Dissipation, persistence, and risk assessment of fluxapyroxad and penthiopyrad residues in perilla leaf (Perilla frutescens var. japonica Hara)

The objective of this study was to determine the residual characteristics and to calculate the persistence of the fungicides fluxapyroxad (15.3% suspension concentrate) and penthiopyrad (20% emulsifiable concentrate) on the leaves of greenhouse-cultivated perilla (Perilla frutescens var. japonica Hara). Fluxapyroxad was diluted 2,000-fold and penthiopyrad was diluted 4,000-fold. Each solution was sprayed 3 times onto crops at 7-d intervals before harvest. Leaf samples were collected at 3 h (0 d), 1, 3, 5 and 7 d after the third and final treatment. The recovery ranges of fluxapyroxad and penthiopyrad and their metabolites were 74.2%–104.1%. Pesticide residue analyses indicated that fluxapyroxad and penthiopyrad residues in perilla leaves dissipated over time. The persistence of fluxapyroxad and penthiopyrad residues 7 d after the final spray were 50.0% ± 4.9% and 44.2% ± 2.8% of those measured 3 h (0 d) after the final spray, respectively. The percent acceptable daily intake (%ADI)—which was assessed according to the daily food intake by Koreans according to age—was < 7.3%. Therefore, it was determined that the health risk was low. The perception that residual pesticides are present in large amounts in perilla leaf has led to consumer concern. However, in this study, the amounts of pesticide in perilla leaf decreased over time. Although it has been hypothesized that the risk of pesticide intake would be higher in younger children, the results actually suggest the opposite. Therefore, the pesticides in question are considered to be safe for use on perilla leaves.


Introduction
Pesticides help to improve and maintain crop quality, and are widely used globally [1] to protect crops against harmful biotic factors such as insect predators, pathogens, and weeds [2,3]. Sprayed pesticides dissipate and are degraded by enzymes, hydrolysis, and exposure to light. However, pesticide residues may persist in and on the surfaces of crops, and can therefore be

Storage stability
The Health & Safety Executive (HSE) [21] reported that a storage stability test should be performed to ensure that the pesticide in the sample does not substantially degrade upon exposure to environmental factors such as temperature, humidity, and light. Thus, the stability of the pesticides in the samples during freezer storage was evaluated. One milliliter of a 2.5 mg L −1 fluxapyroxad standard was mixed with 5 g of untreated perilla leaf, resulting in a final concentration of 0.5 mg kg −1 . Three replicates of 10 g of untreated perilla leaf were fortified with 1 mL of 12.5 mg L −1 M700F002 and 1 mL of 12.5 mg L −1 M700F048, bringing the final concentrations to 2.5 mg kg −1 . For penthiopyrad and its metabolites, each of the three replicated samples of untreated perilla leaf (10 g) was fortified with 1 mL of 5 mg L −1 of the test substances. All samples were stored at −20˚C until analysis. Stabilities were validated by calculating the percent recovery.

Sample preparation
Fluxapyroxad and its metabolites. Five grams of perilla leaf was weighed in a tall beaker (300 mL) and 50 mL acetone was added. The mixture was homogenized in a blender at 10,000 rpm for 3 min. The homogenate was then vacuum-filtered using filter paper (Whatman No. 2 (8 μm), USA) and Celite 545 (Merck, USA). The flask and filter cake were rinsed with 50 mL acetone and the rinsate was combined with the filtrate.
The filtrate was quantitatively transferred to a 1 L separatory funnel, to which 400 mL of distilled water, 100 mL of saturated NaCl solution, and 50 mL of dichloromethane were added in sequence. For metabolite determination, 50 mL of ethyl acetate was substituted for the dichloromethane. The mixture was vigorously agitated on a shaker at 250 rpm for 10 min and allowed to stand until two distinct layers formed. The organic solvent layer was filtered through 20 g of anhydrous Na 2 SO 4 and collected in a 250 mL distillation flask. The partition process was repeated using another 50 mL of dichloromethane (or 50 mL of ethyl acetate in the case of the metabolites), and the organic layers were pooled in the distillation flask. The organic solvent extract was evaporated to dryness at 35˚C on a rotary vacuum evaporator (Buchi, R-205V, Gemany). The residues were redissolved in 5 mL of acetonitrile, filtered with a 0.2 μm syringe filter, and analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) ( Table 1).
Penthiopyrad and its metabolites. Ten grams of perilla leaf was weighed in a tall beaker (300 mL) and 50 mL of acetone was added. The mixture was homogenized in a blender at 10,000 rpm for 2 min. The homogenate was then vacuum-filtered using filter paper (8 μm) and Celite 545. The flask and filter cake were rinsed with 50 mL of acetone and the rinsate was combined with the filtrate.
The filtrate was quantitatively transferred to a 1 L separatory funnel, to which 400 mL of distilled water, 100 mL of saturated NaCl solution, and 50 mL of ethyl acetate were added in sequence. The mixture was vigorously agitated on a funnel shaker at 250 rpm for 5 min and allowed to stand until two distinct layers formed. The organic solvent layer was filtered through 20 g of anhydrous Na 2 SO 4 and collected in a 250 mL distillation flask. The partition process was repeated using another 50 mL of ethyl acetate and the organic layers were pooled in the distillation flask. The organic solvent extract was evaporated to dryness at 35˚C on a rotary vacuum evaporator. The residue was re-dissolved in 5 mL of dichloromethane and subjected to Florisil column chromatography for sample purification.
A chromatographic glass column (35 cm L × 1.1 cm I.D.) was dry-packed with 5 g of activated Florisil at 130˚C over a 5 h period and topped with approximately 2 cm (~2 g) of anhydrous Na 2 SO 4 . Fifty milliliters of dichloromethane were added to the column. When the solvent level reached the top of the anhydrous Na 2 SO 4 , 5 mL of redissolved sample was applied to the column followed by another 5 mL of dichloromethane to wash the sample flask. The column was washed with 50 mL of dichloromethane and the solvent discarded. PAM was eluted with a 50 mL mixture of dichloromethane:acetone (90:10, v/v). Penthiopyrad and 753-A-OH were continually eluted with a 50 mL mixture of dichloromethane:acetone (50:50, v/v). The fractions were collected and concentrated to dryness at 35˚C on a vacuum rotary evaporator. The residue was redissolved in 2 mL of acetone for determination using a gas chromatography-nitrogen phosphorus detector (GC-NPD) ( Table 2).

Total amount of pesticide
The total amount of pesticide was the sum of the concentrations of the parent compound and the metabolites divided by the ratio of the molecular weight of the parent compound and the metabolites (Eqs 1 and 2) (FAO, 2016).

Risk assessment
The risks to human health (percent acceptable daily intake; %ADI) of the test pesticides ingested with perilla leaf were calculated by the ratio of estimated daily intake (EDI) to the acceptable daily intake (ADI). The high residue (HR) and daily food intake of perilla leaf by Koreans according to age were multiplied to obtain the estimated daily intakes (EDIs). The food daily intake of perilla leaf by age for Koreans was derived from the Korea Health Industry Development Institute [22]. The average body weights of Koreans by age were acquired from the Korea Centers for Disease Control and Prevention [23]. The EDIs were calculated using average body weights by age. Eqs 3, 4 and 5 (below) were used for the risk assessment [24]. The ADIs were determined using data from the Pesticide and Veterinary Drugs Information database (http://www.foodsafetykorea.go.kr/residue/main.do) provided by the Korean MFDS. In addition, TMDIs (Theoretical Maximum Daily Intakes) of test pesticides for other crops were calculated using Eq (6) and maximum residue limits (MRLs) were used from the Korean ADI ðmg kg person À 1 Þ ¼ ADI ðmg kg À 1 � body weight day À 1 Þ of test pesticide � average body weight by ageð3Þ EDI ðmg kg person À 1 Þ ¼ amount of test pesticide ðmg kg À 1 Þ � food daily intake ðgÞ ð4Þ TMDIð%Þ ¼ P %ADI of registered all crops ð6Þ  [25] indicated that the recoveries of fluxapyroxad and penthiopyrad from apple, strawberry, tomato, and spinach ranged from 70% to 120%. Five pyrazole pesticides were analyzed in rice, wheat, cucumber, tomato, lettuce, apple, and grape by Dong et al. [26] using the "Quick-Easy-Cheap-Effective-Rugged-Safe" (QuE-ChERS) method and ultra-high-performance liquid chromatography/tandem mass spectrometry (UHPLC-MS/MS). The recovery rates for the five pyrazole fungicides were 75.9%-108.0%. This experiment demonstrates the suitability of the pesticide residue analytical techniques used in the present study.

Storage stability
The storage stability tests for fluxapyroxad, penthiopyrad, and their metabolites in perilla leaf generated recovery rates of 80.5%-98.8% and 77.8%-101.5%, respectively (Table 4). Therefore, the pesticides tested in the present study and their metabolites were stable during storage, as the recovery rates ranged from 70% to 120% [20].

Residual pesticide
The average concentration of fluxapyroxad in perilla leaf on the final day of spraying (3 h) was 20.88 ± 1.58 mg kg −1 (Table 5). Seven days after the final spray, the average fluxapyroxad concentration in perilla leaf was 10.45 ± 1.11 mg kg −1 . Therefore, the fluxapyroxad content in perilla leaf decreased over time. The amount of fluxapyroxad metabolites were less than the LOQ. The JMPR (2013) reported that while the major fluxapyroxad metabolites were M700F002 and M700F048, they were detected at levels below the LOQ. The average penthiopyrad concentration in perilla leaf on the final day of spraying (day 0) was 11.19 ± 0.67 mg  (Table 6). Seven days after the final spray, the average penthiopyrad concentration in perilla leaf was 4.79 ± 0.12 mg kg −1 . There was almost no change in the concentration of 753-A-OH in perilla leaf from days 0 to 7 (Table 7). However, JMPR (2013) reported that the residual concentration of PAM in cabbage sprayed with penthiopyrad steadily increased from day 0 (0.15 ± 0.01 mg kg −1 ) to day 7 (0.30 ± 0.02 mg kg −1 ). According to FAO (2016), the total penthiopyrad concentrations in perilla leaf were 11.53 ± 0.70 mg kg −1 on day 0 and 5.08 ± 0.10 mg kg −1 on day 7. Therefore, in the present study, as well as in previous studies, the residue levels of penthiopyrad decreased over time. The total pesticide concentrations were used to calculate persistence and %ADI. No MRLs have yet been established for fluxapyroxad or penthiopyrad in perilla leaf.

Pesticide residue persistence
Pesticide residue persistence in perilla leaf was determined by collecting and analyzing treated leaf samples 1, 3, 5, and 7 d after the final spray. Fluxapyroxad persistence was calculated according to its concentrations in perilla leaf from day 0 ( Table 7). Relative to day 0 levels, fluxapyroxad persistence levels on days 1 and 7 were 90.1% ± 4.6% and 50.0% ± 4.9%, respectively. Penthiopyrad persistence was calculated using the same method used for fluxapyroxad and was found to be 83.3% ± 7.9% at day 1 and 44.2% ± 2.8% by day 7. Since both test pesticides are pyrazoles, they exhibited similar persistence patterns. Their residues persisted for longer periods of time in perilla leaf than they did in crops like cucumber and squash, whose fruit rapidly increases in weight as they grow [27].

Risk assessment
The human health risks associated with the ingestion of the test pesticides via the crops treated with them were assessed by comparing EDIs and ADIs at an HR level. The %ADIs of fluxapyroxad and penthiopyrad in perilla leaf in Koreans by age are presented in Table 8. The %ADIs tended to increase with age, where the rates of increase in daily food intake are greater than those associated with body weight. From ages 12 to 18 y, the %ADIs of fluxapyroxad and penthiopyrad were 3.7% and 0.5%, respectively. From ages 19 to 29 y, they were 6.4% and 0.9%, respectively. In the latter case, body weight did not significantly increase, while daily food intake doubled relative to the former. The %ADI of the 50-64 y age group was the highest of all measured (7.2% for fluxapyroxad and 1.0% for penthiopyrad, respectively). It is the relative difference in ADI rather than pesticide concentration, body weight, or daily food intake that actually influences %ADI calculations. If the ADI is high, then the %ADI will be low irrespective of pesticide concentration or daily food intake. Additionally, test pesticides without MRLs could not be compared in terms of residual amounts. However, when compared with ADI, the health risk was found to be low. In addition, TMDIs have been calculated to assess safety for fluxapyroxad and penthiopyrad in registered all crops in Korea (Tables 9 and 10). The TMDIs of fluxapyroxad and penthiopyrad were 35.64 and 9.52%, respectively. Therefore, the risk to human health for the test pesticides in registered crops in Korea is considered to be low. Also, considering that the Korean government ensures that TMDIs do not exceed 80% of the ADI for a given pesticide, the test pesticides are considered safe.

Discussion
Pesticide concentrations in perilla leaves were likely affected by various factors, including the physicochemical properties of the pesticide [28], crop morphology [8], and environmental factors such as weather and cultivation conditions [29]. The pesticide concentrations in perilla leaf were high because they are absorbed systemically and translocated in the plant [18,19]. Paterson et al. [30] reported that pesticide translocation and systemic properties significantly influence crop residue levels, while Park et al. [31] stated that systemic pesticides are more persistent than non-systemic ones. In another study, the systemic insecticide imidacloprid was sprayed onto eggplant after bloom and before fruit set. The plants were then analyzed for imidacloprid residues up until harvest, and imidacloprid residues were detected in the fruit following the foliar application of the pesticide because it was translocated from the leaves [32].
Crop morphology also determines residual pesticide concentrations. Residue levels tend to be relatively higher in crops with hairy shoot surfaces [33]. As foliar sprays deposit high concentrations of pesticide on hairy leafy crops like perilla, residual pesticide concentrations in perilla leaf were higher than those in crops with relatively smooth leaves. Moreover, perilla has a comparatively large leaf surface area [8].
Pesticide residue concentrations in fruiting vegetables such as cucumber and squash quickly decrease over time because the pesticides are diluted as the crops rapidly grow [34]. In contrast, perilla leaf grows slowly and does not significantly increase in weight with time. Therefore, there was no dilution effect of the pesticide in perilla leaves and its residue levels remained relatively high. However, it is important to note that the perilla plants used in the present study were cultivated under controlled greenhouse conditions. Consequently, the dilution and rinsing effects of rainfall did not occur, and comparatively larger amounts of the pesticides could be absorbed and translocated by the leaves [35].
Individuals in the youngest age group would have the lowest body weights. Therefore, it is expected that the youngest individuals would experience the highest relative health risk from ingesting perilla leaf containing residues of the test pesticides. Nevertheless, their daily fresh vegetable intake is comparatively low. As fresh vegetable consumption levels increase with age, so does the risk of dietary pesticide exposure. However, the health risks are by no means equally high in all Koreans. Chun and Kang [35] reported that when %ADI is < 10%, the relative risk is low and no further analysis is required. Even when 10 � %ADI � 30, the pesticide residue concentration poses no significant health risk. In the present study, the maximum % ADI was 7.2% and TMDIs of fluxapyroxad and penthiopyrad were 35.64 and 9.52%, respectively. Therefore, the fluxapyroxad and penthiopyrad residue concentrations in perilla leaves pose no significant health risks to Koreans.

Conclusions
The aim of the present study was to determine the dissipation characteristics of two pyrazole fungicides sprayed at time intervals onto perilla leaves under greenhouse conditions. Another objective was to assess the health risks that ingesting perilla leaves treated with these fungicides