8/11/2019 nrde-3

1/5

LETTER doi:10.1038/nature11352

A nuclear Argonaute promotes multigenerationalepigenetic inheritance and germline immortalityBethany A. Buckley1*, Kirk B. Burkhart1*, Sam Guoping Gu2, George Spracklin1, Aaron Kershner3, Heidi Fritz1, Judith Kimble1,3,Andrew Fire2 & Scott Kennedy1

Epigenetic information is frequently erased near the start of eachnew generation1. In some cases, however, epigenetic informationcan be transmitted from parent to progeny (multigenerationalepigenetic inheritance)2. A particularly notable example of thistype of epigenetic inheritance is double-stranded RNA-mediatedgene silencing in Caenorhabditis elegans. This RNA-mediatedinterference (RNAi) can be inherited for more than five genera-tions38. To understand this process, here we conduct a geneticscreen for nematodes defective in transmitting RNAi silencing

signals to future generations. This screen identified the heritableRNAi defective 1 (hrde-1) gene. hrde-1 encodes an Argonauteprotein that associates with small interfering RNAs in the germcells of progeny of animals exposed to double-stranded RNA. Inthe nuclei of these germ cells, HRDE-1 engages the nuclear RNAidefective pathway to direct the trimethylation of histone H3 atLys 9 (H3K9me3) at RNAi-targeted genomic loci and promoteRNAi inheritance. Under normal growth conditions, HRDE-1associates with endogenously expressed short interfering RNAs,which direct nuclear gene silencing in germ cells. In hrde-1-ornuclear RNAi-deficient animals, germline silencing is lost overgenerational time. Concurrently, these animals exhibit steadilyworsening defects in gamete formation and functionthat ultimatelylead to sterility. These results establish that the Argonaute protein

HRDE-1 directs gene-silencing events in germ-cell nuclei that drivemultigenerational RNAi inheritance and promote immortality ofthe germ-cell lineage. We propose that C. elegansuse the RNAiinheritance machinery to transmit epigenetic information, accruedby past generations, into future generations to regulate importantbiological processes.

We conducted a genetic screen to identify factors required formultigenerational RNAi inheritance. We mutagenized animalscarrying a germline green fluorescent protein (gfp) reporter gene,and screened for mutant animals that retained the ability to silence

gfpwhen exposed directly togfpdouble-stranded RNA (dsRNA), butfailed to silence gfp in subsequent generations. Among fourteenmutant alleles fulfilling these criteria, four alleles defined a gene weterm hereheritable RNAi defective 1(hrde-1) (Supplementary Fig. 2).

hrde-1mutant animals silenced GFP expression when exposed togfpdsRNA, but failed to transmit this silencing to subsequent generations(Fig. 1a and Supplementary Fig. 3). Similarly,hrde-1mutants silencedthe germline-expressedoma-1gene when treated directly with oma-1dsRNA, but were defective for silencing inheritance (SupplementaryFig. 4). We conclude that hrde-1 promotes multigenerational RNAisilencing in the germ line.

We mappedhrde-1 to a genomicregion containing the c16c10.3 (alsoknown as wago-9) gene, which is predicted to encode an Argonaute(AGO) protein not known to contribute to gene silencing9,10 (see noteadded in proof).c16c10.3encodes a predicted bipartite nuclear local-ization signal (NLS) and PAZ and PIWI domains (Supplementary

Fig. 5). We found thathrde-1isc16c10.3(Supplementary Fig. 5), anda member of the worm-specific clade of AGO (WAGO) family9.HRDE-1 seemed to be relatively unique among the WAGO proteinsin its contribution to germline RNAi inheritance (Supplementary Fig.6). We constructed a fusion gene betweengfp and a full-length genomiccopy of hrde-1 (gfp::hrde-1). gfp::hrde-1 rescued RNAi inheritance inhrde-12/2 animals, indicating that GFPHRDE-1 is functional(Supplementary Fig. 5). GFPHRDE-1 was expressed in the nuclei ofmale and female germ cells (Fig. 1b and Supplementary Fig. 7). These

data indicate that hrde-1 encodes a germline AGO that localizes to thenucleus.

*These authors contributed equally to this work.

1Laboratory of Genetics, University of Wisconsin-Madison, Madison, Wisconsin 53706, USA. 2Departments of Pathology and Genetics, Stanford University, Stanford, California 94305, USA. 3Howard

Hughes Medical Institute, and Department of Biochemistry, University of Wisconsin-Madison, Madison, Wisconsin 53706, USA.

WT

hrde-1(gg265)

rde-1(ne219)

P0(gfp RNAi) Inheriting F

1

b

ca

d

100%

100%

0%

Developing germ line

GFPH2B

0%

0%

No RNAi

100%

100% 100%

100%

GFPHRDE-1

hrde-1+/

gfpdsRNAP0

F1

F2

F3

0%

22%

0%

98%

97%

1% 96%

hrde-1/

P0oma-1dsRNA

F2F1 F3 F41

10

100

1,000

10,000

F5

GFPH2B

Adult germ line

L1

ChangeinHRDE-1-bound

oma-1siRNA(oma-1

dsRNA)

Figure 1| hrde-1encodes a nuclear AGO protein that acts in inheritinggenerations to promote multigenerational germline RNAi inheritance.a, pie-1::gfp::h2b-expressing animals were exposed to gfpdsRNA. F1progenywere grown in the absence of dsRNA, and GFP expression in oocytes wasvisualized by fluorescencemicroscopy. The percentage of fluorescent animals isindicated.rde-1is required for RNAi17. P0, parental generation; WT, wild-type(n. 100).b, GFPHRDE-1 was visualized by fluorescent microscopy. Smallarrow, L1 animal; inset, magnifications showing GFPHRDE-1 in primordialgerm cells. Large arrow, oocyte nucleus. c, 33FlagHRDE-1 wasimmunoprecipitated with an anti-Flag antibody and co-precipitatinghrde-1RNA was isolated, andoma-1siRNAs were quantified with anoma-1TaqManprobe set. Data are expressed as fold change (6 oma-1RNAi), 1 denotes nochange, and the mean and s.d.are shown (n5 3). d,pie-1::gfp::h2b fluorescencewas scored inhrde-11/2 andhrde-11/1 (scored as one group) or inhrde-12/2

inheritingprogeny. The hrde-12/2 chromosomewas marked with unc-93 (unc-93is ,1.3 cM fromhrde-1), andhrde-1genotypes were inferred by Uncphenotypes. The percentage of fluorescent animals is indicated. Originalmagnification for all images,3630.

2 0 S E P T E M B E R 2 0 1 2 | V O L 4 8 9 | N A T U R E | 4 4 7

Macmillan Publishers Limited. All rights reserved2012

http://www.nature.com/doifinder/10.1038/nature11352http://www.nature.com/doifinder/10.1038/nature11352

8/11/2019 nrde-3

2/5

8/11/2019 nrde-3

3/5

pre-mRNA expression from many germline target genes inhrde-12/2

animals, indicating that the RNAi inheritance machinery silencesgermline target genes co-transcriptionally during the normal courseof reproduction (Fig.3e). These dataindicatethatHRDE-1 contributesto H3K9me3 in the germ line and support a model in which HRDE-1uses endogenous 22G-siRNAs as specificity factors to direct nuclearRNAi in germ cells.

We asked whether HRDE-1-mediated nuclear RNAi at germlinetarget genes was heritable. We out-crossed hrde-12/2 animals withwild-type animals, isolated hrde-12/2 progeny, and conductedH3K9me3 ChIP on these hrde-12/2 animals and their progeny.H3K9me3 at germline target genes was progressively lost over genera-tions in hrde-12/2 animals (Fig. 3f and Supplementary Fig. 14).Similarresults were seen with nrde-12/2 animals (Supplementary Fig. 14).Coincident with a loss of H3K9me3, germline target gene overexpres-sion became more pronounced in late generation hrde-12/2 animals(Supplementary Fig. 15). These data show that the RNAi inheritancemachinery transmits endogenous gene regulatory information acrossgenerational boundaries.

Why an organism might transmit gene regulatory information

across generations is an intriguing question. During the course of our

studies, we noticed that our RNAi inheritance-defective strains wouldperiodically become sterile; stock plates would contain hundreds ofadults, but no progeny. We proposed that the RNAi inheritancemachinery might be required to maintain the integrity of the germ-celllineage. To test this idea, we out-crossed two independently isolatedalleles each of hrde-12/2 and nrde-1/2/42/2 to wild-type and thenmonitored fertility across generations. After out-crossing, hrde-12/2

and nrde-1/2/42/2 animals exhibited near wild-type fertility (earlygenerations), but became sterile in subsequent generations (lategenerations)(Fig.4a andSupplementary Fig. 16). These data show thatRNAi inheritance-defective animals exhibit a mortal germline (Mrt)phenotype16. Animals lacking the somatic AGO NRDE-3 were not Mrt(Supplementary Fig. 17).The Mrtphenotype ofhrde-12/2 animalswastemperature sensitive:hrde-12/2 animals were Mrt at 25 uC, but not at20 uC, indicating that growth at higher temperatures is required toreveal defects associated with HRDE-1 loss (Supplementary Fig. 18).Most late-generation hrde-12/2 mutants (grown at 25 uC) failed toproduce mature oocytes or sperm, showing that one reasonhrde-12/2 animals do not produce progeny is owing to defects ingametogenesis (Fig. 4b and Supplementary Fig. 19). Twenty-six per

cent of late generationhrde-12/2 animals were able to produce sperm

a b

(RPKM)

H3K9m

e3

K10C8.1

nrde-2/4-dependent genes

de

c

HRDE-1-associatedsmall RNAs

26 nt

18 nt

f

54-htab

5.4D51F

8.0520B

6.C3H64Y

21.AL9A16Y

3.21C20F

1.8C01K

2-src R

1

1.8A1

01.B21A17Y

7.3E62T

8-pnr

1-plg

2-dlg

Pre-mRNAmRNA

Foldchange(log2)

RNAabundance

hrde-1

(tm

12

00

)/WT

HRDE-1 targets

H3K9m

e3ChIP(%

ofinput)

0.2

0.4

0.6

0.8

bath-4

5 5.4D51F

8.0520B

6.C3H64Y

21.AL9A16Y

3.21C20F

1.8C01K

2-src R

1

1.8A1

01.B21A17Y

7.3E62T

3-tfe

2-scn

1-til

1-klt

WThrde-1(tm1200)

HRDE-1 targets Non-HRDE-1targets

2

0

2

4

6

8

nrde-2/

WT

0.1

2

820

nrde-2

0.1

2

820

nrde-4

0.1

2

8

20

glp

-4

0.1

2

8

20

nrde-3

N20.1 2 8 20

N20.1 2 8 20

N20.1 2 8 20

N20.1 2 8 20

100

80

60

40

20

0hrde-1siRNA

1210

86420

12,69612,69212,68812,684 12,700 (kb)

Non-HRDE-1targets

0

1

2

3

4

5

6

7

8

9

10

Change

inH3K9me3(WT/hrde-1

)

P0F

1

F2

F6

5,000 0 5,000 (bp)

bath-45

Figure 3| The RNAi inheritance machinery transmits endogenousepigenetic information across generations. a, H3K9me3 in mutant (y-axis)versus wild type (N2,x-axis). Levels of H3K9me3 at C. elegansgenes are theratio of immunoprecipitated nucleosomes to input nucleosomes.smg-1(positive control)8 (blue dot) andnrde-2/4-dependent genes (red) arehighlighted. N2 andnrde-2ChIP-seq data were published previously (GeneExpression Omnibus (GEO) accession number GSE32631)8.b, FlagHRDE-1co-precipitating RNA was 32P-radiolabelled and analysed by polyacrylamidegel electrophoresis (PAGE). nt, nucleotide.c, Example of an HRDE-1 targetgene. RPKM, reads per kilobase (kb) per million mapped reads.d, rtPCRquantification of H3K9me3 ChIP (wild-type orhrde-12/2 animals grown at25 uC ) at 11 genes targeted byhrde-1siRNAs (HRDE-1 targets). Four genes,which are not targeted byhrde-1siRNAs and exhibit NRDE-independentH3K9me3 in the germ line (non-HRDE-1 targets), do not exhibit H3K9me3

loss, showing that H3K9me3 loss inhrde-1 mutants is not simply due to loss ofgerm cells. Data are expressed as the percentage of input DNA recovered byChIP; themean6 s.d.are shown (n5 3). e, Total RNA was isolated from wild-type orhrde-12/2 animals (25 uC) and rtPCR was used to quantify mRNA orpre-mRNA levels from11 HRDE-1-targetedgenes. Threegenes not targeted byHRDE-1 siRNAs, but expressed in germ cells, are also shown. Data arenormalized tonos-3mRNA (germ line only) and expressed as a ratio ( hrde-12/2/wild type); the mean6 s.d. are shown (n5 3).f,dpy-172/2 orhrde-12/2;dpy-172/2 animals were out-crossed five times, and dumpy (Dpy) adultanimals (P0) andadultprogeny(F1, F2, F6) were isolated (20uC) andH3K9me3was quantified by rtPCR. Data are expressed as a ratio (wild-type/hrde-12/2),and mean6 s.e.m. are shown (n5 1 for P0and F1, andn5 3 for F2and F6).Note, in this panel increasedH3K9me3 signalmeansloss of H3K9me3 in hrde-12/2 animals.

LETTER RESEARCH

2 0 S E P T E M B E R 2 0 1 2 | V O L 4 8 9 | N A T U R E | 4 4 9

Macmillan Publishers Limited. All rights reserved2012

8/11/2019 nrde-3

4/5

and oocytes (Fig. 4b). Most of thesegametes, however, are unlikely to befunctional as the fecundity ofhrde-12/2 animals in this late generationwas only 1% of that of wild-type animals (Fig. 4a). Finally, late genera-tion hrde-12/2 animals exhibited a high incidence of male (Him)phenotype, suggesting that loss of RNAi inheritance may cause defectsin chromosome pairing and/or segregation (Supplementary Fig. 19).We conclude thatthe RNAiinheritancemachinery is requiredto main-tain the immortality of the germ line and that, over generations,disabling the RNAi inheritance machinery causes progressive anddiverse defects in germ-cell formation and function.

Here we show that C. elegans possess dedicated regulatory

machinery that promotes an epigenetic memory of RNAi-silencingevents that occurred in distant ancestors (Supplementary Fig. 1).The AGO HRDE-1 is at the heart of this process, binding heritablyexpressed specificity determinants (siRNAs) to direct nuclear RNAiand promote RNAi inheritance in germ cells. Nuclear RNAi alsopromotes RNAi inheritance in somatic tissues7, indicatingthatnucleargene silencing eventsare keydeterminants of RNAi memoryin diversecell types. Finally, we show that the germline RNAi inheritancemachinery transmits endogenous epigenetic information acrossgenerational boundaries while promoting germline immortality (Sup-plementary Fig. 1). Our data suggest a model in which endogenousheritable RNAs that engage HRDE-1 act as specificity factors to directepigenomic maintenance and immortality of the germ-cell lineage.Further work is needed to determine howdefects in epigenomic main-

tenance relate to germline mortality (see Supplementary Discussion).

We note, however, that both processes depend on the samecomplement of factors (hrde-1 and nrde-1/2/4), andin animals lackingthese factors, defects in epigenome maintenance and defects in germ-cell viability are coincident over generational time. Therefore, we pro-pose that one biological function of the RNAi inheritance machinery isto transmit germline immortality small RNAs, selected in previousgenerations for their ability to promote fertility, across generationalboundaries to promote fertility in future generations.

Note added in proof: HRDE-1 was recently shown to act down-stream of piRNAs to direct a multigenerational epigenetic memoryof piRNA silencing in the germ line1921.

METHODS SUMMARYRNAi.RNAi experiments were conducted as described previously11. Theoma-1andpos-1constructs were taken from the Ahringer library.

RNA immunoprecipitation. RNAimmunoprecipitations (RIPs) wereperformedas described previously11, with the exception that adult animals were used for allRIPs. Adult animals were frozen anddounced 10 times before RIP. FlagNRDE-2protein was immunoprecipitated with an anti-Flag M2 antibody (Sigma, A2220).ChIP.ChIP experiments were performed as described previously12, except thatgravid adultanimals were used. Worms were frozen before cross-linking and weredounced 10 times before sonicating. The H3K9me3 antibody was from Upstate(07-523).

oma-1 siRNA TaqMan assay. The TaqMan assay was performed as describedpreviously7. TaqManprobe set1 wasusedin Fig. 2a (see Supplementary Methods).Unless indicated otherwise, the following mutant alleles were used in thisstudy: hrde-1(tm1200), nrde-1(gg088), nrde-2(gg091), nrde-3(gg066) andnrde-4(gg129).

Received 14 December 2011; accepted 28 June 2012.

Published online 18 July 2012.

1. Reik, W., Dean, W. & Walter, J. Epigenetic reprogramming in mammaliandevelopment. Science293,10891093 (2001).

2. Jablonka, E. & Raz, G. Transgenerational epigenetic inheritance:prevalence,mechanisms, andimplicationsfor thestudyof heredity andevolution. Q. Rev. Biol.84,131176 (2009).

3. Fire, A. et al.Potent and specific genetic interference by double-stranded RNA inCaenorhabditis elegans. Nature391,806811 (1998).

4. Grishok,A., Tabara, H. & Mello, C. C. Genetic requirements forinheritance of RNAiin C. elegans. Science287,24942497 (2000).

5. Vastenhouw,N. L.etal. Gene expression:long-termgene silencing by RNAi. Nature442, 882 (2006).

6. Alcazar, R. M., Lin, R. & Fire, A. Z. Transmission dynamics of heritable silencinginduced by double-stranded RNA inCaenorhabditis elegans. Genetics 180,12751288 (2008).

7. Burton,N. O.,Burkhart,K. B. & Kennedy,S. NuclearRNAi maintains heritable genesilencing in Caenorhabditis elegans.Proc. Natl Acad. Sci. USA 108,1968319688(2011).

8. Gu, S. G., Pak, J., Guang, S., Maniar, J. M., Kennedy, S. & Fire, A. Amplification ofsiRNA in Caenorhabditis elegansgenerates a transgenerational sequence-targeted histone H3 lysine 9 methylation footprint.Nature Genet.44,157164(2012).

9. Yigit,E.et al. Analysis of theC. elegansArgonaute family reveals thatdistinctArgonautes act sequentially during RNAi.Cell 127, 747757 (2006).

10. Gu, W.et al. Distinct argonaute-mediated 22G-RNA pathways direct genomesurveillance in the C. elegansgermline.Mol. Cell36,231244 (2009).

11. Guang,S. etal.AnArgonautetransports siRNAsfrom thecytoplasmto thenucleus.Science321,537541 (2008).

12. Guang, S.et al.Small regulatory RNAs inhibit RNA polymerase II during theelongation phase of transcription.Nature465,10971101 (2010).13. Burkhart, K. B.et al. A pre-mRNA-associating factor links endogenous siRNAs to

chromatin regulation.PLoS Genet.7, e1002249 (2011).14. Beanan,M. J. & Strome,S. Characterizationof a germ-lineproliferation mutationin

C. elegans. Development116,755766 (1992).15. Pak, J. & Fire, A. Distinct populations of primary and secondary effectors during

RNAi in C. elegans. Science315,241244 (2007).16. Ahmed, S. & Hodgkin, J. MRT-2 checkpoint protein is required for germline

immortality and telomere replication in C. elegans. Nature403,159164(2000).

17. Tabara, H. et al.Therde-1 gene, RNA interference, and transposon silencing inC.elegans. Cell 99,123132 (1999).

18. Tabara, H., Yigit, E., Siomi, H. & Mello, C. C. The dsRNAbinding protein RDE-4interacts withRDE-1, DCR-1, anda DExH-boxhelicase to direct RNAiin C. elegans.Cell109,861871 (2002).

19. Bagijn, M. P. et al. Function, targets, and evolution of Caenorhabditis eleganspiRNAs.Science http://dx.doi.org/10.1126/science.1220952 (4 June 2012).

20. Ashe, A.et al.piRNAs can trigger a multigenerational epigenetic memory in thegermline of C. elegans. Cell150,8899 (2012).

0

2040

60

80

100

120

140

160

180

200

nrde-1/

nrde-2/

hrde-1/

P1

+/

F1

/

F2

/

F3

/

F4

/

F5

/

Broodsize/animal

WT

P0

/

a

Generation

Sperm and oocyte positive (100%)

Oocytes Sperm

DNAOocytesSperm

*

b

*

*

*

*

*

Oocytes

Sperm

*

Oocytes

Sperm

*

F1

F5

Sperm and oocyte positive (100%)

Sperm only (35%)

Sperm and oocyte positive (100%)

P0

hrde-1+/+

P0hrde-1/

P1hrde-1+/

F1

F5

Sperm and oocyte positive (26%)

Sperm and oocyte negative (30%)

Figure 4| The RNAi inheritance machinery promotes germlineimmortality. a, Animals of indicated genotypes were out-crossed to wild-typetwo to four times, and brood sizes scored across generations at 25 uC. Data aremean6 s.e.m. (n5 5). Note, for unknown reasons,nrde-1and nrde-4mutantsare Mrt at both20 uC and 25 uC, but hrde-1 and nrde-2mutants areMrt only at,25 uC (n5 4).b,hrde-12/2 animals were out-crossed three times, and hrde-11/1 orhrde-12/2 siblings were isolated. Gonads were isolated and stainedwith 49,6-diamidino-2-phenylindole (DAPI; DNA), and immunofluorescence

was used to detect sperm (green) and oocytes (red) (see Methods) in F1and F5generations. Scale bar denotes 100 mm, and this is applicable to all images in b.

RESEARCH LETTER

4 5 0 | N A T U R E | V O L 4 8 9 | 2 0 S E P T E M B E R 2 0 1 2

Macmillan Publishers Limited. All rights reserved2012

http://dx.doi.org/10.1126/science.1220952http://dx.doi.org/10.1126/science.1220952

8/11/2019 nrde-3

5/5

21. Shirayama, M. et al. piRNAs initiate an epigenetic memory of nonself RNA in theC. elegansgermline. Cell150,6577 (2012).

Supplementary Informationis linked to the online version of the paper atwww.nature.com/nature.

AcknowledgementsWe thankP. Anderson, H. Opalicious andD. Wassarmanfordiscussions. We thank S. Ahmed and members of the Ahmed laboratory for sharingunpublished dataconcerning therole of nrde-1 in germline immortality. This workwassupported by grants from the Pew and Shaw scholars programs, and the NationalInstitutes of Health, GM88289 (S.K.), GM37706 (A.F.) and GM069454 (J.K).

Author ContributionsB.B. contributed to Figs 1ac, 2bd and Supplementary Figs 3,4, 5b, 6, 8, 10 and 13. K.B. contributed to Figs 2c, 3df, 4a, b and Supplementary

Figs 1315, 16ac, 17, 18 and 19c. S.G.G. and A.F. contributed to Fig. 3ac,Supplementary Table 2 and Supplementary Figs 11 and 12. G.S. contributed toSupplementary Figs 2, 5a and 16d.A.K. and J.K contributed to Fig. 4b andSupplementary Figs 7 and 19a. H.F. contributed to Fig. 4a and Supplementary Figs16ac, 17, 18 and 19c. S.K. contributed to Figs 1ad, 2a, 3b, Supplementary Table 1andSupplementary Figs2, 6b,9, 10band 19b.S.K., B.B.andK.B. wrote themanuscript.

Author Information ChIP-seqand hrde-1 siRNAdata havebeensubmitted to theGeneExpression Omnibus (GEO) under accession number GSE38041. Reprints andpermissionsinformationis available at www.nature.com/reprints. Theauthors declareno competing financial interests. Readers are welcome to comment on the online

version of this article at www.nature.com/nature. Correspondence and requests formaterials should be addressed to S.K. (sgkennedy@wisc.edu) .

LETTER RESEARCH

2 0 S E P T E M B E R 2 0 1 2 | V O L 4 8 9 | N A T U R E | 4 5 1

M ill P bli h Li it d All i ht d2012

http://www.nature.com/naturehttp://www.nature.com/reprintshttp://www.nature.com/naturemailto:sgkennedy@wisc.edumailto:sgkennedy@wisc.eduhttp://www.nature.com/naturehttp://www.nature.com/reprintshttp://www.nature.com/nature