This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
ARGONAUTE2 Mediates RNA-Silencing AntiviralDefenses against Potato virus X in Arabidopsis1[W][OA]
Marianne Jaubert, Saikat Bhattacharjee, Alexandre F.S. Mello, Keith L. Perry, and Peter Moffett*
Boyce Thompson Institute for Plant Research, Ithaca, New York 14853 (M.J., S.B., P.M.); Department of PlantPathology and Plant-Microbe Biology, Cornell University, Ithaca, New York 14853 (A.F.S.M., K.L.P.); andCentre de Recherche en Amelioration Vegetale, Departement de Biologie, Universite de Sherbrooke,Sherbrooke, Quebec, Canada J1K 2R1 (P.M.)
RNA-silencing mechanisms control many aspects of gene regulation including the detection and degradation of viral RNAthrough the action of, among others, Dicer-like and Argonaute (AGO) proteins. However, the extent to which RNA silencingrestricts virus host range has been difficult to separate from other factors that can affect virus-plant compatibility. Here we showthat Potato virus X (PVX) can infect Arabidopsis (Arabidopsis thaliana), which is normally a nonhost for PVX, if coinfected with asecond virus, Pepper ringspot virus. Here we show that the pepper ringspot virus 12K protein functions as a suppressor of silencingthat appears to enable PVX to infect Arabidopsis. We also show that PVX is able to infect Arabidopsis Dicer-like mutants,indicating that RNA silencing is responsible for Arabidopsis nonhost resistance to PVX. Furthermore, we find that restriction ofPVX on Arabidopsis also depends on AGO2, suggesting that this AGO protein has evolved to specialize in antiviral defenses.
The outcome of a plant-virus interaction is deter-mined by the balance between the ability of the virus toexploit host cellular mechanisms, and the multipledefense mechanisms employed against it by the host.As obligate parasites, viruses are unique in that theirgenetic material is transcribed, translated, and repli-cated entirely inside the host cell. As such, the inabilityof a virus to infect a given plant may be due to anincompatibility between viral proteins and the hosttranslational machinery (Robaglia and Caranta, 2006),inhibition of viral replication (Ishibashi et al., 2007, 2009)and movement (Chisholm et al., 2000; Whitham et al.,2000; Cosson et al., 2010), or by detection and elimina-tion by disease resistance (R) proteins (Moffett, 2009).
Amajor defense mechanism targeting foreign nucleicacids is based on RNA silencing (Ding and Voinnet,2007). RNA silencing refers to a number of relatedcellular processes that employ small RNAs to regulatethe expression of genetic material in a sequence-specificmanner. Small RNAs are generated from dsRNAprecursors by DICER-LIKE (DCL) RNase III-relatedenzymes and subsequently incorporated into RNA-induced silencing complexes (RISCs). The functions of
RISCs are carried out in large part by the activity ofRNAse H-like Argonaute (AGO) proteins. AGO pro-teins bind small RNAs, using them as guides to interactwith homologous RNA molecules, and subsequentlyeffect endonuclease activity, translational repressionof mRNAs, or DNA methylation (Voinnet, 2009). Thehighly structured or dsRNA of viruses is targeted byDCL proteins to generate virus-derived small interfer-ing RNAs (vsiRNAs). These vsiRNAs can be incorpo-rated into virus-induced RISC complexes that targetviral RNAs, thus making RNA silencing a doubly ef-fective antiviral mechanism (Ding and Voinnet, 2007;Omarov et al., 2007). Furthermore, small RNAs canserve as primers for host RNA-dependent RNA poly-merases to generate additional dsRNA targets for DCLenzymes to amplify the silencing signal (Voinnet, 2008;Vaistij and Jones, 2009; Garcia-Ruiz et al., 2010).
In response to the pressure exerted by RNA silenc-ing, plant viruses have evolved viral suppressors ofRNA silencing (VSRs; Diaz-Pendon and Ding, 2008;Alvarado and Scholthof, 2009). Mixed infection by twoviruses can extend the host range, increase virus accu-mulation, and exacerbate viral symptoms (Lathamand Wilson, 2008). In certain cases, this synergisticeffect has been shown to be due to the action, in trans,of the VSR of one of the coinfecting viruses, likelybecause the VSRs of different viruses can target dif-ferent RNA-silencing components (Wu et al., 2010). Anexample of such a synergism is illustrated by thehelper component proteinase from potyviruses thatenhances the accumulation and systemic spreading ofPotato virus X (PVX) in tobacco (Nicotiana tabacum) andsweet potato (Ipomoea batatas; Pruss et al., 1997; Sonodaet al., 2000).
1 This work was supported by the National Science Foundation(grant no. IOB–0343327 to P.M.).
* Corresponding author; e-mail [email protected] author responsible for distribution of materials integral to the
findings presented in this article in accordance with the policydescribed in the Instructions for Authors (www.plantphysiol.org) is:Peter Moffett ([email protected]).
[W] The online version of this article contains Web-only data.[OA] Open Access articles can be viewed online without a sub-
1556 Plant Physiology�, July 2011, Vol. 156, pp. 1556–1564, www.plantphysiol.org � 2011 American Society of Plant Biologists www.plantphysiol.orgon March 9, 2020 - Published by Downloaded from
The Arabidopsis (Arabidopsis thaliana) genome en-codes four DCL and 10 AGO proteins, suggesting asignificant degree of diversification and specializationof RNA-silencing components (Vaucheret, 2008). DCL4and DCL2 have been shown to play major, redundantbut hierarchical, roles in silencing RNA viruses, whileDCL3 also appears to play a minor role in antiviraldefenses (Blevins et al., 2006; Bouche et al., 2006; Deleriset al., 2006; Fusaro et al., 2006; Donaire et al., 2008; Quet al., 2008; Garcia-Ruiz et al., 2010). As such, mutationof dcl2 and dcl4 allows VSR-defective viruses to estab-lish systemic infection and increases susceptibility toTobacco rattle virus (TRV; Deleris et al., 2006; Garcia-Ruizet al., 2010). With respect to the Arabidopsis AGOgenes, hypomorphic ago1 mutants were found to besomewhat more susceptible to Cucumber mosaic virus(CMV; Morel et al., 2002) whereas infection of ago2mutants with Turnip crinkle virus (TCV) and CMV wasshown to result in more severe symptoms, but only asmall and transient effect on virus accumulation (Harveyet al., 2011). At the same time, infection of ago1 and ago7mutants with VSR-defective derivatives of TCV indicatescomplementary roles for AGO1 and AGO7 in antiviraldefenses (Qu et al., 2008; Azevedo et al., 2010). Addition-ally, Ago4-like proteins have been implicated in virusdefense, but only upon induction of an R-gene-mediatedresponse (Bhattacharjee et al., 2009).RNA silencing may also be an important mechanism
of nonhost resistance against viruses, whereby a plant isresistant to all known isolates of a given pathogen.However, the extent to which this may be true has notbeen investigated. PVX is the typemember of the genusPotexvirus (Martelli et al., 2007) and infects severalsolanaceous hosts, but Arabidopsis is considered anonhost for PVX. However, an Arabidopsis-infectingstrain of PVX (PVX 2.7) has been described (Cooperet al., 2003) and used in microarray expression studies(Whitham et al., 2003). We show here that PVX 2.7 caninfect Arabidopsis because of the presence of a secondvirus, Pepper ringspot virus (PepRSV). PepRSV is atobravirus (MacFarlane, 1999, 2010) and as expected,it encodes a VSR, which we demonstrate to be thePepRSV 12K protein. To demonstrate that PVX restric-tion in Arabidopsis is due to RNA-silencing mecha-nisms we show that PVX is able to infect a dcl2 dcl3 dcl4triple mutant. Finally, we show that mutation of AGO2alters the ability of PVX to infect Arabidopsis. Thisreport demonstrates that RNA silencing may representan important mechanism that restricts wild-type, non-host-adapted viruses, and that the AGO2 protein playsan important role in antiviral defenses.
RESULTS
Identification of a Virus Allowing PVX Infectionof Arabidopsis
Sap derived from PVX 2.7-infected Nicotiana cleve-landii leaves (Cooper et al., 2003) was used to rub
inoculate Arabidopsis Columbia-0 (Col-0) plants andsystemic disease symptoms were observed, includingplant stunting, necrosis of rosette leaves, and curvedstems (Fig. 1A). In contrast, Arabidopsis Col-0 plantsinfected with PVX strain UK3 showed no obvious viralsymptoms. Five days post inoculation, high levels ofaccumulation of PVX coat protein (CP) were detectedin protein extracts from leaves inoculated with PVX2.7, but not in leaves inoculated with PVX UK3 (Fig.1B). Both PVX strains infected Nicotiana benthamianawith equal efficiency as judged by PVX CP accumu-lation in inoculated leaves (Fig. 1B). Furthermore, sappreparations of both strains also elicited similar num-bers of local lesions on Rx2 transgenic plants (seebelow, Supplemental Fig. S2), indicating that the two
Figure 1. Infection of Arabidopsis by PVX 2.7. A, Symptoms induced inwild-type Arabidopsis (Col-0) by the PVX 2.7 strain. Three leaves perplant were rub inoculated with sap made from either PVXUK3-infectedtobacco (PVX UK3) or PVX 2.7-infected N. clevelandii (PVX 2.7).Photographs were taken 3 weeks after inoculation. B, Representativeimmunoblots of PVX CP accumulation in Arabidopsis and N. ben-thamiana. Total protein extracts were prepared from inoculated Arab-idopsis or N. benthamiana leaves 5 or 6 d, respectively, afterinoculation, and subjected to SDS-PAGE followed by anti-CP immu-noblotting. Coomassie (left) or Ponceau (right) staining of the sameextracts is shown to demonstrate equal loading. Each lane correspondsto the pool of the three inoculated leaves from a single plant forArabidopsis and a single leaf for N. benthamiana. Three plants pertreatment were tested in each Arabidopsis experiment and the exper-iment was repeated five times for Arabidopsis and twice for N.benthamiana.
ARGONAUTE2 and Antiviral Defense
Plant Physiol. Vol. 156, 2011 1557 www.plantphysiol.orgon March 9, 2020 - Published by Downloaded from
different inocula have similar viral infectious titres. Atthe same time however, rub inoculation of PVX 2.7 onRx transgenic tobacco induced yellow ringspot lesionsand vein necrosis whereas these plants are normallyimmune to PVX and show no visible lesions upon PVXinfection (Supplemental Fig. S1; Bendahmane et al.,1999). Furthermore, in contrast to PVX UK3, PVX 2.7induced necrotic lesions on cowpea (Vigna unguiculata;Supplemental Fig. S1), leading us to consider that PVX2.7 might be contaminated with a second virus.
To determine if another virus was present in thePVX 2.7 sap, we used an expanded version of a mac-roarray developed to identify viruses from infectedplants (Agindotan and Perry, 2007) that consists of70-mer DNA oligonucleotides specific to over 100known viruses along with ribosomal plant RNA hy-bridization controls (K.L. Perry, unpublished data).The presence of specific viruses can be detected byhybridization with a labeled preparation of randomlyprimed, reverse-transcribed, and amplified cDNAsderived from infected plant RNAs (Agindotan andPerry, 2008). As expected, labeled cDNA derived fromPVX 2.7-infected plants hybridized to PVX sequencespresent on the macroarray (Fig. 2). At the same time, arobust signal was detected corresponding to PepRSV(Fig. 2). PepRSV is a bipartite single-stranded positive-sense RNA virus, and a member of the genus Tobravi-rus (MacFarlane, 2010). To confirm the identity ofPepRSV, total RNA from a PVX 2.7-infected Arabi-dopsis leaf was subjected to reverse transcription(RT)-PCR using generic tobravirus PCR primers thatamplify an approximately 898-bp fragment from thetobravirus replicase open reading frame (Jones et al.,2008). The same PCR product was sequenced over aregion of 668 nucleotides and BLASTanalysis resultedin a single hit in GenBank, matching PepRSV (Gen-Bank accession no. L23972) at 667 out of 668 nucleo-tides (Supplemental Fig. S3A). In contrast, the samesequence showed only 55% identity to TRV (Supple-mental Fig. S3B).
To generated sap containing only PepRSV we em-ployed Rx2 transgenic N. benthamiana plants, whichare resistant to PVX (Bhattacharjee et al., 2009). Theseplants also display visible hypersensitive response(HR) lesions upon PVX infection and as such, can beused to ascertain the presence of PVX. PVX 2.7 was
inoculated onto Rx2 N. benthamiana and sap wascollected from systemic tissues. This sap was subse-quently passaged to additional Rx2 plants. Rx2 plantsinoculated with PVX UK3 showed only HR lesions oninoculated leaves. In contrast, plants inoculated withthe PVX 2.7 sap passaged through Rx2 transgenicplants showed systemic viral symptoms, but no HRlesions (Supplemental Fig. S2). This result indicatesthat the latter sap contains only PepRSVand is referredto as such hereafter. To verify the presence of PepRSVin this inoculum, we performed RT-PCR on total RNAfrom N. benthamiana infected with PepRSV sap andamplified the open reading frames encoding thePepRSV-12K and -29K proteins (see below). Sequenceanalysis showed these to be 100% and 99% identicalto corresponding sequences in the published PepRSVsequence (Supplemental Fig. S3). Given that sequenceidentity is low between tobravirus species (MacFarlane,2008), we thus conclude that the contaminating virus inPVX 2.7 is indeed PepRSV.
To demonstrate that PepRSV is responsible for PVXaccumulation in Arabidopsis, we inoculated plantswith a mixture of PVX UK3- and PepRSV-containingsaps. Five days post infection, inoculated leaves werecollected and proteins extracted. Immunoblot analysisrevealed high levels of PVX CP in the presence ofPepRSV (Fig. 3). Moreover, although PepRSV by itselfinduces necrotic symptoms on Arabidopsis, coinfec-tion of both viruses consistently induced more severesymptoms (Fig. 3), indicative of a synergistic interac-tion between the two viruses.
The PepRSV 12K Protein Possesses VSR Activity
The apparent synergistic relationship betweenPepRSV and PVX suggests that coinfection involvesthe inhibition of RNA silencing. The PepRSV 12Kprotein is homologous to the TRV 16K protein thatpossesses VSR activity (Ghazala et al., 2008; Martin-Hernandez and Baulcombe, 2008; Martinez-Priegoet al., 2008). To demonstrate VSR activity in the 12Kprotein we tested it in the standard Agrobacterium-mediated transient expression (agroinfiltration) assaythat takes advantage of the fact that transiently ex-pressed transgenes are subjected to RNA silencing,and that this silencing is suppressed by many VSRs
Figure 2. Detection of PepRSV. TotalRNA from PVX 2.7-inoculated Arabi-dopsis leaves was extracted 5 d afterinoculation, converted to labeledcDNA, and hybridized to the macro-array. Black arrowheads indicate thelocations of the columns and rows withthe spotted control oligonucleotidesspecific to plant ribosomal RNAs. Ar-rows indicate the boxed columns withthe spotted PepRSV- and PVX-specificoligonucleotide probes.
Jaubert et al.
1558 Plant Physiol. Vol. 156, 2011 www.plantphysiol.orgon March 9, 2020 - Published by Downloaded from
(Voinnet et al., 2003). Transient expression of 35S-GFPby agroinfiltration in N. benthamiana leaves results inbright fluorescence visible at 3 d post infiltration (dpi).However, by 5 dpi, GFP expression is silenced andfluorescence only faintly visible (Fig. 4A). Coexpres-sion of GFP with either PepRSV 12K or the well-characterized VSR, P19 (Voinnet et al., 2003), resultedin strong GFP fluorescence at 5 dpi, whereas thePepRSV 29K movement protein had no effect on GFPpersistence (Fig. 4A). PVX encodes three genes re-quired for viral movement, known as the triple geneblock (TGB), including P25, which also possesses VSRactivity (Voinnet et al., 2000; Bayne et al., 2005). Weexpressed 12K with a version of PVX that expresses
GFP but lacks the TGB (PVX-GFPDTGB), to determinewhether the 12K protein could substitute for a lack ofP25 as a VSR. Indeed, GFP fluorescence derived fromPVX-GFPDTGB was significantly augmented by coex-pression with 12K, to a degree that was consistentlyhigher than that afforded by coexpression with P19(Fig. 4B).
Role of RNA-Silencing Components in ArabidopsisResistance to PVX
Given that PepRSVenables PVX to infect Arabidop-sis, we hypothesized that this may be due to efficientelimination of PVX by RNA-silencing mechanismsrather than an intrinsic inability to replicate andmove in this host. To test this, we infected an Arabi-dopsis triple dicer mutant, dcl2 dcl3 dcl4. ArabidopsisDicer mutants allow enhanced accumulation of sev-eral Arabidopsis-infecting RNAviruses (Blevins et al.,2006; Bouche et al., 2006; Deleris et al., 2006; Fusaroet al., 2006; Donaire et al., 2008; Qu et al., 2008; Garcia-Ruiz et al., 2010), and the triple mutant has a dramat-ically reduced ability to produce vsiRNAs (Deleriset al., 2006). At 4 dpi we observed high levels of PVXCP accumulation in infected leaves and after 21 d,
Figure 3. Coinoculation of PVX and PepRSV on Arabidopsis. Threeleaves per plant were rub inoculated with sap containing PVX UK3, sapcontaining PepRSV, or a mixture of both saps. A, Photographs weretaken 3 weeks post infection. Enlargements of the singly or doublyinfected rosettes are shown in the bottom section. B, Representativeimmunoblot analysis of the PVX CPaccumulation in inoculated leaves.Total protein extracts were prepared from inoculated leaves 5 d afterinoculation and subjected to SDS-PAGE followed by anti-CP immuno-blotting. Coomassie staining of the same extracts is shown to demon-strate equal loading. Each lane corresponds to the pool of threeinoculated leaves from the same plant. Three plants per treatmentwere tested in each experiment and the experiment was repeated fivetimes.
Figure 4. VSR activity of PepRSV-12K. A, Persistence of fluorescencefrom GFP upon coexpression with PepRSV-12K. GFP was transientlyexpressed via agroinfiltration from the 35S promoter in N. benthami-ana, together with either 35S:PepRSV-12K:HA, 35S:P19, 35S:PepRSV-29K:HA, or pBin61 empty vector (EV). GFP signal was monitored byUV illumination at 3 and 5 d after infiltration. B, PepRSV-12K enhancesPVX-GFPDTGB inN. benthamiana. A derivative of PVX expressing GFPbut lacking the TGB (PVX-GFPDTGB) was expressed from the 35Spromoter by agroinfiltration, together with either 35S:PepRSV-12K:HA,35S:TBSV-P19, 35S:PepRSV-29K:HA, or pBIN61 empty vector (EV).GFP signal was monitored by UV illumination at 3 and 5 d afterinfiltration.
ARGONAUTE2 and Antiviral Defense
Plant Physiol. Vol. 156, 2011 1559 www.plantphysiol.orgon March 9, 2020 - Published by Downloaded from
plants presented severe disease symptoms with sys-temic necrosis (Fig. 5).
Hypomorphic ago1 and null ago7 mutants are moresusceptible to VSR-defective TCV (Qu et al., 2008;Azevedo et al., 2010). To determine whether wild-typePVX is targeted by similar AGO proteins, PVX wasinoculated onto Arabidopsis lines possessing the hy-pomorphic ago1-27 allele, mutant lines with null al-leles in the AGO2, AGO3, AGO4, AGO5, AGO6, AGO7,AGO8, AGO9, and AGO10 genes, as well as the dcl2dcl3 dcl4 mutant as a control. Twenty-eight days afterinoculation (at least 18 plants for each genotype), novisible virus symptoms were apparent in the systemictissues in any of the 10 agomutants (Fig. 6A). Howeverat 5 dpi, PVX CP accumulation was detected in theinoculated leaves of ago2 mutant plants (Fig. 6B).Despite a lack of visible systemic symptoms followingPVX inoculation, PVX CP was detected by immuno-blotting in the systemic tissues of eight out of 18 PVX-inoculated ago2 plants (compared to zero out of similarnumbers of all other ago mutant genotypes), albeit atlower levels than in the dcl2 dcl3 dcl4mutants (Fig. 7A).In addition, sap collected from systemic tissues ofPVX-infected ago2mutants induced HR lesions on Rx2plants (2/4 times), indicating the presence of PVX inthese plants (Fig. 7B). In contrast no HR lesions wereinduced with sap from PVX-inoculated wild-type
Arabidopsis or any of the other ago mutants (Fig.7B). This result indicates that AGO2 plays a major rolein the antiviral defense against PVX in Arabidopsis.
DISCUSSION
To our knowledge, these results provide the firstdemonstration that RNA silencing plays a major rolein restricting a non-host-adapted virus. We show thatsynergy between PVX and PepRSV enable systemicinfection of PVX in Arabidopsis. Both synergism andexpansion of host range have been shown to occurupon virus coinfection and in certain cases VSRs, suchas helper component proteinase, have been implicatedin this phenomenon (Sonoda et al., 2000; Latham andWilson, 2008). The PepRSV 12K Cys-rich protein ishomologous to the TRV 16K protein, which has pre-viously been shown to function as a VSR (Ghazalaet al., 2008; Martin-Hernandez and Baulcombe, 2008;Martinez-Priego et al., 2008), and we find that PepRSV12K functions efficiently as a VSR in N. benthamianatransient expression assays (Fig. 4).
We did not observe accumulation of a PVX-12Kconstruct in either inoculated or systemic leaves ofArabidopsis (data not shown). This may not be sur-prising since, in addition to a cell-autonomous VSRfunction (Bayne et al., 2005), a critical property of thePVX VSR, the P25 protein, seems to be its ability tosuppress a silencing signal, both locally and system-ically (Voinnet et al., 2000; Himber et al., 2003; Bayneet al., 2005). The silencing signal involves the move-ment of siRNAs (Dunoyer et al., 2010; Molnar et al.,2010) and it is thought that such a mechanism allowsvsiRNAs to accumulate in uninfected cells ahead ofviral movement where they can target viral RNAsupon their entry into the cell. It is unclear how the P25protein inhibits vsiRNA movement, although the useof this strategy may indicate that PVX is particularlysensitive to the effect of having mobile vsiRNAs pre-sent in uninfected cells prior to viral entry. If so, thepresence of vsiRNAs in a cell prior to infection wouldbe predicted to target PVX, via AGO2, before signif-icant amounts of VSR protein can be expressed fromthe PVX genome. However, coinfection of Arabidopsisleaf cells by PepRSV appears to suppress silencingmechanisms, thereby allowing PVX to move into thesecells. Likewise, a block in vsiRNA production in thetriple dicer mutant, or a block in the appropriatevsiRNA effector AGO protein, would accomplish thesame outcome.
The ability of PVX to infect the Arabidopsis tripledicer mutant suggests that the P25 protein may notfully function as an effective VSR in this host. Previousreports have shown that P25 is able to suppresssilencing induced by RNA hairpins in both N. ben-thamiana and Arabidopsis (Dunoyer et al., 2004; Bayneet al., 2005). However the effect of P25 on the spread ofsilencing signals has not been investigated in Arabi-dopsis and the developmental phenotypes induced by
Figure 5. PVX infection of an Arabidopsis triple DICER mutant. A,Symptoms induced by PVX UK3 infection in wild type and dcl2 dcl3dcl4 mutant Arabidopsis. Three leaves per plant were rub inoculatedwith sap containing PVX UK3. Photographs were taken 4 weeks afterinfection. B, Immunoblot analysis of PVX CP accumulation in inocu-lated leaves. Total protein extracts were prepared from inoculatedleaves 6 d after inoculation and subjected to SDS-PAGE followed byanti-CP immunoblotting. Coomassie staining of the same extracts isshown to demonstrate equal loading. Each lane corresponds to a poolof the three inoculated leaves from a single plant. Four plants pertreatment were tested in each experiment and the experiment wasrepeated five times.
Jaubert et al.
1560 Plant Physiol. Vol. 156, 2011 www.plantphysiol.orgon March 9, 2020 - Published by Downloaded from
transgenic expression of P25 in Arabidopsis are milderthan for other VSRs (Dunoyer et al., 2004). Alterna-tively, the inability of PVX to infect Arabidopsis maybe due to a combination of factors, with the inhibitionor ablation of silencing allowing it to overcome athreshold required for systemic infection.Several VSRs directly target AGO1 and either pre-
vent it from loading siRNAs or induce its degradation(Zhang et al., 2006; Baumberger et al., 2007; Azevedoet al., 2010; Giner et al., 2010). The VSRs P1 and P38exert their suppressor of silencing activity throughdirect interaction with AGO1 via their WG/GWmotifs(Azevedo et al., 2010; Giner et al., 2010). Sequenceanalysis of tobravirus 16K homologs indicates thepresence of a conserved GW motif in the N terminiof these proteins (Ghazala et al., 2008). However,attempts to coimmunoprecipitate multiple AGO pro-teins with 12K and 16K did not reveal any interactions(data not shown). Recently, it was reported that thePVX P25 protein is able to interact with AGO1, AGO2,AGO3, and AGO4, as well as destabilizing AGO1(Chiu et al., 2010). Using comparable methodology, weare similarly unable to detect interactions between P25and a number of AGO proteins, including Arabidopsisand tobacco AGO2 (data not shown).It is generally assumed that degradation of viral
RNA by Dicer proteins alone is not sufficient to effectvirus resistance and that their vsiRNA products musteventually be loaded into RISC complexes to targetviral RNAs. The AGO1, AGO2, and AGO5 proteinscan bind vsiRNAs (Zhang et al., 2006; Takeda et al.,2008; Harvey et al., 2011) but, of the 10 Arabidopsis
AGO proteins, genetic analysis has implicated onlyAGO1, AGO2, and possibly AGO7 in virus resistance.The ago1-27 hypomorphic mutant shows a 5-foldgreater accumulation of CMV RNA and both an ago1hypomorphic mutant and an ago7 null mutant allowincreased accumulation of a VSR-defective version ofTCV (Morel et al., 2002; Qu et al., 2008; Azevedo et al.,2010). Infection of ago2 mutants with CMV and TCVresulted in more severe viral symptoms, but littledifference in virus accumulation (Harvey et al., 2011).Surprisingly, we did not observe any difference in PVXaccumulation on either ago1-27 or ago7 mutants (Fig.6). Instead, we observed PVX accumulation in inocu-lated and systemic tissues only in the ago2 mutant(Figs. 6 and 7), indicating that AGO2 is essential forcurtailing PVX on this host. Furthermore, our resultsshow that AGO2 activity can absolutely limit virusinfection rather than simply playing a secondary roleto AGO1, as previously suggested (Harvey et al., 2011).Interestingly, AGO2 is placed in the same phylogeneticclade as AGO7 (Vaucheret, 2008) but unlike ago7,which is involved in tasi-RNA biogenesis, ago2 mu-tants show no obvious phenotype (Vaucheret, 2008).This may indicate that AGO2 has evolved to specializein antiviral defenses.
We do not formally rule out a role for AGO1 inresistance to PVX as the protein encoded by ago1-27may possess sufficient activity to target PVX, but notTCVor CMV. Indeed, although we find that ago2 is theonly single mutant capable of allowing PVX infection,this infection is not as efficient as in the dcl2 dcl3 dcl4mutant (Figs. 5–7), suggesting that at least one other
Figure 6. PVX infection in Arabidopsis ago mutants. A, Symptoms induced by PVX UK3 infection on wild type and ago1-27,ago2-1, ago3-2, ago4-2, ago5, ago6-3, ago7, ago8-1, ago9, and ago10 Arabidopsis mutants. Three leaves per plants were rubinoculated with sap made from PVX UK3-infected tobacco (PVX UK3). Photographs were taken 4 weeks after infection. B,Immunoblot analysis of PVX CPaccumulation in inoculated leaves. Total protein extracts were prepared from inoculated leaves 5d after inoculation and subjected to SDS-PAGE followed by anti-CP immunoblotting. Coomassie staining of the same extracts isshown to demonstrate equal loading. Each lane corresponds to a pool of the three inoculated leaves from a single plant. Thisresult was obtained consistently in five experiments including at least three replicates for each genotype.
ARGONAUTE2 and Antiviral Defense
Plant Physiol. Vol. 156, 2011 1561 www.plantphysiol.orgon March 9, 2020 - Published by Downloaded from
AGO protein acts in PVX antiviral silencing. A sup-plemental or hierarchical role of another AGO protein(s) would not be surprising given the redundancy/complementarity seen with members of the DICERfamily (Deleris et al., 2006; Garcia-Ruiz et al., 2010). Atthe same time AGO2, which is regulated by an endog-enous miRNA (Allen et al., 2005), is expressed at a 4.6-fold higher level in an ago1 hypomorphic mutant(Kurihara et al., 2009), and AGO2 protein levels aresignificantly up-regulated upon infection with TCVand CMV (Harvey et al., 2011). Thus, increased levelsof AGO2may in theory compensate for a lack of AGO1with respect to PVX resistance.
A previous report has shown that the requirementfor AGO1 versus AGO7 for resistance to a VSR-defec-tive version of TCV differed depending on whetherthe virus contained a GFP insert (Qu et al., 2008).Likewise, although AGO1 has been shown to play animportant role in defense against VSR-defective TCVin Arabidopsis (Qu et al., 2008; Azevedo et al., 2010),
we find that down-regulation of NbAgo2 has a muchgreater effect than NbAgo1 on VSR-defective TBSV inN. benthamiana (Scholthof et al., 2011). This suggeststhat different AGO proteins may vary in their effec-tiveness in targeting different viral RNA genomesdepending on as-yet-undefined characteristics of theviral RNA genome. Why certain AGO proteins mightbe more or less important for targeting different vi-ruses is unclear. Recent reports have shown that theregions of a virus most effectively targeted by vsiRNAsdo not necessarily correlate with those regions cor-responding to the most abundant viRNAs (Szittyaet al., 2010) and that AGO proteins show bias in thesiRNAs they bind (Mi et al., 2008; Takeda et al., 2008;Ho et al., 2010). Thus, it is tempting to speculate thattargeting by a given AGO protein may depend ona combination of the availability of the viral genometo RNA-silencing components, determined by RNAstructure, together with the nucleotide compositionof such regions.
CONCLUSION
Our results show that in addition to its well-definedrole in the interaction between viruses and their hosts,RNA silencing is also an important mechanism ofnonhost resistance against non-host-adapted viruses.Furthermore, our results indicate an important role forthe AGO2 protein in antiviral defenses that will lay thegroundwork for future studies on the interplay be-tween different AGOs, DCLs, vsiRNAs, and viruses.
MATERIALS AND METHODS
Plant and Virus Materials
Nicotiana benthamiana, Nicotiana clevelandii, and tobacco (Nicotiana tabacum)
were germinated and grown in a glasshouse maintained at 21�C. Samples of
PVX 2.7-infected N. clevelandii leaves were generously provided by Bret
Cooper (Cooper et al., 2003) and were used to produce sap for further virus
infections. PVX strains were propagated in tobacco or N. clevelandii (UK3) and
N. clevelandii (2.7), respectively. PepRSV was isolated by infecting Rx2 trans-
genic N. benthamiana (Bhattacharjee et al., 2009) with PVX 2.7 sap and
subsequently propagated in N. clevelandii. Virus saps were prepared by
grinding infected or systemic tissue in 0.1 M potassium phosphate buffer, pH
7.5. The concentration of PVX-containing inocula was normalized according
to the immunoblot signal of PVX CP. Arabidopsis (Arabidopsis thaliana) wild
type and mutants are of the Col-0 ecotype and were cultivated in growth
chambers under long-day period (16 h/8 h). All Arabidopsis mutant lines
were of the Col-0 ecotype and have been previously described, including the
triple dicer mutant dcl2-1/dcl3-1/dcl4-2 (Deleris et al., 2006), ago1-27 (Morel
et al., 2002), ago2-1, ago3-2, ago6-3, ago8-1 ago10-2 (Takeda et al., 2008), ago5,
ago7, ago9 (Katiyar-Agarwal et al., 2007), and ago4-2 (Agorio and Vera, 2007).
All virus infections of Arabidopsis were carried out by rubbing sap on
carborundum-dusted leaves of 3-week-old plants with the side of pipette tip.
Immunoblot Analysis
Total proteins were extracted in 8 M urea solution, separated in a 10% SDS-
polyacrylamide gel, and transferred to a polyvinylidene difluoride membrane
by electroblotting. Each sample analyzed corresponds to a pool of inoculated
leaves or systemic tissue of a single plant. For each infection, at least three plants
were tested in more than five different experiments. PVX CP was probed with
Figure 7. Systemic movement of PVX in Arabidopsis RNA-silencingmutants. A, Immunoblot analysis of PVX CP accumulation in Arabi-dopsis systemic tissue. Total protein extracts were prepared from thesystemic leaves of one Arabidopsis plant of each genotype 3 weeks afterPVX inoculation, followed by SDS-PAGE and anti-CP immunoblotting.Coomassie staining of the same extracts is shown to demonstrate equalloading. The triple dicer mutant sample is taken from a separateexperiment but similar amounts of tissue and exposure times wereused. B, Detection of PVX by inoculation of Rx2 transgenic N.benthamiana. Sap was prepared from the systemic tissue of PVXUK3-infected wild type and ago mutant Arabidopsis. The left sectionindicates a representative result (lack of HR lesions) obtained frominoculation with sap from wild type, ago1-27, or mutants in ago3 toago10. The right section depicts a representative result showing HRlesions, indicating the presence of PVX, observed upon inoculationwith sap preparations from individual ago2-1 plants.
Jaubert et al.
1562 Plant Physiol. Vol. 156, 2011 www.plantphysiol.orgon March 9, 2020 - Published by Downloaded from
anti-PVX CP rabbit polyclonal antibodies (Agdia) and subsequently detected by
a goat anti-rabbit Horseradish peroxidase-conjugated polyclonal antibody
followed by chemiluminescence detection (ECL, GE Healthcare).
Cloning and Sequencing of PVX 2.7
Total RNA from PVX 2.7-infected Nicotiana glutinosa or Arabidopsis was
extracted with Trizol reagent according to the manufacturer instructions
(Invitrogen). cDNAs were obtained after RT using a primer targeting the 3#untranslated region of PVX (5#-ATTTATATTATTCTAACAATCAAAC-3#) andPCR (ExpandHi Fidelity Taq DNA polymerase, Roche) with primers targeting
overlapping fragments of PVX, based on the PVX UK3 sequence. PCR
products were cloned into pGEMT (Promega) and sequenced and overlap-
ping sequences were used to assemble a full-length sequence. At least three
independent RT-PCRs were performed for both the Arabidopsis (PVX2.7a)
andN. glutinosa (PVX2.7b) sequences, which have been deposited in GenBank
as accession numbers HQ450387 and HQ450388.
Macroarray Hybridizations and Identification of PepRSV
cDNAwere generated from PVX 2.7-infected Arabidopsis RNA extraction
and used to probe the macroarray membrane as previously described
(Agindotan and Perry, 2007, 2008). The identity of PepRSV was confirmed
by RT-PCR from total RNA isolated from a PVX 2.7-infected Arabidopsis plant
using generic tobravirus PCR primers Tobra-F3 and Tobra-R2, which amplify
an approximately 898-bp fragment, as previously described (Jones et al., 2008).
The latter PCR product was sequenced using the same primers and used to
query sequence databases. The 668 bp that could be read from this sequencing
have been deposited in GenBank as accession number JF906525. The identity
of PepRSV was further confirmed by amplification of the PepRSV-12K and -29K
genes from total Trizol-extracted (Invitrogen) RNA from PepRSV-infected
N. benthamiana. cDNAs were obtained after RTusing polydT primers and PCR
(Expand Hi Fidelity Taq DNA polymerase, Roche) with the primer pairs