7) 135–146www.elsevier.com/locate/yviro

Virology 367 (200

Characterization of small RNAs derived from Citrus exocortis viroid (CEVd)in infected tomato plants

Raquel Martín a, Catalina Arenas b, José-Antonio Daròs c, Alejandra Covarrubias b,José Luis Reyes b, Nam-Hai Chua a,⁎

a Laboratory of Plant Molecular Biology, The Rockefeller University, 1230 York Ave. New York, NY10021, USAb Instituto de Biotecnologia, UNAM. Ave. Universidad 2001 Col. Chamilpa 62210, Cuernavaca, Mor. Mexico de Biotecnología-UNAM, Cuernavaca, Mexico

c Instituto de Biología Molecular y Celular de Plantas (CSIC-Universidad Politécnica de Valencia), Avenida de los Naranjos s/n, 46022 Valencia, Spain

Received 14 March 2007; returned to author for revision 5 April 2007; accepted 7 May 2007Available online 8 June 2007

Abstract

In plants, RNA silencing provides an adaptive immune system that inactivates pathogenic nucleic acids, guided by 21–24-mer RNAs ofpathogen origin. The characterization of pathogen-related small RNAs (sRNAs) is relevant to uncovering the strategies used by pathogens to evadehost defense responses. Several groups have reported the detection of viroid-derived sRNAs during infections, although the origin of these sRNAsand their chemical characteristics were poorly understood. Here, we describe the in vivo cleavage of Citrus exocortis viroid (CEVd) RNA intosRNAs of 21–22 nt, that are phosphorylated at their 5′-end and methylated at 3′. Our studies suggested that the CEVd genomic RNA might be thepredominant in vivo substrate for cleavage by Dicer-like enzyme(s), which preferentially targeted residues mainly located within the right-enddomain of the viroid. Further analysis on the accumulation levels of specific miRNAs controlling major regulators of leaf development and themiRNA pathway and the levels of their target mRNAs provided evidence that the endogenous tomato miRNA pathway was not affected by CEVdinfection.© 2007 Elsevier Inc. All rights reserved.

Keywords: Small RNA; Viroid; DCL cleavage; microRNA pathway

Introduction

RNA silencing (RNAi) is a conserved defense mechanismplants and other eukaryotes use to protect their genomes againstaberrant nucleic acids (Baulcombe, 2004; Ding et al., 2004;Hannon and Conklin, 2004). This mechanism, which is me-diated by small RNAs of 21 to 24 nt, can be executed at thetranscriptional (transcriptional gene silencing or TGS) or post-transcriptional level (post-transcriptional gene silencing orPTGS). Other small RNA (sRNA)-related processes in plantsinclude trans-acting siRNA and microRNA pathways, whichregulate growth and development. PTGS provides plants anadaptive immune system that recognizes pathogenic nucleicacids such as those derived from viruses and inactivates themvia cleavage of their RNAs (Vance and Vaucheret, 2001;

⁎ Corresponding author. Fax: +1 212 327 8327.E-mail address: chua@rockefeller.edu (N.-H. Chua).

0042-6822/$ - see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.virol.2007.05.011

Voinnet, 2001; Waterhouse et al., 2001; Dunoyer and Voinnet,2005). As a counter-attack strategy against this host defensemechanism, many plant viruses encode proteins that interferewith different components of the PTGS machinery (Voinnet,2005; Li and Ding, 2006). Some examples of viral suppressorsof silencing are the tombusviral P19 protein that sequesters 21-bp duplex sRNAs (Silhavy et al., 2002; Lakatos et al., 2004;Sholthof, 2006), the helper component P1/HC-Pro of Turnipmosaic virus (TuMV) which interferes with miRNA-guidedfunctions (Anandalakshmi et al., 1998; Kasschau and Carring-ton, 1998; Kasschau et al., 2003) and the 2b protein of Cu-cumber mosaic virus that blocks Argonaute 1 (AGO1) cleavageactivity (Zhang et al., 2006). Interestingly, some viral RNAs,e.g., the human adenovirus virus associated (VA) RNA, can alsoact as silencing suppressors by interfering with the RNase IIItype Dicer-like protein (DCL) and the RNA-induced silencingcomplex (RISC) activities (Lu and Cullen, 2004; Andersson etal., 2005).

136 R. Martín et al. / Virology 367 (2007) 135–146

A general indicator of the occurrence of RNA silencing in acell is the presence of sRNAs of 21–26 nt (Elbashir et al., 2001;Lagos-Quintana et al., 2001; Lee and Ambros, 2001; Hamiltonet al., 2002; Llave et al., 2002; Borsani et al., 2005). These tinymolecules, generated from cleavage of perfect or imperfectdouble-stranded (ds) RNAs by the plant DCL proteins (Zamoreet al., 2000), guide the cleavage or translational repression oftarget RNAs with complementary sequences. The production ofvirus-derived sRNAs is associated with the activation of hostPTGS by viruses. These sRNAs are suggested to mediate thecleavage and degradation of viral RNAs because high levels ofvirus-derived sRNAs are normally correlated with a decrease inthe virus titer and a reduction in the progression of the infection.However, the accumulation of sRNAs derived from pathogenicnucleic acids is not always a disadvantage for the pathogen,because some animal DNA viruses encode miRNAs that maytarget host transcripts with defense function, thereby aiding theprogression of viral infection (Pffefer et al., 2004, 2005;Sullivan et al., 2005; Sullivan and Ganem, 2005; Cullen, 2006).

Little is still known about the origin, biogenesis and che-mical structure of virus-derived sRNAs. It is generally assumedthat perfect dsRNA intermediates, generated during viral repli-cation or by the action of RNA-dependent RNA polymerases,are the substrates for DCL cleavage during virus-induced PTGS(Dalmay et al., 2000; Mourrain et al., 2000). However, recentevidence suggests that, in the case of RNA viruses, highlystructured viral RNAs may serve as DCL substrates (Molnar etal., 2005). Processing of viral dsRNAs by DCL produces virus-derived sRNAs of 21–25 nt and various modifications at their5′- and 3′-ends. DNA begomoviruses produce small RNAs of21, 22 and 24 nt that are phosphorylated and methylated at their5′- and 3′-ends, respectively (Akbergenov et al., 2006). Bycontrast, Oilseed rape mosaic virus (ORMV) produces sRNAsof 20–22 nt with no modification at their 3′-ends. Indeed,ORMV seems to suppress methylation of the viral siRNAs anda subclass of endogenous small RNAs (Akbergenov et al.,2006).

Viroids are the smallest nucleic-acid-based pathogens infect-ing higher plants. At the molecular level they consist of a naked,covalently closed, single-stranded RNA of small size (250–400nt) that does not encode any open reading frame (Diener, 2003;Tabler and Tsagris, 2004; Flores et al., 2005; Daròs et al., 2006;Ding and Itaya, 2007). Even without encoding any protein,viroids are able to replicate in host cells and spread through thevascular system of plants to establish systemic infection (Dinget al., 1997; Zhu et al., 2001; Qi et al., 2004). Viroid replicationfollows a rolling-circle-like mechanism (Branch and Robertson,1984; Daròs et al., 1994; Navarro et al., 1999) and occurs in thenucleus (family Pospiviroidae) or the chloroplast (family Av-sunviroidae). Of particular interest is the specific secondarystructure that viroids acquire in vivo, owing to the peculiarcharacteristics of its RNA sequence, which is approximately60–70% self-complementary (Sogo et al., 1973). Viroid RNAlikely folds into a rod-like secondary structure defining fivestructural and functional domains in members of the Pospivir-oidae family (Keese and Simons, 1985). These five domainsinclude a left-terminal domain (1), a pathogenicity domain (2), a

central conserved region (3), a variable domain (4) and a right-terminal domain (5) (Fig. S1). Some lines of evidence suggestthat the in vivo existence of this particular secondary structuremay provide binding signals for host protein factors (Gozma-nova et al., 2003; Maniataki et al., 2003; Martínez de Alba et al.,2003) and protect the non-encapsidated viroid genome from anattack by cellular defense responses. Viroid-derived perfect andimperfect double-stranded RNA (dsRNA) species are detectedin host cells during the viroid life cycle, indicating that viroidRNAs are potential activators of RNA silencing. Indeed,working with different viroid species several laboratories havereported the detection of viroid-derived sRNAs during infec-tions (Itaya et al., 2001; Papaefthimiou et al., 2001; Martínez deAlba et al., 2002; Denti et al., 2004; Markarian et al., 2004;Landry and Perreault, 2005) although the origin of these sRNAsand their chemical characteristics are still poorly understood.

Here, we show the specific in vivo cleavage of Citrus exo-cortis viroid (CEVd) RNA into 21–22 nt sRNAs. The majorityof these sRNAs are of the positive (+) polarity, suggesting thatstructured viroid RNAs are the predominant substrates for DCLenzymatic cleavage. The pattern of CEVd RNA cleavage indi-cates the likely existence of domains preferred by DCL. More-over, we have characterized the chemical structure of the CEVd-derived sRNAs produced. In addition, our in vivo assayssuggest that although CEVd affects leaf morphology and plantdevelopment (cellular processes directly controlled by miR-NAs), the endogenous miRNA pathway is not affected inCEVd-infected tomato plants.

Results

High levels of CEVd sRNAs do not decrease viroidaccumulation

To investigate the characteristics and putative roles of CEVd-derived small RNAs (sRNAs), we analyzed the accumulationkinetics of these sRNAs at several stages after infection oftomato seedlings with dimeric CEVd RNAs. Three weeks afterinoculation, systemic tomato leaves began to show light symp-toms correlating with the appearance of the CEVd genomicRNA and the accumulation of CEVd sRNAs (Fig. 1). Weobserved that both genomic and sRNA levels increased duringthe progression of the infection. Time-course experimentsshowed that increasing levels of CEVd-derived sRNAs did notcorrelate with a decrease in the viroid titer.

The majority of the CEVd sRNAs are derived from viroidpositive-strand RNA

To study the origin and chemical characteristics of CEVdsRNAs, we prepared a small RNA library from systemic leavesof CEVd-infected tomato plants. The samples were collected at4 weeks post-inoculation, when the symptoms displayed by theplants were visible, coinciding with high levels of CEVdsRNAs. We screened a total of 7850 clones containing 529fragments derived from the CEVd genome, representing 1.3%of the total sequences of the library (Table 1S). In addition, we

Fig. 1. Accumulation of CEVd RNA species during tomato infection. TotalRNA was extracted from systemic leaves of a non-infected (lane 0) plant anddifferent independently infected plants (lanes 1 to 15) at 1, 2, 3, 4 and 5 w(weeks) after inoculation. Each lane contained 10 μg RNA. rRNAs were used asloading controls. (A) Time-course of accumulation of genomic CEVd RNAduring infection. Blots were hybridized to labeled CEVd (−) transcripts. C- andL-CEVd indicate the circular and linear forms of CEVd, respectively. (B) Time-course of CEVd-related small RNAs (sRNAs) accumulation. Blots werehybridized to labeled full-length CEVd cDNA.

Fig. 2. Changes in (+) and (−) CEVd-derived sRNA levels during CEVdinfection. RNA samples were extracted from CEVd-infected tomato plants at 3,4 and 5 weeks (w) after inoculation. Each lane contained 10 μg of total RNA.Membranes were hybridized with 32P-labeled (−) (upper panel) and (+) (middlepanel) CEVd transcripts to specifically detect CEVd small RNAs of (+) and (−)polarity, respectively. Lanes 1–10 here correspond to samples used in lanes 6–15 of Fig. 1, respectively. rRNAs were used as loading controls (lower panel).

137R. Martín et al. / Virology 367 (2007) 135–146

obtained 19 tomato sequences which were homologs ofannotated Arabidopsis microRNAs (Table 2S). Analysis ofCEVd-specific fragments indicated that 80.4% were derivedfrom the RNA strand of positive polarity, (+) RNA, whereasonly 19.6% were derived from the complementary strand ofnegative polarity, (−) RNA (Table 1S). These data were con-firmed by the specific detection of CEVd sRNAs of (+) and (−)polarities in infected tomato leaves (Fig. 2). At all stages tested,the amount of (+) sRNAs was significantly higher than that of(−) sRNAs. The preferential enrichment of (+) sRNAs suggest-ed that the predominant in vivo substrate of DCL enzymaticcleavage is likely to be a structured CEVd genomic RNA, ratherthan (±) dsRNA intermediates generated during viroid replica-tion or by the action of cellular RNA-dependent RNApolymerases (RDRPs). This result is similar to that recentlyreported for Potato spindle tuber viroid (PSTVd) and someRNA viruses, in which highly structured single-stranded (ss)viroid or viral RNAs were suggested as substrates for DCLcleavage (Molnar et al., 2005; Itaya et al., 2007).

CEVd sRNAs are 21–22 nt in size, 5′ phosphorylated and 3′methylated

A size distribution analysis of the CEVd sRNAs in thelibrary confirmed an average size of predominantly 21 to 22 nt(Fig. 3A). Next, we used enzymatic and chemical methods toprobe the terminal structure of the CEVd (+) sRNAs.

The 5′-nucleotide of small RNAs generated by DICERactivities is phosphorylated (Elbashir et al., 2001). We foundthat alkaline phosphatase treatment caused CEVd sRNAs tomigrate slower than untreated ones in polyacrylamide gel elec-

trophoresis, consistent with the loss of a 5′-terminal phosphategroup (Fig. 3B). These results indicate that CEVd sRNAs arephosphorylated at the 5′-end.

The terminal 3′-nucleotide of Arabidopsis miRNAs andother types of endogenous sRNAs is methylated, resulting inresistance to β-elimination (Li et al., 2005). To test whetherCEVd sRNAs have a similar modification at their 3′-ends, totalRNA samples from CEVd-infected tomato plants were treatedwith sodium periodate (NaIO4) and subjected to β-elimination.Only sRNAs containing both 2′ and 3′ hydroxyl free groups canreact with NaIO4. sRNAs reacted with NaIO4 and β-eliminatedmigrate faster than the unreacted RNA in polyacrylamide gels.With this treatment, we did not observe any change in themigration of the CEVd sRNAs, whereas a 21-nt synthetic RNAwith free 2′-OH and 3′-OH (internal control) migrated fasterthan the non-treated control (Fig. 3C).

CEVd RNA contains preferred regions for DCL cleavage

Based on the sequences of the CEVd sRNAs recovered in thelibrary, we estimated the frequency of cleavage along the CEVdRNA genome (Fig. 4). The profile suggested the likelyexistence of preferred regions for DCL enzymatic cleavage.The CEVd (+) sRNAs mapped to regions mainly located at theright half side of the predicted CEVd RNA structure. To rule outthe possibility that these differences were due to cloningartifacts, we used Northern blot hybridizations to analyze therelative abundance of the CEVd (+) sRNAs derived fromdifferent regions of the viroid genome (Fig. 5A). These assaysconfirmed that sequences of all the CEVd regions wererepresented in the population of CEVd (+) sRNAs, with theright half side being the most abundant. In addition, Northernblot experiments confirmed that the amount of CEVd (+)sRNAs derived from a particular viroid sequence was moreabundant that of the corresponding (−) polarity. Interestingly,among the population of CEVd (−) sRNAs, levels of 21 ntsRNAs were similar to those of the 24 nt sRNAs (Fig. 5B).

Fig. 3. Biochemical characterization of CEVd sRNAs. (A) Left and right panels show the size distribution of the sequenced (+) and (−) CEVd sRNAs, respectively.Numbers below the columns indicate the sRNA sizes and numbers above the columns indicate the number of sRNAs of different sizes. (B and C) CEVd-derivedsRNAs are phosphorylated at the 5′-end and methylated at the 3′-end. (B) RNAs prepared from CEVd-infected tomato plants were treated (+) or not (−) withalkaline phosphatase. (C) For β-elimination assay the samples were treated (+) or not (−) with sodium periodate and β-eliminated. For both panels B and C, eachlane contained 20 μg RNA. The blot was hybridized to a 32P-labeled oligonucleotide complementary to the CEVd RNA sequence contained between nucleotides159 and 180. Similar results were obtained with other labeled oligoprobes complementary to different regions of the CEVd RNA. rRNAs were used as loadingcontrols.

138 R. Martín et al. / Virology 367 (2007) 135–146

Furthermore, the majority of the CEVd (+) sRNAs mappedto the upper strand of the right half side. This asymmetry wasalso found at later stages of infection and conserved in otherindependent infections analyzed (Fig. 5C). Taken together ourresults suggested two possibilities: (1) the right half side ofCEVd might be folded into a secondary structure different fromthe predicted structure of the viroid and this structure might berecognized by DCLs to generate sRNAs, (2) the same structuremight be recognized as a pre-microRNA precursor which wasthen processed to yield sRNA duplexes with differentialstability, e.g., the pairs formed by a mature miRNA and at itscomplementary strand (denominated star strand and indicated asmiRNA*), with the miRNA* being much less stable. To addressthis issue, we tested the in vivo cleavage profile of the CEVdisolated domain formed by the sequence included betweenpositions 135 and 232 of its genomic RNA (CEVd 135–232).After in vivo transient expression, we found that the cleavageprofile of this partial CEVd 135–232 domain was similar tothat obtained for the same domain when present as part ofthe genomic CEVd RNA; CEVd (+) sRNAs were mapped inan asymmetric manner to different regions of the right sideof the viroid sequence (Fig. 5D). There was only onepredicted secondary structure for this CEVd 135–232 domain

based on the minimal free energy folding, and this structurewas similar to that predicted for the CEVd right-end domain.The pattern of cleavage observed, similar to that found inplants infected with CEVd, suggested that this structureddomain of CEVd might be a putative substrate of DCLsin vivo.

CEVd sRNAs do not alter the endogenous miRNA pathway

As a counter defense against host responses, many viralgenomes encode suppressors of gene silencing (Li and Ding,2006); in addition, viral RNAs can function directly as silencingsuppressors (Lu and Cullen, 2004; Andersson et al., 2005).Some viral suppressors, like the potyviral P1/HC-Pro, havebeen shown to cause an accumulation of miRNAs in Arabi-dopsis (Kasschau et al., 2003). microRNAs (miRNAs) areknown to control the expression of genes involved in severaldevelopmental processes (Jones-Rhoades et al., 2006). In thecase of leaf patterning and polarity, miRNAs165/166 regulatethe expression of the class III-HD ZIP (HD-ZIP III) transcrip-tion factors (Rhoades et al., 2002; Kidner and Martienssen,2004) and miR159/319 regulate MYB and TCP (Teosintebranched1, Cycloidea and PCF) transcripts (Palatnik et al.,

Fig. 4. Origin of CEVd sRNAs produced during viroid infection. (A) Frequency of 5′ cleavage at each nucleotide along the CEVd genomic RNA. Positive valuescorrespond to cleavages at the CEVd (+) RNA and negative values correspond to cleavages at the complementary CEVd (−) RNA. The length of the lines indicates thenumber of times each residue has been cleaved by DCLs. The position of each nucleotide in the CEVd genome is indicated. (B) Positions of (+) CEVd sRNAs thathave been recovered five times or more from the library are represented as black boxes alongside the CEVd genome. The thickness of the box is proportional to itsabundance in the library. The sequence contained between nucleotide positions 8 and 29 has been recovered 5 times.

139R. Martín et al. / Virology 367 (2007) 135–146

2003; Millar and Gubler, 2005). Because plants infected withCEVd and other viroids showed strong alteration in leafmorphology, growth and development, it was possible thatviroid infection might alter miRNA biogenesis and/or function.To test this hypothesis, we analyzed the miRNA pathway duringCEVd infections by examining the expression of coordinatesmiR165/166 and miR159/319 (Rhoades et al., 2002; Palatnik etal., 2003) in foliar tissues, at different stages after viroidinoculation, starting from when symptoms become visible. Inaddition, we assayed the expression levels of miR162 and 168,which regulate DCL1 and AGO1 transcripts, respectively; thesetwo transcripts encode two major components of the miRNAbiogenesis and function (Xie et al., 2003; Vaucheret et al., 2004,2006). However, we found no significant changes in theexpression level of all these miRNAs tested at 5 weeks ofinfection (Fig. 6A). Same results were obtained 3 and 4 weeksafter inoculation (data not shown).

To rule out the possibility that the differences were maskedbecause of saturated level of endogenous miRNAs in adultplants, we investigated the expression of transcripts targeted bythese miRNAs. We examined the transcript levels of DCL1 andAGO1, which are targets of miRNAs 162 and 168, respectively,in Arabidopsis thaliana (Xie et al., 2003; Vaucheret et al., 2004,2006). Fig. 6B shows that neither the transcript level of DCL1

nor that of AGO1 was significantly affected by CEVd infection.Similarly, we examined the expression of genes HD-ZIP III,which are targets of miR165/166 (Rhoades et al., 2002), andthat of TCP4, targeted by miR319 in A. thaliana (Palatnik et al.,2003). Fig. 6B shows the transcript level of HD-ZIP III familymembers seemed to be slightly decreased by CEVd infection.However, these differences were detected only for the memberssharing high sequence identity of a specific domain. On theother hand, TCP4 gene expression was not significantly alteredby CEVd. Moreover, we analyzed Asymmetric leaves 1 (AS1)transcript levels, a gene which is also involved in leaf patterningbut not known to be regulated by miRNAs in A. thaliana (Xu etal., 2003). Fig. 6B shows that the AS1 transcript was alsoslightly affected by CEVd infection.

In addition, we analyzed the expression of the tomato Ex-pansin 2 (Exp2) gene, which is down regulated late in PSTVdinfection (Qi and Ding, 2003). Our analysis confirmed that theExp2 transcript level was also decreased by CEVd infectionsalmost 3 times (Fig. 6B).

In some cases, the alteration of miRNA pathway by viralsuppressors results in the accumulation of the star strand ofmiRNAs (miRNA*) (Kasschau et al., 2003), which is veryunstable and almost undetectable under normal conditions. Wefound no detectable levels of the miRNA165* after CEVd

Fig. 5. (A) Relative abundance of CEVd (+) sRNAs in CEVd-infected tomato plants. Blots were hybridized to 20–21-mers probes (P1 to P18) complementary to theCEVd genome. Membranes were exposed for 14 h, except those labeled with asterisk (*) were exposed for 2 days. The location of each oligonucleotide probe (blacklines, P1–P18) is indicated in a schematic representation of the predicted CEVd RNA folding structure (lower panel). Each lane contained 15 μg RNA used to preparethe library. (B) Detection of (+) and (−) CEVd sRNA levels corresponding to the CEVd RNA sequences included between nucleotide positions 160 and 180. Blotswere hybridized to 20–21-mer probes P9 and anti-P9. Right panel: 6-h exposure; left panel: 2-day exposure. Each lane contained 10 μg of total RNA from non-infected (lanes 1 and 2) and CEVd-infected tomato plants (lanes 3 to 9). rRNAs were used as loading controls (lower panels). (C) RNA samples were collected 5weeks post-inoculation. Lanes 1 to 3 correspond to samples of non-infected plants and lanes 4 to 10, samples of different CEVd-infected plants. Each lane contained10 μg RNA and rRNAs were used as loading controls. (D) RNA samples (lanes 1 to 4) from independent N. benthamiana leaves infiltrated with Agrobacteriumcarrying the plasmid pBA:CEVd135–232. Each lane contained 10 μg RNA and rRNAs were used as loading controls. Note that there was no expression of theinfiltrated plasmid in sample 2.

140 R. Martín et al. / Virology 367 (2007) 135–146

infection (Fig. 6C), whereas miRNA165* accumulated to highlevels in transgenic Arabidopsis plants constitutively expressingthe potyviral P1/HC-Pro.

In summary, we could not detect significant differences inthe accumulation levels of all tested miRNAs, between CEVd-

infected and non-infected tomato leaves. Similarly, we did notobserve significant differences in the transcript levels of themajority of the tested miRNA targets. Only the HD-ZIP IIItranscript levels showed a slight decrease and a similar decreasein expression level was also found in the miRNA-independent

Fig. 6. Expression levels of miRNAs and their target transcripts in CEVd-infected tomato plants. (A) Accumulation levels of miRNA162, 168, 165/166,159 and 319/JAW in non-infected (lanes 1 to 3) and infected tomato plants (lanes4 to 8), at 5 weeks after infection. 5S rRNAwas used as a loading control (lowerpanel). (B) Expression levels of Dicer-like 1 (DCL1), Argonaute 1 (AGO1),Expansin 2 (Exp2), genes encoding transcription factors of the class III HD-ZIP (HD-ZIP III), TCP4 and Asymmetric leaves 1 (AS1) genes determined byreal-time RT-PCR analysis in CEVd-infected tomato plants. The analysis wascarried out 5 weeks after CEVd inoculation. Values were first normalized to theActin expression level and then made relative to the mRNA amount in thecontrol (control refers to non-infected plants). Data corresponded to 7independent experiments. Standard errors are indicated. Numbers above thebars correspond to P-values. Asterisks indicate P-values b0.05. Transcriptlevels of AGO1 and HD-ZIP III were determined by two combinations ofprimers (indicated as 1 and 2) amplifying different domains of the genes. (C)Accumulation of miRNA165* 5 weeks after CEVd infection. Col-0:P1/HC-Protransgenic plants were used as a control for miR165* detection. Each lanecontained 10 μg of total RNA from non-infected (lanes 1 to 3) and CEVd-infected tomato plants (lanes 4 to 6). rRNAs were used as loading controls(lower panel).

141R. Martín et al. / Virology 367 (2007) 135–146

AS1mRNA. In addition, there was no accumulation of miR165*in CEVd-infected tomato plants.

These results suggested that CEVd did not significantlyaffect endogenous microRNA pathway in tomato plants.

Discussion

In response to invading nucleic acids, plants use the RNAsilencing machinery to limit multiplication of the pathogenicoffenders. The characterization of pathogen-related sRNAs isessential to uncovering the strategies used by the pathogens tocounter attack these host responses (Akbergenov et al., 2006).Here, we have described the origin and chemical structure ofCEVd-derived sRNAs in vivo and examined some of theputative roles of these sRNAs during pathogenesis.

CEVd sRNAs are mostly derived from structured CEVd (+)RNAs

The origin of pathogen-derived sRNAs is a current subjectof considerable interest. It was generally assumed that (±)dsRNA intermediates, generated during viral replication or bythe action of RDRPs, were the substrates for DCL in vivo(Dalmay et al., 2000; Mourrain et al., 2000). Like someviruses, viroids are RNA pathogens that also replicate via (±)dsRNA intermediates and it was reasonable to speculate thatthese structures might be the origin of viroid-derived sRNAs(Denti et al., 2004). However, this general assumption waschallenged by the experiments of Markarian et al. (2004) whoreported that two CEVd mutants exhibiting severe and mildsymptoms in G. aurantiaca reached similar replication rates,but only the severe strain induced distinctive accumulation ofsRNAs.

The sequence profile obtained in the sRNA library fromCEVd-infected tomato plants showed that the majority of CEVdsRNAs were of the (+) polarity. This observation is consistentwith the notion that genomic CEVd RNA folds into a highlystructured conformation, which is then cleaved by DCL in vivo.This model is similar to that recently suggested for Potatospindle tuber viroid (PSTVd) (Itaya et al., 2007).

Although our Northern blot hybridizations (see Fig. 2)confirmed that CEVd (+) sRNAs were more abundant thanthose of (−) polarity at all stages after infection, it is possiblethat the cloning method used was biased towards a preferentialrecovery of (+) sRNAs. This potential technical issue has beenreported by Pak and Fire (2007) and Sijen et al. (2007) after thecompletion of our work.

The finding that structured viroid RNAs may be substratesfor DCL cleavage is also similar to that described for severalRNA viruses, in which structured virus RNAs are proposed tobe DCL substrates (Molnar et al., 2005). However, this scenariodoes not seem to be generally applicable to all RNA pathogens,because TuMV produces similar levels of sRNAs of bothpolarities in Brassica plants (Ho et al., 2006).

Structured viroid RNAs did not seem to be the exclusivesubstrates for DCL, because a low percentage of (−) polaritysRNAs (less than 20%) were also produced in vivo. These (−)sRNAs might be derived from a substrate different from thatwhich generates (+) sRNAs. We considered it likely that a dif-ferent DCL activity might act on the (±) dsRNA intermediateswhich accumulated to low levels during viroid replication or viathe activity of RDRP to produce these (−) sRNAs. In addition to

142 R. Martín et al. / Virology 367 (2007) 135–146

the (−) sRNAs, the cleavage of (±) dsRNA intermediates byDCL might also generate similar levels of (+) sRNAs. Whereasthe majority of CEVd (+) sRNAs, mostly derived from CEVdgenomic RNA, were 21–22 in size, those sRNAs derived from(±) dsRNA intermediates seemed to have a range of sizes (21–24 nt). This observation is consistent with the hypothesis thatseveral DCL activities may act on different viroid substrates.The sizes and relative accumulation levels of viroid sRNAs ofboth polarities might depend on the specific DCL activity andon the viroid substrate.

Future experiments will be directed at identifying specificDCL activities involved in the processing of viroid substratesand the production of viroid sRNAs.

CEVd sRNAs are likely products of DCL activities

The detection of short RNAs derived from viroid genomes,previously reported by other groups (Papaefthimiou et al., 2001;Itaya et al., 2001; Martínez de Alba et al., 2002; Denti et al.,2004; Markarian et al., 2004; Landry and Perreault, 2005),provides indirect evidence that viroids are putative substrates ofDCL enzymatic cleavage. Our results on the chemical structureof the CEVd (+) sRNAs provide additional indirect evidencethat viroids may be targets of this specific type of RNase IIIenzymes. We showed that the CEVd (+) sRNAs have aphosphate group at the 5′-end, indicating that they are productsof DCL activities. These CEVd sRNAs are modified at the 3′-end with a methyl residue, which renders them resistant to β-elimination. Plant endogenous sRNAs are also methylated at the3′-end under normal conditions (Li et al., 2005), but this mod-ification is not found in all sRNAs derived from plant viruses.sRNAs derived from several families of DNA viruses are alsomethylated at the 3′-end. By contrast, sRNAs derived fromRNA ORMVand some endogenous sRNAs of ORMV-infectedplants also lack this 3′modification because this virus somehowinterferes with the cellular methylation machinery (Akbergenovet al., 2006).

The identification of the DCL activities responsible for theviroid cleavage will provide direct evidence that viroids aretargets of these proteins. Unfortunately, Arabidopsis is not ahost for any known viroid species and although this model plantcontains the replication machinery for many viroids (Daròs andFlores, 2004), so far sRNAs of viroid origin have not beendetected in Arabidopsis transgenic plants expressing viroidsequences (Daròs, unpublished data). Tomato is also a classicaland good model for plant genetics. Although the tomato ge-nome sequencing project is progressing so far no sequenceinformation is yet available on tomato DCLs. This lack of rele-vant sequence information precludes the identification ofspecific DCLs responsible for viroid cleavage. We proposethat several DCL proteins might be involved in the in vivocleavage of the different viroid substrates based on the numberof sRNA sizes found in the CEVd (+) and (−) sRNAs. TheArabidopsis genome encodes several DCLs responsible of theproduction of sRNAs of a number of sizes and all ArabidopsisDCLs (DCL1–4) are involved in virus-related sRNAs produc-tion (Blevins et al., 2006). We believe that the ongoing tomato

genome sequencing project will eventually help to identifyspecific DCLs for viroid cleavage.

Identification of CEVd RNA domains targeted by DCLcleavage

We found a random distribution of CEVd (−) sRNAs alongthe CEVd complementary sequence. By contrast, most of the(+) sRNAs mapped to certain domains of the viroid RNA,suggesting the existence of preferred regions for DCL cleavagein the CEVd genomic RNA. The predicted CEVd RNA secon-dary structure resembles that of a miRNA precursor with stemsof complementary regions separated by bulges and two terminalloops. According to this predicted secondary structure, our stu-dies showed an asymmetric processing of the viroid RNA,suggesting that the right end domain might be recognized andprocessed by DCL as a pre-miRNA, producing sRNAs withasymmetric level of stability, similar to the endogenousmiRNA/miRNA* pairs. Studies with CEVd sequence variantscontaining mutations in this region will help to define thestructural requirements for DCL cleavage.

CEVd sRNAs do not modulate miRNA pathway

To counteract host defense responses, many viruses encodesuppressors of gene silencing, some of which can modulatemiRNAs accumulation (Kasschau et al., 2003; Li and Ding,2006); moreover, non-coding RNAs can also function as silen-cing suppressors (Lu and Cullen, 2004; Andersson et al., 2005).Our analyses showed that CEVd is unlikely to affect thebiogenesis or function of endogenous miRNAs, although thesymptoms displayed by CEVd-infected tomato plants hadraised this possibility. The transcript level of DCL1 or AGO1,the main regulators of miRNA pathway and targets of miR162and 168, respectively (Xie et al., 2003; Vaucheret et al., 2004,2006), did not seem to be significantly affected by CEVdinfections. On the other hand, we observed that the transcriptlevel of genes encoding some transcription factors involved inleaf patterning and development of the shoot apical meristemmight be affected by CEVd infection. This effect could not beattributed to a general alteration of the microRNA pathway byviroids, because the expression of TCP4, targeted bymiRNA319 (Palatnik et al., 2003), was not significantly alteredin CEVd-infected tomato plants. Moreover, we found that theexpression of AS1 (Xu et al., 2003), so far not reported to be atarget of any miRNA, was also slightly decreased by CEVdinfection. It has been demonstrated that lateral organ patterningrequires an interaction between the organ and the shoot apicalmeristem (Reinhardt et al., 2005) and transcription factors of theclass III HD-ZIP family potentially act as a pivotal componentfor leaf patterning–shoot meristem interactions (Byrne, 2006).Several viroid species, including the CEVd used in our studiesand other PSTVd variants, dramatically block development andgrowth of the infected tomato plants and most probably affectthe normal functioning of the apical meristem. It is reasonable toconsider that genes involved in the proper development of themeristem and lateral organs might be directly or indirectly

143R. Martín et al. / Virology 367 (2007) 135–146

affected by CEVd infection. It would be of great interest toinvestigate the expression of these genes in order to elucidatethe basis of viroid infections.

So far, it has not been possible to demonstrate that viroids orviroid sRNAs directly interfere with any cellular process. Ouranalyses show that CEVd does not seem to affect the biogenesisneither the function of endogenous miRNAs, although thesymptoms displayed by CEVd-infected tomato plants hadraised this possibility. Moreover, recent efforts to assign poten-tial silencing suppressor activities to PSTVd have also producednegative results (Itaya et al., 2007).

sRNAs derived from pathogens have been classically relatedto host defense responses; however, the discovery that someanimal viruses encode siRNAs and miRNAs that may silencehost genes challenges this assumption (Cullen, 2006). We ob-served that the accumulation of CEVd sRNAs does not decreasegenomic CEVd RNA levels nor reduce the progression of theinfection. Indeed, there is a positive correlation between in-creasing CEVd sRNA levels and symptom severity, suggestingthat viroid sRNAs could play an important role for viroiddisease progression. This observation has also been reported forother viroids like PSTVd (Papaefthimiou et al., 2001; Itaya etal., 2001; Wang et al., 2004) or Avocado sunblotch viroid(ASBVd) (Markarian et al., 2004), although it is not general forall the viroid species. Particularly, Wang et al. (2004) showedthat the production of viroid sRNAs derived from a non-replicating PSTVd hairpin RNA correlated with the appearanceof symptoms associated with PSTVd infections. Moreover,studies on CEVd pathogenesis showed that a mutation pro-ducing conformational variations in the viroid structure did notaffect viroid replication in Gynura, but the plants accumulatedless viroid sRNAs and displayed milder symptoms (Markarianet al., 2004). Together, these results suggest that viroid-derivedsRNAs might act to mediate down regulation of importantcellular transcripts crucial for normal plant development.

Our structural studies concerning the specific and conservedprocessing of CEVd RNA by DCL in vivo provide new inform-ation regarding the putative viroid genomic RNA regions thatmay encode any of these putative sRNAs. The right-end domainof the predicted structure of CEVd might be a good candidate toencode a putative sRNA. The involvement of the right-enddomain in viroid pathogenesis is consistent with the existence ofviroid sequences outside of the pathogenicity domain that havebeen also reported to affect symptom expression and host range(Loss et al., 1991; Gozmanova et al., 2003; Qi and Ding, 2003).The sequencing of the tomato genome and other viroid planthosts will help to illuminate this issue and advance our know-ledge of viroid pathogenesis.

Materials and methods

Plant materials and viroid sequence variant

Solanum lycopersicum var. Rutgers and Nicotiana benthami-ana plants were grown in soil at 16-h day/8-h night (28 °C).CEVd sequence variant M34917 with a deletion of one Gbetween positions 70 and 74 was used.

Plasmids

pBdCEVd (Daròs and Flores, 2004) was used as templatefor in vitro transcription to produce CEVd inoculum andriboprobes specific for (+) and (−) CEVd. Plasmid pBA:CEVd-135–232 was prepared by cloning the PCR productbetween positions 135 and 232 of the CEVd sequence intopENTR/D-TOPO (Invitrogen) following the manufacturer'sprotocol and finally transferred to pBA-DC gateway binaryvector (Invitrogen) using the Gateway LR clonase (Invitrogen)according to the manufacturer's instructions. The primers usedfor this cloning were 5′ CACCTGCTTCGGTCGCCGCG 3′and 5′ GCGCTTCAGCGACGATCGGATG 3′.

In vitro transcription of viroid RNA and plant inoculation

pBdCEVd was linearized by XhoI digestion and used astemplate for in vitro transcription with MaxiScript T7 RNApolymerase (Ambion) following the manufacturer's protocol.The transcripts were used to mechanically inoculate first trueleaves of 10-day-old tomato plants (100 ng/plant).

Probes preparation

Riboprobes were prepared by in vitro transcription using T3or T7 RNA polymerase (Promega) in the presence of [α−32P]UTP. XhoI- and HindIII-digested pBdCEVds were used astemplates to generate (+) and (−) CEVd transcripts, respec-tively. Unincorporated nucleotides were removed by filtrationthrough Mini Quick Spin Oligo Columns (Roche) according tothe manufacturer's protocol. Strand-specific riboprobes werequantified and the amount of cpm for both probes normalized.CEVd cDNA labeled by the Rediprime II Random PrimeLabeling System (Amersham) was purified with MicroSpinTMS-200 HR columns (Amersham) according to the manufac-turer's protocol. DNA oligonucleotides were 5′-end-labeledwith [γ−32P]ATP by T4 polynucleotide kinase (New EnglandBiolabs) and purified through MicroSpin™ G-25 columns(Amersham).

Cloning of small RNA (sRNA) and identification of coloniescontaining sequences derived from CEVd

Cloning of sRNAs was performed according to Elbashir etal. (2001) with minor modifications. Total RNA was extractedfrom mixed systemic foliar tissues of ten CEVd-infectedtomato plants at 4 weeks after inoculation and separated ondenaturing gels. sRNAs were purified, dephosphorylated andligated to 3′-terminal adapter oligonucleotides. The ligationproducts were then gel purified, 5′ phosphorylated and ligatedto a 5′ adapter. After gel purification, the final ligationproducts were reverse transcribed, amplified by PCR andfinally cloned as concatemers into a commercial TOPO-TAcloning vector (Invitrogen). Individual colonies were screenedfor CEVd-derived inserts by colony hybridization with a full-length CEVd cDNA probe. DNAs from positive clones weresequenced.

144 R. Martín et al. / Virology 367 (2007) 135–146

RNA extraction and analysis

Foliar tissues corresponding to systemic leaves (third leafand younger leaves) of an individual plant were mixed andground on liquid nitrogen preceding RNA extraction. TotalRNA of tomato plants was isolated using Trizol reagent (Invi-trogen) according to the manufacturer's protocol. For detectionof CEVd (+) and (−) RNA, 10 μg of total RNAwas separated on5% PAGE with 8 M urea and (0.5×) TBE. After electrophoresisthe RNAwas electroblotted to a Hybond N+ membrane (Amer-sham) and UV cross-linked. Blot hybridizations were per-formed at 65 °C overnight in UltraHyb-oligo buffer (Ambion)with riboprobes specific for (+) and (−) CEVd (see above).Membranes were washed two times with (2×) SSC, 0.1% SDSfor 30 min and 15 min with (0.1×) SSC, 0.1% SDS at 68 °C. Fordetection of small RNAs, an aliquot of 15 μg RNAwas loadedon 15% PAGE with 8 M urea and (0.5×) TBE. After electro-phoresis, the RNA was transferred to a Hybond N+ membrane(Amersham) by electroblotting and UV cross-linked. Mem-branes were hybridized in ULTRAhyb-oligo Buffer (Ambion)at 42 °C overnight with oligoprobes or cDNA probes andwashed twice in (2×) SSC, 0.1% SDS for 30 min. For detectionof sRNAs of specific polarity, the hybridizations wereperformed at 50 °C and washed with (2×) SSC, 0.1% SDS at50 °C for 30 min, at 55 °C for 30 min and at 59 °C for 45 min.

Phosphatase treatment and β-elimination

For dephosphorylation, 20 μg of total RNAwas treated with20 U of calf intestine alkaline phosphatase (Invitrogen) in 10 μl(1×) buffer for 30 min at 37 °C. For NaIO4 reaction and β-elimination, 20 μg total RNA was dissolved in 17.5 μl boraxbuffer, pH 8.6 (4.375 mM borax, 50 mM boric acid) and 2.5 μl200 mM NaIO4 was added. The mixture was incubated for10 min at room temperature in the dark. 2 μl Glycerol wasadded to quench the unreacted NaIO4 and the mixture incubatedfor an additional 10 min at room temperature. After dryingunder vacuum for 30 min at room temperature, samples weredissolved in 50 μl borax buffer pH 9.5 (33.75 mM borax,50 mM boric acid) and incubated for 90 min at 45 °C. The RNAwas then purified and dissolved in the denaturing gel loadingbuffer.

Quantitative RT-PCR analysis

2 μg Total RNA was used to reverse transcribe target se-quences using oligo(dT) primer and SuperScript III reversetranscriptase (Invitrogen) according to the manufacturer's pro-tocol. PCR was carried out in the presence of the double-stranded DNA-specific dye SYBR green PCR Master Mix(Applied Biosystems). Amplification was monitored in real timewith the 7900 HT sequence detection system (AppliedBiosystems). The PCR primers used for RT-PCRs were basedon the tomato est database. Actin1-Forward: 5′ CCA CGT CACACGGGTGTTATG 3′ and Actin1-Reverse: 5′GATTCGCCTTAG GGT TTA GAG G 3′, based on the pseudogene partialsequence accession U60479. DCL1-Forward: 5′ CTG TAT

CGATGT GTG CAC GAG G 3′ and DCL1-Reverse: 5′ GAGTTC CAATAG AAG AGC TGC TG 3′, based on tomato cloneBI209312. AGO1-1Forward: 5′ CGA ATT CCC CTT GTCAGC GAC C 3′ and AGO1-1Reverse: 5′ CTG ATA GTT GGGTTC TAA AGA TGC 3′, based on estLA26CC11. AGO1-2Forward: 5′ GCT CTG AGT CCC ACG AAA CTG G′ andAGO1-2Reverse: 5′ CCC CTG CCC TGT TGATAG TAC 3′,based on estLC14BA04. HDZIP-1Forward: 5′ GTG AAT CTGTGG TCG TGA GTG G 3′ and HDZIP-1Reverse: 5′ CTC TAGACT CAC GAG GCC GC 3′. HDZIP-2Forward: 5′ AAA GTATGT GAG GTA TAC ACC AG 3′ and HDZIP-2Reverse: 5′ACA CTT GCT TCT GTA ATC GGT C 3′, both based onestBI925551. TCP4-Forward: 5′ GCA CGT AGT TAT CTGTTT AAC TC 3′ and TCP4-Reverse: 5′ GAATAC ATA GGAAAC ACA GCT TG 3′, based on estAW037628. AS1-Forward:5′ TGA AGA GGA TGC TTT GTT GCG AG 3′ and AS1-Reverse: 5′ TTC CAT TTG TTA CCG TGT TTG GC 3′, basedon estBE441125. LeExp2-F: 5′ GAT GGC AAA CTG CTCATG CCA 3′ and LeExp2-R: 5′ TTA GGT AGA GAC GGGTTC GGA G 3′, based on the completed cds AF096776.

Agrobacterium infiltration

Agrobacterium tumefaciens cultures, grown to midlog phase,were resuspended in 10 mM MgCl2 and 150 μM acetosyr-ingone, incubated for 4 h at room temperature and infiltratedinto leaves of N. benthamiana plants. Two days after infil-tration, the inoculated leaves were harvested and subjected toRNA extraction.

Acknowledgments

We are grateful to Dr. Xiujie Wang for the generous andhelpful assistance with the analysis of the library sequences andDr. Xiuren Zhang for the experimental help and valuable sug-gestions. We also thank Drs. Enno Krebbers, Richard Broglie,Barbara Mazur and Bobby Williams for the interesting dis-cussion and critical comments, and Drs. Isabel Abreu andRafael Catalá for critical reading of the manuscript. This workwas partially supported by a grant from DUPONT to N-H.C.and by a postdoctoral fellowship of the Spanish Ministry ofEducation to R.M.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at doi:10.1016/j.virol.2007.05.011.

References

Akbergenov, R., Si-Ammour, A., Blevins, T., Amin, I., Kutter, C., Vanderschu-ren, H., Zhang, P., Gruissem, W., Meins Jr., F., Hohn, T., Pooggin, M.M.,2006. Molecular characterization of geminivirus-derived small RNAs indifferent plant species. Nucleic Acids Res. 34 (2), 462–471.

Anandalakshmi, R., Pruss, G.J., Ge, X., Marathe, R., Mallory, A.C., 1998. Aviral suppressor of gene silencing in plants. Proc. Natl. Acad. Sci. U.S.A. 95,13079–13084.

Andersson, M.G., Haasnoot, P.C., Xu, N., Berenjian, S., Berkhout, B.,

145R. Martín et al. / Virology 367 (2007) 135–146

Akusjarvi, G., 2005. Suppression of RNA interference by adenovirus virus-associated RNA. J. Virol. 79, 9556–9565.

Baulcombe, D., 2004. RNAi silencing in plants. Nature 431, 356–363.Blevins, T., Rajeswaran, R., Shivaprasad, P.V., Beknazariants, D., Si-Ammour,

A., Park, H.S., Vazquez, F., Robertson, D., Meins Jr., F., Hohn, T., Pooggin,M.M., 2006. Four plant Dicers mediate viral small RNA biogenesis andDNA virus induced silencing. Nucleic Acids Res. (21), 6233–6246.

Borsani, O., Zhu, J., Verslues, P.E., Sunkar, R., Zhu, J.K., 2005. EndogenoussiRNAs derived from a pair of natural cis-antisense transcripts regulate salttolerance in Arabidopsis. Cell 123, 1279–1291.

Branch, A.D., Robertson, H.D., 1984. A replication cycle for viroids and othersmall infectious RNAs. Science 223, 450–455.

Byrne, M.E., 2006. Shoot meristem function and leaf polarity: the role of classIII HD-ZIP genes. PLoS Genet. 2 (6), 785–790.

Cullen, B.R., 2006. Viruses and microRNAs. Nat. Genet., Suppl. 38, S25–S30.Dalmay, T., Hamilton, A., Rudd, S., Angell, S., Baulcombe, D.C., 2000. An

RNA-dependent RNA polymerase gene in Arabidopsis is required forposttranscriptional gene silencing mediated by a transgene but not by a virus.Cell 101 (5), 543–553.

Daròs, J.A., Flores, R., 2004. Arabidopsis thaliana has the enzymatic machineryfor replicating representative viroid species of the family Pospiviroidae.Proc. Natl. Acad. Sci. U.S.A. 101, 6792–6797.

Daròs, J.A., Marcos, J.F., Hernandez, C., Flores, R., 1994. Replication of avo-cado sunblotch viroid: evidence for a symmetric pathway with two rollingcircles and hammerhead ribozyme processing. Proc. Natl. Acad. Sci. U.S.A.91, 12813–12817.

Daròs, J.A., Elena, S.F., Flores, R., 2006. Viroids: an Ariadne's thread into theRNA labyrinth. EMBO Rep. 7 (6), 593–598.

Denti, M.A., Boutla, A., Tsagris, M., Tabler, M., 2004. Short interfering RNAsspecific for potato spindle tuber viroid are found in the cytoplasm but not inthe nucleus. Plant J. 37 (5), 762–769.

Diener, T.O., 2003. Discovering viroids—a personal perspective. Nat. Rev.,Microbiol. 1, 75–80.

Ding, B., Itaya, A., 2007. Viroid: a useful model for studying the basic principlesof infection and RNA biology. Mol. Plant-Microb. Interact. 20, 7–20.

Ding, B., Kwon, M.O., Hammond, R., Owens, R., 1997. Cell-to-cell movementof potato spindle tuber viroid. Plant J. 12, 931–936.

Ding, S.W., Li, H., Lu, R., Li, F., Li, W.X., 2004. RNA silencing: a conservedantiviral immunity of plants and animals. Virus Res. 102, 109–115.

Dunoyer, P., Voinnet, O., 2005. The complex interplay between plant virusesand host-silencing pathways. Curr. Opin. Plant Biol. 8, 415–423.

Elbashir, S.M., Lendeckel, W., Tuschl, T., 2001. RNA interference is mediatedby 21- and 22-nucleotide RNAs. Genes Dev. 15, 188–200.

Flores, R., Hernández, C., Martínez de Alba, E., Darós, J.A., Di Serio, F., 2005.Viroids and viroid–host interactions. Annu. Rev. Phytopathol. 43, 117–139.

Gozmanova, M., Denti, M.A., Minkov, I.N., Tsagris, M., Tabler, M., 2003.Characterization of the RNA motif responsible for the specific interaction ofpotato spindle tuber viroid RNA (PSTVd) and the tomato protein Virp1.Nucleic Acids Res. 31 (19), 5534–5543.

Hamilton, A., Voinnet, O., Chappell, L., Baulcombe, D., 2002. Two classes ofshort interfering RNA in RNA silencing. EMBO J. 21, 4671–4679.

Hannon, G.J., Conklin, D.S., 2004. RNA interference by short hairpin RNAsexpressed in expressed in vertebrate cells. Methods Mol. Biol. 257,255–266.

Ho, T., Pallett, D., Rusholme, R., Dalmay, T., Wang, H., 2006. A simplifiedmethod for cloning of short interfering RNAs from Brassica juncea infectedwith Turnip mosaic potyvirus and Turnip crinkle carmovirus. J. Virol.Methods 136 (1–2), 217–223.

Itaya, A., Folimonov, A., Matsuda, Y., Nelson, R.S., Ding, B., 2001. Potatospindle tuber viroid as inducer of RNA silencing in infected tomato. Mol.Plant-Microb. Interact. 14, 1332–1334.

Itaya, A., Zhong, X., Bundschuh, R., Qi, Y., Wang, Y., Takeda, R., Harris, A.R.,Molina, C., Nelson, R.S., Ding, B., 2007. A structured viroid RNA issubstrate for dicer-like cleavage to produce biologically active smallRNAs but is resistant to RISC-mediated degradation. J. Virol. 81 (6),2980–2994.

Jones-Rhoades, M.W., Bartel, D.P., Bartel, B., 2006. MicroRNAs and theirregulatory roles in plants. Annu. Rev. Plant Biol. 57, 19–53.

Kasschau, K.D., Carrington, J.C., 1998. A counterdefensive strategy of plantviruses: suppression of post-transcriptional gene silencing. Cell 95,461–470.

Kasschau, K.D., Xie, Z., Allen, E., Llave, C., Chapman, E.J., Krizan, K.A.,Carrington, J.C., 2003. P1/HC-Pro, a viral suppressor of RNA silencing,interferes with Arabidopsis development and miRNA function. Dev. Cell 4(2), 205–217.

Keese, P., Simons, R.H., 1985. Domains in viroids: evidence of intermolecularRNA rearrangements and their contribution to viroid evolution. Proc. Natl.Acad. Sci. U.S.A. 82, 4582–4586.

Kidner, C.A., Martienssen, R.A., 2004. Spatially restricted microRNA directsleaf polarity through ARGONAUTE 1. Nature 428 (6978), 81–84.

Lagos-Quintana, M., Rauhut, R., Lendeckel, W., Tuschl, T., 2001. Identificationof novel genes coding for small expressed RNAs. Science 294, 853–858.

Lakatos, L., Szittya, G., Silhavy, D., Burgyan, J., 2004. Molecular mechanism ofRNA silencing suppression mediated by p19 protein of tombosviruses.EMBO J. 23 (4), 876–884.

Landry, P., Perreault, J.P., 2005. Identification of a peach latent mosaic viroidhairpin able to act as a Dicer-like substrate. J. Virol. 79 (10), 6540–6543.

Lee, R.C., Ambros, V., 2001. An extensive class of small RNAs in Caenor-habditis elegans. Science 294, 862–864.

Li, F., Ding, S.W., 2006. Virus counterdefense: diverse strategies for evading theRNA-silencing immunity. Annu. Rev. Microbiol. 60, 503–531.

Li, J., Yang, Z., Yu, B., Liu, J., Chen, X., 2005. Methylation protects miRNAsand siRNAs from a 3′-end uridylation activity in Arabidopsis. Curr. Biol.15, 1501–1507.

Llave, C., Kasschau, K.D., Rector, M.A., Carrington, J.C., 2002. Endogenousand silencing-associated small RNAs in plants. Plant Cell 14, 1605–1619.

Loss, P., Schmitz, M., Steger, G., Riesner, D., 1991. Formation of thermo-dynamically metastable structure containing hairpin II is critical forinfectivity of potato spindle tuber viroid RNA. EMBO J. 10, 719–727.

Lu, S., Cullen, B.R., 2004. Adenovirus VA1 noncoding RNA can inhibit smallinterfering RNA and microRNA biogenesis. J. Virol. 78, 12868–12876.

Maniataki, E., Martinez de Alba, A.E., Sagesser, R., Tabler, M., Tsagris, M.,2003. Viroid RNA systemic spread may depend on the interaction of a 71-nucleotide bulged hairpin with the host protein VirP1. RNA 9 (3),346–354.

Markarian, N., Li, H.W., Ding, S.W., Semancik, J.S., 2004. RNA silencing asrelated to viroid induced symptom expression. Arch. Virol. 149 (2),397–406.

Martínez de Alba, A.E., Flores, R., Hernandez, C., 2002. Two chloroplasticviroids induce the accumulation of small RNAs associated with posttran-scriptional gene silencing. J. Virol. 76 (24), 13094–13096.

Martínez de Alba, A.E., Sagesser, R., Tabler, M., Tsagris, M., 2003. Abromodomain-containing protein from tomato specifically binds potatospindle tuber viroid RNA in vitro and in vivo. J. Virol. 77 (17), 9685–9694.

Millar, A.A., Gubler, F., 2005. The Arabidopsis GAMYB-like genes, MYB33and MYB65, are microRNA-regulated genes that redundantly facilitateanther development. Plant Cell 17, 705–721.

Molnar, A., Csorba, T., Lakatos, L., Varallyay, E., Lacomme, C., Burgyan,J., 2005. Plant virus-derived small interfering RNAs originate predomi-nantly from highly structure single-stranded viral RNAs. J. Virol. 79 (12),7812–7818.

Mourrain, P., Beclin, C., Elmayan, T., Feuerbach, F., Gordon, C., Morel, J.B.,Jouette, D., Lacombe, A.M., Nikic, S., Picault, N., Remoue, K., Sanial, M.,Vo, T.A., Vaucheret, H., 2000. Arabidopsis SGS2 and SGS3 genes arerequired for posttranscriptional gene silencing and natural virus resistance.Cell 101 (5), 533–542.

Navarro, J.A., Daròs, J.A., Flores, R., 1999. Complexes containing both polaritystrands of sunblotch viroid: Identification in chloroplasts and characteriza-tion. Virology 253, 77–85.

Pak, J., Fire, A., 2007. Distinct populations of primary and secondary effectorsduring RNAi in C. elegans. Science 315 (5809), 241–244.

Palatnik, J.F., Edwards, A., Wu, X., Schommer, C., Schwab, R., Carrington,J.C., Weigel, D., 2003. Control of leaf morphogenesis by microRNAs.Nature 425, 257–263.

Papaefthimiou, I., Hamilton, A., Denti, M., Baulcombe, D., Tsagris, M., Tabler,M., 2001. Replicating potato spindle tuber viroid RNA is accompanied by

146 R. Martín et al. / Virology 367 (2007) 135–146

short RNA fragments that are characteristic of post-transcriptional genesilencing. Nucleic Acids Res. 29 (11), 2395–2400.

Pffefer, S., Zavolan, M., Grasser, F.A., Chien, M., Russo, J.J., Ju, J., John, B.,Enright, A.J., Marks, D., Sander, C., Tuschl, T., 2004. Identification of virus-encoded microRNAs. Science 304, 734–736.

Pffefer, S., Sewer, A., Lagos-Quintana, M., Sheridan, R., Sander, C., Grasser,F.A., van Dyk, L.F., Ho, C.K., Shuman, S., Chien, M., Russo, J.J., Ju, J.,Randall, G., Lindenbach, B.D., Rice, C.M., Simon, V., Ho, D.D., Zavolan,M., Tuschl, T., 2005. Identification of microRNAs of the herpesvirus family.Nat. Methods 2, 269–276.

Qi, Y., Ding, B., 2003. Inhibition of cell growth and shoot development by aspecific nucleotide sequence in a noncoding viroid RNA. Plant Cell 15 (6),1360–1374.

Qi, Y., Pelissier, T., Itaya, A., Hunt, E., Wassenegger, M., Ding, B., 2004. Directrole of a viroid RNA motif in mediating directional RNA trafficking across aspecific cellular boundary. Plant Cell 16, 1741–1752.

Reinhardt, D., Frenz, M., Mandel, T., Kuhlemeier, C., 2005. Microsurgical andlaser ablation analysis of leaf positioning and dorsoventral patterning intomato. Development 132, 15–26.

Rhoades, M.W., Reinhart, B.J., Lim, L.P., Burge, C.B., Bartel, B., Bartel, D.P.,2002. Prediction of plant microRNA targets. Cell 110, 513–520.

Sholthof, H.B., 2006. The Tombusvirus-encoded P19: from irrelevance toelegance. Nat. Rev., Microbiol. 4 (5), 405–411.

Sijen, T., Steiner, F.A., Thijssen, K.L., Plasterk, R.H., 2007. Secondary siRNAsresult from unprimed RNA synthesis and form a distinct class. Science 315(5809), 244–247.

Silhavy, D., Molnar, A., Lucioli, A., Szittya, G., Hornyik, C., Tavazza, M.,Burgyan, J., 2002. A viral protein suppresses RNA silencing and bindssilencing-generated, 21- to 25-nucleotide double-stranded RNAs. EMBO J.21, 3070–3080.

Sogo, J.M., Koller, T., Diener, T.O., 1973. Potato spindle tuver viroid: X.Visualization and size determination by electron microscopy. Virology 55(1), 70–80.

Sullivan, C.S., Ganem, D., 2005. MicroRNAs and viral infection. Mol. Cell 20,3–7.

Sullivan, C.S., Grundhoff, A.T., Tevethia, S., Pipas, J.M., Ganem, D., 2005.SV40-encoded microRNAs regulate viral gene expression and reducesusceptibility to cytotoxic T cells. Nature 435, 682–686.

Tabler, M., Tsagris, M., 2004. Viroids: petite RNA pathogens with distinguishedtalents. Trends Plant Sci. 9, 339–348.

Vance, V., Vaucheret, H., 2001. RNA silencing in plants—defense and counter-defense. Science 292, 2277–2280.

Vaucheret, H., Vazquez, F., Crete, P., Bartel, D.P., 2004. The action ofARGONAUTE1 in the miRNA pathway and its regulation by themiRNA pathway are crucial for plant development. Genes Dev. 18 (10),1187–1197.

Vaucheret, H., Mallory, A.C., Bartel, D.P., 2006. AGO1 homeostasis entailscoexpression of MIR168 and AGO1 and preferential stabilization ofmiR168 by AGO1. Mol. Cell 22 (1), 129–136.

Voinnet, O., 2001. RNA silencing as a plant immune system against viruses.Trends Genet. 17 (8), 449–459.

Voinnet, O., 2005. Induction and suppression of RNA silencing: insights fromviral infections. Nat. Rev., Genet. 6, 206–220.

Wang, M.B., Bian, X.Y., Wu, L.M., Liu, L.X., Smith, N.A., Isenegger, D., Wu,R.M., Masuta, C., Vance, V.B., Watson, J.M., Rezaian, A., Dennis, E.S.,Waterhouse, P.M., 2004. On the role of RNA silencing in the pathogenicityand evolution of viroids and viral satellites. Proc. Natl. Acad. Sci. U.S.A.101 (9), 3275–3280.

Waterhouse, P.M., Wang, M.B., Lough, T., 2001. Gene silencing as an adaptivedefense against viruses. Nature 11, 834–842.

Xie, Z., Kasschau, K.D., Carrington, J.C., 2003. Negative feedback regulationof Dicer-like1 in Arabidopsis by microRNA-guided mRNA degradation.Curr. Biol. 13 (9), 784–789.

Xu, L., Xu, Y., Dong, A., Sun, Y., Pi, L., Xu, Y., Huang, H., 2003. Novel as1 andas2 defects in leaf adaxial-abaxial polarity reveals the requirement forASYMMETRIC LEAVES1 and 2 and ERECTA functions in specifying leafadaxial identity. Development 130, 4097–4107.

Zamore, P.D., Tuschl, T., Sharp, P.A., Bartel, D.P., 2000. RNAi: double-strandedRNA directs the ATP-dependent cleavage of mRNA at 21 to 23 intervals.Cell 101 (1), 25–33.

Zhang, X., Yuan, Y.R., Pei, Y., Lin, S.S., Tuschl, T., Patel, D.J., Chua, N.H.,2006. Cucumber mosaic virus-encoded 2b suppressor inhibits ArabidopsisArgonaute1 cleavage activity to counter plant defense. Genes Dev. 20 (23),3255–3268.

Zhu, Y., Green, L., Woo, Y.M., Owens, R., Ding, B., 2001. Cellular basis ofpotato spindle tuber viroid systemic movement. Virology 279, 69–77.