22. Saitoh, S., Takahashi, K. & Yanagida, M. Mis6, a fission yeast inner centromere protein, acts during
`G1/S and forms specialized chromatin required for equal segregation. Cell 90, 131–143 (1997).
`23. Kelly, T. J. et al. The fission yeast cdc18 gene product couples S-phase to start and mitosis. Cell 74, 371–
`382 (1993).
`24. Fernandez-Sarabia, M. J., McInery, C., Harris, P., Gordon, C. & Fantes, P. The cell cycle genes cdc22+
`and suc22+ of the fission yeast Schizosaccharomyces pombe encode the large and small subunits of
`ribonucleotide reductase. Mol. Gen. Genet. 238, 241–251 (1993).
`25. Lin, L. & Smith, G. R. Transient, meiosis-induced expression of the rec6 and rec12 genes of
`Schizosaccharomyces pombe. Genetics 136, 769–779 (1994).
`
`Supplementary information is available on Nature’s World-Wide Web site
`(http://www.nature.com) or as paper copy from the London editorial office of Nature.
`
`Acknowledgements
`We thank J. P. Cooper for critical reading of the manuscript and M. Yanagida for CHIP
`method. Y.W. thanks all the members of P.N.’s laboratory for help and discussion,
`particularly J. Hayles, H. Murakami and G. Simchen. Y.W. was supported by JSPS and
`Uehara fellowships and grants from the Ministry of Education, Science and Culture of
`Japan.
`
`Correspondence and requests for materials should be addressed to Y.W.
`(e-mail: ywatanab@ims.u-tokyo.ac.jp).
`
`.................................................................
`Role for a bidentate ribonuclease in
`the initiation step of RNA interference
`
`Emily Bernstein*†, Amy A. Caudy*‡, Scott M. Hammond*§
`& Gregory J. Hannon*
`
`* Cold Spring Harbor Laboratory, and ‡Watson School of Biological Sciences,
`1 Bungtown Road, Cold Spring Harbor, New York 11724, USA
`† Graduate Program in Genetics, State University of New York at Stony Brook,
`Stony Brook, New York, 11794, USA
`§ Genetica, 1 Kendall Square, Building 600, Cambridge, Massachusetts 01239,
`USA
`
`..............................................................................................................................................
`RNA interference (RNAi) is the mechanism through which
`double-stranded RNAs silence cognate genes1–5. In plants, this
`can occur at both the transcriptional and the post-transcriptional
`levels1,2,5; however, in animals, only post-transcriptional RNAi has
`been reported to date. In both plants and animals, RNAi is
`characterized by the presence of RNAs of about 22 nucleotides
`in length that are homologous to the gene that is being sup-
`pressed6–8. These 22-nucleotide sequences
`serve as guide
`sequences that instruct a multicomponent nuclease, RISC, to
`destroy specific messenger RNAs6. Here we identify an enzyme,
`Dicer, which can produce putative guide RNAs. Dicer is a member
`of the RNase III family of nucleases that specifically cleave double-
`stranded RNAs, and is evolutionarily conserved in worms, flies,
`plants, fungi and mammals. The enzyme has a distinctive struc-
`ture, which includes a helicase domain and dual RNase III motifs.
`Dicer also contains a region of homology to the RDE1/QDE2/
`ARGONAUTE family that has been genetically linked to RNAi9,10.
`Biochemical studies have suggested that post-transcriptional
`gene silencing (PTGS) is accomplished by a multicomponent
`nuclease that targets mRNAs for degradation6,8,11. The specificity
`of this complex may derive from the incorporation of a small
`guide sequence that is homologous to the mRNA substrate6.
`These ,22-nucleotide RNAs, originally identified in plants that
`were actively silencing transgenes7, have been produced during
`RNAi in vitro using an extract prepared from Drosophila embryos8.
`Putative guide RNAs can also be produced in extracts from
`Drosophila S2 cells (Fig. 1a). To investigate the mechanism of
`PTGS, we have performed both biochemical fractionation and
`candidate gene approaches to identify the enzymes that execute
`each step of RNAi.
`
`letters to nature
`
`Our previous studies resulted in the partial purification of an
`enzyme complex, RISC, which is an effector nuclease for RNA
`interference6. This enzyme was isolated from Drosophila S2 cells in
`which RNAi had been initiated in vivo by transfection with double-
`stranded RNA (dsRNA). We first investigated whether the RISC
`enzyme, and the enzyme that initiates RNAi through processing of
`dsRNA into 22-nucleotide sequences, are distinct activities. RISC
`activity could be largely cleared from extracts by high-speed
`centrifugation (100,000g for 60 min), whereas the activity that
`produces 22-nucleotide sequences remained in the supernatant
`(Fig. 1b, c). This simple fractionation indicates that RISC and the
`22-nucleotide sequence-generating activity may be separable. How-
`ever, it seems probable that these enzymes interact at some point
`during the silencing process, and it remains possible that initiator
`and effector enzymes share common subunits.
`RNase III family members are among the few nucleases that show
`for dsRNA12. Analysis of
`specificity
`the Drosophila
`and
`Caenorhabditis elegans genomes reveals several types of RNase III
`enzymes. First is the canonical RNase III, which contains a single
`RNase III signature motif and a dsRNA-binding domain (dsRBD;
`for example RNC_CAEEL). Second is a class represented by
`Drosha13, a Drosophila enzyme that contains two RNase III motifs
`and a dsRBD (CeDrosha in C. elegans). A third class contains two
`RNase III signatures and an amino-terminal helicase domain (for
`example, Drosophila CG4792 and CG6493; C. elegans K12H4.8),
`which had been proposed as potential RNAi nucleases14,20. We tested
`representatives of all three classes for the ability to produce discrete
`RNAs of ,22 nucleotides from dsRNA substrates.
`To test the dual RNase III enzymes, we prepared variants of
`Drosha and CG4792 tagged with the T7 epitope. These were
`expressed in transfected S2 cells and isolated by immunoprecipita-
`tion using antibody–agarose conjugates. Treatment of the dsRNA
`with the CG4792 immunoprecipitate yielded fragments of about 22
`nucleotides, similar to those produced in either the S2 or embryo
`extracts (Fig. 2a). Neither the activity in extract nor that in
`immunoprecipitates depended on the sequence of the RNA sub-
`strate, as dsRNAs derived from several genes were processed
`equivalently (see Supplementary Information). Negative results
`
`a
`
`S2 cells Embryo
`M
`
`b
`S10
`
`S100
`
`S10
`
`S100
`
`c
`
`S10
`
`S100
`
`luciferase
`
`cyclin E
`
`Figure 1 Generation of 22-nucleotide sequences and degradation of mRNA by distinct
`enzymatic complexes. a, Extracts prepared from for 0–12 h Drosophila embryos or
`Drosophila S2 cells. Extracts were incubated for 0, 15, 30 or 60 min (left to right) with a
`uniformly labelled dsRNA. A 22-nucleotide marker prepared by in vitro transcription of a
`synthetic template is indicated (M). b, Whole-cell extracts from S2 cells transfected with
`luciferase dsRNA. S10 represents our standard RISC extract6. S100 extracts were
`prepared by additional centrifugation of S10 extracts for 60 min at 100,000g. Assays for
`mRNA degradation6 were performed for 0, 30 or 60 min (left to right in each set) with
`either a single-stranded luciferase mRNA or a single-stranded cyclin E mRNA, as
`indicated. c, S10 or S100 extracts incubated with cyclin E dsRNAs for 0, 60 or 120 min
`(left to right).
`
`NATURE | VOL 409 | 18 JANUARY 2001 | www.nature.com
`
`© 2001 Macmillan Magazines Ltd
`
`363
`
`

`
`22. Saitoh, S., Takahashi, K. & Yanagida, M. Mis6, a fission yeast inner centromere protein, acts during
`G1/S and forms specialized chromatin required for equal segregation. Cell 90, 131–143 (1997).
`23. Kelly, T. J. et al. The fission yeast cdc18 gene product couples S-phase to start and mitosis. Cell 74, 371–
`382 (1993).
`24. Fernandez-Sarabia, M. J., McInery, C., Harris, P., Gordon, C. & Fantes, P. The cell cycle genes cdc22+
`and suc22+ of the fission yeast Schizosaccharomyces pombe encode the large and small subunits of
`ribonucleotide reductase. Mol. Gen. Genet. 238, 241–251 (1993).
`25. Lin, L. & Smith, G. R. Transient, meiosis-induced expression of the rec6 and rec12 genes of
`Schizosaccharomyces pombe. Genetics 136, 769–779 (1994).
`
`Supplementary information is available on Nature’s World-Wide Web site
`(http://www.nature.com) or as paper copy from the London editorial office of Nature.
`
`Acknowledgements
`We thank J. P. Cooper for critical reading of the manuscript and M. Yanagida for CHIP
`method. Y.W. thanks all the members of P.N.’s laboratory for help and discussion,
`particularly J. Hayles, H. Murakami and G. Simchen. Y.W. was supported by JSPS and
`Uehara fellowships and grants from the Ministry of Education, Science and Culture of
`Japan.
`
`Correspondence and requests for materials should be addressed to Y.W.
`(e-mail: ywatanab@ims.u-tokyo.ac.jp).
`
`.................................................................
`Role for a bidentate ribonuclease in
`the initiation step of RNA interference
`
`Emily Bernstein*†, Amy A. Caudy*‡, Scott M. Hammond*§
`& Gregory J. Hannon*
`
`* Cold Spring Harbor Laboratory, and ‡Watson School of Biological Sciences,
`1 Bungtown Road, Cold Spring Harbor, New York 11724, USA
`† Graduate Program in Genetics, State University of New York at Stony Brook,
`Stony Brook, New York, 11794, USA
`§ Genetica, 1 Kendall Square, Building 600, Cambridge, Massachusetts 01239,
`USA
`
`..............................................................................................................................................
`RNA interference (RNAi) is the mechanism through which
`double-stranded RNAs silence cognate genes1–5. In plants, this
`can occur at both the transcriptional and the post-transcriptional
`levels1,2,5; however, in animals, only post-transcriptional RNAi has
`been reported to date. In both plants and animals, RNAi is
`characterized by the presence of RNAs of about 22 nucleotides
`in length that are homologous to the gene that is being sup-
`pressed6–8. These 22-nucleotide sequences
`serve as guide
`sequences that instruct a multicomponent nuclease, RISC, to
`destroy specific messenger RNAs6. Here we identify an enzyme,
`Dicer, which can produce putative guide RNAs. Dicer is a member
`of the RNase III family of nucleases that specifically cleave double-
`stranded RNAs, and is evolutionarily conserved in worms, flies,
`plants, fungi and mammals. The enzyme has a distinctive struc-
`ture, which includes a helicase domain and dual RNase III motifs.
`Dicer also contains a region of homology to the RDE1/QDE2/
`ARGONAUTE family that has been genetically linked to RNAi9,10.
`Biochemical studies have suggested that post-transcriptional
`gene silencing (PTGS) is accomplished by a multicomponent
`nuclease that targets mRNAs for degradation6,8,11. The specificity
`of this complex may derive from the incorporation of a small
`guide sequence that is homologous to the mRNA substrate6.
`These ,22-nucleotide RNAs, originally identified in plants that
`were actively silencing transgenes7, have been produced during
`RNAi in vitro using an extract prepared from Drosophila embryos8.
`Putative guide RNAs can also be produced in extracts from
`Drosophila S2 cells (Fig. 1a). To investigate the mechanism of
`PTGS, we have performed both biochemical fractionation and
`candidate gene approaches to identify the enzymes that execute
`each step of RNAi.
`
`letters to nature
`
`Our previous studies resulted in the partial purification of an
`enzyme complex, RISC, which is an effector nuclease for RNA
`interference6. This enzyme was isolated from Drosophila S2 cells in
`which RNAi had been initiated in vivo by transfection with double-
`stranded RNA (dsRNA). We first investigated whether the RISC
`enzyme, and the enzyme that initiates RNAi through processing of
`dsRNA into 22-nucleotide sequences, are distinct activities. RISC
`activity could be largely cleared from extracts by high-speed
`centrifugation (100,000g for 60 min), whereas the activity that
`produces 22-nucleotide sequences remained in the supernatant
`(Fig. 1b, c). This simple fractionation indicates that RISC and the
`22-nucleotide sequence-generating activity may be separable. How-
`ever, it seems probable that these enzymes interact at some point
`during the silencing process, and it remains possible that initiator
`and effector enzymes share common subunits.
`RNase III family members are among the few nucleases that show
`for dsRNA12. Analysis of
`specificity
`the Drosophila
`and
`Caenorhabditis elegans genomes reveals several types of RNase III
`enzymes. First is the canonical RNase III, which contains a single
`RNase III signature motif and a dsRNA-binding domain (dsRBD;
`for example RNC_CAEEL). Second is a class represented by
`Drosha13, a Drosophila enzyme that contains two RNase III motifs
`and a dsRBD (CeDrosha in C. elegans). A third class contains two
`RNase III signatures and an amino-terminal helicase domain (for
`example, Drosophila CG4792 and CG6493; C. elegans K12H4.8),
`which had been proposed as potential RNAi nucleases14,20. We tested
`representatives of all three classes for the ability to produce discrete
`RNAs of ,22 nucleotides from dsRNA substrates.
`To test the dual RNase III enzymes, we prepared variants of
`Drosha and CG4792 tagged with the T7 epitope. These were
`expressed in transfected S2 cells and isolated by immunoprecipita-
`tion using antibody–agarose conjugates. Treatment of the dsRNA
`with the CG4792 immunoprecipitate yielded fragments of about 22
`nucleotides, similar to those produced in either the S2 or embryo
`extracts (Fig. 2a). Neither the activity in extract nor that in
`immunoprecipitates depended on the sequence of the RNA sub-
`strate, as dsRNAs derived from several genes were processed
`equivalently (see Supplementary Information). Negative results
`
`a
`
`S2 cells Embryo
`M
`
`b
`S10
`
`S100
`
`S10
`
`S100
`
`c
`
`S10
`
`S100
`
`luciferase
`
`cyclin E
`
`Figure 1 Generation of 22-nucleotide sequences and degradation of mRNA by distinct
`enzymatic complexes. a, Extracts prepared from for 0–12 h Drosophila embryos or
`Drosophila S2 cells. Extracts were incubated for 0, 15, 30 or 60 min (left to right) with a
`uniformly labelled dsRNA. A 22-nucleotide marker prepared by in vitro transcription of a
`synthetic template is indicated (M). b, Whole-cell extracts from S2 cells transfected with
`luciferase dsRNA. S10 represents our standard RISC extract6. S100 extracts were
`prepared by additional centrifugation of S10 extracts for 60 min at 100,000g. Assays for
`mRNA degradation6 were performed for 0, 30 or 60 min (left to right in each set) with
`either a single-stranded luciferase mRNA or a single-stranded cyclin E mRNA, as
`indicated. c, S10 or S100 extracts incubated with cyclin E dsRNAs for 0, 60 or 120 min
`(left to right).
`
`NATURE | VOL 409 | 18 JANUARY 2001 | www.nature.com
`
`© 2001 Macmillan Magazines Ltd
`
`363
`
`

`
`letters to nature
`
`were obtained with Drosha and with immunoprecipitates of a
`DExH box helicase (Homeless15; see Fig. 2a and b). Western blotting
`confirmed that each of the tagged proteins was expressed and
`immunoprecipitated similarly (see Supplementary Information).
`Thus, we conclude that CG4792 may carry out the initiation step of
`RNAi by producing guide sequences of about 22 nucleotides from
`dsRNAs. Because of its ability to digest dsRNA into uniformly sized,
`small RNAs, we have named this enzyme Dicer (Dcr). Dicer mRNA
`is expressed in embryos, in S2 cells and in adult flies, which is
`consistent with the presence of functional RNAi machinery in all of
`these contexts (see Supplementary Information).
`An antiserum directed against the carboxy terminus of the Dicer
`protein (Dicer-1, CG4792) could immunoprecipitate a nuclease
`activity from either the Drosophila embryo extracts or from S2 cell
`lysates that produced RNAs of about 22 nucleotides from dsRNA
`substrates (Fig. 2c). The putative guide RNAs that are produced by
`the Dicer-1 enzyme precisely co-migrate with 22-nucleotide
`sequences that are produced in extract, and with 22-nucleotide
`sequences that are associated with the RISC enzyme (Fig. 2d, f). The
`enzyme that produces guide RNAs in Drosophila embryo extracts is
`ATP dependent8. Depletion of this cofactor resulted in a roughly
`sixfold reduction of dsRNA cleavage rate and in the production of
`
`RNAs with a slightly lower mobility. Of note, both Dicer-1 immu-
`noprecipitates and extracts from S2 cells require ATP for the
`production of ,22-nucleotide sequences (Fig. 2d). We did not
`observe the accumulation of lower-mobility products in these cases,
`although we did routinely observe these in ATP-depleted embryo
`extracts. The requirement of this nuclease for ATP is an unusual
`property, and may indicate that unwinding of guide RNAs by the
`helicase domain is required for the enzyme to act catalytically.
`For efficient induction of RNAi in C. elegans and in Drosophila,
`the initiating RNA must be double-stranded and must also be
`several hundred nucleotides in length4. Similarly, Dicer was inactive
`against single-stranded RNAs regardless of length (see Supplemen-
`tary Information). The enzyme could digest both 200- and 500-
`nucleotide dsRNAs, but was significantly less active with shorter
`substrates
`(see
`Supplementary
`Information).
`In contrast,
`Escherichia coli RNase III could digest to completion dsRNAs of
`35 or 22 nucleotides (data not shown). This suggests that the
`substrate preferences of the Dicer enzyme may contribute to, but
`not wholly determine, the size dependence of RNAi.
`To determine whether the Dicer enzyme is involved in RNAi in
`vivo, we depleted Dicer activity from S2 cells and tested the effect on
`dsRNA-induced gene silencing. Transfection of S2 cells with a
`
`Marker
`Control
`RISC
`Dicer IP
`
`f
`
`RISC (Is)
`
`RISC (hs)
`
`e
`
`Total
`
`d
`
`ATP
`
`–
`
`IP
`+
`
`Ext
`+
`–
`
`c
`
`Extract
`Plus peptide
`Immune
`Pre-immune
`Marker
`
`b-gal
`Homeless
`Dicer
`Drosha
`Embryo
`
`a
`
`2
`M S
`
`Helicase
`
`Paz
`
`RIII a RIII b dsrm
`
`RIII a RIII b dsrm
`
`b D
`
`icer
`
`Drosha
`
`Homeless
`
`Helicase
`
`Figure 2 Production of 22-nucleotide sequences by CG4792/Dicer. a, Drosophila S2
`cells transfected with plasmids that direct expression of T7-epitope-tagged versions of
`Drosha, CG4792/Dicer-1 and Homeless or untagged b-galactosidase. Proteins were
`immunoprecipitated and incubated with cyclin E dsRNA for 0 or 60 min. Reactions in
`Drosophila embryo and S2 cell extracts are shown. b, Domain structures of CG4792/
`Dicer-1, Drosha and Homeless. c, Immunoprecipitates prepared from detergent lysates of
`S2 cells using Dicer antiserum. As controls, similar preparations were made with a pre-
`immune serum and an immune serum that had been pre-incubated with an excess of
`antigenic peptide. Cleavage reactions in which each of these precipitates was incubated
`with a ,500 nucleotide fragment of Drosophila cyclin E are shown. An incubation of the
`substrate in Drosophila embryo extract is shown. d, Dicer immunoprecipitates incubated
`with dsRNA substrates in presence or absence of ATP. The same substrate was also
`
`incubated with ATP-added or ATP-depleted S2 extracts. e, Drosophila S2 cells
`transfected with uniformly, 32P-labelled dsRNA corresponding to the first 500 nucleotides
`of GFP. RISC complex was affinity purified using a histidine-tagged version of Drosophila
`Ago-2, a component of the RISC complex (Hammond et al., manuscript in preparation).
`RISC was isolated under ribosome-associated (ls, low salt) or soluble, ribosome-extracted
`(hs, high salt) conditions6. The spectrum of labelled RNAs in the total lysate is shown.
`f, Comparison of guide RNAs produced by incubation of dsRNA with a Dicer
`immunoprecipitate, with guide RNAs present in affinity-purified RISC complex. These co-
`migrate on a gel that has single-nucleotide resolution. The control lane shows an affinity
`selection for RISC from cells transfected with labelled dsRNA, but not with the epitope-
`tagged Drosophila Ago-2.
`
`364
`
`© 2001 Macmillan Magazines Ltd
`
`NATURE | VOL 409 | 18 JANUARY 2001 | www.nature.com
`
`

`
`letters to nature
`
`Supplementary Information), which indicates that these structu-
`rally similar proteins may all share similar biochemical functions.
`Exogenous dsRNAs can affect gene function in early mouse
`embryos18, and our results suggest that this regulation may be
`accomplished by evolutionarily conserved RNAi machinery.
`In addition to RNase III and helicase motifs, searches of the
`PFAM database indicate that each Dicer family member also
`contains a PAZ domain (see Supplementary Information)19,20.
`This sequence was defined on the basis of its conservation in the
`Zwille/ARGONAUTE/Piwi family that has been implicated in RNAi
`by mutations in C. elegans (Rde-1)9 and Neurospora (Qde-2)10.
`Although the function of this domain is unknown, it is notable that
`this region of homology is restricted to two gene families that
`participate in dsRNA-dependent silencing. Both the ARGONAUTE
`and Dicer families have also been implicated in common biological
`processes, namely the determination of stem-cell fates. A hypo-
`morphic allele of carpel factory, a member of the Dicer family in
`Arabidopsis, is characterized by increased proliferation in floral
`meristems16. This phenotype and a number of other characteristic
`features are also shared by Arabidopsis ARGONAUTE (ago1-1)
`mutants21 (C. Kidner and R. Martiennsen, personal communica-
`tion). These genetic analyses provide evidence that RNAi may be
`more than a defensive response to unusual RNAs, but may also have
`integral functions in the regulation of endogenous genes.
`With the identification of Dicer as a potential catalyst of the
`initiation step of RNAi, we have begun to unravel the biochemical
`basis of this unusual mechanism of gene regulation. It is now
`important to determine whether the conserved family members
`from other organisms, particularly mammals, also have a function
`in dsRNA-mediated gene regulation.
`Note added in proof: Yang et al.22 have recently presented evidence
`that guide RNAs are derived directly from dsRNA in Drosophila
`embryos. Fagard et al.23 have recently shown that Arabidopsis Ago1 is
`involved in PTGS.
`M
`
`Methods
`Plasmid constructs
`A full-length complementary DNA encoding Drosha was obtained by polymerase chain
`raaction (PCR) from an expressed sequence tag sequenced by the Berkeley Drosophila
`genome project. The T7 epitope-tag was added to the N terminus of each cDNA by PCR,
`and the tagged cDNAs were cloned into pRIP—a retroviral vector designed specifically for
`expression in insect cells (E. B., unpublished observations). In this vector, expression is
`driven by the Orgyia pseudotsugata IE2 promoter (Invitrogen). As no cDNA was available
`for CG4792/Dicer, a genomic clone was amplified from a bac (bacterial artificial
`chromosome) (BACR23F10; obtained from the BACPAC Resource Center in the
`Deptartment of Human Genetics at the Roswell Park Cancer Institute). We added a T7
`epitope tag at the N terminus of the coding sequence during amplification. We isolated the
`human DICER gene from a cDNA library prepared from HaCaT cells (G.J.H., unpublished
`observations). A T7-tagged version of the complete coding sequence was cloned into
`pCDNA3 (Invitrogen) for expression in human cells (LinX-A).
`
`Cell culture and extract preparation
`We cultured S2 cells at 27 8C in 5% CO2 in Schneider’s insect media supplemented with
`10% heat-inactivated fetal bovine serum (Gemini) and 1% antibiotic–antimycotic
`solution (Gibco BRL). Cells were collected for extract preparation at 107 cells per ml. The
`cells were washed in PBS and resuspended in a hypotonic buffer (10 mM HEPES pH 7.0,
`2 mM MgCl2 and 6 mM b-mercaptoethanol) and lysed. We centrifuged cell lysates at
`20,000g for 20 min. We stored extracts at -80 8C. We reared Drosophila embryos in fly cages
`by standard methodologies and collected them every 12 h. We dechorionated the embryos
`in 50% chlorox bleach and washed them thoroughly with distilled water. Lysis buffer
`(10 mM Hepes, 10 mM KCl, 1.5 mM MgCl2, 0.5 mM EGTA, 10 mM b-glycerophosphate,
`1 mM dithiothreitol (DTT) and 0.2 mM PMSF) was added to the embryos, and extracts
`were prepared by homogenization in a tissue grinder. Lysates were centrifuged for 2 h at
`200,000g, and were frozen at -80 8C. LinX-A cells, a highly transfectable derivative of
`human 293 cells (L. Xie and G.J.H., unpublished observations) were maintained in
`DMEM/10% FCS.
`
`Transfections and immunoprecipitations
`We transfected S2 cells using a calcium phosphate procedure essentially as described6.
`Transfection rates were about 90%, as monitored in controls using an in situ
`b-galactosidase assay. We also transfected LinX-A cells by calcium phosphate
`co-precipitation. For immunoprecipitations, cells (,5 · 106 per immunoprecipitate) were
`
`mixture of dsRNAs homologous to the two Drosophila Dicer genes
`(CG4792 and CG6493) resulted in a roughly 6–7-fold reduction of
`Dicer activity either in whole-cell lysates or in Dicer-1 immuno-
`precipitates (Fig. 3a and b). Transfection with a control dsRNA
`(murine caspase-9) had no effect. Qualitatively similar results were
`seen if Dicer mRNA was examined by northern blotting (data not
`shown). Depletion of Dicer substantially compromised the ability of
`cells to silence an exogenous, green fluorescent protein (GFP)
`transgene by RNAi (Fig. 3c). These results indicate that Dicer may
`be involved in RNAi in vivo. The lack of complete inhibition of
`silencing may result from an incomplete suppression of Dicer or
`may indicate that in vivo guide RNAs may be produced by more
`than one mechanism.
`Our results indicate that the process of RNAi can be divided into
`at least two distinct steps. Initiation of PTGS would occur on
`processing of a dsRNA by Dicer into ,22-nucleotide guide
`sequences, although we cannot formally exclude the possibility
`that another Dicer-associated nuclease may participate in this
`process. These guide RNAs would be incorporated into a distinct
`nuclease complex (RISC) that targets single-stranded mRNAs for
`degradation. An implication of this model
`is that the guide
`sequences are themselves derived directly from the dsRNA that
`triggers the response. In accord with this model, we have shown that
`32P-labelled, exogenous dsRNAs that have been introduced into S2
`cells by transfection are incorporated into the RISC enzyme as 22-
`nuclotide sequences (Fig. 2e).
`A notable feature of the Dicer family is its evolutionary conserva-
`tion. Homologues are found in C. elegans (K12H4.8), Arabidopsis
`(for example, CARPEL FACTORY16, T25K16.4 and AC012328_1),
`mammals
`(Helicase-MOI17)
`and Schizosaccharomyces pombe
`(YC9A_SCHPO) (see Supplementary Information for compari-
`sons). In fact, the human Dicer family member is capable of
`generating ,22-nucleotide RNAs from dsRNA substrates (see
`
`Exp. 3
`
`
`Exp. 2
`
`Exp. 1
`
`Luc dsRNA + control ds
`Luc dsRNA + dicer ds
`GFP dsRNA + control ds
`GFP dsRNA + dicer ds
`
`70
`
`60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`c
`
`Cells expressing GFP (%)
`
`Dicer dsRNA
`casp9 dsRNA
`
`b
`
`Dicer dsRNA
`casp9 dsRNA
`
`a
`
`Figure 3 Dicer participates in RNAi. a, Drosophila S2 cells transfected with dsRNAs
`corresponding to the two Drosophila Dicers (CG4792 and CG6493) or control dsRNA
`corresponding to murine caspase-9 (casp9). Cytoplasmic extracts of these cells were
`tested for Dicer activity. Transfection with Dicer dsRNA reduces activity in lysates
`7.4-fold. b, Dicer-1 antiserum (CG4792) used to prepare immunoprecipitates from S2
`cells (treated as above). Dicer dsRNA reduces the activity of Dicer-1 6.2-fold. c, GFP
`expression of co-transfected cells. Three independent experiments were quantified by
`FACS. A comparison of the relative percentage of GFP-positive cells is shown for control
`(GFP plasmid plus luciferase dsRNA) or silenced (GFP plasmids plus GFP dsRNA)
`populations in cells that had previously been transfected with either control (caspase-9) or
`Dicer dsRNAs.
`
`NATURE | VOL 409 | 18 JANUARY 2001 | www.nature.com
`
`© 2001 Macmillan Magazines Ltd
`
`365
`
`

`
`letters to nature
`
`transfected with various clones, and lysed 3 d later in immunoprecipitate buffer
`(125 Mm KOAc, 1 mM MgOAc, 1 mM CaCl2, 5 mM EGTA, 20 mM HEPES pH 7.0, 1 mM
`DTT and 1% Nonidet P40 plus complete protease inhibitors (Roche)). We centrifuged
`lysates for 10 min at 14,000g, and then added supernatants to T7 antibody-agarose beads
`(Novagen). We performed antibody binding for 4 h at 4 8C. Beads were centrifuged and
`washed three times in lysis buffer, and once in reaction buffer. The Dicer antiserum was
`raised in rabbits using a keyhole limpet haemocyanin-conjugated peptide corresponding
`to the C-terminal eight amino acids of Drosophila Dicer-1 (CG4792).
`
`Cleavage reactions
`Templates to be transcribed to dsRNA were generated by PCR with forward and reverse
`primers, each containing a T7 promoter sequence. RNAs were produced using Riboprobe
`kits (Promega) and were uniformly labelled during the transcription reaction with 32P-
`labelled UTP. Single-stranded RNAs were purified from 1% agarose gels. For cleavage of
`dsRNA, 5 ml of embryo or S2 extracts were incubated for 1 h at 30 8C with dsRNA in a
`reaction containing 20 mM HEPES pH 7.0, 2 mM magnesium acetate, 2 mM DTT, 1 mM
`ATP and 5% Superasin (Ambion). Immunoprecipitates were treated similarly, except that
`a minimal volume of reaction buffer (including ATP and superasin) and dsRNA were
`added to beads that had been washed in reaction buffer. For ATP depletion, Drosophila
`embryo extracts were incubated for 20 min at 30 8C with 2 mM glucose and 0.375 U of
`hexokinase (Roche), before the addition of dsRNA.
`
`Northern and western analysis
`Total RNA was prepared from Drosophila embryos (0–12 h), from adult flies and from S2
`cells using Trizol (Lifetech). We isolated mRNA by affinity selection using magnetic LIGO-
`dT beads (Dynal). RNAs were electrophoresed on denaturing formaldehyde/agarose gels,
`blotted and probed with randomly primed DNAs corresponding to Dicer. For western
`analysis, T7-tagged proteins were immunoprecipitated from whole-cell lysates in immu-
`noprecipitate buffer using agarose-conjugated anti-T7 antibody. Proteins were released
`from the beads by boiling in Laemmli buffer, and were separated by 8% SDS–poly-
`acrylamide gel electrophoresis. After transfer to nitrocellulose, proteins were visualized
`using an HRP-conjugated anti-T7 antibody (Novagen) and chemiluminescent detection
`(Supersignal, Pierce).
`
`RNAi of Dicer
`Drosophila S2 cells were transfected either with a dsRNA corresponding to mouse caspase-
`9 or with a mixture of two dsRNAs corresponding to Drosophila Dicer-1 and Dicer-2
`(CG4792 and CG6493). Two days after the initial transfection, cells were again transfected
`with a mixture containing a GFP expression plasmid and either luciferase dsRNA or GFP
`dsRNA as described6. Cells were assayed for Dicer activity or fluorescence 3 d after the
`second transfection. Quantification of fluorescent cells was done on a Coulter EPICS cell
`sorter, after fixation. Control transfections indicated that Dicer activity was not affected by
`the introduction of caspase-9 dsRNA.
`
`Received 16 October; accepted 14 November 2000.
`
`1. Baulcombe, D. C. RNA as a target and an initiator of post-transcriptional gene silencing in transgenic
`plants. Plant Mol. Biol. 32, 79–88 (1996).
`2. Wassenegger, M. & Pelissier, T. A model for RNA-mediated gene silencing in higher plants. Plant Mol.
`Biol. 37, 349–62 (1998).
`3. Montgomery, M. K. & Fire, A. Double-stranded RNA as a mediator in sequence-specific genetic
`silencing and co-suppression. Trends Genet. 14, 255–258 (1998).
`4. Sharp, P. A. RNAi and double-strand RNA. Genes Dev. 13, 139–141 (1999).
`5. Sijen, T. & Kooter, J. M. Post-transcriptional gene-silencing: RNAs on the attack or on the defense?
`BioEssays 22, 520–531 (2000).
`6. Hammond, S. M., Bernstein, E., Beach, D. & Hannon, G. J. An RNA-directed nuclease mediates post-
`transcriptional gene silencing in Drosophila cells. Nature 404, 293–296 (2000).
`7. Hamilton, A. J. & Baulcombe, D. C. A species of small antisense RNA in posttranscriptional gene
`silencing in plants. Science 286, 950–952 (1999).
`8. Zamore, P. D., Tuschl, T., Sharp, P. A. & Bartel, D. P. RNAi: double-stranded RNA directs the ATP-
`dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell 101, 25–33 (2000).
`9. Tabara, H. et al. The rde-1 gene, RNA interference, and transposon silencing in C. elegans. Cell 99,
`123–132 (1999).
`10. Catalanotto, C., Azzalin, G., Macino, G. & Cogoni, C. Gene silencing in worms and fungi. Nature 404,
`245 (2000).
`11. Tuschl, T., Zamore, P. D., Lehmann, R., Bartel, D. P. & Sharp, P. A. Targeted mRNA degradation by
`double-stranded RNA in vitro. Genes Dev. 13, 3191–3197 (1999).
`12. Nicholson, A. W. Function, mechanism and regulation of bacterial ribonucleases. FEMS Microbiol.
`Rev. 23, 371–390 (1999).
`13. Filippov, V., Solovyev, V., Filippova, M. & Gill, S. S. A novel type of RNase III family proteins in
`eukaryotes. Gene 245, 213–221 (2000).
`14. Bass, B. L. Double-stranded RNA as a template for gene silencing. Cell 101, 235–238 (2000).
`15. Gillespie, D. E. & Berg, C. A. Homeless is required for RNA