Lau IJBCB PiwiReview2010

The International Journal of Biochemistry & Cell Biology 42 (2010) 13341347

Contents lists available at ScienceDirect

The International Journal of Biochemistry& Cell Biology

journa l homepage: www.e lsev ier .com

Small R om

Nelson CDepartment of

a r t i c l

Article history:Available onlin

Keywords:piRNAmiRNAsiRNAgermlinePiwi

, pacryonicell dpathwmiRNmement,A acchanents.the fuals ofents.

1. Introduction

Small RNAs and the RNAi pathway are key components of generegulation systems in eukaryotic cells. In animals, the proper devel-opment of sRNAi pathwdistinct frodevelopmeexhibit an ubecome thetory RNAs.

MicroRNtwo promin(Fig. 1). Whger RNAs mand Sonthewith innate(TEs or transomatic celcertain miRMurchisonHowever, tsiRNAs (endgermline-sp

Tel.: +1 61E-mail add

interacting RNAs (piRNAs). Despite their abundance, piRNAs andendo-siRNAs have only recently been unveiled, and many aspectsof their molecular functions are still undened.

To date, nearly all small regulatory RNAs function in a complex

1357-2725/$ doi:10.1016/j.omatic tissues is intimately dependent on a functioningay (Ambros, 2004). Although germ cells are uniquely

m somatic cells because they form gametes, germ cellnt also appears to require RNAi. However, germ cellsnanticipated diversity of RNAi mechanisms and haverichest environment for the discovery of small regula-

As (miRNAs) and small interfering RNAs (siRNAs) areent classes of small regulatory RNAs in eukaryotesile miRNAs regulate many endogenous target messen-ainly through gene silencing (Bartel, 2009; Carthew

imer, 2009), siRNAs are thought to provide organismsdefenses against viruses and transposable elementssposons). miRNAs were largely rst characterized in

ls, but it was anticipated that germ cells would expressNAs specic to the germline (Mishima et al., 2008;et al., 2007; Ro et al., 2007; Watanabe et al., 2006).he surprises lurking in germ cells were endogenouso-siRNAs) not previously seen in somatic cells, and aecic, abundant class of longer small RNAs called Piwi-

7 724 5455.ress: [email protected].

with an Argonaute (AGO) or PIWI protein. Most animal genomescontain multiple homologs of these proteins. Each possesses a PAZdomain, which binds the 3 end of small RNAs, and a PIWI domain,which folds into an RNase H-like fold and can cleave a target RNAstrand that is matched to the guide RNA bound in the domain(Slicer activity, Tolia and Joshua-Tor, 2007). AGO proteins pre-fer to bind2023nucleotide (nt) miRNAs and siRNAs, while PIWIproteins prefer to bind2431nt long Piwi-interacting RNAs (piR-NAs). miRNAs and siRNAs are processed by the Dicer endonucleasefrom precursors with double-stranded RNA (dsRNA) features; piR-NAs are distinct because they are typically longer in length, theirprecursors, in general, lack dsRNA features, and their productiondoes not require Dicer (Ghildiyal and Zamore, 2009).

Several recent reviews discuss the discovery of these smallregulatory RNAs (Ghildiyal and Zamore, 2009; Malone andHannon, 2009; Klattenhoff and Theurkauf, 2008). The miRNA andsiRNA pathways have also been recently reviewed (Carthew andSontheimer, 2009; Kim et al., 2009; Okamura and Lai, 2008). There-fore, this germline-centric review of small regulatory RNAs willfocus on the two newest classes of germ cell small RNAs: piRNAsand endo-siRNAs. I will discuss the history of how germline devel-opment studies have now intersected with the biology of smallregulatory RNAs. Finally, I will address the current challenges inunderstanding small RNA biology within the context of germ cells

see front matter 2010 Elsevier Ltd. All rights reserved.biocel.2010.03.005NAs in the animal gonad: Guarding gen

. Lau

Biology, Brandeis University, 415 South Street, Waltham, MA 02454, USA

e i n f o

e 19 March 2010

a b s t r a c t

Germ cells must safeguard, apportionensure proper reproduction and embmany genes controlling animal germlinked to the RNA interference (RNAi)RNAs. Germ cells contain microRNAs (RNAs (piRNAs); these are bound byknown to specify germ cell developmdevelopment can also affect small RNsurprising diversity of regulatory meling the spread of transposable elemgermline small RNAs and to identifyarea will likely impact biomedical gothe transposition of mobile DNA elem/ locate /b ioce l

es and guiding development

kage, and deliver their genomes with exquisite precision toc development. Classical genetic approaches have identiedevelopment, but only recently have some of these genes beenay, a gene silencingmechanism centered on small regulatoryAs), endogenous siRNAs (endo-siRNAs), and Piwi-interactingbers of the Piwi/Argonaute protein family. piwi genes werebut we now understand that mutations disrupting germlinecumulation. Small RNA studies in germ cells have revealed aisms and a unifying function for germline genes in control-Future challenges will be to understand the production ofll breadth of gene regulation by these RNAs. Progress in thismanipulating stem cells and preventing diseases caused by

2010 Elsevier Ltd. All rights reserved.

N.C. Lau / The International Journal of Biochemistry & Cell Biology 42 (2010) 13341347 1335

Fig. 1. Gonada ll interRNAs that are

and proposcations for

2. Where i

Our undanimals catode CaenordsRNA triggtal regulatoWightmanshown thatbacteria explong after ttransmissioet al., 2000;

The Plastors that disthese mutasterility thawas carriedteinwas amBoth the Mtor strains wwith dsRNAas rde-3, likwere allelic1999).

In a diffegans mutanmutations iRNA polymgal RdRP gedefend agaifound to berespondingother RdRP1, -2, and -3gene silenciRdRP genesterk laboraRNAi respo2005; Simmeri-1 enhandependentmut-7.

Althouggermline de

ergina plnou

riggewasays iays ton limmedmal cthe

all RNNAsns begar2003of ne006)y RdRsphad Fiy thre dintiafurthans gto e

and ionadolfswline

tinctroteihat ll small RNAs that aregeneratedbyDicer.MicroRNAs (miRNAs) andendogenous smaincorporated into Argonaute (AGO) proteins.

e that germ cell focused studies will have broad impli-understanding animal health and evolution.

t began: RNAi and the nematode germline

erstanding of small RNA-mediated gene regulation inn be traced to pioneering discoveries in the nema-habditis elegans (C. elegans), such as the revelation thaters RNAi (Fire et al., 1998) and that the developmen-r genes lin-4 and let-7 encode miRNAs (Lee et al., 1993;et al., 1993; Reinhart et al., 2000). In C. elegans, it wasRNAi can be induced simply by feeding the nematodesressingdsRNA, and that this effectpersists intoprogenyhe source of the dsRNA trigger is removed, suggestingn and amplication of siRNAs in the germline (GrishokSijen et al., 2001; Timmons and Fire, 1998).

terk laboratory isolatednematodemutants calledmuta-played elevated mobilization of transposons; some ofnts, such as mut-7, displayed a temperature-dependentt was most evident when the homozygous mutationby the hermaphrodite, suggesting that the MUT-7 pro-aternally deposited gene product (Ketting et al., 1999).ello and Plasterk laboratories found that some muta-ere unable to mount an RNAi response when injected, while some RNAi-decient (rde) mutant strains, suchewise displayed elevated transposon mobilization andwith mutator genes (Ketting et al., 1999; Tabara et al.,

rent approach, the Maine laboratory isolated a C. ele-t that was severely defective in oogenesis and carriedn the gene ego-1, a homolog of plant RNA-dependenterases (RdRP, Smardon et al., 2000). Plant and fun-nes are necessary for mounting an RNAi response tonst viruses (Chen, 2009),while in nematodes, ego-1was

an emnomenendogeRNAi tcells, itpathwpathwcommsiRNA-

Forbehindest smof miRC. elegaRNAs ret al.,cationet al., 2ated btripho(Pak anstudy btions aprefere

ToC. elegappliedment,in thegvan Wof germare disAGO pRdRP tnecessary for proper meiosis and mitosis, as well as forto dsRNA to initiate RNAi. In addition to ego-1, threehomologs are encoded in the C. elegans genome: rrf-, and mutations in these genes all exhibit a dampenedng response (Smardonet al., 2000). Convergingon theseas well as uncovering new genes, the Ruvkun and Plas-tories discovered genes that enhance or suppress thense in somatic cells (Kennedy et al., 2004; Kim et al.,er et al., 2002). For example, mutations in rrf-3 and

ce RNAi in neurons, and curiously cause a temperature-sterility defect, analogous to the phenotype seen in

h the links between RNAi, transposon regulation, andvelopment in the nematode were not clear at the time,

certain mRmechanism

3. Not justendo-siRNA

Therstto C. eleganassumptionhigher animor RRF-1. NsiRNAgenewithout anan aggressiferingRNAs (endo-siRNAs) areprocessed intomature single-stranded

g hypothesis was that RNA-based gene silencing phe-ayed a role in germ cell formation through putatives siRNAs and protein factors. It was thought that whilered by exogenous dsRNAs was robust in many somaticalsodampenedwhencompetingwithendogenousRNAin the germline. Thus, mutations in endogenous RNAihat lowered endo-siRNA levels were thought to liberate

iting RNAi factors to be used for effective exogenousiated silencing.haracterization of endo-siRNAs from C. elegans laggedbroader discovery of miRNAs, in part because the earli-A cloning methods focused on the 5 monophosphate

and exogenous siRNAs. However, the endo-siRNAs inecame apparent with cloning techniques that captureddless of the status of the small RNAs 5 end (Ambros; Lee et al., 2006; Pak and Fire, 2007) and in identi-w factors involved in small RNA biogenesis (Duchaine. An examination of secondary siRNA molecules gener-Ps in nematodes indicated that they contain 5 di- and

tes, andmight be unprimed RdRP products frommRNAsre, 2007; Sijen et al., 2007). Finally, a deep sequencinge Bartel laboratory suggested that endo-siRNA popula-verse in C. elegans, with 21-, 22-, and 26-nt species thatlly start with a 5 guanosine (Ruby et al., 2006).er understand the repertoire of endo-siRNAs in theermline, deep sequencing efforts have recently beenmbryos, gametes, mutants affecting germline develop-mmunoprecipitates of nematode AGO proteins residents (Claycombetal., 2009;Guetal., 2009;Hanetal., 2009;inkel et al., 2009). These studies codied two classesendo-siRNAs, 26G RNAs and 22G RNAs (Fig. 2), whichin their length, the factors that generate them, and thens that bind them. Both 26G and 22G RNAs require anikely synthesizes an unprimed antisense transcript to

NAs expressed throughout the genome; however, thefor selecting the specic mRNAs is currently unknown.

nematodes: insects and vertebrate gonadals

discoveredmiRNA, lin-4,was once thought to beuniques because of lack of obvious animal homologs; thismay have also been extended to endo-siRNAs becauseals lack an endogenous RdRP homologous to EGO-1

evertheless, long dsRNAs that are substrates for endo-ration canbe formedby single-strandedRNAprecursorsRdRP (Fig. 1). However, most mammalian cells mountve innate immune response to intracellular dsRNA that

1336 N.C. Lau / The International Journal of Biochemistry & Cell Biology 42 (2010) 13341347

Fig. 2. Gonadawith the expanthis nematodebesidesmiRNA(piRNAs) baseproteins of othin gonads, andinitial steps ofthe predominacontain a 5 mphate. The mosiRNAs/piRNA

is indicative2007), withactivate the2000).

Using demouse oocynot only mstantial num2008). The mwhich the mnumber canlarge invertsive complepseudogeneurations progenerate encomplex (T

To validproled in d

scripts like Ran-GAP and Ppp4r1 were up-regulated 24 foldin mutant oocytes. Ran-GAP is a protein implicated in control-ling the molecular gradient of Ran-GTP, which mediates spindleorganization (Kalab and Heald, 2008). The substantial number

o-siRns, lindleison

NAi cmain008)rtly aortshilaal eb; Ced noneidef embra eo-siROkamrestiof endfunctiootic sp(Murchthat Rthis reet al., 2

Shople repDrosopGhildiy2008a,detector Schcells oOkamuy end2009;

Intel small RNAs specic to the nematode Caenorhabditis elegans. In lineded number of Argonaute-family proteins in the C. elegans genome,s gonad also contains an additional layer of small regulatory RNAss and siRNAs. 21URNAsmaybeorthologous to Piwi-interactingRNAsd on the homology of Piwi-related gene-1 and -2 proteins to PIWIer animals. 26G and 22G RNAs are siRNAs most abundantly detectedappear to require an RNA-dependent RNA Polymerase (RdRP) forbiogenesis. These RNAs are named after their signature length andnt 5 nucleotide in that class. While 21U and 26G RNAs likely oftenonophosphate, 22G RNAs predominantly contain a 5 di- or triphos-del here speculates that 26G and 21U RNAs might serve as primarys that then promote generation of secondary siRNAs (22G RNAs).

of a viral infection (interferon response, SenandSarkar,the notable exception being germ cells, which do notresponse when injected with dsRNA (Svoboda et al.,

ep sequencing of small RNAs from many thousands oftes, the Hannon and Watanabe laboratories identiedany known miRNAs and some piRNAs, but also a sub-ber of endo-siRNAs (Tam et al., 2008; Watanabe et al.,ouse oocyte endo-siRNAs are mainly 2122nt long, ofajority map to transposons; in addition, a signicantalso be mapped to pseudogenes that either contain a

ed repeat (e.g. theRan-GAPpseudogene)or showexten-mentarity to an endogenous gene (e.g. Ppp4r1 and its) (Fig. 1). The groups postulated that these gene cong-duce dsRNA precursors that are processed by Dicer todo-siRNAs that are loaded into theArgonaute-2 (Ago-2)am et al., 2008; Watanabe et al., 2008).ate a function for oocyte endo-siRNAs, mRNAs wereicer-null and ago-2-null oocytes. Predicted target tran-

cells map tomdg1, 297,Ghildiyal eaccount onling TE-assoendo-siRNAet al., 2008siRNAsmapmagnitudesus siRNAsFor examplblood are acorrespondal., 2008). Itranscriptio

4. A small

Studiessmall RNAsin animals;NAs) (Arav2005). ThesinteractingRNAs co-pumals (Aravand Zamorbeen next-al., 2009), wof individuaalso been dworm, sponet al., 2008et al., 2009yielded moorigins.

Piwi-intbased on wI piRNAs prmap uniqustrand bias1100kilobNAs that target genes with microtubule-associatedike Ran-GAP, as well as the pronounced defects in mei-

formation previously observed in dicer-null oocyteset al., 2007; Tanget al., 2007), have led to the suggestionan modulate the microtubule cytoskeleton, althoughs to be tested molecularly (Tam et al., 2008; Watanabe.fter mouse oocyte endo-siRNAs were described, multi-followed with deep sequencing of endo-siRNAs frommelanogaster (D. melanogaster) (Czech et al., 2008;t al., 2008; Kawamura et al., 2008; Okamura et al.,hung et al., 2008). Intriguingly, endo-siRNAs weret only in ovaries, but also in somatic cells like y headsr-2 cultured cells, which are regarded as y somaticryonic origin (Czech et al., 2008; Ghildiyal et al., 2008;

t al., 2008a,b). Nuances on the origin and biogenesis ofNAs in somatic cells have been reviewed (Kim et al.,ura and Lai, 2008; Ghildiyal and Zamore, 2009).ngly, many endo-siRNAs from germline and somaticthe same transposons that are targeted by piRNAs (e.g.

blood, and others; Chung et al., 2008; Czech et al., 2008;t al., 2008; Kawamura et al., 2008). When taking intoy uniquely mapping TE-associated siRNAs or normaliz-ciated siRNA reads to the number of genomic loci, manymatches still accumulate on piRNA cluster loci (Chung; Czech et al., 2008). However, the number of endo-ping topiRNAclusters canbeestimated tobeanorderoflower than piRNAs, and the distribution of piRNAs ver-mapping to consensus TE sequences differ signicantly.e, over 90% of the piRNAs corresponding to mdg1 andntisense matches, while only 60% of the endo-siRNAsing to these elements are antisense matches (Chung ett is unclear if endo-siRNAs are derived from the samenal loci as piRNAs.

piece of piRNA background

in y and zebrash embryos were the rst to describecomplementary to repetitive transposable elementsthese were named repeat-associated siRNAs (rasiR-

in et al., 2001, 2003; Elbashir et al., 2001; Chen et al.,e small RNAs, as a class, have been reclassied as Piwi-RNAs (piRNAs) because these rasiRNAs and other smallrify with PIWI-group proteins from ies and mam-

in et al., 2007a; Malone and Hannon, 2009; Ghildiyale, 2009). The revolution in characterizing piRNAs hasgeneration deep sequencing technology (Morozova ethich has revealed the extraordinary sequence diversityl piRNAs in ies and mammals. Recently, piRNAs haveeeply sequenced in zebrash, frog, monotremes, silk-ges, and sea anemone (Grimson et al., 2008;Murchison; Armisen et al., 2009; Houwing et al., 2007; Kawaoka; Lau et al., 2009a), but studies in y and mouse havest of our understanding of piRNAs and their genomic

eracting RNAs can be classied into two main typeshere the piRNAs originate in the genome (Fig. 3). Classedominantly begin with a 5 uridine (U), and generallyely as clusters to genomic regions with a pronounced, such that single-stranded precursors ranging fromases (kb) are thought to give rise to a huge variety


Fig. 3. Classi ic regipiRNAs. Class esis ofrom mRNAs ( ion (3the ribosome f hat pr

of piRNAs.spermatocygressing thrGirard et al.I piRNA cludivergent trdomains of2006). Maplarge strandloci, sinceproductionever, thesebecausema

Class II pthey have atransposonbegins withond form csense stranbase-pair tosense piRNAone PIWI hare antisendominantly(AGO3) (Bre

When Athe Slicer athe TE tranporation intranscript mpiRNAs incowell undersmodel has(Fig. 3), becamplify piR2009). Althoclear whethclass I overto TEs and cpiRNAs areLau et al., 2

MammaMiwi, Mili,germline, w

Wataiwi2ellsoniaspers Mieralumbtestomidst toAs, tumeclus008)entlyands expternapprrt ofary, a(Lauco pican btoriecation of Piwi-interacting RNAs. Multi-kilobase clusters of piRNAs lying in intergenII piRNAs are thought to be generated by a ping-pong mechanism, whereas biogengenic piRNA precursors) appear to preferentially arise from the 3 untranslated regor access to mRNAs or cryptic promoters may reside upstream of certain 3 UTRs t

These piRNAs are typically abundant in adult rodenttes and spermatids when the male germ cells are pro-ough thepachytenestageofmeiosis (Aravinet al., 2006;, 2006; Lau et al., 2006). Inmammals, someof these classsters exhibit a bi-directional conguration that impliesanscription, with a 0.11kb region separating the twopiRNAs (Aravin et al., 2006; Girard et al., 2006; Lau et al.,ping of class I piRNAs in D. melanogaster also revealed-biased clusters of piRNAs described as master controlmutations near these loci have broad effects on piRNAand transposon control (Brennecke et al., 2007). How-loci also contribute to the production of class II piRNAsny transposon relics have accumulated in these regions.iRNAs generally map to multiple genomic loci becausehigher propensity to match repetitive elements like

s. These piRNAs are composed of one form that oftena 5 U and is antisense to TE messages, while a sec-

ontains adenine (A) at position 10 and is the samed as TE messages (Fig. 3). The class II piRNA forms caneach other in the rst 10nt, such that the U of the anti-matches the A of the sense piRNA. In D. melanogaster,

omolog, Aubergine (Aub), mainly binds piRNAs thatse to TEs, while piRNAs that are sense to TEs are pre-bound by another piRNA-binding protein, Argonaute-3nnecke et al., 2007; Gunawardane et al., 2007).ubpiRNA complexes seek transcripts from active TEs,

2008;Mili, Mgerm cmatogWhenexpresare gensmall nmouseium brcontraII piRNmore npiRNAet al., 2

Recin iesovarieare maferenta cohothe ovmencoamentesteslaboractivity in Aub that is guided by the piRNA can cleavescript to dene the 5 end of a new piRNA for incor-to AGO3. The AGO3piRNA complex reciprocates on aade from master control loci to dene the 5 ends ofrporated intoAub. Additional processing events not yettood help dene the 3 end of the piRNA. This circularbeen called the ping-pong amplication mechanismause two PIWI proteins ping-pong off of each other toNA biogenesis (Kim et al., 2009; Malone and Hannon,ugh Aub and AGO3 clearly bind class II piRNAs, it is noter Piwi, the founding member of PIWI proteins, prefersclass II piRNAs. However, its piRNAs are also antisensean accumulate abundantly evenwhenAGO3-associatedpoorly expressed or absent (Gunawardane et al., 2007;009b).ls also contain three PIWI homologs, and in mice theand Miwi2 proteins are all expressed in the malehereas oocytes appear to express only Mili (Tam et al.,

piRNAs acrmRNAs, suof the mRNadditional cthat remain

5. Nemato

Most anII types; hoelegans andtheir initialdomains ofupstream oeach indivitranscriptio

At rst glength andons give rise to both class I (primary) piRNAs and class II (secondary)f class I piRNAs is unclear. However, a group of class I piRNAs derived UTRs) of the mRNAs, suggesting PIWI proteins might compete withomote transcription of an independent transcript.

nabe et al., 2008). In early stages of spermatogenesis,, and class II piRNAs are expressed in the primordialthat develop and divide by mitosis into primary sper-(Aravin et al., 2003, 2007b, 2008; Carmell et al., 2007).matocytes enter meiosis to become spermatids, theywi as well as a burst of class I piRNAs. Class II piRNAsly more difcult to detect in mouse, perhaps due to theer of early germ cells, whereas class I piRNAs in adult

es are so abundant that they can be detected by ethid-e staining (Aravin et al., 2006; Girard et al., 2006). In

y master control loci, which yield both class I and classhe clusters of mouse class II piRNAs are often distinct,rous, and smaller in genomic coverage than the class Iters (Aravin et al., 2007b, 2008; Kuramochi-Miyagawa., newdistinctions in class I piRNAshave come into focusmice. In contrast to adult mouse testes, D. melanogasterress abundant levels of class II piRNAs, of which manyally deposited into the embryo. Three groups using dif-oaches to partition class II piRNAs have determined thatclass I piRNAs are present in the somatic follicle cells ofnd many derive from a master control locus called a-et al., 2009b; Li et al., 2009; Malone et al., 2009). TheRNAs and bulk of class I piRNA clusters in mouse adulte considered intergenic piRNA clusters. The Lau and Lais have now observed that signicant numbers of class I

oss diverse animals can also derive from many specicch that the piRNAs preferentially map to the 3 UTRsAs (Robine et al., 2009). These recent ndings underlieomplexities in this germline gene regulatory pathwayto be uncovered.

de piRNAs are different: 21U RNAs

imal piRNAs conformto the categories of class I and classwever, an atypical class of piRNAs were revealed in C.named 21URNAs because of their uniform21nt length,5U, and their genomic origin as a cohort from largechromosome IV (Ruby et al., 2006). A consistent motiff the sequence encoding the 21U RNAs was detected fordual 21U specimen, but whether this motif serves as anpromoteror aprocessing signal is currentlyunknown.lance, 21U RNAs might seem like endo-siRNAs in theirthe apparent derivation from either genomic strand, a


hallmark for a dsRNA precursor. However, three groups recentlydemonstrated that 21U RNAs are actually Piwi-interacting RNAs:mutants lacking functional Piwi-related gene-1 (prg-1) lack 21URNAs (Batista et al., 2008; Das et al., 2008;Wang and Reinke, 2008),the immunand the biog2008; Das eistics of 21UPRG-1 exhibacid sequenand concenexhibit temet al., 2008;of 21U RNAposons in thet al., 2008represent a

Are we nPerhaps notabundant sppiRNAs obsNAs becameclass I piRNMiyagawaeof piRNA-lago-2 is genhas been cletranscriptioembryonicproportionet al., 2009a straightfoever, focuseorganelles,yield insigh

6. The ridd

Although(Figs. 1 andpiRNA biogtodes demo(Vagin et aDas et al.,pled from mmodel for abiogenesispiRNAs can2007). Neveprimary piRnism? Howprecursor tnuclease doindening tularly demoKlattenhoff

Interestianimals examodicatioOhara et al.al., 2008). Ilase called DAlthough aessary to stDmHen1/Pi

phenotypes of the null mutant are modest (Horwich et al., 2007;Saito et al., 2007).

Many PIWI-associated factors and genes genetically linkedwithpiRNA biogenesis have been described, but their direct impact on

funcs enA he

u; anLim auc a2006ogs oA b8), w, Kif1et aintentermicsells hor m2009Nishindin

someal gal coocheownon aen deand9; Rproteof thatoid009asilual Tbioga et alar f

ogs ren thrumnadow iNAdingrecisthatalian2006tif dce metecton fas as tPerhcel

er qus inerasechyteneermoprecipitates (IPs) of PRG-1 contain these small RNAs,enesis of 21U RNAs is Dicer-independent (Batista et al.,t al., 2008). Although the length and genomic character-

RNAs are quite divergent from other animal piRNAs,itsmany similaritieswith PIWI proteins besides aminoce similarity. PRG-1 expression is germline-restrictedtrated in germ cell foci called P-granules, and mutantsperature-sensitive sterility due to germcell loss (BatistaDas et al., 2008;WangandReinke, 2008). Curiously, lossexpression only up-regulates the Tc3 class of trans-e worm germline (see below, Batista et al., 2008; Das

). Thus, 21U RNAs and their loci from chromosome IVn extreme conguration of a piRNA class.ear the saturation point for discovery of small RNAs?, since abundant RNAs masked the discovery of the lessecies inpast deep sequencingefforts. Inmice, the class I

cured initial detection of class II piRNAs, but class II piR-apparent in neonatal and young testes libraries, when

As were absent (Aravin et al., 2007b, 2008; Kuramochi-t al., 2008). The Zamore laboratoryhas suggested a classike RNAs detectable in D. melanogaster heads whenetically ablated and the background of endo-siRNAsared away (Ghildiyal et al., 2008). Finally, a new class ofn start site small RNAs have been uncovered in mousestem (ES) cells, but these species represent a minisculeof the sequenced libraries (Seila et al., 2008; Fejes-Toth). Increasing the depth of library sequencing may berward approach to future small RNA discovery. How-d library construction from specic stages or cell types,or biochemical fractions of ribonucleoproteins will alsot into the diversity of small RNAs.

les of germline small RNA biogenesis

our picture of miRNA/siRNA biogenesis is detailed2; Kim et al., 2009), we only have a few insights intoenesis (Fig. 3). Studies in ies, zebrash, and nema-nstrated that piRNA production does not require Dicerl., 2006; Houwing et al., 2007; Batista et al., 2008;2008); this suggests that piRNA biogenesis is uncou-iRNA and siRNA biogenesis in animals. The ping-pongmplication of piRNA biogenesis explains why piRNAdoes not require Dicer and how the prevalent 5 U ofbe dened (Brennecke et al., 2007; Gunawardane et al.,rtheless, several questions remain: How does efcientNA production occur without the ping-pong mecha-exactly are piRNA 3 ends dened? How are piRNA

ranscripts selected? It has been hypothesized that themain-containing proteins Zuc and Squmight play a rolehe3 endof piRNAs; however, this remains to bemolec-nstrated (Pane et al., 2007;Ghildiyal andZamore, 2009;and Theurkauf, 2008; Li et al., 2009).ngly, the 3 ends of piRNAs (and endo-siRNAs) in allmined are processed to contain a 2 hydroxyl (2OMe)n (Houwing et al., 2007; Kirino and Mourelatos, 2007;, 2007; Vagin et al., 2006; Ruby et al., 2006; Grimson etn D. melanogaster, this is mediated by a ribose methy-mHen1/Pimet (Horwich et al., 2007; Saito et al., 2007).

2OMe modication of plant siRNAs and miRNAs is nec-abilize these small RNAs (Yu et al., 2005), the role ofmet and the 2OMe modication is unclear because the

piRNAprotein(the RNand sq2007;Vasa, Zet al.,homolin piRNal., 200(RecQ1(Kotaja

Thetional iproteogerm csingleet al.,2009;The foutudor,abnormmaternand bibeenshThomshas beAGO3,al., 200Tudorizationchromet al., 22009; VindividpiRNANishidmolecuhomol

Givconundated goFirst, hand piRnon-cobe impregionmammet al.,the mosequenbeen dscriptiregion2008).of germ

Othclusterpolympre-papachytgiven gtion still needs to be evaluated. For example, among thecoded by the y genes implicated in piRNA biogenesislicases vasa, spn-E, and armi; the putative nucleases zucd nuage components krimp and mael; Klattenhoff et al.,nd Kai, 2007; Pane et al., 2007; Vagin et al., 2006), onlynd Squ have been shown to interact with Aub (Megosh; Thomson et al., 2008; Pane et al., 2007). The mousef vasa and mael have also been genetically implicatediogenesis (Kuramochi-Miyagawa et al., 2004; Soper ethile several candidate PIWI-associated factors in mice7b, eIF3a) await further physiological characterizationl., 2006; Lau et al., 2006; Unhavaithaya et al., 2008).rest in PIWI proteins has set off a race to identify addi-acting factors. Recently, several laboratories performingof PIWI protein complexes from y, frog, and mouseave all simultaneously identied a protein factor withultiple Tudor domain(s) (Kirino et al., 2009a,b; Laua; Reuter et al., 2009; Vagin et al., 2009; Chen et al.,da et al., 2009; Vasileva et al., 2009; Wang et al., 2009).g member of the Tudor gene family is D. melanogastermutants of which produce sterile progeny and displayerm cell development due to improper localization ofmponents (Thomson and Lasko, 2005). The structure

mical function of the Tudor protein domain itself hastobindsymmetricallydimethylatedarginines (sDMAs;

nd Lasko, 2005), and the presence of multiple sDMAstermined in the N-terminal regions of Mili, Miwi, Aub,Piwi (Chen et al., 2009; Kirino et al., 2009a; Nishida eteuter et al., 2009; Vagin et al., 2009). The association ofins and PIWI proteins may be important for the local-e PIWI/piRNA complex to germ plasm in oocytes or thebody in mammalian sperm (Kirino et al., 2009a,b; Lau

a; Nishida et al., 2009; Reuter et al., 2009; Vagin et al.,eva et al., 2009;Wang et al., 2009). However, ablation ofudor genes in ies and mice causes varied reduction ofenesis and PIWI protein stability (Kirino et al., 2009a,b;l., 2009; Reuter et al., 2009; Vagin et al., 2009). Thus theunction of D. melanogaster Tudor and vertebrate Tudoremains to be fully elucidated.e diversity of small RNA types in animal gonads, severals vex the eld regarding how germ cells and associ-al cells regulate small RNA expression and production.s the transcription of germ cell-specic miRNA, siRNA,precursors regulated? The promoter elements for theseRNA genes are not dened, and motif prediction cane if the transcription start site (TSS) is unknown. Theseparates the piRNA domains in bi-directional adultpiRNA clustersmayharbor promoter elements (Aravin; Girard et al., 2006; Lau et al., 2006). In contrast toetected upstream of 21U RNAs, however, no obviousotif shared by mammalian bi-directional clusters hased. A recent analysis of histone modications and tran-ctors inmouse ES cells has implicatedCpG-rich genomiche promoters of specic miRNA genes (Marson et al.,aps this methodology can also ascertain the promotersl-specic small RNA genes.estions regarding transcriptional regulation of piRNAclude: Which promoting factors and which RNAs mediate piRNA precursor transcription? How areene piRNA clusters regulated differently from adultpiRNA clusters during germ cell development? Does acell express all piRNA clusters simultaneously, or does


it pick one or several piRNA clusters to express, analogous to howlymphocytes or olfactory neurons pick one immunoglobulin geneor odorant receptor gene, respectively, for expression? This lastquestionhas recentlybeenexplored ina single-cell analysisof smallRNAs fromprole of piferent indivexpressed mexpressionbut were dgests that pbetween di

Overall,understandvitro culturbiochemicabeen lackininparticulaline that reHowever, twbeen foundPIWI proteiderives fromlevels of cla2009b). Comfrom the ovboth Siwi (ito expressSince bothsystems wigenesis andpiRNAs into

7. The imp

Althouget al., 2008eral principdevelopmeproliferateprenatal anstem cells (ferentiatingcells providcells to entcritically, pAfterwardsthe sperm iferentiationdelivery venents suchfor buildingPGCs. Mostthat materngerm plasmwhich blast

8. The in

Althougbroadly impsmall RNAsing into focgermline hafor germ ce

in nematodes result in deformed oocytes (Knight and Bass, 2001),whereas dcr-1 mutations specically in the D. melanogaster femalegermline produce GSCs that fail to self-renew but are able to dif-ferentiate into gametes (Hateld et al., 2005; Jin and Xie, 2007).

ons it miRe thrParkgh th

celor Lt al.,so seemalntiatoliferreguuse mNAsnt fronockrs ard fot al.,c prorans7). Inmlinmat8; Ht comay r

n (Hthe

RNAlez a

ougrentbotpro

zygoZdicon ofrmu005)roteiell degg a7). Aave eNAsams veis a, wher, aight

pmere geiRNANAs2004playlevend aXenopus tropicalis oocytes that compared the overallRNAcluster expressionbetweendifferent eggs fromdif-idual females (Lau et al., 2009a). Interestingly, each eggultiple piRNA clusters simultaneously, and the cluster

proles were similar in eggs derived from one motherifferent from eggs from a different mother. This sug-iRNA cluster expression proles can vary signicantlyfferent individual animals.our grasp of piRNA biogenesis still lags behind ouring of miRNA and siRNA biogenesis, partly because ined cells expressing miRNAs and siRNAs have facilitatedl analysis, while culture systems containing piRNAs hadg. SincepiRNAexpression is largely restricted to gonads,r tomeiotic germcells inmammals, ndingamitotic celltained true gonadal characteristics seemed daunting.o insect cell lines derived from ovaries have recently

to endogenously express abundant levels of piRNAs andns. An Ovary Somatic Sheet (OSS) cell line that likelyD.melanogasterovary follicle cells expresses abundant

ss I piRNAs because it only expresses Piwi (Lau et al.,plementing the OSS cells is the BmN4 cell line derivedaries of the silkworm, Bombyx mori, which expresses.e., silkworm Piwi) and BmAGO3; thus this line appearsclass II piRNAs predominantly (Kawaoka et al., 2009).these cell lines also express miRNAs and siRNAs, thesell be invaluable for functional dissection of piRNA bio-determining how cells partition miRNAs, siRNAs, anddifferent AGO/PIWI proteins.

act of small RNAs in germline development

h animal germline development is complex (Cinalli; Lin, 1997; Seydoux and Braun, 2006), some gen-les have emerged. In the earliest stages of embryonicnt, primordial germ cells (PGCs) are sequestered andin a niche that includes somatic support cells. Duringd juvenile development, PGCs transition into germlineGSCs), which proliferate by self-renewing and/or dif-into secondary germ progenitor cells. Somatic supporte signals andmolecules to germ cells that commit germer meiotic phases like leptotene, zygotene, and, mostachytene, when synapsis of sister chromatids occurs., germ cells further differentiate into haploid gametes:nmales and the oocytes in females. During terminal dif-, sperm shed their cytoplasm and streamline into DNActors. In contrast, oocytes bulk up on maternal compo-as energy-rich molecules and materials that are utilizedthe embryo and for establishing the next generation ofanimal oocytes become polarized during growth, suchal components concentrate in special cytoplasm calledand serve as determinants in the embryo to specify

omeres become PGCs.

uence of miRNAs and endo-siRNAs

h it is known that individual small regulatory RNAs canact animal development (Ambros, 2004), the impact ofas a class on germ cell development is only now com-us. Studies where dicer or ago genes are mutated in theve demonstrated that miRNAs and siRNAs are crucialll development. For example, dicer (dcr-1) mutations

Mutatidisruptenanc2005;Althouto germAgo-1(Park ecells althese fdiffereand prdown-

Moand siRdiffereDicer knumberequireTang emeiotisome tal., 200the gerof speral., 200are nothese mablatioof bothing mi(Gonza2008).

Althis appadicer inable toinitialnull (MinjectiMZdiceet al., 2PIWI pgerm cin theal., 200may hby miR

In mmiRNAtheresiRNAsHowevNAs, mdeveloways aendo-sof miRet al.,siRNAsmRNAdcr-2 an loqs, ago-1, and mei-p26, a regulator of ago-1, all likelyNA function and exhibit similar defects in GSC main-ough the loss of GSC self-renewal (Forstemann et al.,et al., 2007; Neumuller et al., 2008; Yang et al., 2007).e role of themiRNApathwaywasdeemed tobe intrinsicls because somatic support cells expressing functionaloqs failed to rescue the maintenance of mutant GSCs2007; Yang et al., 2007), proliferationof somatic supportemsdependentonmiRNAs (Jin andXie, 2007). Together,e germline studies suggest that miRNAs down-regulateion genes in order to promote germ cell self-renewalation. However, the validation of genes predicted to belatedbymiRNAs in the germline remains to be explored.ale and female germ cells equally depend on miRNAsfor proper development, but for reasons that may bem y germ cells. Although conditional oocyte-specicout mice are infertile, ovary morphology and oocytee similar to wild-type mice, suggesting that Dicer is notr mammalian oocyte growth (Murchison et al., 2007;2007). However, the mutant oocytes are incapable ofgression because the spindles are malformed, whileposons are activated (Murchison et al., 2007; Tang etcontrast to females, male mice with dicer mutations in

e exhibit reduced PGC proliferation and a dramatic lossocytes developing into round spermatids (Maatouk etayashi et al., 2008). Nevertheless, these mutant malespletely sterile and produce a few motile sperm, butesult from germ cells that escape the conditional geneayashi et al., 2008). As in ies, the somatic support cellsmale and female mouse germline require a function-pathway to ensure that germ cells properly developnd Behringer, 2009; Hong et al., 2008; Nagaraja et al.,

h the importance of miRNAs in y and mice germ cells, zebrash buck this trend because sh mutants lackingh the zygotic genome and in maternal contribution areduce viable germ cells that can fertilize and form ante (Giraldez et al., 2005). The maternal-zygotic dicer-er) embryos eventually fail to gastrulateproperly, but ana double-stranded miR-403b duplex into the one-celltant embryo rescues embryonic development (Giraldez. A possible explanation for thismay be that piRNAs andns can compensate for miRNAs and siRNAs in zebrashevelopment, since they are likely maternally depositednd do not depend on Dicer for production (Houwing etlternatively,maternal transcripts in zebrashgermcellsvolved mechanisms that obviate a need for regulation(Mishima et al., 2006).mals, it is more difcult to differentiate the impact ofrsus endo-siRNAs on germ cell development becausesingle Dicer that produces both miRNAs and endo-ile all mammalian AGOs bind both miRNAs and siRNAs.Drosha germline mutant, which should only lack miR-indicate if endo-siRNAs are sufcient for germ cell

nt. In D. melanogaster, the miRNA and siRNA path-netically separable because Dcr-2 and Ago-2 produces independently of the Dcr-1 and Ago-1 production(Forstemann et al., 2007; Lee et al., 2004; Okamura; Tomari et al., 2007). However, the role that endo-in y development remains mysteriousendogenous

ls can be altered when endo-siRNAs are depleted, yetgo-2 mutant ies that likely lack most endo-siRNAs do


not appear overtly different from wild-type ies (Lee et al., 2004;Okamura et al., 2004). TE-associated piRNAs might compensatefor TE-associated endo-siRNAs in the germline, and perhaps theendo-siRNA pathway impacts behaviors and responses to naturalselective prtions.

The impmore evideduction or sdecreased 2Argonautessterility, pret al., 2009accumulatiogametogenfail to segrevan Wolfswucts of thegametesansubset of 22gets. Bearinfungi, whichand Moazeway mightcontrast toprimarily eet al., 2009be highly nucleotidylsible degradexhibit mishistonemet

9. A conse

RNAi hanomenon otransgenesably from atransgenesled to the irepetitive ethousandsmutagenic ctrol is essenandRNAi cogenomic inThis modelsiRNAs in f2007) and rfrom both set al., 2005trigger for pcation of Rdnematodesthemodel (Okamura an

Transposince, ultimson replicaRNAi pathwpathway tothe PIWI parst demonhighly elev

and Kuramochi-Miyagawa, 2009). In contrast, mutations affectingendo-siRNA production in ies do not reveal observable defects ingametogenesis (Ghildiyal and Zamore, 2009; Malone and Hannon,2009; Okamura and Lai, 2008). The exact mechanism for how PIWI

s an, buse tgradlyticctivictivt al.,additnsponal tsilenatedthe7; Kili mIAPin Dmetethy

his, 2sm,ges,et a

se reat b

gh aer Detwellingandted tsh onscriic daHowd zilt to eerthto Tion (enfoat arans pAs a

, butt (Baed thdsRNRNATEs,008pathnctiondd th

ly togast, TARo maomete piessures that are typically absent in laboratory condi-

act of endo-siRNAs on the C. elegans germline is muchnt, as many mutations that abrogate endo-siRNA pro-tability cause varying degrees of sterility. Mutants with6G endo-siRNAs (eri-1, rrf-3, and a doublemutant of theT22B3.2 and ZK757.3) exhibit temperature-dependentesumably due to overexpression of target genes (Han). In addition, mutations that affect 22G endo-siRNAn (cde-1, csr-1, drh-3, ego-1, and ekl-1) result in broad

esis defects because mitotic and meiotic chromosomesgate properly (Claycomb et al., 2009; She et al., 2009;inkel et al., 2009; Gu et al., 2009). The protein prod-

se genes appear to coalesce around chromosomes indblastomeres,whileCSR-1, theAGOprotein thatbindsaG endo-siRNAs, directly interacts with chromatin tar-g similarity to heterochromatic siRNAs in plants andguide AGOs tomodulate chromatin dynamics (Grewal

d, 2003), components of the CSR-1 endo-siRNA path-also direct methylation on histone H3. However, inheterochromatin targets in fungi and plants, CSR-1 isnriched on coding genes (Claycomb et al., 2009; She). Interestingly, the dosage and turnover of CSR-1 mayne-tuned, because mutants in cde-1, which encodes atransferase that uridylates 22G endo-siRNAs for pos-ation, have increased levels of 22G endo-siRNAs and

aligned mitotic chromosomes, perhaps due to ectopichylation (She et al., 2009; vanWolfswinkel et al., 2009).

rved role of keeping transposons at bay

d been recognized as a mechanism to explain the phe-f cosuppression, where the silencing of multi-copycauses the silencing of endogenous genes, presum-berrant dsRNAs arising from the repetitive array of(Henikoff, 1998; Birchler et al., 2000). This hypothesisdea that RNAi could also silence endogenous genomiclements like transposons, which can number in theof copies per element in animal genomes. Given theapabilities of transposonmobilization, transposoncon-tial for the long-term tness of almost all organisms,uldhaveevolvedasanelegantmeans to recognize thesevaders with small RNAs (Goodier and Kazazian, 2008).was conrmed by the discovery of transposon-directedungi and plants (Birchler et al., 2000; Zaratiegui et al.,asiRNAs from ies and sh, all of which appear to arisetrands of transposons (Aravin et al., 2001, 2003; Chen). These studies suggested that dsRNA is the commonrogramming RNAi to target transposons, and identi-RP genes important for transposon control in plants andand the endo-siRNA pathway in y further supportedGhildiyal and Zamore, 2009;Malone andHannon, 2009;d Lai, 2008).

son control may be most critical in animal germ cellsately, germ cells are the essential vehicles for transpo-tion. However, instead of relying on the conventionalay, animal germs cells have evolved the PIWI/piRNAregulate transposons. The function and primacy of

thway in controlling transposons in the germline wasstrated by y mutants that lack piRNAs, which showated TE transcript levels in ovaries and testes (Siomi

proteincidatedantisenand deof cataSlicer aSlicer aSaito e

Ining traadditioposonis elevL1 andal., 200and mL1 andalteredof DNADNA mBourccytoplacell sta(Aravin

TheposonsAlthouand othlinks bcompelishedrestriczebraTE tragenom2008).ziwi anthough

Nevferasesformathelp toNAs thC. eleg21U RNmentsmutanproposprimethis dssilenceet al., 2piRNAPIWI fu

BeyrevealepossibmelanoTAHREother tlize telgenerad piRNAs silence transposon transcripts is not fully elu-t it is presumed that PIWI proteins use piRNAs that areo transposons to recognize TE transcripts for cleavageation. This hypothesis is supported by the conservationresidues between PIWI and AGO proteins necessary forty (Tolia and Joshua-Tor, 2007), and the observation ofity in vitro by PIWI/piRNA complexes (Lau et al., 2006;2006).ion to the slicing mechanism postulated for silenc-sons, work on mouse PIWI genes has suggested anranscriptional silencing pathway formaintaining trans-cing. In the testes of mili and miwi2 mutant mice, thereexpression of retrotransposons like the LINE elementLTR element IAP (Aravin et al., 2007b, 2008; Carmell eturamochi-Miyagawa et al., 2008). Interestingly, miwi2utant mice also display loss of DNA methylation atloci, whereas the proles of IAP-targeting piRNAs arenmt3L knockout mice, which lack a putative regulatorhyltransferase activity during de novo establishment oflation patterns in the fetal male germline (Aravin and008). Although Mili and Miwi are predominantly in theMiwi2 is localized in the nucleus in specic fetal germand this localization is dependent on functional Milil., 2008).sults have suggested that Mili and Miwi2 control trans-oth the transcriptional and post-transcriptional levels.biochemical interaction among Mili, Miwi2, Dnmt3L,NAmethyltransferases has not been shown, the geneticen class II piRNAs and DNA methylation events are, and might explain how silencing of TEs can be estab-maintained in somatic cells when piRNA expression iso the germline. In addition to these ndings in mouse,ocytes decient in Ziwi and Zili also display elevatedpts, and it is hypothesized that mobilized TEs causemage that triggers apoptosis (Houwing et al., 2007,ever, it is unknown if DNA methylation is affected ini mutants, and DNA methylation mechanisms are notxist in invertebrates likeD.melanogaster and C. elegans.eless, plant small RNAs clearly guide DNA methyltrans-E loci, and both entities are critical for heterochromatinZaratiegui et al., 2007; Chen, 2009). RdRPs in plants alsorce TE silencing by converting TE transcripts into dsR-e processed into secondary siRNAs. In this respect, theiRNA pathway shares similarity with plants. Very fewppear to target the major Tc1 and Tc3 transposable ele-secondary siRNAs targeting Tc3 are lost in the prg-1tista et al., 2008; Das et al., 2008). Thus, it has beenat 21U RNAs might help signal RdRPs in C. elegans toA production from aberrant TE transcripts (Fig. 2), andbecomes the source of secondary siRNAs that ultimatelyperhaps at the chromatin level (Batista et al., 2008; Das). Whereas TE silencing is a conserved function for theway in animals, the mechanistic details of piRNA andon may be quite varied among animal species.simply controlling TE mobility, y genetics has alsoat TE control can extend to telomere regulation andspeciation of organisms. In the rst example, D.

er telomeres contain specic retrotransposons (HeT-A,T) and TAS repeats that likely recombine with eachintain telomere length (George et al., 2006). To stabi-

re structure, the transposons and TAS repeats appear toRNAs which direct Piwi and Aub to genetically impose


heterochromatic marks onto the telomeres (Savitsky et al., 2006;Yin and Lin, 2007; Klenov et al., 2007). The heterochromatic marksthat are induced by piRNAs then repress telomeric TEs from mobi-lizing and can silence transgenes inserted into nearby telomeres.piRNA pathity to packexpressioning of tranSavitsky et

A secondpathway isdrome of gofrom a crosisolated labhas been dfemales to ra homologying this gen1999). Onlynotion thatthat endowdrome by qet al., 2008;al., 2004). PNAs and hyKalmykova

In additisis mediateas an evolutfertility. Thbe geneticapiRNA contintact and ththe studymcan have a pbarrier thatalso provokprogeny heexpressionmals (Lau etemperatursyndrome (whether pimental con

10. ClassicpiRNA path

Many geously discovScreens fortied aub, cspn-E (Gillemore recen2007). Thegrouped asbut most ofmally (Gilleal., 1997; Lalso affect Rity in the oonucleates gend of the opolarized mdisrupted (C

and Kai, 2007; Pane et al., 2007; Wilson et al., 1996). Interestingly,the proteins from these germline genes all co-localize at the nuage,a perinuclear zone in nurse cells believed to be an active region ofRNA organization (Lim and Kai, 2007; Klattenhoff and Theurkauf,

Klocdo

pmeve shreve

f DNhoffiggemicroesisn obckgrohofftantlocal9), smac

s. Indth mmalltion (piww dgh oSCs ffertipiwind dtraste celnd so007n imh ges lik1 (H

adral to tsilen006logougermrmatregus. piwminayet faene rArav1). Inlocuanti

y abePalumpn-E20041; Tonismncinutatiand

2005way mutants (piwi, aub, spn-E) appear to lose the abil-age telomeres into stable heterochromatin, allowingand mobilization of the telomeric TEs and loss of silenc-sgenes integrated near telomeres (Josse et al., 2007;al., 2006; Yin and Lin, 2007; Klenov et al., 2007).interesting manifestation of TE control by the piRNAhybrid dysgenesis, which describes a genetic syn-nadal atrophy and sterility in hybrid progeny derived

s of wild-type D. melanogaster males and females fromoratory strains (Kidwell et al., 1977). This syndromeetermined to result from an inability of laboratoryepress certain transposons. Some have postulated that-based mechanism could be the system for mediat-etic phenotype (Chaboissier et al., 1998; Jensen et al.,recently have many groups begun to converge on thesiRNAs and piRNAs may be the critical maternal factorsdaughters with the ability to resist the dysgenic syn-uenching TEs (Blumenstiel and Hartl, 2005; BrenneckeChambeyron et al., 2008; Pelisson et al., 2007; Sarot etublished reviews further detail the link between piR-brid dysgenesis (Malone and Hannon, 2009; Shpiz and, 2009).on to the facet of transposon control, hybrid dysgene-d by piRNAs and the PIWI pathway can also be viewedionary force that drives speciation through modulatinge Hannon laboratory demonstrated that y strains canlly identical but epigenetically dissimilar on the basis ofent (Brennecke et al., 2008). Since the piRNA pathway isebulkofpiRNAsarepresent in sterile dysgenichybrids,ay imply that subtle interplays betweenTEs andpiRNAsrofound and rapid impact in generating a reproductivecould ultimately drive animal speciation. This study

es the question ofwhether piRNAprolesmight impactalth, given that individual differences in piRNA clustermay be quite prevalent even between wild-type ani-t al., 2009a). Finally, parameters like reduced growthe and increased parent age can suppress the dysgenicKidwell et al., 1977); thus, it will be interesting to seeRNA proles or TE activity become altered by environ-ditions.

mutants but new twists: y genes affecting theway

nes affecting germline development in y were previ-ered before they were known to affect piRNA function.female sterility and/or disrupted oocyte polarity iden-uff, vasa, zuc, squ (Schupbach and Wieschaus, 1991);spie and Berg, 1995); mael (Clegg et al., 1997); andtly armi (Cook et al., 2004) and krimp (Lim and Kai,se genes all affect piRNA accumulation and can bea class of related mutations where oocytes still form,these oocytes cannot be fertilized and develop abnor-spie and Berg, 1995; Klattenhoff et al., 2007; Clegg etim and Kai, 2007; Chen et al., 2007). These mutationsNA localization mechanisms that help establish polar-cyte. For example, the maternal mRNA for oskar, whicherm plasm, is not properly localized to the posteriorocytes in these mutants, possibly because the normallyicrotubule network between nurse cells and oocytes ishenet al., 2007; Clegg et al., 1997; Cook et al., 2004; Lim

2008;How

develories hafail to pbers oKlattenthen traffecthypothnizatiothe baKlattenble muOskaral., 200tubulepiRNAact wiwhile sregula

Thewith feAlthouadult Gcan betion ofago-1 aIn conin nurscells aet al., 2has beethrougproteinproteinPal-Bhintegrarole inet al., 2an anathe y

Spebut theoocytethe gerdivide(Ste) gcytes (al., 200gene, amarilydestro2006;armi, s2001,al., 200mechaSte sileloqs mmiRNAet al.,and Etkin, 2005; Lim et al., 2009).piRNAs t into the observed phenotypes of germlinent mutations? The Schupbach and Therkauf laborato-own thatmutants that areunable to accumulatepiRNAsnt TEs from mobilizing, which results in elevated num-A breaks in the female germline (Chen et al., 2007;et al., 2007; Pane et al., 2007). The aberrant DNA breaksr damage signaling checkpoint pathways, which in turntubule network organization. Genetic support for thiscomes from the suppression of microtubule disorga-served when checkpoint factors are also mutated inund of a piRNA pathway mutant (Chen et al., 2007;et al., 2007; Pane et al., 2007). However, these dou-

s in checkpoint pathways do not suppress the loss ofization (Klattenhoff and Theurkauf, 2008; Navarro etuggesting that RNA transport, perhaps utilizing micro-hinery, may be a direct role for PIWI proteins andeed, a sea urchin and a frog PIWI homolog both inter-icrotubules (Rodriguez et al., 2005; Lau et al., 2009a),RNApathways have been shown to rely onmicrotubuleBrodersen et al., 2008; Parry et al., 2007).i gene was discovered in separate screens for mutantsividing germline stem cells (Lin and Spradling, 1997).varies initially develop normally in piwi mutants, theail to self-renewandonly differentiate into oocytes thatlized but are incompetent for embryogenesis. This func-draws similarity to the GSC maintenance functions ofcr-1, but themechanism for this function is still obscure.to the cytoplasmic and perinuclear localization of Aubls, Piwi is predominantly nuclear localized in both germmatic support cells (Cox et al., 1998, 2000; Brennecke; Saito et al., 2006; Lau et al., 2009b). Additionally, piwiplicated in regulationofmulti-copy transgene silencingnetic or direct interactions with chromatin-associatede Polycomb group (PcG) proteins and heterochromatinP1; Brower-Toland et al., 2007; Grimaud et al., 2006;et al., 2004; Yin and Lin, 2007). Since PcG proteins arehe self-renewing capacity ofmouse ES cells due to theircing differentiation genes at the chromatin level (Boyer), it is tempting to speculate that piwi might also serves function in the nucleus of GSCs or other stem cells ofline.ogenesis in ies also depends on piwi and aub function,latorymechanismsare somewhatdistinct fromthose inimutants lack spermbecause GSCs fail to self-renew inl niche (Lin andSpradling, 1997), but aubmutant spermil tomaturebecauseproteincrystals encodedbyStellateepeats accumulate and probably poison the spermato-in et al., 2001, 2004; Schmidt et al., 1999; Stapleton etwild-type y testes, the Suppressor of Stellate [Su(Ste)]s with homology to Ste repeats, produces piRNAs pri-sense to Su(Ste), which then presumably direct Aub torrant Ste transcripts (Nishida et al., 2007; Vagin et al.,bo et al., 1994). The oocyte nuage component genes

, zuc, and squ also inuence Ste silencing (Aravin et al.,; Pane et al., 2007; Schmidt et al., 1999; Stapleton etmari et al., 2004), suggesting that the core AubpiRNAoperates similarly in males and females even though

g is male-specic (Palumbo et al., 1994). Interestingly,ons also de-repress Ste silencing in addition to affectingendo-siRNAproduction (Czech et al., 2008; Forstemann; Okamura et al., 2008b), suggesting that either endo-


siRNAs contribute to Ste silencing, or loqs might participate in thefunction of all three major small RNA types in y germ cells.

Two newly characterized D. melanogaster mutants with decitsin the piRNA pathway have recently been shown to display game-togenic defIsolation ofthe use ofgene in protent with thamplicatioond mutanII piRNAs, pducing piRN(Klattenhofevolving vamatin imm42AB piRNAIn contrastinstead appclusters.

Althoughthe ago3 anmutants ismechanismpiRNAs arein these muallowing copiRNA clusttion of piRN(Lau et al., 2erator of cltwo classesare fewer vof AGO3 an

11. PIWI p

Mouse mtherstwavwhich the cofmeiosis, cmatocytes aPIWI homothat PIWI pthe testes ato the apop(Carmell etal., 2004). Mpersists thrin germ celldpp (Aravinet al., 2004and Miwi extal stage deprophase owhile miwiAs themutathrough unnia also proand Lin, 200

Mouse Pspermatogofected. Thunot exactlymelanogaste

are dispensable for female fertility (Carmell et al., 2007; Deng andLin, 2002; Kuramochi-Miyagawa et al., 2004), although long-termfertility studies of homozygous female mutants have not beendescribed. Since female germ cells only appear to express Mili,

roteioocyalianothealianto mdo

alians upsh pmaleed agerm

ascultermntalgerziw

posto aprm cy, thes futddress, sirodt

ore t

ougI proroteior b

ior loelansim(Kla

tinglin emmbe6). Tskarto se

ontraspossmapHouwr theed ach pan and Etkons apath(Cooalizaly dmpoLinransects similar to those seen in aub and piwi mutants.the ago3nullmutant by the Zamore laboratory requiredreverse genetics, possibly due to the location of theximity to heterochromatin (Li et al., 2009). Consis-e hypothesis of AGO3 acting in the ping-pong piRNAn pathway, the bulk of class II piRNAs are lost. A sec-t called rhino was also found to be decient in classrimarily those deriving from master control loci pro-As from both genomic strands, like the 42AB cluster

f et al., 2009). Rhino was determined to be a rapidlyriant of heterochromatin protein 1 (HP1), and chro-unoprecipitation of Rhino indicated its presence at thecluster (Klattenhoff et al., 2009; Vermaak et al., 2005).

to a chromatin silencing role attributed to HP1, Rhinoears necessary to stimulate expression of piRNAs from

much interesting biology remains to be garnered fromd rhino mutants, an important insight revealed by thesethe partitioning of class I and class II piRNA biogenesiss (Klattenhoff et al., 2009; Li et al., 2009). Since class IIdepleted and class I piRNA biogenesis remains intacttants, the class I piRNAs surface in small RNA libraries,ndent classication of the amenco locus as a class Ier. Independent studies looking at maternal contribu-As (Malone et al., 2009) or examining a follicle cell line009b) have also conrmed amenco as a primary gen-ass I piRNAs; however, the functional partition of theof piRNAs in vertebrates remains unclear, since thereertebrate piRNA mutants and the vertebrate orthologsd Rhino are not obvious.

roteins and vertebrate gametogenesis

ale germ cells initially develop synchronously beforeeofmeiosis begins at 10dayspostpartum(dpp), duringlass II pre-pachytene piRNAs are made. After the onsetlass I piRNAs become the dominant species in late sper-nd spermatids. Constitutive gene targeting of the threelogs in mice, Miwi, Mili, and Miwi2, has demonstratedroteins and piRNAs are essential for male fertility, sincere diminished in each knockout mutant due specicallytosis of germ cells and not the somatic support cellsal., 2007; Deng and Lin, 2002; Kuramochi-Miyagawa etili is expressed as early as PGC and GSC formation and

oughout spermatogenesis,whileMiwi2 is only detecteds in the 7 days between 15 days post coitus (dpc) and 3et al., 2008; Carmell et al., 2007; Kuramochi-Miyagawa, 2008). This stage-specic regulation of Miwi2, Mili,pression correlates with the order of the developmen-fects in mutants: early spermatocytes arrest in earlyf meiosis I in miwi2 and mili homozygous knockouts,mutant germ cells arrest in the round spermatid stage.ntmice age, arrestedgermcells undergoapoptosis, and,known causes, earlier spermatocytes and spermatogo-gressively become depleted (Carmell et al., 2007; Deng2; Kuramochi-Miyagawa et al., 2004).IWI mutant neonates contain wild-type numbers ofnia, indicating that GSC proliferation is initially unaf-s, at the physiological level, mouse piwi genes doshare the same stem cell maintenance role of D.

r piwi. Also, in contrast to y piwi, mouse piwi genes

AGO pmousemammesis inmammmitted

Howmammdependzebraand feprovidity andalso msex deronmefemalezili, theweeksundergfew getionallPerhapmay atebrate(Wittb

12. M

Althfor PIWPIWI pizationpostertheD.mbeyondanismsInteresresultsthe nual., 200more oestingresult.

In cto tranpiRNA2006;role fodiscussof whinizatioKloc anmutatipiRNAanismthe locbe highport co

Themote tns and endo-siRNAs may play a redundant role withintes (Tam et al., 2008; Watanabe et al., 2008). However,oogenesis is also fundamentally different from oogen-

r vertebrates and spermatogenesis in general, becauseoogonia proliferate only during gestation and are com-eiosis as primary oocytes at birth.piwigenes impact femalegermlinedevelopment innon-vertebrates, where oogenesis, like in D. melanogaster,

on mitotic proliferation and self-renewal of GSCs? Theiwi homologs, ziwi and zili, are expressed in both malegonads, but mutant analyses of these genes has not

clear answer. Null mutations in ziwi and zili cause steril-cell loss during juvenile development, but this event

inizes zebrash (Houwing et al., 2007, 2008). Zebrashination is not driven by sex chromosomes but by envi-and internal cues; however, a hypomorph of zili allowsm cells to develop into sterile oocytes. Compared toi mutant phenotypes are generally more severe: by 3fertilization (wpf) ziwi mutant germ cells universallyoptosis, while zili mutant gonads at 6 wpf still retain aells expressing Ziwi (Houwing et al., 2007, 2008). Addi-zili hypomorph causing female sterility is male fertile.

ure analysis with the genetically tractable medaka shs how PIWI proteins directly affect oogenesis in ver-nce medaka sex determination is genetically speciedet al., 2002).

o piRNAs than transposon silencing?

h transposon control is the primary biological functionteins and piRNAs, evidence for additional functions forns and piRNAs are also emerging, such as mRNA local-road gene expression control. For example, the speciccalization of oskar mRNA and other maternal factors inogaster oocytemight be regulated by the PIWI pathwayply the activation of DNA damage checkpoint mech-ttenhoff and Theurkauf, 2008; Navarro et al., 2009).y, overexpression of piwi in D. melanogaster femalesbryos with greater Oskar expression, which increases

r of pole cells, the y equivalent of PGCs (Megosh ethis suggests that Piwi may directly interact and recruitmRNA to the posterior end of the egg. It would be inter-e if overexpression of aub or ago3 recapitulates this

st to the >60% ofD.melanogaster piRNAs correspondingons (Yin and Lin, 2007), only 1834% of adult vertebrateto repetitive elements (Aravin et al., 2006; Girard et al.,ing et al., 2007; Lau et al., 2006). So, what might be thebulk of class I piRNAs? Translational control has beens a signicant mechanism for germ cell gene regulation,rt of the cellular mechanism depends upon RNA orga-d localization events (Fig. 4; Mendez and Richter, 2001;in, 2005; Kotaja and Sassone-Corsi, 2007). aub and piwiffecting oskar mRNA localization is one example of theway intersecting with a translational regulation mech-k et al., 2004; Harris and Macdonald, 2001). In addition,tion of PIWI/piRNA complexes in the y germline mayynamic, involving processing bodies and dynein trans-nents (Lim et al., 2009; Navarro et al., 2009).laboratory has proposed that Miwi and Mili may pro-lation of male germline transcripts. Their support for


Fig. 4. Mecha elemtargets may omiRNA/AGO ctransport alonor more broad

this hypothpolysomeswith mRNAof translatioGrivna et aMiwi and Morganelle siRNA procesCorsi, 2007assess the ePIWI homowith compo(Lau et al.,capable of tplete meiosPerhaps theattributed tdrive oocytoocyte matThe prevalegerm cellsand the chrprocessing

If PIWI ppiRNAs fromthis functiomRNAs is ubulges likesubstrates lnistic models of gene regulation by animal germline small RNAs. (A) Transposable

perate at the level of mRNA degradation. (B) Many miRNA targets and potential piRNAomplexes or piRNA/PIWI complexes have the potential to direct target RNAs to sequestratg cytoskeletal tracks. (D) siRNAs and piRNAs may have the capacity to inhibit and possibly at the genomic level through regulation of chromosome segregation (E).

esis includes co-sedimentation of Miwi and Mili within density gradients, the association of Miwi and Milis and proteins that bind mRNA caps, and reduced ratesn initiation inmili knockout testes (Deng and Lin, 2002;l., 2006a,b; Unhavaithaya et al., 2008). In spermatids,ili concentrate in the chromatoid body, a perinuclearmilar to nuage and germ plasm and thought to be ansing center (Kotaja et al., 2006; Kotaja and Sassone-). Although more biochemical studies are needed toffect of Miwi and Mili on translation, the X. tropicalislog, Xiwi, localizes to germ plasm and also associatesnents of the translation machinery, including mRNAs2009a). Additionally, in zebrash, a zili hypomorph isransposon silencing, yet the oocytes are unable to com-is normally, resulting in sterility (Houwing et al., 2008).meiotic defect in this zili hypomorph could also be

o a loss of translation regulation of protein factors thate maturation, since translation regulation is integral touration in Xenopus laevis (Mendez and Richter, 2001).nce of translation regulation mechanisms in animaland the commonalities between nuage, germ plasm,omatoid body suggest there may be a conserved RNAand translation regulation role for PIWI proteins.roteins do regulate germ cell mRNAs, then how might

intergenic clusters or TEs mechanistically modulaten? The substrate recognition rules between piRNAs andnknowndo piRNAs recognize targets with imperfectmiRNAs, or do they only act upon perfectly matchedike siRNAs (Fig. 4)? Two recent papers suggest another

possiblemeprimarily th2009; Saitotrafc jammsuggests mdevelopmethat this pibroadly conto select mmode of gethese two nfurther inve

13. Evolut

RNAi isanimals anpossess AGincluding mestingly, fubased on thas animal Pkingdom oand manysequencesRNAs requiand Gorovset al., 2008)mal germ cent silencing by piRNAs and siRNAs and silencing of certain miRNA

and siRNA targets are also regulated at the level of translation. (C)ion in particular cellular compartments, possibly through active RNAly activate transcription at the chromatin level, either at specic loci,

chanismof gene regulationbyPIWIproteins processinge 3 UTRs of select mRNAs into piRNAs (Robine et al.,et al., 2009). Both studies uncovered theD.melanogasterRNAas a precursor of primary piRNAs,which one study

ay regulate downstream genes that enforce follicle cellnt (Saito et al., 2009). The other study demonstratedRNA biogenesis pathway from the 3 UTRs of mRNAs isserved from ies to vertebrates and may have evolvedany specic mRNAs for piRNA processing as a possiblene regulation (Robine et al., 2009). The implications ofew pathways on germ cell development remain to bestigated.

ionary and concluding perspectives

an ancient mode of gene regulation, conserved fromd plants to archaea and protists. Almost all eukaryotesO and Dicer proteins and a population of small RNAs,iRNAs and siRNAs (Ghildiyal and Zamore, 2009). Inter-ngi and plants lack piRNAs since their AGO proteins,eir amino acid sequence, do not t in the same subcladeIWI proteins (Seto et al., 2007). However, the nebulousf protozoans possess homologs to the PIWI subclade,protozoan small RNAs also serve to target repetitivelike transposons, although the genesis of these smallres a Dicer enzyme (Couvillion et al., 2009; Mochizukiky, 2004; Shi et al., 2004, 2006; Ullu et al., 2005; Zhang. Although plant germ cells are quite different from ani-ells, plants also utilize their own repertoire of specic


small RNAs to regulate germline tissue development (Slotkin et al.,2009). Despite this common theme, the miRNA and siRNA path-ways in animals are distinct from the piRNA pathway in termsof the restricted tissue expression, independence from Dicer forbiogenesis,this distincand more renearly all euand some p

In bilateexample, thsequence le2000). Thisin essentiaing interact2009). Neithexhibit theThis may bof endo-siRof miRNA/mto a muchcompared t

Howevethe level ofmosomes bpresence anless, very litrodent andThis suggesidentity of tposons thatand germlinmals, the pimaster con

Arecentpoint of theportion of s(Grimson evectensis anthree PIWItied basedreads that met al., 2008cells of thessponges anderivativesorganisms2005; Mulleplanarians,sess somatirise to planrelate to the

The planhomologs, sneoblasts, athis animalAlthough fupending, thNAs (mainly(Palakodetismedwi-3 inity of planais not affecare blockeddownof smlevels, sugg

differentiation (Palakodeti et al., 2008), in contrasts to mouse EScells, which rely on miRNAs for differentiation (Bernstein et al.,2003). The fact that planaria and basal animals such as spongesand sea anemones containpluripotentneoblasts/neoblast-like cells

er wral fuissueearchay

omsssintent

the emand etpiRNals ietectain mSincest derativs or oondpiR

For emponalit

in sonGSventatin culti

althandhas

RNAinnimay

wled

graic Laions

ns oms sup7298yan L

nces

V. TheV, Leegeno, Gaivel c;442:A, Boan gerA, Hanse inA, Kletural RBiol 2A, La

small;5:33A, Naded Rents iand possibly in nal regulatory mechanisms. Perhapstion arose with the PIWI pathway evolving separatelycently in animal lineages, because siRNAs are found inkaryotes, while PIWI genes are only known in animalsrotozoans (Seto et al., 2007; Grimson et al., 2008).rian animals, miRNAs can be highly conserved. Fore miRNA let-7 is perfectly conserved at the primaryvel between nematodes and humans (Pasquinelli et al.,may be because miRNAs regulate hundreds of targetsl gene regulatory networks via imperfect base pair-ions between the miRNA and target 3 UTRs (Bartel,er individual endo-siRNAs nor individual piRNAs likelydeep primary sequence conservation seen in miRNAs.e attributed to the scattered biogenesis of multitudesNAs and piRNAs from a precursor versus a single pairiRNA* from a miRNA hairpin precursor, and possiblymore limited regulatory role for piRNAs and siRNAso miRNAs.r, conservation of entire piRNA clusters ismaintained atsyntenywithinmammals. Syntenic alignments of chro-etween mouse, rat, and human illustrate the conservedd conguration of pachytene piRNA clusters. Neverthe-tle primary sequence conservation is detected betweenhuman piRNAs (Girard et al., 2006; Lau et al., 2006).ts that piRNA production is essential, but the sequencehe piRNAs can evolve rapidly, possibly to combat trans-also evolvequickly. AlthoughPIWIprotein componentse functions of piRNAs may be conserved among ani-RNA clusters can vary signicantly (i.e. D. melanogastertrol loci versus C. elegans 21U RNA clusters).analysis of twonon-bilateriananimals that areat abasalevolutionary tree has indicated that the major pro-

mall RNAs in basal animals exhibit qualities of piRNAst al., 2008). The genomes of the anemone Nematostellad the sponge Amphimedon queenslandica each encodehomologs, and class II and class I piRNAs could be iden-upon reads that form ping-pong pairs and clusters ofap in a strand-biased fashion, respectively (Grimson

). Whether these piRNAs are restricted to the germe non-bilaterians remains to be established. However,d anemones contain not typical germ cells, but ratherof somatic pluripotent stem cells, which lend to theseextraordinary regenerative capabilities (Extavour et al.,r, 2006). Interestingly, one class of bilaterian animals,also exhibit extensive regenerative capacities and pos-c pluripotent stem cells called neoblasts, which can givearian germ cells. Small RNA studies of planaria mightbiology of non-bilaterians.arian Schmidtea mediterranea also contains three piwimedwi-1, -2, and -3, which are mainly expressed in thend both piRNAs and miRNAs have been cloned from(Palakodeti et al., 2006, 2008; Friedlnder et al., 2009).ll genomic characterization of planarian piRNAs is stillese piRNAs are a bit longer than other animal piR-3032nt), and several piRNAs also target transposons

et al., 2008). RNAi knockdown studies of smedwi-2 anddicate that these genes affect the regenerative capac-ria after amputation, such that neoblast proliferationted, but differentiation and tissue repair by neoblasts(Palakodeti et al., 2008; Reddien et al., 2005). Knock-

edwi-2 and smedwi-3 also reduces piRNAbut notmiRNAesting that planarian neoblasts rely upon piRNAs for

togethancestulate t

Rescells, mand Thof accepluriporise toand hu(Conraof themammbeen din certdata).neoblaregenepiRNA

Beytrol byaging.ical coabnormlationbetweeming echrom

In mthe heagationbiologysmallunderpgained

Ackno

I amson, Erdiscusscitatiowork i(HD05Wing-

Refere

AmbrosAmbros

endoAravin A

A no2006

Aravin Amali

Aravin Adefe

Aravin Aa naCell

Aravin AThe2003

Aravin Astranelemith abundant populations of piRNAs suggests that annction of the piRNA pathway might have been to mod-regeneration in simple animals.into the biology of pluripotent stem cells, like ES

enable a revolution in treating human diseases (Yuon, 2008). Given the ethical and practical limitationsg embryos, other tissue types have been explored forstem cells, includingmale gonads. Since germ cells givembryo andGSCs are totipotent, investigations onmousemale gonads have also yielded pluripotent stem cellsal., 2008; Kanatsu-Shinohara et al., 2004). The impactA pathway on stem cell function in vertebrates ands only now being explored, and while piRNAs have noted inmouse ES cells, class II piRNAs have been detectedouse spermatogonial stem cell lines (Lau, unpublishedthe piRNA pathway plays an integral role in GSC andvelopment, it is encouraging to speculate that othere processes involving stem cells might be impacted byther germline small RNAs.stem cells, further investigation of transposon con-NAs may also impact our understanding of humanxample, maintaining the stability of genomes is a crit-ent of cancer prevention and avoiding developmental

ies. Comparing TE control in germ cells versus TE regu-matic cells might reveal the differences in pluripotencyCs and somatic stemcells. Perhaps epigenetic program-s during gametogenesis via small RNAs help to establishonformations necessary for proper embryogenesis.cellular organisms, germ cells may not be essential toof the individual, but they are essential to the prop-survival of the species. In the recent past, germ cell

become evidently indispensable for investigations intofunction. As future experiments reveal the molecularngs of how PIWI proteins and piRNAs work, the insightslead to applications useful for human therapies.

gments

teful to Michael Blower, Julie Claycomb, Andrew Grim-i, Sivanne Pearl, Christina Post, and Dianne Schwarz forand critical reading of this manuscript. I apologize foritted due to space and publication constraints. This

ported by a K99/R00 award from theNICHDand theNIH). I dedicate thismanuscript to thememoryofmy father,au, whose passion for science lives on in my studies.

functions of animal microRNAs. Nature 2004;431:3505.RC, Lavanway A, Williams PT, Jewell D. MicroRNAs and other tiny

us RNAs in C. elegans. Curr Biol 2003;13:80718.datzis D, Pfeffer S, Lagos-Quintana M, Landgraf P, Iovino N, et al.lass of small RNAs bind to MILI protein in mouse testes. Nature2037.urchis D. Small RNA guides for de novo DNA methylation in mam-m cells. Genes Dev 2008;22:9705.nnon GJ, Brennecke J. The Piwi-piRNA pathway provides an adaptivethe transposon arms race. Science 2007a;318:7614.nov MS, Vagin VV, Bantignies F, Cavalli G, Gvozdev VA. Dissection ofNA silencing process in the Drosophila melanogaster germ line. Mol

004;24:674250.gos-Quintana M, Yalcin A, Zavolan M, Marks D, Snyder B, et al.RNA prole during Drosophila melanogaster development. Dev Cell750.umova NM, Tulin AV, Vagin VV, Rozovsky YM, Gvozdev VA. Double-NA-mediated silencing of genomic tandem repeats and transposablen the D. melanogaster germline. Curr Biol 2001;11:101727.


Aravin AA, SachidanandamR, Bourchis D, Schaefer C, Pezic D, Toth KF, et al. A piRNApathwayprimedby individual transposons is linked todenovoDNAmethylationin mice. Mol Cell 2008;31:78599.

Aravin AA, Sachidanandam R, Girard A, Fejes-Toth K, Hannon GJ. Developmen-tally regulated piRNA clusters implicate MILI in transposon control. Science2007b;316

Armisen J, Gilcically expXenopus t

Bartel DP. M2009;136:

Batista PJ, Ruband 21U-Relegans. M

Bernstein E, Kessential f

Birchler JA, Bhgenes in a

Blumenstiel JPin the repr2005;102:

Boyer LA, PlatcomplexesNature 20

Brennecke J, Acrete smalDrosophila

Brennecke J, Mgenetic ro2008;322:

Brodersen P, SYY, SiebursiRNAs. Sc

Brower-TolandPIWI asso2007;21:2

Carmell MA, GMIWI2 IsMouse Ma

Carthew RW,2009;136:

Chaboissier Melement, tA 1998;95

Chambeyron Sal. piRNA-Drosophila

Chen C, Jin J,interactommethylate

Chen PY, ManmiRNA pr2005;19:1

Chen X. Small2009;25:2

Chen Y, Panetransposon2007;17:6

Chung WJ, Oksomatic de

Cinalli RM, RaClaycomb JM,

Argonautemosome s

Clegg NJ, Froststrom is rpolarity: p71.

Conrad S, Rennof pluripot

CookHA,Koppis required2004;116:

Couvillion MTSequence,Piwi famil

Cox DN, ChaoconservedDev 1998;

Cox DN, Chaoulates the2000;127:

Czech B, Malosmall inte

Das PP, Bagijnal. Piwi an

press Tc3 transposon mobility in the Caenorhabditis elegans germline. Mol Cell2008;31:7990.

Deng W, Lin H. miwi, a murine homolog of piwi, encodes a cytoplasmic proteinessential for spermatogenesis. Dev Cell 2002;2:81930.

Duchaine TF, Wohlschlegel JA, Kennedy S, Bei Y, Conte Jr D, Pang K, et al. Functionaleomic-medSM, Leotider CG,s in ahanismth K,-transt RNAu S, Mtic int;391:ann Ksortedr-1. CeannKuratioded Rder Mlution2009;JA, Den of;16:1l M, Sved fr;320:lM, Za;10:9DE, B

s andees DevAJ, CoRNA

, Sacmall

z G, Bt andJL, Kaof parSI, Monce 20C, Ba

s are r2006;A, Srand e;455:A, Tabans. ST, Beyc cellsT, Pyhractin006bhirayanauteans gerdanecer-mophilanohAs regans. PN, Maent re;128:SD,r H.;435:K, ChoRNAspermf S.;20:5Luensemaleocrino:7447.hrist MJ, Wilczynska A, Standart N, Miska EA. Abundant and dynam-ressed miRNAs, piRNAs, and other small RNAs in the vertebrateropicalis. Genome Res 2009;19:176675.icroRNAs: target recognition and regulatory functions. Cell

21533.y JG, Claycomb JM, Chiang R, Fahlgren N, Kasschau KD, et al. PRG-1NAs interact to form the piRNA complex required for fertility in C.ol Cell 2008;31:6778.im SY, Carmell MA, Murchison EP, Alcorn H, Li MZ, et al. Dicer isor mouse development. Nat Genet 2003;35:2157.adraMP, BhadraU.Makingnoise about silence: repression of repeatednimals. Curr Opin Genet Dev 2000;10:2116., Hartl DL. Evidence for maternally transmitted small interfering RNAession of transposition in Drosophila virilis. Proc Natl Acad Sci U S A1596570.h K, Zeitlinger J, Brambrink T, Medeiros LA, Lee TI, et al. Polycombrepress developmental regulators in murine embryonic stem cells.

06;441:34953.ravin AA, Stark A, Dus M, Kellis M, Sachidanandam R, et al. Dis-l RNA-generating loci as master regulators of transposon activity in. Cell 2007;128:1089103.alone CD, Aravin AA, Sachidanandam R, Stark A, Hannon GJ. An epi-le for maternally inherited piRNAs in transposon silencing. Science138792.akvarelidze-Achard L, Bruun-Rasmussen M, Dunoyer P, Yamamototh L, et al. Widespread translational inhibition by plant miRNAs andience 2008;320:118590.

B, Findley SD, Jiang L, Liu L, Yin H, Dus M, et al. Drosophilaciates with chromatin and interacts directly with HP1a. Genes Dev30011.irard A, van de Kant HJ, Bourchis D, Bestor TH, de Rooij DG, et al.

Essential for Spermatogenesis and Repression of Transposons in thele Germline. Dev Cell 2007.Sontheimer EJ. Origins and Mechanisms of miRNAs and siRNAs. Cell64255.C, Bucheton A, Finnegan DJ. Copy number control of a transposablehe I factor, a LINE-like element in Drosophila. Proc Natl Acad Sci U S:117815., Popkova A, Payen-Groschene G, Brun C, Laouini D, Pelisson A, etmediated nuclear accumulation of retrotransposon transcripts in thefemale germline. Proc Natl Acad Sci U S A 2008;105:149649.James DA, Adams-Cioaba MA, Park JG, Guo Y, et al. Mouse Piwie identies binding mechanism of Tdrkh Tudor domain to arginined Miwi. Proc Natl Acad Sci U S A 2009;106:2033641.ninga H, Slanchev K, Chien M, Russo JJ, Ju J, et al. The developmentaloles of zebrash as determined by small RNA cloning. Genes Dev28893.RNAs and their roles in plant development. Annu Rev Cell Dev Biol144.A, Schupbach T. Cutoff and aubergine mutations result in retro-upregulation and checkpoint activation in Drosophila. Curr Biol

3742.amura K, Martin R, Lai EC. Endogenous RNA interference provides afense against Drosophila transposons. Curr Biol 2008;18:795802.

ngan P, Lehmann R. Germ cells are forever. Cell 2008;132:55962.Batista PJ, Pang KM, Gu W, Vasale JJ, van Wolfswinkel JC, et al. TheCSR-1 and its 22G-RNA cofactors are required for holocentric chro-

egregation. Cell 2009;139:12334.DM, Larkin MK, Subrahmanyan L, Bryant Z, Ruohola-Baker H. mael-

equired for an early step in the establishment of Drosophila oocyteosterior localization of grk mRNA. Development 1997;124:4661

inger M, Hennenlotter J, Wiesner T, Just L, Bonin M, et al. Generationent stem cells from adult human testis. Nature 2008;456:3449.etschBS,Wu J, TheurkaufWE.TheDrosophila SDE3homologarmitagefor oskar mRNA silencing and embryonic axis specication. Cell

81729., Lee SR, Hogstad B, Malone CD, Tonkin LA, Sachidanandam R, et al.biogenesis, and function of diverse small RNA classes bound to theyproteins of Tetrahymena thermophila. GenesDev2009;23:201632.A, Baker J, Chang L, Qiao D, Lin H. A novel class of evolutionarilygenes dened by piwi are essential for stem cell self-renewal. Genes12:371527.A, Lin H. piwi encodes a nucleoplasmic factor whose activity mod-number and division rate of germline stem cells. Development

50314.ne CD, Zhou R, Stark A, Schlingeheyde C, Dus M, et al. An endogenousrfering RNA pathway in Drosophila. Nature 2008;453:798802.

MP, Goldstein LD, Woolford JR, Lehrbach NJ, Sapetschnig A, etd piRNAs act upstream of an endogenous siRNA pathway to sup-

protRNA

Elbashirnucl

Extavouternmec

Fejes-ToPostshor

Fire A, Xgene1998

Forstemaredice

Forstemmatstran

FriedlnresoUSA

Georgezatio2006

Ghildiyaderi2008

Ghildiya2009

GillespienesiGen

GiraldezMicr8.

Girard Aof s202.

Gonzalemen

Goodiertion

GrewalScie

GrimaudnentCell

Grimsongins2008

Grishokeleg

Grivna Sgeni

Grivna SinteS A 2

Gu W, Sargoeleg

GunawaA sliDros

Han T, MsiRNeleg

Harris Apon2001

HateldBake2005

HayashiMicrand

Henikof1998

Hong X,for fEnds reveals the biochemical niche of C. elegans DCR-1 in multiple small-iated pathways. Cell 2006;124:34354.endeckel W, Tuschl T. RNA interference is mediated by 21- and 22-RNAs. Genes Dev 2001;15:188200.

Pang K, Matus DQ, Martindale MQ. vasa and nanos expression pat-sea anemone and the evolution of bilaterian germ cell specications. Evol Dev 2005;7:20115.

Sotirova V, Sachidanandam R, Assaf G, Hannon GJ, Kapranov P, et al.criptional processing generates a diversity of 5-modied long ands. Nature 2009;457:102832.ontgomery MK, Kostas SA, Driver SE, Mello CC. Potent and specicerference by double-stranded RNA in Caenorhabditis elegans. Nature80611., Horwich MD, Wee L, Tomari Y, Zamore PD. Drosophila microRNAsinto functionally distinct argonaute complexes after production byll 2007;130:28797., Tomari Y, Du T, Vagin VV, Denli AM, BratuDP, et al. NormalmicroRNAnandgerm-line stemcellmaintenance requires Loquacious, a double-NA-binding domain protein. PLoS Biol 2005;3:e236.R, Adamidi C, Han T, Lebedeva S, Isenbarger TA, Hirst M, et al. High-discovery and proling of planarian small RNAs. Proc Natl Acad Sci106:1154651.Baryshe PG, Traverse KL, Celniker SE, Pardue ML. Genomic organi-the Drosophila telomere retrotransposable elements. Genome Res23140.eitz H, Horwich MD, Li C, Du T, Lee S, et al. Endogenous siRNAsom transposons and mRNAs in Drosophila somatic cells. Science107781.morePD. Small silencingRNAs: anexpandinguniverse.NatRevGenet4108.erg CA. Homeless is required for RNA localization in Drosophila ooge-ncodes anewmember of theDE-H family of RNA-dependentATPases.1995;9:2495508.

inalli RM, Glasner ME, Enright AJ, Thomson JM, Baskerville S, et al.s regulate brain morphogenesis in zebrash. Science 2005;308:833

hidanandam R, Hannon GJ, Carmell MA. A germline-specic classRNAs binds mammalian Piwi proteins. Nature 2006;442:199

ehringer RR. Dicer is required for female reproductive tract develop-fertility in the mouse. Mol Reprod Dev 2009.zazian Jr HH. Retrotransposons revisited: the restraint and rehabilita-asites. Cell 2008;135:2335.azed D. Heterochromatin and epigenetic control of gene expression.03;301:798802.ntignies F, Pal-Bhadra M, Ghana P, Bhadra U, Cavalli G. RNAi compo-equired for nuclear clustering of Polycomb group response elements.124:95771.ivastava M, Fahey B, Woodcroft BJ, Chiang HR, King N, et al. Early ori-volution of microRNAs and Piwi-interacting RNAs in animals. Nature11937.ara H, Mello CC. Genetic requirements for inheritance of RNAi in C.

cience 2000;287:24947.ret E, Wang Z, Lin H. A novel class of small RNAs in mouse spermato-. Genes Dev 2006a;20:170914.tila B, Lin H. MIWI associates with translational machinery and PIWI-g RNAs (piRNAs) in regulating spermatogenesis. Proc Natl Acad Sci U;103:1341520.ma M, Conte Jr D, Vasale J, Batista PJ, Claycomb JM, et al. Distinct-mediated 22G-RNA pathways direct genome surveillance in the C.rmline. Mol Cell 2009;36:23144.LS, Saito K, Nishida KM, Miyoshi K, Kawamura Y, Nagami T, et al.ediated mechanism for repeat-associated siRNA 5 end formation ina. Science 2007;315:158790.aran AP, Harkins TT, Bouffard P, Fitzpatrick C, Chu DS, et al. 26G endo-ulate spermatogenic and zygotic gene expression in Caenorhabditis

roc Natl Acad Sci U S A 2009;106:186749.cdonald PM. Aubergine encodes a Drosophila polar granule com-quired for pole cell formation and related to eIF2C. Development282332.Shcherbata HR, Fischer KA, Nakahara K, Carthew RW, Ruohola-Stem cell division is regulated by the microRNA pathway. Nature9748.uva de Sousa Lopes SM, Kaneda M, Tang F, Hajkova P, Lao K, et al.biogenesis is required for mouse primordial germ cell developmentatogenesis. PLoS ONE 2008;3:e1738.Conspiracy of silence among repeated transgenes. Bioessays325.e LJ, McGinnis LK, Nothnick WB, Christenson LK. Dicer1 is essentialfertility and normal development of the female reproductive system.logy 2008;149:620712.


Horwich MD, Li C, Matranga C, Vagin V, Farley G, Wang P, et al. The Drosophila RNAmethyltransferase, DmHen1, modies germline piRNAs and single-strandedsiRNAs in RISC. Curr Biol 2007;17:126572.

Houwing S, Berezikov E, Ketting RF. Zili is required for germ cell differentiation andmeiosis in zebrash. EMBO J 2008;27:270211.

Houwing S, KaA Role forin Zebras

Jensen S, Gassadependen

Jin Z, Xie T2007;17:5

Josse T, Teyssetrans-silenrochromat

Kalab P, Heald2008;121:

Kanatsu-Shinoal. Genera2004;119:

Kawamura Y, Ssmall RNA

Kawaoka S, Hbyx ovary-complexes

Kennedy S, Wregulates R

KettingRF,Havfor transpohelicase an

Kidwell MG,MELANOGand male r

Kim JK, Gabelgenomic a

Kim VN, Han J,2009;10:1

Kirino Y, KimNmethylatiostability. N

Kirino Y, Mou3 termini.

Kirino Y, VourmethylatioRNA 2009

Lau IJBCB PiwiReview2010

Documents

Transcript of Lau IJBCB PiwiReview2010