The seedling system
The results demonstrate clearly that N. attenuata shows inducible defence in very early developmental stages. The application of MeJA induced strongly increased whole plant nicotine concentration, hypocotyl trichome length and trichome density within 4 d. It has been pointed out that PIs are not inducible in seedlings of N. attenuata, but only in later developmental stages (van Dam et al. 2001), indicating that not the entire suite of defence mechanisms is active in seedlings and is hence constrained by plant ontogeny. This fact is also supported by the finding that in Nicotiana sylvestris, nicotine induction after wounding only occurs during the rosette stage, but not during the elongation or flowering stages (Ohnmeiss & Baldwin 2000).
The induced increase of hypocotyl trichome length and density may play an important role for repelling herbivores from seedlings as here, the hypocotyl still represents a relatively high fraction of the plant tissue. Hitherto, increasing trichome densities have only been reported on a larger timescale for leaves emerging after herbivore attack (Agrawal 1999; Traw & Dawson 2002; Traw & Bergelson 2003). Leaves that are already differentiated by the time of herbivore attack are not capable to strengthen their physical defence systems as this can only be performed in developing and non-mature tissues (Mayers & Bazely 1991; Nagata et al. 1999). Clearly, the hypocotyl epidermis cells are still able to differentiate into trichomes, as this tissue is displaying secondary growth. Hence, hypocotyl trichome density may represent a powerful monitoring trait for non-invasive detection of defence induction in similar experiments.
The dynamics of nicotine and trichome induction were different for the two treatments investigated (Fig. 3). While the ‘pure’ wounding treatment (application of H2O) induced both nicotine and trichome formation rapidly, the spit treatment also led to a rapid increase of nicotine content, but trichome density was only increased 3 d after wounding. In addition to the root growth data, which will be discussed later, this may be an indication for the transiently reduced growth potential of the seedling upon herbivoral attack. If the application of spit specifically leads to a transient depletion of metabolites required for growth processes, a delayed elongation of trichomes would be a logical consequence.
It has to be pointed out that all experiments have been conducted at very low light intensity to allow high-resolution root growth monitoring. It is conceivable that at low light intensity, both growth and defence may be carbon limited, and hence it might be problematical to draw conclusions for a trade-off between growth and defence. It is well documented that seedling root growth is suboptimal under the light intensity reached in this cultivation system (Nagel et al. 2006). Yet, the fact that nicotine concentrations in greenhouse-grown seedlings and growth room-cultivated seedlings were comparable rather leads to the speculation that defence mechanisms are realized in a relatively ‘normal’ manner despite the limited availability of photosynthates. This may in turn be realized by the simple fact that the plant body itself is much smaller under low light conditions, and a low amount of nicotine or other defence-related substances consequently results in concentrations similar to those found in high-light-grown plants. A detailed analysis of the relation between light intensity and defence induction would thus surely be interesting in future studies.
Growth reaction after wounding treatments
An immediate root growth response was observed for the wounding treatments. The first rapid decrease clearly points to a hydraulic response of the plant as loosing water or turgor pressure is immediately reducing root growth (Nagel et al. 2006). Apart from the hydraulic response, there is a strong response towards the spit of Manduca larvae, which coincides with the time frame of about 2–3 h that was reported for the systemic increase of JA transported from shoot to root. Baldwin et al. (1997) reported that JA pools in roots are increased systemically (3.5-fold) within 180 min following mechanical wounding. This increase in JA pools of roots could account for the reduction in root growth. The 10-fold higher JA burst following herbivory of M. sexta larvae or the application of their spit to wounds (Kahl et al. 2000) could be responsible for a more pronounced inhibition of root growth. Seventeen hours after the spit treatment, root growth of spit- and H2O-treated plants was comparable again and had even reached the same intensity as that of control plants. Hence, it is conceivable that JA pools or ethylene concentrations in seedling roots have decreased by this time again. However, the ethylene burst elicited by Manduca or their spit may mediate a reduction of root growth as well.
Recently, Schwachtje et al. (2006) reported that a β-subunit of an SNF1-related kinase, GAL83, which is induced via herbivore specific elicitors, regulates root–shoot partitioning of carbohydrates after herbivore attack independently of JA signalling. This change in source–sink relations could in itself act as a signal, as changes in carbohydrate concentrations are known to affect gene regulation (Koch 1996; Smeekens 1998). However, an increase in carbohydrate import to roots would rather result in an increase of root growth as reported by (Nagel et al. 2006), suggesting a down-regulation of root growth via wound signalling by ethylene or JA.
Decreased import of carbohydrates due to destruction of a relevant fraction of photosynthetic tissue can lead to decreased root growth (Aguirrezabal et al. 1994; Nagel et al. 2006). Hence, the permanent destruction of 13% of the source leaf tissue during the wounding treatment may be responsible for the slow recovery of root growth throughout several days (Fig. 5) and for lower growth of H2O- and spit-treated plants compared to control plants in the morning after the treatment day (Fig. 7).
The strongest difference concerning the distribution of REGR between H2O and spit treatment was initially observed about 1 mm behind the root tip in the beginning of the central elongation zone. Here, meristematic activity has practically ceased, which is indicated by increase of REGR. This means that spit-induced growth reduction might mainly be caused by reduction of growth during an early stage of cell expansion. Eight hours later, cells from this region have been displaced towards the end of the growth zone, where now strongest differences are observed. Hence, spit might mostly affect a cell file in an early stage of cell expansion, and the root zone recovers as new cells that are formed in the meristematic part of the root growth zone at the very tip are being displaced into the elongation zone.
Unfortunately, it was impossible to perform leaf and root growth monitoring in response to simulated herbivore attack in comparable stages of plant development. Nevertheless, the reaction of leaf growth was monitored in a later stage of development to gain some insight whether growth reactions might be restricted to roots or whether they might be more pronounced in the aboveground plant part, where herbivore attacks occur and where hence a growth reduction would intuitively seem more meaningful. The results show clearly that the leaf rosette, at least in this developmental stage, shows a weaker reaction than the primary root. If there is any reduction at all towards a single wounding event, it is also of transient nature and seems to prevail for a shorter time than observed in the root. Yet, it has to be pointed out that because of the overall low leaf growth rate during large parts of the diel cycle and because of the high variability among individuals, leaf growth results should be interpreted carefully.
In summary, our results obtained on seedling plants in laboratory conditions clearly indicate that a single herbivore attack at a leaf, which is sensed via the presence of FACs, leads to an immediate growth response in the root. The growth reduction is transient and is regulated by wound-induced JA transported from leaves to roots. Following herbivore attack, the plant diverges its metabolism towards the production of defence substances and structures within hours, leaving a significant imprint in a reduction of growth. The results of Schwachtje et al. (2006) show that in a more realistic situation under higher light intensity, allocation of sugars to roots is increased upon herbivoral attack, leading not to stronger growth, but to increased C storage there. Overall, the picture emerges that not the leaf attacked aboveground, but the belowground root system turns down its spatial expansion and strengthens its function as a ‘safe retreat’ for carbon to produce reproductive growth eventually.