from Dried Tea Leaves to Brewed Tea - ACS Publications - American

Dec 30, 2013 - Canadian Food Inspection Agency, Calgary Laboratory, 3650-36th Street N.W., ... international food safety regulatory agencies, such as ...
2 downloads 0 Views 3MB Size
Article pubs.acs.org/JAFC

Determination of Pesticide Residue Transfer Rates (Percent) from Dried Tea Leaves to Brewed Tea Jian Wang,* Wendy Cheung, and Daniel Leung Canadian Food Inspection Agency, Calgary Laboratory, 3650-36th Street N.W., Calgary, Alberta T2L 2L1, Canada ABSTRACT: This paper presents a study on pesticide residue transfer rates (%) from dried tea leaves to brewed tea. In the study, a brewing procedure simulated the preparation of a hot tea drink as in routine. After brewing, pesticide residues were extracted from brewed tea using a method known as QuEChERS (quick, easy, cheap, effective, rugged, and safe). An UHPLC/ ESI-MS/MS method was developed and validated to identify and quantify up to 172 pesticides in both tea leaves and brewed tea samples. Quantification was achieved using matrix-matched standard calibration curves with isotopically labeled standards or a chemical analogue as internal standards, and the calibration curves consisted of six points (0.4, 2.0, 8.0, 16.0, 24.0, and 40.0 μg/L equivalent in sample). The method was validated at four concentration levels (4.0, 12, 20.0, and 32.0 μg/L equivalent in sample) using five different brewed tea matrices on two separate days per matrix. Method performance parameters included overall recovery, intermediate precision, and measurement uncertainty, which were evaluated according to a nested experimental design. Approximately, 95% of the pesticides studied had recoveries between 81 and 110%, intermediate precision ≤20%, and measurement uncertainty ≤40%. From a pilot study of 44 incurred tea samples, pesticide residues were examined for their ability to transfer from dried tea leaves to brewed tea. Each sample, both tea leaves and brewed tea, was analyzed in duplicate. Pesticides were found to have different transfer rates (%). For example, imidacloprid, methomyl, and carbendazim had transfer rates of 84.9, 83.4, and 92.4%, respectively. KEYWORDS: pesticides, UHPLC/ESI-MS/MS, tea, brewed tea, transfer rate



INTRODUCTION World tea production and consumption in 2010 were about 4.12 and 4.04 million tonnes, respectively. In the same year, tea exports and imports reached 1.67 and 1.80 million tonnes, respectively, worldwide.1 Clearly, tea is one of the most popular beverages in the world, especially in Asia. Tea is rich in phenolics such as catechins and flavonols, which show in vitro and in vivo strong antioxidant activities. It also contains minerals and vitamins, which may increase its antioxidant potential.2 Recent human studies have suggested that tea, especially green tea, might contribute to a reduction of cardiovascular disease and some forms of cancer. These studies also have indicated that tea promoted other health benefits such as antihypertensive effect, body weight control, antibacterial and antiviral activity.2 Consequently, tea is considered a healthful and functional beverage. On the basis of manufacturing processes, tea is classified into three categories: green tea (nonfermented), oolong tea (semifermented), and black (fermented) or red (Pu-Erh) tea. Green tea typically contains more catechins than black or oolong tea. Tea is sometimes scented with various plant essential oils such as lemon, bergamot, rose, or fragrant olive, which impart sweet floral attributes to enhance the natural flavor of the tea. Other types of flavored teas are made by blending tea leaves with flower petals, spices, or dried leaves from plants such as jasmine, chrysanthemum, rosemary, chamomile, and peppermint.3 Application of pesticides during tea cultivation is a common practice for pest and plant disease control. Pesticide residues, which might pose a potential risk for human health due to their subacute and chronic toxicities, could possibly end up in readyto-consume loose teas or tea bags. Therefore, national and/or Published 2013 by the American Chemical Society

international food safety regulatory agencies, such as the European Union and Japan, have set maximum residue limits and established monitoring programs to test for pesticide residues in tea. In Canada, pesticide residues in tea have been monitored under the Canadian National Chemical Residues Monitoring Program and/or the Food Safety Action Plan. Generally, tea is consumed after brewing in hot water. Therefore, pesticide residues may be likely leached from dried tea leaves into brewed tea, and humans are subsequently exposed. In our study, we aimed to investigate pesticide residue transfer rates from tea leaves to brewed tea after brewing by hot water. The challenge was to quantify pesticide residues at parts per billion (ppb) concentration levels in both tea leaves and brewed tea. Some publications have previously reported transfer rates of a few pesticides. For example, chlorpyrifos, ethion, quinalphos, and endosulfan had transfer rates of 1.7−9.2% (by GC with electron capture detection (GC-ECD) and LC with ultraviolet detection (LC-UV));4 fenvalerate, 10−30% (by LC-UV);5 and acetamiprid, 36.8−50.0% (by LC-UV).6 These studies were carried out at relatively high concentration (parts per million (ppm)) levels and for a very limited number of pesticides. In the current study, we used an UHPLC/ESI-MS/MS to analyze up to 172 different pesticides at parts per billion concentrations in both tea and brewed tea samples so as to investigate their transfer rates. Received: Revised: Accepted: Published: 966

September 13, 2013 December 23, 2013 December 30, 2013 December 30, 2013 dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. UHPLC/ESI-MS/MS Parameters and Method Performance for Analysis of Pesticide Residues in Brewed Tea

967

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued

968

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued

969

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued

970

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued

971

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued

972

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued

973

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued

974

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 1. continued



MATERIALS AND METHODS

run time was 12 min. The UHPLC gradient profile is shown in Table 2. MS/MS Conditions. The ion source was a TurboIonSpray or Turbo V electrospray ion source in positive mode. Pause time between mass ranges was 5 ms. Specific mass spectrometric parameters such as dwell time, declustering potential (DP), entrance potential (EP), collision energy (CE), collision cell exit potential (CXP), and multiple reaction monitoring transitions (MRM or Q1 and Q3) are listed in Table 1. Parameters including DP, EP, CE, and CXP were optimized using the Quantitative Optimization bundled with the Analyst software by infusing each individual pesticide standard (10 or 50 μg/L) into the mass spectrometer. The syringe pump (Harvard Apparatus, USA) flow rate was set at 10 μL/min for infusion. Scheduled MRM was set according to the retention time of pesticides using a MRM detection window of 100 s. The total scan time was 1.6211 s, and the duration was 11 min. Other general mass spectrometric parameters are shown in Table 2.

Materials and Reagents. For the pesticide residue transfer rate study, 44 dried tea leaf samples, which had tested positive for one or more pesticides, were obtained from a 2011 targeted survey program under the Food Safety Action Plan in Canada. All samples (50 g per samples) were well blended in CFIA contract laboratories prior to shipment to the Calgary Laboratory for the study. For method development and validation, blank samples of tea leaves (pesticide free), including Earl Grey, oolong, green tea, herbal tea and orange pekoe, were obtained from local markets. Tea bags were cut and emptied into mixing bowls, then mixed to create a composite. Loose tea leaf samples (200−500 g per sample) were blended using a food blender. All samples were stored at room temperature. Ammonium acetate (reagent grade and LC-MS grade) and LC-MS acetonitrile (Chromasolv, 2.5 L) were purchased from Sigma-Aldrich Corp. (Canada). ENVIRO CLEAN extraction columns (6.0 g of anhydrous magnesium sulfate (MgSO4) and 1.5 g of sodium acetate, 50 mL centrifuge tubes) and ENVIRO CLEAN extraction columns ((900 mg of MgSO4, 150 mg C18, and 300 mg of primary−secondary amine (PSA), 15 mL centrifuge tubes) or (900 mg of MgSO4, 150 mg of graphitized black carbon, and 300 mg of PSA, 15 mL centrifuge tubes)) were obtained from United Chemical Technologies, Inc. (Bristol, PA, USA). Acetic acid (glacial acetic acid, reagent grade, 99.7%), acetonitrile (distilled in glass), and methanol (distilled in glass) were obtained from Caledon Laboratories Ltd. (Canada). Water (18.2 MΩ·Cm) used for reagent and sample preparation was from a Barnstead Nanopure system (Thermo Scientific, USA). Pesticide standards (Table 1, column 1) were obtained from EQ Laboratories Inc. (USA), Riedel-de Haen AG (Germany), or Chem Service (USA). Internal standards carbendazim-d4 and carbofuran-d3 were purchased from EQ Laboratories Inc. (USA), and thiabendazole-d4 was from Chemical Synthesis Services (Northern Ireland). LC vials were Mini-UniPrep syringeless filter devices with polypropylene housing and PVDF 0.45 μm membrane (Whatman Inc., USA). UHPLC/ESI-MS/MS Parameters. The LC/ESI-MS/MS system utilized was an Agilent 1200 SL (Agilent, Germany) coupled with an API 5000 LC/MS/MS system (Applied Biosystem, Canada). The system was controlled using Analyst 1.5 software. Mobile phase B was acetonitrile, and mobile phase A was 10 mM ammonium acetate with 2% acetonitrile in water. The column oven temperature was set at 35 °C, and the autosampler temperature was set at 5 °C. Core−Shell Particles Column. The core−shell particles or UHPLC column utilized was a Kinetex C 18, 100 mm × 2.1 mm, 2.6 μm column (Phenomenex, USA). Injection volume was 3 μL, and total

Table 2. UHPLC Gradient Profiles and MS Parameters

Preparation of Standard Solutions. Individual pesticide standard stock solutions were generally prepared at a concentration of 4000.0 μg/mL in methanol. Due to its poor solubility in methanol, carbendazim was prepared at 200.0 μg/mL, and a few pesticides were prepared at 1000.0 or 2000.0 μg/mL (Table 1, column 1). Internal standard stock solution including carbofuran-d3, carbendazim-d4, and 975

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 3. Pesticide Residue Transfer Rate Simulation and Analytical Range Determination

thiabendazole-d4 was prepared at 100.0 μg/mL. Intermediate pesticide standard mix working solutions were prepared at 10.0 μg/mL. Stock and intermediate solutions were stored at −20 °C. Dried Tea Leaves. A six-level pesticide standard mix working solution was prepared by transferring 0.1, 0.5, 2.0, 4.0, 6.0, and 10.0 mL of 10.0 μg/mL intermediate working solution into six separate 50 mL volumetric flasks and making up to volume with methanol. This resulted in 0.02, 0.1, 0.4, 0.8, 1.2, and 2.0 μg/mL standard solutions, which were used to construct matrix-matched standard calibration curves. Internal standard working solution (2.0 μg/mL) was prepared from their respective stock in a mixture of acetonitrile and methanol (50:50, v/v). All working solutions were stored at 4 °C. Brewed Tea. A six-level pesticide standard mix working solution was prepared by transferring 0.1, 0.5, 2.0, 4.0, 6.0, and 10.0 mL of 10.0 μg/mL intermediate working solution into six separate 50 mL volumetric flasks and making up to volume with methanol. This resulted in 0.02, 0.1, 0.4, 0.8, 1.2, and 2.0 μg/mL standard solutions, which were used to construct matrix-matched standard calibration curves. Four-level sample spike pesticide standard working solutions were prepared by transferring 2.0, 6.0, 10.0, and 16.0 mL of 10.0 μg/mL intermediate working solution into separate 50 mL volumetric flasks and making up to volume with methanol. This resulted in 0.4, 1.2, 2.0, and 3.2 μg/mL standard solutions, which were used for sample fortification. Internal standard working solution (0.8 μg/mL) was prepared from their stock in a mixture of acetonitrile and methanol (50:50, v/v). All working solutions were stored at 4 °C. Preparation of Reagent Solutions. Acetonitrile/acetic acid (99 + 1, v/v) was prepared by mixing 990 mL of acetonitrile with 10 mL of acetic acid. Ammonium acetate solution (0.1 M) was prepared by weighing 7.7 g of ammonium acetate and dissolving in 800 mL of water. After transfer into a 1000 mL volumetric flask, the solution was made up to volume with water. Solvent buffer was a mixture of 0.1 M ammonium acetate and methanol (50 + 50, v/v). Sample Preparation and Extraction Procedure. Dried Tea Leaves. Sample extraction and cleanup procedures followed the buffered QuEChERS method.7 An incurred tea sample (5.0 g/sample) was weighed into individual 50 mL polypropylene centrifuge tubes (VWR International, Canada). To each tube was added 250 μL of 2.0 μg/mL internal standard working solution (100.0 μg/kg equivalent in samples) and 15 mL of water. Tubes were capped, and the contents were mixed and allowed to stand for the purpose of hydration. After 30 min of hydration at room temperature, 15 mL of acetonitrile/acetic acid (99 + 1, v/v) was added to individual samples and mixed, followed by the addition of 1.5 g of sodium acetate anhydrous and 6.0 g of magnesium sulfate anhydrous. The centrifuge tubes were capped, and all samples were shaken for 1 min using a Geno/Grinder 2010 at 1500 rpm (SPEX SamplePrep, USA) and then were centrifuged at 3000 rpm (∼2100g) for 3 min using an Allegra 6 centrifuge (Beckman Coulter Inc., USA). Supernatants were transferred (7 mL/sample) into individual 15 mL polypropylene centrifuge tubes that contained 900 mg of MgSO4, 150 mg of graphitized black carbon, and 300 mg of PSA. The centrifuge tubes were capped, shaken for 45 s by hand, and centrifuged at 3000 rpm (∼2100g) for 3 min. Three milliliters of supernatant (1 g sample/3 mL) was transferred into individual 5 mL Pyrex brand centrifuge tubes, precalibrated with 1 mL volume accuracy

(VWR International, Canada). Each of the sample extracts was evaporated to 0.1−0.2 mL, which took approximately 45 min, using an N-EVAP nitrogen evaporator (Organomation Associates Inc., USA) at 30 °C under a stream of nitrogen. The extracts were made up to 0.5 mL with methanol, vortexed for 30 s, and then made up to 1.0 mL with 0.1 M ammonium acetate and vortexed again for 30 s. One hundred microliters of each extract was transferred into a Mini-UniPrep vial (Whatman Inc., USA), and 500 μL of solvent buffer (a mixture of 0.1 M ammonium acetate/methanol, 50 + 50, v/v) was added. The vials were capped, vortexed for 30 s, and pressed to filter. Sample extracts were injected to the UHPLC/ESI-MS/MS. All dried tea leaves (incurred samples) were analyzed in duplicate. Brewed Tea. Dried tea leaves were brewed as follows. Five grams of dried tea leaves were weighed into a 250 mL Erlenmeyer flask, to which 150 mL of boiling water (distilled) was added. The sample was swirled to mix after 3 and 6 min with a total brewing time of 10 min. After the brewing process, the brewed tea samples were cooled by immersing the flask in cold water (∼7 °C) for 20 min. Then, 10 mL of brewed tea was subsampled into individual 50 mL polypropylene centrifuge tubes (VWR International, Canada) for the QuEChERS extraction and cleanup. For the fortification experiment, 100 μL of each four-level sample spike pesticide standard working solution was added into four centrifuge tubes to provide 4.0, 12.0, 20.0, and 32.0 μg/L pesticide equivalents in samples. To each tube was added 100 μL of 0.8 μg/mL internal standard working solution (8.0 μg/kg equivalent in samples). After 10 min at room temperature, 10 mL of acetonitrile/acetic acid (99 + 1, v/v) was added to individual samples and mixed for 45 s, followed by the addition of 1.5 g of sodium acetate anhydrous and 6.0 g of magnesium sulfate anhydrous. Four to six samples were processed at a time and shaken immediately by hand to avoid clumping. All samples were then shaken for 1 min using Geno/Grinder 2010 at 1500 rpm (SPEX SamplePrep, USA) and were centrifuged at 3000 rpm (∼2100g) for 3 min using an Allegra 6 centrifuge (Beckman Coulter Inc., USA). Supernatants were transferred (5 mL/sample) into individual 15 mL polypropylene centrifuge tubes (VWR International, Canada) that contained 900 mg of MgSO4, 150 mg of C18, and 300 mg of PSA. The centrifuge tubes were capped, shaken for 45 s by hand, and centrifuged at 3000 rpm (∼2100g) for 3 min. Two and half milliliters of supernatant was transferred into individual 5 mL Pyrex brand centrifuge tubes, precalibrated with 1 mL volume accuracy (VWR International, Canada). Each of the sample extracts was evaporated to 0.1 − 0.2 mL, which took approximately 30−45 min, using an N-EVAP nitrogen evaporator (Organomation Associates Inc., USA) at 30 °C under a stream of nitrogen. The extracts were made up to 0.5 mL with methanol, vortexed for 30 s, and then made up to 1.0 mL with 0.1 M ammonium acetate and vortexed again for 30 s. Five hundred microliters of each extract was transferred into a Mini-UniPrep vial (Whatman Inc., USA), and the vials were capped, vortexed for 30 s, and pressed to filter. Sample extracts were injected on an UHPLC/ESI-MS/MS system. All brewed tea (incurred) were analyzed in duplicate. Preparation of Matrix-Matched Calibration Standards and Calculation. Dried Tea Leaves. Matrix-matched calibration standards were prepared by adding standards and internal standards to blank sample extracts after sample extraction and cleanup. A blank tea sample (5.0 g/sample) was weighed into a 50 mL centrifuge tube, and the 976

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Figure 1. UHPLC/ESI-MS/MS matrix effects. The 172 pesticides were prepared in matrix extracts (a total of 5 brewed tea matrices) and solvent buffer at a concentration of 8.0 μg/kg equivalent in sample.

Figure 3. (A) Frequency (%) of 12 pesticides found in 44 incurred samples. (B) Number of positive samples, average transfer rate (%), and overall recovery in brewed tea (%). Pesticides: (1) imidacloprid; (2) methomyl; (3) carbendazim; (4) thiabendazole; (5) carbaryl; (6) monocrotophos; (7) clothianidin; (8) thiacloprid; (9) 3-hydroxycarbofuran; (10) thiamethoxam; (11) methidathion; (12) quizalofop. (100.0 μg/kg equivalent in samples). The extracts were made up to 0.5 mL with methanol, vortexed for 30 s, made up to 1.0 mL volume with 0.1 M ammonium acetate and then vortexed again for 30 s. The extracts were diluted six times using the solvent buffer prior to UHPLC/ ESI-MS/MS injection. Brewed Tea. Matrix-matched calibration standards were prepared by adding standards and internal standards to blank sample extracts after sample extraction and cleanup. A blank brewed tea sample (10.0 mL/sample) was weighed into a 50 mL centrifuge tube, and the sample was processed through the extraction procedure as described above for brewed tea. Fifty microliters of each six-level pesticide standard mix working solution was transferred into six separate blank sample extracts, providing 0.4, 2.0, 8.0, 16.0, 24.0, and 40.0 μg/L per standard equivalent in samples. Then, 25 μL of 0.8 μg/mL internal working solution was added to each sample (8.0 μg/L equivalent in samples). The extracts were made up to 0.5 mL with methanol, vortexed for 30 s, made up volume to 1.0 mL with 0.1 M ammonium acetate, and then vortexed again for 30 s. Five hundred microliters of each extract was transferred into a Mini-UniPrep vial (Whatman Inc., USA), and the vials were capped, vortexed for 30 s, and pressed to filter. Sample extracts were analyzed using UHPLC/ESI-MS/MS. To compensate for matrix effects, the oolong tea was used as the standard matrix to prepare matrix-matched calibration standards in both dried tea leaves and brewed tea. Matrix-matched standard calibration curves for each individual pesticide were constructed using MultiQuant. Concentration, (μg/kg or μg/L (ppb)) versus the ratio (analyte area/IS area) of each individual pesticide was plotted. Deuterium-labeled standards carbendazim-d4, carbofuran-d3, and thiabendazole-d4 were used as internal standards for their respective native compounds for quantification. Other pesticides

Figure 2. UHPLC/ESI-MS/MS method performance for analysis of 172 pesticides in brewed tea: (A) overall recovery; (B) intermediate precision; (C) measurement uncertainty. sample was processed through the extraction procedure as described above for dried tea leaves. Two hundred and fifty microliters of each sixlevel pesticide standard mix working solution was added into each of six blank sample extracts, providing 5.0, 25.0, 100.0, 200.0, 300.0, and 500.0 μg/kg per standard equivalent in samples. Then, 50 μL of 2.0 μg/mL internal working solution was added to each sample 977

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 4. Pesticide Residue Transfer Rates (%) from Dried Tea Leaves to Brewed Tea Determined Using UHPLC/ESI-MS/MS

978

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 4. continued

used carbofuran-d3 as an internal standard for quantification. In general, the quadratic function was applied to the calibration curves based on the line of best fit. 1/x weighting was used to improve the accuracy for quantification of pesticide residues at low concentration. Responses for the unknown concentration or fortified samples were compared to the curves to calculate the amount of pesticide residues (μg/kg or μg/L

(ppb)) in samples. Matrix-matched calibration standards were prepared fresh for each batch of samples. Method Validation. The method for brewed tea was validated by a nested experimental design, described elsewhere.8 The main factors of variances associated with method performance or measurement uncertainties of an in-house validated method are concentrations or 979

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 5. Pesticide Ion Ratios from Dried Tea Leaves and Brewed Tea Determined by UHPLC/ESI-MS/MS

980

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Table 5. continued

spike levels of analytes, matrix effects, day-to-day variation, and withinday variation of the method. The last two factors are designated the intermediate precision. In this study, there were a total of five brewed teas, that is, Earl Grey, oolong, green tea, herbal tea, and orange pekoe. For each matrix, samples were spiked at four levels, that is, 4.0, 12.0, 20.0, and 32.0 μg/L, in triplicate. Spike experiments were repeated on two different days or by each of two analysts. Overall recovery, intermediate precision, and measurement uncertainty were calculated using a combined algorithm that consisted of SAS codes (SAS software release 9.1, SAS Institute Inc., USA) along with a Microsoft Excel (Microsoft Office 2002) workbook.

solvent buffer (concentration in vials) for most pesticides studied, which was equivalent to 0.4 μg/L to 40.0 μg/kg in brewed tea as a result of a factor of 2.5 times concentration from brewed tea to final sample extracts. Assuming a pesticide residue will transfer with a rate between 20 and 100%, Table 3 presents four scenarios of the lowest or highest concentration of an incurred pesticide that could possibly be detected in tea leaves or brewed tea when following the procedure described under Materials and Methods. For example, if a pesticide has a transfer rate of 20%, the lowest concentration of a pesticide in tea leaves that can be detected in brewed tea (Table 3A) is 60 μg/kg; and a transfer rate of 100% leads to as low as 12.0 μg/kg in tea leaves (Table 3B). In these cases, when the amount of a pesticide in dried tea leaves happens to be over 40 μg/L in brewed tea (Table 3D), the sample size (tea leaves) can be reduced to fit its level in the range of 0.4−40.0 μg/kg for quantification. Basically, the experiment was designed to detect pesticides in a range from 12.0 to 1200.0 μg/kg in tea leaves when a 100% transfer rate was assumed. QuEChERS. Pesticides were extracted from dried tea leaves (5 g/sample) and brewed tea (10.0 mL/sample) following the buffered QuEChERS method7 or AOAC Official Method 2007.01.9 The complete procedure consisted of three steps including the first step, extraction with acetonitrile containing 1% acetic acid, MgSO4, and sodium acetate; and the second step, cleanup by dispersive solid-phase extraction (d-SPE) using MgSO4, PSA, and graphitized black carbon for dried tea leaves or MgSO4, PSA, and C18 for brewed tea. Graphitized black carbon helps to remove some pigments such as catechins and chlorophyll from tea leaves while leaving polar pesticides behind in the acetonitrile extract. It should be mentioned that the final extracts were still relatively dark in color, especially for black tea. The third step comprised concentration, reconstitution, and filtration. Concentration and reconstitution served as an additional cleanup step to remove particles or pigments, which were precipitated out during the process. Extracts were diluted six times for tea leaves, and no dilution was used for brewed tea (filtered to remove particles) prior to UHPLC/ESI-MS/MS injections. The QuEChERS method proved to be a practical extraction procedure for UHPLC/ESI-MS/MS analysis of pesticide residues in tea. Matrix Effects. The matrix could either enhance or suppress ionization of pesticides during the electrospray process. Its effects might vary from sample to sample and ultimately affect the UHPLC/ESI-MS/MS quantitative results. Previous experiments indicated that the matrix effects were significant in tea leaves,



RESULTS AND DISCUSSION Experimental Design. To study pesticide residue transfer rates from dried tea leaves to brewed tea, a two-stage experiment was deliberately designed. First, tea leaves were brewed, and then the resulting tea was analyzed for pesticide residues. Five grams of dried tea leaves and 150 mL of hot water were used in the brewing process. Data from a few preliminary experiments showed this combination increased method sensitivity or decreased detection limits to low parts per billion. This allowed us to use a 10 mL subsample of brewed tea, a typical QuEChERs sample size, to quantify a pesticide at a concentration as low as 12.0 μg/kg in dried tea leaves. The thermohydrolysis, which may induce the loss of certain pesticides, was not investigated in the current study because we were concerned with the concentrations or amounts of pesticides in brewed tea after the brewing process. The study on thermohydrolysis would be of value, especially for investigation of the production of additional pesticide metabolites. In general, the amounts of incurred pesticides in dried tea leaves, including green tea, oolong tea, and black tea, ranged from 5.0 to 500.0 μg/kg (ppb), as being confirmed by the previous year’s monitoring program results, and some residues were as high as a few millligram per kilogram (ppm). The challenge was to quantify pesticides at parts per billion concentration levels in both dried tea leaves and brewed tea, especially after the brewing process. Some published studies have reported transfer rates for a few pesticides from tea leaves to brewed tea.4−6 These studies were performed at relatively high residue concentrations, that is, parts per million levels, and for a very limited number of pesticides. In the current study, we utilized a highly sensitive UHPLC/ESI-MS/MS system with an appropriate experiment design to analyze up to 172 pesticides at parts per billion concentration levels to understand the transfer rates of a broad spectrum of pesticide residues. The API 5000 LC/MS/MS system used had an analytical range from 1.0 to 100.0 μg/L in 981

dx.doi.org/10.1021/jf404123h | J. Agric. Food Chem. 2014, 62, 966−983

Journal of Agricultural and Food Chemistry

Article

Figure 4. Examples of UHPLC/ESI-MS/MS chromatograms of three common incurred pesticides from one tea and its corresponding brew tea for quantification and confirmation. Imidacloprid transitions: 256.1→209.1 (T1), 256.1→174.9 (T2), and 256.1→84.0 (T3). Methomyl transitions: 163.1→ 87.9 (T1), 163.1→106.0 (T2), and 163.1→58.0 (T3). Carbendazim transitions: 192.1→160.2 (T1), 192.1→132.0 (T2), and 192.1→104.9 (T3).

elsewhere, 7 and the UHPLC/ESI-MS/MS method was revalidated when a core−shell column replaced the conventional C18 column in this study (data not shown). With the core−shell column method for dried tea leaves, 87% of the pesticides had recoveries between 81 and 110%, 94% had an intermediate precision ≤20%, and 89% showed measurement uncertainty ≤40%. The method performance of UHPLC/ESI-MS/MS was very comparable to that of LC/ESI-MS/MS, which was published in 2011.7 The UHPLC/ESI-MS/MS analysis of pesticides in brewed tea was a newly developed method and was therefore fully validated according to a nested design10 to evaluate method performance characteristics including accuracy expressed as overall recovery, intermediate precision, and measurement uncertainty (MU). Four factors, that is, concentrations or spike levels of pesticides,

and only 8−16% of the pesticides showed ion suppression of