Paul Lieberman and Kazuko Nishikura Takada, Andreas Wiedmer, Louise Showe, Artemis ...

14 Pages·2010·4.4 MB·English

RNA:
Editing of Epstein-Barr Virus-encoded
BART6 MicroRNAs Controls Their Dicer
Targeting and Consequently Affects Viral
Latency

Hisashi Iizasa, Bjorn-Erik Wulff, Nageswara
R. Alla, Manolis Maragkakis, Molly Megraw,
Artemis Hatzigeorgiou, Dai Iwakiri, Kenzo
Takada, Andreas Wiedmer, Louise Showe,
Paul Lieberman and Kazuko Nishikura
J. Biol. Chem. 2010, 285:33358-33370.
DD
doi: 10.1074/jbc.M110.138362 originally published online August 17, 2010 oo
ww
nn
lolo
aa
dd
ee
dd
fro fro
mm
h h
Access the most updated version of this article at doi: 10.1074/jbc.M110.138362 ttpttp
://w://w
ww
Find articles, minireviews, Reflections and Classics on similar topics on the JBC Affinity Sites. ww
.jb.jb
cc
.o.o
Alerts: rgrg
• When this article is cited b/ b/
yy
• When a correction for this article is posted g g
uu
eses
t ot o
Click here to choose from all of JBC's e-mail alerts nn
M M
aa
rcrc
hh
Supplemental material: 4 4
http://www.jbc.org/content/suppl/2010/08/30/M110.138362.DC1.html , 2, 2
00
1414
This article cites 68 references, 29 of which can be accessed free at
http://www.jbc.org/content/285/43/33358.full.html#ref-list-1
THEJOURNALOFBIOLOGICALCHEMISTRY VOL.285,NO.43,pp.33358–33370,October22,2010
©2010byTheAmericanSocietyforBiochemistryandMolecularBiology,Inc. PrintedintheU.S.A.
Editing of Epstein-Barr Virus-encoded BART6 MicroRNAs
Controls Their Dicer Targeting and Consequently Affects
Viral Latency*□S
Receivedforpublication,April26,2010,andinrevisedform,August16,2010Published,JBCPapersinPress,August17,2010,DOI10.1074/jbc.M110.138362
HisashiIizasa‡§,Bjorn-ErikWulff‡,NageswaraR.Alla‡,ManolisMaragkakis¶(cid:1),MollyMegraw**,
ArtemisHatzigeorgiou¶,DaiIwakiri§,KenzoTakada§,AndreasWiedmer‡,LouiseShowe‡,PaulLieberman‡,
andKazukoNishikura‡1
Fromthe‡TheWistarInstitute,Philadelphia,Pennsylvania19104,the§InstituteforGeneticMedicine,HokkaidoUniversity,
Sapporo060-0815,Japan,the¶InstituteofMolecularOncology,BiomedicalSciencesResearchCenterAlexanderFleming,16672
Vari-Athens,Greece,the(cid:1)InstituteofComputerScience,MartinLutherUniversityHalle-Wittenberg,06120Halle,Germany,andthe
**DepartmentofGenetics,SchoolofMedicine,UniversityofPennsylvania,Philadelphia,Pennsylvania19104
CertainprimarytranscriptsofmiRNA(pri-microRNAs)un- together with its partner DGCR8 (7, 8) cleaves pri-miRNAs,
dergo RNA editing that converts adenosine to inosine. The releasing60–70-nucleotidepre-miRNAs.Recognitionofcor-
Epstein-Barrvirus(EBV)genomeencodesmultiplemicroRNA rectly processed pre-miRNAs and their nuclear export is
genesofitsown.Herewereportthatprimarytranscriptsof carried out by exportin-5 and RanGTP (9). Cytoplasmic D
o
ebv-miR-BART6 (pri-miR-BART6) are edited in latently Dicertogetherwiththe double-stranded RNA (dsRNA)-bind- w
n
EBV-infected cells. Editing of wild-type pri-miR-BART6 RNAs ingproteinTRBPthencleavespre-miRNAsinto20–22-nucle- loa
d
dramaticallyreducedloadingofmiR-BART6-5pRNAsontothe otidesiRNA-likeduplexes(10,11).Inmostcasesonestrandof ed
microRNA-inducedsilencingcomplex.Editingofamutation- the duplex (called the effective strand) serves as the mature fro
m
containingpri-miR-BART6foundinDaudiBurkittlymphoma miRNA,whereastheotherstrand(calledpassengerstrand)is h
apnrodcensassionpghaorfynmgieRa-lBcAaRrcTin6omRNaACs6.66M-1osctelilmlpinoerstasnutplyp,remssieRd- einligmcionmatpedle.xA(fmteirRiInStCe)g,rmatiiRonNiAnstobltohcekmtriaRnNslAat-iionnduvciaedpasrilteianlcly- ttp://ww
BloAcaRtTed6-i5npthReN3A(cid:2)-sUTsiRlenofceDiDceicremrRthNrAou.TghhemsiuglntiipfilceantacregoeftmsiitRes- cgoetmedplmemReNnAtasroyrbgiuniddiengthseitdeesglroacdaatteidoninoftthaerg3e(cid:1)t-UmTRRNsAosfatfaterr- w.jbc.o
BlaAteRnTcy6.wWaseffuorutnhdertihnavtesmtiigRa-tBedARinT6c-e5llps iRnNvAasriosuusppsrteasgsesthoef bthined“isnege,dmseaqinuleynvciea”t(h1e,45,(cid:1)5h).alfofthemiRNAsequence,called by grg/
EBNA2 viral oncogene required for transition from immuno- u
Epstein-Barr Virus (EBV) causes mononucleosis during e
s
logicallylessresponsivetypeIandtypeIIlatencytothemore acuteandlyticinfectionandalsoestablishesapersistentand t o
n
immunoreactivetypeIIIlatencyaswellasZtaandRtaviralpro- latentinfectioninthehumanhost.LatentlyinfectedEBVhas M
a
teins essential for lytic replication, revealing the regulatory beendemonstratedtobeassociatedwithavarietyofhuman rch
functionofmiR-BART6inEBVinfectionandlatency.Mutation 4
cancers such as Burkitt lymphoma, Hodgkin disease, and , 2
andA-to-Ieditingappeartobeadaptivemechanismsthatantag- 0
nasopharyngealcarcinoma(12,13).Lyticinfectionandtran- 1
4
onizemiR-BART6activities.
sitiontodistinctivestatesoflatency(typeI-III)areregulated
by select viral genes and their interaction with the host im-
munesystem(12,13).VirusgenomesencodemiRNAsoftheir
MicroRNAs (miRNAs)2 play important roles in many pro-
own,andthefirstviralmiRNAwasidentifiedinhumanBcells
cesses including development, differentiation, proliferation,
infected with EBV (14). A total of 23 EBV miRNA genes are
andapoptosis(1,2).CertainmiRNAsactastumorsuppressors
knownandlocatedintheBHRF1andBART(BamH1Aright-
oroncogenesandareassociatedwithmanycancers(3).Primary
ward transcript) regions of the genome (15–17). These EBV
transcripts of miRNA genes (pri-miRNAs) are processed se-
miRNAs have been implicated in regulating the transition
quentially by Drosha and Dicer (4, 5). Nuclear Drosha (6)
from lytic replication to latent infection and in attenuating
antiviral immune responses (18). However, only a limited
*Thisworkwassupported,inwholeorinpart,byNationalInstitutesofHealth numberoftheirtargetshavebeenidentifiedsofar.Theviral
Grants CA010815, GM040536, and HL099342. This work was also sup- miRNAsseemtotargetbothviralandhostcellgenes(18).For
portedbytheEllisonMedicalFoundation(AG-SS-2281-09)andtheCom-
instance,miR-BART2targetstheEBVDNApolymerase,BALF5,
monwealth Universal Research Enhancement Program, Pennsylvania
DepartmentofHealth(toK.N.). perhaps promoting entry of the virus to latency by slowing
□S Theon-lineversionofthisarticle(availableathttp://www.jbc.org)contains downviralreplicationatthetransitionpointfromlytictolatent
supplementalExperimentalProcedures,Table1,andFigs.1–6.
1Towhomcorrespondenceshouldbeaddressed:TheWistarInstitute,3601 infection (19). Down-regulation of the EBV protein LMP1 by
SpruceSt.,Philadelphia,PA19104-4268;Tel.:[phone];Fax:215-898- threeEBVmiRNAs,miR-BART1–5p,miR-BART16,andmiR-
3911;E-mail:[email protected]. BART17–5p, has been reported (20). LMP1 produced during
2Theabbreviationsusedare:miRNA,microRNA;ADAR,adenosinedeami- the EBV type II and III latency controls the NF-(cid:1)(cid:2)signaling
nases acting on RNA; miRISC, miRNA-induced silencing complex; EBV,
Epstein-Barrvirus;qRT,quantitativereal-time. pathwayandgrowthandapoptosisofhostcells.Targetingof
33358 JOURNALOFBIOLOGICALCHEMISTRY VOLUME285•NUMBER43•OCTOBER22,2010
ControlofEBVLatencybyBART6miRNAEditing
hostcellgenesPUMA(p53-up-regulatedmodulatorofapopto- ingandmutationmaybecriticalfortheestablishmentormain-
sis) by miR-BART5 (21) and CXC-chemokine ligand 11 tenanceoflatentEBVinfection.
(CXCL11)bymiR-BHRF1–3(22)havebeenreported.Down-
EXPERIMENTALPROCEDURES
regulationofPUMAmaysuppressapoptosisofvirus-infected
host cells (21), whereas suppression of CXCL11 may shield CellCulture—EBV-transformedlymphoblastoidcelllineGM607
EBV-infectedBcellsfromcytotoxicTcells(22). (GM00607) was obtained from Coriell Institute for Medical
OnetypeofRNAeditinginvolvestheconversionofadeno- Research(Camden,NJ).BurkittlymphomacelllineDaudiwas
sine residues into inosine (A-to-I editing) in dsRNA through obtainedfromAmericanTypeCultureCollection(Manassas,
the action of adenosine deaminase acting on RNA (ADAR). VA).BurkittlymphomaMutuIandMutuIIIandnasopharyn-
Three ADAR gene family members (ADAR1–3) have been gealcarcinomalineC666-1wereusedinourpreviousstudies
identified in humans and rodents (23, 24). The translation (39–41).ThesecelllineswereculturedinRPMI1640(Media-
machineryreadsaninosineasifitwereguanosine,whichcould techInc.,Manassas,VA),supplementedwith100units/mlben-
lead to codon changes (25). Thus, when A-to-I RNA editing zylpenicillin, 100 (cid:3)g/ml of streptomycin sulfate (both from
occurs within a coding sequence, synthesis of proteins not Invitrogen)and10%fetalcalfserum(FCS)(TissueCultureBio-
directly encoded by the genome can result, as demonstrated logical,Tulare,CA).HeLaandHEK293Tcellswereculturedin
with transcripts of glutamate receptor ion channels and 5- Dulbecco’s modified Eagle’s medium (Invitrogen) supple-
HT serotonin receptors (26). However, the most common mentedwith10%FCS.
2C
targets for A-to-I editing are non-coding RNAs that contain AnalysisofinVitroProcessedPri-miRNAProductsbyNorth-
inverted repeats of repetitive elements such as Alu elements ern Blotting—Nonradioactive pri-miR-BART6 RNAs (10
and LINEs located within introns and 3(cid:1)-UTRs (27–30). The fmol) were synthesized by in vitro transcription and pro-
D
biologicalsignificanceofnon-coding,repetitiveRNAeditingis cessedbyDrosha-DGCR8(20ng)and/orDicer-TRBPcom- ow
n
largelyunknown.Furthermore,editingofcertainpri-miRNAs plexes(20ng)asdescribedpreviously(34).ProcessedRNAs lo
a
has been reported (31, 32). A recent survey has revealed that were electrophoresed on a 15% polyacrylamide, 8 M urea de
d
(cid:2)20%ofhumanpri-miRNAsaresubjecttoA-to-IRNAediting gelandtransferredtoaHybondXLmembrane(GEHealth- fro
catalyzedbyADAR1andADAR2(33).Editingofpri-miRNAs care) by electroblotting. Membranes were UV-cross-linked m
h
modulates expression and function of miRNAs (33). For in- (StrataLinker; Stratagene, La Jolla, CA), and hybridized with ttp
stance,A-to-Ieditingofseveraladenosineresidueslocatednear 5(cid:1)-32P-labeledmiRCURYlockednucleicacidprobes(Exiqon ://w
w
theDroshacleavagesitesofpri-miRNA-142resultsininhibi- Inc.,Woburn,MA)andanalyzedbyNorthernblotting.The w
tionoftheprocessingbyDroshaandconsequentdown-regula- hybridizationbuffercontained50%formamide,0.5%SDS,5(cid:3) .jb
c
tionofmaturemiR-142RNAs(34),whereaseditingoftwosites saline/sodiumphosphate/EDTA,5(cid:3)Denhardt’ssolution,and .org
identifiedneartheendloopofthepri-miR-151hairpinstruc- 20(cid:3)g/mlsheared,denatured,salmonspermDNA.Hybridiza- b/
y
tureinhibitstheDicercleavagestep(35).Bycontrast,editingof tionwasconductedat34°C.Membraneswerewashedby2(cid:3) g
u
e
primarytranscriptsofthemiR-376clusterattwositeslocated SSC/0.1% SDS, and hybridized signals were quantified by a st o
withintheseedsequencedoesnotaffecttheirprocessingbut TyphoonImagerSystem. n M
results in expression of mature-edited miR-376 RNAs with Luciferase Reporter Constructs—The human Dicer 3(cid:1)-UTR a
rc
altered seed sequences and consequent silencing of a set of (1498 bp), which contains four miR-BART6-5p binding sites h 4
genesdifferentfromthosetargetedbyuneditedmiR-376RNAs (supplemental Experimental Procedures), were amplified us- , 2
0
1
(36). ing human genomic DNA extracted from GM607 cells and 4
InthisstudywesetouttoexamineeditingofEBVmiRNAsin specific primers hDicerFW (5(cid:1)-GCTACTAGTGATCTTT-
EBV-transformed lymphoblastoid GM607 cells, Burkitt lym- GGCTAAACACCCCAT-3(cid:1)) and hDicerRV (5(cid:1)-GCTGTT-
phoma Daudi cells, and nasopharyngeal carcinoma C666-1 TAAACCTCCAACAAAAAGTGAAACGGC-3(cid:1)). The PCR
cells.HumanlymphoblastoidcellssuchasGM607cellsintype products were inserted into a luciferase reporter vector
IIIlatencyexpressasetofgenesessentialforthisspecificstate (pMIR-REPORTTM Luciferase; Ambion) after digestion with
oflatency,suchasEBNA2andLMP1.Bycontrast,DaudiBur- Spe1andPml1.
kittlymphomacellsintherestrictedsub-typeoftypeIIIlatency Transfections of miR-BART6 RNAs—miR-BART6-5p and
do not express EBNA2 due to the genomic deletion (37, 38). uneditedmiR-BART6–3pRNAduplexesweresynthesizedat
Viral infection in C666-1 nasopharyngeal carcinoma cells is Ambion (Pre-miRTM miRNA). All transfections were carried
associatedwithmorerestrictedformsoftypeIIlatency,which out in triplicate as described previously (42). Briefly, HeLa
expressesonlyalimitednumberofviralgenes,representinga cellswerepre-platedin24-welltissuecultureplates.200ngof
lessimmune-responsivestate(38).Wehavefoundthatprimary luciferasereporterplasmidand200ngofcontrolvectorpMIR-
transcripts of four EBV miRNAs, including miR-BART6, are REPORTTM (cid:2)-galactosidase Control Plasmid (Ambion) were
subjecttoA-to-Iediting.Moreover,wedemonstratethatedit- dilutedinto50(cid:3)lofOpti-MEM(Invitrogen)withorwithout10
ing of pri-miR BART6 RNAs as well as mutations of miR- pmol of miR-BART6 duplex or sequence unrelated control
BART6 RNAs found in latently EBV-infected cells inhibits miRNA,cel-miR-67,ormiR-376afollowedbytheadditionof3
expressionortheirloadingontothefunctionallyactivemiRISC. (cid:3)l of Lipofectamine 2000 (Invitrogen). The transfection mix-
Most significantly, we found that miR-BART6 targets Dicer turewasincubatedatroomtemperaturefor5minfollowedby
andaffectsthelatentstateofEBVviralinfection.Regulationof the addition of DNA/miR-BART6 and further incubated at
themiR-BART6expressionandfunctionthroughA-to-Iedit- roomtemperaturefor20min.Then100(cid:3)loftransfectionmix-
OCTOBER22,2010•VOLUME285•NUMBER43 JOURNALOFBIOLOGICALCHEMISTRY 33359
ControlofEBVLatencybyBART6miRNAEditing
ture was added to the HeLa cells in 500 (cid:3)l of the growth for pri-miR-BART6 expression with 2 (cid:3)g/ml doxycycline
medium. Transfection efficiency monitored by using 5-car- (Sigma).Transfectionefficiencyandexpressionofpri-miRNA
boxylfluorescein-labeled control siRNA (Ambion) was more were determined by turboRFP expression. Protein and total
than 80%. Forty-eight hours after transfection, the luciferase RNA were extracted 48 h after DOX induction. Levels of
activitywasmeasuredusingtheLuciferaseAssaySystem(Pro- maturemiR-BART6-5pwereexaminedbydideoxyoligonucle-
mega,Madison,WI)togetherwith(cid:2)-galactosidaseactivityby otide/primer-extensionassay.
reading the absorbance at 415 nm in a plate reader with the miRISCLoadingAssay—Thetargetprobeswere5(cid:1)-end32P-
(cid:2)-GalactosidaseEnzymeAssaySystem(Promega).Normalized labeledwithT4polynucleotidekinase(NewEnglandBiolabs)
luciferase values divided by the (cid:2)-galactosidase activity were and [(cid:4)-32P]ATP. 5 fmol of 32P-labeled miR-BART6-5p target
statisticallycomparedamongeachgroupbyMann-WhitneyU RNA (5(cid:1)-AACCUACUAUGGAUUGGACCAACCUUACCA-
test. AG-3(cid:1)),BART6–3P-uneditedtarget(5(cid:1)-AACCUAAGCUAA-
Transfections of the miR-BART6-5p Antagomir—C666.1 GGCUAGUCCGAUCCCGCCAAG-3(cid:1)), BART6–3P-edited
or Mutu I cells (1.5 (cid:3) 105 cells) were cultured in 24-well target(5(cid:1)-AACCUAGCCAAGGCUAGUCCGAUCCCCGCC-
plates. The next day cells were transfected with 50 pmol of AAG-3(cid:1)),andpre-miR-BART6RNAs,whichhadbeencleaved
inhibitor of miR-BART6-5p (miScript miRNA inhibitor, from pri-miR-BART6 RNAs with Drosha-DGCR8 and gel-
Qiagen)orAllStarsNegativeControlsiRNA(Qiagen)using3 purified, were incubated with FLAG-tagged Ago2-complex
(cid:3)lofHiperfectTransfectionreagent(Qiagen).After72h,total madefrompermanentlytransfectedHEK293cellsinareac-
RNA was extracted and treated with DNase I. First-strand tion buffer containing 1 unit/(cid:3)l RNasin, 20 mM Tris-HCl
cDNAwassynthesizedusing1(cid:3)goftotalRNAusingmiScript (pH7.6),0.1MNaCl,10%glycerol,2mMDTT,0.2mMPMSF,
Reversetranscriptionkit(Qiagen)orSuperscriptIIIwithran- 1 mM (cid:2)-mercaptoethanol, 3.2 mM MgCl2, 1 mM ATP, 20 mM D
domprimer. creatinephosphate,and1units/(cid:3)lcreatinekinaseat37°Cfor ow
n
TransfectionsoftheDicerTargetingshRNAExpressionVector— 90minasdescribedpreviously(43,44).miRISCloadingprod- lo
a
TosuppressDicerexpression,ashorthairpinexpressionvector ucts (32P-labeled cleaved target RNAs) were electrophoresed de
d
wasused.UsingBLOCK-iTRNAiDesigner(Invitrogen),com- on a 15% polyacrylamide, 8 M urea gel, and quantified by fro
plementaryDNAoligosweredesigned.Forconstructionof TyphoonImager. m
h
DicershRNAplasmids,sense(5(cid:1)-CACCGCAGCTCTGGA- ttp
TCATAATACCCGAAGGTATTATGATCCAGAGCTGC- RESULTS ://w
3(cid:1)) and antisense (5(cid:1)-AAAAGCAGCTCTGGATCATAATA- A-to-I Editing Sites and Mutations Found in EBV Pri- ww
CCTTCGGGTATTATGATCCAGAGCTGC-3(cid:1)) strand oligos miRNAs—We examined the primary transcripts of all 23 .jb
c
weresynthesized.ForconstructionofLacZ2.1Control,sense EBVmiRNAsforA-to-IRNAeditinginlatentlyEBV-infected .org
(5(cid:1)-CACCAAATCGCTGATTTGTGTAGTCGGAGACGAC- human lymphoblastoid GM607, Daudi Burkitt lymphoma, b/
y
TACACAAATCAGCGA-3(cid:1))andantisense(5(cid:1)-AAAATCGC- and C666-1 nasopharyngeal carcinoma cells. We found that g
u
e
TGATTTGTGTAGTCGTCTCCGACTACACAAATCAGC- pri-miR-BHRF1–1, pri-miR-BART6, pri-miR-BART8, and st o
GATTT-3(cid:1)) strand oligos were synthesized. To generate a pri-miR-BART16 are edited at specific sites (Fig. 1A and n M
double-strandedDNA,theseoligoswereannealedandcloned supplementalFig.1A).Althoughtheeditingfrequenciesofpri- a
rc
intopENTER/H1/TOvector(Invitrogen).C666.1orMutuIII miR-BHRF1–1, pri-miR-BART8, and pri-miR-BART16 were h 4
cells(1(cid:3)106)weretransfectedwith1(cid:3)gofvectorDNAusing relatively low (supplemental Fig. 1B), editing of pri-miR- , 2
0
CUY21Pro-Vitro (NEPA GENE., Co Ltd, Ichikawa, Japan). BART6inDaudiandGM607cellsatthe(cid:4)20sitereached50 14
After48h,totalRNAwasextractedandtreatedwithDNaseI. and70%,respectively(Fig.1B).Lowlevelsofeditingofpri-miR-
First-strand cDNA was synthesized using 1 (cid:3)g of total RNA BART6RNAswerealsodetectedinC666-1cells(Fig.1B).We
usingthemiScriptreversetranscriptionkit(Qiagen)orSuper- foundthatthesizeoftheendloopandtheterminalstemofthe
script III with random primers. Transfection efficacy moni- pri-miR-BART6ofDaudiissmallerthanthatofGM607cells
tored by co-transfection of ptdTomato-C1 vector (Clontech) (wild-type) due to deletion of three uridine nucleotides (Fig.
was(cid:2)70–80%. 1A). The same deletion was detected in C666-1 cells, as re-
Induction of Viral miRNA Expression in HEK293T Cells— portedpreviously(16).
The pri-miR-BART6 sequences were PCR-amplified using Because involvement of enzymatically active ADAR1 and
genomicDNAextractedfromGM607cellsandasetofprimers, ADAR2intheRNAeditingmechanismhasbeenestablished(23,
LentiBART6FW(5(cid:1)-GCCTCGAGTGACCTTGTTGGTACT- 24,45),weexaminedtheexpressionofADAR1andADAR2in
TTAAGGTTG-3(cid:1))andLentiBART6-UneditedRV(5(cid:1)-GCGA- GM607, Daudi, and C666-1 cells by Western blotting analysis.
ATTCTGGCCTTGAGTTACTCTAAGGCTA-3(cid:1))containing Although no ADAR2 was detected, abundant expression of
athymidineresidueatthe(cid:4)20siteorLentiBART6-EditedRV ADAR1 (both interferon-inducible p150 and constitutive p110
(5(cid:1)-GCGAATTCTGGCCTTGAGTTACTCCAAGGCTA-3(cid:1)) isoforms)(46)wasfoundinallthreecelllines(supplementalFig.2),
containing a cytidine residue (edited) at the (cid:4)20 site. These indicating that ADAR1 is likely to be responsible for editing of
PCRproductsweredigestedwithXhoIandEcoRI(NewEng- pri-miR-BART6.However,wecannotexcludethepossibilitythat
land Biolabs, Ipswich, MA) and ligated into pTRIPZ vector ADAR2maybealsoabletoeditthissite.
(Open Biosystems, Huntsville, AL). pTRIPZ-derived lentivi- Processing of Pri-miR-BART6 Is Affected by Editing and Mu-
rusesweretransfectedintoHEK293Tinthepresenceofpuro- tation—Manysinglenucleotidepolymorphisms(SNPs)foundin
mycin(Sigma).Permanentlytransfectedcelllineswereinduced humanmiRNAgenesaffectbiogenesisandfunction,suggesting
33360 JOURNALOFBIOLOGICALCHEMISTRY VOLUME285•NUMBER43•OCTOBER22,2010
ControlofEBVLatencybyBART6miRNAEditing
invitroDicerand/orDroshacleav-
age assay products were also ana-
lyzed by Northern blotting using
5p- or 3p strand-specific oligonu-
cleotide probes (Fig. 2C). The effi-
cient conversion of both unedited
and edited wild-type (GM607) pri-
miR-BART6 to pre-miR-BART6
and mature miR-BART6 was de-
tected, indicating that the editing
of wild-type pri-miR-BART6 at
the (cid:4)20 site has no inhibitory
effect on Drosha and Dicer cleav-
age(Fig.2,AandB).Generationof
both 5p and 3p mature miRNAs
fromuneditedandeditedwild-type
pri-miR-BART6 was confirmed by
Northern blotting analysis using
strand-specific probes (Fig. 2C).
Similarly,uneditedDaudi(C666-1)
D
pri-miR-BART6 RNAs were pro- ow
n
cessedtopre-andmaturemiRNAs, lo
a
although Dicer cleavage efficiency de
d
wasreducedto(cid:2)70%oftheuned- fro
itedwild-typelevel,likelyduetothe m
h
deletionofthreeUresidues(Fig.2, ttp
BandC).However,Droshacleavage ://w
w
of edited Daudi (C666-1) pri-miR- w
BART6 was completely blocked .jb
c
(Fig.2,AandC).BindingofDGCR8 .org
toDaudi(C666-1)pre-miR-BART6 b/
y
FIGURE1.A-to-IRNAeditingofpri-miR-BART6RNAs.A,shownarehairpinstructuresofpri-miR-BART6.Two seemedtobeunaffectedbyediting, g
differenthairpinstructuresofpri-miR-BART6(partial),thewild-typefromGM607cellsandamutantfoundin ue
DbceyasuasdendiuBmiunrbtkoeitrtthwaenitdhmCtah6tue6r65e-(cid:1)1mecneiRdllNso,Aafsrteh(5esphmoswaetnnus.reTehameniedRd-3BitpAinRagnTs6tiits–ee3napdsseeensqotursaeinnnedcse((cid:4))ca2or0eunshittieegd)h,lhaigsigh(cid:4)htel1igd.hTihtneedgrrieengernieo.dnM,sisattionudbreiecmaptrieoRd-- amsosbeeilnityfrosmhifatseatssoafyele(EctMroSpAh)orgeesilss st on M
(supplemental Fig. 3). The nearly a
BART6-5pandbothuneditedandedited-3pRNAsarealsoshown.ThreedeletedUnucleotidesareindicated rc
idnebrilvaecdkfbrooxmesGwMit6h0i7n,Dthaeuwdii,ldan-tdypCe66h6a-i1rppinri-smtriuRc-tBuArReT.B6,RDNNAAsasereqsuheonwcinn.gThcherRoNmAaetodgitrianmgssitoef(R(cid:4)T2-P0C)iRsdperotedcutcetds ibdinendtiincgaltoKudnevdailtueedsan((cid:2)d5edintMed)pfroir- h 4, 2
asanA-to-GchangeinthecDNAsequencingchromatogramasindicatedbyredarrows.ThreeTnucleotides, 01
deletedinpri-miR-BART6fromDaudiandC666-1cells,areindicated.Editingfrequencywasestimatedasa miR-BART6 RNAs were estimated 4
percentageestimatedfromtheratioofGpeakoverthesumofGandApeaksofthesequencingchromato- fromanalysisofseveralEMSAgels.
gram.Twoseparatemeasurementsweredone,andidenticalresultswereobtained.
Thus,acombinationofthedeletion
of three U residues and editing at
thattheymaybeassociatedwithdiseases(47).Sequencevari- the (cid:4)20 site appears to inhibit Drosha cleavage of pri-miR-
ations in several EBV pri-miRNAs have also been reported BART6RNAs.
(16),butthesignificanceofmostofthesemutationshasnot TargetingofDicerbymiR-BART6—Theinhibitoryeffectsof
beenevaluated.Wereasonedthateditingatthe(cid:4)20siteand mutationandA-to-Ieditingonprocessingofpri-miR-BART6
themutationsfoundinDaudiandC666-1cellsmayaffectthe RNAsintomaturemiRNAsmayindicatethatthisviralmiRNA
biogenesis of pri-miR-BART6 RNAs. An in vitro pri-miRNA plays a role in regulating EBV infection state in Daudi
processing assay using recombinant Drosha-DGCR8 and and C666-1 cells. For instance, suppression of miR-BART6
Dicer-TRBP complexes (33, 34, 36) was conducted with uni- RNAsmaybenecessaryforEBVtoremainataspecificstate
formly32P-labeleduneditedand“edited”wild-typeandDaudi oflatency.CertainviralmiRNAshavebeenshowntotarget
(C666-1) pri-miR-BART6 RNAs, which were prepared by in genesofthehostcellaswellasgenesofthevirusitself(18).
vitrotranscription.Theeditedpri-miRNAshadanA-to-Gsub- UsingtheDIANA-microTprogram(48)we,therefore,pre-
stitution at the (cid:4)20 site. We had previously shown that the dicted in silico human and EBV target genes for miR-
miRNAprocessingmachineryrecognizesA-to-Gsubstitutions BART6-5pandmiR-BART6–3p(bothuneditedandeditediso-
ofpri-miRNAsasiftheywereA-to-Ichanges(34).Theradio- forms). The candidate target genes were pruned by a species
activepri-,pre-,andmaturemiRNAproductswerequantita- conservationfilterandalsobyacceptingonlygenesthathave
tivelyanalyzedafterfractionationonapolyacrylamidegel(Fig. multipletargetsiteswithinthe3(cid:1)-UTRs.Wefoundnostrong
2,AandB).Inaddition,nonradioactivepri-miRNAsandtheir target gene candidates containing multiple binding sites for
OCTOBER22,2010•VOLUME285•NUMBER43 JOURNALOFBIOLOGICALCHEMISTRY 33361
ControlofEBVLatencybyBART6miRNAEditing
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
://w
w
w
.jb
c
.o
rg
b/
y
g
u
e
s
t o
n
M
a
rc
h
4
, 2
FIGURE2.Invitroprocessingofpri-miR-BART6RNAsbymiRNAprocessorcomplexes.A,effectofeditingonDroshacleavageofwild-typeandmutant 01
pri-miR-BART6RNAswastestedwithuniformly32P-labeledpri-miR-BART6RNAs.Themutantpri-miR-BART6sequencesofDaudiandC666-1areidentical.Thus, 4
itisindicatedasDaudiorC666-1.Uneditedoreditedpri-miR-BART6RNAs(i.e.containinganA-to-Gsubstitutionatthe(cid:4)20site)wassubjectedtotheDrosha
cleavagereactionusingDrosha-DGCR8complex.B,effectofeditingonDicercleavageisshown.TheDrosha-DGCR8reactionproductsweresubjectedtothe
DicercleavagereactionusingtheDicer-TRBPcomplex.AandB,threeindependentassaysweredone.DifferencesanalyzedbyMann-WhitneyUtest:**,p(cid:5)
0.005;***,p(cid:5)0.001.Errorbars,S.E.(n(cid:6)3).C,NorthernblottinganalysisofinvitroprocessedmiR-BART6RNAsisshown.Nonradioactivepri-miR-BART6RNAs
processedinvitrobyDrosha-DGCR8and/orDicer-TRBPcomplexeswereanalyzedbyNorthernblottingusinga32P-labeled5p-or3p-strandspecificoligo
probe.Representativeresultsforuneditedandeditedpri-miR-BART6RNAsofwild-type(GM607)andmutant(Daudi)areshown.
miR-BART6-5por-3pRNAsintheEBVgenome.Bycontrast, 3, A and B, and supplemental Experimental Procedures).
thescreeningidentified14humanstrongcandidatesformiR- Althoughalimitedconservationofsome5psitesforelephant
BART6-5pand3humantargetsformiR-BART6–3pregardless (site1andsite4)orarmadillo(site1andsite2)wasfound,all
of whether miR-BART6–3p RNAs were edited or unedited foursitesidentifiedwereotherwiseuniquetothehumanDicer
(supplementalTable1).Becausethe(cid:4)20editingsiteofmiR- 3(cid:1)-UTR and not evolutionarily conserved even for the chim-
BART6–3pislocatedoutsideoftheseedsequence(Fig.1A),it panzeeDicer3(cid:1)-UTR(datanotshown).Thisisunusualforhigh
wasanticipatedthateditingshouldnotseverelyaffecttheselec- score targets, which often have better species conservation,
tionoftargetgenes. supporting their biological significance. In light of the fact
Most interestingly, Dicer was one of the high-score targets that EBV specifically infects human, it is possible that miR-
for miR-BART6-5p (supplemental Table 1). Because of its BART6-5pevolvedtotargetDicerspecificallyinhumanduring
importance and global effects on many genes via RNAi, we theestablishmentoftheEBV-hostrelationship.
decided to further investigate the targeting of Dicer by miR- In vitro validation experiments were conducted in HeLa
BART6-5p RNAs. Four target binding sites were identified cells(thesecellsareEBV-negativeand,thus,lackpre-exist-
withinthe(cid:2)1.5-kbregionofhumanDicermRNA3(cid:1)-UTR(Fig. ing miR-BART6 RNAs) cotransfected with a luciferase
33362 JOURNALOFBIOLOGICALCHEMISTRY VOLUME285•NUMBER43•OCTOBER22,2010
ControlofEBVLatencybyBART6miRNAEditing
Dicer mRNA 3(cid:1)-UTR sites are in-
deed target sites of miR-BART6-5p
RNAs. To validate the targeting of
the Dicer mRNA by miR-BART6
RNAs in vivo, we then measured
endogenousexpressionlevelsofDicer
inHeLacells.Wefoundasubstan-
tial reduction in the Dicer levels
(3.5-fold) in HeLa cells transfected
with miR-BART6-5p but not with
control cel-miR-67 (Fig. 4A), con-
firminginvivotargetingofDicerby
miR-BART6-5pRNAs.
SuppressionofmiRISCLoadingof
miR-BART6-5p RNAs by Editing—
Tofurtherconfirmtheinvivosilenc-
ing of Dicer by miR-BART6 RNAs,
we prepared two tetracycline-in-
ducible pri-miR-BART6 RNA ex-
pressionconstructsinalentivirus
D
vectorsystem;oneexpressingun- ow
n
editedwild-typepri-miR-BART6and lo
a
theotherexpressingtheeditedpri- de
d
miR-BART6 containing an A-to-G fro
substitutionatthe(cid:4)20editingsite. m
h
HEK293 cells (also EBV-negative ttp
and,thus,lackingpre-existingmiR- ://w
w
BART6-5p RNAs) were infected w
with the lentiviral constructs and .jb
c
subjected to conditional induction .org
ofpri-miR-BART6andconsequent b/
y
mature miR-BART6 RNAs. Very g
u
e
low editing activities have been st o
reported in HEK293 cells (49), and n M
we confirmed that the pri-miR- a
rc
BART6RNAsderivedfromtheun- h 4
FIGURE3.TargetsitesformiR-BART6-5pRNAsidentifiedinthe3(cid:2)-UTRofhumanDicermRNA.A,the edited pri-miRNA expression con- , 20
locationsoffourmiR-BART6-5ptargetsiteslocatedwithinthe3(cid:1)-UTRofhumanDicermRNAareschematically structwerebarelyedited((cid:5)5%,data 14
presented.B,RNAduplexformationbetweentheDicer3(cid:1)-UTRtargetsitesandmiR-BART6-5pRNAsaredia- not shown). Dicer levels were re-
grammed.C,shownisadiagramoftheluciferasereporterplasmidcontainingthefour5pstrandtargetsites.
D,relativeluciferaseactivitiesinHeLacellscotransfectedwiththereportervectorcontaining4(cid:3)5psitesare duced by 70% in HEK293 cells
shown.Twocontrols,thevector-onlytransfection,andcotransfectionwiththeunrelatedsequenceC.elegans infectedwiththeuneditedpri-miR-
miR-67wereconducted.Expressionlevelsoftheluciferasereportergenewerenormalizedbyexpressionlevels
BART6 construct compared with
ofacotransfected(cid:2)-galactosidasereportergene.Threeindependentassayswereconducted.Theluciferase
activitieswerecomparedstatisticallybyMann-WhitneyUtests.Significantdifferencesbetweenvectoronly the vector control (Fig. 4B). The
andmiR-BART6-5por-3pcotransfectedexperimentsareindicatedbyasterisks;***,p(cid:5)0.001.Errorbars,S.E. reduction in the Dicer levels was
(n(cid:6)3).
also detected in HEK293 cells in-
fected with the edited pri-miR-
reporter construct containing the Dicer 3(cid:1)-UTR including BART6 construct. However, the extent of suppression was
thefourtargetsitesofmiR-BART6-5p(reporter-4(cid:3)5p)(Fig. muchless,by25%.Thismayindicatethateditingofthewild-
3C).Theluciferaseexpressionlevelswereclearlysuppressedby typepri-miR-BART6RNAhasnegativeeffectsontheinvivo
miR-BART6-5p (4-fold) but not by miR-BART6–3p in HeLa expressionorfunctionsofmiR-BART6-5pRNAs,althoughno
cellsthatwerecotransfectedwiththe4(cid:3)5pvector(Fig.3D). difference was noted between unedited and edited pri-miR-
Foranunknownreason,ournegativecontrolCaenorhabditis BART6ofwild-type(GM607)ininvitropri-miRNAprocess-
elegans miR-67 RNAs unexpectedly increased the luciferase ing(Fig.2).NosignificantdifferenceinmiR-BART6-5pRNA
expression(Fig.3D).Thisisnotduetononspecificexhaustion levelswasdetectedbetweenHEK293cellsinfectedwiththe
of the miRNA-mediated silencing machinery by this control unedited and edited pri-miR-BART6 expression constructs
miRNA, as the other sequence-unrelated control miRNA, (supplemental Fig. 4), indicating that the stability and/or
humanmiR-376a-5p,hadnoeffectontheluciferaseexpression turnoverofthematuremiR-BART6-5pRNAsisunlikelyto
(data not shown). The results strongly indicate that the four beaffectedbyediting.
OCTOBER22,2010•VOLUME285•NUMBER43 JOURNALOFBIOLOGICALCHEMISTRY 33363
ControlofEBVLatencybyBART6miRNAEditing
BART6 RNAs. This is the first
example of A-to-I editing of a pri-
miRNAaffectingmiRISCloading.
Suppression of Many miRNAs by
miR-BART6-5p—The dramatic re-
duction in the Dicer expression
mediated by miR-BART6-5p (Fig.
4) suggests that it may affect the
biogenesisofmiRNAsglobally.We,
therefore, examined the effects
ofDicersuppressiononexpression
of other miRNAs by miRNA array
analysis. The miRNA levels were
examined in HeLa cells with sub-
stantiallyreducedDicerlevelsafter
transfection with miR-BART6-5p
RNAs (Fig. 4A). Once again, HeLa
cells were used because of the ab-
sence of preexisting miR-BART6
RNAs.Thisstudyrevealedthatlev-
D
elsofatleast69miRNAsweresig- ow
n
nificantly reduced, and 14 of these lo
a
miRNAs showed more than 2-fold de
d
suppression(supplementalFig.5A). fro
Synthesis of these miRNAs may m
h
beparticularlysensitivetotheDicer ttp
concentration.Interestingly,theex- ://w
w
FHIeGLUaRcEel4ls.RtreapnrsefescstieodnwoifthDimceiRr-bByARmTi6R--5BpARRNTA6s-5ispshRoNwAns..TAw,WoceosntetrrnolbelxoptearnimalyesnitssowfeDriececroenxdpurcetsesdio;nHeleLvaeclselilns p19re6sbs-i5opn, omfiRth-r2e0e5–m5piR,NaAnsd, mmiiRR-- w.jbc
(cid:2)Bw,-iatshchtooiunwtlnetrviaesnlassfweWceetriseotnaelrsonorbtmrloaontnsaiftneocartleyedsd.isAwositfuhDmaimcseearqryeuxegpnrracepesh-suioonfnrneilonartmHedEaKlCi2z.e9ed3leTDgacicneelslrsmeixniRpf-er6ec7stse.iAdosnwalietnvhoerilnmsdiasulaiczlisabotlieopnrleecnsoetinnvttirreoudls., 6in2c4r–ea5sped(s.uAplpthleomugehntwaleFdigo.n5oBt),hwavaes by.org/
vectorsforexpressionofuneditedoredited(A-to-Gsubstitutionatthe(cid:4)20site)pri-miR-BART6RNAs.Expres- aconfirmedexplanationforup-reg- gu
scioonntrooflpvrei-cmtoirRd-BirAeRcTts6tRhNeAexspwraesssiinodnuocfendowni-tshile2n(cid:3)cgin/gmvledroifxieydcyncelignaeti(vDeOsXiR).NInAtsh(OepperensBeniocseysotfedmosx)y.Acysculimnem,tahrey ulationofthesethreemiRNAs,one est o
graphofnormalizedDicerexpressionlevelsisalsoshown.AandB,significantdifferenceswereanalyzedby possibilityisthatthegenesregulat- n M
Mann-WhitneyUtests:*,p(cid:5)0.05;**,p(cid:5)0.005;***,p(cid:5)0.001.Errorbars,S.E.(n(cid:6)3). ingtheexpressionofthesemiRNAs a
rc
may be controlled negatively by h 4
Asoneoftheremainingpossibilities,wethoughtthatedit- other miRNAs whose levels are reduced. Our results demon- , 2
0
1
ing might affect the selection and loading of the “effective” strate that suppression of Dicer mediated by miR-BART6-5p 4
strandontothemiRISCcomplex(50).We,therefore,exam- RNAsaffectstheexpressionofalargenumberofmiRNAs.
ined the assembly of functional miRISC from recombinant Modulation of the EBV Latency State by miR-BART6-
FLAG-tagged Ago-2 complexed with Dicer and TRBP and 5p RNAs—We then asked whether Dicer silencing by miR-
either unedited or edited wild-type pre-miR-BART6 RNAs BART6-5p RNAs could control the EBV infection state. To
(Fig.5A)asdescribedpreviously(43,44).Wefoundthatforma- examinethispossibility,wefirstexaminedtherelativeexpres-
tionofthefunctionalmiRISCandconsequentsilencing(cleav- sion levels of miR-BART6-5p strand RNAs and Dicer among
age)ofthe5ptargetRNAwasindeedmuchmoreefficientwith GM607, Daudi, and C666-1 cells by qRT-PCR. Much higher
unedited pre-miR-BART6 than with edited pre-miR-BART6 levelsofmiR-BART6-5pweredetectedinC666-1cells,which
(Fig.5,BandC).LoadingofmiR-BART6–3pstrandRNAsand havemuchlesseditingthanGM607andDaudicells(Fig.6A).
consequent cleavage of their target RNA were extremely in- ThelowlevelsofmiR-BART6-5pinGM607andDaudicellsare
efficient (Fig. 5, B and C), indicating that miR-BART6-5p is consistentwiththehigheditingrateofpri-miR-BART6RNAs
the major effective strand. The cloning frequency for miR- inthesecells(Fig.1B),whichaffectstheirprocessing(Fig.2),
BART6-5pand-3pRNAsinGM607cellsalsoconfirmedthat miRISC formation (Fig. 5C), and consequently the levels of
the5pstrandistheeffectivestrand.Nosequencevariationsin mature miR-BART6-5p RNA. As expected, Dicer levels were
the 5(cid:1) end sequence of the 5p strand was noted among miR- lowestinC666-1cells,ininverserelationtothemiR-BART6-5p
BART6clones,indicatingthateditingatthe(cid:4)20sitedoesnot levels(Fig.6B).Accordingly,wedecidedtoexplorethesignifi-
affecttheDroshacleavagesite(datanotshown).Togetherour canceofDicerrepressionbymiR-BART6-5pRNAsinC666-1
results clearly demonstrate that editing of the wild-type pri- cells. We first attempted to antagonize the miR-BART6-5p
miR-BART6,althoughnotaffectingtheprocessingtopre-and RNAs expressed in C666-1 cells by transfection of a miR-
maturemiRNAs,inhibitstheoverallsilencingeffectsofmiR- BART6-5p antagomir. As expected, the miR-BART6-5p
33364 JOURNALOFBIOLOGICALCHEMISTRY VOLUME285•NUMBER43•OCTOBER22,2010
ControlofEBVLatencybyBART6miRNAEditing
several genes known to be impor-
tantforeitherlyticinfectionorthe
state of latency: EBNA1, EBNA2,
LMP1,Zta,andRta(12,13).EBNA1
is detected in type I, II, and III
latency,whereasEBNA2andLMP1
are usually detected in type III la-
tency (12, 13). EBNA2 is essential
forthetransformationofBlympho-
cytesandplaysacentralroleintype
IIIlatencybyup-regulatingpromot-
ers of all latent EBV genes. Defi-
ciency of the EBNA2 expression is
known in type I and II latency (12,
13).AweakexpressionofLMP1in
typeIIlatencyanditsdeficiencyin
type I latency have been reported
(12,13).Bycontrast,ZtaandRtaare
essential for the initiation of the
lytic EBV infection cycle (12, 13).
D
By antagonizing miR-BART6-5p, ow
n
ZtaandRtaincreasedby2–3-fold, lo
a
indicatingthatmiR-BART6-5pkeeps de
d
thesegeneproductsundercontrol. fro
Furthermore, we noticed substan- m
h
tial up-regulation of EBNA2 onco- ttp
gene expression ((cid:2)5-fold) and ://w
LMP1 ((cid:2)2-fold) by suppression of ww
miR-BART6-5p RNAs, whereas no .jb
c
effects on EBNA1 were observed .org
(Fig.7A). b/
y
The activities of the three viral g
u
e
promoters Cp, Wp, and Qp were st o
alsomonitored(Fig.7B).Transcrip- n M
tionfromCpandWpischaracteris- a
rc
tic of type III latency (51), whereas h 4
the Qp promoter is used in EBV- , 2
0
1
infected cells undergoing type I 4
or II latency (52, 53). We used
qRT-PCR primers specific for
RNAs initiating at Wp, Cp, or Qp
(54). Significant up-regulation of
type III latency-specific Cp and
Wp promoter activities (5.4- and
11-fold,respectively)weredetected
FIGURE5.AssemblyoffunctionalmiRISCswithFLAG-Ago2andpre-miR-BART6RNAs.A,amiRISCloading
in C666-1 cells transfected with
assayofpre-miR-BART6isshown.CleavageofthecognatetargetformiR-BART6-5por-3pRNAsisschemati-
callyshown.ThetargetRNAwas5(cid:1)32P-labeled.B,cleavageofthecognatetargetproduct(17nucleotides) miR-BART6-5p antagomir. On the
guidedbymiR-BART6-5pwassubstantiallymoreefficientwithuneditedpre-miR-BART6RNAsthanwithedited otherhand,Qppromoteractivities
pre-miR-BART6RNAs(leftpanel).Cleavage,althoughveryinefficient,ofbothuneditedandedited3ptarget
associated with type I and type II
wasdetectedonlywithuneditedpre-miR-BART6(middleandrightpanels).C,quantitativesummaryofmiRISC
loadingexperimentsispresented.Thecleavageefficiencywasestimatedbytheratiooftheradioactivityofthe latency were completely abolished;
correctlycleavedbandoverthatoftheuncleavedcontrolband.Significantdifferenceswereanalyzedby thatis,notdetectableincomparison
Mann-WhitneyUtests:***,p(cid:5)0.001.Errorbars,S.E.(n(cid:6)3).
tocontrol.
We then attempted to silence
antagomir substantially decreased the miR-BART6-5p level DicerdirectlybytransfectingaDicertargetingshorthairpinRNA
((cid:2)20-fold)withaconcomitantincreaseintheDicerlevels((cid:2)2- (shRNA) expression vector in C666-1 cells (Fig. 7C). This
fold), indicating that miR-BART6-5p RNAs constantly sup- reducedDicerlevelsby(cid:2)70%,indicatingtheefficiencyofthis
pressandmaintainDiceratthereducedlevelsinC666-1cells DicertargetingsiRNAexpressionvector.Asexpected,changes
(Fig. 7A). We then examined the relative expression levels of inthemarkergeneswerecompletelyoppositetothosenotedin
OCTOBER22,2010•VOLUME285•NUMBER43 JOURNALOFBIOLOGICALCHEMISTRY 33365
ControlofEBVLatencybyBART6miRNAEditing
latencyofC666-1cellsbutalsothetypeIlatencyofMutuIcells
bysuppressinglyticreplicationandalsoinhibitingtransitionof
these cell lines to type III latency, a more immunoresponse-
pronestateoftheviralinfectioncycle.
DISCUSSION
EditingFrequencyofEBVmiRNAs—A-to-Ieditingofaviral
miRNA, KSHV-miR-K12-10 was first implicated because of
identificationofmanycDNAclonescorrespondingtoKSHV-
miR-K12-10 RNAs containing an A-to-G substitution com-
pared with the genomic sequence (55). Additional studies
conducted more recently confirmed that this is indeed due
toA-to-Ieditingatthissiteoftheviraltranscriptharboring
KSHV-miR-K12-10 by ADAR1 (56). Interestingly, the tran-
scriptcouldbeprocessedintotheviralmiRNAaswellasthe
mRNA coding for Kaposin A. A-to-I editing and consequent
recoding of Kaposin A reduced its transforming activity (56).
However, the significance of A-to-I editing of KSHV-miR-
FIGURE6.RelativeexpressionlevelsofmiR-BART6-5pandDicerindiffer-
K12-10RNAsremainsunknown.
entcelllines.A,miR-BART6-5pRNAlevelswereexaminedbyqRT-PCRand
normalizedto(cid:2)-actinmRNAlevel.Threeindependentassaysweredone.Sig- Apart from these reports on KSHV-miR-K12-10 RNAs,
D
bnaifrics,aSn.tEd.(inffe(cid:6)re3n)c.eBs,wDiecreeramnRalNyAzeldevbeylsMwaenrne-WmhointniteoyreUdtbesytsq.R*T,p-P(cid:5)CR0.a0n5d.Enrororr- there has been no additional report on A-to-I editing of viral own
malizedto(cid:2)-actinmRNAlevels.Threeindependentassayswereperformed. miRNAs. In this study we examined EBV miRNAs for A-to-I lo
a
SignificantdifferenceswereanalyzedbyMann-WhitneyUtests.*,p(cid:5)0.05. RNAeditinginGM607Blymphoblastoidcells,DaudiBurkitt de
Errorbars,S.E.(n(cid:6)3). lymphomacells,andC666-1nasopharyngealcarcinomacells. d fro
We found that primary transcripts of four miRNAs, miR- m
h
C666-1 cells transfected with the miR-BART6-5p antagomir; BHRF1-I,miR-BART6,miR-BART8,andmiR-BART16,undergo ttp
thatis,an(cid:2)7-foldreductionintheEBNA2expressionaswellas editing at specific sites. In view of (cid:2)20% of human miRNAs ://w
w
substantialdown-regulationforLMP1,Zta,andRta(Fig.7C), beingsubjecttoA-to-Iediting(33),ourfindingsofeditingof4 w
furtherconfirmingthatcontrolofthesecriticalviralgenesby of23EBVmiRNAsindicatethatbothcellularandviralmiRNAs .jb
c
miR-BART6-5pismediateddirectlyviaitssilencingeffectson aresubjecttoeditingataboutthesamefrequency.Amongfour .org
Dicer. EBV pri-miRNAs, we focused on pri-miR-BART6, which is b/
y
Finally,weexaminedgeneticallyidenticalpairsofBurkitt highlyeditedatthe(cid:4)20siteofthe3pstrandsideofthehairpin g
u
e
lymphomaMutuIandMutuIIIcelllines,whichareintypeI dsRNAstructure. st o
and type III latency, respectively, to assess the function of SuppressionofmiR-BART6ExpressionandmiRISCAssembly n M
miR-BART6-5p and Dicer silencing in B lymphoma cells by A-to-I Editing—We have shown previously that A-to-I a
rc
(non- nasopharyngeal carcinoma cell lines). We found that editingofpri-miRNAscansuppresstheirprocessingtopre- h 4
themiR-BART6-5plevelishigherinMutuIthaninMutuIII miRNAsbyinhibitingDroshacleavageinthenucleus(34)or , 2
0
1
(supplementalFig.6A).Bycontrast,theDicerlevelwaslower suppress processing of pre-miRNAs to mature miRNAs by 4
inMutuIthaninMutuIIIasexpected(supplementalFig.6B).Low inhibitingDicercleavageinthecytoplasm(35).Furthermore,in
levelA-to-Ieditingofpri-miR-BART6RNAswasdetectedonly somecasesA-to-Ieditingofpri-miRNAsresultedinexpression
in Mutu III cells, and no mutations were found in the miR- ofmiRNAswithanaltered(edited)seedsequenceandconse-
BART6 gene of Mutu I and Mutu III cells (data not shown). quentsilencingofasetofgenesdifferentfromthosetargetedby
Thus, it is currently unknown why higher miR-BART6-5p theuneditedversionmiRNAs(36).
expressionisdetectedinMutuIcellsascomparedwithMutu In vitro pri-miR-BART6 processing studies revealed that a
IIIcells.WefirsttransfectedMutuIcellswithmiR-BART6-5p combinationofA-to-Ieditingatthe(cid:4)20site,andthethree-U-
antagomir.AsweobservedinC666-1cells,theantagomireffec- residue deletion mutation, as observed in Daudi Burkitt lym-
tivelyreducedmiR-BART6-5plevels((cid:2)10-fold)andincreased phomaandC666-1nasopharyngealcarcinomacells,blocksthe
Dicerlevels(1.9-fold).Furthermore,significantup-regulation Droshacleavagestepcompletely.Editingofwild-typepri-miR-
ofEBNA2,LMP1,Zta,andRta,aswellasWpandCpactivation, BART6RNAsdidnotaffecttheirprocessing.However,loading
was detected (Fig. 8, A and B). We then transfected Mutu III of miR-BART6-5p onto the functionally active miRISC was
cellswiththeDicershRNAexpressionplasmid,whichsuccess- substantiallyinhibitedbyA-to-Ieditingatthe(cid:4)20site.Editing
fullyrepressedDicerlevels((cid:2)5-fold).Oppositeeffectsofthe of pri-miR-BART6 RNAs reported in this study is the first
miR-BART6-5pantagomirweredetected;this,is,suppression example in which editing suppresses loading of miRNA onto
ofEBNA2,LMP1,Zta,andRta(Fig.8C).Up-regulationofQp miRISC.
activities and down-regulation of Cp and Wp activities were Selection of miR-BART6-5p as an Effective Strand—It has
alsoobserved(Fig.8D). beenreportedthattherelativestabilitiesofthebasepairsatthe
Together,theseresultssuggestthatDicersuppressionmedi- 5(cid:1)endsoftheduplexconsistingoftwomiRNAstrandsdeter-
atedviamiR-BART6-5pRNAsmaintainsnotonlythetypeII minetheselectionoftheeffectivestrand,whichisloadedonto
33366 JOURNALOFBIOLOGICALCHEMISTRY VOLUME285•NUMBER43•OCTOBER22,2010