Drosophila melanogaster retrotransposon and inverted repeat-derived endogenous siRNAs are differentially processed in distinct cellular locations
© The Author(s). 2017
Received: 18 December 2016
Accepted: 7 April 2017
Published: 17 April 2017
Endogenous small interfering (esi)RNAs repress mRNA levels and retrotransposon mobility in Drosophila somatic cells by poorly understood mechanisms. 21 nucleotide esiRNAs are primarily generated from retrotransposons and two inverted repeat (hairpin) loci in Drosophila culture cells in a Dicer2 dependent manner. Additionally, proteins involved in 3’ end processing, such as Symplekin, CPSF73 and CPSR100, have been recently implicated in the esiRNA pathway.
Here we present evidence of overlap between two essential RNA metabolic pathways: esiRNA biogenesis and mRNA 3' end processing. We have identified a nucleus-specific interaction between the essential esiRNA cleavage enzyme Dicer2 (Dcr2) and Symplekin, a component of the core cleavage complex (CCC) required for 3' end processing of all eukaryotic mRNAs. This interaction is mediated by the N-terminal 271 amino acids of Symplekin; CCC factors CPSF73 and CPSF100 do not contact Dcr2. While Dcr2 binds the CCC, Dcr2 knockdown does not affect mRNA 3' end formation. RNAi-depletion of CCC components Symplekin and CPSF73 causes perturbations in esiRNA abundance that correlate with fluctuations in retrotransposon and hairpin esiRNA precursor levels. We also discovered that esiRNAs generated from retrotransposons and hairpins have distinct physical characteristics including a higher predominance of 22 nucleotide hairpin-derived esiRNAs and differences in 3' and 5' base preference. Additionally, retrotransposon precursors and derived esiRNAs are highly enriched in the nucleus while hairpins and hairpin derived esiRNAs are predominantly cytoplasmic similar to canonical mRNAs. RNAi-depletion of either CPSF73 or Symplekin results in nuclear retention of both hairpin and retrotransposon precursors suggesting that polyadenylation indirectly affects cellular localization of Dcr2 substrates.
Together, these observations support a novel mechanism in which differences in localization of esiRNA precursors impacts esiRNA biogenesis. Hairpin-derived esiRNAs are generated in the cytoplasm independent of Dcr2-Symplekin interactions, while retrotransposons are processed in the nucleus.
KeywordsDicer2 mRNA 3’ end processing Endogenous small interfering RNA biogenesis Core cleavage complex Symplekin CPSF73
In Drosophila, independent groups of small RNAs with overlapping function regulate gene expression using transcriptional and post-transcriptional mechanisms. PIWI-interacting RNAs (piRNAs) are found, most notably, in the germ line where they inhibit transposon (Tn) expression by inducing heterochromatin formation at complementary genomic Tn insertion sites [1–8]. Micro RNAs (miRNAs) and endogenous small interfering RNAs (esiRNAs) are expressed ubiquitously; however miRNAs frequently inhibit translation of protein coding genes , while esiRNAs are suggested to inhibit Tn mobility in Drosophila somatic cells [4–6] and potentially target mRNAs for degradation using a cytoplasmic RNAi mechanism [10, 11]. While PIWI mediated Tn repression in germ cells and translational inhibition by miRNAs have been actively investigated, the molecular details of how esiRNAs regulate their targets have not been described.
Twenty-one nucleotide (nt) esiRNAs are generated from double stranded (ds) precursor RNAs by Dicer-2 (Dcr2) and function through association with Argonaute-2 (Ago2) in Drosophila somatic cells [11–16]. esiRNAs produced in Drosophila tissues derive generally from cis-natural antisense transcripts (cis-NATs), inverted repeat containing single stranded RNAs (hairpins (hps)), and retroTns [11, 12, 14, 15]. In contrast, Drosophila culture cells generate esiRNAs predominantly from long terminal repeat (LTR) retrotransposons (retroTns) and hps; few cis-NAT derived esiRNAs are observed in S2 cell derived datasets [12, 13, 17]. Multi-copy LTR and non-LTR retroTns generate esiRNAs that map the entire length of these retroTns. In contrast, hp-derived esiRNAs arise from two loci, defined Esi1 and Esi2, within Drosophila annotated transcripts CR18854 and CG47744, respectively. Esi1 and Esi2 contain multiple inverted repeats, allowing formation of complex dsRNA secondary structures. These loci produce multiple esiRNAs, the most predominant termed Esi1.2 and Esi2.1. Differences between retroTn and hp-derived esiRNA biogenesis have not been previously investigated.
Drosophila LTR and non-LTR retroTns are transcribed in both the sense (S) and antisense (AS) directions from RNA polymerase II-like promoters . S retroTn transcripts are generally polyadenylated while AS transcripts are less likely to contain a poly(A) tail . Because retroTns are polyadenylated, the 3’ ends of potential esiRNA precursors are processed by the core cleavage complex (CCC) containing CPSF73, CPSF100 and Symplekin, [18–20] since this complex cleaves all eukaryotic mRNAs. Potential connections between mRNA 3’ end processing and esiRNA biogenesis are intriguing and have not been previously described.
esiRNAs regulate Tns and additional targets via multiple pathways: A canonical cytoplasmic post-transcriptional RNAi pathway in which esiRNAs hybridize to target mRNAs resulting in translational repression, and/or transcriptional regulation by induction of heterochromatin in the nucleus. mRNA targets of hp derived esiRNAs have been identified  and transcript levels of these targets are elevated in Dcr2 mutant flies,  supporting the post-transcriptional model. Evidence is mounting that Tn derived esiRNAs also mediate heterochromatin formation in Drosophila nuclei [1, 5–7]. Dcr2 catalytic mutants regulate position effect variegation [6, 7], a measure of heterochromatin formation [21, 22]. Additionally, Dcr2 promotes transcription of heat shock genes  and has been observed in the nuclei of Drosophila larvae . These data are consistent with a nuclear pool of Dcr2 that could contribute to transcriptional regulation by induction of heterochromatin in addition to cytoplasmic Dcr2 acting in the RNAi pathway.
To define connections between differential retroTn and hp-derived esiRNA processing and cellular location, and to investigate the potential link between mRNA 3’ end cleavage and esiRNA biogenesis, interactions between CCC components and Dcr2 were characterized and esiRNAs in control and RNAi-depleted Drosophila tissue culture cells were analyzed. These experiments revealed that Dcr2 and the CCC interact, but only in the nucleus, and that the CCC indirectly regulates esiRNA biogenesis by modulating dsRNA precursor levels. Additionally, retroTn- and hp-derived esiRNAs are physically distinct and occupy different subcellular compartments. RetroTn-derived esiRNAs and their precursors are retained in the nucleus while hp-derived esiRNAs and their precursors are exported to the cytoplasm. Collectively, these data support a model in which esiRNAs regulate gene expression and retroTn mobility via diverse compartmentalized mechanisms.
Stable expression of Symplekin mutants, RNAi
Crude nuclear and refined nuclear/cytoplasmic extracts
Crude nuclear extracts were prepared as described . Refined nuclear and cytoplasmic fractions were prepared as described [18, 25] with the following modifications. 500 x 106 cells were collected, washed 2X with cold PBS, re-suspended in 5X the cell pellet volume of hypotonic buffer (10 mM HEPES/KOH, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT) and incubated on ice for 30 min. Cells were lysed with a 15 mL tight glass dounce tissue homogenizer 20X. Lysate was spun at 17 K x g at 4 °C for 10 min to pellet crude nuclei. Supernatant was removed, spun again to remove any residual debris, snap frozen in Liquid N2 and saved as the refined cytoplasmic fraction. The crude nuclear pellet was resuspended in 2 mL S1 buffer (20 mM HEPES/KOH, pH 7.9, 0.88 M sucrose, 5 mM MgCl2, 0.5 mM DTT) layered on top of 15 mL S2 buffer (20 mM HEPES/KOH, pH 7.9, 2 M sucrose, 5 mM MgCl2, 0.5 mM DTT) and spun at 18.5 K x g for 30 min at 4 °C to pellet nuclei. Buffers S1 and S2 were removed. The refined nuclear pellet was resuspended in 500 μL of cold PBS and spun for 5 min at 18.5 K x g at 4 °C to repellet the nuclei. For IPs and western blots, the refined nuclear pellet was lysed in a high salt buffer as described . Both refined nuclear lysate and refined cytoplasmic lysate were dialyzed overnight in Buffer D (20 mM HEPES/KOH, pH 7.9, 20% glycerol, 100 mM KCl, 0.2 mM EDTA, 0.5 mM DTT). For RT-qPCR, RNA was extracted using 1 mL (nuclear pellet) or three volumes (cytoplasmic lysate) Trizol reagent (Ambion).
Immunoprecipitation, western blotting, S1 nuclease assay
Immunoprecipitation of HA-tagged proteins was performed as described . Immunoprecipitation of endogenous proteins was performed as described  using 100 μg of crude nuclear or 175 μg refined cytoplasmic or nuclear extracts. S1 nuclease protection assay was performed as described . Monoclonal and polyclonal HA antibodies (Cat#s MMS-101R and PRB-101C, respectively, Covance) were used for both IP (3 μL) and WB (1:1000). Anti-CPSF73, anti-Symplekin, and anti-CPSF100 antibodies (1:1000) were described previously [18, 26]. Commercial anti-Dcr-2 (Abcam ab4732), anti-Actin (Abcam ab8227), anti-H3 (Cell Signaling 4499), and anti-MEK1/2 (Cell Signaling 8727) were used at manufacturer recommended concentrations. The anti-R2D2 antibody was a generous gift from the Siomi lab .
RT-qPCR from nuclear and cytoplasmic fractions
Nuclear/cytoplasmic enrichment analysis of precursors and esiRNAs by RT-qPCR used Trizol prepped total RNA from the refined fractionation. All samples were column cleaned using the Qiagen miRNeasy Mini Kit (217004) and DNAse treated (Ambion Turbo DNAse # AM 1907) prior to RT. Equal cellular volumes were used in the RT step. RT-qPCR of precursors utilized iScript Reverse Transcription Supermix and SsoAdvanced Universal SYBR Green (Biorad #170884, #1725271, respectively). siRNA RT-qPCR was performed using Taqman Micro RNA RT Kit and Taqman Universal Master Mix (AB #4366596, #4440040, respectively.) Custom small RNA assay numbers and PCR primers are listed in Additional file 1. All qPCR experiments were performed in triplicate.
Immunofluorescence was performed essentially as described . Nuclei were stained with DAPI. Anti-Symplekin  was used at 1:500 and anti-Dcr-2  was used at 1:200. Secondary antibodies were used at 1:1000. Images were obtained on a Zeiss LSM 700, maintaining equal laser strength, gain and 1 AU. The images were processed with ImageJ.
rRNA depletion, library preparation and high-throughput sequencing
rRNA depletion, library preparation and next generation sequencing RNA-seq/small RNA-seq was performed as described . smRNA libraries were constructed from biological duplicates. One biological siRNA-seq and the RNA-seq were performed in technical triplicate.
siRNAs were analyzed using a newly developed pipeline called SMACR (Sequence Mapping, Annotation, and Counting for smRNAs; https://github.com/mrmckain/SMACR). Raw reads were first trimmed using Trimmomatic v.0.33 , with parameters optimized for siRNA data: Adapter trimming using TruSeq3-SE adapters, seed mismatch of 1, palindrome clip threshold of 20, and simple clip threshold of 7; a quality sliding window of 3 basepairs (bp) with a minimal average score of 20; and a minimum length of 19. Trimmed reads were then filtered to remove any longer than 30 bp. Relative abundances were then calculated for all unique trimmed reads. Unique is a read that is different from all others. The unique reads were then mapped to the Drosophila melanogaster genome (Dmel v.6.01; ) using bowtie v.1.1.1  allowing for either 0 or 1 mismatches. The mapping and read abundance information were then merged, estimating reads per million (rpm) for each mapped unique sequence. SMACR can simultaneously read in multiple experimental datasets, including replicates, and maintains each dataset as uniquely identified to the particular experiment and replicate. Annotation coordinates from Dmel v.6.01 for miRNAs, noncoding RNAs, transposons, and two hairpin structures were used to link mapped siRNAs to annotation features. If a siRNA was found to map to more than one feature type, it was disregarded. Abundance (normalized read counts) of siRNAs mapping to a particular feature were totaled and percentages of siRNAs mapping to each feature were calculated for each replicate. 5’ and 3’ nucleotide abundance, siRNA abundance, and relative phasing to the core siRNA for a given mapping site were then analyzed in the final set of siRNAs. Averages include technical triplicates and the biological replicate and standard deviations reflect the standard error of the mean for all four samples. RNA-seq reads from Symplekin and CPSF73 depleted samples were strand specifically mapped using the RNA-seq Unified Mapper (RUM)  and visualized with the University of California Santa Cruz (UCSC) genome browser (http://genome.ucsc.edu, Dm6 assembly, August 2014) [31, 34]. Further analyses were performed as in .
Poly(A)+/- analysis and calculations were performed as in .
mRNA 3’ end processing factor Symplekin interacts with Dcr2
To determine which region of Symplekin interacts with Dcr2, we immunoprecipitated stably expressed HA-tagged Symplekin deletions from Drosophila culture cell lysates. The N-terminal region of Symplekin (amino acids 1-271) clearly interacts with endogenous Dcr2 while the C-terminal region (amino acids 272-1165) does not (Fig. 1b). Reciprocal immunoprecipitation of endogenous Dcr2 reveals co-immunoprecipitation of HA-tagged Symp (1-271) (Fig. 1b).
To further investigate Dcr2-CCC interactions, we used a system in which N- and C-terminal Symplekin mutants are expressed in endogenous Symplekin RNAi-depleted cells. CCC formation is mediated by the C-termini of Symplekin, CPSF73 and CPSF100 . Therefore, exogenous HA-Symp (1-271) does not precipitate endogenous CPSF73 or CPSF100 when full-length endogenous Symplekin is knocked down (Fig. 1d). Unlike the interactions observed with full-length Symplekin, Dcr2 does not immunoprecipitate CPSF73 and CPSF100 under conditions when only HA-Symp (1-271) is present (Fig. 1d). Additionally, while CCCs are formed in cells expressing HA-Symp (272-1165), very little CPSF73 and CPSF100 interact with Dcr2 in these cells (Fig. 1d). These data suggest that the Symplekin N-terminal region interacts with Dcr2 while CPSF73 and CPSF100 are present in this complex via interaction with the C-terminal region of Symplekin .
The Dcr2-CCC complex is functionally distinct from the CCC
Because CCC depletion causes 3’ end misprocessing and Dcr2 interacts with the CCC, we wanted to determine the effects of Dcr2 depletion on mRNA 3’ end processing. First, we mapped the 3’ ends of endogenous Histone 2A (H2A) mRNAs in a Dcr2 depleted sample using an S1 nuclease protection assay. No differences in mRNA 3’ end processing were observed between Dcr2 knockdown and negative control samples (Fig. 2b). Dcr2 depletion does not cause misprocessing of histone mRNA 3’ ends as is observed when CCC components are knocked down (Fig. 2b) [18, 20]. Additionally, an RT-qPCR assay  was used to assess the effects of Dcr2 knockdown on mRNA 3’ end misprocessing of polyadenylated genes. Very little misprocessing of a canonical polyadenylated mRNA (sop) was observed in a Dcr2 depleted sample as compared to the positive Symplekin knockdown control; compare ~4-fold in the Dcr2 knockdown to ~85-fold in the Symplekin RNAi-depleted sample (Fig. 2c). Finally, visual inspection of mapped RNA-seq reads from control (LacZ), Dcr2 and Symplekin knockdown samples reveals mRNA 3’ end misprocessing is only present in the Symplekin RNAi-depleted sample; Dcr2 and LacZ knockdown samples show no or few RNA-seq reads mapping downstream of the IP3K1 gene (Fig 2d). These data support a model in which Dcr2 is not required for mRNA 3’ end processing and that the Dcr2-CCC complex is functionally distinct from the CCC as mRNA 3’ end processing is unaffected in the absence of Dcr2.
Dcr2 interacts with the CCC in the nucleus
Immunoprecipitations of endogenous Dcr2 from refined Dmel-2 nuclear and cytoplasmic fractions show that nuclear Dcr2 co-immunoprecipitates the CCC and R2D2 (Fig. 3b), while cytoplasmic Dcr2 only interacts with R2D2 (Fig. 3b); no interaction between cytoplasmic Dcr2 and the CCC is observed. Nuclear Symplekin co-immunoprecipitates Dcr2 and other CCC components, CPSF73 and CPSF100, but not R2D2 (Fig. 3b). Additionally, cytoplasmic Symplekin does not interact with Dcr2 (Fig. 3b). Together these data support a model in which Dcr2 forms distinct nuclear and cytoplasmic complexes.
The CCC indirectly regulates esiRNA abundance
To investigate the role of the CCC in esiRNA biogenesis, CPSF73, Symplekin and Dcr2 were first independently RNAi-depleted from Drosophila culture cells. Untreated Dmel-2 cells (blank) or those treated with double stranded RNA to a non-Drosophila gene (GFP or LacZ) acted as controls. RNA was then isolated and separated into large (>200 nts) and small (<200 nts) fractions, rRNA depleted and sequenced (Additional file 5). RNA-seq reads were mapped to the Drosophila genome and transcriptome using RUM,  while small RNA (smRNA)-seq reads were mapped and analyzed using a novel pipeline termed Sequence Mapping, Annotation, and Counting for smRNAs or SMACR (Methods and Additional file 5). Only siRNAs and miRNAs were further analyzed. Interestingly, ~40% of RNA-seq reads mapped to multiple locations in the genome (non-unique), although the samples were depleted of rRNAs (Additional file 6). Also, the percentage of non-unique reads changes significantly with knockdown of Dcr2 and Symplekin (Additional file 6). These data support previous claims that Dmel-2 culture cells have undergone Tn expansion [43–46] and indicate that increased numbers of Tns may contribute to higher overall expression of repetitive sequences and abundance of Dcr2 dependent siRNAs. Tn expansion makes Drosophila culture cells an excellent system for studying esiRNAs biogenesis.
To investigate potential explanations for the observed differences between Tn and hp derived esiRNA levels and differential effects of Dcr2 and CCC factor depletion on esiRNAs biogenesis, we first examined esiRNA and precursor levels for both hp loci (Esi 1/2) and individual retroTns: Dm297, mdg1 and jockey, as compared to the LacZ control (Fig. 4b and c). Normalized Esi 1/2 precursor and esiRNA levels in each knockdown sample were divided by the number of Esi 1/2 precursor and esiRNAs in the LacZ control. Esi 1/2 precursor levels decrease when CPSF73 and Symplekin are knocked down and we observe a corresponding decrease of esiRNAs in these samples (Fig. 4b). The number of esiRNAs mapping to mdg1 and jockey increase in response to CCC depletion while esiRNAs generated from Dm297 decrease in these samples (Fig. 4c). RetroTn esiRNA precursors consist of hybridized S and AS retroTn transcripts . We previously reported that both S and AS retroTn transcript levels are elevated in Dcr2 depleted cells . Knockdown of CCC components Symplekin or CPSF73 results in little to no change in sense Dm297, mdg1 or jockey transcript abundance, while the corresponding AS transcript levels are altered significantly (Fig. 4c). Interestingly, retroTn esiRNA levels correlate with perturbations in AS transcript levels in these samples (Fig. 4c). Previously, we hypothesized that the total amount of retroTn Dcr2 substrate in Dmel-2 cells is determined by AS transcript levels as these tend to be limiting . Therefore, changes in retroTn derived esiRNAs levels partially correspond to alterations in retroTn Dcr2 substrate levels. But, because changes in Esi 1/2 and retroTn esiRNA levels do not mirror alterations in substrate abundance (Fig. 4b and c), additional direct affects of CCC-Dcr2 interaction are plausible and require further investigation.
Hp and Tn-derived esiRNAs are physically distinct
Initial characterization of siRNAs and miRNAs in Dcr2 RNAi-depleted, CCC knockdown, and control samples included using SMACR to filter smRNAs by length, 3’, and 5’ base. Generally, there is little variation between depleted samples and few statistically significant differences between Dcr2, CPSF73 and Symplekin RNAi-depleted samples, and controls (Additional file 7). One notable exception is that Dcr2 depletion reduces the percentage of 21 nt Tn and hairpin-derived esiRNAs, while increasing esiRNAs of other sizes (Additional file 7). Knockdown of Dcr2 does not affect miRNA length distributions (Additional file 7), indicating that this molecular phenotype is specific to Dcr2 substrates. This trend is not observed for 3’ and 5’ base preference in Dcr2 knockdown samples (Additional file 7). CCC factor knockdowns have negligible affects on esiRNA size and end nucleotide preference supporting a model that CPSF73 and Symplekin do not directly affect Dcr2 catalytic activity (Additional file 7). Very few statistically significant differences in smRNA length, 3’, or 5’ base are observed between untreated cells and those treated with non-specific dsRNA (LacZ) suggesting that inducing the RNAi response does not affect smRNA size or end nucleotide preference.
Dramatic differences between 3’ base preference are also observed for miRNAs and esiRNAs. The 3’ nucleotide is A for ~75% of miRNAs while C, G and T occur much less frequently at this position (Fig. 6b). Tn generated esiRNAs also predominantly end in an A, but the 3’ nucleotide is more frequently C, G and T than for miRNAs; (Fig. 6b). Interestingly, the 3’ base of hp-derived esiRNAs is G more than 50% of the time, A is observed for ~30% of these esiRNAs, and T and C are present less often at the 3’ end (Fig. 6b). Therefore, while the Tn- and hp-derived esiRNA 3’ nt is generally more diverse than for miRNAs, significant differences are observed between esiRNAs. 5’ nucleotide distributions are more similar for all three smRNA classes. C is the most abundant 5’ nucleotide for miRNAs and esiRNAs, while G and A are also frequently observed (between 25 and 40%, Fig. 6c). T is the least populous 5’ nucleotide for miRNAs and esiRNAs (Fig. 6c). Collectively, these data indicate that esiRNAs processed from Tns and hps have diverse physical characteristics and support a model in which these two precursors are differentially processed in Dmel-2 cells.
RetroTn precursors and esiRNAs are retained in the nucleus
To assess the cellular localization of retroTn, Esi1 and Esi2-derived esiRNAs, we measured levels of the most abundant esiRNAs from these precursors with custom Taqman assays. A cytoplasmic miRNA control (mir2A) is ~2.5 fold enriched in the cytoplasm (Fig. 7b). Strikingly, both Dm297 and mdg1-derived esiRNAs are enriched in the nucleus while Esi2 and Esi1-derived esiRNAs (Esi2.1 and Esi1.2, respectively) localize to the cytoplasm (Fig. 7b). These data correlate with nuclear enrichment of retroTns and nuclear export of hps (Fig. 7a). Together these data support a model in which retroTn dsRNA precursors are retained and processed in the nucleus while single stranded Esi1 and Esi2 precursors are exported to the cytoplasm for Dcr2-dependent generation of esiRNAs.
To investigate the role of CCC components Symplekin and CPSF73 in cellular localization of Dcr2 substrates, a subset of retroTn and hp transcript levels was assessed by RT-qPCR in refined nuclear and cytoplasmic fractions from CCC factor RNAi-depleted samples. Dcr2 knockdown has marginal effects on cellular localization of CG44774, Dm297, and mdg1 RNAs as compared to the GAPDH control. In contrast, Symplekin and CPSF73 RNA-depletions cause slight nuclear retention of hp CG44774 and further nuclear enrichment of retroTn transcripts as compared to GAPDH (Fig. 7c). While Both the CPSF73 and Symplekin knockdowns show similar trends, the effects observed when CPSF73 is RNAi-depleted are more dramatic than when Symplekin is RNAi-depleted. As both CG44774 and CR18854, and retroTn sense transcripts are polyadenylated (Additional file 8 and , respectively), we hypothesize that mRNA 3’ end processing defects in the CCC knockdown samples affect cellular localization of these Dcr2 substrates.
Since the discovery of endogenous small interfering (esiRNAs) in Drosophila, very little progress in understanding their biogenesis and molecular mechanisms of action has been made. Here we provide evidence that components of two major RNA processing pathways, 3’ end processing and esiRNA biogenesis, interact in Drosophila somatic cells, a connection not previously reported. Importantly, we also show that esiRNAs processed from retroTns have different physical characteristics than those generated from hairpins. Double-stranded retroTn RNAs are retained in the nucleus while Esi1/2 hps are exported to the cytoplasm. Together these data support a novel model in which retroTns and hps, both double stranded RNAs cleaved by Dcr2, are differentially processed in Drosophila somatic cells. This is the first evidence that precursor secondary structures potentially contribute to Dcr2 activity in vivo.
To understand the functional implications of CCC-Dcr2 interactions, esiRNA and precursor levels were measured in Symplekin and CPSF73 RNAi-depleted samples. Globally, we observe increased levels of Tn-derived esiRNAs and decreased hp-derived esiRNAs in CCC factor knockdowns (Fig. 4). Examination of specific retroTns and Esi1/2 precursors reveals that changes in esiRNA levels correlate to shifts in precursor abundance in Symplekin and CPSF73 knockdowns (Fig. 4). As we hypothesize that dsRNA retroTn precursor levels are determined AS transcript abundance , more AS transcript would lead to an increase in Dcr2 substrates (and more retroTn-derived esiRNAs), while decreased AS transcript would result in less Dcr2 substrate. Esi1/2 precursors consist of only one S mRNA. Therefore, hp Dcr2 substrate concentration is determined by only CG44774 and CR18854 levels. Once again, the observed lower Esi1/2 esiRNA levels in CCC factor depleted cells correlate with decreased CG44774 and CR18854 transcripts in these samples (Fig. 4). These data support a model in which esiRNA levels are partially influenced by Dcr2 substrate concentration. Substrate levels are affected by CCC factor RNAi-depletion indicating that Symplekin and CPSF73 indirectly determine esiRNA abundance. Because Symplekin and CPSF73 RNAi-depleted samples follow the same trends, we concluded that the CCC is involved in this process.
While hp and retroTn esiRNA levels correlate to S and AS transcript abundance, the number of esiRNAs is always less than the Dcr2 substrate concentration in Symplekin and CPSF73 knockdowns indicating that additional mechanisms must be modulating esiRNA levels in these samples. Although, Dcr2 cleavage site selectivity is unaffected in CCC factor RNAi-depleted samples (Additional file 7), Dcr2 activity could be altered by interaction with the CCC. An additional hypothesis for the observed molecular phenotypes is inefficient nuclear export of retroTn and hp RNAs in Symplekin and CPSF73 RNAi-depleted samples. CCC component knockdowns cause global mRNA 3’ end processing defects (Fig. 2) and how this misprocessing affects cellular localization of retroTn dsRNAs is unknown. However, previous work shows that less polyadenylated RNAs are not effectively exported from the nucleus . Additionally, 3’ end misprocessing of RNAs generated from the Esi2 locus (Fig. 5) might lead to changes in secondary structure that unpredictably affect nuclear export. Inefficient nuclear export of hp RNAs with modified 3’ ends might not change total precursor levels, but could result in less Esi2-derived esiRNAs since cytoplasmic hp precursor levels would be reduced. When Symplekin and CPSF73 are RNAi-depleted, both hp and retroTn dsRNAs are enriched in the nucleus (Fig. 7) supporting the hypothesis that non-polyadenylated RNAs are retained in the nucleus. Taken together, these data support a model in which the CCC indirectly affects the abundance of retroTn- and hp-derived esiRNAs by modulating cellular localization and concentration of Dcr2 substrates (Fig. 8).
Bioinformatic analyses of retroTn- and hp-derived esiRNAs reveals physical distinctions between these groups (Fig. 6). Additionally, retroTn precursors and their corrosponding esiRNAs are highly enriched in the nucleus while hp dsRNAs and their corresponding esiRNAs are cytoplasmic similar to mRNAs (Fig. 7). We hypothesize that these observed disparities are directly related to distinct differences in secondary structures (Fig. 5) and compartmentalization of esiRNA biogenesis factors required to process each structure. dsRNAs derived from S and AS transcription of retroTns  generally result in fully complementary, blunt-ended dsRNAs as many AS retroTn transcripts are poorly polyadenylated . The secondary structures of hps containing multiple inverted repeats are likely variable and complex with frayed ends (Fig. 5). Previous in vitro assays suggest that Dcr2 alone can bind and processively cleave blunt dsRNAs. However, Dcr2 requires a co-factor, Loqs-PD, to process dsRNAs with frayed termini presumably because Loqs-PD allows Dcr2 to bind a substrate internally ; Loqs-PD is cytoplasmic in Drosophila culture cells . Taken together, these data suggest a model in which nuclear retained blunt-ended, fully complementary retroTn precursors can be processed in the nucleus by Dcr2 alone while more complicated hp precursors requiring Loqs-PD are cleaved in the cytoplasm by Dcr2 (Fig. 8). This model is supported by our observations that esiRNAs map the entire length of retroTns (Fig. 3d). Additionally, previous work shows that R2D2 and Dcr2 aggregate in cytoplasmic D2 bodies together with hps .
This model predicts that depletion of Loqs-PD would only affect cleavage of hps precursors and the levels of hp-derived esiRNAs, but not esiRNAs generated from retroTns. Zhou et al. previously reported that depletion of Loqs isoforms reduced the number of esiRNAs derived from both hps and Tns ; however, close examination of the data reveal that retroTn-mapping esiRNAs were unaffected by Loqs knockdown. The most notably affected Tn, Proto-P, is not regulated by the esiRNA pathway .
Our data support a novel model in which esiRNAs are differentially processed from retroTn and hp precursors; retroTn precursors are processed by Dcr2 in the nucleus, while biogenesis of esiRNAs from hp precursors occurs in the cytoplasm. Additionally, Dcr2 clearly interacts with the CCC in the nucleus, but not in the cytoplasm. The CCC indirectly affects esiRNAs biogenesis by regulating Dcr2 substrate levels and directing cellular localization of retroTn and hp RNAs. These data contribute significantly to our understanding of Dcr2 dependent esiRNA production in Drosophila culture cells, but questions regarding Dcr2-CCC complex assembly and function remain. Future studies investigating the role of the Dcr2-CCC complex in both mRNA 3’ end processing and retroTn dsRNA processing will further elucidate molecular details of how these proteins function in Drosophila culture cells.
Core cleavage complex
Cleavage and polyadenylation specificity factor
double stranded RNA
endogenous small interfering RNAs
Sequence Mapping, Annotation, and Counting for smRNAs
We thank Dr. Michael E. Hughes for guidance in bioinformatics analysis. Dr. Mikiko Siomi kindly provided anti-Dcr2 and anti-R2D2 antibodies. Mass spectrometry was performed at the University of North Carolina-Chapel Hill Proteomics Core Facility. We thank Dr. Ambrose Kidd, Dr. Bethany Zolman, Dr. Lon Chubiz and Dr. Wendy Olivas for critical review of the manuscript and members of the Steiniger Lab for thoughtful discussions.
This work was supported by NIH R15 GM107931 and startup funds to M. S. The funding body played no part in the design of the study or collection, analysis, and interpretation of data or in writing the manuscript.
Availability of data and materials
The datasets generated and/or analyzed during the current study are available in the GEO repository, Accession #GSE67725 for LacZ and Dcr2 RNA-seq and smRNA-seq datasets. Accession #GSE82128 for Blank, Symplekin and CPSF73 RNA-seq and smRNA-seq datasets.
AWH, DM, KMB and JMD performed all of the experiments. MRM developed SMACR and provided bioinformatics analyses. AWH, DM, and MS designed experiments. AWH and MS wrote the manuscript. All authors have read and approved the manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Sentmanat MF, Elgin SCR. Ectopic assembly of heterochromatin in Drosophila melanogaster triggered by transposable elements. Proc Natl Acad Sci. 2012;109:14104–9.View ArticlePubMedPubMed CentralGoogle Scholar
- 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.View ArticlePubMedGoogle Scholar
- Gu T, Elgin SCR. Maternal Depletion of Piwi, a Component of the RNAi System, impacts heterochromatin formation in drosophila. Brennecke J, editor. PLoS Genet. 2013;9:e1003780.View ArticlePubMedPubMed CentralGoogle Scholar
- Xie W, Donohue RC, Birchler JA. Quantitatively increased somatic transposition of transposable elements in Drosophila strains compromised for RNAi. PLoS One. 2013;8:e72163.View ArticlePubMedPubMed CentralGoogle Scholar
- Savva YA, Jepson JEC, Chang Y-J, Whitaker R, Jones BC, Laurent GS, et al. RNA editing regulates transposon-mediated heterochromatic gene silencing. Nature Communications. Nature Publishing Group. 2013;4:1–11.
- Fagegaltier D, Bougé A-L, Berry B, Poisot E, Sismeiro O, Coppée J-Y, et al. The endogenous siRNA pathway is involved in heterochromatin formation in Drosophila. Proc Natl Acad Sci. 2009;106:21258–63.View ArticlePubMedPubMed CentralGoogle Scholar
- Haynes KA, Caudy AA, Collins L, Elgin SCR. Element 1360 and RNAi components contribute to HP1-dependent silencing of a pericentric reporter. Curr Biol. 2006;16:2222–7.View ArticlePubMedPubMed CentralGoogle Scholar
- Savva YA, Jepson JEC, Chang Y-J, Whitaker R, Jones BC, St Laurent G, et al. RNA editing regulates transposon-mediated heterochromatic gene silencing. Nat Commun. 2013;4:2745.View ArticlePubMedPubMed CentralGoogle Scholar
- Valencia-Sanchez MA, Liu J, Hannon GJ, Parker R. Control of translation and mRNA degradation by miRNAs and siRNAs. Genes Dev. 2006;20:515–24.View ArticlePubMedGoogle Scholar
- Marques JT, Kim K, Wu P-H, Alleyne TM, Jafari N, Carthew RW. Loqs and R2D2 act sequentially in the siRNA pathway in Drosophila. Nat Struct Mol Biol. 2010;17:24–30.View ArticlePubMedGoogle Scholar
- 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.View ArticlePubMedPubMed CentralGoogle Scholar
- Ghildiyal M, Seitz H, Horwich MD, Li C, Du T, Lee S, et al. Endogenous siRNAs derived from transposons and mRNAs in Drosophila somatic cells. Science. 2008;320:1077–81.View ArticlePubMedPubMed CentralGoogle Scholar
- Kawamura Y, Saito K, Kin T, Ono Y, Asai K, Sunohara T, et al. Drosophila endogenous small RNAs bind to Argonaute 2 in somatic cells. Nature. 2008;453:793–7.View ArticlePubMedGoogle Scholar
- Okamura K, Balla S, Martin R, Liu N, Lai EC. Two distinct mechanisms generate endogenous siRNAs from bidirectional transcription in Drosophila melanogaster. Nat Struct Mol Biol. 2008;15:581–90.View ArticlePubMedPubMed CentralGoogle Scholar
- Okamura K, Chung W-J, Ruby JG, Guo H, Bartel DP, Lai EC. The Drosophila hairpin RNA pathway generates endogenous short interfering RNAs. Nature. 2008;453:803–6.View ArticlePubMedPubMed CentralGoogle Scholar
- Iwasaki S, Sasaki HM, Sakaguchi Y, Suzuki T, Tadakuma H, Tomari Y. Defining fundamental steps in the assembly of the Drosophila RNAi enzyme complex. Nature. 2015;521:533–6.View ArticlePubMedGoogle Scholar
- Russo J, Harrington AW, Steiniger M. Antisense transcription of retrotransposons in drosophila: an origin of endogenous small interfering RNA precursors. Genetics. 2016;202:107–21.View ArticlePubMedGoogle Scholar
- Sullivan KD, Steiniger M, Marzluff WF. A core complex of CPSF73, CPSF100, and Symplekin may form two different cleavage factors for processing of poly(A) and histone mRNAs. Mol Cell. 2009;34:322–32.View ArticlePubMedPubMed CentralGoogle Scholar
- Ryan K, Calvo O, Manley JL. Evidence that polyadenylation factor CPSF-73 is the mRNA 3' processing endonuclease. RNA. 2004;10:565–73.View ArticlePubMedPubMed CentralGoogle Scholar
- Michalski D, Steiniger M. In vivo characterization of the Drosophila mRNA 3' end processing core cleavage complex. RNA. 2015;21:1404–18.View ArticlePubMedPubMed CentralGoogle Scholar
- Sun FL, Cuaycong MH, Elgin SC. Long-range nucleosome ordering is associated with gene silencing in Drosophila melanogaster pericentric heterochromatin. Mol Cell Biol. 2001;21:2867–79.View ArticlePubMedPubMed CentralGoogle Scholar
- Agranat L, Raitskin O, Sperling J, Sperling R. The editing enzyme ADAR1 and the mRNA surveillance protein hUpf1 interact in the cell nucleus. Proc Natl Acad Sci. 2008;105:5028–33.View ArticlePubMedPubMed CentralGoogle Scholar
- Cernilogar FM, Onorati MC, Kothe GO, Burroughs AM, Parsi KM, Breiling A, et al. Chromatin-associated RNA interference components contribute to transcriptional regulation in Drosophila. Nature. 2011;480:391–5.View ArticlePubMedPubMed CentralGoogle Scholar
- Grimaud C, Bantignies F, Pal-Bhadra M, Ghana P, Bhadra U, Cavalli G. RNAi components are required for nuclear clustering of Polycomb group response elements. Cell. 2006;124:957–71.View ArticlePubMedGoogle Scholar
- Nechaev S, Fargo DC, dos Santos G, Liu L, Gao Y, Adelman K. Global analysis of short RNAs reveals widespread promoter-proximal stalling and arrest of Pol II in Drosophila. Science. 2010;327:335–8.View ArticlePubMedGoogle Scholar
- Yang X-C, Burch BD, Yan Y, Marzluff WF, Dominski Z. FLASH, a proapoptotic protein involved in activation of caspase-8, is essential for 3' end processing of histone pre-mRNAs. Mol Cell. 2009;36:267–78.View ArticlePubMedPubMed CentralGoogle Scholar
- Nishida KM, Miyoshi K, Ogino A, Miyoshi T, Siomi H, Siomi MC. Roles of R2D2, a cytoplasmic D2 body component, in the endogenous siRNA pathway in Drosophila. Mol Cell. 2013;49:680–91.View ArticlePubMedGoogle Scholar
- Rogers SL, Rogers GC. Culture of Drosophila S2 cells and their use for RNAi-mediated loss-of-function studies and immunofluorescence microscopy. Nat Protoc. 2008;3:606–11.View ArticlePubMedGoogle Scholar
- Miyoshi K, Okada TN, Siomi H, Siomi MC. Characterization of the miRNA-RISC loading complex and miRNA-RISC formed in the Drosophila miRNA pathway. RNA. 2009;15:1282–91.View ArticlePubMedPubMed CentralGoogle Scholar
- Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30:2114–20.View ArticlePubMedPubMed CentralGoogle Scholar
- dos Santos G, Schroeder AJ, Goodman JL, Strelets VB, Crosby MA, Thurmond J, et al. FlyBase: introduction of the Drosophila melanogaster Release 6 reference genome assembly and large-scale migration of genome annotations. Nucleic Acids Res. 2015;43:D690–7.View ArticlePubMedGoogle Scholar
- 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.View ArticlePubMedPubMed CentralGoogle Scholar
- Grant GR, Farkas MH, Pizarro AD, Lahens NF, Schug J, Brunk BP, et al. Comparative analysis of RNA-Seq alignment algorithms and the RNA-Seq unified mapper (RUM). Bioinformatics. 2011;27:2518–28.PubMedPubMed CentralGoogle Scholar
- Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, et al. The human genome browser at UCSC. Genome Res. 2002;12:996–1006.View ArticlePubMedPubMed CentralGoogle Scholar
- Schönemann L, Kühn U, Martin G, Schäfer P, Gruber AR, Keller W, et al. Reconstitution of CPSF active in polyadenylation: recognition of the polyadenylation signal by WDR33. Genes Dev. 2014;28:2381.View ArticlePubMedPubMed CentralGoogle Scholar
- Chan SL, Huppertz I, Yao C, Weng L, Moresco JJ, Yates JR, et al. CPSF30 and Wdr33 directly bind to AAUAAA in mammalian mRNA 3' processing. Genes Dev. 2014;28:2370–80.View ArticlePubMedPubMed CentralGoogle Scholar
- Sabath I, Skrajna A, Yang X-C, Dadlez M, Marzluff WF, Dominski Z. 3'-End processing of histone pre-mRNAs in Drosophila: U7 snRNP is associated with FLASH and polyadenylation factors. RNA. 2013;19:1726–44.View ArticlePubMedPubMed CentralGoogle Scholar
- Iwasaki S, Kobayashi M, Yoda M, Sakaguchi Y, Katsuma S, Suzuki T, et al. Hsc70/Hsp90 chaperone machinery mediates ATP-dependent RISC loading of small RNA duplexes. Mol Cell. 2010;39:292–9.View ArticlePubMedGoogle Scholar
- Liu Q, Rand TA, Kalidas S, Du F, Kim H-E, Smith DP, et al. R2D2, a bridge between the initiation and effector steps of the Drosophila RNAi pathway Science. Am Assoc Adv Sci. 2003;301:1921–5.Google Scholar
- Tatomer DC, Rizzardi LF, Curry KP, Witkowski AM, Marzluff WF, Duronio RJ. Drosophila symplekin localizes dynamically to the histone locus body and tricellular junctions. Nucleus. 2014;5:613.View ArticlePubMedPubMed CentralGoogle Scholar
- Kim JH, Richter JD. Opposing polymerase-deadenylase activities regulate cytoplasmic polyadenylation. Mol Cell. 2006;24:173–83.View ArticlePubMedGoogle Scholar
- Barnard DC, Ryan K, Manley JL, Richter JD. Symplekin and xGLD-2 are required for CPEB-mediated cytoplasmic polyadenylation. Cell. 2004;119:641–51.View ArticlePubMedGoogle Scholar
- Potter SS, Brorein WJ, Dunsmuir P, Rubin GM. Transposition of elements of the 412, copia and 297 dispersed repeated gene families in Drosophila. Cell. 1979;17:415–27.View ArticlePubMedGoogle Scholar
- Tchurikov NA, Ilyin YV, Skryabin KG, Ananiev EV, Bayev AA, Krayev AS, et al. General properties of mobile dispersed genetic elements in Drosophila melanogaster. Cold Spring Harb Symp Quant Biol. 1981;45 Pt 2:655–65.View ArticlePubMedGoogle Scholar
- Maisonhaute C, Ogereau D, Hua-Van A, Capy P. Amplification of the 1731 LTR retrotransposon in Drosophila melanogaster cultured cells: origin of neocopies and impact on the genome. Gene. 2007;393:116–26.View ArticlePubMedGoogle Scholar
- Wen J, Mohammed J, Bortolamiol-Becet D, Tsai H, Robine N, Westholm JO, et al. Diversity of miRNAs, siRNAs, and piRNAs across 25 Drosophila cell lines. Genome Res. 2014;24:1236.View ArticlePubMedPubMed CentralGoogle Scholar
- Greenleaf AL. Positive patches and negative noodles: linking RNA processing to transcription? Trends Biochem Sci. 1993;18:117–9.View ArticlePubMedGoogle Scholar
- Bentley DL. Rules of engagement: co-transcriptional recruitment of pre-mRNA processing factors. Curr Opin Cell Biol. 2005;17:251–6.View ArticlePubMedGoogle Scholar
- Xiang K, Nagaike T, Xiang S, Kilic T, Beh MM, Manley JL, et al. Crystal structure of the human symplekin-Ssu72-CTD phosphopeptide complex. Nature. 2010;467:729–33.View ArticlePubMedPubMed CentralGoogle Scholar
- Huang XA, Yin H, Sweeney S, Raha D, Snyder M, Lin H. A major epigenetic programming mechanism guided by piRNAs. Dev Cell. 2013;24:502–16.View ArticlePubMedPubMed CentralGoogle Scholar
- Le Thomas A, Rogers AK, Webster A, Marinov GK, Liao SE, Perkins EM, et al. Piwi induces piRNA-guided transcriptional silencing and establishment of a repressive chromatin state. Genes Dev. 2013;27:390–9.View ArticlePubMedPubMed CentralGoogle Scholar
- Rozhkov NV, Hammell M, Hannon GJ. Multiple roles for Piwi in silencing Drosophila transposons. Genes Dev. 2013;27:400–12.View ArticlePubMedPubMed CentralGoogle Scholar
- Sienski G, Dönertas D, Brennecke J. Transcriptional silencing of transposons by Piwi and maelstrom and its impact on chromatin state and gene expression. Cell. 2012;151:964–80.View ArticlePubMedPubMed CentralGoogle Scholar
- Wang SH, Elgin SCR. Drosophila Piwi functions downstream of piRNA production mediating a chromatin-based transposon silencing mechanism in female germ line. Proc Natl Acad Sci. 2011;108:21164–9.View ArticlePubMedPubMed CentralGoogle Scholar
- Huang Y, Carmichael GG. Role of polyadenylation in nucleocytoplasmic transport of mRNA. Mol Cell Biol. 1996;16:1534–42.View ArticlePubMedPubMed CentralGoogle Scholar
- Sinha NK, Trettin KD, Aruscavage PJ, Bass BL. Drosophila Dicer-2 cleavage is mediated by helicase- and dsRNA termini-dependent states that Are Modulated by Loquacious-PD. Mol Cell. 2015;58:406–17.View ArticlePubMedPubMed CentralGoogle Scholar
- Miyoshi K, Miyoshi T, Hartig JV, Siomi H, Siomi MC. Molecular mechanisms that funnel RNA precursors into endogenous small-interfering RNA and microRNA biogenesis pathways in Drosophila. RNA. 2010;16:506–15.View ArticlePubMedPubMed CentralGoogle Scholar
- Zhou R, Czech B, Brennecke J, Sachidanandam R, Wohlschlegel JA, Perrimon N, et al. Processing of Drosophila endo-siRNAs depends on a specific Loquacious isoform. RNA. 2009;15:1886–95.View ArticlePubMedPubMed CentralGoogle Scholar
- Harrington AW, Steiniger M. Bioinformatic analyses of sense and antisense expression from terminal inverted repeat transposons in Drosophila somatic cells. Fly (Austin). 2016;10:1–10.View ArticleGoogle Scholar