At 52 days after planting, all plants were harvested.At harvest, the lettuce plants were gently removed from the soil, thoroughly rinsed with tap water for 5 min and then rinsed with NanoPure water three times. Leaf tissue was carefully separated from vascular and mesophyll tissues Figure S1. Mesophyll and root tissues were ground in liquid nitrogen and lyophilized for 5 days. Part of the freeze-dried mesophyll tissues were sent to UC Davis for metabolomics analysis, and another portion was oven dried at 70 °C for ICP-MS analysis. Since only a small amount of vascular tissue was available, it was only oven-dried for metal analysis.Dried plant tissues were digested with a mixture of 4 mL of H2O2 and 1 mL of plasma pure HNO3 using a microwave oven system at 165 °C for 1 h. The standard reference materials NIST 1547 and 1570a were also digested and analyzed as samples. The recoveries for all elements were between 90 and 99%. Cu and other six important mineral elements were analyzed using inductively coupled plasma mass spectrometry .The freeze-dried lettuce tissues samples were subjected to GC-TOF-MS analysis at the Genome Center Core Services, University of California Davis to identify the metabolites present in lettuce tissues. A description of sample pretreatment, analytical method and instrument has been described by Fiehn et al.Briefly, an Agilent 6890 gas chromatograph containing a Rtx-5Sil MS column with an additional 10 mm integrated guard column was used to run the samples, controlled using Leco ChromaTOF software version 2.32 . Quantification was reported as peak height using the unique ion as default. Metabolites were unambiguously assigned by the BinBase identifier numbers using retention index and mass spectrum as the two most important identification criteria. More details regarding data acquisition, data processing and data reporting are provided in the SI.Partial least-squares discriminant analysis is a supervised clustering method,strawberry gutter system which uses a multiple linear regression technique to maximize the separation between groups and helps to understand which variables carry the class separating information.

PLS-DA was run based on GC-TOFMS data using online resources .Variable Importance in Projection is the weighted sum of the squares of the PLS-DA analysis, which indicates the importance of a variable to the entire model.A variable with a VIP above 1 is regarded as significant.Biological pathway analysis was performed based on all detected metabolites data using MetaboAnalyst 2.0.The impact value threshold calculated for pathway identification was set at 0.1.The concentration of Cu, nutrient element and biomass was statistically analyzed using an independent two sample t test to determine whether concentration levels were significantly different between control and nanopesticides treatments. P-values were calculated with a two-tailed distribution.Cu in leaves increased in a dose-dependent manner in both covered and uncovered soil . Cu increased 82−140 times in vascular and 115−184 times in mesophyll tissues relative to the control, which indicates a high bio-accumulation of copper/nanoparticles in leaf tissues. Even though the leaves were thoroughly washed, copper ions and NPs remained on the surface or were incorporated into leaf tissues; it is likely that washing was not 100% efficient in removing them. Leaf exudates can form weak acids in the presence of water,which can accelerate dissolution of Cu2 nanopesticide, releasing cupric ions as long as the water remains on the leaf. This may result in a pathway for cupric ions to penetrate the epidermis cells and translocate to other tissues. In addition to ionized Cu, nanoparticles smaller than the stomatal diameter may enter past the guard cells. Stomatal diameters range from 8 to 12 μm for several species.Even though trichomes are not abundant on lettuce leaf surfaces, ESEM images taken after 24 h exposure to Cu2 nanopesticides show that many small particles were deposited on the lettuce leaf surface and stomatal cavities. The typical diameter observed for lettuce stomata is 13.1 μm , which is large enough to permit entry to Cu2 nanopesticides aggregates with an observed hydrodynamic diameter of 1530 ± 580 nm. Eichert demonstrated that 43 nm NPs entered stomata and migrated along the surface of stomatal pores.

After passing the stomatal guard cells, the NPs may either attach to cell walls or move between cell walls. For example, Stamenkovic and Gustin showed the majority of foliar Hg was located in epidermal and stomatal cell walls and was rarely found in mesophyll or vascular tissue.However, Hong showed that Ce was present in cucumber root phloem after foliar application of CeO2 NPs.As seen in Table 1, the average [Cu] in control roots is 6.0 mg/kg, while in treated plants, [Cu] in root is 17.5−26.1 mg/kg in covered soil and 34.2−56.9 mg/kg in uncovered soil. Statistical analysis showed all the NP treated plants have significantly higher [Cu] in roots compared to controls, even though application was only foliar. In covered soil, where no direct root uptake could occur, Cu in the roots was translocated from the leaves via phloem loading. Liao et al. showed that some xylem-transported Cu was recirculated to roots via the phloem in chicory and tomato plants.Even though we observed evidence of Cu translocation to the roots, 97−99% of Cu mass was retained in the leaves. In addition, the translocation rate in NP-treated plants was 0.009 to 0.014, which is far lower than that in the control . This indicates plants sequestered most of the Cu in leaves.The threshold level for Cu to induce toxicity in plants is 20−30 mg/kg.However, high concentrations of Cu in lettuce leaves did not cause any visible toxic symptoms throughout the entire exposure period . On the contrary, the leaf biomass significantly increased at low and medium levels for uncovered treatment and medium level for covered treatment . Since a high amount of Cu was retained in leaf tissues but did not induce any toxic symptoms, lettuces likely employ a detoxification mechanism to build tolerance to excess Cu.Using untargeted GC-TOF-MS, a total of 352 compounds were detected in lettuce leaves, and 159 metabolites were identified. To visualize the overall changes in metabolites between control and groups treated at different levels of Cu2 nanopesticides, PLS-DA analyses of all detected compounds were performed.

The score plot shows that all treated groups were clearly separated from the control along the first principal axis , which explained 27% of the total variability. This indicates foliar application of Cu2 nanopesticides significantly altered the metabolite profiles of lettuce leaves. Since there is not much separation in metabolite concentrations between different exposure levels, it is likely that the lowest exposure level already reached a threshold value for a metabolic response. In order to further identify the metabolites responsible for this separation, we performed PLS-DA based on the 159 identified compounds and determined their VIPs . As seen in SI Figure S5, there were 42 metabolites with VIP > 1, which are the ones that play an important role in group separation.Those metabolites include carboxylic acids , amino acids , amines , sugars , fatty acid and other metabolites. An independent two sample t test was also performed to screen metabolites which were significantly different from the controls in all treatment groups. t test results showed that 50 metabolites were significantly different from controls . Twenty one of the metabolites listed in SI Table S1 overlapped with high VIP score metabolites,grow strawberry in containers shown in SI Figure S3. A number of metabolites, that were not responsible for group separation but were significantly modified as determined by a t test, are of special interest. Four metabolites including cis-caffeic acid, chlorogenic acid, 3,4-dihydroxycinnamic acid, dehydroascorbic acid were significantly decreased in leaves exposed to Cu2 nanopesticides compared to the control. In addition, oxalic and threonic acids were also down-regulated. Jansson et al. reported that the dehydroascorbic acid formed was decomposed to oxalic acid and threonic acid by hydrogen peroxide generated from Cu auto oxidation in the presence of oxygen.Therefore, the down-regulation of oxalic and threonic acids is due to decreased dehydroascorbic acid. Pathway analysis indicated that six biological pathways were significantly perturbed : glycine, serine and threonine metabolism; alanine, aspartate and glutamate metabolism; tricarboxylic cycle; pantothenateand coenzyme-A biosynthesis; glycolysis or gluconeogenesis; and pyruvate metabolism. Cu2 NP treated plant leaves exhibited lower levels of TCA cycle intermediates, such as citric, isocitric, and fumaric acids; downregulation of TCA cycle appears to be a clear response. Pidatala et al., also observed that Pb induced TCA cycle disturbance in Vetiver plants. Exposure to Cu2 nanopesticides increased the levels of pyruvic acid 2−5 times compared to the control, indicating three biological pathways in which pyruvate participates were likely perturbed. In previous studies, soluble sugars were highly sensitive to environmental stress; these sugars play an important role in signaling and stress defense.In this study, sucrose, glucose, fructose and hexose concentrations did not change after exposure to Cu2 nanopesticides.

Although treated and control plants can be clearly separated in the PLS-DA analysis, it was not possible to distinguish the different dose levels. It is possible that the lowest dose already reached a threshold that induced changes in the metabolite levels. The same plants that exhibited increased leaf biomass also had the most separation from the control group.As mentioned before, only 1−3% of Cu mass was translocated from leaves to roots, and root biomass was not impacted by foliar exposure to Cu2 nanopesticides. As hypothesized, foliar exposure resulted in minor metabolite profile changes in roots compared to changes in leaves. t test statistical analysis showed that 20 metabolites were significantly different from the control, which is much lower than in leaves . There was noticeable separation between control and treated groups, based on PLS-DA analysis of all detected compounds in roots, but not as clear separation as in leaves . Treated groups were separated from the control along the third principal axis , which explained 8.4% of the total variability . Screened by VIP score, metabolites which were responsible for the separation include several amino acids: proline, glycine, leucine, phenylalanine, methionine, aspartic acid, oxoproline, isoleucine, glutaminc acid, citrulline, threonine, GABA, and serine . Biological pathway analysis indicates that three pathways related to amino acids were disturbed: phenylalanine metabolism; arginine and proline metabolism; and alanine, aspartate and glutamate metabolism. Thus, root metabolism was perturbed less than leaves, and the perturbed pathways were different from those perturbed in leaves. Among the 20 significantly changed root metabolites, 14 overlapped with changed leaf metabolites. This may indicate metabolites produced in the leaves exposed to copper were transported to the roots.Inorganic ions play an important role in many metabolic processes. Previous metabolomics studies considered only organic compounds, which may not provide a comprehensive view of the changes. Compared to the control, only K+ levels in leaves were significantly increased in all Cu2 nanopesticide-treated plants , with no statistically significant changes in root levels after exposure. Lahoavist et al.showed that hydroxyl radical activated the permeable conductance of K+ in Arabidopsis. Demidchik et al.showed that free oxygen radicals regulate plasma membrane Ca2+ and K+ permeable channels in plant root cells. Therefore, it is possible that exposure to Cu2 nanopesticides and released Cu2+ generated hydroxyl radicals that induced K+ imbalance. A number of studies have shown that copper nanoparticles are able to trigger the generation of free radicals in cell membrane.To test this, one-month-old lettuce plants were exposed to 105 mg/100 mL of Cu2 nanopesticides. At day 3, leaf samples were collected and incubated with 2′,7′-dichlorofluorescin diacetate and observed by confocal laser scanning microscopy to localize Reactive Oxygen Species . Representative leaf surfaces ofcontrol and Cu2 nanopesticide treated were imaged by confocal laser scanning microscopy. The green fluorescence signals, which represent the presence of ROS, were found to be much higher in leaves exposed to Cu2 nanopesticide , compared to the control .We hypothesize that lettuces must have antioxidant defenses and copper detoxification mechanisms, since there were no visual symptoms of damage on the leaves. Based on the metabolomics we aimed to identify the underlying mechanisms . Cu Chelation. Previous studies have demonstrated that important mechanisms for plant tolerance of copper are chelation and sequestration, from the upregulated production of organic acids, amino acids, peptides, and polyamines. We found nicotianamine levels were significantly increased in all NPs treated plants, 12−27 times higher than that in controls.