Determination and dietary risk assessment of 284 pesticide residues in local fruit cultivars in Shanghai, China

The presence of pesticide residues has become one of the main risk factors affecting the safety and quality of agro-food. In this study, a multi-residue method for the analysis of 284 pesticides in five local fruit cultivars in Shanghai was developed based on ultrahigh-performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-QTOF/MS). The limits of determination and the limits of quantitation of pesticides were 0.6–10 and 2–30 μg/kg, respectively. A total of 44, 10, 10, 18, and 7 pesticides were detected in strawberries, watermelons, melons, peaches, and grapes, respectively. The pesticide levels in 95.0% of the samples were below the maximum residual limits (MRLs) prescribed by China, and in 66.2% of the samples below the EU MRLs. The dietary risk assessment study showed big differences in the chronic and acute exposure risk values among different Chinese consumer groups. Through fruit consumption, children/females showed higher exposure risks than adults/males. But both the risk values were less than 100%, indicating that potential dietary risk induced by the pesticides was not significant for Chinese consumers. Nevertheless, certain measures are needed for both growers and the government in order to decrease the MRL-exceeding rate of pesticide residues and ensure the quality and safety of fruits for consumers.

Pesticide residues in fruit samples. The developed method was applied to the analysis of 284 pesticides in 260 fruit samples. Matrix-matched calibration curves were used to calculate the concentrations of pesticides in fruits. Pesticide residues were detected in 228 samples (87.7% of the total), mainly fungicides, insecticides and acaricides. The detection rates of pesticide residues in strawberry, watermelon, melon, peach, and grape samples were 93.7%, 82%, 88%, 70%, and 100%, respectively, and more than 56% of the samples contained at least two of the analyzed pesticide residues (Fig. 1). Detailed data of the pesticide residues detected in the fruit samples are shown in Table 1. A total of 44, 10, 10, 18, and 7 pesticides were detected in strawberry, watermelon, melon, peach, and grape samples, respectively ( Table 1). The pesticides with the highest detection frequency were fluopyram  Table 1). The levels of fluopyram residues in 13 strawberry samples (16.3%) were exceeded the Chinese MRL, and in 4 strawberry (5.0%) and 3 melon samples (6.0%) were exceeded the EU MRL (Table 1). In addition, the levels of carbendazim, ethirimol, flonicamid, hexaconazole, prochloraz, propamocarb, thiophanate-methyl, and triadimefon residues in at least one strawberry sample; methoxyfenozide and oxadixyl in at least three watermelon samples; fosthiazate and paclobutrazol in at least one melon sample; acetamiprid, carbendazim, and paclobutrazol in at least two peach samples; and ethirimol, picoxystrobin, and triflumuron in at least two grape samples were exceeded the EU MRLs or exceeded the lower limit of analytical determination prescribed by the EU (Table 1).  Tables 2 and 3.

Discussion
The linearity (> 0.990) was considered acceptable 40 . Most of the LOQ values (82.4%) were below the non-detectable default value (0.01 mg/kg) recommended in the EU regulations 16 . The obtained LOQs were much lower than the Chinese and EU MRLs (0.01-10 mg/kg) for pesticides in strawberries, watermelons, melons, peaches, and grapes 18,22 , indicating that the developed method is sensitive and suitable for comprehensive monitoring of pesticide residues in the fruit samples. The SANTE guidelines recommend that the acceptable mean recoveries are those within the range 70-120%, with an associated RSD ≤ 20% 41 . The accuracy and precision obtained in this study are comparable with those reported in previous studies. Yang et al. 32 determined 50 pesticides in starfruit and Indian jujube using LC-QTOF/MS, and obtained recoveries between 63 and 119%, with RSDs of 0.2-3.2%. Sivaperumal et al. 31 achieved satisfactory recoveries ranging from 74 to 111%, with RSDs below 13.2%. Matrix effect is the combined effect of all components of the sample other than the analytes on the measurement, which can comprehensively suppress or enhance the response of the target compounds 28,33 . The values of matrix effect between -20% and 20% are considered acceptable 33 . Matrix ionization suppression still existed in pesticide analysis, which is in accordance with the results of earlier studies 26,28,32 .
Multiple pesticide residues in fruits are commonly observed. Li et al. 42 found that carbendazim, cyhalothrin, acetamiprid, cypermethrin, imidacloprid, as well as difenoconazole had high detection frequency in peaches. Previous studies have also noted that pesticides, especially fungicides and insecticides such as carbendazim, pyrimethanil, trifloxystrobin, and acetamiprid, had high detection frequency in strawberry fruits in China, Poland, and Turkey 1,30,43 . In this study, the strawberry samples were collected in January when strawberries first appeared on market. In order to increase strawberry yield and maximize returns, growers apply high levels of various pesticides during this period. Chu et al. 30 also noted that the detection rates of pesticides in strawberries collected in January were higher than that in strawberries collected in other months.  www.nature.com/scientificreports/ MRLs are the maximum permissible values of pesticide residues in food. They are established to ensure the proper use of pesticides in agriculture and reduce harmful pesticide intake in humans, and thus protect human health 17 . The overall result revealed that 95.0% of samples were below the MRLs prescribed by the National Standard of China 18 , and 66.2% of samples were below the MRLs prescribed by the EU 16 . Currently, there are many pesticides which are registered for use on fruits in China (China Pesticide Information Network. http:// www. icama. org. cn/ hysj/ index. jhtml). In this study, only 11, 3, 1, 5, and 4 detected pesticides are registered pesticides in strawberries, watermelons, melons, peaches, and grapes, respectively. Moreover, some pesticides had no corresponding residue limits authorized in Chinese regulations but had high detection rates, such as ethirimol in strawberries, watermelons, melons, and grapes, fosthiazate and fluopyram in melons, and triflumuron in grapes. This is a great challenge for the government to monitor the use of pesticides.
The results of chronic and acute dietary exposure assessment are in agreement with the results of Chu et al. 30 . Similar results from studies on chronic dietary risk of pesticides in fruits in Poland and in peaches in China have also been reported 1,42 . They also indicated that the chronic risk values for children were higher than that for adults, but neither exceeded 100%. Through fruit consumption, children had higher chronic and acute exposure risks than adults, and females had higher exposure risks than males, which supports previous findings 30,44,45 . In our study, the evaluated fruits exhibited an acceptably low risk to Chinese consumers, however, other studies indicated that some kind of pesticide residues in fruits showed unacceptable acute risks, especially for infants and children 42,44,45 . Although pesticides with high detection rates do not mean high exposure risks, the potential risks should be paid more attention to the pesticides with high levels (> MRLs) and the pesticides with no In order to decrease the MRL-exceeding rate of pesticide residues in fruits, certain measures, such as increased education of growers, control of the sale and use of pesticides, rigorous monitoring of pesticides before harvest, implementation of integrated pest management methods, as well as improvement of regulations, are urgently needed 17,43 . Both growers and government are responsible for food safety, the applications and monitoring programs for pesticides in domestic products must be responsibly carried out.

Materials and methods
Chemicals and reagents. High-performance liquid chromatography (HPLC) grade methanol and acetonitrile were obtained from Merck (Darmstadt, Germany), ammonium acetate (CH 3 COONH 4 ) was purchased from ANPEL Laboratory Technologies Inc. (Shanghai, China). Analytical grade sodium chloride (NaCl) was purchased from Shanghai Titan Scientific Co., Ltd. (Shanghai, China). Ultrapure water was prepared using a Milli-Q purification system (Millipore, Billerica, MA, USA).
Twenty-one groups of certified pesticide mixed standard solutions (100 µg/mL in HPLC-grade acetonitrile, methanol, toluene, hexane, and acetone) were purchased from Alta Scientific Co., Ltd (Tianjin, China). These stock solutions were stored at -20 ℃ in the dark. From each stock solution, a mixed standard working solution contained all the pesticides was prepared at 1 µg/mL by appropriate dilution with acetone, which was stored at 4 ℃ and renewed every 2 months.  (Table S3) in 2019 and 2020. All the fruit samples were collected with the permission of the farmers. The wet weight for one sample was 3 kg for strawberry, melon, peach, grape, and tomato, and each watermelon sample include three watermelons. All samples were kept cool and the whole fruits of strawberries, watermelons, melons, grapes, and peached (without peach pit) were homogenized in a commercial blender, and then stored at -20 ℃ in the dark until chemical analysis. For peaches, the weight of the peach pit was included when calculating the pesticide residues.   www.nature.com/scientificreports/ Pesticides in fruit samples were extracted using a modified method of Yang et al. 32 . A blended sample (around 10 g) was weighed into a 50 mL polypropylene centrifuge tube, and mixed thoroughly with 10 mL acetonitrile for 20 min using an advanced multi-tube vortexer (Troemner LLC., Thorofare, NJ, USA). Subsequently, in order to improve the extraction efficiency, 5 g of NaCl was added to the tube and vortexed for 1 min. The mixture was centrifuged at 5000 rpm for 5 min in a Thermo Fisher ST 16R centrifuge (Osterode, Germany). After centrifugation, 1 mL of the supernatant was transferred to a 10 mL glass tube, 1 mL of Milli-Q water was added, and the tube was vortexed for 30 s. The extract was then filtered through a 0.22 µm nylon syringe filter (Pall Corp., Port Washington, NY, USA) and 3 μL were injected into the UPLC-QTOF/MS system. Strawberry, watermelon, melon, peach, and grape fruit samples without pesticide residues detected were used as blank samples. Pesticide-spiked blank samples were used as quality control samples. The blank and quality control samples were extracted simultaneously with the fruit samples by the same method. Pesticide residues were quantified by the external standard calibration curve method. The samples were diluted when the pesticide content exceeded the standard calibration curve range. The limit of detection (LOD) and the limit of quantification (LOQ) of the method were defined as 3 and 10 times of spiked blank samples' signal-to-background noise (S/N), respectively. Samples were run in the following order: solvent (acetonitrile/water 1/1, v/v)-calibration curves-blank samples-quality control samples-solvent-fruit samples-calibration curves-solvent. The blank and quality control samples were inserted every 20 fruit samples. UPLC-QTOF/MS analysis. The pesticides were identified and quantified using an ultrahigh-performance liquid chromatography system (Waters Acquity I-Class, Waters Corporation, Milford, MA, USA) coupled to a  The QTOF/MS spectra were acquired in positive electrospray ionization mode (ESI + ) with the following parameters: mass range, 50-1000 m/z; ionspray voltage floating (ISVF), 5500 V; temperature (TEM), 500 ℃; ion source gas (GS1) nebulizer gas pressure, 50 psi; ion source gas (GS2) auxiliary heater gas pressure, 50 psi; curtain gas (CUR), 35 psi; declustering potential (DP), 80 V; collision energy (CE), 35 ± 15 eV. The mass spectrometry analysis was conducted in full scan TOF/MS mode and in MS/MS mode. Detailed instrument conditions are described in Yang et al. 32 .
Method validation. The fruit samples were analyzed by UPLC-QTOF/MS in advance, and the sample detected as pesticide-free was used as the blank matrix sample for the spiking experiment. The validation parameters included linearity, sensitivity, accuracy, precision, specificity, and matrix effect. The linearity was determined using matrix-matched calibration curves, which were obtained by adding mixed pesticide standard solution into the extract of blank matrix at seven concentration levels in the range of 2-200 µg/kg, analyzed in triplicate. The sensitivity was assessed by LODs and LOQs. Method accuracy was evaluated by recovery studies. The blank matrix sample was spiked at two concentration levels (10 and 100 µg/kg) with six replicates for each level, then the spiked samples were extracted according to the procedure as described in** Section Sample preparation. The relative standard deviation (RSD) of the pesticides from the recovery studies were used to evaluate the precision. To assess the specificity, the chromatograms of blank sample and spiked sample at LOD levels were analyzed. The S/N ratios of chromatographic peaks in blank sample had to be lower than that in spiked sample 40 . The matrix effect (ME) was evaluated by comparing the signal intensity of matrix-matched standard with pure solvent standard at the same concentration 28,33 . ME (%) was calculated based on the following equation 28 : ME (%) = (peak area of matrix-matched standard − peak area of solvent standard)/ peak area of solvent standard × 100%.
Dietary exposure risk assessment. The chronic and acute dietary exposure risk values were determined by comparing the value of national estimated daily intake (NEDI) of pesticides with acceptable daily intake (ADI), and by comparing the value of estimated short-term intake (ESTI) of pesticides with acute reference dose (ARfD), respectively, according to the following Eqs. 30,46,47 .
The chronic risk was calculated using the above Eqs. (1) and (2). NEDI (mg/kg·d) indicates the national estimated daily intake; R (mg/kg) is the mean amount of pesticide residues in fruit samples; F (kg/d) is the dietary consumption of fruits in China; bw (kg) is the average body weight; ADI (mg/kg·d) is the acceptable daily intake.
The acute risk was calculated using the above Eqs. (3) and (4). ESTI (mg/kg·d) represents the estimated short-term intake; HR (mg/kg) is the highest amount of pesticide residues in fruit samples; LP (kg/d) is the large portion of fruit consumption in Chinese population; ARfD (mg/kg·d) is the acute reference dose.
In this study, fruit consumption group was divided into three sensitive population groups, including 2-4, 18-30, and 60-70 year old male and female groups. The average body weight and fruit consumption in different groups in China are shown in Table S4. If %ADI or %ARfD value is lower than 100%, the exposure risk is acceptable. The higher the value, the greater the risk. While when the value is higher than 100%, it indicates an unacceptably high risk to consumersy 30,46 .

Approvals and permissions. This study was approved by Shanghai Municipal Agriculture and Rural
Affairs Committee (Approval number: 2019-02-08-00-12-F01144). The experiment was performed in accordance with the regulations (NY/T 789-2004) established by the Ministry of Agriculture and Rural Affairs of the People's Republic of China. All the farms or farmer professional cooperatives are legally registered in Shanghai.

Data availability
All data generated and/or analyzed during the current study are available from the corresponding author on reasonable request.