Abstract
A drawback of recent antiviral therapies based on the transgenic expression of artificial microRNAs (amiRs) is the ease with which viruses generate escape mutations. Here, we show two alternative strategies for improving the effectiveness of resistance in plants. First, we expressed two amiRs complementary to independent targets in the viral genome, and second, we designed amiRs complementary to highly conserved RNA motifs in the viral genome.
TEXT
The transgenic expression of 21-nucleotide (nt)-long artificial microRNAs (amiRs) complementary to viral genomes has been proposed as a new antiviral strategy. Niu et al. (1) used the pre-miRNA159a precursor to engineer amiR159-HCPro, complementary to the HC-Pro cistron of Turnip mosaic virus (TuMV; genus Potyvirus, family Potyviridae). Transgenic expression of this amiR in Arabidopsis thaliana conferred high levels of specific resistance. Similarly, a gene-silencing mechanism (RNA interference [RNAi]) has been used in in vitro assays as an antiviral therapeutic to inhibit the replication of several human viruses (2–4). However, a major issue of these amiR-based antiviral therapies has been the emergence of escape mutant viruses (2, 5–8). These escape variants differ from the wild-type virus by at least one point mutation in the 21-nt target, leading to imperfect matching with the amiR. To evaluate the effectiveness of amiR159-HCPro-mediated resistance in plants, Lafforgue et al. (6) performed an evolution experiment in which multiple independent lineages of TuMV were founded with an ancestral viral clone and allowed to evolve and diversify by serial passages in two different hosts. The first host was a wild-type A. thaliana strain, and the second one was a partially resistant transgenic A. thaliana strain that expressed the amiR at subinhibitory concentrations (line 10-4). Periodically, the evolving populations were used to challenge the resistance of another transgenic A. thaliana line (line 12-4), which was remarkably resistant to the ancestral TuMV. It was found that all lineages, regardless of the plant genotype where they evolved, accumulated mutations in the amiR target and acquired the capacity to successfully infect 12-4 plants. However, viruses evolving in line 10-4 broke resistance approximately seven times faster than viruses evolving in wild-type A. thaliana (6). A subsequent deep-sequencing study of some of the evolved TuMV lineages confirmed that substantial nucleotide diversity was generated at each position of the amiR target along the evolution phase and that the alleles finally breaking the resistance were randomly chosen among the many coexisting ones. The frequency of mutant alleles potentially breaking resistance was significantly higher in virus lineages replicating in the 10-4 line than in wild-type A. thaliana (9). In all cases, resistance-breaking mutations were more common in central positions of the target than in others (e.g., positions 11 and 12) (6). In retrospect, the relatively ease of overcoming the resistance was most likely related to the existence of sequence variations in the 21-nt target of the amiR in TuMV natural populations (G. Lafforgue and S. F. Elena, unpublished results), suggesting that this genomic region is not under strong purifying selection.
The expression of multiple amiRs targeting conserved motifs in viral genomes has also been shown to be a promising solution for both negative (10)- and positive (11)-sense plant RNA viruses, as well as to human retroviruses such as HIV-1 (12), although in the latter case, cross-resistance to the expression of two amiRs was readily shown (13). Here, we have explored two alternative methods to improve the effectiveness of amiR-mediated resistance. In both methods, we tried to minimize the possibility of emergence of escape mutants. Although the generation of an escape mutant is certainly central to the effectiveness of the resistance, factors such as target accessibility (14, 15) and amiR levels (6) should also play an important role. We assume that these additional factors exert a similar effect in all our experiments. The first strategy consisted of combining two amiRs in a single transgenic plant, each one targeting a different sequence in the TuMV genome. The first amiR was the above-mentioned amiR159-HCPro, and the second one was amiR159-TuCP (1), which is complementary to the sequence 5′-GCCAAGAUACGAGCAGAGAGU-3′ within the TuMV CP cistron (positions 8987 to 9007; GenBank accession number AF530055.2). We argued that this approach would expand resistance effectiveness in inverse proportion to the product of the probabilities of generating an escape mutant per amiR target. The mutation rate for the amiR159-HCPro 21-nt target was estimated to be 5.545 × 10−5 per replication event (16); thus, assuming a mutation rate for the amiR159-TuCP target in the same range, the probability of generating the double mutant in a single replication event would be in the range 10−9 to 10−10. The second strategy, which still relied on the expression of a single amiR, was designed to target an RNA sequence under strict purifying selection. In other words, even if produced at the above rate, any mutation in this sequence will render the virus nonviable. To this end, we identified a 29-nt sequence stretch that is strictly conserved among 95 TuMV natural worldwide isolates (full genomes available at GenBank) (Lafforgue and Elena, unpublished). This conserved sequence is in the 3′ end of the CP cistron, which in another potyvirus (Tobacco etch virus) has been shown to be involved in the formation of RNA secondary structures essential for genome replication (17). This new amiR, named amiR159-CPb, is complementary to the sequence 5′-UCGCAUGUAACUUUUCGCAUU-3′ (positions 9323 to 9343; GenBank accession number AF530055.2). New transgenic plants were generated by the floral dip method as described by Niu et al. (1), and all seeds used were homozygous from T4 generations. High and consistent expression of the corresponding amiRs was confirmed by Northern blot analyses.
A large stock of infectious extract was obtained from Nicotiana benthamiana plants inoculated with a plasmid containing TuMV cDNA (6, 18). This amplification step was necessary to overcome the low efficiency of A. thaliana infection by the TuMV cDNA. The extract was obtained by grinding infected tissues in a mortar with liquid N2 and 20 volumes of extraction buffer (50 mM potassium phosphate [pH 7.0], 3% polyethylene glycol 6000). Large numbers of wild-type and 10-4 A. thaliana plants were inoculated with 5 μl buffer containing 10% Carborundum, which was applied to three different leaves by gentle rubbing with a cotton swab soaked in the extract. After inoculation, plants were maintained in a growth chamber (16 h of light at 25°C and 8 h of dark at 24°C). Infected plants from each genotype were pooled and used to prepare two new infectious inocula: wild-type and 10-4. Our previous results showed that amplification in N. benthamiana followed by a single passage in A. thaliana was enough to generate substantial genetic variability in the amiR targets (6, 9). Each infectivity experiment consisted of the inoculation, as described above, of a number of wild-type, amiR159-HCPro (line 12-4), amiR159-HCPro/amiR159-TuCP, and amiR159-CPb A. thaliana plants. Six independent infectivity experiments were performed.
Table 1 shows a summary of the results of the infectivity experiments. The following binomial logistic regression equation was fitted to the infectivity data using generalized linear models: log[Iijkl/(1 − Iijkl)] = β0 + βiSPi + βjTPj + βij(SP × TP)ij + βijkBijk + εijkl, where SP, TP, and B are the genotypes of the plants used as sources of inocula, the genotypes of the plants used as targets (Table 1), and the experimental block (nested within the interaction SP × TP), respectively. Iijkl corresponds to the lth observation of infectivity made in the kth experimental block for inoculum j and in the target plant i. The linear predictor had coefficients β, and each observation is subjected to experimental error εijkl. The model significantly fitted the observed data (omnibus goodness-of-fit test: χ2 = 1247.002, 47 df, P < 0.001). First, we found no significant differences among experimental blocks (χ2 = 20.908, 17 df, P = 0.230), and thus we pooled data from different blocks for making Table 1. Second, the genotype of the target plants had a highly significant overall effect on infectivity (χ2 = 7556.643, 3 df, P < 0.001), meaning that infectivity varies among the different genotypes. Third, we found no overall differences among inocula generated in the wild type (WT) and in 10-4 plants (χ2 = 0.000, 1 df, P = 1.000). However, the significant interaction term (χ2 = 400.617, 3 df, P = 0.028) highlights a subtler effect of inoculum that depends on the genotype of the target plant. Indeed, this effect was due to the higher infectivity of inocula prepared from 10-4 plants for 12-4 plants than inocula prepared from WT plants (see below).
Table 1.
Summary of the infectivity experiments
Genotype of target plant | No.a of plants inoculated or infected with inocula from genotype |
Frequency of infectionb (95% CIc) | |||
---|---|---|---|---|---|
Wild type |
10-4 |
||||
Inoculated | Infected | Inoculated | Infected | ||
Wild-type | 301 | 166 | 311 | 155 | 5.244 × 10−1 (4.849–5.640 × 10−1) |
amiR159-HCPro (12-4) | 522 | 9 | 537 | 26 | 3.393 × 10−2 (2.290–4.496 × 10−2) |
amiR159-HCPro/amiR159-TuCP | 539 | 0 | 475 | 0 | ≤9.843 × 10−4 (0–3.655 × 10−3) |
amiR159-CPb | 535 | 0 | 533 | 0 | ≤9.346 × 10−4 (0–3.471 × 10−3) |
Numbers resulted from adding up the six experimental blocks.
Estimates derived pooling data for both sources of inocula, as no overall difference between them was observed.
Computed using Wald's formula (22).
The last column in Table 1 shows the frequency of infection computed using Laplace's rule-of-succession method (19). This is a Bayesian probabilistic method that takes into consideration a priori knowledge about the possible outcome of any individual trial: success and failure are both likely events, though the probabilities of each outcome are largely unbalanced. The method offers the advantage that for finite sample sizes it does not assign a zero probability to the most unlikely outcome; in this sense, the method provides an upper limit value. TuMV populations showed the largest infectivity in wild-type plants, followed by 12-4 plants. In contrast, the infectivity in the two new transgenic plants was at best 34 to 36 times lower than in 12-4 plants expressing a single amiR targeting a non-strictly conserved RNA sequence. Therefore, we conclude that by combining more than one amiR into the same transgenic plant, or by designing amiRs taking into consideration natural variation of putative target sequences, the likelihood of breaking amiR-mediated resistance by a polymorphic viral population is dramatically reduced. We have previously shown that TuMV populations generated in 10-4 plants contain ca. 100-fold more mutants able to escape from the amiR159-HCPro than those generated in WT plants and thus break resistance more often (9). Indeed, Table 1 confirms this observation, showing that the frequency of infected plants is significantly higher for the 10-4 inocula (Fisher's exact test, P = 0.005). Given this observation, one may think that for this TuMV population the situation would be similar to that in infected plants expressing only the amiR159-TuCP and therefore showing a higher infectivity in the amiR159-HCPro/amiR159-TuCP plants than the TuMV population generated in WT plants. Table 1 shows that this is not the case and that both viral populations behave equally. There are two possible explanations. First, amiR159-TuCP confers a very robust resistance, and second, the amiR159-HCPro escape mutants contained in the population still represents a very minor fraction (less than 10% of that found in the study in reference 9) and thus are not necessarily transmitted during mechanical inoculation (20). An additional consideration is that mutant genotypes escaping amiR159-HCPro may display too low a fitness to face the second amiR159-TuCP amiR.
In our experiments, neither of the two new strategies proved to be more effective than the other against escape mutants; thus, the choice would entirely depend on the availability of strictly conserved RNA sequences in a virus' genome.
ACKNOWLEDGMENTS
We thank Francisca de la Iglesia for excellent technical assistance.
This work was supported by grants RGP12/2008 from the Human Frontier Science Program Organization, PROMETEO2010/019 from the Generalitat Valenciana, and BFU2012-30805 from the Spanish Ministerio de Economía y Competitividad (MINECO) to S.F.E. and by grants 2010TW0015 from CSIC and BIO2011-26741 from MINECO to J.-A.D. F.M. was supported by a fellowship from the Universidad Politécnica de Valencia.
Footnotes
Published ahead of print 22 May 2013
REFERENCES
- 1. Niu Q, Lin SS, Reyes JL, Chen KC, Wu HW, Yeh SD, Chua NH. 2006. Expression of artificial microRNAs in transgenic Arabidopsis thaliana confers virus resistance. Nat. Biotechnol. 24:1420–1428 [DOI] [PubMed] [Google Scholar]
- 2. Boden D, Push O, Lee F, Tucker L, Ramratnam B. 2003. Human immunodeficiency virus type 1 escape from interference. J. Virol. 77:11531–11535 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ge Q, McManus MT, Nguyen T, Shen CH, Sharp PA, Eisen HN, Chen J. 2003. RNA interference of Influenza virus production by directly targeting mRNA for degradation and indirectly inhibiting all viral RNA transcription. Proc. Natl. Acad. Sci. U. S. A. 100:2718–2723 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Krönke J, Kittler R, Buchholz F, Windisch MP, Pietschmann T, Bartenschlager R, Frese M. 2004. Alternative approaches for efficient inhibition of Hepatitis C virus RNA replication by small interfering RNAs. J. Virol. 78:3436–3446 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Gitlin L, Andino R. 2005. Poliovirus escape from RNA interference: short interfering RNA-target recognition and implications for therapeutic approaches. J. Virol. 79:1027–1035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Lafforgue G, Martínez F, Sardanyés J, De la Iglesia F, Niu QW, Lin SS, Solé RV, Chua NH, Daròs JA, Elena SF. 2011. Tempo and mode of plant RNA virus escape from RNA interference-mediated resistance. J. Virol. 85:9686–9695 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Lin SS, Wu HW, Elena SF, Chen KC, Niu QW, Yeh SD, Chen CC, Chua NH. 2009. Molecular evolution of a viral non-coding sequence under the selective pressure of amiRNA-mediated silencing. PLoS Pathog. 5:e1000312. 10.1371/journal.ppat.1000312 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Simón-Mateo C, García JA. 2006. MicroRNA-guided processing impairs Plum pox virus replication, but the virus readily evolves to escape this silencing mechanism. J. Virol. 80:2429–2436 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Martínez F, Lafforgue G, Morelli MJ, González-Candelas F, Chua NH, Daròs JA, Elena SF. 2012. Ultradeep sequencing analysis of population dynamics of virus escape mutants in RNAi-mediated resistant plants. Mol. Biol. Evol. 11:3297–3307 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Kung YJ, Lin SS, Huang YL, Chen TC, Harish SS, Chua NH, Yeh SD. 2012. Multiple artificial microRNAs targeting conserved motifs of the replicase gene confer robust transgenic resistance to negative-sense single-stranded RNA plant virus. Mol. Plant Pathol. 13:303–317 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Fahim M, Millar AA, Wood CC, Larkin PJ. 2012. Resistance to Wheat streak mosaic virus generated by expression of an artificial polycistronic microRNA in wheat. Plant Biotechnol. J. 10:150–163 [DOI] [PubMed] [Google Scholar]
- 12. ter Brake O, Konstantinova P, Ceylan M, Berkhout B. 2006. Silencing of HIV-1 with RNA interference: a multiple shRNA approach. Mol. Ther. 14:883–892 [DOI] [PubMed] [Google Scholar]
- 13. Shah PS, Pham NP, Schaffer DV. 2012. HIV develops indirect cross-resistance to combinatorial RNAi targeting two distinct and spatially distant sites. Mol. Ther. 20:840–848 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Duan CG, Wang CH, Fang RX, Guo HS. 2008. Artificial microRNAs highly accessible to targets confer efficient virus resistance in plants. J. Virol. 82:11084–11095 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Jiang F, Song Y, Han Q, Zhu C, Wen F. 2011. The choice of target site is crucial in artificial miRNA-mediated virus resistance in transgenic Nicotiana tabacum. Physiol. Mol. Plant Pathol. 76:2–8 [Google Scholar]
- 16. De la Iglesia F, Martínez F, Hillung J, Cuevas JM, Gerrish PJ, Daròs JA, Elena SF. 2012. Luria-Delbrück estimation of Turnip mosaic virus mutation rate in vivo. J. Virol. 86:3386–3388 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Haldeman-Cahill R, Daròs JA, Carrington JC. 1998. Secondary structures in the capsid protein coding sequence and 3′ nontranslated region involved in amplification of the Tobacco etch virus genome. J. Virol. 72:4072–4079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Chen CC, Chao CH, Chen CC, Yeh SD, Tsai HT, Chang CA. 2003. Identification of Turnip mosaic virus isolates causing yellow stripe and spot on calla lily. Plant Dis. 87:901–905 [DOI] [PubMed] [Google Scholar]
- 19. Zabell SL. 1989. The rule of succession. Erkenntnis. 31:283–321 [Google Scholar]
- 20. Zwart MP, Daròs JA, Elena SF. 2012. Effect of potyvirus effective population size in inoculated leaves on viral accumulation and the onset of symptoms. J. Virol. 86:9737–9747 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Coburn GA, Cullen BR. 2002. Potent and specific inhibition of Human immunodeficiency virus type 1 replication by RNA interference. J. Virol. 76:9225–9231 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Agresti A, Coull BA. 1998. Approximate is better than “exact” for interval estimation of binomial proportions. Am. Stat. 52:119–126 [Google Scholar]