- Email: [email protected]
MS method for the determination of BADGE-related and BFDGE-related compounds in canned fish food samples based on the formation of [M + NH4]+ aducts
Food Chemistry 135 (2012) 1310–1315
Contents lists available at SciVerse ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
A LC-MS/MS method for the determination of BADGE-related and BFDGE-related compounds in canned fish food samples based on the formation of [M + NH4]+ aducts J. Míguez a,⇑, C. Herrero b, I. Quintás a, C. Rodríguez a, P.G. Gigosos a, O.C. Mariz a a b
Laboratorio de Salud Pública de Lugo, Consellería de Sanidad, Xunta de Galicia, Rua Montevideo 9, 27001 Lugo, Spain Dpto. Química Analítica, Nutrición y Bromatología, Facultad de Ciencias, Universidad de Santiago de Compostela, Augas Férreas s/n, 27002 Lugo, Spain
a r t i c l e
i n f o
Article history: Received 3 November 2011 Received in revised form 17 April 2012 Accepted 25 May 2012 Available online 7 June 2012 Keywords: Bisphenol A diglycidyl ether Bisphenol F diglycidyl ether and derivatives LC-MS/MS analysis Canned fish
a b s t r a c t A new and simple liquid chromatography tandem mass-spectrometry method for the determination of different bisphenol A (BPA) derivatives such as bisphenol A diglycidyl ether (BADGE), bisphenol F diglycidyl ether (BFDGE) and their reaction products with water and hydrochloric acid in different fish food products was developed. The extraction procedure and the chromatographic conditions were optimised for complex food matrices such as fish products. Food samples were homogenised and extracted with a 1:1 solution of acetonitrile-hexane, the solvent was eliminated in a N2 stream and the extract was reconstituted with 0.5 mL of a 0.01 M solution of ammonium formate. The sample solution obtained was directly measured by LC–MS/MS without any further purification under the developed conditions. The use of a mobile phase composed by ammonium formate–methanol in a binary gradient mode produced [M + NH4]+ aducts for the different BADGEs and BFDGEs. These aduct’s fragmentations were employed for the LC–MS/MS quantification of BPA derivatives in canned fish samples. The results of the validation were appropriate: the method was linear for BADGE and its hydrolysed derivatives up to 1000 lg kg 1, for the remaining compounds linearity achieved up to 100 lg kg 1. Quantification limits were in the range 2– 10 lg kg 1. RSD (intra and inter-day) was 6–12% and the recovery was comprised between 89% and 109%. Under the optimised conditions, the chromatographic separation was performed in 8 min per sample. The method was applied to the determination of BADGE, BFDGE and their reaction products in different samples of canned fish from Spanish origin. Migration results obtained were in compliance with the EU regulations. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Epoxy resins, prepolymers that contain two or more epoxyde groups per molecule, are the most employed material for the inner layer in coated cans, large storage envelopes and various types of food packages or containers. Bisphenol A (BPA) derivatives such as bisphenol A diglycidyl ether (BADGE) and bisphenol F diglycidyl ether (BFDGE) are commonly used in internal coatings of food packaging materials, both as a monomer for epoxy resins and epoxybased polymers. Nowadays it is well known that both compounds (as well as their aqueous and hydrochloric derivatives) can produce a genotoxic effect in the consumers. By means of the enzymatic hydrolysis of their epoxy groups, the corresponding hydrolysed derivates (for instance BADGEH2O, BADGE2H2O, BADGEHCl, BADGE2HCl, BADGEHClH2O, BFDGE2H2O and BFDGE2HCl) can be produced as indicated in Fig. 1. Epoxy compounds were reported
⇑ Corresponding author. Tel.: +34 982 292120; fax: +34 982 292004. E-mail address: [email protected] (J. Míguez). 0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2012.05.099
as potential alkylating agents with possible specific cytotoxic actions in tissues affecting cell division. The toxicity of these compounds depends mainly upon fractional concentration of unreacted epoxy groups. Moreover, the chlorohydroxy-derivatives were considered potentially toxic as potential carcinogens because of their structural analogy to the genotoxic monochloropropanediol and other chloropropanols (Satoh, Ohyama, Aoki, Iida, & Nagai, 2004). Due to their toxicity, the European Union has established legislation concerning the specific migration limits (SML) for BADGE and related compounds in foods (Commission Directive 16/EC, 2002; Commission Regulation EC 1895, 2005). Regulation specifies a maximum SML of 9 mg kg 1 for the sum of BADGE and their hydrolysed derivatives (BADGEH2O, BADGE2H2O and BADGEHClH2O) as well as 1 mg kg 1 for the sum of BADGE and their hydrochloric derivatives (BADGEHCl, BADGE2HCl and BADGEHClH2O). The presence of BFDGE has been prohibited since 2005. More studies concerning its toxicological consequences are needed. Different methods were described in the literature for the analysis of BADGE, BFDGE and their derivatives in different foods by
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
1311
Fig. 1. Chemical structures of BPA, BADGE, BFDGE and its derivatives.
chromatographic techniques (Leepipatpiboon, Sae-Khow, & Jayanta, 2005; Yonekubo, Hayakawa, & Sajiki, 2010; Gallart-Ayala, Moyano, & Galcerán, 2010). Several of the published methods used reserved-phase high-performance liquid chromatography with fluorescence detection (RP-HPLC-FD) on C8, C12 or C18 columns using a mixture of acetonitrile–water as mobile phase (Cabodo et al., 2008; Paseiro-Losada, Paz-Abuín, Vázquez-Odériz, SimalLozano, & Simal Gándara, 1991; Paseiro-Losada, Pérez-Lamela, López-Fabal, Sanmatín-Fenollera, & Simal-Lozano, 1997; Petersen, Schaefer, Buckow, Simat, & Steinhart, 2003; Philo, Jickells, Damat, & Castle, 1994; Simal-Gándara, Paz-Abuín, Paseiro-Losada, LópezMahía, & Simal-Lozano, 1992). However, other papers published (Berger & Oehme, 2000; Berger, Oehme, & Girardin, 2001; Cao et al., 2009; Uematsu, Hirata, Suzuki, Iida, & Saito, 2001; Zhang, Xue, Zou, & Dai, 2010) used mass spectrometry as detector due to their sensibility, selective detection and confirming capability. In a recent paper (Gallart-Ayala, Moyano, & Galcerán, 2011) an LC–MS method for the determination of BADGE, BFDGE and their derivatives using [M + NH4]+ aducts as precursor ions was developed. The method presented appropriate analytical characteristics and it was applied to the determination of the analytes in canned vegetables and soft-drink beverages. Spain is the fourth canned fish producer in the world only surpassed by Thailand, China and the United States of America. Galicia is a region located in the NW of Spain characterised by the production of high quality food commodities (such as wine and alcoholic beverages, honey, meat, milk and others). Galicia concentrates almost the 80% of canned fish manufactured in Spain destined to exportation to the European Union and the USA. The objective of this work was to develop an LC–MS/MS method with good sensitivity based on the fragmentation of [M + NH4]+ aducts for quantification and confirmation of BADGE, BFDGE and their hydrolysed and hydrochloric products. The extraction method and the chromatographic conditions were optimised in order to determine BADGE and BFDGE related compounds in a particular food matrix (with high fat content in certain cases) such as fish
products. The described method was applied to measure concentration levels of the above cited analytes in canned fish samples from Spain. In all analysed cases, the results obtained for BADGE, BFDGE and their derivatives were in compliance with European Union regulations. 2. Experimental 2.1. Reagents Bisphenol F diglycidyl ether (BFDGE), bisphenol A diglycidyl ether (BADGE), bisphenol A [2, 3 dihydroxypropyl] glycidyl ether (BADGEH2O), bisphenol A bis [2, 3 dihydroxypropyl] ether (BADGE2H2O), bisphenol A [3 chloro-2 hydroxypropyl] glycidyl ether (BADGEHCl), bisphenol A bis [3 chloro-2 hydroxypropyl] ether (BADGE2HCl), bisphenol A [3 chloro-2 hydroxypropyl] [2, 3 dihydroxipropil] ether (BADGEHClH2O), bisphenol F bis [2, 3 dihydroxypropyl] ether (BFDGE2H2O) and bisphenol F bis [3 chloro-2 hydroxypropyl] ether (BFDGE2HCl) were purchased from Sigma, Germany. Acetonitrile and methanol were HPLC grade from Merck, Germany. Hexane employed for food sample extraction was analytical grade, and ammonium formate was grade mass spectroscopy, both supplied by Merck, Germany. High-purity water used was obtained from a Milli-Q Water system (Millipore, Massachusetts, USA). Samples were filtered through 0.45 lm PVDF hydrophilic nylon filters provided by Millipore, Italy. 2.2. Apparatus An Applied Biosystem QTrap 3200 LC–MS/MS System (Applied Biosystems, Canada) equipped with a TurboIonSpray electrospray ionisation (ESI) source and an Agilent 1200 HPLC system (Agilent, Germany) was employed. Separation was performed at 40 °C on
1312
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
a reverse-phase column Synergy MAX-RP, 100 2.0 mm. i.d., 2.5 lm particle size (Phenomenex, USA). Mass spectra were acquired and treated with the software Analyst 1.5 from Applied Biosystems. Elution was carried out in binary gradient mode. Other equipments used during the extraction procedure were an horizontal shaker OVAN-HL30E, a centrifuge Ependorf 5830R and a nitrogen stream evaporator Zymark Turbo Vap LV. 2.3. Chromatographic conditions The mobile phase was prepared by mixing 0.01 M ammonium formate solution (A) and methanol (B). Before analysis, the column was conditioned for 10 min with 0.01 M ammonium formate– methanol 25:75 to obtain a stable baseline. The binary gradient program employed was the following: for time 0–1 min, the mobile phase was composed of a mixture 25:75 of A:B; between 1 and 4 min, the proportion of B was increased linearly to 95%, to achieve a 5:95 of A:B; for 4–5 min, the mobile phase 5:95 of A:B was maintained; and for time 5–8 min, a column re-equilibration to the initial composition of 25:75 of A:B was performed. Therefore, the total time for the gradient program was 8 min per sample. The flow rate of the mobile phase was 0.2 mL min 1.
and commercial areas by public health inspectors. Due to the specific characteristics of the fish matrices with very high fat content in certain cases (Tuna and mackerel were fish with high fat levels; their values are in the range 20–24% depending on of the species and season; pilchard and anchovies present minor fat amounts, in the range of 6–12%, also according to species and season and mussels have low fat content, approximately 2%), the extraction process was optimised using a 1:1 acetonitrile:n-hexane solution as extractant solvent. This mixture was demonstrated to be useful for achieving the quantitative extraction of the analytes in these food products. The extraction process was performed in shorter that when using other solvents (ethyl acetate) in single matrices such as vegetables and soft-drinks (Gallart-Ayala et al., 2011). Each sample was prepared as follows: the whole content was homogenised including either the oil or water solution in which the sample was in. Two grams of the sample were extracted with a solution of 10 mL acetonitrile and 10 mL n-hexane. The solution obtained was stirred during 10 min in a horizontal shaker at 200 rpm followed by a centrifugation at 1500 rpm for 5 min. After 10 min, 1 mL of acetonitrile layer was evaporated using a nitrogen stream. The extract was reconstituted with 500 lL of mobile phase A and filtered using 0.45 lm PVDF hydrophilic nylon filters into HPLC vials for analysis.
2.4. Mass spectrometry Mass spectrometry data were acquired and processed using the software Analyst 1.5 from Applied Biosystems. Analytes were detected by multiple reaction monitoring mode (MRM) using electrospray ionisation (positive polarity) in a single chromatographic run per sample. Two fragmentation reactions were scanned for each analyte. The MS source parameters were the following: source temperature 450 °C, curtain gas 25 psi, nebulizer gas 60 psi, turbo gas 40 psi, ion spray voltage 5000 V, and collision gas medium. The gas used for the mass spectrometer was 99% pure nitrogen. The optimisation of the analyte dependent MS/MS parameters was performed via direct infusion of standards into the MS at a flow rate of 20 lL min 1. The standard solution were prepared in a 50:50 A:B solution (where A and B was the solvents described in the previous section of chromatographic conditions). Flow injection analysis (FIA) mode was also carried out for each compound in order to optimise the MS source conditions. 2.5. Standard solutions and spiked samples Stock solutions for each standard were prepared at a level of 500 mg L 1 in acetonitrile and they were stored in the darkness at 4 °C. Spiked solutions were prepared by diluting the stock solutions with acetonitrile. The resulting concentrations were 5 lg mL 1 for BADGE, BADGEH2O and BADGE2H2O, and 0.5 lg mL 1 for BADGEHCl, BADGE2HCl, BADGEHClH2O, BFDGE, BFDGE2HCl and BFDGEH2O. In order to take into account the influence of the fish matrix in calibration, in the present case the calibration was performed with spiked blank samples obtained by adding different quantities (comprised between 20 and 400 ll) of spiked solution described above to fish blank samples. The resulting concentrations of calibration curves obtained were 0, 5, 10, 25, 50, 75 and 100 lg kg 1 for all analytes, except BADGE, BADGEH2O and BADGE2H2O whose concentrations were 0, 50, 100, 250, 500, 750 and 1000 lg kg 1. 2.6. Samples and sample preparation Forty-one samples of different canned fish food samples in oil and aqueous media, such as mussels (8), boquerons (11), sardines (4), mackerels (4) and tuna (14), were obtained from supermarkets
2.7. Method validation procedures External calibration method was used for quantitative analysis, calibration curves for each analyte were obtained by plotting area peak versus concentration of the corresponding standard in the range 0–100 lg kg 1 for all analytes, except for BADGE, BADGEH2O and BADGE2H2O in the range 0–1000 lg kg 1. Calibration was performed with spiked blank samples in order to take into account the effect matrix. Concentration of the analytes in the fish samples was calculated on the basis of the regression line obtained from calibration curves. Detection limit, defined as the lowest concentration of analyte in a defined matrix that produced a positive identification using a specified method, was calculated as 3 SD/m, where SD is the standard deviation of 10 measurements of blank solutions and m is the slope of the calibration curve (equivalent to S/N ratio equal to 3). Quantification limit, the lowest concentration of analyte in a defined matrix that achieves a quantitative measurement a specified method, was determined on the basis of 10 SD/m (equivalent to S/N ratio equal to 10). Intra-day and inter-day studies were performed in order to test the reproducibility of the analytical method. Intra-day (n = 6) and inter-day (n = 6) precisions were obtained by the analysis of samples spiked at three concentration levels in the same and in different days (3). For BADGEHCl, BADGE2HCl, BADGEHClH2O, BFDGE, BFDGE2HCl and BFDGEH2O, the fish samples were spiked with 10, 50 and 100 lg kg 1 of standards. For BADGE, BADGE.H2O and BADGE.2H2O samples were spiked with 100, 500 and 1000 lg kg 1 of standards. Since no certified reference material for BDGE, BFDGE and derivatives in canned fish are available, the accuracy of the method was estimated by the recovery of different spiked samples with standard mixtures at the three concentration levels described above. 3. Results and discussion 3.1. Optimisation of chromatographic separation Different experiments were carried out in order to select the adequate mobile phase. After multiple experiments, a mobile phase composed of a solution of 0.01 M ammonium formate (A) and methanol (B) was selected as optimum. The use of this mobile phase produced the best response for the parent and product ions
1313
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
selected in ionisation mode. Different assays which used mobile phases containing formic acid produced lower signal than the one obtained using the selected ammonium formate–methanol, probably this fact is due to ionisation inhibition. Other authors (Zhang et al., 2010) indicated that more than 0.1% of formic acid is necessary to help the ionisation process, and that concentration higher than 0.3% produced an ionisation inhibition (optimum concentration in the range 0.2–0.3%). At the present case, the addition of different quantities of formic acid did not achieve the improvement of the signal, therefore its use was avoided. Other mobile phases evaluated were several solutions obtained from different proportions of acetonitrile–water; in these cases very low responses were obtained, probably due to the fact that quasi-moleculars clusters between the analyte and acetonitrile were not properly formed in ionisation mode (Pardo, Yusá, León, & Pastor, 2006). Therefore, the mobile phase selected for the optimisation of chromatographic separation was ammonium formate–methanol, which is similar to those used in other studies on separation, characterisation and quantification of BADGE, BFDGE and their derivatives (Gallart-Ayala et al., 2010, 2011; Pardo et al., 2006; Yonekubo, Hayakawa, & Sajiki, 2008; Zhang et al., 2010). The gradient program was optimised in order to obtain the suitable separation of analytes with the appropriate stability in the retention times. As can it be seen in Fig. 2, employing the binary gradient previously described in Section 2.3, an adequate separation of the BADGE, BFDGE and their derivatives was obtained in 8 min.
3.2. Ionisation mode and optimisation of MS–MS parameters Both electrospray ionisation (ESI) as well as atmospheric pressure chemical ionisation (APCI) were applied as ionisation modes in liquid chromatography-mass spectroscopy for the determination of BADGE, BFDGE and their derivatives (Gallart-Ayala et al., 2010, 2011; Zhang et al., 2010). Therefore, after different assays performed with ESI and APCI in both positive and negative modes, ESI ionisation in positive mode was selected, because it provided lower detection limits that APCI mode. These results are consistent with those published by Zhang et al., 2010; these authors demonstrated that APCI mode offered worse sensitivity than ESI for the determination of BADGE-related compounds, and especially for
Table 1 Summary of optimised ESI-MS/MS conditions for the determined analytes. MW: molecular weight. DP: declustering potential, CE: collision energy, CXP: collision cell potential. MW
Parent ion
Product ionsa
DP
CE
CXP
BADGE
340.42
358.2
BADGEH2O
358.43
376.2
BADGE2H2O
376.44
394.2
BADGEHCl
376.87
394.2
BADGEHClH2O
394.89
412.1
BFDGE
312.37
330.2
BFDGE2HCl
385.28
402.1
BFDGE2H2O
348.39
366.2
BADGE2HCl
413.33
430.1
191.2 135.1 209.3 191.1 209.2 135.3 227.1 135.2 227.3 135.2 163.3 133.2 199 181.3 133.3 181.4 227.2 135.1
26 26 26 26 21 21 16 16 26 26 16 16 31 31 16 16 21 21
19 41 19 27 23 47 21 47 21 47 17 21 19 31 27 17 21 49
4 4 4 4 4 4 4 4 6 4 4 4 4 4 4 6 4 4
a The first ion indicated for each analyte was employed for quantification purposes while the second ion was used for confirmation.
BADGEHClH2O. In addition, the linear ranges obtained for ESI mode were larger than when APCI were used. The optimisation of ESI conditions is crucial for an adequate MS determination for each analyte. Therefore, the study of different factors influencing the MS signal was carried out. The composition of the eluent has also an important influence in the ionisation effectiveness; consequently, a study over different eluents was performed. The infusion of standards was assayed in different solvents such as mixtures of methanol–water, mixtures of acetonitrile–water and several mixtures with different proportions A:B mobile phase. The best results obtained were achieved for a 50:50 mixture of A:B. The optimisation of the MS–MS parameters to maximise the signal of each analyte was performed. Declustering potencial (DP), collision energy (CE) and collision cell potencial (CXP) were optimised for each compound individually in infusion mode in the range 0–150 V, 0–100 V and 0–60 V, respectively. The optimisation of these parameters, as well as temperature and ion source
Fig. 2. Chromatogram of a canned fish sample spiked with BADGE-related and BFDGE-related compounds as indicated in Section 2.5.
1314
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
art-Ayala et al., 2010). Therefore, in the case at hand, the reproducible formation of ammonium adducts was favored by the use of mobile phase with ammonium formate. No other additives were necessary in order to allow the adduct formation and the efficient fragmentation with a stable signal under tandem mass spectrometry. Under these conditions, appropriate MRM chromatograms were obtained for quantitative ions of the different BADGE and BFDGE derivatives and, the transitions selected were adequate for the quantitative determination of the analytes.
Table 2 Recoveries obtained in canned fish samples at three addition levels. Added (lg kg
1
)
Recovery (%) 100
BADGE BADGEH2O BADGE2H2O
109.51 99.93 101.07
BADGEHCl BADGE2HCl BADGEHClH2O BFDGE BFDGE2H2O BFDGE2HCl
104.99 100.02 99.04 104.95 100.08 100.12
500
1000
96.09 93.84 99.98
10
91.91 90.52 95.23
50
100
97.36 96.04 96.11 94.49 101.96 95.93
94.56 94.84 95.28 89.92 100.21 92.54
3.3. Analytical figures of merit All the analytical figures of merit of the developed method were evaluated at the levels and under the conditions indicated in Method validation procedures section. Different calibrations were performed using fish spiked blank samples in order to take into account the influence of the matrix. The linearity between the area responses of the different target compounds and their concentrations were studied at levels of 0–100 lg kg 1 for all analytes, except for BADGE, BADGEH2O and BADGE2H2O in the range 0– 1000 lg kg 1. Appropriate relationships between known concentrations and measured areas were achieved under these conditions. Linear regression obtained indicated that the method is satisfactory for quantification, with correlation coefficients higher than 0.999, in all cases. Limits of detection (LOD = 3 SD/m) and quantification (LOQ = 10 SD/m) obtained for the different BADGE-related compounds are shown in Table 3. It can be observed that the range of detection limits are between 0.5 and 3.1 lg kg 1 and the quantification limit range is between 1.8 and 10.3 lg kg 1. These results are comparable to other published methods (Gallart-Ayala et al., 2011). Therefore, the present method attains enough sensitivity to determine BAGDE, BFGDE and their hydrolysed and hydrochloric derivatives at lower concentrations than the ones established by the European legislation (Commission Directive 16/EC, 2002; Commission Regulation EC 1895, 2005). Accuracy was studied by conducting recovery assays at three different concentration levels previously indicated. The values obtained are reported in Table 2. The use of calibration with spiked blank samples produced suitable recoveries in the range 89.9– 109.5% which are appropriate for analyte determinations. Reproducibility of the method was tested by intra-day and inter-day precision tests at the three concentration levels indicated above. The results are summarised in Table 3. The values obtained for repeatability (RSDr) and for inter-day precision (RSDR) are satisfactory in the range between 5.8% and 8.3% and between 6.7% and 12%, respectively. These values are satisfactory, although they are
Table 3 Summary of the analytical characteristics of the employed method. LOD: limit of detection, LOQ: limit of quantification, RSDr: Relative standard deviation of intra-day repeatability, RSDR: Relative standard deviation of inter-day precision. LOD (lg kg BADGE BADGEH2O BADGE2H2O BADGEHCl BADGE2HCl BADGEHClH2O BFDGE BFDGE2H2O BFDGE2HCl
3.0 3.1 2.3 2.1 0.8 0.5 2.0 2.1 2.4
1
)
LOQ (lg kg 10.1 10.3 7.2 7.1 2.7 1.8 5.7 6.2 7.3
1
)
RSDr (%)
RSDR (%)
8.0 7.7 8.3 7.8 9.5 7.8 7.6 5.8 7.8
12.0 8.6 9.0 8.5 10.0 8.5 12.0 6.7 9.0
gases, were executed in flow injection analysis by means of analyte standard solutions (1 lg ml 1) infused at a flow rate of 20 ll min 1. The optimum results are summarised in Table 1. Nitrogen was used as curtain gas, nebulizer gas and turbo gas. For ESI-positive mode MS–MS operation, the selection of the ions [M + H]+ and [M + NH4]+ were obtained by MS-scan mode and chosen as precursor ions. Product or daughter ions were selected by product-ion-scan. Quantification was carried out using multiple reaction monitoring mode (MRM) in order to improve selectivity and sensitivity. As indicated in Table 1, one of the product ions (MRM transition 1) was for quantitative purposes, while the other product ion (MRM transition 2) was employed for qualitative ones. The response obtained from the ions [M + H]+ was lower compared to that obtained from the ions [M + NH4]+. Both BADGEs and BFDGEs, under electrospray conditions, tend to form adducts with ammonium instead of the protonated molecule (Gall-
Table 4 Results obtained for the application of the developed method to different fish canned samples. Values for the different analytes detected in four of the analysed samples are in lg kg 1. Analyte
Number of samples analysed
Number of samples in which any analyte was detected
Number of samples with levels higher than permitted by EU Directive [2]
Mussel European anchovy Pilchard Mackerel Tuna Total
8 11
0 0
0 0
4 4 14 41
0 0 4 4
0 0 0 0
PS1-tuna PS2-tuna PS3-tuna PS4-tuna
BADGE
BADGEH2O
BADGE2H2O
BADGEHCl
BADGE2HCl
BADGEHClH2O
BFDGE
BFDGE2H2O
BFDGE2HCl
n.d. n.d. n.d. n.d.
120 n.d. n.d. n.d.
111 247 625 120
n.d. n.d. n.d. n.d.
n.d. n.d. n.d. n.d.
18 61 87 n.d.
n.d. n.d. n.d. n.d.
n.d. n.d. n.d. n.d.
n.d. n.d. n.d. n.d.
n.d.: not detected.
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
slightly higher than other published works in which reproducibility was tested in standards instead of fish samples, avoiding the matrix effect. Under the optimised conditions, the total time of analysis per sample, including sample homogenisation, extraction and chromatographic separation, was 60 min. 3.4. Analysis of samples The applicability of the method was demonstrated by utilising them for the routine analysis of canned fish samples from Spanish origin. 41 different fish samples including mussels (Mytilus edulis), European anchovy (Engraulis encrasicholus), pilchards (Sardina pilchardus), mackerels (Scomber scombrus) and tuna (Thunnus alalunga) were analysed using the method described above. In 90.2% of cases (37 samples) no BADGEs or BFDGEs were detected, in the four remaining samples, all of them of canned tuna, different levels of hydrolysed derivatives of BADGE were discovered (see Table 4). One sample presented measurable levels of BADGE.H2O, BADGE.2H2O and BADGE.HCl.H2O (111, 120 and 18 lg kg 1, respectively). In two other samples both BADGE.2H2O and BADGE.HCl.H2O were detected at the levels 247 and 61 as well as 625 and 87 lg kg 1, respectively. And in one additional sample a concentration level of 120 lg kg 1 of BADGE.2H2O was measured. However, the very low levels founded in these four samples did not achieve the maximum SML established for the EU regulations; therefore, it can be concluded that all the analysed samples are in compliance with EU Directives, demonstrating the safety of these canned fish products. 4. Conclusion A single and fast LC–MS/MS method that can be used for routine analysis of BADGE, BFDGE and their derivatives in canned fish has been developed. The method has been optimised taking into account the specific characteristics of food matrix products and it can be considered complementary of other LC–MS/MS methods using [M + NH4]+ aducts that were developed to determine these analytes in other food products with different matrices such as vegetables and drinks (Gallart-Ayala et al., 2011). The extraction of analytes from fish was performed by means of a simple and fast extraction with a 1:1 solution of acetonitrile:n-hexane. BADGE, BFDGE and their derivatives were measured on the basis of the fragmentation selected products from [M + NH4]+ aducts. The method was validated and it was demonstrated to be suitable, specific, accurate and sensitive enough for the quantification of BADGE, BFDGE and hydrolysed reaction products in order to verify the compliance with the EU Directives. The developed procedure was applied to determine the levels of BADGE, BFDGE and related compounds in different canned fish products from Spain. In all cases the concentrations determined are under the levels permitted by EU regulations concerning the specific migration of these compounds in contact with foodstuffs. Acknowledgements The authors wish to express their gratitude to Dirección Xeral of Innovación and Xestion of Saúde Pública, Consellería de Sanidade, Xunta de Galicia for their support. We also thank the collaboration of technical staff of Public Health Laboratory Lugo who participated in this work.
1315
References Berger, U., & Oehme, M. (2000). Identification of derivatives of bisphenol a diglycidyl ether and novolac glycidyl ether in can coatings by liquid chromatography/ion trap mass spectrometry. Journal of AOAC International, 83, 1367–1376. Berger, U., Oehme, M., & Girardin, L. (2001). Quantification of derivatives of bisphenol A diglycidyl ether (BADGE) and novolac glycidyl ether (NOGE) migrated from can coatings into tuna by HPLC/fluorescence and MS detection. Fresenius Journal of Analytical Chemistry, 369, 115–123. Cabodo, A. G., Aldea, S., Porro, C., Ojea, G., Lago, J., Sobrado, C., et al. (2008). Migration of BADGE (bisphenol A diglycidyl-ether) and BFDGE (bisphenol F diglycidyl-ether) in canned seafood. Food and Chemical Technology, 46, 1674–1680. Cao, X., Dufresne, G., Clement, G., Belisle, S., Robichaud, A., & Beraldin, F. (2009). Levels of bisphenol A diglycidyl ether (BADGE) and bisphenol F diglycidyl ether (BFDGE) in canned liquid infant formula products in Canada and dietary intake. Journal of AOAC International, 92, 1780–1789. Commission Directive 2002/16/EC of 20 February 2002 on the use of certain epoxy derivates in materials and articles intended to come into contact with foodstuffs (2002). Official Journal of the European Communities, L51/27. Gallart-Ayala, H., Moyano, E., & Galcerán, M. T. (2010). Multiple-stage mass spectrometry analysis of bisphenol A diglycidyl ether, bisphenol F diclycidyl ether and their derivatives. Rapid Communications in Mass Spectrometry, 24, 3469–3477. Gallart-Ayala, H., Moyano, E., & Galcerán, M. T. (2011). Fast liquid chromatography– tandem mass spectrometry for the analysis of bisphenol A-diglycidyl ether, bisphenol F-diglycidyl ether and its derivatives in canned food and beverages. Journal of Chromatography A, 1218, 1603–1610. Leepipatpiboon, N., Sae-Khow, O., & Jayanta, S. (2005). Simultaneous determination of bisphenol-A-diglycidyl ether, bisphenol-F-diglycidyl ether, and their derivatives in oil-in-water and aqueous-based canned foods by highperformance liquid chromatography with fluorescence detection. Journal of Chromatography A, 1073, 331–339. Pardo, O., Yusá, V., León, N., & Pastor, A. (2006). Determination of bisphenol diglycidyl ether residues in canned foods by pressurized liquid extraction and liquid chromatography–tandem mass spectrometry. Journal of Chromatography A, 1107, 70–78. Paseiro-Losada, P., Paz-Abuín, S., Vázquez Odériz, L., Simal-Lozano, J., & SimalGándara, J. (1991). Quality control of cured epoxy resins: determination of residual free monomers (m-xylylenediamine and bisphenol A diglycidyl ether) in the finished product. Journal of Chromatography A, 585, 75–81. Paseiro-Losada, P., Pérez-Lamela, C., López-Fabal, F., Sanmatín-Fenollera, P., & Simal-Lozano, J. (1997). Two RP-HPLC sensitive methods to quantify and identify bisphenol A diglycidyl ether and its hydrolysis products. 1. European Union aqueous food simulants. Journal of Agriculture and Food Chemistry, 45, 3493–3500. Petersen, H., Schaefer, A., Buckow, C. A., Simat, T. J., & Steinhart, H. (2003). Determination of bisphenol A diglycidyl ether (BADGE) and its derivatives in food: identification and quantification by internal standard. European Food Research and Technology, 216, 355–364. Philo, M. R., Jickells, S. M., Damat, A. P., & Castle, L. (1994). Stability of plastics monomers in food-simulating liquids under European Union migration test conditions. Journal of Agriculture and Food Chemistry, 42, 1497–1501. Commission Regulation (EC) No 1895/2005 of 18 de November 2005 on the restriction of use of certain epoxy derivates in materials and articles intended to come into contact with foodstuffs (2005). Official Journal of the European Communities, L302, 28–32. Satoh, K., Ohyama, K., Aoki, N., Iida, M., & Nagai, F. (2004). Study on anti-androgenic effects of bisphenol a diglycidyl ether (BADGE), bisphenol F diglycidyl ether (BFDGE) and their derivatives using cells stably transfected with human androgen receptor, AR-EcoScreen. Food and Chemical Toxicology, 42, 983–993. Simal-Gándara, J., Paz-Abuín, S., Paseiro-Losada, P., López-Mahía, P., & SimalLozano, J. (1992). Identification of RP-HPLC peaks of bisphenol F and of bisphenol F diglycidyl ether and its hydrolysis products by thermospray mass spectrometry and gas chromatography/mass spectrometry. Cromatographia, 34, 67–72. Uematsu, Y., Hirata, K., Suzuki, K., Iida, K., & Saito, K. (2001). Chlorohydrins of bisphenol A diglycidyl ether (BADGE) and of bisphenol F diglycidyl ether (BFDGE) in canned foods and ready-to-drink coffees from the Japanese market. Food Additives & Contaminants, 18, 177–185. Yonekubo, J., Hayakawa, K., & Sajiki, J. (2008). Concentrations of bisphenol A, bisphenol A diglycidyl ether, and their derivatives in canned foods in japanese markets. Journal of Agriculture and Food Chemistry, 56, 2041–2047. Zhang, H., Xue, M., Zou, Y., & Dai, Z. (2010). Simultaneous determination of NOGErelated and BADGE-related compounds in canned food by ultra-performance liquid chromatography tandem mass spectrometry. Analytical and Bioanalytical Chemistry, 398, 3165–3174.
Contents lists available at SciVerse ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
A LC-MS/MS method for the determination of BADGE-related and BFDGE-related compounds in canned fish food samples based on the formation of [M + NH4]+ aducts J. Míguez a,⇑, C. Herrero b, I. Quintás a, C. Rodríguez a, P.G. Gigosos a, O.C. Mariz a a b
Laboratorio de Salud Pública de Lugo, Consellería de Sanidad, Xunta de Galicia, Rua Montevideo 9, 27001 Lugo, Spain Dpto. Química Analítica, Nutrición y Bromatología, Facultad de Ciencias, Universidad de Santiago de Compostela, Augas Férreas s/n, 27002 Lugo, Spain
a r t i c l e
i n f o
Article history: Received 3 November 2011 Received in revised form 17 April 2012 Accepted 25 May 2012 Available online 7 June 2012 Keywords: Bisphenol A diglycidyl ether Bisphenol F diglycidyl ether and derivatives LC-MS/MS analysis Canned fish
a b s t r a c t A new and simple liquid chromatography tandem mass-spectrometry method for the determination of different bisphenol A (BPA) derivatives such as bisphenol A diglycidyl ether (BADGE), bisphenol F diglycidyl ether (BFDGE) and their reaction products with water and hydrochloric acid in different fish food products was developed. The extraction procedure and the chromatographic conditions were optimised for complex food matrices such as fish products. Food samples were homogenised and extracted with a 1:1 solution of acetonitrile-hexane, the solvent was eliminated in a N2 stream and the extract was reconstituted with 0.5 mL of a 0.01 M solution of ammonium formate. The sample solution obtained was directly measured by LC–MS/MS without any further purification under the developed conditions. The use of a mobile phase composed by ammonium formate–methanol in a binary gradient mode produced [M + NH4]+ aducts for the different BADGEs and BFDGEs. These aduct’s fragmentations were employed for the LC–MS/MS quantification of BPA derivatives in canned fish samples. The results of the validation were appropriate: the method was linear for BADGE and its hydrolysed derivatives up to 1000 lg kg 1, for the remaining compounds linearity achieved up to 100 lg kg 1. Quantification limits were in the range 2– 10 lg kg 1. RSD (intra and inter-day) was 6–12% and the recovery was comprised between 89% and 109%. Under the optimised conditions, the chromatographic separation was performed in 8 min per sample. The method was applied to the determination of BADGE, BFDGE and their reaction products in different samples of canned fish from Spanish origin. Migration results obtained were in compliance with the EU regulations. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Epoxy resins, prepolymers that contain two or more epoxyde groups per molecule, are the most employed material for the inner layer in coated cans, large storage envelopes and various types of food packages or containers. Bisphenol A (BPA) derivatives such as bisphenol A diglycidyl ether (BADGE) and bisphenol F diglycidyl ether (BFDGE) are commonly used in internal coatings of food packaging materials, both as a monomer for epoxy resins and epoxybased polymers. Nowadays it is well known that both compounds (as well as their aqueous and hydrochloric derivatives) can produce a genotoxic effect in the consumers. By means of the enzymatic hydrolysis of their epoxy groups, the corresponding hydrolysed derivates (for instance BADGEH2O, BADGE2H2O, BADGEHCl, BADGE2HCl, BADGEHClH2O, BFDGE2H2O and BFDGE2HCl) can be produced as indicated in Fig. 1. Epoxy compounds were reported
⇑ Corresponding author. Tel.: +34 982 292120; fax: +34 982 292004. E-mail address: [email protected] (J. Míguez). 0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2012.05.099
as potential alkylating agents with possible specific cytotoxic actions in tissues affecting cell division. The toxicity of these compounds depends mainly upon fractional concentration of unreacted epoxy groups. Moreover, the chlorohydroxy-derivatives were considered potentially toxic as potential carcinogens because of their structural analogy to the genotoxic monochloropropanediol and other chloropropanols (Satoh, Ohyama, Aoki, Iida, & Nagai, 2004). Due to their toxicity, the European Union has established legislation concerning the specific migration limits (SML) for BADGE and related compounds in foods (Commission Directive 16/EC, 2002; Commission Regulation EC 1895, 2005). Regulation specifies a maximum SML of 9 mg kg 1 for the sum of BADGE and their hydrolysed derivatives (BADGEH2O, BADGE2H2O and BADGEHClH2O) as well as 1 mg kg 1 for the sum of BADGE and their hydrochloric derivatives (BADGEHCl, BADGE2HCl and BADGEHClH2O). The presence of BFDGE has been prohibited since 2005. More studies concerning its toxicological consequences are needed. Different methods were described in the literature for the analysis of BADGE, BFDGE and their derivatives in different foods by
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
1311
Fig. 1. Chemical structures of BPA, BADGE, BFDGE and its derivatives.
chromatographic techniques (Leepipatpiboon, Sae-Khow, & Jayanta, 2005; Yonekubo, Hayakawa, & Sajiki, 2010; Gallart-Ayala, Moyano, & Galcerán, 2010). Several of the published methods used reserved-phase high-performance liquid chromatography with fluorescence detection (RP-HPLC-FD) on C8, C12 or C18 columns using a mixture of acetonitrile–water as mobile phase (Cabodo et al., 2008; Paseiro-Losada, Paz-Abuín, Vázquez-Odériz, SimalLozano, & Simal Gándara, 1991; Paseiro-Losada, Pérez-Lamela, López-Fabal, Sanmatín-Fenollera, & Simal-Lozano, 1997; Petersen, Schaefer, Buckow, Simat, & Steinhart, 2003; Philo, Jickells, Damat, & Castle, 1994; Simal-Gándara, Paz-Abuín, Paseiro-Losada, LópezMahía, & Simal-Lozano, 1992). However, other papers published (Berger & Oehme, 2000; Berger, Oehme, & Girardin, 2001; Cao et al., 2009; Uematsu, Hirata, Suzuki, Iida, & Saito, 2001; Zhang, Xue, Zou, & Dai, 2010) used mass spectrometry as detector due to their sensibility, selective detection and confirming capability. In a recent paper (Gallart-Ayala, Moyano, & Galcerán, 2011) an LC–MS method for the determination of BADGE, BFDGE and their derivatives using [M + NH4]+ aducts as precursor ions was developed. The method presented appropriate analytical characteristics and it was applied to the determination of the analytes in canned vegetables and soft-drink beverages. Spain is the fourth canned fish producer in the world only surpassed by Thailand, China and the United States of America. Galicia is a region located in the NW of Spain characterised by the production of high quality food commodities (such as wine and alcoholic beverages, honey, meat, milk and others). Galicia concentrates almost the 80% of canned fish manufactured in Spain destined to exportation to the European Union and the USA. The objective of this work was to develop an LC–MS/MS method with good sensitivity based on the fragmentation of [M + NH4]+ aducts for quantification and confirmation of BADGE, BFDGE and their hydrolysed and hydrochloric products. The extraction method and the chromatographic conditions were optimised in order to determine BADGE and BFDGE related compounds in a particular food matrix (with high fat content in certain cases) such as fish
products. The described method was applied to measure concentration levels of the above cited analytes in canned fish samples from Spain. In all analysed cases, the results obtained for BADGE, BFDGE and their derivatives were in compliance with European Union regulations. 2. Experimental 2.1. Reagents Bisphenol F diglycidyl ether (BFDGE), bisphenol A diglycidyl ether (BADGE), bisphenol A [2, 3 dihydroxypropyl] glycidyl ether (BADGEH2O), bisphenol A bis [2, 3 dihydroxypropyl] ether (BADGE2H2O), bisphenol A [3 chloro-2 hydroxypropyl] glycidyl ether (BADGEHCl), bisphenol A bis [3 chloro-2 hydroxypropyl] ether (BADGE2HCl), bisphenol A [3 chloro-2 hydroxypropyl] [2, 3 dihydroxipropil] ether (BADGEHClH2O), bisphenol F bis [2, 3 dihydroxypropyl] ether (BFDGE2H2O) and bisphenol F bis [3 chloro-2 hydroxypropyl] ether (BFDGE2HCl) were purchased from Sigma, Germany. Acetonitrile and methanol were HPLC grade from Merck, Germany. Hexane employed for food sample extraction was analytical grade, and ammonium formate was grade mass spectroscopy, both supplied by Merck, Germany. High-purity water used was obtained from a Milli-Q Water system (Millipore, Massachusetts, USA). Samples were filtered through 0.45 lm PVDF hydrophilic nylon filters provided by Millipore, Italy. 2.2. Apparatus An Applied Biosystem QTrap 3200 LC–MS/MS System (Applied Biosystems, Canada) equipped with a TurboIonSpray electrospray ionisation (ESI) source and an Agilent 1200 HPLC system (Agilent, Germany) was employed. Separation was performed at 40 °C on
1312
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
a reverse-phase column Synergy MAX-RP, 100 2.0 mm. i.d., 2.5 lm particle size (Phenomenex, USA). Mass spectra were acquired and treated with the software Analyst 1.5 from Applied Biosystems. Elution was carried out in binary gradient mode. Other equipments used during the extraction procedure were an horizontal shaker OVAN-HL30E, a centrifuge Ependorf 5830R and a nitrogen stream evaporator Zymark Turbo Vap LV. 2.3. Chromatographic conditions The mobile phase was prepared by mixing 0.01 M ammonium formate solution (A) and methanol (B). Before analysis, the column was conditioned for 10 min with 0.01 M ammonium formate– methanol 25:75 to obtain a stable baseline. The binary gradient program employed was the following: for time 0–1 min, the mobile phase was composed of a mixture 25:75 of A:B; between 1 and 4 min, the proportion of B was increased linearly to 95%, to achieve a 5:95 of A:B; for 4–5 min, the mobile phase 5:95 of A:B was maintained; and for time 5–8 min, a column re-equilibration to the initial composition of 25:75 of A:B was performed. Therefore, the total time for the gradient program was 8 min per sample. The flow rate of the mobile phase was 0.2 mL min 1.
and commercial areas by public health inspectors. Due to the specific characteristics of the fish matrices with very high fat content in certain cases (Tuna and mackerel were fish with high fat levels; their values are in the range 20–24% depending on of the species and season; pilchard and anchovies present minor fat amounts, in the range of 6–12%, also according to species and season and mussels have low fat content, approximately 2%), the extraction process was optimised using a 1:1 acetonitrile:n-hexane solution as extractant solvent. This mixture was demonstrated to be useful for achieving the quantitative extraction of the analytes in these food products. The extraction process was performed in shorter that when using other solvents (ethyl acetate) in single matrices such as vegetables and soft-drinks (Gallart-Ayala et al., 2011). Each sample was prepared as follows: the whole content was homogenised including either the oil or water solution in which the sample was in. Two grams of the sample were extracted with a solution of 10 mL acetonitrile and 10 mL n-hexane. The solution obtained was stirred during 10 min in a horizontal shaker at 200 rpm followed by a centrifugation at 1500 rpm for 5 min. After 10 min, 1 mL of acetonitrile layer was evaporated using a nitrogen stream. The extract was reconstituted with 500 lL of mobile phase A and filtered using 0.45 lm PVDF hydrophilic nylon filters into HPLC vials for analysis.
2.4. Mass spectrometry Mass spectrometry data were acquired and processed using the software Analyst 1.5 from Applied Biosystems. Analytes were detected by multiple reaction monitoring mode (MRM) using electrospray ionisation (positive polarity) in a single chromatographic run per sample. Two fragmentation reactions were scanned for each analyte. The MS source parameters were the following: source temperature 450 °C, curtain gas 25 psi, nebulizer gas 60 psi, turbo gas 40 psi, ion spray voltage 5000 V, and collision gas medium. The gas used for the mass spectrometer was 99% pure nitrogen. The optimisation of the analyte dependent MS/MS parameters was performed via direct infusion of standards into the MS at a flow rate of 20 lL min 1. The standard solution were prepared in a 50:50 A:B solution (where A and B was the solvents described in the previous section of chromatographic conditions). Flow injection analysis (FIA) mode was also carried out for each compound in order to optimise the MS source conditions. 2.5. Standard solutions and spiked samples Stock solutions for each standard were prepared at a level of 500 mg L 1 in acetonitrile and they were stored in the darkness at 4 °C. Spiked solutions were prepared by diluting the stock solutions with acetonitrile. The resulting concentrations were 5 lg mL 1 for BADGE, BADGEH2O and BADGE2H2O, and 0.5 lg mL 1 for BADGEHCl, BADGE2HCl, BADGEHClH2O, BFDGE, BFDGE2HCl and BFDGEH2O. In order to take into account the influence of the fish matrix in calibration, in the present case the calibration was performed with spiked blank samples obtained by adding different quantities (comprised between 20 and 400 ll) of spiked solution described above to fish blank samples. The resulting concentrations of calibration curves obtained were 0, 5, 10, 25, 50, 75 and 100 lg kg 1 for all analytes, except BADGE, BADGEH2O and BADGE2H2O whose concentrations were 0, 50, 100, 250, 500, 750 and 1000 lg kg 1. 2.6. Samples and sample preparation Forty-one samples of different canned fish food samples in oil and aqueous media, such as mussels (8), boquerons (11), sardines (4), mackerels (4) and tuna (14), were obtained from supermarkets
2.7. Method validation procedures External calibration method was used for quantitative analysis, calibration curves for each analyte were obtained by plotting area peak versus concentration of the corresponding standard in the range 0–100 lg kg 1 for all analytes, except for BADGE, BADGEH2O and BADGE2H2O in the range 0–1000 lg kg 1. Calibration was performed with spiked blank samples in order to take into account the effect matrix. Concentration of the analytes in the fish samples was calculated on the basis of the regression line obtained from calibration curves. Detection limit, defined as the lowest concentration of analyte in a defined matrix that produced a positive identification using a specified method, was calculated as 3 SD/m, where SD is the standard deviation of 10 measurements of blank solutions and m is the slope of the calibration curve (equivalent to S/N ratio equal to 3). Quantification limit, the lowest concentration of analyte in a defined matrix that achieves a quantitative measurement a specified method, was determined on the basis of 10 SD/m (equivalent to S/N ratio equal to 10). Intra-day and inter-day studies were performed in order to test the reproducibility of the analytical method. Intra-day (n = 6) and inter-day (n = 6) precisions were obtained by the analysis of samples spiked at three concentration levels in the same and in different days (3). For BADGEHCl, BADGE2HCl, BADGEHClH2O, BFDGE, BFDGE2HCl and BFDGEH2O, the fish samples were spiked with 10, 50 and 100 lg kg 1 of standards. For BADGE, BADGE.H2O and BADGE.2H2O samples were spiked with 100, 500 and 1000 lg kg 1 of standards. Since no certified reference material for BDGE, BFDGE and derivatives in canned fish are available, the accuracy of the method was estimated by the recovery of different spiked samples with standard mixtures at the three concentration levels described above. 3. Results and discussion 3.1. Optimisation of chromatographic separation Different experiments were carried out in order to select the adequate mobile phase. After multiple experiments, a mobile phase composed of a solution of 0.01 M ammonium formate (A) and methanol (B) was selected as optimum. The use of this mobile phase produced the best response for the parent and product ions
1313
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
selected in ionisation mode. Different assays which used mobile phases containing formic acid produced lower signal than the one obtained using the selected ammonium formate–methanol, probably this fact is due to ionisation inhibition. Other authors (Zhang et al., 2010) indicated that more than 0.1% of formic acid is necessary to help the ionisation process, and that concentration higher than 0.3% produced an ionisation inhibition (optimum concentration in the range 0.2–0.3%). At the present case, the addition of different quantities of formic acid did not achieve the improvement of the signal, therefore its use was avoided. Other mobile phases evaluated were several solutions obtained from different proportions of acetonitrile–water; in these cases very low responses were obtained, probably due to the fact that quasi-moleculars clusters between the analyte and acetonitrile were not properly formed in ionisation mode (Pardo, Yusá, León, & Pastor, 2006). Therefore, the mobile phase selected for the optimisation of chromatographic separation was ammonium formate–methanol, which is similar to those used in other studies on separation, characterisation and quantification of BADGE, BFDGE and their derivatives (Gallart-Ayala et al., 2010, 2011; Pardo et al., 2006; Yonekubo, Hayakawa, & Sajiki, 2008; Zhang et al., 2010). The gradient program was optimised in order to obtain the suitable separation of analytes with the appropriate stability in the retention times. As can it be seen in Fig. 2, employing the binary gradient previously described in Section 2.3, an adequate separation of the BADGE, BFDGE and their derivatives was obtained in 8 min.
3.2. Ionisation mode and optimisation of MS–MS parameters Both electrospray ionisation (ESI) as well as atmospheric pressure chemical ionisation (APCI) were applied as ionisation modes in liquid chromatography-mass spectroscopy for the determination of BADGE, BFDGE and their derivatives (Gallart-Ayala et al., 2010, 2011; Zhang et al., 2010). Therefore, after different assays performed with ESI and APCI in both positive and negative modes, ESI ionisation in positive mode was selected, because it provided lower detection limits that APCI mode. These results are consistent with those published by Zhang et al., 2010; these authors demonstrated that APCI mode offered worse sensitivity than ESI for the determination of BADGE-related compounds, and especially for
Table 1 Summary of optimised ESI-MS/MS conditions for the determined analytes. MW: molecular weight. DP: declustering potential, CE: collision energy, CXP: collision cell potential. MW
Parent ion
Product ionsa
DP
CE
CXP
BADGE
340.42
358.2
BADGEH2O
358.43
376.2
BADGE2H2O
376.44
394.2
BADGEHCl
376.87
394.2
BADGEHClH2O
394.89
412.1
BFDGE
312.37
330.2
BFDGE2HCl
385.28
402.1
BFDGE2H2O
348.39
366.2
BADGE2HCl
413.33
430.1
191.2 135.1 209.3 191.1 209.2 135.3 227.1 135.2 227.3 135.2 163.3 133.2 199 181.3 133.3 181.4 227.2 135.1
26 26 26 26 21 21 16 16 26 26 16 16 31 31 16 16 21 21
19 41 19 27 23 47 21 47 21 47 17 21 19 31 27 17 21 49
4 4 4 4 4 4 4 4 6 4 4 4 4 4 4 6 4 4
a The first ion indicated for each analyte was employed for quantification purposes while the second ion was used for confirmation.
BADGEHClH2O. In addition, the linear ranges obtained for ESI mode were larger than when APCI were used. The optimisation of ESI conditions is crucial for an adequate MS determination for each analyte. Therefore, the study of different factors influencing the MS signal was carried out. The composition of the eluent has also an important influence in the ionisation effectiveness; consequently, a study over different eluents was performed. The infusion of standards was assayed in different solvents such as mixtures of methanol–water, mixtures of acetonitrile–water and several mixtures with different proportions A:B mobile phase. The best results obtained were achieved for a 50:50 mixture of A:B. The optimisation of the MS–MS parameters to maximise the signal of each analyte was performed. Declustering potencial (DP), collision energy (CE) and collision cell potencial (CXP) were optimised for each compound individually in infusion mode in the range 0–150 V, 0–100 V and 0–60 V, respectively. The optimisation of these parameters, as well as temperature and ion source
Fig. 2. Chromatogram of a canned fish sample spiked with BADGE-related and BFDGE-related compounds as indicated in Section 2.5.
1314
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
art-Ayala et al., 2010). Therefore, in the case at hand, the reproducible formation of ammonium adducts was favored by the use of mobile phase with ammonium formate. No other additives were necessary in order to allow the adduct formation and the efficient fragmentation with a stable signal under tandem mass spectrometry. Under these conditions, appropriate MRM chromatograms were obtained for quantitative ions of the different BADGE and BFDGE derivatives and, the transitions selected were adequate for the quantitative determination of the analytes.
Table 2 Recoveries obtained in canned fish samples at three addition levels. Added (lg kg
1
)
Recovery (%) 100
BADGE BADGEH2O BADGE2H2O
109.51 99.93 101.07
BADGEHCl BADGE2HCl BADGEHClH2O BFDGE BFDGE2H2O BFDGE2HCl
104.99 100.02 99.04 104.95 100.08 100.12
500
1000
96.09 93.84 99.98
10
91.91 90.52 95.23
50
100
97.36 96.04 96.11 94.49 101.96 95.93
94.56 94.84 95.28 89.92 100.21 92.54
3.3. Analytical figures of merit All the analytical figures of merit of the developed method were evaluated at the levels and under the conditions indicated in Method validation procedures section. Different calibrations were performed using fish spiked blank samples in order to take into account the influence of the matrix. The linearity between the area responses of the different target compounds and their concentrations were studied at levels of 0–100 lg kg 1 for all analytes, except for BADGE, BADGEH2O and BADGE2H2O in the range 0– 1000 lg kg 1. Appropriate relationships between known concentrations and measured areas were achieved under these conditions. Linear regression obtained indicated that the method is satisfactory for quantification, with correlation coefficients higher than 0.999, in all cases. Limits of detection (LOD = 3 SD/m) and quantification (LOQ = 10 SD/m) obtained for the different BADGE-related compounds are shown in Table 3. It can be observed that the range of detection limits are between 0.5 and 3.1 lg kg 1 and the quantification limit range is between 1.8 and 10.3 lg kg 1. These results are comparable to other published methods (Gallart-Ayala et al., 2011). Therefore, the present method attains enough sensitivity to determine BAGDE, BFGDE and their hydrolysed and hydrochloric derivatives at lower concentrations than the ones established by the European legislation (Commission Directive 16/EC, 2002; Commission Regulation EC 1895, 2005). Accuracy was studied by conducting recovery assays at three different concentration levels previously indicated. The values obtained are reported in Table 2. The use of calibration with spiked blank samples produced suitable recoveries in the range 89.9– 109.5% which are appropriate for analyte determinations. Reproducibility of the method was tested by intra-day and inter-day precision tests at the three concentration levels indicated above. The results are summarised in Table 3. The values obtained for repeatability (RSDr) and for inter-day precision (RSDR) are satisfactory in the range between 5.8% and 8.3% and between 6.7% and 12%, respectively. These values are satisfactory, although they are
Table 3 Summary of the analytical characteristics of the employed method. LOD: limit of detection, LOQ: limit of quantification, RSDr: Relative standard deviation of intra-day repeatability, RSDR: Relative standard deviation of inter-day precision. LOD (lg kg BADGE BADGEH2O BADGE2H2O BADGEHCl BADGE2HCl BADGEHClH2O BFDGE BFDGE2H2O BFDGE2HCl
3.0 3.1 2.3 2.1 0.8 0.5 2.0 2.1 2.4
1
)
LOQ (lg kg 10.1 10.3 7.2 7.1 2.7 1.8 5.7 6.2 7.3
1
)
RSDr (%)
RSDR (%)
8.0 7.7 8.3 7.8 9.5 7.8 7.6 5.8 7.8
12.0 8.6 9.0 8.5 10.0 8.5 12.0 6.7 9.0
gases, were executed in flow injection analysis by means of analyte standard solutions (1 lg ml 1) infused at a flow rate of 20 ll min 1. The optimum results are summarised in Table 1. Nitrogen was used as curtain gas, nebulizer gas and turbo gas. For ESI-positive mode MS–MS operation, the selection of the ions [M + H]+ and [M + NH4]+ were obtained by MS-scan mode and chosen as precursor ions. Product or daughter ions were selected by product-ion-scan. Quantification was carried out using multiple reaction monitoring mode (MRM) in order to improve selectivity and sensitivity. As indicated in Table 1, one of the product ions (MRM transition 1) was for quantitative purposes, while the other product ion (MRM transition 2) was employed for qualitative ones. The response obtained from the ions [M + H]+ was lower compared to that obtained from the ions [M + NH4]+. Both BADGEs and BFDGEs, under electrospray conditions, tend to form adducts with ammonium instead of the protonated molecule (Gall-
Table 4 Results obtained for the application of the developed method to different fish canned samples. Values for the different analytes detected in four of the analysed samples are in lg kg 1. Analyte
Number of samples analysed
Number of samples in which any analyte was detected
Number of samples with levels higher than permitted by EU Directive [2]
Mussel European anchovy Pilchard Mackerel Tuna Total
8 11
0 0
0 0
4 4 14 41
0 0 4 4
0 0 0 0
PS1-tuna PS2-tuna PS3-tuna PS4-tuna
BADGE
BADGEH2O
BADGE2H2O
BADGEHCl
BADGE2HCl
BADGEHClH2O
BFDGE
BFDGE2H2O
BFDGE2HCl
n.d. n.d. n.d. n.d.
120 n.d. n.d. n.d.
111 247 625 120
n.d. n.d. n.d. n.d.
n.d. n.d. n.d. n.d.
18 61 87 n.d.
n.d. n.d. n.d. n.d.
n.d. n.d. n.d. n.d.
n.d. n.d. n.d. n.d.
n.d.: not detected.
J. Míguez et al. / Food Chemistry 135 (2012) 1310–1315
slightly higher than other published works in which reproducibility was tested in standards instead of fish samples, avoiding the matrix effect. Under the optimised conditions, the total time of analysis per sample, including sample homogenisation, extraction and chromatographic separation, was 60 min. 3.4. Analysis of samples The applicability of the method was demonstrated by utilising them for the routine analysis of canned fish samples from Spanish origin. 41 different fish samples including mussels (Mytilus edulis), European anchovy (Engraulis encrasicholus), pilchards (Sardina pilchardus), mackerels (Scomber scombrus) and tuna (Thunnus alalunga) were analysed using the method described above. In 90.2% of cases (37 samples) no BADGEs or BFDGEs were detected, in the four remaining samples, all of them of canned tuna, different levels of hydrolysed derivatives of BADGE were discovered (see Table 4). One sample presented measurable levels of BADGE.H2O, BADGE.2H2O and BADGE.HCl.H2O (111, 120 and 18 lg kg 1, respectively). In two other samples both BADGE.2H2O and BADGE.HCl.H2O were detected at the levels 247 and 61 as well as 625 and 87 lg kg 1, respectively. And in one additional sample a concentration level of 120 lg kg 1 of BADGE.2H2O was measured. However, the very low levels founded in these four samples did not achieve the maximum SML established for the EU regulations; therefore, it can be concluded that all the analysed samples are in compliance with EU Directives, demonstrating the safety of these canned fish products. 4. Conclusion A single and fast LC–MS/MS method that can be used for routine analysis of BADGE, BFDGE and their derivatives in canned fish has been developed. The method has been optimised taking into account the specific characteristics of food matrix products and it can be considered complementary of other LC–MS/MS methods using [M + NH4]+ aducts that were developed to determine these analytes in other food products with different matrices such as vegetables and drinks (Gallart-Ayala et al., 2011). The extraction of analytes from fish was performed by means of a simple and fast extraction with a 1:1 solution of acetonitrile:n-hexane. BADGE, BFDGE and their derivatives were measured on the basis of the fragmentation selected products from [M + NH4]+ aducts. The method was validated and it was demonstrated to be suitable, specific, accurate and sensitive enough for the quantification of BADGE, BFDGE and hydrolysed reaction products in order to verify the compliance with the EU Directives. The developed procedure was applied to determine the levels of BADGE, BFDGE and related compounds in different canned fish products from Spain. In all cases the concentrations determined are under the levels permitted by EU regulations concerning the specific migration of these compounds in contact with foodstuffs. Acknowledgements The authors wish to express their gratitude to Dirección Xeral of Innovación and Xestion of Saúde Pública, Consellería de Sanidade, Xunta de Galicia for their support. We also thank the collaboration of technical staff of Public Health Laboratory Lugo who participated in this work.
1315
References Berger, U., & Oehme, M. (2000). Identification of derivatives of bisphenol a diglycidyl ether and novolac glycidyl ether in can coatings by liquid chromatography/ion trap mass spectrometry. Journal of AOAC International, 83, 1367–1376. Berger, U., Oehme, M., & Girardin, L. (2001). Quantification of derivatives of bisphenol A diglycidyl ether (BADGE) and novolac glycidyl ether (NOGE) migrated from can coatings into tuna by HPLC/fluorescence and MS detection. Fresenius Journal of Analytical Chemistry, 369, 115–123. Cabodo, A. G., Aldea, S., Porro, C., Ojea, G., Lago, J., Sobrado, C., et al. (2008). Migration of BADGE (bisphenol A diglycidyl-ether) and BFDGE (bisphenol F diglycidyl-ether) in canned seafood. Food and Chemical Technology, 46, 1674–1680. Cao, X., Dufresne, G., Clement, G., Belisle, S., Robichaud, A., & Beraldin, F. (2009). Levels of bisphenol A diglycidyl ether (BADGE) and bisphenol F diglycidyl ether (BFDGE) in canned liquid infant formula products in Canada and dietary intake. Journal of AOAC International, 92, 1780–1789. Commission Directive 2002/16/EC of 20 February 2002 on the use of certain epoxy derivates in materials and articles intended to come into contact with foodstuffs (2002). Official Journal of the European Communities, L51/27. Gallart-Ayala, H., Moyano, E., & Galcerán, M. T. (2010). Multiple-stage mass spectrometry analysis of bisphenol A diglycidyl ether, bisphenol F diclycidyl ether and their derivatives. Rapid Communications in Mass Spectrometry, 24, 3469–3477. Gallart-Ayala, H., Moyano, E., & Galcerán, M. T. (2011). Fast liquid chromatography– tandem mass spectrometry for the analysis of bisphenol A-diglycidyl ether, bisphenol F-diglycidyl ether and its derivatives in canned food and beverages. Journal of Chromatography A, 1218, 1603–1610. Leepipatpiboon, N., Sae-Khow, O., & Jayanta, S. (2005). Simultaneous determination of bisphenol-A-diglycidyl ether, bisphenol-F-diglycidyl ether, and their derivatives in oil-in-water and aqueous-based canned foods by highperformance liquid chromatography with fluorescence detection. Journal of Chromatography A, 1073, 331–339. Pardo, O., Yusá, V., León, N., & Pastor, A. (2006). Determination of bisphenol diglycidyl ether residues in canned foods by pressurized liquid extraction and liquid chromatography–tandem mass spectrometry. Journal of Chromatography A, 1107, 70–78. Paseiro-Losada, P., Paz-Abuín, S., Vázquez Odériz, L., Simal-Lozano, J., & SimalGándara, J. (1991). Quality control of cured epoxy resins: determination of residual free monomers (m-xylylenediamine and bisphenol A diglycidyl ether) in the finished product. Journal of Chromatography A, 585, 75–81. Paseiro-Losada, P., Pérez-Lamela, C., López-Fabal, F., Sanmatín-Fenollera, P., & Simal-Lozano, J. (1997). Two RP-HPLC sensitive methods to quantify and identify bisphenol A diglycidyl ether and its hydrolysis products. 1. European Union aqueous food simulants. Journal of Agriculture and Food Chemistry, 45, 3493–3500. Petersen, H., Schaefer, A., Buckow, C. A., Simat, T. J., & Steinhart, H. (2003). Determination of bisphenol A diglycidyl ether (BADGE) and its derivatives in food: identification and quantification by internal standard. European Food Research and Technology, 216, 355–364. Philo, M. R., Jickells, S. M., Damat, A. P., & Castle, L. (1994). Stability of plastics monomers in food-simulating liquids under European Union migration test conditions. Journal of Agriculture and Food Chemistry, 42, 1497–1501. Commission Regulation (EC) No 1895/2005 of 18 de November 2005 on the restriction of use of certain epoxy derivates in materials and articles intended to come into contact with foodstuffs (2005). Official Journal of the European Communities, L302, 28–32. Satoh, K., Ohyama, K., Aoki, N., Iida, M., & Nagai, F. (2004). Study on anti-androgenic effects of bisphenol a diglycidyl ether (BADGE), bisphenol F diglycidyl ether (BFDGE) and their derivatives using cells stably transfected with human androgen receptor, AR-EcoScreen. Food and Chemical Toxicology, 42, 983–993. Simal-Gándara, J., Paz-Abuín, S., Paseiro-Losada, P., López-Mahía, P., & SimalLozano, J. (1992). Identification of RP-HPLC peaks of bisphenol F and of bisphenol F diglycidyl ether and its hydrolysis products by thermospray mass spectrometry and gas chromatography/mass spectrometry. Cromatographia, 34, 67–72. Uematsu, Y., Hirata, K., Suzuki, K., Iida, K., & Saito, K. (2001). Chlorohydrins of bisphenol A diglycidyl ether (BADGE) and of bisphenol F diglycidyl ether (BFDGE) in canned foods and ready-to-drink coffees from the Japanese market. Food Additives & Contaminants, 18, 177–185. Yonekubo, J., Hayakawa, K., & Sajiki, J. (2008). Concentrations of bisphenol A, bisphenol A diglycidyl ether, and their derivatives in canned foods in japanese markets. Journal of Agriculture and Food Chemistry, 56, 2041–2047. Zhang, H., Xue, M., Zou, Y., & Dai, Z. (2010). Simultaneous determination of NOGErelated and BADGE-related compounds in canned food by ultra-performance liquid chromatography tandem mass spectrometry. Analytical and Bioanalytical Chemistry, 398, 3165–3174.