Treating seeds with activators of plant defence generates long-lasting priming of resistance to pests and pathogens

Authors


Author for correspondence:
Michael R. Roberts
Tel: +44 1524 510210
Email: m.r.roberts@lancaster.ac.uk

Summary

  • Priming of defence is a strategy employed by plants exposed to stress to enhance resistance against future stress episodes with minimal associated costs on growth. Here, we test the hypothesis that application of priming agents to seeds can result in plants with primed defences.
  • We measured resistance to arthropod herbivores and disease in tomato (Solanum lycopersicum) plants grown from seed treated with jasmonic acid (JA) and/or β-aminobutryric acid (BABA).
  • Plants grown from JA-treated seed showed increased resistance against herbivory by spider mites, caterpillars and aphids, and against the necrotrophic fungal pathogen, Botrytis cinerea. BABA seed treatment provided primed defence against powdery mildew disease caused by the biotrophic fungal pathogen, Oidium neolycopersici. Priming responses were long-lasting, with significant increases in resistance sustained in plants grown from treated seed for at least 8 wk, and were associated with enhanced defence gene expression during pathogen attack. There was no significant antagonism between different forms of defence in plants grown from seeds treated with a combination of JA and BABA.
  • Long-term defence priming by seed treatments was not accompanied by reductions in growth, and may therefore be suitable for commercial exploitation.

Introduction

When under attack from pests and pathogens, plants are able to mount an array of inducible defence responses, ranging from the rapid synthesis of toxic metabolites and defensive proteins to longer-term morphological changes, such as increases in trichome density. Such induced defences are generally recognised to impose a resource cost on the plant, manifest as reduced growth and reproductive fitness (Cipollini et al., 2003). Attempts to exploit such induced resistance responses via the application of synthetic chemicals that activate defence signalling pathways, such as benzothiadiazoles, have met with rather limited success to date, perhaps in part because these benefits are constrained by the inherent costs of defence (Heil et al., 2000). Besides direct competition for resources (allocation costs), other costs associated with direct activation of defences come in the form of tradeoffs between different forms of defence (Walters & Heil, 2007). The different threats posed by different attackers, such as insect herbivores and biotrophic and necrotrophic pathogens, require different strategies to combat them, and these are regulated by different signalling pathways controlled by phytohormones such as jasmonic acid (JA), salicylic acid (SA) and ethylene. Antagonistic interactions between such defences are well documented, such as the reciprocal negative interactions between JA- and SA-dependent signalling (Kunkel & Brooks, 2002; Glazebrook, 2005). These costs and tradeoffs mean that direct activation of defence provides maximum benefit only under specific circumstances, usually when there is a high threat from a single class of attacker.

As well as direct activation of defence, herbivore and pathogen attacks also result in long-term sensitization of plant-inducible defences to future biotic stress, a phenomenon commonly known as priming. Priming represents a state in which defences are not expressed, but in which the plant is able to respond more rapidly and/or more strongly to attack than other plants which have not experienced previous stress (Conrath et al., 2006). Priming generally provides broad-spectrum enhanced resistance, but with minimal associated costs compared with direct activation of defence (van Hulten et al., 2006). Recent evidence suggests that priming can be an effective mechanism for crop protection in the field (Beckers & Conrath, 2007; Walters et al., 2008). The mechanisms underlying priming are poorly understood, but have been suggested to include increased expression of signalling proteins and transcription factors involved in inducible defence (Bruce et al., 2007; van der Ent et al., 2009), as well as epigenetic changes in defence genes which, following transient activation by an initial stress exposure, switch them into a primed state, poised for a rapid transcriptional response to subsequent attack (Bruce et al., 2007; van den Burg & Takken, 2009).

The establishment of fully activated and primed defence is often spatially and temporally separated in the plant. Infection of a single leaf with a microbial pathogen, for example, can lead to direct activation of defence in the infected leaf and primed defence in other parts of the plant, via the process known as systemic acquired resistance (SAR). Likewise, herbivory activates defences in both local and systemic leaves, and can also prime future direct and indirect defences (Kessler et al., 2006; Frost et al., 2007; Ton et al., 2007). The establishment of primed defence in systemic tissues implies long-range signalling. A number of hormonal and other chemical signals have been identified which can be transported either in the vascular system, or in the atmosphere, to elicit the primed state in receiving tissues (Heil & Ton, 2008; Pieterse et al., 2009). Exogenous application of such compounds, along with some synthetic chemicals, can also activate priming responses (Conrath et al., 2006). As well as influencing future responses in tissues distinct from the site of attack via priming, evidence is accumulating to suggest that the effects of biotic stress on future defence responses can extend to tissues not present at the time of stress perception, and even to subsequent generations (Agrawal et al., 1999; Holeski, 2007; Boyko et al., 2010; Kathiria et al., 2010). Whilst the mechanisms underlying such responses are not yet understood, (Boyko & Kovalchuk, 2011), priming may be part of a phenomenon that can provide long-term adaptive benefits to plants and their offspring. The activation of primed resistance by chemical treatments may therefore provide a simple way of providing crop plants with long-term improvements in stress resistance with minimal impact on productivity. Here, we show that treatment of tomato seeds with JA or beta-aminobutyric acid, (BABA) provides long-lasting increases in herbivore and pathogen resistance in plants grown from them.

Materials and Methods

Plants and seed treatments

Unless otherwise stated, tomato plants (Solanum lycopersicum L., cv Carousel or Money Maker) were germinated and grown in a peat-based compost mixture (Scotts M3) and cultivated in a heated, passively ventilated glasshouse (min 18 ± 2°C, max 25 ± 3°C) with supplementary lighting (Osram greenpower 600 W high-pressure sodium lamps) to a minimum 250 ± 25 μmol m−2 s−1 photosynthetically active radiation at the canopy. A minimum 16 h photoperiod was maintained. Seeds were sown individually in 8-cm-square form pots and watered daily from below to water-holding capacity with excess water being removed from the flood trays. For mildew experiments, plants were re-potted to 13-cm-square form pots. For seed treatments, 20–40 seeds were incubated in the dark at 4°C in aqueous solutions of 3 mM JA (from a 1.2 M stock in ethanol) or 18 mM BABA. Control treatments included 0.25% ethanol where appropriate. Following incubation, seeds were washed twice for 10 min in distilled water before sowing.

Herbivore bioassays

Cultures of red spider mite (Tetranychus urticae Koch) were maintained on tomato plants before the experiments. Adult mites were collected from stock plants and released onto leaves (10 mites per leaf) of tomato plants grown in glasshouses under natural light at Stockbridge Technology Centre (UK). Mites were allowed to feed for 9 d. The extent of plant-mediated resistance to T. urticae was measured by counting, using a binocular microscope, the number of living and dead mites on each plant and the number of mite eggs present. Eggs of tobacco hornworm caterpillars (Manduca sexta L.) were obtained from Chris Apark, University of Bath (UK). Eggs were hatched and caterpillars raised on an artificial diet under controlled environments. Two third-instar larvae were placed on the fifth leaf of 8-wk-old plants and allowed to feed for 4 d before grazed leaf area was measured. Stocks of green peach aphid (Myzus persicae (Sulzer)) were maintained on lettuce plants in a glasshouse. For experiments, five aphids were introduced on to a single leaf and allowed to feed on the plants for 12 d, with population counts conducted at regular intervals.

Botrytis cinerea bioassays

Botrytis cinerea R16 (Faretra & Pollastro, 1991), kindly provided by Monica Höfte (Ghent University), was cultured on potato dextrose agar plates in a plant growth chamber (Percival Scientific, Perry, Iowa, USA) set at 22 ± 1°C with a 10 h light cycle provided by Osram fluora lamps delivering 100 ± 20 μmol m−2 s−1. Conidia were isolated as described in Asselbergh et al. (2007) and resuspended in sterile reverse osmosis water before adding to the inoculation solution (50 mM glucose, 33.5 mM KH2PO4, pH 5) at a final concentration of 106 ml−1. Conidia were pre-germinated for 2.5 h at room temperature. An excised leaf assay was performed (Audenaert et al., 2002) using leaf three or four from a 4-wk-old tomato (cv Carousel) plant and two 5 μl droplets per leaflet. After incubation for 72 h at 22°C in the dark, the infected leaves were imaged and recorded using SPOT Basic imaging software. Lesion diameter was measured using calibrated Screen Calipers (Iconico Inc. http://www.iconico.com). Measurements were converted to area by assuming lesions were circular.

Powdery mildew bioassays

A stock culture of tomato plants (cv Money Maker) heavily infected with Oidium neolycopersici was maintained in the glasshouse. For experimental work, spores were harvested from stock plants by washing four to five leaflets harbouring recently sporulated fungus into a collection tube with the inert solvent Fluorinert® (Pefluor compound FC-43, Apollo Scientific Ltd, Stockport, Cheshire, UK) at a final volume of approx. 5 ml. A suspension of 5 × 104 spores ml−1 in Fluorinert® were sprayed onto the apical and first pair of leaflets on leaves four and five of 4-wk-old tomato (cv Carousel) plants using a modified airbrush (Badger Air-brush Co, Franklin Park, Illinois, USA). Six plants for each treatment were sprayed with 0.75 ml per plant. Inoculated plants were incubated in a glasshouse on a capillary mat flood bench to maintain high humidity around the plant and encourage disease establishment. After 10 d, images of fungal growth were collected by high-resolution charge-coupled device camera through Spot Basic imaging software (SPOT Imaging Solutions, Sterling Heights, Michigan, USA). Additionally, the number of distinct disease colonies for each leaflet was counted by eye. Leaf area was derived using the area selection tool in Image-Pro PLUS (Media Cybernetics Inc, Bethesda, Maryland, USA).

Quantitative reverse transcription polymerase chain reaction (RT-PCR)

Tomato plants (cv Carousel) were infected with B. cinerea as described earlier. Infected leaflets were frozen in liquid nitrogen at 0, 4, 8, 12, 16 and 24 h following application of spores. Plant material was ground to a fine powder using a mortar and pestle and liquid nitrogen. RNA was extracted using a hot phenol method essentially as described by Verwoerd et al. (1989), scaled up accordingly. RNA was purified using Qiagen RNeasy spin columns. First-strand cDNA was synthesized from 4 μg total RNA using reverse transcriptase and primer OG1 (Table 1). PCR was performed in an ABI Prism 7000 cycler (Applied Biosystems, Warrington, Cheshire, UK) using EvaGreen qPCR master mix (qARTA•BIO Inc, Fremont, CA, USA), cDNA corresponding to 40 ng of total RNA and 0.2 μM of each primer in a 25 μl reaction at 95°C for 15 min, followed by 40 two-step cycles at 95°C for 15 s and 60°C for 1 min. All gene-specific primers have been described previously (Table 1). Relative expression levels at each time point were calculated from cycle threshold (CT) values according to the ΔCT method (Applied Biosystems User Bulletin #2) using the tomato Ubi3 gene as a reference.

Table 1.   Oligonucleotide primers used for cDNA synthesis and quantitative real-time PCR
 Forward primerReverse primerReference
OG1 GAGAGAGGATCCTCGAGT(15)Worrall et al. (1998)
AOS2TCTCTTCCTCTTCCTTCTCTTCACCCGCCGGGTATAGTCCTGGTAGATAFowler et al. (2009)
Ubi3ACTCTTGCCGACTACAACATCCCTCCTTACGAAGCCTCTGAACCFowler et al. (2009)
PR1b1CCGTGCAATTGTGGGTGTCGAGTTGCGCCAGACTACTTGAGTFlors et al. (2007)
Pin IIGGATTTAGCGGACTTCCTTCTGATGCCAAGGCTTGTACTAGAGAATGPeiffer et al. (2009)

Results

Seed treatment with JA enhances herbivore and disease resistance

To determine whether plants might respond to priming agents at the seed stage, seeds of tomato plants (Solanum lycopersicum cv Carousel) were soaked in a 3 mM solution of JA before germination. To examine the effect of the seed treatment on JA-dependent herbivore resistance responses, 7- to 10-wk-old plants grown from control and JA-treated seed were challenged with different pest species, as detailed in the ‘Materials and Methods’ section. The red spider mite (T. urticae) is an important commercial pest of tomato. One week following introduction of spider mites, the visual damage caused by feeding activity was noticeably lower in JA seed-treated plants than in controls, and both the populations (Fig. 1a) and reproductive rate (Fig. 1b) of mites measured 9 d after infestation were significantly reduced by JA seed treatment. Feeding of tobacco hornworm (Manduca sexta) caterpillars was also reduced in JA seed-treated plants (Fig. 1c), and populations of the green peach aphid (M. persicae) were significantly lower on plants grown from treated seed than on controls (Fig. 1d). We also assessed the effect of the seed treatment on JA-dependent disease resistance responses using a bioassay for infection by the necrotrophic fungal pathogen, B. cinerea. Measurements of lesion areas following Botrytis inoculation showed that plants grown from JA-treated seed are significantly more resistant to disease (Fig. 1e, Supporting Information, Table S1).

Figure 1.

Seed treatment with jasmonic acid (JA) enhances plant resistance to herbivory and disease. (a, b) Tomato (Solanum lycopersicum) plants grown from control or 3 mM JA-treated seed were grown in a glasshouse and 10 adult female red spider mites (Tetranychus urticae) introduced onto a single leaflet of each plant when they were 7 wk old. Mite performance was measured by counting live mites (a) and eggs laid (b) 9 d later. Data show means + SEM (= 10). Differences were significant by a Student’s t-test (= 0.0012 for survival and < 0.0001 for egg laying). (c) Leaf area measured after Manduca sexta caterpillar grazing. Data show means + SEM (= 10). Differences were significant by a Student’s t-test (= 0.011). (d) Aphid (Myzus persicae) population on leaflets of tomato plants grown from JA-treated seed as a percentage of the population on control plants. (e) Areas of disease lesions on leaves of 4-wk-old tomato plants 3 d following inoculation with Botrytis cinerea. Data show means + SEM (= 8 leaves with 10 inoculations per leaf). Populations are significantly different by a Mann–Whitney U-test (< 0.0001).

JA seed treatment has minimal impact on growth and development

Direct activation of plant defence is commonly associated with reduced growth, an ecological cost which is minimized by priming of defences (van Hulten et al., 2006). To determine the impact of the JA seed treatment on growth and development, a range of traits were measured. We observed a delay in germination of seed treated with 1–5 mM JA of approx. 1 d compared with control treated seeds (data not shown), but final germination percentage was not significantly altered (Fig. 2a). At JA concentrations above 10 mM, however, inhibition of germination became significant (Fig. 2a). Furthermore, 3 mM JA seed treatment caused a reduction in growth of the primary root relative to controls in seedlings grown axenically on agar (Fig. 2b). This is consistent with the known role of JA in root growth regulation (Wasternack, 2007). However, over the longer term, we observed no differences in growth and development between plants grown from control and those grown from JA-treated seed. Examples of characteristics determined include plant height (Fig. 2c), and fruit DW (Fig. 2d).

Figure 2.

Jasmonic acid (JA) seed treatment has minimal impact on growth and development. (a) Germination of seeds of tomato (Solanum lycopersicum) following treatment with JA over a range of concentrations. The illustrated data are means of 10 replicate populations ± SEM, and the line is a fitted dose response. The dashed lines represent the 95% confidence interval. (b) JA inhibition of primary root growth of seedlings grown for 10 d on Murashige and Skoog (MS) agar. Data show mean root length ± SEM (= 20). Means are significantly different by Student’s t-test (< 0.0001). There were no significant effects on the vegetative or reproductive growth of tomato grown from seed treated with 3 mM JA. Plots show plant height (c) and fruit DW (d) measurements taken at successive weekly harvests. Data are means of eight replicate plants ± SEM, and the lines are fitted growth-response curves. Open circles, control; closed circles, JA. Neither the fitted growth response nor two-way ANOVA (treatment × harvest) showed any significant difference between control plants and those grown from JA-treated seed.

The priming agent β-aminobutyric acid influences plant pathogen responses when applied as a seed treatment

Whilst necrotrophic fungal pathogens such as B. cinerea are generally controlled by JA-dependent pathways, resistance against biotrophic pathogens is associated with SA-dependent pathways (Glazebrook, 2005). Various compounds are known which can prime SA-dependent resistance. One, BABA, is a nonprotein amino acid which is well known for its ability to prime a range of stress resistance responses in plants, including resistance against biotrophic pathogens (Zimmerli et al., 2000). We treated tomato seed with BABA to test whether it might also prime disease resistance when applied to the seed. As for seed treatments with JA, BABA treatment caused no reduction in plant growth. Plants grown from treated seed were challenged with an important pathogen of tomato, powdery mildew (O. neolycopersici). Plants grown from treated seed suffered significantly lower degrees of colonization by powdery mildew (Fig. 3a), suggesting that, like JA, BABA is able to prime defence responses in the growing plant when applied before germination.

Figure 3.

Effects of β-aminobutryric acid (BABA) and jasmonic acid (JA) seed treatments on disease. Tomato (Solanum lycopersicum) plants were grown in a glasshouse from seeds treated with a control solution, 3 mM JA, 18 mM BABA, or 3 mM JA + 18 mM BABA, and challenged with either Oidium neolycopersici (a) or Botrytis cinerea (b). Open bars, no JA; closed bars, + JA. (a) Plants were inoculated with spores of O. neolycopersici when they were 4 wk old. Bars show mean colony number cm–2 leaf area + SEM (= six leaves × three leaflets), 2 wk after inoculation. From two-way ANOVA, the effect of JA is not significant (= 0.49), whereas the effect of BABA treatment is highly significant (= 0.0008). The JA × BABA interaction was also not significant (= 0.188). (b) Botrytis cinerea lesion areas were measured as in Fig. 1. Bars show mean lesion area + SEM (= 8 leaves with 10 inoculations per leaf). From two-way ANOVA, the effect of JA is significant at < 0.0001, BABA at = 0.0002, but the JA × BABA interaction was not significant (= 0.121). Letters above bars indicate statistically significantly different means as determined by a Tukey post-hoc test.

Tradeoffs between different resistance mechanisms are minimized by seed treatment-induced priming

Because JA- and SA-dependent signalling pathways can be antagonistic when directly activated (Kunkel & Brooks, 2002; Glazebrook, 2005), we were interested to determine whether there may be interactions between JA and BABA seed priming treatments in the context of resistance against necrotrophic and biotrophic pathogens. We found that a seed treatment with JA alone did not alter resistance to powdery mildew, consistent with the idea that JA-induced resistance responses are not effective against biotrophs (Fig. 3a). Importantly, however, the data also indicate that JA seed treatment does not increase susceptibility to powdery mildew via negative cross-talk. Furthermore, inclusion of JA in a combined treatment with BABA did not significantly impact on the ability of BABA to enhance resistance (Fig. 3a). We also performed the reciprocal experiment to test whether BABA priming might act antagonistically with JA-induced resistance against the necrotroph, B. cinerea. In this case, we did detect increased susceptibility to disease in BABA-treated plants, but found no interaction between treatments in two-way ANOVA when JA increased resistance, indicating that, although in isolation BABA treatment can be antagonistic towards basal resistance against Botrytis, it did not prevent JA from priming resistance (Fig. 3b, Table S2).

JA-induced priming of Botrytis resistance depends on JA, ethylene and ABA signalling, and is associated with increased JA-dependent gene expression

To determine the influence of different plant hormones on JA priming of pathogen resistance, we examined the effects of JA seed treatments on resistance to Botrytis infection in tomato mutants disrupted in JA, ethylene and ABA responses. In contrast to results from wildtype plants, Fig. 4 shows that JA seed treatment was unable to increase Botrytis resistance in JL5 (def1) and Never ripe plants, which are deficient in JA biosynthesis and ethylene perception, respectively (Lanahan et al., 1994; Howe et al., 1996). Intriguingly, we found that in the ABA-deficient mutant, flacca, which is more resistant to Botrytis (Audenaert et al., 2002), JA seed treatment increased susceptibility to disease (Fig. 4).

Figure 4.

Jasmonic acid (JA) seed treatment-induced Botrytis resistance depends on JA, ethylene and ABA signalling. Mean lesion areas in leaves from treated tomato (Solanum lycopersicum) plants relative to control untreated plants from wildtype (Carousel) and the JL5/def1, Never ripe and flacca mutants (open bars, control; closed bars, JA). Data show means + SE from several independent experiments (Carousel, = 9; JL5/def1, = 6; Never Ripe, = 4; flacca, = 3), in which relative lesion areas were determined from 60–80 replicate measurements per genotype.

A common feature of defence priming is that plants in the primed state exhibit more rapid and/or stronger transcriptional responses to stress. To investigate whether this might be a mechanism underlying the seed treatment-induced disease resistance, Botrytis inoculated leaves were sampled over a 24-h time course for gene expression analysis. Quantitative real-time PCR (qPCR) was used to monitor expression of a number of well-known defence-related genes from tomato, including several regulated by JA. Although there was substantial inter-experiment variation in the exact timing of increases in expression, we found that Botrytis infection consistently resulted in early transcriptional activation of the JA biosynthetic gene ALLENE OXIDE SYNTHASE 2 (AOS2), mid-phase activation of a JA-responsive defence gene PROTEINASE INHIBITOR II (PinII) and late activation of the pathogenesis-related gene PR1b1. For AOS2 and PR1b1, we found no consistent difference between the timing or peak expression levels between control and JA seed-treated plants, but in the case of the JA-dependent defence gene, PinII, we observed higher expression in JA seed-treated plants in all three replicate experiments performed. Representative data illustrating these responses are shown in Fig. 5.

Figure 5.

Transcriptional responses of defence genes to Botrytis infection are altered by jasmonic acid (JA) seed treatment. Relative steady-state mRNA levels of the genes AOS2, PinII and PR1b1 measured by real-time PCR during a time course following inoculation of tomato (Solanum lycopersicum) leaves with Botrytis cinerea. Open bars, data from control untreated plants; closed bars, data from JA seed-treated plants. Error bars represent ± SD determined from three technical replicate assays. Data shown are representative of three independent time course experiments.

Discussion

Entry into a primed state enhances plant resistance to future stress episodes with minimal costs to growth and development, and may therefore be a desirable trait to exploit commercially. Here, we show that seeds are receptive to agents that establish a primed state for pest and disease resistance, resulting in long-term increases in resistance to biotic attackers, including a range of arthropod herbivores and fungal pathogens. Seed treatment with JA reduced the performance of all three herbivores tested. Importantly, these species are representative of the three major herbivore feeding guilds. JA is well-known for its role in defence against chewing insects (Wasternack, 2007), including lepidopteran larvae, as exemplified here by M. sexta (Howe et al., 1996), and against cell content feeders such as spider mites (Li et al., 2002). Its role in defence against aphids and other phloem feeding herbivores is less clear, since these insects tend to activate SA-dependent responses in the plant. However, this appears to be a decoy strategy by which aphids repress plant responses regulated by JA that provide more effective resistance (Walling, 2008).

Consistent with its role in defence against necrotrophic pathogens, we also found that JA seed treatment enhances resistance to B. cinerea. We used bioassays for Botrytis infection in a range of hormone mutants to begin to dissect the mechanism underlying the induced resistance afforded by JA seed treatment. As may be expected, we found that both JA and ethylene signalling are required for elevated resistance in plants grown from JA-treated seed. Our experiments do not distinguish between a requirement for these signals in perception of the seed treatment and subsequent expression of defence in infected leaves, and nor do they distinguish between degrees of basal resistance in the mutants. Nevertheless, our data show clear differences in the impact of JA treatment between lines, and is consistent with a model in which the seed treatment acts via the typical JA- and ethylene-dependent pathways for defence against necrotrophic pathogens. ABA is known as a negative regulator of disease resistance (Ton et al., 2009), and the ABA-deficient tomato mutant, sitiens, is hyper-resistant to Botrytis (Audenaert et al., 2002). Surprisingly, we consistently found that JA seed treatment increased susceptibility to Botrytis in the ABA-deficient flacca mutant. Whilst the explanation for this response remains unclear, our data suggest a complex interaction between JA and ABA signalling during Botrytis infection.

β-Aminobutryric acid can prime resistance in many plant species against a range of stresses, including both necrotrophic and biotrophic pathogens (Zimmerli et al., 2000, 2001; Ton & Mauch-Mani, 2004; Ton et al., 2005). When applied as a seed treatment in tomato, we found that it improved resistance against powdery mildew disease caused by the biotrophic fungal pathogen, O. neolycopersici, presumably via effects on SA-dependent resistance, as seen in other biotrophic and hemibiotrophic pathosystems (Zimmerli et al., 2000; Ton et al., 2005). However, we found that BABA seed treatments failed to promote Botrytis resistance. This contrasts with reports of BABA priming resistance to various necrotrophic pathogens, including B. cinerea, in other plant species when applied as a soil drench (Zimmerli et al., 2001; Ton & Mauch-Mani, 2004).

Importantly, we found that biotic stress resistance afforded by seed treatments was long-lasting, with significant effects on herbivore resistance evident in plants at least 8–9 wk old. Although other examples of the use of seed treatments to improve biotic stress resistance exist in the literature, those we are aware of measured short-term responses occurring in seedlings a few days old rather than mature plants (Jensen et al., 1998; Latunde-Dada & Lucas, 2001; Shailasree et al., 2001; Buzi et al., 2004). A long-lasting effect on stress resistance as seen here is suggestive of a priming response rather than constitutive activation of defences. Apart from some early effects of JA seed treatments on germination and seedling root growth, which are consistent with known effects of JA on these processes (Wasternack, 2007), we were unable to detect long-term effects on vegetative and reproductive growth at the concentration used here. Constitutive activation of plant defence is commonly associated with reductions in growth and reproductive fitness (Heil et al., 2000; Cipollini et al., 2003; Walters & Heil, 2007). Priming of defences, on the other hand, minimizes these costs whilst improving future resistance to attack (van Hulten et al., 2006), consistent with the effects of the seed treatments employed here. Measurements of defence gene expression also support the idea that JA seed treatment primes future JA-dependent defences. Expression of the genes assayed was similar in control and treated plants before biotic stress, which argues against constitutive activation of defence, but at least in the case of PinII, induced levels of transcripts in response to Botrytis infection were significantly elevated in JA seed-treated plants. Perhaps significantly, proteinase inhibitors, including PinII, have recently been demonstrated to be essential for resistance against Botrytis (El Oirdi et al., 2011).

The antagonism between the JA- and SA-dependent pathogen resistances responses assayed here via their effects on Botrytis and powdery mildew, respectively, is well studied (Kunkel & Brooks, 2002; Glazebrook, 2005). Although the exact mechanism by which this antagonism arises is still to be fully elucidated, it is clear that interactions occur at the signalling level when endogenous JA and SA concentrations are elevated. It might therefore be predicted that priming, rather than constitutive activation of JA- and SA-dependent resistances, would have minimal impact on the other response. This prediction was found to be broadly correct in our experiments, since neither treatment prevented the ability of the other to increase resistance against the corresponding pathogen in combined JA and BABA seed treatments. However, BABA seed treatment alone tended to increase susceptibility to Botrytis in the absence of a JA seed treatment. One possible explanation for this observation is that BABA priming of SA-dependent defence can interfere with the endogenous JA-dependent resistance against necrotrophic pathogens. However, BABA primes resistance against Botrytis and other necrotrophs via SA-independent mechanisms in other systems (Zimmerli et al., 2001; Ton & Mauch-Mani, 2004). In Arabidopsis, one of these SA-independent mechanisms is via priming of ABA-dependent responses (Ton & Mauch-Mani, 2004). The suppression of resistance to Botrytis by ABA in tomato (Audenaert et al., 2002) may provide an alternative explanation for the negative effect of BABA seed treatment on Botrytis resistance. In either case, our data indicate that increased susceptibility to Botrytis following BABA seed treatment is overcome when plants are also primed for JA responses.

Long-term effects of the environment on plant genomes have recently attracted growing attention, particularly with regard to epigenetic mechanisms for the regulation of gene expression. A number of studies of mutants affected in DNA methylation and histone modification, which function as epigenetic regulators of gene expression, show that genome-wide changes in chromatin status have pleiotropic consequences, ranging from development to stress tolerance (Reinders et al., 2009; Kim et al., 2010; Luo et al., 2011), including JA-dependent responses (Zhou et al., 2005; Wu et al., 2008; Berr et al., 2010). Since the level of transcriptional activity of stress-related genes is maintained by epigenetic marks, it follows that changes in chromatin modifications as a consequence of stress may be one mechanism by which priming may operate. For example, van den Burg & Takken (2009) recently put forward a model for priming during SAR based on histone replacement at defence-related gene promoters following pathogen recognition, and Jaskiewicz et al. (2011) showed that changes in histone acetylation and methylation were correlated with priming during SAR. It is possible, therefore, that the long-term priming effects we observe here as a consequence of JA and BABA seed treatments may be mediated via epigenetic modifications of JA- and BABA-responsive genes in embryonic tissues during imbibition. In this way, these genes could become more responsive to JA- and SA-dependent signalling pathways during biotic attack on the growing plant. Interestingly, there is also mounting evidence for transgenerational changes in the sensitivity of plant resistance responses mediated by stress-induced epigenetic changes (Boyko et al., 2010; Kathiria et al., 2010; Scoville et al., 2011).

The control of pests and pathogens in crop plants by synthetic pesticides with direct toxic activity is becoming increasingly less desirable, and the use of more environmentally-friendly approaches are required for a more sustainable future. The use of elicitors of plant defences, or ‘plant activators’ as they have been termed, has been proposed as an alternative approach to crop protection (Vallad & Goodman, 2004; Bruce, 2010). However, commercial success in this area is currently limited. Priming of natural plant defences, as well as minimizing yield penalties, should be compatible with other facets of integrated pest management strategies, such as the use of biological control. Since seed treatments are economically more attractive than chemical application to plants in the field and require no action by growers, the approaches we have described here may have useful applications in agriculture and horticulture.

Acknowledgements

The work presented here was funded by grants from the UK Natural and Environment Research Council, the Biotechnology and Biological Sciences Research Council and the UK government Department for Food and Rural Affairs. We also acknowledge the support of the Horticultural Development Company which provided a Fellowship to J.P.M. We thank Mr Phil Nott for technical assistance with herbivore cultures and bioassays.

Ancillary