Two-step cleavage of hairpin RNA with 5' overhangs by human DICER
- 6.6k Downloads
DICER is an RNase III family endoribonuclease that processes precursor microRNAs (pre-miRNAs) and long double-stranded RNAs, generating microRNA (miRNA) duplexes and short interfering RNA duplexes with 20~23 nucleotides (nts) in length. The typical form of pre-miRNA processed by the Drosha protein is a hairpin RNA with 2-nt 3' overhangs. On the other hand, production of mature miRNA from an endogenous hairpin RNA with 5' overhangs has also been reported, although the mechanism for this process is unknown.
In this study, we show that human recombinant DICER protein (rDICER) processes a hairpin RNA with 5' overhangs in vitro and generates an intermediate duplex with a 29 nt-5' strand and a 23 nt-3' strand, which was eventually cleaved into a canonical miRNA duplex via a two-step cleavage. The previously identified endogenous pre-miRNA with 5' overhangs, pre-mmu-mir-1982 RNA, is also determined to be a substrate of rDICER through the same two-step cleavage.
The two-step cleavage of a hairpin RNA with 5' overhangs shows that DICER releases double-stranded RNAs after the first cleavage and binds them again in the inverse direction for a second cleavage. These findings have implications for how DICER may be able to interact with or process differing precursor structures.
KeywordsCleavage Product miRNA Duplex RISC Loading Complex Recombinant Dicer dsRNA Binding Motif
DICER plays a key role in RNA interference pathways through the biogenesis of microRNA (miRNA) and small interfering RNA (siRNA) [1, 2, 3]. Most miRNA genes are transcribed as long primary transcripts (pri-miRNAs) where stem-loop structures with mature miRNA sequences embedded in the arm of a stem are cleaved by the Drosha nuclear microprocessor complex releasing a precursor miRNA (pre-miRNA) hairpin [4, 5]. The cleavage site is determined mainly by the distance (~11 bp) from the stem-single stranded RNA junction of pri-miRNA and most pre-miRNAs have 2 nt-3' overhangs . Pre-miRNAs, exported into the cytoplasm by Exportin-5 and Ran-GTP , are processed by the RISC loading complex (RLC) into 20~23 nt duplexes where the RNase III enzyme DICER plays a central role together with the double stranded (ds) RNA-binding proteins TRBP and PACT and the miRNA-associated RNA-induced silencing complex (miRISC) core component Argonaute-2 (AGO2) [8, 9, 10]. miRNA duplexes processed by RLC are finally loaded to miRISC as a double stranded-structure  and separated into the functional guide strand, which is complementary to the target, and the passenger strand, which is subsequently degraded [12, 13]. Strand selection of the functional guide strand by AGO2 depends on the thermodynamic stabilities of the base pairs at the 5' ends of the two strands [12, 14, 15]. Duplexes of siRNA or miRNA produced by DICER can be loaded in either direction to Argonaute [16, 17, 18]. Indeed, the mature miRNA either in the 5' or 3' strands can be harboured from pre-miRNA [19, 20, 21]. On the other hand, endogenous human AGO2 can bind directly to pre-miRNAs in DICER-knockout cells . Recently, it was reported that human DICER is not essential for loading dsRNAs to AGO2 but functions in pre-selection of effective siRNAs for handoff to AGO2 .
Human DICER is a ~220 kDa protein consisting of several domains; an N-terminal DExH-box RNA helicase-like domain, a DUF283 domain, a PAZ domain, two RNase III domains (RIIIa and RIIIb), and a dsRNA binding motif domain (DARM) . The two RNase III domains of DICER form a single dsRNA processing center via intramolecular dimerization which together cleave the opposite strands of the dsRNA, generating dinucleotide-long 3' overhangs on both ends . The crystal structure of Dicer from Giardia intestinalis showed that the hydrophobic pocket of the PAZ domain was responsible for the binding of the 3' dinucleotide overhangs of the substrate and the connector helix between the PAZ domain and RNase III domain functioned as a molecular ruler measuring the distance from the 3' end of pre-miRNA to the cleavage site [26, 27]. However, 3'-dinucleotide dsRNA overhangs are not essential for binding with DICER . When the 3' overhang is removed, DICER can still cleave dsRNA through interaction with the remaining 5' overhang . This is consistent with MacRae et al. who found that the recombinant Dicer protein of Giardia intestinalis could cleave the dsRNA with 5' overhangs . However, they used perfectly matched dsRNAs with no gap, which might resemble an endogenous siRNA precursor. An additional study by Flores-Jasso et al. showed that human recombinant DICER protein could nick either strand of a mononucleotide-5' overhanged pre-miRNA with some strand preferences . Despite this, the detailed step mechanism for pre-miRNA cleavage, especially for the pre-miRNA with 5' overhangs, is not yet elucidated.
An alternative nuclear pathway of pre-miRNA biogenesis was described where a short intron with a hairpin can be spliced and debranched into pre-miRNA hairpin mimics (mirtrons) [30, 31, 32]. This processing pathway uses intron splicing machinery instead of the Drosha endonuclease; miRNA precursors generated from intronic sequences (debranched mirtrons) are believed to be incorporated into the canonical miRNA pathway as a substrate of DICER. Interestingly, mouse pre-mir-1982 is a mirtron with an 11 nt tail at the 5' end , although most mammalian mirtron are hairpin structures with single nucleotide overhangs at both ends [32, 33, 34]. Mature mouse miR-1982* miRNA emerges without 11 nt-5' overhangs from deep sequencing data of murine cells [33, 35] while the elimination mechanism of this 11 nt-5' tail is still unknown.
In this paper, we investigated the detailed processing pattern of hairpin RNAs containing 5' overhangs by human recombinant DICER. We show here that human recombinant DICER is able to process hairpin RNA with 5' overhangs and two-step cleavage by DICER forms the mature miRNA duplex from the hairpin RNAs. Additionally, pre-mmu-mir-1982 RNA, which is a natural hairpin RNA with 5' overhangs, is also processed by a two-step cleavage mediated by human recombinant DICER protein in vitro.
Results and Discussion
Processing of the pre-miRNA by recombinant DICER protein
Processing of the hairpin RNA with 5' overhangs, RNA-I, by recombinant DICER protein
To analyze band 1 at a longer incubation time, the RNA-I was incubated with the purified rDICER and the cleavage reactions were performed for 30 min and 16 hours. Surprisingly, band 1 was detected at 30 min incubation but disappeared after 16-hours incubation. On the other hand, band 2 continued to accumulate (see Figure 3E). These results showed that band 1, which seems to be a product of first processing by rDICER, disappeared following an extended incubation time.
Cloning of products from RNA-I processed by recombinant DICER protein
In this research, we found that hairpin RNA with trinucleotide-5' overhangs was cleaved into a 23-22 nt RNA duplex through two-step processing by rDICER. This could not be detected if we used only end-labelled RNA or label-incorporated RNA as a substrate for rDICER as reported previously [27, 29]. In the first step, rDICER processes the hairpin RNA with 5' overhangs to dsRNA with 29 nt and 23 nt. Our results indicate the first processed dsRNA binds again in the inverse direction with the same or a different rDICER molecule and is again effective cleaved. The results are consistent with the previous report that human DICER protein binds either 3' ends of dsRNA strand on the PAZ domain and cleaves dsRNA at the ~23 nt position from the binding end . In the sequential process described here, dsRNAs with 29 nt and 23 nt gradually increased and then stabilized at a steady level, followed by rapid increase of 23-22 nt duplexes (see Figure 3B and 3C). This indicates that dsRNAs with 29 nt and 23 nt are processed at a constant rate and 23-22 nt duplexes are belatedly processed at a similar rate. This suggests that the binding and processing of miRNA duplexes and releasing of the duplex from rDICER occur at the same rate. The bidirectional binding of processed dsRNA by DICER could result in directional presentation of dsRNA to Argonaute [16, 17, 18].
The effect of 5' overhangs in substrate cleavage by recombinant DICER protein
We generated three different hairpin RNAs based on pre-hsa-mir-21 RNA: pre-mir-21 RNA, RNA-I and RNA-II (see Figure 2A, 3A and Additional file 1: Figure S1A, respectively). RNA-II is the same as pre-mir-21 RNA except for 5' addition. Using probe-1 to detect the cleavage product from the 5' strand for Northern blotting, both bands of mature miRNA in Figure 2B and band 1 (first cleavage product) in Figure 3B emerged after 20-min incubation and increased gradually. On the other hand, band 8 (probable first cleavage product) in Figure S1B was detected faintly after 30-min incubation but increased quite slowly (see Additional file 1: Figures S1B and S1D). Band 9 in Figure S1C was also detected in a similar manner as band 8 (see Additional file 1: Figure S1C). Our results indicated that longer 5' overhangs with stable stem structures could reduce the efficiency or rate of substrate cleavage.
Processing of an endogenous hairpin RNA with 5' overhangs, pre-mmu-mir-1982 RNA, by recombinant DICER protein
In addition to miRNA maturation, mammalian DICER also processes other kinds of small RNAs including endo-siRNAs [33, 37, 38, 39]. It has been reported that mammalian endo-siRNAs are processed from various precursors including long hairpin RNAs and naturally formed dsRNAs resulting from bidirectional transcripts or antisense transcripts from pseudogenes. Although the complete structure of these precursor dsRNAs remains unclear, it seems likely that they have diverse 5' and 3' structures. Our results indicate DICER tolerance for 5' substrate overhang, potentially increasing the range of small RNA substrates that DICER can process. Recently, it was reported that the AGO2 protein could bind not only siRNAs and miRNAs but longer RNAs and pre-miRNAs [22, 40]. However, most endogenous AGO2 proteins bind miRNAs  and the RISC requires 3' overhangs for the dsRNA incorporation [3, 42]. As this research shows, DICER could process pre-miRNAs, longer dsRNAs and hairpin RNAs with 5' overhangs into dsRNA with 3' overhangs, which might be subsequently loaded with canonical length to the RISC. Further experimentation is required to connect our findings with the AGO loading mechanism.
We show human rDICER recognizes and processes a hairpin RNA bearing a trinucleotide-5' overhang, and the two-step cleavage by rDICER forms canonical miRNA duplexes from the hairpin RNAs. It indicates that human rDICER functions as a molecular ruler by anchoring the 3' end of both the hairpin RNA with 5' overhangs and the 5' strand in the intermediate duplex. Moreover, an endogenously-expressed pre-miRNA with 5' overhangs, pre-mmu-mir-1982, also can be utilized as a substrate of rDICER and processed into a canonical miRNA duplex by the two-step cleavage reaction. While pre-mmu-mir-1982 RNA is a naturally expressed pre-miRNA [33, 35], this 5'-overhanged structure is not a suitable substrate for nuclear export by Exportin-5  and, assuming the absence of possible alternative export pathways, may not be presented to cytoplasmic DICER in the cells. However, it is worth noting a recent report, that mammalian DICER might be located in the nucleus and associate with ribosomal DNA chromatin . We have also observed human DICER localized in both cytoplasm and nucleus (unpublished data, Ando et al.). These findings raise the intriguing possibility that nuclear DICER could process hairpin RNA with 5'-overhangs, like pre-mmu-mir-1982 RNA.
The two-step cleavage of a hairpin RNA with 5' overhangs shows that rDICER can release dsRNAs after the first cleavage and binds them again in the inverse direction for a second cleavage. The DICER protein's ability to release and bind dsRNA again indicates DICER could be capable of binding and processing dsRNA multiple times during short RNA maturation. DICER has recently been linked to the processing of diverse non-coding RNA precursors with as-yet undetermined structures. The experiments performed above suggest DICER has considerable flexibility in processing precursors, contributing to an ability to generate various short RNA products for incorporation into functional RISCs.
Preparation of hairpin RNA substrates
Pre-hsa-mir-21 RNA (pre-mir-21), pre-miRNA mimic hairpin RNA (RNA-I) and pre-mmu-mir-1982 RNA used in this study were generated by in vitro transcription using the Ampliscribe T7 High Yield Transcription kits (Epicentre) according to manufacturer's instructions. We made double-stranded DNA templates with T7 RNA polymerase promoter sequence by overlap-PCR using the following oligonucleotide pair; pre-mir-21-sense 5'-taatacgactcactatagAGCTTATCAGACTGATGTTGACTG-3' and pre-mir-21-antisense 5'-ACAGCCCATCGACTGGTGTTGCCATGAGATTCAACAGTCAACATC-3', RNAI-sense 5'-taatacgactcactatagggTGTCGGGTAGCTTATCAGACTGATGTTGA-3' and RNAI-antisense 5'-TGTCAGACAGCCCATCGACTGGTGTTGCCATGAGATTCAACAGTCAACA-3', pre-mmu-mir-1982-sense 5'-taatacgactcactataGTTGGTATTGCTTGGGAGGGTCCTGGGGAGGGGAGTT-3' and pre-mmu-mir-1982-antisense 5'-CTGTGGGAGAACATAGGGTGAGAGGTTGGGGTGCCAGAACTCCCCTCCCCA-3'. The overlapped sequences are underlined and the lower-case characters show the sequence of the T7 RNA polymerase promoter. In vitro transcription reactions were performed at 37°C overnight. Transcripts were run on 10% denaturing polyacrylamide gels in 0.5x TBE (45 mM Tris-borate, 1 mM EDTA), gel-excised, eluted from the gel in 1 M NaCl at 4°C overnight, and precipitated with ethanol. The pellet was resuspended in an appropriate volume of water and stored into the freezer at -30°C. Before use, RNA substrates were heated to 70°C for 5 min and then slowly cooled to room temperature.
Affinity purification of recombinant FLAG-DICER fusion proteins
We assembled a full-length cDNA of human DICER1 protein from HeLa total RNA. This cDNA sequence was identical to the coding sequence cited in the Swiss-Prot Protein Database (Swiss-Prot) [Swiss-Prot: Q9UPY3]. N-terminally FLAG-tagged human DICER1 protein was purified from 293T cells transfected with the plasmid pCA-FLAG-DICER1. This vector contained the full-length human DICER1 protein FLAG-tagged at the amino terminus in a pCA-FLAG-DEST vector . We purified the recombinant FLAG-DICER1 fusion protein (rDICER) using ANTI-FLAG M2-Agarose Affinity Gel (Sigma) and eluted by 0.1 M Glycine-HCl (pH3.5). Then, the eluate was neutralized by Tris-HCl (pH8.0). The average yield was 50-100 μg of the active form of rDICER protein from 1 × 108 culture cell. Purified rDICER protein was detected by Coomassie Brilliant Blue (CBB) staining and Western blotting using anti-DICER (H212, Santa Cruz) antibody to check for successful homogenous purification (see Figures 1A and 1B). The contamination of known DICER-binding proteins in rDICER samples was checked by Western blotting using anti-FLAG (M2, Sigma), anti-AGO2 (07-590, Upstate) and anti-TRBP (ab42018, Abcam) antibody, respectively (see Figure 1C).
Processing of RNA substrates using recombinant DICER enzyme
The affinity-purified rDICER protein (2 pmol) was incubated with 45 pmol of RNA substrates (pre-mir-21 RNA, RNA-I or pre-mmu-mir-1982 RNA) in 1x reaction buffer (300 mM NaCl, 50 mM Tris-HCl, 20 mM HEPES, 5 mM MgCl2, pH 9.0) and 40 units of RNaseOUT (Invitrogen). These mixtures were incubated at 37°C for the indicated times. The reactions were purified by phenol-chloroform extraction followed by sodium acetate-ethanol precipitation at -20°C. The RNA pellet was resuspended in water at a final concentration of approximately 1 pmol/μl.
rDICER-processed RNAs (1 pmol) were separated on 7 M urea-denaturing 20% polyacrylamide gels, then blotted onto Hybond-N+ membranes (GE Healthcare) using a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad). Hybridization was performed in Church buffer (0.5 M NaHPO4, pH 7.2, 1 mM EDTA and 7% SDS) containing 106 c.p.m./ml of each 32P-labelled probe for 14 h. The membranes were washed in 2x SSC, and the signals were detected by autoradiography. All experiments were repeated and replicated consistently.
The probe sequences in this study were as follows: probe-1 (5'-TCAACATCAGTCTGATAAGCTA-3'), probe-2 (5'-ACAGCCCATCGACTGGTGTTG-3'), probe-3 (5'-CCATGAGATTCAACAG-3'), probe-4 (5'-CCTCCCCAGGACCCTCCCAA-3') and probe-5 (5'-CTGTGGGAGAACATAGGGTGAGA-3'). The probes were 5'-end labelled using T4 polynucleotide kinase (TaKaRa Bio) with [γ-32P] ATP (6000Ci/mmol) at 37°C for 4 h.
Cloning of cleavage products
rDICER-processed RNAs (1 pmol) were separated on 7 M urea-denaturing 15% polyacrylamide gels, then the gel was stained by SYBR Gold (Invitrogen). The band around 23 nt was excised from the gel and purified as described above. The purified RNA was cloned by the Small RNA cloning kit (TaKaRa Bio) and sequenced by capillary sequencing.
5'-end labelling of the transcript
For the 5'-end labelling, RNA (45 pmol) was dephosphorylated with CIP at 37°C for 60 min. The reaction was inactivated by phenol-chloroform extraction and precipitated by sodium acetate-ethanol at -20°C. The pellet was resuspended in an appropriate volume of water. The dephosphorylated transcript was 5' end-labelled using T4 polynucleotide kinase (TaKaRa Bio) with [γ-32P] ATP (3000Ci/mmol) at 37°C for 4 h. The 5'-end labeled transcript was PAGE-purified as described above and the RNA pellet was resuspended in water at a final concentration of approximately 0.5 pmol/μl. One microliter of this was used for the processing reaction by rDICER. These processed samples were run on 7.5 M urea-denaturing 20% polyacrylamide gels in 1x TBE buffer with RNA molecular marker or the products of alkaline hydrolysis of the same RNA molecule. The alkaline hydrolysis ladder was generated by incubating the labelled RNA in alkaline hydrolysis buffer (Ambion) at 100°C for 10 min. The signals were detected by autoradiography and quantified using ImageJ software (National Institutes of Health; http://rsb.info.nih.gov/ij/). The signal intensities were calculated as the mean of pixel value of selected area.
We thank Drs. Yasuhiro Tomaru, Timo Lassmann and Masanori Suzuki for their helpful discussion. We also thank Dr. Junichi Yano (Nippon Shinyaku Co. LTd., Kyoto) for the gift of materials. We acknowledge to RIKEN GeNAS for their support of our sequencing data production.
This work was supported by a Research Grant for the RIKEN Omics Science Center from the Ministry of Education, Culture, Sports, Science and Technology of Japan to YH.
- 21.Burroughs AM, Ando Y, de Hoon MJ, Tomaru Y, Nishibu T, Ukekawa R, Funakoshi T, Kurokawa T, Suzuki H, Hayashizaki Y: A comprehensive survey of 3' animal miRNA modification events and a possible role for 3' adenylation in modulating miRNA targeting effectiveness. Genome Res. 2010, 20 (10): 1398-1410. 10.1101/gr.106054.110PubMedCentralCrossRefPubMedGoogle Scholar
- 23.Sakurai K, Amarzguioui M, Kim DH, Alluin J, Heale B, Song MS, Gatignol A, Behlke MA, Rossi JJ: A role for human Dicer in pre-RISC loading of siRNAs. Nucleic Acids Res. 2010Google Scholar
- 40.Burroughs AM, Ando Y, Hoon ML, Tomaru Y, Suzuki H, Hayashizaki Y, Daub CO: Deep-sequencing of human Argonaute-associated small RNAs provides insight into miRNA sorting and reveals Argonaute association with RNA fragments of diverse origin. RNA Biol. 2011Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.