Elevated CO2 Reduces the Resistance and Tolerance of Tomato Plants to Helicoverpa armigera by Suppressing the JA Signaling Pathway

Both resistance and tolerance, which are two strategies that plants use to limit biotic stress, are affected by the abiotic environment including atmospheric CO2 levels. We tested the hypothesis that elevated CO2 would reduce resistance (i.e., the ability to prevent damage) but enhance tolerance (i.e., the ability to regrow and compensate for damage after the damage has occurred) of tomato plants to the cotton bollworm, Helicoverpa armigera. The results showed that elevated CO2 reduced resistance by decreasing the jasmonic acid (JA) level and activities of lipoxygenase, proteinase inhibitors, and polyphenol oxidase in wild-type (WT) plants infested with H. armigera. Consequently, the activities of total protease, trypsin-like enzymes, and weak and active alkaline trypsin-like enzymes increased in the midgut of H. armigera when fed on WT plants grown under elevated CO2. Unexpectedly, the tolerance of the WT to H. armigera (in terms of photosynthetic rate, activity of sucrose phosphate synthases, flower number, and plant biomass and height) was also reduced by elevated CO2. Under ambient CO2, the expression of resistance and tolerance to H. armigera was much greater in wild type than in spr2 (a JA-deficient genotype) plants, but elevated CO2 reduced these differences of the resistance and tolerance between WT and spr2 plants. The results suggest that the JA signaling pathway contributes to both plant resistance and tolerance to herbivorous insects and that by suppressing the JA signaling pathway, elevated CO2 will simultaneously reduce the resistance and tolerance of tomato plants.


Introduction
In the last 250 years, atmospheric carbon dioxide (CO 2 ) has risen from 280 ppm to greater than 390 ppm, and is anticipated to reach at least 550 ppm by year 2050 [1]. Because elevated CO 2 increases the carbon to nitrogen (C:N) ratio and reduces the N content in the tissue of most plant species, elevated CO 2 is expected to alter plant synthesis of phenolics, terpenes, and other secondary metabolites [2,3]. Such changes in C:N and in the content of secondary metabolites will alter the nutritional quality and palatability of host plants for herbivores and could therefore affect the performance of herbivorous insects [4].
Plants have evolved a variety of mechanisms to reduce the negative impacts of herbivory [5,6]. When damaged by herbivorous insects, plants can produce herbivore-deterrent metabolites or defensive proteins to limit the damage [7]. This kind of induced defense (i.e., resistance) is energy and resource costly, however, and cannot be maintained at high levels throughout the growing season [8]. An alternative to resistance is tolerance, which compensates for tissue loss after insect attack [9]. In expressing tolerance, plants reallocate energy and resources from undamaged to damaged tissues (for example, by increasing sucrose-transport enzymes in the damaged tissues) and increase photosynthetic rates and growth parameters [10,11]. Although researchers generally assume that there is a trade-off between resistance and tolerance (i.e., plants with high resistance have low tolerance and vice versa), the relationship between plant resistance and tolerance to herbivores varies among studies and often depends on the plant species, soil resource, and environment [12,13].
Elevated CO 2 is likely to increase constitutive levels of defensive metabolites, including phenolics and tannins, in plant leaves [2,14], and such increases in phenolics and tannins have an negative influence on the development and fitness of chewing herbivorous insects [15]. However, the induced phenolic compounds are decreased by elevated CO 2 when responding to damage of insect [16]. Additionally, jasmonic acid (JA) signaling defense (JA is considered as the most important defense hormone involved in resistance against chewing insects) has been reported to be suppressed by elevated CO 2 [17], and CO 2 -induced decreases in the expression of downstream genes of JA pathway (i.e., proteinase inhibitors) increased the consumption of soybean leaves by herbivorous insects [18].
Little is known about how CO 2 affects plant tolerance to herbivores but the possible effects of resource availability on tolerance have been described by three classic models or hypotheses. The compensatory continuum hypothesis (CCH) predicts that plants growing in resourcerich or low-competition environments will be more tolerant to herbivores than those growing in resource-poor, stressful environments [19]. The main rival to the CCH is the growth rate model (GRM), which predicts that plants grow at a low relative growth rate will be more tolerant than plants grow at a high relative growth rate, because, unlike plants growing in stress-free environments, plants growing in stressful environments are not growing at their maximum rate and therefore have the potential to increase their growth rate [19]. The limiting resource model (LRM) predicts that tolerance will depend on the particular resource that is limiting plant fitness and how acquisition of that resource is affected by herbivory; according to the LRM, the relative effects of a stressful vs. a stress-free environment on tolerance will therefore depend on the nature of the resource [20].
Some researchers have reported that elevated CO 2 increased plant susceptibility to herbivorous insects [21,22,23], while others found that elevated CO 2 increased compensatory growth in response to artificial herbivory, i.e., in response to researcher removal of buds from cotton plants [24,25]. Elevated CO 2 could possibly affect the re-growth ability or tolerance by increasing C:N and by decreasing the N concentration of plant tissues [26]. Although research has established that JA plays a crucial role in plant resistance to herbivorous insects [27] and that plant tolerance and resistance are not independent [13], it is still unclear how tolerance is affected by the JA signaling pathway and how the JA signaling pathway, and therefore resistance and tolerance, are affected by elevated CO 2 .
Using the cotton bollworm, Helicoverpa armigera, and tomato, the current study investigated the relationships between elevated CO 2 , insect herbivory, the JA pathway, and plant tolerance and resistance. Helicoverpa armigera is a leaf-chewing insect that causes great damage to cotton, tomato, and many other crops in Northern China [28]. To determine whether elevated CO 2 influenced resistance and tolerance of tomato plants by affecting the JA pathway, we used the JA pathway-impaired mutant spr2. We tested the hypothesis that, by altering the JA signaling pathway, elevated CO 2 would reduce tomato plant resistance (the ability to prevent damage) against H. armigera while enhancing plant tolerance (the ability to re-grow after H. armigera damage). Our specific aims were to determine (1) whether elevated CO 2 affects the JA-dependent defense of tomato plants and the midgut enzyme activities of cotton bollworm associated with different tomato genotypes, and (2) whether elevated CO 2 affects tolerance (in terms of re-growth ability, as indicated by photosynthetic rate, sucrose phosphate synthases, sucrose synthases, biomass, flower number, height, and branch length) of wild-type and spr2 plants after H. armigera damage.  of Zoology, Chinese Academy of Science (CAS) in Xiaotangshan County, Beijing, China (40u119N, 116u249E). The atmospheric CO 2 concentration treatments were: (1) current atmospheric CO 2 levels (375 ml/L) (''ambient CO 2 ''), and (2) doubled ambient CO 2 levels (750 ml/L) (''elevated CO 2 ''). Four OTCs were used for each CO 2 concentration treatment. During the period from seedling emergence to the harvesting of tomato plants, CO 2 concentrations were monitored and adjusted with an infrared CO 2 analyzer (Ventostat 8102; Telaire, Goleta, CA, USA) once every 20 min to maintain the CO 2 concentrations. The automatic-control system for adjusting the levels of CO 2 concentration, as well as specifications for the OTC, are detailed in Chen and Ge [29].

Tomato plants
Wild-type (WT) tomato plants (Lycopersicum esculentum cv. Castlemart) and jasmonate-deficient mutant plants (spr2) were kindly provided by Professor C. Li of the Institute of Genetics and Developmental Biology, the Chinese Academy of Sciences. The JA-biosynthesis mutant, suppressor of prosystemin-mediated responses2 (spr2), reduces chloroplast v3 fatty acid desaturase, which impairs the synthesis of JA [30]. WT tomato was the parent for the spr2 mutant. After they had grown in sterilized soil for 2 weeks, the tomato seedlings were individually transplanted into large plastic pots (20 cm diameter and 22 cm height) containing sterilized loamy field soil (organic carbon, 25 g/kg; N, 500 mg/kg; P, 200 mg/kg; K, 300 mg/kg) and placed in OTCs on 23 May 2010. Each OTC contained 40 plants. Pots were rearranged randomly within each OTC once every week. No chemical fertilizers and insecticides were used. Water was added to each pot once every 2 days.

Herbivore treatment
In its native habitat in China, Helicoverpa armigera is one of the most abundant of the lepidopteran herbivores on L. esculentum cv. Castlemart. H. armigera larvae tend to attack leaves and fruits. Thirty-six days after the tomato plants had been placed in the OTCs, 40 plants of each genotype at each CO 2 level ( = 10 plants per OTC) were randomly selected and infested with H. armigera. Six 4 th -stage H. armigera larvae were placed on each of two leaves at mid-plant height, and the leaves were caged (80 mesh gauze); corresponding leaves of control plants were caged in the same way. After 2 days, the larvae were removed and were weighed and analyzed for midgut proteases as described later in the Methods. The leaves of eight randomly selected plants infested with larvae and of eight noninfested plants were harvested and immediately stored in liquid nitrogen (Fig. 1).

Plant photosynthesis
Net photosynthetic rate was determined in the presence or absence of herbivores on four 5-week-old plants for each genotype at each CO 2 level. Gas exchange was measured on the distal portion of the leaf blade of undamaged leaves using a Li-6400 with a red/blue LED light source (6400-02B; Li-Cor, Lincoln, NE, USA). The CO 2 concentration of the incoming air was adjusted to 400 mmol mol 21 CO 2 or 750 mmol mol 21 . Relative humidity corresponded to ambient conditions. Before gas exchange was measured, photosynthetic active radiation (PAR) for the leaf in the measuring cuvette was increased in steps to 1200 mmol m 22 s 21 .-When the CO 2 assimilation rate was stable for at least 2 min, a light response curve was recorded. Measurements were carried out after herbivores had been allowed to feed on plants for 2 days.

JA measurement
A 0.5 g sample of fresh leaves was ground to a fine powder on ice. The powder was mixed with 4 ml of 80% methanol (80 methanol/20 water, V/V) and kept at 220uC for 12 h, and then added to 6 ml of [9,10]-dihydro-JA (dihydro-JA) for use as an internal standard. The total extracted preparation was centrifuged at 8,000 g for 20 min. The condensed endogenous JA were extracted according to procedures described by ren et al., [31].
Endogenous JA and its internal standards (dihydro-JA) were analyzed using full GC/MS scans. Retention times were identified using Xcalibur 1.2, the NIST 2003 mass library. Endogenous JA was measured by GC-MS selected ion monitoring (SIM). The characteristic ions (m/z) were 151/224 for JA and 153/226 for the internal standard (dihydro-JA).
Plant enzyme activity (SPS, SS, LOX, PIs, PPO, POD, and PAL) For analyses of sucrose synthases (SS) and sucrose phosphate synthases (SPS), 0.5 g of fresh tomato leaves was homogenized for 1.5 min at 4uC in 1:10 (fresh weight/buffer volume ratio) 100 mMphosphate buffer, pH 7.4, containing 100 Mm KCl and 1 mM EDTA. Homogenates were centrifuged at 10 000 g for 10 min, and the supernatants were subjected to SS and SPS analysis. SS and SPS were analyzed according to the protocol of the reagent kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu Province, China).
About 0.1 g of frozen leaf tissue was used for determination of plant defensive enzyme activity, including the activities of lipoxygenase (LOX), proteinase inhibitors (PIs), polyphenol oxidase (PPO), peroxidase (POD), and phenylalanine ammonia lyase (PAL). Extract was obtained from individual leaflets by grinding them in a 50 mM Tris HCl buffer (pH 7.8, 3 ml/g of leaf tissue) containing 7% polyvinylpolypyrrolidine, 1.67 mM phenylthiourea, 0.3 M KCl, and 0.4 mM ascorbic acid. This extract was immediately frozen for later use. For assays, the thawed extract was centrifuged at 13,000 g for 10 min, and enzyme activity was measured in the supernatant. LOX assays were performed in a 1ml reaction mixture containing 20 ml of supernatant and 980 ml of reaction buffer. The reaction buffer consisted of 0.113 g of linoleic acid, which had been dissolved in 3 ml of methanol in 100 ml pH 7.8 Tris-HCl buffer (w/v, pH 7.0), and 800 ml of Tween-20. PIs were measured as in Thaler et al. [32]. PPO activity was measured as in Mahanta et al. [33]. POD activity was determined according to Tegelberg et al. [34]. Reaction mixtures for determination of PAL activity contained 950 ml of 0.02 mM Lphenylalanine (0.33 g of L-phenylalanine dissolved in 100 mL of Tris-HCl, pH 7.8) and 50 ml of supernatant. The changes in absorbance were detected with a spectrophotometer at 290 nm (SPEKTRAmaxHPlus; Molecular Devices, Sunnyvale, California, USA).

Midgut proteases of H. armigera
After they had fed on plants for 48 h, 30 larvae of uniform size from each treatment were selected. Larvae were anaesthetized by chilling on ice, and their midguts were removed. The midgut tissue was cut longitudinally, and the external surface of the midgut tissue was cleaned by placing it in 0.15 mol/L NaCl. Finally, the tissues were stored at 280uC.
All enzyme substrates were obtained from Sigma-Aldrich. Weak alkaline trypsin-like enzyme (WATE) activity with Na-tosil-Larginin methyl ester (TAME) substrates was assayed spectrophotometrically at 248 nm according to the methods of Erlanger et al.  [35]. The substrates were initially dissolved in dimethyl formamide (DMF) and further dissolved in 20 mM Tris-HCl buffer, pH 8.5.
Active alkaline trypsin-like enzyme (AATE) activity was measured according to Erlanger et al. [35] by using N-benzoyl-DL-arginin-r-nitroanilide (BApNA) and 7.5% (v/v) dimethyl sulfoxide (DMSO) dissolved in Glycine-NaOH buffer (0.1 mM, pH 10.5) at a final concentration of 1.4 mM. The mixtures were incubated for 20 min at 30uC, and the reactions were stopped by adding 0.5 ml of 30%(v/v) acetic acid. The enzyme activity was measured at 406 nm.
Chymotrypsin activity in extracts was determined spectrophotometrically according to Blackwood et al. [36], based on the hydrolysis of N-benzoyl-L-tyrosin ethyl ester (BTEE), which was determined at 256 nm (l M in 10% (v/v) methanol with 0.15 M NaC1). Tris-HCl buffer (0.2 M, pH 8.5) was used. One unit of enzyme activity was defined as 1 mol of BTEE hydrolyzed per min. The extinction coefficient of BTEE was 964.
Total protease was measured with azocasein as the substrate. Azocasein (20 mg ml 21 ) was dissolved in 0.15 M NaC1. The sample was then dissolved in 0.2 M glycine-NaOH buffer (pH 8.0). The reaction was run by adding 300 ml of reaction buffer to 0.3 ml of the azocasein solution and incubating the mixture for 24 h at 30uC. The reaction was stopped by adding 0.6 ml of 20% (W/V) trichloroacetic acid (TCA). Samples from each time point were centrifuged at 10000 g for 15 min at 4uC, and the absorbance of the supernatant was measured once at 366 nm. One absorbance unit from the mixture was defined as one unit of azocasein under the given assay conditions.
Mean relative growth rate of H. armigera H. armigera larvae were weighed with an automatic electrobalance before and after they had fed on tomato plants. Mean relative growth rate (MRGR) was calculated following Chen et al. [37]: MRGR = (ln W 2 2ln W 1 )/t, where W 1 is the initial weight;

Resistance and tolerance analysis
The following variables were considered to be measures of resistance: the activities of LOX, PIs, PPO, POD, and PAL in plants; the activities of midgut enzymes in H. armigera; and the MRGR of H. armigera.
Tolerance was measured by comparing the parameters on plants that were or were not exposed to H. armigera and that were exposed to ambient or elevated CO 2 . In the short term, the following variables were considered to be measures of tolerance: photosynthetic rate, sucrose phosphate synthases and sucrose phosphate synthases. In the long term, the following variables were considered to be measures of tolerance (i.e. re-growth ability): total branch length per plant, plant height, cumulative number of flowers per plant, and final root and shoot mass. These data were obtained 60 days after the larvae had been added from five plants from each combination of genotype and herbivore treatment per OTC ( = 20 plants per OTC and 160 plants in total). JA content was considered to be one measure of both tolerance and resistance.

Statistical analysis
A split-split plot design was used to analyze the univariate responses of the measured variables (i.e., plant traits, chemical components, phytohormones) (ANOVA, PASW, 2009). In the following ANOVA model, CO 2 and block (a pair of OTCs with ambient and elevated CO 2 ) were the main effects, tomato genotype was the subplot effect, and herbivore level was the subsubplot effect: where C is the CO 2 treatment (i = 2), B is the block (j = 4), G is the tomato genotype (k = 2), and H is the herbivore treatment (l = 2).   www.r-project.org/) was used to calculate the standard error of the difference value of parameters between herbivore infested and uninfested treatment.

Photosynthesis and growth
Relative to ambient CO 2 and in the absence of H. armigera, under elevated CO 2 , the photosynthetic rate were significantly increased by 87.9% and 43.8%, biomass by 31.6% and 19.8%, flower number by 51.1% and 53.0%, total branch length by 27.5% and 31.8%, and plant height by 56.7% and 44.5% for WT and spr2 plants, respectively. The root to shoot ratio (R:S) were significantly decreased by 27.2% for WT plants and by 8.2% for spr2 plants under elevated CO 2 (Fig. 2).

SPS and SS activity
Relative to ambient CO 2 and in the absence of H. armigera, SPS activity were significantly increased by 2.2-fold for WT plants and by 4.4-fold for spr2 plants under elevated CO 2 (Fig. 3A) but SS activity of either WT or spr2 plants were not affected (Fig. 3B). SPS and SS activities were unaffected by genotype (Fig. 3).

JA level and defense enzyme activity
Genotype, H.armigera infestation, the interaction between CO 2 and genotype, the interaction between genotype and H.armigera infestation, as well as the interaction among CO 2 level, genotype and H.armigera infestation, significantly affected plant JA content (Table S1). JA levels and LOX, PI, and PPO activities of WT plants were decreased when damaged by H. Armigera under elevated CO 2 (Fig. 4A-D). Elevated CO 2 also decreased LOX and PI activity of undamaged WT plants. POD and PAL activity of undamaged plants in both genotypes were higher under elevated CO 2 (Fig. 4E, F).
H. armigera significantly increased the levels of JA and all defensive proteins in WT plants under ambient CO 2 (Table S1, Fig. 4). Under elevated CO 2 , in contrast, H. armigera only increased JA levels and PI activity in WT plants.

Midgut proteases and weight of H. armigera
Total Proteolysis was affected by CO 2 and the interaction between CO 2 and genotype (Table S2). H. armigera that consumed leaves of WT plants grown under elevated CO 2 had substantially higher gut protease activities than those that consumed leaves of WT plants grown under ambient CO 2 (Table S2, Fig. 5). However, elevated CO 2 did not affect gut protease activity in H. armigera that consumed spr2 foliage. Under ambient CO 2 , H. armigera that consumed leaves of WT plants had higher total protease and CTE activity than those that consumed spr2 foliage (Fig. 5A, B). Moreover, H. armigera that consumed leaves of WT plants had a lower MRGR than those that consumed spr2 foliage under ambient CO 2 (Fig. 6).

Regrowth and sucrose transportation after H. armigera attack
For WT plants under ambient CO 2 , H. armigera increased the photosynthetic rate and SPS and SS activities but did not significantly affect biomass, flower number, height, and total Figure 8. The major results and conclusion of this study were summarized. Elevated CO 2 decreased resistance and tolerance of WT to H. armigera. In contrast, the resistance and tolerance of spr2 were not changed by elevated CO 2 . Elevated CO 2 reduces the resistance and tolerance of WT plants by suppressing the JA signaling pathway. doi:10.1371/journal.pone.0041426.g008 branch length. For WT plants under elevated CO 2 , however, H. armigera not only reduced the photosynthetic rate but also decreased plant biomass by 13.7% and flower number by 24.9% (Fig. 7D, F). For spr2 plants under ambient CO 2 , H. armigera reduced plant biomass by 20.1%, flower number by 24.2%, and plant height by 11.6% (Fig. 7D, F, G). For spr2 plants under elevated CO 2 , H. armigera reduced plant biomass by 18.8%, flower number by 33.8%, and total branch length by 14.6% (Fig. 7D, F,  H). In addition, H. armigera increased the R:S of both genotypes under both CO 2 levels (Fig. 7E).

Discussion
The effects of elevated CO 2 on resistance and tolerance ability of tomato plants to H.armigera attack Many studies have evaluated the relationship between plant resistance and tolerance to herbivores [38,39], but little information is available regarding how the relationship between tolerance and resistance is affected by an abiotic stress such as global CO 2 enrichment. Our results suggested that elevated CO 2 decreased tomato plant resistance against H. armigera by suppressing the critical defensive signal molecule JA and JA-pathway-related defensive enzymes. Our results also indicated that tomato plants grown under elevated CO 2 are less tolerant to H. armigera than plants grown under ambient CO 2 . Phenotypic plasticity is a principal means by which plants cope with biotic or abiotic stress [40], and the decreased resistance and tolerance to herbivores under elevated CO 2 in this study suggests that elevated CO 2 reduces the phenotypic plasticity of plant response to herbivorous insect attack.
Prior studies have revealed that elevated CO 2 increases total nonstructural carbohydrates in plant tissues and that the excess C is probably allocated to the increased synthesis of secondary metabolites, such as terpenes and phenolics [41,42], which in turn can reduce the development of chewing insects [43]. Furthermore, PAL is known to be a principle enzyme involved in a rate-limiting step of phenolic biosynthetic process [41]. Our results with both genotypes of tomato plants were consistent with the previous finding that elevated CO 2 increases PAL activity in plants (Fig. 4F). Generally, attack by chewing insects induces a complex set of defense responses in plants [44]. In WT plants under ambient CO 2 , H. armigera attack increased anti-oxidant enzymes (in terms of POD), followed by triggered JA signaling pathway defense (in terms of JA and LOX) and caused plants to increase PIs and PPO activities ( Fig. 4A-E). To the undamaged WT plants, elevated CO 2 decreased LOX and PIs activity. Additionally, when damaged by H.armigera, elevated CO 2 reduced LOX activity and JA level as well as PI and PPO activity in WT plants (Fig. 4A-D). It seems that elevated CO 2 tends to impair the JA-dependent defense induced by H. armigera. Although responses to elevated CO 2 by plants and insects are species-specific [14], the current results are not the first to show that elevated CO 2 changes the plant-insect interaction by modifying the JA-dependent pathway. In soybean plants, elevated CO 2 also suppressed the JA signaling pathway and increased susceptibility to the Japanese beetle, Popillia japonica [17,18].
Proteinase inhibitors (PIs) of plants are able to reduce the feeding fitness of chewing insects by suppressing insect gut proteases [45]. WT plants grown under elevated CO 2 had reduced PI activity, the reduced resistance resulted in increased gut protease activities for H. armigera (Fig. 5); these results may explain our previous finding that H. armigera consumed more wheat biomass when the wheat was grown under elevated vs. ambient CO 2 conditions [46]. Although increases in gut protease activities may result in increased consumption under elevated CO 2 , the H. armigera MRGR did not change (Fig. 6). Perhaps the increased consumption only complemented feeding and enabled the insect to maintain development and growth when consuming leaves grown under elevated CO 2 , i.e., on leaves with a reduced N concentration.
Plant tolerance to herbivorous insects can depend on the availability of particular resources such as C resources [9]. Elevated CO 2 increases C assimilation and causes re-allocation of C (especially sucrose) in plant tissue [47]. In the transport of sucrose from leaves to sink tissues via phloem, SPS and SS are key regulatory enzymes [48]. Because elevated CO 2 significantly increases plant growth and C metabolism (Fig. 2), the CCH hypothesis would predict that plant tolerance to herbivores would be increased in the resource-rich, elevated-CO 2 environment. The CCH hypothesis, however, was not supported by the current study. Under ambient CO 2 , WT plants expressed substantial tolerance to H. armigera attack in that herbivory increased C assimilation and sucrose synthesis (as indicated by SPS activity in leaves) and transportation (as indicated by SS activity in roots) such that plant biomass and other parameters measured at week 8 did not differ between plants with and without H. armigera infestation. This suggests that WT plants can completely compensate for H. armigera damage under ambient CO 2 . Under elevated CO 2 , in contrast, H. armigera consumption of WT plants reduced photosynthesis, biomass, flower number, and plant height and did not increase SPS and SS activity. We conclude that elevated CO 2 reduces the tolerance of WT tomato plants to H. armigera. Furthermore, our results are consistent with the GRM hypothesis, which predicts that plants growing under conditions that promote a high growth rate will be less tolerant to herbivores than plant growing under conditions that reduce the growth rate [23].

The interaction between elevated CO 2 and jasmonate signalling
In addition to the essential role of the JA pathway in resistance against herbivorous insects, JA has been found to regulate the interaction between tolerance and resistance in Nicotiana attenuata against Manduca sexta [49]. While WT plants in the current study exhibited substantial tolerance against H. armigera under ambient CO 2 , spr2 mutant plants did not, i.e., tolerance under ambient CO 2 was much lower in spr2 than in the WT (Fig. 7). This is consistent with that finding that treatment of Populus with JA increased C transport to the roots, nutrient uptake, and regrowth capacity, and therefore increased tolerance [50]. Results concerning the effect of the JA pathway on tolerance, however, have been inconsistent. N. attenuata and its JA-deficient genotype (asLOX) did not differ in capsule production after simulated herbivore attack [9], and the tolerance to defoliation did not differ between WT Arabidopsis and the overexpressing-JA genotype JMT [51]. This indicates that tolerance to herbivorous insects may depend on some mechanisms other than the JA pathway. Under elevated CO 2 , the tolerance of spr2 plants to H. armigera was not lower than WT plants except the total branch length (Fig. 7A-G). This result confirmed that JA pathway plays an important role in tolerance of plants when attacked by H.armigera under elevated CO 2 . Our study suggests that the suppression of JA pathway may be one of reasons why elevated CO 2 decreased both resistance and tolerance of tomato when damaged by H. armigera.

Conclusion
This study has generated a number of significant findings. First, the results support the view that the JA signaling pathway is important in both resistance and tolerance to chewing insects. Second, the results are consistent with the GRM model, which predicts that tolerance will be greater in resource-limited than in resource-unlimited environments. Third, a trade-off between resistance and tolerance as predicted by classical theory was not evident in our study [52]. Finally and perhaps most importantly, the results suggest that plants may suffer greater damage from herbivorous insects if levels of atmospheric CO 2 continue to increase (Fig. 8).

Supporting Information
Table S1 F and P values from MANOVAs for the effect of CO 2 level, tomato genotype and H. armigera infestation on growth traits, photosynthesis, and foliar chemical components of two tomato genotypes. (DOC)