The RNA Sensor RIG-I Dually Functions as an Innate Sensor and Direct Antiviral Factor for ...

36 Pages·2015·4.14 MB·English

by  Seiichi Sato

Article
The RNA Sensor RIG-I Dually Functions as an Innate
Sensor and Direct Antiviral Factor for Hepatitis B
Virus
Graphical Abstract Authors
SeiichiSato,KaiLi,...,YasuhitoTanaka,
AkinoriTakaoka
Correspondence
[email protected]
In Brief
ThesensingmechanismofhepatitisB
virus(HBV)andthesubsequentsignaling
eventsremaintobefullyclarified.Sato
andcolleaguesdemonstratethattheRNA
sensorRIG-InotonlysensesHBV
pregenomicRNAtopreferentiallyinduce
typeIIIinterferonproductionbutalso
counteractstheinteractionofviral
polymerasewiththepregenomicRNAfor
antiviraldefenseagainstHBV.
Highlights
d TypeIIIIFNsarepredominantlyinducedinhuman
hepatocytesduringHBVinfection
d RIG-IsensestheHBVgenotypeA,B,andCfortheinduction
oftypeIIIIFNs
d The50-εregionofHBVpgRNAisakeyelementfortheRIG-I-
mediatedrecognition
d RIG-IcounteractstheinteractionofHBVPwithpgRNAto
suppressviralreplication
Satoetal.,2015,Immunity42,123–132
January20,2015ª2015ElsevierInc.
http://dx.doi.org/10.1016/j.immuni.2014.12.016
Immunity
Article
The RNA Sensor RIG-I Dually Functions
as an Innate Sensor and Direct Antiviral Factor
for Hepatitis B Virus
SeiichiSato,1,2,10KaiLi,1,2,10TakeshiKameyama,1,2,10TakayaHayashi,3,10YujiIshida,4ShukoMurakami,5
TsunamasaWatanabe,5SayukiIijima,5YuSakurai,6KoichiWatashi,7SusumuTsutsumi,5YusukeSato,6HidetakaAkita,6
TakajiWakita,7CharlesM.Rice,8HideyoshiHarashima,6MichinoriKohara,9YasuhitoTanaka,5andAkinoriTakaoka1,2,*
1DivisionofSignalinginCancerandImmunology,InstituteforGeneticMedicine,HokkaidoUniversity,Sapporo,Hokkaido060-0815,Japan
2MolecularMedicalBiochemistryUnit,BiologicalChemistryandEngineeringCourse,GraduateSchoolofChemicalSciencesand
Engineering,HokkaidoUniversity,Sapporo,Hokkaido060-0815,Japan
3ResearchCenterforInfection-AssociatedCancer,InstituteforGeneticMedicine,HokkaidoUniversity,Sapporo,Hokkaido060-0815,Japan
4PhoenixBioCo.,Ltd.,Higashihiroshima,Hiroshima739-0046,Japan
5DepartmentofVirologyandLiverUnit,NagoyaCityUniversityGraduateSchoolofMedicalSciences,Nagoya,Aichi467-8601,Japan
6LaboratoryofInnovativeNanomedicine,FacultyofPharmaceuticalSciences,HokkaidoUniversity,Sapporo,Hokkaido060-0812,Japan
7DepartmentofVirologyII,NationalInstituteofInfectiousDiseases,Tokyo208-0011,Japan
8LaboratoryofVirologyandInfectiousDisease,TheRockefellerUniversity,NewYork,NY10065,USA
9DepartmentofMicrobiologyandCellBiology,TokyoMetropolitanInstituteofMedicalScience,Setagaya-ku,Tokyo156-8506,Japan
10Co-firstauthor
*Correspondence:[email protected]
http://dx.doi.org/10.1016/j.immuni.2014.12.016
SUMMARY zer et al., 2012; Revill and Yuan, 2013). Around four hundred
million people worldwide are persistently infected with HBV,
Host innate recognition triggers key immune re- whichisamajorcausativefactorassociatedwithnotonlyinflam-
sponses for viral elimination. The sensing mecha- mationbutalsocirrhosisandevencanceroftheliver.Currently,
nism of hepatitis B virus (HBV), a DNA virus, and interferon(IFN)andnucleoside/nucleotideanalogsareavailable
thesubsequentdownstreamsignalingeventsremain for HBV treatment (Rehermann and Nascimbeni, 2005; Hale-
to be fully clarified. Here we found that type III but goua-De Marzio and Hann, 2014). However, the long-term
responseratesarestillnotsatisfactory.Elucidationofhostim-
not type I interferons are predominantly induced in
muneresponseagainstHBVinfectioniscrucialforbetterunder-
human primary hepatocytes in response to HBV
standing of the pathological processes and viralelimination to
infection, through retinoic acid-inducible gene-I
controlHBVinfection.
(RIG-I)-mediated sensing of the 50-ε region of HBV
ThetypeIIFNs,IFN-aandIFN-b,arerepresentativecytokines
pregenomicRNA.Inaddition,RIG-Icouldalsocoun-
thatelicithostinnateimmuneresponsesagainstviralinfections.
teract the interaction of HBV polymerase (P protein) Inaddition,anotherIFNfamily,typeIIIIFNs(IFN-l,alsoknownas
with the 50-ε region in an RNA-binding dependent IL-28andIL-29)exhibitspotentantiviralactivitysimilartoIFN-a
manner,whichconsistentlysuppressedviralreplica- andIFN-b(Sheppardetal.,2003;Kotenko,2011;Kotenkoetal.,
tion.Liposome-mediateddelivery andvector-based 2003).ProductionoftypeIandtypeIIIIFNsismassivelyinduced
expression of this ε region-derived RNA in liver inmanytypesofcellsuponinfectionwithvariousviruses,which
abolished the HBV replication in human hepato- isknowntobemediatedbytheactivationofpattern-recognition
receptors (PRRs). During virus infection, virus-derived nucleic
cyte-chimeric mice. These findings identify an
acids(bothRNAandDNA)aremainlysensedbycertainPRRs,
innate-recognitionmechanismbywhichRIG-Idually
suchasretinoicacid-induciblegene-I(RIG-I)(Yoneyamaetal.,
functions as an HBV sensor activating innate
2004;Choietal.,2009;Chiuetal.,2009;Ablasseretal.,2009),
signaling and to counteract viral polymerase in
melanomadifferentiation-associatedgene5(MDA5)(Yoneyama
humanhepatocytes.
etal.,2005),cyclicGMP-AMPsynthase(cGAS)(Sunetal.,2013),
andIFN-g-inducibleprotein16(IFI16)(Unterholzneretal.,2010).
Particularly, RIG-I is a key PRR that can detect virus-derived
INTRODUCTION RNAsinthecytoplasmduringinfectionwithavarietyofviruses,
such as influenza virus, hepatitis C virus (HCV), and measles
HepatitisBvirus(HBV)isahepatotropicvirusoftheHepadnavir- virus,whicharecloselyrelatedtohumandiseasepathogenesis
idae family and contains a circular, partially double-stranded (RehwinkelandReiseSousa,2010).BindingofRIG-Itoitsligand
DNA genome of about 3.2 k base pairs that is replicated via RNAs, such as 50-triphosphorylated RNA or short double-
reverse transcription of a pregenomic RNA (pgRNA). HBV stranded RNAs (Takeuchi and Akira, 2009; Hornung et al.,
causes hepatic inflammation associated with substantial 2006), activates the downstream signaling pathways in a
morbidityworldwide(RehermannandNascimbeni,2005;Prot- mannerdependentontheadaptorproteinmitochondrialantiviral
Immunity42,123–132,January20,2015ª2015ElsevierInc. 123
signalingprotein(MAVS;alsoknownasIPS-1,VISA,orCardif) cytes(PHH)invitro24hrafterinfectionwithHBV-C(Figure1C);
(TakeuchiandAkira,2009),leadingtotheinductionoftheIFN- however,neitheroftypeInortypeIIIFNtestedwasinduced(Fig-
regulatoryfactor-3(IRF-3)andNF-kB-dependentgeneexpres- uresS1DandS1E).Althoughitisdifficulttosimplycomparethe
sionandthesubsequentproductionoftypeIandtypeIIIIFNs amountofIFNinducedbydifferenttypesofvirus,theinductionof
and inflammatory cytokines (Takeuchi and Akira, 2009). Thus, IFN-l1, l2, and l3 mRNAs in response to HBV infection was
RIG-I sensing of viral RNA is a crucial process to activate the muchweakerthanthatofNewcastlediseasevirus(NDV)infec-
antiviralinnateresponsestolimitviralreplicationandtheactiva- tion(Figure1C).Inthisregard,inordertoruleoutthepossibility
tionofadaptiveimmunity(TakeuchiandAkira,2009). thattheIFN-lresponseisduetocontaminantsintheinocula,we
Asforthevirusesthatareknowntobetheleadingcauseofhe- usedLamivudine (LAM),anHBV inhibitor, in thisassay. Treat-
paticinflammation, RIG-IisthemajorPRRthatinitiates innate mentwithLAMinhibitedIFN-lmRNAinductioninresponseto
immuneresponsesagainstHCV.RIG-IsensingofHCVismedi- HBV infection in PHH (Figure 1C), suggesting that the IFN
atedthroughitsrecognitionofthepoly-U/UCmotifoftheHCV response is actually induced by HBV replication. Furthermore,
RNAgenome30nontranslatedregion,whichleadstotheactiva- weanalyzedHepG2-sodium taurocholatecotransporting poly-
tionoftypeIIFNresponse(Saitoetal.,2008).Ontheotherhand, peptide (NTCP)-C4 cell line (Iwamoto et al., 2014) stably ex-
earlierstudieshaveshownthattheinnateimmuneactivationis pressing human NTCP, a functional receptor for HBV (Yan
impaired and the induction of type I IFNs such as IFN-a or etal.,2012),andconfirmedthatIFN-l1andIFN-induciblegenes
IFN-bishardlydetectedinanimalmodelsofHBVinfection,as suchasOAS2andRSAD2,butnotIFN-b,wereinducedinthese
comparedwithHCVinfection(Wielandetal.,2004;Nakagawa cellsafterinfectionwithHBV-C,andthattheseinductionswere
et al., 2013). However, it is still not fully clarified how HBV is abolishedbytreatmentwithLAM(Figure1D).Tonextassessthe
recognizedbyhumanhepatocytesandtheroleoftypeIIIIFNs innateimmuneresponsesinvivoduringHBVinfection,weex-
aswell. ploitedseverecombinedimmunodeficiencymicethatcarrythe
HerewereportthatHBVinfectionpredominatelyinducestype urokinase-type plasminogen activator transgene controlled by
III,butnottypeI,IFNgeneinduction,whichismediatedbyRIG-I an albumin promoter (uPA+/+/SCID mice), in which more than
throughitsrecognitionofthe50-εregionofHBV-derivedpgRNA. 70% of murine hepatocytes were replaced by human hepato-
WealsoshowthatRIG-IcancounteracttheinteractionofHBV cytes (Tateno et al., 2004) (hereinafter referred to as chimeric
polymerase (P protein) with the 50-ε region of pgRNA in an mice).Afterthechimericmicewereintravenouslyinfectedwith
RNA-binding dependent manner, resulting in the suppression HBV-C,whichwasderivedfrompatientswithchronichepatitis,
of HBV replication. Furthermore, liposome-mediated delivery theexpressionoftypeIIIIFNmRNAsincreasedinthelivertissue,
and expression of the 50-ε region-derived RNA in liver sup- whereas IFN-a4and IFN-bmRNAs were not upregulated (Fig-
pressedtheHBVreplicationinvivoinchimericmicewithhuman- ure 1E). In parallel with this type III IFN response, we also
ized livers. Thus, our findings demonstrate the innate defense observed the expression of IFN-inducible genes, such as
mechanismsbasedontheviralRNA-RIG-Iinteraction,whereby CXCL10,OAS2,andRSAD2,inthehumanliveroftheseinfected
RIG-IfunctionsnotonlyasaHBVsensorfortheactivationofIFN mice(Figure1E).ThesefindingsindicatethatamoderatetypeIII
responsebutalsoasadirectantiviralfactor. butnottypeIortypeIIIFNresponseisactivatedinhumanhepa-
tocytesinresponsetoHBVinfection.
RESULTS
HBV-InducedTypeIIIIFNExpressionDependsonRIG-I
TypeIIIIFNsArePredominantlyInducedinHepatocytes Wenextdeterminedwhichsensor-mediatedsignalingpathway
duringHBVInfection is responsible for the HBV-induced type III IFN response. As
To investigate the innate immune activation during HBV infec- HBVisaDNAvirus(RehermannandNascimbeni,2005;Protzer
tion,weexaminedtypeIandtypeIIIIFNresponsesinhumanhe- etal.,2012;RevillandYuan,2013),weassessedthecontribution
patocytes.Consistentwiththepreviousreports(Wielandetal., of previously reported cytosolic DNA sensors including RIG-I
2004;Nakagawaetal.,2013),wehardlyobservedtheinduction (Chiuetal.,2009;Ablasseretal.,2009;Choietal.,2009),IFI16
oftypeIIFNs,IFN-a4,andIFN-binresponsetotransfectionwith (Unterholzneretal.,2010),andcGAS(Sunetal.,2013)inhuman
plasmids carrying 1.24-fold the HBV genome of three major hepatocytes.KnockdownanalysesrevealedthatIFN-l1induc-
different genotypes, Ae (HBV-Ae), Bj (HBV-Bj), and C (HBV-C) tion in HepG2 or Huh-7 cells by plasmid transfection for
(Figure1AandFigureS1Aavailableonline)atleastuptoseven HBV-CorHBV-Ae,respectively,wassuppressedbytheknock-
days after transfection, although the expression of HBV RNAs downofRIG-I,butnotthatoftheothersensors(Figures2A,S2A
was detectable (Figure S1B). On the other hand, type III IFN, andS2B).To furtherconfirm theinvolvement of RIG-Iin HBV-
IFN-l1,wasinducedinallofthethreetypesofhumanhepato- triggered type III IFN response, we measured IFN-l1 mRNA
cyte cell line tested (Figures 1A and S1A). In HepG2 cells, expression induced by plasmid expression in Huh-7.5 cells
HBV-C shows the highest IFN-l1 response, which was also thatcarryadominant-negativemutantRIG-Iallelethatprevents
confirmed byELISA, albeit weakly (Figures 1Aand 1B). More- RIG-I signaling (Saito et al., 2007), as compared with Huh-7
over,IFN-l1inculturesupernatantcouldinhibitvesicularstoma- cells that have an intact RIG-I pathway. Huh-7.5 cells failed
titisvirus(VSV)replicationinplaquereductionassay,aswellas to induce IFN-l1 mRNA expression in response to HBV-Ae
HBV replication (Figure S1C), indicating the physiological rele- genomeplasmidtransfection,asinthecaseofstimulationwith
vance of the induced IFN-l1 to antiviral activities. Consistent 50-triphosphate RNA (3pRNA), a RIG-I ligand (Takeuchi and
withtheseresults,weobservedthesignificantinductionofnot Akira,2009; Hornungetal.,2006) (Figure2B).Inconcordance
onlyIFN-l1butalsoIFN-l2and-l3inprimaryhumanhepato- withthisresult,knockdownoftripartitemotifcontainingprotein
124 Immunity42,123–132,January20,2015ª2015ElsevierInc.
Figure1. IFN-lInductioninHumanHepatocytesinResponsetoHBVInfection
(A)QuantitativeRT-PCR(qRT-PCR)analysisofIFNL1(left),IFNA4(middle),andIFNB1(right)mRNAattheindicatedtimesaftertransfectionwith1.24-foldthe
HBVgenome(genotypeAe,Bj,orC)oremptyvector(Mock)inHepG2cells.
(B)ELISAofIFN-l1at48or72hraftertransfectionwiththeHBVgenomeinHepG2cells.Thedotlineindicatestheminimumdetectableamount(31.2pg/ml)of
IFN-l1bytheELISAkit.ND,notdetected,indicatesbelowtheminimumdetectableamount.
(C)qRT-PCRanalysisofIFNL1,IFNL2,andIFNL3mRNAat24hrafterinfectionwithHBV,NDV(multiplicityofinfection=10)ormock((cid:2)),ormedia-treated
Lamivudine(LAM)ascontrolinprimaryhumanhepatocytes(PHH).ThemRNAcopynumber(±SD)ofeachsubtypeoftypeIIIIFNper1mgtotalRNAuponHBV-C
infectionisasfollows:IFNL1(83,197.6±6,241.4)andIFNL2/3(409,280.6±119,676.2).
(D)TimecourseanalysesbyqRT-PCRofIFNL1,OAS2,RSAD2,IFNB1mRNA,andpgRNAattheindicatedtimesafterHBVinfectioninHepG2-hNTCP-C4cells.
TheeffectofLamivudinetreatmentwasalsoanalyzed.
(E)qRT-PCRanalysisofIFNL1,IFNL2,IFNL3,IFNA4,IFNB1,CXCL10,OAS2,andRSAD2mRNAoflivertissuesat4or5weeksafterinfectionwithHBV-Cin
chimericmice.((cid:2)),uninfectedmice.Redlinesrepresentthemeanofeachdataset.*p<0.05and**p<0.01versuscontrol.RE,relativeexpression.(A–D)Dataare
presentedasmeanandSD(n=3)andarerepresentativeofatleastthreeindependentexperiments.SeealsoFigureS1.
25(TRIM25),MAVS,TANK-bindingkinase1(TBK1),andIRF-3, sultedinthesuppressionofIFN-l1mRNAinductioninHepG2
all of which are signaling molecules essentially involved in the cells in response to transfection with the HBV-C genome. On
RIG-I-mediated IFN pathway (Takeuchi and Akira, 2009), re- theotherhand,suchaneffectwasnotobservedincellstreated
Immunity42,123–132,January20,2015ª2015ElsevierInc. 125
Figure2. RIG-I-DependentIFN-lInduction
inResponsetoHBVInfection
(A)HepG2cellstreatedwithcontrolsiRNA(Con-
trol)orsiRNAtargetingRIG-I,IFI16,orcGASwere
transfected with the HBV-C genome for 48 or
72hr.TheamountofIFN-l1weremeasuredby
ELISA.Thedotlineindicatestheminimumcyto-
kineexpressiondetected(31.2pg/ml)ofIFN-l1by
theELISAkit.ND,notdetected,indicatesbelow
detectableconcentrations(left),andknockdown
efficiency was analyzed by immunoblotting (IB)
(right).
(B)qRT-PCRanalysisofIFNL1mRNAinHuh-7or
Huh-7.5 cells transfected with the HBV-Ae
genome(at24hraftertransfection)orstimulated
with3pRNA(1mg/ml)for6hr.
(C)HepG2cellstreatedwithcontrolsiRNA(Con-
trol)ortheindicatedsiRNAsweretransfectedwith
the HBV-C genome. At 48 hr after transfection,
totalRNAsweresubjectedtoqRT-PCRanalysis
forIFNL1.
(D)qRT-PCRanalysisofIFNL1mRNAinsiRNA-
treatedPHHat24hrpostinfectionwithindicated
HBVgenotype.Mock,emptyvector-transfected.
((cid:2)), uninfected. Data were normalized to the
expression of GAPDH. Data are presented as
meanandSD(n=3)andarerepresentativeofat
least three independent experiments. *p < 0.05
and **p < 0.01 versus control in (B) or HBV-in-
fectedcontrolgroupin(A,C,andD).NS,notsig-
nificant.SeealsoFigureS2.
witheitherTRIF(alsoknownasTICAM-1)orMYD88siRNA(Fig- the expression of all of these RNA transcripts and tested its
ures2CandS2C).Inaddition,weconfirmedthattheknockdown effect on HBV-induced IFN-l1 expression. As shown in Fig-
ofRIG-IandMAVSabolishedIFN-l1inductioninPHHinfected ure 3B, knockdown with this siRNA (Figure S3A) suppressed
witheachgenotype(Figure2D).Furthermore,wealsoconfirmed IFN-l1inductioninHuh-7cellstransfectedwithHBV-Ae.Next,
by knockdown assay that the induction of IFN-l1 and OAS2 todeterminewhichoftheseHBVRNAtranscriptsis/areinvolved
mRNAinHepG2-hNTCP-C4cellsinresponsetoinfectionwith intheRIG-I-mediatedIFN-l1induction,wepreparedexpression
HBV-CwasdependentonRIG-I(FigureS2E).Thesedataindi- vectors to express each of these four viral transcripts in
catethatIFN-l1geneinductionduringHBVinfectiondepends HEK293T cells that are often used to analyze RIG-I signaling
largelyonRIG-Isignalingpathway. pathway in human cells. As a result, it is only the longest
3.5kbtranscript,thatis,pgRNA,thathasthepotentialtoelicit
The50-εRegionofHBVpgRNAIsaKeyElementfor a significant induction of IFN-l1 mRNA (Figures 3C and S3B).
RIG-I-DependentIFN-l1Induction ItwasalsoconfirmedbyknockdownanalysiswithpgRNA-tar-
RIG-Icanrecognizenotonlyvirus-derivedRNAbutalsoDNAin geted siRNA, which showed significant suppression of IFN-l1
the cytoplasm (Yoneyama et al., 2004; Choi et al., 2009; Chiu inductioninHepG2cellstransfectedwithHBV-Ae(FigureS3C).
etal.,2009;Ablasseretal.,2009).TofurtherclarifyhowRIG-I Theseresultssuggestthat50-1.1kbregionofHBVpgRNAiscrit-
recognizes HBV, we first examined either or both of which icalfortheactivationofRIG-IpathwaytoinduceIFN-l1expres-
nucleic acid (DNA and RNA) derived from HBV-infected cells sion.Ontheotherhand,theremainingthreetranscripts,which
canactivateIFN-l1geneexpression.Transfectionwithnucleic alsocontainthesamesequenceofpartofthis1.1kbregionof
acidfractionsextractedfromHBVinfectedHuh-7cellsafterpre- HBV pgRNA at the 30 end of their transcripts, failed to induce
treatment with RNase A, but not DNase I resulted in marked IFN-l1mRNA(Figure3C).AnartificiallydeletedformofpgRNA,
inhibition of the IFNL1 promoter activation, suggesting that which lacks this overlapping sequence at the 30-region (D3),
virus-derivedRNAsmightbecandidatesoftheRIG-Iliganddur- showed IFN-l1 induction, whereas such response was not
ingHBVinfection(Figure3A). observedfor anothermutant pgRNAlacking itatthe 50-region
The HBV genome comprises a partially double-stranded (D5)(Figure3D).Thesedataalsosupportapossibleimportant
3.2kbDNA.DuringalifecycleofHBVinhepatocytes,itscova- role of the 50-overlapping sequence of HBV pgRNA for RIG-I-
lentlyclosedcircularDNA(cccDNA)istranscribedtogenerate mediatedIFN-l1induction.
fourmajorRNAspecies:the3.5,2.4,2.1,and0.7kbviralRNA The50-endofHBVpgRNAisknowntocontaintheencapsida-
transcripts (Rehermann and Nascimbeni, 2005; Protzer et al., tionsequence,called‘‘epsilon(ε),’’whichtakesastem-loopsec-
2012;RevillandYuan,2013).WecreatedansiRNAtosuppress ondary structure (Junker-Niepmann et al., 1990; Pollack and
126 Immunity42,123–132,January20,2015ª2015ElsevierInc.
Figure3. RIG-IActivationIsMediatedbyIts
Recognition of the 50-ε Region of HBV
pgRNA
(A) Luciferase activity of an IFN-l1 reporter
plasmid after 24 hr of stimulation with
nucleic acids (2 mg/ml) extracted from Huh-7
cells transfected with control plasmid (Mock)
ortheHBV-AegenomewithorwithoutRNaseA
or DNase I treatment. RLU, relative luciferase
units.
(B) Huh-7 cells treated with control or HBV
RNA-targeted siRNA were transfected with the
HBV-Ae genome or mock. After 24 hr of trans-
fection,totalRNAsweresubjectedtoqRT-PCR
forIFNL1.
(CandD)Aschematicrepresentationoffourtypes
ofHBVRNAs,pgRNA(3.5kb),2.4kb,2.1kb,and
0.7 kb RNAs in (C), and two deleted forms of
pgRNA,D5andD3,in(D).Theoverlappingregion
is shown in blue. qRT-PCR analysis of IFNL1
mRNAofHEK293Tcellsafter24hroftransfection
withtheindicatedexpressionvectors.Datawere
normalized to the amount of each HBV RNA
expression(CandD).
(E)AschematicrepresentationofpgRNA,εRNA,
or control RNA (ContRNA) (left). HEK293T cells
treated with control or RIG-I siRNA were
unstimulated (Mock) or stimulated with εRNA
for 12 hr. Total RNAs were subjected to qRT-
PCRforIFNL1(middle).qRT-PCRanalysisofIFNL1mRNAinHEK293Tcellsafter12hrofstimulationwithεRNAorContRNA(right).EachoftheRNAswas
preparedbyinvitrotranscription.
(F)HEK293Tcellsweretransfectedwitheachplasmidforstem-loopmutantsofpgRNA(5STM,3STM,or5and3STM),andthensubjectedtoqRT-PCRanalysis
asdescribedin(C).*p<0.05and**p<0.01versuscontrolin(A,B,andE)orversus3.5Kin(C,D,andF).NS,notsignificant.SeealsoFigureS3.
Ganem, 1993; Knaus and Nassal, 1993; Jeong et al., 2000). ContRNA,inHEK293Tcells(Figure4A,top).Similarly,endoge-
Therefore,wehypothesizedthatthis50-εstructuremightconfer nousRIG-IinteractedwithεRNAalbeitweakly(Figure4A,bot-
apossiblepathogen-associatedmolecularpattern(PAMP)motif tom). We also demonstrated the intracellular colocalization of
for RIG-I recognition. To test this hypothesis, we stimulated RIG-IwithεRNAinHuh-7.5cells(Figure4B).Inaddition,RNA-
HEK293TandHepG2cellswiththeεregion-derivedRNA(here- binding protein immunoprecipitation (RIP) assay revealed that
after called εRNA). Consequently, IFN-l1 mRNA was signifi- thefulllengthofHBVpgRNAwasdetectedintheRIG-I-immuno-
cantly induced, which was dependent on RIG-I, while such a precipitatedcomplex,andD5pgRNAandD3pgRNAwerealso
responsewasnotdetecteduponstimulationwiththeequivalent detected(FigureS4A),whichisseeminglyinconsistentwiththe
length of RNA that is derived from HBV pgRNA but does not results by the functional assay (Figures 3C, 3D, 3F and S3C).
containanyεelement(ContRNA)(Figure3EandS3D).Wealso Theseresultssuggestthattheεregionisrequiredforitsinterac-
confirmedRIG-I-dependentIRF-3activationinresponsetostim- tionwithRIG-I,butonlythe50-εregionisnecessarytoactivate
ulationwithεRNA(FiguresS3DandS3E).Duetotheoverlapping RIG-I pathway. We further tried to determine which region of
sequenceof50-and30-endsofHBVpgRNAasmentionedabove, RIG-ImediatesitsinteractionwithHBVpgRNA.BothRIPassay
thisεelementisfoundatbothendsofpgRNA.Wenextgener- andRNApull-downassaywithseveraldeletionmutantsofRIG-I
atedseveralmutantformsofHBVpgRNA,eachofwhichcarries showedthattheC-terminalportionofRIG-I(C-RIG)includingits
mutationswithin50-or30-εregionorbothtodisruptthestem- helicasedomainandrepressordomain(RD)exceptforCARDs
loop structure (5STM, 3STM, or 5 and 3STM, respectively). In canbindtoHBVpgRNA(Figure4C;FiguresS4BandS4C).In
concordancewiththeresultsshowninFigures3Cand3Dand addition, gel shift assay showed that the interaction of HBV
S3B, IFN-l1 mRNA induction was detected upon expression εRNA or pgRNA was impaired with the RD or C-RIG mutant,
ofthe3STMtranscriptthathasanintact50-εregion,assimilar respectively, each of which carries a point mutation (K888E)
to that of intact 3.5-kb pgRNA (Figure 3F). In contrast, either that abolishes its RNA-binding activity (Cui et al., 2008) (Fig-
5STMor5and3STMdidnotshowsignificantresponse.These ure4D).AsimilarresultwasalsoobtainedbyRIPassay,wherein
findings indicate that the 50-ε region of HBV pgRNA is critical thewild-type(WT)C-RIG,butnottheK888Emutant,wascoim-
forIFN-l1inductionpossiblythroughtherecognitionbyRIG-I. munoprecipitatedwithHBVpgRNA(FigureS4D),likeHCVRNA
thatwaspreviouslyreportedtointeractwithRIG-I(FigureS4E).
RIG-IInteractswiththeε-RegionofpgRNA WealsoconfirmedtheinteractionofHBVpgRNAwithendoge-
Next,weassessedtheinteractionofRIG-Iwiththeεregionof nousRIG-IinHepG2cells,whereasitsinteractionwithothernu-
HBV pgRNA, that is, εRNA. Pull-down assays showed that cleicacidsensors,suchasIFI16andMDA5(Yoneyamaetal.,
Flag-tagged RIG-I was coprecipitated with εRNA,but notwith 2005), was not detected (Figure 4E). These data indicate that
Immunity42,123–132,January20,2015ª2015ElsevierInc. 127
Figure4. RIG-IInteractswiththeεRegion
ofpgRNA
(A) RNA pull-down assay showing the binding
activityoftheindicatedRNAstoFlag-taggedRIG-I
(Flag-RIG-I)inHEK293Tcells(top)orendogenous
RIG-IinHepG2cells(bottom).
(B)FRETanalysisfortheinteractionofYFP-tagged
RIG-I (YFP-RIG-I) with rhodamine (ROX)-conju-
gatedεRNA(εRNA-ROX)orContRNA(ContRNA-
ROX). Representative fluorescence images of
YFP,ROX,andFRETC/YFP(theratioofcorrected
FRET(FRETC)toYFP).Arrowheadsindicatearea
showing high FRET efficiency. Scale bar repre-
sents20mm.Right,dotplotofFRETC/YFPratio
(smallhorizontalbars,mean).
(C)RIPassaywithHEK293Tcelllysatesexpress-
ingseveralFlag-taggeddeletionmutantsofRIG-I
andpgRNAexpressionvectorbyusinganti-Flag
antibody.ImmunoprecipitatedpgRNAwasquan-
titatedbyqRT-PCRandnormalizedtotheamount
ofimmunoprecipatedproteins(FigureS4C)andis
represented as fraction of input RNA prior to
immunoprecipitation(percentageinput).
(D) Gel-shift analysis of complex formation be-
tweenεRNAandrecombinantRIG-IRD(WT)orRD
(K888E).Arrowheadsdenotepositionofunbound
RNAandRNA–RIG-Icomplexes.
(E) RIP assay with HepG2 cell lysates prepared after 48 hr of transfection of the HBV-C genome by using anti-RIG-I, anti-IFI16, anti-MDA5, or control
immunoglobulinG.TheimmunoprecipitatedpgRNAwasmeasuredbyqRT-PCR(top)asdescribedin(C).Whole-cellexpressionandimmunoprecipitated
amountsofRIG-I,IFI16,andMDA5(bottom).DataarepresentedasmeanandSD(n=3)andarerepresentativeofatleastthreeindependentexperiments.
*p<0.05and**p<0.01versuscontrolin(BandE).NS,notsignificant.SeealsoFigureS4.
the50-εregionofviralpgRNAfunctionsasanHBV-associated K888E(Cuietal.,2008)mutantcouldnotinhibitthebindingof
molecular pattern to be specifically recognized by RIG-I and PproteinwithpgRNA(Figures5DandS5C).Inaddition,treat-
cantriggerIFN-lresponse. mentwithrecombinantIFN-l1inHuh-7.5cellsupregulatedthe
amountofthemutantRIG-Iprotein(T55I)(FigureS5D),resulting
RIG-IExertsanAntiviralActivitybyCounteractingthe inapartialinhibitionofthePproteininteractionwithpgRNA,and
InteractionofHBVPolymerasewithpgRNA this inhibitory effect was abrogated by RIG-I knockdown (Fig-
WenextassessedthecontributionofRIG-Ipathwayinantiviral ure5E).Infact,FRETanalysisshowedthatthePprotein-εRNA
defense against HBV infection. RIG-I knockdown in PHH re- interaction was significantly suppressed by expression of the
sulted in a higher HBV genome copy number at 10 days after RIG-IRD(WT)alone,butnotthemutantRD(K888E)(FigureS5E).
infection with HBV-C, as compared with PHH treated with Furthermore,HBVreplicationwasalsosuppressedbyexpres-
control siRNA (Figure 5A). A similar observation was made for sionoftheRIG-IRD(WT)inHuh-7.5cells,whereinanyIFNinduc-
RIG-I siRNA-treated HuS-E/2 cells (Figure S5A). These results tionisnotobserved,whilethemutantRD(K888E)didnotaffect
indicateanimplicatedroleofRIG-Iasaninnatesensortoacti- viral replication (Figure 5F). These findings revealed another
vate antiviral response against HBV infection. On the other aspectofRIG-Iasadirectantiviralfactorthroughitsinterference
hand, it has been previously reported that the 50-ε region of withthebindingofHBVPproteintopgRNAinanIFNpathway-
HBVpgRNAisimportanttoserveasabindingsiteofviralPpro- independentmanner.
tein for initiating reverse transcription (Bartenschlager and
Schaller, 1992). As consistentwith this, weshowed that the P TheεRNARestrictsHBVReplicationinHuman
protein interacts with εRNA in Huh-7.5 and HEK293T cells, by Hepatocyte-ChimericMice
fluorescence resonance energy transfer (FRET) analysis (Fig- Lastly,basedontheaboveresults,wetriedtoharnessthether-
ure 5B) and RNA pull-down assay (Figure S5B), respectively. apeuticpotentialofthePprotein-interactingεRNAforthecontrol
These findings facilitated us to examine whether RIG-I could ofHBVinfection.Avectorwasdesignedtoincludea63bpDNA
block the access of P protein toward the ε region. As we ex- oligo, which is transcribed into an εRNA. We confirmed in the
pected, recombinant RIG-I protein suppressed the interaction invitroexperimentsusingHuh-7.5cellsthatεRNAinducedby
of P protein with pgRNA in a dose-dependent manner (Fig- thisvector-drivenexpressioniscapabletofunctionasadecoy
ure5C).SuchaninhibitoryeffectwasalsoobservedinHuh-7.5 RNA to interfere with the binding of HBV P protein to pgRNA
cells by expression of WT RIG-I, as well as its T55I (Sumpter andtoinhibitviralreplicationinanIFN-independentmanner(Fig-
etal.,2005;Saitoetal.,2007)orK270A(Takahasietal.,2008) ures6Aand6B,left).Ontheotherhand,εRNAdidnotshowany
mutant (Figures 5D and S5C), both of which are not able to difference in HCV replication as compared with control (Fig-
induceligand-dependentactivationofthedownstreamsignaling ure 6B, right). In order to evaluate the therapeutic efficacy of
but retain their RNA-binding activities. On the other hand, the εRNA in vivo, we exploited HBV infection model of human
128 Immunity42,123–132,January20,2015ª2015ElsevierInc.
Figure 5. RIG-I Functions as an Antiviral
FactorbyCounteractingtheInteractionof
HBVPProteinwithpgRNA
(A)qPCRanalysisofcopynumbersofencapsi-
dated HBV DNA (left) and qRT-PCR analysis of
RIG-I(middle)andIFNL1mRNA(right)incontrolor
RIG-IsiRNA-treatedPHHafter10daysofinfection
withHBV-C.
(B)FRETanalysisfortheinteractionbetweenYFP-
taggedPprotein(YFP-P)orYFPandεRNA-ROX
asdescribedinFigure4B.Scalebarrepresents
20 mm. Arrowheads indicate area showing high
FRETefficiency.
(C)HEK293TcelllysatesexpressingpgRNAand
HA-taggedPprotein(HA-P)wereincubatedwith
the indicated amount of recombinant RIG-I
(rRIG-I).TheinteractionofpgRNAwithHA-Pwas
analyzedbyRIPassayandqRT-PCRanalysisas
describedinFigure4C.
(D)CelllysatesfromHuh-7.5cellsexpressingHBV
pgRNA, HA-P,andFlag-RIG-Ioritsmutantsas
indicated were subjected to RIP assay for the
characterization of the capability of RIG-I to
counteracttheinteractionofpgRNAwithHA-P,as
describedinFigure4C.
(E)TheeffectofrIFN-l1treatmentontheinterac-
tion of pgRNA with HA-P in Huh-7.5 cells was
assessedbyRIPassay.Huh-7.5cellsexpressing
bothpgRNAandHA-PweretreatedwithrIFN-l1
(100ng/ml)for24hr,andsubjectedtoRIPassay
asdescribedinFigure4C.RIG-Idependencywas
alsodeterminedbyRIG-Iknockdownanalysis.
(F) Huh-7.5 cells were transfected with an
expressionvectorforRIG-IRD(WT)orRD(K888E),
togetherwiththeHBV-Aegenome.After72hrof
transfection,copynumbersofencapsidatedHBV
DNA were measured (left), as described in (A).
ExpressionofFlag-RIG-IRD(WT)andRD(K888E)
(right).*p<0.05and**p<0.01versuscontrol.NS,
notsignificant.SeealsoFigureS5.
hepatocyte-chimericmice.HBV-infectedmiceunderwentintra- loopofHBVpgRNA(Figures1,2,3,and4).Inthisrespect,there
venousadministrationwiththeεRNAexpressionvectorloaded havealsobeenseveralreportsshowingthatHBVXorPprotein
inaliposomalcarrier,amultifunctionalenvelope-typenanode- interactswithMAVSorcompetesforDDX3bindingwithTBK1,
vice (MEND) for efficient delivery, for 2 weeks. Treatment with respectively (Wei et al., 2010; Wang and Ryu, 2010; Yu et al.,
εRNA-MENDsignificantlysuppressedtheelevationofthenum- 2010), and inhibits RIG-I-mediated type I IFN pathway, which
berofviralgenomecopiesintheserabylessthanonetenthof possibly enables HBV to evade from antiviral innate immune
thoseforcontrolmice(Figure6C).Consistently,immunofluores- response.ThiswouldmirrortheimportantroleofRIG-I-mediated
cenceanalysesshowedthattheexpressionofHBVcoreantigen signaling for antiviral defense against HBV infection, although
(HBcAg) in the liver tissues of εRNA-MEND-treated chimeric furtherinvestigationwillberequiredtodeterminewhetherother
micewasremarkablyreducedascomparedwiththoseofcontrol sensingmoleculesexceptforRIG-Iareengagedintheactivation
mice(Figure6D). of innate responses in other cell types including dendritic cell
subsets. Interestingly, Lu et al. have recently showed that the
DISCUSSION genotypeDofHBVissensedbyMDA5,butnotRIG-I,whichis
basedonlyupontheanalyseswithHBVgenome(2-fold)plasmid
Theinnateimmunesystemactsasafrontlineofhostdefense transfectioninasinglecelllineHuh-7(LuandLiao,2013).Inthis
against viral infection. In this step, PRRs play a crucial role in respect,wepresumethatsuchseeminglycontradictoryresults
the recognition of invading viruses. In particular, nucleic acid mightarisemainlyfromthedifferenceinHBVgenotype:Ithas
sensingofvirusesiscentraltotheinitiationofantiviralimmune beenreportedthatthegenotypeDisphylogeneticallydifferent
responses.Inthisstudy,wetriedtoseekforarelevantnucleic fromthegenotypesA,B,andC,whichweanalyzedinthisstudy
acidsensor(s)forHBVandtocharacterizetheIFNresponsedur- (Katoetal.,2002).
ing HBV infection. As a result, we have identified RIG-I as an Inaddition,accordingtoourresults(Figure1CandS1C),HBV-
important innate sensor of HBV to predominantly induce type inducedtypeIIIIFNresponsedoesnotseemtobeefficientas
IIIIFNsinhepatocytesthroughitsrecognitionofthe50-εstem- comparedwiththecasewithNDVinfection.Wespeculatethat
Immunity42,123–132,January20,2015ª2015ElsevierInc. 129
Figure 6. Inhibition of HBV Replication by
εRNA
(A)Huh-7.5cellsweretransfectedwithexpression
vectors for HA-P and εRNA, together with the
HBV-Ae genome. RIP assay was performed to
evaluate the effect of εRNA on the interaction
between HA-P and pgRNA, as described in
Figure5D.
(B)CopynumbersofencapsidatedHBVDNAin
Huh-7.5cellsexpressingtheHBV-Aegenomeand
εRNA,asdeterminedbyqPCR(left).HCVrepli-
cationinHuh-7.5.1/Rep-Feo-1bcellsexpressing
εRNA,asdeterminedbyluciferaseassay(right).
(C)HBV-infectedmicewereintravenouslyadmin-
istratedwiththeεRNAexpressionvector(εRNA-
MEND)oremptyvector(Control-MEND)loadedin
liposomalcarrieratadoseof0.5mg/kgofbody
weightevery2daysfor14days.SerumHBVDNA
inHBV-infectedchimericmicewasdeterminedby
qPCR(n=3pergroup).Day0indicatesthetimeof
theinitiationofadministration.
(D)Immunofluorescenceimagingwasperformed
forthedetectionofHBcAg (red)andhumanal-
bumin(green)intheliversectionsofHBV-infected
chimeric mice at 14 days after treatment with
εRNA-MEND or Control-MEND as described in
ExperimentalProcedures.Dataarepresentedas
meanandSD(n=3)andarerepresentativeofat
leastthreeindependentexperiments.**p<0.01
versuscontrol.NS,notsignificant.
the weakness of the IFN response during HBV infection might poly-U/UCtractinthe30 nontranslatedregionofHCVgenome
attribute at least in part to these viral evasions from host-cell (Saito et al., 2008) and 50 terminal region of influenza virus
control, which would be supported by our preliminary data genome (Baum et al., 2010) were previously reported to be
showing that one HBV mutant, which generates viral RNAs directly or indirectly critical for viral replication (You and Rice,
including pgRNA but lacks the ability to express whole viral 2008;Huang etal.,2005;Moeller etal.,2012). Inthisrespect,
proteins including HBV X and P proteins, can induce higher one could envisage that such an exquisite targeting by RIG-I
amounts of IFN-l1 than intact HBV (Figure S1G). In relevance would confer a unique machinery to ensure efficient antiviral
withthis,ourpresentdataindicatethattheinteraction ofHBV activitiesofRIG-I.Therefore,RIG-Iislikelytoplaydualrolesas
P protein with the 50-ε stem-loop affects the RIG-I-mediated aninnatesensorandasadirectantiviraleffectorforhostdefense
recognition of viral pgRNA and the subsequent downstream duringviralinfection.
signaling events, which might likely suppress the induction of InrelationtotheevaluationoftheexperimentsshowninFig-
IFN-ls.ThismightprovideanaspectofHBVPproteininterms ures6Cand6D,weadditionallyanalyzedthefollowingpoints:
ofviralevasionfromRIG-Iactivation.Asforthemechanismfor When we treated HepG2 cells with εRNA-MEND or Control-
the preferential induction of type III IFNs in hepatocytes in MEND,inbothcaseswehardlydetectedthemassiveinduction
responsetoHBV,aswellasHCV(Nakagawaetal.,2013;Park of cytokines such as TNF, IL6, and CXCL10 (data not shown).
et al., 2012), we might speculate the existence of a hepato- This was further confirmed by analyzing SCID mice injected
cyte-specific factor(s), which is selectively involved in type III withεRNA-MENDorControl-MEND (datanotshown).Inaddi-
IFNgeneinduction,althoughthisissuemeritsfurtherinvestiga- tion, εRNA-MEND has the specific effect on the replication of
tionincludingepigeneticevaluationofhumanhepatocytes.We HBV, but not HCV in Huh-7.5 cells (Figure 6B). These data
also found that either of the 50-or 30-ε region of pgRNA could suggest that the results (Figures 6C and 6D) might not be
interactwithRIG-Ibutitwasonlythe50-εregionthatcontributed mainlyinfluencedbymassiveproductionofantiviralcytokines,
totheinductionofIFN-l1(Figures3DandS4A).Inthisrespect, although the cross-reactivity of cytokines should be still care-
wepresumethatsomecofactor(s)mightadditionallydetermine fully considered. Therefore, it is presumed that the effect of
thepreferentialuseofthe50-εregionforRIG-Iactivation;how- εRNA might be based on not only its antagonistic activity but
ever,itwouldbeanextinterestingissuetobesolved.Inaddition also its cytokine-inducing activity. These findings might afford
tothis,ourdatademonstratedahitherto-unidentifiedfunctionof a new therapeutic modality in replace of conventional antiviral
RIG-IasadirectantiviralfactoragainstHBVinfection(Figure5). drugsthathavebeenreportedtohavearisktodevelopdrug-
Mechanistically,RIG-Iwasfoundtocounteracttheaccessibility resistance HBV (Song et al., 2012). The present study might
of HBV P protein to the 50-ε stem-loop of pgRNA, which is an provide a better approach to the strategy for development of
important process for the initiation of viral replication (Bar- nucleicacidmedicineandofferanattractiveclinicaloptionfor
tenschlagerandSchaller,1992).Asisthecasewiththis,several thetherapy againstnotonlyHBVbutalsopossibly other virus
viralPAMPsknowntoberecognizedbyRIG-I,forexample,the infections.
130 Immunity42,123–132,January20,2015ª2015ElsevierInc.
EXPERIMENTALPROCEDURES AUTHORCONTRIBUTIONS
InfectionofHumanHepatocyte-ChimericMicewithHBVandInVivo S.S.,K.L.,T.K.,andT.H.carriedoutmostoftheexperimentsandanalyzed
TreatmentwithεRNA data.Y.T.providedmaterialsofHBVanddesignedtheprotocolforinfection.
To generate human hepatocyte-chimeric mice, uPA+/+/SCID mice, were Y.I.conductedHBVinfectioninchimericmiceandpreparationofhumanhepa-
transplantedwithcommerciallyavailablecryopreservedhumanhepatocytes tocytes,andS.M.,T.Watanabe,S.I.,S.T.,andY.T.performedtherelatedanal-
(a 2-year-old Hispanic female; BD Bioscience) as described previously ysisandcontributedtotheinterpretationoftheresults.C.M.R.offeredcritical
(Tateno et al., 2004). Chimeric mice were intravenously infected with adviceonthewholemanuscriptandprovidedapairofHuh-7andHuh-7.5cell
HBV-C (106copiespermouse)derived frompatient withchronic hepatitis lines.K.W.andT.WakitaprovidedHepG2-hNTCP-C4cells.T.K.carriedout
(Sugiyamaetal.,2006).TotalRNAswereisolatedfromlivertissuesat0,4, FRET analysis. Y. Sakurai, Y. Sato, H.A., and H.H. syntheses of plasmid-
or 5 weeks after infection and subjected to quantitative RT-PCR (qRT- loaded MEND,andM.K.contributedtoestablishmentofthe protocols for
PCR). In preparation for εRNA treatment, HBV infection at 3 weeks after in vivo treatment. A.T. supervised the project, designed experiments and
infection was confirmed bymeasuring thenumberofviralgenome copies wrotethemanuscriptwithcriticalinputfromthecoauthors,andallauthors
in the sera of HBV-infected chimeric mice by qPCR analysis, and at contributedtodiscussingtheresults.
4-weekpostinfection,εRNAexpressionvectororemptyvectorloadedina
liposomalcarrier,amultifunctionalenvelope-typenanodevice(MEND),was ACKNOWLEDGMENTS
administered intravenously at a dose of 0.5 mg/kg of body weight (n = 3
pergroup)everytwodaysfor14days.Serumsamplesweresubjectedto WethankM.HijikataandK.ShimotohnoforHuS-E/2cellline,N.Sakamotofor
qPCR for the quantification of DNA copy numbers of HBV as described Huh-7.5.1/Rep-Feo-1bcells,J.MiyazakiforpCAGGSvector,A.Miyawakifor
previously (Nakagawa et al., 2013). All animal protocols described in this pCAGGS-YFPvector,T.Fujitafortheluciferasereporterplasmidp-55C1BLuc,
studywereperformedinaccordancewiththeGuidefortheCareandUse A.TakadaforVSV,H.KidaforNDV,H.IshizuandT.Hirosefortechnicaladvice
ofLaboratoryAnimalsand approved bythe Animal Welfare Committee of onRIPassay,andA.Bergthaler,A.Shlomai,T.Saito,andM.Yasudaforcritical
PhoenixBioCo.,Ltd. readingofthemanuscriptandhelpfuladvice.Theauthorsaregratefulforfinan-
cialsupportsfromtheMinistryofHealth,LabourandWelfareofJapan(Grant-
HBVInfectioninHumanHepatocytesandQuantificationof
in-AidtoA.T.andY.T.),theMinistryofEducation,Culture,Sports,Scienceand
EncapsidatedHBVDNA
TechnologyofJapan(Grant-in-AidforScientificResearch[A][25253030]to
HBVinfectioninPHHorHepG2-hNTCP-C4cellswasperformedat10or100 A.T., Grant-in-Aid for Scientific Research on Innovative Areas [25115502,
genomeequivalentspercell,respectively,inthepresenceof4%PEG8000at 23112701]toA.T.,Grant-in-AidforYoungScientists[B][25870015]toS.S.),
37(cid:3)Cfor24hraspreviouslydescribed(Sugiyamaetal.,2006;Watashietal., IRYOHOJINSHADANJIKOKAI(H.TanakaandN.Takayanagi)toA.T.,the
2013;Iwamotoetal.,2014).Lamivudine(50mM;Sigma)wasaddedinHuh-7 KatoMemorialBioscienceFoundationtoA.T.,theYasudaMedicalFoundation
culturemediaduringHBVproduction.HBVDNAwaspurifiedfromintracel- toA.T.,theTakedaScienceFoundationtoA.T.,andTheWaksmanFoundation
lular core particles as described previously (Turelli et al., 2004; Fujiwara ofJapantoA.T.Y.I.isanemployeeofPhoenixBio.
etal.,2005).Briefly,cellsweresuspendedinlysisbuffercontaining50mM
Tris-HCl (pH 7.4), 1 mM EDTA, and 1% NP-40. Nuclei were pelleted by Received:June23,2014
centrifugation at 4(cid:3)C and 15,000 rpm for 5 min. The supernatant was Accepted:December3,2014
adjustedto6mMMgOAcandtreatedwithDNaseI(200mg/ml)andofRNase Published:December31,2014
A (100 mg/ml) for 3 hr at 37(cid:3)C. The reaction was stopped by addition of
10mMEDTAandthemixturewasincubatedfor15minat65(cid:3)C.Aftertreat- REFERENCES
mentwithproteinaseK(200mg/ml),1%SDS,and100mMNaClfor2hrat
37(cid:3)C,viralnucleicacidswereisolatedbyphenol:chloroform:isoamylalcohol
Ablasser, A.,Bauernfeind,F.,Hartmann,G.,Latz,E.,Fitzgerald,K.A., and
(25:24:1) extraction and ethanol precipitation with 20 mg glycogen. Copy
Hornung,V.(2009).RIG-I-dependentsensingofpoly(dA:dT)throughthein-
numbersofHBVDNAweremeasuredbyqPCRwiththeindicatedprimers
duction of an RNA polymerase III-transcribed RNA intermediate. Nat.
(TableS3).
Immunol.10,1065–1072.
qRT-PCRAnalysis Bartenschlager,R.,andSchaller,H.(1992).Hepadnaviralassemblyisinitiated
bypolymerasebindingtotheencapsidationsignalintheviralRNAgenome.
TotalRNAswereisolatedfromculturecellsorfrozenlivertissueusingIsogen
EMBOJ.11,3413–3420.
(NipponGene)andweretreatedwithDNaseI(Invitrogen).cDNAwasprepared
fromtotalRNAsusingReverTraAce(TOYOBO).qRT-PCRwasperformedus- Baum,A.,Sachidanandam,R.,andGarcı´a-Sastre,A.(2010).Preferenceof
ingSYBRPremixExTaq(TAKARA)andanalyzedonaStepOnePlusreal-time RIG-IforshortviralRNAmoleculesininfectedcellsrevealedbynext-genera-
PCR system (Applied Biosystems). Detailed information about the primers tionsequencing.Proc.Natl.Acad.Sci.USA107,16303–16308.
usedhereisshowninTableS3.Datawerenormalizedtotheexpressionof Chiu,Y.H.,Macmillan,J.B.,andChen,Z.J.(2009).RNApolymeraseIIIdetects
GAPDHorHBVRNAsforeachsample. cytosolicDNAandinducestypeIinterferonsthroughtheRIG-Ipathway.Cell
138,576–591.
RIPAssay
Choi,M.K.,Wang,Z.,Ban,T.,Yanai,H.,Lu,Y.,Koshiba,R.,Nakaima,Y.,
After 2 hr incubation with the antibody as indicated for immunopre-
Hangai,S.,Savitsky,D.,Nakasato,M.,etal.(2009).Aselectivecontribution
cipitation, cell lysates were mixed with Protein-G Dynabeads (Invitrogen)
oftheRIG-I-likereceptorpathwaytotypeIinterferonresponsesactivated
and further incubated for 1 hr with gentle shaking. After washing three
bycytosolicDNA.Proc.Natl.Acad.Sci.USA106,17870–17875.
times,theprecipitatedRNAswereanalyzedbyqRT-PCRwithappropriate
Cui,S.,Eisena¨cher,K.,Kirchhofer,A.,Brzo´zka,K.,Lammens,A.,Lammens,
primers to detect the target RNA. The amount of immunoprecipitated
K., Fujita, T., Conzelmann, K.K., Krug, A., and Hopfner, K.P. (2008). The
RNAsisrepresentedasthepercentileoftheamountofinputRNA(percent-
C-terminal regulatory domain is the RNA 50-triphosphate sensor of RIG-I.
age input). The detail was described in the Supplemental Experimental
Mol.Cell29,169–179.
Procedures.
Fujiwara,K.,Tanaka,Y.,Paulon,E.,Orito,E.,Sugiyama,M.,Ito,K.,Ueda,R.,
SUPPLEMENTALINFORMATION Mizokami,M.,andNaoumov,N.V.(2005).NoveltypeofhepatitisBvirusmuta-
tion:replacementmutationinvolvingahepatocytenuclearfactor1bindingsite
SupplementalInformationincludesfivefigures,fourtables,andSupplemental tandemrepeatinchronichepatitisBvirusgenotypeE.J.Virol.79,14404–14410.
ExperimentalProceduresandcanbefoundwiththisarticleonlineathttp://dx. Halegoua-DeMarzio,D.,andHann,H.W.(2014).Thenandnow:theprogressin
doi.org/10.1016/j.immuni.2014.12.016. hepatitisBtreatmentoverthepast20years.WorldJ.Gastroenterol.20,401–413.
Immunity42,123–132,January20,2015ª2015ElsevierInc. 131