Epigenetic Silencing of P-Element Reporter Genes Induced by Transcriptionally Active Domains of Constitutive Heterochromatin in Drosophila melanogaster

Reporter genes inserted via P-element integration into different locations of the Drosophila melanogaster genome have been routinely used to monitor the functional state of chromatin domains. It is commonly thought that P-element-derived reporter genes are subjected to position effect variegation (PEV) when transposed into constitutive heterochromatin because they acquire heterochromatin-like epigenetic modifications that promote silencing. However, sequencing and annotation of the D. melanogaster genome have shown that constitutive heterochromatin is a genetically and molecularly heterogeneous compartment. In fact, in addition to repetitive DNAs, it harbors hundreds of functional genes, together accounting for a significant fraction of its entire genomic territory. Notably, most of these genes are actively transcribed in different developmental stages and tissues, irrespective of their location in heterochromatin. An open question in the genetic and molecular studies on PEV in D. melanogaster is whether functional heterochromatin domains, i.e., heterochromatin harboring active genes, are able to silence reporter genes therein transposed or, on the contrary, can drive their expression. In this work, we provide experimental evidence showing that strong silencing of the Pw+ reporters is induced even when they are integrated within or near actively transcribed loci in the pericentric regions of chromosome 2. Interestingly, some Pw+ reporters were found insensitive to the action of a known PEV suppressor. Two of them are inserted within Yeti, a gene expressed in the deep heterochromatin of chromosome 2 which carries active chromatin marks. The difference sensitivity to suppressors-exhibited Pw+ reporters supports the view that different epigenetic regulators or mechanisms control different regions of heterochromatin. Together, our results suggest that there may be more complexity regarding the molecular mechanisms underlying PEV.

Based on their structure and transposition mechanism, eukaryotic TEs are classified into two main classes. Class I contains retrotransposons that are mobilized via RNA transposition intermediates, whereas class II elements transpose either using a "cut and paste" mechanism (Subclass I), rolling-circle replication [6], or single strand excision followed by extrachromosomal replication [7]. TEs have shaped eukaryotic genomes [8] including regions of constitutive heterochromatin where a build-up of TEs has been documented in many evolutionarily distant organisms [9,10].
Constitutive heterochromatin has been originally defined based on cytological features, i.e., the compact state during all phases of the cell cycle and C-banding staining. In addition, low gene density, low rate of meiotic recombination, high content of repetitive DNAs and specific epigenetic signatures are hallmarks of the constitutive heterochromatin [11]. However, the lack of functional genes is no more considered a general feature of constitutive heterochromatin of D. melanogaster and other species [11]. Indeed, genome sequencing and annotation have shown that this peculiar genomic compartment contains hundreds of transcriptionally active genes, both unique and repetitive [12][13][14], which together account for a large fraction of the genomic territory of D. melanogaster constitutive heterochromatin [11].
Among the Drosophila TEs that have been extensively characterized both at the genetic and molecular levels, P and I elements are naturally occurring transposons which have been found to insert themselves into the constitutive heterochromatin of D. melanogaster and eventually target the genes therein located [15][16][17][18][19][20][21][22].
Beside the above applications, engineered P-elements represent useful tools to monitor the functional state of diverse chromatin domains of D. melanogaster genome. When inserted into regions of constitutive heterochromatin, P-associated reporters undergo strong silencing or expressed following a mosaic pattern, in which some cells fully or partially display the reporter phenotype (known as variegated phenotype), while other cells do not [33][34][35]. The variegated phenotypes produced by transgene insertions into constitutive heterochromatin are very similar to those usually ascribed to position effect variegation (PEV) caused by chromosome rearrangements that relocate a euchromatic gene nearby to constitutive heterochromatin [36]. It is generally believed that the PEV phenotypes displayed by a variety of P-element-reporter genes reflect the silenced state of a given heterochromatin region where the elements are inserted. The variegated phenotype of P-reporter genes inserted in constitutive heterochromatin can be suppressed by known genetic modifiers of PEV [21] and by the Y chromosome, a known suppressor of PEV [37]. P-element-based vectors containing the mini-white gene driven by an hsp70 promoter were also used to characterize the structure of chromatin domains of the D. melanogaster fourth chromosome [38,39]. The results of these works have shown that this chromosome contains heterochromatic domains showing a variegating phenotype alternating with euchromatic domains showing a fully pigmented eye phenotype.
The resistance to exogenous nuclease digestion [40] and to methylation [41] exhibited by P-element reporters inserted within constitutive heterochromatin and their relatively ordered nucleosome arrays [42] together suggested that heterochromatic silencing is due to the loss of sites for binding of transcription factors and/or RNA polymerase.
In conclusion, based on the aforementioned results, P-element reporter genes have been considered a useful tool to investigate the genetic and molecular bases underlying the PEV phenomenon.
Histone modifications have been shown to play a role in the occurrence of epigenetic silencing and thus contribute to PEV [36]. Notably, actively transcribed genes in constitutive heterochromatin must be accessible to RNA pol II and cis-acting transcriptional regulators, although they have regulatory requirements different from those of euchromatic genes [43]. In particular, combinations of negative and active histone modification marks, together with the contribution of key epigenetic regulators, such as the HP1 and SU(VAR)3-9 proteins, are likely to contribute to establishing gene expression in constitutive heterochromatin [11,[44][45][46][47][48][49][50][51][52].
Here, we asked whether a given P-element reporter gene inserted into constitutive heterochromatin, within or nearby a transcribed gene, would "sense" the active chromatin state of the surrounding domain, with its expression being driven by the regulatory ele-Genes 2023, 14, 12 3 of 17 ments of that region or, on the contrary, would it undergo PEV. To answer this question, we examined 12 lines each containing a mini-white reporter gene carried by a P-element (from here indicated as Pw + reporters) cytologically and molecularly mapped to the pericentric heterochromatin of chromosome 2 (Table 1 [21,35,53]).
Our results provide new hints on the mechanisms of epigenetic silencing induced by constitutive heterochromatic in D. melanogaster.

Drosophila Strains
Fly cultures and crosses were carried out at 25 • C in standard cornmeal yeast medium. The KV lines, each carrying a single SUPor-P element [35], were gifts of Robert Levis. The Su(var)205 2 and aubergne QC42 mutants were gifts of Sarah Elgin, while the Su(var)30 71 and Su(var)3-9 1 were gifts of Gunter Reuter. The LP1 mutant was generated by Messina et al. [53].

Mapping of the P White+ (Pw + ) Reporters
The flanking sequences of the KV lines were obtained by inverse PCR (done by Robert Levis) and their locations are reported in FlyBase. The mapping of 19.74.3 and LP1 lines was reported by Messina et al. [53].

Complementation Test with Df(2Rh)MS41-10
Pw + /SM1, Cy females were crossed to Df(2Rh)MS41-10/SM1, Cy males, and the progeny was scored for the presence of Cy + and Cy flies. The expected ratio between Cy and Cy + among the offspring of these crosses was 2:1 because Cy is homozygous lethal. The absence of the Cy + progeny indicated that a given Pw + insert was lethal over the deficiency. KV00363 and LP1 were found to be lethal over Df(2Rh)MS41-10 (all progeny showing the Cy phenotype), while for the other inserts the ratio between Cy and Cy + phenotypes was close to 2:1. For each complementation test, over 100 adults were counted.

Drosophila Strains
Fly cultures and crosses were carried out at 25 °C in standard cornme medium. The KV lines, each carrying a single SUPor-P element [35], were gifts o Levis. The Su(var)205 2 and aubergne QC42 mutants were gifts of Sarah Elgin, w Su(var) 30 71 and Su(var)3-9 1 were gifts of Gunter Reuter. The LP1 mutant was gene Messina et al. [53].

Mapping of the P white+ (Pw + ) Reporters
The flanking sequences of the KV lines were obtained by inverse PCR ( Robert Levis) and their locations are reported in FlyBase. The mapping of 19.74.3 lines was reported by Messina et al. [53].

Complementation Test with Df(2Rh)MS41-10
Pw + /SM1, Cy females were crossed to Df(2Rh)MS41-10/SM1, Cy males, progeny was scored for the presence of Cy + and Cy flies. The expected ratio bet and Cy + among the offspring of these crosses was 2:1 because Cy is homozygou The absence of the Cy + progeny indicated that a given Pw + insert was lethal deficiency. KV00363 and LP1 were found to be lethal over Df(2Rh)MS41-10 (all showing the Cy phenotype), while for the other inserts the ratio between Cy phenotypes was close to 2:1. For each complementation test, over 100 adults were

Eye Pigment Assays
The extraction of red eye pigment was performed according to Ephrussi and Herold [55], as described in Prozzillo et al. [56]. For each genotype, three replicate samples of 10 heads were performed. Absorbance at 480 nm was then measured using a 96-well plate in a VICTOR Multilabel Plate Reader spectrophotometer (PerkinElmer, Waltham, MA, USA). Photographs of representative adult fly eyes were taken using a Nikon SMZ745T stereoscopic microscope (Minato, Tokyo, Japan) equipped with a digital C-mount camera.

Distribution of Epigenetic Marks
The distribution of the epigenetic marks in the regions of constitutive heterochromatin is available in Genome Browser of FlyBase and was also retrieved from the ModEncode data.

Analysis of P-Element Insertions
First, we asked whether Pw + reporters inserted into actively transcribed domains of constitutive heterochromatin are subjected to PEV. To answer this question, among a collection of KV lines [21] showing strong mini-white silencing, we selected 10 lines with elements located within the pericentric portions of chromosome 2 ( Figure 2). These regions have been extensively characterized at both genetic and molecular levels [11,14,20,43,53,[57][58][59][60][61]. Each KV line carries a single SUPor-P element which has two reporter genes, the mini-white gene and the intron-less yellow gene, as well as two "suppressor of hairy wing" [Su(Hw)] binding sites flanking the mini-white gene [62]. This design allows a robust expression of the yellow reporter and, together with the presence of the Y chromosome, facilitates the recovery of SUPor-P pericentric insertions that would be otherwise missed due to the strong silencing of the mini-white reporter [21,63].
The analysis of the insertion sites showed that the SUPor-P element are located within or close to genes that are expressed (Table 1). In particular, as shown in Figure 3, according to FlyBase and FlyAtlas, the genes linked to the inserted elements are expressed during different developmental stages and in the eyes of both male and female. We also examined the LP1 line carrying a Pw + reporter into the Yeti gene and its progenitor insert of line 19.74.3, which maps about 2 kb downstream Yeti ([53], Table 1). In addition to the mini-white reporter, these elements contain Fab-7 insulator sequences characteristic of the bithorax complex [64].
All the P-insertions mapping to the right arm of chromosome 2 heterochromatin (2Rh) were tested in complementation analyses with Df(2Rh)MS41-10. This deletion lacks almost the entire mitotic heterochromatin of 2Rh, spanning from the deep region h40 to the very distal 46 [58,59,65], together with all the known genes, including rolled, CG40191, CG41265, Yeti, dpr21, Haspin and p120 ctn (See Figure 1, from Marsano et al. [11]). These results showed that, with the exception of LP1 and KV00363 (lethal alleles of Yeti), the analyzed P-inserts in 2Rh were fully viable over Df(2Rh)MS41-10. The insert in line KV00462 which maps to 2Lh was homozygous viable. Thus, it is conceivable that, while the expression of Yeti was clearly affected, the expression of the genes linked to the other insertions was not significantly perturbed. Indeed, in the LP1 homozygotes no Yeti transcripts were detected [53].
Together, from these analyses it emerged that, despite their location in expressed domains of constitutive heterochromatin, all the examined lines exhibited a strong silencing of the Pw + reporters. The males of some lines (KV00524, KV00158, KV00384, KV00376 and 19.74.3) exhibited a slightly stronger red-eye pigmentation compared to that shown by the females of the same line (Figure 2), an effect that could be due to the PEV suppression exerted by the Y chromosome [36,37].   All the P-insertions mapping to the right arm of chromosome 2 heterochromatin (2Rh) were tested in complementation analyses with Df(2Rh)MS41-10. This deletion lacks almost the entire mitotic heterochromatin of 2Rh, spanning from the deep region h40 to the very distal 46 [58,59,65], together with all the known genes, including rolled, CG40191, CG41265, Yeti, dpr21, Haspin and p120 ctn (See Figure 1, from Marsano et al. [11]). These results showed that, with the exception of LP1 and KV00363 (lethal alleles of Yeti), the analyzed P-inserts in 2Rh were fully viable over Df(2Rh)MS41-10. The insert in line KV00462 which maps to 2Lh was homozygous viable. Thus, it is conceivable that, while the expression of Yeti was clearly affected, the expression of the genes linked to the other insertions was not significantly perturbed. Indeed, in the LP1 homozygotes no Yeti transcripts were detected [53].
Together, from these analyses it emerged that, despite their location in expressed domains of constitutive heterochromatin, all the examined lines exhibited a strong silencing of the Pw + reporters. The males of some lines (KV00524, KV00158, KV00384, KV00376 and 19.74.3) exhibited a slightly stronger red-eye pigmentation compared to that shown by the females of the same line (Figure 2), an effect that could be due to the PEV suppression exerted by the Y chromosome [36,37].

Experiments with Genetic Modifiers of PEV
Loss-of-function mutations resulting in dominant suppression of PEV were shown to identify crucial epigenetic regulators, such as the HP1 protein [36]. Thus, first we investigated whether the variegation of Pw + reporters showed by the lines under investigation can be suppressed by Su(var)205 and Su(var)307, two dominant PEV suppressors [36]. To do this, we crossed w/w; Su(var)205 2 /SM6, Cy,S; +/+ or w/w; +/+; Su(var)3-7 1 /TM3, Sb females to males of 10 selected Pw + reporters (Figure 2) balanced over the SM1, Cy chromosome ( Figure 1; Section 2).
The results of this analysis are summarized in Figure 4. Su(var)205 2 clearly suppressed the variegation of Pw+ reporters in lines KV00158, KV00376, KV00369 and KV00171. However, this was not the case for the remaining five Pw + reporters in lines KV00249, KV00299, KV00363, KV00524 and LP1, which were quite insensitive to the suppression. Su(var)307 1 suppressed most Pw + reporters, although with a variable efficiency. However, the reporters of LP1 and KV00299 lines showed only a slight suppression effect, while the one of KV00363 was clearly insensitive.
Moreover, mutations in piwi, aubergine or homeless genes, encoding components of the piRNA pathway [66,67], also suppress the silencing of Pw + inserted in heterochromatin as a result of reduction of H3 Lys9 methylation and delocalization of HP1. Thus, we also tested the effect of the aubergine QC42 mutant allele on the variegation of the Pw + reporters under investigation (see Section 2 for the scheme of crosses). As shown in Figure 4, the trend of the effect was very similar to that found with Su(var)205 2 in that the Pw + reporters of KV00249, KV00299, KV00363 and LP1 lines were insensitive to suppression, with the exception of KV00524.  The product of the wild-type Su(var)205 gene, the heterochromatin-associated HP1 protein, interacts with the histone methyltransferase (HMTase) encoded by the wild-type Su(var)3-9 gene. The SU(VAR)3-9 methyltransferase selectively methylates histone H3 at lysine 9 (H3-K9), and it is generally accepted that H3K9me stabilizes the binding of HP1a to chromatin [68][69][70]. The Su(var)3-9 mutant alleles are known to be dominant modifiers of PEV [36]. We then tested the effect of the Su(var)3-9 1 mutation on the variegation of Pw + reporters from KV00249, KV00299, KV00363, LP1 and 19.74.3 lines. As shown in Figure 5, the variegation of Pw + reporters in KV00249, KV00299, KV00363 and LP1 lines was insensitive to suppression by Su(var)3-9 1 , whereas the Pw + variegation in the 19.74.3 was efficiently suppressed. In conclusion, using different PEV suppressors (Su(var)205 2 , Su(var)3-9 1 and aubergine QC4 ), we obtained comparable results with the same Pw + reporters (Figures 4 and 5). was efficiently suppressed, while that of KV00363, LP1, KV00299 and KV00249 was unaffected. For the scheme of crosses see Figure 1 and Materials and Methods.

Discussion
It is generally believed that PEV is induced by transcriptionally inactive constitutive heterochromatin. The aim of this work was to further investigate this aspect by studying the heterochromatin-induced silencing on a number of Pw + reporters inserted into different positions of the pericentromeric regions of chromosome 2, within or nearby a transcribed gene (Table 1; Figures 3 and 6). We found that all the examined reporters subjected to strong PEV (Figure 2) map within or close to genes/domains that support an appreciable level of transcription during development and in the eye (Figure 3). Moreover, some reporters were suppressed by classical dominant PEV suppressors encoding epigenetic factors/regulators required for heterochromatin establishment, while other reporters in lines KV00249, KV00299, KV00363 and LP1 were less sensitive or even insensitive to the action of the suppressors (Figures 4-6). In general, there appears to be no significant correlation between the mini-white expression levels for each Pw + reporter line and the transcription levels of the heterochromatic gene(s) linked to it. For example, CG40191 is on average more expressed than dpr21, and yet the respective inserts have the same behavior towards suppression.
Although the heterogeneity of the genetic background might partially contribute to explaining the different sensitivity to epigenetic regulators tested, these results suggest that the chromatin state characteristic of transcriptionally active heterochromatin may not be sufficient to guarantee the proper expression of a given reporter gene inserted within or nearby. In other words, the regulatory requirements for the expression of the autochthonous genes in heterochromatin seem to be different, or even antipodal, from those needed for adventitious DNA sequences, such as the Pw + reporters. However, the observation that the silencing of some reporters is abolished by the tested suppressors also suggests that the product of wild-type alleles of some suppressors, on one hand, can induce the silencing of the reporters, but, on the other hand, may be required for, or is compatible with, the expression of the genes located in pericentric heterochromatin. The effect of the product of wild-type alleles of suppressors could be performed directly or indirectly.
Indeed, the expression of light and rolled heterochromatic genes of chromosome 2 was found to be affected when moved away from their native location to euchromatin by chromosome rearrangements [70,71]. Moreover, genetic and molecular studies have shown that the proper expression of light, rolled, and other heterochromatic genes such as Rpl15 and Dbp80 depends on the HP1a protein [72][73][74][75]. In accordance, large-scale mapping experiments carried out in Drosophila Kc cells have shown that HP1a and SU(VAR)3-9 are associated with pericentric genes, such as light and rolled genes [44,46,47].
Together, the available findings support the idea that histone marks, together with HP1a and other epigenetic regulators, are involved to ensure the expression of genes located in constitutive heterochromatin. However, some apparently contradictory results [44,51] indicated that differences in the epigenetic regulation of heterochromatic genes may exist between in vivo and in vitro systems.
The insensitivity to suppressors showed by Pw + reporters in KV363 and LP1 lines is intriguing for several reasons. Firstly, Yeti is an efficiently expressed gene in the heterochromatin of chromosome 2 ( Figure 2 [11,53,[76][77][78][79]) and carries active chromatin marks (Table 1). Secondly, the P-insertion of the 19.74.3 line, the progenitor of LP1, maps roughly 2 kb downstream the 5 of Yeti (Figure 6), but its silencing is efficiently suppressed by Su(var)205 2 , Su(var)307 1 , Su(var)309 2 and aubergne QC42 (Figures 4 and 5). Thirdly, Su(var)205, Su(var)309 and Su(var)307 are among the most general and effective PEV suppressors and were already found to efficiently erase the silencing of many pericentric reporters [36,80]. Finally, the distribution of HP1 and SU(VAR)309 proteins along the pericentric regions was found to be enriched in the heterochromatin of 2R where Yeti is located [44,46,47].  The molecular bases underpinning insensitivity to suppressors exhibited by Pw + reporters in KV363 and LP1 lines are currently elusive. A trivial explanation for the apparently strong silencing could be that the Pw + element in these lines is structurally defective so that the expression of the mini-white is compromised. However, we found that in both lines the element can still transpose in the presence of the Delta-2-3 element, a known source of the P-transposase, giving rise to de novo fully pigmented red-eye insertions (not show). This result excludes the occurrence of significant DNA sequence alterations in the Pw+ element of both KV363 and LP1 lines.
It is possible that the chromatin state of the Yeti gene is regulated by epigenetic regulators/mechanisms different from those controlled by the wild-type alleles of the tested PEV suppressors.
We did not find recurrent repeated sequences flanking the reporters, suggesting the absence of specific relationships between location and sensitivity to suppression. However, two non-coding RNAs, CR44042 and CR44043, were associated with Yeti and l(2)41Ab genes ( Figure 6), respectively, with the Pw + reporter of KV00299 being inserted within CR44043. It could be possible that the transcription of CR44042 and CR44043 may interfere somehow with that of the reporters.
A consistent body of experimental evidence suggested that nascent RNA may play a direct role in transcription and chromatin regulation, in that it may function as regulator of its own expression, giving rise to positive-feedback loops in which active gene expression states can be maintained [81]. This model could explain the resistance to all PEV suppressor exhibited by the Pw + reporters in LP1 and KV00363 lines. We hypothesize that in wild-type conditions the Yeti-encoded RNA interacts with its own gene at the level of DNA/chromatin, acting as an epigenetic insulator or as a positive effector, giving rise to an active chromatin state which allows a stable expression of Yeti. However, following a P-insertional mutation, a consequent lack of functional Yeti transcripts (or a low level of transcription) would shutdown the expression of the gene disrupting its active chromatin state with the acquisition of a closed chromatin structure. As a consequence, the Pw + reporters inserted within Yeti would become silenced (Figure 7). Such positive RNA feedback mechanism, if existing, should be insensitive to the action of classical PEV suppressors.

Conclusions
In conclusion, we have shown that transcriptionally active domains of constitutive heterochromatin can induce epigenetic silencing of Pw + reporters. Whatever the genetic bases of this phenomenon are, our results represent a remarkable paradox suggesting that there may be more complexity regarding the molecular mechanisms underlying PEV.