Background

Jasmonates, including jasmonic acid (JA) and the biologically active intermediates and derivatives of the JA biosynthetic pathway, are powerful regulators of plant development and inducible resistance. By mediating signal transduction they influence changes in expression profiles of a wide range of genes involved in plant defence [1]. Induction of JA-related response has often been linked to tissue damage, and the important roles of JA signalling in defence against bacterial and fungal infections or caterpillar attack are well documented (for reviews [2–4]). More recent research, however, provides evidence for the activation of JA-mediated defence upon attack by phloem-sucking insects, such as aphids and silverleaf whitefly nymphs, which try to avoid tissue damage during feeding [5–10]. Phloem feeders possess stylet-like mouthparts, which they use to ingest phloem sap. During penetration of plant tissue the stylet is manoeuvred through plant tissue until it is finally anchored in a sieve tube element. Here it can stay for several hours or even days, facilitating a continuous sap supply. By avoiding extensive tissue wounding, aphids minimize the risk of inducing defence responses in the attacked plant while depriving it of assimilates. In the case of a massive infestation, the loss of nutrients interferes with plant growth and development, and may eventually lead to plant death. Constitutive or transient activation of JA-related responses is known to enhance a plant's resistance to phloem feeders, including aphids [11–13].

JA is biosynthesized from polyunsaturated fatty acids released from chloroplast membranes via a series of enzymatic reactions usually referred to as the octadecanoid pathway. In pathogen-free laboratory conditions, a non-functional JA pathway does not result in any disturbance in normal vegetative growth. In a more natural environment, however, mutant plants that do not synthesize JA are more susceptible to pathogen attack because they fail to activate JA-dependent defences [45]. However, their involvement in the activation of plant defence has not been assessed yet. Strong up-regulation of these genes in wt plants attacked by B. brassicae suggests that they play an important role in defence against aphids. The regulatory function of BTB and TAZ domain containing proteins has not been established yet, but BTB and TAZ domain protein (BT2) have been identified as essential components of the TELOMERASE ACTIVATOR1 (TAC1)-mediated telomerase activation pathway [46]. Telomerase activity is high in plants in rapidly dividing cells and reproductive organs. The induction of BT2 and BT5 in the non-challenged aos plants suggests that these genes are under negative regulation of JA. All five BTB and TAZ proteins (BT1-BT5) are known to be readily induced by H2O2 and SA treatments [47].

The glutaredoxin family protein GRX480, whose induction was eliminated in the infested aos plants, was recently identified as a regulator of JA/SA cross talk. It interacts with TGA transcription factors to antagonize expression of JA-responsive genes in an NPR1-dependent manner [48]. Our results indicate that the induction of GRX480 upon B. brassicae attack is dependent on JA levels.

The expression of EDS5 in both non-challenged and aphid-attacked plants shows that JA levels also influence it. This is in contrast to previous reports, which describe solitary SA signalling based regulation of the EDS5 gene [49]. Our results suggest that regulation of EDS5 is more complex than previously thought.

Additional signals are involved in regulation of the response to B. brassicae infestation

Some genes, whose expression in non-challenged plants was clearly dependent on JA responded to infestation in the aos mutant despite the lack of JA-derived signals, even though their induction was not as extensive as the induction observed in wt plants. This indicates that, in addition to JA, some other signalling mechanisms are involved in the regulation of these transcripts upon B. brassicae infestation. It is well established that the activation of invader-specific responses in plants attacked by insects is mediated by cross-talk between different signalling pathways [38]. In the case of insect infestation, in addition to JA, phytohormones such as salicylic acid (SA), ethylene (ET) and abscisic acid (ABA) play major roles in coordinating the induction of appropriate defences [26, 50]. Thus SA, ET or ABA are likely regulators of the defence responses in the absence of JA for genes such as trypsin inhibitors (ATTI1 and At1g73260), TAT3, CYP79B2, PR4 or ASA1.

Induction of JAZ repressors desensitizes fou2 response to B. brassicae attack

The transcriptional profile of the non-challenged fou2 genotype mimics the profile of wt plants that manifest induced defence [33]. In our studies many of the genes that have been shown to be involved in the response to aphid attack in wt plants were up-regulated in the non-challenged fou2 mutant, often showing similar or stronger intensity of changes compared to attacked wt plants (Table 1 and Additional file 7 Table S5). A similar induction of transcription factors and defence-related genes was observed by Bonaventure and co-workers [33]. However, in contrast to the previously observed reaction of fou2 to wounding [17], further induction of these transcripts upon infestation was much weaker than observed in wt plants. A similar lack of stress responses resulting from prolonged high endogenous JA levels was observed in potato plants subjected to wounding and water stress. Although several of the genes involved in JA biosynthesis are induced by JA thereby creating a positive feedback loop [51], there exists also a negative regulatory feedback loop protecting the plants from the adverse effects of their own defence. The constitutive up-regulation of the JA synthesis pathway in the fou2 mutant probably triggers this negative feedback loop, leading to desensitization of processes involved in the activation of the aphid-induced defence. JAZ family proteins act to repress transcription of JA-inducible genes and thus modulate JA-mediated plant responses [52, 53]. The high induction of several JAZ genes in the fou2 mutant (Additional file 3 Table S1) indicates activation of the desensitization mechanism and may explain the reduced responsiveness of fou2 plants challenged with B. brassicae. The negative regulation of JA responses is delayed and takes effect some time after the proceeding induction [45]. The hyper activation of JA biosynthesis genes in fou2 plants shortly after mechanical wounding that was observed by Bonaventure and co-workers [17] was not observed by us after 72 h of sustained B. brassicae infestation. This might be due to a stealthy manner of aphid feeding that causes only minimal tissue damage. The induction of the wound-specific JA responses in aphid-infested plants is therefore much weaker than in mechanically wounded plants. In addition, the high level of JAZ repressors may also tune the JA-regulated transcriptional changes in the aphid-attacked fou2 plants after 72 h.

Aphid fitness is comparable on wt and aos genotypes but reduced on fou2

Despite the reduced responsiveness of a wide range of defence-linked genes in the aos mutant, we did not observe any improvement in aphid fitness in comparison to wt plants. This may seem surprising as JA signalling seems to be important for plant defence mechanisms induced upon infestation. In contrast to our results, Ellis and co-workers observed increased growth of green peach aphid (Myzus persicae) populations on the coi1-16 mutant that had defects in JA signalling [13]. However the coi1-16 line carries an additional mutation that might have influenced M. persicae responses observed by Ellis and co-workers. This mutation lies in the PENETRATION2 (PEN2) gene encoding a glycoside hydrolase and renders the PEN2 protein with highly reduced stability [54]. PEN2 is required for indole glucosinolate-dependent pathogen-induced callose deposition [55]. As accumulation of callose is one of the defence mechanisms against aphid infestation [7], the pen2-4 mutation, present in coi1-16 line, may contribute to the increased susceptibility of coi1-16 plants to infestation with M. persicae.

It is also conceivable that the expressional changes of JA-regulated genes observed by us in the aphid-infested aos mutant were sufficient to sustain the same level of aphid resistance/susceptibility as is present in wt plants. It should be noted that many genes known to be regulated by SA, ABA or auxin signalling were up-regulated in aos plants. Several of these can be involved in defence against B. brassicae infestation and influence aphid fitness.

As revealed by the insect fitness tests, physiological changes resulting from the fou2 mutation render plants more resistant to infestation than wt, despite the reduced intensity of the aphid-induced responses. As the observed resistance was not based on feeding deterrence, it is most probably based on antibiosis. Various defence-related responses that are constitutively activated in fou2 plants, e.g. high expression of plant defensin proteins (PDFs), pathogenesis-related proteins (PR) or protease inhibitors, can exhibit an antibiotic effect on insect pests. The latter, for example, can disturb digestion and absorption of food in the insect gut [27]. Moreover, the high activity of LOX enzyme in fou2 plants can increase production of reactive lipid peroxides, cause oxidative damage to the insect gut and significantly decrease the nutritive quality of dietary proteins [56]. It should be noted, however, that the mechanism responsible for the manifestation of the fou2 phenotype is not fully understood. Therefore, we cannot eliminate the possibility that other, unknown, features of fou2 could play a role in mediating aphid resistance.

Conclusions

A comparison of transcriptional profiles of non-challenged aos, fou2 and wt plants allowed us to identify more than 200 genes whose expression profiles in non-challenged plants were dependent on endogenous jasmonate status. Most of these transcripts were up-regulated in fou2 and down-regulated in aos mutants, which points to a positive regulatory function of JA-derived compounds. Many of the jasmonate-dependent genes were connected to regulation of transcription, defence responses, redox balance and cell wall modification.

Upon infestation with Brevicoryne brassicae, the responsiveness of many genes was changed in aos and fou2 plants. Genes attributed to GO categories connected to the regulation of transcription and responses to stress were generally less induced in both mutants. In contrast, transcripts classified as involved in cell division and development, cell wall modification and transport were more induced or not as much down-regulated in the mutants compared to wt. The observed changes in aphid-mediated responsiveness of aos had, however, no noticeable impact on aphid fitness. This may indicate that the induced responses, although weaker than in wt, were strong enough to keep the same level of resistance. Alternatively, responses were mainly induced locally, so that the aphids could benefit from frequent changes of feeding places. In the fou2 mutant, several genes involved in defence against B. brassicae were induced in non-challenged plants. As a consequence, the transcriptional profile of non-challenged fou2 resembled the aphid-induced profile of wt. Although additional B. brassicae mediated regulation of already induced genes was limited, the aphids' reproduction rate was negatively influenced by the fou2 mutation. As an array of defensive responses is constitutively activated in fou2 plants, the feeding aphids could not move to a leaf area where the response was not induced, as they could in the case of wt plants.

Our results indicate that JA-regulated responses are important in defining susceptibility of a plant to infestation with aphids. As shown in this study, JA-derived compounds are powerful regulators of a range of defensive responses exhibited by plants attacked by aphids.

Methods

Plant material

The Arabidopsis thaliana Columbia-0 ecotype (Col-0) single seeds line used in the experiment has been derived from seeds produced by Lehle Seeds (Round Rock, USA; Catalogue No. WT-2-8, Seed Lot No. GH195-1). The aos mutant was the one described in [15]. The fou2 mutant was kindly donated by Prof. Edward Farmer (University of Lausanne, Switzerland). Both mutants are in Col-0 background. Seeds were sterilized according to standard procedures and plants were initially grown aseptically on agar medium containing MS basal salt mixture (Sigma), 3% (v/w) sucrose, and 0.7% (v/w) agar (pH 5.7) to assure uniform germination. After 15 days, seedlings were moved to 6 cm diameter pots (3 seedlings per pot) filled with a sterile soil mix (1.0 part soil, and 0.5 part horticultural perlite). Plants were kept in growth chambers Vötsch VB 1514 (Vötsch Industrietechnik GmbH, Germany) under the following conditions: a 8/16 h (light/dark) photoperiod at 22°C/18°C, 40%/70% relative humidity, and 70/0 μmol m-2s-1 light intensity. A short time day was applied to prevent plants from bolting. For aphid fitness experiments, plants were sown directly to pots with soil (one plant per pot) and kept in chambers under a 16/8 h (light/dark) photoperiod.

Insects

Brevicoryne brassicae was reared on Brassica napus or Brassica oleracea plants in a growth chamber with a 16/8 h (light/dark) photoperiod at 22°C/18°C, 40%/70% relative humidity, and 70/0 μmol m-2s-1 light intensity.

Infestation experiments

Thirty-two-day-old plants (17 days after transferring to soil) had 8 fully developed leaves. Each plant was infested with 32 wingless aphids (4 per leaf), which were transferred to leaves with a fine paintbrush. Infested plants and aphid-free controls were kept in plexiglass cylinders as described in [9]. Plants were harvested 72 h after infestation between the 6th and 8th hour of the light photoperiod. Four biological replicates were run, each sampled from 15 individual plants. Whole rosettes were cut at the hypocotyls and aphids were removed by washing with Milli-Q-filtered water. Harvested material was immediately frozen in liquid nitrogen.

RNA isolation, cDNA synthesis and microarray experiments

All procedures were done as described in [7]. Custom-designed, full genome Arabidopsis oligonucleotide microarrays printed at the Norwegian Microarray Consortium (Trondheim, Norway) were used in all experiments.

Quantitative real-time PCR

For qRT-PCR analysis, the total RNA was DNAse treated using DNA-free™ Kit (Applied Biosystems), while the QuantiTect® kit (QIAGEN) was used for cDNA synthesis. A LightCycler 480 System and the corresponding SYBR Green I Master mix (Roche Diagnostics GmbH) were used in a three-step programme including (1) preincubation at 95°C for 5 min; (2) 40 cycles of amplification consisting of 95°C for 10 s, 55°C or 60°C for 10 s and 72°C for 10 s; and (3) melting curve analysis by heating from 65°C to 97°C with a ramp rate of 2.2°C/s. Each 20 μl reaction contained 0.5 μM of each forward and reverse primer (for gene-specific primer sequences used in qRT-PCR, see Additional file 9 Table S7), and cDNA quantity corresponding to 50 ng of RNA. LinRegPCR software [57] was used to determine the PCR reaction efficiency for each sample and the efficiencies for each primer set were calculated by averaging the efficiency values obtained from the individual samples. Relative expression ratios of the targeted genes were calculated and normalized to TIP41-like gene (At4g34270) [58] with the use of REST 2008 software [59]. The qRT-PCR analysis was performed with the use of three biological replicates.

Statistical analysis of microarray data

The microarray experiment was a 2-by-3 factorial, with the factors as plant type (wt, aos mutant or fou2 mutant) and treatment (infested or not infested). Each experimental condition, i.e. each combination of factors, was represented by four biological replicates. Seven different direct comparisons of the experimental conditions, using four replicates (each representing 15 individual plants) for each comparison, were made with the use of microarray data sets. However, only data from microarrays with very good technical quality were used for further analyses. (Figure 1 shows the direct comparisons that were made and the comparisons for which only three replicates were of good enough technical quality). Note that using this setup means that the same biological replicate will occur on two different microarrays. Also note that experimental conditions that were not compared directly can still be contrasted, but with lower efficiency than the direct comparisons.

The microarray data for each array were filtered and normalized as discussed in [7]. To make statistical inferences about differential regulation between experimental conditions, the limma package [60] was used. In each comparison of experimental conditions a q-value [61] was calculated for each gene. For a gene to be considered differentially regulated at a statistically significant level, its q-value had to be lower than 0.05. In effect this controlled the false discovery rate (FDR) [62–64] of the comparison at a 0.05 level.

Aphid fitness experiments

B. brassicae fitness on aos and fou2 mutants in comparison to wt Col-0 was evaluated in experiments assessing aphid asexual fecundity. Two first instar nymphs were placed on each plant and plants were placed in plexiglass cages (3 plants per cage). Eleven cages (33 plants) were used for each genotype tested. After 13 days, aphid progeny numbers in each cage were counted. To compare aphid counts for the different plant types, a two-tailed Wilcoxon rank sum test was used with a significance level of 0.05.

Electrical Penetration Graph

The EPG technique was used to monitor aphid feeding behaviour [65]. An eight-channel GIGA-8 direct current amplifier (Wageningen University, The Netherlands) was used for simultaneous recordings of eight individual wingless Brevicoryne brassicae aphids feeding on eight plants (4 wt plants and 4 fou2 mutants). The aphids originated from a colony kindly donated by Prof. Gary Thompson (Oklahoma State University) propagated on Brassica oleracea plants. Before the start of an experiment, the aphids were starved for 4 h and immediately before wiring, an individual aphid's dorsum was cleaned of wax with the help of a paintbrush hair, and a thin gold wire (12.7 μm diameter, 2-4 cm long) was glued to the dorsum with silver paint (Ted Pella). The other end of the wire was connected to an EPG probe and an output wire from the EPG monitor was inserted into the soil in which the plant was rooted. Plants used in EPG experiments were 3 to 4 weeks old, and did not reach the bolting stage. During experiments plants and insets were kept inside a Faraday cage at constant light conditions and 22°C. The waveform recordings were analysed using the EPG analysis software PROBE 3.0 (W.F. Tjallingii, Wageningen University, The Netherlands). The experiments were repeated several times to obtain a total of 30 biological replicates for fou2 and 34 for wt. A Wilcoxon rank sum test was used to compare the amount of time B. brassiae spent on different feeding behaviours as measured with EPG.