A field experiment with elevated atmospheric CO2-mediated changes to C4 crop-herbivore interactions

The effects of elevated CO2 (E-CO2) on maize and Asian corn borer (ACB), Ostrinia furnacalis, in open-top chambers were studied. The plants were infested with ACB and exposed to ambient and elevated (550 and 750 μl/l) CO2. E-CO2 increased the plant height and kernel number per ear. The plants had lower nitrogen contents and higher TNC: N ratios under E-CO2 than at ambient CO2. The response of plant height to E-CO2 was significantly dampened in plants with ACB infestation. However, the weight gain of the survivors declined in plants grown under E-CO2. Moreover, the plant damage caused by ACB was not different among the treatments. Overwintering larvae developed under E-CO2 had a lower supercooling point than those developed under ambient CO2. The results indicated that there was a positive effect of E-CO2 on the accumulation of maize biomass, i.e., the “air-fertilizer” effect, which led to a nutritional deficiency in the plants. The fitness-related parameters of ACB were adversely affected by the CO2-mediated decreased in plant nutritional quality, and ACB might alter its food consumption to compensate for these changes. Larval damage to maize under E-CO2 appears to be offset by this “air-fertilizer” effect, with reductions in larval fitness.

. Depending on the species, the herbivore-induced response of plants may vary in different growth environments 13 .
Because of the elevated CO 2 -mediated dilution of the N content in plants results in a nutritional deficiency for protein-limited insect pests 14 , many insects compensate for the changes in plant quality by altering their food intake, which results in more severe damage or defoliation to the host plants 15,16 .
The C 3 and C 4 pathways of photosynthesis are distinct, and C 4 plants are less sensitive to E-CO 2 than C 3 plants 17,18 . Therefore, an elevation in CO 2 theoretically would not influence the rate of photosynthesis in C 4 plants 19 , however, research has found that the rate of photosynthesis and the above ground biomass increased in C 4 plants [20][21][22] . There are several explanations for this phenomenon 23 . A report showed that C 3 grasses were more nutritious and had higher levels of proteins, nonstructural carbohydrates, and water and lower levels of fibre and toughness than C 4 grasses under E-CO 2 24 . Maize is a C 4 plant and is the most important food and feed crop in China. The Asian corn borer (ACB), Ostrinia furnacalis (Guenée), is a key pest of maize and causes yield losses of 30% in various agro-climatic regions 25 . The ACB overwinters as fully developed larvae that are found in the maize stalks, cobs, and weed stems, or in a spun-silk covering in the plant debris 26 .
Numerous studies have been performed to predict the effects of rising CO 2 concentrations on C 3 crop-pest interactions in agriculture, but few studies have examined the effects on the C 4 crops and their insect pests 27 . Similar to the C 3 plants, E-CO 2 also changes the interactions between maize and its pests. Yin et al. (2010) 16 reported that the growth, development and consumption of Helicoverpa armigera (Hbn.) changed when it was fed maize grains grown under E-CO 2 . The exposure of maize plants and Chilo partellus (Swinhoe) to E-CO 2 levels not only affected the growth and yield of maize, but also affected the development of the insect in the open-top chambers 28 .
The cold hardiness of an insect species is measured by its supercooling point (SCP), which influences the density of the overwintering population 29 . The SCP is influenced by various factors, including the host plant species and nutritional quality and the contents of water and of the cryoprotective substances in the larval body [30][31][32][33] . An understanding of the effects of E-CO 2 on the SCP could provide direct evidence for the cold hardiness of the overwintering ACB populations in environments with future climate change.
The objectives of this study were to examine the effects of E-CO 2 on the development and abundance of the 1 st -and 2 nd -generation ACBs and on the damage caused by the ACBs to the maize plants grown in open-top chambers. Additionally, the effect of E-CO 2 on the cold hardiness of overwintering larvae was evaluated with tests to determine the supercooling points of diapause larvae. The information in this study on the performance and abundance of insects on plants (insect-plant interactions) under elevated levels of CO 2 is as important as understanding the changes in herbivorous damage to agricultural commodities caused by global climate change.

Results
Maize plant chemistry. A negative effect of E-CO 2 on the N content in maize plants was observed in the experiment (Table 1). Compared with the maize plants grown under ambient CO 2 , the N content significantly decreased by 8.0% and 17.0% for leaves and stalks, respectively, in maize plant grown under 750 μ l/l CO 2 ( Table 1). The maize plants grown under 750 μ l/l CO 2 also had a 4.5% decrease in the N content of leaves compared with the maize plants grown under 550 μ l/l CO 2 . There were no significant differences between the ambient and the 550 μ l/l CO 2 levels for the N content in leaves or stalks. A positive effect of E-CO 2 on the C content of the maize plants was observed in the experiment ( Table 1). The response of total non-structural carbohydrates (TNCs) including soluble sugars and starch, was consistent, with a significant effect of E-CO 2 found only in the 750 μ l/l CO 2 treatment, for which the increasing was approximately 15.7% for soluble sugars and 8.3% for starch in leaves and 14.2% for soluble sugars and 11.9% for starch in stalks compared with maize plants grown under an ambient CO 2 condition. The TNC: N ratio was significantly different among the treatments (Table 1). Compared with the ambient CO 2 condition, the TNC: N ratio increased by 8.5 and 18.5% in leaves and 16.6% and 35.3% in stalks under E-CO 2 levels (550 and 750 μ l/l, respectively). Although the water content declined in the plants grown under elevated levels of CO 2 , the treatments were not significantly different ( Table 1).

Fitness of ACB larvae.
For the 1 st generation of ACB, the larval survival among the CO 2 treatments was not affected ( Table 2). However for the 2 nd -generation, the survival of the larvae decreased by 16.5% in 2012 and 21.0% in 2013 in the plant grown under elevated CO 2 compared with the maize plants grown under ambient CO 2 . The average weight gain per larva (2 nd generatrion, diapause) declined significantly by 13.0% and 16.1%, respectively, when the larvae were fed maize plants grown under the two elevated levels of CO 2 (550 and 750 μ l/l), compared with the ambient CO 2 (Table 3). Additionally, the cold hardiness of the diapause larvae was significantly affected by the CO 2 concentrations. The average SCP of the diapause larvae that developed on the maize plants grown under E-CO 2 (750 μ l/l) was slightly lower (approximately 0.31 °C lower) than that for the larvae that developed on the maize plants grown under ambient CO 2 .
The fecundity of the moths that developed from the overwintering larvae was marginally affected by the CO 2 levels (Table 3). Although the number of eggs laid by the females that developed from the overwintering larvae declined as the atmospheric CO 2 concentration increased, the difference was not significant among the treatments.
Maize plant growth and ACB damage ratings. The results revealed that the E-CO 2 had a positive effect on the growth of plants, whereas the ACB infestation affected plant growth negatively (Fig. 1). A significant "air fertilizer" effect of E-CO 2 was observed on maize plant growth (2012: F 2,6 = 4.92, P < 0.05,   Table 3. Body weight and supercooling points (SCP) (mean ± SE) for overwintering larvae and female fecundity of Ostrinia furnacalis from different CO 2 levels in 2012. Means within a column followed by different letters are significantly different (CONTRASTS test, P < 0.05).
6.54% with an ACB infestation at the whorl stage compared with the control. Overall, the ACB damage at the whorl stage completely dampened the positive height response of the maize plants to the E-CO 2 ( Fig. 1), which suggested that the larvae altered their food consumption to compensate for the changes in the quality of the maize plants. Additionally, the grain yields of the maize plants were significantly affected by the concentration of CO 2 ( Table 4). The average number of kernels per ear increased by 4.3% and 4.5% in 2012 and 2013, respectively, under E-CO 2 (750 μ l/l) compared with the ambient CO 2 .
There was no significant difference in the weight per 100-kernels among the treatments. These results demonstrated the positive effect of E-CO 2 on maize grain production. The damage ratings of the ACB were unaffected by the concentration of CO 2 during the two years (Table 5). There were no significant differences in the number of tunnels per plant or in the length of cavities per plant among the treatments.

Discussion
The effects of E-CO 2 on maize plants have been assessed in a number of places with variable conditions. A few studies have suggested that maize plant are insensitive or less sensitive to elevated levels of CO 2 in the absence of drought and heat 34,35 . Moreover, most studies have revealed that E-CO 2 has a    positive effect on the maize plant, i.e., maize plant are likely to have greater rates of photosynthesis and above ground biomass accumulation in addition to reduced transpirational water losses and increased water-use efficiencies [36][37][38][39] . The number of seed is also greater with E-CO 2 than that of plants grown under ambient CO 2 level (+ 5.0%) 40 . The stimulation of the photosynthesis and growth of maize plants under E-CO 2 typically results in a reduction in leaf N content, or an increased in the TNC: protein ratio of maize grains 16 . The results of the present study were in accordance with previous studies that showed positive effects of E-CO 2 on plant biomass, i.e. the heights of plant and the kernels per ear increased in the E-CO 2 environment. Additionally, the chemical changes in the maize plants followed the general pattern of plants in responses to E-CO 2 , such as a decrease in the nitrogen content and an increase in the total nonstructural carbohydrates 41 . The TNC: N ratio significantly increased in response to E-CO 2 in the present study, which suggested that the nutritional quality of the maize plant was reduced when the maize was grown under an E-CO 2 condition. Similar research found that C 4 grasses were poor host plants primarily because of their lower level of nutrient, higher level of fibre and greater toughness 42 .
Despite the direct response of many insects to E-CO 2 , the changes in the performance of herbivorous insects are intimately correlated with changes in the quality of food plants grown under E-CO 2 conditions. The nitrogen (protein) content of the host plant is only one limiting nutrient for insect herbivores 43 , and a number of chemical compositions affect the nutritional quality of host plants 44 . A decrease in the foliar nitrogen content of host plants can affect the rates of development and survival of insect herbivores 45 . Moreover, insects that can compensate for the CO 2 -mediated dilution of foliar nitrogen by increasing the rate of feeding will experience retarded growth and will be subject to predation for a longer period of time (the slow-growth-high-mortality hypothesis) 46 . In present study, the survival decreased and the average weight gain per larva declined significantly when the larvae fed on the maize plants grown under E-CO 2 , possibly because of the CO 2 -mediated declines in the nitrogen content of the maize plants. These results indicated that the ACB is a protein-limited insect; the development and survival of the larvae were adversely affected by the CO 2 -mediated reduced suitability and nutritional quality of the host maize plants. Prominent among the many factors that affect the amount of plant tissues consumed by insect herbivores is that of the suitability and nutritional quality of their host plants. Studies have found that some leaf-chewing herbivores perform compensatory feeding and by increase the intake of foliage with a lower nitrogen content to meet their nutritional requirements under an E-CO 2 environment 27,45 . As a consequence, levels of damage or defoliation increase. By contrast, the plants may be damaged less and have more undamaged foliar area when the E-CO 2 causes an increase in plant biomass, and reduces the plant fitness-mediated population density of insect herbivores 47 . In the present study, the survival of the ACB larvae declined in the maize plants exposed to E-CO 2 compared with the ambient treatment level. The E-CO 2 also reduced the suppression of maize plant height caused by the ACB infestation. Therefore, the damage caused by the ACB to maize will be offset by the "air-fertilizer" effect for the plant and the reduced fitness of the insect herbivores on the host plants.
The ACB overwinters as fully developed larvae in maize stalks in northern China. The overwintering larval population (2 nd -generation) has an important role in the overwinter survival of the ACB, and in the regulation of the population for the subsequent year 48 . The SCP of diapause larvae is largely related to the larval cold hardiness. In the present study, the diapause larvae that developed on the maize plants grown under 750 μ l/l CO 2 weighed less and had a slightly lower SCP than the larvae that developed on the maize plants grown under ambient CO 2 . These results suggested that there was a positive effect of   E-CO 2 on the cold hardiness of diapausing larvae, and many reports have shown that the host plants play a pivotal role in the coldhardiness of insect herbivore. For example, the average SCP was significantly lower for the 3 rd instar larvae of beet armyworm, Spodoptera exigua (Hbn.), that developed from cabbage than those that developed from pakehoi, shallot and spinach 49 . Similar evidence was found for the hemlock looper, Lambdina fiscellaria (Guenée) 50 . The host plant quality affected the overwintering success of the leaf beetle, Chrysomela lapponica (L.) 32 , the hypothesis to explain this result was that the high water content in the high-weight beetles of C. lapponica might be the primary cause of the increased winter mortality 51 . In the present study, the larval body weight was lower and the increase in the SCP occurred under E-CO 2 conditions; however, these results might also be associated with the nutritional quality and the lower water content of the maize plants that were grown under E-CO 2 than those that were grown in ambient conditions. In this study, the survival rate of the overwintering larvae (data not shown) was not influenced by the E-CO 2 although there was a small decrease in the SCPs of the larvae. Additionally, the number of eggs laid per female that developed from the overwintering larvae declined at E-CO 2 treatments although the difference was not statistically significant compared with the ambient condition. Taken together, there was insufficient evidence to conclude that the ACB exhibited a direct response to the elevated levels of CO 2 . . Three levels of CO 2 were applied continuously, i.e., ambient CO 2 (∼ 390 μ l/l) and E-CO 2 (550 μ l/l and 750 μ l/l), which represented the current and predicted levels of CO 2 in future years, respectively 52,53 , each treatment was replicated four times, for a total of twelve chambers in the experiment. The air was continuously distributed from the blowers into the chambers through a water curtain cooling system into perforated polyethylene ducts inside the chamber base-wall at 10 cm above the level of the soil. The CO 2 was added to the inlet airstreams in the chambers of the elevated treatments to reach the target CO 2 concentrations (550 μ l/l and 750 μ l/l). The concentrations were monitored and adjusted with a CO 2 sensor (JQAW-8VACD, ColliHigh Company, Beijing, China) once every 60 s to ensure relatively stable levels of CO 2 . The actual mean CO 2 concentrations in the chambers were 542 ± 14 μ l/l and 746 ± 15 μ l/l for the two E-CO 2 levels, whereas in the ambient chambers, the concentration was ∼ 390 μ l/l. The concentrations in the ambient chambers were monitored but were not controlled. The automatic-control system for adjusting the CO 2 concentration was similar to that described by Chen et al. (2005) 54  Insect stocks and plant infestation. The ACB neonates used in this study were obtained from a laboratory colony that originated from a field population collected every year, which was maintained on a regular artificial diet for ACBs 26 for 3-4 generations in the Institute of Plant Protection, Chinese Academy of Agricultural Science.

Methods
The plants were infested when they developed to the whorl and silking stages, which represent the 1 st -and the 2 nd -generation infestations in nature, respectively. Before the first infestation, each chamber was separated into two plots with screen to prevent the larvae from transferring between the two plots. The two plots in a chamber were used as treatments with either the 1 st -generation infestation or the 2 nd -generation infestation. Each plant was infested with 50 neonates (< 24 h) of the ACB with traditional artificial infestation techniques similar to those described by He et al. (2000) 55 on 29 and 28 June (the 1 st -generation) and on 8 and 7August (the 2 nd -generation) in 2012 and 2013. To avoid exposing the neonates to high temperature and direct sunlight, the infestations were applied during the late afternoon to evening. Nine chambers were used for the ACB infestation, which included three for each level of CO 2 , and another three chambers were used for the assays on chemical composition of the maize leaves. Supercooling points and overwinter status. The larvae were collected from the dissected maize plants during the autumn harvest, introduced into plastic centrifuge tubes with an air hole punched through the bottom (1.5 ml) and placed into cartons maintained, in the open-top chambers during the winter. The supercooling points of the larvae after overwintering were determined on 18 January 2012. A tube was constructed by removing the bottom part from a micro-centrifuge tube (0.5 ml), which was used to position larvae that were connected to a multichannel temperature recorder (TMC-40A, designed by the Institute of Agro-meteorology, Chinese Academy of Agricultural Sciences, Beijing). These tubes were then placed in the temperature test chamber (Heraeus-Votsch VM 04/100) at 0 °C to equilibrate for 24 h before cooling at a rate of 1 °C/min until a temperature of − 40 °C was reached. The lowest temperature reached before the release of the latent heat of fusion was recorded as the supercooling point (SCP).
Twenty-four larvae were tested for each treatment, and each treatment was replicated three times. The larvae that were not unsupercooled with the same origin were then placed individually into modified 5-ml centrifuge tube (two holes 1 mm in diameter were made in each lid; sterilized) with a piece of wet cotton as a moisture source and a piece of corrugated paper as a cryptic habitat. Finally, these larvae were reared to pupation at 26 °C, 70% RH, and with a 16:8 h (L:D) photoperiod. The newly emerged moths from each treatment were transferred in pairs to an oviposition cage (11 cm × 8 cm × 8 cm), which was covered with a piece of waxed paper as an oviposition substrate 26 . The number of eggs laid per female was recorded daily.
Chemical compositions of maize leaves. One entire unfolded ear leaf and stalk was collected from each maize plant in the OTCs for tissue samples for the chemical composition assays on the 75 th day after sowing. Ten maize plants were selected at random from each of the three CO 2 treatments, on three separate occasions, for a total of 30 leaves and 30 stalks per treatment. The water content, as a proportion of fresh weight, was calculated after the maize leaves and stalks were dried at 80 °C for 72 h. The total non-structural carbohydrates (primarily soluble sugars and starch) were analysed using the method of Tissue and Wright (1995) 56 . The nitrogen content was assayed using a Kjeltec N analyser (Model KDY-9830; Foss automated Kjeltec instruments, Beijing, China).

Data analyses.
One-way analyses of variance (ANOVAs) were used to analyse the effects of elevated CO 2 on the chemical compositions of maize leaves and stalks, larval survival per maize plant, larval body weights, supercooling points, kernel numbers, 100-kernel weights, maize plant heights, the ACB damage ratings, and the number of eggs . All data were analysed with a general linear model procedure (PROC GLM) (SAS Institute 2001). Differences in the treatments were compared using CONTRASTS test. The significance level was set at P < 0.05. Before the analyses, the data were subjected to standard transformations to improve their normality and the homogeneity of variance. The percentage data were arcsine transformed to meet the assumptions of homogeneity of variance.