Quick screening of priority beta-agonists in urine using automated TurboFlow™ - LC/Exactive mass spectrometry Thorsten Bernsmann, Peter Fuerst, Michal Godula
To cite this version:
Thorsten Bernsmann, Peter Fuerst, Michal Godula. Quick screening of priority beta-agonists in urine using automated TurboFlow™ - LC/Exactive mass spectrometry. Food Additives and Contaminants, 2011, 28 (10), pp.1352-1363. 10.1080/19440049.2011.619504. hal-00743048
HAL Id: hal-00743048 https://hal.archives-ouvertes.fr/hal-00743048 Submitted on 18 Oct 2012
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Food Additives and Contaminants
This paper describes a method for the determination of priority β agonists in urine based on a fully automated sample preparation procedure using the online TurboFlow™ chromatography clean up step and determination on the Orbitrap™ mass analyzer technology. The principle of the method after enzymatic hydrolysis over night on a small column packed with a special stationary phase (TurboFlow™) while flushing away sample matrix and interfering compounds. Thereafter the analytes are transferred onto an analytical column and detected by id chromatography/high resolution mass spectrometry in full scan mode at a resolution of R=50,000 FWHM (full width at half maximum) and in HCD (Higher Energy Collisional Dissociation) scan mode at a resolving power of 10,000 FWHM. The optimization of each step of theFor Peer Review Only method, such as selection of the TurboFlowTM and analytical column as well as sample loading and elution parameters were performed using a standard solution containing salbutamol, clenbuterol and mabuterol at a veloped automated sample preparation significantly improved the throughput and efficiency of the previous used screening method and resulted in a considerable reduction in analysis time. Validation experiments including 24 β agonists in urine gave 0.35 ug/L. The repeatability of analyses for urine samples spiked at 0.5 ug/L was within the range of 5 26% and recoveries for all compounds were found to be within 89 107%.
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 1 of 26 Food Additives and Contaminants
1 2 3 4 5 6 7 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 2 of 26
1 2 Quick screening of priority beta-agonists in urine using automated 3 4 TurboFlow TM - LC/Exactive mass spectrometry 5 6 7 8 Thorsten Bernsmann a*, Peter Fürst a and Michal Godula b 9 10 11 a 12 Chemical and Veterinary Analytical Institute Münsterland Emscher Lippe, Joseph König 13 Straße 40, 48147 Münster, Germany 14 For Peer Review Only b 15 Thermo Fisher Scientific, Slunečná 27, 10000 Praha 10, Czech Republic 16 17 18 19 Abstract 20 21 22 This paper describes a method for the determination of priority β agonists in urine based on a 23 24 fully automated sample preparation procedure using an online TurboFlow TM chromatography 25 TM 26 clean up step and determination with Orbitrap mass analyzer technology. The principle of 27 28 the method was the enrichment of the β agonists after enzymatic hydrolysis overnight on a 29 small column packed with a special stationary phase (TurboFlow TM ) while flushing away 30 31 sample matrix and interfering compounds. Thereafter the analytes were transferred onto an 32 33 analytical column and detected by liquid chromatography/high resolution mass spectrometry 34 35 in full scan mode at a resolution of R=50,000 FWHM (full width at half maximum) and in 36 HCD (Higher Energy Collisional Dissociation) scan mode at a resolving power of 10,000 37 38 FWHM. The optimization of each step of the method, such as selection of the TurboFlow TM 39 40 and analytical column as well as sample loading and elution parameters were performed using 41 42 a standard solution containing salbutamol, clenbuterol and mabuterol at a concentration of 100 43 44 g/L. The developed automated sample preparation significantly improved the throughput and 45 efficiency of the previous used screening method and resulted in a considerable reduction in 46 47 analysis time. Validation experiments including 24 β agonists in urine gave decision limits 48 49 (CCα) between 0.05 0.35 g/L. The repeatability of analyses for urine samples spiked at 0.5 50 51 g/L was within the range of 5 26% and recoveries for all compounds were found to be 52 within 89 107%. 53 54 55 Key Words: Beta agonists, urine, TurboFlow TM , Orbitrap TM mass analyzer technology, 56 57 Exactive MS 58 59 60 * Corresponding author. Email: thorsten.bernsmann@cvua mel.de http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 3 of 26 Food Additives and Contaminants
1 2 3 4 5 6 Introduction 7 8 9 Beta agonists are synthetically produced compounds that, in addition to their bronchodilatory 10 and tocolytic effects, can promote live weight gain in food producing animals. There have 11 12 been documented cases when consumption of liver and meat from animals illegally treated 13 14 with clenbuterolFor has resulted Peer in serious humanReview intoxication (Botsoglou Only et al. 2001). Due to 15 16 their adverse effects, the use of clenbuterol and its analogues from the β agonist group has 17 been banned by the European Union (EU 1996) and other regulatory agencies worldwide. 18 19 Monitoring programs have shown that β agonists are still illegally used by food producers 20 21 (Fiori et al. 2002, Mazzanti et al. 2003). Moreover, newly developed analogues with modified 22 23 structures are obviously being continuously introduced in routine practice. Many papers have 24 25 been published in the past describing the analysis of β agonists in various matrices using GC 26 MS. Typically are those methods based on the determination of compounds after 27 28 derivatization by GC MS (Montrade et al. 1993, Damasceno et al. 2000, Henze et al. 2001) or 29 30 GC MS/MS (Biancotto et al. 1999, Amendola et al. 2002, Bocca et al. 2003). In most cases 31 32 the sample preparation steps include time consuming evaporation of water based re extracts 33 or solid phase extraction (SPE) clean up steps. 34 35 36 37 In order to avoid derivatization steps the recently published methods use liquid 38 39 chromatography coupled to tandem mass spectrometry (MS/MS) instruments. In the case of 40 41 liver and kidney samples, the β agonists are typically extracted from the tissue after enzymatic 42 digestion and the extracts are cleaned up using liquid liquid extraction and SPE (Fesser et al. 43 44 2005). Beta agonists from urine are generally alkaline extracted after enzymatic hydrolysis 45 46 and acidic re extraction is then performed. After the evaporation of the acidic re extract and 47 48 reconstitution in the mobile phase the analytes are determined using electrospray ionization 49 (ESI) LC MS/MS (Thevis et al. 2003). Alternatively, the clean up of the sample can be 50 51 improved using SPE and as an alternative to ESI LC MS/MS atmospheric pressure chemical 52 53 ionization (APCI) is also used for LC MS/MS determination (Dickson et al. 2005). The very 54 55 efficient alternative to the commonly used SPE approach is the use of molecularly imprinted 56 57 polymer (MIP) columns. Several papers have been published using MIP columns to 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 4 of 26
1 2 effectively remove sample matrix and to allow sub ng/mL determination of β agonists in 3 4 various matrices (Widestrand et al. 2004, Fiori et al. 2005, Kootstra et al. 2005). Although 5 6 most of the above cited approaches make it possible to reach the required detection limits their 7 main limitation is the time consuming and expensive column clean up step. 8 9 10 11 Based on the annual national production figures, the EU Commission stipulates sampling 12 13 levels and frequencies for β agonists and other veterinary drugs in animal products and 14 For Peer Review Only 15 matrices in order to check for illegally administered substances and in the case of authorized 16 compounds for compliance with prescribed withdrawal periods. Compared to time and effort 17 18 to fulfill these requirements, the number of positive findings is relatively low. This indicates 19 20 that there is a clear need for quick and simple screening methods to routinely and accurately 21 22 control levels of β agonists in samples of animal origin, such as urine, plasma, and tissues. 23 24 25 The principle of TurboFlow TM chromatography is the separation of analytes from the matrix 26 27 using specific columns packed with large particles. In combination with high linear velocity of 28 29 the mobile phase the conditions are induced on the column that allow the retention of smaller 30 31 molecules (i.e. veterinary drugs) while the large molecules (such as proteins and lipids) are 32 passing through the column unretained (Quinn and Takarewski 1997). The application of 33 34 TurboFlow TM chromatography has been already described on different examples of analyzing 35 36 specific groups of analytes in a range of matrices (milk, honey or pig tissues) (Mottier et al, 37 38 2008, Krebber et al 2009, Stolker et al. 2010). 39 40 41 The use of mass spectrometers based on high resolution and high mass accuracy 42 43 measurements offers certain benefits over traditional tandem mass spectrometry. As 44 45 documented already (Kaufman et al 2011), the selectivity and detection limits obtained by 46 high resolution mass spectrometer is comparable or better than those obtained by triple 47 48 quadrupole mass spectrometers. The advantage of the high resolution MS approach is the 49 50 simple operation with no specific compound setup and the fact that the data acquisition is 51 52 done most often in full scan mode. This allows not only to detect targeted analytes but also to 53 54 potentially search for non targeted analytes (metabolites e.g.). The importance of the various 55 factors affecting the quality of data produced by high resolution mass spectrometers, 56 57 especially the resolving power and mass accuracy achieved has been also discussed 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 5 of 26 Food Additives and Contaminants
1 2 (Kaufmann et al 2010, Kellman et al 2009).To our knowledge, till today no work has been 3 4 documented analyzing a complete set of β agonists in urine using the combination of the 5 TM 6 automated TurboFlow sample preparation and detection based on the use of high resolution 7 Orbitrap mass spectrometer. The newly developed approach results in a simplification of 8 9 current methodologies and a considerable reduction in analysis time. 10 11 12 Materials and Methods 13 14 For Peer Review Only 15 Samples 16 The urine samples from cattle and pigs were routinely taken within the frame of the national 17 18 residue control plan. Until analysis t he samples were stored at 18 °C. 19 20 Chemicals and Reagents: 21 22 Acetonitrile, methanol, isopropanol (all picograde) were purchased from Promochem (Wesel, 23 Germany) . Ultrapure water was produced by reverse osmoses. Acetic acid glacial (purity AR 24 25 A) and ammonium acetate (purity AR U) were obtained from Biosolve (Valkenswaard, The 26 27 Netherlands) and an aqueous solution of ß glucuronidase Type H 2 (from Helix pomatia ) with 28 29 an activity of 130,200 units/mL came from Sigma Aldrich (St. Louis, USA). 30 31 32 Reference standards 33 34 Beta agonist standards: clenbuterol, fenoterol, isoxsuprine, ritodrin, salbutamol, salmeterol, 35 36 terbutalin were purchased from Sigma Aldrich (St. Louis, USA). Brombuterol, 37 38 chlorbrombuterol, cimbuterol, clencyclohexerol, clenhexerol, clenpenterol, clenproperol, 39 hydroxymethylclenbuterol, mabuterol, mapenterol, tulobuterol were purchased from Witega 40 41 (Berlin, Germany). Ractopamin hydrochlorid was purchased from Dr. Ehrensdorfer 42 43 (Augsburg, Germany). Carbuterol, metoprolol, pirbuterol and zilpaterol were obtained from 44 45 the former Federal Institute for Consumer Health Protection and Veterinary Medicine (now 46 Federal Office of Consumer Protection and Food Safety, Berlin, Germany). Cimaterol was 47 48 purchased from Tocris bioscience (Bristol, UK). 49 50 51 52 Sample Preparation: 53 54 Urine samples were centrifuged at 3500 rpm for 10 min. One mL of the supernatant was 55 diluted with 580 L water, 400 L 50 mM ammonium acetate buffer (pH 5.2) and 20 L ß 56 57 glucoronidase solution. The mixture was kept at 37 °C over night in a shaker. After the 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 6 of 26
1 2 hydrolysis step the mixture was centrifuged again at 3500 rpm. The supernatant was filtrated 3 4 through a 0.45 m filter into an autosampler vial. 5 6 7 TLX-Exactive analysis 8 9 The online clean up of urine samples was carried out using a TurboFlow TM system TLX 10 11 Aria 1 (Thermo Fisher Scientific, San Jose, CA, USA) consisting of loading and eluting 12 TM 13 pumps, two valve switching modules and automatic liquid sampler. TurboFlow online 14 For Peer Review Only 15 sample preparation was carried out using different columns: C18 P XL 0.5 mm I.D x 50mm 16 length; C18 XL 0.5x50mm; C8 0.5x50mm, Cyclone 0.5x50mm, Cyclone P 0.5x50mm; 17 18 Cyclone MAX 0.5x50mm (all Thermo Fisher Scientific, Franklin, MA, USA). Separation of 19 20 the compounds was carried out on a reversed phase phenyl hexyl analytical column (150 mm 21 22 x 4.6 mm x 3.5 m particle size (Eclipse Plus Phenyl Hexyl, Agilent Technologies, Santa 23 Clara, CA, USA). The injected sample volume was 50 to 100 L. 24 25 26 TM 27 The online TurboFlow system Aria 1 (schematics shown in Figure 1) was connected to a 28 TM 29 single stage mass spectrometer LC MS Exactive with an HCD cell, equipped with the 30 31 heated electrospray interface HESI II (Thermo Fisher Scientific, San Jose, CA, USA). The 32 Exactive was operated in positive ionization mode using the following MS parameters: 33 34 electrospray, voltage: 4.0 kV; sheath gas: 65 arbitrary units; auxiliary gas: 15 arbitrary units; 35 36 capillary temperature 250 °C, heater temperature 300 °C, capillary voltage 60 V, tube lens 37 38 voltage 120 V. These settings are a compromise between the different source parameters for 39 each analyte and gave the best sensitivity over the whole group. The other instrument 40 41 parameters were automatically tuned for maximum ion signal. The system was operated in full 42 43 scan mode at a resolving power of 50,000 FWHM and in HCD scan mode at a resolving 44 45 power of 10,000 FWHM with an HCD voltage of 30 V. No specific lock mass was used for 46 internal mass axis correction. Mobile phases used in the Aria 1 system were as follows: 47 48 Loading pump: 49 50 A: 10 mM ammonium acetate in water, adjusted to pH 8 with ammonium hydroxide 51 52 B: methanol 53 54 C: isopropanol:acetonitrile:cyclohexane (30:30:40, v/v/v) 55 D: 0.1% acetic acid in water 56 57 Eluting pump: 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 7 of 26 Food Additives and Contaminants
1 2 A: 0.1% acetic acid in water 3 4 B: methanol 5 TM 6 The typical conditions used for the online sample clean up on a TurboFlow column and the 7 chromatographic separation on the analytical column are shown in Table 1. 8 9 10 11 12 Results and Discussion 13 14 For Peer Review Only 15 The simplification and increased throughput of current analytical methods is of major concern 16 in many labs performing routine analyses for residues of veterinary drugs in samples of 17 18 animal origin. The automated online sample preparation technique based on the high 19 20 turbulent flow liquid chromatography (HTLC) has the potential to offer significant advantage 21 22 over manually performed procedures. The principle of the approach is the enrichment of 23 analytes on a small column packed with special stationary phase (TurboFlow TM ) while 24 25 flushing away sample matrix and interfering compounds. Thereafter, the analytes are 26 27 transferred onto an analytical column and detected by mass spectrometry as shown in Figure 28 29 1. However, before successful application in routine analysis, it is necessary to optimize the 30 31 performance of the whole system. 32 33 34 The most important steps in the optimization procedure are: 35 TM 36 (I) TurboFlow column selection – evaluation of different types of columns with respect 37 38 to the retention of analytes 39 (II) Sample loading parameters – optimization of the pH of the loading mobile phase to 40 41 retain the analytes on the TurboFlow TM column 42 43 (III) Sample elution parameters – optimization of the elution mobile phase pH, flow rate 44 TM 45 and time during the elution step from the TurboFlow column 46 (IV) Analytical column focusing parameters – optimization of initial gradient composition 47 48 to effectively trap and focus all compounds of interest during the chromatographic 49 50 separation step. 51 52 53 54 The full optimization of each step of the method was performed using a standard solution 55 containing salbutamol, clenbuterol and mabuterol each at a level of 100 g/L. The three 56 57 compounds were selected from all conceivable β agonists due to their physico chemical 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 8 of 26
1 2 properties and strong dependency on the pH of the mobile phase which substantially 3 4 influences their behavior in the chromatographic system. 5 6 7 The selection of the TurboFlow TM column with a suitable stationary phase was the first 8 9 optimization step. The following TurboFlow TM columns were tested with respect to their 10 11 retention capabilities: 12 13 C18 P XL 0.5x50mm 14 For Peer Review Only 15 C18 XL 0.5x50mm 16 C8 0.5x50mm 17 18 Cyclone 0.5x50mm 19 20 Cyclone P 0.5x50mm 21 22 Cyclone MAX 0.5x50mm 23 24 25 The standard solution containing the three compounds was injected on the different 26 TM 27 TurboFlow columns using water with 10 mM of ammonium acetate as loading solvent at a 28 29 flow rate of 1.5 mL/min. Elution from the column was performed with a 200 L loop filled 30 31 with a solution of acetonitrile:water 80/20 v/v with 0.1% formic acid added. The extracted ion 32 chromatograms (EIC) of the [M+H] + ions for the three test compounds salbutamol, 33 34 clenbuterol and mabuterol shown in Figure 2 document how the column selectivity influences 35 36 the retention of the tested compounds. From all tested columns, the Cyclone P column was 37 38 identified as the most promising one due to the fact that only a small portion of salbutamol 39 was not retained. The other compounds (mabuterol and clenbuterol) were fully retained. On 40 41 all other columns salbutamol was not retained at all due to its amphoteric properties leading to 42 43 charging of the compound under the applied conditions. 44 45 46 The next step to optimize was the composition of the loading mobile phase. As already 47 48 mentioned, due to the amphoteric nature of some β agonists, it was necessary to select the 49 50 optimal pH of the mobile phase used during this step. 51 52 The following mobile phase compositions were tested at a flow rate of 1.5 mL/min: 53 54 water + 0.1% formic acid, pH 3 55 water + 10 mM ammonium acetate, adjusted to pH 6 56 57 water + 10 mM ammonium acetate, adjusted to pH 8 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 9 of 26 Food Additives and Contaminants
1 2 The chromatograms shown in Figure 3 demonstrate how the pH influences the retention of 3 4 compounds on the Cyclone P column. Clenbuterol and mabuterol are fully retained at all pH 5 6 conditions tested. However, at lower pH values (3 6) salbutamol is not fully retained on the 7 column. The efficiency of salbutamol retention is only about 60%. At pH 8 more than 85% of 8 9 salbutamol stays on the column. For this reason, a loading phase with pH 8 was found optimal 10 11 for all tested analytes. 12 13 14 For Peer Review Only 15 After successful enrichment of the analytes it is important to provide their efficient transfer 16 onto an analytical column. The transfer from the TurboFlow TM column to an analytical 17 18 column is achieved by the elution with a solvent located in the fixed loop with a typical 19 20 volume of 100 200 L (see Figure 1). The solvent filled in the loop typically consists of the 21 22 organic phase (methanol or acetonitrile) with a certain water percentage. The aim of this 23 optimization step is to use the mixture with the lowest possible organic phase percentage and 24 25 to accomplish the elution with the lowest possible flow rate. Those two conditions are 26 27 required to allow efficient focusing of the analytes at the head of the analytical column. To 28 29 optimize the elution composition and flow rate a mobile phase containing methanol:water + 30 31 0.1 % formic acid at different ratios was used. Methanol was chosen due to its higher polarity 32 compared to acetonitrile, expected better solubility for compounds and better compatibility 33 34 with LC MS solvents. 35 36 37 38 The solvent elution ratios used were from 100% to 50% of methanol in 10% increments. The 39 results of the optimization are illustrated in Figure 4. When the percentage of methanol in the 40 41 elution solvent dropped down to 70%, clenbuterol was no longer completely eluted from the 42 TM 43 TurboFlow column. At 50% of methanol about 25% of clenbuterol was retained on the 44 45 column. Consequently, the flow rate of the mobile phase during elution was tested and it was 46 found that the elution of all compounds from the column took about 1 min when a flow rate of 47 48 100 L/min was applied. At 75 L/min this process took already 2 min. Therefore, for the 49 50 next experiments, a flow rate of 100 L/min was selected as optimal condition and 70% of 51 52 methanol in the eluting solvent was used. 53 54 55 The last step in the method optimization was the evaluation of the combinations of flow rates 56 57 and mobile phase compositions to efficiently focus analytes in the front part of the analytical 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 10 of 26
1 2 column. The initial conditions tested were: 3 4 5 6 Loading pump: methanol : water + 0.1 % formic acid (70:30, v/v), flow rate 100 L/min 7 Eluting pump: methanol : water with 0.1% formic acid (2:98, v/v), flow rate 700 L/min. 8 9 Both solvent flows were combined in the T piece on the valve 2. Using these conditions, a 10 11 peak broadening of salbutamol was observed on the analytical column, indicating improper 12 13 focusing. Further optimization was then performed to improve the focusing step to obtain 14 For Peer Review Only 15 good peak shapes for all compounds. Different mobile phases, pH, flow ratios from eluting 16 and analytical pump and percentages of the organic phase in the HPLC mobile phase were 17 18 tested. However, at all settings employed, the peak shape of salbutamol was still not 19 20 satisfactory. Consequently, the modification of the mobile phase with acetic acid instead of 21 22 formic acid was tested. Probably due to its better interactions with the stationary phase of the 23 phenyl based column, acetic acid at the concentration levels of 0.1% in the water phases 24 25 significantly improved the focusing step and provided excellent peak shapes for all analyzed 26 TM 27 agonists (see Figure 5). The final conditions that were optimized for the online TurboFlow 28 29 sample preparation are summarized in Table 2. 30 31 32 Method validation 33 34 It is necessary to ensure the quality and comparability of analytical results generated by a 35 36 newly developed method before its use in official residue control. This should be achieved by 37 38 using quality assurance systems and especially by performing a validation according to 39 internationally recognized and harmonized procedures and performance criteria. Thus, 40 41 following the optimization of the TurboFlow TM and HPLC column conditions, the method 42 43 was thoroughly validated by following validation criteria laid down in Decision 2002/657/EC . 44 45 Where the Decision has set no criteria, such as for using a high resolution mass spectrometer 46 coupled with liquid chromatography, respective criteria were taken from the Document 47 48 SANCO/10684/2009 which lays down the method validation and quality control procedures 49 50 for pesticide residues analysis in food and feed. For the identification of the different analytes 51 + 52 the retention time and the [M+H] ion in full scan mode (based on Decision 2002/657/EC) at 53 54 a resolution of 50,000 FWHM were used. For confirmation purposes other ions, such as [M 55 + H2O +H] or other adduct ions like sodium or ammonium in full scan mode or a product ion in 56 57 HCD mode were utilized. The most sensitive ions with their exact masses used for 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 11 of 26 Food Additives and Contaminants
1 2 identification (mass accuracy < 5 ppm; one fragment ion), unequivocal confirmation and 3 4 quantification are listed in Table 3. Extracted ion chromatograms of the recorded ions of a 5 6 standard solution (0.5 g/L) and a spiked urine sample (0.1 g/L) are shown in Figures 6 & 7. 7 8 9 A matrix considering validation protocol was used which takes into account the uncertainty 10 11 due to a potential matrix influence. (Jülicher et al. 1998; 1998 a; Gowik et. al. 1998). To build 12 13 up the experimental design and to calculate all relevant validation parameters the software 14 For Peer Review Only 15 program “Interval” from the company “quo data” was used. The factors that may influence the 16 measurements were determined based on practical considerations. Two factors were chosen 17 18 which could have an impact on the analytical results. 19 20 21 22 As can be seen from Table 4, the validation experiment considered species (urine from 23 different cows and pigs collected in the frame of the National Residue Control Plan) as a 24 25 matrix related factor and two analysts as leading factors. For this design a factorial plan was 26 27 calculated by the programme, which required a randomised performance of the experiment. 28 29 The experimental plan was replicated for each concentration level. The concentration levels, 30 31 which were chosen for each compound (in matrix), were 0.00 (blank sample), 0.10, 0.25, 0.50, 32 1.0, 1.5 and 2.0 g/L. For this purpose, pretested blank urine samples from cows and pigs 33 34 were spiked with the different concentrations of the β agonists. After spiking, the urine 35 36 samples were prepared in the same way as described above. In total, the validation study was 37 38 based on 112 different analyses for each substance. The model allows the estimation of 39 critical concentrations for given α errors and the calculation of the corresponding power 40 41 function which are important to evaluate the performance of an analytical method. The 42 43 relevant parameters are: 44
45 46 • Decision limit (CC α): This means the limit at and above which it can be concluded 47 48 with an error probability of α that a sample is non compliant (Decision 2002/657/EC) 49 50 • Detection capability (CC β): This describes the smallest content of the substance that 51 52 may be detected, identified and/or quantified in a sample with an error probability of β. 53 In the case of substances for which no permitted limits have been established, the 54 55 detection capability is the lowest concentration at which a method is able to detect 56 57 truly contaminated samples with a statistical certainty of 1 – β (Decision 2002/657/EC) 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 12 of 26
1 2 • In house validation (Decision 2002/657/EC) 3 4 5 6 Table 5 summarizes the data obtained from the method validation study. The decision limits 7 8 and detection capabilities calculated range from 0.05 0.35 g /L and from 0.13 0.70 g /L 9 urine, respectively. The repeatability of sample analyses spiked at 0.5 g /L is within the 10 11 range of 5 26% whilst within laboratory reproducibility (precision obtained in the same 12 13 laboratory under stipulated conditions concerning e.g. method, test materials, operators, 14 For Peer Review Only 15 environment over long time intervals) range between 11 31%. Recoveries for all compounds 16 were between 89 and 107%. The TLX system showed less than 1 % cross contamination. The 17 18 main source for the cross contamination was the auto sampler where an intensive washing 19 20 procedure with an acetic acid/methanol solution after each injection is necessary, which can 21 22 be done during the chromatographic run. 23 24 25 The validation has proven that the developed method is fit for purpose. It allows the 26 27 determination of 24 β agonists in urine at the concentration of interest whilst considerably 28 29 reducing the manual work load for the laboratory technicians. Moreover, due to the substantial 30 31 shorter analysis time as a result of the unattended automatic sample preparation process, the 32 laboratory’s sample throughput can be significantly increased. According to our calculations, 33 34 the overall sample preparation time excluding the enzymatic hydrolysis which is necessary to 35 36 release bound forms of β agonists is approximately 30 min. This is more than 3x less time 37 38 than what is needed for a typical method based on simple sample preparation (Thevis et al 39 2003). It is worthy of mention that the flexibility of the overall system makes it possible to 40 41 perform also traditional HPLC or UHPLC separations without any hardware modifications. 42 43 As already mentioned in the introduction, the possibility of acquiring the MS data in full scan 44 45 mode offers the future extension of the method to screen for other groups of analytes as well 46 47 as search for metabolites or degradation products. 48 Acknowledgements 49 50 The authors would like to acknowledge Dr. Francois Espourteille from Thermo Fisher 51 TM 52 Scientific, Franklin, MA (USA) for his help with the initial TurboFlow method 53 54 optimization and column selection 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 13 of 26 Food Additives and Contaminants
1 2 References 3 4 5 Amendola L, Colamonici C, Rossi F, Botre F. 2002. Determination of clenbuterol in human 6 7 urine by GC–MS–MS–MS: confirmation analysis in antidoping control . Journal of 8 9 Chromatography B 773 : 7 16 10 11 Biancotto G, Angeletti R, Traldi P, Guidugli F. Positive chemical ionization and tandem mass 12 13 spectrometric fragmentation for the gas chromatographic analysis of beta agonists using the 14 ion trap technique.For 1999. JournalPeer of Mass ReviewSpectrometry 34: 1346 53 Only 15 16 17 Bocca B, Fiori M, Cartoni C, Brambilla G. 2003. Simultaneous determination of Zilpaterol 18 and other beta agonists in calf eye by gas chromatography/tandem mass spectrometry Journal 19 20 of AOAC International 86: 8 14 21 22 23 Botsoglou, N.A., Fletouris D.J., Drug Residues in Food. Pharmacology, Food Safety and 24 Analysis, Marcel Dekker: New York, 2001 25 26 27 Damasceno L, Ventura R, Ortu˜no J, Segura J. 2000. Derivatization procedures for the 28 detection of beta(2) agonists by gas chromatographic/mass spectrometric analysis. Journal of 29 30 Mass Spectrometry 35 : 1285 94 31 32 Dickson L, MacNeil JD, Lee S, Fesser A. 2005. Determination of β agonist residues in bovine 33 34 urine using liquid chromatography tandem mass spectrometry. Journal of AOAC International 35 36 88(1) : 46 56 37 38 EU COUNCIL DIRECTIVE 96/22/EC of 29 April 1996, OJ L 125, 23.5.1996 39 40 41 Fesser A, Dickson LC, MacNeil JD, Patterson JR, Lee S, Gedir R. 2005 Determination of β 42 agonists in liver and retina by liquid chromatography tandem mass spectrometry Journal of 43 44 AOAC International 88 (1): 61 69 45 46 Fiori M, Cartoni C, Bocca B, Brambilla G. 2002. The Use of Nonendcapped C18 Columns in 47 48 the Cleanup of Clenbuterol and a New Adrenergic Agonist from Bovine Liver by Gas 49 50 Chromatography–Tandem Mass Spectrometry Analysis. Journal of Chromatographic Science 51 52 40: 92 96 53 54 Fiori M, Civitareale C, Mirante S, Magarò E, Brambilla G. 2005 Evaluation of two different 55 56 clean up steps, to minimise ion suppression phenomena in ion trap liquid chromatography– 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 14 of 26
1 2 tandem mass spectrometry for the multi residue analysis of beta agonists in calves urine. 3 4 Analytica Chimica Acta , 529(1-2): 207 210 5 6 Gowik P, Jülicher B, Uhlig B. Multi residue method for non steroidal anti inflammatory 7 8 drugs in plasma using high performance liquid chromatography–photodiode array detection: 9 10 Method description and comprehensive in house validation. Journal of Chromatography B 11 12 716 (1998) : 221 232 13 14 Henze MK, OpfermannFor G, PeerSpahn Langguth Review H, Schänzer W. 2001. Only Screening of beta 2 15 16 agonists and confirmation of fenoterol, orciprenaline, reproterol and terbutaline with gas 17 chromatography mass spectrometry as tetrahydroisoquinoline derivatives. Journal of 18 19 Chromatography B 751: 93 105 20 21 Jülicher B, Gowik P, Uhlig S. 1998. Assessment of detection methods in trace analysis by 22 23 means of a statistically based in house validation concept. Analyst 123: 173 179 24 25 Jülicher B, P Gowik P, Uhlig S. 1998 a. A top down in house validation based approach for 26 27 the investigation of the measurement uncertainty using fractional factorial experiments. 28 29 Analyst 124: 537 545 30 31 Kaufmann A, Butcher P, Maden K, Walker S, Widmer M. 2011. Quantitative and 32 33 confirmative performance of liquid chromatography coupled to high resolution mass 34 35 spectrometry compared to tandem mass spectrometry Rapid Communication in Mass 36 37 Spectrometry, 25, 1–14 38 39 Kaufmann A, Butcher P, Maden K, Walker S, Widmer M. 2011. Development of an 40 41 improved high resolution mass spectrometry based multi residue method for veterinary drugs 42 in various food matrices 2010 Anaytica Chimica Acta, doi: 10.1016/j.aca.2010.11.034 43 44 45 Kellmann M, Muenster H, Zomer P, Mol HJ. 2009. Full Scan MS in Comprehensive 46 47 Qualitative and Quantitative Residue Analysis in Food and Feed Matrices: How Much 48 Resolving Power is Required? Journal of the American Society for Mass Spectrometry 20 (8): 49 50 1464–1476 51 52 Kootstra PR, Kuijpers CJPF, Wubs KL, Van Doorn D, Sterk SS, Van Ginkel LA, Stephany 53 54 RW. 2005. The analysis of beta agonists in bovine muscle using molecular imprinted 55 56 polymers with ion trap LCMS screening. Analytica Chimica Acta 529(1-2): 75 81 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 15 of 26 Food Additives and Contaminants
1 2 Krebber R, Hoffend FJ, Ruttmann F. 2009. Simple and rapid determination of enrofloxacin 3 4 and ciprofloxacin in edible tissues by turbulent flow chromatography/tandem mass 5 6 spectrometry (TFC–MS/MS). Analytica Chimica Acta 637 , 208–213 7 8 Mazzanti G, Daniele C, Boatt G, Manca G, Brambilla G, Loizzo A. 2003. New β adrenergic 9 10 agonists used illicitly as growth promoters in animal breeding: chemical and 11 12 pharmacodynamic studies. Toxicology 187 : 91 99 13 14 Montrade MP, ForLe Bizec B, Peer Monteau F, SiliartReview B, Andre F. 1993. Only Multiresidue analysis for β 15 16 agonistic drugs in urine of meat producing animals by gas chromatography mass 17 spectrometry. Analytica Chimica Acta 275: 253 268 18 19 20 Mottier P, Hammel YA, Gremaud E, Guy PA. 2008. Quantitative high throughput analysis of 21 16 (fluoro)quinolones in honey using automated extraction by turbulent flow chromatography 22 23 coupled to liquid chromatography tandem mass spectrometry. Journal of Agricultural and 24 25 Food Chemistry 56(1): 35 43 26 27 Quinn HM, Takarewski JJ 1997, Int. PatentWO 97/16724 28 29 30 Stolker A, Peters R, Zuiderent R, DiBussolo JM, Martins C. 2010. Fully automated screening 31 of veterinary drugs in milk by turbulent flow chromatography and tandem mass spectrometry. 32 33 Analytical Bioanalytical Chemistry , 397(7) : 2841–2849. 34 35 Thevis M, Opfermann G, Schänzer W. 2003. Liquid chromatography/electrospray ionization 36 37 tandem mass spectrometric screening and confirmation methods for b2 agonists in human or 38 39 equine urine. Journal of Mass Spectrometry 38 :1197 1206 40 41 Widstrand C, Larsson F, Fiori M, Civitareale C, Mirante S, Brambilla G. 2004. Evaluation of 42 43 MISPE for the multi residue extraction of β agonists from calves urine. Journal of 44 45 Chromatography B 804 (1): 85 91 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 16 of 26
1 2 Figure 1. Schematics of the online TurboFlow TM clean up system 3 4 5 6 7 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 * Corresponding author. Email: thorsten.bernsmann@cvua mel.de http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 17 of 26 Food Additives and Contaminants
1 2 Figure 2. The influence of TurboFlow TM column chemistry on the retention of the β agonists 3 4 used for optimization (extracted ion chromatograms, EIC). Columns with different stationary 5 6 phases were used (C18P, C18, C8, Cyclone, Cyclone P, Cyclone Max). Cyclone P shows the 7 best retention behavior for all compounds. 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 18 of 26
1 2 Figure 3. The pH optimization of the mobile phase chosen for the loading of the β agonists 3 TM 4 used for optimization (extracted ion chromatograms, EIC) on the TurboFlow column. 5 6 Increasing the pH from 3 to 8 (left to right) improves the retention of salbutamol significantly. 7 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 19 of 26 Food Additives and Contaminants
1 2 Figure 4. Loop filling optimization (100% to 50% of methanol in the loop) for the β agonists 3 4 used for optimization (extracted ion chromatograms, EIC). At 70% methanol clenbuterol is 5 6 not fully eluted anymore. 7 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 20 of 26
1 2 Figure 5. Improper mobile phases composition showing poor refocusing of the analytes 3 4 (extracted ion chromatograms, EIC) on the analytical column (left). Peak shapes after 5 6 optimization step (right) 7 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 21 of 26 Food Additives and Contaminants
1 2 Figure 6. EICs of all 24 analysed β agonists in a standard solution at a concentration of 0.5 3 4 ug/L. (numbering of analyts see Table 3). 5 6 7 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 22 of 26
1 2 Figure 7. EICs of all 24 analyzed β agonists in a urine sample spiked at 0.1 ug/L. (numbering 3 4 of compounds see Table 3) 5 6 7 8 9 10 11 12 13 14 For Peer Review Only 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 23 of 26 Food Additives and Contaminants
1 2 Table 1: TurboFlow TM loading and eluting conditions 3 4 Valve Valve 5 Loading pump Eluting pump 6 1 2 7 Step Start Sec Flow Grad %A %B %C %D Tee Loop Flow Grad %A %B 8 9 1 00:00 30 1.5 Step 100 Out 0.8 Step 98 2 10 2 00:30 120 0.1 Step 100 T In 0.9 Step 98 2 11 12 3 02:30 30 1.0 Step 80 20 In 0.8 Ramp 70 30 13 4 03:00 120 1.0 Step 80 20 In 0.8 Ramp 55 45 14 For Peer Review Only 15 5 05:00 320 1.0 Step 80 20 In 0.8 Ramp 20 80 16 6 10.20 180 1.0 Step 70 30 In 0.8 Ramp 5 95 17 18 7 13.20 120 1.0 Step 70 30 In 0.8 Step 5 95 19 20 8 15.20 160 1.5 Step 100 Out 0,8 Step 98 2 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 * Corresponding author. Email: thorsten.bernsmann@cvua mel.de http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 24 of 26
1 2 Table 2: Optimized conditions for online TurboFlow TM sample preparation 3 TM 4 TurboFlow Column Cyclone P, 0.5 x 50 mm 5 6 Loading conditions water with10 mM NH 4Ac, pH 8, 1.5 mL/min, 30 s 7 8 Eluting conditions methanol : 0.1% acetic acid, 70/30, 0.10 mL/min, 120 s 9 Wash step 1. methanol/0.1% acetic acid, 80:20, 1.5mL/min 10 11 2. acetonitrile/acetone/isopropanol; 30/30/40 (v/v/v), 1.5 12 13 mL/min, 30 s 14 For Peer Review Only 15 Analytical column Eclipse Plus Phenyl Hexyl, 4.6 mm x 150 mm x 3.5 m 16 17 HPLC mobile phase A: water + 0.1% acetic acid; B: methanol 18 For gradient see Table 1 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 25 of 26 Food Additives and Contaminants
1 2 Table 3: List of analysed compounds with molecular formula and exact masses for 3 identification and quantification 4 5 6 Molecular Quant. Qual. HCD 7 No. Name formula Ion Ion Ion 8 1 clenproperol C11H16N2OCl2 263.0713 265.0364 245.0607 9 10 2 brombuterol C12H18N2OBr2 366.9818 366.9838 350.9729 11 3 salbutamol C13H21NO3 240.15940 148.0757 12 13 4 mabuterol C13H18N2OClF3 311.1133 313.1104 237.0400 14 5 clenbuterol For Peer C12H18N2OCl2 Review 277.0869 Only 279.0837 203.0138 15 16 6 ractopamin C18H23NO3 302.1751 284.1645 284.1645 17 7 terbutalin C12H19NO3 226.1438 152.0705 152.0705 18 8 isoxsuprine C18H23NO3 302.1751 284.1645 284.1645 19 20 9 pirbuterol C12H20N2O3 241.15467 223.1441 21 10 fenoterol C17H21NO4 304.1543 326.1363 286.1438 22 23 11 salbutamol C12H18NOCl 228.1150 230.1119 154.0415 24 12 zilpaterol C14H19N3O2 262.1550 244.1440 202.0976 25 26 13 cimbuterol C13H19N3O 234.1601 160.0867 160.0867 27 14 ritodrin C17H21NO3 288.15940 270.1489 270.1489 28 29 15 clencyclohexerol C14H20N2O2Cl2 319.0975 321.0942 301.0868 30 16 clenpenterol C13H20N2OCl2 291.1026 293.0996 203.0682 31 32 17 mapenterol C14H20N2OClF3 325.1289 327.0396 217.0337 33 18 carbuterol C13H21N3O3 268.1656 194.0916 34 35 19 clenhexerol C14H22N2OCl2 305.1182 307.1154 287.1076 36 20 chlorbrombuterol C12H18N2OClBr 323.0343 321.0364 248.9608 37 38 21 salmeterol C25H37NO4 416.27954 438.2615 398.2690 39 22 cimaterol C12H17N3O 220.14440 143.0602 40 23 metoprolol C15H25NO3 268.19072 250.1802 250.1802 41 42 24 hydroxymethylclenbuterol C12H18Cl2N2O2 293.08240 295.0790 275.0712 43 Quant. Ion: Quantification ion is the most intensive ion in the full scan experiment. Mostly it 44 + 45 is the [M+H] ion. 46 Qual. Ion: Qualification ion 47 HCD Ion is the most intensive ion in the HCD experiment. 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Food Additives and Contaminants Page 26 of 26
1 2 Table 4: Details of validation experiment 3 4 Analyte(s): brombuterol; carbuterol; chlorbrombuterol; cimaterol; 5 cimbuterol; clenbuterol; clencyclohexerol; clenhexerol; 6 clenisopenterol; clenpenterol; clenproperol; fenoterol; 7 8 hydroxymethylclenbuterol; isoxsuprin; mabuterol; 9 mapenterol; metoprolol; pirbuterol; ractopamin; ritodrin; 10 salbutamol; salmeterol; terbutalin; tulobuterol; zilpaterol 11 Number of factors: 1 + 1 leading factor 12 Number of leading factors 2 13 14 Number of runs:For Peer 16 Review Only 15 Number of concentrations 7 16 Total number of analyses: 112 17 Runs per day: 8 18 19 Leading factor: persons (2) 20 matrix related factor: species (urine from different cows and pigs) 21 Sample preparation dilution 22 technique: 23 TM 24 Analytical technique: TurboFlow HRMS (Exactive) 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected] Page 27 of 26 Food Additives and Contaminants
1 2 Table 5: Results of the method validation 3 4 Within laboratory 5 CC α CC β Repeatability (%) reproducibility (%) Recovery 6 Compound name ( g/L) ( g/L) (0.5 g/L) (0.5 g/L) (%) 7 8 brombuterol 0.19 0.46 20 24 97 9 carbuterol 0.22 0.58 25 27 100 10 chlorbrombuterol 0.14 0.38 14 26 99 11 12 cimaterol 0.13 0.35 15 19 100 13 cimbuterol 0.08 0.20 9 13 100 14 Forclenbuterol Peer 0.14 0.33 Review 13 Only 25 98 15 16 clencyclohexerol 0.08 0.21 9 12 100 17 clenhexerol 0.22 0.49 22 23 98 18 clenpenterol 0.20 0.49 22 25 99 19 20 clenproperol 0.10 0.24 11 15 100 21 fenoterol 0.11 0.24 11 14 99 22 hydroxymethylclenbuterol 0.09 0.21 9 14 100 23 24 isoxsuprine 0.10 0.25 11 18 100 25 mabuterol 0.09 0.23 9 17 100 26 mapenterol 0.10 0.24 10 18 99 27 28 metoprolol 0.10 0.30 11 25 100 29 pirbuterol 0.35 0.70 11 29 89 30 ractopamin 0.05 0.13 5 13 100 31 32 ritodrin 0.05 0.13 6 12 100 33 salbutamol 0.24 0.60 26 27 103 34 salmeterol 0.16 0.42 18 21 99 35 36 terbutalin 0.20 0.52 22 26 101 37 tulobuterol 0.10 0.32 11 31 107 38 zilpaterol 0.08 0.21 9 14 100 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
http://mc.manuscriptcentral.com/tfac Email: [email protected]