3.1. Method performance
The developed method for the analysis of abamectin and difenoconazole using LC-MS/MS showed good selectivity, once no interfering compounds were observed in the same retention time of analytes. Linearity was evaluated using analytical curves in the matrices (strawberry flower and pollen) with good response in the concentration range of 0.5–1,00 µg L− 1. Peak areas were used as responses, and the method was shown to be linear and determination coefficients (R2) were greater than 0.98 for both pesticides, with deviation for each concentration ≤ 20%. The parameters of the analytical curves, the detection, and the quantification limits for the method and system are shown in Table 1. The matrix effect for abamectin and difenoconazole in the flower samples was less than 100%, indicating that there was a suppression of ionization. (Matuszewski et al. 2003; 2006). As for the pollen samples, both analytes showed values close to 100%, indicating that the response in the solvent and the matrix we're the same, and no effect was observed.
In all cases, the threshold (± 0.1) established by SANTE guidelines (SANTE/11813/ 2017) was achieved and ISO. The confirmation of the analytical parameters was carried out through the acquisition of the MS / MS transitions and a comparison of their intensity proportions, taking into account that the relative proportion between the transitions must be ≤ 30%. The selectivity of the proposed method was tested by the extraction and analysis of pure extracts from strawberry flowers and pollen-free from pesticides, to establish the absence of signs at the time of elution to target the pesticides and thus demonstrating that neither the matrix nor the compounds present in the sample gave false positives.
Table 1
Analytical parameters for LC-MS/MS analysis of abamectin (ABA) and difenoconazole (DIF) in strawberry flower and pollen.
Matrix
|
|
Linear equation
|
(r2)
|
Linearitya
|
ME (%)
|
LC-MS/MSa
|
|
Methodb
|
LOD
|
LOQ
|
|
LOD
|
LOQ
|
Flower
|
ABA
|
y = 73.66x-978.3
|
0.993
|
0.5 to 1000
|
80.48
|
0.15
|
0.50
|
|
0.30
|
1.00
|
DIF
|
y = 5447.6x + 41872
|
0.995
|
0.5 to 1000
|
83.21
|
0.15
|
0.50
|
|
0.30
|
1.00
|
Pollen
|
ABA
|
y = 95.35x + 332.4
|
0.994
|
0.5 to 1000
|
100.0
|
0.15
|
0.50
|
|
0.30
|
1.00
|
DIF
|
y = 6403.4x-147626
|
0.994
|
0.5 to 1000
|
97.80
|
0.15
|
0.50
|
|
0.30
|
1.00
|
ME = matrix effect |
aµg L− 1 |
bng g− 1 |
The optimization of the QuEChERS method (SM1) was done for the three versions of the method: original, acetate, and citrate, and the clean-up optimization was done using different salts, such as magnesium sulfate, PSA, C18, and activated carbon. Of the tested methods, for the flower matrix, the best performance was observed for the original QuEChERS method, which used as extraction salts 0.40 g of MgSO4 + 0.10g NaCl and as clean-up salts 150 mg MgSO4 + 50 mg PSA, in which recoveries were 99.7% for difenoconazole and 61.1% for abamectin. For the pollen matrix, the best performance was also observed for the original QuEChERS method, while for the clean-up, the “original + C18” method was the one that presented the best performance and used 150 mg MgSO4 + 50 mg PSA + 50 mg C18, and recoveries were 109.1% for difenoconazole and 108.1% for abamectin and adequate RSD values were achieved (0.2-7%).
After selecting the method for extracting the analytes, the precision and accuracy for the two matrices were evaluated using three concentration levels: low (5 µg L− 1), medium (100 µg L− 1), and high (750 µg L− 1), and the results obtained are shown in Table 2.
Table 2
Accuracy and precision of the validated method for abamectin and difenoconazole determination in strawberry flower and pollen using the QuEChERS extraction method and LC-MS/MS analysis.
Matrix
|
|
Level
|
Accuracy (%)a
|
Precision (RSD %)
|
Intra-daya
|
Inter-daya
|
Flower
|
Abamectin
|
Low
|
80.84
|
3.78
|
6.19
|
Medium
|
88.15
|
0.91
|
3.14
|
High
|
92.82
|
0.25
|
0.22
|
Difenoconazole
|
Low
|
100.99
|
1.29
|
1.80
|
Medium
|
108.34
|
1.11
|
0.90
|
High
|
103.32
|
1.09
|
2.28
|
Pollen
|
Abamectin
|
Low
|
108.71
|
7.31
|
6.55
|
Medium
|
90.21
|
3.90
|
4.17
|
High
|
90.43
|
1.14
|
1.03
|
Difenoconazole
|
Low
|
96.65
|
0.72
|
1.33
|
Medium
|
97.17
|
0.24
|
1.41
|
High
|
105.23
|
0.17
|
1.11
|
a n = 5 |
Studies have used the modified QuEChERS method for detecting abamectin and difenoconazole in bee pollen samples. Friedle et al (2021) used the modified QuEChERS method to detect more than 260 pesticides in pollen samples and the method showed LOQ of 3 ng g− 1 and recovery of 87% for difenoconazole. The maximum and minimum concentrations of difenoconazole detected in the samples were 48 and 1.5 ng g− 1, respectively. Other studies have determined difenoconazole in bee pollen samples using the modified QuEChERS method and evaluating different clean-up agents; the best method showed recoveries of 96 and 89% for spiking levels of 5 and 50 µg kg− 1 of difenoconazole; accuracy less than 20%; LOQ of 5 µg kg− 1 (Vázquez et al. 2015); Wiest et al (2011) used the citrate QuEChERS method to detect abamectin and other contaminants in pollen. For abamectin, the method presented LOD of 10.2 6 ng g− 1 and LOQ of 30.6 ng g− 1 and recoveries in the range of 81–112%, and abamectin was not detected in real samples. For flowers, there is still very little work on these and other pesticides in the literature, but there are some that corroborate with the extraction technique for identification and quantification in strawberries (Oshita et al. 2014) and validation of pesticides in processed fruit by UHPLC/MS-MS (Valera et al. 2020). Studies involving the determination of abamectin and difenoconazole in bees have been developed using different QuEChERS extraction salts tested in this study. Prado et al (2020), using the acetate QuEChERS method, present abamectin recovery of 89.4% in the high-level spiking (100 ng g− 1) and for difenoconazole 95.5% using spiking levels from (1 to 100 ng g− 1) and LOQ of 0.01 ng g− 1.
Thus, in the present work, using the proposed method for pollen and strawberry flowers, was possible to detect concentrations below the MRL levels, with good linearity, quantitation limits as well as accuracy inside the recommended range of 80–120% and the precision below 20%. The application of this method is discussed below.
3.2. Monitoring
The commercial insecticide with abamectin (a.i) was not detected in any strawberry flower or pollen samples from the hives. The absence of this compound can be focused first, in the lower agronomic dose (Kraft@ 36), which corresponds to 13.5 g ha− 1 in mass, indicated for strawberry fields. Besides this, for this crop, it is allowed only two applications in the period of 14 days; in contrast, the fungicide containing difenoconazole (a.i) mass is 20 g ha− 1 with six possible applications over the same period (Syngenta, Score® CE). This information suggests, preliminarily, that the detection of fungicide is probably higher than for the insecticide considering the dose and application frequency.
However, an important factor to be considered is also the chemical properties and the persistence (1/2 life period) of each investigated pesticide. For difenoconazole, the half-life of the active ingredient in the terrestrial environment is about 85 days, whereas for abamectin is only 1 day. For abamectin, the behavior in plants is related to photolysis with no residues, where the avermectin B1a component can be considered as more representative for environmental monitoring (EFSA 2008). In this sense, abamectin residues in the field can be considered as commonly low (below 0.025 ppm), with no persistence or accumulation observed in the environment (Lasota and Dybas 1990). Considering the bee’s matrices, for pollen and bee samples (A. mellifera), Wiest et al (2011) have no abamectin detection in any sample in a multi-residue method. Besides the environmental matrices tested (pollen and flowers) have no residues of abamectin, for exposed stingless bees, Prado et al. (2020), have been observed the uptake of the commercial product (Kraft) via topic and oral exposure. This uptake can alert for the exposure of the commercial product via spray drift and toxicity related not only to abamectin but also the inactive ingredients of the commercial formulation over the native bees and other non-target insects.
As mentioned above, in opposition to abamectin, difenoconazole was observed in the majority of flowers and pollen samples. The results for difenoconazole in strawberry flowers are depicted in Fig. 2 in a concentration range of < LOQ (1 ng g− 1) to 7.53 ng g− 1 demonstrating the capacity of the strawberry plant to take up this fungicide and accumulate. The sampling campaigns were proceeded during the wet season considering the Brazilian weather, where the fungi proliferation is pronounced. In this case, the application of consecutive treatments of difenoconazole can be evidenced. As observed in this figure, in November 2018, difenoconazole was detected in only three points (P1, P4, and P9), while for other sampling campaigns this compound could be found in more sampling points. In January, difenoconazole was not detected only at P2.
The highest concentration found in the strawberry flowers was 7.53 ng g− 1, which occurred in March at P10 of area 2, almost equal to P4 of area 1, with 7.06 ng g− 1 for the same period. In Area 1, is possible to note that the points located in the superior part of the strawberry field (P1 and P4) have presented higher concentrations of difenoconazole compared with other sampling points. An experimental design of the difenoconazole application in strawberry fields has detected a rapid dissipation of this fungicide after pulverization in leaves, but also an increase of the residual amounts in fruit after consecutive applications (250 g L− 1) with 14 days intervals (Heleno et al. 2014). This systemic effect can explain the general increase of concentrations through the sampling campaigns and mainly observed in P4. In another study, Sun et al. (2015) also have observed that the half-life of difenoconazole increases in consecutive applications from 3.65 (1 application) to 6.30 (2 applications) days in strawberry fields. When compared with other pesticides, in laboratory experiments, using a soil field rate application of difenoconazole in rice plants (Oryza sativa L.), Ge et al. (2017) have been observed that this fungicide has a greater half-life than thiamethoxam and imidacloprid, however with lower bioaccumulation factor (BCF) and translocation factor (TFs) compared with those neonicotinoids pesticides.
Figure 2 Concentration (ng g− 1) of difenoconazole in strawberry flowers sampled in Areas 1 (P1 to P5) and 2 (P6 to P10) in different sampling campaigns: November/18; January/19; and March/19. Missing data are below the limit of quantification (< LOQ).
In Brazil, the maximum residual limit for difenoconazole in strawberry fields is 0.5 µg g− 1, with foliar application with a security interval of 1 day (ANVISA 2021b). Converting the maximum value found in the strawberry flower (P10), we have 0.00706 µg g− 1 of strawberry, which corresponds to an amount 70 times lower than the MRL level allowed for fruits. In this sense, for human health, the concentration levels found in the present study are well below the harmful maximum limits. In other countries, the MRL is also higher than the detected in flowers as in European Union (0.4 µg g− 1) (EFSA 2011).
For bees, the exposure and toxicity of pesticides applied in crops occur during the pollination process, but the magnitude of pesticides' risk must also consider the landscape and the diversity of visited plants (McCart et al. 2017). For strawberry fields, Antunes et al. (2007), has observed from 15.9 to 18.6 visits of T. angustula per flower, per hour, when the field is surrounded by 4 hives, the same number of hives displayed for the present investigation and that can contribute significantly for the results described below.
In a multi-residue method for the monitoring of 81 pesticides in pollen and bees (A. mellifera), Saibt (2017) has detected only difenoconazole (16 ng g− 1) in pollen samples from the Rio Grande do Sul State, Brazil. The exposure to difenoconazole (Score 250 EC 0.2 L ha− 1) in apple (Malus domestica) field has been also detected in bee (A. mellifera) products, where the pollen contamination was about 43 ng g− 1, with the detected concentration related to the capacity of fungicides to be fixed by sugars, amino acids or proteins (Kubik et al. 2000). Other studies involving this fungicide detection in pollen (A. mellifera) has included: Friedle et al (2021) in a concentration range of 0.02 to 48 ng g− 1; and Vásquez et al (2015) with a concentration of 45 ng g− 1, levels below the mostly of detected concentrations observed in the present study.
Considering the hives arranged next to Area 1, pollen samples (Fig. 3) have presented impressing high concentration of difenoconazole, especially the H1 hive in January/19 (456 ng g− 1). The accumulation of this fungicide in the pollen collected by T. angustula can be associated with the visit of these stingless bees in many strawberry fields located in the Bom Repouso region reaching this bee flight range of 500 m (van Nieuwstadt & Iraheta 1996). Flight activity of T. angustula can also be dependent on temperature, where warm weather (above 19.6oC) can allow a major activity and the collection of pollen to the hives (Marlebo-Souza and Halak 2016). This behavior can contribute to the major concentrations detected in January/19 samples which are observed as the greatest strawberry blossom in the field and bee's activity. Besides the increase of activity, the conditions inside the hives as protection against sunlight, temperature control, and anti-bactericide properties can contribute significantly to the accumulation and preservation of this pesticide.
Figure 3 Concentration (ng g− 1) of difenoconazole in pollen from T. angustula hives located nearby strawberry fields: Areas 1 (H1 and H2) and 2 (H3 to H5) in different sampling campaigns: November/18; January/19; and March/19. Missing data are below the limit of quantification (< LOQ).
In addition to strawberry fields, other plants can also be visited by T. angustula on area, where Asteraceae and Fabaceae are indicated in the literature as the favorite plant families for this species (Braga et al. 2012). A preliminary study (not published) in developing in our lab to investigate the diversity of pollen in samples from these hives and had demonstrated that about 90% of identified pollen is from strawberry plants.
3.3. Risk Assessment
Once the pollen is transported to the hive, the bee's contamination path reaches another configuration, changing from contact to oral. In this sense, it is important to highlight that pollen acts as a main food when worker bees are in their first two weeks of life, and as a supplement, after two weeks of their lifetime. So, while the honey supply is responsible for providing energy to bees, pollen is considered an important source of minerals, vitamins, and proteins (Vit, et al. 2004). All these exposures and further toxicity by the food supply can affect the colony's health, bringing sublethal effects and impact the maintenance of the colony through the effect over larvae development (Leite et al. 2018).
In this sense, for the risk assessment, once the exposure to difenoconazole was observed, the program BeeRex was used for estimated if the RQ exceeds the levels of concern (0.4 for acute risk). As mentioned above, for RQ calculations the variables considered were product application rate (Kg a.i. ha− 1), taken into account the maximum recommended dose and the oral LD50 for 48h for T. angustula (µg a.i. bee− 1). At the same time, a calculation considering CAE/LD50 was also investigated.
For this calculation, once there is no T. angustula toxicity data for this compound in literature, we have consulted the literature data for Apis mellifera, where the oral toxicity endpoint (LD50) registered is 177 µg a.i. bee− 1 (EFSA 2011) and 33.48 µg a.i. bee− 1 (SHARDA BRASIL 2019). Considering those two data, the LD50 resulting average was 105.24 µg a.i. bee− 1. However, due to the absence of toxicity data for this native bee, an adjustment of 10 times sensibility was made, considering the LD50 amount observed for Apis, (Arena and Sgolastra 2014), and applied for T. angustula, resulting in a final LD50 of 10.52 µg a.i. bee− 1.
Thus, considering the difenoconazole application rate of 20 g ha− 1 by foliar spray and adding the empirical data of residual amount detected in pollen the risk quotient was estimated, corresponding to Tier 1. In this sense, including this estimated endpoint and the application rate in strawberry fields, the RQ for difenoconazole exposure was estimated at 6.8 (CAE/LD50) and 194.03 (CAE – BeeREX), both surpassing the level of concern (< 0.4). However, when the levels detected in pollen samples are considered point by point, only the H1 hive sampled in January /19 has presented the risk (0.43).
Checking this scenario, we can verify that the ideal risk assessment must consider most complex analysis (Tier 2), including another aspects as behavior and physical aspects of T. angustula. When we consider the foraging efforts of pollen and nectar, bees can be exposed to pesticides by direct contact with contaminated flowers surface. This exposure route can be considered as lethal and break the colony balance, once foraging bees compose the majority of colony, representing 83% of the individuals in a hive of T. angustula (Prato et al. 2013); according to literature, this species makes about 40 daily flights (Vida Natural), touching an average of 40 thousand flowers, and that on each flight it carries, on average, 40 mg of pollen to the hive (Fujiyoshi, H. - Beekeeping Department Yamada). Using the highest concentration of difenoconazole detected in the strawberry flower (7.53 ng g− 1), and that the bee can carry, on average, 40 mg of pollen per flight, we have to, per flight, it “comes in contact” with 0.3012 ng of difenoconazole, still, considering the average of 40 daily flights, we have daily physical exposure to 12.048 ng bee day− 1. It is important to note that this value should be a little lower, since the pollen's contact area with the bee is not exactly the value of the mass it carries, therefore, it is an approximation.
Considering this previous risk assessment results, and, due to the absence of toxicity data for T. angustula, we cannot rule out the possible harm that the presence of this pesticide can cause in the bees’ lives, as behavioral changes, decreased birth rates, a lower expectation of life, and others. Therefore, a study of the real impacts for this stingless bee species is necessary, considering the uptake, lethal and sublethal effects under laboratory and field studies and involving adults and all brood (eggs, larvae, and pupae), mainly when the systemic pesticides are considered.