ACCEPTED. 2 compromises both sirna and mirna mediated pathways 3 Tibor Csorba 1,2, Aurelie Bovi 3, Tamás Dalmay 3 *, József Burgyán 1 * Hungary

Transcription

1 JVI Accepts, published online ahead of print on August 0 J. Virol. doi:./jvi.0-0 Copyright 0, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved. 1 The p1 subunit of Tobacco mosaic virus replicase is a potent silencing suppressor and compromises both sirna and mirna mediated pathways Tibor Csorba 1,, Aurelie Bovi, Tamás Dalmay *, József Burgyán 1 * 1 Agricultural Biotechnology Center, Plant Biology Institute, P.O. Box, H-01 Gödöllõ, Hungary Department of Genetics, Eötvös Lóránd University, Budapest, Hungary University of East Anglia, Norwich UK Running title: Suppression of RNA silencing by viral protein 1 The number of words 1 Abstract: words, Main text: 0 characters with spaces 1 Keywords: cr-tmv, p1 silencing suppressor, small RNA binding, small RNA methylation 1 *Correspondence: 1 József Burgyán 1 1 H-01 Gödöllı, P. O. Box., Hungary TEL: ---1; FAX: burgyan@abc.hu Tamás Dalmay University of East Anglia, Norwich UK T.Dalmay@uea.ac.uk 1

2 Summary One of the functions of RNA silencing in plants is to defend against molecular parasites such as viruses, retrotranspozons, and transgenes. Plant viruses are inducers as well as targets of RNA silencing based antiviral defense. Replication intermediates or folded viral RNAs activate RNA silencing generating small interfering (si) RNAs, which are the key players in the antiviral response. Viruses are able to counteract RNA silencing by expressing silencing suppressor proteins. It has been shown that many of the identified silencing suppressor proteins bind long dsrna or sirnas and thereby prevent assembly of the silencing effector complexes. In this study we have shown that the 1 kda replicase subunit (p1) of crucifer-infecting Tobacco mosaic virus (cr-tmv) is a potent silencing suppressor protein. We found that the p1 protein preferentially binds to double-stranded nt sirna and micro (mi) RNA intermediates having nt overhangs inhibiting the incorporation of sirna and mirna into silencing related complexes (e.g.: RISC) both in vitro and in planta, but cannot interfere with the previously programmed RISC complexes. In addition, our results also suggest that the virus infection and/or sequestration of the sirna and mirna molecules by p1 enhance mirnas accumulation despite preventing their methylation. However, the p1 silencing suppressor does not prevent the methylation of certain mirnas in hst-1 mutants where the nuclear export of mirnas is compromised. Introduction RNA silencing is a highly conserved mechanism in eukaryotes and endogenous silencing pathways has important roles in gene regulation at the transcriptional, RNA stability and translational levels (). Although it operates through diverse pathways, RNA silencing relies on a set of core processes, which may not entirely overlap. The triggers of RNA silencing are highly structured ssrna or dsrna molecules, which are processed into - nt short interfering RNA (sirna) or microrna (mirna) duplexes by the RNase III-type DICER enzymes. Then small RNAs are incorporated into a ribonucleoprotein complex termed RNA induced silencing complex (RISC) (0). Assembly of RISC involves guide strand incorporation and passenger strand elimination to program active RISC (1, ). In Drosophila embryo extract, DICER-RD

3 (DCR-RD) complexed with duplex sirna serves as an initiator of RISC assembly. This complex then interacts with an AGO-containing multiprotein complex and cleaves the passenger strand of the sirna duplex to form the single-stranded (ss) sirna containing 0S holo-risc, which catalyses sequence-specific cleavage of target RNA (1,, ). In higher plants AGO1 protein is the slicer component of RISC () and it recruits small silencing-related RNAs such as mirnas, trans-acting sirnas, transgene-derived sirnas () and virus-derived sirnas (). However, other members of the AGO family and different Dicer-like (DCL) and RNA-dependent RNA polymerase (RDR) proteins are also involved in diverse RNA silencing pathways. In the nucleus, DCL-, RDR-, AGO- and NRPD1-dependent silencing is associated with DNA cytosine methylation, heterochromatin-associated modifications of histone H tails and transcriptional silencing (1,,,, 0, ). Post-transcriptional RNA silencing related pathways for endogenous gene regulation involve mirnas (DCL1-dependent) or ta-sirnas (DCL- and RDR-dependent) (0). Virus-induced RNA silencing is triggered by dsrna intermediates of cytoplasmically replicating viruses, RDR1- or RDR-dependent formation of dsrna, or structured regions of viral RNAs (, ). Virus-induced silencing leads to the sequence-specific degradation of viral RNA () and generation of a mobile silencing signal that activates RNA silencing in non-infected cells (, ). Systemic invasion of plants by viruses requires effective mechanisms to suppress RNA silencing. To counteract antiviral RNA silencing, most plant and many animal viruses have evolved silencing suppressor proteins (, ). The large majority of plant viral suppressors of RNA silencing bind either to long dsrna or to sirna/mirna duplexes (,,,, 1). However, two exceptions are also reported; the b protein of Cucumoviruses () that binds to and inactivates the AGO1 effector protein of silencing and the P0 RNA silencing suppressor of the Poleroviruses (0, 1), which does not have RNA binding activity (J Burgyan, unpublished results). The crucifer-infecting strain of Tobacco mosaic virus (cr-tmv) is belonging to the Tobamovirus genus. Tobamoviruses are a group of rod-shaped plant viruses with undivided positive sense RNA genome encoding at least four proteins. The genomic RNA of cr-tmv directs the translation of a 1kDa protein and its read-through product, a kda protein, which are involved in the replication of the viral RNA. The other two proteins, the kda movement protein and the 1kDa coat protein are translated from individual coterminal subgenomic RNAs

4 (1). The p1 protein is composed of three domains: a methyltransferase domain (MT) - aa, an intervening region (IR) and a helicase domain (HEL) - aa, whereas the K protein possesses an additional RNA-dependent RNA polymerase motif. Little is known about the precise mechanism of action of p1 upon infection. It has been demonstrated that the p1 protein of TMV OM strain (corresponding to the p1 of cr-tmv) forms a heterodimer replicase complex with p1 and two or more host proteins (). Only one specific interaction between the p1 and p1 was identified, which is in the C- terminal half of p1 IR and N-terminal portion of p1 HEL domain (). The ratio between p1 and p1 is 1:1, although they are expressed in a :1 ratio during infection (). The biological function of this excess amount of p1 remains to be determined. A substitution mutant of TMV-L strain in which the amber stop codon of p1 was replaced by tyrosine codon thus expressing only the p1 readthrough product was shown to replicate in vivo in the absence of the p1 protein. The growth rate of this mutant virus was about one-tenth of the rate of wild- type (). All these results imply that ORF1 product besides being involved in replication complex actions has other functions as well. Indeed the other analogous protein p of Tomato mosaic virus (ToMV) has been shown to have silencing suppressor function (). It was also shown that p does not suppress the activity of the pre-existing, sequence-specific silencing machinery, suggesting that p blocks the utilization of small RNAs in the formation of new effector complexes (), however, the molecular bases of the suppression remained to be determined. Here we demonstrated that p1 has strong silencing suppressor activity and we explored in details the molecular mechanism of silencing suppression by p1 protein. Our findings demonstrated that p1 prevents the si/mirnas assembly into RISC complexes inhibiting the development of virus or transgene induced silencing activity. In contrast p1 has no effect on the slicer activity of preassembled RISC, both in vitro and in vivo. In addition, we also demonstrated that p1 interferes with mirna and tasirna mediated pathways, however this interference is not a general effect instead it likely depends on the spatial and temporal coexpression of p1 and mirnas. MATERIALS AND METHODS Plasmid constructs.

5 The full length infectious cdna clone of the puc1-cr-tmv was prepared previously (T. Dalmay unpublished results) and the puc1-cr-tmv- p1 mutant virus clone has been prepared PCR mutagenesis substituting the the amber stop codon of p1 (TAG) to tyrosin (TAT). Suppressor p1 was PCR amplified using appropriate primer pairs and cloned into pbin1 and used for agroinfiltration assays. The following constructs used for different transient assays were described previously S-GFP, S-GFP-IR, S-sigma (), S-His-HC-Pro (), GFP-Cym, GFP-PoLV (). GFP-1.1 and GFP-1. were kindly provided by O. Voinnet and were described previously (). Construct S-p1-His was prepared amplifying the p1 ORF by PCR using a forward primer containing a start codon (italics) ( - ATGGCACAATTTCAACAAACAATTGAC- ) and with a reverse primer containing RGS(His) epitop codons (underlined) and the stop codon (italics) ( - CTAGTGATGGTGATGGTGATGCGATCCTCTTTGTATCCCCGCTTCAACTCTATACATG TC- ), and then this fragment cloned into SmaI-cleaved BIN1 vector. S-p1 was prepared as S-p1-His except that the His tag was omitted from the revers primer. For Pri-miR1c construct we amplified the pri-mir1 sequence from cdna of A. thaliana with the forward ( - TGAGCGCACTATCGGACATCAAATAC- ) and reverse primers ( - TAAACGCGTGATATTGGCACGGCTC- ), and cloned it into pbin1-smai vector. Agrobacterium tumefaciens infiltration A. tumefaciens CC1/pBIN1 harboring the appropriate plasmid (pbin-gfp, pbin-hc-pro, pbin-p1, pbin-sigma) was infiltrated according to the method described previously (). The GF-IR construction was made in phannibal vector () by amplifying GFP sequences with forward primers -GAACTCGAGATGAGTAAAGGAGAAGAACTTTTCAC- or - GAGGTACCCGTGTCTTGTAGTTCCCGTCATC- and reverse primers - GAATCTAGAATGAGTAAAGGAGAAGAACTTTTCAC- or - GAATCGATCGTGTCTTGTAGTTCCCGTCATC-. The GF-IR construction (bearing an intron between the GF and FG sequences) was transfered into the pbin-1 vector and used in infiltration assays. GFP expressing N. benthamiana 1C line was co-infiltrated with S-GFP (OD 00 = 0.) or S-GF-IR (OD 00 = 0.) and suppressor protein constructs at OD 00 = 0.. Total RNA was extracted from the infiltrated pathes and analysed by northern and western blots.

6 Non infected or Cym1stop infected N. benthamiana plants was co-infiltrated with sirna- or mirna- sensor constructs ()()at OD 00 = 0.1 and suppressor protein constructs (OD 00 = 0.) and the pri-mir1c construct was infiltrated at OD 00 = 0.. RNA isolation and hybridization analyses In vitro transcription of the puc1-cr-tmv or puc1-cr-tmv- p1 viral constructs from PmlI linearized DNA templates and the inoculation of RNA transcripts onto Nicotiana benthamiana and Arabidopsis thaliana Columbia-0 ecotype plants were performed as described earlier (1). In vitro RNA transcripts were capped with a cap analogue (New England Biolabs, Hitchin, United Kingdom) as previously described. Total RNA from mock- and cr-tmvinfected plants was isolated using Trizol reagent. RNA extraction was performed dpi for N. benthamiana and 1 dpi for A. thaliana. RNA gel blot analysis was performed as described in (). Quantitative real-time RT-PCR analysis Expression of DCL1 gene in crtmv-infected and healthy control wild type or mutant plants was estimated by quantitative real-time RT-PCR using the SYBR Green assay (Applied Biosystems) and the Rotor-Gene 000 (Corbett research) system. RT-PCR was performed with DCL1 forward primer ( -ACAACTGCTGCTTGGAAGGTTTTTCAACCTTTGC- ) and DCL1 reverse primer ( -GCATTGGAAGTGTCTCTGGTGTCACCATGG- ) (amplicon region between 01-0 bases of the DCL1 mrna). cdna was synthesized in µl reaction volume using QuantiTect Reverse Transcription Kit (Qiagen). The qrt-pcr thermal profile was: C for min, and 0 cycles of C for 1 s and 0 C for 0 s. The U snrna was used as internal standard in quantitative PCR analysis for each reaction. The primers for U was: forward primer ( - GTCCCTTCGGGGACATCCGATAAAATTGGAACG- ) and reverse primer ( - AAAATTTGGACCATTTCTCGATTTATGCGTG- ). The quantitative real-time RT-PCR analysis was repeated times, and the average was taken to determine level of DCL1 mrna. Protein analysis Infiltrated leaf tissues were homogenized dpi in extraction buffer ( mm Tris-HCl, ph., M urea, % sodium dodecyl sulfate, and % β-mercaptoethanol). Samples were boiled and cell debris was removed by centrifugation at 1,000 g at C for min. Supernatants were resolved

7 on SDS-PAGE % and subjected to Western blot analysis. The proteins were visualized using anti-gfp, anti-his and anti-ha antibodies by chemiluminescence (ECL kit, Amersham) according to the manufacturer s instructions. Commercially available antibodies were used for detection of ergfp, xhis-tagged proteins. Ponceau red staining was used to check the global protein content of the samples. In vitro RNA silencing Drosophila lysate preparation, target RNA labeling and sirna annealing were described previously (1). In a ml reaction, ml of lysate and sirna in nm final concentrations were used in 1 lysis buffer containing % v/v of glycerol. GFP target RNA was in vitro transcribed with T polymerase in the presence of P-UTP and used in 0. nm final concentration. In direct competition assays, reactions were incubated for 1 hour. In active RISC assays, sirna and the extract were incubated for 0 minutes to allow RISC assembly, and then target RNA and suppressor proteins were added to the reaction. Samples were de-proteinized and RNA was analyzed on an % denaturing gel. Native gel electrophoresis Native gel electrophoresis for separation of silencing complexes was essentially as described () with some modifications. In direct competition assays, in vitro reactions that were used for target cleavage assays were incubated 0 minutes with nm P-labeled sirna and suppressor protein, diluted with µl of loading buffer (1 lysis buffer, % of Ficoll 00) and a part of it analyzed on a.% (:1 acrylamide:bisacrylamide) native acrylamide gel. Gels were dried, exposed to a storage phosphor screen and bands were quantified with a Genius Image Analyzer (Syngene). Detection of and OH on the last nucleotide of mirnas Periodate treatment and β-elimination were performed as previously described (). Gel mobility shift assay. For RNA binding reactions, labeled ssrna or dsrna (0. nm) were incubated with agrobacterium infiltrated N.benthamiana leaf extract containing ~1 µg total protein or the relevant dilutions. Binding reactions and the mobility shift assays were carried out as described (). RESULTS P1 protein of cr-tmv is a potent silencing suppressor

8 Kubota et al. (0) reported that the ToMV p protein had silencing suppressor activity suggesting that homologous proteins in the closely related TMV species are also silencing suppressors. In the strain of TMV, which infects A. thaliana (cr-tmv) the corresponding protein is encoded by the p1 gene. Cr-TMV infected A. thaliana plant shows aberrant leaf development (Figure 1A) - typical for plants carrying mutations in small RNA pathways that supports the hypothesis that p1 of cr-tmv is a silencing suppressor protein. To get insight in the nature of silencing suppressor activity of p1, previously described systemically silenced GFP transgenic A. thaliana AmpxGFP and Amp plants (1) were inoculated with cr-tmv. 1 days post inoculation (dpi) GFP fluorescence was assessed in the virus infected plants. Figure 1C,D show that TMV infection results in reversion of GFP expression in both Amp and AmpxGFP plants. However, in the AmpxGFP plants GFP expressed only in and around the vein. These observations suggest that the suppressor protein expressed during infection cr-tmv cannot suppress silencing completely in plants where silencing is robust as this is the case in AmpxGFP plants. To explore the molecular bases of silencing suppression of p1, S-p1 and S-p1-His were co-expressed with S-GFP or/and S-GF-IR in N. benthamiana leaves. The well characterized HC-Pro of Tobacco etch virus (TEV) that specifically binds to nt sirnas and Reovirus sigma that exclusively binds to long dsrna (), were used as controls in agroinfiltration assay (). The transiently expressed p1 suppressed GFP silencing with the same efficiency as S-HC-Pro based on the bright fluorescence in the co-infiltrated areas (Figure 1F). The suppression of GFP silencing by p1 was also very effective in the presence of S-GF-IR expression (Figure 1F). To confirm the visual observations we also checked the GFP mrna and GFP specific sirna accumulation. Total RNA samples were extracted from the infiltrated area 0 hours after infiltration and analyzed on Northern blot (Figure 1G). In the presence of p1 or HC-Pro the GFP mrna accumulated at high level, even when the S-GF- IR was simultaneously co-expressed, while the level of GFP mrna was strongly reduced when GFP was co-expressed with GF-IR or expressed alone in the absence of suppressors (Figure 1G). An RNA band corresponding to the GF-IR transcript was observed only in the presence of sigma protein that inhibits DICER activity by dsrna binding () but not in the presence of HC-Pro or p1 (Figure 1G, upper panel, compare lanes,, with ) indicating that p1 does not interferes with DICER activity similarly to HC-Pro that has been demonstrated previously

9 (). As we expected GFP sirna accumulated in the leaf infiltrated only with GFP (Figure 1G lower panel hybridised with GFP probe), while the presence of HC-Pro, p1 or sigma protein abolished the accumulation of GFP sirnas. In contrast, when GFP was coinfiltrated with GF-IR very high amount of GFP sirna accumulated in either the presence or the absence of p1 and Hc-Pro (Figure 1G, lower panel, lanes,,). The accumulation of sirna was inhibited only in the presence of sigma, which compromises DICER activity (). This result demonstrated that p1 inhibits the accumulation of sirna when the RNA silencing triggred by sense GFP transcript. However, p1 did not inhibit the processing of GF-IR dsrna to sirnas but interferes with the silencing machinery downstream to sirna generation. We also tested the effect of p1 in the accumulation of RDR dependent secondary sirnas (). We have shown that p1 was able to inhibt the accumulation of secondary sirnas detected by P-specific probe (Figure 1G compare lane 1 to lane ). The inhibition of secondary sirna (P specific sirna) accumulation by p1 was very efficient regardless that the GFP silencing was triggered by GFP alone or GFP + GF-IR expression (Figure 1G compare lanes 1 to and -). It is worthy to note, that the majority of primary sirnas were nt long and only one third of sirnas were / nt long when the silencing was triggered by GF-IR (Figure 1G lanes,, middle panel, GF specific sirnas). In contrast the majority of secondary sirnas (P specific) were / nt long, suggesting that they are the products of DCL and DCL, respectively. We also analyzed the accumulation of virus derived sirna in cr-tmv infected plants. Total RNA was extracted from N. benthamiana at dpi and from A. thaliana at 1 dpi. High amount of viral sirna accumulated in the virus infected but not in the mock inoculated plants (Figure 1E) suggesting that p1 does not compromise the generation of sirnas. P1 inhibits sirna-directed RNA cleavage and the assembly of silencing related complexes in Drosophila embryo extracts in dose-dependent manner The Drosophila embryo extract based RNA silencing system had been successfully used to characterize silencing suppressor proteins (). This system allows the analysis of RNA-silencing complex formation and the cleavage activity of programmed RISC complex (, ). To better understand how p1 silencing protein works we tested the sirna programmed RISC activity in the presence of p1 protein. Repeated attempts to express p1 in bacteria were not successful therefore p1 were expressed in plants using binary expression vector (). We set up two sets

10 of reactions for RISC cleavage assay. In the direct competition assay the inducer sirna, the labeled target RNA containing the sequence complementary to the inducer sirna, the Drosophila embryo extract and the p1 containing plant extract made from S-p1 agroinfiltrated N. benthamiana leaves at dpi were added simultaneously. RISC activity was measured by quantifying the amount of -cleveage product of the target RNA ( nt) over a dilution series of the suppressor protein extract (Figure, lines -). For control reaction we used empty vector-infiltrated plant extract at the highest concentration used for p1 (Figure, line 1), and we performed a reaction without embryo extract as negative control (line ). P1 was able to inhibit the target cleavage (Figure lanes -), likely preventing the assembly of RISC complex. To analyze the effect of p1 on preassembled RISC activity (indirect competition) we pre-incubated the Drosophila embryo extract with inducer sirna for minutes and then the target and the suppressor containing plant extract or mock-infiltrated plant extract were added (Figure, lines -1). P1 proved to be a potent inhibitor of RISC-cleavage at higher concentrations (IC 0 at fold dilution) in the direct competition assay but had no effect on preassembled RISC activity. Activity of programmed RISC was refractory to p1 suppressor protein regardless of the concentration. These results suggested that the sirna-sequestration model may also explain the mechanism of p1 mediated suppression as was demonstrated for p1, HC-Pro and p (,,, 1, ). RNA silencing initiator complex formation is impaired in vitro in the presence of p1 protein To test our hypothesis we analyzed the assembly of the RNA silencing complexes with sirnas by electrophoretic mobility shift assay developed by Pham et al (0) and adapted for suppressor protein assays by Lakatos et al (0). This technique is used to study the formation of the various stages of silencing complexes (sirna-dcr-rd, RISC loading complex and RISC). In the direct competition assay we mixed inducer sirna, Drosophila embryo extract and p1-infiltrated plant extracts simultaneously (Figure lines -). In the indirect competition assay the embryo extract were pre-incubated with the labeled sirna for minutes and then p1-infiltrated plant extract was added at the same dilutions (Figure lines 1-). As control reactions we incubated embryo extract with labeled sirna (line ), or we added empty vectorinfiltrated plant extract at the highest concentration used for direct and indirect competitions (lines and 1, respectively). In the direct competition experiment, p1 was able to inhibit the

11 RISC-assembly up to 1:1 dilution of plant extract, while in the indirect assay p1 did not compromise the preassembled RISC in any dilutions. It is worth noting that a newly formed complex was observed running slightly above the sirna-dcr-rd complex (Figure, compare line with lines and respectively). The accumulation of this complex gradually diminished as the concentration of p1 decreased. We hypothesized that this was probably a p1-sirna complex containing some unidentified cellular factors. These results and the observation of p1-sirna formation strongly suggested that p1 was able to bind ds sirnas and inhibited RNA silencing via sirna sequestration. P1 does not inhibit the programmed sirisc and mirisc in vivo The in vitro data showed that p1 could not interfere with the preassembled RISC activity but prevented the new RISC complexes to be formed. To test the effect of p1 on the si/mirna programmed RISC in planta we used GFP based sensor systems developed previously (,, ). N. benthamiana plants were infected with the p1 silencing suppressor mutant (Cym1stop) of Cymbidium ringspot virus (). The recovered leaves of Cym1stop infected plants developed two weeks after inoculation showing mild or no symptoms and contain RISC complexes which cleave the CymRSV specific sequences (). These leaves were infiltrated with the GFP-Cym sirna-sensor construct and as negative control we used a GFP-PoLV sensor, which expresses a GFP mrna containing a Pothos latent virus (PoLV) sequence at the UTR (). This second GFP sensor is not expected to be silenced because there are no RISC complexes programmed against the PoLV sequence. Both GFP-Cym and GFP-PoLV were expressed at high level after agro-infiltration in the leaves of non infected plants (Figure A lines 1,). However, when recovered leaves of Cym1stop infected plants were infiltrated, the GFP-Cym sensor mrna was efficiently targeted by viral sirna programmed RISC, reducing the level of sensor RNA and GFP expression. On contrary, the level of heterologous sensor Cym-PoLV remained high (Figure A lines,). The detection of cleavage product of GFP-Cym further confirmed the activity of virus sirna programmed RISC. As we expected from in vitro experiments the co-expression of p1-his with the sensor GFP-Cym was not able to inhibit the cleavage of sensor RNA similarly to His-HC-Pro (Figure A lanes -) and p1 (). To find out what is the effect of p1-his on mirna loaded RISC activity a similar experiment was carried out using a cleavable sensor (GFP-1.1) and a non-cleavable mutant sensor (GFP-

12 ) constructs (). The GFP-1.1 and GFP-1. constructs were infiltrated on the two sides of the same leaf, to ensure the same level of mir-1 loaded RISC and GFP fluorescence was monitored at dpi. Consistent to the previous results, the GFP mrna and protein levels were reduced in the GFP-1.1 sensor infiltrated patch (Figure B, lane 1) compared to the GFP- 1. control (Figure B lane ). This expression pattern was the same in the presence of p1 or the control HC-Pro (Figure B lanes -). The expression of the His-HC-Pro and p1-his proteins was demonstrated by Western blot analysis. These results strengthen our in vitro findings and demonstrate that p1 is not able to inhibit preloaded si- or mirisc activity in vivo, as observed in the in vitro Drosophila embryo system. p1 binds ds sirnas with size specific manner To explore the affinity of p1 to sirnas we carried out a more detailed analysis of the RNA binding affinity of this protein. Electrophoretic mobility shift assays were performed using labeled synthetic single- or double-stranded RNA oligonucleotides in different sizes and diluted plant protein extract of S-p1 infiltrated N. benthamiana leaf. In these experiements, the well characterized S-p1 of Carnation Italian ringspot virus () infiltrated N. benthamiana leaf extract was used as control (data not shown). P1 did not show any single stranded RNA bindig activity irrespectively of the length of RNA (data not shown). The relative dissociation constant (K r ) of p1 for ds sirna was in the same range as p1 infiltrated extract (not shown). p1 bound to nt RNA duplexes with nt overhang with the highest affinity (Figure A,C, squares). The size and the overhangs of the sirna proved to be important because the binding affinity was slightly reduced when 1 nt blunt (Figure A,C, circles) or nt blunt (Figure C, up triangles) duplex RNAs were used, K r/1 = and K r/ =, respectively (K r of p1 for nt long sirna is considered to be 1). The affinity for nt sirna species (Figure C, down triangles) was much lower: K r/ =. P1 did not bind the other RNAs tested: 1nt RNA with nt overhang nt RNA with nt overhang, nt duplex RNA (Figure C) and 1-, -, nt single-stranded RNAs (data not shown). Since viral sirnas are predominantly - nt long, these results suggest a specific adaptation of the virus to counteract the antiviral silencing machinery. To analyze the natural function of p1 during the course of virus infection, we tested the sirna binding affinity of plant extract derived from wt cr-tmv infected plants. Figure E shows that the plant extract from wt virus infected plants has the same binding activity as 1

13 transiently expressed p1: it binds nt sirnas but does not bind nt sirnas. The mobility of p1-sirna complex is the same as wt virus protein-sirna complex suggesting that p1 is the only viral protein, which binds sirna and the p viral replicase does not have sirna binding activity. We were not able to test directly of p for RNA binding because a S-p- HA construct failed to express detectable amount of p-ha. However, indirect evidence suggesting that p did not bind sirnas was obtained. We created a mutant cr-tmv in which the amber stop codon of p1 was substituted by a tyrosine codon. The mutant virus (cr-tmv- 1) was able to replicate, although at lower rate than the wt (not shown) and the protein extract from cr-tmv- 1 infected plants did not show sirna binding activity (Figure F). This result suggests that the p readtrough product of p1 doesn t have sirna binding activity. p1 bind mirna duplexes and interferes with the mirna accumulation and the terminal methylation It is becoming clear that viruses not only suppress RNA silencing but the virus encoded silencing suppressors can also interfere with cellular functions that are controlled by plant small RNAs. This interference can contribute to viral symptoms as suggested previously for P1/HC-Pro of Turnip mosaic virus (). However, considering our findings that p1 has a specific affinity to bona fide nt ds-sirna only in a narrow size range, it was not predictable whether p1 was able to inhibit mirna pathways by binding ds mirna intermediates or not since these mirna/mirna* duplexes contain mismatches and bulges, which can modify the structure of mirna duplexes compared to a perfectly matched sirna duplex. To this end we tested synthetic mir1a, mir1b and mir1c RNA duplexes, having slightly different predicted structures ()()() and mi1 sirna having a perfectly matched star strand, in p1 binding assay. We found that the binding affinity of p1 to the three mirna/mirna* (mir1a, mir1b, mir1c) duplexes are in the same range, (K r =1.1, 1. and 1., respectively) and only slightly reduced as compared to the 1 sirna perfect duplex. (K r =1) (Figure B and D). To explore the possible effect of p1 on mirna pathways, A. thaliana Col-0 plants were infected with cr-tmv. After 1 days we observed severe mottling of systemically infected leaves and the edges of the leaves became serrated (Figure 1A) similar to those plants in which the mirna pathways were compromised by mutation () or silencing suppressor proteins were expressed (1). This observation suggested that p1 may also interfere with mirna pathways. 1

14 We hypothesized that if p1 could bind mirna/mirna* duplexes in vivo, the accumulation of mirna* would increase because the mirna duplexes would be stabilized by p1 binding, similarly to the previously reported silencing suppressors, which bind mirna duplexes (,, ). Indeed, upon cr-tmv infection both mature and star strands of all tested endogenous mirnas (mirc, mir1a, mir1c, mir1a, mir) accumulated to a higher level compared to the mock-infected control plants (Figure A). This observations are in line with our in vitro findings and strengthens the idea that p1 can interfere with mirna pathways by sequestering mirna duplexes and stabilizing them. However, virus infection itself without silencing suppressor protein can also increase the accumulation of mirnas (J. Burgyan unpublished results). Interestingly the accumulation of mir1 was particularly high in cr-tmv infected plants. This result prompted us to analyze the expression of AGO1, which is controlled by mir1 since it was recently reported that the expression of mir1 and its target AGO1 are coregulated as consequence of AGO1 homeostasis (). Indeed, the expression of AGO1 in cr- TMV infected plants was also elevated similarly to mir1 and the high level of AGO1 was neither influenced by the lack of methylation in hen1-1 plant () nor in plant hst-1 () in which the mirna nuclear export is affected () (Figure B). It is worth to note that although mir1 (Figure A), which regulates DCL1 accumulated at higher level in virus infected plants we did not observe significant downregulation of DCL1 mrna using quantitative RT-PCR analysis (data not shown). The finding that the mir1* also accumulated at higher level in virus infected plants (Figure A) suggests that p1 binds mir1/ mir1* duplexes thus prevents the mir1 mediated downregulation of DCL1. The effect of virus infection on the accumulation of trans-acting sirnas (tasirnas) was particularly interesting since tasirna biogenesis involves an initial mirna-guided cleavage of primary TAS RNA transcript, which identifies the starting point for RDR-directed complementary strand synthesis. Then the double-stranded product of RDR is processed by DCL to mature tasirnas. These tasirnas guide RISC complexes to cleave mrnas having complementary sequences to tasirna ()()(). We expected that p1 expressed during virus infection may interfer with the biogenesis of tasirnas similarly to other silencing suppressor proteins expressed transgenically (). The mirnas (mir for sir and mir0 for TASD+) involved in tasirna biogenesis accumulated higher similarly to the other tested mirnas in cr-tmv infected plants than in mock inoculated non infected plants (Figure A). 1

15 However, the accumulation of tasirnas (sir and TASD+) regulated by these mirnas was reduced, compared to mock inoculated plants (Figure A). These results indicate an inhibitory effect of p1 on tasirna biogenesis. We have also showed that p1 recognized the nt overhangs of sirna duplexes predicting that p1 may interfere with end methylation of endogenous small RNAs and viral sirnas. Plant srnas are likely methylated by HEN1 methyltransferase at their terminal nucleotide at the -hydroxyl group (, ). The methylation appears to protect them from oligouridilation and subsequent degradation () and it is present in all species of small RNA family (sirna, mirna, tasirna, sense- and hairpin transgene derived, transposon- and repeat derived sirnas). The methylation status of small RNAs can be assessed by treatment of the RNAs with sodium- periodate followed by β-elimination. The free OH groups are sensitive to the chemical modification and results in the elimination of the last nucleotide from the RNA. The resulting molecules will migrate faster in gel electrophoresis. To test whether p1 is able to interfere with the methylation of small RNAs, we checked the methylation status of several mirnas, tasirnas and viral sirnas generated upon virus- infection or in agroinfiltration transient assay. In virus infected A. thaliana leaf RNA extract the tested mirnas and tasirnas showed a partial inhibition in the methylation level as compared to mock-inoculated plants (Figure A). All mirnas tested were fully methylated in mock- inoculated plants, while mirnas - both mature and star strands - derived from cr-tmv infected plants were partially non methylated., which was indicated the by the faster migration of mirnas that underwent β-elimination (Figure A). Interestingly the methylation of mirc, mir1c and mir1a* were only little effected in the presence of cr-tmv. Moreover the mature as well as the star strands of mira migrated in three different size classes in virus infected plants incontrast to mock inoculated plants where only the nt mira was accumulated. The appearance of the and nt species are not DCL or DCL dependent (supplementary Figure 1). Similarly to mira the and nt long species of mir1a, mir1a, mirc, mir1a, mir1c or shorter forms of mir1c were also observed (supplementary Figure 1). The partial inhibition of mirna and viral sirna methylation in the virus infected plants were somewhat surprising, since HEN1 suggested to be localized in the nucleus, while cr-tmv replicates exclusively in cytoplasm (). Furthermore, we can rule out the exsistence of other unidentified methylase, which may operates in the cytoplasm since non of the mirnas an 1

16 sirnas were methylated in the in hen1-1 plants. These findings suggested that HEN1 is active either in the nucleus or in the cytoplasm. Moreover, the ability of virus to interfere with mirna methylation may suggest that mirnas exported from the nucleus to the cystoplasm are in both methylated and non-methylated forms. To better understand the effect of virus infection on methylation of mirnas and sirnas hst-1 plants - in which the mirna nuclear export is affected () - were also infected with cr-tmv. The virus infection resulted in elevated accumulation of all tested mirna including both mature and star strands. More importantly these mirnas were completely methylated suggesting that the virus infection could not inhibit the methylation of mirnas (Figure A). This finding is could be the consequence of the fact that the nuclear export of mirnas to the cytoplasm is compromised. In contrast to mirnas the methylation of virus sirnas were partially inhibited similarly to wt suggesting that the inactivated HST has no role in virus sirna biogeneis. The methylation status of tasirnas (sir) was similar to that of mirnas in wt plants. sir was fully methylated in the mock inoculated and partially in the virus infected plants (we were not able to analyse TASD+ RNA, because it accumulated at a very low level in virus infected plants). However, sir behaved differently in virus infected hst-1 plants in which the sir methylation was partially inhibited, suggesting that sir duplexes were available for p1 since they generated in the cytoplasm similarly to viral sirnas. These results further support the HEN1 activity in the cytoplasm. We have shown that p1 efficiently binds ds sirnas and mirna intermediates, which strongly suggests that the small RNA binding activity of p1 is responsible for the inhibition of small RNA methylation. To further support our hypothesis p1 was coexpressed with S-pri- mirna1c hairpin (that generates mirna1c duplexes) using agroinfiltration. As a control S-pri-miRNA1c was also co-infiltrated with S-HC-Pro or empty vector. When S-primiRNA1c was co-infiltrated with the empty vector the generated mir1c mature and mir1c RNAs were resistant o the β-elimination treatment. It is worthy to note that the amount of mirna in the infiltrated patch is in high excess as compared to the natural mirna levels, but even in this case the methylation was 0% (Figure A lanes,), suggesting that this step in mirna biogenesis is very efficient. Importantly, β-elimination reaction was complete as demonstrated by the synthetic oligonucleotide controls (Figure A lanes -1). The co-expressed p1 partially inhibited mir1c methylation (Figure A lanes,) similarly to virus infected plants (Figure ) and to HC-Pro (Figure lanes,). Interestingly, the mir1c* strand was fully 1

17 methylated in p1 infiltrated plants, suggesting that one of the two termini of mir1c/mir1c is not protected by p1 and it is accessible for HEN1 mediated methylation. (Figure A lines,). We also tested the effecet of p1 on transgene-derived sirnas and found that p1 was able to interfere mostly with the methylation of nt long species, affected only very slightly the methylation of nt sirna and did not interfere with the methylation of nt sirna (Figure B lane ). Our finding that p1 binds nt sirnas with much lower efficiency than nt sirna could explain why p1 did not block the methylation of nt sirnas. DISCUSSION The p1 protein of cr-tmv is the homologous gene of the previously reported p1 protein of TMV, which is the subunit of viral replicase complex (). The p1 forms complex with the p1 protein, the readthrough product of p1, at a one to one ratio.(). However, it was also shown that p1 expressed approximately in times molar excess to p1 suggesting that the p1 protein has other important functions in the virus life cycle (). Indeed, our finding demonstrated that the p1 of cr-tmv (the corresponding protein of p1) is also functioning as a silencing suppressor protein apart from its function in the virus replication. The combination of replicase and the silencing suppression functions of the same protein are likely to be advantageous for the virus, since the replicating RNA is probably the most exposed form of the viral genome to silencing mediated RNA degradation. In this work we present an extensive characterization of silencing function of cr-tmv p1 protein. The p1 protein of cr-tmv efficiently suppress RNA silencing by sirna sequestration The analysis of the effect of p1 silencing suppresor in agroinfiltration assays indicated that p1 inhibits RNA silencing triggered by both sense and dsrnas. We propose that the molecular basis of inhibition in both cases relies on the ability of p1 to sequester sirna duplexes. Indeed, when GFP RNA was coexpressed with GF-IR and p1 the Dicer activity was clearly not inhibited, although p1 efficiently blocked GFP RNA targeting probaly by inhibiting the sirna incorporation in RISC complexes via sirna binding. Moreover, we also showed in this

18 experiment that p1 inhibited the accumlation of secondary sirnas (see Figure 1G P-specific sirnas), which requires the action of RDR (). We also suggest that p1 inhibits the action of RDR in the case of GFP sense RNA silencing by binding the minute amount of primary sirna, which is essential for RDR mediated amplification of RNA silencing through the production of secondary sirnas. Agreeing with our observation it was shown very recently that several silencing suppressor proteins expressed transgenically inhibited RDR dependent steps in the RNA silencing patways (). In line with our sirna sequestration model we also showed that p1 inhibited RNA silencing in quantitative manner, which can explain our observation that the silencing of GFP expression in Amp plants was completely, while in Amp x GFP plants it was only partially suppressed. In the case of Amp x GFP the silencing is much more robust and generates much higher level of GFP specific sirnas (1) and it is very likely that the high amount of sirna can t be sequestered completely by p1. This may suggests that in the veins of AmpxGFP plants the level of GFP sirnas is lower (as low as in Amp plants) than in the interveinal tissue. Our in vivo and in vitro studies further demonstrated that p1 acts upstream to RISC programming step, since the loaded sirisc and mirisc cannot be inhibited by the presence of the p1 protein. This feature resembles to the ToMV suppressor p, which had been suggested to suppress GFP silencing but cannot interfere with the pre-programmed RNA silencing machinery (). Indeed, our results demonstrated that molecular bases of silencing suppression by p1 is the sequestration of sirnas thus preventing the assembly of sirna containing effector complexes in the antiviral silencing response. We have shown that p1 selects for bona fide sirna for binding similarly to p1 of tombusviruses and HC-Pro of TEV (, ). This finding further supports the model of sirna-risc mediated targeting of viral RNA (, ) in contrast to the theoretical possibility that dicer activity alone could control the efficient viral invasion. Since p1 is part of the replicase complex it is interesting that p1 does not bind nt long double-stranded or single-stranded RNA molecules. Although, it is possible that in the replicase complex p1 has different RNA binding activity or an other component of the viral replicase complex binds the replicating long viral RNAs. Cr-TMV infection enhances the accumulation of mirnas and downregulates tasirnas 1

19 P1 efficiently binds ds mirna intermediates in vitro and the simultaneous accumulation of mirna/mirna* in cr-tmv infected plant suggests that p1 binds mirna duplexes also in planta. The symptoms of cr-tmv infected plants may indicate that the mirna pathways are compromised. The in vitro band shift assay demonstrated that the mirna mismatches did not alter significantly the binding affinity of p1 protein to mirna duplexes, which is in line with the finding that the accumulation of tested mirnas and mirna*s were increased in virus infected plants. It is not clear whether this is due to stabilization of the mirna duplexes in the mirna-p1 complex or the large amount of viral sirnas alter the synthesis-degradation equilibrium of mirna biogenesis. The mirnas are produced by DCL1 protein, which is regulated through a feed-back loop by mir1 RNA (1). Elevated level of DCL1 mrna was detected in dcl1 and hen1-1 mutant plants, or in plants expressing a viral suppressor TuMV P1/HC-Pro, which inhibits mirna-guided degradation of DCL1 mrna (1). In cr-tmv infected leaves both mir1a and mir1a*, similarly to other mirna/mirna* pairs accumulated at a higher level. Their methylation was partially blocked suggesting that p1 binds mir1a/mir1a* duplexes in vivo, which can lead to stabilization thus the increased accumulation of these mirnas. Alternatively or additionally, the elevated level of mirnas might be also the consequence of enhanced DCL1 enzyme activity upon the viral infection and not simply the result of p1 expression. However, p1 may mask the effect of enhanced mirna levels by sequestering many of the mirna duplexes including mir1, thus blocking the down-regulation of DCL1 mrna. Mir1a and its corresponding target AGO1 accumulated particularly high level during the development of virus infection, suggesting that AGO1 homeostasis () keeps the balance between the mir1 and AGO1 expressions. In contrast to the enhanced accummulation of mirnas the accumulation of tested tasirnas was reduced in cr-tmv infected plants. Although the level of mirnas including mir and mir0 was higher in virus infected plants, their biological activity was likely inhibited by p1 sequestration as suggested by the low level of tasirna accummulation, which requires the cleavage of TAS1 and TAS RNAs madiated by mir and mir0 respectively. In addition to inhibition of mir and mir0 action p1 may interfer with tasirna biogenesis at an additional stage. Indeed we showed that the tested tasirnas were only partially methylated in cr- TMV infected plants, demonstrating that the virus infection interfers with tasirna mediated 1

20 pathways at different steps. Our findings are in line with recent observations that tasirnas (sir) are downregulated and partially unmethylated by in an other tobamovirus Oilseed rape mosaic virus (ORMV) infected plants (). P1 inhibits the methylation of small RNAs Previous results demonstrated that transgenically expressed silencing suppressor proteins, which are able to bind sirnas inhibit terminal methylation (). In addition, inhibition HEN1 mediated methylation of small RNAs has been observed in (ORMV) infected Arabidopsis plants (1)(). We have shown in this study that the methylation of viral sirnas, transgene derived sirnas and mirnas is inhibited by p1 likely through the binding of mi/sirna duplexes. Our results also demonstrated that HEN1 operates both in the nucleus and cytoplasma since cr-tmv as many other positive strand RNA viruses replicates in cytoplasm and the virus derived sirnas were partially methylated (Figure ). Importantly, we can exclude the activity an other not unidentified methylase, which may operates in the cytoplasm since all mirnas an sirnas were not methylated in the hen1-1 plants. The ability of the replicating virus to interfere with mirna methylation suggests that mirnas are exported from the nucleus to the cytoplasm in both methylated and non-methylated forms. The infection of hst-1 plants - in which the mirna nuclear export might be compromised - resulted in the elevated accumulation of all tested mirna and these mirnas were completely methylated. This demonstrated that the virus infection could not inhibit the methylation of mirnas, likely because they were separated in different compartments while the methylation of virus sirnas were partially inhibited (as in wt plants) since they accumulated in the same compartment where the virus replicated. Indeed the activity of HEN1 in the cytoplasm was supported by a recent report, which demonstrates that HEN1 is present both in the nucleus and the cytoplasm (1). Another puzzling result is the asymmetrical methylation of the mirnas upon virus infection. In some cases the mature strand is protected by the methyl group and the star strand is not (mir1/mir1*, not shown), while in other cases we observed the opposite situation where the mir1c methylation was partially blocked but the mir1* was not (Figure ). The asymmetrical methylation could be a consequence of two step kinetics of the HEN1 enzyme. In this scenario the suppressors binds the mirna after the first step in which one of the strands is

21 methylated. It is also possible that the p1-mirna complex is not symmetrical, therefore HEN1 and/or p1 recognize the ends of mirna duplexes asymmetrically. The difference between the suppression of methylation of different species could be due to a spatial and temporal compartmentalization of the different mirnas and p1 protein. We also observed that the transgene derived nt sirnas were sensitive to the β-elimination reaction in the p1 infiltrated tissue indicating that they were not methylated, while the nt sirnas were fully methylated. These results are likely the consequence of the preferential binding of nt sirnas versus nt by the p1 suppressor protein further supporting our finding that p1 binds sirna duplexes in a size specific manner. In conclusion, the multifunctional p1 protein of cr-tmv is a very potent silencing suppressor protein, which blocks the intermediate steps of antiviral and endogenous silencing pathways preventing the assembly of DCL enzymes generated si/mirna duplexes into effector molecules such as RDR, HEN1 and RISC. ACKNOWLEDGEMENTS We are grateful to Olivier Voinnet providing the mir1.1 and mir1. sensor constructs. We thank to David Baulcombe for seeds of 1C plants. We also thank to György Szittya for critical reading and to Loránt Lakatos for valuable advices during the work. T.Cs. was recipient of grant of Ministry of Educations and Culture for off broad Hungarians (Number: A/-000). J.B. was recipient of grant of Royal Society for visiting scientist. This research was supported by grants from the Hungarian Scientific Research Fund (OTKA; T0 and OTKA; NK0), and the SIROCCO EU project LSHG-CT FIGURE LEGENDES Figure 1: The effects of cr-tmv infection and the expression of p1 silencing suppressor. Infection of Arabidopsis thaliana with cr-tmv leads to strong phenotype after weeks of

22 infection (A) compared to the non-infected plants (B). Reversion of silenced GFP in cr-tmv infected transgenic AmpxGFP (C) and Amp (D) plants at 1 days post inoculation. GFP fluorescence was assessed in the plants under UV light using a dissecting microscope and photographs were taken at 1 days after inoculation. (E) Accumulation of virus specific sirnas during cr-tmv infection. Total RNA was extracted from cr-tmv infected N. bethamiana at dpi and A. thaliana plants at 1 dpi and separated on denaturing agarose (for viral RNAs) and in 1% denaturing polyacrylamide gel (for sirnas) (F) Suppression of RNA silencing by p1 in N. benthamiana GFP1C leaves were agroinfiltrated with S-GFP, S-GF-IR, S-HcPro and S-p1, as indicated. The GFP fluorescence was monitored under UV light and photographs were taken at days after agroinfiltration. (G) P1 silencing suppressor inhibits the GFP RNA degradation but does not impair primary sirnas production. Leaves of N. benthamiana GFP1C line were infiltrated with S-GFP (lane 1), S-GFP and S-GF-IR (lane ), S-GFP and HC-Pro (lane ), S-GFP and p1 (lane ), S-GFP, S- GF-IR and HcPro (lane ), S-GFP, S-GF-IR and p1 (lane ), S-GFP, S-GF-IR and sigma (lane ). The RNA samples extracted 0 hours after infiltration were subjected to Norther analysis using appropriate probes to detect GFP mrna and GF-IR RNA, and GFP, GF and P specific sirnas. Figure : p1 protein inhibits sirna-guided target cleavage. P1 inhibits target cleavage in vitro in the direct competition assay (lines -), but did not interfere with pre-programmed RISC activity (indirect competition assay, lines -1). In the direct competition assay Drosophila embryo extract, target RNA (0. nm), labelled sirna (nm) and either empty-vector infiltrated (line 1) or p1-infiltrated N. benthamiana plant extracts were used at different dilutions and all components were added simultaneously. In the indirect competition assay sirna were incubated with the Drosophila embryo extract for minutes then target RNA and p1-infiltrated N. benthamiana plant extracts were added in different dilutions. The effect of p1 on RISC mediated cleavage was monitored by the detection of cleavage products.

23 Figure : In vitro RISC formation is inhibited by p1 protein. In the direct competition assay Drosophila embryo extract, labeled sirna and pbin-empty vector infiltrated (line ) or p1- infiltrated plant extract dilutions (lines -) were added at the same time. In the indirect competition reactions embryo extracts were preincubated with labelled sirna, prior to addition of the pbin-infiltrated (line 1) or p1-infiltrated plant extracts (lines 1-). Control reactions were labeled sirna only (line 1), embryo extract and labeled sirna (line ), sirna and pbinp1 infiltrated or empty-vector infiltrated plant extract (line and, respectively). The different forms of sirna containing silencing related complexes were separated in.% native acrylamide gel. The position of RISC, RISC loading complex (RLC), sirna-dicer-rd and p1-sirnas are indicated. Figure : p1 protein does not inhibit preassembled sirisc or mirisc activity in vivo. (A) GFP-Cym or GFP-PolV sensor constructs had been infiltrated in Cym1stop infected recovery leaves. GFP mrna and protein were analyzed at dpi in northern blot and western immunoblot assays, respectively. (B) Pre-programmed mirisc activity is not affected by the presence of p1. The sensor construct GFP-1.1, bearing a perfect target site for mir1 is cleaved irrespectively of the presence of the p1 protein. Non-cleveable GFP-1. sensor, which bears a mutation in the target site was used as control. HC-Pro coinfiltration was also used as a control. For western blot analysis anti-gfp and anti-his antibodies were applied. Figure : Affinity of p1 protein to different RNA duplexes. nt bona fide sirna and 1 nt blunt-ended RNA duplex (A), or nt sirna and mir1 duplexes (B), were incubated with a dilution series of p1 infiltrated plant extract and loaded on % native 0.xTBE gel. (C) Determination of relative binding affinity of p1 extract for different RNA molecules nt sirna (black squares), 1 nt blunt (red circles), nt blunt (green up-triangles), nt sirna (blue down-triangles), nt sirna (cyan diamonds), dsrna (purple plus), 1 nt sirna (grey cross). (D) Relative affinity of p1 protein for mirna duplexes mir1a (red circle), mir1b (green up-triangles) and mir1c (blue down-triangles) compared with the sir1 (black squares). (E) Binding affinity using pbin-p1 infiltrated or cr-tmv infected plant extract for nt sirna and for nt dsrna. Complexes formed run at the same mobility. Control

24 reactions are shown without protein extract or with empty vector infiltrated plant extract. (F) Plant extract infected with mutant virus not coding for p1 do not show any binding. Figure : Accumulation and terminal methylation of mirnas and sirnas upon cr-tmv infection. (A) Elevated accumulation of the different mirnas and mirna*s have been observed in cr-tmv infected plants. Total RNAs were extracted from cr-tmv-infected or non-infected wild type or mutant A. thaliana plants as indicated. For hybridization labelled LNA oligonucleotides complementary to indicated mirna and mirna* were used. For virus-derived sirnas and for U as loading control, labeled complementary transcripts were applied as probes. For the identification of methylation status of mirnas and sirnas total RNAs from infected and non-infected plants were subjected (+) or not (-) to the β-elimination reactions. (B) Northern analysis of AGO1 mrna from mock and virus infected A. thaliana wild type and mutant plants. Mouse total RNA was used as control. Figure : The effect of p1 protein on the methylation of GFP-derived sirna and overexpressed mirna 1c. Total RNA extracts were subjected (+) or not (-) to β -elimination reactions and blots were hybridized as in Figure. (A) GFP-IR derived sirnas are fully methylated in the absence of suppressor proteins. HC-Pro completelly inhibits the methylation of nt long sirnas and partially inhibit the methylation of nt sirnas (line ). P1 protein partially interferes with methylation of and nt long sirna species (line ). Neither suppressor blocks the methylation of nt sirna molecules (lines, ). Non-methylated synthetic RNA oligos were used as positive control for β-elimination reactions (lines, ). (B) Overexpressed mir1c and mir1c* are fully methylated in the absence of silencing suppressors (line ). HC-Pro and p1 proteins partially inhibit the methylation of mir1c (lines and ), but not the mir1c* Positive controls for the β-elimination test are shown in lines, Akbergenov, R., A. Si-Ammour, T. Blevins, I. Amin, C. Kutter, H. Vanderschuren, P. Zhang, W. Gruissem, F. Meins, Jr., T. Hohn, and M. M. Pooggin. 0. Molecular characterization of geminivirus-derived small RNAs in different plant species. Nucleic Acids Res :-1.

25 Alefelder, S., B. K. Patel, and F. Eckstein. 1. Incorporation of terminal phosphorothioates into oligonucleotides. Nucleic Acids Res :-.. Allen, E., Z. Xie, A. M. Gustafson, and J. C. Carrington. 0. microrna-directed phasing during trans-acting sirna biogenesis in plants. Cell 1:-.. Baulcombe, D. 0. RNA silencing in plants. Nature 1:-.. Baumberger, N., and D. C. Baulcombe. 0. Arabidopsis ARGONAUTE1 is an RNA Slicer that selectively recruits micrornas and short interfering RNAs. Proc Natl Acad Sci U S A :-.. Blevins, T., R. Rajeswaran, P. V. Shivaprasad, D. Beknazariants, A. Si-Ammour, H. S. Park, F. Vazquez, D. Robertson, F. Meins, Jr., T. Hohn, and M. M. Pooggin. 0. Four plant Dicers mediate viral small RNA biogenesis and DNA virus induced silencing. Nucleic Acids Res :-.. Boutet, S., F. Vazquez, J. Liu, C. Beclin, M. Fagard, A. Gratias, J. B. Morel, P. Crete, X. Chen, and H. Vaucheret. 0. Arabidopsis HEN1. A Genetic Link between Endogenous mirna Controlling Development and sirna Controlling Transgene Silencing and Virus Resistance. Curr Biol 1:-.. Brodersen, P., and O. Voinnet. 0. The diversity of RNA silencing pathways in plants. Trends Genet :-0.. Chao, J. A., J. H. Lee, B. R. Chapados, E. W. Debler, A. Schneemann, and J. R. Williamson. 0. Dual modes of RNA-silencing suppression by Flock House virus protein B. Nat Struct Mol Biol 1:-.. Chapman, E. J., A. I. Prokhnevsky, K. Gopinath, V. V. Dolja, and J. C. Carrington. 0. Viral RNA silencing suppressors inhibit the microrna pathway at an intermediate step. Genes Dev 1:1-.. Chen, X. 0. MicroRNA biogenesis and function in plants. FEBS Lett : Dalmay, T., A. Hamilton, E. Mueller, and D. C. Baulcombe. 00. Potato virus X amplicons in arabidopsis mediate genetic and epigenetic gene silencing. Plant Cell 1:-. 1. Dalmay, T., L. Rubino, J. Burgyan, A. Kollar, and M. Russo. 1. Functional analysis of cymbidium ringspot virus genome. Virology 1: Dorokhov Yu, L., P. A. Ivanov, V. K. Novikov, A. A. Agranovsky, S. Morozov, V. A. Efimov, R. Casper, and J. G. Atabekov. 1. Complete nucleotide sequence and genome organization of a tobamovirus infecting cruciferae plants. FEBS Lett 0:-. 1. Dunoyer, P., C. H. Lecellier, E. A. Parizotto, C. Himber, and O. Voinnet. 0. Probing the microrna and small interfering RNA pathways with virus-encoded suppressors of RNA silencing. Plant Cell 1: Fang, Y., and D. L. Spector. 0. Identification of Nuclear Dicing Bodies Containing Proteins for MicroRNA Biogenesis in Living Arabidopsis Plants. Curr Biol.. Goregaoker, S. P., D. J. Lewandowski, and J. N. Culver. 01. Identification and functional analysis of an interaction between domains of the 1/1-kDa replicaseassociated proteins of tobacco mosaic virus. Virology :-. 1. Haley, B., G. Tang, and P. D. Zamore. 0. In vitro analysis of RNA interference in Drosophila melanogaster. Methods 0: Herr, A. J., M. B. Jensen, T. Dalmay, and D. C. Baulcombe. 0. RNA polymerase IV directs silencing of endogenous DNA. Science 0:-.. Hull, R. 0. Matthews' Plant Virology, Fourth ed. Academic Press, San Diego, California, USA.

26 Ishikawa, M., T. Meshi, F. Motoyoshi, N. Takamatsu, and Y. Okada. 1. In vitro mutagenesis of the putative replicase genes of tobacco mosaic virus. Nucleic Acids Res 1:1-0.. Kanno, T., B. Huettel, M. F. Mette, W. Aufsatz, E. Jaligot, L. Daxinger, D. P. Kreil, M. Matzke, and A. J. Matzke. 0. Atypical RNA polymerase subunits required for RNA-directed DNA methylation. Nat Genet :1-.. Kasschau, K. D., Z. Xie, E. Allen, C. Llave, E. J. Chapman, K. A. Krizan, and J. C. Carrington. 0. P1/HC-Pro, a viral suppressor of RNA silencing, interferes with Arabidopsis development and mirna unction. Dev Cell :-.. Kubota, K., S. Tsuda, A. Tamai, and T. Meshi. 0. Tomato mosaic virus replication protein suppresses virus-targeted posttranscriptional gene silencing. J Virol :-.. Lakatos, L., T. Csorba, V. Pantaleo, E. J. Chapman, J. C. Carrington, Y. P. Liu, V. V. Dolja, L. F. Calvino, J. J. Lopez-Moya, and J. Burgyan. 0. Small RNA binding is a common strategy to suppress RNA silencing by several viral suppressors. Embo J :-0.. Lakatos, L., G. Szittya, D. Silhavy, and J. Burgyan. 0. Molecular mechanism of RNA silencing suppression mediated by p1 protein of tombusviruses. Embo J :-. Epub 0 Feb 1.. Li, J., Z. Yang, B. Yu, J. Liu, and X. Chen. 0. Methylation protects mirnas and sirnas from a '-end uridylation activity in Arabidopsis. Curr Biol 1:1-.. Lichner, Z., D. Silhavy, and J. Burgyan. 0. Double-stranded RNA-binding proteins could suppress RNA interference-mediated antiviral defences. J Gen Virol :-0.. Lippman, Z., and R. Martienssen. 0. The role of RNA interference in heterochromatic silencing. Nature 1: Mallory, A. C., and H. Vaucheret. 0. Functions of micrornas and related small RNAs in plants. Nat Genet Suppl:S Matranga, C., Y. Tomari, C. Shin, D. P. Bartel, and P. D. Zamore. 0. Passengerstrand cleavage facilitates assembly of sirna into Ago-containing RNAi enzyme complexes. Cell 1:0-.. Merai, Z., Z. Kerenyi, S. Kertesz, M. Magna, L. Lakatos, and D. Silhavy. 0. Double-stranded RNA binding may be a general plant RNA viral strategy to suppress RNA silencing. J Virol 0:-.. Moissiard, G., E. A. Parizotto, C. Himber, and O. Voinnet. 0. Transitivity in Arabidopsis can be primed, requires the redundant action of the antiviral Dicer-like and Dicer-like, and is compromised by viral-encoded suppressor proteins. Rna 1:1-.. Molnar, A., T. Csorba, L. Lakatos, E. Varallyay, C. Lacomme, and J. Burgyan. 0. Plant virus-derived small interfering RNAs originate predominantly from highly structured single-stranded viral RNAs. J Virol :1-.. Onodera, Y., J. R. Haag, T. Ream, P. C. Nunes, O. Pontes, and C. S. Pikaard. 0. Plant nuclear RNA polymerase IV mediates sirna and DNA methylation-dependent heterochromatin formation. Cell 1:1-.. Palatnik, J. F., E. Allen, X. Wu, C. Schommer, R. Schwab, J. C. Carrington, and D. Weigel. 0. Control of leaf morphogenesis by micrornas. Nature :-.. Pantaleo, V., G. Szittya, and J. Burgyan. 0. Molecular Bases of Viral RNA Targeting by Viral Small Interfering RNA-Programmed RISC. J Virol 1:-0.

27 Parizotto, E. A., P. Dunoyer, N. Rahm, C. Himber, and O. Voinnet. 0. In vivo investigation of the transcription, processing, endonucleolytic activity, and functional relevance of the spatial distribution of a plant mirna. Genes Dev 1:-.. Park, M. Y., G. Wu, A. Gonzalez-Sulser, H. Vaucheret, and R. S. Poethig. 0. Nuclear processing and export of micrornas in Arabidopsis. Proc Natl Acad Sci U S A : Pazhouhandeh, M., M. Dieterle, K. Marrocco, E. Lechner, B. Berry, V. Brault, O. Hemmer, T. Kretsch, K. E. Richards, P. Genschik, and V. Ziegler-Graff. 0. F- box-like domain in the polerovirus protein P0 is required for silencing suppressor function. Proc Natl Acad Sci U S A : Pfeffer, S., P. Dunoyer, F. Heim, K. E. Richards, G. Jonard, and V. Ziegler-Graff. 0. P0 of beet Western yellows virus is a suppressor of posttranscriptional gene silencing. J Virol :1-.. Pham, J. W., J. L. Pellino, Y. S. Lee, R. W. Carthew, and E. J. Sontheimer. 0. A Dicer--dependent 0s complex cleaves targeted mrnas during RNAi in Drosophila. Cell 1:-.. Qi, Y., A. M. Denli, and G. J. Hannon. 0. Biochemical specialization within Arabidopsis RNA silencing pathways. Mol Cell 1:-.. Schwarz, D. S., G. Hutvagner, T. Du, Z. Xu, N. Aronin, and P. D. Zamore. 0. Asymmetry in the assembly of the RNAi enzyme complex. Cell :1-.. Silhavy, D., and J. Burgyan. 0. Effects and side-effects of viral RNA silencing suppressors on short RNAs. Trends Plant Sci :-.. Silhavy, D., A. Molnar, A. Lucioli, G. Szittya, C. Hornyik, M. Tavazza, and J. Burgyan. 0. A viral protein suppresses RNA silencing and binds silencing-generated, - to -nucleotide double-stranded RNAs. Embo J : Sunkar, R., and J. K. Zhu. 0. Novel and stress-regulated micrornas and other small RNAs from Arabidopsis. Plant Cell 1: Szittya, G., A. Molnar, D. Silhavy, C. Hornyik, and J. Burgyan. 0. Short defective interfering RNAs of tombusviruses are not targeted but trigger post-transcriptional gene silencing against their helper virus. Plant Cell 1:-.. Tomari, Y., T. Du, B. Haley, D. S. Schwarz, R. Bennett, H. A. Cook, B. S. Koppetsch, W. E. Theurkauf, and P. D. Zamore. 0. RISC assembly defects in the Drosophila RNAi mutant armitage. Cell : Tomari, Y., and P. D. Zamore. 0. Perspective: machines for RNAi. Genes Dev 1:-. 1. Vargason, J., G. Szittya, J. Burgyan, and T. M. Hall. 0. Size selective recognition of sirna by an RNA silencing suppressor. Cell :-.. Vaucheret, H., A. C. Mallory, and D. P. Bartel. 0. AGO1 homeostasis entails coexpression of MIR1 and AGO1 and preferential stabilization of mir1 by AGO1. Mol Cell :1-.. Vazquez, F., H. Vaucheret, R. Rajagopalan, C. Lepers, V. Gasciolli, A. C. Mallory, J. L. Hilbert, D. P. Bartel, and P. Crete. 0. Endogenous trans-acting sirnas Regulate the Accumulation of Arabidopsis mrnas. Mol Cell 1:-.. Voinnet, O. 0. Induction and suppression of RNA silencing: insights from viral infections. Nat Rev Genet :-.

28 Wang, X. J., J. L. Reyes, N. H. Chua, and T. Gaasterland. 0. Prediction and identification of Arabidopsis thaliana micrornas and their mrna targets. Genome Biol :R.. Watanabe, T., A. Honda, A. Iwata, S. Ueda, T. Hibi, and A. Ishihama. 1. Isolation from tobacco mosaic virus-infected tobacco of a solubilized template-specific RNAdependent RNA polymerase containing a 1K/1K protein heterodimer. J Virol :-0.. Wesley, S. V., C. A. Helliwell, N. A. Smith, M. B. Wang, D. T. Rouse, Q. Liu, P. S. Gooding, S. P. Singh, D. Abbott, P. A. Stoutjesdijk, S. P. Robinson, A. P. Gleave, A. G. Green, and P. M. Waterhouse. 01. Construct design for efficient, effective and high-throughput gene silencing in plants. Plant J :1-0.. Xie, Z., E. Allen, N. Fahlgren, A. Calamar, S. A. Givan, and J. C. Carrington. 0. Expression of Arabidopsis MIRNA genes. Plant Physiol 1:-.. Xie, Z., E. Allen, A. Wilken, and J. C. Carrington. 0. DICER-LIKE functions in trans-acting small interfering RNA biogenesis and vegetative phase change in Arabidopsis thaliana. Proc Natl Acad Sci U S A : Xie, Z., L. K. Johansen, A. M. Gustafson, K. D. Kasschau, A. D. Lellis, D. Zilberman, S. E. Jacobsen, and J. C. Carrington. 0. Genetic and Functional Diversification of Small RNA Pathways in Plants. PLoS Biol :E. 1. Xie, Z., K. D. Kasschau, and J. C. Carrington. 0. Negative Feedback Regulation of Dicer-Like1 in Arabidopsis by microrna-guided mrna Degradation. Curr Biol 1:-.. Ye, K., L. Malinina, and D. J. Patel. 0. Recognition of small interfering RNA by a viral suppressor of RNA silencing. Nature :.. Yoshikawa, M., A. Peragine, M. Y. Park, and R. S. Poethig. 0. A pathway for the biogenesis of trans-acting sirnas in Arabidopsis. Genes Dev 1:-.. Yu, B., E. J. Chapman, Z. Yang, J. C. Carrington, and X. Chen. 0. Transgenically expressed viral RNA silencing suppressors interfere with microrna methylation in Arabidopsis. FEBS Lett 0:1-.. Yu, B., Z. Yang, J. Li, S. Minakhina, M. Yang, R. W. Padgett, R. Steward, and X. Chen. 0. Methylation as a crucial step in plant microrna biogenesis. Science 0:-.. Zhang, X., Y. R. Yuan, Y. Pei, S. S. Lin, T. Tuschl, D. J. Patel, and N. H. Chua. 0. Cucumber mosaic virus-encoded b suppressor inhibits Arabidopsis Argonaute1 cleavage activity to counter plant defense. Genes Dev :-.. Zilberman, D., X. Cao, and S. E. Jacobsen. 0. ARGONAUTE control of locusspecific sirna accumulation and DNA and histone methylation. Science :1-.

29 A C F GFP GFP GFP + GF-IR B D GFP + HC-Pro GFP + p1 G 0nt E nt nt 0nt nt 0nt nt GFP N. b. A. th. mock GFP + HC-Pro cr-tmv GFP + p1 mock cr-tmv GFP + sigma GFP + GF-IR GFP + GF-IR + HC-Pro cr-tmv genomic RNA rrna cr-tmv sirna rrna N. benthamiana GFP1C GFP + GF-IR GFp + GF-IR GFP + GF-IR + HC-Pro GFP + GF-IR + p1 GFP + GF-IR + p1-his GFP + GF-IR + p1 GFP + GF-IR + sigma 1 GFP mrna GF-IR RNA rrna GF sirna P sirna GFP sirna rrna Figure 1.

30 M 1xvector - Direct competition 1 1: 1: 1: 1: Indirect competition 1 1: 1: 1: 1: 1 1 p1 dilution target RNA cleavage product Figure.

31 p1 Nb-extr pbin Nb-extr Drosphila extr RISC RLC sirna-dcr -RD p1-sirna complex unbound sirna 1: 1: Direct comp. Indirect comp. 1: 1: 1: 1:1 1: 1: 1: 1: 1: 1: 1:1 1: 1: 1: Figure.

32 A GFP-Cym - - HC-Pro p1 GFP-PolV GFP-Cym GFP-PolV GFP-Cym GFP-PolV GFP-Cym GFP-PolV 1 non Cym1stop infected infected plants GFP mrna cleveage product rrna GFP p1 HC-Pro protein loading B GFP-1.1 GFP-1. GFP HC-Pro GFP p1 GFP HC-Pro GFP-1. + HC-Pro GFP p1 GFP-1. + p1 1 GFP mrna rrna GFP protein p1 HC-Pro protein loading Figure.

33 A B 1: 1: 1: 1:1 1: 1: 1:1 1: 1:1 - - P1-siRNA complex C C F cr-tmv P1-siRNA complex free RNA Figure. 1:1 cr-tmv 1 1: cr-tmv 1: p1 1 1:1 A - mock cr-tmv 1 1: 1: 1: 1 1: 1: 1: mock - p1 0,1 1: nt dsrna nt sirna 0,01 p1 relative concentration 1 1: 1: 1: 1 1: 1: 1: E 1E- 1: 1 1 0,1 1: 0,01 p1 relative concentration 0 1: 1E- 1: : 0 1: 0 0 1:1 0 sir1 mir1a mir1b mir1c 0 1: 0 Bound fraction (%) nt sirna 1nt blunt nt blunt nt sirna nt sirna nt dsrna 1nt sirna 1 1: 0 p1 Mock Bound fraction (%) D - T P E C - D E Free RNA 1: - mir1a 1: 1: 1: 1:1 1: 1: 1:1 1: 1:1 sir1 1nt blunt dsrna 1 1: 1: 1: 1:1 1: 1: 1:1 1: 1 1: 1: 1: 1:1 1: 1: 1:1 1: nt sirna

34 A Col-0 hen1-1 hst cr-tmv β -elimination mir1c mir1c* mir1a mir1a* mira mira* mirc mirc* mir1a mir1a* mir sir mir0 TASD+ sirna cr-tmv U B Mouse RNA Col-0 hen1-1 hst-1 Mock cr-tmv Mock cr-tmv Mock cr-tmv AGO1 Cleaved AGO1 Figure. rrna

35 A β-eliminaton B nt nt nt rrna nt rrna pri-mir1c pri-mir1c + HC-Pro pri-mir1c + p1 synthetic mir1c synthetic mir1c * β-elimination nt 0nt nt Mock Mock GFP-IR GFP-IR + HC-Pro GFP-IR + P1 synth. GFP sirna GFP sirnas -mir1c -mir1c* Figure.