MicroRNAs Genomics, Biogenesis,Mechanism,and Function

Cell,Vol.116,281–297,January23,2004,Copyright2004byCellPress

MicroRNAs:Genomics,

Biogenesis,Mechanism,andFunction

Review

1,2,

*

1

WhiteheadInstituteforBiomedicalResearch

9CambridgeCenter

Cambridge,Massachusetts02142

2

DepartmentofBiology

MassachusettsInstituteofTechnology

Cambridge,Massachusetts02139

MicroRNAs(miRNAs)areendogenousف22ntRNAs

thatcanplayimportantregulatoryrolesinanimalsand

plantsbytargetingmRNAsforcleavageortransla-

ghtheyescapednoticeuntil

relativelyrecently,miRNAscompriseoneofthemore

abundantclassesofgeneregulatorymoleculesin

multicellularorganismsandlikelyinfluencetheoutput

ofmanyprotein-codinggenes.

Inaninvestigationinspiringforbothitsperseverance

anditsscientificinsight,VictorAmbrosandcolleagues,

RosalindLeeandRhondaFeinbaum,discoveredthat

lin-4,s

larvaldevelopment,doesnotcodeforaproteinbut

insteadproducesapairofsmallRNAs(Leeetal.,1993).

OneRNAisapproximately22ntinlength,andtheother

isapproximately61nt;thelongeronewaspredictedto

foldintoastemloopproposedtobetheprecursorof

rosandRuvkunlabsthenno-

ticedthattheselin-4RNAshadantisensecomplemen-

taritytomultiplesitesinthe3ЈUTRofthelin-14gene

(Leeetal.,1993;Wightmanetal.,1993).Thiscomple-

mentarityfellinaregionofthe3ЈUTRpreviouslypro-

posedtomediatetherepressionoflin-14bythelin-4

geneproduct(Wightmanetal.,1991).TheRuvkunlab

wentontodemonstratetheimportanceofthesecom-

plementarysitesforregulationoflin-14bylin-4,showing

alsothatthisregulationsubstantiallyreducesthe

amountofLIN-14proteinwithoutnoticeablechange

er,thesediscoveries

supportedamodelinwhichthelin-4RNAspairtothe

lin-143ЈUTRtospecifytranslationalrepressionofthe

lin-14messageaspartoftheregulatorypathwaythat

triggersthetransitionfromcelldivisionsofthefirstlarval

stagetothoseofthesecond(Leeetal.,1993;Wightman

etal.,1993).

Theshorterlin-4RNAisnowrecognizedasthefound-

ingmemberofanabundantclassoftinyregulatoryRNAs

calledmicroRNAsormiRNAs(Lagos-Quintanaetal.,

2001;Lauetal.,2001;LeeandAmbros,2001).The

breadthandimportanceofmiRNA-directedgeneregula-

tionarecomingintofocusasmoremiRNAsandtheir

-

centlydiscoveredmiRNAfunctionsincludecontrolof

cellproliferation,celldeath,andfatmetabolisminflies

(Brenneckeetal.,2003;Xuetal.,2003),neuronalpat-

terninginnematodes(JohnstonandHobert,2003),mod-

*Correspondence:dbartel@

ulationofhematopoieticlineagedifferentiationinmam-

mals(Chenetal.,2004),andcontrolofleafandflower

developmentinplants(AukermanandSakai,2003;

Chen,2003;Emeryetal.,2003;Palatniketal.,2003).

Computationalapproachesforfindingmessagescon-

trolledbymiRNAsindicatethattheseexamplesrepre-

sentaverysmallfractionofthetotal(Rhoadesetal.,

2002;Enrightetal.,2003;Lewisetal.,2003;Starket

al.,2003).

Thisreviewhighlightswhathasbeenlearnedabout

miRNAsinthedecadesincethereportofthelin-4RNA

ortopicsdiscussed

aremiRNAgenomics,miRNAbiogenesis,miRNAregula-

torymechanisms,andtherolesofmiRNAsingeneregu-

latorypathways.

Genomics:ThemiRNAGenes

Forsevenyearsafterthediscoveryofthelin-4RNA,the

genomicsofthistypeoftinyregulatoryRNAappeared

simple:therewasnoevidenceforlin-4-likeRNAsbe-

yondnematodesandnosignofanysimilarnoncoding

lchangeduponthedis-

coverythatlet-7,shetero-

chronicpathway,encodedasecondف22ntregulatory

-7RNAactstopromotethetransitionfrom

late-larvaltoadultcellfatesinthesamewaythatthe

lin-4RNAactsearlierindevelopmenttopromotethe

progressionfromthefirstlarvalstagetothesecond

(Reinhartetal.,2000;Slacketal.,2000).Furthermore,

homologsofthelet-7geneweresoonidentifiedinthe

humanandflygenomes,andlet-7RNAitselfwasde-

tectedinhuman,Drosophila,andelevenotherbilateral

animals(Pasquinellietal.,2000).

Becauseoftheircommonrolesincontrollingthetim-

ingofdevelopmentaltransitions,thelin-4andlet-7

RNAsweredubbedsmalltemporalRNAs(stRNAs),with

anticipationthatadditionalregulatoryRNAsofthistype

wouldbediscovered(Pasquinellietal.,2000).Indeed,

lessthanoneyearlater,threelabscloningsmallRNAs

fromflies,worms,andhumancellsreportedatotalof

overonehundredadditionalgenesfortinynoncoding

RNAs,approximately20newgenesinDrosophila,ap-

proximately30inhuman,andapproximately60in

worms(Lagos-Quintanaetal.,2001;Lauetal.,2001;

LeeandAmbros,2001).TheRNAproductsofthese

genesresembledthelin-4andlet-7stRNAsinthatthey

wereف22ntendogenouslyexpressedRNAs,potentially

processedfromonearmofastemloopprecursor(Figure

1),andtheyweregenerallyconservedinevolution—

somequitebroadly,othersonlyinmorecloselyrelated

ike

lin-4andlet-7RNAs,manyofthenewlyidentifiedف22

ntRNAswerenotexpressedindistinctstagesofdevel-

opmentandinsteadweremorelikelytobeexpressed

etermmicroRNAwas

usedtorefertothestRNAsandalltheothertinyRNAs

withsimilarfeaturesbutunknownfunctions(Lagos-

Quintanaetal.,2001;Lauetal.,2001;LeeandAmbros,

2001).Intensifiedcloningeffortshaverevealednumer-

Cell

282

esofMetazoanmiRNAs

Shownarepredictedstemloopsinvolvingthe

maturemiRNAs(red)andflankingsequence.

ThemiRNAs*(blue)arealsoshownincases

wheretheyhavebeenexperimentallyidenti-

fied(Limetal.,2003a).

(A)Predictedstemloopsofthefounding

miRNAs,lin-4andlet-7RNAs(Leeetal.,1993;

Reinhartetal.,2000).Theprecisesequences

ofthematuremiRNAsweredefinedbyclon-

ing(Lauetal.,2001).-

gansstemloops,butclosehomologsofboth

havebeenfoundinfliesandmammals(Pas-

quinellietal.,2000;Lagos-Quintanaetal.,

2001,2002).

(B)ExamplesofmiRNAsfromothermetazoan

genes,mir-1,mir-34,re

sstemloops,butclosehomo-

logsofthesemiRNAshavebeenfoundinflies

andmammals(Lagos-Quintanaetal.,2001,

2002;Lauetal.,2001;LeeandAmbros,2001).

(C)ExamplesofmiRNAsfromplantgenes,

MIR165a,MIR172a2,re

Arabidopsisstemloops,butclosehomologs

ofthesemiRNAshavebeenfoundinriceand

otherplants(Parketal.,2002;Reinhartetal.,

2002;Palatniketal.,2003).

ousadditionalmiRNAgenesinmammals,fish,worms,

andflies(Lagos-Quintanaetal.,2002,2003;Mourelatos

etal.,2002;Ambrosetal.,2003b;Aravinetal.,2003;

Dostieetal.,2003;Houbaviyetal.,2003;Kimetal.,

2003;Limetal.,2003a,2003b;Michaeletal.,2003).A

registryhasbeensetuptocatalogthemiRNAsand

facilitatethenamingofnewlyidentifiedgenes(Griffiths-

Jones,2004).

slin-4andlet-7,mostmiRNAgenes

comefromregionsofthegenomequitedistantfrom

previouslyannotatedgenes,implyingthattheyderive

fromindependenttranscriptionunits(Lagos-Quintana

etal.,2001;Lauetal.,2001;LeeandAmbros,2001).

Nonetheless,,aboutaquarterof

thehumanmiRNAgenes)areintheintronsofpre-

repreferentiallyinthesameorientation

asthepredictedmRNAs,suggestingthatmostofthese

miRNAsarenottranscribedfromtheirownpromoters

butareinsteadprocessedfromtheintrons,asseenalso

formanysnoRNAs(Aravinetal.,2003;Lagos-Quintana

etal.,2003;Laietal.,2003;Limetal.,2003a).This

arrangementprovidesaconvenientmechanismforthe

-

latoryscenariosareeasytoimagineinwhichsuchcoor-

dinateexpressioncouldbeuseful,whichwouldexplain

theconservedrelationshipsbetweenmiRNAsandhost

ingexampleofthisconservationinvolves

mir-7,foundintheintronofhnRNPKinbothinsects

andmammals(Aravinetal.,2003).

OthermiRNAgenesareclusteredinthegenomewith

anarrangementandexpressionpatternimplyingtran-

scriptionasamulti-cistronicprimarytranscript(Lagos-

Quintanaetal.,2001;Lauetal.,2001).Althoughthe

majorityofwormandhumanmiRNAgenesareisolated

andnotclustered(Limetal.,2003a,2003b),overhalf

oftheknownDrosophilamiRNAsareclustered(Aravin

etal.,2003).ThemiRNAswithinagenomicclusterare

often,thoughnotalways,relatedtoeachother;and

relatedmiRNAsaresometimesbutnotalwaysclustered

(Lagos-Quintanaetal.,2001;Lauetal.,2001).Orthologs

slin-4andlet-7areclusteredintheflyand

humangenomesandarecoexpressed,sometimesfrom

thesameprimarytranscript,leadingtotheideathatthe

genomicseparationoflin-4fromlet-7innematodes

mightbeuniquetothewormlineage(Aravinetal.,2003;

Bashirullahetal.,2003;Sempereetal.,2003).Thisexam-

Review

283

pleillustratesthepossibilitythatevenincaseswhere

clusteredgeneshavenoapparentsequencehomology,

theymaysharefunctionalrelationships.

Someofthemoreinterestinggenomiclocationsof

mir-10geneliesintheAntennapediacomplexofinsects

andintheorthologouslocationsintwoHoxclustersof

mammals,whereasthemir-iab-4geneiswithinthein-

sectBithoraxcluster(Aravinetal.,2003;Lagos-

Quintanaetal.,2003).Inlightoftherolesofothergenes

oftheHoxclusters,theHoxmiRNAsareespeciallygood

candidatesforhavinginterestingfunctionsinanimal

nterestinglociincludethemir-15a-

mir-16cluster,whichfallswithinaregionofhumanchro-

mosome13thoughttoharboratumorsuppressorgene

becauseitisthesiteofthemostcommonstructural

aberrationsinbothmantlecelllymphomaandBcell

chroniclymphocyticleukemia(Lagos-Quintanaetal.,

2001;Calinetal.,2002).

NearlyalloftheclonedmiRNAsareconservedin

closelyrelatedanimals,suchashumanandmouse,or

ae(Lagos-Quintanaetal.,2003;

Limetal.,2003a,2003b).Thisstatementremainstrue

evenwhenignoringevolutionaryconservationasacrite-

ealso

conservedmorebroadlyamongtheanimallineages

(Ambrosetal.,2003b;Aravinetal.,2003;Lagos-

Quintanaetal.,2003;Limetal.,2003a).Forinstance,

smiRNAshaveeasily

recognizedhomologsamongthehumanmiRNAs(Lim

etal.,2003a).Whencomparingdistantlineages,consid-

erableexpansionorcontractionofgenefamiliesisap-

parent,themoststrikingexamplebeingthelet-7family,

sandat

least15inhuman,butonlyoneinDrosophila(Pasquinelli

etal.,2000;Aravinetal.,2003;Laietal.,2003;Limet

al.,2003a).

Genomics:miRNAExpression

example,slin-4

andlet-7RNAshavestage-specificexpressionindevel-

opmentasifthey,too,functionasstRNAs(Pasquinelli

etal.,2000;Lauetal.,2001;Lagos-Quintanaetal.,2002;

Bashirullahetal.,2003;Limetal.,2003a).Otherinterest-

ingexamplesincludemiR-1,whichisprimarilyfoundin

themammalianheart(LeeandAmbros,2001;Lagos-

Quintanaetal.,2002);miR-122,whichisprimarilyinthe

liver(Lagos-Quintanaetal.,2002);miR-223,whichis

primarilyinthegranulocytesandmacrophagesof

mousebonemarrow(Chenetal.,2004);miRNAsofthe

mir-35–mir-42cluster,whicharepreferentiallyintheC.

elegansembryo(Lauetal.,2001);andthoseofthemir-

290–mir-295cluster,whichareexpressedinmouseem-

bryonicstemcellsbutnotindifferentiatedcells(Hou-

baviyetal.,2003).Expressionarraytechnologyhasbeen

adaptedtoexaminemiRNAsandhasrevealeddistinct

expressionpatternsindifferentdevelopmentalstages

orregionsofthemammalianbrain(Krichevskyetal.,

2003).Withallthedifferentgenesandexpressionpat-

terns,itisreasonabletoproposethateverymetazoan

celltypeateachdevelopmentalstagemighthavea

distinctmiRNAexpressionprofile—providingampleop-

portunityfor“micromanaging”theoutputofthetran-

scriptome.

AnotherremarkableaspectofmiRNAexpressionis

example,miR-2,miR-52,andmiR-58areeachpresent

onaverageatmorethan50,000moleculesperadult

wormcell—agreaterabundancethantheU6snRNAof

thespliceosome(Limetal.,2003a).Whetherthishigh

expressionisattributabletoveryrobusttranscription

RNAsare

tance,miR-124

ispresentintheadultwormonaverageat800molecules

percell(Limetal.,2003a).Thisloweraveragelevel

(thoughstillhigherthanthatofthetypicalmRNA)might

beduetolowexpressioninmanycellsorhighexpres-

dingthatthemouseor-

thologofmiR-124isnearlyexclusivelyexpressedinthe

brainsupportsthelatterexplanation(Lagos-Quintana

etal.,2002).

Genomics:ComputationalApproaches

andGeneNumber

TherehasbeensomespeculationastowhymiRNAs

werenotdiscoveredearlier;theanswerisclearlynotthat

NAsandtheirassociatedproteins

appeartobeoneofthemoreabundantribonucleopro-

eless,miRNAswhose

expressionisrestrictedtononabundantcelltypesor

specificenvironmentalconditionscouldstillbemissed

,computationalapproacheshave

beendevelopedtocomplementexperimentalap-

rlyon,

homologysearcheshaverevealedorthologsandpara-

logsofknownmiRNAgenes(Pasquinellietal.,2000;

Lagos-Quintanaetal.,2001;Lauetal.,2001;Leeand

Ambros,2001).Anothersimpleapproachhasbeento

searchthevicinityofknownmiRNAgenesforother

stemloopsthatmightrepresentadditionalgenesofa

genomiccluster(Lauetal.,2001;Aravinetal.,2003;

Seitzetal.,2003;Ohleretal.,2004).Thisstrategyis

importantbecausesomeofthemostrapidlyevolving

miRNAgenesarepresentastandemarrayswithinop-

eron-likeclusters,andthedivergentsequencesofthese

genesmakethemrelativelydifficulttospotusingthe

moregeneralapproaches.

Gene-findingapproachesthatdonotdependonho-

mologyorproximitytoknowngeneshavealsobeen

developedandappliedtoentiregenomes(Ambrosetal.,

2003b;Gradetal.,2003;Laietal.,2003;Limetal.,

2003a).Theytypicallystartbyidentifyingconservedge-

nomicsegmentsthatbothfalloutsideofpredictedpro-

tein-codingregionsandpotentiallycouldformstem

loopsandthenscorethesecandidatemiRNAstemloops

forthepatternsofconservationandpairingthatcharac-

,thetwomostsensi-

tivecomputationalscoringtoolsareMiRscan,whichhas

beensystematicallyappliedtonematodeandvertebrate

candidates(Limetal.,2003a,2003b),andmiRseeker,

whichhasbeensystematicallyappliedtoinsectcandi-

dates(Laietal.,2003).BothMiRscanandmiRseeker

haveidentifieddozensofgenesthatweresubsequently

(orconcurrently)eof

theirrelativelyhighsensitivity,MiRscanandmiRseeker

Cell

284

havealsoenabledreasonablyfirmestimatesofthenum-

berofmiRNAgenesinthegenomesofhuman(200–255

miRNAgenes;Limetal.,2003b),s(103–120

genes;Limetal.,2003a;Ohleretal.,2004),andDrosoph-

ila(96–124genes;Laietal.,2003).Ineachspecies,these

numbersrepresentnearly1%ofthepredictedgenesin

thegenome,afractionsimilartothatofotherlargegene

familieswithregulatoryroles,suchasthehomeodomain

transcription-factorfamily.

TheseestimatesimplythatthemajorityofmiRNA

geneshavenowbeenfoundinthemammalianandnem-

atodelineages—s,whereap-

proximately100miRNAgeneshavebeenidentified.

(Thistallyisconservativeinthatitexcludessomere-

portedgenesthatappeartobequestionable[Ohleret

al.,2004].)InDrosophila,77genes,representing71

uniquemiRNAs,havebeenreliablyidentified(Aravinet

al.,2003;Laietal.,2003),andinhumans,approximately

175genes,representingapproximately145unique

miRNAs,haveeitherbeenvalidatedinhumancellsor

identifiedbasedontheirhomologytogenesvalidated

inmouseorzebrafish(miRNARegistry,release3.0;Grif-

fiths-Jones,2004).Whenconsideringthenumberof

miRNAsremainingtobeidentifiedorvalidatedinthese

species,itisimportanttorememberthatgenenumber

estimatesbyMiRscanandmiRseekerrestontheas-

sumptionthatthestemloopsoftherare,difficult-to-

clonemiRNAswillshowpatternsofconservationand

pairingresemblingthoseoftheabundant,easilycloned

-

gans,forwhichtherewasareassuringlackofcorrelation

betweenthenumberoftimesanmiRNAwasclonedand

itsMiRscanscore(Limetal.,2003a).

Ifinsteadadisproportionatenumberofdifficult-to-

clonemiRNAsarealsodifficulttoidentifycomputation-

ally,thenestimatesofthenumberofmiRNAgenesin

ghtbethesituation

inhumans—perhapsbecausethevertebrategenomes

thefirst109miRNAsclonedfrommammalshavereadily

identifiablehomologsinthegenomeofpufferfish(Fugu

ripens),whichenabledMiRscananalysistoidentify81

(74%)ofthesegenesbyscoringstemloopsconserved

inhuman,mouse,andfish(Limetal.,2003b).Extrapolat-

ingfromthissensitivityandthenumberofadditional

candidateswithscoresmatchingtheknownmiRNAs,

anupperboundonthenumberofhumanmiRNAgenes

wascalculatedtobe255(Limetal.,2003b).However,

morerecentlyidentifiedmammalianmiRNAgenesap-

pearrelativelylesslikelytobeconservedinfish,particu-

larlythosegenesclonedfromembryonicstemcellsand

mammalianbrainandthe14miRNAcandidatesresiding

inalargeimprintedcluster(Houbaviyetal.,2003;Kim

etal.,2003;Seitzetal.,2003).Theserecentdatasuggest

thatthemoredifficult-to-clonemammalianmiRNAsare

lesslikelytobeconservedinfishandthuslesslikelyto

havebeenidentifiedcomputationally,whichimpliesthat

aconfidentupperboundonthenumberofhumangenes

isdifficulttodetermineusinganalysesthatextended

tofishandthat255istoolowavalueforthisupper

bound—althoughitstillmightexceedtheactualnumber

ofhumanmiRNAgenes.

Genomics:miRNAsinPlants

CloningofsmallRNAsfromplantshasalsorevealed

miRNAs,althoughthemultitudeofother21to24nt

RNAsfoundinplantssometimescomplicatedtheirinitial

classification(Llaveetal.,2002a;Metteetal.,2002;Park

etal.,2002;Reinhartetal.,2002).Likethemetazoan

miRNAs,theplantmiRNAs(1)areendogenouslyex-

pressedف22ntRNAspotentiallyprocessedfromone

armoffoldbackprecursors,(2)aregenerallyconserved

inevolution,and(3)comefromregionsofthegenome

distinctfrompreviouslyannotatedgenes(Reinhartet

al.,2002).Todate,20uniqueArabidopsismiRNAshave

beenreported;afewarecloselyrelatedtoeachother,

andthusthereportedgenesrepresent15distinctmiRNA

families.(BartelandBartel,2003;Palatniketal.,2003).

Becausesomecouldbederivedfrommultiplegenomic

loci,the20miRNAscouldrepresentmorethan40Arabi-

ologysearchesbasedonthe

clonedgenesalsorevealnumerouspotentialparalogs

withapointsubstitutionortwointhepredictedmiRNA.

Additionalgenefamiliesarelikelytobefoundwhenthe

cloningofsmallplantRNAsisscaledupandcomputa-

appearsthat,asinanimals,asubstantialfractionofthe

generegulatorymoleculesinplantscouldbeRNArather

thanprotein.

ThediscoveryofmiRNAsinbothplantsandanimals

suggeststhatthisclassofnoncodingRNAshasbeen

modulatinggeneexpressionsinceatleastthelastcom-

monancestoroftheselineages(Reinhartetal.,2002).

Nonetheless,plantandanimalmiRNAsdifferinsome

aspects,whichappeartoberelatedtodifferencesin

tnotabledifferencesarein

themiRNAstemloops;theplantpredictedfoldbacks

aremuchmorevariableinsizeandtypicallylargerthan

thoseofanimals(Figure1;foramorecomprehensive

lookatplantmiRNApredictedstemloops,seeonline

supplementalmaterialofReinhartetal.,2002).More

subtledifferencesincludesomewhatmorepairingbe-

tweenthemiRNAandtheotherarmofthestemloopin

plantscomparedtoanimals,atighterdistributionof

plantmiRNAlengthsthatcenterson21ntratherthan

the22–23ntlengthsmostoftenseeninanimals,and

perhapsastrongerpreferenceforaUatthe5Јterminus

oftheplantmiRNAs(Lauetal.,2001;Reinhartetal.,

2002;BartelandBartel,2003).Thesedifferences,to-

getherwiththeabsenceofreportsthatparticularmiRNA

genesareconservedbetweenplantsandanimals,leave

opentheprospectthatmiRNAgenesaroseindepen-

dentlyineachofthesemulticellularlineages,aftertheir

lastcommonancestor(whichisthoughttohavebeen

unicellular).Eveninthisscenarioofdualorigins,the

presenceofmiRNAsinallplantandanimalspecies

examinedthusfarsuggestsearlyoriginsinbothlin-

eages,perhapsprecedingandfacilitatingthedevelop-

mentalpatterningneededformulticellularbodyplans.

Biogenesis:miRNATranscription

A693bpgenomicfragmentrescuesthelin-4deficiency,

implyingthatalltheelementsrequiredfortheregulation

andinitiationoftranscriptionarelocatedinthisshort

fragment(Leeetal.,1993).However,littleisknownre-

gardingthesetranscriptionalprocessesforlin-4orany

Review

285

RNAsresidinginintrons

arelikelytosharetheirregulatoryelementsandprimary

re-

mainingmiRNAgenes,presumablytranscribedfrom

theirownpromoters,noprimarytranscriptshavebeen

eless,theseprimarymiRNAtran-

scripts,calledpri-miRNAs(Leeetal.,2002),aregener-

allythoughttobemuchlongerthantheconservedstem

loopscurrentlyusedtodefinemiRNAgenes,assug-

gestedbythefollowing:(1)theideathatclustered

miRNAstemloopsaretranscribedfromasingleprimary

transcript(Lagos-Quintanaetal.,2001;Lauetal.,2001),

(2)matchesbetweenmiRNAsandlengthyESTsinthe

databases(Lagos-Quintanaetal.,2002;Aukermanand

Sakai,2003),(3)RT-PCRexperimentsamplifyinglarge

fragmentsofthepri-miRNAs(Leeetal.,2002;Aravinet

al.,2003).

ThetwocandidateRNApolymerasesforpri-miRNA

roducesthe

mRNAsandsomenoncodingRNAs,includingthesmall

nucleolarRNAs(snoRNAs)andfourofthesmallnuclear

RNAs(snRNAs)ofthespliceosome,whereaspolIIIpro-

ducessomeoftheshorternoncodingRNAs,including

tRNAs,5SribosomalRNA,

miRNAsprocessedfromtheintronsofprotein-coding

followingobservationsprovideindirectevidencethat

manyoftheothermiRNAsalsoarepolIIproducts,even

thoughmostofthemetazoanmiRNAgenesdonothave

theclassicalsignalsforpolyadenylylation(Ohleretal.,

2004):(1)Thepri-miRNAscanbequitelong,morethan

one1kb,whichislongerthantypicalpolIIItranscripts.

(2)Thesepresumedpri-miRNAsoftenhaveinternalruns

ofuridineresidues,whichwouldbeexpectedtoprema-

turelyterminatepolIIItranscription.(3)ManymiRNAs

aredifferentiallyexpressedduringdevelopment,asis

observedoftenforpolIIbutnotpolIIIproducts.(4)

Fusionsthatplacetheopenreadingframeofareporter

proteindownstreamfromthe5ЈportionofmiRNAgenes

leadtorobustreporterproteinexpression,suggesting

thatmiRNAprimarytranscriptsarecappedpolIItran-

esofsuchfusionsincludeartificialre-

porterconstructsdesignedtoinvestigatetheregulation

ofmiRNAexpression(Johnsonetal.,2003;Johnston

andHobert,2003)andanaturalchromosometransloca-

tionlinkedtoanaggressiveBcellleukemia,inwhicha

truncatedMYCgeneisfusedtothe5Јportionofmir-

142(Gauwerkyetal.,1989;Lagos-Quintanaetal.,2002).

AlthoughtheseobservationsindicatethatmanymiRNAs

arepolIItranscripts,othersmightstillbepolIIItran-

scripts,justasmostbutnotallsnRNAsarepolIIprod-

cexpressionofmiR-142andothermiRNAs

fromapolIIIpromoterproducesefficientlyandprecisely

processedmiRNAsthatfunctioninvivo(Chenetal.,

2004),indicatingthatthereisnoobligatelinkbetween

theidentityofthepolymeraseanddownstreammiRNA

processingorfunction.

Biogenesis:miRNAMaturation

Thecurrentmodelformaturationofthemammalian

ststepisthe

nuclearcleavageofthepri-miRNA,whichliberatesa

ف60–70ntstemloopintermediate,knownasthemiRNA

precursor,orthepre-miRNA(Leeetal.,2002;Zengand

Cullen,2003).Thisprocessingisperformedbythe

DroshaRNaseIIIendonuclease,whichcleavesboth

strandsofthestematsitesnearthebaseoftheprimary

stemloop(Leeetal.,2003)(Figure2B,step2).Drosha

cleavestheRNAduplexwithastaggeredcuttypicalof

RNaseIIIendonucleases,andthusthebaseofthepre-

miRNAstemloophasa5Јphosphateandف2nt3Ј

overhang(Basyuketal.,2003;Leeetal.,2003).This

pre-miRNAisactivelytransportedfromthenucleusto

thecytoplasmbyRan-GTPandtheexportreceptorEx-

portin-5(Yietal.,2003;Lundetal.,2004)(Figure2B,

step3).

ThenuclearcutbyDroshadefinesoneendofthe

erendisprocessedinthecyto-

plasmbytheenzymeDicer(Leeetal.,2003).Dicer,also

anRNaseIIIendonuclease,wasfirstrecognizedforits

roleingeneratingthesmallinterferingRNAs(siRNAs)

thatmediateRNAinterference(RNAi)(Bernsteinetal.,

2001)andwaslatershowntoplayaroleinmiRNAmatu-

ration(Grishoketal.,2001;Hutva

´

gneretal.,2001;Ket-

tingetal.,2001).Accordingtothecurrentmodelof

miRNAmaturation,Dicerperformsanactivityinmeta-

zoanmiRNAmaturationsimilartothatwhichitperforms

whenchoppingupdouble-strandedRNAduringRNAi:

Itfirstrecognizesthedouble-strandedportionofthe

pre-miRNA,perhapswithparticularaffinityfora5Јphos-

phateand3Јoverhangatthebaseofthestemloop.

Then,atabouttwohelicalturnsawayfromthebaseof

thestemloop,

cleavagebyDicerlopsofftheterminalbasepairsand

loopofthepre-miRNA,leavingthe5Јphosphateandف2

nt3ЈoverhangcharacteristicofanRNaseIIIandproduc-

ingansiRNA-likeimperfectduplexthatcomprisesthe

maturemiRNAandsimilar-sizedfragmentderivedfrom

theopposingarmofthepre-miRNA(Figure2B,step4).

Thefragmentsfromtheopposingarm,calledthe

miRNA*sequences(Lauetal.,2001),arefoundinlibrar-

iesofclonedmiRNAsbuttypicallyatmuchlowerfre-

quencythanarethemiRNAs(Lagos-Quintanaetal.,

2002;Aravinetal.,2003;Limetal.,2003a).Forexample,

inaneffortthatidentifiedover3400clonesrepresenting

smiRNAs,only38clonesrepresenting14

miRNAs*werefound(Limetal.,2003a).Thisapproxi-

mately100-folddifferenceincloningfrequencyindi-

catesthatthemiRNA:miRNA*duplexisgenerallyshort-

livedcomparedtothemiRNAsinglestrand.

Accordingtothecurrentmodel,thespecificityofthe

initialcleavagemediatedbyDroshadeterminesthecor-

rectregisterofcleavagewithinthemiRNAprecursor

andthusdefinesbothmatureendsofthemiRNA(Lee

etal.,2003).ThisideathatDrosha,notDicer,imparts

thespecificityisappealingbecausestudieshaveshown

thatgenericdouble-strandedRNAisrefractoryto

DroshacleavageandthatDicerprogressivelychops

upanRNAdoublestrand,irrespectiveofitssequence

(Zamoreetal.,2000;Bernsteinetal.,2001;Elbashiretal.,

2001a;Zhangetal.,2002).ThedeterminantsofDrosha

recognitionarelargelyundefinedbutincludethesec-

ondarystructureatthebaseoftheprimarystemloop

aswellassomeelementsflankingthestemloopbut

generallywithin125ntofthemiRNA(Leeetal.,2003;

Chenetal.,2004).

ThisstepwisescenarioformiRNAmaturationisbased

Cell

286

genesisofmiRNAsandsiRNAs

(A)ThebiogenesisofaplantmiRNA(steps1–6;seetextfordetails)anditshetero-silencingoflociunrelatedtothatfromwhichitoriginated

(step7).Thepre-miRNAintermediates(bracketed),thoughttobeveryshort-lived,NA(red)is

incorporatedintotheRISC(step6),whereasthemiRNA*(blue)isdegraded(hatchedsegment).Amonophosphate(P)marksthe5Јterminus

ofeachfragment.

(B)ThebiogenesisofametazoanmiRNA(steps1–6;seetextfordetails)anditshetero-silencingoflociunrelatedtothatfromwhichit

originated(step7).

(C)ThebiogenesisofanimalsiRNAs(steps1–6;seetextfordetails)andtheirauto-silencingofthesame(orsimilar)locifromwhichthey

originated(step7).

primarilyontheinvestigationofmammalianDroshaand

Dicerfunction(Leeetal.,2002,2003).Thenotionthat

itappliestoothermetazoanspeciesissupportedby

slin-4RNA,

whichappearstobeanexcellentmatch(withinthereso-

lutionofnucleasemapping)tothatexpectedforthe

lin-4pre-miRNA(Leeetal.,1993).Furthermore,pre-

sumedpre-miRNAsfornumerousmiRNAscanbede-

tectedonNorthernblots,andwhenexaminedinthe

contextofreducedDiceractivity,thesepre-miRNAs

invariablyincreaseinabundance,aswouldbeexpected

ifDicerwasresponsiblefortheirprocessing(Grishoket

al.,2001;Hutva

´

gneretal.,2001;Kettingetal.,2001;Lee

andAmbros,2001;Limetal.,2003a).Finally,thegeneral

existenceofthemiRNA:miRNA*duplexissupportedby

thecloningofnumerousmiRNAs*innematodesand

flies,althoughformostmiRNAgenes,anexperimentally

identifiedmiRNA*hasnotyetbeenreported.

ThecloningofafewmiRNAs*inplantsalsopointsto

atransientmiRNA:miRNA*duplex(Reinhartetal.,2002).

However,thebiogenesisofthisduplexappearstodiffer

inplants(Figure2A).Mostnotably,pre-miRNAshave

notbeencompellinglydetectedinplants—notevenin

plantswithcrippledDCL1,aDicer-likeproteinknown

toassistinmiRNAmaturation(Reinhartetal.,2002).

Thelackofpre-miRNAinthesedcl1-9plants(formerly

knownascaf-1plants),togetherwiththeapparentnu-

clearlocalizationoftheDCL1protein(Pappetal.,2003),

suggeststhatDCL1providestheDroshafunctionality

inplants,makingthefirstcutthatsetstheregisterfor

miRNAmaturation(Figure2A,step2).DCL1(oranother

enzymeyettobeidentified)thenmakesthesecondcut,

whichcorrespondstometazoanDicercleavage,before

themiRNAleavesthenucleus(Figure2A,step3).A

coupledsecondcutinthenucleuswouldexplainwhy

pre-miRNA-likeRNAsdonotaccumulatetodetectable

dalsoexplainwhyectopicnuclear

butnotcytoplasmicexpressionofP19,aplantviral

proteinthatinhibitssilencingbysequesteringsiRNA

duplexes,preventsmiRNAaccumulation(Pappetal.,

Review

287

2003).PerhapsHASTY,theplantorthologofExportin-5,

isresponsibleforexportingthemiRNA:miRNA*duplex

fromthenucleus,whichwouldexplainthepleiotropic

developmentalphenotypesofhastymutants(Bollman

etal.,2003;Yietal.,2003;Lundetal.,2004)(Figure2A,

step4).

Biogenesis:RISCAssembly

Followingcleavageandnucleocytoplasmicexport,the

miRNApathwayofplantsandanimalsappearstobe

biochemicallyindistinguishablefromthecentralsteps

ofRNAsilencingpathwaysknownasposttranscriptional

genesilencing(PTGS)inplants,quellinginfungi,and

,understandingmiRNAbiogene-

sisandfunctionhasbeengreatlyfacilitatedbyanalogy

andcontrasttothesiRNAsofRNAi,

lightofthesebiochemicalconnections,thediscovery

oflin-4anditsregulationoflin-14canbeconsideredin

hindsightasthefirstcharacterizationofanRNAi-like

phenomenoninanimals.

ToillustratethecommonalitybetweenmiRNAsand

siRNAs,theRNAipathwayisbrieflyoutlinedhere(and

depictedinFigure2C).Thepathwaybeginswithlong

double-strandedRNA,eitherabimolecularduplexoran

extendedhairpin,thateitherisartificiallyintroducedinto

thecelloranimalduringageneknockdownexperiment

(Fireetal.,1998)orisnaturallygenerated—fromsense

andantisensegenomictranscripts,orperhapsfromthe

activityofacellularRNA-dependentRNApolymerase

(foundinplants,fungi,andnematodes,butnotflies

ormammals)orasanintermediateofviralreplication

(CogoniandMacino,1999;Kettingetal.,1999;Dalmay

etal.,2000;Mourrainetal.,2000;Smardonetal.,2000;

Aravinetal.,2001,2003;Lietal.,2002).Thedouble-

strandedRNAisprocessedbyDicerintomanyف22nt

siRNAs(HamiltonandBaulcombe,1999;Hammondet

al.,2000;Parrishetal.,2000;Zamoreetal.,2000;Grishok

etal.,2001;Kettingetal.,2001;KnightandBass,2001)

(Figure2C,steps2–4).AlthoughthesesiRNAsareini-

tiallyshortdouble-strandedspecieswith5Јphosphates

and2nt3ЈoverhangscharacteristicofRNaseIIIcleav-

ageproducts,theyeventuallybecomeincorporatedas

single-strandedRNAsintoaribonucleoproteincomplex,

knownastheRNA-inducedsilencingcomplex(RISC)

(Hammondetal.,2000;Elbashiretal.,2001a,2001b;

Nyka

¨

nenetal.,2001;Martinezetal.,2002;Schwarzet

al.,2002)(Figure2C,step6).TheRISCidentifiestarget

messagesbasedonperfect(ornearlyperfect)comple-

mentaritybetweenthesiRNAandthemRNA,andthen

theendonucleaseoftheRISCcleavesthemRNAat

asitenearthemiddleofthesiRNAcomplementarity,

measuringfromthe5ЈendofthesiRNAandcutting

betweenthenucleotidespairingtoresidues10and11

ofthesiRNA(Elbashiretal.,2001a,2001b).Similarpath-

wayshavebeenproposedforgenesilencinginplants

andfungi(HamiltonandBaulcombe,1999;Vanceand

Vaucheret,2001;Pickfordetal.,2002).

TheRISChasbeenpurifiedfromflyandhumancells

andinbothcasescontainsamemberoftheArgonaute

proteinfamily,whichisthoughttobeacorecomponent

ofthecomplex(Hammondetal.,2001;Hutva

´

gnerand

Zamore,2002;Martinezetal.,2002).Thisfitsnicelywith

previousgeneticdatashowingthatArgonauteproteins

RDE-1,QDE2,andAGO1arecrucialforRNAiandanalo-

gousprocessesinworms,fungi,andplants,respectively

(Tabaraetal.,1999;Catalanottoetal.,2000;Fagardet

al.,2000).Argonauteanditshomologsareapproxi-

mately100kDaproteinsthataresometimescalledPPD

proteinsbecausetheyallsharethePAZandPIWIdo-

mains(Ceruttietal.,2000).ThePAZdomain(firstrecog-

nizedinPiwi,Argonaute,andZwille/Pinheadproteins)

hasastablefoldwhenisolatedfromtherestofthe

protein,whichhasa␤barrelcorethattogetherwith

asideappendageappearstobindweaklytosingle-

strandedRNAsatleast5ntinlengthandalsotodouble-

strandedRNA(Lingeletal.,2003;Songetal.,2003;Yan

etal.,2003).Thisdualbindingabilitysuggeststhatthe

Argonauteproteincouldbedirectlyassociatedwiththe

siRNAbeforeandafteritrecognizesthemRNAtarget.

OtherRISC-associatedproteinsincludethesus-

pectedRNAbindingproteinsVIGandFragileX-related

proteinandthenucleaseTudor-SN,noneofwhichhave

definedrolesintheRISC(Caudyetal.,2002,2003;

Ishizukaetal.,2002).Theseproteinsdonotcopurify

withRISCinallpurificationschemesandtheirstoichi-

sthey

arealsocorecomponentsoftheRISCthatdonotremain

a-

tively,theycouldbeaccessoryfactorsthatmodifythe

ion

thatRISCcomesindifferentsubtypesisalreadysup-

portedbythenumberofArgonautefamilymembers

foundindifferentspecies,s,

andthepreferentialgeneticorbiochemicalassociation

ofdifferentfamilymemberswithdifferenttypesofsi-

lencingRNAs(Grishoketal.,2001;Caudyetal.,2002;

Zilbermanetal.,2003).TheRISCendonuclease,known

asSlicer,hasnotbeenidentified,suggestingthatit

mightbepresentinsub-stoichiometricamountsand

onlyrecruitedaftertheothercomponentsofRISChave

rpossibility

isthatoneoftheidentifiedRISCcomponentsprovides

theSliceractivitybymeansofanunrecognized

nucleasedomain.

MicroRNAswerefirstreportedtoresideinthemiRNA

ribonucleoproteincomplex(miRNP),whichinhumans

includestheproteinseIF2C2,thehelicaseGemin3,and

Gemin4(Mourelatosetal.,2002).eIF2C2isahuman

Argonautehomologandwaslaterfoundtobeaconstit-

uentofthehumansiRNA-programmedRISC(Martinez

etal.,2002).Furthermore,thehumanlet-7miRNAis

associatedwitheIF2C2andcapableofspecifyingcleav-

ageofanartificialtargetwithperfectcomplementarity

tothemiRNA(Hutva

´

gnerandZamore,2002).Thus,the

miRNPpossessesthesalientpropertiesthatdefinethe

RISC(Hutva

´

gnerandZamore,2002),andalthoughit

mightlaterbeshowntorepresentaparticularsubtype

ofRISC,

perspectiveisfurthersupportedbythedemonstration

thatplantmiRNAscandirectcleavageoftheirnatural

targets(Llaveetal.,2002b;Tangetal.,2003)andthat

siRNAsoriginallydesignedtospecifycleavagecanalso

mediatetranslationalrepression(Doenchetal.,2003;

Zengetal.,2003).

WhenthemiRNAstrandofthemiRNA:miRNA*duplex

isloadedintotheRISC,themiRNA*appearstobe

enisthemechanism

Cell

288

ionsofSmallSilencingRNAs

(A)rrowheadindicatessiteofcleavage.

(B)TranslationalrepressionspecifiedbymiRNAsorsiRNAs.

(C)Transcriptionalsilencing,thoughttobespecifiedbyheterochromaticsiRNAs.

forchoosingwhichofthetwostrandsenterstheRISC?

Theanswerlargelyliesintherelativestabilityofthetwo

endsoftheduplex:forbothsiRNAandmiRNAduplexes,

thestrandthatenterstheRISCisnearlyalwaystheone

whose5Јendislesstightlypaired(Khvorovaetal.,2003;

Schwarzetal.,2003).Thisobservationsuggeststhata

helicase-likeenzyme(yettobeidentified)samplesthe

endsoftheduplexmultipletimes—usuallyreleasingthe

endbeforebeginningtoproductivelyunwindtheduplex

butoccasionallyunwindingtheduplex,resultingina

strongbiasforproductiveunwindingattheeasierend

(Khvorovaetal.,2003;Schwarzetal.,2003)(Figures

2A–2C,steps5).Thiselegantruleforpredictingwhich

strandoftheduplexwillentertheRISCwasinitially

formulatedbasedonobservationsandexperimentsin

animalsystems,butitalsoappliestoplantsiRNAs

(Khvorovaetal.,2003)dictive

valueforthevastmajorityofplantandanimalmiRNAs

stronglyimpliestheexistenceofthemiRNA:miRNA*du-

plexasatransientintermediateinthebiogenesisofall

miRNAs,eventhoseforwhichamiRNA*hasnotyet

wvertebrateandinsectgenes,both

strandsofthemiRNAduplexaccumulateatfrequencies

suggestingthatbothentertheRISC,raisingthepros-

pectthateitherorbothmightbefunctional(Lagos-

Quintanaetal.,2002;Krichevskyetal.,2003;Schwarz

etal.,2003).Theserarecasescanbereconciledwith

theasymmetricloadingoftheRISCbecausetheends

oftheseduplexeshavenearlyequivalentstabilitiesat

theirends;foreachRISCassembled,thehelicaseloads

onlyonestrandofeachduplexbutchooseseachstrand

withsimilarfrequency(Schwarzetal.,2003).

Mechanism:mRNACleavage

MicroRNAscandirecttheRISCtodownregulategene

expressionbyeitheroftwoposttranscriptionalmecha-

nisms:mRNAcleavageortranslationalrepression(Fig-

ures3Aand3B).Accordingtotheprevailingmodel,the

choiceofposttranscriptionalmechanismsisnotdeter-

minedbywhetherthesmallsilencingRNAoriginated

asansiRNAoramiRNAbutinsteadisdeterminedby

theidentityofthetarget:Onceincorporatedintoacyto-

plasmicRISC,themiRNAwillspecifycleavageifthe

mRNAhassufficientcomplementaritytothemiRNA,or

itwillrepressproductivetranslationifthemRNAdoes

nothavesufficientcomplementaritytobecleavedbut

doeshaveasuitableconstellationofmiRNAcomple-

mentarysites(Hutva

´

gnerandZamore,2002;Zenget

al.,2002,2003;Doenchetal.,2003).Althoughthismodel

isgenerallysupportedbyexperimentaltests,highly

functionalsiRNAsandmetazoanmiRNAshavese-

quence-compositiondifferencescenteringatpositions

12and13,whichmightpointtoinherentdifferential

sequencepreferencesforthetworespectivemodesof

repression(Khvorovaetal.,2003).Furthermore,aper-

plexingobservationhascomefromthestudyofaplant

miRNA,miR172,whichappearstoregulateAPETALA2

viatranslationalrepressiondespitethenear-perfect

complementaritybetweenthemiRNAanditssingle

complementarysiteintheAPETALA2ORF(Aukerman

andSakai,2003;Chen,2003).

WhenamiRNAguidescleavage,thecutisatprecisely

thesamesiteasthatseenforsiRNA-guidedcleavage,

i.e.,betweenthenucleotidespairingtoresidues10and

11ofthemiRNA(Elbashiretal.,2001a;Hutva

´

gnerand

Zamore,2002;Llaveetal.,2002b;Kasschauetal.,2003).

Theregisterofcleavagedoesnotchangewhenthe

miRNAisnotperfectlypairedtothetargetatits5Ј

terminus(Kasschauetal.,2003;Palatniketal.,2003).

Therefore,thecutsiteappearstobedeterminedrelative

tomiRNAresidues,notmiRNA:

cleavageofthemRNA,themiRNAremainsintactand

canguidetherecognitionanddestructionofadditional

messages(Hutva

´

gnerandZamore,2002;Tangetal.,

2003).

Mechanism:TranslationalRepression

Fromthebeginning,itwasproposedthatlin-4RNA

slin-

thesimplestinterpretationoftheob-

servationthatlin-4RNAexpressioncoincideswitha

dropinLIN-14proteinwithoutachangeinlin-14mRNA

(Wightmanetal.,1993).Thesurprisecamelater,when

itwasshownthatthepolysomeprofileoflin-14mRNA

atthefirstlarvalstageisindistinguishablefromthat

atlaterlarvalstages,whenLIN-14proteinlevelshave

dropped(OlsenandAmbros,1999).Thesameistrue

forlin-28mRNA,anothermessagetargetedbylin-4RNA

(Seggersonetal.,2002).Twopossibilitieswereputfor-

wardtoexplaintheseresults(OlsenandAmbros,1999).

Thelin-4RNAmightrepresstranslationatastepafter

translationinitiation,inamannerthatdoesnotperceiv-

ablyalterthedensityoftheribosomesonthemessage,

e.g.,bytheslowingorstallingofalltheribosomeson

rnativepossibilityisthattranslation

continuesatthesameratebutisnonproductivebecause

thenewlysynthesizedpolypeptideisspecificallyde-

review,bothofthesemechanisticpossi-

Review

289

bilitiesarelumpedtogetherastranslationalrepression,

asiscommonpractice,eventhoughinthesecondpossi-

bilitypolypeptidesynthesisperseisnotrepressed.A

bettermechanisticunderstandingoflin-4-specified

translationalrepressionawaitsthedevelopmentofan

invitrosystemthatfaithfullyrecapitulateslin-4regula-

tionofitstargets.

Extendingtheanalysisofpolysomeprofilesbeyond

slin-4regulationwillbeimportantforlearning

whetherthepostinitiationmechanismappliesmoregen-

erallytotranslationalrepressionmediatedbyother

,evidencefortranslationalrepression

ofanymetazoanmiRNAtargetsotherthanthoseoflin-4

isscantbecausethefateofthemessengerRNAduring

miRNA-mediatedregulationhasnotyetbeenmonitored

eless,severalindirect

linesofevidencesupportthenotionthatmetazoan

miRNAsotherthanlin-4RNAtypicallymediatetransla-

tionalrepressionratherthanmRNAcleavage:First,

othermetazoanmiRNAs,aswellassiRNAs,canrepress

theexpressionofheterologousreportertranscripts

withoutdecreasingmRNAlevels,ifthesemessages

containeitherthenaturalmiRNAcomplementarysites

fromthemiRNAtarget(Brenneckeetal.,2003)ormulti-

pleartificialcomplementarysitesthathavebulgesor

mismatchesattheircenterwhenpairedtothemiRNA,

suchthatthepatternofbasepairingresemblesthat

foundbetweenthelet-7RNAanditsnaturalcomplemen-

s,lin-413ЈUTR(Zengetal.,

2002,2003;Doenchetal.,2003).Second,thelet-7-pro-

grammedRISCendogenoustohumancellsdoesnot

cleaveanRNAfragmentcontainingthelet-7comple-

slin-41(Hutva

´

gnerand

Zamore,2002).Third,thereisadifferencebetween

plantsandanimalswithregardtotheextentofcomple-

mentaritybetweenthemiRNAsandmRNAs(Rhoades

etal.,2002).Becausenear-perfectcomplementarityis

thoughttoberequiredforRISC-mediatedcleavagebut

nottranslationalrepression,thelowerdegreeofcomple-

mentarityseeninanimalssuggeststhattranslational

repressionismoreprevalentinanimalsthaninplants.

Nonetheless,itwouldbeprematuretoconcludethat

moremetazoanmiRNAregulatorytargetsaretransla-

singlylittle

complementarityappearstobeneededtospecifyde-

tectableRISC-mediatedcleavageinmammaliancells

(Jacksonetal.,2003),suggestingthatitwillnotbelong

beforenaturalexamplesofmiRNA-directedmRNA

cleavagewillbereportedinanimals.

ThecooperativeactionofmultipleRISCsappears

toprovidethemostefficienttranslationalinhibition

(Doenchetal.,2003).Thisexplainsthepresenceofmulti-

plemiRNAcomplementarysitesinmostgenetically

identifiedtargetsofmetazoanmiRNAs(Leeetal.,1993;

Wightmanetal.,1993;Reinhartetal.,2000;Abrahanteet

al.,2003;Linetal.,2003).Thecomputationallyidentified

metazoantargetsalsohavemultiplesites,butthispat-

ternisuninformativebecausethepresenceofmultiple

siteswasacriterionfortheiridentification(Brennecke

etal.,2003;Lewisetal.,2003;Starketal.,2003).Al-

thoughonlyasmallfractionofthemiRNA-mRNAregula-

torypairsareknowninanyanimal,therearealready

instancesinwhichdifferentmiRNAspecieshavebeen

proposedtoregulatethesametargets(Reinhartetal.,

2000;Abrahanteetal.,2003;Linetal.,2003).These

examples,andtheanalogytootherbiologicalregulatory

systems,mostnotablytranscriptionalregulation,have

ledtothegeneralexpectationthatasthelistofknown

metazoanmiRNA:mRNAregulatoryinteractionsbe-

comesmorecomprehensive,combinatorialcontrolwill

beseentobecommon,ifnotthenorm.

Thecomplementarysitesfortheknownmetazoan

targetsresideinthe3Јasmightreflecta

mechanisticpreference,perhapsenablingthebound

complexestoavoidthemRNA-clearingactivityofthe

ll,numerousotherexamplesofeukary-

otictranslationregulationaremediatedthrough3ЈUTR

elements(KuerstenandGoodwin,2003).Alternatively,

itmightreflectabiasinthewaythatmetazoanmiRNA

targetsandcomplementarysitesarediscovered:The

lin-4:lin-14precedentmighthavedirectedsubsequent

searchestothe3ЈUTRs,andconservedcomplementary

sitesareeasiertodistinguishintheUTRs,awayfrom

theconfoundingsequenceconservationoftheORFs.

ThereportedsiRNA-mediatedtranslationalrepression

fromasingleimperfectcomplementarysiteintheORF

ofamammalianreporterconstruct(Saxenaetal.,2003)

illustrateswhyitwouldbeprematuretoconcludethat

mostmetazoanmiRNAregulationismediatedthrough

multiplecomplementarysitesinthe3ЈUTRs.

AmongthedozensofmiRNA-targetrelationshipsthat

havebeenexamined,therehasbeennoevidencefor

miRNAsdirectingupregulationofgeneexpression.

ThesefindingsareconsistentwiththeideathatmiRNAs

areallactingwithinasilencingcomplex,namelythe

miRNAsarelimitedtofunctioningwithin

RISCcomplexes,thereisstilltheprospectthatsome

miRNAsmightspecifymorethanjustposttranscriptional

repression;somemighttargetDNAfortranscriptional

silencing(Figure3C).ArgonauteproteinsandsiRNAs

areassociatedwithDNAmethylationandsilencingin

plants(Metteetal.,2000;Hamiltonetal.,2002;Zilber-

manetal.,2003),heterochromatinformationinfungi

(Halletal.,2002;ReinhartandBartel,2002;Volpeetal.,

2002),andDNArearrangementsinciliates(Mochizuki

etal.,2002).Eachoftheseexamplessuggeststheexis-

Asarenot

involvedinDNAsilencing,itwillbeinterestingtolearn

howtheyavoidenteringthenuclearRISC,particularly

inplants,whereprocessingappearstobecompleted

inthenucleus.

Mechanism:TargetRecognition

Theimportanceofcomplementaritytothe5Јportionof

metazoanmiRNAshasbeensuspectedsincetheobser-

vationthatthelin-14UTRhas“coreelements”ofcom-

plementaritytothe5Јregionofthelin-4miRNA(Wight-

manetal.,1993).Morerecentobservationssupportthis

idea:(1)Residues2–8ofseveralinvertebratemiRNAs

areperfectlycomplementaryto3ЈUTRelementspre-

viouslyshowntomediateposttranscriptionalrepression

(Lai,2002).(2)WithinthemiRNAcomplementarysites

ofthefirstvalidatedtargetsofinvertebratemiRNAs,

mRNAresiduesthatpair(sometimesimperfectly)toresi-

dues2–8ofthemiRNAareperfectlyconservedinor-

thologousmessagesofotherspecies,andacontiguous

helixofatleastsixbasepairsisnearlyalwaysseenin

Cell

290

thisregion(Starketal.,2003).(3)Residues2–8ofthe

miRNAarethemostconservedamonghomologous

metazoanmiRNAs(Lewisetal.,2003;Limetal.,2003a).

(4)WhenpredictingtargetsofmammalianmiRNAs,re-

quiringperfectpairingtotheheptamerspanningresi-

dues2–8ofthemiRNAismuchmoreproductivethan

isrequiringpairingtoanyotherheptamerofthemiRNA

(Lewisetal.,2003).Pairingtothis5Јcoreregionalso

appearstodisproportionallygovernthespecificityof

siRNA-mediatedmRNAcleavage(Jacksonetal.,2003;

Puschetal.,2003),andthesameistrueforaplant

miRNAthatmediatesmRNAcleavage(Reinhart,Mal-

lory,Tang,Zamore,Barton,D.B.,unpublished).

Whyiscomplementaritytothe5Јendofthesmall

RNAuniversallyimportant,regardlessofthemechanism

ofgeneregulation?OnepossibilityisthattheRISCpre-

sentsonlythiscoreregiontonucleatepairingtothe

tationoftheseف7nucleotidesprear-

rangedinthegeometryofanA-formhelixwouldprefer-

entiallyenhancetheaffinitywithmatchedmRNAseg-

tationofapreformedhelicalsegmentof

thislengthwouldbeareasonablecompromisebetween

thetopologicaldifficultiesassociatedwithlongerprear-

rangedhelicalgeometryandthedropininitialbinding

scenario,mismatcheswiththecoreregioninhibitinitial

targetrecognitionandthuspreventcleavageortransla-

tionalrepressionregardlessofthedegreeofcomple-

e

issufficientadditionalpairingaftertheremainderofthe

miRNAisallowedtoparticipate,-

ever,corepairingsupplementedbyjustafewflanking

pairsappearstobesufficienttomediatetranslational

repressionincooperationwithotherRISCsboundto

themessage(Lewisetal.,2003).Interestingly,theability

oftheArgonautePAZdomaintobindbothdouble-and

single-strandedRNAs(Lingeletal.,2003;Songetal.,

2003;Yanetal.,2003),mentionedearlier,wouldmake

itasuitablecandidateforpresentingthecoreandstabi-

lizingthecorepairing.

Mechanism:DistinctionsbetweenmiRNAs

andsiRNAs

BecausemiRNAsandendogenoussiRNAshavea

sharedcentralbiogenesis(Figures2Band2C,steps4–6)

andcanperforminterchangeablebiochemicalfunctions

(Figures3Aand3B),thesetwoclassesofsilencingRNAs

cannotbedistinguishedbyeithertheirchemicalcompo-

eless,important

distinctionscanbemade,particularlyinregardtotheir

origin,evolutionaryconservation,andthetypesofgenes

thattheysilence(Figured2Band2C,steps1–3and

7;BartelandBartel,2003):First,miRNAsderivefrom

genomiclocidistinctfromotherrecognizedgenes,

whereassiRNAsoftenderivefrommRNAs,transposons,

viruses,orheterochromaticDNA(Figure2,steps1).

Second,miRNAsareprocessedfromtranscriptsthat

canformlocalRNAhairpinstructures,whereassiRNAs

areprocessedfromlongbimolecularRNAduplexesor

extendedhairpins(Figure2,steps2).Third,asingle

miRNA:miRNA*duplexisgeneratedfromeachmiRNA

hairpinprecursormolecule,whereasamultitudeof

siRNAduplexesaregeneratedfromeachsiRNAprecur-

sormolecule,leadingtomanydifferentsiRNAsaccumu-

latingfrombothstrandsofthisextendeddsRNA(Figure

2).Fourth,miRNAsequencesarenearlyalwayscon-

servedinrelatedorganisms,whereasendogenous

ypesof

differencesarethebasisofpracticalguidelinesfordis-

tinguishingandannotatingnewlydiscoveredmiRNAs

andendogenoussiRNAs(Ambrosetal.,2003a).

Althoughmuchremainstobelearnedaboutthebio-

logicaltargetsofmiRNAsandendogenoussiRNAs,a

fifthdistinctioncanbemadebetweenthesetwoclasses

ofsilencingRNAs:endogenoussiRNAstypicallyspecify

“auto-silencing,”inthattheyspecifythesilencingof

thesamelocus(orverysimilarloci)fromwhichthey

originate,whereasmiRNAsspecify“hetero-silencing,”

inthattheyareproducedfromgenesthatspecifythe

silencingofverydifferentgenes(Figure2,steps7).Natu-

ralexamplesofauto-silencingincludethesilencingof

viruses,transposons,andtheheterochromaticouterre-

rexampleistheDrosoph-

ilaSu(Ste)repeats,whichgeneratesiRNAsthatsilence

theSu(Ste)repeatsthemselvesaswellastheverysimilar

Stellategenes(Aravinetal.,2001).Atfirstglance,

miR-127andmiR-136mightseemtobeexceptionsto

thisprinciplebecausetheyoriginatefromtheantisense

strandoftheirpresumptivetarget,theRtl1mRNA(Seitz

etal.,2003).However,becausethesegeneslieinan

imprintedlocus,inwhichthemiRNAsareexpressed

fromthematernalchromosomeandtheRtl1mRNAis

expressedfromthepaternalchromosome,these

miRNAscanstillbethoughtofasspecifyinghetero-

fthdistinctionexplainsthegreaterse-

xtent

thatthesiRNAscomefromthesamelocithattheytarget,

amutationaleventthatchangesthesequenceofthe

siRNAwouldalsochangethesequenceofitsregulatory

target,andsiRNAregulationwouldbepreserved—an

unusualcaseofmaintaininganimportantfunctionwith-

contrast,amutationinamiRNAwouldrarelybeaccom-

paniedbysimultaneouscompensatorychangesatthe

lociofitstargets,andthusselectionpressurewould

preservethemiRNAsequence.

WiththesedistinctionsbetweenthemiRNAsandthe

endogenoussiRNAsinmind,itisperhapsworthconsid-

eringhowtoclassifythesmallRNAsthatarisefrom

constructsintroducedintocellsforthepurposeofgene

NAsprocessedfrom

theextendeddouble-strandedregionsoflong,inverted

therextremeare

approximately22ntRNAsprocessedfrompre-miRNA-

azoancasesinwhichthese

stemloopsincludethedeterminantsforthesequential

processingbyDroshathenDicer,classificationisagain

simple;r,clas-

sificationislessclearforRNAsderivingfromtheshort

hairpinconstructstypicallyusedforknockdownsin

mammaliancells(Dykxhoornetal.,2003),whosepro-

cessingisunlikelytoinvolveDroshaandevenmightnot

involveDicer.

Function:RegulatoryRolesofmiRNAs

Themostpressingquestiontoarisefromthediscovery

ofthehundredsofdifferentmiRNAsis,whatareallthese

Review

291

NAsandTheirFunctions:ExamplesforwhichPhenotypicConsequencesofDisruptedorEctopicmiRNARegulationAreKnown

miRNA

Nematodes

lin-4RNA

let-7RNA

lsy-6RNA

Insects

bantammiRNA

miR-14

Mammals

miR-181

Plants

miR165/166

miR172

miR-JAW

miR159

TargetGene(s)

Ce

Ce

Ce

Ce

Ce

lin-14probabletranscriptionfactor

lin-28coldshockdomainprotein

lin-41probableRNA-bindingprotein

hbl-1transcriptionfactor

cog-1transcriptionfactor

BiologicalRoleofmiRNA/TargetGene

Timingofearlylarvaldevelopmentaltransitions

Timingofearlylarvaldevelopmentaltransitions

Timingoflatelarvaldevelopmentaltransitions

Timingoflatelarvaldevelopmentaltransitions

Left/rightasymmetryofchemoreceptorexpression

Apoptosisandgrowthcontrolduringdevelopment

Apoptosisandfatmetabolism

Hematopoieticdifferentiation

Axialmeristeminititionandleafdevelopment

Flowerdevelopment;timingtransitiontoflowering

Leafdevelopment,embryonicpatterning

Leafdevelopment

Refs

1,2

3

4,5

6,7

8

9

10

11

12-14

15-18

19

12,15,19

Dmhidpro-apoptoticprotein

unknown

unknown

AtREVandrelelatedtranscriptionfactors

AtAP2andrelatedtranscriptionfactors

AtTCP4andreleatedtranscriptionfactors

AtMYB33andrelatedtranscriptionfactors

Speciesabbreviations:Caenorhabditiselegans,Ce;Drosophilamelanogaster,Dm;Arabidopsisthaliana,At.

1(Leeetal.,1993);2(Wightmanetal.,1993);3(Mossetal.,1997);4(Reinhartetal.,2000);5(Slacketal.,2000);6(Abrahanteetal.,2003);

7(Linetal.,2003);8(JohnstonandHobert,2003);9(Brenneckeetal.,2003);10(Xuetal.,2003);11(Chenetal.,2004);12(Rhoadesetal.,

2002);13(Tangetal.,2003);14(Emeryetal.,2003);15(Parketal.,2002);16(Kasschauetal.,2003);17(Chen,2003);18(Aukermanand

Sakai,2003);19(Palatniketal.,2003)

tinynoncodingRNAsdoing?Forlin-4,let-7,andseveral

othermiRNAsidentifiedbyforwardgenetics,crucial

cluestotheirfunctionandregulatorytargetscameeven

beforetheirstatusasnoncodingRNAgeneswasdiscov-

ered(MeneelyandHerman,1979;Chalfieetal.,1981;

Ambros,1989;Weigeletal.,2000;Hipfneretal.,2002;

AukermanandSakai,2003;Brenneckeetal.,2003;

JohnstonandHobert,2003;Xuetal.,2003).Theseand

othermiRNAsthathavereportedfunctionsbasedonin

eof

thesecases,functionwasdeterminedbythephenotypic

consequencesofamutatedmiRNAoranalteredmiRNA

complementarysite,eitherofwhichcandisruptmiRNA

rcases,functionwasinferredfrom

theeffectsofmutationsortransgenicconstructsthat

leadtoectopicexpressionofthemiRNA.

ForthevastmajorityofmiRNAs,thephenotypiccon-

sequencesofdisruptedoralteredmiRNAregulationare

r,computationalapproachesarebe-

ingdevelopedtofindtheregulatorytargetsofthe

miRNAs,providingcluestomiRNAfunctionbasedon

theknownrolesofthesetargets(Rhoadesetal.,2002;

Enrightetal.,2003;Lewisetal.,2003;Starketal.,2003).

Computationallypredictedtargetssupportedbysubse-

quentexperimentsorindependentphylogeneticevi-

erimentssupporting

theidentityofthesetargetstypicallyfallintotwoclasses.

IncaseswherethemiRNAisthoughttospecifymRNA

cleavage,thecleavageproductscanbereverse-tran-

scribed,cloned,andsequenced;apreponderanceof

sequencesthatendpreciselyatthepredictedsiteof

cleavageprovidesexperimentalvalidationthatthis

mRNAisacleavagetargetofthecomplementarymiRNA

(Llaveetal.,2002b;Kasschauetal.,2003;Xieetal.,

2003).Toenabledetectionofbothtranslationalrepres-

sionandmRNAcleavage,heterologousreporterassays

canbeusedinwhichthemiRNAcomplementarysites

arefusedtoareportergeneandexpressionisexamined

relativetocontrolconstructs,orinthepresenceand

absenceofthemiRNA(Lewisetal.,2003;Starketal.,

2003).Cautioniswarrantedwheninterpretingreporter

assaysthatinvolvemultimerizationofthemiRNAcom-

plementarysite(s)becausesuchanassaysucceeded

invalidatingamiRNAcomplementarysitethatwasmis-

takenlytakenfromagenethatwasunrelatedtothe

intendedtargetbutsimilarlyannotated(Kawasakiand

Taira,2003a,2003b).Apositiveresultintheheterolo-

gousreporterassayindicatesthatdeterminantsneeded

formiRNAregulationareindeedpresentwithinthe

mRNAfragmentfusedtothereporter,whichtogether

withevolutionaryconservationofboththemiRNAand

itscomplementarysitescanprovidereasonableevi-

se,sucha

hypothesisisconsiderablystrengthenedwithevidence

ofcoincidentexpressionofthemiRNAanditstargetin

theanimalorplant,orexperimentsthatexaminethe

effectsofmanipulatingthemiRNAoritscomplementary

siteinitsnativeinvivocontext.

Function:RolesofPlantmiRNAs

Inplants,miRNAshaveapropensitytopairtomRNAs

withnear-perfectcomplementarity,enablingconvincing

targetstobereadilypredictedformostknownplant

miRNAs(Rhoadesetal.,2002;BartelandBartel,2003).

EvolutionaryconservationofthemiRNA:mRNApairing

inArabidopsisandrice,togetherwithexperimentalevi-

denceshowingthatmiRNAscandirectcleavageoftar-

getedmRNAs,supportsthevalidityofthesepredictions

(Llaveetal.,2002a;Rhoadesetal.,2002;Kasschauet

al.,2003;Tangetal.,2003).TheknownplantmiRNAs

havearemarkablepenchantfortargetingtranscription

factorgenefamilies,particularlythosewithknownor

suspectedrolesindevelopmentalpatterningorcelldif-

ferentiation(Rhoadesetal.,2002;Tables1and2).This

explainsthepleiotropicdevelopmentalphenotypesof

plantsmutantinDCL1(CAF)andHEN1,genesknown

toinfluencemiRNAaccumulation,andAGO1,agene

thatmightbeinvolvedinmiRNAfunction(Bohmertetal.,

Cell

292

NAsandTheirFunctions:ExamplesforwhichStrongEvidencefortheValidityoftheRegulatoryTargetHasBeenReported

butthePhenotypicConsequencesofDisruptedorEctopicmiRNARegulationAreNotYetKnown

miRNA

Insects

miR-7

TargetGene(s)BiologicalRoleofTargetGene(s)

Interprets

Interprets

Interprets

Promotes

Promotes

Promotes

Refs

DmHLHm3basicHLHtranscriptionalrepressor

DmhairybasicHLHtranscriptionalrepressor

Dmm4Brdfamilyprotein

miR-14familyDmgrimantagonistofcaspaseinhibitor

Dmreaperantagonistofcaspaseinhibitor

Dmsickleantagonistofcaspaseinhibitor

Mammals

miR-1HsBrain-derivedneurotrophicfactor(BDNF)

HsGlucose-6-phosphatedehydrogenase(G6PD)

miR-19aHsPtdIns(3,4,5)P3phosphatase(PTEN)

miR-23aHsStromalcell-derivedfactor1(SDF-1)

HsBRN-3bPOU-domaintranscriptionfactor

miR-26aHsSMAD-1transcriptionalco-modulator

miR-34HsDelta1transmembraneprotein

HsNotch1transmembranereceptorforDelta

miR-101HsENX-1polycombgene

HsN-MYCbasicHLHtranscriptionfactor

miR-130HsMacrophagecolonystimulatingfactor-1(MCSF)

Plants

miR170/171AtSCL6-III,-IV&relatedtranscriptionfactors

miR156/157AtSPL2&relatedtranscriptionfactors

miR160AtARF10,ARF17&relatedtranscriptionfactors

miR167AtARF8&ARF6transcriptionfactors

miR164AtCUC1,CUC2&relatedtranscriptionfactors

miR169AtCBF-HAP2DNA-bindingproteins

miR162AtDCL1Dicer-likeRNaseIII

Notch-mediateddecisionsinneuronaldevelopment1,2

Notch-mediateddecisionsinneuronaldevelopment2

Notch-mediateddecisionsinneuronaldevelopment2

apoptosis2

apoptosis2

apoptosis2

3

3

3

3

3

3

3

3

3

3

3

4-7

6,8

6,8

6,8,9

6,8

6

10,11

Growthfactor;neuronaldevelopment

Oxidativestressresistance

Tumorsuppressorgene

Growth&localizationofhematopoieticprogenitorcells

Nueronaldevelopment

RegulatesTGF-dependentgeneexpression

ActivatesNotchduringcell-fatedecisions

Cell-fatedecisionsduringdevelopment

Proliferationofhemotpoeiticcellsandothergeneregulation

Proto-oncogene;celldifferentiation&proliferation

Mononuclearphagocyticlineageregulation

Relatedtogenesforrootradialpatterning

Relatedtogenesforfloralmeristemidentity

Relatedtogenesforauxinresponse&development

Relatedtogenesforauxinresponse&development

Shootapicalmeristemformation&organseparation

unknown

miRNAbiogenesis

Themetazoanregulatorytargentregulatorytargetslistedwere

predictedcomputationallythensupportedwithindependentphylogeneticand/sabbreviations:Drosophila

melanogaster,Dm;human,Hs;Arabidopsisthaliana,At.

1(Lai,2002);2(Starketal.,2003);3(Lewisetal.,2003);4(Llaveetal.,2002a);5(Reinhartetal.,2002);6(Rhoadesetal.,2002);7(Llaveet

al.,2002b);8(Kasschauetal.,2003);9(Parketal.,2002);10(Xieetal.,2003);11(BartelandBartel,2003)

1998;Jacobsenetal.,1999;Parketal.,2002;Reinhartet

al.,2002;Schaueretal.,2002).Ofthefewpredicted

planttargetsthatarenottranscriptionfactors,twoare

DCL1andAGO1,suggestinganegativefeedbackmech-

anismthatcontrolstheexpressionofthesegeneswith

knownorsuspectedrolesinmiRNAbiogenesisand

function(Rhoadesetal.,2002;BartelandBartel,2003;

Xieetal.,2003).

WhyaretheresomanytargetsoftheplantmiRNAs

transcriptionfactorsthathavebeenimplicatedinthe

controlofplantdevelopment?Themodelputforwardto

answerthisquestionproposesthatmanyplantmiRNAs

functionduringcellulardifferentiationbymediatingthe

degradationofkeyregulatorygenetranscriptsinspe-

cificdaughtercelllineages(Rhoadesetal.,2002;Figure

4).Forexample,duringdifferentiation,certaingenes

specifyingalessdifferentiatedstatemightneedtobe

nbeachievedbyrepressingtranscrip-

tion;however,ageneisnotfullyoffuntilitsmessage

,tomorequicklystopex-

pressionofsuchagene,thedifferentiatingcellcande-

ployamiRNAthatspecifiesthecleavageofthatmRNA.

Theactiveclearingofthelingeringregulatorymessages

(orofnewmessagesgeneratedbycontinuedtranscrip-

tion)couldenablerapiddaughtercelldifferentiation

withouthavingtodependonregulatorygeneshavingcon-

respect,miRNAreg-

ulationwouldbeanalogoustoubiquitin-dependentpro-

teindegradation,exceptthatspecificmRNAs,rather

thanproteins,aretargetedfordegradation.

Thismodelconcurswiththeobservationthatamuta-

tiondisruptingthemiRNAcomplementarysiteofPHB

mRNAleadstoamoreexpansivedistributionofthe

message,asifitwerenolongerbeingclearedfromcells

expressingthemiRNA(McConnelletal.,2001;Rhoades

etal.,2002).Italsoexplainswhysomanyoftheinitially

identifiedtargetgenesspecifyformationandidentityof

,plantstemcells(Tables1and2)—these

arepreciselythegenesthatwouldneedtobeturned

elalsowould

applytoscenarioslaterindifferentiationortocases

wherethedaughtercellischoosingamongtwoormore

differentiatedstates,whichwouldexplainthetargeting

oftheothertranscriptsthathaveregulatoryroleslater

ntofcautionintryingtodeduce

thegeneralrolesofplantmiRNAsisthattheknown

setofplantmiRNAsisenrichedinthemoreabundant

miRNAsofplanttissuesandorgansandthusmightnot

mple,miRNAsspecifyingan

undifferentiatedstatewouldhavebeenlesslikelytobe

clonedbecausemostcellsofplantorgansaretypi-

callydifferentiated.

Function:RolesofAnimalmiRNAs

Computationalmethodshaverecentlybeendeveloped

toidentifythetargetsofDrosophilaandmammalian

miRNAs(Enrightetal.,2003;Lewisetal.,2003;Starket

al.,2003).Thesemethodssearchformultipleconserved

regionsofmiRNAcomplementaritywithin3Ј-

Review

293

gModelfortheRolesofmiRNAsthatTargetthe

MessagesofTranscriptionFactorsduringPlantDevelopment

Followingcelldivision,thedaughtercellsinheritmRNAsfromthe

precursorcell(step1).Adifferentiatingdaughtercell(cellonright)

expressesnewtranscriptionfactormessages(green)aswellasa

miRNA(red)complementarytomessagesthatmustbecleared(blue)

inorderforthecelltoprogresstothedifferentiatedstate(step2).

ThemiRNAdirectsthecleavageoftargetmessages,preventing

prolongedorinappropriateexpressionofthetranscriptionalregula-

tor,thusenablingtherapiddifferentiationofthedaughtercell(step

3).(FigureredrawnfromRhoadesetal.,2002,copyrightedbyCell

press,usedwithpermission.)

tifyingtargetsinanimalshasbeenamoredifficulttask

thaninplantsbecauseinanimalstherearefarfewer

mRNAswithnear-perfectcomplementaritytomiRNAs.

Thismakestheanalysisnoisier—muchmoreproneto

rmore,evolutionaryconservation

wasusedasacriterionfortargetidentificationinani-

mals,andthusitcouldnotbeusedasameansto

eless,theex-

perimentalsupportachievedforamajorityofthepredic-

tionstestedisencouraging(Table2),andtherearecom-

pellingreasonstotakeseriouslytheremaininguntested

mple,inoneoftheflystudies,there

werestrikingclustersoffunctionallyrelatedgenes

amongthetoppredictions(Starketal.,2003).Themost

notableexampleswereNotchtargetgenesformiR-7,

proapoptoticgenesformiR-2,andasetofenzymes

involvedinbranched-chainaminoaciddegradationfor

ammalianstudy,over400regulatory

targetswerepredictedwhenusingparametercutoffs

thatgaveasignal-to-noiseratioof3.2:1(Lewisetal.,

2003).Thissignal:noiseratiowasseenonlywhen

restrictingthemiRNAstothosemostconservedamong

mammalsandfish,andonlywhendemandingperfect

complementaritytothemostconservedportionof

miRNAs(the7ntcoresegmentcomprisingresidues2–8

ofthemiRNAs),observationsthatwouldbeexceedingly

difficulttoexplainifmostoftheidentifiedmessages

werenotrelevanttargetsofthemiRNAs.

TheabilitytoidentifyhundredsofmiRNAtargetswith

confidencethatmostofthepredictedtargetsareau-

thenticenablestheanalysisofthetypesofgenesmost

commonlytargetedbymammalianmiRNAs(Lewisetal.,

2003).Asinplants,thepredictedtargetsaresignificantly

enrichedingenesinvolvedintranscriptionalregulation,

suggestingthatthemodelproposedfortherolesof

manyplantmiRNAs(Figure4)couldalsobeoperating

eless,thisenrichmentfortranscrip-

tionalregulatorsisfarlesspronouncedinmammals,

andonlyaminorityofthepredictedmammaliantargets

dictedtargetsrep-

resentasurprisinglybroaddiversityofmolecularfunc-

,incontrasttothe

plantmiRNAs,mostmammalianmiRNAsdonotappear

tobeprimarilyinvolvedattheupperlevelsofthegene

regulatorycascadesbutinsteadappeartobeoperating

atmanylevelstoregulatetheexpressionofadiverse

setofgenes,manyofwhichdonotgoontodirectly

influencetheexpressionofothergenes(Lewisetal.,

2003).

Function:TheQuestionofSpecificity

AlthoughcurrentlistsofpredictedmiRNAtargetspro-

videinsightsandhypothesesforthousandsoffollow-

upexperiments,theycouldbefarfromcomprehensive.

Forexample,intheanimalstudies,thecomputational

methodsusedevolutionaryconservationtodistinguish

miRNAtargetsitesfromthemultitudeof3ЈUTRseg-

mentsthatotherwisewouldscoreequallywellwithre-

gardtothequalityandstabilityofbasepairing(Lewis

etal.,2003;Starketal.,2003).Thecell,ontheother

hand,cannotusethefilterofevolutionaryconservation

ismeanthat

manyoftheseothermRNAswouldinfactbetargeted

ifexpressedinthesamecellsasthecognatemiRNAs?

Perhapsnot—perhapsmiRNAbasepairingisnotthe

nsormRNA

structurecouldrestrictmiRNPaccessibilitytotheUTRs.

Butifthisweregenerallytrue,siRNAknockdownexperi-

mentsmightbeexpectedtohaveamuchlowersuccess

nsormRNAstructurecouldalsofacilitate

recognitionoftheauthenticmRNAtargetsbymeans

ofelementsinthemRNAsthathavethusfarescaped

didateforsuchaproteinistheFragile

X-relatedprotein,aDrosophilaRISCcomponentthatis

relatedtoproteinsknowntobindspecificmRNAs

(Caudyetal.,2002;Ishizukaetal.,2002).

Thealternativeidea—thatthequalityandstabilityof

basepairingisinfacttheprimarydeterminantofspeci-

ficity—ll,thiscomple-

mentarityrequirementincludesa7ntperfectornear-

perfectcorematchnearthe5ЈterminusofthemiRNA

(Lai,2002;Lewisetal.,2003;Starketal.,2003),which

byitselfwouldrepresentadegreeofspecificitycompa-

rabletothatoftheDNAsitesrecognizedbymanytran-

goutsidethe7ntcoresite,al-

thoughperhapslessimportantthanoncethought,

aschromatinstructurelimitsthepossibilitiesfortran-

scription-factorbinding,therestrictedsetofgenes

transcribedineachcelllimitswhichgenesofthege-

Cell

294

thesamewaythatthecooperativeactionofmultiple

transcriptionfactorsincreasesthespecificityoftheir

control,thecooperativeactionofhomotypicandhetero-

typicmiRNA:UTRinteractionswouldprovideanaddi-

tionalmechanismtoincreasespecificityofmiRNAcon-

ethesemechanismsforincreasingthe

regulatoryspecificity,thenotionthattarget-siterecogni-

tionisprimarilydeterminedbymultipleinstancesof

7ntcorecomplementaritywouldimplythatmiRNAs

influencetheexpressionofaremarkablylargenumber

ofdifferentmRNAs(Lewisetal.,2003).

The“manytargets”hypothesisisembracedandpar-

tiallyrationalizedinaproposalthatthemiRNAmilieu,

uniquetoeachcelltype,providesimportantcontextfor

theevolutionofallmRNAsequencesandisproductively

usedtodampentheutilizationofthousandsofmRNAs

(.-,unpublished).FormRNAsthat

shouldnotbeexpressedinaparticularcelltype,

miRNAsreduceproteinproductiontoinconsequential

ultisequivalenttoadiscreteoffswitch,

andthusthesemessages,whichincludetargetsofTable

1,canbethoughtofas“switchtargets.”Inadditionto

theseclassicaltargets,atleastthreeothercategories

ofmRNAscanbeimagined:Formessagescalled“tuning

targets,”miRNAscouldadjustproteinoutputinaman-

nerthatallowsforcustomizedexpressionindifferent

celltypesyetamoreuniformlevelwithineachcelltype.

OthermRNAscouldbesimplybystanders,“neutraltar-

gets,”forwhichdownregulationbymiRNAsistolerated

y,when

thinkingabouttheeffectsofthemiRNAmilieuonthe

evolutionofmRNAsequences,itisalsousefultocon-

sider“antitargets,”messagesunderselectivepressure

toavoidfortuitouscomplementaritytothemultitudeof

miRNAsinthecellswheretheyareexpressed,either

becausesuchcomplementaritywouldinappropriately

dampentheirexpressionorbecauseitwouldtitratethe

miRNAsawayfromtheirpropertargets.

Whilemolecularbiologistswillhavetheirhandsfull

identifyingandcharacterizingadditionalinstances

wheremiRNAsareplayingtheclassicalroleofdiscrete

generegulatoryswitches,computationalandsystems

biologistswillhavetocontendwiththeprospectthata

substantialfractionofallanimalmRNAscouldhavetheir

preciselevelofexpressiondefinedbymiRNAregulation.

TotheextentthatthemiRNAsdirecttranslationalre-

pressionratherthanmRNAcleavage,thisregulationwill

beinvisibletothemostpowerfultoolofthesystems

biologist,e-

less,inonlytwoyearssincetheabundanceofmiRNA

geneswasreported,therehasbeenrapidprogressin

catalogingthemiRNAgenes,determiningtheirexpres-

sionpatterns,andidentifyingtheirregulatorytargets,

providinghopethatthegoalofaccuratelyintegrating

theirfunctionintomodelsofmetazoangeneregulatory

circuitrycanonedayberealized.

Acknowledgments

Ithankmembersofmylab,,,,T.

Tuschl,,andmanyothercolleaguesfortheirinputand

stimulatingdiscussionsoverthepastfewyears,,B.

Bartel,,ce,andothersforhelpfulcommentson

miRNAsinmylabiscurrentlysupportedby

grantsfromtheNIHandtheAlexanderandMargaretStewartTrust.

References

Abrahante,J.E.,Daul,A.L.,Li,M.,Volk,M.L.,Tennessen,J.M.,Miller,

E.A.,andRougvie,A.E.(2003).TheCaenorhabditiseleganshunch-

back-likegenelin-57/hbl-1controlsdevelopmentaltimeandisregu-

4,625–637.

Ambros,V.(1989).Ahierarchyofregulatorygenescontrolsalarva-

57,49–57.

Ambros,V.,Bartel,B.,Bartel,D.P.,Burge,C.B.,Carrington,J.C.,

Chen,X.,Dreyfuss,G.,Eddy,S.R.,Griffiths-Jones,S.,Marshall,

M.,etal.(2003a).

9,277–279.

Ambros,V.,Lee,R.C.,Lavanway,A.,Williams,P.T.,andJewell,D.

(2003b).s.

.13,807–818.

Aravin,A.A.,Naumova,N.M.,Tulin,A.V.,Vagin,V.V.,Rozovsky,Y.M.,

andGvozdev,V.A.(2001).Double-strandedRNA-mediatedsilencing

ofgenomictandemrepeatsandtransposableelementsinDrosoph-

.11,1017–1027.

Aravin,A.A.,Lagos-Quintana,M.,Yalcin,A.,Zavolan,M.,Marks,D.,

Snyder,B.,Gaasterland,T.,Meyer,J.,andTuschl,T.(2003).The

smallRNAprofileduringDrosophilamelanogasterdevelopment.

5,337–350.

Aukerman,M.J.,andSakai,H.(2003).Regulationoffloweringtime

andfloralorganidentitybyaMicroRNAanditsAPETALA2-liketarget

ell10,10.

Bartel,B.,andBartel,D.P.(2003).MicroRNAs:Attherootofplant

development?PlantPhysiol.132,709–717.

Bashirullah,A.,Pasquinelli,A.E.,Kiger,A.A.,Perrimon,N.,Ruvkun,

G.,andThummel,C.S.(2003).Coordinateregulationofsmalltempo-

.

259,1–8.

Basyuk,E.,Suavet,F.,Doglio,A.,Bordonne,R.,andBertrand,E.

(2003).Humanlet-7stem-loopprecursorsharborfeaturesofRNase

cAcidsRes.31,6593–6597.

Bernstein,E.,Caudy,A.A.,Hammond,S.M.,andHannon,G.J.(2001).

RoleforabidentateribonucleaseintheinitiationstepofRNAinter-

409,295–296.

Bohmert,K.,Camus,I.,Bellini,C.,Bouchez,D.,Caboche,M.,and

Benning,C.(1998).AGO1definesanovellocusofArabidopsiscon-

.17,170–180.

Bollman,K.M.,Aukerman,M.J.,Park,M.Y.,Hunter,C.,Berardini,

T.Z.,andPoethig,R.S.(2003).HASTY,theArabidopsisortholog

ofexportin5/MSN5,regulatesphasechangeandmorphogenesis.

Development130,1493–1504.

Brennecke,J.,Hipfner,D.R.,Stark,A.,Russell,R.B.,andCohen,

S.M.(2003).bantamencodesadevelopmentallyregulatedmi-

croRNAthatcontrolscellproliferationandregulatestheproapo-

113,25–36.

Calin,G.A.,Dumitru,C.D.,Shimizu,M.,Bichi,R.,Zupo,S.,Noch,

E.,Aldler,H.,Rattan,S.,Keating,M.,Rai,K.,etal.(2002).Frequent

deletionsanddown-regulationofmicro-RNAgenesmiR15and

.

99,15524–15529.

Catalanotto,C.,Azzalin,G.,Macino,G.,andCogoni,C.(2000).Gene

404,245.

Caudy,A.A.,Myers,M.,Hannon,G.J.,andHammond,S.M.(2002).

FragileX-relatedproteinandVIGassociatewiththeRNAinterfer-

ev.16,2491–2496.

Caudy,A.A.,Ketting,R.F.,Hammond,S.M.,Denli,A.M.,Bathoorn,

A.M.,Tops,B.B.,Silva,J.M.,Myers,M.M.,Hannon,G.J.,andPlas-

terk,R.H.(2003).AmicrococcalnucleasehomologueinRNAief-

425,411–414.

Cerutti,L.,Mian,N.,andBateman,A.(2000).Domainsingenesilenc-

ingandcelldifferentiationproteins:thenovelPAZdomainandredef-

.25,481–482.

Chalfie,M.,Horvitz,H.R.,andSulston,J.E.(1981).Mutationsthat

24,59–69.

Chen,X.(2003).AMicroRNAasaTranslationalRepressorofAPET-

Review

295

hedonline

September11,2003.10.1126/science.1088060.

Chen,C.Z.,Li,L.,Lodish,H.F.,andBartel,D.P.(2004).MicroRNAs

e303,83–86.

Cogoni,C.,andMacino,G.(1999).GenesilencinginNeurospora

crassarequiresaproteinhomologoustoRNA-dependentRNApoly-

399,166–169.

Dalmay,T.,Hamilton,A.,Rudd,S.,Angell,S.,andBaulcombe,D.C.

(2000).AnRNA-dependentRNApolymeraseinArabidopsisisre-

quiredforposttranscriptionalgenesilencingmediatedbyatrans-

101,543–553.

Doench,J.G.,Peterson,C.P.,andSharp,P.A.(2003).siRNAscan

ev.17,438–442.

Dostie,J.,Mourelatos,Z.,Yang,M.,Sharma,A.,andDreyfuss,G.

(2003).NumerousmicroRNPsinneuronalcellscontainingnovelmi-

9,631–632.

Dykxhoorn,D.M.,Novina,C.D.,andSharp,P.A.(2003).Killingthe

messenger:.

CellBiol.4,457–467.

Elbashir,S.M.,Leneckel,W.,andTuschl,T.(2001a).RNAinterfer-

ev.

15,188–200.

Elbashir,S.M.,Martinez,J.,Patkaniowska,A.,Lendeckel,W.,and

Tuschl,T.(2001b).FunctionalanatomyofsiRNAsformediatingeffi-

.

20,6877–6888.

Emery,J.F.,Floyd,S.K.,Alvarez,J.,Eshed,Y.,Hawker,N.P.,Izhaki,

A.,Baum,S.F.,andBowman,J.L.(2003).Radialpatterningofarabi-

.

13,1768–1774.

Enright,A.J.,John,B.,Gaul,U.,Tuschl,T.,Sander,C.,andMarks,

D.S.(2003).Biol.5,R1.

Fagard,M.,Boutet,S.,Morel,J.B.,Bellini,C.,andVaucheret,H.

(2000).AGO1,QDE-2,andRDE-1arerelatedproteinsrequiredfor

post-transcriptionalgenesilencinginplants,quellinginfungi,and

97,11650–

11654.

Fire,A.,Xu,S.,Montgomery,M.K.,Kostas,S.A.,Driver,S.E.,and

Mello,C.C.(1998).Potentandspecificgeneticinterferencebydou-

391,806–811.

Gauwerky,C.E.,Huebner,K.,Isobe,M.,Nowell,P.C.,andCroce,

C.M.(1989).ActivationofMYCinamaskedt(8;17)translocation

86,8867–8871.

Grad,Y.,Aach,J.,Hayes,G.D.,Reinhart,B.J.,Church,G.M.,Ruvkun,

G.,andKim,J.(2003).Computationalandexperimentalidentification

11,1253–1263.

Griffiths-Jones,S.(2004).cAcidsRes.

32,D109–D111.

Grishok,A.,Pasquinelli,A.E.,Conte,D.,Li,N.,Parrish,S.,Ha,I.,

Baillie,D.L.,Fire,A.,Ruvkun,G.,andMello,C.C.(2001).Genesand

mechanismsrelatedtoRNAinterferenceregulateexpressionofthe

sdevelopmentaltiming.

Cell106,23–34.

Hall,I.M.,Shankaranarayana,G.D.,Noma,K.,Ayoub,N.,Cohen,

A.,andGrewal,S.I.(2002).Establishmentandmaintenanceofa

e297,2232–2237.

Hamilton,A.J.,andBaulcombe,D.C.(1999).Anovelspeciesof

e

286,950–952.

Hamilton,A.,Voinnet,O.,Chappell,L.,andBaulcombe,D.(2002).

.

21,4671–4679.

Hammond,S.C.,Bernstein,E.,Beach,D.,andHannon,G.J.(2000).

AnRNA-directednucleasemediatesposttranscriptionalgenesi-

404,293–296.

Hammond,S.M.,Boettcher,S.,Caudy,A.C.,Kobayashi,R.,and

Hannon,G.J.(2001).Argonaute2,alinkbetweengeneticandbio-

e293,1146–1150.

Hipfner,D.R.,Weigmann,K.,andCohen,S.M.(2002).Thebantam

cs161,1527–1537.

Houbaviy,H.B.,Murray,M.F.,andSharp,P.A.(2003).Embryonic

5,351–358.

Hutva

´

gner,G.,andZamore,P.D.(2002).AmicroRNAinamultiple-

e297,2056–2060.

Hutva

´

gner,G.,McLachlan,J.,Pasquinelli,A.E.,Balint,E.,Tuschl,T.,

andZamore,P.D.(2001).AcellularfunctionfortheRNA-interference

enzymeDicerinthematurationofthelet-7smalltemporalRNA.

Science293,834–838.

Ishizuka,A.,Siomi,M.C.,andSiomi,H.(2002).ADrosophilafragile

XproteininteractswithcomponentsofRNAiandribosomalproteins.

GenesDev.16,2497–2508.

Jackson,A.L.,Bartz,S.R.,Schelter,J.,Kobayashi,S.V.,Burchard,

J.,Mao,M.,Li,B.,Cavet,G.,andLinsley,P.S.(2003).Expression

hnol.

21,635–637.

Jacobsen,S.E.,Running,M.P.,andMeyerowitz,E.M.(1999).Disrup-

tionofanRNAhelicase/RNAseIIIgeneinArabidopsiscausesunreg-

pment126,5231–

5243.

Johnson,S.M.,Lin,S.Y.,andSlack,F.J.(2003).Thetimeofappear-

slet-7microRNAistranscriptionallycontrolled

.

259,364–379.

Johnston,R.J.,andHobert,O.(2003).AmicroRNAcontrollingleft/

426,

845–849.

Kasschau,K.D.,Xie,Z.,Allen,E.,Llave,C.,Chapman,E.J.,Krizan,

K.A.,andCarrington,J.C.(2003).P1/HC-Pro,aviralsuppressorof

RNAsilencing,interfereswithArabidopsisdevelopmentandmiRNA

4,205–217.

Kawasaki,H.,andTaira,K.(2003a).Hes1isatargetofmicroRNA-

23duringretinoic-acid-inducedneuronaldifferentiationofNT2cells.

Nature423,838–842.

Kawasaki,H.,andTaira,K.(2003b).retraction:Hes1isatargetof

microRNA-23duringretinoic-acid-inducedneuronaldifferentiation

426,100.

Ketting,R.F.,Haverkamp,T.H.,vanLuenen,H.G.,andPlasterk,R.H.

(1999).s,requiredfortransposonsilencingand

RNAinterference,isahomologofWernersydromehelicaseand

99,133–141.

Ketting,R.F.,Fischer,S.E.J.,Bernstein,E.,Sijen,T.,Hannon,G.J.,

andPlasterk,R.H.A.(2001).DicerfunctionsinRNAinterferenceand

insynthesisofsmallRNAinvolvedindevelopmentaltiminginC.

ev.15,2654–2659.

Khvorova,A.,Reynolds,A.,andJayasena,S.D.(2003).Functional

115,209–216.

Kim,J.,Krichevsky,A.,Grad,Y.,Hayes,G.D.,Kosik,K.S.,Church,

G.M.,andRuvkun,G.(2003).IdentificationofmanymicroRNAsthat

.

hedonlineDecember22,2003,10.1073/

pnas.2333854100.

Knight,S.W.,andBass,B.L.(2001).ArolefortheRNaseIIIenzyme

DCR-1inRNAinterferenceandgermlinedevelopment,incaeno-

e293,2269–2271.

Krichevsky,A.M.,King,K.S.,Donahue,C.P.,Khrapko,K.,andKosik,

K.S.(2003).AmicroRNAarrayrevealsextensiveregulationofmi-

9,1274–1281.

Kuersten,S.,andGoodwin,E.B.(2003).Thepowerofthe3ЈUTR:

.4,626–637.

Lagos-Quintana,M.,Rauhut,R.,Lendeckel,W.,andTuschl,T.

(2001).Identificationofnovelgenescodingforsmallexpressed

e294,853–858.

Lagos-Quintana,M.,Rauhut,R.,Yalcin,A.,Meyer,J.,Lendeckel,

W.,andTuschl,T.(2002).Identificationoftissue-specificmicroRNAs

.12,735–739.

Lagos-Quintana,M.,Rauhut,R.,Meyer,J.,Borkhardt,A.,and

Cell

296

Tuschl,T.(2003).

9,175–179.

Lai,E.C.(2002).MicroRNAsarecomplementaryto3ЈUTRmotifs

.

30,363–364.

Lai,E.C.,Tomancak,P.,Williams,R.W.,andRubin,G.M.(2003).

-

nomeBiol4:R42,1–20.

Lau,N.C.,Lim,L.P.,Weinstein,E.G.,andBartel,D.P.(2001).An

abundantclassoftinyRNAswithprobableregulatoryrolesin

e294,858–862.

Lee,R.C.,andAmbros,V.(2001).AnextensiveclassofsmallRNAs

e294,862–864.

Lee,R.C.,Feinbaum,R.L.,andAmbros,V.(1993).s

heterochronicgenelin-4encodessmallRNAswithantisensecom-

75,843–854.

Lee,Y.,Jeon,K.,Lee,J.T.,Kim,S.,andKim,V.N.(2002).MicroRNA

maturation:

J.21,4663–4670.

Lee,Y.,Ahn,C.,Han,J.,Choi,H.,Kim,J.,Yim,J.,Lee,J.,Provost,

P.,Radmark,O.,Kim,S.,andKim,V.N.(2003).ThenuclearRNase

425,415–419.

Lewis,B.P.,Shih,I.,Jones-Rhoades,M.W.,Bartel,D.P.,andBurge,

C.B.(2003).115,

787–798.

Li,H.,Li,W.X.,andDing,S.W.(2002).Inductionandsuppression

e296,1319–1321.

Lim,L.P.,Lau,N.C.,Weinstein,E.G.,Abdelhakim,A.,Yekta,S.,

Rhoades,M.W.,Burge,C.B.,andBartel,D.P.(2003a).Themicro-

ev.17,991–1008.

Lim,L.P.,Glasner,M.E.,Yekta,S.,Burge,C.B.,andBartel,D.P.

(2003b).e299,1540.

Lin,S.Y.,Johnson,S.M.,Abraham,M.,Vella,M.C.,Pasquinelli,A.,

Gamberi,C.,Gottlieb,E.,andSlack,F.J.(2003).s

hunchbackhomolog,hbl-1,controlstemporalpatterningandisa

4,639–650.

Lingel,A.,Simon,B.,Izaurralde,E.,andSattler,M.(2003).Structure

andnucleic-acidbindingoftheDrosophilaArgonaute2PAZdomain.

Nature426,465–469.

Llave,C.,Kasschau,K.D.,Rector,M.A.,andCarrington,J.C.(2002a).

Cell14,1605–1619.

Llave,C.,Xie,Z.,Kasschau,K.D.,andCarrington,J.C.(2002b).

CleavageofScarecrow-likemRNAtargetsdirectedbyaclassof

e297,2053–2056.

Lund,E.,Gu

¨

ttinger,S.,Calado,A.,Dahlberg,J.E.,andKutay,U.

(2004).e303,95–98.

Martinez,J.,Patkaniowska,A.,Urlaub,H.,Luhrmann,R.,andTuschl,

T.(2002).Single-strandedantisensesiRNAsguidetargetRNAcleav-

110,563–574.

McConnell,J.R.,Emery,J.,Eshed,Y.,Bao,N.,Bowman,J.,and

Barton,M.K.(2001).RoleofPHABULOSAandPHAVOLUTAinde-

411,709–713.

Meneely,P.M.,andHerman,R.K.(1979).Lethals,sterilesanddefi-

cienciesinaregionoftheXchromosomeofCaenorhabditiselegans.

Genetics92,99–115.

Mette,M.F.,Aufsatz,W.,Winden,J.v.d.,Matzke,M.A.,andMatzke,

A.J.(2000).Transcriptionalsilencingandpromotermethylationtrig-

.19,5194–5201.

Mette,M.F.,vanderWinden,J.,Matzke,M.,andMatzke,A.J.(2002).

ShortRNAscanidentifynewcandidatetransposableelementfami-

hysiol.130,6–9.

Michael,M.Z.,O’Connor,S.M.,vanHolstPellekaan,N.G.,Young,

G.P.,andJames,R.J.(2003).Reducedaccumulationofspecific

Res.1,882–91.

Mochizuki,K.,Fine,N.A.,Fujisawa,T.,andGorovsky,M.A.(2002).

Analysisofapiwi-relatedgeneimplicatessmallRNAsingenome

110,689–699.

Moss,E.G.,Lee,R.C.,andAmbros,V.(1997).Thecoldshockdomain

sandis

88,637–646.

Mourelatos,Z.,Dostie,J.,Paushkin,S.,Sharma,A.,Charroux,B.,

Abel,L.,Rappsilber,J.,Mann,M.,andDreyfuss,G.(2002).miRNPs:a

novelclassofribonucleoproteinscontainingnumerousmicroRNAs.

GenesDev.16,720–728.

Mourrain,P.,Beclin,C.,Elmayan,T.,Feuerbach,F.,Godon,C.,

Morel,J.B.,Jouette,D.,Lacombe,A.M.,Nikic,S.,Picault,N.,etal.

(2000).ArabidopsisSGS2andSGS3genesarerequiredforposttran-

101,

533–542.

Nyka

¨

nen,A.,Haley,B.,andZamore,P.D.(2001).ATPrequirements

andsmallinterferingRNAstructureintheRNAinterferencepathway.

Cell107,309–321.

Ohler,U.,Yekta,S.,Lim,L.P.,Bartel,D.P.,andBurge,C.B.(2004).

Patternsofflankingsequenceconservationandacharacteristic

,inpress.

Olsen,P.H.,andAmbros,V.(1999).Thelin-4regulatoryRNAcontrols

developmentaltiminginCaenorhabditiselegansbyblockingLIN-14

.216,

671–680.

Palatnik,J.F.,Allen,E.,Wu,X.,Schommer,C.,Schwab,R.,Carring-

ton,J.C.,andWeigel,D.(2003).Controlofleafmorphogenesisby

425,257–263.

Papp,I.,Mette,M.F.,Aufsatz,W.,Daxinger,L.,Schauer,S.E.,Ray,

A.,vanderWinden,J.,Matzke,M.,andMatzke,A.J.(2003).Evidence

fornuclearprocessingofplantmicroRNAandshortinterferingRNA

hysiol.132,1382–1390.

Park,W.,Li,J.,Song,R.,Messing,J.,andChen,X.(2002).CARPEL

FACTORY,aDicerhomolog,andHEN1,anovelprotein,actinmi-

.12,1484–

1495.

Parrish,S.,Fleenor,J.,Xu,S.,Mello,C.,andFire,A.(2000).Func-

tionalanatomyofadsRNAtrigger:differentialrequirementforthe

6,1077–1087.

Pasquinelli,A.E.,Reinhart,B.J.,Slack,F.,Martindale,M.Q.,Kuroda,

M.,Maller,B.,Srinivasan,A.,Fishman,M.,Hayward,D.,Ball,E.,et

al.(2000).Conservationacrossanimalphylogenyofthesequence

andtemporalregulationofthe21nucleotidelet-7heterochronic

408,86–89.

Pickford,A.S.,Catalanotto,C.,Cogoni,C.,andMacino,G.(2002).

.46,277–303.

Pusch,O.,Boden,D.,Silbermann,R.,Lee,F.,Tucker,L.,andRam-

ratnam,B.(2003).Nucleotidesequencehomologyrequirementsof

cAcidsRes.31,6444–6449.

Reinhart,B.J.,andBartel,D.P.(2002).SmallRNAscorrespondto

e297,1831.

Reinhart,B.J.,Slack,F.J.,Basson,M.,Bettinger,J.C.,Pasquinelli,

A.E.,Rougvie,A.E.,Horvitz,H.R.,andRuvkun,G.(2000).The21

nucleotidelet-7RNAregulatesdevelopmentaltiminginCaenorhab-

403,901–906.

Reinhart,B.J.,Weinstein,E.G.,Rhoades,M.W.,Bartel,B.,andBar-

tel,D.P.(2002).ev.16,1616–1626.

Rhoades,M.W.,Reinhart,B.J.,Lim,L.P.,Burge,C.B.,Bartel,B.,

andBartel,D.P.(2002).

110,513–520.

Saxena,S.,Jonsson,Z.O.,andDutta,A.(2003).SmallRNAswith

a-

tionsforoff-targetactivityofsmallinhibitoryRNAinmammalian

.278,44312–44319.

Schauer,S.E.,Jacobsen,S.E.,Meinke,D.W.,andRay,A.(2002).

DICER-LIKE1:blindmenandelephantsinArabidopsisdevelopment.

TrendsPlantSci.7,487–491.

Schwarz,D.S.,Hutva

´

gner,G.,Haley,B.,andZamore,P.D.(2002).

EvidencethatsiRNAsfunctionasguides,notprimers,intheDro-

10,537–548.

Schwarz,D.S.,Hutvagner,G.,Du,T.,Xu,Z.,Aronin,N.,andZamore,

Review

297

P.D.(2003).AsymmetryintheassemblyoftheRNAienzymecom-

115,199–208.

Seggerson,K.,Tang,L.,andMoss,E.G.(2002).Twogeneticcircuits

represstheCaenorhabditiselegansheterochronicgenelin-28after

.243,215–225.

Seitz,H.,Youngson,N.,Lin,S.P.,Dalbert,S.,Paulsen,M.,Bachelle-

rie,J.P.,Ferguson-Smith,A.C.,andCavaille,J.(2003).Imprinted

microRNAgenestranscribedantisensetoareciprocallyimprinted

.34,261–262.

Sempere,L.F.,Sokol,N.S.,Dubrovsky,E.B.,Berger,E.M.,and

Ambros,V.(2003).TemporalregulationofmicroRNAexpressionin

Drosophilamelanogastermediatedbyhormonalsignalsandbroad-

.259,9–18.

Slack,F.J.,Basson,M.,Liu,Z.,Ambros,V.,Horvitz,H.R.,andRuv-

kun,G.(2000).shetero-

chronicpathwaybetweenthelet-7regulatoryRNAandtheLIN-29

5,659–669.

Smardon,A.,Spoerke,J.M.,Stacey,S.C.,Klein,M.E.,Mackin,N.,

andMaine,E.M.(2000).EGO-1isrelatedtoRNA-directedRNApoly-

meraseandfunctionsingerm-linedevelopmentandRNAinterfer-

.10,169–178.

Song,J.J.,Liu,J.,Tolia,N.H.,Schneiderman,J.,Smith,S.K.,Mar-

tienssen,R.A.,Hannon,G.J.,andJoshua-Tor,L.(2003).Thecrystal

structureoftheArgonaute2PAZdomainrevealsanRNAbinding

.10,1026–1032.

Stark,A.,Brennecke,J.,Russell,R.B.,andCohen,S.M.(2003).Iden-

ol.1,E60.

Tabara,H.,Sarkissian,M.,Kelly,W.G.,Fleenor,J.,Grishok,A.,Tim-

mons,L.,Fire,A.,andMello,C.C.(1999).Therde-1gene,RNA

interference,99,

123–132.

Tang,G.,Reinhart,B.J.,Bartel,D.P.,andZamore,P.D.(2003).A

ev.

17,49–63.

Vance,V.,andVaucheret,H.(2001).RNAsilencinginplants-defense

e292,2277–2280.

Volpe,T.,Kidner,C.,Hall,I.,Teng,G.,Grewal,S.,andMartienssen,

R.(2002).HeterochromaticsilencingandhistoneH3lysine9methyl-

e297,1833–1837.

Weigel,D.,Ahn,J.H.,Blazquez,M.A.,Borevitz,J.O.,Christensen,

S.K.,Fankhauser,C.,Ferrandiz,C.,Kardailsky,I.,Malancharuvil,

E.J.,Neff,M.M.,etal.(2000).

Physiol.122,1003–1013.

Wightman,B.,Burglin,T.R.,Gatto,J.,Arasu,P.,andRuvkun,G.

(1991).Negativeregulatorysequencesinthelin-143Ј-untranslated

regionarenecessarytogenerateatemporalswitchduringCaeno-

ev.5,1813–1824.

Wightman,B.,Ha,I.,andRuvkun,G.(1993).Posttranscriptional

regulationoftheheterochronicgenelin-14bylin-4mediatestempo-

75,855–862.

Xie,Z.,Kasschau,K.D.,andCarrington,J.C.(2003).Negativefeed-

backregulationofDicer-like1inArabidopsisbymicroRNA-guided

.13,784–789.

Xu,P.,Vernooy,S.Y.,Guo,M.,andHay,B.A.(2003).TheDrosophila

microRNAmir-14suppressescelldeathandisrequiredfornormal

.13,790–795.

Yan,K.S.,Yan,S.,Farooq,A.,Han,A.,Zeng,L.,andZhou,M.M.

(2003).StructureandconservedRNAbindingofthePAZdomain.

Nature426,468–474.

Yi,R.,Qin,Y.,Macara,I.G.,andCullen,B.R.(2003).Exportin-5

mediatesthenuclearexportofpre-microRNAsandshorthairpin

ev.17,3011–3016.

Zamore,P.D.,Tuschl,T.,Sharp,P.A.,andBartel,D.P.(2000).RNAi:

double-strandedRNAdirectstheATP-dependentcleavageof

101,25–33.

Zeng,Y.,andCullen,B.R.(2003).Sequencerequirementsformicro

9,112–123.

Zeng,Y.,Wagner,E.J.,andCullen,B.R.(2002).Bothnaturaland

designedmicroRNAscaninhibittheexpressionofcognatemRNAs

9,1327–1333.

Zeng,Y.,Yi,R.,andCullen,B.R.(2003).MicroRNAsandsmallin-

terferingRNAscaninhibitmRNAexpressionbysimilarmechanisms.

100,9779–9784.

Zhang,H.,Kolb,F.A.,Brondani,V.,Billy,E.,andFilipowicz,W.(2002).

HumanDicerpreferentiallycleavesdsRNAsattheirterminiwithout

.21,5875–5885.

Zilberman,D.,Cao,X.,andJacobsen,S.E.(2003).ARGONAUTE4

controloflocus-specificsiRNAaccumulationandDNAandhistone

e299,716–719.