Love, A.J., Geri, C., Laird, J., Carr, C., Yun, B.-W., Loake, G.J., Tada, Y., Sadanandom, A. and Milner, J.J. (2012) Cauliflower mosaic virus protein P6 inhibits signaling responses to salicylic acid and regulates innate immunity. PLoS ONE, 7 (10). e47535. ISSN 1932-620. http://eprints.gla.ac.uk/70627/ Deposited on: 24 October 2012 Enlighten – Research publications by members of the University of Glasgow http://eprints.gla.ac.uk CauliflowermosaicvirusProtein P6 Inhibits Signaling Responses to Salicylic Acid and Regulates Innate Immunity Andrew J. Love1¤a, Chiara Geri1,2, Janet Laird1, Craig Carr1, Byung-Wook Yun3¤b, Gary J. Loake3, Yasuomi Tada4¤c, Ari Sadanandom1¤d, Joel J. Milner1* 1PlantScienceResearchTheme,SchoolofLifeSciencesandInstituteofMolecular,CellularandSystemsBiology,CollegeofMedical,VeterinaryandLifeSciences, UniversityofGlasgow,Glasgow,Scotland,UnitedKingdom,2IstitutodiBiologiaeBiotechnologiaAgraria,ConsiglioNazionaleDelleRicherche,Pisa,Italy,3Instituteof MolecularPlantSciences,UniversityofEdinburgh,King’sBuildings,Edinburgh,UnitedKingdom,4DepartmentofBiology,DukeUniversity,Durham,NorthCarolina, UnitedStatesofAmerica Abstract Cauliflowermosaicvirus(CaMV)encodesamultifunctionalproteinP6thatisrequiredfortranslationofthe35SRNAandalso acts as a suppressor of RNA silencing. Here we demonstrate that P6 additionally acts as a pathogenicity effector of an uniqueandnoveltype,modifyingNPR1(akeyregulatorofsalicylicacid(SA)-andjasmonicacid(JA)-dependentsignaling) andinhibitingSA-dependentdefenceresponsesWefindthatthattransgene-mediatedexpressionofP6inArabidopsisand transient expression in Nicotiana benthamiana has profound effects on defence signaling, suppressing expression of representativeSA-responsivegenesandincreasingexpressionofrepresentativeJA-responsivegenes.Relativetowild-type ArabidopsisP6-expressingtransgenicshadgreatlyreducedexpressionofPR-1followingSA-treatment,infectionbyCaMVor inoculation with an avirulent bacterial pathogen Pseudomonas syringae pv tomato (Pst). Similarly transient expression in NicotianabenthamianaofP6(includingamutantformdefectiveintranslationaltransactivationactivity)suppressedPR-1a transcript accumulation in responseto Agrobacterium infiltration and followingSA-treatment. Aswell assuppressing the expressionofrepresentativeSA-regulatedgenes,P6-transgenicArabidopsisshowedgreatlyenhancedsusceptibilitytoboth virulentandavirulentPst(titreselevated10to30-foldcomparedtonon-transgeniccontrols)butreducedsusceptibilityto thenecrotrophicfungusBotrytiscinerea.NecrosisfollowingSA-treatmentorinoculationwithavirulentPstwasreducedand delayedinP6-transgenics.NPR1animportantregulatorofSA/JAcrosstalk,wasmorehighlyexpressedinthepresenceofP6 and introduction of the P6 transgene into a transgenic line expressing an NPR1:GFP fusion resulted in greatly increased fluorescenceinnucleievenintheabsenceofSA.ThusinthepresenceofP6aninactiveformofNPR1ismislocalizedinthe nucleuseveninuninducedplants.TheseresultsdemonstratethatP6isanewtypeofpathogenicityeffectorproteinthat enhances susceptibility tobiotrophic pathogens bysuppressing SA- butenhancing JA-signalingresponses. Citation:LoveAJ,GeriC,LairdJ,CarrC,YunB-W,etal.(2012)CauliflowermosaicvirusProteinP6InhibitsSignalingResponsestoSalicylicAcidandRegulates InnateImmunity.PLoSONE7(10):e47535.doi:10.1371/journal.pone.0047535 Editor:FrederikBo¨rnke,Friedrich-Alexander-UniversityErlangen-Nurenberg,Germany ReceivedJune26,2012;AcceptedSeptember12,2012;PublishedOctober11,2012 Copyright: (cid:1) 2012 Love et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited. Funding:ThisworkwassupportedbyBBSRC(www.bbsrc.ac.uk)grantsBB/D017319(toJJMandAS)andBB/C005561(toJJM,ASandGJL).AJLwassupportedby BB/D017319;JLwassupportedbyBB/D017319andBB/C005561.Thefundershadnoroleinstudydesign,datacollectionandanalysis,decisiontopublish,or preparationofthemanuscript. CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist. *E-mail:[email protected] ¤aCurrentaddress:CellandMolecularSciences,JamesHuttonResearchInstitute,Invergowrie,Dundee,UnitedKingdom ¤bCurrentaddress:DivisionofPlantBiosciences,SchoolofAppliedBiosciences,CollegeofAgriculture&LifeScience,KyungpookNationalUniversity,Daegu, SouthKorea ¤cCurrentaddress:LifeScienceResearchCenter,KagawaUniversity,Kagawa,Japan ¤dCurrentaddress:SchoolofBiologicalandBiomedicalSciences,DurhamUniversity,Durham,UnitedKingdom Introduction Plant viruses are amongst the least genetically complex pathogensbutmustneverthelessmaintaintheabilitytoovercome It is a paradigm that any micro-organism that functions as an host-defences. Much recent attention has focused on the role of effective pathogenmustpossessmechanismseithertosuppressor RNA-silencing in anti-virus defence and the majority of plant evade the armoury of host defence responses. This is particularly virusesencodesilencingsuppressorproteins(VSSPs)[4].Aswellas true for biotrophic plant pathogens because they can only RNA-silencing, many plant viruses stimulate basal defence maintainasuccessfulinfectioninthepresenceoflivinghostcells. responses in compatible hosts; these include the global activation Plant pathogenic bacteria introduce dozens of pathogen-encoded ofSalicylicAcid(SA)-responsiveandchangesintheabundanceof effectorsintothehostviatheTypeIIIsecretionsystem(TTSS)[1] JasmonicAcid(JA)-responsivegenes[5,6].Exogenousapplication and biotrophic fungal or oomycete pathogens may deliver of SA or analogues can reduce accumulation or long-distance hundreds of effectors across the haustorial membrane [2,3] that movementofanumberofviruses,e.g.Tobaccomosaicvirus(TMV) suppress ormodify plant defence responses. PLOSONE | www.plosone.org 1 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 andCucumbermosaicvirus(CMV[7–9]andlongdistancemovement plants. We also show that in P6-expressing transgenic plants, of Cauliflower mosaic virus (CaMV) is inhibited in Arabidopsis susceptibilitytovirulentandavirulentstrainsofPseudomonassyringae (Arabidopsis thaliana) mutants in which SA-responsive defence pv tomato (Pst) is greatly enhanced, but susceptibility to the pathways are constitutively activated [10]. CMV protein 2b, necrotrophicfungalpathogenBotrytiscinereaisgreatlyreduced.We aVSSP,hasalsobeenimplicatedasmodifyingbothJA-andSA- concludethatP6isanovelanduniquepathogenicityeffectorthat responses implying a possible link between silencing, SA and JA specifically targets basal defence by profoundly altering signaling [6,11], although in Arabidopsis SA-mediated antiviral defence responsestoSA andJA. againstTMVandCMVappearstoactpartiallyindependentlyof DICERactivity[12].Theseresultsdemonstratetheimportanceof Results SA and JA signaling in defence responses following infection by either compatibleor incompatible viruses. Expression of CaMV-P6 from a transgene in Arabidopsis CaMVisaplantpararetroviruswithan8.0kbpDNAgenome and transiently in N. benthamiana suppresses and a worldwide distribution [13]. Hostsinclude members ofthe transcriptional up-regulation of SA-dependent marker Cruciferae [13] and infections of Arabidopsis (Arabidopsis thaliana) genes in response to infection make an excellent model pathosystem with which to study To test if P6 might be interfering with SA-signaling during signaling and defence responses during compatible host-virus CaMV infection we inoculated transgenic Arabidopsis plants interactions. In Arabidopsis, CaMV-infection strongly stimulates whichconstitutively expresses high levels ofP6 [28] withCaMV, activationofgenesresponsivetoethylene,reactiveoxygenspecies and compared the expression of PR-1 (a reliable marker for SA- and SA [14]. Up-regulation of markers such as PR-1, BGL2 and dependent defence) [10,31] with CaMV-infected non-transgenic PR-5isdependentonfunctionalSA-signalingpathways.However, (NT) plants in the same ecotypic (Ler) background. Transcripts although Arabidopsis mutants in which SA-responsive defence were measured by Real Time qPCR in samples harvested at pathwaysareconstitutivelyactivatedshowenhancedresistanceto 14dpi a time at which we have previously showed that SA- CaMV,genotypesthatcannotaccumulateSAe.g.NahGandsid2- responsive markers are strongly up-regulated in CaMV-infected 2 do not show any corresponding enhanced susceptibility [10]. wild-type plants [14]. In a typical P6-transgenic line (A7) virus Therefore, in wild-type Arabidopsis plants (but not in mutants in levelsat14dpi(measuredbyq-PCR)were2664%thoseinNT; which Systemic Acquired Resistance (SAR) is constitutively we have previously reported similarly reduced virus titres in P6- activated) SA-dependent defence responses triggered by CaMV- transgenic lines [26]. In NT plants, CaMV-infection stimulated infectionmaybeineffectiveinrestrictingvirusmultiplicationand the expected large increase in abundance of PR-1 transcripts. In spread. contrast, PR-1 transcripts accumulated in A7 to levels that were Measured on a whole plant basis, SA-responsive transcripts only2.660.3%thelevelsininfectedNTandonlyslightlygreater increaseinabundanceovertimefollowingCaMV-inoculationbut than inuninfected NT plants (figure 1A). Although virus titres in leaf-by-leaf the situation is more complex. In individual leaves at the P6-transgenic plants were lower than in NT, as we have theveryearlystagesofvirusinvasion,levelsofPR-1transcriptsare reported previously levels of PR-1 transcripts are not directly high but they decrease as virus titres within the leaf increase proportionaltovirustitresbutrisesharplyaround8dpi(atwhich [15,16].Aplausibleexplanationforthisstronginversecorrelation timeviruslevelsarestilllow)[14,15];wethereforethinkitunlikely between virus titre and PR-1 expression in individual leaves, and that the large reduction in PR-1 transcript levels in the P6- for the lack of enhanced susceptibility in SA-deficient genotypes transgenic relative to NT lines is attributable to the much more [10], is that a virus-encoded protein may be suppressing SA- modest relative reductionin virustitres. signaling ininfected cells. To determine whether the effect was specific to CaMV- CaMV gene VI encodes P6, a 62kD multifunctional nuclear- infection, we quantified PR-1 transcripts 48h after inoculation cytoplasmic shuttle protein [17] with an essential role in virus with Pseudomonas syringae pv tomato (Pst) (AvrB), which is avirulent replication [13]. P6 acts as a translational transactivator (TAV) on Ler and triggers a strong gene-for-gene response [32]. promotingtranslationofdownstreamopenreadingframes(ORFs) Transcript levels in Pst (AvrB)-inoculated leaves of NT plants on the polycistronic 35S mRNA through a non-canonical showed the anticipated large increase compared to uninoculated mechanism [13,18]. It also associates with actin filaments and controls (figure 1B). In leaves of P6-transgenic plants inoculated playsaroleinviruscell-to-cellmovement[19,20].GeneVIisalso with Pst (AvrB), PR-1 transcripts only accumulated to levels that the major genetic determinant of pathogenicity [21–23] and in were approximately one tenth those in Pst (AvrB)-inoculated NT. some hosts P6 functions as an avirulence determinant triggering Thus, expression of P6 froma transgene inhibitedtheexpression a hypersensitive response (HR) [24,25]. Consistent with a role as of a SA-dependent marker gene both in response to a virus- a pathogenicity determinant, P6 can also act as a suppressor of infectionandinagene-for-gene responsetoa bacterialinfection. RNA-silencing, interfering with DICER activity through an To assess if this novel function was limited to Arabidopsis or interaction with nuclear protein DRB4 [26,27]. Here we have could be manifested in other plant species we took advantage of investigatedwhetherP6mightplayyetanotherrolebysuppressing the ability of Agrobacterium (Agrobacterium tumefaciens) to trigger SA-dependent defence responses. Pathogen-Associated Molecular Pattern (PAMP)-driven defence We have previously reported that expression of P6 from responseswheninfiltratedintoNicotianaleaves[33].Wetransiently a transgene in Arabidopsis results in a symptom-like phenotype expressed P6 in Nicotiana benthamiana by infiltrating leaves with andmodifiesethyleneandAuxinsignalingresponses[28–30].We Agrobacterium carrying an appropriate Ti binary expression now show that in transgenic Arabidopsis and when transiently vector and measured the expression of NbPR-1a a homologue of expressed in Nicotiana benthamiana, P6 dramatically decreases the theSA-responsivemarkergenesNtPR-1ainN.tabacumandAtPR-1 abundance of representative SA-responsive transcripts and in- in Arabidopsis [34]. Transcript levels were measured in leaves creasestheabundanceofJA-responsivetranscripts.P6expression agroinfiltrated with Ti binary plasmid pJO-BJI [28] in which results in increased levels and altered subcellular localization of expression of P6 is driven from a 35S promoter; controls NONEXPRESSOROFPATHOGENESISRELATED1 (NPR1), comprised leaves agroinfiltrated with the empty parent vector a key regulatory protein that controls many aspects of SAR in pJO530 (EV). The expression of P6 in agroinfiltrated leaves was PLOSONE | www.plosone.org 2 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 Figure1.QuantificationofAtPR-1andNbPR-1atranscriptsinresponsetopathogens.(A)PR-1transcriptsinmock-inoculatedandCaMV- infectedNT(Ler)andP6-transgenic(A7)Arabidopsis,determinedbyqPCR.Barsshowmeanlevels(inarbitraryunits)of3independentbiological sampleseachcomprisingpooledtissuefrom3plants.Sampleswereharvested14dpi.Errorbarsshowstandarddeviations.(B)PR-1transcriptsin uninoculatedcontrolsandPst(AvrB)inoculatedNT(Ler)andP6-transgenic(A7)Arabidopsis,determinedbyqPCR.Barsshowmeanlevels(inarbitrary units)of3independentbiologicalsampleseachcomprisingpooledinoculatedleavesfrom3plants.Sampleswereharvested48hafterinfiltration. Error bars show standard deviations. (C) PR-1a transcripts, determined by qPCR, in N. benthamiana leaves harvested 48h after agroinfiltration. Sampleswere(U)uninfiltratedleaves,andleavesinfiltratedwithAgrobacteriumcarryingthefollowingvectors(EV)pJO530,(pJO-BJI)pJO-BJI,(P6myc) pGWB-P6myc,(P6Y305P)pGWB-P6Y305P.Barsshowmeanlevels(inarbitraryunits)of3independentbiologicalsampleseachcomprising3pooled infiltratedleafsections.Errorbarsshowstandarddeviations. doi:10.1371/journal.pone.0047535.g001 confirmed in western blots using anti-P6 antibody; these gave of expression of P6Y305P and P6myc, NbPR-1a transcript levels a strong immunoreactive band of the expected size in leaves were essentially identical to those with pJO-BJI and ,10% the infiltratedwithpJO-BJI,butnotinuninfiltratedcontrolsorleaves levelinplantsagroinfiltratedwithEV.ThereforetheabilityofP6 agroinfiltrated with EV (supplementary figureS1A). Levels of P6 to suppress expression of a SA-responsive marker gene does not in N. benthamiana leaves were slightly higher than in the P6- require functional TAVactivity. overexpressingtransgenicArabidopsislineA7;wehavepreviously shownthesetobesimilartolevelsinCaMV-infectedArabidopsis P6 reduces levels of SA-responsive gene transcripts in [28]. SA-treated and in untreated plants Agroinfiltration with EV stimulated an increase in NbPR-1a TheverylowlevelsofAtPR-1andNbPR-1atranscriptsinplants transcripts of more than three orders of magnitude compared to exposedtopathogenssuggestedthatP6wasmostlikelyinterfering uninfiltrated controls, reflecting the PAMP-driven engagement of with SA-signal transduction. We therefore tested the responsive- basal defence [33]. In leaves agroinfiltrated with pJO-BJI (for ness of P6-transgenic plants to exogenous application of SA. NT transient expression of P6) the NbPR-1a transcripts were more (Ler) and two independent P6-transgenic lines, A7 and B6 were abundant than in the untreated control, but were however sprayedwith1.0mMSA.Wethenquantifiedtranscriptsofthree consistently ,10-fold less abundant than in EV-agroinfiltrated representative SA-responsive marker genes PR-1, BGL2 and leaves(figure1C).TheseresultstakentogetherindicatethatP6is AOX1A at intervals up to 24h after treatment (figure 2A, B and able to suppress representative SA-dependent responses to C). SA-treatment of NT elicited the expected strong time- different pathogens in two unrelated plant species, Arabidopsis dependent increase in PR-1 transcripts (up to a maximum of and N.benthamiana. ,700-fold) and a slightly smaller but somewhat more rapid P6 plays an essential role in CaMV-replication facilitating the increase in BGL2 transcripts. In contrast, in P6-transgenics levels translationofdownstreamORFsonthepolycistronic35SRNAby ofbothtranscriptsremainedverylowfollowingapplicationofSA. translational transactivation (TAV). To determine whether the SA stimulated a more modest increase (up to 4-fold) in AOX1A TAVactivityofP6wasrequiredforsuppressingPAMP-responsive transcript levels in NT; again in P6-transgenics levels increased NbPR-1a expression, we agroinfiltrated N. benthamiana with onlyslightlyfollowingapplicationofSA.Totakeadvantageofthe a construct expressing a mutant form P6Y305P, in which the sensitivity of qPCR, which allows accurate quantification of conserved Tyrosine at amino acid 305 has been substituted for transcriptlevels overa rangeof severalordersof magnitude,and a Proline, a change that abolishes TAV activity [35]. Since to allow accurate comparison of transcript levels in untreated as P6Y305P was constructed in a different binary vector pGWB17 well as SA-treated plants we plotted the data on a log scale [36],weusedasanadditionalcontrolathirdconstructP6mycthat (figure 2D and E). Levels of BGL2 and PR-1 transcripts in expresseswild-typeP6(withaC-terminalmyctag)fromthesame untreatedP6transgeniclineswereabout1to2 logslowerthanin pGWB17 vector. P6 protein levels in P6myc and P6Y305P- untreated NT plants. Following SA treatment transcript levels infiltrated leaves were much lower than in leaves agroinfiltrated showedsimilartime-dependentincreasesrelativetountreatedplantsin with pJO-BJI and although P6myc and P6Y305P proteins were all three backgrounds but at each of the individual time point, readilydetectableinwesternBlotsusingananti-mycantibodythey absolute levels in the P6-transgenic lines were always 1 to 2 logs gaveonlyfaintbandsusinganti-P6antibodiesmakingitdifficultto lower than in NT plants. For AOX1a transcripts, differences in accuratelycompareexpressionofthethreeconstructsinthesame levels between P6-transgenic and NT plants were much smaller blot. To better compare expression from the three constructs we (figure 2C) andcomparisonon a log scalewas not informative. quantified P6 transcripts by qPCR (Supplementary figureS1B) To establish whether a similar effect was occurring in N. Transcript levels from P6myc and P6Y305P were 7% and 15% benthamiana,leaveswereagroinfiltratedwitheitherpJO-BJIorEV. respectivelythosefrompJO-BJIbutdespitethemuchlowerlevels After48h (toallowexpression ofP6)theywere sprayedwithSA PLOSONE | www.plosone.org 3 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 Figure2.QuantificationoftranscriptsofSA-responsivegenesinP6-expressingandnon-expressingplantsfollowingSA-treatment. (A)PR-1atranscripts(inarbitraryunits),determinedbyqPCR,inP6-transgenic(A7andB6)andNT(Ler)Arabidopsisfollowingtreatmentwith1.0mM SA.Eachpointrepresentsmeanlevelsof3independentbiologicalsampleseachcomprisingpooledtissuefrom3plantsharvestedatintervalsfrom 0to24h.Errorbarsshowstandarddeviations.(B)BGL2transcripts(inarbitraryunits),determinedbyqPCR,inP6-transgenic(A7andB6)andnon transgenic(Ler)Arabidopsisfollowingtreatmentwith1.0mMSAasinfigure2Aabove.(C)AOX1atranscripts(inarbitraryunits),determinedbyqPCR, inP6-transgenic(A7andB6)andnontransgenic(Ler)Arabidopsisfollowingtreatmentwith1.0mMSAasinfigure2Aabove.(D)Datapresentedin (A)butwithabundanceofPR-1atranscriptsplottedonalogarithmicscale.Errorbars(positiveonly)showstandarddeviations.(E)Datapresentedin (B)butwithabundanceofBGL2transcriptsplottedonalogarithmicscale.Errorbars(positiveonly)showstandarddeviations.(F)PR-1atranscripts, determined by qPCR, in N. benthamiana leaves followingtreatmentwith 1.0mM SA. Leaveswere infiltrated with Agrobacterium carrying either emptyvector(pJO530)orabinaryvectorexpressingP6(pJO-BJI)asinfigure1Cabove.After48hleavesweresprayedwith1mMSA;controlswere left untreated. Tissue was harvested 12h later and PR-1a transcripts were quantitated by qPCR. Bars show mean levels (in arbitrary units) of 3 independentbiologicalsampleseachcomprising3pooledinfiltratedleafsections.Errorbarsshowstandarddeviations. doi:10.1371/journal.pone.0047535.g002 andNbPR-1atranscriptswerequantified12hlater(figure2F).In toleavestransientlyexpressingP6(similartotheresultsshownin agroinfiltrated plants not treated with SA, NbPR-1a transcript figure1C).SA-treatmentstimulatedafurther4-to5-foldincrease levelswereapproximately10-foldhigherinEVcontrolscompared in transcripts (over and above the PAMP-driven response to PLOSONE | www.plosone.org 4 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 Agrobacterium) in both EV and P6-expressing plants. However, P6 expression enhances expression of JA-responsive levels of NbPR-1a transcripts in SA-treated leaves that were markers transiently expressing P6 were still ,6-fold lower than in SA- SA- and JA-dependent responses are coordinately but antago- treatedEV-plants(andonlyabouthalfthelevelinEV-plantsthat nistically regulated [37] and transgene-mediated expression of had not been SA-treated). Therefore in N. benthamiana as with other VSSPs has been reported to lead to changes in the Arabidopsis, accumulation of transcripts of a representative SA- expressionofJA-responsivegenes[6,38].Wethereforequantified responsive marker gene was greatly reduced in leaves transiently transcripts of three representative JA-responsive marker genes expressing P6,although expression remainedSA-responsive. VSP1, VSP2 and THI2.1, plus AOS1 which encodes a JA-bio- NTArabidopsisalsorespondedtoSA-treatmentbydeveloping synthetic enzyme. Transcripts were between 14- and 160-fold necroticpatchesonleaves;thesewereeasilyidentifiable24hafter more abundant in the P6-transgenics than in NT (figure 4). To SA-treatmentwhentheleaveswerestainedwithTrypanBlue.In determine whether the P6-transgenic plants retained JA-respon- contrast P6-transgenic lines never developed necrotic lesions sivenessoverandabovethealreadyenhancedlevelsofexpression, following SA-treatment (figure 3A). This suggests that in the we treated plants with 10mM JA and quantified transcripts over presenceofP6,pathwaysinvolvedinthepromotionofcelldeath the subsequent 24h. JA-treatment stimulated an increase in in responsetoSAare inhibited. abundance of all four transcripts both in P6 transgenic lines and NT(figure4–tofacilitatecomparisonoftranscriptlevels,dataare plotted on a log scale). However, transcripts of VSP1, VSP2, and THI2.1 were consistently ,2 logs more abundant in the P6- transgenicsthaninNTthroughoutthe24hafterJAtreatment.In contrast, with AOS1 although JA treatment stimulated an early increaseinabundanceintheP6transgenics,thiswasmuchmore modestthaninNT,sothatby3 haftertreatmenttranscriptlevels inP6-transgenicsandinNTweresimilar.VSP1,VSP2andTHI2.1 are all regulated downstream of JA (via the COI1 pathway) whereas AOS1 encodes an enzyme required for JA biosynthesis whoseexpressionispresumablyfeedback-regulatedbyJAlevels– thismayaccount forthedifferences inresponse characteristics. Expression of P6 suppresses gene-for-gene and basal defences against P. syringae, but enhances resistance against a necrotrophic fungal pathogen SA is a central regulator of defence against biotrophic pathogens. Given the inhibitory effect of P6 on transcript levels ofSA-responsivegenes,weanticipatedthatexpressionofP6might enhancesusceptibilitytobiotrophicpathogensotherthanCaMV. Wethereforemeasuredgrowthofvirulent(DC3000)andavirulent (AvrB)strainsofPstandofthehrpAmutantthatcarriesadefectin theTTSS[39]andisunabletodelivereffectorsintothehostcell. Bacterial titres were determined in P6 transgenics and in NT at intervals up to 4 d after infiltration (figure 5). Titres of both virulentandavirulentstrainswereconsistently5to30-foldhigher in P6-transgenics than in NT. The differences in titre were somewhat dependent on time after inoculation. However, multi- variate ANOVA indicated that across all three time points bacterial growth rates were very significantly higher in A7 than in NT (p,0.001 for DC3000 and p,0.01 for AvrB). The hprA mutantgrewpoorlyinbothNTandA7buttitresinthelatterwere still typically elevated by about 5-fold. Statistical analysis again indicated that these differences in growth were highly significant across all time points (p,0.001). As a further control, we compared bacterial growth in A7 with growth in sid2 a mutant which lacks a functional ISOCHORISMATE SYNTHASE 1 gene (making it unable to synthesize SA in response to infection) and exhibits a well-documented enhanced susceptibility to Pst [40]. TitresofbothavirulentandvirulentPstweresignificantlyhigherin Figure 3. Development of necrosis (HR) on leaves of NT (Ler) and P6-transgenic (A7 & B6) Arabidopsis. (A) Effect of SA- A7 than in sid2-2 (p,0.01), particularly for the virulent isolate treatment. Panels show from top to bottom: untreated leaves and DC3000 (supplementary figureS2A), suggesting that additional treated with 1.0mM SA. (B) Effect of inoculation with virulent or factors over and above the suppression of SA-signaling may be avirulent Pst. Panels show from top to bottom, uninfiltrated controls, involved intheenhanced susceptibility of P6-transgenics. leavesinfiltratedwithPst(DC3000)andleavesinfiltratedwithPst(AvrB). P6 transgenic plants of line A7 are chlorotic and dwarfed. To BlackarrowsindicateareasofleafinfiltratedwithPstAvrB(bottomrow). NecrosiswasvisualizedbystainingleaveswithTrypanBlue24hafter eliminate any possibility that the increased susceptibility to Pst SA-treatmentorinfiltrationwithPst. mightbedependentontheinsertionsiteofthetransgeneorwholly doi:10.1371/journal.pone.0047535.g003 orpartiallyanindirectconsequenceofthedwarfphenotyperather PLOSONE | www.plosone.org 5 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 Figure4.QuantificationoftranscriptsofJAresponsivegenesinP6-transgenicandnontransgenicplants.(A)VSP1(B)VSP2,(C)THI2.1 and(D)AOS1transcriptlevelsdeterminedbyqPCR,inP6-transgenic(A7andB6)andnontransgenic(Ler-0)Arabidopsisfollowingtreatmentwith 10mM JA. Transcript levels are in Arbitrary Units and are plotted on a logarithmic scale. Each point represents mean levels of 3 independent biologicalsampleseachcomprisingpooledtissuefrom3plantsharvestedatintervalsfrom0to24h.Errorbars(positive)showstandarddeviations. doi:10.1371/journal.pone.0047535.g004 thantheinhibitionofSA-signaling,wemeasuredbacterialgrowth delayed by 6–12h and infiltrated areas stained less strongly with in a different P6-transgenic line B6 and also in line b2-3, TrypanBlue(figure3B)indicatingthatP6expressionreducesthe aderivativeofA7thatwasobtainedfromasuppressorscreen[29]. HRtriggered bya gene-for-gene interaction anddelaysitsonset. This line is homozygous for a recessive mutant allele cse-2 at an In contrast to biotrophs, defence against necrotrophs is independent locus to the transgene. The presence of cse-2 greatly regulated predominantly through JA. Since transcripts of JA- ameliorates thechloroticdwarf phenotypeof theparent A7[29]. responsivemarkerswereelevatedinP6-transgenics,weinoculated Compared to NT, B6 showed the same significantly enhanced plantswiththenecrotrophicfungusB.cinereaandestimatedfungal susceptibilitytobothDC3000andAvrBasdidA7(supplementary growth visually (figure 5D).Compared toLer,bothP6-transgenic figure S2B) demonstrating that the effect is similar in two lines showed visibly reduced growth of fungal hyphae and much independent transgenic lines. Bacterial growth in cse-2 (in a NT less chlorosis at 5 dpi, indicating that they were much less Lerbackground)wassimilartoLerwhereasinb2-3(cse-2inanA7 susceptible to infection. This is consistent with P6 enhancing JA- background) titres were essentially identical to those in A7 signaling responses. (Supplementary figureS2B). Therefore, the increased susceptibil- ity of the P6-transgenics isnot attributable to thechlorotic dwarf P6-transgenic plants accumulate lower levels of SA, but phenotype. retain pathogen-responsiveness Since development of SA-induced necrosis was delayed in P6- The differences in gene-expression and pathogen-susceptibility transgenicplants,weinoculatedwithahightitreofPst(AvrB)and between P6-transgenic and NT plants following SA- and JA- assessed the development of HR in response to an incompatible treatments could reflect effects on SA-biosynthesis or accumula- pathogen. NT developed a fairly rapid HR, detectable by strong tion. We measured levels of free SA and SA-b-glucoside autofluorescenceunderUV-illumination18hafterinfiltrationand (conjugated SA) in infected leaves of NT and P6-transgenics at 24h by a visible necrosis readily identifiable by Trypan Blue beforeand48hafterinoculationwithavirulentPst(AvrB)(figure6). staining.Incontrast,inP6transgenicplantsautofluorescencewas ConcentrationsoffreeSAinuninoculatedP6-transgenicswere4-6 PLOSONE | www.plosone.org 6 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 perleaf)ofPst(AvrB)atintervalsfrom0to4daysafterinfiltrationof leaveswith1.26103cfuofbacteria.Titresareplottedonalogscaleand show mean and standard deviation of colony numbers from 10 individualleaves.(C)Titres(incfuperleaf)ofPsthrpAatintervalsfrom 0to4daysafterinfiltrationofleaveswith1.26103cfuofbacteria.Data areplottedonalogscaleandshowmeanandstandarddeviationof colony numbers from 10 individual leaves. (D) Photographs of representativeleavesfromnon-transgenic(Ler)andP6-transgenic(A7 andB6)plants5daysafterinfectionwithB.cinerea. doi:10.1371/journal.pone.0047535.g005 foldlowerthaninNTandlevelsofSA-b-glucosidewere3–4-fold lower.Infectionstimulateda1.4–1.8-foldincreaseinlevelsoffree SAinP6transgenics,anda2.2-foldincreaseinNT.LevelsofSA- b-glucoside in the P6-transgenics increased 10-fold following infection and although levels remained below those in infected NT,theywerestillmuchhigherthaninuninfectedNT.ThusP6 expression reduces the overall accumulation of both free and conjugatedSAininfectedplants,althoughlevelsinP6-transgenic plants appeared to retain inducibility in response to a gene-for- geneinteraction. P6 expression alters the expression and subcellular localization of NPR1 NPR1 is a key regulator of host responses to infection by biotrophic pathogens [41,42]. NPR1 plays a central role in regulating many of the responses to both SA and JA, notably by activating transcription of a battery of genes in response torising SA-levels and by modulating JA responses via the COI1- dependent pathway [42,43]. npr1 mutants show enhanced susceptibility to both virulent and avirulent strains of Pst and accumulate very low levels of PR-1 and BGL2 transcripts in response to infection or SA-treatment [44]. The similarities between these aspects of the phenotypes of npr1 mutants and P6 transgeniclinessuggestedapotentialinvolvementforNPR1inthe action of P6. We therefore compared levels of NPR1 in P6- transgenic and NT plants in western blots probed with an anti- NPR1antibody(figure7A).Abandof,60kDcorresponding to theexpectedsizeof NPR1wasvisibleinP6-transgenicsandvery faintly in NT plants. Following treatment with SA the relative intensity of the band was increased in both transgenic and NT plants,butitremainedmuchstrongerandappearedasadoublet in the P6-transgenic lines. We assume that the band(s) must correspond to authentic NPR1 since they were undetectable in extracts froma null mutantnpr1-1even afterSA-treatment. Todetermine whether increased levels ofNPR1 proteinmight be a result of greater transcript levels we measured NPR1 transcripts by qPCR and found that they were approximately 7- fold more abundant in P6-transgenics than in Ler (figure 7B). To determinewhetherP6mightbeactingattheproteinlevel,e.g.by stabilizing NPR1 against proteolysis, we transiently co-expressed HA-taggedNPR1andeitherP6mycorasacontrolEV(pGWB17) in N. benthamiana. NPR1 levels were quantified in western blots using an anti-HA antibody (figure 7C). The abundance of NPR1:HA was unaltered when co-expressed with P6 compared toEVco-expression controls;thereforeP6 doesnotappeartobe affecting protein stability. Most likely the greater levels of NPR1 when P6 is present are a consequence of enhanced gene- expression. NPR1 undergoes a redox-dependent translocation from cyto- Figure 5. Pathogen growth in P6-transgenic (A7, B6) and NT plasmtonucleusinresponsetoincreasingconcentrationsofSA;it (Ler) Arabidopsis. (A) Titres (in cfu per leaf) of Pst (DC3000) at is the presence of activated NPR1 in the nucleus that activates intervalsfrom0to4daysafterinfiltrationofleaveswith1.26103cfuof transcription of genes of the PR1 regulon [45]. To investigate bacteria.Titresareplottedonalogscaleandshowmean(6standard whetherP6mightbeaffectingthelocalizationand/oractivationof deviation)ofcolonynumbersfrom10individualleaves.(B)Titres(incfu NPR1, we used a transgenic reporter line that constitutively PLOSONE | www.plosone.org 7 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 Figure7.TheeffectsofexpressionofP6onNPR1.(A)Western blotsofproteinextractedfromnpr1mutant,P6-transgenic(A7,B6)and Figure 6. Levels of free and conjugated SA in P6-transgenic NT(Ler)plants,separatedbypolyacrylamidegelelectrophoresis.Tissue andnon-transgenicArabidopsisfollowinginoculationwithPst was harvested from plants either before (4 lanes on left) or 12hours (AvrB).(A)freeSA,(B)SA-conjugates(SA-b-glucoside)inleavesofNT after(4lanesonright)treatmentwith1.0mMSA.Upperpanelshows (Ler)andP6-transgenic(A7andB6)plantsinoculatedonasingleleaf blots probed with antibody to NPR1 and bands visualized by with 1.26103cfu of bacteria. Leaves from uninoculated plants were chemiluminescence.Barsonleftindicatemobilityofmolecularweight used as controls. Each sample comprised the pooled tissue from 10 markers;arrowindicatesexpectedmobilityofNPR1.Lowerpanelshows leavesharvested48hafterinoculation.Barsshowmeanvaluesfrom3 Ponceau-stained loading control; arrow indicates mobility of Rubisco samples,errorbarsindicatestandarderror. LargeSubunit(RBCL).(B)WesternblotsshowingNPR1accumulatingin doi:10.1371/journal.pone.0047535.g006 N. benthamiana leaves following agroinfiltration with a binary vector expressing NPR1 (HA-tagged at the N-terminus) under the control of a35Spromoter.Upperpanelshowsblotsprobedwithanti-HAantibody expresses a NPR1:GFP fusion [46] to monitor its subcellular and bands visualized by chemiluminescence. Lower panel shows location before and after SA-treatment. Under our growth Ponceau-stainedloadingcontrol–arrowindicatesmobilityofRubisco conditions, mature 35S::NPR1:GFP transgenic plants were dark LargeSubunit(RBCL).(Lane1)HA:NPR1co-infiltratedwithemptybinary green and dwarfed with necrotic micro-lesions, a phenotype vector pGWB17; (lane2) HA:NPR1 co-infiltrated with P6 expressing reminiscent of mutants with constitutively activated SA-mediated binaryvectorpGWB-P6myc.(C)NPR1transcriptlevels(inarbitraryunits) determinedbyqPCRinP6-transgenic(A7andB6)andNT(Ler)plants. responses e.g. cpr5 [47]. Progeny of crosses between Barsrepresentmeanlevels of3 independentbiologicalsampleseach 35S::NPR1:GFP and either of the P6-transgenic lines were larger comprising pooled tissue from 3 plants. Error bars show standard thanbothparentsanddidnotdevelopnecroticlesions.Evidently deviations.(D) Confocalmicroscope imagesof representative pairs of the P6 transgene suppresses the cpr5-like phenotype which guard cells from transgenic plants expressing an NPR1:GFP fusion. presumably results from constitutive overexpression of Panels a–c are in a NT background; panels d-f are in a P6-transgenic NPR1:GFP. In turn, the latter suppresses the chlorotic dwarf background (the progeny of a typical cross between the NPR1:GFP transgenic and B6). Panels a & d show samples from uninduced phenotype whichresultsfromectopic overexpression of P6. seedlings(infiltratedwithwater).Panelsb&eandc&fshowsamples Cotyledons from soil-grown seedlings were infiltrated with from seedlings following 5min and 40min (respectively) infiltration 1 mMSAtoinducedefenceresponsesandexaminedbyconfocal with1.0mMSA.Arrowsindicatethenucleus. microscopy to determine the subcellular location of NPR1:GFP. doi:10.1371/journal.pone.0047535.g007 Uninduced controls were infiltrated with water. figure 7D shows PLOSONE | www.plosone.org 8 October2012 | Volume 7 | Issue 10 | e47535 DefenceSuppressionbyCaMVP6 representative imagesofguardcells (asusedtoshownuclearand a compensatory relationship between RNA-silencing and stress- cytoplasmic localization of NPR1:GFP by Mou et al. [46]). In and defence-response pathways. However, three other VSSPs uninducedplantswiththewild-typebackground,fluorescencewas CMV2bandthegeminivirus-encodedC2andbC1,allsuppressed visible in the cytoplasm with no obvious accumulation in the JA-responsive gene expression [6,48,49], exactly the opposite nuclei. Following 5 min treatment with SA, much of the effect to HC-Pro (and P6). Therefore, it is probably an over- fluorescencewasnowlocatedinthenuclei,andby40minnuclei simplificationtodirectlylinksuppressionofsilencingwithageneral appeared strongly fluorescent (figure 7D panels a, b and c). To up-regulation of stressand defence responses. confirmthatthesiteofNPR1localizationwasindeedinnuclei,we P6playsanessentialroleinCaMVreplicationbyfacilitatingthe carried out DAPI (49,6-diamidino-2-phenylindole) staining and non-canonicaltranslationofdownstreamORFsonthe35SRNA observed colocalization of GFP and DAPI fluorescence in SA- [13]. Since P6Y305P, a mutant form that is defective in TAV induced plants (supplementary figureS3). When we repeated the activity, suppressed PAMP-responsive expression of NbPR-1a as infiltrations on several independent progeny of crosses between efficiently as wild-type P6, TAV activity (which localizes to the 35S::NPR1:GFPandeitherA7orB6(i.e.seedlingscontainingboth central region of P6) does not appeared to be required for the P6 and 35S::NPR1:GFP transgenes), the nuclei always exhibited suppressionofSA-signaling.Replication(TAV)andpathogenicity strongGFPfluorescenceevenintheuninduced(water-infiltrated) functions may therefore be at least partially separated. Recent leaves(paneld);SA-treatmenthadnoobviousfurthereffectonits work in our laboratory (C. Carr, J. Laird and JJ Milner, location (panels e and f). We conclude that P6 stimulates the unpublisheddata,2012)indicatesthatsequencesthatareessential accumulation of enhanced amounts of NPR1, which are for both of the pathogenicity functions (suppression of silencing apparently targeted to the nucleus even in uninduced plants and suppression of PR1a expression) map within the N-terminal (althoughpresumablyinaformthatcannotstimulatetranscription 112 amino acid domain. Thus the two activities may be of PRgenes). functionally linked. Kobayashi and Hohn [51,52] introduced in- frame deletions tothe same N-terminal domain resulting in virus Discussion mutants that were replication-competent but deficient in long- distance movement – again this is consistent with our proposed ExpressioninplantaofP6,apathogenicitydeterminantencoded link between SA-dependent defence responses and long-distance by CaMV, strongly inhibited SA-dependent responses including movement of CaMV. The N-terminal region of P6 is an the expression of representative pathogenesis-related genes, avirulence domain in some Arabidopsis ecotypes and in at least hypersensitive cell death, and basal defence against biotrophic twoNicotianaspecies[20,24,25],consistentwiththezig-zagmodel bacterialpathogens.Wehavepreviouslyshowedthatlongdistance for the evolution of defence mechanisms in plants [53] whereby movementofCaMVisalmostcompletelyinhibitedinArabidopsis effectorsthatsuppressbasaldefencethemselvesbecometargetsfor mutantsthataccumulatehighlevelsofSAandshowconstitutively effector-triggered immunity. activated SAR [10]. Therefore SA-dependent defence responses, P6interactsdirectlywithnuclearproteinDRB4andisbelieved in particularly when they are pre-engaged (as in these mutants), to suppress RNA-silencing by modifying DICER activity [27]. If mustbecapableofrestrictinglongdistancemovementofCaMV. silencingsuppressionandSA-signalingsuppressionarelinked,itis During natural CaMV-infections of wild-type plants an effect of not clear which activity underlies the other. Shivaprasad etal. theaccumulationofP6ininfectedleavesmightbetosuppressthe reported that P6 interferes with RDR6-dependent siRNA and activation of these SA-dependent defence responses thereby tasiRNApathways[54],andthatsmallRNApopulationsundergo facilitating systemic spread. global changes during CaMV-infection [54–56]. Some of these P6 is a multifunctional protein one of whose activities is smallRNAsmightbeinvolvedinregulatingSA-andJA-signaling suppressing RNA silencing [26,27]. In addition to affecting SA- therebyprovidingamechanismthroughwhichP6mightsuppress responses,P6alsohadaprofoundeffectonJA-signaling,strongly SA-signaling. This hypothesis is consistent with our observation up-regulatingrepresentativeJA-responsivemarkergenes.SA-and that in the presence of P6, despite overall levels of SA- and JA- JA-signaling pathways are regulated in a mutually antagonistic responsivetranscriptsbeingprofoundlyaffected,responsivenessto manner through mechanisms that at least partially rely on the exogenousapplicationofSAorJAwasstillpresent.Analternative activity of NPR1 [37,43]. There have been several reports of model puts SA-signaling as the primary target for P6, silencing VSSPs modifying JA-responsive gene-expression. Transgene- suppression perhaps being a secondary consequence of an effect mediated expression of two geminivirus VSSPs C2 and bC1 on the SA-responsive RNA-amplification loop. RDR1, an SA- suppressed JA-dependent responses, albeit by different mechan- responsive RNA-dependent RNA polymerase [57] is a potential isms [48,49]. HC-Pro from a potyvirus and 2b from CMV had link between silencing and SA-mediated defence [11]. However profound effects on the patterns of expression of JA-responsive although RDR1 plays an important role in the generation of genesintheirhosts[6,38,50].However,incontrasttoP6,noneof siRNAsinCMV-infectedplants[58,59],RDR6,whoseexpression theseVSSPssignificantlyalteredtheglobalpatternsofexpression is not SA-responsive, may be the more important target for P6 of SA-responsive genes and they seemed to be more specifically [54]. targetedatJA-signaling.P6thereforeappearstobeuniqueinthis respect, being the prototype member of a novel class of Sequence analysis of P6 predicts several putative functional pathogenicity effectors that that directly targets SA-signaling. P6 domains (e.g. RNA-binding, Zn-finger) but no homologues are accumulates to high levels during CaMV infection [13,28] known outside the closely related members of the Caulimovirus providing a plausible explanation for our previous observation and Soymovirus families of the Caulimoviridae. Apart from these thatinindividualleaves,virusinvasioncoincideswithadecreasein small groups of viruses, the TAV-dependent translation of theabundance of transcripts ofSA-responsive genes [15]. a polycistronic mRNA has no obvious equivalent elsewhere in Transgene-mediated expression of VSSPs can have pleiotropic eukaryotes.AlthoughTAVanddefencesuppressiondomainsmay effects on developmental and stress responses. Endres etal. [38] have evolved independently and their evolutionary lineage suggested that the global activation of wounding-, JA-, cold- and remainsobscure, P6appearstobeafunctionallyandstructurally heat-responsive genes by potyvirus HC-Pro might indicate unique protein. PLOSONE | www.plosone.org 9 October2012 | Volume 7 | Issue 10 | e47535