Cues from a specialist herbivore increase tolerance to defoliation in tomato

Introduction

Plants have evolved adaptations to minimize negative effects and to maximize recovery in response to a variety of biotic and abiotic stresses (Rosenthal & Janzen 1979). While most plant species respond to herbivore damage by producing defence compounds that directly and negatively affect the performance of herbivores, many species have also evolved tolerance traits that increase regrowth capacity and/or reproduction (reviewed by Karban & Baldwin 1997; Stowe et al. 2000; Tiffin 2000; Fornoni 2011; Orians, Thorn & Gomez 2011). Tolerance may be a particularly important strategy against specialist herbivores that have evolved to either tolerate or detoxify the plant's direct chemical defences (Karban & Baldwin 1997; Strauss & Agrawal 1999), as is the case for Manduca sexta, a specialist on Solanaceae (Wink & Theile 2002).

In general, a plant's capacity to tolerate damage largely depends upon its inherent growth rate and extent of stored resources prior to the damage event (van der Meijden, Wijn & Verkaar 1988; Huhta, Tuomi & Rautio 2000b; Huhta et al. 2000a; Stowe et al. 2000; Tiffin 2000; Wise & Abrahamson 2005; Schwachtje & Baldwin 2008; Fornoni 2011; Hochwender et al. 2012; Zhao & Chen 2012). These are considered constitutive traits. Additionally, there are herbivore-inducible mechanisms that can increase tolerance, such as increase in photosynthetic rates, activation of dormant meristems and compensatory growth (reviewed in Tiffin 2000). More recently, it has been hypothesized that plants can also increase their tolerance to herbivores by reallocating resources away from tissues under attack and using those resources for future regrowth after the herbivory event. Evidence of herbivore-induced resource reallocation includes increased transport of recently fixed carbon to roots and/or stems in response to both herbivores and jasmonates (Holland, Cheng & Crossley 1996; Babst et al. 2005, 2008; Schwachtje et al. 2006; Henkes et al. 2008; Kaplan et al. 2008; Gomez et al. 2012; Moreira, Zas & Sampedro 2012; Poveda et al. 2012; Ferrieri et al. 2013). Similarly, jasmonic acid treatment of roots modifies carbon partitioning in Nicotiana (Henkes et al. 2008). Remobilization of nitrogen in response to (simulated) herbivory has also been shown (Newingham, Callaway & BassiriRad 2007; Gomez et al. 2010). For example, Newingham, Callaway & BassiriRad (2007) observed a shift in allocation of nitrogen towards the leaves after root herbivory in Centaurea maculosa. Some evidence suggests that these herbivore-induced resource shifts might support storage in inaccessible tissues. In Populus, for example, jasmonates increase protein concentration in stem tissue (Beardmore, Wetzel & Kalous 2000). This storage role of reallocated resources might support increased regrowth after the attack, thereby increasing tolerance to herbivory. Additionally, resource shifts can also support defence production in local and systemic tissues (Arnold et al. 2004; Ferrieri et al. 2013).

In the tomato system specifically, we have previously shown herbivory-induced resource allocation away from treated leaves and into storage organs (Gomez et al. 2010, 2012) along with rapid changes in the concentrations of primary metabolites locally and systemically (Steinbrenner et al. 2011; Gomez et al. 2012). Interestingly, cues of the specialist herbivore, Manduca sexta, largely induced metabolites associated with carbon and nitrogen transport while cues of a generalist herbivore, Helicoverpa zea, induced many more metabolites that are precursors of defence induction (Steinbrenner et al. 2011). This suggests that plants might cope with some specialist herbivores by enhancing regrowth capacity instead of direct chemical defence; defences that tend to have less impact on specialist herbivores. Perhaps when cues signal the potential for extensive defoliation, not uncommon among specialist herbivores, plants undergo a rapid and specific response that minimizes the loss of resources to the herbivore and maximizes the amount of resources available for later regrowth. Despite the increasing evidence that invertebrate herbivory triggers resource reallocation, whether such changes result in improved tolerance expressed as increase regrowth and flowering production is underexamined (but see Schwachtje et al. 2006).

Here, we examined whether cues of M. sexta increase the regrowth capacity and initial flower production of a cultivated tomato variety. First, we hypothesized that during a five-day damage treatment period, plants in the simulated herbivory treatment (damage + M. sexta regurgitant) would exhibit a decrease in growth compared with mechanically damaged plants (damage + water) or undamaged control plants. This would be consistent with our previous work demonstrating shifts in resource allocation from leaves to stems and roots in tomato. Secondly, we hypothesized that, if induced sequestration of resources increases tolerance, subsequent defoliation would result in faster and more regrowth and greater initial flower production in response to herbivore cues as compared to mechanically damaged or undamaged control plants.

Materials and methods

Induction Treatments

A total of 45 plants were divided equally among three experimental groups: an undamaged control group and two damage treatments. In the two damage treatments (hereafter referred to as ‘induction treatments’), three leaves (leaves 2, 3 and 4, counting down from the apex) were damaged for five consecutive days (Fig. 1; Induction phase). This time frame was chosen to mimic herbivores with limited movement feeding on the same plant for several days. Previous experiments have shown that this amount of time elicits quantifiable changes in tomato's primary metabolism (T. Korpita, unpublished data). Leaf 1 was defined as the newest leaf that was over 50% expanded. A fabric trace wheel was used to make puncture wounds along the edge of every leaflet of each of the three leaves. After damage, 80 μL of either Manduca sexta regurgitant or deionized water were applied with a paintbrush (‘Manduca’ and ‘Water’ treatment, respectively). Caterpillar regurgitant is well known to elicit a specific response compared with mechanical damage (Kessler & Baldwin 2002; Halitschke et al. 2003). Caterpillar regurgitant was collected from fifth instar M. sexta larvae fed on tomato plants for at least 48 h and stored at −80 °C until use. The water treatment was designed to test for the effects of mechanical damage alone.

Results

Leaves in all treatments experienced a decline in chlorophyll content during the induction phase. However, the magnitude of the decline differed among treatments as indicated by the overall significant induction treatment effect (one-way manova; Wilks' λ = 0·662; F_6,80 = 3·06; P = 0·01; Fig. 2a). This was driven by a significant induction treatment effect on leaf 2 (young fully expanded leaf F_2,42 = 4·15; P = 0·02) and a marginally significant effect (F_2,42 = 3·01; P = 0·06) on leaf 0, the youngest expanding leaf. The chlorophyll content in leaf 2 of plants treated with M. sexta regurgitant declined significantly more over time than in control plants. Chlorophyll in regurgitant-treated plants declined 36% by day 6, whereas that in undamaged control plants it declined 23%. Chlorophyll content in mechanically damaged plants treated with water was not significantly different to either control or regurgitant-treated plants (Fig. 2a). In leaf 0, this decline was 28% in regurgitant-treated plants compared with 16% in control plants (Fig. 2a). However, induction did not affect chlorophyll content on leaf 5 (mature, fully expanded, undamaged source leaf (F_2,42 = 2·27; P = 0·12)).

There was a strong effect of the induction treatment on plant stem height: width growth ratio after 5 days of damage (F_2,32 = 8·97; P < 0·001). Control, undamaged plants had a higher ratio, driven by greater apical growth and reduced stem diameter (Fig. 2b). On average, by the end of the induction treatment, plants treated with M. sexta regurgitant were about 33% shorter but had a 222% thicker stem than undamaged control plants. Mechanically damaged plants treated with water did not significantly differ from control or regurgitant-treated plants in apical growth (height) or stem width growth.

Damage prior to complete defoliation (induction treatment) had a significant and positive effect on regrowth capacity and initial flower production. Plants previously challenged with Manduca regurgitant showed the greatest regrowth capacity, producing overall more leaves than control, undamaged plants (Fig. 3; Induction treatment effect F_2,27 = 5·17; P = 0·01). This effect was visible soon after the defoliation treatment was applied. Specifically, 1-day post-defoliation, plants in the two damage treatments had more new leaf buds than undamaged plants (Fig. 3). Moreover, the damage effects on the amount of leaves produced post-defoliation did not vary over time (no time by treatment interaction; Fig. 3). At harvest, day 23 post-defoliation, plants in the Manduca treatment had an average of 23·8 ± 1·1 regrown leaves compared with 21·9 ± 0·8 and 20·2 ± 1·4 regrown leaves in the Water and Control treatments, respectively (Fig. 3).

There was no significant effect of the induction treatment on total plant biomass. Interestingly, while the biomass M. sexta regurgitant-treated plants tended to be lower than controls at the end of the induction treatment (day 6), they were higher at the final harvest (23 days post-defoliation; Table 1). In addition, there was no significant induction treatment effect on the relative allocation to vegetative tissues.

The induction treatment significantly affected initial flower production at harvest (Fig. 4; F_2,27 = 3·93; P = 0·03). At the time of harvest, plants treated with Manduca regurgitant had produced an average of 3·6 flowers per plant, while undamaged controls produced 2·0 flowers per plant; initial flower production on plants mechanically damaged (treated with water) showed intermediate values (2·8 flowers per plant, Fig. 4).

Discussion

Our results show that simulated herbivory by the specialist M. sexta elicited rapid changes in a tomato cultivar resulting in greater regrowth capacity and initial flower production after a severe defoliation event. The effect of mechanical damage alone was similar but weaker than those elicited by herbivore cues. To our knowledge, this is one of the first studies to show that herbivore cues can elicit a change in resource allocation that may ultimately increase a plant's tolerance to herbivory. There are several implications of our results.

First, herbivore cues caused a shift in the pattern of growth that could be linked to the increase in tolerance. Although both leaf chlorophyll and apical growth declined during the induction phase, stem width growth did not. The observed decline in chlorophyll content in younger leaves [the main pool of leaf nitrogen (Evans 1989)] and the tendency for less biomass accumulation following treatment with M. sexta regurgitant support a shift in allocation patterns away from immediate growth (Gomez et al. 2010, 2012). This pattern is consistent with the general observation that younger leaves are more inducible (Arnold et al. 2004; Gutbrodt et al. 2011). Additionally, plants in the Manduca treatment exhibited a larger increase in basal stem diameter than did mechanically damaged or control plants. Perhaps, storage of carbon and proteins in the basal part of the stem saved those resources from M. sexta and made them available for subsequent regrowth after defoliation. Importantly, we note that the effects can be subtle; even though there was no significant change in total biomass or relative biomass allocation among the treatments, the small changes in basal stem diameter were associated with significant differences in the timing of regrowth (see discussion below). To our knowledge, no previous work has examined this issue.

Historically, decreases in apical growth following herbivory have been assumed to be related to allocation costs of plant defence traits (Green & Ryan 1972; Herms & Mattson 1992; Karban & Baldwin 1997; Zavala & Baldwin 2006). While this may be a component of the reduced apical growth during the induction phase, our results suggest that greater allocation to stem tissue may have contributed to this pattern. Importantly, given the increasing evidence for short-term resource sequestration induced by herbivores (reviewed by Orians, Thorn & Gomez 2011), more attention should focus on how leaf-feeding herbivores alter patterns of resource allocation to growth, storage and defence.

Secondly, we showed that changes in resource allocation caused by herbivory alter a plant's regrowth capacity following defoliation, indicating that improving short-term tolerance is a part of the integrated plant responses induced by herbivory. Plants that had been exposed to regurgitant by the specialist M. sexta were clearly able to recover from the loss of all their leaf tissue better than previously undamaged plants. We note that this effect was observed even 1-day post-damage, suggesting that herbivore cues were altering patterns of apical dominance. Not only did they produce new leaves at a faster rate, but also had higher initial flower production. Plant tolerance to leaf damage by herbivores is well studied (Paige 1999; Strauss & Agrawal 1999; Tiffin 2000). Traditionally, it has been viewed as a function of stored reserves and release from apical dominance through increased photosynthesis and growth of new tissues following the loss of the apical meristem (Tschaplinski & Blake 1989a,b; Huhta, Tuomi & Rautio 2000b; Tiffin 2000). A rapid herbivore-induced shift in growth during or just after the actual herbivory event may be another important mechanism of tolerance. This would not require an investment of resources in the absence of herbivory, as would other tolerance characteristics, such as constitutive storage. Based on work in tomato and other systems (Holland, Cheng & Crossley 1996; Schwachtje et al. 2006; Babst et al. 2008; Kaplan et al. 2008; reviewed by Erb et al. 2009; Anten & Pierik 2010), we conclude that the mechanisms of increased tolerance in response to M. sexta are likely driven by changes in carbon and nitrogen allocation to sites of short-term storage (especially sites near sites of new shoot growth) and by hormonal changes that affect patterns of apical dominance.

Finally, we suggest a well-tailored response to the type of herbivory simulated in this study is relevant in natural settings provided that plants have time to regrow and produce seeds before the season ends. Some herbivores, like M. sexta, can defoliate entire plants once they reach later instars. Similarly, gregarious feeders can also cause considerable damage during later developmental stages. We predict that these herbivores will trigger changes in growth that will increase regrowth capacity once the threat is no longer present (Orians, Thorn & Gomez 2011). Perhaps the observed induced mechanisms to increase tolerance are a common response to specialist herbivores that are well adapted to their host plant's defence compounds. Not only are they capable of detoxifying or tolerating the defence chemicals of their host plants (as is the case with M. sexta on Solanaceae) (Wink & Theile 2002), but also they are likely to continue feeding on the same plants (Wittstock et al. 2004) and thus provide a reliable cue to the plant. In general, induced secondary defences are expected to be more important against generalist herbivores (Orians, Thorn & Gomez 2011). In addition, an induced tolerance response may be important in plants subject to periodic insect outbreaks if plants have the appropriate anatomical structures for storage of resources. Specifically, we predict that plant species that experience periodic outbreaks (Barbosa, Letourneau & Agrawal 2012; Zvereva, Zverev & Kozlov 2012) will, in response to cues of those herbivores (perhaps above a critical threshold), undergo a metabolic reorganization that increases resource allocation to inaccessible tissues and failure to respond will compromise regrowth and other fitness measures when outbreaks do occur.

In summary, these results provide a new perspective on the induced responses of plants, and to our knowledge, provide the first evidence of specialist herbivore cues increasing regrowth after a defoliation event. While this induced mechanism of tolerance may potentially be ecologically relevant, it is important to note that this study was conducted on a domesticated cultivar grown in the greenhouse with regular fertilization. Future studies should focus on determining whether this is a common response to specialist herbivory across systems, in both wild and cultivated plant species across a range of ecologically relevant environments. They should also elucidate the details of the underlying mechanisms that contribute to the documented increase in tolerance. A study that simultaneously documented post-herbivory changes in resource allocation to growth, defence and storage would go a long way towards determining the relative significance of this induced response.