- Top of page
- Materials and Methods
- Supporting Information
The green peach aphid (GPA), Myzus persicae, is one of the most destructive pests on cultivated crops worldwide (Blackman & Eastop, 2000). GPA causes feeding damage and, more importantly, is the vector of many different plant viruses (Ng & Perry, 2004; Hogenhout et al., 2008). Insect herbivores, including aphids, have often specialized to colonize one or a few related plant species, whereas only a few herbivores, such as GPA, can colonize diverse plant species. Therefore, most plants can defend themselves effectively against the majority of insect herbivores. Moreover, insects are probably required to modulate a variety of plant processes to facilitate colonization. However, the mechanisms by which plants defend themselves against insect colonization and how aphids modulate plant processes are not fully understood.
Aphids possess specialized mouthparts, named stylets, which are developed for the piercing of plant tissues and the ingestion of sap, and allow them to feed from phloem tissue (Tjallingii, 2006). Access to this tissue is gained following extensive probing by the stylets of epidermal and parenchymal cell layers, before the establishment of a successful feeding site in the phloem sieve element (Tjallingii & Esch, 1993). Once established, feeding can be maintained for several hours (Tjallingii, 1995).
In plants, small RNAs (sRNAs) regulate changes in gene expression in response to a variety of biotic and abiotic stimuli (Sunkar & Zhu, 2004; Fujii et al., 2005; Ruiz-Ferrer & Voinnet, 2009; Katiyar-Agarwal & Jin, 2010). It has long been known that components of sRNA pathways play an extensive role in antiviral defence (Ding & Voinnet, 2007). More recently, sRNA pathways have been implicated in resistance to bacteria, fungi, nematodes and insects (Navarro et al., 2006; Pandey & Baldwin, 2007; Hewezi et al., 2008; Pandey et al., 2008; Ellendorff et al., 2009). sRNAs modify gene expression by acting at both the transcriptional and post-transcriptional levels (Voinnet, 2009). RNA-induced silencing is initiated by double-stranded RNA (dsRNA), which can occur as a stem-loop precursor, or a longer dsRNA molecule generated by either bidirectional transcription or the action of an RNA-dependent RNA polymerase (RDR) on a single-stranded RNA (ssRNA) template (Ruiz-Ferrer & Voinnet, 2009). In Arabidopsis, segments of dsRNA are cleaved into 18–24-nucleotide (nt) sRNA duplexes by one or a combination of four Dicer-like (DCL) endoribonucleases. Following methylation of the 2-nt 3′ overhang by the methyltransferase HUA ENHANCER1 (HEN1; Yu et al., 2005), sRNA can be exported from the nucleus before incorporation into an RNA-induced silencing complex (RISC) containing one of 10 Argonaute (AGO) proteins (Vazquez et al., 2010). The sRNA guides the RISC to either cleave or repress the translation of target transcripts bearing sufficient homology to the loaded sRNA.
sRNAs can be divided into subgroups depending on their source and mode of processing (Vazquez et al., 2010). Small interfering RNA (siRNA) is processed from segments of long, perfectly complementary dsRNA, which may be derived from pathogens (e.g. viruses) or generated from loci throughout the genome, but especially from highly repetitive regions (Rabinowicz et al., 2003; Matzke et al., 2007). The latter is consistent with the known role for siRNAs in directing heterochromatic silencing of genomic regions harbouring mobile genetic elements (Matzke et al., 2007). MicroRNAs (miRNAs) are a class of largely 21-nt sRNAs derived from imperfectly complementary stem-loop precursors. miRNAs are excised from their precursors by DCL1 (Park et al., 2002; Kurihara & Watanabe, 2004), although the rate and fidelity of this excision is dependent on the cofactors SERRATE (SE) and HYPONASTIC LEAVES 1 (HYL1; Dong et al., 2008). miRNAs are subject to methylation by HEN1 and are exported from the nucleus via both HASTY (HST)-dependent and independent mechanisms (Park et al., 2005). At some point, there is unravelling of the duplex into its component miR and complementary miR* strands, before one strand is selectively incorporated into RISC. AGO1 is the dominant slicer of the miRNA pathway (Baumberger & Baulcombe, 2005), although a proportion is reported to act through AGO7 or AGO10 (Brodersen et al., 2008; Montgomery et al., 2008).
The miRNA pathway is known to play a significant role in the regulation of the defence response that occurs following challenge by the bacterial biotroph Pseudomonas syringae (Navarro et al., 2006; Zhang et al., 2011) and the pathogen-associated molecular pattern (PAMP) flg22 (Li et al., 2010). The defence pathways activated in response to attack from chewing herbivores are also governed by sRNAs. The growth of Manduca sexta (tobacco hornworm) larvae is enhanced on Nicotiana attenuata lacking RDR1 (Pandey & Baldwin, 2007). In this interaction, RDR1-dependent siRNAs are required to coordinate a defence response involving nicotine biosynthesis and the jasmonic acid (JA) and ethylene (ET) signalling pathways (Pandey et al., 2008).
Aphid infestations elicit transcriptional reprogramming in host plants, despite causing little visible feeding damage (Moran et al., 2002; Couldridge et al., 2007; Kusnierczyk et al., 2007, 2008; Gao et al., 2010). In one study, these changes were more pronounced than those elicited by fungal or bacterial pathogens, or a leaf-chewing lepidopteran pest (De Vos et al., 2005). miRNAs, in particular, are known to target large families of transcription factors. Infestation by several aphid species also results in large-scale changes in the transcription factor profile of infested tissue (Kusnierczyk et al., 2008; Gao et al., 2010; Sattar et al., 2012). Given these observations and the known involvement of sRNAs in defence responses against pathogens and a chewing herbivore, we speculated that sRNAs may play a similarly important role in coordinating the complex and large-scale response to aphids.
GPA effectively colonizes members of the family Brassicaceae, including the model plant Arabidopsis thaliana. Here, we report that Arabidopsis plants deficient in miRNA processing show increased resistance to GPA. This resistance is partly a result of the enhanced production of the phytoalexin camalexin, which is known to play a role in plant defence against bacterial and fungal microbial pathogens. Camalexin is produced at GPA stylet penetration sites, and this plant compound accumulates in aphids fed on plants and an artificial diet containing camalexin. Progeny production is reduced in aphids exposed to camalexin, whereas aphids produce more progeny on plants compromised in camalexin production. Together, this work uncovers a novel role for camalexin in modifying insect reproductive ability.
- Top of page
- Materials and Methods
- Supporting Information
In this study, we have shown that GPA produces significantly less progeny on Arabidopsis plants that aberrantly process miRNAs. Plants unable to process miRNAs respond to aphid infestation with increased induction of PAD3 and production of camalexin. Aphids are more successful on the Arabidopsis pad3 and cyp79b2/cyp79b3 mutants defective in camalexin production. In addition, camalexin is present in the phloem stream and aphids raised on miRNA pathway mutants accumulate more camalexin than aphids raised on control plants. Aphids produce less progeny on artificial diets containing camalexin, indicating that this phytoalexin reduces the reproductive ability of GPA. Finally, aphid fecundity is partially restored for aphids raised on dcl1/pad3 mutants relative to dcl1.
Our finding that aphids were less successful on dcl1 plants was initially unexpected, as pathogen and insect performances have been shown to increase on silencing-deficient hosts (Deleris et al., 2006; Pandey & Baldwin, 2007). Indeed, type III secretion system (T3SS)-deficient P. syringae (which normally reproduces poorly on Arabidopsis) shows increased proliferation on Arabidopsis miRNA pathway mutants, but not on Arabidopsis plants defective in other silencing pathways (Navarro et al., 2008). Similarly, Pseudomonas fluorescens and Escherichia coli, which do not normally infect Arabidopsis, can multiply on Arabidopsis miRNA pathway mutants (Navarro et al., 2008). In addition, some RNA silencing mutants are hypersusceptible to infection by the vascular fungus Verticillium (Ellendorff et al., 2009). More specifically, for insects, an RDR1-silenced line of Nicotiana attenuata (irRdR1) is more susceptible to larvae of the solanaceous specialist Manduca sexta (Pandey & Baldwin, 2007). Nonetheless, there are several examples of increased resistance of Arabidopsis miRNA mutants to pathogens and pests. Both Arabidopsis miRNA and siRNA pathway mutants exhibit increased resistance to the cyst nematode Heterodera schachtii (Hewezi et al., 2008), and dcl1 plants are resistant to tumour formation following stab inoculation with tumorigenic Agrobacterium (Dunoyer et al., 2006). This may be expected, as miRNAs are integral players in plant development, and cyst nematodes and Agrobacterium reprogramme plant development to generate cysts and galls, respectively, which provide feeding and replication sites for these plant colonizers. Thus, our observation that aphids do less well on Arabidopsis miRNA mutants may be a consequence of the highly specialized feeding mode of aphids. GPA does not form noticeable galls, but may still need to modulate specific developmental or basic plant defence processes that are regulated by miRNAs in order to establish long-term feeding sites. The salivary components that aphids release into cells whilst they navigate to the phloem and during phloem feeding (Will et al., 2007; Mutti et al., 2008; De Vos & Jander, 2009; Bos et al., 2010; Pitino & Hogenhout, 2013) may induce these modulations. We propose that the GPA colonization efficiency of Arabidopsis is enhanced by the ability of this aphid to modulate specific plant processes that are regulated by miRNAs.
dcl1 plants display greater resistance to GPA infestation, and our data suggest that this is a result, in part, of the hyperactivation of the camalexin defence pathway. By contrast, this pathway is only modestly induced in aphid-susceptible Col-0 and dcl2/3/4 plants. One possibility is that factors that act as brakes or suppressors of defence hyperactivation in Col-0 or dcl2/3/4 are ineffective or absent in dcl1 plants. Suppressors of hyperactivation may be protein effectors present in aphid saliva that can modify aspects of host physiology and suppress defensive mechanisms. Therefore, host proteins involved in camalexin production or specific miRNAs involved in the management of this pathway may be targets for as yet uncharacterized aphid salivary effectors. Indeed, effectors from a plant pathogen are capable of interfering with host miRNA processing (Navarro et al., 2008). Another possibility is that plants actively manage their response through the induction of specific miRNA species that target transcripts involved in the camalexin pathway. This control mechanism would be largely disabled in dcl1 plants. As large quantities of camalexin are toxic to Arabidopsis cells in culture (Rogers et al., 1996), this dampening effect may represent a form of plant self-defence.
In Arabidopsis, some miRNAs target transcripts related to secondary metabolism. One group of miRNAs (miR160, miR167, miR390, miR393) is specifically related to auxin signalling (Zhang et al., 2011), which is linked to camalexin and glucosinolate biosynthesis. In addition, miR393 has a role in the plant immune response as it is induced following exposure to the PAMP flg22 (Navarro et al., 2006; Li et al., 2010), and following inoculation of both virulent and avirulent strains of P. syringae pv. tomato (Pst; Zhang et al., 2011). It has also been reported that miR393 has a role in resource allocation between the glucosinolate and camalexin pathways (Robert-Seilaniantz et al., 2011).
Aphids transmit one-third of c. 800 described plant viruses (Ng & Perry, 2004; Hogenhout et al., 2008). Many of these viruses encode suppressor molecules which block antiviral RNA silencing (Ding & Voinnet, 2007) and can interfere with the miRNA pathway during infection (Chapman et al., 2004). Silencing suppression is crucial to promote virus infectivity; however, suppression of the miRNA pathway might have a negative impact on the fecundity of the aphid vectors through the mechanisms described here. The relationship between virus and insect will strongly determine the outcome of this tritrophic interaction. Viruses that are acquired rapidly and transmitted by aphids will benefit from plant behaviour that discourages aphid settling (Mauck et al., 2010). By contrast, viruses that require longer acquisition times, such as those that are phloem limited, may act to extend aphid feeding time at a particular feeding site (Eigenbrode et al., 2002).
Our qRT-PCR assays indicated that aphid-resistant dcl1 plants increase transcription of an ET-responsive gene relative to susceptible Col-0 and dcl2/3/4 plants following aphid colonization. Fecundity assays confirmed the involvement of ET signalling in resistance, as aphid performance was improved significantly on ein2 mutants. Our result, showing no change in aphid fecundity on etr1, is consistent with previous studies in which the performances of GPA and Brevicoryne brassicae were either unaffected or reduced on etr1 mutants (Mewis et al., 2005, 2006). Other laboratories have demonstrated that saliva-induced aphid resistance is independent of EIN2 and ET signalling (De Vos & Jander, 2009), whereas EIN2 is known to be critical for resistance to GPA following treatment with the bacterial protein harpin (Dong et al., 2004; Liu et al., 2011). It remains a possibility that altered regulation of this signalling mechanism contributes to the dcl1 resistance phenotype.
Aphid fecundity was increased on the pad3 and cyp79b2/cyp79b3 mutants relative to Col-0. By contrast, aphid performance was unchanged on the cyp81f2 mutant. Taken together, these results indicate that, under our experimental conditions, the production of camalexin is a major resistance factor. This is in contrast with the observations of Pegadaraju et al. (2005), who found no statistically significant increase in GPA colonization ability on pad3 mutants. In addition, Kim et al. (2008) found no change in fecundity of aphids raised on cyp79b2/cyp79b3 mutants relative to wild-type plants. However, in both cases, nonaged aphids were exposed to the mutant plants for a relatively short period, that is 2–5 d, whereas, in the experiments reported herein, the nymphs were born on the mutant plants and reared on these plants to adulthood (c. 16 d), during the course of which they began to produce nymphs themselves. Thus, differences in the experimental procedures may account for the different outcomes. Indeed, the dcl1 resistance phenotype was absent when experiments were carried out following a previously published protocol (Pegadaraju et al., 2005; Fig. S7). It is also possible that the aphid colonies maintained by different laboratories have varying susceptibilities to different phytochemicals. Our results are in agreement with those of Kusnierczyk et al. (2008), who found that B. brassicae (cabbage aphid) is more successful on pad3 relative to wild-type Arabidopsis when both plants are pretreated with UV light to induce camalexin production. In these experiments, aged nymphs were raised on test plants for 13 d, a protocol very similar to our own assay. Furthermore, aphids produce less progeny on artificial diets containing camalexin compared with control diets, confirming that camalexin has a negative impact on GPA performance. This indicates an unsuspected depth to camalexin function beyond antifungal and antibacterial defence. This work also highlights the extensive role of the miRNA-mediated regulation of secondary metabolic defence pathways with relevance to resistance against an aphid pest.