Generalist insects behave in a jasmonate-dependent manner on their host plants, leaving induced areas quickly and staying longer on distant parts

Abstract

Plants are sessile, so have evolved sensitive ways to detect attacking herbivores and sophisticated strategies to effectively defend themselves. Insect herbivory induces synthesis of the phytohormone jasmonic acid which activates downstream metabolic pathways for various chemical defences such as toxins and digestion inhibitors. Insects are also sophisticated animals, and many have coevolved physiological adaptations that negate this induced plant defence. Insect behaviour has rarely been studied in the context of induced plant defence, although behavioural adaptation to induced plant chemistry may allow insects to bypass the host's defence system. By visualizing jasmonate-responsive gene expression within whole plants, we uncovered spatial and temporal limits to the systemic spread of plant chemical defence following herbivory. By carefully tracking insect movement, we found induced changes in plant chemistry were detected by generalist Helicoverpa armigera insects which then modified their behaviour in response, moving away from induced parts and staying longer on uninduced parts of the same plant. This study reveals that there are plant-wide signals rapidly generated following herbivory that allow insects to detect the heterogeneity of plant chemical defences. Some insects use these signals to move around the plant, avoiding localized sites of induction and staying ahead of induced toxic metabolites.

1. Introduction

During the course of evolutionary history, plants have established intricate antagonistic relationships with herbivores. Their sessile nature has forced plants to develop sophisticated ways of detecting threats and multiple interrelated strategies to defend themselves. Biochemical strategies range from the production of toxic and bitter tasting chemicals to kill or repel attackers [1,2], to proteinases that inhibit the digestion of plant tissue [3], to programmed cell death that sacrifices non-essential plant parts [4]. Biochemical defences induced in plants by the onset of herbivorous insect feeding are predominantly initiated by the synthesis of the phytohormone jasmonic acid (JA) in response to localized cell damage and chemical elicitors in insect saliva [5]. In a major branch of the JA response pathway, JA conjugates with the amino acid isoleucine [6] and the conjugate (JA-Ile) binds to an F-box protein receptor, COI1 [7], causing ubiquitination and degradation of JAZ protein repressors [8,9] and liberation of their target transcription factors, such as the MYC class of transcription factor [10], from the repressor complex (figure 1). The archetypal transcription factor associated with jasmonate responses, MYC2, activates various downstream metabolic pathways [11] involved in plant defence against herbivores [12,13]. An unknown mechanism signals an insect attack to the rest of the plant and may initiate systemic jasmonate biosynthesis in undamaged leaves [14]. Vascular architecture plays a key role in the systemic induction of plants [15]; however the spatial and temporal scale of systemic change within plants following insect attack is still poorly understood.

Figure 1.

A simplified schematic of jasmonic acid (JA) synthesis and activation of the MYC branch of JA-isoleucine-induced transcriptional responses in A. thaliana.

Insects are also sophisticated organisms, and many have evolved physiological adaptations that negate induced plant defences [16]. Behavioural adaptation also may be crucial for bypassing the host's induced defence strategies. Optimal foraging theory suggests that plant-eating insects need to move away from their feeding sites in order to avoid induced negative effects on food quality [17]; a strategy which results in a more dispersed pattern of damage to plants [18–20], and which in turn influences the expression of induced plant defence [21]. In a single field study on insect movement in relation to induced plant defence, feeding insects were found to move away from whole plants before the negative consequences of induction could have an effect [22]. The authors of a recent study of wild tobacco plants concluded that leafhoppers select host plants by eavesdropping on JA-mediated signalling [23]. Both these field studies indicate that herbivore host selection and movement are influenced by events early in the JA-mediated plant response. While there is obvious benefit to an insect in removing itself from harmful chemicals, the plant may also benefit from increased herbivore movement [17]: more dispersed feeding damage may be tolerated or compensated for more easily, and may lead to epigenetic resistance in the next generation of plants [24]. Also, there is an energetic cost to the insect that is associated with moving [25], and it may be more vulnerable to exposure [26] or predation [27]. While our understanding of insect-induced defence pathways in host plants has increased significantly in recent years [5], studies explicitly linking herbivore behaviour to jasmonate-induced changes in plant chemistry are few and all infer movement from feeding damage [19,21,23]. A major obstacle to these studies has been the inability to map the induced changes in real time, within plants at the small spatial scale relevant to most insect herbivores. We have met this challenge and gone on to quantify herbivore movement in relation to early induced changes in plants by direct observation.

Here, we have used the model insect–plant systems of the polyphagous caterpillar Helicoverpa armigera (Hübner) (Lepidoptera: Noctuidae), and the oligophagous Plutella xylostella (Linnaeus) (Lepidoptera: Plutellidae), on Arabidopsis thaliana (L. Heynh) (Brassicaceae) to answer the question: is herbivore movement on hosts driven by induced plant defence? Arabidopsis thaliana is a well-established model for understanding plant responses to herbivory and other biotic and abiotic stressors. Helicoverpa armigera is a widespread pest feeding on a large range of host plants [28], whereas P. xylostella, also a widespread pest, is a specialist feeder on the Brassica family of plants [29].

2. Results

(a). Herbivore movement within elicited plants

To determine whether insects avoid plant parts affected by induced chemical changes, we conducted bioassays of H. armigera and P. xylostella on A. thaliana wild ecotype Col-0 (wild-type, WT) and the loss-of-function jasmonate signalling mutant myc2 in a Col-0 background [30]. Arabidopsis thaliana leaves are arranged in spiral patterns and were numbered by order of appearance (figure 2a), and are connected by vasculature [31]. First instars of H. armigera that were placed on WT leaves elicited by larval feeding 1 h earlier (leaf 11), or on a neighbouring leaf (leaf 6), left more quickly, whereas insects placed on a distant leaf (leaf 7) stayed longer, as compared with insects placed on any of these leaves of unelicited WT or elicited myc2 plants (figure 2a,b). The difference was significant between the time insects stayed on elicited and neighbouring leaves versus distant leaves of WT plants (log-rank test [32], n = 59, χ² = 5.18 on 2 d.f., p = 0.028). There was no significant difference between the staying times of H. armigera on any leaf of a control WT or myc2 plant, so for clarity, data for all three leaves of these plant treatments have been combined in figure 2b (uncombined plots in electronic supplementary material, figure S1). In contrast, caterpillars of the specialist herbivore, P. xylostella showed no significant difference in time taken to move from the leaf of placement with regard to plant genotype, plant elicitation, or starting leaf (figure 2a,c). These results show clearly that the generalist-feeding H. armigera responded to MYC-dependent signalling within the plant by changing its behaviour whereas the specialist P. xylostella did not. The data (figure 2b) suggest that the signal responded to by H. armigera is plant-wide, but the insect's behavioural response is opposite on elicited and neighbouring leaves (i.e. insects move away) compared to distant leaves (i.e. insects stayed).

Figure 2.

The effect of prior insect feeding (elicitation) on the behaviour of a generalist and specialist herbivorous insect on A. thaliana. (a) Diagram of a four-to-five week old A. thaliana. Leaves numbered by order of appearance and shaded by number of vascular connections joining leaves to leaf 11, with the closest being darkest. Leaves 3, 6, 16 and 19 are joined to leaf 11 by one vascular connection, leaves 1, 8 and 14 by two successive vascular connections, leaves 2, 4, 9 and 13 by three successive vascular connections, leaves 5, 7, 10, 12 and 17 by four vascular connections and leaves 15, 18 and 20 by five successive vascular connections. In our assays, leaf 11 was the elicited leaf, leaf 6 the neighbouring (directly connected) leaf, and leaf 7 the distant leaf; (b) Kaplan–Meier plots of day old H. armigera larvae staying on their leaf of placement over 60 min. Leaf treatments (leaves onto which larvae were placed) were: leaf 11, elicited by H. armigera feeding 1 h beforehand (elicited, n = 20); or leaf 6, neighbouring the elicited leaf (neighbouring, n = 18); or leaf 7, distant to the elicited leaf (distant, n = 21), in WT Col-0 background; the same leaves of myc2 plants (data from all leaves combined (myc2, n = 45)); and the same leaves of control WT plants that were not elicited (data from all leaves combined (control, n = 45)); and (c) Kaplan–Meier plots of second instar P. xylostella larvae staying on their leaf of placement over 60 min (leaf treatments are the same as for (b)).

(b). The induced state of leaves within plants following elicitation by H. armigera

To confirm the induced state of leaves used in the behavioural assay reported above, real-time quantitative polymerase chain reaction (RT-qPCR) with reverse transcription  for two jasmonate-responsive genes up-regulated early following herbivory was performed on the test/elicited (leaf 11), neighbouring (leaf 6) and distant (leaf 7) leaves of elicited WT and myc2 plants and the same numbered leaves on unelicited WT plants. JAZ1 [8,9] (At1g19180) encodes a protein that represses the MYC2 [11] (At1g32640) transcription factor and prevents downstream signalling and synthesis of defensive metabolites in unattacked plants. The degradation of JAZ following herbivory and JA-Ile synthesis (figure 1) results in the up-regulation of the JAZ gene. Expression of JAZ1 increased more than twofold in the elicited leaves of WT and myc2 mutant plants, but not in neighbouring or distant leaves, when compared with the same leaves from control plants (figure 3a). The expression of MYC2 also increased more than twofold in the WT elicited leaves, but not in the elicited leaves of myc2 mutants or in any neighbouring or distant leaves (figure 3b). These results confirmed the induced state of the elicited (leaf 11) leaves as expected. However, there was no upregulation of jasmonate-induced genes in leaves neighbouring WT induced leaves (i.e. in leaf 6). This was unexpected because insects left neighbouring leaves as quickly as elicited leaves during the bioassays (figure 2b) and we expected to see some evidence of induction in these neighbouring leaves. Several studies have shown that the mobile signal activating systemic jasmonate defence pathways in plant parts distant from the local area of damage is both dependent on and constrained by plant vasculature [15,19,33,34]. If this is the case, induced gene expression in neighbouring leaves may have been concentrated around the main vein and went undetected in our whole leaf samples because of the dilution effect of sampling whole leaves for RT-qPCR. Hence, we used a reporter-gene method for visualizing induction within leaves in more detail.

Figure 3.

Jasmonate-dependent gene expression in leaves of A. thaliana following elicitation by H. armigera 1st instars. Fold change in (a) JAZ1 and (b)MYC2 gene expression in elicited leaf 11, neighbouring leaf 6, and distant leaf 7 of WT and myc2 plants, where leaf 11 was elicited using a single feeding event by a first instar H. armigera 1 h previously, compared to the same leaves of WT unfed-on plants. White bar, elicited leaf; grey bar, neighbouring leaf; black bar, distant leaf.

(c). The real-time, fine-scale pattern of induced gene expression within leaves following elicitation by H. armigera

To determine whether systemic induction in leaves was constrained to the main veins and nearby cells, we examined the spatial distribution of induced tissue within leaves in situ in real time using transgenic A. thaliana plants expressing firefly luciferase (LUC) under the control of an early jasmonate-responsive promoter gene A70 (At5g56980) [35]. A bioluminescent LUC signal appeared around the feeding site in the elicited leaf 30 min post-feeding (pf), peaked around 2–3 h pf, and lasted approximately 10 h pf (figure 4). No A70:LUC expression was detected in any part of the neighbouring or distant leaves within 2 h pf (the time period of our behavioural assays). However, subsequently the main vein of the neighbouring leaf 6 showed transient gene expression at 8 h pf and the main vein of leaf 8 at 12 h and again at 34 h pf (figure 4). This result suggests that systemic induction does develop over time, even following a single feeding event as elicitor, and that the pattern of systemic induction follows vascular connectivity to ‘sink’ leaves.

Figure 4.

Luciferase imaging of the jasmonate-responsive A70 gene in whole plants following a single feeding event by a 1st instar H. armigera on leaf 11 of an A70:LUC plant. (a) Bioluminescent signal from a plant constitutively expressing luciferase; A. thaliana Landsberg erecta FLC:LUC was used as a positive control and photographed at the same time as the A70:LUC images. (b) The A70:LUC plant photographed in darkness 1.5–2.0 h, 7.5–8.0 h, 11.5–12.0 h and 33.5–34.0 h post-feeding. (b) The mask placed over the background delimits the plant leaf area and has been applied uniformly to all images.

(d). Is the plant signal that leads to movement away from plant parts a volatile?

To determine whether the unknown plant signal responsible for the movement of H. armigera away from leaves neighbouring elicited leaves was a volatile emitted from the elicited leaf, rather than a signal associated with the vasculature, we observed the behaviour of insects on neighbouring leaves placed adjacent to elicited leaves on another plant, i.e. where the elicited and neighbouring leaves were not on the same plant and therefore not vascularly connected. There was no significant difference between the time taken for the larvae to leave a leaf adjacent to an elicited leaf and a leaf adjacent to an unelicited leaf (figure 5) (log-rank test [32], n = 52, p = 0.463). From this result, we concluded that the signal for moving away from induced plant parts was unlikely to be a volatile.

Figure 5.

The movement of 1st instar H. armigera from leaves positioned next to but not vascularly connected to elicited leaves. (a) The arrangement of two A. thaliana WT plants so that leaf 11 of one is next to leaf 6 of the other; leaf 11 was elicited by a single feeding event of a 1st instar H. armigera and 1 h later a second 1st instar H. armigera placed on leaf 6. (b) Kaplan–Meier plots of larvae staying on their leaf of placement next to elicited or control leaves over 60 min (n = 23, 29). (b) Solid line, adjacent to an unelicited leaf; dashed line, adjacent to an elicited leaf.

(e). The movement of Helicoverpa armigera between plants of differing jasmonate-signalling ability

Lastly, we tested whether or not jasmonate-induced defence pathways in A. thaliana affected movement between plants as well as within plants as shown above. We measured the movement of first instar H. armigera away from plants of differing jasmonate-signalling ability in the laboratory by pairing either a WT A. thaliana, a mutant silenced in the jasmonate signalling genes MYC2, MYC3 and MYC4, or a transgenic overexpressing 35S:MYC2 plant with WT plants. A single first instar H. armigera was placed on leaf 11 of each starting plant of the pair. After 4 days, 30 per cent of larvae placed on WT, 0 per cent of larvae placed on myc2, myc3, myc4 and 60 per cent of larvae placed on 35S:MYC2, were on the second receiver plant (n = 10 for each pairing) (figure 6). There was a significant association between the genotype of the starting plant and the distribution of larvae at the end of the assay after 4 days (χ² = 14.56, d.f. = 4, p = 0.006). The largest contributors to the significant χ² result were that fewer larvae than expected moved from myc2, myc3, myc4 to the paired WT plant, and more larvae moved from the 35S:MYC2 plant. This assay provides further evidence of the movement of H. armigera away from jasmonate and MYC-dependent induced change in host plants.

Figure 6.

The movement of 1st instar H. armigera between A. thaliana plants over 4 days when the starting plant where larvae were placed on leaf 11 was a (a) WT, (b) myc2,myc3,myc4 mutant, or (c) a 35S:MYC2 transgenic plant.

3. Discussion

Taken together, our results clearly demonstrate an induced behavioural response by generalist insects to jasmonate and MYC2-dependent changes in plants following herbivory. We have shown that insects can react to early-induced changes in plants, before the synthesis of anti-herbivore compounds such as proteinase inhibitors, or increased levels of glucosinolates, are thought to occur [2,3]. The insects’ moving or staying behaviour is innate, as our experimental insects had no previous experience of plants, and must provide a role in enabling the insect to adapt to its changing environment on the plant. We have shown that a generalist insect can minimize contact with defended or soon-to-be defended plant parts by leaving them quickly and staying longer on the undefended parts of attacked plants. Insects appear to be moving ahead of the ‘induction front’ of defensive biochemical changes within plants. A moving target model of induced plant defence has been proposed, where the heterogeneity of induced plant qualities is associated with increased plant fitness [36]. Such a strategy may be difficult for herbivores to cope with if the heterogeneity is unpredictable [37]. But our study suggests that some insects are able to detect this heterogeneity and respond accordingly, thereby modifying their foraging to match the defences of the plant. The role of herbivore behaviour in modifying the costs and benefits of induced plant changes in relation to both the herbivore and the plant needs further attention.

Avoiding induced tissue early would give insects time for their own physiological adaptations to defensive metabolites [16] to develop. Previously, we have found that first instar H. armigera spend less time feeding and more time moving than third instars, but despite this, first instars have a higher relative growth rate [38]. Moving on from and therefore avoiding induced plant tissue could contribute to this higher growth rate. Behavioural resistance to induced defences would be expected to benefit physiologically vulnerable insects more than those adapted or habituated to the changed plant environment. The specialist P. xylostella, which has evolved physiological resistance to defensive chemicals in Brassica plants [29], did not demonstrate avoidance of induced plant parts in our experiments. For generalist insects such as H. armigera, as some physiological resistance and/or tolerance develops over an insect's lifetime [39], younger insects would be the most likely to display this evolved behaviour in response to induced plant chemistry.

The nature of the early warning signal to insects is unknown. Our experiments suggest that JA does not directly lead to insect avoidance [19] as the JA signalling mutant myc2 synthesizes JA [30] but the insects did not show the behaviour on this plant. Moreover, insects stayed longer on the distant leaves of attacked plants than on the same leaves of control unattacked plants, suggesting one signal that is differentially expressed (i.e. able to increase and decrease from its steady state) or two different signals; one that initiates egress from induced parts, and another that promotes staying on the uninduced parts of attacked plants. Given the millions of years of coevolution of insects and plants, insect sensory receptors are likely to be sensitive to smaller changes in plant metabolism than our laboratory methods can detect. During our behavioural assays, significantly more ‘movers’ did not feed (62%) than fed before they moved (χ² goodness of fit, χ² = 4.404, p = 0.036, n = 82), therefore the leaving response has not evolved as a behaviour that occurs after a fixed amount of feeding. From this and other studies [19,33,34], there is consistent evidence that plant vasculature plays a central role in the propagation of induction signals from the feeding site. An exogenous volatile signal cannot be ruled out, but appears unlikely unless it is working in concert with a vascular-dependent signal. The volatile methyl jasmonate is rarely emitted by green plant tissues despite its endogenous accumulation [40]. Green leaf volatiles (GLVs) involved in plant defence are not synthesized via the MYC2-regulated branch of the JA pathway [40] whereas the insect moving and staying behaviour is dependent on a functional MYC2 gene. Further, we have shown previously that first instar H. armigera are neither attracted nor repelled by (Z)-3-hexenyl acetate, an ubiquitous GLV which affects early instar foraging by modifying phototaxis [41].

Our results clearly demonstrate a behavioural mechanism used by physiologically vulnerable insects to maximize their fitness in the face of myriad induced plant defences. Herbivorous insects can quickly perceive induced changes in plant tissue which signal information about the quality of the plant as a food source not just at that time, but also in the future. In insect–plant interactions, induced plant defence should be considered as a process that varies at a fine spatial scale in ways that are biologically relevant. The heterogeneity of plant responses provides a basis for the evolution of behavioural traits in herbivores, such as induced movement, that limit fitness costs. The ability of insects to predict and avoid detrimental defensive plant chemistry may mean that some coevolutionary relationships between plants and other organisms need to be re-interpreted. A priority for future research is to determine how widespread induced behavioural change is within insect and other groups, and to find the nature of the ‘early warning’ plant signal that elicits the changes.

4. Material and methods

(a). Insects and plants

Helicoverpa armigera were obtained from a culture maintained at The University of Queensland at 25°C in a 12 L : 12 D cycle [42] and supplemented with wild-caught moths each summer. Plutella xylostella were obtained from a recently established culture maintained on Brassica plants at The University of Queensland. Single A. thaliana plants of the WT Columbia-0, loss of function mutants myc2 [30] and myc2, myc3, myc4 [13], and transgenic 35S:MYC2 [30] and A70:LUC genotypes were grown from seed in 200 ml pots in an illuminated growth chamber with a 14 L : 10 D cycle at 18°C and used when four-to-five week old with approximately 20 leaves. A fine mesh stocking was used to cover the potting mix, and leaves of interest were marked with pins beside but not touching the leaves. Transgenic A. thaliana plants expressing firefly luciferase (LUC) under the control of an early JA-responsive promoter gene A70 (At5g56980) [35] were constructed as follows: Promega pGL4-11 was digested with Fse1, ends were polished then digested with Kpn1. The luc2P containing fragment was cloned into pCAMBIA1302 digested with Kpn1 and Pml1 to generate pC1LUCP. A 1631 bp promoter fragment of A70 was PCR amplified from Col-0 genomic DNA using the primers: A70F + KpnI AATCCATGGTCAACCGTAAAAGGTCGGTGTAG and A70R + NcoI AACTCCATGGTTGGGTTGTGTTTTATGTTGGTTTTG. The PCR product was digested with Kpn1/Nco1 and cloned into Kpn1/Nco1 cut pC1LUCP and sequence verified. Transgenic homozygous A70:LUC lines were generated in Col-5 background by standard methods.

(b). Elicitation of Arabidopsis thaliana plants

We elicited one leaf per plant (leaf 11) of A. thaliana WT and myc2 by allowing a day-old starved H. armigera or second instar P. xylostella to make a single feeding hole (approx. 0.3–0.4 mm²). Young H. armigera eat small ‘exploratory’ feeding holes before they settle into extended feeding cycles [42], so a single feeding event is a minimal but biologically relevant elicitor of induction.

(c). Behavioural assay of insects on elicited plants

We tested whether insects move away from plant parts that have been induced by earlier insect feeding. One leaf (leaf 11) per plant of A. thaliana WT and mutant myc2 was elicited by insect feeding as above. One hour after elicitation, and coinciding with maximum JA accumulation within the leaf [43,44], a second day-old starved H. armigera or second instar P. xylostella larva was placed on either the elicited leaf number 11, its neighbouring, vascularly connected leaf number 6 or a distant leaf number 7 (figure 2a) and observed for 1 h. To test whether or not the plant signal leading to the movement of H. armigera from leaves neighbouring elicited leaves was a volatile, WT plants were arranged in pairs such that leaf 11 of one plant was adjacent to leaf 6 of another unelicited plant (figure 5). One hour after elicitation of leaf 11 by a first instar H. armigera, a second first instar H. armigera larva was placed on the adjacent leaf 6 and observed until it left the leaf or for 1 h, whichever came first. All behavioural assays were undertaken in the laboratory maintained at approximately 25°C and ambient RH. The windows were boarded so no natural light could enter and lighting was by two overhead fluorescent tubes. For all behavioural assays, time to event (event = leaving) analyses using Kaplan–Meier methods with right censoring was used to analyse the likelihood of insects leaving the leaves they were placed on [32,45]. Data deposited in the Dryad repository: doi:10.5061/dryad.cf87p.

(d). Amplification of MYC2 and JAZ1 transcripts and real-time RT-qPCR

To sample leaves without inducing further gene expression, entire plants were flash frozen in situ by inverting into liquid nitrogen for 15 s and then sat in a bath of liquid nitrogen while individual leaves were detached by snapping off with forceps. Frozen leaves were ground individually and total RNA was extracted using NucleoSpin RNA Plant kits (Machery and Nagel) according to the manufacturer's instructions, and quantified using a NanoDrop ND-1000 Spectrophotometer. Complementary DNA (cDNA) was synthesized according to the SuperScript III First-Strand cDNA Synthesis protocol (Invitrogen) using oligo(dT)₁₅ primer (Promega). RT-qPCR was performed in duplicate on cDNA samples for JAZ1 and MYC2 genes. Gene-specific primer sequences are given in electronic supplementary material, table S1. Each reaction contained 20 ng cDNA, 200 nM forward primer, 200 nM reverse primer, 5 µl SensiMix SYBR (Bioline) and water to 10 µl. Disassociation curves were checked for single peaks. Technical replicates were averaged and the primer efficiency (PE) for each gene calculated. Transcript abundance was determined using the comparative cycle threshold (CT) method; JAZ1 and MYC2 gene expression was normalized against 18S using the formula PE_gene^–avCTgene/PE_18S^–avCT18S. Raw 18S CT values were analysed by ANOVA to ensure 18S expression did not vary significantly between treatments. Gene expression data shown in figure 3 was analysed by ANOVA with respect to plant type and leaf number and pair-wise comparison of leaf number within plants was performed using the p-value adjustment method of Holm. Statistical software used was R [45]. Data deposited in the Dryad repository: doi:10.5061/dryad.cf87p.

(e). Bioluminescent imaging of whole plants expressing A70:LUC

At least 24 h prior to assays, plants were sprayed morning and afternoon with 300 µl of freshly prepared 1 mM luciferin (d-Luciferin sodium salt, Goldbio) in water to establish a substrate balance within the plant and avoid visualizing previously accumulated luciferase. The bioluminescent signal was captured using a Pixis 1024B Ultra-low light camera and WinView32 imaging software (Princeton Instruments, USA). After focusing the image under ambient light, untreated plants were photographed in the dark for 30 min exposure to check that they were not bioluminescent prior to treatment. After treatment, plants were photographed in the dark using sequential 30 min exposures. Four plants were photographed simultaneously: an untreated A70:LUC plant (negative control) and a constitutively LUC expressing A. thaliana Landsberg erecta FLC:LUC construct [46] (positive control) were visualized simultaneously with two treated A70:LUC plants.