Comparison of dimethylformamide dialkylacetal derivatization reagents for the analysis of heterocyclic amines in meat extracts by gas chromatography–mass spectrometry

Analytica Chimica Acta 545 (2005) 209–217

Comparison of dimethylformamide dialkylacetal derivatization reagents for the analysis of heterocyclic amines in meat extracts by gas chromatography–mass spectrometry E. Barcel´o-Barrachina, F.J. Santos, L. Puignou ∗ , M.T. Galceran Departament de Qu´ımica Anal´ıtica, Universitat de Barcelona, Avda. Diagonal 647, 08028 Barcelona, Spain Received 16 March 2005; received in revised form 3 May 2005; accepted 3 May 2005 Available online 1 June 2005

Abstract A simple and selective methodology for the analysis of heterocyclic amines (HAs) in a meat extract by gas chromatography–mass spectrometry (GC–MS) is proposed. A comparative study of several HAs derivatization procedures based on the formation of Schiff bases using N,N-dimethylformamide dialkylacetals reagents was performed. Optimization of the reaction conditions was carried out, such as reagent volume (2 ␮l to derivatize 1 ng of HAs) temperature (100 ◦ C) and time (10 min). After that, the GC–MS working parameters for the analysis of HAs derivatives were also studied in order to achieve the best chromatographic separation with the maximum sensitivity. Among the derivatization reagents, N,N-dimethylformamide di-tert-butylacetal (DMF-DtBA) was selected because it provided the best yield in the derivatization process, and consequently the best sensitivity in the GC–MS method. Quality parameters such as limits of detection and repeatability were established using a meat extract sample. A complete validation of the methodology was achieved using a laboratory reference material previously analyzed in a European interlaboratory exercise. © 2005 Elsevier B.V. All rights reserved. Keywords: Heterocyclic amines; Gas chromatography; Derivatization

1. Introduction Heterocyclic amines (HAs) are a family of compounds that have been isolated and identified from cooked foods [1]. They are potent mutagens in Salmonella assays and many of them have also proved to be carcinogenic in rodent feeding studies. For these reasons, it is important to determine the amount of these mutagens present in cooked foods in order to estimate the intakes and the risks to human health. As HAs are present at low concentration level in complex matrices, it is very important to develop sensitive and selective methodologies to analyze these compounds in food and feed samples. In the last few decades, many research groups have been working on the development of extraction and purification methods for HAs determination [2] using chromatographic ∗

Corresponding author. Fax: +34 934021233. E-mail address: [email protected] (L. Puignou).

0003-2670/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2005.05.002

techniques. Among these, liquid chromatography (LC) has been the most widely used separation technique with detection systems such as ultraviolet-diode array detector (UV-DAD) [3,4] or coupled to mass spectrometry (MS) [5,6]. Gas chromatography (GC) has also been used for HAs analysis due to its simplicity, separation efficiency, and high sensitivity and specificity when it is coupled to mass spectrometry. However, a derivatization step to increase the volatility of the amines is required. In the last 15 years, several derivatization methods have been proposed. The method most widely used was the alkylation of the primary amino group by means of 3,5-bistrifluoromethylbenzyl bromide, proposed by Murray and co-workers [7–12]. Although it provided high sensitivity, the derivative products were a mixture of mono- and di-alkylated forms, and besides only a few HAs could be efficiently derivatized. Another reagent proposed in the literature is heptafluorobutyric acid anhydride [13] to perform an acylation reaction that has traditionally

210

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

been used to derivatize aromatic amines, but due to the acidity of the acylated aminoimidazoazaarenes, this method was unsuitable. Then, in order to overcome this problem, several laboratories recommended carrying out a second derivatization step after the acylation that consisted in the methylation of the amino group by means of diazomethane [11,14–18] or dimethylformamide dimethylacetal [19]. However, these methods have only been applied for the analysis of a few HAs. Friesen and co-workers reported another derivatization procedure by means of bis-pentafluorobenzyl bromide [20–24], but this was only used for determination of PhIP from DNA adducts. In the last few years, the use of N,N-dimethylformamide dimethylacetal (DMF-DMA) [25] and N-methyl-N-(tert-butyldimethylsilyl)trifluoroacetamide [26] has been proposed, allowing the simultaneous analysis of a higher number of HAs in combustion smoke [27], river water [28] and cooked food [26,29] samples. Nevertheless, both procedures have been applied to the analysis of HAs in several samples without a complete validation of the methodologies. In fact, the formation of HAs in cooked meat samples varies depending on the cooking technique, temperature and the degree of doneness; so analysis methods might not be comparable since concentration levels of HAs can be very different. Moreover, the complexity of food matrices can greatly influence the analytical results and consequently validation of the methodology should be necessary. Among the above-mentioned derivatization procedures, the formation of N,N-dimethylformamide dialkylacetal derivatives offers some advantages. For instance, the derivatization reaction is performed in a single step, the excess of reagent can be easily removed by evaporation, and the derivatives present high stability. In this work and in order to carry out HAs derivatization for GC–MS analysis, three new reagents such as DMF-diethylacetal (DMF-DEA), -diisopropylacetal (DMF-DIPA) and -di-tert-butylacetal (DMF-DtBA) together with DMF-DMA have been studied. First, optimization of the reaction conditions were performed, such as reagent volume, derivatization temperature and time. After that, the GC–ion trap mass spectrometric (ITMS) parameters for the analysis of HAs derivatives were also studied in order to achieve the best chromatographic separation and the maximum sensitivity. In order to achieve a complete validation of the method, the optimum conditions were used for the quantitative determination of HAs in a lyophilized meat extract using a laboratory reference material [30]. The results were compared with those obtained in a previous European interlaboratory exercise [31].

3-methylimidazo[4,5-f]quinoline (IQ), 2-amino-3,4-dimethylimidazo[4,5-f]quinoline (MeIQ), 2-amino-3, 8-dimethylimidazo[4,5-f]quinoxaline (MeIQx), 2-amino-6-methyldipyrido[1,2-a:3 ,2 -d]imidazole (Glu-P-1), 2-amino-3,4,8trimethylimidazo[4,5-f]quinoxaline (4,8-DiMeIQx), 2– amino-3,7,8-trimethylimidazo[4,5-f]quinoxaline (7,8-DiMeIQx), 3-amino-1,4-dimethyl-5H-pyrido[4,3-b]indole (Trp-P-1), 3-amino-1-methyl-5H-pyrido[4,3-b]indole (TrpP-2), 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), 2-amino-9H-pyrido[2,3-b]indole (A␣C) and 2-amino-3-methyl-9H-pyrido[2,3-b]indole (MeA␣C), obtained from Toronto Research Chemicals Inc. (Toronto, Canada). p-Terphenyl was used as internal standard and it was purchased by Fluka (Buchs, Switzerland). The derivatization reagents used in this work were: N,N-dimethylformamide dimethylacetal, purchased from Pierce (Rockford, USA); N,N-dimethylformamide diethylacetal and N,N-dimethylformamide diisopropylacetal, that were all from Fluka (Buchs); N,N-dimethylformamide di-tert-butylacetal, that was from Aldrich (Steinhem, Germany). Molecular sieves were purchased from Fluka (Buchs). Before use, this material was activated at 200 ◦ C and it was added to the vials that contained the reagents and solvents used in the analysis in order to prevent the presence of moisture. Anhydrous methanol was purchased from Aldrich (Steinhem) and HPLC grade ethyl acetate was from Romil Chemicals (Shepshed, UK). Gradient grade acetonitrile and methanol were purchased from Merck (Darmstadt, Germany). Water was purified in an Elix-Milli Q system (Millipore Co., Bedford, MA, USA). Analytical grade ammonia solution (25%) and formic acid (98%) were obtained from Merck (Darmstadt) and ammonium formate from Fluka (Buchs). Diatomaceous earth extraction cartridges (Extrelut-20) and refill material were provided by Merck (Darmstadt); PRS sodium form (500 mg) and endcapped C18 (100 and 500 mg) Bond Elut cartridges, coupling pieces and stopcocks were from Varian (Harbor City, USA). Derivatization reaction vials of 5 ml were purchased from Supelco (Bellefonte, USA). HA methanolic stock standard solutions of 100 ␮g g−1 were prepared and used for further dilutions. Although the methanol used was anhydrous, molecular sieves were added at each vial in order to avoid the presence of moisture. A standard mixture of all amines over 3 ppm was used to establish the quality parameters of the methodology, and it also contained a little quantity of molecular sieves.

2. Experimental

2.2. Sample treatment

2.1. Materials and chemicals

A lyophilized meat extract was prepared from a commercial beef extract (Bovril® ), which was spiked before the lyophilization with DMIP, IQ, MeIQ, MeIQx, 4,8-DiMeIQx, Trp-P-1, Trp-P-2, PhIP, A␣C and MeA␣C at a level of ∼70 ng g−1 extract [30]. The same lyophilized meat extract

The HAs studied, which are shown in Fig. 1, were 2-amino-1,6-dimethylimidazo[4,5-b]pyridine (DMIP), 2aminodipyrido[1,2-a:3 ,2 -d]imidazole (Glu-P-2), 2-amino-

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

211

Fig. 1. Structures of the HAs studied.

but without spiked HAs was used for establishing the quality parameters of the developed method. To extract the analytes from these samples, a previously described clean-up method [32] was used. Briefly, 1 g of beef extract sample was homogenized in 12 ml of 1 M NaOH and mixed with diatomaceous earth. The amines were eluted from the extraction column, containing the diatomaceous earth mixture, directly to a propanesulfonic acid (PRS) cartridge using 75 ml of ethyl acetate. It was dried and rinsed with 6 ml of 0.01 M HCl, 15 ml of MeOH–0.1 M HCl (6:4) and 2 ml of water, which contained the less-polar compounds. After adding 25 ml of water, the combined acidic washing solution was neutralized with 500 ␮l of ammonia. It was passed through a C18 (500 mg) cartridge and the amines retained were eluted, using 1.4 ml of methanol–ammonia solution (9:1, v/v) providing the less-polar extract. The PRS column was then coupled to a C18 (100 mg) cartridge, and after that the most polar amines were eluted from the cationic exchanger with 20 ml of 0.5 M ammonium acetate solution at pH 8.5. The adsorbed HAs were then eluted from C18 , using 0.8 ml of methanol–ammonia solution (9:1, v/v) providing the polar

extract. The two extracts were evaporated to dryness under a stream of nitrogen, and after that the residue was redissolved in anhydrous methanol for the next derivatization step. A Supelco Visiprep and a Visidry SPE vacuum manifold (Supelco, Gland, Switzerland) were used for manipulations with solid-phase extraction cartridges and solvent evaporation, respectively. 2.3. HAs derivatization procedure for GC analysis An aliquot of methanolic solution containing the HAs was mixed in the reaction vial with the derivatization reagent, and the mixture was heated at 100 ◦ C for 10 min in an aluminum block. The first 5 min of reaction were performed capping the vial, to enhance the reactivity between the amines and the derivatization reagent. Once the solution was at room temperature, it was evaporated to dryness under a stream of nitrogen. Methanol anhydrous (over 200 ␮l) was added to the vial before the evaporation was completed, in order to concentrate the analytes at the bottom of the vial, and thus to maximize the recovery of the analytes. After that, the residue

212

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

was dissolved in 30 ␮l of ethyl acetate, which contained the internal standard (p-terphenyl). The solution was stored at 4 ◦ C prior to GC analysis. 2.4. GC–MS instrumentation and conditions GC analysis was carried out with a Varian 3800 CX capillary gas chromatograph coupled to a Varian Saturn 2200 ion trap mass spectrometer (Sugar Land, TX, USA). The chromatographic separation was performed using a DB-5 (5% phenyl–95% methylpolysilixane) 30 m × 0.25 mm i.d. fused-silica capillary GC column (J&W Scientific, Folsom, CA, USA) of 0.25 ␮m particle size. Injector temperature was kept at 280 ◦ C and the sample injection was done in splitless mode (1 min). Oven temperature program was: 70 ◦ C (held for 1 min) to 240 ◦ C at 25 ◦ C/min (held for 1 min) and to 300 ◦ C at 10 ◦ C/min (held for 5 min). Helium was used as carrier gas at a constant flow-rate of 1 ml/min held by electronic pressure control at 90 ◦ C. The ion trap mass spectrometer was operated in the following conditions: ion trap temperature was set at 220 ◦ C, injection temperature was 280 ◦ C, and transfer line temperature was 290 ◦ C. Data acquisition was performed in full scan mode from m/z 50 to 290 at 0.7 s/scan (3 ␮scan/scan). The instrument was tuned in positive electron ionization mode (EI) using perfluorotributylamine (FC-43) according to the manufacturer’s recommendations to achieve the best sensitivity. Parameters such as automatic gain control (AGC) and multiplier (1350 V, 105 gain) were set by automatic tuning. The electron energy was 70 eV and the emission current value 100 ␮A. Varian Workstation Version 6.20 software was used for data acquisition and processing of the results.

3. Results and discussion New derivatization procedures based on the formation of Schiff bases by means of several N,N-dimethylformamide dialkylacetals reagents were studied. Fig. 2 shows the different alkyl chains of the studied derivatization reagents, and

Fig. 3. Optimization of experimental conditions for derivatization of GluP-2. (A) Influence of reagent volume to a solution containing 100 ng of each HA. (B) Influence of the reaction temperature a solution containing 60 ng of each HA.

a reaction scheme for the nucleophilic substitution to form the N-dimethylaminomethylene derivative is also included. However, this reaction can present some problems related to the presence of moisture and the possible secondary reactions that produce a decrease in the yield of the derivatization. In order to overcome these negative factors, some experiments were carried out to establish suitable conditions for the preparation of HAs derivatives. 3.1. Optimization of HAs derivatization procedure The most important condition to achieve a suitable yield in the derivatization is to avoid the presence of moisture in

Fig. 2. Reaction between an heterocyclic amine and a N,N-dimethylformamide dialkylacetal by means of a nucleophylic substitution of the primary amino group to the acetal group.

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

213

Fig. 4. Extracted ion chromatograms of HAs derivatives of a standard solution of 3.3 ng/␮l using DMF-DtBA as derivatization reagent.

both the methanolic solutions of HAs and in the bottles that contained the reagent, so mixing molecular sieves with these solutions before the reaction allowed to remove efficiently the water traces. The reaction temperature, the reaction time and the volume of reagent used have a great influence in the derivatization process, so a systematic optimization of these parameters was performed. The initial values used were those given in

the literature [25], and besides these studies were performed using the four reagents in order to compare their behavior in the derivatization. The first parameter to study was the influence of reagent volume. Solutions containing 100 ng of each HA were mixed with volumes from 20 to 300 ␮l of reagent, and the reaction was performed at 100 ◦ C during 15 min. As an example, Fig. 3A shows the results obtained for Glu-P-2 using the four dialkylacetals, where it can be observed that

214

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

the optimum value was 200 ␮l, so in this case, 2 ␮l of reagent were necessary to derivatize 1 ng of HAs in the mixture. The second parameter to optimize was the reaction time. To carry out this study, solutions containing 60 ng of each HA were mixed with 120 ␮l of reagent at a 100 ◦ C, and the reaction time was varied from 5 to 30 min. Maximum derivatization yield was obtained after 10 min. Finally, in order to optimize the reaction temperature, solutions containing 60 ng of each HA were mixed with 120 ␮l of reagent and they were kept during 10 min at different temperatures from 60 to 120 ◦ C. As Fig. 3B shows, for Glu-P-2 the optimum value was 100 ◦ C, because at higher temperature a possible degradation of the reagents or partial evaporation of the derivatives could occur, producing an important decrease in the derivatives formation. Moreover, in all the experiments and for all the HAs, it was observed that the DMF-DtBA derivatives provided the maximum response. 3.2. Performance of the GC–MS methodology for the analysis of HAs Once the derivatization procedure was established, the study of the parameters concerning the GC–ITMS instrument was performed. The most important factors to optimize were the temperatures of injector, transfer line and ion trap, because the derivatives of some polar HAs had been observed as provided large tailing peaks. Thus, the injector temperature was varied from 250 to 290 ◦ C, and the selected value was 280 ◦ C. The transfer line temperature was studied from 270 to 300 ◦ C, and the final value was 290 ◦ C. The same occurred for ion trap temperature, in which it was required to maintain it at 220 ◦ C. In these conditions, peak tailing was decreased considerably. In Fig. 4, the GC–MS chromatograms of a derivatized standard solution using DMF-DtBA are given. As can be seen, a complete chromatographic separation was achieved in 22 min. Full scan spectra of HAs derivatives using electronic impact ionization were obtained. As an example, Fig. 5 shows the spectra corresponding to two representative HAs, TrpP-2 and IQ. It was observed that the molecular ion [M]+ of all derivatives was the base peak of the spectra except for Trp-P-1 and Trp-P-2 derivatives, in which the base peak was [M − 15]+ . Moreover, further fragmentation occurred providing ions of relative abundance <50% that were very similar for all analytes because these fragments were due to consecutive losses of the derivatized primary amino group. For instance, m/z [M − 15]+ corresponded to the loss of • CH ; [M − 30]+ was the loss of • C H ; [M − 44]+ was the 3 2 6 loss of • N(CH3 )2 ; [M − 56]+ was the loss of • C N(CH3 )2 ; [M − 71]+ was the loss of • N=CH N(CH3 )2 . In order to obtain the chromatogram corresponding to each analyte, the m/z selected was the base peak of each spectrum, and these are given in Table 1. For Trp-P-1 and Trp-P-2, two ions were taking into account for quantification purposes because they had high relative abundances. The chromatographic peak areas corresponding to the selected ions provided for each HA

Fig. 5. Full scan spectra of the (A) Trp-P-2 and (B) IQ derivatives, scanning from m/z 50 to 290.

a linear working concentration range from 0.3 to 10 ng/␮l. Moreover, Table 1 indicates the confirmation ions used for HAs determinations that corresponded to the [M − 56]+ as characteristic fragment ion. Once the optimal conditions for GC–MS analysis were established, the four reagents were compared in order to evaluate the sensitivity of the methodology in terms of signal-to-noise ratio. It was observed (see Fig. 3) that DMFDtBA gave the highest response. Moreover, the reagent proposed in the bibliography, DMF-DMA, was the one that provided the lowest yield in the derivatization reaction, giving a decrease in the response of three to eight times respect

Table 1 m/z Corresponding to HAs derivatives, used for quantitation and confirmation procedures Compounds

Quantitative ion, m/z (abundance %)

Confirmation ion [M − 56]+ , m/z (abundance %)

DMIP Glu-P-2 Glu-P-1 A␣C MeA␣C Trp-P-1 Trp-P-2 IQ MeIQ MeIQx 7,8-DiMeIQx 4,8-DiMeIQx PhIP

217 (100) 239 (100) 253 (100) 238 (100) 252 (100) 266 (81) + 251 (100) 252 (94) + 237 (100) 253 (100) 267 (100) 268 (100) 282 (100) 282 (100) 279 (100)

161 (19) 183 (22) 197 (18) 182 (13) 196 (12) 210 (27) 196 (49) 197 (15) 211 (14) 212 (27) 226 (35) 226 (25) 223 (14)

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

215

Fig. 6. Extracted ion chromatograms corresponding to HAs derivatives after their extraction from a lyophilized meat extract, using DMF-DtBA as derivatization reagent. (A) Polar extract and (B) less-polar extract.

DMF-DtBA derivatives. On the other hand, for the same HAs amount, signal-to-noise ratio of DMF-DEA and DMF-DIPA derivatives were slightly higher than DMF-DMA, but not as good as DMF-DtBA. This fact could be explained because a favoured nucleophilic substitution occurred between

the HA and the acetal group of the reagent (see Fig. 2). Actually, a more stabilized anion R O− is released in direct proportion to the degree of branching in the alkyl chain R in the derivatization reagent, and as a consequence, an improvement in the derivatization yield can be obtained.

216

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

Therefore, DMF-DtBA was proposed as derivatization reagent to perform HAs determinations by GC–MS. 3.3. Analysis of HAs in a laboratory reference material The determination of HAs in a lyophilized meat extract was performed in order to validate the proposed GC–MS method because this sample was proposed as laboratory reference material and it was previously analyzed in an interlaboratory comparison exercise. A preliminary comparative study was carried out between the proposed new reagent DMFDtBA and DMF-DMA, which is the derivatization reagent recommended in the literature. As was expected, DMF-DtBA provided higher sensitivity than DMF-DMA, and besides the chromatograms presented fewer interfering peaks. Because of the influence of sample matrix, the DMF-DtBA amount for derivatization was re-optimized. Actually, although an exhaustive clean-up was performed before derivatization, the reagent was able to react with some other compounds present in the matrix. So, the volume of reagent required to derivatize the HAs could be higher than the volume used for HAs in a standard solution. Some experiments with meat extract were performed using increasing volumes of DMF-DtBA until 10fold the required quantity of reagent for standard solutions. In this type of sample, a ratio of 10 ␮l for a solution containing 1 ng of each HA was necessary to achieve the highest signal. Under these new conditions, the analysis of HAs in the meat sample was performed. Some quality parameters of the method were established using a meat extract without detectable quantities of HAs. Limits of detection defined as the concentration of the analytes in the sample, which causes a peak with a signal-to-noise ratio of 3, were also determined. In order to calculate their values, a meat extract without detectable quantities of HAs was spiked with each HA at low levels and the compounds were extracted with the established clean-up procedure. Under these conditions, limits of detection of the method for these compounds ranged from 0.9 to 10.3 ng g−1 of sample except for PhIP and Trp-P-2 that were slightly higher. Moreover, repeatability or short-term precision was determined by six replicate analysis of this meat extract spiked with 100 ng of each HA. Relative standard deviations of the chromatographic peak area relative to that of the internal standard were calculated for all analytes, and they varied from 6.0 to 12.0%, showing that this method provides the values of repeatability required for an accurate analysis of HAs. Quantitation of HAs in the lyophilized meat extract was performed by standard addition method, at four spiking levels. Six independent replicates on three separate days and from different bottles were analyzed following the experimental procedure described in the interlaboratory exercise above mentioned [31]. The clean-up procedure used is described in Section 2.2, and the final extracts were derivatized using DMF-DtBA as indicated before. As an example, the extracted ion chromatograms corresponding to polar and lesspolar extracts are shown in Fig. 6, where it can be observed

Table 2 Quantitation of HAs in a lyophilized meat extract by standard addition method, using DMF-DtBA as derivatization reagent for GC–MS analysis, and comparison with results obtained in an interlaboratory exercise between six European laboratories [31] (95% confidence interval) HAs

Concentration of HAs (ng g−1 meat extract) GC–MS

DMIP IQ MeIQ MeIQx 4,8-DiMeIQx Trp-P-2 Trp-P-1 PhIP A␣C MeA␣C

59.4 66.9 67.7 82.9 59.6 57.6 71.1 65.5 57.6 77.2

± ± ± ± ± ± ± ± ± ±

7.3 6.6 9.0 15.8 9.0 7.5 9.7 7.3 10.7 13.1

Interlaboratory exercise 68.1 64.5 63.6 64.5 71.6 68.6 71.4 71.9 72.0 70.8

± ± ± ± ± ± ± ± ± ±

9.1 10.9 11.6 15.6 13.8 19.3 9.2 10.0 8.2 9.3

the complexity of the matrix by the high amounts of interfering peaks that appeared in the less-polar extract. However, none of these peaks avoided the determination of the analytes. The results of the quantitation by GC–MS are given in Table 2, as well as those obtained in the intercomparison exercise [31]. In this case, different methodologies based in LC–MS and LC with electrochemical detection were used to perform the HAs analysis. Moreover, it must be taken into account that these LC methods were previously validated using other laboratory reference materials such as test solutions and other lyophilized meat extracts, thus giving a high reliability in the provided quantitation results. The concentration values of HAs in the sample determined by the proposed GC–MS methodology were in agreement with those given in the interlaboratory exercise, and moreover, the relative standard deviations varied from 9.8 to 19.0%, which were significantly better than for the interlaboratory values, i.e. 12.8–28.1%. Consequently, these results demonstrated the applicability of this new GC–MS method for determination of HAs in complex food samples such as the analyzed meat extract.

4. Conclusions As a selective methodology is required for HAs analysis in food, a simple derivatization procedure for GC–MS analysis was developed. Four N,N-dimethylformamide dialkylacetals were evaluated as derivatization reagents. Optimization of the reaction conditions were established, concluding that the higher derivatization yield for the four reagents tested was obtained at 100 ◦ C, 10 min and adding a ratio of 2 ␮l of reagent in a solution containing 1 ng of each HA. DMF-DtBA provided the highest sensitivity in HAs determination. The applicability of the method was studied using a lyophilized meat extract. Good repeatability between 6 and 12% and low enough limits of detection for quantitative purpose ranged from 0.9 to 10.3 ng g−1 of sample were determined. The quantitative determination of HAs in this laboratory reference material, which was analyzed in an interlaboratory

E. Barcel´o-Barrachina et al. / Analytica Chimica Acta 545 (2005) 209–217

exercise, was performed. The results obtained by the new procedure showed a good agreement with those given using other validated methodologies based on LC and LC–MS, which demonstrates the potentiality of the GC–MS method for the analysis of complex food matrices such as meat extracts.

Acknowledgements This work has been carried out with support from the European Commission, Priority 5 on Food Quality and Safety (Contract no. FOOD-CT-2003-506820 Specific Targeted Project), “Heat-generated food toxicants—identification, characterisation and risk minimisation”. This publication reflects the author’s views and not necessarily those of the EC. The information in this document is provided as is and no guarantee or warranty is given that the information is fit for any particular purpose. The user thereof uses the information at its sole risk and liability. Financial support was also provided by the Ministerio de Ciencia y Tecnolog´ıa, project AGL2000-0948. The authors also wish to thank Dr. Andr´es Lukach for his valuable help during the optimization of the derivatization procedure.

References [1] P. Pais, M.G. Knize, J. Chromatogr. B 747 (2000) 139. [2] F. Toribio, L. Puignou, M.T. Galceran, J. Chromatogr. B 747 (2000) 171. [3] G. Gross, Carcinogenesis 11 (1990) 1597. [4] G. Gross, A. Gr¨uter, J. Chromatogr. 592 (1992) 271. [5] E. Barcel´o-Barrachina, E. Moyano, L. Puignou, M.T. Galceran, J. Chromatogr. A 1023 (2004) 67. [6] R.D. Holland, J. Taylor, L. Schoenbachler, R.C. Jones, J.P. Freeman, D.W. Miller, B.G. Lake, N.J. Gooderham, R.J. Turesky, Chem. Res. Toxicol. 17 (2004) 1121. [7] S. Murray, N.J. Gooderham, A.R. Boobis, D.S. Davies, Carcinogenesis 9 (1988) 321. [8] S. Murray, N.J. Gooderham, A.R. Boobis, D.S. Davies, Carcinogenesis 10 (1989) 763.

217

[9] A.M. Lynch, M.G. Knize, A.R. Boobis, N.J. Gooderham, D.S. Davies, S. Murray, Cancer Res. 52 (1992) 6216. [10] S. Murray, A.M. Lynch, M.G. Knize, N.J. Gooderham, J. Chromatogr. 616 (1993) 211. [11] S. Murray, B.G. Lake, S. Gray, A.J. Edwards, C. Springall, E.A. Bowey, G. Williamson, A.R. Boobis, N.J. Gooderham, Carcinogenesis 22 (2001) 1413. [12] W.G. Stillwell, R. Sinha, S.R. Tannenbaum, Carcinogenesis 23 (2002) 831. [13] N. Loprieno, G. Boncristiani, G. Loprieno, Food Chem. Toxicol. 29 (1991) 377. [14] R. Reistad, O.J. Rossland, Food Chem. Toxicol. 35 (1997) 945. [15] R. Reistad, S.H. Nyholm, L.S. Haug, G. Becher, J. Alexander, Biomarkers 4 (1999) 263. [16] S. Hegstad, E. Lundanes, R. Reistad, L.S. Haug, G. Becher, J. Alexander, Chromatographia 52 (2000) 499. [17] C. Magagnoti, F. Orsi, R. Barnati, N. Celli, D. Rotilio, R. Fanelli, L. Airoldi, Int. J. Cancer 88 (2000) 1. [18] S. Hegstad, R. Reistad, L.S. Haug, J. Alexander, Pharmacol. Toxicol. 90 (2002) 333. [19] T.A. Sasaki, J.M. Wilkins, J.B. Forehand, S.C. Moldoveanu, Anal. Lett. 34 (2001) 1749. [20] M.D. Friesen, Environ. Health Perspect. 99 (1993) 179. [21] M.D. Friesen, K. Kaderlik, D.X. Lin, L. Garren, H. Bartsch, N.P. Lang, F.F. Kadlubar, Chem. Res. Toxicol. 7 (1994) 733. [22] M.D. Friesen, D.A. Cummings, L. Garren, R. Butler, H. Bartsch, H.A.J. Schut, Carcinogenesis 17 (1996) 67. [23] H.A.J. Schut, D.A. Cummings, M.H.E. Smale, S. Josyula, M.D. Friesen, Mutat. Res. Fundam. Mol. Mech. Mut. 376 (1997) 185. [24] M.D. Friesen, N. Rothman, P.T. Strickland, Cancer Lett. 173 (2001) 43. [25] H. Kataoka, K. Kijima, J. Chromatogr. A 767 (1997) 187. [26] S. Casal, E. Mendes, J.O. Fernandes, M.B.P.P. Oliveira, M.A. Ferreira, J. Chromatogr. A 1040 (2004) 105. [27] H. Kataoka, K. Kijima, G. Maruo, Bull. Environ. Contam. Toxicol. 60 (1998) 60. [28] H. Kataoka, H. Hayatsu, G. Hietsch, H. Steinkellner, S. Nishioka, S. Narimatsu, S. Knasmuller, T. Hayatsu, Mutat. Res. 466 (2000) 27. [29] H. Kataoka, S. Nishioka, M. Kobayashi, T. Hanaoka, S. Tsugane, Bull. Environ. Contam. Toxicol. 69 (2002) 682. [30] E. Bermudo, R. Busquets, E. Barcel´o-Barrachina, L. Puignou, F.J. Santos, M.T. Galceran, J. Chromatogr. B 802 (2004) 61. [31] F.J. Santos, E. Barcel´o-Barrachina, F. Toribio, L. Puignou, M.T. Galceran, E. Persson, K. Skog, C. Messner, M. Murkovic, U. Nabinger, A. Ristic, J. Chromatogr. B 802 (2004) 69. [32] M.T. Galceran, E. Moyano, L. Puignou, P. Pais, J. Chromatogr. A 730 (1996) 185.