Waters Application Notes. Food Testing

Transcription

1 Waters Application Notes Food Testing

2 THE ONLY THING YOU LL FIND DIFFICULT TO QUANTIFY ARE THE POSSIBILITIES. INTRODUCING XEVO TQ-XS Your laboratory is being challenged to expand the scope of ultimate sensitivity analysis. Don t let complex matrices and low concentration levels stand in the way. The fast-track to simplifying your most complex analyses with highly repeatable results awaits at waters.com/xevotqxs PHARMACEUTICAL HEALTH SCIENCES FOOD ENVIRONMENTAL CHEMICAL MATERIALS 216 Waters Corporation. Xevo, Waters and The Science of What s Possible are registered trademarks of Waters Corporation.

3 Ensuring consistent, reliable safety in the global food supply In these recently published application notes, learn how Waters comprehensive analytical solutions are helping food testing laboratories to identify diverse chemical compounds, meet compliance requirements, increase productivity, and most importantly, help ensure public safety. [ PESTICIDES ] UPLC and APGC Multiresidue Pesticide Analysis on a Single Tandem Quadrupole MS Platform Analysis of a large suite of pesticides in fruit and vegetable extracts using both LC and GC on the same tandem quadrupole MS platform at legislatively relevant limits, with less than 3 min needed to switch between chromatographic inlets...5 Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS A generic QuEChERS sample extraction procedure was used to extract the pesticides from the fruit and vegetable samples, followed by rapid and high resolution UPLC separation and trace level detection of pesticides using the Xevo TQ-S micro...15 Highly Sensitive Analysis of Polar Pesticides in Food Matrices Using the ACQUITY UPLC System coupled to the Xevo TQ-XS Triple Quadrupole Mass Spectrometer, we achieved sub-parts per billion detection (ppb) limits for glyphosate, glufosinate, and aminomethylphosphonic acid (AMPA)...25 Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides in Chili Powder All pesticides were analyzed on a reverse phase column within 12 minutes. Two MRM transitions for each pesticide were monitored using both ESI positive and negative modes (deploying polarity switching)...27 Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra Simultaneous acquisition of MRMs and RADAR full-scan data provides quantitative and qualitative information in single injection [ ALLERGENS ] Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs Targeted UPLC-MS/MS analysis of food allergens in a variety of complex food matrices, including chocolate, ice cream, tomato sauce, and cookies...51 Identification and Quantitative Analysis of Egg Allergen Peptides Using Data Independent Ion Mobility MS A discovery proteomic workflow has been applied to determine marker peptides that can be used for the quantitative analysis of allergenic proteins within food...57 [ VETERINARY DRUGS ] Determination of Beta Agonists in Animal Tissues and Urine Using Liquid Chromatography-Tandem Quadrupole Mass Spectrometry A specific, targeted method for ß-agonists determination in a range of sample types that meets requirements for both official control and food business operators due diligence testing...65 Determination of the Metabolites of Nitrofuran Antibiotics in Animal Tissues and Food Products UPLC-MS/MS method facilitates routine screening and confirmation for official control purposes with a high degree of confidence, but also meets the requirements of pre-export testing, which often demands lower limits of quantification...74 Rapid, Simple, and Effective Clean-up of Bovine Liver Samples Prior to UPLC-MS/MS Multiresidue Veterinary Drug Analysis A novel reversed-phase sorbent, Oasis PRiME HLB, is used for highly effective removal of both phospholipids and fats from bovine liver extracts prior to UPLC-MS/MS analysis...79 Selective and Sensitive Screening of Chloramphenicol in Milk Mass detection, coupled with liquid chromatography offers increased selectivity, which helps differentiate between chloramphenicol and contaminating matrix and eliminates the need for time-consuming derivatization typically required for screening of chloramphenicol by GC [ 1 ]

4 [ FOOD CONTACT MATERIALS ] Quantifying Primary Aromatic Amines (PAAs) in Polyamide Kitchenware Sensitive LC-MS/MS method for the analysis of 23 common PAAs with very easy sample preparation. The Xevo TQ-S micro provides sensitive detection levels well below the regulation on plastics (EU 1/211)...89 Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System Simple workflow and methodology for the screening and characterization of food contact materials and other packaging materials...98 [ NATURAL TOXINS ] Mycotoxin Analysis in Cereal Grains Simple modified QuEChERS extraction protocol and ssmple clean-up strategies suitable for multiresidue mycotoxins analysis by UPLC-MS/MS...17 Analysis of Cyanotoxins, Including Microcystins in Drinking and Surface Waters UPLC-MS/MS method for the analysis of 1 cyanotoxins, including important microcystins, using standard addition, without the need for extraction at concentrations well below the WHO provisional guideline of 1 µg/l [ FOOD PROFILING/RESEARCH ] Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector This application note demonstrates the utility of DART (direct analysis in real time) and ACQUITY QDa mass detection for rapid, accurate, and cost-effective flavor characterization, for monitoring product quality, and reducing production costs Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado Using APGC-HRMS Using high resolution mass spectrometry (HRMS) and atmospheric pressure chemical ionization following gas chromatography (APGC), the accurate mass of the molecular ion and generated fragments were used to propose an identification of the unknown compound Discrimination of Different Unifloral Honeys Using an Untargeted High-Definition Mass Spectrometry Metabolomic Workflow In this study we investigated whether untargeted metabolomics, using UPLC with HDMS and multivariate statistics, could differentiate honeys of different botanical origin. This metabolomics workflow is emerging as a powerful approach for the discovery of biomarkers to tackle food fraud Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix In this application note, learn how mass detection can be deployed in food QC lab environments to provide the selectivity and sensitivity necessary to detect low levels of compounds in complex matrices, such as orange juice [ FOOD INGREDIENTS ] Fast Analysis of Isoflavones in Dietary Supplements USP Method Transfer onto a UHPLC System Demonstrates the transfer of the USP method onto an ACQUITY Arc UHPLC System. The analysis time with the ACQUITY Arc System is only 18 minutes, including column wash and equilibration. The ACQUITY QDa Mass Detector was used to expedite the method transfer described in this study A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products A fast, reliable, and sensitive method for the identification and quantification of sweeteners for food and beverage manufacturers using ACQUITY Arc UHPLC and ACQUITY QDa mass detection [ 2 ]

5 [ PESTICIDES ] [ 3 ]

6 [ PESTICIDES ] [ 4 ]

7 [ APPLICATION NOTE ] UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform Kari Organtini, 1 Gareth Cleland, 1 Eimear McCall, 2 and Simon Hird 2 1 Waters Corporation, Milford, MA, USA 2 Waters Corporation, Wilmslow, UK APPLICATION BENEFITS Using the Xevo TQ-S micro Tandem Quadupole Mass Spectrometer with the Universal Source for pesticide analysis allows: UPLC and APGC analysis of the sample extracts on a single tandem quadrupole mass spectrometer. Analysis of large suites of pesticides in a single injection per chromatographic inlet. Analysis of fruit and vegetable matrices at legislatively relevant levels of.1 mg/kg. Easy generation of methods using the Quanpedia Database. WATERS SOLUTIONS ACQUITY UPLC H-Class System Atmospheric Pressure Gas Chromatography (APGC) Xevo TQ-S micro DisQuE QuEChERS, AOAC Method Sample Preparation Kit, Pouches MassLynx MS Software Quanpedia Database TargetLynx XS Application Manager KEYWORDS LC, GC, pesticide residue analysis, MRL, QuEChERS, GC-MS/MS, LC-MS/MS AIM Demonstrate analysis of a large suite of pesticides in fruit and vegetable extracts using both LC and GC on the same tandem quadrupole MS platform at legislatively relevant limits. INTRODUCTION Hundreds of pesticides are commercially available and approved for use on various fruit and vegetable plants, to prevent pest infestation and improve shelf-life of fresh produce. Maximum Residue Levels (MRLs) are set at the highest level of pesticide that the relevant regulatory body would expect to find in that crop when it has been treated in line with good agricultural practice. In the EU, if a pesticide is not explicitly mentioned in the MRL legislation, a default MRL is used for enforcement. This default value is set to be equal to the limit of quantification (LOQ) achievable with the analytical methods used for analysis. National authorities control and enforce MRLs by testing samples for pesticide residue levels using analytical surveillance programs. These programs check for compliance with MRLs, assess dietary exposure, and check for use of unauthorized pesticides. The food industry also carries out its own due diligence analyses. Mass spectrometry coupled with both gas (GC) and liquid chromatography (LC) is needed to provide comprehensive analysis of a wide range of pesticide residues with sufficient sensitivity to meet global MRL regulations. The use of Quick, Easy, Cheap, Efficient, Rugged, and Safe (QuEChERS) sample extraction and clean up has streamlined analytical efficiencies for multi residue analyses. 1 The advantage of ultra performance liquid chromatography (UPLC) coupled with tandem quadrupole mass spectrometry (MS/MS) for multi residue pesticide analysis is widely reported. 2 More recently the use of GC-MS/MS operated at atmospheric pressure (APGC) has been shown to offer significant improvements in performance over electron impact (EI) for challenging pesticides, in terms of selectivity, specificity, and speed of analysis. 3,4 UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform [ 5 ]

8 [ APPLICATION NOTE ] The APGC source ionizes compounds using a corona discharge at atmospheric pressure in an APCI-like manner. Therefore, this ionization mechanism is a much softer technique than classic electron impact (EI) ionization and produces larger amounts of intact parent ions, especially in the case of fragile or easily fragmented compounds. APGC ionization can occur using two mechanisms; proton transfer (wet source) or charge transfer (dry source). In proton transfer ionization, [M+H] + ions are formed, whereas in charge transfer ionization, M + ions are formed. In this application note, a single workflow for the multi residue analysis of pesticides is demonstrated on a variety of fruit and vegetable samples. Utilizing the universal source of Waters Xevo TQ-S micro allows for LC and GC analyses to be completed on the same tandem quadrupole MS instrument, with less than 3 minutes needed to switch between chromatographic inlets. The performance of the method will be highlighted in terms of sensitivity, repeatability, and linearity for both LC and GC in compliance with the SANTE guidelines (11945/215) for pesticide analysis. 5 EXPERIMENTAL The LC and GC suites of pesticides analyzed in this study (listed in the Appendix) were chosen to cover a wide range of different pesticide classes and chemistries. The multi residue MS/MS methods were generated using Quanpedia, with separate databases utilized for generation of the LC and GC methods. Each database contains MRMs and retention time information for each compound. When the MS method is generated the MRM function windows are automatically set for each compound. For the UPLC method, a window of 1 minute was placed around each compound's expected retention time. For the APGC method, a window of 3 seconds was used due to the narrower peak widths exhibited in GC analysis. In addition to the MS methods, TargetLynx data processing methods and the LC inlet method were also generated through the Quanpedia Database. Sample extraction and cleanup Celery, lemon, corn, and kale samples were purchased at a local grocery store. Samples were chosen to be representative of different types of matrix complexity from different commodity groups, including high water content (celery and kale), high acid content (lemon), and high starch/protein with low water content (corn). Samples were immediately homogenized in a food processer and frozen until sample preparation was performed. QuEChERS extraction was performed according to the official AOAC method 27.1 using the DisQuE QuEChERS, AOAC Method Sample Preparation Kit (P/N ). 6 Figure 1 highlights the sample extraction. Weigh 15 g of homogenized sample Add 15 ml acidified acetonitrile (1 acetic acid) Add pouch of AOAC DisQuE salts Shake 1 min, centrifuge for 5 min at 4 rpm Sample MgSO4 PSA GCB Volume Part number Celery 15 mg 25 mg 7.5 mg 1 ml Lemon 15 mg 25 mg 1 ml Corn 15 mg 25 mg 1 ml Kale 9 mg 15 mg 15 mg 6 ml Table 1. dspe cleanup conditions used for each sample matrix. Perform DisQuE cleanup according to Table 1 Shake for 1 min, centrifuge for 5 min at 6 rpm For LC analysis: µl of supernatant + 4 µl water For GC analysis: Evaporate µl supernatant and reconstitute to µl in hexane Figure 1. DisQuE sample extraction method. [ 6 ] UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform

9 LC-MS/MS conditions LC system: ACQUITY UPLC H-Class Column: ACQUITY BEH C µm 2.1 x mm Column temp.: 45 C Injection volume: 5 µl Flow rate:.45 ml/min Mobile phase A: Water + 1 mm ammonium acetate Mobile Phase B: Methanol + 1 mm ammonium acetate Gradient: Time (min) A B MS system: Xevo TQ-S micro Ionization mode: ESI+ Capillary voltage: 1 kv Desolvation temp.: 5 C Desolvation gas flow: L/hr Source temp.: 15 C GC-MS/MS conditions GC system: 789A Autosampler: CTC PAL Column: 3 m x.25 mm x.25 µm Rxi-5MS Carrier gas: Helium Flow rate: 2. ml/min Injection: Splitless Injector temp.: 28 C Injection volume: 1 µl Makeup gas: Nitrogen at 25 ml/min Transfer line temp.: 32 C Oven program: Rate Temp. Hold ( C/min) ( C) (min) MS system: Xevo TQ-S micro Ionization mode: API+ Ionization mechanism: Corona current: Cone gas flow: Auxiliary gas flow: Proton transfer (3 vials of water in source) 2 µa for first 3.5 min 3. µa for rest of run L/hr 25 L/hr Source temp.: 15 C UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform [ 7 ]

10 [ APPLICATION NOTE ] RESULTS AND DISCUSSION METHOD MANAGEMENT USING THE QUANPEDIA DATABASE Working with methods involving large numbers of compounds can be time consuming when done manually and is prone to errors when setting up time segmented acquisition. Quanpedia is a compound centric database, typically used for method generation, but can also function as a method management tool. Initial methods for this analysis were generated using existing UPLC and APGC databases (Figure 2). Retention time changes resulting from further method development or method changes wereupdated in the database. This allowed for immediate and automatic updates to be made in the MS and processing methods by just re-generating the methods in three simple clicks. Figure 2. Quanpedia databases that were used to manage the methods for both UPLC and APGC analysis demonstrating the three click workflow of method generation. [ 8 ] UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform

11 RAPID AND ROBUST DATA AQUISITION For successful analysis of large numbers of pesticides and their metabolites, it is important that the mass spectrometer can maintain sufficient sensitivity while acquiring MRM transitions with a fast scan speed to provide enough data points across each chromatographic peak (e.g. minumum of 12 points per peak). The fast scanning speeds of the TQ-S micro allow for this robust and rapid data acquisition while maintaining large retention time windows to accommodate any shift in retention time due to column maintenance (GC) or chromatography changes caused by the different matrices.6 Figure 3 highlights one of the busiest sections of the APGC MS Method. In this example, flutolanil is just one of approximately 3 pesticides (set across 3 channels, each acquiring at least two transitions per compound) eluting in a 1.5 minute time window. The dwell time calculated by the autodwell function to collect a minimum of 12 points per peak was.6 s. The resulting chromatogram of three replicate injections of.1 mg/kg of flutolanil in celery matrix can be seen in Figure 3. Even with the fast scanning speed, 19 points were collected across the peak and the RSD of three consecutive injections in matrix was 5.2. The same is true for the UPLC method used for this analysis. 324 > 282 (Flutolanil») 1.67e5 S/N:PtP= Time Figure 3. Demonstration of the fast scanning of the Xevo TQ-S micro demonstrating retention of peak quality at a fast scan time. UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform [9]

12 [ APPLICATION NOTE ] PESTICIDES IN MATRIX Matrix matched standards were prepared in celery, lemon, corn, and kale over a range of.1 to.5 mg/kg and replicate injections made using the UPLC and APGC methods. A summed MRM overlay of a selection of pesticides can be seen in Figure 4, showing.1 mg/kg in celery extract from both the (A) APGC and (B) UPLC analyses. The data were fitted with the best fit calibration; for the UPLC data, the response was shown to be linear whereas the APGC response over the range investigated was non-linear and so was fitted with a quadratic calibration. The majority of the compounds in both analysis methods had correlation coefficient (R 2 ) values of.995 or greater. Figure 5 shows the matrix matched calibration curves and the peak response at.1 mg/kg of a representative pesticide from each analysis method in the four matrices. Residuals from triplicate injections at each calibration point were within ±2. Ion ratios were also shown to be within 3 tolerance of the reference values. A B APGC UPLC Figure 4. Overlay of a selection of pesticides at.1 mg/kg analyzed in a celery extract on A. APGC, and B. UPLC. Figure 5. Matrix matched calibration curves and chromatograms for standards at.1 mg/kg for peaks from: A. APGC analysis of leptophos in celery and lemon; and B. UPLC analysis of carbofuran in corn and kale. [ 1 ] UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform

13 For convenience, all sample extracts were spiked at the default MRL of.1 mg/kg. Figure 6 demonstrates the percentage of pesticides in each method detected in the spiked matrices at.1 mg/kg. However many pesticides could also be detected at.1 mg/kg as demonstrated in Figure 5 showing leptophos (APGC compound) and carbofuran (UPLC compound) in the different matrices. The precision of the measurements was excellent with more than 9 of the detected pesticides exhibiting RSDs of peak area of less than 1 (n=3). The exception was the APGC analysis of the kale matrix which had more than 8 of pesticides exhibiting RSDs less than 1 (Figure 7). Percent detected in matrix 12. APGC UPLC Celery Lemon Corn Kale Solvent Figure 6. The percentage of pesticides detected in the.1 mg/kg standard for each matrix using both APGC and UPLC. 12. RSD in matrix APGC RSD <5 APGC RSD <1 UPLC RSD <5 UPLC RSD < Celery Lemon Corn Kale Solvent Figure 7. Percentage of compounds detected at.1 mg/kg in each matrix and associated RSDs. UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform [ 11 ]

14 [ APPLICATION NOTE ] CONCLUSIONS Complex multi residue pesticide analysis was demonstrated using both UPLC and APGC analysis on the same tandem quadrupole instrument (Xevo TQ-S micro). Instrument methods were generated and maintained using Quanpedia databases making method generation and maintenance fast and simple. Although the multi residue methods contained approximately 2 compounds each, the reliable scanning speed of the TQ-S micro produced accurate and precise measurements. The performance for the determination of pesticide residues analyzed in four matrices of varying complexity complied with the SANTE guidelines for pesticide residue analysis. Detection at the EU default maximum residue limit of.1 mg/kg was easily achieved for >99 of pesticides analyzed with good precision (RSDs <1) for most analytes in the food samples. Having the flexibility of the Universal Source architecture to provide access to both UPLC-MS/MS and GC-MS/MS on the same instrument, allows for an increase of laboratory efficiency, while maintaining required sensitivity and repeatability. References 1. D Shah, E McCall, G Cleland. Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides in a Complex Matrix Without Compromising Data Quality. Waters Application Note no EN. January, T Kovalczuk, M Jech, J Poustka, J Hajslova. UPLC-MS/MS: A Novel Challenge in Multiresidue Pesticide Analysis in Food, Analytica Chimica Acta, 577, M Tienstra, T Portoles, F Hernandez, J G J Mol. Fast Gas Chromatographic Residue Analysis in Animal Feed Using Split Injection and Atmospheric Pressure Chemical Ionisation Tandem Mass Spectrometry. J. Chrom. A. 1422, October, L Cherta, T Portoles, J Beltran, E Pitarch, J G Mol, F Hernandez. Application of Gas Chromatography- Mass Spectrometry with Atmospheric Pressure Chemical Ionization for the Determination of Multiclass Pesticides in Fruits and Vegetables. J. Chrom. A. 1314: , November, European Commission. SANTE/11945/215. Guidance Document on Analytical Quality Control and Method Validation Procedures for Pesticides Residues Analysis in Food and Feed. 215, rev.. 6. AOAC Official Method 27.1: Pesticide Residues in Foods by Acetonitrile Extraction and Partitioning with Magnesium Sulfate Waters, Xevo, ACQUITY, UPLC, MassLynx, and The Science of What's Possible are registered trademarks of Waters Corporation. DisQuE, Quanpedia, and TargetLynx are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. May EN AG-PDF [ 12 ] Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform

15 Appendix Pesticides in APGC Method 2-Phenylphenol Diclobenil Oxyfluorfen 4,4'-Methoxychlor olefin Dicloran Paclobutrazol Acetochlor Dimethachlor Parathion Acrinathrin Diphenamid Pebulate Alachlor Diphenylamine Penconazole Allidochlor Edifenphos Pendimethalin Anthraquinone Endosulfan ether Pentachloroaniline Atrazine Endosulfan II Pentachlorobenzonitrile Azinphos-ethyl Endosulfan sulfate Pentachlorothioanisole Azinphos-methyl Endrin aldehyde Permethrin, cis- Benfluralin EPN Permethrin, trans- Bifenthrin Ethalfluralin Phenothrin 1 Bioallethrin Ethion Phenothrin 2 Biphenyl Ethylan Phorate Bromfenvinphos Etofenprox Phosalone Bromfenvinphos-methyl Etridazole Phosmet Bromophos-ethyl Fenamiphos Piperonyl butoxide Bromophos-methyl Fenarimol Pirimiphos-ethyl Bromopropylate Fenchlorphos Pirimiphos-methyl Bupirimate Fenitrothion Prochloraz Captafol Fenpropathrin Procymidone Captan Fenson Prodiamine Carbophenothion Fenthion Profenofos Carfentrazone ethyl Fenvalerate 1 Profluralin Chlorfenapyr Fenvalerate 2 Propachlor Chlorfenvinphos Fipronil Propanil Chlorobenzilate Fluazifop-P-butyl Propisochlor Chloroneb Fluchloralin Propyzamide Chlorothalonil Flucythrinate 1 Prothiofos Chlorpropham Flucythrinate 2 Pyraclofos Chlorpyrifos Fludioxonil Pyrazophos Chlorpyrifos-methyl Fluquinconazole Pyridaben Chlorthal-dimethyl Flusilazole Pyridaphenthion Chlorthiophos 1 Flutolanil Pyrimethanil Chlorthiophos 2 Flutriafol Pyriproxyfen Chlorthiophos 3 Folpet Quinalphos Chlozolinate Fonofos Resmethrin 1 Clomazone Hexachlorobenzene Sulfotep Coumaphos Hexazinone Sulprofos Cycloate Iodofenfos tau-fluvalinate 1 Cyfluthrin 1 Iprodione tau-fluvalinate 2 Cyfluthrin 2 Isazophos Tebuconazole Cyfluthrin 3 Isodrin Tebufenpyrad Cyfluthrin 4 Isopropalin Tefluthrin Cyhalothrin, lambda- Lenacil Terbacil Cypermethrin 1 Leptophos Terbufos Cypermethrin 2 Linuron Terbutylazine Cypermethrin 3 Malathion Tetrachloroaniline, 2,3,5,6- Cypermethrin 4 Metalaxyl Tetrachlorvinphos Cyprodinil Metazachlor Tetradifon DDD, o,p'- Methacrifos Tetramethrin 1 DDD, p,p'- Methoxychlor Tetramethrin 2 DDE, o,p'- Methyl parathion Tolclofos-methyl DDE, p,p'- Metolachlor Tolylfluanid DDT, o,p'- Mevinphos Transfluthrin DDT, p,p'- MGK Triadimefon Deltamethrin MGK Triadimenol Diallate Myclobutanil Triallate Diazinon N-(2;4-Dimethylphenyl)formamide Triazophos Dichlofluanid Nitralin Triflumizole Dichloroaniline, 3,4'- Nitrofen Trifluralin Dichlorobenzophenone, 4,4'- Oxadiazon Vinclozolin UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform [ 13 ]

16 [ APPLICATION NOTE ] Pesticides in UPLC Method Abamectin Etoxazole Nuarimol Acephate Famoxadone Omethoate Acetamiprid Fenamidone Oxadixyl Acibenzolar-S-methyl Fenarimol Oxamyl Aldicarb Fenazaquin Paclobutrazol Aldicarb sulfone Fenbuconazole Penconazole Aldicarb sulfoxide Fenhexamid Pencycuron Ametryn Fenobucarb Phenmedipham Aminocarb Fenoxycarb Picoxystrobin Amitraz Fenpropimorph Piperonyl butoxide Azoxystrobin Fenpyroximat Pirimicarb Benalaxyl Fenuron Procloraz Bendiocarb Fipronil Promecarb Benfuracarb Flonicamid Prometon Benzoximate Flufenacet Prometryn Bifenazate Flufenoxuron Propamocarb Bitertanol Fluomethuron Propargite Boscalid Fluoxastrobin Propham Bromuconazole I Fluquinconazole Propiconazole Bromuconazole II Flusilazole Propoxur Bupirimate Flutolanil Prothioconazole Buprofezin Flutriafol Pymetrozine Butafenacil Forchlorfenuron Pyracarbolid Butocarboxim Formetanate HCL Pyraclostrobin Butoxycarboxim Fuberidazole Pyridaben Carbaryl Furalaxyl Pyrimethanil Carbendazim Furathiocarb Pyriproxifen Carbetamide Hexaconazole Quinoxyfen Carbofuran Hexythiazox Rotenone Carbofuran-3-hydroxy Hydramethylnon Secbumeton Carboxin Imazalil Siduron Carfentrazone-ethyl Imidacloprid Simetryn Chlorantraniliprole Indoxacarb Spinetoram Chlorfluazuron Ipconazole Spinosad A Chloroxuron Iprovalicarb I Spinosad D Chlortoluron Iprovalicarb II Spirodiclofen Clethodim I Isocarbofos Spirotetramat Clofentezine Isoprocarb Spiroxamine I Clothianidin Isoproturon Spiroxamine II Cyazofamid Kresoxim-methyl Sulfentrazone Cycluron Linuron Tebuconazole Cymoxanil Lufenuron Tebufenozide Cyproconazole I Mandipropamid Tebufenpyrad Cyproconazole II Mefenacet Tebuthiuron Cyprodinil Mepanipyrim Teflubenzuron Cyromazine Mepronil Temephos Desmedipham Mesotrione Terbumeton Diclobutrazol Metaflumizone Terbutryn Dicrotophos Metalaxyl Tetraconazole Diethofencarb Metconazole Thiabendazole Difenoconazole Methabenzthiazuron Thiacloprid Diflubenzuron Methamidophos Thiamethoxam Dimethoate Methiocarb Thidiazuron Dimethomorph I Methomyl Thiobencarb Dimethomorph II Methoprotryne Thiophanate-methyl Dimoxystrobin Methoxyfenozide Triadimefon Diniconazole Metobromuron Triadimenol Dinotefuran Metribuzin Trichlorfon Dioxacarb Mevinphos I Tricyclazole Diuron Mevinphos II Trifloxystrobin Emamectin benzoate Mexacarbate Triflumizole Epoxiconazole Monocrotophos Triflumuron Etaconazole Monolinuron Triticonazole Ethiofencarb Myclobutanil Vamidothion Ethiprole Neburon Zoxamide Ethirimol Nitenpyram Ethofumesate Novaluron [ 14 ] UPLC and APGC Multi Residue Pesticide Analysis on a Single Tandem Quadrupole Mass Spectrometer Platform

17 [ APPLICATION NOTE ] Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS Dimple Shah, Naren Meruva, and Gareth Cleland Waters Corporation, Milford, MA, USA APPLICATION BENEFITS Multiresidue method for 19 pesticides in fruit and vegetable commodities showing excellent recoveries and precision at or below EU MRL. Detection of incurred residues confirmed in compliance with SANTE 11945/215 guidelines. Method readily extendable for additional pesticides of interest, using DisQuE variations, MS/MS method database, and easily updated reporting criteria. WATERS SOLUTIONS ACQUITY UPLC H-Class System Xevo TQ-S micro INTRODUCTION Pesticides are widely used in agricultural farming across the world. Pesticide residue levels in food products are regulated and closely monitored. Most countries have established legislation imposing Maximum Residue Limits (MRLs) for pesticide residues in food commodities which require analytical techniques that are sensitive, selective, accurate, and robust. Multiresidue analysis is challenging due to the low limits of detection required to achieve MRL compliance for a diverse range of pesticides in a wide range of food commodities. There are currently in excess of pesticides commercially available, and laboratories are under increasing pressure to widen the scope of their analytical methods for routine pesticide monitoring. In this application note, we describe the development of a multiresidue method for the routine determination of 19 pesticide residues in various fruit and vegetable matrices using UPLC-MS/MS. White grapes and green beans were selected for this application due to their relatively increased complexity, since they contain chlorophylls which are known to affect the extraction efficiency of pesticides from food matrices. A generic QuEChERS sample extraction procedure was used to extract the pesticides from the fruit and vegetable samples, followed by rapid and high resolution UPLC separation and trace level detection of pesticides using the Xevo TQ-S micro. ACQUITY BEH C18 Column MassLynx MS Software TargetLynx Application Manager Quanpedia Database DisQuE QuEChERS Dispersive Solid Phase Extraction KEYWORDS Pesticides, green beans, white grapes, QuEChERS, food safety, MRL Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS [ 15 ]

18 [ APPLICATION NOTE ] EXPERIMENTAL UPLC conditions LC system: ACQUITY UPLC H-Class Column: ACQUITY BEH C µm, 2.1 x mm Column temp.: 45 C Injection volume: 1 µl Flow rate:.45 ml/min Mobile phase A: 1 mm Ammonium acetate (ph 5) in water Mobile phase B: 1 mm Ammonium acetate (ph 5) in methanol Weak needle wash: 5/5 Water/methanol (v/v) Strong needle wash: 1/9 Methanol/water (v/v) Seal wash: 9/1 Water/methanol Gradient: Time Flow rate (min) (ml/min) A B Curve Initial MS conditions MS system: Xevo TQ-S micro Two MRM transitions for each compound were obtained from the Quanpedia Database 1 which contains a compendium of methods, and monitored for all of the pesticides that were studied. The data were acquired and processed using MassLynx MS Software v.4.1 with TargetLynx XS Application Manager. Standards Waters LC Multiresidue Pesticide Standards Kit (p/n ) was used to make a mix of calibration standards. The stock solution of 1 µg/ml was prepared by combining µl from each ampoule. Sample description The green beans and white grape samples investigated in this study were purchased from a local supermarket. Sample preparation 15 g of homogenized samples were extracted with 15 ml of 1 glacial acetic acid in acetonitrile, followed by the addition of QuEChERS AOAC material (DisQuE Pouches, p/n ). The tube was shaken for 1 minute and centrifuged at 37 rpm for 5 minutes. Then µl of the extract was taken and diluted to 1 ml with water before LC-MS/MS analysis. To study linearity, solvent and matrix matched standards (MMS) calibration curves were created by spiking the pesticide mix from.1 to.5 mg/kg (1 ppb to 5 ppb) in solvent and the matrices, respectively. Ionization mode: Capillary voltage: ESI+ 1. kv Desolvation temp.: 5 C Desolvation gas flow: L/Hr Source temp.: 15 C [ 16 ] Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS

19 RESULTS AND DISCUSSION All pesticides were analyzed on an ACQUITY UPLC BEH C18 Column. For each pesticide, two MRM transitions were monitored, and AutoDwell was applied. Autodwell is a feature used in MassLynx Software to ensure that sufficient points across each chromatographic peak are achieved. A user simply enters the average peak width and number of points required, and the software automatically calculates the dwell time required to achieve the minimum number of points across the peak specified. With the rapid acquisition rate of the Xevo TQ-S micro, 38 MRMs were monitored with a 1 minute wide retention time window with at least 12 data points across the peak. False negatives are potentially avoided by extending the acquisition range. A B Figure 1 shows an overlay of chromatograms (vertically linked) for all pesticides spiked at 1 ppb (.1 mg/kg) in the green beans and white grapes. LINEARITY Linearity was studied with nine different levels of matrixmatched standards calibration. The concentrations of the calibration levels ranged from 1 to 5 ppb (sample equivalent to.1 to.5 mg/kg). A majority of the compounds (96) showed a linear response with correlation coefficients >.99 in both matrices. Example calibration curves of imidacloprid in white grapes and pyraclostrobin in green beans are shown in Figure 2. Figure 1. Total ion chromatogram of pesticides in A. green beans, and B. white grapes. Compound name: Imidacloprid Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None A 55 RECOVERY Method recovery was evaluated by spiking the reference standards in the samples and quantifying against the matrix-matched calibration curve. Green bean and white grape samples were pre-spiked with all of the pesticides at 1 ppb (.1 mg/kg) in triplicate. The samples were extracted and quantified against the matrix matched calibration curve. Recoveries were calculated using TargetLynx XS Software. Recoveries for most of the pesticides (97 in white grapes and 96 in green beans) fell within the acceptable tolerance of 7 to 12 range (DG SANTE/11945/215) 2 in both samples. The precision in terms of RSD for most compounds (93) in green beans and white grapes were less than 2. The use of a suitable internal standard will significantly improve repeatability for those analytes. Recoveries for all the pesticides in green beans and white grapes are shown in Appendix A. Response Response ppb Compound name: Pyraclostrobin Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None B ppb Figure 2. Matrix matched calibration curve of A. imidacloprid in white grapes, and B. pyraclostrobin in green beans. Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS [ 17 ]

20 [ APPLICATION NOTE ] SAMPLE ANALYSIS To determine incurred residues, the white grape and green bean samples were prepared as described in the Sample preparation section and analyzed. From the obtained results, carbendazim, propamocarb, and pyrimethanil were observed in the green bean sample and quantified below.7 mg/kg. Boscalid, cyprodinil, fenhexaid, imidacloprid, methoxyfenozide, pyraclostrobin, and trifloxystrobin were observed in the grape sample and quantified at less than.25 mg/kg. For accurate quantification of incurred residues, a standard addition technique can be employed. Figure 3 shows an example of an incurred residue found in the green bean sample. In order to avoid false identification, it is important to check the retention time tolerance and ion ratios of the incurred residues. TargetLynx automatically calculates ion ratios and provides accurate quantification for all incurred residues. All of the incurred pesticide residues detected in the green bean and white grape samples were identified in accordance with the criteria specified in the European Commission SANTE document 11945/215, 2 [retention time (±.1 minute) and ion ratios (<3 )] and compared against the reference. Figure 3. Carbendazim identified in the green bean sample. [ 18 ] Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS

21 CONCLUSIONS A broad range of pesticides (19) of interest were successfully analyzed using an easy sample preparation procedure (QuEChERS) and UPLC-MS/MS determination in green bean and white grape samples. This method can easily detect all of the listed pesticides at.1 mg/kg which is at or below the maximum residue level limit (MRL). More than 95 of the pesticides were detected with recoveries within the range of 7 to 12 using a QuEChERS extraction procedure. TargetLynx Application Manager provides efficient and automated data processing tools for calculating recoveries and quantification of incurred residues. References 1. Quanpedia Database: A Compendium of Compounds and Analyses for Rapid and Simple Multi-Residue LC-MS/MS Method Development. Waters Technology Brief No. 7244en Guidance document on the analytical quality control and method validation procedures for pesticides residues analysis in food and feed. 3. SANTE/11945/215: food/plant/docs/pesticides_mrl_guidelines_ wrkdoc_11945.pdf Waters, ACQUITY, UPLC, Xevo, MassLynx, and The Science of What's Possible are registered trademarks of Waters Corporation. TargetLynx and DisQuE are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. June EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS [ 19 ]

22 [ APPLICATION NOTE ] Appendix A Name RT White grapes Recovery.1 mg/kg (n=3) RSD Recovery.1 mg/kg (n=3) Green grapes RSD Acephate Acetamiprid Acibenzolar-S-methyl Aldicarb Aldicarb sulfone Aldicarb sulfoxide Ametryn Aminocarb Azoxystrobin Benalaxyl Bendiocarb Benzoximate Bifenazate Bitertanol Boscalid 8.67 * Bromuconazole I Bromuconazole II Bupirimate Buprofezin Butafenacil Butocarboxim Butoxycarboxim Carbaryl Carbendazim * Carbetamide Carbofuran Carbofuran-3-hydroxy Carboxin Carfentrazone-ethyl Chlorantraniliprole Chlorfluazuron Chloroxuron Chlortoluron Clethodim I Clethodim II Clofentezine Clothianidin Cyazofamid Cycluron Cymoxanil Cyproconazole I [ 2 ] Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS

23 Name RT White grapes Recovery.1 mg/kg (n=3) RSD Recovery.1 mg/kg (n=3) Green grapes RSD Cyproconazole II Cyprodinil 9.66 * Cyromazine Desmedipham Diclobutrazol Dicrotophos Difenoconazole I Difenoconazole II Diflubenzuron Dimethoate Dimethomorph I Dimethomorph II Dimoxystrobin Diniconazole Dinotefuran Diuron Emamectin benzoate I Emamectin benzoate II Epoxiconazole Eprinomectin Etaconazole I Etaconazole II Ethiofencarb Ethiprole Ethofumesate Famoxadone Fenamidone Fenarimol Fenazaquin Fenbuconazole Fenhexamid 9.16 * Fenobucarb Fenoxycarb Fenpropimorph Fenpyroximat Fenuron Fipronil Flonicamid Flufenacet Flufenoxuron Fluomethuron Fluoxastrobin Fluquinconazole Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS [ 21 ]

24 [ APPLICATION NOTE ] Name RT White grapes Recovery.1 mg/kg (n=3) RSD Recovery.1 mg/kg (n=3) Green grapes RSD Flusilazole Flutolanil Flutriafol Forchlorfenuron Formetanate HCL Fuberidazole Furalaxyl Furathiocarb Halofenzoide Hexaconazole Hexythiazox Hydramethylnon Imazalil Imidacloprid 4.21 * Indoxacarb Ipconazole I Ipconazole II Iprovalicarb I Iprovalicarb II Isocarbofos Isoprocarb Isoproturon Ivermectine Kresoxim-methyl Linuron Lufenuron Mandipropamid Mefenacet Mepanipyrim Mepronil Mesotrione Metalaxyl Metconazole Methabenzthiazuron Methamidophos Methiocarb Methomyl Methoprotryne Methoxyfenozide 8.85 * Metobromuron Metribuzin Mevinphos I Mevinphos II [ 22 ] Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS

25 Name RT White grapes Recovery.1 mg/kg (n=3) RSD Recovery.1 mg/kg (n=3) Green grapes RSD Mexacarbate Monocrotophos Monolinuron Myclobutanil Neburon Nitenpyram Novaluron Omethoate Oxadixyl Oxamyl Paclobutrazol Penconazole Pencycuron Phenmedipham Picoxystrobin Piperonyl butoxide Pirimicarb Procloraz Promecarb Prometon Prometryn Propamocarb * Propargite Propham Propiconazole I Propiconazole II Propoxur Pymetrozine Pyracarbolid Pyraclostrobin 9.97 * Pyridaben Pyrimethanil * * Pyriproxifen Secbumeton Siduron Simetryn Spinosad A Spirodiclofen Spiromesifen Spirotetramat Spiroxamine I Spiroxamine II Sulfentrazone Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS [ 23 ]

26 [ APPLICATION NOTE ] Name RT White grapes Recovery.1 mg/kg (n=3) RSD Recovery.1 mg/kg (n=3) Green grapes RSD Tebuconazole Tebufenozide Tebuthiuron Teflubenzuron Terbumeton Terbutryn Tetraconazole Thiabendazole Thiacloprid Thiamethoxam Thidiazuron Thiobencarb Triadimefon Triadimenol Trichlorfon Tricyclazole Trifloxystrobin 1.42 * Triflumizole Triflumuron Triticonazole Vamidothion Zoxamide * Incurred residue - Not detected at spiked level [ 24 ] Multiresidue Analysis of Pesticides in Fruits and Vegetables Using UPLC-MS/MS

27 [ TECHNOLOGY BRIEF ] Highly Sensitive Analysis of Polar Pesticides in Food Matrices on the Xevo TQ-XS Benjamin Wuyts,¹ Dimple Shah,² Euan Ross,¹ Jonathan Fox,¹ Eimear McCall¹ ¹Waters Corporation, Wilmslow, UK, ²Waters Corporation, Milford, MA, USA While discussions on the toxicological concerns of glyphosate and associated compounds continue, maximum residue limits (MRLs) are enforced globally, varying on commodity of interest.¹ GOAL Develop a sensitive LC-MS/MS method for the robust analysis of highly polar pesticides in food commodities without the need of derivatization. mg/kg LLOQ AMPA.2 Glufosinate.9 Glyphosate.1 BACKGROUND Glyphosate is a non-selective broad spectrum herbicide, which accounts for more than 6 of global herbicide sales. While discussions on the toxicological concerns of glyphosate and associated compounds continue, maximum residue limits (MRLs) are enforced globally, varying on commodity of interest.¹ Due to certain physicochemical characteristics of glyphosate, its metabolites and similar compounds, generic catch-all multiresidue methods are not appropriate. Traditionally, derivatization has been utilized to allow for chromatographic retention on reverse phase columns. While good sensitivity and repeatability have been reported using this technique,² there is an interest in direct analysis to avoid timely sample preparation and derivatization. Ion chromatography offers an option for the direct analysis of such polar compounds; however, specialized equipment and the use of non-ms compatible solvent additives can be required. Figure 1. Example of chromatography for 1. AMPA, 2. glufosinate, and 3. glyphosate, respectively at.2 mg/l in solvent. THE SOLUTION Using Waters latest tandem quadrupole technology, the Xevo TQ-XS, the sensitivity and repeatability was investigated for the simplified analysis of underivatized polar pesticides. A mixed mode, weak ion exchange/reverse phase analytical column allowed for MS friendly mobile phase of 5 mm aqueous ammonium formate (ph 2.9) to be readily used in electrospray negative ion mode.³ Waters Xevo TQ-XS, coupled with the ACQUITY UPLC System achieves sub parts per billion detection limits for glyphosate, glufosinate and aminomethylphosphonic acid (AMPA), as shown in Figure 1. Here an example of the chromatography obtained from the analysis of the three analytes of interest is shown at.2 mg/l in water, while inset shows the limit of quantitation achieved for each analyte based on signal-to-noise (S/N) 1. Highly Sensitive Analysis of Polar Pesticides in Food Matrices on the Xevo TQ-XS [ 25 ]

28 [ TECHNOLOGY BRIEF ] System robustness, in terms of precision was observed from the analysis of a selection of spiked extracts of cereal-based foods and beverage. Extracts were prepared using a simple aqueous extraction, pass through SPE, spiked at.25 mg/ kg, filtered and injected into the ACQUITY UPLC System coupled to Xevo TQ-XS. A summary of the results, run over almost 17 hours, and acquired from replicate injections in a single batch, is provided in Table 1, while an example of the chromatography achieved for each analyte in lentils is shown in Figure 2. Linearity was evaluated for all analytes, in solvent and matrix. Excellent linearity was achieved using calibration curves bracketing the analytical run. An example is shown in Figure 3, where two curves are overlaid showing an R².995 with residuals <15 for all concentration levels (.5 to 1 mg/ kg). When comparing the slopes of the solvent and matrix matched calibration curves, significant matrix effects (namely ion suppression) were observed, especially for AMPA, which could be mitigated by the use of stable isotope analogue internal standards. Table 1. Summary of precision data (RSD) from the analysis of extracts from a v ariety of cereal based products, spiked at.25 mg/kg. Wheat Lentils Barley Beer n = Glyphosate Glufosinate AMPA Time Figure 2. Example of chromatography for 1. AMPA, 2. glufosinate and 3. glyphosate, respectively at.25 mg/kg in lentils. 3 SUMMARY A fast and simple method has been demonstrated for the analysis of highly polar pesticides in extracts from a variety of foodstuffs. The sensitivity of the Xevo TQ-XS provides excellent limits of detection, suitable for monitoring MRL compliance for these challenging analytes and good quantitative performance, in terms of precision and calibration characteristics, even in the absence of internal standard, in compliance with SANTE 11954/215 guidelines. References Ehling S and Reddy T M. J Agric Food Chem. 63: , Chamkasem N, Morris C, Harmon T. J Reg Sci. 2: 2-26, 21 Compound name: Glyphosate Correlation coefficient: r = , r^2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Residual Response ppb Figure 3. Example of calibration achieved for glyphosate in solvent from.5 to 1 mg/l. ppb Waters, ACQUITY UPLC, and The Science of What s Possible are registered trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 216 Waters Corporation. Produced in the U.S.A. October EN TC-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 26 ] Highly Sensitive Analysis of Polar Pesticides in Food Matrices on the Xevo TQ-XS

29 [ APPLICATION NOTE ] Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides in a Complex Matrix Without Compromising Data Quality Dimple Shah, 1 Eimear McCall, 2 and Gareth Cleland 1 1 Waters Corporation, Milford, MA, USA 2 Waters Corporation, Wilmslow, UK APPLICATION BENEFITS Single method for confirmation and quantification of a large number of pesticides in chili powder. The Xevo TQ-S micro provides reliable, robust, and reproducible data at fast acquisition rates. Automated quantification of samples with Auto Addition and standard addition features. WATERS SOLUTIONS ACQUITY UPLC H-Class System Xevo TQ-S micro ACQUITY UPLC BEH C 18 Column INTRODUCTION Hundreds of compounds are routinely used for crop protection across the globe. With increasing global trade there is a requirement for rapid multiresidue screening and quantification methods to efficiently determine residue violations and protect consumers. Effective multi-residue methods rely on management of the acquisition of a large number of MRM transitions. Setting up overlapping MRM windows based around the retention time of each analyte ensures that no time is wasted acquiring other transitions for compounds that have yet to elute. This optimizes the time spent acquiring data to maximize sensitivity while ensuring a sufficient number of data points across peaks to give good precision. One of the objectives of this method was to implement relatively wide MRM windows without any loss in performance. This removes the need to make regular checks on retention time drift, and avoids having to make adjustments to the acquisition method before each analysis. In this application note, we describe a single, fast method for confirmation and quantification of more than 425 pesticides. All pesticides were analyzed on a reverse phase column within 12 minutes. Two MRM transitions for each pesticide were monitored using both ESI positive and negative modes (deploying polarity switching). The excellent performance of the method has been demonstrated at very low concentrations in chili powder, which is a very complex matrix. Standard curves in solvent and matrix, easy-to-use instrument and software features to calculate the incurred residues in a chili sample, and data showing robustness after a large number of injections will be presented. MassLynx MS Software DisQuE QuEChERS Dispersive Solid Phase Extraction LC Multiresidue Pesticide Standards Kit KEY WORDS Pesticides, chili, QuECheRS, food safety Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides [ 27 ]

30 [ APPLICATION NOTE ] EXPERIMENTAL UPLC conditions LC system: Time (min) Flow rate (ml/min) A B Curve Initial Table 1. UPLC method for pesticide analysis. MS conditions MS system: Ionization mode: ESI +/- Capillary voltage: Desolvation temp.: 5 C Desolvation gas flow: Xevo TQ-S micro 1 kv(+) and.5 kv(-) L/hr Source temp.: 15 C Acquisition: ACQUITY UPLC H-Class Column: ACQUITY BEH C 18, 2.1 x mm, 1.7 µm Column temp.: 45 C Injection volume: 1 µl Flow rate:.45 ml/min Mobile phase A: 1 mm Ammonium Acetate, ph 5, (P/N ) in water Mobile phase B: Weak needle wash Strong needle wash: Seal wash: 1 mm Ammonium acetate in methanol 5/5 Water/methanol (v/v) 9/1 Methanol/water (v/v) 9/1 Water/methanol Multiple Reaction Monitoring (MRM) MS methods and data acquisition Two MRM transitions for each pesticide were generated using Quanpedia except for fipronil, where 1 MRM transition was used. Methods are shown in Table 2. The data were acquired using MassLynx Software, and processed using TargetLynx XS Application Manager. Method MRM transitions Dwell time Average peak width (sec) n=3 Average data points across peak n=3 A 16 Autodwell B 859 Autodwell Table 2. Methods A and B showing the number of MRM transitions and dwell times in the respective methods. Also listed are the average peak widths and data points across the peak of furalaxyl for three replicates of a chili sample spiked at 1 µg/kg. Standards Waters LC Multiresidue Pesticide Standards Kit (P/N ) was used to make a mix of standards. The standards kit has a collection of 24 compounds in 1 different ampoules at individual concentrations of µg/ml. The stock solution of 1 µg/ml was prepared by combining µl from each ampoule. The working standards were further diluted with acetonitrile. For the solvent calibration curve and standard addition work, acetonitrile was spiked with 24 pesticides at 1, 5, 1, 25, 5,, 25, 5, and µg/kg (ppb) (concentrations equate to sample). Sample preparation The sample preparation followed the protocol described in a previous application note. 1 A DisQuE QuEChERS (CEN method 15662) 2 was used to prepare all samples. Briefly, two grams of chili powder was weighed into a centrifuge tube. The sample was mixed with 8 ml of water and vortexed for 3 seconds. The mixture was extracted with 1 ml of acetonitrile followed by the addition of QuEChERS CEN material (4 g MgSO4, 1 g NaCl, and 1.5 g sodium citrate). The resulting mixture was shaken for 1 minute. The tube was then centrifuged at 4 rpm for 5 minutes, and the supernatant was placed into vials for analysis. For the matrix match spiked (MMS) calibration, a chili sample was spiked with 24 pesticides at the same level as the solvent calibration range. These spiked levels equaled the concentrations in the chili sample. [ 28 ] Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides

31 RESULTS AND DISCUSSION The chili sample extract was analyzed for the presence of pesticides using an ACQUITY UPLC H-Class System coupled with the Xevo TQ-S micro Mass Spectrometer. Novel SpaceWire technology facilitates faster acquisition speeds with Xcellerate Ion Transfer (XIT ). To achieve the additional sensitivity, the instrument is integrated with StepWave Ion Guide Technology. StepWave effectively removes the neutral molecules, providing additional sensitivity, and improving robustness. Figure 1 shows the Xevo TQ-S micro, along with the StepWave ion guide. In order to demonstrate the fast acquisition rate and excellent data quality of the MS instrument, a chili extract was post spiked with pesticides at 1 µg/kg (concentration equates to sample) and analyzed using Methods A and B, as described in Table 2. Method A contains 8 pesticides having two MRM transitions for each compound (16 MRMs total). Method B contains 43 pesticides (1 SRM, and 429 with two MRM transitions, totaling 859 MRMs). Both methods were enabled with AutoDwell functionality, at the click of a button. AutoDwell allows MassLynx MS Software to optimize the dwell time automatically for each compound depending on its retention time, as well as the peak width and required data points across the peak, as defined by the user. In this experiment, 1-minute wide acquisition windows were selected for both methods which eliminates the regular checks for retention time drift (due to matrix interferences) and simplifies inter- and/or intra- laboratory method transfer. Figure 1. Xevo TQ-S micro (left) and StepWave ion guide diagram (right). Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides [ 29 ]

32 [ APPLICATION NOTE ] Figure 2 shows a screen shot of part of Method B. In this method, more than pesticides (of the 43 being monitored) eluted between 9 and 1 minutes. Furalaxyl (RT 9.1 min) has eluted within this crowded region. As shown in Table 2, furalaxyl had an average peak width and data points across the peak of 4.4 sec and 1 data points respectively (Method B), for three replicates of the chili sample spiked at 1 µg/kg. Similar results were observed with Method A where fewer transitions (16) were monitored in the method. Despite the large number of compounds in Method B, the data quality was not compromised in the complex matrix. 1-minute wide window Figure 2. Screen capture of Method B. Figure 3 shows chromatograms of furalaxyl acquired by Methods A and B. The peak area difference between Methods A and B which contained 16 and 859 MRM transitions respectively were minimal. Despite Method B having significantly more MRMs than Method A, both methods yielded similar area counts (<2 deviation), which is an acceptable deviation for injection of replicate matrix. Both methods have shown a minimum of 1 data points across the peak, a typical requisite for accurate quantification. Method A 16 MRMs Method B 859 MRMs Furalaxyl 32.1>95 Furalaxyl 32.1>95 Figure 3. Chromatograms of furalaxyl showing peak area count (Methods A and B) in a spiked chili powder sample at 1 µg/kg. [ 3 ] Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides

33 Linearity Linearity over the required working range was determined in solvent (data not shown) and matrix matched calibration curves using both MS methods. Figure 4 shows excellent agreement and linearity for matrix matched calibration curve of methoxyfenozide (RT 9.56 min) from 1 to µg/kg (ppb) using methods A and B. As can be seen in the Figure 4, Methods A and B showed good linearity with an r 2 value >.999 over the entire specified range and similar slope of the equation. Compound name: Methoxyfenozide Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Method A 25 2 Response Conc Compound name: Methoxyfenozide Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Method B Response ppb Figure 4. Matrix match spiked calibration curves of methoxyfenozide using Methods A and B. Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides [ 31 ]

34 [ APPLICATION NOTE ] Robustness study To assess the repeatability and robustness of the method, 3 injections of the spiked chili sample (25 µg/kg) were analyzed by Method B. Figure 5 shows the RSD of the peak area for example pesticide residues. As shown in Figure 5, all positive and negative ionized compound showed good RSD (3.3 to 13.9) over 3 injections. Despite the large number of MRM transitions employing polarity switching, Method B showed excellent reproducibility and quantification at a lower concentration in a complex matrix like chili RSD Figure 5. RSD of 3 injections of the example pesticides spiked in chill sample at 25 µg/kg and analyzed by Method B. Standard addition using the Auto Addition feature A few incurred residues were found when the chili sample was screened with Method B. The standard addition method was employed to calculate the concentration of incurred residues in the chili sample. Matrix matched samples were prepared using the Auto Addition functionality of the UPLC System, automatically enabling the repeatable mixing of multiple aliquots from several vials within a single injection. The extract was diluted directly in the sample loop before injection. In this study, Auto Addition provided an automated calibration curve for the chili sample by spiking blank chili extract with various concentration of pesticides. Figure 6 shows the Auto Addition setup sample list created in MassLynx Software. 1 st vial 1 st volume 2 nd vial 2 nd volume Figure 6. Sample list created in MassLynx showing Auto Addition setup. As shown in Figure 6, vials 1:6 and 1:4 contain acetonitrile and a blank chili extract respectively. A mix of pesticide standards in acetonitrile, ranging from 1 to 5 µg/kg (ppb), were placed in individual vials (1:6 to 1:25). Figure 6 also shows the injection order that starts with drawing 2.5 µl of non-spiked extracted chili sample followed by 2.5 µl of acetonitrile. The rest of the injections start by drawing 2.5 µl of the non-spiked extracted chili sample and 2.5 µl of various concentrations of pesticide standards. In this instance, a 5-µL injection total volume was kept constant to ensure accurate and reproducible injections. It also maintains a constant amount of matrix to allow for accurate quantification. [ 32 ] Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides

35 After the data acquisition, incurred residues were then automatically quantified using the Standard Addition functionality in TargetLynx XS Software. Figure 7 shows an example of the calculated concentration of pyraclostrobin (12.35 ppb) in a chili sample using the standard addition approach. Figure 7. TargetLynx XS showing quantification of incurred pyraclostrobin in a chili sample by standard addition method. Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides [ 33 ]

36 [ APPLICATION NOTE ] CONCLUSIONS Waters LC Multi-Residue Pesticide Standards Kit is a quality assured collection of 24 compounds that eliminates the need for monotonous manual measurement of individual standards, thus saving the chemist time and laboratory resources. A large number of MRM transitions with a one-minute wide acquisition window in Method B eliminates the manual retention time check process and allows for easy transfer of methods between laboratories. References 1. D Shah, E Goh, J Burgess. Rapid analysis of Sudan and other prohibited dyes in chili powder using the ACQUITY UPLC H-Class System with Xevo TQD. Waters application note no en. March, DisQuE Dispersive Sample Preparation brochure no en. July, 215. The fast scanning speed of the Xevo TQ-S micro provides enough data points across the peak for accurate quantification within high volume multi-residue analysis. A combination of the Auto Addition and standard addition features facilitates automated quantification of incurred residues which reduces labor and the need for a blank sample. Excellent linearity, robustness, and sensitivity were achieved in a complex matrix such as chili powder. Waters, ACQUITY UPLC, UPLC, Xevo, and The Science of What s Possible are registered trademarks of Waters Corporation. DisQuE, Quanpedia, XIT, and StepWave are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 216 Waters Corporation. Produced in the U.S.A. January EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 34 ] Single LC-MS/MS Method for Confirmation and Quantification of Over 4 Pesticides

37 [ APPLICATION NOTE ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra with Simultaneous Acquisition of Qualitative Full-Spectrum MS and MS/MS Data Dimple Shah, 1 Mark Benvenuti, 1 Antonietta Gledhill, 2 P. M. N. Rajesh, 3 and Jennifer A. Burgess 1 1 Waters Corporation, Milford, MA, USA 2 Waters Corporation, Manchester, UK 3 Waters Corporation, Bangalore, India APPLICATION BENEFITS Multiple pesticide residues can be detected simultaneously at legislative limits in okra samples using the ACQUITY UPLC H-Class System coupled with Xevo TQD MS. Quantitative and qualitative information can be achieved in a single injection. RADAR technology enables simultaneous full-scan data to be acquired, providing important information on matrix background ions that could potentially interfere with the analysis. PICS (product ion confirmation scan) provide additional confirmation for compound identification through acquisition of MS/MS spectra in the same injection. INTRODUCTION Okra is an important vegetable of the tropical countries and a popular diet component in several countries including India. According to the Food and Agriculture Organization of the United Nations (FAO), 1 India is one of the largest okra producers in the world and it produced approximately 5,8 tons of okra in 21 and 211. Okra is susceptible to a variety of pests and diseases 2 and a wide-range of pesticides are used to treat okra plants in India. Legislative limits are in place for the presence of pesticides in domestically produced, imported, or exported okra. 3 It is, therefore, very important to monitor the presence of commonly used pesticides in okra at legislative limits. According to the PRiF (Pesticide Residues in Food) report, import controls under regulation (EC) No 669/29 have been increased for okra imported from India because of the frequent detection of pesticide residues, mainly monocrotophos. The consignment is supposed to be rejected if it is noncompliant with MRLs (Maximum Residue Limits). Since July 1, 212, the frequency of testing consignments has been increased from 1 to 5. With this frequent testing, monocrotophos, triazophos, and acetamiprid were found at.2 mg/kg in okra samples from India, while the MRL for these compounds is.1 mg/kg. 4 In this application note, a multi-residue analysis method for the detection of 212 pesticides in okra is presented. For a complete list of all pesticides, see Appendix A. WATERS SOLUTIONS ACQUITY UPLC H-Class System Xevo TQD ACQUITY UPLC HSS T3 Column MassLynx Software Methods A multi-residue MS method for the pesticides was created using Waters Xevo TQD Quanpedia database. All of the pesticides were analyzed under ESI+ or ESI- mode using rapid polarity switching. Full-scan data were acquired in order to assess any matrix effects and the use of two MRMs and product ion confirmation scans were acquired to confirm and quantify the pesticide residues. KEY WORDS Pesticides, okra, QuECheRS, food safety Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra [ 35 ]

38 [ APPLICATION NOTE ] EXPERIMENTAL UPLC conditions LC system: ACQUITY UPLC H-Class Column: ACQUITY HSS T3 2.1 x mm, 1.8 µm Column temp.: 45 C Injection volume: 1 µl Flow rate:.45 ml/min Mobile phase A: 1 mm ammonium acetate (ph 5) in water Mobile phase B: 1 mm ammonium acetate (ph 5) in methanol Weak needle wash: 5/5 Water/ methanol (v/v) Strong needle wash: 1/9 Methanol/ water (v/v) Seal wash: 9/1 water/methanol Time Flow rate (min) (ml/min) A B Curve Initial Table 1. UPLC method for pesticide analysis. Standard preparation Pesticide standards were purchased either from Sigma-Aldrich, Fisher Scientific, or AccuStandard. A mix of all pesticides at 4 ng/ml was prepared in acetonitrile and stored at 4 C. Sample preparation QuEChERS is a popular method worldwide for the multi-residue analysis of pesticides in fruits and vegetables. The AOAC official method 27.1, was used to prepare okra samples that were purchased at a local supermarket. Briefly, okra samples were homogenized in water and 15 grams of homogenate was collected into a 5-mL centrifuge tube. Samples were extracted with acidified acetonitrile and mixed with MgSO 4 and NaCl (Tube 1). The tube was shaken for a minute and centrifuged at 15 rcf for 1 minute. After centrifugation, the matrix cleanup was accomplished by dispersive solid phase extraction (d-spe) by using 5 mg of primary secondary amine (PSA), 5 mg of C 18 bonded silica, 15 mg of MgSO 4, and 7.5 mg of graphitized carbon black (GCB). 5 1 ml of supernatant from Tube 1 was added to d-spe cleanup tube and centrifuged at 15 rcf for 1 minute. 1 ml of this extract was evaporated to dryness and reconstituted in 2 µl of 4/6 acetonitrile/water spiked with internal standard. RESULTS AND DISCUSSION All of the pesticides were successfully detected at 1 ppb (.1 mg/kg) in okra sample. For all of the pesticides, Appendix A lists the ionization mode, retention time, and whether or not the compound was detected in a prespike 1 ppb sample, as well as the 1 ppb pre-spike sample. Figure 1 shows an overlay of the total ion chromatogram (TIC) of all the pesticides at 1 ppb in okra sample. MS conditions MS system: Xevo TQD Ionization mode: ESI+/ESI- Capillary voltage: 3 kv Desolvation temp.: 5 C Desolvation gas flow: L/Hr Figure 1. Overlay of MRM chromatograms of all pesticides at 1 ppb (.1 mg/kg) in okra. Source temp.: 15 C [ 36 ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra

39 Solvent and matrix match spiked calibration (MMS) curves were prepared at concentrations that equated to the range 1 ppb to 5 ppb (i.e..1 to.5 mg/kg of okra) and injected in triplicate. The majority of the compounds showed linearity with R 2 values greater than.99 in both the solvent and MMS curves. Ethoxyquin, milbemectin A3, and A4, oxadiazon, spiromesifen, and terbufos showed R 2 values greater than.97 for both solvent and MMS curves. However, fipronil, phorate, and thiabendazole showed R 2 values greater than.97 in MMS curves only. Figures 2 and 3 show calibration curves and residuals for an example compound (triazophos) in solvent and matrix respectively. Compound name: Triazophos Correlation coefficient: r =.99876, r 2 = Calibration curve: * x Response type: Internal Std ( Ref 233 ), Area * ( IS Conc. / IS Area ) Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None 5. Residual. -5. ng/ml 3 Response 2 ng/ml Figure 2. Calibration curve of triazophos in solvent from 1 ppb to 5 ng/ml (.1 to.5 mg/kg). Compound name: Triazophos Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: Internal Std ( Ref 233 ), Area * ( IS Conc. / IS Area ) Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None 5. Residual. -5. ng/ml 4 Response 3 2 ng/ml Figure 3. Matrix-match spiked calibration curve of triazophos in okra sample from 1 ppb to 5 ppb (.1 to.5 mg/kg). Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra [ 37 ]

40 [ APPLICATION NOTE ] To evaluate the recovery, accuracy, and precision of the method, studies were carried out on spiked samples. Okra samples were pre-spiked with all the pesticides at 1 ppb (.1 mg/kg) in triplicate, extracted, and quantified against the MMS calibration curve. Recoveries were calculated using TargetLynx Software. The recoveries reported are without any internal standard correction. As shown in Figure 4 (A, B, C, and D), recoveries for all of the pesticides ranged from 25 to 15. Relative standard deviations (RSDs, shown as error bars in Figure 4) for most compounds were <2. The RSDs for 34 compounds were found to be higher than 2. Use of an internal standard would be likely to significantly improve repeatability for those analytes. A ,4-D 6-Benzyl Adenine Acephate Acetamiprid Acetochlor Acifluorfen Aldicarb Aldicarb sulfone Aldicarb sulfoxide Allethrin Atrazine Atrazine-desethyl Atrazine-desisopropyl Azoxystrobin Barban Benalaxyl Bendiocarb Benfuracarb Bensulfuron methyl Bifenazate Bitertanol Boscalid Bromacil Bromodiolone Buprofezin Butachlor Carbaryl Carbendazim Carbofuran Carbofuran-3-hydroxy Carbofuran-3-keto Carbosulfan Carboxin Chlorantraniliprole Chlorfenvinphos Chlormiuron ethyl Chlorpyrifos Chlorsulfuron Clothianidin Coumachlor Coumatetralyl Cruformate Cyazofamid Cycloxydim Cymoxanil Cyprazine Recovery Cyproconazole I Cyproconazole II Diafenthiuron Diazinon Dichlofluanid Dichlorvos B Diclofop methyl Difenoconazole I Difenoconazole II Difenoxuron Diflubenzuron Diflufenican Dimethoate Dimethomorph I Dimethomorph II Diniconazole Dinotefuran Dioxathion Diuron DMSA Edifenphos Emamec n benzoate Ethiofencarb Ethion Ethoxyquin Ethoxysulfuron Etrimfos Famoxadone Fenamidone Fenamiphos Fenarimol Fenazaquin Fenchlorphos oxon Fenobucarb Fenoxaprop- p Ethyl Fenoxycarb Fenpropathrin Fenpyroximat Fenthion Fenthion-sulfone Fenthion-sulfoxide Fipronil fipronil carboximide Fipronil desulfinyl Fipronil sulphide Fipronil sulphone Flonicamid Fluazafop-P-butyl Fluazifop Flubendazole Flufenacet Flufenoxuron Flufenzine fluopicolid Fluopyram Flusilazole Halosulfuron methyl Haloxyfop Hexaconazole Hexazinone Recovery C Recovery Hexythiazox Imazalil Imazaquin Imazosulfuron Imidacloprid Indoxacarb Iodosulfuron methyl Iprobenphos Iprodione Iprovalicarb Isoprothiolane Isoproturon Linuron Lufenuron Malaoxon Malathion Mandipropamid Mesosulfuron-methyl Metaflumizone Metalaxyl Methabenzthiazuron Methidathion Methiocarb Methomyl Metolachlor + S Metoxuron Metribuzin Metsulfuron methyl Mevinphos I Mevinphos II Milbemec n A3 Milbemec n A4 Molinate Monocrotophos Monolinuron Mycobutanil Novaluron Omethoate Oryzalin Oxadiargyl Oxadiazon Oxamyl Oxycarboxin Oxyflurofen Paclobutrazol Paraoxon-methyl Parathion ethyl Penconazole Pencycuron Pendimethalin Phenthoate Phorate Phorate sulfone Phorate sulfoxide 18 D 15 Recovery Phosalone Phosmet Phosphamidon Picoxystrobin Pirimicarb Pirimiphos-methyl pre lachlor Procloraz Profenofos Propanil Propetamphos Propiconazole Propoxur Pyraclostrobin Pyridalyl Pyrimethanil Pyriproxifen Pyrithiobac Na Quinalphos Quizalfop- Free acid Quizalfop- P-ethyl Rimsulfuron Simazine Spinosad A Spinosad D Spiromesifen Spirotetramat Spiroxamine Sulfosulfuron Tebuconazole Temephos Terbufos Tetraconazole Tetradifon Thiabendazole Thiacloprid Thiamethoxam Thiobencarb Thiodicarb Thiophanate methyl Tralkoxidym Triadimefon Triadimenol Triallate Triazophos Trichlorfon Tricyclazole Tridemorph Trifloxystrobin Triflumizole Tri conazole Vamidothion Figure 4. Recovery for 212 pesticides in okra sample at 1 ppb (.1 mg/kg). [ 38 ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra

41 Matrix effects Matrix effects for all of the pesticides were calculated by taking the ratio of the slope of the MMS calibration curve to the slope of solvent calibration curve. A percent variation of + 2 was considered as no matrix effect as this variation is close to the repeatability values. 6 Values between + 2 to + 5 were considered as a medium matrix effect, and a strong matrix effect was considered to be values greater than Figure 5 shows levels of the matrix effect that were observed in the analysis of okra for all pesticides. A strong matrix effect was observed for the majority of compounds, demonstrating that the analysis of okra samples poses a challenge in regards to high matrix complexity. Even with these high matrix effects, all compounds can easily be detected at legislative limits and quantified using the matrix-matched calibration curve No matrix effect Medium matrix effect Strong matrix effect Figure 5. Matrix effects observed for okra sample. Understanding matrix effects RADAR To further understand the impact of co-eluting matrix components that can compete with an analyte of interest during the ionization process, RADAR technology enables the simultaneous acquisition of full spectrum data during quantitative MS/MS analysis. Figure 6 shows an example of the use of RADAR technology. In Figure 6A, the base peak intensity (BPI) chromatogram from the full-scan background data for the okra sample is shown. At 5.8 minutes, close to the retention time of dimethoate (Figure 6B and 6C), high matrix interference was observed. The spectrum at 5.8 minute showed an intense ion at m/z (Figure 6D). A B C D! Figure 6. Use of RADAR Technology: (A) Full scan background data for okra sample, (B) and C) MRM transitions of dimethoate, (D) spectrum at retention time of dimethoate. Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra [ 39 ]

42 [ APPLICATION NOTE ] This interferent potentially has a large impact on the detection of dimethoate and a 48 ion suppression effect was observed for dimethoate. In the case of aldicarb, however, matrix interference was minimal (.4) and the RADAR data (Figure 7) showed no evidence of interferences at the retention time of aldicarb (6.13 minutes). The spectrum at the retention time of aldicarb has been expanded and zoomed in the inset (Figure 7D), clearly demonstrating that there was a much higher response from co-extracted matrix ions at the retention time of dimethoate compared to aldicarb. These data clearly demonstrate the usefulness of RADAR technology in assessing the matrix background and its potential effect on ion enhancement or suppression. A B C!!!! D Figure 7. Use of RADAR technology. (A) Full-scan background data for okra sample, (B) and (C) MRM transitions of aldicarb, D) Spectrum at retention time of aldicarb. The inset has been zoomed to show lower level response compared to the spectrum at the retention time of dimethoate. [ 4 ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra

43 Product ion confirmation (PICs) In complex matrices, situations arise where closely-related compounds such as metabolites or matrix interferences show responses for the target compounds of interest, even in MRM mode. This can lead to ambiguity and may require an additional qualitative experiment. An alternative is to employ a product ion confirmation scan (PICs) within the quantitative MRM experiment. PICs can be used to confirm peak identity through automatic acquisition of an MS/MS spectrum after the apex of the peak has eluted. PICs, in combination with TargetLynx, provides additional confirmation of the compounds of interest through comparison of the acquired MS/MS spectrum to a reference spectrum. Figure 8 shows the TargetLynx results from the comparison of the atrazine MS/MS spectrum obtained from PICS in an okra sample versus the reference spectrum, which was obtained from MS/MS analysis of the standard in solvent. Figure 8. Product ion confirmation (PICs) data for atrazine in okra sample. Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra [ 41 ]

44 [ APPLICATION NOTE ] CONCLUSIONS The combination of ACQUITY UPLC H-Class System with the Xevo TQD tandem mass spectrometer can detect pesticides below the legislative limit in okra samples. Even though a strong matrix effect was observed for many compounds, detection and quantification at the legislative limit was achieved. Simultaneous acquisition of MRMs and RADAR full-scan data provides quantitative and qualitative information in single injection. Product ion confirmation (PICs) increases confidence in compound assignments, which proves highly useful when working with complex matrices. References procedureokraeu.pdf 4. Documents/Results2and2Reports/212/Q Final.pdf 5. S J Lehotay, et al. Comparison of QuEChERS sample preparation methods for the analysis of pesticide residues in fruits and vegetables. J Chromatogr A (16): p SANCO/1684/29. Method validation and quality control procedures for pesticide residues analysis in food and feed. Document no. SANCO/3131/ F Carmen F, MJMartinez-Bueno, L Ana, AR Fernandez-Alba. Pesticide residue analysis of fruit juices by LC-MS/MS direct injection. One year pilot survey. Waters, ACQUITY UPLC, UPLC, Xevo, MassLynx, and The Science of What s Possible are registered trademarks of Waters Corporation. TargetLynx, Quanpedia, and RADAR are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 213 Waters Corporation. Produced in the U.S.A. September en AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 42 ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra

45 Appendix A In order to determine that the method was fit-for-purpose for the analytes listed, the analysis of prespiked samples at 1 ppb (.1 mg/kg) and 1 ppb (.1 mg/kg) was undertaken. All compounds were detected at 1 ppb. Those compounds that were also detected at 1 ppb are indicated in the fourth column. Some early eluting compounds showed compromised peak shapes, owing to the sample diluent (4 acetonitrile). Signal-to-noise improvements (and therefore lower LODs) can be gained from reducing the organic content of the sample diluent, however, some risk lies with ensuring that non-polar analytes remain in solution. For this work 4 organic was utilized. Atrazine desethyldesisopropyl, dinotefuran, methamidophos, and oxydemeton methyl showed compromised chromatographic peaks. In addition, for seven compounds, the second transition peak was not apparent at the lowest level. These compounds are shown by an asterisk in the table below. Name Ionization mode Retention time (minute) Pre-spike level detected 2,4-D ESI ppb 6-Benzyl Adenine* ESI ppb Acephate ESI ppb Acetachlor ESI ppb Acetamiprid ESI ppb Acifluorfen* ESI ppb Aldicarb ESI ppb Aldicarb sulfone ESI ppb Aldicarb Sulfoxide ESI ppb Allethrin ESI ppb Atrazine ESI ppb Atrazine ESI ppb desethyldesisopropyl Atrazine desisopropyl ESI ppb Atrazine-desethyl ESI ppb Azoxystrobin ESI ppb Barban/Barbamate* ESI ppb Bendiocarb ESI ppb Benalaxyl ESI ppb Benfuracarb ESI ppb Bensulfuron methyl ESI ppb Bifenazate ESI ppb Bitertanol ESI ppb Boscalid ESI ppb Bromacil ESI ppb Bromodialone ESI ppb Buprofezin ESI ppb Butachlor ESI ppb Carbaryl ESI ppb Carbendazim ESI ppb Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra [ 43 ]

46 [ APPLICATION NOTE ] Name Ionization mode Retention time (minute) Pre-spike level detected Carbofuran ESI ppb Carbofuran 3 keto ESI ppb Carbofuran-3-hydroxy ESI ppb Carbosulfan ESI ppb Carboxin ESI ppb Chlorantraniliprole ESI ppb Chlorfenvinphos ESI ppb Chlorimuron ethyl ESI ppb Chlorpyriphos / ESI ppb Dursban Chlorsulfuron ESI ppb Clothianidin ESI ppb Coumachlor ESI ppb Coumatetralyl ESI ppb Cruformate ESI ppb Cyazofamide/ ESI ppb cyazofamid Cycloxidim ESI ppb Cymoxanil ESI ppb Cyprazine ESI ppb Cyproconazole I ESI ppb Cyproconazole II ESI ppb Diafenthiuron ESI ppb Diazinon ESI ppb Dichlofluanid ESI ppb Dichlorvos ESI ppb Diclofop methyl ESI ppb Difenconazole I ESI ppb Difenconazole II ESI ppb Difenoxuron ESI ppb Diflubenzuron ESI ppb Diflufenican ESI ppb Dimethoate ESI ppb Dimethomorph I ESI ppb Dimethomorph II ESI ppb Diniconazole ESI ppb Dinotefuran ESI ppb Dioxathion ESI ppb Diuron ESI ppb DMSA ESI ppb Edifenfos ESI ppb [ 44 ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra

47 Name Ionization mode Retention time (minute) Pre-spike level detected Emamectin Benzoate ESI ppb Ethiofencarb ESI ppb Ethion ESI ppb Ethoxyquin ESI ppb Ethoxysulfuron ESI ppb Etrimphos ESI ppb Famoxadone ESI ppb Fenamidone ESI ppb Fenamiphos ESI ppb Fenarimole ESI ppb Fenazaquin ESI ppb Fenchlorphos-oxon ESI ppb Fenobucarb ESI ppb Fenoxaprop-p-ethyl ESI ppb Fenoxycarb ESI ppb Fenpropathrin ESI ppb Fenpyroximate ESI ppb Fenthion ESI ppb Fenthion sulfoxide ESI ppb Fenthion-sulfone ESI ppb Fipronil* ESI ppb Fipronil carboximide ESI ppb Fipronil desulfinyl ESI ppb Fipronil sulfone ESI ppb Fipronil sulphide ESI ppb Flonicamid ESI ppb Fluazifop ESI ppb Fluazifop-p-butyl ESI ppb Flubendazole ESI ppb Flufenacet ESI ppb Flufennoxuron ESI ppb (flufenoxuron) Flufenzine * ESI ppb Fluopicolide ESI ppb Fluopyram ESI ppb Flusilazole ESI ppb Halosulfuron-methyl ESI ppb Haloxyfop ESI ppb Hexaconazole ESI ppb Hexazinone ESI ppb Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra [ 45 ]

48 [ APPLICATION NOTE ] Name Ionization mode Retention time (minute) Pre-spike level detected Hexythiazox ESI ppb Imazalil ESI ppb Imazaquin ESI ppb Imazosulfuron ESI ppb Imidachloprid ESI ppb Indoxacarb ESI ppb Iodosulfuran-methyl ESI ppb Iprobenfos ESI ppb Iprodione ESI ppb Iprovalicarb ESI ppb Isoprothiolane ESI ppb Isoproturon ESI ppb Linuron ESI ppb Lufenuron ESI ppb Malaoxon ESI ppb Malathion ESI ppb Mandipropamid ESI ppb Mesosulfuron methyl ESI ppb Metaflumizone ESI ppb Metalaxyl ESI ppb Methabenzthiazuron ESI ppb Methamidophos ESI ppb Methidathion ESI ppb Methiocarb ESI ppb Methomyl ESI ppb Metolachlor + ESI ppb S-metolachlor Metoxuron ESI ppb Metribuzin ESI ppb Metsulfuron methyl ESI ppb Mevinphos I ESI ppb Mevinphos II ESI ppb Milbemectin A3* ESI ppb Milbemectin A4 * ESI ppb Molinate ESI ppb Monocrotophos ESI ppb Monolinuron ESI ppb Mycobutanil ESI ppb Novaluron ESI ppb Omethoate ESI ppb [ 46 ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra

49 Name Ionization mode Retention time (minute) Pre-spike level detected Oryzalin ESI ppb Oxadiargyl ESI ppb Oxadiazon ESI ppb Oxamyl ESI ppb Oxycarboxin ESI ppb Oxydemeton methyl ESI ppb Oxyfluorfen ESI ppb Paclobutrazole ESI ppb Parathion ethyl ESI ppb Paraxon methyl ESI ppb Penconazole ESI ppb Pencycuron ESI ppb Pendimethalin ESI ppb Phenthoate ESI ppb Phorate ESI ppb Phorate sulfone ESI ppb Phorate sulfoxide ESI ppb Phosalone ESI ppb phosmet ESI ppb Phosphamidon ESI ppb Picoxystrobin ESI ppb Pirimiphos methyl ESI ppb Pretilachlor ESI ppb Primicarb ESI ppb Prochloraz ESI ppb Profenofos ESI ppb Propanil ESI ppb Propetamphos ESI ppb Propiconazole (Tilt) ESI ppb Propoxur ESI ppb Pyraclostrobin ESI ppb Pyridalyl ESI ppb Pyrimethanil ESI ppb Pyriproxyfen ESI ppb Pyrithiobac sodium ESI ppb Quinalphos ESI ppb Quizalfop free acid ESI ppb Quizalfop-p-ethyl ESI ppb Rimsulfuron ESI ppb Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra [ 47 ]

50 [ APPLICATION NOTE ] Name Ionization mode Retention time (minute) Pre-spike level detected Simazine ESI ppb Spinosad A ESI ppb Spinosad D ESI ppb Spiromesifen ESI ppb Spirotetramat ESI ppb Spiroxamine ESI ppb Sulfosulfuron ESI ppb Tebuconazole ESI ppb Temephos ESI ppb Terbufos ESI ppb Tetraconazole ESI ppb Tetradifon ESI ppb Thiabendazole ESI ppb Thiacloprid ESI ppb Thiobencarb ESI ppb Thiodicarb ESI ppb Thiomethoxam ESI ppb (Thiamethoxam) Thiophanate methyl ESI ppb Tralkoxydim ESI ppb Triademefon ESI ppb Triademenol ESI ppb Triallate ESI ppb Triazophos ESI ppb Trichlorfon ESI ppb Tricyclazole ESI ppb Tridemorph ESI ppb Trifloxystrobin ESI ppb Triflumizole ESI ppb Triticonazole ESI ppb Vamidathion (Vamidothion) ESI ppb [ 48 ] Routine UPLC-MS/MS Quantification of Pesticide Residues in Okra

51 [ ALLERGENS ] [ 49 ]

52 [ ALLERGENS ] [ 5 ]

53 [ APPLICATION NOTE ] Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs Using UPLC-MS/MS Mélanie Planque, 1 Antonietta Wallace, 2 and Nathalie Gillard 1 1 CER Groupe, Marloie, Belgium; 2 Waters Corporation, Wilmslow, UK APPLICATION BENEFITS Sensitive multi-allergen method using UPLC-MS/MS. Allergens monitored in this method were assessed from the recommended levels provided by VITAL (Voluntary Incidental Trace Allergen Labelling) and the AOAC SMPR for food allergens (216.2). This multi-allergen detection method has the lowest limits of quantification available to date (expressed in total proteins and not soluble proteins). INTRODUCTION Food allergy is a worldwide health problem affecting both adults and children. To avoid allergic reactions, allergens must be totally excluded from the diet. Consequently, allergic customers can only refer to mandatory labeling to try and avoid coming into contact with the food allergen. However, the undeclared presence of these allergens is still widespread. To help food industries in the management of hidden allergens, sensitive, specific quantitative, and robust analytical methods need to be developed. Traditionally techniques such as ELISA and PCR have been used for routine analysis, but in recent years, there has been increasing interest in the utility of LC-MS based methods. In March 216, AOAC released the first standard method performance requirements (SMPR) specifically for the analysis of four food allergens using LC-MS/MS. 1 The detection levels tested are benchmarked against the levels stated in the AOAC SMPR and VITAL (Voluntary Incidental Trace Allergen Labelling) 2 reference doses. In this application note, we describe the targeted analysis of four food allergens in a variety of matrices using Waters ACQUITY UPLC System and Xevo TQ-S. A. B. WATERS SOLUTIONS ACQUITY UPLC System ACQUITY UPLC BEH13 BEH Column Xevo TQ-S MassLynx MS Software C. D. KEYWORDS Proteomics, allergens, LC-MS/MS, egg, peanut, milk, soybean Figure 1. 1A. MBP-fusion protein of the major peanut allergen Ara h 2: DOI: 1.221/pdb3ob4/pdb; 1B. Bovine allergen Bos d 2 in the trigonal space group P3221:DOI: 1.221/pdb4wfu/pdb; 1C. NMR solution structure of soybean allergen Gly m 4: DOI: 1.221/pdb2k7h/pdb; 1D. Crystal structure of uncleaved ovalbumin at 1.95 angstroms resolution: DOI: 1.221/ pdb1ova/pdb. Images courtesy of the RSCB Protein Data Bank. Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs [ 51 ]

54 [ APPLICATION NOTE ] EXPERIMENTAL This method is based on a single protocol applicable to the different tested allergens and foodstuffs. Details on the sample preparation step are described elsewhere. 3 The four allergens investigated in this method were milk (Bos Taurus), egg (Gallus gallus chicken), peanut (Arachis hypogaea) and soybean (Glycine Max (Glycine hispida). The protocol was tested on processed and complex food matrices including chocolate, ice cream, tomato sauce, and cookies. LC conditions LC system: ACQUITY UPLC Column: ACQUITY UPLC BEH13, 2.1 x 15 mm Column temp.: 4 C Sample temp.: 1 C Injection volume: 2 µl Flow rate:.2 ml/min Mobile phase A: Water +.1 formic acid Mobile phase B: Acetonitrile +.1 formic acid Gradient: to 1 min: 86 A; 1 to 16.5 min: 86 to 6 A; 16.5 to 16.6 min: 6 to A; 16.6 to 21 min: A; 21. to 21.1 min: to 86 A; 21.1 to 24 min: 86 A MS conditions MS system: Xevo TQ-S Data solutions Skyline (MacCoss Lab) UniProt MassLynx Table 1. Multiple reaction monitoring (MRM) parameters for the identification of milk, egg, soybean, and peanut proteins by ACQUTIY UPLC and Xevo TQ-S. Food Egg Peptide RT* (min) Precursor (charge state) (m/z) Product ion (fragment) Collision energy (ev) (y12+) 26 GGLEPINFQTAADQAR (++) (y1+) (y12+) (y9+) 31 LTEWTSSNVMEER (++) (y8+) (y7+) (y11+) 33 ISQAVHAAHAEINEAGR (++) (y1+) (y9+) (y7+) 19 EALQPIHDLADEAISR (+++) 69.3 (y6+) (y12++) (y9+) 15 NIPFAEYPTYK (++) 58.3 (y4+) (y9++) (y7+) 12 NIGELGVEK (++) (y6+) (y5+) (y7+) 15 YLLDLLPAAASHR (+++) (y11++) (y7++) (y7+) 14 NFLINETAR (++) 73.4 (y6+) (y5+) 16 (Table 1 continues on the next page.) Ionization mode : Capillary voltage: Collision gas flow: Cone voltage: Cone gas flow: Desolvation flow: ESI+ in MRM mode 2. kv.12 ml/min 35 V 15 L/h 12 L/h Source temp.: 15 C Desolvation temp.: 5 C [ 52 ] Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs

55 (Table 1 continued.) Food Peanut Soybean Milk Peptide RT* (min) Precursor (charge state) (m/z) NTLEAAFNAEFNEIR (++) RPFYSNAPQEIFIQQGR (+++) FNLAGNHEQEFLR (+++) TANELNLLILR (++) ISTLNSLTLPALR (++) EAFGVNMQIVR (++) ELINLATMCR (++) LITLAIPVNKPGR (+++) HQGLPQEVLNENLLR (+++) FFVAPFPEVFGK (++) YLGYLEQLLR (++) NAVPITPTLNR (++) VYVEELKPTPEGDLEILLQK (+++) VLVLDTDYK (++) LSFNPTQLEEQCHI Product ion (fragment) Collision energy (ev) (y9+) (y8+) (y7+) (y6+) (y1++) (b7+) (y5+) (y1++) (y9++) (y8+) (y7+) (y6+) (y9+) (y8+) (y7+) (y7+) (y6+) (y5+) (y7+) (y6+) (y (y7+) (y11++) (y9++) (y7+) (y6+) (b4+) (y9+) (y8+) (y6+) (y7+) (y6+) (y5+) (y8+) (y8++) (b3+) (y16++) (y14++) (y11++) (y7+) (y6+) (y5+) (y1+) 26 N-terminal peptide (++) (y7+) (y1++) 27 *Retention time (RT) is in sauce. Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs [ 53 ]

56 [ APPLICATION NOTE ] RESULTS AND DISCUSSION The software package Skyline, was used for in silico enzymatic digestion of food allergen proteins and to help produce potential MRMs for the experiment. From the list produced by Skyline, each MRM was analyzed using the ACQUITY UPLC System coupled to the Xevo TQ-S for sensitivity and reproducibility (in different food matrices). Milk (casein) FFVAPFPEVFGK Milk proteins.5 mg/kg 692.9> >92.5 Cookie Sauce Ice cream Chocolate I= I= I= I= S/N=32 S/N=116 S/N=91 S/N=73 Time Time Time Time I= I= S/N=1 S/N=66 I= I= S/N=62 S/N=37 Time Time Time Time In this method a total of 23 peptides and 69 MRMs were included as part of the analysis, although no regulations as yet state what determines a positive identification of an allergenic protein. (e.g. number of proteins and peptides to be monitored). For egg and milk, peptides representative of the different components of the egg: egg white (ovalbumin) and the yolk (vitellogenin), milk (casein), and whey (β-lactoglobulin), are included in the method. METHOD SENSITIVITY Current regulations address the analytical levels of detection for gluten, and so for the allergens monitored in this method, levels were assessed from the recommendation levels provided by VITAL and the AOAC SMPR for food allergens. Egg (white) Peanut TANELNLLILR Milk (whey) VLVLVTDYK Egg proteins 3.4 mg/kg Peanut proteins 2.5 mg/kg Milk proteins 5 mg/kg 844.4> > > > > >754.4 I= S/N=7 Time I= S/N=38 Time Time Time I= I= S/N=13 S/N= 28 I= S/N=1 Time I= S/N=25 Time I= I= I= I= S/N=1 S/N=132 S/N=46 S/N= 57 Time Time I= I= S/N=11 S/N=22 Time Time Time Time I= I= S/N= 12 S/N= 31 I= S/N= 9 I= S/N= 32 Time Time Time Time I= I= I= I= S/N=28 S/N=22 S/N=16 S/N=1 Time Time Time Time I= S/N=6 Time I= S/N=3 Time I= S/N=16 Time I= S/N=32 Time For each allergen, a single, common LOQ was determined for all targeted matrices (Figure 2). For each peptide, two MRM transitions in allergen-free matrices and incurred matrices were shown to demonstrate the specificity of the method and to confirm detection of the food allergens at the LOQ. The LOQ was defined as the minimum concentration giving a signalto-noise ratio (S/N) of 1 for the most intense MRM transition of the targeted food allergen. The sensitivity of detection for the food allergen peptides was determined on the worst case, mainly processed cookies. The LOQs recorded are:.5 mg milk proteins/kg for caseins, 5 mg milk proteins/kg for whey, 3.4 mg egg proteins/kg for egg white, 3.8 mg egg proteins/kg for egg yolk, 2.5 mg/kg for peanut proteins, and 5 mg/kg for soybean proteins. Egg (yolk) NFLINETAR Soybean EAFGVNMQIVR Egg proteins 3.8 mg/kg Soy proteins 5 mg/kg 539.3> > > >76.4 Time Time I= S/N= I= S/N= I= S/N=21 I= S/N=3 Time Time Time Time I= S/N=8 I= S/N= I= S/N=7 I= S/N=18 Time Time Time I= S/N= I= S/N= I= S/N=14 I= S/N=2 Time Time Time I= S/N= Time I= S/N=11 Time I= S/N=15 Time I= S/N=2 Figure 2. Chromatograms of the two higher MRM transitions of milk casein peptide FFVAPFPEVFGK, whey milk peptide VLVLDTDYK, and peanut peptide TANELNLLILR. Egg white peptide GGLEPINFQTAADQAR, egg yolk peptide NFLINETAR, and soy peptide EAFGVNMQIVR in chocolate, ice cream, tomato sauce, and cookies. Data of incurred or processed matrices at the limit of quantification are presented without any data treatment. Time [ 54 ] Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs

57 METHOD LINEARITY Linearity and matrix effects were tested by analyzing three independent foodstuff preparations (incurred chocolate and ice cream and processed cookies and sauce) that contained different concentrations of milk, egg, soy, and peanut food allergen proteins (Figure 3). Although the matrix effect and the effect of the thermal process were not the same for both targeted peptides from the same food allergen, the linear coefficient of regression supported the reliability of the method even the absence of an internal standard. Figure 3. Linear regression of peptide peak area of the higher MRM in function of the concentration of food allergen proteins performed in three independent replicates in incurred tomato sauce, chocolate, ice cream, and processed cookies. The linearity was controlled for each food allergen: milk casein FFVAPFPEVFGK (692.9>92.5) and YLGYLEQLLR (634.4>771.5); whey milk VLVLDTDYK (533.3>853.4) and LSFNPTQLEEQC[+57] HI (858.4>928.4) (carbamidomethylation of cysteine amino acids by addition of iodoacetamide before an enzymatic digestion to block the onset of disulfur bridges); egg white GGLEPINFQTAADQAR (844.4>666.3) and ISQAVHAAHAEINEAGR (887.5>167.5); egg yolk NFLINETAR (539.3>73.4) and EALQPIHDLADEAISR (593.3>668.8); peanut TANELNLLILR (635.4>741.5) and FNLAGNHEQEFLR (525.6>6.8; and soybean EAFGVNMQIVR (632.9>76.4) and LITAIPVNKPGR (464.6>583.4). Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs [ 55 ]

58 [ APPLICATION NOTE ] CONCLUSIONS Sensitive detection of food allergens (milk casein, whey, egg white, egg yolk, peanut, and soybean) was achieved by analyzing food allergen peptides using the ACQUITY UPLC System coupled to the Xevo TQ-S. In keeping with food production requirements, the targeted matrices were processed (tomato sauce, cookies) or incurred (chocolate, ice cream). This multi-allergen detection method has the lowest limits of quantification available to date (expressed in total proteins and not soluble proteins):.5 mg milk proteins/kg for caseins, 5 mg milk proteins/kg for whey, 3.4 mg egg proteins/kg for egg white, 3.8 mg egg proteins/kg for egg yolk, 2.5 mg peanut proteins/kg, and 5 mg soybean proteins/kg. References 1. AOAC SMPR 216.2, Standard method performance requirements for Detection and Quantitation of Selected Food Allergens M Planque, T Arnould, M Dieu, P Delahaut, P Renard, N Gillard. Advances in ultra-high performance liquid chromatography coupled to tandem mass spectrometry for sensitive detection of several food allergens in complex and processed foodstuffs. J Chrom A. 1464: , 216. While matrix effects can be observed from the data shown, further work will involve the inclusion of internal standards in order to make the method quantitative. Waters, ACQUITY UPLC, Xevo, MassLynx, and The Science of What's Possible are registered trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. February EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 56 ] Targeted and Sensitive Detection of Food Allergens in Complex and Processed Foodstuffs

59 [ APPLICATION NOTE ] Identification and Quantitative Analysis of Egg Allergen Peptides Using Data Independent Ion Mobility Mass Spectrometry Lee A Gethings, 1 Nathalie Gillard, 2 Antonietta Wallace, 1 and Valery Dumont 2 1 Waters Corporation, Wilmslow, UK; 2 CER Groupe, Marloie, Belgium APPLICATION BENEFITS HDMS E provides both qualitative and quantitative information in a single experiment. Utilizing ion mobility as part of the workflow provides enhanced specificity and therefore confidence of identifications returned. Method provides potential means for multi-allergen detection. INTRODUCTION Food allergies arise from an abnormal immunological response to certain foods. Proteins are the main candidates for triggering allergic reactions. Egg-based proteins are one of the most frequent causes of adverse reactions in food. Since many processed foods contain egg as a raw ingredient, the ability to assess changes in protein structure and detection through the manufacturing cycle is important. Food allergen analysis using LC-MS/MS is a current hot topic for many food scientists and there are two approaches that can be used to generate a quantitative method. The first approach is to perform in silico digestion of proteins, based on fasta sequences available in databases (e.g. UniProt), providing a list of potential peptides and MRM transitions. This methodology requires further investigation to determine which MRMs are the most specific and sensitive at detectable response levels for postfood processing, sample treatment, and during the ionization process. The alternative is to perform a discovery omics experiment using a high resolution instrument, such as a QTof mass spectrometer and use the data observed from this experiment to generate a targeted method. In this study, the second approach has been applied and focuses on identifying and quantifying known allergenic proteins from raw and cooked egg samples. Proteins extracted from raw and cooked egg samples were digested using trypsin and label-free protein expression data were acquired with Waters SYNAPT G2-Si using an ion mobility data independent approach (whereby the collision energy was switched between low and elevated energy states during alternate scans). WATERS SOLUTIONS SYNAPT G2-Si ACQUITY UPLC M-Class System Progenesis QI for Proteomics Utilizing ion mobility as part of the workflow provides enhanced specificity and therefore confidence of identifications returned, even in the presence of complex matrices, such as processed food samples. Precursor and product ions were associated by means of retention and drift time alignment. Although egg proteins were the focus of this work, other allergenic proteins that are also extracted using the sample preparation could be investigated, providing a potential means for multi-allergen detection. KEYWORDS Allergens, foodomics, proteomics, discovery omics, QTof, ion mobility Identification and Quantitative Analysis of Egg Allergen Peptides [ 57 ]

60 [ APPLICATION NOTE ] The acquired data were processed using Progenesis QI for Proteomics and searched against a Gallus Gallus (Uniprot) database. The results generated allowed for relative quantification to be established. The results of this study showed that a significant proportion of proteins identified were expressed when comparing cooked and raw egg sample sets, which included known allergenic proteins (e.g. apovitellenin I). Peptides identified in both sample sets allowed for MRM transitions to be generated and a quantifiable value assigned. EXPERIMENTAL Sample preparation Proteins were extracted from egg-based samples using phosphate buffer saline (PBS) and a BCA assay used to determine initial protein concentrations and afterward normalized proteins concentration to 1 mg/ml. Samples were reduced and alkylated before overnight digestion using trypsin. Prior to LC-MS analysis, samples were filtered using a.22 µm filter to remove any particulates and diluted appropriately using.1 formic acid (Figure 1). LC-MS conditions Label-free LC-MS was used for qualitative and quantitative peptide analyses. Experiments were conducted using a 9 min gradient from 5 to 4 acetonitrile (.1 formic acid) at 3 nl/min using an ACQUITY UPLC M-Class System configured with an ACQUITY UPLC Peptide BEH C18 nanoacquity Column 1K psi, 13Å, 1.7 µm, 75 µm X 15 mm, part no Data were acquired in data independent analysis (DIA) utilizing a SYNAPT G2-Si Mass Spectrometer enabled with ion mobility functionality. /*G5:2 Samples H*,(77 Raw egg ##I6(77 Cooked egg ##I&(>25&I6B Cookie (spiked) J1!1/(*I FAPAS cake Protein extraction with PBS!./ BCA.1(*22*3 assay Bioinformatics The LC-MS peptide data were processed and searched with Progenesis QI for Proteomics Software. A species specific Gallus Gallus (Uniprot) database was used. Fixed and variable modifications included carbamidomethyl C and met-oxidation respectively in addition to a protein false discovery rate of 4. 4"35$&+6&72$&#' Tryptic digest Bioinformatics 8*9:(;"(8 Label <=/(>?@(8<@A1<A=/<=/B free LC-HDMS E Qualitative / quantitative analysis Figure 1. Experimental design study for egg allergen proteins. ACQUITY UPLC M-Class System with the SYNAPT G2-Si. [ 58 ] Identification and Quantitative Analysis of Egg Allergen Peptides

61 RESULTS AND DISCUSSION PROTEIN IDENTIFICATION AND QUANTIFICATION FOR RAW AND COOKED EGG Proteins extracted from raw and cooked egg Table 1. Typical allergenic proteins identified and quantified from raw and cooked egg extracts. samples were analyzed to identify, quantify, Raw Cooked and investigate the variance between potential Allergenic protein (ng/µl) (ng/µl) allergenic markers. A total of 95 and 84 proteins P5: Ovomucoid (OVM) were identified for raw and cooked respectively. A subset of those proteins identified are P112: Ovalbumin (OVA) highlighted in Table 1 with their respective P2659: Apovitellenin (APO) amounts found in both the raw and cooked P2789: Ovotransferrin (OVT) egg extracts. P698: Lysozyme (LYS) The SYNAPT G2-Si utilized data independent analysis with ion mobility (HDMS E ) enabled. The advantage of HDMS E mode is that it maximizes the number of identified proteins through increased peak capacity and overall specificity. Example low energy (relating to precursor ions) and high energy (relating to fragment ions) are presented in Figure 2. Ion mobility was enabled to provide enhanced specificity for the experiment. This results in cleaner spectra (important when analyzing complex food matrices), and provides the ability to separate similar species, as shown in Figure 3, where example spectra are shown with ion mobility deactivated (off) and activated (on). Low collision energy (Precursor ions) Ramped collision energy (Fragment ions) Figure 2. OVM diagnostic marker identifiable at 1 ppm (m/z , ELAAVSVDCSEYPKPDCTAEDRPLCGSDNK). Identification and Quantitative Analysis of Egg Allergen Peptides [ 59 ]

62 [ APPLICATION NOTE ] IMS OFF Raw Cooked IMS ON Raw Cooked Figure 3. Effect of IMS separation demonstrated for the overlapping OVT peptide (m/z , AIANNEADAISLDGG). Upper traces show overlapping species for both raw and cooked egg samples. The implementation of IMS (lower trace) allows separation of precursors with the same m/z, resulting in the identification of two distinct species. Using the SYNAPT G2-Si in HDMS E mode made it possible to obtain high peptide sequence coverage in the presence of cake matrix (Figure 4). Figure 4. P112, ovalbumin (OVA), and gallus gallus chicken peptides identified in the presence of cake matrix (51.8 sequence coverage). [ 6 ] Identification and Quantitative Analysis of Egg Allergen Peptides

63 UTILIZING ION MOBILITY TO REDUCE BACKGROUND MATRIX EFFECTS To assess the capabilities and advantages of implementing ion mobility into the analytical workflow, a dilution series of cooked egg ranging from 5 to 1 ppm was spiked into cookie matrix, which was maintained at the same concentration throughout (Figure 5). 7 OVA 16 OVM 25 coverage Intensity PPM PPM coverage APO PPM Intensity 1: TOF MS ES+ BPI 2.82e Time Figure 5. Identified allergenic proteins for serially diluted cooked egg in the presence of cookie matrix. Number of peptides (black), sum of intensity for the top three most abundant peptides (grey), and sequence coverage (blue). An example chromatogram is provided for 5 ppm of cooked egg in the presence of the cookie matrix. FAPAS TEST SAMPLE CAKE MIX Test material was supplied for FAPAS Proficiency Test material T277 in the form of a cake mix obtained from a retail source which was free from egg and milk but contained gluten. Royal icing sugar was used to introduce egg white protein. The major allergen identified was ovalbumin (OVA), and the sequence coverage (percentage of peptides identified that make up the protein sequence) in the presence of cake matrix was high at Initial results from the FAPAS proficiency study were conducted using ELISA, which quantified egg between 39.6 to 62.1 ppm (mean = 47.7 ppm). The LC-MS label-free experiments corresponded with the ELISA findings, quantifying at 58 ppm. The nature of HDMS E also allows for multi-allergens to be detected and quantified as part of this experiment. Identification and Quantitative Analysis of Egg Allergen Peptides [ 61 ]

64 [ APPLICATION NOTE ] CONCLUSIONS A discovery proteomic workflow has been applied to determine marker peptides that can be used for the quantitative analysis of allergenic proteins within food. A label-free proteomic approach has been applied for the analysis of egg-based allergens, by implementing HDMS E to provide both qualitative and quantitative information in a single experiment. Ion mobility as part of the workflow is shown to provide enhanced specificity and therefore confidence of identifications returned, even in the presence of complex matrices, such as processed food samples. References 1. Commission Directive No 27/68/EC of 27 November 27 amending Annex IIIa to Directive 2/13/EC (Official Journal, L31, 28/11/27, 1 14). 2. Li GZ, Vissers JP, Silva JC, Golick D, Gorenstein MV, Geromanos SJ. Database searching and accounting of multiplexed precursor and product ion spectra from the data independent analysis of simple and complex peptide mixtures. Proteomics. Mar;9(6): , 29. Although only egg proteins were the focus of this work, other proteins relating to other allergens were observed, providing a potential means for multi-allergen detection. Waters, SYNAPT, ACQUITY UPLC, Progenesis, and The Science of What's Possible are registered trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. February EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 62 ] Identification and Quantitative Analysis of Egg Allergen Peptides

65 [ VETERINARY DRUGS ] [ 63 ]

66 [ VETERINARY DRUGS ] [ 64 ]

67 [ APPLICATION NOTE ] Determination of Beta Agonists in Animal Tissues and Urine Using Liquid Chromatography-Tandem Quadrupole Mass Spectrometry Simon Hird 1 and James Jowett 2 1 Waters Corporation, Wilmslow, UK; 2 Fera Science Ltd, York, UK APPLICATION BENEFITS A specific, targeted method for β-agonists determination in a range of sample types that meets requirements for both official control and food business operators due diligence testing. INTRODUCTION Beta-adrenergic receptor agonists (β-agonists) are synthetic compounds that mimic some of the effects of naturally-occurring compounds by binding to beta-receptors on the surface of cells within the muscle, fat, and other tissues of animals. Some β-agonists are used in human medicine for the treatment of conditions such as asthma, but they have also been used in livestock production to enhance growth and alter body composition. 1 The structures of some of the key β-agonists are shown in Figure 1. Salbutamol Zilpaterol Clenbuterol Mabuterol HO H 3 C CH 3 HN CH 3 HO OH HO N NH O NH CH 3 CH 3 Cl H 2 N OH NH CH 3 CH 3 CH 3 Cl Cimaterol H 2 N Cl F F F HO Turbutaline NH CH 3 CH3 CH 3 H 2 N CH 3 OH Salmeterol N OH NH HO CH 3 NH CH 3 CH3 HO O Isoxsuprine HO CH 3 HO NH Clenproperol Cimbuterol HO Ractopamine OH OH Cl H H N Cl HO O HN CH 3 CH 3 HO H 3 C CH 3 H N CH 3 NH OH WATERS SOLUTIONS ACQUITY UPLC I-Class System HO CH 3 Figure 1. Structures of some of the key β-agonists. H 3 C CH 3 N H OH N H N H Xevo TQ-S micro CORTECS UPLC C18+ Column MassLynx MS Software TargetLynx Application Manager KEYWORDS UPLC-MS/MS, Beta agonists, targeted analysis, ractopamine, clenbuterol Although the administration of β-agonists as growth-promoting agents in food-producing animals is banned in many countries due to concerns over human health, there are exceptions. 2 Ractopamine and zilpaterol are authorized for the production of some animals (e.g. cattle and pigs) in a limited number of countries such as the U.S., Canada, and Brazil. The Joint FAO/WHO Expert Committee on Food Additives (JECFA) has established Maximum Residue Limits (MRLs) for ractopamine in cattle and pig muscle (both 1 µg/kg), which have been adopted by the Codex Alimentarius Commission (Codex) and implemented in some countries (e.g. Canada). Determination of Beta Agonists in Animal Tissues and Urine [ 65 ]

68 [ APPLICATION NOTE ] Others have alternative limits for ractopamine, such as the tolerances set in the U.S. (3 and 5 µg/kg in cattle and pig muscle, respectively). However, many countries disagree with the Codex standards and are restricting or banning meat products containing β-agonists. In the EU, MRLs only exist for clenbuterol (cattle and horse). Other β-agonists are prohibited for use in food producing animals. As no Minimum Required Performance Limits (MRPLs) were set for these banned β-agonists, 3 the EU relies on Recommended Concentrations (RCs) to indicate required performance of associated analytical methods. In contrast to MRPLs, RCs have no legal standing and are used for guidance only. Importing countries can set even lower limits for testing based upon trading decisions to provide better warranties to their customers and to gain commercial advantage. For example, Russia established an action level of.1 µg/kg for ractopamine in imported beef consignments; one hundred times lower than Codex MRL. However, as Russia has a zero tolerance policy, if a laboratory found any confirmed residues at lower concentrations, the consignment would be rejected. This drives down the analytical limits for those involved with pre-export testing. Most countries have established strict surveillance programs for official control purposes to check compliance with regulatory limits, for both domestic and imported produce. Monitoring compliance within these limits requires the use of highly sensitive and selective analytical methodology based on liquid chromatography-tandem mass spectrometry (LC-MS/MS). As the frequency of detection of residues is typically low, samples tend to be screened using a reduced but adequate level of quality control. This includes the use of stable isotope analogs as surrogates to monitor and mitigate the impact of any sample-to-sample variability in matrix effects. Any suspect positives are re-analyzed with the same method but with additional analytical quality control suitable for quantification (e.g. multi-level matrix-extracted calibration). This method had also been validated as suitable for confirmation 4 with values for Decision Limit (CCα) as low as reasonably achievable. EXPERIMENTAL Sample preparation Extracts were generated by Fera using a validated method based on methods developed by the EURL BVL in Germany. 5,6,7 Samples of meat and fish were extracted by blending with an aqueous buffer/methanol solution and subjected to enzyme hydrolysis to cleave any conjugated β-agonists, and to help solubilize any residues. Solid-phase extraction (SPE) was carried out on a mixed mode-type cartridge prior to determination using UPLC-MS/MS. Samples of urine were extracted by SPE after enzyme hydrolysis. Stable isotope analogues were available for many of the analytes, and were added to all samples prior to extraction. Matrix-extracted standards were prepared at the Screening Target Concentration (STC) when screening and at multiple levels, typically over the range.5x to 1x STC, for any confirmatory analysis. Typically, STC values are set at half the maximum limit, which for β-agonists varies (see Table 1). For example, the STC for clenbuterol in animal muscle was set at.5 µg/kg, whereas that for terbutaline was set at 5. µg/kg. Sample extracts were filtered prior to subsequent analysis by UPLC-MS/MS. [ 66 ] Determination of Beta Agonists in Animal Tissues and Urine

69 UPLC conditions UPLC system: ACQUITY UPLC I-Class with FTN Autosampler Column: CORTECS UPLC C µm, 2.1 mm Mobile phase A:.1 formic acid (aq.) Mobile phase B:.1 formic acid in acetonitrile Flow rate:.6 ml/min Injection volume: 5 µl MS conditions MS system: Xevo TQ-S micro Ionization mode: ESI + Capillary voltage: 1. kv Desolvation temp.: 6 C Desolvation gas flow: L/Hr Source temp.: 14 C Cone gas flow: L/Hr Column temp.: 4 C Sample temp.: 1 C Run time: Time (min) A B Curve min The two MRM transitions that showed the best selectivity were used for each of the β-agonists and one for each stable isotope analogue. The data were acquired using MassLynx 4.1 Software and processed using TargetLynx XS Application Manager. Table 2 summarizes conditions for all MRM transitions and retention times. Cone voltage was set to 4 V throughout, with the exception of the in-source fragment at m/z 22.1, which was selected as a precursor ion for cimaterol (75 V). The optimum dwell time was set automatically using the Autodwell function based upon 4s wide peaks and 12 data points per peak. Table 1. STC values (µg/kg) for β-agonists in difficult sample types. Compound Animal muscle Animal offal Poultry offal Urine Terbutaline (TER) Cimaterol (CIM) Salbutamol (SAL) Zilpaterol (ZIL) Cimbuterol (CIMB) Hydroxymethyl Clenbuterol (HMC) Clenproperol (CLE) Ractopamine (RAC) Clenbuterol (CLEN) Tulobuterol (TUL) Brombuterol (BRO) Isoxsuprine (ISO) Mabuterol (MAB) Clenpenterol (CLENP) Mapenterol (MAP) Salmeterol (SALM) Determination of Beta Agonists in Animal Tissues and Urine [ 67 ]

70 [ APPLICATION NOTE ] Table 2. Retention times and MRM parameters for β-agonists and stable isotope analogues (quantatitive transitions in bold). Compound Retention time (min) Terbutaline 1.57 MRM 226.1> >125. CE (ev) Terbutaline-d > Cimaterol > >143. Cimaterol-d > Salbutamol > >148. Salbutamol-d > Zilpaterol > >22.1 Zilpaterol-d > Cimbuterol > >143. Cimbuterol-d > Hydroxymethyl Clenbuterol 2.15 Clenproperol > > > >132. Clenproperol-d > Ractopamine > >164. Ractopamine-d > Clenbuterol > >25.1 Clenbuterol-d > Tulobuterol > >172. Tulobuterol-d > Brombuterol 2.72 Isoxsuprine 2.81 Mabuterol > > > > > >237.1 Mabuterol-d > Clenpenterol 2.83 Mapenterol > > > >237.1 Mapenterol-d > Salmeterol > >38.3 Salmeterol-d > [ 68 ] Determination of Beta Agonists in Animal Tissues and Urine

71 RESULTS AND DISCUSSION The objective of this work was to evaluate the performance of the Xevo TQ-S micro instrument for the determination of 16 β-agonists in appropriate matrices rather than development and validation of a new method. Analysis was restricted to batches of various animal target tissues (e.g. liver), edible products (e.g. muscle), and urine collected from animals on the farm. Performance was evaluated based on its sensitivity for analytes in context with international regulations, the absence of isobaric interference, precision of UPLC-MS/MS measurements, accuracy and precision through the analysis of spiked samples and proficiency test materials, as well as compliance with identification and quantification criteria. The LC conditions were modified from those reported in an earlier Waters application note. 8 The CORTECS C18+ Column is a charged surface, high efficiency C18 column, based on a solid-core particle that delivers excellent peak shape for these basic analytes at low ph, while providing sufficient retention of the most polar β-agonists, both at a lower operating pressure than columns packed with fully porous sub-2 µm particles. The gradient was adjusted to provide sufficient chromatographic resolution of all 16 β-agonists from isobaric interference on both transitions. The selection of MRM transitions and optimization of critical parameters was performed by infusion of individual solutions of all the analytes and evaluation of the data by IntelliStart Software to automatically create acquisition and processing methods. Excellent sensitivity and selectivity was demonstrated by the response for each analyte peaks detected from the analysis of matrix-extracted standards prepared at 1 x STC in a range of different sample types: bovine and poultry liver, salmon, and bovine urine. See Figure 2 and Figure 3 for examples. ZIL m/z 262>185 RAC m/z 32>164 SAL m/z 24>148 CLE m/z 365>132 CIM m/z 22>13 HMC m/z 293>23 TER m/z 226>17 CIMB m/z 243>16 ISO m/z 32>17 SALM m/z 416>38 BRO m/z 367>293 MAP m/z 325>237 TUL m/z 228>154 MAB m/z 311>237 CLEN m/z 277>23 CLENP m/z 291>23 Figure 2. Chromatograms showing β-agonists from analysis of matrix-extracted standards of bovine liver at the STC. Determination of Beta Agonists in Animal Tissues and Urine [ 69 ]

72 [ APPLICATION NOTE ] ZIL m/z 262>185 RAC m/z 32>164 SAL m/z 24>148 CLE m/z 365>132 CIM m/z 22>13 HMC m/z 293>23 TER m/z 226>17 CIMB m/z 243>16 ISO m/z 32>17 SALM m/z 416>38 BRO m/z 367>293 MAP m/z 325>237 TUL m/z 228>154 MAB m/z 311>237 CLEN m/z 277>23 CLENP m/z 291>23 Figure 3. Chromatograms showing β-agonists from analysis of matrix-extracted standards of bovine urine at the STC. Figure 4 shows the detection of ractopamine spiked into pork meat at.5 µg/kg. This demonstrates the suitability of the method for checking compliance with the Russian action level for imported meat of.1 µg/kg ractopamine. No interfering compounds were detected at the retention times of the β-agonists or stable isotope analogues in all of the tested blank samples, with the exception of the second transition for clenproperol in liver samples. RAC m/z 32>164 RAC m/z 32>164 RAC m/z 32>17 RAC m/z 32>17 RAC- d6 m/z 38>168 RAC- d6 m/z 38>168 Figure 4. Chromatograms showing ractopamine and stable isotope analogue from analysis of matrix-extracted standard of pork meat spiked at.5 µg/kg and associated blank. Carryover was observed to be typically very low, with the exception of the late eluting salmeterol. A maximum of 1.3 carryover was observed for salmeterol from measurements in solvent blanks after injection of standards, but this proved to be a consistent background in the system equivalent to ca. 1 of the STC level. [ 7 ] Determination of Beta Agonists in Animal Tissues and Urine

73 Measurement precision, as determined from the peak areas of replicate (n=25) injections of β-agonists matrix extracted standards (1x STC), was good for the majority of compounds (<5 RSD). Precision for salmeterol and brombuterol was worse in some matrices, having lower area counts due to lower recovery. When concentration was estimated using appropriate stable isotope analogues as internal standards, precision of the measurements was observed to improve for all matrices tested (RSD <4). Retention times of the analytes from these replicate injections were very stable showing a maximum shift of ±.1 minutes. Ion ratios were found to be within tolerance with the exception of clenproperol in liver samples, where there was an isobaric interference on the second transition, so an alternative MRM transition would be needed for confirmatory analyses. Table 3. Precision data (RSDs) for measurement of β-agonists in the different matrices. Compound Bovine liver Porcine liver Salmon Bovine urine Terbutaline (TER) Cimaterol (CIM) Salbutamol (SAL) Zilpaterol (ZIL) Cimbuterol (CIMB) Hydroxymethyl clenbuterol (HMC) Clenproperol (CLE) Ractopamine (RAC) Clenbuterol (CLEN) Tulobuterol (TUL) Brombuterol (BRO) Isoxsuprine (ISO) Mabuterol (MAB) Clenpenterol (CLENP) Mapenterol (MAP) Salmeterol (SALM) To assess performance of the method for confirmatory purposes, accuracy, precision, and identification criteria were evaluated through the analysis of incurred proficiency test materials and spiked blank samples. The linearity of response for two of the β-agonists in bovine liver matrix was evaluated using a bracketed calibration over a suitable concentration range of:.25 to 5. µg/kg for ractopamine, and.5 to 1. µg/kg for clenbuterol. The coefficients of determination were satisfactory (r 2 >.99), and the residuals <4. Determination of Beta Agonists in Animal Tissues and Urine [ 71 ]

74 [ APPLICATION NOTE ] Clenbuterol Ractopamine Compound name: Clenbuterol Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: Internal Std (Ref 4), Area * (IS Conc./IS Area) Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Compound name: Ractopamine Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: Internal Std (Ref 2), Area * (IS Conc./IS Area) Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Residual ug/kg Residual ug/kg Response Response ug/kg ug/kg Figure 5. Calibration graphs for clenbuterol and ractapamine prepared in bovine liver matrix. Three bovine liver proficiency test materials (BVL, Germany), contaminated with β-agonists, were prepared in duplicate and the extracts were analyzed by UPLC-MS/MS in triplicate (n=6). Sample A was found to contain both ractopamine (1.1 µg/kg, 2.6 RSD) and clenbuterol (.27 µg/kg, 5.8 RSD); Sample B just ractopamine (3.69 µg/kg, 3.7 RSD); and Sample C contained clenbuterol only (.18 µg/kg, 3.4 RSD). The mean measured concentrations of ractopamine and clenbuterol compared well with the assigned values in all three samples. In addition two samples of bovine liver, previously shown to be blank, were spiked with ractopamine (1.25 µg/kg) and clenbuterol (.2 µg/kg) and analyzed in triplicate. Mean recovery and precision was (1.9 RSD) and 11 (2.1 RSD), respectively, after correction for losses using the appropriate stable isotope analogue as an internal standard. Requirements given in EU Commission Decision 22/65/7/EC for identification were met (European Commission, 22). The range of calculated ion ratios for ractopamine (.95 to 1.) and clenbuterol (.62 to.64), from the replicate measurements of the incurred samples, were well within the tolerance of ±2 of the reference values for ractopamine (.8 to 1.21) and clenbuterol (.51 to.71), respectively. The retention times for ractopamine (2.3 to 2.31 min) and clenbuterol (2.49 min) showed little variation and were well within acceptance tolerance of ±2.5 (2.25 to 2.37 min and 2.43 to 2.55 min). [ 72 ] Determination of Beta Agonists in Animal Tissues and Urine

75 CONCLUSIONS The ACQUITY UPLC I-Class System combined with the Xevo TQ-S micro provides excellent sensitivity for the detection, identification, and quantification of β-agonists in a range of products. This method can be used for both screening and confirmation for official control purposes with a high degree of confidence, but also to meet the challenging requirements of pre-export testing, which often demands lower limits of quantification. Scientists must validate the method in their own laboratories and demonstrate that their calculated Detection Capability (CCβ) and Decision Limit (CCα) values are fit for purpose (i.e. as low as reasonably achievable). Acknowledgements The authors gratefully acknowledge Fera Science Ltd. for providing the extracts for the β-agonists analysis. References 1. Centner TJ, Alvey JC, Stelzleni AM. Beta agonists in livestock feed: status, health concerns, and international trade. J Animal Sci (9): Nino AMM, Granja RH, Wanschel AC, Salerno AG. The challenges of ractopamine use in meat production for export to European Union and Russia. Food Control : European Commission. Commission Decision 23/181/EC of 13 March 23 amending Decision 22/657/EC as regards the setting of minimum required performance limits (MRPLs) for certain residues in food of animal origin. Official Journal of the European Communities. 23 L71: European Commission. Commission Decision 22/657/EC of 12 August 22 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Official Journal of the European Communities. 22 L221: Radeck W and Gowik P. Validation of a multiresidue method for confirmation and quantification of ß agonists in liver in accordance with CD 22/657EC. First International Symposium on Recent Advances in Food Analysis, Prague, Radeck W and Gowik P. Validation of a multiresidue method for confirmation and quantification of 22 beta-agonists in urine. Fifth International Symposium on Hormone and Veterinary Drug Residue Analysis, Antwerp, Radeck W. Confirmatory method for the determination of 3 ß-agonists in muscle using LC-MS/MS. 28, unpublished results. 8. Ren Y and Yang J. Determination of Common Beta-Agonist Residues in Meat Products by UPLC-MS/MS. Waters Application Note No en Waters, ACQUITY, UPLC, Xevo, MassLynx, CORTECS, and The Science of What's Possible are registered trademarks of Waters Corporation. TargetLynx and IntelliStart are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. June EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: Determination of Beta Agonists in Animal Tissues and Urine [ 73 ]

76 [ APPLICATION NOTE ] Determination of the Metabolites of Nitrofuran Antibiotics in Animal Tissues and Food Products by UPLC-MS/MS Simon Hird 1 and Richard Ginn 2 1 Waters Corporation, Wilmslow, UK; 2 Fera Science Ltd, York, UK APPLICATION BENEFITS Specific, targeted method for nitrofurans determination in a range of sample types that meet requirements for both official control and food business operators pre-export testing. INTRODUCTION Nitrofurans (NFs) are a group of broad spectrum antibiotics. Due to health concerns, nitrofurans are now prohibited for use in food-producing animals in most jurisdictions. They are still authorized for human medicine and for the treatment of non-food animals. Nitrofurans are widely manufactured, sold, and hence available for misuse. 1 There have been frequent findings of nitrofuran residues in honey, poultry, and aquaculture products imported to EU countries, which has led to product recalls, border rejections, and de-listed suppliers. Violations have resulted in implementation of emergency measures that have necessitated mandatory pre-export testing (PET), widespread voluntary pre-harvest tests (PHT), and an increase both in the analysis of imports at border control within the EU and in the frequency of European Commission Food and Veterinary Office (FVO) visits. 2 The EU Minimum Required Performance Limit (MRPL) in poultry meat and aquaculture is 1 µg/kg for each of the four nitrofurans, measured as their WATERS SOLUTIONS ACQUITY UPLC H-Class System Xevo TQ-S micro MS System ACQUITY UPLC BEH C18 Column MassLynx MS Software TargetLynx Application Manager KEYWORDS Nitrofurans, veterinary drugs, antibiotics, furazolidone, 3-amino-2-oxazolidinone (AOZ), nitrofurazone, semicarbazide (SCA), furaltadone, 3-amino-5-morpholinomethyl- 2-oxazolidinone (AMOZ), nitrofurantoin, 1-aminohydantoin (AHD) respective tissue-bound metabolites (EU Commission Decision 23). 3 MRPLs are the minimum content of an analyte in a sample, which at least has to be detected and confirmed. They are also the reference point for action (Action Levels) when evaluating food consignments. Laboratories must demonstrate that their calculated Detection Capability (CCβ) and Decision Limit (CCα) values are at or below the MRPL. 4 Although enforcement action is only taken when a residue exceeds the MRPL, noncompliant samples below the MRPL must still be monitored. Suppliers and importers can set even lower limits for PET based upon trading decisions to provide better warranties to their customers and to gain commercial advantage. Meeting these requirements requires the continued development of highly sensitive and specific analytical methodology based upon UPLC -MS/MS. Previous studies have demonstrated that parent nitrofurans deplete rapidly in animals and that they are extensively metabolized to tissue-bound metabolites. 5 Methods have been described for various animal tissues such as kidney for official control, muscle, honey, shrimp, eggs, and milk for assessing risk to the consumer. Parent nitrofurans are only sought in medicated feeds used for animal production and aquaculture. [ 74 ] Determination of the Metabolites of Nitrofuran Antibiotics in Animal Tissues and Food Products

77 Commonly sought parent nitrofurans and associated metabolites include: furazolidone as 3-amino-2-oxazolidinone (AOZ), nitrofurazone as semicarbazide (SCA), furaltadone as 3-amino-5-morpholinomethyl-2-oxazolidinone (AMOZ) and nitrofurantoin as 1-aminohydantoin (AHD). Derivatization improves extraction efficiency from the acidic aqueous phase into the organic solvent, imparts functionality amenable to reversed-phase chromatography for both SPE and UPLC, and improves electrospray ionization efficiency. EXPERIMENTAL Sample preparation Extracts were generated by Fera Science, Ltd. using a method that was originally developed as part of the FoodBRAND Project 5 but subsequently modified for routine surveillance. Although newer technology has enabled improvements in performance, the principles behind procedure have changed little. Tails, shells, and coverings were removed from shrimp samples as SCA has been found to be present naturally in the shells. Breadcrumbs were removed from any breaded scampi, fish and chicken products as SCA is formed during the baking of breaded crust. Any ice glaze was excluded as it is no longer considered part of the food. The sample was washed to remove unbound sources of SCA. Acid hydrolysis was used to cleave the metabolites bound to proteins or other tissues and to hydrolyse the side-chain of any intact parent drug present, as well as to solubilize any residues. After adjusting the ph, solid-phase extraction (SPE) was carried out on a C18-type cartridge prior to derivatization with 2-nitrobenzaldehyde. For the analysis of the kidney samples, extracts were diluted 5-fold. Stable isotope analogues for each of the analogues were added to all samples. Matrix-extracted calibrants were prepared at a Screening Target Concentration (STC) of.5 µg/kg, when screening, and at multiple levels, typically over the range.25 to 5. µg/ kg, for any confirmatory analysis. Sample extracts were filtered prior to subsequent analysis by UPLC-MS/MS. UPLC conditions UPLC system: ACQUITY UPLC H-Class with FTN autosampler Column: ACQUITY UPLC BEH C µm, 2.1 mm Mobile phase A:.5 mm ammonium formate (aq.) Mobile phase B: methanol Flow rate:.45 ml/min Injection volume: 5 µl Column temp.: 45 C Sample temp.: 15 C MS conditions MS system: Xevo TQ-S micro Source: Electrospray Ionization mode: ESI+ Capillary voltage:.5 kv Desolvation temp.: 65 C Desolvation gas flow: L/hr Source temp.: 15 C Cone gas flow: 15 L/hr Collision gas flow:.15 ml/min Run time: 11 min Nebulizer gas pressure: 7 Bar Time (min) A B Curve Two MRM transitions were used for each of the nitrofuran metabolites and one for each stable isotope analogue, all determined as the nitrophenyl (NP-) derivatives. Data were acquired using MassLynx MS Software and processed using TargetLynx XS Application Manager. Table 1 summarizes conditions for all MRM transitions including the retention times. The optimum dwell time was set automatically using the autodwell function based upon 4 s wide peaks and 12 data points per peak. Determination of the Metabolites of Nitrofuran Antibiotics in Animal Tissues and Food Products [ 75 ]

78 [ APPLICATION NOTE ] Table 1. MRM parameters for nitrofuran metabolites and stable isotope analogues. Compound Retention time NP-AHD 2.82 MRM 249.1> >14. Cone (V) NP-AHD- 13 C > NP-AOZ > >14. NP-AOZ-d > NP-SCA 3.3 NP-SCA- 13 C 15 N > > > >291.1 NP-AMOZ 335.1> NP-AMOZ-d > CE (ev) RESULTS AND DISCUSSION The objective of this work was to evaluate the performance of the Xevo TQ-S micro MS system for nitrofurans analysis rather than development and validation of a new method. Analysis was restricted to batches of different sample types relevant to routine testing. Performance was evaluated based upon sensitivity for the analytes, absence of isobaric interference, precision of the UPLC-MS/MS measurements, compliance with identification and typical quantification criteria, and accuracy and precision through analysis of proficiency test materials. The UPLC conditions were taken from an earlier Waters application note no en, 6 but the gradient was adjusted to provide sufficient chromatographic resolution of NP-AHD from isobaric interference on both transitions. The selection of MRM transitions and optimization of critical parameters was performed by infusion of individual solutions of all the analytes and evaluation of the data by IntelliStart Software to automatically create acquisition and processing methods. Excellent sensitivity and selectivity was demonstrated by the response for each analyte peak detected from the analysis of matrix-extracted calibrants prepared at the STC of.5 µg/kg in a range of different sample types: prawn, fish, poultry muscle, bovine kidney, egg, and honey, as shown in Figure 1. No interfering compounds were detected at the retention times of the analytes in all the tested blank samples. Precision, as measured from the peak areas of replicate (n=8) injections of NP-nitrofuran metabolite standards in solvent (1 ng/ml), varied from 1.7 to 3.5 RSD. Overall repeatability, as determined from the peak areas of the stable isotope analogues added to each of the replicate (n=2) extractions of incurred honey (1 µg/kg), varied from 2.8 to 13 RSD. Linearity in matrix (honey) was evaluated over the concentration range.2 to.5 µg/kg for each of the analytes. The coefficients of determination were satisfactory (r 2 >.99), and the residuals <15. 1 portions of a FAPAS honey proficiency test material contaminated with AMOZ only were prepared and analyzed in duplicate. The mean measured concentration of AMOZ (1.41 µg/kg) compared well with the assigned value (1.5 µg/kg). The precision for the calculated AMOZ concentration in the replicates of the FAPAS test material was 2.9 RSD. Requirements stipulated in EU Commission Decision 22/65/7/EC for identification were met. The range of calculated AMOZ ion ratios (.81 to.84) was well within tolerance of ±2 (.66 to.99), and the AMOZ retention time from the replicates of the incurred sample showed little variation (.1 RSD; 3.88 to 3.89 min), well within the acceptance tolerance of ±2.5 (3.79 to 3.98 min). [ 76 ] Determination of the Metabolites of Nitrofuran Antibiotics in Animal Tissues and Food Products

79 A. Prawn 31_3_215_Data117 2-NP-AHD 31_3_215_Data117 2-NP-AMOZ B. Egg 31_3_215_Data133 2-NP-AHD 31_3_215_Data133 2-NP-AMOZ 2.92 min min 2.9 min min 31_3_215_Data117 2-NP-AHD 31_3_215_Data117 2-NP-AMOZ 31_3_215_Data133 2-NP-AHD 31_3_215_Data133 2-NP-AMOZ min min min min 31_3_215_Data117 2-NP-AHD-13C3 31_3_215_Data117 2-NP-AMOZ-D5 31_3_215_Data133 2-NP-AHD-13C3 31_3_215_Data133 2-NP-AMOZ-D min min min min 31_3_215_Data117 2-NP-AOZ 31_3_215_Data117 2-NP-SCA 31_3_215_Data133 2-NP-AOZ 31_3_215_Data133 2-NP-SCA min min min min 31_3_215_Data117 2-NP-AOZ 31_3_215_Data117 2-NP-SCA 31_3_215_Data133 2-NP-AOZ 31_3_215_Data133 2-NP-SCA 2.72 min min min min 31_3_215_Data117 2-NP-AOZ-D4 31_3_215_Data117 2-NP-SCA-13C115N2 31_3_215_Data133 2-NP-AOZ-D4 31_3_215_Data133 2-NP-SCA-13C115N2 min min min min C. Chicken muscle 31_3_215_Data72 2-NP-AHD 31_3_215_Data72 2-NP-AMOZ D. Honey 31_3_215_Data85 2-NP-AHD 31_3_215_Data85 2-NP-AMOZ min min min min 31_3_215_Data72 2-NP-AHD 31_3_215_Data72 2-NP-AMOZ 31_3_215_Data85 2-NP-AHD 31_3_215_Data85 2-NP-AMOZ 2.7 min min min min 31_3_215_Data72 2-NP-AHD-13C3 31_3_215_Data72 2-NP-AMOZ-D5 31_3_215_Data85 2-NP-AHD-13C3 31_3_215_Data85 2-NP-AMOZ-D min min min min 31_3_215_Data72 2-NP-AOZ 31_3_215_Data72 2-NP-SCA 31_3_215_Data85 2-NP-AOZ 31_3_215_Data85 2-NP-SCA 31_3_215_Data72 2-NP-AOZ min 31_3_215_Data72 2-NP-SCA min 31_3_215_Data85 2-NP-AOZ min 31_3_215_Data85 2-NP-SCA min _3_215_Data72 2-NP-AOZ-D4 min 31_3_215_Data72 2-NP-SCA-13C115N2 min 31_3_215_Data85 2-NP-AOZ-D4 min 31_3_215_Data85 2-NP-SCA-13C115N2 min min min min min E. Fish 31_3_215_Data59 2-NP-AHD 31_3_215_Data59 2-NP-AMOZ F. Bovine kidney 31_3_215_Data51 2-NP-AHD 31_3_215_Data51 2-NP-AMOZ min min min min 31_3_215_Data59 2-NP-AHD 31_3_215_Data59 2-NP-AMOZ 31_3_215_Data51 2-NP-AHD 31_3_215_Data51 2-NP-AMOZ 2.72 min min min 4.3 min 31_3_215_Data59 2-NP-AHD-13C3 31_3_215_Data59 2-NP-AMOZ-D5 31_3_215_Data51 2-NP-AHD-13C3 31_3_215_Data51 2-NP-AMOZ-D min min min min 31_3_215_Data59 2-NP-AOZ 31_3_215_Data59 2-NP-SCA 31_3_215_Data51 2-NP-AOZ 31_3_215_Data51 2-NP-SCA 3.1 min min min min 31_3_215_Data59 2-NP-AOZ 31_3_215_Data59 2-NP-SCA 31_3_215_Data51 2-NP-AOZ 31_3_215_Data51 2-NP-SCA 2.71 min min min min 31_3_215_Data59 2-NP-AOZ-D4 31_3_215_Data59 2-NP-SCA-13C115N2 31_3_215_Data51 2-NP-AOZ-D4 31_3_215_Data51 2-NP-SCA-13C115N2 min min min min Figure 1. Chromatograms showing nitrofuran metabolites and stable isotope analogues, as NP-derivatives, from analysis of matrix-extracted calibrants of: A. Prawn., B. Egg, C. Chicken muscle, D. Honey, E. Fish, and F. Bovine kidney at the STC (.5 µg/kg). Determination of the Metabolites of Nitrofuran Antibiotics in Animal Tissues and Food Products [ 77 ]

80 [ APPLICATION NOTE ] Compound name: 2-NP-AHD Compound name: 2-NP-AMOZ Correlation coefficient: r =.9989, r 2 = Correlation coefficient: r = , r 2 = Calibration curve: * x Calibration curve: * x Response type: Internal Std (Ref 2), Area * (IS Conc./IS Area) Response type: Internal Std (Ref 8), Area * (IS Conc./IS Area) Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None g/kg g/kg Compound name: 2-NP-AOZ Compound name: 2-NP-SCA Correlation coefficient: r =.99962, r 2 = Correlation coefficient: r =.99939, r 2 = Calibration curve: * x Calibration curve: * x Response type: Internal Std (Ref 4), Area * (IS Conc./IS Area) Response type: Internal Std (Ref 6), Area * (IS Conc./IS Area) Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None g/kg g/kg Residual Response Residual Response Figure 2. Calibration graphs for the nitrofuran metabolites prepared in honey matrix. CONCLUSIONS The ACQUITY UPLC H-Class System coupled with the Xevo TQ-S micro MS System provides sufficient sensitivity for detection, identification, and quantification of nitrofuran metabolites in a range of products. This method facilitates routine screening and confirmation for official control purposes with a high degree of confidence, but also meets the requirements of pre-export testing, which often demands lower limits of quantification. Residual Response Residual Response g/kg g/kg g/kg g/kg References 1. Vass M, Hruska K, Franek M. Nitrofuran antibiotics: A review on the application, prohibition and residual analysis. Veterinarni Medicina (9): Points J, Thorburn Burns D, Walker MJ. Forensic issues in the analysis of trace nitrofuran veterinary residues in food of animal origin. Food Control : EU Commission Decision 22/657/EC of 12 August 22 implementing EU Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Official Journal of the European Communities. 22 L221: European Commission Decision 23/181/EC of 13 March 23 amending EU Commission Decision 22/657/EC in regards to the setting of minimum required performance limits (MRPLs) for certain residues in food of animal origin. Official Journal of the European Communities. 23 L71: Cooper KM, Mulder PPJ, van Rhijn JA, Kovacsics L, McCracken RJ, Young PB, et al. Depletion of four nitrofuran antibiotics and their tissue-bound metabolites in porcine tissues and determination using LC-MS/MS and HPLC-UV. Food Additives & Contaminants. 25 Part A 25: Morphet J, Hancock P. Application of ACQUITY TQD for the Analysis of Nitrofuran Veterinary Drug Residues in Shrimp. Waters Application Note no en. September, 27. Acknowledgements The authors gratefully acknowledge Fera Science Ltd for provision of the extracts for nitrofuran analysis. Waters, ACQUITY UPLC, Xevo, UPLC, MassLynx, and The Science of What's Possible are registered trademarks of Waters Corporation. TargetLynx and IntelliStart are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 216 Waters Corporation. Produced in the U.S.A. December EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 78 ] Determination of the Metabolites of Nitrofuran Antibiotics in Animal Tissues and Food Products

81 [ APPLICATION NOTE ] Rapid, Simple, and Effective Clean-up of Bovine Liver Samples Prior to UPLC-MS/MS Multiresidue Veterinary Drugs Analysis Michael S. Young and Kim Van Tran Waters Corporation, Milford, MA, USA APPLICATION BENEFITS Efficient, timesaving multiclass/ multiresidue methodology Simple, rapid, and effective sample clean-up suitable for a diverse range of analytes Fast, sensitive UPLC-MS/MS analysis OVERVIEW In order to ensure public health and safety, reliable analytical methods are necessary to determine veterinary drug residue levels in edible tissue samples such as beef liver. The compounds of interest range from highly polar water-soluble compounds to very non-polar fat-soluble compounds. In order to maximize throughput and minimize costs it is desirable to determine the widest possible range of veterinary drug residues in tissue samples with a single analytical method. WATERS SOLUTIONS ACQUITY UPLC I-Class System Xevo TQ-XS Mass Spectrometer Oasis PRiME HLB Cartridge for SPE Clean-up INTRODUCTION Tissue samples, such as bovine muscle and liver, are typically extracted with an acetonitrile based solvent for LC-MS determination of veterinary drug residues. Among the most significant co-extracted substances are fats and polar lipids, particularly phospholipids (lecithin). A gram of bovine liver typically contains about 45 mg of fat, about half the amount usually present in muscle tissue, but still significant. Bovine liver is also a very good source of dietary lecithin (phospholipids); a gram of liver contains about 25 mg of phospholipids, about four times the amount typically found in muscle. Fats can be removed from the acetonitrile based tissue extracts by liquid extraction with hexane or with SPE with octadecyl silica (C18). Although C18 is effective for removal of most non-polar lipids, it does not remove phospholipids. Excessive amounts of phospholipids can shorten LC column life, contribute to ion-suppression, and contaminate the mass spectrometer. In this study a novel reversed-phase sorbent, Oasis PRiME HLB, is used for highly effective removal of both phospholipids and fats from bovine liver extracts prior to LC-MS/MS analysis. With the new sorbent recoveries of veterinary drugs were similar to results obtained using C18 for clean-up. However, greater than 95 of phospholipids and greater than 85 of fats were effectively removed from the tissue extracts after the simple pass-through SPE procedure. KEYWORDS UPLC-MS/MS, Oasis PRiME HLB Cartridges, veterinary drugs, beef liver Rapid, Simple, and Effective Clean-up of Bovine Liver Samples [ 79 ]

82 [ APPLICATION NOTE ] EXPERIMENTAL UPLC conditions LC system: Column: Mobile phase: Injection vol.: 7 µl Injection mode: Column temp.: 3 C ACQUITY UPLC I-Class with Fixed- Loop Sample Manager ACQUITY UPLC CSH C18, 1.7 µm, 2.1 mm x mm I.D. A:.1 formic in water B:.1 formic acid in 5:5 acetonitrile/methanol partial loop injection Weak needle wash: 1:9 acetonitrile:water (6 µl) Strong needle wash: 5:3:4 water:acetonitrile: IPA (2 µl) Seal wash: Gradient: 1:9 acetonitrile: water Time Flow A B (ml/min) MS conditions Mass spectrometer: Xevo TQ-XS Mode: Positive Ion Electrospray Source temp.: 15 C Desolvation temp.: 4 C Desolvation gas flow: L/Hr Cone gas flow: 3 L/Hr Collision gas flow:.15 ml/min Data management: MassLynx v4.1 Compound MRM Amocixicillin 366.2> >114.1 Ampicillin 35.2> >16.1 Amprolium 243.3> >94.1 Bacitracin A 712.2> >191.1 Ceftiofur 524.3> >285. Chlorotetracycline 479.3> >462.2 Clopidol 192.1> >128. Clorsulon 378> >344. Cloxacillin > >277.1 Danofloxacin 358.2> >96. Desethlylene Ciprofloxacin 35.9> >288.1 Erythromycin 734.7> >576.5 Eprinomectin 915.6> Famphur 326.> >93. Fenbendazole 3.> >159. Flunixin 297.2> >279. Ivermection 892.6> >569.4 Levamisole 25.> >9.8 Melengestrol Acetate 397.4> >279. Monesin 693.7> >461.1 Morantel 221.2> >18. Moxidectin 64.> >498.3 Noviobiocin 613.1> >396. n-methyl-1 3-propanediamine 89.1> >58.2 Oxfendazole 316.2> >284. Oxteracyline 461.4> >365. Penicillin G 335.2> >158.1 Progesterone 315.2> >97. Ractopamine 32.2> >284.2 Sulfachlorpyridazine 285.> >92.1 Sulfadimethoxine 311.1> >92. Sulfamethazine 279.1> >124.1 Sulfaquinoxaline 31.1> >92.2 Tetracycline 445.1> >41.1 Thiabendazole 22.> >131. Tilmicosin > >696.5 Tripelennamine 256.1> >91. Tylosin 916.5> >11.1 Zilpaterol 262.2> >185.1 Cone (V) Collision (ev) RT (min) Table 1. MRM transitions (primary transition first) and instrument parameters used for this study; also listed are the observed retention times (RT) for the compounds [ 8 ] Rapid, Simple, and Effective Clean-up of Bovine Liver Samples

83 Sample preparation 1. Initial Extraction/Precipitation: A 2 g sample of tissue was placed into a 15 ml centrifuge tube containing ceramic homogenizer balls (a Bertin Technologies Precellys Evolution Homogenizer was used for this step). For standards or QC samples the samples were spiked with appropriate amounts of desired analytes. 1 ml.2 formic acid in 85:15 acetonitrile/water was added and the samples were homogenized/extracted for 1.5 minutes. The tubes were then centrifuged at 32 rcf for 5 minutes. Note: The extraction/precipitation step gives good recovery of most compounds of interest but also extracts significant amounts of fats and phospholipids. 2. Pass-through SPE clean-up: An Oasis PRiME HLB Cartridge (6 cc, 2 mg) was mounted on a pre-cleaned vacuum manifold. Cartridge conditioning is NOT required, and was NOT performed. The vacuum was set to 2 psi. A.6 ml portion of the supernatant was passed-through the Oasis PRiME Cartridge and discarded. Collection tubes were then installed and a 1 ml portion of the supernatant was passed-through the Oasis PRiME Cartridge and collected. A 2 µl aliquot of the pass-through clean-up sample was taken and diluted with 4 µl of 1 mm ammonium formate buffer (ph 4.5) prior to UPLC-MS/MS analysis. RESULTS Figure 1 shows the recovery data obtained from replicate analysis of spiked tissue samples (n = 6). Matrix effects averaged about 4. The chromatograms shown in Figure 2 show the effectiveness of the Oasis PRiME HLB Cartridge for removal of 95 of phospholipids from the beef liver extracts. The cartridge also removes more than 9 of hexane extractable fat ppb ppb Figure 1. Recovery data from spiked beef liver sample for low level (1 ng/g in blue) and high level ( ng/g) in red. No Clean-up OASIS PRiME HLB Clean-up Time Figure 2. LC-MS/MS chromatograms showing effective removal of 95 of phospholipids from beef liver extract Rapid, Simple, and Effective Clean-up of Bovine Liver Samples [ 81 ]

84 [ APPLICATION NOTE ] DISCUSSION The procedure utilized in this study was developed from methods presented previously. 1,2 Although the overall method recoveries averaged above 7 percent, lower recovery was observed for some of the more polar compound classes, such as tetracyclines. Unfortunately, no single solvent extraction step will be highly efficient for all target compounds. For most of the lower recovered compounds the signal response and reproducibility are acceptable for target screening analysis. It is important to understand the contribution of the sample cleanup to any observed recovery losses. The SPE recovery data shown in Figure 3 were obtained from beef liver samples spiked after solvent extraction and prior to SPE clean-up. These data indicate that, for most of the compounds, the Oasis PRiME HLB Cartridge clean-up contributes little to the observed recovery losses. However, for ivermectin, monensin, moxidectin, and novabiocin, the post extraction cleanup did introduce measurable recovery losses. More information on these analytes will be presented in future work Figure 3. Recovery of veterinary compounds from blank beef liver extracts spiked after initial extraction and prior to Oasis PRiME HLB pass-through clean-up. CONCLUSIONS A simple and effective extraction/protein precipitation procedure was developed for screening analysis of bovine liver tissue for a wide range of veterinary drugs A simple pass-through clean-up protocol using Oasis PRiME HLB Cartridges was employed to remove greater than 9 of fats and phospholipids from the initial extracts References 1. M. Young and K. Tran, Oasis PRiME HLB Cartridge for Effective Clean-up of Meat Extracts Prior to Multi-Residue Veterinary Drug UPLC-MS Analysis, Waters Application Brief, S. Lehotay, High-Throughput Screening Analysis by UHPLC-MS/MS of >6 Veterinary Drugs in Animal Tissues, 125th AOAC Annual Meeting, Presentation 233, 21 September, 211. The sample preparation methodology produced an extract that was free of particulates and required no subsequent filtration prior to LC-MS analysis Consistent recoveries were observed for a wide range of veterinary drugs using the simple one-step pass-through clean-up protocol with Oasis PRiME HLB Cartridges Waters, The Science of What s Possible, ACQUITY UPLC, MassLynx, Xevo, and Oasis are registered trademarks of Waters Corporation. CSH is a trademark of the Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. March EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 82 ] Rapid, Simple, and Effective Clean-up of Bovine Liver Samples

85 [ APPLICATION NOTE ] Selective and Sensitive Screening of Chloramphenicol in Milk Using the ACQUITY UPLC I-Class System and the ACQUITY QDa Mass Detector Eimear McCall, 1 Sara Stead, 1 Dominic Roberts, 1 and Jennifer A. Burgess 2 1 Waters Corporation, Manchester, UK 2 Waters Corporation, Milford, MA, USA APPLICATION BENEFITS A rapid screening method for chloramphenicol in milk. Quantitative screening of samples with detection below regulatory limits. Simple and sensitive detection of a regulated antibiotic. Easy implementation of mass detection within existing LC workflows. INTRODUCTION Chloramphenicol is an inexpensive broad spectrum antibiotic. In certain susceptible individuals, chloramphenicol has been associated with toxic effects. It is suspected to be both a carcinogen and genotoxic compound, and cases of fatal aplastic anemia (depression of bone marrow) are widely documented. Therefore, its use in food producing animals, insects, and aquaculture is prohibited in the EU, Canada, the U.S., along with a number of Asian and South American countries. Within the EU, Commission Decision 23/181/EC requires that any method for the detection of chloramphenicol in food have a minimum required performance limit (MRPL) of.3 µg.kg -1. Dairy testing laboratories traditionally screen large volume samples using enzyme-linked immunosorbent assay (ELISA), where a quick turnaround time is essential given the short shelf life of the product. Although ELISA allows for sensitive and rapid screening of chloramphenicol, one drawback is the significant level of false positive results, where a false positive level as high as 16 has been reported in certain instances. 1 WATERS SOLUTIONS Oasis HLB Column ACQUITY UPLC HSS C18 Column Due to cross reaction of the immunosorbent with interfering matrix, chloramphenicol can often be misdetected in milk resulting in the unnecessary destruction of the product. Mass detection offers increased selectivity, which helps differentiate between chloramphenicol and contaminating matrix. Furthermore, mass detection coupled with liquid chromatography (LC) eliminates the need for time-consuming derivatization typically required for screening of chloramphenicol by gas chromatography. The ACQUITY QDa Mass Detector can easily be integrated into LC workflows, with the ease of use that does not require specialist mass spectrometry knowledge. ACQUITY UPLC I-Class System ACQUITY QDa Mass Detector KEYWORDS Chloramphenicol, antibiotic, mass detection, milk, veterinary drug Selective and Sensitive Screening of Chloramphenicol in Milk [ 83 ]

86 [ APPLICATION NOTE ] EXPERIMENTAL UPLC conditions UPLC system: Run time: Column: Mobile phase: ACQUITY UPLC I-Class 5 min ACQUITY UPLC HSS C µm, 2.1 x mm 1 mm ammonium acetate in 45:55 methanol:water Sample preparation Semi-skimmed cow s milk (5 g) samples were weighed into 25-mL centrifuge tubes and fortified to allow for.3 µg.kg -1 in the sample. Fat removal and protein precipitation were completed by adding hexane (5 ml) and 7 acetonitrile solution (15 ml) to the sample. Samples were shaken then centrifuged for 1 min at 3,9 G (4 C). The lower acetonitrile layer (5 ml) was removed and diluted to a total volume of 15 ml with 1 methanol. Injection volume: 1 µl MS conditions MS system: ACQUITY QDa Ionization mode: ESI- SIR channel: m/z 321. Cone voltage: 15 V Oasis HLB 3 cc (6 mg) Cartridges were conditioned with methanol (2 ml) and water (4 ml). The diluted samples (15 ml) were loaded on to the cartridge, which was then washed with 5 methanol solution (6 ml). Chloramphenicol was eluted using methanol (6 ml). This eluate was subsequently evaporated under nitrogen, reconstituted with 1 methanol solution (3 µl), and transferred for LC-MS analysis. Probe temp.: Capillary voltage: Sampling rate: Default (6 C) Default (.8 kv) Default (5 Hz) Preparation of standards Blank milk samples (unfortified) were prepared through the sample method outlined. Once dried under nitrogen, the blank matrix samples were reconstituted using standard solutions of different concentrations to allow for a six-point calibration curve over the range equating to.75 to 1 µg.kg -1. RESULTS AND DISCUSSION Prior to the analysis of the milk extracts, the optimum source conditions were investigated for chloramphenicol, where the pre-optimized source parameters of the ACQUITY QDa Detector provided for the most efficient and sensitive detection. Chloramphenicol was most sensitive in negative ionization mode with a voltage of 15 applied to the cone. Using the method described, a matrix-matched calibration curve allowed for accurate quantification of milk samples fortified at the EU MRPL. Excellent linearity (R 2 >.995) was observed for chloramphenicol over the selected working range of.75 to 1 µg.kg, -1 as shown in Figure 1. Compound name: CAP Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: Null, Axis trans: None 2. Residual. -2. ng/ml Response 2 ng/ml Figure 1. Matrix-matched calibration curve with duplicate injections, equating to range of.75 to 1 µg kg -1. [ 84 ] Selective and Sensitive Screening of Chloramphenicol in Milk

87 In order to assess matrix effects, the matrix matched calibration curve was compared to a solvent calibration curve over the same range. Figure 2 shows an overlay of the matrix-matched and solvent calibration curves. From this overlay, matrix effects (ion enhancement) are evident. This can also be demonstrated statistically using the slope of each of the calibration Slope sample Slope standard curves ( x ), where matrix effects of 177 are calculated, confirming ion enhancement. Despite the matrix effects observed, the accurate quantitative screening of chloramphenicol in milk is readily achievable with the use of the matrix-matched calibration curve. Replicate blank samples of milk were analyzed, where no false positives were detected. An example chromatogram for an extracted blank is shown in Figure 3A, where only baseline noise was observed. Four replicate milk samples were fortified at the EU MRPL, extracted and analyzed in duplicate (n= 8), in the absence of expensive deuterated internal standard. The resultant samples were quantified against the matrixmatched calibration curve. Figure 3B shows the chloramphenicol peak observed for replicates at the fortification level (.3 µg.kg -1 ), where good peak-to-peak signal-to-noise ratio (S/N) of 5 was achieved. Peak area B. MM milk y = x R² = A. Solvent y = x R² = Concentration (ppb) Figure 2. Overlay of calibration curves 2A. solvent (1 methanol), and 2B. matrix matched (in milk) equating to.75 to 1 µg.kg -1. A. Blank milk sample B. Chloramphenicol in milk at.3 µg.kg -1 Figure 3. Selected ion recording (SIR) chromatogram of chloramphenicol (m/z 321) for: 3A. Blank milk extracts (n= 2); 3B. Milk fortified at MRPL (.3 µg.kg -1 ). Replicate injections (n= 8) are overlaid with S/N 5. Selective and Sensitive Screening of Chloramphenicol in Milk [ 85 ]

88 [ APPLICATION NOTE ] As shown in Figure 3B, good repeatability was observed for the retention time. This is also interpreted statistically, along with recoveries, in Table 1. An average recovery of 72.7 was determined, which is acceptable in accordance with European Commission Decision 22/657/EC. Excellent repeatability was obtained over eight replicates at the regulatory limit with relative standard deviation (RSD) of <4.6. This demonstrates that the method is both robust and sensitive for the quantitative screening of chloramphenicol in milk at the regulatory limit. Table 1. Recoveries (n= 8) achieved from SPE method, showing good repeatability of sample preparation and instrument. Replicate number Recovery Retention time (min) Average Std dev RSD CONCLUSIONS An accurate and robust method has been developed for the reliable quantitative screening of chloramphenicol in milk. This method has proven to achieve levels of detection that can easily meet regulatory requirements, in the absence of an internal standard, while showing excellent repeatability at the EU MRPL of.3 µg.kg -1. The additional discrimination that mass detection affords results in increased specificity, where no false positives were detected. References 1. V Gaudin, N Cadieu, O Maris. IVth International Symposium on Hormone and Veterinary Drug Residue Analysis, Belgium, 22. Waters, ACQUITY UPLC, and The Science of What s Possible are registered trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 214 Waters Corporation. Produced in the U.S.A. November EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 86 ] Selective and Sensitive Screening of Chloramphenicol in Milk

89 [ FOOD CONTACT MATERIALS ] [ 87 ]

90 [ FOOD CONTACT MATERIALS ] [ 88 ]

91 [ APPLICATION NOTE ] Quantifying Primary Aromatic Amines in Polyamide Kitchenware Using the ACQUITY UPLC I-Class System and Xevo TQ-S micro Steven Haenen and Marijn Van Hulle Waters Corporation, Brussels, Belgium APPLICATION BENEFITS Single method for analysis of 23 PAAs No need for ion-pairing reagents, or the removal of acetic acid from the sample extract prior to analysis Sensitive detection at levels well below the EU guidelines with Xevo TQ-S micro Triple Quadrupole Mass Spectrometry INTRODUCTION Primary Aromatic Amines (PAAs) are a class of compounds of which the simplest form is aniline (Figure 1). PAAs are substances that are used, for example, in the production of certain colorants, so-called azo pigments, notably in the color range yellow orange red. Whereas a large number of PAAs are safe for human health, some PAAs are known human carcinogens. For kitchenware, paper napkins, baker s bags with colorful print and other printed items that come in contact with food, some PAAs may pose a health risk, if they are transferred to the food. Compound Mass Structure Aniline 93 o-toluidine 17 2,4-Diaminotoluene 122 WATERS SOLUTIONS ACQUITY UPLC I-Class System Xevo TQ-S micro o-anisidine 123 Figure 1. Chemical structures of some PAAs. ACQUITY UPLC HSS T3 Column MassLynx MS Software TargetLynx XS Application Manager KEYWORDS PAAs, primary aromatic amines, kitchenware, utensils, migration, food contact materials, FCMs Quantifying Primary Aromatic Amines in Polyamide Kitchenware [ 89 ]

92 [ APPLICATION NOTE ] Because of the potential health risks, specific migration limits (SMLs) are put in place. 1 According to the regulation on plastics EU 1/211: Plastic materials and articles shall not release primary aromatic amines, excluding those appearing in Table 1 of Annex I, in a detectable quantity into food or food simulant. The detection limit is.1 mg of substance per kg of food or food simulant. The detection limit applies to the sum of primary aromatic amines released. The provisions in Regulation 1/211 state that for primary aromatic amine migration from polyamide kitchenware, only one migration test will be carried out, if this first extract is compliant with the summed SML (SML(T)) of.1 mg/kg. However, if this first simulant extract exceeds the permitted SML(T), two subsequent migration studies are required. 2 This PAAs migration testing is conducted with simulant B, 3 (w/v) acetic acid, as it has been demonstrated that this simulant represents the worst case for the migration of PAAs from polyamide kitchenware. 3 PAAs are small, basic compounds, which are ionized with low ph. As a result of their basic properties and the 3 acetic acidic sample solvent, some PAAs don t focus well on the head of the column, resulting in poor peak shape and/or loss of retention. In order to improve chromatographic retention ion-pairing reagents are often used. 2 Unfortunately these reagents have a negative impact on the electrospray sensitivity and are to be avoided where possible. In this application note we describe a LC-MS/MS method for the analysis of 23 common PAAs in kitchenware after migration using Waters ACQUITY UPLC I-Class System coupled to a Xevo TQ-S micro Mass Spectrometer. The described method does not use an ion-pair reagent to improve chromatographic retention. [ 9 ] Quantifying Primary Aromatic Amines in Polyamide Kitchenware

93 EXPERIMENTAL UPLC conditions UPLC system: ACQUITY UPLC I-Class Sample manager: Flow-Through Needle Column: ACQUITY UPLC HSS T3, 1.8 µm, 2.1 x mm Mobile phase A: Water Mobile phase B: Methanol Column temp.: 45 C Sample temp.: 1 C Flow rate:.4 ml/min Run time: 15 min Injection volume: 2 µl Gradient: min 5 B 1 min B 12 min 5 B 12.1 min 5 B 15 min 5 B MS conditions MS system: Xevo TQ-S micro Ionization mode: ESI + Capillary 2 kv voltage: Desolvation temp.: 6 C Desolvation gas flow: 12 L/hr Source temp.: 15 C Acquisition: Multiple Reaction Monitoring (MRM) MS methods and data acquisition Two MRM transitions were used, unless otherwise stated. The dwell times were chosen automatically using the built-in points-per-peak calculator in the MS method. The data were acquired using MassLynx v. 4.1 Software, and processed using TargetLynx XS Application Manager. Table 1 summarizes all MRM transitions. Figure 2 shows the retention time windows of the MRM method. Table 1. Overview of MRM transitions for all 23 PAAs. Compound Transitions Cone voltage (V) Collision energy (ev) Aniline 93.8> o-toluidine 2,4-Diaminotoluene o-anisidine 4-Chloroaniline 3-Chloro-o-toluidine 2,4,5-Trimethyl aniline 2-Methoxy-5-methylaniline 4-Chloro-2-methylaniline 2-Amino naphthalene 2-Methyl-5-nitroaniline 4-Aminobiphenyl 2-Aminobiphenyl Benzidine 4-Phenyl azoaniline 4,4'-Diamino diphenylmethane 4,4'-Oxydianiline 3,3'-Dimethyl benzidine 17.8> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > ,4'-Thiodianiline > o-amino azotoluene 226.> ,3'-Dimethyl-4,4'- diaminodiphenylmethane 227.> ,3'-Dimethoxy benzidine 3,3'-Dichloro benzidine 4,4'-Methylene bis (2-chloroaniline) 245.> > > > > > Quantifying Primary Aromatic Amines in Polyamide Kitchenware [ 91 ]

94 [ APPLICATION NOTE ] Figure 2. Retention time windows for the PAAs acquisition method. Standards A mixed standard solution containing all PAAs at a concentration of µg/ml was used. The working standards were further diluted with the 3 acetic acid food stimulant solution. For the solvent calibration a dilution series starting at ng/ml down to a level of.78 ng/ml was made. Sample preparation Nine polyamide kitchenware utensils were extracted with a 3 acetic acid solution according to the procedure described in the EU 1/211 guidelines. 1 [ 92 ] Quantifying Primary Aromatic Amines in Polyamide Kitchenware

95 RESULTS AND DISCUSSION UPLC METHOD DEVELOPMENT Because of the basic properties of PAAs, and the fact that acetic acid is used as a migration stimulant, some PAAs don t focus well on the head of the column, resulting in poor peak shape and/or loss of retention. Aniline elutes early and is therefore prone to this effect. As a result, some literature references cite the use of ion-pair reagents. 2 Adding ammmonium hydroxide to the 3 acetic acid samples prior to injection, the ph of the sample is increased and the polar and weakly basic PAAs such as aniline will be in their neutral form. A volume of 1 µl of a 25 NH4OH solution was added to 1 ml of sample. This approach resulted in more robust results and is therefore preferred over the use of ion-pair reagent. Figure 3 shows a chromatogram of aniline with an unchanged ph (top) and adjusted ph (bottom). The neutralization of the ph drastically improves the peak shape of aniline, without the need for ion-pairing reagent. Figure 3. Chromatogram of aniline in 3 acetic acid food stimulant without (top) and with (bottom) ph adjustment. Figure 4. Chromatograms of all 23 PAAs. Quantifying Primary Aromatic Amines in Polyamide Kitchenware [ 93 ]

96 [ APPLICATION NOTE ] LINEARITY Calibration curves were prepared from.78 ng/ml to ng/ml for all compounds. An example is given for aniline (Figure 5). For each calibration curve, a linear regression and a 1/X weighting was applied. All compounds show good linearity across the range of concentrations as well as excellent residual values. Compound name: Aniline Correlation coefficient: r =.99993, r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None Residual ppb 8 Response ppb Figure 5. Calibration curve (bottom) and residuals plot (top) for aniline in the range.78 to ng/ml. [ 94 ] Quantifying Primary Aromatic Amines in Polyamide Kitchenware

97 Acidified mobile phases aid in the protonation of compounds and therefore improve the sensitivity in positive ion electrospray. As no acid was added to the mobile phases, we investigated whether a post-column addition (PCA) with formic acid would be beneficial. Using the Xevo TQ-S micro s built-in IntelliStart fluidics, a solution of 2 formic acid was infused at a constant flow rate of 2 µl/min into the UPLC flow exiting the column. As such the formic acid solution was diluted 2-fold with the mobile phase, resulting in a final concentration of.1 of formic acid going into the ESI source. Figure 6 shows how this PCA was configured in the acquisition method, while Figure 7 shows the chromatograms for a selection of PAAs with (top trace) and without (bottom trace) this post-column addition. For better interpretation, the intensity axes have been linked. As can be seen from the chromatograms, the sensitivity is significantly improved when formic acid is added to the eluent. Figure 6. Post-column addition in the MS acquisition method. Figure 7. Increase in sensitivity with the use of a formic acid post-column addition (top), and without (bottom), illustrated for: A. aniline, B. o-toluidine, C. 4-Chloroaniline, D. 2,4,5-Trimethylaniline, E. 2-Methoxy-5-methylaniline, and F. 4-Chloro-2-methylanaline. Quantifying Primary Aromatic Amines in Polyamide Kitchenware [ 95 ]

98 [ APPLICATION NOTE ] Table 2 summarizes the quantitation limits (LOQ) for all compounds using this PCA approach. The LOQ is defined as the concentration giving rise to a signal-to-noise (S/N) value of 1:1. For the calculation of S/N, raw data was used and the peak-to-peak algorithm was applied. An extrapolation was made in most cases, as the reported S/N values were still significantly high, even at the lowest reported standard level of.78 ng/ml. Calculated LOQs below 2 pg/ml are not mentioned specifically but are cut off at this level. The reported LOQ concentrations range between 2 pg/ml and 3 pg/ml. MATRIX EFFECTS Internal standards were not used in this method. Therefore it was investigated whether the food simulant extract leads to ion suppression. One of the samples was spiked to a final concentration of 1 ppb and this sample was compared with a standard dissolved in the same food stimulant solution. All spike recoveries were within 9 to 17, indicating that matrix effects were low to non-existing for the 23 compounds under investigation. Compound S/N ratio LOQ (ng/ml) Aniline o-toluidine 768 <.2 2,4-Diaminotoluene o-anisidine Chloroaniline ,4,5-Trimethyl aniline 693 <.2 2-Methoxy-5-methylaniline 1444 <.2 4-Chloro-2-methylaniline 353 <.2 2-Amino naphthalene 1858 <.2 2-Methyl-5-nitroaniline Aminobiphenyl Aminobiphenyl Benzidine 559 <.2 4-Phenyl azoaniline 1931 <.2 4,4'-Diamino diphenylmethane 1353 <.2 4,4'-Oxydianiline ,3'-Dimethyl benzidine ,4'-Thiodianiline 2582 <.2 o-amino azotoluene 1746 <.2 3,3'-Dimethyl-4,4'-diaminodiphenylmethane 1818 <.2 3,3'-Dimethoxy benzidine 528 <.2 3,3'-Dichloro benzidine 926 <.2 4,4'-Methylene bis (2-chloroaniline) 1522 <.2 Table 2. Calculated S/N values at.78 ng/ml and estimated LOQ values for all 23 PAAs investigated. KITCHENWARE SAMPLES Using the external calibration curves, nine kitchenware samples were quantified. Except for aniline and 4,4'-diamino diphenylmethane found in all nine samples at levels between.4 to 1.1 ppb and.4 to.11 ppb, respectively, no other PAAs were detected. Figure 8 shows the chromatograms of aniline in the sample containing.4 ppb and of 4,4'-diamino diphenylmethane in the sample containing.4 ppb. As can be seen sensitivity was excellent at these sub ppb level Koopman 19 F1:MRM of 1 channel,es Koopman 18 F15:MRM of 2 channels,es+ neutralized > e+5 neutralized > e min min Figure 8. Chromatograms of aniline in kitchenware samples present at.4 ppb (left), and of 4,4'-Diamino diphenylmethane in the sample containing.4 ppb (right). [ 96 ] Quantifying Primary Aromatic Amines in Polyamide Kitchenware

99 CONCLUSIONS We have demonstrated a sensitive method for 23 PAAs with very easy sample preparation. The addition of ammonium hydroxide as neutralizing agent, and a post-column addition of formic acid into the Xevo TQ-S micro via IntelliStart s built-in fluidics resulted in a very sensitive assay which could reach sub ppb levels. Linearity was observed over a large range and up to ppb. The samples were all below detection limits except for aniline which was detected at.4 to 1.1 ppb, and 4,4'-diamino diphenylmethane which was detected at.4 to.11 ppb. The total PAAs content for all samples was below the SML(T) of.1 mg/kg as stipulated in the regulations EU 1/211. References 1. Commission Regulation (EU) No 1/211 of 14 January 211 on plastic materials and articles intended to come into contact with food. 2. LB-NA EN-N, Technical guidelines on testing the migration of primary aromatic amines from polyamide kitchenware and of formaldehyde from melamine kitchenware. 3. Analysis of primary aromatic amines (PAA) in black nylon kitchenware 214. Selected samples from the Norwegian Market. Waters, Xevo, ACQUITY UPLC, UPLC, MassLynx, and The Science of What s Possible are registered trademarks of Waters Corporation. TargetLynx and IntelliStart are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 216 Waters Corporation. Produced in the U.S.A. September EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: Quantifying Primary Aromatic Amines in Polyamide Kitchenware [ 97 ]

100 [ TECHNICAL NOTE ] Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System Baiba Čabovska Waters Corporation, Milford, MA, USA TECHNOLOGY BENEFITS Simple MS methodology using high-resolution mass spectrometry (HRMS) that can be adopted for cosmetics, 3D printing media, food, and pharmaceutical packaging extractable applications. Streamlines the structural elucidation process for packaging extracts by utilizing MS E data of accurate mass precursor and fragment ion information on a single software platform A rapid and automated way to evaluate information for an unknown component (m/z) by ranking the possible elemental compositions, and searching databases for likely structures ranked based on fragmentation matching. WATERS SOLUTIONS UNIFI Scientific Information System Ion Mobility Mass Spectrometry Quadrupole Time-of-Flight Mass Spectrometry KEYWORDS extractables, screening, elucidation, multivariate statistics, scientific library, MS, E HDMS E INTRODUCTION Characterization of packaging, food contact materials, medical devices, and many other consumables used in various industries is becoming more and more important due to ever-increasing global regulations. The initial step in characterizing extractables from packaging includes targeted screening, i.e. testing the extracts for known compounds. This is a well-established process and can be performed in various ways by using analytical techniques ranging from GC-FID-MS to LC-UV-MS. However, the final packaging may have impurities present from starting materials and additional degradants such as those formed during the molding process. The structural elucidation of unknowns is typically a very complex and time-consuming process that requires the analyst to have a high level of expertise. Waters UNIFI Scientific Information System provides a simple workflow that includes scientific library creation, multivariate statistical analysis, elucidation, and reporting. This single platform Informatics solution enables analysts to evaluate complex data in a more efficient way through simplifying data review and facilitating the decision-making process. DISCUSSION As shown in Figure 1, the workflow starts with a non-targeted, data independent analysis (MS E or HDMS E ) acquired on a quadrupole time-of-flight mass spectrometer (QToF) or on an ion mobility QTof mass spectrometer (IMS-QTof). The QTof MS is operated in the alternate scanning MS E mode (where the E represents elevated collision energy), as this technique provides two MS scan functions for data acquisition in one analytical run. The first scan function acquires MS data using low collision energy and collects information on the precursor ions in the sample. For the second scan function the collision energy is ramped from low to high energy which allows for the collection of fragment ions over a wide m/z range. With IMS-QTof, an additional dimension of separation is achieved by the inclusion of ion mobility, thus achieving High Definition Mass Spectrometry E (HDMS E ). These types of data acquisitions allow simultaneous collection of precursor and fragment ion information, which is crucial when doing elucidation for unknown compounds. In extractables testing complete information about sample extract is rarely available. Therefore after the targeted screening, the elucidation steps in nontargeted screening are essential. [ 98 ] Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System

101 The sample separation prior to MS analysis can be performed by liquid chromatography using UPLC, by gas chromatography using APGC, as well as by convergence chromatography using UPC. 2 MS E or HDMS E componentized data Non-targeted screening Binary compare Multivariate statistics Scientific library Targeted screening/ Quantitation Discovery tool: Elemental composition ChemSpider search Fragment Match Report Figure 1. Screening workflow in UNIFI. Prior to starting data analysis, the user can create a scientific library based on knowledge of expected compounds in the sample extract, i.e. if the starting compounds in the formulation of a plastic material are known, or a literature search has provided list of compounds that are typically encountered in similar types of packaging. Additionally, regulations provide lists of compounds that are either allowed or prohibited in certain types of packaging, (e.g. food contact materials). The scientific library (Figure 2.) can contain as much information as is available. The most common information typically included is the compound name, its molecular formula, structure, item tag, and fragmentation information. For ion mobility data, the information needed in screening would be the collisional cross section value (CCS). 1 More extensive information about each compound in the library can reduce the number of false positives during targeted screening analysis. Examples of additional information that could be added to a UNIFI scientific library include MS spectra and other relevant documents (e.g. MSDS, articles, SOPs). Once the data has been acquired, UNIFI uses a target list created by the user from the library to process the raw data and search for compounds which match acceptance criteria. It is also possible to create target lists manually, if required. The user can set up processing criteria such as retention time and mass accuracy tolerances. Subsequently, it is possible to review the proposed identifications based on the number of expected fragments versus the number of expected fragments found, expected and observed CCS values for IMS data with CCS delta (), isotope intensity matches in ppm or, among other parameters. Figure 2. UNIFI scientific library screen shot. Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System [ 99 ]

102 [ TECHNICAL NOTE ] UNIFI allows each user to have a customized workflow which displays information in the preferred dashboard for review (Figure 3). It is possible to review the spectrum for precursor and fragments. Extracted ion chromatograms (XIC) can be displayed for all precursors as well as for fragments. Summary plots can be used to verify the presence or absence, or changes in intensity of a target in other injections. If the appropriate standards have been analyzed during the sample run and a calibration curve is available, it is possible to quantify the identified targets in the analysis. RT, Mass accuracy Precursor ion spectrum XIC Fragment ion spectrum Figure 3. Example data review window. After reviewing the identified targets for false positives and removing them from the identified list, the next question to be answered is What else is in my sample? or What are the differences between a blank extract and sample extract or between these two samples?. UNIFI has two tools for comparison and statistical analysis. The first one, Binary Compare, allows the user to compare two injections. One injection must be labeled as a reference sample, in this case, an extraction blank. [ ] Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System

103 Masses in the reference spectra and the unknown spectra are considered to be the same component if they are within the specified mass and retention time tolerances. The comparison can be presented graphically as a mirror image of base peak intensity chromatograms (BPI), total ion count chromatograms (TIC), or as a table of candidate masses (Figure 4). Also the spectra of the compound in the reference sample can be displayed in comparison to the unknown sample. The column labeled Match Type shows whether the candidate is present only in the unknown sample or in the reference sample, or both. The corresponding match types would be Unknown Unique, Reference Unique, or Common. Typically, compounds that are unique to the sample and absent in the reference sample would be of most interest in an analysis. Reference Sample Difference Figure 4. Binary Compare plot, table, and spectra. Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System [ 11 ]

104 [ TECHNICAL NOTE ] If there is a need to compare more than two samples or groups of samples, UNIFI provides Principal Component Analysis (PCA) and other models for data reduction and evaluation by an integrated workflow with statistical software package- EZInfo. PCA is a statistical tool which allows the reduction of a large set of multivariate data into uncorrelated variables called principal components. The differences among the groups of samples are emphasized by Projection to Latent Structures Discriminant Analysis (PLS- DA) model (Figure 5), where a sample group is specified. PLS-DA models the quantitative relationships between the variables X (predictors) and Y (responses) for all of the sample groups. Subsequently, Orthogonal Projection to Latent Structures Discriminant analysis (OPLS-DA) plot demonstrates the differences between two groups.² The data points (markers) in the loadings plot and S-plot are called Accurate Mass/Retention Time pairs (AMRTs). Individual markers that contribute to the biggest differences between the samples can be selected from either the loadings plot or the S-plot and transferred back to the discovery tool for elucidation. When transferring the selected markers, labels can be added to make the data easier to sort and to keep track of markers for different sample groups. When an individual marker is selected from the marker matrix table, a TrendPlot is displayed, allowing the analyst to quickly evaluate its presence in the other samples or injections (Figure 6). PLS-DA OPLS-DA Loadings plot S-plot Figure 5. Example of statistical plots provided with UNIFI s multivariate analysis tools. Figure 6. Markers from statistical analysis and a trendplot. [ 12 ] Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System

105 Once the markers are selected either from Binary Compare or from MVA analysis, UNIFI s Discovery tool can be used to find the possible identity of the ion. The Discovery tool automatically combines all of the analytical information contained in the data: accurate mass, isotope pattern, fragmentation in the high collision energy channel with the structural database search. The Results table shows the possible molecular formulas for the ion, corresponding structures from a ChemSpider search, and a number of fragments that can be matched to each structure based on fragmentation data. Information returned from the search (Figure 7) also includes the number of citations and synonyms used for each structure. Many polymer additives have common names like Irganoxes and Tinuvins, which helps in further narrowing down the possible compound choices. When the decision for the compound identity is made, the chosen structure and name can be assigned to the candidate mass ion. Assignment will change the identification status of the candidate to identified. All of the elucidated compounds can be added to UNIFI s scientific library to be used in subsequent targeted screening analysis. One of the final steps of the analysis is to create a report. A report template can be embedded in the UNIFI Analysis method which can be used for similar types of analysis (Figure 8). The report can be customized to include all the relevant information such as analysis method, processing parameters, chromatograms, spectra, and identified compound summary tables, among others. Figure 7. UNIFI Summary table for Discovery toolset. Figure 8. Report examples for analysis information and target summary. Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System [ 13 ]

106 [ APPLICATION TECHNICAL NOTE NOTE ] ] CONCLUSIONS The UNIFI Scientific Information System provides analysts with a well-established workflow for extractable screening analysis. The UNIFI workflow starts with the scientific library for targeted screening, followed by statistical analysis for the determination of markers or relevant compounds. The Discovery tool automatically utilizes information-rich raw data for elemental compositions, followed by a structural database search and fragmentation assignments. This integrated workflow reduces the amount of time required for extractables screening analysis with structural elucidation. References 1. M McCullagh, V Hanot, and S Goscinny. Use of Ion Mobility Spectral Cleanup and Collision Cross Section Values to Increase Confidence and Efficiency in Pesticide Residues Screening Strategies. Waters application note no. 7258en. June, B Cabovska. Non-targeted screening analysis of packaging extracts using the UNIFI Scientific Information System. Waters application note no en. March, 215. Waters, UNIFI, UPLC, High Definition Mass Spectrometry, HDMS, UPC,2 and The Science of What s Possible are registered trademarks of Waters Corporation. TrendPlot is a trademark of Waters Corporation. 216 Waters Corporation. Produced in the U.S.A. April EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 14 ] Screening Workflow for Extractables Testing Using the UNIFI Scientific Information System

107 [ NATURAL TOXINS ] [ 15 ]

108 [ NATURAL TOXINS ] [ 16 ]

109 [ TECHNOLOGY BRIEF ] Oasis PRiME HLB Cartridges and DisQuE QuEChERS Products for UPLC-MS/MS Mycotoxin Analysis in Cereal Grains Michael S. Young and Kim Tran Waters Corporation, Milford, MA, USA The Oasis PRiME HLB Cartridge and DisQuE Products for QuEChERS provide simple and effective extraction and cleanup of cereal grains prior to UPLC-MS/MS analysis. GOAL A simple modified QuEChERS extraction protocol and simple clean-up strategies suitable for multiresidue mycotoxins analysis by UPLC-MS/MS. 1. Deoxynivalenol (75 ng/g) 2. Aflatoxin G2 (1 ng/g) 3. Aflatoxin G1 (1 ng/g) 4. Aflatoxin B2 (1 ng/g) 5. Aflatoxin B1 (1 ng/g) 6. Fumonisin B1 (8 ng/g) 7. HT-2 Toxin (5 ng/g) 8. Ochratoxin B (3 ng/g) 9. T-2 Toxin (5 ng/g) 1. Fumonisin B2 (2 ng/g) 11. Zearalenone ( ng/g) 12. Ochratoxin A (3 ng/g) , BACKGROUND Mycotoxins are toxic compounds produced by molds or other fungi that can grow on foodstuffs intended for domestic animal or human consumption. Ingestion of food containing only parts-per-billion (µg/kg) concentration of some mycotoxins may cause severe illness. Therefore, sensitive and reliable analytical methods are required to determine mycotoxins in foods and feeds. Cereal grains, such as wheat, rice, and maize are important examples of these types of foods. Many of the natural constituents of these grains are potential interferences for LC-MS/MS analysis. Although proteins and starches are removed during the QuEChERS extraction by partition, precipitation, and centrifugation, significant amounts of fat and lecithins (phospholipids) are co-extracted along with the target mycotoxins. Time Figure 1. Ion chromatograms obtained from a wheat flour sample fortified with 12 mycotoxins at the indicated levels. The presence of these co-extracted substances can lead to interference in the LC-MS analysis, contamination of the analytical column and other components of the UPLC system, and contamination of the mass-spectrometer itself. Fats have traditionally been removed from QuEChERS extracts using cumbersome hexane defatting steps or by the use of reversed-phase sorbents such as C18 silica. Although these techniques may be effective for fat removal, neither of these procedures removes phosholipids. Oasis PRiME HLB Cartridges and DisQuE QuEChERS Products for UPLC-MS/MS Mycotoxin Analysis in Cereal Grains [ 17 ]

110 [ TECHNOLOGY BRIEF ] THE SOLUTION Pass-through clean-up with Oasis PRiME HLB cartridge for fat and phospholipid removal and dspe (dispersive SPE) for removal of residual sugars and other polars. The recoveries of the mycotoxins are not compromised using these cleanup protocols. UPLC Conditions UPLC system: Column: Mobile phase: ACQUITY UPLC I-Class (FTN) CORTECS UPLC T3, 1.6 µm, x 2.1 mm A:.5 formic acid, 5 mm ammonium formate in water B:.5 formic acid, 5 mm ammonium formate in 5:5 methanol/acetonitrile EXPERIMENTAL QuEChERS extraction. Cereal grain flours were purchased at a local grocery store. A 2 g sample was weighed into a 5 ml centrifuge tube. 1 ml water and 1 ml 1:9 formic acid/ acetonitrile were added and the sample was placed on automated shaker for 1 hour. Then, QuEChERS salts (contents of DisQuE pouch for CEN, pn ) were added and the tube was shaken vigorously by hand for 1 minute. After centrifugation (32 rcf for 5 minutes), a portion of the supernatant was taken for cleanup. Clean-up. An Oasis PRiME HLB Cartridge (3 cc, 15 mg, pn ) was mounted on a pre-cleaned vacuum manifold set to minimal vacuum (approx 2 psi). No cartridge conditioning is required or was performed. A.4 ml aliquot of the supernatant was passed-through the Oasis PRiME Cartridge and discarded. Then a 1 ml portion of the supernatant was passed through the cartridge and collected. The collected extract was then transferred to a 2 ml dspe tube (pn ) containing a mixture of sorbents. After centrifugation (1 minute at 135 rcf), a 5 µl aliquot was taken, evaporated under a gentle nitrogen stream, and reconstituted in 25 µl of 15:85 acetonitrile/water. Injection vol.: 1 µl Column temp.: 3 C Needle wash and Sample 1 formic acid, 1 mm citric acid in 1:1:1:1 manager purge: water/methanol/isopropanol/acetonitrile Seal wash: 1:9 methanol: 1:9 methanol:water MS Conditions MS system: Ionization mode: Source temp.: Desolvation temp.: Desolvation gas flow: Cone gas flow: Data management: Gradient: Time (min) Flow rate (ml/min) A B Initial Xevo TQ-S micro ESI+ 15 C 5 C L/Hr 3 L/Hr MassLynx v4.1 INSTRUMENTAL CONDITIONS UPLC-MS/MS conditions are presented below. Table 1 presents the target compound list, MRM transitions, and mass-spectral conditions used for this study. Six point matrix-matched calibration curves run bracketing the target levels showed good linearity (R2>.99) for all compounds. RESULTS AND DISCUSSION Recoveries were measured for 12 mycotoxins at a low and high level. The high level was consistent with EU maximum permitted levels for aflatoxins, fumonisins, ochratoxin A, and zearalenone, and recommended levels for T2 and HT2 toxins (see Figure 1). The low level was 1/4 X the high level (.25 ng/g for aflatoxins). In wheat flour, total method recoveries for both the low and high level spiked samples were better than 8 for all target compounds except for zearalenone (73). Similar performance was observed for whole rice and maize flours. [ 18 ] Oasis PRiME HLB Cartridges and DisQuE QuEChERS Products for UPLC-MS/MS Mycotoxin Analysis in Cereal Grains

111 Very little of any recovery loss for any of the toxins was caused by the pass-through or dspe clean-up steps. Figure 1 shows ion chromatograms obtained from a sample spiked at.25 ng/g (ppb) in wheat flour. Figure 2 shows chromatograms that illustrate the effectiveness of the Oasis PRiME Cartridge for phosholipid removal; greater than 95 of phospholipids and greater than 8 of total fats were removed using the Oasis PRiME Cartridge. Table 1. Target compounds, MRM transitions, and mass spectral conditions used for this study (note: For HT-2 and T-2 toxins the ammonium adducts were used for MRMs; to enhance this response, ammonium formate was incorporated in the mobile phase). Mycotoxin Retention (min) Deoxynivalenol 3.2 Aflatoxin G2 4.8 Aflatoxin G Aflatoxin B2 5.1 Aflatoxin B Fumonisin B1 5.7 HT-2 Toxin 5.72 Ochratoxin B 6.6 T-2 Toxin 6.3 Fumonisin B Zearalenone 6.57 Ochratoxin A 6.6 MRM transitions > > > > > > > > > > > > > > > > > > > > > > > > Cone (V) Collision (ev) No 78234"516 Clean-up Oasis!"#$#&'$()*+, PRiME HLB Pass-through &"##-./1234"516 Clean-up 6/8#6/83$6$9# phospholipids Figure 2. Pass-through clean-up using Oasis PRiME HLB; nearly complete removal of phospholipids from QuEChERS extract of wheat flour. CONCLUSIONS An improved QuEChERS extraction procedure was shown to be effective for simultaneous extraction of 12 mycotoxins from wheat, rice, and maize flours. Pass-through cleanup with an Oasis PRiME HLB Cartridge effectively removed greater than 8 of fats and greater than 95 of phospholipids from the QuEChERS extract. Polar co-extractables were effectively removed from the QuEChERS extract using a mixed sorbent dspe clean-up. LOQ acceptable for meeting EU regulations was achieved using the sample preparation protocol prior to LC-MS/MS analysis using Xevo TQ-S micro Mass Spectrometer. Time Waters, The Science of What s Possible, ACQUITY UPLC, CORTECS, Xevo, Oasis, and MassLynx are registered trademarks of WatersCorporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. January EN RF-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: Oasis PRiME HLB Cartridges and DisQuE QuEChERS Products for UPLC-MS/MS Mycotoxin Analysis in Cereal Grains [ 19 ]

112 [ APPLICATION NOTE ] The Analysis of Cyanotoxins, Including Microcystins, in Drinking and Surface Waters by Liquid Chromatography-Tandem Quadrupole Mass Spectrometry Julie Degryse, 1 Marijn Van Hulle, 2 and Simon Hird 3 1 Watergroep, Heverlee, Belgium; 2 Waters Corporation, Brussels, Belgium; 3 Waters Corporation, Wilmslow, UK APPLICATION BENEFITS Direct injection without the need for sample preparation Sensitive detection at sub-ppb levels with the Xevo TQ-S Use of standard addition to avoid expensive internal standards INTRODUCTION Cyanobacteria (blue-green algae) are photosynthetic organisms found in both marine and freshwater environments. The occurrence of algal blooms has increased over time, and anthropogenic input of phosphorus and nitrogen into natural water bodies are contributors to this increase in toxic algal blooms worldwide. Many species of cyanobacteria produce secondary metabolites, some of which are toxic to higher organisms. Microcystins are a group of non-ribosomally synthesized cyclic heptapeptides, of which microcystin-lr (MC-LR) is the most common (Figure 1). About 9 microcystin variants have been identified, including a number of microcystin-lr homologues, some of which have similar toxicity to MC-LR. Their hepatotoxicity may cause serious damage to the liver. Microcystins can be introduced to tissues of organisms through the diet or by ingestion of contaminated water. Microcystins are readily water-soluble so once ingested they accumulate in the liver via the bile acid transport system. The risk to human health due to the increasing presence of these toxins is of concern. 1,2 WATERS SOLUTIONS ACQUITY UPLC I-Class System ACQUITY UPLC BEH C 18 Column Xevo TQ-S MassLynx MS Software TargetLynx XS Application Manager KEYWORDS Cyanotoxins, cylindrospermopsin (CYL), anatoxin-a (ANA), nodularin (NOD), microcystins, microcystin-lr (MC-LR), MC-DeRR, MC-RR, MC-YR, MC-LY, MC-LW, MC-LF Microcystin-LR O CH 3 CH 3 Cylindrospermopsin O S O HO O H 3 C CH 3 H 3 C H NH H O HN H N N H O HN O HN NH 2 OH H N OH H N HN O CH 3 O O O CH 3 N NH H N O OH H 3 C O O CH 2 C H 3 NH NH O CH 3 Anatoxin-a H N H H 3 C H O Figure 1. Structures of three of the cyanotoxins analyzed in this study. [ 11 ] The Analysis of Cyanotoxins, Including Microcystins, in Drinking and Surface Waters

113 The World Health Organization (WHO) provisional guideline value for MC-LR of 1 µg/l, which was deemed protective of total microcystins, has been adopted as a basis for national standards or guideline values for drinking water in many countries. 3 Some countries set limits for microcystins in water bodies used for recreation and these tend to be at higher concentrations (typically between 1 and µg/l). Other cyanotoxins, such as cylindrospermopsin and anatoxin-a (Figure 1), are regulated in some countries by the use of guidance values and health alert limits. Detection of microcystins is a challenging problem for water testing laboratories. Historically, antibody-based enzyme-linked immunosorbent assays (ELISAs) have been used to monitor microcystin concentrations in water and in extracted bloom samples. However, ELISAs do not differentiate between free microcystin and its conjugates or between different congeners of microcystins. Hence, many analyses conducted using ELISA are susceptible to generation of false positive results for microcystins. Other cyanotoxins, anatoxins and cylindrospermopsin, are structurally and biologically different to MC-LR, so would require separate ELISA tests. In response, many laboratories have turned to liquid chromatography- tandem mass spectrometry (LC-MS/MS) for the concurrent analysis of a range of cyanotoxins. This application note describes a method for the analysis of drinking and surface waters for 1 well-known cyanotoxins; cylindrospermopsin (CYL), anatoxin-a (ANA) and microcystins nodularin (NOD), microcystin-lr (MC-LR), MC-DeRR, MC-RR, MC-YR, MC-LY, MC-LW, and MC-LF. The method uses direct injection of 5 µl of water sample with Waters ACQUITY UPLC I-Class System coupled to Xevo TQ-S. EXPERIMENTAL UPLC conditions Drinking and surface water samples were injected as supplied without any further preparation. UPLC system: ACQUITY UPLC I-Class with FTN autosampler Column: ACQUITY UPLC BEH C µm, 2.1 x mm MS conditions MS system: Xevo TQ-S Source: Electrospray Ionization mode: ESI+ Capillary voltage: 3. kv Desolvation temp.: 6 C Mobile phase A:.1 formic acid (aq.) Desolvation gas flow: L/Hr Mobile phase B: acetonitrile Source temp.: 15 C Flow rate:.4 ml/min Cone gas flow: 15 L/Hr Injection volume: 5 µl Collision gas flow:.15 ml/min Column temp.: 5 C Nebulizer gas pressure: 7 Bar Sample temp.: 1 C Run time: 7.5 min Time (min) A B Curve ACQUITY UPLC I-Class System with the Xevo TQ-S. The Analysis of Cyanotoxins, Including Microcystins, in Drinking and Surface Waters [ 111 ]

114 [ APPLICATION NOTE ] Data were was acquired using MassLynx MS Software and processed using TargetLynx XS Application Manager. The selection of MRM transitions and optimization of critical parameters was performed by infusion of individual solutions of all the analytes and evaluation of the data by IntelliStart Software to automatically create acquisition and processing methods. Two MRM transition were used, unless otherwise stated. Table 1 summarizes conditions for all MRM transitions including the retention times. The optimum dwell time was set automatically using the Autodwell function based upon 4s wide peaks and 16 data points per peak. Table 1. MRM parameters for cyanotoxins (quantatitive transitions in bold). Compound Retention time (min) MRM Cone (V) Collision energy (ev) CYL > > ANA > > MC-DeRR > > MC-RR > > NOD > > MC-YR > > MC-LR > > MC-LY > > MC-LW > > MC-LF > > RESULTS AND DISCUSSION Excellent sensitivity and selectivity was demonstrated by the response for each analyte peaks detected from the analysis of water (as shown in Figure 2). No interfering compounds were detected at the retention times of the analytes in all the tested blank samples. All compounds can be easily detected at the 25 ng/l level, well below the WHO provisional guideline value for MC-LR of 1 µg/l and demonstrates the method would be suitable for determination at the lowest of the concentration required to meet guidance/action levels for drinking water adopted by four states in the U.S. ( ng/l), without the need for on-line enrichment. Linearity in drinking and surface water was evaluated over the concentration range 1 to ng/l for each of the analytes. Figure 3 shows a selection of calibration graphs, each an overlay of 2 calibration curves, bracketing the injections of the spiked water samples. The coefficients of determination were satisfactory (r2>.99) and the residuals <2. [ 112 ] The Analysis of Cyanotoxins, Including Microcystins, in Drinking and Surface Waters

115 Short-term repeatability was evaluated from replicate injections of drinking water (DW) and surface water (SW), previously spiked at ng/l. Precision, as measured from the peak areas of replicate (n=5) injections varied from.5 to 5. RSD. The method's accuracy was evaluated by measuring the concentration of eight of the cyanotoxins in drinking water and surface water samples, spiked at, 2, and 4 ng/l five times, using a bracketed calibration series of standards prepared in drinking water. The accuracy of the concentrations measured in both sets of water samples was poor. Measured values were consistently underestimated in drinking water and highly variable in the surface water, indicating a presence of ion suppression due to matrix effects. Such matrix effects can be compensated for by the use of internal standards, but since stable isotope analogues for microcystins are not widely available, and those that do exist are very expensive, standard addition offers an alternative means of quantification. The standard addition methodology was used for the quantification of cyanotoxins in recreational lake water. Standard addition is a type of quantitative analysis whereby known amounts of standard are added directly to test portions of each water sample to be analyzed. This can be done in an automated manner using the Auto Addition feature in the ACQUITY UPLC Sample Manager driver, which provides the option to add known amounts of cyanotoxin standard solution to vials of the sample. Data from the analysis of the sample and test portions of the sample spiked with known amounts of standard solution was processed to provide a concentration of any cyanotoxins in the sample. The standard addition calculations were carried out automatically in the TargetLynx XS Application Manager. Figures 4 and 5 show the selection of key parameters in the MassLynx sample list and in the TargetLynx 2 MC-LF MC-LW MC-LY MC-LR MC-YR NOD MC-RR MC-DeRR ANA CYL Figure 2. Chromatograms showing cyanotoxins from analysis of a standard prepared at 25 ng/l in drinking water (quantitative transition only see Table 1). Residual Response Residual Response MC-LR - Conc NOD Figure 3. Calibration graphs for a selection of cyanotoxins prepared in drinking water. Conc Conc - Conc Residual Response Residual Response CYL - Conc MC-LR Conc - Conc XS method. In the sample list, select Standard for sample type of all test portions of the sample. Add the spiked concentration in the concentration column. Finally assign a sample group for each set of samples. In the TargetLynx XS processing method, check the Use Standard Addition option. As a result of the processing of the samples, a TargetLynx XS dataset is created, displaying the calculated concentration in the summary table, as well as the standard addition calibration curve (Figure 6). The concentration of MC-LR in the PT sample, calculated by standard addition, was 724 ng/l, which was in good agreement with the assigned value of 72 ng/l. Time Conc Figure 4. Parameters for Standard Addition in the MassLynx sample list. The Analysis of Cyanotoxins, Including Microcystins, in Drinking and Surface Waters [ 113 ]

116 [ APPLICATION NOTE ] Concentration of MC-LR in PT sample = 724 ng/l Figure 5. Parameters for Standard Addition in the Targetlynx XS method. Figure 6. Standard Addition results in TargetLynx XS showing the calculated concentration in the table, calibration graph, and chromatograms (quantitative transition only). CONCLUSIONS This application note describes the performance of a method for the analysis of 1 cyanotoxins, including important microcystins, by UPLC-MS/MS, using direct injection on an ACQUITY UPLC I-Class System coupled to Xevo TQ-S. This method provides fast and reliable quantitation of the presence of cyanotoxins in water samples using standard addition, without the need for extraction, at concentrations well below the WHO provisional guideline value for MC-LR of 1 µg/l. The response was stable for up to 15 hours of analysis during the validation study, providing good repeatability (RSD <5). Calibration characteristics, linearity and residuals, were excellent over the concentration range studied. To overcome matrix effects, in the absence of internal standards, standard addition provided an accurate means for quantitation of microcystins in lake water, as demonstrated by successful determination of MC-LR in an independent Aquacheck proficiency test lake water sample. References 1. Schmidt JR, Wilhelm SW, Boyer GL. The Fate of Microcystins in the Environment and Challenges for Monitoring. Toxins 214 6: Testai E, Buratti FM, Funari, E et al., 216. Review and analysis of occurrence, exposure and toxicity of cyanobacteria toxins in food. EFSA Supporting Publication 216; 13(2):EN 998, 39 pp. 3. WHO (23) Cyanobacterial toxins: Microcystin-LR in drinking-water. Background document for preparation of WHO Guidelines for drinking-water quality. Geneva, World Health Organization (WHO/SDE/WSH/3.4/57). ACKNOWLEDGEMENTS Waters gratefully acknowledges the Watergroep, Heverlee, Belgium. Waters, Xevo, ACQUITY UPLC, MassLynx, and The Science of What's Possible are registered trademarks of Waters Corporation. TargetLynx and IntelliStart are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. April EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 114 ] The Analysis of Cyanotoxins, Including Microcystins, in Drinking and Surface Waters

117 [ FOOD PROFILING/RESEARCH ] [ 115 ]

118 [ FOOD PROFILING/RESEARCH ] [ 116 ]

119 [ APPLICATION NOTE ] Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector Kari Organtini, Naren Meruva, and Gareth Cleland Waters Corporation, Milford, MA, USA APPLICATION BENEFITS Direct analysis of food flavors using mass spectrometry detection Real-time screening of volatile and semi-volatile compounds Little to no sample preparation No chromatography required Ease of use INTRODUCTION Characterization of flavors in food products is critical for a consumer s perception of aroma and taste, food quality, and product branding. Maintaining consistent flavor in food products requires quality control testing to ensure the right flavor is applied to the product at the right levels and that the product is free from contamination and off-odors. Analysis of flavor compounds is challenging and time consuming as it requires several lab-based measurements. Based on the flavor source (e.g. natural versus synthetic), flavor formulations can contain several hundred compounds of varying solubility, volatility, concentrations, and stability. A combination of gas chromatography (GC) and liquid chromatography (LC) based methods integrated with mass spectrometry (MS) are typically applied for the identification and quantification of food flavors. These procedures require isolation and concentration of flavor compounds from their food matrix prior to instrument analysis. This application note demonstrates the utility of DART (direct analysis in real time) and ACQUITY QDa mass detection for rapid, accurate, and cost-effective flavor characterization, for monitoring product quality, and reducing production costs. DART-MS was applied to characterize volatile and semi-volatile flavors in whiskey, chewing gum, and tobacco samples with different characteristic flavors. The flexibility of sample introduction using an ambient ionization technique like DART combined with the sensitivity of the ACQUITY QDa Detector reduce analysis times, and increases confidence in results compared to conventional methods, making this system more suitable for quality control applications in the food, beverage, and consumer goods industries. 1-3 WATERS SOLUTIONS ACQUITY QDa Detector MassLynx MS Software KEYWORDS Flavors, aroma, volatiles, food QC, food authenticity, tobacco, DART, mass detection, ambient ionization Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector [ 117 ]

120 [ APPLICATION NOTE ] EXPERIMENTAL Table 1. DART and mass detection method parameters used to acquire data for whiskey, chewing gums, and tobacco samples. DART method Whiskey Gum Tobacco Ion mode Positive Positive Positive Run temp. ( C) Sampling speed (mm/sec) Exit grid voltage (V) NA NA QDa method Whiskey Gum Tobacco Ion mode Cone voltage (V) Mass range (amu) Sampling frequency (Hz) to 5 5 to 5 5 to Sample analysis Whiskey, chewing gum, and chewing tobacco samples were purchased commercially. No sample preparation was performed on the samples prior to analysis with the DART-MS system. Three different sampling techniques were used for the sample types (Figures 1A 1C). Chewing gum with four different flavors (original mint, spearmint, tropical twist, and cinnamon) were evaluated in this study. The chewing gum was sampled as a solid by holding a piece of gum with tweezers in the path of the heated ionizing helium beam generated by the DART source (Figure 1A). Whiskey samples (Scotch, Bourbon, Canadian, and Irish) were sampled as liquids by spotting 5 µl of whiskey onto Quickstrip cards that were moved through the helium beam in an automated fashion (Figure 1B). Chewing tobacco samples with four different flavors (Natural, Classic Mint, Classic Straight, and Classic Wintergreen) were investigated by DART-MS. The aroma sampling (Figure 1C) of chewing tobacco was performed by analyzing the volatiles in the headspace above the sample. In this case, the open chewing tobacco can was placed directly underneath the flow of the helium beam. (A) Solid sampling (B) Liquid sampling (C) Aroma sampling Sampling cone DART source (D) Top down view QDa Mass Detector Flavored sample Figure 1. DART-MS sampling options: (A) Solid sampling; (B) Liquid sampling using Quickstrip; (C) Aroma sampling; and (D) Top-down view of the ceramic tube pulling ions into the ACQUITY QDa. [ 118 ] Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector

121 RESULTS AND DISCUSSION METHOD DEVELOPMENT AND OPTIMIZATION For operating DART and the ACQUITY QDa, a few important parameters must be optimized for each sample type analyzed, including cone voltage and DART sampling temperature. When optimizing the cone voltage, it must be decided whether to use a low cone voltage (~5 to 1 V), or a high cone voltage (~3 V). Using a high cone voltage can induce fragmentation of the compounds which can be beneficial for generating unique mass profiles of the samples. The effects of high and low cone voltages are demonstrated in the whiskey samples shown in Figure 2. Sampling temperature is also a very important parameter that must be optimized for each sample type. Compound desorption and ionization efficiency are dependent on this parameter. Figure 3 shows the mass spectra collected from the same whiskey samples at four different temperatures. Some ions are only present at certain temperatures while the intensities of the other ions vary dependent upon temperature. CV 3 CV 1 Figure 2. Mass spectrum of whiskey sampled at a cone voltage of (top) 3 V and (bottom) 1 V. 4 C 35 C 3 C 25 C Figure 3. Whiskey mass profiles at different sampling temperatures ranging from (top) 4 C to (bottom) 25 C. Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector [ 119 ]

122 [ APPLICATION NOTE ] WHISKEY ANALYSIS Different brands of whiskey were tested using optimized DART-MS conditions. Whiskey flavor is heavily influenced by the fermentation and aging process during production. Grain type, barrel type/char, and length of aging all contribute to making each brand distinct. Figure 4 demonstrates the unique fingerprint generated by DART-MS for four different brands of whiskey. Some ions are common among a few of the samples (m/z 342, 198, 127, 97), but vary by relative intensity and ratio. Other ions are unique to one particular sample type (m/z 29 and 249 in bourbon). The bourbon sample has the most unique flavor profile of the four whiskey samples. The nominal mass was used for tentative identifications of some of these ions for characterizing the flavor profiles. The bourbon sample contained a very significant peak at m/z 29, that may be attributed to sinapaldehyde, a breakdown product of lignin resulting from the wooden aging barrel. The Scotch blend, Canadian, and Irish whiskies all contain an ion at m/z of 145. This ion could be attributed to ethyl hexanoate, which imparts a fruity odor. This DART-MS analysis method cannot be used to identify and confirm unknown flavor compounds, but it can be used as a tool to rapidly identify differences between samples and sample types. To confirm the flavor identifications indicated in the DART-MS data, reference flavor standards and conventional methods for flavor analysis should be used. Scotch Blend Bourbon Canadian Irish Figure 4. Mass spectral profile of four different brands of whiskey analyzed using DART-MS. CHEWING GUM ANALYSIS The addition of the proper flavors and sweeteners to the gum base is a critical step in the manufacturing of high quality chewing gums with long lasting flavor. Original mint, spearmint, tropical fruit, and cinnamon are the most preferred flavors by consumers today. These four flavored chewing gums of the same brand were directly analyzed by holding the gum samples in between the DART source and the ACQUITY QDa inlet. [ 12 ] Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector

123 As observed in Figure 5, DART-MS generated a unique mass spectral fingerprint for the four different flavored chewing gum products. The peaks at m/z 159, 236, and 237 were common to all of the chewing gum samples and may be attributed to the gum matrix. The cinnamon mass profile shows the presence of its characteristic flavor ingredient, cinnamaldehyde (m/z 133). Ions 172, 212, and 27 represent cooling agents that are typically used in tropical fruit, spearmint, and original mint gums. The mass profile of spearmint flavored gum shows the presence of characteristic flavor, carvone (m/z 15). Methyl salicylate (m/z 153) was also observed in both the spearmint and original gum samples. Cinnamon Tropical Fruit Spearmint Original Mint Figure 5. Mass spectral profile of four different flavors of chewing gum analyzed using DART-MS. CHEWING TOBACCO ANALYSIS Open tobacco cans were placed between the DART source and the ACQUITY QDa inlet for direct analysis of headspace volatiles. Figure 6 demonstrates the mass spectral fingerprint generated by DART-MS for wintergreen, straight, mint, and natural flavored smokeless tobacco products. The nicotine peak (m/z 163, protonated molecular ion) dominates the volatile headspace of these tobacco samples irrespective of the applied flavor. Nicotine is readily volatile at room temperature (vapor pressure 5.5 Pa at 25 C) and is the base peak in the mass profiles. Ions with m/z 8, 86, 18, 132, 136, 276, and 325 were observed in all of the tobacco samples and may be attributed to the tobacco matrix. Methyl salicylate (m/z 152) is the characterizing flavor of wintergreen and its protonated molecular ion (m/z 153) was observed in the mass profile of the wintergreen tobacco product. Ethyl salicylate (m/z 167, protonated molecular ion), a characteristic flavor ingredient was detected in the straight flavored tobacco samples. The mass profile of the mint tobacco sample showed trace levels of several flavor markers indicating the use of a natural source of peppermint oil including compounds such as limonene (m/z 137), menthone (m/z 155), traces of methyl salicylate (m/z 153), and ethyl salicylate (m/z 167). No major flavor markers were identified in the natural tobacco samples which has the simplest and relatively low intensity mass profile suggesting that not many flavors had been added to this product. Wintergreen Straight Mint Natural Figure 6. Mass spectral profile of four different flavors of chewing tobacco analyzed using DART-MS. Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector [ 121 ]

124 [ APPLICATION NOTE ] CONCLUSIONS This application note demonstrates the utility of DART and ACQUITY QDa mass detection for rapid monitoring and cost effective fingerprinting of food flavors to determine product authenticity and for quality control. Three sample types were analyzed using three different sample introduction techniques for their flavor profiles. The mass spectral fingerprints for the different samples in each of the sample groups (whiskey, gum, and tobacco) were unique to each sample. For the whiskey samples, potential flavor components influencing the flavors of the types of whiskey sampled were detected. The gum and tobacco samples exhibited flavor components typically associated with the particular samples, such as methyl salicylate in spearmint flavored products. References 1. K Organtini, G Astarita, and G Cleland. Rapid screening of fatty acids in oil supplements using ambient ionization (DART) and mass detection, Waters Application Note en, K Organtini and Gareth Cleland. Rapid profiling and authentication of cinnamon samples using ambient ionization (DART) and single quadrupole mass spectrometry. Waters Application Note en, F Li. Screening of PDE-5 inhibitors in illicit dietary supplements using DART source equipped with Waters ACQUITY QDa mass detector.ion Sense Application Note, 216. The flexibility of sample introduction using an ambient ionization technique like DART combined with the ACQUITY QDa Mass Detector reduces overall analysis and decision making times in R&D and manufacturing QC environments. The elimination of the need for sample preparation also provides the benefit of having a system located in the R&D and QC labs for rapid monitoring and quality control applications in the food, beverage and consumer goods industries. Waters, ACQUITY, QDa, MassLynx, and The Science of What s Possible are trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. May EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 122 ] Real-Time Analysis of Flavors Using DART and the ACQUITY QDa Mass Detector

125 [ APPLICATION NOTE ] Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado Using APGC-HRMS Lauren Mullin, 1 Dana Krueger, 2 and Joe Romano 1 1 Waters Corporation, Milford MA, USA 2 Krueger Food Laboratories Inc., Chelmsford, MA, USA APPLICATION BENEFITS Generation of accurate mass measurements of low- and high-energy spectra across a wide m/z range aided in identification of known FAMEs and elucidation of the unknown compound. Soft ionization using APGC resulted in preservation of the molecular ion followed by fragmentation to provide more comprehensive molecular detail. Integrated elucidation tools in UNIFI Discovery Toolset facilitated the search of hundreds of online databases, facilitating elemental composition and fragment match. INTRODUCTION Avocado is a major tropical fruit with various known health benefits when consumed as part of a balanced diet. 1 Its high lipid fraction in particular contains phytosterols, tocopherols, and most notably omega fatty acids. 1 Profiling this fatty acid content is often performed via gas chromatography (GC) following derivatization into fatty acid methyl esters (FAMEs). This is done to reduce the fatty acid polarity and neutralize the carboxyl functionality so that structural differences can be exploited via chromatography. 2 Chromatographic resolution is sufficient and reproducible enough for various FAMEs to be identified by retention time alone, and some methods will implement the use of non-mass selective detection such as flame ionization detection (FID). However the appearance of additional unexpected constituents in the derivitized lipid fraction can complicate the chromatogram. Identification of unknowns in foodstuffs is a major challenge for food companies who need to determine whether an unknown compound may be harmful to consumers. In this application note, we describe a FAMEs analysis of an avocado sample that contained a previously unknown constituent resulting from an unexpected chromatographic peak. Through the use of high resolution mass spectrometry (HRMS) and atmospheric pressure chemical ionization following gas chromatography (APGC), the accurate mass of the molecular ion and generated fragments were used to propose an identification of the compound. WATERS SOLUTIONS Xevo G2-XS QTof UNIFI Scientific Information System Atmospheric Pressure Gas Chromatography (APGC) KEYWORDS Avocado, fatty acids, FAMEs, HRMS, APGC, Universal Source Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado [ 123 ]

126 [ APPLICATION NOTE ] EXPERIMENTAL GC conditions GC system: A789 Column: DB-5MS 3 m x.25 mm.25 µm (J&W) Injection mode: Splitless Liner: Gooseneck Splitless, deactivated (Restek) Column pneumatics: Constant flow Column flow: 1.2 ml/min Injector temp.: 28 GC oven temp. ramp: Temp. Temp. ramp Hold time C 4 min 24 C 5 C/min 15 min Total run time: 47 min MS conditions MS system: Xevo G2-XS QTof Ionization mode: API + An avocado extract (NIST, Gaithersburg, MD, described below) was diluted 1: (v/v) in heptanes and 1 µl was injected for analysis on the Waters APGC-Xevo G2-XS QTof System. For the analysis of relatively polar compounds, as described in this work, a vial of water was placed into the enclosed ionization source in order to generate protons and yield the preferential formation of [M+H] + ions upon vaporization (Figure 1). Ionization is achieved through coronal discharge in the APGC source. Alternating states of low and elevated collision energy occurred during the chromatographic run, and the resulting accurate mass spectrum is time-aligned for precursor and fragment ions (MS E ). Both spectra were used in the identification of known FAMEs and structural elucidation of the unknown peak. A search list in the analysis method of known FAMEs was generated by importing structures obtained on ChemSpider ( for the following compounds: C18:2 methyl lineoleate, C18:1 trans-methyl elaidate, C18:1 cismethyl oleate, C16:1 methyl palmitoleate, and C16: methyl palmitate. Structural import provided the UNIFI Scientific Library search list with the resulting molecular formula and exact mass, which was automatically used to generate extracted ion chromatograms, as well as propose fragment structures from the high collision energy data. Acquisition mode: Acquisition range: Low collision energy: MS E 5 to m/z 6 ev High collision energy: 2 to 5 ev Source temp.: 15 C Interface temp.: 31 C Corona current: 5. µa Cone voltage: 3 V Cone gas: 11 L/hr Auxillary gas: 3 L/hr Make-up gas: 3 L/hr Figure 1. Schematic of protonation as it occurs with the APGC Source. Sample description The fat from the samples was prepared by acid hydrolysis Mojonnier extraction according to AOAC The recovered fat was converted to FAMEs with NaOH/MeOH/ BF3 according to AOAC The final FAME extract was diluted in heptanes. Data management UNIFI Scientific Information System [ 124 ] Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado

127 RESULTS AND DISCUSSION IDENTIFICATION OF KNOWN FAMEs In order to initially assess the sample and effectiveness of the analytical approach, previously identified FAMEs (named in the EXPERIMENTAL section) from a GC-FID analysis of the sample were identified. Identification of the 5 FAMEs was based on exact mass measurements within +/- 3 ppm mass error. Figure 2 shows the results from the search list as a table (top) summarizing key pieces of data for each compound, such as retention time, mass error, and adducts observed in the spectrum. Adducts monitored included the predominant [M+H] + ion, as well as the loss of an electron (displayed as e) which results from a charge transfer reaction, and the lesser observed [M+N] + ion. Due to the collection of data independent spectra, the full molecular character can be captured without intervention in the acquisition method; adducts are specified in the processing method which can be revised at any time following analysis. For each compound, the extracted ion chromatogram (XIC) as well as low and high energy spectra can be seen for the predominant ion observed. Structures were also used to interrogate the high energy spectrum (the bottom spectrum shown) producing likely structural fragmentation matches from the high energy spectra. Figure 2. Summary of all identifications based on the target list, screened against the full spectrum data that were acquired. Below the table view, the XIC, low-energy spectrum, and high-energy spectrum for the selected component, C18:1 cis-methyl oleate, are shown. One of the proposed fragment structures is displayed in the spectrum view above the m/z peak. Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado [ 125 ]

128 [ APPLICATION NOTE ] Further interrogation of the fragmentation of the FAMEs in the high energy spectra yielded some key structural characteristics. Figure 3 shows the high energy spectrum for C18:1 cis-methyl oleate in an expanded and annotated view. The presence of the traditionally observed (i.e. as seen using electron ionization GC-MS) loss of the terminal ether group is seen at m/z (mass error -.2 mda, or -.89 ppm). To the left of this fragment, we see a collection of alkyl chain breakages. Within these breakages, clusters spanning ~6 m/z indicate the process of hydride abstraction, 3 where the chain will lose two subsequent hydrogen atoms. In combining all of these fragments, the complete structure of the compound can be further confirmed. Close-up of high energy fragmentation spectra for C18 cis-methyl oleate Fragmentation of alkyl chain and hydride abstraction around chain pieces Loss of terminal ester group (parent-32 Da; m/z ), as seen in NIST Library Near elimination of parent molecule (m/z ) Figure 3. Accurate mass fragments observed for C18:1 cis-methyl oleate. STRUCTURAL ELUCIDATION OF UKNOWN PEAK Following the retention order of the known FAMEs, an unknown prominent peak was observed in the BPI (base peak intensity) chromatogram, eluting after the known FAMEs. Investigation of the spectrum under this peak resulted in a proposed molecular mass of m/z. Figure 4 shows the overlaid XIC of this mass, relative to the other FAMEs. C16: methyl palmitate C16:1 methyl palmitoleate C18:1 methyl palmitoleate/ trans-methyl elaidate C18:2 methyl lineoleate? Figure 4. Known FAMEs and an unknown peak at minutes. [ 126 ] Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado

129 Structural elucidation of the unknown peak was initialized with UNIFI by selecting the parent ion mass from the list of m/z from which an assignment was not made against the screening list of FAMEs (Figure 5). The candidate mass of interest ( m/z) was selected, and following a right-mouse click, and the option to Elucidate was chosen, launching the Discovery Toolset. The Discovery Toolset within UNIFI utilizes a combination of elemental composition calculation, isotope fidelity computational scoring, ChemSpider database searching, and fragment matching of high energy data. Figure 5. Candidate mass at m/z as it appears prior to structural elucidation. Lock mass correction has already been applied during the initial automated data processing, allowing for elucidation tools in the Discovery Toolset to be immediately applied. For the unknown peak, the top result for a proposed identification was a 3-pentadecenylphenol species, shown in Figure 6. The proposed structure was found to have a high number of fragment formulas suggested, further supporting the likelihood of this identification. Further interrogation of the high energy spectra (Figure 7) shows similar alkyl chain breakage and hydride abstraction, as seen with the FAMEs the proposed structure also contains an alkyl chain tail, increasing confidence in the identification. However, the ester loss was seen for C18:1 cis-methyl oleate (Figure 3). Instead, the loss of H2O (parent-18 m/z; m/z) can be seen off of the ring portion of the structure, and most importantly, the intact benzene ring itself is apparent as a fragment (Figure 7). Online searching of this compound against FooDB ( the world s largest compendium of food related compounds, showed that this compound has been found in the lipid portion of cashew nuts and Gingko biloba fruit (Figure 8). 4 Although not indicated to be in avocado explicitly, the nature of the compound identification with highly confirmed spectral matching, combined with the presence of this compound in lipid portions of other fruits supported a tentative match. Assignment of the previously unknown mass with the name from ChemSpider was performed within UNIFI by selecting Assign below the ChemSpider result view (Figure 6). Though not commercially available at the time of this writing, final confirmation would typically be performed through the analysis of a pure chemical standard. Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado [ 127 ]

130 [ APPLICATION NOTE ] Figure 5. Candidate mass at m/z as it appears prior to structural elucidation. Lock mass correction has already been applied during the initial automated data processing, allowing for the elucidation tools in the Discovery Toolset to be immediately applied. Figure 6. Discovery Toolset results for a proposed structure as found based on elemental composition of spectral peak at , ChemSpider search of this composition, and comparison of high energy spectra to possible fragments from the structure. Blue molecule icons in spectral view (high energy spectra only) indicate a fragment match for that mass. A total of 63 proposed fragments were found for the spectra against this structure. [ 128 ] Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado

131 Intact benzene ring Loss of H 2 O (parent-18 Da; m/z ) Remaining parent molecule (m/z ) Figure 7. High energy spectrum close-up for unknown peak with proposed structure and fragment assignments. Figure 8. FooDB search result for 3[(8Z)-8-pentadecen-1-yl]phenol. 4 Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado [ 129 ]

132 [ APPLICATION NOTE ] CONCLUSIONS A lipid extract of avocado was found to contain an unexpected and unknown component. The extract was subsequently analyzed for both known FAMEs and for additional compounds. The use of full spectral acquisition, with alternating collision energy states (MS E ), afforded the ability to both interrogate the data for known compounds using fragment matching, as well as searching for any additional masses of interest. Elucidation of the unknowns was made possible by interrogating the high quality accurate mass data into the easily accessible UNIFI Discovery Toolkit. A proposed identification was made for the unknown peak based on the use of the spectra and assigned within the analysis. Future profiling of avocado samples can be performed with this additional compound in mind, and any future unknown peaks can be elucidated using this same basic approach. References 1. Duarte PF, Chaves MA, Borges CD, Barboza Mendonca CR. Avocado: characteristics, health benefits and uses. Food Techn (4): html?tablepage= chap7.pdf 4. Acknowledgment The authors would like to thank Melissa Phillips of NIST for providing the avocado powder samples and for her thoughtful review of this work. Waters, UNIFI, Xevo, and The Science of What's Possible are registered trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. May EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 13 ] Structural Elucidation of an Unknown Compound in the Fatty Acid Methyl Esters (FAMEs) Extract of Avocado

133 [ APPLICATION NOTE ] Discrimination of Different Unifloral Honeys Using an Untargeted High-Definition Mass Spectrometry Metabolomic Workflow Antonietta Wallace, Joanne Connolly, Sara Stead, and Simon Hird Waters Corporation, Wilmslow, UK APPLICATION BENEFITS Differentiation of honeys of different botanical origin. Verification of the selection of markers for different botanical origin in honey by UPLC-MS/MS. WATERS SOLUTIONS ACQUITY UPLC I-Class System SYNAPT G2-Si HDMS MassLynx MS Software Progenesis QI Software Xevo TQ-S TargetLynx XS Application Manager KEYWORDS Honey, metabolomics, unifloral honey, polyfloral honey, heather honey, Manuka, Kanuka, leptosperin, principal component analysis (PCA), orthogonal projection to latent structures discriminant analysis (OPLS-DA) INTRODUCTION Food fraud is a collective term which describes a substitution, addition, alteration; or a misrepresentation, deliberate and intentional of food ingredients or of food packaging; or false or misleading statements formulated concerning a product for economic gain. 1 Recent scandals have highlighted that food fraud can also result in major food safety issues. Honey is a high value food commodity frequently subjected to fraudulent mislabeling and adulteration. The non-compliances detected during the European Commission (EC) 215 coordinated control plan were mostly related to adulteration with sugar (6), and to the declaration of the botanical source (7). Non-compliances related to the declaration of the geographical origin were less frequent (2), but were considered more difficult to detect. 2 Analytical methods used in the quality control of honey have recently been reviewed. 3 The nutraceutical properties of honey arise from specific chemical compositions, which vary according to botanical origin. These properties also confer distinct sensory profiles to various types of unifloral honeys. For these reasons, the price of unifloral honeys is higher than that of polyfloral honey. Adulteration in terms of the dilution of honeys of high value floral origin with those of lower value has increased in recent years. For example, the great commercial relevance attributed to Manuka honey, due to the promotion of its perceived increased antibacterial activity and health benefits, as well as its limited availability, has led to a high frequency of fraud. Guidance on labeling requirements and defining the characteristics of Manuka honey has been established. 4 Identification of the floral origin of honey is typically achieved by melissopalynological analysis based on pollen characterization, complemented by sensory and physico-chemical analysis. However pollen identification requires a high degree of skill, in some cases gives erroneous results and can fail to detect closely related species (e.g. Manuka and Kanuka). The simultaneous detection of multiple components using spectroscopic and spectrometric techniques, coupled with statistical analysis, is a promising approach to achieve botanical discrimination. 5 LC-HRMS is one technique that has been used extensively for metabolic profiling in the food and beverage industry 6,7,8 including for the analysis of honey. 9 Discrimination of Different Unifloral Honeys [ 131 ]

134 [ APPLICATION NOTE ] In this study we investigated whether untargeted metabolomics, using Waters UltraPerformance Liquid Chromatography (UPLC ) coupled with High Definition Mass Spectrometry (HDMS ) and multivariate statistics, could differentiate honeys of different botanical origin. HDMS combines ion mobility spectrometry with high resolution mass spectrometry to allow researchers to analyze ions differentiated by size, shape, and charge, as well as mass. The selection of markers of different botanical origin was verified using a targeted method based upon UPLC coupled with tandem quadrupole (TQ) mass spectrometry. EXPERIMENTAL Authentic samples of rape (3), heather (9), buckwheat (5), and Manuka (8) unifloral honeys were obtained from indisputable sources. Each floral class contained separate samples from different countries (Norway, Denmark, Lithuania, Poland, and New Zealand) and different years of collection giving multiple biological replicates per floral class. Honey samples were subjected to minimal sample preparation: honey (.5 g) was diluted (1 ml methanol/1 formic acid in water, 5:5 v/v), shaken, sonicated (2 min) and centrifuged. Samples were analyzed in a randomized order including interspersed pooled honey samples for QC purposes. Part 1: UPLC-HDMS Samples were analyzed in triplicate by HDMS E acquisition on a SYNAPT G2-Si Mass Spectrometer using electrospray in positive and negative ion modes. HDMS E data-independent analysis provides accurate mass measurements of all detectable ions; both precursors and fragments. Chromatographic and drift time alignment of precursor and fragment ion data aids assignment of fragment ions to parent ions of similar mass or retention time. UPLC conditions UPLC system: ACQUITY UPLC I-Class with FTN autosampler Column: ACQUITY UPLC BEH C18, 1.7 µm, 2.1 mm Mobile phase A: Mobile phase B: Flow rate: 1 mm Ammonium acetate (aq.) Acetonitrile.5 ml/min Injection volume: 5 and 9 µl Column temp.: 45 C Sample temp.: 5 C Run time: 1 2 min Time (min)a B Curve MS conditions MS system: Acquisition mode: SYNAPT G2-Si HDMS Acquisition mass range: 5 to 12 m/z Scan time: Lockmass: ESI+, ESI-; resolution mode; HDMS E.1 s Collision energy ramp: 15 to 55 ev Capillary voltage: m/z (Leucine enkephalin) 3.1 kv Desolvation temp.: 6 C Desolvation gas flow: 8 L/Hr Source temp.: 13 C Cone voltage: Cone gas flow: Nebulizer gas pressure: Drift gas: IMS wave velocity range: IMS wave height: 3 V 35 L/Hr 5 Bar N2 65 m/s 4 V IMS gas flow: 9 ml/min Acquisition software: MassLynx v4.1 Bioinformatics software: Progenesis QI (2.2) with EZinfo (MKS Data Analytics Solutions, Sweden) and access to various databases: HMDB, Metlin, and MassBank. [ 132 ] Discrimination of Different Unifloral Honeys

135 Part 2: UPLC-MS/MS UPLC conditions UPLC system: ACQUITY UPLC I-Class with FTN autosampler Column: ACQUITY UPLC CSH Phenylhexyl 1.7 µm, 2.1 mm Mobile phase A:.1 formic acid (aq.) Mobile phase B: Acetonitrile Flow rate:.4 ml/min Injection volume: 5 µl Column temp.: 4 C Sample temp.: 15 C Run time: 9 min Time (min) A B Curve MS conditions MS system: Xevo TQ-S Acquisition mode: ESI+ Capillary voltage: 2. kv Desolvation temp.: 5 C Desolvation gas flow: L/Hr Source temp.: 15 C Cone gas flow: 15 L/Hr Collision gas flow:.15 ml/min Nebulizer gas pressure: 7 Bar MRM transitions 581>211 and 581>323 for leptosperin: [from Kato et al. (214)] Cone voltage: 34 V Collision energy: 16 and 22eV, respectively (optimized using a Manuka honey sample) Dwell times: 35 ms Acquisition software: MassLynx v.4.1 Processing software: TargetLynx XS Application Manager RESULTS AND DISCUSSION PART 1: UPLC-HDMS Although the base peak intensity chromatogram (BPI) chromatogram of the pooled QC honey sample from the negative ion, low energy, HDMS E experiment (Figure 1) shows a number of peaks, it fails to illustrate the complexity of the honey sample extracts. The inset shows the co-elution of numerous metabolites, which hinders their identification. A more in-depth assessment of the data is required. ESI negative ion mode Low energy m/z Time Figure 1. Typical base peak intensity (BPI) chromatogram from the analysis of the pooled QC honey sample by UPLC-MS (ESI negative ion, low energy, HDMS E data) and the 3D view of a subset of the data. The two data sets acquired in ESI positive and in ESI negative modes were imported into Progenesis QI Data Analysis Software to search for and identify unique markers of botanical origin. With this software, significant differences between different unifloral honeys were observed by comparing multiple samples, employing a user-friendly workflow approach. Time Discrimination of Different Unifloral Honeys [ 133 ]

136 [ APPLICATION NOTE ] After alignment, peak detection and deconvolution in Progenesis QI, principal component analysis (PCA) was used for initial exploration of the 9 compounds extracted from the ESI negative ion HDMS E data in order to determine whether there were any outliers in the data, and also see how well the samples were grouped. No attempt is made to apply any classification of the data into different unifloral honeys at this point; PCA enables variances within the dataset to be assessed to see whether sample-to-sample variation is higher than any differences between the classes. After export into EZInfo, PCA was carried out after Pareto scaling. Scaling means shrinking or stretching variance of individual variables to ensure the score values are not so small compared to the loadings that they are not visible in a plot. Pareto scaling is commonly used to reduce the influence of intense peaks while emphasizing weaker peaks that may have more significance. The corresponding loadings of intense signals will be reduced and loadings from weak signals will be increased due to Pareto scaling. The PCA scores plot generated in EZInfo from the ESI negative ion HDMS E data for unifloral honeys showed separation into distinctive clusters of pooled QC, buckwheat, heather, rape, and Manuka (Figure 2). All pooled QC samples were found to be tightly located within the center of the PCA scores plot which indicates good reproducibility of the method and the absence of any bias introduced during the processing of the data. Buckwheat Heather Manuka Pool Rape Figure 2. Principal component analysis (PCA) scores plot from EZInfo (ESI negative ion HDMS E data). In order to help better understand the separation between types of unifloral honey and to identify potential characteristic markers for each unifloral honey, supervised multivariate analysis was performed using orthogonal projection to latent structures discriminant analysis (OPLS-DA). OPLS-DA is a supervised technique where the compound ions are classified into the available groups using regression and prediction methods. The advantage of OPLS-DA is that it shows which variables responsible for class discrimination and so is easier to interpret the results. The OPLS-DA scores plots generated from the ESI negative ion HDMS E data for the unifloral honeys (Figure 3) shows the classification of the compound ions into the honey groups; pooled QC, buckwheat, heather, rape, and Manuka. Buckwheat Heather Manuka Pool Rape Figure 3. OPLS-DA scores plot from EZInfo (ESI negative ion HDMS E data). [ 134 ] Discrimination of Different Unifloral Honeys

137 An S-plot can also be created, as shown in Figure 4, to quickly highlight those features responsible for the difference between pairs of unifloral honeys with the highest confidence and contribution (area highlighted by red line in Figure 4). Selected features from the S-plot can then be imported back into Progenesis QI and tagged, allowing them to be filtered and viewed independently. Figure 4 shows the S-plot for Manuka and heather honeys from the ESI negative ion HDMS E data. It is also possible to compare a group of interest (e.g. Manuka) with all other honey samples by creating a new group of all other unifloral honeys. The most significant markers from the S-plot were tagged and the data imported data back into Progenesis QI for verification and further evaluation (e.g. identification). Manuka markers Confidence Importance Heather markers Figure 4. S-plot from EZInfo showing a comparison of Manuka and heather honeys (ESI negative ion HDMS E data). After importing the data for the tagged markers from EZinfo back into Progenesis QI, the data was filtered to verify the selection of markers prior to their identification. Markers were selected for further investigation that showed a significant difference (Anova p-value <.1) and a 5 fold or greater increase in abundance for one of the types of unifloral honeys. Using the Progenesis QI search engine, Progenesis MetaScope, queries were made of publicly available databases (Figure 5). Search parameters were customized to maximize all aspects of the data acquired to the database being searched. Possible identifications for each marker were ranked on an overall score based on: mass error, isotope similarity (calculated from the comparison of the measured isotope distribution for the compound versus the theoretical based on the compound formula), and fragmentation score. To improve confidence in the compound identification, theoretical fragmentation of the candidate list of compounds was performed, and the resulting in silico fragmentation matched against the measured/observed fragments for a compound. High spectral specificity (spectral cleanup) was observed due to the ability of the data analysis software to time align and drift align spectra from the four dimensional HDMS E data. Discrimination of Different Unifloral Honeys [ 135 ]

138 [ APPLICATION NOTE ] Identifier entry from database searched Figure 5. An example of a tentative identification of a marker compound from the database search. Some of the metabolites identified were highlighted as being able to differentiate the Manuka honey samples, which have been reported as Manuka markers previously. 1 The standardized abundance profiles for three of these marker compounds are represented graphically in Figure 6. This provides information about how the compound abundances are changing across the different types of unifloral honey and information about the reliability of the changes seen. Identifications assigned, along with statistical values, are shown in the accompanying table. All other uniflorals Manuka Methyl syringate Leptosin Tentative ID m/z Formula Mass error (ppm) RT (mins) Ion P value Fold change (mean) Methyl syringate C 1 H 12 O [M-H] - <1.1E Leptosin C 22 H 32 O [M-H] - <1.1E Leptosperin C 23 H 34 O [M-H] - <1.1E-16 1 Figure 6. Review of standardized abundance profiles and assignment of identity for three markers of Manuka honey as displayed in Progenesis QI Software. [ 136 ] Discrimination of Different Unifloral Honeys

139 PART 2: UPLC-MS/MS Independent verification of one of the markers of Manuka honey, leptosperin, which was identified from the UPLC-HDMS data, was afforded by targeted UPLC-MS/MS in MRM mode. Figure 7 shows MRM chromatograms illustrating the detection of leptosperin in a sample of Manuka honey that is absent from the sample of heather honey. Manuka honey Heather honey 581> > > > Time Time Figure 7. MRM chromatograms from the determination of leptosperin in samples of Manuka and heather honeys. Discrimination of Different Unifloral Honeys [ 137 ]

140 [ APPLICATION NOTE ] CONCLUSIONS The metabolomics workflow is emerging as a powerful approach for the discovery of biomarkers to tackle food fraud. UPLC provides high efficiency separations and comprehensive, unbiased HDMS E acquisition provides information-rich data including accurate mass and isotope distribution for precursor and fragment ions. The addition of ion mobility offers increased peak capacity, separation of isomers and spectral cleanup. The Progenesis QI workflow provides an easy-to-use, scalable system for analysis of food metabolomic data including accurate peak alignment and peak picking, classification of samples using multivariate statistical analysis, quantification of relative abundance of markers for each class and identification of markers from database searches supported by structural elucidation tools. Acknowledgements Waters kindly acknowledges Fera Science Ltd for analysis on the Xevo TQ-S. References 1. Spink and Moyer. Defining the public health threat of food fraud. J Food Sci. 76 (9): R157 R163, docs/official-controls_food-fraud_honey_controlplan-results.pdf 3. Pita-Calvo et al. Analytical methods used in the quality control of honey. J Agric. Food Chem. 65: 69 73, manuka-honey/ 5. Spiteri et al. Combination of 1 H NMR and chemometrics to discriminate manuka honey from other floral honey types from Oceania. Food Chem. 217: , Jandric et al. Assessment of fruit juice authenticity using UPLC QToF MS: A metabolomics approach. Food Chem. 148: 7 17, Dai et al. Nontargeted analysis using ultraperformance liquid chromatographyquadrupole time-of-flight mass spectrometry uncovers the effects of harvest season on the metabolites and taste quality of tea (Camellia sinensis L.). J Agric. Food Chem. 63: , Black et al. A comprehensive strategy to detect the fraudulent adulteration of herbs: The oregano approach. Food Chem. 21: , Jandric et al. An investigative study on discrimination of honey of various floral and geographical origins using UPLC-QToF MS and multivariate data analysis. Food Control. 72: , Kato Y et al. (214). Plausible authentication of manuka honey and related products by measuring leptosperin with methyl syringate. J Agric. Food Chem. 62(27): 64 7, 214. Waters, ACQUITY UPLC, UPLC, SYNAPT, Xevo, Progenesis, MassLynx, High Definition Mass Spectrometry, HDMS, and The Science of What's Possible are registered trademarks of Waters Corporation. TargetLynx is a trademark of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. April EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 138 ] Discrimination of Different Unifloral Honeys

141 [ APPLICATION NOTE ] Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix Richard C. Daw Waters Corporation, Morrisville, NC, USA APPLICATION BENEFITS Easy-to-use UPLC -MS method to detect naringin and carbendazim in orange juice. Utilize mass detection to get confirmatory information about the compounds detected to aid identification. Use of multi-dimensional chromatography (trap-and-elute with at column dilution) to enrich the sample for increased sensitivity. INTRODUCTION Regulatory agencies have been testing orange juice for pesticide residues and adulteration for decades. In early 215, the U.S. FDA announced that it was investigating reports that low levels of a common fungicide, carbendazim, were being detected in imported orange juice. Naringin, a flavanone commonly found in very low levels in orange juice and high levels in grapefruit juice, is commonly tested to ensure that the orange juice is not being adulterated with grapefruit juice. At the time of writing, the testing limits for these compounds were 1 ppm for carbendazim and.1 ppm for naringin. 1 Time-consuming HPLC/UV methods are deployed in testing labs as a first pass test for these compounds. These methods are neither sensitive nor selective and therefore samples are typically re-directed to centralized MS/MS labs for further testing when any peak elutes near the expected retention time. To address the challenges for routine testing, a simplified, easy-to-use mass detection method can be used. Waters ACQUITY QDa Mass Detector is a compact single quadrupole mass spectrometer that is as intuitive and as easy to use as an optical detector. It is pre-optimized to work with most samples without the tuning or sample specific optimization necessary for most mass spectrometers. Mass detection can be deployed in QC lab environments to provide the selectivity and sensitivity necessary to detect low levels of compounds in complex matrices, such as orange juice. WATERS SOLUTIONS ACQUITY UPLC H-Class System BEH (Ethylene Bridged Hybrid) Technology ACQUITY QDa Mass Detector The sensitivity challenges of the aforementioned HPLC-UV method can benefit from multi-dimensional UPLC, which is an emerging technology that can provide increased peak capacity and increased sensitivity. One approach to multi-dimensional chromatography is to use trap-and-elute with at-column dilution to focus large amounts of material on a trap and then to elute to a chromatography column. Sample enrichment can be achieved as a result of deploying this approach in the lab. Empower 3 CDS Software KEYWORDS Naringin, carbendazim, multi-dimensional chromatography, 2D, mass detection, orange juice, trap-and-elute Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix [ 139 ]

142 [ APPLICATION NOTE ] EXPERIMENTAL Sample preparation Standards were prepared from ppm naringin in water and ppm carbendazim in methanol standard stock solutions. A matrix blank was prepared by centrifuging a sample of low pulp orange juice for 2 minutes in a bench top centrifuge and collecting the supernatant. A 1 ppm solvent standard was prepared by mixing µl of each standard stock solution with 8 µl of water and vortexing to mix. Lower concentration solvent standards were prepared by serial dilution with water. A combined 1 ppm matrix standard was prepared by mixing µl of each standard stock solution with 8 µl of matrix blank and vortexing to mix. Lower concentration matrix standards were prepared by serial dilution with matrix blank. For analysis, each solvent standard and matrix standard was diluted and filtered before injection. A µl aliquot of each standard was subsequently mixed with 3 µl of mobile phase B (1 mm ammonium acetate in methanol) and 6 µl of mobile phase A (1 mm ammonium acetate in water), vortexed and filtered through a.2-µm GHP filter (Waters Acrodisc Minispike Syringe Filter, GHP, 13-mm.2-µm, WAT97962). A 2-µL aliquot was injected using a standard 15-µL needle and -µl syringe. For injections over 1 µl, a 5-µL extension loop was with the previously described configuration. UPLC conditions (1D) UPLC System: ACQUITY UPLC H-Class with CM-A Detector: PDA and ACQUITY QDa (in series) Column: ACQUITY UPLC BEH C18, 1.7 µm, 2.1 x 15 mm (P/N: ) Mobile phase A: 1 mm Ammonium acetate in water Mobile phase B: 1 mm Ammonium acetate in methanol Seal wash: 9/1 Water/methanol Purge/needle wash: 5/5 Water/methanol Flow rate:.4 ml/min Isocratic flow: 61/39 A:B Run time: 3 min Injection volume: 2 µl Sample temp.: 4 C Column temp.: 6 C Detector: ACQUITY UPLC epda Wavelength: 254 nm Sampling rate: 1 pts/sec UPLC conditions (2D) UPLC System: ACQUITY UPLC H-Class with 2D Technology, with at-column dilution Detector: PDA and ACQUITY QDa (in series) Column: ACQUITY UPLC BEH C18, 1.7 µm, 2.1 x 15 mm (P/N: ) Trap column: XBridge C18 Direct Connect HP, 1 µm, 2.1 x 3 mm (P/N: ) QSM (1st dimension pump):.2 ml/min 1 mm Ammonium ISM (dilution pump): BSM (elution pump): acetate in water 1.8 ml/min 1 mm Ammonium acetate in water.4 ml/min, 61/39 Mobile phase A/B as above All other conditions are per the LC conditions (1D). MS conditions MS system: ACQUITY QDa Ionization modes: ESI+ (carbendazim) and ESI- (naringin) Probe temp.: 3 C Capillary voltages: 1.5kV (ESI+),.8kV (ESI-) Sampling rate: 5 pts/sec Acquisition range: to 6 m/z Target compounds Name RT ESI SIR CV SIR CV (min) mode (par) (par) (frag) (frag) Carbendazim 2.19 ESI m/z 1 V m/z 4 V Naringin 2.38 ESI m/z 25 V m/z 6 V [ 14 ] Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix

143 Multi-dimensional chromatography is possible on the ACQUITY UPLC H-Class System by using a CM-A with the optional 6-port valves. Extra dilution/elution pumps can be configured with the ACQUITY UPLC H-Class System using either Empower 3 or MassLynx Software. Trap-and-elute is a technique that allows for focusing material on a first dimension trap column before transferring to the second dimension chromatography column for analysis. This can be useful to capture sections of an existing chromatogram or to aid in sample enrichment (or both). At-column dilution (ACD) is a technique that allows for large volume injections of relatively strong sample diluents. By tee-ing in a second low organic mobile phase at a larger flow rate using a dilution pump, the slug of injected material can be easily refocused onto the head of a column (Figure 1a). When used in combination with trap-and-elute, an elution pump can be used to back-flush the focused band of material from the trap for separation on a UPLC column (Figure 1b). An XBridge C18 Direct Connect HP Column was used to trap-and-elute large volumes of naringin and carbendazim before transferring the material to the UPLC column in an attempt to improve the sensitivity of the method. 1A BSM 1B BSM Column TRAP Waste Column TRAP Waste ISM QSM/FTN ISM QSM/FTN Figure 1A. Picture of tubing layout. Figure 1B. Picture of tubing layout. For both methods, simultaneous positive and negative single ion recordings were collected at the m/z values previously described. The cone voltage was optimized for the parent base ion and then pushed to high levels to induce in-source fragmentation of each compound of interest. The response of the parent ion and the fragment are consistent from injection to injection and so were compared to aid in peak identification. Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix [ 141 ]

144 [ APPLICATION NOTE ] RESULTS AND DISCUSSION Under isocratic conditions, calibration curves were prepared for both analytes using mass detection. Solvent standards were prepared in the range of.412 ppm to 1 ppm (Table 1). Calibration curves were produced for 3.33 to 1 ppm for carbendazim (linear, R 2 =.99) and.412 to 1 ppm for naringin (quadratic, R 2 = 1.) (Figure 2). Matrix standards were prepared at.123 ppm and 1 ppm levels. Precision was tested for the solvent and matrix standards at the required limits of.123 ppm (naringin) and 1 ppm (carbendazim). Precision was found to be 1.1 (solvent) and 3.3 (matrix) for naringin and 4.7 (solvent) and 5. (matrix) for carbendazim. These results met the previously described limits. Table 1. Calibration data for solvent standards. 1 n=6 at.123 and 1. levels. 2 Naringin area uses SIR m/z 192.4, Carbendazim data uses SIR m/z Concentration (PPM) Naringin avg. area (n=3) 2 Naringin precision (n=3) 1 Naringin 16/92 Carbendazim avg. area (n=3) 2 Carbendazim precision (n=3) ND ND ND ND ND ND ND ND ND ND ND ND Carbendazim 271/ y = -5,142.67x 2 + 1,899,815.46x + 36, Concentration (PPM) R = 1. Response Naringin (SIR m/z 192.4) 5 4 Carbendazim (SIR m/z 579.7) Response y = 4,52.18x - 1, R =.99 Concentration (PPM) Figure 2. Calibration curve for solvent standards. [ 142 ] Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix

145 Mass detection of the solvent standards produced slightly different results than those observed for the matrix standards. For naringin solvent standards, a predominant peak was observed at 2.4 minutes; however, in the matrix standard, additional peaks were observed at.9 and 1. minutes. For carbendazim the differences were more pronounced. Specifically the SIR channel corresponding to the parent base ion for carbendazim (m/z 579.7) produced a single peak for the solvent standard, while an additional peak eluting just before the main peak was observed in the matrix standard (Figure 3). Naringin (SIR m/z 192.4) Solvent standards (.123 to 1 PPM) Naringin (SIR m/z 192.4) Matrix standards (.123 PPM) 3.6x x x1 6 3.x x x x x Intensity 2.x x x x x1 6 1.x1 6 Intensity x1 5 6.x1 5 4.x1 5 2.x Minutes Carbendazim (SIR m/z 579.7) Solvent standards (3.33 to 1 ppm) Minutes Carbendazim (SIR m/z 579.7) Matrix standards (1 ppm) Intensity Intensity Minutes Minutes Figure 3. Overlay of solvent and matrix standards. Ion ratios can be used to confirm the identity of peaks by comparing the response from the parent ion to a fragment ion. As previously described, different SIR channels were collected using the ACQUITY QDa Mass Detector with a higher cone voltage used to induce fragmentation of the parent ion. Using these conditions, the ion ratios of (parent, CV 1) to (fragment, CV 4) for naringin was found to be.9 (Table 1). The ratio was repeatable at the.1 ppm limit for naringin and consistent across the dynamic range for solvent and matrix standards. For carbendazim this approach produced an ion ratio of.6 for (parent, CV 25) to (fragment, CV 6) (Table 1). This ratio was also repeatable at the 1 ppm limit for carbendazim and consistent across the dynamic range for solvent and matrix standards. Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix [ 143 ]

146 [ APPLICATION NOTE ] Given the consistency of the ion ratios these values can be used to confirm identification of the analytes in a complex matrix, where other analytes with the same m/z values may be present (Figure 3). For example, overlay of the parent and fragment SIR channels for carbendazim demonstrate the differences in the ion ratios for the various peaks observed in the matrix standard. As shown in Figure 4, the carbendazim peak at 2.2 minutes has a ratio of approximately.6, consistent with the ion ratio observed for the solvent standards. An additional interference peak at 2. minutes has a ratio of >1 and the peaks between.6 and 1.2 minutes have a ratio of ~. These results provide confirmation that the peak at 2.2 minutes is carbendazim, and that the other predominant peaks are most likely components in the matrix. This approach allows for more reliable identification of known components in a complex matrix. To increase sensitivity, a variety of larger injections of solvent and matrix standards were attempted. However, larger injections often lead to strong solvents effects and poor peak shape and decreased resolution, most notable in high organic diluents. As the diluent and the mobile phase matched, the effects of the increased injection volume were muted; however the resolution between naringin and carbendazim suffered and the peak tailing increased to unacceptable levels. When At-Column dilution (ACD) was deployed, larger injections (25 µl instead of 2 µl) gave acceptable resolution and peak tailing, which extended the sensitivity of each component. The higher injection volumes were used to evaluate the dynamic range and sensitivity of the multi-dimensional approach using solvent standards with ACD and trap-and-elute. Calibration curves were produced over the range of.37 to 1 ppm for carbendazim (linear, R2 = 1.) and from.137 to 5 ppm for naringin (quadratic, R2 = 1.) (Table 2, Figure 5). Precision was improved with this approach and was found to be 1. (solvent) and.6 (matrix) for naringin, and 1.6 (solvent) and 2.3 (matrix) for carbendazim. Recovery was 95.7 for naringin and 89.7 for carbendazim (Table 3). Ion ratios matched those produced during the original isocratic run. Intensity Blue = m/z 579 Black = m/z 271 Ion ratio 579/271 = Ion ratio 579/271 = ~1.4 Ion ratio 579/271 = Minutes Figure 4. Overlay of ions. Concentration (PPM) Naringin avg. area (n=3) 1 Naringin 16/92 Carbendazim avg. area (n=3) 1 Carbendazim 271/ ND ND ND ND ND ND Table 2. Table of calibration data for solvent standards with ACD. 1 Naringin area uses SIR m/z 192.4, Carbendazim data uses SIR m/z Response Naringin SIR m/z (with ACD) y = -2,526,484.85x ,342,458.65x + 259,95.5 Concentration (PPM) R = 1. Response y = 39,729.4x - 3, R = 1. Carbendazim SIR m/z (with ACD) Concentration (PPM) Figure 5. Calibration curve for solvent standards with ACD. [ 144 ] Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix

147 Injection # Naringin RT Naringin solvent std. area (.123 PPM) Naringin matrix std. area (.123 PPM) Average RSD Injection # Carbendazim RT Naringin solvent std. area (1 PPM) Naringin matrix std. area (1 PPM) Average RSD Table 3. Table of precision solvent and matrix standards with ACD. 1 Naringin area uses SIR m/z 192.4, Carbendazim data uses SIR m/z Naringin solvent standards (.137 to 1 ppm) Naringin matrix standards (.123 ppm) x x x x x Intensity 7.x1 6 6.x1 6 5.x1 6 4.x1 6 Intensity x x1 6 1.x Minutes Minutes Carbendazim solvent standards (.37 to 1 ppm) Carbendazim matrix standards (1 ppm) Intensity Intensity Minutes Minutes Figure 6. Overlay of solvent and matrix standards with ACD. Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix [ 145 ]

148 [ APPLICATION NOTE ] Preparing samples in organic solvents is something that is appealing to food testing labs because higher organic diluents allow for reduced loss of material from poor solubility and interactions with glassware. ACD can allow injections of pure organic diluents because the dilution pump refocuses the material on the trap before eluting to the chromatographic column. Figure 7 shows an overlay of a 1 ppm matrix standard prepared in 6/4 water/meoh with 1mM ammonium acetate and another prepared in methanol. Regardless of the diluent, the same peak shape, resolution, and sensitivity is achieved. Intensity Minutes Figure 7. Overlay of matrix standards prepared with different solvents and analyzed with ACD. CONCLUSIONS An isocratic method has been developed for the analysis of naringin and carbendazim in orange juice. The ratio of parent and fragment ions were also used to confirm peak identity. Multi-dimensional techniques including at-column dilution (ACD) with trap-and-elute allowed the extension of the dynamic range to lower levels and improved precision. With at-column dilution, increased sensitivity was achieved by enabling the injection of larger volumes without loss of resolution or peak shape problems due to strong solvent effects. These results show how a greater dynamic range and improved precision can be achieved on a low cost mass detector in combination with ACD with trap-and-elute approach. References 1. US Food and Drug Administration (212), Orange Juice Products and Carbendazim: Addendum to FDA Letter to the Juice Products Association, Washington, DC). 2. M Twohig, D A Kruger, A Gledhill, J Burgess. Adulteration in Fruit Juices: A Solution to a Common Problem. (212), Waters Application Note no en. 3. Direct Injection of Orange Juice for the Rapid Detection of the Fungicide Carbendazim using ACQUITY UPLC I-Class with Xevo TQ-S. (212) Waters Technology Brief no en. Waters, QDa, ACQUITY, UPLC, XBridge, MassLynx, and Empower are registered trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 217 Waters Corporation. Produced in the U.S.A. January EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 146 ] Strategies for On-Line Sample Enrichment and Confirmation of Analytes in a Complex Food Matrix

149 [ FOOD INGREDIENTS ] [ 147 ]

150 [ FOOD INGREDIENTS ] [ 148 ]

151 [ APPLICATION NOTE ] Fast Analysis of Isoflavones in Dietary Supplements USP Method Transfer onto a UHPLC System Jinchuan Yang, Mark Benvenuti, and Gareth Cleland Waters Corporation, Milford, MA, USA APPLICATION BENEFITS Reduce analysis time of isoflavones from 74 min to 18 min. Easily transfer the standard USP isoflavones method to UHPLC with the aid of mass detection. Confirm peak identities more confidently using mass detection. INTRODUCTION Isoflavones are found primarily in plants of soy (Glycine max), red clover (Trifolium pretense), and Kudzu (Pueraria lobata). The 12 major isoflavones found in these plants are daidzein, glycitein, genistein, and their respective glucoside and malonyl- and acetyl- glucoside derivatives. The structures of 12 isoflavones and an internal standard, apigenin, are shown in Figure 1. These hormone-like compounds are often used in remedies to reduce menopausal and post-menopausal symptoms. They are even associated with low breast cancer rate in Asia and the retarded progression of Alzheimer s disease. Standard methods for isoflavones in dietary supplements have been established by organizations such as USP 1 and AOAC. 2 These methods use reversed-phase LC with C18 columns and ultraviolet and visible light (UV-Vis) spectroscopy for separation and quantitation. Because of the close structural similarity of these compounds, the chromatographic run times of these methods are typically over 7 minutes long. It is highly desirable to develop a more rapid isoflavone analysis method. WATERS SOLUTIONS ACQUITY Arc UHPLC System This application note demonstrates the transfer of the USP method onto Waters ACQUITY Arc UHPLC System. The analysis time with the ACQUITY Arc System is only 18 minutes, including column wash and equilibration. Waters ACQUITY QDa Mass Detector was used to expedite the method transfer described in this study. The benefits of mass detection in peak identification and method optimization are also highlighted. ACQUITY QDa Mass Detector 2998 Photodiode Array (PDA) Detector CORTECS C 18 Column Empower 3 CDS Software KEYWORDS Isoflavones, dietary supplement, mass detection, soy, Glycine max, red clover, Trifolium pretense, Kudzu, Pueraria lobata, daidzein, glycitein, genistein Fast Analysis of Isoflavones in Dietary Supplements [ 149 ]

152 [ APPLICATION NOTE ] Genistein Genistin Acetyl Genistin Malonyl Genistin Daidzein Daidzin Acetyl Daidzin Malonyl Daidzin Glycitein Glycitin Acetyl Glycitin Malonyl Glycitin Apigenin Figure 1. Structures of isoflavones in this study. EXPERIMENTAL Sample preparation The standards, daidzin, glycitin, genistin, daidzein, glycitein, genistein, and apigenin, were purchased from ChromaDex (Irvine, CA) and INDOFINE Chemical (Hillsborough Township, NJ). Defatted powdered Soy RS was purchased from US Pharmacopeia (Rockville, MD). NIST SRM 3238 was purchased from NIST (Gaithersburg, MD). Isoflavone dietary supplement samples from major brands were purchased from online retail stores. The standard and sample solutions were prepared the same way as in the USP isoflavone method. 1 Sample solutions were further diluted with an acetonitrile water mixture (2/3 by volume) to various levels to fit the calibration range. The concentration of internal standard was always kept constant at 4 ppm. [ 15 ] Fast Analysis of Isoflavones in Dietary Supplements

153 UHPLC conditions UHPLC system: ACQUITY Arc MS conditions MS system: ACQUITY QDa (Performance) Detector: 2998 PDA Ionization mode: ESI + Software: Empower 3 Capillary voltage:.8 kv Column: CORTECS C µm, 3. x mm Column temp.: 3 C Mobile phase A: Water with.1 formic acid Mobile phase B: Acetonitrile with.1 formic acid Injection volume: 2. µl Flow rate: 1.8 ml/min Run time: 18. min UV detection: 26 nm UV resolution: 1.2 nm Elution gradient Time Flow rate (min) (ml/min) A Curve 1 Initial Cone voltage: 15 V Probe temp.: 6 C SIR masses: Table 1. Masses of isoflavone molecular ions. Analyte SIR mass (Daltons) Daidzin Glycitin 447. Genistin Daidzein Glycitein 285. Genistein 27.9 Apigenin (IS) 27.9 Malonyl Daidzin 53.4 Malonyl Glycitin Acetyl Daidzin Acetyl Glycitin 489. Malonyl Genistin 519. Acetyl Genistin RESULTS AND DISCUSSION USP METHOD TRANSFER AND OPTIMIZATION ONTO ACQUITY ARC UHPLC The USP method (isoflavones powder extract) 1 was transferred to an ACQUITY Arc UHPLC System with a CORTECS C18 Column (2.7 µm, 3 x mm), p/n The CORTECS Column s 2.7 µm packing material is solid-core particle, which provides higher separation efficiency and lower back pressure than a fully porous particle column of equivalent particle size. The USP method gradient elution program was converted to a new gradient elution program using Waters ACQUITY UPLC Column Calculator. 3 The column parameters in the USP method (5 µm, 3. x 25 mm) and the parameters of the CORTECS C18 Column (2.7 µm, 3. x mm), as well as the USP method's 74 minutes gradient elution program were entered in the column calculator, and a 18 minutes gradient elution program that is equivalent to the USP method was calculated. The mobile phase additive was changed from.5 phosphoric acid to.1 formic acid, which is a more MS friendly additive. The factory default ACQUITY QDa Detector instrument parameters were used without any modification. The column temperature of 4 C was tested, but was later optimized to 3 C to meet the USP suitability criteria on peak resolution. Figure 2 shows the chromatograms of the USP defatted powdered soy RS, unheated and heated, and their Single Ion Recording (SIR) traces that were obtained using the ACQUITY QDa Mass Detector. Mass detection was used to confirm the peak identities. Fast Analysis of Isoflavones in Dietary Supplements [ 151 ]

154 [ APPLICATION NOTE ] AU Minutes Figure 2. UV chromatograms of unheated defatted soy (black, the bottom trace) and heated defatted soy (red trace next to the bottom), and the corresponding mass detector SIR traces for every compound. Peak ID: 1. Daidzin; 2. Glycitin; 3. Genistin; 4. Malonyl Daidzin; 5. Malonyl Glycitin; 6. Acetyl Daidzin; 7. Acetyl Glycitin; 8. Malonyl Genistin; 9. Daidzein; 1. Glycitein; 11. Acetyl Genistin; 12. Genistein; 13. Apigenin. METHOD TRANSFER AND OPTIMIZATION USING MASS DETECTION Since the acetyl and malonyl isoflavone standards were not commercially available, the peak assignment of these compounds were carried out using a reference material and a pattern matching method as described in the USP standard. Heat treatment of defatted soy (DFS) can convert the malonyl forms to the acetyl forms. By comparing the chromatograms of the unheated DFS and the heated DFS and their reference chromatograms, one can assign the peak IDs to those acetyl and malonyl isoflavones. However, this pattern matching approach is not reliable, especially when LC conditions, such as the column, the mobile phase additives, or the LC system are changed. The ACQUITY QDa detects the ions that are formed in electrospray ionization (ESI) at unit mass resolution (.7 Da). Table 1 lists the molecular ions of these isoflavones. Using mass detection, we were able to selectively detect these compounds and eliminate any possible interference from closely eluting compounds. Genistein and apigenin have the same mass, but their peaks were well separated by chromatography (see Figure 2). The addition of mass detection (ACQUITY QDa Detector) enabled unambiguous assignment of peak IDs to the correct acetyl and malonyl isoflavones without resorting to individual standards. The ACQUITY QDa Detector also sped up the method optimization because the selective detection of every compound allowed us to monitor the retention time (RT) changes of all compounds simultaneously with high confidence. This greatly saved the number of injections in method optimization. More details on how the ACQUITY QDa Detector benefits the method transfer are discussed in a separate application note. 4 METHOD PERFORMANCE CHARACTERISTICS Table 2 shows the UV calibration results of the standards. Apigenin was used as the internal standard for calibration and quantitation of all compounds. The square of the correlation coefficients (R 2 ) between the relative responses (peak area ratio) and the concentration of standards in solutions (ppm) for all compounds were better than.999. The retention time relative standard deviations for all compounds were less than.12. Table 2. Isoflavones retention times and their relative standard deviation, calibration equations, R 2, and linear ranges. RT Range Analyte Min RSD () Equation R 2 (ppm) Daidzin Y = 2.9x1-1 X + 3.9x Glycitin Y = 2.42x1-1 X x Genistin Y = 4.53x1-1 X x Daidzein Y = 2.6x1-1 X x Glycitein Y = 4.96x1-1 X + 1.7x Genistein Y = 4.91x1-1 X x [ 152 ] Fast Analysis of Isoflavones in Dietary Supplements

155 Table 3 shows the isoflavone results for the NIST 3238 SRM and the comparison to its certified and reference values. A relative difference of <11 was obtained for the genistin, glycitin, daidzin, genistein, and glycitein. The result of daidzein was 15 higher than the NIST value. A literature search found that a high daidzein value was also reported elsewhere. 5 The accuracy for the daidzein, genistein, and glycitein were also evaluated by a spiking experiment (Table 4). Recoveries of 98 to 11 were obtained for these compounds. ANALYSIS OF ISOFLAVONE DIETARY SUPPLEMENTS The isoflavone content in four isoflavone dietary supplement samples were measured by the fast 18-minute UHPLC-UV method described above. Sample types included tablets, capsules, and soy powder. The USP calibration and quantitation protocols 1 were followed in the data processing. Table 3. Comparison of determined isoflavone values to the certified and reference values of NIST 3238 SRM. NIST value (mg/kg) Determined value (mg/kg) Relative difference Genistin 127± Glycitin 376± Daidzin 134± Daidzein 241± Genistein 18± Glycitein 211± Table 4. Recovery results from a spiking experiment. Original value (mg/kg) Spiked level (mg/kg) Determined value (mg/kg) Recovery () Daidzein Genistein Glycitein The same conversion factors for the acetyl and malonyl derivatives that are specified in the USP HPLC-UV method were used in the analyses. Table 5 shows the determined individual and the total isoflavones, as well as the label claimed total isoflavone contents. For easy comparison, the label claim values were converted to concentration (mg/kg). Two of the three samples (C and D) showed good agreement between the determined values and their label claim values, while one sample (B) contained much less measured total isoflavone content then claimed on its label. The reason for such low total isoflavone content is unknown. Table 5. Isoflavone contents in dietary supplements and their label claim values. Sample B C D E (mg/kg) Mean RSD Mean RSD Mean RSD Mean RSD Daidzin Glycitin Genistin Malonyl Daidzin Acetyl Daidzin Acetyl Glycitin Malonyl Genistin Daidzein Glycitein Acetyl Genistin Genistin Total Isoflavones 12,543 3,6 1, Label Claim Value >31,25 >25, 867 to 2,6 no claim Fast Analysis of Isoflavones in Dietary Supplements [ 153 ]

156 [ APPLICATION NOTE ] CONCLUSIONS The USP method for isoflavones was successfully transferred to an ACQUITY Arc UHPLC System with a 2998 PDA Detector. The total analysis time per injection on the UHPLC system was 18 minutes, which was significantly shorter than the 74 minutes for the USP method. This corresponds to a three times increase in the analysis throughput, and a 75 cost savings for solvents used. The ACQUITY QDa Mass Detector provided excellent detection selectivity, which is a great asset in method transfer and development, as well as in the isoflavone analysis of unknowns and challenging samples where potential interference risk is high. Analysis of isoflavones in three dietary supplements showed compliance in the label claims for two samples. Low isoflavone content was found in one of the three tested samples for unknown reason. References 1. USP Monograph. Powdered Soy Isoflavones Extract, USP39 NF34 S1 [6841], The United States Pharmacopeial Convention. 2. AOAC Official Method 28.3 Total soy isoflavones in dietary supplements, supplement ingredients, and soy foods HPLC with UV detection First Action 28, AOAC International K Fountain. Transferring USP Compendia HPLC Methods to UPLC Technology for Routine Generic Drug Analysis, Waters application note, no EN, J Yang, M Benvenuti, G Cleland. Fast Analysis of Isoflavones Benefits of Mass Detection in Method Transfer and Sample Analysis, Waters application note, no EN, LX Zhang, CQ Burdette, MM Phillips, CA Rimmer, and RK Marcus. Determination of isoflavone content in SRM 3238 using liquid chromatographyparticle beam/electron ionization mass spectrometry. J AOAC Int. 215; 98(6): Waters, ACQUITY, QDa, CORTECS, Empower, and The Science of What's Possible are registered trademarks of Waters Corporation. Arc is a trademark of Waters Corporation. All other trademarks are the property of their respective owners. 216 Waters Corporation. Produced in the U.S.A. December EN AG-PDF Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F: [ 154 ] Fast Analysis of Isoflavones in Dietary Supplements

157 [ APPLICATION NOTE ] A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products Using the ACQUITY H-Class System and ACQUITY QDa Mass Detector Mark Benvenuti, Dimple Shah, and Jennifer A. Burgess Waters Corporation, Milford, MA, USA APPLICATION BENEFITS The ACQUITY QDa Detector facilitates simultaneous analysis of sweeteners in single run. Quantification of natural and artificial non-nutritive sweeteners in less than 1 minutes. Mass detection provides information-rich orthogonal detection for co-eluting compounds. The ACQUITY QDa Mass Detector provides sensitive and selective determination of UV transparent sweeteners. WATERS SOLUTIONS ACQUITY UPLC H-Class System ACQUITY QDa Mass Detector MassLynx MS Software INTRODUCTION Sugars are renowned for their sweet taste and are often added to manufactured foods to enhance human perception of flavor. Due to the negative health effects of excessive consumption of sugar, alternative non-nutritive sweeteners are commonly used in food and beverage products. Examples include soft drinks, table-top sweeteners, chocolates, dairy products, and many other so-called diet foods. In many cases it is a combination of various sweeteners that are used to impart overall sweetness to these products. Aspartame, saccharin, acesulfame-k, neotame, and sucralose are approved artificial sweeteners by the U.S. FDA. 1 Compounds such as rebaudioside A and stevioside, which originated from the South American Stevia plant, are also becoming more popular in the U.S. as a sweetener. In 21, the European Food Safety Authority (EFSA), approved the use of stevioside as a sweetener. The European Union (EU) Directive 94/35/EC, along with four amendments: 96/83/EC, 23/115/EC, 26/52/EC, and 29/163/EU restrict the level of sweeteners in specific types of food. The EU Commission Regulation 1129/211 lists the maximum level of the sweeteners permissible in various food products. Hence the determination of the amount of these sweeteners in foods is also important in order to ensure consistency in product quality. The most common method for the detection of sweeteners is HPLC coupled to a UV detector. This configuration enables the detection of some sweeteners such as acesulfame-k, aspartame, saccharin, and neotame. However cyclamate and sucralose cannot be analyzed by UV because they lack a chromophore. The ability to analyze all of these sweeteners using a single method with mass detection would be ideal. Waters has developed the ACQUITY QDa Mass Detector to allow food and beverage scientists to incorporate mass detection into their existing chromatographic workflows. The ACQUITY QDa not only allows for the detection of all sweeteners in a single run, but it also brings improved discrimination to the analysis, thereby eliminating the requirement of baseline separation of all compounds. KEYWORDS The combination of Waters ACQUITY UPLC H-Class System with the ACQUITY QDa Mass Detector is extremely beneficial for food and beverage Aspartame, saccharin, rebaudioside, manufacturers for the identification and quantification of sweeteners stevioside, Stevia, acesulfame-k, in their products using a single analysis method. neotame, sucralose, mass detection In this application note, a fast, reliable and sensitive method was developed to analyze sweeteners in food and beverage samples. A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products [ 155 ]

158 [ APPLICATION NOTE ] EXPERIMENTAL UPLC conditions UPLC system: Column: ACQUITY UPLC H-Class ACQUITY UPLC HSS T3 1.8 µm, 2.1 x mm Standard preparation Stock solutions of each of the sweeteners were prepared in water. A series of working solutions were also prepared in water with a concentration range from.4 to 3 mg/l. Column temp.: 4 C Injection volume: 2 µl Flow rate: Mobile phase A: Mobile phase B: Wash solvent: Purge wash: Seal wash:.4 ml/min Water +.1 formic acid Acetonitrile.1 formic acid 5/5 water/methanol (v/v) 1/9 acetonitrile/water (v/v) 1/9 acetonitrile/water Table 1. UPLC gradient method for sweeteners analysis. Time (min) Flow rate (ml/min) A B Curve Initial Sample preparation A total of seven different samples were analyzed in this work. All of the samples were purchased from local markets. The samples included two beverages (diet cola and diet tea), two table-top sweeteners, a ready-made pudding, hard candy, and diet marmalade. The table-top sweeteners (~1 g) were dissolved in ml water and further diluted at two additional levels: 1:2 and 1:1. All three dilution levels were injected to cover the different concentrations of sweeteners in this sample. The candy (5.281 g) and pudding (1.244 g) were dissolved separately in ml water and diluted 1:1. These dilution levels were injected. Marmalade (1.432 g) was dissolved in ml water and analyzed. The beverages were filtered through a.22-µm PVDF filter, diluted 1:2 with water, and analyzed. Detector conditions Detector 1: Wavelength: Analog channels: Scan rate: Detector 2: Ionization mode: ACQUITY UPLC PDA Scanning 21 to 4 nm at 214 and 254 nm 1 pts/sec ACQUITY QDa (Performance) ESI- Probe temp.: 5 C Capillary voltage: Sampling rate: Acquisition type: Cone voltage:.6 kv 5 pts/sec Full scan m/z 15 to 1 (centroid) Ramp from 5 to 4 V SIR channels: See Table 2 [ 156 ] A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products

159 RESULTS AND DISCUSSION The chemical structures for the eight sweeteners analyzed in this study are shown in Figure 1. These sweeteners were separated on a reversed-phase column and detected using the ACQUITY QDa Mass Detector. The retention times, along with the Single Ion Recording (SIR) mass-to-charge ratio (m/z), and cone voltages of these sweeteners are shown in Table 2. Sodium Cyclamate Saccharin Sucralose Aspartame Acesulfame K Stevioside Neotame Rebaudioside A (Reb A) Figure 1. Chemical structures of the sweeteners analyzed. Table 2. Retention times, SIR m/z and cone voltages for the sweeteners. Compounds SIR m/z Cone Voltage (V) RT (min) Acesulfame K Aspertame Cyclamate Neotame Rebaudioside A Saccharin Stevioside Sucralose Figure 2 shows the SIR chromatograms of the individual sweetener standards in solvent. As shown in Figure 2, there are two co-eluting pairs. Aspartame (3.2) and sucralose (3.2) eluted at the same time, while rebaudioside A (4.11) and stevioside (4.14) eluted at very nearly the same time. However, due to the different mass of aspartame (m/z 293) and sucralose (m/z 395), mass detection was able to separate and detect these compounds. Likewise, rebaudioside A (m/z 965.5) and stevioside (m/z 83.3) were co-eluting, but they were still able to be individually detected using the ACQUITY QDa Mass Detector. Acesulfame K, RT 1.4 SIR m/z: 162 Sucralose, RT 3.2 SIR m/z: 395 Saccharin, RT 1.34 SIR m/z: 182 Rebaudioside A, RT 4.11 SIR m/z: Cyclamate, RT 1.77 SIR m/z: 178 Stevioside, RT 4.14 SIR m/z: 83.3 Aspartame, RT 3.2 SIR m/z: 293 Neotame, RT 4.43 SIR m/z: 377 Figure 2. SIR chromatograms of all sweeteners in a solvent standard. A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products [ 157 ]

160 [ APPLICATION NOTE ] Example calibration curves of two of the co-eluting sweeteners are shown in Figure 3. Even though the compounds are not chromatographically resolved, mass detection offers the selectivity that enables accurate quantification results. The discrimination offered by mass detection further eliminated tedious method development and/or method optimization steps that would be necessary to attempt to obtain baseline separation of the co-eluting components. Mass detection offers the opportunity to distinguish between sweeteners that cannot be not detected by UV due to the lack of a chromophore group in the structures. Even though cyclamate and sucralose are both UV transparent compounds, their detection can be easily achieved with the ACQUITY QDa Mass Detector. Figure 4 shows the SIR chromatograms of cyclamate (RT 1.77) and sucralose (RT 3.2), along with the UV channel at 254 nm at 5 mg/l. Compound name: Reb A Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None A 4 Response mg/l Compound name: Stevioside Correlation coefficient: r = , r 2 = Calibration curve: * x Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None B 15 Response mg/l Figure 3. Calibration curve of co-eluting pairs: A. rebaudioside, and B. stevioside. A : Sucralose SIR m/z 395 B: Cyclamate SIR m/z 178 C : UV at 254 nm Figure 4. SIR chromatogram of A. sucralose, B. cyclamate, and C. UV chromatogram at 254 nm of a standard mix at 5 mg/l. [ 158 ] A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products

161 Calibration curves were created for the target analytes. Table 3 lists the optimum calibration range, correlation coefficient, and curve fit used for each of the compounds studied. The correlation constant (R 2 ) was >.995 for all of the sweeteners. ANALYSIS OF SAMPLES The seven samples were prepared as previously described, and analyzed in duplicate. Table 4 shows the quantification results of all seven samples. The first table-top sweetener contained acesulfame K (98 mg/kg), aspartame (15,8 mg/kg), and saccharin (3798 mg/kg). The second table-top sweetener was found to contain rebaudioside A (315 mg/kg) and stevioside (63 mg/kg). With the discrimination and sensitivity of mass detection it was easily apparent that both of these co-eluting sweeteners were present in the sample, even though one was present at a much lower level and was not labeled to be present. Figure 5 shows the SIR chromatograms of rebaudioside A and stevioside in a solvent standard and in the second table-top sweetener. Table 3. Calibration range, correlation coefficient values and curve fit (R 2 ) for the eight sweeteners. Analyte Calibration range (mg/l) Correlation coefficient (R 2 ) Acesulfame K Aspartame Cyclamate Neotame Rebaudioside A Saccharin Stevioside Sucralose Standard Standard Curve fit Quadratic, 1/X weighting Quadratic, 1/X weighting Quadratic, 1/X weighting Quadratic, 1/X weighting Linear, 1/X weighting Quadratic, 1/X weighting Linear, 1/X weighting Quadratic, 1/X weighting Aspartame was detected in the diet cola and diet tea at 3 mg/l and 489 mg/l respectively. Sucralose was found in the hard candy (145 mg/kg) and in the marmalade (297 mg/kg). The amount of sucralose found in the marmalade was within the EU Commission Regulation 1129/211. Table-top sweetener 2 Table-top sweetener 2 Figure 5. SIR chromatogram of stevioside and rebaudioside A in solvent and sample (tabletop sweetener 2). Table 4. Quantification results of the seven samples tested. Sample Acesulfame K Aspartame Rebaudioside A Saccharin Stevioside Sucralose Unit Diet cola 3 48 mg/l Candy 145 mg/kg Diet tea 379 mg/l Marmalade 297 mg/kg Pudding mg/kg Table-top sweetener mg/kg Table-top sweetener mg/kg A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products [ 159 ]

162 [ APPLICATION NOTE ] CONCLUSIONS Reference This application note has described a method for the separation, detection, 1. D J Yang, B Chen. Simultaneous determination of non-nutritive sweeteners in foods by HPLC/ESI-MS. J Agric Food Chem. 57(8): , 29. and quantification of natural and artificial non-nutritive sweeteners in less than 1 minutes using the ACQUITY UPLC H-Class System coupled with the ACQUITY QDa Mass Detector. The ACQUITY QDa Mass Detector provides single method to analyze both UV-transparent and non-transparent sweeteners, and this method can be incorporated into existing workflows with or without UV detection. Mass detection provides a much higher level of analyte discrimination and it does not require baseline separation of analytes with different masses. This helps to minimize method development times and removes the need for separate injections of individual standards to verify compound identifications. Waters, ACQUITY, UPLC, MassLynx, QDa, and The Science of What s Possible are registered trademarks of Waters Corporation. All other trademarks are the property of their respective owners. 214 Waters Corporation. Produced in the U.S.A. November EN AG-PDF [ 16 ] A Method for the Rapid and Simultaneous Analysis of Sweeteners in Various Food Products Waters Corporation 34 Maple Street Milford, MA 1757 U.S.A. T: F:

163 YOU REDEFINE WHAT S ROUTINE. With VionTM IMS QTof, you get cleaner spectra, additional data for confident identifications, and streamlined analytical workflows. This is ion mobility for everyone. Make it part of your routine by bringing Vion IMS QTof to your lab. Visit waters.com/vion PHARMACEUTICAL HEALTH SCIENCES FOOD ENVIRONMENTAL CHEMICAL MATERIALS 216 Waters Corporation. Waters and the Science of What s Possible are registered trademarks of Waters Corporation. Vion is a trademark of Waters Corporation.