Báo cáo y học: "cGlobal analysis of microRNA target gene expression reveals the potential roles of microRNAs in maintaining tissue identity" ppsx

Genome Biology 2005, 6:P14
Deposited research article
Global analysis of microRNA target gene expression reveals the
potential roles of microRNAs in maintaining tissue identity
Zhenbao Yu, Zhaofeng Jian, Shi-Hsiang Shen, Enrico Purisima, and Edwin
Wang
Addresses: Biotechnology Research Institute, National Research Council of Canada, Montréal, Québec, H4P 2R2, Canada.
Correspondence: Edwin Wang. E-mail: Zhenbao Yu. E-mail:
comment
reviews
reports
deposited research
interactions
information
refereed research
.deposited research
AS A SERVICE TO THE RESEARCH COMMUNITY, GENOME BIOLOGY PROVIDES A 'PREPRINT' DEPOSITORY
TO WHICH ANY ORIGINAL RESEARCH CAN BE SUBMITTED AND WHICH ALL INDIVIDUALS CAN ACCESS
FREE OF CHARGE. ANY ARTICLE CAN BE SUBMITTED BY AUTHORS, WHO HAVE SOLE RESPONSIBILITY FOR
THE ARTICLE'S CONTENT. THE ONLY SCREENING IS TO ENSURE RELEVANCE OF THE PREPRINT TO
GENOME BIOLOGY'S SCOPE AND TO AVOID ABUSIVE, LIBELLOUS OR INDECENT ARTICLES. ARTICLES IN THIS SECTION OF
THE JOURNAL HAVE NOT BEEN PEER-REVIEWED. EACH PREPRINT HAS A PERMANENT URL, BY WHICH IT CAN BE CITED.
RESEARCH SUBMITTED TO THE PREPRINT DEPOSITORY MAY BE SIMULTANEOUSLY OR SUBSEQUENTLY SUBMITTED TO
GENOME BIOLOGY OR ANY OTHER PUBLICATION FOR PEER REVIEW; THE ONLY REQUIREMENT IS AN EXPLICIT CITATION
OF, AND LINK TO, THE PREPRINT IN ANY VERSION OF THE ARTICLE THAT IS EVENTUALLY PUBLISHED. IF POSSIBLE, GENOME
BIOLOGY WILL PROVIDE A RECIPROCAL LINK FROM THE PREPRINT TO THE PUBLISHED ARTICLE.
Posted: 19 December 2005
Genome Biology 2005, 6:P14
The electronic version of this article is the complete one and can be
found online at />© 2005 BioMed Central Ltd
Received: 13 December 2005

This is the first version of this article to be made available publicly.
This article was submitted to Genome Biology for peer review.
This information has not been peer-reviewed. Responsibility for the findings rests solely with the author(s).
 1
GlobalanalysisofmicroRNAtargetgeneexpressionrevealsthepotentialrolesof
microRNAsinmaintainingtissueidentity

ZhenbaoYu
*
,ZhaofengJian,Shi-HsiangShen,EnricoPurisima,andEdwinWang
*

BiotechnologyResearchInstitute,NationalResearchCouncilofCanada,Montréal,
Québec,H4P2R2,Canada

*
CorrespondenceshouldbeaddressedtoE.W.()andZ.Y.
(
)
 2
Abstract
Background:MicroRNAsarenon-codingsmallRNAsof~22nucleotidesthatregulate
thegeneexpressionbybase-paringwithtargetmRNAs,leadingtomRNAcleavageor
translationalrepression.ItiscurrentlyestimatedthatmicroRNAsaccountfor~1%of
predictedgenesinhighereukaryoticgenomesandthatupto30%ofgenesmightbe
regulatedbymicroRNAs.However,onlyveryfewmicroRNAshavebeenfunctionally
characterizedandthegeneralfunctionsofmicroRNAsarenotgloballystudied.
Results:WesystematicallyanalyzedtheexpressionpatternsofmicroRNAtargetsusing
severalpublicmicroarrayprofilesandfoundthattheexpressionlevelsofmicroRNA
targetsaresignificantlylowerinallmouseandDrosophilatissuesthanintheembryos

andthatmicroRNAtargetsaredramaticallyexcludedfromthetissue-specifically
expressedgenegroups.
Conclusion:TheseresultsstronglysuggestthattheglobalfunctionsofmicroRNAsare
largelyinvolvedindrivingtissuedifferentiationandmaintainingtissueidentityrather
thanintissue-specificphysiologicalfunctions.Inaddition,thesefindingsimplythat
disruptionofmicroRNAfunctionsmightcausedelineationofdifferentiatedcells,a
crucialsteptowardscarcinogenesis.

 3
Background
MicroRNAs(miRNAs),encodedinthechromosomalDNAandtranscribedaslonger
stem-loopprecursors,termedpri-miRNAs,arenon-codingsmall(21-23nucleotide)
RNAsthatregulatetheexpressionoftargetmRNAs(reviewedin[1-4]).Upon
transcription,pri-miRNAisconvertedtomaturemiRNAduplexthroughsequential
processingbyRNaseIIIfamilyofendonucleasesDroshaandDicer[3,4].Onestrand

of
theprocessedduplexisincorporatedintoasilencingcomplex

andguidedtotarget
sequencesbybase-pairing(reviewedin[5,6]).Thisresultsinthe

cleavageoftarget
mRNAsorrepressionoftheirproductive

translation[5,6].Inthepastfewyears,several
hundredmiRNAswereidentifiedinanimalsandplants[7-18].Itiscurrentlyestimated
thatmiRNAsaccountfor~1%ofpredictedgenesinhighereukaryoticgenomes[19].

DespitethelargenumberofidentifiedmiRNAs,onlyahandfulofthemhavebeen

functionallycharacterized.Forexample,lin-4andlet-7regulatethetimingoflarval
developmentinC.elegans[20,21].Lsy-6andmiR-273actsequentiallytocontrolthe
left/rightasymmetricgeneexpressioninC.eleganschemosensoryneurons[22].Bantam
promotescellproliferationandinhibitsapoptosisinDrosophila[23].MiR-14suppresses
celldeathandregulatesfatmetabolism[24].MiR-181potentiatesB-celldifferentiation
[25].Thesefindings,togetherwiththecomplicatedexpressionpatternsandlargenumber
ofpredictedtargets,implythatmiRNAsmayregulateabroadrangeofphysiologicaland
developmentalprocesses.

 4
IdentificationofthetargetsofeachmiRNAiscrucialforunderstandingthebiological
functionofmiRNAs.Accumulatingempiricalevidencehasrevealedtheimportanceof
the5-terminalsegmentofmiRNAswith6-8nucleotidesinlength,called“seed”region,
formiRNAfunction[26-29].Forexample,systematicalsinglenucleotidemutation
studiesdemonstratedthatbase-pairingofmiRNAstotheirtargetswith7nucleotidesat
the5-terminusofmiRNAsfromposition2toposition8isessentialandsometimes
sufficientformiRNAstoknockdowntheirtargetexpression[26].Basedonthese
discoveries,severalcomputationalmethodshavebeendevelopedtosearchformiRNA
targets[30-39].Mostofthesemethodshavebeenbiologicallyvalidatedandprovedtobe
veryefficientandaccurate.Theaccuracyofthesemethodshasalsobeenprovedbylarge
scalegeneexpressionprofilestudies[40,41].Inonestudy,Limetal.[40]reportedthat
transfectionsofmiR-1andmiR-124intoHeLacellsrespectivelycauseddown-regulation
oflargenumbersoftargetmRNAsandmajority(76%and88%respectively)ofdown-
regulatedmRNAsshowedasegmentwith6nucleotidescomplementarytothe5’-
terminusofthetransfectedmiRNAs(the“seed”sequence).Inanotherstudy,Krutzfeldt
etal.[41]demonstratedthatknockdownofmiRNA-122byintravenousadministrationof
miRNA“antagomirs”ledtoupregulationanddownregulationofalargenumberofgenes
inliver.Theyfoundthatthe3’-untranslatedregionsofupregulatedgenesarestrongly
enrichedinmiRNA-122“seed”-matchmotifs,whereasdownregulatedgenesaredepleted
inthesemotifs[41].


Thesemethodshaveyieldedalargenumberofcandidatetargetsinbothplantsand
animals.TheestimatedhumanmiRNAtargetscanaccountforuptoonethirdofhuman
 5
genes[35].ThediversityandabundanceofmiRNAtargetsreflectthatmiRNAsandtheir
targetsappeartoformacomplexregulatorynetwork.Forexample,asinglemiRNAcan
regulatehundredsofmRNAsandasinglemRNAcanbetargetedbyseveraldifferent
miRNAs.

Basedonitsbiochemicalfunction,thebiologicalfunctionsofamiRNAshoulddepend
onthecombinationofitsactiontoeachofallitstargetsfortheirexpression.
Theoretically,thetissueswithlowleveloftheexpressionofthetargetsofamiRNAare
probablythetissuesinwhichthemiRNAisfunctionallyinvolved.Systematicalanalysis
ofgeneexpressionprofileshasbeenprovedtobevaluableforstudiesondiverse
biologicalprocesses[42-48].TounderstandtheglobalroleofthesenumerousmiRNAs,
weundertookaglobalanalysisoftheexpressionofmRNAtargetsinhuman,mouseand
Drosophilausingseveralpublicdatasets[49-51].Wefoundthattheexpressionlevelsof
miRNAtargetsaresignificantlylowerinallmouseandDrosophilatissuesthaninthe
embryos.Wealsofoundthatthepercentageofthenumberoftissue-specifically
expressedmiRNAtargetsissignificantlylowerthanthatofubiquitouslyexpressed
miRNAtargets.ThesefindingsstronglysuggestthatmiRNAsplayamostimportantrole
indrivingtissueterminaldifferentiationandparticularlyinmaintainingtissueidentity
ratherthanindeterminingorregulatingtissue-specificphysiologicalfunctions.
 6
Results
ExpressionlevelofmiRNAtargetsistissue-dependent
SincemiRNAfunctiondependsonthecombinationofitsactionstoeachofallitstargets
fortheirexpression,tounderstandtheglobalroleofthesenumerousmiRNAs,we
undertookaglobalanalysisoftheexpressionofmRNAtargetsinhuman,mouseand
Drosophilausingseveralpublicdatasets.Wefirstanalyzedthemicroarrayexpression

datacontaining~10,000genesover41humantissuespublishedbyJohnsonetal.[50].
WecomparedtherelativeexpressionlevelofthetotaltargetsofindividualmiRNAs
acrossthe41humantissues.ForeachmiRNA,wecouldfindthetissuesinwhichits
functionsmaybeinvolvedbysearchingforthetissueswhichhavelowerexpressionlevel
ofitstotaltargets.SinceeachmiRNAhasmanytargetsandtheabsoluteexpression
levelsofthesetargetsareverydifferent,tomakeeachtargetequallycontributetothe
comparison,wefirstrankedeachgeneover41humantissuesaccordingitsexpression
levelsintherespectivetissues(seemethods).Alowerranknumbermeansalower
expressionlevel.ForeachmiRNA,ineachtissue,wecountedthenumberofitstargets
[35]ateachrankposition(TableS1).Bycomparingthedistributionoftheranknumber
ofthetargetsbetweendifferenttissues,wecouldfindtherelativeexpressionlevelsofthe
totaltargetsofamiRNAineachtissuecomparedtoothertissues.Thismethodcould
avoidtheeffectofthebiasoftheabsoluteexpressionlevelsofthemiRNAtargetsonthe
analysis.Figure1ashowsatypicalresultforthedistributionoftheranknumberofmiR-
128atargets[35]inliverandbrain.Inliver,thenumberofmiR-128atargetswithalower
ranknumberisobviouslymorethanthatofthosewithahigherranknumber.Incontrast,
inbrain,theresultisreversed.ThissuggeststhattheoverallexpressionlevelofmiR-
 7
128atargetsinliverislowerthanthatinbrain.Toobtainaquickoverview,wegrouped
thetargetsintotwosets,onewithranknumbersfrom1-20andtheotherwithrank
numbersfrom22-41(seeinsetinFigure1a).WethencalculatedtheRRvalue(see
Methods),N
Rank1-20
/N
Rank22-41
.AhigherRRvaluemeanslowerexpressionlevelofthe
miRNAtargets.ARRvaluemorethanonesuggeststhattheexpressionlevelofthe
targetsofamiRNAinatissueismostlikelytobelowerthanthemedianexpressionlevel
ofthetargetsinalltissues.Forexample,theRRvalueformiR-128ais2.1(197targets/
92targets)inliverand0.57(104targets/184targets)inbrain,suggestingalower

expressionlevelofthemiR-128atargetsinliverthanthatinbrain.Totally,weanalyzed
55miRNAs,eachofwhichhaveatleast55targetspresentedinthemicroarraydataset
(average180targets/miRNA),across41tissues.Wealsodidthesameanalysisfortotal
genespresentinthemicroarraydataset.TheRRvaluesareshowninTableS1.TheRR
valueoftargetgenesforeachmiRNAinatissuewasnormalizedbytheRRvalueoftotal
genesinthesametissueandthenplottedasafunctionofmiRNAsandtissues
respectively(Figure1band1c).Asexpected,foreachindividualmiRNA,theRRvalues
indifferenttissuesareequallydistributedaroundone(thenumberofthetissueswitha
RRvaluemorethanoneissimilartothenumberofthetissueswithaRRvaluelessthan
one)(Figure1B).ForeachmiRNA,thetissueswithhighestRRvaluescouldbefound
fromthisfigureandTableS1,andtheyaremostlikelytobethetissuesinwhichthis
mRNAisfunctionallyinvolved.

However,whenwelookedatthedistributionoftheRRvaluesineachtissue(Figure1c),
tooursurprise,wefoundadramaticdifferencebetweendifferenttissues.Insometissues,
 8
thepreponderanceofmiRNAshaveaRR>1.Conversely,insometissues,RR<1foran
overwhelmingfractionofthemiRNAs.Thissuggeststhattheoverallexpressionlevelof
miRNAtargetsisquitedepletedinsometissuesandenhancedinothers.Forexample,in
bonemarrow,54ofthe55miRNAshaveaRRvaluemorethanone,whereasinbrain,
noneofthemhasaRRvaluemorethanone,indicatingthattheoverallexpressionlevel
ofmiRNAtargetsinbonemarrowislowerthanthatinbrainnomatterwhichmiRNAit
is.SimilarresultswereobtainedwhenusingthemiRNAtargetspublishedbyJohnetal.
[31](FigureS1).
 
ExpressionlevelsofmiRNAtargetsarelowerindifferentiatedtissuesthanin
embryosinbothmammalianandfly
WenextanalyzedtheexpressionofmiRNAtargetsin55mousesamplesusingthegene
expressionprofiledatapublishedbyZhangetal.[51]andthedatasetofmicroRNA
targetspublishedbyJohnetal.[31].Similartowhatwefoundinhumantissues,theRR

valueforallofthe141miRNAsinmousebonemarrowisgreaterthanone(Figure.2a),
suggestingthattheexpressionlevelofmiRNAtargetsinthistissueisobviouslylower
thanthemedianlevelacrosstheothertissues.Asimilarresultwasobservedinother
hematopoiticcells-richorlymphocytes-richtissues,suchasthymus,spleenandlymph
node(Figure2band2c).Interestingly,wefoundanobviouscorrelationofthe
distributionpatternofRRvaluesandtheclusteroftissueproperty(Figure2band2c).
Mostimportantly,theexpressionlevelsofmiRNAtargetsinembryo,embryoheadand
placentaaresignificantlyhigherthanthatinothertissues(Figure2band2c).For
example,in12.5-dayembryo,allofthe141miRNAshaveaRR

valuebelowone.
 9

Tofurtherconfirmtheobservation,wedirectlycomparedtheaverageexpressionvalueof
thetotal2276miRNAtargetsinthemousetissues.Consistingwiththeresultabove
derivedfromrankinganalysis,theaverageexpressionlevelofmiRNAtargetgenesis
higherin12.5-dayembryothanthatinanytissuesexceptforthe14.5-dayembryohead
andcortex(Figure3).Forexample,thelevelsinbonemarrow,spleen,thymusandlymph
nodearerespectivelymorethantwotimeslowerthanthatin12.5-dayembryo.

Wethenfocusedonthecomparisonofmousetissuestoembryo.Todoso,wecounted
thetotalnumberofmiRNAtargetswhoseexpressionlevelislowerinagiventissuethan
thatin12.5-dayembryo(N
<E12.5
)anddivideditbythetotalnumberofmiRNAtargets
whoseexpressionlevelishigherthanthatin12.5-dayembryo(N
>E12.5
).Asshownin
Figure4a,inalltissuesexceptfor14.5-dayembryohead,thelowerexpressedtarget
numberismorethanhigherexpressedtargetnumber(N

<E12.5
/N
>E12.5
>1).Asacontrol,
wecarriedoutthesamecalculationsforallgenes.Weseethat,foralltissues,theN
<E12.5
/
N
>E12.5
valueoftotalgenesislowerthanthatofmiRNAtargets(Figure4a).Resembling
statisticaltests(see“Methods”fordetails)demonstratedthatthedifferenceissignificant
(P<0.0002foralmostallofthetissues,TableS2).Tofurtherconfirmtheobservation,we
performedthesameanalysiswiththetotalmiRNAtargetspublishedbyLewisetal.[35]
andKreketal.[33]respectively.Wefoundthesimilarpatterns(Figure4band4c).
Figure4DshowsthatthedataobtainedusingeachsetofthemiRNAtargetspublishedby
eachofthethreegroupsarehighlycorrelated.Thissupportsboththequalityofthe
originaldataandouranalysismethod.
 10

Takentogether,theoverallexpressionlevelofmiRNAtargetsineverydifferentiated
mousetissuesislowerthanthatinmouseembryo,suggestingthatmiRNAsmayplayan
importantrolefordeterminingthefateoftissuedifferentiationduringembryo
development,andmaintainingthetissueidentityinthelaterstage.

Todetermineiftheobservedexpressionpatternisconservedinotherspecies,we
analyzedthepublishedgeneexpressionprofileover75stagesofthewholelifecycleof
Drosophila[49].AsshowninFigure5aand5b,comparedto23-24hembryo,whilethe
ratiooflowerexpressedmiRNAtargetstohigherexpressedmiRNAtargetsremainsthe
sameduringembryoperiod,itdramaticallyincreasesstartingfromlarvalperiodsand
lastingtoadulthood(p<0.0002foralllarvaandmaleadult,seeTableS3formore

details).Thereisaclearcorrespondence(R=0.93)betweenthedatacalculatedusingthe
targetsetspublishedbyEnrightetal.[30]andStarketal.[30,30,38]respectively(Figure
5c).ThisdatastronglysuggestthatmiRNAsplayimportantrolesfordeterminingthe
timingoftissuedifferentiationduringlarvaperiodofDrosophiladevelopmentand
maintainingthetissueidentityduringtheadulthood.

Itshouldbenotedthatthehumanmicroarraydataset[50]weusedinthisstudydoesnot
containhumanembryoandalargescalegeneexpressionprofilecontaininghuman
embryoisnotavailable.Consequently,wecouldnotperformthecomparisonofthe
expressionlevelsofhumanmiRNAtargetsbetweenhumantissuesandhumanembryos.

 11
MiRNAsmorefrequentlytargetubiquitouslyexpressedgenesthantissue-specific
genes

TodetermineifmiRNAsareinvolvedintissue-specificphysiologicalfunctions,we
analyzedthetissue-specificityofmiRNAtargetexpressionusingthemicroarraydata
representing21622mousegenes[51]including2276predictedmiRNAtargets[31].Each
ofthesegeneswasexpressedinatleastoneofthe55mousetissues[51].Boththe21622
genesandthe2276miRNAtargetswereclassifiedinto55groupsaccordingtothe
numberoftissues(between1to55)inwhichthegenewasexpressed.Wecountedthe
numbersofmiRNAtargetsandtotalgenesineachgrouprespectively.Todetermineif
themiRNAtargetsareenrichedinorexcludedfromsomegroups,wecomparedthe
percentageofmiRNAtargetstothetotalgenesineachgroupwiththepercentage
(10.5%)ofthetotalmiRNAtargets(2276)tothetotalgenes(21622).AsshowninFigure
6,amongthegenesthatareexpressedubiquitously,thetargetsofmiRNAsareover-
represented.Forexample,ingenesfoundingroups45-55(i.e.,genesfoundinalmostall
thetissues),approximately20%aremiRNAtargets,aroundtwicethefractionofmiRNA
targetsinthewholelistofgenes(10.5%).Incontrast,amonggenesthatareexpressedin
asmallnumberoftissues,miRNAtargetsareunder-represented.Forexample,amongthe

genesthatarespecificallyexpressedinonly1-4tissues,thefractionofmiRNAtargets
presentisabouthalforlessthanthatinthegeneralgenepopulation.Sincetissue-
specificallyexpressedgenesaremostlyinvolvedintissue-specificphysiological
functions,thisobservationsuggeststhatmiRNAsarepreferablyinvolvedindetermining
andmaintainingtissueidentityratherthanplayingatissue-specificphysiologicalrole.By
 12
thatwemeanthetargetsofmiRNAsarecommontomosttissues.Amongthedifferent
tissues,varyingcombinationsofthesecommongenesaresuppressedbythemiRNA
resultinginthespecifictissuetype.Asananalogy,thecommongenesarelikeastarting
blockofmarblewiththemiRNAsbeingthesculptorthatchiselsawayeverythingthatis
notneededforthefinalfigure.

 13
Discussion
Inthisstudy,wedevelopedamethodfortheanalysisoftheglobalexpressionpatternsof
miRNAtargets.Inthismethod,weconsideredthebiasoftheabsoluteexpressionlevels
ofeachmiRNAtargetsandlimiteditseffectonthecalculationbyrankingeachgeneover
allsamplesaccordingitsexpressionlevelsintherespectivetissuesandcountingthetotal
numberoftargetsateachrankingposition.BycomparingtheRRvaluesdefinedinthis
studybetweeneachsamples,weareabletofindthetissuesinwhichmiRNAtargetsare
lessexpressed.BasedonthebiochemicalfunctionofmiRNAs,thetissueswithlowest
expressionlevelsofitstargetsarelikelytobethetissuesinwhichamiRNAis
functionallyinvolved.Therefore,inthisstudy,wefirstprovideatoolfortheprediction
ofmiRNAfunctionsthroughanalysisoftheirtargetexpression.

Moreimportantly,wefoundthattheexpressionlevelofmiRNAtargetsindifferentiated
tissuesissignificantlylowerthanthatinembryosinbothmammalianandflyandthat
miRNAsmorefrequentlytargetubiquitouslyexpressedgenesthantissue-specificgenes.
ThesefindingsstronglysuggestthatmiRNAsplayamostimportantroleindrivingtissue
terminaldifferentiationandmaintainingtissueidentityratherthanindeterminingor

regulatingtissue-specificphysiologicalfunctions.Previousstudiessuggestthat10%to
30%ofhumangenesarepotentialmiRNAtargets[31,34].However,analysisofthe
specificgeneontology(GO)molecularfunctionclassificationamongthepredicted
targetscouldnotrevealanyspecificbiologicalfunctionsofanimalmiRNAssincethe
animalmiRNAtargetspopulatedmanyGOfunctionalcategories[31,34].Only~13%of
mammalianmiRNAtargetspredictedbyLewisetal.wereinvolvedindevelopment
 14
accordingtotheGObiologicalprocesscategories[34].Failingtopredictthefunctionsof
miRNAtargetsthroughGOanalysismaybesimplycausedbytheevolvingstageofthe
classificationofGOfunctioncategories.Alternatively,onprinciple,thefunctionsof
miRNAscouldnotbepredictedbytheGOfunctioncategoriesoftheirtargetsbecausethe
expressionandthereforethefunctionsoftheirtargetsareproposedtobeturneddownbut
notinducedbymiRNAexpressionandGOanalysiscanonlytellthefunctionofagroup
ofgenesinaGOfunctioncategorywhentheyareexpressedorup-regulatedbutnotthat
whentheyaredown-regulated.Ourstudiesdemonstratethatstatisticalanalysisofthe
expressionofmiRNAtargetscanreflecttheglobalfunctionsofmiRNAs.The
statisticallylowerexpressionlevelofmiRNAtargetsinmaturedtissuesthaninembryo
demonstratesthatmiRNAsplayanimportantrolenotonlyfordeterminingtissuefate
duringembryodevelopmentbutalsoformaintainingidentityandpreventing
dedifferentiationofmaturedtissues.

OurstudiesalsodemonstratethattheoverallexpressionlevelsofmiRNAtargetscould
beclusteredinlargepartaccordingtotheiranatomiclocationsorphysiologicalfunctions.
Forexample,theaverageexpressionlevelsofmiRNAtargetsinlymphoidandmyeloid
tissues(lymphnode,thymus,spleenandbonemarrow)aremuchlowerthanthatinmost
ofothertissues.ThisresultindicatesthatmiRNAsmightalsoplayaveryimportantrole
forthedifferentiationofhematopoitieticlineagefrommyeloidandlymphoidprogenitors
inadditiontothatforthedifferentiationoftissuesfromembryocells.

 15

Uptonow,themolecularmechanismdeterminingthelowerexpressionlevelofmiRNA
targetsindifferentiatedtissuesthanthatinembryosremainstobeelucidated.One
potentialreasonisthatthemiRNAexpressionlevelislowerinembryos.Thisistruefor
zebrafish.Recently,Wienholdsetal.havereportedthatmostzebrafishmiRNAswerenot
detectedduringearlydevelopment[52].However,theresultsregardingtotheglobal
expressionpatternsofmiRNAsduringmouseorDrosophilaembryodevelopmentare
currentlynotavailablealthoughthedifferentialexpressionofmiRNAsindifferenttissues
wereclearlydemonstrated[53-62].AnotherpossibilityisthattheactivityofmiRNA
machineryislowerinembryosthaninothertissues.Forexample,Yangetal.[63]
reportedthatdicer,animportantproteinforbothmiRNAbiogenesisandmiRNA
function,startsexpressionin7-dayoldmouseembryosandremainsstablethrough17-
dayembryos.

Inthisstudy,wealsofoundthatmiRNAtargetsaresignificantlyenrichedinubiquitously
expressedgenesandlargelyexcludedfromtissue-restrictivelyexpressedgenes.This
resultimpliesthatmiRNAsplaylessimportantrolefortissue-specificphysiological
functions.Instead,themajorbiologicalfunctionofmiRNAsistodeterminethefateof
tissuedifferentiationandmaintainthetissueidentity.

OurstudyalsoindicatesthatreductionofmiRNAexpressionmightcausedelineationof
differentiatedcells,acrucialsteptowardscarcinogenesis.Thisisconsistentwiththe
recentdiscoveries[64,65].Forexample,morethan50%ofhumanmiRNAsarelocatedin
chromosomeregionsinvolvedinhumancancers[64].Mostrecently,miRNAexpression
 16
profileshasrevealedthatmostofmiRNAshadlowerexpressionlevelsintumors
comparedwithnormaltissues[65].Hence,understandingtheglobalroleofmiRNAsin
maintaininglineageofdifferentiatedtissuesandcellshasgreatimpactonthestudiesof
miRNAsincancergenetics.
 17


Methods
Datasetsusedinthisstudy
ThedatasetsusedinthisstudyincludetwocompletelistsofhumanmiRNAtargets
publishedbyLewisetal.[35]andJohnetal.[31],threecompletelistsofmousemiRNA
targetspublishedbyJohnetal.[31],Lewisetal.[35]andKreketal.[33],twocomplete
listsofDrosophilamiRNAtargetspublishedbyEnrightetal.[30]andStarketal.
[30,30,38],amicroarrayexpressiondatasetformorethan10,000humangenesin41
tissuesand

celllines[50],amicroarrayexpressiondatasetfornearly40,000knownand
predictedmRNAsin55mousetissues[51]andamicroarrayexpressiondatafornearly
one-thirdofallDrosophilagenesduringthewholelifecycle[49,51].

RankingofmiRNAtargetgenes
TostudythecorrelationoftheexpressionofmiRNAsandtheirtargets,weanalyzedthe
microarrayexpressiondatacontaining~10,000genesover41humantissuespublished
byJohnsonetal.[50]byrankingeachgeneoveralltissuesaccordingtoitsexpression
levelintherespectivetissueasdescribedpreviously[40].Forexample,ifageneis
expressedlessinadefinedtissuethaninanyothertissues,theranknumberofthisgene
inthistissueis1.Similarly,ranknumber41meansthattheexpressionlevelofagenein
atissueishigherthanthatinanyothertissues.ForeachmiRNA,wecollectedallofits
predictedtargetspresentedinthemicroarraydatasetandobtainedtheranknumberof
eachofitstargetsinanygiventissues.ThemiRNAtargetsweusedinthisanalysisare
 18
fromthedatasetspublishedbyLewisetal.[34]andJohnetal.[31]respectively.On
average,180targetspermiRNAcouldbefoundinthemicroarraydataset.

Tofacilitateamoreglobalview,wealsogroupedthegenesintotwosets,onewithlower
halfofranknumbersandtheotherwithhigherhalf.Whenthereisanoddnumberof
ranks,themiddlemostrankisexcluded.Inourspecificexamplewith41tissuesamples,

onesetconsistsofranks1-20,andtheotherfrom22-41,excludingrank21.Forany
miRNAinanytissue,wecountedthetotalnumberofitstargetswithinlower-ranksetand
divideditbythetotalnumberofitstargetswithinhigher-ranksettoyieldthecalculated
rankratio,RR.Forexample,foramiRNAinthe41-tissueset,RR=N
Rank1-20
/N
Rank22-41
.
TheRRvalueisanindicatorofthepreferentialtissueexpressionofagivenmiRNA’s
targetgenes.AnRRvaluegreaterthanonemeansthatthemajorityofexpressedtargets
ofamiRNAinthistissuehavealowerexpressionlevelthanthemedianlevelof
expressionofthemiRNA’stargetsacrossallthetissues.IftheRR

valueofamiRNAis
greaterinaparticulartissuethanthatinanyothers,theexpressionlevelofthetargetsof
thismiRNAinthistissueisverylikelytobethelowestamongthe41humantissues.We
alsodidthesameanalysisfortotalgenespresentedinthemicroarraydataset.TheRR
valueoftargetgenesforeachmiRNAinatissuewasnormalizedbytheRRvalueoftotal
genesinthesametissueandthenplottedasafunctionoftissuesandmiRNAs
respectively.TheRRvalueprovidesaglobaldescriptorofthetissuedistributionofan
miRNA’stargetgenesratherthantheexpressionlevelsofindividualgenes.Itdoesnot
providegene-specificinformationbutallowstheextractionofglobaltrendsofagroupof
genes(miRNAtargetgenes)amidthenoisydataforindividualgenes.
 19

Thismethodwasalsousedtoanalyzeamicroarraydatasetcontaining55mousesamples.
Inthiscase,RR=N
Rank1-27
/N
Rank29-55

.

ComparisonoftheexpressionlevelofmiRNAtargetsintheembryosofmouseand
Drosophilawiththatintheirtissues
TocomparetheexpressionlevelofmiRNAtargetsinmousetissueswiththatinmouse
embryo,wecountedthetotalnumbersofmiRNAtargetswhoseexpressionlevelislower
andhigherrespectivelyinadefinedtissuethanthatin12.5-dayembryo,andcalculated
theratiooflower-expressedtargetstothehigher-expressedtargets.Aratiomorethanone
meansthatthenumberoflower-expressedtargetsismorethanhigher-expressedtargets.
Toobtainthestatisticalsignificanceofthisratio,weperformedresamplingstatistical
tests.Ineachresampling,werandomlysub-pooledthesamenumberofgenesasthe
numberofmiRNAtargetsfromthepooloftotalgenesthatarelower-andhigher-
expressedgenesinadefinedtissuetotheembryo.Wecalculatedtheratiooflower-
expressedgenestothehigher-expressedgenesinthissub-poolanddefineditasR
random
.
WetestedthenullhypothesisR
random
≥R
mirna
byperforming5,000timesofresampling
tests.Werejectedthehypothesisifp<0.05.

AsimilarmethodwasusedtocomparetheexpressionofmiRNAtargetsinthedifferent
periodsofDrosophilalifecyclewith23-24hembryousingamicroarraydataset
publishedbyArbeitmanetal.[49,51],exceptthattheratioofnumberofmiRNAtargets
whoseexpressionlevelinadefinedperiodistwo-foldlowerthanthatin23-24hembryo
 20
tothatofmiRNAtargetswhoseexpressionlevelistwotimeshigherthanthatin23-24h
embryowasrepresented.Thestatisticalanalysisforthissetwasconductedasdescribed

above.

Analysisofthetissue-specificityofmiRNAtargetexpression
UsingthemicroarraydatabasepublishedbyZhangetal.[51],whichcontains21622
mousegenesincluding2276predictedmiRNAtargets,weanalyzedthetissue-specificity
ofmiRNAtargetexpression.Themousegenesareclassifiedinto55groupsaccordingto
thenumberoftissues(1to55)inwhichagenewasexpressed.Thetotalnumbersof
miRNAtargetsandtotalgenesineachgroupwerecountedrespectively.Thepercentage
ofmiRNAtargetstothetotalgenesineachgroupwascalculatedandcomparedtothe
percentage(10.52%)ofthetotalmiRNAtargets(2276)tothetotalgenes(21622)for
determiningifthemiRNAtargetsareenrichedinorexcludedfromtherespectivegroup.

 21
ReferenceList

 1. AmbrosV:microRNAs:tinyregulatorswithgreatpotential.Cell2001,
107:823-826.
 2. BartelDP:MicroRNAs:genomics,biogenesis,mechanism,andfunction.Cell
2004,116:281-297.
 3. CullenBR:TranscriptionandprocessingofhumanmicroRNAprecursors.
MolCell2004,16:861-865.
 4. KimVN:MicroRNAbiogenesis:coordinatedcroppinganddicing.NatRev
MolCellBiol2005,6:376-385.
 5. CarmellMA,HannonGJ:RNaseIIIenzymesandtheinitiationofgene
silencing.NatStructMolBiol2004,11:214-218.
 6. MeisterG,TuschlT:Mechanismsofgenesilencingbydouble-strandedRNA.
Nature2004,431:343-349.
 7. Lagos-QuintanaM,RauhutR,LendeckelW,TuschlT:Identificationofnovel
genescodingforsmallexpressedRNAs.Science2001,294:853-858.
 8. LauNC,LimLP,WeinsteinEG,BartelDP:AnabundantclassoftinyRNAs

withprobableregulatoryrolesinCaenorhabditiselegans.Science2001,
294:858-862.
 9. LeeRC,AmbrosV:AnextensiveclassofsmallRNAsinCaenorhabditis
elegans.Science2001,294:862-864.
 10. Lagos-QuintanaM,RauhutR,YalcinA,MeyerJ,LendeckelW,TuschlT:
Identificationoftissue-specificmicroRNAsfrommouse.CurrBiol2002,
12:735-739.
 11. BentwichI,AvnielA,KarovY,AharonovR,GiladS,BaradO,BarzilaiA,Einat
P,EinavU,MeiriEetal.:Identificationofhundredsofconservedand
nonconservedhumanmicroRNAs.NatGenet2005,37:766-770.
 12. BerezikovE,GuryevV,vandeBJ,WienholdsE,PlasterkRH,CuppenE:
Phylogeneticshadowingandcomputationalidentificationofhuman
microRNAgenes.Cell2005,120:21-24.
 13. GradY,AachJ,HayesGD,ReinhartBJ,ChurchGM,RuvkunG,KimJ:
ComputationalandexperimentalidentificationofC.elegansmicroRNAs.
MolCell2003,11:1253-1263.
 22
 14. Jones-RhoadesMW,BartelDP:Computationalidentificationofplant
microRNAsandtheirtargets,includingastress-inducedmiRNA.MolCell
2004,14:787-799.
 15. KimJ,KrichevskyA,GradY,HayesGD,KosikKS,ChurchGM,RuvkunG:
IdentificationofmanymicroRNAsthatcopurifywithpolyribosomesin
mammalianneurons.ProcNatlAcadSciUSA2004,101:360-365.
 16. Lagos-QuintanaM,RauhutR,MeyerJ,BorkhardtA,TuschlT:NewmicroRNAs
frommouseandhuman.RNA2003,9:175-179.
 17. LimLP,LauNC,WeinsteinEG,AbdelhakimA,YektaS,RhoadesMW,Burge
CB,BartelDP:ThemicroRNAsofCaenorhabditiselegans.GenesDev2003,
17:991-1008.
 18. SunkarR,GirkeT,JainPK,ZhuJK:Cloningandcharacterizationof
microRNAsfromrice.PlantCell2005,17:1397-1411.

 19. Griffiths-JonesS:ThemicroRNARegistry.NucleicAcidsRes2004,32:D109-
D111.
 20. LeeRC,FeinbaumRL,AmbrosV:TheC.elegansheterochronicgenelin-4
encodessmallRNAswithantisensecomplementaritytolin-14.Cell1993,
75:843-854.
 21. ReinhartBJ,SlackFJ,BassonM,PasquinelliAE,BettingerJC,RougvieAE,
HorvitzHR,RuvkunG:The21-nucleotidelet-7RNAregulatesdevelopmental
timinginCaenorhabditiselegans.Nature2000,403:901-906.
 22. JohnstonRJ,HobertO:AmicroRNAcontrollingleft/rightneuronal
asymmetryinCaenorhabditiselegans.Nature2003,426:845-849.
 23. BrenneckeJ,HipfnerDR,StarkA,RussellRB,CohenSM:bantamencodesa
developmentallyregulatedmicroRNAthatcontrolscellproliferationand
regulatestheproapoptoticgenehidinDrosophila.Cell2003,113:25-36.
 24. XuP,VernooySY,GuoM,HayBA:TheDrosophilamicroRNAMir-14
suppressescelldeathandisrequiredfornormalfatmetabolism.CurrBiol
2003,13:790-795.
 25. ChenCZ,LiL,LodishHF,BartelDP:MicroRNAsmodulatehematopoietic
lineagedifferentiation.Science2004,303:83-86.
 26. BrenneckeJ,StarkA,RussellRB,CohenSM:PrinciplesofmicroRNA-target
recognition.PLoSBiol2005,3:e85.
 27. DoenchJG,SharpPA:SpecificityofmicroRNAtargetselectionin
translationalrepression.GenesDev2004,18:504-511.
 23
 28. KiriakidouM,NelsonPT,KouranovA,FitzievP,BouyioukosC,MourelatosZ,
HatzigeorgiouA:Acombinedcomputational-experimentalapproachpredicts
humanmicroRNAtargets.GenesDev2004,18:1165-1178.
 29. KloostermanWP,WienholdsE,KettingRF,PlasterkRH:Substrate
requirementsforlet-7functioninthedevelopingzebrafishembryo.Nucleic
AcidsRes2004,32:6284-6291.
 30. EnrightAJ,JohnB,GaulU,TuschlT,SanderC,MarksDS:MicroRNAtargets

inDrosophila.GenomeBiol2003,5:R1.
 31. JohnB,EnrightAJ,AravinA,TuschlT,SanderC,MarksDS:Human
MicroRNAtargets.PLoSBiol2004,2:e363.
 32. KiriakidouM,NelsonPT,KouranovA,FitzievP,BouyioukosC,MourelatosZ,
HatzigeorgiouA:Acombinedcomputational-experimentalapproachpredicts
humanmicroRNAtargets.GenesDev2004,18:1165-1178.
 33. KrekA,GrunD,PoyMN,WolfR,RosenbergL,EpsteinEJ,MacMenaminP,da
P,I,GunsalusKC,StoffelMetal.:CombinatorialmicroRNAtarget
predictions.NatGenet2005,37:495-500.
 34. LewisBP,ShihIH,Jones-RhoadesMW,BartelDP,BurgeCB:Predictionof
mammalianmicroRNAtargets.Cell2003,115:787-798.
 35. LewisBP,BurgeCB,BartelDP:Conservedseedpairing,oftenflankedby
adenosines,indicatesthatthousandsofhumangenesaremicroRNAtargets.
Cell2005,120:15-20.
 36. RhoadesMW,ReinhartBJ,LimLP,BurgeCB,BartelB,BartelDP:Prediction
ofplantmicroRNAtargets.Cell2002,110:513-520.
 37. RobinsH,LiY,PadgettRW:IncorporatingstructuretopredictmicroRNA
targets.ProcNatlAcadSciUSA2005,102:4006-4009.
 38. StarkA,BrenneckeJ,RussellRB,CohenSM:IdentificationofDrosophila
MicroRNAtargets.PLoSBiol2003,1:E60.
 39. WangXJ,ReyesJL,ChuaNH,GaasterlandT:Predictionandidentificationof
ArabidopsisthalianamicroRNAsandtheirmRNAtargets.GenomeBiol
2004,5:R65.
 40. LimLP,LauNC,Garrett-EngeleP,GrimsonA,SchelterJM,CastleJ,BartelDP,
LinsleyPS,JohnsonJM:MicroarrayanalysisshowsthatsomemicroRNAs
downregulatelargenumbersoftargetmRNAs.Nature2005,433:769-773.
 41. KrutzfeldtJ,RajewskyN,BraichR,RajeevKG,TuschlT,ManoharanM,Stoffel
M:SilencingofmicroRNAsinvivowith'antagomirs'.Nature2005.
 24
 42. CheungVG,SpielmanRS,EwensKG,WeberTM,MorleyM,BurdickJT:

Mappingdeterminantsofhumangeneexpressionbyregionalandgenome-
wideassociation.Nature2005,437:1365-1369.
 43. HuminieckiL,WolfeKH:Divergenceofspatialgeneexpressionprofiles
followingspecies-specificgeneduplicationsinhumanandmouse.Genome
Res2004,14:1870-1879.
 44. MorleyM,MolonyCM,WeberTM,DevlinJL,EwensKG,SpielmanRS,
CheungVG:Geneticanalysisofgenome-widevariationinhumangene
expression.Nature2004,430:743-747.
 45. RifkinSA,HouleD,KimJ,WhiteKP:Amutationaccumulationassayreveals
abroadcapacityforrapidevolutionofgeneexpression.Nature2005,
438:220-223.
 46. RodwellGE,SonuR,ZahnJM,LundJ,WilhelmyJ,WangL,XiaoW,Mindrinos
M,CraneE,SegalEetal.:Atranscriptionalprofileofaginginthehuman
kidney.PLoSBiol2004,2:e427.
 47. StolcV,GauharZ,MasonC,HalaszG,vanBatenburgMF,RifkinSA,HuaS,
HerremanT,TongprasitW,BarbanoPEetal.:Ageneexpressionmapforthe
euchromaticgenomeofDrosophilamelanogaster.Science2004,306:655-660.
 48. StuartJM,SegalE,KollerD,KimSK:Agene-coexpressionnetworkforglobal
discoveryofconservedgeneticmodules.Science2003,302:249-255.
 49. ArbeitmanMN,FurlongEE,ImamF,JohnsonE,NullBH,BakerBS,Krasnow
MA,ScottMP,DavisRW,WhiteKP:Geneexpressionduringthelifecycleof
Drosophilamelanogaster.Science2002,297:2270-2275.
 50. JohnsonJM,CastleJ,Garrett-EngeleP,KanZ,LoerchPM,ArmourCD,Santos
R,SchadtEE,StoughtonR,ShoemakerDD:Genome-widesurveyofhuman
alternativepre-mRNAsplicingwithexonjunctionmicroarrays.Science2003,
302:2141-2144.
 51. ZhangW,MorrisQD,ChangR,ShaiO,BakowskiMA,MitsakakisN,
MohammadN,RobinsonMD,ZirngiblR,SomogyiEetal.:Thefunctional
landscapeofmousegeneexpression.JBiol2004,3:21.
 52. WienholdsE,KloostermanWP,MiskaE,Alvarez-SaavedraE,BerezikovE,de

BruijnE,HorvitzHR,KauppinenS,PlasterkRH:MicroRNAexpressionin
zebrafishembryonicdevelopment.Science2005,309:310-311.
 53. BabakT,ZhangW,MorrisQ,BlencoweBJ,HughesTR:ProbingmicroRNAs
withmicroarrays:tissuespecificityandfunctionalinference.RNA2004,
10:1813-1819.