Mutations in Dref, Trf2, and Ken affect the telomeric retrotransposons HeT-A and TART

We investigated whether mutations in Dref, Trf2, and Ken affected the expression of the telomeric retrotransposons HeT-A and TART. As mentioned, Pzg has been found in the same protein complexes as DREF and KEN. TRF2 (TATA-box-binding protein (TBP) related factor 2) has been found in a complex with NURF and DREF. It was demonstrated that the recognition of promoters of replication-related genes by TRF2 depended on the presence of DREF that directly binds to specific DNA motifs [13]. If the function of Pzg at the telomeres depends on the action of the DREF complex, we would expect similar effects of mutant alleles of Trf2 and Dref over the telomere retrotransposons. Therefore, the mutant alleles included in this study are Dref ^KG0994 , Trf ²⁶⁰⁰⁷¹ , and Ken ¹ mutants.

We have previously observed that null mutants of pzg do not affect the expression of the HeT-A retrotransposon [6], although the hypomorph mutant Z4 ^7.1 affects its expression and genomic copy number [6]. We included two different pzg mutant alleles in these studies, the pzg hypomorph (Z4 ^7.1 ) and a pzg null mutant (pzg ⁶⁶ ), in order to investigate whether they affected the expression of the TART retrotransposon. For all mutants, we analyzed the levels of HeT-A and TART mRNA by quantitative real-time PCR. Because the number of copies of the telomeric retrotransposons varies among stocks, we normalized the HeT-A and TART mRNA data by the number of copies of the retrotransposon in each stock, to obtain the level of expression of each copy of HeT-A and TART. Moreover, the presence of a higher copy number of one of the telomere retrotransposons in a mutant allele could be indicative of an increased expression and rate of terminal transposition.

To determine the number of copies of HeT-A and TART in each stock, we extracted genomic DNA from third-instar larvae without salivary glands. Dref ^KG0994 , Trf ²⁶⁰⁰⁷¹ , and Ken ¹ mutant alleles did not show differences in HeT-A copy number (Figure 1A), but in the Z4 ^7.1 /Z4 ^7.1 hypomorph mutant, an increase in HeT-A copy number was observed, as we have previously demonstrated [6]. Results for the TART element are different; Dref ^KG0994 , Trf ²⁶⁰⁰⁷ , ¹ and the Z4 ^7.1 /Z4 ^7.1 alleles show a significant increase of TART copies in their genomes, while Ken ¹ and the null allele of pzg, pzg ⁶⁶ , do not show a significant change in the number of copies of TART in their genomes (Figure 1A).

Next, we analyzed the mRNA levels of HeT-A and TART in the same mutants. To obtain the expression data we extracted mRNA from whole third-instar larvae and analyzed them by quantitative real-time PCR. Unlike HeT-A, the TART element does not show an increase in transcription in any of the pzg mutant alleles after normalizing the data. On the other hand, in accordance with the genomic copy number obtained, a significant increase in TART expression was observed for the Dref ^KG0994 and the Trf ²⁶⁰⁰⁷¹ alleles. These same alleles show a similar behavior for the expression of the HeT-A retrotransposon although less accentuated than for its telomere partner (Figure 1B,C). These observations are in accordance with a possible link between the role of Pzg and the DREF/TRF2 protein complexes in the control of the telomere retrotransposons. Finally, the Ken ¹ allele also shows an increase in TART expression although no effect in HeT-A expression was observed (Figure 1B,C). These results indicate that Dref ^KG0994 , Trf ²⁶⁰⁰⁷¹ , and Ken ¹ mutant alleles affect gene expression differently at the telomere in Drosophila.

The TART promoter contains DREF binding sequences

After showing that the DREF/TRF2 complex had a role in controlling the expression of the telomeric retrotransposons, HeT-A, TART, and TAHRE, we searched for the presence of the DREF binding sequence (5′-TATCGATA) along the sequence of the telomeric retrotransposons [16]. We did not find the DREF motif in the sequence of HeT-A and TAHRE. However, we were able to identify two DREF binding sequences in TART; one in the 5′ UTR around 170 bp downstream of the transcription start site and another around 600 bp upstream the end of the 3′ UTR. To investigate whether these binding sites were conserved among the different TART subfamilies we performed a nucleotide sequence alignment using ClustalW software with the available copies in the databases (Figure 2). The sequence alignment revealed that the DREF binding site at the 5′ UTR was highly conserved among all TART subfamilies (Figure 2A) while the 3′ UTR sequence was only present in the TART A subfamily (Figure 2B). The DREF binding site at the TART 5′ UTR is located at the TART promoter that drives sense transcription, while the 3′ UTR binding site lies around the area where a putative TART anti-sense promoter has been proposed [17].

HeT-A, TART and TAHRE contain KEN binding sites

Similarly, we searched for KEN binding sites (5′-GAGAAAK, K = G/T) [18] in HeT-A, TART, and TAHRE. We found that this sequence was present in the three telomeric retrotransposons. HeT-A has a KEN binding site at the 5′ UTR that is conserved in three of the six analyzed sequences (Figure 3A). For this analysis, we only used the complete HeT-A sequences available in the databases (for more information, see [19]). TART has two KEN binding motifs inside the Gag coding sequence (ORF1); one is present in the three TART subfamilies, while the other is only present in the TART A subfamily (Figure 3B,C). Finally, in TAHRE, we found six copies of the KEN binding site (data not shown), three at the 5′ UTR, one at the reverse transcriptase domain (ORF 2), and two at the 3′ UTR. We could not perform a sequence multi-alignment analysis for TAHRE, since only one full sequence is available in databases [20].

Both these observations, the alterations in the levels of expression of the telomere retrotransposons in mutant alleles of Dref, Trf2, and Ken and the presence of both DREF and KEN motifs in the sequences of the different telomere retrotransposons, suggest that the complexes that these proteins nucleate are able to bind to the telomeric array, and therefore that they are susceptible to regulate the expression of the telomeric retrotransposons. In future, in vivo evidence for these bindings will confirm the regulatory role of the DREF/TRF2 and Ken protein complexes at the telomere array.

Dref and Trf2 control the expression of HeT-A and TART

Although a previous study had reported that no DREF binding sites were present at the HTT array [10], we identified two DREF binding sites in the TART element of D. melanogaster; one at the 5′UTR, highly conserved among TART subfamilies, and another at the 3′UTR, only conserved in the TART A subfamily (Figure 2). The high conservation of the 5′UTR DREF binding site, suggests the existence of selective pressure acting at this DNA binding motif. Interestingly, the TART A subfamily, the one that contains both DREF binding motifs, is the more abundant of the TART subfamilies in D. melanogaster[21]. In addition to the presence of the DREF DNA binding motif in the sequence of the TART element, the recruitment of the DREF/TRF2 complex at the telomeres has been recently demonstrated [10].

We have observed an increase of transcription of both HeT-A and TART retrotransposons in the Dref ^KG09294 and Trf ²⁶⁰⁰⁷¹ mutant alleles (Figure 1C). In agreement with the presence of DREF binding sites in the TART element, we find that the difference in TART transcription for the Dref ^KG09294 and Trf ²⁶⁰⁰⁷¹ mutant alleles is especially significant. Moreover, both the Dref ^KG09294 and Trf ²⁶⁰⁰⁷¹ mutant alleles contain a higher copy number of TART elements, suggesting that this increase in transcription is translated in an increase in terminal transposition. In a Dref ^KG09294 and Trf ²⁶⁰⁰⁷¹ mutant background, the intermingled nature of HeT-A and TART at the HTT array [22] could help to spread the possible change in chromatin structure caused by the lack of binding of DREF/TRF2 at the TART sequences towards the HeT-A promoter, provoking an indirect increase in the transcription of HeT-A despite not containing DREF binding sites (Figure 1C).

The increase in expression from the HTT array in the mutant alleles of Dref and Trf2 indicates that in a wild-type situation this protein complex behaves as a repressor of the telomere retrotransposons. A priori, we were expecting DREF/TRF2 to activate the telomere expression in synchrony with cell proliferation [11, 13]. Nevertheless, in accordance with our results, the DREF/TRF2 complex has also been reported to contribute to gene repression.

The DREF/TRF2 complex competes for DNA binding with the BEAF (boundary element-associated factor) because both DREF and BEAF have overlapping binding sites (BEAF binds to 5′-CGATA motifs) [23]. Moreover, it was shown that both proteins are able to target the same promoter [24]. Boundary elements have the ability to insulate a transgene from its chromosomal context by blocking enhancer-promoter interactions and heterochromatin spreading [25]. In D. melanogaster, BEAF associates preferentially with active transcribed genes [26]. The presence of BEAF at DREF-regulated genes leads to a decrease in the deposition of the heterochromatic mark H3K9me3, de-repressing those genes from the surrounding heterochromatin [24, 27, 28]. Our results are compatible with this scenario, in which equilibrium between the binding of both complexes would be necessary to obtain a fine-tuned regulation of the expression of the telomere retrotransposons. The binding of BEAF to the DREF binding sites at the TART promoter would protect the telomere retrotransposon from a repressive environment, and the binding of DREF would protect the genome from an excessive transcription, transposition, and telomere elongation. In a DREF mutant background, a major occupancy of BEAF at the HTT array would create an opening of the surrounding chromatin and, consequently, also increase the levels of HeT-A expression.

Our initial hypothesis was that Pzg could be exerting its role at the telomeres through its interaction with the DREF/TRF2 complex. In contrast with the increase in TART transcription observed in the Dref ^KG09294 and Trf ²⁶⁰⁰⁷¹ mutant alleles, the results obtained with two pzg mutant alleles, Z4 ^7.1 hypomorph and pzg ⁶⁶ null, show no change in the expression level of the TART retrotransposon. This result suggests that although the DREF/TRF2 has been related with Pzg in other situations, in this case they might be acting independently.

More chromatin modifiers have been related with the DREF/TRF2 complex, like the linker Histone H1, which is also involved in heterochromatin and transposable element gene expression as well as the NURF subunit ISWI (involved in ATP dependent nucleosome sliding) [29–32]. The loss of Trf2 could also affect the telomere chromatin through a secondary effect on these other heterochromatin components. Future studies involving additional mutant alleles of the mentioned chromatin components will help to elucidate the molecular mechanism by which the DREF/TRF2 complex affects the telomeres.

KEN is a repressor of TART

The JAK/STAT signaling pathway is responsible for the activation of the immune response genes [18, 33]. KEN is a well-defined repressor of the JAK/STAT pathway and competes with STAT for the binding to a target gene [33]. Recently, NURF301 and Pzg were also found to be recruited by KEN and repress the JAK/STAT pathway [14, 15]. When we analyzed the effect of the Ken mutant allele Ken ¹ in Drosophila telomeres, an increase in TART expression was observed but no effects on HeT-A expression were detected (Figure 1C). Accordingly, we found that the KEN binding sequences are not conserved in all the analyzed HeT-A sequences in opposition to the high conservation of at least one of the KEN binding sites in the TART sequence (Figure 3). A possible explanation could be that KEN might be involved in the recruitment of the chromatin-remodeling complex NURF to TART sequences repressing its transcription. In a Ken mutant background, the NURF complex is not recruited to the telomeres, leading to a relaxation of the chromatin. The fact that no effect on HeT-A expression is observed in Ken mutants suggests that the degree of chromatin relaxation is weaker than the one observed in Dref and Trf2 mutants. As with the DREF/TRF2 complex, we have found that the pzg alleles here analyzed do not show a change in transcription equivalent to the one observed in the Ken ¹ mutant. This result suggests a possible independent role of pzg and the Ken at Drosophila telomeres. We do not know whether the role of Ken controlling the expression of the TART element keeps any relationship with the JAK/STAT pathway, but it is possible that independent roles of Ken from the JAK/STAT pathway exist and that the control of telomere transcription could be the first one to be described.

Fly stocks and crosses

Fly stocks were maintained and crosses preformed at 25°C on standard Drosophila corn meal medium. w ¹¹¹⁸ and ry ⁵⁰⁶ were used as control, depending on the genetic background of each strain. w ¹¹¹⁸ ; Trf ²⁶⁰⁰⁷¹ /FM7c, ry ⁵⁰⁶ ; Dref ^KG09294 /CyO and ry ⁵⁰⁶ ; Ken ¹ /CyO were obtained from Bloomington Stock Center. w ¹¹¹⁸ ; Trf ²⁶⁰⁰⁷¹ /FM7c balancer was changed to FM7c-GFP to allow the selection of hemizygous males to perform the experiments. The three mutations correspond to P-element insertions inside the coding region. The hypomorph line Z4 ^7.1 /TM3Sb, lacking the promoter region, came from Harald Eggert and Harald Saumweber and the line pzg ⁶⁶ /TM6 was provided by Anja Nagel, this line is a null mutant that results in embryonic and early larval lethality.

Sequence alignments

The sequence alignments were carried out using ClustalW software. For the accession number of the HeT-A sequences used in the alignments see [19] and for the accession number of TART and TAHRE sequences see [21].

Genomic DNA extraction

Genomic DNA was extracted from adult flies to quantify the number of HeT-A and TART copies in each strain. Ten third-instar larvae without salivary glands were homogenized in 200 μl solution A (0.1 M Tris-HCl pH9.0, 0.1 M EDTA and 1% SDS) and incubated at 70°C for 30 min. 28 μl 8 M KAc were added and the samples incubated for 30 min on ice. Cell debris was harvested at maximum speed for 15 min at 4°C. The supernatant was transferred to a new tube and the DNA precipitated by adding 0.5 volumes isopropanol and centrifuging at 15.000 rpm for 5 min. Pelleted DNA was washed with 1 volume 70% ethanol and centrifuged. Finally, the DNA pellet was air-dried, and re-suspended in 50 μl 1× TE by rotating overnight at 4°C. After genomic DNA extraction, the number of copies was determined by quantitative real-time PCR using 2 ng of DNA per reaction.

Because certain endoreplication of the telomeric sequences exists in salivary glands, we depleted the samples from this tissue in order to have a copy number that would reflect the real copy number of the adult organism.

Primers used for TART amplification were, for TART_5′ UTR_F, GATAATAGCCGATAAGCCCGCCA, and for TART_5′ UTR_R, AAGACACAGCGGTTGATCGATATG. Primers used for HeT-A amplification were HeT-A_3′ UTR_F (CCCCGCCAGAAGGACGGA) and HeT-A_3′ UTR_R (TGTTGCAAGTGGCGCGCA). Primers used for actin amplification were Actin_F (GCGCCCTTACTCTTTCACCA) and Actin_R (ATGTCACGGACGATTTCACG).

RNA extraction and cDNA synthesis

Total RNA was isolated from ten whole third-instar larvae and extracted using RNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol. RNase Free DNase Set (Qiagen) was used to remove genomic DNA contaminations as follows: one on column during the extraction accordingly to manufacturer’s protocol, and two in solution for 2 hours at 37°C. RNA was cleaned by precipitation and its quality assessed using NanoDrop spectrophotometry.

One microgram of RNA was reverse transcribed into cDNA using Transcriptor First Strand cDNA Synthesis Kit (Roche) with oligo(dT) primers, and the expression of the different transcripts analyzed by quantitative real-time PCR. For each fly strain, two independent RNA extractions were prepared and analyzed independently three times. Primers used for TART amplification were, for TART_5′ UTR_F, GATAATAGCCGATAAGCCCGCCA, and for TART_5′ UTR_R, AAGACACAGCGGTTGATCGATATG. Primers used for HeT-A amplification were HeT-A_3′ UTR_F (CCCCGCCAGAAGGACGGA) and HeT-A_3′ UTR_R (TGTTGCAAGTGGCGCGCA). Primers used for actin amplification were Actin_F (GCGCCCTTACTCTTTCACCA) and Actin_R (ATGTCACGGACGATTTCACG).

We did not deplete the samples of salivary glands in the expression analysis, since the telomere retrotransposons are not expressed in this tissue [34].

Quantitative real-time PCR

Quantitative Real Time-PCR was performed to determine HeT-A and TART copy number and expression. The iQ5 Multicolor Real-Time PCR Detection System was used and the iQ™ SYBR^® Green Supermix (BioRad) was used to prepare the reactions. Relative levels of HeT-A and TART expression were determined using the threshold cycle and normalized to actin levels. Three independent experiments of two samples each strain were performed.