Current Trends in Mass Spectrometry March 2020 Chromatographyonline.Com

Volume 18 Number 1 March 2020 www.chromatographyonline.com

Identifying and Quantifying PFAS in Water with EPA Method 8327

Automated Analysis of Trace Elements in Single Cells Using ICP-MS

Expediting Drug Discovery with Affinity Selection– Mass Spectrometry

GCxGC–TOF-MS for Comprehensive Fingerprinting of Volatiles in Food

Separation of Eight Cannabinoids

With the recent legalization of both medicinal and recreational marijuana in the United States, analysis of individual cannabinoids has captured the public’s interest at a new level. As such, many new cannabis products are now available, i.e., edibles, vaporizers, and extracts to name a few. The increased marketability of the product has incited consumers to take a greater interest in the quality and craft ability of the products being sold. Through the quantification of individual cannabinoids, the consumer can make an informed decision about the possible effects they could expect from the products they purchase. Therefore, the need for accurate, robust, and affordable analysis tools are of the upmost importance.

With health, safety, and edibles dosing as the primary motivation, Hamilton Company developed an HPLC method that isolates Separation of Eight Cannabinoids eight major cannabinoids. The HxSil C18 (3 µm) column provides

an accurate, cost effective, and robust solution that can be used 6 in any HPLC system. 40 3 2

1 Column Information 30 4 Packing Material HxSil, 3 µm 5 mAU Part Number 79641 20 7 8 Dimensions 150 x 4.6 mm

Chromatographic Conditions 10 0–10 min, 78–92% B Gradient 10–15 min, 78% B Temperature Ambient 0 2 3 4 5 6 7 8 9

Injection Volume 5 μL Time (minutes)

Detection UV at 230 Compounds: Eluent A 20 mM NH COOH pH 3.5 1: Cannabidivarin (CBDV) 5: Cannabigerol (CBG) 4 2: Cannabidiol (CBD) 6: Cannabinol (CBN) Eluent B Acetonitrile 3: Cannabidiolic Acid (CBDA) 7: Δ-9-Tetrahydrocannabinol (Δ-9-THC) 4: Cannabigerolic Acid (CBGA) 8: Δ-9-Tetrahydrocannabinolic Acid (Δ-9-THCA) Flow Rate 1.0 mL/min

Author: Adam L. Moore, PhD, Hamilton Company ©2019 Hamilton Company. All rights reserved. All other trademarks are owned and/or registered by Hamilton Company in the U.S. and/or other countries. Lit. No. L80098 — 08/2019

Web: www.hamiltoncompany.com Hamilton Americas & Pacific Rim Hamilton Europe, Asia & Africa Hamilton Company Inc. Hamilton Central Europe S.R.L. USA: 800-648-5950 4970 Energy Way str. Hamilton no. 2-4 Reno, Nevada 89502 USA 307210 Giarmata, Romania Europe: +40-356-635-055 Tel: +1-775-858-3000 Tel: +40-356-635-055 Fax: +1-775-856-7259 Fax: +40-356-635-060 To find a representative in your area, please visit hamiltoncompany.com/contacts. [email protected] [email protected]

HAM1477_Separation-of-8-Cannabinoids-App-Note_R1_FNL.indd 1 8/13/19 12:40 PM 4 Current Trends In Mass Spectrometry March 2020 chromatographyonline.com

MANUSCRIPTS: For manuscript preparation guidelines, see chromatographyonline. 485F US Highway One South, Digital Production Manager com/lcgc-author-guidelines, or call The Editor, (732) 596-0276. LCGC welcomes Suite 210 Sabina Advani unsolicited articles, manuscripts, photographs, illustrations, and other materials Iselin, NJ 08830 [email protected] but cannot be held responsible for their safekeeping or return. Every precaution (732) 596-0276 Fax: (732) 647-1235 Managing Editor, Special Projects is taken to ensure accuracy, but LCGC cannot accept responsibility for the Kaylynn Chiarello-Ebner accuracy of information supplied herein or for any opinion expressed. [email protected] PUBLISHING/SALES SUBSCRIPTIONS: For subscription and circulation information: LCGC, P.O. Box MARKETING/OPERATIONS 457, Cranbury, NJ 08512-0457, or email [email protected]. Delivery of Vice President/Group Publisher Michael J. Tessalone LCGC outside the United States is 14 days after printing. (LCGC Europe and LCGC [email protected] Marketing Manager Asia Pacific are available free of charge to users and specifiers of chromatographic Brianne Pangaro equipment in Western Europe and Asia and Australia, respectively.) Publisher [email protected] Stephanie Shaffer [email protected] C.A.S.T. Data and List Information CHANGE OF ADDRESS: Send change of address to LCGC, P.O. Box 457, Melissa Stillwell Cranbury, NJ 08512-0457; alternately, send change via e-mail to Associate Publisher [email protected] [email protected] or go to the following URL: Edward Fantuzzi http://mmhpubs.mmhgroup.com/Welcome.aspx?pubid=LCGC [email protected] Reprints Alexandra Rockenstein Allow four to six weeks for change. PUBLICATIONS MAIL AGREEMENT No. 40612608. [email protected] Return all undeliverable Canadian addresses to: IMEX Global Solutions, P.O. Box 25542, Sales Manager Brianne Molnar London, ON, N6C 6B2, CANADA. Canadian GST number: R-124213133RT001. [email protected] Audience Development Manager Jessica Stariha C.A.S.T. DATA AND LIST INFORMATION: Contact Melissa Stillwell, Account Executive [email protected] tel. (218) 740-6831, e-mail [email protected]. Timothy Edson [email protected] CORPORATE REPRINTS: Contact Michael J. Tessalone, Senior Director, Digital Media Chairman & Founder e-mail: [email protected] Michael Kushner Mike Hennessy Sr [email protected] INTERNATIONAL LICENSING: Contact Kim Scaffidi, Vice Chairman e-mail: [email protected] Jack Lepping EDITORIAL President & CEO CUSTOMER INQUIRIES: Customer inquiries can be forwarded Editorial Director Mike Hennessy Jr directly to MJH Life Sciences, Attn: Subscriptions, 2 Clarke Drive, Suite Laura Bush 100, Cranbury, NJ 08512; e-mail: [email protected] [email protected] Chief Financial Officer Neil Glasser, CPA/CFE Managing Editor John Chasse Executive Vice President, ecycled R Pa [email protected] Operations % p 0 e r

5 1

Tom Tolvé

0 e

t - 2 s Senior Technical Editor 0 a % W Senior Vice President, Content P r Jerome Workman os e t Consum [email protected] Silas Inman

Associate Editor Senior Vice President, Cindy Delonas I.T. & Enterprise Systems © 2020 MultiMedia Pharma Sciences, LLC. All rights reserved. No part of this publication [email protected] John Moricone may be reproduced or transmitted in any form or by any means, electronic or mechanical including by photocopy, recording, or information storage and retrieval without permission in Creative Director, Publishing Senior Vice President, writing from the publisher. Authorization to photocopy items for internal/educational or person- Ray Pelesko Audience Generation al use, or the internal/educational or personal use of specific clients is granted by MultiMedia [email protected] & Product Fulfillment Pharma Sciences, LLC. for libraries and other users registered with the Copyright Clearance Joy Puzzo Senior Art Director Center, 222 Rosewood Dr., Danvers, MA 01923, (978) 750-8400, fax (978) 646-8700, or Gwendolyn Salas Vice President, Human Resources visit http://www.copyright.com online. For uses beyond those listed above, please direct your [email protected] & Administration written request to Permission Dept. email: [email protected] Shari Lundenberg Graphic Designer MultiMedia Pharma Sciences, LLC. provides certain customer contact data (such as custom- Courtney Soden Vice President, Business er’s name, addresses, phone numbers, and e-mail addresses) to third parties who wish to pro- [email protected] Intelligence mote relevant products, services, and other opportunities that may be of interest to you. If you do Chris Hennessy not want MultiMedia Pharma Sciences, LLC. to make your contact information available to third CONTENT MARKETING Vice President, Marketing parties for marketing purposes, simply email [email protected] and a customer service Amy Erdman representative will assist you in removing your name from MultiMedia Pharma Sciences, LLC. lists. Custom Content Writer Executive Creative Director, LCGC North America does not verify any claims or other information appearing in any of Allissa Marrapodi Creative Services [email protected] the advertisements contained in the publication, and cannot take responsibility for any losses Jeff Brown or other damages incurred by readers in reliance of such content. Webcast Operations Manager To subscribe, email [email protected]. Kristen Moore [email protected]

Project Manager Vania Oliveira [email protected] chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 5

March 2020

Articles

Towards Automated Routine Analysis of the Distribution of Trace Elements in Single Cells Using ICP-MS 6 M. Corte-Rodríguez, R. Álvarez-Fernández García, P. García-Cancela, M. Montes-Bayón, J. Bettmer, and D.J. Kutscher Analysis of the compositional variation in living cells is essential for understanding biological processes. Single-cell elemental analysis by triple-quadrupole ICP-MS is emerging as a selective, highly sensitive, and potentially high-throughput technique for the study of constitutive elements, and uptake of metallodrugs (or metal-containing nanomaterials) in single cells.

Rapid Quantitation of PFAS in Non-Potable Waters 11

Ruth Marfil-Vega and Brahm Prakash The presence of per- and polyfluoralkyl substances (PFAS) in water is an important health and environmental concern. Liquid chromatography–mass spectrometry (LC–MS) has been established as the most suitable technology for monitoring these substances. A method is described, using EPA 8327, for PFAS analysis in groundwater, surface water, and wastewater.

Affinity Selection-Mass Spectrometry: Defining the Bioactive Compounds in Complex Mixtures of Natural Products and Combinatorial Libraries 18

Richard B. van Breemen Drug discovery using high-throughput screening of discreet compounds, and the discovery of natural products with pharmacological mechanisms of action, rely on bioassay-guided fractionation analysis. Recent applications of affinity selection–mass spectrometry (AS-MS) are useful for exploring the discovery of ligands to membrane-bound proteins and RNA targets.

GC×GC–TOF-MS and Comprehensive Fingerprinting of Volatiles in Food: Capturing the Signature of Quality 26

Federico Stilo, Erica Liberto, Carlo Bicchi, Stephen E. Reichenbach, and Chiara Cordero Food quality differences are dependent on botanical and geographical origins of primary food ingredients as well as storage and handling. Quality assessment for food materials, including cocoa and olive oil, is demonstrated by applying two-dimensional gas chromatography (GC×GC) combined with time-of-flight mass spectrometry (TOF-MS) and pattern recognition.

Departments

Products �� 33

Cover image courtesy of Oleksii Sergieiev/stock.adobe.com. 6 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Towards Automated Routine Analysis of the Distribution of Trace Elements in Single Cells Using ICP-MS Cell-to-cell variability is known to be of crucial importance to understand different biological processes. Studying the individual variations of, for example, trace elements in cell populations can be carried out only by means of single cell analysis techniques, and, for this aim, single cell elemental analysis by ICP-MS is emerging as a selective, highly sensitive, and potentially high-throughput technique. The study of constitutive elements, and uptake of metallodrugs (or metal-containing nanomaterials), are of special interest and importance. In this work, a highly efficient sample introduction system and a triple-quadrupole (TQ) ICP-MS, in combination with a dedicated microflow autosampler, was used for single cell analysis. This setup enables the transport of intact single cells to the inductively coupled plasma used as an ion source, with transport efficiencies over 50% for different cell lines. The analysis of single cells using an element selective detection system, such as an ICP-MS, allows one to study the elemental content of both intrinsic and exogenous elements. For some of the intrinsic elements, such as phosphorus, typically occurring spectral interferences need to be removed by using a triple-quadrupole based ICP-MS system. The strategy to analyze concentration distributions at single cell level will be presented for yeast cells. The selected combination of instruments enables fully automated, unattended, and potentially high-throughput analysis of single cells.

M. Corte-Rodríguez, R. Álvarez-Fernández García, P. García-Cancela, M. Montes-Bayón, J. Bettmer, and D.J. Kutscher

he analysis of elements in biological systems has a long tra- might be considered useful for this purpose as well. Similar to dition, due to its importance for the understanding of their single-particle ICP-MS developed by Degueldre and associates in T functions. Because of the increasing sensitivity of mod- 2003, the concept of single cell ICP-MS is based on introducing a ern analytical techniques like inductively coupled plasma-mass diluted cell suspension into the ionization source in combination spectrometry (ICP-MS), the determination of various elements with short integration times (between µs and a few ms) on the de- in small individual objects like cells, called single cell analysis, is tector side (1). Once a cell enters the ionization source, it produces now possible. This significant improvement can provide insights a plume of ions that can be registered at the detector as a short into the biological variation of the elemental composition within signal of approximately 500 µs duration. The signal intensity for a cell population. Other important information can be gathered a certain isotope is related to the corresponding elemental mass for the cellular uptake of nanomaterials or pharmaceutical drugs. within the cell and the frequency of signals (usually called events) Single cell ICP-MS usually refers to the introduction of a cell correlates to the cell number concentration in the suspension. suspension via a nebulization system, although laser ablation It was first shown by Li and colleagues, and forms the concep- chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 7 tual basis of recent approaches (2). Most washing solution, and, therefore, 2% nitric cytometry). Similarly, the other cell lines critical of such a system is the sample in- acid is used for this purpose. All internal (A2780 and GM04312) were washed troduction through nebulization. First, it parts of the autosampler are inert and with a TBS solution, and finally diluted requires the transport of intact cells until metal-free. One sample can be analyzed to approximately 25,000 cells per mL. they reach the plasma, and, second, the in a total time of less than 7 min, with an transport efficiency should be as high as effective measurement time of up to 3 min. Data Treatment possible. Recent developments presented As previously reported (4,6), each data set combinations of nebulizers and spray Preparation of Cell Samples was averaged, and the data points higher chambers that permit high efficiencies Lyophilized yeast samples were resus- than three or five standard deviations over of up to 100% (3–5). In this work, we will pended in water, washed twice by cen- the mean were considered as cell or par- present a highly efficient sample intro- trifugation, and diluted to a final con- ticle events. This procedure was iterated, duction system, and a triple- quadrupole centration of around 50,000 yeast cells after removal of the events, until no new (TQ) ICP-MS in combination with a ded- per mL in water (determined by flow data points above the threshold were de- icated microflow autosampler. Transport efficiencies will be illustrated on several examples. Finally, we will demonstrate the application to the analysis of yeast SMART LABS CHOOSE samples, in order to demonstrate the in- tracellular incorporation of elements by ON-SITE GAS GENERATION the commercial producer.

Experimental Instrumentation For all ICP-MS measurements, an iCAP TQ ICP-MS (Thermo Fisher Scientific) was used. For measurements of phosphorus, selenium, and chromium, TQ-O2 mode was selected (mass shift from 31P+ to 31P16O+, 80Se+ to 80Se16O+ and 52Cr+ to 52Cr16O+, respectively, after reaction with oxygen in the reaction cell). For single cell measurements, the instrument was fitted to the high sensitivity single-cell sample introduction system for ICP-MS (Glass Expansion). The data were acquired using time-resolved analysis mode at a dwell time of 5 ms. All ICP-TQ-MS parameters are summarized in Table I. Samples and rinsing solutions were introduced into the plasma at a flow rate of 10 µL/min, using the MVX-7100 µL workstation (Teledyne CETAC Technologies). H2, N2 AND ZERO AIR ON-DEMAND This autosampler system is connected • Consistent Purity • Cost Effective to the iCAP TQ through a trigger cable, which allows the unsupervised analysis of • Consistent Pressure • Eliminates Cylinder sample sequences. This system also offers • Proven Safe Storage and Delivery the possibility of diluting and mixing the Issues samples before the injection. Briefly, the workstation allows to place the samples in a cooled holder, where they are aspirated through an inert sample probe. Only the •Chicago, IL Philadelphia, PA measured sample volume needs to be •March 3-5 • Pittcon taken, because the sample is then brought •March 22-26 to the injection loop. After settling the ACS Spring sample volume into the loop, a carrier flow pushes it to the sample introduction system. The carrier flow is also used as +1.203.949.8697 www.ProtonOnSite.com

LCGC-CTMS Island Half Page Ad – Smart Labs – March Shows.indd 1 1/29/20 12:21 PM 8 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Table I: Instrumental ICP-MS parameters for single cell analysis ICP-MS Parameters Parameter Value 500 Instrument iCAP TQ ICP-MS (a) RF Power (W) 1250 400 (cts) +

Plasma gas flow rate (L/min) 14.0 Au 300 197

Auxiliary flow rate (L/min) 0.8 200 Carrier gas flow rate (L/min) 0.5

Intensity 100 O2 Cell gas flow rate (mL/min) 0.3 0 High sensitivity single-cell sample intro- 0 10 20 30 40 50 60 Nebulizer and spray chamber duction system Time (min) Sheath gas flow rate (L/min) 0.31 100 (b) Sample introduction MVX-7100 workstation 80 Data Acquisition Parameters 60 Data acquisition mode Time-resolved analysis 40 Sample flow rate (µL/min) 10 Frequency Dwell time (ms) 5 20

Run time (s) 120 - 180 0 0 100 200 300 400 500 600 700 Isotopes monitored 31P+|31P16O+, 80Se+|80Se16O+, 52Cr+|52Cr16O+ Intensity197Au+ (cts)

Figure 1: (a) Time-resolved signal and Table II: Obtained transport efficiencies for different cell lines (b) histogram for the measurement of a Cell Line (Cell Number Concentration, Transport Efficiency (%) suspension of 24000 particles per mL of the (per mL) 30 nm gold nanoparticle standard NIST 8012. A2780 (25,000) 50 - 58 GM04312 (25,000) 85 - 95 Yeast cells (50,000) 62 - 69

2000 (a) (cts)

tected. All single-cell or particle signals In order to mimic the transport effi- + 1500 higher than three standards deviations ciency of cells (an important factor for Eu 153 above their mean were discarded as mul- the determination of the cell number 1000 tiple-cell or particle events. concentration), europium-loaded cal- 500 ibration beads (Fluidigm) were used. Intensity

Results and Discussion This calibration standard contains 3.3 0 5 0 20 40 60 Transport Efficiency of the x 10 natural isotopic europium-loaded Time (min)

Sample Introduction System polystyrene beads per mL, with a par- 50 The transport efficiency of the liquid ticle size of 3 µm. These particles are (b) elemental standards for mass determi- better comparable to cells in terms of 40 nation was calculated using the 30 nm particle size and matrix composition 30 gold nanoparticles reference material than gold nanoparticles. After 10 times 20 NIST 8012, as previously reported (4). A dilution in water, a typical measurement typical measurement of gold nanopar- of polystyrene calibration beads and its Frequency 10 ticles and its corresponding histogram corresponding histogram is shown in 0 0 500 1000 1500 2000 is shown in Figure 1. The average sig- Figure 2. In this case, 174 particles per Intensity153Eu+ (cts) nal was 99±10 counts per nanoparticle, minute were detected, with an average and 124 nanoparticles per minute were intensity of 818 ± 101 counts per particle. detected. The transport efficiency was The transport efficiency was 51% for the Figure 2: (a) Time-resolved signal and (b) calculated by dividing the number of calibration beads, showing that there is histogram for the measurement of a 33000 detected particles by the number of no significant change in the transport beads per mL suspension of polystyrene particles in the suspension, which was efficiency when calculated with gold calibration beads loaded with Eu. calculated using the certified values of nanoparticles or polystyrene beads. flow cytometry (4). This determination gold concentration and particle size. For the determination of the cell served as a reference for the following The transport efficiency calculated in transport efficiency, the strategy applied single cell ICP-MS experiments. In a this manner resulted to be 55% for 30 was initially to determine the cell num- recent work, we suggested, either after nm gold nanoparticles. ber concentration of the suspension by the uptake of a non-toxic terbium-con- chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 9 taining complex or the detection of the constitutive element iron, for counting cell events by ICP-MS (4). Since the ICP- TQ-MS offers the sensitive detection of 10000 (cts) +

other constitutive elements like phos- O 16 phorous, we figured out that this element P was better suited as a general ICP-MS cell 31 5000 marker. Due to its ubiquitous presence in biological systems (in form of DNA, RNA, Intensity 0 phosphate, phospholipids, and so forth), 0 20 40 60 the detection of 31P+ with a mass shift to Time (min) 31 16 + P O after reaction with oxygen turned Figure 3: Time-resolved signal of phosphorus for the measurement of a 25,000 GM04312 cells out to be a good indicator for the presence per mL. and number of cells. Figure 3 shows a typical example of the time-resolved analysis of GM04312 cells in the single cell mode. Cell events could be clearly distinguished 100 90 from the background using the criteria 80 mentioned in paragraph “data treat- 70 ment.” The number of detected cell events 60 was then set into relation to the theoret- 50 ically expected cell number, as obtained 40 30 from the flow cytometric experiments, 20 and the resulting ratios were equal to the 10

transport efficiencies. Efﬁciency (%) Transport 0 The different cell lines showed the 25 000 50 000 100 000 transport efficiencies, as summarized Cell number concentration (per mL) in Table II. Differences between the cell types could be observed, and might be Figure 4: Transport efficiency of yeast cells in function of the cell number concentration. explainable by the differing cell shapes, robustness, and sizes. The day-to-day variations were in the order of 5% absolute, and the transport efficiency was virtually independent from the intro- 2000 (a)

duced cell number concentration (Fig- (cts) + 1500 ure 4). Initial experiments on the effect O 16 of cell sedimentation showed that any P influence was negligible up to 30 min 31 1000 analysis time. In any case, the sample introduction system used offers in prin- 500 ciple a resuspending step. Quantification Intensity 0 of different elements in individual cells 0 20 40 60 was carried out after external calibration Time (min) using single elemental standards, and the required transport efficiency for the 100 (b) aqueous standard was obtained by cor- (cts) +

relating it to a standard of gold nanopar- O ticles (NIST 6012 or NIST 6013) (4,7). 16 Se

80 50 Analysis of Commercially Available Yeast Samples Nutritional yeast is sold commercially Intensity as a food product, but yeast with added 0 0 20 40 60 elements like selenium, calcium, chro- Time (min) mium, and others can be found in various pharmacies as food supplement products. Figure 5: (a) Time-resolved signal of phosphorus and (b) selenium for the measurement of a They should support the supply of essen- selenized yeast sample. 10 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com tial elements to the human body in case of deficiency. The control of quality of these products is essential, and here we applied the single cell ICP-MS approach to investi- 8000 (a) gate the cellular incorporation of the added (cts) + elements. As examples, the results on com- O 6000 16 mercially available yeast products enriched P in selenium and chromium are shown. 31 4000 The strategy was to determine the total number of cells by measuring phosphorus 2000 and then compare it with the number of cell Intensity 0 events in which the sought elements were 0 20 40 60 detectable (7). In order to discriminate be- Time (min) tween cell events and background signals, the procedure as described in the paragraph (b) data treatment was followed. 800 (cts) Typical time-resolved measurements + O 600 for the elements phosphorus and selenium 16 are shown in Figure 5. Phosphorus again Cr 52 served as a cell marker, and the resulting 400 ratio between cell events containing selenium and the number of cells detected by 200 monitoring phosphorus resulted in 62%. Intensity 0 That means that only two-thirds of the 0 20 40 60 yeast cells incorporated selenium, at least Time (min) above the detection limit for selenium (0.16 Figure 6: (a) Time-resolved signal of phosphorus and (b) chromium for the measurement of a fg per cell [7]). The detected amounts of se- yeast sample enriched in chromium. lenium per cell were between 0.5 and 30 fg. It can be concluded that an active cell incorporation occurred during the production of Conclusions (4) M. Corte-Rodríguez, R. Álvarez-Fernán- the so-called “selenized yeast.” The analysis of the content of trace elements dez García, E. Blanco, J. Bettmer, and M. In the case of the yeast enriched with at an individual cell level is possible using Montes-Bayón, Anal. Chem. 89, 11491– chromium, the same strategy was followed. single cell ICP-MS with a dedicated sample 11497 (2017). Figure 6 reflects the observations for mon- introduction system. The technique allows (5) P.E. Verboket, O. Borovinskaya, N. Meyer, itoring phosphorus and chromium in the to screen a high number of individual cells D. Günther, and P.S. Dittrich, Anal. Chem. single cell mode. It becomes clearly visible in a short period of time, so that excellent 86, 6012–6018 (2014). that much less cell events were observable counting statistics are achieved, allowing a (6) M. Corte-Rodríguez, E. Blanco-González, for the chromium trace. Furthermore, the true overview on whether a given element J. Bettmer, and M. Montes-Bayón, Anal. background signals were relatively high, is distributed homogenously within cell Chem. 91, 15532–15538 (2019). indicating the presence of dissolved chro- population, included in the cells or not, or (7) R. Álvarez-Fernández García, M. mium in the cell suspension, even after evaluation of differences in intake or me- Corte-Rodríguez, M. Macke, K.L. LeB- the applied washing steps. These findings tabolism under certain conditions. Further- lanc, Z. Mester, M. Montes-Bayón, and J. showed that either the yeast cells were not more, accurate quantification of the amount Bettmer, Analyst, (2020) DOI: 10.1039/ actively incorporating chromium, or the of metal per cell is possible after calibration. C9AN01565E cells were just mixed with an unknown This information can be valuable in a va- chromium species. riety of applications, such as structural bi- M. Corte-Rodríguez is with the In any case, these preliminary results ology, clinical research, and biotechnology. Institute for Analytical Chemistry at the demonstrate that single cell ICP-TQ-MS University of Vienna, in Vienna, Austria. can deliver important information in the References R. Álvarez-Fernández García, control of yeast-based products enriched (1) C. Degueldre and P.-Y. Favarger, Coll. P. García-Cancela, M. Montes- with elements. Apart from quantitative Surf. A Physicochem. Eng. Asp. 217, Bayón, and J. Bettmero are with data, it can provide a fast tool to distinguish 137–142 (2003). the Department of Physical and Analytical between incorporated and extracellular el- (2) F. Li, D.W. Armstrong, and R.S. Houk, Chemistry at the University of Oviedo, in ements, as shown on the example of chro- Anal. Chem. 77, 1407–1413 (2005). Oviedo, Spain. D.J. Kutscher is with mium. Finally, it might be generally useful (3) H. Wang, M. Wang, B. Wang, L. Zheng, Thermo Fisher Scientific, in Bremen, Germany. for controlling the production of commer- H. Chen, Z. Chai, and W. Feng, Anal. Bio- Direct correspondence to: daniel.kutscher@ cial products based on yeast. anal. Chem. 409, 1415–1423 (2017). thermofisher.com chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 11

Rapid Quantitation of PFAS in Non-Potable Waters Environmental scientists have documented a dramatic increase in the public interest about the occurrence of per- and polyfluoralkyl substances (PFAS) in water. Liquid chromatography mass spectrometry-based PFAS detection (LC–MS) has been established as the most suitable technology for monitoring these substances. Implementing robust methodologies for targeted monitoring, identification, and emergency response to emerging PFAS is challenging. Despite a vast PFAS-related scientific knowledge base, a lack of access and training for LC–MS instruments is a primary hurdle for expanded monitoring and identification. This paper presents data acquired according to draft method EPA 8327. This method requires minimal sample preparation and adequately assesses PFAS in various environmental samples, including groundwater, surface water, and wastewater. Ruth Marfil-Vega and Brahm Prakash

er- and polyfluoralkyl substances (PFAS) are a class of chain compounds has been confirmed by the changes of PFAS more than 5,000 man-made and commercially available fingerprint in wastewaters (7). P chemicals (1). PFAS are stable, and have water, oil, grease, The occurrence and fate of PFAS in the environment has been and heat-resistant properties. These substances have been exten- studied by the scientific community for two decades. Analytical sively used in the manufacturing of industrial (such as firefight- methods have been developed and normally validated for specific ing foams) and household (such as carpets, clothing, cookware, studies, including variable lists of target compounds. Liquid chro- and food packages) products since the 1940s. The properties of matography mass spectrometry (LC–MS) detection (either triple PFAS that assist manufacturers in producing durable products quadrupole or high-resolution instruments) is required for the ex- are also responsible for their persistence in the environment tremely low detection limits required for quantification of known and human body. Therefore, PFAS are resistant to biodegrada- PFAS and the identification of new ones in environmental samples. tion and other conventional remediation technologies. Possible In the United States, LC–MS-based PFAS methods have had a lim- health effects in humans and animals from exposure to PFAS ited reach in the environmental field outside of research laborato- have been reported, including increased risk of cancer, endo- ries. However, the recent emergence of public concern regarding crine, and development disruption (2,3). the presence of PFAS in the environment, especially in water, have Known PFAS are currently divided into ten subclasses (4). required laboratories to provide rapid and accurate results to rele- Perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic vant stakeholders. Figure 1 describes various workflows that labora- acid (PFOS) have been historically the most studied chemicals. tories may need to implement, depending on the end-goal for PFAS PFOA and PFOS belong to perfluoroalkyl carboxylic acids analysis. Nakayama and associates published an extensive review

(PFCAs; base structure: CnF2n+1COOH) and perfluoroalkane of the latest global trends for quantitation of PFAS and screening of sulfonic acids (PFSAs: CnF2n+1SO3H) classes, respectively. new compounds in 2019 (8). In this review, research methods allow Hexafluoropropylene oxide (HFPO), from the perfluoroether for more flexibility, especially as it applies to samples that can be carboxylic acids class (PFECAs; base structure: CnF2n+1-O-Cm- analyzed using a single method and pre-treatment step. F2m+1-COOH), also known as GenX, has received focused at- Over the past year, the US Environmental Protection Agency tention since its recent discovery in the environment (5). GenX (EPA) published several standardized methods to partially ad- was introduced in the market as a replacement of PFOA and dress the current needs of environmental labs and other relevant PFOS, together with other shorter chained PFAS, as a result of stakeholders. Namely, standardized methods that allow for faster the PFOA Stewardship Program. This program involved eight turn-around-times. In November 2018, EPA method 537.1 (De- major PFAS manufacturers in the United States that voluntarily termination of selected per- and polyfluorinated alkyl substances phased out PFOA and PFOS-related compounds by 2015 (6). in drinking water by solid phase extraction and liquid chroma- The shift in PFAS manufacturing processes towards the shorter tography–tandem mass spectrometry [LC–MS/MS]) (9) was pub- 12 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Materials and Methods Reagents

Potable Non-Potable Solid A commercial mixture of PFAS, includ- Waters Waters Samples ing targets and isotopically labeled compounds, from Wellington Laboratories Targeted Targeted Unknown PFAS Known PFAS Unknown PFAS (Discovery) were used (Catalog no. PFCA-24PAR and (Suspects) MPFCA-24ES). These standards were di-

Quantitation (*) Screening luted to working standards with 95:5 (v/v) acetonitrile:water, in accordance to EPA

Standardized Method 8327. Methods Research (EPA, ASTM) Methods Instrumental Conditions The LC–MS/MS analysis of 43 PFAS (24 targets and 19 surrogates, listed in Table I) High LC-Triple LC-Triple resolution was performed using a Shimadzu Nexera Quadrupole Quadrupole mass with low with high spectrometer UHPLC system coupled with LCMS-8050 sensitivity sensitivity (ex. QTOF) triple quadrupole mass spectrometer. A

INSTRUMENT SENSITIVITY detailed description of the LC–MS/MS

(*) Identiﬁcation and conﬁrmation of compounds in suspects or discovery analysis parameters is included in Table II. A Shim- requires library spectrum matching and comparison of rentention time with authentic standards. pack GIST Phenyl-Hexyl, 2.1 mm × 100 mm and 3.0-μm particle size (Shimadzu Figure 1: Guidance tool: select sample type, type of compounds, goal of analysis, and type of method. P/N: 227-30713-03) analytical column lished. EPA 537.1 was an update of EPA 537 current version of DOD’s Quality Systems was used to conduct the analysis, along (determination of selected perfluorinated Manual (Version 5.3) (12). with a Shimadzu Shim-pack XR-ODS 50 alkyl acids in drinking water by solid phase Draft method EPA 8327 (Per-and poly- mm × 3.0 mm x 2.2-μm (Shimadzu P/N: extraction and liquid chromatography–tan- fluoroalkyl substances [PFAS] using ex- 228-41606-92) as delay column. The delay dem mass spectrometry [LC–MS/MS]) (10) ternal standard calibration and multiple column was installed for minimizing pos- to include four additional compounds: HF- reaction monitoring [MRM] liquid chro- sible PFAS contamination from the solvents PO-DA (GenX), ADONA, 11Cl-PF3OUdS matography–tandem mass spectrometry being used. Mobile Phases A and B were the and 9Cl-PF3ONS. These two methods were [LC–MS/MS]) was published in the sum- same as those listed in EPA method 8327 for only validated for the analysis of drinking mer of 2019 (13). Draft method EPA 8327 a binary gradient: 20 mM ammonium ac- water; hence, their performance was not is similar to ASTM 7979 (Standard test etate in 95:5 water:acetonitrile and 10 mM demonstrated in non-potable waters. In method for determination of per- and ammonium acetate in 95:5 acetonitrile:wa- December 2019, the EPA published a new polyfluoroalkyl substances in water, ter for A and B, respectively. A 0.3 mL/min method for the analysis of PFAS in drinking sludge, influent, effluent, and wastewa- flow rate was used in the 21 min gradient, water, EPA method 533 (Determination of ter by liquid chromatography–tandem an allowed modification of the gradient per- and polyfluoroalkyl substances in mass spectrometry [LC–MS/MS]), origi- listed in EPA method 8327. The total run drinking water by isotope dilution anion nally published in 2016 (14). This method time of 21 min included re-equilibration for exchange solid phase extraction and liq- expands the list of target compounds as both the delay and the analytical column. uid chromatography–tandem mass spec- well as sample types available for analysis, This gradient (described in Table II) was trometry) (11). This method expanded including non-potable waters. The scope developed to ensure maximum resolution the target compounds to focus on the draft method EPA 8327 is the quantifi- between peaks in the shortest time possible shorted chain substances. Table I lists the cation of 28 selected PFAS with minimal with minimum co-elution of isomers. target compounds included in each of the sample preparation (similar to “dilute All compound specific parameters, in- methods published by EPA for analysis of and shoot” approaches) in non-potable cluding precursor ion, product ion, and PFAS in drinking water. The Department waters (ground water, surface water, and collision energies were optimized using of Defense (DOD) is also invested in the wastewater). The elimination of solid flow injection analysis (FIA), bypassing standardization of analytical methods for phase extraction in this method reduces the analytical column using Lab Solutions the assessment of PFAS in drinking water. analysis times and experimental bias, software. The lower ESI heater temperature While the DOD requires laboratories to use and minimizes PFAS contamination selected reduces HF loss and minimizes method EPA 537 or EPA 537.1 for drinking during sample pre-treatment. This paper false identification of fluorotelomer acids. water analysis, the DOD allows in-house presents the results from assessing PFAS The MRM transitions are listed in Table III. developed methods for the analysis of other in non-potable waters (groundwater, sur- A total of 66 MRMs was monitored. The environmental samples. These methods are face water, and wastewater) according targets monitored in this work are anionic, acceptable provided that the laboratories to draft method EPA 8327, and aims to hence, analyzed by electrospray ionization demonstrate adherence and compliance demonstrate the reliable and robust per- in negative mode. However, new research with the quality criteria established in the formance of the method. suggests cationic and zwitterionic PFAS chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 13

Table I: Target analytes included in methods EPA 537/537.1, EPA 8327, and ASTM 7979-19 EPA 8327/ Chemical Acronym CAS Number EPA 537.1 ASTM 7979-19 EPA 533 11-chloroeicosafluoro-3-oxaundecane-1-sulfonic acid 11Cl-PF3OUdS 763051-92-9 X (*) X Fluorotelomer sulphonic acid 4:2 4:2 FTS 757124-72-4 X X Fluorotelomer sulfonate 6:2 6:2 FTS 27619-97-2 X X Fluorotelomer sulfonate 8:2 8:2 FTS 39108-34-4 X X 9-chlorohexadecafluoro-3-oxanone-1-sulfonic acid 9Cl-PF3ONS 756426-58-1 X (*) X 4,8-dioxa-3H-perfluorononanoic acid ADONA 919005-14-4 X (*) X Perfluorooctanesulfonamide FOSA 754-91-6 X Hexafluoropropylene oxide dimer acid HFPO-DA 13252-13-6 X (*) X N-ethyl perfluorooctanesulfonamidoacetic acid NEtFOSAA X X N-methyl perfluorooctanesulfonamidoacetic acid NMeFOSAA X X Nonafluoro-3,6-dioxaheptanoic acid NFDHA 151772-58-6 X Perfluorobutanoic acid PFBA 375-22-4 X X Perfluorobutanesulfonic acid PFBS 375-73-5 X X X Perfluorodecanoic acid PFDA 335-76-2 X X X Perfluorododecanoic acid PFDoA 307-55-1 X X X Perfluorodecyl sulfonate PFDS 335-77-3 X Perfluoroheptanoic acid PFHpA 375-85-9 X X X Perfluoroheptyl sulfonate PFHpS 375-92-8 X X Perfluoro(2-ethoxyethane)sulfonic acid PFEESA 113507-82-7 X Perfluorohexanoic acid PFHxA 307-24-4 X X X Perfluorohexanesulfonic acid PFHxS 355-46-4 X X X Perfluor-3-methoxypropanoic acid PFMPA 377-73-1 X Perfluoro-4-methoxybutanoic acid PFMBA 863090-89-5 X Perfluorononanoic acid PFNA 375-95-1 X X X Perfluorononane sulfonate PFNS 17202-41-4 X Perfluorooctanoic acid PFOA 335-67-1 X X X Perfluorooctanesulfonic acid PFOS 1763-23-1 X X X Perfluoropentanoic acid PFPeA 2706-90-3 X X Perfluorohexane sulfonate PFPeS 68259-08-5 X X Perfluorotetradecanoic acid PFTeDA 376-06-7 X X Perfluorotridecanoic acid PFTrDA 72629-94-8 X X Perfluoroundecanoic acid PFUnA 2058-94-8 X X X (*) Compounds included in EPA 537.1; all others included in both EPA 537 and 537.1 might ionize better in positive mode ng/L, with the injection solvent consisting with isotopically labeled surrogates, and (15). The ultra-fast acquisition rate of 555 of 50:50 water:methanol, with 0.1% acetic vortexed for 2 min. The samples were MRM/sec of the LCMS-8050 used in this acid in order to match the injection solvent then filtered through 0.2 μm syringe work allows for the inclusion of more tar- for the extracted samples. Filtration was not filters, and analyzed by LC–MS/MS. gets (to be analyzed in positive and neg- performed on the calibration standards. ative mode), without compromising the Quality control samples (QCs) were pre- Results and Discussion sensitivity in a single run. Additional de- pared as the calibration standards. Calibration Curve and tails about the method and configuration Quality Control Samples can be found elsewhere (16). Sample Preparation Data presented in this paper were acquired Replicated samples, including reagent in two batches, and analyzed in consecutive Calibration Curve and water, groundwater, surface water, and dates. Each batch included the analysis of 39 Quality Control Samples wastewater were analyzed during this injections (9 calibration standards, 4 blanks The working standards were used to create study. Each sample was diluted 50:50 with [reagent, method, and travel blanks], 6 QCs a calibration curve ranging from 5 to 200 methanol and 0.1% acetic acid, spiked [LLOQ, LCS, and CCV] injected along the 14 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Table II: Instrumental Conditions (Shimadzu LCMS-8050) Parameter Value Shim-pack XR-ODS (3 mm ID. x 50 mm L., 2.2-μm) Delay column P/N 228-41606-92 Shim-pack GIST Phenyl-Hexyl (2.1 mm ID. x 100 mm L., Analytical column 3-μm) P/N 227-30713-03 Column oven temperature 40 °C Injection volume 30 µL A: 20 mmol ammonium acetate in 5 % (v/v) acetonitrile Mobile phase in reagent water; B: 10 mmol ammonium acetate in 95 % (v/v) acetonitrile in reagent water

LC Flow rate 0.3 mL/min (Nexera) Time (minutes) % B 0 0 1 20 6 50 Gradient 14 100 17 100 18 0 21 0 Run time 21 minutes Nebulizing gas flow 5 L/min Heating gas flow 15 L/min Interface temperature 300 °C MS/MS Desolvation line temperature 100 °C (Shimadzu LCMS-8050) Heat block temperature 200 °C Drying gas flow 5 L/min Acquisition cycle time 21 min Total MRMs 66 sequence according to EPA requirements, verification standards (QCs) were injected for the PFCAs surrogates was observed for and 20 samples [reagent water, ground in the two batches consecutively analyzed the targets. For the targets, including the water, surface water and wastewater]). for this study. Surrogates and targets were other classes not shown in Figure 4, results Calibration curves were performed for spiked at 80 ng/L (in sample equivalent). were well within EPA’s acceptable range all PFAS targets, using a nine-point calibra- Figure 3 summarizes the concentrations (±30% of the expected value) and even tion curve, ranging from 5 to 200 ng/L. Fig- of all surrogates, grouped by PFAS class, the stricter bracket of ±20%. The percent ure 2 shows a total ion chromatogram and from EPA 8327 in the eight QCs samples; relative standard deviation was <20% for MRMs from a 5 ng/L standard; this figure the red solid line represents the true concen- all compounds (results not included here) demonstrates the separation and peak tration (80 ng/L), and the dotted red lines in the two batches analyzed. The results shape of targets at the lowest concentration set the limits of the acceptable concentra- summarized in Figures 3 and 4 demon- included in the calibration curve. tion range according to EPA guidelines strate the robustness and reproducibility of The linearity of the curve was deter- (±30% of the expected value, equivalent the method validated in this work. It is also mined using a 1/x weighting factor, and to 56 and 104 ng/L). For the PFCAs class, noteworthy that maintenance was not re- not forcing through zero. Excellent linearity the concentration range increases with quired for maintaining the performance of was obtained with correlation coefficients carbon chain length; no clear trends were the instrument after the analysis of at least (r2) greater than 0.99 for all analytes or tran- observed for the other classes of PFAS. 40 environmental samples (approximately sitions in the two batches analyzed for col- All results, except for one injection of 28 h of operation). lecting the data presented in this paper. Re- d3-NMeFOSAA, were within ±30% of the siduals of each standards were within ±30% expected value, and, moreover, most to the Environmental Samples difference from its true concentration. More results were within ±20% of the expected Forty environmental samples were ana- detailed information about the calibration value. Figure 4 summarizes the concen- lyzed in two consecutive batches. Figure curve results can be found elsewhere (16). trations of the eleven PFCAs included in 5 summarizes the surrogate (spiked at 160 A total of eight continuing calibration the method. Similar trend to that observed ng/L) recoveries from 16 representative chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 15

Table III: MRM transitions, retention times and collision energies of compounds from method EPA 8327 Compound Retention Time (minutes) Transition (m/z) Collision energy (V) PFBA 3.34 213 > 169 9 MPFBA 3.34 217 > 172 9 PFPeA 3.94 263 > 219 8 M5PFPeA 3.94 268 > 223 8 327 > 307 18 4-2 FTS 4.44 327 >81 35 M4-2 FTS 4.44 329 > 309 20 313 > 269 9 PFHxA 4.68 313 >119 21 M5PFHxA 4.68 318 > 273 11 299 > 80 30 PFBS 4.71 299 >99 28 M3PFBS 4.81 302 > 80 34 363 > 319 9 PFHpA 5.40 363 >169 16 M4PFHpA 5.40 367 > 322 10 349 > 80 42 PFPeS 5.61 349 >99 30 427 > 407 23 6-2 FTS 5.80 427 >81 39 M6-2 FTS 5.80 429 >409 22 413 > 369 10 PFOA 6.05 413 >169 17 M8PFOA 6.05 421 > 376 10 399 > 80 43 PFHxS 6.30 399 >99 22 402 > 80 49 M3PFHxS 6.31 403 >84 49 463 > 419 11 PFNA 6.64 463 >219 16 M9PFNA 6.64 472 > 427 12 527 > 507 26 8-2 FTS 6.93 527 >81 49 529 > 509 26 M8-2 FTS 6.93 527 >81 49 449 > 80 51 PFHpS 6.93 449 >99 37 570 > 419 21 N-MeFOSAA 7.25 570 >483 16 d3M N-MeFOSAA 7.24 573 > 419 20 513 > 468.9 11 PFDA 7.19 413 >219 17 M6PFDA 7.19 519 > 474 11 584 > 419 20 N-EtFOSAA 7.46 584 >483 16 M N-EtFOSAA 7.46 589 > 419 21 499 > 80 54 PFOS 7.48 499 >99 38 M8PHOS 7.48 507 > 80 55 563 > 519 12 PFUdA 7.70 563 >269 16 M7PFUdA 7.69 570 > 525 12 549 > 80 54 PFNS 8.01 549 >99 44 Table III Continued on 16 16 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Table III Continued From Page 15 613 > 569 12 PFDoA 8.18 613 >169 21 MPFDoA 8.18 615 > 570 11 FOSA 8.50 498 > 78 43 M8FOSA 8.50 506 > 78 48 599 > 80 55 PFDS 8.52 599 >99 50 663 > 619 12 PFTriA 8.66 663 >169 27 713 > 669 13 PFTeDA 9.15 713 >169 27 M2PFTeDA 9.13 715 > 670 15

within the acceptable ranges (70-130%). The percent recoveries in the groundwa- 1200 ter samples were consistently higher than 1100 1000 in the other sample types; however, they 900 were still within limits accepted by EPA, 800 700 and ranged between 95% and 123%, except 600 for M2PFTreA (138%). As samples were 500 400 analyzed randomly, the difference could 300 unts au be attributed to matrix enhancement in 200 100 the groundwater samples. EPA method 0 8327 utilizes an external calibration; the 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 matrix enhancement could be minimized ime min by using an internal calibration or isotopic Figure 2: TIC (black) chromatograms and MRM transitions (other colors) of all PFAS in EPA dilution for calculating the PFAS concen- Method 8327 at a concentration of 5 ng/L. trations. The %RSDs were less than 10% (half of the acceptable criteria listed in EPA method 8327), and no major differences were observed across samples types. The

MPB reproducibility in the results from the MPPe surrogate recoveries corroborates the ro- MP MP bustness of the system inferred from the MP results of the QCs. MP MP MPn Conclusions MP Multiple analytical methods for the mon- MPre itoring of PFAS have been developed and MPB MP validated over the past two decades. How-

nentratin ng MP ever, one of the challenges environmental M-- M-- stakeholders face for understanding the M-- global occurrence of PFAS and the ef- d-Me fectiveness of mitigation strategies is the d-t M lack of standardized methods. Secondly, the recent increase in PFAS by the public urrgates requires laboratories to provide results Figure 3: Concentration of surrogates included in method EPA 8327 measured in replicated QCs quickly. This paper demonstrates the re- standards (n=8) injected during two consecutive batches (total number of injections, including liable and robust quantification of PFAS, calibration standards, QCs samples and samples: 78). according to latest EPA method for the analysis of environmental samples differ- samples collected from various sources 8327. Samples were analyzed in a random ent from drinking water, with minimal (reagent water, groundwater, surface order in the two batches. Surrogates %re- sample preparation; hence, it allows labo- water, and wastewater), and prepared as covery ranged between 82% and 92% in ratories to increase the sample throughput, described in the Materials and Methods reagent water, groundwater, surface water, and ensure stakeholders receive informa- section, in accordance to EPA method and wastewater; these values are well tion in a timely manner. chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 17

(10) J.A. Shoemaker and D.R. Tettenhorst. Method 537.1: Determination of selected perfluorinated alkyl acids in drinking PB water by solid phase extraction and liq- PPe uid chromatography/ tandem mass spec- P trometry (LC/MS/MS). (US EPA, Office of P P Research and Development, Washington, P D.C., Version 1.0, November 2018). P (11) L. Rosenblum and S. Wendelken. Method Pn 533: Determination of Per- And Polyfluo- nentratin ng P roalkyl Substances in Drinking Water by Pri Isotope Dilution Anion Exchange Solid Pre Phase Extraction and Liquid Chromatog- raphy/Tandem Mass Spectrometry. (US Ps EPA, Office of Research and Development, Figure 4: Concentration of PFCAs included in method EPA 8327 measured in replicated QCs Washington, D.C., Version 1.0, December standards (n=8) injected during two consecutive batches (total number of injections, including 2019). calibration standards, QCs samples and samples: 78). (12) US Department of Defense, Department of Defense Consolidated Quality Systems Manual (QSM) for Environmental Labo- Reagent ater rundater urfae ater asteater ratories Version 5.3. Available at https:// www.denix.osd.mil/edqw/documents/ manuals/qsm-version-5-3-final/. Accessed January 2nd, 2020. (13) US Environmental Protection Agency,

Reer Method 8327 Per- and polyfluoroalkyl substances (PFAS) using external standard calibration and multiple reaction monitoring (MRM) liquid chromatography/ tandem mass spectrometry (LC/

MPB MPB MP MP MP MP M MS/MS). Available at https://www.epa. MPPe MP MP MP MPnMP MPre M-- M-- M-- d-t d-Me gov/sites/production/files/2019-06/documents/proposed_method_8327_proce- Figure 5: Percent recovery of surrogates with corresponding %RSD in replicated (n=4) samples dure.pdf. Accessed January 2nd, 2020. from each matrix type: reagent water, groundwater, surface water, and wastewater. (14) ASTM D7979-19, Standard Test Method References (6) US Environmental Protection Agency, Fact for Determination of Per- and Polyfluoroal- (1) US Food and Drug Administration, Per and Sheet: 2010/2015 PFOA Stewardship Pro- kyl Substances in Water, Sludge, Influent, Polyfluoroalkyl Substances (PFAS). Available gram. Available at https://www.epa.gov/ Effluent, and Wastewater by Liquid Chro- at https://www.fda.gov/food/chemicals/ assessing-and-managing-chemicals-un- matography Tandem Mass Spectrometry and-polyfluoroalkyl-substances-pfas. Ac- der-tsca/fact-sheet-20102015-pfoa-stew- (LC/MS/MS). (ASTM International, West cessed January 2nd, 2020. ardship-program#what. Accessed January Conshohocken, PA, 2019) (2) US Environmental Protection Agency, PFOA, 2nd, 2020. (15) G. Munoz, P. Ray, S. Mejia-Avendaño, S.V. PFOS and Other PFASs. Available at https:// (7) E.F. Houtz, R. Sutton, J.S. Park, and M. Sed- Duy, D.T. Do, J. Liu, and S. Sauvé, Anal. www.epa.gov/pfas/basic-information-pfas. lak, Water Res. 95, 142–149 (2016). Chim. Acta, 1034, 74–84 (2018). Accessed January 2nd, 2020. (8) S.F. Nakayama, M. Yoshikane, Y. Onoda, (16) Analysis of Per- and Polyfluoroalkyl Sub- (3) Centers for Disease Control and Prevention, Y. Nishihama, M. Iwai-Shimada, M. Takagi, stances (PFAS) Specified in EPA M8327 Per- and Polyfluorinated Substances (PFAS) Y. Kobayashi, and T. Isobe, Trends Anal. using the LCMS-8050 Triple Quadrupole Factsheet. Available at https://www.cdc.gov/ Chem. 151, 115410 (2019). Mass Spectrometer, Shimadzu Scientific biomonitoring/PFAS_FactSheet.html. Ac- (9) J.A. Shoemaker, P.E. Grimmett, and B.K. Instruments Application Note. cessed January 2nd, 2020. Boutin. Method 537: Determination of (4) M. Sun, E. Arevalo, M. Strynar, A. Lindstrom, selected perfluorinated alkyl acids in Ruth Marfil-Vega is the Environmental M. Richardson, B. Kearns, A. Pickett, C. drinking water by solid phase extraction Marketing Manager at Shimazdu Scientific Smith, and R.U. Knappe. Environ. Sci. Technol. and liquid chromatography/ tandem Instruments, in Columbia, Maryland. Brahm Lett. 3(12), 415–419 (2016). mass spectrometry (LC/MS/MS). (US Prakash is a Strategic Scientist at Shimazdu (5) Z. Wang, J.C. DeWitt, C.P. Higgins, and I.T. EPA, Office of Research and Develop- Scientific Instruments, in Columbia, Maryland. Cousins, Environ. Sci. Technol. 51, 2508– ment, Washington, D.C., Version 1.1, Direct correspondence to: rmmarfilvega@shi- 2518 (2017). September 2009). madzu.com 18 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Affinity Selection-Mass Spectrometry: Defining the Bioactive Compounds in Complex Mixtures of Natural Products and Combinatorial Libraries

Drug discovery from combinatorial libraries typically utilizes high-throughput screening of discreet compounds, and the discovery of natural products with pharmacological mechanisms of action relies on bioassay-guided fractionation. Both processes can be expedited through the application of affinity selection-mass spectrometry (AS-MS). AS-MS includes a family of MS-based affinity screening methods, including pulsed ultrafiltration (PUF)-AS-MS, size exclusion chromatography AS-MS, and magnetic microbead affinity selection screening (MagMASS). All AS-MS approaches begin by incubating a pharmacologically important receptor with a mixture of possible ligands, separating the ligand-receptor complexes from non-binding molecules (the approaches differ in this separation step), and then using LC–MS to characterize the affinity-extracted ligands. The speed, selectivity, and sensitivity of mass spectrometry and ultrahigh-pressure liquid chromatography (UHPLC)-compatible MS ionization techniques, like electrospray and atmospheric pressure chemical ionization, make AS-MS ideal for characterizing ligands. Recent applications of AS-MS include allosteric as well as orthosteric ligand discovery, and finding ligands to membrane-bound proteins and RNA targets.

Richard B. van Breemen

he development of combinatorial chemistry in the approaches, usually involving multiwall plate technol- 1990s facilitated the rapid synthesis of large num- ogies, which enabled the screening of more than 10,000 T bers of drug-like molecules, called combinatorial compounds per day. Today, most high-throughput libraries, for evaluation in pharmacological assays. The screening approaches test one compound at a time using subsequent demand for faster and less expensive assays rapid assays based on absorption or fluorescence such drove the development of high-throughput screening as fluorescence polarization (1), but growing in pop- ADVANTAGE See What It Can Do for You and Your Lab

• Technical Articles • Education & Instruction & Applications • Online Tools & Calculators • Videos & ChromaBLOGraphy • Product Selection • FAQs & Troubleshooting Assistance

Pure Chromatography 20 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Table I: Comparison of features and capabilities of the AS-MS techniques size exclusion (SEC) AS-MS, pulsed ultrafiltration (PUF) AS-MS, and magnetic microbead affinity selection screening (MagMASS). Feature SEC PUF MagMASS Rapid natural products screening √ √ √ Rapid combinatorial library screening √ √ √ No interference from fluorescent or UV chromophores √ √ √ No radioisotope or fluorescent labels required √ √ √ Minimum consumption of receptors and cofactors √ √ √ Compatible with any binding buffer √ √ √ Targets include any macromolecule (proteins, RNA, and similar compounds) √ √ √ Allosteric ligand screening √ √ √ Combinatorial library mixture screening √ √ √ Natural product mixture screening √ √ √ Solution-phase receptor screening √ √ X Solid-phase (insoluble) receptor screening X X √ Membrane-bound receptor screening √ √ √ Ranking of ligand mixtures by affinity to receptor √ √ √

- Ligand - Non-ligand 2D CHROMATOGRAPHY - Receptor

m B RR Extract + Receptor RR Denature complexes & separate ligands UHPLC-MS & MS/MS

Retention time B RR R RR Extract + Receptor Remove unbound Elute ligands compounds

Figure 1: Comparison of SEC-AS-MS and PUF-AS-MS screening. After incubation of a receptor with a mixture of compounds, the large ligand– receptor complexes are separated from the small unbound compounds using either size exclusion chromatography (SEC) or ultrafiltration. The isolated ligand-receptor complexes are denatured to release the ligands either off-line or during the final UHPLC–MS/MS analysis step. ularity are higher-throughput mass advantage is particularly important vance knowledge of the identities of spectrometry (MS)-based screening when screening natural products, the compounds in a mixture, these approaches that screen combinato- which often contain strong chromo- screening methods are ideal for nat- rial library mixtures. phores. The need for more chemi- ural products. Although the utility By relying on mass selectivity, MS- cal diversity in screening programs of mass spectrometry for natural based screening eliminates the need than is available from combinatorial products drug discovery was recog- for radiolabels or chromophores. In libraries has also renewed inter- nized at least 20 years ago (2), most particular, false positive results that ests in natural products as sources MS-based drug discovery assays can be caused by test compounds of diverse chemical structures. In used in the pharmaceutical industry that exhibit interfering absorbance addition, since mass spectrome- remain focused on combinatorial or fluorescence are avoided. This try-based assays do not require ad- library screening (3). chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 21

The discovery of natural products with specific pharmacological mechanisms of action has been labor 100 RXRα-MBP intensive, time consuming, and not N easily adapted to a high-through- MBP Control No protein Control O put formats. The standard approach C today is bioassay-guided fraction- LG100268 OH 50 ation (4), which begins with a phar- Ketoconazole macologically active natural prod- Internal standard uct mixture such as an extract of a plant, fungus, or a microbial cul- reaive rene 0.0 ture. The extract is partitioned with 1.0 2.0 3.0 solvents, and the active partition is Reenin ime min fractionated using chromatography. After assaying each fraction, the ac- Figure 2: SEC-AS-MS screening for ligands of retinoid X receptor (RXR)-α expressed recombinantly tive fraction is subfractionated and with a maltose binding protein (MBP) tag. The known ligand LG100268 (used as a positive re-assayed, until active compounds control) was detected in the assay using RXRα-MBP but not in the controls using MBP alone or no are isolated for spectroscopic anal- protein. Ketoconazole was added after the SEC separation as an internal standard to normalize ysis and identification. The appli- and compare the reversed phase UHPLC–MS chromatograms. cation of mass spectrometry-based screening can expedite this process gands to receptors and enzymes. The large ligand-receptor complexes by condensing the reiterative frac- Other AS-MS approaches that have elute first. tionation/bioassay process into not yet been so widely adopted in- The ligands are released from the a single step. Because these mass clude affinity chromatography-MS receptor using organic solvent or a spectrometry-based screening as- (9), affinity capillary electrophore- pH change that denatures the re- says share a common affinity sep- sis-MS (10), frontal affinity chro- ceptor. This denaturing step can be aration step (5), they have become matography-MS (11), direct affinity carried out prior to LC–MS analysis, known as affinity selection-mass screening of ligand-receptor com- or directly on the reversed-phase spectrometry (AS-MS). plexes using electrospray ionization column. For maximum sensitivity, and ultrahigh resolution MS (12), 2-dimensional SEC-reversed-phase Affinity Selection-Mass and matrix-assisted laser desorption chromatography is used, in which Spectrometry (AS-MS) MS of ligands affinity-captured by the ligand-receptor complexes A variety of mass spectrome- an immobilized target (13). eluting from the SEC column are try-based screening methods have directed on-line onto the reversed been invented based on the affinity Size exclusion -phase column for the final LC–MS of active compounds for a pharma- Chromatography (SEC)-AS-MS or UHPLC–MS step. Besides serving cological target. All of these meth- During SEC-AS-MS, a mixture to separate the ligands from each ods utilize affinity interaction (the of compounds is incubated with a other and from the receptor, the re- binding of a ligand to a receptor) pharmacological receptor and al- versed-phase chromatographic sep- to facilitate the isolation of active lowed to come to equilibrium (Fig- aration also removes the binding compounds from complex mixtures, ure 1). Any binding buffer may be buffer, which is often incompati- and then MS to characterize the af- used, including non-volatile buffers ble with mass spectrometric anal- finity-extracted ligands. The most that are incompatible with mass ysis. The SEC separation must be popular and successful of these spectrometry, since the buffer will carried out quickly and at reduced methods include size-exclusion be removed during the final stage of temperatures (usually 4 °C), because chromatography (SEC)-AS-MS (6), reverse phase LC–MS analysis. The ligands begin to dissociate from the pulsed ultrafiltration (PUF)-AS-MS ligand-receptor complex, which is receptor immediately and can be- (7), and magnetic microbead affin- much larger than the unbound li- come lost into the size exclusion ity selection MS (MagMASS) (8). In gands, is separated from the low stationary phase. addition to the speed, selectivity, mass, non-binding compounds SEC-LC–MS was invented by and sensitivity of mass spectrom- using SEC (14). Also known as gel Kaur and co-workers (15), and was etry, the availability of LC-com- permeation chromatography, SEC originally used to screen peptide li- patible MS ionization techniques separates molecules according to braries for ligands of a protein tar- like electrospray and atmospheric size as they pass through a station- get. Since then, SEC-AS-MS been pressure chemical ionization have ary phase containing particles with automated to include 2-D chroma- enabled these AS-MS methods to pores that allow small molecules to tography (SEC-reversed-phase) and characterize a wide variety of li- enter but exclude large molecules. has been applied to a wide variety of 22 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

As a natural products application of PUF-AS-MS, an extract of ginger (Zingiber officinale Roscoe) was screened for ligands of the in- 100 O OH 10-gingerol flammation target cyclooxygenase-2 MeO 13.7 (CH2)8CH3 (COX-2) (23) (Figure 3). Ginger roots 10-gingerol HO have been used traditionally to treat inflammation, and had been re- 8-gingerol ported to inhibit COX. To identify 50 9.8 the pharmacologically active com- Control pounds in ginger, PUF-AS-MS was 6-shogaol COX-2 used to screen a chloroform partition 10.8 of a methanol extract of ginger roots reaive rene 6-gingerol for COX-2 ligands. A series of gin- 6.0 gerols and shogaols were determined 0 using PUF-AS-MS to bind to COX-2 5 10 15 20 (Figure 3). Competitive binding ex- Reenin ime min periments using PUF-AS-MS with ginger and the known synthetic in- Figure 3: PUF-MS screening of an extract of ginger for ligands to the inflammation target hibitor celecoxib indicated that the cyclooxygenase-2 (COX-2). After pulsed ultrafiltration, the ultrafiltrates of experiment (incubations ginger ligands bound to the active with functional COX-2) and control (incubation without receptor) were analyzed using negative site of COX-2. Additional functional ion electrospray mass spectrometry with reversed phase HPLC separation. Gingerols and enzyme assays confirmed that these shogaols were detected as COX-2 ligands based on enhancement of the LC–MS signal relative to ginger compounds inhibited COX-2 the control. (Adapted from van Breemen, et al. [23] and used with permission.) but not COX-1 with IC-50 values in the low micromolar range. targets (16,17). In this way, pharma- ground noise of SEC-AS-MS assays Invented in the van Breemen ceutical researchers can screen over is characteristically lower than other laboratory in the 1990s (24), PUF 100,000 combinatorial library com- AS-MS screening approaches. In this -AS-MS has been adopted for use pounds per day, in mixtures rou- example, UHPLC–MS was used in- by both academic (25,26) and in- tinely containing up to 2,600 com- stead of HPLC for faster separations dustrial researchers (27) to screen pounds at a time (18,19). In natural (Figure 2). combinatorial libraries in pools of product applications, SEC-AS-MS up to 2,700 compounds each (28), has been used to identify pharma- Pulsed Ultrafiltration as well as natural product extracts cologically active compounds in (PUF)-AS-MS (29, 30) for the discovery of ligands complex botanical extracts such PUF-AS-MS screening begins with to a wide variety of macromolecular as traditional Chinese medicines the incubation of a mixture of po- receptors (31,32). Advantages and (20,21). Some of the advantages and tential ligands with a macromolec- disadvantages of PUF-AS-MS are disadvantages of SEC-AS-MS are ular target in solution, such as an listed in Table I summarized in Table I. enzyme or receptor. The ligand-re- As an example of SEC-AS-MS ceptor complexes are then sepa- Magnetic Microbead Affinity data, reversed-phase LC–MS chro- rated from the unbound compounds Selection Screening (MagMASS) matograms for the screening for using ultrafiltration. As with SEC- A newer AS-MS method, MagMASS ligands of the nuclear receptor, ret- AS-MS, the ultrafiltration separa- was first reported in 2008 by Choi inoid X receptor (RXRα), are shown tion of bound ligands from unbound and van Breemen (33) for screening in Figure 2. RXRα is under inves- compounds should be carried out combinatorial library mixtures and tigation as a drug discovery target quickly, and at reduced temperature, natural product extracts for ligands for anti-inflammation therapy as to minimize dissociation and loss of to an immobilized target. Recently, well as for cancer treatment and ligand. The ligand-receptor com- MagMASS became the first AS-MS prevention. The chromatograms plexes are denatured, usually using approach to incorporate multititer represent two negative control incu- water and acetonitrile acidified with well plates, which facilitated auto- bations and a positive control. Note a volatile organic acid such as formation and enhanced the through- the intense chromatographic peak mic acid, and the released ligands put (8,34). MagMASS has also been for the known ligand LG100268 in are trapped on a reversed-phase applied to the screening of complex the positive control, and almost no HPLC or UHPLC column for char- botanical extracts (34). signal or background for the neg- acterization using (UHPLC–MS and The MagMASS process (Figure 4) ative control incubations. The back- MS/MS (Figure 1) (22). begins by immobilizing a pharma- chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 23 cological receptor to magnetic beads either covalently or non-covalently. A variety of bead surface chemis- tries are available from commercial sources for covalently immobilizing - Ligand receptors. Non-covalent immobili- - Non-ligand zation is also possible by utilizing - Receptor interactions such as immobilized - Magnetic microbead streptavidin binding biotinylated receptors, immobilized nickel ions Experiment binding with receptors containing Int. Std. His-tags, and immobilized amylose Control interacting with receptors contain- Magnet Magnet ing maltose binding protein. Retention time After immobilization of the re- ncae ah e ceptor, the magnetic microbeads are incubated with mixtures of potential ligands (Figure 4). A washing Figure 4: For MagMASS, the receptor/target is immobilized on magnetic microbeads, which are step is used to remove the unbound incubated with a compound mixture containing possible ligands. A magnetic field is applied to compounds while a magnetic field separate the beads containing the receptor-ligands complexes from the unbound compounds, retains the beads containing the re- which are washed away. The ligand-receptor complexes are then disrupted using a denaturing ceptor-ligand complexes. Ligands solution. A magnetic field is used to separate the beads from the released ligands, and the ligands are then released using a denaturing are analyzed using UHPLC–MS/MS. solvent or a pH change for analysis using (UHPLC–MS and (UHPLC– binatorial libraries, the masses and additional structure determination MS/MS. To control for non-specific elemental compositions of each using spectroscopic techniques such binding, beads without receptor are compound are already known and as NMR. usually incubated in parallel. See may be entered into a database. Table I for a summary of advantages By using high resolution accurate Comparison of AS-MS and disadvantages of MagMASS and mass measurements during screen- Approaches and Novel comparison with SEC-AS-MS and ing, the elemental compositions and Applications PUF-AS-MS. isotope patterns of ligands may be Advantages of all AS-MS methods As an example of applying Mag- compared with the database either include the diversity of receptors MASS to the screening of complex manually or automatically using that may be screened, the ability botanical extracts, human estrogen commercially available metabolo- to screen complex natural product receptor-β was immobilized cova- mics software. For additional con- extracts as well as combinatorial li- lently on aldehyde functionalized firmation of ligand identity and braries, compatibility with any in- magnetic beads and used to screen for distinguishing among isomeric cubation buffer, low consumption of for botanical estrogens (Figure 5) compounds, tandem mass spec- receptors and cofactors, elimination (33). Following incubation with extra may be compared with those of of radiolabels and chromophores, tracts of hops (Humulus lupulus L.), standards. lack of interference from matrix, the non-binding compounds were The identification of ligands and lower costs and faster screening washed away using buffer, and then from natural product extracts is than the current standard of screen- bound ligands were released by de- more challenging, but also begins ing discreet compounds (Table I). naturing the protein with methanol. with high resolution MS and MS/ Due to the procedural step of sep- The natural product 8-prenylnar- MS analyses to determine elemen- arating ligand-receptor complexes ingenin was identified as the most tal composition and structural fea- from unbound compounds, all active estrogen in hops (Figure 5). tures. By comparing the measured AS-MS methods share the common elemental compositions and tandem disadvantage of reduced sensitiv- AS-MS Data Analysis mass spectra with those of known ity for rapidly dissociating ligands. For all AS-MS approaches, high natural products in databases in a The reversed-phase chromatography resolution tandem mass spectrom- process known as dereplication, li- separation is a relatively slow step in eters are preferred, such as quad- gand structures may be determined the AS-MS process, but this step has rupole time-of-flight or orbitrap quickly. Reversed-phase UHPLC been accelerated by implementing mass spectrometers, so that elemen- retention times, obtained during UHPLC in place of HPLC. tal compositions of ligands may be screening step, enable novel natural Although SEC-AS-MS and PUF- determined. When screening com- product ligands to be collected for AS-MS enable solution-phase 24 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

technology that facilitates the rapid discovery of natural products with 2e6 pharmacological activities from complex mixtures such as botanical and microbial extracts. By Estrogen receptor- eliminating the need for spectro-

Negative Control scopic detection, AS-MS eliminates 2e6 Xanthohumol many limitations of conventional m 339 high-throughput screening, such as matrix interference or false pos- itives due to strong absorbance or 1e6 fluorescence of test compounds. In addition, AS-MS offers a variety or unique applications that are not easily addressed or have not 5e5 8-Prenylnaringenin been achieved using conventional high-throughput screening such as reaive rene 6-Prenylnaringenin

Isoanthohumol discovering allosteric ligands, ligands to RNA, and ligands to mem- Naringenin (internal standard) brane-bound receptors. Although 0 invented over 20 years ago (15,24), 12 14 16 18 20 22 24 26 28 many of the unique applications Reenin ime min of AS-MS have only recently been demonstrated. Therefore, expect Figure 5: An extract of hops (Humulus lupulus L.) was screened for ligands to the human many more innovations and natu- estrogen receptor-β (ER-β) using MagMASS with positive ion electrospray LC–MS. The prenylated ral products applications of AS-MS flavonone 8-prenylnaringenin was identified as the most potent estrogen in the extract. (Adapted during the next 20 years. from Choi and van Breemen [33] and used with permission.) References screening, MagMASS requires re- in order of affinity for the recep- (1) W.A. Lea, and A. Simeonov, Expert ceptor immobilization, which has tor. Note that the conventional Opin. Drug Discov. 6, 17–32 (2011). the potential disadvantage of al- approach to ranking ligands for (2) J. Liu, J.E. Burdette, H. Xu, C. Gu, tering binding properties (Table binding affinity requires individ- R.B. van Breemen, K.P. Bhat, N. I). On the other hand, receptor ual assays of each compound for Booth, A.I. Constantinou, J.M. Pez- immobilization for MagMASS can competition with a radiolabeled zuto, H.H. Fong, N.R. Farnsworth, facilitate the screening of mem- ligand. Unlike competition bind- and J.L. Bolton, J. Agric. Food Chem. brane-bound or poorly soluble re- ing assays, AS-MS may be used to 49, 2472–2479 (2001). ceptors. In an online configuration discover ligands that bind to al- (3) S.S. Walker, D. Degen, E. Nickbarg, of PUF-AS-MS that uses MS-com- losteric sites (37,38), and provide D. Carr, A. Soriano, M. Mandal, R.E. patible buffers without column lead compounds for targets pre- Painter, P. Sheth, L. Xiao, X. Sher, chromatography, PUF-AS-MS viously considered “undruggable” N. Murgolo, J. Su, D.B. Olsen, R.H. has the advantage of enabling the (39). Use of enzymes and receptors Ebright, and K. Young, ACS Chem. measurement of affinity constants, embedded in microsomes (40) or Biol. 12, 1346–1352 (2017). stoichiometry of binding, and en- cell membranes (41,42) has also (4) M.G. Weller, Sensors (Basel) 12, zyme kinetics for ligand–receptor been demonstrated. AS-MS is not 9181–9209 (2012). interactions (35,36). limited to screening for ligand to (5) Y.G. Shin, and R.B. van Breemen, AS-MS offers substantial im- protein receptors, as other mac- Biopharmaceut. Drug Dispos. 22, provements in speed and conve- romolecular targets such as RNA 353–372 (2001). nience over alternative screening molecules have been used (43,44). (6) F. Touti, Z.P. Gates, A. Bandyopd- approaches, and many novel ap- hyay, G. Lautrette, and B.L. Pen- plications have been reported just Conclusions telute, Nat. Chem. Biol. 15, 410–418 recently (Table I). For example, In addition to providing a faster (2019). AS-MS may be used to screen an and less expensive alternative to (7) B.M. Johnson, D. Nikolic, and R.B. equimolar mixture of natural prod- conventional screening of discreet van Breemen, Mass Spectrom. Rev. ucts (31) or combinatorial library compounds for interactions with 21, 76-86 (2002). compounds (37) and, in a single receptors, AS-MS addresses the (8) M.D. Rush, E.M. Walker, G. Prehna, experiment, rank order the ligands important and yet unmet need for T. Burton, and R.B. van Breemen, J. chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 25

Am. Soc. Mass Spectrom. 28, 479– (21) F. Fei, H. Sun, X. Cheng, J. Liu, J. Chem. Soc. 126, 15495–15503 448 (2017). Wang, Q. Li, and Y. Zhang, Chin. (2004). (9) M.A. Kelly, H. Liang, I.I. Sytwu, I. Med. 13, 49 (2018). (38) C.E. Whitehurst, N. Nazef, D.A. Vlattas, N.L. Lyons, B.R. Bowen, (22) B.M. Johnson, D. Nikolic, and R.B. Annis, Y. Hou, D.M. Murphy, P. Spac- and L.P. Wennogle, Biochemistry van Breemen, Mass Spectrom. Rev. ciapoli, Z. Yao, M.R. Ziebell, C.C. 35, 11747–11755 (19 96). 21, 76–86 (2002). Cheng, G.W. Shipps Jr, J.S. Felsch, (10) Y.-H. Chu, L.Z. Avila, J. Gao, and (23) R.B. van Breemen, Y. Tao, and W. Li, D. Lau, and H.M. Nash, J. Biomol. G.M. Whitesides, Ace. Chem. Res. Fitoterapia 82, 38–43 (2011). Screen. 11, 194–207 (2006). 28, 461–468 (1995). (24) R.B. van Breemen, C.H. Huang, D. (39) W.L. Petrilli, G.C. Adam, R.S. Erd- (11) E.S.M. Ng, F. Yang, A. Kameyama, Nikolic, C.P. Woodbury, Y.Z. Zhao, mann, P. Abeywickrema, V. Agnani, M.M. Palcic, O. Hindsgaul, and D.C. and D.L. Venton, Anal. Chem. 69, X. Ai, J. Baysarowich, N. Byrne, J.P. Schriemer, Anal. Chem. 77, 6125– 2159–2164 (1997). US Patent No. Caldwell, W. Chang, E. DiNunzio, 6133 (2005). US 5872015 (filed May 10, 1996) Z. Feng, R. Ford, S. Ha, Y. Huang, (12) H. Vu, N.B. Pham, and R.J. Quinn, (25) R. Munigunti, V. Mulabagal, and A.I. B. Hubbard, J.M. Johnston, M. Ka- Biomol. Screen. 13, 265–2 7 5 Calderón, J. Pharm. Biomed. Anal. vana, J.-M. Lisnock, R. Liang, J. Lu, (2008). 55, 265–271 (2011). Z. Lu, J. Meng, P. Orth, O. Palyha, (13) E.C. VanderPorten, M.D. Scholle, J. (26) J.L. Zhou, Z.M. Qian, Y.D. Luo, G. Parthasarathy, S.P. Salowe, S. Sherrill, J.C. Tran, and Y. Liu, SLAS D. Tang, H. Chen, L. Yi, and P. Li, Sharma, J. Shipman, S.M. Soisson, Discov. 22, 1211–1217 (2017). Biomed. Chromatogr. 22, 116 4 – A.M. Strack, H. Youm, K. Zhao, D.L. (14) S. Mathur, J.D. Park, D.H. Kim, and 1172 (2008). Zink, H. Zokian, G.H. Addona, K. R.W. Hartmann. J. Biomol. Screen. (27) M.B. Beverly, P. West, and R.K. Ju- Akinsanya, J.R. Tata, Y. Xiong, and 10, 30–35 (2005). lian, Comb. Chem. High Through- J.E. Imbriglio, Cell Chem. Biol. 26, (15) S. Kaur, L. McGuire, D. Tang, G. put Screen. 5, 65–73 (2002). 32–40 (2020). Dollinger, and V. Huebner, J. Pro- (28) E.K. Comess, J.E. Schurdak, M.J. (40) Y.G. Shin, J.L. Bolton, and R.B. tein Chem. 16, 505–511 (19 97). Voorbach, M. Coen, J.D. Trumbull, van Breemen, Comb. Chem. High (16) T.N. O’Connell, J. Ramsay, S.F. H. Yang, L. Gao, H. Tang, X. Cheng, Throughput Screen. 5, 59– 6 4 Rieth, M.J. Shapiro, and J.G. Stroh. C.G. Lerner, J.O. McCall, D.J. Burns, (2002). Anal. Chem. 86, 7413–7420 (2014). and B.A. Beutel, J. Biomol. Screen. (41) Y. Lu, S. Qin, B. Zhang, A. Dai, X. (17) V.D. Kutilek, C.L. Andrews, M.P. 11, 743–754 (2006). Cai, M. Ma, Z.-G. Gao, D. Yang, Richards, Z. Xu, T. Sun T, Y. Chen, A. (29) H. Li, F. Song, J. Xing, R. Tsao, Z. R.C. Stevens, K.A. Jacobson, M.-W. Hashke, N. Smotrov, R. Fernandez, Liu, and S. Liu, J. Am. Soc. Mass Wang, and W. Shui, Anal. Chem. 91, E.B. Nickbarg, C. Chamberlin, B. Spectrom. 20, 1496–1503 (2009). 8162–8169 (2019). Sauvagnat, P.J. Curran, R. Boinay, P. (30) Y. Choi, Y. Jung, and S.N. Kim, Mol- (42) S. Qin, M. Meng, D. Yang, W. Bai, Saradjian, S.J. Allen, N. Byrne, N.L. ecules 20, 13753–13763 (2015). Y. Lu, Y. Peng, G. Song, Y. Wu, Q. Elsen, R.E. Ford, D.L. Hall, M. Korn- (31) Y. Sun, C. Gu, X. Liu, W. Liang, P. Zhou, S. Zhao, X. Huang, J.D. Mc- ienko, K.W. Rickert, S. Sharma, J.M. Yao, J.L. Bolton, and R.B. van Bree- Corvy, X. Cai, A. Dai, B.L. Roth, M.A. Shipman, K.J. Lumb, K. Coleman, men, J. Am. Soc. Mass Spectrom. Hanson, Z.-J. Liu, M.-W. Wang, R.C. P.J. Dandliker, I. Kariv, and B. Beu- 16, 271–279 (2005). Stevens, and W. Shui, Chem. Sci. 9, tel, J. Biomol. Screen. 21, 608–619 (32) D. Liu, J. Guo, Y. Luo, D.J. Broderick, 3192–3199 (2018). (2016). M.I. Schimerlik, J.M. Pezzuto, and (43) N.F. Rizvi, and E.B. Nickbarg, Meth- (18) Y. Deng, G.W. Shipps, Jr, L. Zhao, R.B. van Breemen, Anal. Chem. 79, ods 167, 28–38 (2019). M.A. Siddiqui, J. Popovici-Muller, 9398–9402 (2007). (44) N.F. Rizvi, J.A. Howe, A. Nahvi, D.J. P.J. Curran, J.S. Duca, A.W. Hruza, (33) Y. Choi, and R.B. van Breemen, Klein, T.O. Fischmann, H.-Y. Kim, T.O. Fischmann, V.S. Madison, Comb. Chem. High Throughput M.A. McCoy, S.S. Walker, A. Hruza, R. Zhang, C.W. McNemar, T.W. Screen. 11, 1–6 (2008). M.P. Richards, C. Chamberlin, P. Mayhood, R. Syto, A. Annis, P. (34) M.D. Rush, E.M. Walker, T. Burton, Saradjian, M.T. Butko, G. Mercado, Kirschmeier, E.M. Lees, D.A. Parry, and R.B. van Breemen, J. Nat. Prod. J. Burchard, C. Strickland, P.J. Dan- and W.T. Windsor, Bioorg. Med. 79, 2898–2902 (2016). dliker, G.F. Smith, and E.B. Nick- Chem. Lett. 24, 199–203 (2014). (35) C. Gu, D. Nikolic, J. Lai, X. Xu, and barg, ACS Chem. Biol. 13, 820–831 (19) X. Huang, C.C. Cheng, T.O. Fis- R.B. van Breemen, Combin. Chem. (2018) chmann, J.S. Duca, X. Yang, M. High Throughput Screen. 2, 353– Richards, and G.W. Shipps, Jr., 359 (1999). ACS Med. Chem. Lett. 3, 123–128 (36) B.M. Johnson, D. Nikolic, and R.B. Richard van Breemen is with (2012). van Breemen, Mass Spectrom. Rev. the Linus Pauling Institute at Oregon (20) Y. Fu, H.Y. Mo, W. Gao, J.Y. Hong, J. 21, 76–86 (2002). State University in Corvallis, Oregon. Lu, P. Li, and J. Chen, Anal. Bioanal. (37) D.A. Annis, N. Nazef, C.C. Chuang, Direct correspondence to: richard.van- Chem. 406, 4987–4995 (2014). M.P. Scott, and H.M. Nash, J. Am. [email protected] 26 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

GC×GC–TOF-MS and Comprehensive Fingerprinting of Volatiles in Food: Capturing the Signature of Quality Chemical fingerprinting can provide evidence for quality differences resulting from botanical and geographical origins of primary food ingredients, post-harvest practices, production processes (such as traditional versus industrial processes), and the shelf-life evolution of finished products. This article discusses the strategic role and potential of comprehensive two-dimensional gas chromatography (GC×GC) combined with time-of-flight mass spectrometry (TOF-MS) and pattern recognition, using template matching for data processing, to unravel the quality traits of high-quality food products. Practical examples dealing with high-quality cocoa and extra-virgin olive oil are described.

Federico Stilo, Erica Liberto, Carlo Bicchi, Stephen E. Reichenbach, and Chiara Cordero

fter the second industrial revolution, the need for the primary ingredients, post‑harvest practices, production pro- standardization of quality for food production led to cesses (traditional versus industrial), and shelf-life evolution Athe introduction of periodic inspections, standardiza- of finished products. Illustrative examples on how compre- tion of procedures, and quality controls (1,2). The Interna- hensive chemical fingerprinting processes can be strategically tional Organization for Standardization (ISO) 9000 is proba- valuable for characterizing the quality traits of food from our bly the best known modern international standard for quality recent research are highlighted here. This article focuses on management (3). Current European Union (EU) policy on the potentials of comprehensive two-dimensional gas chro- food quality aims to protect products characterized by unique matography coupled to time-of-flight mass spectrometry features linked to their geographical origin, as well as tradi- (GC×GC–TOF-MS) to generate highly informative finger- tional expertise. With this policy, product names are granted prints of chemicals from complex fractions, including potent by a geographical indication (GI), a protected geographical odorants responsible for the aroma signature, and technologi- indication (PGI), or a protected designation of origin (PDO), cal markers that indicate the impact of the production process. if they have specificities strictly linked to the place where they When complex two-dimensional (2D) patterns of chemicals are produced, including the compositional characteristics of are explored by dedicated pattern recognition algorithms, a raw materials, the climate, and the traditional processes of high level of information about sensory profile, product au- manufacturing and transformation (4). thenticity, and technological impact is revealed (5,6). An analytical platform or method capable of capturing The first example deals with high-quality cocoa, a food the chemical traits of a food that are related to its perceived commodity of global economic interest, and the second ex- quality (mainly sensory quality, raw material authenticity, and ample deals with extra-virgin olive oil, an important local processing impact) will contribute to the quality assessment commodity in many Mediterranean countries, including process ,while also providing a foundation for consumer-tai- Spain, Italy, Greece, and Tunisia. Both of these commodities lored strategies to improve a product’s acceptance and loyalty. have an intrinsic “added value” related to their flavor profile Chemical fingerprinting can provide evidence for quality and perceived quality that are 80–90% a result of aroma-active differences arising from botanical and geographical origins of compounds (7). chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 27

Experimental Procedures Chemicals and Samples The mixture of n-alkanes (n-C9 to n-C25) for calibrating linear retention indices (IT) was from Sigma-Aldrich. The IT solution was prepared in cyclohexane at a concentration of 100 mg/L. Cocoa samples were from Gobino srl, and were selected on the basis of their peculiar sensory profile from high-quality productions of different geographic origins. Processing was performed in three replicated batches, using time and temperature protocols from 100–130 °C between 20–40 min. Processing was opti- Figure 1: Chemical classes distribution within the cocoa volatilome based on the list of targeted mized for each origin, and driven by a de- analytes reliably identified in the selected samples. sirable flavor development. Hot-air roasting was conducted in a vertical roaster, SPME device was exposed to 500 mg of polydimethylsiloxane, 7% phenyl, 7% designed by Bühler AG. cocoa in a headspace glass vial (20 mL) for cyanopropyl; 2 m × 0.1 mm dc, 0.10-μm Cocoa samples were frozen in liquid ni- 30 min at 50 °C, or 100 mg of olive oil in a df) from J&W (Agilent). SPME thermal trogen immediately after each step of pro- headspace glass vial (20 mL) for 60 min at desorption into the GC injector port was cessing, and then stored at -80 °C. Before 40 °C. Extracted analytes were recovered under the following conditions: split–split- headspace analysis, samples were ground by thermal desorption of the fiber into less injector in split mode at 250 °C, split in a laboratory mill up to approximately the S/SL injection port of the GC system ratio 1:20. The carrier gas was helium at 300 µm (Grindomix GM200, Retsch); at 250 °C for 5 min. a constant flow of 1.3 mL/min. The oven particle size homogeneity was verified by temperature program was from 40 °C (2 visual inspection. GC×GC–TOF-MS Conditions min) to 240 °C at 3.5 °C/min (10 min). Extra-virgin olive (EVO) oil samples GC×GC analyses were performed on an The n-alkanes liquid sample solution T were collected within the Italian “Violin” Agilent 7890B GC system coupled with a for I S determination was analyzed under Project (Valorization of Italian OLive prod- Bench TOF-Select system (Markes Inter- the following conditions: split–splitless ucts through INnovative analytical tools- national), featuring Tandem Ionization injector in split mode, split ratio: 1:50, in- AGER Fondazioni in rete per la Ricerca that provides variable-energy electron jector temperature: 250 °C, and injection Agroalimentare), and selected for their ionization. The ion source and transfer volume: 1 µL. sensory profile by an expert panel. Exam- line were set at 270 °C. The MS optimiza- ples cited in this paper refer to a commercial tion option was set to operate with a mass Raw Data Acquisition, EVO oil with a PGI quality label (Azienda range between 40–300 m/z; data acquisition 2D-Data Processing and Statistics Agricola Mori Concetta, PGI Toscano, ol- frequency was 50 Hz for each channel; fila- Data were acquired by TOF-DS software ives Mariolo cultivar, San Casciano in Val ment voltage was set at 1.60 V. Electron ion- (Markes International), and processed di Pesa, Firenze, Italy) and a PDO product ization energies explored were 70 and 12 eV. using GC Image version 2.8 (GC Image, (Azienda Agricola Leone Sabino, Don Gio- The system was equipped with a two- LLC). Statistical analysis used XLStat acchino Gran Cru, DOP Terra di Bari Cas- stage KT 2004 loop thermal modulator (Addinsoft). tel del Monte, 100% Coratina olives cultivar, (Zoex Corporation), cooled with liquid Canosa di Puglia, Italy). nitrogen controlled by Optimode V.2 Results and Discussion (SRA Instruments). The hot jet pulse time Cocoa Origin and its

Headspace Solid-Phase was set at 250 ms, modulation period (PM) Distinctive Chemical Signature Microextraction (SPME) Devices was 4 s for cocoa and 3.5 s for olive oil, Cocoa (Theobroma Cacao L. Malvaceae and Sampling Conditions and cold-jet total flow was progressively family) is the main raw ingredient for choc- A divinylbenzene/carboxen/polydimeth- reduced with a linear function from 40% olate production (9). It is native to tropical ylsiloxane 1-cm SPME fiber from Supelco of mass flow controller (MFC) at initial forests of the South American continent, was used for HS-SPME sampling. The conditions to 8% at the end of the run. although recent statistics indicate that standard in-fiber procedure (8) was ad- most of the production is concentrated in opted to preload the IS (α-thujone) onto GC×GC Columns and Settings Africa, with Cote d’Ivoire and Ghana cov- the fiber before sampling. A 5.0-µL solu- The column set was configured as fol- ering about 56% of the global production, tion of IS (α-thujone at 100 mg/L in diethyl lows: 1D SolGel-Wax column (100% followed by Indonesia (12%), Nigeria (6%), phthalate) was placed into a 20-mL glass polyethylene glycol; 30 m × 0.25 mm dc, and Cameroon (6%). vial, and subjected to HS-SPME at 50 °C 0.25-μm df) from SGE Analytical Science, Cocoa and chocolate are considered for 5 min. After the IS loading step, the coupled with a 2D OV1701 column (86% comfort foods ,and are consumed world- 28 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Figure 2: Scores plot resulting from the normalized response of 595 reliable peak-regions from (a) raw and (b) roasted cocoas of different origin (CH-Chontalpa/Mexico, VEN-Venezuela, CO-Colombia, EC-Ecuador, JA-Java, TRI-Trinidad, ST-Sao Tomè). Adapted with permission from reference 21.

unique signatures of volatiles from commercial, high-quality cocoa intermedi- ates, with a pattern recognition strategy based on GC×GC–MS that extends the investigation potential to both untargeted and targeted analytes (21). The approach is based on the template matching prin- ciple, and is named untargeted/targeted (UT) fingerprinting (22,23). Samples were from different geographical provenience (Mexico, Ecuador, Venezuela, Colombia, Java, Trinidad, and Sao Tomè), selected by experts for their unique aroma profiles, and were studied along early steps of processing as raw, roasted, steamed, and ground nibs. For some origins, cocoa liquor was also in- Figure 3: Samples discrimination based on three variables: x-axis linalool, y-axis cluded to evaluate its aroma signature for 2,3,5-trimethylpyrazine, and bubble-size 2-pentylfuran selected by regression tree analysis. chocolate products design. Adapted with permission from reference 21. The fraction of volatiles was extracted by automated headspace solid-phase microex- wide for their pleasant sensory profile of and informative fingerprinting of such a traction (HS-SPME) and on-line analyzed unique and complex flavors. This complex- complex fraction. Deeper insights on the by GC×GC–MS in a system equipped with ity arises from multiple interconnected bio- quali-quantitative distribution of volatiles a loop-type thermal modulator. Within 595 chemical and chemical reactions occurring would help in delineating an origin-spe- detectable analytes, delineated by unique at post-harvest stage where pedoclimatic cific aroma blueprint, inform about the 2D peak-regions and covering most of the conditions and farming practices play a seasonal variations or the effect of climate chemical dimensions of cocoa volatilome, major role (10,11). Later in the processing changes, or help chocolate manufacturers about 200 compounds were tentatively 1 chain, roasting, conching, and tempering in designing tailored blends evoking pe- identified on the basis of D IT and MS develop the flavor profile and the distinctive culiar aroma notes. spectral similarity, with authentic standards sensory signature of chocolate (12–14). GC×GC–MS exploits the potential of or with spectra collected in commercial Cocoa’s complex aroma is modulated two separation dimensions with the addi- (Wiley 7n and NIST 2015) and in-house by a series of potent odorants (15–17), tional orthogonal information provided by databases (23). Figure 1 shows chemical whose specific quali‑quantitative distri- MS. This results in: (i) increased separation classes distribution within the cocoa vol- bution within the bulk of several hun- power; (ii) meaningful 2D patterns with atilome. Esters dominate the volatile frac- dreds of volatiles has been identified as a analytes structurally ordered in the chro- tion and, together with alcohols and acids, distinctive aroma signature, also referred matographic space; and (iii) enhanced sen- bring information about fermentation and to as an aroma blueprint. From this per- sitivity as a result of band focusing in space its impact on primary metabolites (mainly spective, GC×GC–MS would be the ana- obtained by cryogenic modulation (18–20). sugars and amino acids). Within the het- lytical technique of choice for an accurate In a recent study, we investigated the erocycles, the subset of alkyl pyrazines is of chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 29

Figure 4: Heat map representing the quantitation results for a selection of key odorants. Data are expressed as (a) absolute concentration in the sample and (b) by odor activity value (OAV) computed as the coefficient of the concentration of an odorant (mg/kg) versus its odor threshold (mg/ kg). Adapted with permission from reference 24. great relevance, because they provide infor- The total explained variability represented chain and branched fatty acids (acetic acid, mation about the technological impact on by the first two principal components (F1 butanoic acid, 2-methylpropanoic acid, and some precursors present in raw cocoa, and and F2) is about 44% for raw and 41% for 3-methylbutanoic acid), whose presence at for the most odor‑active compounds, they roasted beans. high concentrations can impart off-flavors bring the earthy/roasty notes to the global Supervised approaches, at this stage, as a result of their rancid, sour, and sweaty aroma. Alkyl pyrazines are formed from may help in defining or selecting highly notes. Strecker aldehydes (2- and 3-methyl- the early stages of processing during bean discriminating variables. A classification butanal), formed during fermentation and drying, and later by roasting and steaming model based on three variables, 2-pentylfu- roasting, impress malty and buttery notes, (9). Carbonyls (aldehydes and ketones) are ran, 2,3,5-trimethylpyrazine, and linalool, and phenylacetaldehyde, derived from formed mainly from fatty acids precursors is capable of discriminating between cocoa L-phenylalanine (L-Phe), imparts a pleas- by oxidative (chemical and enzymatic) re- nib origins. Figure 3 shows how samples ant honey-like note. actions. The sub-group of Streker aldehydes could be discriminated in the Cartesian Other key-aromas are esters (eth- (2- and 3-methylbutanal, methylpropanal, space of three variables: x-axis linalool, yl-2-methylbutanoate: fruity; 2-phenylethyl and phenylacetaldehyde) are fundamental y-axis 2,3,5-trimethylpyrazine, and bub- acetate: flowery), linear alcohols (2-hepta- for cocoa and chocolate flavor modulating ble-size 2-pentylfuran (21). This model nol: citrusy), phenyl propanoids derivatives the malty, buttery, and honey-like notes. confirms what unsupervised exploration (2-phenylethanol: flowery), and sulfur-de- The quali-quantitative distribution of by PCA showed (Figure 2b). Samples from rived compounds (dimethyl trisulfide). known and unknown volatiles, for exam- Ecuador and Colombia are aligned along Accurate quantitation of these ana- ple, the volatile metabolome fingerprint, the x-axis (low linalool content) together lytes, performed by multiple headspace is potentially informative, and helps in the with the Venezuela samples. Java is char- extraction SPME–GC–MS with flame discrimination and differentiation of geo- acterized by a low pyrazines signature ionization detection (FID) (24), are visu- graphical origin and manufacturing stage. (2,3,5-trimethylpyrazine is one of the most alized as a heat map for a subset of origins As an example, unsupervised multivariate origin sensitive), but clustered together with in Figure 4. Odorants were quantified by analysis, such as principal component anal- Chontalpa/Mexico and Sao Tomè for their external calibration and FID-predicted ysis (PCA), can be applied to reveal the nat- lower amount of linalool. Trinidad has an relative response factors. Amounts are re- ural conformation (groups) of the analyzed intermediate position between the two ported for nibs and cocoa mass in Figure samples and helps in localizing informative groups, but with a relatively high amount 4a and the odor activity value (OAV) is chemical features responsible for cocoa dis- of 2-pentylfuran and trimethyl pyrazine. shown in Figure 4b. OAV is computed as crimination. Figures 2a and 2b show the Key aroma compound signatures (15–17) the coefficient of the concentration of an scores plot resulting from the normalized are of particular interest for perceived qual- odorant (mg/kg) versus its odor threshold response of 595 reliable peak-regions from ity. These signatures are buried within the (mg/kg). OAV is a useful parameter for raw (2a) and roasted (2b) cocoas of different bulk of the cocoa volatilome, but their infor- discriminating odorants from interfering origin (CH-Chontalpa/Mexico, VEN-Ven- mation is strategic for the confectionery in- components. Below an OAV of 1, which ezuela, CO-Colombia, EC-Ecuador, JA- dustry and can be used in new origin selec- is generally used as a threshold value, it is Java, TR-Trinidad, ST-Sao Tomè). Samples tion and blending. Key aroma compounds assumed that an odorant does not play a are grouped into three main clusters: Ec- include alkyl pyrazines (2,3,5-trimethylpyr- role in eliciting its characteristic quality. uador‑Venezuela-Colombia (blue circles), azine, 2-ethyl-3,5-dimethylpyrazine, and However, several more parameters need to Chontalpa/Mexico-Sao Tomè-Java (red 3,5-diethyl-2-methylpyrazine), which are be considered to judge odor activity of vol- circles), and Trinidad (green circles) (21). responsible for the earthy notes, and short- atiles. The capture of chemical complexity 30 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com

Figure 5: 2D-patterns from Venezuela cocoa ([a] roasted and [b] steamed) together with a comparative visualization rendered as colorized fuzzy ratio (c). The zoomed area (orange rectangle) highlights the pyrazines signature (d).

dihydroxy-6-methyl(4H)-pyran-4-one (DDMP) is an effective marker of processing (21): its concentration in cocoa mass is, on average, two orders of magnitude higher than in cocoa nibs. To confirm the great flexibility of GC×GC in cocoa volatilome fingerprinting, it is interesting to explore the effect of steaming on roasted cocoa beans with simple and intuitive tools. Steaming is conducted on shelled cocoa beans after roasting. It is an on-line process conducted with overheated‑steam that lowers the bacte- rial charge of the beans. However, steaming impacts cocoa aroma and its process parameters have to be carefully tuned to avoid off-flavor formation or the loss of key odorants. From this per- spective, comparative visualization based on datapoint features (26) could be of great help. This is a pointwise approach where chromatograms are compared point‑by‑point or pixel‑by‑pixel. In a GC×GC–MS chromatogram, every datapoint corresponds to a detector event. With this approach, each datapoint is a feature, and therefore datapoint features at the same retention times are implicitly matched. Figure 5 shows the 2D-patterns from Ven- ezuela cocoa (Figure 5a roasted and Figure 5b steamed). Their comparative visualization is rendered as colorized fuzzy ratio (Figure 5c); within the zoomed area (orange rectangle), the pyrazines signature is visible (Figure 5d). This approach gives prompt information on analytes (relative) variations over the patterns Figure 6: 2D chromatogram of (a) a PGI Toscano EVO oil, and in and, by color codes (green, red, and gray), informs about their zoomed areas the lipoxygenase (LOX) signature (b), linear saturated higher or lower abundance in the samples. In this specific case, and unsaturated aldehydes (c). steaming has an impact on pyrazines quantitative distribution; those with short alkyl substituents are colored by green pixels of volatile patterns by comprehensive approaches is currently one indicating that they were more abundant in the analyzed image of the most effective strategies available (25). (steamed beans). Conversely, red colored pixels indicate analytes Hierarchical clustering based on Euclidean distances helps in and chromatographic regions where the detector response was sample discrimination based on their aroma profiles. Key odor- higher in the reference (roasted beans) image. ants, such as 2-methylpropanic acid, 3-methylbutanoic acid, ace- toin, and 2-phenylethanol, show a homogeneous trend across all Italian Extra-Virgin Olive Oil: samples. A similar behavior is seen for other odorants, such as How to Capture Volatiles Chemical Complexity isoamyl acetate, γ-butyrolactone, and 2-acetyl pyrrole, that clus- Consumers appreciate high-quality olive oil, whether virgin oil ter independently from the others. As expected, 2,3-dihydro-3,5- (VO) or EVO oil, for its health benefits and pleasant and dis- chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 31

attributes, origin (29,30), or ripening sta- tus of olives (6). The volatile fraction of olive oil is complex, and connoted by high chemical dimensionality (31), a parameter defined by Giddings to describe the degree of order or disorder in multidimensional separations. The presence of several chemical functionalities, also represented by ho- mologous series, generated by the multiple chemical and biochemical reactions occurring to olives primary metabolites, results in complex 2D patterns that require, for accurate fingerprinting, high-resolution separations and orthogonal detection by mass spectrometry. MS is fundamental Figure 7: 2D chromatograms of a high quality PDO Don Gioacchino Gran Cru Terra di Bari EVO here to extract the information on ana- oil acquired at (a) 70 eV and (b) at 12 eV. Spectra for (5E)-3-ethyl-1,5-octadiene and (E,E)-3,7- lyte fragmentation patterns for a reliable decadiene at 70 and 12 eV are reported in (c) and (d). identification. Figure 6a shows the 2D chromatogram of a PGI Toscano EVO tinctive flavor. The objective quality and and promotion to counteract frauds and oil. The number of detectable 2D-peaks purity of this product is regulated by in- illicit operations. with a signal-to-noise ratio (S/N) thresh- ternational organizations, including the Undoubtedly, quality control issues are old of 100 is about 750 and, for 180 of EU, International Oil Council (IOC), and partially related to the lack of powerful them, reliable identification was possi- Codex Alimentarius. If from one side, and informative analytical methods capable by matching 1D IT and MS spectrum EVO oil adulteration can be assessed by ble of supporting objective classification of with those collected in commercial and characterizing the nonvolatile fraction, for olive oils, in particular for the distinction in-house databases (5,6). Reference com- example, fatty acid methyl esters (FAMEs) between VO and EVO oils based on their pounds were also adopted in the case of profiles, sterol and triterpene dialcohol aroma signatures. Olive oil is currently the key odorants related to positive attributes composition, wax content, and presence of only food product with officially regulated or coded defects. conjugated dienes and trienes, the aroma sensory attributes; standardized protocols Within the separated volatiles, the quality can be evaluated by panel testing (27,28) for sensory assessment are estab- lipoxygenase (LOX) signature (Fig- without any objective analytical protocol lished by law, and highly trained panel- ure 6b) is fundamental to define fresh- supporting the sensory assessment. ists are involved in this task. Based on the green and fruity notes, the positive at- Olive oil production is located mainly presence or absence and the intensities of tributes. C6 unsaturated alcohols and in Europe with Spain, Italy, and Greece coded defects and, on the other side, of aldehydes, (Z)-2-hexenal, (E)-2-hexenal, as the key players. Italian production, “fruity” perception, evaluated by smelling hexanol, (Z)-3-hexenol, (E)-2-hexenol, which is estimated at approximately 50% and tasting, virgin olive oil is then classi- (E,Z)-2,4-hexadienal, and (E,E)-2,4-hexa- of Spanish production, is characterized fied into three categories: EVO, VO, and dienal, are formed from linoleic and lin- by a relatively high number of PDO oils, lampante oil. olenic acids oxidative cleavage. Figure 6c with 46 registered products compared In this scenario, GC×GC–MS rep- illustrates the linear saturated and unsat- with 31 for Spain (European Commis- resents an analytical tool with great po- urated aldehydes, together with a few ke- sion – DOOR database http://ec.europa. tential to describe the chemical signature tones likely representing hydroperoxides eu/agriculture/quality/door/list.html). of sensory quality markers, including po- (that is, primary products of lipid oxida- Quality labeling gives consumers a sense tent odorants related to oil-coded defects, tion) cleavage products. This last group of of tradition, while improving trust and that is fusty/muddy sediment, musty/ analytes informs generally about shelf‑life confidence for a high-quality product. humid/earthy, winey/vinegary, or ran- evolution, with increasing concentrations Although these products must comply cid. Previous studies by our research team of potent odorants that bring rancid and for higher quality standards compared explored and confirmed the possibility of fatty notes. to ordinary production, the EU Parlia- delineating meaningful patterns of po- For an accurate and informative fin- ment, through the resolution of 14 Janu- tent odorants eliciting the coded defects gerprinting of EVO oil volatiles, an ad- ary 2014 on “food crisis, fraud in the food in olive oils (5). At the same time, thanks ditional dimension at the detection level chain and the control thereof,” revealed to the separation power and enhanced could be of help. Instruments capable of that olive oil is, among others, one of the resolution achievable with a comprehen- acquiring variable‑energy electron ion- top foods subjected to adulteration. Con- sive 2D-GC analysis, additional chemical ization spectra by time‑switching between sequently, the EU action plan focused on information can be collected and rational- two ionization energies across every sin- specific aspects related to quality, control, ized before correlating it with oil sensory gle analytical run represent an interesting 32 Current Trends in Mass Spectrometry March 2020 chromatographyonline.com option to extend method dimensionality. Conclusions speculations, but can also be concrete strat- Research on variable-energy electron ion- The concept of quality of food is complex egies to improve competitive advantages in ization spectra acquisition also showed ben- and connoted by different meanings. Food a complex food market. efits for the identification of large isomeric compositional complexity offers the oppor- species in unresolved complex mixtures tunity, through informative and reliable Acknowledgments (UCMs) of motor oil samples (32) and for analytical protocols, to objectify the most The research on cocoa was carried out light volatile organic compounds (VOCs) relevant characteristics of quality. Multidi- thanks to the financial support of Firmen- from human blood (33). The operation of mensional analytical platforms, and in par- ich S.A. Geneva, Switzerland. The research the ion source at low energies (12–16 eV) al- ticular those implementing comprehensive on olive oil was supported by Progetto Ager lowed enhanced intensity for structure-in- two‑dimensional separations with mass – Fondazioni in rete per la riscerca agroal- dicating fragments, which can improve spectrometry, have the intrinsic potential imentare. Project acronym Violin - Valo- method specificity. of delineating meaningful chemical finger- rization of Italian olive products through Recently, Freyre and colleagues (34) prints that can be explored for their targeted innovative analytical tools. proposed a tile-based Fisher ratio analysis and untargeted feature distribution. and a discovery-based investigation strat- To achieve the higher level of informa- References egy to detect exogenous analytes (spiked tion exploiting the quality concepts related (1) J.C. Wood and M.C. Wood, Eds., Henry at 50 ppm) in diesel fuel samples. Their to food origin, harvest practices, techno- Ford: Critical Evaluations in Business and strategy was based on data processing after logical processes impact, and flavor profile, Management (Routledge, London, United tandem signals fusion. Our research group meaningful analyte patterns must be delin- Kingdom, and New York, New York, 2003). explored the cocoa complex volatilome by eated. Data processing tools are therefore (2) J. Papp, Quality Management in the Im- GC×GC–TOF-MS and variable‑energy fundamental. Flexibility is a key character- aging Sciences (Elsevier Health Sciences, electron ionization proposing a novel istic for algorithms and workflows, and data 2014), p. 372. approach based on the UT fingerprint- patterns should be explored from different (3) www.iso.org ing strategy but performed on fused data perspectives in relation to the objectives. If (4) https://ec.europa.eu/info/food-farm- streams (hard and soft ionization energies the full pattern—for example, the chemi- ing-fisheries/food-safety-and-quality/ together) for effective cross‑comparative cal signature—of untargeted and targeted certification/quality-labels_en analysis of samples (23). analytes may help in describing chemical (5) G. Purcaro, C. Cordero, E. Liberto, C. Bic- Tandem data streams are complemen- differences between samples of different or- chi, and L.S. Conte, J. Chromatogr. A 1334, tary in terms of both spectral information/ igin, such as cocoa origins characterization, 101–111 (2014). fragmentation pattern dissimilarity and alternatively, datapoint features with visual (6) F. Magagna, L. Valverde-Som, C. Ruíz-Sam- absolute response. The first characteristic comparisons may help in locating single an- blás, L. Cuadros‑Rodríguez, S.E. Reichen- is of help when 70 eV spectra lack molecu- alyte variations between sample pairs, such bach, C. Bicchi, et al., Anal. Chim. Acta lar ions and structurally informative frag- as cocoa processing. 936, 245–258 (2016). ments, while the differential response from Algorithm flexibility should be accom- (7) C. Spence, Psychologist 23(9), 720–723 the two channels opens new perspectives panied by adequate tools for data align- (2010). in terms of dynamic range and linearity. ment, with effective transform functions (8) Y. Wang, J. O’Reilly, Y. Chen, and J. Pawl- Although at lower energies (12–16 eV), the (36) that support re-alignment of datasets iszyn, J. Chromatogr. A 1072(1), 13–17 absolute number of ionized molecules is re- acquired across wide time ranges, in analyt- (2005). duced, benefits are evident for background ical batches affected by random instrumen- (9) A.C. Aprotosoaie, S. Vlad Luca, and A. noise intensity and S/N. The latter benefit tal fluctuations, or with multidimensional Miron, Compr. Rev. Food Sci. Food Saf. is accrued for analytes showing a reduced detectors, with variable-energy electron 15(1), 73–91 (2016). fragmentation at lower energies. ionization TOF-MS, for example, for EVO (10) N. Camu, T. De Winter, S.K. Addo, J.S. Figure 7 shows the volatile 2D pattern of oil volatile patterns. Takrama, H. Bernaert, and L. De Vuyst, a high-quality PDO Don Gioacchino Gran In the scenario of linking meaningful J. Sci. Food Agric. 88(13), 2288–2297 Cru Terra di Bari EVO oil acquired at 70 chemical signatures to food quality con- (2008). eV (Figure 7a) and at 12 eV (Figure 7b). cepts, the role of data processing software is (11) G.V. de M. Pereira, V.T. Soccol, and C.R. The enlarged areas correspond to the elu- central. Analysts should drive the explora- Soccol, Current Opinion in Food Science tion region of unsaturated alkanes char- tion of complexity with simple and intuitive 7, 50–57 (2016). acteristic of early stages of olive ripening tools with an understanding of how data is (12) J.E. Kongor, M. Hinneh, D. Van de Walle, (35). They include: 3,4-diethyl-1,5-hexadi- preprocessed and treated along the steps of E.O. Afoakwa, P. Boeckx, and K. Dewet- ene (RS + SR), 3,4-diethyl-1,5-hexadiene the data analysis workflow to fully under- tinck, Food Res. Int. 82, 44–52 (2015). (meso), (5Z) and (5E)-3-ethyl-1,5-octadi- stand and appropriately use results. (13) R. Nazaruddin, H. Osman, S. Mamot, S. ene, (E,Z)- and (E,E)-3,7-decadiene, and Interestingly, these research efforts have Wahid, and A. Nor, J. Food Process Preserv. (E)-4,8-dimethyl- 1,3,7-nonatriene (6,35). broken down barriers between academic 30, 280–298 (2006). Spectra for (5E)-3-Ethyl-1,5-octadiene and and industrial research, indicating that (14) R. Saltini, R. Akkerman, and (E,E)-3,7-decadiene at 70 and 12 eV are re- these analytical tools and data mining con- S. Frosch, Food Control 29(1), ported in Figures 7c and 7d. cepts are not only academic exercises and 167–187 (2013). chromatographyonline.com March 2020 Current Trends in Mass Spectrometry 33 PRODUCTS & RESOURCES 3-MCPD analyzer Digital ion trap mass spectrometer An automated GC–MS-based system from Gerstel is designed to deter- Shimadzu Scientific’s MALDImini-1 digital ion trap (DIT) mass spectrom- mine 3-MCPD, 2-MCPD, and glycidyl fatty acid esters in edible oil, meeting eter (MS) is designed to fit in a space the size of a piece of paper while the requirements of standard providing high-sensitivity measure- ISO, AOCS, and DGF methods. ments and structural and qualitative According to the company, sam- analyses over a wide mass range, ples are automatically prepared even with sub-microliter sample and analyzed, including analyte volumes. According to the company, derivatization and evaporation of the system’s digital ion trap uses excess reagent and solvent, for rectangular wave RF to allow ion best limits of determination and trapping up to 70,000 Da. system stability. Shimadzu Scientific Instruments, Gerstel, Inc., Columbia, MD. Linthicum, MD. www.ssi.shimadzu.com www.gerstel.com

Hydrogen lab server Mass spectrometer Proton OnSite’s hydrogen lab server is designed to produce up Thermo Fisher’s Orbitrap Eclipse Tribrid mass spectrometer is designed to 18.8 standard liters with advancements that improve system sensitivity and speed over pre- per minute (equivalent vious generations of platforms to four cylinders) of ultra- through its high performance high purity hydrogen gas and flexibility. According to the per day. According to the company, the system extends company, the lab server structural analysis up to m/z senses demand, and adjusts 8000, enabling the isolation production accordingly. and selective dissociation of Proton OnSite, protein complexes into their Wallingford, CT. individual components. www.protononsite.com Thermo Fisher Scientific, San Jose, CA. www.thermofisher.com

(15) P. Schnermann and P. Schieberle, J. Agric. Food Chem. (27) Commission of the European Communities, Commission Regula- 45(3), 867–872 (1997). tion (Eec) No 2568/91 (Official Journal of the European Commu- (16) F. Frauendorfer and P. Schieberle, J. Agric. Food Chem. 54, nities. 1991), pp. 1–83. 5521–5529 (2006). (28) International Oil Council, COI/T.20/DOC.15/Rev.10 Sensory Analy- (17) F. Frauendorfer and P. Schieberle, J. Agric. Food Chem. 56, sis of Olive Oil - Method for the Organoleptic Assessment of Virgin 10244–10251 (2008). Olive Oil (2018). (18) H.J. Cortes, B. Winniford, J. Luong, and M. Pursch, J. Sep. (29) I. Lukić, S. Carlin, I. Horvat, and U. Vrhovsek, Food Chem. 270, Sci. 32(5–6), 883–904 (2009). 403–414 (2019). (19) M.S. Klee, J. Cochran, M. Merrick, and L.M. Blumberg, J. (30) L.T. Vaz-Freire, M.D.R.G. da Silva, and A.M.C. Freitas, Anal. Chim. Chromatogr. A 1383, 151–159 (2015). Acta 633(2), 263–270 (2009). (20) M. Adahchour, J. Beens, and U.A.T. Brinkman, J. Chro- (31) J.C. Giddings, J. Chromatogr. A 703(1–2), 3–15 (1995). matogr. A 1186(1–2), 67–108 (2008). (32) M.S. Alam, C. Stark, and R.M. Harrison, Anal. Chem. 88(8), 4211– (21) F. Magagna, A. Guglielmetti, E. Liberto, S.E. Reichenbach, 4220 (2016). E. Allegrucci, G. Gobino, et al., J. Agric. Food Chem. 65(30), (33) L.M. Dubois, K.A. Perrault, P.H. Stefanuto, S. Koschinski, M. Ed- 6329–6341 (2017). wards, L. McGregor, et al., J. Chromatogr. A 1501, 117–127 (2017). (22) S.E. Reichenbach, P.W. Carr, D.R. Stoll, and Q. Tao, J. Chro- (34) C.E. Freye, N.R. Moore, and R.E. Synovec, J. Chromatogr. A 1537, matogr. A 1216(16), 3458–3466 (2009). 99–108 (2018). (23) C. Cordero, A. Guglielmetti, C. Bicchi, E. Liberto, L. Baroux, (35) F. Angerosa, L. Camera, N. D’Alessandro, and G. Mellerio, J. Agric. P. Merle, et al., J. Chromatogr. A (in press) https://doi. Food Chem. 46(2), 648–653 (1998). org/10.1016/j.chroma.2019.03.025 (36) D.W. Rempe, S.E. Reichenbach, Q. Tao, C. Cordero, W.E. Rathbun, (24) C. Cordero, A. Guglielmetti, B. Sgorbini, C. Bicchi, E. Alle- and C.A. Zini, Anal. Chem. 88(20), 10028–10035 (2016). grucci, G. Gobino, et al., Anal. Chim. Acta 1052, 190–201 (2019). Federico Stilo, Erica Liberto, Carlo Bicchi, and (25) C. Cordero, J. Kiefl, S.E. Reichenbach, and C. Bicchi, TrAC - Chiara Cordero are with the Department of Drug Science and Trends in Analytical Chemistry 113, 364–378 (2019). Technology at the University of Turin, in Turin, Italy. Stephen E. (26) S.E. Reichenbach, X. Tian, C. Cordero, and Q. Tao, J. Chro- Reichenbach is a Professor of Computer Science and Engineering matogr. A 1226, 140–148 (2012). at the University of Nebraska, in Lincoln, Nebraska. Direct correspondence to: [email protected] FOOD AND BEVERAGE

Accurate Transfer of Viscous Samples for Completely Automated Extraction and LC–MS Determination of Mycotoxins in Edible Oils Fredrick D. Foster, John R. Stuff, Laurel A. Vernarelli, Jacqueline A. Whitecavage, Gerstel, Inc.,

The accurate and precise transfer of liquid samples is critical to the quality of analytical results. Liquid samples with high viscosities pose several challenges to achieving accurate and precise delivery. Automated, accurate transfer of viscous liquids can help improve the quality of the overall analytical procedure and the resulting data while freeing the analyst from performing a difficult and tedious manual task. A robotic autosampler for GC and HPLC was used in this work to perform a variety of sample preparation techniques. An analytical balance was included for weight verification of liquid transfers. The performance of a new heated liquid syringe tool that allows viscous liquid samples to be accurately transferred was examined. Resulting weight verification data for edible oil samples are provided. Good accuracy and precision for automated transfer of viscous samples was demonstrated. The system Figure 1: MPS robotic/roboticPRO sampler configured with Gerstel enables completely automated extraction of mycotoxins from edible automated sample preparation options. oils combined with LC–MS/MS analysis of the extract using a single automated analysis setup under integrated control software. 6. Supernatant is transferred to a clean, empty, 10 mL vial. Experimental 7. 4 mL hexane is added. Extra virgin olive oil (cold pressed), sesame oil (pure), flax oil (organic, 8. The vial content is agitated for 10 min. at 2000 rpm. pure, unrefined, cold pressed), and sunflower oil (virgin, cold 9. The vial is centrifuged for 5 min at 2000 g. pressed), were purchased from local markets. A range of aflatoxin- 10. An aliquot of the lower layer is transferred to an empty, round spiked edible oil samples were prepared using dilutions of the bottom vial. mycotoxin mix stock solution. A (95:5) acetonitrile:formic acid (v:v) 11. The extract is evaporated to dryness at 45 °C. extraction solution was prepared. 12. The residue is taken up in methanol:water. 13. An aliquot of the reconstituted extract is injected—or filtered and Instrumentation injected—to the LC-QQQ. All automated PrepSequences were performed using a Gerstel

MPS robotic/roboticPRO dual head sampler fitted with Gerstel CF- Analysis conditions (for a complete overview, please see Ref. 2) 200 centrifuge, balance (weighing option), mVAP multi-position LC run time: 14 min; Injection volume: 2.0 μL (loop over-fill technique); evaporation unit, quickMIX module, five-position dilutor option, a Column temperature: 30 °C. MS with electrospray ionization, positive heated agitator, and Gerstel heated liquid syringe module (HLM) as mode. shown in Figure 1. All analyses were performed on an Agilent 1260 HPLC, coupled to Results and Discussion an Agilent Ultivo triple quadrupole mass spectrometer with jet stream Raising the temperature of a viscous sample decreases its viscosity. electrospray source. Sample injections were performed using the Both the sample and the syringe being used to transfer the sample

Gerstel roboticPRO sampler with LC–MS tool into a six-port injection must be heated to achieve reliable and accurate transfer of viscous valve outfitted with a 2-μL stainless steel loop. samples. Increasing the temperature of propylene glycol from 30 to 40 °C was shown to lead to an improvement in transfer volume Automated Prep Sequence accuracy from 68.4% to 98.4%. A manual method for liquid–liquid extraction of mycotoxins from edible Identical volumes of olive oil (cP = 40 at 38 °C), sesame oil oils (1) was automated. (cP = 41 at 35 °C), flax oil (cP = 29 at 38 °C), and sunflower oil (cP 1. The sample is incubated at 60 °C for 10 min. = 49 at 25 °C) were placed into the heated agitator at 60 °C for 10 2. Edible oil is sampled into an empty vial. min, and replicate aliquots of each were then transferred to individual 3. (95:5) acetonitrile: formic acid (v/v) is added. vials using the HLM (65 °C). In Table I, the resulting precision and 4. The vial content is agitated for 10 min. at 2000 rpm. accuracy data for replicate transfers of each edible oil are shown. 5. The vial is centrifuged for 5 min at 2000 g. Figure 2 shows representative mass chromatograms resulting from

34 THE APPLICATION NOTEBOOK – MARCH 2020 FOOD AND BEVERAGE

Table I. Precision and accuracy of edible oil transfer using Table II. % Recovery of mycotoxins from edible oil using the the HLM automated extraction procedure

Sesame Oil Sunflower Replicate Olive Oil [g] Flax Oil [g] % Recovery Aflatoxin B1 Aflatoxin B2 Aflatoxin G1 Aflatoxin G2 [g] Oil [g] Olive Oil 95.9 103 95.7 129 1 1.3199 1.3305 1.343 1.3281 Sesame Oil 83.8 93.7 92.1 131 2 1.3192 1.3298 1.3431 1.3298 Flax Oil 87.7 104 88.0 126 3 1.3189 1.3312 1.3438 1.3301 Sunflower 4 1.3180 1.3276 1.3439 1.3284 84.3 101 90.2 101 Oil mean 1.3190 1.3298 1.3435 1.3291 Ave. % 87.9 101 91.5 122 SD 0.000787 0.00156 0.000465 0.000997 Recovery % CV 0.0597 0.1172 0.0346 0.0750 % Diff from -5.45 -4.68 -3.70 -4.72 Conclusions Theo. As a result of this study, we were able to show: • An extraction procedure for mycotoxins in edible oils was readily extracted sunflower oil spiked with mycotoxin at concentrations of 10 automated using the Gerstel MPS roboticPRO sampler, including ng/mL (Aflatoxins B1, G1) and 2.5 ng/mL (Aflatoxins B2, G2). introduction of the extract to LC–MS/MS and analysis based on Mycotoxins were shown to be reproducibly extracted from edible Agilent Ultivo triple quadrupole mass spectrometer. • Viscous edible oil samples can be transferred accurately and precisely using the Gerstel HLM.

x101 Aflatoxin B1: +ESI MRM Frag=190.0V [email protected] (313.3 -> 285.1) 071619_1010.d • Mycotoxins can be reproducibly extracted from edible oil 6.799 5.5 313.61 5 4.5 4 samples using an automated extraction procedure with an 3.5 3 2.5 2 average precision of 4.3 % (range: 2.4 %–6.9 %RSD). 1.5 1 0.5 0 1 • The recovery of mycotoxins from edible oil samples using the x10 Aflatoxin B2: +ESI MRM Frag=190.0V [email protected] (315.1 -> 287.1) 071619_1010.d 6.476 3 172.23 2.75 2.5 automated extraction procedure and LC–MS/MS analysis 2.25 2 1.75 1.5 1.25 averaged 101 % (range: 87.9 %–122 %). 1 0.75 0.5 0.25 0 x101 Aflatoxin G1: +ESI MRM Frag=180.0V [email protected] (329.1 -> 243.1) 071619_1010.d 6.140 326.79 6 References 5

3 (1) T. Eom et al., Food Addit Contam A, 34(11), 2011–2022 (2017).

2 1 (2) F D. Foster, J. R. Stuff, L. A. Vernarelli, and J. A. Whitecavage, Gerstel 0 x101 Aﬂatoxin G2: +ESI MRM Frag=190.0V [email protected] (331.1 -> 313.1) 071619_1010.d 5.796 1.4 66.47 application note 207. http://www.Gerstel.com/pdf/AppNote-207.pdf

1.2 1

0.8 0.6

0.4

0.2 0 4 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8 8.1 8.2 8.3 8.4 8.5 8.6 8.7

Figure 2: Stacked view of mass chromatograms resulting from extraction of aflatoxin-spiked sunflower oil. oil samples using the automated procedure. Replicate extracts of olive oil spiked with mycotoxins were analyzed by LC–MS resulting in an average precision of 4.3 % (range: 2.4 %–6.9 %RSD). The recovery of mycotoxins using the automated extraction workflow was assessed by comparing results obtained from mycotoxin-spiked edible oil samples with those resulting from spiking the extracts of Gerstel, Inc. blank edible oil samples after the extraction. In Table II, the resulting % 701 Digital Drive, Suite J, Linthicum, MD 21090 recovery for each mycotoxin from each examined edible oil is shown. tel. (800) 413-8160, mail [email protected] US: www.Gerstelus.com; International: www.gerstel.com

THE APPLICATION NOTEBOOK – MARCH 2020 35 Day in … Day out …

Extraction, derivatization, addition of standards Something you can rely on

Solid Phase Extraction (SPE), Filtration The GERSTEL MPS handles Your Sample Preparation and Introduction efficiently and reliably. Our solutions Evaporative concentration are intelligently automated to your specifications. (mVAP) No programming, just Setup and Start by mouse-click. Agitation, quickMIX Your MPS works day and night, using less solvent and without anyone watching over it.

Centrifuge GERSTEL Solutions for GC/MS and LC/MS with Application Support at your Service. MAESTRO PrepAhead Productivity What can we do for you?

[email protected] (800) 413-8160 · [email protected] www.gerstel.com www.gerstelus.com