Skip to main content

Modeling early germline immunization after horizontal transfer of transposable elements reveals internal piRNA cluster heterogeneity



A fraction of all genomes is composed of transposable elements (TEs) whose mobility needs to be carefully controlled. In gonads, TE activity is repressed by PIWI-interacting RNAs (piRNAs), a class of small RNAs synthesized by heterochromatic loci enriched in TE fragments, called piRNA clusters. Maintenance of active piRNA clusters across generations is secured by maternal piRNA inheritance providing the memory for TE repression. On rare occasions, genomes encounter horizontal transfer (HT) of new TEs with no piRNA targeting them, threatening the host genome integrity. Naïve genomes can eventually start to produce new piRNAs against these genomic invaders, but the timing of their emergence remains elusive.


Using a set of TE-derived transgenes inserted in different germline piRNA clusters and functional assays, we have modeled a TE HT in Drosophila melanogaster. We have found that the complete co-option of these transgenes by a germline piRNA cluster can occur within four generations associated with the production of new piRNAs all along the transgenes and the germline silencing of piRNA sensors. Synthesis of new transgenic TE piRNAs is linked to piRNA cluster transcription dependent on Moonshiner and heterochromatin mark deposition that propagates more efficiently on short sequences. Moreover, we found that sequences located within piRNA clusters can have different piRNA profiles and can influence transcript accumulation of nearby sequences.


Our study reveals that genetic and epigenetic properties, such as transcription, piRNA profiles, heterochromatin, and conversion efficiency along piRNA clusters, could be heterogeneous depending on the sequences that compose them. These findings suggest that the capacity of transcriptional signal erasure induced by the chromatin complex specific of the piRNA cluster can be incomplete through the piRNA cluster loci. Finally, these results have revealed an unexpected level of complexity that highlights a new magnitude of piRNA cluster plasticity fundamental for the maintenance of genome integrity.


Transposable elements (TEs) are mobile endogenous genetic elements predominantly transmitted through vertical transfer (i.e., from one generation to the next) like any other gene. In most metazoan germlines, transposition is repressed during gametogenesis by specialized 23–29-nt small RNAs associated with PIWI proteins [1,2,3]. These small RNAs have been named PIWI-interacting RNAs or piRNAs. They are produced from numerous heterochromatic loci, called piRNA clusters, that mainly contain TE fragments coming from ancient insertions serving as libraries of mobile sequences to repress. Among the 140 ovarian piRNA clusters of Drosophila melanogaster, only a few of them (the flamenco somatic and the 42AB, 38C, 80F, 20A germline clusters) have been extensively studied to identify factors involved in piRNA-dependent silencing [4,5,6,7,8,9]. It is well established that germline piRNA clusters are transcribed by a specialized RNA polymerase II complex that contains Moonshiner (Moon) recruited to the locus through its interaction with a complex containing the HP1 homolog Rhino (Rhi), Deadlock (Del), and Cutoff (Cuff) forming the RDC complex [6, 10]. Furthermore, only maternally inherited piRNAs participate in the transgenerational memory of TE sequences to repress as no paternally inherited piRNAs have been found [11, 12]. Although our understanding of the molecular events involved in the maintenance of active piRNA clusters through generations has expanded substantially, major gaps still exist especially in the early events of functional piRNA cluster establishment.

In gonads, 21-nt siRNAs can be also synthesized from dual-stranded piRNA clusters [13]. They are usually produced from double-stranded transcripts that are recognized by the Dcr-2 nuclease, and once loaded on the Ago2 protein, they can induce the cleavage of complementary RNA targets [14]. However, their germline function is not clear as they can be dispensable without significantly affecting viability, fertility, TE repression, and piRNA cluster maintenance [15].

Rarely, TEs are also transmitted through horizontal transfer (HT) corresponding to DNA transmission between unrelated individuals. It has been however noted that HTs are more frequent than originally thought [16,17,18] raising the question of how and by which dynamics new piRNAs are produced by naïve genomes in the absence of maternal inheritance of complementary piRNAs. One of the best documented HT is the one of P element that has successfully invaded the genome of natural populations of D. melanogaster within two decades during the twentieth century and of D. simulans since the beginning of the twenty-first century [19,20,21]. In D. melanogaster, the subtelomere of the X chromosome (cytological site 1A) is a hot spot of P insertions [22,23,24,25]. This locus also known as telomeric-associated sequences (X-TAS) is also one of the Drosophila piRNA clusters (hereafter called cluster 1A) that can be dispensable in laboratory environments [26]. Cluster 1A contains repeats with regions sharing homologies with the autosomal subtelomeric piRNA clusters of the 2R and the 3R chromosomes (clusters 60F and 100F) and a 0.9-kb region called T3 not present elsewhere in the genome, a unique feature among all known piRNA clusters [26]. From a P copy inserted in cluster 1A, piRNAs derived from P are produced in the female germline capable of repressing euchromatic active P elements [11, 27,28,29,30,31]. Moreover, lacZ encoding P transgenes inserted in cluster 1A (e.g., P(lArB) in P-1152 strain) have been shown to silence female germline expression of another P-lacZ transgene located in euchromatin [32, 33]. This euchromatic P-lacZ has served as a reliable reporter system (or “piRNA sensor”) for studying functional piRNA biology as its silencing depends on piRNA biogenesis factors [5, 12, 34,35,36,37,38] but not on siRNA biogenesis factors [15, 34]. Indeed, using this sensor, we have shown that the mechanism of repression is accomplished according to an ON/OFF mode, where egg chambers show either strong (ON) or no (OFF) lacZ silencing [34]. We have also shown that when subtelomeric P(lArB) transgenes were paternally inherited, the number of fully repressed egg chambers in the first generation is low and increases progressively to reach a maximum level of repression after five generations [34].

We have also found that a naïve locus made of seven tandemly repeated P(lacW) transgenes, in the strain BX2, that is maintained as a non-piRNA producer over the years, can be fully converted into a stable piRNA cluster in one generation by maternally inherited piRNAs matching the whole length of the transgenes, uncovering a stable case of epigenetic conversion called paramutation [12]. The switch from a naïve locus to an active one able to produce piRNAs, hereafter referred to as “conversion,” is associated with an enrichment of H3K9me3 [5, 39]. Such functional conversion can occur when the locus producing maternal piRNAs is located on different chromosomes and is partially homologous to P(lacW) [15]. Moreover, ovarian small RNA analyses revealed that conversion of the full length of P(lacW) can be completed when tested after the third generation [15]. These results along with others suggested that the piRNA machinery is able to eventually co-opt an unknown sequence from the maternal piRNA repertoire to produce de novo piRNAs of this new sequence [15, 40,41,42]. A similar scenario could happen when a naïve genome faces HT of new TEs that insert into piRNA clusters. At first, the TE copy newly integrated into a piRNA cluster is surrounded by sequences that are targeted by maternal piRNAs (Additional file 1: Fig. S1). In fine, new piRNAs against this TE will be synthesized and able to repress active euchromatic copies. The rareness of such event and the repetitive nature of piRNA clusters have made it difficult to directly address the precise latency and the identification of early molecular events involved in these co-option processes (Additional file 1: Fig. S1C).

We report here a study where we have modeled a TE neo insertion into a piRNA cluster in a naïve genome to question the kinetics of production of new specific piRNAs and their capacity to repress from the first generation, necessary to protect genome integrity. To model such event, we have used several transgenes, derived from the P transposon, inserted in different piRNA clusters or inserted in euchromatin working as piRNA sensors. Using the paternal origin of transgenes inserted in piRNA clusters, we have been able to correlate the emergence of new piRNA production with their silencing capacities using functional assays from the very first generation. We have also shown that the kinetics of co-option by piRNA cluster leading to the conversion of a sequence could depend on intrinsic properties such as its length. We have identified that all regions of the transgene are converted concomitantly with the same efficiency at each generation, but this conversion is restricted to sequences nested in piRNA clusters as previously shown [43] revealing an active mechanism preventing cis-propagation of piRNA clusters to their flanking regions [44]. By studying more specifically a germline subtelomeric piRNA cluster, cluster 1A, from the P-1152 strain containing P(lArB) and T3 sequences, we have identified that heterogeneity can be observed inside piRNA clusters as they can exhibit different rates of conversion and different piRNA profiles (symmetrical and asymmetrical dual-strand clusters) associated with chromatin and transcription variations along the locus. Altogether, this study brings new insights into piRNA cluster dynamics.


Functional conversion of paternally inherited subtelomeric transgenes completed within four generations is associated with piRNA synthesis

It was previously observed that silencing of a lacZ sensor induced by subtelomeric P(lArB) transgenes inserted in cluster 1A was female germline-specific, with a maternal effect that showed variegated ON/OFF lacZ egg chambers repression (between 80 and 100% of repressed egg chambers) and dependent on the piRNA biogenesis pathway [12, 34, 36, 38, 45]. By contrast, paternally inherited P(lArB) induced lacZ silencing in few ovarian egg chambers in the first generation (between 5 and 35%) that increased in subsequent generations [34]. The progressive increase in the number of repressed egg chambers per ovary suggested that the amount of lacZ piRNAs per ovary produced by the subtelomeric P(lArB) was proportionally increasing at each generation.

To test this model, we have set up reciprocal crosses between the P-1152 and the Canton strains. P-1152 contains two P(lArB) transgenes inserted in cluster 1A (Fig. 1A, B). Canton lacks cluster 1A (Δ-1A strain) and is devoid of P-derived transgenes (Additional file 1: Fig. S2A, Table 1, Additional file 2: Table S1 [26]). However, Canton is carrying the autosomic subtelomeric piRNA clusters (clusters 60F and 100F) that produce piRNAs targeting the common regions between cluster 1A and the autosomal clusters (T1, T2, T4, and INV-4, Fig. 1A [26]). For these crosses, P(lArB) were first paternally or maternally inherited and then maternally maintained as hemizygous in the successive generations to establish the paternal and the maternal lineages (PI and MI, respectively) (Fig. 1C). Four independent replicate lines were generated for each lineage. Ovarian lacZ silencing was assayed at each generation by crossing PI and MI females with males containing the lacZ sensor (Additional file 1: Fig. S2B). Ovarian X-gal staining of the progenies allowed to quantify the level of conversion of P(lArB) into an active piRNA-producing locus. When P(lArB) were paternally inherited (PI), the first generation showed a limited amount of repressed egg chambers (G1; 9.25%) that progressively reached the same level of repression as the maternal lineage after four generations (> 90%, Fig. 1D and Additional file 2: Table S2), reminiscent of previous results with another lacZ sensor [34]. In parallel, to correlate the level of repression with the production of subtelomeric P(lArB) piRNAs, ovarian small RNAs were sequenced, and the normalized 23–29-nt P(lArB) small RNAs were analyzed at each generation. At G1, the number of 23–29-nt small RNAs in PI females was low compared to MI females increasing progressively at each generation reaching a plateau at G4 corresponding to the same amount of 23–29-nt small RNAs than in MI females (Fig. 1E, F, Additional file 1: Fig. S3A and Additional file 2: Table S3). This correlates with the lacZ repression (Fig. 1D). The enrichment in uridine as the first nucleotide for the 23–29-nt small RNAs (1U bias) (Fig. 1F) and an enrichment of ping-pong pairs (Fig. 1E), two signatures of germline piRNA biogenesis, together with previous mutant analyses reinforced the assumption that these small RNAs are genuine piRNAs [5, 12, 34, 36, 38]. Therefore, these results correlate the lacZ silencing efficiency with piRNA amount. Furthermore, the distribution of piRNAs on sense and antisense along P(lArB) corresponds to a dual-strand piRNA cluster profile (Fig. 1E). These results were reproduced in different genetic backgrounds (Canton and w1118, Additional file 1: Fig. S4) and with another P derived transgene inserted in the autosomal subtelomeric piRNA cluster 100F (Additional file 1: Fig. S5). Interestingly, piRNAs synthesis was homogeneously increasing for all regions whatever their position within P(lArB) in PI, with roughly the same kinetics over generations. The most distal parts of the transgene, adjacent to sequences targeted by maternally inherited piRNAs, do not show a quicker conversion process than the internal parts. These results support the hypothesis that the kinetics of conversion was roughly the same independently of their position within the transgene, their sequence origin (Drosophila or E. coli), or their nature (genes or P-derived sequences) (Fig. 2, Additional file 1: Fig. S3C-G). This is in accordance with the capacity of Rhino and its partners to erase internal transcriptional signals, leading most likely to a uniform production of piRNAs independently of the sequence origin [4, 6, 46]. Taken together, these results support the idea that the percentage of egg chambers in an ON or OFF state for lacZ expression (Fig. 1D) reflects the piRNA amount across paternal inheritance (Fig. 1F).

Fig. 1
figure 1

Heterogeneity inside cluster 1A. A Subtelomeric cluster 1A on the X chromosome in the P-1152 strain composed of several repeats (n) containing solo LTRs of INV-4 (gray), T1 (blue), T2 (pink), and T4 (yellow) domains also found in autosomal subtelomeric piRNA clusters and the T3 domain (0.9 kb) found only in cluster 1A (green). One of the repeats contains the P(lArB) transgenes (18 kb) (asterisks). B The P(lArB) transgene includes the 5′ and 3′ P-derived sequences, a plasmid sequence, rosy, Adh, lacZ (under the control of the P promoter), and the two transcriptional terminators (tCG and tHs). Black arrows represent the sense of transcription. C Paternal (PI) and maternal (MI) inheritances of cluster 1A were obtained by two reciprocal crosses between P-1152 and Canton strains. The maternal alleles are above the fraction. P-1152 carries P(lArB) and T3 in cluster 1A (“P(larB); T3 + ”), absent in Canton (“Δ-1A”). D Ovarian lacZ repression in the PI and MI P(lArB) lineages. Values represent the mean of four sublines with standard deviation. E Experimental design showing regions complementary to maternal piRNAs (small black lines) in G1. Maternal piRNAs in PI are produced by autosomal subtelomeres. Below are the size distributions of normalized 20–29-nt counts and the relative frequency (z-score) of overlapping sense-antisense small RNA pairs in the subsets of 23–29-nt RNAs matching P(lArB) and T3, showing enrichment of 10 nucleotides overlaps. The sense piRNAs are in red, and the antisense are in blue. F, G Normalized 23–29-nt reads mapping to P(lArB) (F) and T3 (G) of the MI H and PI D sublines. The percentage of 23–29-nt P(lArB) and T3 RNAs beginning with a 5′ uridine (1U bias) and characteristic of piRNAs are indicated for G4. Note that P(lArB) (18 kb), X subtelomeric repeat (1.8 kb), and T3 (0.9 kb) are not drawn to scale

Table 1 Strains used in this study
Fig. 2
figure 2

Normalized 23–29-nt reads on P(lArB) regions in MI and PI lineages. Normalized ovarian 23–29-nt reads mapping to different regions of P(lArB): plasmid (A), rosy (B), Adh (C), lacZ (D) genes, and the P-derived sequences (E) in the MI H and PI D sublines (Additional file 2: Table S3). All the sequences, either exogenous (lacZ, plasmid, and P) or endogenous (rosy, Adh) from the Drosophila genome, have similar kinetics of producing a progressive increase number of 23–29-nt small RNAs over generations

Consistent with an HT of TE into a naïve genome, G1 males of P(lArB)-PI contribute only to DNA transgenic copies without contributing to complementary piRNAs inheritance. Moreover, our results suggest that maternally inherited piRNAs produced by the autosomal subtelomeres in Δ-1A females can target and convert progressively across generations the subtelomeric repeats surrounding P(lArB) insertions (Fig. 1E, F and Additional file 1: Fig. S1).

P(lArB) and T3 show distinct conversion dynamics

Cluster 1A is composed of repeats shared with other piRNA clusters in autosomal subtelomeres and of a unique 0.9-kb T3 domain not found outside of cluster 1A (Fig. 1A) [26]. In view of the progressively increasing number of P(lArB) piRNAs in the paternal lineage (Fig. 1F), we hypothesized the same dynamic over four generations for a paternally inherited T3 domain. To test this, ovarian small RNA libraries were reanalyzed by aligning the 23–29-nt reads to the T3 domain. Unexpectedly, the same amount of normalized T3 23–29-nt small RNAs was found in both paternal and maternal lineages from the first generation (Fig. 1E, G, Additional file 1: Fig. S3B and Additional file 2: Table S3). This was observed using the Canton or w1118 Δ-1A strain genetic background (Additional file 1: Fig. S4D). As in our previous work, the 23–29-nt T3 small RNAs were mostly produced from one strand leading to an asymmetrical dual-strand piRNA cluster, showing an enrichment in uridine at first nucleotide (1U bias) [26] and an overlapping of 10 nucleotides bias among the small RNA pairs (a ping-pong signature, Fig. 1E, G).

To test if T3 small RNAs were functional for repression from the first generation, we designed a P-derived transgenic sensor using the red fluorescent protein (RFP) reporter gene transcriptionally fused to the T3 domain and expressed under the control of the UASp sequences (pRFP-T3). RFP expression was induced in female germline by the Gal4 protein expressed under the control of the nanos (nos) promoter (Fig. 3A). Female germline expression of pRFP-T3 was observed in the Δ-1A w1118 background confirming the absence of other T3 piRNA sources (Fig. 3B). When the P-1152 strain was used as a donor of T3 small RNAs, almost complete silencing of pRFP-T3 was observed induced by small RNAs produced by both the T3 domain and P-derived sequences of P(lArB) (Additional file 1: Fig. S6A and B). To avoid this, we used the Oregon strain, as a donor of T3 small RNAs [26], because this strain is devoid of P-derived sequences (P transgene or natural P element) and carries cluster 1A [1, 26]. Ovarian pRFP-T3 expression was repressed when cluster 1A of the Oregon strain was maternally inherited (100% of RFP repressed egg chambers (n = 1008), Fig. 3C and Additional file 1: Fig. S6C) suggesting that small RNAs produced from T3 are indeed fully functional germline repressors. Moreover, the paternally inherited T3 locus from Oregon was also able to strongly repress pRFP-T3 from the first generation (99.6% of RFP repressed egg chambers (n = 748), Fig. 3D). Unlike P(lArB), T3 paternal allele is functionally converted in a single generation in all cells, and in direct correlation with the amount of T3 small RNAs detected in PI G1 females (Fig. 1E, G and Additional file 1: Fig. S3B). To confirm that the pRFP-T3 silencing was piRNA mediated, we knocked down germline expression of Piwi and Nxf2, two co-transcriptional silencing factors of the piRNA pathway, Bootlegger (Boot) that recruits nuclear export factors like Nxf3-Nxt1 to piRNA cluster loci [1, 47] and Moon, a subunit specific of germline piRNA cluster RNA polymerase [10]. Figure 3E shows that RFP silencing was strongly affected by the knockdown of these factors supporting the notion that 23–29-nt T3 small RNAs are functional piRNAs targeting pRFP-T3 reporter in the female germline. These knockdowns were also affecting lacZ sensor silencing induced by subtelomeric P(lArB) (Additional file 1: Fig. S7).

Fig. 3
figure 3

Cluster 1A relies on the germline piRNA pathway. A Schematic representation of the experimental cross: Oregon flies containing cluster 1A carrying T3 and the nos-Gal4 germline driver were crossed with w1118 flies devoid of cluster 1A (Δ-1A) but expressing the pRFP-T3 sensor. B Strong ovarian germline RFP expression of progenies from control nos-Gal4 females crossed with males encoding the pRFP-T3 sensor in the absence of cluster 1A. Maternally (C) or paternally (D) inherited T3 strongly represses ovarian germline expression of the pRFP-T3 piRNA sensor. E Ovarian pRFP-T3 repression of maternally inherited T3 is strongly affected by germline knockdown of piwi, nxf2, boot, and moon (piwi-KD, nxf2-KD, boot-KD, moon-KD). Knockdown for white served as control. Repression was assayed by counting the percentage of RFP-silenced egg chambers at stages 8–10. The total numbers of counted egg chambers are indicated in parenthesis. RFP expression is in red, and DAPI staining, indicating nuclei, is in white. Parental crosses are indicated above micrographs

Thus, paternally inherited P(lArB) and T3 are piRNA-producing sequences that display different rates of conversion as well as different piRNA distribution profiles (symmetric dual-strand for P(lArB) and asymmetrical dual-strand for T3), although they are located in the same locus and dependent on the same piRNA pathway. Functionally, these results might indicate that piRNA cluster activation is dependent on maternal piRNA inheritance at each generation that targets the flanking subtelomeric regions (Additional file 1: Fig. S1), but also on properties of sequences present within the locus.

Sequence length could influence the conversion efficiency

The contrasting conversion rate observed in G1 between the paternally inherited 0.9-kb T3 domain and the 18-kb P(lArB) suggests that short sequences could be converted more efficiently than longer ones. To test whether the sequence length could influence the frequency of conversion, we used several strains: the seven tandemly repeated P(lacW) transgenes in the BX2 strain that can be converted into an active piRNA cluster by complementary maternal piRNAs [12, 15], the P-1152 strain, and the RS3 strain that carries the P(RS3) transgene inserted in the autosomal 3R subtelomeric piRNA cluster, cluster 100F (Fig. 4, Table 1 and Additional file 2: Table S1). Both the P(lArB) and P(lacW) transgenes encode the 3.5-kb E. coli lacZ gene and the 1.8-kb bacterial plasmid backbone (Fig. 4A, C). The P(RS3) and P(lacW) transgenes both encode the 4.1-kb white gene (Fig. 4B, D). In addition, P(lArB), P(RS3), and P(lacW) have in common the 5′ (0.58 kb) and 3′ (0.23 kb) distal regions of the P element.

Fig. 4
figure 4

Conversion of the P(lacW) transgenes by P(lArB) or P(RS3). Diagrams of P(lArB) inserted in cluster 1A (A) and P(RS3) inserted in cluster 100F (B). piRNAs produced by both clusters are represented by small colored lines below the transgenes. Crosses to convert the seven P(lacW) transgenes inserted in tandem by P(lArB) (C) or by P(RS3) (D). Complementary maternal piRNAs produced by either P(lArB) or P(RS3) are indicated above the P(lacW) scheme. Normalized reads of 23–29-nt mapping to lacZ (E, H), white (F, I), and plasmid sequence (G, J). When P(lacW) transgenes are activated by P(lArB), the density of 23–29-nt small RNAs between G1 and G4 is similar for lacZ (E) and the plasmid sequence (G) and 2.25-fold higher for white (red box, F). When P(lacW) transgenes are activated by P(RS3), the density ratio of 23–29-nt small RNAs is close to 1 between G1 and G7 for all the domains of P(lacW) (1.3 for lacZ (H), 0.8 for white (I) and 1.3 for the 1.8-kb plasmid region (red box, J) that is not targeted by maternal piRNAs in G1). The density of normalized 23–29-nt reads per kb (reads/kb) and the fraction of 1U bias at 5′ are indicated in each panel. The P(lArB) (18 kb), P(RS3) (6 kb), and P(lacW) (10.7 kb) transgenes are not drawn to scale

In this set of experiments, hemizygous P(lArB) and P(RS3) females (donors of piRNAs) were crossed to BX2 males hemizygous for P(lacW) (Fig. 4 and Additional file 1: Fig. S8). The G1 progenies that paternally inherited the P(lacW) transgenes and maternally inherited the piRNAs from either P(lArB) or P(RS3), but not the subtelomeric transgenes, were then crossed to each other for several generations (Fig. 4C, D and Additional file 1: Fig. S8). Previously, ovarian 23–29-nt RNAs from G3 up to G10 mapping to the different regions of P(lacW) were identified at a time when the complete conversion was reached, that is to say when stable paramutation already occurred [15]. Here, to question the conversion establishment, ovarian 23–29-nt RNAs were analyzed as soon as G1 and up to G4 or G7 generations. piRNAs matching lacZ and plasmid sequences maternally inherited from P(lArB) were able to convert in G1 the complementary regions in P(lacW) (Fig. 4E, G). The lacZ gene, expressed as a transcriptional fusion with P first exons, and the white gene of P(lacW) were also converted from the first generation by complementary piRNAs synthesized by the maternal P(RS3) allele (density ratio of 1.3 and 0.8 between G1 and G7, respectively, Fig. 4H, I, 100% of lacZ sensor silencing Fig. 4H and Additional file 1: Fig. S8). Strikingly, the piRNA population in the G1 progeny was limited for the 4.1 kb white sequence absent in P(lArB) (658 reads/kb) that significantly increased in G4 (1480 reads/kb) (ratio of 2.25 between G1 and G4, Fig. 4F), while the piRNAs for the 1.8-kb plasmid sequence not included in P(RS3) were already detected at G1 (ratio of 1.3 between G1 and G7, Fig. 4J).

Therefore, combining these results with the fact that the 18-kb P(lArB) transgene requires several generations to be fully converted when it is paternally inherited (Fig. 1) and that the 0.9-kb T3 conversion occurs into only one generation, we suggest that the efficiency of conversion of a new sequence inserted into a piRNA cluster targeted by maternally inherited piRNAs, but not targeted itself, could depend on its length.

We have also tested whether the conversion rate could be influenced by the nucleotide composition of sequences. One hypothesis is that the T3 and plasmid sequences might be enriched in some dinucleotides compared to the P(lArB) and white sequence. To address this, we have computed the dinucleotide content of the four sequences. We have found that the content of dinucleotides is quite consistent between sequences with different lengths. We only noticed a slight bias toward AT/TA dinucleotides for the T3 sequence, but not for the others including the plasmid, that is converted with the same efficiency as T3 (Additional file 1: Fig. S9A, Additional file 2: Table S5). To understand the potential role of this bias on sequence conversion rate, we have included it in a linear regression model (piRNA.density ~ sequence.length + AT or TA) but obtained poor p-values (0.665 and 0.199, respectively) to conclude. Therefore, it is unlikely that dinucleotide content, in itself, might explain the difference of conversion efficiency between the considered sequences.

Finally, we have tested if small RNAs with imperfect mapping (i.e., 3 mismatches) with the references sequences could exist in the parental strains that could participate to the one-generation conversion of T3 and the plasmid, while absent or less abundant for P(lArB) and white gene. The parental strains are Canton and w1118 for T3 and P(lArB) (Fig. 1 and Additional file 1: Fig. S4), P-1152 for white gene included in P(lacW) transgene (Fig. 4F), and RS3 for plasmid included in P(lacW) (Fig. 4J). For T3, 3 mismatch piRNAs were identified in the Canton and w1118 parental strains (1563 and 1378 reads/kb, respectively) (Additional file 1: Fig. S9B and C, Additional file 2: Table S6). This result was expected, as the L subtelomeric repeats (left arms of the autosomal chromosomes shown to be piRNA clusters [1, 26]) contain a small domain with similarities with T3 [26]. It has been shown that the Piwi protein from a sponge specie can be tolerant to mismatches for the piRNA target binding but requires extensive pairing for the endonuclease activity preventing unwanted mRNA targeting [48]. In Drosophila, the 3 mismatches piRNAs mapping to T3 in Canton and w1118 do not have a ping-pong signature (Additional file 1: Fig. S9F and G) and are not functional for the silencing of the pRFP-T3 piRNA sensor (Fig. 3B, Additional file 1: Fig. S6B). It can be also noted that the region sharing similarities with the L subtelomeres does not show a high piRNA density of 0 mismatch T3 piRNAs (see the read counts around position 600 in Fig. 1E, Additional file 1: Fig. S4D, S9B, C). This suggests that the L piRNAs do not participate to the conversion of the T3 domain. Few 3 mismatch piRNAs were identified mapping to the plasmid in RS3 (8.8 reads/kb), and with the same order of magnitude for the white gene in P-1152 (5.8 reads/kb) and to P(lArB) in the Canton, w1118 (8.1 and 9.5 reads/kb, Additional file 1: Fig. S9B, C, D, E and Additional File 2: Table S6). Therefore, no common feature concerning the role of 3 mismatch small RNAs was observed between T3 and the plasmid. Altogether, these results strongly suggest that the conversion efficiency of a sequence, not targeted in G1 but flanked by sequences targeted by the maternal piRNA pool depends, at least in part, on its length: in one generation, a low frequency of conversion can occur for sequences longer than 4 kb (i.e., white or P(lArB)), whereas the high frequency of conversion occurs for shorter sequences (i.e., plasmid sequence or T3).

Conversion is restricted to sequences embedded within pre-existing piRNA clusters

The above results defined the conversion of loci surrounded on both sides by sequences targeted by maternal piRNAs. We then examined whether such conversion could also spread onto adjacent genomic sequences. Few 23–29-nt small RNAs flanking the insertion site of P(lacW) were detected that were not increasing between G1 and G4 (Additional file 1: Fig. S10A and B). The majority of them correspond to the transcribed strand of Ago1 where the array of P(lacW) is inserted, suggesting that they were produced primarily by phasing without amplification [49]. The same analysis was performed on CG17636, the first gene on the X chromosome, close to cluster 1A. Few and unchanged 23–29-nt matching CG17636 were identified between G1 and G4 (Additional file 1: Fig. S10C). Similar results were observed on endogenous homologous loci present in P(lArB) transgenes (Additional file 1: Fig. S11A, B, C). We conclude from these results that spreading of conversion from transgene sequences in cis as observed here and in earlier studies [43] or in trans outside the piRNA clusters is very limited suggesting the existence of a tight control that restricts piRNA cluster spreading and defines precisely their borders [44], like the transcription of genes flanking cluster 42AB or cluster 80F [10].

Cluster 1A is a heterogeneous piRNA cluster

Although P(lArB) and T3 are located in the same cluster 1A of the P-1152 strain, their piRNA profiles and kinetics of conversion when paternally inherited are different (Fig. 1E–G). To understand early molecular events occurring in G1, we first analyzed the ovarian heterochromatin throughout cluster 1A. Using chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR, with primers shown in Fig. 5A, Additional file 1: Table S7), a high trimethylated Lysine 9 of Histone 3 (H3K9me3) enrichment was found on P(lArB), when maternally inherited as compared to paternally inherited (Fig. 5B), confirming previous observations [5]. On T3, high H3K9me3 enrichment was observed in both PI and MI, with the overall level of H3K9me3 on PI being higher than on P(lArB) (Fig. 5B). Therefore, maternally inherited piRNAs can induce H3K9me3 enrichment on all sequences of cluster 1A. However, in PI G1, H3K9me3 enrichment is heterogeneous along the 1A locus, from weak on lacZ to high on T3, consistent with their piRNA productions and silencing of lacZ and pRFP-T3 piRNA sensors (Figs. 1, 2, and 3).

Fig. 5
figure 5

Chromatin state and steady-state transcription of cluster 1A. A Schematic representation of one of the subtelomeric repeats of cluster 1A of the P-1152 strain, containing the PlArB transgenes (indicated by an asterisk). Yellow arrows indicate the position of qPCR primers. The X subtelomeric repeats (1.8 kb) and the P(lArB) (18 kb) are not drawn to scale. B H3K9me3 ovarian enrichment on three different regions of P(lArB) and of T3 in maternal and paternal inheritance (MI, red; PI, blue) measured by ChIP-qPCR in G1. The signal was normalized to the 42AB region highly enriched in H3K9me3 marks. ChIP experiments were performed on three independent biological samples. P-values were calculated using a bilateral t-test (n = 3). B Ovarian RNA accumulations of G1 P(lArB) and T3 in P-1152 (T3(P-1152)) and T3 in Oregon (T3(Oregon)) were measured by RT-qPCR in control KD (white-KD) and moon-KD and normalized to the expression of RpL32 gene. P-values were calculated using a one-way ANOVA test followed by a Tukey HSD test (n = 3). ns, not significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (Additional file 2: Table S8)

We next asked whether the chromatin variations observed in P(lArB) and T3 and piRNAs synthesis were correlated with the RNA steady state of cluster 1A. For this, we analyzed the ovarian RNA accumulation across the locus by RT-qPCR experiments. Using the same primers (Fig. 5A), RNA accumulation of P(lArB) and T3 in the P-1152 strain were found higher in PI than in MI (see white-KD control in Fig. 5C). After four generations, differences between PI and MI were no longer detected, as the system reached an equilibrium similar to the maternal lineage (Fig. 1F, G, and Additional file 1: Fig. S12A). One explanation is that, in MI, high amounts of P(lArB) and T3 piRNAs address the complementary P(lArB) and T3 transcripts to the piRNA biogenesis inducing their degradation. In PI, low amounts of P(lArB) piRNAs (Fig. 1E, F) could instead prevent their recognition as piRNA precursors allowing accumulation of P(lArB) transcripts (Fig. 5C). Surprisingly, the high amount of T3 piRNA in PI (Fig. 1E, G) was not correlated with low accumulation of T3 containing transcripts (Fig. 5C). We therefore wondered if the presence of P(lArB) could affect T3 RNA accumulation. As cluster 1A of Oregon was used to obtain P(lArB) insertion in the P-1152 strain [50], we measured T3 RNA accumulation in the Oregon strain (“T3(Oregon)”) and found unexpectedly that T3 RNA steady state was unchanged between MI and PI contrary to T3 in the vicinity of P(lArB) (“T3(P-1152)”) (Fig. 5C). The size and structure of germline piRNA cluster transcripts are still unknown; however, we assumed that chimeric transcripts can exist between different domains of piRNA clusters. Based on this assumption, in T3(Oregon) PI, chimeric transcripts containing T3 and the other subtelomeric domains (INV-4, T1, T2, T4) can be targeted by piRNAs produced by the autosomal subtelomeric piRNA clusters. These transcripts are then directed to the piRNA degradation pathway with the same efficiency as in MI (Fig. 5C). In T3(P-1152) PI, the chimeric transcripts containing T3 and subtelomeres are processed as described above leading to the production of piRNAs observed in Fig. 1G. In addition, the chimeric transcripts containing T3 and P(lArB) can accumulate because they are not efficiently targeted by maternally inherited piRNAs that lack T3 and P(lArB) (Fig. 5C).

We then tested if the non-canonical transcription specific of germline piRNA cluster was directly required for transcription of P(lArB) and T3 in both genomic contexts (T3(P-1152) and T3(Oregon)). Germline knockdown of moonshiner (moon-KD, Additional file 1: Fig. S12B and C), involved heterochromatic transcription of dual-strand germline clusters [10], exhibited a clear increase of P(lArB) and T3 RNA steady state compared to the control white-KD MI (Fig. 5C). These results indicate that cluster 1A transcription is Moon dependent, that these transcripts are funneled to the piRNA machinery, and that the presence of P(lArB) affects the RNA steady-state of T3 in T3(P-1152) (Figs. 3E and 5C). One explanation is that transcriptional signals (initiation, termination) of P(lArB) transgenes might not be totally erased by the piRNA cluster chromatin and could therefore influence adjacent T3 RNA accumulation.

According to dual-strand piRNA synthesis (Fig. 1E), accumulation of P(lArB) transcripts in G1 PI (Fig. 5C) could lead to the production of double-stranded RNAs that could be potentially processed into siRNAs in the absence of maternally inherited piRNAs. Consistent with a bioRxiv preprint from Luo et al. [51], we have questioned whether siRNAs were produced in parallel to piRNAs. We have compared the kinetics of occurrence of siRNAs and piRNAs during the P(lArB) conversion process in PI over the four generations (Fig. 1F). Indeed, a high amount of P(lArB) siRNAs is accumulated in G1 PI that persists across the first 4 generations, whereas piRNAs require the 4 generations to reach the plateau (Fig. 6A and Additional file 1: Fig. S13). The same profile of small RNA distribution was detected for all regions of the transgene in the Canton background and in the w1118 genetic background (Additional file 1: Fig. S13A, B, and D). No such siRNA amount was found in the MI lineage, where P(lArB) was converted a long time ago (Fig. 6B and Additional file 1: Fig. S13C). Importantly, functional assays indicate that these transgenic siRNAs are not functional for the silencing of the P(lacZ) reporter (Fig. 1D, Additional file 2: Table S2). To complete this observation, we have also looked at siRNAs corresponding to the white gene when activated by P(lArB) (Fig. 4F). In this context, white siRNAs were produced from the first generation with a less spectacular abundance compared to white piRNAs, than in the case of P(lArB) G1 conversion, and that accumulate in G4 (Fig. 6C). Thus, the presence of siRNAs could precede the production of piRNAs, but this is not a general phenomenon. Their emergence can be also the result of accumulation in G1 PI of transcripts that are not targeted by maternal piRNAs and become the subtract of Dcr-2 endonuclease, in accordance with the fact that siRNAs were shown to be dispensable for germline piRNA cluster maintenance, silencing of piRNA sensor and paramutation [15, 34].

Fig. 6
figure 6

siRNAs and piRNAs abundance during conversion. Size distribution of ovarian small RNAs isolated from the P(lArB)-PI, subline D (A), or the P(lArB)-MI subline H (B) matching to T3 and P(lArB) in G1, G2, G3, and G4. When paternally inherited, P(lArB) is converted for piRNA synthesis progressively across generations, while T3 is converted from the first generation (G1). C Size distribution of small RNAs isolated in the G1 and G4 progenies of females containing P(lArB) crossed with males containing P(lacW) mapped on the white sequence. In this context, white gene is progressively converted (Fig. 4F). The ratio number of normalized 23–29-nt over 21 nt is indicated for the P(lArB) panel


It is now well established that, although TEs are harmful to host genomes, they are also a source of evolution providing a large range of genetic opportunities, like new tissue gene-expression or novel genes [52]. Few of these events can be positively selected which has led to the notion of “molecular domestication” [53]. TEs have the capacity to transpose into or close to each other [54] forming highly enriched regions of TE fragments [55]. Some of these regions were shown to encode piRNAs involved in TE silencing in gonads of Metazoans [1, 3], the final domestication event granting TEs control of their own mobility. Accordingly, our understanding of the biochemistry and the genetics of piRNA biogenesis is well advanced [56]. In Drosophila, maternally inherited piRNAs can activate loci containing transgenes derived from P or I elements [12, 43, 57] and a euchromatic transgene can be activated for piRNA synthesis when its transcription is directed by the promoter of HeT-A, one of the telomeric TEs [58]. Early studies on hybrid dysgenesis revealed how newly horizontally transferred P elements could be maintained in Drosophila populations because of their capacity to insert into cluster 1A [11, 22,23,24, 30, 31, 34, 45]. This piRNA cluster can be lost under laboratory environment without affecting fly survival despite the fact that P could amplify and become stabilized most likely by new insertions into other piRNA clusters [26]. Analogous cases were found with a deletion of the flamenco locus that somatically derepressed the ZAM element leading to its neo insertion into a germline piRNA cluster [40] or after transgenesis of Penelope within D. melanogaster genome and its subsequent integration into a piRNA cluster [59]. These new insertions were in fine able to produce new piRNAs. However, how and at which rate newly inserted TE copy lose euchromatic identity to acquire piRNA cluster ones was unknown.

Model of TE co-option by piRNA cluster: the secret of a successful horizontal transfer

Horizontal transfers of TEs been described in various organisms highlighting that such events can occur more frequently than originally thought [16]. However, the mechanisms and kinetics involved in taming these new invaders are not well understood. It has been modeled that upon a new insertion into a piRNA cluster, TE sequences are embedded in a locus that is targeted by maternal piRNAs. In our work, the paternally inherited P(lArB) transgenes inserted in cluster 1A or the white gene of P(lacW) activated by the P-1152 strain recapitulate a TE, in length, inserted into a piRNA cluster in a naïve genome. Previously, it was shown in D. simulans that 20–40 generations are required to efficiently produce new piRNAs directed against the P element when monitored from the beginning of the invasion under environmental stress [60]. This result is in accordance with an evolutionary model that was based on several independent studies on the P element in D. melanogaster [61] and on the capacity of the P cytotype to take place across generations [62]. Other simulation studies have described three phases during TE invasion: a proliferative period followed by the occurrence and segregation of cluster insertions and finally fixation of cluster insertions [41]. The latter step requires reprogramming of piRNA cluster properties as shown for the ZAM element in Drosophila [40]. Once integrated, our results suggest that at each generation, egg chamber conversion will occur, and the TE copies will acquire stable piRNA cluster marks (Fig. 7) licensing non-canonical transcription [6, 9]. Once the conversion occurs, this epigenetic state is stably transmitted to the next generations. Altogether our work indicates that the long-term success of a new TE invasion depends principally on its capacity to insert into a piRNA cluster, because once inserted, any further transposition will be impeded. In addition, it has been proposed that some euchromatic copies of the new invaders might be targeted by the new piRNAs and contribute also to the production of piRNAs [63, 64]. A number of studies have also suggested that heterochromatic regions, like pericentromeric ones containing piRNA clusters, show a large accumulation of TEs as well as strong suppression of recombination [65, 66] limiting most likely deletions within inserted TE copies. Thus, while speculative but also based on our observations, some of the successful TE HT might also drive positive selection toward non-autonomous deleted copies, instead of full-length copies, inserted into piRNA clusters. Moreover, the efficiency of conversion which seems to depend on the size of the sequences (Figs. 1 and 4) might be interpreted as a safeguard that might exist to define piRNA cluster borders, avoiding deleterious propagation on flanking euchromatic regions. Future works will be necessary to identify factors involved in defining piRNA cluster borders and canonical vs non-canonical piRNA-related transcription. Finally, this work highlights that maternal piRNA legacy is a fundamental prerequisite for genome stability that has to be established and then inherited at each generation in order to maintain active maternal and paternal piRNA cluster alleles.

Fig. 7
figure 7

Model of sequence cooption revealing the existence of piRNA production heterogeneity within a piRNA cluster. A Maternally inherited piRNAs maintain active germline piRNA cluster at each generation (left), whereas paternal gametes transmit their DNA but no piRNAs to their progenies (right). B When a newly horizontally transferred TE inserts into a piRNA cluster, it is not efficiently targeted at the first generation (G1) by the maternal piRNA pool, even if the region is transcribed, resulting in a non-piRNA cluster conversion. This context is modeled in this study by the paternally inherited P-derived transgene, P(lArB) (red arrow), having a size closed to some autonomous TE but lacking complementary piRNAs. However, although paternally inherited and lacking homologous piRNAs, shorter sequences can be efficiently converted in G1, like T3 (brown arrow). C Four generations (G4) can be sufficient to convert full-length TE, leading to new piRNA production. This conversion can lead to the synthesis of piRNAs with different piRNA profiles throughout the locus. Sense and antisense schematic piRNA profiles are extrapolated from this study, Marie et al., and Asif-Laidin et al. [26, 35]. Arrows symbolize TE fragments or repeats found in piRNA clusters, red and blue are for the sense and antisense piRNAs, and darker colors represent the “new” piRNAs, i.e., identified in a given generation (G1 or G4) and absent in the parents

Non-allelic paramutations between piRNA clusters occur at each generation to maintain genome stability

Because piRNA clusters are composed for the most part of fragments of TEs, it is difficult to conduct an extensive analysis of their dynamics. We concentrated our effort on cluster 1A, because strains devoid of this locus exist, T3 is restricted to the X subtelomere and it is long enough to conduct simple molecular analysis [26]. Assuming that T3 can recapitulate to some extent the properties of the whole native cluster 1A, our analysis indicates that maternally inherited piRNAs targeting redundant subtelomeric sequences, such as INV-4, T1, T2, and T4, produced by the autosomal subtelomeres, are able to epigenetically convert paternally inherited T3 sequence from the first generation. Such epigenetic conversion, dependent on maternal piRNAs able to convert allelic and non-allelic loci, is known as paramutation [12, 15]. It might be the keystone to convert paternal loci of germline piRNA clusters at each generation. In addition, studies of piRNA cluster through several Drosophila species identified that piRNA clusters are genomic regions evolving rapidly due, in part, to recurrent chromosomal rearrangements [63, 67]. Once again, non-allelic paramutation appears to be an efficient mechanism to maintain the mandatory pool of piRNA producers activated by maternally inherited piRNAs produced by any loci to maintain genome integrity. These results reveal the selection forces that might exist to preserve hundreds of piRNA clusters with redundant sequences that secure efficient TE control at each generation (Fig. 7).

Germline piRNA clusters can be heterogeneous entities

In Drosophila, germline piRNA clusters can be dual-strand or uni-strand clusters. So far, most descriptions of piRNA clusters have considered these loci of several dozens of kb as epigenetically and functionally homogeneous regions. The cluster 42AB, the largest dual-strand cluster, is H3K9me3 enriched that recruits Rhi, which binds to Deadlock interacting with Moon allowing initiation of transcription on both strands. On the contrary, cluster 20A is a uni-strand cluster whose transcription is Moon-independent and mildly enriched in H3K9me3 [6, 10]. Here, we have dissected the molecular and genetic properties of cluster 1A by focusing on two domains (P(lArB) and T3) which are separated by about 500 bp from each other. Although they are both H3K9me3 enriched and Moon dependent, our results indicate that they differ by the ratio of sense vs antisense piRNA distribution. P(lArB) results in a symmetric dual-strand region, while T3 is an asymmetric dual-strand domain (Fig. 7). This observation might indicate that piRNA clusters are not strictly epigenetically determined but may involve genetic properties, which are not completely erased by the piRNA cluster chromatin. In addition, they differ by the kinetics of conversion when they are paternally inherited in the absence of maternal complementary piRNA inheritance. These results have revealed an unexpected feature for piRNA clusters like cluster 1A in that they can be rather heterogeneous (Fig. 7). Is this property unique to cluster 1A? In the light of our results and inspection of that of others, cluster 80F displays also a heterogeneous piRNA profile distribution being a dual-strand cluster for the most part of the locus and a uni-strand/asymmetric cluster in the 3′ region [6, 10, 47]. Furthermore, in the absence of moon, germline 42AB and 38C1 dual-strand piRNA clusters display an increasing number of piRNAs localized to the 3′ region of cluster 42AB and to the whole cluster 38C1 [10]. Therefore, it would be interesting to design piRNA sensors for different parts of these clusters to functionally test their silencing in this context. RT-qPCR data revealed that Moon is required for the transcription of P(lArB) and T3 with an increase of about twofold of RNA accumulation for P(lArB) and T3. Finally, our data suggest that the primary properties of a sequence to become part of a piRNA cluster, in the absence of complementary maternally inherited piRNAs, could depend on the genomic environment, on the composition of chromatin but also on intrinsic features of sequences, for instance, transcriptional signals. Recently, Kipferl, a DNA-binding protein, has been shown to recruit Rhi at piRNA clusters through specific sequences and is required for some functional piRNA clusters [68]. This underlines a genetic and an epigenetic part for piRNA cluster determination. Therefore, by dissecting cluster 1A, we uncovered heterogeneity within large piRNA clusters revealing that piRNA cluster loci are more intricate regions than originally thought.


Using transgenic tools inserted into piRNA clusters, piRNA sensors, cluster 1A, and the fact that no paternally inherited piRNAs are transmitted to the progenies, we could address the dynamics of production of new piRNAs from sequences located within piRNA cluster but absent from the maternal piRNA repertoire. This situation is encountered in the case of horizontal transfer of TEs. During our analysis on cluster 1A as a model of horizontal transfer events, we identified that this cluster is able to respond within 4 generations to co-opt a sequence of about a TE size, not belonging to the maternal piRNA repertoire, whose transcription become moon dependent leading to the synthesis of new piRNAs. Our study also identified that a piRNA cluster can be heterogeneous, as different domains can display different piRNA profiles leading to symmetric next to asymmetric dual-strand piRNA cluster regions. As piRNA clusters contain TE fragments, these results lead us to suggest that some of these fragments might have an influence on the piRNA cluster biology leading to the heterogeneity that we observed.


Drosophila strains

P-1152 (FBti0005700) strain carries two enhancer trap P(lArB) transgenes that contain an in-frame translational fusion of the E. coli lacZ gene to the second exon of the P transposase gene [50]. They are inserted in one of the subtelomeric repeats of the X chromosome that is one of the piRNA clusters, we named cluster 1A (Fig. 1A, B) [32, 45, 50]. This strain also contains the autosomic subtelomeric repeats on the 3R chromosome, that is a piRNA cluster (cluster 100F) [26] and is mutant for the white gene (wc allele). The Canton strain is devoid of P element and cluster 1A [26], hereafter referred to as a Δ-1A strain. The Canton line used is marked by a yellow allele affecting body color pigmentation and is derived from the classical laboratory strain. The BX2 strain contains seven P(lacW) transgenes (FBtp0000204) tandemly inserted in euchromatin on the second chromosome [69]. P(RS3) is a P-FRT-white transgene (FBtp0001534) inserted in the subtelomere of the 3R chromosomal arm (cluster 100F) [70]. The lacZ piRNA sensor (BQ16, FBtp0000154) carries a euchromatic P-lacZ fusion enhancer trap transgene strongly expressing lacZ in the germline [71]. When P-1152 females are crossed by lacZ piRNA sensor males, lacZ is strongly repressed in the female germline [32, 34]. Oregon strain is a classical laboratory strain devoid of the P element with the X (cluster 1A), 2R (cluster 60F), and 3R (cluster 100F) subtelomeric piRNA clusters [26]. Its X chromosome was present in the strain used to generate the P-1152 strain [50].

The pRFP-T3 transgenic strain carries an insertion of the RFP reporter gene transcriptionally fused to the T3 domain. It was obtained after transformation into the w1118 strain, a Δ-1A strain. This T3 piRNA sensor was obtained by cloning the PCR-amplified RFP gene (using primers: 5′-AGGTACCATGCCCAAGAAGAAGCGCAAGGTGGCCTCCTCCGAGGACGTCATCAAG-3′ and 5′-ATCTAGATTAGGCGCCGGTGGAGTGGC-3′) into the pUASp plasmid [72] digested with KpnI and XbaI, followed by inserting into the XbaI site, the PCR amplified T3 sequence (primers used: 5′-AATCTAGACCCAGCAAATTTATGGATAAAC-3′ and 5′-ATTCTAGACCTAATTTTTGGCAAAGTTGTAC-3′). Female germline expression of pRFP-T3 is driven by the PBac{w[+ mW.hs] = GreenEye.nosGAL4} (FBtp0056000) transgene. Germline-specific knockdowns (KD) were performed by crossing short hairpin or RNA interference transgenes [10, 47, 73] directed against genes under the UAS sequence with the Gal4 maternal driver under the promoter of the nos gene (FBtp0056000 transgene). Additional file 2: Table S1 summarizes the characteristics of each strain used in this study.

Experimental conditions and measurements of lacZ and RFP silencing

All crosses were performed at 25 °C. The y and w marker genes present on the Canton (Δ-1A) and P-1152 strains, respectively, were used to control the subtelomere segregation in successive generations as well as to identify rare recombination events. We generated hemizygous females for the subtelomeric P(lArB) transgenes with a maternal (MI) or paternal (PI) inheritance (Fig. 1C) that were then crossed with Canton males in successive generations. In parallel, silencing capacities of MI and PI lines were measured using a germline functional assay at each generation for four independent sublines by crossing hemizygous females at each generation with males containing the lacZ piRNA sensor (Additional file 1: Fig. S2B). Ovarian lacZ expressions were assayed using X-gal staining and the repression was quantified as previously described by determining the percentage of egg chambers with no lacZ expression in the germline [34]. Ovarian RFP expressions were assayed by fixing ovaries in 5% formaldehyde during 6 min, washing in 1X PBS (3 times), and overnight incubating in PBS/DAPI. The ovaries were then washed 3 times in PBS 1X and then spread on a slide with a mounting medium (80% glycerol in PBS, 4% propylgalate) [74]. The silencing capacity of subtelomeric T3 piRNAs was quantified by determining the percentage of repressed egg chambers carrying a pRFP-T3 reporter transgene. The quantification was restricted to stages 8–10 of egg chambers where the RFP expression was intense and reproducible. Images were acquired with an Axio-ApoTome (Zeiss) and ZEN2 software.

Ovarian small RNA sequencing and analysis

Total RNA extraction, small RNA libraries preparation, and sequencing were performed as previously described [15]. Sequence reads in fastq format were trimmed from the adapter sequence 5′-TGGAATTCTCGGGTGCCAAG-3′ and matched the reference sequences using Bowtie [75]. Annotation of small RNA libraries is described in Additional file 2: Table S4. Small RNAs matching the reference sequences with 0 mismatches were retained for subsequent analysis, except when notified. Sequence length distributions and small RNA mapping were generated from bowtie alignments using Python and R ( scripts, which were wrapped and run in a Galaxy instance publicly available at Tools used in this study may be downloaded from this Galaxy instance. For library comparisons, read counts were normalized (normalization factor) relative to the number of sequence reads aligning to the D. melanogaster genome but not to miscRNAs (including rRNA and snoRNA) or tRNAs [12]. Similar results were obtained, when tested by normalizing the read counts to one million miRNA reads or to one million D. melanogaster reads (Additional file 1: Fig. S14A, Table S4A). Furthermore, no global differences among libraries were identified when comparing the ratio of 23–29-nt reads over the miRNA reads between several libraries (Additional file 1: Fig. S14B, Additional file 2: Table S4).


For each sample, 2 μg of total RNA was treated with DNase (Fermentas). One microgram of DNase-treated RNA was used for reverse transcription (RT) using random hexamer primers (Fermentas). Real-time qPCR was performed on triplicates of each sample using primers referred to in Additional file 2: Table S7. The RpL32 gene was used as a reference. The same series of dilution of a mix of different RT preparations was used to normalize the quantity of transcripts in all RT preparations leading to standard quantity (Sq) values. Variations between technical triplicates were very low when compared to variations between biological replicates. The mean of the three technical replicates was then systematically used (meanSq). For each biological sample, we calculated the ratio meanSq(gene)/meanSq(Reference Gene) to normalize the transcript quantity. Then, the mean of each sample ratio was used to compare the two conditions.


For each sample, 50 to 100 pairs of ovaries from 2- to 3-day-old females were manually dissected in cold 1X PBS, crossed-linked with 1.8% formaldehyde, quenched with glycine, washed with 1X PBS, and collected by centrifugation. Pellets were flash-frozen in liquid nitrogen. The ovaries were ground with a pestle and lysed in ChIP lysis buffer (NaCl 100 mM, Tris pH8 50 mM, EDTA 5 mM, 1% SDS). Lysates were sonicated using Bioruptor (Bioruptor Standard Diagenode) three times during 15 min (30 s ON, 30 s OFF) and cleared by centrifugation. Five percent of cleared lysate was set aside to serve as input samples, and the remainder was divided in two equal portions and incubated at 4 °C with antibodies overnight under gentle rotation (anti-H3K9me3 from Active Motif Cat#39161, negative IgG control from Diagenode Low Cell ChIP-kit Cat#803-015). Magnetic beads coupled to G protein (Dynabeads Protein G, Invitrogen Cat#10003D) were washed two times in low salt wash buffer (0.1% SDS, 1% Triton X-100, EDTA 2 mM, Tris pH8 20 mM, NaCl 150 mM), transferred into immunoprecipitated lysate and incubated at 4 °C during 1 h under gentle rotation. The beads were washed two times in low-salt wash buffer, one time in high-salt wash buffer (0.1% SDS, 1% Triton X-100, EDTA 2 mM, Tris pH8 20 mM, NaCl 500 mM), two times in LiCl wash buffer (LiCl 0.25 M, 1% NP-40, 1% Na deoxycholate, EDTA 1 mM, Tris pH8 10 mM), and two times in TE buffer (Tris pH8 10 mM, EDTA 1 mM). Decrosslinking, elution, and DNA purification were performed using an IPure kit (Diagenode Cat# C03010015).

Real-time PCRs were performed on duplicates for each biological sample leading to cycle threshold (Ct) values. Variations between technical duplicates were very low compared to variations between biological replicates. The mean of the two technical replicates was then systematically used (meanCt). For each sample, the IP fraction was normalized beside input to take account of sample preparation difference as follows: ΔCt [normalized ChIP] = (meanCt [ChIP] − (meanCt [Input] − Log2 (Input dilution factor))) where meanCt [ChIP] is the Ct value measure for immunoprecipitated samples, and meanCt [Input] is the Ct value measure for input and input dilution factor corresponds to the chromatin fraction set aside for input (in this experiment, 5% of chromatin fraction was set aside, thus input dilution factor was 20). To confirm specific antibody signals compared to the negative control, fold enrichment was calculated for each sample as follows: fold enrichment = 2(−ΔCt [normalized ChIP] − ΔCt [normalized NS]. Fold enrichment values are then normalized on the 42AB region, used as a control region.

Availability of data and materials

Small RNA sequencing data have been deposited at the European Nucleotide Archive (ENA) under accession numbers PRJEB11491 for GRH12 (, [15, 76]), PRJEB19350 for GRH129 and GRH133 (, [26, 77]), PRJNA899103 ( [78]), and PRJNA935551 ( [79]).



Transposable element


PIWI-interacting RNA


Horizontal transfer


Telomeric-associated sequences


Base pair




Reverse transcriptase


Quantitative PCR


Chromatin immunoprecipitation


Red fluorescent protein


Paternal inheritance


Maternal inheritance













RDC complex:

Rhi-Del-Cuff complex


Long terminal repeat




  1. Brennecke J, Aravin AA, Stark A, Dus M, Kellis M, Sachidanandam R, et al. Discrete small RNA-generating loci as master regulators of transposon activity in Drosophila. Cell. 2007;128:1089–103.

    Article  CAS  PubMed  Google Scholar 

  2. Girard A, Sachidanandam R, Hannon GJ, Carmell MA. A germline-specific class of small RNAs binds mammalian Piwi proteins. Nature. 2006;442:199–202.

    Article  PubMed  Google Scholar 

  3. Gunawardane LS, Saito K, Nishida KM, Miyoshi K, Kawamura Y, Nagami T, et al. A slicer-mediated mechanism for repeat-associated siRNA 5’ end formation in Drosophila. Science. 2007;315:1587–90.

    Article  CAS  PubMed  Google Scholar 

  4. Chen Y-CA, Stuwe E, Luo Y, Ninova M, Le Thomas A, Rozhavskaya E, et al. Cutoff suppresses RNA polymerase II termination to ensure expression of piRNA precursors. Mol Cell. 2016;63:97–109.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Le Thomas A, Stuwe E, Li S, Du J, Marinov G, Rozhkov N, et al. Transgenerationally inherited piRNAs trigger piRNA biogenesis by changing the chromatin of piRNA clusters and inducing precursor processing. Genes Dev. 2014;28:1667–80.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Mohn F, Sienski G, Handler D, Brennecke J. The rhino-deadlock-cutoff complex licenses noncanonical transcription of dual-strand piRNA clusters in Drosophila. Cell. 2014;157:1364–79.

    Article  CAS  PubMed  Google Scholar 

  7. Rangan P, Malone CD, Navarro C, Newbold SP, Hayes PS, Sachidanandam R, et al. piRNA production requires heterochromatin formation in Drosophila. Curr Biol. 2011;21:1373–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Rozhkov NV, Hammell M, Hannon GJ. Multiple roles for Piwi in silencing Drosophila transposons. Genes Dev. 2013;27:400–12.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Zhang Z, Wang J, Schultz N, Zhang F, Parhad SS, Tu S, et al. The HP1 homolog rhino anchors a nuclear complex that suppresses piRNA precursor splicing. Cell. 2014;157:1353–63.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Andersen PR, Tirian L, Vunjak M, Brennecke J. A heterochromatin-dependent transcription machinery drives piRNA expression. Nature. 2017;549:54–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Brennecke J, Malone CD, Aravin AA, Sachidanandam R, Stark A, Hannon GJ. An epigenetic role for maternally inherited piRNAs in transposon silencing. Science. 2008;322:1387–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. de Vanssay A, Bougé A-L, Boivin A, Hermant C, Teysset L, Delmarre V, et al. Paramutation in Drosophila linked to emergence of a piRNA-producing locus. Nature. 2012;490:112–5.

    Article  PubMed  Google Scholar 

  13. Czech B, Malone CD, Zhou R, Stark A, Schlingeheyde C, Dus M, et al. An endogenous small interfering RNA pathway in Drosophila. Nature. 2008;453:798–802.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Matranga C, Tomari Y, Shin C, Bartel DP, Zamore PD. Passenger-strand cleavage facilitates assembly of siRNA into Ago2-containing RNAi enzyme complexes. Cell. 2005;123:607–20.

    Article  CAS  PubMed  Google Scholar 

  15. Hermant C, Boivin A, Teysset L, Delmarre V, Asif-Laidin A, van den Beek M, et al. Paramutation in Drosophila requires both nuclear and cytoplasmic actors of the piRNA pathway and induces cis-spreading of piRNA production. Genetics. 2015;201:1381–96.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Gilbert C, Feschotte C. Horizontal acquisition of transposable elements and viral sequences: patterns and consequences. Curr Opin Genet Dev. 2018;49:15–24.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Hill T, Betancourt AJ. Extensive exchange of transposable elements in the Drosophila pseudoobscura group. Mob DNA. 2018;9:20.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Loreto ELS, Carareto CMA, Capy P. Revisiting horizontal transfer of transposable elements in Drosophila. Heredity (Edinb). 2008;100:545–54.

    Article  CAS  PubMed  Google Scholar 

  19. Bingham PM, Kidwell MG, Rubin GM. The molecular basis of P-M hybrid dysgenesis: the role of the P element, a P-strain-specific transposon family. Cell. 1982;29:995–1004.

    Article  CAS  PubMed  Google Scholar 

  20. Anxolabéhère D, Kidwell MG, Periquet G. Molecular characteristics of diverse populations are consistent with the hypothesis of a recent invasion of Drosophila melanogaster by mobile P elements. Mol Biol Evol. 1988;5:252–69.

    PubMed  Google Scholar 

  21. Kofler R, Hill T, Nolte V, Betancourt AJ, Schlötterer C. The recent invasion of natural Drosophila simulans populations by the P-element. Proc Natl Acad Sci USA. 2015;112:6659–63.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Ronsseray S, Lehmann M, Anxolabéhère D. Copy number and distribution of P and I mobile elements in Drosophila melanogaster populations. Chromosoma. 1989;98:207–14.

    Article  CAS  PubMed  Google Scholar 

  23. Karpen GH, Spradling AC. Analysis of subtelomeric heterochromatin in the Drosophila minichromosome Dp1187 by single P element insertional mutagenesis. Genetics. 1992;132:737–53.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Zhang S, Pointer B, Kelleher ES. Rapid evolution of piRNA-mediated silencing of an invading transposable element was driven by abundant de novo mutations. Genome Res. 2020;30:566–75.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Ajioka JW, Eanes WF. The accumulation of P-elements on the tip of the X chromosome in populations of Drosophila melanogaster. Genet Res. 1989;53:1–6.

    Article  CAS  PubMed  Google Scholar 

  26. Asif-Laidin A, Delmarre V, Laurentie J, Miller WJ, Ronsseray S, Teysset L. Short and long-term evolutionary dynamics of subtelomeric piRNA clusters in Drosophila. DNA Res. 2017;24:459–72.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Reiss D, Josse T, Anxolabehere D, Ronsseray S. Aubergine mutations in Drosophila melanogaster impair P cytotype determination by telomeric P elements inserted in heterochromatin. Mol Genet Genomics. 2004;272:336–43.

    Article  CAS  PubMed  Google Scholar 

  28. Srivastav SP, Rahman R, Ma Q, Pierre J, Bandyopadhyay S, Lau NC. Har-P, a short P-element variant, weaponizes P-transposase to severely impair Drosophila development. Elife. 2019;8:e49948.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Marin L, Lehmann M, Nouaud D, Izaabel H, Anxolabéhère D, Ronsseray S. P-element repression in Drosophila melanogaster by a naturally occurring defective telomeric P copy. Genetics. 2000;155:1841–54.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Ronsseray S, Lehmann M, Anxolabéhère D. The maternally inherited regulation of P elements in Drosophila melanogaster can be elicited by two P copies at cytological site 1A on the X chromosome. Genetics. 1991;129:501–12.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Ronsseray S, Lehmann M, Nouaud D, Anxolabéhère D. The regulatory properties of autonomous subtelomeric P elements are sensitive to a suppressor of variegation in Drosophila melanogaster. Genetics. 1996;143:1663–74.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Roche SE, Rio DC. Trans-silencing by P elements inserted in subtelomeric heterochromatin involves the Drosophila Polycomb group gene, Enhancer of zeste. Genetics. 1998;149:1839–55.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Ronsseray S, Marin L, Lehmann M, Anxolabéhère D. Repression of hybrid dysgenesis in Drosophila melanogaster by combinations of telomeric P-element reporters and naturally occurring P elements. Genetics. 1998;149:1857–66.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Josse T, Teysset L, Todeschini A-L, Sidor CM, Anxolabéhère D, Ronsseray S. Telomeric trans-silencing: an epigenetic repression combining RNA silencing and heterochromatin formation. PLoS Genet. 2007;3:1633–43.

    Article  CAS  PubMed  Google Scholar 

  35. Marie PP, Ronsseray S, Boivin A. From embryo to adult: piRNA-mediated silencing throughout germline development in Drosophila. G3 (Bethesda). 2017;G3(7):505–16.

    Article  Google Scholar 

  36. Muerdter F, Olovnikov I, Molaro A, Rozhkov NV, Czech B, Gordon A, et al. Production of artificial piRNAs in flies and mice. RNA. 2012;18:42–52.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Poyhonen M, de Vanssay A, Delmarre V, Hermant C, Todeschini AL, Teysset L, et al. Homology-dependent silencing by an exogenous sequence in the Drosophila germline. G3 (Bethesda). 2012;G3(2):331–8.

    Article  Google Scholar 

  38. Todeschini A-L, Teysset L, Delmarre V, Ronsseray S. The epigenetic trans-silencing effect in Drosophila Involves maternally-transmitted small RNAs whose production depends on the piRNA pathway and HP1. PLoS One. 2010;5:e11032.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Casier K, Delmarre V, Gueguen N, Hermant C, Viode E, Vaury C, et al. Environmentally-induced epigenetic conversion of a piRNA cluster. Elife. 2019;8:e39842.

    Article  PubMed  PubMed Central  Google Scholar 

  40. Duc C, Yoth M, Jensen S, Mouniée N, Bergman CM, Vaury C, et al. Trapping a somatic endogenous retrovirus into a germline piRNA cluster immunizes the germline against further invasion. Genome Biol. 2019;20:127.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Kofler R. Dynamics of transposable element invasions with piRNA clusters. Mol Biol Evol. 2019;36:1457–72.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Zanni V, Eymery A, Coiffet M, Zytnicki M, Luyten I, Quesneville H, et al. Distribution, evolution, and diversity of retrotransposons at the flamenco locus reflect the regulatory properties of piRNA clusters. Proc Natl Acad Sci USA. 2013;110:19842–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Olovnikov I, Ryazansky S, Shpiz S, Lavrov S, Abramov Y, Vaury C, et al. De novo piRNA cluster formation in the Drosophila germ line triggered by transgenes containing a transcribed transposon fragment. Nucleic Acids Res. 2013;41:5757–68.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Radion E, Sokolova O, Ryazansky S, Komarov P, Abramov Y, Kalmykova A. The integrity of piRNA clusters is abolished by insulators in the Drosophila germline. Genes. 2019;10:209.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Josse T, Maurel-Zaffran C, de Vanssay A, Teysset L, Todeschini A-L, Delmarre V, et al. Telomeric trans-silencing in Drosophila melanogaster: tissue specificity, development and functional interactions between non-homologous telomeres. PLoS One. 2008;3:e3249.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Klattenhoff C, Xi H, Li C, Lee S, Xu J, Khurana JS, et al. The Drosophila HP1 homolog Rhino is required for transposon silencing and piRNA production by dual-strand clusters. Cell. 2009;138:1137–49.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. ElMaghraby MF, Andersen PR, Pühringer F, Hohmann U, Meixner K, Lendl T, et al. A heterochromatin-specific RNA export pathway facilitates piRNA production. Cell. 2019;178:964-979.e20.

    Article  CAS  PubMed  Google Scholar 

  48. Anzelon TA, Chowdhury S, Hughes SM, Xiao Y, Lander GC, MacRae IJ. Structural basis for piRNA targeting. Nature. 2021;597:285–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. Mohn F, Handler D, Brennecke J. piRNA-guided slicing specifies transcripts for Zucchini-dependent, phased piRNA biogenesis. Science. 2015;348:812–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  50. O’Kane CJ, Gehring WJ. Detection in situ of genomic regulatory elements in Drosophila. Proc Natl Acad Sci USA. 1987;84:9123–7.

    Article  PubMed  PubMed Central  Google Scholar 

  51. Luo Y, He P, Kanrar N, Toth KF, Aravin AA. Maternally inherited siRNAs initiate piRNA cluster formation. preprint in bioRxiv. 2022.

  52. Jangam D, Feschotte C, Betrán E. Transposable element domestication as an adaptation to evolutionary conflicts. Trends Genet. 2017;33:817–31.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Miller WJ, McDonald JF, Nouaud D, Anxolabéhère D. Molecular domestication–more than a sporadic episode in evolution. Genetica. 1999;107:197–207.

    Article  CAS  PubMed  Google Scholar 

  54. Gao C, Xiao M, Ren X, Hayward A, Yin J, Wu L, et al. Characterization and functional annotation of nested transposable elements in eukaryotic genomes. Genomics. 2012;100:222–30.

    Article  CAS  PubMed  Google Scholar 

  55. Bergman CM, Quesneville H, Anxolabéhère D, Ashburner M. Recurrent insertion and duplication generate networks of transposable element sequences in the Drosophila melanogaster genome. Genome Biol. 2006;7:R112.

    Article  PubMed  PubMed Central  Google Scholar 

  56. Sato K, Siomi MC. The piRNA pathway in Drosophila ovarian germ and somatic cells. Proc Jpn Acad Ser B Phys Biol Sci. 2020;96:32–42.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  57. Shpiz S, Ryazansky S, Olovnikov I, Abramov Y, Kalmykova A. Euchromatic transposon insertions trigger production of novel Pi- and endo-siRNAs at the target sites in the Drosophila germline. PLoS Genet. 2014;10:e1004138.

    Article  PubMed  PubMed Central  Google Scholar 

  58. Radion E, Ryazansky S, Akulenko N, Rozovsky Y, Kwon D, Morgunova V, et al. Telomeric retrotransposon HeT-A contains a bidirectional promoter that initiates divergent transcription of piRNA precursors in Drosophila germline. J Mol Biol. 2017;429:3280–9.

    Article  CAS  PubMed  Google Scholar 

  59. Rozhkov NV, Schostak NG, Zelentsova ES, Yushenova IA, Zatsepina OG, Evgen’ev MB. Evolution and dynamics of small RNA response to a retroelement invasion in Drosophila. Mol Biol Evol. 2013;30:397–408.

    Article  CAS  PubMed  Google Scholar 

  60. Kofler R, Senti K-A, Nolte V, Tobler R, Schlötterer C. Molecular dissection of a natural transposable element invasion. Genome Res. 2018;28:824–35.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  61. Kelleher ES, Azevedo RBR, Zheng Y. The evolution of small-RNA-mediated silencing of an invading transposable element. Genome Biol Evol. 2018;10:3038–57.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. Coen D, Lemaitre B, Delattre M, Quesneville H, Ronsseray S, Simonelig M, et al. DrosophilaP element: transposition, regulation and evolution. Genetica. 1994;93:61–78.

    Article  CAS  PubMed  Google Scholar 

  63. Gebert D, Neubert LK, Lloyd C, Gui J, Lehmann R, Teixeira FK. Large Drosophila germline piRNA clusters are evolutionarily labile and dispensable for transposon regulation. Mol Cell. 2021;81:3965-3978.e5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  64. Akulenko N, Ryazansky S, Morgunova V, Komarov PA, Olovnikov I, Vaury C, et al. Transcriptional and chromatin changes accompanying de novo formation of transgenic piRNA clusters. RNA. 2018;24:574–84.

    Article  PubMed  PubMed Central  Google Scholar 

  65. Kent TV, Uzunović J, Wright SI. Coevolution between transposable elements and recombination. Philos Trans R Soc Lond B Biol Sci. 2017;372(1736):20160458.

    Article  PubMed  PubMed Central  Google Scholar 

  66. Rizzon C, Marais G, Gouy M, Biémont C. Recombination rate and the distribution of transposable elements in the Drosophila melanogaster genome. Genome Res. 2002;12:400–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  67. Wierzbicki F, Kofler R, Signor S. Evolutionary dynamics of piRNA clusters in Drosophila. Molecular Ecology. 2023;32:1306–22.

    Article  CAS  PubMed  Google Scholar 

  68. Baumgartner L, Handler D, Platzer SW, Yu C, Duchek P, Brennecke J. The Drosophila ZAD zinc finger protein Kipferl guides Rhino to piRNA clusters. eLife. 2022;11:e80067.

    Article  PubMed  PubMed Central  Google Scholar 

  69. Dorer DR, Henikoff S. Expansions of transgene repeats cause heterochromatin formation and gene silencing in Drosophila. Cell. 1994;77:993–1002.

    Article  CAS  PubMed  Google Scholar 

  70. Dufourt J, Dennis C, Boivin A, Gueguen N, Théron E, Goriaux C, et al. Spatio-temporal requirements for transposable element piRNA-mediated silencing during Drosophila oogenesis. Nucleic Acids Res. 2014;42:2512–24.

    Article  CAS  PubMed  Google Scholar 

  71. Lemaitre B, Ronsseray S, Coen D. Maternal repression of the P element promoter in the germline of Drosophila melanogaster: a model for the P cytotype. Genetics. 1993;135:149–60.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  72. Rørth P. Gal4 in the Drosophila female germline. Mech Dev. 1998;78:113–8.

    Article  PubMed  Google Scholar 

  73. Ni J-Q, Zhou R, Czech B, Liu L-P, Holderbaum L, Yang-Zhou D, et al. A genome-scale shRNA resource for transgenic RNAi in Drosophila. Nat Methods. 2011;8:405–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  74. Lacoste J, Soula H, Burg A, Audibert A, Darnat P, Gho M, et al. A neural progenitor mitotic wave is required for asynchronous axon outgrowth and morphology. eLife. 2022;11:e75746.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  75. Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 2009;10:R25.

    Article  PubMed  PubMed Central  Google Scholar 

  76. Paramutation in Drosophila requires both nuclear and cytoplasmic actors of the piRNA pathway and induces cis-spreading of piRNA production. NCBI; 2015. Accessed 24 Oct 2015.

  77. Small RNA from D. melanogaster female gonads. NCBI; 2017. Accessed 1 May 2017.

  78. Modelling early germline immunization after horizontal transfer of transposable elements reveals internal piRNA cluster heterogeneity. NCBI; 2022. Accessed 7 Nov 2022.

  79. Ovarian small RNA library for w1118 strain. NCBI; 2023. Accessed 16 Feb 2023.

Download references


We thank the Bloomington Stock Center and the TRiP at Harvard Medical School (NIH/NIGMS R01-GM084947) for providing fly stocks, for providing valuable databases, and the ARTbio/Bioinformatics analyses platform at the IBPS led by C. Antoniewski for the development of valuable tools ( We thank Sandra Duharcourt, Ana Maria Vallés, and all members of the TErBio team for the helpful discussions, Ritha Zamy for the technical assistance, and Camille Lespinet for the Drosophila drawings.


This work was supported by fellowships from the Ministère de l’Enseignement Supérieur et de la Recherche to A.A.L., K.C., and E.V.; by a PhD fellowship from the i-Bio Initiative for interdisciplinary research to Z.Z.; by a fellowships from the Ligue Nationale Contre le Cancer to K.C.; and by grants from the Association de la Recherche contre le Cancer (Fondation ARC PJA20131200470 to S.R., PJA 20191209395 to L.T.), the Fondation pour la Recherche Médicale (FRM DEP20131128532 to S.R.), the Association Nationale de la Recherche (ANR projects “plastisipi” to S.R. and BiopiC ANR-21-CE12-0022-01 to C.C.), and the University Pierre et Marie Curie/Sorbonne Université (Emergence EME1223) to L.T. K.C. Thank you to the Réseau André Picard and the Société Française de Biologie du Développement for their supports on travel fellowship grants.

Author information

Authors and Affiliations



L.T. and S.R. conceived the project and designed the research. A.A.L., K.C., Z.Z., A.B., E.V., V.D., and L.T. performed the experiments. A.A.L., K.C., C.C., A.B., S.R., and L.T. analyzed the data. L.T. wrote the original manuscript and submitted the sequencing data. A.A.L., C.C., K.C., A.B., and S.R. revised the manuscript. S.R., C.C., and L.T. raised funding for this project. The authors read and approved the final manuscript.

Corresponding author

Correspondence to Laure Teysset.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Fig. S1.

Mapping of complementary maternally inherited piRNAs on a piRNA cluster locus. Fig. S2. Parental strains and experimental schemes. Fig. S3. Concomitant conversion of all the regions of the P transgene in the MI H and PI B sublines. Fig. S4. Conversion of cluster1A in the Δ-1Aw1118 background. Fig. S5. Conversion of the P transgene inserted in the autosomal cluster100F. Fig. S6. P-derived sequences of P are able to repress ovarian expression of pRFP. Fig. S7. Ovarian lacZ silencing induced by P is impaired by germline knockdown of genes involved in piRNA biology. Fig. S8. Crosses used for the conversion of the P transgenes by P or P. Fig. S9. Study of composition of converted sequences. Fig. S10. No cis-conversion of flanking regions of the P transgene clusters or of cluster 1A. Fig. S11. No trans-conversion of endogenous homologous sequences. Fig. S12. RT-qPCR experiments of P and T3 in P-1152. Fig. S13. siRNAs and piRNAs abundance during conversion of cluster 1A. Fig. S14. Comparison of methods for small RNA libraries normalizations.

Additional file 2: Table S1.

List of Drosophila strains. Table S2. Ovarian lacZ repression for the maternally and paternally inherited Ptransgenes. Table S3. Normalized number of unique mappers of 23 to 29 reads in the 3 sublines sequenced. Table S4. Annotation of small RNA libraries. Table S5. Percentage of dinucleotides content. Table S6. Number of 21 ntand 23-29 ntreads with 0 or 3 mismatches mapping to P, T3, white and plasmid. Table S7. Primers used in this study. Table S8. P-values calculated for the H3K9me3 ovarian enrichment by ChIP-qPCR.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Asif-Laidin, A., Casier, K., Ziriat, Z. et al. Modeling early germline immunization after horizontal transfer of transposable elements reveals internal piRNA cluster heterogeneity. BMC Biol 21, 117 (2023).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: