Abstract
QuEChERS sample preparation was optimized and validated using solvent extraction with 10% (v/v) acetic acid-containing acetonitrile in the presence of four salts (anh. MgSO4, NaCl, sodium citrate tribasic dihydrate and sodium citrate dibasic sesquihydrate) and dispersive solid-phase extraction with mixed sorbents (octadecylsilane, primary and secondary amine and silica sorbents) for an ultra high performance liquid chromatography–tandem mass spectrometric determination of nine mycotoxins in brown rice: aflatoxins (AFB1, AFB2, AFG1 and AFG2), fumonisins (FB1 and FB2), deoxynivalenol, ochratoxin A and zearalenone (ZON). Our developed method allows for the determination of trace levels of mycotoxins with method detection limits in the range of 1.4–25 µg/kg, below the maximum limits of EU regulations, and with an acceptable accuracy and precision, and recoveries in the range of 81–101% with relative standard deviations of 5–19% over a mycotoxin concentration range of 5.0–1,000 µg/kg. Six out of fourteen real samples of brown rice were found to be contaminated with at least one of these mycotoxins, ranging from 2.49–5.41 µg/kg of FB1, 4.33 ± 0.04 µg/kg of FB2 and 6.10–14.88 µg/kg of ZON.
Introduction
Rice (Oryza sativa L.), categorized as a cereal, is a source of energy for more than half the world's population, especially in Asian countries. In general, rice is best known as milled rice, in which the germ and bran are removed from brown rice to improve its sensory qualities and storage stability (1, 2). However, rice germ and rice bran are beneficial nutrients and are sources of bioactive compounds, such as γ-oryzanols in bran or tocopherols in germ, with a strong antioxidant activity. Therefore, brown rice is considered a healthy food (3). During inappropriate harvesting or storage, brown rice may be contaminated with fungal mycotoxins.
The main mycotoxins include aflatoxin, ochratoxin, zearalenone (ZON), deoxynivalenol (DON) and fumonisin. Their forms of toxicity are classified as: carcinogenic, mutagenic, immunosuppressive and teratogenic for aflatoxin B1 (AFB1), aflatoxin B2 (AFB2), aflatoxin G1 (AFG1) and aflatoxin G2 (AFG2); kidney and liver toxicity for ochratoxin A (OTA); estrogenic effect for ZON; immunosuppressant for DON; and esophageal tumors and liver toxicity for fumonisin B1 (FB1) and fumonisin B2 (FB2) (4–6). According to the European Commission Regulation (EC) No. 1881/2006, which sets maximum levels for certain contaminants in foodstuffs in the European Union (EU), the maximum limits (ML) of mycotoxins in cereals and/or rice are regulated as follows: 5.0 µg/kg for AFB1; 10.0 µg/kg for the sum of AFB1, AFB2, AFG1 and AFG2; 5.0 µg/kg for OTA; 100 µg/kg for ZON; 1,250 µg/kg for DON; and 2,000 µg/kg for the sum of FB1 and FB2. Therefore, an effective analytical method to determine trace levels of mycotoxins in brown rice is necessary for consumer food safety.
High-performance liquid chromatography (HPLC) has been widely used for the analytical separation and determination of mycotoxins in a variety of samples, such as snacks (7), spices (8), bread (9), eggs (10), wheat flour (11), apple juice (12), dried fruit (13) and corn or maize (14–16). HPLC detectors for mycotoxin determination include UV-visible detectors or photodiode array detectors (12, 17, 18); fluorescence detectors (14, 17, 19–21); and mass spectrometers (MS), either MS or tandem MS (MS-MS) (5, 15, 16, 19, 22–29).
Ultra HPLC (UHPLC) is a powerful alternative separation technique to conventional HPLC (30) because UHPLC provides a higher resolution and faster separation owing to the smaller particle size of its stationary phase, typically <2 µm. In addition, UHPLC systems consume a smaller amount of the mobile phase, leading to less organic waste. The hyphenated technique of UHPLC with MS or MS-MS is widely used for the quantitative determination and identification of compounds in samples with complex matrices. UHPLC–MS-MS has previously been used to determine mycotoxins in various samples, including rice, corn, wheat and peanut (5); rice, wheat, oats, barley and corn (23); nuts and seeds (26); and eggs (28). However, prior to HPLC–MS analysis, careful sample preparation is required to remove matrices that may interfere with the detection of mycotoxins, reduce separation efficiency, shorten the column life or contaminate the MS interface.
Previous studies of sample preparation techniques for the determination of mycotoxins have involved simple solvent extraction (16, 22, 23), ultrasonic extraction (15) or solvent extraction and cleanup using an immunoaffinity column (IAC) (5, 20) or solid-phase extraction (SPE) (24). The disadvantages of these procedures include the use of a large amount of organic solvent, matrix interferences for a single step of solvent extraction, long time requirements due to the multiple steps of IAC or SPE, and expensive SPE cartridges. Recently, the QuEChERS (quick, easy, cheap, effective, rugged and safe) sample preparation method, a simple two-step technique based on solvent extraction in the presence of salts and dispersive-SPE (d-SPE) for clean-up, has been reported for the determination of mycotoxins in various samples using the following procedures: acetonitrile (ACN) extraction and d-SPE using octadecylsilane (C18) sorbent (31, 32); ACN extraction without d-SPE (7, 9, 33); ACN (5% v/v formic acid, FA) extraction without d-SPE (26); ACN (1% v/v FA) extraction without d-SPE (8); ACN (1% v/v acetic acid, HOAc) extraction without d-SPE (34); methanol (MeOH) with 1% v/v HOAc extraction without d-SPE (10, 28); ACN:MeOH extraction and d-SPE using primary and secondary amine (PSA) sorbent (11); and ACN (10% v/v FA) extraction and d-SPE using mixed C18, alumina-neutral and PSA sorbents (35). Compared with conventional solvent extraction and SPE, each step of QuEChERS sample preparation is easily performed. A small amount (5–10 mL) of organic solvent and <1 g of cheap d-SPE sorbent are applied in a plastic centrifuge tube, then vortexed and centrifuged for 2–6 min. This method can be partially automated for high-throughput analysis demands. Therefore, the QuEChERS sample preparation method provides the advantages suggested by its name.
Our preliminary study on QuEChERS sample preparation for the determination of mycotoxins in brown rice was performed using a QuEChERS procedure similar to that reported for milled rice analysis in a previous work (35). However, the mycotoxin extraction recovery was found to be <50%, which may be due to the presence of more complex matrices in brown rice than in milled rice. In order to minimize these matrices for acceptable recovery, the reduction of sample sizes with suitable sorbents and other factors for QuEChERS were required, particularly for brown rice. It should be noted that a quantitative multi-mycotoxins method using UHPLC–MS-MS was previously reported for rice and other samples such as maize, wheat and peanut (5). However, the solvent extraction and Myco6in1™ IAC cleanup was used for sample preparation in previous work, while the modified QuEChERS sample preparation was used in this work with more complex matrices in the brown rice than in milled rice as previously mentioned. Up to date, previous work on the QuEChERS sample preparation for brown rice has not been reported, and therefore this work aimed to optimize and validate QuEChERS sample preparation for the determination of mycotoxins in brown rice prior to UHPLC–MS-MS analysis.
Experimental
Chemicals and reagents
ACN and MeOH were purchased from JT Baker (Center Valley, PA, USA), chloroform from VWR (Fontenay-sur-Bios, France), benzene from Carlo Erba (Ronado, MI, USA), ethyl acetate (EtOAc) from JT Baker (Phillipsburg, NJ, USA), HOAc, FA and silica from Merck (Darmstadt, Germany), ammonium formate from Acros Organics (NJ, USA), anhydrous magnesium sulfate (anh. MgSO4) from Panreac (Barcelona, Spain), sodium chloride (NaCl) from Univar (Ingleburn, NSW, Australia), sodium citrate tribasic dihydrate and sodium citrate dibasic sesquihydrate from Merck, and PSA, C18 and graphitized carbon black (GCB) from Supelco® (Bellefonte, PA, USA). Brown rice samples were obtained from supermarkets in Bangkok (Thailand). Eight solid mycotoxin standards (AFB1, AFB2, AFG1, AFG2, FB1, OTA, ZON and DON) were purchased from Sigma-Aldrich (St. Louis, MO, USA). A solution standard of 50 µg/mL FB2 in ACN: water (50/50, v/v) was obtained from Fluka (Buchs, Switzerland). Sulfamethoxazole (SMX), used as an internal standard (IS) for all mycotoxins, was purchased from Sigma-Aldrich.
Preparation of standard solutions
Stock solutions of mycotoxin standards at 40 mg/L were individually prepared in benzene:ACN (98 : 2, v/v) for AFB1, AFB2, AFG1 and AFG2; benzene:HOAc (99 : 1, v/v) for OTA; EtOAc:MeOH (95:5, v/v) for DON; and ACN-water (50/50, v/v) for FB1. The stock solution of ZON at 400 mg/L was prepared in benzene. A stock solution of the SMX IS was prepared at 1 g/L in MeOH. A working mixed-standard solution was prepared by diluting individual stock solutions in MeOH to yield the desired amounts of each mycotoxin and 25 µg/L SMX.
QuEChERS sample preparation
Brown rice samples were blended and homogenized to a powder-like consistency. A powder sample (1.0 g) was weighed into a 50-mL centrifuge tube and 5 mL of water was added and mixed. To this suspension was added 5 mL of 10% (v/v) HOAc in ACN and vortexed for 1 min at high speed. Four salts (2.0 g anh. MgSO4, 0.50 g NaCl, 0.50 g sodium citrate tribasic dihydrate and 0.25 g sodium citrate dibasic sesquihydrate) were added into the mixture and then vigorously shaken by hand for 1 min. The mixture was then centrifuged for 5 min at 1,911×g and then 2 mL of supernatant was transferred into a 15-mL centrifuge tube containing 300 mg anh. MgSO4, 50 mg C18, 25 mg PSA and 25 mg silica. This portion was shaken and centrifuged, and then 1 mL of the supernatant was evaporated to dryness under a stream of nitrogen gas. An aliquot was reconstituted in 1 mL of water with a 1:1 (v/v) ratio of 0.1% (v/v) FA:MeOH and 0.5 µg/L of an SMX IS. The extracted solutions were filtered through 0.22-µm PTFE syringe filters prior to UHPLC–MS-MS analysis.
UHPLC–MS-MS conditions
A Model 1290 UHPLC system (Agilent Technologies, CA, USA), consisting of a vacuum degasser, binary pump, Agilent jet weaver, autosampler and a column oven, was utilized in this work. The UHPLC column used was an ACQUITY UPLC® HSS T3 C18, 100 × 2.1 mm ID, 1.8 µm (Waters, MA, USA), with a mobile-phase flow rate of 0.3 mL/min. The column temperature was controlled at 40 °C. Standard solutions or extracted solutions were injected at 10 µL. Mobile phase A was 0.5% (v/v) FA in water containing 5 mM ammonium formate, and mobile phase B was MeOH. The elution gradient profile was as follows: 0 min: 25% B; 7 min: 85% B; 8 min: 85% B; 9 min: 25% B; and 14 min: 25% B, yielding a data collection time of 9 min and a total run time of 14 min.
A Model 6490 (Agilent Technologies) triple quadrupole mass analyzer with electrospray ionization (ESI) and MassHunter software processing was used. ESI was operated in both the positive and negative modes with multiple-reaction monitoring (MRM) as shown in Table I. The SMX IS was set at the transition 254.0 > 91.9 in positive polarity mode using a collision energy (CE) of 25 eV and cell accelerator voltage (CAV) of 3 V. Additional MS conditions were as follows: 3,000 V capillary voltage, 1,000 V nozzle voltage, 16 L/min gas flow, 150 °C gas temperature, 20 psi nebulizer pressure, 11 L/min sheath gas flow, 400 °C sheath gas temperature, 380 V fragmentor and a 50 ms dwell time.
Table I.
MRM MS-MS parameters for each mycotoxin
| Analyte | Molecular ion |
Product ion 1 (quantitative) | Product ion 2 (qualitative) | CE1 | CE2 | CAV1 | CAV2 | |
|---|---|---|---|---|---|---|---|---|
| Form | m/z | |||||||
| AFB1 | [M + H]+ | 313.1 | 241.0 | 284.9 | 40 | 21 | 3 | 7 |
| AFB2 | [M + H]+ | 315.1 | 287.0 | 259.0 | 25 | 29 | 3 | 3 |
| AFG1 | [M + H]+ | 329.1 | 243.1 | 310.9 | 26 | 19 | 3 | 7 |
| AFG2 | [M + H]+ | 331.1 | 312.9 | 189.0 | 24 | 46 | 7 | 7 |
| FB1 | [M + H]+ | 722.4 | 334.3 | 352.1 | 45 | 40 | 5 | 5 |
| FB2 | [M + H]+ | 706.4 | 336.3 | 318.3 | 40 | 41 | 5 | 7 |
| DON | [M + H]+ | 297.1 | 249.1 | 203.1 | 7 | 9 | 3 | 5 |
| OTA | [M − H]− | 402.1 | 358.1 | 166.8 | 16 | 39 | 5 | 3 |
| ZON | [M − H]− | 317.1 | 131.0 | 174.9 | 30 | 24 | 3 | 7 |
Results
Optimization of the UHPLC–MS-MS conditions
The molecular ion of each mycotoxin was observed using MS2 scan mode. The mass to charge ratios (m/z) were set to cover the molecular mass of each mycotoxin. Both positive and negative polarity modes were used, but only the mode yielding the higher sensitivity was chosen for each particular analyte. Agilent MassHunter Optimizer software was used to find the product ions and suitable CE of each mycotoxin. The product ion yielding the highest abundance was selected as a quantitative m/z ion, while the lower-abundance ion was selected as a qualitative m/z ion. Then, MRM mode was used to find the suitable CE. The MS conditions used are shown in Table I. The other MS parameters were also optimized, such as the capillary voltage, nozzle voltage, gas flow, gas temperature, nebulizer pressure, sheath gas flow, sheath gas temperature, fragmentor and CAV.
In previous works on the HPLC–MS-MS determination of mycotoxins, various mobile phases have been used, such as MeOH–water and ACN–water containing FA (10, 11); ammonium formate (28); HOAc (9); ammonium acetate (34); and FA and ammonium formate (7, 8, 26, 27, 33), with gradient elution. In this work, nine versions of mobile phase A (water) containing 0.1, 0.5 or 1% (v/v) FA and 0, 5 or 10 mM ammonium formate were varied while keeping the mobile phase B as MeOH. The results showed that water containing 0.1% (v/v) FA and 0–10 mM ammonium formate gave poor tailing peaks for FB1 and FB2, and water containing 0.1–1% (v/v) FA without ammonium formate gave a lower sensitivity for the detection of AFB1, AFB2, AFG1 and AFG2. To achieve symmetric peak shapes and a high detection sensitivity, water containing 0.5% (v/v) FA and 5 mM ammonium formate was chosen for mobile phase A. The A:B mobile phase gradient profile for the elution of mycotoxins was also optimized according to the chromatogram given in Figure 1, showing a selectivity for the detection of all mycotoxins in the MRM mode of the MS-MS analysis.
Figure 1.
UHPLC–MS-MS MRM chromatograms of mycotoxins standard (10 µg/kg of AFB1, AFB2, AFG1 and AFG2; 20 µg/kg of FB1, FB2, OTA and ZON; 200 µg/kg of DON; 0.5 µg/kg of SMX).
QuEChERS optimization
The QuEChERS optimization was performed as described in the Experimental section (QuEChERS sample preparation), except that ACN was varied with and without FA or HOAc at 1, 5 and 10% v/v, and with buffering salts for the extraction of a blank sample (three batches) of brown rice spiked with the nine standard mycotoxins at known levels ACN. It can be seen from Figure 2 that ACN with 10% (v/v) HOAc gives better recovery within 82–94% for all analytes. Therefore, ACN with 10% (v/v) HOAc was chosen as the extraction solvent for studying on the effect of salts on extraction recovery.
Figure 2.
Average recovery (n = 3) of mycotoxins obtained from the QuEChERS sample preparation using ACN with and without acid for solvent extraction of a blank sample spiked with nine mycotoxins (50 µg/kg for AFB1, AFB2, AFG1, AFG2, FB1 and FB2; 200 µg/kg for OTA and ZON; 400 µg/kg for DON). Other QuEChERS conditions are given in Experimental section (QuEChERS sample preparation).
In this work, anh. MgSO4 and NaCl (4:1 (w/w) ratio) with and without buffer (citrate or acetate) were compared for QuEChERS extraction using the three different salt mixtures (SM): I (2.0 g anh. MgSO4 and 0.50 g NaCl), II (SMI plus 0.50 g sodium citrate tribasic dihydrate and 0.25 g sodium citrate dibasic sesquihydrate) and III (SMI plus 0.50 g sodium acetate). In most cases, the highest recovery was obtained in the order SMIII > SMII > SMI, except for FB1, FB2 and OTA, where the order of highest recovery was SMII > SMIII; in particular, SMIII yielded 49 ± 1% for FB1 and 75 ± 1% for FB2. It should be noted that the concentration and ratio of anh. MgSO4:NaCl in SMII was varied as (g:g) 0.50:0.125, 1.0:0.25, 4.0:1.0 and 2.0:0.50. However, by setting acceptable recoveries in the range of 80–110%, only one (DON), three (DON, FB1 and ZON) and seven (all except FB1 and FB2) out of the nine mycotoxins were found to be within an acceptable range for anh. MgSO4:NaCl (g:g) ratios of 0.50:0.125, 1.0:0.25 and 4.0:1.0, respectively, while the remaining recovery data were <80%. Therefore, the SM containing anh. MgSO4:NaCl:sodium citrate tribasic dihydrate:sodium citrate dibasic sesquihydrate (2.0:0.5:0.5:0.25, g:g:g:g) was chosen in this work, providing recoveries of 82–94% for all mycotoxins.
Various volume ratios (mL) of water and extraction solvent (ACN with 10% v/v HOAc), 5:10, 10:5 and 10:10, were compared with a ratio of 5:5. However, only four (FB1, FB2, OTA and ZON), five (DON, FB1, FB2, OTA and ZON) and four (FB1, FB2, OTA and ZON) out of the nine mycotoxins were found to be within an acceptable recovery range of 80–110% for water to extraction solvent v/v ratios of 5:10, 10:5 and 10:10, respectively, while the remaining recovery data were <80%.
To remove matrices, such as pigments and other components, d-SPE was performed using a 50-mg single sorbent (C18, PSA, GCB or silica) and 100-mg mixtures (mg:mg) of C18 and other sorbents: Sorbents I (C18:PSA 50:50), II (C18:GCB 50:50), III (C18:silica 50:50), IV (C18:PSA:GCB 50:25:25), V (C18:PSA:silica 50:25:25) and VI (C18:GCB:silica 50 : 25 : 25). Results are shown in Figure 3.
Figure 3.
Average recovery (n = 3) of mycotoxins obtained from the QuEChERS sample preparation using seven different sorbents for d-SPE of a blank sample spiked with nine mycotoxins at known amounts (given in Figure 2). Other QuEChERS conditions are given in Experimental section (QuEChERS sample preparation).
According to the QuEChERS standard method, with a 5- to 10-mL ACN solvent extraction step, sample sizes of 5–15 g are typically used. In this work, brown rice samples of 5, 2 and 1 g were used with 5 mL of ACN for extraction. A darker brown ACN extract, poorer baseline chromatogram and lower recovery of all mycotoxins were observed with sample sizes of 5 and 2 g, due to the interferences of the complex matrix in brown rice. Therefore, 1-g brown rice samples were extracted with 5 mL of ACN in this work.
Method validation
Analytical limits
According to the modern IUPAC recommendation (36), the limit of detection (LOD) is defined as the minimum analyte concentration that can be discriminated from the blank, controlling the risks of false positives and false negative, and therefore, may be expressed by Equation (1);
| (1) |
where A is the slope of the linear plot between the signal against the analyte concentration, Sy/x is the residual standard deviation, h0 is the leverage for the blank sample, as described in ref. (36), and I is the number of calibration samples. Similar concepts also apply to the limit of quantitation (LOQ) with the factor of 10 instead of 3.3 in Equation (1), to ensure a maximum relative prediction uncertainty of 10.
In order to obtain the LOD and LOQ values from a matrix-matched calibration plot, the blank sample was extracted using the QuEChERS sample preparation as described in the QuEChERS sample preparation section. The known amounts of mycotoxins in the QuEChERS extract were determined by UHPLC–MS-MS as described in UHPLC–MS-MS conditions. It should be noted that the method detection limit (MDL) and the method quantitation limit (MQL) are estimated from the LOD and LOQ values, respectively, with the QuEChERS sample preparation for 1 g of sample and 5 mL of the final extraction solvent. Results are shown in Table II.
Table II.
Analytical limits (LOD, LOQ, MDL and MQL) obtained from the IS calibration results
| Analyte | Conc. range (µg/L) | Calibration plot |
MDL (µg/kg) | MQL (µg/kg) | ML (µg/kg) | ||||
|---|---|---|---|---|---|---|---|---|---|
| Slopea | Sy/x | h0 | LOD (µg/L) | LOQ (µg/L) | |||||
| AFB1 | 0.06–2.5 | 5.82 ± 0.14 | 0.43 | 0.102 | 0.27 | 0.82 | 1.4 | 4.1 | 5.0 for AFB1; total of 10.0 for aflatoxins |
| AFB2 | 0.05–2.5 | 3.18 ± 0.09 | 0.27 | 0.100 | 0.31 | 0.94 | 1.6 | 4.7 | |
| AFG1 | 0.04–2.5 | 4.78 ± 0.14 | 0.44 | 0.099 | 0.33 | 1.0 | 1.7 | 5.0 | |
| AFG2 | 0.06–2.5 | 2.72 ± 0.09 | 0.29 | 0.102 | 0.39 | 1.2 | 2.0 | 6.0 | |
| FB1 | 0.25–5.0 | 0.858 ± 0.018 | 0.11 | 0.109 | 0.48 | 1.5 | 2.4 | 7.5 | Total of 2,000 for fumonisins |
| FB2 | 0.20–5.0 | 0.844 ± 0.020 | 0.12 | 0.106 | 0.54 | 1.7 | 2.7 | 8.5 | |
| OTA | 0.35–5.0 | 0.293 ± 0.006 | 0.038 | 0.114 | 0.47 | 1.5 | 2.4 | 7.5 | 5.0 |
| ZON | 0.15–5.0 | 0.335 ± 0.012 | 0.079 | 0.103 | 0.85 | 2.6 | 4.3 | 13 | 100 |
| DON | 6.0–50 | 0.0372 ± 0.0009 | 0.051 | 0.112 | 5.0 | 15 | 25 | 75 | 1,250 |
MDL and MQL were estimated from the LOD and LOQ values using 1 g of sample with 5 mL of the final extraction solvent (dilution factor of 5).
aData are obtained from four triplicate concentration levels.
Matrix effect and linearity
To investigate matrix effects, which may interfere by increasing or decreasing the detection signal of analytes, two curves for the IS calibration and matrix-matched IS calibration were compared using five triplicate concentration levels of the mycotoxin standard (N1 of 5 and N2 of 5). The former curve was established using standards dissolved in 50:50 (v/v) of water with 0.1% (v/v) FA:MeOH (data not shown), while the latter curve was constructed using standards spiked into the QuEChERS extract of blank samples. Results of calibration parameters obtained are shown in Table III. Slope values were calculated for the linear fit of the two calibration curves, and the relative difference in slopes (%Δ) was found to be 119, 144, 128, 109, −75, −66, −80, 1.6 and 103% for AFB1, AFB2, AFG1, AFG2, FB1, FB2, OTA, ZON and DON, respectively. Note that %Δ is calculated from 100 × (slopeS − slopeM)/slopeM, where subscripts M and S indicate the matrix-matched and standard calibration methods, respectively. In addition, comparison of the slopes of the two linear regression plots using the experimental t value (texp), as given in previous works (36, 37), gave texp values of 36, 45, 35, 42, 54, 66, 60, 0.77 and 28 for AFB1, AFB2, AFG1, AFG2, FB1, FB2, OTA, ZON and DON, respectively. According to the F-test as a best linearity indicator, as recommended by IUPAC (36, 38), the experimental F values (Fexp) were calculated, where Fexp is defined as the ratio of residual variance and squared pure error .
Table III.
Calibration parameters for quantitative analysis: slope, intercept, correlation coefficient (r2), residual varience (Sy/x2), squared pure error (Sy2) and experimental F value (Fexp=Sy/x2/Sy2)
| Analyte | Conc. range (µg/L) | Calibration plot |
|||||
|---|---|---|---|---|---|---|---|
| Slopea | Intercepta | r2 | Fexp | ||||
| AFB1 | 1.0–10 | 5.51 ± 0.09 | 0.2 ± 0.5 | 0.997 | 1.2 | 2.9 | 0.41 |
| AFB2 | 1.0–10 | 2.95 ± 0.04 | 1.0 ± 0.3 | 0.997 | 0.31 | 0.63 | 0.49 |
| AFG1 | 1.0–10 | 4.63 ± 0.08 | 0.4 ± 0.5 | 0.996 | 1.0 | 2.3 | 0.43 |
| AFG2 | 1.2–10 | 2.90 ± 0.04 | 0.1 ± 0.3 | 0.998 | 0.25 | 0.58 | 0.43 |
| FB1 | 2.0–20 | 0.900 ± 0.013 | −0.10 ± 0.16 | 0.997 | 0.11 | 0.25 | 0.44 |
| FB2 | 2.0–20 | 0.885 ± 0.009 | −0.27 ± 0.11 | 0.999 | 0.050 | 0.11 | 0.45 |
| OTA | 2.0–20 | 0.296 ± 0.004 | −0.09 ± 0.05 | 0.997 | 0.011 | 0.010 | 1.1 |
| ZON | 2.6–20 | 0.318 ± 0.005 | −0.02 ± 0.07 | 0.996 | 0.018 | 0.021 | 0.86 |
| DON | 15–200 | 0.0367 ± 0.0006 | −0.15 ± 0.07 | 0.997 | 0.021 | 0.028 | 0.75 |
aData are obtained from five triplicate concentration levels. Critical F(0.05,13,10) value of 2.9.
Accuracy and precision
The accuracy and precision of this quantitative analysis method were evaluated by spiking mycotoxin standards into a blank sample of brown rice, with 10 batches for each of the three concentration levels, and the mycotoxin content after QuEChERS extraction was determined by UHPLC–MS-MS analysis using the matrix-matched calibration method. Results of the accuracy and precision, expressed by average recovery and relative standard deviation (RSD), respectively, are shown in Table IV.
Table IV.
Accuracy and precision in the QuEChERS extraction recovery of mycotoxins spiked in blank samples at three levels (n = 10)
| Analyte | Conc. (µg/kg) | % Recovery (% RSD) |
Overall |
||||
|---|---|---|---|---|---|---|---|
| Acceptable criteria | Day 1 | Day2 | Day3 | P-value for SD | % Recovery (% RSD) | ||
| AFB1 | 5.0 | 60–115 (23) | 92 (18) | 83 (20) | 89 (19) | 0.45 | 88 (19) |
| 25 | 80–110 (18) | 91 (10) | 91 (12) | 94 (9) | 0.63 | 92 (10) | |
| 50 | 80–110 (17) | 87 (9) | 83 (6) | 91 (8) | 0.086 | 87 (9) | |
| AFB2 | 5.0 | 60–115 (23) | 90 (17) | 83 (17) | 95 (11) | 0.15 | 89 (16) |
| 25 | 80–110 (18) | 91 (10) | 92 (9) | 97 (8) | 0.25 | 93 (9) | |
| 50 | 80–110 (17) | 84 (10) | 83 (7) | 87 (8) | 0.34 | 85 (9) | |
| AFG1 | 5.0 | 60–115 (23) | 81 (16) | 95 (18) | 89 (18) | 0.15 | 88 (18) |
| 25 | 80–110 (18) | 89 (9) | 90 (11) | 91 (9) | 0.87 | 90 (9) | |
| 50 | 80–110 (17) | 82 (10) | 83 (9) | 88 (8) | 0.16 | 84 (9) | |
| AFG2 | 6.0 | 60–115 (23) | 86 (19) | 89 (19) | 91 (14) | 0.74 | 89 (17) |
| 25 | 80–110 (18) | 88 (7) | 94 (9) | 87 (10) | 0.15 | 90 (9) | |
| 50 | 80–110 (17) | 84 (8) | 83 (8) | 85 (9) | 0.86 | 84 (8) | |
| FB1 | 10 | 60–115 (21) | 95 (13) | 98 (16) | 97 (14) | 0.88 | 97 (13) |
| 50 | 80–110 (17) | 91 (7) | 91 (12) | 93 (7) | 0.68 | 92 (9) | |
| 100 | 80–110 (15) | 90 (6) | 90 (5) | 93 (7) | 0.35 | 91 (6) | |
| FB2 | 10 | 60–115 (21) | 86 (18) | 88 (18) | 86 (18) | 0.93 | 87 (17) |
| 50 | 80–110 (17) | 89 (9) | 90 (11) | 90 (8) | 0.93 | 89 (9) | |
| 100 | 80–110 (15) | 88 (7) | 89 (6) | 92 (7) | 0.30 | 90 (7) | |
| OTA | 10 | 60–115 (21) | 86 (17) | 83 (13) | 85 (16) | 0.89 | 84 (15) |
| 50 | 80–110 (17) | 93 (9) | 92 (11) | 91 (8) | 0.95 | 92 (9) | |
| 100 | 80–110 (15) | 87 (7) | 86 (7) | 90 (9) | 0.41 | 88 (8) | |
| ZON | 13 | 80–110 (20) | 94 (12) | 95 (12) | 91 (4) | 0.67 | 93 (13) |
| 50 | 80–110 (17) | 97 (8) | 98 (10) | 96 (9) | 0.86 | 97 (9) | |
| 100 | 80–110 (15) | 100 (4) | 101 (4) | 96 (9) | 0.20 | 99 (6) | |
| DON | 75 | 80–110 (16) | 85 (11) | 85 (9) | 86 (12) | 0.90 | 85 (10) |
| 500 | 80–110 (12) | 88 (8) | 92 (8) | 91 (7) | 0.57 | 90 (8) | |
| 1,000 | 80–110 (11) | 86 (5) | 88 (4) | 89 (5) | 0.22 | 88 (5) | |
P-value of > 0.05 indicates no significant difference the different day precisions at a 95% confidence level.
Application to real samples
The QuEChERS method developed and validated was used for UHPLC–MS-MS determination of mycotoxins in 14 brown rice samples collected from supermarkets in Bangkok. Results are shown in Table V.
Table V.
Mycotoxins contamination in brown rice samples
| Sample | Range of contamination (µg/kg) |
||||||||
|---|---|---|---|---|---|---|---|---|---|
| AFB1 | AFB2 | AFG1 | AFG2 | FB1 | FB2 | DON | OTA | ZON | |
| 1 | nd | nd | nd | nd | 5.41 ± 0.02a | 4.33 ± 0.04a | nd | nd | 6.25 ± 0.11a |
| 2 | nd | nd | nd | nd | nd | nd | nd | nd | 14.88 ± 0.20 |
| 3 | nd | nd | nd | nd | 3.51 ± 0.04a | nd | nd | nd | nd |
| 4 | nd | nd | nd | nd | 2.64 ± 0.04a | nd | nd | nd | nd |
| 5 | nd | nd | nd | nd | nd | nd | nd | nd | 6.10 ± 0.14a |
| 6 | nd | nd | nd | nd | 2.49 ± 0.05a | nd | nd | nd | nd |
| 7 to 14 | nd | nd | nd | nd | nd | nd | nd | nd | nd |
nd, not detected, otherwise the data are shown as the mean ± SD, derived from duplicate runs.
aThe determined amount is between MDL and MQL.
Discussion
Optimization of UHPLC conditions
In this work, the mobile-phase gradient profile was optimized using either FA:MeOH or formate buffer:MeOH mobile phase. However, the latter mobile phase was suitable for UHPLC separation of mycotoxins, where the formate buffer contain 0.5% (v/v) FA and 5 mM ammonium formate in water. In comparison with previous work using the 0.1% (v/v) FA:MeOH mobile phase for UHPLC separation of mycotoxins (5), similar MRM chromatograms of our nine mycotoxins were obtained, except for the retention time order FB2 < OTA < ZON for this work, while OTA ≈ ZON < FB2 for previous work. The differences in the analyte elution order may be due to differences in the surface chemistry of the analytical columns between Acquity HSS T3 C18 and Acquity BEH C18 used in this and previous work (5), respectively, and also possibly due to the differences in the mobile-phase gradient programing used in these two works. It should be noted from this work that the same order FB2 < OTA < ZON was obtained using FA:MeOH and formate buffer:MeOH mobile phase, but the better symmetric peak shapes and higher detection sensitivity was observed using the latter mobile phase.
QuEChERS optimization
In this work, three (OTA, FB1 and FB2) out of the nine mycotoxins used are weak acids, while the rest are neutral. In previous work on QuEChERS sample preparation for the determination of acidic analytes, the organic solvent extraction step was mostly performed with the addition of FA (8, 26, 35) or HOAc (34) in the organic solvent with (26, 34, 35) or without (8) buffering salts. Therefore, in this work, ACN, which is commonly used in the solvent extraction step of QuEChERS, was tested with and without acid and with buffering salts for the extraction of a blank sample of brown rice spiked with the nine standard mycotoxins. As can be seen in Figure 2, increasing the FA or HOAc content in ACN from 0 to 10% (v/v) provides a better extraction efficiency for the three acidic mycotoxins (FB1, FB2 and OTA) with similar extraction efficiencies (recoveries of 79–104%) for the six neutral mycotoxins. These results were within the acceptable range of recoveries (80–110%) for analyte concentrations in the range of 50–400 µg/kg (39), except for the recoveries of 72 ± 1% for DON using ACN without acid, and 42 ± 2% for AFB1 and 43 ± 2% for AFG1 when using ACN with 10% (v/v) FA. This difference in recoveries for neutral and acidic mycotoxins occurred because FA or HOAc promotes the extraction of the neutral form of acidic mycotoxins into the ACN phase. However, ACN with 10% (v/v) FA gave much poorer recoveries for AFB1 and AFG1 (42–43% recovery) than did ACN with 5% (v/v) FA (79–80% recovery) or 10% (v/v) HOAc (84–85% recovery). The determination of AFB1 and AFG1 levels using ACN with 10 and 5% v/v FA or 10% (v/v) HOAc were repeated on three different days where the same trends were obtained on each separate day. The use of ACN with 10% v/v FA resulted in a darker brown extract from the brown rice, implying that the increased pigment matrices reduced the extraction efficiency or may have interfered with the detection of AFB1 and AFG1. Therefore, ACN with 10% (v/v) HOAc, giving recoveries within 82–94% for all mycotoxins, was chosen as the extraction solvent for all subsequent work.
MgSO4 and NaCl salts are widely used (typically at a 4:1 w/w ratio) to induce phase separation in the QuEChERS solvent during the extraction step (7, 8, 26, 27, 31–34). To maintain an aqueous buffer solution, additional salts may be included with these two salts, such as sodium citrate tribasic dihydrate and sodium citrate dibasic sesquihydrate for a citrate buffer, according to the Committee of European Normalization (European EN 15662 Method) and other previous works (26, 27, 34, 35); or sodium acetate for an acetate buffer, according to standard methods from the Association of Official Analytical Chemists (AOAC Official Method 2007.01) and other previous works (10, 28). In comparison with MgSO4 and NaCl salts with an acetate buffer, MgSO4 and NaCl salts with a citrate buffer giving the better recovery in this work was found to be consistent with previous work on QuEChERS sample preparation for determination of mycotoxins in milled rice (35).
In this work, the content of water in brown rice is <10% (w/w). Therefore, the purposes of addition of water before extraction are to hydrate and swell the matrices in the sample and to reduce the interaction between analytes and matrices, leading to a good extraction efficiency (8) similar to previous work on QuEChERS sample preparation of various sample with low water content (8, 35).
It should be noted that in previous work on the determination of mycotoxins, a reduced sample size was also reported to be optimal for spices, such as red chili and white and black pepper (1 g with 5 mL ACN) (8), nuts and seeds (2 g with 5 mL ACN and 5% v/v FA) (26) and snacks (2 g with 10 mL ACN) (7). In this work, a reduced sample size of 1-g brown rice with 5 mL of ACN was found to be suitable for QuEChERS extraction, while a sample size of 10 g with 10 mL ACN was used for the QuEChERS extraction of milled rice (35) due to the presence of less complex matrices in comparison with brown rice.
From the results in Figure 3, the C18 sorbent was found to give recoveries in the range of 80–99% for all mycotoxins (Figure 3), but three (FB1, FB2 and OTA), three (AFB1, AFB2 and OTA) and two (AFB1 and AFG1) out of the nine mycotoxins (data not shown) were recovered at <80% when using PSA, GCB and silica, respectively. In particular, recovery values of 27 ± 3% for FB1, 30 ± 2% for FB2 and 69 ± 2% for OTA were found with the PSA sorbent, and 59 ± 2% was found for OTA with the GCB sorbent. However, it was noticed that, after the d-SPE clean-up step, the C18 sorbent gave a darker brown extract than did the other three sorbents, implying a higher amount of pigment remained in the extract. In comparison with the single sorbent and other mixed sorbents in Figure 3, Sorbent V was found to provide recoveries within an acceptable range, while three (FB1, FB2 and OTA), three (AFB1, AFB2 and OTA), three (AFB1, AFG1 and ZON), four (FB1, FB2, OTA and ZON) and six (AFB1, AFB2, AFG1, FB1, OTA and ZON) out of the nine mycotoxins were recovered at <80% with Sorbents I, II, III, IV and VI, respectively. Therefore, Sorbent V was used for the d-SPE step.
Method validation
Analytical limits
As shown in Table II, this method allows for the determination of trace levels of mycotoxins with MDL and MQL values in the range of 1.4–25 and 4.1–75 µg/kg, respectively, which were all below the ML of EU regulation except for OTA with an MQL of 7.5 µg/kg that was slightly higher than the ML of 5.0 µg/kg.
Matrix effect and linearity
The acceptable criteria of %Δ for each analyte are within ± 20% (8, 28) for all analytes. Using the t-test with one-tail t-coefficient at a 95% confidence level with six degrees of freedom (N1 + N2-4) for statistical comparison of the two analytical methods, a higher texp for almost all of the analytes than the critical t value of 2.4 was obtained, except for ZON, which indicated a significant difference in the slope between the two calibration methods. Both the %Δ and t-test results in this work imply that quantitative analysis should be performed using the matrix-matched calibration method.
From the results in Table III, the observed Fexp value of less than the critical F(0.05,13,10) value of 2.9 with five triplicate concentration levels indicated a good linearity of the calibration plot of each analyte, with r2 > 0.996. Several previous works have also reported the use of a matrix-matched calibration method (7, 8, 28, 31, 32). It should be noted that the recovery values reported in the previous QuEChERS optimization section are obtained from the matrix-matched calibration method as well.
Accuracy and precision
For each day analysis of each analyte, average recoveries in Table IV were obtained with satisfactory accuracy with values in the range of 81–101% for analyte concentrations of 5.0–1,000 µg/kg. Furthermore, all of the recovery data fell within the criteria for acceptable recovery (39) for the respective analyte concentrations. The recovery RSD value for intraday precision was obtained using 10 batches from the QuEChERS sample preparation method, while that for the interday precision was evaluated using three days. The statistical analysis of variance (ANOVA) with single-factor analysis at a 95% confidence level, revealed no significant differences between the precision obtained on each different day (P-value > 0.05) for each mycotoxin (Table IV). Therefore, the overall % RSD for intraday precision was calculated using a single data set (n = 90 for each analyte), and the % RSD for interday precision was equal to that for intraday precision. Satisfactory RSD values for inter- and intraday precision (overall RSD) were obtained, falling in the range of 5–19% for analyte concentrations of 5.0–1,000 µg/kg, with all of the RSD data falling within an acceptable range with values of 11–23% (39).
Application to real samples
From the results in Table V, six out of fourteen samples were found to be contaminated with at least one mycotoxin. FB1, FB2 and ZON were detected at 2.49–5.41 µg/kg (four samples), 4.33 ± 0.04 µg/kg (one sample) and 6.10–14.88 µg/kg (three samples), which are all lower than the ML established by EU regulations.
Conclusion
QuEChERS sample preparation was optimized and validated for the rapid and simultaneous UHPLC–MS-MS determination of nine mycotoxins in brown rice. Solvent extraction of brown rice (1.0 g) was performed with 10% (v/v) HOAc-containing ACN (5.0 mL) in the presence of four salts (2.0 g anh.MgSO4, 0.50 g NaCl, 0.50 g sodium citrate tribasic dihydrate and 0.25 g sodium citrate dibasic sesquihydrate) and with d-SPE clean-up using a mixture of three sorbents (50 mg C18, 25 mg PSA and 25 mg silica). This developed method provided an MDL of 1.4–25 µg/kg and an MQL of 4.1–75 µg/kg; a matrix-matched IS calibration method with good linearity using the F-test indicator, and acceptable accuracy and precision with a recovery of 81–101% and RSD of 5–19% for 5.0–1,000 µg/kg, respectively. In addition, 6 out of 14 real samples of brown rice were found to be contaminated with at least one mycotoxin. Therefore, given an MDL below the maximum limits of EU regulation, as well as acceptable accuracy and precision, this validated QuEChERS sample preparation method can be used as an alternative sample preparation method for the UHPLC–MS-MS determination of nine mycotoxins in brown rice.
Acknowledgments
The authors gratefully acknowledge the financial support provided by the 90th Anniversary of the Chulalongkorn University Fund (Ratchadaphiseksomphot Endowment Fund), Chulalongkorn University (GCUGR11255725085M).
References
- 1.Paiva F.F., Vanier N.L., Berrios J.D.J., Pan J., Villanova F.d.A., Takeoka G. et al. ; Physicochemical and nutritional properties of pigmented rice subjected to different degrees of milling; Journal of Food Composition and Analysis, (2014); 35(1): 10–17. [Google Scholar]
- 2.Pereira-Caro G., Cros G., Yokota T., Crozier A.; Phytochemical profiles of black, red, brown, and white rice from the Camargue Region of France; Journal of Agricultural and Food Chemistry, (2013); 61(33): 7976–7986. [DOI] [PubMed] [Google Scholar]
- 3.Moongngarm A., Daomukda N., Khumpika S.; Chemical compositions, phytochemicals, and antioxidant capacity of rice bran, rice bran layer, and rice germ; APCBEE Procedia, (2012); 2(0): 73–79. [Google Scholar]
- 4.Kumar V., Basu M.S., Rajendran T.P.; Mycotoxin research and mycoflora in some commercially important agricultural commodities; Crop Protection, (2008); 27(6): 891–905. [Google Scholar]
- 5.Tang Y., Lin H., Chen Y., Su W., Wang S., Chiueh L. et al. ; Development of a quantitative multi-mycotoxin method in rice, maize, wheat and Peanut Using UPLC-MS/MS; Food Analytical Methods, (2013); 6(3): 727–736. [Google Scholar]
- 6.Richard J.L.; Some major mycotoxins and their mycotoxicoses—an overview; International Journal of Food Microbiology, (2007); 119(1–2): 3–10. [DOI] [PubMed] [Google Scholar]
- 7.Rubert J., Fapohunda S.O., Soler C., Ezekiel C.N., Mañes J., Kayode F.; A survey of mycotoxins in random street-vended snacks from Lagos, Nigeria, using QuEChERS-HPLC-MS/MS; Food Control, (2013); 32(2): 673–677. [Google Scholar]
- 8.Yogendrarajah P., Van Poucke C., De Meulenaer B., De Saeger S.; Development and validation of a QuEChERS based liquid chromatography tandem mass spectrometry method for the determination of multiple mycotoxins in spices; Journal of Chromatography A, (2013); 1297(0): 1–11. [DOI] [PubMed] [Google Scholar]
- 9.Paíga P., Morais S., Oliva-Teles T., Correia M., Delerue-Matos C., Duarte S.C. et al. ; Extraction of ochratoxin A in bread samples by the QuEChERS methodology; Food Chemistry, (2012); 135(4): 2522–2528. [DOI] [PubMed] [Google Scholar]
- 10.Capriotti A.L., Cavaliere C., Piovesana S., Samperi R., Laganà A.; Multiclass screening method based on solvent extraction and liquid chromatography–tandem mass spectrometry for the determination of antimicrobials and mycotoxins in egg; Journal of Chromatography A, (2012); 1268(0): 84–90. [DOI] [PubMed] [Google Scholar]
- 11.Sospedra I., Blesa J., Soriano J.M., Mañes J.; Use of the modified quick easy cheap effective rugged and safe sample preparation approach for the simultaneous analysis of type A- and B-trichothecenes in wheat flour; Journal of Chromatography A, (2010); 1217(9): 1437–1440. [DOI] [PubMed] [Google Scholar]
- 12.Al-Hazmi N.A.; Determination of Patulin and Ochratoxin A using HPLC in apple juice samples in Saudi Arabia; Saudi Journal of Biological Sciences, (2010); 17(4): 353–359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Azaiez I., Giusti F., Sagratini G., Mañes J., Fernández-Franzón M.; Multi-mycotoxins analysis in dried fruit by LC/MS/MS and a modified QuEChERS procedure; Food Analytical Methods, (2014); 7(4): 935–945. [Google Scholar]
- 14.Pleadin J., Sokolović M., Perši N., Zadravec M., Jaki V., Vulić A.; Contamination of maize with deoxynivalenol and zearalenone in Croatia; Food Control, (2012); 28(1): 94–98. [Google Scholar]
- 15.Li C., Wu Y.L., Yang T., Huang-Fu W.G.; Rapid determination of fumonisins B1 and B2 in corn by liquid chromatography-tandem mass spectrometry with ultrasonic extraction; Journal of Chromatographic Science, (2012); 50(1): 57–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Dohnal V., Ježková A., Polišenská I., Kuca K.; Determination of fumonisins in milled corn grains using HPLC-MS; Journal of Chromatographic Science, (2010); 48(8): 680–684. [DOI] [PubMed] [Google Scholar]
- 17.Jiménez M., Mateo R.; Determination of mycotoxins produced by Fusarium isolates from banana fruits by capillary gas chromatography and high-performance liquid chromatography; Journal of Chromatography A, (1997); 778(1–2): 363–372. [DOI] [PubMed] [Google Scholar]
- 18.Wu R.-N., Dang Y.-L., Niu L., Hu H.; Application of matrix solid-phase dispersion–HPLC method to determine patulin in apple and apple juice concentrate; Journal of Food Composition and Analysis, (2008); 21(7): 582–586. [Google Scholar]
- 19.Bertuzzi T., Rastelli S., Mulazzi A., Donadini G., Pietri A.; Mycotoxin occurrence in beer produced in several European countries; Food Control, (2011); 22(12): 2059–2064. [Google Scholar]
- 20.Vidal A., Marín S., Ramos A.J., Cano-Sancho G., Sanchis V.; Determination of aflatoxins, deoxynivalenol, ochratoxin A and zearalenone in wheat and oat based bran supplements sold in the Spanish market; Food and Chemical Toxicology, (2013); 53(0): 133–138. [DOI] [PubMed] [Google Scholar]
- 21.Marina Martins H., Almeida I., Marques M.F., Guerra M.M.; Fumonisins and deoxynivalenol in corn-based food products in Portugal; Food and Chemical Toxicology, (2008); 46(7): 2585–2587. [DOI] [PubMed] [Google Scholar]
- 22.Soleimany F., Jinap S., Abas F.; Determination of mycotoxins in cereals by liquid chromatography tandem mass spectrometry; Food Chemistry, (2012); 130(4): 1055–1060. [Google Scholar]
- 23.Soleimany F., Jinap S., Faridah A., Khatib A.; A UPLC–MS/MS for simultaneous determination of aflatoxins, ochratoxin A, zearalenone, DON, fumonisins, T-2 toxin and HT-2 toxin, in cereals; Food Control, (2012); 25(2): 647–653. [Google Scholar]
- 24.Malysheva S.V., Diana Di Mavungu J., Boonen J., De Spiegeleer B., Goryacheva I.Y., Vanhaecke L. et al. ; Improved positive electrospray ionization of patulin by adduct formation: usefulness in liquid chromatography–tandem mass spectrometry multi-mycotoxin analysis; Journal of Chromatography A, (2012); 1270(0): 334–339. [DOI] [PubMed] [Google Scholar]
- 25.Berthiller F., Schuhmacher R., Buttinger G., Krska R.; Rapid simultaneous determination of major type A- and B-trichothecenes as well as zearalenone in maize by high performance liquid chromatography–tandem mass spectrometry; Journal of Chromatography A, (2005); 1062(2): 209–216. [DOI] [PubMed] [Google Scholar]
- 26.Arroyo-Manzanares N., Huertas-Pérez J.F., Gámiz-Gracia L., García-Campaña A.M.; A new approach in sample treatment combined with UHPLC-MS/MS for the determination of multiclass mycotoxins in edible nuts and seeds; Talanta, (2013); 115(0): 61–67. [DOI] [PubMed] [Google Scholar]
- 27.Arroyo-Manzanares N., García-Campaña A.M., Gámiz-Gracia L.; Multiclass mycotoxin analysis in Silybum marianum by ultra high performance liquid chromatography–tandem mass spectrometry using a procedure based on QuEChERS and dispersive liquid–liquid microextraction; Journal of Chromatography A, (2013); 1282(0): 11–19. [DOI] [PubMed] [Google Scholar]
- 28.Frenich A.G., Romero-González R., Gómez-Pérez M.L., Vidal J.L.M.; Multi-mycotoxin analysis in eggs using a QuEChERS-based extraction procedure and ultra-high-pressure liquid chromatography coupled to triple quadrupole mass spectrometry; Journal of Chromatography A, (2011); 1218(28): 4349–4356. [DOI] [PubMed] [Google Scholar]
- 29.Juan C., Ritieni A., Mañes J.; Determination of trichothecenes and zearalenones in grain cereal, flour and bread by liquid chromatography tandem mass spectrometry; Food Chemistry, (2012); 134(4): 2389–2397. [DOI] [PubMed] [Google Scholar]
- 30.Fekete S., Schappler J., Veuthey J.-L., Guillarme D.; Current and future trends in UHPLC; TrAC Trends in Analytical Chemistry, (2014); 63(0): 2–13. [Google Scholar]
- 31.Ferreira I., Fernandes J.O., Cunha S.C.; Optimization and validation of a method based in a QuEChERS procedure and gas chromatography–mass spectrometry for the determination of multi-mycotoxins in popcorn; Food Control, (2012); 27(1): 188–193. [Google Scholar]
- 32.Rodríguez-Carrasco Y., Berrada H., Font G., Mañes J.; Multi-mycotoxin analysis in wheat semolina using an acetonitrile-based extraction procedure and gas chromatography–tandem mass spectrometry; Journal of Chromatography A, (2012); 1270(0): 28–40. [DOI] [PubMed] [Google Scholar]
- 33.Rubert J., Dzuman Z., Vaclavikova M., Zachariasova M., Soler C., Hajslova J.; Analysis of mycotoxins in barley using ultra high liquid chromatography high resolution mass spectrometry: Comparison of efficiency and efficacy of different extraction procedures; Talanta, (2012); 99(0): 712–719. [DOI] [PubMed] [Google Scholar]
- 34.Fernandes P.J., Barros N., Câmara J.S.; A survey of the occurrence of ochratoxin A in Madeira wines based on a modified QuEChERS extraction procedure combined with liquid chromatography–triple quadrupole tandem mass spectrometry; Food Research International, (2013); 54(1): 293–301. [Google Scholar]
- 35.Koesukwiwat U., Sanguankaew K., Leepipatpiboon N.; Evaluation of a modified QuEChERS method for analysis of mycotoxins in rice; Food Chemistry, (2014); 153(0): 44–51. [DOI] [PubMed] [Google Scholar]
- 36.Olivieri A.C.; Practical guidelines for reporting results in single- and multi-component analytical calibration: a tutorial; Analytica Chimica Acta, (2015); 868(0): 10–22. [DOI] [PubMed] [Google Scholar]
- 37.Andrade J.M., Estévez-Pérez M.G.; Statistical comparison of the slopes of two regression lines: a tutorial; Analytica Chimica Acta, (2014); 838(0): 1–12. [DOI] [PubMed] [Google Scholar]
- 38.Danzer K., Currie L.A.; Guidelines for Calibration in Analytical Chemistry. Part 1. Fundamentals and Single Component Calibration; Pure and Applied Chemistry, (1998); 70(4): 993–1014. [Google Scholar]
- 39.Gustavo González A., Ángeles Herrador M.; A practical guide to analytical method validation, including measurement uncertainty and accuracy profiles; TrAC Trends in Analytical Chemistry, (2007); 26(3): 227–238. [Google Scholar]