Changes in melon plant phytochemistry impair Aphis gossypii growth and weight under elevated CO2

Elevated CO2 (eCO2) modifies plant primary and secondary metabolism that subsequently impacts herbivore insect performance due to changes in its nutritional requirements. This laboratory study evaluated interactions between Aphis gossypii Glover (Hemiptera: Aphididae) and melon (Cucumis melo L., Cucurbitaceae), previously acclimated two or six weeks to different CO2 levels, eCO2 (700 ppm) or ambient CO2 (400 ppm). Under eCO2, melon plants decreased nitrogen foliar concentration and increased carbon to nitrogen ratio, independently of acclimation period, significantly reducing the content of some amino acids (alanine, asparagine, glycine, isoleucine, lysine, serine, threonine, and valine) and increasing the carbohydrate (sucrose) content in melon leaves. The dilution in some essential amino acids for aphid nutrition could have aggravated the reduction in A. gossypii population growth reared on melon previously acclimated two weeks to eCO2, as well as the loss of aphid body mass from two successive generations of A. gossypii reared under eCO2 on plants previously acclimated two or six weeks to eCO2. The response to eCO2 of phloem feeders, such as aphids, is actually variable, but this study highlights a negative response of A. gossypii to this climate change driver. Potential implications on control of this pest in a global change scenario are discussed.

Non-structural carbohydrates can act as phagostimulants 7 , i.e. compounds that tasted by aphids can stimulate and sustain its feeding 26 . Sucrose is a major transport sugar, the most abundant carbohydrate in phloem sap and the most effective phagostimulant for herbivorous insects. Numerous species from Cucurbitaceae family also transport raffinose, stachyose and higher order oligosaccharides. Polyols (sugar alcohols) are also abundant in phloem 18 . However, sugars are not limiting nutrient source for aphid feeding 15,18 . The effect of eCO 2 on plant carbohydrate content is species-specific thus, most of the plants show an increase in carbohydrates content 14,20,27 meanwhile in others, soluble sugars are not affected by eCO 2 24 . In general, the increase in plant biomass and the accumulation of C-based compounds due to eCO 2 could dilute the concentration of foliar proteins; finally counteracting the positive effect that the boost in phagostimulatory activity due to carbohydrates increment produces in herbivorous insects 7 .
Aphids are very sensitive to changes in quality and quantity of their nutritional requirements 24 . Therefore, aphid responses to eCO 2 -mediated effects on host plant quality and quantity nutrient compounds are particularly variable, and could be either positive 22,28,29 , negative 12,13,24,30,31 or not significantly modified 17,25 , comparing to aphid performance under current CO 2 concentration.
Most of the studies investigating the effects of eCO 2 on agricultural crops have been focused on grains, predominantly cereals but also some legumes 32 . However, few studies have analyzed the effect of climate change on plant-herbivore interactions in horticultural crops 13,33 . This is the first time a research focused on the impact of increasing atmospheric CO 2 on melon plants (Cucumis melo L., Cucurbitaceae) under a climate change scenario. The cotton aphid, Aphis gossypii Glover (Hemiptera: Aphididae) is one of the principal pest species colonizing almost one hundred of plant species, actually one of the most important aphid pests on cucurbits. Aphis gossypii is originated from warmer regions, but can also survive northern winters in greenhouses 34 . Milder winters under climatic change could increase winter survival of insect pests and rates of herbivory 3 , therefore intensifying the damage of cotton aphid. Elevated CO 2 has been found to affect A. gossypii feeding on cotton (Gossypium hirsutum L.), ingesting more phloem sap due to a higher plant C:N ratio and lower levels of amino acids, although no change in the mean relative growth rate was found when compared eCO 2 to ambient CO 2 (aCO 2 ) 25 .
The objectives of our study were to analyze: (1) if eCO 2 changes melon plant biomass and biochemistry, specifically amino acids and soluble carbohydrates content; (2) in consequence, if eCO 2 mediated changes on plants could affect aphid performance, and; (3) whether a longer acclimation period to eCO 2 could impact more severely both plants and aphids. For that purpose, we analyzed the effect of eCO 2 on A. gossypii body mass and colony growth rate, reared on melon plants previous acclimated during two or six weeks to different CO 2 regimes, eCO 2 (700 ppm) or aCO 2 (400 ppm).

Materials and methods
Melon plants and aphids. Biological material production and experiment setup were conducted in the Institute of Agricultural Sciences of the Spanish National Research Council (ICA-CSIC, Madrid, Spain). Melon cv. Sancho (Syngenta Seeds B.V., Enkhuizen, The Netherlands) plants were used in the experiments. After germination in darkness above wet filter paper in a Petri dish, seedlings were transplanted at one week old with a mixture of equal parts of soil substrate (GoV4, Jiffy International, A.S. Norway) and vermiculite (No. 3, Asfaltex S.A., Barcelona, Spain) to 11 × 11 × 12 cm pots. Plants were placed since seedling in the plant growth chamber at 24:20 °C temperature, 60:100% RH and 16:8 h (L:D) photoperiod until CO 2 acclimation. Plants were watered on alternate days (680 mL/plant-week). A NPK 20-20-20 fertilizer (Miller Chemical & Fertilizer Corp., Pennsylvania, USA) was added to the irrigation water (1 g/L).
The clonal A. gossypii colony at the laboratory was initiated from a single virginiparous apterae collected from melon in El Ejido, Spain, in 1998. Aphid colonies were reared on melon plants for several generations inside rearing cages in environmental growth chamber at optimal development conditions of 23:18 °C temperature, 60-80% RH and 14:10 h (L:D) photoperiod. Aphids were synchronized prior the bioassays to guarantee age homogeneity (10-11 days old) at the time of the experiment. Plant acclimation to CO 2 . Two walk-in climate chambers were used for plant acclimation to CO 2 with identical conditions of 24:20 °C temperature, 60-70% RH, 14:10 h (L:D) photoperiod, and 310 ± 3 µmol m −2 s −1 light intensity at canopy level (GreenPower LED production dr/b/fr 150, Philips, Eindhoven, The Netherlands); but with different CO 2 atmospheric concentrations, one chamber with eCO 2 -700 ppm (703.28 ± 1.81 ppm) and the other with aCO 2 -400 ppm (409.89 ± 1.40 ppm). Temperature and humidity data were recorded every hour with a data logger (Tinytag Ultra 2, Gemini Data Loggers, UK) in each chamber. CO 2 concentration was monitored in aCO 2 chamber with a datalogger device (Rotronic AG CP11, Bassersdorf, Swirtzeland), while eCO 2 chamber incorporated a system that automatically regulated and recorded the chamber gas concentration.
One-week-old melon plants were divided into four sets and two of them were placed in eCO 2 or aCO 2 chamber respectively for six weeks of acclimation period, whereas the remaining two sets were maintained in the general plant growth chamber (see conditions above). Two weeks before the beginning of plant measurements, these sets were transferred to eCO 2 and aCO 2 chambers respectively for two weeks of acclimation period 35 . All plants were 7-weeks-old when experiments started and insect experimental units were maintained in their respective CO 2 treatment chambers during the bioassays. or six weeks to aCO 2 or eCO 2 respectively was concluded, five plants per treatment were randomly collected for destructive sampling to assess the effects of plant exposure to CO 2 on total C and N plant concentration. Melon stems and leaves were analyzed separately as plant chemical composition can differ within plants, and these specific niches could subsequently affect aphid performance in a different manner 36 . Stems and leaves separately were cut (pieces of 1-2 cm) and dried in a drying-oven (Selecta, Barcelona, Spain) for 48 h at 60 °C. They were then milled into powder with an analytical grinder (YellowLine A10, IKA-WERKE, Germany). Total C and N concentration was determined using an Organic Elemental Analyzer-NC Soil Analyzer (Flash 2000, Thermo scientific, Waltham, USA) 37 at the Analysis of Soils, Plants and Waters Service in ICA-CSIC. C:N ratio was calculated by dividing the concentration of C by the concentration of N for each sample. This experiment was repeated twice, obtaining finally ten replicates of leaf and stem samples respectively.

Plant biomass.
To assess the effect on plant weight to the exposure of CO 2 , ten plants per treatment were randomly collected for destructive sampling when the previous acclimation period to two or 6 weeks to aCO 2 or eCO 2 was concluded. Plants were separated in stems and leaves, then, samples were maintained at − 20 °C and the day before the freeze-drying, they were deep-frozen at − 80 °C. Once freeze-dried (Epsilon 2-4 LSCplus freeze dryer, Christ, Osterode am Harz, Germany), samplings were weighed on an analytical balance (model AB204, Mettler Toledo, Greifensee, Switzerland) to calculate their dry weight and kept in a desiccator to analyze amino acids and carbohydrates pigments at a later time.
Amino acids and carbohydrates content. Free amino acids and carbohydrates were obtained adapting the extraction method and the Gas Chromatography Mass Spectrometry analysis of plant samples (n = 6), leaves and stems separately, from the protocol described on supplementary information in Corrales et al. 38 , at the Metabolomic Service in Centro de Biotecnología y Genómica de Plantas (CBGP, UPM-INIA, Madrid, Spain). Amino acids and carbohydrates were measured from tissue extraction, instead of phloem sap collection that could be a priori better related with sap-feeding insects, after investigating that tissue extraction has been shown to be a reliable indicator on the relative composition of amino acids and some carbohydrates (e.g. sucrose) in other crops, such as lucerne (Medicago sativa L.) 21,23 , barley (Hordeum vulgare L.) 39 , and spinach (Spinacia oleracea L.) 40 .
Aphid growth and performance. Effects of CO 2 on Aphis gossypii adult weight. In order to calculate aphid body mass, we weighed plots of 50 synchronized first (F1) and second (F2) generation adults (7-days-old), exposed to the different CO 2 concentration on melon plants previously acclimated to aCO 2 or eCO 2 for 2 or 6 weeks. To get F1 A. gossypii adults, the day that the previous plant acclimation period to two or six weeks to aCO 2 or eCO 2 concluded, 100 adults of A. gossypii from the synchronized rearing, were placed distributed in 18 clip-cages in a plant of each treatment. Twenty-four hours later, adults were removed and onset nymphs were left to develop themselves exposed to the CO 2 conditions determined for each treatment. Seven days later, when nymphs had already reached adulthood (F1 adults), we proceeded to weigh them. To get F2 adults, 100 adult aphids from the first generation of each treatment were placed in another acclimated plant of the same treatment. Twenty-four hours later, adults were removed and onset nymphs (start of the A. gossypii second generation) were left to develop themselves exposed to the different conditions of CO 2 . Seven days later, when nymphs had already reached adulthood, we proceeded to weigh F2 adults. For each aphid generation, groups of 50 adults (n = 12) were made to calculate aphid average weight. Adults were anesthetized with CO 2 and then weighed (fresh weight) on an analytical balance (Mettler AE166 DeltaRange, Greifensee, Switzerland). Then, samples were put in an oven-drier at 60 °C for 24 h and weighed again (dry weight).
Effects of CO 2 on Aphis gossypii colony performance. The day of the beginning of bioassays, when melon plants had been previously acclimated two weeks to aCO 2 or eCO 2 , two synchronized adults (12-days-old) were placed in each plant (ten plants per treatment), in order to acclimate aphids to the respective CO 2 concentration. Twenty-four hours later, aphids were removed except six nymphs per plant. Each plant was covered with a fine mesh, to facilitate aphid dispersal in the plant but avoiding contamination between plants. After 7 days, when nymphs had already reached adulthood, only two were left per plant. The offspring of these previously acclimated adult females were counted at 14 and 21 days. For the last count, samples were frozen to facilitate a later counting due to the massive number of aphids. The aphid colony growth rate on each plant was calculated as the difference between number of aphids on a given day and the number of aphids on the previous count day 30 . In our case, analysis was performed with the increase in population number from day 7 to 14, and from day 14 to 21, so this value was calculated between weeks.
Statistical analysis. To determine the effects of CO 2 concentration, acclimation period to CO 2 and their interaction, all plant data (C and N concentration, biomass, amino acids and carbohydrates content) and aphid adult weight data were subjected to the two-way analysis of variance (ANOVA) using the General Linear Model module in IBM SPSS Statistics 22.0.0.0 software (package for Windows, 64-bit edition, Chicago, USA). Whenever interaction between factors was statistically significant (P < 0.05), a post hoc LSD test was performed for pairwise comparisons. To achieve normality and homoscedasticity of some parameters, data was transformed by sqrt(x + 0.5) or log(x + 1). Aphis gossypii colony performance were analyzed by Student t-test (P ≤ 0.05) with the same statistical software.

Results
Total carbon and nitrogen concentration. Nitrogen concentration in melon leaves was significantly affected by CO 2 concentration, being significantly lower under eCO 2 than under aCO 2 (F 1,32 = 13.065; P = 0.001), whereas C concentration in melon leaves was not affected by the CO 2 concentration level (F 1,32 = 0.003, P = 0.959). Statistically significant differences were also found in C:N ratio in melon leaves due to CO 2 concentration. A significant increase in foliar C:N ratio due to the dilution in N concentration occurred under eCO 2 compared to aCO 2 (F 1,32 = 7.873; P = 0.008) (Fig. 1, Supplementary Table S1). In contrast, acclimation period did not affect C nor N concentration in melon leaves. Neither the C and N concentration nor C:N ratio in melon stems were significantly affected by CO 2 concentration or acclimation period (Fig. 1, Supplementary Table S1). and, (c) C:N ratio from leaves and stems of melon plants measured after a previous acclimation period (A.P.) to two or six weeks to elevated CO 2 (eCO 2 ) (700 ppm) or ambient CO 2 (aCO 2 ) (400 ppm). Mean values ± SE are shown (n = 9). **(P ≤ 0.01) and ***(P ≤ 0.001) when statistically significant differences were found, ns (no statistically significant differences) (Two-way ANOVA and LSD tests). Melon leaf C:N ratio data were transformed by log (x + 1). Amino acids content. In total, 18 individual amino acids were detected in melon leaves, whereas only 15 were determined in melon stems. Elevated CO 2 compared to aCO 2 significantly decreased the concentration of Alanine (49%), Asparagine (65%), Glycine (71%), Isoleucine (44%), Lysine (76%), Serine (59%), Threonine (50%), and Valine (55%) in melon leaves (Fig. 3, Supplementary Table S2). Methionine significantly increased its concentration by 95% under 6 weeks of acclimation period compared to 2 weeks (Fig. 4, Supplementary  Table S2).
In melon stems, effects of both factors CO 2 and acclimation period were less pronounced than in melon leaves. Elevated CO 2 significantly increased Tyrosine content (93%), whereas Tryptophan content was reduced under eCO 2 (51%) compared to aCO 2 (Fig. 3, Supplementary Table 2). After 6 weeks of acclimation period a significant reduction of Threonine content (67%) was scored compared to 2 weeks of acclimation period (Fig. 4, Supplementary Table S2). The interaction between CO 2 concentration and acclimation period was significantly different on Asparagine content in melon stems (Supplementary Table S2).
Carbohydrates content. The content of sugars on melon leaves and stems was significantly affected by CO 2 concentration or by acclimation period to CO 2 . There was no interaction between the two factors (Supplementary Table S3). Elevated CO 2 compared to aCO 2 significantly increased the concentration of sucrose (86%) and significantly decreased the concentration of mannitol (47%), sorbitol (37%) and xylitol (63%) in melon leaves (Fig. 3, Supplementary Table S3). Fructose, maltose and trehalose were significantly affected by acclimation period on melon leaves, increasing their concentration by 141%, 650% and 854% respectively after 6 weeks of acclimation period compared to 2 weeks (Fig. 4, Supplementary Table S3).
Effects of CO 2 on Aphis gossypii adult weight. Dry body mass of F1 A. gossypii adults was significantly affected by the CO 2 concentration level and by the acclimation period of 2 and 6 weeks. F1 aphid body mass significantly decreased under eCO 2 compared to aCO 2 (F 1,43 = 23.044, P ≤ 0.001). Furthermore, F1 aphid body mass was significantly lower when aphids fed on plants previous exposed to a longer acclimation period of 6 weeks compared to the shorter acclimation period of 2 weeks (F 1,43 = 10.940, P = 0.002) (Fig. 5).
Dry body mass of F2 A. gossypii adults was significantly affected by the CO 2 concentration level depending on the acclimation period of 2 and 6 weeks (F2: acclimation period × CO 2 : F 1,41 = 22.992, P ≤ 0.001). Due to the significant interaction, data was analysed by LSD pairwise comparison. There was a significant decrease in the body mass when aphids fed on melon plants previously acclimated for 6 weeks to eCO 2 (119.27 ± 6.29 µg) compared to 2 weeks under eCO 2 (165.27 ± 4.14 µg). Furthermore, there was a significant loss of weight on Figure 2. Melon biomass. Dry weight of leaves and stems of melon plants measured after a previous acclimation period (A.P.) to two or six weeks to elevated CO 2 (eCO 2 ) (700 ppm) or ambient CO 2 (aCO 2 ) (400 ppm). Mean values ± SE are shown (n = 10). ***(P ≤ 0.001) when statistically significant differences were found, ns (no significant differences) (Two-way ANOVA and LSD tests).  (Fig. 5). Consequently, the dry weight of the aphid grown on plants previously acclimated during 6 weeks at eCO 2 decreased for both A. gossypii generations, whereas this effect was more pronounced in the second generation.
Effects of CO 2 on Aphis gossypii colony performance. Aphis gossypii population performance differed when colony was reared under aCO 2 or eCO 2 conditions on plants previously acclimated for 2 weeks to the respective CO 2 concentration. No statistical differences were observed in day 14 (Colony growth rate: t = − 1.550, df = 18, P = 0.139). However, the number of aphids decreased by 23% in the A. gossypii colony under eCO 2 in day 21, with fewer nymphs number, and subsequently less colony growth rate, compared to the colony developed under aCO 2 concentration (Nymphs growth rate: t = − 2.675, df = 18, P = 0.015; Colony growth rate: t = − 2.486, df = 18, P = 0.023) (Fig. 6, Supplementary Table S4).

Discussion
This research emphasizes how the increase in atmospheric CO 2 , main driver of climate change, generates changes in plant nutritional quality and subsequently, influences pest insect performance. We mainly focused on how different plant acclimation to eCO 2 modified the content of carbohydrates and amino acids of melon plants and affected A. gossypii body mass and population growth. Under eCO 2 , melon plants decreased N foliar concentration and increased C:N ratio, independently of acclimation period. Elevated CO 2 leaded to changes in primary metabolites, significantly reducing the content of some amino acids and increasing some carbohydrates. Few carbohydrates were influenced by acclimation period, increasing their content under longer exposure to experimental climate conditions. Due to the importance of amino acids for aphid nutrition, the dilution of the foliar content of some essential amino acids could have aggravated the reduction in the population growth of A. gossypii reared on melon plants previous acclimated two weeks to eCO 2 , and the loss of aphid body mass from two successive generations of A. gossypii reared under eCO 2 on plants previous acclimated 2 or 6 weeks. Furthermore, the drop in aphid body mass was more pronounced when reared longer period under eCO 2 (6 weeks of acclimation compared to 2 weeks), and more marked in the second generation of A. gossypii compared to the first generation.
Atmospheric CO 2 enrichment usually promotes an increase in plant biomass 6,8,10,13,23,29,32,41 , as occurred in melon plants under eCO 2 in our experiment. Furthermore, the significant decrease in foliar N concentration Results of the effect of CO 2 concentration (melon leaf and stem separately) by Two-way ANOVA and LSD tests are denoted by asterisks: *(P ≤ 0.05), **(P ≤ 0.01) and ***(P ≤ 0.001) when statistically significant differences were found.  19 , we used foliar amino acids and carbohydrates in order to relate biochemical compounds with nutritional quality of melon plants for aphids under eCO 2 . In our study, five essential amino acids: isoleucine, lysine, threonine and valine (in leaves) and tryptophan (in stems), as well as other amino acids, such as alanine, asparagine, glycine, and serine (in leaves), reduced their content under eCO 2 , consequently affecting the N quality and therefore, the nutritional value of melon plants for A. gossypii reared under these climatic conditions. In consonance with the research performed by Sun et al. 25 , alanine, glycine, lysine, threonine and tryptophan content also decreased in cotton phloem sap under eCO 2 , forcing A. gossypii to ingest more phloem sap to satisfy its nutritional requirements. Alanine was also significantly reduced, although histidine and tryptophan increased their concentrations, in wheat (Triticum aestivum L.) under eCO 2 27 . Ryan et al. 12 observed that arginine, aspartate (aspartic acid), glutamine and valine in the pasture grass Schedonorus arundinaceus Schreb were correlated with Rhopalosiphum padi L. abundance, but only valine appeared to decrease due to eCO 2 , explaining to some extent the decrease in aphid performance. Accordingly, Myzus persicae Sulzer was negatively affected by the decrease in individual amino acid concentrations in oilseed rape (Brassica napus L.) 20 .

Scientific Reports
When evaluating the role of important amino acids for Aphis fabae Scopoli growth, alanine and proline were considered primarily phagostimulants, and serine also stimulated the (artificial) diet intake 44 . When histidine, methionine, threonine, valine, and possibly tryptophan, were lacking in artificial diets to test different clones of A. fabae nutritional requirements, its individual fitness was impaired 19 . The lack of histidine, isoleucine or methionine reduced the feeding rate of M. persicae, decreasing its growth rate 45 . Aphid feeding behaviour, development, fecundity and size could be impaired on N-deficient plants [46][47][48][49][50] . Therefore, the reduction in the content of some amino acids could have altered A. gossypii feeding requirements and partially explain the negative effect on aphid weight and colony growth under eCO 2 . First and second generation were examined separately. *(P ≤ 0.05), **(P ≤ 0.01) and ***(P ≤ 0.001) when statistically significant differences were found, ns (no significant differences) (Two-way ANOVA and LSD tests). A.P. = Acclimation Period to the respective CO 2 concentration.  www.nature.com/scientificreports/ Accordingly to our results, Rhopalosiphum maidis Fitch decreased its body weight, fecundity and intrinsic population growth rate when reared on barley under eCO 2 due to a significant reduction in crude protein, total amino acids and most of the free amino acids concentrations 24 . However, our results differ from those of Jiang et al. 28 , in which A. gossypii increased its fecundity, body weight and population abundance under eCO 2 , due to an increase in free amino acids and soluble proteins in cotton plants. To explain the divergent aphid performance, we should take into account: (1) the different host plant (melon vs cotton), (2) different clones of the same aphid species may have diverse amino acids requirements 19 , (3) the variation in the pattern of essential amino acids synthetized by the bacterial endosymbiont Buchnera between the different aphid clones 19,27 .
Unlike general predictions, a tendency of increase the relative abundance of essential amino acids for aphids was observed in barley phloem under eCO 2 , subsequently improving the performance of R. padi 22 . This cereal aphid also increased its relative growth rate in spring wheat under eCO 2 due to an increase in most of all the individual amino acids concentrations in phloem sap 20 . Furthermore, eCO 2 could differently change the content in foliar amino acids depending on the crop resistance to aphids, reducing the content in moderate resistant genotypes and increasing the content in resistant genotypes 23 .
Plant non-structural carbohydrates, such as starch and soluble sugars, usually increased under eCO 2 7 . Among soluble sugars, sucrose is not only a key phagostimulant for herbivorous insects, but is also important for aphid growth and development. Thus, the increase in sucrose levels in host plants could potentially enhance aphid performance 51 . However, under eCO 2 the rising in carbohydrates content can dilute the N nutrients required and finally counteract the positive effects of their increase 7 . This unbalance between carbohydrates and amino acids could increase aphids consumption rates due to compensatory feeding, finally increasing the plant damage 7,11 .
The eCO 2 -effect on the content of each soluble sugar depends on the host plant and could modify aphid performance differently. Elevated CO 2 increase the concentration of fructose, mannitol and trehalose in wheat 11 and this change in host plant quality could have produced the increase in R. padi weight 27 . In barley, the total soluble sugar and glucose, fructose and sucrose contents were not affected by eCO 2 , but a reduction in crude protein and amino acids content, could have influenced aphid feeding, leading to a decrease in R. maidis fresh body weight, fecundity and intrinsic population growth rate 24 . Due to CO 2 enrichment, fructose and glucose concentrations increased in spring wheat but sucrose remains unchanged compared to aCO 2 , and together with a significant increase in individual amino acids concentrations, positively affected R. padi relative growth rate. Sucrose and individual amino acids did not change significantly their concentrations in oilseed rape under eCO 2 , finally negatively affected M. persicae relative growth rate 20 .
In our study, eCO 2 significantly increased sucrose content on melon leaves and stems. Galactose, maltose and trehalose, aphid feeding stimulants as sucrose 52 also increased their content in melon stems under eCO 2 . However, the concentrations of sugar alcohols mannitol, sorbitol and xylitol were reduced in melon plants under eCO 2 . Mannitol and sorbitol are organic osmolytes that protect aphids and whiteflies from osmotic stress and not-optimal developmental temperatures 53 . Therefore, a reduction in these polyols and some essential amino acids in melon leaves under eCO 2 , could potentially have impaired A. gossypii performance, decreasing aphid weight and the colony growth rate.
Our research showed how changes in plant biochemistry due to eCO 2 have negatively affected A. gossypii performance. However, it is difficult to generalize the effect of eCO 2 for phloem-feeders 12 , because their responses to eCO 2 are heterogeneous and, in the case of aphids, the effects could be species-specific 17,54 or even genotypespecific 23,43 . In fact, aphid populations under eCO 2 could decrease, in accordance with our results 13,17,30,31,42,55,56 , but also increase 17,55 , or even being unaffected by eCO 2 17 . In contrast, other insect feeding guilds responds more homogenously to eCO 2 . For instance, due to the dilution of N under eCO 2 , chewing insects show compensatory feeding, increasing their food consumption due to the lower food quality 14 . While leaf-chewers do not seem to have adverse effects on development and pupal weight under eCO 2 14 , leaf-miners decrease their abundance and increase their development time 7 .
We observed that the acclimation period did not significantly affect melon C and N concentration and biomass. However, the differences between eCO 2 and aCO 2 observed under 6 weeks of acclimation were usually higher than under 2 weeks. The content of some amino acids and carbohydrates significantly increased after 6 weeks of acclimation compared to 2 weeks. Furthermore, the difference in F1 and F2 aphid biomass between eCO 2 and aCO 2 was greater under 6 weeks of acclimation than under 2 weeks. In general, the effects of CO 2 on plants and aphids were higher under longer exposure; although a short period of previous plant acclimation to CO 2 could be enough to detect eCO 2 effects on aphid biomass. Klaiber et al. 30,31 also showed that changes in plant and aphids were higher after longer exposure of plants to eCO 2 .
In our study, we observed differences in some amino acids and carbohydrates depending on the plant part analysed (stems or leaves) under eCO 2 . This could indicate differences in the aphid niche establishment and the consequent plant colonization 36 . Further research analysing sequential sampling at distinct melon growth stages or after different moments of plant colonization by aphids could provide in-depth information about how aphid infestation could lead to change in plant biochemistry under eCO 2 and the subsequent impacts on plant-aphid interactions 36,57 .
In conclusion, although the change in nutritional quality of melon plants under eCO 2 has damaged A. gossypii performance, and this could be thinkable as positive for pest control, the changes in foliar amino acids and carbohydrates content could make plants more palatable for other herbivore insects, or even produce a different effect on other aphid species, even on other aphid clones, which also feed on cucurbits. Therefore, further research is needed to elucidate the effect of eCO 2 on melon crop and its associated herbivorous insects, ideally analysed in open-chambers or in a free air CO 2 enrichment facility in order to generate more realistic predictions about how climate change affects trophic interactions in agroecosystems.

Data availability
The datasets generated during the current study are available from the corresponding author on reasonable request. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.