APPLICATINS Rapid, High Resolution Analysis of Aflatoxin Extracts from Peanut Butter Using Kinetex Core-Shell Technology and Strata SPE Sky Countryman, Shahana Huq, Terrell Mathews, et al. Phenomenex, Inc., 411 Madrid Ave., Torrance, CA 90501 USA A two-stage Solid Phase Extraction (SPE) procedure using Strata Florisil and Strata Silica is effective in removing detector interfering contaminants from the difficult sample matrix, peanut butter, while maintaining absolute recoveries above 80 %. Improved separation of aflatoxins is achieved with Kinetex 2.6 µm PFP core-shell columns, compared to traditional C18 columns, allowing for rapid run times under 2 minutes with great precision, and accuracy. Introduction Mycotoxins are secondary metabolites generated by several kinds of molds. Aflatoxins are a subclass of mycotoxins mainly produced by Aspergillus flavus and Aspergillus parasiticus which can contaminate a wide spectrum of foodstuffs. Due to their high toxicity and carcinogenicity, aflatoxins are of major concern for food producers, the food processing industry, and consumers. Most countries have legislation setting maximum permissible limits for aflatoxins, which are in the low micrograms per kilogram for food matrices. 1,2 Two types of limits exist, one for aflatoxin B1 and the other for the sum of the four aflatoxins (B1, B2, G1 and G2). The most frequently used analytical protocol for aflatoxins is HPLC coupled with either fluorescence or mass spectrometric detection. 3,4 The toxins are solvent extracted from food matrices and the extracts cleaned either by immunoaffinity (IAC) or solidphase extraction (SPE) columns prior to analysis. Sample cleanup with IAC suffers from several disadvantages, the major ones being inactivation of the affinity sites or masking of the toxin structure by matrix contaminants from food samples resulting in lower recovery yields. 5 Several sorbents have been used for the SPE cleanup of solvent extracts from food matrices including C18, silica, florisil, phenylsilica and functionalized polymers. It has been reported that SPE provides cleanliness comparable to IAC. In this technical note, we demonstrate a new SPE protocol for aflatoxins analysis from peanut butter which gives ultra-clean extracts. Samples are analyzed by LC/MS/MS using the new Kinetex core-shell particle technology achieving full resolution of all compounds in less than 2 minutes. LC/MS/MS systems were used: Applied Biosystems API 3000 with TurboIonSpray Source and 3200 Q TRAP with Turbo V Ion Source. To 5 g of peanut butter, 40 ml of methanol/water (80:20) containing 0.2 g of sodium chloride was added and the mixture was mechanically stirred for 2 hours. The residual solids were filtered off on a Whatman filter paper and rinsed three times with 5 ml methanol. The combined extracts were dried over anhydrous magnesium sulfate, the solvent removed under nitrogen at 45 C and the residue reconstituted in 500 µl of methanol/water (80:20). Since this process was quite time consuming, peanut butter matrix extracts were prepared by extracting larger quantities of un-spiked peanut butter (50-100 g). These extracts were refrigerated and then could be spiked with a known amount of aflatoxin as needed. Extracts were spiked with aflatoxin standards at 50 ppb for LC/MS/ MS for use with SPE. Data generated on the API 3000 LC/MS/MS was done at Phenomenex, Torrance, CA. Work on the API 3200 Q TRAP was done at AB/ MDS Analytical Technologies in Toronto, Canada. Results and Discussion Peanut butter is extremely complex consisting of carbohydrates, proteins, vitamins, phytosterols, poly/mono-unsaturated fatty acids and several inorganic ingredients. Solvent extraction will co-extract many of these constituents with the aflatoxins. Therefore, a thorough sample cleanup protocol is required to eliminate these matrix components, especially when parts per billion level aflatoxins are to be quantitated. In addition, the lactone ring present in aflatoxins can be cleaved under basic conditions. The G1 and G2 isomers are especially prone to this because they contain two such rings (Figure 1). Moreover, aflatoxins possess high melting points and are stable to heat in the dry condition, but when moisture is present, heating may lead to cleavage of the ring as well as decarboxylation. n the other hand, strong acids convert aflatoxins to adducts, as for example, trifluoroacetic acid forms the addition compound across the double bond of the furan ring. Thus, care must be execised in isolating the aflatoxins before the analytical step Experimental Conditions Aflatoxin standards were obtained from Sigma, St. Louis, M. Peanut butter with no aflatoxins was purchased from a local grocery store. The Strata Florisil and Silica cartridges as well as the Kinetex PFP columns were obtained from Phenomenex, Torrance, CA. All solvents and buffers are reagent grade from Sigma. Two Page 1 of 8
APPLICATINS Figure 1. Structures of Aflatoxins and their Hydrophobicity Indices (log P) and Molecular Weights Aflatoxin B 1 CH 3 log P= 1.22 MF= C17H1206 MW= 312 Aflatoxin B 2 CH 3 log P= 9 MF= C17H1406 MW= 314 When we combine the high efficiency Kinetex particle with the PFP chemistry, we were able to separate all four aflatoxins in <2 minutes, with the entire cycle time less than 5 minutes. When working with complex matrices such as peanut butter, there is always a concern that interferences can lead to ion suppression or enhancement, ultimately resulting in poor quantitation. The use of the twostage SPE cleanup removed a majority of potential contaminants and performing the separation under gradient conditions (50 to 95 % methanol) removed any additional hydrophobic contaminants that could lead to sample carryover. Mycotoxin M1 was used as an internal standard to improve quantitation. Figure 2. LC/MS/MS Chromatogram of Aflatoxin Standards at 50 ppb Aflatoxin G 1 CH 3 log P= 1.19 MF= C17H1207 MW= 328 Aflatoxin G 2 CH 3 log P= 6 MF= C17H1407 MW= 330 Counts 1377 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 0 0.5 App ID 18642 1.5 2.5 min Food safety testing often requires extremely fast turn around time between samples to allow proper action to be taken. When looking to achieve fast chromatography, there has been a lot of discussion in the past several years on the use of sub-2 µm or other high resolution columns. Since backpressure generated by the sub-2 µm materials is higher than the pressure limits of the standard HPLC system, many labs may have trouble implementing these fast LC solutions. To overcome this challenge, Phenomenex has developed the new Kinetex HPLC columns that utilize core-shell technology to provide sub-2 µm performance at backpressures compatible with standard HPLC instruments. Kinetex columns have a solid 1.9 µm core with a 0.35 µm porous shell, giving a total particle diameter of 2.6 µm. The 2.6 µm particle gives backpressures similar to a 3 µm particle, but the improved mass transfer kinetics of the 0.35 µm porous shell significantly improve resolving power giving sub-2 µm performance on any HPLC system. Most of the published work on aflatoxins reference the use of C18 type phases. When using C18 phases, resolution problems between B2 and G1 are common. The conjugated aromatic structure of aflatoxins suggests that they would be well retained through π-π type interactions using phenyl-based phases. The Kinetex PFP is a pentafluorophenyl phase that is highly prone to electrophilic interactions due to the delocalized electrons in π orbitals above and below the planar ring. Solutes containing aromatic rings participate in a stacking interaction with the benzene ring of the PFP ligand resulting in increased interaction and resolution. Page 2 of 8 Column: Kinetex 2.6 μm PFP Dimensions: 50 x 2.1 mm Part No.: 00B-4477-AN Mobile Phase: A: 0.1 % Formic acid and 5 mm Ammonium acetate in Water B: 0.1 % Formic acid and 5 mm Ammonium acetate in Methanol Gradient: Time (min) % B Time (min) % B 0 50 1 95 0.25 50 2.5 95 Flow Rate: 400 µl/min Temperature: 25 C Detection: Mass Spectrometer (MS) Sample: 1. Aflotoxins M1(IS) 2. Aflotoxins G2 3. Aflotoxins G1 4. Aflotoxins B2 5. Aflotoxins B1 2 70 2.51 50 4.4 50 The recovery yields from the first stage solvent-extraction of peanut butter samples were found to be in the 85-90 % range. This suggested that about 10-15 % of the aflatoxins were lost due to incomplete extraction and/or degradation during the evaporation step. When performing peanut butter extractions it is important to account for this loss when deciding on the mass of sample that must be extracted to meet the required detection limits. The recovery from the entire extraction and cleanup process ranged from 73-85 %. These values were calculated by multiplying the % recovery from the liquid-liquid extraction procedure by the % recovery from the two-stage SPE cleanup.
APPLICATINS SPE Procedure Strata Florisil 500 mg/3 ml Strata Silica (Si-1) 200 mg/3 ml Part No.: 8B-S013-HBJ Part No.: 8B-S012-FBJ Condition: No conditioning was performed Condition: 2 x 3 ml of hexane as this led to reduced recoveries of aflatoxins Load: A 1.5 ml aliquot of peanut butter extract was spiked with afl atoxin standards and loaded Wash: 1. 2 x 3 ml of methanol/water (80:20) 2. 2 x 3 ml of 100 % methanol Elute: 2 x 3 ml of acetone/water/0.5 % formic acid (96:3.5:0.5) The combined eluate was dried under nitrogen and the residue reconstituted in 2 ml of 1:1 hexane/chloroform and loaded onto the Strata Silica cartridge for further cleanup The two-stage SPE protocol presented here was designed to remove the maximum amount of matrix contamination. The peanut butter extract from methanol/ water (80:20) was directly loaded on the Strata Florisil and thoroughly washed with a four-step wash sequence. This protocol was expected to eliminate most hydrophilic impurities. However, since peanut butter contains a high amount of fatty acids, an optional second cleanup step was added to remove additional impurities. The eluate from the Strata Florisil was reconstituted in 1:1 chloroform/hexane and loaded onto a Strata Silica cartridge. The aflatoxins were recovered by non-retentive SPE, while the impurities were presumably retained on the cartridge. Post-column infusion studies were conducted to determine the effectiveness of the two-stage SPE cleanup. Figure 3 shows the total ion current when a blank (30 % acetonitrile) is injected while post-column infusing all four aflatoxins. As expected, no suppression was observed. When a blank peanut butter matrix, which has undergone no cleanup procedure, was injected, there was a significant suppression region from about 0.2- min (Figure 4). Suppression in this region interferes with the internal standard M1 and results in poor sensitivity and quantitation. In Figure 5 the same peanut butter extract is re-analyzed after cleanup using the twostage SPE procedure. The suppression region is significantly reduced, now only affecting peaks eluting in the 0.30 0.60 minute window. Since none of our target analytes or internal standard elute in this region, no suppression was observed. It should be noted that the ion suppression work was performed on an API 3000 system with a TurboIonSpray source. The TurboIonSpray source will show a greater degree of ion suppression and take longer for signals to recover making good sample cleanup even more critical. Although suppression studies have not been performed with the Turbo V source on the 3200 Q TRAP, it is presumed that suppression would be further reduced allowing for shorter run times. Load: 2 ml of reconstituted sample from the Strata Florisil Wash: 1. 2 x 2 ml of methanol/chloroform (1:1) 2. 1 x 1 ml of methanol/chloroform (1:1) Load and wash solutions from the silica SPE were pooled together and dried down under nitrogen and reconstituted in 500 µl of the mobile phase used for LC/UV or LC/ MS analysis Figure 3. Infusion of Aflatoxin Standard (G2, B2, G1, B1) and Injection of Solvent Blank (30 % ACN/Water) 535 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 Figure 4. Infusion of Aflatoxin Standard (G2, B2, G1, B1) and Injection of Extracted Peanut Butter Matrix before SPE Cleanup 475 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 475 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 0.2 0.4 0.6 0.8 1.2 1.4 1.6 1.8 2.2 2.4 2.6 2.8 min 0.2 0.4 0.6 0.8 1.2 1.4 1.6 1.8 2.2 2.4 2.6 2.8 min Figure 5. Infusion of Aflatoxin Standard (G2, B2, G1, B1) and Injection of Peanut Butter Matrix after Two-Stage SPE Cleanup App ID 18476 App ID 18476 App ID 18476 20 0 0.2 0.4 0.6 0.8 1.2 1.4 1.6 1.8 2.2 2.4 2.6 2.8 min Page 3 of 8
APPLICATINS Calibration curves for aflatoxins using the ABI 3200 Q TRAP were generated from 5 to 500 ppb. When the concentration range is reduced to 5 100 ppb, the data fits well to a linear equation with R2 values > 0.993 for all analytes. When the concentration ranges are extended to 500 ppb, a quadratic fit now yields the best quantitative results (Figure 6). At 500 ppb the response (height in cps) is well within the linear dynamic range of the detector suggesting that the suppression is originating in the ion source. This phenomenon has been observed in the past for other compounds and is not well understood. Further work must be done to better characterize and explain this observation. Four replicate QC spiked samples were prepared at the 50 ppb level and processed using the Florisil SPE procedure only. Two calibration curves were prepared, one in mobile phase and another in a blank peanut butter matrix that had been processed through the Florisil SPE cartridges. Accuracy for the QC spiked samples was between 132.8 144.3 % with % CVs <6.5 (Table 3). This data suggested that even though the accuracy was not as good as expected the method was highly reproducible. Previously published work using Florisil as a clean up technique for aflatoxins demonstrated recoveries in the mid 80 % range for peanut samples spiked in the 50 ng/g level. 6 The relatively high recoveries observed here were thought to result from poor correlation in response from the QC samples to the calibration curve made up in mobile phase. To better account for matrix interferences, calibration standards were prepared in a blank peanut butter matrix that had been processed using the Florisil SPE cartridges. QC sample accuracy improved dramatically to 94.3-97.6 % when quantitated using the extracted calibration curve (Table 4). The improvement in accuracy highlights the importance of using appropriate matrix matched calibration standards when looking do quantitative analysis by LC/MS/ MS. If this method were applied to other food products such as spices or nuts, it would be important to develop a similar food matrix blank for calibration purposes. Signal to noise at the 5 ppb level was still very good for all compounds in peanut butter matrix indicating the method was suitable for control of aflatoxins well below world wide detection limits (Table 5). If even lower detection limits were required, a larger initial amount of peanut butter could be extracted to achieve ppt level analysis. Table 1 LC/MS/MS Transitions Aflatoxin Molecular Ion Primary MRM Confirmation MRM B1 313.1 241.1 28 B2 315.1 259.1 287.2 G1 329.1 24 31 G2 331.1 245.1 31 M1 329.1 27 25 Table 2 Recovery Data of Aflatoxin Extraction and Two-Stage SPE Cleanup Aflatoxin* G2 B2 G1 B1 Liquid-Liquid Extraction Recovery 90 % 85 % 87 % 88 % 2-Stage SPE Recovery 89 % 97 % 84 % 97 % Absolute Recovery 80 % 82 % 73 % 85 % * n=3 for all samples Page 4 of 8
APPLICATINS Figure 6. Non-extracted Standard Calibration Curves 2 1 2 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 B1 R = 0.9880 B2 R = 0.9972 1 11.5 1 10.5 1 5.4 5.2 4.8 4.6 4.4 9.5 4.2 8.5 3.8 7.5 6.5 5.5 3.6 3.4 3.2 2.8 2.6 2.4 2.2 4.5 3.5 2.5 1.5 G1 R = 0.9970 1.8 1.6 1.4 1.2 0.8 0.6 0.4 G2 R = 0.9961 0.5 0.2 Table 3 Performance of QC spiked samples at 50 ng/ml quantitated against the non-extracted calibration curve Analyte S/N * %CV** Recovery *** Aflatoxin B1 275.9 141.7 Aflatoxin B2 302.2 4.7 144.3 Aflatoxin G1 224.5 3.2 132.8 Aflatoxin G2 18 6.5 136.4 * Signal to Noise (S/N) being the peak height divided by the noise measured at three standard deviations of the noise. ** % CV estimates from 4 replicate samples, with 2 injections for each sample (n=8) *** Recovery was obtained using non-extracted standard calibration curve. Page 5 of 8
APPLICATINS Figure 7. Spiked Standard Calibration Curves 2 1 2 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 B1 R = 0.9980 B2 R = 0.9987 1 11.5 1 10.5 1 5.4 5.2 4.8 4.6 4.4 9.5 4.2 8.5 3.8 7.5 6.5 5.5 3.6 3.4 3.2 2.8 2.6 2.4 2.2 4.5 3.5 2.5 1.5 G1 R = 0.9990 1.8 1.6 1.4 1.2 0.8 0.6 0.4 G2 R = 0.9984 0.5 0.2 Table 4 Performance of QC spiked samples at 50 ng/ml quantitated against the extracted calibration curve Analyte S/N * %CV** Recovery *** Aflatoxin B1 275.9 4.8 96.7 Aflatoxin B2 302.2 4.5 94.3 Aflatoxin G1 224.5 3.1 97.6 Aflatoxin G2 18 6.3 94.9 * Signal to Noise (S/N) being the peak height divided by the noise measured at three standard deviations of the noise. ** % CV estimates from 4 replicate samples, with 2 injections for each sample (n=8) *** Recovery was obtained using spiked standard calibration curve Page 6 of 8
APPLICATINS Table 5 Representative Signal to Noise Ratios at 5 ng/ml spiked Standard level Analyte S/N * %CV** Recovery *** Aflatoxin B1 185.5 5.6 97.8 Aflatoxin B2 63.5 3.4 99.2 Aflatoxin G1 120.5 5.4 96.2 Aflatoxin G2 49.5 1 97.5 * Signal to Noise (S/N) being the peak height divided by the noise measured at three standard deviations of the noise. ** % CV estimates from 2 replicate samples, with 2 injections for each sample (n=4). Conclusions The current work describes a quick and easy method for cleanup and analysis of aflatoxins from peanut butter samples. The twostage SPE procedure removes a majority of matrix interferences observed in LC/MS/MS chromatography. The SPE phases also provide a good alternative to immunoaffinity products. Combining the high efficiency Kinetex core-shell particle with the highly selective PFP chemistry enabled ultra fast separation of all four aflatoxins on any LC/MS/MS system configuration. Precision and accuracy of this method were excellent when using a matrix matched calibration curve. The estimated detection limits for each analyte are well below 5 ppb, which are more than sufficient to allow the method to be used for either confirmation or quantitation. Although this work focused only on the analysis of peanut butter, future work will be done to apply this method to other food products including peanuts and maize. To facilitate sample sharing between labs, sample stability will also be evaluated to determine the best way to preserve and ship samples. Acknowledgements We would like to specially thank David Lavorato, Yun Yun Zou, Adam Latawiec, and Andre Schreiber from Applied Biosystems in Toronto, Canada for running samples, interpreting data, and providing advice regarding this project. References 1. H.P. van Egmond, Analytical Bioanalytical Chemistry 2004, 378, 1152-1160. 2. Food and Agriculture rganization. Worldwide Regulations for Mycotoxins in Food and Feed, 2003, FA Food and Nutr. Paper, 2004, 81. 3. M.C. Spanjer, P.M. Rensen and J.M. Scholten, Food Additives and Contaminants, 2008, 25, 472-489. 4. V.A. Vega, J. AAC International 2005, 88, 1383-1386. 5. M. Castegnaro, M. Tozlovanu, C. Wild, An Molinie, A. Sylla and A. Pfohl- Leszkowicz, Mol. Nutr. Food Res. 2006, 50, 480-487. 6. V. Sobolev, J. Agric. Food Chem. 2007, 55, 2136-2141. Page 7 of 8
APPLICATINS Australia t: 02-9428-6444 f: 02-9428-6445 auinfo@phenomenex.com Austria t: 01-319-1301 f: 01-319-1300 anfrage@phenomenex.com Belgium t: +31 (0)30-2418700 f: +31 (0)30-2383749 beinfo@phenomenex.com Canada t: (800) 543-3681 f: (310) 328-7768 info@phenomenex.com Denmark t: 4824 8048 f: 4810 6265 dkinfo@phenomenex.com France t: 01 30 09 21 10 f: 01 30 09 21 11 franceinfo@phenomenex.com Germany t: 06021-58830-0 f: 06021-58830-11 anfrage@phenomenex.com Ireland t: 01 247 5405 f: +44 1625-501796 eireinfo@phenomenex.com Italy t: 051 6327511 f: 051 6327555 italiainfo@phenomenex.com Luxembourg t: +31 (0)30-2418700 f: +31 (0)30-2383749 nlinfo@phenomenex.com Netherlands t: 030-2418700 f: 030-2383749 nlinfo@phenomenex.com Strata rdering Information Sample Preparation Part No. Description Unit 8B-S013-HBJ Strata Florisil (500 mg/3 ml) 50/box 8B-S012-FBJ Strata Silica (200 mg/3 ml) 100/box Kinetex rdering Information 2.6 μm Minibore Columns (mm) 50 x 2.1 100 x 2.1 150 x 2.1 PFP 00B-4477-AN 00D-4477-AN 00F-4477-AN 2.6 μm Solvent Saver Midbore Columns (mm) 50 x 100 x 150 x PFP 00B-4477-Y0 00D-4477-Y0 00F-4477-Y0 2.6 μm Analytical Columns (mm) 50 x 4.6 100 x 4.6 150 x 4.6 PFP 00B-4477-E0 00D-4477-E0 00F-4477-E0 KrudKatcher Ultra In-Line Filter rdering Information Disposable in-line filter fits virtually all UHPLC/HPLC columns to 4.6 mm ID. Extremely low dead volume minimizes sample peak dispersion. Pressure rated to 20,000 psi(1,375 bar). KrudKatcher Ultra In-Line Filter Part No. Description Unit AF0-8497 KrudKatcher Ultra In-Line Filter, 0.5 µm Porosity x 04 in. ID 3/pk Installation wrench not provided. KrudKatcher Ultra requires 5 /16 in. wrench. New Zealand t: 09-4780951 f: 09-4780952 nzinfo@phenomenex.com Puerto Rico t: (800) 541-HPLC f: (310) 328-7768 info@phenomenex.com United Kingdom t: 01625-501367 f: 01625-501796 ukinfo@phenomenex.com All other countries: Corporate ffice USA t: (310) 212-0555 f: (310) 328-7768 info@phenomenex.com www.phenomenex.com Phenomenex products are available worldwide. For the distributor in your country, contact Phenomenex USA, International Department at international@phenomenex.com. If Strata SPE products and Kinetex core-shell HPLC columns do not provide at least equivalent results and separations as compared to products of similiar dimensions, phase, and particle size, return the product with your comparative data within 45 days for a FULL REFUND. Trademarks Strata is a registered trademark of Phenomenex, Inc. Kinetex and SecurityGuard are trademarks of Phenomenex, Inc. TurbolonSpray and Q TRAP are registered trademarks, and API 3000, API 3200, and TURB V are trademarks of Life Technologies Corporation and its affiliated company Applied Biosystems. Florisil is a registered trademark U.S. Silica Co. Disclaimer Phenomenex is not affiliated with Applied Biosystems or U.S. Silica Co. Subject to Phenomenex Standard Terms & Conditions, which may be viewed at www.phenomenex.com/termsandconditions. 2009 Phenomenex, Inc. All rights reserved. TN68830409_L Page 8 of 8