Second-generation shRNA libraries covering the mouse and human genomes

Jose M Silva^1, 4, Mamie Z Li^2, 4, Ken Chang^1, 4, Wei Ge³, Michael C Golding¹, Richard J Rickles², Despina Siolas¹, Guang Hu², Patrick J Paddison¹, Michael R Schlabach², Nihar Sheth¹, Jeff Bradshaw³, Julia Burchard³, Amit Kulkarni³, Guy Cavet³, Ravi Sachidanandam¹, W Richard McCombie¹, Michele A Cleary³, Stephen J Elledge² & Gregory J Hannon¹

¹ Cold Spring Harbor Laboratory, Watson School of Biological Sciences, 1 Bungtown Road, Cold Spring Harbor, New York 11724, USA.

² Department of Genetics, Center for Genetics and Genomics, Brigham and Women's Hospital, Harvard Medical School, Room 158D, NRB, 77 Avenue Louis Pasteur, Boston, Massachusetts 02115, USA.

³ Rosetta Inpharmatics LLC, 401 Terry Ave. North, Seattle, Washington 98109, USA.

⁴ These authors contributed equally to this work.

The RNAi machinery can be programmed by exogenous or endogenous sources of double-stranded RNA. The most well-characterized source of endogenous triggers is microRNA (miRNA) genes^{6, 7}. It was initially assumed that miRNAs were transcribed from the genome as short hairpin RNAs⁸ (shRNAs) that were directly processed by Dicer to yield the mature small RNAs that enter RISC^{9, 10, 11, 12}. Over the past year, however, a different picture has emerged. Numerous studies have shown that, in animals, miRNAs are transcribed by RNA polymerase II to generate long primary polyadenylated RNAs (pri-miRNAs)^{13, 14}. Through mechanisms not yet fully understood, the pri-miRNA is recognized and cleaved at a specific site by the nuclear Microprocessor complex^{15, 16, 17, 18, 19}. This contains an RNase III family enzyme, Drosha, that cleaves the hairpin to produce a miRNA precursor (pre-miRNA) of 70–90 nucleotides (nt) with a 2-nt 3' overhang¹⁵. This distinctive structure signals transport of the pre-miRNA to the cytoplasm by a mechanism mediated by Exportin-5 (refs. 20,21). Only then is the pre-miRNA recognized by Dicer and cleaved to produce a mature miRNA. This probably involves recognition of the 2-nt 3' overhang created by Drosha to focus Dicer cleavage at a single site 22 nt from the end of the hairpin^{22, 23}.

Expression of a simple, 29-bp hairpin from a U6 small nucleolar RNA (snRNA) promoter can induce effective suppression of target genes when delivered either transiently or stably from integrated constructs^{26, 29, 30}. Longer hairpin structures are more effective inhibitors of gene expression than are shorter structures with stems of 19–21 nt. All these constructs, however, were designed to express a pre-miRNA hairpin, an intermediate in miRNA biogenesis, rather than a transcript that closely resembles a primary miRNA. Effective suppression could be achieved by redesigning an endogenous miRNA, miR-30, such that its targeting sequence was directed against a reporter gene²⁸. We sought to compare directly the abundance of small RNAs produced from vectors with simple hairpin structures versus those that more closely resemble a natural miRNA. Because the efficient ectopic expression of endogenous miRNAs requires substantial flanking sequence³¹, we developed a vector in which sequences from a remodeled miR30 were flanked by 125 bases of 5' and 3' sequence derived from the primary transcript. Incorporation of appropriate cloning sites into this vector required altering only three positions in the precursor. We inserted this cassette into a vector equivalent to that in which we constructed our first-generation shRNA library (pSM1) and called the new shRNA vector pSM2. To distinguish the second-generation shRNAs from those in our first-generation library, we call them shRNA-mirs.

Figure 1. Design and structure of shRNA-mir cassettes. (a) The structures of several silencing triggers, including an siRNA, a portion of the shRNA precursor (as generated from our first-generation design in pSM1) and a segment of the shRNA-mir precursor produced by pSM2, are compared. The sequence of the target site (sense orientation) from firefly luciferase (luc1309, see c) is shown in blue (passenger strand), with the guide strand shown in red. For pSM2, mapped potential cleavage sites for Dicer and Drosha are indicated by blue and red lines, respectively. (b) Northern blotting was used to detect the mature small RNA produced after transfection of HEK 293 cells with shRNA and shRNA-mir cassettes expressed from pSM1 and pSM2, respectively, by the U6 snRNA promoter, but no substantial accumulation of pre-miRNA was observed. Transfection rates were normalized using a codelivered DsRed expression plasmid. (c) The ability of five different promoters (human tRNAval, human H1 RNA, human U6 snRNA, MSCV LTR and human CMV IE) to drive shRNA-mir expression and to silence luciferase in transient transfections was tested. Two different shRNAs were used, a highly efficient shRNA (luc1309) and a less efficient shRNA (luc311). In each case, the level of firefly luciferase was normalized to that of a nontargeted Renilla luciferase. Controls with empty vectors lacking a hairpin insert are also shown. Full Figure and legend (39K)

To test the performance of pSM2 relative to pSM1, we used both vectors to express a sequence targeting firefly luciferase. We inserted the sequence such that an identical mature small RNA would be generated from each construct after processing in vivo (Supplementary Fig. 2 online). Of primary concern was the overall amount of mature small RNA generated from each construct, because dose-response experiments for shRNAs indicate that suppression correlates very well with the amount of RNA delivered²², particularly at the relatively low doses expected to be achieved by expression from transfected or integrated constructs as compared with directly transfected synthetic RNAs. We transfected 293 cells with pSM1-luc and pSM2-luc, prepared RNA and assayed the processed small RNA by northern blotting. Cells transfected with pSM2-luc contained 12 times more small RNA than did cells transfected with pSM1-luc (Fig. 1b).

Figure 2. Construction of the second-generation library. (a) The pSM2c vector contains a U6 promoter, a U6 terminator after miR3', a self inactivating retroviral backbone; two bacterial antibiotic resistance markers (kanamycin and chloramphenicol); a protein-dependent origin (RK6); a mammalian selectable marker (puromycin) driven by the PGK promoter; a homology region (HR2) for use in bacterial recombination; and a randomly generated 60-mer barcode sequence. The shRNA-mir inserts were cloned between the 5' and 3' flanking sequences derived from the miR-30 primary transcript. The nucleotide positions for sites in an excised version of an empty vector (no shRNA or barcode) are given. (b) Construction of the second-generation libraries began with the generation of a derivative of pSM2. A barcoded prelibrary was generated by the ligation of PCR-amplified random 60-mers into FseI-AvrII–cleaved pSM2 to generate a barcoded library pool (upper right). The barcoded pSM2 was converted into a shRNA library by insertion of PCR-amplified shRNA inserts prepared by in situ synthesis in pools of 22,000 into the EcoRI-XhoI–cleaved prelibrary (upper right). Packaged phage were amplified and used to infect BUN25, which express Cre recombinase and pir1-116 for pSM2 replication. Each excision event gave rise to a KanR CmR colony. These were pooled and used for preparation of library DNA, which was transformed into BW F'DOT. Individual colonies were selected for sequence analysis. Full Figure and legend (72K)

Construction of the library proceeded stochastically using a highly parallel in situ synthesis approach for oligonucleotide production (Fig. 2b). We synthesized groups of 22,000 oligonucleotides, each containing a different shRNA-mir cassette, on glass-slide microarrays³⁹. We eluted populations from the arrays and amplified them by PCR. To ensure efficient cloning, we inserted the pSM2 backbone into a lambda phage backbone such that it was flanked by loxP sites. -pSM2 contains unique XhoI and EcoRI sites for subcloning amplified hairpins and unique FseI and AvrII sites for inserting barcode 60-mers. For -pSM2 barcodes, we first used a mixed library of random 60-nt sequences amplified with a primer set that included one primer with an FseI site and one primer with a T7 promoter followed by the AvrII site. We 'lysogenized' amplified -pSM2 libraries containing barcodes into a strain that we constructed for this purpose, DH10_KP, which has a wild-type pir1 gene and the lambda repressor, cI, to allow -pSM2 to replicate as a 42-kb plasmid. We selected 10⁸ chloramphenicol- (Cm^R) and kanamycin-resistant (Kan^R) lysogens to serve as a barcoded library pool. We purified the barcoded -pSM2 libraries using cesium chloride gradient centrifugation; cleaved them with EcoRI and XhoI; ligated them to gel-purified, EcoRI-XhoI–cleaved, pooled shRNA-mir inserts from an individual chip; and packaged them. The average library size was 5 10⁷ recombinants per pool. To generate pSM2 library plasmid pools, we used the phage to infect an Escherichia coli strain that we constructed, BUN25, which expresses both Cre recombinase and the pir1-116 gene, needed for high-copy RK6 replication. We then transformed pooled plasmid libraries into a mating-competent host strain (BW F'DOT) and sequenced individual clones at random. Between 25% and 50% of clones had perfect inserts; we selected these and saved them as an arrayed set. We monitored accumulation of new clones from each pool dynamically; once a pool began to yield fewer unique clones per sequencing run, we halted sequencing and resynthesized the pool without the sequences that we had already obtained. We reiteratively synthesized 70 chips to maximize unique sequencing. Also, once we obtained three or more verified shRNAs for any given gene, we withdrew the remaining shRNAs targeting that gene from the population selected for resynthesis.

Table 1. Coverage by functional group Full Table

To assess more thoroughly the performance of the second-generation libraries, we compared results from the proteasome assay for 36 shRNA-mir expression plasmids to suppression of target RNAs as measured by semiquantitative RT-PCR. We transfected HeLa cells with plasmids with 80% efficiency, as measured by reference to a cotransfected reporter plasmid. Despite this incomplete transfection, all but 6 of the shRNA-mirs reduced the levels of their target RNAs by 60% or more, and 13 of the 36 shRNA-mirs suppressed their targets by 80%, the theoretical maximum (Fig. 3c and Supplementary Table 1). We obtained similar results with an additional 12 shRNAs that did not target proteasome subunits (data not shown). These studies also indicated that in some cases (e.g., pSMB3), the functional assay did not show a large activation of the reporter despite substantial suppression of the targeted mRNA (Fig. 3c). Therefore, the functional assay slightly underestimated the efficacy of the library.

	Top

	Top

Construction of the lysogenic strain DH10_KP and excision strain BUN25.

Strain DH10_KP [mcrA(mrr-hsdRMS-mcrBC)80 lacZM15lacX74 deoR recA1 endA1 ara139(ara, leu)7697 galU galK ^- rpsL nupG tonA -pir1-npt] containing cI and the pir1 gene was constructed to 'lysogenize' the SM2 barcode library before introduction of the hairpin fragments. To generate this strain, we constructed _KP containing the pir1 and Kan^R genes. To generate _KP, we amplified the pir1 gene from BW23473 using primers MZL393 and MZL51 and cloned it into the pCR2.1 TOPO TA cloning vector. We excised the pir1 gene from this clone on a BamHI fragment and ligated it into BamHI-cleaved pSE356, which contains an npt gene and a BamHI restriction site flanked by two 1-kb DNA fragments⁴² to generate pSE356pirWT. We recombined the pir1 Kan^R fragment onto wild-type by amplifying on LE392/pSE356pirWT, collected the resulting phage and used them to infect DH10. We infected 100 l of DH10 cells with 10⁶ plaque-forming units (PFUs) at 30 °C for 30 min in Luria broth plus 10 mM MgSO₄, diluted them with 900 l of Luria broth, incubated them at 30 °C for 2 h with shaking and plated them on Luria broth containing 50 g ml^-1 kanamycin at 37 °C overnight to select _KP lysogens. We tested the ability of lysogens to lysogenize vectors containing R6K origins of replication as extrachromosomal elements. We selected a strain capable of doing this and named it DH10_KP.

Strain BUN25 [F' traD36 lacI^q (lacZ)M15 proA⁺B⁺/e14^-(McrA^-) (lac-proAB) thi gyrA96 (Nal^r) endA1 hsdR17 (r_k^-m_k⁺) relA1 glnV44 -cre-npt umuC::pir116-Frt sbcDC-Frt] containing pir1-116 and cre was constructed to allow the conversion of SM2 shRNA libraries into pSM2 shRNA libraries. We generated a PCR fragment containing pir1-116 using primers MZL393 and MZL51 (Supplementary Table 3 online), cleaved it with BamHI and ligated it into BamHI-cleaved pUC18 to generate pML284. We isolated a fragment containing BstBI-Frt-cat-Frt-NdeI (filled-in) from KD3 (ref. 43) and inserted it into the SmaI site of pML284 to generate pML334. We isolated an HpaI fragment containing UmuDC from pSE117 and cloned it into pBluescript XhoI (filled in)-EcoRV to generate pML236. We eliminated one of the BamHI site on pML236 by digesting it with PstI-XbaI, filling in with T4 DNA polymerase and ligating. We isolated the Frt-cat-Frt-pir116 sequence from pML334 as a KpnI-SacI (filled-in) fragment and ligated it into MluI-BamHI (filled-in)-cleaved pML236-BamHI to generate pML346. We integrated the 3.8-kb KpnI-SacI UmuDC-Frt-cat-Frt-pir116-UmuC fragment from pML346 into BNN132/pML104 by homologous recombination using the recombinase expressed from pML104 and confirmed it by colony PCR. We removed the cat gene by FLP-mediated excision in vivo using pCP20 (ref. 44), which expresses the FLP recombinase, to generate BUN24. We amplified a cassette that has Frt-cat-Frt flanked by 50 bp of homology to sbcD and 50 bp of homology to sbcC by primers MZL493 and MZL494 using KD3 as a template and used this cassette to replace sbcD and part of sbcC on BNN132 by homologous recombination. The deletions were confirmed by colony PCR. The strain was named BNN132sbcDC-Frt-cat-Frt. We then used a pair of outside primers (MZL495 and MZL496) that gave about 500 bp of homology to the upstream of sbcD and 500 bp of homology to the sbcC to amplify a PCR product from BNN132sbcDC-Frt-cat-Frt to recombine onto the sbcDC region of BUN24. The resulting strain was named BUN25 and is used to stabilize inverted repeats in E. coli⁴⁵.

We inserted a pair of loxP-NotI-loxP duplexed oligos (MZL524 and MZL525) into the pSM2 BstXI site to generate pSM2c-loxP. We inserted a second pair of duplexed oligos (MZL541 and MZL542), carrying the proper restriction sites for cloning barcodes into SM2, into the BbsI-MluI sites of pSM2c-loxP to create pML375. We digested ACT2 with NotI, gel-purified the arms and ligated them to NotI-digested pML375 to generate SM2. We packaged the ligation mixture using MaxPlax lambda packaging extracts from Epicentre. We selected a SM2 lysogen by infecting 200 l of BW23473 cell (A₆₀₀ = 0.8) with 100 l of SM2 packaging mix in the presence of 10 mM MgSO₄ and 0.2% (w/v) maltose, incubated it at 30 °C for 30 min, added 900 l of Luria broth, incubated it at 30 °C for 2 h with shaking to express the Cm^R marker and plated it on Luria broth containing 17 g ml^-1 of chloramphenicol at 30 °C overnight. The proper recombinants were confirmed by restriction analysis. Oligonucleotide sequences are given in Supplementary Table 3.

We amplified the 60-bp barcode oligos using barcode primers 1 and 2 (Supplementary Table 3). For PCR, we used 0.1 pmol of barcodes, 50 pmol of each primer, 25 nmol of each dNTP and 2.5 U of Taq DNA polymerase and the following conditions: 94 °C for 45 s; 13 cycles of 94 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s; 72 °C for 10 min. We pooled ten PCR reactions together, purified them using a QIAquick PCR purification kit, digested them with BamHI, EcoRI, XhoI and SalI to remove these sites in the barcodes and gel–purified them. We digested the purified barcodes with FseI and AvrII and then purified them using the QIAquick gel extraction kit. We mixed 2 g of FseI-AvrII–digested SM2 ligated with 10 ng of FseI-AvrII–digested barcodes with 1 ligation buffer and 0.5 l of T4 DNA ligase in a 5-l final volume and incubated it at 16 °C overnight. We packaged the ligation mixture and amplified it. The size of the SM2-barcode library was 4.2 10⁷. We used 20 ml of DH10_KP cells (A₆₀₀ = 1) to lysogenize 2 10⁹ of SM2-barcode library as 42-kb plasmids. We mixed the cell and the phage in the presence of 10 mM MgSO₄ and 0.2% (w/v) maltose, incubated the mixture at 30 °C for 30 min and added 200 ml of Luria broth to recover at 30 °C for 1 h by shaking. We concentrated the mixture by centrifugation at 4,000 r.p.m. for 20 min, resuspended the pellet in 3 ml of Luria broth, plated the mixture on 10 large Luria broth plates with 17 g ml^-1 of chloramphenicol and incubated it at 30 °C overnight. We scraped the cells from plates and grew them in 3 l of terrific broth containing 17 g ml^-1 of chloramphenicol overnight. We prepared supercoiled SM2-barcode library DNA using cesium chloride. The lysogenization efficiency was 30%.

To collect oligonucleotides, we treated microarrays for 2 h with 2–3 ml of 35% NH₄OH solution (Fisher Scientific) at room temperature. We transferred the solution to 1.5-ml microcentrifuge tubes and subjected it to speed vacuum drying at medium heat (55 °C) overnight. We resuspended the dried material in 200 l of RNase- and DNase-free water and carried out PCR amplification in 50-l reaction volumes using Invitrogen Platinum Pfx DNA Polymerase. To obtain a sufficient amount of PCR product, we did four 50-l reactions for each sample. Each reaction contained 2 Pfx PCR amplification buffer, 0.3 mM of each dNTP, 1 mM MgSO₄, 0.3 M of each primer, 0.5 PCR enhancer solution, 0.5 units of Platinum Pfx DNA Polymerase and 10 l of template DNA. The primers used for amplification were 5'-mir30-PCR-xhoI-F and 3'-mir30-PCR-ecorI-R (sequences available on request). After an initial denaturation step of 94 °C for 5 min, DNA amplification occurred through 25 cycles of denaturing at 94 °C for 45 s and annealing and extension at 68 °C for 1 min and 15 s, followed by a final 7-min extension at 68 °C. We then combined the four reactions into one tube, cleaned up the PCR product using the QIAGEN MinElute PCR Purification Kit and eluted it in a total volume of 26 l.

We digested the SM2 barcode library and shRNA PCR products with EcoRI and XhoI overnight and gel-purified them. We set up ligations with shDNA oligos using 1.5 g of XhoI-EcoRI–cleaved vector, 8–10 ng of XhoI-EcoRI–cleaved inserts generated from the PCR of shDNA oligos from the parallel microarray synthesis, 1 l of 10 ligation buffer, 0.5 l of T4 DNA ligase and water to a final volume of 10 l. We incubated the ligation mixtures at 16 °C overnight and packaged them. We typically observed 30- to 90-fold stimulation of PFUs and 2 10⁷ to 8 10⁷ PFUs total for each library pool. We typically amplified 2 10⁷ PFUs for each pool to generate a stock. To verify the ligation efficiency, we excised 10 l of package mix by infecting 100 l of BUN25 (A₆₀₀ = 0.5) and selected colonies on Luria broth with 30 g ml^-1 of chloramphenicol at 30 °C overnight. We carried out colony PCR using forward and reverse primers (Supplementary Table 3) and typically observed 85–95% correctly sized inserts, with some containing multiple inserts. To generate plasmid DNA from these libraries, we typically excised 5 10⁷ PFUs through infection of BUN25 cells as described earlier. The cells were scraped from plates and grown in 2 l of Luria broth plus 13 g l^-1 of circle growth at 37 °C for 7–8 h. We used the cesium chloride method to prepare DNA. DNAs were transformed into BW23474 F'DOT SbcC, and individual clones were sequenced using primer5' (sequence available on request). Correct clones were individually rearrayed to form the final library.

We transfected 293 cells in 10-cm dishes at 60% confluency with 15 g of shRNA plasmid DNA along with 5 g of pDsRed-N1 (Clontech) using TransIT-LT1 (Mirus). At 48 h after transfection, we confirmed transfection efficiency by estimating the percentage of cells expressing DsRed (80%) and then extracted and purified total RNA using Trizol. We carried out small RNA northern blots as described⁴⁶ using 30 g of RNA per lane. For hairpins targeting EGFP at starting position 481 (Fig. 1b), northern probes were DNA oligos corresponding to the antisense strand (sequence available on request) of the mature RNA.

We grew bacteria cultures in 96-well plates for 36 h in GS96 medium (Bio101). We extracted plasmid DNA using Qiagen Ultrapure plasmids minipreps in a 96-well plate format. We determined DNA concentrations by mixing an aliquot of each sample with picogreen (Molecular Probes) and determining fluorescence on a Victor2 plate reader. We plated 1 10⁵ HEK 293T cells per ml in 96-well optic plates (Corning). For the proteasome assay, we cotransfected cells in each well with 12.5 ng of pDsRed-N1 (Clontech), 5 ng of the Zsprosensor (Clontech) and 75 ng of each individual shRNA construct using 0.3 l of LT-1 (Mirus) transfection reagent. After 24 h, we replaced the transfection medium. After 72 h, we removed the medium and replaced it with phosphate-buffered saline in order to read fluorescence. Fluorescence signals were read on a Victor2 plate reader. Signals in the green channel were normalized to transfection efficiency using customized scripts with fluorescence in the red channel serving as a normalization criterion. Cut-offs were assigned by using control shRNA transfections to determine the range for a negative outcome.

We seeded 0.5 10⁵ HeLa cells per well in 24-well plates and transfected them 24 h later with 1 g per well of the appropriate plasmid. We delivered each plasmid to four wells using Lipofectamine 2000 (Invitrogen) in accordance with the manufacturer's protocols. We determined transfection efficiency by parallel transfection of a GFP-expressing plasmid and assayed the percentage of fluorescent cells by flow cytometry. For analysis of target gene mRNA knock-down, we collected cell lysates 24 h after transfection and prepared total RNA using RNeasy columns (Qiagen) in accordance with the manufacturer's protocols. We quantified mRNA by real-time PCR of reverse transcription products, using available Applied Biosystems TaqMan primer probe sets, and determined the percent mRNA remaining by comparison with mRNA levels from cells treated with transfection reagent alone.

	Top

1. Nakayashiki, H. et al. RNA silencing as a tool for exploring gene function in ascomycete fungi. Fungal Genet. Biol. 42, 275–283 (2005). | PubMed  | ISI | ChemPort |
2. Tang, G. & Galili, G. Using RNAi to improve plant nutritional value: from mechanism to application. Trends Biotechnol. 22, 463–469 (2004). | Article | PubMed  | ISI | ChemPort |
3. Dasgupta, R. & Perrimon, N. Using RNAi to catch Drosophila genes in a web of interactions: insights into cancer research. Oncogene 23, 8359–8365 (2004). | Article | PubMed  | ChemPort |
4. Fraser, A. Towards full employment: using RNAi to find roles for the redundant. Oncogene 23, 8346–8352 (2004). | Article | PubMed  | ChemPort |
5. Silva, J., Chang, K., Hannon, G.J. & Rivas, F.V. RNA-interference-based functional genomics in mammalian cells: reverse genetics coming of age. Oncogene 23, 8401–8409 (2004). | Article | PubMed  | ChemPort |
6. Bartel, D.P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116, 281–297 (2004). | Article | PubMed  | ISI | ChemPort |
7. He, L. & Hannon, G.J. MicroRNAs: small RNAs with a big role in gene regulation. Nat. Rev. Genet. 5, 522–531 (2004). | Article | PubMed  | ISI | ChemPort |
8. Reinhart, B.J. et al. The 21-nucleotide let-7 RNA regulates developmental timing in Caenorhabditis elegans. Nature 403, 901–906 (2000). | Article | PubMed  | ISI | ChemPort |
9. Ketting, R.F. et al. Dicer functions in RNA interference and in synthesis of small RNA involved in developmental timing in C. elegans. Genes Dev. 15, 2654–2659 (2001). | Article | PubMed  | ISI | ChemPort |
10. Grishok, A. et al. Genes and mechanisms related to RNA interference regulate expression of the small temporal RNAs that control C. elegans developmental timing. Cell 106, 23–34 (2001). | Article | PubMed  | ISI | ChemPort |
11. Knight, S.W. & Bass, B.L. A role for the RNase III enzyme DCR-1 in RNA interference and germ line development in Caenorhabditis elegans. Science 293, 2269–2271 (2001). | Article | PubMed  | ISI | ChemPort |
12. Hutvagner, G. et al. A cellular function for the RNA-interference enzyme Dicer in the maturation of the let-7 small temporal RNA. Science 293, 834–838 (2001). | Article | PubMed  | ISI | ChemPort |
13. Lee, Y. et al. MicroRNA genes are transcribed by RNA polymerase II. EMBO J. 23, 4051–4060 (2004). | Article | PubMed  | ChemPort |
14. Cai, X., Hagedorn, C.H. & Cullen, B.R. Human microRNAs are processed from capped, polyadenylated transcripts that can also function as mRNAs. RNA 10, 1957–1966 (2004). | Article | PubMed  | ISI | ChemPort |
15. Lee, Y. et al. The nuclear RNase III Drosha initiates microRNA processing. Nature 425, 415–419 (2003). | Article | PubMed  | ISI | ChemPort |
16. Denli, A.M., Tops, B.B., Plasterk, R.H., Ketting, R.F. & Hannon, G.J. Processing of primary microRNAs by the Microprocessor complex. Nature 432, 231–235 (2004). | Article | PubMed  | ISI | ChemPort |
17. Landthaler, M., Yalcin, A. & Tuschl, T. The human DiGeorge syndrome critical region gene 8 and its D. melanogaster homolog are required for miRNA biogenesis. Curr. Biol. 14, 2162–2167 (2004). | Article | PubMed  | ISI | ChemPort |
18. Han, J. et al. The Drosha-DGCR8 complex in primary microRNA processing. Genes Dev. 18, 3016–3027 (2004). | Article | PubMed  | ISI | ChemPort |
19. Gregory, R.I. et al. The Microprocessor complex mediates the genesis of microRNAs. Nature 432, 235–240 (2004). | Article | PubMed  | ISI | ChemPort |
20. Yi, R., Qin, Y., Macara, I.G. & Cullen, B.R. Exportin-5 mediates the nuclear export of pre-microRNAs and short hairpin RNAs. Genes Dev. 17, 3011–3016 (2003). | Article | PubMed  | ISI | ChemPort |
21. Lund, E., Guttinger, S., Calado, A., Dahlberg, J.E. & Kutay, U. Nuclear export of microRNA precursors. Science 303, 95–98 (2004). | Article | PubMed  | ISI | ChemPort |
22. Siolas, D. et al. Synthetic shRNAs as potent RNAi triggers. Nat. Biotechnol. 23, 227–231 (2005). | Article | PubMed  | ISI | ChemPort |
23. Song, J.J. et al. The crystal structure of the Argonaute2 PAZ domain reveals an RNA binding motif in RNAi effector complexes. Nat. Struct. Biol. 10, 1026–1032 (2003). | Article | PubMed  | ISI | ChemPort |
24. Schwarz, D.S. et al. Asymmetry in the assembly of the RNAi enzyme complex. Cell 115, 199–208 (2003). | Article | PubMed  | ISI | ChemPort |
25. Khvorova, A., Reynolds, A. & Jayasena, S.D. Functional siRNAs and miRNAs exhibit strand bias. Cell 115, 209–216 (2003). | Article | PubMed  | ISI | ChemPort |
26. Paddison, P.J. et al. A resource for large-scale RNA-interference-based screens in mammals. Nature 428, 427–431 (2004). | Article | PubMed  | ISI | ChemPort |
27. Berns, K. et al. A large-scale RNAi screen in human cells identifies new components of the p53 pathway. Nature 428, 431–437 (2004). | Article | PubMed  | ISI | ChemPort |
28. Zeng, Y., Wagner, E.J. & Cullen, B.R. Both natural and designed micro RNAs can inhibit the expression of cognate mRNAs when expressed in human cells. Mol. Cell 9, 1327–1333 (2002). | Article | PubMed  | ISI | ChemPort |
29. Paddison, P.J., Caudy, A.A., Bernstein, E., Hannon, G.J. & Conklin, D.S. Short hairpin RNAs (shRNAs) induce sequence-specific silencing in mammalian cells. Genes Dev. 16, 948–958 (2002). | Article | PubMed  | ISI | ChemPort |
30. Westbrook, T.F. et al. A genetic screen for candidate tumor suppressors identifies REST. Cell 121, 837–848 (2005). | Article | PubMed  | ISI | ChemPort |
31. Chen, C.Z., Li, L., Lodish, H.F. & Bartel, D.P. MicroRNAs modulate hematopoietic lineage differentiation. Science 303, 83–86 (2004). | Article | PubMed  | ISI | ChemPort |
32. Zeng, Y. & Cullen, B.R. Sequence requirements for micro RNA processing and function in human cells. RNA 9, 112–123 (2003). | Article | PubMed  | ISI | ChemPort |
33. Kawasaki, H. & Taira, K. Short hairpin type of dsRNAs that are controlled by tRNA(Val) promoter significantly induce RNAi-mediated gene silencing in the cytoplasm of human cells. Nucleic Acids Res. 31, 700–707 (2003). | Article | PubMed  | ISI | ChemPort |
34. Brummelkamp, T.R., Bernards, R. & Agami, R. A system for stable expression of short interfering RNAs in mammalian cells. Science 296, 550–553 (2002). | Article | PubMed  | ISI | ChemPort |
35. Zheng, L. et al. An approach to genomewide screens of expressed small interfering RNAs in mammalian cells. Proc. Natl. Acad. Sci. USA 101, 135–140 (2004). | Article | PubMed  | ChemPort |
36. Dickins, R.A. et al. Probing tumor phenotypes using stable and regulated synthetic microRNA precursors. Nat. Genet., advance online publication XX XXX 2005 (10.1038/ngXXX). [date and doi for lowe]
37. Stegmeier, F., Hu, G., Rickles, R.J., Hannon, G.J. & Elledge, S.J. A lentiviral microRNA-based system for single copy Pol II regulated RNAi in mammalian cells. Proc. Natl. Acad. Sci. USA 102, 13212–13217 (2005). | Article | PubMed  | ChemPort |
38. Li, M.Z. & Elledge, S.J. MAGIC, an in vivo genetic method for the rapid construction of recombinant DNA molecules. Nat. Genet. 37, 311–319 (2005). | Article | PubMed  | ISI | ChemPort |
39. Cleary, M.A. et al. Production of complex nucleic acid libraries using highly parallel in situ oligonucleotide synthesis. Nat. Methods 1, 241–248 (2004). | Article | PubMed  | ISI | ChemPort |
40. Li, X. et al. Generation of destabilized green fluorescent protein as a transcription reporter. J. Biol. Chem. 273, 34970–34975 (1998). | Article | PubMed  | ISI | ChemPort |
41. Carmell, M.A. & Hannon, G.J. RNase III enzymes and the initiation of gene silencing. Nat. Struct. Mol. Biol. 11, 214–218 (2004). | Article | PubMed  | ChemPort |
42. Elledge, S.J. & Walker, G.C. Phasmid vectors for identification of genes by complementation of Escherichia coli mutants. J. Bacteriol. 162, 777–783 (1985). | PubMed  | ISI | ChemPort |
43. Datsenko, K.A. & Wanner, B.L. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6640–6645 (2000). | Article | PubMed  | ChemPort |
44. Cherepanov, P.P. & Wackernagel, W. Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene 158, 9–14 (1995). | Article | PubMed  | ISI | ChemPort |
45. Chalker, A.F., Leach, D.R. & Lloyd, R.G. Escherichia coli sbcC mutants permit stable propagation of DNA replicons containing a long palindrome. Gene 71, 201–205 (1988). | Article | PubMed  | ISI | ChemPort |
46. Caudy, A.A., Myers, M., Hannon, G.J. & Hammond, S.M. Fragile X-related protein and VIG associate with the RNA interference machinery. Genes Dev. 16, 2491–2496 (2002). | Article | PubMed  | ISI | ChemPort |

	Top

Competing interests statement:  The authors declare that they have no competing financial interests.