One-step extraction and analysis of 45 contaminants of emerging concern using QuEChERS methodology and HR-MS in radish leaves and roots

The scarcity of freshwater has led to a considerable increase of the reuse of reclaimed wastewater for irrigation of field crops [1,2]. This practice potentially exposes agricultural produce to a large variety of xenobiotic compounds including contaminants of emerging concern (CECs) which have been widely recognized to be present in wastewater [3]. Common approaches for the extraction of CECs from crops rely on solid-liquid extraction [4], assisted solvent extraction [5], ultra-sound solvent extraction [6] and recently QuEChERS (QUick, Easy, CHeap, Effective, Rugged and Safe) [[7], [8]–9]. Here, eight QuEChERS-based methodologies were compared for their suitability to determine 45 CECs in roots and leaves of soil-grown radish. The key points of the method development were:• The development of two single-step analytical methods specific for radish root and leaves, after testing eight different approaches using QuEChERS extraction for the quantitation of 45 CECs. The analytical methodology selected requires minimal time and solvent, making it cost-effective.• Methods validation were performed at five concentrations levels (2, 5, 10, 50 and 200 ng g−1), with low limits of quantification between 0.01 and 0.32 ng g−1.• The two optimized methodologies may be applied to identify large number of compounds of different families in radish crop. However, validation will be needed to quantify compounds different from the target compounds of this paper.


Method details
The extraction of contaminants of emerging concern (CECs) from plants tissues is a challenging procedure due to the complexity of the matrix. Plants, including crops, contain lipids, proteins, fatty acids and a wide range of components [10 , 11] that could negatively interfere with the analytical performance, particularly in the ionization of MS-based detection [12] . In lasts years, the QuEChERS use for the extraction of organic compounds in biological matrices has grown. Although originally developed for pesticide residues analysis in food stuff [13] , some method modifications have been implemented in order to adapt QuEChERS to extract diverse compounds of interest from various matrices [14] . Here, after testing several QuEChERS variants, two simple but efficient approaches were developed to extract CECs from root and leaves of radish. The differences between roots and leaves matrices called for the use of two QuEChERS methods differing in the salt composition: for roots, the original extraction salts (OR), also known as non-buffered salts while CEN 15662 (EN) the bufferedsalts was chosen for leaves. Our validated analytical methodology is time-efficient, due to the reduced number of steps in the sample treatment, uses little organic solvent, and is sensitive enough to detect trace levels in radish by using LC-QToF-MS. Target analytes were selected among the most reported CECs in reclaimed wastewater and taking into account their wide diversities in terms of physicochemical properties.

Matrix preparation
For validation purposes, bunches of radish plants (leaves and roots) were bought from a local organic supermarket (Barcelona, Spain). Then, the whole plants were carefully hand-washed to remove any soil particles and roots were separated from leaves and frozen for at -20 °C for 48h, separately. Consequently, the tissues were lyophilized (LyoAlfa 6 system, Telstar Technologies, Terrassa, Spain) and ground to powder using a knife mill (Grindomix GM 200, Retsch, GmbH, Haan, Germany) and finally stored at -20 °C for the method development.

Extraction procedure
Eight QuEChERS analytical protocols were tested with two different extraction solvents with and without addition of 0.5% HCOOH, two QuEChERS salts and one dSPE clean-up ( Fig. 1 ) They were applied to roots and leaves selecting the best approach for each matrix. All tests were performed in triplicate (n = 3). The original method by Anastassiades et.al., was modified as follows: First, 1g of previously lyophilized and ground samples, were placed into a 50 mL polypropylene centrifuge tube. The sample was hydrated adding 9 mL of water and vortexed for 2 min. After 1 h, 50 μL standard mixture (2 μg mL −1 in MeOH) were added to achieve a final concentration of 10 ng g −1 .
Then, sample was vortexed for 5 min and then allowed to stand for 1 h. Next, 10 mL of extraction solvent were added and the sample was vortexed for 2 min. QuEChERS salt was added, and the tube was hand-shaken in order to avoid the formation of MgSO 4 agglomerates followed by another vortex. The sample was centrifuged for 10 min at 40 0 0 rpm at 4 ˚C, the organic phase (top layer) was transferred to a glass tube and kept overnight at -20 °C to induce the precipitation fatty acids, proteins, chlorophyll and sugars, which would have interfered on the analysis [17 , 18] . Six milliliters of the organic phase were carefully aspirated and transferred directly into the PSA-containing tube (900 mg MgSO 4 , 150 mg PSA, 150 mg C18), immediately hand-shaken for 30 s and vortexed for 2 min. Next, the suspension was centrifuged for 5 min at 40 0 0 rpm at 4 °C. Finally, 1 mL supernatant was transferred into a HPLC glass vial. Then, sample was evaporated until dryness under a gentle N 2 stream and reconstituted in 1 mL of 10% MeOH prior to LC-MS/MS analysis. In method 1, 3, 5 and 7, after keeping extracts at -20 °C overnight, 1 mL of the organic layer (supernatant) were analyzed without the clean-up step. In the selected methods, method 3 and 5, no clean-up step was considered as greater recoveries were achieved.

LC-MS/MS acquisition and data analysis
Analysis of final extracts was performed on SCIEX ExionLC TM AD chromatograph coupled with a SCIEX X500R QTOF (Sciex, Redwood city, CA, U.S.) with Turbo V TM electrospray Ionization (ESI). The ion mode was selected based on highest sensitivity of the molecular ion. The injection volume was 10 μL with an auto-sampler temperature set to 8 °C. The chromatographic run time was 13 min and chromatographic separation was achieved on a Hibar R HR Purospher R STAR RP-C18 column (100 mm x 2.1 mm i.d., 2 μm particle size, Merck, Darmstadt, Germany), at 40 °C. Mobile phases + ESI The settings of the ion source were as follows: ion spray voltage was set to 5500 and -4500 V for ( + ESI) and (-ESI), respectively; source temperature and nitrogen gas flows were set to 550 °C and 55 psi, respectively, curtain gas 30 psi, and collision gas (CAD) 7. For MRM HR the precursor ion selection consisted of one TOF-MS survey (100-950 Da for 120 ms of Accumulation time (AT); Declustering Potential (DP) and Collision Energy (CE) were set to 80 and 10 V and -80 and -10 V, for ( + )ESI and (-)ESI, respectively. The Guided MRM HR tool from SCIEX was used for optimizing the high-resolution transitions (see Table 1 ).
To maintain the mass accuracy of the MS detection, the instrument was automatically recalibrated during batch acquisitions by infusing reserpine reference standard (C 33 H 40 N 2 O 9 , m/z 609.2807) in ( + ESI) or a cluster of sodium trifluoroacetate (detection of the cluster (TFA-Na) 5 + TFA − at m/z 792.8596) for (-ESI). The instrument provided a resolving power at Full Width Half Maximum of between 31,0 0 0 and 44,0 0 0 at m/z 132.9049 and 829.5395 with a mass error of 0.4 ppm.

Method validation
The protocols selected for validation were method 3 for roots and method 5 for leaves. It included the determination of accuracy, intra-day precision linearity, matrix effect (ME) and limits of detection (LOD) and quantification (LOQ). Accuracy was assessed in recovery studies for each compound at five spiking levels (n = 3) (2, 5, 10, 50 and 200 ng g −1 ). Recoveries were calculated as the ratio between the peak area in the extract from spiked radish root (or leaves) and the peak area in a blank radish root (or leaves) extract spiked at the same concentration levels. Blanks consisting of the radish tissues used in the validation study fortified only with the IS mixture were also analyzed to confirm the absence of the target analytes in this matrix. Intra-day precision or repeatability, was evaluated based on the relative standard deviation (RSD %) from the recovery data (n = 5). Good precision values were considered less than 20 % deviation for most of the compounds. Otherwise, LOD (the lowest concentration that could be distinguished of the matrix signal with a S/N greater than 3) and LOQ (the lowest concentration with a signal response that could be quantified with a S/N greater than 10 and an RSD ≤ 20%) were determined using the matrix matched calibration curves by linear regression [19 , 20] . To determine ME, blank matrixes (n = 3) were extracted following the selected protocol for both matrices and then spiked at the same concentration levels from the recovery studies. ME was determined comparing the peak areas from the spiked sample with peak areas from the standards in solvent (H 2 O/MeOH, 90:10, v/v) at the same concentration levels [21] . High influence of the matrix compounds in the analysis of an analyte was considered when ME was higher than |40% | [22] . Radish root and leaves are complex matrices due to the high effect of its compounds, resulting in high marked values in the ME. This phenomenon has been studied largely and it is well known that working with ESI the reduction of the signal is related with the ionization of the sample when it is transformed from liquid to gas and depending on the polarity of each compound the effect is different in every case [12 , 23 , 24] . The use of isotopically labelled Internal Standards (I.S.) serves for compensate the matrix effect (signal enhancement/suppression) but also improves accuracy and precision [25] . A matrix-matched calibration curve (CC) was elaborated using blank radishes and spiking them using at least eight different curve points, ranging from 0.05 to 300 ng mL −1 in dry weight radish. Linearity was accepted when coefficient correlation (r 2 ) was ≥ 0.99. Table 2 shows the values obtained in the middle validation point (10 ng g −1 ), whereas Tables S1 and S2 contain the results of the other validation points. Globally, analytes were recovered satisfactory with a range between 70 and 120 % in most of the cases. However, overall recoveries of 5-methyl-2H-benzotriazole and carbamazepine-10,11-epoxide (53-94 and 25-82 %, respectively) were fairly low in roots at the 5 validation points. Otherwise, in leaves ciprofloxacin was the compound showing the lowest recoveries (13,14 %). The values of the ME differ between compounds but the CECs showing high enhancement in root were acetaminophen and 4-hydroxydiclofenac (421 and 451 %, respectively) whereas biggest signal suppression were observed in fenofibrate and gemfibrozil ((-93) and (-83), respectively). On the other hand, the highest signal enhancement in radish leaves were found in diltiazem (246 %) and There are only few previous studies that used QuEChERS for the extraction of CECs in radish tissues ( Table 3 ) [8 , 9 , 26 , 27] . To our knowledge there are no studies using HRMS for a such matrix. We only found two studies for detection/identification of metabolites using cell cultures or under hydroponic conditions using a QToF and Ion mobility-QToF, respectively [28 , 29] . The results of our   That allows skipping the clean-up step, therefore cheapening the total cost of the method, reducing the sample treatment time and avoiding potential analytes loses due to the use of extra salts. Finally, the acquisition took place using the QToF-MS instrument in MRMHR mode. This acquisition mode shows greater selectivity than LR-MS instruments due to the use of high resolution, and a sensitivity comparable to QQQ and QTrap instruments, resulting in similar LODs and LOQs [30] . The key point here is the development of a specific approach for radish roots and another for radish leaves which improves terms in accuracy, precision and ME.

Method applicability
To test the applicability of the method, radish plants were growth in controlled conditions according to the same procedure reported elsewhere and watered using artificial contaminated water at 10 ng g −1 . Briefly, Radish seeds were sown in pots (4 seeds per pot, n = 6 pots) and after 5 days from germination, the seedling were regularly irrigated every two days for 20 days with 200 mL of artificial contaminated water at 10 ng g −1 . Control samples were irrigated only with tap water. The use of fertilizers was needed to ensure a correct crop growth. After 21 days, crops were harvested and hand-washed to remove any soil particles. Then, roots were separated from leaves and individually frozen at -20 °C for 48 h, freeze-dried, and prepared according the validation procedure. Among the validated compounds, ten CECs have been satisfactorily quantified in both radish roots and leaves. Results and frequency of detection (FoD) are reported in Table 4 . Four compounds were detected in both, roots and leaves (Climbazole, metoprolol, propranolol, verapamil). While furosemide, gemfibrozil and irbesartan were detected only in the roots, the leaves were merely positive for bisphenol A, ibuprofen and ketoprofen. The compound with the highest accumulation in root was furosemide (6 ng g −1 ), while verapamil in leaves (6 ng g −1 ). Moreover, propranolol (1.1 ng g −1 ) and bisphenol A (1.9 ng g −1 ) are the compounds showing the lowest concentration in roots and leaves, respectively.

Declaration of Competing Interest
The Authors confirm that there are no conflicts of interest.

Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi: 10. 1016/j.mex.2021.101308 .