle) and SF sterility (percentage of dead embryos, solid squares) as a function of RZI copy number.

View larger version (15K): In this window In a new window Download PPT slide   Figure 3. Mating schemes used to discriminate the zygotic antisense effect from parental effects. Transgenes were followed by the colored-eye phenotype they confer. Tg, transgene (one, two, or three copies); control, isogenic but nontransgenic strain; +, wild-type nontransgenic chromosome (for clarity a single autosomal chromosome is shown); I, inducer strain; R, reactive strain; SF, dysgenic sterile female. In all crosses SF females were isogenic except for the transgenic insertions. The long-term (A) and short-term (B) maternal effects were investigated by using Tg/+ females as mothers of SF females; transgenes were transmitted through either n maternal generations (A) or one maternal generation (B) before the dysgenic cross. The maternal effect was evaluated by subtracting the percentage sterility of Tg- SF females from the percentage sterility of control SF females in both crosses. (C) Zygotic effect and paternal effect were investigated using Tg/+ males as fathers of SF females. The zygotic asRNA effect was evaluated by subtracting the percentage sterility of Tg+ SF females from the percent-sterility of their Tg- SF sisters. The paternal effect was evaluated by subtracting the percentage sterility of Tg- SF females from the percentage sterility of control SF females.

Statistical analyses were performed on SF sterility values, as well as on the data of zygotic and maternal and paternal effects, by using the Statgraphics Plus 3.1 software (Statistical Graphics Corporation). ANOVA was done to determine if the zygotic and maternal and paternal effect values—which are differences between two sterility levels (see text)—were statistically different from zero (at P < 0.01 or < 0.05).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Establishing transgenic lines:
Various constructs expressing either antisense RNA (asRNAs) targeting the IF ORF1 and/or sense RNAs (sRNAs) corresponding to the same sequence were designed (Fig 1). Because germline expression of IF is restricted to females (CHABOISSIER et al. 1990 ; LACHAUME et al. 1992 ; SELEME et al. 1999 ), asRNAs were driven by the promoter of the actin5C gene, which is strongly expressed during oogenesis (FYRBERG et al. 1983 ). sRNAs were produced from the weak internal IF promoter present in some constructs.

To have the opportunity to use either I or R strains to deliver one, two, or three copies of each transgene to the SF females, the following procedure was followed. First, every independent transgenic insertion was established in the reactive strain w^K as a homozygous line and assigned to a chromosome. Second, identical transgenes inserted on different chromosomes were combined to yield lines homozygous for two or three different insertions (MATERIALS AND METHODS). Among these transgenic lines no variegation of the colored-eye phenotype associated with the mini-white selection marker was observed, suggesting that no transgenes were subject to a repressive effect such as those mediated by the Polycomb group of proteins or the modifiers of position effect variegation and that increasing copy number did not result in heterochromatinization of the transgenic marker. Third, to have the set of lines in I and R states, the R transgenic lines were contaminated by functional IFs as previously described (LACHAUME and PINON 1993 ). Finally, Northern blot analysis performed on ovarian extracts from all transgenic lines confirmed that the transgenes strongly expressed transgenic asRNAs with the expected length. By contrast, sRNAs were faintly or not detected (not shown). The phenotypic effects observed with each construct are depicted in Fig 1, which gives a brief overview of the results discussed below.

SF sterility results from IF mobilization:
The effect of RZI, which produces asRNA from +1 to +1104, on IF expression and the sterile female phenotype was evaluated. The mere presence of this transgene, however, might modify the level of reactivity and its maternal transmission in the transgenic lines (JENSEN et al. 1995 ). To avoid this bias, RZI was transmitted from homozygous males. R males carrying the RZI construct were crossed to R females carrying the IF-lacZ construct (Fig 1) to give R female progeny, the ovaries of which were assessed for transgenic asRNAs and ß-galactosidase without interference from naturally occuring IFs. The SF females, daughters of I RZI males crossed to R w^K females, were checked in parallel for their sterility without interference from the IF-lacZ construct (Fig 2).

Increasing the copy number of RZI increased the amount of asRNAs (Fig 2A). This clearly rules out any cosuppression between the various copies of RZI. Moreover, there was a correlation between the expression of RZI in R ovaries, as defined by the level of asRNAs, and (1) the IF expression level in R ovaries, as monitored by the IF-lacZ reporter, and (2) the degree of sterility of isogenic transgenic SF females without the IF-lacZ target. The increasing amount of asRNAs correlated with a decrease in both the amount of ovarian transgenic ß-galactosidase and the SF lethality level (Fig 2B). This correlation demonstrates a link between the expression level of IF and the appearance of SF sterility, the main phenotypic trait of hybrid dysgenesis.

Detection and assessment of zygotic and parental repressive effects induced by RZI:
Is RZI also able to induce a maternal effect on SF females in addition to the zygotic asRNA effect described above? To address this point, the mating schemes described in Fig 3 were used.

The aim was to obtain simultaneously transgenic (Tg⁺, or colored-eyed) as well as nontransgenic (Tg^-, or white-eyed) SF females derived from the same heterozygous parents. Their respective levels of sterility could be compared with that of an SF control, which had no transgene in its ancestry. The number of maternal generations through which the transgene was transmitted varies according to each protocol: n (Fig 3A), 1 (Fig 3B), or 0 (Fig 3C). If the level of sterility is lower in the Tg^- females than in control females, this rescue should be ascribed to the presence of the transgene in the previous generation(s)—that is, to a maternal effect [either accumulated over several generations (Fig 3A) or in one generation (Fig 3B)] or to a putative paternal effect (Fig 3C). In contrast, a zygotic effect is assessed by comparing the sterility levels of Tg⁺ and Tg^- sibling SF females when there is no maternal effect, that is, when the transgene is transmitted through males to SF females (Fig 3C).

The RZI transgene induces zygotic, maternal, and paternal effects:
Ten generations after transgene establishment in an R or I context, the various effects of one, two, and three copies of RZI were evaluated. The sterility levels of the different SF females were checked daily during the first week of egg laying. Average values obtained each day from the two 3RZI lines are shown in Fig 4.

View larger version (27K): In this window In a new window Download PPT slide   Figure 4. Inhibition of SF sterility by RZI transgenes. Ten generations after establishment of the RZI lines, the phenotypic effect of three copies of RZI was evaluated according to mating schemes A and B in Fig 3. Transmission of only two copies of RZI was analyzed through zero maternal generation (mating type C, Fig 3), as paternal transmission loses the insertion on the X chromosome. SF sterility was recorded daily and in triplicate over the first week of egg laying. Results are expressed in percentage sterility (see MATERIALS AND METHODS). Control SF females, dashed lines; solid symbols, transgenic (Tg+) SF females; open symbols, nontransgenic (Tg-) SF females. Shaded areas, maternal effects (A and B) and paternal effect (C).

The SF control exhibited high sterility, indicating that the reactivity level was not affected by our experimental conditions. However, after maternal transmission of the transgenes over several generations (Fig 4A), transgenic (Tg⁺) SF females exhibited full reversion of the sterility phenotype (<10% dead embryos) as early as the first day of egg laying. Strikingly, total repression of hybrid dysgenesis was also recorded in the nontransgenic SF females (Tg^-). This proved the existence of a maternal repressive effect, induced only by the presence of the transgenes within the maternal genome.

To test whether the maternal effect developed fully within a single generation, or increased over several generations, the protocol of Fig 3B was applied: three copies of RZI were transmitted through one maternal generation. The results (Fig 4B) show that a maternal effect was still observed in the Tg^- SF females, but that it was weaker: the effect was significant only from the fourth day of egg laying and it rescued even fewer embryos at day 7 (compare with Fig 4A). Thus the maternal effect induced by RZI is a quantitative effector that accumulates over generations.

Paternal transmission (Fig 3C) could not be tested for three copies of RZI, because all lines had a transgene insertion on the chromosome X that is not transmitted from father to daughters; instead, the zygotic effect of two copies of RZI was evaluated. A pure paternal effect, distinct from a zygotic effect, was identified as the difference of repression between the Tg^- and control SF females, as the SF females aged (Fig 4C). This paternal inheritance of repression, which has not been previously described, proves that some inherited effectors of regulation can pass through male meiosis.

Zygotic and maternal effects increased with the number of RZI copies, but whether the paternal effect follows the same rule was not as clear (Table 1). At first glance, zygotic and maternal and paternal effects seemed to be additive; however, individual and precise evaluation of each effect became difficult (1) when individual values fell below 10% and lost statistical significance (for examples, see Table 1) or (2) when one effect reached a value >80% and masked any other effect (in Fig 4A the accumulated maternal effect is so high that the zygotic effect—the difference between the Tg⁺ and Tg^- curves—of three copies of RZI can no longer be observed). In addition, all these effects were dynamic and varied as the SF females aged (Fig 4). To compare the various experimental conditions described here, the percentages of rescued embryos were combined within strains (from day 1 to 7) and over strains according to the transgene copy number (see MATERIALS AND METHODS) to give the "average effect of the average line" (see Table 1, RZI).

The paternal and maternal effects of RZI suggest that epigenetic mechanisms are involved in the regulation of hybrid dysgenesis by this transgene. However, none of the previous hypotheses (see Introduction) could be excluded: (1) a defective and repressive ORF1 protein that was accumulated from our construct; (2) a putative activator of IF expression that was depleted by our construct; or (3) an epigenetic phenomenon induced by the introduction of IF sequences into the Drosophila genome, whatever their expression abilities might be. To discriminate among these hypotheses, we analyzed the effects of three new constructs: RZAP, 8S, and BB (Fig 1).

Zygotic effects are not induced by a defective ORF1 protein but by specific asRNAs:
Construct RZAP was derived from RZI, but had no functional IF promoter and was further mutated in the initiation codons at 187 and 586 (legend to Fig 1). Construct BB expressed asRNA, like RZAP, but contained only the central part of ORF1 (from 290 to 1104). Construct 8S was derived from RZI but had no actin5C promoter.

Table 1 shows that RZAP and RZI induced similar zygotic effects, but 8S and BB did not. Results with BB were unexpected; this construct expressed asRNA but none of the 16 transgenic lines ever tested positive for a zygotic effect. In fact, RZAP and RZI expressed the same asRNAs, while BB expressed shorter asRNAs that hybridized with 724 bases of their target, but neither with the 5'-untranslated region (UTR), nor the ATG187 region, nor the 1014–1104 segment of IF (Fig 1). This suggests that any of these parameters, or some combination of them, is required to develop a zygotic effect.

8S contains the same IF sequences as RZAP and RZI, but it did not express asRNA; its lack of a zygotic effect confirms, a contrario, that the zygotic effect observed with RZI is due to asRNAs longer than those produced by BB. It also rules out the hypothesis that a truncated protein encoded by ORF1 in RZI accounts for the zygotic and paternal effects.

Genetic features of the maternal effect:
A maternal effect was detected for all constructs—except for BB—and it increased with the copy number of the transgene and was additive with the zygotic effect (Table 1). One construct, 8S, exhibited only a maternal effect, an effect also developed by the IZ transgenes of JENSEN et al. 1995 that contain a complete ORF1 (from nucleotide 187 to 1464, compared to 187 to 1104 for 8S). Thus a full-length ORF1 was not required to induce the maternal repressive effect. However, the fact that BB had no maternal effect indicated either that the 5'-UTR (and/or the 1014–1104 part of ORF1) of the IF was responsible for the maternal effect or that the BB construct, despite its 724 bp from IF, was not long enough to mediate such a maternal effect.

Initial quantitative analysis showed that the maternal effect increased with strain aging. The stability of all effects was therefore tested. In fact, the efficiency of the maternal effect increased after transgene establishment, over at least 100 generations for RZI, while the zygotic and paternal effects remained constant (Fig 5). The kinetics suggest that this increase has no other boundary than the rescue of all embryos laid by SF females. This cumulative property is also shared, to a lesser extent, by reactivity, but distinguishes the maternal effect from the paternal effect.

View larger version (25K): In this window In a new window Download PPT slide   Figure 5. Evolution of zygotic, maternal, and paternal effects as a function of strain aging. The effects of a single copy of transgene were evaluated 10, 20, 50, and 100 generations after transgene establishment using protocols in Fig 3A and Fig C. The evaluation was repeated three times in parallel for each line. Average values from all the lines at our disposal are expressed in percentage of embryos rescued by zygotic (hatched box), maternal (open box), or paternal (solid box) effect.

The stability of the maternal effect was evaluated one, two, or three generations after transgene removal (Fig 6). Two generations after removal, the repressive effect was still evident (20% of all embryos rescued). At the third generation, however, repression of hybrid dysgenesis was significant (rescue > 10%; P < 0.05) only for the maternal effect induced by two or three copies of the transgene: 15 ± 5, 18 ± 7, and 19 ± 9% of maternally rescued embryos over the week when derived from the 2RZI, 3RZI, and 2RZAP lines, respectively (Fig 6B and Fig C). However, when only the 6th and 7th days of egg laying were considered, the maternal effect induced by a single copy of the transgene was higher than 10% at P < 0.05 (not shown). Altogether, these results demonstrated a surprisingly long-lasting memory of the maternal effect after transgene removal.

View larger version (28K): In this window In a new window Download PPT slide   Figure 6. Perdurance of the maternal effect of each transgene. The maternal effect was evaluated for all transgenic lines in Tg- SF females, as shown in A. Sterility of Tg- SF females was recorded during the first week of egg laying, when the reactive mother was heterozygous for the transgene (G0), or one, two, or three generations after transgene removal (G1, G2, and G3, respectively). At each generation, the percentage sterility of these Tg- SF females was compared to the percentage sterility of an SF control, as in Fig 3. (B) Perdurance of maternal effect induced by one, two, or three copies of RZI. (C) Perdurance of maternal effect induced by one or two copies of RZAP, or one copy of 8S.

Paternal effect, maternal effect, and reactivity:
The paternal effect was induced by RZI and RZAP, but was markedly weaker than the maternal effect (Table 1). In fact, it was never significant before the 5th day of egg laying (Fig 4C and Fig 7), which may explain why it has never been described before; however, the paternal effect induced by two copies of the transgenes could rescue as many as 40% of all embryos laid by SF females on the 7th day of laying. The very existence of such an effect supports previous hypotheses on the involvement of epigenetic mechanisms in the regulation of IF. The paternal effect is, however, different from the maternal effect: (1) it is not affected by strain aging (above and Fig 5); (2) it appears only in lines already exhibiting a zygotic effect; and (3) it is not detected in the 8S lines even though they exhibit a maternal effect (Table 1). Thus the paternal effect is not just the mirror image of the maternal effect in a different meiosis. However, as a noncytoplasmic mechanism, it might be a component of the maternal effect. The paternal effect increased with female aging, like the maternal effect, while reactivity decreased (Fig 7). The apparently increasing efficacy of RZI, RZAP, and 8S during female aging might, then, be a consequence of this decrease of reactivity, while the actual transgene efficacy remained constant.

View larger version (24K): In this window In a new window Download PPT slide   Figure 7. Effect of SF female aging on the paternal effect, maternal effect, and reactivity. Dashed line, evolution of the reactivity level of SF females as a function of their aging (average value from 16 independent experiments). Continuous lines: maternal (m, solid symbols) and paternal (p, open symbols) effects of a single copy of RZI, RZAP, 8S, and BB, as a function of aging in young SF females. Data are from lines arbitrarily chosen 20 generations after transgene establishment.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

IF mobilization is responsible for the hybrid dysgenesis trait:
AsRNAs were used previously to demonstrate the involvement of the P element in P-M hybrid dysgenesis (SIMMONS et al. 1996 ). Here, this approach was used to demonstrate the involvement of another transposable element, the IF, in I-R hybrid dysgenesis. First, when the RZI and RZAP transgenes are inherited from I males, the zygotic component of the repressive effect reflects this "antisense" targeting of the IF transcripts within the germline of SF females, since 8S does not exert any zygotic repressive activity. Second, the comparison between RZAP and 8S clearly demonstrates that asRNAs are able to downregulate IF expression during oogenesis. Third, a comparison of RZAP and BB suggests that asRNAs must hybridize with the 5'-UTR and/or AUG187 and/or the 1014–1104 segment of the IF transcript to be fully effective. Fourth, the higher antisense efficiency of RZI compared to RZAP may be a consequence of dsRNA formation in the case of RZI, since dsRNAs have been shown to have efficient repressive properties in Drosophila as well as in Caenorhabditis elegans through a process called "RNA interference" (FIRE et al. 1998 ; KENNERDELL and CARTHEW 1998 ; HAMMOND et al. 2000 ).

How is IF responsible for SF lethality? Is it through expression of IF proteins? Is it a consequence of insertional mutagenesis on the host genome? After a dysgenic cross, the IF invades the R genome. During this process IF is expressed in the female germline for 1 to 10 generations (LACHAUME and PINON 1993 ), and the number of IFs increases until it reaches 10 to 15 copies per haploid genome. Sterility, however, can be recorded for many more generations (PELISSON and BREGLIANO 1987 ). Furthermore, overexpression of IF-encoded proteins in transgenic flies is not lethal (JENSEN et al. 1995 ; BUSSEAU et al. 1998 ), even though ORF2 encodes a putative DNA endonuclease known to be active in the related LINE L1Hs (FENG et al. 1996 ). These data suggest that IF expression, insertional mutagenesis, and DNA breaks induced during retrotransposition of the IF in ovaries cannot induce a high rate of SF sterility. Alternatively, IF insertions in chromosomal breaks arising at the beginning of meiotic recombination may induce a VAMOS response with lethal potential (LAURENCON et al. 1997 ). If so, IF mobilization may then be considered one of the many signals that can activate the VAMOS recombination-repair system in the female germline of D. melanogaster. This may be similar to the induction of the SOS system during Tn10 transposition (ROBERTS and KLECKNER 1988 ).

Regulation of the IF: involvement of epigenetic mechanisms:
IF sequences are strong mediators of their own repression through sense and antisense transcription (JENSEN et al. 1999A , JENSEN et al. 1999B ; this work). However, asRNA-producing constructs exhibit more than a zygotic effect: repression of I-R hybrid dysgenesis by the RZI and RZAP transgenes reveals additional maternal and paternal controls on IF mobilization. The maternal effect induced by IF sequences is not only efficient but also accumulates for 100 generations and still exists two generations after transgene removal. The paternal effect is transmitted by sperm devoid of the transgene. These results argue for epigenetic control of IF.

On the one hand, if maternal and paternal effects are related mechanistically, then at least part of the maternal effect is not obtained through accumulation of a regulatory molecule in the cytoplasm of the oocyte. On the other hand, reactivity is also able to vary slowly over generations through a cumulative maternal effect (BUCHETON and PICARD 1978 ) and may share a mechanistic basis with the parental effects described here.

Maternally inherited effects have already been reported in Drosophila for epigenetic controls such as those mediated by the Polycomb group (CAVALLI and PARO 1998 ). Paternal imprinting has also been described for the enhancer of the position effect variegation gene, E(var)3-93D (DORN et al. 1993 ). Such a possibility is further strengthened by the fact that the IF-lacZ transgene inserted at euchromatic loci exhibited a variegated pattern of expression during oogenesis (TATOUT et al. 1994 ). All these results, as well as the possible involvement of (1) IF promoter in transcriptional silencing (CHABOISSIER et al. 1998 ) and (2) IF transcripts in post-transcriptional silencing (JENSEN et al. 1999A , JENSEN et al. 1999B ), point toward epigenetic mechanisms as key effectors in the control of IF transposition and autoregulation. This raises several prospective questions. What are the molecular mechanisms that are able to trigger and establish such a repression? How is this repression maintained and transmitted?

Setting up the maternal effect through nucleic acid pairing?
Length, transcriptional ability, and features of transgenes are important parameters in the setting of the maternal effect. Transgenes with various lengths of IF (from 186 bp to 1.5 kbp) have been tested (JENSEN et al. 1995 , JENSEN et al. 1999A , JENSEN et al. 1999B ; CHABOISSIER et al. 1998 ; this work) with a wide range of efficacies: the shorter they are, the less effective they are. This might indicate that hybridization of nucleic acids is involved in the process. Indeed, JENSEN et al. 1999B propose that the maternal effect is induced by the presence of IF sense transcripts, which in turn induces a phenomenon analogous to cosuppression in plants (MATZKE and MATZKE 1995 ).

If the direct involvement of euchromatic transgenic sequences in hybridization still has to be proven, there are several limitations to the use of RNA as sole effector of this repression. First, abundant IF sense transcripts already exist in Drosophila; they are encoded in all strains by defective elements located in pericentromeric heterochromatin (VAURY et al. 1990 ; BUCHETON et al. 1992 ). Second, impairing sense transcription, as with the RZAP construct, should suppress the maternal effect, which it does not. A residual promoter activity might still exist in RZAP, but then it is difficult to understand why this construct has such pronounced maternal and paternal effects. This may be achieved through the production of aberrant RNAs from cryptic and adjacent promoters; such RNAs, like chimeric RNAs, have already been shown to mediate silencing (METTE et al. 1999 ; reviewed in WASSENEGGER and PELISSIER 1998 ).

Involvement of DNA sequences, through length and genomic copy number, could be consistent with a mechanism implicated in the control of dispersed repetitive elements like the IFs. Curiously, specific features of IF are required. Deletion of the 3' end of ORF1 from nucleotide 1104 to 1464 does not abolish the maternal effect (this work: RZI, RZAP, 8S), whereas further deletion (BB) of the 5' end (from 1 to 290) and 3' end (from 1014 to 1104) does. Transgenes limited to the untranslated leader of IF (from 1 to 186) have very low but still significant efficacy (CHABOISSIER et al. 1998 ), which suggests that the 5' end of IF is one of many segments able to interact with functional IFs to induce silencing.

Maintenance of the silencing:
Transgenes do not need to interact directly with the functional IFs in SF females to repress I-R hybrid dysgenesis: (a) paternal and maternal effects are observed in SF females without the transgene, and (b) two generations without the transgene decrease but do not abolish the maternal effect. If these effects are associated with a local change in chromatin structure at the IFs insertion points, they should be "remembered" through mitoses and meioses after transgene(s) removal.

One attractive hypothesis is that the maternal and paternal effects rely on the defective IF elements located at the pericentric heterochromatin to maintain the effects initiated by the transgenes. Strikingly, it has been shown that a heterochromatic P-element P1A, located near the telomere of the X chromosome, interacts with euchromatic P elements and induces trans-silencing (RONSERAY et al. 1996 ; ROCHE and RIO 1998 ). In fact, endogenous DNA sequences might be key mediators in homology-dependent gene silencing (PAL-BHADRA et al. 1999 ). Thus, (1) if transgenes such as RZI or RZAP are able to change the chromatin structure of the pericentric defective I elements, and (2) if these changes modify the level of expression of functional IFs even after transgene removal, it would be interesting to assess the effect of mutations affecting the trans-silencing of P elements. Involvement of the population of defective heterochromatic IFs in the regulation of euchromatic IFs might also explain the particular inheritance of reactivity.

Whatever the precise mechanism by which IF repression is achieved, the growing evidence that transposable elements are regulated through epigenetic phenomena opens a great field of investigation for an unsolved problem: how the host genome deals with advantages brought by transposable elements despite their tendency to invade and mutate the genome. Epigenetic effects might then be viewed as a real buffer against the massive expansion of transposable elements in eukaryotic genomes.

	FOOTNOTES

¹ These authors contributed equally to this work.
² Present address: Institut Gustave Roussy, CNRS UMR 1573, 39 rue Camille Desmoulins, 94805 Villejuif Cedex, France.

	ACKNOWLEDGMENTS

We thank R. M. Molina for her helpful assistance in performing statistical analysis. This work was supported by University Claude Bernard—Lyon1, and by grants 6577 and 9856 from "Association pour la Recherche contre le Cancer" (ARC). E.G. received a fellowship from the Ministère de l'Enseignement Supérieur de la Recherche et des Technologies (MESRT). C.T. received a fellowship from the European Community through the "Molecular Tools for Biodiversity" fund.

Manuscript received February 3, 1999; Accepted for publication August 28, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BIÉMONT, C., 1986  Polymorphism of the mdg-1 and I mobile elements in Drosophila melanogaster.. Chromosoma 93:393-397.

BLACKMANN, R. K., R. GRIMAILA, M. M. D. KOEHLER, and W. M. GELBART, 1987  Mobilization of hobo elements residing within the Decapentaplegic gene complex: suggestion of a new hybrid dysgenesis system in Drosophila melanogaster.. Cell 49:497-505[Medline].

BRÉGLIANO, J.-C., and M. G. KIDWELL, 1983 Hybrid dysgenesis determinants, pp. 363–410 in Mobile DNA, edited by D. E. BERG and M. M. HOWE. American Society for Microbiology, Washington, DC.

BRÉGLIANO, J. C., A. LAURENÇON, and F. DEGROOTE, 1995  Evidence for an inducible repair-recombination system in the female germ line of Drosophila melanogaster. I. Induction by inhibitors of nucleotide synthesis and by gamma rays. Genetics 141:571-578[Abstract].

BUCHETON, A., 1978  Non-mendelian female sterility in Drosophila melanogaster: influence of aging and thermic treatments. I. Evidence for a partly inheritable effect of these two factors. Heredity 41:357-369[Medline].

BUCHETON, A., 1979a  Non mendelian female sterility in Drosophila melanogaster: influence of ageing and thermic treatments. II. Action of thermic treatments on the sterility of SF females and on the reactivity of reactive females. Biologie Cellulaire 34:43-49.

BUCHETON, A., 1979b  Non mendelian female sterility in Drosophila melanogaster: influence of ageing and thermic treatments. III. Cumulative effects induced by these factors. Genetics 93:131-142[Abstract/Free Full Text].

BUCHETON, A. and G. PICARD, 1978  Non-mendelian female sterility in Drosophila melanogaster: hereditary transmission of reactivity levels. Heredity 40:207-223.

BUCHETON, A., J. M. LAVIGE, G. PICARD, and P. L'HERITIER, 1976  Non mendelian female sterility in Drosophila melanogaster: quantitative variations in the efficiency of inducer and reactive strains. Heredity 36:305-314[Medline].

BUCHETON, A., C. VAURY, M.-C. CHABOISSIER, P. ABAD, and A. PÉLISSON et al., 1992  I elements and the Drosophila genome. Genetica 86:175-190[Medline].

BUSSEAU, I., S. MALINSKY, M. BALAKIREVA, M.-C. CHABOISSIER, and D. TENINGES et al., 1998  A genetically marked I element in Drosophila melanogaster can be mobilized when ORF2 is provided in trans.. Genetics 148:267-275[Abstract/Free Full Text].

CAVALLI, G. and R. PARO, 1998  The Drosophila Fab-7 chromosomal element conveys epigenetic inheritance during mitosis and meiosis. Cell 93:505-518[Medline].

CHABOISSIER, M.-C., I. BRUSSEAU, J. PROSSER, D. J. FINNEGAN, and A. BUCHETON, 1990  Identification of a potential RNA intermediate for transposition of the LINE-like element I factor in Drosophila melanogaster.. EMBO J. 9:3557-3563[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].

COGONI, C., J. T. IRELAN, M. SCHUMACHER, T. J. SCMIDHAUSER, and E. U. SELKER et al., 1996  Transgene silencing of the al-1 gene in vegetative cells of Neurospora is mediated by cytoplasmic effector and does not depend on DNA-DNA interactions or DNA methylation. EMBO J. 15:3153-3163[Medline].

DAVID, J., 1959  Etude quantitative du développement de la drosophile élevée en milieu axenique. Bull. Soc. Biol. Fr. Belg. 93:472.

DE LA ROCHE SAINT ANDRÉ, C. and J. C. BRÉGLIANO, 1998  Evidence for a multistep control in transposition of I factor in Drosophila melanogaster.. Genetics 148:1875-1884[Abstract/Free Full Text].

DORN, R., V. KRAUSS, G. REUTER, and H. SAUMWEBER, 1993  The enhancer of position-effect variegation of Drosophila, E(var)3-93D, codes for a chromatin protein containing a conserved domain common to several transcriptional regulators. Proc. Natl. Acad. Sci. USA 90:11376-11380[Abstract/Free Full Text].

ECHOLS, H. and M. F. GOODMAN, 1990  Mutation induced by DNA damage: a many protein affair. Mutat. Res. 236:301-311[Medline].

FAWCETT, D. H., C. K. LISTER, E. KELLETT, and D. J. FINNEGAN, 1986  Transposable elements controlling I-R hybrid dysgenesis in D. melanogaster are similar to mammalian LINEs. Cell 47:1007-1015[Medline].

FENG, Q., J. V. MORAN, H. H. KAZAZIAN, JR., and J. D. BOEKE, 1996  Human L1 retrotransposon encodes a conserved endonuclease required for retrotransposition. Cell 87:905-916[Medline].

FINNEGAN, D. J., 1989  Eukaryotic transposable element and genome evolution. Trends Genet. 5:103-107[Medline].

FINNEGAN, D. J., 1992 Transposable elements pp. 1096–1107 in The Genome of Drosophila melanogaster, edited by D. L. LINDSLEY and G. G. ZIMM. Academic Press, New York.

FIRE, A., S. Q. XU, M. K. MONTGOMERY, S. A. KOTAS, and S. E. DRIVER et al., 1998  Potent and specific genetic interference by double-stranded RNA in C. elegans.. Nature 391:806-811[Medline].

FYRBERG, E. A., J. W. MAHAFFEY, B. J. BOND, and N. DAVIDSON, 1983  Transcripts of the six Drosophila actin genes accumulate in a stage- and tissue-specific manner. Cell 33:115-123[Medline].

GRANT, S. R., 1999  Dissecting the mechanisms of post-transcriptional gene silencing: divide and conquer. Cell 96:303-306[Medline].

HAMMOND, S. M., E. BERNSTEIN, D. BEACH, and G. J. HANNON, 2000  An RNA-directed nuclease mediates post-transcriptional gene silencing in Drosophila cells. Nature 404:293-296[Medline].

HAZELRIGG, T., R. LEVIS, and G. M. RUBIN, 1984  Transformation of white locus DNA in Drosophila: dosage compensation, zeste interaction, and position effects. Cell 36:469-481[Medline].

JENSEN, S. and T. HEIDMANN, 1991  An indicator gene for detection of germline retrotransposition in transgenic Drosophila demonstrates RNA-mediated transposition of the LINE I element. EMBO J. 10:1927-1937[Medline].

JENSEN, S., L. CAVAREC, M.-P. GASSAMA, and T. HEIDMANN, 1995  Defective I elements introduced into Drosophila as transgenes can regulate reactivity and prevent I-R hybrid dysgenesis. Mol. Gen. Genet. 248:381-390[Medline].

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

JENSEN, S., M.-P. GASSAMA, and T. HEIDMANN, 1999b  Cosuppression of I transposon activity in Drosophila by I-containing sense and antisense transgenes. Genetics 153:1767-1774[Abstract/Free Full Text].

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].

KIDWELL, M. G. and D. LISCH, 1997  Transposable elements as sources of variations in animals and plants. Proc. Natl. Acad. Sci. USA 94:7704-7711[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].

KUAN, C. T. and I. TESSMAN, 1991  LexA protein of Escherichia coli represses expression of the Tn5 transposase gene. J. Bacteriol. 173:6406-6410[Abstract/Free Full Text].

LACHAUME, P. and H. PINON, 1993  Germ-line expression of the I factor, a functional LINE from the fruit fly Drosophila melanogaster, is positively regulated by reactivity, a peculiar cellular state. Mol. Gen. Genet. 240:277-285[Medline].

LACHAUME, P., K. BOUHIDEL, M. MESURE, and H. PINON, 1992  Spatial and temporal expression of the I factor during oogenesis in Drosophila melanogaster.. Development 115:729-735[Abstract].

LAURENÇON, A., F. GAY, J. DUCAU, and J.-C. BRÉGLIANO, 1997  Evidence for an inducible repair-recombination system in the female germ line of Drosophila melanogaster. III. Correlation between reactivity levels, crossover frequency and repair efficiency. Genetics 146:1333-1344[Abstract].

LAVIGE, J. M., 1986  I-R system of hybrid dysgenesis in Drosophila melanogaster: further data on the arrest of development of the embryos from SF females. Biol. Cell 56:207-216.

LIM, J. K. and M. J. SIMMONS, 1994  Gross chromosome rearrangements mediated by transposable elements in Drosophila melanogaster.. Bioessays 16:269-275[Medline].

LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, San Diego.

LÜNING, K. G., 1981  Genetics of inbred Drosophila melanogaster.. Hereditas 95:181-188.

MATZKE, M. A. and A. J. M. MATZKE, 1995  How and why do plants inactivate homologous (trans)genes? Plant Physiol. 107:679-685[Medline].

MCLEAN, C., A. BUCHETON, and D. J. FINNEGAN, 1993  The 5' untranslated region of the I factor, a long interspersed nuclear element-like retrotransposon of Drosophila melanogaster, contains an internal promoter and sequences that regulate expression. Mol. Cell. Biol. 13:1042-1050[Abstract/Free Full Text].

METTE, M. F., J. VAN DER WINDEN, M. A. MATZKE, and A. J. M. MATZKE, 1999  Production of aberrant promoter transcripts contributes to the methylation and silencing of unlinked homologous promoters in trans.. EMBO J. 18:241-248[Medline].

O'NEILL, R. J. W., M. J. O'NEILL, and J. A. M. GRAVES, 1998  Undermethylation associated with retroelement activation and chromosome remodeling in an interspecific mammalian hybrid. Nature 393:68-72[Medline].

PAL-BHADRA, M., U. BHADRA, and J. A. BIRCHLER, 1997  Cosuppression in Drosophila: gene silencing of alcohol dehydrogenase by white-Adh transgenes is Polycomb dependent. Cell 90:479-490[Medline].

PAL-BHADRA, M., U. BHADRA, and J. A. BIRCHLER, 1999  Cosuppression of nonhomologous transgenes in Drosophila involves mutually related endogenous sequences. Cell 99:35-46[Medline].

PÉLISSON, A. and J.-C. BRÉGLIANO, 1987  Evidence for rapid limitation of the I element copy number in a genome submitted to several generations of I-R hybrid dysgenesis in Drosophila melanogaster.. Mol. Gen. Genet. 207:306-313.

PÉLISSON, A., D. J. FINNEGAN, and A. BUCHETON, 1991  Evidence for retrotransposition of the I factor, a LINE element of Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 88:4907-4910[Abstract/Free Full Text].

PICARD, G., 1978  Non mendelian female sterility in Drosophila melanogaster: further data on chromosomal contamination. Mol. Gen. Genet. 164:235-247.

ROBERTS, D. and N. KLECKNER, 1988  Tn10 transposition promotes RecA-dependent induction of a lambda prophage. Proc. Natl. Acad. Sci. USA 85:6037-6041[Abstract/Free Full Text].

ROCHE, S. E. and D. C. RIO, 1998  Trans-silencing by P elements inserted in subtelomeric heterochromatin involves the Drosophila Polycomb group gene, Enhancer of zeste.. Genetics 149:1839-1855[Abstract/Free Full Text].

RONSERAY, S., M. LEHMANN, D. NOUAUD, and D. ANXOLABEHERE, 1996  The regulatory properties of autonomous subtelomeric P elements are sensitive to a Suppressor of variegation in Drosophila melanogaster.. Genetics 143:1663-1674[Abstract].

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

RUBIN, G. M., M. G. KIDWELL, and P. M. BINGHAM, 1982  The molecular basis of P-M hybrid dysgenesis: the nature of induced mutations. Cell 29:987-994[Medline].

SAMBROOK, S., E. I. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Ed. 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SAN MIGUEL, P., A. TIKHONOV, Y.-K. JIN, N. MOTCHOULSKAIA, and D. ZAKHAROV et al., 1996  Nested retrotransposons in the intergenic regions of the maize genome. Science 274:765-768[Abstract/Free Full Text].

SELEME, M., I. BUSSEAU, S. MALINSKY, A. BUCHETON, and D. TENINGES, 1999  High-frequency retrotransposition of a marked I factor in Drosophila melanogaster correlates with a dynamic expression pattern of the ORF1 protein in the cytoplasm of oocytes. Genetics 151:761-771[Abstract/Free Full Text].

SELKER, E. U., 1999  Gene silencing: repeats that count. Cell 97:157-160[Medline].

SIMMONS, M. J., J. D. RAYMOND, C. D. GRIMES, C. BELINCO, and B. C. HAAKE et al., 1996  Repression of hybrid dysgenesis in Drosophila melanogaster by heat-shock-inducible sense and antisense P-element constructs. Genetics 144:1529-1544[Abstract].

TATOUT, C., M. DOCQUIER, P. LACHAUME, M. MESURE, and P. LECHER et al., 1994  Germ-line expression of a functional LINE from Drosophila melanogaster: fine characterization allows for potential investigations of trans-regulators. Int. J. Dev. Biol. 38:27-33[Medline].

TATOUT, C., E. GAUTHIER, and H. PINON, 1998  Rapid evaluation in Escherichia coli of antisense RNAs and ribozymes. Lett. Appl. Microbiol. 27:297-301[Medline].

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[Medline].

UDOMKIT, A., S. FORBES, C. MCLEAN, I. ARKHIPOVA, and D. J. FINNEGAN, 1996  Control of expression of the I factor, a LINE-like transposable element in Drosophila melanogaster.. EMBO J. 15:3174-3181[Medline].

VAURY, C., P. ABAD, A. PÉLISSON, A. LENOIR, and A. BUCHETON, 1990  Molecular characteristics of the heterochromatic I elements from a reactive strain of Drosophila melanogaster.. J. Mol. Evol. 31:424-431[Medline].

VOYTAS, D. F. and G. J. P. NAYLOR, 1998  Rapid flux in plant genomes. Nat. Genet. 20:6-7[Medline].

WASSENEGGER, M. and T. PELISSIER, 1998  A model for RNA-mediated gene silencing in higher plants. Plant Mol. Biol. 37:349-362[Medline].

This article has been cited by other articles:

				S. Robin, S. Chambeyron, A. Bucheton, and I. Busseau Gene Silencing Triggered by Non-LTR Retrotransposons in the Female Germline of Drosophila melanogaster Genetics, June 1, 2003; 164(2): 521 - 531. [Abstract] [Full Text] [PDF]

				S. Jensen, M.-P. Gassama, X. Dramard, and T. Heidmann Regulation of I-Transposon Activity in Drosophila: Evidence for Cosuppression of Nonhomologous Transgenes and Possible Role of Ancestral I-Related Pericentromeric Elements Genetics, November 1, 2002; 162(3): 1197 - 1209. [Abstract] [Full Text] [PDF]