During the experiment, three plants died . Figure 4.2 shows the mean mass balance for the systems at the end of the experiment, depicting the fractions of the spiked 14C present in plant tissues, in the used nutrient solution, and as unaccounted activity. The unaccounted activity reflected the 14C that was not found in the nutrient solution at the time of solution renewal or in the plant tissues after harvest and may include losses via unidentified processes, such as volatilization, microbial mineralization in the nutrient solution , or stomatal release. Activity in each fraction varied across compounds and to a lesser degree across plant species, suggesting specificity to uptake. Figure 4.3 shows the cumulative 14C dissipation from the nutrient solution as calculated from the difference in activity in the solution at the beginning and end of each 3 d interval of solution renewal, representing 14C loss from plant uptake and other processes. Dissipation followed the decreasing order of BPA > NP > DCL > NPX for all treatments and occurred at a similar rate throughout the 21 d cultivation. The presence of plants significantly increased the dissipation of PPCP/EDCs from the nutrient solution, except for NP. For example, the initial concentration of 14C-DCL in the nutrient solution was 105.3 ± 0.3 dpm/mL, but it decreased to only 32.8 ± 1.9 dpm/mL after 3 d in the presence of lettuce,vertical indoor farming while 91.2 ± 3.2 dpm/mL remained in the no-plant control. Lettuce and collards treatments had different levels of chemical dissipation in the nutrient solution. For example, the overall dissipation of BPA in the lettuce treatment was 69.1 ± 8.7%, as compared to 88.4 ± 5.3% in the collards treatment .

Different compounds also dissipated at different rates. For instance, in the presence of collards, the cumulative loss was 88.4 ± 5% for BPA, 55.6 ± 11.8% for DCL, and 45.5 ± 4.3% for NPX. The dissipation of NP in the solutions with plants was found to be similar to that in the no-plant control, especially for the lettuce treatment . The loss of NP from the no-plant control was likely associated with volatilization, as continuous aeration was used to maintain the oxygen level in the nutrient solution throughout the experiment. The Henry’s Law constant for NP is 1.09 × 104 atm m3 mol-1 , suggesting a tendency for volatilization. The loss of NP in the solution was found to be insignificant, as all of the spiked 14C was found in the solution , and asolvent rinse of the system showed little sorption of 14C-NP on the container wall . Doucette et al. found that in a hydroponic set up, about 13% of the spiked NP was lost to volatilization in the absence of plants. The increased volatilization losses in the current study were likely due to specific aeration and temperature conditions used. Despite volatilization losses, significant amounts of 14C were detected in plant tissues, suggesting that both collards and lettuce accumulated NP .Noureddin et al. studied the uptake of 5 mg/L BPA from hydroponic solution by water convolvulus and found that approximately 75% of the spiked BPA was removed after 3 d. This removal was comparable to that observed for BPA with lettuce in this study, but was smaller than that with collards . CalderónPreciado et al. evaluated hydroponic uptake of triclosan, hydrocinnamic acid, tonalide, ibuprofen, naproxen, and clofibric acid by lettuce and spath and showed that the removal of NPX from solution was about 70% for lettuce and 10% for spath after 3 d. In comparison, Matamoros et al. observed less than 10% removal of NPX after 3 d of hydroponic growth with wetland plants , while 46% removal of NPX was measured in the collards treatment in the present study. Matamoros et al. also showed that DCL did not dissipate appreciably in treatments with wetland plants, which was in contrast to the high removal of DCL by leafy vegetables observed in this study . It is likely that the smaller plant mass and the use of non-aerated nutrient solution in the earlier study contributed to the limited plant uptake.

The range of variation suggests that plant species, along with other factors such as plant mass and environmental conditions, affect the actual accumulation of PPCP/EDCs into plant tissues.Plant tissues were collected after 21 d of cultivation, rinsed with deionized water, and separated into roots, stems, new leaves, and original leaves for analysis of both extractable and non-extractable 14C. Table 4.1 shows concentrations of 14C in plant tissues, expressed as parent-equivalents. In agreement with the dissipation trends in solution, plant accumulation followed the decreasing order of BPA > NP > DCL > NPX. Concentrations based on dry plant mass ranged from 0.22 ± 0.03 to 12.12 ± 1.91 ng/g in leaves and stems. Statistical analysis showed that the accumulation in leaves and stems was not significantly different between lettuce and collards, or among the different compounds. In contrast, roots accumulated significantly more 14C than all the other plant tissues, with concentrations that ranged from 71.08 ± 12.12 to 926.89 ± 212.89 ng/g. Accumulation of 14C in plant tissues exhibited several apparent trends. In whole collards plants, significantly greater accumulation was found for the neutral compounds BPA and NP than the anionic compounds DCL and NPX , suggesting that the charge state of PPCP/EDCs may greatly influence plant uptake . Similar effects have been frequently observed for anionic herbicides, and are attributed to exclusion of negatively charged molecules by cell membranes . Between lettuce and collards, lettuce significantly accumulated less PPCP/EDC when all test compounds were pooled , although the difference for individual compounds was not significant . Accumulation of BPA or NP in plant roots was significantly higher for collards than lettuce , while portion of DCL accumulated into lettuce and collards roots was not significantly different. Analysis of tissue extracted with solvent showed that essentially all of the 14C was non-extractable; only the root samples from NP-collards treatment contained a detectable fraction of 14C in extracts .

Combustion of extracted plant tissues confirmed that almost all 14C remained as non-extractable residue, one possible endpoint for xenobiotics taken up by plants . Only a few studies have examined the plant uptake of some of the same PPCP/EDCs considered in this study. Wu et al. grew iceberg lettuce and spinach for 21 d in hydroponic solution initially spiked with a suite of 19 PPCPs, including DCL and NPX, each at 500 ng/L and found no detectableresidues of DCL or NPX, except for NPX in spinach at 0.04 ng/g. Calderón-Preciado et al. analyzed apple tree leaves and alfalfa from fields irrigated with water containing BPA, DCL, and NPX. DCL was detected at 0.354 ng/g in apple leaves and 0.198 ng/g in alfalfa; NPX was detected at 0.043 ng/g and 0.04 ng/g, respectively. The low concentrations found in these studies generally agree with the findings of this study, but there is some variation in the tendency for specific compounds to accumulate. This variation may be partly attributed to the different analytical approaches. In other studies, uptake of PPCP/EDCs by plants was evaluated using non-labeled compounds, and accumulation was measured by targeted chromatographic analysis for the extractable parent compound. The use of 14C-labeled compounds in the current study should have provided “worst-case” estimates of human exposure,vertical growing towers as the concentrations included nonextractable residue and likely also included transformation products. Transformation products may be an important component of potential risk since the metabolites of some PPCP/EDCs have higher biological activity than their parents and studies have shown that a large portion of PPCP/EDCs that are taken up by plants may be transformed in vivo . A translocation factor , which was the total 14C in stems, new leaves, and original leaves divided by the 14C in roots, was calculated . These TFs were consistently very small, demonstrating the poor translocation of these PPCP/EDCs from roots to upper tissues after uptake. The TF values followed the decreasing order of NPX > DCL > NP > BPA, the opposite observed for plant accumulation. Lettuce displayed lower TFs than collards for the same PPCP/EDCs. For example, the mean TF for BPA was only 0.010 ± 0.003 for lettuce, but was 0.051 ± 0.008 for collards. The much greater accumulation of PPCP/EDCs in roots, as compared to leaves, has been observed in previous studies. For instance, Herklotz et al. found that the levels in leaves were 0.00952 – 0.00503 of those in roots for cabbage grown in nutrient solution spiked with carbamazepine, salbutamol, sulfamethoxazole, and trimethoprim.

Doucette et al. reported that the accumulation of NP in leaves was 0.0233 – 0.0167 of that in the roots of crested wheatgrass grown in solution. The poor translocation of the selected PPCP/EDCs from roots to leaves may be attributed to several factors. The compounds considered in this study have moderately high hydrophobicity with log Kow from 3.35 to 4.48 . Translocation of organic compounds within plants generally decreases with increasing hydrophobicity . Also, roots have higher lipid content than most other plant tissues, and neutral compounds have been shown to be preferentially distributed in tissues with high lipid content . In addition, the rapid conversion of 14C residue to the non-extractable form, as discussed above, may be another important factor for the negligible transfer from roots to other plant tissues. The use of 14C labeling, while giving unique information such as the total chemical accumulation in plant tissues, did not provide insights on the chemical composition of the accumulated residue. It is likely that some of the PPCP/EDCs were transformed in the nutrient solution before they were taken up by plants. The used nutrient solution from hydroponic cultivation was subjected to fractionation on HPLC to characterize the portions of 14C existing as parent compound and transformation products . It is evident that different PPCP/EDCs were transformed to different degrees in the nutrient solution and the presence of plants generally enhanced the transformation. In the no-plant control of DCL and NPX, the majority of 14C was in the form of the parent compound , while the percentage of 14C in the SPE aqueous filtrate or eluted on HPLC prior to the parent compound was very small . The presence of lettuce or collards did not increase the transformation of DCL or NPX, with the exception of the DCL-lettuce treatment, where 93.8 ± 6.2% of the recovered activity was detected in the SPE aqueous filtrate. In contrast, BPA and NP were extensively transformed, even in the absence of plants, and transformation was accelerated in the presence of a plant. For example, 50.3 ± 24.3% of the recovered 14C was identified as the parent in the BPA no plant control, but collards and lettuce treatments had no detectable BPA. In the presence of a plant, 14C was detected in the aqueous filtrate and in HPLC eluent prior to the retention time for BPA . Extensive transformation of NP was also observed; all of the 14C from lettuce or collards cultivation was found in the aqueous phase of the extraction . The fraction of activity in aqueous phases may be attributed to transformation products that were not retained by the HLB cartridge or solvent phase during solvent extraction . Preliminary experiments showed that an average of 93.6% of 14C-BPA, 84.5% of 14C-DCL, and 92.0% of 14C-NPX were recovered from the HLB cartridges and 97.8% of the spiked 14C-NP was recovered in the solvent phase, while the activity in aqueous phases were below detection. Therefore, 14C in the SPE aqueous filtrate for BPA, DCL, and NPX, or in the aqueous phase for NP, was likely from polar transformation products containing the 14C label. The detection of transformation products in used solution suggests that some of the 14C found in plant tissues may be from transformation products formed in the nutrient solution prior to plant uptake. The demonstrated accumulation of PPCP/EDCs into leafy vegetables suggests a potential risk to humans through dietary uptake. To assess whether the concentrations detected in plant tissues in this study may present a potential human health risk, an individual’s annual exposure was estimated using values from the U.S. Environmental Protection Agency for average daily consumption of leafy vegetables .