Sample Preparation and Matrix Effects in the Detection of Chemical Residues in Foods Steven J. Lehotay, HyeYoung Kwon, and Lucía Geis-Asteggiante Hyeyoung Kwon Lucía GeisAsteggiante Rural Development Administration National Academy of Agricultural Science Suwon, South Korea Cátedra de Farmacognosia y Productos Naturales, DQO; UdelaR Montevideo, Uruguay Presentation Outline I. QuEChERS & Update II. Matrix Effects (GC and LC-API-MS) III. Pesticide Experiments & Results IV. Veterinary Drug Residue Results V. Conclusions Multiclass, Multiresidue Analytical Approach: extraction clean-up Quick Easy QuEChERS method Cheap Effective Rugged quantification identification Safe GC-MS LC-MS Slide Devised by Katerina Mastovska QuEChERS Approach Tomato Grape Spinach Strawberry 1) Shake sample with solvent and salts 2) Centrifuge for 1 min Spinach Tomato Grape Strawberry 4) Centrifuge for 1 min 5) Analyze Pesticides 3) Mix a portion with a sorbent Different QuEChERS Methods 2003 Anastassiades et al. 2005 Lehotay et al. 2007 Anastassiades et al. Original AOAC 2007.01 CEN 15662 10-15 g sub sample 10-15 g sub sample 10-15 g sub sample 10-15 mL 1% HOAc in MeCN 10-15 mL MeCN shake shake centrifuge shake & centrifuge 0.4 g/mL anh.MgSO4 0.1 g/mL NaCl 0.1g/mL Na3Cit2H2O 0.05 g/mL Na2Cit1.5H2O 0.4 g/mL anh.MgSO4 0.1 g/mL NaOAc shake shake shake 0.4 g/mL anh.MgSO4 0.1 g/mL NaCl 150 mg/mL anh.MgSO4 25 mg/mL PSA 10-15 mL MeCN Option: + 50 mg C18 & 7.5 mg GCB centrifuge 150 mg/mL anh.MgSO4 50 mg/mL PSA shake & centrifuge shake Option: centrifuge + 50 mg C18 & 150 mg/mL anh.MgSO4 2.5-7.5 25 mg/mL PSA mg GCB shake & centrifuge Option: Scale-Up & Conc. in Toluene Slide Devised by Urairat Koesukwiwat QuEChERS Update Steve and Angelo “Interviewed” by Ron Majors QuEChERS a Sample Preparation Technique that is “Catching On”: An Up-to-Date Interview with the Inventors, LC GC North America, July, 2010 The QuEChERS Revolution, LC GC Europe, Sept., 2010 Available from ChromatographyOnline.com QuEChERS in the Literature Number 100 90 80 70 60 50 40 30 20 10 0 363 Total Citations to 1st Paper 257 Total QuEChERS & d-SPE Pubs According to ISI Web of KnowledgeSM * Year 2003 2004 2005 2006 2007 2008 2009 2010 2011 *search conducted on May 3, 2011 What’s New with QuEChERS? • • • • • • • • • More vendors and formats (e.g. DPX) GCB or ChloroFiltr for chlorophyll reduction Type of MgSO4: can use 97% purity Shakers – e.g. Spex and Glas-Col Automation with robotic autosampler “Unified” method to undergo AOAC update d-SPE has a life of its own More applications (e.g. PAHs) Veterinary drug residue methods Vendors of QuEChERS Products b e k o lu t Disposable Pipette Extraction (DPX) Patented in 2003 by William Brewer, University of South Carolina Comparison of ChloroFiltr (16B) and GCB Recoveries in Lettuces (QuEChERS w/MeCN) 50 mg/mL 16B 7.5 mg/mL GCB 100% 60% 40% 20% ac hl C or hl or py rif os Fe nt hi on C yp ro Th di ia ni be l nd az Te ol bu e co na zo le H ep t zo lin lo il Vi nc ha l on e ro t an C hl o Li nd Q ui nt oz en CB e 0% H Recovery 80% Unified QuEChERS Method 1 g sample per 1 mL of MeCN w/ 1% HOAc for fruits and vegetables add internal standard per g sample, add 0.4 g anh. MgSO4 + 0.1 g anh. NaOAc shake or blend centrifuge per mL of the upper layer: 150 mg MgSO4 + 50 mg PSA + 50 mg C18 + 7.5 mg GCB mix and centrifuge QuEChERS for Grains, Nuts, Doughs 2.5 - 5 g sample + 10 mL water* + 10 mL MeCN + internal standards shake for 1 h add 4 g MgSO4 + 1 g NaCl shake vigorously for 1 min centrifuge for 1 min *15 mL water for 5 g of rice 1 mL of the upper layer + 150 mg PSA + 50 mg C18 + 150 mg MgSO4 mix for 30 s centrifuge for 1 min H C Q ui M B nt ire o p, ze x p n ci o '-D e s- ,p D Ch '- E D l o H rd DT ep a C C hl tac ne hl or h or py lo py r r L i B r if in fo ro o d s m s-m an E n C op e e do yp r op thy su er yl l lfa m e ate n th S r Tr ulf in O ifl at xy u r e Pi rim P flu ali n ip erm or fe ho e n s- thr m i F e et n Vi nt hy C ncl hio l ou oz n m ol ap i n ho s Recovery 6 Data Sets from 3 Types of Flaxseeds QuEChERS of Milled Flaxseeds 100% Why not correct for recoveries? 80% 60% 40% 20% 0% n = 26 Biggest Problem with LC- and GC- MS(/MS) plus costs to purchase and maintain, and facility requirements, and downtime, and need for more expertise due to greater complexity It is still a pain! Experiment to Assess Matrix Effects • 33 LC- and/or GC- amenable pesticides • 4 matrices (apple, orange, spinach, and rices) • 20 different sources of each commodity • Calibration standards from 10-350 ng/g in each commodity/source and reagent-only • Analyte protectants added to GC standards • Analytical sequences conducted on API-3000 LC-(ESI+)-MS/MS and LP-GC/ToF-MS (10 µL PTV) • Matrix effects calculated (vs. I.S. and not) How to Calculate (Estimate) Matrix Effects GC-TOF/MS of Atrazine in Rice 2.0E+06 Reagent Stds Peak Area 1.8E+06 1.6E+06 Matrix Stds (n=20) 1.4E+06 Linear (Reagent Stds) Linear (Matrix Stds (n=20)) 1.2E+06 1.0E+06 y = 4612x + 29698 R2 = 0.9986 y = 4142x + 4558 R2 = 0.9987 8.0E+05 6.0E+05 4.0E+05 %ME = (4612-4142)/4142 = 11% 2.0E+05 0.0E+00 0 50 100 150 200 250 Conc. (ng/g) 300 350 400 Isotopically-Labeled Internal Standard is Ideal GC-TOF/MS of Atrazine in Rice Peak Area vs. Atrazine-d5 1.2 Reagent Stds 1.0 Matrix Stds (n=20) 0.8 Linear y(Reagent = 0.00264x + 0.00132 Stds) 2 R = 1.000 Linear (Matrix Stds (n=20)) y = 0.00258x + 0.00124 R2 = 1.0000 0.6 0.4 0.2 %ME = (264-258)/258 = 2% 0.0 0 50 100 150 200 250 Conc. (ng/g) 300 350 400 Alternative Calculation Method GC-TOF/MS of Atrazine in Rice (w/o I.S.) 40% 20% 10% 0% -10% -20% White Rices (R13-20) ME = 13% ± 10% -30% -40% R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 Matrix Effect 30% Brown Rices (R1-R12) Sample Effect of Isotopically-Labeled IS 40% 20% 10% But the I.S. is not infallible Brown Rices (R1-R12) 0% -10% -20% White Rices (R13-20) ME = 2% ± 3% -30% -40% R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 Matrix Effect 30% GC-TOF/MS of Atrazine vs. IS in Rice Sample Results for LP-GC/ToF (w/o I.S.) 100% Apple Orange Spinach Rice 80% Matrix Effects 60% Matrix-Induced Enhancement 40% 20% 0% -20% -40% Ethoprophos Atrazine Linuron Azoxystrobin Results for LC-MS/MS (w/o I.S.) 10% 0% Matrix Effects -10% -20% -30% Ion Suppression -40% -50% Apple Orange Spinach Rice -60% Ethoprophos Atrazine Linuron Azoxystrobin Results for LC-MS/MS (w/ I.S.) 30% 20% Matrix Effects 10% 0% -10% -20% normalization to atrazine-d5 helps more than just for atrazine -30% -40% Apple Orange Spinach Rice -50% Ethoprophos Atrazine Linuron Azoxystrobin Results for LC-MS/MS (w/ I.S.) 100% 80% %ME: orange > spinach > rice = apple Matrix Effect 60% 40% 20% 0% -20% -40% -60% -80% -100% 2.0 Apple Orange Spinach Rice Linear (Rice) Linear4.0 (Orange) 3.0 5.0 6.0 7.0 8.0 9.0 Linear (Spinach) Retention Time (min) Linear (Apple) 10.0 11.0 12.0 Results for LP-GC/ToF (w/ I.S.) 200% 175% Matrix Effect 150% 125% 100% 75% 50% Apple Orange cypermethrin Spinach Rice Linear (Rice) Linear (Orange) Linear (Spinach) Linear (Apple) azoxystrobin 25% 0% -25% -50% 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 Retention Time (min) %ME: orange >> rice > spinach = apple Analyte Protectants Protectant Strongly interact with active sites in GC system (inlet, column and ion source) to decrease degradation and adsorption of co-injected analytes. Sharper peaks, less tailing, more ruggedness, lower LOD HO O O HO HO OH OH HO O HO OH OH HO OH HO ethylglycerol 1 mg/mL OH gulonolactone 0.1 mg/mL O OH HO sorbitol 0.1 mg/mL shikimic acid 0.05 mg/mL Effect of Analyte Protectants w/ analyte protectants w/o analyte protectants Anastassiades, Maštovská, Lehotay, J. Chromatogr. A, 1015, 163-184 (2003) Combination of Analyte Protectants for GC Pesticide Residue Analysis O HO OH ethylglycerol (10 g) Signal enhancement: moderate strong HO O HO HO O OH gulonolactone (1 g) OH HO OH OH OH sorbitol (1 g) retention time K. Mastovska, S.J. Lehotay, M. Anastassiades, Anal. Chem., 77, 8129-8137 (2005) Conclusions of Pesticides Study • Matrix effects aren’t so bad in QuEChERS with LCand GC- MS(/MS) analyses, but worse in citrus • In terms of matrix effects, one apple is much like another, and oranges are alike, too, but apples aren’t like oranges, they’re like plums, etc. • Analyte protectants in GC improve results, but matrix-matching still needed for late-eluters, especially in citrus. • Isotopically-labeled internal standards work best to overcome matrix effects, but not perfectly, and they even help reduce effects for other analytes. Comparison of 6 Vet. Drug Methods 1) Mol et al. (Rikilt – The Netherlands) 2) Martos et al. (U. Guelph – ON, Canada) 3) Lehotay et al. (USDA-ARS – Wyndmoor, PA) 4) Leepipatpiboon et al. (Chulalongkorn U., Thailand) 5) Stubbings et al. (FERA – York, UK) 6) Kaufmann et al. (Switzerland) All methods gave similar qualitative MS/MS screening capabilities with nearly all 60 of the analytes meeting identification criteria at ½ “tolerance” level in kidney. Speed, cost, ease of use and ruggedness become the differentiating aspects. 60 Vet. Drugs in Beef Kidney vs. SMZ-d6 200% 180% 160% 140% 120% 100% Mol 80% Martos 60% ARS 40% Thai 20% Stubbings e lin Te tr ac yc lfa s Su es sa m id co id e s+ Lin w th M ac ro l Gr o Fl u or oq ui pr om no lo n rin sp o al o Ce ph + be ta -la ct am ot or s 0% Analysis of Incurred Kidney (2 g) Flunixin vs. Flunixin-d3 IS in Incurred Kidney 700 Concentration (ng/g) 5 min shake; w/o acid 600 5 min shake; w/ acid 500 30 min shake; w/o acid 30 min shake w/ acid 400 60 min heat; w/o acid 60 min heat; w/ acid 300 200 100 0 1 2 3 4 5 6 7 Kidney Sample 8 9 10 Analysis of Incurred Kidney (2 g) Sulfamethazine vs. SMZ-d6 IS in Incurred Kidney Concentration (ng/g) 250 5 min shake; w/o acid 200 5 min shake; w/ acid 30 min shake; w/o acid 150 30 min shake w/ acid 60 min heat; w/o acid 100 60 min heat; w/ acid 50 0 1 2 3 4 5 6 Kidney Sample 7 8 9 10 Analysis of Incurred Kidney (2 g) Penicillin G vs. PenG-d7 IS in Incurred Kidney Concentration (ng/g) 60 5 min shake; w/o acid 30 min shake; w/o acid 60 min heat; w/o acid 50 40 5 min shake; w/ acid 30 min shake w/ acid 60 min heat; w/ acid No need for long shake nor heat, so fast and cool is fine! 30 Spikes were ok, but incurred penicillins can’t tolerate acids 20 10 0 1 2 3 4 5 6 7 Kidney Sample 8 9 10 Fast Method for Vet. Drug Residues 2 g tissue in a 50 mL tube add IS mix (SMZ-d6; flunixin-d3; PenG-d7) add 10 mL of 4/1 (v/v) MeCN/water vortex briefly, shake for 5 min centrifuge for 5 min >3500 rcf supernatant + 500 mg C18 + 10 mL hexane sat’d w/MeCN; mix for 30 s, centrifuge for 5 min > 3500 rcf; aspirate hexane to waste evaporate 5 mL extract to 1 mL final vol. filter extract with the Mini-UniPrepTM UHPLC-MS/MS analysis Streamlined Method Validation Needs: • Trueness (Recoveries at ≥3 Levels, n > 5) • Precision (Repeatability & Reproducibility) • Ruggedness (Multi-day, Multi-Analyst, etc.) • Selectivity (Interferences in Blanks?) • Range (calibration and matrix effects) • Detection limits (MDL, LOD, LOQ, LOI) • Qualitative (False Negatives/Positives) Can We Meet All Needs in 3 Days? 3-Day Validation Experiment Day 1: • Analyst 1 in hot Lab, Reagents A, 10 matrix blanks from different sources, 6 spikes at 3 levels each in 6 matrices + 4 spikes each at same levels in mixed matrices (1 in glass tubes); 5-point calibration each in mixed matrix and reagent-only stds; reagent blk = 0-Std inj’d after high std to check for carry-over Days 2 and 3: • Analysts 2 & 3 in cooler labs repeat using Reagents B & C with different sources of matrices Veterinary Drug Residues Conclusions • The streamlined method has met validation criteria for most drugs in a 3-day validation for qualitative identification screening purposes. • Sample throughput is 60 samples/day by 1 chemist for UHPLC-MS/MS analysis. • The method is being implemented for routine monitoring of cattle (so far) by the USDA labs. • Quantitation is acceptable for ≈75% of the drugs, but enforcement requires 2nd analysis anyway. • The new streamlined method still needs a cool name. Acknowledgments Michelangelo Anastassiades Katerina Mastovska Alan Lightfield Urairat (Oil) Koesukwiwat Terry Dutko and others at USDA FSIS US-Israel Binational Agricultural Research and Development Grant US-4273-09 QuEChERS Baby Picture This is what happens when two fathers who hate to do dishes have a baby together Muchas Gracias! Efficient Pesticide Residue Analysis Collect Appropriate Sample (cut and store in dry ice?) Drive Mobile Lab to the Field or Send Sample to the Lab Homogenize and Subsample – Add QC Spk for Sample Processing Weigh 15 g Sample into Tube (Place ≈200 g for Cold Storage) Use QuEChERS to Extract (Add I.S.) including d-SPE or DPX Cleanup 250 μL to Mini Uni-Prep Vial - Add QC Std - Transfer 150 µL to Vial w/ Insert for GC; Add Mobile Phase Diluent for LC and Filter in Vial Inject in (DSI-)LP-GC/MS(-MS) (Need for Analyte Protectants?) Inject in UPLC-MS/MS (Need for Matrix-Matching?) Incorporate Data into LIMS for QA/QC Review and Report Conclusions QuEChERS is a well-proven, fast sample preparation method for hundreds of pesticide residues in different types of food matrices. UPLC-MS/MS can provide 10 min analysis of hundreds of LC-amenable pesticides. LP-GC/MS can also provide 10 min analysis of hundreds of GC-amenable pesticides. Currently, the HUGE sample throughput limitation is data processing and review! QuEChERS as a Teenager QuEChERS is no longer a baby, born of two fathers, it is a teenager influenced by friends, some you can trust and others you can’t. The QuEChERS approach is still learning its potential and limitations in the big world. QuEChERS concepts are easy and fast to try in your application(s) – no big loss if it fails. Recovery experiments alone are not enough to validate methods – use incurred samples, proficiency testing, and/or interlab trials. Dispersive-SPE • Why use an SPE apparatus for “chemical filtration?” • Dispersive-SPE involves the mixing of the sorbent with the extract in a mini-centrifuge tube to retain matrix interferants, but not analytes. QuEChERS Features and Impact • A single extract can be prepared in 10 min or a batch of 20 in an hr by a single analyst with ≈$1-3 of disposable materials per sample and generate <12 mL nonchlorinated solvent waste. • Consistently high recoveries (mostly 90-110% with RSDs < 10%) of a wide range of GC- and LC-amenable pesticides are achieved from many food matrices. • Countless labs have implemented QuEChERS successfully for up to 500 pesticides in food and increased efficiency (faster, less labor, lower cost, less waste, saves space, less labware, higher throughput). • QuEChERS concepts have spread to other applications. What is QuEChERS? www.quechers.com http://en.wikipedia.org/wiki/Quechers The Quechers method is a streamlined approach that makes it easier and less expensive for analytical chemists to examine pesticide residues in food. The name is a portmanteau word formed from "Quick, Easy, Cheap, Effective, Rugged, and Safe." QuEChERS History 2000-2002 2002 EPRW-Rome 2003 Publication MgSO4 Dispersive-SPE Analyte Protectants Limitations of QuEChERS? • • • • • • • • Too Many Modified Versions Cereals require a separate protocol Still Problems with captan, folpet, captafol Spices and oils give problems Works best with modern MS systems Need PTV or solvent exchange for low LOD Matrix effects in complicated matrices Even simpler sample prep possible QuEChERS Update NOTE: Recoveries were normalized to results without ice. Syringeless Filters Mini-UniPrepTM (Whatman) 1) Place unfiltered sample (max. 0.5 mL) in chamber. 2) Compress filter plunger into sample chamber. Clean filtrate fills reservoir bottom up. 3) Place the Mini-UniPrepTM vial in an autosampler. aqueous samples: PVDF (polyvinylidenefluoride) filter QuEChERS Sample Prep (1) weigh 15 g homogenized sample into a 50 mL tube (2) add spiking and I.S. solutions, and vortex for 1 min; (3) add 15 mL of MeCN with 1% HOAc; shake for 30 s; (4) add 6 g of anh. MgSO4 and 1.5 g of anh. NaOAc; (5) shake the tube immediately for 1 min; (6) centrifuge the tube at 3,250 rcf for 2 min; (7) transfer 1 mL extract to d-SPE tube containing 150 mg anh. MgSO4 + 50 mg PSA + 50 mg C-18 + 7.5 mg GCB; (8) mix for 30 s and centrifuge at 3,250 rcf for 2 min; (9) transfer 0.5 mL into an autosampler vial; (10) add 50 L of the QC and analyte protectants mixture and 50 μL MeCN (to make sample volumes equal those of the calibration standards), and (11) conduct LP-GC/MS-MS analysis.
© Copyright 2020