Herbivore-induced jasmonic acid bursts in leaves of Nicotiana attenuata mediate short-term reductions in root growth


A. Walter. Fax: +49 2461 61 2492; e-mail: a.walter@fz-juelich.de


Root growth in Nicotiana attenuata is transiently reduced after application of oral secretions (OS) of Manduca sexta larvae to wounds in leaves. Feeding of M. sexta or OS elicitation is known to result in jasmonic acid (JA) and ethylene bursts, and activates a suite of defence responses. Because both plant hormones are known to strongly reduce root growth, their activation might account for the observed reduction of root growth following herbivory. To test this hypothesis, we measured primary root growth with digital image sequence processing at high temporal resolution in antisense-lipoxygenase 3 (asLOX3) and inverted repeat-coronatin-insensitive 1 (irCOI1) seedlings which are impaired in JA biosynthesis and perception, respectively, and wild-type (WT) seedlings. Higher root growth rates in irCOI1 compared with WT were observed after OS elicitation. The dynamics of wound-induced root growth reduction coincide with the dynamics of root growth reduction induced by external application of methyl JA. In an experiment with 1-methylcyclopropen (1-MCP), a potent ethylene receptor blocker, no wounding-specific difference between growth of 1-MCP-treated plants and non-treated plants was observed, suggesting that wound-induced endogenous JA and not ethylene mediates the wounding-specific reduction in root growth. Yet, inhibiting the ethylene response by applying 1-MCP led to markedly increased root growth compared with that of control plants, indicating that ethylene normally suppresses plant growth in N. attenuata seedlings.


When the specialist lepidopteran herbivore Manduca sexta attacks Nicotiana attenuata, the immediate growth reductions that are observed are more severe in the root than in the shoot (Hummel et al. 2007). Merely wounding leaves mechanically also leads to transient root growth reduction, but more pronounced effects are found when oral secretion (OS) and regurgitants of M. sexta (OS) are applied to puncture wounds (OS elicitation). The interaction between the insect and its natural host plant is well studied at different levels (Kessler & Baldwin 2002) and the central role played by jasmonic acid (JA) in the signalling cascade has been shown elegantly in recent years (Baldwin 1998; Halitschke & Baldwin 2003; Li et al. 2004). Yet, as a number of different signalling systems are induced upon herbivore attack and as growth is controlled by a network of factors, it is unclear what triggers the process that results in reduced growth.

In most plants, jasmonates (JA and other oxylipin derivates) play a major role in mediating induced defence responses to herbivore attack. Jasmonates are involved not only in plant defence against pathogens and herbivores but also in plant development processes such as root growth, fruit ripening, tendril coiling, tuberization, reproductive development and senescence (Creelman & Mullet 1997; Lorenzo & Solano 2005). JA and its derivates are known to be potent root growth inhibitors and many studies have used this property to isolate mutants deficient in their JA signalling, such as jar1 (Staswick, Su & Howell 1992), jin1 and jin4 (Berger, Bell & Mullet 1996), and coronatin-insensitive 1 (coi1) (Feys et al. 1994). Yet, these findings resulted from experiments in which JAs were applied exogenously to roots, often at non-physiological concentrations (Tung et al. 1996; Toro, Martin-Closas & Pelacho 2003; Uppalapati et al. 2005; Paschold, Halitschke & Baldwin 2007). Whether the herbivory elicited reductions in root growth are mediated by changes in endogenous JA levels has not yet been shown. Zavala & Baldwin (2006) recently reported that a JA-deficient mutant (asNaLOX3) outperformed wild-type (WT) plants in terms of growth and fitness when plants were undamaged and grown under glasshouse conditions, suggesting growth is regulated via JA signalling.

Upon herbivore attack and OS elicitation, both JA- and ethylene signalling are strongly activated and levels of both phytohormones accumulate above the levels that are elicited by mechanical wounding alone (Kahl et al. 2000). Hence, it is conceivable that a reduction of root growth is mediated by ethylene, which is know to be a potent root growth inhibitor (Abeles, Morgan & Saltveit 1992; Visser & Pierik 2007). It is also conceivable that the immediate cause of growth reduction upon herbivore attack is related neither to JA nor to ethylene directly but, rather, to the diversion of metabolites from primary (growth-sustaining) to secondary (defence-related) metabolism (Herms & Mattson 1992), which in turn may be mediated by hormonal signalling.

The aim of this study was to determine if either ethylene- or JA signalling mediates the reduction in root growth that is associated with herbivore attack. To elucidate this question, we first analysed the effect of 1-methylcyclopropen (1-MCP) on root growth in N. attenuata. 1-MCP is a non-competitive inhibitor that persistently binds to ethylene receptors, inhibiting plants from reacting to ethylene (Sisler, Dupille & Serek 1996). In subsequent experiments, we tested whether root growth dynamics are altered in transgenic N. attenuata plants impaired in JA signalling. We used an antisense-lipoxygenase 3 (asLOX3) line, in which the specific LOX that supplies the JA biosynthetic cascade with fatty acid hydroperoxides is silenced and has reduced JA levels after wounding and OS elicitation compared with the WT plants (Halitschke & Baldwin 2003), and a COI1-silenced line [inverted repeat-COI1-(irCOI1)], in which the F-box protein that mediates JA perception is silenced by expression of an inverted repeat construct (Paschold et al. 2007).


Plants and cultivation system for 1-MCP experiments

We used seeds of an inbred line of N. attenuata Torr. Ex Watts (22 generation) as a WT genotype originating from a natural population in Utah. Seeds were treated with smoke and were sterilized as described by Krügel et al. (2002). Five seeds were planted into square Petri dishes (120 × 120 × 17 mm) half-filled with sterile, solidified 1% Phytagel (w/v) with full-strength Gamborg B5 Medium (Duchefa, Haarlem, the Netherlands). Petri dishes were wrapped with one layer of fabric tape (Micropore, 3M Health Care, Neuss, Germany) to guarantee gas exchange and were placed at an 85° angle to ensure that roots grew along the bottom. The shoots grew in the air-filled volume of the Petri dish. Seedlings were treated as they grew in this cultivation system. Plants were grown in a cultivation room under 26 °C during the light phase and 22 °C during the dark phase. They were exposed to a photon flux density of 85 µmol m−2 s−1 between 0600 and 2000 h.

Root growth

Growth velocity of the primary root tip (VTip) of each individual seedling was quantified using a ruler in the 1-MCP experiment. In this experiment, we were interested whether we can restore the negative spit effect on root growth, demonstrated (Hummel et al. 2007) by blocking the ethylene receptors with the application of 1-MCP; hence, we chose the ‘classical’ monitoring method in order to gain a higher number of replicates per experiment. The ‘microrhizotron’ set-up used for high-resolution monitoring of root growth described further below is limited to only one root at the same time, which increases the time of experimentation significantly. Each day at 1300 h, the position of the root tip was marked with a pen on the rear side of the Petri dish. For our analysis of root growth, we included only plants with a root length between 25 and 35 mm on the treatment day. All suitable plants were treated and VTip was averaged within each Petri dish. The number of Petri dishes was used as the replicate number for each treatment; the same holds true for other analyses unless otherwise specified.

1-MCP application

Two days before treatment with gaseous 1-MCP, five holes (Ø 0.5 cm) were melted with a soldering iron into the upper border of the Petri dishes to increase gas diffusion into the dishes. Subsequently they were transferred into a tightly sealed glass chamber (220 × 340 × 550 mm) with a total volume of 41.3 l. 1-MCP was obtained as SmartFresh (Agrofresh Inc/Rohm and Haas, Spring House, USA), containing 0.14% 1-MCP (w/w). Then, 1.32 g SmartFresh was put into a glass beaker that was placed into the glass chamber. After the chamber was tightly sealed, 100 mL water was added via a silicon tube and a syringe from the exterior of the glass chamber to prevent any loss of 1-MCP gas. 1-MCP gas was released through the reaction of SmartFresh powder with water, yielding a concentration of 45 µL L−1 within the glass chamber (Ma et al. 2003). To guarantee a homogeneous distribution of the 1-MCP gas, we placed an 8 cm fan into the glass chamber. Every 24 h, the glass chamber was opened to mark the position of the primary root tips at the rear of each Petri dish. 1-MCP treatments were immediately renewed after each measurement. The same set-up was utilized for the controls, although the 1-MCP was not placed into the glass beaker.

Plant genotypes and cultivation system for JA experiments

A slightly different plant cultivation and growth monitoring system was used for analysing the effects of JA on root growth. Here, it was necessary to wound leaves and to monitor root growth with higher temporal resolution. In this so-called ‘microrhizotron’ set-up, which is described in more detail elsewhere (Nagel, Schurr & Walter 2006; Hummel et al. 2007), shoots grew outside the Petri dish, which was almost completely filled with Phytagel. After the Phytagel had cooled, three small holes were melted into the upper border of the Petri dish; immediately after piercing, 2-day-old sterile seedlings were transplanted into each hole. Holes were sealed with sterilized silicon fat (Baysilone-Paste, Bayer, Leverkusen, Germany) to prevent contamination. Petri dishes were then placed at an 85° angle to ensure that roots grew along the bottom. Plants were grown in the same cultivation room as described above under 26 °C during the light phase and 22 °C during the dark phase. They were exposed to a photon flux density of 85 µmol m−2 s−1 between 0600 and 2000 h. Petri dishes were not placed into the sealed glass chamber.

In addition to the WT plants described above, we used transgenic N. attenuata plants isogenic with the WT plants that were silenced in jasmonate biosynthesis and perception. One line was transformed with a construct carrying a fragment of the N. attenuata LOX3 gene in an antisense (as) orientation (asLOX3 plants) which accumulates about 50% of the JA that WT plants do after OS elicitation (Halitschke & Baldwin 2003). Another line (irCOI1) was transformed to express a fragment of the N. attenuata gene coding for the F-box protein, COI1, in an inverted repeat (ir) orientation and was dramatically inhibited in its ability to perceive jasmonates (Paschold et al. 2007).

High-resolution root growth monitoring

Ten days after germination, experiments were performed with plants grown in the microrhizotron set-up. Every 20 s, an image of the primary root growth zone was taken with a charge-coupled device (CCD) camera (Sony XC-ST50, Sony, Köln, Germany) having a resolution of 740 × 480 pixels, which corresponds to a total area of 4.4 × 2.9 mm. Infrared illumination (λ = 940 nm) enabled images to be acquired during the dark phase as well. The camera was equipped with a low-pass infrared filter (RG, Schott, Mainz, Germany) to block visible irradiation. Each replicate was measured for at least 24 h (one diel cycle). The root tips were followed via a tracking algorithm; this algorithm controlled a set of x-y moving stages that repositioned the entire Petri dish and hence the root tip, when it approached the border of the image field (Hummel et al. 2007). Root tip velocities (VTip) were calculated via image sequence processing algorithms described in more detail elsewhere (Schmundt et al. 1998; Walter et al. 2002; Walter, Feil & Schurr 2003). Root growth data were normalized by dividing the VTip values with VTip at the time point of treatment (1300 h) to make treatments easier to compare. Temperature of the Phytagel of one representative dish was recorded once per minute using a thermocouple attached to a data logger (Delta-t Devices Ltd, Cambridge, England). In all experiments, roots were exposed to light during the day. This surely affected root growth. However, all results were compared with control plants which were exposed to the same conditions, making sure that effects induced by treatments were not masked by root growth reduction due to the light reaching the roots. Moreover, in an earlier study, the effect of light on root growth in this special cultivation system was tested explicitly by comparing root growth in a light-shielded set-up with root growth in an uncovered set-up (Nagel et al. 2006; fig. 4 of this reference). As long as shoots were growing outside, even when a light intensity of 300 µmol m−2 s−1 was applied to the root, only minor reductions of root growth compared with the shaded situation were observed. Yet, strongly decreased root growth rates were observed when seedlings were grown in the way that is still common practice in other labs, with shoots inside the Petri dish. Then, also shoot growth was reduced and shoot morphology strongly altered, suggesting that under these circumstances, growth-inhibiting amounts of ethylene are produced (Eliasson & Bollmark 1988). As light intensity applied here was fairly low and as the fraction of light reaching the root was almost negligible due to total reflection of most of the light coming from above and being reflected from the almost vertical bottom and lid areas of the Petri dish, we chose not to shield roots from ambient light, hereby improving temperature control of the seedling and ease of handling during experimental manipulations.

Wounding and OS-elicitation treatments

To simulate herbivory, both primary leaves were squeezed with tweezers to produce small puncture holes. By this procedure, a maximum of 2% of the total leaf area (including the cotyledons) was damaged. Immediately after wounding, 1 µL H2O (buffered to pH 7.8 by 50 mm phosphate buffer; H2O) or 1 µL M. sexta larval OS (diluted 1:5 with phosphate buffer) was applied to the wounds. For the 1-MCP experiment, wherein shoots were grown inside the Petri dishes, the Petri dishes were opened and the treatments were performed within sterile clean benches to avoid any contamination.

In methyl jasmonate (MeJA) experiments, 500 ng (MeJA, Sigma-Aldrich, Steinheim, Germany) dissolved in 1 µL lanolin paste (Sigma-Aldrich) was applied to the undamaged primary leaf. Control plants received the same amount of pure lanolin paste only. All treatments were performed at 1300 h.

JA analysis

Eight to 10 seedlings (entire shoots) were pooled to produce approximately 150 mg samples per replicate. These pooled samples were immediately shock frozen in liquid N2 after harvest. For JA extraction, plant material was transferred to FastPrep tubes containing 900 mg of FastPrep matrix (BIO 101, Vista, USA). One millilitre of ethyl acetate mixed with 100 ng of D2-JA was used as an internal standard for JA analysis and was added to each sample. Samples were then homogenized on a FastPrep homogenizer (Savant Instruments, Holbrook, USA). After centrifugation at 12 100 g for 20 min at 4 °C, extraction was repeated with 1 mL ethyl acetate. The supernatants were combined and then evaporated to dryness on a vacuum concentrator. The dried residue was dissolved in 300 µL 70% (v/v) methanol and subsequently centrifuged at 12 100 g for 10 min.

Measurements were conducted on a 1200 L liquid chromatography–triple quadrupole mass spectrometry system (Varian, Palo Alto, USA). At a flow rate of 0.1 mL min−1, 15 µL of each sample was injected onto a ProntoSIL column (C18; 5 µm, 50 × 2 mm, Bischoff, Germany) attached to a pre-column (C18, 4 × 2 mm, Phenomenex, USA). A mobile phase composed of solvent A (0.05% formic acid) and solvent B (0.05% formic acid in methanol) was used in a gradient mode for the separation. The mass spectrometer was operated in a negative electrospray ionization mode. The most abundant and characteristic fragment ion was chosen for quantification (Wu et al. 2007).

Nicotine analysis

Shoots from the 1-MCP treatment were harvested and immediately shock frozen in liquid N2. Plant material was ground frozen with a pestle in a 1.5 mL tube. The samples were extracted with 400 µL of 40% MeOH (v/v) containing 0.5% acetic acid (v/v) and vigorously shaken at 23 °C for 2 h. After extraction, the samples were centrifuged (10 min, 12 100 g), filtered (0.45 µm) and transferred into a vial for HPLC analysis. Nicotine was separated by a Merck-Hitachi HPLC system (Darmstadt, Germany) on a Multosphere column (120 RP 18-HP 3 µm, 250 × 4 mm, CS-Chromatographie Service, Langerwehe, Germany) and detected with a diode array detector at 254 nm as described in Hummel et al. (2007).

Statistical analysis

Data were analysed with Statistica, version 6.0 (StatSoft Inc, Tulsa, USA). A repeated measures two-factorial analysis of variance (anova) was used to analyse VTip of the 1-MCP experiment, while a one-factorial anova was used to detect the first significant difference in time between MeJA-treated and control plants. For the kinetics of the nicotine and JA concentrations, a two-factorial anova was performed. All anovas were followed by Fisher's protected least significant difference test for post hoc comparisons.


Root growth when perception of the OS-elicited ethylene is inhibited

Experiments were performed in which M. sexta OS was applied to leaf wounds and in which 1-MCP was applied to inhibit the plants' perception of the wound- and OS-induced ethylene bursts ( Figs 1 & 2). Prior exposure of plants to 1-MCP led to an overall increase of root growth in all treatments (Fig. 1): while the velocity of 1-MCP-treated roots increased from 4 to about 8 mm d−1 within 1 d (Fig. 1b), untreated plants showed a slight increase to only about 5.5 mm d−1, which is a clear developmental effect. As seedlings were grown with shoots inside the Petri dish in this experiment, the difference between root growths at day 1 in Fig. 1 depicts clearly how strong ethylene – produced in the commonly accepted cultivation system with seedlings completely enclosed in Petri dishes – reduces plant growth performance.

Figure 1.

Root growth of Nicotiana attenuata after pre-treatment with 1-methylcyclopropen (1-MCP) (day 0) and following simulated herbivory [oral secretion (OS) elicitation performed on day 1]. Primary leaves were wounded with tweezers and immediately supplied with 1 µL water (H2O) or OS and regurgitant of Manduca sexta (OS) diluted 1:5 with phosphate buffer. (a) Control plants. (b) 1-MCP treated plants (mean values ± SE; n = 10).

Figure 2.

Fresh mass of Nicotiana attenuata after 1-methylcyclopropen (1-MCP) pre-treatment and simulated herbivory (harvested 3 d after 1-MCP application). Primary leaves were wounded with tweezers and immediately supplied with 1 µL water (H2O) or oral secretions (OS) and regurgitant of Manduca sexta (OS) diluted 1:5 with phosphate buffer. (a) Total root length (mean values ± SE, n = 9). (b) Root fresh mass (mean values ± SE, n = 7). (c) Shoot fresh mass (mean values ± SE, n = 7). Asterisks indicate significant differences between wounding treatments (*P < 0.05; ***P < 0.001).

The OS-induced specific growth effect is comparable in both untreated and 1-MCP-treated plants, respectively, and is superimposed on the developmental acceleration and ethylene-induced deceleration of root growth in the same manner: OS-elicited ‘control’ plants (Fig. 1a) showed a reduction in root growth that was transiently more pronounced than in plants treated with H2O. Exactly the same reaction pattern was observed in plants treated with 1-MCP (Fig. 1b): wounding led to a decrease of root growth compared with non-wounded control plants and this decrease was more pronounced for 2 d if OS rather than buffer was applied to the wounds and the differences were all highly significant (P < 0.001). This demonstrates that ethylene is not specifically involved in the decrease of growth after herbivore attack but suggests that it might generally suppress growth in plants cultivated within agar-filled Petri dishes. Gelling agents like Phytagel or agar increase the diffusion resistance and ethylene might accumulate within the roots, thereby reducing root growth. Moreover, the presumably higher ethylene levels in the root systems due to the cultivation system might deafen the response to ethylene. A further increase in ethylene concentration induced by herbivory might not have an additional effect.

The observations of the elongation of the primary root were consistent with the responses observed in the analysis of total root length, root and shoot fresh mass (Fig. 2): the significantly faster root growth we observed in treated 1-MCP plants compared with control plants led to significantly longer roots within 3 d (F1,58 = 96.886, P < 0.001; Fig. 2a). Non-wounded treated 1-MCP plants had significantly longer roots (26%) than control plants (P < 0.001), and H2O and OS treatments led to longer roots in plants pre-treated with 1-MCP compared with control plants (17%, P < 0.001; 16%, P < 0.001, respectively). Furthermore, 1-MCP treatments also significantly increased root fresh mass (F1,42 = 5.693, P = 0.022; Fig. 2b). The masses of non-damaged plants when pre-treated with 1-MCP (14%) were significantly higher that the masses of control plants (P = 0.026). No significant differences in shoot biomass were found between treatments (Fig. 2c), but the biomass of plants treated with 1-MCP tended to be higher than the biomass of non-treated plants.

De novo nicotine biosynthesis was significantly increased in both wounding treatments compared with the control – with and without application of 1-MCP (F2,25 = 42.393, P < 0.001; Fig. 3). When OS was applied to wounds, more nicotine was found in plants treated with 1-MCP than in non-treated plants (1.17-fold difference; P = 0.01), consistent with the observation that OS-elicited ethylene attenuates the de novo biosynthesis of nicotine (Kahl et al. 2000; Winz & Baldwin 2001).

Figure 3.

Nicotine concentrations of Nicotiana attenuata after 1-methylcyclopropen (1-MCP) pre-treatment and simulated herbivory (harvested 48 h after wounding) (FM: fresh mass). Primary leaves were wounded with tweezers and immediately supplied with 1 µL water (H2O) or oral secretions (OS) and regurgitant of Manduca sexta (OS) diluted 1:5 with phosphate buffer (mean values ± SE; n = 5). Different letters indicate significant differences at P < 0.05 level.

Root growth responses to OS elicitation in plants with inhibited JA responses

As shown above (Fig. 1a), a singular application of OS led to a transient reduction of root growth in plants that were OS-elicited compared with plants that were wounded but only treated with buffer. This transient reduction, which was not mediated by ethylene (Fig. 1b) and lasted approximately 1 d, had already been shown in an earlier study (Hummel et al. 2007). Hence, we wondered if, at a high temporal resolution, JA affects root growth during the 24 h immediately following wound treatment, in the same way that wounding induced a JA burst that was dramatically amplified when OS was applied to wounds ( Fig. 4). Root growth was strongly correlated with the temperature of the Phytagel (Fig. 5a). Temperature strongly affects root growth and it was previously demonstrated that root growth is not endogenously controlled by diel growth patterns (Pahlavanian & Silk 1988; Walter et al. 2002; Hummel et al. 2007). Hence, the diel fluctuations affecting root growth in all treatments can be considered to be caused by the prevailing temperature regime. The time series analyses of WT, asLOX3, and irCOI1 plants were similar (Fig. 5), with the strongest differences found between 1600 and 2000 h when irCOI1 plants grew slightly more slowly than WT plants.

Figure 4.

Jasmonic acid concentrations in seedlings of Nicotiana attenuata after simulated herbivory (FM: fresh mass). Primary leaves were wounded with tweezers and immediately supplied with 1 µL water (H2O) or oral secretions (OS) and regurgitant of Manduca sexta (OS) diluted 1:5 with phosphate buffer and harvested from 0 to 90 min after wounding (mean values ± SE; n = 3–5). Different letters indicate significant differences at P < 0.05 level.

Figure 5.

Normalized values of root growth velocities (VTip, norm) conducted with high resolution from different jasmonic acid-mutants of Nicotiana attenuata after simulated herbivore attack. Primary leaves were wounded with tweezers and immediately supplied with 1 µL water (H2O) or oral secretions (OS) and regurgitant of Manduca sexta (OS) diluted 1:5 with phosphate buffer. (a) Temperature course of Phytagel over 24 h during the experiments. Light was switched on at 0600 h and switched off at 2000 h (night period gray shaped). (b) Wild-type plants (WT, mean values ± SE, n = 15). (c) Antisense-lipoxygenase 3 (asLOX3) plants (mean values ± SE, n = 12). (d) Inverted repeat-coronatin-insensitive 1 (irCOI1) plants (mean values ± SE, n = 5).

When WT plants were wounded, they displayed the reaction pattern indicated in Fig. 1a; a significantly stronger reaction was observed when OS was applied to wounds compared with when H2O was applied (Fig. 5b). Root growth in OS-treated WT plants was 33% that of control plants 7 h after wounding (just before dark phase; Fig. 5b). Root growth in H2O-treated plants was 51% that of control plants 7 h after wounding (Fig. 5b). asLOX3 plants, which display a 50% reduction in JA levels, showed very similar reaction patterns (Fig. 5c): root growth was reduced stronger in OS-treated plants compared with H2O-treated plants. In comparison to control plants, growth was reduced to 61 and 44% of that in H2O-treated and OS-treated plants 7 h after wounding, respectively.

In irCOI1 lines, wounded plants of both treatments reached VTip-values of 60% of control plants at the end of the day (Fig. 5d). Differences between H2O- and OS-treated plants were not significant, indicating that JA perception is crucial for the specific root growth reduction observed in OS-treated plants to take place.

Independent of the genotype, an immediate decrease of VTip during 65 ± 4 min after wounding was observed with values of 64 ± 8% of the initial VTip, independent of the specific wounding treatment or the genotype (Fig. 5). This immediate ‘first’ decrease was followed by a phase of steady VTip-values and remained relatively stable for 57 ± 10 min when averaged over all treatments. Thereafter, a ‘second’ decrease in root growth was observed, lasting several hours until the end of the day. This second decrease is primarily responsible for the growth reductions described above.

High-resolution root growth analysis after MeJA application

To elucidate MeJA-elicited changes in root growth, MeJA was applied to primary leaves and root growth was measured at high temporal resolution (Fig. 6a). MeJA dramatically reduced root growth and a first significant difference between VTip-values of control and treatment plants was observed 97 min after MeJA application (P = 0.028). After this point in time, the VTip of MeJA-treated plants was significantly reduced throughout the entire measurement period, which is best displayed by depicting the ratio of VTip of MeJA-treated and VTip of control plants (Fig. 6b). This ratio decreased almost linearly from wounding until the beginning of the night, reaching values between 40 and 50% throughout the night.

Figure 6.

Normalized values of root growth velocities (VTip) of Nicotiana attenuata after treatment with 500 ng methyl jasmonate (MeJA) dissolved in 1 µL lanolin paste. (a) 24 h overview. Light was switched on at 0600 h and switched off at 2130 h (night period gray shaped). (b) VTip-Ratio : VTip of MeJA-treated plants divided by VTip of control plants (mean values ± SE, n = 4).


Inhibiting ethylene perception increases root growth but not the response to OS elicitation

Mechanical wounding of N. attenuata leaves transiently increases endogenous JA and ethylene levels, and when the OS of the specialist lepidopteran larvae M. sexta is applied to the wounds, these transient increases rise dramatically (Kahl et al. 2000; Schittko, Preston & Baldwin 2000). Ethylene is known to reduce root elongation considerably (Abeles et al. 1992; Tholen, Voesenek & Poorter 2004; Visser & Pierik 2007). In Petri dish cultivation systems, it can occur in high concentrations and lead to strong root growth reduction (Buer, Wasteneys & Masle 2003). As our results show (Figs 1 & 2), 1-MCP alleviates this reduction significantly. These are consistent with results from Arabidopsis and Epipremnum (Muller et al. 1997; Ma et al. 2003), where it was shown previously that 1-MCP leads to an increase in root growth. Similar results were obtained in Brassica napus by using another ethylene receptor inhibitor (1-cyclopropenylmethyl; Saleh-Lakha et al. 2004). The reduction of root growth due to OS (Hummel et al. 2007) was still observed when 1-MCP was applied, clearly indicating that the wound-induced ethylene burst is not inducing this reaction.

In N. attenuata, wounding or elicitation with MeJA induces de novo biosynthesis of nicotine used for defence (Steppuhn et al. 2004); however, other resistance traits are regulated by JAs, such as trypsin protease inhibitors, diterpene glycosides and volatile emissions involved in indirect defence (Halitschke et al. 2000; Steppuhn et al. 2004). We tested whether ethylene alters nicotine concentrations of our seedling system after wounding, as was previously demonstrated for N. attenuata at later developmental stages by Kahl et al. (2000). The results clearly demonstrate that ethylene attenuates the nicotine response when OS is applied to wounds, suggesting that the mechanism reported by Kahl et al. (2000) and Winz & Baldwin (2001) is already active in the seedling stage of N. attenuata. The ecological interpretation of this attenuated nicotine response is reviewed in Kessler & Baldwin (2002). They hypothesize that the M. sexta-induced ethylene burst adaptively reduces nicotine production, as the larvae is alkaloid tolerant and can use increased levels of nicotine for its own defence against parasitoids or predators.

Inhibition of JA-signalling alters the OS-elicited root growth response

JA up-regulates the expressions of many constitutively expressed defence genes but down-regulates many genes involved in growth (Creelman & Mullet 1997; Halitschke et al. 2003). Hence, we tested whether transgenic N. attenuata plants, impaired in their JA-signalling pathways, display altered root growth in response to OS elicitation. We first monitored root growth without any damage treatment and found that asLOX3 and WT plants showed no difference concerning the diel variation of root growth activity, which is strongly affected by temperature as previously demonstrated by Walter & Hummel (2008).

LOX3 is the specific wound and herbivory-induced isoform of the LOX gene family in N. attenuata, which is involved in the JA biosynthesis. asLOX3 plants are impaired in the wound-induced biosynthesis of JA but are unaltered in their unelicited constitutive levels (Halitschke & Baldwin 2003). Hence, unsurprisingly, root growth is comparable between WT and asLOX3 seedlings without wounding (Fig. 5). This result contrasts the results of Zavala & Baldwin (2006), who reported that the growth and fitness of asLOX3 plants was better than that of undamaged WT plants. However, their plants were glasshouse grown in soil-filled pots. Environmental stresses (temperature, light) or biotic stresses (elicitation by soil microbes) are also mediated via the JA pathway, which might have down-regulated growth in WT plants (Creelman & Mullet 1997; Pozo, Van Loon & Pieterse 2004; Yadav et al. 2005).

COI1, an F-box protein, is required for the perception of JA, and several studies have demonstrated that root growth of COI1-deficient plants is relatively insensitive to MeJA treatments (Feys et al. 1994; Xie et al. 1998; Li et al. 2004; Paschold et al. 2007). Diel root growth in irCOI1 plants was similar to that in WT and asLOX3 plants; however, in the afternoon, irCOI1 plants grew more slowly than did WT plants, suggesting that COI1 might play a role in controlling root growth at least during the afternoon.

After wounding, we observed in all genotypes an immediate and steep reduction of root growth for 65 min until growth temporarily stabilized again, as previously described in Hummel et al. (2007). This first rapid decrease points to a hydraulic response as a loss of water and turgor pressure immediately reduces root growth (Nagel et al. 2006); however, the immediate reduction in root growth could also have been mediated via electrical signalling (Fromm & Lautner 2007). A second negative growth response was observed approximately 2 h after wounding, which coincides remarkably well with the ‘decreasing’ root growth dynamics of MeJA-treated plants (Fig. 6a). This suggests that the second decrease in root growth might be mediated via wound-elicited JA, imported from the wound sites in leaves to the roots. Baldwin et al. (1997) reported that JA pools in roots of N. sylvestris increase 3.5-fold 180 min after mechanical leaf wounding, and Zhang & Baldwin (1997), by exogenously applying 2-14C-labelled JA to leaves, demonstrated that JA is directly transported from the leaves to the roots and could account for the systemic increase in root JA pools after leaf wounding. Moreover, Thorpe et al. (2007) reported a rapid transport of 11C-labelled MeJA from leaves to roots in N. tabacum within 60 min. The intensity with which root growth is reduced the second time depends on the specific wounding treatment and the genotype. The root growth of WT plants was more strongly reduced when OS was applied to wounds than when seedlings were only mechanically wounded, confirming previous results of Hummel et al. (2007). The root growth of asLOX3 plants was less reduced compared with that of WT plants in both wounding treatments; nevertheless, the application of OS also led to strong reductions in root growth. These results are consistent with the endogenous JA concentration levels of asLOX3 plants after wounding, which are also amplified by the application of OS, although not to the same extent as in WT plants (Halitschke & Baldwin 2003). In this study, the wounding treatments were applied in a single event, having the character of a pulse-chase experiment. In nature, herbivores are continuously feeding on leaves, which can be considered as a series of mechanical wounds resulting in a permanently elevated JA content. Thus, root growth might be reduced as long as the herbivore is feeding. This would strongly affect root system development, nutrient acquisition and the competitive strength of the plant. Especially in highly productive environments, intraspecific competition results in high fitness costs for attacked plants compared with their unattacked conspecifics (van Dam & Baldwin 1998, 2001). However, a constantly elevated level of JA induced through herbivory might be integrated differently by the root system and could trigger different root growth reactions. A detailed analysis of the relation between the signal duration and the response would be interesting in future studies.

In irCOI1 lines, wounding leads to an immediate hydraulic response; however, the second response is clearly less intense compared with WT and asLOX3 plants. Only negligible differences in root growth of OS- and H2O-treated plants were observed, suggesting that irCOI1 plants are insensitive to wound-induced JA changes in the plants. This confirms findings by Paschold et al. (2007), who directly applied different concentrations of MeJA to roots and showed that irCOI1 plants are remarkably growth-insensitive to JA. These results provide further evidence that (1) the second decrease in root growth is mediated by JA; and (2) the more pronounced reduction of root growth after application of OS to wounds results from the higher OS-elicited JA levels.

To conclude, our study demonstrates that JA signalling is involved in mediating root growth inhibition and it demonstrates how the shoot may regulate root growth reduction. Whether this growth reduction is a direct response to JA or whether it is triggered by another hormone interacting with JA is a matter of future research. Yet, evidence is provided that endogenous JA acts as a distress signal, slowing vegetative growth during defence responses.


We would like to thank Eva Rothe, Jinsong Wu and Norbert Kirchgessner for their assistance. We are also grateful to Jon F. Fobes (AgroFresh Inc) for the generous gift of 1-MCP. We would like to acknowledge stimulating discussions with all members of the Virtual Institute for Biotic Interactions. This work was financially supported by the Helmholtz Association of German Research Centers.