SEARCH

SEARCH BY CITATION

Keywords:

  • herbivore resistance;
  • Nicotiana attenuata;
  • RNA-directed RNA polymerase;
  • transcriptional regulation;
  • RdR1;
  • nicotine biosynthesis

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

Small RNAs are important regulators of plant development and resistance to viruses. To determine whether small RNAs mediate defense responses to herbivore attack, we silenced the expression of three RNA-directed RNA polymerases (RdRs) in the native tobacco Nicotiana attenuata by virus-induced gene silencing. Larvae of the leaf-chewing solanaceous specialist Manduca sexta grew faster on the RdR1-silenced plants than on empty vector (EV) controls; silencing RdR3 and 2 had little to no effect on larval performance. NaRdR1 transcripts were strongly elicited when puncture wounds were treated with M. sexta oral secretions (OS) to simulate herbivore attack, and with SA and JA, phytohormones that are elicited by herbivore attack. Stably silencing RdR1 by transforming N. attenuata with an inverted-repeat RdR1 construct produced plants (irRdR1) that grew normally but were highly susceptible to both M. sexta larvae and the cell-content-feeder Tupiocoris notatus. When irRdR1 lines were planted into N. attenuata’s native habitat in the Great Basin Desert (Utah, USA), they were highly susceptible to herbivore attack, due to deficiencies in direct rather than indirect defenses. Microarray analysis revealed the downregulation of ADC and ODC genes, which supply substrates for synthesizing the chemical defense compound nicotine; irRdR1 lines failed to accumulate nicotine after attack. We conclude that RdR1 mediates herbivore resistance, and infer that the small RNAs produced by RdR1 are probably involved in orchestrating some of the rapid metabolic adjustments required for plants to survive herbivore attack in their natural habitats. The experiment highlights the value of carrying out ‘real-world’ tests of gene function early in the discovery process.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

Attack from herbivorous insects elicits a large-scale reconfiguration of plant metabolism (Kessler and Baldwin, 2004; Walling, 2000). Because insects are highly mobile, the responses spread rapidly throughout a plant and are frequently elicited by herbivore-specific elicitors introduced into wounds during feeding (Korth, 2003; Voelckel and Baldwin, 2004b). The responses include the production and activation of direct defenses, such as toxins, digestibility reducers and anti-feedants, that directly protect plants (Duffey and Stout, 1996), as well as indirect defenses that recruit natural enemies from the plant’s surroundings to attack feeding herbivores (Turlings and Tumlinson, 1992; De Moraes et al., 1998; Kessler and Baldwin, 2001). A large-scale, transcriptional response, which precedes and presumably mediates many of the induced defense responses, is also elicited by herbivore attack. For instance, in Populus, 1728 genes are differentially regulated after attack by forest tent caterpillars, Malacosoma disstria (Ralph et al., 2006). On the other hand, in Arabidopsis, only 114 genes are differentially regulated after attack by Pieris rapae larvae (Reymond et al., 2004). Although this transcriptional response did not differ substantially among plants attacked by lepidopteran larvae that are known to be Brassica specialists and generalists, the transcriptional responses to herbivore attack in the Solanaceous taxa are known to be highly herbivore-specific (Voelckel and Baldwin, 2004a,b; Voelckel et al., 2004) and species-specific (Qu et al., 2004; Schmidt et al., 2005). Transcriptional responses, such as the defense responses, are known to spread from the attack site throughout the plant rapidly and systemically follow vascular connections between attacked and unattacked plant parts (Schittko and Baldwin, 2003; Van Dam et al., 2001). How these large and rapidly dispersed transcriptional responses to herbivore attack are coordinated remains unknown, except that phytohormone signaling, in particular oxylipin signaling, is involved (Halitschke et al., 2004; Howe et al., 1996; Reymond et al., 2004; Ryan and Pearce, 2003). Both the speed and extent of the responses suggest that transcriptional regulation by the movement of proteins or phytohormones may not be the only mechanism.

RNA silencing is emerging as a fundamental regulatory process affecting many layers of endogenous gene expression (Voinnet, 2002): non-coding small RNAs appear to be important regulators of gene expression in both plants and animals (Bartel, 2004; Mallory and Vaucheret, 2004). RNA silencing has been shown to be essential for plant development and differentiation processes, such as embryonic and vegetative organ formation, leaf morphogenesis and flower development (Chen, 2004; Emery et al., 2003; Llave et al., 2002; Mallory et al., 2004; Palatnik et al., 2003). Small RNAs are now regarded as key elements that, depending on their source, the RNA and the nature of the interaction with the target nucleic acid, trigger chromosomal modifications (Aufsatz et al., 2002; Martienssen, 2003; Mette et al., 2000), post-transcriptional gene silencing (Dalmay et al., 2000; Mourrain et al., 2000; Yu and Kumar, 2003) and a translational blockade. Of the small RNAs, a large class of 18–26 nucleotide long RNAs [small-interfering (si)RNAs and micro (mi)RNAs] is involved in defense against viruses, in post-transcriptional gene silencing, and in regulating developmental genes through mRNA degradation or translational repression (Bartel, 2004; Pickford and Cogoni, 2003). The role of endogenous RNA interference in the mediation of responses to herbivore attack remains unexplored. Moreover, none of the small RNA-mediated phenotypes have been examined in organisms living in real-world settings. Given that components of the endogenous RNAi system are sensitive to abiotic stresses (Borsani et al., 2005), it is not clear how consistently the responses will be expressed in organisms growing in complex environments.

All of the RNA-silencing pathways involve cleaving double-stranded (ds) RNA into short 21–26 nucleotide RNAs (Baulcombe, 2004). The dsRNA molecules are produced by RNA-directed RNA polymerases (RdRps; Pickford and Cogoni, 2003), or RdRs according to the new nomenclature (Wassenegger and Krczal, 2006), in Caenorhabditis elegans (Sijen et al., 2001), fungi (Cogoni and Macino, 1999) and plants (Dalmay et al., 2000; Mourrain et al., 2000), suggesting that they mediate mechanisms important for genetic regulation. As plant miRNAs are similar to siRNAs, miRNAs may be serving as primers that allow RdRs to generate dsRNA (Tang et al., 2003; but see Petersen and Albrechtsen, 2005). Alternatively, miRNAs (along with RdRs) may be involved in additional processes (Allen et al., 2005): miRNAs may mediate the formation of pre-siRNAs, and then, using these pre-siRNAs, RdRs may form dsRNAs, and, after degradation, the remaining strand of dsRNA, siRNA, regulates the mRNA targets (Allen et al., 2005).

Three functionally distinct RdRs have been reported in Arabidopsis, tomato (Solanum lycopersicum) and Nicotiana (Dalmay et al., 2000; Mourrain et al., 2000; Schiebel et al., 1998; Yang et al., 2004). In separate but overlapping processes (Pickford and Cogoni, 2003), RdR1 and 6 (SDE1/SGS2 and their natural variants) are thought to be involved in virus resistance and post-transcriptional gene-silencing mechanisms; however, the role of the third RdR (RdR2) is not understood (reviewed by Wassenegger and Krczal, 2006). The transitivity of the RNA-silencing signal depends on the activity of the RdRs (Himber et al., 2003). The RdR6 homolog is required for the cell to perceive the silencing signal but not to produce or transport it (Schwach et al., 2005). RdR1 and 6 are elicited by salicylic acid (SA) treatment in tobacco and Arabidopsis, but these responses appear to be confined specifically to viral defense (Yang et al., 2004; Yu et al., 2003). There is clearly much more to be learned about the role that RdRs play in RNA silencing, and silencing the expression of these key enzymes in order to examine the phenotypes of RdR-silenced plants represents a valuable means of uncovering the function of RdRs.

Here we explore the role of RdRs in mediating defense responses to herbivore attack in a model ecological expression system whose defense responses are well characterized. Nicotiana attenuata is a native of the southwestern United States and grows in the immediate post-fire environment. It times seed germination from a long-lived seed bank by responding to pyrolysis products found in wood smoke (Preston and Baldwin, 1999). Because the plant ‘chases’ the ephemeral post-fire environment in time, the herbivore community is forced to re-establish itself with each new plant population (Baldwin, 2001). The resulting unpredictability of the composition of herbivore communities has probably selected for the plant’s vast array of inducible defenses. These allow the plant to shape its resistance traits according to the herbivore community that attacks it at a given location (Kessler and Baldwin, 2004; Voelckel and Baldwin, 2004a). How the plant tailors its defense responses is best understood in the case of attack from the larvae of the specialist lepidopteran herbivore Manduca sexta, the species that regularly accounts for the majority of the leaf area lost to herbivores in native populations (Kessler and Baldwin, 2002).

Attack from this larvae results in the differential regulation of about 500 N. attenuata genes, which can be crudely classified into functional categories such as photosynthesis, electron transport, primary metabolism, signaling, cytoskeleton, secondary metabolism, DNA-binding proteins, stress responsive factors, etc (Heidel and Baldwin, 2004; Hermsmeier et al., 2001; Hui et al., 2003; Schmidt et al., 2005). These large-scale, rapid, herbivore-specific transcriptional responses can be elicited by applying M. sexta oral secretions (OS) to puncture wounds (Halitschke et al., 2001; Roda et al., 2004). Eight fatty acid–amino acid conjugates present in M. sexta OS are necessary and sufficient to elicit the response (Halitschke et al., 2001; Roda et al., 2004), which in turn requires jasmonate (JA) signaling. The importance of JA signaling is apparent from the highly attenuated response in plants transformed to silence the specific lypoxygenase (NaLox3) that supplies hydroperoxide substrates for JA biosynthesis (Halitschke and Baldwin, 2003). JA signaling transcriptionally elicits a number of potent direct defenses including the neurotoxin nicotine (Winz and Baldwin, 2001). Once its genes for nicotine biosynthesis have been silenced, N. attenuata is highly vulnerable to herbivores in nature (Steppuhn et al., 2004). Yet how these rapidly activated responses are regulated at molecular levels remains unknown.

Here we explore the involvement of RNA interference in regulating herbivore-induced plant defense responses by independently silencing the expression of the three RdRs present in the N. attenuata genome. Our study highlights the value of testing the functional significance of a gene under real-world circumstances at an early stage in the discovery process. In the study of traits mediating pathogen resistance, tests of real-world significance are not usually conducted until most of the mechanistic details mediating the response are understood. Using an herbivore susceptibility screen with virus-induced gene-silenced (VIGS) plants in glasshouse experiments, we identified an RdR gene (RdR1) that, when silenced, increased the susceptibility of plants to M. sexta attack. Plants were produced that were stably silenced for RdR1 expression by Agrobacterium-mediated transformation and susceptible to M. sexta attack as well as to attack from a cell-content-feeding herbivore, Tupiochoris notatus. The stably silenced plants were planted into N. attenuata’s native habitat and further characterized.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

Isolation of RdR1 from N. attenuata

We used a PCR-based approach with N. attenuata DNA and isolated the complete coding region of RdR1. NaRdR1 has high sequence similarity (>90%) with its corresponding homolog from N. tabacum and N. benthamiana, but no similarity with other RdRs from closely related species. Screening a cDNA library prepared from N. attenuata leaves after 24 h of continuous M. sexta attack did not yield any positive clones, indicating that RdR1 is expressed at very low levels (Yang et al., 2004). The probes used in the screening were obtained by PCR amplification of 434 bp fragments of RdR1 from N. attenuata genomic DNA. Similarly, partial NaRdR2 and 3 sequences were obtained by the PCR amplification of 350 and 353 bp fragments from the genomic DNA of N. attenuata. The primer sequences are given in Experimental procedures. The RdR1 and partial RdR2 and 3 sequences from N. attenuata were aligned with the other characterized RdR genes from closely related Solanaceous species, namely N. benthamiana, N. tabacum and Solanum lycopersicum (formerly Lycopersicon esculentum), and the six putative RdRs from Arabidopsis, and a phylogenetic tree was prepared (Figure 1).

image

Figure 1.  Phylogenetic analysis of N. attenuata RdR1. The full-length NaRdR1 and partial NaRdR2 and 3 sequences were aligned with sequences of six RdR genes from Arabidopsis (AtRdR1, 2, 3, 4, 5 and AtSGS2), three RdR genes from N. benthamiana, and N. tabacum and S. lycopersicum (L. esculentum) RdR genes. Distance values were calculated according to the neighbor-joining method with 1000 bootstrap replicates.

Download figure to PowerPoint

Virus-induced gene silencing (VIGS) of RdR genes and M. sexta performance

In order to study the role of the three RdR genes in defense against herbivores, we silenced the expression the three RdR genes independently using a VIGS system optimized for N. attenuata (Saedler and Baldwin, 2004), in which unique fragments of the three RdR genes (Figure S1) were expressed in a TRV vector and agro-inoculated into the plants (Ratcliff et al., 2001). We monitored the progress of silencing by observing the bleaching in plants inoculated with pTVPDS constructs containing a 206 bp fragment of the N. benthamiana phytoene desaturase gene (Figure S2). The first bleaching symptoms appeared 9 days after inoculation, and all five replicate plants showed bleaching symptoms at day 11. A quantitative real-time PCR analysis revealed that all three RdR genes were silenced (Figure 3). In RdR1-silenced plants, RdR1 transcript levels were reduced by more than 50%, while levels of RdR2 and 3 were unaffected. In RdR2-silenced plants, RdR2 transcript levels were reduced by 81%, while RdR1 transcript levels were unchanged and RdR3 transcripts increased by 58%. When transcripts of RdR3 were silenced, RdR3 levels were reduced by 64%, while RdR2 transcript levels were unaltered and RdR1 levels were reduced by 28%. Thus, quantitatively, the RdR1-silencing was completely specific, effecting no change in non-target RdR genes, and largely specific for RdR2 and 3, with only minor changes in the non-target RdR genes. Just as there were no apparent abnormalities in plant morphology, there were no significant variations in the height of plants with the silencing constructs compared to empty vector (EV) control plants even 30 days after inoculation (Figure S2; anova, F3,93 = 2.628; = 0.0549).

image

Figure 3. M. sexta performance after virus-induced gene silencing (VIGS) of the three N. attenuata RdR genes. The bar graphs represent the relative transcript levels of (a) RdR1, (b) RdR2 and (c) RdR3 in the empty vector (EV) and the three ptvRdR lines. All comparisons were made relative to transcript levels in EV (set to 1). The line graph represents the performance of M. sexta larvae on EV and the various RdR lines. Larval mass was measured at 4-day intervals. **Significantly different from EV at ≤ 0.001.

Download figure to PowerPoint

To examine the roles of RdR genes in mediating direct defenses against M. sexta attack, we measured larval growth on the three RdR-silenced lines as well as on the EV control plants in an experiment that included 20 replicate plants per line (Figure 3). After 12 days of continuous feeding, the larvae growing on the RdR1-silenced plants gained more than twice as much weight as those growing on EV control plants (anova, F1,38 = 17.28; = 0.0002), and those growing on RdR3-silenced lines gained 1.5 times as much weight as those growing on EV control plants (anova, F1,38 = 10.461; = 0.0025). Larval performance on the RdR2-silenced plants compared to EV controls did not differ significantly (anova, F1,38 = 3.685; = 0.0625). As the TRV-RdR3 construct also reduced the accumulation of RdR1 transcripts (by 28%), the significant increase in larval performance could be due to the silencing of RdR1 rather than of RdR3.

RdR1 expression patterns

Because M. sexta larvae performed best on RdR1-silenced plants, we characterized RdR1 transcript accumulation in WT plants by quantitative real-time PCR (Figure 2 and Figure S3). Enhanced RdR1 levels were found when the WT plants were subjected to various elicitors. Elicitation with jasmonic acid (JA), salicylic acid (SA) and OS increased RdR1 transcript accumulation up to 10-fold within 1 h, but no change in transcript levels was recorded when plants were only wounded (Figure 2). This suggests that RdR1 is involved in direct plant defense against herbivores.

image

Figure 2. Characterization of RdR1 from N. attenuata. Time-course analysis of RdR1 transcript induction with various elicitors. High accumulations were of induced NaRdR1 transcripts was recorded within 1–4 h of elicitation with 1 mm JA (dashed line, solid squares), 2 mm SA (dotted lines, open circles) and simulated herbivory (wounding with fabric pattern wheel and immediate treatment with 20 μl M. sexta OS; solid lines, solid triangles), but not after mechanical wounding alone (thin line with cross). At time 0, plants were induced with the various elicitors, and induced NaRdR1 levels were compared to constitutive levels at the time of elicitation.

Download figure to PowerPoint

Stably silencing RdR1 increases performance of herbivores from two feeding guilds

Plants stably silenced in RdR1 expression were produced by transforming N. attenuata with the same RdR1 sequence used in the TRV-RdR1 construct in an inverted repeat orientation using Agrobacterium-mediated transformation (Kruegel et al., 2002). Transformed plants were subjected to high-throughput phenotype screening (Kruegel et al., 2002), and transgene incorporation as a single copy was verified by Southern blot analysis (Figure S4). Two independently transformed lines harboring a single insert (234–10 and 265–7; Figure S5) were analyzed for RdR1 transcript levels after OS elicitation: neither had accumulated any elicited transcripts (Figure 4). As RdR1 has been associated with resistance to plant viruses in N. tabacum (Xie et al., 2001), we used a virus susceptibility screen as a phenotypic test of RdR1 silencing at a functional level. Both the lines were highly susceptible to tobamo-viruses, whose consequences include highly impaired growth, and rapid senescence and death (Figure S6).

image

Figure 4. NaRdR1 expression in two independently transformed homozygous lines harboring a single copy of a fragment of NaRdR1 in an inverted-repeat construct. Both lines (234-10, dashed line; 265-7, dotted line ) failed to accumulate NaRdR1 transcripts after OS elicitation.

Download figure to PowerPoint

To investigate the role of RdR1 in plant defense against insect herbivores, we challenged the irRdR1 plants with two different herbivores that regularly attack N. attenuata in its native habitat, M. sexta and T. notatus, and compared their performance to that on WT or stably transformed empty vector (EV) plants. No difference was recorded in the mass of M. sexta larvae growing on WT or EV plants (Figure 5a; anova, F1,30 = 0.117; P = 0.735; a comparison of WT and EV plants is presented in Table S1: no differences were observed). M. sexta larvae grew faster on plants of line 234–10 (Figure 5a; anova, F2,38 = 4.32; P = 0.02) than on WT plants (Fisher’s PLSD  = 0.0068) or EV control plants (Fisher’s PLSD = 0.0256). Similarly, M. sexta larvae performed better on plants of line 265–7 (anova, F2,44 = 4.487; P = 0.016) than on WT plants (Fisher’s PLSD = 0.0064) or EV control plants (Fisher’s PLSD = 0.033). Also, both irRdR1 lines were more severely damaged than were WT or EV plants (Figure 5a; anova, F3,50 = 6.9; P = 0.0006). When all three genotypes were challenged with T. notatus, both irRdR1 lines suffered significantly more damage than did controls (Figure 5b; anova, F3,50 = 3.846; P = 0.0149). The damage to WT plants did not differ from the damage to EV plants (Fisher’s PLSD = 0.7882).

image

Figure 5. irRdR1 lines were highly susceptible to herbivores from two different feeding guilds. (a) The performance of M. sexta larvae on WT (♦), EV (bsl00001) and two irRdR1 lines (bsl00066, 234–10; bsl00084, 265–7) was recorded during 11 days of continuous feeding. Larval mass was recorded at intervals of 2 days. The bar graph (inset) represents the total amount of damage (percentage of leaf area) at the end of 11 days. (b) The performance of T. notatus on different lines in terms of percentage of damage after 6 days of insect attack. **Significantly different from WT and EV plants at ≤ 0.01; *significantly different from WT and EV plants at ≤ 0.05.

Download figure to PowerPoint

Silencing RdR1 increases N. attenuata’s vulnerability to herbivores in their native habitats due to deficiencies in direct rather than indirect defenses

In order to determine the ecological relevance of RdR1 silencing, we examined the performance of both transformed lines in the plant’s native habitat in the Great Basin Desert (Utah, USA). irRdR1 plants did not defend themselves well against the native herbivore community. Mirids, beetles and grasshoppers were the main herbivores attacking the plants at their release sites, with mirids causing the most damage. The total canopy area damaged was recorded 10 days after release in the field. Because comparisons in the laboratory showed that the EV behaved like WT, and no differences between the WT and EV controls were observed (showed above and in Table S1), we used WT plants as controls in the field experiments. Plants from line 234–10 suffered damage to 15% more leaf area than did WT plants (Figure 6; n = 10 pairs, paired t-test, t = 3.672, P = 0.0051). In a subsequent assessment after another 5 days, herbivory had started to decline, but irRdR1 plants still suffered more damage than did WT plants (n = 10 pairs, paired t-test, t = 2.435, P = 0.0377). A third and final assessment was made 5 days after the second assessment. The decline in herbivory continued, with consistently more damage to irRdR1 plants than to WT controls (Figure 6; n = 10 pairs, paired t-test, t = 3.744, P = 0.0046). Similar patterns were seen for the plants of line 265–7 in comparison to WT controls (Figure 6). The contribution of the various herbivorous taxa to the total canopy damage at given times is presented in pie charts in Figure 6. There were no apparent abnormalities in plant morphology or significant differences in plant height between irRdR1 and WT plants even in the field (line 234–10, n = 10 pairs, paired t-test, t = 0.0556, P = 0.592; line 265–7, n = 9 pairs, paired t-test, t = 1.604, P = 0.1474).

image

Figure 6. Performance of irRdR1 lines in natural habitats. Every irRdR1 plant from each line (234–10 and 265–7) was individually paired with a WT plant. Plants were transplanted to the field at day 0 and were undamaged. Cumulative damage on irRdR1 plants (dotted lines) and WT plants (solid lines ) was monitored three times, at 10, 15 and 20 days after transplanting to the field. The individual pie charts represent the contribution of particular herbivores to the total herbivory (measured in terms of total plant area damaged). Pie charts in the upper panel of the line graph are for irRdR1 lines and those in the lower panel are for WT. **Significantly different from WT plants at ≤ 0.01; *significantly different from WT plants at ≤ 0.05.

Download figure to PowerPoint

Because Geocoris pallens predators have been shown to be attracted to herbivore-attacked plants by the release of volatile organic compounds (VOCs) into the atmosphere, making VOCs an effective indirect defense in nature (Kessler and Baldwin, 2001), we performed a Manduca egg predation assay using the field-grown lines. As with M. sexta caterpillar attack, when N. attenuata leaves are elicited with OS, they release VOCs that attract G. pallens from the surroundings; these insects then consume the affixed eggs. Predation on M. sexta eggs was < 0.05% (both on WT and line 234–10) before OS elicitation just 48 h after the eggs were affixed. As OS elicitation is known to attract G. pallens predators, we treated plants by mechanically wounding the first stem leaf and immediately applying OS. A 12–16% increase in predation was observed 48 h after elicitation, but there were no differences between WT and line 234–10 (paired t-test, n = 10 pairs, = 0.56, P = 0.591). After 72 h of elicitation, predation rates increased to 42–44%, but still no differences in M. sexta egg predation between WT and line 234–10 were observed (paired t-test, n = 10 pairs, = 0.19, = 0.853).

Transcriptional responses of RdR1-silenced plants

In order to understand how the herbivore-induced transcriptional responses are altered in RdR1-silenced plants, we performed microarray analysis using a custom microarray that was developed to characterize the M. sexta-induced responses in N. attenuata (Halitschke and Baldwin, 2003; Voelckel and Baldwin, 2004a,b). We hybridized arrays using RNA extracted from irRdR1 and WT plants grown in their natural habitat in the Great Basin Desert , and elicited plant responses by treating puncture wounds with M. sexta OS 2 h prior to tissue harvest. The microarray analysis was replicated with independent biological samples. For each of the two microarrays, RNA was extracted from three replicate WT and line 234–10 plants. The OS-induced transcripts levels were significantly altered for 45 genes in irRdR1 plants compared to WT plants (Figure S7).

NaLOX3 is the specific lipoxygenase in N. attenuata required for JA signaling, which in turn mediates many of N. attenuata’s induced defense responses. Because silencing the NaLOX3 gene also decreases the resistance of N. attenuata to M. sexta attack (Halitschke and Baldwin, 2003), as was observed for RdR1-silenced plants, we compared the transcriptional responses of irRdR1 plants to OS elicitation with the responses of as-lox3 plants (Halitschke and Baldwin, 2003). There was very little overlap in the transcriptional responses of the two genotypes. Photosynthetic genes were the most commonly downregulated genes. The gene Sn-1 and dioxygenase-like proteins (stress-responsive genes) were upregulated in both irRdR1 and as-lox3 lines. On the other hand, as-lox3 responses substantially differed from irRdR1 responses. The signature JA-responsive genes such as HPL (hydroperoxide lyase, responsible for production of volatiles), TD (threonine deaminase) and NaPI (proteinase inhibitor) were downregulated in as-lox3 but not in irRdR1 plants. This suggests that the susceptibility of irRdR1 plants is probably not due to impaired JA signaling.

The microarray analysis revealed that the cell wall extensin precursor and other stress-responsive genes (phosphatase 2C, superoxide dismutase and pathogenesis-related proteins) commonly upregulated by OS elicitation in WT plants were downregulated in RdR1-silenced plants, and this may account for the greater susceptibility of these plants to herbivore attack. Aldolase-like protein and threonine synthase (primary metabolism) as well as gibberellin-2-oxidase and gibberellic-acid-insensitive (GAI)-like protein (hormone metabolism), and polyphenol oxidase and Solanumnigrum prosystemin (secondary metabolism and signaling) were upregulated. Serine protease inhibitors (Pin2) along with the subtilisin-like proteinase were also upregulated. How these relate to the herbivore phenotype is not known.

The key biosynthetic genes for putrescine, arginine decarboxylase (ADC) and ornithine decarboxylase (ODC) were strongly downregulated in irRdR1 compared to WT plants, as were those for nitrate reductase and inorganic phosphatase. As these genes supply the nicotine biosynthetic pathway, we hypothesized that OS-elicited nicotine production was impaired by RdR1 silencing.

RdR1 silencing reduces elicited nicotine production

Nicotine is one of the most important induced direct defense compounds in N. attenuata (Steppuhn et al., 2004). Nicotine was measured in non-wounded and OS-elicited WT and irRdR1 plants. Compared to WT plants, nicotine levels in irRdR1 plants were reduced by 51 and 35% in lines 234–10 and 265–7, respectively, 72 h after elicitation – the time of maximum wound-induced nicotine accumulation (Figure 7a; anova, F5,24 = 4.485, P = 0.005, Fisher’s PLSD < 0.05). The induced nicotine levels in WT were more than twice those of uninduced WT plants (Fisher’s PLSD ≤ 0.0005), whereas the elicited nicotine levels in irRdR1 lines did not differ significantly from those in untreated irRdR1 lines or untreated WT plants (Fisher’s PLSD > 0.05). Similarly, under field conditions, the elicited nicotine levels in line 234–10 were only 41% of those of WT plants (paired t-test, n = 5 pairs, = 3.02, = 0.039), and those in line 265–7 were only 28% of those of WT plants (paired t-test, = 5 pairs, = 3.518, = 0.024).

image

Figure 7. Silencing RdR1 inhibits elicited nicotine production. (a) (upper panel) Fold downregulation of the putrescine biosynthesis enzymes ODC and ADC, along with nitrate reductase (NR); levels of nicotine from field- (middle panel) and glasshouse-grown (lower panel) WT and irRdR1 lines. (b) A simplified scheme for nicotine production. **Significantly different from WT plants at P ≤ 0.01; *significantly different from WT plants at ≤ 0.05.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

Here we expand the functional arena of plant RNA-directed RNA polymerase 1 (RdR1) from post-transcriptional gene silencing and virus resistance (Dalmay et al., 2000; Mourrain et al., 2000; Xie et al., 2001; Yang et al., 2004; Yu et al., 2003) to the regulation of traits mediating the direct defenses of plants against herbivores; N. attenuata plants whose RdR1 gene was silenced by VIGS or stable transformation were highly susceptible to insect herbivores from different feeding guilds in the glasshouse as well as in their native habitat. This demonstration of the ‘real-world’ relevance of the RdR1 phenotype is significant from both mechanistic and pedagogical perspectives.

RdR genes form an important component of the small RNA-regulatory machinery in siRNA biogenesis (Xie et al., 2004). The role of RdR genes in plant defense against viruses and post-transcriptional gene silencing has been studied using mutant screens and challenges with known viruses in the laboratory (Dalmay et al., 2000; Schwach et al., 2005; Vaistij et al., 2002), but, to our knowledge, none of the studies has evaluated in their natural habitat plants whose RdR genes or any other component of the small RNA regulatory system have been silenced. RdR genes act in the biogenesis and transport of siRNAs, which may function epigenetically. In maize, RdR2 has recently been shown to be essential for paramutation, a process by which epigenetic information is transferred to the next generation (Alleman et al., 2006). Although RdR1 has not been strictly associated with paramutations, the function of siRNA in cells may be regarded as an epigenetic process. Epigenetic traits are highly affected by the environment; accordingly, plants silenced for some components of small RNA production or activity may behave differently in natural environments than under controlled conditions. This difference is why it is important to determine whether small RNA-driven phenotypes can be reproducibly identified in organisms growing in complex environments.

A subtext to this study is that it highlights the value of carrying out ‘real-world’ tests of gene function early in the discovery process. Such real-world tests are typically conducted after most of the mechanistic details of a problem have been resolved. Postponing proof-of-function studies can lead to some interesting situations. For example, the pathogen resistance protein, PR1, which has been used for more than three decades of research in plant–pathogen interactions as a ‘reliable marker for pathogen resistance’, has never been shown to be necessary or sufficient for resistance to any pathogen in a ‘real-world’ setting (Niderman et al., 1995). Postponing real-world functional tests is an artifact of the educational chasm that several decades ago split the biology departments of most universities into cell and molecular discern and organismic sub-divisions. As was eloquently illustrated in a recent proteomic analysis of Pseudomonas fluorescens (Knight et al., 2006), organismic-level functional understanding is essential for gene function, and an intensive period of ‘cross-chasm’ training will thus be required for the next generation of biologists.

While this work establishes the real-world significance of RdR1 function, much additional work is needed to understand the mechanisms by which RdR1 mediates herbivore resistance. Herbivore-specific elicitors in OS, or the endogenous signals they elicit, rapidly increase the accumulation of RdR1 transcripts. Microarray analysis revealed very little overlap between the OS-elicited transcriptional signatures in NaLOX3- and RdR1-silenced plants, and N. attenuata’s indirect defenses, which also require intact jasmonate signaling (Halitschke and Baldwin, 2003), are unaffected in RdR1-silenced plants. These results suggest that RdR1 mediates processes that are either jasmonate-independent or downstream of jasmonate signaling. In the microarray analysis, alkaloid biosynthesis genes were found to be downregulated, indicating that the nicotine biosynthetic pathway was affected. Experiments with nicotine-silenced N. attenuata plants grown in natural habitats have demonstrated the importance of this direct defense (Steppuhn et al., 2004). Induced nicotine production may be affected in two ways: either the regulators of ADC and ODC may be direct targets, or the siRNAs may regulate the influx of nitrogen in the metabolism of defense-related compounds. As evidence, in addition to ADC and ODC, nitrate reductase (NR) was also found to be downregulated. The molecular basis of induced alkaloid biogenesis is not well understood, because little is known about OS-inducible transcription factors and repressors. We propose that herbivory elicits RdR1 activity, which then amplifies siRNA genesis. These siRNAs target constitutive repressors of alkaloid biosynthesis, which induce nicotine production. However, when RdR1 is silenced, these repressors are not degraded effectively; therefore, irRdR1 plants cannot produce sufficient nicotine, and are, in turn, susceptible to herbivores.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

Plant and insect material

Wild-type N. attenuata plants were from 17th or 22nd generation inbred lines of seeds originally collected from a native population in Utah (USA). All the plants were grown under the conditions described by Kruegel et al. (2002) and Halitschke and Baldwin (2003). All experiments utilized 28–32-day-old plants at the rosette stage. Eggs of Manduca sexta L. (Lepidoptera) were obtained from North Carolina State University (Raleigh, NC, USA) and Tupiocoris notatus (Hemiptera: Miridae) were collected from the field station in Utah in the south western United States and maintained at the Department of Molecular Ecology, Max Planck Institute for Chemical Ecology, Jena, Germany.

The accession numbers for the other RdRs used in this study are Y10403 (LeRdR), AJ011576 (NtRdR), AY574374 (NbRdR1m), AY722009 (NbRdR2), AY722008 (NbSDE1), At1g14790 (AtRdR1), At4g11130 (AtRdR2), AF239718 (AtRdR6/SGS2), At2g19910 (AtRdR3), At2g19920 (AtRdR4) and At2g19930 (AtRdR5).

Isolating N. attenuata RdR genes

A PCR-based strategy was used to clone the RdR genes from N. attenuata. Genomic DNA was extracted according to the procedures described by Bubner et al. (2004). To isolate RdR1, identical or complementary PCR primers were designed from the corresponding RdR1 sequences from relatives of N. attenuata: N. benthamiana and N. tabaccum. Single bands were gel-purified using GFX PCR DNA and a gel band purification kit (Amersham Biosciences; http://www5.amershambiosciences.com/) according to the manufacturer’s instructions, and sequenced. Sequences were aligned with the corresponding cDNA sequence. Fragments giving a positive alignment were considered to be exons. Sequences that did not match the corresponding homologs were tentative introns and subjected to Fourier analysis to determine whether they were non-coding (Tiwari et al., 1997). The following primer pairs were designed for the PCR amplification of gene fragments of RdR2 and RdR3 from N. attenuata: RDR3-32 (5′-GCGGCGGTCGACGTGCTGCAAGGATGGGTCAG-3′) and RDR4-34 (5′-GCGGCGGGATCCCTTGGTAATATTAAGCATCCTG-3′) for RdR2; RDR31-32 (5′-GCGGCGGTCGACTGAACCGGCAAATAGTAACC-3′) and RDR32-31 (5′-GCGGCGGGATCCAAGCTCACCTAATTCATCC-3′) for RdR3 (SalI and BamHI sites underlined). Gene sequences from various RdR genes were aligned using MegAlign (DNASTAR, http://www.dnastar.com). Using MEGA 3 (Molecular Evolutionary Genetic Analysis; Kumar et al., 2004), a neighbor-joining tree was built. Support for the groups was evaluated using 1000 bootstrap replicates (Wu et al., 2006).

Expression analysis by quantitative real-time PCR

RdR1 and RdR6 in Arabidopsis and N. benthamiana are known to be elicited by SA treatments and virus attack, and their role in viral defense has been demonstrated (Yang et al., 2004; Yu et al., 2003). To determine whether herbivore attack elicits RdR1 transcriptional changes, we wounded plants and immediately treated the puncture wounds with M. sexta OS, which are known to activate the herbivore-specific responses in N. attenuata (Halitschke et al., 2001). The second fully expanded (+2; Van Dam et al., 2001) leaves of three or four rosette-stage plants were wounded by rolling a fabric pattern wheel down the leaf lamina and immediately treating the resulting puncture wounds with 20 μl OS (diluted 1:1 with distilled water) as described by Halitschke et al. (2001) or with water at each time point (described below). Because JA treatment also elicits herbivore-specific responses, we measured RdR1 transcript pools in WT plants sprayed until run-off with 1 mm JA. We also determined the effect of SA on RdR1 levels by spraying plants with a 2 mm SA solution until run-off (Yang et al., 2004). To determine the kinetics of RdR1 transcript accumulation, treated +2 leaves were harvested from three or four replicate plants at 0, 1, 4, 8, 12, 24 and 48 h after each treatment. To determine whether RdR1 transcripts accumulated diurnally, we harvested leaves from untreated +2 nodes of four replicate plants at 4:00, 8:00, 12:00, 16:00, 20:00 and 24:00.

Total RNA was extracted using the Trizol method. Total RNA was reverse-transcribed to prepare first-strand cDNA using the SuperScript first-strand synthesis system for real-time PCR, with oligo(dT) as primers, following the manufacturer’s protocol (Invitrogen; http://www.invitrogen.com/). SYBR green assays were developed (qPCR core kit for SYBR Green I, Eurogentec (http://www.eurogentec.be), following the manufacturer’s protocol) to test the efficiency of the amplicon. All the quantitative real-time PCR assays were performed with cDNA corresponding to 100 ng RNA before transcription as a template, and amplified using qPCR™ core reagent kit (Eurogentec) and gene-specific primers and probes. Each biological replicate was used twice on the qPCR plate. The 2−ΔΔCT method was used for data analysis (Bubner et al., 2004). To simplify data interpretation, expression levels in control plants (0 h treatment) were fixed to 1, and relative expression levels were calculated with respect to this reference value. To determine the RdR levels in the silenced lines, gene-specific primers were designed outside the region used in the silencing constructs. All the gene-specific primers and probes were designed with Primer Express software (Applied Biosystems, http://www.appliedbiosystems.com).

Silencing the RdR genes of N. attenuata

The virus-induced gene-silencing (VIGS) system based on the tobacco rattle virus (TRV) was used to independently silence the three RdR genes in N. attenuata (Ratcliff et al., 2001; Saedler and Baldwin, 2004). PCR fragments of 434, 350 and 353 bp for RdR1, RdR2 and RdR3, respectively (Figure S2), obtained using the primer pair combinations of RDR1-31 (5′- GCGGCGGTCGACTATGATCCAGTGAGGTGGC-3′) and RDR2-31 (5′- GCGGCGGGATCCATCCACACTGAATTATCCC-3′) for RdR1 (SalI and BamHI sites underlined), RDR3-32 and RDR4-34 for RdR2, and RDR31-32 and RDR32-31 for RdR3 (primer sequences given above) were digested with SalI and BamHI (site also in the amplified RdR1 sequence) and cloned in plasmid pTV00 (Ratcliff et al., 2001) cut with the same enzymes. The inserts of the resulting TRV-based vectors pTVRDR1, pTVRDR2 and pTVRDR31 (all 5.9 kb) for RdR1, RdR2 and RdR3, respectively, were sequenced. Empty vector (EV; pTV00) constructs served as controls for this experiment. Inoculating plants with a TRV vector containing a 206 bp fragment of phytoene desaturase (PDS) gene from N. benthaminana (pTVPDS), which bleaches plants as the PDS gene is silenced, allowed us to monitor the progress of the VIGS (Figure S2). Forty-five replicate plants per construct were inoculated with each of the three RdR constructs and an EV control construct, and five replicate plants were inoculated with PDS constructs. The plants used for inoculation of the VIGS constructs were 25–28-day-old N. attenuata WT plants, at the rosette stage. Growth conditions at the start of the experiment were 20°C, 65% relative humidity, and no light for 2 days, after which light levels were returned to normal (400–1000 μmol m−2 sec−1, 16/8 h light/dark).

In order to produce plants stably silenced for RdR1 expression, an RdR1 gene fragment was cloned in an inverted repeat orientation in a pRESC5 transformation vector as described by Steppuhn et al. (2004) and Bubner et al. (2006). Transformation using Agrobacterium tumefaciens was performed as described by Kruegel et al. (2002). T1 plants were screened for hygromycin resistance. Homozygosity was determined by segregation analysis of T2 plants. Quantitative real-time PCR was used to quantify transcript accumulation, as described above, and Southern analysis was used for copy number determination of the transgene. Two independently transformed homozygous lines (234–10 and 265–7), each containing a single insertion of the transgene, were further characterized in the T2 generation. As RdR1 is thought to be required for resistance to viruses, we used a virus assay as an additional positive control for RdR1 silencing. The +1, +2 and +3 leaves of the WT and transgenic plants were inoculated with tomato mosaic virus. Leaves of three replicate plants of each line at the rosette stage of growth (28 days after germination) were rubbed with corborundum powder, and 50 μl of viral material suspended in phosphate buffer was applied to the abraded leaves. Equal numbers of plants from each line were rubbed with corborundum powder and treated with 50 μl of phosphate buffer without any virus as mock control. Plants were monitored for symptom development for 12–14 days.

Insect performance assays

The performance of M. sexta larvae was evaluated for the VIGS-silenced plants and the stably transformed lines in separate experiments. A freshly enclosed larva was placed on the +2 leaf of each replicate plant, 14–15 days after inoculation with one of the four pTRV constructs (three RdR constructs and the one empty vector control) under the temperature and light conditions required for VIGS as described above. Bleaching symptoms appeared in all five replicate plants 11 days after inoculation with pTVPDS. Larval mass was recorded every 4th day for 12 days, and 20 replicate plants per construct were used. Performance assays were also carried out for each transgenic line (234–10, 265–7, EV) and for WT plants in a separate experiment with 10–16 replicate plants per line, grown under normal glasshouse conditions. In this experiment, larvae were allowed to feed for 11 days and data were recorded every 2 days for 11 days starting from day 3. The percentage of total damage to each transgenic line and to WT plants was estimated at the end of the assay.

T. notatus performance assays were conducted on stably transformed lines. Twelve to 14 replicate plants from four genotypes (234–10, 265–7, EV and WT), all in the early elongation stage of growth, 35 days after germination, were enclosed in a completely randomized manner in a mesh tent. Mirids are highly mobile and move readily among the plants in a tent. Approximately 250 adult mirids were released into the tents and allowed to feed on the plants. After 6 days, the percentage of leaf area damaged was measured.

Performance under field conditions

WT and transgenic irRdR lines were planted into natural habitats of N. attenuata in the south western United States. Plants were placed in a watered field plot at the Lytle Preserve research station (Santa Clara, UT, USA) in a paired design. Seeds of WT and irRdR1 plants were germinated on agar plates. The plates were then kept at 25°C/16 h in the light (200 µm m−2 sec−1) and 20°C/8 h in the dark. After 10 days, seedlings were transferred to Jiffy 703 pots (1 3/4 × 1 3/4 inch, AlwaysGrows, http://www.alwaysgrows.com) which had been soaked in borax solution (0.4 mg/45 ml water). The seedlings were fertilized with iron solution (stock solution: 2.78 g FeSO4·7 H2O and 3.93 g Titriplex in 1 l H2O, diluted 100-fold for fertilization) after 7 days. After 3–4 weeks, plants were transferred to the field plot. They were allowed to gradually adapt to the environmental conditions of the Great Basin Desert (high sun exposure and low relative humidity) over 2 weeks in a mesh tent. Ten to twelve irRdR1–WT pairs of adapted seedlings of the same size were transplanted to a nearby, watered field plot. Seedlings were watered every other day for 2 weeks until roots were established in the native soil. Releases of the transformed plants were conducted under APHIS notification number 06-003-08n. The plants were colonized by the natural herbivore community for 3 weeks and the study was terminated after 28 days. All the capsules were plucked off and destroyed, and all the plants in and around the plantation plot were removed and destroyed to comply with 7CFR 340.4, the legal statute which governs the release of transgenic organisms. The leaves were scrutinized at intervals of 5 days for the characteristic damage of the various herbivores that commonly attack N. attenuata in Utah (Figure S8), such as mirids, grasshoppers, beetles, etc., and total herbivory was estimated as a percentage of the total canopy area.

M. sexta eggs are laid singly as well as in clusters of 4–7 on the underside of the leaf surface. In order to determine whether N. attenuata’s indirect defenses, specifically its ability to attract predators with OS-elicited VOCs, were altered in irRdR1 plants, we conducted a predation assay (Kessler and Baldwin, 2001): five M. sexta eggs were glued using natural glue (known to have no effects on VOC production) on the second stem leaf of 10 replicate WT and irRdR1 plant pairs which had not been previously attacked by M. sexta. Predation rates were measured twice at an interval of 24 h after the M. sexta eggs were attached. Because OS elicitation mimics the release of VOCs that normally occurs after larval attack and the VOCs attract G. pallens predators that prey on M. sexta eggs and larvae, the first stem leaves were elicited with M. sexta OS and the number of eggs consumed 24, 48 and 72 h after elicitation was determined.

Microarray analysis

Microarray analysis for samples derived from the above-described field study was performed using microarrays enriched with M. sexta-responsive N. attenuata genes that had been previously used to characterize the insect-induced responses in N. attenuata (Voelckel and Baldwin, 2004a,b) in accordance with the MIAME guidelines. A similar hybridization strategy was adapted as described by Halitschke and Baldwin (2003), in which samples from NaLOX3-silenced plants (as-lox3) were Cy3-labeled and hybridized against WT Cy5-labeled samples. In the field, for each chip, the second fully expanded ( + 2) leaves of three plants of the 234–10 line and the WT controls were elicited with M. sexta OS as described above. The treated leaves were harvested after 2 h. Total RNA was extracted from three biological replicate plants, and an equal amount of RNA from each replicate was used for each chip. RNA from the treated 234–10 plants was labeled with Cy3; RNA from the treated WT control was labeled with Cy5. Approximately 400 µg total RNA was used in each labeling reaction. The whole procedure was replicated, and two arrays were hybridized. Microarray data were lowess-normalized with the MIDAS package (Microarray Data Analysis System, Institute for Genome Research, Washington DC, USA). The quadruplicate spots for each gene were analyzed for significant differences using a t-test at a confidence level (α) of 0.05, and a threshold of a 1.5-fold change in expression ratio was used. A gene was regarded as differentially regulated if it met both the criteria in both microarrays. In some genes where the values were present only for one channel, the data were evaluated for differences from the signal-to-noise ratio, and if the intensity was >2.5 times the signal-to-noise ratio, the gene was regarded as differentially regulated.

Analysis of secondary metabolites

Secondary metabolites were analyzed using HPLC as described by Steppuhn et al. (2004). Briefly, leaf samples from the field (zero or first stem leaves) as well as glasshouse-grown plants (second fully expanded leaves) from five or six replicate plants for each line and WT were isolated 72 h after OS elicitation. Samples (approximately 100 mg) were extracted with 2:3 methanol: 0.5% acetic acid (v/v) and injected into the HPLC system. A standard curve was generated using a dilution series of nicotine, and nicotine levels were quantified.

Statistical analysis

Data (arcsine-transformed wherever they did not meet assumptions of normality) were analyzed with StatView (Abacus Concepts Inc., http://www.statview.com). Insect assays in the glasshouse were analyzed by anova. All the field data or data derived from samples from field were tested with a paired t-test because a transgenic line and WT control plant were planted as a single pair in all the field experiments.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

This research was supported by the Max Planck Society. We thank E. Wheeler for editorial comments, Drs R. Halitschke and M. Wassenegger for stimulating initial discussions, K. Gase for initial isolation of partial gene fragments and preparing silencing constructs, T. Hahn, W. Kroeber, S. Allmann and S. Kutschbach for invaluable assistance with the microarray hybridization and scanning, VIGS experiments and plant transformation, C. Diezel for assistance in insect assays, D. Kessler for patient training in field work and pictures of herbivore damage, and Dr E. Gaquerel for invaluable help with the field work.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

Accession numbers NCBI accession numbers DQ988990 (NaRdR1), DQ988991 (partial NaRdR2) and DQ988992 (partial NaRdR3).

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  9. Supporting Information

Supporting Method Figure 1. Characteristic damage symptoms of different herbivore species that attack N. attenuate plants in Utah (pictures by D. Kessler). Figure S1. Partial sequences of the three N. attenuate RdRs used to make silencing constructs. Figure S2. No morphological abnormalities were seen after silencing the three RdRs with VIGS during (A) rosette- (B) flowering-stage growth. Bleaching symptoms resulting from silencing phytoene desaturase (PDS) show the spread of the VIGS throughout the plant. Figure S3. Relative expression of RdR1 in WT N. attenuate. (A) RdR1 expression was detected in all plant parts; (B) The constitutive levels did not exhibit a diurnal pattern of accumulation. Figure S4. Southern analysis of the two independently transformed irRdR1 lines showing single insertion. Genomic DNA (10 μg) from individual plants was digested with ECoR1 and blotted onto a nylon membrane. The blot was hybridized with a PCR fragment of the hygromycin phosphotransferase II gene, specific for the selective marker on the T-DNA. Figure S5. Growth phenotypes of the two independently transformed lines. No differences in rosette size (A; ANOVA, F3,16 = 1.396, P=0.2803) and stalk length (B and C; ANOVA, F3,20 = 0.268, P=0.847) were observed. Figure S6. irRdR1 silenced lines are highly susceptible to viruses. When rosette leaves were inoculated with tomato-mosaic virus, growth of the irRdR1 lines was severely impaired and plants of both lines rapidly senesced and died. Figure S7. Microarray analysis of significantly (from two replicate arrays hybridized with cDNA from 3 pooled plants of each line) differentially OS-regulated genes in irRdR1 compared to WT plants. +/- from zero represent up- or down-regulation. Table S1. Analysis of Variance (ANOVA) for growth phenotypes of WT and EV ?transformed plants used in this study.

FilenameFormatSizeDescription
TPJ_3030_sm_fig1.tif1612KSupporting info item
TPJ_3030_sm_fig2.TIF2420KSupporting info item
TPJ_3030_sm_fig4.TIF387KSupporting info item
TPJ_3030_sm_fig5a.TIF2735KSupporting info item
TPJ_3030_sm_fig5b_c.TIF3055KSupporting info item
TPJ_3030_sm_fig6.TIF4165KSupporting info item
TPJ_3030_sm_fig7.pdf13KSupporting info item
TPJ_3030_sm_table1.doc27KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.