Make Illumina part of your DNA Sign up for PNAS Online eTocs
Link: Info for AuthorsLink: Editorial BoardLink: AboutLink: SubscribeLink: AdvertiseLink: ContactLink: Sitemap Link: PNAS Home
Proceedings of the National Academy of Sciences
Link: Current Issue "" Link: Archives "" Link: Online Submission ""  Link: Advanced Search

Published online on December 12, 2006, 10.1073/pnas.0604627103
PNAS | December 19, 2006 | vol. 103 | no. 51 | 19593-19598
OPEN ACCESS ARTICLE


This Article
Free via Open Access: OA
Right arrow OA Abstract
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Supporting Information
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a colleague
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Moissiard, G.
Right arrow Articles by Voinnet, O.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Moissiard, G.
Right arrow Articles by Voinnet, O.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg  
What's this?

 Previous Article  | Table of Contents |  Next Article 

BIOLOGICAL SCIENCES / PLANT BIOLOGY
RNA silencing of host transcripts by cauliflower mosaic virus requires coordinated action of the four Arabidopsis Dicer-like proteins

Guillaume Moissiard, and Olivier Voinnet*

Institut de Biologie Moléculaire des Plantes, Centre National de la Recherche Scientifique, Unité Propre de Recherche 2357, 12 Rue du Général Zimmer, 67084 Strasbourg Cedex, France

Edited by David Baulcombe, The Sainsbury Laboratory, Norwich, United Kingdom, and approved September 20, 2006 (received for review June 3, 2006)


   Abstract
Top
 Abstract
Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 
RNA silencing is an ancient mechanism of gene regulation with important antiviral roles in plants and insects. Although induction of RNA silencing by RNA viruses has been well documented in plants, the interactions between DNA viruses and the host silencing machinery remain poorly understood. We investigate this question with cauliflower mosaic virus (CaMV), a dsDNA virus that expresses its genome through the polycistronic 35S RNA, which carries an unusually extensive secondary structure known as translational leader. We show that CaMV-derived siRNAs accumulate in turnip- and Arabidopsis-infected plants and that the leader is a major, albeit not exclusive, source for those molecules. Biogenesis of leader-derived siRNA requires the coordinated and hierarchical action of the four Arabidopsis Dicer-like (DCL) proteins. Our study also uncovers a "facilitating" role exerted by the microRNA biosynthetic enzyme DCL1 on accumulation of DCL2-, DCL3-, and DCL4-dependent siRNAs derived from the 35S leader. This feature of DCL1 defines a small RNA biosynthetic pathway that might have relevance for endogenous gene regulation. Several leader-derived siRNAs were found to bear near-perfect sequence complementarity to Arabidopsis transcripts, and, using a sensor transgene, we provide direct evidence that at least one of those molecules acts as a bona fide siRNA in infected turnip. Extensive bioinformatics searches identified >100 transcripts potentially targeted by CaMV-derived siRNAs, several of which are effectively down-regulated during infection. The implications of virus-directed silencing of host gene expression are discussed.

In RNA silencing, dsRNA or imperfect stem-loop RNA molecules are processed by homologs of the RNase-III enzyme Dicer into 21- to 24-nt RNAs, referred to as siRNAs and microRNAs (miRNAs) (1). The model plant Arabidopsis thaliana exhibits a broad spectrum of endogenous RNA silencing pathways that involve four specialized Dicer-like (DCL) proteins (2, 3). DCL1 catalyzes cleavage of intergenic/intronic fold-back precursors to release miRNAs. These molecules incorporate into an RNA-induced silencing complex (RISC) to promote cleavage or translational repression of cellular transcripts carrying miRNA-complementary sites (4). DCL3 produces 24-nt siRNAs guiding heterochromatin formation and transcriptional repression of transposon and DNA repeats (5). DCL4 has roles in RNA interference (6) and processes noncoding RNA precursors into 21-nt trans-acting siRNAs that control developmental timing and leaf polarity (7, 8). DCL2 synthesizes a 24-nt natural antisense transcript siRNA that regulates stress responses (9), and, in plants with compromised DCL4 activity, it alternately processes {approx}22-nt siRNAs from trans-acting siRNA precursors (7, 8).

Besides its endogenous functions, RNA silencing also has antiviral roles in plants and insects (10, 11). dsRNA from replication intermediates or viral RNA with extensive fold-back structure are presumed sources of virus-derived small RNA (vsRNA). vsRNAs are thought to guide endonucleolytic cleavage of viral genomes/transcripts, after their incorporation into a RISC. This model for antiviral silencing is strongly supported by the findings that most plant viruses produce suppressor proteins targeting DCL, RISC, or small RNA (sRNA) activities (10, 12). We recently showed that the siRNA products of DCL4 and DCL2 redundantly recruit an antiviral RISC to mediate defense against three distinct RNA viruses (13).

Although induction and suppression of RNA silencing have now been extensively documented with plant RNA viruses, little information is available regarding viruses with DNA genomes. Recent studies show that infections by ssDNA geminiviruses trigger accumulation of different vsRNA classes (14). This likely reflects the action of distinct DCLs that possibly use dsRNA resulting from annealing of converging sense/antisense transcripts as substrate (15). However, the picture is less clear in the case of viruses with dsDNA genomes, such as cauliflower mosaic virus (CaMV), type member of the Caulimovirus genus (supergroup: pararetrovirus). Despite reports of CaMV-induced transgene silencing (16, 17), there has been no account of vsRNA accumulation in infected plants. Additionally, convergent transcription is not part of pararetroviral replication such that the possible origin of vsRNAs is unknown. The CaMV circular dsDNA ({approx}8 kb) is replicated by reverse transcription of an RNA intermediate. RNA polymerase II produces two nuclear viral transcripts, the 35S and 19S RNAs, with the former carrying an extensive fold-back structure at its 5' end (the translational leader) that ensures ribosomal shunting required for expression of all ORFs within the polycistronic 35S RNA (18).

We provide here an analysis of the interaction between CaMV and the Arabidopsis RNA silencing machinery. We identify the origin, size, and biosynthetic pathways required for accumulation of the predominant CaMV-derived vsRNA species found in infected tissues. This analysis also revealed that vsRNAs can potentially down-regulate the expression of many host genes through nucleotide sequence-specific interactions. These results reveal a new layer of complexity in the plant antiviral RNA silencing mechanism.


   Results
Top
 Abstract
 Results
Discussion
 Materials and Methods
 Acknowledgements
 References
 
The 35S Leader Is a Major Source of CaMV-Derived sRNAs, Which Accumulate as Discrete, 21- to 22-nt and 24-nt Species in Infected Plants. Turnip plants (Brassica rapa rapifera) and Arabidopsis ecotype Col-0 were inoculated with CaMV (Fig. 1A). Systemically infected tissues were harvested 21 days postinoculation (dpi), and Northern blot analyses carried out on low-molecular-weight RNA to assay for accumulation of vsRNA. By using a DNA probe corresponding to the full-length viral genome, 24-nt and 21- to 22-nt RNAs were detected in both plants, although the latter RNA species was less abundant in Arabidopsis than in turnip (Fig. 1B). To investigate the origin of those vsRNAs, low-molecular-weight RNA from infected Arabidopsis (21 dpi) was sequentially hybridized with a set of {approx}1,000-bp DNA probes covering the entire CaMV genome (Fig. 1C, diagram). This analysis revealed a strong vsRNA signal originating from the 5' end of the 35S RNA (Fig. 1C, lanes 1a and 1b); a less abundant signal was also detected with a probe corresponding to the 5' end of the shorter 19S RNA (Fig. 1C, lane 6). Under these hybridization conditions, there were no appreciable signals from any of the remaining probes specific for the other regions of the CaMV genome (Fig. 1C, lanes 2–5, 7, and 8).


Figure 1
View larger version (37K):
[in this window]
[in a new window]

  		Fig. 1. CaMV infection in B. rapa and Arabidopsis triggers accumulation of vsRNAs derived from the 35S RNA leader sequence. (A) Organization of the double-stranded circular DNA genome of CaMV and the two 35S and 19S viral RNAs. The leader sequence is in the 5' end of the 35S RNA. (B) RNA blot analyses of low-molecular-weight RNAs from CaMV-infected plants. The probe used covered the entire CaMV genome. 5S, ethidium bromide staining of 5S ribosomal RNA. (C) Schematics of the 35S and 19S viral RNAs and RNA blot analyses of low-molecular-weight RNAs derived from different regions of the CaMV genome (nucleotide coordinates are indicated) by using DNA probes. Only a 1,500-nt fragment between nucleotides 4700 and 6264 was not probed. The probe against the leader sequence (1b) allowed detection of mainly 24-nt vsRNAs at short exposures ({approx}2 h). Longer exposures ({approx}20 h) revealed 21-nt vsRNA accumulation. (D) Predicted strong secondary structure of the 35S RNA leader (Vienna RNA folding package). Col-0, Columbia.

The strong vsRNA signal from the 5' end of the 35S RNA likely originated from the extensive fold-back structure of the 35S leader (Fig. 1D). To test this idea, we used an oligonucleotide probe (1a) spanning 40 nt located just upstream of the stem-loop leader structure or a 600-nt DNA probe (1b) exactly spanning the entire stem-loop sequence. The vsRNA signal was detected only with the 600-nt probe (Fig. 1C). As observed with the full-length CaMV DNA probe, the 24-nt siRNA species detected with the leader-specific probe was prominent compared with the 21- to 22-nt species, which was nevertheless evident after longer exposures (Fig. 1C, lane 1b lower). These results identify the translational leader as a major source of CaMV-derived sRNAs and suggest that this stem-loop is normally processed by distinct DCL activities.

DCL2, DCL3, and DCL4, but Not DCL1, Efficiently Process the 35S Leader into Specifically Sized vsRNAs. To characterize the genetic requirements for biogenesis of leader-derived vsRNAs, we did Northern blot analyses of low-molecular-weight RNAs extracted from CaMV-infected Arabidopsis carrying dcl knockout mutations (Col-0 ecotype, 21 dpi). We initially used the same leader-specific probe as used in Fig. 1C. The prominent 24-nt siRNA species was as abundant in dcl2 and dcl4 mutants as in WT plants. However, it was lost in dcl3 mutants, in which only 21-nt siRNA accumulated (Fig. 2A). The vsRNA pattern was unaltered in rdr2 and rdr6 mutants (Fig. 2A), suggesting a limited contribution of de novo dsRNA synthesis to vsRNA accumulation, in agreement with the extensive fold-back structure of the 35S leader (Fig. 1D).


Figure 2
View larger version (50K):
[in this window]
[in a new window]

  		Fig. 2. DCL dependencies of vsRNA patterns and CaMV accumulation in Arabidopsis silencing mutants. (A) Northern blots of low-molecular-weight RNA accumulating in CaMV-infected silencing mutants of Arabidopsis. Probe was specific for the entire leader sequence. The numbers of inoculated plants showing systemic infection are from two separate experiments involving five plants each. (B) Same as in A, but the probe was an end-labeled oligonucleotide corresponding to the sRCC1 vsRNA (sequence on top) identified through the procedure described in Fig. 3. (C) Same as in B, but in the dcl1–9 and hen1 mutants. (D) Same as in C, but with two independent alleles of dcl1. The blot was exposed longer to allow detection of vsRNA accumulation (Lower). (E) Immunoblot of CaMV CP (a product of the 35S RNA) accumulation in various RNA silencing mutants and combinations thereof. Proteins were extracted from the samples used in A–D. 5S, ethidium bromide staining of 5S ribosomal RNA; Tot prot: total protein revealed by Coomassie staining. (F) Northern blot analysis of high-molecular-weight RNA extracted from the samples used in A–D. (Right) Composite comparing effects of dcl1–9 and triple dcl2-dcl3-dcl4 mutations.

The very strong 24-nt siRNA signal from the leader-specific probe hampered further investigation of the 21-nt siRNA species. To solve this problem and improve vsRNA resolution, similar analyses were performed with a 20-nt probe specific for a small region of the 35S leader, processed as a discrete vsRNA species. The procedure for identifying this vsRNA, coined "sRCC1" (Fig. 2B, diagram), is described below. As shown in Fig. 2B, the sRCC1 probe detected a 24-nt RNA signal and a less abundant, 21-nt RNA signal in WT plants. The vsRNA patterns in various dcl and rdr mutants were similar to those detected with the probe spanning the entire leader sequence (Fig. 2, compare A with B). To further dissect vsRNA biogenesis, these experiments were repeated in double and triple dcl mutants. These analyses confirmed that biogenesis of 24-nt vsRNA was DCL3 dependent, because this species was lost in dcl3, dcl2-dcl3, dcl3-dcl4, and dcl2-dcl3-dcl4 mutants (Fig. 2B). The 21-nt siRNA accumulation was lost in dcl4 and dcl2-dcl4 mutants, suggesting that this RNA species is processed by DCL4 (Fig. 2B). Analysis of dcl3-dcl4 mutants showed strong accumulation of 22-nt RNAs, also detectable in infected WT Col-0 plants (Fig. 2B). Synthesis of the 22-nt RNA likely required DCL2 because it was lost in dcl2-dcl4, dcl2-dcl3 double mutants and in dcl2-dcl3-dcl4 triple mutants, the latter accumulating very low levels of 21-nt vsRNA (Fig. 2B). This residual processing of vsRNA could be explained by the inefficient action of DCL1 (the only active DCL in the triple dcl2-dcl3-dcl4 knockout) or by suboptimal activities of other Arabidopsis RNase-III, as proposed recently (19). These vsRNA accumulation patterns and DCL dependencies were not specific to sRCC1, because they were also detected with other 20-nt probes corresponding to other regions of the leader (Fig. 3A and data not shown). We conclude that, in natural infection contexts, processing of the 35S RNA leader results from the combined action of DCL2, -3, and -4 producing 22-, 24-, and 21-nt vsRNAs, respectively.


Figure 3
View larger version (35K):
[in this window]
[in a new window]

  		Fig. 3. The leader-derived sRCC1 vsRNA down-regulates accumulation of an Arabidopsis transcript and acts as a bona fide siRNA in infected turnip. (A) Discrete, leader-derived sRNAs, as detected by Northern blot of low-molecular-weight RNA from CaMV-infected Arabidopsis. Probes used were oligonucleotides complementary to indicated sequences. sRNAs in black were not found to exhibit significant complementarity to Arabidopsis transcripts as assessed by BLAST searches. sRNAs in blue were found to exhibit near-perfect complementarity to Arabidopsis transcripts (BLAST search results only). (B) Schematic of predicted sRCC1 target position in At1g76950. (C) sQPCR analysis of At1g76950 transcript accumulation in infected (I) and mock-inoculated (M) Arabidopsis. No RT, control reaction performed without reverse transcriptase; UBI, amplification control using ubiquitin-specific primers. (D) sRCC1 accumulation as assessed by Northern blot of low-molecular-weight RNA in CaMV-infected and mock-inoculated turnip leaves. (E) Schematic of sRCC1 sensor transgene constructs and their transient expression in CaMV-infected and noninfected (Mock) turnip leaves at 15 dpi. LB and RB, left and right T-DNA borders, respectively. (F) Immunoblot analysis of GFP accumulation in the tissues depicted in E. e.v., empty vector used as a negative control in infiltrations.

DCL1, but Not HEN1, Facilitates the Accumulation of All Species of Leader-Derived vsRNA. We tested vsRNA accumulation in plants compromised for miRNA biogenesis/accumulation, employing alleles of the dcl1 and hen1 mutations, respectively [Lansberg erecta (Laer) ecotype, 28 dpi]. The levels of the 21- to 22- and 24-nt species of sRCC1 were reduced in the dcl1–9 allele (Fig. 2C). By contrast, both vsRNA species accumulated normally in the hen1 mutant, as compared with WT (Fig. 2C). dcl1–9 is a T-DNA insertion mutant in the RNA-binding domain of DCL1 (20). To rule out that reduction in vsRNA levels was a peculiarity of this allele, a missense allele of DCL1, dcl1–7, was assayed. dcl1–7 carries a Pro-to-Ser substitution within the DExH motif (21). As shown in Fig. 2D Upper, the sRCC1 levels were similarly decreased in dcl1–9 and dcl1–7 plants. Longer exposures showed that 21-, 22-, and 24-nt vsRNA were still detectable in both dcl1 backgrounds (Fig. 2D Lower), agreeing with the partial loss-of-function nature of the mutations. The difference between the effects of hen1 and dcl1 on vsRNA levels was not peculiar to sRCC1; it was also observed with a probe spanning the entire leader and with other 20-nt probes specific for other leader-derived vsRNAs (Fig. 3A and data not shown). We rule out that the particular effect of dcl1 results from pleiotropy because the hen1 mutation also causes strong developmental anomalies that overlap with those of dcl1 (22). These results suggest that DCL1, although not directly involved in the processing of leader-derived vsRNA (Fig. 2B, track dcl2-dcl3-dcl4), specifically facilitates their biogenesis by the three other DCLs.

Effects of dcl Mutations on CaMV Accumulation. We sought to determine the effects of single and combined dcl mutations on CaMV accumulation. We monitored viral particle formation in immunoblot analyses of the 35S RNA-encoded coat protein (CP). Viral titers (21 dpi) were not significantly different between WT plants and single dcl2, dcl3, and dcl4 mutants (Fig. 2E Upper Left); the same observation was made in double dcl2-dcl3, dcl2-dcl4, and dcl3-dcl4 mutants (Fig. 2E Upper Right). However, the virion levels were higher in triple dcl2-dcl3-dcl4 mutants (Fig. 2E Upper Right). Comparisons of WT Laer plants and hen1 and dcl1 mutants (28 dpi) did not reveal any significant change in CaMV titers (Fig. 2E Lower). Northern blots of RNA extracted from the same tissues showed that 35S RNA levels were significantly higher in dcl2-dcl3-dcl4 mutants, whereas they remained similar to those of WT infected plants in all other single and combination dcl mutants (Fig. 2F and data not shown). We conclude from those analyses that the combined inactivation of DCL2, -3, and -4, which causes a near-complete loss of vsRNA processing (Fig. 2B), is necessary to promote hypersusceptibility to CaMV in Arabidopsis. By contrast, loss of DCL1 activity, which only partially reduces vsRNA accumulation (Fig. 2D), is not sufficient to promote hypersusceptibility to the virus.

A Leader-Derived vsRNA Induces Down-Regulation of an Endogenous Gene Belonging to the Regulator of Chromosome Condensation (RCC1) Family. Several mammalian DNA viruses were shown to produce sRNAs from discrete, genomic stem-loop structures (reviewed in ref. 23). Those molecules appear to play important roles in cis by regulating viral genome expression, and, in at least one example, they were also shown to inhibit expression of host transcripts in trans (24). To test whether, likewise, some CaMV-derived vsRNA could inhibit expression of specific host transcripts, we subjected the entire 35S RNA leader sequence to a BLAST search against Arabidopsis cDNAs and ESTs, looking for possible 18- to 25-nt microhomologies. By using a procedure allowing a maximum of two mismatches, three Arabidopsis transcripts were retrieved (Fig. 3A), among which was the mRNA for At1g76950. The predicted product of At1g76950 awaits functional characterization, but it contains a zinc-finger motif and a domain characteristic of the members of the RCC1 family of proteins that bind to chromatin and interact with the GTP-binding factor RAN. Expression profiling indicates that At1g76950 is mainly expressed in flower buds (TAIR gene expression resource). Sequence alignment revealed a near-perfect complementarity between the 5' UTR of At1g76950 and a 20-nt sequence located in the descending arm of the 35S leader (Fig. 3A, section 4, and B). Using a complementary oligonucleotide probe, we confirmed, by Northern blot analysis, the accumulation of a corresponding vsRNA (sRCC1) in CaMV-infected but not in mock-inoculated Arabidopsis (Fig. 2 B–D and Fig. 4A). The sequence alignment (Fig. 3B) suggested that sRCC1 could act as a bona fide siRNA to promote cleavage of At1g76950 transcripts during infections. Semiquantitative RT-PCR (sQPCR) analyses indeed revealed that At1g76950 accumulation was dramatically reduced in CaMV-infected but not in mock-inoculated Arabidopsis (Fig. 3C).


Figure 4
View larger version (17K):
[in this window]
[in a new window]

  		Fig. 4. Tens of Arabidopsis transcripts are potentially targeted by sRCC1 and other leader-derived vsRNAs. (A) sRCC1 accumulation in infected tissues of the At1g76950 T-DNA insertion mutant, as assessed by Northern blot of low-molecular-weight RNA. The last panel shows CaMV CP accumulation assessed by immunoblot. (B) sQPCR analysis of At1g75330 and At3g52500 transcript levels in leaves of CaMV-infected (I) and mock-inoculated (M) plants at 21 dpi (33 amplification cycles, Upper). (Lower) QPCR analysis of At1g75330 transcript levels in WT and dcl2-dcl3-dcl4 mutant backgrounds. For each sample, mRNA levels were normalized to that of Actin2 (At3g18780). Error bars represent standard deviation from three independent experiments involving triplicate PCRs. (C) The sRCC1 sequence was queried for complementarity to Arabidopsis transcripts by using miRU settings with decreasing stringencies. Mism, mismatch; G:U, G:U wobbles allowances. The transcript sets retrieved under high stringency conditions were always contained within those obtained under lower stringency settings. The target transcript number reached a plateau of 90. Sequence alignments and identities of predicted targets are in Table 1 and Data Set 1, which are published as supporting information on the PNAS web site.

To validate that sRCC1 could directly guide sequence-specific transcript cleavage in natural infection contexts, GFP-sensor constructs were engineered in which the 3' UTR contained either the authentic target sequence found within At1g76950 or a modified version carrying five nucleotide substitutions predicted to abolish annealing and cleavage by sRCC1 (Fig. 3E). The constructs were delivered by Agrobacterium leaf infiltration into CaMV-infected turnip plants in which accumulation of sRCC1 was confirmed at 15 dpi (Fig. 3D). Five days after agroinfiltration, GFP expression was monitored under UV light, and its accumulation was quantified by immunoblot analyses (Fig. 3 E and F). These levels were consistently higher with GFP-RCC1mut, carrying mismatches, than with GFP-RCC1wt, bearing the authentic sRCC1target (Fig. 3F Left). Both constructs were similarly expressed in uninfected turnip plants (Fig. 3F Right). Addition of the sRCC1-complementary region of At1g76950 was sufficient to promote sequence-specific down-regulation of a reporter mRNA during CaMV infection. These findings strongly suggest that reduction of At1g76950 levels in Arabidopsis results from direct action of the CaMV-derived sRCC1.

Leader-Derived sRNAs Potentially Down-Regulate a Large Number of Arabidopsis Transcripts. To examine whether down-regulation of At1g76950 affected virus accumulation, we infected a T-DNA insertion mutant of At1g76950 with CaMV. The infection kinetics and CaMV accumulation (21 dpi) were unchanged in the mutant compared with WT-infected plants, as were vsRNA levels (Fig. 4A and data not shown). Therefore, the At1g76950 knockout did not impact the infection, suggesting that either targeting of At1g76950 by sRCC1 had no incidence (at least under laboratory conditions) or there is functional redundancy among At1g76950 homologs. Another possibility was that sRCC1 and other vsRNAs normally target other host genes in addition to At1g76950, so that their combined inactivation would have been required to appreciably affect viral titers. To test this idea, we subjected the deduced 20-nt sequence of sRCC1 to the miRU algorithm, designed to identify putative Arabidopsis target transcripts of queried, 19- to 28-nt RNAs. miRU takes into account the thermodynamics and 5' seed-pairing requirements for optimal miRNA and siRNA activities in plants (25). By allowing a maximum of two nucleotide mismatches and only one possible G:U wobble, an additional set of 14 putative sRCC1 mRNA targets was retrieved (see Table 1), with At1g75330 and At3g52500 (encoding, respectively, an ornithine carbamoyltransferase/ornithine transcarbamylase and an aspartyl protease) ranking the highest scores. sQPCR analyses indicated that both transcripts were down-regulated in CaMV-infected but not mock-inoculated Arabidopsis (Fig. 4B Upper). This down-regulation likely resulted from the specific effects of CaMV-derived siRNAs (i.e., sRCC1) because quantitative RT-PCR (QPCR) analyses showed that the decrease in At1g75330 transcript levels was much less pronounced between infected and noninfected dcl2-dcl3-dcl4 mutant plants than between infected and noninfected WT plants (Fig. 4B Lower), consistent with the fact that accumulation of leader-derived vsRNA is strongly reduced in dcl2-dcl3-dcl4 (Fig. 2B).

The full spectrum of host mRNA potentially targeted by sRCC1 was revealed by using miRU searches with increasing tolerance for base mismatches and G:U wobbles, giving a total of 90 candidates, 65% of which carried the sRCC1 target sequence within 5' or 3' UTRs (Fig. 4C, Table 1, and Data Set 1). Gene ontology analysis showed that the majority of identified targets are involved in basic cell metabolism, with a preponderance of chloroplast-targeted products (Fig. 5, which is published as supporting information on the PNAS web site). A similar gene set was retrieved if the queried sRCC1 sequence had an additional 3 nt on either 5' or 3' ends, taking into account the fact that the sRNA also accumulates as 22- and 24-nt species (data not shown). Further miRU analyses (low stringency settings) identified >30 potential hits with similar ontology and UTR biases for vsRNA no. 2; there were no or few hits with the other vsRNAs detected by Northern blot (Fig. 3A and data not shown). We conclude that discrete vsRNA species derived from the 35S RNA leader may account for down-regulation of large numbers of Arabidopsis transcripts through 20- to 24-nt stretches of sequence complementarity.


   Discussion
Top
 Abstract
 Results
 Discussion
Materials and Methods
 Acknowledgements
 References
 
We show that the CaMV 35S RNA leader contributes significantly to vsRNA production. However, the nonexhaustive nature of our mapping approach does not rule out participation of other genomic regions, one of which, located at the 5' end of the 19S RNA, was not further investigated in this study. The results with the 35S leader echo those of previous studies showing that the majority of vsRNAs cloned from tombusvirus-infected plants mapped to partially base-paired hairpins found in positive genomic RNA strands (26). Thus, with both DNA and RNA viruses, imperfect intramolecular, rather than perfect intermolecular RNA base-pairing might strongly stimulate antiviral DCL activities. The analysis of dcl mutants indicates that the 35S leader is efficiently processed into 21- and 24-nt siRNAs by DCL4 and -3, respectively. The action of DCL2, generating 22-nt siRNAs, was mostly evident when DCL4 was genetically inactivated, whereas siRNA processing by DCL1 was vastly suboptimal. We recently found that a similar consortium of hierarchical DCL activities accounts for vsRNA accumulation from three distinct cytoplasmic RNA viruses (13) despite the fact that the replication strategies and subcellular localization of their genomes are drastically dissimilar to those of CaMV minichromosomes, transcribed in the nucleus. These observations raise important issues regarding our current views of DCL localization/function because all four Arabidopsis DCLs (with the possible exception of DCL2) have been assigned strict nuclear localizations based on experiments with GFP-reporter gene fusions (5): either those GFP fusions do not provide realistic localization information, or some dsRNA generated by RNA viruses is taken into the nucleus for processing into siRNAs. A third possibility is that DCLs might relocalize during infections.

In RNA virus infections, combined inactivation of DCL4 and DCL2 was both necessary and sufficient to promote hypersusceptibility, whereas DCL3 had no antiviral effects (13). This was not the case with CaMV, because increased viral titers were observed only when DCL4, DCL2, and DCL3 were simultaneously inactivated. Thus, the 24-nt siRNA product of DCL3 apparently impacts CaMV, but not RNA virus accumulation. To explain the difference we envisage two nonmutually exclusive hypotheses. First, it could be that CaMV accumulation is merely restricted through dicing of the 35S RNA leader, with limited contribution of an antiviral RISC, as suggested to explain the resistance of extensively base-paired viroid genomes to degradation by viroid-derived siRNAs (27). This would explain why enhanced CaMV accumulation was evident only when DCL2, -3, and -4 were simultaneously inactivated because they individually contributed strongly to leader-derived vsRNA production. Alternatively, given the contribution of DCL3 to cytosine/histone methylation at endogenous loci, 24-nt vsRNAs could contribute to transcriptional inactivation of CaMV minichromosomes. In this case, inactivating both transcriptional (DCL3) and posttranscriptional (DCL4 and -2) barriers to CaMV accumulation would be required to promote hypersusceptibility.

A second, notable difference between RNA virus and CaMV infections is the facilitating effect of DCL1 on vsRNA accumulation. This feature of DCL1 is unlikely to involve its vsRNA-processing activity because it was suboptimal, as revealed in dcl2-dcl3-dcl4 triple mutants. We also rule out that synthesis of CaMV-derived vsRNA requires endogenous miRNA-directed functions because vsRNA levels were unaffected by the hen1 mutation, which promotes a dramatic destabilization of miRNAs (28). A possible explanation lies in the fact that DCL1 not only processes mature miRNAs from pre-miRNAs, but it is also required in the nucleus to generate pre-miRNA from long, primary miRNA transcripts (29), a step catalyzed in animal cells by the RNase-III enzyme Drosha (30). Because the nuclear 35S RNA has features of pri-miRNA, it is conceivable that DCL1 excises the 35S leader as a cognate pre-miRNA structure. Excision of the stem-loop might then facilitate its optimal processing by the three other DCLs. Those effects of nuclear DCL1 were not previously detected during RNA virus infections possibly because, unlike the 35S RNA, genomic transcripts of RNA viruses accumulate in the cytoplasm. In any case, those findings define a sRNA biosynthetic pathway in plants that involves the coordinated action of all four DCLs. This pathway may also be recruited endogenously. We note, for instance, that several intergenic sRNA loci that form stem-loop structures in Arabidopsis have been classified as miRNA genes based on a stringent requirement of DCL1 for sRNA accumulation but, in several cases, the impact of DCL2, -3, and -4 was not tested. We also note that the DCL1-dependent release of miR163 from premiR163 was coincident with the biogenesis of at least two more sRNAs with ill-defined biosynthetic requirements (29).

Tomato plants expressing viroid-derived inverted repeat transgenes (producing viroid siRNAs) develop symptoms that phenocopy those of authentic viroid infections (27). This observation, together with recent studies carried out with mammalian DNA viruses has substantiated the idea that infections by viral and subviral pathogens might promote host gene-specific knockdown owing to sequence homology with vsRNAs (10, 23, 31). Our findings with CaMV provide direct experimental support to this hypothesis in plants. Hence, production of sRCC1 may account, on its own, for the down-regulation of many Arabidopsis transcripts. Given the density and diversity of vsRNA produced from the 35S leader, dozens, perhaps hundreds, of host genes might be affected. Virus-induced silencing of host gene expression might partly account for the profound modifications in cell metabolism that are commonly elicited by diverse plant viruses (32). Host gene silencing might also contribute to the expression of symptoms, the molecular bases of which are currently poorly understood. This specific question is complicated in the case of CaMV because responses to CaMV infection in Arabidopsis are influenced by many parameters, including vegetative–floral transition, that are themselves altered in some RNA silencing mutants (33). In addition, the CaMV P6 protein is, on its own, an important symptom determinant (34).

One issue is whether host gene-targeting by CaMV occurs fortuitously (i.e., resulting in colateral damages) or represents a bona fide viral strategy to facilitate infection. One way to address this issue would be to measure the impact of vsRNA sequence polymorphisms on virus fitness, in a given, genetically tractable host context. However, the translational leader is extremely conserved in sequence and structure among CaMV strains, a feature that likely reflects its fundamental roles in CaMV genome expression. Engineering small sequence or structural alterations in the 35S RNA leader also can have strong detrimental impact on translation efficacy, and consequently, on virus fitness (35, 36). Codepletions of host mRNAs predicted as "sensitive" vsRNA targets (e.g., using multiple T-DNA insertions) could provide an alternative method to address this issue. Our analysis with sRCC1 uncovered a significant bias for vsRNA matches toward 5' and 3' UTR regions, which, unlike coding regions, are often divergent in sequence and size among transcripts from closely related species. Assuming that host-gene silencing plays important roles for CaMV, it is possible that 5' or 3' UTR polymorphisms might contribute to the differences in infection phenotypes observed among host plant species.


   Materials and Methods
Top
 Abstract
 Results
 Discussion
 Materials and Methods
Acknowledgements
 References
 
Plant Material. All mutant combinations were based on dcl1-9, dcl1–7, dcl2–1, dcl3–1, dcl4–2, rdr2, rdr6–15, and hen1–1, described previously (5, 8, 37). The dcl2-dcl3 and dcl2-dcl4, dcl3-dcl4, and dcl2-dcl3-dcl4 triple mutant lines were as described (13). The At1g76950 T-DNA insertion mutant line was ordered at the Nottingham Arabidopsis Stock Centre (Loughborough, U.K.). Plants were germinated in standard greenhouse conditions for 3 weeks and transferred into short-days conditions for a further 2–3 weeks to prevent bolting, on which infections were carried out.

Viruses and Infection Assays. Arabidopsis and turnip infections were carried out with CaMV virions (strain Cabb B-JI). Virions were obtained by infecting B. rapa rapifera with SalI-linearized pMD324-CaMV-JI plasmid. Infected leaves were harvested at 21 dpi, ground in liquid nitrogen, and stored at –80°C. All infections of B. rapa ({approx}3 weeks old) and Arabidopsis (rosette leaves, {approx}5–6 weeks old, before bolting) were performed with a fresh sap prepared from 100 mg of ground leaf material diluted into 300 µl of sterile water. The results reported in this study were exclusively obtained from systemic, infected tissues. Samples were collected at 21 dpi for mutants in the Col-0 ecotype and at 28 dpi for mutants in the Laer ecotype.

Nucleic Acids. Total RNA from leaves was extracted by using TRIzol (Invitrogen), precipitated with isopropanol, and redissolved in 50% formamide. Northern blot analyses of low- and high-molecular-weight RNA were performed with 10 and 5 µg of total RNA, respectively, as described (13). Radiolabeled probes for detection of the whole CaMV genome or the full-length leader were made by random priming reactions (Promega, Madison, WI) in the presence of [{alpha}-32P]dCTP. The probe specific for sRCC1 was a DNA oligonucleotide complementary to the vsRNA sequence, end-labeled with [{gamma}-32P]ATP by using T4 polynucleotide kinase (New England Biolabs).

Immunoblot. Total proteins from CaMV-infected plants were extracted in Laemmli buffer, resolved by SDS/PAGE, and transferred by electroblotting onto a PVDF membrane (Immobilon-P; Millipore). CP antiserum (Mario Keller, Institut de Biologie Moléculaire des Plantes) was used at a 1:15,000 dilution and revealed by a second peroxidase-conjugated antibody (Biosys) through enhanced chemiluminescence (Lumi-light PLUS; Roche).

sQPCR and QPCR Analyses. Reactions were as described (6) and involved 1 µg of total RNA. DNaseI and reverse transcription reactions (SuperScript; Promega) were performed according to the manufacturer. Ubiquitin-specific (sQPCR) and actin-specific (QPCR) amplifications were used to adjust the amount of cDNA within each sample analysis.

GFP Sensor and Transient Expression Experiments. The ER-mGFP5 cDNA was PCR-amplified with forward and reverse primers. Reverse primers contained the authentic sRCC1 target sequence, as found in At1g76950 (GFP-RCC1wt) or a derivative with mismatches (GFP-RCC1mut). PCR products were mobilized into pBin61 binary vector transformed into Agrobacterium strain GV3101. Agrobacterium cells were grown and induced as described (38) and diluted in 10 mM MgCl2 to a final OD600 of 0.8. Agroinfiltration was carried out in CaMV-infected B. rapa or in mock-inoculated leaves, at 15 dpi. At 5 d after agroinfiltration, total proteins were extracted from the infiltrated regions for immunoblot analyses. Pictures were taken under a Nikon SMZ15000 dissecting microscope coupled to a 100-W epifluorescence module.


   Acknowledgements
Top
 Abstract
 Results
 Discussion
 Materials and Methods
 Acknowledgements
References
 
We thank members of the O.V. laboratory, especially P. Brodersen and A. Deleris, for helpful discussions and Richard Wagner's team for plant care. This work was supported by a PhD fellowship from the Mininstry of Research (to G.M.). The project is an objective of the Silencing RNA: organisers and coordinators of complexity in eukaryotic organisms (SIROCCO) Integrated Project of the European Union.


   Footnotes
 

Abbreviations: dpi, days postinoculation; RISC, RNA-induced silencing complex; DCL, Dicer-like; miRNA, microRNA; CaMV, cauliflower mosaic virus; sRNA, small RNA; vsRNA, virus-derived sRNA; CP, coat protein; Laer, Lansberg erecta; QPCR, quantitative RT-PCR; sQPCR, semiquantitative RT-PCR.

*To whom correspondence should be addressed. E-mail: olivier.voinnet{at}ibmp-ulp.u-strasbg.fr

Freely available online through the PNAS open access option.

Author contributions: O.V. designed research; G.M. performed research; G.M. and O.V. analyzed data; and O.V. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS direct submission.

© 2006 by The National Academy of Sciences of the USA


   References
Top
 Abstract
 Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 

  1. Baulcombe, D. (2004) Nature 431, 356–363.[CrossRef][Medline]
  2. Brodersen, P & Voinnet, O. (2006) Trends Genet 22, 268–280.[CrossRef][ISI][Medline]
  3. Vaucheret, H. (2006) Genes Dev 20, 759–771.[Abstract/Free Full Text]
  4. Bartel, DP. (2004) Cell 116, 281–297.[CrossRef][ISI][Medline]
  5. Xie, Z, Johansen, LK, Gustafson, AM, Kasschau, KD, Lellis, AD, Zilberman, D, Jacobsen, SE & Carrington, JC. (2004) PLoS Biol 2, E104.[CrossRef][Medline]
  6. Dunoyer, P, Himber, C & Voinnet, O. (2005) Nat Genet 37, 1356–1360.[CrossRef][ISI][Medline]
  7. Gasciolli, V, Mallory, AC, Bartel, DP & Vaucheret, H. (2005) Curr Biol 15, 1494–1500.[CrossRef][ISI][Medline]
  8. Xie, Z, Allen, E, Wilken, A & Carrington, JC. (2005) Proc Natl Acad Sci USA 102, 12984–12989.[Abstract/Free Full Text]
  9. Borsani, O, Zhu, J, Verslues, PE, Sunkar, R & Zhu, JK. (2005) Cell 123, 1279–1291.[CrossRef][ISI][Medline]
  10. Voinnet, O. (2005) Nat Rev Genet 6, 206–220.[CrossRef][ISI][Medline]
  11. Wang, XH, Aliyari, R, Li, WX, Li, HW, Kim, K, Carthew, RW, Atkinson, P & Ding, SW. (2006) Science 312, 452–454.[Abstract/Free Full Text]
  12. Silhavy, D & Burgyan, J. (2004) Trends Plants Sci 9, 76–83.
  13. Deleris, A, Gallego-Batolome, J, Bao, J, Kasschau, KD, Carrington, JC & Voinnet, O. (2006) Science 313, 68–71.[Abstract/Free Full Text]
  14. Akbergenov, R, Si-Ammour, A, Blevins, T, Amin, I, Kutter, C, Vanderschuren, H, Zhang, P, Gruissem, W, Meins, F Jr, Hohn, T & Pooggin, MM. (2006) Nucleic Acids Res 34, 462–471.[Abstract/Free Full Text]
  15. Chellappan, P, Vanitharani, R & Fauquet, CM. (2004) J Virol 78, 7465–7477.[Abstract/Free Full Text]
  16. Al-Kaff, NS, Kreike, MM, Covey, SN, Pitcher, R, Page, AM & Dale, PJ. (2000) Nat Biotechnol 18, 995–999.[CrossRef][ISI][Medline]
  17. Al-Kaff, NS, Covey, SN, Kreike, MM, Page, AM, Pinder, R & Dale, PJ. (1998) Science 279, 2113–2115.[Abstract/Free Full Text]
  18. Ryabova, LA & Hohn, T. (2000) Genes Dev 14, 817–829.[Abstract/Free Full Text]
  19. Bouche, N, Lauressergues, D, Gasciolli, V & Vaucheret, H. (2006) EMBO J 25, 3347–3356.[CrossRef][ISI][Medline]
  20. Jacobsen, SE, Running, MP & Meyerowitz, EM. (1999) Development (Cambridge, UK) 126, 5231–5243.[Abstract]
  21. Golden, TA, Schauer, SE, Lang, JD, Pien, S, Mushegian, AR, Grossniklaus, U, Meinke, DW & Ray, A. (2002) Plant Physiol 130, 808–822.[Abstract/Free Full Text]
  22. Park, W, Li, J, Song, R, Messing, J & Chen, X. (2002) Curr Biol 12, 1484–1495.[CrossRef][ISI][Medline]
  23. Cullen, BR. (2006) Nat Genet 38, Suppl, S25–S30.[CrossRef][Medline]
  24. Gupta, A, Gartner, JJ, Sethupathy, P, Hatzigeorgiou, AG & Fraser, NW. (2006) Nature 442, 82–85.[Medline]
  25. Zhang, Y. (2005) Nucleic Acids Res 33, W701–W704.[Abstract/Free Full Text]
  26. Szittya, G, Molnar, A, Silhavy, D, Hornyik, C & Burgyan, J. (2002) Plant Cell 14, 359–372.[Abstract/Free Full Text]
  27. Wang, MB, Bian, XY, Wu, LM, Liu, LX, Smith, NA, Isenegger, D, Wu, RM, Masuta, C, Vance, VB & Watson, JM, et al. (2004) Proc Natl Acad Sci USA 101, 3275–3280.[Abstract/Free Full Text]
  28. Yu, B, Yang, Z, Li, J, Minakhina, S, Yang, M, Padgett, RW, Steward, R & Chen, X. (2005) Science 307, 932–935.[Abstract/Free Full Text]
  29. Kurihara, Y & Watanabe, Y. (2004) Proc Natl Acad Sci USA 101, 12753–12758.[Abstract/Free Full Text]
  30. Lee, Y, Ahn, C, Han, J, Choi, H, Kim, J, Yim, J, Lee, J, Provost, P, Radmark, O, Kim, S & Kim, VN. (2003) Nature 425, 415–419.[CrossRef][Medline]
  31. Dunoyer, P & Voinnet, O. (2005) Curr Opin Plant Biol 8, 415–423.[CrossRef][ISI][Medline]
  32. Whitham, SA, Quan, S, Chang, HS, Cooper, B, Estes, B, Zhu, T, Wang, X & Hou, YM. (2003) Plant J 33, 271–283.[CrossRef][ISI][Medline]
  33. Leisner, SM, Turgeon, R & Howell, SH. (1993) Plant Cell 5, 191–202.[Abstract]
  34. Yu, W, Murfett, J & Schoelz, JE. (2003) Mol Plant Microbe Interact 16, 35–42.[ISI][Medline]
  35. Dominguez, DI, Ryabova, LA, Pooggin, MM, Schmidt-Puchta, W, Futterer, J & Hohn, T. (1998) J Biol Chem 273, 3669–3678.[Abstract/Free Full Text]
  36. Pooggin, MM, Hohn, T & Futterer, J. (2000) J Biol Chem 275, 17288–17296.[Abstract/Free Full Text]
  37. Allen, E, Xie, Z, Gustafson, AM & Carrington, JC. (2005) Cell 121, 207–221.[CrossRef][ISI][Medline]
  38. Voinnet, O, Rivas, S, Mestre, P & Baulcombe, D. (2003) Plant J 33, 949–956.[CrossRef][ISI][Medline]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg    What's this?


This article has been cited by other articles in HighWire Press-hosted journals:


Home page
 J. Virol.Home page
X. Zhang, G. C. Segers, Q. Sun, F. Deng, and D. L. Nuss
Characterization of Hypovirus-Derived Small RNAs Generated in the Chestnut Blight Fungus by an Inducible DCL-2-Dependent Pathway
J. Virol., March 15, 2008; 82(6): 2613 - 2619.
[Abstract] [Full Text] [PDF]

Home page
 J. Virol.Home page
A. Valli, G. Dujovny, and J. A. Garcia
Protease Activity, Self Interaction, and Small Interfering RNA Binding of the Silencing Suppressor P1b from Cucumber Vein Yellowing Ipomovirus
J. Virol., January 15, 2008; 82(2): 974 - 986.
[Abstract] [Full Text] [PDF]

Home page
 J. Gen. Virol.Home page
A. J. Love, J. Laird, J. Holt, A. J. Hamilton, A. Sadanandom, and J. J. Milner
Cauliflower mosaic virus protein P6 is a suppressor of RNA silencing
J. Gen. Virol., December 1, 2007; 88(12): 3439 - 3444.
[Abstract] [Full Text] [PDF]

Home page
 DNA ResHome page
Y. Tagami, N. Inaba, N. Kutsuna, Y. Kurihara, and Y. Watanabe
Specific Enrichment of miRNAs in Arabidopsis thaliana Infected with Tobacco mosaic virus
DNA Res, December 1, 2007; (2007) dsm022v1.
[Abstract] [Full Text] [PDF]

Home page
 Plant Physiol.Home page
H. B. Scholthof
Heterologous Expression of Viral RNA Interference Suppressors: RISC Management
Plant Physiology, December 1, 2007; 145(4): 1110 - 1117.
[Full Text] [PDF]

Home page
 J. Virol.Home page
Q.-S. Du, C.-G. Duan, Z.-H. Zhang, Y.-Y. Fang, R.-X. Fang, Q. Xie, and H.-S. Guo
DCL4 Targets Cucumber Mosaic Virus Satellite RNA at Novel Secondary Structures
J. Virol., September 1, 2007; 81(17): 9142 - 9151.
[Abstract] [Full Text] [PDF]

Home page
 J. Virol.Home page
J. Qu, J. Ye, and R. Fang
Artificial MicroRNA-Mediated Virus Resistance in Plants
J. Virol., June 15, 2007; 81(12): 6690 - 6699.
[Abstract] [Full Text] [PDF]

This Article
Free via Open Access: OA
Right arrow OA Abstract
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Supporting Information
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a colleague
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Moissiard, G.
Right arrow Articles by Voinnet, O.
Right arrow Search for Related Content
PubMed</