Functional Group Properties and Position Drive Differences in Xenobiotic Plant Uptake Rates, but Metabolism Shares a Similar Pathway

Plant uptake of xenobiotic compounds is crucial for phytoremediation (including green stormwater infrastructure) and exposure potential during crop irrigation with recycled water. Experimentally determining the plant uptake for every relevant chemical is impractical; therefore, illuminating the role of specific functional groups on the uptake of trace organic contaminants is needed to enhance predictive power. We used benzimidazole derivatives to probe the impact of functional group electrostatic properties and position on plant uptake and metabolism using the hydroponic model plant Arabidopsis thaliana. The greatest plant uptake rates occurred with an electron-withdrawing functional group at the 2 position; however, uptake was still observed with an electron-donating group. An electron-donating group at the 1 position significantly slowed uptake for both benzimidazole- and benzotriazole-based molecules used in this study, indicating possible steric effects. For unsubstituted benzimidazole and benzotriazole structures, the additional heterocyclic nitrogen in benzotriazole increased plant uptake rates compared to benzimidazole. Analysis of quantitative structure–activity relationship parameters for the studied compounds implicates energy-related molecular descriptors as uptake drivers. Despite significantly varied uptake rates, compounds with different functional groups yielded shared metabolites, including an impact on endogenous glutathione production. Although the topic is complex and influenced by multiple factors in the field, this study provides insights into the impact of functional groups on plant uptake, with implications for environmental fate and consumer exposure.


Chemicals
Benzimidazole and benzotriazole-based compounds were used in this work because they are taken up by plants, 1,2 with benzotriazole uptake known to exceed transpiration rate, 1 and their base structures vary by only a single nitrogen in the heterocyclic ring. Benzimidazole derivatives are commonly used fungicides [3][4][5] and benzotriazole is a widely used corrosion inhibitor; 6 both have high solubilities (Table S1) and are present in environmental waters. 7 Chemicals used in these experiments include (Table S1) Then 5 g sucrose (Research Products International) and DI water to 1 L was added. The pH was checked and adjusted to a 5.7 as needed with potassium hydroxide or hydrochloric acid. Before experimental use, the medium was filter-sterilized using a bottle top filter (Corning #431118, 0.22 µm pore size) into an autoclaved bottle.

Seed Sterilization Procedure
A previously published seed sterilization procedure was used with Arabidopsis thaliana Columbia ecotype "0" (Col-0) seeds, 18 with the following minor modifications: • Instead of conducting the procedures over a flame, seed sterilization was conducted in a biosafety cabinet.
• Rather than 50 µL of seed, between 10 and 50 µL of seed were used, depending on the quantity of plants required for the experiment.

Arabidopsis Growth Procedure
A previously published Arabidopsis growth procedure 18 was used to grow up the sterilized seeds, before exposure to isothiazolinones, with the following modifications: • 30 +/-2 seeds per box were used, placed into boxes with a pipette tip and visually counted • Growth chamber temperatures were 23°C during the light period and 21°C during the dark period • Plants were grown for 11-13 days before exposure to benzimidazole or benzotriazole based compounds. The average dry biomass of plants (per box) after this growth period was 0.03±0.005g (ave±stdev).

Plant Exposure Experiment Details
The exposure experiments were modeled on previous work. 1,17,18 After a 11-13 day period of growth in unspiked sterile hydroponic medium, the boxes were taken from the growth chamber into a biological safety cabinet and the following procedures conducted using sterile technique.
A master mix of medium (i.e., a large volume of medium was prepared to ensure consistency rather than each box being prepared separately) was spiked with the benzimidazole or benzotriazole of interest (one compound per treatment). 3-4 samples were taken from the master mix, at 0.6 mL each, and filtered with nylon filters (0.2 µm, 13 mm diameter, mdi SY13NN) into LC vials. These medium samples were frozen at -20°C at the end of each timepoint and kept frozen until analysis.
After master mix sampling for the t=0 timepoint, the microporous tape was removed from each plant Magenta box and the box tilted to allow for the medium to leave the box while the plant tissue remained in the box. The box lid was then removed and freshly spiked plant growth medium was added to each box, at 25 mL per box. The box lid and microporous tape were then replaced.
Additionally, an abiotic control was created at t=0 for each treatment. Each control replicate consisted of the same amount of master mix of medium as the plant boxes (25 mL), pipetted into a washed and autoclaved Magenta box. The lid was replaced and microporous tape applied in the same manner as the plant boxes. Each treatment and control was conducted at n=3-4.
Except for sampling, boxes were kept in the Percival growth chamber alternating between 16 hours light at 23° C and 8 hours dark at 21° C. Relative humidity was maintained at 50%. All sampling was conducted using sterile technique in the biosafety cabinet, as described above.

Neural network plant uptake model
The authors of a previously published plant uptake model focused on TSCF prediction 8 generously provided their model's TSCF predictions of BZ and Carb-BZ by inputting those molecules' log KOW, molecular weight, hydrogen bond donor, hydrogen bond acceptor, rotatable bonds, and polar surface area properties into the model.

Plant Tissue Harvest Details
The microporous tape was removed from each box and the box tilted with the lid still on to allow for the medium to drain out while the plants were retained in the box. The box was then inverted onto a clean paper towel, and the box removed. The plant tissue was gently patted with the paper towel to remove any remaining medium. Clean tweezers were then used to move the tissue into 1.5 or 2 mL microcentrifuge tubes with locking lids, with the tweezers cleaned with ethanol between each box. Plant tissue was then frozen at -20 °C until overnight freeze drying and extraction (below).

Plant Extraction for Metabolomics
The following procedure from previous work 1,18 was used after overnight freeze drying: A single stainless steel homogenization bead (5 mm) and 1.0 mL of 1:1 methanol/water solution were added to freeze-dried plant tissues in a microcentrifuge tube. The tubes were frozen at −80 °C for 30 min. Samples were thawed and placed on a Retsch mixer mill for 5 min at 30 Hz. The samples were then sonicated for 10 min, vortexed for 1 min, and centrifuged at 10 000 × g for 10 min. Following centrifugation the supernatant was removed with a 22G x 1 ½ BD precision glide needles needle syringe and filtered through a 0.2 μm, 13 mm diameter PFTE filter (mdi) into an empty autosampler vial. The extraction procedure was repeated sequentially two additional times by adding only 0.5 mL (rather than 1.0 mL) of the methanol:water solution for each subsequent extraction and otherwise exactly repeating the extraction procedure (i.e., homogenization, sonication, vortex, centrifugation, filtration). All three fractions were combined in a single autosampler vial for analysis.   QSAR parameters were determined based on each compound's ground state equilibrium geometry in water (we note that these computations can be impacted by the redox conditions, which can introduce some uncertainty). Computations were performed using a B97X-D/6-31G*  at a resolution of 0.0002 e/au 3 . A total of 33 descriptors were obtained directly from Spartan and describe the electrical, quantum, and geometric molecular properties of a given compound (Tables   S7 and S8).  Notes on these results: For a limited number of the compounds, data were collected to 10 days after exposure to the chemicals. We focused on the 48h period in the main presentation because contact time between plants and water is often limited (e.g., irrigation, bioinfiltration). These results revealed further separation of some of the 'moderate' removal group of compounds. 2A-BZ maintains the moderate removal found at 48 hours, ending at a C/C0 of 58% after 10 days. In contrast, the C/C0 of Carb-BZ was 25% at 10 days, and BZ is almost completely removed by 10 days (C/C0=5%). We also note differences within compounds in the 'moderate' removal group at the 48h point. For example, although 2-CBZ and 1-ABZ group together when compared with the 'no removal' and 'greatest removal' groups, at 48 hours each have a significantly higher C/C0 than Carb-BZ and 2-ABZ (both compounds vs. Carb-BZ p=0.03, both compounds vs. 2-ABZ p=0.0007), and BZ's C/C0 is significantly higher than 2-ABZ (p=0.02). Please also note that plant uptake is focused on depletion from the hydroponic medium; subsequent in planta degradation of compounds taken up by plants is not treated separately. Plant degradation of compounds outside the plant (i.e., via root exudates) has been previously shown to be not significant for other related TOrCs under aseptic hydroponic conditions. 1      Figure  S3) was available. These compounds are noted with an * below. Second-order curve fits (using least-squares regression) had equal to or higher r 2 values for all compounds than first-or zeroorder curve fits except for BZ (first-order r 2 = 0.97 vs. second-order r 2 = 0.95); 2Cl-BZ (r 2 = 0.81 for first-, second-, and zero-order); and 2N-BZ (first-order r 2 = 0.99 vs. second-order r 2 = 0.96), but the second-order rate values for these two compounds are provided below for comparative purposes. Second-order kinetics have been previously reported for similar hydroponic plant uptake kinetics results. [17][18][19] Note that rate constants here are unitless because the values are derived from relative concentration (C/C0) values.        Table S7. For the parameters above:

LC-MS/MS and MRM Transition Details
E HOMO is the Energy of the highest occupied molecular orbital (eV) E LUMO is the Energy of the lowest unoccupied molecular orbital (eV) Log P is the partitioning coeffient Min LocIonPot is the Minimum local ionization potential at the electron density surface (eV) Property Min(Surface2) is Minimum displayed local ionization potential mapped between 6-15 eV

Metabolomics via high-resolution mass spectrometry (overall summary)
Extracted plant tissues were analyzed on a Thermo Q-Exactive High-Resolution Mass Spectrometer, run in MS as well as both positive and negative data-dependent MS/MS modes.
The chromatography and method parameters from the Agilent Triple Quadrupole MS BZ method (Table S3) was used on the Q-Exactive for these samples. Q-Exactive data of the metabolomics samples were analyzed via Compound Discoverer 3.1. The workflow "Untargeted Metabolomics with Statistics Detect Unknowns with ID using Online Databases and mzLogic" was used. The results were filtered to remove background with the "background is false" filter, and to those compounds with a p-value ≤0.05 in the fold-change ratio (the ratio of the peak area of the treatment plant tissue extracts to the peak area in the positive control plant tissue extracts) for the compound of interest. Filtered results were then sorted from greatest to least peak area ratio.
Results with a peak area ratio ≥100 were selected for further analysis. Those compounds were then sorted by retention time, and similar retention time compounds (within 0.07 min) were grouped as likely in-source fragments. The compound was listed as a proposed structure if the accurate mass deviation between the proposed compound and measured m/z was <10ppm.

Compound Discoverer Analysis
Compound Discoverer (Thermo Scientific) analysis was run April 7, 2021 (Compound Discoverer version used not noted) using the .RAW files produced by the Q-Exactive on September 23, 2020. Run as samples were the three carbendazim-exposed plant tissue extract sample MS scan files, the seven benzimidazole-exposed plant tissue extract sample MS scan files, the seven 2-cyanobenzimidazole-exposed plant tissue extract sample MS scan files, and  Note the gear symbol with a minus sign on "Detect Compounds" and "Normalize Areas" indicates obsolete nodes as of February 9, 2023 when the screenshot was taken, but there were functional on April 7, 2021 when the analysis was run.
Workflow node details:  A much more restrictive 100-fold or greater change cutoff along with p-value of ≤0.05 was used for metabolite analysis in this paper. Features decreased in peak area in fungicide-exposed vs. unexposed plants were not examined further in this work, but HRMS data for these features is available upon request.

METABOLITE DETAILS
*ppm 68 = ((difference between measured m/z and exact mass of proposed ionized formula, in atomic mass units)/exact mass of proposed ionized formula)x10 6

Metabolite Mass Spectra and Structures
Metabolites in each category are arranged from smallest to largest retention time.
(NOTE: all proposed metabolites and fragment structures are drawn unionized. Exact masses of proposed ionized formulas presented in the tables above correspond to the ionization state given in the numbers presented in the spectra-generally plus or minus a proton from the structure drawn below).

Metabolites Shared Between All Three Fungicides
Gamma  Table) Unknown M427  Table) 118.05327 Accurate Mass: Benzimidazole (parent compound)  Table) Cyano-hydrolyzed CN-BZ   (Link to Summary Table)