RNAi Factors are Present and Active in Human Cell Nuclei

Summary RNAi is widely appreciated as a powerful regulator of mRNA translation in the cytoplasm of mammalian cells. However, the presence and activity of RNAi factors in the mammalian nucleus has been the subject of considerable debate. Here we show that Argonaute-2 (Ago2) and RNAi factors Dicer, TRBP and TRNC6A/GW182 are in the human nucleus and associate together in multi-protein complexes. Small RNAs can silence nuclear RNA and guide site-specific cleavage of the targeted RNA, demonstrating that RNAi can function in the human nucleus. Nuclear Dicer is active and miRNAs are bound to nuclear Ago2, consistent with the existence of nuclear miRNA pathways. Notably, we do not detect loading of duplex small RNAs in nuclear extracts and known loading factors are absent. These results extend RNAi into the mammalian nucleus and suggest that regulation of RNAi via small RNA loading of Ago2 differs between the cytoplasm and the nucleus.

Lipofectamine RNAiMAX (Invitrogen) was used to deliver siRNAs into HeLa cells following the manufacturer's recommended protocol in OptiMEM low serum medium (Invitrogen). Growth media was changed to full medium after 24 h. Transfected cells were harvested 72 h after transfection for qPCR and RACE analyses, 36 h after transfection for in vitro Ago2 cleavage assays, and 48 h after transfection for FISH analysis. Sequences of siRNAs used are listed in Table S1.

Nuclear and cytoplasmic cell fractions
Cells were harvested with trypsin-EDTA solution (Invitrogen), washed with PBS then resuspended in ice-cold hypotonic lysis buffer (HLB) (10 mM Tris-HCl, pH 7.5, 10 mM NaCl, 3 mM MgCl 2 , 0.3% NP-40) supplemented with 1% Protease Inhibitor Cocktail Set I (Calbiochem), 1 mM sodium fluoride and 1 mM sodium orthovanadate at a final of 1 mL/75 mg wet cell pellet. After incubation on ice for 15 min and pipetting and vortexing, lysate was spun at 4°C at 800xg for 5 min. The supernatant was kept as cytoplasmic extract and NaCl and glycerol were added to a final of 140 mM and 10%, respectively. Pelleted nuclei were washed 3x with ice-cold HLB by 5 min incubation on ice, pipetting and vortexing, then spinning at 4°C at 100xg for 2 min. For nuclear extracts, nuclei were resuspended in ice-cold nuclear lysis buffer (NLB) (20 mM Tris-HCl, pH 7.5, 0.15 M NaCl, 3 mM MgCl 2 , 0.3% NP-40, 10% glycerol) supplemented with 1% Protease Inhibitor Cocktail Set I, 1 mM sodium fluoride and 1 mM sodium orthovanadate at a final of 0.5 mL/75 mg of original wet cell pellet weight (1/2 the volume of cytoplasmic fraction). Nuclei were sonicated on ice at 20% power 3x for 15 sec in 4 mL volumes. After high speed centrifugation at 4°C to remove insoluble cell debris, the soluble fraction was kept as nuclear extract. All extracts were aliquoted, flash-frozen in liquid nitrogen, then stored at -80°C for later use.

Optimization of nuclei isolation protocol
A protocol similar to the above described for nuclear and cytoplasmic fraction preparation was followed except detergent was either omitted or TWEEN-20, NP-40 or Triton X-100 nonionic detergents were included in HLB at the indicated concentrations. After the final nuclei wash, a fraction of the nuclei were aliquoted for visualization by fluorescence microscopy while the remaining nuclei were resuspended in NLB and sonicated to prepare nuclear extracts as described above. Nuclei were prepared for microscopy by washing 1x in ice-cold PBS then incubating in PBS + 1 µM ER Tracker Red (Invitrogen) for 20 min on ice. Nuclei were diluted 10-fold in ice-cold PBS + 4% paraformaldahyde and incubated on ice 10 min.
Nuclei were resuspended by pipetting and spotted on glass slides. After partial air-drying, one drop of Vectashield Hard Set Mounting Medium with DAPI (Vector Laboratories, H-1500) was added, a coverslip added and mounting media allowed to harden at room temperature for 15 min.
Nuclei were visualized with a 60x objective lens and DAPI and TRITC filters on a widefield epifluorescence Deltavision microscope. Z-sections were taken at 0.15 µm thickness.
Images were deconvoluted by blind deconvolution using AutoQuant X3 (Media Cybernetics), stacked and ER tracker staining pseudo-colored yellow in ImageJ for visualization.

Western blot analysis
Cell extracts were prepared as described above. For comparing nuclear and cytoplasmic fractions by Western blot, the same cell equivalents of extract were separated by electrophoresis (1/2 the volume of nuclear extract for every 1 volume of cytoplasmic). In general, loading equal amounts of total protein is unsatisfactory for comparing the nuclear and cytoplasmic levels of specific proteins since there is approximately 4-fold more total protein in cytoplasmic extracts. Protein was separated on 4-20% gradient SDS-PAGE TGX pre-cast gels (Biorad) at 100 V for 75 min. After gel electrophoresis, proteins were transferred to nitrocellulose membrane (Hybond-C Extra, GE Healthcare Life Sciences) at 100 V for 90 min. Membranes were blocked for 30 min at room temperature with 5% milk protein in PBS + 0.05% TWEEN-20 (PBST).
After primary antibody incubation, membranes were washed 3x for 5 min at room temperature with PBST then incubated for 30-45 min at room temperature with HRPconjugated anti-mouse at 1:10000 (Jackson Laboratories, 715-035-150) or anti-rabbit at 1:5000 (Jackson Laboratories, 711-035-152) in PBST + 5% milk. Membranes were washed again 3x for 15 min in PBST at room temperature, then protein bands visualized using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific). For quantification of RNAi factor protein levels from Western blots of cellular fractions, films were scanned and bands quantified using ImageJ.

Chromatographic and ammonium sulfate fractionation of cell extracts
Nuclear extracts were prepared as described above but concentrated 2-fold by using 1/2 the standard amount of NLB during nuclei resuspension and sonication. For size-exclusion chromatography, 2 mL of nuclear extract was either treated with 50 µg of RNase A or 200 units of SUPERase-In (Ambion) for 1 h at room temperature. Samples were then 0.45 µm filtered and injected onto a Superdex 200 HiLoad 16/60 FPLC column (Amersham Pharmacia) that was pre-equilibrated with FPLC buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 3 mM MgCl 2 , 5% glycerol). Protein was eluted off the column with FPLC buffer and the elution collected in fractions at 4°C that were then snap-frozen in liquid nitrogen and stored at -80°C. Western blot analysis of fractions was performed as described above. For subsequent fractionation by anion exchange, size-exclusion fractions were concentrated 3-fold and diluted to 0.1 M NaCl and injected onto a Mono-Q FPLC column (Amersham Pharmacia) equilibrated with FPLC buffer at 0.1 M NaCl. Elution was performed at room temperature with a linear gradient from 0.1 to 1 M NaCl in FPLC buffer. Fractions were collected and snapfrozen in liquid nitrogen. Western blot analysis was performed as described above.
Ammonium sulfate cuts were performed by addition of saturated ammonium sulfate solution to extract up to the indicated percentage and incubated on ice for 15 min.
Precipitated protein was pelleted by centrifugation at 18,000xg for 20 min at 4°C. Supernatant was kept and additional ammonium sulfate added up to the next indicated percentage and incubation and centrifugation repeated. Precipitated protein pellets were resuspended in identical volumes of SDS loading buffer, resolved by SDS-PAGE and probed by Western blot as described above.

Analysis of siRNA-mediated RNA knock-down in cellular compartments
HeLa cells were transfected in quadruplicate with 25 nM siRNA then harvested 72 h later with trypsin-EDTA solution. Cells were washed with PBS then counted with a hemocytometer.
Pelleted chromatin was washed 3x with ice-cold MWS then 1 mL Trizol (Invitrogen) was added to the final chromatin pellet. To process cytoplasmic and nucleoplasmic fractions, precipitates were pelleted at 18000xg for 15 min at 4°C then 1 mL Trizol added to each pellet.
Samples in Trizol were heated to 70°C with vortexing until completely dissolved, then cooled to room temperature. To each sample 0.2 mL chloroform:isoamyl alcohol (24:1) was added, samples vortexed, then spun at 18000xg for 10 min. The top aqueous layer was collected and RNA precipitated by addition of 1 volume of isopropanol and incubation at -20°C overnight. RNA was pelleted by spinning at 18000xg, washed with 70% ethanol, then air dried and prepared for quantitative PCR or 5' RACE.

Quantitative PCR
Identical volumes of RNA (representing approximately the same number of cells and ranging from 1-2 µg of RNA) that were prepared from cellular fractions above were treated with 2 units of DNase I (Worthington) in 9.5 µL of DNase I buffer (10 mM Tris-HCl, pH 7.0, 10 mM NaCl, 2 mM MgCl 2 , 0.5 mM CaCl 2 ) for 15 min at room temperature to degrade any genomic DNA contamination. Afterwards, 0.5 µL of a 50 mM EDTA, 10 mM EGTA solution was added and DNase I heat-inactivated at 70°C for 10 min. Treated RNAs were reverse-transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) in a final volume of 20 µL. Quantitative PCR (qPCR) was performed using iTaq Supermix with ROX (Biorad) with ~10-20 ng of cDNA as template.
Data were normalized relative to measured GAPDH levels in each cellular compartment. Because no normalization control exists that is the same level across cellular compartments, and because spike-in controls can have variability in performance (data not shown), comparisons among treatments were only performed within each cellular compartment and not across cellular compartments (ie. chromatin to chromatin, not cytoplasm to chromatin). Primers used in qPCR are listed in Table S1.

Rapid Amplification of cDNA Ends (RACE)
RACE was performed using the GeneRacer Kit (Invitrogen) following the manufacturer's recommended protocol. cDNA was prepared from ~1 µg RNA from each cellular fraction (prepared as described above) by reverse-transcription (RT) reaction using random primers.
The 5' end of cDNA was amplified using Platinum Taq DNA polymerase (Invitrogen) and specific primer sets for Malat1 and RPL30 (Table S1). The thermal cycling condition was 94°C for 2 min, followed by 5 cycles of 94°C for 30 sec and 72°C for 1 min, 5 cycles of 94°C for 30 sed and 70°C for 1 min, and 25 cycles of 94°C for 30 sec, 65-66°C for 30 sec, and 68°C for 1 min, followed by final extension of 68°C for 10 min. PCR products were analyzed on 1.2-1.5% agarose gel (Fig. S6B-C). Major PCR products on gels were excised and cloned intopCR 4-TOPO vectors, transformed into TOP10 chemically competent cells, then sequenced to map the 5' cleavage sites. Chromatin-associated cleavage products for RPL30 were below our detection limit (Fig. S6C), despite performing nested PCR with multiple primer sets (data not shown). This result may reflect the relatively low level of RPL30 mRNA actually associated with chromatin. In addition to relative starting levels of targeted RNA, detection of cleavage products will also depend upon other factors like the rate of cleavage product formation and degradation.

In vitro Ago2 cleavage assay
HeLa cells were either untreated or transfected with 25 nM siLuc or siLuc_mm (see Table S1) then harvested 36 h later and nuclear and cytoplasmic extracts prepared as described above.
Reactions were incubated at 30°C for 1.5 hr with periodic mixing and RNA collected by addition of 0.5 µL of 0.5 M EDTA then phenol extracted. Extracted RNA was precipitated with 9 volumes of 2% LiClO 4 in acetone, washed with acetone, resuspended in 90% formamide, 1x Tris-Borate EDTA (TBE) buffer, boiled for 3 min, then resolved on a 15% denaturing polyacrylamide (7 M urea, 1x TBE, 2% glycerol) sequencing gel. The gel was dried and exposed to a phosphorimager screen overnight to visualize radioactive bands. Target

Immuno-fluorescence and co-localization analysis
Immuno-fluorescence was performed similarly to that previously described (Ohrt et al., 2012; Spector, 2011) with modifications. Briefly, cells were grown for 16-24 h to 50-70% confluency on 35 mm dishes (MatTek Corporation, P35GCOL-1.5-14-C) with a 14 mm glass bottom of 1.5 mm thickness. Cells were washed with PBS then fixed in freshly made 2% formaldehyde in PBS or 4% paraformaldehyde in PBS for 15 min at 20°C, or in 70% ethanol for at least 30 min at 4°C. Fixed cells were washed 3x in PBS for 10 min at 4°C. Cells were permeabilized in PBS containing 0.2% Triton X-100 and 1% normal goat serum (NGS) for 5-10 min on ice then washed 3x with ice-cold PBS + 1% NGS at 4°C for 10 min.
The cells were incubated in primary antibody for 1 hr at room temperature or overnight at 4°C. Primary antibodies were diluted in PBS + 1% NGS and incubated for 1 h at room temperature or at 4°C overnight: anti-Ago2 at 1:100 (Abcam, ab57113), anti-Ago2 at Cells were imaged with a 60x objective lens and DAPI, FITC and TRITC filters on a wide-field epifluorescence Deltavision deconvolution microscope. Z-sections were taken at 0.1, 0.15 or 0.2 µm for at least 6 µm thickness. Images were blind deconvoluted using AutoQuant X3 (Media Cybernetics) and stacked and analyzed in Imaris (Bitplane).
Alternatively, some samples were imaged by Andor spin disc confocal microscopy. The Colocalization channel of Ago2 (FITC, green) and TNRC6A (TRITC, red) was calculated using Imaris software based on the correlation of the strength of linear relation between the red channel and the green channel and the threshold levels for calculation of co-localization were selected above background signals.

Fluorescence in situ hybridization (FISH)
Cells were grown for 16-24hrs to 50-70% confluence on 35mm MatTek dishes as described for immuno-fluorescence above. Cells were transfected with 25 nM siLuc or siMalat1 as Vectashield Hard Set Mounting Medium with DAPI (Vector Laboratories, H-1500) was then added, covered with a coverslip and allowed to harden for 15 min at room temperature. Cells were imaged and analyzed the same as described above for immuno-fluorescence but using DAPI and Cy-5 filters. Z-stacks were taken with 0.15 or 0.2 µm slices and 6 µm thickness.

Small RNA-seq library preparation and sequencing
For whole cell small RNA sequencing, T47D cells were harvested and then dissolved in TRIzol (Sigma). For sequencing of small RNA from cell nuclei, pure nuclei were isolated first as described above. Nuclei were then dissolved in TRIzol. Small RNA was isolated using miRNeasy Mini Kit (Qiagen) and then treated with DNase I (Worthington) for 20 min at 37°C to remove any contaminating genomic DNA. The small RNA-seq library was made using Illumina Small RNA Truseq kit following the manufacturer's recommended protocol. RNA-seq libraries were sequenced on a Illumina HiSeq 2000 as per manufacturer's instructions for single-end 1x50. Approximately, 30 million raw reads (averaged over two replicates) were obtained per sample. Reads with low quality score were removed and reads that passed the filtration were trimmed by removing the adaptor sequence. Trimmed reads shorter than 15 nt were also excluded from analysis. Filtered and trimmed reads were aligned to the human genome (hg19) using Bowtie2 by allowing up to two mismatches to the reference sequence.
Up to 10 different alignments per read were permitted. For reads that were aligned to multiple positions in the reference genome, the single aligned read with the fewest mismatches was selected using a Perl script. If multiple reads still remained, the original read would be disregarded. Approximately 70% of raw reads were successfully aligned. Finally, the aligned reads were again mapped to the UCSC miRNA database and miRBase (mature miRNAs) to search for possible miRNA hits.

Sequencing of AGO2-associated small RNA in cell nuclei
RNA Immunoprecipitation (RIP) assay using nuclear lysate (T47D cells) was performed as described (Chu et al., 2010;Matsui et al., 2013) using anti-human Ago2 antibody (Wako Chemical, 015-22031). A non-specific mouse IgG antibody was used as a control. The isolated RNA was loaded on a 10% polyacrylamide gel and the small RNA fraction (15nt-40nt) was cut out and extracted. The small RNA was then subjected to polyadenylation by using a poly(A) tailing kit (Ambion). 3′-deoxy-ATP (cordycepin triphosphate, Jena Biosciences) was introduced 10 min after the initiation of the polyadenylation reaction for 3′end blocking and tail length limitation (performed by Helicos, Inc.). The Direct RNA Sequencing (DRS) was carried out on a single molecule Helicos sequencer. Raw sequencing data was filtered to remove low quality reads and subsequently aligned using the Helisphere package, a software designed to specifically analyze sequencing data generated from the Helicos sequencer. The aligned reads were again mapped to the UCSC miRNA database and miRBase (mature miRNAs) to search for possible miRNA hits.

In vitro Dicer processing assay
Dicer was immunoprecipitated using 3 µg of anti-Dicer (Abcam, ab14601) antibody (or 3 µg normal mouse IgG as control), 50 µL Protein G Plus/Protein A agarose (Calbiochem) and 200 µL of HeLa nuclear or cytoplasmic extract (prepared as described above) with rotation at room temperature for 1 h. Resin was washed 3x with 0.5 mL IPWB 300 (20 mM Tris-HCl, pH Reactions were incubated at 32°C for 1 hr with periodic mixing and RNA collected by addition of 0.5 µL of 0.5 M EDTA and 5 µg yeast tRNA then phenol extracted. Extracted RNA was precipitated with 9 volumes of 2% LiClO 4 in acetone, washed with acetone, resuspended in 90% formamide, 1x Tris-Borate EDTA (TBE) buffer, boiled for 3 min, then resolved on a 15% denaturing polyacrylamide (7 M urea, 1x TBE, 2% glycerol) sequencing gel. The gel was dried and exposed to a phosphorimager screen overnight to visualize radioactive bands. Pre-miR-19a sequence from miRBase was used prepare in vitro transcription template. DNA seqeunces below were annealed and used for in vitro transcription (Epicentre Ampliscribe T7 In Vitro Transcription Kit) following the manufacturers protocol.

GCAGGCCACCATCAGTTTTGCATAGATTTGCACAACTACATTCTTCTTGTAGTGCAACTA TGCAAAACTAACAGAGGACTGCtatagtgagtcgtattag
In vitro Ago2 small RNA loading assay Duplex siRNA or single-strand guide RNA radiolabeled at the 5' end was incubated with cell extracts supplemented with 1 mM ATP for 1 h at room temperature with rotation. For reactions using mismatch-containing siLuc, the standard antisense siLuc strand was radiolabeled and annealed to sense strand with the indicated base position mismatches (see Table S1) then gel-purified before use. When indicated, phosphocreatine (10 mM) and creatine kinase (100 µg/mL) were added as an ATP regeneration system. Ago2 was immunoprecipitated using 2µg of anti-Ago2 antibody (Abcam, ab57113) and 40 µL of Protein G Plus/Protein A agarose (Calbiochem) with rotation for 1 h at room temperature. Resin was washed 3x with 0.8 mL IPWB 500 (20 mM Tris-HCl, pH 7.5, 4 mM MgCl 2 , 0.5 M NaCl, 0.05% NP-40) and 1x with 0.5 mL IPDB (20 mM Tris-HCl, pH 7.5, 3 mM MgCl 2 , 0.15 M NaCl) at room temperature. Co-precipitated RNA was collected by addition of 0.5 µL of 0.5 M EDTA and 5 µg yeast tRNA then phenol-chloroform extraction. Extracted RNA was precipitated with    . Gels for separation of RACE products are shown with the cleavage product band highlighted. Cleavage product bands were excised and sequenced to unambiguously validate sequence-specific siRNAdirected cleavage at the expected phosphodiester bond of the targeted RNA. 5' ends of sequenced RACE products are shown below each gel for each cell fraction. RACE was performed at least twice and cloning and sequencing performed several times per band for each experiment. HeLa cells were treated with (C) a mock siRNA (siLuc) or (D) siRNA against Malat-1 (siMalat1) then processed for FISH analysis (see Methods) 48 h later. Z-stacks 6 µm thick were taken. Figure S5, Related to Figure 6: Controls for Ago2 loading deficiency in the human cell nucleus. (A) Incubation of single-stranded siLuc guide RNA with T47D cytoplasmic and nuclear extracts followed by Ago2 immunoprecipitation, or with Ago2 after immunoprecipitation. Single-stranded guide RNA does not efficiently bind Ago2 in extracts presumably due to rapid degradation. If Ago2 is first immunoprecipitated to wash away contaminating nucleases, single-stranded guide RNA readily binds and co-immunoprecipitates. (B) Degradation kinetics of duplex siLuc in HeLa cytoplasmic or nuclear extracts. (+) SUPERase-In = addition of 50 unites of SUPERase-In RNase inhibitor. (C) Ago2 in vitro loading assay using siLuc siRNA dual-labeled on both guide and passenger strands. (D) Immunoprecipitation of Ago2 from HeLa cytoplasmic and nuclear extracts followed by coomassie staining shows equivalent amounts of Ago2 capture from both extracts. (E) Ago2 in vitro loading assay using radiolabeled siLuc and the addition of ATP regeneration system (see Extended Methods) in cytoplasmic or nuclear extracts.